| 1 |
.\" Copyright 1993, Rickard E. Faith <[email protected]> |
| 2 |
.\" Copyright 1994, Andries E. Brouwer <[email protected]> |
| 3 |
.\" Copyright 2002-2016, Michael Kerrisk <[email protected]> |
| 4 |
.\" Copyright, the authors of the Linux man-pages project |
| 5 |
.\" |
| 6 |
.\" SPDX-License-Identifier: Linux-man-pages-copyleft |
| 7 |
.\" |
| 8 |
.TH mount 2 2026-02-11 "Linux man-pages 6.18" |
| 9 |
.SH NAME |
| 10 |
mount \- mount filesystem |
| 11 |
.SH LIBRARY |
| 12 |
Standard C library |
| 13 |
.RI ( libc ,\~ \-lc ) |
| 14 |
.SH SYNOPSIS |
| 15 |
.nf |
| 16 |
.B "#include <sys/mount.h>" |
| 17 |
.P |
| 18 |
.BI "int mount(const char *" source ", const char *" target , |
| 19 |
.BI " const char *" filesystemtype ", unsigned long " mountflags , |
| 20 |
.BI " const void *_Nullable " data ); |
| 21 |
.fi |
| 22 |
.SH DESCRIPTION |
| 23 |
.BR mount () |
| 24 |
attaches the filesystem specified by |
| 25 |
.I source |
| 26 |
(which is often a pathname referring to a device, |
| 27 |
but can also be the pathname of a directory or file, |
| 28 |
or a dummy string) to the location (a directory or file) |
| 29 |
specified by the pathname in |
| 30 |
.IR target . |
| 31 |
.P |
| 32 |
Appropriate privilege (Linux: the |
| 33 |
.B CAP_SYS_ADMIN |
| 34 |
capability) is required to mount filesystems. |
| 35 |
.P |
| 36 |
Values for the |
| 37 |
.I filesystemtype |
| 38 |
argument supported by the kernel are listed in |
| 39 |
.I /proc/filesystems |
| 40 |
(e.g., "btrfs", "ext4", "jfs", "xfs", "vfat", "fuse", |
| 41 |
"tmpfs", "cgroup", "proc", "mqueue", "nfs", "cifs", "iso9660"). |
| 42 |
Further types may become available when the appropriate modules |
| 43 |
are loaded. |
| 44 |
.P |
| 45 |
The |
| 46 |
.I data |
| 47 |
argument is interpreted by the different filesystems. |
| 48 |
Typically it is a string of comma-separated options |
| 49 |
understood by this filesystem. |
| 50 |
See |
| 51 |
.BR mount (8) |
| 52 |
for details of the options available for each filesystem type. |
| 53 |
This argument may be specified as NULL, if there are no options. |
| 54 |
.P |
| 55 |
A call to |
| 56 |
.BR mount () |
| 57 |
performs one of a number of general types of operation, |
| 58 |
depending on the bits specified in |
| 59 |
.IR mountflags . |
| 60 |
The choice of which operation to perform is determined by |
| 61 |
testing the bits set in |
| 62 |
.IR mountflags , |
| 63 |
with the tests being conducted in the order listed here: |
| 64 |
.IP \[bu] 3 |
| 65 |
Remount an existing mount: |
| 66 |
.I mountflags |
| 67 |
includes |
| 68 |
.BR MS_REMOUNT . |
| 69 |
.IP \[bu] |
| 70 |
Create a bind mount: |
| 71 |
.I mountflags |
| 72 |
includes |
| 73 |
.BR MS_BIND . |
| 74 |
.IP \[bu] |
| 75 |
Change the propagation type of an existing mount: |
| 76 |
.I mountflags |
| 77 |
includes one of |
| 78 |
.BR MS_SHARED , |
| 79 |
.BR MS_PRIVATE , |
| 80 |
.BR MS_SLAVE , |
| 81 |
or |
| 82 |
.BR MS_UNBINDABLE . |
| 83 |
.IP \[bu] |
| 84 |
Move an existing mount to a new location: |
| 85 |
.I mountflags |
| 86 |
includes |
| 87 |
.BR MS_MOVE . |
| 88 |
.IP \[bu] |
| 89 |
Create a new mount: |
| 90 |
.I mountflags |
| 91 |
includes none of the above flags. |
| 92 |
.P |
| 93 |
Each of these operations is detailed later in this page. |
| 94 |
Further flags may be specified in |
| 95 |
.I mountflags |
| 96 |
to modify the behavior of |
| 97 |
.BR mount (), |
| 98 |
as described below. |
| 99 |
.\" |
| 100 |
.SS Additional mount flags |
| 101 |
The list below describes the additional flags that can be specified in |
| 102 |
.IR mountflags . |
| 103 |
Note that some operation types ignore some or all of these flags, |
| 104 |
as described later in this page. |
| 105 |
.\" |
| 106 |
.\" FIXME 2.6.25 Added MS_I_VERSION, which needs to be documented. |
| 107 |
.\" commit 7a224228ed79d587ece2304869000aad1b8e97dd |
| 108 |
.\" (This is a per-superblock flag) |
| 109 |
.\" |
| 110 |
.TP |
| 111 |
.BR MS_DIRSYNC " (since Linux 2.5.19)" |
| 112 |
Make directory changes on this filesystem synchronous. |
| 113 |
(This property can be obtained for individual directories |
| 114 |
or subtrees using |
| 115 |
.BR chattr (1).) |
| 116 |
.TP |
| 117 |
.BR MS_LAZYTIME " (since Linux 4.0)" |
| 118 |
.\" commit 0ae45f63d4ef8d8eeec49c7d8b44a1775fff13e8 |
| 119 |
.\" commit fe032c422c5ba562ba9c2d316f55e258e03259c6 |
| 120 |
.\" commit a26f49926da938f47561f386be56a83dd37a496d |
| 121 |
Reduce on-disk updates of inode timestamps (atime, mtime, ctime) |
