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
28 #ifdef WIN32
29 #include <windows.h>
30 #include <shlwapi.h>
31 #endif
32
33 #include "file.h"
34
35 #ifndef lint
36 FILE_RCSID("@(#)$File: magic.c,v 1.123 2023/12/29 18:04:48 christos Exp $")
37 #endif /* lint */
38
39 #include "magic.h"
40
41 #include <stdlib.h>
42 #include <unistd.h>
43 #include <string.h>
44 #ifdef QUICK
45 #include <sys/mman.h>
46 #endif
47 #include <limits.h> /* for PIPE_BUF */
48
49 #if defined(HAVE_UTIMES)
50 # include <sys/time.h>
51 #elif defined(HAVE_UTIME)
52 # if defined(HAVE_SYS_UTIME_H)
53 # include <sys/utime.h>
54 # elif defined(HAVE_UTIME_H)
55 # include <utime.h>
56 # endif
57 #endif
58
59 #ifdef HAVE_UNISTD_H
60 #include <unistd.h> /* for read() */
61 #endif
62
63 #ifndef PIPE_BUF
64 /* Get the PIPE_BUF from pathconf */
65 #ifdef _PC_PIPE_BUF
66 #define PIPE_BUF pathconf(".", _PC_PIPE_BUF)
67 #else
68 #define PIPE_BUF 512
69 #endif
70 #endif
71
72 file_private void close_and_restore(const struct magic_set *, const char *, int,
73 const struct stat *);
74 file_private int unreadable_info(struct magic_set *, mode_t, const char *);
75 file_private const char* get_default_magic(void);
76 #ifndef COMPILE_ONLY
77 file_private const char *file_or_fd(struct magic_set *, const char *, int);
78 #endif
79
80 #ifndef STDIN_FILENO
81 #define STDIN_FILENO 0
82 #endif
83
84 #ifdef WIN32
85 /* HINSTANCE of this shared library. Needed for get_default_magic() */
86 static HINSTANCE _w32_dll_instance = NULL;
87
88 static void
_w32_append_path(char ** hmagicpath,const char * fmt,...)89 _w32_append_path(char **hmagicpath, const char *fmt, ...)
90 {
91 char *tmppath;
92 char *newpath;
93 va_list ap;
94
95 va_start(ap, fmt);
96 if (vasprintf(&tmppath, fmt, ap) < 0) {
97 va_end(ap);
98 return;
99 }
100 va_end(ap);
101
102 if (access(tmppath, R_OK) == -1)
103 goto out;
104
105 if (*hmagicpath == NULL) {
106 *hmagicpath = tmppath;
107 return;
108 }
109
110 if (asprintf(&newpath, "%s%c%s", *hmagicpath, PATHSEP, tmppath) < 0)
111 goto out;
112
113 free(*hmagicpath);
114 free(tmppath);
115 *hmagicpath = newpath;
116 return;
117 out:
118 free(tmppath);
119 }
120
121 static void
_w32_get_magic_relative_to(char ** hmagicpath,HINSTANCE module)122 _w32_get_magic_relative_to(char **hmagicpath, HINSTANCE module)
123 {
124 static const char *trypaths[] = {
125 "%s/share/misc/magic.mgc",
126 "%s/magic.mgc",
127 };
128 LPSTR dllpath;
129 size_t sp;
130
131 dllpath = calloc(MAX_PATH + 1, sizeof(*dllpath));
132
133 if (!GetModuleFileNameA(module, dllpath, MAX_PATH))
134 goto out;
135
136 PathRemoveFileSpecA(dllpath);
137
138 if (module) {
139 char exepath[MAX_PATH];
140 GetModuleFileNameA(NULL, exepath, MAX_PATH);
141 PathRemoveFileSpecA(exepath);
142 if (stricmp(exepath, dllpath) == 0)
143 goto out;
144 }
145
146 sp = strlen(dllpath);
147 if (sp > 3 && stricmp(&dllpath[sp - 3], "bin") == 0) {
148 _w32_append_path(hmagicpath,
149 "%s/../share/misc/magic.mgc", dllpath);
150 goto out;
151 }
152
153 for (sp = 0; sp < __arraycount(trypaths); sp++)
154 _w32_append_path(hmagicpath, trypaths[sp], dllpath);
155 out:
