x25519 1.0.6 → 1.0.7
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGES.md +7 -0
- data/Gemfile +1 -0
- data/LICENSE +499 -160
- data/README.md +34 -4
- data/ext/x25519_precomputed/fp25519_x64.c +843 -740
- data/ext/x25519_precomputed/fp25519_x64.h +64 -45
- data/ext/x25519_precomputed/table_ladder_x25519.h +514 -263
- data/ext/x25519_precomputed/x25519_precomputed.h +6 -1
- data/ext/x25519_precomputed/x25519_x64.c +210 -208
- data/lib/x25519/version.rb +1 -1
- data/x25519.gemspec +1 -1
- metadata +3 -3
data/README.md
CHANGED
|
@@ -7,9 +7,9 @@
|
|
|
7
7
|
[appveyor-image]: https://ci.appveyor.com/api/projects/status/4s05bcae0mow85v1?svg=true
|
|
8
8
|
[appveyor-link]: https://ci.appveyor.com/project/tarcieri/x25519
|
|
9
9
|
[docs-image]: https://img.shields.io/badge/yard-docs-blue.svg
|
|
10
|
-
[docs-link]: http://www.rubydoc.info/gems/x25519/1.0.
|
|
11
|
-
[license-image]: https://img.shields.io/badge/License-LGPL%
|
|
12
|
-
[license-link]: https://www.gnu.org/licenses/lgpl-
|
|
10
|
+
[docs-link]: http://www.rubydoc.info/gems/x25519/1.0.6
|
|
11
|
+
[license-image]: https://img.shields.io/badge/License-LGPL%20v2.1-blue.svg
|
|
12
|
+
[license-link]: https://www.gnu.org/licenses/lgpl-2.1
|
|
13
13
|
[gitter-image]: https://badges.gitter.im/badge.svg
|
|
14
14
|
[gitter-link]: https://gitter.im/crypto-rb/Lobby
|
|
15
15
|
|
|
@@ -32,6 +32,36 @@ Curve25519.
|
|
|
32
32
|
[rfc7748_precomputed]: https://github.com/armfazh/rfc7748_precomputed
|
|
33
33
|
[ed25519 gem]: https://github.com/crypto-rb/ed25519
|
|
34
34
|
|
|
35
|
+
### Is it any good?
|
|
36
|
+
|
|
37
|
+
[Yes.](http://news.ycombinator.com/item?id=3067434)
|
|
38
|
+
|
|
39
|
+
### What is it useful for?
|
|
40
|
+
|
|
41
|
+
X25519 is a key exchange/agreement algorithm generally used as a low-level
|
|
42
|
+
building block in cryptographic protocols.
|
|
43
|
+
|
|
44
|
+
### Can I use X25519 to encrypt things?
|
|
45
|
+
|
|
46
|
+
Please use [RbNaCl::Box] or the (experimental) [XSTREAM] library if you would
|
|
47
|
+
like to use X25519 for public-key encryption. Otherwise, the X25519 algorithm
|
|
48
|
+
is not directly useful for encryption without a higher-level encryption protocol
|
|
49
|
+
built on top of it.
|
|
50
|
+
|
|
51
|
+
[RbNaCl::Box]: https://github.com/crypto-rb/rbnacl/wiki/Public-Key-Encryption
|
|
52
|
+
[XSTREAM]: https://github.com/miscreant/xstream
|
|
53
|
+
|
|
54
|
+
## Help and Discussion
|
|
55
|
+
|
|
56
|
+
Have questions? Want to suggest a feature or change? Join a discussion group:
|
|
57
|
+
|
|
58
|
+
* [Crypto.rb Gitter]: web-based chat about Ruby crypto projects including **x25519**.
|
|
59
|
+
* [Crypto.rb Google Group]: join via web or email ([crypto-rb+subscribe@googlegroups.com])
|
|
60
|
+
|
|
61
|
+
[Crypto.rb Gitter]: https://gitter.im/crypto-rb/Lobby
|
|
62
|
+
[Crypto.rb Google Group]: https://groups.google.com/forum/#!forum/crypto-rb
|
|
63
|
+
[crypto-rb+subscribe@googlegroups.com]: mailto:crypto-rb+subscribe@googlegroups.com?subject=subscribe
|
|
64
|
+
|
|
35
65
|
## Requirements
|
|
36
66
|
|
|
37
67
|
**x25519.rb** is supported on and tested against the following platforms:
|
|
@@ -260,7 +290,7 @@ The optimized [rfc7748_precomputed] implementation was designed by:
|
|
|
260
290
|
Copyright (c) 2017-2018 Armando Faz, Tony Arcieri
|
|
261
291
|
|
|
262
292
|
This gem is available as open source under the terms of the
|
|
263
|
-
GNU Lesser General Public License
|
|
293
|
+
GNU Lesser General Public License v2.1 ([LICENSE](https://www.gnu.org/licenses/lgpl-2.1.txt))
|
|
264
294
|
|
|
265
295
|
## Code of Conduct
|
|
266
296
|
|
|
@@ -1,13 +1,16 @@
|
|
|
1
1
|
/**
|
|
2
|
-
* Copyright (c) 2017 Armando Faz <armfazh@ic.unicamp.br>.
|
|
2
|
+
* Copyright (c) 2017 Armando Faz <armfazh@ic.unicamp.br>. All Rights Reserved.
|
|
3
3
|
* Institute of Computing.
|
|
4
4
|
* University of Campinas, Brazil.
|
|
5
5
|
*
|
|
6
|
-
*
|
|
7
|
-
*
|
|
8
|
-
*
|
|
6
|
+
* Copyright (C) 2018 Jason A. Donenfeld <Jason@zx2c4.com>. All Rights Reserved.
|
|
7
|
+
* Copyright (C) 2018 Samuel Neves <sneves@dei.uc.pt>. All Rights Reserved.
|
|
8
|
+
*
|
|
9
|
+
* This program is free software: you can redistribute it and/or modify
|
|
10
|
+
* it under the terms of the GNU Lesser General Public License as
|
|
11
|
+
* published by the Free Software Foundation, version 2 or greater.
|
|
9
12
|
*
|
|
10
|
-
* This program is distributed in the hope that it will be useful, but
|
|
13
|
+
* This program is distributed in the hope that it will be useful, but
|
|
11
14
|
* WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
12
15
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
13
16
|
* Lesser General Public License for more details.
|
|
@@ -17,197 +20,175 @@
|
|
|
17
20
|
*/
|
|
18
21
|
#include "fp25519_x64.h"
|
|
19
22
|
|
|
20
|
-
int compare_bytes(uint8_t* A, uint8_t* B,unsigned int num_bytes)
|
|
21
|
-
{
|
|
22
|
-
unsigned int i=0;
|
|
23
|
-
uint8_t ret=0;
|
|
24
|
-
for(i=0;i<num_bytes;i++)
|
|
25
|
-
{
|
|
26
|
-
ret += A[i]^B[i];
|
|
27
|
-
}
|
|
28
|
-
return ret;
|
|
29
|
-
}
|
|
30
|
-
|
|
31
|
-
int compare_EltFp25519_1w_x64(uint64_t *A, uint64_t *B)
|
|
32
|
-
{
|
|
33
|
-
return compare_bytes((uint8_t*)A,(uint8_t*)B,SIZE_ELEMENT_BYTES);
|
|
34
|
-
}
|
|
35
|
-
|
|
36
23
|
/**
|
|
37
24
|
*
|
|
38
|
-
* @param c Two 512-bit products:
|
|
39
|
-
* @param a Two 256-bit integers:
|
|
40
|
-
* @param b Two 256-bit integers:
|
|
25
|
+
* @param c Two 512-bit products: c0[0:7]=a0[0:3]*b0[0:3] and c1[8:15]=a1[4:7]*b1[4:7]
|
|
26
|
+
* @param a Two 256-bit integers: a0[0:3] and a1[4:7]
|
|
27
|
+
* @param b Two 256-bit integers: b0[0:3] and b1[4:7]
|
|
41
28
|
*/
|
|
42
|
-
void mul2_256x256_integer_x64(uint64_t *const c, uint64_t *const a,
|
|
43
|
-
{
|
|
29
|
+
void mul2_256x256_integer_x64(uint64_t *const c, uint64_t *const a,
|
|
30
|
+
uint64_t *const b) {
|
|
44
31
|
#ifdef __BMI2__
|
|
45
32
|
#ifdef __ADX__
|
|
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
|
-
|
|
106
|
-
|
|
107
|
-
|
|
33
|
+
__asm__ __volatile__(
|
|
34
|
+
"xorl %%r14d, %%r14d ;"
|
|
35
|
+
"movq (%1), %%rdx; " /* A[0] */
|
|
36
|
+
"mulx (%2), %%r8, %%r12; " /* A[0]*B[0] */ "xorl %%r10d, %%r10d ;" "movq %%r8, (%0) ;"
|
|
37
|
+
"mulx 8(%2), %%r10, %%rax; " /* A[0]*B[1] */ "adox %%r10, %%r12 ;"
|
|
38
|
+
"mulx 16(%2), %%r8, %%rbx; " /* A[0]*B[2] */ "adox %%r8, %%rax ;"
|
|
39
|
+
"mulx 24(%2), %%r10, %%rcx; " /* A[0]*B[3] */ "adox %%r10, %%rbx ;"
|
|
40
|
+
/*******************************************/ "adox %%r14, %%rcx ;"
|
|
41
|
+
|
|
42
|
+
"movq 8(%1), %%rdx; " /* A[1] */
|
|
43
|
+
"mulx (%2), %%r8, %%r9; " /* A[1]*B[0] */ "adox %%r12, %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
44
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[1]*B[1] */ "adox %%r10, %%r9 ;" "adcx %%r9, %%rax ;"
|
|
45
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[1]*B[2] */ "adox %%r8, %%r11 ;" "adcx %%r11, %%rbx ;"
|
|
46
|
+
"mulx 24(%2), %%r10, %%r12; " /* A[1]*B[3] */ "adox %%r10, %%r13 ;" "adcx %%r13, %%rcx ;"
|
|
47
|
+
/*******************************************/ "adox %%r14, %%r12 ;" "adcx %%r14, %%r12 ;"
|
|
48
|
+
|
|
49
|
+
"movq 16(%1), %%rdx; " /* A[2] */ "xorl %%r10d, %%r10d ;"
|
|
50
|
+
"mulx (%2), %%r8, %%r9; " /* A[2]*B[0] */ "adox %%rax, %%r8 ;" "movq %%r8, 16(%0) ;"
|
|
51
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[2]*B[1] */ "adox %%r10, %%r9 ;" "adcx %%r9, %%rbx ;"
|
|
52
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[2]*B[2] */ "adox %%r8, %%r11 ;" "adcx %%r11, %%rcx ;"
|
|
53
|
+
"mulx 24(%2), %%r10, %%rax; " /* A[2]*B[3] */ "adox %%r10, %%r13 ;" "adcx %%r13, %%r12 ;"
|
|
54
|
+
/*******************************************/ "adox %%r14, %%rax ;" "adcx %%r14, %%rax ;"
|
|
55
|
+
|
|
56
|
+
"movq 24(%1), %%rdx; " /* A[3] */ "xorl %%r10d, %%r10d ;"
|
|
57
|
+
"mulx (%2), %%r8, %%r9; " /* A[3]*B[0] */ "adox %%rbx, %%r8 ;" "movq %%r8, 24(%0) ;"
|
|
58
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[3]*B[1] */ "adox %%r10, %%r9 ;" "adcx %%r9, %%rcx ;" "movq %%rcx, 32(%0) ;"
|
|
59
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[3]*B[2] */ "adox %%r8, %%r11 ;" "adcx %%r11, %%r12 ;" "movq %%r12, 40(%0) ;"
|
|
60
|
+
"mulx 24(%2), %%r10, %%rbx; " /* A[3]*B[3] */ "adox %%r10, %%r13 ;" "adcx %%r13, %%rax ;" "movq %%rax, 48(%0) ;"
|
|
61
|
+
/*******************************************/ "adox %%r14, %%rbx ;" "adcx %%r14, %%rbx ;" "movq %%rbx, 56(%0) ;"
|
|
62
|
+
|
|
63
|
+
"movq 32(%1), %%rdx; " /* C[0] */
|
|
64
|
+
"mulx 32(%2), %%r8, %%r12; " /* C[0]*D[0] */ "xorl %%r10d, %%r10d ;" "movq %%r8, 64(%0);"
|
|
65
|
+
"mulx 40(%2), %%r10, %%rax; " /* C[0]*D[1] */ "adox %%r10, %%r12 ;"
|
|
66
|
+
"mulx 48(%2), %%r8, %%rbx; " /* C[0]*D[2] */ "adox %%r8, %%rax ;"
|
|
67
|
+
