1 /*
2 * Copyright (c) Ian F. Darwin 1986-1995.
3 * Software written by Ian F. Darwin and others;
4 * maintained 1995-present by Christos Zoulas and others.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice immediately at the beginning of the file, without modification,
11 * this list of conditions, and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
20 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 */
28 /*
29 * compress routines:
30 * zmagic() - returns 0 if not recognized, uncompresses and prints
31 * information if recognized
32 * uncompress(method, old, n, newch) - uncompress old into new,
33 * using method, return sizeof new
34 */
35 #include "file.h"
36
37 #ifndef lint
38 FILE_RCSID("@(#)$File: compress.c,v 1.64 2009/05/08 17:41:58 christos Exp $")
39 #endif
40
41 #include "magic.h"
42 #include <stdlib.h>
43 #ifdef HAVE_UNISTD_H
44 #include <unistd.h>
45 #endif
46 #include <string.h>
47 #include <errno.h>
48 #include <sys/ioctl.h>
49 #ifdef HAVE_SYS_WAIT_H
50 #include <sys/wait.h>
51 #endif
52 #if defined(HAVE_SYS_TIME_H)
53 #include <sys/time.h>
54 #endif
55 #if defined(HAVE_ZLIB_H) && defined(HAVE_LIBZ)
56 #define BUILTIN_DECOMPRESS
57 #include <zlib.h>
58 #endif
59
60 private const struct {
61 const char magic[8];
62 size_t maglen;
63 const char *argv[3];
64 int silent;
65 } compr[] = {
66 { "\037\235", 2, { "gzip", "-cdq", NULL }, 1 }, /* compressed */
67 /* Uncompress can get stuck; so use gzip first if we have it
68 * Idea from Damien Clark, thanks! */
69 { "\037\235", 2, { "uncompress", "-c", NULL }, 1 }, /* compressed */
70 { "\037\213", 2, { "gzip", "-cdq", NULL }, 1 }, /* gzipped */
71 { "\037\236", 2, { "gzip", "-cdq", NULL }, 1 }, /* frozen */
72 { "\037\240", 2, { "gzip", "-cdq", NULL }, 1 }, /* SCO LZH */
73 /* the standard pack utilities do not accept standard input */
74 { "\037\036", 2, { "gzip", "-cdq", NULL }, 0 }, /* packed */
75 { "PK\3\4", 4, { "gzip", "-cdq", NULL }, 1 }, /* pkzipped, */
76 /* ...only first file examined */
77 { "BZh", 3, { "bzip2", "-cd", NULL }, 1 }, /* bzip2-ed */
78 { "LZIP", 4, { "lzip", "-cdq", NULL }, 1 },
79 { "\3757zXZ\0",6,{ "xz", "-cd", NULL }, 1 }, /* XZ Utils */
80 };
81
82 private size_t ncompr = sizeof(compr) / sizeof(compr[0]);
83
84 #define NODATA ((size_t)~0)
85
86
87 private ssize_t swrite(int, const void *, size_t);
88 private size_t uncompressbuf(struct magic_set *, int, size_t,
89 const unsigned char *, unsigned char **, size_t);
90 #ifdef BUILTIN_DECOMPRESS
91 private size_t uncompressgzipped(struct magic_set *, const unsigned char *,
92 unsigned char **, size_t);
93 #endif
94
95 protected int
file_zmagic(struct magic_set * ms,int fd,const char * name,const unsigned char * buf,size_t nbytes)96 file_zmagic(struct magic_set *ms, int fd, const char *name,
97 const unsigned char *buf, size_t nbytes)
98 {
99 unsigned char *newbuf = NULL;
100 size_t i, nsz;
101 int rv = 0;
102 int mime = ms->flags & MAGIC_MIME;
103
104 if ((ms->flags & MAGIC_COMPRESS) == 0)
105 return 0;
106
107 for (i = 0; i < ncompr; i++) {
108 if (nbytes < compr[i].maglen)
109 continue;
110 if (memcmp(buf, compr[i].magic, compr[i].maglen) == 0 &&
111 (nsz = uncompressbuf(ms, fd, i, buf, &newbuf,
112 nbytes)) != NODATA) {
113 ms->flags &= ~MAGIC_COMPRESS;
114 rv = -1;
115 if (file_buffer(ms, -1, name, newbuf, nsz) == -1)
116 goto error;
117
118 if (mime == MAGIC_MIME || mime == 0) {
119 if (file_printf(ms, mime ?
