| 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 1.68 2011/12/08 12:38:24 rrt 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 | #ifndef __MINGW32__ |
| 49 | #include <sys/ioctl.h> |
| 50 | #endif |
| 51 | #ifdef HAVE_SYS_WAIT_H |
| 52 | #include <sys/wait.h> |
| 53 | #endif |
| 54 | #if defined(HAVE_SYS_TIME_H) |
| 55 | #include <sys/time.h> |
| 56 | #endif |
| 57 | #if defined(HAVE_ZLIB_H) && defined(HAVE_LIBZ) |
| 58 | #define BUILTIN_DECOMPRESS |
| 59 | #include <zlib.h> |
| 60 | #endif |
| 61 | |
| 62 | private const struct { |
| 63 | const char magic[8]; |
| 64 | size_t maglen; |
| 65 | const char *argv[3]; |
| 66 | int silent; |
| 67 | } compr[] = { |
| 68 | { "\037\235", 2, { "gzip", "-cdq", NULL }, 1 }, /* compressed */ |
| 69 | /* Uncompress can get stuck; so use gzip first if we have it |
| 70 | * Idea from Damien Clark, thanks! */ |
| 71 | { "\037\235", 2, { "uncompress", "-c", NULL }, 1 }, /* compressed */ |
| 72 | { "\037\213", 2, { "gzip", "-cdq", NULL }, 1 }, /* gzipped */ |
| 73 | { "\037\236", 2, { "gzip", "-cdq", NULL }, 1 }, /* frozen */ |
| 74 | { "\037\240", 2, { "gzip", "-cdq", NULL }, 1 }, /* SCO LZH */ |
| 75 | /* the standard pack utilities do not accept standard input */ |
| 76 | { "\037\036", 2, { "gzip", "-cdq", NULL }, 0 }, /* packed */ |
| 77 | { "PK\3\4", 4, { "gzip", "-cdq", NULL }, 1 }, /* pkzipped, */ |
| 78 | /* ...only first file examined */ |
| 79 | { "BZh", 3, { "bzip2", "-cd", NULL }, 1 }, /* bzip2-ed */ |
| 80 | { "LZIP", 4, { "lzip", "-cdq", NULL }, 1 }, |
| 81 | { "\3757zXZ\0",6,{ "xz", "-cd", NULL }, 1 }, /* XZ Utils */ |
| 82 | { "LRZI", 4, { "lrzip", "-dqo-", NULL }, 1 }, /* LRZIP */ |
| 83 | }; |
| 84 | |
| 85 | #define NODATA ((size_t)~0) |
| 86 | |
| 87 | private ssize_t swrite(int, const void *, size_t); |
| 88 | #if HAVE_FORK |
| 89 | private size_t ncompr = sizeof(compr) / sizeof(compr[0]); |
| 90 | private size_t uncompressbuf(struct magic_set *, int, size_t, |
| 91 | const unsigned char *, unsigned char **, size_t); |
| 92 | #ifdef BUILTIN_DECOMPRESS |
| 93 | private size_t uncompressgzipped(struct magic_set *, const unsigned char *, |
| 94 | unsigned char **, size_t); |
| 95 | #endif |
| 96 | |
| 97 | protected int |
| 98 | file_zmagic(struct magic_set *ms, int fd, const char *name, |
| 99 | const unsigned char *buf, size_t nbytes) |
| 100 | { |
| 101 | unsigned char *newbuf = NULL; |
| 102 | size_t i, nsz; |
| 103 | int rv = 0; |
| 104 | int mime = ms->flags & MAGIC_MIME; |
| 105 | |
| 106 | if ((ms->flags & MAGIC_COMPRESS) == 0) |
| 107 | return 0; |
| 108 | |
| 109 | for (i = 0; i < ncompr; i++) { |
| 110 | if (nbytes < compr[i].maglen) |
| 111 | continue; |
| 112 | if (memcmp(buf, compr[i].magic, compr[i].maglen) == 0 && |
