xref: /freebsd/bin/pax/sel_subs.c (revision 5b31cc94b10d4bb7109c6b27940a0fc76a44a331)
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 #endif /* not lint */
38 #include <sys/cdefs.h>
39 #include <sys/types.h>
40 #include <sys/time.h>
41 #include <sys/stat.h>
42 
43 #include <ctype.h>
44 #include <grp.h>
45 #include <pwd.h>
46 #include <stdio.h>
47 #include <stdlib.h>
48 #include <string.h>
49 #include <strings.h>
50 
51 #include "pax.h"
52 #include "sel_subs.h"
53 #include "extern.h"
54 
55 static int str_sec(const char *, time_t *);
56 static int usr_match(ARCHD *);
57 static int grp_match(ARCHD *);
58 static int trng_match(ARCHD *);
59 
60 static TIME_RNG *trhead = NULL;		/* time range list head */
61 static TIME_RNG *trtail = NULL;		/* time range list tail */
62 static USRT **usrtb = NULL;		/* user selection table */
63 static GRPT **grptb = NULL;		/* group selection table */
64 
65 /*
66  * Routines for selection of archive members
67  */
68 
69 /*
70  * sel_chk()
71  *	check if this file matches a specified uid, gid or time range
72  * Return:
73  *	0 if this archive member should be processed, 1 if it should be skipped
74  */
75 
76 int
77 sel_chk(ARCHD *arcn)
78 {
79 	if (((usrtb != NULL) && usr_match(arcn)) ||
80 	    ((grptb != NULL) && grp_match(arcn)) ||
81 	    ((trhead != NULL) && trng_match(arcn)))
82 		return(1);
83 	return(0);
84 }
85 
86 /*
87  * User/group selection routines
88  *
89  * Routines to handle user selection of files based on the file uid/gid. To
90  * add an entry, the user supplies either the name or the uid/gid starting with
91  * a # on the command line. A \# will escape the #.
92  */
93 
94 /*
95  * usr_add()
96  *	add a user match to the user match hash table
97  * Return:
98  *	0 if added ok, -1 otherwise;
99  */
100 
101 int
102 usr_add(char *str)
103 {
104 	u_int indx;
105 	USRT *pt;
106 	struct passwd *pw;
107 	uid_t uid;
108 
109 	/*
110 	 * create the table if it doesn't exist
111 	 */
112 	if ((str == NULL) || (*str == '\0'))
113 		return(-1);
114 	if ((usrtb == NULL) &&
115  	    ((usrtb = (USRT **)calloc(USR_TB_SZ, sizeof(USRT *))) == NULL)) {
116 		paxwarn(1, "Unable to allocate memory for user selection table");
117 		return(-1);
118 	}
119 
120 	/*
121 	 * figure out user spec
122 	 */
123 	if (str[0] != '#') {
124 		/*
125 		 * it is a user name, \# escapes # as first char in user name
126 		 */
127 		if ((str[0] == '\\') && (str[1] == '#'))
128 			++str;
129 		if ((pw = getpwnam(str)) == NULL) {
130 			paxwarn(1, "Unable to find uid for user: %s", str);
131 			return(-1);
132 		}
133 		uid = (uid_t)pw->pw_uid;
134 	} else
135 		uid = (uid_t)strtoul(str+1, NULL, 10);
136 	endpwent();
137 
138 	/*
139 	 * hash it and go down the hash chain (if any) looking for it
140 	 */
141 	indx = ((unsigned)uid) % USR_TB_SZ;
142 	if ((pt = usrtb[indx]) != NULL) {
143 		while (pt != NULL) {
144 			if (pt->uid == uid)
145 				return(0);
146 			pt = pt->fow;
147 		}
148 	}
149 
150 	/*
151 	 * uid is not yet in the table, add it to the front of the chain
152 	 */
153 	if ((pt = (USRT *)malloc(sizeof(USRT))) != NULL) {
154 		pt->uid = uid;
155 		pt->fow = usrtb[indx];
156 		usrtb[indx] = pt;
157 		return(0);
158 	}
159 	paxwarn(1, "User selection table out of memory");
160 	return(-1);
161 }
162 
163 /*
164  * usr_match()
165  *	check if this files uid matches a selected uid.