| 122 |
by maintaining these changes only in memory. |
| 123 |
The on-disk timestamps are updated only when: |
| 124 |
.RS |
| 125 |
.IP \[bu] 3 |
| 126 |
the inode needs to be updated for some change unrelated to file timestamps; |
| 127 |
.IP \[bu] |
| 128 |
the application employs |
| 129 |
.BR fsync (2), |
| 130 |
.BR syncfs (2), |
| 131 |
or |
| 132 |
.BR sync (2); |
| 133 |
.IP \[bu] |
| 134 |
an undeleted inode is evicted from memory; |
| 135 |
or |
| 136 |
.IP \[bu] |
| 137 |
more than 24 hours have passed since the inode was written to disk. |
| 138 |
.RE |
| 139 |
.IP |
| 140 |
This mount option significantly reduces writes |
| 141 |
needed to update the inode's timestamps, especially mtime and atime. |
| 142 |
However, in the event of a system crash, the atime and mtime fields |
| 143 |
on disk might be out of date by up to 24 hours. |
| 144 |
.IP |
| 145 |
Examples of workloads where this option could be of significant benefit |
| 146 |
include frequent random writes to preallocated files, |
| 147 |
as well as cases where the |
| 148 |
.B MS_STRICTATIME |
| 149 |
mount option is also enabled. |
| 150 |
(The advantage of combining |
| 151 |
.B MS_STRICTATIME |
| 152 |
and |
| 153 |
.B MS_LAZYTIME |
| 154 |
is that |
| 155 |
.BR stat (2) |
| 156 |
will return the correctly updated atime, but the atime updates |
| 157 |
will be flushed to disk only in the cases listed above.) |
| 158 |
.TP |
| 159 |
.B MS_MANDLOCK |
| 160 |
Permit mandatory locking on files in this filesystem. |
| 161 |
(Mandatory locking must still be enabled on a per-file basis, |
| 162 |
as described in |
| 163 |
.BR fcntl (2).) |
| 164 |
Since Linux 4.5, |
| 165 |
.\" commit 95ace75414f312f9a7b93d873f386987b92a5301 |
| 166 |
this mount option requires the |
| 167 |
.B CAP_SYS_ADMIN |
| 168 |
capability and a kernel configured with the |
| 169 |
.B CONFIG_MANDATORY_FILE_LOCKING |
| 170 |
option. |
| 171 |
Mandatory locking has been fully deprecated in Linux 5.15, so |
| 172 |
this flag should be considered deprecated. |
| 173 |
.TP |
| 174 |
.B MS_NOATIME |
| 175 |
Do not update access times for (all types of) files on this filesystem. |
| 176 |
.TP |
| 177 |
.B MS_NODEV |
| 178 |
Do not allow access to devices (special files) on this filesystem. |
| 179 |
.TP |
| 180 |
.B MS_NODIRATIME |
| 181 |
Do not update access times for directories on this filesystem. |
| 182 |
This flag provides a subset of the functionality provided by |
| 183 |
.BR MS_NOATIME ; |
| 184 |
that is, |
| 185 |
.B MS_NOATIME |
| 186 |
implies |
| 187 |
.BR MS_NODIRATIME . |
| 188 |
.TP |
| 189 |
.B MS_NOEXEC |
| 190 |
Do not allow programs to be executed from this filesystem. |
| 191 |
.\" (Possibly useful for a filesystem that contains non-Linux executables. |
| 192 |
.\" Often used as a security feature, e.g., to make sure that restricted |
| 193 |
.\" users cannot execute files uploaded using ftp or so.) |
| 194 |
.TP |
| 195 |
.B MS_NOSUID |
| 196 |
Do not honor set-user-ID and set-group-ID bits or file capabilities |
| 197 |
when executing programs from this filesystem. |
| 198 |
In addition, SELinux domain |
| 199 |
transitions require the permission |
| 200 |
.IR nosuid_transition , |
| 201 |
which in turn needs |
| 202 |
also the policy capability |
| 203 |
.IR nnp_nosuid_transition . |
| 204 |
.\" (This is a security feature to prevent users executing set-user-ID and |
| 205 |
.\" set-group-ID programs from removable disk devices.) |
| 206 |
.TP |
| 207 |
.B MS_RDONLY |
| 208 |
Mount filesystem read-only. |
| 209 |
.TP |
| 210 |
.BR MS_REC " (since Linux 2.4.11)" |
| 211 |
Used in conjunction with |
| 212 |
.B MS_BIND |
| 213 |
to create a recursive bind mount, |
| 214 |
and in conjunction with the propagation type flags to recursively change |
| 215 |
the propagation type of all of the mounts in a subtree. |
| 216 |
See below for further details. |
| 217 |
.TP |
| 218 |
.BR MS_RELATIME " (since Linux 2.6.20)" |
| 219 |
When a file on this filesystem is accessed, |
| 220 |
update the file's last access time (atime) only if the current value |
| 221 |
of atime is less than or equal to the file's last modification time (mtime) |
| 222 |
or last status change time (ctime). |
| 223 |
This option is useful for programs, such as |
| 224 |
.BR mutt (1), |
| 225 |
that need to know when a file has been read since it was last modified. |
| 226 |
Since Linux 2.6.30, the kernel defaults to the behavior provided |
| 227 |
by this flag (unless |
| 228 |
.B MS_NOATIME |
