• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
2  * All rights reserved.
3  *
4  * This package is an SSL implementation written
5  * by Eric Young (eay@cryptsoft.com).
6  * The implementation was written so as to conform with Netscapes SSL.
7  *
8  * This library is free for commercial and non-commercial use as long as
9  * the following conditions are aheared to.  The following conditions
10  * apply to all code found in this distribution, be it the RC4, RSA,
11  * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
12  * included with this distribution is covered by the same copyright terms
13  * except that the holder is Tim Hudson (tjh@cryptsoft.com).
14  *
15  * Copyright remains Eric Young's, and as such any Copyright notices in
16  * the code are not to be removed.
17  * If this package is used in a product, Eric Young should be given attribution
18  * as the author of the parts of the library used.
19  * This can be in the form of a textual message at program startup or
20  * in documentation (online or textual) provided with the package.
21  *
22  * Redistribution and use in source and binary forms, with or without
23  * modification, are permitted provided that the following conditions
24  * are met:
25  * 1. Redistributions of source code must retain the copyright
26  *    notice, this list of conditions and the following disclaimer.
27  * 2. Redistributions in binary form must reproduce the above copyright
28  *    notice, this list of conditions and the following disclaimer in the
29  *    documentation and/or other materials provided with the distribution.
30  * 3. All advertising materials mentioning features or use of this software
31  *    must display the following acknowledgement:
32  *    "This product includes cryptographic software written by
33  *     Eric Young (eay@cryptsoft.com)"
34  *    The word 'cryptographic' can be left out if the rouines from the library
35  *    being used are not cryptographic related :-).
36  * 4. If you include any Windows specific code (or a derivative thereof) from
37  *    the apps directory (application code) you must include an acknowledgement:
38  *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
39  *
40  * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
41  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
42  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
43  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
44  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
45  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
46  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
47  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
48  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
49  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
50  * SUCH DAMAGE.
51  *
52  * The licence and distribution terms for any publically available version or
53  * derivative of this code cannot be changed.  i.e. this code cannot simply be
54  * copied and put under another distribution licence
55  * [including the GNU Public Licence.] */
56 
57 #include <openssl/bn.h>
58 
59 #include <assert.h>
60 #include <limits.h>
61 
62 #include <openssl/err.h>
63 
64 #include "internal.h"
65 
66 
67 #if !defined(BN_CAN_DIVIDE_ULLONG) && !defined(BN_CAN_USE_INLINE_ASM)
68 // bn_div_words divides a double-width |h|,|l| by |d| and returns the result,
69 // which must fit in a |BN_ULONG|.
bn_div_words(BN_ULONG h,BN_ULONG l,BN_ULONG d)70 static BN_ULONG bn_div_words(BN_ULONG h, BN_ULONG l, BN_ULONG d) {
71   BN_ULONG dh, dl, q, ret = 0, th, tl, t;
72   int i, count = 2;
73 
74   if (d == 0) {
75     return BN_MASK2;
76   }
77 
78   i = BN_num_bits_word(d);
79   assert((i == BN_BITS2) || (h <= (BN_ULONG)1 << i));
80 
81   i = BN_BITS2 - i;
82   if (h >= d) {
83     h -= d;
84   }
85 
86   if (i) {
87     d <<= i;
88     h = (h << i) | (l >> (BN_BITS2 - i));
89     l <<= i;
90   }
91   dh = (d & BN_MASK2h) >> BN_BITS4;
92   dl = (d & BN_MASK2l);
93   for (;;) {
94     if ((h >> BN_BITS4) == dh) {
95       q = BN_MASK2l;
96     } else {
97       q = h / dh;
98     }
99 
100     th = q * dh;
101     tl = dl * q;
102     for (;;) {
103       t = h - th;
104       if ((t & BN_MASK2h) ||
105           ((tl) <= ((t << BN_BITS4) | ((l & BN_MASK2h) >> BN_BITS4)))) {
106         break;
107       }
108       q--;
109       th -= dh;
110       tl -= dl;
111     }
112     t = (tl >> BN_BITS4);
113     tl = (tl << BN_BITS4) & BN_MASK2h;
114     th += t;
115 
116     if (l < tl) {
117       th++;
118     }
119     l -= tl;
120     if (h < th) {
121       h += d;
122       q--;
123     }
124     h -= th;
125 
126     if (--count == 0) {
127       break;
128     }
129 
130     ret = q << BN_BITS4;
131     h = (h << BN_BITS4) | (l >> BN_BITS4);
132     l = (l & BN_MASK2l) << BN_BITS4;
133   }
134 
135   ret |= q;
136   return ret;
137 }
138 #endif  // !defined(BN_CAN_DIVIDE_ULLONG) && !defined(BN_CAN_USE_INLINE_ASM)
139 
bn_div_rem_words(BN_ULONG * quotient_out,BN_ULONG * rem_out,BN_ULONG n0,BN_ULONG n1,BN_ULONG d0)140 static inline void bn_div_rem_words(BN_ULONG *quotient_out, BN_ULONG *rem_out,
141                                     BN_ULONG n0, BN_ULONG n1, BN_ULONG d0) {
142   // GCC and Clang generate function calls to |__udivdi3| and |__umoddi3| when
143   // the |BN_ULLONG|-based C code is used.
