xref: /freebsd/bin/pax/ftree.c (revision 3c4ba5f55438f7afd4f4b0b56f88f2bb505fd6a6)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1992 Keith Muller.
5  * Copyright (c) 1992, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * This code is derived from software contributed to Berkeley by
9  * Keith Muller of the University of California, San Diego.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer in the
18  *    documentation and/or other materials provided with the distribution.
19  * 3. Neither the name of the University nor the names of its contributors
20  *    may be used to endorse or promote products derived from this software
21  *    without specific prior written permission.
22  *
23  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33  * SUCH DAMAGE.
34  */
35 
36 #ifndef lint
37 #if 0
38 static char sccsid[] = "@(#)ftree.c	8.2 (Berkeley) 4/18/94";
39 #endif
40 #endif /* not lint */
41 #include <sys/cdefs.h>
42 __FBSDID("$FreeBSD$");
43 
44 #include <sys/types.h>
45 #include <sys/time.h>
46 #include <sys/stat.h>
47 #include <unistd.h>
48 #include <string.h>
49 #include <stdio.h>
50 #include <errno.h>
51 #include <stdlib.h>
52 #include <fts.h>
53 #include "pax.h"
54 #include "ftree.h"
55 #include "extern.h"
56 
57 /*
58  * routines to interface with the fts library function.
59  *
60  * file args supplied to pax are stored on a single linked list (of type FTREE)
61  * and given to fts to be processed one at a time. pax "selects" files from
62  * the expansion of each arg into the corresponding file tree (if the arg is a
63  * directory, otherwise the node itself is just passed to pax). The selection
64  * is modified by the -n and -u flags. The user is informed when a specific
65  * file arg does not generate any selected files. -n keeps expanding the file
66  * tree arg until one of its files is selected, then skips to the next file
67  * arg. when the user does not supply the file trees as command line args to
68  * pax, they are read from stdin
69  */
70 
71 static FTS *ftsp = NULL;		/* current FTS handle */
72 static int ftsopts;			/* options to be used on fts_open */
73 static char *farray[2];			/* array for passing each arg to fts */
74 static FTREE *fthead = NULL;		/* head of linked list of file args */
75 static FTREE *fttail = NULL;		/* tail of linked list of file args */
76 static FTREE *ftcur = NULL;		/* current file arg being processed */
77 static FTSENT *ftent = NULL;		/* current file tree entry */
78 static int ftree_skip;			/* when set skip to next file arg */
79 
80 static int ftree_arg(void);
81 
82 /*
83  * ftree_start()
84  *	initialize the options passed to fts_open() during this run of pax
85  *	options are based on the selection of pax options by the user
86  *	fts_start() also calls fts_arg() to open the first valid file arg. We
87  *	also attempt to reset directory access times when -t (tflag) is set.
88  * Return:
89  *	0 if there is at least one valid file arg to process, -1 otherwise
90  */
91 
92 int
93 ftree_start(void)
94 {
95 	/*
96 	 * Set up the operation mode of fts, open the first file arg. We must
97 	 * use FTS_NOCHDIR, as the user may have to open multiple archives and
98 	 * if fts did a chdir off into the boondocks, we may create an archive
99 	 * volume in a place where the user did not expect to.
100 	 */
101 	ftsopts = FTS_NOCHDIR;
102 
103 	/*
104 	 * optional user flags that effect file traversal
105 	 * -H command line symlink follow only (half follow)
106 	 * -L follow symlinks (logical)
107 	 * -P do not follow symlinks (physical). This is the default.
108 	 * -X do not cross over mount points
109 	 * -t preserve access times on files read.
110 	 * -n select only the first member of a file tree when a match is found
111 	 * -d do not extract subtrees rooted at a directory arg.