"mulx 56(%2), %%r10, %%rcx; " /* C[0]*D[3] */ "adox %%r10, %%rbx ;"
|
|
68
|
+
/*******************************************/ "adox %%r14, %%rcx ;"
|
|
69
|
+
|
|
70
|
+
"movq 40(%1), %%rdx; " /* C[1] */ "xorl %%r10d, %%r10d ;"
|
|
71
|
+
"mulx 32(%2), %%r8, %%r9; " /* C[1]*D[0] */ "adox %%r12, %%r8 ;" "movq %%r8, 72(%0);"
|
|
72
|
+
"mulx 40(%2), %%r10, %%r11; " /* C[1]*D[1] */ "adox %%r10, %%r9 ;" "adcx %%r9, %%rax ;"
|
|
73
|
+
"mulx 48(%2), %%r8, %%r13; " /* C[1]*D[2] */ "adox %%r8, %%r11 ;" "adcx %%r11, %%rbx ;"
|
|
74
|
+
"mulx 56(%2), %%r10, %%r12; " /* C[1]*D[3] */ "adox %%r10, %%r13 ;" "adcx %%r13, %%rcx ;"
|
|
75
|
+
/*******************************************/ "adox %%r14, %%r12 ;" "adcx %%r14, %%r12 ;"
|
|
76
|
+
|
|
77
|
+
"movq 48(%1), %%rdx; " /* C[2] */ "xorl %%r10d, %%r10d ;"
|
|
78
|
+
"mulx 32(%2), %%r8, %%r9; " /* C[2]*D[0] */ "adox %%rax, %%r8 ;" "movq %%r8, 80(%0);"
|
|
79
|
+
"mulx 40(%2), %%r10, %%r11; " /* C[2]*D[1] */ "adox %%r10, %%r9 ;" "adcx %%r9, %%rbx ;"
|
|
80
|
+
"mulx 48(%2), %%r8, %%r13; " /* C[2]*D[2] */ "adox %%r8, %%r11 ;" "adcx %%r11, %%rcx ;"
|
|
81
|
+
"mulx 56(%2), %%r10, %%rax; " /* C[2]*D[3] */ "adox %%r10, %%r13 ;" "adcx %%r13, %%r12 ;"
|
|
82
|
+
/*******************************************/ "adox %%r14, %%rax ;" "adcx %%r14, %%rax ;"
|
|
83
|
+
|
|
84
|
+
"movq 56(%1), %%rdx; " /* C[3] */ "xorl %%r10d, %%r10d ;"
|
|
85
|
+
"mulx 32(%2), %%r8, %%r9; " /* C[3]*D[0] */ "adox %%rbx, %%r8 ;" "movq %%r8, 88(%0);"
|
|
86
|
+
"mulx 40(%2), %%r10, %%r11; " /* C[3]*D[1] */ "adox %%r10, %%r9 ;" "adcx %%r9, %%rcx ;" "movq %%rcx, 96(%0) ;"
|
|
87
|
+
"mulx 48(%2), %%r8, %%r13; " /* C[3]*D[2] */ "adox %%r8, %%r11 ;" "adcx %%r11, %%r12 ;" "movq %%r12, 104(%0) ;"
|
|
88
|
+
"mulx 56(%2), %%r10, %%rbx; " /* C[3]*D[3] */ "adox %%r10, %%r13 ;" "adcx %%r13, %%rax ;" "movq %%rax, 112(%0) ;"
|
|
89
|
+
/*******************************************/ "adox %%r14, %%rbx ;" "adcx %%r14, %%rbx ;" "movq %%rbx, 120(%0) ;"
|
|
90
|
+
:
|
|
91
|
+
: "r" (c), "r" (a), "r" (b)
|
|
92
|
+
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx",
|
|
93
|
+
"%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14"
|
|
94
|
+
);
|
|
108
95
|
#else
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
|
|
112
|
-
|
|
113
|
-
|
|
114
|
-
|
|
115
|
-
|
|
116
|
-
|
|
117
|
-
|
|
118
|
-
|
|
119
|
-
|
|
120
|
-
|
|
121
|
-
|
|
122
|
-
|
|
123
|
-
|
|
124
|
-
|
|
125
|
-
|
|
126
|
-
|
|
127
|
-
|
|
128
|
-
|
|
129
|
-
|
|
130
|
-
|
|
131
|
-
|
|
132
|
-
|
|
133
|
-
|
|
134
|
-
|
|
135
|
-
|
|
136
|
-
|
|
137
|
-
|
|
138
|
-
|
|
139
|
-
|
|
140
|
-
|
|
141
|
-
|
|
142
|
-
|
|
143
|
-
|
|
144
|
-
|
|
145
|
-
|
|
146
|
-
|
|
147
|
-
|
|
148
|
-
|
|
149
|
-
|
|
150
|
-
|
|
151
|
-
|
|
152
|
-
|
|
153
|
-
|
|
154
|
-
|
|
155
|
-
|
|
156
|
-
|
|
157
|
-
|
|
158
|
-
|
|
159
|
-
|
|
160
|
-
|
|
161
|
-
|
|
162
|
-
|
|
163
|
-
|
|
164
|
-
|
|
165
|
-
|
|
166
|
-
|
|
167
|
-
|
|
168
|
-
|
|
169
|
-
|
|
170
|
-
|
|
171
|
-
|
|
172
|
-
|
|
173
|
-
|
|
174
|
-
|
|
175
|
-
|
|
176
|
-
|
|
177
|
-
|
|
178
|
-
|
|
179
|
-
|
|
180
|
-
|
|
181
|
-
|
|
182
|
-
|
|
183
|
-
|
|
184
|
-
|
|
185
|
-
|
|
186
|
-
|
|
187
|
-
|
|
188
|
-
|
|
189
|
-
|
|
190
|
-
|
|
191
|
-
|
|
192
|
-
|
|
193
|
-
|
|
194
|
-
|
|
195
|
-
|
|
196
|
-
|
|
197
|
-
|
|
198
|
-
|
|
199
|
-
|
|
200
|
-
"adcq %%rdx, %%rcx \n\t" "movq %%rcx, 120(%0) \n\t"
|
|
201
|
-
:
|
|
202
|
-
: "r" (c), "r" (a), "r" (b)
|
|
203
|
-
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx", "%r8",
|
|
204
|
-
"%r9", "%r10", "%r11", "%r12", "%r13", "%r14"
|
|
205
|
-
);
|
|
96
|
+
__asm__ __volatile__(
|
|
97
|
+
"movq (%1), %%rdx; " /* A[0] */
|
|
98
|
+
"mulx (%2), %%r8, %%r12; " /* A[0]*B[0] */ "movq %%r8, (%0) ;"
|
|
99
|
+
"mulx 8(%2), %%r10, %%rax; " /* A[0]*B[1] */ "addq %%r10, %%r12 ;"
|
|
100
|
+
"mulx 16(%2), %%r8, %%rbx; " /* A[0]*B[2] */ "adcq %%r8, %%rax ;"
|
|
101
|
+
"mulx 24(%2), %%r10, %%rcx; " /* A[0]*B[3] */ "adcq %%r10, %%rbx ;"
|
|
102
|
+
/*******************************************/ "adcq $0, %%rcx ;"
|
|
103
|
+
|
|
104
|
+
"movq 8(%1), %%rdx; " /* A[1] */
|
|
105
|
+
"mulx (%2), %%r8, %%r9; " /* A[1]*B[0] */ "addq %%r12, %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
106
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[1]*B[1] */ "adcq %%r10, %%r9 ;"
|
|
107
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[1]*B[2] */ "adcq %%r8, %%r11 ;"
|
|
108
|
+
"mulx 24(%2), %%r10, %%r12; " /* A[1]*B[3] */ "adcq %%r10, %%r13 ;"
|
|
109
|
+
/*******************************************/ "adcq $0, %%r12 ;"
|
|
110
|
+
|
|
111
|
+
"addq %%r9, %%rax ;"
|
|
112
|
+
"adcq %%r11, %%rbx ;"
|
|
113
|
+
"adcq %%r13, %%rcx ;"
|
|
114
|
+
"adcq $0, %%r12 ;"
|
|
115
|
+
|
|
116
|
+
"movq 16(%1), %%rdx; " /* A[2] */
|
|
117
|
+
"mulx (%2), %%r8, %%r9; " /* A[2]*B[0] */ "addq %%rax, %%r8 ;" "movq %%r8, 16(%0) ;"
|
|
118
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[2]*B[1] */ "adcq %%r10, %%r9 ;"
|
|
119
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[2]*B[2] */ "adcq %%r8, %%r11 ;"
|
|
120
|
+
"mulx 24(%2), %%r10, %%rax; " /* A[2]*B[3] */ "adcq %%r10, %%r13 ;"
|
|
121
|
+
/*******************************************/ "adcq $0, %%rax ;"
|
|
122
|
+
|
|
123
|
+
"addq %%r9, %%rbx ;"
|
|
124
|
+
"adcq %%r11, %%rcx ;"
|
|
125
|
+
"adcq %%r13, %%r12 ;"
|
|
126
|
+
"adcq $0, %%rax ;"
|
|
127
|
+
|
|
128
|
+
"movq 24(%1), %%rdx; " /* A[3] */
|
|
129
|
+
"mulx (%2), %%r8, %%r9; " /* A[3]*B[0] */ "addq %%rbx, %%r8 ;" "movq %%r8, 24(%0) ;"
|
|
130
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[3]*B[1] */ "adcq %%r10, %%r9 ;"
|
|
131
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[3]*B[2] */ "adcq %%r8, %%r11 ;"
|
|
132
|
+
"mulx 24(%2), %%r10, %%rbx; " /* A[3]*B[3] */ "adcq %%r10, %%r13 ;"
|
|
133
|
+
/*******************************************/ "adcq $0, %%rbx ;"
|
|
134
|
+
|
|
135
|
+
"addq %%r9, %%rcx ;" "movq %%rcx, 32(%0) ;"
|
|
136
|
+
"adcq %%r11, %%r12 ;" "movq %%r12, 40(%0) ;"
|
|
137
|
+
"adcq %%r13, %%rax ;" "movq %%rax, 48(%0) ;"
|
|
138
|
+
"adcq $0, %%rbx ;" "movq %%rbx, 56(%0) ;"
|
|
139
|
+
|
|
140
|
+
"movq 32(%1), %%rdx; " /* C[0] */
|
|
141
|
+
"mulx 32(%2), %%r8, %%r12; " /* C[0]*D[0] */ "movq %%r8, 64(%0) ;"
|
|
142
|
+
"mulx 40(%2), %%r10, %%rax; " /* C[0]*D[1] */ "addq %%r10, %%r12 ;"
|
|
143
|
+
"mulx 48(%2), %%r8, %%rbx; " /* C[0]*D[2] */ "adcq %%r8, %%rax ;"
|
|
144
|
+
"mulx 56(%2), %%r10, %%rcx; " /* C[0]*D[3] */ "adcq %%r10, %%rbx ;"
|
|
145
|
+
/*******************************************/ "adcq $0, %%rcx ;"
|
|
146
|
+
|
|
147
|
+
"movq 40(%1), %%rdx; " /* C[1] */
|
|
148
|
+
"mulx 32(%2), %%r8, %%r9; " /* C[1]*D[0] */ "addq %%r12, %%r8 ;" "movq %%r8, 72(%0) ;"
|
|
149
|
+
"mulx 40(%2), %%r10, %%r11; " /* C[1]*D[1] */ "adcq %%r10, %%r9 ;"
|
|
150
|
+
"mulx 48(%2), %%r8, %%r13; " /* C[1]*D[2] */ "adcq %%r8, %%r11 ;"
|
|
151
|
+
"mulx 56(%2), %%r10, %%r12; " /* C[1]*D[3] */ "adcq %%r10, %%r13 ;"
|
|
152
|
+
/*******************************************/ "adcq $0, %%r12 ;"
|
|
153
|
+
|
|
154
|
+
"addq %%r9, %%rax ;"
|
|
155
|
+
"adcq %%r11, %%rbx ;"
|
|
156
|
+
"adcq %%r13, %%rcx ;"
|
|
157
|
+
"adcq $0, %%r12 ;"
|
|
158
|
+
|
|
159
|
+
"movq 48(%1), %%rdx; " /* C[2] */
|
|
160
|
+
"mulx 32(%2), %%r8, %%r9; " /* C[2]*D[0] */ "addq %%rax, %%r8 ;" "movq %%r8, 80(%0) ;"
|
|
161
|
+
"mulx 40(%2), %%r10, %%r11; " /* C[2]*D[1] */ "adcq %%r10, %%r9 ;"
|
|
162
|
+
"mulx 48(%2), %%r8, %%r13; " /* C[2]*D[2] */ "adcq %%r8, %%r11 ;"
|
|
163
|
+
"mulx 56(%2), %%r10, %%rax; " /* C[2]*D[3] */ "adcq %%r10, %%r13 ;"
|
|
164
|
+
/*******************************************/ "adcq $0, %%rax ;"
|
|
165
|
+
|
|
166
|
+
"addq %%r9, %%rbx ;"
|
|
167
|
+
"adcq %%r11, %%rcx ;"
|
|
168
|
+
"adcq %%r13, %%r12 ;"
|
|
169
|
+
"adcq $0, %%rax ;"
|
|
170
|
+
|
|
171
|
+
"movq 56(%1), %%rdx; " /* C[3] */
|
|
172
|
+
"mulx 32(%2), %%r8, %%r9; " /* C[3]*D[0] */ "addq %%rbx, %%r8 ;" "movq %%r8, 88(%0) ;"
|
|
173
|
+
"mulx 40(%2), %%r10, %%r11; " /* C[3]*D[1] */ "adcq %%r10, %%r9 ;"
|
|
174
|
+
"mulx 48(%2), %%r8, %%r13; " /* C[3]*D[2] */ "adcq %%r8, %%r11 ;"
|
|
175
|
+
"mulx 56(%2), %%r10, %%rbx; " /* C[3]*D[3] */ "adcq %%r10, %%r13 ;"
|
|
176
|
+
/*******************************************/ "adcq $0, %%rbx ;"
|
|
177
|
+
|
|
178
|
+
"addq %%r9, %%rcx ;" "movq %%rcx, 96(%0) ;"
|
|
179
|
+
"adcq %%r11, %%r12 ;" "movq %%r12, 104(%0) ;"
|
|
180
|
+
"adcq %%r13, %%rax ;" "movq %%rax, 112(%0) ;"
|
|
181
|
+
"adcq $0, %%rbx ;" "movq %%rbx, 120(%0) ;"
|
|
182
|
+
:
|
|
183
|
+
: "r" (c), "r" (a), "r" (b)
|
|
184
|
+
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx",
|
|
185
|
+
"%r8", "%r9", "%r10", "%r11", "%r12", "%r13"
|
|
186
|
+
);
|
|
206
187
|
#endif
|
|
207
188
|
#else /* Without BMI2 */
|
|
208
|
-
|
|
209
|
-
|
|
210
|
-
|
|
189
|
+
/**
|
|
190
|
+
* TODO: Multiplications using MULQ instruction.
