Commit | Line | Data |
---|---|---|
b55540f6 | 1 | /* -*-c-*- |
b55540f6 | 2 | * |
3 | * Generate passphrases from word lists | |
4 | * | |
5 | * (c) 2000 Straylight/Edgeware | |
6 | */ | |
7 | ||
45c0fd36 | 8 | /*----- Licensing notice --------------------------------------------------* |
b55540f6 | 9 | * |
10 | * This file is part of Catacomb. | |
11 | * | |
12 | * Catacomb is free software; you can redistribute it and/or modify | |
13 | * it under the terms of the GNU Library General Public License as | |
14 | * published by the Free Software Foundation; either version 2 of the | |
15 | * License, or (at your option) any later version. | |
45c0fd36 | 16 | * |
b55540f6 | 17 | * Catacomb is distributed in the hope that it will be useful, |
18 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
19 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
20 | * GNU Library General Public License for more details. | |
45c0fd36 | 21 | * |
b55540f6 | 22 | * You should have received a copy of the GNU Library General Public |
23 | * License along with Catacomb; if not, write to the Free | |
24 | * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, | |
25 | * MA 02111-1307, USA. | |
26 | */ | |
27 | ||
b55540f6 | 28 | /*----- Header files ------------------------------------------------------*/ |
29 | ||
30 | #include "config.h" | |
31 | ||
32 | #include <ctype.h> | |
33 | #include <errno.h> | |
34 | #include <math.h> | |
35 | #include <stdio.h> | |
36 | #include <stdlib.h> | |
37 | #include <string.h> | |
38 | ||
39 | #include <mLib/alloc.h> | |
40 | #include <mLib/bits.h> | |
41 | #include <mLib/darray.h> | |
42 | #include <mLib/dstr.h> | |
43 | #include <mLib/mdwopt.h> | |
44 | #include <mLib/quis.h> | |
45 | #include <mLib/report.h> | |
46 | #include <mLib/sym.h> | |
47 | ||
48 | #include "grand.h" | |
49 | #include "noise.h" | |
50 | #include "rand.h" | |
51 | ||
52 | /*----- Global state ------------------------------------------------------*/ | |
53 | ||
54 | static unsigned min = 0, max = 256; /* Word length bounds */ | |
8a26239f | 55 | static unsigned minbits = 128, maxbits = UINT_MAX; /* Acceptable entropy */ |
b55540f6 | 56 | static unsigned count = 1; /* How many passphrases to make */ |
57 | ||
352783dd MW |
58 | static const char |
59 | all_wchars[] = "'abcdefghijklmnopqrstuvwxyz", | |
60 | *wchars = all_wchars; | |
b55540f6 | 61 | |
62 | typedef struct ppgen_ops { | |
63 | const char *name; /* Name of the generator */ | |
64 | void *(*init)(void); /* Initialize generator */ | |
65 | void (*scan)(FILE */*fp*/, void */*p*/); /* Scan an input word list */ | |
66 | void (*endscan)(void */*p*/); /* Scanning phase completed */ | |
67 | double (*gen)(dstr */*d*/, grand */*r*/, void */*p*/); | |
68 | /* Emit word and return entropy */ | |
69 | void (*done)(void */*p*/); /* Close down generator */ | |
70 | } ppgen_ops; | |
71 | ||
72 | /*----- Word list ---------------------------------------------------------*/ | |
73 | ||
74 | #ifndef STRING_V | |
75 | # define STRING_V | |
76 | DA_DECL(string_v, char *); | |
77 | #endif | |
78 | ||
79 | typedef struct wlist { | |
80 | string_v sv; | |
81 | sym_table tab; | |
82 | char *buf; | |
83 | double logp; | |
84 | } wlist; | |
85 | ||
86 | static void *wordlist_init(void) | |
87 | { | |
88 | wlist *w = xmalloc(sizeof(wlist)); | |
89 | sym_create(&w->tab); | |
90 | w->logp = 0; | |
91 | return (w); | |
92 | } | |
93 | ||
94 | static void wordlist_scan(FILE *fp, void *p) | |
95 | { | |
96 | wlist *w = p; | |
