Electroneum
Loading...
Searching...
No Matches
curve25519-donna-sse2.h
Go to the documentation of this file.
1/*
2 Public domain by Andrew M. <liquidsun@gmail.com>
3 See: https://github.com/floodyberry/curve25519-donna
4
5 SSE2 curve25519 implementation
6*/
7
8#include <emmintrin.h>
9typedef __m128i xmmi;
10
11typedef union packedelem8_t {
12 unsigned char u[16];
13 xmmi v;
14} packedelem8;
15
20
25
26/* 10 elements + an extra 2 to fit in 3 xmm registers */
27typedef uint32_t bignum25519[12];
28typedef packedelem32 packed32bignum25519[5];
29typedef packedelem64 packed64bignum25519[10];
30
31static const packedelem32 bot32bitmask = {{0xffffffff, 0x00000000, 0xffffffff, 0x00000000}};
32static const packedelem32 top32bitmask = {{0x00000000, 0xffffffff, 0x00000000, 0xffffffff}};
33static const packedelem32 top64bitmask = {{0x00000000, 0x00000000, 0xffffffff, 0xffffffff}};
34static const packedelem32 bot64bitmask = {{0xffffffff, 0xffffffff, 0x00000000, 0x00000000}};
35
36/* reduction masks */
37static const packedelem64 packedmask26 = {{0x03ffffff, 0x03ffffff}};
38static const packedelem64 packedmask25 = {{0x01ffffff, 0x01ffffff}};
39static const packedelem32 packedmask2625 = {{0x3ffffff,0,0x1ffffff,0}};
40static const packedelem32 packedmask26262626 = {{0x03ffffff, 0x03ffffff, 0x03ffffff, 0x03ffffff}};
41static const packedelem32 packedmask25252525 = {{0x01ffffff, 0x01ffffff, 0x01ffffff, 0x01ffffff}};
42
43/* multipliers */
44static const packedelem64 packednineteen = {{19, 19}};
45static const packedelem64 packednineteenone = {{19, 1}};
46static const packedelem64 packedthirtyeight = {{38, 38}};
47static const packedelem64 packed3819 = {{19*2,19}};
48static const packedelem64 packed9638 = {{19*4,19*2}};
49
50/* 121666,121665 */
51static const packedelem64 packed121666121665 = {{121666, 121665}};
52
53/* 2*(2^255 - 19) = 0 mod p */
54static const packedelem32 packed2p0 = {{0x7ffffda,0x3fffffe,0x7fffffe,0x3fffffe}};
55static const packedelem32 packed2p1 = {{0x7fffffe,0x3fffffe,0x7fffffe,0x3fffffe}};
56static const packedelem32 packed2p2 = {{0x7fffffe,0x3fffffe,0x0000000,0x0000000}};
57
58static const packedelem32 packed32packed2p0 = {{0x7ffffda,0x7ffffda,0x3fffffe,0x3fffffe}};
59static const packedelem32 packed32packed2p1 = {{0x7fffffe,0x7fffffe,0x3fffffe,0x3fffffe}};
60
61/* 4*(2^255 - 19) = 0 mod p */
62static const packedelem32 packed4p0 = {{0xfffffb4,0x7fffffc,0xffffffc,0x7fffffc}};
63static const packedelem32 packed4p1 = {{0xffffffc,0x7fffffc,0xffffffc,0x7fffffc}};
64static const packedelem32 packed4p2 = {{0xffffffc,0x7fffffc,0x0000000,0x0000000}};
65
66static const packedelem32 packed32packed4p0 = {{0xfffffb4,0xfffffb4,0x7fffffc,0x7fffffc}};
67static const packedelem32 packed32packed4p1 = {{0xffffffc,0xffffffc,0x7fffffc,0x7fffffc}};
68
69/* out = in */
70DONNA_INLINE static void
71curve25519_copy(bignum25519 out, const bignum25519 in) {
72 xmmi x0,x1,x2;
73 x0 = _mm_load_si128((xmmi*)in + 0);
74 x1 = _mm_load_si128((xmmi*)in + 1);
75 x2 = _mm_load_si128((xmmi*)in + 2);
76 _mm_store_si128((xmmi*)out + 0, x0);
77 _mm_store_si128((xmmi*)out + 1, x1);
78 _mm_store_si128((xmmi*)out + 2, x2);
79}
80
81/* out = a + b */
82DONNA_INLINE static void
83curve25519_add(bignum25519 out, const bignum25519 a, const bignum25519 b) {
84 xmmi a0,a1,a2,b0,b1,b2;
85 a0 = _mm_load_si128((xmmi*)a + 0);
86 a1 = _mm_load_si128((xmmi*)a + 1);
87 a2 = _mm_load_si128((xmmi*)a + 2);
88 b0 = _mm_load_si128((xmmi*)b + 0);
89 b1 = _mm_load_si128((xmmi*)b + 1);
90 b2 = _mm_load_si128((xmmi*)b + 2);
91 a0 = _mm_add_epi32(a0, b0);
92 a1 = _mm_add_epi32(a1, b1);
93 a2 = _mm_add_epi32(a2, b2);
94 _mm_store_si128((xmmi*)out + 0, a0);
95 _mm_store_si128((xmmi*)out + 1, a1);
96 _mm_store_si128((xmmi*)out + 2, a2);
97}
98
99#define curve25519_add_after_basic curve25519_add_reduce
100DONNA_INLINE static void
101curve25519_add_reduce(bignum25519 out, const bignum25519 a, const bignum25519 b) {
102 xmmi a0,a1,a2,b0,b1,b2;
103 xmmi c1,c2,c3;
104 xmmi r0,r1,r2,r3,r4,r5;
105
106 a0 = _mm_load_si128((xmmi*)a + 0);
107 a1 = _mm_load_si128((xmmi*)a + 1);
108 a2 = _mm_load_si128((xmmi*)a + 2);
109 b0 = _mm_load_si128((xmmi*)b + 0);
110 b1 = _mm_load_si128((xmmi*)b + 1);
111 b2 = _mm_load_si128((xmmi*)b + 2);
112 a0 = _mm_add_epi32(a0, b0);
113 a1 = _mm_add_epi32(a1, b1);
114 a2 = _mm_add_epi32(a2, b2);
115
116 r0 = _mm_and_si128(_mm_unpacklo_epi64(a0, a1), bot32bitmask.v);
117 r1 = _mm_srli_epi64(_mm_unpacklo_epi64(a0, a1), 32);
118 r2 = _mm_and_si128(_mm_unpackhi_epi64(a0, a1), bot32bitmask.v);
119 r3 = _mm_srli_epi64(_mm_unpackhi_epi64(a0, a1), 32);
120 r4 = _mm_and_si128(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), bot32bitmask.v);
121 r5 = _mm_srli_epi64(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), 32);
122
123 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
124 c1 = _mm_srli_epi64(r1, 25); c2 = _mm_srli_epi64(r3, 25); r1 = _mm_and_si128(r1, packedmask25.v); r3 = _mm_and_si128(r3, packedmask25.v); r2 = _mm_add_epi64(r2, c1); r4 = _mm_add_epi64(r4, c2); c3 = _mm_slli_si128(c2, 8);
125 c1 = _mm_srli_epi64(r4, 26); r4 = _mm_and_si128(r4, packedmask26.v); r5 = _mm_add_epi64(r5, c1);
126 c1 = _mm_srli_epi64(r5, 25); r5 = _mm_and_si128(r5, packedmask25.v); r0 = _mm_add_epi64(r0, _mm_unpackhi_epi64(_mm_mul_epu32(c1, packednineteen.v), c3));
127 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
128
129 _mm_store_si128((xmmi*)out + 0, _mm_unpacklo_epi64(_mm_unpacklo_epi32(r0, r1), _mm_unpacklo_epi32(r2, r3)));
130 _mm_store_si128((xmmi*)out + 1, _mm_unpacklo_epi64(_mm_unpackhi_epi32(r0, r1), _mm_unpackhi_epi32(r2, r3)));
131 _mm_store_si128((xmmi*)out + 2, _mm_unpackhi_epi32(r4, r5));
132}
133
134DONNA_INLINE static void
135curve25519_sub(bignum25519 out, const bignum25519 a, const bignum25519 b) {
136 xmmi a0,a1,a2,b0,b1,b2;
137 xmmi c1,c2;
138 xmmi r0,r1;
139
140 a0 = _mm_load_si128((xmmi*)a + 0);
141 a1 = _mm_load_si128((xmmi*)a + 1);
142 a2 = _mm_load_si128((xmmi*)a + 2);
143 a0 = _mm_add_epi32(a0, packed2p0.v);
144 a1 = _mm_add_epi32(a1, packed2p1.v);
145 a2 = _mm_add_epi32(a2, packed2p2.v);
146 b0 = _mm_load_si128((xmmi*)b + 0);
147 b1 = _mm_load_si128((xmmi*)b + 1);
148 b2 = _mm_load_si128((xmmi*)b + 2);
149 a0 = _mm_sub_epi32(a0, b0);
150 a1 = _mm_sub_epi32(a1, b1);
151 a2 = _mm_sub_epi32(a2, b2);
152
153 r0 = _mm_and_si128(_mm_shuffle_epi32(a0, _MM_SHUFFLE(2,2,0,0)), bot32bitmask.v);
154 r1 = _mm_and_si128(_mm_shuffle_epi32(a0, _MM_SHUFFLE(3,3,1,1)), bot32bitmask.v);
155
156 c1 = _mm_srli_epi32(r0, 26);
157 c2 = _mm_srli_epi32(r1, 25);
158 r0 = _mm_and_si128(r0, packedmask26.v);
159 r1 = _mm_and_si128(r1, packedmask25.v);
160 r0 = _mm_add_epi32(r0, _mm_slli_si128(c2, 8));
161 r1 = _mm_add_epi32(r1, c1);
162
163 a0 = _mm_unpacklo_epi64(_mm_unpacklo_epi32(r0, r1), _mm_unpackhi_epi32(r0, r1));
164 a1 = _mm_add_epi32(a1, _mm_srli_si128(c2, 8));
165
166 _mm_store_si128((xmmi*)out + 0, a0);
167 _mm_store_si128((xmmi*)out + 1, a1);
168 _mm_store_si128((xmmi*)out + 2, a2);
169}
170
171DONNA_INLINE static void
172curve25519_sub_after_basic(bignum25519 out, const bignum25519 a, const bignum25519 b) {
173 xmmi a0,a1,a2,b0,b1,b2;
174 xmmi c1,c2,c3;
175 xmmi r0,r1,r2,r3,r4,r5;
176
177 a0 = _mm_load_si128((xmmi*)a + 0);
178 a1 = _mm_load_si128((xmmi*)a + 1);
179 a2 = _mm_load_si128((xmmi*)a + 2);
180 a0 = _mm_add_epi32(a0, packed4p0.v);
181 a1 = _mm_add_epi32(a1, packed4p1.v);
182 a2 = _mm_add_epi32(a2, packed4p2.v);
183 b0 = _mm_load_si128((xmmi*)b + 0);
184 b1 = _mm_load_si128((xmmi*)b + 1);
185 b2 = _mm_load_si128((xmmi*)b + 2);
186 a0 = _mm_sub_epi32(a0, b0);
187 a1 = _mm_sub_epi32(a1, b1);
188 a2 = _mm_sub_epi32(a2, b2);
