1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
|
/*-
* Copyright (c) 1992 Keith Muller.
* Copyright (c) 1992, 1993
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Keith Muller of the University of California, San Diego.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#ifndef lint
#if 0
static char sccsid[] = "@(#)file_subs.c 8.1 (Berkeley) 5/31/93";
#endif
static const char rcsid[] =
"$Id: file_subs.c,v 1.10 1998/05/15 06:27:39 charnier Exp $";
#endif /* not lint */
#include <sys/types.h>
#include <sys/time.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <string.h>
#include <stdio.h>
#include <errno.h>
#include <sys/uio.h>
#include <stdlib.h>
#include "pax.h"
#include "extern.h"
static int
mk_link __P((register char *,register struct stat *,register char *, int));
/*
* routines that deal with file operations such as: creating, removing;
* and setting access modes, uid/gid and times of files
*/
#define FILEBITS (S_ISVTX | S_IRWXU | S_IRWXG | S_IRWXO)
#define SETBITS (S_ISUID | S_ISGID)
#define ABITS (FILEBITS | SETBITS)
/*
* file_creat()
* Create and open a file.
* Return:
* file descriptor or -1 for failure
*/
#if __STDC__
int
file_creat(register ARCHD *arcn)
#else
int
file_creat(arcn)
register ARCHD *arcn;
#endif
{
int fd = -1;
mode_t file_mode;
int oerrno;
/*
* assume file doesn't exist, so just try to create it, most times this
* works. We have to take special handling when the file does exist. To
* detect this, we use O_EXCL. For example when trying to create a
* file and a character device or fifo exists with the same name, we
* can accidently open the device by mistake (or block waiting to open)
* If we find that the open has failed, then figure spend the effort to
* figure out why. This strategy was found to have better average
* performance in common use than checking the file (and the path)
* first with lstat.
*/
file_mode = arcn->sb.st_mode & FILEBITS;
if ((fd = open(arcn->name, O_WRONLY | O_CREAT | O_TRUNC | O_EXCL,
file_mode)) >= 0)
return(fd);
/*
* the file seems to exist. First we try to get rid of it (found to be
* the second most common failure when traced). If this fails, only
* then we go to the expense to check and create the path to the file
*/
if (unlnk_exist(arcn->name, arcn->type) != 0)
return(-1);
for (;;) {
/*
* try to open it again, if this fails, check all the nodes in
* the path and give it a final try. if chk_path() finds that
* it cannot fix anything, we will skip the last attempt
*/
if ((fd = open(arcn->name, O_WRONLY | O_CREAT | O_TRUNC,
file_mode)) >= 0)
break;
oerrno = errno;
if (chk_path(arcn->name,arcn->sb.st_uid,arcn->sb.st_gid) < 0) {
sys_warn(1, oerrno, "Unable to create %s", arcn->name);
return(-1);
}
}
return(fd);
}
/*
* file_close()
* Close file descriptor to a file just created by pax. Sets modes,
* ownership and times as required.
* Return:
* 0 for success, -1 for failure
*/
#if __STDC__
void
file_close(register ARCHD *arcn, int fd)
#else
void
file_close(arcn, fd)
register ARCHD *arcn;
int fd;
#endif
{
int res = 0;
if (fd < 0)
return;
if (close(fd) < 0)
sys_warn(0, errno, "Unable to close file descriptor on %s",
arcn->name);
/*
* set owner/groups first as this may strip off mode bits we want
* then set file permission modes. Then set file access and
* modification times.
*/
if (pids)
res = set_ids(arcn->name, arcn->sb.st_uid, arcn->sb.st_gid);
/*
* IMPORTANT SECURITY NOTE:
* if not preserving mode or we cannot set uid/gid, then PROHIBIT
* set uid/gid bits
*/
if (!pmode || res)
arcn->sb.st_mode &= ~(SETBITS);
if (pmode)
set_pmode(arcn->name, arcn->sb.st_mode);
if (patime || pmtime)
set_ftime(arcn->name, arcn->sb.st_mtime, arcn->sb.st_atime, 0);
}
/*
* lnk_creat()
* Create a hard link to arcn->ln_name from arcn->name. arcn->ln_name
* must exist;
* Return:
* 0 if ok, -1 otherwise
*/
#if __STDC__
int
lnk_creat(register ARCHD *arcn)
#else
int
lnk_creat(arcn)
register ARCHD *arcn;
#endif
{
struct stat sb;
/*
* we may be running as root, so we have to be sure that link target
* is not a directory, so we lstat and check
*/
if (lstat(arcn->ln_name, &sb) < 0) {
sys_warn(1,errno,"Unable to link to %s from %s", arcn->ln_name,
arcn->name);
return(-1);
}
if (S_ISDIR(sb.st_mode)) {
pax_warn(1, "A hard link to the directory %s is not allowed",
arcn->ln_name);
return(-1);
}
return(mk_link(arcn->ln_name, &sb, arcn->name, 0));
}
/*
* cross_lnk()
* Create a hard link to arcn->org_name from arcn->name. Only used in copy
* with the -l flag. No pax_warning or error if this does not succeed (we will
* then just create the file)
* Return:
* 1 if copy() should try to create this file node
* 0 if cross_lnk() ok, -1 for fatal flaw (like linking to self).
