1 /* 2 * Copyright (c) 1989, 1993, 1994 3 * The Regents of the University of California. All rights reserved. 4 * 5 * This code is derived from software contributed to Berkeley by 6 * Chris Newcomb. 7 * 8 * Redistribution and use in source and binary forms, with or without 9 * modification, are permitted provided that the following conditions 10 * are met: 11 * 1. Redistributions of source code must retain the above copyright 12 * notice, this list of conditions and the following disclaimer. 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in the 15 * documentation and/or other materials provided with the distribution. 16 * 3. Neither the name of the University nor the names of its contributors 17 * may be used to endorse or promote products derived from this software 18 * without specific prior written permission. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 23 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 30 * SUCH DAMAGE. 31 */ 32 33 #ifndef lint 34 static const char copyright[] = 35 "@(#) Copyright (c) 1989, 1993, 1994\n\ 36 The Regents of the University of California. All rights reserved.\n"; 37 #endif /* not lint */ 38 39 #ifndef lint 40 #if 0 41 static const char sccsid[] = "@(#)du.c 8.5 (Berkeley) 5/4/95"; 42 #endif 43 #endif /* not lint */ 44 #include <sys/cdefs.h> 45 __FBSDID("$FreeBSD$"); 46 47 #include <sys/param.h> 48 #include <sys/queue.h> 49 #include <sys/stat.h> 50 51 #include <err.h> 52 #include <errno.h> 53 #include <fnmatch.h> 54 #include <fts.h> 55 #include <libutil.h> 56 #include <locale.h> 57 #include <stdint.h> 58 #include <stdio.h> 59 #include <stdlib.h> 60 #include <string.h> 61 #include <sysexits.h> 62 #include <unistd.h> 63 64 static SLIST_HEAD(ignhead, ignentry) ignores; 65 struct ignentry { 66 char *mask; 67 SLIST_ENTRY(ignentry) next; 68 }; 69 70 static int linkchk(FTSENT *); 71 static void usage(void); 72 static void prthumanval(int64_t); 73 static void ignoreadd(const char *); 74 static void ignoreclean(void); 75 static int ignorep(FTSENT *); 76 static void siginfo(int __unused); 77 78 static int nodumpflag = 0; 79 static int Aflag; 80 static long blocksize, cblocksize; 81 static volatile sig_atomic_t info; 82 83 int 84 main(int argc, char *argv[]) 85 { 86 FTS *fts; 87 FTSENT *p; 88 off_t savednumber, curblocks; 89 off_t threshold, threshold_sign; 90 int ftsoptions; 91 int depth; 92 int Hflag, Lflag, aflag, sflag, dflag, cflag; 93 int hflag, lflag, ch, notused, rval; 94 char **save; 95 static char dot[] = "."; 96 97 setlocale(LC_ALL, ""); 98 99 Hflag = Lflag = aflag = sflag = dflag = cflag = hflag = 100 lflag = Aflag = 0; 101 102 save = argv; 103 ftsoptions = FTS_PHYSICAL; 104 savednumber = 0; 105 threshold = 0; 106 threshold_sign = 1; 107 cblocksize = DEV_BSIZE; 108 blocksize = 0; 109 depth = INT_MAX; 110 SLIST_INIT(&ignores); 111 112 while ((ch = getopt(argc, argv, "AB:HI:LPasd:cghklmnrt:x")) != -1) 113 switch (ch) { 114 case 'A': 115 Aflag = 1; 116 break; 117 case 'B': 118 errno = 0; 119 cblocksize = atoi(optarg); 120 if (errno == ERANGE || cblocksize <= 0) { 121 warnx("invalid argument to option B: %s", 122 optarg); 123 usage(); 124 } 125 break; 126 case 'H': 127 Hflag = 1; 128 Lflag = 0; 129 break; 