1/* crypto/bn/bn_nist.c */
2/*
3 * Written by Nils Larsch for the OpenSSL project
4 */
5/* ====================================================================
6 * Copyright (c) 1998-2005 The OpenSSL Project.  All rights reserved.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 *
12 * 1. Redistributions of source code must retain the above copyright
13 *    notice, this list of conditions and the following disclaimer.
14 *
15 * 2. Redistributions in binary form must reproduce the above copyright
16 *    notice, this list of conditions and the following disclaimer in
17 *    the documentation and/or other materials provided with the
18 *    distribution.
19 *
20 * 3. All advertising materials mentioning features or use of this
21 *    software must display the following acknowledgment:
22 *    "This product includes software developed by the OpenSSL Project
23 *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
24 *
25 * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
26 *    endorse or promote products derived from this software without
27 *    prior written permission. For written permission, please contact
28 *    openssl-core@openssl.org.
29 *
30 * 5. Products derived from this software may not be called "OpenSSL"
31 *    nor may "OpenSSL" appear in their names without prior written
32 *    permission of the OpenSSL Project.
33 *
34 * 6. Redistributions of any form whatsoever must retain the following
35 *    acknowledgment:
36 *    "This product includes software developed by the OpenSSL Project
37 *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
38 *
39 * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
40 * EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
41 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
42 * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE OpenSSL PROJECT OR
43 * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
44 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
45 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
46 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
47 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
48 * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
49 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
50 * OF THE POSSIBILITY OF SUCH DAMAGE.
51 * ====================================================================
52 *
53 * This product includes cryptographic software written by Eric Young
54 * (eay@cryptsoft.com).  This product includes software written by Tim
55 * Hudson (tjh@cryptsoft.com).
56 *
57 */
58
59#include "bn_lcl.h"
60#include <openssl/local/cryptlib.h>
61
62
63#define BN_NIST_192_TOP	(192+BN_BITS2-1)/BN_BITS2
64#define BN_NIST_224_TOP	(224+BN_BITS2-1)/BN_BITS2
65#define BN_NIST_256_TOP	(256+BN_BITS2-1)/BN_BITS2
66#define BN_NIST_384_TOP	(384+BN_BITS2-1)/BN_BITS2
67#define BN_NIST_521_TOP	(521+BN_BITS2-1)/BN_BITS2
68
69/* pre-computed tables are "carry-less" values of modulus*(i+1) */
70#if BN_BITS2 == 64
71static const BN_ULONG _nist_p_192[][BN_NIST_192_TOP] = {
72	{0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFEULL,0xFFFFFFFFFFFFFFFFULL},
73	{0xFFFFFFFFFFFFFFFEULL,0xFFFFFFFFFFFFFFFDULL,0xFFFFFFFFFFFFFFFFULL},
74	{0xFFFFFFFFFFFFFFFDULL,0xFFFFFFFFFFFFFFFCULL,0xFFFFFFFFFFFFFFFFULL}
75	};
76static const BN_ULONG _nist_p_192_sqr[] = {
77	0x0000000000000001ULL,0x0000000000000002ULL,0x0000000000000001ULL,
78	0xFFFFFFFFFFFFFFFEULL,0xFFFFFFFFFFFFFFFDULL,0xFFFFFFFFFFFFFFFFULL
79	};
80static const BN_ULONG _nist_p_224[][BN_NIST_224_TOP] = {
81	{0x0000000000000001ULL,0xFFFFFFFF00000000ULL,
82	 0xFFFFFFFFFFFFFFFFULL,0x00000000FFFFFFFFULL},
83	{0x0000000000000002ULL,0xFFFFFFFE00000000ULL,
84	 0xFFFFFFFFFFFFFFFFULL,0x00000001FFFFFFFFULL} /* this one is "carry-full" */
85	};
86static const BN_ULONG _nist_p_224_sqr[] = {
87	0x0000000000000001ULL,0xFFFFFFFE00000000ULL,
88	0xFFFFFFFFFFFFFFFFULL,0x0000000200000000ULL,
89	0x0000000000000000ULL,0xFFFFFFFFFFFFFFFEULL,
90	0xFFFFFFFFFFFFFFFFULL
91	};
92static const BN_ULONG _nist_p_256[][BN_NIST_256_TOP] = {
93	{0xFFFFFFFFFFFFFFFFULL,0x00000000FFFFFFFFULL,
94	 0x0000000000000000ULL,0xFFFFFFFF00000001ULL},
95	{0xFFFFFFFFFFFFFFFEULL,0x00000001FFFFFFFFULL,