*/
#if __STDC__
int
cross_lnk(register ARCHD *arcn)
#else
int
cross_lnk(arcn)
register ARCHD *arcn;
#endif
{
/*
* try to make a link to original file (-l flag in copy mode). make sure
* we do not try to link to directories in case we are running as root
* (and it might succeed).
*/
if (arcn->type == PAX_DIR)
return(1);
return(mk_link(arcn->org_name, &(arcn->sb), arcn->name, 1));
}
/*
* chk_same()
* In copy mode if we are not trying to make hard links between the src
* and destinations, make sure we are not going to overwrite ourselves by
* accident. This slows things down a little, but we have to protect all
* those people who make typing errors.
* Return:
* 1 the target does not exist, go ahead and copy
* 0 skip it file exists (-k) or may be the same as source file
*/
#if __STDC__
int
chk_same(register ARCHD *arcn)
#else
int
chk_same(arcn)
register ARCHD *arcn;
#endif
{
struct stat sb;
/*
* if file does not exist, return. if file exists and -k, skip it
* quietly
*/
if (lstat(arcn->name, &sb) < 0)
return(1);
if (kflag)
return(0);
/*
* better make sure the user does not have src == dest by mistake
*/
if ((arcn->sb.st_dev == sb.st_dev) && (arcn->sb.st_ino == sb.st_ino)) {
pax_warn(1, "Unable to copy %s, file would overwrite itself",
arcn->name);
return(0);
}
return(1);
}
/*
* mk_link()
* try to make a hard link between two files. if ign set, we do not
* complain.
* Return:
* 0 if successful (or we are done with this file but no error, such as
* finding the from file exists and the user has set -k).
* 1 when ign was set to indicates we could not make the link but we
* should try to copy/extract the file as that might work (and is an
* allowed option). -1 an error occurred.
*/
#if __STDC__
static int
mk_link(register char *to, register struct stat *to_sb, register char *from,
int ign)
#else
static int
mk_link(to, to_sb, from, ign)
register char *to;
register struct stat *to_sb;
register char *from;
int ign;
#endif
{
struct stat sb;
int oerrno;
/*
* if from file exists, it has to be unlinked to make the link. If the
* file exists and -k is set, skip it quietly
*/
if (lstat(from, &sb) == 0) {
if (kflag)
return(0);
/*
* make sure it is not the same file, protect the user
*/
if ((to_sb->st_dev==sb.st_dev)&&(to_sb->st_ino == sb.st_ino)) {
pax_warn(1, "Unable to link file %s to itself", to);
return(-1);;
}
/*
* try to get rid of the file, based on the type
*/
if (S_ISDIR(sb.st_mode)) {
if (rmdir(from) < 0) {
sys_warn(1, errno, "Unable to remove %s", from);
return(-1);
}
} else if (unlink(from) < 0) {
if (!ign) {
sys_warn(1, errno, "Unable to remove %s", from);
return(-1);
}
return(1);
}
}
/*
* from file is gone (or did not exist), try to make the hard link.
* if it fails, check the path and try it again (if chk_path() says to
* try again)
*/
for (;;) {
if (link(to, from) == 0)
break;
oerrno = errno;
if (chk_path(from, to_sb->st_uid, to_sb->st_gid) == 0)
continue;
if (!ign) {
sys_warn(1, oerrno, "Could not link to %s from %s", to,
from);
return(-1);
}
return(1);
}
/*
* all right the link was made
*/
return(0);
}
/*
* node_creat()
* create an entry in the file system (other than a file or hard link).
