openssl: update to 1.0.2d
[tomato.git] / release / src / router / openssl / crypto / bn / bn_div.c
blob72e6ce3f74c067acfcda28580b185a6337ba38a0
1 /* crypto/bn/bn_div.c */
2 /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
3 * All rights reserved.
5 * This package is an SSL implementation written
6 * by Eric Young (eay@cryptsoft.com).
7 * The implementation was written so as to conform with Netscapes SSL.
9 * This library is free for commercial and non-commercial use as long as
10 * the following conditions are aheared to. The following conditions
11 * apply to all code found in this distribution, be it the RC4, RSA,
12 * lhash, DES, etc., code; not just the SSL code. The SSL documentation
13 * included with this distribution is covered by the same copyright terms
14 * except that the holder is Tim Hudson (tjh@cryptsoft.com).
16 * Copyright remains Eric Young's, and as such any Copyright notices in
17 * the code are not to be removed.
18 * If this package is used in a product, Eric Young should be given attribution
19 * as the author of the parts of the library used.
20 * This can be in the form of a textual message at program startup or
21 * in documentation (online or textual) provided with the package.
23 * Redistribution and use in source and binary forms, with or without
24 * modification, are permitted provided that the following conditions
25 * are met:
26 * 1. Redistributions of source code must retain the copyright
27 * notice, this list of conditions and the following disclaimer.
28 * 2. Redistributions in binary form must reproduce the above copyright
29 * notice, this list of conditions and the following disclaimer in the
30 * documentation and/or other materials provided with the distribution.
31 * 3. All advertising materials mentioning features or use of this software
32 * must display the following acknowledgement:
33 * "This product includes cryptographic software written by
34 * Eric Young (eay@cryptsoft.com)"
35 * The word 'cryptographic' can be left out if the rouines from the library
36 * being used are not cryptographic related :-).
37 * 4. If you include any Windows specific code (or a derivative thereof) from
38 * the apps directory (application code) you must include an acknowledgement:
39 * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
41 * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
42 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
43 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
44 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
45 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
46 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
47 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
48 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
49 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
50 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
51 * SUCH DAMAGE.
53 * The licence and distribution terms for any publically available version or
54 * derivative of this code cannot be changed. i.e. this code cannot simply be
55 * copied and put under another distribution licence
56 * [including the GNU Public Licence.]
59 #include <stdio.h>
60 #include <openssl/bn.h>
61 #include "cryptlib.h"
62 #include "bn_lcl.h"
64 /* The old slow way */
65 #if 0
66 int BN_div(BIGNUM *dv, BIGNUM *rem, const BIGNUM *m, const BIGNUM *d,
67 BN_CTX *ctx)
69 int i, nm, nd;
70 int ret = 0;
71 BIGNUM *D;
73 bn_check_top(m);
74 bn_check_top(d);
75 if (BN_is_zero(d)) {
76 BNerr(BN_F_BN_DIV, BN_R_DIV_BY_ZERO);
77 return (0);
80 if (BN_ucmp(m, d) < 0) {
81 if (rem != NULL) {
82 if (BN_copy(rem, m) == NULL)
83 return (0);
85 if (dv != NULL)
86 BN_zero(dv);
87 return (1);
90 BN_CTX_start(ctx);
91 D = BN_CTX_get(ctx);
92 if (dv == NULL)
93 dv = BN_CTX_get(ctx);
94 if (rem == NULL)
95 rem = BN_CTX_get(ctx);
96 if (D == NULL || dv == NULL || rem == NULL)
97 goto end;
99 nd = BN_num_bits(d);
100 nm = BN_num_bits(m);
101 if (BN_copy(D, d) == NULL)
102 goto end;
103 if (BN_copy(rem, m) == NULL)
104 goto end;
107 * The next 2 are needed so we can do a dv->d[0]|=1 later since