96	 0x0000000000000000ULL,0xFFFFFFFE00000002ULL},
97	{0xFFFFFFFFFFFFFFFDULL,0x00000002FFFFFFFFULL,
98	 0x0000000000000000ULL,0xFFFFFFFD00000003ULL},
99	{0xFFFFFFFFFFFFFFFCULL,0x00000003FFFFFFFFULL,
100	 0x0000000000000000ULL,0xFFFFFFFC00000004ULL},
101	{0xFFFFFFFFFFFFFFFBULL,0x00000004FFFFFFFFULL,
102	 0x0000000000000000ULL,0xFFFFFFFB00000005ULL},
103	};
104static const BN_ULONG _nist_p_256_sqr[] = {
105	0x0000000000000001ULL,0xFFFFFFFE00000000ULL,
106	0xFFFFFFFFFFFFFFFFULL,0x00000001FFFFFFFEULL,
107	0x00000001FFFFFFFEULL,0x00000001FFFFFFFEULL,
108	0xFFFFFFFE00000001ULL,0xFFFFFFFE00000002ULL
109	};
110static const BN_ULONG _nist_p_384[][BN_NIST_384_TOP] = {
111	{0x00000000FFFFFFFFULL,0xFFFFFFFF00000000ULL,0xFFFFFFFFFFFFFFFEULL,
112	 0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL},
113	{0x00000001FFFFFFFEULL,0xFFFFFFFE00000000ULL,0xFFFFFFFFFFFFFFFDULL,
114	 0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL},
115	{0x00000002FFFFFFFDULL,0xFFFFFFFD00000000ULL,0xFFFFFFFFFFFFFFFCULL,
116	 0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL},
117	{0x00000003FFFFFFFCULL,0xFFFFFFFC00000000ULL,0xFFFFFFFFFFFFFFFBULL,
118	 0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL},
119	{0x00000004FFFFFFFBULL,0xFFFFFFFB00000000ULL,0xFFFFFFFFFFFFFFFAULL,
120	 0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL},
121	};
122static const BN_ULONG _nist_p_384_sqr[] = {
123	0xFFFFFFFE00000001ULL,0x0000000200000000ULL,0xFFFFFFFE00000000ULL,
124	0x0000000200000000ULL,0x0000000000000001ULL,0x0000000000000000ULL,
125	0x00000001FFFFFFFEULL,0xFFFFFFFE00000000ULL,0xFFFFFFFFFFFFFFFDULL,
126	0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL
127	};
128static const BN_ULONG _nist_p_521[] =
129	{0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,
130	0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,
131	0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,
132	0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,
133	0x00000000000001FFULL};
134static const BN_ULONG _nist_p_521_sqr[] = {
135	0x0000000000000001ULL,0x0000000000000000ULL,0x0000000000000000ULL,
136	0x0000000000000000ULL,0x0000000000000000ULL,0x0000000000000000ULL,
137	0x0000000000000000ULL,0x0000000000000000ULL,0xFFFFFFFFFFFFFC00ULL,
138	0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,
139	0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,0xFFFFFFFFFFFFFFFFULL,
140	0xFFFFFFFFFFFFFFFFULL,0x000000000003FFFFULL
141	};
142#elif BN_BITS2 == 32
143static const BN_ULONG _nist_p_192[][BN_NIST_192_TOP] = {
144	{0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFE,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
145	{0xFFFFFFFE,0xFFFFFFFF,0xFFFFFFFD,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
146	{0xFFFFFFFD,0xFFFFFFFF,0xFFFFFFFC,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF}
147	};
148static const BN_ULONG _nist_p_192_sqr[] = {
149	0x00000001,0x00000000,0x00000002,0x00000000,0x00000001,0x00000000,
150	0xFFFFFFFE,0xFFFFFFFF,0xFFFFFFFD,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF
151	};
152static const BN_ULONG _nist_p_224[][BN_NIST_224_TOP] = {
153	{0x00000001,0x00000000,0x00000000,0xFFFFFFFF,
154	 0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
155	{0x00000002,0x00000000,0x00000000,0xFFFFFFFE,
156	 0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF}
157	};
158static const BN_ULONG _nist_p_224_sqr[] = {
159	0x00000001,0x00000000,0x00000000,0xFFFFFFFE,
160	0xFFFFFFFF,0xFFFFFFFF,0x00000000,0x00000002,
161	0x00000000,0x00000000,0xFFFFFFFE,0xFFFFFFFF,
162	0xFFFFFFFF,0xFFFFFFFF
163	};
164static const BN_ULONG _nist_p_256[][BN_NIST_256_TOP] = {
165	{0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0x00000000,
166	 0x00000000,0x00000000,0x00000001,0xFFFFFFFF},
167	{0xFFFFFFFE,0xFFFFFFFF,0xFFFFFFFF,0x00000001,
168	 0x00000000,0x00000000,0x00000002,0xFFFFFFFE},
169	{0xFFFFFFFD,0xFFFFFFFF,0xFFFFFFFF,0x00000002,
170	 0x00000000,0x00000000,0x00000003,0xFFFFFFFD},
171	{0xFFFFFFFC,0xFFFFFFFF,0xFFFFFFFF,0x00000003,
172	 0x00000000,0x00000000,0x00000004,0xFFFFFFFC},
173	{0xFFFFFFFB,0xFFFFFFFF,0xFFFFFFFF,0x00000004,
174	 0x00000000,0x00000000,0x00000005,0xFFFFFFFB},
175	};
