summaryrefslogtreecommitdiff
path: root/src/nspawn/nspawn-mount.c
blob: 85e81b43fe27e147668517502f502f5920a6c5d9 (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
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
/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/

/***
  This file is part of systemd.

  Copyright 2015 Lennart Poettering

  systemd is free software; you can redistribute it and/or modify it
  under the terms of the GNU Lesser General Public License as published by
  the Free Software Foundation; either version 2.1 of the License, or
  (at your option) any later version.

  systemd is distributed in the hope that it will be useful, but
  WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  Lesser General Public License for more details.

  You should have received a copy of the GNU Lesser General Public License
  along with systemd; If not, see <http://www.gnu.org/licenses/>.
***/

#include <sys/mount.h>

#include "util.h"
#include "rm-rf.h"
#include "strv.h"
#include "path-util.h"
#include "mkdir.h"
#include "label.h"
#include "set.h"
#include "cgroup-util.h"

#include "nspawn-mount.h"

CustomMount* custom_mount_add(CustomMount **l, unsigned *n, CustomMountType t) {
        CustomMount *c, *ret;

        assert(l);
        assert(n);
        assert(t >= 0);
        assert(t < _CUSTOM_MOUNT_TYPE_MAX);

        c = realloc(*l, (*n + 1) * sizeof(CustomMount));
        if (!c)
                return NULL;

        *l = c;
        ret = *l + *n;
        (*n)++;

        *ret = (CustomMount) { .type = t };

        return ret;
}

void custom_mount_free_all(CustomMount *l, unsigned n) {
        unsigned i;

        for (i = 0; i < n; i++) {
                CustomMount *m = l + i;

                free(m->source);
                free(m->destination);
                free(m->options);

                if (m->work_dir) {
                        (void) rm_rf(m->work_dir, REMOVE_ROOT|REMOVE_PHYSICAL);
                        free(m->work_dir);
                }

                strv_free(m->lower);
        }

        free(l);
}

int custom_mount_compare(const void *a, const void *b) {
        const CustomMount *x = a, *y = b;
        int r;

        r = path_compare(x->destination, y->destination);
        if (r != 0)
                return r;

        if (x->type < y->type)
                return -1;
        if (x->type > y->type)
                return 1;

        return 0;
}

int bind_mount_parse(CustomMount **l, unsigned *n, const char *s, bool read_only) {
        _cleanup_free_ char *source = NULL, *destination = NULL, *opts = NULL;
        const char *p = s;
        CustomMount *m;
        int r;

        assert(l);
        assert(n);

        r = extract_many_words(&p, ":", EXTRACT_DONT_COALESCE_SEPARATORS, &source, &destination, NULL);
        if (r < 0)
                return r;
        if (r == 0)
                return -EINVAL;

        if (r == 1) {
                destination = strdup(source);
                if (!destination)
                        return -ENOMEM;
        }

        if (r == 2 && !isempty(p)) {
                opts = strdup(p);
                if (!opts)
                        return -ENOMEM;
        }

        if (!path_is_absolute(source))
                return -EINVAL;

        if (!path_is_absolute(destination))
                return -EINVAL;

        m = custom_mount_add(l, n, CUSTOM_MOUNT_BIND);
        if (!m)
                return log_oom();

        m->source = source;
        m->destination = destination;
        m->read_only = read_only;
        m->options = opts;

        source = destination = opts = NULL;
        return 0;
}

int tmpfs_mount_parse(CustomMount **l, unsigned *n, const char *s) {
        _cleanup_free_ char *path = NULL, *opts = NULL;
        const char *p = s;
        CustomMount *m;
        int r;

        assert(l);
        assert(n);
        assert(s);

        r = extract_first_word(&p, &path, ":", EXTRACT_DONT_COALESCE_SEPARATORS);
        if (r < 0)
                return r;
        if (r == 0)
                return -EINVAL;

        if (isempty(p))
                opts = strdup("mode=0755");
        else
                opts = strdup(p);
        if (!opts)
                return -ENOMEM;

        if (!path_is_absolute(path))
                return -EINVAL;

        m = custom_mount_add(l, n, CUSTOM_MOUNT_TMPFS);
        if (!m)
                return -ENOMEM;

        m->destination = path;
        m->options = opts;

        path = opts = NULL;
        return 0;
}

static int tmpfs_patch_options(
                const char *options,
                bool userns, uid_t uid_shift, uid_t uid_range,
                const char *selinux_apifs_context,
                char **ret) {

        char *buf = NULL;