108 * BN_lshift1 will only work once there is a value :-)
110 BN_zero(dv);
111 if (bn_wexpand(dv, 1) == NULL)
112 goto end;
113 dv->top = 1;
115 if (!BN_lshift(D, D, nm - nd))
116 goto end;
117 for (i = nm - nd; i >= 0; i--) {
118 if (!BN_lshift1(dv, dv))
119 goto end;
120 if (BN_ucmp(rem, D) >= 0) {
121 dv->d[0] |= 1;
122 if (!BN_usub(rem, rem, D))
123 goto end;
125 /* CAN IMPROVE (and have now :=) */
126 if (!BN_rshift1(D, D))
127 goto end;
129 rem->neg = BN_is_zero(rem) ? 0 : m->neg;
130 dv->neg = m->neg ^ d->neg;
131 ret = 1;
132 end:
133 BN_CTX_end(ctx);
134 return (ret);
137 #else
139 # if !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM) \
140 && !defined(PEDANTIC) && !defined(BN_DIV3W)
141 # if defined(__GNUC__) && __GNUC__>=2
142 # if defined(__i386) || defined (__i386__)
144 * There were two reasons for implementing this template:
145 * - GNU C generates a call to a function (__udivdi3 to be exact)
146 * in reply to ((((BN_ULLONG)n0)<<BN_BITS2)|n1)/d0 (I fail to
147 * understand why...);
148 * - divl doesn't only calculate quotient, but also leaves
149 * remainder in %edx which we can definitely use here:-)
151 * <appro@fy.chalmers.se>
153 # undef bn_div_words
154 # define bn_div_words(n0,n1,d0) \
155 ({ asm volatile ( \
156 "divl %4" \
157 : "=a"(q), "=d"(rem) \
158 : "a"(n1), "d"(n0), "g"(d0) \
159 : "cc"); \
160 q; \
162 # define REMAINDER_IS_ALREADY_CALCULATED
163 # elif defined(__x86_64) && defined(SIXTY_FOUR_BIT_LONG)
165 * Same story here, but it's 128-bit by 64-bit division. Wow!
166 * <appro@fy.chalmers.se>
168 # undef bn_div_words
169 # define bn_div_words(n0,n1,d0) \
170 ({ asm volatile ( \
171 "divq %4" \
172 : "=a"(q), "=d"(rem) \
173 : "a"(n1), "d"(n0), "g"(d0) \
174 : "cc"); \
175 q; \
177 # define REMAINDER_IS_ALREADY_CALCULATED
178 # endif /* __<cpu> */
179 # endif /* __GNUC__ */
180 # endif /* OPENSSL_NO_ASM */
183 * BN_div computes dv := num / divisor, rounding towards
184 * zero, and sets up rm such that dv*divisor + rm = num holds.
185 * Thus:
186 * dv->neg == num->neg ^ divisor->neg (unless the result is zero)
187 * rm->neg == num->neg (unless the remainder is zero)
188 * If 'dv' or 'rm' is NULL, the respective value is not returned.
190 int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
191 BN_CTX *ctx)
193 int norm_shift, i, loop;
194 BIGNUM *tmp, wnum, *snum, *sdiv, *res;
195 BN_ULONG *resp, *wnump;
196 BN_ULONG d0, d1;
197 int num_n, div_n;
198 int no_branch = 0;
201 * Invalid zero-padding would have particularly bad consequences so don't
202 * just rely on bn_check_top() here (bn_check_top() works only for
203 * BN_DEBUG builds)
205 if ((num->top > 0 && num->d[num->top - 1] == 0) ||
206 (divisor->top > 0 && divisor->d[divisor->top - 1] == 0)) {
207 BNerr(BN_F_BN_DIV, BN_R_NOT_INITIALIZED);
208 return 0;
211 bn_check_top(num);
212 bn_check_top(divisor);
214 if ((BN_get_flags(num, BN_FLG_CONSTTIME) != 0)
215 || (BN_get_flags(divisor, BN_FLG_CONSTTIME) != 0)) {
216 no_branch = 1;
219 bn_check_top(dv);
220 bn_check_top(rm);
221 /*- bn_check_top(num); *//*
222 * 'num' has been checked already
224 /*- bn_check_top(divisor); *//*
225 * 'divisor' has been checked already
228 if (BN_is_zero(divisor)) {
229 BNerr(BN_F_BN_DIV, BN_R_DIV_BY_ZERO);
230 return (0);
233 if (!no_branch && BN_ucmp(num, divisor) < 0) {
234 if (rm != NULL) {
235 if (BN_copy(rm, num) == NULL)
236 return (0);
238 if (dv != NULL)
239 BN_zero(dv);
240 return (1);
243 BN_CTX_start(ctx);
244 tmp = BN_CTX_get(ctx);
245 snum = BN_CTX_get(ctx);
246 sdiv = BN_CTX_get(ctx);
247 if (dv == NULL)
248 res = BN_CTX_get(ctx);
249 else
250 res = dv;
251 if (sdiv == NULL || res == NULL || tmp == NULL || snum == NULL)
252 goto err;
254 /* First we normalise the numbers */
255 norm_shift = BN_BITS2 - ((BN_num_bits(divisor)) % BN_BITS2);
256 if (!(BN_lshift(sdiv, divisor, norm_shift)))
257 goto err;
258 sdiv->neg = 0;
259 norm_shift += BN_BITS2;
260 if (!(BN_lshift(snum, num, norm_shift)))
261 goto err;
262 snum->neg = 0;
264 if (no_branch) {
266 * Since we don't know whether snum is larger than sdiv, we pad snum
267 * with enough zeroes without changing its value.
