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
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
|
/*-
* 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.
*
* $Id: buf_subs.c,v 1.7 1997/02/22 14:04:20 peter Exp $
*/
#ifndef lint
static char const sccsid[] = "@(#)buf_subs.c 8.2 (Berkeley) 4/18/94";
#endif /* not lint */
#include <sys/types.h>
#include <sys/time.h>
#include <sys/stat.h>
#include <sys/param.h>
#include <stdio.h>
#include <errno.h>
#include <unistd.h>
#include <stdlib.h>
#include <string.h>
#include "pax.h"
#include "extern.h"
/*
* routines which implement archive and file buffering
*/
#define MINFBSZ 512 /* default block size for hole detect */
#define MAXFLT 10 /* default media read error limit */
/*
* Need to change bufmem to dynamic allocation when the upper
* limit on blocking size is removed (though that will violate pax spec)
* MAXBLK define and tests will also need to be updated.
*/
static char bufmem[MAXBLK+BLKMULT]; /* i/o buffer + pushback id space */
static char *buf; /* normal start of i/o buffer */
static char *bufend; /* end or last char in i/o buffer */
static char *bufpt; /* read/write point in i/o buffer */
int blksz = MAXBLK; /* block input/output size in bytes */
int wrblksz; /* user spec output size in bytes */
int maxflt = MAXFLT; /* MAX consecutive media errors */
int rdblksz; /* first read blksize (tapes only) */
off_t wrlimit; /* # of bytes written per archive vol */
off_t wrcnt; /* # of bytes written on current vol */
off_t rdcnt; /* # of bytes read on current vol */
/*
* wr_start()
* set up the buffering system to operate in a write mode
* Return:
* 0 if ok, -1 if the user specified write block size violates pax spec
*/
#if __STDC__
int
wr_start(void)
#else
int
wr_start()
#endif
{
buf = &(bufmem[BLKMULT]);
/*
* Check to make sure the write block size meets pax specs. If the user
* does not specify a blocksize, we use the format default blocksize.
* We must be picky on writes, so we do not allow the user to create an
* archive that might be hard to read elsewhere. If all ok, we then
* open the first archive volume
*/
if (!wrblksz)
wrblksz = frmt->bsz;
if (wrblksz > MAXBLK) {
pax_warn(1, "Write block size of %d too large, maximium is: %d",
wrblksz, MAXBLK);
return(-1);
}
if (wrblksz % BLKMULT) {
pax_warn(1, "Write block size of %d is not a %d byte multiple",
wrblksz, BLKMULT);
return(-1);
}
/*
* we only allow wrblksz to be used with all archive operations
*/
blksz = rdblksz = wrblksz;
if ((ar_open(arcname) < 0) && (ar_next() < 0))
return(-1);
wrcnt = 0;
bufend = buf + wrblksz;
bufpt = buf;
return(0);
}
/*
* rd_start()
* set up buffering system to read an archive
* Return:
* 0 if ok, -1 otherwise
*/
#if __STDC__
int
rd_start(void)
#else
int
rd_start()
#endif
{
/*
* leave space for the header pushback (see get_arc()). If we are
* going to append and user specified a write block size, check it
* right away
*/
buf = &(bufmem[BLKMULT]);
if ((act == APPND) && wrblksz) {
if (wrblksz > MAXBLK) {
pax_warn(1,"Write block size %d too large, maximium is: %d",
wrblksz, MAXBLK);
return(-1);
}
if (wrblksz % BLKMULT) {
pax_warn(1, "Write block size %d is not a %d byte multiple",
wrblksz, BLKMULT);
return(-1);
}
}
/*
* open the archive
*/
if ((ar_open(arcname) < 0) && (ar_next() < 0))
return(-1);
bufend = buf + rdblksz;
bufpt = bufend;
rdcnt = 0;
return(0);
}
/*
* cp_start()
* set up buffer system for copying within the file system
*/
#if __STDC__
void
cp_start(void)
#else
void
cp_start()
#endif
{
buf = &(bufmem[BLKMULT]);
rdblksz = blksz = MAXBLK;
}
/*
* appnd_start()
* Set up the buffering system to append new members to an archive that
* was just read. The last block(s) of an archive may contain a format
* specific trailer. To append a new member, this trailer has to be
* removed from the archive. The first byte of the trailer is replaced by
* the start of the header of the first file added to the archive. The
* format specific end read function tells us how many bytes to move
* backwards in the archive to be positioned BEFORE the trailer. Two
* different postions have to be adjusted, the O.S. file offset (e.g. the
* position of the tape head) and the write point within the data we have
* stored in the read (soon to become write) buffer. We may have to move
* back several records (the number depends on the size of the archive
* record and the size of the format trailer) to read up the record where
* the first byte of the trailer is recorded. Trailers may span (and
* overlap) record boundries.