120 " compressed-encoding=" : " (") == -1)
121 goto error;
122 }
123
124 if ((mime == 0 || mime & MAGIC_MIME_ENCODING) &&
125 file_buffer(ms, -1, NULL, buf, nbytes) == -1)
126 goto error;
127
128 if (!mime && file_printf(ms, ")") == -1)
129 goto error;
130 rv = 1;
131 break;
132 }
133 }
134 error:
135 if (newbuf)
136 free(newbuf);
137 ms->flags |= MAGIC_COMPRESS;
138 return rv;
139 }
140
141 /*
142 * `safe' write for sockets and pipes.
143 */
144 private ssize_t
swrite(int fd,const void * buf,size_t n)145 swrite(int fd, const void *buf, size_t n)
146 {
147 ssize_t rv;
148 size_t rn = n;
149
150 do
151 switch (rv = write(fd, buf, n)) {
152 case -1:
153 if (errno == EINTR)
154 continue;
155 return -1;
156 default:
157 n -= rv;
158 buf = CAST(const char *, buf) + rv;
159 break;
160 }
161 while (n > 0);
162 return rn;
163 }
164
165
166 /*
167 * `safe' read for sockets and pipes.
168 */
169 protected ssize_t
sread(int fd,void * buf,size_t n,int canbepipe)170 sread(int fd, void *buf, size_t n, int canbepipe)
171 {
172 ssize_t rv, cnt;
173 #ifdef FIONREAD
174 int t = 0;
175 #endif
176 size_t rn = n;
177
178 if (fd == STDIN_FILENO)
179 goto nocheck;
180
181 #ifdef FIONREAD
182 if ((canbepipe && (ioctl(fd, FIONREAD, &t) == -1)) || (t == 0)) {
183 #ifdef FD_ZERO
184 for (cnt = 0;; cnt++) {
185 fd_set check;
186 struct timeval tout = {0, 100 * 1000};
187 int selrv;
188
189 FD_ZERO(&check);
190 FD_SET(fd, &check);
191
192 /*
193 * Avoid soft deadlock: do not read if there
194 * is nothing to read from sockets and pipes.
195 */
196 selrv = select(fd + 1, &check, NULL, NULL, &tout);
197 if (selrv == -1) {
198 if (errno == EINTR || errno == EAGAIN)
199 continue;
200 } else if (selrv == 0 && cnt >= 5) {
201 return 0;
202 } else
203 break;
204 }
205 #endif
206 (void)ioctl(fd, FIONREAD, &t);
207 }
208
209 if (t > 0 && (size_t)t < n) {
210 n = t;
211 rn = n;
212 }
213 #endif
214
215 nocheck:
216 do
217 switch ((rv = read(fd, buf, n))) {
218 case -1:
219 if (errno == EINTR)
220 continue;
221 return -1;
222 case 0:
223 return rn - n;
224 default:
225 n -= rv;
226 buf = ((char *)buf) + rv;
227 break;
228 }
229 while (n > 0);
230 return rn;
231 }
232
233 protected int
file_pipe2file(struct magic_set * ms,int fd,const void * startbuf,size_t nbytes)234 file_pipe2file(struct magic_set *ms, int fd, const void *startbuf,
235 size_t nbytes)
236 {
237 char buf[4096];
238 ssize_t r;
239 int tfd, te;
240
241 (void)strlcpy(buf, "/tmp/file.XXXXXX", sizeof buf);
242 #ifndef HAVE_MKSTEMP
243 {
244 char *ptr = mktemp(buf);
245 tfd = open(ptr, O_RDWR|O_TRUNC|O_EXCL|O_CREAT, 0600);
246 r = errno;
247 (void)unlink(ptr);
248 errno = r;
249 }
250 #else
251 tfd = mkstemp(buf);
252 te = errno;
253 (void)unlink(buf);
254 errno = te;
255 #endif
256 if (tfd == -1) {
257 file_error(ms, errno,
258 "cannot create temporary file for pipe copy");
259 return -1;
260 }
261
262 if (swrite(tfd, startbuf, nbytes) != (ssize_t)nbytes)
263 r = 1;
264 else {
265 while ((r = sread(fd, buf, sizeof(buf), 1)) > 0)
266 if (swrite(tfd, buf, (size_t)r) != r)
267 break;
268 }
269
270 switch (r) {
271 case -1:
272 file_error(ms, errno, "error copying from pipe to temp file");
273 return -1;
274 case 0:
275 break;
276 default:
277 file_error(ms, errno, "error while writing to temp file");
278 return -1;
279 }
280
281 /*
282 * We duplicate the file descriptor, because fclose on a
283 * tmpfile will delete the file, but any open descriptors
284 * can still access the phantom inode.
