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
36FILE_RCSID("@(#)$File: magic.c,v 1.112 2020/06/08 19:44:10 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
72private void close_and_restore(const struct magic_set *, const char *, int,
73    const struct stat *);
74private int unreadable_info(struct magic_set *, mode_t, const char *);
75private const char* get_default_magic(void);
76#ifndef COMPILE_ONLY
77private 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() */
86static HINSTANCE _w32_dll_instance = NULL;
87
88static void
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;
117out:
118	free(tmppath);
119}
120
121static void
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);
155out:
156	free(dllpath);
157}
158
159/* Placate GCC by offering a sacrificial previous prototype */
160BOOL WINAPI DllMain(HINSTANCE, DWORD, LPVOID);
161
162BOOL WINAPI
163DllMain(HINSTANCE hinstDLL, DWORD fdwReason,
164    LPVOID lpvReserved __attribute__((__unused__)))
165{
166	if (fdwReason == DLL_PROCESS_ATTACH)
167		_w32_dll_instance = hinstDLL;
168	return 1;
169}
170#endif
171
172private const char *
173get_default_magic(void)
174{
175	static const char hmagic[] = "/.magic/magic.mgc";
176	static char *default_magic;
177	char *home, *hmagicpath;
178
179#ifndef WIN32
180	struct stat st;
181
182	if (default_magic) {
183		free(default_magic);
184		default_magic = NULL;
185	}
186	if ((home = getenv("HOME")) == NULL)
187		return MAGIC;
188
189	if (asprintf(&hmagicpath, "%s/.magic.mgc", home) < 0)
190		return MAGIC;
191	if (stat(hmagicpath, &st) == -1) {
192		free(hmagicpath);
193		if (asprintf(&hmagicpath, "%s/.magic", home) < 0)
194			return MAGIC;
195		if (stat(hmagicpath, &st) == -1)
196			goto out;
197		if (S_ISDIR(st.st_mode)) {
198			free(hmagicpath);
199			if (asprintf(&hmagicpath, "%s/%s", home, hmagic) < 0)
200				return MAGIC;
201			if (access(hmagicpath, R_OK) == -1)
202				goto out;
203		}
204	}
205
206	if (asprintf(&default_magic, "%s:%s", hmagicpath, MAGIC) < 0)
207		goto out;
208	free(hmagicpath);
209	return default_magic;
210out:
211	default_magic = NULL;
212	free(hmagicpath);
213	return MAGIC;
214#else
215	hmagicpath = NULL;
216
217	if (default_magic) {
218		free(default_magic);
219		default_magic = NULL;
220	}
221
222	/* First, try to get a magic file from user-application data */
223	if ((home = getenv("LOCALAPPDATA")) != NULL)
224		_w32_append_path(&hmagicpath, "%s%s", home, hmagic);
225
226	/* Second, try to get a magic file from the user profile data */
227	if ((home = getenv("USERPROFILE")) != NULL)
228		_w32_append_path(&hmagicpath,
229		    "%s/Local Settings/Application Data%s", home, hmagic);
230
231	/* Third, try to get a magic file from Common Files */
232	if ((home = getenv("COMMONPROGRAMFILES")) != NULL)
233		_w32_append_path(&hmagicpath, "%s%s", home, hmagic);
234
235	/* Fourth, try to get magic file relative to exe location */
236        _w32_get_magic_relative_to(&hmagicpath, NULL);
237
238	/* Fifth, try to get magic file relative to dll location */
239        _w32_get_magic_relative_to(&hmagicpath, _w32_dll_instance);
240
241	/* Avoid MAGIC constant - it likely points to a file within MSys tree */
242	default_magic = hmagicpath;
243	return default_magic;
244#endif
245}
246
247public const char *
248magic_getpath(const char *magicfile, int action)
249{
