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
|
/* $OpenBSD: xinstall.c,v 1.42 2004/10/04 05:21:27 jsg Exp $ */
/* $NetBSD: xinstall.c,v 1.9 1995/12/20 10:25:17 jonathan Exp $ */
/*
* Copyright (c) 1987, 1993
* The Regents of the University of California. All rights reserved.
*
* 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. 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
static char copyright[] =
"@(#) Copyright (c) 1987, 1993\n\
The Regents of the University of California. All rights reserved.\n";
#endif /* not lint */
#ifndef lint
#if 0
static char sccsid[] = "@(#)xinstall.c 8.1 (Berkeley) 7/21/93";
#endif
static char rcsid[] = "$OpenBSD: xinstall.c,v 1.42 2004/10/04 05:21:27 jsg Exp $";
#endif /* not lint */
#include <sys/param.h>
#include <sys/wait.h>
#include <sys/mman.h>
#include <sys/stat.h>
#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <fcntl.h>
#include <grp.h>
#include <paths.h>
#include <pwd.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sysexits.h>
#include <utime.h>
#include "pathnames.h"
#define DIRECTORY 0x01 /* Tell install it's a directory. */
#define SETFLAGS 0x02 /* Tell install to set flags. */
#define NOCHANGEBITS (UF_IMMUTABLE | UF_APPEND | SF_IMMUTABLE | SF_APPEND)
#define BACKUP_SUFFIX ".old"
struct passwd *pp;
struct group *gp;
int dobackup, docompare, dodir, dopreserve, dostrip, safecopy;
int mode = S_IRWXU|S_IRGRP|S_IXGRP|S_IROTH|S_IXOTH;
char pathbuf[MAXPATHLEN], tempfile[MAXPATHLEN];
char *suffix = BACKUP_SUFFIX;
uid_t uid;
gid_t gid;
void copy(int, char *, int, char *, off_t, int);
int compare(int, const char *, size_t, int, const char *, size_t);
void install(char *, char *, u_long, u_int);
void install_dir(char *);
void strip(char *);
void usage(void);
int create_newfile(char *, struct stat *);
int create_tempfile(char *, char *, size_t);
int file_write(int, char *, size_t, int *, int *, int);
void file_flush(int, int);
int
main(int argc, char *argv[])
{
struct stat from_sb, to_sb;
void *set;
u_int32_t fset;
u_int iflags;
int ch, no_target;
char *flags, *to_name, *group = NULL, *owner = NULL;
iflags = 0;
while ((ch = getopt(argc, argv, "BbCcdf:g:m:o:pSs")) != -1)
switch((char)ch) {
case 'C':
docompare = 1;
break;
case 'B':
suffix = optarg;
/* fall through; -B implies -b */
case 'b':
dobackup = 1;
break;
case 'c':
/* For backwards compatibility. */
break;
case 'f':
flags = optarg;
if (strtofflags(&flags, &fset, NULL))
errx(EX_USAGE, "%s: invalid flag", flags);
iflags |= SETFLAGS;
break;
case 'g':
group = optarg;
break;
case 'm':
if (!(set = setmode(optarg)))
errx(EX_USAGE, "%s: invalid file mode", optarg);
mode = getmode(set, 0);
free(set);
break;
case 'o':
owner = optarg;
break;
case 'p':
docompare = dopreserve = 1;
break;
case 'S':
safecopy = 1;
break;
case 's':
dostrip = 1;
break;
case 'd':
dodir = 1;
break;
case '?':
default:
usage();
}
argc -= optind;
argv += optind;
/* some options make no sense when creating directories */
if ((safecopy || docompare || dostrip) && dodir)
usage();
/* must have at least two arguments, except when creating directories */
if (argc < 2 && !dodir)
usage();
/* need to make a temp copy so we can compare stripped version */
if (docompare && dostrip)
safecopy = 1;
/* get group and owner id's */
if (group && !(gp = getgrnam(group)) && !isdigit(*group))
