summaryrefslogtreecommitdiffstats
path: root/src/shared/mount-util.c
blob: b19b3849aaa68b6cf378a067d12b202c197eb723 (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
/* SPDX-License-Identifier: LGPL-2.1-or-later */

#include <errno.h>
#include <stdlib.h>
#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/statvfs.h>
#include <unistd.h>

#include "alloc-util.h"
#include "extract-word.h"
#include "fd-util.h"
#include "fileio.h"
#include "fs-util.h"
#include "hashmap.h"
#include "libmount-util.h"
#include "mount-util.h"
#include "mountpoint-util.h"
#include "parse-util.h"
#include "path-util.h"
#include "set.h"
#include "stat-util.h"
#include "stdio-util.h"
#include "string-util.h"
#include "strv.h"

int mount_fd(const char *source,
             int target_fd,
             const char *filesystemtype,
             unsigned long mountflags,
             const void *data) {

        char path[STRLEN("/proc/self/fd/") + DECIMAL_STR_MAX(int)];

        xsprintf(path, "/proc/self/fd/%i", target_fd);
        if (mount(source, path, filesystemtype, mountflags, data) < 0) {
                if (errno != ENOENT)
                        return -errno;

                /* ENOENT can mean two things: either that the source is missing, or that /proc/ isn't
                 * mounted. Check for the latter to generate better error messages. */
                if (proc_mounted() == 0)
                        return -ENOSYS;

                return -ENOENT;
        }

        return 0;
}

int mount_nofollow(
                const char *source,
                const char *target,
                const char *filesystemtype,
                unsigned long mountflags,
                const void *data) {

        _cleanup_close_ int fd = -1;

        /* In almost all cases we want to manipulate the mount table without following symlinks, hence
         * mount_nofollow() is usually the way to go. The only exceptions are environments where /proc/ is
         * not available yet, since we need /proc/self/fd/ for this logic to work. i.e. during the early
         * initialization of namespacing/container stuff where /proc is not yet mounted (and maybe even the
         * fs to mount) we can only use traditional mount() directly.
         *
         * Note that this disables following only for the final component of the target, i.e symlinks within
         * the path of the target are honoured, as are symlinks in the source path everywhere. */

        fd = open(target, O_PATH|O_CLOEXEC|O_NOFOLLOW);
        if (fd < 0)
                return -errno;

        return mount_fd(source, fd, filesystemtype, mountflags, data);
}

int umount_recursive(const char *prefix, int flags) {
        int n = 0, r;
        bool again;

        /* Try to umount everything recursively below a
         * directory. Also, take care of stacked mounts, and keep
         * unmounting them until they are gone. */

        do {
                _cleanup_(mnt_free_tablep) struct libmnt_table *table = NULL;
                _cleanup_(mnt_free_iterp) struct libmnt_iter *iter = NULL;

                again = false;

                r = libmount_parse("/proc/self/mountinfo", NULL, &table, &iter);
                if (r < 0)
                        return log_debug_errno(r, "Failed to parse /proc/self/mountinfo: %m");

                for (;;) {
                        struct libmnt_fs *fs;
                        const char *path;

                        r = mnt_table_next_fs(table, iter, &fs);
                        if (r == 1)
                                break;
                        if (r < 0)
                                return log_debug_errno(r, "Failed to get next entry from /proc/self/mountinfo: %m");

                        path = mnt_fs_get_target(fs);
                        if (!path)
                                continue;

                        if (!path_startswith(path, prefix))
                                continue;

                        if (umount2(path, flags | UMOUNT_NOFOLLOW) < 0) {
                                log_debug_errno(errno, "Failed to umount %s, ignoring: %m", path);
                                continue;
                        }

                        log_debug("Successfully unmounted %s", path);

                        again = true;
                        n++;

                        break;
                }
        } while (again);

        return n;
}

static int get_mount_flags(
                struct libmnt_table *table,
                const char *path,
                unsigned long *ret) {

        _cleanup_close_ int fd = -1;
        struct libmnt_fs *fs;
        struct statvfs buf;
        const char *opts;
        int r;

