xref: /freebsd/bin/pax/sel_subs.c (revision 058ac3e8063366dafa634d9107642e12b038bf09)
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 __FBSDID("$FreeBSD$");
43 
44 #include <sys/types.h>
45 #include <sys/time.h>
46 #include <sys/stat.h>
47 
48 #include <ctype.h>
49 #include <grp.h>
50 #include <pwd.h>
51 #include <stdio.h>
52 #include <stdlib.h>
53 #include <string.h>
54 #include <strings.h>
55 
56 #include "pax.h"
57 #include "sel_subs.h"
58 #include "extern.h"
59 
60 static int str_sec(const char *, time_t *);
61 static int usr_match(ARCHD *);
62 static int grp_match(ARCHD *);
63 static int trng_match(ARCHD *);
64 
65 static TIME_RNG *trhead = NULL;		/* time range list head */
66 static TIME_RNG *trtail = NULL;		/* time range list tail */
67 static USRT **usrtb = NULL;		/* user selection table */
68 static GRPT **grptb = NULL;		/* group selection table */
69 
70 /*
71  * Routines for selection of archive members
72  */
73 
74 /*
75  * sel_chk()
76  *	check if this file matches a specified uid, gid or time range
77  * Return:
78  *	0 if this archive member should be processed, 1 if it should be skipped
79  */
80 
81 int
82 sel_chk(ARCHD *arcn)
83 {
84 	if (((usrtb != NULL) && usr_match(arcn)) ||
85 	    ((grptb != NULL) && grp_match(arcn)) ||
86 	    ((trhead != NULL) && trng_match(arcn)))
87 		return(1);
88 	return(0);
89 }
90 
91 /*
92  * User/group selection routines
93  *
94  * Routines to handle user selection of files based on the file uid/gid. To
95  * add an entry, the user supplies either the name or the uid/gid starting with
96  * a # on the command line. A \# will escape the #.
97  */
98 
99 /*
100  * usr_add()
101  *	add a user match to the user match hash table
102  * Return:
103  *	0 if added ok, -1 otherwise;
104  */
105 
106 int
107 usr_add(char *str)
108 {
109 	u_int indx;
110 	USRT *pt;
111 	struct passwd *pw;
112 	uid_t uid;
113 
114 	/*
115 	 * create the table if it doesn't exist
116 	 */
117 	if ((str == NULL) || (*str == '\0'))
118 		return(-1);
119 	if ((usrtb == NULL) &&
120  	    ((usrtb = (USRT **)calloc(USR_TB_SZ, sizeof(USRT *))) == NULL)) {
121 		paxwarn(1, "Unable to allocate memory for user selection table");
122 		return(-1);
123 	}
124 
125 	/*
126 	 * figure out user spec
127 	 */
128 	if (str[0] != '#') {
129 		/*
130 		 * it is a user name, \# escapes # as first char in user name
131 		 */
132 		if ((str[0] == '\\') && (str[1] == '#'))
133 			++str;
134 		if ((pw = getpwnam(str)) == NULL) {
135 			paxwarn(1, "Unable to find uid for user: %s", str);
136 			return(-1);
137 		}
138 		uid = (uid_t)pw->pw_uid;
139 	} else
140 		uid = (uid_t)strtoul(str+1, NULL, 10);
141 	endpwent();
142 
143 	/*
144 	 * hash it and go down the hash chain (if any) looking for it
145 	 */
146 	indx = ((unsigned)uid) % USR_TB_SZ;
147 	if ((pt = usrtb[indx]) != NULL) {
148 		while (pt != NULL) {
149 			if (pt->uid == uid)
150 				return(0);
151 			pt = pt->fow;
152 		}
153 	}
154 
155 	/*
156 	 * uid is not yet in the table, add it to the front of the chain
157 	 */
158 	if ((pt = (USRT *)malloc(sizeof(USRT))) != NULL) {
159 		pt->uid = uid;
160 		pt->fow = usrtb[indx];
161 		usrtb[indx] = pt;
162 		return(0);
163 	}
164 	paxwarn(1, "User selection table out of memory");
165 	return(-1);
166 }
167 
168 /*
169  * usr_match()
170  *	check if this files uid matches a selected uid.
