xref: /freebsd/bin/pax/sel_subs.c (revision d4eeb02986980bf33dd56c41ceb9fc5f180c0d47)
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 #include <pwd.h>
48 #include <grp.h>
49 #include <stdio.h>
50 #include <string.h>
51 #include <strings.h>
52 #include <stdlib.h>
53 #include "pax.h"
54 #include "sel_subs.h"
55 #include "extern.h"
56 
57 static int str_sec(char *, time_t *);
58 static int usr_match(ARCHD *);
59 static int grp_match(ARCHD *);
60 static int trng_match(ARCHD *);
61 
62 static TIME_RNG *trhead = NULL;		/* time range list head */
63 static TIME_RNG *trtail = NULL;		/* time range list tail */
64 static USRT **usrtb = NULL;		/* user selection table */
65 static GRPT **grptb = NULL;		/* group selection table */
66 
67 /*
68  * Routines for selection of archive members
69  */
70 
71 /*
72  * sel_chk()
73  *	check if this file matches a specified uid, gid or time range
74  * Return:
75  *	0 if this archive member should be processed, 1 if it should be skipped
76  */
77 
78 int
79 sel_chk(ARCHD *arcn)
80 {
81 	if (((usrtb != NULL) && usr_match(arcn)) ||
82 	    ((grptb != NULL) && grp_match(arcn)) ||
83 	    ((trhead != NULL) && trng_match(arcn)))
84 		return(1);
85 	return(0);
86 }
87 
88 /*
89  * User/group selection routines
90  *
91  * Routines to handle user selection of files based on the file uid/gid. To
92  * add an entry, the user supplies either then name or the uid/gid starting with
93  * a # on the command line. A \# will escape the #.
94  */
95 
96 /*
97  * usr_add()
98  *	add a user match to the user match hash table
99  * Return:
100  *	0 if added ok, -1 otherwise;
101  */
102 
103 int
104 usr_add(char *str)
105 {
106 	u_int indx;
107 	USRT *pt;
108 	struct passwd *pw;
109 	uid_t uid;
110 
111 	/*
112 	 * create the table if it doesn't exist
113 	 */
114 	if ((str == NULL) || (*str == '\0'))
115 		return(-1);
116 	if ((usrtb == NULL) &&
117  	    ((usrtb = (USRT **)calloc(USR_TB_SZ, sizeof(USRT *))) == NULL)) {
118 		paxwarn(1, "Unable to allocate memory for user selection table");
119 		return(-1);
120 	}
121 
122 	/*
123 	 * figure out user spec
124 	 */
125 	if (str[0] != '#') {
126 		/*
127 		 * it is a user name, \# escapes # as first char in user name
128 		 */
129 		if ((str[0] == '\\') && (str[1] == '#'))
130 			++str;
131 		if ((pw = getpwnam(str)) == NULL) {
132 			paxwarn(1, "Unable to find uid for user: %s", str);
133 			return(-1);
134 		}
135 		uid = (uid_t)pw->pw_uid;
136 	} else
137 #		ifdef NET2_STAT
138 		uid = (uid_t)atoi(str+1);
139 #		else
140 		uid = (uid_t)strtoul(str+1, NULL, 10);
141 #		endif
142 	endpwent();
143 
144 	/*
145 	 * hash it and go down the hash chain (if any) looking for it
146 	 */
147 	indx = ((unsigned)uid) % USR_TB_SZ;
148 	if ((pt = usrtb[indx]) != NULL) {
149 		while (pt != NULL) {
150 			if (pt->uid == uid)
151 				return(0);
152 			pt = pt->fow;
153 		}
154 	}
155 
156 	/*
157 	 * uid is not yet in the table, add it to the front of the chain
158 	 */
159 	if ((pt = (USRT *)malloc(sizeof(USRT))) != NULL) {
160 		pt->uid = uid;
161 		pt->fow = usrtb[indx];
162 		usrtb[indx] = pt;
163 		return(0);
164 	}
165 	paxwarn(1, "User selection table out of memory");
166 	return(-1);
167 }
168 
169 /*
170  * usr_match()
171  *	check if this files uid matches a selected uid.