        /* Get the mount flags for the mountpoint at "path" from "table". We have a fallback using statvfs()
         * in place (which provides us with mostly the same info), but it's just a fallback, since using it
         * means triggering autofs or NFS mounts, which we'd rather avoid needlessly.
         *
         * This generally doesn't follow symlinks. */

        fs = mnt_table_find_target(table, path, MNT_ITER_FORWARD);
        if (!fs) {
                log_debug("Could not find '%s' in mount table, ignoring.", path);
                goto fallback;
        }

        opts = mnt_fs_get_vfs_options(fs);
        if (!opts) {
                *ret = 0;
                return 0;
        }

        r = mnt_optstr_get_flags(opts, ret, mnt_get_builtin_optmap(MNT_LINUX_MAP));
        if (r != 0) {
                log_debug_errno(r, "Could not get flags for '%s', ignoring: %m", path);
                goto fallback;
        }

        /* MS_RELATIME is default and trying to set it in an unprivileged container causes EPERM */
        *ret &= ~MS_RELATIME;
        return 0;

fallback:
        fd = open(path, O_PATH|O_CLOEXEC|O_NOFOLLOW);
        if (fd < 0)
                return -errno;

        if (fstatvfs(fd, &buf) < 0)
                return -errno;

        /* The statvfs() flags and the mount flags mostly have the same values, but for some cases do
         * not. Hence map the flags manually. (Strictly speaking, ST_RELATIME/MS_RELATIME is the most
         * prominent one that doesn't match, but that's the one we mask away anyway, see above.) */

        *ret =
                FLAGS_SET(buf.f_flag, ST_RDONLY) * MS_RDONLY |
                FLAGS_SET(buf.f_flag, ST_NODEV) * MS_NODEV |
                FLAGS_SET(buf.f_flag, ST_NOEXEC) * MS_NOEXEC |
                FLAGS_SET(buf.f_flag, ST_NOSUID) * MS_NOSUID |
                FLAGS_SET(buf.f_flag, ST_NOATIME) * MS_NOATIME |
                FLAGS_SET(buf.f_flag, ST_NODIRATIME) * MS_NODIRATIME;

        return 0;
}

/* Use this function only if you do not have direct access to /proc/self/mountinfo but the caller can open it
 * for you. This is the case when /proc is masked or not mounted. Otherwise, use bind_remount_recursive. */
int bind_remount_recursive_with_mountinfo(
                const char *prefix,
                unsigned long new_flags,
                unsigned long flags_mask,
                char **deny_list,
                FILE *proc_self_mountinfo) {

        _cleanup_set_free_free_ Set *done = NULL;
        _cleanup_free_ char *simplified = NULL;
        int r;

        assert(prefix);
        assert(proc_self_mountinfo);

        /* Recursively remount a directory (and all its submounts) read-only or read-write. If the directory is already
         * mounted, we reuse the mount and simply mark it MS_BIND|MS_RDONLY (or remove the MS_RDONLY for read-write
         * operation). If it isn't we first make it one. Afterwards we apply MS_BIND|MS_RDONLY (or remove MS_RDONLY) to
         * all submounts we can access, too. When mounts are stacked on the same mount point we only care for each
         * individual "top-level" mount on each point, as we cannot influence/access the underlying mounts anyway. We
         * do not have any effect on future submounts that might get propagated, they might be writable. This includes
         * future submounts that have been triggered via autofs.
         *
         * If the "deny_list" parameter is specified it may contain a list of subtrees to exclude from the
         * remount operation. Note that we'll ignore the deny list for the top-level path. */

        simplified = strdup(prefix);
        if (!simplified)
                return -ENOMEM;

        path_simplify(simplified, false);

        done = set_new(&path_hash_ops);
        if (!done)
                return -ENOMEM;

        for (;;) {
                _cleanup_set_free_free_ Set *todo = NULL;
                _cleanup_(mnt_free_tablep) struct libmnt_table *table = NULL;
                _cleanup_(mnt_free_iterp) struct libmnt_iter *iter = NULL;
                bool top_autofs = false;
                char *x;
                unsigned long orig_flags;

                todo = set_new(&path_hash_ops);
                if (!todo)
                        return -ENOMEM;

                rewind(proc_self_mountinfo);

                r = libmount_parse("/proc/self/mountinfo", proc_self_mountinfo, &table, &iter);
                if (r < 0)
                        return log_debug_errno(r, "Failed to parse /proc/self/mountinfo: %m");

                for (;;) {
                        struct libmnt_fs *fs;
                        const char *path, *type;