144   //
145   // GCC bugs:
146   //   * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=14224
147   //   * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=43721
148   //   * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54183
149   //   * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58897
150   //   * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65668
151   //
152   // Clang bugs:
153   //   * https://llvm.org/bugs/show_bug.cgi?id=6397
154   //   * https://llvm.org/bugs/show_bug.cgi?id=12418
155   //
156   // These issues aren't specific to x86 and x86_64, so it might be worthwhile
157   // to add more assembly language implementations.
158 #if defined(BN_CAN_USE_INLINE_ASM) && defined(OPENSSL_X86)
159   __asm__ volatile("divl %4"
160                    : "=a"(*quotient_out), "=d"(*rem_out)
161                    : "a"(n1), "d"(n0), "rm"(d0)
162                    : "cc");
163 #elif defined(BN_CAN_USE_INLINE_ASM) && defined(OPENSSL_X86_64)
164   __asm__ volatile("divq %4"
165                    : "=a"(*quotient_out), "=d"(*rem_out)
166                    : "a"(n1), "d"(n0), "rm"(d0)
167                    : "cc");
168 #else
169 #if defined(BN_CAN_DIVIDE_ULLONG)
170   BN_ULLONG n = (((BN_ULLONG)n0) << BN_BITS2) | n1;
171   *quotient_out = (BN_ULONG)(n / d0);
172 #else
173   *quotient_out = bn_div_words(n0, n1, d0);
174 #endif
175   *rem_out = n1 - (*quotient_out * d0);
176 #endif
177 }
178 
179 // BN_div computes "quotient := numerator / divisor", rounding towards zero,
180 // and sets up |rem| such that "quotient * divisor + rem = numerator" holds.
181 //
182 // Thus:
183 //
184 //     quotient->neg == numerator->neg ^ divisor->neg
185 //        (unless the result is zero)
186 //     rem->neg == numerator->neg
187 //        (unless the remainder is zero)
188 //
189 // If |quotient| or |rem| is NULL, the respective value is not returned.
190 //
191 // This was specifically designed to contain fewer branches that may leak
192 // sensitive information; see "New Branch Prediction Vulnerabilities in OpenSSL
193 // and Necessary Software Countermeasures" by Onur Acıçmez, Shay Gueron, and
194 // Jean-Pierre Seifert.
BN_div(BIGNUM * quotient,BIGNUM * rem,const BIGNUM * numerator,const BIGNUM * divisor,BN_CTX * ctx)195 int BN_div(BIGNUM *quotient, BIGNUM *rem, const BIGNUM *numerator,
196            const BIGNUM *divisor, BN_CTX *ctx) {
197   int norm_shift, loop;
198   BIGNUM wnum;
199   BN_ULONG *resp, *wnump;
200   BN_ULONG d0, d1;
201   int num_n, div_n;
202 
203   // This function relies on the historical minimal-width |BIGNUM| invariant.
204   // It is already not constant-time (constant-time reductions should use
205   // Montgomery logic), so we shrink all inputs and intermediate values to
206   // retain the previous behavior.
207 
208   // Invalid zero-padding would have particularly bad consequences.
209   int numerator_width = bn_minimal_width(numerator);
210   int divisor_width = bn_minimal_width(divisor);
211   if ((numerator_width > 0 && numerator->d[numerator_width - 1] == 0) ||
212       (divisor_width > 0 && divisor->d[divisor_width - 1] == 0)) {
213     OPENSSL_PUT_ERROR(BN, BN_R_NOT_INITIALIZED);
214     return 0;
215   }
216 
217   if (BN_is_zero(divisor)) {
218     OPENSSL_PUT_ERROR(BN, BN_R_DIV_BY_ZERO);
219     return 0;
220   }
221 
222   BN_CTX_start(ctx);
223   BIGNUM *tmp = BN_CTX_get(ctx);
224   BIGNUM *snum = BN_CTX_get(ctx);
225   BIGNUM *sdiv = BN_CTX_get(ctx);
226   BIGNUM *res = NULL;
227   if (quotient == NULL) {
228     res = BN_CTX_get(ctx);
229   } else {
230     res = quotient;
231   }
232   if (sdiv == NULL || res == NULL) {
233     goto err;
234   }
235 
236   // First we normalise the numbers
237   norm_shift = BN_BITS2 - (BN_num_bits(divisor) % BN_BITS2);
238   if (!BN_lshift(sdiv, divisor, norm_shift)) {
239     goto err;
240   }
241   bn_set_minimal_width(sdiv);
242   sdiv->neg = 0;
243   norm_shift += BN_BITS2;
244   if (!BN_lshift(snum, numerator, norm_shift)) {
245     goto err;
246   }
247   bn_set_minimal_width(snum);
248   snum->neg = 0;
249 
250   // Since we don't want to have special-case logic for the case where snum is
251   // larger than sdiv, we pad snum with enough zeroes without changing its
252   // value.