172  * Return:
173  *	0 if this archive member should be processed, 1 if it should be skipped
174  */
175 
176 static int
177 usr_match(ARCHD *arcn)
178 {
179 	USRT *pt;
180 
181 	/*
182 	 * hash and look for it in the table
183 	 */
184 	pt = usrtb[((unsigned)arcn->sb.st_uid) % USR_TB_SZ];
185 	while (pt != NULL) {
186 		if (pt->uid == arcn->sb.st_uid)
187 			return(0);
188 		pt = pt->fow;
189 	}
190 
191 	/*
192 	 * not found
193 	 */
194 	return(1);
195 }
196 
197 /*
198  * grp_add()
199  *	add a group match to the group match hash table
200  * Return:
201  *	0 if added ok, -1 otherwise;
202  */
203 
204 int
205 grp_add(char *str)
206 {
207 	u_int indx;
208 	GRPT *pt;
209 	struct group *gr;
210 	gid_t gid;
211 
212 	/*
213 	 * create the table if it doesn't exist
214 	 */
215 	if ((str == NULL) || (*str == '\0'))
216 		return(-1);
217 	if ((grptb == NULL) &&
218  	    ((grptb = (GRPT **)calloc(GRP_TB_SZ, sizeof(GRPT *))) == NULL)) {
219 		paxwarn(1, "Unable to allocate memory fo group selection table");
220 		return(-1);
221 	}
222 
223 	/*
224 	 * figure out user spec
225 	 */
226 	if (str[0] != '#') {
227 		/*
228 		 * it is a group name, \# escapes # as first char in group name
229 		 */
230 		if ((str[0] == '\\') && (str[1] == '#'))
231 			++str;
232 		if ((gr = getgrnam(str)) == NULL) {
233 			paxwarn(1,"Cannot determine gid for group name: %s", str);
234 			return(-1);
235 		}
236 		gid = gr->gr_gid;
237 	} else
238 #		ifdef NET2_STAT
239 		gid = (gid_t)atoi(str+1);
240 #		else
241 		gid = (gid_t)strtoul(str+1, NULL, 10);
242 #		endif
243 	endgrent();
244 
245 	/*
246 	 * hash it and go down the hash chain (if any) looking for it
247 	 */
248 	indx = ((unsigned)gid) % GRP_TB_SZ;
249 	if ((pt = grptb[indx]) != NULL) {
250 		while (pt != NULL) {
251 			if (pt->gid == gid)
252 				return(0);
253 			pt = pt->fow;
254 		}
255 	}
256 
257 	/*
258 	 * gid not in the table, add it to the front of the chain
259 	 */
260 	if ((pt = (GRPT *)malloc(sizeof(GRPT))) != NULL) {
261 		pt->gid = gid;
262 		pt->fow = grptb[indx];
263 		grptb[indx] = pt;
264 		return(0);
265 	}
266 	paxwarn(1, "Group selection table out of memory");
267 	return(-1);
268 }
269 
270 /*
271  * grp_match()
272  *	check if this files gid matches a selected gid.
273  * Return:
274  *	0 if this archive member should be processed, 1 if it should be skipped
275  */
276 
277 static int
278 grp_match(ARCHD *arcn)
279 {
280 	GRPT *pt;
281 
282 	/*
283 	 * hash and look for it in the table
284 	 */
285 	pt = grptb[((unsigned)arcn->sb.st_gid) % GRP_TB_SZ];
286 	while (pt != NULL) {
287 		if (pt->gid == arcn->sb.st_gid)
288 			return(0);
289 		pt = pt->fow;
290 	}
291 
292 	/*
293 	 * not found
294 	 */
295 	return(1);
296 }
297 
298 /*
299  * Time range selection routines
300  *
301  * Routines to handle user selection of files based on the modification and/or
302  * inode change time falling within a specified time range (the non-standard
303  * -T flag). The user may specify any number of different file time ranges.
304  * Time ranges are checked one at a time until a match is found (if at all).
305  * If the file has a mtime (and/or ctime) which lies within one of the time
306  * ranges, the file is selected. Time ranges may have a lower and/or an upper
307  * value. These ranges are inclusive. When no time ranges are supplied to pax
308  * with the -T option, all members in the archive will be selected by the time
309  * range routines. When only a lower range is supplied, only files with a
310  * mtime (and/or ctime) equal to or younger are selected. When only an upper
311  * range is supplied, only files with a mtime (and/or ctime) equal to or older
312  * are selected. When the lower time range is equal to the upper time range,
313  * only files with a mtime (or ctime) of exactly that time are selected.
314  */
315 
316 /*
317  * trng_add()
318  *	add a time range match to the time range list.
319  *	This is a non-standard pax option. Lower and upper ranges are in the
320  *	format: [yy[mm[dd[hh]]]]mm[.ss] and are comma separated.