| 229 |
was specified), and the |
| 230 |
.B MS_STRICTATIME |
| 231 |
flag is required to obtain traditional semantics. |
| 232 |
In addition, since Linux 2.6.30, |
| 233 |
the file's last access time is always updated if it |
| 234 |
is more than 1 day old. |
| 235 |
.\" Matthew Garrett notes in the patch that added this behavior |
| 236 |
.\" that this lets utilities such as tmpreaper (which deletes |
| 237 |
.\" files based on last access time) work correctly. |
| 238 |
.TP |
| 239 |
.BR MS_SILENT " (since Linux 2.6.17)" |
| 240 |
Suppress the display of certain |
| 241 |
.RI ( printk ()) |
| 242 |
warning messages in the kernel log. |
| 243 |
This flag supersedes the misnamed and obsolete |
| 244 |
.B MS_VERBOSE |
| 245 |
flag (available since Linux 2.4.12), which has the same meaning. |
| 246 |
.TP |
| 247 |
.BR MS_STRICTATIME " (since Linux 2.6.30)" |
| 248 |
Always update the last access time (atime) when files on this |
| 249 |
filesystem are accessed. |
| 250 |
(This was the default behavior before Linux 2.6.30.) |
| 251 |
Specifying this flag overrides the effect of setting the |
| 252 |
.B MS_NOATIME |
| 253 |
and |
| 254 |
.B MS_RELATIME |
| 255 |
flags. |
| 256 |
.TP |
| 257 |
.B MS_SYNCHRONOUS |
| 258 |
Make writes on this filesystem synchronous (as though |
| 259 |
the |
| 260 |
.B O_SYNC |
| 261 |
flag to |
| 262 |
.BR open (2) |
| 263 |
was specified for all file opens to this filesystem). |
| 264 |
.TP |
| 265 |
.BR MS_NOSYMFOLLOW " (since Linux 5.10)" |
| 266 |
.\" dab741e0e02bd3c4f5e2e97be74b39df2523fc6e |
| 267 |
Do not follow symbolic links when resolving paths. |
| 268 |
Symbolic links can still be created, |
| 269 |
and |
| 270 |
.BR readlink (1), |
| 271 |
.BR readlink (2), |
| 272 |
.BR realpath (1), |
| 273 |
and |
| 274 |
.BR realpath (3) |
| 275 |
all still work properly. |
| 276 |
.P |
| 277 |
From Linux 2.4 onward, some of the above flags are |
| 278 |
settable on a per-mount basis, |
| 279 |
while others apply to the superblock of the mounted filesystem, |
| 280 |
meaning that all mounts of the same filesystem share those flags. |
| 281 |
(Previously, all of the flags were per-superblock.) |
| 282 |
.P |
| 283 |
The per-mount-point flags are as follows: |
| 284 |
.IP \[bu] 3 |
| 285 |
Since Linux 2.4: |
| 286 |
.BR MS_NODEV ", " MS_NOEXEC ", and " MS_NOSUID |
| 287 |
flags are settable on a per-mount-point basis. |
| 288 |
.IP \[bu] |
| 289 |
Additionally, since Linux 2.6.16: |
| 290 |
.B MS_NOATIME |
| 291 |
and |
| 292 |
.BR MS_NODIRATIME . |
| 293 |
.IP \[bu] |
| 294 |
Additionally, since Linux 2.6.20: |
| 295 |
.BR MS_RELATIME . |
| 296 |
.P |
| 297 |
The following flags are per-superblock: |
| 298 |
.BR MS_DIRSYNC , |
| 299 |
.BR MS_LAZYTIME , |
| 300 |
.BR MS_MANDLOCK , |
| 301 |
.BR MS_SILENT , |
| 302 |
and |
| 303 |
.BR MS_SYNCHRONOUS . |
| 304 |
.\" And MS_I_VERSION? |
| 305 |
The initial settings of these flags are determined on the first |
| 306 |
mount of the filesystem, and will be shared by all subsequent mounts |
| 307 |
of the same filesystem. |
| 308 |
Subsequently, the settings of the flags can be changed |
| 309 |
via a remount operation (see below). |
| 310 |
Such changes will be visible via all mounts associated |
| 311 |
with the filesystem. |
| 312 |
.P |
| 313 |
Since Linux 2.6.16, |
| 314 |
.B MS_RDONLY |
| 315 |
can be set or cleared on a per-mount-point basis as well as on |
| 316 |
the underlying filesystem superblock. |
| 317 |
The mounted filesystem will be writable only if neither the filesystem |
| 318 |
nor the mount point are flagged as read-only. |
| 319 |
.\" |
| 320 |
.SS Remounting an existing mount |
| 321 |
An existing mount may be remounted by specifying |
| 322 |
.B MS_REMOUNT |
| 323 |
in |
| 324 |
.IR mountflags . |
| 325 |
This allows you to change the |
| 326 |
.I mountflags |
| 327 |
and |
| 328 |
.I data |
| 329 |
of an existing mount without having to unmount and remount the filesystem. |
| 330 |
.I target |
| 331 |
should be the same value specified in the initial |
| 332 |
.BR mount () |
| 333 |
call. |
| 334 |
.P |
| 335 |
The |
| 336 |
.I source |
| 337 |
and |
| 338 |
.I filesystemtype |
| 339 |
arguments are ignored. |
| 340 |
.P |
| 341 |
The |
| 342 |
.I mountflags |
| 343 |
and |
| 344 |
.I data |
| 345 |
arguments should match the values used in the original |
| 346 |
.BR mount () |
| 347 |
call, except for those parameters that are being deliberately changed. |
| 348 |
.P |
| 349 |
The following |
| 350 |
.I mountflags |