253   if (snum->width <= sdiv->width + 1) {
254     if (!bn_wexpand(snum, sdiv->width + 2)) {
255       goto err;
256     }
257     for (int i = snum->width; i < sdiv->width + 2; i++) {
258       snum->d[i] = 0;
259     }
260     snum->width = sdiv->width + 2;
261   } else {
262     if (!bn_wexpand(snum, snum->width + 1)) {
263       goto err;
264     }
265     snum->d[snum->width] = 0;
266     snum->width++;
267   }
268 
269   div_n = sdiv->width;
270   num_n = snum->width;
271   loop = num_n - div_n;
272   // Lets setup a 'window' into snum
273   // This is the part that corresponds to the current
274   // 'area' being divided
275   wnum.neg = 0;
276   wnum.d = &(snum->d[loop]);
277   wnum.width = div_n;
278   // only needed when BN_ucmp messes up the values between width and max
279   wnum.dmax = snum->dmax - loop;  // so we don't step out of bounds
280 
281   // Get the top 2 words of sdiv
282   // div_n=sdiv->width;
283   d0 = sdiv->d[div_n - 1];
284   d1 = (div_n == 1) ? 0 : sdiv->d[div_n - 2];
285 
286   // pointer to the 'top' of snum
287   wnump = &(snum->d[num_n - 1]);
288 
289   // Setup to 'res'
290   res->neg = (numerator->neg ^ divisor->neg);
291   if (!bn_wexpand(res, loop + 1)) {
292     goto err;
293   }
294   res->width = loop - 1;
295   resp = &(res->d[loop - 1]);
296 
297   // space for temp
298   if (!bn_wexpand(tmp, div_n + 1)) {
299     goto err;
300   }
301 
302   // if res->width == 0 then clear the neg value otherwise decrease
303   // the resp pointer
304   if (res->width == 0) {
305     res->neg = 0;
306   } else {
307     resp--;
308   }
309 
310   for (int i = 0; i < loop - 1; i++, wnump--, resp--) {
311     BN_ULONG q, l0;
312     // the first part of the loop uses the top two words of snum and sdiv to
313     // calculate a BN_ULONG q such that | wnum - sdiv * q | < sdiv
314     BN_ULONG n0, n1, rm = 0;
315 
316     n0 = wnump[0];
317     n1 = wnump[-1];
318     if (n0 == d0) {
319       q = BN_MASK2;
320     } else {
321       // n0 < d0
322       bn_div_rem_words(&q, &rm, n0, n1, d0);
323 
324 #ifdef BN_ULLONG
325       BN_ULLONG t2 = (BN_ULLONG)d1 * q;
326       for (;;) {
327         if (t2 <= ((((BN_ULLONG)rm) << BN_BITS2) | wnump[-2])) {
328           break;
329         }
330         q--;
331         rm += d0;
332         if (rm < d0) {
333           break;  // don't let rm overflow
334         }
335         t2 -= d1;
336       }
337 #else  // !BN_ULLONG
338       BN_ULONG t2l, t2h;
339       BN_UMULT_LOHI(t2l, t2h, d1, q);
340       for (;;) {
341         if (t2h < rm ||
342             (t2h == rm && t2l <= wnump[-2])) {
343           break;
344         }
345         q--;
346         rm += d0;
347         if (rm < d0) {
348           break;  // don't let rm overflow
349         }
350         if (t2l < d1) {
351           t2h--;
352         }
353         t2l -= d1;
354       }
355 #endif  // !BN_ULLONG
356     }
357 
358     l0 = bn_mul_words(tmp->d, sdiv->d, div_n, q);
359     tmp->d[div_n] = l0;
360     wnum.d--;
361     // ingore top values of the bignums just sub the two
362     // BN_ULONG arrays with bn_sub_words
363     if (bn_sub_words(wnum.d, wnum.d, tmp->d, div_n + 1)) {
364       // Note: As we have considered only the leading
365       // two BN_ULONGs in the calculation of q, sdiv * q
366       // might be greater than wnum (but then (q-1) * sdiv
367       // is less or equal than wnum)
368       q--;
369       if (bn_add_words(wnum.d, wnum.d, sdiv->d, div_n)) {
370         // we can't have an overflow here (assuming
371         // that q != 0, but if q == 0 then tmp is
372         // zero anyway)
373         (*wnump)++;
374       }
375     }
376     // store part of the result
377     *resp = q;
378   }
379 
380   bn_set_minimal_width(snum);
381 
382   if (rem != NULL) {
383     // Keep a copy of the neg flag in numerator because if |rem| == |numerator|
384     // |BN_rshift| will overwrite it.