176static const BN_ULONG _nist_p_256_sqr[] = {
177	0x00000001,0x00000000,0x00000000,0xFFFFFFFE,
178	0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFE,0x00000001,
179	0xFFFFFFFE,0x00000001,0xFFFFFFFE,0x00000001,
180	0x00000001,0xFFFFFFFE,0x00000002,0xFFFFFFFE
181	};
182static const BN_ULONG _nist_p_384[][BN_NIST_384_TOP] = {
183	{0xFFFFFFFF,0x00000000,0x00000000,0xFFFFFFFF,0xFFFFFFFE,0xFFFFFFFF,
184	 0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
185	{0xFFFFFFFE,0x00000001,0x00000000,0xFFFFFFFE,0xFFFFFFFD,0xFFFFFFFF,
186	 0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
187	{0xFFFFFFFD,0x00000002,0x00000000,0xFFFFFFFD,0xFFFFFFFC,0xFFFFFFFF,
188	 0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
189	{0xFFFFFFFC,0x00000003,0x00000000,0xFFFFFFFC,0xFFFFFFFB,0xFFFFFFFF,
190	 0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
191	{0xFFFFFFFB,0x00000004,0x00000000,0xFFFFFFFB,0xFFFFFFFA,0xFFFFFFFF,
192	 0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF},
193	};
194static const BN_ULONG _nist_p_384_sqr[] = {
195	0x00000001,0xFFFFFFFE,0x00000000,0x00000002,0x00000000,0xFFFFFFFE,
196	0x00000000,0x00000002,0x00000001,0x00000000,0x00000000,0x00000000,
197	0xFFFFFFFE,0x00000001,0x00000000,0xFFFFFFFE,0xFFFFFFFD,0xFFFFFFFF,
198	0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF
199	};
200static const BN_ULONG _nist_p_521[] = {0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,
201	0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,
202	0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,
203	0xFFFFFFFF,0x000001FF};
204static const BN_ULONG _nist_p_521_sqr[] = {
205	0x00000001,0x00000000,0x00000000,0x00000000,0x00000000,0x00000000,
206	0x00000000,0x00000000,0x00000000,0x00000000,0x00000000,0x00000000,
207	0x00000000,0x00000000,0x00000000,0x00000000,0xFFFFFC00,0xFFFFFFFF,
208	0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,
209	0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,0xFFFFFFFF,
210	0xFFFFFFFF,0xFFFFFFFF,0x0003FFFF
211	};
212#else
213#error "unsupported BN_BITS2"
214#endif
215
216
217static const BIGNUM _bignum_nist_p_192 =
218	{
219	(BN_ULONG *)_nist_p_192[0],
220	BN_NIST_192_TOP,
221	BN_NIST_192_TOP,
222	0,
223	BN_FLG_STATIC_DATA
224	};
225
226static const BIGNUM _bignum_nist_p_224 =
227	{
228	(BN_ULONG *)_nist_p_224[0],
229	BN_NIST_224_TOP,
230	BN_NIST_224_TOP,
231	0,
232	BN_FLG_STATIC_DATA
233	};
234
235static const BIGNUM _bignum_nist_p_256 =
236	{
237	(BN_ULONG *)_nist_p_256[0],
238	BN_NIST_256_TOP,
239	BN_NIST_256_TOP,
240	0,
241	BN_FLG_STATIC_DATA
242	};
243
244static const BIGNUM _bignum_nist_p_384 =
245	{
246	(BN_ULONG *)_nist_p_384[0],
247	BN_NIST_384_TOP,
248	BN_NIST_384_TOP,
249	0,
250	BN_FLG_STATIC_DATA
251	};
252
253static const BIGNUM _bignum_nist_p_521 =
254	{
255	(BN_ULONG *)_nist_p_521,
256	BN_NIST_521_TOP,
257	BN_NIST_521_TOP,
258	0,
259	BN_FLG_STATIC_DATA
260	};
261
262
263const BIGNUM *BN_get0_nist_prime_192(void)
264	{
265	return &_bignum_nist_p_192;
266	}
267
268const BIGNUM *BN_get0_nist_prime_224(void)
269	{
270	return &_bignum_nist_p_224;
271	}
272
273const BIGNUM *BN_get0_nist_prime_256(void)
274	{
275	return &_bignum_nist_p_256;
276	}
277
278const BIGNUM *BN_get0_nist_prime_384(void)
279	{
280	return &_bignum_nist_p_384;
281	}
282
283const BIGNUM *BN_get0_nist_prime_521(void)
284	{
285	return &_bignum_nist_p_521;
286	}
287
288
289static void nist_cp_bn_0(BN_ULONG *buf, BN_ULONG *a, int top, int max)
290	{
291	int i;
292	BN_ULONG *_tmp1 = (buf), *_tmp2 = (a);
293
294#ifdef BN_DEBUG
295	OPENSSL_assert(top <= max);
296#endif
297	for (i = (top); i != 0; i--)
298		*_tmp1++ = *_tmp2++;
299	for (i = (max) - (top); i != 0; i--)
300		*_tmp1++ = (BN_ULONG) 0;
301	}
302
303static void nist_cp_bn(BN_ULONG *buf, BN_ULONG *a, int top)
304	{
305	int i;
306	BN_ULONG *_tmp1 = (buf), *_tmp2 = (a);
307	for (i = (top); i != 0; i--)
308		*_tmp1++ = *_tmp2++;
309	}
310
311#if BN_BITS2 == 64
312#define bn_cp_64(to, n, from, m)	(to)[n] = (m>=0)?((from)[m]):0;
313#define bn_64_set_0(to, n)		(to)[n] = (BN_ULONG)0;
314/*
315 * two following macros are implemented under assumption that they
316 * are called in a sequence with *ascending* n, i.e. as they are...