                        r = mnt_table_next_fs(table, iter, &fs);
                        if (r == 1)
                                break;
                        if (r < 0)
                                return log_debug_errno(r, "Failed to get next entry from /proc/self/mountinfo: %m");

                        path = mnt_fs_get_target(fs);
                        type = mnt_fs_get_fstype(fs);
                        if (!path || !type)
                                continue;

                        if (!path_startswith(path, simplified))
                                continue;

                        /* Ignore this mount if it is deny-listed, but only if it isn't the top-level mount
                         * we shall operate on. */
                        if (!path_equal(path, simplified)) {
                                bool deny_listed = false;
                                char **i;

                                STRV_FOREACH(i, deny_list) {
                                        if (path_equal(*i, simplified))
                                                continue;

                                        if (!path_startswith(*i, simplified))
                                                continue;

                                        if (path_startswith(path, *i)) {
                                                deny_listed = true;
                                                log_debug("Not remounting %s deny-listed by %s, called for %s",
                                                          path, *i, simplified);
                                                break;
                                        }
                                }
                                if (deny_listed)
                                        continue;
                        }

                        /* Let's ignore autofs mounts.  If they aren't
                         * triggered yet, we want to avoid triggering
                         * them, as we don't make any guarantees for
                         * future submounts anyway.  If they are
                         * already triggered, then we will find
                         * another entry for this. */
                        if (streq(type, "autofs")) {
                                top_autofs = top_autofs || path_equal(path, simplified);
                                continue;
                        }

                        if (!set_contains(done, path)) {
                                r = set_put_strdup(&todo, path);
                                if (r < 0)
                                        return r;
                        }
                }

                /* If we have no submounts to process anymore and if
                 * the root is either already done, or an autofs, we
                 * are done */
                if (set_isempty(todo) &&
                    (top_autofs || set_contains(done, simplified)))
                        return 0;

                if (!set_contains(done, simplified) &&
                    !set_contains(todo, simplified)) {
                        /* The prefix directory itself is not yet a mount, make it one. */
                        r = mount_nofollow(simplified, simplified, NULL, MS_BIND|MS_REC, NULL);
                        if (r < 0)
                                return r;

                        orig_flags = 0;
                        (void) get_mount_flags(table, simplified, &orig_flags);

                        r = mount_nofollow(NULL, simplified, NULL, (orig_flags & ~flags_mask)|MS_BIND|MS_REMOUNT|new_flags, NULL);
                        if (r < 0)
                                return r;

                        log_debug("Made top-level directory %s a mount point.", prefix);

                        r = set_put_strdup(&done, simplified);
                        if (r < 0)
                                return r;
                }

                while ((x = set_steal_first(todo))) {

                        r = set_consume(done, x);
                        if (IN_SET(r, 0, -EEXIST))
                                continue;
                        if (r < 0)
                                return r;

                        /* Deal with mount points that are obstructed by a later mount */
                        r = path_is_mount_point(x, NULL, 0);
                        if (IN_SET(r, 0, -ENOENT))
                                continue;
                        if (r < 0) {
                                if (!ERRNO_IS_PRIVILEGE(r))
                                        return r;

                                /* Even if root user invoke this, submounts under private FUSE or NFS mount points
                                 * may not be acceessed. E.g.,
                                 *
                                 * $ bindfs --no-allow-other ~/mnt/mnt ~/mnt/mnt
                                 * $ bindfs --no-allow-other ~/mnt ~/mnt
                                 *
                                 * Then, root user cannot access the mount point ~/mnt/mnt.
                                 * In such cases, the submounts are ignored, as we have no way to manage them. */
                                log_debug_errno(r, "Failed to determine '%s' is mount point or not, ignoring: %m", x);
                                continue;
                        }

