summaryrefslogtreecommitdiff
path: root/src/basic/copy.c
blob: 00b60df4f40ffdac73f27b7fd453056f640aad42 (plain)
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
/* SPDX-License-Identifier: LGPL-2.1+ */
/***
  Copyright 2014 Lennart Poettering
***/

#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/sendfile.h>
#include <sys/stat.h>
#include <sys/xattr.h>
#include <time.h>
#include <unistd.h>

#include "alloc-util.h"
#include "btrfs-util.h"
#include "chattr-util.h"
#include "copy.h"
#include "dirent-util.h"
#include "fd-util.h"
#include "fileio.h"
#include "fs-util.h"
#include "io-util.h"
#include "macro.h"
#include "missing.h"
#include "mount-util.h"
#include "string-util.h"
#include "strv.h"
#include "time-util.h"
#include "umask-util.h"
#include "user-util.h"
#include "xattr-util.h"

#define COPY_BUFFER_SIZE (16U*1024U)

/* A safety net for descending recursively into file system trees to copy. On Linux PATH_MAX is 4096, which means the
 * deepest valid path one can build is around 2048, which we hence use as a safety net here, to not spin endlessly in
 * case of bind mount cycles and suchlike. */
#define COPY_DEPTH_MAX 2048U

static ssize_t try_copy_file_range(
                int fd_in, loff_t *off_in,
                int fd_out, loff_t *off_out,
                size_t len,
                unsigned int flags) {

        static int have = -1;
        ssize_t r;

        if (have == 0)
                return -ENOSYS;

        r = copy_file_range(fd_in, off_in, fd_out, off_out, len, flags);
        if (have < 0)
                have = r >= 0 || errno != ENOSYS;
        if (r < 0)
                return -errno;

        return r;
}

enum {
        FD_IS_NO_PIPE,
        FD_IS_BLOCKING_PIPE,
        FD_IS_NONBLOCKING_PIPE,
};

static int fd_is_nonblock_pipe(int fd) {
        struct stat st;
        int flags;

        /* Checks whether the specified file descriptor refers to a pipe, and if so if O_NONBLOCK is set. */

        if (fstat(fd, &st) < 0)
                return -errno;

        if (!S_ISFIFO(st.st_mode))
                return FD_IS_NO_PIPE;

        flags = fcntl(fd, F_GETFL);
        if (flags < 0)
                return -errno;

        return FLAGS_SET(flags, O_NONBLOCK) ? FD_IS_NONBLOCKING_PIPE : FD_IS_BLOCKING_PIPE;
}

int copy_bytes_full(
                int fdf, int fdt,
                uint64_t max_bytes,
                CopyFlags copy_flags,
                void **ret_remains,
                size_t *ret_remains_size) {

        bool try_cfr = true, try_sendfile = true, try_splice = true;
        int r, nonblock_pipe = -1;
        size_t m = SSIZE_MAX; /* that is the maximum that sendfile and c_f_r accept */

        assert(fdf >= 0);
        assert(fdt >= 0);

        /* Tries to copy bytes from the file descriptor 'fdf' to 'fdt' in the smartest possible way. Copies a maximum
         * of 'max_bytes', which may be specified as UINT64_MAX, in which no maximum is applied. Returns negative on
         * error, zero if EOF is hit before the bytes limit is hit and positive otherwise. If the copy fails for some
         * reason but we read but didn't yet write some data an ret_remains/ret_remains_size is not NULL, then it will
         * be initialized with an allocated buffer containing this "remaining" data. Note that these two parameters are
         * initialized with a valid buffer only on failure and only if there's actually data already read. Otherwise
         * these parameters if non-NULL are set to NULL. */

        if (ret_remains)
                *ret_remains = NULL;
        if (ret_remains_size)
                *ret_remains_size = 0;

        /* Try btrfs reflinks first. This only works on regular, seekable files, hence let's check the file offsets of
         * source and destination first. */
        if ((copy_flags & COPY_REFLINK)) {
                off_t foffset;

                foffset = lseek(fdf, 0, SEEK_CUR);
                if (foffset >= 0) {
                        off_t toffset;

                        toffset = lseek(fdt, 0, SEEK_CUR);
                        if (toffset >= 0) {

                                if (foffset == 0 && toffset == 0 && max_bytes == UINT64_MAX)
                                        r = btrfs_reflink(fdf, fdt); /* full file reflink */
                                else
                                        r = btrfs_clone_range(fdf, foffset, fdt, toffset, max_bytes == UINT64_MAX ? 0 : max_bytes); /* partial reflink */
                                if (r >= 0) {
                                        off_t t;

