summaryrefslogtreecommitdiff
path: root/src/basic/mount-util.c
blob: bfa04394fe966084135c1a69f5c8303d58ab360f (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
/***
  This file is part of systemd.

  Copyright 2010 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 <errno.h>
#include <stdlib.h>
#include <string.h>
#include <sys/mount.h>
#include <sys/stat.h>
#include <sys/statvfs.h>
#include <unistd.h>

#include "alloc-util.h"
#include "escape.h"
#include "fd-util.h"
#include "fileio.h"
#include "hashmap.h"
#include "mount-util.h"
#include "parse-util.h"
#include "path-util.h"
#include "set.h"
#include "stdio-util.h"
#include "string-util.h"

static int fd_fdinfo_mnt_id(int fd, const char *filename, int flags, int *mnt_id) {
        char path[strlen("/proc/self/fdinfo/") + DECIMAL_STR_MAX(int)];
        _cleanup_free_ char *fdinfo = NULL;
        _cleanup_close_ int subfd = -1;
        char *p;
        int r;

        if ((flags & AT_EMPTY_PATH) && isempty(filename))
                xsprintf(path, "/proc/self/fdinfo/%i", fd);
        else {
                subfd = openat(fd, filename, O_CLOEXEC|O_PATH);
                if (subfd < 0)
                        return -errno;

                xsprintf(path, "/proc/self/fdinfo/%i", subfd);
        }

        r = read_full_file(path, &fdinfo, NULL);
        if (r == -ENOENT) /* The fdinfo directory is a relatively new addition */
                return -EOPNOTSUPP;
        if (r < 0)
                return -errno;

        p = startswith(fdinfo, "mnt_id:");
        if (!p) {
                p = strstr(fdinfo, "\nmnt_id:");
                if (!p) /* The mnt_id field is a relatively new addition */
                        return -EOPNOTSUPP;

                p += 8;
        }

        p += strspn(p, WHITESPACE);
        p[strcspn(p, WHITESPACE)] = 0;

        return safe_atoi(p, mnt_id);
}

int fd_is_mount_point(int fd, const char *filename, int flags) {
        union file_handle_union h = FILE_HANDLE_INIT, h_parent = FILE_HANDLE_INIT;
        int mount_id = -1, mount_id_parent = -1;
        bool nosupp = false, check_st_dev = true;
        struct stat a, b;
        int r;

        assert(fd >= 0);
        assert(filename);

        /* First we will try the name_to_handle_at() syscall, which
         * tells us the mount id and an opaque file "handle". It is
         * not supported everywhere though (kernel compile-time
         * option, not all file systems are hooked up). If it works
         * the mount id is usually good enough to tell us whether
         * something is a mount point.
         *
         * If that didn't work we will try to read the mount id from
         * /proc/self/fdinfo/<fd>. This is almost as good as
         * name_to_handle_at(), however, does not return the
         * opaque file handle. The opaque file handle is pretty useful
         * to detect the root directory, which we should always
         * consider a mount point. Hence we use this only as
         * fallback. Exporting the mnt_id in fdinfo is a pretty recent
         * kernel addition.
         *
         * As last fallback we do traditional fstat() based st_dev
         * comparisons. This is how things were traditionally done,
         * but unionfs breaks this since it exposes file
         * systems with a variety of st_dev reported. Also, btrfs
         * subvolumes have different st_dev, even though they aren't
         * real mounts of their own. */

        r = name_to_handle_at(fd, filename, &h.handle, &mount_id, flags);
        if (r < 0) {
                if (errno == ENOSYS)
                        /* This kernel does not support name_to_handle_at()
                         * fall back to simpler logic. */
                        goto fallback_fdinfo;
                else if (errno == EOPNOTSUPP)
                        /* This kernel or file system does not support
                         * name_to_handle_at(), hence let's see if the
                         * upper fs supports it (in which case it is a
                         * mount point), otherwise fallback to the
                         * traditional stat() logic */
                        nosupp = true;
                else
                        return -errno;
        }

        r = name_to_handle_at(fd, "", &h_parent.handle, &mount_id_parent, AT_EMPTY_PATH);
        if (r < 0) {
                if (errno == EOPNOTSUPP) {
                        if (nosupp)
                                /* Neither parent nor child do name_to_handle_at()?
                                   We have no choice but to fall back. */
                                goto fallback_fdinfo;
                        else
                                /* The parent can't do name_to_handle_at() but the
                                 * directory we are interested in can?
                                 * If so, it must be a mount point. */
                                return 1;
                } else
                        return -errno;
        }