                        /* Try to reuse the original flag set */
                        orig_flags = 0;
                        (void) get_mount_flags(table, x, &orig_flags);

                        r = mount_nofollow(NULL, x, NULL, (orig_flags & ~flags_mask)|MS_BIND|MS_REMOUNT|new_flags, NULL);
                        if (r < 0)
                                return r;

                        log_debug("Remounted %s read-only.", x);
                }
        }
}

int bind_remount_recursive(
                const char *prefix,
                unsigned long new_flags,
                unsigned long flags_mask,
                char **deny_list) {

        _cleanup_fclose_ FILE *proc_self_mountinfo = NULL;
        int r;

        r = fopen_unlocked("/proc/self/mountinfo", "re", &proc_self_mountinfo);
        if (r < 0)
                return r;

        return bind_remount_recursive_with_mountinfo(prefix, new_flags, flags_mask, deny_list, proc_self_mountinfo);
}

int bind_remount_one_with_mountinfo(
                const char *path,
                unsigned long new_flags,
                unsigned long flags_mask,
                FILE *proc_self_mountinfo) {

        _cleanup_(mnt_free_tablep) struct libmnt_table *table = NULL;
        unsigned long orig_flags = 0;
        int r;

        assert(path);
        assert(proc_self_mountinfo);

        rewind(proc_self_mountinfo);

        table = mnt_new_table();
        if (!table)
                return -ENOMEM;

        r = mnt_table_parse_stream(table, proc_self_mountinfo, "/proc/self/mountinfo");
        if (r < 0)
                return r;

        /* Try to reuse the original flag set */
        (void) get_mount_flags(table, path, &orig_flags);

        r = mount_nofollow(NULL, path, NULL, (orig_flags & ~flags_mask)|MS_BIND|MS_REMOUNT|new_flags, NULL);
        if (r < 0)
                return r;

        return 0;
}

int mount_move_root(const char *path) {
        assert(path);

        if (chdir(path) < 0)
                return -errno;

        if (mount(path, "/", NULL, MS_MOVE, NULL) < 0)
                return -errno;

        if (chroot(".") < 0)
                return -errno;

        if (chdir("/") < 0)
                return -errno;

        return 0;
}

int repeat_unmount(const char *path, int flags) {
        bool done = false;

        assert(path);

        /* If there are multiple mounts on a mount point, this
         * removes them all */

        for (;;) {
                if (umount2(path, flags) < 0) {

                        if (errno == EINVAL)
                                return done;

                        return -errno;
                }

                done = true;
        }
}

int mode_to_inaccessible_node(
                const char *runtime_dir,
                mode_t mode,
                char **ret) {

        /* This function maps a node type to a corresponding inaccessible file node. These nodes are created
         * during early boot by PID 1. In some cases we lacked the privs to create the character and block
         * devices (maybe because we run in an userns environment, or miss CAP_SYS_MKNOD, or run with a
         * devices policy that excludes device nodes with major and minor of 0), but that's fine, in that
         * case we use an AF_UNIX file node instead, which is not the same, but close enough for most
         * uses. And most importantly, the kernel allows bind mounts from socket nodes to any non-directory
         * file nodes, and that's the most important thing that matters.
         *
         * Note that the runtime directory argument shall be the top-level runtime directory, i.e. /run/ if
         * we operate in system context and $XDG_RUNTIME_DIR if we operate in user context. */

        _cleanup_free_ char *d = NULL;
        const char *node = NULL;

        assert(ret);

        if (!runtime_dir)
                runtime_dir = "/run";

        switch(mode & S_IFMT) {
                case S_IFREG:
                        node = "/systemd/inaccessible/reg";
                        break;

                case S_IFDIR:
                        node = "/systemd/inaccessible/dir";
                        break;

                case S_IFCHR:
                        node = "/systemd/inaccessible/chr";
                        break;

                case S_IFBLK:
                        node = "/systemd/inaccessible/blk";
                        break;

                case S_IFIFO:
                        node = "/systemd/inaccessible/fifo";
                        break;

                case S_IFSOCK:
                        node = "/systemd/inaccessible/sock";
                        break;
        }
        if (!node)
                return -EINVAL;

        d = path_join(runtime_dir, node);
        if (!d)
                return -ENOMEM;