112 	 */
113 	if (Lflag)
114 		ftsopts |= FTS_LOGICAL;
115 	else
116 		ftsopts |= FTS_PHYSICAL;
117 	if (Hflag)
118 		ftsopts |= FTS_COMFOLLOW;
119 	if (Xflag)
120 		ftsopts |= FTS_XDEV;
121 
122 	if ((fthead == NULL) && ((farray[0] = malloc(PAXPATHLEN+2)) == NULL)) {
123 		paxwarn(1, "Unable to allocate memory for file name buffer");
124 		return(-1);
125 	}
126 
127 	if (ftree_arg() < 0)
128 		return(-1);
129 	if (tflag && (atdir_start() < 0))
130 		return(-1);
131 	return(0);
132 }
133 
134 /*
135  * ftree_add()
136  *	add the arg to the linked list of files to process. Each will be
137  *	processed by fts one at a time
138  * Return:
139  *	0 if added to the linked list, -1 if failed
140  */
141 
142 int
143 ftree_add(char *str, int chflg)
144 {
145 	FTREE *ft;
146 	int len;
147 
148 	/*
149 	 * simple check for bad args
150 	 */
151 	if ((str == NULL) || (*str == '\0')) {
152 		paxwarn(0, "Invalid file name argument");
153 		return(-1);
154 	}
155 
156 	/*
157 	 * allocate FTREE node and add to the end of the linked list (args are
158 	 * processed in the same order they were passed to pax). Get rid of any
159 	 * trailing / the user may pass us. (watch out for / by itself).
160 	 */
161 	if ((ft = (FTREE *)malloc(sizeof(FTREE))) == NULL) {
162 		paxwarn(0, "Unable to allocate memory for filename");
163 		return(-1);
164 	}
165 
166 	if (((len = strlen(str) - 1) > 0) && (str[len] == '/'))
167 		str[len] = '\0';
168 	ft->fname = str;
169 	ft->refcnt = 0;
170 	ft->chflg = chflg;
171 	ft->fow = NULL;
172 	if (fthead == NULL) {
173 		fttail = fthead = ft;
174 		return(0);
175 	}
176 	fttail->fow = ft;
177 	fttail = ft;
178 	return(0);
179 }
180 
181 /*
182  * ftree_sel()
183  *	this entry has been selected by pax. bump up reference count and handle
184  *	-n and -d processing.
185  */
186 
187 void
188 ftree_sel(ARCHD *arcn)
189 {
190 	/*
191 	 * set reference bit for this pattern. This linked list is only used
192 	 * when file trees are supplied pax as args. The list is not used when
193 	 * the trees are read from stdin.
194 	 */
195 	if (ftcur != NULL)
196 		ftcur->refcnt = 1;
197 
198 	/*
199 	 * if -n we are done with this arg, force a skip to the next arg when
200 	 * pax asks for the next file in next_file().
201 	 * if -d we tell fts only to match the directory (if the arg is a dir)
202 	 * and not the entire file tree rooted at that point.
203 	 */
204 	if (nflag)
205 		ftree_skip = 1;
206 
207 	if (!dflag || (arcn->type != PAX_DIR))
208 		return;
209 
210 	if (ftent != NULL)
211 		(void)fts_set(ftsp, ftent, FTS_SKIP);
212 }
213 
214 /*
215  * ftree_notsel()
216  *	this entry has not been selected by pax.
217  */
218 
219 void
220 ftree_notsel(void)
221 {
222 	if (ftent != NULL)
223 		(void)fts_set(ftsp, ftent, FTS_SKIP);
224 }
225 
226 /*
227  * ftree_chk()
228  *	called at end on pax execution. Prints all those file args that did not
229  *	have a selected member (reference count still 0)
230  */
231 
232 void
233 ftree_chk(void)
234 {
235 	FTREE *ft;
236 	int wban = 0;
237 
238 	/*
239 	 * make sure all dir access times were reset.
240 	 */
241 	if (tflag)
242 		atdir_end();
243 
244 	/*
245 	 * walk down list and check reference count. Print out those members
246 	 * that never had a match
247 	 */
248 	for (ft = fthead; ft != NULL; ft = ft->fow) {
249 		if ((ft->refcnt > 0) || ft->chflg)
250 			continue;
251 		if (wban == 0) {
252 			paxwarn(1,"WARNING! These file names were not selected:");
253 			++wban;
254 		}
255 		(void)fprintf(stderr, "%s\n", ft->fname);
256 	}
257 }
258 
259 /*
260  * ftree_arg()
261  *	Get the next file arg for fts to process. Can be from either the linked
262  *	list or read from stdin when the user did not them as args to pax. Each
263  *	arg is processed until the first successful fts_open().