189
190 r0 = _mm_and_si128(_mm_unpacklo_epi64(a0, a1), bot32bitmask.v);
191 r1 = _mm_srli_epi64(_mm_unpacklo_epi64(a0, a1), 32);
192 r2 = _mm_and_si128(_mm_unpackhi_epi64(a0, a1), bot32bitmask.v);
193 r3 = _mm_srli_epi64(_mm_unpackhi_epi64(a0, a1), 32);
194 r4 = _mm_and_si128(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), bot32bitmask.v);
195 r5 = _mm_srli_epi64(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), 32);
196
197 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
198 c1 = _mm_srli_epi64(r1, 25); c2 = _mm_srli_epi64(r3, 25); r1 = _mm_and_si128(r1, packedmask25.v); r3 = _mm_and_si128(r3, packedmask25.v); r2 = _mm_add_epi64(r2, c1); r4 = _mm_add_epi64(r4, c2); c3 = _mm_slli_si128(c2, 8);
199 c1 = _mm_srli_epi64(r4, 26); r4 = _mm_and_si128(r4, packedmask26.v); r5 = _mm_add_epi64(r5, c1);
200 c1 = _mm_srli_epi64(r5, 25); r5 = _mm_and_si128(r5, packedmask25.v); r0 = _mm_add_epi64(r0, _mm_unpackhi_epi64(_mm_mul_epu32(c1, packednineteen.v), c3));
201 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
202
203 _mm_store_si128((xmmi*)out + 0, _mm_unpacklo_epi64(_mm_unpacklo_epi32(r0, r1), _mm_unpacklo_epi32(r2, r3)));
204 _mm_store_si128((xmmi*)out + 1, _mm_unpacklo_epi64(_mm_unpackhi_epi32(r0, r1), _mm_unpackhi_epi32(r2, r3)));
205 _mm_store_si128((xmmi*)out + 2, _mm_unpackhi_epi32(r4, r5));
206}
207
208DONNA_INLINE static void
209curve25519_sub_reduce(bignum25519 out, const bignum25519 a, const bignum25519 b) {
210 xmmi a0,a1,a2,b0,b1,b2;
211 xmmi c1,c2,c3;
212 xmmi r0,r1,r2,r3,r4,r5;
213
214 a0 = _mm_load_si128((xmmi*)a + 0);
215 a1 = _mm_load_si128((xmmi*)a + 1);
216 a2 = _mm_load_si128((xmmi*)a + 2);
217 a0 = _mm_add_epi32(a0, packed2p0.v);
218 a1 = _mm_add_epi32(a1, packed2p1.v);
219 a2 = _mm_add_epi32(a2, packed2p2.v);
220 b0 = _mm_load_si128((xmmi*)b + 0);
221 b1 = _mm_load_si128((xmmi*)b + 1);
222 b2 = _mm_load_si128((xmmi*)b + 2);
223 a0 = _mm_sub_epi32(a0, b0);
224 a1 = _mm_sub_epi32(a1, b1);
225 a2 = _mm_sub_epi32(a2, b2);
226
227 r0 = _mm_and_si128(_mm_unpacklo_epi64(a0, a1), bot32bitmask.v);
228 r1 = _mm_srli_epi64(_mm_unpacklo_epi64(a0, a1), 32);
229 r2 = _mm_and_si128(_mm_unpackhi_epi64(a0, a1), bot32bitmask.v);
230 r3 = _mm_srli_epi64(_mm_unpackhi_epi64(a0, a1), 32);
231 r4 = _mm_and_si128(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), bot32bitmask.v);
232 r5 = _mm_srli_epi64(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), 32);
233
234 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
235 c1 = _mm_srli_epi64(r1, 25); c2 = _mm_srli_epi64(r3, 25); r1 = _mm_and_si128(r1, packedmask25.v); r3 = _mm_and_si128(r3, packedmask25.v); r2 = _mm_add_epi64(r2, c1); r4 = _mm_add_epi64(r4, c2); c3 = _mm_slli_si128(c2, 8);
236 c1 = _mm_srli_epi64(r4, 26); r4 = _mm_and_si128(r4, packedmask26.v); r5 = _mm_add_epi64(r5, c1);
237 c1 = _mm_srli_epi64(r5, 25); r5 = _mm_and_si128(r5, packedmask25.v); r0 = _mm_add_epi64(r0, _mm_unpackhi_epi64(_mm_mul_epu32(c1, packednineteen.v), c3));
238 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
239
240 _mm_store_si128((xmmi*)out + 0, _mm_unpacklo_epi64(_mm_unpacklo_epi32(r0, r1), _mm_unpacklo_epi32(r2, r3)));
241 _mm_store_si128((xmmi*)out + 1, _mm_unpacklo_epi64(_mm_unpackhi_epi32(r0, r1), _mm_unpackhi_epi32(r2, r3)));
242 _mm_store_si128((xmmi*)out + 2, _mm_unpackhi_epi32(r4, r5));
243}
244
245
246DONNA_INLINE static void
247curve25519_neg(bignum25519 out, const bignum25519 b) {
248 xmmi a0,a1,a2,b0,b1,b2;
249 xmmi c1,c2,c3;
250 xmmi r0,r1,r2,r3,r4,r5;
251
252 a0 = packed2p0.v;
253 a1 = packed2p1.v;
254 a2 = packed2p2.v;
255 b0 = _mm_load_si128((xmmi*)b + 0);
256 b1 = _mm_load_si128((xmmi*)b + 1);
257 b2 = _mm_load_si128((xmmi*)b + 2);
258 a0 = _mm_sub_epi32(a0, b0);
259 a1 = _mm_sub_epi32(a1, b1);
260 a2 = _mm_sub_epi32(a2, b2);
261
262 r0 = _mm_and_si128(_mm_unpacklo_epi64(a0, a1), bot32bitmask.v);
263 r1 = _mm_srli_epi64(_mm_unpacklo_epi64(a0, a1), 32);
264 r2 = _mm_and_si128(_mm_unpackhi_epi64(a0, a1), bot32bitmask.v);
265 r3 = _mm_srli_epi64(_mm_unpackhi_epi64(a0, a1), 32);
266 r4 = _mm_and_si128(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), bot32bitmask.v);
267 r5 = _mm_srli_epi64(_mm_unpacklo_epi64(_mm_setzero_si128(), a2), 32);
268
269 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
270 c1 = _mm_srli_epi64(r1, 25); c2 = _mm_srli_epi64(r3, 25); r1 = _mm_and_si128(r1, packedmask25.v); r3 = _mm_and_si128(r3, packedmask25.v); r2 = _mm_add_epi64(r2, c1); r4 = _mm_add_epi64(r4, c2); c3 = _mm_slli_si128(c2, 8);
271 c1 = _mm_srli_epi64(r4, 26); r4 = _mm_and_si128(r4, packedmask26.v); r5 = _mm_add_epi64(r5, c1);
272 c1 = _mm_srli_epi64(r5, 25); r5 = _mm_and_si128(r5, packedmask25.v); r0 = _mm_add_epi64(r0, _mm_unpackhi_epi64(_mm_mul_epu32(c1, packednineteen.v), c3));
273 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
274
275 _mm_store_si128((xmmi*)out + 0, _mm_unpacklo_epi64(_mm_unpacklo_epi32(r0, r1), _mm_unpacklo_epi32(r2, r3)));
276 _mm_store_si128((xmmi*)out + 1, _mm_unpacklo_epi64(_mm_unpackhi_epi32(r0, r1), _mm_unpackhi_epi32(r2, r3)));
277 _mm_store_si128((xmmi*)out + 2, _mm_unpackhi_epi32(r4, r5));
278}
279
280
281/* Multiply two numbers: out = in2 * in */
282static void
283curve25519_mul(bignum25519 out, const bignum25519 r, const bignum25519 s) {
284 xmmi m01,m23,m45,m67,m89;
285 xmmi m0123,m4567;
286 xmmi s0123,s4567;
287 xmmi s01,s23,s45,s67,s89;
288 xmmi s12,s34,s56,s78,s9;
289 xmmi r0,r2,r4,r6,r8;
290 xmmi r1,r3,r5,r7,r9;
291 xmmi r119,r219,r319,r419,r519,r619,r719,r819,r919;
292 xmmi c1,c2,c3;
293
294 s0123 = _mm_load_si128((xmmi*)s + 0);
295 s01 = _mm_shuffle_epi32(s0123,_MM_SHUFFLE(3,1,2,0));
296 s12 = _mm_shuffle_epi32(s0123, _MM_SHUFFLE(2,2,1,1));
297 s23 = _mm_shuffle_epi32(s0123,_MM_SHUFFLE(3,3,2,2));
298 s4567 = _mm_load_si128((xmmi*)s + 1);
299 s34 = _mm_unpacklo_epi64(_mm_srli_si128(s0123,12),s4567);
300 s45 = _mm_shuffle_epi32(s4567,_MM_SHUFFLE(3,1,2,0));
301 s56 = _mm_shuffle_epi32(s4567, _MM_SHUFFLE(2,2,1,1));
302 s67 = _mm_shuffle_epi32(s4567,_MM_SHUFFLE(3,3,2,2));
303 s89 = _mm_load_si128((xmmi*)s + 2);
304 s78 = _mm_unpacklo_epi64(_mm_srli_si128(s4567,12),s89);
305 s89 = _mm_shuffle_epi32(s89,_MM_SHUFFLE(3,1,2,0));
306 s9 = _mm_shuffle_epi32(s89, _MM_SHUFFLE(3,3,2,2));
307
308 r0 = _mm_load_si128((xmmi*)r + 0);
309 r1 = _mm_shuffle_epi32(r0, _MM_SHUFFLE(1,1,1,1));
310 r1 = _mm_add_epi64(r1, _mm_and_si128(r1, top64bitmask.v));
311 r2 = _mm_shuffle_epi32(r0, _MM_SHUFFLE(2,2,2,2));
312 r3 = _mm_shuffle_epi32(r0, _MM_SHUFFLE(3,3,3,3));
313 r3 = _mm_add_epi64(r3, _mm_and_si128(r3, top64bitmask.v));
314 r0 = _mm_shuffle_epi32(r0, _MM_SHUFFLE(0,0,0,0));
315 r4 = _mm_load_si128((xmmi*)r + 1);
316 r5 = _mm_shuffle_epi32(r4, _MM_SHUFFLE(1,1,1,1));
317 r5 = _mm_add_epi64(r5, _mm_and_si128(r5, top64bitmask.v));
318 r6 = _mm_shuffle_epi32(r4, _MM_SHUFFLE(2,2,2,2));
319 r7 = _mm_shuffle_epi32(r4, _MM_SHUFFLE(3,3,3,3));
320 r7 = _mm_add_epi64(r7, _mm_and_si128(r7, top64bitmask.v));
321 r4 = _mm_shuffle_epi32(r4, _MM_SHUFFLE(0,0,0,0));
322 r8 = _mm_load_si128((xmmi*)r + 2);
323 r9 = _mm_shuffle_epi32(r8, _MM_SHUFFLE(3,1,3,1));
324 r9 = _mm_add_epi64(r9, _mm_and_si128(r9, top64bitmask.v));
325 r8 = _mm_shuffle_epi32(r8, _MM_SHUFFLE(3,0,3,0));
326
327 m01 = _mm_mul_epu32(r1,s01);
328 m23 = _mm_mul_epu32(r1,s23);
329 m45 = _mm_mul_epu32(r1,s45);
330 m67 = _mm_mul_epu32(r1,s67);
331 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r3,s01));
332 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r3,s23));
333 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r3,s45));
334 m89 = _mm_mul_epu32(r1,s89);
335 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r5,s01));