|
|
191
|
+
**/
|
|
211
192
|
#endif
|
|
212
193
|
}
|
|
213
194
|
|
|
@@ -216,140 +197,186 @@ void mul2_256x256_integer_x64(uint64_t *const c, uint64_t *const a, uint64_t *co
|
|
|
216
197
|
* @param c
|
|
217
198
|
* @param a
|
|
218
199
|
*/
|
|
219
|
-
void sqr2_256x256_integer_x64(uint64_t *const c, uint64_t *const a)
|
|
220
|
-
{
|
|
200
|
+
void sqr2_256x256_integer_x64(uint64_t *const c, uint64_t *const a) {
|
|
221
201
|
#ifdef __BMI2__
|
|
222
|
-
|
|
223
|
-
|
|
224
|
-
|
|
225
|
-
|
|
226
|
-
|
|
227
|
-
|
|
228
|
-
|
|
229
|
-
|
|
230
|
-
|
|
231
|
-
|
|
232
|
-
|
|
233
|
-
|
|
234
|
-
|
|
235
|
-
|
|
236
|
-
|
|
237
|
-
|
|
238
|
-
|
|
239
|
-
|
|
240
|
-
|
|
241
|
-
|
|
242
|
-
|
|
243
|
-
|
|
244
|
-
|
|
245
|
-
|
|
246
|
-
|
|
247
|
-
|
|
248
|
-
|
|
249
|
-
|
|
250
|
-
|
|
251
|
-
|
|
252
|
-
|
|
253
|
-
|
|
254
|
-
|
|
255
|
-
|
|
256
|
-
|
|
257
|
-
|
|
258
|
-
|
|
259
|
-
|
|
260
|
-
|
|
261
|
-
|
|
262
|
-
|
|
263
|
-
|
|
264
|
-
|
|
265
|
-
|
|
266
|
-
|
|
267
|
-
|
|
268
|
-
|
|
269
|
-
|
|
270
|
-
|
|
271
|
-
|
|
272
|
-
|
|
273
|
-
|
|
274
|
-
|
|
275
|
-
|
|
276
|
-
|
|
277
|
-
|
|
278
|
-
|
|
279
|
-
|
|
280
|
-
|
|
281
|
-
|
|
282
|
-
|
|
283
|
-
|
|
284
|
-
|
|
285
|
-
|
|
286
|
-
|
|
287
|
-
|
|
288
|
-
|
|
289
|
-
|
|
290
|
-
|
|
291
|
-
|
|
292
|
-
|
|
293
|
-
|
|
294
|
-
|
|
295
|
-
|
|
296
|
-
|
|
297
|
-
|
|
298
|
-
|
|
299
|
-
|
|
300
|
-
|
|
301
|
-
|
|
302
|
-
|
|
303
|
-
|
|
304
|
-
|
|
305
|
-
|
|
306
|
-
|
|
307
|
-
|
|
308
|
-
|
|
309
|
-
|
|
310
|
-
|
|
311
|
-
|
|
312
|
-
|
|
313
|
-
|
|
314
|
-
|
|
315
|
-
|
|
316
|
-
|
|
317
|
-
|
|
318
|
-
|
|
319
|
-
|
|
320
|
-
|
|
321
|
-
|
|
322
|
-
|
|
323
|
-
|
|
324
|
-
|
|
325
|
-
|
|
326
|
-
|
|
327
|
-
|
|
328
|
-
|
|
329
|
-
|
|
330
|
-
|
|
331
|
-
|
|
332
|
-
|
|
333
|
-
|
|
334
|
-
|
|
335
|
-
|
|
336
|
-
|
|
337
|
-
|
|
338
|
-
|
|
339
|
-
|
|
340
|
-
|
|
341
|
-
|
|
342
|
-
|
|
343
|
-
|
|
344
|
-
|
|
345
|
-
|
|
346
|
-
|
|
347
|
-
|
|
348
|
-
|
|
202
|
+
#ifdef __ADX__
|
|
203
|
+
__asm__ __volatile__(
|
|
204
|
+
"movq (%1), %%rdx ;" /* A[0] */
|
|
205
|
+
"mulx 8(%1), %%r8, %%r14 ;" /* A[1]*A[0] */ "xorl %%r15d, %%r15d;"
|
|
206
|
+
"mulx 16(%1), %%r9, %%r10 ;" /* A[2]*A[0] */ "adcx %%r14, %%r9 ;"
|
|
207
|
+
"mulx 24(%1), %%rax, %%rcx ;" /* A[3]*A[0] */ "adcx %%rax, %%r10 ;"
|
|
208
|
+
"movq 24(%1), %%rdx ;" /* A[3] */
|
|
209
|
+
"mulx 8(%1), %%r11, %%r12 ;" /* A[1]*A[3] */ "adcx %%rcx, %%r11 ;"
|
|
210
|
+
"mulx 16(%1), %%rax, %%r13 ;" /* A[2]*A[3] */ "adcx %%rax, %%r12 ;"
|
|
211
|
+
"movq 8(%1), %%rdx ;" /* A[1] */ "adcx %%r15, %%r13 ;"
|
|
212
|
+
"mulx 16(%1), %%rax, %%rcx ;" /* A[2]*A[1] */ "movq $0, %%r14 ;"
|
|
213
|
+
/*******************************************/ "adcx %%r15, %%r14 ;"
|
|
214
|
+
|
|
215
|
+
"xorl %%r15d, %%r15d;"
|
|
216
|
+
"adox %%rax, %%r10 ;" "adcx %%r8, %%r8 ;"
|
|
217
|
+
"adox %%rcx, %%r11 ;" "adcx %%r9, %%r9 ;"
|
|
218
|
+
"adox %%r15, %%r12 ;" "adcx %%r10, %%r10 ;"
|
|
219
|
+
"adox %%r15, %%r13 ;" "adcx %%r11, %%r11 ;"
|
|
220
|
+
"adox %%r15, %%r14 ;" "adcx %%r12, %%r12 ;"
|
|
221
|
+
"adcx %%r13, %%r13 ;"
|
|
222
|
+
"adcx %%r14, %%r14 ;"
|
|
223
|
+
|
|
224
|
+
"movq (%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[0]^2 */
|
|
225
|
+
/********************/ "movq %%rax, 0(%0) ;"
|
|
226
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
227
|
+
"movq 8(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[1]^2 */
|
|
228
|
+
"adcq %%rax, %%r9 ;" "movq %%r9, 16(%0) ;"
|
|
229
|
+
"adcq %%rcx, %%r10 ;" "movq %%r10, 24(%0) ;"
|
|
230
|
+
"movq 16(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[2]^2 */
|
|
231
|
+
"adcq %%rax, %%r11 ;" "movq %%r11, 32(%0) ;"
|
|
232
|
+
"adcq %%rcx, %%r12 ;" "movq %%r12, 40(%0) ;"
|
|
233
|
+
"movq 24(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[3]^2 */
|
|
234
|
+
"adcq %%rax, %%r13 ;" "movq %%r13, 48(%0) ;"
|
|
235
|
+
"adcq %%rcx, %%r14 ;" "movq %%r14, 56(%0) ;"
|
|
236
|
+
|
|
237
|
+
|
|
238
|
+
"movq 32(%1), %%rdx ;" /* B[0] */
|
|
239
|
+
"mulx 40(%1), %%r8, %%r14 ;" /* B[1]*B[0] */ "xorl %%r15d, %%r15d;"
|
|
240
|
+
"mulx 48(%1), %%r9, %%r10 ;" /* B[2]*B[0] */ "adcx %%r14, %%r9 ;"
|
|
241
|
+
"mulx 56(%1), %%rax, %%rcx ;" /* B[3]*B[0] */ "adcx %%rax, %%r10 ;"
|
|
242
|
+
"movq 56(%1), %%rdx ;" /* B[3] */
|
|
243
|
+
"mulx 40(%1), %%r11, %%r12 ;" /* B[1]*B[3] */ "adcx %%rcx, %%r11 ;"
|
|
244
|
+
"mulx 48(%1), %%rax, %%r13 ;" /* B[2]*B[3] */ "adcx %%rax, %%r12 ;"
|
|
245
|
+
"movq 40(%1), %%rdx ;" /* B[1] */ "adcx %%r15, %%r13 ;"
|
|
246
|
+
"mulx 48(%1), %%rax, %%rcx ;" /* B[2]*B[1] */ "movq $0, %%r14 ;"
|
|
247
|
+
/*******************************************/ "adcx %%r15, %%r14 ;"
|
|
248
|
+
|
|
249
|
+
"xorl %%r15d, %%r15d;"
|
|
250
|
+
"adox %%rax, %%r10 ;" "adcx %%r8, %%r8 ;"
|
|
251
|
+
"adox %%rcx, %%r11 ;" "adcx %%r9, %%r9 ;"
|
|
252
|
+
"adox %%r15, %%r12 ;" "adcx %%r10, %%r10 ;"
|
|
253
|
+
"adox %%r15, %%r13 ;" "adcx %%r11, %%r11 ;"
|
|
254
|
+
"adox %%r15, %%r14 ;" "adcx %%r12, %%r12 ;"
|
|
255
|
+
"adcx %%r13, %%r13 ;"
|
|
256
|
+
"adcx %%r14, %%r14 ;"
|
|
257
|
+
|
|
258
|
+
"movq 32(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* B[0]^2 */
|
|
259
|
+
/********************/ "movq %%rax, 64(%0) ;"
|
|
260
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 72(%0) ;"
|
|
261
|
+
"movq 40(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* B[1]^2 */
|
|
262
|
+
"adcq %%rax, %%r9 ;" "movq %%r9, 80(%0) ;"
|
|
263
|
+
"adcq %%rcx, %%r10 ;" "movq %%r10, 88(%0) ;"
|
|
264
|
+
"movq 48(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* B[2]^2 */
|
|
265
|
+
"adcq %%rax, %%r11 ;" "movq %%r11, 96(%0) ;"
|
|
266
|
+
"adcq %%rcx, %%r12 ;" "movq %%r12, 104(%0) ;"
|
|
267
|
+
"movq 56(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* B[3]^2 */
|
|
268
|
+
"adcq %%rax, %%r13 ;" "movq %%r13, 112(%0) ;"
|
|
269
|
+
"adcq %%rcx, %%r14 ;" "movq %%r14, 120(%0) ;"
|
|
270
|
+
:
|
|
271
|
+
: "r" (c), "r" (a)
|
|
272
|
+
: "memory", "cc", "%rax", "%rcx", "%rdx",
|
|
273
|
+
"%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14", "%r15"
|
|
274
|
+
);
|
|
275
|
+
#else /* Without ADX */
|
|
276
|
+
__asm__ __volatile__(
|
|
277
|
+
"movq 8(%1), %%rdx ;" /* A[1] */
|
|
278
|
+
"mulx (%1), %%r8, %%r9 ;" /* A[0]*A[1] */
|
|
279
|
+
"mulx 16(%1), %%r10, %%r11 ;" /* A[2]*A[1] */
|
|
280
|
+
"mulx 24(%1), %%rcx, %%r14 ;" /* A[3]*A[1] */
|
|
281
|
+
|
|
282
|
+
"movq 16(%1), %%rdx ;" /* A[2] */
|
|
283
|
+
"mulx 24(%1), %%r12, %%r13 ;" /* A[3]*A[2] */
|
|
284
|
+
"mulx (%1), %%rax, %%rdx ;" /* A[0]*A[2] */
|
|
285
|
+
|
|
286
|
+
"addq %%rax, %%r9 ;"
|
|
287
|
+
"adcq %%rdx, %%r10 ;"
|
|
288
|
+
"adcq %%rcx, %%r11 ;"
|
|
289
|
+
"adcq %%r14, %%r12 ;"
|
|
290
|
+
"adcq $0, %%r13 ;"
|
|
291
|
+
"movq $0, %%r14 ;"
|
|
292
|
+
"adcq $0, %%r14 ;"
|
|
293
|
+
|
|
294
|
+
"movq (%1), %%rdx ;" /* A[0] */
|
|
295
|
+
"mulx 24(%1), %%rax, %%rcx ;" /* A[0]*A[3] */
|
|
296
|
+
|
|
297
|
+
"addq %%rax, %%r10 ;"
|
|
298
|
+
"adcq %%rcx, %%r11 ;"
|
|
299
|
+
"adcq $0, %%r12 ;"
|
|
300
|
+
"adcq $0, %%r13 ;"
|
|
301
|
+
"adcq $0, %%r14 ;"
|
|
302
|
+
|
|
303
|
+
"shldq $1, %%r13, %%r14 ;"
|
|
304
|
+
"shldq $1, %%r12, %%r13 ;"
|
|
305
|
+
"shldq $1, %%r11, %%r12 ;"
|
|
306
|
+
