1 .. SPDX-License-Identifier: GPL-2.0 1 .. SPDX-License-Identifier: GPL-2.0
2 2
3 ===== 3 =====
4 Tmpfs 4 Tmpfs
5 ===== 5 =====
6 6
7 Tmpfs is a file system which keeps all of its 7 Tmpfs is a file system which keeps all of its files in virtual memory.
8 8
9 9
10 Everything in tmpfs is temporary in the sense 10 Everything in tmpfs is temporary in the sense that no files will be
11 created on your hard drive. If you unmount a t 11 created on your hard drive. If you unmount a tmpfs instance,
12 everything stored therein is lost. 12 everything stored therein is lost.
13 13
14 tmpfs puts everything into the kernel internal 14 tmpfs puts everything into the kernel internal caches and grows and
15 shrinks to accommodate the files it contains a 15 shrinks to accommodate the files it contains and is able to swap
16 unneeded pages out to swap space, if swap was !! 16 unneeded pages out to swap space. It has maximum size limits which can
17 mount. tmpfs also supports THP. !! 17 be adjusted on the fly via 'mount -o remount ...'
18 18
19 tmpfs extends ramfs with a few userspace confi !! 19 If you compare it to ramfs (which was the template to create tmpfs)
20 explained further below, some of which can be !! 20 you gain swapping and limit checking. Another similar thing is the RAM
21 fly using a remount ('mount -o remount ...') o !! 21 disk (/dev/ram*), which simulates a fixed size hard disk in physical
22 filesystem can be resized but it cannot be res !! 22 RAM, where you have to create an ordinary filesystem on top. Ramdisks
23 usage. tmpfs also supports POSIX ACLs, and ext !! 23 cannot swap and you do not have the possibility to resize them.
24 trusted.*, security.* and user.* namespaces. r <<
25 cannot modify any parameter for a ramfs filesy <<
26 filesystem is how much memory you have availab <<
27 used so to not run out of memory. <<
28 <<
29 An alternative to tmpfs and ramfs is to use br <<
30 (/dev/ram*), which allows you to simulate a bl <<
31 To write data you would just then need to crea <<
32 this ramdisk. As with ramfs, brd ramdisks cann <<
33 configured in size at initialization and you c <<
34 Contrary to brd ramdisks, tmpfs has its own fi <<
35 block layer at all. <<
36 24
37 Since tmpfs lives completely in the page cache !! 25 Since tmpfs lives completely in the page cache and on swap, all tmpfs
38 all tmpfs pages will be shown as "Shmem" in /p !! 26 pages will be shown as "Shmem" in /proc/meminfo and "Shared" in
39 free(1). Notice that these counters also inclu 27 free(1). Notice that these counters also include shared memory
40 (shmem, see ipcs(1)). The most reliable way to 28 (shmem, see ipcs(1)). The most reliable way to get the count is
41 using df(1) and du(1). 29 using df(1) and du(1).
42 30
43 tmpfs has the following uses: 31 tmpfs has the following uses:
44 32
45 1) There is always a kernel internal mount whi 33 1) There is always a kernel internal mount which you will not see at
46 all. This is used for shared anonymous mapp 34 all. This is used for shared anonymous mappings and SYSV shared
47 memory. 35 memory.
48 36
49 This mount does not depend on CONFIG_TMPFS. 37 This mount does not depend on CONFIG_TMPFS. If CONFIG_TMPFS is not
50 set, the user visible part of tmpfs is not 38 set, the user visible part of tmpfs is not built. But the internal
51 mechanisms are always present. 39 mechanisms are always present.
52 40
53 2) glibc 2.2 and above expects tmpfs to be mou 41 2) glibc 2.2 and above expects tmpfs to be mounted at /dev/shm for
54 POSIX shared memory (shm_open, shm_unlink). 42 POSIX shared memory (shm_open, shm_unlink). Adding the following
55 line to /etc/fstab should take care of this 43 line to /etc/fstab should take care of this::
56 44
57 tmpfs /dev/shm tmpfs defaul 45 tmpfs /dev/shm tmpfs defaults 0 0
58 46
59 Remember to create the directory that you i 47 Remember to create the directory that you intend to mount tmpfs on
60 if necessary. 48 if necessary.
