7765e926 |
1 | /* -*-c-*- |
2 | * |
8823192f |
3 | * $Id: exp.h,v 1.1.4.1 2004/03/20 00:13:31 mdw Exp $ |
7765e926 |
4 | * |
5 | * Generalized exponentiation |
6 | * |
7 | * (c) 2001 Straylight/Edgeware |
8 | */ |
9 | |
10 | /*----- Licensing notice --------------------------------------------------* |
11 | * |
12 | * This file is part of Catacomb. |
13 | * |
14 | * Catacomb is free software; you can redistribute it and/or modify |
15 | * it under the terms of the GNU Library General Public License as |
16 | * published by the Free Software Foundation; either version 2 of the |
17 | * License, or (at your option) any later version. |
18 | * |
19 | * Catacomb is distributed in the hope that it will be useful, |
20 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
21 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
22 | * GNU Library General Public License for more details. |
23 | * |
24 | * You should have received a copy of the GNU Library General Public |
25 | * License along with Catacomb; if not, write to the Free |
26 | * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, |
27 | * MA 02111-1307, USA. |
28 | */ |
29 | |
30 | /*----- Revision history --------------------------------------------------* |
31 | * |
32 | * $Log: exp.h,v $ |
8823192f |
33 | * Revision 1.1.4.1 2004/03/20 00:13:31 mdw |
34 | * Projective coordinates for prime curves |
35 | * |
7765e926 |
36 | * Revision 1.1 2001/06/16 13:00:59 mdw |
37 | * New generic exponentation code. Includes sliding-window simultaneous |
38 | * exponentiation. |
39 | * |
40 | */ |
41 | |
42 | #ifdef CATACOMB_EXP_H |
43 | # error "Multiple inclusion of <catacomb/exp.h>" |
44 | #endif |
45 | |
46 | #define CATACOMB_EXP_H |
47 | |
48 | #ifdef __cplusplus |
49 | extern "C" { |
50 | #endif |
51 | |
52 | /*----- Header files ------------------------------------------------------*/ |
53 | |
54 | #include <stddef.h> |
55 | |
56 | #include <mLib/alloc.h> |
57 | |
58 | #ifndef CATACOMB_MP_H |
59 | # include "mp.h" |
60 | #endif |
61 | |
62 | /*----- Data structures ---------------------------------------------------*/ |
63 | |
64 | typedef struct exp_simulscan { |
65 | mpw w; |
66 | size_t len; |
67 | const mpw *v; |
68 | } exp_simulscan; |
69 | |
70 | typedef struct exp_simul { |
71 | unsigned b; |
72 | size_t o, n; |
73 | exp_simulscan *s; |
74 | } exp_simul; |
75 | |
76 | /*----- Macros provided ---------------------------------------------------*/ |
77 | |
78 | /* --- Parameters --- */ |
79 | |
80 | #ifndef EXP_WINSZ /* Sliding window size */ |
81 | # define EXP_WINSZ 4 /* Predefine if you need to */ |
82 | #endif |
83 | |
84 | /* --- These are determined from the window size --- */ |
85 | |
86 | #define EXP_TABSZ (1 << EXP_WINSZ) |
87 | #define EXP_THRESH (((MPW_BITS / EXP_WINSZ) << 2) + 1) |
88 | |
89 | /* --- Required operations --- * |
90 | * |
91 | * The macros here are independent of the underlying group elements. You |
92 | * must provide the necessary group operations and other definitions. The |
93 | * group operation is assumed to be written multiplicatively. |
94 | * |
95 | * @EXP_TYPE@ The type of a group element, e.g., @mp *@. |
96 | * |
97 | * @EXP_COPY(d, x)@ Makes @d@ be a copy of @x@. |
98 | * |
99 | * @EXP_DROP(x)@ Discards the element @x@, reclaiming any |