errx(EX_NOUSER, "unknown group %s", group);
gid = (group) ? ((gp) ? gp->gr_gid : (gid_t)strtoul(group, NULL, 10)) : (gid_t)-1;
if (owner && !(pp = getpwnam(owner)) && !isdigit(*owner))
errx(EX_NOUSER, "unknown user %s", owner);
uid = (owner) ? ((pp) ? pp->pw_uid : (uid_t)strtoul(owner, NULL, 10)) : (uid_t)-1;
if (dodir) {
for (; *argv != NULL; ++argv)
install_dir(*argv);
exit(EX_OK);
/* NOTREACHED */
}
no_target = stat(to_name = argv[argc - 1], &to_sb);
if (!no_target && S_ISDIR(to_sb.st_mode)) {
for (; *argv != to_name; ++argv)
install(*argv, to_name, fset, iflags | DIRECTORY);
exit(EX_OK);
/* NOTREACHED */
}
/* can't do file1 file2 directory/file */
if (argc != 2)
errx(EX_OSERR, "Target: %s", argv[argc-1]);
if (!no_target) {
if (stat(*argv, &from_sb))
err(EX_OSERR, "%s", *argv);
if (!S_ISREG(to_sb.st_mode))
errx(EX_OSERR, "%s: %s", to_name, strerror(EFTYPE));
if (to_sb.st_dev == from_sb.st_dev &&
to_sb.st_ino == from_sb.st_ino)
errx(EX_USAGE, "%s and %s are the same file", *argv, to_name);
}
install(*argv, to_name, fset, iflags);
exit(EX_OK);
/* NOTREACHED */
}
/*
* install --
* build a path name and install the file
*/
void
install(char *from_name, char *to_name, u_long fset, u_int flags)
{
struct stat from_sb, to_sb;
struct utimbuf utb;
int devnull, from_fd, to_fd, serrno, files_match = 0;
char *p;
(void)memset((void *)&from_sb, 0, sizeof(from_sb));
(void)memset((void *)&to_sb, 0, sizeof(to_sb));
/* If try to install NULL file to a directory, fails. */
if (flags & DIRECTORY || strcmp(from_name, _PATH_DEVNULL)) {
if (stat(from_name, &from_sb))
err(EX_OSERR, "%s", from_name);
if (!S_ISREG(from_sb.st_mode))
errx(EX_OSERR, "%s: %s", from_name, strerror(EFTYPE));
/* Build the target path. */
if (flags & DIRECTORY) {
(void)snprintf(pathbuf, sizeof(pathbuf), "%s/%s",
to_name,
(p = strrchr(from_name, '/')) ? ++p : from_name);
to_name = pathbuf;
}
devnull = 0;
} else {
devnull = 1;
}
if (stat(to_name, &to_sb) == 0) {
/* Only compare against regular files. */
if (docompare && !S_ISREG(to_sb.st_mode)) {
docompare = 0;
warnx("%s: %s", to_name, strerror(EFTYPE));
}
} else if (docompare) {
/* File does not exist so silently ignore compare flag. */
docompare = 0;
}
if (safecopy) {
to_fd = create_tempfile(to_name, tempfile, sizeof(tempfile));
if (to_fd < 0)
err(EX_OSERR, "%s", tempfile);
} else if (docompare && !dostrip) {
if ((to_fd = open(to_name, O_RDONLY, 0)) < 0)
err(EX_OSERR, "%s", to_name);
} else {
if ((to_fd = create_newfile(to_name, &to_sb)) < 0)
err(EX_OSERR, "%s", to_name);
}
if (!devnull) {
if ((from_fd = open(from_name, O_RDONLY, 0)) < 0) {
serrno = errno;
(void)unlink(safecopy ? tempfile : to_name);
errx(EX_OSERR, "%s: %s", from_name, strerror(serrno));
}
if (docompare && !safecopy) {
files_match = !(compare(from_fd, from_name,
(size_t)from_sb.st_size, to_fd,
to_name, (size_t)to_sb.st_size));
/* Truncate "to" file for copy unless we match */
if (!files_match) {
(void)close(to_fd);
if ((to_fd = create_newfile(to_name, &to_sb)) < 0)
err(EX_OSERR, "%s", to_name);
}
}
if (!files_match)
copy(from_fd, from_name, to_fd,
safecopy ? tempfile : to_name, from_sb.st_size,
((off_t)from_sb.st_blocks * S_BLKSIZE < from_sb.st_size));
}
if (dostrip) {
strip(safecopy ? tempfile : to_name);
/*
* Re-open our fd on the target, in case we used a strip
* that does not work in-place -- like gnu binutils strip.