        /* On new kernels unprivileged users are permitted to create 0:0 char device nodes (because they also
         * act as whiteout inode for overlayfs), but no other char or block device nodes. On old kernels no
         * device node whatsoever may be created by unprivileged processes. Hence, if the caller asks for the
         * inaccessible block device node let's see if the block device node actually exists, and if not,
         * fall back to the character device node. From there fall back to the socket device node. This means
         * in the best case we'll get the right device node type — but if not we'll hopefully at least get a
         * device node at all. */

        if (S_ISBLK(mode) &&
            access(d, F_OK) < 0 && errno == ENOENT) {
                free(d);
                d = path_join(runtime_dir, "/systemd/inaccessible/chr");
                if (!d)
                        return -ENOMEM;
        }

        if (IN_SET(mode & S_IFMT, S_IFBLK, S_IFCHR) &&
            access(d, F_OK) < 0 && errno == ENOENT) {
                free(d);
                d = path_join(runtime_dir, "/systemd/inaccessible/sock");
                if (!d)
                        return -ENOMEM;
        }

        *ret = TAKE_PTR(d);
        return 0;
}

#define FLAG(name) (flags & name ? STRINGIFY(name) "|" : "")
static char* mount_flags_to_string(long unsigned flags) {
        char *x;
        _cleanup_free_ char *y = NULL;
        long unsigned overflow;

        overflow = flags & ~(MS_RDONLY |
                             MS_NOSUID |
                             MS_NODEV |
                             MS_NOEXEC |
                             MS_SYNCHRONOUS |
                             MS_REMOUNT |
                             MS_MANDLOCK |
                             MS_DIRSYNC |
                             MS_NOATIME |
                             MS_NODIRATIME |
                             MS_BIND |
                             MS_MOVE |
                             MS_REC |
                             MS_SILENT |
                             MS_POSIXACL |
                             MS_UNBINDABLE |
                             MS_PRIVATE |
                             MS_SLAVE |
                             MS_SHARED |
                             MS_RELATIME |
                             MS_KERNMOUNT |
                             MS_I_VERSION |
                             MS_STRICTATIME |
                             MS_LAZYTIME);

        if (flags == 0 || overflow != 0)
                if (asprintf(&y, "%lx", overflow) < 0)
                        return NULL;

        x = strjoin(FLAG(MS_RDONLY),
                    FLAG(MS_NOSUID),
                    FLAG(MS_NODEV),
                    FLAG(MS_NOEXEC),
                    FLAG(MS_SYNCHRONOUS),
                    FLAG(MS_REMOUNT),
                    FLAG(MS_MANDLOCK),
                    FLAG(MS_DIRSYNC),
                    FLAG(MS_NOATIME),
                    FLAG(MS_NODIRATIME),
                    FLAG(MS_BIND),
                    FLAG(MS_MOVE),
                    FLAG(MS_REC),
                    FLAG(MS_SILENT),
                    FLAG(MS_POSIXACL),
                    FLAG(MS_UNBINDABLE),
                    FLAG(MS_PRIVATE),
                    FLAG(MS_SLAVE),
                    FLAG(MS_SHARED),
                    FLAG(MS_RELATIME),
                    FLAG(MS_KERNMOUNT),
                    FLAG(MS_I_VERSION),
                    FLAG(MS_STRICTATIME),
                    FLAG(MS_LAZYTIME),
                    y);
        if (!x)
                return NULL;
        if (!y)
                x[strlen(x) - 1] = '\0'; /* truncate the last | */
        return x;
}

int mount_verbose_full(
                int error_log_level,
                const char *what,
                const char *where,
                const char *type,
                unsigned long flags,
                const char *options,
                bool follow_symlink) {