336 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r5,s23));
337 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r3,s67));
338 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r7,s01));
339 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r5,s45));
340 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r7,s23));
341 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r9,s01));
342
343 /* shift up */
344 m89 = _mm_unpackhi_epi64(m67,_mm_slli_si128(m89,8));
345 m67 = _mm_unpackhi_epi64(m45,_mm_slli_si128(m67,8));
346 m45 = _mm_unpackhi_epi64(m23,_mm_slli_si128(m45,8));
347 m23 = _mm_unpackhi_epi64(m01,_mm_slli_si128(m23,8));
348 m01 = _mm_unpackhi_epi64(_mm_setzero_si128(),_mm_slli_si128(m01,8));
349
350 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r0,s01));
351 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r0,s23));
352 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r0,s45));
353 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r0,s67));
354 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r2,s01));
355 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r2,s23));
356 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r4,s23));
357 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r0,s89));
358 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r4,s01));
359 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r2,s45));
360 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r2,s67));
361 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r6,s01));
362 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r4,s45));
363 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r6,s23));
364 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r8,s01));
365
366 r219 = _mm_mul_epu32(r2, packednineteen.v);
367 r419 = _mm_mul_epu32(r4, packednineteen.v);
368 r619 = _mm_mul_epu32(r6, packednineteen.v);
369 r819 = _mm_mul_epu32(r8, packednineteen.v);
370 r119 = _mm_shuffle_epi32(r1,_MM_SHUFFLE(0,0,2,2)); r119 = _mm_mul_epu32(r119, packednineteen.v);
371 r319 = _mm_shuffle_epi32(r3,_MM_SHUFFLE(0,0,2,2)); r319 = _mm_mul_epu32(r319, packednineteen.v);
372 r519 = _mm_shuffle_epi32(r5,_MM_SHUFFLE(0,0,2,2)); r519 = _mm_mul_epu32(r519, packednineteen.v);
373 r719 = _mm_shuffle_epi32(r7,_MM_SHUFFLE(0,0,2,2)); r719 = _mm_mul_epu32(r719, packednineteen.v);
374 r919 = _mm_shuffle_epi32(r9,_MM_SHUFFLE(0,0,2,2)); r919 = _mm_mul_epu32(r919, packednineteen.v);
375
376 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r919,s12));
377 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r919,s34));
378 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r919,s56));
379 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r919,s78));
380 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r719,s34));
381 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r719,s56));
382 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r719,s78));
383 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r719,s9));
384 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r519,s56));
385 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r519,s78));
386 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r519,s9));
387 m67 = _mm_add_epi64(m67,_mm_mul_epu32(r819,s89));
388 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r319,s78));
389 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r319,s9));
390 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r619,s89));
391 m89 = _mm_add_epi64(m89,_mm_mul_epu32(r919,s9));
392 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r819,s23));
393 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r819,s45));
394 m45 = _mm_add_epi64(m45,_mm_mul_epu32(r819,s67));
395 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r619,s45));
396 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r619,s67));
397 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r419,s67));
398 m23 = _mm_add_epi64(m23,_mm_mul_epu32(r419,s89));
399 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r219,s89));
400 m01 = _mm_add_epi64(m01,_mm_mul_epu32(r119,s9));
401
402 r0 = _mm_unpacklo_epi64(m01, m45);
403 r1 = _mm_unpackhi_epi64(m01, m45);
404 r2 = _mm_unpacklo_epi64(m23, m67);
405 r3 = _mm_unpackhi_epi64(m23, m67);
406 r4 = _mm_unpacklo_epi64(m89, m89);
407 r5 = _mm_unpackhi_epi64(m89, m89);
408
409 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
410 c1 = _mm_srli_epi64(r1, 25); c2 = _mm_srli_epi64(r3, 25); r1 = _mm_and_si128(r1, packedmask25.v); r3 = _mm_and_si128(r3, packedmask25.v); r2 = _mm_add_epi64(r2, c1); r4 = _mm_add_epi64(r4, c2); c3 = _mm_slli_si128(c2, 8);
411 c1 = _mm_srli_epi64(r4, 26); r4 = _mm_and_si128(r4, packedmask26.v); r5 = _mm_add_epi64(r5, c1);
412 c1 = _mm_srli_epi64(r5, 25); r5 = _mm_and_si128(r5, packedmask25.v); r0 = _mm_add_epi64(r0, _mm_unpackhi_epi64(_mm_mul_epu32(c1, packednineteen.v), c3));
413 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
414
415 m0123 = _mm_unpacklo_epi32(r0, r1);
416 m4567 = _mm_unpackhi_epi32(r0, r1);
417 m0123 = _mm_unpacklo_epi64(m0123, _mm_unpacklo_epi32(r2, r3));
418 m4567 = _mm_unpacklo_epi64(m4567, _mm_unpackhi_epi32(r2, r3));
419 m89 = _mm_unpackhi_epi32(r4, r5);
420
421 _mm_store_si128((xmmi*)out + 0, m0123);
422 _mm_store_si128((xmmi*)out + 1, m4567);
423 _mm_store_si128((xmmi*)out + 2, m89);
424}
425
426DONNA_NOINLINE static void
427curve25519_mul_noinline(bignum25519 out, const bignum25519 r, const bignum25519 s) {
428 curve25519_mul(out, r, s);
429}
430
431#define curve25519_square(r, n) curve25519_square_times(r, n, 1)
432static void
433curve25519_square_times(bignum25519 r, const bignum25519 in, int count) {
434 xmmi m01,m23,m45,m67,m89;
435 xmmi r0,r1,r2,r3,r4,r5,r6,r7,r8,r9;
436 xmmi r0a,r1a,r2a,r3a,r7a,r9a;
437 xmmi r0123,r4567;
438 xmmi r01,r23,r45,r67,r6x,r89,r8x;
439 xmmi r12,r34,r56,r78,r9x;
440 xmmi r5619;
441 xmmi c1,c2,c3;
442
443 r0123 = _mm_load_si128((xmmi*)in + 0);
444 r01 = _mm_shuffle_epi32(r0123,_MM_SHUFFLE(3,1,2,0));
445 r23 = _mm_shuffle_epi32(r0123,_MM_SHUFFLE(3,3,2,2));
446 r4567 = _mm_load_si128((xmmi*)in + 1);
447 r45 = _mm_shuffle_epi32(r4567,_MM_SHUFFLE(3,1,2,0));
448 r67 = _mm_shuffle_epi32(r4567,_MM_SHUFFLE(3,3,2,2));
449 r89 = _mm_load_si128((xmmi*)in + 2);
450 r89 = _mm_shuffle_epi32(r89,_MM_SHUFFLE(3,1,2,0));
451
452 do {
453 r12 = _mm_unpackhi_epi64(r01, _mm_slli_si128(r23, 8));
454 r0 = _mm_shuffle_epi32(r01, _MM_SHUFFLE(0,0,0,0));
455 r0 = _mm_add_epi64(r0, _mm_and_si128(r0, top64bitmask.v));
456 r0a = _mm_shuffle_epi32(r0,_MM_SHUFFLE(3,2,1,2));
457 r1 = _mm_shuffle_epi32(r01, _MM_SHUFFLE(2,2,2,2));
458 r2 = _mm_shuffle_epi32(r23, _MM_SHUFFLE(0,0,0,0));
459 r2 = _mm_add_epi64(r2, _mm_and_si128(r2, top64bitmask.v));
460 r2a = _mm_shuffle_epi32(r2,_MM_SHUFFLE(3,2,1,2));
461 r3 = _mm_shuffle_epi32(r23, _MM_SHUFFLE(2,2,2,2));
462 r34 = _mm_unpackhi_epi64(r23, _mm_slli_si128(r45, 8));
463 r4 = _mm_shuffle_epi32(r45, _MM_SHUFFLE(0,0,0,0));
464 r4 = _mm_add_epi64(r4, _mm_and_si128(r4, top64bitmask.v));
465 r56 = _mm_unpackhi_epi64(r45, _mm_slli_si128(r67, 8));
466 r5619 = _mm_mul_epu32(r56, packednineteen.v);
467 r5 = _mm_shuffle_epi32(r5619, _MM_SHUFFLE(1,1,1,0));
468 r6 = _mm_shuffle_epi32(r5619, _MM_SHUFFLE(3,2,3,2));
469 r78 = _mm_unpackhi_epi64(r67, _mm_slli_si128(r89, 8));
470 r6x = _mm_unpacklo_epi64(r67, _mm_setzero_si128());
471 r7 = _mm_shuffle_epi32(r67, _MM_SHUFFLE(2,2,2,2));
472 r7 = _mm_mul_epu32(r7, packed3819.v);
473 r7a = _mm_shuffle_epi32(r7, _MM_SHUFFLE(3,3,3,2));
474 r8x = _mm_unpacklo_epi64(r89, _mm_setzero_si128());
475 r8 = _mm_shuffle_epi32(r89, _MM_SHUFFLE(0,0,0,0));
476 r8 = _mm_mul_epu32(r8, packednineteen.v);