171  * Return:
172  *	0 if this archive member should be processed, 1 if it should be skipped
173  */
174 
175 static int
176 usr_match(ARCHD *arcn)
177 {
178 	USRT *pt;
179 
180 	/*
181 	 * hash and look for it in the table
182 	 */
183 	pt = usrtb[((unsigned)arcn->sb.st_uid) % USR_TB_SZ];
184 	while (pt != NULL) {
185 		if (pt->uid == arcn->sb.st_uid)
186 			return(0);
187 		pt = pt->fow;
188 	}
189 
190 	/*
191 	 * not found
192 	 */
193 	return(1);
194 }
195 
196 /*
197  * grp_add()
198  *	add a group match to the group match hash table
199  * Return:
200  *	0 if added ok, -1 otherwise;
201  */
202 
203 int
204 grp_add(char *str)
205 {
206 	u_int indx;
207 	GRPT *pt;
208 	struct group *gr;
209 	gid_t gid;
210 
211 	/*
212 	 * create the table if it doesn't exist
213 	 */
214 	if ((str == NULL) || (*str == '\0'))
215 		return(-1);
216 	if ((grptb == NULL) &&
217  	    ((grptb = (GRPT **)calloc(GRP_TB_SZ, sizeof(GRPT *))) == NULL)) {
218 		paxwarn(1, "Unable to allocate memory fo group selection table");
219 		return(-1);
220 	}
221 
222 	/*
223 	 * figure out user spec
224 	 */
225 	if (str[0] != '#') {
226 		/*
227 		 * it is a group name, \# escapes # as first char in group name
228 		 */
229 		if ((str[0] == '\\') && (str[1] == '#'))
230 			++str;
231 		if ((gr = getgrnam(str)) == NULL) {
232 			paxwarn(1,"Cannot determine gid for group name: %s", str);
233 			return(-1);
234 		}
235 		gid = gr->gr_gid;
236 	} else
237 		gid = (gid_t)strtoul(str+1, NULL, 10);
238 	endgrent();
239 
240 	/*
241 	 * hash it and go down the hash chain (if any) looking for it
242 	 */
243 	indx = ((unsigned)gid) % GRP_TB_SZ;
244 	if ((pt = grptb[indx]) != NULL) {
245 		while (pt != NULL) {
246 			if (pt->gid == gid)
247 				return(0);
248 			pt = pt->fow;
249 		}
250 	}
251 
252 	/*
253 	 * gid not in the table, add it to the front of the chain
254 	 */
255 	if ((pt = (GRPT *)malloc(sizeof(GRPT))) != NULL) {
256 		pt->gid = gid;
257 		pt->fow = grptb[indx];
258 		grptb[indx] = pt;
259 		return(0);
260 	}
261 	paxwarn(1, "Group selection table out of memory");
262 	return(-1);
263 }
264 
265 /*
266  * grp_match()
267  *	check if this files gid matches a selected gid.
268  * Return:
269  *	0 if this archive member should be processed, 1 if it should be skipped
270  */
271 
272 static int
273 grp_match(ARCHD *arcn)
274 {
275 	GRPT *pt;
276 
277 	/*
278 	 * hash and look for it in the table
279 	 */
280 	pt = grptb[((unsigned)arcn->sb.st_gid) % GRP_TB_SZ];
281 	while (pt != NULL) {
282 		if (pt->gid == arcn->sb.st_gid)
283 			return(0);
284 		pt = pt->fow;
285 	}
286 
287 	/*
288 	 * not found
289 	 */
290 	return(1);
291 }
292 
293 /*
294  * Time range selection routines
295  *
296  * Routines to handle user selection of files based on the modification and/or
297  * inode change time falling within a specified time range (the non-standard
298  * -T flag). The user may specify any number of different file time ranges.
299  * Time ranges are checked one at a time until a match is found (if at all).
300  * If the file has a mtime (and/or ctime) which lies within one of the time
301  * ranges, the file is selected. Time ranges may have a lower and/or an upper
302  * value. These ranges are inclusive. When no time ranges are supplied to pax
303  * with the -T option, all members in the archive will be selected by the time
304  * range routines. When only a lower range is supplied, only files with a
305  * mtime (and/or ctime) equal to or younger are selected. When only an upper
306  * range is supplied, only files with a mtime (and/or ctime) equal to or older
307  * are selected. When the lower time range is equal to the upper time range,
308  * only files with a mtime (or ctime) of exactly that time are selected.
309  */
310 
311 /*
312  * trng_add()
313  *	add a time range match to the time range list.
314  *	This is a non-standard pax option. Lower and upper ranges are in the
315  *	format: [[[[[cc]yy]mm]dd]HH]MM[.SS] and are comma separated.
316  *	Time ranges are based on current time, so 1234 would specify a time of
317  *	12:34 today.