264  * Return:
265  *	0 when the next arg is ready to go, -1 if out of file args (or EOF on
266  *	stdin).
267  */
268 
269 static int
270 ftree_arg(void)
271 {
272 	char *pt;
273 
274 	/*
275 	 * close off the current file tree
276 	 */
277 	if (ftsp != NULL) {
278 		(void)fts_close(ftsp);
279 		ftsp = NULL;
280 	}
281 
282 	/*
283 	 * keep looping until we get a valid file tree to process. Stop when we
284 	 * reach the end of the list (or get an eof on stdin)
285 	 */
286 	for(;;) {
287 		if (fthead == NULL) {
288 			/*
289 			 * the user didn't supply any args, get the file trees
290 			 * to process from stdin;
291 			 */
292 			if (fgets(farray[0], PAXPATHLEN+1, stdin) == NULL)
293 				return(-1);
294 			if ((pt = strchr(farray[0], '\n')) != NULL)
295 				*pt = '\0';
296 		} else {
297 			/*
298 			 * the user supplied the file args as arguments to pax
299 			 */
300 			if (ftcur == NULL)
301 				ftcur = fthead;
302 			else if ((ftcur = ftcur->fow) == NULL)
303 				return(-1);
304 			if (ftcur->chflg) {
305 				/* First fchdir() back... */
306 				if (fchdir(cwdfd) < 0) {
307 					syswarn(1, errno,
308 					  "Can't fchdir to starting directory");
309 					return(-1);
310 				}
311 				if (chdir(ftcur->fname) < 0) {
312 					syswarn(1, errno, "Can't chdir to %s",
313 					    ftcur->fname);
314 					return(-1);
315 				}
316 				continue;
317 			} else
318 				farray[0] = ftcur->fname;
319 		}
320 
321 		/*
322 		 * Watch it, fts wants the file arg stored in an array of char
323 		 * ptrs, with the last one a null. We use a two element array
324 		 * and set farray[0] to point at the buffer with the file name
325 		 * in it. We cannot pass all the file args to fts at one shot
326 		 * as we need to keep a handle on which file arg generates what
327 		 * files (the -n and -d flags need this). If the open is
328 		 * successful, return a 0.
329 		 */
330 		if ((ftsp = fts_open(farray, ftsopts, NULL)) != NULL)
331 			break;
332 	}
333 	return(0);
334 }
335 
336 /*
337  * next_file()
338  *	supplies the next file to process in the supplied archd structure.
339  * Return:
340  *	0 when contents of arcn have been set with the next file, -1 when done.
341  */
342 
343 int
344 next_file(ARCHD *arcn)
345 {
346 	int cnt;
347 	time_t atime;
348 	time_t mtime;
349 
350 	/*
351 	 * ftree_sel() might have set the ftree_skip flag if the user has the
352 	 * -n option and a file was selected from this file arg tree. (-n says
353 	 * only one member is matched for each pattern) ftree_skip being 1
354 	 * forces us to go to the next arg now.
355 	 */
356 	if (ftree_skip) {
357 		/*
358 		 * clear and go to next arg
359 		 */
360 		ftree_skip = 0;
361 		if (ftree_arg() < 0)
362 			return(-1);
363 	}
364 
365 	/*
366 	 * loop until we get a valid file to process
367 	 */
368 	for(;;) {
369 		if ((ftent = fts_read(ftsp)) == NULL) {
370 			/*
371 			 * out of files in this tree, go to next arg, if none
372 			 * we are done
373 			 */
374 			if (ftree_arg() < 0)
375 				return(-1);
376 			continue;
377 		}
378 
379 		/*
380 		 * handle each type of fts_read() flag
381 		 */
382 		switch(ftent->fts_info) {
383 		case FTS_D:
384 		case FTS_DEFAULT:
385 		case FTS_F:
386 		case FTS_SL:
387 		case FTS_SLNONE:
388 			/*
389 			 * these are all ok
390 			 */
391 			break;
392 		case FTS_DP:
393 			/*
394 			 * already saw this directory. If the user wants file
395 			 * access times reset, we use this to restore the
396 			 * access time for this directory since this is the
397 			 * last time we will see it in this file subtree
398 			 * remember to force the time (this is -t on a read
399 			 * directory, not a created directory).
