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. All advertising materials mentioning features or use of this software 17 * must display the following acknowledgement: 18 * This product includes software developed by the University of 19 * California, Berkeley and its contributors. 20 * 4. Neither the name of the University nor the names of its contributors 21 * may be used to endorse or promote products derived from this software 22 * without specific prior written permission. 23 * 24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 34 * SUCH DAMAGE. 35 */ 36 37 #ifndef lint 38 static const char copyright[] = 39 "@(#) Copyright (c) 1989, 1993, 1994\n\ 40 The Regents of the University of California. All rights reserved.\n"; 41 #endif /* not lint */ 42 43 #ifndef lint 44 #if 0 45 static const char sccsid[] = "@(#)du.c 8.5 (Berkeley) 5/4/95"; 46 #endif 47 #endif /* not lint */ 48 #include <sys/cdefs.h> 49 __FBSDID("$FreeBSD$"); 50 51 #include <sys/param.h> 52 #include <sys/queue.h> 53 #include <sys/stat.h> 54 55 #include <err.h> 56 #include <errno.h> 57 #include <fnmatch.h> 58 #include <fts.h> 59 #include <math.h> 60 #include <stdio.h> 61 #include <stdlib.h> 62 #include <string.h> 63 #include <sysexits.h> 64 #include <unistd.h> 65 66 #define KILO_SZ(n) (n) 67 #define MEGA_SZ(n) ((n) * (n)) 68 #define GIGA_SZ(n) ((n) * (n) * (n)) 69 #define TERA_SZ(n) ((n) * (n) * (n) * (n)) 70 #define PETA_SZ(n) ((n) * (n) * (n) * (n) * (n)) 71 72 #define KILO_2_SZ (KILO_SZ(1024ULL)) 73 #define MEGA_2_SZ (MEGA_SZ(1024ULL)) 74 #define GIGA_2_SZ (GIGA_SZ(1024ULL)) 75 #define TERA_2_SZ (TERA_SZ(1024ULL)) 76 #define PETA_2_SZ (PETA_SZ(1024ULL)) 77 78 #define KILO_SI_SZ (KILO_SZ(1000ULL)) 79 #define MEGA_SI_SZ (MEGA_SZ(1000ULL)) 80 #define GIGA_SI_SZ (GIGA_SZ(1000ULL)) 81 #define TERA_SI_SZ (TERA_SZ(1000ULL)) 82 #define PETA_SI_SZ (PETA_SZ(1000ULL)) 83 84 unsigned long long vals_si [] = {1, KILO_SI_SZ, MEGA_SI_SZ, GIGA_SI_SZ, TERA_SI_SZ, PETA_SI_SZ}; 85 unsigned long long vals_base2[] = {1, KILO_2_SZ, MEGA_2_SZ, GIGA_2_SZ, TERA_2_SZ, PETA_2_SZ}; 86 unsigned long long *valp; 87 88 typedef enum { NONE, KILO, MEGA, GIGA, TERA, PETA, UNIT_MAX } unit_t; 89 90 int unitp [] = { NONE, KILO, MEGA, GIGA, TERA, PETA }; 91 92 SLIST_HEAD(ignhead, ignentry) ignores; 93 struct ignentry { 94 char *mask; 95 SLIST_ENTRY(ignentry) next; 96 }; 97 98 static int linkchk(FTSENT *); 99 static void usage(void); 100 void prthumanval(double); 101 unit_t unit_adjust(double *); 102 void ignoreadd(const char *); 103 void ignoreclean(void); 104 int ignorep(FTSENT *); 105 106 int 107 main(int argc, char *argv[]) 108 { 109 FTS *fts; 110 FTSENT *p; 111 long blocksize, savednumber = 0; 112 int ftsoptions; 113 int listall; 114 int depth; 115 int Hflag, Lflag, Pflag, aflag, sflag, dflag, cflag, hflag, ch, notused, rval; 116 char **save; 117 static char dot[] = "."; 118 119 Hflag = Lflag = Pflag = aflag = sflag = dflag = cflag = hflag = 0; 120 121 save = argv; 122 ftsoptions = 0; 123 depth = INT_MAX; 124 SLIST_INIT(&ignores); 125 126 while ((ch = getopt(argc, argv, "HI:LPasd:chkrx")) != -1) 127 