* If successful, sets uid/gid modes and times as required.
* Return:
* 0 if ok, -1 otherwise
*/
#if __STDC__
int
node_creat(register ARCHD *arcn)
#else
int
node_creat(arcn)
register ARCHD *arcn;
#endif
{
register int res;
register int ign = 0;
register int oerrno;
register int pass = 0;
mode_t file_mode;
struct stat sb;
/*
* create node based on type, if that fails try to unlink the node and
* try again. finally check the path and try again. As noted in the
* file and link creation routines, this method seems to exhibit the
* best performance in general use workloads.
*/
file_mode = arcn->sb.st_mode & FILEBITS;
for (;;) {
switch(arcn->type) {
case PAX_DIR:
res = mkdir(arcn->name, file_mode);
if (ign)
res = 0;
break;
case PAX_CHR:
file_mode |= S_IFCHR;
res = mknod(arcn->name, file_mode, arcn->sb.st_rdev);
break;
case PAX_BLK:
file_mode |= S_IFBLK;
res = mknod(arcn->name, file_mode, arcn->sb.st_rdev);
break;
case PAX_FIF:
res = mkfifo(arcn->name, file_mode);
break;
case PAX_SCK:
/*
* Skip sockets, operation has no meaning under BSD
*/
pax_warn(0,
"%s skipped. Sockets cannot be copied or extracted",
arcn->name);
return(-1);
case PAX_SLK:
if ((res = symlink(arcn->ln_name, arcn->name)) == 0)
return(0);
break;
case PAX_CTG:
case PAX_HLK:
case PAX_HRG:
case PAX_REG:
default:
/*
* we should never get here
*/
pax_warn(0, "%s has an unknown file type, skipping",
arcn->name);
return(-1);
}
/*
* if we were able to create the node break out of the loop,
* otherwise try to unlink the node and try again. if that
* fails check the full path and try a final time.
*/
if (res == 0)
break;
/*
* we failed to make the node
*/
oerrno = errno;
if ((ign = unlnk_exist(arcn->name, arcn->type)) < 0)
return(-1);
if (++pass <= 1)
continue;
if (chk_path(arcn->name,arcn->sb.st_uid,arcn->sb.st_gid) < 0) {
sys_warn(1, oerrno, "Could not create: %s", arcn->name);
return(-1);
}
}
/*
* we were able to create the node. set uid/gid, modes and times
*/
if (pids)
res = set_ids(arcn->name, arcn->sb.st_uid, arcn->sb.st_gid);
else
res = 0;
/*
* IMPORTANT SECURITY NOTE:
* if not preserving mode or we cannot set uid/gid, then PROHIBIT any
* set uid/gid bits
*/
if (!pmode || res)
arcn->sb.st_mode &= ~(SETBITS);
if (pmode)
set_pmode(arcn->name, arcn->sb.st_mode);
if (arcn->type == PAX_DIR) {
/*
* Dirs must be processed again at end of extract to set times
* and modes to agree with those stored in the archive. However
* to allow extract to continue, we may have to also set owner
* rights. This allows nodes in the archive that are children
* of this directory to be extracted without failure. Both time
* and modes will be fixed after the entire archive is read and
* before pax exits.
*/
if (access(arcn->name, R_OK | W_OK | X_OK) < 0) {
if (lstat(arcn->name, &sb) < 0) {
sys_warn(0, errno,"Could not access %s (stat)",
arcn->name);
set_pmode(arcn->name,file_mode | S_IRWXU);
} else {
/*
* We have to add rights to the dir, so we make
* sure to restore the mode. The mode must be
* restored AS CREATED and not as stored if
* pmode is not set.
*/
set_pmode(arcn->name,
((sb.st_mode & FILEBITS) | S_IRWXU));
if (!pmode)
arcn->sb.st_mode = sb.st_mode;
}
/*
* we have to force the mode to what was set here,
* since we changed it from the default as created.
*/
add_dir(arcn->name, arcn->nlen, &(arcn->sb), 1);
} else if (pmode || patime || pmtime)
add_dir(arcn->name, arcn->nlen, &(arcn->sb), 0);
}
if (patime || pmtime)
set_ftime(arcn->name, arcn->sb.st_mtime, arcn->sb.st_atime, 0);
return(0);
}
/*
* unlnk_exist()
* Remove node from file system with the specified name. We pass the type
* of the node that is going to replace it. When we try to create a
* directory and find that it already exists, we allow processing to
* continue as proper modes etc will always be set for it later on.
