1MOUNT.CIFS(8) MOUNT.CIFS(8)
2
6 mount.cifs - mount using the Common Internet File System (CIFS)
7
9 mount.cifs {service} {mount-point} [-o options]
10 This tool is part of the cifs-utils suite.
12 mount.cifs mounts a CIFS or SMB3 filesystem from Linux. It is usually
14 invoked indirectly by the mount(8) command when using the "-t cifs"
15 option. This command only works in Linux, and the kernel must support
16 the cifs filesystem. The SMB3 protocol is the successor to the CIFS
17 (SMB) protocol and is supported by most Windows servers, Azure (cloud
18 storage), Macs and many other commercial servers and Network Attached
19 Storage appliances as well as by the popular Open Source server Samba.
20 The mount.cifs utility attaches the UNC name (exported network
22 resource) specified as service (using //server/share syntax, where
23 "server" is the server name or IP address and "share" is the name of
24 the share) to the local directory mount-point.
25 Options to mount.cifs are specified as a comma-separated list of
27 key=value pairs. It is possible to send options other than those listed
28 here, assuming that the cifs filesystem kernel module (cifs.ko) sup‐
29 ports them. Unrecognized cifs mount options passed to the cifs vfs ker‐
30 nel code will be logged to the kernel log.
31 mount.cifs causes the cifs vfs to launch a thread named cifsd. After
33 mounting it keeps running until the mounted resource is unmounted (usu‐
34 ally via the umount utility).
35 mount.cifs -V command displays the version of cifs mount helper.
37 modinfo cifs command displays the version of cifs module.
39
41 username=arg|user=arg
42 specifies the username to connect as. If this is not given, then
43 the environment variable USER is used.
44 Earlier versions of mount.cifs also allowed one to specify the
46 username in a user%password or workgroup/user or work‐
47 group/user%password to allow the password and workgroup to be
48 specified as part of the username. Support for those alternate
49 username formats is now deprecated and should no longer be used.
50 Users should use the discrete password= and domain= to specify
51 those values. While some versions of the cifs kernel module
52 accept user= as an abbreviation for this option, its use can
53 confuse the standard mount program into thinking that this is a
54 non-superuser mount. It is therefore recommended to use the full
55 username= option name.
56 password=arg|pass=arg
58 specifies the CIFS password. If this option is not given then
59 the environment variable PASSWD is used. If the password is not
60 specified directly or indirectly via an argument to mount,
61 mount.cifs will prompt for a password, unless the guest option
62 is specified.
63 Note that a password which contains the delimiter character
65 (i.e. a comma ',') will fail to be parsed correctly on the com‐
66 mand line. However, the same password defined in the PASSWD
67 environment variable or via a credentials file (see below) or
68 entered at the password prompt will be read correctly.
69 credentials=filename|cred=filename
71 specifies a file that contains a username and/or password and
72 optionally the name of the workgroup. The format of the file is:
73 username=value
75 password=value
76 domain=value
77 This is preferred over having passwords in plaintext in a shared
79 file, such as /etc/fstab . Be sure to protect any credentials
80 file properly.
81 uid=arg
83 sets the uid that will own all files or directories on the
84 mounted filesystem when the server does not provide ownership
85 information. It may be specified as either a username or a
86 numeric uid. When not specified, the default is uid 0. The
87 mount.cifs helper must be at version 1.10 or higher to support
88 specifying the uid in non-numeric form. See the section on FILE
89 AND DIRECTORY OWNERSHIP AND PERMISSIONS below for more informa‐
90 tion.
91 forceuid
93 instructs the client to ignore any uid provided by the server
94 for files and directories and to always assign the owner to be
95 the value of the uid= option. See the section on FILE AND DIREC‐
96 TORY OWNERSHIP AND PERMISSIONS below for more information.
97 cruid=arg
99 sets the uid of the owner of the credentials cache. This is pri‐
100 marily useful with sec=krb5. The default is the real uid of the
101 process performing the mount. Setting this parameter directs the
102 upcall to look for a credentials cache owned by that user.
103 gid=arg
105 sets the gid that will own all files or directories on the
106 mounted filesystem when the server does not provide ownership
107 information. It may be specified as either a groupname or a
108 numeric gid. When not specified, the default is gid 0. The
109 mount.cifs helper must be at version 1.10 or higher to support
110 specifying the gid in non-numeric form. See the section on FILE
111 AND DIRECTORY OWNERSHIP AND PERMISSIONS below for more informa‐
112 tion.
113 forcegid
115 instructs the client to ignore any gid provided by the server
116 for files and directories and to always assign the owner to be
117 the value of the gid= option. See the section on FILE AND DIREC‐
118 TORY OWNERSHIP AND PERMISSIONS below for more information.
119 idsfromsid
121 Extract uid/gid from special SID instead of mapping it. See the
122 section on FILE AND DIRECTORY OWNERSHIP AND PERMISSIONS below
123 for more information.
