xref: /netbsd-src/usr.bin/du/du.c (revision cac8e449158efc7261bebc8657cbb0125a2cfdde)
1 /*	$NetBSD: du.c,v 1.33 2008/07/30 22:03:40 dsl Exp $	*/
2 
3 /*
4  * Copyright (c) 1989, 1993, 1994
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Chris Newcomb.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34 
35 #include <sys/cdefs.h>
36 #ifndef lint
37 __COPYRIGHT("@(#) Copyright (c) 1989, 1993, 1994\
38  The Regents of the University of California.  All rights reserved.");
39 #endif /* not lint */
40 
41 #ifndef lint
42 #if 0
43 static char sccsid[] = "@(#)du.c	8.5 (Berkeley) 5/4/95";
44 #else
45 __RCSID("$NetBSD: du.c,v 1.33 2008/07/30 22:03:40 dsl Exp $");
46 #endif
47 #endif /* not lint */
48 
49 #include <sys/types.h>
50 #include <sys/stat.h>
51 
52 #include <dirent.h>
53 #include <err.h>
54 #include <errno.h>
55 #include <fts.h>
56 #include <util.h>
57 #include <stdio.h>
58 #include <stdlib.h>
59 #include <string.h>
60 #include <unistd.h>
61 #include <limits.h>
62 
63 int	linkchk(dev_t, ino_t);
64 void	prstat(const char *, int64_t);
65 void	usage(void);
66 
67 int hflag;
68 long blocksize;
69 
70 int
71 main(int argc, char *argv[])
72 {
73 	FTS *fts;
74 	FTSENT *p;
75 	int64_t totalblocks;
76 	int ftsoptions, listfiles;
77 	int depth;
78 	int Hflag, Lflag, aflag, ch, cflag, dflag, gkmflag, nflag, rval, sflag;
79 	const char *noargv[2];
80 
81 	Hflag = Lflag = aflag = cflag = dflag = gkmflag = nflag = sflag = 0;
82 	totalblocks = 0;
83 	ftsoptions = FTS_PHYSICAL;
84 	depth = INT_MAX;
85 	while ((ch = getopt(argc, argv, "HLPacd:ghkmnrsx")) != -1)
86 		switch (ch) {
87 		case 'H':
88 			Hflag = 1;
89 			Lflag = 0;
90 			break;
91 		case 'L':
92 			Lflag = 1;
93 			Hflag = 0;
94 			break;
95 		case 'P':
96 			Hflag = Lflag = 0;
97 			break;
98 		case 'a':
99 			aflag = 1;
100 			break;
101 		case 'c':
102 			cflag = 1;
103 			break;
104 		case 'd':
105 			dflag = 1;
106 			depth = atoi(optarg);
107 			if (depth < 0 || depth > SHRT_MAX) {
108 				warnx("invalid argument to option d: %s",
109 					optarg);
110 				usage();
111 			}
112 			break;
113 		case 'g':
114 			blocksize = 1024 * 1024 * 1024;
115 			gkmflag = 1;
116 			break;
117 		case 'h':
118 			hflag = 1;
119 			break;
120 		case 'k':
121 			blocksize = 1024;
122 			gkmflag = 1;
123 			break;
124 		case 'm':
125 			blocksize = 1024 * 1024;
126 			gkmflag = 1;
127 			break;
128 		case 'n':
129 			nflag = 1;
130 			break;
131 		case 'r':
132 			break;
133 		case 's':
134 			sflag = 1;
135 			break;
136 		case 'x':
137 			ftsoptions |= FTS_XDEV;
138 			break;
139 		case '?':
140 		default:
141 			usage();
142 		}
143 	argc -= optind;
144 	argv += optind;
145 
146 	/*
147 	 * XXX
148 	 * Because of the way that fts(3) works, logical walks will not count
149 	 * the blocks actually used by symbolic links.  We rationalize this by
150 	 * noting that users computing logical sizes are likely to do logical
151 	 * copies, so not counting the links is correct.  The real reason is
152 	 * that we'd have to re-implement the kernel's symbolic link traversing
153 	 * algorithm to get this right.  If, for example, you have relative
154 	 * symbolic links referencing other relative symbolic links, it gets
155 	 * very nasty, very fast.  The bottom line is that it's documented in
156 	 * the man page, so it's a feature.
157 	 */
158 	if (Hflag)
159 		ftsoptions |= FTS_COMFOLLOW;
160 	if (Lflag) {
161 		ftsoptions &= ~FTS_PHYSICAL;
162 		ftsoptions |= FTS_LOGICAL;
163 	}
164 
165 	listfiles = 0;
166 	if (aflag) {
167 		if (sflag || dflag)
168 			usage();
169 		listfiles = 1;
170 	} else if (sflag) {
171 		if (dflag)
172 			usage();
173 		depth = 0;
174 	}
175 
176 	if (!*argv) {
177 		noargv[0] = ".";
178 		noargv[1] = NULL;
179 		argv = __UNCONST(noargv);
180 	}
181 
182 	if (!gkmflag)
183 		(void)getbsize(NULL, &blocksize);
184 	blocksize /= 512;
185 
186 	if ((fts = fts_open(argv, ftsoptions, NULL)) == NULL)
187 		err(1, "fts_open `%s'", *argv);
188 
189 	for (rval = 0; (p = fts_read(fts)) != NULL;) {
190 		if (nflag) {
191 			switch (p->fts_info) {
192 			case FTS_NS:
193 			case FTS_SLNONE:
194 				/* nothing */
195 				break;
196 			default:
197 				if (p->fts_statp->st_flags & UF_NODUMP) {
198 					fts_set(fts, p, FTS_SKIP);
199 					continue;
200 				}
201 			}
202 		}
203 		switch (p->fts_info) {
204 		case FTS_D:			/* Ignore. */
205 			break;
206 		case FTS_DP:
207 			p->fts_parent->fts_number +=
208 			    p->fts_number += p->fts_statp->st_blocks;
209 			if (cflag)
210 				totalblocks += p->fts_statp->st_blocks;
211 			/*
212 			 * If listing each directory, or not listing files
213 			 * or directories and this is post-order of the
214 			 * root of a traversal, display the total.