* We first calculate which record has the first byte of the trailer. We
* move the OS file offset back to the start of this record and read it
* up. We set the buffer write pointer to be at this byte (the byte where
* the trailer starts). We then move the OS file pointer back to the
* start of this record so a flush of this buffer will replace the record
* in the archive.
* A major problem is rewriting this last record. For archives stored
* on disk files, this is trival. However, many devices are really picky
* about the conditions under which they will allow a write to occur.
* Often devices restrict the conditions where writes can be made writes,
* so it may not be feasable to append archives stored on all types of
* devices.
* Return:
* 0 for success, -1 for failure
*/
#if __STDC__
int
appnd_start(off_t skcnt)
#else
int
appnd_start(skcnt)
off_t skcnt;
#endif
{
register int res;
off_t cnt;
if (exit_val != 0) {
pax_warn(0, "Cannot append to an archive that may have flaws.");
return(-1);
}
/*
* if the user did not specify a write blocksize, inherit the size used
* in the last archive volume read. (If a is set we still use rdblksz
* until next volume, cannot shift sizes within a single volume).
*/
if (!wrblksz)
wrblksz = blksz = rdblksz;
else
blksz = rdblksz;
/*
* make sure that this volume allows appends
*/
if (ar_app_ok() < 0)
return(-1);
/*
* Calculate bytes to move back and move in front of record where we
* need to start writing from. Remember we have to add in any padding
* that might be in the buffer after the trailer in the last block. We
* travel skcnt + padding ROUNDED UP to blksize.
*/
skcnt += bufend - bufpt;
if ((cnt = (skcnt/blksz) * blksz) < skcnt)
cnt += blksz;
if (ar_rev((off_t)cnt) < 0)
goto out;
/*
* We may have gone too far if there is valid data in the block we are
* now in front of, read up the block and position the pointer after
* the valid data.
*/
if ((cnt -= skcnt) > 0) {
/*
* watch out for stupid tape drives. ar_rev() will set rdblksz
* to be real physical blocksize so we must loop until we get
* the old rdblksz (now in blksz). If ar_rev() fouls up the
* determination of the physical block size, we will fail.
*/
bufpt = buf;
bufend = buf + blksz;
while (bufpt < bufend) {
if ((res = ar_read(bufpt, rdblksz)) <= 0)
goto out;
bufpt += res;
}
if (ar_rev((off_t)(bufpt - buf)) < 0)
goto out;
bufpt = buf + cnt;
bufend = buf + blksz;
} else {
/*
* buffer is empty
*/
bufend = buf + blksz;
bufpt = buf;
}
rdblksz = blksz;
rdcnt -= skcnt;
wrcnt = 0;
/*
* At this point we are ready to write. If the device requires special
* handling to write at a point were previously recorded data resides,
* that is handled in ar_set_wr(). From now on we operate under normal
* ARCHIVE mode (write) conditions
*/
if (ar_set_wr() < 0)
return(-1);
act = ARCHIVE;
return(0);
out:
pax_warn(1, "Unable to rewrite archive trailer, cannot append.");
return(-1);
}
/*
* rd_sync()
* A read error occurred on this archive volume. Resync the buffer and
* try to reset the device (if possible) so we can continue to read. Keep
* trying to do this until we get a valid read, or we reach the limit on
* consecutive read faults (at which point we give up). The user can
* adjust the read error limit through a command line option.