321  *	Time ranges are based on current time, so 1234 would specify a time of
322  *	12:34 today.
323  * Return:
324  *	0 if the time range was added to the list, -1 otherwise
325  */
326 
327 int
328 trng_add(char *str)
329 {
330 	TIME_RNG *pt;
331 	char *up_pt = NULL;
332 	char *stpt;
333 	char *flgpt;
334 	int dot = 0;
335 
336 	/*
337 	 * throw out the badly formed time ranges
338 	 */
339 	if ((str == NULL) || (*str == '\0')) {
340 		paxwarn(1, "Empty time range string");
341 		return(-1);
342 	}
343 
344 	/*
345 	 * locate optional flags suffix /{cm}.
346 	 */
347 	if ((flgpt = strrchr(str, '/')) != NULL)
348 		*flgpt++ = '\0';
349 
350 	for (stpt = str; *stpt != '\0'; ++stpt) {
351 		if ((*stpt >= '0') && (*stpt <= '9'))
352 			continue;
353 		if ((*stpt == ',') && (up_pt == NULL)) {
354 			*stpt = '\0';
355 			up_pt = stpt + 1;
356 			dot = 0;
357 			continue;
358 		}
359 
360 		/*
361 		 * allow only one dot per range (secs)
362 		 */
363 		if ((*stpt == '.') && (!dot)) {
364 			++dot;
365 			continue;
366 		}
367 		paxwarn(1, "Improperly specified time range: %s", str);
368 		goto out;
369 	}
370 
371 	/*
372 	 * allocate space for the time range and store the limits
373 	 */
374 	if ((pt = (TIME_RNG *)malloc(sizeof(TIME_RNG))) == NULL) {
375 		paxwarn(1, "Unable to allocate memory for time range");
376 		return(-1);
377 	}
378 
379 	/*
380 	 * by default we only will check file mtime, but the user can specify
381 	 * mtime, ctime (inode change time) or both.
382 	 */
383 	if ((flgpt == NULL) || (*flgpt == '\0'))
384 		pt->flgs = CMPMTME;
385 	else {
386 		pt->flgs = 0;
387 		while (*flgpt != '\0') {
388 			switch(*flgpt) {
389 			case 'M':
390 			case 'm':
391 				pt->flgs |= CMPMTME;
392 				break;
393 			case 'C':
394 			case 'c':
395 				pt->flgs |= CMPCTME;
396 				break;
397 			default:
398 				paxwarn(1, "Bad option %c with time range %s",
399 				    *flgpt, str);
400 				free(pt);
401 				goto out;
402 			}
403 			++flgpt;
404 		}
405 	}
406 
407 	/*
408 	 * start off with the current time
409 	 */
410 	pt->low_time = pt->high_time = time(NULL);
411 	if (*str != '\0') {
412 		/*
413 		 * add lower limit
414 		 */
415 		if (str_sec(str, &(pt->low_time)) < 0) {
416 			paxwarn(1, "Illegal lower time range %s", str);
417 			free(pt);
418 			goto out;
419 		}
420 		pt->flgs |= HASLOW;
421 	}
422 
423 	if ((up_pt != NULL) && (*up_pt != '\0')) {
424 		/*
425 		 * add upper limit
426 		 */
427 		if (str_sec(up_pt, &(pt->high_time)) < 0) {
428 			paxwarn(1, "Illegal upper time range %s", up_pt);
429 			free(pt);
430 			goto out;
431 		}
432 		pt->flgs |= HASHIGH;
433 
434 		/*
435 		 * check that the upper and lower do not overlap
436 		 */
437 		if (pt->flgs & HASLOW) {
438 			if (pt->low_time > pt->high_time) {
439 				paxwarn(1, "Upper %s and lower %s time overlap",
440 					up_pt, str);
441 				free(pt);
442 				return(-1);
443 			}
444 		}
445 	}
446 
447 	pt->fow = NULL;
448 	if (trhead == NULL) {
449 		trtail = trhead = pt;
450 		return(0);
451 	}
452 	trtail->fow = pt;
453 	trtail = pt;
454 	return(0);
455 
456     out:
457 	paxwarn(1, "Time range format is: [yy[mm[dd[hh]]]]mm[.ss][/[c][m]]");
458 	return(-1);
459 }
460 
461 /*
462  * trng_match()
463  *	check if this files mtime/ctime falls within any supplied time range.