"shldq $1, %%r10, %%r11 ;"
|
|
307
|
+
"shldq $1, %%r9, %%r10 ;"
|
|
308
|
+
"shldq $1, %%r8, %%r9 ;"
|
|
309
|
+
"shlq $1, %%r8 ;"
|
|
310
|
+
|
|
311
|
+
/********************/ "mulx %%rdx, %%rax, %%rcx ; " /* A[0]^2 */
|
|
312
|
+
/********************/ "movq %%rax, 0(%0) ;"
|
|
313
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
314
|
+
"movq 8(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ; " /* A[1]^2 */
|
|
315
|
+
"adcq %%rax, %%r9 ;" "movq %%r9, 16(%0) ;"
|
|
316
|
+
"adcq %%rcx, %%r10 ;" "movq %%r10, 24(%0) ;"
|
|
317
|
+
"movq 16(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ; " /* A[2]^2 */
|
|
318
|
+
"adcq %%rax, %%r11 ;" "movq %%r11, 32(%0) ;"
|
|
319
|
+
"adcq %%rcx, %%r12 ;" "movq %%r12, 40(%0) ;"
|
|
320
|
+
"movq 24(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ; " /* A[3]^2 */
|
|
321
|
+
"adcq %%rax, %%r13 ;" "movq %%r13, 48(%0) ;"
|
|
322
|
+
"adcq %%rcx, %%r14 ;" "movq %%r14, 56(%0) ;"
|
|
323
|
+
|
|
324
|
+
"movq 40(%1), %%rdx ;" /* B[1] */
|
|
325
|
+
"mulx 32(%1), %%r8, %%r9 ;" /* B[0]*B[1] */
|
|
326
|
+
"mulx 48(%1), %%r10, %%r11 ;" /* B[2]*B[1] */
|
|
327
|
+
"mulx 56(%1), %%rcx, %%r14 ;" /* B[3]*B[1] */
|
|
328
|
+
|
|
329
|
+
"movq 48(%1), %%rdx ;" /* B[2] */
|
|
330
|
+
"mulx 56(%1), %%r12, %%r13 ;" /* B[3]*B[2] */
|
|
331
|
+
"mulx 32(%1), %%rax, %%rdx ;" /* B[0]*B[2] */
|
|
332
|
+
|
|
333
|
+
"addq %%rax, %%r9 ;"
|
|
334
|
+
"adcq %%rdx, %%r10 ;"
|
|
335
|
+
"adcq %%rcx, %%r11 ;"
|
|
336
|
+
"adcq %%r14, %%r12 ;"
|
|
337
|
+
"adcq $0, %%r13 ;"
|
|
338
|
+
"movq $0, %%r14 ;"
|
|
339
|
+
"adcq $0, %%r14 ;"
|
|
340
|
+
|
|
341
|
+
"movq 32(%1), %%rdx ;" /* B[0] */
|
|
342
|
+
"mulx 56(%1), %%rax, %%rcx ;" /* B[0]*B[3] */
|
|
343
|
+
|
|
344
|
+
"addq %%rax, %%r10 ;"
|
|
345
|
+
"adcq %%rcx, %%r11 ;"
|
|
346
|
+
"adcq $0, %%r12 ;"
|
|
347
|
+
"adcq $0, %%r13 ;"
|
|
348
|
+
"adcq $0, %%r14 ;"
|
|
349
|
+
|
|
350
|
+
"shldq $1, %%r13, %%r14 ;"
|
|
351
|
+
"shldq $1, %%r12, %%r13 ;"
|
|
352
|
+
"shldq $1, %%r11, %%r12 ;"
|
|
353
|
+
"shldq $1, %%r10, %%r11 ;"
|
|
354
|
+
"shldq $1, %%r9, %%r10 ;"
|
|
355
|
+
"shldq $1, %%r8, %%r9 ;"
|
|
356
|
+
"shlq $1, %%r8 ;"
|
|
357
|
+
|
|
358
|
+
/********************/ "mulx %%rdx, %%rax, %%rcx ; " /* B[0]^2 */
|
|
359
|
+
/********************/ "movq %%rax, 64(%0) ;"
|
|
360
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 72(%0) ;"
|
|
361
|
+
"movq 40(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ; " /* B[1]^2 */
|
|
362
|
+
"adcq %%rax, %%r9 ;" "movq %%r9, 80(%0) ;"
|
|
363
|
+
"adcq %%rcx, %%r10 ;" "movq %%r10, 88(%0) ;"
|
|
364
|
+
"movq 48(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ; " /* B[2]^2 */
|
|
365
|
+
"adcq %%rax, %%r11 ;" "movq %%r11, 96(%0) ;"
|
|
366
|
+
"adcq %%rcx, %%r12 ;" "movq %%r12, 104(%0) ;"
|
|
367
|
+
"movq 56(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ; " /* B[3]^2 */
|
|
368
|
+
"adcq %%rax, %%r13 ;" "movq %%r13, 112(%0) ;"
|
|
369
|
+
"adcq %%rcx, %%r14 ;" "movq %%r14, 120(%0) ;"
|
|
370
|
+
:
|
|
371
|
+
: "r" (c), "r" (a)
|
|
372
|
+
: "memory", "cc", "%rax", "%rcx", "%rdx",
|
|
373
|
+
"%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14"
|
|
374
|
+
);
|
|
375
|
+
#endif
|
|
349
376
|
#else /* Without BMI2 */
|
|
350
|
-
|
|
351
|
-
|
|
352
|
-
|
|
377
|
+
/**
|
|
378
|
+
* TODO: Multiplications using MULQ instruction.
|
|
379
|
+
**/
|
|
353
380
|
#endif
|
|
354
381
|
}
|
|
355
382
|
|
|
@@ -358,467 +385,543 @@ void sqr2_256x256_integer_x64(uint64_t *const c, uint64_t *const a)
|
|
|
358
385
|
* @param c
|
|
359
386
|
* @param a
|
|
360
387
|
*/
|
|
361
|
-
void red_EltFp25519_2w_x64(uint64_t *const c, uint64_t *const a)
|
|
362
|
-
{
|
|
388
|
+
void red_EltFp25519_2w_x64(uint64_t *const c, uint64_t *const a) {
|
|
363
389
|
#ifdef __BMI2__
|
|
364
390
|
#ifdef __ADX__
|
|
365
|
-
|
|
366
|
-
|
|
367
|
-
|
|
368
|
-
|
|
369
|
-
|
|
370
|
-
|
|
371
|
-
|
|
372
|
-
|
|
373
|
-
|
|
374
|
-
|
|
375
|
-
|
|
376
|
-
|
|
377
|
-
|
|
378
|
-
|
|
379
|
-
|
|
380
|
-
|
|
381
|
-
|
|
382
|
-
|
|
383
|
-
|
|
384
|
-
|
|
385
|
-
|
|
386
|
-
|
|
387
|
-
|
|
391
|
+
__asm__ __volatile__(
|
|
392
|
+
"movl $38, %%edx; " /* 2*c = 38 = 2^256 */
|
|
393
|
+
"mulx 32(%1), %%r8, %%r10; " /* c*C[4] */ "xorl %%ebx, %%ebx ;" "adox (%1), %%r8 ;"
|
|
394
|
+
"mulx 40(%1), %%r9, %%r11; " /* c*C[5] */ "adcx %%r10, %%r9 ;" "adox 8(%1), %%r9 ;"
|
|
395
|
+
"mulx 48(%1), %%r10, %%rax; " /* c*C[6] */ "adcx %%r11, %%r10 ;" "adox 16(%1), %%r10 ;"
|
|
396
|
+
"mulx 56(%1), %%r11, %%rcx; " /* c*C[7] */ "adcx %%rax, %%r11 ;" "adox 24(%1), %%r11 ;"
|
|
397
|
+
/****************************************/ "adcx %%rbx, %%rcx ;" "adox %%rbx, %%rcx ;"
|
|
398
|
+
"clc ;"
|
|
399
|
+
"mulx %%rcx, %%rax, %%rcx ; " /* c*C[4] */
|
|
400
|
+
"adcx %%rax, %%r8 ;"
|
|
401
|
+
"adcx %%rcx, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
402
|
+
"adcx %%rbx, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
403
|
+
"adcx %%rbx, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
404
|
+
"mov $0, %%ecx ;"
|
|
405
|
+
"cmovc %%edx, %%ecx ;"
|
|
406
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, (%0) ;"
|
|
407
|
+
|
|
408
|
+
"mulx 96(%1), %%r8, %%r10; " /* c*C[4] */ "xorl %%ebx, %%ebx ;" "adox 64(%1), %%r8 ;"
|
|
409
|
+
"mulx 104(%1), %%r9, %%r11; " /* c*C[5] */ "adcx %%r10, %%r9 ;" "adox 72(%1), %%r9 ;"
|
|
410
|
+
"mulx 112(%1), %%r10, %%rax; " /* c*C[6] */ "adcx %%r11, %%r10 ;" "adox 80(%1), %%r10 ;"
|
|
411
|
+
"mulx 120(%1), %%r11, %%rcx; " /* c*C[7] */ "adcx %%rax, %%r11 ;" "adox 88(%1), %%r11 ;"
|
|
412
|
+
/*****************************************/ "adcx %%rbx, %%rcx ;" "adox %%rbx, %%rcx ;"
|
|
413
|
+
"clc ;"
|
|
414
|
+
"mulx %%rcx, %%rax, %%rcx ; " /* c*C[4] */
|
|
415
|
+
"adcx %%rax, %%r8 ;"
|
|
416
|
+
"adcx %%rcx, %%r9 ;" "movq %%r9, 40(%0) ;"
|
|
417
|
+
"adcx %%rbx, %%r10 ;" "movq %%r10, 48(%0) ;"
|
|
418
|
+
"adcx %%rbx, %%r11 ;" "movq %%r11, 56(%0) ;"
|
|
419
|
+
"mov $0, %%ecx ;"
|
|
420
|
+
"cmovc %%edx, %%ecx ;"
|
|
421
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 32(%0) ;"
|
|
422
|
+
:
|
|
423
|
+
: "r" (c), "r" (a)
|
|
424
|
+
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11"
|
|
425
|
+
);
|
|
388
426
|
#else
|
|
389
|
-
|
|
390
|
-
|
|
391
|
-
|
|
392
|
-
|
|
393
|
-
|
|
394
|
-
|
|
395
|
-
|
|
396
|
-
|
|
397
|
-
|
|
398
|
-
|
|
399
|
-
|
|
400
|
-
|
|
401
|
-
|
|
402
|
-
|
|
403
|
-
|
|
404
|
-
|
|
405
|
-
|
|
406
|
-
|
|
407
|
-
|
|
408
|
-
|
|
409
|
-
|
|
410
|
-
|
|
411
|
-
|
|
412
|
-
|
|
413
|
-
|
|
414
|
-
|
|
415
|
-
|
|
416
|
-
|
|
417
|
-
|
|
418
|
-
|
|
419
|
-
|
|
420
|
-
|
|
421
|
-
|
|
422
|
-
|
|
423
|
-
|
|
424
|
-
|
|
425
|
-
|
|
426
|
-
|
|
427
|
+
__asm__ __volatile__(
|
|
428
|
+
"movl $38, %%edx ; " /* 2*c = 38 = 2^256 */
|
|
429
|
+
"mulx 32(%1), %%r8, %%r10 ;" /* c*C[4] */
|
|
430
|
+
"mulx 40(%1), %%r9, %%r11 ;" /* c*C[5] */ "addq %%r10, %%r9 ;"
|
|
431
|
+
"mulx 48(%1), %%r10, %%rax ;" /* c*C[6] */ "adcq %%r11, %%r10 ;"
|
|
432
|
+
"mulx 56(%1), %%r11, %%rcx ;" /* c*C[7] */ "adcq %%rax, %%r11 ;"
|
|
433
|
+
/****************************************/ "adcq $0, %%rcx ;"
|
|
434
|
+
"addq (%1), %%r8 ;"
|
|
435
|
+
"adcq 8(%1), %%r9 ;"
|
|
436
|
+
"adcq 16(%1), %%r10 ;"
|
|
437
|
+