                                        /* This worked, yay! Now — to be fully correct — let's adjust the file pointers */
                                        if (max_bytes == UINT64_MAX) {

                                                /* We cloned to the end of the source file, let's position the read
                                                 * pointer there, and query it at the same time. */
                                                t = lseek(fdf, 0, SEEK_END);
                                                if (t < 0)
                                                        return -errno;
                                                if (t < foffset)
                                                        return -ESPIPE;

                                                /* Let's adjust the destination file write pointer by the same number
                                                 * of bytes. */
                                                t = lseek(fdt, toffset + (t - foffset), SEEK_SET);
                                                if (t < 0)
                                                        return -errno;

                                                return 0; /* we copied the whole thing, hence hit EOF, return 0 */
                                        } else {
                                                t = lseek(fdf, foffset + max_bytes, SEEK_SET);
                                                if (t < 0)
                                                        return -errno;

                                                t = lseek(fdt, toffset + max_bytes, SEEK_SET);
                                                if (t < 0)
                                                        return -errno;

                                                return 1; /* we copied only some number of bytes, which worked, but this means we didn't hit EOF, return 1 */
                                        }
                                }

                                log_debug_errno(r, "Reflinking didn't work, falling back to non-reflink copying: %m");
                        }
                }
        }

        for (;;) {
                ssize_t n;

                if (max_bytes <= 0)
                        return 1; /* return > 0 if we hit the max_bytes limit */

                if (max_bytes != UINT64_MAX && m > max_bytes)
                        m = max_bytes;

                /* First try copy_file_range(), unless we already tried */
                if (try_cfr) {
                        n = try_copy_file_range(fdf, NULL, fdt, NULL, m, 0u);
                        if (n < 0) {
                                if (!IN_SET(n, -EINVAL, -ENOSYS, -EXDEV, -EBADF))
                                        return n;

                                try_cfr = false;
                                /* use fallback below */
                        } else if (n == 0) /* EOF */
                                break;
                        else
                                /* Success! */
                                goto next;
                }

                /* First try sendfile(), unless we already tried */
                if (try_sendfile) {
                        n = sendfile(fdt, fdf, NULL, m);
                        if (n < 0) {
                                if (!IN_SET(errno, EINVAL, ENOSYS))
                                        return -errno;

                                try_sendfile = false;
                                /* use fallback below */
                        } else if (n == 0) /* EOF */
                                break;
                        else
                                /* Success! */
                                goto next;
                }

                /* Then try splice, unless we already tried. */
                if (try_splice) {

                        /* splice()'s asynchronous I/O support is a bit weird. When it encounters a pipe file
                         * descriptor, then it will ignore its O_NONBLOCK flag and instead only honour the
                         * SPLICE_F_NONBLOCK flag specified in its flag parameter. Let's hide this behaviour here, and
                         * check if either of the specified fds are a pipe, and if so, let's pass the flag
                         * automatically, depending on O_NONBLOCK being set.
                         *
                         * Here's a twist though: when we use it to move data between two pipes of which one has
                         * O_NONBLOCK set and the other has not, then we have no individual control over O_NONBLOCK
                         * behaviour. Hence in that case we can't use splice() and still guarantee systematic
                         * O_NONBLOCK behaviour, hence don't. */

                        if (nonblock_pipe < 0) {
                                int a, b;

                                /* Check if either of these fds is a pipe, and if so non-blocking or not */
                                a = fd_is_nonblock_pipe(fdf);
                                if (a < 0)
                                        return a;

                                b = fd_is_nonblock_pipe(fdt);
                                if (b < 0)
                                        return b;

                                if ((a == FD_IS_NO_PIPE && b == FD_IS_NO_PIPE) ||
                                    (a == FD_IS_BLOCKING_PIPE && b == FD_IS_NONBLOCKING_PIPE) ||
                                    (a == FD_IS_NONBLOCKING_PIPE && b == FD_IS_BLOCKING_PIPE))

                                        /* splice() only works if one of the fds is a pipe. If neither is, let's skip
                                         * this step right-away. As mentioned above, if one of the two fds refers to a
                                         * blocking pipe and the other to a non-blocking pipe, we can't use splice()
                                         * either, hence don't try either. This hence means we can only use splice() if
                                         * either only one of the two fds is a pipe, or if both are pipes with the same
                                         * nonblocking flag setting. */

