xref: /freebsd/bin/pax/sel_subs.c (revision 1db64f89363c97858961c4df0b7d02f3223723cf)
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 #include <sys/types.h>
37 #include <sys/time.h>
38 #include <sys/stat.h>
39 
40 #include <ctype.h>
41 #include <grp.h>
42 #include <pwd.h>
43 #include <stdio.h>
44 #include <stdlib.h>
45 #include <string.h>
46 #include <strings.h>
47 
48 #include "pax.h"
49 #include "sel_subs.h"
50 #include "extern.h"
51 
52 static int str_sec(const char *, time_t *);
53 static int usr_match(ARCHD *);
54 static int grp_match(ARCHD *);
55 static int trng_match(ARCHD *);
56 
57 static TIME_RNG *trhead = NULL;		/* time range list head */
58 static TIME_RNG *trtail = NULL;		/* time range list tail */
59 static USRT **usrtb = NULL;		/* user selection table */
60 static GRPT **grptb = NULL;		/* group selection table */
61 
62 /*
63  * Routines for selection of archive members
64  */
65 
66 /*
67  * sel_chk()
68  *	check if this file matches a specified uid, gid or time range
69  * Return:
70  *	0 if this archive member should be processed, 1 if it should be skipped
71  */
72 
73 int
74 sel_chk(ARCHD *arcn)
75 {
76 	if (((usrtb != NULL) && usr_match(arcn)) ||
77 	    ((grptb != NULL) && grp_match(arcn)) ||
78 	    ((trhead != NULL) && trng_match(arcn)))
79 		return(1);
80 	return(0);
81 }
82 
83 /*
84  * User/group selection routines
85  *
86  * Routines to handle user selection of files based on the file uid/gid. To
87  * add an entry, the user supplies either the name or the uid/gid starting with
88  * a # on the command line. A \# will escape the #.
89  */
90 
91 /*
92  * usr_add()
93  *	add a user match to the user match hash table
94  * Return:
95  *	0 if added ok, -1 otherwise;
96  */
97 
98 int
99 usr_add(char *str)
100 {
101 	u_int indx;
102 	USRT *pt;
103 	struct passwd *pw;
104 	uid_t uid;
105 
106 	/*
107 	 * create the table if it doesn't exist
108 	 */
109 	if ((str == NULL) || (*str == '\0'))
110 		return(-1);
111 	if ((usrtb == NULL) &&
112  	    ((usrtb = (USRT **)calloc(USR_TB_SZ, sizeof(USRT *))) == NULL)) {
113 		paxwarn(1, "Unable to allocate memory for user selection table");
114 		return(-1);
115 	}
116 
117 	/*
118 	 * figure out user spec
119 	 */
120 	if (str[0] != '#') {
121 		/*
122 		 * it is a user name, \# escapes # as first char in user name
123 		 */
124 		if ((str[0] == '\\') && (str[1] == '#'))
125 			++str;
126 		if ((pw = getpwnam(str)) == NULL) {
127 			paxwarn(1, "Unable to find uid for user: %s", str);
128 			return(-1);
129 		}
130 		uid = (uid_t)pw->pw_uid;
131 	} else
132 		uid = (uid_t)strtoul(str+1, NULL, 10);
133 	endpwent();
134 
135 	/*
136 	 * hash it and go down the hash chain (if any) looking for it
137 	 */
138 	indx = ((unsigned)uid) % USR_TB_SZ;
139 	if ((pt = usrtb[indx]) != NULL) {
140 		while (pt != NULL) {
141 			if (pt->uid == uid)
142 				return(0);
143 			pt = pt->fow;
144 		}
145 	}
146 
147 	/*
148 	 * uid is not yet in the table, add it to the front of the chain
149 	 */
150 	if ((pt = (USRT *)malloc(sizeof(USRT))) != NULL) {
151 		pt->uid = uid;
152 		pt->fow = usrtb[indx];
153 		usrtb[indx] = pt;
154 		return(0);
155 	}
156 	paxwarn(1, "User selection table out of memory");
157 	return(-1);
158 }
159 
160 /*
161  * usr_match()
162  *	check if this files uid matches a selected uid.