166  * Return:
167  *	0 if this archive member should be processed, 1 if it should be skipped
168  */
169 
170 static int
171 usr_match(ARCHD *arcn)
172 {
173 	USRT *pt;
174 
175 	/*
176 	 * hash and look for it in the table
177 	 */
178 	pt = usrtb[((unsigned)arcn->sb.st_uid) % USR_TB_SZ];
179 	while (pt != NULL) {
180 		if (pt->uid == arcn->sb.st_uid)
181 			return(0);
182 		pt = pt->fow;
183 	}
184 
185 	/*
186 	 * not found
187 	 */
188 	return(1);
189 }
190 
191 /*
192  * grp_add()
193  *	add a group match to the group match hash table
194  * Return:
195  *	0 if added ok, -1 otherwise;
196  */
197 
198 int
199 grp_add(char *str)
200 {
201 	u_int indx;
202 	GRPT *pt;
203 	struct group *gr;
204 	gid_t gid;
205 
206 	/*
207 	 * create the table if it doesn't exist
208 	 */
209 	if ((str == NULL) || (*str == '\0'))
210 		return(-1);
211 	if ((grptb == NULL) &&
212  	    ((grptb = (GRPT **)calloc(GRP_TB_SZ, sizeof(GRPT *))) == NULL)) {
213 		paxwarn(1, "Unable to allocate memory fo group selection table");
214 		return(-1);
215 	}
216 
217 	/*
218 	 * figure out user spec
219 	 */
220 	if (str[0] != '#') {
221 		/*
222 		 * it is a group name, \# escapes # as first char in group name
223 		 */
224 		if ((str[0] == '\\') && (str[1] == '#'))
225 			++str;
226 		if ((gr = getgrnam(str)) == NULL) {
227 			paxwarn(1,"Cannot determine gid for group name: %s", str);
228 			return(-1);
229 		}
230 		gid = gr->gr_gid;
231 	} else
232 		gid = (gid_t)strtoul(str+1, NULL, 10);
233 	endgrent();
234 
235 	/*
236 	 * hash it and go down the hash chain (if any) looking for it
237 	 */
238 	indx = ((unsigned)gid) % GRP_TB_SZ;
239 	if ((pt = grptb[indx]) != NULL) {
240 		while (pt != NULL) {
241 			if (pt->gid == gid)
242 				return(0);
243 			pt = pt->fow;
244 		}
245 	}
246 
247 	/*
248 	 * gid not in the table, add it to the front of the chain
249 	 */
250 	if ((pt = (GRPT *)malloc(sizeof(GRPT))) != NULL) {
251 		pt->gid = gid;
252 		pt->fow = grptb[indx];
253 		grptb[indx] = pt;
254 		return(0);
255 	}
256 	paxwarn(1, "Group selection table out of memory");
257 	return(-1);
258 }
259 
260 /*
261  * grp_match()
262  *	check if this files gid matches a selected gid.
263  * Return:
264  *	0 if this archive member should be processed, 1 if it should be skipped
265  */
266 
267 static int
268 grp_match(ARCHD *arcn)
269 {
270 	GRPT *pt;
271 
272 	/*
273 	 * hash and look for it in the table
274 	 */
275 	pt = grptb[((unsigned)arcn->sb.st_gid) % GRP_TB_SZ];
276 	while (pt != NULL) {
277 		if (pt->gid == arcn->sb.st_gid)
278 			return(0);
279 		pt = pt->fow;
280 	}
281 
282 	/*
283 	 * not found
284 	 */
285 	return(1);
286 }
287 
288 /*
289  * Time range selection routines
290  *
291  * Routines to handle user selection of files based on the modification and/or
292  * inode change time falling within a specified time range (the non-standard
293  * -T flag). The user may specify any number of different file time ranges.
294  * Time ranges are checked one at a time until a match is found (if at all).
295  * If the file has a mtime (and/or ctime) which lies within one of the time
296  * ranges, the file is selected. Time ranges may have a lower and/or an upper
297  * value. These ranges are inclusive. When no time ranges are supplied to pax
298  * with the -T option, all members in the archive will be selected by the time
299  * range routines. When only a lower range is supplied, only files with a
300  * mtime (and/or ctime) equal to or younger are selected. When only an upper
301  * range is supplied, only files with a mtime (and/or ctime) equal to or older
302  * are selected. When the lower time range is equal to the upper time range,
303  * only files with a mtime (or ctime) of exactly that time are selected.
304  */
305 
306 /*
307  * trng_add()
308  *	add a time range match to the time range list.
309  *	This is a non-standard pax option. Lower and upper ranges are in the
310  *	format: [[[[[cc]yy]mm]dd]HH]MM[.SS] and are comma separated.
311  *	Time ranges are based on current time, so 1234 would specify a time of
312  *	12:34 today.