100 | * memory it used. |
101 | * |
102 | * @EXP_MUL(a, x)@ Multiplies @a@ by @x@ (writing the result |
103 | * back to @a@). |
104 | * |
8823192f |
105 | * @EXP_FIX(x)@ Makes @x@ be a canonical representation of |
106 | * its value. All multiplications have the |
107 | * right argument canonical. |
108 | * |
7765e926 |
109 | * @EXP_SQR(a)@ Multiplies @a@ by itself. |
110 | * |
111 | * @EXP_SETMUL(d, x, y)@ Sets @d@ to be the product of @x@ and @y@. |
112 | * The value @d@ has not been initialized. |
113 | * |
114 | * @EXP_SETSQR(d, x)@ Sets @d@ to be the square of @x@. |
115 | * |
116 | * Only @EXP_TYPE@, @EXP_MUL@ and @EXP_SQR@ are required for simple |
117 | * exponentation. Sliding window and simultaneous exponentation require all |
118 | * of the operations. |
119 | */ |
120 | |
121 | #ifndef EXP_TYPE |
122 | # error "EXP_TYPE not defined for <catacomb/exp.h>" |
123 | #endif |
124 | |
125 | /* --- @EXP_SIMPLE@ --- * |
126 | * |
127 | * Arguments: @a@ = the result object, initially a multiplicative identity |
128 | * @g@ = the object to exponentiate |
129 | * @x@ = the exponent, as a multiprecision integer |
130 | * |
131 | * Use: Performs a simple left-to-right exponentiation. At the end |
132 | * of the code, the answer is left in @a@; @g@ and @x@ are |
133 | * unchanged. |
134 | */ |
135 | |
136 | #define EXP_SIMPLE(a, g, x) do { \ |
137 | mpscan sc; \ |
138 | unsigned sq = 0; \ |
139 | \ |
140 | /* --- Begin scanning --- */ \ |
141 | \ |
142 | mp_rscan(&sc, x); \ |
143 | if (!MP_RSTEP(&sc)) \ |
144 | goto exp_simple_exit; \ |
145 | while (!MP_RBIT(&sc)) \ |
146 | MP_RSTEP(&sc); \ |
147 | \ |
148 | /* --- Do the main body of the work --- */ \ |
149 | \ |
8823192f |
150 | EXP_FIX(g); \ |
7765e926 |
151 | for (;;) { \ |
152 | EXP_MUL(a, g); \ |
153 | sq = 0; \ |
154 | for (;;) { \ |
155 | if (!MP_RSTEP(&sc)) \ |
156 | goto exp_simple_done; \ |
157 | sq++; \ |
158 | if (MP_RBIT(&sc)) \ |
159 | break; \ |
160 | } \ |
161 | while (sq--) EXP_SQR(a); \ |
162 | } \ |
163 | \ |
164 | /* --- Do a final round of squaring --- */ \ |
165 | \ |
166 | exp_simple_done: \ |
167 | while (sq--) EXP_SQR(a); \ |
168 | exp_simple_exit:; \ |
169 | } while (0) |
170 | |
171 | /* --- @EXP_WINDOW@ --- * |
172 | * |
173 | * Arguments: @a@ = the result object, initially a multiplicative identity |
174 | * @g@ = the object to exponentiate |
175 | * @x@ = the exponent, as a multiprecision integer |
176 | * |
177 | * Use: Performs a sliding-window exponentiation. At the end of the |
178 | * code, the answer is left in @a@; @g@ and @x@ are unchanged. |
179 | */ |
180 | |
181 | #define EXP_WINDOW(a, g, x) do { \ |
182 | EXP_TYPE *v; \ |
183 | EXP_TYPE g2; \ |
184 | unsigned i, sq = 0; \ |
185 | mpscan sc; \ |
186 | \ |
187 | /* --- Get going --- */ \ |
188 | \ |
189 | mp_rscan(&sc, x); \ |
190 | if (!MP_RSTEP(&sc)) \ |
191 | goto exp_window_exit; \ |
192 | \ |
193 | /* --- Do the precomputation --- */ \ |
194 | \ |
8823192f |
195 | EXP_FIX(g); \ |
7765e926 |
196 | EXP_SETSQR(g2, g); \ |
8823192f |
197 | EXP_FIX(g2); \ |
7765e926 |
198 | v = xmalloc(EXP_TABSZ * sizeof(EXP_TYPE)); \ |
199 | EXP_COPY(v[0], g); \ |
8823192f |
200 | for (i = 1; i < EXP_TABSZ; i++) { \ |
7765e926 |
201 | EXP_SETMUL(v[i], v[i - 1], g2); \ |
8823192f |
202 | EXP_FIX(v[i]); \ |