        if (userns && uid_shift != 0) {
                assert(uid_shift != UID_INVALID);

                if (options)
                        (void) asprintf(&buf, "%s,uid=" UID_FMT ",gid=" UID_FMT, options, uid_shift, uid_shift);
                else
                        (void) asprintf(&buf, "uid=" UID_FMT ",gid=" UID_FMT, uid_shift, uid_shift);
                if (!buf)
                        return -ENOMEM;

                options = buf;
        }

#ifdef HAVE_SELINUX
        if (selinux_apifs_context) {
                char *t;

                if (options)
                        t = strjoin(options, ",context=\"", selinux_apifs_context, "\"", NULL);
                else
                        t = strjoin("context=\"", selinux_apifs_context, "\"", NULL);
                if (!t) {
                        free(buf);
                        return -ENOMEM;
                }

                free(buf);
                buf = t;
        }
#endif

        *ret = buf;
        return !!buf;
}

int mount_all(const char *dest,
              bool use_userns, bool in_userns,
              uid_t uid_shift, uid_t uid_range,
              const char *selinux_apifs_context) {

        typedef struct MountPoint {
                const char *what;
                const char *where;
                const char *type;
                const char *options;
                unsigned long flags;
                bool fatal;
                bool userns;
        } MountPoint;

        static const MountPoint mount_table[] = {
                { "proc",      "/proc",          "proc",   NULL,        MS_NOSUID|MS_NOEXEC|MS_NODEV,                              true,  true  },
                { "/proc/sys", "/proc/sys",      NULL,     NULL,        MS_BIND,                                                   true,  true  },   /* Bind mount first */
                { NULL,        "/proc/sys",      NULL,     NULL,        MS_BIND|MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_REMOUNT, true,  true  },   /* Then, make it r/o */
                { "sysfs",     "/sys",           "sysfs",  NULL,        MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV,                    true,  false },
                { "tmpfs",     "/dev",           "tmpfs",  "mode=755",  MS_NOSUID|MS_STRICTATIME,                                  true,  false },
                { "tmpfs",     "/dev/shm",       "tmpfs",  "mode=1777", MS_NOSUID|MS_NODEV|MS_STRICTATIME,                         true,  false },
                { "tmpfs",     "/run",           "tmpfs",  "mode=755",  MS_NOSUID|MS_NODEV|MS_STRICTATIME,                         true,  false },
                { "tmpfs",     "/tmp",           "tmpfs",  "mode=1777", MS_STRICTATIME,                                            true,  false },
#ifdef HAVE_SELINUX
                { "/sys/fs/selinux", "/sys/fs/selinux", NULL, NULL,     MS_BIND,                                                   false, false },  /* Bind mount first */
                { NULL,              "/sys/fs/selinux", NULL, NULL,     MS_BIND|MS_RDONLY|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_REMOUNT, false, false },  /* Then, make it r/o */
#endif
        };

        unsigned k;
        int r;

        for (k = 0; k < ELEMENTSOF(mount_table); k++) {
                _cleanup_free_ char *where = NULL, *options = NULL;
                const char *o;

                if (in_userns != mount_table[k].userns)
                        continue;

                where = prefix_root(dest, mount_table[k].where);
                if (!where)
                        return log_oom();

                r = path_is_mount_point(where, AT_SYMLINK_FOLLOW);
                if (r < 0 && r != -ENOENT)
                        return log_error_errno(r, "Failed to detect whether %s is a mount point: %m", where);

                /* Skip this entry if it is not a remount. */
                if (mount_table[k].what && r > 0)
                        continue;

                r = mkdir_p(where, 0755);
                if (r < 0) {
                        if (mount_table[k].fatal)
                                return log_error_errno(r, "Failed to create directory %s: %m", where);

                        log_warning_errno(r, "Failed to create directory %s: %m", where);
                        continue;
                }

                o = mount_table[k].options;
                if (streq_ptr(mount_table[k].type, "tmpfs")) {
                        r = tmpfs_patch_options(o, use_userns, uid_shift, uid_range, selinux_apifs_context, &options);
                        if (r < 0)
                                return log_oom();
                        if (r > 0)
                                o = options;
                }

                if (mount(mount_table[k].what,
                          where,
                          mount_table[k].type,
                          mount_table[k].flags,
                          o) < 0) {

                        if (mount_table[k].fatal)
                                return log_error_errno(errno, "mount(%s) failed: %m", where);

