xref: /freebsd/bin/pax/sel_subs.c (revision 3047fefe49f57a673de8df152c199de12ec2c6d3)
1 /*-
2  * Copyright (c) 1992 Keith Muller.
3  * Copyright (c) 1992, 1993
4  *	The Regents of the University of California.  All rights reserved.
5  *
6  * This code is derived from software contributed to Berkeley by
7  * Keith Muller of the University of California, San Diego.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  * 1. Redistributions of source code must retain the above copyright
13  *    notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 3. All advertising materials mentioning features or use of this software
18  *    must display the following acknowledgement:
19  *	This product includes software developed by the University of
20  *	California, Berkeley and its contributors.
21  * 4. Neither the name of the University nor the names of its contributors
22  *    may be used to endorse or promote products derived from this software
23  *    without specific prior written permission.
24  *
25  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35  * SUCH DAMAGE.
36  */
37 
38 #ifndef lint
39 #if 0
40 static char sccsid[] = "@(#)sel_subs.c	8.1 (Berkeley) 5/31/93";
41 #endif
42 #endif /* not lint */
43 #include <sys/cdefs.h>
44 __FBSDID("$FreeBSD$");
45 
46 #include <sys/types.h>
47 #include <sys/time.h>
48 #include <sys/stat.h>
49 #include <pwd.h>
50 #include <grp.h>
51 #include <stdio.h>
52 #include <string.h>
53 #include <strings.h>
54 #include <unistd.h>
55 #include <stdlib.h>
56 #include "pax.h"
57 #include "sel_subs.h"
58 #include "extern.h"
59 
60 static int str_sec(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 then 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 #		ifdef NET2_STAT
141 		uid = (uid_t)atoi(str+1);
142 #		else
143 		uid = (uid_t)strtoul(str+1, NULL, 10);
144 #		endif
145 	endpwent();
146 
147 	/*
148 	 * hash it and go down the hash chain (if any) looking for it
149 	 */
150 	indx = ((unsigned)uid) % USR_TB_SZ;
151 	if ((pt = usrtb[indx]) != NULL) {
152 		while (pt != NULL) {
153 			if (pt->uid == uid)
154 				return(0);
155 			pt = pt->fow;
156 		}
157 	}
158 
159 	/*
160 	 * uid is not yet in the table, add it to the front of the chain
161 	 */
162 	if ((pt = (USRT *)malloc(sizeof(USRT))) != NULL) {
163 		pt->uid = uid;
164 		pt->fow = usrtb[indx];
165 		usrtb[indx] = pt;
166 		return(0);
167 	}
168 	paxwarn(1, "User selection table out of memory");
169 	return(-1);
170 }
171 
172 /*
173  * usr_match()
174  *	check if this files uid matches a selected uid.
175  * Return:
176  *	0 if this archive member should be processed, 1 if it should be skipped
177  */
178 
179 static int
180 usr_match(ARCHD *arcn)
181 {
182 	USRT *pt;
183 
184 	/*
185 	 * hash and look for it in the table
186 	 */
187 	pt = usrtb[((unsigned)arcn->sb.st_uid) % USR_TB_SZ];
188 	while (pt != NULL) {
189 		if (pt->uid == arcn->sb.st_uid)
190 			return(0);
191 		pt = pt->fow;
192 	}
193 
194 	/*
195 	 * not found
196 	 */
197 	return(1);
198 }
199 
200 /*
201  * grp_add()
202  *	add a group match to the group match hash table
203  * Return:
204  *	0 if added ok, -1 otherwise;
205  */
206 
207 int
208 grp_add(char *str)
209 {
210 	u_int indx;
211 	GRPT *pt;
212 	struct group *gr;
213 	gid_t gid;
214 
215 	/*
216 	 * create the table if it doesn't exist
217 	 */
218 	if ((str == NULL) || (*str == '\0'))
219 		return(-1);
220 	if ((grptb == NULL) &&
221  	    ((grptb = (GRPT **)calloc(GRP_TB_SZ, sizeof(GRPT *))) == NULL)) {
222 		paxwarn(1, "Unable to allocate memory fo group selection table");
223 		return(-1);
224 	}