                                        try_splice = false;
                                else
                                        nonblock_pipe = a == FD_IS_NONBLOCKING_PIPE || b == FD_IS_NONBLOCKING_PIPE;
                        }
                }

                if (try_splice) {
                        n = splice(fdf, NULL, fdt, NULL, m, nonblock_pipe ? SPLICE_F_NONBLOCK : 0);
                        if (n < 0) {
                                if (!IN_SET(errno, EINVAL, ENOSYS))
                                        return -errno;

                                try_splice = false;
                                /* use fallback below */
                        } else if (n == 0) /* EOF */
                                break;
                        else
                                /* Success! */
                                goto next;
                }

                /* As a fallback just copy bits by hand */
                {
                        uint8_t buf[MIN(m, COPY_BUFFER_SIZE)], *p = buf;
                        ssize_t z;

                        n = read(fdf, buf, sizeof buf);
                        if (n < 0)
                                return -errno;
                        if (n == 0) /* EOF */
                                break;

                        z = (size_t) n;
                        do {
                                ssize_t k;

                                k = write(fdt, p, z);
                                if (k < 0) {
                                        r = -errno;

                                        if (ret_remains) {
                                                void *copy;

                                                copy = memdup(p, z);
                                                if (!copy)
                                                        return -ENOMEM;

                                                *ret_remains = copy;
                                        }

                                        if (ret_remains_size)
                                                *ret_remains_size = z;

                                        return r;
                                }

                                assert(k <= z);
                                z -= k;
                                p += k;
                        } while (z > 0);
                }

        next:
                if (max_bytes != (uint64_t) -1) {
                        assert(max_bytes >= (uint64_t) n);
                        max_bytes -= n;
                }
                /* sendfile accepts at most SSIZE_MAX-offset bytes to copy,
                 * so reduce our maximum by the amount we already copied,
                 * but don't go below our copy buffer size, unless we are
                 * close the limit of bytes we are allowed to copy. */
                m = MAX(MIN(COPY_BUFFER_SIZE, max_bytes), m - n);
        }

        return 0; /* return 0 if we hit EOF earlier than the size limit */
}

static int fd_copy_symlink(
                int df,
                const char *from,
                const struct stat *st,
                int dt,
                const char *to,
                uid_t override_uid,
                gid_t override_gid,
                CopyFlags copy_flags) {

        _cleanup_free_ char *target = NULL;
        int r;

        assert(from);
        assert(st);
        assert(to);

        r = readlinkat_malloc(df, from, &target);
        if (r < 0)
                return r;

        if (symlinkat(target, dt, to) < 0)
                return -errno;

        if (fchownat(dt, to,
                     uid_is_valid(override_uid) ? override_uid : st->st_uid,
                     gid_is_valid(override_gid) ? override_gid : st->st_gid,
                     AT_SYMLINK_NOFOLLOW) < 0)
                return -errno;

        return 0;
}

static int fd_copy_regular(
                int df,
                const char *from,
                const struct stat *st,
                int dt,
                const char *to,
                uid_t override_uid,
                gid_t override_gid,
                CopyFlags copy_flags) {

        _cleanup_close_ int fdf = -1, fdt = -1;
        struct timespec ts[2];
        int r, q;

        assert(from);
        assert(st);
        assert(to);

        fdf = openat(df, from, O_RDONLY|O_CLOEXEC|O_NOCTTY|O_NOFOLLOW);
        if (fdf < 0)
                return -errno;

        fdt = openat(dt, to, O_WRONLY|O_CREAT|O_EXCL|O_CLOEXEC|O_NOCTTY|O_NOFOLLOW, st->st_mode & 07777);
        if (fdt < 0)
                return -errno;

        r = copy_bytes(fdf, fdt, (uint64_t) -1, copy_flags);
        if (r < 0) {
                (void) unlinkat(dt, to, 0);
                return r;
        }

        if (fchown(fdt,
                   uid_is_valid(override_uid) ? override_uid : st->st_uid,
                   gid_is_valid(override_gid) ? override_gid : st->st_gid) < 0)
                r = -errno;

        if (fchmod(fdt, st->st_mode & 07777) < 0)
                r = -errno;

        ts[0] = st->st_atim;
        ts[1] = st->st_mtim;
        (void) futimens(fdt, ts);
        (void) copy_xattr(fdf, fdt);