*/
close(to_fd);
if ((to_fd = open(safecopy ? tempfile : to_name, O_RDONLY,
0)) < 0)
err(EX_OSERR, "stripping %s", to_name);
}
/*
* Compare the (possibly stripped) temp file to the target.
*/
if (safecopy && docompare) {
int temp_fd = to_fd;
struct stat temp_sb;
/* Re-open to_fd using the real target name. */
if ((to_fd = open(to_name, O_RDONLY, 0)) < 0)
err(EX_OSERR, "%s", to_name);
if (fstat(temp_fd, &temp_sb)) {
serrno = errno;
(void)unlink(tempfile);
errx(EX_OSERR, "%s: %s", tempfile, strerror(serrno));
}
if (compare(temp_fd, tempfile, (size_t)temp_sb.st_size, to_fd,
to_name, (size_t)to_sb.st_size) == 0) {
/*
* If target has more than one link we need to
* replace it in order to snap the extra links.
* Need to preserve target file times, though.
*/
if (to_sb.st_nlink != 1) {
utb.actime = to_sb.st_atime;
utb.modtime = to_sb.st_mtime;
(void)utime(tempfile, &utb);
} else {
files_match = 1;
(void)unlink(tempfile);
}
(void) close(temp_fd);
}
}
/*
* Move the new file into place if doing a safe copy
* and the files are different (or just not compared).
*/
if (safecopy && !files_match) {
/* Try to turn off the immutable bits. */
if (to_sb.st_flags & (NOCHANGEBITS))
(void)chflags(to_name, to_sb.st_flags & ~(NOCHANGEBITS));
if (dobackup) {
char backup[MAXPATHLEN];
(void)snprintf(backup, MAXPATHLEN, "%s%s", to_name,
suffix);
/* It is ok for the target file not to exist. */
if (rename(to_name, backup) < 0 && errno != ENOENT) {
serrno = errno;
unlink(tempfile);
errx(EX_OSERR, "rename: %s to %s: %s", to_name,
backup, strerror(serrno));
}
}
if (rename(tempfile, to_name) < 0 ) {
serrno = errno;
unlink(tempfile);
errx(EX_OSERR, "rename: %s to %s: %s", tempfile,
to_name, strerror(serrno));
}
/* Re-open to_fd so we aren't hosed by the rename(2). */
(void) close(to_fd);
if ((to_fd = open(to_name, O_RDONLY, 0)) < 0)
err(EX_OSERR, "%s", to_name);
}
/*
* Preserve the timestamp of the source file if necessary.
*/
if (dopreserve && !files_match) {
utb.actime = from_sb.st_atime;
utb.modtime = from_sb.st_mtime;
(void)utime(to_name, &utb);
}
/*
* Set owner, group, mode for target; do the chown first,
* chown may lose the setuid bits.
*/
if ((gid != (gid_t)-1 || uid != (uid_t)-1) && fchown(to_fd, uid, gid)) {
serrno = errno;
(void)unlink(to_name);
errx(EX_OSERR, "%s: chown/chgrp: %s", to_name, strerror(serrno));
}
if (fchmod(to_fd, mode)) {
serrno = errno;
(void)unlink(to_name);
errx(EX_OSERR, "%s: chmod: %s", to_name, strerror(serrno));
}
/*
* If provided a set of flags, set them, otherwise, preserve the
* flags, except for the dump flag.
