1 /*
2 * Copyright (c) Christos Zoulas 2003.
3 * All Rights Reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 * notice immediately at the beginning of the file, without modification,
10 * this list of conditions, and the following disclaimer.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
14 *
15 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
19 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25 * SUCH DAMAGE.
26 */
27 #include "file.h"
28
29 #ifndef lint
30 FILE_RCSID("@(#)$File: funcs.c,v 1.92 2017/04/07 20:10:24 christos Exp $")
31 #endif /* lint */
32
33 #include "magic.h"
34 #include <stdarg.h>
35 #include <stdlib.h>
36 #include <string.h>
37 #include <ctype.h>
38 #if defined(HAVE_WCHAR_H)
39 #include <wchar.h>
40 #endif
41 #if defined(HAVE_WCTYPE_H)
42 #include <wctype.h>
43 #endif
44 #if defined(HAVE_LOCALE_H)
45 #include <locale.h>
46 #endif
47
48 #ifndef SIZE_MAX
49 #define SIZE_MAX ((size_t)~0)
50 #endif
51
52 #include "php.h"
53 #include "main/php_network.h"
54
55 #ifndef PREG_OFFSET_CAPTURE
56 # define PREG_OFFSET_CAPTURE (1<<8)
57 #endif
58
59 protected int
file_printf(struct magic_set * ms,const char * fmt,...)60 file_printf(struct magic_set *ms, const char *fmt, ...)
61 {
62 va_list ap;
63 int len;
64 char *buf = NULL, *newstr;
65
66 va_start(ap, fmt);
67 len = vspprintf(&buf, 0, fmt, ap);
68 va_end(ap);
69
70 if (ms->o.buf != NULL) {
71 len = spprintf(&newstr, 0, "%s%s", ms->o.buf, (buf ? buf : ""));
72 if (buf) {
73 efree(buf);
74 }
75 efree(ms->o.buf);
76 ms->o.buf = newstr;
77 } else {
78 ms->o.buf = buf;
79 }
80 return 0;
81 }
82
83 /*
84 * error - print best error message possible
85 */
86 /*VARARGS*/
87 private void
file_error_core(struct magic_set * ms,int error,const char * f,va_list va,size_t lineno)88 file_error_core(struct magic_set *ms, int error, const char *f, va_list va,
89 size_t lineno)
90 {
91 char *buf = NULL;
92
93 /* Only the first error is ok */
94 if (ms->event_flags & EVENT_HAD_ERR)
95 return;
96 if (lineno != 0) {
97 efree(ms->o.buf);
98 ms->o.buf = NULL;
99 file_printf(ms, "line %" SIZE_T_FORMAT "u:", lineno);
100 }
101
102 vspprintf(&buf, 0, f, va);
103 va_end(va);
104
105 if (error > 0) {
106 file_printf(ms, "%s (%s)", (*buf ? buf : ""), strerror(error));
107 } else if (*buf) {
108 file_printf(ms, "%s", buf);
109 }
110
111 if (buf) {
112 efree(buf);
113 }
114
115 ms->event_flags |= EVENT_HAD_ERR;
116 ms->error = error;
117 }
118
119 /*VARARGS*/
120 protected void
file_error(struct magic_set * ms,int error,const char * f,...)121 file_error(struct magic_set *ms, int error, const char *f, ...)
122 {
123 va_list va;
124 va_start(va, f);
125 file_error_core(ms, error, f, va, 0);
126 va_end(va);
127 }
128
129 /*
130 * Print an error with magic line number.
131 */
132 /*VARARGS*/
133 protected void
file_magerror(struct magic_set * ms,const char * f,...)134 file_magerror(struct magic_set *ms, const char *f, ...)
