xref: /freebsd/bin/pax/sel_subs.c (revision 2008043f386721d58158e37e0d7e50df8095942d)
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[] = "@(#)sel_subs.c	8.1 (Berkeley) 5/31/93";
39 #endif
40 #endif /* not lint */
41 #include <sys/cdefs.h>
42 #include <sys/types.h>
43 #include <sys/time.h>
44 #include <sys/stat.h>
45 
46 #include <ctype.h>
47 #include <grp.h>
48 #include <pwd.h>
49 #include <stdio.h>
50 #include <stdlib.h>
51 #include <string.h>
52 #include <strings.h>
53 
54 #include "pax.h"
55 #include "sel_subs.h"
56 #include "extern.h"
57 
58 static int str_sec(const char *, time_t *);
59 static int usr_match(ARCHD *);
60 static int grp_match(ARCHD *);
61 static int trng_match(ARCHD *);
62 
63 static TIME_RNG *trhead = NULL;		/* time range list head */
64 static TIME_RNG *trtail = NULL;		/* time range list tail */
65 static USRT **usrtb = NULL;		/* user selection table */
66 static GRPT **grptb = NULL;		/* group selection table */
67 
68 /*
69  * Routines for selection of archive members
70  */
71 
72 /*
73  * sel_chk()
74  *	check if this file matches a specified uid, gid or time range
75  * Return:
76  *	0 if this archive member should be processed, 1 if it should be skipped
77  */
78 
79 int
80 sel_chk(ARCHD *arcn)
81 {
82 	if (((usrtb != NULL) && usr_match(arcn)) ||
83 	    ((grptb != NULL) && grp_match(arcn)) ||
84 	    ((trhead != NULL) && trng_match(arcn)))
85 		return(1);
86 	return(0);
87 }
88 
89 /*
90  * User/group selection routines
91  *
92  * Routines to handle user selection of files based on the file uid/gid. To
93  * add an entry, the user supplies either the name or the uid/gid starting with
94  * a # on the command line. A \# will escape the #.
95  */
96 
97 /*
98  * usr_add()
99  *	add a user match to the user match hash table
100  * Return:
101  *	0 if added ok, -1 otherwise;
102  */
103 
104 int
105 usr_add(char *str)
106 {
107 	u_int indx;
108 	USRT *pt;
109 	struct passwd *pw;
110 	uid_t uid;
111 
112 	/*
113 	 * create the table if it doesn't exist
114 	 */
115 	if ((str == NULL) || (*str == '\0'))
116 		return(-1);
117 	if ((usrtb == NULL) &&
118  	    ((usrtb = (USRT **)calloc(USR_TB_SZ, sizeof(USRT *))) == NULL)) {
119 		paxwarn(1, "Unable to allocate memory for user selection table");
120 		return(-1);
121 	}
122 
123 	/*
124 	 * figure out user spec
125 	 */
126 	if (str[0] != '#') {
127 		/*
128 		 * it is a user name, \# escapes # as first char in user name
129 		 */
130 		if ((str[0] == '\\') && (str[1] == '#'))
131 			++str;
132 		if ((pw = getpwnam(str)) == NULL) {
133 			paxwarn(1, "Unable to find uid for user: %s", str);
134 			return(-1);
135 		}
136 		uid = (uid_t)pw->pw_uid;
137 	} else
138 		uid = (uid_t)strtoul(str+1, NULL, 10);
139 	endpwent();
140 
141 	/*
142 	 * hash it and go down the hash chain (if any) looking for it
143 	 */
144 	indx = ((unsigned)uid) % USR_TB_SZ;
145 	if ((pt = usrtb[indx]) != NULL) {
146 		while (pt != NULL) {
147 			if (pt->uid == uid)
148 				return(0);
149 			pt = pt->fow;
150 		}
151 	}
152 
153 	/*
154 	 * uid is not yet in the table, add it to the front of the chain
155 	 */
156 	if ((pt = (USRT *)malloc(sizeof(USRT))) != NULL) {
157 		pt->uid = uid;
158 		pt->fow = usrtb[indx];
159 		usrtb[indx] = pt;
160 		return(0);
161 	}
162 	paxwarn(1, "User selection table out of memory");
163 	return(-1);
164 }
165 
166 /*
167  * usr_match()
168  *	check if this files uid matches a selected uid.