* Returns:
* 0 on success, and -1 on failure
*/
#if __STDC__
int
rd_sync(void)
#else
int
rd_sync()
#endif
{
register int errcnt = 0;
register int res;
/*
* if the user says bail out on first fault, we are out of here...
*/
if (maxflt == 0)
return(-1);
if (act == APPND) {
pax_warn(1, "Unable to append when there are archive read errors.");
return(-1);
}
/*
* poke at device and try to get past media error
*/
if (ar_rdsync() < 0) {
if (ar_next() < 0)
return(-1);
else
rdcnt = 0;
}
for (;;) {
if ((res = ar_read(buf, blksz)) > 0) {
/*
* All right! got some data, fill that buffer
*/
bufpt = buf;
bufend = buf + res;
rdcnt += res;
return(0);
}
/*
* Oh well, yet another failed read...
* if error limit reached, ditch. o.w. poke device to move past
* bad media and try again. if media is badly damaged, we ask
* the poor (and upset user at this point) for the next archive
* volume. remember the goal on reads is to get the most we
* can extract out of the archive.
*/
if ((maxflt > 0) && (++errcnt > maxflt))
pax_warn(0,"Archive read error limit (%d) reached",maxflt);
else if (ar_rdsync() == 0)
continue;
if (ar_next() < 0)
break;
rdcnt = 0;
errcnt = 0;
}
return(-1);
}
/*
* pback()
* push the data used during the archive id phase back into the I/O
* buffer. This is required as we cannot be sure that the header does NOT
* overlap a block boundry (as in the case we are trying to recover a
* flawed archived). This was not designed to be used for any other
* purpose. (What software engineering, HA!)
* WARNING: do not even THINK of pback greater than BLKMULT, unless the
* pback space is increased.
*/
#if __STDC__
void
pback(char *pt, int cnt)
#else
void
pback(pt, cnt)
char *pt;
int cnt;
#endif
{
bufpt -= cnt;
bcopy(pt, bufpt, cnt);
return;
}
/*
* rd_skip()
* skip foward in the archive during a archive read. Used to get quickly
* past file data and padding for files the user did NOT select.
* Return:
* 0 if ok, -1 failure, and 1 when EOF on the archive volume was detected.
*/
#if __STDC__
int
rd_skip(off_t skcnt)
#else
int
rd_skip(skcnt)
off_t skcnt;
#endif
{
off_t res;
off_t cnt;
off_t skipped = 0;
/*
* consume what data we have in the buffer. If we have to move foward
* whole records, we call the low level skip function to see if we can
* move within the archive without doing the expensive reads on data we
* do not want.
*/
if (skcnt == 0)
return(0);
res = MIN((bufend - bufpt), skcnt);
bufpt += res;
skcnt -= res;
/*
* if skcnt is now 0, then no additional i/o is needed
*/
if (skcnt == 0)
return(0);
/*
* We have to read more, calculate complete and partial record reads
* based on rdblksz. we skip over "cnt" complete records
*/
res = skcnt%rdblksz;
cnt = (skcnt/rdblksz) * rdblksz;
/*
* if the skip fails, we will have to resync. ar_fow will tell us
* how much it can skip over. We will have to read the rest.