385     int neg = numerator->neg;
386     if (!BN_rshift(rem, snum, norm_shift)) {
387       goto err;
388     }
389     if (!BN_is_zero(rem)) {
390       rem->neg = neg;
391     }
392   }
393 
394   bn_set_minimal_width(res);
395   BN_CTX_end(ctx);
396   return 1;
397 
398 err:
399   BN_CTX_end(ctx);
400   return 0;
401 }
402 
BN_nnmod(BIGNUM * r,const BIGNUM * m,const BIGNUM * d,BN_CTX * ctx)403 int BN_nnmod(BIGNUM *r, const BIGNUM *m, const BIGNUM *d, BN_CTX *ctx) {
404   if (!(BN_mod(r, m, d, ctx))) {
405     return 0;
406   }
407   if (!r->neg) {
408     return 1;
409   }
410 
411   // now -|d| < r < 0, so we have to set r := r + |d|.
412   return (d->neg ? BN_sub : BN_add)(r, r, d);
413 }
414 
bn_reduce_once(BN_ULONG * r,const BN_ULONG * a,BN_ULONG carry,const BN_ULONG * m,size_t num)415 BN_ULONG bn_reduce_once(BN_ULONG *r, const BN_ULONG *a, BN_ULONG carry,
416                         const BN_ULONG *m, size_t num) {
417   assert(r != a);
418   // |r| = |a| - |m|. |bn_sub_words| performs the bulk of the subtraction, and
419   // then we apply the borrow to |carry|.
420   carry -= bn_sub_words(r, a, m, num);
421   // We know 0 <= |a| < 2*|m|, so -|m| <= |r| < |m|.
422   //
423   // If 0 <= |r| < |m|, |r| fits in |num| words and |carry| is zero. We then
424   // wish to select |r| as the answer. Otherwise -m <= r < 0 and we wish to
425   // return |r| + |m|, or |a|. |carry| must then be -1 or all ones. In both
426   // cases, |carry| is a suitable input to |bn_select_words|.
427   //
428   // Although |carry| may be one if it was one on input and |bn_sub_words|
429   // returns zero, this would give |r| > |m|, violating our input assumptions.
430   assert(carry == 0 || carry == (BN_ULONG)-1);
431   bn_select_words(r, carry, a /* r < 0 */, r /* r >= 0 */, num);
432   return carry;
433 }
434 
bn_reduce_once_in_place(BN_ULONG * r,BN_ULONG carry,const BN_ULONG * m,BN_ULONG * tmp,size_t num)435 BN_ULONG bn_reduce_once_in_place(BN_ULONG *r, BN_ULONG carry, const BN_ULONG *m,
436                                  BN_ULONG *tmp, size_t num) {
437   // See |bn_reduce_once| for why this logic works.
438   carry -= bn_sub_words(tmp, r, m, num);
439   assert(carry == 0 || carry == (BN_ULONG)-1);
440   bn_select_words(r, carry, r /* tmp < 0 */, tmp /* tmp >= 0 */, num);
441   return carry;
442 }
443 
bn_mod_sub_words(BN_ULONG * r,const BN_ULONG * a,const BN_ULONG * b,const BN_ULONG * m,BN_ULONG * tmp,size_t num)444 void bn_mod_sub_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b,
445                       const BN_ULONG *m, BN_ULONG *tmp, size_t num) {
446   // r = a - b
447   BN_ULONG borrow = bn_sub_words(r, a, b, num);
448   // tmp = a - b + m
449   bn_add_words(tmp, r, m, num);
450   bn_select_words(r, 0 - borrow, tmp /* r < 0 */, r /* r >= 0 */, num);
451 }
452 
bn_mod_add_words(BN_ULONG * r,const BN_ULONG * a,const BN_ULONG * b,const BN_ULONG * m,BN_ULONG * tmp,size_t num)453 void bn_mod_add_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b,
454                       const BN_ULONG *m, BN_ULONG *tmp, size_t num) {
455   BN_ULONG carry = bn_add_words(r, a, b, num);
456   bn_reduce_once_in_place(r, carry, m, tmp, num);
457 }
458 
bn_div_consttime(BIGNUM * quotient,BIGNUM * remainder,const BIGNUM * numerator,const BIGNUM * divisor,BN_CTX * ctx)459 int bn_div_consttime(BIGNUM *quotient, BIGNUM *remainder,
460                      const BIGNUM *numerator, const BIGNUM *divisor,
461                      BN_CTX *ctx) {
462   if (BN_is_negative(numerator) || BN_is_negative(divisor)) {
463     OPENSSL_PUT_ERROR(BN, BN_R_NEGATIVE_NUMBER);
464     return 0;
465   }
466   if (BN_is_zero(divisor)) {
467     OPENSSL_PUT_ERROR(BN, BN_R_DIV_BY_ZERO);
468     return 0;
469   }
470 
471   // This function implements long division in binary. It is not very efficient,
472   // but it is simple, easy to make constant-time, and performant enough for RSA
473   // key generation.