317 */
318#define bn_cp_32_naked(to, n, from, m)	(((n)&1)?(to[(n)/2]|=((m)&1)?(from[(m)/2]&BN_MASK2h):(from[(m)/2]<<32))\
319						:(to[(n)/2] =((m)&1)?(from[(m)/2]>>32):(from[(m)/2]&BN_MASK2l)))
320#define bn_32_set_0(to, n)		(((n)&1)?(to[(n)/2]&=BN_MASK2l):(to[(n)/2]=0));
321#define bn_cp_32(to,n,from,m)		((m)>=0)?bn_cp_32_naked(to,n,from,m):bn_32_set_0(to,n)
322#else
323#define bn_cp_64(to, n, from, m) \
324	{ \
325	bn_cp_32(to, (n)*2, from, (m)*2); \
326	bn_cp_32(to, (n)*2+1, from, (m)*2+1); \
327	}
328#define bn_64_set_0(to, n) \
329	{ \
330	bn_32_set_0(to, (n)*2); \
331	bn_32_set_0(to, (n)*2+1); \
332	}
333#if BN_BITS2 == 32
334#define bn_cp_32(to, n, from, m)	(to)[n] = (m>=0)?((from)[m]):0;
335#define bn_32_set_0(to, n)		(to)[n] = (BN_ULONG)0;
336#endif
337#endif /* BN_BITS2 != 64 */
338
339
340#define nist_set_192(to, from, a1, a2, a3) \
341	{ \
342	bn_cp_64(to, 0, from, (a3) - 3) \
343	bn_cp_64(to, 1, from, (a2) - 3) \
344	bn_cp_64(to, 2, from, (a1) - 3) \
345	}
346
347int BN_nist_mod_192(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
348	BN_CTX *ctx)
349	{
350	int      top = a->top, i;
351	int      carry;
352	register BN_ULONG *r_d, *a_d = a->d;
353	BN_ULONG t_d[BN_NIST_192_TOP],
354	         buf[BN_NIST_192_TOP],
355		 c_d[BN_NIST_192_TOP],
356		*res;
357	size_t   mask;
358	static const BIGNUM _bignum_nist_p_192_sqr = {
359		(BN_ULONG *)_nist_p_192_sqr,
360		sizeof(_nist_p_192_sqr)/sizeof(_nist_p_192_sqr[0]),
361		sizeof(_nist_p_192_sqr)/sizeof(_nist_p_192_sqr[0]),
362		0,BN_FLG_STATIC_DATA };
363
364	field = &_bignum_nist_p_192; /* just to make sure */
365
366 	if (BN_is_negative(a) || BN_ucmp(a,&_bignum_nist_p_192_sqr)>=0)
367		return BN_nnmod(r, a, field, ctx);
368
369	i = BN_ucmp(field, a);
370	if (i == 0)
371		{
372		BN_zero(r);
373		return 1;
374		}
375	else if (i > 0)
376		return (r == a) ? 1 : (BN_copy(r ,a) != NULL);
377
378	if (r != a)
379		{
380		if (!bn_wexpand(r, BN_NIST_192_TOP))
381			return 0;
382		r_d = r->d;
383		nist_cp_bn(r_d, a_d, BN_NIST_192_TOP);
384		}
385	else
386		r_d = a_d;
387
388	nist_cp_bn_0(buf, a_d + BN_NIST_192_TOP, top - BN_NIST_192_TOP, BN_NIST_192_TOP);
389
390	nist_set_192(t_d, buf, 0, 3, 3);
391	carry = (int)bn_add_words(r_d, r_d, t_d, BN_NIST_192_TOP);
392	nist_set_192(t_d, buf, 4, 4, 0);
393	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_192_TOP);
394	nist_set_192(t_d, buf, 5, 5, 5)
395	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_192_TOP);
396
397	if (carry > 0)
398		carry = (int)bn_sub_words(r_d,r_d,_nist_p_192[carry-1],BN_NIST_192_TOP);
399	else
400		carry = 1;
401
402	/*
403	 * we need 'if (carry==0 || result>=modulus) result-=modulus;'
404	 * as comparison implies subtraction, we can write
405	 * 'tmp=result-modulus; if (!carry || !borrow) result=tmp;'
406	 * this is what happens below, but without explicit if:-) a.
