• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* Copyright (c) 2015, Google Inc.
2  *
3  * Permission to use, copy, modify, and/or distribute this software for any
4  * purpose with or without fee is hereby granted, provided that the above
5  * copyright notice and this permission notice appear in all copies.
6  *
7  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
8  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
9  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
10  * SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
11  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
12  * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
13  * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */
14 
15 /* This code is mostly taken from the ref10 version of Ed25519 in SUPERCOP
16  * 20141124 (http://bench.cr.yp.to/supercop.html). That code is released as
17  * public domain but this file has the ISC license just to keep licencing
18  * simple.
19  *
20  * The field functions are shared by Ed25519 and X25519 where possible. */
21 
22 #include <openssl/curve25519.h>
23 
24 #include <string.h>
25 
26 #include "../internal.h"
27 #include "internal.h"
28 
29 
30 #if defined(BORINGSSL_X25519_X86_64)
31 
32 typedef struct { uint64_t v[5]; } fe25519;
33 
34 /* These functions are defined in asm/x25519-x86_64.S */
35 void x25519_x86_64_work_cswap(fe25519 *, uint64_t);
36 void x25519_x86_64_mul(fe25519 *out, const fe25519 *a, const fe25519 *b);
37 void x25519_x86_64_square(fe25519 *out, const fe25519 *a);
38 void x25519_x86_64_freeze(fe25519 *);
39 void x25519_x86_64_ladderstep(fe25519 *work);
40 
fe25519_setint(fe25519 * r,unsigned v)41 static void fe25519_setint(fe25519 *r, unsigned v) {
42   r->v[0] = v;
43   r->v[1] = 0;
44   r->v[2] = 0;
45   r->v[3] = 0;
46   r->v[4] = 0;
47 }
48 
49 /* Assumes input x being reduced below 2^255 */
fe25519_pack(unsigned char r[32],const fe25519 * x)50 static void fe25519_pack(unsigned char r[32], const fe25519 *x) {
51   fe25519 t;
52   t = *x;
53   x25519_x86_64_freeze(&t);
54 
55   r[0] = (uint8_t)(t.v[0] & 0xff);
56   r[1] = (uint8_t)((t.v[0] >> 8) & 0xff);
57   r[2] = (uint8_t)((t.v[0] >> 16) & 0xff);
58   r[3] = (uint8_t)((t.v[0] >> 24) & 0xff);
59   r[4] = (uint8_t)((t.v[0] >> 32) & 0xff);
60   r[5] = (uint8_t)((t.v[0] >> 40) & 0xff);
61   r[6] = (uint8_t)((t.v[0] >> 48));
62 
63   r[6] ^= (uint8_t)((t.v[1] << 3) & 0xf8);
64   r[7] = (uint8_t)((t.v[1] >> 5) & 0xff);
65   r[8] = (uint8_t)((t.v[1] >> 13) & 0xff);
66   r[9] = (uint8_t)((t.v[1] >> 21) & 0xff);
67   r[10] = (uint8_t)((t.v[1] >> 29) & 0xff);
68   r[11] = (uint8_t)((t.v[1] >> 37) & 0xff);
69   r[12] = (uint8_t)((t.v[1] >> 45));
70 
71   r[12] ^= (uint8_t)((t.v[2] << 6) & 0xc0);
72   r[13] = (uint8_t)((t.v[2] >> 2) & 0xff);
73   r[14] = (uint8_t)((t.v[2] >> 10) & 0xff);
74   r[15] = (uint8_t)((t.v[2] >> 18) & 0xff);
75   r[16] = (uint8_t)((t.v[2] >> 26) & 0xff);
76   r[17] = (uint8_t)((t.v[2] >> 34) & 0xff);
77   r[18] = (uint8_t)((t.v[2] >> 42) & 0xff);