135 {
136 va_list va;
137 va_start(va, f);
138 file_error_core(ms, 0, f, va, ms->line);
139 va_end(va);
140 }
141
142 protected void
file_oomem(struct magic_set * ms,size_t len)143 file_oomem(struct magic_set *ms, size_t len)
144 {
145 file_error(ms, errno, "cannot allocate %" SIZE_T_FORMAT "u bytes",
146 len);
147 }
148
149 protected void
file_badseek(struct magic_set * ms)150 file_badseek(struct magic_set *ms)
151 {
152 file_error(ms, errno, "error seeking");
153 }
154
155 protected void
file_badread(struct magic_set * ms)156 file_badread(struct magic_set *ms)
157 {
158 file_error(ms, errno, "error reading");
159 }
160
161
162 static int
checkdone(struct magic_set * ms,int * rv)163 checkdone(struct magic_set *ms, int *rv)
164 {
165 if ((ms->flags & MAGIC_CONTINUE) == 0)
166 return 1;
167 if (file_printf(ms, "\n- ") == -1)
168 *rv = -1;
169 return 0;
170 }
171
172 /*ARGSUSED*/
173 protected int
file_buffer(struct magic_set * ms,php_stream * stream,const char * inname,const void * buf,size_t nb)174 file_buffer(struct magic_set *ms, php_stream *stream, const char *inname, const void *buf,
175 size_t nb)
176 {
177 int m = 0, rv = 0, looks_text = 0;
178 const unsigned char *ubuf = CAST(const unsigned char *, buf);
179 unichar *u8buf = NULL;
180 size_t ulen;
181 const char *code = NULL;
182 const char *code_mime = "binary";
183 const char *type = "application/octet-stream";
184 const char *def = "data";
185 const char *ftype = NULL;
186
187 if (nb == 0) {
188 def = "empty";
189 type = "application/x-empty";
190 goto simple;
191 } else if (nb == 1) {
192 def = "very short file (no magic)";
193 goto simple;
194 }
195
196 if ((ms->flags & MAGIC_NO_CHECK_ENCODING) == 0) {
197 looks_text = file_encoding(ms, ubuf, nb, &u8buf, &ulen,
198 &code, &code_mime, &ftype);
199 }
200
201 #ifdef __EMX__
202 if ((ms->flags & MAGIC_NO_CHECK_APPTYPE) == 0 && inname) {
203 m = file_os2_apptype(ms, inname, buf, nb);
204 if ((ms->flags & MAGIC_DEBUG) != 0)
205 (void)fprintf(stderr, "[try os2_apptype %d]\n", m);
206 switch (m) {
207 case -1:
208 return -1;
209 case 0:
210 break;
211 default:
212 return 1;
213 }
214 }
215 #endif
216
217 #if PHP_FILEINFO_UNCOMPRESS
218 if ((ms->flags & MAGIC_NO_CHECK_COMPRESS) == 0) {
219 m = file_zmagic(ms, stream, inname, ubuf, nb);
220 if ((ms->flags & MAGIC_DEBUG) != 0)
221 (void)fprintf(stderr, "[try zmagic %d]\n", m);
222 if (m) {
223 goto done_encoding;
224 }
225 }
226 #endif
227 /* Check if we have a tar file */
228 if ((ms->flags & MAGIC_NO_CHECK_TAR) == 0) {
229 m = file_is_tar(ms, ubuf, nb);
230 if ((ms->flags & MAGIC_DEBUG) != 0)
231 (void)fprintf(stderr, "[try tar %d]\n", m);
232 if (m) {
233 if (checkdone(ms, &rv))
234 goto done;
235 }
236 }
237
238 /* Check if we have a CDF file */
239 if ((ms->flags & MAGIC_NO_CHECK_CDF) == 0) {
240 php_socket_t fd;
241 if (stream && SUCCESS == php_stream_cast(stream, PHP_STREAM_AS_FD, (void **)&fd, 0)) {
242 m = file_trycdf(ms, fd, ubuf, nb);
243 if ((ms->flags & MAGIC_DEBUG) != 0)
244 (void)fprintf(stderr, "[try cdf %d]\n", m);
245 if (m) {
246 if (checkdone(ms, &rv))
247 goto done;
248 }
249 }
250 }
251
252 /* try soft magic tests */
253 if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
254 m = file_softmagic(ms, ubuf, nb, NULL, NULL, BINTEST,
255 looks_text);
256 if ((ms->flags & MAGIC_DEBUG) != 0)
257 (void)fprintf(stderr, "[try softmagic %d]\n", m);
258 if (m) {
259 #ifdef BUILTIN_ELF
260 if ((ms->flags & MAGIC_NO_CHECK_ELF) == 0 && m == 1 &&
261 nb > 5 && fd != -1) {
262 /*
263 * We matched something in the file, so this
264 * *might* be an ELF file, and the file is at
265 * least 5 bytes long, so if it's an ELF file
266 * it has at least one byte past the ELF magic
267 * number - try extracting information from the
268 * ELF headers that cannot easily * be
269 * extracted with rules in the magic file.