| 351 |
can be changed: |
| 352 |
.BR MS_LAZYTIME , |
| 353 |
.\" FIXME |
| 354 |
.\" MS_LAZYTIME seems to be available only on a few filesystems, |
| 355 |
.\" and on ext4, it seems (from experiment that this flag |
| 356 |
.\" can only be enabled (but not disabled) on a remount. |
| 357 |
.\" The following code in ext4_remount() (kernel 4.17) seems to |
| 358 |
.\" confirm this: |
| 359 |
.\" |
| 360 |
.\" if (*flags & SB_LAZYTIME) |
| 361 |
.\" sb->s_flags |= SB_LAZYTIME; |
| 362 |
.BR MS_MANDLOCK , |
| 363 |
.BR MS_NOATIME , |
| 364 |
.BR MS_NODEV , |
| 365 |
.BR MS_NODIRATIME , |
| 366 |
.BR MS_NOEXEC , |
| 367 |
.BR MS_NOSUID , |
| 368 |
.BR MS_RELATIME , |
| 369 |
.BR MS_RDONLY , |
| 370 |
.B MS_STRICTATIME |
| 371 |
(whose effect is to clear the |
| 372 |
.B MS_NOATIME |
| 373 |
and |
| 374 |
.B MS_RELATIME |
| 375 |
flags), |
| 376 |
and |
| 377 |
.BR MS_SYNCHRONOUS . |
| 378 |
Attempts to change the setting of the |
| 379 |
.\" See the definition of MS_RMT_MASK in include/uapi/linux/fs.h, |
| 380 |
.\" which excludes MS_DIRSYNC and MS_SILENT, although SB_DIRSYNC |
| 381 |
.\" and SB_SILENT are split out as per-superblock flags in do_mount() |
| 382 |
.\" (Linux 4.17 source code) |
| 383 |
.B MS_DIRSYNC |
| 384 |
and |
| 385 |
.B MS_SILENT |
| 386 |
flags during a remount are silently ignored. |
| 387 |
Note that changes to per-superblock flags are visible via |
| 388 |
all mounts of the associated filesystem |
| 389 |
(because the per-superblock flags are shared by all mounts). |
| 390 |
.P |
| 391 |
Since Linux 3.17, |
| 392 |
.\" commit ffbc6f0ead47fa5a1dc9642b0331cb75c20a640e |
| 393 |
if none of |
| 394 |
.BR MS_NOATIME , |
| 395 |
.BR MS_NODIRATIME , |
| 396 |
.BR MS_RELATIME , |
| 397 |
or |
| 398 |
.B MS_STRICTATIME |
| 399 |
is specified in |
| 400 |
.IR mountflags , |
| 401 |
then the remount operation preserves the existing values of these flags |
| 402 |
(rather than defaulting to |
| 403 |
.BR MS_RELATIME ). |
| 404 |
.P |
| 405 |
Since Linux 2.6.26, the |
| 406 |
.B MS_REMOUNT |
| 407 |
flag can be used with |
| 408 |
.B MS_BIND |
| 409 |
to modify only the per-mount-point flags. |
| 410 |
.\" See https://lwn.net/Articles/281157/ |
| 411 |
This is particularly useful |
| 412 |
for setting or clearing the "read-only" flag on a mount |
| 413 |
without changing the filesystem parameters of the underlying filesystem. |
| 414 |
The |
| 415 |
.I data |
| 416 |
argument is ignored. |
| 417 |
The existing per-mount-point flags of the mount point |
| 418 |
are cleared |
| 419 |
and replaced with those in |
| 420 |
.IR mountflags . |
| 421 |
This means that |
| 422 |
if you wish to preserve |
| 423 |
any existing per-mount-point flags, |
| 424 |
you need to include them in |
| 425 |
.IR mountflags , |
| 426 |
along with the per-mount-point flags you wish to set |
| 427 |
(or with the flags you wish to clear missing). |
| 428 |
Specifying |
| 429 |
.I mountflags |
| 430 |
as: |
| 431 |
.P |
| 432 |
.in +4n |
| 433 |
.EX |
| 434 |
MS_REMOUNT | MS_BIND | MS_RDONLY |
| 435 |
.EE |
| 436 |
.in |
| 437 |
.P |
| 438 |
will make access through this mount point read-only |
| 439 |
(clearing all other per-mount-point flags), |
| 440 |
without affecting other mounts of this filesystem. |
| 441 |
.\" |
| 442 |
.SS Creating a bind mount |
| 443 |
If |
| 444 |
.I mountflags |
| 445 |
includes |
| 446 |
.B MS_BIND |
| 447 |
(available since Linux 2.4), |
| 448 |
.\" since Linux 2.4.0-test9 |
| 449 |
then perform a bind mount. |
| 450 |
A bind mount makes a file or a directory subtree visible at |
| 451 |
another point within the single directory hierarchy. |
| 452 |
Bind mounts may cross filesystem boundaries and span |
| 453 |
.BR chroot (2) |
| 454 |
jails. |
| 455 |
.P |
| 456 |
The |
| 457 |
.I filesystemtype |
| 458 |
and |
| 459 |
.I data |
| 460 |
arguments are ignored. |
| 461 |
.P |
| 462 |
The remaining bits (other than |
| 463 |
.BR MS_REC , |
| 464 |
described below) in the |
| 465 |
.I mountflags |
| 466 |
argument are also ignored. |
| 467 |
(The bind mount has the same mount options as |
| 468 |
the underlying mount.) |
| 469 |
However, see the discussion of remounting above, |
| 470 |
for a method of making an existing bind mount read-only. |
| 471 |
.P |
| 472 |
By default, when a directory is bind mounted, |
| 473 |
only that directory is mounted; |
| 474 |
if there are any submounts under the directory tree, |
| 475 |
they are not bind mounted. |
| 476 |
If the |
| 477 |
.B MS_REC |
| 478 |