203 | } \ |
7765e926 |
204 | EXP_DROP(g2); \ |
205 | \ |
206 | /* --- Skip top-end zero bits --- * \ |
207 | * \ |
208 | * If the initial step worked, there must be a set bit somewhere, so \ |
209 | * keep stepping until I find it. \ |
210 | */ \ |
211 | \ |
212 | while (!MP_RBIT(&sc)) \ |
213 | MP_RSTEP(&sc); \ |
214 | \ |
215 | /* --- Now for the main work --- */ \ |
216 | \ |
217 | for (;;) { \ |
218 | unsigned l = 1; \ |
219 | unsigned z = 0; \ |
220 | \ |
221 | /* --- The next bit is set, so read a window index --- * \ |
222 | * \ |
223 | * Reset @i@ to zero and increment @sq@. Then, until either I read \ |
224 | * @WINSZ@ bits or I run out of bits, scan in a bit: if it's clear, \ |
225 | * bump the @z@ counter; if it's set, push a set bit into @i@, \ |
226 | * shift it over by @z@ bits, bump @sq@ by @z + 1@ and clear @z@. \ |
227 | * By the end of this palaver, @i@ is an index to the precomputed \ |
228 | * value in @v@. \ |
229 | */ \ |
230 | \ |
231 | i = 0; \ |
232 | sq++; \ |
233 | while (l < EXP_WINSZ && MP_RSTEP(&sc)) { \ |
234 | l++; \ |
235 | if (!MP_RBIT(&sc)) \ |
236 | z++; \ |
237 | else { \ |
238 | i = ((i << 1) | 1) << z; \ |
239 | sq += z + 1; \ |
240 | z = 0; \ |
241 | } \ |
242 | } \ |
243 | \ |
244 | /* --- Do the squaring --- * \ |
245 | * \ |
246 | * Remember that @sq@ carries over from the zero-skipping stuff \ |
247 | * below. \ |
248 | */ \ |
249 | \ |
250 | while (sq--) EXP_SQR(a); \ |
251 | \ |
252 | /* --- Do the multiply --- */ \ |
253 | \ |
254 | EXP_MUL(a, v[i]); \ |
255 | \ |
256 | /* --- Now grind along through the rest of the bits --- */ \ |
257 | \ |
258 | sq = z; \ |
259 | for (;;) { \ |
260 | if (!MP_RSTEP(&sc)) \ |
261 | goto exp_window_done; \ |
262 | if (MP_RBIT(&sc)) \ |
263 | break; \ |
264 | sq++; \ |
265 | } \ |
266 | } \ |
267 | \ |
268 | /* --- Do a final round of squaring --- */ \ |
269 | \ |
270 | exp_window_done: \ |
271 | while (sq--) EXP_SQR(a); \ |
272 | for (i = 0; i < EXP_TABSZ; i++) \ |
273 | EXP_DROP(v[i]); \ |
274 | xfree(v); \ |
275 | exp_window_exit:; \ |
276 | } while (0) |
277 | |
278 | /* --- @EXP_SIMUL@ --- * |
279 | * |
280 | * Arguments: @a@ = the result object, initially a multiplicative identity |
281 | * @f@ = pointer to a vector of base/exp pairs |
282 | * @n@ = the number of base/exp pairs |
283 | * |
284 | * Use: Performs a simultaneous sliding-window exponentiation. The |
285 | * @f@ table is an array of structures containing members @base@ |
286 | * of type @EXP_TYPE@, and @exp@ of type @mp *@. |
287 | */ |
288 | |
289 | #define EXP_SIMUL(a, f, n) do { \ |
290 | size_t i, j, jj, k; \ |
291 | size_t vn = 1 << (EXP_WINSZ * n), m = (1 << n) - 1; \ |
292 | EXP_TYPE *v = xmalloc(vn * sizeof(EXP_TYPE)); \ |
293 | exp_simul e; \ |
294 | unsigned sq = 0; \ |
295 | \ |
296 | /* --- Fill in the precomputed table --- */ \ |
297 | \ |
298 | j = 1; \ |
299 | for (i = 0; i < n; i++) { \ |
300 | EXP_COPY(v[j], f[n - 1 - i].base); \ |
8823192f |
301 | EXP_FIX(v[j]); \ |
7765e926 |
302 | j <<= 1; \ |
303 | } \ |
304 | k = n * EXP_WINSZ; \ |
305 | jj = 1; \ |
306 | for (; i < k; i++) { \ |
307 | EXP_SETSQR(v[j], v[jj]); \ |
8823192f |
308 | EXP_FIX(v[j]); \ |
7765e926 |
309 | j <<= 1; jj <<= 1; \ |
310 | } \ |
311 | for (i = 1; i < vn; i <<= 1) { \ |
8823192f |