        _cleanup_free_ char *fl = NULL, *o = NULL;
        unsigned long f;
        int r;

        r = mount_option_mangle(options, flags, &f, &o);
        if (r < 0)
                return log_full_errno(error_log_level, r,
                                      "Failed to mangle mount options %s: %m",
                                      strempty(options));

        fl = mount_flags_to_string(f);

        if ((f & MS_REMOUNT) && !what && !type)
                log_debug("Remounting %s (%s \"%s\")...",
                          where, strnull(fl), strempty(o));
        else if (!what && !type)
                log_debug("Mounting %s (%s \"%s\")...",
                          where, strnull(fl), strempty(o));
        else if ((f & MS_BIND) && !type)
                log_debug("Bind-mounting %s on %s (%s \"%s\")...",
                          what, where, strnull(fl), strempty(o));
        else if (f & MS_MOVE)
                log_debug("Moving mount %s → %s (%s \"%s\")...",
                          what, where, strnull(fl), strempty(o));
        else
                log_debug("Mounting %s (%s) on %s (%s \"%s\")...",
                          strna(what), strna(type), where, strnull(fl), strempty(o));

        if (follow_symlink)
                r = mount(what, where, type, f, o) < 0 ? -errno : 0;
        else
                r = mount_nofollow(what, where, type, f, o);
        if (r < 0)
                return log_full_errno(error_log_level, r,
                                      "Failed to mount %s (type %s) on %s (%s \"%s\"): %m",
                                      strna(what), strna(type), where, strnull(fl), strempty(o));
        return 0;
}

int umount_verbose(
                int error_log_level,
                const char *what,
                int flags) {

        assert(what);

        log_debug("Umounting %s...", what);

        if (umount2(what, flags) < 0)
                return log_full_errno(error_log_level, errno,
                                      "Failed to unmount %s: %m", what);

        return 0;
}

int mount_option_mangle(
                const char *options,
                unsigned long mount_flags,
                unsigned long *ret_mount_flags,
                char **ret_remaining_options) {

        const struct libmnt_optmap *map;
        _cleanup_free_ char *ret = NULL;
        const char *p;
        int r;

        /* This extracts mount flags from the mount options, and store
         * non-mount-flag options to '*ret_remaining_options'.
         * E.g.,
         * "rw,nosuid,nodev,relatime,size=1630748k,mode=700,uid=1000,gid=1000"
         * is split to MS_NOSUID|MS_NODEV|MS_RELATIME and
         * "size=1630748k,mode=700,uid=1000,gid=1000".
         * See more examples in test-mount-utils.c.
         *
         * Note that if 'options' does not contain any non-mount-flag options,
         * then '*ret_remaining_options' is set to NULL instead of empty string.
         * Note that this does not check validity of options stored in
         * '*ret_remaining_options'.
         * Note that if 'options' is NULL, then this just copies 'mount_flags'
         * to '*ret_mount_flags'. */

        assert(ret_mount_flags);
        assert(ret_remaining_options);

        map = mnt_get_builtin_optmap(MNT_LINUX_MAP);
        if (!map)
                return -EINVAL;

        p = options;
        for (;;) {
                _cleanup_free_ char *word = NULL;
                const struct libmnt_optmap *ent;

                r = extract_first_word(&p, &word, ",", EXTRACT_UNQUOTE);
                if (r < 0)
                        return r;
                if (r == 0)
                        break;

                for (ent = map; ent->name; ent++) {
                        /* All entries in MNT_LINUX_MAP do not take any argument.
                         * Thus, ent->name does not contain "=" or "[=]". */
                        if (!streq(word, ent->name))
                                continue;

                        if (!(ent->mask & MNT_INVERT))
                                mount_flags |= ent->id;
                        else if (mount_flags & ent->id)
                                mount_flags ^= ent->id;

                        break;
                }

                /* If 'word' is not a mount flag, then store it in '*ret_remaining_options'. */
                if (!ent->name && !strextend_with_separator(&ret, ",", word, NULL))
                        return -ENOMEM;
        }

        *ret_mount_flags = mount_flags;
        *ret_remaining_options = TAKE_PTR(ret);

        return 0;
}