225 
226 	/*
227 	 * figure out user spec
228 	 */
229 	if (str[0] != '#') {
230 		/*
231 		 * it is a group name, \# escapes # as first char in group name
232 		 */
233 		if ((str[0] == '\\') && (str[1] == '#'))
234 			++str;
235 		if ((gr = getgrnam(str)) == NULL) {
236 			paxwarn(1,"Cannot determine gid for group name: %s", str);
237 			return(-1);
238 		}
239 		gid = gr->gr_gid;
240 	} else
241 #		ifdef NET2_STAT
242 		gid = (gid_t)atoi(str+1);
243 #		else
244 		gid = (gid_t)strtoul(str+1, NULL, 10);
245 #		endif
246 	endgrent();
247 
248 	/*
249 	 * hash it and go down the hash chain (if any) looking for it
250 	 */
251 	indx = ((unsigned)gid) % GRP_TB_SZ;
252 	if ((pt = grptb[indx]) != NULL) {
253 		while (pt != NULL) {
254 			if (pt->gid == gid)
255 				return(0);
256 			pt = pt->fow;
257 		}
258 	}
259 
260 	/*
261 	 * gid not in the table, add it to the front of the chain
262 	 */
263 	if ((pt = (GRPT *)malloc(sizeof(GRPT))) != NULL) {
264 		pt->gid = gid;
265 		pt->fow = grptb[indx];
266 		grptb[indx] = pt;
267 		return(0);
268 	}
269 	paxwarn(1, "Group selection table out of memory");
270 	return(-1);
271 }
272 
273 /*
274  * grp_match()
275  *	check if this files gid matches a selected gid.
276  * Return:
277  *	0 if this archive member should be processed, 1 if it should be skipped
278  */
279 
280 static int
281 grp_match(ARCHD *arcn)
282 {
283 	GRPT *pt;
284 
285 	/*
286 	 * hash and look for it in the table
287 	 */
288 	pt = grptb[((unsigned)arcn->sb.st_gid) % GRP_TB_SZ];
289 	while (pt != NULL) {
290 		if (pt->gid == arcn->sb.st_gid)
291 			return(0);
292 		pt = pt->fow;
293 	}
294 
295 	/*
296 	 * not found
297 	 */
298 	return(1);
299 }
300 
301 /*
302  * Time range selection routines
303  *
304  * Routines to handle user selection of files based on the modification and/or
305  * inode change time falling within a specified time range (the non-standard
306  * -T flag). The user may specify any number of different file time ranges.
307  * Time ranges are checked one at a time until a match is found (if at all).
308  * If the file has a mtime (and/or ctime) which lies within one of the time
309  * ranges, the file is selected. Time ranges may have a lower and/or a upper
310  * value. These ranges are inclusive. When no time ranges are supplied to pax
311  * with the -T option, all members in the archive will be selected by the time
312  * range routines. When only a lower range is supplied, only files with a
313  * mtime (and/or ctime) equal to or younger are selected. When only a upper
314  * range is supplied, only files with a mtime (and/or ctime) equal to or older
315  * are selected. When the lower time range is equal to the upper time range,
316  * only files with a mtime (or ctime) of exactly that time are selected.
317  */
318 
319 /*
320  * trng_add()
321  *	add a time range match to the time range list.
322  *	This is a non-standard pax option. Lower and upper ranges are in the
323  *	format: [yy[mm[dd[hh]]]]mm[.ss] and are comma separated.
324  *	Time ranges are based on current time, so 1234 would specify a time of
325  *	12:34 today.
326  * Return:
327  *	0 if the time range was added to the list, -1 otherwise
328  */
329 
330 int
331 trng_add(char *str)
332 {
333 	TIME_RNG *pt;
334 	char *up_pt = NULL;
335 	char *stpt;
336 	char *flgpt;
337 	int dot = 0;
338 
339 	/*
340 	 * throw out the badly formed time ranges
341 	 */
342 	if ((str == NULL) || (*str == '\0')) {
343 		paxwarn(1, "Empty time range string");
344 		return(-1);
345 	}
346 
347 	/*
348 	 * locate optional flags suffix /{cm}.