312 | for (j = 1; j < i; j++) { \ |
7765e926 |
313 | EXP_SETMUL(v[j + i], v[j], v[i]); \ |
8823192f |
314 | EXP_FIX(v[j + i]); \ |
315 | } \ |
7765e926 |
316 | } \ |
317 | \ |
318 | /* --- Set up the bitscanners --- * \ |
319 | * \ |
320 | * Got to use custom scanners, to keep them all in sync. \ |
321 | */ \ |
322 | \ |
323 | e.n = n; \ |
324 | e.b = 0; \ |
325 | e.s = xmalloc(n * sizeof(*e.s)); \ |
326 | e.o = 0; \ |
327 | for (i = 0; i < n; i++) { \ |
328 | MP_SHRINK(f[i].exp); \ |
329 | e.s[i].len = MP_LEN(f[i].exp); \ |
330 | e.s[i].v = f[i].exp->v; \ |
331 | if (e.s[i].len > e.o) \ |
332 | e.o = e.s[i].len; \ |
333 | } \ |
334 | \ |
335 | /* --- Skip as far as a nonzero column in the exponent matrix --- */ \ |
336 | \ |
337 | do { \ |
338 | if (!e.o && !e.b) \ |
339 | goto exp_simul_done; \ |
340 | i = exp_simulnext(&e, 0); \ |
341 | } while (!(i & m)); \ |
342 | \ |
343 | /* --- Now for the main work --- */ \ |
344 | \ |
345 | for (;;) { \ |
346 | unsigned l = 1; \ |
347 | unsigned z = 0; \ |
348 | \ |
349 | /* --- Just read a nonzero column, so read a window index --- * \ |
350 | * \ |
351 | * Clear high bits of @i@ and increment @sq@. Then, until either I \ |
352 | * read @WINSZ@ columns or I run out, scan in a column and append \ |
353 | * it to @i@. If it's zero, bump the @z@ counter; if it's nonzero, \ |
354 | * bump @sq@ by @z + 1@ and clear @z@. By the end of this palaver, \ |
355 | * @i@ is an index to the precomputed value in @v@, followed by \ |
356 | * @n * z@ zero bits. \ |
357 | */ \ |
358 | \ |
359 | sq++; \ |
360 | while (l < EXP_WINSZ && (e.o || e.b)) { \ |
361 | l++; \ |
362 | i = exp_simulnext(&e, i); \ |
363 | if (!(i & m)) \ |
364 | z++; \ |
365 | else { \ |
366 | sq += z + 1; \ |
367 | z = 0; \ |
368 | } \ |
369 | } \ |
370 | \ |
371 | /* --- Do the squaring --- * \ |
372 | * \ |
373 | * Remember that @sq@ carries over from the zero-skipping stuff \ |
374 | * below. \ |
375 | */ \ |
376 | \ |
377 | while (sq--) EXP_SQR(a); \ |
378 | \ |
379 | /* --- Do the multiply --- */ \ |
380 | \ |
381 | i >>= (z * n); \ |
382 | EXP_MUL(a, v[i]); \ |
383 | \ |
384 | /* --- Now grind along through the rest of the bits --- */ \ |
385 | \ |
386 | sq = z; \ |
387 | for (;;) { \ |
388 | if (!e.o && !e.b) \ |
389 | goto exp_simul_done; \ |
390 | if ((i = exp_simulnext(&e, 0)) != 0) \ |
391 | break; \ |
392 | sq++; \ |
393 | } \ |
394 | } \ |
395 | \ |
396 | /* --- Do a final round of squaring --- */ \ |
397 | \ |
398 | exp_simul_done: \ |
399 | while (sq--) EXP_SQR(a); \ |
8823192f |
400 | for (i = 1; i < vn; i++) \ |
7765e926 |
401 | EXP_DROP(v[i]); \ |
402 | xfree(v); \ |
403 | } while (0) |
404 | |
405 | /*----- Functions provided ------------------------------------------------*/ |
406 | |
407 | /* --- @exp_simulnext@ --- * |
408 | * |
409 | * Arguments: @exp_simul *e@ = pointer to state structure |
410 | * @size_t x@ = a current accumulator |
411 | * |
412 | * Returns: The next column of bits. |
413 | * |
414 | * Use: Scans the next column of bits for a simultaneous |
415 | * exponentiation. |
416 | */ |
417 | |
418 | extern size_t exp_simulnext(exp_simul */*e*/, size_t /*x*/); |
419 | |
420 | /*----- That's all, folks -------------------------------------------------*/ |
421 | |
422 | #ifdef __cplusplus |
423 | } |
424 | #endif |