130 case 'I': 131 ignoreadd(optarg); 132 break; 133 case 'L': 134 Lflag = 1; 135 Hflag = 0; 136 break; 137 case 'P': 138 Hflag = Lflag = 0; 139 break; 140 case 'a': 141 aflag = 1; 142 break; 143 case 's': 144 sflag = 1; 145 break; 146 case 'd': 147 dflag = 1; 148 errno = 0; 149 depth = atoi(optarg); 150 if (errno == ERANGE || depth < 0) { 151 warnx("invalid argument to option d: %s", 152 optarg); 153 usage(); 154 } 155 break; 156 case 'c': 157 cflag = 1; 158 break; 159 case 'g': 160 hflag = 0; 161 blocksize = 1073741824; 162 break; 163 case 'h': 164 hflag = 1; 165 break; 166 case 'k': 167 hflag = 0; 168 blocksize = 1024; 169 break; 170 case 'l': 171 lflag = 1; 172 break; 173 case 'm': 174 hflag = 0; 175 blocksize = 1048576; 176 break; 177 case 'n': 178 nodumpflag = 1; 179 break; 180 case 'r': /* Compatibility. */ 181 break; 182 case 't' : 183 if (expand_number(optarg, &threshold) != 0 || 184 threshold == 0) { 185 warnx("invalid threshold: %s", optarg); 186 usage(); 187 } else if (threshold < 0) 188 threshold_sign = -1; 189 break; 190 case 'x': 191 ftsoptions |= FTS_XDEV; 192 break; 193 case '?': 194 default: 195 usage(); 196 /* NOTREACHED */ 197 } 198 199 argc -= optind; 200 argv += optind; 201 202 /* 203 * XXX 204 * Because of the way that fts(3) works, logical walks will not count 205 * the blocks actually used by symbolic links. We rationalize this by 206 * noting that users computing logical sizes are likely to do logical 207 * copies, so not counting the links is correct. The real reason is 208 * that we'd have to re-implement the kernel's symbolic link traversing 209 * algorithm to get this right. If, for example, you have relative 210 * symbolic links referencing other relative symbolic links, it gets 211 * very nasty, very fast. The bottom line is that it's documented in 212 * the man page, so it's a feature. 213 */ 214 215 if (Hflag) 216 ftsoptions |= FTS_COMFOLLOW; 217 if (Lflag) { 218 ftsoptions &= ~FTS_PHYSICAL; 219 ftsoptions |= FTS_LOGICAL; 220 } 221 222 if (!Aflag && (cblocksize % DEV_BSIZE) != 0) 223 cblocksize = howmany(cblocksize, DEV_BSIZE) * DEV_BSIZE; 224 225 if (aflag + dflag + sflag > 1) 226 usage(); 227 if (sflag) 228 depth = 0; 229 230 if (!*argv) { 231 argv = save; 232 argv[0] = dot; 233 argv[1] = NULL; 234 } 235 236 if (blocksize == 0) 237 (void)getbsize(¬used, &blocksize); 238 239 if (!Aflag) { 240 cblocksize /= DEV_BSIZE; 241 blocksize /= DEV_BSIZE; 242 } 243 244 if (threshold != 0) 245 threshold = howmany(threshold / DEV_BSIZE * cblocksize, 246 blocksize); 247 248 rval = 0; 249 250 (void)signal(SIGINFO, siginfo); 251 252 if ((fts = fts_open(argv, ftsoptions, NULL)) == NULL) 253 err(1, "fts_open"); 254 255 while ((p = fts_read(fts)) != NULL) { 256 switch (p->fts_info) { 257 case FTS_D: /* Ignore. */ 258 if (ignorep(p)) 259 fts_set(fts, p, FTS_SKIP); 260 break; 261 case FTS_DP: 262 if (ignorep(p)) 263 break; 264 265 curblocks = Aflag ? 