flag is also specified, then a recursive bind mount operation is performed: |
| 479 |
all submounts under the |
| 480 |
.I source |
| 481 |
subtree (other than unbindable mounts) |
| 482 |
are also bind mounted at the corresponding location in the |
| 483 |
.I target |
| 484 |
subtree. |
| 485 |
.\" |
| 486 |
.SS Changing the propagation type of an existing mount |
| 487 |
If |
| 488 |
.I mountflags |
| 489 |
includes one of |
| 490 |
.BR MS_SHARED , |
| 491 |
.BR MS_PRIVATE , |
| 492 |
.BR MS_SLAVE , |
| 493 |
or |
| 494 |
.B MS_UNBINDABLE |
| 495 |
(all available since Linux 2.6.15), |
| 496 |
then the propagation type of an existing mount is changed. |
| 497 |
If more than one of these flags is specified, an error results. |
| 498 |
.P |
| 499 |
The only other flags that can be specified while changing |
| 500 |
the propagation type are |
| 501 |
.B MS_REC |
| 502 |
(described below) and |
| 503 |
.B MS_SILENT |
| 504 |
(which is ignored). |
| 505 |
.P |
| 506 |
The |
| 507 |
.IR source , |
| 508 |
.IR filesystemtype , |
| 509 |
and |
| 510 |
.I data |
| 511 |
arguments are ignored. |
| 512 |
.P |
| 513 |
The meanings of the propagation type flags are as follows: |
| 514 |
.TP |
| 515 |
.B MS_SHARED |
| 516 |
Make this mount shared. |
| 517 |
Mount and unmount events immediately under this mount will propagate |
| 518 |
to the other mounts that are members of this mount's peer group. |
| 519 |
Propagation here means that the same mount or unmount will automatically |
| 520 |
occur under all of the other mounts in the peer group. |
| 521 |
Conversely, mount and unmount events that take place under |
| 522 |
peer mounts will propagate to this mount. |
| 523 |
.TP |
| 524 |
.B MS_PRIVATE |
| 525 |
Make this mount private. |
| 526 |
Mount and unmount events do not propagate into or out of this mount. |
| 527 |
.TP |
| 528 |
.B MS_SLAVE |
| 529 |
If this is a shared mount that is a member of a peer group |
| 530 |
that contains other members, convert it to a slave mount. |
| 531 |
If this is a shared mount that is a member of a peer group |
| 532 |
that contains no other members, convert it to a private mount. |
| 533 |
Otherwise, the propagation type of the mount is left unchanged. |
| 534 |
.IP |
| 535 |
When a mount is a slave, |
| 536 |
mount and unmount events propagate into this mount from |
| 537 |
the (master) shared peer group of which it was formerly a member. |
| 538 |
Mount and unmount events under this mount do not propagate to any peer. |
| 539 |
.IP |
| 540 |
A mount can be the slave of another peer group |
| 541 |
while at the same time sharing mount and unmount events |
| 542 |
with a peer group of which it is a member. |
| 543 |
.TP |
| 544 |
.B MS_UNBINDABLE |
| 545 |
Make this mount unbindable. |
| 546 |
This is like a private mount, |
| 547 |
and in addition this mount can't be bind mounted. |
| 548 |
When a recursive bind mount |
| 549 |
.RB ( mount () |
| 550 |
with the |
| 551 |
.B MS_BIND |
| 552 |
and |
| 553 |
.B MS_REC |
| 554 |
flags) is performed on a directory subtree, |
| 555 |
any unbindable mounts within the subtree are automatically pruned |
| 556 |
(i.e., not replicated) |
| 557 |
when replicating that subtree to produce the target subtree. |
| 558 |
.P |
| 559 |
By default, changing the propagation type affects only the |
| 560 |
.I target |
| 561 |
mount. |
| 562 |
If the |
| 563 |
.B MS_REC |
| 564 |
flag is also specified in |
| 565 |
.IR mountflags , |
| 566 |
then the propagation type of all mounts under |
| 567 |
.I target |
| 568 |
is also changed. |
| 569 |
.P |
| 570 |
For further details regarding mount propagation types |
| 571 |
(including the default propagation type assigned to new mounts), see |
| 572 |
.BR mount_namespaces (7). |
| 573 |
.\" |
| 574 |
.SS Moving a mount |
| 575 |
If |
| 576 |
.I mountflags |
| 577 |
contains the flag |
| 578 |
.B MS_MOVE |
| 579 |
(available since Linux 2.4.18), |
| 580 |
then move a subtree: |
| 581 |
.I source |
| 582 |
specifies an existing mount and |
| 583 |
.I target |
| 584 |
specifies the new location to which that mount is to be relocated. |
| 585 |
The move is atomic: at no point is the subtree unmounted. |
| 586 |
.P |
| 587 |
The remaining bits in the |
| 588 |
.I mountflags |
| 589 |
argument are ignored, as are the |
| 590 |
.I filesystemtype |
| 591 |
and |
| 592 |
.I data |
| 593 |
arguments. |
| 594 |
.\" |
| 595 |
.SS Creating a new mount |
| 596 |
If none of |
| 597 |
.BR MS_REMOUNT , |
| 598 |
.BR MS_BIND , |
| 599 |