97 | dstr d = DSTR_INIT; | |
98 | unsigned f = 0; | |
99 | ||
100 | for (;;) { | |
101 | int ch = getc(fp); | |
102 | if (ch == EOF || isspace(ch)) { | |
103 | DPUTZ(&d); | |
104 | if (f && d.len >= min && d.len <= max) | |
105 | sym_find(&w->tab, d.buf, d.len + 1, sizeof(sym_base), 0); | |
106 | f = 0; | |
107 | DRESET(&d); | |
108 | if (ch == EOF) | |
109 | break; | |
110 | continue; | |
111 | } | |
112 | ch = tolower(ch); | |
113 | if (strchr(wchars, ch)) { | |
114 | DPUTC(&d, ch); | |
115 | f = 1; | |
116 | } | |
117 | } | |
118 | ||
119 | dstr_destroy(&d); | |
120 | } | |
121 | ||
122 | static void wordlist_endscan(void *p) | |
123 | { | |
124 | wlist *w = p; | |
125 | size_t buflen = 0; | |
126 | sym_iter i; | |
127 | sym_base *b; | |
128 | char *q; | |
129 | ||
130 | for (sym_mkiter(&i, &w->tab); (b = sym_next(&i)) != 0; ) | |
131 | buflen += b->len; | |
132 | w->buf = xmalloc(buflen); | |
133 | q = w->buf; | |
134 | DA_CREATE(&w->sv); | |
135 | for (sym_mkiter(&i, &w->tab); (b = sym_next(&i)) != 0; ) { | |
136 | memcpy(q, SYM_NAME(b), b->len); | |
137 | DA_PUSH(&w->sv, q); | |
138 | q += b->len; | |
139 | } | |
140 | sym_destroy(&w->tab); | |
141 | w->logp = log(DA_LEN(&w->sv))/log(2); | |
142 | } | |
143 | ||
144 | static double wordlist_gen(dstr *d, grand *r, void *p) | |
145 | { | |
146 | wlist *w = p; | |
147 | uint32 i = r->ops->range(r, DA_LEN(&w->sv)); | |
148 | DPUTS(d, DA(&w->sv)[i]); | |
149 | return (w->logp); | |
150 | } | |
151 | ||
152 | static void wordlist_done(void *p) | |
153 | { | |
154 | wlist *w = p; | |
155 | xfree(w->buf); | |
156 | DA_DESTROY(&w->sv); | |
157 | xfree(w); | |
158 | } | |
159 | ||
160 | static ppgen_ops wordlist_ops = { | |
161 | "wordlist", | |
162 | wordlist_init, wordlist_scan, wordlist_endscan, wordlist_gen, wordlist_done | |
163 | }; | |
164 | ||
165 | /*----- Markov word model -------------------------------------------------*/ | |
166 | ||
167 | enum { | |
168 | C_START = 27, | |
169 | C_END, | |
170 | VECSZ | |
171 | }; | |
172 | ||
173 | typedef struct node { | |
174 | uint32 count; | |
175 | uint32 p[VECSZ]; | |
176 | } node; | |
177 | ||
178 | static void *markov_init(void) | |
179 | { | |
180 | node (*model)[VECSZ][VECSZ][VECSZ] = xmalloc(sizeof(*model)); | |
181 | unsigned i, j, k, l; | |
45c0fd36 | 182 | |
b55540f6 | 183 | for (i = 0; i < VECSZ; i++) { |
184 | for (j = 0; j < VECSZ; j++) { | |
185 | for (k = 0; k < VECSZ; k++) { | |
186 | node *n = &(*model)[i][j][k]; | |
187 | n->count = 0; | |
188 | for (l = 0; l < VECSZ; l++) | |
189 | n->p[l] = 0; | |
190 | } | |
191 | } | |
192 | } | |
193 | ||
194 | return (model); | |
195 | } | |
196 | ||
197 | static void markov_scan(FILE *fp, void *p) | |
198 | { | |
199 | node (*model)[VECSZ][VECSZ][VECSZ] = p; | |
200 | unsigned i = C_START, j = C_START, k = C_START, l = C_END; | |
201 | ||
202 | for (;;) { | |
203 | int ch = getc(fp); | |
204 | const char *q; | |
205 | node *n = &(*model)[i][j][k]; | |
206 | ||
207 | if (ch == EOF || isspace(ch)) { | |
208 | if (l != C_END) { | |
209 | l = C_END; | |
210 | n->count++; | |
211 | n->p[l]++; | |
212 | i = j = k = C_START; | |
213 | } | |
214 | if (ch == EOF) | |
215 | break; | |
216 | continue; | |
217 | } | |
218 | ||
219 | if ((q = strchr(wchars, tolower(ch))) == 0) | |
220 | continue; | |
221 | l = q - wchars; | |
222 | n->count++; | |
223 | n->p[l]++; | |
224 | i = j; j = k; k = l; | |
225 | } | |
226 | } | |
227 | ||
228 | static double markov_gen(dstr *d, grand *r, void *p) | |
229 | { | |