*/
if (fchflags(to_fd,
flags & SETFLAGS ? fset : from_sb.st_flags & ~UF_NODUMP)) {
if (errno != EOPNOTSUPP || (from_sb.st_flags & ~UF_NODUMP) != 0)
warnx("%s: chflags: %s", to_name, strerror(errno));
}
(void)close(to_fd);
if (!devnull)
(void)close(from_fd);
}
/*
* copy --
* copy from one file to another
*/
void
copy(int from_fd, char *from_name, int to_fd, char *to_name, off_t size,
int sparse)
{
ssize_t nr, nw;
int serrno;
char *p, buf[MAXBSIZE];
/* Rewind file descriptors. */
if (lseek(from_fd, (off_t)0, SEEK_SET) == (off_t)-1)
err(EX_OSERR, "lseek: %s", from_name);
if (lseek(to_fd, (off_t)0, SEEK_SET) == (off_t)-1)
err(EX_OSERR, "lseek: %s", to_name);
/*
* Mmap and write if less than 8M (the limit is so we don't totally
* trash memory on big files. This is really a minor hack, but it
* wins some CPU back. Sparse files need special treatment.
*/
if (!sparse && size <= 8 * 1048576) {
volatile size_t siz;
if ((p = mmap(NULL, (size_t)size, PROT_READ, MAP_PRIVATE,
from_fd, (off_t)0)) == MAP_FAILED) {
serrno = errno;
(void)unlink(to_name);
errx(EX_OSERR, "%s: %s", from_name, strerror(serrno));
}
if (size)
madvise(p, size, MADV_SEQUENTIAL);
siz = (size_t)size;
if ((nw = write(to_fd, p, siz)) != siz) {
serrno = errno;
(void)unlink(to_name);
errx(EX_OSERR, "%s: %s",
to_name, strerror(nw > 0 ? EIO : serrno));
}
(void) munmap(p, (size_t)size);
} else {
int sz, rem, isem = 1;
struct stat sb;
/*
* Pass the blocksize of the file being written to the write
* routine. if the size is zero, use the default S_BLKSIZE.
*/
if (fstat(to_fd, &sb) != 0 || sb.st_blksize == 0)
sz = S_BLKSIZE;
else
sz = sb.st_blksize;
rem = sz;
while ((nr = read(from_fd, buf, sizeof(buf))) > 0) {
if (sparse)
nw = file_write(to_fd, buf, nr, &rem, &isem, sz);
else
nw = write(to_fd, buf, nr);
if (nw != nr) {
serrno = errno;
(void)unlink(to_name);
errx(EX_OSERR, "%s: %s",
to_name, strerror(nw > 0 ? EIO : serrno));
}
}
if (sparse)
file_flush(to_fd, isem);
if (nr != 0) {
serrno = errno;
(void)unlink(to_name);
errx(EX_OSERR, "%s: %s", from_name, strerror(serrno));
}
}
}
/*
* compare --
* compare two files; non-zero means files differ
*/
int
compare(int from_fd, const char *from_name, size_t from_len, int to_fd,
const char *to_name, size_t to_len)
{
caddr_t p1, p2;
size_t length, remainder;
off_t from_off, to_off;
int dfound;
if (from_len != to_len)
return(1);
/*
* Compare the two files being careful not to mmap
* more than 8M at a time.