156 free(dllpath);
157 }
158
159 #ifndef BUILD_AS_WINDOWS_STATIC_LIBARAY
160 /* Placate GCC by offering a sacrificial previous prototype */
161 BOOL WINAPI DllMain(HINSTANCE, DWORD, LPVOID);
162
163 BOOL WINAPI
DllMain(HINSTANCE hinstDLL,DWORD fdwReason,LPVOID lpvReserved)164 DllMain(HINSTANCE hinstDLL, DWORD fdwReason,
165 LPVOID lpvReserved __attribute__((__unused__)))
166 {
167 if (fdwReason == DLL_PROCESS_ATTACH)
168 _w32_dll_instance = hinstDLL;
169 return 1;
170 }
171 #endif
172 #endif
173
174 file_private const char *
get_default_magic(void)175 get_default_magic(void)
176 {
177 static const char hmagic[] = "/.magic/magic.mgc";
178 static char *default_magic;
179 char *home, *hmagicpath;
180
181 #ifndef WIN32
182 struct stat st;
183
184 if (default_magic) {
185 free(default_magic);
186 default_magic = NULL;
187 }
188 if ((home = getenv("HOME")) == NULL)
189 return MAGIC;
190
191 if (asprintf(&hmagicpath, "%s/.magic.mgc", home) < 0)
192 return MAGIC;
193 if (stat(hmagicpath, &st) == -1) {
194 free(hmagicpath);
195 if (asprintf(&hmagicpath, "%s/.magic", home) < 0)
196 return MAGIC;
197 if (stat(hmagicpath, &st) == -1)
198 goto out;
199 if (S_ISDIR(st.st_mode)) {
200 free(hmagicpath);
201 if (asprintf(&hmagicpath, "%s/%s", home, hmagic) < 0)
202 return MAGIC;
203 if (access(hmagicpath, R_OK) == -1)
204 goto out;
205 }
206 }
207
208 if (asprintf(&default_magic, "%s:%s", hmagicpath, MAGIC) < 0)
209 goto out;
210 free(hmagicpath);
211 return default_magic;
212 out:
213 default_magic = NULL;
214 free(hmagicpath);
215 return MAGIC;
216 #else
217 hmagicpath = NULL;
218
219 if (default_magic) {
220 free(default_magic);
221 default_magic = NULL;
222 }
223
224 /* Before anything else, try to get a magic file from user HOME */
225 if ((home = getenv("HOME")) != NULL)
226 _w32_append_path(&hmagicpath, "%s%s", home, hmagic);
227
228 /* First, try to get a magic file from user-application data */
229 if ((home = getenv("LOCALAPPDATA")) != NULL)
230 _w32_append_path(&hmagicpath, "%s%s", home, hmagic);
231
232 /* Second, try to get a magic file from the user profile data */
233 if ((home = getenv("USERPROFILE")) != NULL)
234 _w32_append_path(&hmagicpath,
235 "%s/Local Settings/Application Data%s", home, hmagic);
236
237 /* Third, try to get a magic file from Common Files */
238 if ((home = getenv("COMMONPROGRAMFILES")) != NULL)
239 _w32_append_path(&hmagicpath, "%s%s", home, hmagic);
240
241 /* Fourth, try to get magic file relative to exe location */
242 _w32_get_magic_relative_to(&hmagicpath, NULL);
243
244 /* Fifth, try to get magic file relative to dll location */
245 _w32_get_magic_relative_to(&hmagicpath, _w32_dll_instance);
246
247 /* Avoid MAGIC constant - it likely points to a file within MSys tree */
248 default_magic = hmagicpath;
249 return default_magic;
250 #endif
251 }
252
253 file_public const char *
magic_getpath(const char * magicfile,int action)254 magic_getpath(const char *magicfile, int action)
255 {
256 if (magicfile != NULL)
257 return magicfile;
258
259 magicfile = getenv("MAGIC");
260 if (magicfile != NULL)
261 return magicfile;
262