407	 */
408	mask  = 0-(size_t)bn_sub_words(c_d,r_d,_nist_p_192[0],BN_NIST_192_TOP);
409	mask &= 0-(size_t)carry;
410	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
411	nist_cp_bn(r_d, res, BN_NIST_192_TOP);
412	r->top = BN_NIST_192_TOP;
413	bn_correct_top(r);
414
415	return 1;
416	}
417
418typedef BN_ULONG (*bn_addsub_f)(BN_ULONG *,const BN_ULONG *,const BN_ULONG *,int);
419
420#define nist_set_224(to, from, a1, a2, a3, a4, a5, a6, a7) \
421	{ \
422	bn_cp_32(to, 0, from, (a7) - 7) \
423	bn_cp_32(to, 1, from, (a6) - 7) \
424	bn_cp_32(to, 2, from, (a5) - 7) \
425	bn_cp_32(to, 3, from, (a4) - 7) \
426	bn_cp_32(to, 4, from, (a3) - 7) \
427	bn_cp_32(to, 5, from, (a2) - 7) \
428	bn_cp_32(to, 6, from, (a1) - 7) \
429	}
430
431int BN_nist_mod_224(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
432	BN_CTX *ctx)
433	{
434	int	top = a->top, i;
435	int	carry;
436	BN_ULONG *r_d, *a_d = a->d;
437	BN_ULONG t_d[BN_NIST_224_TOP],
438	         buf[BN_NIST_224_TOP],
439		 c_d[BN_NIST_224_TOP],
440		*res;
441	size_t   mask;
442	union { bn_addsub_f f; size_t p; } u;
443	static const BIGNUM _bignum_nist_p_224_sqr = {
444		(BN_ULONG *)_nist_p_224_sqr,
445		sizeof(_nist_p_224_sqr)/sizeof(_nist_p_224_sqr[0]),
446		sizeof(_nist_p_224_sqr)/sizeof(_nist_p_224_sqr[0]),
447		0,BN_FLG_STATIC_DATA };
448
449
450	field = &_bignum_nist_p_224; /* just to make sure */
451
452 	if (BN_is_negative(a) || BN_ucmp(a,&_bignum_nist_p_224_sqr)>=0)
453		return BN_nnmod(r, a, field, ctx);
454
455	i = BN_ucmp(field, a);
456	if (i == 0)
457		{
458		BN_zero(r);
459		return 1;
460		}
461	else if (i > 0)
462		return (r == a)? 1 : (BN_copy(r ,a) != NULL);
463
464	if (r != a)
465		{
466		if (!bn_wexpand(r, BN_NIST_224_TOP))
467			return 0;
468		r_d = r->d;
469		nist_cp_bn(r_d, a_d, BN_NIST_224_TOP);
470		}
471	else
472		r_d = a_d;
473
474#if BN_BITS2==64
475	/* copy upper 256 bits of 448 bit number ... */
476	nist_cp_bn_0(t_d, a_d + (BN_NIST_224_TOP-1), top - (BN_NIST_224_TOP-1), BN_NIST_224_TOP);
477	/* ... and right shift by 32 to obtain upper 224 bits */
478	nist_set_224(buf, t_d, 14, 13, 12, 11, 10, 9, 8);
479	/* truncate lower part to 224 bits too */
480	r_d[BN_NIST_224_TOP-1] &= BN_MASK2l;
481#else
482	nist_cp_bn_0(buf, a_d + BN_NIST_224_TOP, top - BN_NIST_224_TOP, BN_NIST_224_TOP);
483#endif
484	nist_set_224(t_d, buf, 10, 9, 8, 7, 0, 0, 0);
485	carry = (int)bn_add_words(r_d, r_d, t_d, BN_NIST_224_TOP);
486	nist_set_224(t_d, buf, 0, 13, 12, 11, 0, 0, 0);
487	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_224_TOP);
488	nist_set_224(t_d, buf, 13, 12, 11, 10, 9, 8, 7);
489	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_224_TOP);
490	nist_set_224(t_d, buf, 0, 0, 0, 0, 13, 12, 11);
491	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_224_TOP);
492
493#if BN_BITS2==64
494	carry = (int)(r_d[BN_NIST_224_TOP-1]>>32);
495#endif
496	u.f = bn_sub_words;
497	if (carry > 0)
498		{
499		carry = (int)bn_sub_words(r_d,r_d,_nist_p_224[carry-1],BN_NIST_224_TOP);
500#if BN_BITS2==64
501		carry=(int)(~(r_d[BN_NIST_224_TOP-1]>>32))&1;
502#endif
503		}
504	else if (carry < 0)
505		{
506		/* it's a bit more comlicated logic in this case.