61 49
62 This mount is _not_ needed for SYSV shared 50 This mount is _not_ needed for SYSV shared memory. The internal
63 mount is used for that. (In the 2.3 kernel 51 mount is used for that. (In the 2.3 kernel versions it was
64 necessary to mount the predecessor of tmpfs 52 necessary to mount the predecessor of tmpfs (shm fs) to use SYSV
65 shared memory.) 53 shared memory.)
66 54
67 3) Some people (including me) find it very con 55 3) Some people (including me) find it very convenient to mount it
68 e.g. on /tmp and /var/tmp and have a big sw 56 e.g. on /tmp and /var/tmp and have a big swap partition. And now
69 loop mounts of tmpfs files do work, so mkin 57 loop mounts of tmpfs files do work, so mkinitrd shipped by most
70 distributions should succeed with a tmpfs / 58 distributions should succeed with a tmpfs /tmp.
71 59
72 4) And probably a lot more I do not know about 60 4) And probably a lot more I do not know about :-)
73 61
74 62
75 tmpfs has three mount options for sizing: 63 tmpfs has three mount options for sizing:
76 64
77 ========= =================================== 65 ========= ============================================================
78 size The limit of allocated bytes for th 66 size The limit of allocated bytes for this tmpfs instance. The
79 default is half of your physical RA 67 default is half of your physical RAM without swap. If you
80 oversize your tmpfs instances the m 68 oversize your tmpfs instances the machine will deadlock
81 since the OOM handler will not be a 69 since the OOM handler will not be able to free that memory.
82 nr_blocks The same as size, but in blocks of 70 nr_blocks The same as size, but in blocks of PAGE_SIZE.
83 nr_inodes The maximum number of inodes for th 71 nr_inodes The maximum number of inodes for this instance. The default
84 is half of the number of your physi 72 is half of the number of your physical RAM pages, or (on a
85 machine with highmem) the number of 73 machine with highmem) the number of lowmem RAM pages,
86 whichever is the lower. 74 whichever is the lower.
87 ========= =================================== 75 ========= ============================================================
88 76
89 These parameters accept a suffix k, m or g for 77 These parameters accept a suffix k, m or g for kilo, mega and giga and
90 can be changed on remount. The size parameter 78 can be changed on remount. The size parameter also accepts a suffix %
91 to limit this tmpfs instance to that percentag 79 to limit this tmpfs instance to that percentage of your physical RAM:
92 the default, when neither size nor nr_blocks i 80 the default, when neither size nor nr_blocks is specified, is size=50%
93 81
94 If nr_blocks=0 (or size=0), blocks will not be 82 If nr_blocks=0 (or size=0), blocks will not be limited in that instance;
95 if nr_inodes=0, inodes will not be limited. I 83 if nr_inodes=0, inodes will not be limited. It is generally unwise to
96 mount with such options, since it allows any u 84 mount with such options, since it allows any user with write access to
97 use up all the memory on the machine; but enha 85 use up all the memory on the machine; but enhances the scalability of
98 that instance in a system with many CPUs makin 86 that instance in a system with many CPUs making intensive use of it.