263 return action == FILE_LOAD ? get_default_magic() : MAGIC;
264 }
265
266 file_public struct magic_set *
magic_open(int flags)267 magic_open(int flags)
268 {
269 return file_ms_alloc(flags);
270 }
271
272 file_private int
unreadable_info(struct magic_set * ms,mode_t md,const char * file)273 unreadable_info(struct magic_set *ms, mode_t md, const char *file)
274 {
275 if (file) {
276 /* We cannot open it, but we were able to stat it. */
277 if (access(file, W_OK) == 0)
278 if (file_printf(ms, "writable, ") == -1)
279 return -1;
280 #ifndef WIN32
281 if (access(file, X_OK) == 0)
282 if (file_printf(ms, "executable, ") == -1)
283 return -1;
284 #else
285 /* X_OK doesn't work well on MS-Windows */
286 {
287 const char *p = strrchr(file, '.');
288 if (p && (stricmp(p, ".exe")
289 || stricmp(p, ".dll")
290 || stricmp(p, ".bat")
291 || stricmp(p, ".cmd")))
292 if (file_printf(ms, "writable, ") == -1)
293 return -1;
294 }
295 #endif
296 }
297 if (S_ISREG(md))
298 if (file_printf(ms, "regular file, ") == -1)
299 return -1;
300 if (file_printf(ms, "no read permission") == -1)
301 return -1;
302 return 0;
303 }
304
305 file_public void
magic_close(struct magic_set * ms)306 magic_close(struct magic_set *ms)
307 {
308 if (ms == NULL)
309 return;
310 file_ms_free(ms);
311 }
312
313 /*
314 * load a magic file
315 */
316 file_public int
magic_load(struct magic_set * ms,const char * magicfile)317 magic_load(struct magic_set *ms, const char *magicfile)
318 {
319 if (ms == NULL)
320 return -1;
321 return file_apprentice(ms, magicfile, FILE_LOAD);
322 }
323
324 #ifndef COMPILE_ONLY
325 /*
326 * Install a set of compiled magic buffers.
327 */
328 file_public int
magic_load_buffers(struct magic_set * ms,void ** bufs,size_t * sizes,size_t nbufs)329 magic_load_buffers(struct magic_set *ms, void **bufs, size_t *sizes,
330 size_t nbufs)
331 {
332 if (ms == NULL)
333 return -1;
334 return buffer_apprentice(ms, RCAST(struct magic **, bufs),
335 sizes, nbufs);
336 }
337 #endif
338
339 file_public int
magic_compile(struct magic_set * ms,const char * magicfile)340 magic_compile(struct magic_set *ms, const char *magicfile)
341 {
342 if (ms == NULL)
343 return -1;
344 return file_apprentice(ms, magicfile, FILE_COMPILE);
345 }
346
347 file_public int
magic_check(struct magic_set * ms,const char * magicfile)348 magic_check(struct magic_set *ms, const char *magicfile)
349 {
350 if (ms == NULL)
351 return -1;
352 return file_apprentice(ms, magicfile, FILE_CHECK);
353 }
354
355 file_public int
magic_list(struct magic_set * ms,const char * magicfile)356 magic_list(struct magic_set *ms, const char *magicfile)
357 {
358 if (ms == NULL)
359 return -1;
360 return file_apprentice(ms, magicfile, FILE_LIST);
361 }
362
363 file_private void
close_and_restore(const struct magic_set * ms,const char * name,int fd,const struct stat * sb)364 close_and_restore(const struct magic_set *ms, const char *name, int fd,
365 const struct stat *sb)
366 {
367 if (fd == STDIN_FILENO || name == NULL)
368 return;
369 (void) close(fd);
370
371 if ((ms->flags & MAGIC_PRESERVE_ATIME) != 0) {
372 /*
373 * Try to restore access, modification times if read it.