                        log_warning_errno(errno, "mount(%s) failed, ignoring: %m", where);
                }
        }

        return 0;
}

static int parse_mount_bind_options(const char *options, unsigned long *mount_flags, char **mount_opts) {
        const char *p = options;
        unsigned long flags = *mount_flags;
        char *opts = NULL;

        assert(options);

        for (;;) {
                _cleanup_free_ char *word = NULL;
                int r = extract_first_word(&p, &word, ",", 0);
                if (r < 0)
                        return log_error_errno(r, "Failed to extract mount option: %m");
                if (r == 0)
                        break;

                if (streq(word, "rbind"))
                        flags |= MS_REC;
                else if (streq(word, "norbind"))
                        flags &= ~MS_REC;
                else {
                        log_error("Invalid bind mount option: %s", word);
                        return -EINVAL;
                }
        }

        *mount_flags = flags;
        /* in the future mount_opts will hold string options for mount(2) */
        *mount_opts = opts;

        return 0;
}

static int mount_bind(const char *dest, CustomMount *m) {
        struct stat source_st, dest_st;
        const char *where;
        unsigned long mount_flags = MS_BIND | MS_REC;
        _cleanup_free_ char *mount_opts = NULL;
        int r;

        assert(m);

        if (m->options) {
                r = parse_mount_bind_options(m->options, &mount_flags, &mount_opts);
                if (r < 0)
                        return r;
        }

        if (stat(m->source, &source_st) < 0)
                return log_error_errno(errno, "Failed to stat %s: %m", m->source);

        where = prefix_roota(dest, m->destination);

        if (stat(where, &dest_st) >= 0) {
                if (S_ISDIR(source_st.st_mode) && !S_ISDIR(dest_st.st_mode)) {
                        log_error("Cannot bind mount directory %s on file %s.", m->source, where);
                        return -EINVAL;
                }

                if (!S_ISDIR(source_st.st_mode) && S_ISDIR(dest_st.st_mode)) {
                        log_error("Cannot bind mount file %s on directory %s.", m->source, where);
                        return -EINVAL;
                }

        } else if (errno == ENOENT) {
                r = mkdir_parents_label(where, 0755);
                if (r < 0)
                        return log_error_errno(r, "Failed to make parents of %s: %m", where);
        } else {
                log_error_errno(errno, "Failed to stat %s: %m", where);
                return -errno;
        }

        /* Create the mount point. Any non-directory file can be
         * mounted on any non-directory file (regular, fifo, socket,
         * char, block).
         */
        if (S_ISDIR(source_st.st_mode))
                r = mkdir_label(where, 0755);
        else
                r = touch(where);
        if (r < 0 && r != -EEXIST)
                return log_error_errno(r, "Failed to create mount point %s: %m", where);

        if (mount(m->source, where, NULL, mount_flags, mount_opts) < 0)
                return log_error_errno(errno, "mount(%s) failed: %m", where);

        if (m->read_only) {
                r = bind_remount_recursive(where, true);
                if (r < 0)
                        return log_error_errno(r, "Read-only bind mount failed: %m");
        }

        return 0;
}

static int mount_tmpfs(
                const char *dest,
                CustomMount *m,
                bool userns, uid_t uid_shift, uid_t uid_range,
                const char *selinux_apifs_context) {

        const char *where, *options;
        _cleanup_free_ char *buf = NULL;
        int r;

        assert(dest);
        assert(m);

        where = prefix_roota(dest, m->destination);

        r = mkdir_p_label(where, 0755);
        if (r < 0 && r != -EEXIST)
                return log_error_errno(r, "Creating mount point for tmpfs %s failed: %m", where);

        r = tmpfs_patch_options(m->options, userns, uid_shift, uid_range, selinux_apifs_context, &buf);
        if (r < 0)
                return log_oom();
        options = r > 0 ? buf : m->options;

        if (mount("tmpfs", where, "tmpfs", MS_NODEV|MS_STRICTATIME, options) < 0)
                return log_error_errno(errno, "tmpfs mount to %s failed: %m", where);

        return 0;
}

static char *joined_and_escaped_lower_dirs(char * const *lower) {
        _cleanup_strv_free_ char **sv = NULL;

        sv = strv_copy(lower);
        if (!sv)
                return NULL;

        strv_reverse(sv);

        if (!strv_shell_escape(sv, ",:"))
                return NULL;

        return strv_join(sv, ":");
}

static int mount_overlay(const char *dest, CustomMount *m) {
        _cleanup_free_ char *lower = NULL;
        const char *where, *options;
        int r;

        assert(dest);
        assert(m);

        where = prefix_roota(dest, m->destination);

        r = mkdir_label(where, 0755);
        if (r < 0 && r != -EEXIST)
                return log_error_errno(r, "Creating mount point for overlay %s failed: %m", where);