169  * Return:
170  *	0 if this archive member should be processed, 1 if it should be skipped
171  */
172 
173 static int
174 usr_match(ARCHD *arcn)
175 {
176 	USRT *pt;
177 
178 	/*
179 	 * hash and look for it in the table
180 	 */
181 	pt = usrtb[((unsigned)arcn->sb.st_uid) % USR_TB_SZ];
182 	while (pt != NULL) {
183 		if (pt->uid == arcn->sb.st_uid)
184 			return(0);
185 		pt = pt->fow;
186 	}
187 
188 	/*
189 	 * not found
190 	 */
191 	return(1);
192 }
193 
194 /*
195  * grp_add()
196  *	add a group match to the group match hash table
197  * Return:
198  *	0 if added ok, -1 otherwise;
199  */
200 
201 int
202 grp_add(char *str)
203 {
204 	u_int indx;
205 	GRPT *pt;
206 	struct group *gr;
207 	gid_t gid;
208 
209 	/*
210 	 * create the table if it doesn't exist
211 	 */
212 	if ((str == NULL) || (*str == '\0'))
213 		return(-1);
214 	if ((grptb == NULL) &&
215  	    ((grptb = (GRPT **)calloc(GRP_TB_SZ, sizeof(GRPT *))) == NULL)) {
216 		paxwarn(1, "Unable to allocate memory fo group selection table");
217 		return(-1);
218 	}
219 
220 	/*
221 	 * figure out user spec
222 	 */
223 	if (str[0] != '#') {
224 		/*
225 		 * it is a group name, \# escapes # as first char in group name
226 		 */
227 		if ((str[0] == '\\') && (str[1] == '#'))
228 			++str;
229 		if ((gr = getgrnam(str)) == NULL) {
230 			paxwarn(1,"Cannot determine gid for group name: %s", str);
231 			return(-1);
232 		}
233 		gid = gr->gr_gid;
234 	} else
235 		gid = (gid_t)strtoul(str+1, NULL, 10);
236 	endgrent();
237 
238 	/*
239 	 * hash it and go down the hash chain (if any) looking for it
240 	 */
241 	indx = ((unsigned)gid) % GRP_TB_SZ;
242 	if ((pt = grptb[indx]) != NULL) {
243 		while (pt != NULL) {
244 			if (pt->gid == gid)
245 				return(0);
246 			pt = pt->fow;
247 		}
248 	}
249 
250 	/*
251 	 * gid not in the table, add it to the front of the chain
252 	 */
253 	if ((pt = (GRPT *)malloc(sizeof(GRPT))) != NULL) {
254 		pt->gid = gid;
255 		pt->fow = grptb[indx];
256 		grptb[indx] = pt;
257 		return(0);
258 	}
259 	paxwarn(1, "Group selection table out of memory");
260 	return(-1);
261 }
262 
263 /*
264  * grp_match()
265  *	check if this files gid matches a selected gid.
266  * Return:
267  *	0 if this archive member should be processed, 1 if it should be skipped
268  */
269 
270 static int
271 grp_match(ARCHD *arcn)
272 {
273 	GRPT *pt;
274 
275 	/*
276 	 * hash and look for it in the table
277 	 */
278 	pt = grptb[((unsigned)arcn->sb.st_gid) % GRP_TB_SZ];
279 	while (pt != NULL) {
280 		if (pt->gid == arcn->sb.st_gid)
281 			return(0);
282 		pt = pt->fow;
283 	}
284 
285 	/*
286 	 * not found
287 	 */
288 	return(1);
289 }
290 
291 /*
292  * Time range selection routines
293  *
294  * Routines to handle user selection of files based on the modification and/or
295  * inode change time falling within a specified time range (the non-standard
296  * -T flag). The user may specify any number of different file time ranges.
297  * Time ranges are checked one at a time until a match is found (if at all).
298  * If the file has a mtime (and/or ctime) which lies within one of the time
299  * ranges, the file is selected. Time ranges may have a lower and/or an upper
300  * value. These ranges are inclusive. When no time ranges are supplied to pax
301  * with the -T option, all members in the archive will be selected by the time
302  * range routines. When only a lower range is supplied, only files with a
303  * mtime (and/or ctime) equal to or younger are selected. When only an upper
304  * range is supplied, only files with a mtime (and/or ctime) equal to or older
305  * are selected. When the lower time range is equal to the upper time range,
306  * only files with a mtime (or ctime) of exactly that time are selected.
307  */
308 
309 /*
310  * trng_add()
311  *	add a time range match to the time range list.
312  *	This is a non-standard pax option. Lower and upper ranges are in the
313  *	format: [[[[[cc]yy]mm]dd]HH]MM[.SS] and are comma separated.
314  *	Time ranges are based on current time, so 1234 would specify a time of
315  *	12:34 today.