| 113 | (nsz = uncompressbuf(ms, fd, i, buf, &newbuf, |
| 114 | nbytes)) != NODATA) { |
| 115 | ms->flags &= ~MAGIC_COMPRESS; |
| 116 | rv = -1; |
| 117 | if (file_buffer(ms, -1, name, newbuf, nsz) == -1) |
| 118 | goto error; |
| 119 | |
| 120 | if (mime == MAGIC_MIME || mime == 0) { |
| 121 | if (file_printf(ms, mime ? |
| 122 | " compressed-encoding=" : " (") == -1) |
| 123 | goto error; |
| 124 | } |
| 125 | |
| 126 | if ((mime == 0 || mime & MAGIC_MIME_ENCODING) && |
| 127 | file_buffer(ms, -1, NULL, buf, nbytes) == -1) |
| 128 | goto error; |
| 129 | |
| 130 | if (!mime && file_printf(ms, ")") == -1) |
| 131 | goto error; |
| 132 | rv = 1; |
| 133 | break; |
| 134 | } |
| 135 | } |
| 136 | error: |
| 137 | free(newbuf); |
| 138 | ms->flags |= MAGIC_COMPRESS; |
| 139 | return rv; |
| 140 | } |
| 141 | #endif |
| 142 | /* |
| 143 | * `safe' write for sockets and pipes. |
| 144 | */ |
| 145 | private ssize_t |
| 146 | swrite(int fd, const void *buf, size_t n) |
| 147 | { |
| 148 | ssize_t rv; |
| 149 | size_t rn = n; |
| 150 | |
| 151 | do |
| 152 | switch (rv = write(fd, buf, n)) { |
| 153 | case -1: |
| 154 | if (errno == EINTR) |
| 155 | continue; |
| 156 | return -1; |
| 157 | default: |
| 158 | n -= rv; |
| 159 | buf = CAST(const char *, buf) + rv; |
| 160 | break; |
| 161 | } |
| 162 | while (n > 0); |
| 163 | return rn; |
| 164 | } |
| 165 | |
| 166 | |
| 167 | /* |
| 168 | * `safe' read for sockets and pipes. |
| 169 | */ |
| 170 | protected ssize_t |
| 171 | sread(int fd, void *buf, size_t n, int canbepipe __attribute__ ((unused))) |
| 172 | { |
| 173 | ssize_t rv; |
| 174 | #ifdef FD_ZERO |
| 175 | ssize_t cnt; |
| 176 | #endif |
| 177 | #ifdef FIONREAD |
| 178 | int t = 0; |
| 179 | #endif |
| 180 | size_t rn = n; |
| 181 | |
| 182 | if (fd == STDIN_FILENO) |
| 183 | goto nocheck; |
| 184 | |
| 185 | #ifdef FIONREAD |
| 186 | if ((canbepipe && (ioctl(fd, FIONREAD, &t) == -1)) || (t == 0)) { |
| 187 | #ifdef FD_ZERO |
| 188 | for (cnt = 0;; cnt++) { |
| 189 | fd_set check; |
| 190 | struct timeval tout = {0, 100 * 1000}; |
| 191 | int selrv; |
| 192 | |
| 193 | FD_ZERO(&check); |
| 194 | FD_SET(fd, &check); |
| 195 | |
| 196 | /* |
| 197 | * Avoid soft deadlock: do not read if there |
| 198 | * is nothing to read from sockets and pipes. |
| 199 | */ |
| 200 | selrv = select(fd + 1, &check, NULL, NULL, &tout); |
| 201 | if (selrv == -1) { |
| 202 | if (errno == EINTR || errno == EAGAIN) |
| 203 | continue; |
| 204 | } else if (selrv == 0 && cnt >= 5) { |
| 205 | return 0; |
| 206 | } else |
| 207 | break; |
| 208 | } |
| 209 | #endif |
| 210 | (void)ioctl(fd, FIONREAD, &t); |
| 211 | } |
| 212 | |
| 213 | if (t > 0 && (size_t)t < n) { |