250	if (magicfile != NULL)
251		return magicfile;
252
253	magicfile = getenv("MAGIC");
254	if (magicfile != NULL)
255		return magicfile;
256
257	return action == FILE_LOAD ? get_default_magic() : MAGIC;
258}
259
260public struct magic_set *
261magic_open(int flags)
262{
263	return file_ms_alloc(flags);
264}
265
266private int
267unreadable_info(struct magic_set *ms, mode_t md, const char *file)
268{
269	if (file) {
270		/* We cannot open it, but we were able to stat it. */
271		if (access(file, W_OK) == 0)
272			if (file_printf(ms, "writable, ") == -1)
273				return -1;
274		if (access(file, X_OK) == 0)
275			if (file_printf(ms, "executable, ") == -1)
276				return -1;
277	}
278	if (S_ISREG(md))
279		if (file_printf(ms, "regular file, ") == -1)
280			return -1;
281	if (file_printf(ms, "no read permission") == -1)
282		return -1;
283	return 0;
284}
285
286public void
287magic_close(struct magic_set *ms)
288{
289	if (ms == NULL)
290		return;
291	file_ms_free(ms);
292}
293
294/*
295 * load a magic file
296 */
297public int
298magic_load(struct magic_set *ms, const char *magicfile)
299{
300	if (ms == NULL)
301		return -1;
302	return file_apprentice(ms, magicfile, FILE_LOAD);
303}
304
305#ifndef COMPILE_ONLY
306/*
307 * Install a set of compiled magic buffers.
308 */
309public int
310magic_load_buffers(struct magic_set *ms, void **bufs, size_t *sizes,
311    size_t nbufs)
312{
313	if (ms == NULL)
314		return -1;
315	return buffer_apprentice(ms, RCAST(struct magic **, bufs),
316	    sizes, nbufs);
317}
318#endif
319
320public int
321magic_compile(struct magic_set *ms, const char *magicfile)
322{
323	if (ms == NULL)
324		return -1;
325	return file_apprentice(ms, magicfile, FILE_COMPILE);
326}
327
328public int
329magic_check(struct magic_set *ms, const char *magicfile)
330{
331	if (ms == NULL)
332		return -1;
333	return file_apprentice(ms, magicfile, FILE_CHECK);
334}
335
336public int
337magic_list(struct magic_set *ms, const char *magicfile)
338{
339	if (ms == NULL)
340		return -1;
341	return file_apprentice(ms, magicfile, FILE_LIST);
342}
343
344private void
345close_and_restore(const struct magic_set *ms, const char *name, int fd,
346    const struct stat *sb)
347{
348	if (fd == STDIN_FILENO || name == NULL)
349		return;
350	(void) close(fd);
351
352	if ((ms->flags & MAGIC_PRESERVE_ATIME) != 0) {
353		/*
354		 * Try to restore access, modification times if read it.
355		 * This is really *bad* because it will modify the status
356		 * time of the file... And of course this will affect
357		 * backup programs
358		 */
359#ifdef HAVE_UTIMES
360		struct timeval  utsbuf[2];
361		(void)memset(utsbuf, 0, sizeof(utsbuf));
362		utsbuf[0].tv_sec = sb->st_atime;
363		utsbuf[1].tv_sec = sb->st_mtime;
364
365		(void) utimes(name, utsbuf); /* don't care if loses */
366#elif defined(HAVE_UTIME_H) || defined(HAVE_SYS_UTIME_H)
367		struct utimbuf  utbuf;
368
369		(void)memset(&utbuf, 0, sizeof(utbuf));
370		utbuf.actime = sb->st_atime;
371		utbuf.modtime = sb->st_mtime;
372		(void) utime(name, &utbuf); /* don't care if loses */
373#endif
374	}
375}
376
377#ifndef COMPILE_ONLY
378
379/*
380 * find type of descriptor
381 */
382public const char *
383magic_descriptor(struct magic_set *ms, int fd)
384{
385	if (ms == NULL)
386		return NULL;
387	return file_or_fd(ms, NULL, fd);
388}
389
390/*
391 * find type of named file