*/
if (ar_fow(cnt, &skipped) < 0)
return(-1);
res += cnt - skipped;
rdcnt += skipped;
/*
* what is left we have to read (which may be the whole thing if
* ar_fow() told us the device can only read to skip records);
*/
while (res > 0L) {
cnt = bufend - bufpt;
/*
* if the read fails, we will have to resync
*/
if ((cnt <= 0) && ((cnt = buf_fill()) < 0))
return(-1);
if (cnt == 0)
return(1);
cnt = MIN(cnt, res);
bufpt += cnt;
res -= cnt;
}
return(0);
}
/*
* wr_fin()
* flush out any data (and pad if required) the last block. We always pad
* with zero (even though we do not have to). Padding with 0 makes it a
* lot easier to recover if the archive is damaged. zero paddding SHOULD
* BE a requirement....
*/
#if __STDC__
void
wr_fin(void)
#else
void
wr_fin()
#endif
{
if (bufpt > buf) {
bzero(bufpt, bufend - bufpt);
bufpt = bufend;
(void)buf_flush(blksz);
}
}
/*
* wr_rdbuf()
* fill the write buffer from data passed to it in a buffer (usually used
* by format specific write routines to pass a file header). On failure we
* punt. We do not allow the user to continue to write flawed archives.
* We assume these headers are not very large (the memory copy we use is
* a bit expensive).
* Return:
* 0 if buffer was filled ok, -1 o.w. (buffer flush failure)
*/
#if __STDC__
int
wr_rdbuf(register char *out, register int outcnt)
#else
int
wr_rdbuf(out, outcnt)
register char *out;
register int outcnt;
#endif
{
register int cnt;
/*
* while there is data to copy copy into the write buffer. when the
* write buffer fills, flush it to the archive and continue
*/
while (outcnt > 0) {
cnt = bufend - bufpt;
if ((cnt <= 0) && ((cnt = buf_flush(blksz)) < 0))
return(-1);
/*
* only move what we have space for
*/
cnt = MIN(cnt, outcnt);
bcopy(out, bufpt, cnt);
bufpt += cnt;
out += cnt;
outcnt -= cnt;
}
return(0);
}
/*
* rd_wrbuf()
* copy from the read buffer into a supplied buffer a specified number of
* bytes. If the read buffer is empty fill it and continue to copy.
* usually used to obtain a file header for processing by a format
* specific read routine.
* Return
* number of bytes copied to the buffer, 0 indicates EOF on archive volume,
* -1 is a read error
*/
#if __STDC__
int
rd_wrbuf(register char *in, register int cpcnt)
#else
int
rd_wrbuf(in, cpcnt)
register char *in;
register int cpcnt;
#endif
{
register int res;
register int cnt;
register int incnt = cpcnt;
/*
* loop until we fill the buffer with the requested number of bytes
*/
while (incnt > 0) {
cnt = bufend - bufpt;
if ((cnt <= 0) && ((cnt = buf_fill()) <= 0)) {
/*
* read error, return what we got (or the error if
* no data was copied). The caller must know that an
* error occured and has the best knowledge what to
* do with it
*/
if ((res = cpcnt - incnt) > 0)
return(res);
return(cnt);
}
/*
* calculate how much data to copy based on whats left and
* state of buffer
*/
cnt = MIN(cnt, incnt);
bcopy(bufpt, in, cnt);
bufpt += cnt;
incnt -= cnt;
in += cnt;
}
return(cpcnt);
}
/*
* wr_skip()
* skip foward during a write. In other words add padding to the file.
* we add zero filled padding as it makes flawed archives much easier to
* recover from. the caller tells us how many bytes of padding to add
* This routine was not designed to add HUGE amount of padding, just small
* amounts (a few 512 byte blocks at most)
* Return:
* 0 if ok, -1 if there was a buf_flush failure
*/
#if __STDC__
int
wr_skip(off_t skcnt)
#else
int
wr_skip(skcnt)
off_t skcnt;
#endif
{
register int cnt;
/*
* loop while there is more padding to add
*/
while (skcnt > 0L) {
cnt = bufend - bufpt;
if ((cnt <= 0) && ((cnt = buf_flush(blksz)) < 0))
return(-1);
cnt = MIN(cnt, skcnt);
bzero(bufpt, cnt);
bufpt += cnt;
skcnt -= cnt;
}
return(0);
}
/*
* wr_rdfile()
* fill write buffer with the contents of a file. We are passed an open
* file descriptor to the file an the archive structure that describes the
* file we are storing. The variable "left" is modified to contain the
* number of bytes of the file we were NOT able to write to the archive.
