xref: /freebsd/contrib/file/src/magic.c (revision c36e54bb328697af1e6113812caecbd3bac89fe0)
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 1.93 2015/04/15 23:47:58 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 #ifdef HAVE_LIMITS_H
48 #include <limits.h>	/* for PIPE_BUF */
49 #endif
50 
51 #if defined(HAVE_UTIMES)
52 # include <sys/time.h>
53 #elif defined(HAVE_UTIME)
54 # if defined(HAVE_SYS_UTIME_H)
55 #  include <sys/utime.h>
56 # elif defined(HAVE_UTIME_H)
57 #  include <utime.h>
58 # endif
59 #endif
60 
61 #ifdef HAVE_UNISTD_H
62 #include <unistd.h>	/* for read() */
63 #endif
64 
65 #ifndef PIPE_BUF
66 /* Get the PIPE_BUF from pathconf */
67 #ifdef _PC_PIPE_BUF
68 #define PIPE_BUF pathconf(".", _PC_PIPE_BUF)
69 #else
70 #define PIPE_BUF 512
71 #endif
72 #endif
73 
74 private void close_and_restore(const struct magic_set *, const char *, int,
75     const struct stat *);
76 private int unreadable_info(struct magic_set *, mode_t, const char *);
77 private const char* get_default_magic(void);
78 #ifndef COMPILE_ONLY
79 private const char *file_or_fd(struct magic_set *, const char *, int);
80 #endif
81 
82 #ifndef	STDIN_FILENO
83 #define	STDIN_FILENO	0
84 #endif
85 
86 #ifdef WIN32
87 /* HINSTANCE of this shared library. Needed for get_default_magic() */
88 static HINSTANCE _w32_dll_instance = NULL;
89 
90 static void
91 _w32_append_path(char **hmagicpath, const char *fmt, ...)
92 {
93 	char *tmppath;
94         char *newpath;
95 	va_list ap;
96 
97 	va_start(ap, fmt);
98 	if (vasprintf(&tmppath, fmt, ap) < 0) {
99 		va_end(ap);
100 		return;
101 	}
102 	va_end(ap);
103 
104 	if (access(tmppath, R_OK) == -1)
105 		goto out;
106 
107 	if (*hmagicpath == NULL) {
108 		*hmagicpath = tmppath;
109 		return;
110 	}
111 
112 	if (asprintf(&newpath, "%s%c%s", *hmagicpath, PATHSEP, tmppath) < 0)
113 		goto out;
114 
115 	free(*hmagicpath);
116 	free(tmppath);
117 	*hmagicpath = newpath;
118 	return;
119 out:
120 	free(tmppath);
121 }
122 
123 static void
124 _w32_get_magic_relative_to(char **hmagicpath, HINSTANCE module)
125 {
126 	static const char *trypaths[] = {
127 		"%s/share/misc/magic.mgc",
128 		"%s/magic.mgc",
129 	};
130 	LPSTR dllpath;
131 	size_t sp;
132 
133 	dllpath = calloc(MAX_PATH + 1, sizeof(*dllpath));
134 
135 	if (!GetModuleFileNameA(module, dllpath, MAX_PATH))
136 		goto out;
137 
138 	PathRemoveFileSpecA(dllpath);
139 
140 	sp = strlen(dllpath);
141 	if (sp > 3 && stricmp(&dllpath[sp - 3], "bin") == 0) {
142 		_w32_append_path(hmagicpath,
143 		    "%s/../share/misc/magic.mgc", dllpath);
144 		goto out;
145 	}
146 
147 	for (sp = 0; sp < __arraycount(trypaths); sp++)
148 		_w32_append_path(hmagicpath, trypaths[sp], dllpath);
149 out:
150 	free(dllpath);
151 }
152 
153 /* Placate GCC by offering a sacrificial previous prototype */
154 BOOL WINAPI DllMain(HINSTANCE, DWORD, LPVOID);
155 
156 BOOL WINAPI
157 DllMain(HINSTANCE hinstDLL, DWORD fdwReason,
158     LPVOID lpvReserved __attribute__((__unused__)))
159 {
160 	if (fdwReason == DLL_PROCESS_ATTACH)
161 		_w32_dll_instance = hinstDLL;
162 	return TRUE;
163 }
164 #endif
165 
166 private const char *
167 get_default_magic(void)
168 {
169 	static const char hmagic[] = "/.magic/magic.mgc";
170 	static char *default_magic;
171 	char *home, *hmagicpath;
172 
173 #ifndef WIN32
174 	struct stat st;
175 
176 	if (default_magic) {
177 		free(default_magic);