477 r9 = _mm_shuffle_epi32(r89, _MM_SHUFFLE(2,2,2,2));
478 r9x = _mm_slli_epi32(_mm_shuffle_epi32(r89, _MM_SHUFFLE(3,3,3,2)), 1);
479 r9 = _mm_mul_epu32(r9, packed3819.v);
480 r9a = _mm_shuffle_epi32(r9, _MM_SHUFFLE(2,2,2,2));
481
482 m01 = _mm_mul_epu32(r01, r0);
483 m23 = _mm_mul_epu32(r23, r0a);
484 m45 = _mm_mul_epu32(r45, r0a);
485 m45 = _mm_add_epi64(m45, _mm_mul_epu32(r23, r2));
486 r23 = _mm_slli_epi32(r23, 1);
487 m67 = _mm_mul_epu32(r67, r0a);
488 m67 = _mm_add_epi64(m67, _mm_mul_epu32(r45, r2a));
489 m89 = _mm_mul_epu32(r89, r0a);
490 m89 = _mm_add_epi64(m89, _mm_mul_epu32(r67, r2a));
491 r67 = _mm_slli_epi32(r67, 1);
492 m89 = _mm_add_epi64(m89, _mm_mul_epu32(r45, r4));
493 r45 = _mm_slli_epi32(r45, 1);
494
495 r1 = _mm_slli_epi32(r1, 1);
496 r3 = _mm_slli_epi32(r3, 1);
497 r1a = _mm_add_epi64(r1, _mm_and_si128(r1, bot64bitmask.v));
498 r3a = _mm_add_epi64(r3, _mm_and_si128(r3, bot64bitmask.v));
499
500 m23 = _mm_add_epi64(m23, _mm_mul_epu32(r12, r1));
501 m45 = _mm_add_epi64(m45, _mm_mul_epu32(r34, r1a));
502 m67 = _mm_add_epi64(m67, _mm_mul_epu32(r56, r1a));
503 m67 = _mm_add_epi64(m67, _mm_mul_epu32(r34, r3));
504 r34 = _mm_slli_epi32(r34, 1);
505 m89 = _mm_add_epi64(m89, _mm_mul_epu32(r78, r1a));
506 r78 = _mm_slli_epi32(r78, 1);
507 m89 = _mm_add_epi64(m89, _mm_mul_epu32(r56, r3a));
508 r56 = _mm_slli_epi32(r56, 1);
509
510 m01 = _mm_add_epi64(m01, _mm_mul_epu32(_mm_slli_epi32(r12, 1), r9));
511 m01 = _mm_add_epi64(m01, _mm_mul_epu32(r34, r7));
512 m23 = _mm_add_epi64(m23, _mm_mul_epu32(r34, r9));
513 m01 = _mm_add_epi64(m01, _mm_mul_epu32(r56, r5));
514 m23 = _mm_add_epi64(m23, _mm_mul_epu32(r56, r7));
515 m45 = _mm_add_epi64(m45, _mm_mul_epu32(r56, r9));
516 m01 = _mm_add_epi64(m01, _mm_mul_epu32(r23, r8));
517 m01 = _mm_add_epi64(m01, _mm_mul_epu32(r45, r6));
518 m23 = _mm_add_epi64(m23, _mm_mul_epu32(r45, r8));
519 m23 = _mm_add_epi64(m23, _mm_mul_epu32(r6x, r6));
520 m45 = _mm_add_epi64(m45, _mm_mul_epu32(r78, r7a));
521 m67 = _mm_add_epi64(m67, _mm_mul_epu32(r78, r9));
522 m45 = _mm_add_epi64(m45, _mm_mul_epu32(r67, r8));
523 m67 = _mm_add_epi64(m67, _mm_mul_epu32(r8x, r8));
524 m89 = _mm_add_epi64(m89, _mm_mul_epu32(r9x, r9a));
525
526 r0 = _mm_unpacklo_epi64(m01, m45);
527 r1 = _mm_unpackhi_epi64(m01, m45);
528 r2 = _mm_unpacklo_epi64(m23, m67);
529 r3 = _mm_unpackhi_epi64(m23, m67);
530 r4 = _mm_unpacklo_epi64(m89, m89);
531 r5 = _mm_unpackhi_epi64(m89, m89);
532
533 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
534 c1 = _mm_srli_epi64(r1, 25); c2 = _mm_srli_epi64(r3, 25); r1 = _mm_and_si128(r1, packedmask25.v); r3 = _mm_and_si128(r3, packedmask25.v); r2 = _mm_add_epi64(r2, c1); r4 = _mm_add_epi64(r4, c2); c3 = _mm_slli_si128(c2, 8);
535 c1 = _mm_srli_epi64(r4, 26); r4 = _mm_and_si128(r4, packedmask26.v); r5 = _mm_add_epi64(r5, c1);
536 c1 = _mm_srli_epi64(r5, 25); r5 = _mm_and_si128(r5, packedmask25.v); r0 = _mm_add_epi64(r0, _mm_unpackhi_epi64(_mm_mul_epu32(c1, packednineteen.v), c3));
537 c1 = _mm_srli_epi64(r0, 26); c2 = _mm_srli_epi64(r2, 26); r0 = _mm_and_si128(r0, packedmask26.v); r2 = _mm_and_si128(r2, packedmask26.v); r1 = _mm_add_epi64(r1, c1); r3 = _mm_add_epi64(r3, c2);
538
539 r01 = _mm_unpacklo_epi64(r0, r1);
540 r45 = _mm_unpackhi_epi64(r0, r1);
541 r23 = _mm_unpacklo_epi64(r2, r3);
542 r67 = _mm_unpackhi_epi64(r2, r3);
543 r89 = _mm_unpackhi_epi64(r4, r5);
544 } while (--count);
545
546 r0123 = _mm_shuffle_epi32(r23, _MM_SHUFFLE(2,0,3,3));
547 r4567 = _mm_shuffle_epi32(r67, _MM_SHUFFLE(2,0,3,3));
548 r0123 = _mm_or_si128(r0123, _mm_shuffle_epi32(r01, _MM_SHUFFLE(3,3,2,0)));
549 r4567 = _mm_or_si128(r4567, _mm_shuffle_epi32(r45, _MM_SHUFFLE(3,3,2,0)));
550 r89 = _mm_shuffle_epi32(r89, _MM_SHUFFLE(3,3,2,0));
551
552 _mm_store_si128((xmmi*)r + 0, r0123);
553 _mm_store_si128((xmmi*)r + 1, r4567);
554 _mm_store_si128((xmmi*)r + 2, r89);
555}
556
557DONNA_INLINE static void
558curve25519_tangle32(packedelem32 *out, const bignum25519 x, const bignum25519 z) {
559 xmmi x0,x1,x2,z0,z1,z2;
560
561 x0 = _mm_load_si128((xmmi *)(x + 0));
562 x1 = _mm_load_si128((xmmi *)(x + 4));
563 x2 = _mm_load_si128((xmmi *)(x + 8));
564 z0 = _mm_load_si128((xmmi *)(z + 0));
565 z1 = _mm_load_si128((xmmi *)(z + 4));
566 z2 = _mm_load_si128((xmmi *)(z + 8));
567
568 out[0].v = _mm_unpacklo_epi32(x0, z0);
569 out[1].v = _mm_unpackhi_epi32(x0, z0);
570 out[2].v = _mm_unpacklo_epi32(x1, z1);
571 out[3].v = _mm_unpackhi_epi32(x1, z1);
572 out[4].v = _mm_unpacklo_epi32(x2, z2);
573}
574
575DONNA_INLINE static void
576curve25519_untangle32(bignum25519 x, bignum25519 z, const packedelem32 *in) {
577 xmmi t0,t1,t2,t3,t4,zero;
578
579 t0 = _mm_shuffle_epi32(in[0].v, _MM_SHUFFLE(3,1,2,0));
580 t1 = _mm_shuffle_epi32(in[1].v, _MM_SHUFFLE(3,1,2,0));
581 t2 = _mm_shuffle_epi32(in[2].v, _MM_SHUFFLE(3,1,2,0));
582 t3 = _mm_shuffle_epi32(in[3].v, _MM_SHUFFLE(3,1,2,0));
583 t4 = _mm_shuffle_epi32(in[4].v, _MM_SHUFFLE(3,1,2,0));
584 zero = _mm_setzero_si128();
585 _mm_store_si128((xmmi *)x + 0, _mm_unpacklo_epi64(t0, t1));
586 _mm_store_si128((xmmi *)x + 1, _mm_unpacklo_epi64(t2, t3));
587 _mm_store_si128((xmmi *)x + 2, _mm_unpacklo_epi64(t4, zero));
588 _mm_store_si128((xmmi *)z + 0, _mm_unpackhi_epi64(t0, t1));
589 _mm_store_si128((xmmi *)z + 1, _mm_unpackhi_epi64(t2, t3));
590 _mm_store_si128((xmmi *)z + 2, _mm_unpackhi_epi64(t4, zero));
591}
592
593DONNA_INLINE static void
594curve25519_add_reduce_packed32(packedelem32 *out, const packedelem32 *r, const packedelem32 *s) {
595 xmmi r0,r1,r2,r3,r4;
596 xmmi s0,s1,s2,s3,s4,s5;
597 xmmi c1,c2;
598
599 r0 = _mm_add_epi32(r[0].v, s[0].v);
600 r1 = _mm_add_epi32(r[1].v, s[1].v);
601 r2 = _mm_add_epi32(r[2].v, s[2].v);
602 r3 = _mm_add_epi32(r[3].v, s[3].v);
603 r4 = _mm_add_epi32(r[4].v, s[4].v);
604
605 s0 = _mm_unpacklo_epi64(r0, r2); /* 00 44 */
606 s1 = _mm_unpackhi_epi64(r0, r2); /* 11 55 */
607 s2 = _mm_unpacklo_epi64(r1, r3); /* 22 66 */
608 s3 = _mm_unpackhi_epi64(r1, r3); /* 33 77 */
609 s4 = _mm_unpacklo_epi64(_mm_setzero_si128(), r4); /* 00 88 */
610 s5 = _mm_unpackhi_epi64(_mm_setzero_si128(), r4); /* 00 99 */
611
612 c1 = _mm_srli_epi32(s0, 26); c2 = _mm_srli_epi32(s2, 26); s0 = _mm_and_si128(s0, packedmask26262626.v); s2 = _mm_and_si128(s2, packedmask26262626.v); s1 = _mm_add_epi32(s1, c1); s3 = _mm_add_epi32(s3, c2);
613 c1 = _mm_srli_epi32(s1, 25); c2 = _mm_srli_epi32(s3, 25); s1 = _mm_and_si128(s1, packedmask25252525.v); s3 = _mm_and_si128(s3, packedmask25252525.v); s2 = _mm_add_epi32(s2, c1); s4 = _mm_add_epi32(s4, _mm_unpackhi_epi64(_mm_setzero_si128(), c2)); s0 = _mm_add_epi32(s0, _mm_unpacklo_epi64(_mm_setzero_si128(), c2));
614 c1 = _mm_srli_epi32(s2, 26); c2 = _mm_srli_epi32(s4, 26); s2 = _mm_and_si128(s2, packedmask26262626.v); s4 = _mm_and_si128(s4, packedmask26262626.v); s3 = _mm_add_epi32(s3, c1); s5 = _mm_add_epi32(s5, c2);
615 c1 = _mm_srli_epi32(s3, 25); c2 = _mm_srli_epi32(s5, 25); s3 = _mm_and_si128(s3, packedmask25252525.v); s5 = _mm_and_si128(s5, packedmask25252525.v); s4 = _mm_add_epi32(s4, c1); s0 = _mm_add_epi32(s0, _mm_or_si128(_mm_slli_si128(c1, 8), _mm_srli_si128(_mm_add_epi32(_mm_add_epi32(_mm_slli_epi32(c2, 4), _mm_slli_epi32(c2, 1)), c2), 8)));
616 c1 = _mm_srli_epi32(s0, 26); c2 = _mm_srli_epi32(s2, 26); s0 = _mm_and_si128(s0, packedmask26262626.v); s2 = _mm_and_si128(s2, packedmask26262626.v); s1 = _mm_add_epi32(s1, c1); s3 = _mm_add_epi32(s3, c2);
617
618 out[0].v = _mm_unpacklo_epi64(s0, s1); /* 00 11 */
619 out[1].v = _mm_unpacklo_epi64(s2, s3); /* 22 33 */
620 out[2].v = _mm_unpackhi_epi64(s0, s1); /* 44 55 */
621 out[3].v = _mm_unpackhi_epi64(s2, s3); /* 66 77 */