374 * This is really *bad* because it will modify the status
375 * time of the file... And of course this will affect
376 * backup programs
377 */
378 #ifdef HAVE_UTIMES
379 struct timeval utsbuf[2];
380 (void)memset(utsbuf, 0, sizeof(utsbuf));
381 utsbuf[0].tv_sec = sb->st_atime;
382 utsbuf[1].tv_sec = sb->st_mtime;
383
384 (void) utimes(name, utsbuf); /* don't care if loses */
385 #elif defined(HAVE_UTIME_H) || defined(HAVE_SYS_UTIME_H)
386 struct utimbuf utbuf;
387
388 (void)memset(&utbuf, 0, sizeof(utbuf));
389 utbuf.actime = sb->st_atime;
390 utbuf.modtime = sb->st_mtime;
391 (void) utime(name, &utbuf); /* don't care if loses */
392 #endif
393 }
394 }
395
396 #ifndef COMPILE_ONLY
397
398 /*
399 * find type of descriptor
400 */
401 file_public const char *
magic_descriptor(struct magic_set * ms,int fd)402 magic_descriptor(struct magic_set *ms, int fd)
403 {
404 if (ms == NULL)
405 return NULL;
406 return file_or_fd(ms, NULL, fd);
407 }
408
409 /*
410 * find type of named file
411 */
412 file_public const char *
magic_file(struct magic_set * ms,const char * inname)413 magic_file(struct magic_set *ms, const char *inname)
414 {
415 if (ms == NULL)
416 return NULL;
417 return file_or_fd(ms, inname, STDIN_FILENO);
418 }
419
420 file_private const char *
file_or_fd(struct magic_set * ms,const char * inname,int fd)421 file_or_fd(struct magic_set *ms, const char *inname, int fd)
422 {
423 int rv = -1;
424 unsigned char *buf;
425 struct stat sb;
426 ssize_t nbytes = 0; /* number of bytes read from a datafile */
427 int ispipe = 0;
428 int okstat = 0;
429 off_t pos = CAST(off_t, -1);
430
431 if (file_reset(ms, 1) == -1)
432 goto out;
433
434 /*
435 * one extra for terminating '\0', and
436 * some overlapping space for matches near EOF
437 */
438 #define SLOP (1 + sizeof(union VALUETYPE))
439 if ((buf = CAST(unsigned char *, malloc(ms->bytes_max + SLOP))) == NULL)
440 return NULL;
441
442 switch (file_fsmagic(ms, inname, &sb)) {
443 case -1: /* error */
444 goto done;
445 case 0: /* nothing found */
446 break;
447 default: /* matched it and printed type */
448 rv = 0;
449 goto done;
450 }
451
452 #ifdef WIN32
453 /* Place stdin in binary mode, so EOF (Ctrl+Z) doesn't stop early. */
454 if (fd == STDIN_FILENO)
455 _setmode(STDIN_FILENO, O_BINARY);
456 #endif
457 if (inname != NULL) {
458 int flags = O_RDONLY|O_BINARY|O_NONBLOCK|O_CLOEXEC;
459 errno = 0;
460 if ((fd = open(inname, flags)) < 0) {
461 okstat = stat(inname, &sb) == 0;
462 #ifdef WIN32
463 /*
464 * Can't stat, can't open. It may have been opened in
465 * fsmagic, so if the user doesn't have read permission,
466 * allow it to say so; otherwise an error was probably
467 * displayed in fsmagic.
