1QEMU virtio-fs shared file system daemon
2========================================
3
4Synopsis
5--------
6
7**virtiofsd** [*OPTIONS*]
8
9Description
10-----------
11
12Share a host directory tree with a guest through a virtio-fs device.  This
13program is a vhost-user backend that implements the virtio-fs device.  Each
14virtio-fs device instance requires its own virtiofsd process.
15
16This program is designed to work with QEMU's ``--device vhost-user-fs-pci``
17but should work with any virtual machine monitor (VMM) that supports
18vhost-user.  See the Examples section below.
19
20This program must be run as the root user.  Upon startup the program will
21switch into a new file system namespace with the shared directory tree as its
22root.  This prevents "file system escapes" due to symlinks and other file
23system objects that might lead to files outside the shared directory.  The
24program also sandboxes itself using seccomp(2) to prevent ptrace(2) and other
25vectors that could allow an attacker to compromise the system after gaining
26control of the virtiofsd process.
27
28Options
29-------
30
31.. program:: virtiofsd
32
33.. option:: -h, --help
34
35  Print help.
36
37.. option:: -V, --version
38
39  Print version.
40
41.. option:: -d
42
43  Enable debug output.
44
45.. option:: --syslog
46
47  Print log messages to syslog instead of stderr.
48
49.. option:: -o OPTION
50
51  * debug -
52    Enable debug output.
53
54  * flock|no_flock -
55    Enable/disable flock.  The default is ``no_flock``.
56
57  * log_level=LEVEL -
58    Print only log messages matching LEVEL or more severe.  LEVEL is one of
59    ``err``, ``warn``, ``info``, or ``debug``.  The default is ``info``.
60
61  * norace -
62    Disable racy fallback.  The default is false.
63
64  * posix_lock|no_posix_lock -
65    Enable/disable remote POSIX locks.  The default is ``posix_lock``.
66
67  * readdirplus|no_readdirplus -
68    Enable/disable readdirplus.  The default is ``readdirplus``.
69
70  * source=PATH -
71    Share host directory tree located at PATH.  This option is required.
72
73  * timeout=TIMEOUT -
74    I/O timeout in seconds.  The default depends on cache= option.
75
76  * writeback|no_writeback -
77    Enable/disable writeback cache. The cache alows the FUSE client to buffer
78    and merge write requests.  The default is ``no_writeback``.
79
80  * xattr|no_xattr -
81    Enable/disable extended attributes (xattr) on files and directories.  The
82    default is ``no_xattr``.
83
84.. option:: --socket-path=PATH
85
86  Listen on vhost-user UNIX domain socket at PATH.
87
88.. option:: --fd=FDNUM
89
90  Accept connections from vhost-user UNIX domain socket file descriptor FDNUM.
91  The file descriptor must already be listening for connections.
92
93.. option:: --thread-pool-size=NUM
94
95  Restrict the number of worker threads per request queue to NUM.  The default
96  is 64.
97
98.. option:: --cache=none|auto|always
99
100  Select the desired trade-off between coherency and performance.  ``none``
101  forbids the FUSE client from caching to achieve best coherency at the cost of
102  performance.  ``auto`` acts similar to NFS with a 1 second metadata cache
103  timeout.  ``always`` sets a long cache lifetime at the expense of coherency.
104
105Examples
106--------
107
108Export ``/var/lib/fs/vm001/`` on vhost-user UNIX domain socket
109``/var/run/vm001-vhost-fs.sock``:
110
111::
112
113  host# virtiofsd --socket-path=/var/run/vm001-vhost-fs.sock -o source=/var/lib/fs/vm001
114  host# qemu-system-x86_64 \
115      -chardev socket,id=char0,path=/var/run/vm001-vhost-fs.sock \
116      -device vhost-user-fs-pci,chardev=char0,tag=myfs \
117      -object memory-backend-memfd,id=mem,size=4G,share=on \
118      -numa node,memdev=mem \
119      ...
120  guest# mount -t virtiofs myfs /mnt
121