285 */
286 if ((fd = dup2(tfd, fd)) == -1) {
287 file_error(ms, errno, "could not dup descriptor for temp file");
288 return -1;
289 }
290 (void)close(tfd);
291 if (lseek(fd, (off_t)0, SEEK_SET) == (off_t)-1) {
292 file_badseek(ms);
293 return -1;
294 }
295 return fd;
296 }
297
298 #ifdef BUILTIN_DECOMPRESS
299
300 #define FHCRC (1 << 1)
301 #define FEXTRA (1 << 2)
302 #define FNAME (1 << 3)
303 #define FCOMMENT (1 << 4)
304
305 private size_t
uncompressgzipped(struct magic_set * ms,const unsigned char * old,unsigned char ** newch,size_t n)306 uncompressgzipped(struct magic_set *ms, const unsigned char *old,
307 unsigned char **newch, size_t n)
308 {
309 unsigned char flg = old[3];
310 size_t data_start = 10;
311 z_stream z;
312 int rc;
313
314 if (flg & FEXTRA) {
315 if (data_start+1 >= n)
316 return 0;
317 data_start += 2 + old[data_start] + old[data_start + 1] * 256;
318 }
319 if (flg & FNAME) {
320 while(data_start < n && old[data_start])
321 data_start++;
322 data_start++;
323 }
324 if(flg & FCOMMENT) {
325 while(data_start < n && old[data_start])
326 data_start++;
327 data_start++;
328 }
329 if(flg & FHCRC)
330 data_start += 2;
331
332 if (data_start >= n)
333 return 0;
334 if ((*newch = CAST(unsigned char *, malloc(HOWMANY + 1))) == NULL) {
335 return 0;
336 }
337
338 /* XXX: const castaway, via strchr */
339 z.next_in = (Bytef *)strchr((const char *)old + data_start,
340 old[data_start]);
341 z.avail_in = CAST(uint32_t, (n - data_start));
342 z.next_out = *newch;
343 z.avail_out = HOWMANY;
344 z.zalloc = Z_NULL;
345 z.zfree = Z_NULL;
346 z.opaque = Z_NULL;
347
348 /* LINTED bug in header macro */
349 rc = inflateInit2(&z, -15);
350 if (rc != Z_OK) {
351 file_error(ms, 0, "zlib: %s", z.msg);
352 return 0;
353 }
354
355 rc = inflate(&z, Z_SYNC_FLUSH);
356 if (rc != Z_OK && rc != Z_STREAM_END) {
357 file_error(ms, 0, "zlib: %s", z.msg);
358 return 0;
359 }
360
361 n = (size_t)z.total_out;
362 (void)inflateEnd(&z);
363
364 /* let's keep the nul-terminate tradition */
365 (*newch)[n] = '\0';
366
367 return n;
368 }
369 #endif
370
371 private size_t
uncompressbuf(struct magic_set * ms,int fd,size_t method,const unsigned char * old,unsigned char ** newch,size_t n)372 uncompressbuf(struct magic_set *ms, int fd, size_t method,
373 const unsigned char *old, unsigned char **newch, size_t n)
374 {
375 int fdin[2], fdout[2];
376 ssize_t r;
377
378 #ifdef BUILTIN_DECOMPRESS
379 /* FIXME: This doesn't cope with bzip2 */
380 if (method == 2)