392 */
393public const char *
394magic_file(struct magic_set *ms, const char *inname)
395{
396	if (ms == NULL)
397		return NULL;
398	return file_or_fd(ms, inname, STDIN_FILENO);
399}
400
401private const char *
402file_or_fd(struct magic_set *ms, const char *inname, int fd)
403{
404	int	rv = -1;
405	unsigned char *buf;
406	struct stat	sb;
407	ssize_t nbytes = 0;	/* number of bytes read from a datafile */
408	int	ispipe = 0;
409	int	okstat = 0;
410	off_t	pos = CAST(off_t, -1);
411
412	if (file_reset(ms, 1) == -1)
413		goto out;
414
415	/*
416	 * one extra for terminating '\0', and
417	 * some overlapping space for matches near EOF
418	 */
419#define SLOP (1 + sizeof(union VALUETYPE))
420	if ((buf = CAST(unsigned char *, malloc(ms->bytes_max + SLOP))) == NULL)
421		return NULL;
422
423	switch (file_fsmagic(ms, inname, &sb)) {
424	case -1:		/* error */
425		goto done;
426	case 0:			/* nothing found */
427		break;
428	default:		/* matched it and printed type */
429		rv = 0;
430		goto done;
431	}
432
433#ifdef WIN32
434	/* Place stdin in binary mode, so EOF (Ctrl+Z) doesn't stop early. */
435	if (fd == STDIN_FILENO)
436		_setmode(STDIN_FILENO, O_BINARY);
437#endif
438	if (inname != NULL) {
439		int flags = O_RDONLY|O_BINARY|O_NONBLOCK;
440		errno = 0;
441		if ((fd = open(inname, flags)) < 0) {
442			okstat = stat(inname, &sb) == 0;
443			if (okstat && S_ISFIFO(sb.st_mode))
444				ispipe = 1;
445#ifdef WIN32
446			/*
447			 * Can't stat, can't open.  It may have been opened in
448			 * fsmagic, so if the user doesn't have read permission,
449			 * allow it to say so; otherwise an error was probably
450			 * displayed in fsmagic.
451			 */
452			if (!okstat && errno == EACCES) {
453				sb.st_mode = S_IFBLK;
454				okstat = 1;
455			}
456#endif
457			if (okstat &&
458			    unreadable_info(ms, sb.st_mode, inname) == -1)
459				goto done;
460			rv = 0;
461			goto done;
462		}
463	}
464
465	if (fd != -1) {
466		okstat = fstat(fd, &sb) == 0;
467		if (okstat && S_ISFIFO(sb.st_mode))
468			ispipe = 1;
469		if (inname == NULL)
470			pos = lseek(fd, CAST(off_t, 0), SEEK_CUR);
471	}
472
473	/*
474	 * try looking at the first ms->bytes_max bytes
475	 */
476	if (ispipe) {
477		if (fd != -1) {
478			ssize_t r = 0;
479
480			while ((r = sread(fd, RCAST(void *, &buf[nbytes]),
481			    CAST(size_t, ms->bytes_max - nbytes), 1)) > 0) {
482				nbytes += r;
483				if (r < PIPE_BUF) break;
484			}
485		}
486
487		if (nbytes == 0 && inname) {
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 if (fd != -1) {
496		/* Windows refuses to read from a big console buffer. */
497		size_t howmany =
498#if defined(WIN32)
499		    _isatty(fd) ? 8 * 1024 :
500#endif
501		    ms->bytes_max;
502		if ((nbytes = read(fd, RCAST(void *, 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, okstat ? &sb : NULL, inname, buf, CAST(size_t, nbytes)) == -1)
514		goto done;
515	rv = 0;
516done:
517	free(buf);
518	if (fd != -1) {
519		if (pos != CAST(off_t, -1))
520			(void)lseek(fd, pos, SEEK_SET);
521		close_and_restore(ms, inname, fd, &sb);
522	}
523out:
524	return rv == 0 ? file_getbuffer(ms) : NULL;
525}
526
527
528public const char *
529magic_buffer(struct magic_set *ms, const void *buf, size_t nb)
530{
531	if (ms == NULL)
532		return NULL;
533	if (file_reset(ms, 1) == -1)
534		return NULL;
535	/*
536	 * The main work is done here!