163  * Return:
164  *	0 if this archive member should be processed, 1 if it should be skipped
165  */
166 
167 static int
168 usr_match(ARCHD *arcn)
169 {
170 	USRT *pt;
171 
172 	/*
173 	 * hash and look for it in the table
174 	 */
175 	pt = usrtb[((unsigned)arcn->sb.st_uid) % USR_TB_SZ];
176 	while (pt != NULL) {
177 		if (pt->uid == arcn->sb.st_uid)
178 			return(0);
179 		pt = pt->fow;
180 	}
181 
182 	/*
183 	 * not found
184 	 */
185 	return(1);
186 }
187 
188 /*
189  * grp_add()
190  *	add a group match to the group match hash table
191  * Return:
192  *	0 if added ok, -1 otherwise;
193  */
194 
195 int
196 grp_add(char *str)
197 {
198 	u_int indx;
199 	GRPT *pt;
200 	struct group *gr;
201 	gid_t gid;
202 
203 	/*
204 	 * create the table if it doesn't exist
205 	 */
206 	if ((str == NULL) || (*str == '\0'))
207 		return(-1);
208 	if ((grptb == NULL) &&
209  	    ((grptb = (GRPT **)calloc(GRP_TB_SZ, sizeof(GRPT *))) == NULL)) {
210 		paxwarn(1, "Unable to allocate memory fo group selection table");
211 		return(-1);
212 	}
213 
214 	/*
215 	 * figure out user spec
216 	 */
217 	if (str[0] != '#') {
218 		/*
219 		 * it is a group name, \# escapes # as first char in group name
220 		 */
221 		if ((str[0] == '\\') && (str[1] == '#'))
222 			++str;
223 		if ((gr = getgrnam(str)) == NULL) {
224 			paxwarn(1,"Cannot determine gid for group name: %s", str);
225 			return(-1);
226 		}
227 		gid = gr->gr_gid;
228 	} else
229 		gid = (gid_t)strtoul(str+1, NULL, 10);
230 	endgrent();
231 
232 	/*
233 	 * hash it and go down the hash chain (if any) looking for it
234 	 */
235 	indx = ((unsigned)gid) % GRP_TB_SZ;
236 	if ((pt = grptb[indx]) != NULL) {
237 		while (pt != NULL) {
238 			if (pt->gid == gid)
239 				return(0);
240 			pt = pt->fow;
241 		}
242 	}
243 
244 	/*
245 	 * gid not in the table, add it to the front of the chain
246 	 */
247 	if ((pt = (GRPT *)malloc(sizeof(GRPT))) != NULL) {
248 		pt->gid = gid;
249 		pt->fow = grptb[indx];
250 		grptb[indx] = pt;
251 		return(0);
252 	}
253 	paxwarn(1, "Group selection table out of memory");
254 	return(-1);
255 }
256 
257 /*
258  * grp_match()
259  *	check if this files gid matches a selected gid.
260  * Return:
261  *	0 if this archive member should be processed, 1 if it should be skipped
262  */
263 
264 static int
265 grp_match(ARCHD *arcn)
266 {
267 	GRPT *pt;
268 
269 	/*
270 	 * hash and look for it in the table
271 	 */
272 	pt = grptb[((unsigned)arcn->sb.st_gid) % GRP_TB_SZ];
273 	while (pt != NULL) {
274 		if (pt->gid == arcn->sb.st_gid)
275 			return(0);
276 		pt = pt->fow;
277 	}
278 
279 	/*
280 	 * not found
281 	 */
282 	return(1);
283 }
284 
285 /*
286  * Time range selection routines
287  *
288  * Routines to handle user selection of files based on the modification and/or
289  * inode change time falling within a specified time range (the non-standard
290  * -T flag). The user may specify any number of different file time ranges.
291  * Time ranges are checked one at a time until a match is found (if at all).
292  * If the file has a mtime (and/or ctime) which lies within one of the time
293  * ranges, the file is selected. Time ranges may have a lower and/or an upper
294  * value. These ranges are inclusive. When no time ranges are supplied to pax
295  * with the -T option, all members in the archive will be selected by the time
296  * range routines. When only a lower range is supplied, only files with a
297  * mtime (and/or ctime) equal to or younger are selected. When only an upper
298  * range is supplied, only files with a mtime (and/or ctime) equal to or older
299  * are selected. When the lower time range is equal to the upper time range,
300  * only files with a mtime (or ctime) of exactly that time are selected.
301  */
302 
303 /*
304  * trng_add()
305  *	add a time range match to the time range list.
306  *	This is a non-standard pax option. Lower and upper ranges are in the
307  *	format: [[[[[cc]yy]mm]dd]HH]MM[.SS] and are comma separated.
308  *	Time ranges are based on current time, so 1234 would specify a time of
309  *	12:34 today.