        (void) mkdir_p_label(m->source, 0755);

        lower = joined_and_escaped_lower_dirs(m->lower);
        if (!lower)
                return log_oom();

        if (m->read_only) {
                _cleanup_free_ char *escaped_source = NULL;

                escaped_source = shell_escape(m->source, ",:");
                if (!escaped_source)
                        return log_oom();

                options = strjoina("lowerdir=", escaped_source, ":", lower);
        } else {
                _cleanup_free_ char *escaped_source = NULL, *escaped_work_dir = NULL;

                assert(m->work_dir);
                (void) mkdir_label(m->work_dir, 0700);

                escaped_source = shell_escape(m->source, ",:");
                if (!escaped_source)
                        return log_oom();
                escaped_work_dir = shell_escape(m->work_dir, ",:");
                if (!escaped_work_dir)
                        return log_oom();

                options = strjoina("lowerdir=", lower, ",upperdir=", escaped_source, ",workdir=", escaped_work_dir);
        }

        if (mount("overlay", where, "overlay", m->read_only ? MS_RDONLY : 0, options) < 0)
                return log_error_errno(errno, "overlay mount to %s failed: %m", where);

        return 0;
}

int mount_custom(
                const char *dest,
                CustomMount *mounts, unsigned n,
                bool userns, uid_t uid_shift, uid_t uid_range,
                const char *selinux_apifs_context) {

        unsigned i;
        int r;

        assert(dest);

        for (i = 0; i < n; i++) {
                CustomMount *m = mounts + i;

                switch (m->type) {

                case CUSTOM_MOUNT_BIND:
                        r = mount_bind(dest, m);
                        break;

                case CUSTOM_MOUNT_TMPFS:
                        r = mount_tmpfs(dest, m, userns, uid_shift, uid_range, selinux_apifs_context);
                        break;

                case CUSTOM_MOUNT_OVERLAY:
                        r = mount_overlay(dest, m);
                        break;

                default:
                        assert_not_reached("Unknown custom mount type");
                }

                if (r < 0)
                        return r;
        }

        return 0;
}

static int mount_legacy_cgroup_hierarchy(const char *dest, const char *controller, const char *hierarchy, bool read_only) {
        char *to;
        int r;

        to = strjoina(strempty(dest), "/sys/fs/cgroup/", hierarchy);

        r = path_is_mount_point(to, 0);
        if (r < 0 && r != -ENOENT)
                return log_error_errno(r, "Failed to determine if %s is mounted already: %m", to);
        if (r > 0)
                return 0;

        mkdir_p(to, 0755);

        /* The superblock mount options of the mount point need to be
         * identical to the hosts', and hence writable... */
        if (mount("cgroup", to, "cgroup", MS_NOSUID|MS_NOEXEC|MS_NODEV, controller) < 0)
                return log_error_errno(errno, "Failed to mount to %s: %m", to);

        /* ... hence let's only make the bind mount read-only, not the
         * superblock. */
        if (read_only) {
                if (mount(NULL, to, NULL, MS_BIND|MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, NULL) < 0)
                        return log_error_errno(errno, "Failed to remount %s read-only: %m", to);
        }
        return 1;
}

static int mount_legacy_cgroups(
                const char *dest,
                bool userns, uid_t uid_shift, uid_t uid_range,
                const char *selinux_apifs_context) {

        _cleanup_set_free_free_ Set *controllers = NULL;
        const char *cgroup_root;
        int r;

        cgroup_root = prefix_roota(dest, "/sys/fs/cgroup");

        /* Mount a tmpfs to /sys/fs/cgroup if it's not mounted there yet. */
        r = path_is_mount_point(cgroup_root, AT_SYMLINK_FOLLOW);
        if (r < 0)
                return log_error_errno(r, "Failed to determine if /sys/fs/cgroup is already mounted: %m");
        if (r == 0) {
                _cleanup_free_ char *options = NULL;

                r = tmpfs_patch_options("mode=755", userns, uid_shift, uid_range, selinux_apifs_context, &options);
                if (r < 0)
                        return log_oom();

                if (mount("tmpfs", cgroup_root, "tmpfs", MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME, options) < 0)
                        return log_error_errno(errno, "Failed to mount /sys/fs/cgroup: %m");
        }

        if (cg_unified() > 0)
                goto skip_controllers;