99 87
100 If nr_inodes is not 0, that limited space for <<
101 extended attributes: "df -i"'s IUsed and IUse% <<
102 <<
103 tmpfs blocks may be swapped out, when there is <<
104 tmpfs has a mount option to disable its use of <<
105 <<
106 ====== ====================================== <<
107 noswap Disables swap. Remounts must respect t <<
108 By default swap is enabled. <<
109 ====== ====================================== <<
110 <<
111 tmpfs also supports Transparent Huge Pages whi <<
112 configured with CONFIG_TRANSPARENT_HUGEPAGE an <<
113 your system (has_transparent_hugepage(), which <<
114 The mount options for this are: <<
115 <<
116 ================ ============================= <<
117 huge=never Do not allocate huge pages. <<
118 huge=always Attempt to allocate huge page <<
119 huge=within_size Only allocate huge page if it <<
120 Also respect madvise(2) hints <<
121 huge=advise Only allocate huge page if re <<
122 ================ ============================= <<
123 <<
124 See also Documentation/admin-guide/mm/transhug <<
125 sysfs file /sys/kernel/mm/transparent_hugepage <<
126 be used to deny huge pages on all tmpfs mounts <<
127 force huge pages on all tmpfs mounts for testi <<
128 <<
129 tmpfs also supports quota with the following m <<
130 <<
131 ======================== ===================== <<
132 quota User and group quota <<
133 is enabled on the mou <<
134 system quota files th <<
135 usrquota User quota accounting <<
136 on the mount. <<
137 grpquota Group quota accountin <<
138 on the mount. <<
139 usrquota_block_hardlimit Set global user quota <<
140 usrquota_inode_hardlimit Set global user quota <<
141 grpquota_block_hardlimit Set global group quot <<
142 grpquota_inode_hardlimit Set global group quot <<
143 ======================== ===================== <<
144 <<
145 None of the quota related mount options can be <<
146 <<
147 Quota limit parameters accept a suffix k, m or <<
148 and can't be changed on remount. Default globa <<
149 effect for any and all user/group/project exce <<
150 quota entry for user/group/project id is being <<
151 first time an inode with a particular id owner <<
152 the mount. In other words, instead of the limi <<
153 they are initialized with the particular value <<
154 options. The limits can be changed for any use <<
155 normally can be. <<
156 <<
157 Note that tmpfs quotas do not support user nam <<
158 translation is done if quotas are enabled insi <<
159 88
160 tmpfs has a mount option to set the NUMA memor 89 tmpfs has a mount option to set the NUMA memory allocation policy for
161 all files in that instance (if CONFIG_NUMA is 90 all files in that instance (if CONFIG_NUMA is enabled) - which can be
162 adjusted on the fly via 'mount -o remount ...' 91 adjusted on the fly via 'mount -o remount ...'
163 92
164 ======================== ===================== 93 ======================== ==============================================
165 mpol=default use the process alloc 94 mpol=default use the process allocation policy
166 (see set_mempolicy(2) 95 (see set_mempolicy(2))
167 mpol=prefer:Node prefers to allocate m 96 mpol=prefer:Node prefers to allocate memory from the given Node
168 mpol=bind:NodeList allocates memory only 97 mpol=bind:NodeList allocates memory only from nodes in NodeList
169 mpol=interleave prefers to allocate f 98 mpol=interleave prefers to allocate from each node in turn
170 mpol=interleave:NodeList allocates from each n 99 mpol=interleave:NodeList allocates from each node of NodeList in turn
171 mpol=local prefers to allocate m 100 mpol=local prefers to allocate memory from the local node
172 ======================== ===================== 101 ======================== ==============================================
173 102
174 NodeList format is a comma-separated list of d 103 NodeList format is a comma-separated list of decimal numbers and ranges,
175 a range being two hyphen-separated decimal num 104 a range being two hyphen-separated decimal numbers, the smallest and
176 largest node numbers in the range. For exampl 105 largest node numbers in the range. For example, mpol=bind:0-3,5,7,9-15
177 106
178 A memory policy with a valid NodeList will be 107 A memory policy with a valid NodeList will be saved, as specified, for
179 use at file creation time. When a task alloca 108 use at file creation time. When a task allocates a file in the file
180 system, the mount option memory policy will be 109 system, the mount option memory policy will be applied with a NodeList,
181 if any, modified by the calling task's cpuset 110 if any, modified by the calling task's cpuset constraints
182 [See Documentation/admin-guide/cgroup-v1/cpuse 111 [See Documentation/admin-guide/cgroup-v1/cpusets.rst] and any optional flags,
183 listed below. If the resulting NodeLists is t 112 listed below. If the resulting NodeLists is the empty set, the effective
184 memory policy for the file will revert to "def 113 memory policy for the file will revert to "default" policy.
185 114
186 NUMA memory allocation policies have optional 115 NUMA memory allocation policies have optional flags that can be used in
187 conjunction with their modes. These optional 116 conjunction with their modes. These optional flags can be specified
188 when tmpfs is mounted by appending them to the 117 when tmpfs is mounted by appending them to the mode before the NodeList.