313  * Return:
314  *	0 if the time range was added to the list, -1 otherwise
315  */
316 
317 int
318 trng_add(char *str)
319 {
320 	TIME_RNG *pt;
321 	char *up_pt = NULL;
322 	char *stpt;
323 	char *flgpt;
324 	int dot = 0;
325 
326 	/*
327 	 * throw out the badly formed time ranges
328 	 */
329 	if ((str == NULL) || (*str == '\0')) {
330 		paxwarn(1, "Empty time range string");
331 		return(-1);
332 	}
333 
334 	/*
335 	 * locate optional flags suffix /{cm}.
336 	 */
337 	if ((flgpt = strrchr(str, '/')) != NULL)
338 		*flgpt++ = '\0';
339 
340 	for (stpt = str; *stpt != '\0'; ++stpt) {
341 		if ((*stpt >= '0') && (*stpt <= '9'))
342 			continue;
343 		if ((*stpt == ',') && (up_pt == NULL)) {
344 			*stpt = '\0';
345 			up_pt = stpt + 1;
346 			dot = 0;
347 			continue;
348 		}
349 
350 		/*
351 		 * allow only one dot per range (secs)
352 		 */
353 		if ((*stpt == '.') && (!dot)) {
354 			++dot;
355 			continue;
356 		}
357 		paxwarn(1, "Improperly specified time range: %s", str);
358 		goto out;
359 	}
360 
361 	/*
362 	 * allocate space for the time range and store the limits
363 	 */
364 	if ((pt = (TIME_RNG *)malloc(sizeof(TIME_RNG))) == NULL) {
365 		paxwarn(1, "Unable to allocate memory for time range");
366 		return(-1);
367 	}
368 
369 	/*
370 	 * by default we only will check file mtime, but the user can specify
371 	 * mtime, ctime (inode change time) or both.
372 	 */
373 	if ((flgpt == NULL) || (*flgpt == '\0'))
374 		pt->flgs = CMPMTME;
375 	else {
376 		pt->flgs = 0;
377 		while (*flgpt != '\0') {
378 			switch(*flgpt) {
379 			case 'M':
380 			case 'm':
381 				pt->flgs |= CMPMTME;
382 				break;
383 			case 'C':
384 			case 'c':
385 				pt->flgs |= CMPCTME;
386 				break;
387 			default:
388 				paxwarn(1, "Bad option %c with time range %s",
389 				    *flgpt, str);
390 				free(pt);
391 				goto out;
392 			}
393 			++flgpt;
394 		}
395 	}
396 
397 	/*
398 	 * start off with the current time
399 	 */
400 	pt->low_time = pt->high_time = time(NULL);
401 	if (*str != '\0') {
402 		/*
403 		 * add lower limit
404 		 */
405 		if (str_sec(str, &(pt->low_time)) < 0) {
406 			paxwarn(1, "Illegal lower time range %s", str);
407 			free(pt);
408 			goto out;
409 		}
410 		pt->flgs |= HASLOW;
411 	}
412 
413 	if ((up_pt != NULL) && (*up_pt != '\0')) {
414 		/*
415 		 * add upper limit
416 		 */
417 		if (str_sec(up_pt, &(pt->high_time)) < 0) {
418 			paxwarn(1, "Illegal upper time range %s", up_pt);
419 			free(pt);
420 			goto out;
421 		}
422 		pt->flgs |= HASHIGH;
423 
424 		/*
425 		 * check that the upper and lower do not overlap
426 		 */
427 		if (pt->flgs & HASLOW) {
428 			if (pt->low_time > pt->high_time) {
429 				paxwarn(1, "Upper %s and lower %s time overlap",
430 					up_pt, str);
431 				free(pt);
432 				return(-1);
433 			}
434 		}
435 	}
436 
437 	pt->fow = NULL;
438 	if (trhead == NULL) {
439 		trtail = trhead = pt;
440 		return(0);
441 	}
442 	trtail->fow = pt;
443 	trtail = pt;
444 	return(0);
445 
446     out:
447 	paxwarn(1, "Time range format is: [[[[[cc]yy]mm]dd]HH]MM[.SS][/[c][m]]");
448 	return(-1);
449 }
450 
451 /*
452  * trng_match()
453  *	check if this files mtime/ctime falls within any supplied time range.
454  * Return:
455  *	0 if this archive member should be processed, 1 if it should be skipped
456  */
457 
458 static int
459 trng_match(ARCHD *arcn)
460 {
461 	TIME_RNG *pt;
462 
463 	/*
464 	 * have to search down the list one at a time looking for a match.
465 	 * remember time range limits are inclusive.
