1 /*****************************************************************************
3 * Copyright (c) 2012, Intel Corporation *
5 * All rights reserved. *
7 * Redistribution and use in source and binary forms, with or without *
8 * modification, are permitted provided that the following conditions are *
11 * * Redistributions of source code must retain the above copyright *
12 * notice, this list of conditions and the following disclaimer. *
14 * * Redistributions in binary form must reproduce the above copyright *
15 * notice, this list of conditions and the following disclaimer in the *
16 * documentation and/or other materials provided with the *
19 * * Neither the name of the Intel Corporation nor the names of its *
20 * contributors may be used to endorse or promote products derived from *
21 * this software without specific prior written permission. *
24 * THIS SOFTWARE IS PROVIDED BY INTEL CORPORATION ""AS IS"" AND ANY *
25 * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE *
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR *
27 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL INTEL CORPORATION OR *
28 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, *
29 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, *
30 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR *
31 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF *
32 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING *
33 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS *
34 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. *
36 ******************************************************************************
37 * Developers and authors: *
38 * Shay Gueron (1, 2), and Vlad Krasnov (1) *
39 * (1) Intel Corporation, Israel Development Center, Haifa, Israel *
40 * (2) University of Haifa, Israel *
41 *****************************************************************************/
48 * See crypto/bn/asm/rsaz-avx2.pl for further details.
50 void rsaz_1024_norm2red_avx2(void *red
, const void *norm
);
51 void rsaz_1024_mul_avx2(void *ret
, const void *a
, const void *b
,
52 const void *n
, BN_ULONG k
);
53 void rsaz_1024_sqr_avx2(void *ret
, const void *a
, const void *n
, BN_ULONG k
,
55 void rsaz_1024_scatter5_avx2(void *tbl
, const void *val
, int i
);
56 void rsaz_1024_gather5_avx2(void *val
, const void *tbl
, int i
);
57 void rsaz_1024_red2norm_avx2(void *norm
, const void *red
);
60 # define ALIGN64 __attribute__((aligned(64)))
61 #elif defined(_MSC_VER)
62 # define ALIGN64 __declspec(align(64))
63 #elif defined(__SUNPRO_C)
65 # pragma align 64(one,two80)
67 /* not fatal, might hurt performance a little */
71 ALIGN64
static const BN_ULONG one
[40] = {
72 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
73 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
76 ALIGN64
static const BN_ULONG two80
[40] = {
77 0, 0, 1 << 22, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
78 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
81 void RSAZ_1024_mod_exp_avx2(BN_ULONG result_norm
[16],
82 const BN_ULONG base_norm
[16],
83 const BN_ULONG exponent
[16],
84 const BN_ULONG m_norm
[16], const BN_ULONG RR
[16],
87 unsigned char storage
[320 * 3 + 32 * 9 * 16 + 64]; /* 5.5KB */
88 unsigned char *p_str
= storage
+ (64 - ((size_t)storage
% 64));
89 unsigned char *a_inv
, *m
, *result
;
90 unsigned char *table_s
= p_str
+ 320 * 3;
91 unsigned char *R2
= table_s
; /* borrow */
95 if ((((size_t)p_str
& 4095) + 320) >> 12) {
98 m
= p_str
+ 320 * 2; /* should not cross page */
100 m
= p_str
; /* should not cross page */
101 result
= p_str
+ 320;
102 a_inv
= p_str
+ 320 * 2;
105 rsaz_1024_norm2red_avx2(m
, m_norm
);
106 rsaz_1024_norm2red_avx2(a_inv
, base_norm
);
107 rsaz_1024_norm2red_avx2(R2
, RR
);
109 rsaz_1024_mul_avx2(R2
, R2
, R2
, m
, k0
);
110 rsaz_1024_mul_avx2(R2
, R2
, two80
, m
, k0
);
113 rsaz_1024_mul_avx2(result
, R2
, one
, m
, k0
);
114 /* table[1] = a_inv^1 */
115 rsaz_1024_mul_avx2(a_inv
, a_inv
, R2
, m
, k0
);
117 rsaz_1024_scatter5_avx2(table_s
, result
, 0);
118 rsaz_1024_scatter5_avx2(table_s
, a_inv
, 1);
120 /* table[2] = a_inv^2 */
121 rsaz_1024_sqr_avx2(result
, a_inv
, m
, k0
, 1);
122 rsaz_1024_scatter5_avx2(table_s
, result
, 2);
124 /* this is almost 2x smaller and less than 1% slower */
125 for (index
= 3; index
< 32; index
++) {
126 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
127 rsaz_1024_scatter5_avx2(table_s
, result
, index
);
130 /* table[4] = a_inv^4 */
131 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