318  * Return:
319  *	0 if the time range was added to the list, -1 otherwise
320  */
321 
322 int
323 trng_add(char *str)
324 {
325 	TIME_RNG *pt;
326 	char *up_pt = NULL;
327 	char *stpt;
328 	char *flgpt;
329 	int dot = 0;
330 
331 	/*
332 	 * throw out the badly formed time ranges
333 	 */
334 	if ((str == NULL) || (*str == '\0')) {
335 		paxwarn(1, "Empty time range string");
336 		return(-1);
337 	}
338 
339 	/*
340 	 * locate optional flags suffix /{cm}.
341 	 */
342 	if ((flgpt = strrchr(str, '/')) != NULL)
343 		*flgpt++ = '\0';
344 
345 	for (stpt = str; *stpt != '\0'; ++stpt) {
346 		if ((*stpt >= '0') && (*stpt <= '9'))
347 			continue;
348 		if ((*stpt == ',') && (up_pt == NULL)) {
349 			*stpt = '\0';
350 			up_pt = stpt + 1;
351 			dot = 0;
352 			continue;
353 		}
354 
355 		/*
356 		 * allow only one dot per range (secs)
357 		 */
358 		if ((*stpt == '.') && (!dot)) {
359 			++dot;
360 			continue;
361 		}
362 		paxwarn(1, "Improperly specified time range: %s", str);
363 		goto out;
364 	}
365 
366 	/*
367 	 * allocate space for the time range and store the limits
368 	 */
369 	if ((pt = (TIME_RNG *)malloc(sizeof(TIME_RNG))) == NULL) {
370 		paxwarn(1, "Unable to allocate memory for time range");
371 		return(-1);
372 	}
373 
374 	/*
375 	 * by default we only will check file mtime, but the user can specify
376 	 * mtime, ctime (inode change time) or both.
377 	 */
378 	if ((flgpt == NULL) || (*flgpt == '\0'))
379 		pt->flgs = CMPMTME;
380 	else {
381 		pt->flgs = 0;
382 		while (*flgpt != '\0') {
383 			switch(*flgpt) {
384 			case 'M':
385 			case 'm':
386 				pt->flgs |= CMPMTME;
387 				break;
388 			case 'C':
389 			case 'c':
390 				pt->flgs |= CMPCTME;
391 				break;
392 			default:
393 				paxwarn(1, "Bad option %c with time range %s",
394 				    *flgpt, str);
395 				free(pt);
396 				goto out;
397 			}
398 			++flgpt;
399 		}
400 	}
401 
402 	/*
403 	 * start off with the current time
404 	 */
405 	pt->low_time = pt->high_time = time(NULL);
406 	if (*str != '\0') {
407 		/*
408 		 * add lower limit
409 		 */
410 		if (str_sec(str, &(pt->low_time)) < 0) {
411 			paxwarn(1, "Illegal lower time range %s", str);
412 			free(pt);
413 			goto out;
414 		}
415 		pt->flgs |= HASLOW;
416 	}
417 
418 	if ((up_pt != NULL) && (*up_pt != '\0')) {
419 		/*
420 		 * add upper limit
421 		 */
422 		if (str_sec(up_pt, &(pt->high_time)) < 0) {
423 			paxwarn(1, "Illegal upper time range %s", up_pt);
424 			free(pt);
425 			goto out;
426 		}
427 		pt->flgs |= HASHIGH;
428 
429 		/*
430 		 * check that the upper and lower do not overlap
431 		 */
432 		if (pt->flgs & HASLOW) {
433 			if (pt->low_time > pt->high_time) {
434 				paxwarn(1, "Upper %s and lower %s time overlap",
435 					up_pt, str);
436 				free(pt);
437 				return(-1);
438 			}
439 		}
440 	}
441 
442 	pt->fow = NULL;
443 	if (trhead == NULL) {
444 		trtail = trhead = pt;
445 		return(0);
446 	}
447 	trtail->fow = pt;
448 	trtail = pt;
449 	return(0);
450 
451     out:
452 	paxwarn(1, "Time range format is: [[[[[cc]yy]mm]dd]HH]MM[.SS][/[c][m]]");
453 	return(-1);
454 }
455 
456 /*
457  * trng_match()
458  *	check if this files mtime/ctime falls within any supplied time range.
459  * Return:
460  *	0 if this archive member should be processed, 1 if it should be skipped
461  */
462 
463 static int
464 trng_match(ARCHD *arcn)
465 {
466 	TIME_RNG *pt;
467 
468 	/*
469 	 * have to search down the list one at a time looking for a match.
470 	 * remember time range limits are inclusive.