124 port=arg
126 sets the port number on which the client will attempt to contact
127 the CIFS server. If this value is specified, look for an exist‐
128 ing connection with this port, and use that if one exists. If
129 one doesn't exist, try to create a new connection on that port.
130 If that connection fails, return an error. If this value isn't
131 specified, look for an existing connection on port 445 or 139.
132 If no such connection exists, try to connect on port 445 first
133 and then port 139 if that fails. Return an error if both fail.
134 netbiosname=arg
136 When mounting to servers via port 139, specifies the RFC1001
137 source name to use to represent the client netbios machine dur‐
138 ing the netbios session initialization.
139 servern=arg
141 Similar to netbiosname except it specifies the netbios name of
142 the server instead of the client. Although rarely needed for
143 mounting to newer servers, this option is needed for mounting to
144 some older servers (such as OS/2 or Windows 98 and Windows ME)
145 since when connecting over port 139 they, unlike most newer
146 servers, do not support a default server name. A server name can
147 be up to 15 characters long and is usually uppercased.
148 file_mode=arg
150 If the server does not support the CIFS Unix extensions this
151 overrides the default file mode.
152 dir_mode=arg
154 If the server does not support the CIFS Unix extensions this
155 overrides the default mode for directories.
156 ip=arg|addr=arg
158 sets the destination IP address. This option is set automati‐
159 cally if the server name portion of the requested UNC name can
160 be resolved so rarely needs to be specified by the user.
161 domain=arg|dom=arg|workgroup=arg
163 Sets the domain (workgroup) of the user. If no domains are
164 given, the empty domain will be used. Use domainauto to automat‐
165 ically guess the domain of the server you are connecting to.
166 domainauto
168 When using NTLM authentication and not providing a domain via
169 domain, guess the domain from the server NTLM challenge. This
170 behavior used to be the default on kernels older than 2.6.36.
171 guest don't prompt for a password.
173 iocharset
175 Charset used to convert local path names to and from Unicode.
176 Unicode is used by default for network path names if the server
177 supports it. If iocharset is not specified then the nls_default
178 specified during the local client kernel build will be used. If
179 server does not support Unicode, this parameter is unused.
180 ro mount read-only.
182 rw mount read-write.
184 setuids
186 If the CIFS Unix extensions are negotiated with the server the
187 client will attempt to set the effective uid and gid of the
188 local process on newly created files, directories, and devices
189 (create, mkdir, mknod). If the CIFS Unix Extensions are not
190 negotiated, for newly created files and directories instead of
191 using the default uid and gid specified on the the mount, cache
192 the new file's uid and gid locally which means that the uid for
193 the file can change when the inode is reloaded (or the user
194 remounts the share).
195 nosetuids
197 The client will not attempt to set the uid and gid on on newly
198 created files, directories, and devices (create, mkdir, mknod)
199 which will result in the server setting the uid and gid to the
200 default (usually the server uid of the user who mounted the
201 share). Letting the server (rather than the client) set the uid
202 and gid is the default. If the CIFS Unix Extensions are not
203 negotiated then the uid and gid for new files will appear to be
204 the uid (gid) of the mounter or the uid (gid) parameter speci‐
205 fied on the mount.
206 perm Client does permission checks (vfs_permission check of uid and
208 gid of the file against the mode and desired operation), Note
209 that this is in addition to the normal ACL check on the target
210 machine done by the server software. Client permission checking
211 is enabled by default.
212 noperm Client does not do permission checks. This can expose files on
214 this mount to access by other users on the local client system.
215 It is typically only needed when the server supports the CIFS
216 Unix Extensions but the UIDs/GIDs on the client and server sys‐
217 tem do not match closely enough to allow access by the user
218 doing the mount. Note that this does not affect the normal ACL
219 check on the target machine done by the server software (of the
220 server ACL against the user name provided at mount time).
221 dynperm
223 Instructs the server to maintain ownership and permissions in
224 memory that can't be stored on the server. This information can
225 disappear at any time (whenever the inode is flushed from the
226 cache), so while this may help make some applications work, it's
227 behavior is somewhat unreliable. See the section below on FILE
228 AND DIRECTORY OWNERSHIP AND PERMISSIONS for more information.
229 cache=arg
231 Cache mode. See the section below on CACHE COHERENCY for
232 details. Allowed values are:
233 · none - do not cache file data at all
235 · strict - follow the CIFS/SMB2 protocol strictly
237 · loose - allow loose caching semantics
239 The default in kernels prior to 3.7 was loose. As of kernel 3.7
241 the default is strict.
