/src/libressl/crypto/bn/bn_div.c
Line | Count | Source (jump to first uncovered line) |
1 | | /* $OpenBSD: bn_div.c,v 1.25 2017/01/29 17:49:22 beck Exp $ */ |
2 | | /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com) |
3 | | * All rights reserved. |
4 | | * |
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. |
8 | | * |
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). |
15 | | * |
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. |
22 | | * |
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)" |
40 | | * |
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. |
52 | | * |
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.] |
57 | | */ |
58 | | |
59 | | #include <stdio.h> |
60 | | |
61 | | #include <openssl/opensslconf.h> |
62 | | |
63 | | #include <openssl/bn.h> |
64 | | #include <openssl/err.h> |
65 | | |
66 | | #include "bn_lcl.h" |
67 | | |
68 | | #if !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM) \ |
69 | | && !defined(BN_DIV3W) |
70 | | # if defined(__GNUC__) && __GNUC__>=2 |
71 | | # if defined(__i386) || defined (__i386__) |
72 | | /* |
73 | | * There were two reasons for implementing this template: |
74 | | * - GNU C generates a call to a function (__udivdi3 to be exact) |
75 | | * in reply to ((((BN_ULLONG)n0)<<BN_BITS2)|n1)/d0 (I fail to |
76 | | * understand why...); |
77 | | * - divl doesn't only calculate quotient, but also leaves |
78 | | * remainder in %edx which we can definitely use here:-) |
79 | | * |
80 | | * <appro@fy.chalmers.se> |
81 | | */ |
82 | | #undef bn_div_words |
83 | | # define bn_div_words(n0,n1,d0) \ |
84 | | ({ asm volatile ( \ |
85 | | "divl %4" \ |
86 | | : "=a"(q), "=d"(rem) \ |
87 | | : "a"(n1), "d"(n0), "g"(d0) \ |
88 | | : "cc"); \ |
89 | | q; \ |
90 | | }) |
91 | | # define REMAINDER_IS_ALREADY_CALCULATED |
92 | | # elif defined(__x86_64) && defined(_LP64) |
93 | | /* |
94 | | * Same story here, but it's 128-bit by 64-bit division. Wow! |
95 | | * <appro@fy.chalmers.se> |
96 | | */ |
97 | | # undef bn_div_words |
98 | | # define bn_div_words(n0,n1,d0) \ |
99 | 0 | ({ asm volatile ( \ |
100 | 0 | "divq %4" \ |
101 | 0 | : "=a"(q), "=d"(rem) \ |
102 | 0 | : "a"(n1), "d"(n0), "g"(d0) \ |
103 | 0 | : "cc"); \ |
104 | 0 | q; \ |
105 | 0 | }) |
106 | | # define REMAINDER_IS_ALREADY_CALCULATED |
107 | | # endif /* __<cpu> */ |
108 | | # endif /* __GNUC__ */ |
109 | | #endif /* OPENSSL_NO_ASM */ |
110 | | |
111 | | |
112 | | /* BN_div computes dv := num / divisor, rounding towards |
113 | | * zero, and sets up rm such that dv*divisor + rm = num holds. |
114 | | * Thus: |
115 | | * dv->neg == num->neg ^ divisor->neg (unless the result is zero) |
116 | | * rm->neg == num->neg (unless the remainder is zero) |
117 | | * If 'dv' or 'rm' is NULL, the respective value is not returned. |
118 | | */ |
119 | | static int |
120 | | BN_div_internal(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor, |
121 | | BN_CTX *ctx, int ct) |
122 | 0 | { |
123 | 0 | int norm_shift, i, loop; |
124 | 0 | BIGNUM *tmp, wnum, *snum, *sdiv, *res; |