78   r[19] = (uint8_t)((t.v[2] >> 50));
79 
80   r[19] ^= (uint8_t)((t.v[3] << 1) & 0xfe);
81   r[20] = (uint8_t)((t.v[3] >> 7) & 0xff);
82   r[21] = (uint8_t)((t.v[3] >> 15) & 0xff);
83   r[22] = (uint8_t)((t.v[3] >> 23) & 0xff);
84   r[23] = (uint8_t)((t.v[3] >> 31) & 0xff);
85   r[24] = (uint8_t)((t.v[3] >> 39) & 0xff);
86   r[25] = (uint8_t)((t.v[3] >> 47));
87 
88   r[25] ^= (uint8_t)((t.v[4] << 4) & 0xf0);
89   r[26] = (uint8_t)((t.v[4] >> 4) & 0xff);
90   r[27] = (uint8_t)((t.v[4] >> 12) & 0xff);
91   r[28] = (uint8_t)((t.v[4] >> 20) & 0xff);
92   r[29] = (uint8_t)((t.v[4] >> 28) & 0xff);
93   r[30] = (uint8_t)((t.v[4] >> 36) & 0xff);
94   r[31] = (uint8_t)((t.v[4] >> 44));
95 }
96 
fe25519_unpack(fe25519 * r,const uint8_t x[32])97 static void fe25519_unpack(fe25519 *r, const uint8_t x[32]) {
98   r->v[0] = x[0];
99   r->v[0] += (uint64_t)x[1] << 8;
100   r->v[0] += (uint64_t)x[2] << 16;
101   r->v[0] += (uint64_t)x[3] << 24;
102   r->v[0] += (uint64_t)x[4] << 32;
103   r->v[0] += (uint64_t)x[5] << 40;
104   r->v[0] += ((uint64_t)x[6] & 7) << 48;
105 
106   r->v[1] = x[6] >> 3;
107   r->v[1] += (uint64_t)x[7] << 5;
108   r->v[1] += (uint64_t)x[8] << 13;
109   r->v[1] += (uint64_t)x[9] << 21;
110   r->v[1] += (uint64_t)x[10] << 29;
111   r->v[1] += (uint64_t)x[11] << 37;
112   r->v[1] += ((uint64_t)x[12] & 63) << 45;
113 
114   r->v[2] = x[12] >> 6;
115   r->v[2] += (uint64_t)x[13] << 2;
116   r->v[2] += (uint64_t)x[14] << 10;
117   r->v[2] += (uint64_t)x[15] << 18;
118   r->v[2] += (uint64_t)x[16] << 26;
119   r->v[2] += (uint64_t)x[17] << 34;
120   r->v[2] += (uint64_t)x[18] << 42;
121   r->v[2] += ((uint64_t)x[19] & 1) << 50;
122 
123   r->v[3] = x[19] >> 1;
124   r->v[3] += (uint64_t)x[20] << 7;
125   r->v[3] += (uint64_t)x[21] << 15;
126   r->v[3] += (uint64_t)x[22] << 23;
127   r->v[3] += (uint64_t)x[23] << 31;
128   r->v[3] += (uint64_t)x[24] << 39;
129   r->v[3] += ((uint64_t)x[25] & 15) << 47;
130 
131   r->v[4] = x[25] >> 4;
132   r->v[4] += (uint64_t)x[26] << 4;
133   r->v[4] += (uint64_t)x[27] << 12;
134   r->v[4] += (uint64_t)x[28] << 20;
135   r->v[4] += (uint64_t)x[29] << 28;
136   r->v[4] += (uint64_t)x[30] << 36;
137   r->v[4] += ((uint64_t)x[31] & 127) << 44;
138 }
139 
fe25519_invert(fe25519 * r,const fe25519 * x)140 static void fe25519_invert(fe25519 *r, const fe25519 *x) {
141   fe25519 z2;
142   fe25519 z9;
143   fe25519 z11;
144   fe25519 z2_5_0;
145   fe25519 z2_10_0;
146   fe25519 z2_20_0;
147   fe25519 z2_50_0;
148   fe25519 z2_100_0;
149   fe25519 t;
150   int i;
151 
152   /* 2 */ x25519_x86_64_square(&z2, x);
153   /* 4 */ x25519_x86_64_square(&t, &z2);
154   /* 8 */ x25519_x86_64_square(&t, &t);
155   /* 9 */ x25519_x86_64_mul(&z9, &t, x);
156   /* 11 */ x25519_x86_64_mul(&z11, &z9, &z2);
157   /* 22 */ x25519_x86_64_square(&t, &z11);
158   /* 2^5 - 2^0 = 31 */ x25519_x86_64_mul(&z2_5_0, &t, &z9);
159 
160   /* 2^6 - 2^1 */ x25519_x86_64_square(&t, &z2_5_0);
161   /* 2^20 - 2^10 */ for (i = 1; i < 5; i++) { x25519_x86_64_square(&t, &t); }