| 214 | n = t; |
| 215 | rn = n; |
| 216 | } |
| 217 | #endif |
| 218 | |
| 219 | nocheck: |
| 220 | do |
| 221 | switch ((rv = read(fd, buf, n))) { |
| 222 | case -1: |
| 223 | if (errno == EINTR) |
| 224 | continue; |
| 225 | return -1; |
| 226 | case 0: |
| 227 | return rn - n; |
| 228 | default: |
| 229 | n -= rv; |
| 230 | buf = ((char *)buf) + rv; |
| 231 | break; |
| 232 | } |
| 233 | while (n > 0); |
| 234 | return rn; |
| 235 | } |
| 236 | |
| 237 | protected int |
| 238 | file_pipe2file(struct magic_set *ms, int fd, const void *startbuf, |
| 239 | size_t nbytes) |
| 240 | { |
| 241 | char buf[4096]; |
| 242 | ssize_t r; |
| 243 | int tfd; |
| 244 | #ifdef HAVE_MKSTEMP |
| 245 | int te; |
| 246 | #endif |
| 247 | |
| 248 | (void)strlcpy(buf, "/tmp/file.XXXXXX", sizeof buf); |
| 249 | #ifndef HAVE_MKSTEMP |
| 250 | { |
| 251 | char *ptr = mktemp(buf); |
| 252 | tfd = open(ptr, O_RDWR|O_TRUNC|O_EXCL|O_CREAT, 0600); |
| 253 | r = errno; |
| 254 | (void)unlink(ptr); |
| 255 | errno = r; |
| 256 | } |
| 257 | #else |
| 258 | tfd = mkstemp(buf); |
| 259 | te = errno; |
| 260 | (void)unlink(buf); |
| 261 | errno = te; |
| 262 | #endif |
| 263 | if (tfd == -1) { |
| 264 | file_error(ms, errno, |
| 265 | "cannot create temporary file for pipe copy"); |
| 266 | return -1; |
| 267 | } |
| 268 | |
| 269 | if (swrite(tfd, startbuf, nbytes) != (ssize_t)nbytes) |
| 270 | r = 1; |
| 271 | else { |
| 272 | while ((r = sread(fd, buf, sizeof(buf), 1)) > 0) |
| 273 | if (swrite(tfd, buf, (size_t)r) != r) |
| 274 | break; |
| 275 | } |
| 276 | |
| 277 | switch (r) { |
| 278 | case -1: |
| 279 | file_error(ms, errno, "error copying from pipe to temp file"); |
| 280 | return -1; |
| 281 | case 0: |
| 282 | break; |
| 283 | default: |
| 284 | file_error(ms, errno, "error while writing to temp file"); |
| 285 | return -1; |
| 286 | } |
| 287 | |
| 288 | /* |
| 289 | * We duplicate the file descriptor, because fclose on a |
| 290 | * tmpfile will delete the file, but any open descriptors |
| 291 | * can still access the phantom inode. |
| 292 | */ |
| 293 | if ((fd = dup2(tfd, fd)) == -1) { |
| 294 | file_error(ms, errno, "could not dup descriptor for temp file"); |
| 295 | return -1; |
| 296 | } |
| 297 | (void)close(tfd); |
| 298 | if (lseek(fd, (off_t)0, SEEK_SET) == (off_t)-1) { |
| 299 | file_badseek(ms); |
| 300 | return -1; |
| 301 | } |
| 302 | return fd; |
| 303 | } |
| 304 | #if HAVE_FORK |
| 305 | #ifdef BUILTIN_DECOMPRESS |
| 306 | |
| 307 | #define FHCRC (1 << 1) |
| 308 | #define FEXTRA (1 << 2) |
| 309 | #define FNAME (1 << 3) |
| 310 | #define FCOMMENT (1 << 4) |
| 311 | |
| 312 | private size_t |
| 313 | uncompressgzipped(struct magic_set *ms, const unsigned char *old, |