381 return uncompressgzipped(ms, old, newch, n);
382 #endif
383 (void)fflush(stdout);
384 (void)fflush(stderr);
385
386 if ((fd != -1 && pipe(fdin) == -1) || pipe(fdout) == -1) {
387 file_error(ms, errno, "cannot create pipe");
388 return NODATA;
389 }
390 switch (fork()) {
391 case 0: /* child */
392 (void) close(0);
393 if (fd != -1) {
394 (void) dup(fd);
395 (void) lseek(0, (off_t)0, SEEK_SET);
396 } else {
397 (void) dup(fdin[0]);
398 (void) close(fdin[0]);
399 (void) close(fdin[1]);
400 }
401
402 (void) close(1);
403 (void) dup(fdout[1]);
404 (void) close(fdout[0]);
405 (void) close(fdout[1]);
406 #ifndef DEBUG
407 if (compr[method].silent)
408 (void)close(2);
409 #endif
410
411 (void)execvp(compr[method].argv[0],
412 (char *const *)(intptr_t)compr[method].argv);
413 #ifdef DEBUG
414 (void)fprintf(stderr, "exec `%s' failed (%s)\n",
415 compr[method].argv[0], strerror(errno));
416 #endif
417 exit(1);
418 /*NOTREACHED*/
419 case -1:
420 file_error(ms, errno, "could not fork");
421 return NODATA;
422
423 default: /* parent */
424 (void) close(fdout[1]);
425 if (fd == -1) {
426 (void) close(fdin[0]);
427 /*
428 * fork again, to avoid blocking because both
429 * pipes filled
430 */
431 switch (fork()) {
432 case 0: /* child */
433 (void)close(fdout[0]);
434 if (swrite(fdin[1], old, n) != (ssize_t)n) {
435 #ifdef DEBUG
436 (void)fprintf(stderr,
437 "Write failed (%s)\n",
438 strerror(errno));
439 #endif
440 exit(1);
441 }
442 exit(0);
443 /*NOTREACHED*/
444
445 case -1:
446 #ifdef DEBUG
447 (void)fprintf(stderr, "Fork failed (%s)\n",
448 strerror(errno));
449 #endif
450 exit(1);
451 /*NOTREACHED*/
452
453 default: /* parent */
454 break;
455 }
456 (void) close(fdin[1]);
457 fdin[1] = -1;
458 }
459
460 if ((*newch = (unsigned char *) malloc(HOWMANY + 1)) == NULL) {
461 #ifdef DEBUG
462 (void)fprintf(stderr, "Malloc failed (%s)\n",
463 strerror(errno));
464 #endif
465 n = 0;
466 goto err;
467 }
468 if ((r = sread(fdout[0], *newch, HOWMANY, 0)) <= 0) {
469 #ifdef DEBUG
470 (void)fprintf(stderr, "Read failed (%s)\n",
471 strerror(errno));
472 #endif
473 free(*newch);
474 n = 0;
475 newch[0] = '\0';
476 goto err;
477 } else {
478 n = r;
479 }
480 /* NUL terminate, as every buffer is handled here. */
481 (*newch)[n] = '\0';
482 err:
483 if (fdin[1] != -1)
484 (void) close(fdin[1]);
485 (void) close(fdout[0]);
486 #ifdef WNOHANG
487 while (waitpid(-1, NULL, WNOHANG) != -1)
488 continue;
489 #else
490 (void)wait(NULL);
491 #endif
492 (void) close(fdin[0]);
493
494 return n;
495 }
496 }
497