178 		default_magic = NULL;
179 	}
180 	if ((home = getenv("HOME")) == NULL)
181 		return MAGIC;
182 
183 	if (asprintf(&hmagicpath, "%s/.magic.mgc", home) < 0)
184 		return MAGIC;
185 	if (stat(hmagicpath, &st) == -1) {
186 		free(hmagicpath);
187 		if (asprintf(&hmagicpath, "%s/.magic", home) < 0)
188 			return MAGIC;
189 		if (stat(hmagicpath, &st) == -1)
190 			goto out;
191 		if (S_ISDIR(st.st_mode)) {
192 			free(hmagicpath);
193 			if (asprintf(&hmagicpath, "%s/%s", home, hmagic) < 0)
194 				return MAGIC;
195 			if (access(hmagicpath, R_OK) == -1)
196 				goto out;
197 		}
198 	}
199 
200 	if (asprintf(&default_magic, "%s:%s", hmagicpath, MAGIC) < 0)
201 		goto out;
202 	free(hmagicpath);
203 	return default_magic;
204 out:
205 	default_magic = NULL;
206 	free(hmagicpath);
207 	return MAGIC;
208 #else
209 	hmagicpath = NULL;
210 
211 	if (default_magic) {
212 		free(default_magic);
213 		default_magic = NULL;
214 	}
215 
216 	/* First, try to get a magic file from user-application data */
217 	if ((home = getenv("LOCALAPPDATA")) != NULL)
218 		_w32_append_path(&hmagicpath, "%s%s", home, hmagic);
219 
220 	/* Second, try to get a magic file from the user profile data */
221 	if ((home = getenv("USERPROFILE")) != NULL)
222 		_w32_append_path(&hmagicpath,
223 		    "%s/Local Settings/Application Data%s", home, hmagic);
224 
225 	/* Third, try to get a magic file from Common Files */
226 	if ((home = getenv("COMMONPROGRAMFILES")) != NULL)
227 		_w32_append_path(&hmagicpath, "%s%s", home, hmagic);
228 
229 	/* Fourth, try to get magic file relative to exe location */
230         _w32_get_magic_relative_to(&hmagicpath, NULL);
231 
232 	/* Fifth, try to get magic file relative to dll location */
233         _w32_get_magic_relative_to(&hmagicpath, _w32_dll_instance);
234 
235 	/* Avoid MAGIC constant - it likely points to a file within MSys tree */
236 	default_magic = hmagicpath;
237 	return default_magic;
238 #endif
239 }
240 
241 public const char *
242 magic_getpath(const char *magicfile, int action)
243 {
244 	if (magicfile != NULL)
245 		return magicfile;
246 
247 	magicfile = getenv("MAGIC");
248 	if (magicfile != NULL)
249 		return magicfile;
250 
251 	return action == FILE_LOAD ? get_default_magic() : MAGIC;
252 }
253 
254 public struct magic_set *
255 magic_open(int flags)
256 {
257 	return file_ms_alloc(flags);
258 }
259 
260 private int
261 unreadable_info(struct magic_set *ms, mode_t md, const char *file)
262 {
263 	if (file) {
264 		/* We cannot open it, but we were able to stat it. */
265 		if (access(file, W_OK) == 0)
266 			if (file_printf(ms, "writable, ") == -1)
267 				return -1;
268 		if (access(file, X_OK) == 0)
269 			if (file_printf(ms, "executable, ") == -1)
270 				return -1;
271 	}
272 	if (S_ISREG(md))
273 		if (file_printf(ms, "regular file, ") == -1)
274 			return -1;
275 	if (file_printf(ms, "no read permission") == -1)
276 		return -1;
277 	return 0;
278 }
279 
280 public void
281 magic_close(struct magic_set *ms)
282 {
283 	if (ms == NULL)
284 		return;
285 	file_ms_free(ms);
286 }
287 
288 /*
289  * load a magic file
290  */
291 public int
292 magic_load(struct magic_set *ms, const char *magicfile)
293 {
294 	if (ms == NULL)
295 		return -1;
296 	return file_apprentice(ms, magicfile, FILE_LOAD);
297 }
298 
299 #ifndef COMPILE_ONLY
300 /*
301  * Install a set of compiled magic buffers.