125 | 0 | BN_ULONG *resp, *wnump; |
126 | 0 | BN_ULONG d0, d1; |
127 | 0 | int num_n, div_n; |
128 | 0 | int no_branch = 0; |
129 | | |
130 | | /* Invalid zero-padding would have particularly bad consequences |
131 | | * in the case of 'num', so don't just rely on bn_check_top() for this one |
132 | | * (bn_check_top() works only for BN_DEBUG builds) */ |
133 | 0 | if (num->top > 0 && num->d[num->top - 1] == 0) { |
134 | 0 | BNerror(BN_R_NOT_INITIALIZED); |
135 | 0 | return 0; |
136 | 0 | } |
137 | | |
138 | 0 | bn_check_top(num); |
139 | |
|
140 | 0 | if (ct) |
141 | 0 | no_branch = 1; |
142 | |
|
143 | 0 | bn_check_top(dv); |
144 | 0 | bn_check_top(rm); |
145 | | /* bn_check_top(num); */ /* 'num' has been checked already */ |
146 | 0 | bn_check_top(divisor); |
147 | |
|
148 | 0 | if (BN_is_zero(divisor)) { |
149 | 0 | BNerror(BN_R_DIV_BY_ZERO); |
150 | 0 | return (0); |
151 | 0 | } |
152 | | |
153 | 0 | if (!no_branch && BN_ucmp(num, divisor) < 0) { |
154 | 0 | if (rm != NULL) { |
155 | 0 | if (BN_copy(rm, num) == NULL) |
156 | 0 | return (0); |
157 | 0 | } |
158 | 0 | if (dv != NULL) |
159 | 0 | BN_zero(dv); |
160 | 0 | return (1); |
161 | 0 | } |
162 | | |
163 | 0 | BN_CTX_start(ctx); |
164 | 0 | tmp = BN_CTX_get(ctx); |
165 | 0 | snum = BN_CTX_get(ctx); |
166 | 0 | sdiv = BN_CTX_get(ctx); |
167 | 0 | if (dv == NULL) |
168 | 0 | res = BN_CTX_get(ctx); |
169 | 0 | else |
170 | 0 | res = dv; |
171 | 0 | if (tmp == NULL || snum == NULL || sdiv == NULL || res == NULL) |
172 | 0 | goto err; |
173 | | |
174 | | /* First we normalise the numbers */ |
175 | 0 | norm_shift = BN_BITS2 - ((BN_num_bits(divisor)) % BN_BITS2); |
176 | 0 | if (!(BN_lshift(sdiv, divisor, norm_shift))) |
177 | 0 | goto err; |
178 | 0 | sdiv->neg = 0; |
179 | 0 | norm_shift += BN_BITS2; |
180 | 0 | if (!(BN_lshift(snum, num, norm_shift))) |
181 | 0 | goto err; |
182 | 0 | snum->neg = 0; |
183 | |
|
184 | 0 | if (no_branch) { |
185 | | /* Since we don't know whether snum is larger than sdiv, |
186 | | * we pad snum with enough zeroes without changing its |
187 | | * value. |
188 | | */ |
189 | 0 | if (snum->top <= sdiv->top + 1) { |
190 | 0 | if (bn_wexpand(snum, sdiv->top + 2) == NULL) |
191 | 0 | goto err; |
192 | 0 | for (i = snum->top; i < sdiv->top + 2; i++) |
193 | 0 | snum->d[i] = 0; |
194 | 0 | snum->top = sdiv->top + 2; |
195 | 0 | } else { |
196 | 0 | if (bn_wexpand(snum, snum->top + 1) == NULL) |
197 | 0 | goto err; |
198 | 0 | snum->d[snum->top] = 0; |
199 | 0 | snum->top ++; |
200 | 0 | } |
201 | 0 | } |
202 | | |
203 | 0 | div_n = sdiv->top; |
204 | 0 | num_n = snum->top; |
205 | 0 | loop = num_n - div_n; |
206 | | /* Lets setup a 'window' into snum |
207 | | * This is the part that corresponds to the current |
208 | | * 'area' being divided */ |
209 | 0 | wnum.neg = 0; |
210 | 0 | wnum.d = &(snum->d[loop]); |
211 | 0 | wnum.top = div_n; |
212 | | /* only needed when BN_ucmp messes up the values between top and max */ |
213 | 0 | wnum.dmax = snum->dmax - loop; /* so we don't step out of bounds */ |