        /* The parent can do name_to_handle_at() but the
         * directory we are interested in can't? If so, it
         * must be a mount point. */
        if (nosupp)
                return 1;

        /* If the file handle for the directory we are
         * interested in and its parent are identical, we
         * assume this is the root directory, which is a mount
         * point. */

        if (h.handle.handle_bytes == h_parent.handle.handle_bytes &&
            h.handle.handle_type == h_parent.handle.handle_type &&
            memcmp(h.handle.f_handle, h_parent.handle.f_handle, h.handle.handle_bytes) == 0)
                return 1;

        return mount_id != mount_id_parent;

fallback_fdinfo:
        r = fd_fdinfo_mnt_id(fd, filename, flags, &mount_id);
        if (r == -EOPNOTSUPP)
                goto fallback_fstat;
        if (r < 0)
                return r;

        r = fd_fdinfo_mnt_id(fd, "", AT_EMPTY_PATH, &mount_id_parent);
        if (r < 0)
                return r;

        if (mount_id != mount_id_parent)
                return 1;

        /* Hmm, so, the mount ids are the same. This leaves one
         * special case though for the root file system. For that,
         * let's see if the parent directory has the same inode as we
         * are interested in. Hence, let's also do fstat() checks now,
         * too, but avoid the st_dev comparisons, since they aren't
         * that useful on unionfs mounts. */
        check_st_dev = false;

fallback_fstat:
        /* yay for fstatat() taking a different set of flags than the other
         * _at() above */
        if (flags & AT_SYMLINK_FOLLOW)
                flags &= ~AT_SYMLINK_FOLLOW;
        else
                flags |= AT_SYMLINK_NOFOLLOW;
        if (fstatat(fd, filename, &a, flags) < 0)
                return -errno;

        if (fstatat(fd, "", &b, AT_EMPTY_PATH) < 0)
                return -errno;

        /* A directory with same device and inode as its parent? Must
         * be the root directory */
        if (a.st_dev == b.st_dev &&
            a.st_ino == b.st_ino)
                return 1;

        return check_st_dev && (a.st_dev != b.st_dev);
}

/* flags can be AT_SYMLINK_FOLLOW or 0 */
int path_is_mount_point(const char *t, int flags) {
        _cleanup_close_ int fd = -1;
        _cleanup_free_ char *canonical = NULL, *parent = NULL;

        assert(t);

        if (path_equal(t, "/"))
                return 1;

        /* we need to resolve symlinks manually, we can't just rely on
         * fd_is_mount_point() to do that for us; if we have a structure like
         * /bin -> /usr/bin/ and /usr is a mount point, then the parent that we
         * look at needs to be /usr, not /. */
        if (flags & AT_SYMLINK_FOLLOW) {
                canonical = canonicalize_file_name(t);
                if (!canonical)
                        return -errno;

                t = canonical;
        }

        parent = dirname_malloc(t);
        if (!parent)
                return -ENOMEM;

        fd = openat(AT_FDCWD, parent, O_DIRECTORY|O_CLOEXEC|O_PATH);
        if (fd < 0)
                return -errno;

        return fd_is_mount_point(fd, basename(t), flags);
}

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

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

        do {
                _cleanup_fclose_ FILE *proc_self_mountinfo = NULL;

                again = false;
                r = 0;

                proc_self_mountinfo = fopen("/proc/self/mountinfo", "re");
                if (!proc_self_mountinfo)
                        return -errno;

                for (;;) {
                        _cleanup_free_ char *path = NULL, *p = NULL;
                        int k;

                        k = fscanf(proc_self_mountinfo,
                                   "%*s "       /* (1) mount id */
                                   "%*s "       /* (2) parent id */
                                   "%*s "       /* (3) major:minor */
                                   "%*s "       /* (4) root */
                                   "%ms "       /* (5) mount point */
                                   "%*s"        /* (6) mount options */
                                   "%*[^-]"     /* (7) optional fields */
                                   "- "         /* (8) separator */
                                   "%*s "       /* (9) file system type */
                                   "%*s"        /* (10) mount source */
                                   "%*s"        /* (11) mount options 2 */
                                   "%*[^\n]",   /* some rubbish at the end */
                                   &path);
                        if (k != 1) {
                                if (k == EOF)
                                        break;

                                continue;
                        }

                        r = cunescape(path, UNESCAPE_RELAX, &p);
                        if (r < 0)
                                return r;

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

                        if (umount2(p, flags) < 0) {
                                r = -errno;
                                continue;
                        }

                        again = true;
                        n++;