537	 * We have the file name and/or the data buffer to be identified.
538	 */
539	if (file_buffer(ms, -1, NULL, NULL, buf, nb) == -1) {
540		return NULL;
541	}
542	return file_getbuffer(ms);
543}
544#endif
545
546public const char *
547magic_error(struct magic_set *ms)
548{
549	if (ms == NULL)
550		return "Magic database is not open";
551	return (ms->event_flags & EVENT_HAD_ERR) ? ms->o.buf : NULL;
552}
553
554public int
555magic_errno(struct magic_set *ms)
556{
557	if (ms == NULL)
558		return EINVAL;
559	return (ms->event_flags & EVENT_HAD_ERR) ? ms->error : 0;
560}
561
562public int
563magic_getflags(struct magic_set *ms)
564{
565	if (ms == NULL)
566		return -1;
567
568	return ms->flags;
569}
570
571public int
572magic_setflags(struct magic_set *ms, int flags)
573{
574	if (ms == NULL)
575		return -1;
576#if !defined(HAVE_UTIME) && !defined(HAVE_UTIMES)
577	if (flags & MAGIC_PRESERVE_ATIME)
578		return -1;
579#endif
580	ms->flags = flags;
581	return 0;
582}
583
584public int
585magic_version(void)
586{
587	return MAGIC_VERSION;
588}
589
590public int
591magic_setparam(struct magic_set *ms, int param, const void *val)
592{
593	if (ms == NULL)
594		return -1;
595	switch (param) {
596	case MAGIC_PARAM_INDIR_MAX:
597		ms->indir_max = CAST(uint16_t, *CAST(const size_t *, val));
598		return 0;
599	case MAGIC_PARAM_NAME_MAX:
600		ms->name_max = CAST(uint16_t, *CAST(const size_t *, val));
601		return 0;
602	case MAGIC_PARAM_ELF_PHNUM_MAX:
603		ms->elf_phnum_max = CAST(uint16_t, *CAST(const size_t *, val));
604		return 0;
605	case MAGIC_PARAM_ELF_SHNUM_MAX:
606		ms->elf_shnum_max = CAST(uint16_t, *CAST(const size_t *, val));
607		return 0;
608	case MAGIC_PARAM_ELF_NOTES_MAX:
609		ms->elf_notes_max = CAST(uint16_t, *CAST(const size_t *, val));
610		return 0;
611	case MAGIC_PARAM_REGEX_MAX:
612		ms->regex_max = CAST(uint16_t, *CAST(const size_t *, val));
613		return 0;
614	case MAGIC_PARAM_BYTES_MAX:
615		ms->bytes_max = *CAST(const size_t *, val);
616		return 0;
617	default:
618		errno = EINVAL;
619		return -1;
620	}
621}
622
623public int
624magic_getparam(struct magic_set *ms, int param, void *val)
625{
626	if (ms == NULL)
627		return -1;
628	switch (param) {
629	case MAGIC_PARAM_INDIR_MAX:
630		*CAST(size_t *, val) = ms->indir_max;
631		return 0;
632	case MAGIC_PARAM_NAME_MAX:
633		*CAST(size_t *, val) = ms->name_max;
634		return 0;
635	case MAGIC_PARAM_ELF_PHNUM_MAX:
636		*CAST(size_t *, val) = ms->elf_phnum_max;
637		return 0;
638	case MAGIC_PARAM_ELF_SHNUM_MAX:
639		*CAST(size_t *, val) = ms->elf_shnum_max;
640		return 0;
641	case MAGIC_PARAM_ELF_NOTES_MAX:
642		*CAST(size_t *, val) = ms->elf_notes_max;
643		return 0;
644	case MAGIC_PARAM_REGEX_MAX:
645		*CAST(size_t *, val) = ms->regex_max;
646		return 0;
647	case MAGIC_PARAM_BYTES_MAX:
648		*CAST(size_t *, val) = ms->bytes_max;
649		return 0;
650	default:
651		errno = EINVAL;
652		return -1;
653	}
654}
655