622 out[4].v = _mm_unpackhi_epi64(s4, s5); /* 88 99 */
623}
624
625DONNA_INLINE static void
626curve25519_add_packed32(packedelem32 *out, const packedelem32 *r, const packedelem32 *s) {
627 out[0].v = _mm_add_epi32(r[0].v, s[0].v);
628 out[1].v = _mm_add_epi32(r[1].v, s[1].v);
629 out[2].v = _mm_add_epi32(r[2].v, s[2].v);
630 out[3].v = _mm_add_epi32(r[3].v, s[3].v);
631 out[4].v = _mm_add_epi32(r[4].v, s[4].v);
632}
633
634DONNA_INLINE static void
635curve25519_sub_packed32(packedelem32 *out, const packedelem32 *r, const packedelem32 *s) {
636 xmmi r0,r1,r2,r3,r4;
637 xmmi s0,s1,s2,s3;
638 xmmi c1,c2;
639
640 r0 = _mm_add_epi32(r[0].v, packed32packed2p0.v);
641 r1 = _mm_add_epi32(r[1].v, packed32packed2p1.v);
642 r2 = _mm_add_epi32(r[2].v, packed32packed2p1.v);
643 r3 = _mm_add_epi32(r[3].v, packed32packed2p1.v);
644 r4 = _mm_add_epi32(r[4].v, packed32packed2p1.v);
645 r0 = _mm_sub_epi32(r0, s[0].v); /* 00 11 */
646 r1 = _mm_sub_epi32(r1, s[1].v); /* 22 33 */
647 r2 = _mm_sub_epi32(r2, s[2].v); /* 44 55 */
648 r3 = _mm_sub_epi32(r3, s[3].v); /* 66 77 */
649 r4 = _mm_sub_epi32(r4, s[4].v); /* 88 99 */
650
651 s0 = _mm_unpacklo_epi64(r0, r2); /* 00 44 */
652 s1 = _mm_unpackhi_epi64(r0, r2); /* 11 55 */
653 s2 = _mm_unpacklo_epi64(r1, r3); /* 22 66 */
654 s3 = _mm_unpackhi_epi64(r1, r3); /* 33 77 */
655
656 c1 = _mm_srli_epi32(s0, 26); c2 = _mm_srli_epi32(s2, 26); s0 = _mm_and_si128(s0, packedmask26262626.v); s2 = _mm_and_si128(s2, packedmask26262626.v); s1 = _mm_add_epi32(s1, c1); s3 = _mm_add_epi32(s3, c2);
657 c1 = _mm_srli_epi32(s1, 25); c2 = _mm_srli_epi32(s3, 25); s1 = _mm_and_si128(s1, packedmask25252525.v); s3 = _mm_and_si128(s3, packedmask25252525.v); s2 = _mm_add_epi32(s2, c1); r4 = _mm_add_epi32(r4, _mm_srli_si128(c2, 8)); s0 = _mm_add_epi32(s0, _mm_slli_si128(c2, 8));
658
659 out[0].v = _mm_unpacklo_epi64(s0, s1); /* 00 11 */
660 out[1].v = _mm_unpacklo_epi64(s2, s3); /* 22 33 */
661 out[2].v = _mm_unpackhi_epi64(s0, s1); /* 44 55 */
662 out[3].v = _mm_unpackhi_epi64(s2, s3); /* 66 77 */
663 out[4].v = r4;
664}
665
666DONNA_INLINE static void
667curve25519_sub_after_basic_packed32(packedelem32 *out, const packedelem32 *r, const packedelem32 *s) {
668 xmmi r0,r1,r2,r3,r4;
669 xmmi s0,s1,s2,s3,s4,s5;
670 xmmi c1,c2;
671
672 r0 = _mm_add_epi32(r[0].v, packed32packed4p0.v);
673 r1 = _mm_add_epi32(r[1].v, packed32packed4p1.v);
674 r2 = _mm_add_epi32(r[2].v, packed32packed4p1.v);
675 r3 = _mm_add_epi32(r[3].v, packed32packed4p1.v);
676 r4 = _mm_add_epi32(r[4].v, packed32packed4p1.v);
677 r0 = _mm_sub_epi32(r0, s[0].v); /* 00 11 */
678 r1 = _mm_sub_epi32(r1, s[1].v); /* 22 33 */
679 r2 = _mm_sub_epi32(r2, s[2].v); /* 44 55 */
680 r3 = _mm_sub_epi32(r3, s[3].v); /* 66 77 */
681 r4 = _mm_sub_epi32(r4, s[4].v); /* 88 99 */
682
683 s0 = _mm_unpacklo_epi64(r0, r2); /* 00 44 */
684 s1 = _mm_unpackhi_epi64(r0, r2); /* 11 55 */
685 s2 = _mm_unpacklo_epi64(r1, r3); /* 22 66 */
686 s3 = _mm_unpackhi_epi64(r1, r3); /* 33 77 */
687 s4 = _mm_unpacklo_epi64(_mm_setzero_si128(), r4); /* 00 88 */
688 s5 = _mm_unpackhi_epi64(_mm_setzero_si128(), r4); /* 00 99 */
689
690 c1 = _mm_srli_epi32(s0, 26); c2 = _mm_srli_epi32(s2, 26); s0 = _mm_and_si128(s0, packedmask26262626.v); s2 = _mm_and_si128(s2, packedmask26262626.v); s1 = _mm_add_epi32(s1, c1); s3 = _mm_add_epi32(s3, c2);
691 c1 = _mm_srli_epi32(s1, 25); c2 = _mm_srli_epi32(s3, 25); s1 = _mm_and_si128(s1, packedmask25252525.v); s3 = _mm_and_si128(s3, packedmask25252525.v); s2 = _mm_add_epi32(s2, c1); s4 = _mm_add_epi32(s4, _mm_unpackhi_epi64(_mm_setzero_si128(), c2)); s0 = _mm_add_epi32(s0, _mm_unpacklo_epi64(_mm_setzero_si128(), c2));
692 c1 = _mm_srli_epi32(s2, 26); c2 = _mm_srli_epi32(s4, 26); s2 = _mm_and_si128(s2, packedmask26262626.v); s4 = _mm_and_si128(s4, packedmask26262626.v); s3 = _mm_add_epi32(s3, c1); s5 = _mm_add_epi32(s5, c2);
693 c1 = _mm_srli_epi32(s3, 25); c2 = _mm_srli_epi32(s5, 25); s3 = _mm_and_si128(s3, packedmask25252525.v); s5 = _mm_and_si128(s5, packedmask25252525.v); s4 = _mm_add_epi32(s4, c1); s0 = _mm_add_epi32(s0, _mm_or_si128(_mm_slli_si128(c1, 8), _mm_srli_si128(_mm_add_epi32(_mm_add_epi32(_mm_slli_epi32(c2, 4), _mm_slli_epi32(c2, 1)), c2), 8)));
694 c1 = _mm_srli_epi32(s0, 26); c2 = _mm_srli_epi32(s2, 26); s0 = _mm_and_si128(s0, packedmask26262626.v); s2 = _mm_and_si128(s2, packedmask26262626.v); s1 = _mm_add_epi32(s1, c1); s3 = _mm_add_epi32(s3, c2);
695
696 out[0].v = _mm_unpacklo_epi64(s0, s1); /* 00 11 */
697 out[1].v = _mm_unpacklo_epi64(s2, s3); /* 22 33 */
698 out[2].v = _mm_unpackhi_epi64(s0, s1); /* 44 55 */
699 out[3].v = _mm_unpackhi_epi64(s2, s3); /* 66 77 */
700 out[4].v = _mm_unpackhi_epi64(s4, s5); /* 88 99 */
701}
702
703DONNA_INLINE static void
704curve25519_tangle64_from32(packedelem64 *a, packedelem64 *b, const packedelem32 *c, const packedelem32 *d) {
705 xmmi c0,c1,c2,c3,c4,c5,t;
706 xmmi d0,d1,d2,d3,d4,d5;
707 xmmi t0,t1,t2,t3,t4,zero;
708
709 t0 = _mm_shuffle_epi32(c[0].v, _MM_SHUFFLE(3,1,2,0));
710 t1 = _mm_shuffle_epi32(c[1].v, _MM_SHUFFLE(3,1,2,0));
711 t2 = _mm_shuffle_epi32(d[0].v, _MM_SHUFFLE(3,1,2,0));
712 t3 = _mm_shuffle_epi32(d[1].v, _MM_SHUFFLE(3,1,2,0));
713 c0 = _mm_unpacklo_epi64(t0, t1);
714 c3 = _mm_unpackhi_epi64(t0, t1);
715 d0 = _mm_unpacklo_epi64(t2, t3);
716 d3 = _mm_unpackhi_epi64(t2, t3);
717 t = _mm_unpacklo_epi64(c0, d0); a[0].v = t; a[1].v = _mm_srli_epi64(t, 32);
718 t = _mm_unpackhi_epi64(c0, d0); a[2].v = t; a[3].v = _mm_srli_epi64(t, 32);
719 t = _mm_unpacklo_epi64(c3, d3); b[0].v = t; b[1].v = _mm_srli_epi64(t, 32);
720 t = _mm_unpackhi_epi64(c3, d3); b[2].v = t; b[3].v = _mm_srli_epi64(t, 32);
721
722 t0 = _mm_shuffle_epi32(c[2].v, _MM_SHUFFLE(3,1,2,0));
723 t1 = _mm_shuffle_epi32(c[3].v, _MM_SHUFFLE(3,1,2,0));
724 t2 = _mm_shuffle_epi32(d[2].v, _MM_SHUFFLE(3,1,2,0));
725 t3 = _mm_shuffle_epi32(d[3].v, _MM_SHUFFLE(3,1,2,0));
726 c1 = _mm_unpacklo_epi64(t0, t1);
727 c4 = _mm_unpackhi_epi64(t0, t1);
728 d1 = _mm_unpacklo_epi64(t2, t3);
729 d4 = _mm_unpackhi_epi64(t2, t3);
730 t = _mm_unpacklo_epi64(c1, d1); a[4].v = t; a[5].v = _mm_srli_epi64(t, 32);
731 t = _mm_unpackhi_epi64(c1, d1); a[6].v = t; a[7].v = _mm_srli_epi64(t, 32);
732 t = _mm_unpacklo_epi64(c4, d4); b[4].v = t; b[5].v = _mm_srli_epi64(t, 32);
733 t = _mm_unpackhi_epi64(c4, d4); b[6].v = t; b[7].v = _mm_srli_epi64(t, 32);
734
735 t4 = _mm_shuffle_epi32(c[4].v, _MM_SHUFFLE(3,1,2,0));
736 zero = _mm_setzero_si128();
737 c2 = _mm_unpacklo_epi64(t4, zero);
738 c5 = _mm_unpackhi_epi64(t4, zero);
739 t4 = _mm_shuffle_epi32(d[4].v, _MM_SHUFFLE(3,1,2,0));
740 d2 = _mm_unpacklo_epi64(t4, zero);
741 d5 = _mm_unpackhi_epi64(t4, zero);
742 t = _mm_unpacklo_epi64(c2, d2); a[8].v = t; a[9].v = _mm_srli_epi64(t, 32);
743 t = _mm_unpacklo_epi64(c5, d5); b[8].v = t; b[9].v = _mm_srli_epi64(t, 32);
744}
745
746DONNA_INLINE static void
747curve25519_tangle64(packedelem64 *out, const bignum25519 x, const bignum25519 z) {
748 xmmi x0,x1,x2,z0,z1,z2,t;
749
750 x0 = _mm_load_si128((xmmi *)x + 0);
751 x1 = _mm_load_si128((xmmi *)x + 1);
752 x2 = _mm_load_si128((xmmi *)x + 2);
753 z0 = _mm_load_si128((xmmi *)z + 0);
754 z1 = _mm_load_si128((xmmi *)z + 1);
755 z2 = _mm_load_si128((xmmi *)z + 2);
756
757 t = _mm_unpacklo_epi64(x0, z0); out[0].v = t; out[1].v = _mm_srli_epi64(t, 32);
758 t = _mm_unpackhi_epi64(x0, z0); out[2].v = t; out[3].v = _mm_srli_epi64(t, 32);
759 t = _mm_unpacklo_epi64(x1, z1); out[4].v = t; out[5].v = _mm_srli_epi64(t, 32);
760 t = _mm_unpackhi_epi64(x1, z1); out[6].v = t; out[7].v = _mm_srli_epi64(t, 32);
761 t = _mm_unpacklo_epi64(x2, z2); out[8].v = t; out[9].v = _mm_srli_epi64(t, 32);
762}
763
764DONNA_INLINE static void