* it is important that we always write EXACTLY the number of bytes that
* the format specific write routine told us to. The file can also get
* bigger, so reading to the end of file would create an improper archive,
* we just detect this case and pax_warn the user. We never create a bad
* archive if we can avoid it. Of course trying to archive files that are
* active is asking for trouble. It we fail, we pass back how much we
* could NOT copy and let the caller deal with it.
* Return:
* 0 ok, -1 if archive write failure. a short read of the file returns a
* 0, but "left" is set to be greater than zero.
*/
#if __STDC__
int
wr_rdfile(ARCHD *arcn, int ifd, off_t *left)
#else
int
wr_rdfile(arcn, ifd, left)
ARCHD *arcn;
int ifd;
off_t *left;
#endif
{
register int cnt;
register int res = 0;
register off_t size = arcn->sb.st_size;
struct stat sb;
/*
* while there are more bytes to write
*/
while (size > 0L) {
cnt = bufend - bufpt;
if ((cnt <= 0) && ((cnt = buf_flush(blksz)) < 0)) {
*left = size;
return(-1);
}
cnt = MIN(cnt, size);
if ((res = read(ifd, bufpt, cnt)) <= 0)
break;
size -= res;
bufpt += res;
}
/*
* better check the file did not change during this operation
* or the file read failed.
*/
if (res < 0)
sys_warn(1, errno, "Read fault on %s", arcn->org_name);
else if (size != 0L)
pax_warn(1, "File changed size during read %s", arcn->org_name);
else if (fstat(ifd, &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 copy to archive",
arcn->org_name);
*left = size;
return(0);
}
/*
* rd_wrfile()
* extract the contents of a file from the archive. If we are unable to
* extract the entire file (due to failure to write the file) we return
* the numbers of bytes we did NOT process. This way the caller knows how
* many bytes to skip past to find the next archive header. If the failure
* was due to an archive read, we will catch that when we try to skip. If
* the format supplies a file data crc value, we calculate the actual crc
* so that it can be compared to the value stored in the header
* NOTE:
* We call a special function to write the file. This function attempts to
* restore file holes (blocks of zeros) into the file. When files are
* sparse this saves space, and is a LOT faster. For non sparse files
* the performance hit is small. As of this writing, no archive supports
* information on where the file holes are.
* Return:
* 0 ok, -1 if archive read failure. if we cannot write the entire file,
* we return a 0 but "left" is set to be the amount unwritten
*/
#if __STDC__
int
rd_wrfile(ARCHD *arcn, int ofd, off_t *left)
#else
int
rd_wrfile(arcn, ofd, left)
ARCHD *arcn;
int ofd;
off_t *left;
#endif
{
register int cnt = 0;
register off_t size = arcn->sb.st_size;
register int res = 0;
register char *fnm = arcn->name;
int isem = 1;
int rem;
int sz = MINFBSZ;
struct stat sb;
u_long crc = 0L;
/*
* pass the blocksize of the file being written to the write routine,
* if the size is zero, use the default MINFBSZ
*/
if (fstat(ofd, &sb) == 0) {
if (sb.st_blksize > 0)
sz = (int)sb.st_blksize;
} else
sys_warn(0,errno,"Unable to obtain block size for file %s",fnm);
rem = sz;
*left = 0L;
/*
* Copy the archive to the file the number of bytes specified. We have
* to assume that we want to recover file holes as none of the archive
* formats can record the location of file holes.