230 | node (*model)[VECSZ][VECSZ][VECSZ] = p; | |
231 | unsigned i = C_START, j = C_START, k = C_START, l; | |
232 | double logp = 0; | |
233 | double log2 = log(2); | |
234 | ||
235 | for (;;) { | |
236 | node *n = &(*model)[i][j][k]; | |
237 | uint32 z = r->ops->range(r, n->count); | |
238 | for (l = 0; z >= n->p[l]; z -= n->p[l++]) | |
239 | ; | |
240 | logp -= log((double)n->p[l]/(double)n->count)/log2; | |
241 | if (l == C_END) | |
242 | break; | |
243 | DPUTC(d, wchars[l]); | |
244 | i = j; j = k; k = l; | |
245 | } | |
246 | ||
247 | return (logp); | |
248 | } | |
249 | ||
250 | static void markov_done(void *p) | |
251 | { | |
252 | node (*model)[VECSZ][VECSZ][VECSZ] = p; | |
253 | xfree(model); | |
254 | } | |
255 | ||
256 | static ppgen_ops markov_ops = { | |
257 | "markov", | |
258 | markov_init, markov_scan, 0, markov_gen, markov_done | |
259 | }; | |
260 | ||
261 | /*----- Main code ---------------------------------------------------------*/ | |
262 | ||
263 | static ppgen_ops *ppgentab[] = { | |
264 | &markov_ops, | |
265 | &wordlist_ops, | |
266 | 0 | |
267 | }; | |
268 | ||
269 | static void version(FILE *fp) | |
270 | { | |
271 | pquis(fp, "$, Catacomb version " VERSION "\n"); | |
272 | } | |
273 | ||
274 | static void usage(FILE *fp) | |
275 | { | |
276 | pquis(fp, "\ | |
8a26239f MW |
277 | Usage: $ [-p] [-b MIN[-MAX]] [-g GEN] [-n COUNT]\n\ |
278 | \t[-r [MIN-]MAX] WORDLIST...\n \ | |
b55540f6 | 279 | "); |
280 | } | |
281 | ||
282 | static void help(FILE *fp) | |
283 | { | |
284 | ppgen_ops **ops; | |
285 | version(fp); | |
286 | fputc('\n', fp); | |
287 | usage(fp); | |
288 | pquis(fp, "\n\ | |
289 | Generates random passphrases with the requested level of entropy. Options\n\ | |
290 | supported are:\n\ | |
291 | \n\ | |
292 | -h, --help Show this help text.\n\ | |
293 | -v, --version Show the program's version number.\n\ | |
294 | -u, --usage Show a terse usage summary.\n\ | |
8a26239f | 295 | -b, --bits=MIN[-MAX] Minimum and maximum bits of entropy.\n\ |
b55540f6 | 296 | -g, --generator=GEN Use passphrase generator GEN.\n\ |
297 | -n, --count=COUNT Generate COUNT passphrases.\n\ | |
298 | -p, --probability Show -log_2 of probability for each phrase.\n\ | |
299 | -r, --range=[MIN-]MAX Supply minimum and maximum word lengths.\n\ | |
300 | \n\ | |
301 | Generators currently available:"); | |
302 | for (ops = ppgentab; *ops; ops++) | |
303 | fprintf(fp, " %s", (*ops)->name); | |
304 | fputc('\n', fp); | |
305 | } | |
306 | ||
307 | int main(int argc, char *argv[]) | |
308 | { | |
309 | ppgen_ops *ops = ppgentab[0]; | |
310 | unsigned f = 0; | |
311 | void *ctx; | |
312 | dstr d = DSTR_INIT; | |
313 | dstr dd = DSTR_INIT; | |
314 | unsigned i; | |
315 | ||
16efd15b | 316 | #define f_bogus 1u |
317 | #define f_showp 2u | |
b55540f6 | 318 | |
319 | ego(argv[0]); | |
320 | for (;;) { | |
321 | static struct option opts[] = { | |
322 | { "help", 0, 0, 'h' }, | |
323 | { "version", 0, 0, 'v' }, | |
324 | { "usage", 0, 0, 'u' }, | |
352783dd | 325 | { "no-apostrophe", 0, 0, 'A' }, |
b55540f6 | 326 | { "bits", OPTF_ARGREQ, 0, 'b' }, |
327 | { "generator", OPTF_ARGREQ, 0, 'g' }, | |
328 | { "count", OPTF_ARGREQ, 0, 'n' }, | |
329 | { "probability", 0, 0, 'p' }, | |
330 | { "range", OPTF_ARGREQ, 0, 'r' }, | |
331 | { 0, 0, 0, 0 } | |
332 | }; | |
352783dd | 333 | int i = mdwopt(argc, argv, "hvu Ab:g:n:pr:", opts, 0, 0, 0); |
b55540f6 | 334 | |
335 | if (i < 0) | |
336 | break; | |
337 | switch (i) { | |
338 | case 'h': | |
339 | help(stdout); | |