316  * Return:
317  *	0 if the time range was added to the list, -1 otherwise
318  */
319 
320 int
321 trng_add(char *str)
322 {
323 	TIME_RNG *pt;
324 	char *up_pt = NULL;
325 	char *stpt;
326 	char *flgpt;
327 	int dot = 0;
328 
329 	/*
330 	 * throw out the badly formed time ranges
331 	 */
332 	if ((str == NULL) || (*str == '\0')) {
333 		paxwarn(1, "Empty time range string");
334 		return(-1);
335 	}
336 
337 	/*
338 	 * locate optional flags suffix /{cm}.
339 	 */
340 	if ((flgpt = strrchr(str, '/')) != NULL)
341 		*flgpt++ = '\0';
342 
343 	for (stpt = str; *stpt != '\0'; ++stpt) {
344 		if ((*stpt >= '0') && (*stpt <= '9'))
345 			continue;
346 		if ((*stpt == ',') && (up_pt == NULL)) {
347 			*stpt = '\0';
348 			up_pt = stpt + 1;
349 			dot = 0;
350 			continue;
351 		}
352 
353 		/*
354 		 * allow only one dot per range (secs)
355 		 */
356 		if ((*stpt == '.') && (!dot)) {
357 			++dot;
358 			continue;
359 		}
360 		paxwarn(1, "Improperly specified time range: %s", str);
361 		goto out;
362 	}
363 
364 	/*
365 	 * allocate space for the time range and store the limits
366 	 */
367 	if ((pt = (TIME_RNG *)malloc(sizeof(TIME_RNG))) == NULL) {
368 		paxwarn(1, "Unable to allocate memory for time range");
369 		return(-1);
370 	}
371 
372 	/*
373 	 * by default we only will check file mtime, but the user can specify
374 	 * mtime, ctime (inode change time) or both.
375 	 */
376 	if ((flgpt == NULL) || (*flgpt == '\0'))
377 		pt->flgs = CMPMTME;
378 	else {
379 		pt->flgs = 0;
380 		while (*flgpt != '\0') {
381 			switch(*flgpt) {
382 			case 'M':
383 			case 'm':
384 				pt->flgs |= CMPMTME;
385 				break;
386 			case 'C':
387 			case 'c':
388 				pt->flgs |= CMPCTME;
389 				break;
390 			default:
391 				paxwarn(1, "Bad option %c with time range %s",
392 				    *flgpt, str);
393 				free(pt);
394 				goto out;
395 			}
396 			++flgpt;
397 		}
398 	}
399 
400 	/*
401 	 * start off with the current time
402 	 */
403 	pt->low_time = pt->high_time = time(NULL);
404 	if (*str != '\0') {
405 		/*
406 		 * add lower limit
407 		 */
408 		if (str_sec(str, &(pt->low_time)) < 0) {
409 			paxwarn(1, "Illegal lower time range %s", str);
410 			free(pt);
411 			goto out;
412 		}
413 		pt->flgs |= HASLOW;
414 	}
415 
416 	if ((up_pt != NULL) && (*up_pt != '\0')) {
417 		/*
418 		 * add upper limit
419 		 */
420 		if (str_sec(up_pt, &(pt->high_time)) < 0) {
421 			paxwarn(1, "Illegal upper time range %s", up_pt);
422 			free(pt);
423 			goto out;
424 		}
425 		pt->flgs |= HASHIGH;
426 
427 		/*
428 		 * check that the upper and lower do not overlap
429 		 */
430 		if (pt->flgs & HASLOW) {
431 			if (pt->low_time > pt->high_time) {
432 				paxwarn(1, "Upper %s and lower %s time overlap",
433 					up_pt, str);
434 				free(pt);
435 				return(-1);
436 			}
437 		}
438 	}
439 
440 	pt->fow = NULL;
441 	if (trhead == NULL) {
442 		trtail = trhead = pt;
443 		return(0);
444 	}
445 	trtail->fow = pt;
446 	trtail = pt;
447 	return(0);
448 
449     out:
450 	paxwarn(1, "Time range format is: [[[[[cc]yy]mm]dd]HH]MM[.SS][/[c][m]]");
451 	return(-1);
452 }
453 
454 /*
455  * trng_match()
456  *	check if this files mtime/ctime falls within any supplied time range.
457  * Return:
458  *	0 if this archive member should be processed, 1 if it should be skipped
459  */
460 
461 static int
462 trng_match(ARCHD *arcn)
463 {
464 	TIME_RNG *pt;
465 
466 	/*
467 	 * have to search down the list one at a time looking for a match.
468 	 * remember time range limits are inclusive.