474 
475   int ret = 0;
476   BN_CTX_start(ctx);
477   BIGNUM *q = quotient, *r = remainder;
478   if (quotient == NULL || quotient == numerator || quotient == divisor) {
479     q = BN_CTX_get(ctx);
480   }
481   if (remainder == NULL || remainder == numerator || remainder == divisor) {
482     r = BN_CTX_get(ctx);
483   }
484   BIGNUM *tmp = BN_CTX_get(ctx);
485   if (q == NULL || r == NULL || tmp == NULL ||
486       !bn_wexpand(q, numerator->width) ||
487       !bn_wexpand(r, divisor->width) ||
488       !bn_wexpand(tmp, divisor->width)) {
489     goto err;
490   }
491 
492   OPENSSL_memset(q->d, 0, numerator->width * sizeof(BN_ULONG));
493   q->width = numerator->width;
494   q->neg = 0;
495 
496   OPENSSL_memset(r->d, 0, divisor->width * sizeof(BN_ULONG));
497   r->width = divisor->width;
498   r->neg = 0;
499 
500   // Incorporate |numerator| into |r|, one bit at a time, reducing after each
501   // step. At the start of each loop iteration, |r| < |divisor|
502   for (int i = numerator->width - 1; i >= 0; i--) {
503     for (int bit = BN_BITS2 - 1; bit >= 0; bit--) {
504       // Incorporate the next bit of the numerator, by computing
505       // r = 2*r or 2*r + 1. Note the result fits in one more word. We store the
506       // extra word in |carry|.
507       BN_ULONG carry = bn_add_words(r->d, r->d, r->d, divisor->width);
508       r->d[0] |= (numerator->d[i] >> bit) & 1;
509       // |r| was previously fully-reduced, so we know:
510       //      2*0 <= r <= 2*(divisor-1) + 1
511       //        0 <= r <= 2*divisor - 1 < 2*divisor.
512       // Thus |r| satisfies the preconditions for |bn_reduce_once_in_place|.
513       BN_ULONG subtracted = bn_reduce_once_in_place(r->d, carry, divisor->d,
514                                                     tmp->d, divisor->width);
515       // The corresponding bit of the quotient is set iff we needed to subtract.
516       q->d[i] |= (~subtracted & 1) << bit;
517     }
518   }
519 
520   if ((quotient != NULL && !BN_copy(quotient, q)) ||
521       (remainder != NULL && !BN_copy(remainder, r))) {
522     goto err;
523   }
524 
525   ret = 1;
526 
527 err:
528   BN_CTX_end(ctx);
529   return ret;
530 }
531 
bn_scratch_space_from_ctx(size_t width,BN_CTX * ctx)532 static BIGNUM *bn_scratch_space_from_ctx(size_t width, BN_CTX *ctx) {
533   BIGNUM *ret = BN_CTX_get(ctx);
534   if (ret == NULL ||
535       !bn_wexpand(ret, width)) {
536     return NULL;
537   }
538   ret->neg = 0;
539   ret->width = width;
540   return ret;
541 }
542 
543 // bn_resized_from_ctx returns |bn| with width at least |width| or NULL on
544 // error. This is so it may be used with low-level "words" functions. If
545 // necessary, it allocates a new |BIGNUM| with a lifetime of the current scope
546 // in |ctx|, so the caller does not need to explicitly free it. |bn| must fit in
547 // |width| words.
bn_resized_from_ctx(const BIGNUM * bn,size_t width,BN_CTX * ctx)548 static const BIGNUM *bn_resized_from_ctx(const BIGNUM *bn, size_t width,
549                                          BN_CTX *ctx) {
550   if ((size_t)bn->width >= width) {
551     // Any excess words must be zero.