270 */
271 m = file_tryelf(ms, fd, ubuf, nb);
272 if ((ms->flags & MAGIC_DEBUG) != 0)
273 (void)fprintf(stderr, "[try elf %d]\n",
274 m);
275 }
276 #endif
277 if (checkdone(ms, &rv))
278 goto done;
279 }
280 }
281
282 /* try text properties */
283 if ((ms->flags & MAGIC_NO_CHECK_TEXT) == 0) {
284
285 m = file_ascmagic(ms, ubuf, nb, looks_text);
286 if ((ms->flags & MAGIC_DEBUG) != 0)
287 (void)fprintf(stderr, "[try ascmagic %d]\n", m);
288 if (m) {
289 if (checkdone(ms, &rv))
290 goto done;
291 }
292 }
293
294 simple:
295 /* give up */
296 m = 1;
297 if (ms->flags & MAGIC_MIME) {
298 if ((ms->flags & MAGIC_MIME_TYPE) &&
299 file_printf(ms, "%s", type) == -1)
300 rv = -1;
301 } else if (ms->flags & MAGIC_APPLE) {
302 if (file_printf(ms, "UNKNUNKN") == -1)
303 rv = -1;
304 } else if (ms->flags & MAGIC_EXTENSION) {
305 if (file_printf(ms, "???") == -1)
306 rv = -1;
307 } else {
308 if (file_printf(ms, "%s", def) == -1)
309 rv = -1;
310 }
311 done:
312 if ((ms->flags & MAGIC_MIME_ENCODING) != 0) {
313 if (ms->flags & MAGIC_MIME_TYPE)
314 if (file_printf(ms, "; charset=") == -1)
315 rv = -1;
316 if (file_printf(ms, "%s", code_mime) == -1)
317 rv = -1;
318 }
319 #if PHP_FILEINFO_UNCOMPRESS
320 done_encoding:
321 #endif
322 free(u8buf);
323 if (rv)
324 return rv;
325
326 return m;
327 }
328
329 protected int
file_reset(struct magic_set * ms)330 file_reset(struct magic_set *ms)
331 {
332 if (ms->mlist[0] == NULL) {
333 file_error(ms, 0, "no magic files loaded");
334 return -1;
335 }
336 if (ms->o.buf) {
337 efree(ms->o.buf);
338 ms->o.buf = NULL;
339 }
340 if (ms->o.pbuf) {
341 efree(ms->o.pbuf);
342 ms->o.pbuf = NULL;
343 }
344 ms->event_flags &= ~EVENT_HAD_ERR;
345 ms->error = -1;
346 return 0;
347 }
348
349 #define OCTALIFY(n, o) \
350 /*LINTED*/ \
351 (void)(*(n)++ = '\\', \
352 *(n)++ = (((uint32_t)*(o) >> 6) & 3) + '0', \
353 *(n)++ = (((uint32_t)*(o) >> 3) & 7) + '0', \
354 *(n)++ = (((uint32_t)*(o) >> 0) & 7) + '0', \
355 (o)++)
356
357 protected const char *
file_getbuffer(struct magic_set * ms)358 file_getbuffer(struct magic_set *ms)
359 {
360 char *pbuf, *op, *np;
361 size_t psize, len;
362
363 if (ms->event_flags & EVENT_HAD_ERR)
364 return NULL;
365
366 if (ms->flags & MAGIC_RAW)
367 return ms->o.buf;
368
369 if (ms->o.buf == NULL)
370 return NULL;
371
372 /* * 4 is for octal representation, + 1 is for NUL */
373 len = strlen(ms->o.buf);
374 if (len > (SIZE_MAX - 1) / 4) {
375 file_oomem(ms, len);
376 return NULL;
377 }
378 psize = len * 4 + 1;
379 if ((pbuf = CAST(char *, erealloc(ms->o.pbuf, psize))) == NULL) {
380 file_oomem(ms, psize);
381 return NULL;
382 }
383 ms->o.pbuf = pbuf;
384
385 #if defined(HAVE_WCHAR_H) && defined(HAVE_MBRTOWC) && defined(HAVE_WCWIDTH)
386 {
387 mbstate_t state;
388 wchar_t nextchar;
389 int mb_conv = 1;
390 size_t bytesconsumed;
391 char *eop;
392 (void)memset(&state, 0, sizeof(mbstate_t));
393
394 np = ms->o.pbuf;
395 op = ms->o.buf;
396 eop = op + len;
397
398 while (op < eop) {
399 bytesconsumed = mbrtowc(&nextchar, op,
400 (size_t)(eop - op), &state);
401 if (bytesconsumed == (size_t)(-1) ||
402 bytesconsumed == (size_t)(-2)) {
403 mb_conv = 0;
404 break;
405 }
406
407 if (iswprint(nextchar)) {
408 (void)memcpy(np, op, bytesconsumed);
409 op += bytesconsumed;
410 np += bytesconsumed;
411 } else {
412 while (bytesconsumed-- > 0)
413 OCTALIFY(np, op);
414 }
415 }
416 *np = '\0';
417
418 /* Parsing succeeded as a multi-byte sequence */
419 if (mb_conv != 0)
420 return ms->o.pbuf;
421 }
422 #endif
423
424 for (np = ms->o.pbuf, op = ms->o.buf; *op;) {
425 if (isprint((unsigned char)*op)) {
426 *np++ = *op++;
427 } else {
428 OCTALIFY(np, op);
429 }
430 }
431 *np = '\0';
432 return ms->o.pbuf;
433 }
434
435 protected int
file_check_mem(struct magic_set * ms,unsigned int level)436 file_check_mem(struct magic_set *ms, unsigned int level)
437 {
438 size_t len;
439
440 if (level >= ms->c.len) {
441 len = (ms->c.len = 20 + level) * sizeof(*ms->c.li);
442 ms->c.li = CAST(struct level_info *, (ms->c.li == NULL) ?