*/
from_off = to_off = (off_t)0;
remainder = from_len;
do {
length = MIN(remainder, 8 * 1048576);
remainder -= length;
if ((p1 = mmap(NULL, length, PROT_READ, MAP_PRIVATE,
from_fd, from_off)) == MAP_FAILED)
err(EX_OSERR, "%s", from_name);
if ((p2 = mmap(NULL, length, PROT_READ, MAP_PRIVATE,
to_fd, to_off)) == MAP_FAILED)
err(EX_OSERR, "%s", to_name);
if (length) {
madvise(p1, length, MADV_SEQUENTIAL);
madvise(p2, length, MADV_SEQUENTIAL);
}
dfound = memcmp(p1, p2, length);
(void) munmap(p1, length);
(void) munmap(p2, length);
from_off += length;
to_off += length;
} while (!dfound && remainder > 0);
return(dfound);
}
/*
* strip --
* use strip(1) to strip the target file
*/
void
strip(char *to_name)
{
int serrno, status;
char * volatile path_strip;
if (issetugid() || (path_strip = getenv("STRIP")) == NULL)
path_strip = _PATH_STRIP;
switch (vfork()) {
case -1:
serrno = errno;
(void)unlink(to_name);
errx(EX_TEMPFAIL, "forks: %s", strerror(serrno));
case 0:
execl(path_strip, "strip", to_name, (char *)NULL);
warn("%s", path_strip);
_exit(EX_OSERR);
default:
if (wait(&status) == -1 || !WIFEXITED(status))
(void)unlink(to_name);
}
}
/*
* install_dir --
* build directory hierarchy
*/
void
install_dir(char *path)
{
char *p;
struct stat sb;
int ch;
for (p = path;; ++p)
if (!*p || (p != path && *p == '/')) {
ch = *p;
*p = '\0';
if (stat(path, &sb)) {
if (errno != ENOENT || mkdir(path, 0777) < 0) {
err(EX_OSERR, "%s", path);
/* NOTREACHED */
}
}
if (!(*p = ch))
break;
}
if (((gid != (gid_t)-1 || uid != (uid_t)-1) && chown(path, uid, gid)) ||
chmod(path, mode)) {
warn("%s", path);
}
}
/*
* usage --
* print a usage message and die
*/
void
usage(void)
{
(void)fprintf(stderr, "\
usage: install [-bCcdpSs] [-B suffix] [-f flags] [-g group] [-m mode] [-o owner]\n source [...] target [...]\n");
exit(EX_USAGE);
/* NOTREACHED */
}
/*
* create_tempfile --
* create a temporary file based on path and open it
*/
int
create_tempfile(char *path, char *temp, size_t tsize)
{
char *p;
(void)strncpy(temp, path, tsize);
temp[tsize - 1] = '\0';
if ((p = strrchr(temp, '/')))
p++;
else
p = temp;
(void)strncpy(p, "INS@XXXXXXXXXX", &temp[tsize - 1] - p);
temp[tsize - 1] = '\0';
return(mkstemp(temp));
}
/*
* create_newfile --
* create a new file, overwriting an existing one if necessary
*/
int
create_newfile(char *path, struct stat *sbp)
{
char backup[MAXPATHLEN];
/*
* Unlink now... avoid ETXTBSY errors later. Try and turn
* off the append/immutable bits -- if we fail, go ahead,
* it might work.
*/
if (sbp->st_flags & (NOCHANGEBITS))
(void)chflags(path, sbp->st_flags & ~(NOCHANGEBITS));
if (dobackup) {
(void)snprintf(backup, MAXPATHLEN, "%s%s", path, suffix);
/* It is ok for the target file not to exist. */
if (rename(path, backup) < 0 && errno != ENOENT)
err(EX_OSERR, "rename: %s to %s (errno %d)", path, backup, errno);
} else
(void)unlink(path);
return(open(path, O_CREAT | O_RDWR | O_TRUNC, S_IRUSR | S_IWUSR));
}
/*
* 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.
*/
int
file_write(int fd, char *str, size_t cnt, int *rem, int *isempt, int sz)
{
char *pt;
char *end;
size_t wcnt;
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) {
warn("lseek");
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) {
warn("write");
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).
*/
void
file_flush(int fd, int isempt)
{
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) {
warn("Failed seek on file");
return;
}
if (write(fd, blnk, 1) < 0)
warn("Failed write to file");
return;
}
|