765curve25519_tangleone64(packedelem64 *out, const bignum25519 x) {
766 xmmi x0,x1,x2;
767
768 x0 = _mm_load_si128((xmmi *)(x + 0));
769 x1 = _mm_load_si128((xmmi *)(x + 4));
770 x2 = _mm_load_si128((xmmi *)(x + 8));
771
772 out[0].v = _mm_shuffle_epi32(x0, _MM_SHUFFLE(0,0,0,0));
773 out[1].v = _mm_shuffle_epi32(x0, _MM_SHUFFLE(1,1,1,1));
774 out[2].v = _mm_shuffle_epi32(x0, _MM_SHUFFLE(2,2,2,2));
775 out[3].v = _mm_shuffle_epi32(x0, _MM_SHUFFLE(3,3,3,3));
776 out[4].v = _mm_shuffle_epi32(x1, _MM_SHUFFLE(0,0,0,0));
777 out[5].v = _mm_shuffle_epi32(x1, _MM_SHUFFLE(1,1,1,1));
778 out[6].v = _mm_shuffle_epi32(x1, _MM_SHUFFLE(2,2,2,2));
779 out[7].v = _mm_shuffle_epi32(x1, _MM_SHUFFLE(3,3,3,3));
780 out[8].v = _mm_shuffle_epi32(x2, _MM_SHUFFLE(0,0,0,0));
781 out[9].v = _mm_shuffle_epi32(x2, _MM_SHUFFLE(1,1,1,1));
782}
783
784DONNA_INLINE static void
785curve25519_swap64(packedelem64 *out) {
786 out[0].v = _mm_shuffle_epi32(out[0].v, _MM_SHUFFLE(1,0,3,2));
787 out[1].v = _mm_shuffle_epi32(out[1].v, _MM_SHUFFLE(1,0,3,2));
788 out[2].v = _mm_shuffle_epi32(out[2].v, _MM_SHUFFLE(1,0,3,2));
789 out[3].v = _mm_shuffle_epi32(out[3].v, _MM_SHUFFLE(1,0,3,2));
790 out[4].v = _mm_shuffle_epi32(out[4].v, _MM_SHUFFLE(1,0,3,2));
791 out[5].v = _mm_shuffle_epi32(out[5].v, _MM_SHUFFLE(1,0,3,2));
792 out[6].v = _mm_shuffle_epi32(out[6].v, _MM_SHUFFLE(1,0,3,2));
793 out[7].v = _mm_shuffle_epi32(out[7].v, _MM_SHUFFLE(1,0,3,2));
794 out[8].v = _mm_shuffle_epi32(out[8].v, _MM_SHUFFLE(1,0,3,2));
795 out[9].v = _mm_shuffle_epi32(out[9].v, _MM_SHUFFLE(1,0,3,2));
796}
797
798DONNA_INLINE static void
799curve25519_untangle64(bignum25519 x, bignum25519 z, const packedelem64 *in) {
800 _mm_store_si128((xmmi *)(x + 0), _mm_unpacklo_epi64(_mm_unpacklo_epi32(in[0].v, in[1].v), _mm_unpacklo_epi32(in[2].v, in[3].v)));
801 _mm_store_si128((xmmi *)(x + 4), _mm_unpacklo_epi64(_mm_unpacklo_epi32(in[4].v, in[5].v), _mm_unpacklo_epi32(in[6].v, in[7].v)));
802 _mm_store_si128((xmmi *)(x + 8), _mm_unpacklo_epi32(in[8].v, in[9].v) );
803 _mm_store_si128((xmmi *)(z + 0), _mm_unpacklo_epi64(_mm_unpackhi_epi32(in[0].v, in[1].v), _mm_unpackhi_epi32(in[2].v, in[3].v)));
804 _mm_store_si128((xmmi *)(z + 4), _mm_unpacklo_epi64(_mm_unpackhi_epi32(in[4].v, in[5].v), _mm_unpackhi_epi32(in[6].v, in[7].v)));
805 _mm_store_si128((xmmi *)(z + 8), _mm_unpackhi_epi32(in[8].v, in[9].v) );
806}
807
808DONNA_INLINE static void
809curve25519_mul_packed64(packedelem64 *out, const packedelem64 *r, const packedelem64 *s) {
810 xmmi r1,r2,r3,r4,r5,r6,r7,r8,r9;
811 xmmi r1_2,r3_2,r5_2,r7_2,r9_2;
812 xmmi c1,c2;
813
814 out[0].v = _mm_mul_epu32(r[0].v, s[0].v);
815 out[1].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[1].v), _mm_mul_epu32(r[1].v, s[0].v));
816 r1_2 = _mm_slli_epi32(r[1].v, 1);
817 out[2].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r1_2 , s[1].v), _mm_mul_epu32(r[2].v, s[0].v)));
818 out[3].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[3].v), _mm_add_epi64(_mm_mul_epu32(r[1].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r[2].v, s[1].v), _mm_mul_epu32(r[3].v, s[0].v))));
819 r3_2 = _mm_slli_epi32(r[3].v, 1);
820 out[4].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[4].v), _mm_add_epi64(_mm_mul_epu32(r1_2 , s[3].v), _mm_add_epi64(_mm_mul_epu32(r[2].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r3_2 , s[1].v), _mm_mul_epu32(r[4].v, s[0].v)))));
821 out[5].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[5].v), _mm_add_epi64(_mm_mul_epu32(r[1].v, s[4].v), _mm_add_epi64(_mm_mul_epu32(r[2].v, s[3].v), _mm_add_epi64(_mm_mul_epu32(r[3].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r[4].v, s[1].v), _mm_mul_epu32(r[5].v, s[0].v))))));
822 r5_2 = _mm_slli_epi32(r[5].v, 1);
823 out[6].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[6].v), _mm_add_epi64(_mm_mul_epu32(r1_2 , s[5].v), _mm_add_epi64(_mm_mul_epu32(r[2].v, s[4].v), _mm_add_epi64(_mm_mul_epu32(r3_2 , s[3].v), _mm_add_epi64(_mm_mul_epu32(r[4].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r5_2 , s[1].v), _mm_mul_epu32(r[6].v, s[0].v)))))));
824 out[7].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[7].v), _mm_add_epi64(_mm_mul_epu32(r[1].v, s[6].v), _mm_add_epi64(_mm_mul_epu32(r[2].v, s[5].v), _mm_add_epi64(_mm_mul_epu32(r[3].v, s[4].v), _mm_add_epi64(_mm_mul_epu32(r[4].v, s[3].v), _mm_add_epi64(_mm_mul_epu32(r[5].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r[6].v, s[1].v), _mm_mul_epu32(r[7].v , s[0].v))))))));
825 r7_2 = _mm_slli_epi32(r[7].v, 1);
826 out[8].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[8].v), _mm_add_epi64(_mm_mul_epu32(r1_2 , s[7].v), _mm_add_epi64(_mm_mul_epu32(r[2].v, s[6].v), _mm_add_epi64(_mm_mul_epu32(r3_2 , s[5].v), _mm_add_epi64(_mm_mul_epu32(r[4].v, s[4].v), _mm_add_epi64(_mm_mul_epu32(r5_2 , s[3].v), _mm_add_epi64(_mm_mul_epu32(r[6].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r7_2 , s[1].v), _mm_mul_epu32(r[8].v, s[0].v)))))))));
827 out[9].v = _mm_add_epi64(_mm_mul_epu32(r[0].v, s[9].v), _mm_add_epi64(_mm_mul_epu32(r[1].v, s[8].v), _mm_add_epi64(_mm_mul_epu32(r[2].v, s[7].v), _mm_add_epi64(_mm_mul_epu32(r[3].v, s[6].v), _mm_add_epi64(_mm_mul_epu32(r[4].v, s[5].v), _mm_add_epi64(_mm_mul_epu32(r[5].v, s[4].v), _mm_add_epi64(_mm_mul_epu32(r[6].v, s[3].v), _mm_add_epi64(_mm_mul_epu32(r[7].v, s[2].v), _mm_add_epi64(_mm_mul_epu32(r[8].v, s[1].v), _mm_mul_epu32(r[9].v, s[0].v))))))))));
828
829 r1 = _mm_mul_epu32(r[1].v, packednineteen.v);
830 r2 = _mm_mul_epu32(r[2].v, packednineteen.v);
831 r1_2 = _mm_slli_epi32(r1, 1);
832 r3 = _mm_mul_epu32(r[3].v, packednineteen.v);
833 r4 = _mm_mul_epu32(r[4].v, packednineteen.v);
834 r3_2 = _mm_slli_epi32(r3, 1);
835 r5 = _mm_mul_epu32(r[5].v, packednineteen.v);
836 r6 = _mm_mul_epu32(r[6].v, packednineteen.v);
837 r5_2 = _mm_slli_epi32(r5, 1);
838 r7 = _mm_mul_epu32(r[7].v, packednineteen.v);
839 r8 = _mm_mul_epu32(r[8].v, packednineteen.v);
840 r7_2 = _mm_slli_epi32(r7, 1);
841 r9 = _mm_mul_epu32(r[9].v, packednineteen.v);
842 r9_2 = _mm_slli_epi32(r9, 1);
843
844 out[0].v = _mm_add_epi64(out[0].v, _mm_add_epi64(_mm_mul_epu32(r9_2, s[1].v), _mm_add_epi64(_mm_mul_epu32(r8, s[2].v), _mm_add_epi64(_mm_mul_epu32(r7_2, s[3].v), _mm_add_epi64(_mm_mul_epu32(r6, s[4].v), _mm_add_epi64(_mm_mul_epu32(r5_2, s[5].v), _mm_add_epi64(_mm_mul_epu32(r4, s[6].v), _mm_add_epi64(_mm_mul_epu32(r3_2, s[7].v), _mm_add_epi64(_mm_mul_epu32(r2, s[8].v), _mm_mul_epu32(r1_2, s[9].v))))))))));
845 out[1].v = _mm_add_epi64(out[1].v, _mm_add_epi64(_mm_mul_epu32(r9 , s[2].v), _mm_add_epi64(_mm_mul_epu32(r8, s[3].v), _mm_add_epi64(_mm_mul_epu32(r7 , s[4].v), _mm_add_epi64(_mm_mul_epu32(r6, s[5].v), _mm_add_epi64(_mm_mul_epu32(r5 , s[6].v), _mm_add_epi64(_mm_mul_epu32(r4, s[7].v), _mm_add_epi64(_mm_mul_epu32(r3 , s[8].v), _mm_mul_epu32(r2, s[9].v)))))))));
846 out[2].v = _mm_add_epi64(out[2].v, _mm_add_epi64(_mm_mul_epu32(r9_2, s[3].v), _mm_add_epi64(_mm_mul_epu32(r8, s[4].v), _mm_add_epi64(_mm_mul_epu32(r7_2, s[5].v), _mm_add_epi64(_mm_mul_epu32(r6, s[6].v), _mm_add_epi64(_mm_mul_epu32(r5_2, s[7].v), _mm_add_epi64(_mm_mul_epu32(r4, s[8].v), _mm_mul_epu32(r3_2, s[9].v))))))));
847 out[3].v = _mm_add_epi64(out[3].v, _mm_add_epi64(_mm_mul_epu32(r9 , s[4].v), _mm_add_epi64(_mm_mul_epu32(r8, s[5].v), _mm_add_epi64(_mm_mul_epu32(r7 , s[6].v), _mm_add_epi64(_mm_mul_epu32(r6, s[7].v), _mm_add_epi64(_mm_mul_epu32(r5 , s[8].v), _mm_mul_epu32(r4, s[9].v)))))));
848 out[4].v = _mm_add_epi64(out[4].v, _mm_add_epi64(_mm_mul_epu32(r9_2, s[5].v), _mm_add_epi64(_mm_mul_epu32(r8, s[6].v), _mm_add_epi64(_mm_mul_epu32(r7_2, s[7].v), _mm_add_epi64(_mm_mul_epu32(r6, s[8].v), _mm_mul_epu32(r5_2, s[9].v))))));