310  * Return:
311  *	0 if the time range was added to the list, -1 otherwise
312  */
313 
314 int
315 trng_add(char *str)
316 {
317 	TIME_RNG *pt;
318 	char *up_pt = NULL;
319 	char *stpt;
320 	char *flgpt;
321 	int dot = 0;
322 
323 	/*
324 	 * throw out the badly formed time ranges
325 	 */
326 	if ((str == NULL) || (*str == '\0')) {
327 		paxwarn(1, "Empty time range string");
328 		return(-1);
329 	}
330 
331 	/*
332 	 * locate optional flags suffix /{cm}.
333 	 */
334 	if ((flgpt = strrchr(str, '/')) != NULL)
335 		*flgpt++ = '\0';
336 
337 	for (stpt = str; *stpt != '\0'; ++stpt) {
338 		if ((*stpt >= '0') && (*stpt <= '9'))
339 			continue;
340 		if ((*stpt == ',') && (up_pt == NULL)) {
341 			*stpt = '\0';
342 			up_pt = stpt + 1;
343 			dot = 0;
344 			continue;
345 		}
346 
347 		/*
348 		 * allow only one dot per range (secs)
349 		 */
350 		if ((*stpt == '.') && (!dot)) {
351 			++dot;
352 			continue;
353 		}
354 		paxwarn(1, "Improperly specified time range: %s", str);
355 		goto out;
356 	}
357 
358 	/*
359 	 * allocate space for the time range and store the limits
360 	 */
361 	if ((pt = (TIME_RNG *)malloc(sizeof(TIME_RNG))) == NULL) {
362 		paxwarn(1, "Unable to allocate memory for time range");
363 		return(-1);
364 	}
365 
366 	/*
367 	 * by default we only will check file mtime, but the user can specify
368 	 * mtime, ctime (inode change time) or both.
369 	 */
370 	if ((flgpt == NULL) || (*flgpt == '\0'))
371 		pt->flgs = CMPMTME;
372 	else {
373 		pt->flgs = 0;
374 		while (*flgpt != '\0') {
375 			switch(*flgpt) {
376 			case 'M':
377 			case 'm':
378 				pt->flgs |= CMPMTME;
379 				break;
380 			case 'C':
381 			case 'c':
382 				pt->flgs |= CMPCTME;
383 				break;
384 			default:
385 				paxwarn(1, "Bad option %c with time range %s",
386 				    *flgpt, str);
387 				free(pt);
388 				goto out;
389 			}
390 			++flgpt;
391 		}
392 	}
393 
394 	/*
395 	 * start off with the current time
396 	 */
397 	pt->low_time = pt->high_time = time(NULL);
398 	if (*str != '\0') {
399 		/*
400 		 * add lower limit
401 		 */
402 		if (str_sec(str, &(pt->low_time)) < 0) {
403 			paxwarn(1, "Illegal lower time range %s", str);
404 			free(pt);
405 			goto out;
406 		}
407 		pt->flgs |= HASLOW;
408 	}
409 
410 	if ((up_pt != NULL) && (*up_pt != '\0')) {
411 		/*
412 		 * add upper limit
413 		 */
414 		if (str_sec(up_pt, &(pt->high_time)) < 0) {
415 			paxwarn(1, "Illegal upper time range %s", up_pt);
416 			free(pt);
417 			goto out;
418 		}
419 		pt->flgs |= HASHIGH;
420 
421 		/*
422 		 * check that the upper and lower do not overlap
423 		 */
424 		if (pt->flgs & HASLOW) {
425 			if (pt->low_time > pt->high_time) {
426 				paxwarn(1, "Upper %s and lower %s time overlap",
427 					up_pt, str);
428 				free(pt);
429 				return(-1);
430 			}
431 		}
432 	}
433 
434 	pt->fow = NULL;
435 	if (trhead == NULL) {
436 		trtail = trhead = pt;
437 		return(0);
438 	}
439 	trtail->fow = pt;
440 	trtail = pt;
441 	return(0);
442 
443     out:
444 	paxwarn(1, "Time range format is: [[[[[cc]yy]mm]dd]HH]MM[.SS][/[c][m]]");
445 	return(-1);
446 }
447 
448 /*
449  * trng_match()
450  *	check if this files mtime/ctime falls within any supplied time range.
451  * Return:
452  *	0 if this archive member should be processed, 1 if it should be skipped
453  */
454 
455 static int
456 trng_match(ARCHD *arcn)
457 {
458 	TIME_RNG *pt;
459 
460 	/*
461 	 * have to search down the list one at a time looking for a match.
462 	 * remember time range limits are inclusive.