400 			 */
401 			if (!tflag || (get_atdir(ftent->fts_statp->st_dev,
402 			    ftent->fts_statp->st_ino, &mtime, &atime) < 0))
403 				continue;
404 			set_ftime(ftent->fts_path, mtime, atime, 1);
405 			continue;
406 		case FTS_DC:
407 			/*
408 			 * fts claims a file system cycle
409 			 */
410 			paxwarn(1,"File system cycle found at %s",ftent->fts_path);
411 			continue;
412 		case FTS_DNR:
413 			syswarn(1, ftent->fts_errno,
414 			    "Unable to read directory %s", ftent->fts_path);
415 			continue;
416 		case FTS_ERR:
417 			syswarn(1, ftent->fts_errno,
418 			    "File system traversal error");
419 			continue;
420 		case FTS_NS:
421 		case FTS_NSOK:
422 			syswarn(1, ftent->fts_errno,
423 			    "Unable to access %s", ftent->fts_path);
424 			continue;
425 		}
426 
427 		/*
428 		 * ok got a file tree node to process. copy info into arcn
429 		 * structure (initialize as required)
430 		 */
431 		arcn->skip = 0;
432 		arcn->pad = 0;
433 		arcn->ln_nlen = 0;
434 		arcn->ln_name[0] = '\0';
435 		arcn->sb = *(ftent->fts_statp);
436 
437 		/*
438 		 * file type based set up and copy into the arcn struct
439 		 * SIDE NOTE:
440 		 * we try to reset the access time on all files and directories
441 		 * we may read when the -t flag is specified. files are reset
442 		 * when we close them after copying. we reset the directories
443 		 * when we are done with their file tree (we also clean up at
444 		 * end in case we cut short a file tree traversal). However
445 		 * there is no way to reset access times on symlinks.
446 		 */
447 		switch(S_IFMT & arcn->sb.st_mode) {
448 		case S_IFDIR:
449 			arcn->type = PAX_DIR;
450 			if (!tflag)
451 				break;
452 			add_atdir(ftent->fts_path, arcn->sb.st_dev,
453 			    arcn->sb.st_ino, arcn->sb.st_mtime,
454 			    arcn->sb.st_atime);
455 			break;
456 		case S_IFCHR:
457 			arcn->type = PAX_CHR;
458 			break;
459 		case S_IFBLK:
460 			arcn->type = PAX_BLK;
461 			break;
462 		case S_IFREG:
463 			/*
464 			 * only regular files with have data to store on the
465 			 * archive. all others will store a zero length skip.
466 			 * the skip field is used by pax for actual data it has
467 			 * to read (or skip over).
468 			 */
469 			arcn->type = PAX_REG;
470 			arcn->skip = arcn->sb.st_size;
471 			break;
472 		case S_IFLNK:
473 			arcn->type = PAX_SLK;
474 			/*
475 			 * have to read the symlink path from the file
476 			 */
477 			if ((cnt = readlink(ftent->fts_path, arcn->ln_name,
478 			    PAXPATHLEN - 1)) < 0) {
479 				syswarn(1, errno, "Unable to read symlink %s",
480 				    ftent->fts_path);
481 				continue;
482 			}
483 			/*
484 			 * set link name length, watch out readlink does not
485 			 * always NUL terminate the link path
486 			 */
487 			arcn->ln_name[cnt] = '\0';
488 			arcn->ln_nlen = cnt;
489 			break;
490 		case S_IFSOCK:
491 			/*
492 			 * under BSD storing a socket is senseless but we will
493 			 * let the format specific write function make the
494 			 * decision of what to do with it.
495 			 */
496 			arcn->type = PAX_SCK;
497 			break;
498 		case S_IFIFO:
499 			arcn->type = PAX_FIF;
500 			break;
501 		}
502 		break;
503 	}
504 
505 	/*
506 	 * copy file name, set file name length
507 	 */
508 	arcn->nlen = l_strncpy(arcn->name, ftent->fts_path, sizeof(arcn->name) - 1);
509 	arcn->name[arcn->nlen] = '\0';
510 	arcn->org_name = ftent->fts_path;
511 	return(0);
512 }
513