552     assert(bn_fits_in_words(bn, width));
553     return bn;
554   }
555   BIGNUM *ret = bn_scratch_space_from_ctx(width, ctx);
556   if (ret == NULL ||
557       !BN_copy(ret, bn) ||
558       !bn_resize_words(ret, width)) {
559     return NULL;
560   }
561   return ret;
562 }
563 
BN_mod_add(BIGNUM * r,const BIGNUM * a,const BIGNUM * b,const BIGNUM * m,BN_CTX * ctx)564 int BN_mod_add(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *m,
565                BN_CTX *ctx) {
566   if (!BN_add(r, a, b)) {
567     return 0;
568   }
569   return BN_nnmod(r, r, m, ctx);
570 }
571 
BN_mod_add_quick(BIGNUM * r,const BIGNUM * a,const BIGNUM * b,const BIGNUM * m)572 int BN_mod_add_quick(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
573                      const BIGNUM *m) {
574   BN_CTX *ctx = BN_CTX_new();
575   int ok = ctx != NULL &&
576            bn_mod_add_consttime(r, a, b, m, ctx);
577   BN_CTX_free(ctx);
578   return ok;
579 }
580 
bn_mod_add_consttime(BIGNUM * r,const BIGNUM * a,const BIGNUM * b,const BIGNUM * m,BN_CTX * ctx)581 int bn_mod_add_consttime(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
582                          const BIGNUM *m, BN_CTX *ctx) {
583   BN_CTX_start(ctx);
584   a = bn_resized_from_ctx(a, m->width, ctx);
585   b = bn_resized_from_ctx(b, m->width, ctx);
586   BIGNUM *tmp = bn_scratch_space_from_ctx(m->width, ctx);
587   int ok = a != NULL && b != NULL && tmp != NULL &&
588            bn_wexpand(r, m->width);
589   if (ok) {
590     bn_mod_add_words(r->d, a->d, b->d, m->d, tmp->d, m->width);
591     r->width = m->width;
592     r->neg = 0;
593   }
594   BN_CTX_end(ctx);
595   return ok;
596 }
597 
BN_mod_sub(BIGNUM * r,const BIGNUM * a,const BIGNUM * b,const BIGNUM * m,BN_CTX * ctx)598 int BN_mod_sub(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *m,
599                BN_CTX *ctx) {
600   if (!BN_sub(r, a, b)) {
601     return 0;
602   }
603   return BN_nnmod(r, r, m, ctx);
604 }
605 
bn_mod_sub_consttime(BIGNUM * r,const BIGNUM * a,const BIGNUM * b,const BIGNUM * m,BN_CTX * ctx)606 int bn_mod_sub_consttime(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
607                          const BIGNUM *m, BN_CTX *ctx) {
608   BN_CTX_start(ctx);
609   a = bn_resized_from_ctx(a, m->width, ctx);
610   b = bn_resized_from_ctx(b, m->width, ctx);
611   BIGNUM *tmp = bn_scratch_space_from_ctx(m->width, ctx);
612   int ok = a != NULL && b != NULL && tmp != NULL &&
613            bn_wexpand(r, m->width);
614   if (ok) {
615     bn_mod_sub_words(r->d, a->d, b->d, m->d, tmp->d, m->width);
616     r->width = m->width;
617     r->neg = 0;
618   }
619   BN_CTX_end(ctx);
620   return ok;
621 }
622 
BN_mod_sub_quick(BIGNUM * r,const BIGNUM * a,const BIGNUM * b,const BIGNUM * m)623 int BN_mod_sub_quick(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
624                      const BIGNUM *m) {
625   BN_CTX *ctx = BN_CTX_new();
626   int ok = ctx != NULL &&
627            bn_mod_sub_consttime(r, a, b, m, ctx);
628   BN_CTX_free(ctx);
629   return ok;
630 }
631 
BN_mod_mul(BIGNUM * r,const BIGNUM * a,const BIGNUM * b,const BIGNUM * m,BN_CTX * ctx)632 int BN_mod_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *m,
633                BN_CTX *ctx) {
634   BIGNUM *t;
635   int ret = 0;
636 
637   BN_CTX_start(ctx);
638   t = BN_CTX_get(ctx);
639   if (t == NULL) {
640     goto err;
641   }
642 
643   if (a == b) {
644     if (!BN_sqr(t, a, ctx)) {
645       goto err;
646     }
647   } else {
648     if (!BN_mul(t, a, b, ctx)) {
649       goto err;
650     }
651   }
652 
653   if (!BN_nnmod(r, t, m, ctx)) {
654     goto err;
655   }
656 
657   ret = 1;
658 
659 err:
660   BN_CTX_end(ctx);
661   return ret;
662 }
663 