214 | 0 | wnum.flags = snum->flags | BN_FLG_STATIC_DATA; |
215 | | |
216 | | /* Get the top 2 words of sdiv */ |
217 | | /* div_n=sdiv->top; */ |
218 | 0 | d0 = sdiv->d[div_n - 1]; |
219 | 0 | d1 = (div_n == 1) ? 0 : sdiv->d[div_n - 2]; |
220 | | |
221 | | /* pointer to the 'top' of snum */ |
222 | 0 | wnump = &(snum->d[num_n - 1]); |
223 | | |
224 | | /* Setup to 'res' */ |
225 | 0 | res->neg = (num->neg ^ divisor->neg); |
226 | 0 | if (!bn_wexpand(res, (loop + 1))) |
227 | 0 | goto err; |
228 | 0 | res->top = loop - no_branch; |
229 | 0 | resp = &(res->d[loop - 1]); |
230 | | |
231 | | /* space for temp */ |
232 | 0 | if (!bn_wexpand(tmp, (div_n + 1))) |
233 | 0 | goto err; |
234 | | |
235 | 0 | if (!no_branch) { |
236 | 0 | if (BN_ucmp(&wnum, sdiv) >= 0) { |
237 | | /* If BN_DEBUG_RAND is defined BN_ucmp changes (via |
238 | | * bn_pollute) the const bignum arguments => |
239 | | * clean the values between top and max again */ |
240 | 0 | bn_clear_top2max(&wnum); |
241 | 0 | bn_sub_words(wnum.d, wnum.d, sdiv->d, div_n); |
242 | 0 | *resp = 1; |
243 | 0 | } else |
244 | 0 | res->top--; |
245 | 0 | } |
246 | | |
247 | | /* if res->top == 0 then clear the neg value otherwise decrease |
248 | | * the resp pointer */ |
249 | 0 | if (res->top == 0) |
250 | 0 | res->neg = 0; |
251 | 0 | else |
252 | 0 | resp--; |
253 | |
|
254 | 0 | for (i = 0; i < loop - 1; i++, wnump--, resp--) { |
255 | 0 | BN_ULONG q, l0; |
256 | | /* the first part of the loop uses the top two words of |
257 | | * snum and sdiv to calculate a BN_ULONG q such that |
258 | | * | wnum - sdiv * q | < sdiv */ |
259 | | #if defined(BN_DIV3W) && !defined(OPENSSL_NO_ASM) |
260 | | BN_ULONG bn_div_3_words(BN_ULONG*, BN_ULONG, BN_ULONG); |
261 | | q = bn_div_3_words(wnump, d1, d0); |
262 | | #else |
263 | 0 | BN_ULONG n0, n1, rem = 0; |
264 | |
|
265 | 0 | n0 = wnump[0]; |
266 | 0 | n1 = wnump[-1]; |
267 | 0 | if (n0 == d0) |
268 | 0 | q = BN_MASK2; |
269 | 0 | else /* n0 < d0 */ |
270 | 0 | { |
271 | | #ifdef BN_LLONG |
272 | | BN_ULLONG t2; |
273 | | |
274 | | #if defined(BN_DIV2W) && !defined(bn_div_words) |
275 | | q = (BN_ULONG)(((((BN_ULLONG)n0) << BN_BITS2)|n1)/d0); |
276 | | #else |
277 | | q = bn_div_words(n0, n1, d0); |
278 | | #endif |
279 | | |
280 | | #ifndef REMAINDER_IS_ALREADY_CALCULATED |
281 | | /* |
282 | | * rem doesn't have to be BN_ULLONG. The least we |
283 | | * know it's less that d0, isn't it? |
284 | | */ |
285 | | rem = (n1 - q * d0) & BN_MASK2; |
286 | | #endif |
287 | | t2 = (BN_ULLONG)d1*q; |
288 | | |
289 | | for (;;) { |
290 | | if (t2 <= ((((BN_ULLONG)rem) << BN_BITS2) | |
291 | | wnump[-2])) |
292 | | break; |
293 | | q--; |
294 | | rem += d0; |
295 | | if (rem < d0) break; /* don't let rem overflow */ |
296 | | t2 -= d1; |
297 | | } |
298 | | #else /* !BN_LLONG */ |
299 | 0 | BN_ULONG t2l, t2h; |
300 | |
|
301 | 0 | q = bn_div_words(n0, n1, d0); |
302 | | #ifndef REMAINDER_IS_ALREADY_CALCULATED |
303 | | rem = (n1 - q*d0)&BN_MASK2; |
304 | | #endif |
305 | |
|
306 | 0 | #if defined(BN_UMULT_LOHI) |