242 nostrictsync
244 Do not ask the server to flush on fsync(). Some servers perform
245 non-buffered writes by default in which case flushing is redun‐
246 dant. In workloads where a client is performing a lot of small
247 write + fsync combinations and where network latency is much
248 higher than the server latency, this brings a 2x performance
249 improvement. This option is also a good candidate in scenarios
250 where we want performance over consistency.
251 handlecache
253 (default) In SMB2 and above, the client often has to open the
254 root of the share (empty path) in various places during mount,
255 path revalidation and the statfs(2) system call. This option
256 cuts redundant round trip traffic (opens and closes) by simply
257 keeping the directory handle for the root around once opened.
258 nohandlecache
260 Disable caching of the share root directory handle.
261 handletimeout=arg
263 The time (in milliseconds) for which the server should reserve
264 the handle after a failover waiting for the client to reconnect.
265 When mounting with resilienthandles or persistenthandles mount
266 option, or when their use is requested by the server (continuous
267 availability shares) then this parameter overrides the server
268 default handle timeout (which for most servers is 120 seconds).
269 rwpidforward
271 Forward pid of a process who opened a file to any read or write
272 operation on that file. This prevent applications like wine(1)
273 from failing on read and write if we use mandatory brlock style.
274 mapchars
276 Translate six of the seven reserved characters (not backslash,
277 but including the colon, question mark, pipe, asterik, greater
278 than and less than characters) to the remap range (above
279 0xF000), which also allows the CIFS client to recognize files
280 created with such characters by Windows's Services for Mac. This
281 can also be useful when mounting to most versions of Samba
282 (which also forbids creating and opening files whose names con‐
283 tain any of these seven characters). This has no effect if the
284 server does not support Unicode on the wire. Please note that
285 the files created with mapchars mount option may not be accessi‐
286 ble if the share is mounted without that option.
287 nomapchars
289 (default) Do not translate any of these seven characters.
290 mapposix
292 Translate reserved characters similarly to mapchars but use the
293 mapping from Microsoft "Services For Unix".
294 intr currently unimplemented.
296 nointr (default) currently unimplemented.
298 hard The program accessing a file on the cifs mounted file system
300 will hang when the server crashes.
301 soft (default) The program accessing a file on the cifs mounted file
303 system will not hang when the server crashes and will return
304 errors to the user application.
305 noacl Do not allow POSIX ACL operations even if server would support
307 them.
308 The CIFS client can get and set POSIX ACLs (getfacl, setfacl) to
310 Samba servers version 3.0.10 and later. Setting POSIX ACLs
311 requires enabling both CIFS_XATTR and then CIFS_POSIX support in
312 the CIFS configuration options when building the cifs module.
313 POSIX ACL support can be disabled on a per mount basis by speci‐
314 fying noacl on mount.
315 cifsacl
317 This option is used to map CIFS/NTFS ACLs to/from Linux permis‐
318 sion bits, map SIDs to/from UIDs and GIDs, and get and set Secu‐
319 rity Descriptors.
320 See section on CIFS/NTFS ACL, SID/UID/GID MAPPING, SECURITY
322 DESCRIPTORS for more information.
323 backupuid=arg
325 File access by this user shall be done with the backup intent
326 flag set. Either a name or an id must be provided as an argu‐
327 ment, there are no default values.
328 See section ACCESSING FILES WITH BACKUP INTENT for more details.
330 backupgid=arg
332 File access by users who are members of this group shall be done
333 with the backup intent flag set. Either a name or an id must be
334 provided as an argument, there are no default values.
335 See section ACCESSING FILES WITH BACKUP INTENT for more details.
337 nocase Request case insensitive path name matching (case sensitive is
339 the default if the server supports it).
340 ignorecase
342 Synonym for nocase.
343 sec=arg
345 Security mode. Allowed values are:
346 · none - attempt to connection as a null user (no name)
348 · krb5 - Use Kerberos version 5 authentication
350 · krb5i - Use Kerberos authentication and forcibly enable packet
352 signing
353 · ntlm - Use NTLM password hashing
355 · ntlmi - Use NTLM password hashing and force packet signing
357 · ntlmv2 - Use NTLMv2 password hashing
359 · ntlmv2i - Use NTLMv2 password hashing and force packet signing
361 · ntlmssp - Use NTLMv2 password hashing encapsulated in Raw
363 NTLMSSP message
364 · ntlmsspi - Use NTLMv2 password hashing encapsulated in Raw
366 NTLMSSP message, and force packet signing
367 The default in mainline kernel versions prior to v3.8 was
369 sec=ntlm. In v3.8, the default was changed to sec=ntlmssp.
370 If the server requires signing during protocol negotiation, then
372 it may be enabled automatically. Packet signing may also be
373 enabled automatically if it's enabled in /proc/fs/cifs/Securi‐
374 tyFlags.