"adcq 24(%1), %%r11 ;"
|
|
438
|
+
"adcq $0, %%rcx ;"
|
|
439
|
+
"mulx %%rcx, %%rax, %%rcx ;" /* c*C[4] */
|
|
440
|
+
"addq %%rax, %%r8 ;"
|
|
441
|
+
"adcq %%rcx, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
442
|
+
"adcq $0, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
443
|
+
"adcq $0, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
444
|
+
"mov $0, %%ecx ;"
|
|
445
|
+
"cmovc %%edx, %%ecx ;"
|
|
446
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, (%0) ;"
|
|
447
|
+
|
|
448
|
+
"mulx 96(%1), %%r8, %%r10 ;" /* c*C[4] */
|
|
449
|
+
"mulx 104(%1), %%r9, %%r11 ;" /* c*C[5] */ "addq %%r10, %%r9 ;"
|
|
450
|
+
"mulx 112(%1), %%r10, %%rax ;" /* c*C[6] */ "adcq %%r11, %%r10 ;"
|
|
451
|
+
"mulx 120(%1), %%r11, %%rcx ;" /* c*C[7] */ "adcq %%rax, %%r11 ;"
|
|
452
|
+
/*****************************************/ "adcq $0, %%rcx ;"
|
|
453
|
+
"addq 64(%1), %%r8 ;"
|
|
454
|
+
"adcq 72(%1), %%r9 ;"
|
|
455
|
+
"adcq 80(%1), %%r10 ;"
|
|
456
|
+
"adcq 88(%1), %%r11 ;"
|
|
457
|
+
"adcq $0, %%rcx ;"
|
|
458
|
+
"mulx %%rcx, %%rax, %%rcx ;" /* c*C[4] */
|
|
459
|
+
"addq %%rax, %%r8 ;"
|
|
460
|
+
"adcq %%rcx, %%r9 ;" "movq %%r9, 40(%0) ;"
|
|
461
|
+
"adcq $0, %%r10 ;" "movq %%r10, 48(%0) ;"
|
|
462
|
+
"adcq $0, %%r11 ;" "movq %%r11, 56(%0) ;"
|
|
463
|
+
"mov $0, %%ecx ;"
|
|
464
|
+
"cmovc %%edx, %%ecx ;"
|
|
465
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 32(%0) ;"
|
|
466
|
+
:
|
|
467
|
+
: "r" (c), "r" (a)
|
|
468
|
+
: "memory", "cc", "%rax", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11"
|
|
469
|
+
);
|
|
427
470
|
#endif
|
|
428
471
|
#else /* Without BMI2 */
|
|
429
|
-
|
|
472
|
+
/* [TODO] */
|
|
430
473
|
#endif
|
|
431
474
|
}
|
|
432
475
|
|
|
433
|
-
void mul_256x256_integer_x64(uint64_t *const c, uint64_t *const a, uint64_t *const b)
|
|
434
|
-
{
|
|
476
|
+
void mul_256x256_integer_x64(uint64_t *const c, uint64_t *const a, uint64_t *const b) {
|
|
435
477
|
#ifdef __BMI2__
|
|
436
478
|
#ifdef __ADX__
|
|
437
|
-
|
|
438
|
-
|
|
439
|
-
|
|
440
|
-
|
|
441
|
-
|
|
442
|
-
|
|
443
|
-
|
|
444
|
-
|
|
445
|
-
|
|
446
|
-
|
|
447
|
-
|
|
448
|
-
|
|
449
|
-
|
|
450
|
-
|
|
451
|
-
|
|
452
|
-
|
|
453
|
-
|
|
454
|
-
|
|
455
|
-
|
|
456
|
-
|
|
457
|
-
|
|
458
|
-
|
|
459
|
-
|
|
460
|
-
|
|
461
|
-
|
|
462
|
-
|
|
463
|
-
|
|
464
|
-
|
|
465
|
-
|
|
466
|
-
|
|
467
|
-
|
|
468
|
-
|
|
469
|
-
|
|
470
|
-
);
|
|
479
|
+
__asm__ __volatile__(
|
|
480
|
+
"movq (%1), %%rdx; " /* A[0] */
|
|
481
|
+
"mulx (%2), %%r8, %%r9; " /* A[0]*B[0] */ "xorl %%r10d, %%r10d ;" "movq %%r8, (%0) ;"
|
|
482
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[0]*B[1] */ "adox %%r9, %%r10 ;" "movq %%r10, 8(%0) ;"
|
|
483
|
+
"mulx 16(%2), %%r12, %%r13; " /* A[0]*B[2] */ "adox %%r11, %%r12 ;"
|
|
484
|
+
"mulx 24(%2), %%r14, %%rdx; " /* A[0]*B[3] */ "adox %%r13, %%r14 ;" "movq $0, %%rax ;"
|
|
485
|
+
/*******************************************/ "adox %%rdx, %%rax ;"
|
|
486
|
+
|
|
487
|
+
"movq 8(%1), %%rdx; " /* A[1] */
|
|
488
|
+
"mulx (%2), %%r8, %%r9; " /* A[1]*B[0] */ "xorl %%r10d, %%r10d ;" "adcx 8(%0), %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
489
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[1]*B[1] */ "adox %%r9, %%r10 ;" "adcx %%r12, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
490
|
+
"mulx 16(%2), %%r12, %%r13; " /* A[1]*B[2] */ "adox %%r11, %%r12 ;" "adcx %%r14, %%r12 ;" "movq $0, %%r8 ;"
|
|
491
|
+
"mulx 24(%2), %%r14, %%rdx; " /* A[1]*B[3] */ "adox %%r13, %%r14 ;" "adcx %%rax, %%r14 ;" "movq $0, %%rax ;"
|
|
492
|
+
/*******************************************/ "adox %%rdx, %%rax ;" "adcx %%r8, %%rax ;"
|
|
493
|
+
|
|
494
|
+
"movq 16(%1), %%rdx; " /* A[2] */
|
|
495
|
+
"mulx (%2), %%r8, %%r9; " /* A[2]*B[0] */ "xorl %%r10d, %%r10d ;" "adcx 16(%0), %%r8 ;" "movq %%r8, 16(%0) ;"
|
|
496
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[2]*B[1] */ "adox %%r9, %%r10 ;" "adcx %%r12, %%r10 ;" "movq %%r10, 24(%0) ;"
|
|
497
|
+
"mulx 16(%2), %%r12, %%r13; " /* A[2]*B[2] */ "adox %%r11, %%r12 ;" "adcx %%r14, %%r12 ;" "movq $0, %%r8 ;"
|
|
498
|
+
"mulx 24(%2), %%r14, %%rdx; " /* A[2]*B[3] */ "adox %%r13, %%r14 ;" "adcx %%rax, %%r14 ;" "movq $0, %%rax ;"
|
|
499
|
+
/*******************************************/ "adox %%rdx, %%rax ;" "adcx %%r8, %%rax ;"
|
|
500
|
+
|
|
501
|
+
"movq 24(%1), %%rdx; " /* A[3] */
|
|
502
|
+
"mulx (%2), %%r8, %%r9; " /* A[3]*B[0] */ "xorl %%r10d, %%r10d ;" "adcx 24(%0), %%r8 ;" "movq %%r8, 24(%0) ;"
|
|
503
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[3]*B[1] */ "adox %%r9, %%r10 ;" "adcx %%r12, %%r10 ;" "movq %%r10, 32(%0) ;"
|
|
504
|
+
"mulx 16(%2), %%r12, %%r13; " /* A[3]*B[2] */ "adox %%r11, %%r12 ;" "adcx %%r14, %%r12 ;" "movq %%r12, 40(%0) ;" "movq $0, %%r8 ;"
|
|
505
|
+
"mulx 24(%2), %%r14, %%rdx; " /* A[3]*B[3] */ "adox %%r13, %%r14 ;" "adcx %%rax, %%r14 ;" "movq %%r14, 48(%0) ;" "movq $0, %%rax ;"
|
|
506
|
+
/*******************************************/ "adox %%rdx, %%rax ;" "adcx %%r8, %%rax ;" "movq %%rax, 56(%0) ;"
|
|
507
|
+
:
|
|
508
|
+
: "r" (c), "r" (a), "r" (b)
|
|
509
|
+
: "memory", "cc", "%rax", "%rdx", "%r8",
|
|
510
|
+
"%r9", "%r10", "%r11", "%r12", "%r13", "%r14"
|
|
511
|
+
);
|
|
471
512
|
#else
|
|
472
|
-
|
|
473
|
-
|
|
474
|
-
|
|
475
|
-
|
|
476
|
-
|
|
477
|
-
|
|
478
|
-
|
|
479
|
-
|
|
480
|
-
|
|
481
|
-
|
|
482
|
-
|
|
483
|
-
|
|
484
|
-
|
|
485
|
-
|
|
486
|
-
|
|
487
|
-
|
|
488
|
-
|
|
489
|
-
|
|
490
|
-
|
|
491
|
-
|
|
492
|
-
|
|
493
|
-
|
|
494
|
-
|
|
495
|
-
|
|
496
|
-
|
|
497
|
-
|
|
498
|
-
|
|
499
|
-
|
|
500
|
-
|
|
501
|
-
|
|
502
|
-
|
|
503
|
-
|
|
504
|
-
|
|
505
|
-
|
|
506
|
-
|
|
507
|
-
|
|
508
|
-
|
|
509
|
-
|
|
510
|
-
|
|
511
|
-
|
|
512
|
-
|
|
513
|
-
|
|
514
|
-
|
|
515
|
-
|
|
516
|
-
|
|
517
|
-
|
|
518
|
-
|
|
519
|
-
|
|
520
|
-
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx", "%r8",
|
|
521
|
-
"%r9", "%r10", "%r11", "%r12", "%r13", "%r14"
|
|
522
|
-
);
|
|
513
|
+
__asm__ __volatile__(
|
|
514
|
+
"movq (%1), %%rdx; " /* A[0] */
|
|
515
|
+
"mulx (%2), %%r8, %%r12; " /* A[0]*B[0] */ "movq %%r8, (%0) ;"
|
|
516
|
+
"mulx 8(%2), %%r10, %%rax; " /* A[0]*B[1] */ "addq %%r10, %%r12 ;"
|
|
517
|
+
"mulx 16(%2), %%r8, %%rbx; " /* A[0]*B[2] */ "adcq %%r8, %%rax ;"
|
|
518
|
+
"mulx 24(%2), %%r10, %%rcx; " /* A[0]*B[3] */ "adcq %%r10, %%rbx ;"
|
|
519
|
+
/*******************************************/ "adcq $0, %%rcx ;"
|
|
520
|
+
|
|
521
|
+
"movq 8(%1), %%rdx; " /* A[1] */
|
|
522
|
+
"mulx (%2), %%r8, %%r9; " /* A[1]*B[0] */ "addq %%r12, %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
523
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[1]*B[1] */ "adcq %%r10, %%r9 ;"
|
|
524
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[1]*B[2] */ "adcq %%r8, %%r11 ;"
|
|
525
|
+
"mulx 24(%2), %%r10, %%r12; " /* A[1]*B[3] */ "adcq %%r10, %%r13 ;"
|
|
526
|
+
/*******************************************/ "adcq $0, %%r12 ;"
|
|
527
|
+
|
|
528
|
+
"addq %%r9, %%rax ;"
|
|
529
|
+
"adcq %%r11, %%rbx ;"
|
|
530
|
+
"adcq %%r13, %%rcx ;"
|
|
531
|
+
"adcq $0, %%r12 ;"
|
|
532
|
+
|
|
533
|
+
"movq 16(%1), %%rdx; " /* A[2] */
|
|
534
|
+
"mulx (%2), %%r8, %%r9; " /* A[2]*B[0] */ "addq %%rax, %%r8 ;" "movq %%r8, 16(%0) ;"