.BR MS_MOVE , |
| 600 |
.BR MS_SHARED , |
| 601 |
.BR MS_PRIVATE , |
| 602 |
.BR MS_SLAVE , |
| 603 |
or |
| 604 |
.B MS_UNBINDABLE |
| 605 |
is specified in |
| 606 |
.IR mountflags , |
| 607 |
then |
| 608 |
.BR mount () |
| 609 |
performs its default action: creating a new mount. |
| 610 |
.I source |
| 611 |
specifies the source for the new mount, and |
| 612 |
.I target |
| 613 |
specifies the directory at which to create the mount point. |
| 614 |
.P |
| 615 |
The |
| 616 |
.I filesystemtype |
| 617 |
and |
| 618 |
.I data |
| 619 |
arguments are employed, and further bits may be specified in |
| 620 |
.I mountflags |
| 621 |
to modify the behavior of the call. |
| 622 |
.\" |
| 623 |
.SH RETURN VALUE |
| 624 |
On success, zero is returned. |
| 625 |
On error, \-1 is returned, and |
| 626 |
.I errno |
| 627 |
is set to indicate the error. |
| 628 |
.SH ERRORS |
| 629 |
The error values given below result from filesystem type independent |
| 630 |
errors. |
| 631 |
Each filesystem type may have its own special errors and its |
| 632 |
own special behavior. |
| 633 |
See the Linux kernel source code for details. |
| 634 |
.TP |
| 635 |
.B EACCES |
| 636 |
A component of a path was not searchable. |
| 637 |
(See also |
| 638 |
.BR path_resolution (7).) |
| 639 |
.TP |
| 640 |
.B EACCES |
| 641 |
Mounting a read-only filesystem was attempted without giving the |
| 642 |
.B MS_RDONLY |
| 643 |
flag. |
| 644 |
.IP |
| 645 |
The filesystem may be read-only for various reasons, including: |
| 646 |
it resides on a read-only optical disk; |
| 647 |
it is resides on a device with a physical switch that has been set to |
| 648 |
mark the device read-only; |
| 649 |
the filesystem implementation was compiled with read-only support; |
| 650 |
or errors were detected when initially mounting the filesystem, |
| 651 |
so that it was marked read-only |
| 652 |
and can't be remounted as read-write (until the errors are fixed). |
| 653 |
.IP |
| 654 |
Some filesystems instead return the error |
| 655 |
.B EROFS |
| 656 |
on an attempt to mount a read-only filesystem. |
| 657 |
.TP |
| 658 |
.B EACCES |
| 659 |
The block device |
| 660 |
.I source |
| 661 |
is located on a filesystem mounted with the |
| 662 |
.B MS_NODEV |
| 663 |
option. |
| 664 |
.\" mtk: Probably: write permission is required for MS_BIND, with |
| 665 |
.\" the error EPERM if not present; CAP_DAC_OVERRIDE is required. |
| 666 |
.TP |
| 667 |
.B EBUSY |
| 668 |
An attempt was made to stack a new mount directly on |
| 669 |
top of an existing mount point that was created in this |
| 670 |
mount namespace with the same |
| 671 |
.I source |
| 672 |
and |
| 673 |
.IR target . |
| 674 |
.TP |
| 675 |
.B EBUSY |
| 676 |
.I source |
| 677 |
cannot be remounted read-only, |
| 678 |
because it still holds files open for writing. |
| 679 |
.TP |
| 680 |
.B EFAULT |
| 681 |
One of the pointer arguments points outside the user address space. |
| 682 |
.TP |
| 683 |
.B EINVAL |
| 684 |
.I source |
| 685 |
had an invalid superblock. |
| 686 |
.TP |
| 687 |
.B EINVAL |
| 688 |
A remount operation |
| 689 |
.RB ( MS_REMOUNT ) |
| 690 |
was attempted, but |
| 691 |
.I source |
| 692 |
was not already mounted on |
| 693 |
.IR target . |
| 694 |
.TP |
| 695 |
.B EINVAL |
| 696 |
A move operation |
| 697 |
.RB ( MS_MOVE ) |
| 698 |
was attempted, but the mount tree under |
| 699 |
.I source |
| 700 |
includes unbindable mounts and |
| 701 |
.I target |
| 702 |
is a mount that has propagation type |
| 703 |
.BR MS_SHARED . |
| 704 |
.TP |
| 705 |
.B EINVAL |
| 706 |
A move operation |
| 707 |
.RB ( MS_MOVE ) |
| 708 |
was attempted, but the parent mount of |
| 709 |
.I source |
| 710 |
mount has propagation type |
| 711 |
.BR MS_SHARED . |
| 712 |
.TP |
| 713 |
.B EINVAL |
| 714 |
A move operation |
| 715 |
.RB ( MS_MOVE ) |
| 716 |
was attempted, but |
| 717 |
.I source |
| 718 |
was not a mount, or was \[aq]/\[aq]. |
| 719 |
.TP |
| 720 |
.B EINVAL |
| 721 |
A bind operation |
| 722 |
.RB ( MS_BIND ) |
| 723 |
was requested where |
| 724 |
.I source |
| 725 |
referred a mount namespace magic link (i.e., a |
| 726 |
.IR /proc/ pid /ns/mnt |
| 727 |
magic link or a bind mount to such a link) |
| 728 |
and the propagation type of the parent mount of |
| 729 |
.I target |
| 730 |
was |
| 731 |
.BR MS_SHARED , |
| 732 |
.\" See commit 8823c079ba7136dc1948d6f6dcb5f8022bde438e |
| 733 |
but propagation of the requested bind mount could lead to a circular |
| 734 |