471 	 */
472 	pt = trhead;
473 	while (pt != NULL) {
474 		switch(pt->flgs & CMPBOTH) {
475 		case CMPBOTH:
476 			/*
477 			 * user wants both mtime and ctime checked for this
478 			 * time range
479 			 */
480 			if (((pt->flgs & HASLOW) &&
481 			    (arcn->sb.st_mtime < pt->low_time) &&
482 			    (arcn->sb.st_ctime < pt->low_time)) ||
483 			    ((pt->flgs & HASHIGH) &&
484 			    (arcn->sb.st_mtime > pt->high_time) &&
485 			    (arcn->sb.st_ctime > pt->high_time))) {
486 				pt = pt->fow;
487 				continue;
488 			}
489 			break;
490 		case CMPCTME:
491 			/*
492 			 * user wants only ctime checked for this time range
493 			 */
494 			if (((pt->flgs & HASLOW) &&
495 			    (arcn->sb.st_ctime < pt->low_time)) ||
496 			    ((pt->flgs & HASHIGH) &&
497 			    (arcn->sb.st_ctime > pt->high_time))) {
498 				pt = pt->fow;
499 				continue;
500 			}
501 			break;
502 		case CMPMTME:
503 		default:
504 			/*
505 			 * user wants only mtime checked for this time range
506 			 */
507 			if (((pt->flgs & HASLOW) &&
508 			    (arcn->sb.st_mtime < pt->low_time)) ||
509 			    ((pt->flgs & HASHIGH) &&
510 			    (arcn->sb.st_mtime > pt->high_time))) {
511 				pt = pt->fow;
512 				continue;
513 			}
514 			break;
515 		}
516 		break;
517 	}
518 
519 	if (pt == NULL)
520 		return(1);
521 	return(0);
522 }
523 
524 /*
525  * str_sec()
526  *	Convert a time string in the format of [[[[[cc]yy]mm]dd]HH]MM[.SS] to
527  *	seconds UTC. Tval already has current time loaded into it at entry.
528  * Return:
529  *	0 if converted ok, -1 otherwise
530  */
531 
532 static int
533 str_sec(const char *p, time_t *tval)
534 {
535 	struct tm *lt;
536 	const char *dot, *t;
537 	size_t len;
538 	int bigyear;
539 	int yearset;
540 
541 	yearset = 0;
542 	len = strlen(p);
543 
544 	for (t = p, dot = NULL; *t; ++t) {
545 		if (isdigit((unsigned char)*t))
546 			continue;
547 		if (*t == '.' && dot == NULL) {
548 			dot = t;
549 			continue;
550 		}
551 		return(-1);
552 	}
553 
554 	lt = localtime(tval);
555 
556 	if (dot != NULL) {			/* .SS */
557 		if (strlen(++dot) != 2)
558 			return(-1);
559 		lt->tm_sec = ATOI2(dot);
560 		if (lt->tm_sec > 61)
561 			return(-1);
562 		len -= 3;
563 	} else
564 		lt->tm_sec = 0;
565 
566 	switch (len) {
567 	case 12:				/* cc */
568 		bigyear = ATOI2(p);
569 		lt->tm_year = (bigyear * 100) - 1900;
570 		yearset = 1;
571 		/* FALLTHROUGH */
572 	case 10:				/* yy */
573 		if (yearset) {
574 			lt->tm_year += ATOI2(p);
575 		} else {
576 			lt->tm_year = ATOI2(p);
577 			if (lt->tm_year < 69)		/* hack for 2000 ;-} */
578 				lt->tm_year += (2000 - 1900);
579 		}
580 		/* FALLTHROUGH */
581 	case 8:					/* mm */
582 		lt->tm_mon = ATOI2(p);
583 		if ((lt->tm_mon > 12) || !lt->tm_mon)
584 			return(-1);
585 		--lt->tm_mon;			/* time struct is 0 - 11 */
586 		/* FALLTHROUGH */
587 	case 6:					/* dd */
588 		lt->tm_mday = ATOI2(p);
589 		if ((lt->tm_mday > 31) || !lt->tm_mday)
590 			return(-1);
591 		/* FALLTHROUGH */
592 	case 4:					/* HH */
593 		lt->tm_hour = ATOI2(p);
594 		if (lt->tm_hour > 23)
595 			return(-1);
596 		/* FALLTHROUGH */
597 	case 2:					/* MM */
598 		lt->tm_min = ATOI2(p);
599 		if (lt->tm_min > 59)
600 			return(-1);
601 		break;
602 	default:
603 		return(-1);
604 	}
605 
606 	/* convert broken-down time to UTC clock time seconds */
607 	if ((*tval = mktime(lt)) == -1)
608 		return(-1);
609 	return(0);
610 }
611