1tmpfs(5) File Formats Manual tmpfs(5)
2
6 tmpfs - a virtual memory filesystem
7
9 The tmpfs facility allows the creation of filesystems whose contents
10 reside in virtual memory. Since the files on such filesystems typi‐
11 cally reside in RAM, file access is extremely fast.
12 The filesystem is automatically created when mounting a filesystem with
14 the type tmpfs via a command such as the following:
15 $ sudo mount -t tmpfs -o size=10M tmpfs /mnt/mytmpfs
17 A tmpfs filesystem has the following properties:
19 • The filesystem can employ swap space when physical memory pressure
21 demands it.
22 • The filesystem consumes only as much physical memory and swap space
24 as is required to store the current contents of the filesystem.
25 • During a remount operation (mount -o remount), the filesystem size
27 can be changed (without losing the existing contents of the filesys‐
28 tem).
29 If a tmpfs filesystem is unmounted, its contents are discarded (lost).
31 Mount options
33 The tmpfs filesystem supports the following mount options:
34 size=bytes
36 Specify an upper limit on the size of the filesystem. The size
37 is given in bytes, and rounded up to entire pages.
38 The size may have a k, m, or g suffix for Ki, Mi, Gi (binary
40 kilo (kibi), binary mega (mebi), and binary giga (gibi)).
41 The size may also have a % suffix to limit this instance to a
43 percentage of physical RAM.
44 The default, when neither size nor nr_blocks is specified, is
46 size=50%.
47 nr_blocks=blocks
49 The same as size, but in blocks of PAGE_CACHE_SIZE.
50 Blocks may be specified with k, m, or g suffixes like size, but
52 not a % suffix.
53 nr_inodes=inodes
55 The maximum number of inodes for this instance. The default is
56 half of the number of your physical RAM pages, or (on a machine
57 with highmem) the number of lowmem RAM pages, whichever is
58 smaller.
59 Inodes may be specified with k, m, or g suffixes like size, but
61 not a % suffix.
62 mode=mode
64 Set initial permissions of the root directory.
65 gid=gid (since Linux 2.5.7)
67 Set the initial group ID of the root directory.
68 uid=uid (since Linux 2.5.7)
70 Set the initial user ID of the root directory.
71 huge=huge_option (since Linux 4.7.0)
73 Set the huge table memory allocation policy for all files in
74 this instance (if CONFIG_TRANSPARENT_HUGE_PAGECACHE is enabled).
75 The huge_option value is one of the following:
77 never Do not allocate huge pages. This is the default.
79 always Attempt to allocate huge pages every time a new page is
81 needed.
82 within_size
84 Only allocate huge page if it will be fully within
85 i_size. Also respect fadvise(2) and madvise(2) hints
86 advise Only allocate huge pages if requested with fadvise(2) or
88 madvise(2).
89 deny For use in emergencies, to force the huge option off from
91 all mounts.
92 force Force the huge option on for all mounts; useful for test‐
94 ing.
95 mpol=mpol_option (since Linux 2.6.15)
97 Set the NUMA memory allocation policy for all files in this in‐
98 stance (if CONFIG_NUMA is enabled).
99 The mpol_option value is one of the following:
101 default
103 Use the process allocation policy (see set_mempolicy(2)).
104 prefer:node
106 Preferably allocate memory from the given node.
107 bind:nodelist
109 Allocate memory only from nodes in nodelist.
110 interleave
112 Allocate from each node in turn.
113 interleave:nodelist
115 Allocate from each node of in turn.
116 local Preferably allocate memory from the local node.
118 In the above, nodelist is a comma-separated list of decimal num‐
120 bers and ranges that specify NUMA nodes. A range is a pair of
121 hyphen-separated decimal numbers, the smallest and largest node
122 numbers in the range. For example, mpol=bind:0-3,5,7,9-15.
123
125 The tmpfs facility was added in Linux 2.4, as a successor to the older
126 ramfs facility, which did not provide limit checking or allow for the
127 use of swap space.
128
130 In order for user-space tools and applications to create tmpfs filesys‐
131 tems, the kernel must be configured with the CONFIG_TMPFS option.
132 The tmpfs filesystem supports extended attributes (see xattr(7)), but
134 user extended attributes are not permitted.
135 An internal shared memory filesystem is used for System V shared memory
137 (shmget(2)) and shared anonymous mappings (mmap(2) with the MAP_SHARED
138 and MAP_ANONYMOUS flags). This filesystem is available regardless of
139 whether the kernel was configured with the CONFIG_TMPFS option.
140 A tmpfs filesystem mounted at /dev/shm is used for the implementation
142 of POSIX shared memory (shm_overview(7)) and POSIX semaphores
143 (sem_overview(7)).
144 The amount of memory consumed by all tmpfs filesystems is shown in the
146 Shmem field of /proc/meminfo and in the shared field displayed by
147 free(1).
148 The tmpfs facility was formerly called shmfs.
150
152 df(1), du(1), memfd_create(2), mmap(2), set_mempolicy(2), shm_open(3),
153 mount(8)
154 The kernel source files Documentation/filesystems/tmpfs.txt and Docu‐
156 mentation/admin-guide/mm/transhuge.rst.
157Linux man-pages 6.04 2023-02-05 tmpfs(5)