302  */
303 public int
304 magic_load_buffers(struct magic_set *ms, void **bufs, size_t *sizes,
305     size_t nbufs)
306 {
307 	if (ms == NULL)
308 		return -1;
309 	return buffer_apprentice(ms, (struct magic **)bufs, sizes, nbufs);
310 }
311 #endif
312 
313 public int
314 magic_compile(struct magic_set *ms, const char *magicfile)
315 {
316 	if (ms == NULL)
317 		return -1;
318 	return file_apprentice(ms, magicfile, FILE_COMPILE);
319 }
320 
321 public int
322 magic_check(struct magic_set *ms, const char *magicfile)
323 {
324 	if (ms == NULL)
325 		return -1;
326 	return file_apprentice(ms, magicfile, FILE_CHECK);
327 }
328 
329 public int
330 magic_list(struct magic_set *ms, const char *magicfile)
331 {
332 	if (ms == NULL)
333 		return -1;
334 	return file_apprentice(ms, magicfile, FILE_LIST);
335 }
336 
337 private void
338 close_and_restore(const struct magic_set *ms, const char *name, int fd,
339     const struct stat *sb)
340 {
341 	if (fd == STDIN_FILENO || name == NULL)
342 		return;
343 	(void) close(fd);
344 
345 	if ((ms->flags & MAGIC_PRESERVE_ATIME) != 0) {
346 		/*
347 		 * Try to restore access, modification times if read it.
348 		 * This is really *bad* because it will modify the status
349 		 * time of the file... And of course this will affect
350 		 * backup programs
351 		 */
352 #ifdef HAVE_UTIMES
353 		struct timeval  utsbuf[2];
354 		(void)memset(utsbuf, 0, sizeof(utsbuf));
355 		utsbuf[0].tv_sec = sb->st_atime;
356 		utsbuf[1].tv_sec = sb->st_mtime;
357 
358 		(void) utimes(name, utsbuf); /* don't care if loses */
359 #elif defined(HAVE_UTIME_H) || defined(HAVE_SYS_UTIME_H)
360 		struct utimbuf  utbuf;
361 
362 		(void)memset(&utbuf, 0, sizeof(utbuf));
363 		utbuf.actime = sb->st_atime;
364 		utbuf.modtime = sb->st_mtime;
365 		(void) utime(name, &utbuf); /* don't care if loses */
366 #endif
367 	}
368 }
369 
370 #ifndef COMPILE_ONLY
371 
372 /*
373  * find type of descriptor
374  */
375 public const char *
376 magic_descriptor(struct magic_set *ms, int fd)
377 {
378 	if (ms == NULL)
379 		return NULL;
380 	return file_or_fd(ms, NULL, fd);
381 }
382 
383 /*
384  * find type of named file
385  */
386 public const char *
387 magic_file(struct magic_set *ms, const char *inname)
388 {
389 	if (ms == NULL)
390 		return NULL;
391 	return file_or_fd(ms, inname, STDIN_FILENO);
392 }
393 
394 private const char *
395 file_or_fd(struct magic_set *ms, const char *inname, int fd)
396 {
397 	int	rv = -1;
398 	unsigned char *buf;
399 	struct stat	sb;
400 	ssize_t nbytes = 0;	/* number of bytes read from a datafile */
401 	int	ispipe = 0;
402 	off_t	pos = (off_t)-1;
403 
404 	if (file_reset(ms) == -1)
405 		goto out;
406 
407 	/*
408 	 * one extra for terminating '\0', and
409 	 * some overlapping space for matches near EOF
410 	 */
411 #define SLOP (1 + sizeof(union VALUETYPE))
412 	if ((buf = CAST(unsigned char *, malloc(HOWMANY + SLOP))) == NULL)
413 		return NULL;
414 
415 	switch (file_fsmagic(ms, inname, &sb)) {
416 	case -1:		/* error */
417 		goto done;