132 rsaz_1024_scatter5_avx2(table_s
, result
, 4);
133 /* table[8] = a_inv^8 */
134 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
135 rsaz_1024_scatter5_avx2(table_s
, result
, 8);
136 /* table[16] = a_inv^16 */
137 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
138 rsaz_1024_scatter5_avx2(table_s
, result
, 16);
139 /* table[17] = a_inv^17 */
140 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
141 rsaz_1024_scatter5_avx2(table_s
, result
, 17);
144 rsaz_1024_gather5_avx2(result
, table_s
, 2);
145 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
146 rsaz_1024_scatter5_avx2(table_s
, result
, 3);
148 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
149 rsaz_1024_scatter5_avx2(table_s
, result
, 6);
151 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
152 rsaz_1024_scatter5_avx2(table_s
, result
, 12);
154 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
155 rsaz_1024_scatter5_avx2(table_s
, result
, 24);
157 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
158 rsaz_1024_scatter5_avx2(table_s
, result
, 25);
161 rsaz_1024_gather5_avx2(result
, table_s
, 4);
162 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
163 rsaz_1024_scatter5_avx2(table_s
, result
, 5);
165 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
166 rsaz_1024_scatter5_avx2(table_s
, result
, 10);
168 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
169 rsaz_1024_scatter5_avx2(table_s
, result
, 20);
171 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
172 rsaz_1024_scatter5_avx2(table_s
, result
, 21);
175 rsaz_1024_gather5_avx2(result
, table_s
, 6);
176 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
177 rsaz_1024_scatter5_avx2(table_s
, result
, 7);
179 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
180 rsaz_1024_scatter5_avx2(table_s
, result
, 14);
182 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
183 rsaz_1024_scatter5_avx2(table_s
, result
, 28);
185 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
186 rsaz_1024_scatter5_avx2(table_s
, result
, 29);
189 rsaz_1024_gather5_avx2(result
, table_s
, 8);
190 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
191 rsaz_1024_scatter5_avx2(table_s
, result
, 9);
193 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
194 rsaz_1024_scatter5_avx2(table_s
, result
, 18);
196 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
197 rsaz_1024_scatter5_avx2(table_s
, result
, 19);
200 rsaz_1024_gather5_avx2(result
, table_s
, 10);
201 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
202 rsaz_1024_scatter5_avx2(table_s
, result
, 11);
204 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
205 rsaz_1024_scatter5_avx2(table_s
, result
, 22);
207 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
208 rsaz_1024_scatter5_avx2(table_s
, result
, 23);
211 rsaz_1024_gather5_avx2(result
, table_s
, 12);
212 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
213 rsaz_1024_scatter5_avx2(table_s
, result
, 13);
215 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
216 rsaz_1024_scatter5_avx2(table_s
, result
, 26);
218 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
219 rsaz_1024_scatter5_avx2(table_s
, result
, 27);
222 rsaz_1024_gather5_avx2(result
, table_s
, 14);
223 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
224 rsaz_1024_scatter5_avx2(table_s
, result
, 15);
226 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 1);
227 rsaz_1024_scatter5_avx2(table_s
, result
, 30);
229 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
230 rsaz_1024_scatter5_avx2(table_s
, result
, 31);
233 /* load first window */
234 p_str
= (unsigned char *)exponent
;
235 wvalue
= p_str
[127] >> 3;
236 rsaz_1024_gather5_avx2(result
, table_s
, wvalue
);
240 while (index
> -1) { /* loop for the remaining 127 windows */
242 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 5);
244 wvalue
= *((unsigned short *)&p_str
[index
/ 8]);
245 wvalue
= (wvalue
>> (index
% 8)) & 31;
248 rsaz_1024_gather5_avx2(a_inv
, table_s
, wvalue
); /* borrow a_inv */
249 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
252 /* square four times */
253 rsaz_1024_sqr_avx2(result
, result
, m
, k0
, 4);
255 wvalue
= p_str
[0] & 15;
257 rsaz_1024_gather5_avx2(a_inv
, table_s
, wvalue
); /* borrow a_inv */
258 rsaz_1024_mul_avx2(result
, result
, a_inv
, m
, k0
);
260 /* from Montgomery */
261 rsaz_1024_mul_avx2(result
, result
, one
, m
, k0
);
263 rsaz_1024_red2norm_avx2(result_norm
, result
);
265 OPENSSL_cleanse(storage
, sizeof(storage
));
269 * See crypto/bn/rsaz-x86_64.pl for further details.