215 			 */
216 			if (p->fts_level <= depth
217 			    || (!listfiles && !p->fts_level))
218 				prstat(p->fts_path, p->fts_number);
219 			break;
220 		case FTS_DC:			/* Ignore. */
221 			break;
222 		case FTS_DNR:			/* Warn, continue. */
223 		case FTS_ERR:
224 		case FTS_NS:
225 			warnx("%s: %s", p->fts_path, strerror(p->fts_errno));
226 			rval = 1;
227 			break;
228 		default:
229 			if (p->fts_statp->st_nlink > 1 &&
230 			    linkchk(p->fts_statp->st_dev, p->fts_statp->st_ino))
231 				break;
232 			/*
233 			 * If listing each file, or a non-directory file was
234 			 * the root of a traversal, display the total.
235 			 */
236 			if (listfiles || !p->fts_level)
237 				prstat(p->fts_path, p->fts_statp->st_blocks);
238 			p->fts_parent->fts_number += p->fts_statp->st_blocks;
239 			if (cflag)
240 				totalblocks += p->fts_statp->st_blocks;
241 		}
242 	}
243 	if (errno)
244 		err(1, "fts_read");
245 	if (cflag)
246 		prstat("total", totalblocks);
247 	exit(rval);
248 }
249 
250 void
251 prstat(const char *fname, int64_t blocks)
252 {
253 	if (hflag) {
254 		char buf[5];
255 		int64_t sz = blocks * 512;
256 
257 		humanize_number(buf, sizeof(buf), sz, "", HN_AUTOSCALE,
258 		    HN_B | HN_NOSPACE | HN_DECIMAL);
259 
260 		(void)printf("%s\t%s\n", buf, fname);
261 	} else
262 		(void)printf("%lld\t%s\n",
263 		    (long long)howmany(blocks, (int64_t)blocksize),
264 		    fname);
265 }
266 
267 int
268 linkchk(dev_t dev, ino_t ino)
269 {
270 	static struct entry {
271 		dev_t	dev;
272 		ino_t	ino;
273 	} *htable;
274 	static int htshift;  /* log(allocated size) */
275 	static int htmask;   /* allocated size - 1 */
276 	static int htused;   /* 2*number of insertions */
277 	static int sawzero;  /* Whether zero is in table or not */
278 	int h, h2;
279 	uint64_t tmp;
280 	/* this constant is (1<<64)/((1+sqrt(5))/2)
281 	 * aka (word size)/(golden ratio)
282 	 */
283 	const uint64_t HTCONST = 11400714819323198485ULL;
284 	const int HTBITS = CHAR_BIT * sizeof(tmp);
285 
286 	/* Never store zero in hashtable */
287 	if (dev == 0 && ino == 0) {
288 		h = sawzero;
289 		sawzero = 1;
290 		return h;
291 	}
292 
293 	/* Extend hash table if necessary, keep load under 0.5 */
294 	if (htused<<1 >= htmask) {
295 		struct entry *ohtable;
296 
297 		if (!htable)
298 			htshift = 10;   /* starting hashtable size */
299 		else
300 			htshift++;   /* exponential hashtable growth */
301 
302 		htmask  = (1 << htshift) - 1;
303 		htused = 0;
304 
305 		ohtable = htable;
306 		htable = calloc(htmask+1, sizeof(*htable));
307 		if (!htable)
308 			err(1, "calloc");
309 
310 		/* populate newly allocated hashtable */
311 		if (ohtable) {
312 			int i;
313 			for (i = 0; i <= htmask>>1; i++)
314 				if (ohtable[i].ino || ohtable[i].dev)
315 					linkchk(ohtable[i].dev, ohtable[i].ino);
316 			free(ohtable);
317 		}
318 	}
319 
320 	/* multiplicative hashing */
321 	tmp = dev;
322 	tmp <<= HTBITS>>1;
323 	tmp |=  ino;
324 	tmp *= HTCONST;
325 	h  = tmp >> (HTBITS - htshift);
326 	h2 = 1 | ( tmp >> (HTBITS - (htshift<<1) - 1)); /* must be odd */
327 
328 	/* open address hashtable search with double hash probing */
329 	while (htable[h].ino || htable[h].dev) {
330 		if ((htable[h].ino == ino) && (htable[h].dev == dev))
331 			return 1;
332 		h = (h + h2) & htmask;
333 	}
334 
335 	/* Insert the current entry into hashtable */
336 	htable[h].dev = dev;
337 	htable[h].ino = ino;
338 	htused++;
339 	return 0;
340 }
341 
342 void
343 usage(void)
344 {
345 
346 	(void)fprintf(stderr,
347 		"usage: du [-H | -L | -P] [-a | -d depth | -s] [-cghkmnrx] [file ...]\n");
348 	exit(1);
349 }
350