BN_mod_sqr(BIGNUM * r,const BIGNUM * a,const BIGNUM * m,BN_CTX * ctx)664 int BN_mod_sqr(BIGNUM *r, const BIGNUM *a, const BIGNUM *m, BN_CTX *ctx) {
665   if (!BN_sqr(r, a, ctx)) {
666     return 0;
667   }
668 
669   // r->neg == 0,  thus we don't need BN_nnmod
670   return BN_mod(r, r, m, ctx);
671 }
672 
BN_mod_lshift(BIGNUM * r,const BIGNUM * a,int n,const BIGNUM * m,BN_CTX * ctx)673 int BN_mod_lshift(BIGNUM *r, const BIGNUM *a, int n, const BIGNUM *m,
674                   BN_CTX *ctx) {
675   BIGNUM *abs_m = NULL;
676   int ret;
677 
678   if (!BN_nnmod(r, a, m, ctx)) {
679     return 0;
680   }
681 
682   if (m->neg) {
683     abs_m = BN_dup(m);
684     if (abs_m == NULL) {
685       return 0;
686     }
687     abs_m->neg = 0;
688   }
689 
690   ret = bn_mod_lshift_consttime(r, r, n, (abs_m ? abs_m : m), ctx);
691 
692   BN_free(abs_m);
693   return ret;
694 }
695 
bn_mod_lshift_consttime(BIGNUM * r,const BIGNUM * a,int n,const BIGNUM * m,BN_CTX * ctx)696 int bn_mod_lshift_consttime(BIGNUM *r, const BIGNUM *a, int n, const BIGNUM *m,
697                             BN_CTX *ctx) {
698   if (!BN_copy(r, a)) {
699     return 0;
700   }
701   for (int i = 0; i < n; i++) {
702     if (!bn_mod_lshift1_consttime(r, r, m, ctx)) {
703       return 0;
704     }
705   }
706   return 1;
707 }
708 
BN_mod_lshift_quick(BIGNUM * r,const BIGNUM * a,int n,const BIGNUM * m)709 int BN_mod_lshift_quick(BIGNUM *r, const BIGNUM *a, int n, const BIGNUM *m) {
710   BN_CTX *ctx = BN_CTX_new();
711   int ok = ctx != NULL &&
712            bn_mod_lshift_consttime(r, a, n, m, ctx);
713   BN_CTX_free(ctx);
714   return ok;
715 }
716 
BN_mod_lshift1(BIGNUM * r,const BIGNUM * a,const BIGNUM * m,BN_CTX * ctx)717 int BN_mod_lshift1(BIGNUM *r, const BIGNUM *a, const BIGNUM *m, BN_CTX *ctx) {
718   if (!BN_lshift1(r, a)) {
719     return 0;
720   }
721 
722   return BN_nnmod(r, r, m, ctx);
723 }
724 
bn_mod_lshift1_consttime(BIGNUM * r,const BIGNUM * a,const BIGNUM * m,BN_CTX * ctx)725 int bn_mod_lshift1_consttime(BIGNUM *r, const BIGNUM *a, const BIGNUM *m,
726                              BN_CTX *ctx) {
727   return bn_mod_add_consttime(r, a, a, m, ctx);
728 }
729 
BN_mod_lshift1_quick(BIGNUM * r,const BIGNUM * a,const BIGNUM * m)730 int BN_mod_lshift1_quick(BIGNUM *r, const BIGNUM *a, const BIGNUM *m) {
731   BN_CTX *ctx = BN_CTX_new();
732   int ok = ctx != NULL &&
733            bn_mod_lshift1_consttime(r, a, m, ctx);
734   BN_CTX_free(ctx);
735   return ok;
736 }
737 
BN_div_word(BIGNUM * a,BN_ULONG w)738 BN_ULONG BN_div_word(BIGNUM *a, BN_ULONG w) {
739   BN_ULONG ret = 0;
740   int i, j;
741 
742   if (!w) {
743     // actually this an error (division by zero)
744     return (BN_ULONG) - 1;
745   }
746 
747   if (a->width == 0) {
748     return 0;
749   }
750 
751   // normalize input for |bn_div_rem_words|.
752   j = BN_BITS2 - BN_num_bits_word(w);
753   w <<= j;
754   if (!BN_lshift(a, a, j)) {
755     return (BN_ULONG) - 1;
756   }
757 
758   for (i = a->width - 1; i >= 0; i--) {
759     BN_ULONG l = a->d[i];
760     BN_ULONG d;
761     BN_ULONG unused_rem;
762     bn_div_rem_words(&d, &unused_rem, ret, l, w);
763     ret = l - (d * w);
764     a->d[i] = d;
765   }
766 
767   bn_set_minimal_width(a);
768   ret >>= j;
769   return ret;
770 }
771 
BN_mod_word(const BIGNUM * a,BN_ULONG w)772 BN_ULONG BN_mod_word(const BIGNUM *a, BN_ULONG w) {
773 #ifndef BN_CAN_DIVIDE_ULLONG
774   BN_ULONG ret = 0;
775 #else
776   BN_ULLONG ret = 0;
777 #endif
778   int i;
779 
780   if (w == 0) {
781     return (BN_ULONG) -1;
782   }
783 
784 #ifndef BN_CAN_DIVIDE_ULLONG
785   // If |w| is too long and we don't have |BN_ULLONG| division then we need to
786   // fall back to using |BN_div_word|.