463 	 */
464 	pt = trhead;
465 	while (pt != NULL) {
466 		switch(pt->flgs & CMPBOTH) {
467 		case CMPBOTH:
468 			/*
469 			 * user wants both mtime and ctime checked for this
470 			 * time range
471 			 */
472 			if (((pt->flgs & HASLOW) &&
473 			    (arcn->sb.st_mtime < pt->low_time) &&
474 			    (arcn->sb.st_ctime < pt->low_time)) ||
475 			    ((pt->flgs & HASHIGH) &&
476 			    (arcn->sb.st_mtime > pt->high_time) &&
477 			    (arcn->sb.st_ctime > pt->high_time))) {
478 				pt = pt->fow;
479 				continue;
480 			}
481 			break;
482 		case CMPCTME:
483 			/*
484 			 * user wants only ctime checked for this time range
485 			 */
486 			if (((pt->flgs & HASLOW) &&
487 			    (arcn->sb.st_ctime < pt->low_time)) ||
488 			    ((pt->flgs & HASHIGH) &&
489 			    (arcn->sb.st_ctime > pt->high_time))) {
490 				pt = pt->fow;
491 				continue;
492 			}
493 			break;
494 		case CMPMTME:
495 		default:
496 			/*
497 			 * user wants only mtime checked for this time range
498 			 */
499 			if (((pt->flgs & HASLOW) &&
500 			    (arcn->sb.st_mtime < pt->low_time)) ||
501 			    ((pt->flgs & HASHIGH) &&
502 			    (arcn->sb.st_mtime > pt->high_time))) {
503 				pt = pt->fow;
504 				continue;
505 			}
506 			break;
507 		}
508 		break;
509 	}
510 
511 	if (pt == NULL)
512 		return(1);
513 	return(0);
514 }
515 
516 /*
517  * str_sec()
518  *	Convert a time string in the format of [[[[[cc]yy]mm]dd]HH]MM[.SS] to
519  *	seconds UTC. Tval already has current time loaded into it at entry.
520  * Return:
521  *	0 if converted ok, -1 otherwise
522  */
523 
524 static int
525 str_sec(const char *p, time_t *tval)
526 {
527 	struct tm *lt;
528 	const char *dot, *t;
529 	size_t len;
530 	int bigyear;
531 	int yearset;
532 
533 	yearset = 0;
534 	len = strlen(p);
535 
536 	for (t = p, dot = NULL; *t; ++t) {
537 		if (isdigit((unsigned char)*t))
538 			continue;
539 		if (*t == '.' && dot == NULL) {
540 			dot = t;
541 			continue;
542 		}
543 		return(-1);
544 	}
545 
546 	lt = localtime(tval);
547 
548 	if (dot != NULL) {			/* .SS */
549 		if (strlen(++dot) != 2)
550 			return(-1);
551 		lt->tm_sec = ATOI2(dot);
552 		if (lt->tm_sec > 61)
553 			return(-1);
554 		len -= 3;
555 	} else
556 		lt->tm_sec = 0;
557 
558 	switch (len) {
559 	case 12:				/* cc */
560 		bigyear = ATOI2(p);
561 		lt->tm_year = (bigyear * 100) - 1900;
562 		yearset = 1;
563 		/* FALLTHROUGH */
564 	case 10:				/* yy */
565 		if (yearset) {
566 			lt->tm_year += ATOI2(p);
567 		} else {
568 			lt->tm_year = ATOI2(p);
569 			if (lt->tm_year < 69)		/* hack for 2000 ;-} */
570 				lt->tm_year += (2000 - 1900);
571 		}
572 		/* FALLTHROUGH */
573 	case 8:					/* mm */
574 		lt->tm_mon = ATOI2(p);
575 		if ((lt->tm_mon > 12) || !lt->tm_mon)
576 			return(-1);
577 		--lt->tm_mon;			/* time struct is 0 - 11 */
578 		/* FALLTHROUGH */
579 	case 6:					/* dd */
580 		lt->tm_mday = ATOI2(p);
581 		if ((lt->tm_mday > 31) || !lt->tm_mday)
582 			return(-1);
583 		/* FALLTHROUGH */
584 	case 4:					/* HH */
585 		lt->tm_hour = ATOI2(p);
586 		if (lt->tm_hour > 23)
587 			return(-1);
588 		/* FALLTHROUGH */
589 	case 2:					/* MM */
590 		lt->tm_min = ATOI2(p);
591 		if (lt->tm_min > 59)
592 			return(-1);
593 		break;
594 	default:
595 		return(-1);
596 	}
597 
598 	/* convert broken-down time to UTC clock time seconds */
599 	if ((*tval = mktime(lt)) == -1)
600 		return(-1);
601 	return(0);
602 }
603