                        break;
                }

        } while (again);

        return r ? r : n;
}

static int get_mount_flags(const char *path, unsigned long *flags) {
        struct statvfs buf;

        if (statvfs(path, &buf) < 0)
                return -errno;
        *flags = buf.f_flag;
        return 0;
}

int bind_remount_recursive(const char *prefix, bool ro) {
        _cleanup_set_free_free_ Set *done = NULL;
        _cleanup_free_ char *cleaned = NULL;
        int r;

        /* 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 migt be writable. This includes future
         * submounts that have been triggered via autofs. */

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

        path_kill_slashes(cleaned);

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

        for (;;) {
                _cleanup_fclose_ FILE *proc_self_mountinfo = NULL;
                _cleanup_set_free_free_ Set *todo = NULL;
                bool top_autofs = false;
                char *x;
                unsigned long orig_flags;

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

                proc_self_mountinfo = fopen("/proc/self/mountinfo", "re");
                if (!proc_self_mountinfo)
                        return -errno;

                for (;;) {
                        _cleanup_free_ char *path = NULL, *p = NULL, *type = NULL;
                        int k;

                        k = fscanf(proc_self_mountinfo,
                                   "%*s "       /* (1) mount id */
                                   "%*s "       /* (2) parent id */
                                   "%*s "       /* (3) major:minor */
                                   "%*s "       /* (4) root */
                                   "%ms "       /* (5) mount point */
                                   "%*s"        /* (6) mount options (superblock) */
                                   "%*[^-]"     /* (7) optional fields */
                                   "- "         /* (8) separator */
                                   "%ms "       /* (9) file system type */
                                   "%*s"        /* (10) mount source */
                                   "%*s"        /* (11) mount options (bind mount) */
                                   "%*[^\n]",   /* some rubbish at the end */
                                   &path,
                                   &type);
                        if (k != 2) {
                                if (k == EOF)
                                        break;

                                continue;
                        }

                        r = cunescape(path, UNESCAPE_RELAX, &p);
                        if (r < 0)
                                return r;

                        /* 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(cleaned, p);
                                continue;
                        }

                        if (path_startswith(p, cleaned) &&
                            !set_contains(done, p)) {

                                r = set_consume(todo, p);
                                p = NULL;

                                if (r == -EEXIST)
                                        continue;
                                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, cleaned)))
                        return 0;

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

                        orig_flags = 0;
                        (void) get_mount_flags(cleaned, &orig_flags);
                        orig_flags &= ~MS_RDONLY;

                        if (mount(NULL, prefix, NULL, orig_flags|MS_BIND|MS_REMOUNT|(ro ? MS_RDONLY : 0), NULL) < 0)
                                return -errno;

                        x = strdup(cleaned);
                        if (!x)
                                return -ENOMEM;

                        r = set_consume(done, x);
                        if (r < 0)
                                return r;
                }

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

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

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

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

                        if (mount(NULL, x, NULL, orig_flags|MS_BIND|MS_REMOUNT|(ro ? MS_RDONLY : 0), NULL) < 0)
                                return -errno;

                }
        }
}

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;
}

bool fstype_is_network(const char *fstype) {
        static const char table[] =
                "afs\0"
                "cifs\0"
                "smbfs\0"
                "sshfs\0"
                "ncpfs\0"
                "ncp\0"
                "nfs\0"
                "nfs4\0"
                "gfs\0"
                "gfs2\0"
                "glusterfs\0"
                "pvfs2\0" /* OrangeFS */
                "ocfs2\0"
                ;

        const char *x;

        x = startswith(fstype, "fuse.");
        if (x)
                fstype = x;

        return nulstr_contains(table, fstype);
}

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;
        }
}

const char* mode_to_inaccessible_node(mode_t mode) {
        /* This function maps a node type to the correspondent inaccessible node type.
         * Character and block inaccessible devices may not be created (because major=0 and minor=0),
         * in such case we map character and block devices to the inaccessible node type socket. */
        switch(mode & S_IFMT) {
                case S_IFREG:
                        return "/run/systemd/inaccessible/reg";
                case S_IFDIR:
                        return "/run/systemd/inaccessible/dir";
                case S_IFCHR:
                        if (access("/run/systemd/inaccessible/chr", F_OK) == 0)
                                return "/run/systemd/inaccessible/chr";
                        return "/run/systemd/inaccessible/sock";
                case S_IFBLK:
                        if (access("/run/systemd/inaccessible/blk", F_OK) == 0)
                                return "/run/systemd/inaccessible/blk";
                        return "/run/systemd/inaccessible/sock";
                case S_IFIFO:
                        return "/run/systemd/inaccessible/fifo";
                case S_IFSOCK:
                        return "/run/systemd/inaccessible/sock";
        }
        return NULL;
}