375 seal Request encryption at the SMB layer. The encryption algorithm
377 used is AES-128-CCM. Requires SMB3 or above (see vers).
378 rdma Connect directly to the server using SMB Direct via a RDMA
380 adapter. Requires SMB3 or above (see vers).
381 resilienthandles
383 Enable resilient handles. If the server supports it, keep opened
384 files across reconnections. Requires SMB2.1 (see vers).
385 noresilienthandles
387 (default) Disable resilient handles.
388 persistenthandles
390 Enable persistent handles. If the server supports it, keep
391 opened files across reconnections. Persistent handles are also
392 valid across servers in a cluster and have stronger guarantees
393 than resilient handles. Requires SMB3 or above (see vers).
394 nopersistenthandles
396 (default) Disable persistent handles.
397 snapshot=time
399 Mount a specific snapshot of the remote share. time must be a
400 positive integer identifying the snapshot requested (in
401 100-nanosecond units that have elapsed since January 1, 1601, or
402 alternatively it can be specified in GMT format e.g.
403 @GMT-2019.03.27-20.52.19). Supported in the Linux kernel start‐
404 ing from v4.19.
405 nobrl Do not send byte range lock requests to the server. This is nec‐
407 essary for certain applications that break with cifs style
408 mandatory byte range locks (and most cifs servers do not yet
409 support requesting advisory byte range locks).
410 forcemandatorylock
412 Do not use POSIX locks even when available via unix extensions.
413 Always use cifs style mandatory locks.
414 locallease
416 Check cached leases locally instead of querying the server.
417 sfu When the CIFS or SMB3 Unix Extensions are not negotiated,
419 attempt to create device files and fifos in a format compatible
420 with Services for Unix (SFU). In addition retrieve bits 10-12 of
421 the mode via the SETFILEBITS extended attribute (as SFU does).
422 In the future the bottom 9 bits of the mode mode also will be
423 emulated using queries of the security descriptor (ACL). [NB:
424 requires version 1.39 or later of the CIFS VFS. To recognize
425 symlinks and be able to create symlinks in an SFU interoperable
426 form requires version 1.40 or later of the CIFS VFS kernel mod‐
427 ule.
428 mfsymlinks
430 Enable support for Minshall+French symlinks (see
431 http://wiki.samba.org/index.php/UNIX_Extensions#Minshall.2BFrench_symlinks).
432 This option is ignored when specified together with the sfu
433 option. Minshall+French symlinks are used even if the server
434 supports the CIFS Unix Extensions.
435 echo_interval=n
437 sets the interval at which echo requests are sent to the server
438 on an idling connection. This setting also affects the time
439 required for a connection to an unresponsive server to timeout.
440 Here n is the echo interval in seconds. The reconnection happens
441 at twice the value of the echo_interval set for an unresponsive
442 server. If this option is not given then the default value of
443 60 seconds is used. The minimum tunable value is 1 second and
444 maximum can go up to 600 seconds.
445 serverino
447 Use inode numbers (unique persistent file identifiers) returned
448 by the server instead of automatically generating temporary
449 inode numbers on the client. Although server inode numbers make
450 it easier to spot hardlinked files (as they will have the same
451 inode numbers) and inode numbers may be persistent (which is
452 useful for some software), the server does not guarantee that
453 the inode numbers are unique if multiple server side mounts are
454 exported under a single share (since inode numbers on the
455 servers might not be unique if multiple filesystems are mounted
456 under the same shared higher level directory). Note that not all
457 servers support returning server inode numbers, although those
458 that support the CIFS Unix Extensions, and Windows 2000 and
459 later servers typically do support this (although not necessar‐
460 ily on every local server filesystem). Parameter has no effect
461 if the server lacks support for returning inode numbers or
462 equivalent. This behavior is enabled by default.
463 noserverino
465 Client generates inode numbers itself rather than using the
466 actual ones from the server.
467 See section INODE NUMBERS for more information.
469 posix|unix|linux
471 (default) Enable Unix Extensions for this mount. Requires CIFS
472 (vers=1.0) or SMB3.1.1 (vers=3.1.1) and a server supporting
473 them.
474 noposix|nounix|nolinux
476 Disable the Unix Extensions for this mount. This can be useful
477 in order to turn off multiple settings at once. This includes
478 POSIX acls, POSIX locks, POSIX paths, symlink support and
479 retrieving uids/gids/mode from the server. This can also be use‐
480 ful to work around a bug in a server that supports Unix Exten‐
481 sions.
482 See section INODE NUMBERS for more information.
484 nouser_xattr
486 Do not allow getfattr/setfattr to get/set xattrs, even if server
487 would support it otherwise. The default is for xattr support to
488 be enabled.