269 if (snum->top <= sdiv->top + 1) {
270 if (bn_wexpand(snum, sdiv->top + 2) == NULL)
271 goto err;
272 for (i = snum->top; i < sdiv->top + 2; i++)
273 snum->d[i] = 0;
274 snum->top = sdiv->top + 2;
275 } else {
276 if (bn_wexpand(snum, snum->top + 1) == NULL)
277 goto err;
278 snum->d[snum->top] = 0;
279 snum->top++;
283 div_n = sdiv->top;
284 num_n = snum->top;
285 loop = num_n - div_n;
287 * Lets setup a 'window' into snum This is the part that corresponds to
288 * the current 'area' being divided
290 wnum.neg = 0;
291 wnum.d = &(snum->d[loop]);
292 wnum.top = div_n;
294 * only needed when BN_ucmp messes up the values between top and max
296 wnum.dmax = snum->dmax - loop; /* so we don't step out of bounds */
298 /* Get the top 2 words of sdiv */
299 /* div_n=sdiv->top; */
300 d0 = sdiv->d[div_n - 1];
301 d1 = (div_n == 1) ? 0 : sdiv->d[div_n - 2];
303 /* pointer to the 'top' of snum */
304 wnump = &(snum->d[num_n - 1]);
306 /* Setup to 'res' */
307 res->neg = (num->neg ^ divisor->neg);
308 if (!bn_wexpand(res, (loop + 1)))
309 goto err;
310 res->top = loop - no_branch;
311 resp = &(res->d[loop - 1]);
313 /* space for temp */
314 if (!bn_wexpand(tmp, (div_n + 1)))
315 goto err;
317 if (!no_branch) {
318 if (BN_ucmp(&wnum, sdiv) >= 0) {
320 * If BN_DEBUG_RAND is defined BN_ucmp changes (via bn_pollute)
321 * the const bignum arguments => clean the values between top and
322 * max again
324 bn_clear_top2max(&wnum);
325 bn_sub_words(wnum.d, wnum.d, sdiv->d, div_n);
326 *resp = 1;
327 } else
328 res->top--;
332 * if res->top == 0 then clear the neg value otherwise decrease the resp
333 * pointer
335 if (res->top == 0)
336 res->neg = 0;
337 else
338 resp--;
340 for (i = 0; i < loop - 1; i++, wnump--, resp--) {
341 BN_ULONG q, l0;
343 * the first part of the loop uses the top two words of snum and sdiv
344 * to calculate a BN_ULONG q such that | wnum - sdiv * q | < sdiv
346 # if defined(BN_DIV3W) && !defined(OPENSSL_NO_ASM)
347 BN_ULONG bn_div_3_words(BN_ULONG *, BN_ULONG, BN_ULONG);
348 q = bn_div_3_words(wnump, d1, d0);
349 # else
350 BN_ULONG n0, n1, rem = 0;
352 n0 = wnump[0];
353 n1 = wnump[-1];
354 if (n0 == d0)
355 q = BN_MASK2;
356 else { /* n0 < d0 */
358 # ifdef BN_LLONG
359 BN_ULLONG t2;
361 # if defined(BN_LLONG) && defined(BN_DIV2W) && !defined(bn_div_words)
362 q = (BN_ULONG)(((((BN_ULLONG) n0) << BN_BITS2) | n1) / d0);
363 # else
364 q = bn_div_words(n0, n1, d0);
365 # ifdef BN_DEBUG_LEVITTE
366 fprintf(stderr, "DEBUG: bn_div_words(0x%08X,0x%08X,0x%08\
367 X) -> 0x%08X\n", n0, n1, d0, q);
368 # endif
369 # endif
371 # ifndef REMAINDER_IS_ALREADY_CALCULATED
373 * rem doesn't have to be BN_ULLONG. The least we
374 * know it's less that d0, isn't it?