468 */
469 if (!okstat && errno == EACCES) {
470 sb.st_mode = S_IFBLK;
471 okstat = 1;
472 }
473 #endif
474 if (okstat &&
475 unreadable_info(ms, sb.st_mode, inname) == -1)
476 goto done;
477 rv = 0;
478 goto done;
479 }
480 #if O_CLOEXEC == 0 && defined(F_SETFD)
481 (void)fcntl(fd, F_SETFD, FD_CLOEXEC);
482 #endif
483 }
484
485 if (fd != -1) {
486 okstat = fstat(fd, &sb) == 0;
487 if (okstat && S_ISFIFO(sb.st_mode))
488 ispipe = 1;
489 if (inname == NULL)
490 pos = lseek(fd, CAST(off_t, 0), SEEK_CUR);
491 }
492
493 /*
494 * try looking at the first ms->bytes_max bytes
495 */
496 if (ispipe) {
497 if (fd != -1) {
498 ssize_t r = 0;
499
500 while ((r = sread(fd, RCAST(void *, &buf[nbytes]),
501 CAST(size_t, ms->bytes_max - nbytes), 1)) > 0) {
502 nbytes += r;
503 if (r < PIPE_BUF) break;
504 }
505 }
506
507 if (nbytes == 0 && inname) {
508 /* We can not read it, but we were able to stat it. */
509 if (unreadable_info(ms, sb.st_mode, inname) == -1)
510 goto done;
511 rv = 0;
512 goto done;
513 }
514
515 } else if (fd != -1) {
516 /* Windows refuses to read from a big console buffer. */
517 size_t howmany =
518 #ifdef WIN32
519 _isatty(fd) ? 8 * 1024 :
520 #endif
521 ms->bytes_max;
522 if ((nbytes = read(fd, RCAST(void *, buf), howmany)) == -1) {
523 if (inname == NULL && fd != STDIN_FILENO)
524 file_error(ms, errno, "cannot read fd %d", fd);
525 else
526 file_error(ms, errno, "cannot read `%s'",
527 inname == NULL ? "/dev/stdin" : inname);
528 goto done;
529 }
530 }
531
532 (void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */
533 if (file_buffer(ms, fd, okstat ? &sb : NULL, inname, buf, CAST(size_t, nbytes)) == -1)
534 goto done;
535 rv = 0;
536 done:
537 free(buf);
538 if (fd != -1) {
539 if (pos != CAST(off_t, -1))
540 (void)lseek(fd, pos, SEEK_SET);
541 close_and_restore(ms, inname, fd, &sb);
542 }
543 out:
544 return rv == 0 ? file_getbuffer(ms) : NULL;
545 }
546
547
548 file_public const char *
magic_buffer(struct magic_set * ms,const void * buf,size_t nb)549 magic_buffer(struct magic_set *ms, const void *buf, size_t nb)
550 {
551 if (ms == NULL)
552 return NULL;
553 if (file_reset(ms, 1) == -1)
554 return NULL;
555 /*
556 * The main work is done here!
557 * We have the file name and/or the data buffer to be identified.