466 	 */
467 	pt = trhead;
468 	while (pt != NULL) {
469 		switch(pt->flgs & CMPBOTH) {
470 		case CMPBOTH:
471 			/*
472 			 * user wants both mtime and ctime checked for this
473 			 * time range
474 			 */
475 			if (((pt->flgs & HASLOW) &&
476 			    (arcn->sb.st_mtime < pt->low_time) &&
477 			    (arcn->sb.st_ctime < pt->low_time)) ||
478 			    ((pt->flgs & HASHIGH) &&
479 			    (arcn->sb.st_mtime > pt->high_time) &&
480 			    (arcn->sb.st_ctime > pt->high_time))) {
481 				pt = pt->fow;
482 				continue;
483 			}
484 			break;
485 		case CMPCTME:
486 			/*
487 			 * user wants only ctime checked for this time range
488 			 */
489 			if (((pt->flgs & HASLOW) &&
490 			    (arcn->sb.st_ctime < pt->low_time)) ||
491 			    ((pt->flgs & HASHIGH) &&
492 			    (arcn->sb.st_ctime > pt->high_time))) {
493 				pt = pt->fow;
494 				continue;
495 			}
496 			break;
497 		case CMPMTME:
498 		default:
499 			/*
500 			 * user wants only mtime checked for this time range
501 			 */
502 			if (((pt->flgs & HASLOW) &&
503 			    (arcn->sb.st_mtime < pt->low_time)) ||
504 			    ((pt->flgs & HASHIGH) &&
505 			    (arcn->sb.st_mtime > pt->high_time))) {
506 				pt = pt->fow;
507 				continue;
508 			}
509 			break;
510 		}
511 		break;
512 	}
513 
514 	if (pt == NULL)
515 		return(1);
516 	return(0);
517 }
518 
519 /*
520  * str_sec()
521  *	Convert a time string in the format of [[[[[cc]yy]mm]dd]HH]MM[.SS] to
522  *	seconds UTC. Tval already has current time loaded into it at entry.
523  * Return:
524  *	0 if converted ok, -1 otherwise
525  */
526 
527 static int
528 str_sec(const char *p, time_t *tval)
529 {
530 	struct tm *lt;
531 	const char *dot, *t;
532 	size_t len;
533 	int bigyear;
534 	int yearset;
535 
536 	yearset = 0;
537 	len = strlen(p);
538 
539 	for (t = p, dot = NULL; *t; ++t) {
540 		if (isdigit((unsigned char)*t))
541 			continue;
542 		if (*t == '.' && dot == NULL) {
543 			dot = t;
544 			continue;
545 		}
546 		return(-1);
547 	}
548 
549 	lt = localtime(tval);
550 
551 	if (dot != NULL) {			/* .SS */
552 		if (strlen(++dot) != 2)
553 			return(-1);
554 		lt->tm_sec = ATOI2(dot);
555 		if (lt->tm_sec > 61)
556 			return(-1);
557 		len -= 3;
558 	} else
559 		lt->tm_sec = 0;
560 
561 	switch (len) {
562 	case 12:				/* cc */
563 		bigyear = ATOI2(p);
564 		lt->tm_year = (bigyear * 100) - 1900;
565 		yearset = 1;
566 		/* FALLTHROUGH */
567 	case 10:				/* yy */
568 		if (yearset) {
569 			lt->tm_year += ATOI2(p);
570 		} else {
571 			lt->tm_year = ATOI2(p);
572 			if (lt->tm_year < 69)		/* hack for 2000 ;-} */
573 				lt->tm_year += (2000 - 1900);
574 		}
575 		/* FALLTHROUGH */
576 	case 8:					/* mm */
577 		lt->tm_mon = ATOI2(p);
578 		if ((lt->tm_mon > 12) || !lt->tm_mon)
579 			return(-1);
580 		--lt->tm_mon;			/* time struct is 0 - 11 */
581 		/* FALLTHROUGH */
582 	case 6:					/* dd */
583 		lt->tm_mday = ATOI2(p);
584 		if ((lt->tm_mday > 31) || !lt->tm_mday)
585 			return(-1);
586 		/* FALLTHROUGH */
587 	case 4:					/* HH */
588 		lt->tm_hour = ATOI2(p);
589 		if (lt->tm_hour > 23)
590 			return(-1);
591 		/* FALLTHROUGH */
592 	case 2:					/* MM */
593 		lt->tm_min = ATOI2(p);
594 		if (lt->tm_min > 59)
595 			return(-1);
596 		break;
597 	default:
598 		return(-1);
599 	}
600 
601 	/* convert broken-down time to UTC clock time seconds */
602 	if ((*tval = mktime(lt)) == -1)
603 		return(-1);
604 	return(0);
605 }
606