|
|
535
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[2]*B[1] */ "adcq %%r10, %%r9 ;"
|
|
536
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[2]*B[2] */ "adcq %%r8, %%r11 ;"
|
|
537
|
+
"mulx 24(%2), %%r10, %%rax; " /* A[2]*B[3] */ "adcq %%r10, %%r13 ;"
|
|
538
|
+
/*******************************************/ "adcq $0, %%rax ;"
|
|
539
|
+
|
|
540
|
+
"addq %%r9, %%rbx ;"
|
|
541
|
+
"adcq %%r11, %%rcx ;"
|
|
542
|
+
"adcq %%r13, %%r12 ;"
|
|
543
|
+
"adcq $0, %%rax ;"
|
|
544
|
+
|
|
545
|
+
"movq 24(%1), %%rdx; " /* A[3] */
|
|
546
|
+
"mulx (%2), %%r8, %%r9; " /* A[3]*B[0] */ "addq %%rbx, %%r8 ;" "movq %%r8, 24(%0) ;"
|
|
547
|
+
"mulx 8(%2), %%r10, %%r11; " /* A[3]*B[1] */ "adcq %%r10, %%r9 ;"
|
|
548
|
+
"mulx 16(%2), %%r8, %%r13; " /* A[3]*B[2] */ "adcq %%r8, %%r11 ;"
|
|
549
|
+
"mulx 24(%2), %%r10, %%rbx; " /* A[3]*B[3] */ "adcq %%r10, %%r13 ;"
|
|
550
|
+
/*******************************************/ "adcq $0, %%rbx ;"
|
|
551
|
+
|
|
552
|
+
"addq %%r9, %%rcx ;" "movq %%rcx, 32(%0) ;"
|
|
553
|
+
"adcq %%r11, %%r12 ;" "movq %%r12, 40(%0) ;"
|
|
554
|
+
"adcq %%r13, %%rax ;" "movq %%rax, 48(%0) ;"
|
|
555
|
+
"adcq $0, %%rbx ;" "movq %%rbx, 56(%0) ;"
|
|
556
|
+
:
|
|
557
|
+
: "r" (c), "r" (a), "r" (b)
|
|
558
|
+
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx",
|
|
559
|
+
"%r8", "%r9", "%r10", "%r11", "%r12", "%r13"
|
|
560
|
+
);
|
|
523
561
|
#endif
|
|
524
562
|
#else /* Without BMI2 */
|
|
525
|
-
|
|
526
|
-
|
|
527
|
-
|
|
563
|
+
/**
|
|
564
|
+
* TODO: Multiplications using MULQ instruction.
|
|
565
|
+
**/
|
|
528
566
|
#endif
|
|
529
567
|
}
|
|
530
568
|
|
|
531
|
-
void sqr_256x256_integer_x64(uint64_t *const c, uint64_t *const a)
|
|
532
|
-
{
|
|
569
|
+
void sqr_256x256_integer_x64(uint64_t *const c, uint64_t *const a) {
|
|
533
570
|
#ifdef __BMI2__
|
|
534
|
-
|
|
535
|
-
|
|
536
|
-
|
|
537
|
-
|
|
538
|
-
|
|
539
|
-
|
|
540
|
-
|
|
541
|
-
|
|
542
|
-
|
|
543
|
-
|
|
544
|
-
|
|
545
|
-
|
|
546
|
-
|
|
547
|
-
|
|
548
|
-
|
|
549
|
-
|
|
550
|
-
|
|
551
|
-
|
|
552
|
-
|
|
553
|
-
|
|
554
|
-
|
|
555
|
-
|
|
556
|
-
|
|
557
|
-
|
|
558
|
-
|
|
559
|
-
|
|
560
|
-
|
|
561
|
-
|
|
562
|
-
|
|
563
|
-
|
|
564
|
-
|
|
565
|
-
|
|
566
|
-
|
|
567
|
-
|
|
568
|
-
|
|
569
|
-
|
|
570
|
-
|
|
571
|
-
|
|
572
|
-
|
|
573
|
-
|
|
574
|
-
|
|
575
|
-
|
|
576
|
-
|
|
577
|
-
|
|
578
|
-
|
|
579
|
-
|
|
580
|
-
|
|
581
|
-
|
|
582
|
-
|
|
583
|
-
|
|
584
|
-
|
|
585
|
-
|
|
586
|
-
|
|
587
|
-
|
|
588
|
-
|
|
589
|
-
|
|
590
|
-
|
|
591
|
-
|
|
592
|
-
|
|
593
|
-
|
|
594
|
-
|
|
595
|
-
|
|
596
|
-
|
|
597
|
-
|
|
598
|
-
|
|
599
|
-
|
|
571
|
+
#ifdef __ADX__
|
|
572
|
+
__asm__ __volatile__(
|
|
573
|
+
"movq (%1), %%rdx ;" /* A[0] */
|
|
574
|
+
"mulx 8(%1), %%r8, %%r14 ;" /* A[1]*A[0] */ "xorl %%r15d, %%r15d;"
|
|
575
|
+
"mulx 16(%1), %%r9, %%r10 ;" /* A[2]*A[0] */ "adcx %%r14, %%r9 ;"
|
|
576
|
+
"mulx 24(%1), %%rax, %%rcx ;" /* A[3]*A[0] */ "adcx %%rax, %%r10 ;"
|
|
577
|
+
"movq 24(%1), %%rdx ;" /* A[3] */
|
|
578
|
+
"mulx 8(%1), %%r11, %%r12 ;" /* A[1]*A[3] */ "adcx %%rcx, %%r11 ;"
|
|
579
|
+
"mulx 16(%1), %%rax, %%r13 ;" /* A[2]*A[3] */ "adcx %%rax, %%r12 ;"
|
|
580
|
+
"movq 8(%1), %%rdx ;" /* A[1] */ "adcx %%r15, %%r13 ;"
|
|
581
|
+
"mulx 16(%1), %%rax, %%rcx ;" /* A[2]*A[1] */ "movq $0, %%r14 ;"
|
|
582
|
+
/*******************************************/ "adcx %%r15, %%r14 ;"
|
|
583
|
+
|
|
584
|
+
"xorl %%r15d, %%r15d;"
|
|
585
|
+
"adox %%rax, %%r10 ;" "adcx %%r8, %%r8 ;"
|
|
586
|
+
"adox %%rcx, %%r11 ;" "adcx %%r9, %%r9 ;"
|
|
587
|
+
"adox %%r15, %%r12 ;" "adcx %%r10, %%r10 ;"
|
|
588
|
+
"adox %%r15, %%r13 ;" "adcx %%r11, %%r11 ;"
|
|
589
|
+
"adox %%r15, %%r14 ;" "adcx %%r12, %%r12 ;"
|
|
590
|
+
"adcx %%r13, %%r13 ;"
|
|
591
|
+
"adcx %%r14, %%r14 ;"
|
|
592
|
+
|
|
593
|
+
"movq (%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[0]^2 */
|
|
594
|
+
/********************/ "movq %%rax, 0(%0) ;"
|
|
595
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
596
|
+
"movq 8(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[1]^2 */
|
|
597
|
+
"adcq %%rax, %%r9 ;" "movq %%r9, 16(%0) ;"
|
|
598
|
+
"adcq %%rcx, %%r10 ;" "movq %%r10, 24(%0) ;"
|
|
599
|
+
"movq 16(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[2]^2 */
|
|
600
|
+
"adcq %%rax, %%r11 ;" "movq %%r11, 32(%0) ;"
|
|
601
|
+
"adcq %%rcx, %%r12 ;" "movq %%r12, 40(%0) ;"
|
|
602
|
+
"movq 24(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[3]^2 */
|
|
603
|
+
"adcq %%rax, %%r13 ;" "movq %%r13, 48(%0) ;"
|
|
604
|
+
"adcq %%rcx, %%r14 ;" "movq %%r14, 56(%0) ;"
|
|
605
|
+
:
|
|
606
|
+
: "r" (c), "r" (a)
|
|
607
|
+
: "memory", "cc", "%rax", "%rcx", "%rdx",
|
|
608
|
+
"%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14", "%r15"
|
|
609
|
+
);
|
|
610
|
+
#else /* Without ADX */
|
|
611
|
+
__asm__ __volatile__(
|
|
612
|
+
"movq 8(%1), %%rdx ;" /* A[1] */
|
|
613
|
+
"mulx (%1), %%r8, %%r9 ;" /* A[0]*A[1] */
|
|
614
|
+
"mulx 16(%1), %%r10, %%r11 ;" /* A[2]*A[1] */
|
|
615
|
+
"mulx 24(%1), %%rcx, %%r14 ;" /* A[3]*A[1] */
|
|
616
|
+
|
|
617
|
+
"movq 16(%1), %%rdx ;" /* A[2] */
|
|
618
|
+
"mulx 24(%1), %%r12, %%r13 ;" /* A[3]*A[2] */
|
|
619
|
+
"mulx (%1), %%rax, %%rdx ;" /* A[0]*A[2] */
|
|
620
|
+
|
|
621
|
+
"addq %%rax, %%r9 ;"
|
|
622
|
+
"adcq %%rdx, %%r10 ;"
|
|
623
|
+
"adcq %%rcx, %%r11 ;"
|
|
624
|
+
"adcq %%r14, %%r12 ;"
|
|
625
|
+
"adcq $0, %%r13 ;"
|
|
626
|
+
"movq $0, %%r14 ;"
|
|
627
|
+
"adcq $0, %%r14 ;"
|
|
628
|
+
|
|
629
|
+
"movq (%1), %%rdx ;" /* A[0] */
|
|
630
|
+
"mulx 24(%1), %%rax, %%rcx ;" /* A[0]*A[3] */
|
|
631
|
+
|
|
632
|
+
"addq %%rax, %%r10 ;"
|
|
633
|
+
"adcq %%rcx, %%r11 ;"
|
|
634
|
+
"adcq $0, %%r12 ;"
|
|
635
|
+
"adcq $0, %%r13 ;"
|
|
636
|
+
"adcq $0, %%r14 ;"
|
|
637
|
+
|
|
638
|
+
"shldq $1, %%r13, %%r14 ;"
|
|
639
|
+
"shldq $1, %%r12, %%r13 ;"
|
|
640
|
+
"shldq $1, %%r11, %%r12 ;"
|
|
641
|
+
"shldq $1, %%r10, %%r11 ;"
|
|
642
|
+
"shldq $1, %%r9, %%r10 ;"
|
|
643
|
+
"shldq $1, %%r8, %%r9 ;"
|
|
644
|
+
"shlq $1, %%r8 ;"
|
|
645
|
+
|
|
646
|
+
/********************/ "mulx %%rdx, %%rax, %%rcx ;" /* A[0]^2 */
|
|
647
|
+
/********************/ "movq %%rax, 0(%0) ;"
|
|
648
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, 8(%0) ;"
|
|
649
|
+
"movq 8(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[1]^2 */
|
|
650
|
+
"adcq %%rax, %%r9 ;" "movq %%r9, 16(%0) ;"
|
|
651
|
+
"adcq %%rcx, %%r10 ;" "movq %%r10, 24(%0) ;"
|
|
652
|
+
"movq 16(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[2]^2 */
|
|
653
|
+
"adcq %%rax, %%r11 ;" "movq %%r11, 32(%0) ;"
|
|
654
|
+
"adcq %%rcx, %%r12 ;" "movq %%r12, 40(%0) ;"
|
|
655
|
+
"movq 24(%1), %%rdx ;" "mulx %%rdx, %%rax, %%rcx ;" /* A[3]^2 */
|
|
656
|
+
"adcq %%rax, %%r13 ;" "movq %%r13, 48(%0) ;"
|
|
657
|
+
"adcq %%rcx, %%r14 ;" "movq %%r14, 56(%0) ;"
|
|
658
|
+
:
|
|
659
|
+
: "r" (c), "r" (a)
|
|
660
|
+
: "memory", "cc", "%rax", "%rcx", "%rdx",
|
|
661
|
+
"%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14"
|
|
662
|
+
);
|
|
663
|
+
#endif
|
|
600
664
|
#else /* Without BMI2 */
|
|
601
|
-
|
|
602
|
-
|
|
603
|
-
|
|
665
|
+
/**
|
|
666
|
+
* TODO: Multiplications using MULQ instruction.