489 nodfs Do not follow Distributed FileSystem referrals. IO on a file not
491 stored on the server will fail instead of connecting to the tar‐
492 get server transparently.
493 noautotune
495 Use fixed size for kernel recv/send socket buffers.
496 nosharesock
498 Do not try to reuse sockets if the system is already connected
499 to the server via an existing mount point. This will make the
500 client always make a new connection to the server no matter what
501 he is already connected to. This can be useful in simulating
502 multiple clients connecting to the same server, as each mount
503 point will use a different TCP socket.
504 noblocksend
506 Send data on the socket using non blocking operations (MSG_DONT‐
507 WAIT flag).
508 rsize=bytes
510 Maximum amount of data that the kernel will request in a read
511 request in bytes. Maximum size that servers will accept is typi‐
512 cally 8MB for SMB3 or later dialects. Default requested during
513 mount is 4MB. Prior to the 4.20 kernel the default requested was
514 1MB. Prior to the SMB2.1 dialect the maximum was usually 64K.
515 wsize=bytes
517 Maximum amount of data that the kernel will send in a write
518 request in bytes. Maximum size that servers will accept is typi‐
519 cally 8MB for SMB3 or later dialects. Default requested during
520 mount is 4MB. Prior to the 4.20 kernel the default requested was
521 1MB. Prior to the SMB2.1 dialect the maximum was usually 64K.
522 bsize=bytes
524 Override the default blocksize (1MB) reported on SMB3 files
525 (requires kernel version of 5.1 or later). Prior to kernel ver‐
526 sion 5.1, the blocksize was always reported as 16K instead of
527 1MB (and was not configurable) which can hurt the performance of
528 tools like cp and scp (especially for uncached I/O) which decide
529 on the read and write size to use for file copies based on the
530 inode blocksize. bsize may not be less than 16K or greater than
531 16M.
532 max_credits=n
534 Maximum credits the SMB2 client can have. Default is 32000. Must
535 be set to a number between 20 and 60000.
536 fsc Enable local disk caching using FS-Cache for CIFS. This option
538 could be useful to improve performance on a slow link, heavily
539 loaded server and/or network where reading from the disk is
540 faster than reading from the server (over the network). This
541 could also impact the scalability positively as the number of
542 calls to the server are reduced. But, be warned that local
543 caching is not suitable for all workloads, for e.g., read-once
544 type workloads. So, you need to consider carefully the situa‐
545 tion/workload before using this option. Currently, local disk
546 caching is enabled for CIFS files opened as read-only.
547 NOTE: This feature is available only in the recent kernels that
549 have been built with the kernel config option CON‐
550 FIG_CIFS_FSCACHE. You also need to have cachefilesd daemon
551 installed and running to make the cache operational.
552 multiuser
554 Map user accesses to individual credentials when accessing the
555 server. By default, CIFS mounts only use a single set of user
556 credentials (the mount credentials) when accessing a share. With
557 this option, the client instead creates a new session with the
558 server using the user's credentials whenever a new user accesses
559 the mount. Further accesses by that user will also use those
560 credentials. Because the kernel cannot prompt for passwords,
561 multiuser mounts are limited to mounts using sec= options that
562 don't require passwords.
563 With this change, it's feasible for the server to handle permis‐
565 sions enforcement, so this option also implies noperm . Further‐
566 more, when unix extensions aren't in use and the administrator
567 has not overridden ownership using the uid= or gid= options,
568 ownership of files is presented as the current user accessing
569 the share.
570 actimeo=arg
572 The time (in seconds) that the CIFS client caches attributes of
573 a file or directory before it requests attribute information
574 from a server. During this period the changes that occur on the
575 server remain undetected until the client checks the server
576 again.
577 By default, the attribute cache timeout is set to 1 second. This
579 means more frequent on-the-wire calls to the server to check
580 whether attributes have changed which could impact performance.
581 With this option users can make a tradeoff between performance
582 and cache metadata correctness, depending on workload needs.
583 Shorter timeouts mean better cache coherency, but frequent
584 increased number of calls to the server. Longer timeouts mean a
585 reduced number of calls to the server but looser cache
586 coherency. The actimeo value is a positive integer that can hold
587 values between 0 and a maximum value of 2^30 * HZ (frequency of
588 timer interrupt) setting.
589 noposixpaths
591 If unix extensions are enabled on a share, then the client will
592 typically allow filenames to include any character besides '/'
593 in a pathname component, and will use forward slashes as a path‐
594 name delimiter. This option prevents the client from attempting
595 to negotiate the use of posix-style pathnames to the server.
596 posixpaths
598 Inverse of noposixpaths .
599 prefixpath=arg
601 It's possible to mount a subdirectory of a share. The preferred
602 way to do this is to append the path to the UNC when mounting.