507		 * if bn_add_words yields no carry, then result
508		 * has to be adjusted by unconditionally *adding*
509		 * the modulus. but if it does, then result has
510		 * to be compared to the modulus and conditionally
511		 * adjusted by *subtracting* the latter. */
512		carry = (int)bn_add_words(r_d,r_d,_nist_p_224[-carry-1],BN_NIST_224_TOP);
513		mask = 0-(size_t)carry;
514		u.p = ((size_t)bn_sub_words&mask) | ((size_t)bn_add_words&~mask);
515		}
516	else
517		carry = 1;
518
519	/* otherwise it's effectively same as in BN_nist_mod_192... */
520	mask  = 0-(size_t)(*u.f)(c_d,r_d,_nist_p_224[0],BN_NIST_224_TOP);
521	mask &= 0-(size_t)carry;
522	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
523	nist_cp_bn(r_d, res, BN_NIST_224_TOP);
524	r->top = BN_NIST_224_TOP;
525	bn_correct_top(r);
526
527	return 1;
528	}
529
530#define nist_set_256(to, from, a1, a2, a3, a4, a5, a6, a7, a8) \
531	{ \
532	bn_cp_32(to, 0, from, (a8) - 8) \
533	bn_cp_32(to, 1, from, (a7) - 8) \
534	bn_cp_32(to, 2, from, (a6) - 8) \
535	bn_cp_32(to, 3, from, (a5) - 8) \
536	bn_cp_32(to, 4, from, (a4) - 8) \
537	bn_cp_32(to, 5, from, (a3) - 8) \
538	bn_cp_32(to, 6, from, (a2) - 8) \
539	bn_cp_32(to, 7, from, (a1) - 8) \
540	}
541
542int BN_nist_mod_256(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
543	BN_CTX *ctx)
544	{
545	int	i, top = a->top;
546	int	carry = 0;
547	register BN_ULONG *a_d = a->d, *r_d;
548	BN_ULONG t_d[BN_NIST_256_TOP],
549	         buf[BN_NIST_256_TOP],
550		 c_d[BN_NIST_256_TOP],
551		*res;
552	size_t   mask;
553	union { bn_addsub_f f; size_t p; } u;
554	static const BIGNUM _bignum_nist_p_256_sqr = {
555		(BN_ULONG *)_nist_p_256_sqr,
556		sizeof(_nist_p_256_sqr)/sizeof(_nist_p_256_sqr[0]),
557		sizeof(_nist_p_256_sqr)/sizeof(_nist_p_256_sqr[0]),
558		0,BN_FLG_STATIC_DATA };
559
560	field = &_bignum_nist_p_256; /* just to make sure */
561
562 	if (BN_is_negative(a) || BN_ucmp(a,&_bignum_nist_p_256_sqr)>=0)
563		return BN_nnmod(r, a, field, ctx);
564
565	i = BN_ucmp(field, a);
566	if (i == 0)
567		{
568		BN_zero(r);
569		return 1;
570		}
571	else if (i > 0)
572		return (r == a)? 1 : (BN_copy(r ,a) != NULL);
573
574	if (r != a)
575		{
576		if (!bn_wexpand(r, BN_NIST_256_TOP))
577			return 0;
578		r_d = r->d;
579		nist_cp_bn(r_d, a_d, BN_NIST_256_TOP);
580		}
581	else
582		r_d = a_d;
583
584	nist_cp_bn_0(buf, a_d + BN_NIST_256_TOP, top - BN_NIST_256_TOP, BN_NIST_256_TOP);
585
586	/*S1*/
587	nist_set_256(t_d, buf, 15, 14, 13, 12, 11, 0, 0, 0);
588	/*S2*/
589	nist_set_256(c_d, buf, 0, 15, 14, 13, 12, 0, 0, 0);
590	carry = (int)bn_add_words(t_d, t_d, c_d, BN_NIST_256_TOP);
591	/* left shift */
592		{
593		register BN_ULONG *ap,t,c;
594		ap = t_d;
595		c=0;
596		for (i = BN_NIST_256_TOP; i != 0; --i)
597			{
598			t= *ap;
599			*(ap++)=((t<<1)|c)&BN_MASK2;
600			c=(t & BN_TBIT)?1:0;
601			}
602		carry <<= 1;
603		carry  |= c;
604		}
605	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_256_TOP);
606	/*S3*/
607	nist_set_256(t_d, buf, 15, 14, 0, 0, 0, 10, 9, 8);
608	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_256_TOP);
609	/*S4*/
610	nist_set_256(t_d, buf, 8, 13, 15, 14, 13, 11, 10, 9);
611	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_256_TOP);
612	/*D1*/
613	nist_set_256(t_d, buf, 10, 8, 0, 0, 0, 13, 12, 11);
614	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
615	/*D2*/
616	nist_set_256(t_d, buf, 11, 9, 0, 0, 15, 14, 13, 12);
