xref: /OK3568_Linux_fs/kernel/Documentation/filesystems/automount-support.rst (revision 4882a59341e53eb6f0b4789bf948001014eff981)
1*4882a593Smuzhiyun.. SPDX-License-Identifier: GPL-2.0
2*4882a593Smuzhiyun
3*4882a593Smuzhiyun=================
4*4882a593SmuzhiyunAutomount Support
5*4882a593Smuzhiyun=================
6*4882a593Smuzhiyun
7*4882a593Smuzhiyun
8*4882a593SmuzhiyunSupport is available for filesystems that wish to do automounting
9*4882a593Smuzhiyunsupport (such as kAFS which can be found in fs/afs/ and NFS in
10*4882a593Smuzhiyunfs/nfs/). This facility includes allowing in-kernel mounts to be
11*4882a593Smuzhiyunperformed and mountpoint degradation to be requested. The latter can
12*4882a593Smuzhiyunalso be requested by userspace.
13*4882a593Smuzhiyun
14*4882a593Smuzhiyun
15*4882a593SmuzhiyunIn-Kernel Automounting
16*4882a593Smuzhiyun======================
17*4882a593Smuzhiyun
18*4882a593SmuzhiyunSee section "Mount Traps" of  Documentation/filesystems/autofs.rst
19*4882a593Smuzhiyun
20*4882a593SmuzhiyunThen from userspace, you can just do something like::
21*4882a593Smuzhiyun
22*4882a593Smuzhiyun	[root@andromeda root]# mount -t afs \#root.afs. /afs
23*4882a593Smuzhiyun	[root@andromeda root]# ls /afs
24*4882a593Smuzhiyun	asd  cambridge  cambridge.redhat.com  grand.central.org
25*4882a593Smuzhiyun	[root@andromeda root]# ls /afs/cambridge
26*4882a593Smuzhiyun	afsdoc
27*4882a593Smuzhiyun	[root@andromeda root]# ls /afs/cambridge/afsdoc/
28*4882a593Smuzhiyun	ChangeLog  html  LICENSE  pdf  RELNOTES-1.2.2
29*4882a593Smuzhiyun
30*4882a593SmuzhiyunAnd then if you look in the mountpoint catalogue, you'll see something like::
31*4882a593Smuzhiyun
32*4882a593Smuzhiyun	[root@andromeda root]# cat /proc/mounts
33*4882a593Smuzhiyun	...
34*4882a593Smuzhiyun	#root.afs. /afs afs rw 0 0
35*4882a593Smuzhiyun	#root.cell. /afs/cambridge.redhat.com afs rw 0 0
36*4882a593Smuzhiyun	#afsdoc. /afs/cambridge.redhat.com/afsdoc afs rw 0 0
37*4882a593Smuzhiyun
38*4882a593Smuzhiyun
39*4882a593SmuzhiyunAutomatic Mountpoint Expiry
40*4882a593Smuzhiyun===========================
41*4882a593Smuzhiyun
42*4882a593SmuzhiyunAutomatic expiration of mountpoints is easy, provided you've mounted the
43*4882a593Smuzhiyunmountpoint to be expired in the automounting procedure outlined separately.
44*4882a593Smuzhiyun
45*4882a593SmuzhiyunTo do expiration, you need to follow these steps:
46*4882a593Smuzhiyun
47*4882a593Smuzhiyun (1) Create at least one list off which the vfsmounts to be expired can be
48*4882a593Smuzhiyun     hung.
49*4882a593Smuzhiyun
50*4882a593Smuzhiyun (2) When a new mountpoint is created in the ->d_automount method, add
51*4882a593Smuzhiyun     the mnt to the list using mnt_set_expiry()::
52*4882a593Smuzhiyun
53*4882a593Smuzhiyun             mnt_set_expiry(newmnt, &afs_vfsmounts);
54*4882a593Smuzhiyun
55*4882a593Smuzhiyun (3) When you want mountpoints to be expired, call mark_mounts_for_expiry()
56*4882a593Smuzhiyun     with a pointer to this list. This will process the list, marking every
57*4882a593Smuzhiyun     vfsmount thereon for potential expiry on the next call.
58*4882a593Smuzhiyun
59*4882a593Smuzhiyun     If a vfsmount was already flagged for expiry, and if its usage count is 1
60*4882a593Smuzhiyun     (it's only referenced by its parent vfsmount), then it will be deleted
61*4882a593Smuzhiyun     from the namespace and thrown away (effectively unmounted).
62*4882a593Smuzhiyun
63*4882a593Smuzhiyun     It may prove simplest to simply call this at regular intervals, using
64*4882a593Smuzhiyun     some sort of timed event to drive it.
65*4882a593Smuzhiyun
66*4882a593SmuzhiyunThe expiration flag is cleared by calls to mntput. This means that expiration
67*4882a593Smuzhiyunwill only happen on the second expiration request after the last time the
68*4882a593Smuzhiyunmountpoint was accessed.
69*4882a593Smuzhiyun
70*4882a593SmuzhiyunIf a mountpoint is moved, it gets removed from the expiration list. If a bind
71*4882a593Smuzhiyunmount is made on an expirable mount, the new vfsmount will not be on the
72*4882a593Smuzhiyunexpiration list and will not expire.
73*4882a593Smuzhiyun
74*4882a593SmuzhiyunIf a namespace is copied, all mountpoints contained therein will be copied,
75*4882a593Smuzhiyunand the copies of those that are on an expiration list will be added to the
76*4882a593Smuzhiyunsame expiration list.
77*4882a593Smuzhiyun
78*4882a593Smuzhiyun
79*4882a593SmuzhiyunUserspace Driven Expiry
80*4882a593Smuzhiyun=======================
81*4882a593Smuzhiyun
82*4882a593SmuzhiyunAs an alternative, it is possible for userspace to request expiry of any
83*4882a593Smuzhiyunmountpoint (though some will be rejected - the current process's idea of the
84*4882a593Smuzhiyunrootfs for example). It does this by passing the MNT_EXPIRE flag to
85*4882a593Smuzhiyunumount(). This flag is considered incompatible with MNT_FORCE and MNT_DETACH.
86*4882a593Smuzhiyun
87*4882a593SmuzhiyunIf the mountpoint in question is in referenced by something other than
88*4882a593Smuzhiyunumount() or its parent mountpoint, an EBUSY error will be returned and the
89*4882a593Smuzhiyunmountpoint will not be marked for expiration or unmounted.
90*4882a593Smuzhiyun
91*4882a593SmuzhiyunIf the mountpoint was not already marked for expiry at that time, an EAGAIN
92*4882a593Smuzhiyunerror will be given and it won't be unmounted.
93*4882a593Smuzhiyun
94*4882a593SmuzhiyunOtherwise if it was already marked and it wasn't referenced, unmounting will
95*4882a593Smuzhiyuntake place as usual.
96*4882a593Smuzhiyun
97*4882a593SmuzhiyunAgain, the expiration flag is cleared every time anything other than umount()
98*4882a593Smuzhiyunlooks at a mountpoint.
99