switch (ch) { 128 case 'H': 129 Hflag = 1; 130 break; 131 case 'I': 132 ignoreadd(optarg); 133 break; 134 case 'L': 135 if (Pflag) 136 usage(); 137 Lflag = 1; 138 break; 139 case 'P': 140 if (Lflag) 141 usage(); 142 Pflag = 1; 143 break; 144 case 'a': 145 aflag = 1; 146 break; 147 case 's': 148 sflag = 1; 149 break; 150 case 'd': 151 dflag = 1; 152 errno = 0; 153 depth = atoi(optarg); 154 if (errno == ERANGE || depth < 0) { 155 warnx("invalid argument to option d: %s", optarg); 156 usage(); 157 } 158 break; 159 case 'c': 160 cflag = 1; 161 break; 162 case 'h': 163 putenv("BLOCKSIZE=512"); 164 hflag = 1; 165 valp = vals_base2; 166 break; 167 case 'k': 168 hflag = 0; 169 putenv("BLOCKSIZE=1024"); 170 break; 171 case 'r': /* Compatibility. */ 172 break; 173 case 'x': 174 ftsoptions |= FTS_XDEV; 175 break; 176 case '?': 177 default: 178 usage(); 179 } 180 181 argc -= optind; 182 argv += optind; 183 184 /* 185 * XXX 186 * Because of the way that fts(3) works, logical walks will not count 187 * the blocks actually used by symbolic links. We rationalize this by 188 * noting that users computing logical sizes are likely to do logical 189 * copies, so not counting the links is correct. The real reason is 190 * that we'd have to re-implement the kernel's symbolic link traversing 191 * algorithm to get this right. If, for example, you have relative 192 * symbolic links referencing other relative symbolic links, it gets 193 * very nasty, very fast. The bottom line is that it's documented in 194 * the man page, so it's a feature. 195 */ 196 197 if (Hflag + Lflag + Pflag > 1) 198 usage(); 199 200 if (Hflag + Lflag + Pflag == 0) 201 Pflag = 1; /* -P (physical) is default */ 202 203 if (Hflag) 204 ftsoptions |= FTS_COMFOLLOW; 205 206 if (Lflag) 207 ftsoptions |= FTS_LOGICAL; 208 209 if (Pflag) 210 ftsoptions |= FTS_PHYSICAL; 211 212 listall = 0; 213 214 if (aflag) { 215 if (sflag || dflag) 216 usage(); 217 listall = 1; 218 } else if (sflag) { 219 if (dflag) 220 usage(); 221 depth = 0; 222 } 223 224 if (!*argv) { 225 argv = save; 226 argv[0] = dot; 227 argv[1] = NULL; 228 } 229 230 (void) getbsize(¬used, &blocksize); 231 blocksize /= 512; 232 233 rval = 0; 234 235 if ((fts = fts_open(argv, ftsoptions, NULL)) == NULL) 236 err(1, "fts_open"); 237 238 while ((p = fts_read(fts)) != NULL) { 239 switch (p->fts_info) { 240 case FTS_D: /* Ignore. */ 241 if (ignorep(p)) 242 fts_set(fts, p, FTS_SKIP); 243 break; 244 case FTS_DP: 245 if (ignorep(p)) 246 break; 247 248 p->fts_parent->fts_number += 249 p->fts_number += p->fts_statp->st_blocks; 250 251 if (p->fts_level <= depth) { 252 if (hflag) { 253 (void) prthumanval(howmany(p->fts_number, blocksize)); 254 (void) printf("\t%s\n", p->fts_path); 255 } else { 256 (void) printf("%ld\t%s\n", 257 howmany(p->fts_number, blocksize), 258 p->fts_path); 259 } 260 } 261 break; 262 case FTS_DC: /* Ignore. */ 263 break; 264 case FTS_DNR: /* Warn, continue. */ 265 case FTS_ERR: 266 case FTS_NS: 267 warnx("%s: %s", p->fts_path, strerror(p->fts_errno)); 268 rval = 1; 269 break; 270 default: 271 if (ignorep(p)) 272 break; 273 274 if (p->fts_statp->st_nlink > 1 && linkchk(p)) 275 break; 276 277 if (listall || p->fts_level == 0) { 278 if (hflag) { 279 (void) prthumanval(howmany(p->fts_statp->st_blocks, 280 blocksize)); 281 (void) printf("\t%s\n", p->fts_path); 282 } else { 283 (void) printf("%qd\t%s\n", 284 (long long)howmany(p->fts_statp->st_blocks, blocksize), 285 p->fts_path); 286 } 287 } 288 289 p->fts_parent->fts_number += p->fts_statp->st_blocks; 290 } 291 savednumber = p->fts_parent->fts_number; 292 } 293 294 if (errno) 295 err(1, "fts_read"); 296 297 if (cflag) { 298 if (hflag) { 299 (void) prthumanval(howmany(savednumber, blocksize)); 300 (void) printf("\ttotal\n"); 301 } else { 302 (void) printf("%ld\ttotal\n", howmany(savednumber, blocksize)); 303 } 304 } 305 306 ignoreclean(); 307 exit(rval); 308 } 309 310 static int 311 linkchk(FTSENT *p) 312 { 313 struct links_entry { 314 struct links_entry *next; 315 struct links_entry *previous; 316 int links; 317 dev_t dev; 318 ino_t ino; 319 }; 320 static const size_t links_hash_initial_size = 8192; 321 static struct links_entry **buckets; 322 static struct links_entry *free_list; 323 static size_t number_buckets; 324 static unsigned long number_entries; 325 static char stop_allocating; 326 struct links_entry *le, **new_buckets; 327 struct stat *st; 328 size_t i, new_size; 329 int count, hash; 330 331 st = p->fts_statp; 332 333 /* If necessary, initialize the hash table. */ 334 if (buckets == NULL) { 335 number_buckets = links_hash_initial_size; 336 buckets = malloc(number_buckets * sizeof(buckets[0])); 337 if (buckets == NULL) 338 errx(1, "No memory for hardlink detection"); 339 for (i = 0; i < number_buckets; i++) 340 buckets[i] = NULL; 341 } 342 343 /* If the hash table is getting too full, enlarge it. */ 344 if (number_entries > number_buckets * 10 && !stop_allocating) { 345 new_size = number_buckets * 2; 346 new_buckets = malloc(new_size * sizeof(struct links_entry *)); 347 count = 0; 348 349 /* Try releasing the free list to see if that helps. */ 350 if (new_buckets == NULL && free_list != NULL) { 351 while (free_list != NULL) { 352 le = free_list; 353 free_list = le->next; 354 free(le); 355 } 356 new_buckets = malloc(new_size * sizeof(new_buckets[0])); 357 } 358 359 if (new_buckets == NULL) { 360 stop_allocating = 1; 361 warnx("No more memory for tracking hard links"); 362 } else { 363 memset(new_buckets, 0, 364 new_size * sizeof(struct links_entry *)); 365 for (i = 0; i < number_buckets; i++) { 366 while (buckets[i] != NULL) { 367 /* Remove entry from old bucket. */ 368 le = buckets[i]; 369 buckets[i] = le->next; 370 371 /* Add entry to new bucket. */ 372 hash = (le->dev ^ le->ino) % new_size; 373 374 if (new_buckets[hash] != NULL) 375 new_buckets[hash]->previous = 376 le; 377 le->next = new_buckets[hash]; 378 le->previous = NULL; 379 new_buckets[hash] = le; 380 } 381 } 382 free(buckets); 383 buckets = new_buckets; 384 number_buckets = new_size; 385 } 386 } 387 388 /* Try to locate this entry in the hash table. */ 389 hash = ( st->st_dev ^ st->st_ino ) % number_buckets; 390 for (le = buckets[hash]; le != NULL; le = le->next) { 391 if (le->dev == st->st_dev && le->ino == st->st_ino) { 392 /* 393 * Save memory by releasing an entry when we've seen 394 * all of it's links. 