* Return:
* 0 is ok to proceed, no file with the specified name exists
* -1 we were unable to remove the node, or we should not remove it (-k)
* 1 we found a directory and we were going to create a directory.
*/
#if __STDC__
int
unlnk_exist(register char *name, register int type)
#else
int
unlnk_exist(name, type)
register char *name;
register int type;
#endif
{
struct stat sb;
/*
* the file does not exist, or -k we are done
*/
if (lstat(name, &sb) < 0)
return(0);
if (kflag)
return(-1);
if (S_ISDIR(sb.st_mode)) {
/*
* try to remove a directory, if it fails and we were going to
* create a directory anyway, tell the caller (return a 1)
*/
if (rmdir(name) < 0) {
if (type == PAX_DIR)
return(1);
sys_warn(1,errno,"Unable to remove directory %s", name);
return(-1);
}
return(0);
}
/*
* try to get rid of all non-directory type nodes
*/
if (unlink(name) < 0) {
sys_warn(1, errno, "Could not unlink %s", name);
return(-1);
}
return(0);
}
/*
* chk_path()
* We were trying to create some kind of node in the file system and it
* failed. chk_path() makes sure the path up to the node exists and is
* writeable. When we have to create a directory that is missing along the
* path somewhere, the directory we create will be set to the same
* uid/gid as the file has (when uid and gid are being preserved).
* NOTE: this routine is a real performance loss. It is only used as a
* last resort when trying to create entries in the file system.
* Return:
* -1 when it could find nothing it is allowed to fix.
* 0 otherwise
*/
#if __STDC__
int
chk_path( register char *name, uid_t st_uid, gid_t st_gid)
#else
int
chk_path(name, st_uid, st_gid)
register char *name;
uid_t st_uid;
gid_t st_gid;
#endif
{
register char *spt = name;
struct stat sb;
int retval = -1;
/*
* watch out for paths with nodes stored directly in / (e.g. /bozo)
*/
if (*spt == '/')
++spt;
for(;;) {
/*
* work foward from the first / and check each part of the path
*/
spt = strchr(spt, '/');
if (spt == NULL)
break;
*spt = '\0';
/*
* if it exists we assume it is a directory, it is not within
* the spec (at least it seems to read that way) to alter the
* file system for nodes NOT EXPLICITLY stored on the archive.
* If that assumption is changed, you would test the node here
* and figure out how to get rid of it (probably like some
* recursive unlink()) or fix up the directory permissions if
* required (do an access()).
*/
if (lstat(name, &sb) == 0) {
*(spt++) = '/';
continue;
}
/*
* the path fails at this point, see if we can create the
* needed directory and continue on
*/
if (mkdir(name, S_IRWXU | S_IRWXG | S_IRWXO) < 0) {
*spt = '/';
retval = -1;
break;
}
/*
* we were able to create the directory. We will tell the
* caller that we found something to fix, and it is ok to try
* and create the node again.
*/
retval = 0;
if (pids)
(void)set_ids(name, st_uid, st_gid);
/*
* make sure the user doen't have some strange umask that
* causes this newly created directory to be unusable. We fix
* the modes and restore them back to the creation default at
* the end of pax
*/
if ((access(name, R_OK | W_OK | X_OK) < 0) &&
(lstat(name, &sb) == 0)) {
set_pmode(name, ((sb.st_mode & FILEBITS) | S_IRWXU));
add_dir(name, spt - name, &sb, 1);
}
*(spt++) = '/';
continue;
}
return(retval);
}
/*
* set_ftime()
* Set the access time and modification time for a named file. If frc is
* non-zero we force these times to be set even if the the user did not
* request access and/or modification time preservation (this is also
* used by -t to reset access times).
* When ign is zero, only those times the user has asked for are set, the
* other ones are left alone. We do not assume the un-documented feature
* of many utimes() implementations that consider a 0 time value as a do
* not set request.
*/
#if __STDC__
void
set_ftime(char *fnm, time_t mtime, time_t atime, int frc)
#else
void
set_ftime(fnm, mtime, atime, frc)
char *fnm;
time_t mtime;
time_t atime;
int frc;
#endif
{
static struct timeval tv[2] = {{0L, 0L}, {0L, 0L}};
struct stat sb;
tv[0].tv_sec = (long)atime;
tv[1].tv_sec = (long)mtime;
if (!frc && (!patime || !pmtime)) {
/*
* if we are not forcing, only set those times the user wants
* set. We get the current values of the times if we need them.