*/
while (size > 0L) {
cnt = bufend - bufpt;
/*
* if we get a read error, we do not want to skip, as we may
* miss a header, so we do not set left, but if we get a write
* error, we do want to skip over the unprocessed data.
*/
if ((cnt <= 0) && ((cnt = buf_fill()) <= 0))
break;
cnt = MIN(cnt, size);
if ((res = file_write(ofd,bufpt,cnt,&rem,&isem,sz,fnm)) <= 0) {
*left = size;
break;
}
if (docrc) {
/*
* update the actual crc value
*/
cnt = res;
while (--cnt >= 0)
crc += *bufpt++ & 0xff;
} else
bufpt += res;
size -= res;
}
/*
* if the last block has a file hole (all zero), we must make sure this
* gets updated in the file. We force the last block of zeros to be
* written. just closing with the file offset moved foward may not put
* a hole at the end of the file.
*/
if (isem && (arcn->sb.st_size > 0L))
file_flush(ofd, fnm, isem);
/*
* if we failed from archive read, we do not want to skip
*/
if ((size > 0L) && (*left == 0L))
return(-1);
/*
* some formats record a crc on file data. If so, then we compare the
* calculated crc to the crc stored in the archive
*/
if (docrc && (size == 0L) && (arcn->crc != crc))
pax_warn(1,"Actual crc does not match expected crc %s",arcn->name);
return(0);
}
/*
* cp_file()
* copy the contents of one file to another. used during -rw phase of pax
* just as in rd_wrfile() we use a special write function to write the
* destination file so we can properly copy files with holes.
*/
#if __STDC__
void
cp_file(ARCHD *arcn, int fd1, int fd2)
#else
void
cp_file(arcn, fd1, fd2)
ARCHD *arcn;
int fd1;
int fd2;
#endif
{
register int cnt;
register off_t cpcnt = 0L;
register int res = 0;
register char *fnm = arcn->name;
register int no_hole = 0;
int isem = 1;
int rem;
int sz = MINFBSZ;
struct stat sb;
/*
* check for holes in the source file. If none, we will use regular
* write instead of file write.
*/
if (((off_t)(arcn->sb.st_blocks * BLKMULT)) >= arcn->sb.st_size)
++no_hole;
/*
* pass the blocksize of the file being written to the write routine,
* if the size is zero, use the default MINFBSZ
*/
if (fstat(fd2, &sb) == 0) {
if (sb.st_blksize > 0)
sz = sb.st_blksize;
} else
sys_warn(0,errno,"Unable to obtain block size for file %s",fnm);
rem = sz;
/*
* read the source file and copy to destination file until EOF
*/
for(;;) {
if ((cnt = read(fd1, buf, blksz)) <= 0)
break;
if (no_hole)
res = write(fd2, buf, cnt);
else
res = file_write(fd2, buf, cnt, &rem, &isem, sz, fnm);
if (res != cnt)
break;
cpcnt += cnt;
}
/*
* check to make sure the copy is valid.
*/
if (res < 0)
sys_warn(1, errno, "Failed write during copy of %s to %s",
arcn->org_name, arcn->name);
else if (cpcnt != arcn->sb.st_size)
pax_warn(1, "File %s changed size during copy to %s",
arcn->org_name, arcn->name);
else if (fstat(fd1, &sb) < 0)
sys_warn(1, errno, "Failed stat of %s", arcn->org_name);
else if (arcn->sb.st_mtime != sb.st_mtime)
pax_warn(1, "File %s was modified during copy to %s",
arcn->org_name, arcn->name);
/*
* if the last block has a file hole (all zero), we must make sure this
* gets updated in the file. We force the last block of zeros to be
* written. just closing with the file offset moved foward may not put
* a hole at the end of the file.
*/
if (!no_hole && isem && (arcn->sb.st_size > 0L))
file_flush(fd2, fnm, isem);
return;
}
/*
* buf_fill()
* fill the read buffer with the next record (or what we can get) from
* the archive volume.