603 However, it's also possible to do the same by setting this
604 option and providing the path there.
605 vers=arg
607 SMB protocol version. Allowed values are:
608 · 1.0 - The classic CIFS/SMBv1 protocol.
610 · 2.0 - The SMBv2.002 protocol. This was initially introduced in
612 Windows Vista Service Pack 1, and Windows Server 2008. Note
613 that the initial release version of Windows Vista spoke a
614 slightly different dialect (2.000) that is not supported.
615 · 2.1 - The SMBv2.1 protocol that was introduced in Microsoft
617 Windows 7 and Windows Server 2008R2.
618 · 3.0 - The SMBv3.0 protocol that was introduced in Microsoft
620 Windows 8 and Windows Server 2012.
621 · 3.02 or 3.0.2 - The SMBv3.0.2 protocol that was introduced in
623 Microsoft Windows 8.1 and Windows Server 2012R2.
624 · 3.1.1 or 3.11 - The SMBv3.1.1 protocol that was introduced in
626 Microsoft Windows 10 and Windows Server 2016.
627 · 3 - The SMBv3.0 protocol version and above.
629 · default - Tries to negotiate the highest SMB2+ version sup‐
631 ported by both the client and server.
632 If no dialect is specified on mount vers=default is used. To
634 check Dialect refer to /proc/fs/cifs/DebugData
635 Note too that while this option governs the protocol version
637 used, not all features of each version are available.
638 The default since v4.13.5 is for the client and server to nego‐
640 tiate the highest possible version greater than or equal to 2.1.
641 In kernels prior to v4.13, the default was 1.0. For kernels
642 between v4.13 and v4.13.5 the default is 3.0.
643 --verbose
645 Print additional debugging information for the mount. Note that
646 this parameter must be specified before the -o . For example:
647 mount -t cifs //server/share /mnt --verbose -o user=username
649
651 It's generally preferred to use forward slashes (/) as a delimiter in
652 service names. They are considered to be the "universal delimiter"
653 since they are generally not allowed to be embedded within path compo‐
654 nents on Windows machines and the client can convert them to back‐
655 slashes (\) unconditionally. Conversely, backslash characters are
656 allowed by POSIX to be part of a path component, and can't be automati‐
657 cally converted in the same way.
658 mount.cifs will attempt to convert backslashes to forward slashes where
660 it's able to do so, but it cannot do so in any path component following
661 the sharename.
662
664 When Unix Extensions are enabled, we use the actual inode number pro‐
665 vided by the server in response to the POSIX calls as an inode number.
666 When Unix Extensions are disabled and serverino mount option is enabled
668 there is no way to get the server inode number. The client typically
669 maps the server-assigned UniqueID onto an inode number.
670 Note that the UniqueID is a different value from the server inode num‐
672 ber. The UniqueID value is unique over the scope of the entire server
673 and is often greater than 2 power 32. This value often makes programs
674 that are not compiled with LFS (Large File Support), to trigger a glibc
675 EOVERFLOW error as this won't fit in the target structure field. It is
676 strongly recommended to compile your programs with LFS support (i.e.
677 with -D_FILE_OFFSET_BITS=64) to prevent this problem. You can also use
678 noserverino mount option to generate inode numbers smaller than 2 power
679 32 on the client. But you may not be able to detect hardlinks properly.
680
682 With a network filesystem such as CIFS or NFS, the client must contend
683 with the fact that activity on other clients or the server could change
684 the contents or attributes of a file without the client being aware of
685 it. One way to deal with such a problem is to mandate that all file
686 accesses go to the server directly. This is performance prohibitive
687 however, so most protocols have some mechanism to allow the client to
688 cache data locally.
689 The CIFS protocol mandates (in effect) that the client should not cache
691 file data unless it holds an opportunistic lock (aka oplock) or a
692 lease. Both of these entities allow the client to guarantee certain
693 types of exclusive access to a file so that it can access its contents
694 without needing to continually interact with the server. The server
695 will call back the client when it needs to revoke either of them and
696 allow the client a certain amount of time to flush any cached data.
697 The cifs client uses the kernel's pagecache to cache file data. Any I/O
699 that's done through the pagecache is generally page-aligned. This can
700 be problematic when combined with byte-range locks as Windows' locking
701 is mandatory and can block reads and writes from occurring.
702 cache=none means that the client never utilizes the cache for normal
704 reads and writes. It always accesses the server directly to satisfy a
705 read or write request.
706 cache=strict means that the client will attempt to follow the CIFS/SMB2
708 protocol strictly. That is, the cache is only trusted when the client
709 holds an oplock. When the client does not hold an oplock, then the
710 client bypasses the cache and accesses the server directly to satisfy a
711 read or write request. By doing this, the client avoids problems with
712 byte range locks. Additionally, byte range locks are cached on the
713 client when it holds an oplock and are "pushed" to the server when that
714 oplock is recalled.