443 emalloc(len) :
444 erealloc(ms->c.li, len));
445 if (ms->c.li == NULL) {
446 file_oomem(ms, len);
447 return -1;
448 }
449 }
450 ms->c.li[level].got_match = 0;
451 #ifdef ENABLE_CONDITIONALS
452 ms->c.li[level].last_match = 0;
453 ms->c.li[level].last_cond = COND_NONE;
454 #endif /* ENABLE_CONDITIONALS */
455 return 0;
456 }
457
458 protected size_t
file_printedlen(const struct magic_set * ms)459 file_printedlen(const struct magic_set *ms)
460 {
461 return ms->o.buf == NULL ? 0 : strlen(ms->o.buf);
462 }
463
464 protected int
file_replace(struct magic_set * ms,const char * pat,const char * rep)465 file_replace(struct magic_set *ms, const char *pat, const char *rep)
466 {
467 zval patt;
468 int opts = 0;
469 pcre_cache_entry *pce;
470 zend_string *res;
471 zend_string *repl;
472 int rep_cnt = 0;
473
474 (void)setlocale(LC_CTYPE, "C");
475
476 opts |= PCRE_MULTILINE;
477 convert_libmagic_pattern(&patt, (char*)pat, strlen(pat), opts);
478 if ((pce = pcre_get_compiled_regex_cache(Z_STR(patt))) == NULL) {
479 zval_ptr_dtor(&patt);
480 rep_cnt = -1;
481 goto out;
482 }
483 zval_ptr_dtor(&patt);
484
485 repl = zend_string_init(rep, strlen(rep), 0);
486 res = php_pcre_replace_impl(pce, NULL, ms->o.buf, strlen(ms->o.buf), repl, -1, &rep_cnt);
487
488 zend_string_release(repl);
489 if (NULL == res) {
490 rep_cnt = -1;
491 goto out;
492 }
493
494 strncpy(ms->o.buf, ZSTR_VAL(res), ZSTR_LEN(res));
495 ms->o.buf[ZSTR_LEN(res)] = '\0';
496
497 zend_string_release(res);
498
499 out:
500 (void)setlocale(LC_CTYPE, "");
501 return rep_cnt;
502 }
503
504 protected file_pushbuf_t *
file_push_buffer(struct magic_set * ms)505 file_push_buffer(struct magic_set *ms)
506 {
507 file_pushbuf_t *pb;
508
509 if (ms->event_flags & EVENT_HAD_ERR)
510 return NULL;
511
512 if ((pb = (CAST(file_pushbuf_t *, emalloc(sizeof(*pb))))) == NULL)
513 return NULL;
514
515 pb->buf = ms->o.buf;
516 pb->offset = ms->offset;
517
518 ms->o.buf = NULL;
519 ms->offset = 0;
520
521 return pb;
522 }
523
524 protected char *
file_pop_buffer(struct magic_set * ms,file_pushbuf_t * pb)525 file_pop_buffer(struct magic_set *ms, file_pushbuf_t *pb)
526 {
527 char *rbuf;
528
529 if (ms->event_flags & EVENT_HAD_ERR) {
530 efree(pb->buf);
531 efree(pb);
532 return NULL;
533 }
534
535 rbuf = ms->o.buf;
536
537 ms->o.buf = pb->buf;
538 ms->offset = pb->offset;
539
540 efree(pb);
541 return rbuf;
542 }
543
544 /*
545 * convert string to ascii printable format.
546 */
547 protected char *
file_printable(char * buf,size_t bufsiz,const char * str)548 file_printable(char *buf, size_t bufsiz, const char *str)
549 {
550 char *ptr, *eptr;
551 const unsigned char *s = (const unsigned char *)str;
552
553 for (ptr = buf, eptr = ptr + bufsiz - 1; ptr < eptr && *s; s++) {
554 if (isprint(*s)) {
555 *ptr++ = *s;
556 continue;
557 }
558 if (ptr >= eptr - 3)
559 break;
560 *ptr++ = '\\';
561 *ptr++ = ((CAST(unsigned int, *s) >> 6) & 7) + '0';
562 *ptr++ = ((CAST(unsigned int, *s) >> 3) & 7) + '0';
563 *ptr++ = ((CAST(unsigned int, *s) >> 0) & 7) + '0';
564 }
565 *ptr = '\0';
566 return buf;
567 }
568