*/
if (lstat(fnm, &sb) == 0) {
if (!patime)
tv[0].tv_sec = (long)sb.st_atime;
if (!pmtime)
tv[1].tv_sec = (long)sb.st_mtime;
} else
sys_warn(0,errno,"Unable to obtain file stats %s", fnm);
}
/*
* set the times
*/
if (utimes(fnm, tv) < 0)
sys_warn(1, errno, "Access/modification time set failed on: %s",
fnm);
return;
}
/*
* set_ids()
* set the uid and gid of a file system node
* Return:
* 0 when set, -1 on failure
*/
#if __STDC__
int
set_ids(char *fnm, uid_t uid, gid_t gid)
#else
int
set_ids(fnm, uid, gid)
char *fnm;
uid_t uid;
gid_t gid;
#endif
{
if (chown(fnm, uid, gid) < 0) {
sys_warn(1, errno, "Unable to set file uid/gid of %s", fnm);
return(-1);
}
return(0);
}
/*
* set_pmode()
* Set file access mode
*/
#if __STDC__
void
set_pmode(char *fnm, mode_t mode)
#else
void
set_pmode(fnm, mode)
char *fnm;
mode_t mode;
#endif
{
mode &= ABITS;
if (chmod(fnm, mode) < 0)
sys_warn(1, errno, "Could not set permissions on %s", fnm);
return;
}
/*
* file_write()
* Write/copy a file (during copy or archive extract). This routine knows
* how to copy files with lseek holes in it. (Which are read as file
* blocks containing all 0's but do not have any file blocks associated
* with the data). Typical examples of these are files created by dbm
* variants (.pag files). While the file size of these files are huge, the
* actual storage is quite small (the files are sparse). The problem is
* the holes read as all zeros so are probably stored on the archive that
* way (there is no way to determine if the file block is really a hole,
* we only know that a file block of all zero's can be a hole).
* At this writing, no major archive format knows how to archive files
* with holes. However, on extraction (or during copy, -rw) we have to
* deal with these files. Without detecting the holes, the files can
* consume a lot of file space if just written to disk. This replacement
* for write when passed the basic allocation size of a file system block,
* uses lseek whenever it detects the input data is all 0 within that
* file block. In more detail, the strategy is as follows:
* While the input is all zero keep doing an lseek. Keep track of when we
* pass over file block boundries. Only write when we hit a non zero
* input. once we have written a file block, we continue to write it to
* the end (we stop looking at the input). When we reach the start of the
* next file block, start checking for zero blocks again. Working on file
* block boundries significantly reduces the overhead when copying files
* that are NOT very sparse. This overhead (when compared to a write) is
* almost below the measurement resolution on many systems. Without it,
* files with holes cannot be safely copied. It does has a side effect as
* it can put holes into files that did not have them before, but that is
* not a problem since the file contents are unchanged (in fact it saves
* file space). (Except on paging files for diskless clients. But since we
* cannot determine one of those file from here, we ignore them). If this
* ever ends up on a system where CTG files are supported and the holes
* are not desired, just do a conditional test in those routines that
* call file_write() and have it call write() instead. BEFORE CLOSING THE
* FILE, make sure to call file_flush() when the last write finishes with
* an empty block. A lot of file systems will not create an lseek hole at
* the end. In this case we drop a single 0 at the end to force the
* trailing 0's in the file.
* ---Parameters---
* rem: how many bytes left in this file system block
* isempt: have we written to the file block yet (is it empty)
* sz: basic file block allocation size
* cnt: number of bytes on this write
* str: buffer to write
* Return:
* number of bytes written, -1 on write (or lseek) error.