349 	 */
350 	if ((flgpt = strrchr(str, '/')) != NULL)
351 		*flgpt++ = '\0';
352 
353 	for (stpt = str; *stpt != '\0'; ++stpt) {
354 		if ((*stpt >= '0') && (*stpt <= '9'))
355 			continue;
356 		if ((*stpt == ',') && (up_pt == NULL)) {
357 			*stpt = '\0';
358 			up_pt = stpt + 1;
359 			dot = 0;
360 			continue;
361 		}
362 
363 		/*
364 		 * allow only one dot per range (secs)
365 		 */
366 		if ((*stpt == '.') && (!dot)) {
367 			++dot;
368 			continue;
369 		}
370 		paxwarn(1, "Improperly specified time range: %s", str);
371 		goto out;
372 	}
373 
374 	/*
375 	 * allocate space for the time range and store the limits
376 	 */
377 	if ((pt = (TIME_RNG *)malloc(sizeof(TIME_RNG))) == NULL) {
378 		paxwarn(1, "Unable to allocate memory for time range");
379 		return(-1);
380 	}
381 
382 	/*
383 	 * by default we only will check file mtime, but usee can specify
384 	 * mtime, ctime (inode change time) or both.
385 	 */
386 	if ((flgpt == NULL) || (*flgpt == '\0'))
387 		pt->flgs = CMPMTME;
388 	else {
389 		pt->flgs = 0;
390 		while (*flgpt != '\0') {
391 			switch(*flgpt) {
392 			case 'M':
393 			case 'm':
394 				pt->flgs |= CMPMTME;
395 				break;
396 			case 'C':
397 			case 'c':
398 				pt->flgs |= CMPCTME;
399 				break;
400 			default:
401 				paxwarn(1, "Bad option %c with time range %s",
402 				    *flgpt, str);
403 				goto out;
404 			}
405 			++flgpt;
406 		}
407 	}
408 
409 	/*
410 	 * start off with the current time
411 	 */
412 	pt->low_time = pt->high_time = time(NULL);
413 	if (*str != '\0') {
414 		/*
415 		 * add lower limit
416 		 */
417 		if (str_sec(str, &(pt->low_time)) < 0) {
418 			paxwarn(1, "Illegal lower time range %s", str);
419 			(void)free((char *)pt);
420 			goto out;
421 		}
422 		pt->flgs |= HASLOW;
423 	}
424 
425 	if ((up_pt != NULL) && (*up_pt != '\0')) {
426 		/*
427 		 * add upper limit
428 		 */
429 		if (str_sec(up_pt, &(pt->high_time)) < 0) {
430 			paxwarn(1, "Illegal upper time range %s", up_pt);
431 			(void)free((char *)pt);
432 			goto out;
433 		}
434 		pt->flgs |= HASHIGH;
435 
436 		/*
437 		 * check that the upper and lower do not overlap
438 		 */
439 		if (pt->flgs & HASLOW) {
440 			if (pt->low_time > pt->high_time) {
441 				paxwarn(1, "Upper %s and lower %s time overlap",
442 					up_pt, str);
443 				(void)free((char *)pt);
444 				return(-1);
445 			}
446 		}
447 	}
448 
449 	pt->fow = NULL;
450 	if (trhead == NULL) {
451 		trtail = trhead = pt;
452 		return(0);
453 	}
454 	trtail->fow = pt;
455 	trtail = pt;
456 	return(0);
457 
458     out:
459 	paxwarn(1, "Time range format is: [yy[mm[dd[hh]]]]mm[.ss][/[c][m]]");
460 	return(-1);
461 }
462 
463 /*
464  * trng_match()
465  *	check if this files mtime/ctime falls within any supplied time range.
466  * Return:
467  *	0 if this archive member should be processed, 1 if it should be skipped
468  */
469 
470 static int
471 trng_match(ARCHD *arcn)
472 {
473 	TIME_RNG *pt;
474 
475 	/*
476 	 * have to search down the list one at a time looking for a match.
477 	 * remember time range limits are inclusive.