418 	case 0:			/* nothing found */
419 		break;
420 	default:		/* matched it and printed type */
421 		rv = 0;
422 		goto done;
423 	}
424 
425 #ifdef WIN32
426 	/* Place stdin in binary mode, so EOF (Ctrl+Z) doesn't stop early. */
427 	if (fd == STDIN_FILENO)
428 		_setmode(STDIN_FILENO, O_BINARY);
429 #endif
430 
431 	if (inname == NULL) {
432 		if (fstat(fd, &sb) == 0 && S_ISFIFO(sb.st_mode))
433 			ispipe = 1;
434 		else
435 			pos = lseek(fd, (off_t)0, SEEK_CUR);
436 	} else {
437 		int flags = O_RDONLY|O_BINARY;
438 		int okstat = stat(inname, &sb) == 0;
439 
440 		if (okstat && S_ISFIFO(sb.st_mode)) {
441 #ifdef O_NONBLOCK
442 			flags |= O_NONBLOCK;
443 #endif
444 			ispipe = 1;
445 		}
446 
447 		errno = 0;
448 		if ((fd = open(inname, flags)) < 0) {
449 #ifdef WIN32
450 			/*
451 			 * Can't stat, can't open.  It may have been opened in
452 			 * fsmagic, so if the user doesn't have read permission,
453 			 * allow it to say so; otherwise an error was probably
454 			 * displayed in fsmagic.
455 			 */
456 			if (!okstat && errno == EACCES) {
457 				sb.st_mode = S_IFBLK;
458 				okstat = 1;
459 			}
460 #endif
461 			if (okstat &&
462 			    unreadable_info(ms, sb.st_mode, inname) == -1)
463 				goto done;
464 			rv = 0;
465 			goto done;
466 		}
467 #ifdef O_NONBLOCK
468 		if ((flags = fcntl(fd, F_GETFL)) != -1) {
469 			flags &= ~O_NONBLOCK;
470 			(void)fcntl(fd, F_SETFL, flags);
471 		}
472 #endif
473 	}
474 
475 	/*
476 	 * try looking at the first HOWMANY bytes
477 	 */
478 	if (ispipe) {
479 		ssize_t r = 0;
480 
481 		while ((r = sread(fd, (void *)&buf[nbytes],
482 		    (size_t)(HOWMANY - nbytes), 1)) > 0) {
483 			nbytes += r;
484 			if (r < PIPE_BUF) break;
485 		}
486 
487 		if (nbytes == 0) {
488 			/* We can not read it, but we were able to stat it. */
489 			if (unreadable_info(ms, sb.st_mode, inname) == -1)
490 				goto done;
491 			rv = 0;
492 			goto done;
493 		}
494 
495 	} else {
496 		/* Windows refuses to read from a big console buffer. */
497 		size_t howmany =
498 #if defined(WIN32) && HOWMANY > 8 * 1024
499 				_isatty(fd) ? 8 * 1024 :
500 #endif
501 				HOWMANY;
502 		if ((nbytes = read(fd, (char *)buf, howmany)) == -1) {
503 			if (inname == NULL && fd != STDIN_FILENO)
504 				file_error(ms, errno, "cannot read fd %d", fd);
505 			else
506 				file_error(ms, errno, "cannot read `%s'",
507 				    inname == NULL ? "/dev/stdin" : inname);
508 			goto done;
509 		}
510 	}
511 
512 	(void)memset(buf + nbytes, 0, SLOP); /* NUL terminate */
513 	if (file_buffer(ms, fd, inname, buf, (size_t)nbytes) == -1)
514 		goto done;
515 	rv = 0;
516 done:
517 	free(buf);
518 	if (pos != (off_t)-1)
519 		(void)lseek(fd, pos, SEEK_SET);
520 	close_and_restore(ms, inname, fd, &sb);
521 out:
522 	return rv == 0 ? file_getbuffer(ms) : NULL;
523 }
524 
525 
526 public const char *
527 magic_buffer(struct magic_set *ms, const void *buf, size_t nb)
528 {
529 	if (ms == NULL)
530 		return NULL;
531 	if (file_reset(ms) == -1)
532 		return NULL;
533 	/*
534 	 * The main work is done here!