787   if (w > ((BN_ULONG)1 << BN_BITS4)) {
788     BIGNUM *tmp = BN_dup(a);
789     if (tmp == NULL) {
790       return (BN_ULONG)-1;
791     }
792     ret = BN_div_word(tmp, w);
793     BN_free(tmp);
794     return ret;
795   }
796 #endif
797 
798   for (i = a->width - 1; i >= 0; i--) {
799 #ifndef BN_CAN_DIVIDE_ULLONG
800     ret = ((ret << BN_BITS4) | ((a->d[i] >> BN_BITS4) & BN_MASK2l)) % w;
801     ret = ((ret << BN_BITS4) | (a->d[i] & BN_MASK2l)) % w;
802 #else
803     ret = (BN_ULLONG)(((ret << (BN_ULLONG)BN_BITS2) | a->d[i]) % (BN_ULLONG)w);
804 #endif
805   }
806   return (BN_ULONG)ret;
807 }
808 
BN_mod_pow2(BIGNUM * r,const BIGNUM * a,size_t e)809 int BN_mod_pow2(BIGNUM *r, const BIGNUM *a, size_t e) {
810   if (e == 0 || a->width == 0) {
811     BN_zero(r);
812     return 1;
813   }
814 
815   size_t num_words = 1 + ((e - 1) / BN_BITS2);
816 
817   // If |a| definitely has less than |e| bits, just BN_copy.
818   if ((size_t) a->width < num_words) {
819     return BN_copy(r, a) != NULL;
820   }
821 
822   // Otherwise, first make sure we have enough space in |r|.
823   // Note that this will fail if num_words > INT_MAX.
824   if (!bn_wexpand(r, num_words)) {
825     return 0;
826   }
827 
828   // Copy the content of |a| into |r|.
829   OPENSSL_memcpy(r->d, a->d, num_words * sizeof(BN_ULONG));
830 
831   // If |e| isn't word-aligned, we have to mask off some of our bits.
832   size_t top_word_exponent = e % (sizeof(BN_ULONG) * 8);
833   if (top_word_exponent != 0) {
834     r->d[num_words - 1] &= (((BN_ULONG) 1) << top_word_exponent) - 1;
835   }
836 
837   // Fill in the remaining fields of |r|.
838   r->neg = a->neg;
839   r->width = (int) num_words;
840   bn_set_minimal_width(r);
841   return 1;
842 }
843 
BN_nnmod_pow2(BIGNUM * r,const BIGNUM * a,size_t e)844 int BN_nnmod_pow2(BIGNUM *r, const BIGNUM *a, size_t e) {
845   if (!BN_mod_pow2(r, a, e)) {
846     return 0;
847   }
848 
849   // If the returned value was non-negative, we're done.
850   if (BN_is_zero(r) || !r->neg) {
851     return 1;
852   }
853 
854   size_t num_words = 1 + (e - 1) / BN_BITS2;
855 
856   // Expand |r| to the size of our modulus.
857   if (!bn_wexpand(r, num_words)) {
858     return 0;
859   }
860 
861   // Clear the upper words of |r|.
862   OPENSSL_memset(&r->d[r->width], 0, (num_words - r->width) * BN_BYTES);
863 
864   // Set parameters of |r|.
865   r->neg = 0;
866   r->width = (int) num_words;
867 
868   // Now, invert every word. The idea here is that we want to compute 2^e-|x|,
869   // which is actually equivalent to the twos-complement representation of |x|
870   // in |e| bits, which is -x = ~x + 1.
871   for (int i = 0; i < r->width; i++) {
872     r->d[i] = ~r->d[i];
873   }
874 
875   // If our exponent doesn't span the top word, we have to mask the rest.
876   size_t top_word_exponent = e % BN_BITS2;
877   if (top_word_exponent != 0) {
878     r->d[r->width - 1] &= (((BN_ULONG) 1) << top_word_exponent) - 1;
879   }
880 
881   // Keep the minimal-width invariant for |BIGNUM|.
882   bn_set_minimal_width(r);
883 
884   // Finally, add one, for the reason described above.
885   return BN_add(r, r, BN_value_one());
886 }
887