558 */
559 if (file_buffer(ms, -1, NULL, NULL, buf, nb) == -1) {
560 return NULL;
561 }
562 return file_getbuffer(ms);
563 }
564 #endif
565
566 file_public const char *
magic_error(struct magic_set * ms)567 magic_error(struct magic_set *ms)
568 {
569 if (ms == NULL)
570 return "Magic database is not open";
571 return (ms->event_flags & EVENT_HAD_ERR) ? ms->o.buf : NULL;
572 }
573
574 file_public int
magic_errno(struct magic_set * ms)575 magic_errno(struct magic_set *ms)
576 {
577 if (ms == NULL)
578 return EINVAL;
579 return (ms->event_flags & EVENT_HAD_ERR) ? ms->error : 0;
580 }
581
582 file_public int
magic_getflags(struct magic_set * ms)583 magic_getflags(struct magic_set *ms)
584 {
585 if (ms == NULL)
586 return -1;
587
588 return ms->flags;
589 }
590
591 file_public int
magic_setflags(struct magic_set * ms,int flags)592 magic_setflags(struct magic_set *ms, int flags)
593 {
594 if (ms == NULL)
595 return -1;
596 #if !defined(HAVE_UTIME) && !defined(HAVE_UTIMES)
597 if (flags & MAGIC_PRESERVE_ATIME)
598 return -1;
599 #endif
600 ms->flags = flags;
601 return 0;
602 }
603
604 file_public int
magic_version(void)605 magic_version(void)
606 {
607 return MAGIC_VERSION;
608 }
609
610 file_public int
magic_setparam(struct magic_set * ms,int param,const void * val)611 magic_setparam(struct magic_set *ms, int param, const void *val)
612 {
613 if (ms == NULL)
614 return -1;
615 const size_t v = *CAST(const size_t *, val);
616 switch (param) {
617 case MAGIC_PARAM_INDIR_MAX:
618 ms->indir_max = CAST(uint16_t, v);
619 return 0;
620 case MAGIC_PARAM_NAME_MAX:
621 ms->name_max = CAST(uint16_t, v);
622 return 0;
623 case MAGIC_PARAM_ELF_PHNUM_MAX:
624 ms->elf_phnum_max = CAST(uint16_t, v);
625 return 0;
626 case MAGIC_PARAM_ELF_SHNUM_MAX:
627 ms->elf_shnum_max = CAST(uint16_t, v);
628 return 0;
629 case MAGIC_PARAM_ELF_SHSIZE_MAX:
630 ms->elf_shsize_max = v;
631 return 0;
632 case MAGIC_PARAM_ELF_NOTES_MAX:
633 ms->elf_notes_max = CAST(uint16_t, v);
634 return 0;
635 case MAGIC_PARAM_REGEX_MAX:
636 ms->regex_max = CAST(uint16_t, v);
637 return 0;
638 case MAGIC_PARAM_BYTES_MAX:
639 ms->bytes_max = v;
640 return 0;
641 case MAGIC_PARAM_ENCODING_MAX:
642 ms->encoding_max = v;
643 return 0;
644 case MAGIC_PARAM_MAGWARN_MAX:
645 ms->magwarn_max = v;
646 return 0;
647 default:
648 errno = EINVAL;
649 return -1;
650 }
651 }
652
653 file_public int
magic_getparam(struct magic_set * ms,int param,void * val)654 magic_getparam(struct magic_set *ms, int param, void *val)
655 {
656 if (ms == NULL)
657 return -1;
658 switch (param) {
659 case MAGIC_PARAM_INDIR_MAX:
660 *CAST(size_t *, val) = ms->indir_max;
661 return 0;
662 case MAGIC_PARAM_NAME_MAX:
663 *CAST(size_t *, val) = ms->name_max;
664 return 0;
665 case MAGIC_PARAM_ELF_PHNUM_MAX:
666 *CAST(size_t *, val) = ms->elf_phnum_max;
667 return 0;
668 case MAGIC_PARAM_ELF_SHNUM_MAX:
669 *CAST(size_t *, val) = ms->elf_shnum_max;
670 return 0;
671 case MAGIC_PARAM_ELF_SHSIZE_MAX:
672 *CAST(size_t *, val) = ms->elf_shsize_max;
673 return 0;
674 case MAGIC_PARAM_ELF_NOTES_MAX:
675 *CAST(size_t *, val) = ms->elf_notes_max;
676 return 0;
677 case MAGIC_PARAM_REGEX_MAX:
678 *CAST(size_t *, val) = ms->regex_max;
679 return 0;
680 case MAGIC_PARAM_BYTES_MAX:
681 *CAST(size_t *, val) = ms->bytes_max;
682 return 0;
683 case MAGIC_PARAM_ENCODING_MAX:
684 *CAST(size_t *, val) = ms->encoding_max;
685 return 0;
686 case MAGIC_PARAM_MAGWARN_MAX:
687 *CAST(size_t *, val) = ms->magwarn_max;
688 return 0;
689 default:
690 errno = EINVAL;
691 return -1;
692 }
693 }
694