715 cache=loose allows the client to use looser protocol semantics which
717 can sometimes provide better performance at the expense of cache
718 coherency. File access always involves the pagecache. When an oplock or
719 lease is not held, then the client will attempt to flush the cache soon
720 after a write to a file. Note that that flush does not necessarily
721 occur before a write system call returns.
722 In the case of a read without holding an oplock, the client will
724 attempt to periodically check the attributes of the file in order to
725 ascertain whether it has changed and the cache might no longer be
726 valid. This mechanism is much like the one that NFSv2/3 use for cache
727 coherency, but it particularly problematic with CIFS. Windows is quite
728 "lazy" with respect to updating the LastWriteTime field that the client
729 uses to verify this. The effect is that cache=loose can cause data cor‐
730 ruption when multiple readers and writers are working on the same
731 files.
732 Because of this, when multiple clients are accessing the same set of
734 files, then cache=strict is recommended. That helps eliminate problems
735 with cache coherency by following the CIFS/SMB2 protocols more
736 strictly.
737 Note too that no matter what caching model is used, the client will
739 always use the pagecache to handle mmap'ed files. Writes to mmap'ed
740 files are only guaranteed to be flushed to the server when msync() is
741 called, or on close().
742 The default in kernels prior to 3.7 was loose. As of 3.7, the default
744 is strict.
745
747 This option is used to work with file objects which posses Security
748 Descriptors and CIFS/NTFS ACL instead of UID, GID, file permission
749 bits, and POSIX ACL as user authentication model. This is the most com‐
750 mon authentication model for CIFS servers and is the one used by Win‐
751 dows.
752 Support for this requires both CIFS_XATTR and CIFS_ACL support in the
754 CIFS configuration options when building the cifs module.
755 A CIFS/NTFS ACL is mapped to file permission bits using an algorithm
757 specified in the following Microsoft TechNet document:
758 http://technet.microsoft.com/en-us/library/bb463216.aspx
760 In order to map SIDs to/from UIDs and GIDs, the following is required:
762 · a kernel upcall to the cifs.idmap utility set up via
764 request-key.conf(5)
765 · winbind support configured via nsswitch.conf(5) and smb.conf(5)
767 Please refer to the respective manpages of cifs.idmap(8) and win‐
769 bindd(8) for more information.
770 Security descriptors for a file object can be retrieved and set
772 directly using extended attribute named system.cifs_acl. The security
773 descriptors presented via this interface are "raw" blobs of data and
774 need a userspace utility to either parse and format or to assemble it
775 such as getcifsacl(1) and setcifsacl(1) respectively.
776 Some of the things to consider while using this mount option:
778 · There may be an increased latency when handling metadata due to addi‐
780 tional requests to get and set security descriptors.
781 · The mapping between a CIFS/NTFS ACL and POSIX file permission bits is
783 imperfect and some ACL information may be lost in the translation.
784 · If either upcall to cifs.idmap is not setup correctly or winbind is
786 not configured and running, ID mapping will fail. In that case uid
787 and gid will default to either to those values of the share or to the
788 values of uid and/or gid mount options if specified.
789
791 For an user on the server, desired access to a file is determined by
792 the permissions and rights associated with that file. This is typically
793 accomplished using ownership and ACL. For a user who does not have
794 access rights to a file, it is still possible to access that file for a
795 specific or a targeted purpose by granting special rights. One of the
796 specific purposes is to access a file with the intent to either backup
797 or restore i.e. backup intent. The right to access a file with the
798 backup intent can typically be granted by making that user a part of
799 the built-in group Backup Operators. Thus, when this user attempts to
800 open a file with the backup intent, open request is sent by setting the
801 bit FILE_OPEN_FOR_BACKUP_INTENT as one of the CreateOptions.
802 As an example, on a Windows server, a user named testuser, cannot open
804 this file with such a security descriptor:
805 REVISION:0x1
807 CONTROL:0x9404
808 OWNER:Administrator
809 GROUP:Domain Users
810 ACL:Administrator:ALLOWED/0x0/FULL
811 But the user testuser, if it becomes part of the Backup Operators
813 group, can open the file with the backup intent.
814 Any user on the client side who can authenticate as such a user on the
816 server, can access the files with the backup intent. But it is desir‐
817 able and preferable for security reasons amongst many, to restrict this
818 special right.
819 The mount option backupuid is used to restrict this special right to a
821 user which is specified by either a name or an id. The mount option
822 backupgid is used to restrict this special right to the users in a
823 group which is specified by either a name or an id. Only users matching
824 either backupuid or backupgid shall attempt to access files with backup
825 intent. These two mount options can be used together.