        q = close(fdt);
        fdt = -1;

        if (q < 0) {
                r = -errno;
                (void) unlinkat(dt, to, 0);
        }

        return r;
}

static int fd_copy_fifo(
                int df,
                const char *from,
                const struct stat *st,
                int dt,
                const char *to,
                uid_t override_uid,
                gid_t override_gid,
                CopyFlags copy_flags) {
        int r;

        assert(from);
        assert(st);
        assert(to);

        r = mkfifoat(dt, to, st->st_mode & 07777);
        if (r < 0)
                return -errno;

        if (fchownat(dt, to,
                     uid_is_valid(override_uid) ? override_uid : st->st_uid,
                     gid_is_valid(override_gid) ? override_gid : st->st_gid,
                     AT_SYMLINK_NOFOLLOW) < 0)
                r = -errno;

        if (fchmodat(dt, to, st->st_mode & 07777, 0) < 0)
                r = -errno;

        return r;
}

static int fd_copy_node(
                int df,
                const char *from,
                const struct stat *st,
                int dt,
                const char *to,
                uid_t override_uid,
                gid_t override_gid,
                CopyFlags copy_flags) {
        int r;

        assert(from);
        assert(st);
        assert(to);

        r = mknodat(dt, to, st->st_mode, st->st_rdev);
        if (r < 0)
                return -errno;

        if (fchownat(dt, to,
                     uid_is_valid(override_uid) ? override_uid : st->st_uid,
                     gid_is_valid(override_gid) ? override_gid : st->st_gid,
                     AT_SYMLINK_NOFOLLOW) < 0)
                r = -errno;

        if (fchmodat(dt, to, st->st_mode & 07777, 0) < 0)
                r = -errno;

        return r;
}

static int fd_copy_directory(
                int df,
                const char *from,
                const struct stat *st,
                int dt,
                const char *to,
                dev_t original_device,
                unsigned depth_left,
                uid_t override_uid,
                gid_t override_gid,
                CopyFlags copy_flags) {

        _cleanup_close_ int fdf = -1, fdt = -1;
        _cleanup_closedir_ DIR *d = NULL;
        struct dirent *de;
        bool created;
        int r;

        assert(st);
        assert(to);

        if (depth_left == 0)
                return -ENAMETOOLONG;

        if (from)
                fdf = openat(df, from, O_RDONLY|O_DIRECTORY|O_CLOEXEC|O_NOCTTY|O_NOFOLLOW);
        else
                fdf = fcntl(df, F_DUPFD_CLOEXEC, 3);
        if (fdf < 0)
                return -errno;

        d = fdopendir(fdf);
        if (!d)
                return -errno;
        fdf = -1;

        r = mkdirat(dt, to, st->st_mode & 07777);
        if (r >= 0)
                created = true;
        else if (errno == EEXIST && (copy_flags & COPY_MERGE))
                created = false;
        else
                return -errno;

        fdt = openat(dt, to, O_RDONLY|O_DIRECTORY|O_CLOEXEC|O_NOCTTY|O_NOFOLLOW);
        if (fdt < 0)
                return -errno;

        r = 0;

        FOREACH_DIRENT_ALL(de, d, return -errno) {
                struct stat buf;
                int q;

                if (dot_or_dot_dot(de->d_name))
                        continue;

                if (fstatat(dirfd(d), de->d_name, &buf, AT_SYMLINK_NOFOLLOW) < 0) {
                        r = -errno;
                        continue;
                }

                if (S_ISDIR(buf.st_mode)) {
                        /*
                         * Don't descend into directories on other file systems, if this is requested. We do a simple
                         * .st_dev check here, which basically comes for free. Note that we do this check only on
                         * directories, not other kind of file system objects, for two reason:
                         *
                         * • The kernel's overlayfs pseudo file system that overlays multiple real file systems
                         *   propagates the .st_dev field of the file system a file originates from all the way up
                         *   through the stack to stat(). It doesn't do that for directories however. This means that
                         *   comparing .st_dev on non-directories suggests that they all are mount points. To avoid
                         *   confusion we hence avoid relying on this check for regular files.
                         *
                         * • The main reason we do this check at all is to protect ourselves from bind mount cycles,
                         *   where we really want to avoid descending down in all eternity. However the .st_dev check
                         *   is usually not sufficient for this protection anyway, as bind mount cycles from the same
                         *   file system onto itself can't be detected that way. (Note we also do a recursion depth
                         *   check, which is probably the better protection in this regard, which is why
                         *   COPY_SAME_MOUNT is optional).
                         */