|
|
667
|
+
**/
|
|
604
668
|
#endif
|
|
605
669
|
}
|
|
606
670
|
|
|
607
|
-
void red_EltFp25519_1w_x64(uint64_t *const c, uint64_t *const a)
|
|
608
|
-
{
|
|
671
|
+
void red_EltFp25519_1w_x64(uint64_t *const c, uint64_t *const a) {
|
|
609
672
|
#ifdef __BMI2__
|
|
610
673
|
#ifdef __ADX__
|
|
611
|
-
|
|
612
|
-
|
|
613
|
-
|
|
614
|
-
|
|
615
|
-
|
|
616
|
-
|
|
617
|
-
|
|
618
|
-
|
|
619
|
-
|
|
620
|
-
|
|
621
|
-
|
|
622
|
-
|
|
623
|
-
|
|
624
|
-
|
|
674
|
+
__asm__ __volatile__(
|
|
675
|
+
"movl $38, %%edx ;" /* 2*c = 38 = 2^256 */
|
|
676
|
+
"mulx 32(%1), %%r8, %%r10 ;" /* c*C[4] */ "xorl %%ebx, %%ebx ;" "adox (%1), %%r8 ;"
|
|
677
|
+
"mulx 40(%1), %%r9, %%r11 ;" /* c*C[5] */ "adcx %%r10, %%r9 ;" "adox 8(%1), %%r9 ;"
|
|
678
|
+
"mulx 48(%1), %%r10, %%rax ;" /* c*C[6] */ "adcx %%r11, %%r10 ;" "adox 16(%1), %%r10 ;"
|
|
679
|
+
"mulx 56(%1), %%r11, %%rcx ;" /* c*C[7] */ "adcx %%rax, %%r11 ;" "adox 24(%1), %%r11 ;"
|
|
680
|
+
/****************************************/ "adcx %%rbx, %%rcx ;" "adox %%rbx, %%rcx ;"
|
|
681
|
+
"clc ;"
|
|
682
|
+
"mulx %%rcx, %%rax, %%rcx ;" /* c*C[4] */
|
|
683
|
+
"adcx %%rax, %%r8 ;"
|
|
684
|
+
"adcx %%rcx, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
685
|
+
"adcx %%rbx, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
686
|
+
"adcx %%rbx, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
687
|
+
"mov $0, %%ecx ;"
|
|
688
|
+
"cmovc %%edx, %%ecx ;"
|
|
689
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, (%0) ;"
|
|
690
|
+
:
|
|
691
|
+
: "r" (c), "r" (a)
|
|
692
|
+
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11"
|
|
693
|
+
);
|
|
625
694
|
#else
|
|
626
|
-
|
|
627
|
-
|
|
628
|
-
|
|
629
|
-
|
|
630
|
-
|
|
631
|
-
|
|
632
|
-
|
|
633
|
-
|
|
634
|
-
|
|
635
|
-
|
|
636
|
-
|
|
637
|
-
|
|
638
|
-
|
|
639
|
-
|
|
640
|
-
|
|
641
|
-
|
|
642
|
-
|
|
643
|
-
|
|
644
|
-
|
|
645
|
-
|
|
646
|
-
|
|
695
|
+
__asm__ __volatile__(
|
|
696
|
+
"movl $38, %%edx ;" /* 2*c = 38 = 2^256 */
|
|
697
|
+
"mulx 32(%1), %%r8, %%r10 ;" /* c*C[4] */
|
|
698
|
+
"mulx 40(%1), %%r9, %%r11 ;" /* c*C[5] */ "addq %%r10, %%r9 ;"
|
|
699
|
+
"mulx 48(%1), %%r10, %%rax ;" /* c*C[6] */ "adcq %%r11, %%r10 ;"
|
|
700
|
+
"mulx 56(%1), %%r11, %%rcx ;" /* c*C[7] */ "adcq %%rax, %%r11 ;"
|
|
701
|
+
/****************************************/ "adcq $0, %%rcx ;"
|
|
702
|
+
"addq (%1), %%r8 ;"
|
|
703
|
+
"adcq 8(%1), %%r9 ;"
|
|
704
|
+
"adcq 16(%1), %%r10 ;"
|
|
705
|
+
"adcq 24(%1), %%r11 ;"
|
|
706
|
+
"adcq $0, %%rcx ;"
|
|
707
|
+
"mulx %%rcx, %%rax, %%rcx ;" /* c*C[4] */
|
|
708
|
+
"addq %%rax, %%r8 ;"
|
|
709
|
+
"adcq %%rcx, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
710
|
+
"adcq $0, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
711
|
+
"adcq $0, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
712
|
+
"mov $0, %%ecx ;"
|
|
713
|
+
"cmovc %%edx, %%ecx ;"
|
|
714
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, (%0) ;"
|
|
715
|
+
:
|
|
716
|
+
: "r" (c), "r" (a)
|
|
717
|
+
: "memory", "cc", "%rax", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11"
|
|
718
|
+
);
|
|
647
719
|
#endif
|
|
648
720
|
#else /* Without BMI2 */
|
|
649
|
-
|
|
650
|
-
|
|
651
|
-
|
|
721
|
+
/**
|
|
722
|
+
* TODO: Multiplications using MULQ instruction.
|
|
723
|
+
**/
|
|
652
724
|
#endif
|
|
653
725
|
}
|
|
654
726
|
|
|
655
|
-
inline void add_EltFp25519_1w_x64(uint64_t *const c, uint64_t *const a, uint64_t *const b)
|
|
656
|
-
|
|
657
|
-
|
|
658
|
-
|
|
659
|
-
|
|
660
|
-
|
|
661
|
-
|
|
662
|
-
|
|
663
|
-
|
|
664
|
-
|
|
665
|
-
|
|
666
|
-
|
|
667
|
-
|
|
668
|
-
|
|
669
|
-
|
|
670
|
-
|
|
671
|
-
|
|
672
|
-
|
|
673
|
-
|
|
674
|
-
|
|
675
|
-
|
|
676
|
-
|
|
677
|
-
: "r" (c), "r" (a), "r" (b)
|
|
678
|
-
: "memory","cc", "%rax", "%rcx", "%r8", "%r9"
|
|
679
|
-
);
|
|
727
|
+
inline void add_EltFp25519_1w_x64(uint64_t *const c, uint64_t *const a, uint64_t *const b) {
|
|
728
|
+
#ifdef __ADX__
|
|
729
|
+
__asm__ __volatile__(
|
|
730
|
+
"mov $38, %%eax ;"
|
|
731
|
+
"xorl %%ecx, %%ecx ;"
|
|
732
|
+
"movq (%2), %%r8 ;" "adcx (%1), %%r8 ;"
|
|
733
|
+
"movq 8(%2), %%r9 ;" "adcx 8(%1), %%r9 ;"
|
|
734
|
+
"movq 16(%2), %%r10 ;" "adcx 16(%1), %%r10 ;"
|
|
735
|
+
"movq 24(%2), %%r11 ;" "adcx 24(%1), %%r11 ;"
|
|
736
|
+
"cmovc %%eax, %%ecx ;"
|
|
737
|
+
"xorl %%eax, %%eax ;"
|
|
738
|
+
"adcx %%rcx, %%r8 ;"
|
|
739
|
+
"adcx %%rax, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
740
|
+
"adcx %%rax, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
741
|
+
"adcx %%rax, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
742
|
+
"mov $38, %%ecx ;"
|
|
743
|
+
"cmovc %%ecx, %%eax ;"
|
|
744
|
+
"addq %%rax, %%r8 ;" "movq %%r8, (%0) ;"
|
|
745
|
+
:
|
|
746
|
+
: "r" (c), "r" (a), "r" (b)
|
|
747
|
+
: "memory", "cc", "%rax", "%rcx", "%r8", "%r9", "%r10", "%r11"
|
|
748
|
+
);
|
|
680
749
|
#else
|
|
681
|
-
|
|
682
|
-
|
|
683
|
-
|
|
684
|
-
|
|
685
|
-
|
|
686
|
-
|
|
687
|
-
|
|
688
|
-
|
|
689
|
-
|
|
690
|
-
|
|
691
|
-
|
|
692
|
-
|
|
693
|
-
|
|
694
|
-
|
|
695
|
-
|
|
696
|
-
|
|
697
|
-
|
|
698
|
-
|
|
699
|
-
|
|
700
|
-
: "memory","cc", "%rax", "%rcx", "%r8", "%r9"
|
|
701
|
-
);
|
|
750
|
+
__asm__ __volatile__(
|
|
751
|
+
"mov $38, %%eax ;"
|
|
752
|
+
"movq (%2), %%r8 ;" "addq (%1), %%r8 ;"
|
|
753
|
+
"movq 8(%2), %%r9 ;" "adcq 8(%1), %%r9 ;"
|
|
754
|
+
"movq 16(%2), %%r10 ;" "adcq 16(%1), %%r10 ;"
|
|
755
|
+
"movq 24(%2), %%r11 ;" "adcq 24(%1), %%r11 ;"
|
|
756
|
+
"mov $0, %%ecx ;"
|
|
757
|
+
"cmovc %%eax, %%ecx ;"
|
|
758
|
+
"addq %%rcx, %%r8 ;"
|
|
759
|
+
"adcq $0, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
760
|
+
"adcq $0, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
761
|
+
"adcq $0, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
762
|
+
"mov $0, %%ecx ;"
|
|
763
|
+
"cmovc %%eax, %%ecx ;"
|
|
764
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, (%0) ;"
|
|
765
|
+
:
|
|
766
|
+
: "r" (c), "r" (a), "r" (b)
|
|
767
|
+
: "memory", "cc", "%rax", "%rcx", "%r8", "%r9", "%r10", "%r11"
|
|
768
|
+
);
|
|
702
769
|
#endif
|
|
703
770
|
}
|
|
704
771
|
|
|
705
|
-
inline void sub_EltFp25519_1w_x64(uint64_t *const
|
|
706
|
-
|
|
707
|
-
|
|
708
|
-
|
|
709
|
-
|
|
710
|
-
|
|
711
|
-
|
|
712
|
-
|
|
713
|
-
|
|
714
|
-
|
|
715
|
-
|
|
716
|
-
|
|
717
|
-
|
|
718
|
-
|
|
719
|
-
|
|
720
|
-
|
|
721
|
-
|
|
722
|
-
|
|
723
|
-
|
|
724
|
-
|
|
725
|
-
:
|
|
726
|
-
: "r" (c), "r" (a), "r" (b)
|
|
727
|
-
: "memory","cc", "%rax", "%rcx", "%r8", "%r9"
|
|
728
|
-
);
|
|
772
|
+
inline void sub_EltFp25519_1w_x64(uint64_t *const c, uint64_t *const a, uint64_t *const b) {
|
|
773
|
+
__asm__ __volatile__(
|
|
774
|
+
"mov $38, %%eax ;"
|
|
775
|
+
"movq (%1), %%r8 ;" "subq (%2), %%r8 ;"
|
|
776
|
+
"movq 8(%1), %%r9 ;" "sbbq 8(%2), %%r9 ;"
|
|
777
|
+
"movq 16(%1), %%r10 ;" "sbbq 16(%2), %%r10 ;"
|
|
778
|
+
"movq 24(%1), %%r11 ;" "sbbq 24(%2), %%r11 ;"
|
|
779
|
+
"mov $0, %%ecx ;"
|
|
780
|
+