266 howmany(p->fts_statp->st_size, cblocksize) : 267 howmany(p->fts_statp->st_blocks, cblocksize); 268 p->fts_parent->fts_bignum += p->fts_bignum += 269 curblocks; 270 271 if (p->fts_level <= depth && threshold <= 272 threshold_sign * howmany(p->fts_bignum * 273 cblocksize, blocksize)) { 274 if (hflag) { 275 prthumanval(p->fts_bignum); 276 (void)printf("\t%s\n", p->fts_path); 277 } else { 278 (void)printf("%jd\t%s\n", 279 (intmax_t)howmany(p->fts_bignum * 280 cblocksize, blocksize), 281 p->fts_path); 282 } 283 } 284 if (info) { 285 info = 0; 286 (void)printf("\t%s\n", p->fts_path); 287 } 288 break; 289 case FTS_DC: /* Ignore. */ 290 break; 291 case FTS_DNR: /* Warn, continue. */ 292 case FTS_ERR: 293 case FTS_NS: 294 warnx("%s: %s", p->fts_path, strerror(p->fts_errno)); 295 rval = 1; 296 break; 297 default: 298 if (ignorep(p)) 299 break; 300 301 if (lflag == 0 && p->fts_statp->st_nlink > 1 && 302 linkchk(p)) 303 break; 304 305 curblocks = Aflag ? 306 howmany(p->fts_statp->st_size, cblocksize) : 307 howmany(p->fts_statp->st_blocks, cblocksize); 308 309 if (aflag || p->fts_level == 0) { 310 if (hflag) { 311 prthumanval(curblocks); 312 (void)printf("\t%s\n", p->fts_path); 313 } else { 314 (void)printf("%jd\t%s\n", 315 (intmax_t)howmany(curblocks * 316 cblocksize, blocksize), 317 p->fts_path); 318 } 319 } 320 321 p->fts_parent->fts_bignum += curblocks; 322 } 323 savednumber = p->fts_parent->fts_bignum; 324 } 325 326 if (errno) 327 err(1, "fts_read"); 328 329 if (cflag) { 330 if (hflag) { 331 prthumanval(savednumber); 332 (void)printf("\ttotal\n"); 333 } else { 334 (void)printf("%jd\ttotal\n", (intmax_t)howmany( 335 savednumber * cblocksize, blocksize)); 336 } 337 } 338 339 ignoreclean(); 340 exit(rval); 341 } 342 343 static int 344 linkchk(FTSENT *p) 345 { 346 struct links_entry { 347 struct links_entry *next; 348 struct links_entry *previous; 349 int links; 350 dev_t dev; 351 ino_t ino; 352 }; 353 static const size_t links_hash_initial_size = 8192; 354 static struct links_entry **buckets; 355 static struct links_entry *free_list; 356 static size_t number_buckets; 357 static unsigned long number_entries; 358 static char stop_allocating; 359 struct links_entry *le, **new_buckets; 360 struct stat *st; 361 size_t i, new_size; 362 int hash; 363 364 st = p->fts_statp; 365 366 /* If necessary, initialize the hash table. */ 367 if (buckets == NULL) { 368 number_buckets = links_hash_initial_size; 369 buckets = malloc(number_buckets * sizeof(buckets[0])); 370 if (buckets == NULL) 371 errx(1, "No memory for hardlink detection"); 372 for (i = 0; i < number_buckets; i++) 373 buckets[i] = NULL; 374 } 375 376 /* If the hash table is getting too full, enlarge it. */ 377 if (number_entries > number_buckets * 10 && !stop_allocating) { 378 new_size = number_buckets * 2; 379 new_buckets = calloc(new_size, sizeof(struct links_entry *)); 380 381 /* Try releasing the free list to see if that helps. */ 382 if (new_buckets == NULL && free_list != NULL) { 383 while (free_list != NULL) { 384 le = free_list; 385 free_list = le->next; 386 free(le); 387 } 388 new_buckets = calloc(new_size, sizeof(new_buckets[0])); 389 } 390 391 if (new_buckets == NULL) { 392 stop_allocating = 1; 393 warnx("No more memory for tracking hard links"); 394 } else { 395 for (i = 0; i < number_buckets; i++) { 396 while (buckets[i] != NULL) { 397 /* Remove entry from old bucket. */ 398 le = buckets[i]; 399 buckets[i] = le->next; 400 401 /* Add entry to new bucket. */ 402 hash = (le->dev ^ le->ino) % new_size; 403 404 if (new_buckets[hash] != NULL) 405 new_buckets[hash]->previous = 406 le; 407 le->next = new_buckets[hash]; 408 le->previous = NULL; 409 new_buckets[hash] = le; 410 } 411 } 412 free(buckets); 413 buckets = new_buckets; 414 number_buckets = new_size; 415 } 416 } 417 418 /* Try to locate this entry in the hash table. */ 419 hash = ( st->st_dev ^ st->st_ino ) % number_buckets; 420 for (le = buckets[hash]; le != NULL; le = le->next) { 421 if (le->dev == st->st_dev && le->ino == st->st_ino) { 422 /* 423 * Save memory by releasing an entry when we've seen 424 * all of its links. 425 */ 426 if (--le->links <= 0) { 427 if (le->previous != NULL) 428 le->previous->next = le->next; 429 if (le->next != NULL) 430 le->next->previous = le->previous; 431 if (buckets[hash] == le) 432 buckets[hash] = le->next; 433 number_entries--; 434 /* Recycle this node through the free list */ 435 if (stop_allocating) { 436 free(le); 437 } else { 438 le->next = free_list; 439 free_list = le; 440 } 441 } 442 return (1); 443 } 444 } 445 446 if (stop_allocating) 447 return (0); 448 449 /* Add this entry to the links cache. */ 450 if (free_list != NULL) { 451 /* Pull a node from the free list if we can. */ 452 le = free_list; 453 free_list = le->next; 454 } else 455 /* Malloc one if we have to. */ 456 le = malloc(sizeof(struct links_entry)); 457 if (le == NULL) { 458 stop_allocating = 1; 459 warnx("No more memory for tracking hard links"); 460 return (0); 461 } 462 le->dev = st->st_dev; 463 le->ino = st->st_ino; 464 le->links = st->st_nlink - 1; 465 number_entries++; 466 le->next = buckets[hash]; 467 le->previous = NULL; 468 if (buckets[hash] != NULL) 469 buckets[hash]->previous = le; 470 buckets[hash] = le; 471 return (0); 472 } 473 474 static void 475 prthumanval(int64_t bytes) 476 { 477 char buf[5]; 478 479 bytes *= cblocksize; 480 if (!Aflag) 481 bytes *= DEV_BSIZE; 482 483 humanize_number(buf, sizeof(buf), bytes, "", HN_AUTOSCALE, 484 HN_B | HN_NOSPACE | HN_DECIMAL); 485 486 (void)printf("%4s", buf); 487 } 488 489 static void 490 usage(void) 491 { 492 (void)fprintf(stderr, 493 "usage: du [-Aclnx] [-H | -L | -P] [-g | -h | -k | -m] " 494 "[-a | -s | -d depth] [-B blocksize] [-I mask] " 495 "[-t threshold] [file ...]\n"); 496 exit(EX_USAGE); 497 } 498 499 static void 500 ignoreadd(const char *mask) 501 { 502 struct ignentry *ign; 503 504 ign = calloc(1, sizeof(*ign)); 505 if (ign == NULL) 506 errx(1, "cannot allocate memory"); 507 ign->mask = strdup(mask); 508 if (ign->mask == NULL) 509 errx(1, "cannot allocate memory"); 510 SLIST_INSERT_HEAD(&ignores, ign, next); 511 } 512 513 static void 514 ignoreclean(void) 515 { 516 struct ignentry *ign; 517 518 while (!SLIST_EMPTY(&ignores)) { 519 ign = SLIST_FIRST(&ignores); 520 SLIST_REMOVE_HEAD(&ignores, next); 521 free(ign->mask); 522 free(ign); 523 } 524 } 525 526 static int 527 ignorep(FTSENT *ent) 528 { 529 struct ignentry *ign; 530 531 if (nodumpflag && (ent->fts_statp->st_flags & UF_NODUMP)) 532 return 1; 533 SLIST_FOREACH(ign, &ignores, next) 534 if (fnmatch(ign->mask, ent->fts_name, 0) != FNM_NOMATCH) 535 return 1; 536 return 0; 537 } 538 539 static void 540 siginfo(int sig __unused) 541 { 542 543 info = 1; 544 } 545