478 	 */
479 	pt = trhead;
480 	while (pt != NULL) {
481 		switch(pt->flgs & CMPBOTH) {
482 		case CMPBOTH:
483 			/*
484 			 * user wants both mtime and ctime checked for this
485 			 * time range
486 			 */
487 			if (((pt->flgs & HASLOW) &&
488 			    (arcn->sb.st_mtime < pt->low_time) &&
489 			    (arcn->sb.st_ctime < pt->low_time)) ||
490 			    ((pt->flgs & HASHIGH) &&
491 			    (arcn->sb.st_mtime > pt->high_time) &&
492 			    (arcn->sb.st_ctime > pt->high_time))) {
493 				pt = pt->fow;
494 				continue;
495 			}
496 			break;
497 		case CMPCTME:
498 			/*
499 			 * user wants only ctime checked for this time range
500 			 */
501 			if (((pt->flgs & HASLOW) &&
502 			    (arcn->sb.st_ctime < pt->low_time)) ||
503 			    ((pt->flgs & HASHIGH) &&
504 			    (arcn->sb.st_ctime > pt->high_time))) {
505 				pt = pt->fow;
506 				continue;
507 			}
508 			break;
509 		case CMPMTME:
510 		default:
511 			/*
512 			 * user wants only mtime checked for this time range
513 			 */
514 			if (((pt->flgs & HASLOW) &&
515 			    (arcn->sb.st_mtime < pt->low_time)) ||
516 			    ((pt->flgs & HASHIGH) &&
517 			    (arcn->sb.st_mtime > pt->high_time))) {
518 				pt = pt->fow;
519 				continue;
520 			}
521 			break;
522 		}
523 		break;
524 	}
525 
526 	if (pt == NULL)
527 		return(1);
528 	return(0);
529 }
530 
531 /*
532  * str_sec()
533  *	Convert a time string in the format of [yy[mm[dd[hh]]]]mm[.ss] to gmt
534  *	seconds. Tval already has current time loaded into it at entry.
535  * Return:
536  *	0 if converted ok, -1 otherwise
537  */
538 
539 static int
540 str_sec(char *str, time_t *tval)
541 {
542 	struct tm *lt;
543 	char *dot = NULL;
544 
545 	lt = localtime(tval);
546 	if ((dot = strchr(str, '.')) != NULL) {
547 		/*
548 		 * seconds (.ss)
549 		 */
550 		*dot++ = '\0';
551 		if (strlen(dot) != 2)
552 			return(-1);
553 		if ((lt->tm_sec = ATOI2(dot)) > 61)
554 			return(-1);
555 	} else
556 		lt->tm_sec = 0;
557 
558 	switch (strlen(str)) {
559 	case 10:
560 		/*
561 		 * year (yy)
562 		 * watch out for year 2000
563 		 */
564 		if ((lt->tm_year = ATOI2(str)) < 69)
565 			lt->tm_year += 100;
566 		str += 2;
567 		/* FALLTHROUGH */
568 	case 8:
569 		/*
570 		 * month (mm)
571 		 * watch out months are from 0 - 11 internally
572 		 */
573 		if ((lt->tm_mon = ATOI2(str)) > 12)
574 			return(-1);
575 		--lt->tm_mon;
576 		str += 2;
577 		/* FALLTHROUGH */
578 	case 6:
579 		/*
580 		 * day (dd)
581 		 */
582 		if ((lt->tm_mday = ATOI2(str)) > 31)
583 			return(-1);
584 		str += 2;
585 		/* FALLTHROUGH */
586 	case 4:
587 		/*
588 		 * hour (hh)
589 		 */
590 		if ((lt->tm_hour = ATOI2(str)) > 23)
591 			return(-1);
592 		str += 2;
593 		/* FALLTHROUGH */
594 	case 2:
595 		/*
596 		 * minute (mm)
597 		 */
598 		if ((lt->tm_min = ATOI2(str)) > 59)
599 			return(-1);
600 		break;
601 	default:
602 		return(-1);
603 	}
604 	/*
605 	 * convert broken-down time to GMT clock time seconds
606 	 */
607 	if ((*tval = mktime(lt)) == -1)
608 		return(-1);
609 	return(0);
610 }
611