307 | 0 | BN_UMULT_LOHI(t2l, t2h, d1, q); |
308 | | #elif defined(BN_UMULT_HIGH) |
309 | | t2l = d1 * q; |
310 | | t2h = BN_UMULT_HIGH(d1, q); |
311 | | #else |
312 | | { |
313 | | BN_ULONG ql, qh; |
314 | | t2l = LBITS(d1); |
315 | | t2h = HBITS(d1); |
316 | | ql = LBITS(q); |
317 | | qh = HBITS(q); |
318 | | mul64(t2l, t2h, ql, qh); /* t2=(BN_ULLONG)d1*q; */ |
319 | | } |
320 | | #endif |
321 | |
|
322 | 0 | for (;;) { |
323 | 0 | if ((t2h < rem) || |
324 | 0 | ((t2h == rem) && (t2l <= wnump[-2]))) |
325 | 0 | break; |
326 | 0 | q--; |
327 | 0 | rem += d0; |
328 | 0 | if (rem < d0) |
329 | 0 | break; /* don't let rem overflow */ |
330 | 0 | if (t2l < d1) |
331 | 0 | t2h--; |
332 | 0 | t2l -= d1; |
333 | 0 | } |
334 | 0 | #endif /* !BN_LLONG */ |
335 | 0 | } |
336 | 0 | #endif /* !BN_DIV3W */ |
337 | |
|
338 | 0 | l0 = bn_mul_words(tmp->d, sdiv->d, div_n, q); |
339 | 0 | tmp->d[div_n] = l0; |
340 | 0 | wnum.d--; |
341 | | /* ingore top values of the bignums just sub the two |
342 | | * BN_ULONG arrays with bn_sub_words */ |
343 | 0 | if (bn_sub_words(wnum.d, wnum.d, tmp->d, div_n + 1)) { |
344 | | /* Note: As we have considered only the leading |
345 | | * two BN_ULONGs in the calculation of q, sdiv * q |
346 | | * might be greater than wnum (but then (q-1) * sdiv |
347 | | * is less or equal than wnum) |
348 | | */ |
349 | 0 | q--; |
350 | 0 | if (bn_add_words(wnum.d, wnum.d, sdiv->d, div_n)) |
351 | | /* we can't have an overflow here (assuming |
352 | | * that q != 0, but if q == 0 then tmp is |
353 | | * zero anyway) */ |
354 | 0 | (*wnump)++; |
355 | 0 | } |
356 | | /* store part of the result */ |
357 | 0 | *resp = q; |
358 | 0 | } |
359 | 0 | bn_correct_top(snum); |
360 | 0 | if (rm != NULL) { |
361 | | /* Keep a copy of the neg flag in num because if rm==num |
362 | | * BN_rshift() will overwrite it. |
363 | | */ |
364 | 0 | int neg = num->neg; |
365 | 0 | BN_rshift(rm, snum, norm_shift); |
366 | 0 | if (!BN_is_zero(rm)) |
367 | 0 | rm->neg = neg; |
368 | 0 | bn_check_top(rm); |
369 | 0 | } |
370 | 0 | if (no_branch) |
371 | 0 | bn_correct_top(res); |
372 | 0 | BN_CTX_end(ctx); |
373 | 0 | return (1); |
374 | | |
375 | 0 | err: |
376 | 0 | bn_check_top(rm); |
377 | 0 | BN_CTX_end(ctx); |
378 | 0 | return (0); |
379 | 0 | } |
380 | | |
381 | | int |
382 | | BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor, |
383 | | BN_CTX *ctx) |
384 | 0 | { |
385 | 0 | int ct = ((BN_get_flags(num, BN_FLG_CONSTTIME) != 0) || |
386 | 0 | (BN_get_flags(divisor, BN_FLG_CONSTTIME) != 0)); |
387 | |
|
388 | 0 | return BN_div_internal(dv, rm, num, divisor, ctx, ct); |
389 | 0 | } |
390 | | |
391 | | int |
392 | | BN_div_nonct(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor, |
393 | | BN_CTX *ctx) |
394 | 0 | { |
395 | 0 | return BN_div_internal(dv, rm, num, divisor, ctx, 0); |
396 | 0 | } |
397 | | |
398 | | int |
399 | | BN_div_ct(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor, |
400 | | BN_CTX *ctx) |
401 | 0 | { |
402 | 0 | return BN_div_internal(dv, rm, num, divisor, ctx, 1); |
403 | 0 | } |