849 out[5].v = _mm_add_epi64(out[5].v, _mm_add_epi64(_mm_mul_epu32(r9 , s[6].v), _mm_add_epi64(_mm_mul_epu32(r8, s[7].v), _mm_add_epi64(_mm_mul_epu32(r7 , s[8].v), _mm_mul_epu32(r6, s[9].v)))));
850 out[6].v = _mm_add_epi64(out[6].v, _mm_add_epi64(_mm_mul_epu32(r9_2, s[7].v), _mm_add_epi64(_mm_mul_epu32(r8, s[8].v), _mm_mul_epu32(r7_2, s[9].v))));
851 out[7].v = _mm_add_epi64(out[7].v, _mm_add_epi64(_mm_mul_epu32(r9 , s[8].v), _mm_mul_epu32(r8, s[9].v)));
852 out[8].v = _mm_add_epi64(out[8].v, _mm_mul_epu32(r9_2, s[9].v));
853
854 c1 = _mm_srli_epi64(out[0].v, 26); c2 = _mm_srli_epi64(out[4].v, 26); out[0].v = _mm_and_si128(out[0].v, packedmask26.v); out[4].v = _mm_and_si128(out[4].v, packedmask26.v); out[1].v = _mm_add_epi64(out[1].v, c1); out[5].v = _mm_add_epi64(out[5].v, c2);
855 c1 = _mm_srli_epi64(out[1].v, 25); c2 = _mm_srli_epi64(out[5].v, 25); out[1].v = _mm_and_si128(out[1].v, packedmask25.v); out[5].v = _mm_and_si128(out[5].v, packedmask25.v); out[2].v = _mm_add_epi64(out[2].v, c1); out[6].v = _mm_add_epi64(out[6].v, c2);
856 c1 = _mm_srli_epi64(out[2].v, 26); c2 = _mm_srli_epi64(out[6].v, 26); out[2].v = _mm_and_si128(out[2].v, packedmask26.v); out[6].v = _mm_and_si128(out[6].v, packedmask26.v); out[3].v = _mm_add_epi64(out[3].v, c1); out[7].v = _mm_add_epi64(out[7].v, c2);
857 c1 = _mm_srli_epi64(out[3].v, 25); c2 = _mm_srli_epi64(out[7].v, 25); out[3].v = _mm_and_si128(out[3].v, packedmask25.v); out[7].v = _mm_and_si128(out[7].v, packedmask25.v); out[4].v = _mm_add_epi64(out[4].v, c1); out[8].v = _mm_add_epi64(out[8].v, c2);
858 c2 = _mm_srli_epi64(out[8].v, 26); out[8].v = _mm_and_si128(out[8].v, packedmask26.v); out[9].v = _mm_add_epi64(out[9].v, c2);
859 c2 = _mm_srli_epi64(out[9].v, 25); out[9].v = _mm_and_si128(out[9].v, packedmask25.v); out[0].v = _mm_add_epi64(out[0].v, _mm_mul_epu32(c2, packednineteen.v));
860 c1 = _mm_srli_epi64(out[0].v, 26); c2 = _mm_srli_epi64(out[4].v, 26); out[0].v = _mm_and_si128(out[0].v, packedmask26.v); out[4].v = _mm_and_si128(out[4].v, packedmask26.v); out[1].v = _mm_add_epi64(out[1].v, c1); out[5].v = _mm_add_epi64(out[5].v, c2);
861}
862
863DONNA_INLINE static void
864curve25519_square_packed64(packedelem64 *out, const packedelem64 *r) {
865 xmmi r0,r1,r2,r3;
866 xmmi r1_2,r3_2,r4_2,r5_2,r6_2,r7_2;
867 xmmi d5,d6,d7,d8,d9;
868 xmmi c1,c2;
869
870 r0 = r[0].v;
871 r1 = r[1].v;
872 r2 = r[2].v;
873 r3 = r[3].v;
874
875 out[0].v = _mm_mul_epu32(r0, r0);
876 r0 = _mm_slli_epi32(r0, 1);
877 out[1].v = _mm_mul_epu32(r0, r1);
878 r1_2 = _mm_slli_epi32(r1, 1);
879 out[2].v = _mm_add_epi64(_mm_mul_epu32(r0, r2 ), _mm_mul_epu32(r1, r1_2));
880 r1 = r1_2;
881 out[3].v = _mm_add_epi64(_mm_mul_epu32(r0, r3 ), _mm_mul_epu32(r1, r2 ));
882 r3_2 = _mm_slli_epi32(r3, 1);
883 out[4].v = _mm_add_epi64(_mm_mul_epu32(r0, r[4].v), _mm_add_epi64(_mm_mul_epu32(r1, r3_2 ), _mm_mul_epu32(r2, r2)));
884 r2 = _mm_slli_epi32(r2, 1);
885 out[5].v = _mm_add_epi64(_mm_mul_epu32(r0, r[5].v), _mm_add_epi64(_mm_mul_epu32(r1, r[4].v), _mm_mul_epu32(r2, r3)));
886 r5_2 = _mm_slli_epi32(r[5].v, 1);
887 out[6].v = _mm_add_epi64(_mm_mul_epu32(r0, r[6].v), _mm_add_epi64(_mm_mul_epu32(r1, r5_2 ), _mm_add_epi64(_mm_mul_epu32(r2, r[4].v), _mm_mul_epu32(r3, r3_2 ))));
888 r3 = r3_2;
889 out[7].v = _mm_add_epi64(_mm_mul_epu32(r0, r[7].v), _mm_add_epi64(_mm_mul_epu32(r1, r[6].v), _mm_add_epi64(_mm_mul_epu32(r2, r[5].v), _mm_mul_epu32(r3, r[4].v))));
890 r7_2 = _mm_slli_epi32(r[7].v, 1);
891 out[8].v = _mm_add_epi64(_mm_mul_epu32(r0, r[8].v), _mm_add_epi64(_mm_mul_epu32(r1, r7_2 ), _mm_add_epi64(_mm_mul_epu32(r2, r[6].v), _mm_add_epi64(_mm_mul_epu32(r3, r5_2 ), _mm_mul_epu32(r[4].v, r[4].v)))));
892 out[9].v = _mm_add_epi64(_mm_mul_epu32(r0, r[9].v), _mm_add_epi64(_mm_mul_epu32(r1, r[8].v), _mm_add_epi64(_mm_mul_epu32(r2, r[7].v), _mm_add_epi64(_mm_mul_epu32(r3, r[6].v), _mm_mul_epu32(r[4].v, r5_2 )))));
893
894 d5 = _mm_mul_epu32(r[5].v, packedthirtyeight.v);
895 d6 = _mm_mul_epu32(r[6].v, packednineteen.v);
896 d7 = _mm_mul_epu32(r[7].v, packedthirtyeight.v);
897 d8 = _mm_mul_epu32(r[8].v, packednineteen.v);
898 d9 = _mm_mul_epu32(r[9].v, packedthirtyeight.v);
899
900 r4_2 = _mm_slli_epi32(r[4].v, 1);
901 r6_2 = _mm_slli_epi32(r[6].v, 1);
902 out[0].v = _mm_add_epi64(out[0].v, _mm_add_epi64(_mm_mul_epu32(d9, r1 ), _mm_add_epi64(_mm_mul_epu32(d8, r2 ), _mm_add_epi64(_mm_mul_epu32(d7, r3 ), _mm_add_epi64(_mm_mul_epu32(d6, r4_2), _mm_mul_epu32(d5, r[5].v))))));
903 out[1].v = _mm_add_epi64(out[1].v, _mm_add_epi64(_mm_mul_epu32(d9, _mm_srli_epi32(r2, 1)), _mm_add_epi64(_mm_mul_epu32(d8, r3 ), _mm_add_epi64(_mm_mul_epu32(d7, r[4].v), _mm_mul_epu32(d6, r5_2 )))));
904 out[2].v = _mm_add_epi64(out[2].v, _mm_add_epi64(_mm_mul_epu32(d9, r3 ), _mm_add_epi64(_mm_mul_epu32(d8, r4_2), _mm_add_epi64(_mm_mul_epu32(d7, r5_2 ), _mm_mul_epu32(d6, r[6].v)))));
905 out[3].v = _mm_add_epi64(out[3].v, _mm_add_epi64(_mm_mul_epu32(d9, r[4].v ), _mm_add_epi64(_mm_mul_epu32(d8, r5_2), _mm_mul_epu32(d7, r[6].v))));
906 out[4].v = _mm_add_epi64(out[4].v, _mm_add_epi64(_mm_mul_epu32(d9, r5_2 ), _mm_add_epi64(_mm_mul_epu32(d8, r6_2), _mm_mul_epu32(d7, r[7].v))));
907 out[5].v = _mm_add_epi64(out[5].v, _mm_add_epi64(_mm_mul_epu32(d9, r[6].v ), _mm_mul_epu32(d8, r7_2 )));
908 out[6].v = _mm_add_epi64(out[6].v, _mm_add_epi64(_mm_mul_epu32(d9, r7_2 ), _mm_mul_epu32(d8, r[8].v)));
909 out[7].v = _mm_add_epi64(out[7].v, _mm_mul_epu32(d9, r[8].v));
910 out[8].v = _mm_add_epi64(out[8].v, _mm_mul_epu32(d9, r[9].v));
911
912 c1 = _mm_srli_epi64(out[0].v, 26); c2 = _mm_srli_epi64(out[4].v, 26); out[0].v = _mm_and_si128(out[0].v, packedmask26.v); out[4].v = _mm_and_si128(out[4].v, packedmask26.v); out[1].v = _mm_add_epi64(out[1].v, c1); out[5].v = _mm_add_epi64(out[5].v, c2);
913 c1 = _mm_srli_epi64(out[1].v, 25); c2 = _mm_srli_epi64(out[5].v, 25); out[1].v = _mm_and_si128(out[1].v, packedmask25.v); out[5].v = _mm_and_si128(out[5].v, packedmask25.v); out[2].v = _mm_add_epi64(out[2].v, c1); out[6].v = _mm_add_epi64(out[6].v, c2);
914 c1 = _mm_srli_epi64(out[2].v, 26); c2 = _mm_srli_epi64(out[6].v, 26); out[2].v = _mm_and_si128(out[2].v, packedmask26.v); out[6].v = _mm_and_si128(out[6].v, packedmask26.v); out[3].v = _mm_add_epi64(out[3].v, c1); out[7].v = _mm_add_epi64(out[7].v, c2);
915 c1 = _mm_srli_epi64(out[3].v, 25); c2 = _mm_srli_epi64(out[7].v, 25); out[3].v = _mm_and_si128(out[3].v, packedmask25.v); out[7].v = _mm_and_si128(out[7].v, packedmask25.v); out[4].v = _mm_add_epi64(out[4].v, c1); out[8].v = _mm_add_epi64(out[8].v, c2);
916 c2 = _mm_srli_epi64(out[8].v, 26); out[8].v = _mm_and_si128(out[8].v, packedmask26.v); out[9].v = _mm_add_epi64(out[9].v, c2);
917 c2 = _mm_srli_epi64(out[9].v, 25); out[9].v = _mm_and_si128(out[9].v, packedmask25.v); out[0].v = _mm_add_epi64(out[0].v, _mm_mul_epu32(c2, packednineteen.v));
918 c1 = _mm_srli_epi64(out[0].v, 26); c2 = _mm_srli_epi64(out[4].v, 26); out[0].v = _mm_and_si128(out[0].v, packedmask26.v); out[4].v = _mm_and_si128(out[4].v, packedmask26.v); out[1].v = _mm_add_epi64(out[1].v, c1); out[5].v = _mm_add_epi64(out[5].v, c2);
919}
920
921
922/* Take a little-endian, 32-byte number and expand it into polynomial form */
923static void
924curve25519_expand(bignum25519 out, const unsigned char in[32]) {
925 uint32_t x0,x1,x2,x3,x4,x5,x6,x7;
926
927 x0 = *(uint32_t *)(in + 0);
928 x1 = *(uint32_t *)(in + 4);
929 x2 = *(uint32_t *)(in + 8);
930 x3 = *(uint32_t *)(in + 12);
931 x4 = *(uint32_t *)(in + 16);
932 x5 = *(uint32_t *)(in + 20);
933 x6 = *(uint32_t *)(in + 24);
934 x7 = *(uint32_t *)(in + 28);
935
936 out[0] = ( x0 ) & 0x3ffffff;