dependency that might prevent the mount namespace from ever being freed. |
| 735 |
.TP |
| 736 |
.B EINVAL |
| 737 |
.I mountflags |
| 738 |
includes more than one of |
| 739 |
.BR MS_SHARED , |
| 740 |
.BR MS_PRIVATE , |
| 741 |
.BR MS_SLAVE , |
| 742 |
or |
| 743 |
.BR MS_UNBINDABLE . |
| 744 |
.TP |
| 745 |
.B EINVAL |
| 746 |
.I mountflags |
| 747 |
includes |
| 748 |
.BR MS_SHARED , |
| 749 |
.BR MS_PRIVATE , |
| 750 |
.BR MS_SLAVE , |
| 751 |
or |
| 752 |
.B MS_UNBINDABLE |
| 753 |
and also includes a flag other than |
| 754 |
.B MS_REC |
| 755 |
or |
| 756 |
.BR MS_SILENT . |
| 757 |
.TP |
| 758 |
.B EINVAL |
| 759 |
An attempt was made to bind mount an unbindable mount. |
| 760 |
.TP |
| 761 |
.B EINVAL |
| 762 |
In an unprivileged mount namespace |
| 763 |
(i.e., a mount namespace owned by a user namespace |
| 764 |
that was created by an unprivileged user), |
| 765 |
a bind mount operation |
| 766 |
.RB ( MS_BIND ) |
| 767 |
was attempted without specifying |
| 768 |
.RB ( MS_REC ), |
| 769 |
which would have revealed the filesystem tree underneath one of |
| 770 |
the submounts of the directory being bound. |
| 771 |
.TP |
| 772 |
.B ELOOP |
| 773 |
Too many links encountered during pathname resolution. |
| 774 |
.TP |
| 775 |
.B ELOOP |
| 776 |
A move operation was attempted, and |
| 777 |
.I target |
| 778 |
is a descendant of |
| 779 |
.IR source . |
| 780 |
.TP |
| 781 |
.B EMFILE |
| 782 |
(In case no block device is required:) |
| 783 |
Table of dummy devices is full. |
| 784 |
.TP |
| 785 |
.B ENAMETOOLONG |
| 786 |
A pathname was longer than |
| 787 |
.BR MAXPATHLEN . |
| 788 |
.TP |
| 789 |
.B ENODEV |
| 790 |
.I filesystemtype |
| 791 |
not configured in the kernel. |
| 792 |
.TP |
| 793 |
.B ENOENT |
| 794 |
A pathname was empty or had a nonexistent component. |
| 795 |
.TP |
| 796 |
.B ENOMEM |
| 797 |
The kernel could not allocate a free page to copy filenames or data into. |
| 798 |
.TP |
| 799 |
.B ENOSPC |
| 800 |
Number of mounts will exceed the limit after this syscall |
| 801 |
in some mount namespaces. |
| 802 |
This limit is controlled by |
| 803 |
.IR /proc/sys/fs/mount-max , |
| 804 |
see |
| 805 |
.BR proc_sys_fs (5) |
| 806 |
for details of this limit. |
| 807 |
that this limit applies both |
| 808 |
to the mount namespace where the mount operation occurs |
| 809 |
and to any mount namespaces to which the mount will propagate. |
| 810 |
See |
| 811 |
.BR mount_namespaces (7) |
| 812 |
for details of mount namespaces and mount propagation. |
| 813 |
.TP |
| 814 |
.B ENOTBLK |
| 815 |
.I source |
| 816 |
is not a block device (and a device was required). |
| 817 |
.TP |
| 818 |
.B ENOTDIR |
| 819 |
.IR target , |
| 820 |
or a prefix of |
| 821 |
.IR source , |
| 822 |
is not a directory. |
| 823 |
.TP |
| 824 |
.B ENXIO |
| 825 |
The major number of the block device |
| 826 |
.I source |
| 827 |
is out of range. |
| 828 |
.TP |
| 829 |
.B EPERM |
| 830 |
The caller does not have the required privileges. |
| 831 |
.TP |
| 832 |
.B EPERM |
| 833 |
An attempt was made to modify |
| 834 |
.RB ( MS_REMOUNT ) |
| 835 |
the |
| 836 |
.BR MS_RDONLY , |
| 837 |
.BR MS_NOSUID , |
| 838 |
or |
| 839 |
.B MS_NOEXEC |
| 840 |
flag, or one of the "atime" flags |
| 841 |
.RB ( MS_NOATIME , |
| 842 |
.BR MS_NODIRATIME , |
| 843 |
.BR MS_RELATIME ) |
| 844 |
of an existing mount, but the mount is locked; |
| 845 |
see |
| 846 |
.BR mount_namespaces (7). |
| 847 |
.TP |
| 848 |
.B EROFS |
| 849 |
Mounting a read-only filesystem was attempted without giving the |
| 850 |
.B MS_RDONLY |
| 851 |
flag. |
| 852 |
See |
| 853 |
.BR EACCES , |
| 854 |
above. |
| 855 |
.\" |
| 856 |
.SH STANDARDS |
| 857 |
Linux. |
| 858 |
.SH HISTORY |
| 859 |
The definitions of |
| 860 |
.BR MS_DIRSYNC , |
| 861 |
.BR MS_MOVE , |
| 862 |
.BR MS_PRIVATE , |
| 863 |
.BR MS_REC , |
| 864 |
.BR MS_RELATIME , |
| 865 |
.BR MS_SHARED , |
| 866 |
.BR MS_SLAVE , |
| 867 |
.BR MS_STRICTATIME , |
| 868 |
and |
| 869 |
.B MS_UNBINDABLE |
| 870 |
were added to glibc headers in glibc 2.12. |
| 871 |
.P |
| 872 |
Since Linux 2.4 a single filesystem can be mounted at |
| 873 |
multiple mount points, and multiple mounts can be stacked |
| 874 |
on the same mount point. |
| 875 |
.\" Multiple mounts on same mount point: since Linux 2.3.99pre7. |
| 876 |
.P |
| 877 |
The |
| 878 |
.I mountflags |
| 879 |
argument may have the magic number 0xC0ED |
| 880 |
.RB ( MS_MGC_VAL ) |
| 881 |
in the top 16 bits. |
| 882 |
(All of the other flags discussed in DESCRIPTION |