469 	 */
470 	pt = trhead;
471 	while (pt != NULL) {
472 		switch(pt->flgs & CMPBOTH) {
473 		case CMPBOTH:
474 			/*
475 			 * user wants both mtime and ctime checked for this
476 			 * time range
477 			 */
478 			if (((pt->flgs & HASLOW) &&
479 			    (arcn->sb.st_mtime < pt->low_time) &&
480 			    (arcn->sb.st_ctime < pt->low_time)) ||
481 			    ((pt->flgs & HASHIGH) &&
482 			    (arcn->sb.st_mtime > pt->high_time) &&
483 			    (arcn->sb.st_ctime > pt->high_time))) {
484 				pt = pt->fow;
485 				continue;
486 			}
487 			break;
488 		case CMPCTME:
489 			/*
490 			 * user wants only ctime checked for this time range
491 			 */
492 			if (((pt->flgs & HASLOW) &&
493 			    (arcn->sb.st_ctime < pt->low_time)) ||
494 			    ((pt->flgs & HASHIGH) &&
495 			    (arcn->sb.st_ctime > pt->high_time))) {
496 				pt = pt->fow;
497 				continue;
498 			}
499 			break;
500 		case CMPMTME:
501 		default:
502 			/*
503 			 * user wants only mtime checked for this time range
504 			 */
505 			if (((pt->flgs & HASLOW) &&
506 			    (arcn->sb.st_mtime < pt->low_time)) ||
507 			    ((pt->flgs & HASHIGH) &&
508 			    (arcn->sb.st_mtime > pt->high_time))) {
509 				pt = pt->fow;
510 				continue;
511 			}
512 			break;
513 		}
514 		break;
515 	}
516 
517 	if (pt == NULL)
518 		return(1);
519 	return(0);
520 }
521 
522 /*
523  * str_sec()
524  *	Convert a time string in the format of [[[[[cc]yy]mm]dd]HH]MM[.SS] to
525  *	seconds UTC. Tval already has current time loaded into it at entry.
526  * Return:
527  *	0 if converted ok, -1 otherwise
528  */
529 
530 static int
531 str_sec(const char *p, time_t *tval)
532 {
533 	struct tm *lt;
534 	const char *dot, *t;
535 	size_t len;
536 	int bigyear;
537 	int yearset;
538 
539 	yearset = 0;
540 	len = strlen(p);
541 
542 	for (t = p, dot = NULL; *t; ++t) {
543 		if (isdigit((unsigned char)*t))
544 			continue;
545 		if (*t == '.' && dot == NULL) {
546 			dot = t;
547 			continue;
548 		}
549 		return(-1);
550 	}
551 
552 	lt = localtime(tval);
553 
554 	if (dot != NULL) {			/* .SS */
555 		if (strlen(++dot) != 2)
556 			return(-1);
557 		lt->tm_sec = ATOI2(dot);
558 		if (lt->tm_sec > 61)
559 			return(-1);
560 		len -= 3;
561 	} else
562 		lt->tm_sec = 0;
563 
564 	switch (len) {
565 	case 12:				/* cc */
566 		bigyear = ATOI2(p);
567 		lt->tm_year = (bigyear * 100) - 1900;
568 		yearset = 1;
569 		/* FALLTHROUGH */
570 	case 10:				/* yy */
571 		if (yearset) {
572 			lt->tm_year += ATOI2(p);
573 		} else {
574 			lt->tm_year = ATOI2(p);
575 			if (lt->tm_year < 69)		/* hack for 2000 ;-} */
576 				lt->tm_year += (2000 - 1900);
577 		}
578 		/* FALLTHROUGH */
579 	case 8:					/* mm */
580 		lt->tm_mon = ATOI2(p);
581 		if ((lt->tm_mon > 12) || !lt->tm_mon)
582 			return(-1);
583 		--lt->tm_mon;			/* time struct is 0 - 11 */
584 		/* FALLTHROUGH */
585 	case 6:					/* dd */
586 		lt->tm_mday = ATOI2(p);
587 		if ((lt->tm_mday > 31) || !lt->tm_mday)
588 			return(-1);
589 		/* FALLTHROUGH */
590 	case 4:					/* HH */
591 		lt->tm_hour = ATOI2(p);
592 		if (lt->tm_hour > 23)
593 			return(-1);
594 		/* FALLTHROUGH */
595 	case 2:					/* MM */
596 		lt->tm_min = ATOI2(p);
597 		if (lt->tm_min > 59)
598 			return(-1);
599 		break;
600 	default:
601 		return(-1);
602 	}
603 
604 	/* convert broken-down time to UTC clock time seconds */
605 	if ((*tval = mktime(lt)) == -1)
606 		return(-1);
607 	return(0);
608 }
609