                        if (FLAGS_SET(copy_flags, COPY_SAME_MOUNT)) {
                                if (buf.st_dev != original_device)
                                        continue;

                                r = fd_is_mount_point(dirfd(d), de->d_name, 0);
                                if (r < 0)
                                        return r;
                                if (r > 0)
                                        continue;
                        }

                        q = fd_copy_directory(dirfd(d), de->d_name, &buf, fdt, de->d_name, original_device, depth_left-1, override_uid, override_gid, copy_flags);
                } else if (S_ISREG(buf.st_mode))
                        q = fd_copy_regular(dirfd(d), de->d_name, &buf, fdt, de->d_name, override_uid, override_gid, copy_flags);
                else if (S_ISLNK(buf.st_mode))
                        q = fd_copy_symlink(dirfd(d), de->d_name, &buf, fdt, de->d_name, override_uid, override_gid, copy_flags);
                else if (S_ISFIFO(buf.st_mode))
                        q = fd_copy_fifo(dirfd(d), de->d_name, &buf, fdt, de->d_name, override_uid, override_gid, copy_flags);
                else if (S_ISBLK(buf.st_mode) || S_ISCHR(buf.st_mode) || S_ISSOCK(buf.st_mode))
                        q = fd_copy_node(dirfd(d), de->d_name, &buf, fdt, de->d_name, override_uid, override_gid, copy_flags);
                else
                        q = -EOPNOTSUPP;

                if (q == -EEXIST && (copy_flags & COPY_MERGE))
                        q = 0;

                if (q < 0)
                        r = q;
        }

        if (created) {
                struct timespec ut[2] = {
                        st->st_atim,
                        st->st_mtim
                };

                if (fchown(fdt,
                           uid_is_valid(override_uid) ? override_uid : st->st_uid,
                           gid_is_valid(override_gid) ? override_gid : st->st_gid) < 0)
                        r = -errno;

                if (fchmod(fdt, st->st_mode & 07777) < 0)
                        r = -errno;

                (void) copy_xattr(dirfd(d), fdt);
                (void) futimens(fdt, ut);
        }

        return r;
}

int copy_tree_at(int fdf, const char *from, int fdt, const char *to, uid_t override_uid, gid_t override_gid, CopyFlags copy_flags) {
        struct stat st;

        assert(from);
        assert(to);

        if (fstatat(fdf, from, &st, AT_SYMLINK_NOFOLLOW) < 0)
                return -errno;

        if (S_ISREG(st.st_mode))
                return fd_copy_regular(fdf, from, &st, fdt, to, override_uid, override_gid, copy_flags);
        else if (S_ISDIR(st.st_mode))
                return fd_copy_directory(fdf, from, &st, fdt, to, st.st_dev, COPY_DEPTH_MAX, override_uid, override_gid, copy_flags);
        else if (S_ISLNK(st.st_mode))
                return fd_copy_symlink(fdf, from, &st, fdt, to, override_uid, override_gid, copy_flags);
        else if (S_ISFIFO(st.st_mode))
                return fd_copy_fifo(fdf, from, &st, fdt, to, override_uid, override_gid, copy_flags);
        else if (S_ISBLK(st.st_mode) || S_ISCHR(st.st_mode) || S_ISSOCK(st.st_mode))
                return fd_copy_node(fdf, from, &st, fdt, to, override_uid, override_gid, copy_flags);
        else
                return -EOPNOTSUPP;
}

int copy_tree(const char *from, const char *to, uid_t override_uid, gid_t override_gid, CopyFlags copy_flags) {
        return copy_tree_at(AT_FDCWD, from, AT_FDCWD, to, override_uid, override_gid, copy_flags);
}

int copy_directory_fd(int dirfd, const char *to, CopyFlags copy_flags) {
        struct stat st;

        assert(dirfd >= 0);
        assert(to);

        if (fstat(dirfd, &st) < 0)
                return -errno;

        if (!S_ISDIR(st.st_mode))
                return -ENOTDIR;

        return fd_copy_directory(dirfd, NULL, &st, AT_FDCWD, to, st.st_dev, COPY_DEPTH_MAX, UID_INVALID, GID_INVALID, copy_flags);
}

int copy_directory(const char *from, const char *to, CopyFlags copy_flags) {
        struct stat st;

        assert(from);
        assert(to);