395 */ 396 if (--le->links <= 0) { 397 if (le->previous != NULL) 398 le->previous->next = le->next; 399 if (le->next != NULL) 400 le->next->previous = le->previous; 401 if (buckets[hash] == le) 402 buckets[hash] = le->next; 403 number_entries--; 404 /* Recycle this node through the free list */ 405 if (stop_allocating) { 406 free(le); 407 } else { 408 le->next = free_list; 409 free_list = le; 410 } 411 } 412 return (1); 413 } 414 } 415 416 if (stop_allocating) 417 return (0); 418 419 /* Add this entry to the links cache. */ 420 if (free_list != NULL) { 421 /* Pull a node from the free list if we can. */ 422 le = free_list; 423 free_list = le->next; 424 } else 425 /* Malloc one if we have to. */ 426 le = malloc(sizeof(struct links_entry)); 427 if (le == NULL) { 428 stop_allocating = 1; 429 warnx("No more memory for tracking hard links"); 430 return (0); 431 } 432 le->dev = st->st_dev; 433 le->ino = st->st_ino; 434 le->links = st->st_nlink - 1; 435 number_entries++; 436 le->next = buckets[hash]; 437 le->previous = NULL; 438 if (buckets[hash] != NULL) 439 buckets[hash]->previous = le; 440 buckets[hash] = le; 441 return (0); 442 } 443 444 /* 445 * Output in "human-readable" format. Uses 3 digits max and puts 446 * unit suffixes at the end. Makes output compact and easy to read, 447 * especially on huge disks. 448 * 449 */ 450 unit_t 451 unit_adjust(double *val) 452 { 453 double abval; 454 unit_t unit; 455 unsigned int unit_sz; 456 457 abval = fabs(*val); 458 459 unit_sz = abval ? ilogb(abval) / 10 : 0; 460 461 if (unit_sz >= UNIT_MAX) { 462 unit = NONE; 463 } else { 464 unit = unitp[unit_sz]; 465 *val /= (double)valp[unit_sz]; 466 } 467 468 return (unit); 469 } 470 471 void 472 prthumanval(double bytes) 473 { 474 unit_t unit; 475 476 bytes *= 512; 477 unit = unit_adjust(&bytes); 478 479 if (bytes == 0) 480 (void)printf(" 0B"); 481 else if (bytes > 10) 482 (void)printf("%3.0f%c", bytes, "BKMGTPE"[unit]); 483 else 484 (void)printf("%3.1f%c", bytes, "BKMGTPE"[unit]); 485 } 486 487 static void 488 usage(void) 489 { 490 (void)fprintf(stderr, 491 "usage: du [-H | -L | -P] [-a | -s | -d depth] [-c] [-h | -k] [-x] [-I mask] [file ...]\n"); 492 exit(EX_USAGE); 493 } 494 495 void 496 ignoreadd(const char *mask) 497 { 498 struct ignentry *ign; 499 500 ign = calloc(1, sizeof(*ign)); 501 if (ign == NULL) 502 errx(1, "cannot allocate memory"); 503 ign->mask = strdup(mask); 504 if (ign->mask == NULL) 505 errx(1, "cannot allocate memory"); 506 SLIST_INSERT_HEAD(&ignores, ign, next); 507 } 508 509 void 510 ignoreclean(void) 511 { 512 struct ignentry *ign; 513 514 while (!SLIST_EMPTY(&ignores)) { 515 ign = SLIST_FIRST(&ignores); 516 SLIST_REMOVE_HEAD(&ignores, next); 517 free(ign->mask); 518 free(ign); 519 } 520 } 521 522 int 523 ignorep(FTSENT *ent) 524 { 525 struct ignentry *ign; 526 527 SLIST_FOREACH(ign, &ignores, next) 528 if (fnmatch(ign->mask, ent->fts_name, 0) != FNM_NOMATCH) 529 return 1; 530 return 0; 531 } 532