376 rem = (n1 - q * d0) & BN_MASK2;
377 # endif
378 t2 = (BN_ULLONG) d1 *q;
380 for (;;) {
381 if (t2 <= ((((BN_ULLONG) rem) << BN_BITS2) | wnump[-2]))
382 break;
383 q--;
384 rem += d0;
385 if (rem < d0)
386 break; /* don't let rem overflow */
387 t2 -= d1;
389 # else /* !BN_LLONG */
390 BN_ULONG t2l, t2h;
392 q = bn_div_words(n0, n1, d0);
393 # ifdef BN_DEBUG_LEVITTE
394 fprintf(stderr, "DEBUG: bn_div_words(0x%08X,0x%08X,0x%08\
395 X) -> 0x%08X\n", n0, n1, d0, q);
396 # endif
397 # ifndef REMAINDER_IS_ALREADY_CALCULATED
398 rem = (n1 - q * d0) & BN_MASK2;
399 # endif
401 # if defined(BN_UMULT_LOHI)
402 BN_UMULT_LOHI(t2l, t2h, d1, q);
403 # elif defined(BN_UMULT_HIGH)
404 t2l = d1 * q;
405 t2h = BN_UMULT_HIGH(d1, q);
406 # else
408 BN_ULONG ql, qh;
409 t2l = LBITS(d1);
410 t2h = HBITS(d1);
411 ql = LBITS(q);
412 qh = HBITS(q);
413 mul64(t2l, t2h, ql, qh); /* t2=(BN_ULLONG)d1*q; */
415 # endif
417 for (;;) {
418 if ((t2h < rem) || ((t2h == rem) && (t2l <= wnump[-2])))
419 break;
420 q--;
421 rem += d0;
422 if (rem < d0)
423 break; /* don't let rem overflow */
424 if (t2l < d1)
425 t2h--;
426 t2l -= d1;
428 # endif /* !BN_LLONG */
430 # endif /* !BN_DIV3W */
432 l0 = bn_mul_words(tmp->d, sdiv->d, div_n, q);
433 tmp->d[div_n] = l0;
434 wnum.d--;
436 * ingore top values of the bignums just sub the two BN_ULONG arrays
437 * with bn_sub_words
439 if (bn_sub_words(wnum.d, wnum.d, tmp->d, div_n + 1)) {
441 * Note: As we have considered only the leading two BN_ULONGs in
442 * the calculation of q, sdiv * q might be greater than wnum (but
443 * then (q-1) * sdiv is less or equal than wnum)
445 q--;
446 if (bn_add_words(wnum.d, wnum.d, sdiv->d, div_n))
448 * we can't have an overflow here (assuming that q != 0, but
449 * if q == 0 then tmp is zero anyway)
451 (*wnump)++;
453 /* store part of the result */
454 *resp = q;
456 bn_correct_top(snum);
457 if (rm != NULL) {
459 * Keep a copy of the neg flag in num because if rm==num BN_rshift()
460 * will overwrite it.
462 int neg = num->neg;
463 BN_rshift(rm, snum, norm_shift);
464 if (!BN_is_zero(rm))
465 rm->neg = neg;
466 bn_check_top(rm);
468 if (no_branch)
469 bn_correct_top(res);
470 BN_CTX_end(ctx);
471 return (1);
472 err:
473 bn_check_top(rm);
474 BN_CTX_end(ctx);
475 return (0);
477 #endif