        if (lstat(from, &st) < 0)
                return -errno;

        if (!S_ISDIR(st.st_mode))
                return -ENOTDIR;

        return fd_copy_directory(AT_FDCWD, from, &st, AT_FDCWD, to, st.st_dev, COPY_DEPTH_MAX, UID_INVALID, GID_INVALID, copy_flags);
}

int copy_file_fd(const char *from, int fdt, CopyFlags copy_flags) {
        _cleanup_close_ int fdf = -1;
        int r;

        assert(from);
        assert(fdt >= 0);

        fdf = open(from, O_RDONLY|O_CLOEXEC|O_NOCTTY);
        if (fdf < 0)
                return -errno;

        r = copy_bytes(fdf, fdt, (uint64_t) -1, copy_flags);

        (void) copy_times(fdf, fdt);
        (void) copy_xattr(fdf, fdt);

        return r;
}

int copy_file(const char *from, const char *to, int flags, mode_t mode, unsigned chattr_flags, CopyFlags copy_flags) {
        int fdt = -1, r;

        assert(from);
        assert(to);

        RUN_WITH_UMASK(0000) {
                fdt = open(to, flags|O_WRONLY|O_CREAT|O_CLOEXEC|O_NOCTTY, mode);
                if (fdt < 0)
                        return -errno;
        }

        if (chattr_flags != 0)
                (void) chattr_fd(fdt, chattr_flags, (unsigned) -1);

        r = copy_file_fd(from, fdt, copy_flags);
        if (r < 0) {
                close(fdt);
                (void) unlink(to);
                return r;
        }

        if (close(fdt) < 0) {
                unlink_noerrno(to);
                return -errno;
        }

        return 0;
}

int copy_file_atomic(const char *from, const char *to, mode_t mode, unsigned chattr_flags, CopyFlags copy_flags) {
        _cleanup_free_ char *t = NULL;
        int r;

        assert(from);
        assert(to);

        r = tempfn_random(to, NULL, &t);
        if (r < 0)
                return r;

        r = copy_file(from, t, O_NOFOLLOW|O_EXCL, mode, chattr_flags, copy_flags);
        if (r < 0)
                return r;

        if (copy_flags & COPY_REPLACE) {
                r = renameat(AT_FDCWD, t, AT_FDCWD, to);
                if (r < 0)
                        r = -errno;
        } else
                r = rename_noreplace(AT_FDCWD, t, AT_FDCWD, to);
        if (r < 0) {
                (void) unlink(t);
                return r;
        }

        return 0;
}

int copy_times(int fdf, int fdt) {
        struct timespec ut[2];
        struct stat st;
        usec_t crtime = 0;

        assert(fdf >= 0);
        assert(fdt >= 0);

        if (fstat(fdf, &st) < 0)
                return -errno;

        ut[0] = st.st_atim;
        ut[1] = st.st_mtim;

        if (futimens(fdt, ut) < 0)
                return -errno;

        if (fd_getcrtime(fdf, &crtime) >= 0)
                (void) fd_setcrtime(fdt, crtime);

        return 0;
}

int copy_xattr(int fdf, int fdt) {
        _cleanup_free_ char *bufa = NULL, *bufb = NULL;
        size_t sza = 100, szb = 100;
        ssize_t n;
        int ret = 0;
        const char *p;

        for (;;) {
                bufa = malloc(sza);
                if (!bufa)
                        return -ENOMEM;

                n = flistxattr(fdf, bufa, sza);
                if (n == 0)
                        return 0;
                if (n > 0)
                        break;
                if (errno != ERANGE)
                        return -errno;

                sza *= 2;

                bufa = mfree(bufa);
        }

        p = bufa;
        while (n > 0) {
                size_t l;

                l = strlen(p);
                assert(l < (size_t) n);

                if (startswith(p, "user.")) {
                        ssize_t m;

                        if (!bufb) {
                                bufb = malloc(szb);
                                if (!bufb)
                                        return -ENOMEM;
                        }

                        m = fgetxattr(fdf, p, bufb, szb);
                        if (m < 0) {
                                if (errno == ERANGE) {
                                        szb *= 2;
                                        bufb = mfree(bufb);
                                        continue;
                                }

                                return -errno;
                        }

                        if (fsetxattr(fdt, p, bufb, m, 0) < 0)
                                ret = -errno;
                }

                p += l + 1;
                n -= l + 1;
        }

        return ret;
}