535 	 * We have the file name and/or the data buffer to be identified.
536 	 */
537 	if (file_buffer(ms, -1, NULL, buf, nb) == -1) {
538 		return NULL;
539 	}
540 	return file_getbuffer(ms);
541 }
542 #endif
543 
544 public const char *
545 magic_error(struct magic_set *ms)
546 {
547 	if (ms == NULL)
548 		return "Magic database is not open";
549 	return (ms->event_flags & EVENT_HAD_ERR) ? ms->o.buf : NULL;
550 }
551 
552 public int
553 magic_errno(struct magic_set *ms)
554 {
555 	if (ms == NULL)
556 		return EINVAL;
557 	return (ms->event_flags & EVENT_HAD_ERR) ? ms->error : 0;
558 }
559 
560 public int
561 magic_setflags(struct magic_set *ms, int flags)
562 {
563 	if (ms == NULL)
564 		return -1;
565 #if !defined(HAVE_UTIME) && !defined(HAVE_UTIMES)
566 	if (flags & MAGIC_PRESERVE_ATIME)
567 		return -1;
568 #endif
569 	ms->flags = flags;
570 	return 0;
571 }
572 
573 public int
574 magic_version(void)
575 {
576 	return MAGIC_VERSION;
577 }
578 
579 public int
580 magic_setparam(struct magic_set *ms, int param, const void *val)
581 {
582 	switch (param) {
583 	case MAGIC_PARAM_INDIR_MAX:
584 		ms->indir_max = (uint16_t)*(const size_t *)val;
585 		return 0;
586 	case MAGIC_PARAM_NAME_MAX:
587 		ms->name_max = (uint16_t)*(const size_t *)val;
588 		return 0;
589 	case MAGIC_PARAM_ELF_PHNUM_MAX:
590 		ms->elf_phnum_max = (uint16_t)*(const size_t *)val;
591 		return 0;
592 	case MAGIC_PARAM_ELF_SHNUM_MAX:
593 		ms->elf_shnum_max = (uint16_t)*(const size_t *)val;
594 		return 0;
595 	case MAGIC_PARAM_ELF_NOTES_MAX:
596 		ms->elf_notes_max = (uint16_t)*(const size_t *)val;
597 		return 0;
598 	default:
599 		errno = EINVAL;
600 		return -1;
601 	}
602 }
603 
604 public int
605 magic_getparam(struct magic_set *ms, int param, void *val)
606 {
607 	switch (param) {
608 	case MAGIC_PARAM_INDIR_MAX:
609 		*(size_t *)val = ms->indir_max;
610 		return 0;
611 	case MAGIC_PARAM_NAME_MAX:
612 		*(size_t *)val = ms->name_max;
613 		return 0;
614 	case MAGIC_PARAM_ELF_PHNUM_MAX:
615 		*(size_t *)val = ms->elf_phnum_max;
616 		return 0;
617 	case MAGIC_PARAM_ELF_SHNUM_MAX:
618 		*(size_t *)val = ms->elf_shnum_max;
619 		return 0;
620 	case MAGIC_PARAM_ELF_NOTES_MAX:
621 		*(size_t *)val = ms->elf_notes_max;
622 		return 0;
623 	default:
624 		errno = EINVAL;
625 		return -1;
626 	}
627 }
628