        controllers = set_new(&string_hash_ops);
        if (!controllers)
                return log_oom();

        r = cg_kernel_controllers(controllers);
        if (r < 0)
                return log_error_errno(r, "Failed to determine cgroup controllers: %m");

        for (;;) {
                _cleanup_free_ char *controller = NULL, *origin = NULL, *combined = NULL;

                controller = set_steal_first(controllers);
                if (!controller)
                        break;

                origin = prefix_root("/sys/fs/cgroup/", controller);
                if (!origin)
                        return log_oom();

                r = readlink_malloc(origin, &combined);
                if (r == -EINVAL) {
                        /* Not a symbolic link, but directly a single cgroup hierarchy */

                        r = mount_legacy_cgroup_hierarchy(dest, controller, controller, true);
                        if (r < 0)
                                return r;

                } else if (r < 0)
                        return log_error_errno(r, "Failed to read link %s: %m", origin);
                else {
                        _cleanup_free_ char *target = NULL;

                        target = prefix_root(dest, origin);
                        if (!target)
                                return log_oom();

                        /* A symbolic link, a combination of controllers in one hierarchy */

                        if (!filename_is_valid(combined)) {
                                log_warning("Ignoring invalid combined hierarchy %s.", combined);
                                continue;
                        }

                        r = mount_legacy_cgroup_hierarchy(dest, combined, combined, true);
                        if (r < 0)
                                return r;

                        r = symlink_idempotent(combined, target);
                        if (r == -EINVAL) {
                                log_error("Invalid existing symlink for combined hierarchy");
                                return r;
                        }
                        if (r < 0)
                                return log_error_errno(r, "Failed to create symlink for combined hierarchy: %m");
                }
        }

skip_controllers:
        r = mount_legacy_cgroup_hierarchy(dest, "none,name=systemd,xattr", "systemd", false);
        if (r < 0)
                return r;

        if (mount(NULL, cgroup_root, NULL, MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_STRICTATIME|MS_RDONLY, "mode=755") < 0)
                return log_error_errno(errno, "Failed to remount %s read-only: %m", cgroup_root);

        return 0;
}

static int mount_unified_cgroups(const char *dest) {
        const char *p;
        int r;

        assert(dest);

        p = strjoina(dest, "/sys/fs/cgroup");

        r = path_is_mount_point(p, AT_SYMLINK_FOLLOW);
        if (r < 0)
                return log_error_errno(r, "Failed to determine if %s is mounted already: %m", p);
        if (r > 0) {
                p = strjoina(dest, "/sys/fs/cgroup/cgroup.procs");
                if (access(p, F_OK) >= 0)
                        return 0;
                if (errno != ENOENT)
                        return log_error_errno(errno, "Failed to determine if mount point %s contains the unified cgroup hierarchy: %m", p);

                log_error("%s is already mounted but not a unified cgroup hierarchy. Refusing.", p);
                return -EINVAL;
        }

        if (mount("cgroup", p, "cgroup", MS_NOSUID|MS_NOEXEC|MS_NODEV, "__DEVEL__sane_behavior") < 0)
                return log_error_errno(errno, "Failed to mount unified cgroup hierarchy to %s: %m", p);

        return 0;
}

int mount_cgroups(
                const char *dest,
                bool unified_requested,
                bool userns, uid_t uid_shift, uid_t uid_range,
                const char *selinux_apifs_context) {

        if (unified_requested)
                return mount_unified_cgroups(dest);
        else
                return mount_legacy_cgroups(dest, userns, uid_shift, uid_range, selinux_apifs_context);
}

int mount_systemd_cgroup_writable(
                const char *dest,
                bool unified_requested) {

        _cleanup_free_ char *own_cgroup_path = NULL;
        const char *systemd_root, *systemd_own;
        int r;

        assert(dest);

        r = cg_pid_get_path(NULL, 0, &own_cgroup_path);
        if (r < 0)
                return log_error_errno(r, "Failed to determine our own cgroup path: %m");

        /* If we are living in the top-level, then there's nothing to do... */
        if (path_equal(own_cgroup_path, "/"))
                return 0;

        if (unified_requested) {
                systemd_own = strjoina(dest, "/sys/fs/cgroup", own_cgroup_path);
                systemd_root = prefix_roota(dest, "/sys/fs/cgroup");
        } else {
                systemd_own = strjoina(dest, "/sys/fs/cgroup/systemd", own_cgroup_path);
                systemd_root = prefix_roota(dest, "/sys/fs/cgroup/systemd");
        }