937 out[1] = ((((uint64_t)x1 << 32) | x0) >> 26) & 0x1ffffff;
938 out[2] = ((((uint64_t)x2 << 32) | x1) >> 19) & 0x3ffffff;
939 out[3] = ((((uint64_t)x3 << 32) | x2) >> 13) & 0x1ffffff;
940 out[4] = (( x3) >> 6) & 0x3ffffff;
941 out[5] = ( x4 ) & 0x1ffffff;
942 out[6] = ((((uint64_t)x5 << 32) | x4) >> 25) & 0x3ffffff;
943 out[7] = ((((uint64_t)x6 << 32) | x5) >> 19) & 0x1ffffff;
944 out[8] = ((((uint64_t)x7 << 32) | x6) >> 12) & 0x3ffffff;
945 out[9] = (( x7) >> 6) & 0x1ffffff;
946 out[10] = 0;
947 out[11] = 0;
948}
949
950/* Take a fully reduced polynomial form number and contract it into a
951 * little-endian, 32-byte array
952 */
953static void
954curve25519_contract(unsigned char out[32], const bignum25519 in) {
955 bignum25519 ALIGN(16) f;
956 curve25519_copy(f, in);
957
958 #define carry_pass() \
959 f[1] += f[0] >> 26; f[0] &= 0x3ffffff; \
960 f[2] += f[1] >> 25; f[1] &= 0x1ffffff; \
961 f[3] += f[2] >> 26; f[2] &= 0x3ffffff; \
962 f[4] += f[3] >> 25; f[3] &= 0x1ffffff; \
963 f[5] += f[4] >> 26; f[4] &= 0x3ffffff; \
964 f[6] += f[5] >> 25; f[5] &= 0x1ffffff; \
965 f[7] += f[6] >> 26; f[6] &= 0x3ffffff; \
966 f[8] += f[7] >> 25; f[7] &= 0x1ffffff; \
967 f[9] += f[8] >> 26; f[8] &= 0x3ffffff;
968
969 #define carry_pass_full() \
970 carry_pass() \
971 f[0] += 19 * (f[9] >> 25); f[9] &= 0x1ffffff;
972
973 #define carry_pass_final() \
974 carry_pass() \
975 f[9] &= 0x1ffffff;
976
977 carry_pass_full()
978 carry_pass_full()
979
980 /* now t is between 0 and 2^255-1, properly carried. */
981 /* case 1: between 0 and 2^255-20. case 2: between 2^255-19 and 2^255-1. */
982 f[0] += 19;
983 carry_pass_full()
984
985 /* now between 19 and 2^255-1 in both cases, and offset by 19. */
986 f[0] += (1 << 26) - 19;
987 f[1] += (1 << 25) - 1;
988 f[2] += (1 << 26) - 1;
989 f[3] += (1 << 25) - 1;
990 f[4] += (1 << 26) - 1;
991 f[5] += (1 << 25) - 1;
992 f[6] += (1 << 26) - 1;
993 f[7] += (1 << 25) - 1;
994 f[8] += (1 << 26) - 1;
995 f[9] += (1 << 25) - 1;
996
997 /* now between 2^255 and 2^256-20, and offset by 2^255. */
998 carry_pass_final()
999
1000 #undef carry_pass
1001 #undef carry_full
1002 #undef carry_final
1003
1004 f[1] <<= 2;
1005 f[2] <<= 3;
1006 f[3] <<= 5;
1007 f[4] <<= 6;
1008 f[6] <<= 1;
1009 f[7] <<= 3;
1010 f[8] <<= 4;
1011 f[9] <<= 6;
1012
1013 #define F(i, s) \
1014 out[s+0] |= (unsigned char )(f[i] & 0xff); \
1015 out[s+1] = (unsigned char )((f[i] >> 8) & 0xff); \
1016 out[s+2] = (unsigned char )((f[i] >> 16) & 0xff); \
1017 out[s+3] = (unsigned char )((f[i] >> 24) & 0xff);
1018
1019 out[0] = 0;
1020 out[16] = 0;
1021 F(0,0);
1022 F(1,3);
1023 F(2,6);
1024 F(3,9);
1025 F(4,12);
1026 F(5,16);
1027 F(6,19);
1028 F(7,22);
1029 F(8,25);
1030 F(9,28);
1031 #undef F
1032}
1033
1034/* if (iswap) swap(a, b) */
1035DONNA_INLINE static void
1036curve25519_swap_conditional(bignum25519 a, bignum25519 b, uint32_t iswap) {
1037 const uint32_t swap = (uint32_t)(-(int32_t)iswap);
1038 xmmi a0,a1,a2,b0,b1,b2,x0,x1,x2;
1039 xmmi mask = _mm_cvtsi32_si128(swap);
1040 mask = _mm_shuffle_epi32(mask, 0);
1041 a0 = _mm_load_si128((xmmi *)a + 0);
1042 a1 = _mm_load_si128((xmmi *)a + 1);
1043 b0 = _mm_load_si128((xmmi *)b + 0);
1044 b1 = _mm_load_si128((xmmi *)b + 1);
1045 b0 = _mm_xor_si128(a0, b0);
1046 b1 = _mm_xor_si128(a1, b1);
1047 x0 = _mm_and_si128(b0, mask);
1048 x1 = _mm_and_si128(b1, mask);
1049 x0 = _mm_xor_si128(x0, a0);
1050 x1 = _mm_xor_si128(x1, a1);
1051 a0 = _mm_xor_si128(x0, b0);
1052 a1 = _mm_xor_si128(x1, b1);
1053 _mm_store_si128((xmmi *)a + 0, x0);
1054 _mm_store_si128((xmmi *)a + 1, x1);
1055 _mm_store_si128((xmmi *)b + 0, a0);
1056 _mm_store_si128((xmmi *)b + 1, a1);
1057
1058 a2 = _mm_load_si128((xmmi *)a + 2);
1059 b2 = _mm_load_si128((xmmi *)b + 2);
1060 b2 = _mm_xor_si128(a2, b2);
1061 x2 = _mm_and_si128(b2, mask);
1062 x2 = _mm_xor_si128(x2, a2);
1063 a2 = _mm_xor_si128(x2, b2);
1064 _mm_store_si128((xmmi *)b + 2, a2);
1065 _mm_store_si128((xmmi *)a + 2, x2);
1066}
1067
1068/* out = (flag) ? out : in */
1069DONNA_INLINE static void
1070curve25519_move_conditional_bytes(uint8_t out[96], const uint8_t in[96], uint32_t flag) {
1071 xmmi a0,a1,a2,a3,a4,a5,b0,b1,b2,b3,b4,b5;
1072 const uint32_t nb = flag - 1;
1073 xmmi masknb = _mm_shuffle_epi32(_mm_cvtsi32_si128(nb),0);
1074 a0 = _mm_load_si128((xmmi *)in + 0);
1075 a1 = _mm_load_si128((xmmi *)in + 1);
1076 a2 = _mm_load_si128((xmmi *)in + 2);
1077 b0 = _mm_load_si128((xmmi *)out + 0);
1078 b1 = _mm_load_si128((xmmi *)out + 1);
1079 b2 = _mm_load_si128((xmmi *)out + 2);
1080 a0 = _mm_andnot_si128(masknb, a0);
1081 a1 = _mm_andnot_si128(masknb, a1);
1082 a2 = _mm_andnot_si128(masknb, a2);
1083 b0 = _mm_and_si128(masknb, b0);
1084 b1 = _mm_and_si128(masknb, b1);
1085 b2 = _mm_and_si128(masknb, b2);
1086 a0 = _mm_or_si128(a0, b0);
1087 a1 = _mm_or_si128(a1, b1);
1088 a2 = _mm_or_si128(a2, b2);
1089 _mm_store_si128((xmmi*)out + 0, a0);
1090 _mm_store_si128((xmmi*)out + 1, a1);
1091 _mm_store_si128((xmmi*)out + 2, a2);
1092
1093 a3 = _mm_load_si128((xmmi *)in + 3);
1094 a4 = _mm_load_si128((xmmi *)in + 4);
1095 a5 = _mm_load_si128((xmmi *)in + 5);
1096 b3 = _mm_load_si128((xmmi *)out + 3);
1097 b4 = _mm_load_si128((xmmi *)out + 4);
1098 b5 = _mm_load_si128((xmmi *)out + 5);
1099 a3 = _mm_andnot_si128(masknb, a3);
1100 a4 = _mm_andnot_si128(masknb, a4);
1101 a5 = _mm_andnot_si128(masknb, a5);
1102 b3 = _mm_and_si128(masknb, b3);
1103 b4 = _mm_and_si128(masknb, b4);
1104 b5 = _mm_and_si128(masknb, b5);
1105 a3 = _mm_or_si128(a3, b3);
1106 a4 = _mm_or_si128(a4, b4);
1107 a5 = _mm_or_si128(a5, b5);
1108 _mm_store_si128((xmmi*)out + 3, a3);
1109 _mm_store_si128((xmmi*)out + 4, a4);
1110 _mm_store_si128((xmmi*)out + 5, a5);
1111}
1112
bignum25519
uint32_t bignum25519[10]
Definition curve25519-donna-32bit.h:8
carry_pass_full
#define carry_pass_full()
carry_pass_final
#define carry_pass_final()
curve25519_mul_noinline
#define curve25519_mul_noinline
Definition curve25519-donna-32bit.h:152
xmmi
__m128i xmmi
Definition curve25519-donna-sse2.h:9
packed64bignum25519
packedelem64 packed64bignum25519[10]
Definition curve25519-donna-sse2.h:29
packedelem8
union packedelem8_t packedelem8
packedelem64
union packedelem64_t packedelem64
carry_pass_full
#define carry_pass_full()
packed32bignum25519
packedelem32 packed32bignum25519[5]
Definition curve25519-donna-sse2.h:28
F
#define F(i, s)
packedelem32
union packedelem32_t packedelem32
DONNA_NOINLINE
#define DONNA_NOINLINE
Definition ed25519-donna-portable.h:22
DONNA_INLINE
#define DONNA_INLINE
Definition ed25519-donna-portable.h:21
ALIGN
#define ALIGN(x)
Definition ed25519-donna-portable.h:23
rct::zero
key zero()
Definition rctOps.h:70
a
const GenericPointer< typename T::ValueType > T2 T::AllocatorType & a
Definition pointer.h:1124
t2
t2
Definition pow22523.h:103
t1
t1
Definition pow22523.h:58
t0
t0
Definition pow22523.h:53
t3
t3
Definition pow225521.h:103
uint32_t
unsigned int uint32_t
Definition stdint.h:126
int32_t
signed int int32_t
Definition stdint.h:123
uint8_t
unsigned char uint8_t
Definition stdint.h:124
uint64_t
unsigned __int64 uint64_t
Definition stdint.h:136
packedelem32_t
Definition curve25519-donna-sse2.h:16
packedelem32_t::u
uint32_t u[4]
Definition curve25519-donna-sse2.h:17
packedelem32_t::v
xmmi v
Definition curve25519-donna-sse2.h:18
packedelem64_t
Definition curve25519-donna-sse2.h:21
packedelem64_t::v
xmmi v
Definition curve25519-donna-sse2.h:23
packedelem64_t::u
uint64_t u[2]
Definition curve25519-donna-sse2.h:22
packedelem8_t
Definition curve25519-donna-sse2.h:11
packedelem8_t::v
xmmi v
Definition curve25519-donna-sse2.h:13
packedelem8_t::u
unsigned char u[16]
Definition curve25519-donna-sse2.h:12