340 | exit(0); | |
341 | case 'v': | |
342 | version(stdout); | |
343 | exit(0); | |
344 | case 'u': | |
345 | usage(stdout); | |
346 | exit(0); | |
352783dd MW |
347 | case 'A': |
348 | wchars = all_wchars + 1; | |
349 | break; | |
b55540f6 | 350 | case 'b': { |
351 | char *p; | |
8a26239f MW |
352 | minbits = strtoul(optarg, &p, 0); |
353 | if (*p == '-') | |
354 | maxbits = strtoul(p + 1, &p, 0); | |
355 | else | |
356 | maxbits = UINT_MAX; | |
357 | if (*p || minbits > maxbits) | |
358 | die(EXIT_FAILURE, "bad entropy range `%s'", optarg); | |
b55540f6 | 359 | } break; |
360 | case 'g': { | |
361 | ppgen_ops **p; | |
362 | size_t n = strlen(optarg); | |
363 | ops = 0; | |
364 | for (p = ppgentab; *p; p++) { | |
365 | if (strncmp(optarg, (*p)->name, n) == 0) { | |
366 | if (!(*p)->name[n]) { | |
367 | ops = *p; | |
368 | break; | |
369 | } else if (ops) | |
370 | die(EXIT_FAILURE, "ambiguous generator name `%s'", optarg); | |
371 | ops = *p; | |
372 | } | |
373 | } | |
374 | if (!ops) | |
375 | die(EXIT_FAILURE, "unknown generator name `%s'", optarg); | |
376 | } break; | |
377 | case 'n': { | |
378 | char *p; | |
379 | unsigned long n = strtoul(optarg, &p, 0); | |
380 | if (*p) | |
381 | die(EXIT_FAILURE, "bad integer `%s'", optarg); | |
382 | count = n; | |
383 | } break; | |
384 | case 'p': | |
385 | f |= f_showp; | |
386 | break; | |
387 | case 'r': { | |
388 | char *p; | |
389 | unsigned long n = min, nn = max; | |
390 | nn = strtoul(optarg, &p, 0); | |
391 | if (*p == '-') { | |
392 | n = nn; | |
393 | nn = strtoul(p + 1, &p, 0); | |
394 | } | |
9b87a466 | 395 | if (*p || min > max) |
b55540f6 | 396 | die(EXIT_FAILURE, "bad range string `%s'", optarg); |
397 | min = n; max = nn; | |
398 | } break; | |
399 | default: | |
400 | f |= f_bogus; | |
401 | break; | |
402 | } | |
403 | } | |
404 | ||
405 | argc -= optind; | |
406 | argv += optind; | |
407 | if ((f & f_bogus) || !argc) { | |
408 | usage(stderr); | |
409 | exit(EXIT_FAILURE); | |
410 | } | |
411 | ||
412 | rand_noisesrc(RAND_GLOBAL, &noise_source); | |
413 | rand_seed(RAND_GLOBAL, 160); | |
414 | ||
415 | ctx = ops->init(); | |
416 | while (*argv) { | |
417 | if (strcmp(*argv, "-") == 0) | |
8b2a1fd1 | 418 | ops->scan(stdin, ctx); |
b55540f6 | 419 | else { |
420 | FILE *fp = fopen(*argv, "r"); | |
421 | if (!fp) { | |
422 | die(EXIT_FAILURE, "error opening file `%s': %s", | |
423 | *argv, strerror(errno)); | |
424 | } | |
425 | ops->scan(fp, ctx); | |
426 | fclose(fp); | |
427 | } | |
428 | argv++; | |
429 | } | |
430 | if (ops->endscan) | |
431 | ops->endscan(ctx); | |
432 | ||
8a26239f | 433 | for (i = 0; !count || i < count; ) { |
b55540f6 | 434 | double logp = 0; |
435 | DRESET(&d); | |
8a26239f | 436 | while (logp < minbits) { |
b55540f6 | 437 | double pp; |
438 | DRESET(&dd); | |
439 | pp = ops->gen(&dd, &rand_global, ctx); | |
440 | if (!pp || dd.len < min || dd.len > max) | |
441 | continue; | |
442 | if (logp) | |
443 | DPUTC(&d, ' '); | |
444 | DPUTD(&d, &dd); | |
445 | logp += pp; | |
446 | } | |
8a26239f MW |
447 | if (logp >= (double)maxbits + 1) |
448 | continue; | |
b55540f6 | 449 | dstr_write(&d, stdout); |
450 | if (f & f_showp) | |
451 | printf(" [%g]", logp); | |
452 | fputc('\n', stdout); | |
8a26239f | 453 | i++; |
b55540f6 | 454 | } |
455 | ||
456 | ops->done(ctx); | |
457 | dstr_destroy(&d); | |
458 | dstr_destroy(&dd); | |
459 | return (0); | |
460 | } | |
461 | ||
462 | /*----- That's all, folks -------------------------------------------------*/ |