*/
#if __STDC__
int
file_write(int fd, char *str, register int cnt, int *rem, int *isempt, int sz,
char *name)
#else
int
file_write(fd, str, cnt, rem, isempt, sz, name)
int fd;
char *str;
register int cnt;
int *rem;
int *isempt;
int sz;
char *name;
#endif
{
register char *pt;
register char *end;
register int wcnt;
register char *st = str;
/*
* while we have data to process
*/
while (cnt) {
if (!*rem) {
/*
* We are now at the start of file system block again
* (or what we think one is...). start looking for
* empty blocks again
*/
*isempt = 1;
*rem = sz;
}
/*
* only examine up to the end of the current file block or
* remaining characters to write, whatever is smaller
*/
wcnt = MIN(cnt, *rem);
cnt -= wcnt;
*rem -= wcnt;
if (*isempt) {
/*
* have not written to this block yet, so we keep
* looking for zero's
*/
pt = st;
end = st + wcnt;
/*
* look for a zero filled buffer
*/
while ((pt < end) && (*pt == '\0'))
++pt;
if (pt == end) {
/*
* skip, buf is empty so far
*/
if (lseek(fd, (off_t)wcnt, SEEK_CUR) < 0) {
sys_warn(1,errno,"File seek on %s",
name);
return(-1);
}
st = pt;
continue;
}
/*
* drat, the buf is not zero filled
*/
*isempt = 0;
}
/*
* have non-zero data in this file system block, have to write
*/
if (write(fd, st, wcnt) != wcnt) {
sys_warn(1, errno, "Failed write to file %s", name);
return(-1);
}
st += wcnt;
}
return(st - str);
}
/*
* file_flush()
* when the last file block in a file is zero, many file systems will not
* let us create a hole at the end. To get the last block with zeros, we
* write the last BYTE with a zero (back up one byte and write a zero).
*/
#if __STDC__
void
file_flush(int fd, char *fname, int isempt)
#else
void
file_flush(fd, fname, isempt)
int fd;
char *fname;
int isempt;
#endif
{
static char blnk[] = "\0";
/*
* silly test, but make sure we are only called when the last block is
* filled with all zeros.
*/
if (!isempt)
return;
/*
* move back one byte and write a zero
*/
if (lseek(fd, (off_t)-1, SEEK_CUR) < 0) {
sys_warn(1, errno, "Failed seek on file %s", fname);
return;
}
if (write(fd, blnk, 1) < 0)
sys_warn(1, errno, "Failed write to file %s", fname);
return;
}
/*
* rdfile_close()
* close a file we have beed reading (to copy or archive). If we have to
* reset access time (tflag) do so (the times are stored in arcn).
*/
#if __STDC__
void
rdfile_close(register ARCHD *arcn, register int *fd)
#else
void
rdfile_close(arcn, fd)
register ARCHD *arcn;
register int *fd;
#endif
{
/*
* make sure the file is open
*/
if (*fd < 0)
return;
(void)close(*fd);
*fd = -1;
if (!tflag)
return;
/*
* user wants last access time reset
*/
set_ftime(arcn->org_name, arcn->sb.st_mtime, arcn->sb.st_atime, 1);
return;
}
/*
* set_crc()
* read a file to calculate its crc. This is a real drag. Archive formats
* that have this, end up reading the file twice (we have to write the
* header WITH the crc before writing the file contents. Oh well...
* Return:
* 0 if was able to calculate the crc, -1 otherwise
*/
#if __STDC__
int
set_crc(register ARCHD *arcn, register int fd)
#else
int
set_crc(arcn, fd)
register ARCHD *arcn;
register int fd;
#endif
{
register int i;
register int res;
off_t cpcnt = 0L;
u_long size;
unsigned long crc = 0L;
char tbuf[FILEBLK];
struct stat sb;
if (fd < 0) {
/*
* hmm, no fd, should never happen. well no crc then.
*/
arcn->crc = 0L;
return(0);
}
if ((size = (u_long)arcn->sb.st_blksize) > (u_long)sizeof(tbuf))
size = (u_long)sizeof(tbuf);
/*
* read all the bytes we think that there are in the file. If the user
* is trying to archive an active file, forget this file.
*/
for(;;) {
if ((res = read(fd, tbuf, size)) <= 0)
break;
cpcnt += res;
for (i = 0; i < res; ++i)
crc += (tbuf[i] & 0xff);
}
/*
* safety check. we want to avoid archiving files that are active as
* they can create inconsistant archive copies.
*/
if (cpcnt != arcn->sb.st_size)
pax_warn(1, "File changed size %s", arcn->org_name);
else if (fstat(fd, &sb) < 0)
sys_warn(1, errno, "Failed stat on %s", arcn->org_name);
else if (arcn->sb.st_mtime != sb.st_mtime)
pax_warn(1, "File %s was modified during read", arcn->org_name);
else if (lseek(fd, (off_t)0L, SEEK_SET) < 0)
sys_warn(1, errno, "File rewind failed on: %s", arcn->org_name);
else {
arcn->crc = crc;
return(0);
}
return(-1);
}
|