        /* Make our own cgroup a (writable) bind mount */
        if (mount(systemd_own, systemd_own,  NULL, MS_BIND, NULL) < 0)
                return log_error_errno(errno, "Failed to turn %s into a bind mount: %m", own_cgroup_path);

        /* And then remount the systemd cgroup root read-only */
        if (mount(NULL, systemd_root, NULL, MS_BIND|MS_REMOUNT|MS_NOSUID|MS_NOEXEC|MS_NODEV|MS_RDONLY, NULL) < 0)
                return log_error_errno(errno, "Failed to mount cgroup root read-only: %m");

        return 0;
}

int setup_volatile_state(
                const char *directory,
                VolatileMode mode,
                bool userns, uid_t uid_shift, uid_t uid_range,
                const char *selinux_apifs_context) {

        _cleanup_free_ char *buf = NULL;
        const char *p, *options;
        int r;

        assert(directory);

        if (mode != VOLATILE_STATE)
                return 0;

        /* --volatile=state means we simply overmount /var
           with a tmpfs, and the rest read-only. */

        r = bind_remount_recursive(directory, true);
        if (r < 0)
                return log_error_errno(r, "Failed to remount %s read-only: %m", directory);

        p = prefix_roota(directory, "/var");
        r = mkdir(p, 0755);
        if (r < 0 && errno != EEXIST)
                return log_error_errno(errno, "Failed to create %s: %m", directory);

        options = "mode=755";
        r = tmpfs_patch_options(options, userns, uid_shift, uid_range, selinux_apifs_context, &buf);
        if (r < 0)
                return log_oom();
        if (r > 0)
                options = buf;

        if (mount("tmpfs", p, "tmpfs", MS_STRICTATIME, options) < 0)
                return log_error_errno(errno, "Failed to mount tmpfs to /var: %m");

        return 0;
}

int setup_volatile(
                const char *directory,
                VolatileMode mode,
                bool userns, uid_t uid_shift, uid_t uid_range,
                const char *selinux_apifs_context) {

        bool tmpfs_mounted = false, bind_mounted = false;
        char template[] = "/tmp/nspawn-volatile-XXXXXX";
        _cleanup_free_ char *buf = NULL;
        const char *f, *t, *options;
        int r;

        assert(directory);

        if (mode != VOLATILE_YES)
                return 0;

        /* --volatile=yes means we mount a tmpfs to the root dir, and
           the original /usr to use inside it, and that read-only. */

        if (!mkdtemp(template))
                return log_error_errno(errno, "Failed to create temporary directory: %m");

        options = "mode=755";
        r = tmpfs_patch_options(options, userns, uid_shift, uid_range, selinux_apifs_context, &buf);
        if (r < 0)
                return log_oom();
        if (r > 0)
                options = buf;

        if (mount("tmpfs", template, "tmpfs", MS_STRICTATIME, options) < 0) {
                r = log_error_errno(errno, "Failed to mount tmpfs for root directory: %m");
                goto fail;
        }

        tmpfs_mounted = true;

        f = prefix_roota(directory, "/usr");
        t = prefix_roota(template, "/usr");

        r = mkdir(t, 0755);
        if (r < 0 && errno != EEXIST) {
                r = log_error_errno(errno, "Failed to create %s: %m", t);
                goto fail;
        }

        if (mount(f, t, NULL, MS_BIND|MS_REC, NULL) < 0) {
                r = log_error_errno(errno, "Failed to create /usr bind mount: %m");
                goto fail;
        }

        bind_mounted = true;

        r = bind_remount_recursive(t, true);
        if (r < 0) {
                log_error_errno(r, "Failed to remount %s read-only: %m", t);
                goto fail;
        }

        if (mount(template, directory, NULL, MS_MOVE, NULL) < 0) {
                r = log_error_errno(errno, "Failed to move root mount: %m");
                goto fail;
        }

        (void) rmdir(template);

        return 0;

fail:
        if (bind_mounted)
                (void) umount(t);

        if (tmpfs_mounted)
                (void) umount(template);
        (void) rmdir(template);
        return r;
}

VolatileMode volatile_mode_from_string(const char *s) {
        int b;

        if (isempty(s))
                return _VOLATILE_MODE_INVALID;

        b = parse_boolean(s);
        if (b > 0)
                return VOLATILE_YES;
        if (b == 0)
                return VOLATILE_NO;

        if (streq(s, "state"))
                return VOLATILE_STATE;

        return _VOLATILE_MODE_INVALID;
}