271 void rsaz_512_mul(void *ret
, const void *a
, const void *b
, const void *n
,
273 void rsaz_512_mul_scatter4(void *ret
, const void *a
, const void *n
,
274 BN_ULONG k
, const void *tbl
, unsigned int power
);
275 void rsaz_512_mul_gather4(void *ret
, const void *a
, const void *tbl
,
276 const void *n
, BN_ULONG k
, unsigned int power
);
277 void rsaz_512_mul_by_one(void *ret
, const void *a
, const void *n
, BN_ULONG k
);
278 void rsaz_512_sqr(void *ret
, const void *a
, const void *n
, BN_ULONG k
,
280 void rsaz_512_scatter4(void *tbl
, const BN_ULONG
*val
, int power
);
281 void rsaz_512_gather4(BN_ULONG
*val
, const void *tbl
, int power
);
283 void RSAZ_512_mod_exp(BN_ULONG result
[8],
284 const BN_ULONG base
[8], const BN_ULONG exponent
[8],
285 const BN_ULONG m
[8], BN_ULONG k0
, const BN_ULONG RR
[8])
287 unsigned char storage
[16 * 8 * 8 + 64 * 2 + 64]; /* 1.2KB */
288 unsigned char *table
= storage
+ (64 - ((size_t)storage
% 64));
289 BN_ULONG
*a_inv
= (BN_ULONG
*)(table
+ 16 * 8 * 8);
290 BN_ULONG
*temp
= (BN_ULONG
*)(table
+ 16 * 8 * 8 + 8 * 8);
291 unsigned char *p_str
= (unsigned char *)exponent
;
295 /* table[0] = 1_inv */
304 rsaz_512_scatter4(table
, temp
, 0);
306 /* table [1] = a_inv^1 */
307 rsaz_512_mul(a_inv
, base
, RR
, m
, k0
);
308 rsaz_512_scatter4(table
, a_inv
, 1);
310 /* table [2] = a_inv^2 */
311 rsaz_512_sqr(temp
, a_inv
, m
, k0
, 1);
312 rsaz_512_scatter4(table
, temp
, 2);
314 for (index
= 3; index
< 16; index
++)
315 rsaz_512_mul_scatter4(temp
, a_inv
, m
, k0
, table
, index
);
317 /* load first window */
320 rsaz_512_gather4(temp
, table
, wvalue
>> 4);
321 rsaz_512_sqr(temp
, temp
, m
, k0
, 4);
322 rsaz_512_mul_gather4(temp
, temp
, table
, m
, k0
, wvalue
& 0xf);
324 for (index
= 62; index
>= 0; index
--) {
325 wvalue
= p_str
[index
];
327 rsaz_512_sqr(temp
, temp
, m
, k0
, 4);
328 rsaz_512_mul_gather4(temp
, temp
, table
, m
, k0
, wvalue
>> 4);
330 rsaz_512_sqr(temp
, temp
, m
, k0
, 4);
331 rsaz_512_mul_gather4(temp
, temp
, table
, m
, k0
, wvalue
& 0x0f);
334 /* from Montgomery */
335 rsaz_512_mul_by_one(result
, temp
, m
, k0
);
337 OPENSSL_cleanse(storage
, sizeof(storage
));
342 # if defined(PEDANTIC) || defined(__DECC) || defined(__clang__)
343 static void *dummy
= &dummy
;