"cmovc %%eax, %%ecx ;"
|
|
781
|
+
"subq %%rcx, %%r8 ;"
|
|
782
|
+
"sbbq $0, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
783
|
+
"sbbq $0, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
784
|
+
"sbbq $0, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
785
|
+
"mov $0, %%ecx ;"
|
|
786
|
+
"cmovc %%eax, %%ecx ;"
|
|
787
|
+
"subq %%rcx, %%r8 ;" "movq %%r8, (%0) ;"
|
|
788
|
+
:
|
|
789
|
+
: "r" (c), "r" (a), "r" (b)
|
|
790
|
+
: "memory", "cc", "%rax", "%rcx", "%r8", "%r9", "%r10", "%r11"
|
|
791
|
+
);
|
|
729
792
|
}
|
|
730
793
|
|
|
731
|
-
|
|
732
|
-
|
|
794
|
+
/**
|
|
795
|
+
* Multiplication by a24 = (A+2)/4 = (486662+2)/4 = 121666
|
|
796
|
+
**/
|
|
797
|
+
inline void mul_a24_EltFp25519_1w_x64(uint64_t *const c, uint64_t *const a) {
|
|
733
798
|
#ifdef __BMI2__
|
|
734
|
-
|
|
735
|
-
|
|
736
|
-
|
|
737
|
-
|
|
738
|
-
|
|
739
|
-
|
|
740
|
-
|
|
741
|
-
|
|
742
|
-
|
|
743
|
-
|
|
744
|
-
|
|
745
|
-
|
|
746
|
-
|
|
747
|
-
|
|
748
|
-
|
|
749
|
-
|
|
750
|
-
|
|
751
|
-
|
|
752
|
-
|
|
753
|
-
|
|
754
|
-
|
|
755
|
-
: "r" (c), "r" (a), "r" (a24)
|
|
756
|
-
: "cc", "%rax", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11"
|
|
757
|
-
);
|
|
799
|
+
const uint64_t a24 = 121666;
|
|
800
|
+
__asm__ __volatile__(
|
|
801
|
+
"movq %2, %%rdx ;"
|
|
802
|
+
"mulx (%1), %%r8, %%r10 ;"
|
|
803
|
+
"mulx 8(%1), %%r9, %%r11 ;" "addq %%r10, %%r9 ;"
|
|
804
|
+
"mulx 16(%1), %%r10, %%rax ;" "adcq %%r11, %%r10 ;"
|
|
805
|
+
"mulx 24(%1), %%r11, %%rcx ;" "adcq %%rax, %%r11 ;"
|
|
806
|
+
/***************************/ "adcq $0, %%rcx ;"
|
|
807
|
+
"movl $38, %%edx ;" /* 2*c = 38 = 2^256 mod 2^255-19*/
|
|
808
|
+
"mulx %%rcx, %%rax, %%rcx ;"
|
|
809
|
+
"addq %%rax, %%r8 ;"
|
|
810
|
+
"adcq %%rcx, %%r9 ;" "movq %%r9, 8(%0) ;"
|
|
811
|
+
"adcq $0, %%r10 ;" "movq %%r10, 16(%0) ;"
|
|
812
|
+
"adcq $0, %%r11 ;" "movq %%r11, 24(%0) ;"
|
|
813
|
+
"mov $0, %%ecx ;"
|
|
814
|
+
"cmovc %%edx, %%ecx ;"
|
|
815
|
+
"addq %%rcx, %%r8 ;" "movq %%r8, (%0) ;"
|
|
816
|
+
:
|
|
817
|
+
: "r" (c), "r" (a), "r" (a24)
|
|
818
|
+
: "memory", "cc", "%rax", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11"
|
|
819
|
+
);
|
|
758
820
|
#else /* Without BMI2 */
|
|
759
|
-
|
|
760
|
-
|
|
761
|
-
|
|
821
|
+
/**
|
|
822
|
+
* TODO: Multiplications using MULQ instruction.
|
|
823
|
+
**/
|
|
762
824
|
#endif
|
|
763
825
|
}
|
|
764
826
|
|
|
765
|
-
void inv_EltFp25519_1w_x64(uint64_t *const
|
|
766
|
-
|
|
767
|
-
|
|
768
|
-
|
|
769
|
-
|
|
770
|
-
|
|
771
|
-
|
|
772
|
-
|
|
773
|
-
|
|
774
|
-
|
|
775
|
-
|
|
776
|
-
|
|
777
|
-
|
|
778
|
-
|
|
779
|
-
|
|
780
|
-
|
|
781
|
-
|
|
782
|
-
|
|
783
|
-
|
|
784
|
-
|
|
785
|
-
|
|
786
|
-
|
|
787
|
-
|
|
788
|
-
|
|
789
|
-
|
|
790
|
-
|
|
791
|
-
|
|
792
|
-
|
|
793
|
-
|
|
794
|
-
|
|
795
|
-
|
|
796
|
-
|
|
797
|
-
|
|
798
|
-
|
|
799
|
-
|
|
800
|
-
|
|
801
|
-
|
|
802
|
-
|
|
803
|
-
|
|
804
|
-
|
|
805
|
-
|
|
806
|
-
|
|
807
|
-
|
|
808
|
-
|
|
809
|
-
|
|
810
|
-
|
|
811
|
-
|
|
812
|
-
|
|
813
|
-
sqrn_EltFp25519_1w_x64(T[2],5);
|
|
814
|
-
mul_EltFp25519_1w_x64(T[1], T[1], T[2]);
|
|
827
|
+
void inv_EltFp25519_1w_x64(uint64_t *const c, uint64_t *const a) {
|
|
828
|
+
#define sqrn_EltFp25519_1w_x64(A, times)\
|
|
829
|
+
counter = times;\
|
|
830
|
+
while ( counter-- > 0) {\
|
|
831
|
+
sqr_EltFp25519_1w_x64(A);\
|
|
832
|
+
}
|
|
833
|
+
|
|
834
|
+
EltFp25519_1w_Buffer_x64 buffer_1w;
|
|
835
|
+
EltFp25519_1w_x64 x0, x1, x2;
|
|
836
|
+
uint64_t * T[5];
|
|
837
|
+
uint64_t counter;
|
|
838
|
+
|
|
839
|
+
T[0] = x0;
|
|
840
|
+
T[1] = c; /* x^(-1) */
|
|
841
|
+
T[2] = x1;
|
|
842
|
+
T[3] = x2;
|
|
843
|
+
T[4] = a; /* x */
|
|
844
|
+
|
|
845
|
+
copy_EltFp25519_1w_x64(T[1], a);
|
|
846
|
+
sqrn_EltFp25519_1w_x64(T[1], 1);
|
|
847
|
+
copy_EltFp25519_1w_x64(T[2], T[1]);
|
|
848
|
+
sqrn_EltFp25519_1w_x64(T[2], 2);
|
|
849
|
+
mul_EltFp25519_1w_x64(T[0], a, T[2]);
|
|
850
|
+
mul_EltFp25519_1w_x64(T[1], T[1], T[0]);
|
|
851
|
+
copy_EltFp25519_1w_x64(T[2], T[1]);
|
|
852
|
+
sqrn_EltFp25519_1w_x64(T[2], 1);
|
|
853
|
+
mul_EltFp25519_1w_x64(T[0], T[0], T[2]);
|
|
854
|
+
copy_EltFp25519_1w_x64(T[2], T[0]);
|
|
855
|
+
sqrn_EltFp25519_1w_x64(T[2], 5);
|
|
856
|
+
mul_EltFp25519_1w_x64(T[0], T[0], T[2]);
|
|
857
|
+
copy_EltFp25519_1w_x64(T[2], T[0]);
|
|
858
|
+
sqrn_EltFp25519_1w_x64(T[2], 10);
|
|
859
|
+
mul_EltFp25519_1w_x64(T[2], T[2], T[0]);
|
|
860
|
+
copy_EltFp25519_1w_x64(T[3], T[2]);
|
|
861
|
+
sqrn_EltFp25519_1w_x64(T[3], 20);
|
|
862
|
+
mul_EltFp25519_1w_x64(T[3], T[3], T[2]);
|
|
863
|
+
sqrn_EltFp25519_1w_x64(T[3], 10);
|
|
864
|
+
mul_EltFp25519_1w_x64(T[3], T[3], T[0]);
|
|
865
|
+
copy_EltFp25519_1w_x64(T[0], T[3]);
|
|
866
|
+
sqrn_EltFp25519_1w_x64(T[0], 50);
|
|
867
|
+
mul_EltFp25519_1w_x64(T[0], T[0], T[3]);
|
|
868
|
+
copy_EltFp25519_1w_x64(T[2], T[0]);
|
|
869
|
+
sqrn_EltFp25519_1w_x64(T[2], 100);
|
|
870
|
+
mul_EltFp25519_1w_x64(T[2], T[2], T[0]);
|
|
871
|
+
sqrn_EltFp25519_1w_x64(T[2], 50);
|
|
872
|
+
mul_EltFp25519_1w_x64(T[2], T[2], T[3]);
|
|
873
|
+
sqrn_EltFp25519_1w_x64(T[2], 5);
|
|
874
|
+
mul_EltFp25519_1w_x64(T[1], T[1], T[2]);
|
|
815
875
|
#undef sqrn_EltFp25519_1w_x64
|
|
816
876
|
}
|
|
817
877
|
|
|
818
|
-
|
|
819
|
-
|
|
820
|
-
|
|
821
|
-
|
|
822
|
-
|
|
878
|
+
/**
|
|
879
|
+
* Given C, a 256-bit number, fred_EltFp25519_1w_x64 updates C
|
|
880
|
+
* with a number such that 0 <= C < 2**255-19.
|
|
881
|
+
* Contributed by: Samuel Neves.
|
|
882
|
+
**/
|
|
883
|
+
inline void fred_EltFp25519_1w_x64(uint64_t *const c) {
|
|
884
|
+
__asm__ __volatile__ (
|
|
885
|
+
/* First, obtains a number less than 2^255. */
|
|
886
|
+
"btrq $63, 24(%0) ;"
|
|
887
|
+
"sbbl %%ecx, %%ecx ;"
|
|
888
|
+
"andq $19, %%rcx ;"
|
|
889
|
+
"addq %%rcx, (%0) ;"
|
|
890
|
+
"adcq $0, 8(%0) ;"
|
|
891
|
+
"adcq $0, 16(%0) ;"
|
|
892
|
+
"adcq $0, 24(%0) ;"
|
|
893
|
+
|
|
894
|
+
"btrq $63, 24(%0) ;"
|
|
895
|
+
"sbbl %%ecx, %%ecx ;"
|
|
896
|
+
"andq $19, %%rcx ;"
|
|
897
|
+
"addq %%rcx, (%0) ;"
|
|
898
|
+
"adcq $0, 8(%0) ;"
|
|
899
|
+
"adcq $0, 16(%0) ;"
|
|
900
|
+
"adcq $0, 24(%0) ;"
|
|
901
|
+
|
|
902
|
+
/* Then, in case the number fall into [2^255-19, 2^255-1] */
|
|
903
|
+
"cmpq $-19, (%0) ;"
|
|
904
|
+
"setaeb %%al ;"
|
|
905
|
+
"cmpq $-1, 8(%0) ;"
|
|
906
|
+
"setzb %%bl ;"
|
|
907
|
+
"cmpq $-1, 16(%0) ;"
|
|
908
|
+
"setzb %%cl ;"
|
|
909
|
+
"movq 24(%0), %%rdx ;"
|
|
910
|
+
"addq $1, %%rdx ;"
|
|
911
|
+
"shrq $63, %%rdx ;"
|
|
912
|
+
"andb %%bl, %%al ;"
|
|
913
|
+
"andb %%dl, %%cl ;"
|
|
914
|
+
"test %%cl, %%al ;"
|
|
915
|
+
"movl $0, %%eax ;"
|
|
916
|
+
"movl $19, %%ecx ;"
|
|
917
|
+
"cmovnz %%rcx, %%rax ;"
|
|
918
|
+
"addq %%rax, (%0) ;"
|
|
919
|
+
"adcq $0, 8(%0) ;"
|
|
920
|
+
"adcq $0, 16(%0) ;"
|
|
921
|
+
"adcq $0, 24(%0) ;"
|
|
922
|
+
"btrq $63, 24(%0) ;"
|
|
923
|
+
:
|
|
924
|
+
: "r"(c)
|
|
925
|
+
: "memory", "cc", "%rax", "%rbx", "%rcx", "%rdx"
|
|
926
|
+
);
|
|
823
927
|
}
|
|
824
|
-
|