826
828 The core CIFS protocol does not provide unix ownership information or
829 mode for files and directories. Because of this, files and directories
830 will generally appear to be owned by whatever values the uid= or gid=
831 options are set, and will have permissions set to the default file_mode
832 and dir_mode for the mount. Attempting to change these values via
833 chmod/chown will return success but have no effect.
834 When the client and server negotiate unix extensions, files and direc‐
836 tories will be assigned the uid, gid, and mode provided by the server.
837 Because CIFS mounts are generally single-user, and the same credentials
838 are used no matter what user accesses the mount, newly created files
839 and directories will generally be given ownership corresponding to
840 whatever credentials were used to mount the share.
841 If the uid's and gid's being used do not match on the client and
843 server, the forceuid and forcegid options may be helpful. Note however,
844 that there is no corresponding option to override the mode. Permissions
845 assigned to a file when forceuid or forcegid are in effect may not
846 reflect the the real permissions.
847 When unix extensions are not negotiated, it's also possible to emulate
849 them locally on the server using the dynperm mount option. When this
850 mount option is in effect, newly created files and directories will
851 receive what appear to be proper permissions. These permissions are not
852 stored on the server however and can disappear at any time in the
853 future (subject to the whims of the kernel flushing out the inode
854 cache). In general, this mount option is discouraged.
855 It's also possible to override permission checking on the client alto‐
857 gether via the noperm option. Server-side permission checks cannot be
858 overridden. The permission checks done by the server will always corre‐
859 spond to the credentials used to mount the share, and not necessarily
860 to the user who is accessing the share.
861
863 The variable USER may contain the username of the person to be used to
864 authenticate to the server. The variable can be used to set both user‐
865 name and password by using the format username%password.
866 The variable PASSWD may contain the password of the person using the
868 client.
869 The variable PASSWD_FILE may contain the pathname of a file to read the
871 password from. A single line of input is read and used as the password.
872
874 This command may be used only by root, unless installed setuid, in
875 which case the noexec and nosuid mount flags are enabled. When
876 installed as a setuid program, the program follows the conventions set
877 forth by the mount program for user mounts, with the added restriction
878 that users must be able to chdir() into the mountpoint prior to the
879 mount in order to be able to mount onto it.
880 Some samba client tools like smbclient(8) honour client-side configura‐
882 tion parameters present in smb.conf. Unlike those client tools,
883 mount.cifs ignores smb.conf completely.
884
886 The primary mechanism for making configuration changes and for reading
887 debug information for the cifs vfs is via the Linux /proc filesystem.
888 In the directory /proc/fs/cifs are various configuration files and
889 pseudo files which can display debug information and performance sta‐
890 tistics. There are additional startup options such as maximum buffer
891 size and number of buffers which only may be set when the kernel cifs
892 vfs (cifs.ko module) is loaded. These can be seen by running the mod‐
893 info utility against the file cifs.ko which will list the options that
894 may be passed to cifs during module installation (device driver load).
895 For more information see the kernel file fs/cifs/README. When configur‐
896 ing dynamic tracing (trace-cmd) note that the list of SMB3 events which
897 can be enabled can be seen at: /sys/kernel/debug/tracing/events/cifs/.
898
900 The use of SMB2.1 or later (including the latest dialect SMB3.1.1) is
901 recommended for improved security and SMB1 is no longer requested by
902 default at mount time. Old dialects such as CIFS (SMB1, ie vers=1.0)
903 have much weaker security. Use of CIFS (SMB1) can be disabled by mod‐
904 probe cifs disable_legacy_dialects=y.
905
907 Mounting using the CIFS URL specification is currently not supported.
908 The credentials file does not handle usernames or passwords with lead‐
910 ing space.
911 Note that the typical response to a bug report is a suggestion to try
913 the latest version first. So please try doing that first, and always
914 include which versions you use of relevant software when reporting bugs
915 (minimum: mount.cifs (try mount.cifs -V), kernel (see /proc/version)
916 and server type you are trying to contact.
917
919 This man page is correct for version 2.18 of the cifs vfs filesystem
920 (roughly Linux kernel 5.0).
921
923 cifs.upcall(8), getcifsacl(1), setcifsacl(1)
924 Documentation/filesystems/cifs.txt and fs/cifs/README in the Linux ker‐
926 nel source tree may contain additional options and information.
927
929 Steve French
930 The maintainer of the Linux cifs vfs is Steve French. The maintainer of
932 the cifs-utils suite of user space tools is Pavel Shilovsky. The Linux
933 CIFS Mailing list is the preferred place to ask questions regarding
934 these programs.
935 MOUNT.CIFS(8)