617	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
618	/*D3*/
619	nist_set_256(t_d, buf, 12, 0, 10, 9, 8, 15, 14, 13);
620	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
621	/*D4*/
622	nist_set_256(t_d, buf, 13, 0, 11, 10, 9, 0, 15, 14);
623	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_256_TOP);
624
625	/* see BN_nist_mod_224 for explanation */
626	u.f = bn_sub_words;
627	if (carry > 0)
628		carry = (int)bn_sub_words(r_d,r_d,_nist_p_256[carry-1],BN_NIST_256_TOP);
629	else if (carry < 0)
630		{
631		carry = (int)bn_add_words(r_d,r_d,_nist_p_256[-carry-1],BN_NIST_256_TOP);
632		mask = 0-(size_t)carry;
633		u.p = ((size_t)bn_sub_words&mask) | ((size_t)bn_add_words&~mask);
634		}
635	else
636		carry = 1;
637
638	mask  = 0-(size_t)(*u.f)(c_d,r_d,_nist_p_256[0],BN_NIST_256_TOP);
639	mask &= 0-(size_t)carry;
640	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
641	nist_cp_bn(r_d, res, BN_NIST_256_TOP);
642	r->top = BN_NIST_256_TOP;
643	bn_correct_top(r);
644
645	return 1;
646	}
647
648#define nist_set_384(to,from,a1,a2,a3,a4,a5,a6,a7,a8,a9,a10,a11,a12) \
649	{ \
650	bn_cp_32(to, 0, from,  (a12) - 12) \
651	bn_cp_32(to, 1, from,  (a11) - 12) \
652	bn_cp_32(to, 2, from,  (a10) - 12) \
653	bn_cp_32(to, 3, from,  (a9) - 12)  \
654	bn_cp_32(to, 4, from,  (a8) - 12)  \
655	bn_cp_32(to, 5, from,  (a7) - 12)  \
656	bn_cp_32(to, 6, from,  (a6) - 12)  \
657	bn_cp_32(to, 7, from,  (a5) - 12)  \
658	bn_cp_32(to, 8, from,  (a4) - 12)  \
659	bn_cp_32(to, 9, from,  (a3) - 12)  \
660	bn_cp_32(to, 10, from, (a2) - 12)  \
661	bn_cp_32(to, 11, from, (a1) - 12)  \
662	}
663
664int BN_nist_mod_384(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
665	BN_CTX *ctx)
666	{
667	int	i, top = a->top;
668	int	carry = 0;
669	register BN_ULONG *r_d, *a_d = a->d;
670	BN_ULONG t_d[BN_NIST_384_TOP],
671	         buf[BN_NIST_384_TOP],
672		 c_d[BN_NIST_384_TOP],
673		*res;
674	size_t	 mask;
675	union { bn_addsub_f f; size_t p; } u;
676	static const BIGNUM _bignum_nist_p_384_sqr = {
677		(BN_ULONG *)_nist_p_384_sqr,
678		sizeof(_nist_p_384_sqr)/sizeof(_nist_p_384_sqr[0]),
679		sizeof(_nist_p_384_sqr)/sizeof(_nist_p_384_sqr[0]),
680		0,BN_FLG_STATIC_DATA };
681
682
683	field = &_bignum_nist_p_384; /* just to make sure */
684
685 	if (BN_is_negative(a) || BN_ucmp(a,&_bignum_nist_p_384_sqr)>=0)
686		return BN_nnmod(r, a, field, ctx);
687
688	i = BN_ucmp(field, a);
689	if (i == 0)
690		{
691		BN_zero(r);
692		return 1;
693		}
694	else if (i > 0)
695		return (r == a)? 1 : (BN_copy(r ,a) != NULL);
696
697	if (r != a)
698		{
699		if (!bn_wexpand(r, BN_NIST_384_TOP))
700			return 0;
701		r_d = r->d;
702		nist_cp_bn(r_d, a_d, BN_NIST_384_TOP);
703		}
704	else
705		r_d = a_d;
706
707	nist_cp_bn_0(buf, a_d + BN_NIST_384_TOP, top - BN_NIST_384_TOP, BN_NIST_384_TOP);
708
709	/*S1*/
710	nist_set_256(t_d, buf, 0, 0, 0, 0, 0, 23-4, 22-4, 21-4);
711		/* left shift */
712		{
713		register BN_ULONG *ap,t,c;
714		ap = t_d;
715		c=0;
716		for (i = 3; i != 0; --i)
717			{
718			t= *ap;
719			*(ap++)=((t<<1)|c)&BN_MASK2;
720			c=(t & BN_TBIT)?1:0;
721			}
722		*ap=c;
723		}
724	carry = (int)bn_add_words(r_d+(128/BN_BITS2), r_d+(128/BN_BITS2),
725		t_d, BN_NIST_256_TOP);
726	/*S2 */
727	carry += (int)bn_add_words(r_d, r_d, buf, BN_NIST_384_TOP);
728	/*S3*/
729	nist_set_384(t_d,buf,20,19,18,17,16,15,14,13,12,23,22,21);
730	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