* Return:
* Number of bytes of data in the read buffer, -1 for read error, and
* 0 when finished (user specified termination in ar_next()).
*/
#if __STDC__
int
buf_fill(void)
#else
int
buf_fill()
#endif
{
register int cnt;
static int fini = 0;
if (fini)
return(0);
for(;;) {
/*
* try to fill the buffer. on error the next archive volume is
* opened and we try again.
*/
if ((cnt = ar_read(buf, blksz)) > 0) {
bufpt = buf;
bufend = buf + cnt;
rdcnt += cnt;
return(cnt);
}
/*
* errors require resync, EOF goes to next archive
*/
if (cnt < 0)
break;
if (ar_next() < 0) {
fini = 1;
return(0);
}
rdcnt = 0;
}
exit_val = 1;
return(-1);
}
/*
* buf_flush()
* force the write buffer to the archive. We are passed the number of
* bytes in the buffer at the point of the flush. When we change archives
* the record size might change. (either larger or smaller).
* Return:
* 0 if all is ok, -1 when a write error occurs.
*/
#if __STDC__
int
buf_flush(register int bufcnt)
#else
int
buf_flush(bufcnt)
register int bufcnt;
#endif
{
register int cnt;
register int push = 0;
register int totcnt = 0;
/*
* if we have reached the user specified byte count for each archive
* volume, prompt for the next volume. (The non-standrad -R flag).
* NOTE: If the wrlimit is smaller than wrcnt, we will always write
* at least one record. We always round limit UP to next blocksize.
*/
if ((wrlimit > 0) && (wrcnt > wrlimit)) {
pax_warn(0, "User specified archive volume byte limit reached.");
if (ar_next() < 0) {
wrcnt = 0;
exit_val = 1;
return(-1);
}
wrcnt = 0;
/*
* The new archive volume might have changed the size of the
* write blocksize. if so we figure out if we need to write
* (one or more times), or if there is now free space left in
* the buffer (it is no longer full). bufcnt has the number of
* bytes in the buffer, (the blocksize, at the point we were
* CALLED). Push has the amount of "extra" data in the buffer
* if the block size has shrunk from a volume change.
*/
bufend = buf + blksz;
if (blksz > bufcnt)
return(0);
if (blksz < bufcnt)
push = bufcnt - blksz;
}
/*
* We have enough data to write at least one archive block
*/
for (;;) {
/*
* write a block and check if it all went out ok
*/
cnt = ar_write(buf, blksz);
if (cnt == blksz) {
/*
* the write went ok
*/
wrcnt += cnt;
totcnt += cnt;
if (push > 0) {
/* we have extra data to push to the front.
* check for more than 1 block of push, and if
* so we loop back to write again
*/
bcopy(bufend, buf, push);
bufpt = buf + push;
if (push >= blksz) {
push -= blksz;
continue;
}
} else
bufpt = buf;
return(totcnt);
} else if (cnt > 0) {
/*
* Oh drat we got a partial write!
* if format doesnt care about alignment let it go,
* we pax_warned the user in ar_write().... but this means
* the last record on this volume violates pax spec....
*/
totcnt += cnt;
wrcnt += cnt;
bufpt = buf + cnt;
cnt = bufcnt - cnt;
bcopy(bufpt, buf, cnt);
bufpt = buf + cnt;
if (!frmt->blkalgn || ((cnt % frmt->blkalgn) == 0))
return(totcnt);
break;
}
/*
* All done, go to next archive
*/
wrcnt = 0;
if (ar_next() < 0)
break;
/*
* The new archive volume might also have changed the block
* size. if so, figure out if we have too much or too little
* data for using the new block size
*/
bufend = buf + blksz;
if (blksz > bufcnt)
return(0);
if (blksz < bufcnt)
push = bufcnt - blksz;
}
/*
* write failed, stop pax. we must not create a bad archive!
*/
exit_val = 1;
return(-1);
}
|