| 883 |
occupy the low order 16 bits of |
| 884 |
.IR mountflags .) |
| 885 |
Specifying |
| 886 |
.B MS_MGC_VAL |
| 887 |
was required before Linux 2.4, |
| 888 |
but since Linux 2.4 is no longer required and is ignored if specified. |
| 889 |
.P |
| 890 |
The original |
| 891 |
.B MS_SYNC |
| 892 |
flag was renamed |
| 893 |
.B MS_SYNCHRONOUS |
| 894 |
in 1.1.69 |
| 895 |
when a different |
| 896 |
.B MS_SYNC |
| 897 |
was added to |
| 898 |
.IR <mman.h> . |
| 899 |
.P |
| 900 |
Before Linux 2.4 an attempt to execute a set-user-ID or set-group-ID program |
| 901 |
on a filesystem mounted with |
| 902 |
.B MS_NOSUID |
| 903 |
would fail with |
| 904 |
.BR EPERM . |
| 905 |
Since Linux 2.4 the set-user-ID and set-group-ID bits are |
| 906 |
just silently ignored in this case. |
| 907 |
.\" The change is in patch-2.4.0-prerelease. |
| 908 |
.\" |
| 909 |
.SH NOTES |
| 910 |
.SS Mount namespaces |
| 911 |
Starting with Linux 2.4.19, Linux provides mount namespaces. |
| 912 |
A mount namespace is the set of filesystem mounts that |
| 913 |
are visible to a process. |
| 914 |
Mount namespaces can be (and usually are) |
| 915 |
shared between multiple processes, |
| 916 |
and changes to the namespace (i.e., mounts and unmounts) by one process |
| 917 |
are visible to all other processes sharing the same namespace. |
| 918 |
(The pre-2.4.19 Linux situation can be considered as one in which |
| 919 |
a single namespace was shared by every process on the system.) |
| 920 |
.P |
| 921 |
A child process created by |
| 922 |
.BR fork (2) |
| 923 |
shares its parent's mount namespace; |
| 924 |
the mount namespace is preserved across an |
| 925 |
.BR execve (2). |
| 926 |
.P |
| 927 |
A process can obtain a private mount namespace if: |
| 928 |
it was created using the |
| 929 |
.BR clone (2) |
| 930 |
.B CLONE_NEWNS |
| 931 |
flag, |
| 932 |
in which case its new namespace is initialized to be a |
| 933 |
.I copy |
| 934 |
of the namespace of the process that called |
| 935 |
.BR clone (2); |
| 936 |
or it calls |
| 937 |
.BR unshare (2) |
| 938 |
with the |
| 939 |
.B CLONE_NEWNS |
| 940 |
flag, |
| 941 |
which causes the caller's mount namespace to obtain a private copy |
| 942 |
of the namespace that it was previously sharing with other processes, |
| 943 |
so that future mounts and unmounts by the caller are invisible |
| 944 |
to other processes (except child processes that the caller |
| 945 |
subsequently creates) and vice versa. |
| 946 |
.P |
| 947 |
For further details on mount namespaces, see |
| 948 |
.BR mount_namespaces (7). |
| 949 |
.\" |
| 950 |
.SS Parental relationship between mounts |
| 951 |
Each mount has a parent mount. |
| 952 |
The overall parental relationship of all mounts defines |
| 953 |
the single directory hierarchy seen by the processes within a mount namespace. |
| 954 |
.P |
| 955 |
The parent of a new mount is defined when the mount is created. |
| 956 |
In the usual case, |
| 957 |
the parent of a new mount is the mount of the filesystem |
| 958 |
containing the directory or file at which the new mount is attached. |
| 959 |
In the case where a new mount is stacked on top of an existing mount, |
| 960 |
the parent of the new mount is the previous mount that was stacked |
| 961 |
at that location. |
| 962 |
.P |
| 963 |
The parental relationship between mounts can be discovered via the |
| 964 |
.IR /proc/ pid /mountinfo |
| 965 |
file (see below). |
| 966 |
.\" |
| 967 |
.SS \f[I]/proc/\f[]pid\f[I]/mounts\f[] and \f[I]/proc/\f[]pid\f[I]/mountinfo\f[] |
| 968 |
The Linux-specific |
| 969 |
.IR /proc/ pid /mounts |
| 970 |
file exposes the list of mounts in the mount |
| 971 |
namespace of the process with the specified ID. |
| 972 |
The |
| 973 |
.IR /proc/ pid /mountinfo |
| 974 |
file exposes even more information about mounts, |
| 975 |
including the propagation type and mount ID information that makes it |
| 976 |
possible to discover the parental relationship between mounts. |
| 977 |
See |
| 978 |
.BR proc (5) |
| 979 |
and |
| 980 |
.BR mount_namespaces (7) |
| 981 |
for details of this file. |
| 982 |
.SH SEE ALSO |
| 983 |
.BR mountpoint (1), |
| 984 |
.BR chroot (2), |
| 985 |
.BR FS_IOC_SETFLAGS (2const), |
| 986 |
.BR mount_setattr (2), |
| 987 |
.BR pivot_root (2), |
| 988 |
.BR umount (2), |
| 989 |
.BR mount_namespaces (7), |
| 990 |
.BR path_resolution (7), |
| 991 |
.BR findmnt (8), |
| 992 |
.BR lsblk (8), |
| 993 |
.BR mount (8), |
| 994 |
.BR umount (8) |