| 314 | unsigned char **newch, size_t n) |
| 315 | { |
| 316 | unsigned char flg = old[3]; |
| 317 | size_t data_start = 10; |
| 318 | z_stream z; |
| 319 | int rc; |
| 320 | |
| 321 | if (flg & FEXTRA) { |
| 322 | if (data_start+1 >= n) |
| 323 | return 0; |
| 324 | data_start += 2 + old[data_start] + old[data_start + 1] * 256; |
| 325 | } |
| 326 | if (flg & FNAME) { |
| 327 | while(data_start < n && old[data_start]) |
| 328 | data_start++; |
| 329 | data_start++; |
| 330 | } |
| 331 | if(flg & FCOMMENT) { |
| 332 | while(data_start < n && old[data_start]) |
| 333 | data_start++; |
| 334 | data_start++; |
| 335 | } |
| 336 | if(flg & FHCRC) |
| 337 | data_start += 2; |
| 338 | |
| 339 | if (data_start >= n) |
| 340 | return 0; |
| 341 | if ((*newch = CAST(unsigned char *, malloc(HOWMANY + 1))) == NULL) { |
| 342 | return 0; |
| 343 | } |
| 344 | |
| 345 | /* XXX: const castaway, via strchr */ |
| 346 | z.next_in = (Bytef *)strchr((const char *)old + data_start, |
| 347 | old[data_start]); |
| 348 | z.avail_in = CAST(uint32_t, (n - data_start)); |
| 349 | z.next_out = *newch; |
| 350 | z.avail_out = HOWMANY; |
| 351 | z.zalloc = Z_NULL; |
| 352 | z.zfree = Z_NULL; |
| 353 | z.opaque = Z_NULL; |
| 354 | |
| 355 | /* LINTED bug in header macro */ |
| 356 | rc = inflateInit2(&z, -15); |
| 357 | if (rc != Z_OK) { |
| 358 | file_error(ms, 0, "zlib: %s", z.msg); |
| 359 | return 0; |
| 360 | } |
| 361 | |
| 362 | rc = inflate(&z, Z_SYNC_FLUSH); |
| 363 | if (rc != Z_OK && rc != Z_STREAM_END) { |
| 364 | file_error(ms, 0, "zlib: %s", z.msg); |
| 365 | return 0; |
| 366 | } |
| 367 | |
| 368 | n = (size_t)z.total_out; |
| 369 | (void)inflateEnd(&z); |
| 370 | |
| 371 | /* let's keep the nul-terminate tradition */ |
| 372 | (*newch)[n] = '\0'; |
| 373 | |
| 374 | return n; |
| 375 | } |
| 376 | #endif |
| 377 | |
| 378 | private size_t |
| 379 | uncompressbuf(struct magic_set *ms, int fd, size_t method, |
| 380 | const unsigned char *old, unsigned char **newch, size_t n) |
| 381 | { |
| 382 | int fdin[2], fdout[2]; |
| 383 | ssize_t r; |
| 384 | pid_t pid; |
| 385 | |
| 386 | #ifdef BUILTIN_DECOMPRESS |
| 387 | /* FIXME: This doesn't cope with bzip2 */ |
| 388 | if (method == 2) |
| 389 | return uncompressgzipped(ms, old, newch, n); |
| 390 | #endif |
| 391 | (void)fflush(stdout); |
| 392 | (void)fflush(stderr); |
| 393 | |
| 394 | if ((fd != -1 && pipe(fdin) == -1) || pipe(fdout) == -1) { |
| 395 | file_error(ms, errno, "cannot create pipe"); |
| 396 | return NODATA; |
| 397 | } |
| 398 | switch (pid = fork()) { |
| 399 | case 0: /* child */ |
| 400 | (void) close(0); |
| 401 | if (fd != -1) { |
| 402 | (void) dup(fd); |
| 403 | (void) lseek(0, (off_t)0, SEEK_SET); |
| 404 | } else { |