464  * Return:
465  *	0 if this archive member should be processed, 1 if it should be skipped
466  */
467 
468 static int
469 trng_match(ARCHD *arcn)
470 {
471 	TIME_RNG *pt;
472 
473 	/*
474 	 * have to search down the list one at a time looking for a match.
475 	 * remember time range limits are inclusive.
476 	 */
477 	pt = trhead;
478 	while (pt != NULL) {
479 		switch(pt->flgs & CMPBOTH) {
480 		case CMPBOTH:
481 			/*
482 			 * user wants both mtime and ctime checked for this
483 			 * time range
484 			 */
485 			if (((pt->flgs & HASLOW) &&
486 			    (arcn->sb.st_mtime < pt->low_time) &&
487 			    (arcn->sb.st_ctime < pt->low_time)) ||
488 			    ((pt->flgs & HASHIGH) &&
489 			    (arcn->sb.st_mtime > pt->high_time) &&
490 			    (arcn->sb.st_ctime > pt->high_time))) {
491 				pt = pt->fow;
492 				continue;
493 			}
494 			break;
495 		case CMPCTME:
496 			/*
497 			 * user wants only ctime checked for this time range
498 			 */
499 			if (((pt->flgs & HASLOW) &&
500 			    (arcn->sb.st_ctime < pt->low_time)) ||
501 			    ((pt->flgs & HASHIGH) &&
502 			    (arcn->sb.st_ctime > pt->high_time))) {
503 				pt = pt->fow;
504 				continue;
505 			}
506 			break;
507 		case CMPMTME:
508 		default:
509 			/*
510 			 * user wants only mtime checked for this time range
511 			 */
512 			if (((pt->flgs & HASLOW) &&
513 			    (arcn->sb.st_mtime < pt->low_time)) ||
514 			    ((pt->flgs & HASHIGH) &&
515 			    (arcn->sb.st_mtime > pt->high_time))) {
516 				pt = pt->fow;
517 				continue;
518 			}
519 			break;
520 		}
521 		break;
522 	}
523 
524 	if (pt == NULL)
525 		return(1);
526 	return(0);
527 }
528 
529 /*
530  * str_sec()
531  *	Convert a time string in the format of [yy[mm[dd[hh]]]]mm[.ss] to gmt
532  *	seconds. Tval already has current time loaded into it at entry.
533  * Return:
534  *	0 if converted ok, -1 otherwise
535  */
536 
537 static int
538 str_sec(char *str, time_t *tval)
539 {
540 	struct tm *lt;
541 	char *dot = NULL;
542 
543 	lt = localtime(tval);
544 	if ((dot = strchr(str, '.')) != NULL) {
545 		/*
546 		 * seconds (.ss)
547 		 */
548 		*dot++ = '\0';
549 		if (strlen(dot) != 2)
550 			return(-1);
551 		if ((lt->tm_sec = ATOI2(dot)) > 61)
552 			return(-1);
553 	} else
554 		lt->tm_sec = 0;
555 
556 	switch (strlen(str)) {
557 	case 10:
558 		/*
559 		 * year (yy)
560 		 * watch out for year 2000
561 		 */
562 		if ((lt->tm_year = ATOI2(str)) < 69)
563 			lt->tm_year += 100;
564 		str += 2;
565 		/* FALLTHROUGH */
566 	case 8:
567 		/*
568 		 * month (mm)
569 		 * watch out months are from 0 - 11 internally
570 		 */
571 		if ((lt->tm_mon = ATOI2(str)) > 12)
572 			return(-1);
573 		--lt->tm_mon;
574 		str += 2;
575 		/* FALLTHROUGH */
576 	case 6:
577 		/*
578 		 * day (dd)
579 		 */
580 		if ((lt->tm_mday = ATOI2(str)) > 31)
581 			return(-1);
582 		str += 2;
583 		/* FALLTHROUGH */
584 	case 4:
585 		/*
586 		 * hour (hh)
587 		 */
588 		if ((lt->tm_hour = ATOI2(str)) > 23)
589 			return(-1);
590 		str += 2;
591 		/* FALLTHROUGH */
592 	case 2:
593 		/*
594 		 * minute (mm)
595 		 */
596 		if ((lt->tm_min = ATOI2(str)) > 59)
597 			return(-1);
598 		break;
599 	default:
600 		return(-1);
601 	}
602 	/*
603 	 * convert broken-down time to GMT clock time seconds
604 	 */
605 	if ((*tval = mktime(lt)) == -1)
606 		return(-1);
607 	return(0);
608 }
609