731	/*S4*/
732	nist_set_384(t_d,buf,19,18,17,16,15,14,13,12,20,0,23,0);
733	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
734	/*S5*/
735	nist_set_384(t_d, buf,0,0,0,0,23,22,21,20,0,0,0,0);
736	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
737	/*S6*/
738	nist_set_384(t_d,buf,0,0,0,0,0,0,23,22,21,0,0,20);
739	carry += (int)bn_add_words(r_d, r_d, t_d, BN_NIST_384_TOP);
740	/*D1*/
741	nist_set_384(t_d,buf,22,21,20,19,18,17,16,15,14,13,12,23);
742	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_384_TOP);
743	/*D2*/
744	nist_set_384(t_d,buf,0,0,0,0,0,0,0,23,22,21,20,0);
745	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_384_TOP);
746	/*D3*/
747	nist_set_384(t_d,buf,0,0,0,0,0,0,0,23,23,0,0,0);
748	carry -= (int)bn_sub_words(r_d, r_d, t_d, BN_NIST_384_TOP);
749
750	/* see BN_nist_mod_224 for explanation */
751	u.f = bn_sub_words;
752	if (carry > 0)
753		carry = (int)bn_sub_words(r_d,r_d,_nist_p_384[carry-1],BN_NIST_384_TOP);
754	else if (carry < 0)
755		{
756		carry = (int)bn_add_words(r_d,r_d,_nist_p_384[-carry-1],BN_NIST_384_TOP);
757		mask = 0-(size_t)carry;
758		u.p = ((size_t)bn_sub_words&mask) | ((size_t)bn_add_words&~mask);
759		}
760	else
761		carry = 1;
762
763	mask  = 0-(size_t)(*u.f)(c_d,r_d,_nist_p_384[0],BN_NIST_384_TOP);
764	mask &= 0-(size_t)carry;
765	res   = (BN_ULONG *)(((size_t)c_d&~mask) | ((size_t)r_d&mask));
766	nist_cp_bn(r_d, res, BN_NIST_384_TOP);
767	r->top = BN_NIST_384_TOP;
768	bn_correct_top(r);
769
770	return 1;
771	}
772
773#define BN_NIST_521_RSHIFT	(521%BN_BITS2)
774#define BN_NIST_521_LSHIFT	(BN_BITS2-BN_NIST_521_RSHIFT)
775#define BN_NIST_521_TOP_MASK	((BN_ULONG)BN_MASK2>>BN_NIST_521_LSHIFT)
776
777int BN_nist_mod_521(BIGNUM *r, const BIGNUM *a, const BIGNUM *field,
778	BN_CTX *ctx)
779	{
780	int	top = a->top, i;
781	BN_ULONG *r_d, *a_d = a->d,
782		 t_d[BN_NIST_521_TOP],
783		 val,tmp,*res;
784	size_t	mask;
785	static const BIGNUM _bignum_nist_p_521_sqr = {
786		(BN_ULONG *)_nist_p_521_sqr,
787		sizeof(_nist_p_521_sqr)/sizeof(_nist_p_521_sqr[0]),
788		sizeof(_nist_p_521_sqr)/sizeof(_nist_p_521_sqr[0]),
789		0,BN_FLG_STATIC_DATA };
790
791	field = &_bignum_nist_p_521; /* just to make sure */
792
793 	if (BN_is_negative(a) || BN_ucmp(a,&_bignum_nist_p_521_sqr)>=0)
794		return BN_nnmod(r, a, field, ctx);
795
796	i = BN_ucmp(field, a);
797	if (i == 0)
798		{
799		BN_zero(r);
800		return 1;
801		}
802	else if (i > 0)
803		return (r == a)? 1 : (BN_copy(r ,a) != NULL);
804
805	if (r != a)
806		{
807		if (!bn_wexpand(r,BN_NIST_521_TOP))
808			return 0;
809		r_d = r->d;
810		nist_cp_bn(r_d,a_d, BN_NIST_521_TOP);
811		}
812	else
813		r_d = a_d;
814
815	/* upper 521 bits, copy ... */
816	nist_cp_bn_0(t_d,a_d + (BN_NIST_521_TOP-1), top - (BN_NIST_521_TOP-1),BN_NIST_521_TOP);
817	/* ... and right shift */
818	for (val=t_d[0],i=0; i<BN_NIST_521_TOP-1; i++)
819		{
820		tmp = val>>BN_NIST_521_RSHIFT;
821		val = t_d[i+1];
822		t_d[i] = (tmp | val<<BN_NIST_521_LSHIFT) & BN_MASK2;
823		}
824	t_d[i] = val>>BN_NIST_521_RSHIFT;
825	/* lower 521 bits */
826	r_d[i] &= BN_NIST_521_TOP_MASK;
827
828	bn_add_words(r_d,r_d,t_d,BN_NIST_521_TOP);
829	mask = 0-(size_t)bn_sub_words(t_d,r_d,_nist_p_521,BN_NIST_521_TOP);
830	res  = (BN_ULONG *)(((size_t)t_d&~mask) | ((size_t)r_d&mask));
831	nist_cp_bn(r_d,res,BN_NIST_521_TOP);
832	r->top = BN_NIST_521_TOP;
833	bn_correct_top(r);
834
835	return 1;
836	}
837