162   /* 2^10 - 2^0 */ x25519_x86_64_mul(&z2_10_0, &t, &z2_5_0);
163 
164   /* 2^11 - 2^1 */ x25519_x86_64_square(&t, &z2_10_0);
165   /* 2^20 - 2^10 */ for (i = 1; i < 10; i++) { x25519_x86_64_square(&t, &t); }
166   /* 2^20 - 2^0 */ x25519_x86_64_mul(&z2_20_0, &t, &z2_10_0);
167 
168   /* 2^21 - 2^1 */ x25519_x86_64_square(&t, &z2_20_0);
169   /* 2^40 - 2^20 */ for (i = 1; i < 20; i++) { x25519_x86_64_square(&t, &t); }
170   /* 2^40 - 2^0 */ x25519_x86_64_mul(&t, &t, &z2_20_0);
171 
172   /* 2^41 - 2^1 */ x25519_x86_64_square(&t, &t);
173   /* 2^50 - 2^10 */ for (i = 1; i < 10; i++) { x25519_x86_64_square(&t, &t); }
174   /* 2^50 - 2^0 */ x25519_x86_64_mul(&z2_50_0, &t, &z2_10_0);
175 
176   /* 2^51 - 2^1 */ x25519_x86_64_square(&t, &z2_50_0);
177   /* 2^100 - 2^50 */ for (i = 1; i < 50; i++) { x25519_x86_64_square(&t, &t); }
178   /* 2^100 - 2^0 */ x25519_x86_64_mul(&z2_100_0, &t, &z2_50_0);
179 
180   /* 2^101 - 2^1 */ x25519_x86_64_square(&t, &z2_100_0);
181   /* 2^200 - 2^100 */ for (i = 1; i < 100; i++) {
182     x25519_x86_64_square(&t, &t);
183   }
184   /* 2^200 - 2^0 */ x25519_x86_64_mul(&t, &t, &z2_100_0);
185 
186   /* 2^201 - 2^1 */ x25519_x86_64_square(&t, &t);
187   /* 2^250 - 2^50 */ for (i = 1; i < 50; i++) { x25519_x86_64_square(&t, &t); }
188   /* 2^250 - 2^0 */ x25519_x86_64_mul(&t, &t, &z2_50_0);
189 
190   /* 2^251 - 2^1 */ x25519_x86_64_square(&t, &t);
191   /* 2^252 - 2^2 */ x25519_x86_64_square(&t, &t);
192   /* 2^253 - 2^3 */ x25519_x86_64_square(&t, &t);
193 
194   /* 2^254 - 2^4 */ x25519_x86_64_square(&t, &t);
195 
196   /* 2^255 - 2^5 */ x25519_x86_64_square(&t, &t);
197   /* 2^255 - 21 */ x25519_x86_64_mul(r, &t, &z11);
198 }
199 
mladder(fe25519 * xr,fe25519 * zr,const uint8_t s[32])200 static void mladder(fe25519 *xr, fe25519 *zr, const uint8_t s[32]) {
201   fe25519 work[5];
202 
203   work[0] = *xr;
204   fe25519_setint(work + 1, 1);
205   fe25519_setint(work + 2, 0);
206   work[3] = *xr;
207   fe25519_setint(work + 4, 1);
208 
209   int i, j;
210   uint8_t prevbit = 0;
211 
212   j = 6;
213   for (i = 31; i >= 0; i--) {
214     while (j >= 0) {
215       const uint8_t bit = 1 & (s[i] >> j);
216       const uint64_t swap = bit ^ prevbit;
217       prevbit = bit;
218       x25519_x86_64_work_cswap(work + 1, swap);
219       x25519_x86_64_ladderstep(work);
220       j -= 1;
221     }
222     j = 7;
223   }
224 
225   *xr = work[1];
226   *zr = work[2];
227 }
228 
x25519_x86_64(uint8_t out[32],const uint8_t scalar[32],const uint8_t point[32])229 void x25519_x86_64(uint8_t out[32], const uint8_t scalar[32],
230                   const uint8_t point[32]) {
231   uint8_t e[32];
232   OPENSSL_memcpy(e, scalar, sizeof(e));
233 
234   e[0] &= 248;
235   e[31] &= 127;
236   e[31] |= 64;
237 
238   fe25519 t;
239   fe25519 z;
240   fe25519_unpack(&t, point);
241   mladder(&t, &z, e);
242   fe25519_invert(&z, &z);
243   x25519_x86_64_mul(&t, &t, &z);
244   fe25519_pack(out, &t);
245 }
246 
247 #endif  /* BORINGSSL_X25519_X86_64 */
248