| 405 | (void) dup(fdin[0]); |
| 406 | (void) close(fdin[0]); |
| 407 | (void) close(fdin[1]); |
| 408 | } |
| 409 | |
| 410 | (void) close(1); |
| 411 | (void) dup(fdout[1]); |
| 412 | (void) close(fdout[0]); |
| 413 | (void) close(fdout[1]); |
| 414 | #ifndef DEBUG |
| 415 | if (compr[method].silent) |
| 416 | (void)close(2); |
| 417 | #endif |
| 418 | |
| 419 | (void)execvp(compr[method].argv[0], |
| 420 | (char *const *)(intptr_t)compr[method].argv); |
| 421 | #ifdef DEBUG |
| 422 | (void)fprintf(stderr, "exec `%s' failed (%s)\n", |
| 423 | compr[method].argv[0], strerror(errno)); |
| 424 | #endif |
| 425 | exit(1); |
| 426 | /*NOTREACHED*/ |
| 427 | case -1: |
| 428 | file_error(ms, errno, "could not fork"); |
| 429 | return NODATA; |
| 430 | |
| 431 | default: /* parent */ |
| 432 | (void) close(fdout[1]); |
| 433 | if (fd == -1) { |
| 434 | (void) close(fdin[0]); |
| 435 | /* |
| 436 | * fork again, to avoid blocking because both |
| 437 | * pipes filled |
| 438 | */ |
| 439 | switch (fork()) { |
| 440 | case 0: /* child */ |
| 441 | (void)close(fdout[0]); |
| 442 | if (swrite(fdin[1], old, n) != (ssize_t)n) { |
| 443 | #ifdef DEBUG |
| 444 | (void)fprintf(stderr, |
| 445 | "Write failed (%s)\n", |
| 446 | strerror(errno)); |
| 447 | #endif |
| 448 | exit(1); |
| 449 | } |
| 450 | exit(0); |
| 451 | /*NOTREACHED*/ |
| 452 | |
| 453 | case -1: |
| 454 | #ifdef DEBUG |
| 455 | (void)fprintf(stderr, "Fork failed (%s)\n", |
| 456 | strerror(errno)); |
| 457 | #endif |
| 458 | exit(1); |
| 459 | /*NOTREACHED*/ |
| 460 | |
| 461 | default: /* parent */ |
| 462 | break; |
| 463 | } |
| 464 | (void) close(fdin[1]); |
| 465 | fdin[1] = -1; |
| 466 | } |
| 467 | |
| 468 | if ((*newch = (unsigned char *) malloc(HOWMANY + 1)) == NULL) { |
| 469 | #ifdef DEBUG |
| 470 | (void)fprintf(stderr, "Malloc failed (%s)\n", |
| 471 | strerror(errno)); |
| 472 | #endif |
| 473 | n = 0; |
| 474 | goto err; |
| 475 | } |
| 476 | if ((r = sread(fdout[0], *newch, HOWMANY, 0)) <= 0) { |
| 477 | #ifdef DEBUG |
| 478 | (void)fprintf(stderr, "Read failed (%s)\n", |
| 479 | strerror(errno)); |
| 480 | #endif |
| 481 | free(*newch); |
| 482 | n = 0; |
| 483 | newch[0] = '\0'; |
| 484 | goto err; |
| 485 | } else { |
| 486 | n = r; |
| 487 | } |
| 488 | /* NUL terminate, as every buffer is handled here. */ |
| 489 | (*newch)[n] = '\0'; |
| 490 | err: |
| 491 | if (fdin[1] != -1) |
| 492 | (void) close(fdin[1]); |
| 493 | (void) close(fdout[0]); |
| 494 | #ifdef WNOHANG |
| 495 | while (waitpid(pid, NULL, WNOHANG) != -1) |
| 496 | continue; |
| 497 | #else |
| 498 | (void)wait(NULL); |
| 499 | #endif |
| 500 | (void) close(fdin[0]); |
| 501 | |
| 502 | return n; |
| 503 | } |
| 504 | } |
| 505 | #endif |