189 See Documentation/admin-guide/mm/numa_memory_p 118 See Documentation/admin-guide/mm/numa_memory_policy.rst for a list of
190 all available memory allocation policy mode fl 119 all available memory allocation policy mode flags and their effect on
191 memory policy. 120 memory policy.
192 121
193 :: 122 ::
194 123
195 =static is equivalent to 124 =static is equivalent to MPOL_F_STATIC_NODES
196 =relative is equivalent to 125 =relative is equivalent to MPOL_F_RELATIVE_NODES
197 126
198 For example, mpol=bind=static:NodeList, is the 127 For example, mpol=bind=static:NodeList, is the equivalent of an
199 allocation policy of MPOL_BIND | MPOL_F_STATIC 128 allocation policy of MPOL_BIND | MPOL_F_STATIC_NODES.
200 129
201 Note that trying to mount a tmpfs with an mpol 130 Note that trying to mount a tmpfs with an mpol option will fail if the
202 running kernel does not support NUMA; and will 131 running kernel does not support NUMA; and will fail if its nodelist
203 specifies a node which is not online. If your 132 specifies a node which is not online. If your system relies on that
204 tmpfs being mounted, but from time to time run 133 tmpfs being mounted, but from time to time runs a kernel built without
205 NUMA capability (perhaps a safe recovery kerne 134 NUMA capability (perhaps a safe recovery kernel), or with fewer nodes
206 online, then it is advisable to omit the mpol 135 online, then it is advisable to omit the mpol option from automatic
207 mount options. It can be added later, when th 136 mount options. It can be added later, when the tmpfs is already mounted
208 on MountPoint, by 'mount -o remount,mpol=Polic 137 on MountPoint, by 'mount -o remount,mpol=Policy:NodeList MountPoint'.
209 138
210 139
211 To specify the initial root directory you can 140 To specify the initial root directory you can use the following mount
212 options: 141 options:
213 142
214 ==== ================================== 143 ==== ==================================
215 mode The permissions as an octal number 144 mode The permissions as an octal number
216 uid The user id 145 uid The user id
217 gid The group id 146 gid The group id
218 ==== ================================== 147 ==== ==================================
219 148
220 These options do not have any effect on remoun 149 These options do not have any effect on remount. You can change these
221 parameters with chmod(1), chown(1) and chgrp(1 150 parameters with chmod(1), chown(1) and chgrp(1) on a mounted filesystem.
222 151
223 152
224 tmpfs has a mount option to select whether it 153 tmpfs has a mount option to select whether it will wrap at 32- or 64-bit inode
225 numbers: 154 numbers:
226 155
227 ======= ======================== 156 ======= ========================
228 inode64 Use 64-bit inode numbers 157 inode64 Use 64-bit inode numbers
229 inode32 Use 32-bit inode numbers 158 inode32 Use 32-bit inode numbers
230 ======= ======================== 159 ======= ========================
231 160
232 On a 32-bit kernel, inode32 is implicit, and i 161 On a 32-bit kernel, inode32 is implicit, and inode64 is refused at mount time.
233 On a 64-bit kernel, CONFIG_TMPFS_INODE64 sets 162 On a 64-bit kernel, CONFIG_TMPFS_INODE64 sets the default. inode64 avoids the
234 possibility of multiple files with the same in 163 possibility of multiple files with the same inode number on a single device;
235 but risks glibc failing with EOVERFLOW once 33 164 but risks glibc failing with EOVERFLOW once 33-bit inode numbers are reached -
236 if a long-lived tmpfs is accessed by 32-bit ap 165 if a long-lived tmpfs is accessed by 32-bit applications so ancient that
237 opening a file larger than 2GiB fails with EIN 166 opening a file larger than 2GiB fails with EINVAL.
238 167
239 168
240 So 'mount -t tmpfs -o size=10G,nr_inodes=10k,m 169 So 'mount -t tmpfs -o size=10G,nr_inodes=10k,mode=700 tmpfs /mytmpfs'
241 will give you tmpfs instance on /mytmpfs which 170 will give you tmpfs instance on /mytmpfs which can allocate 10GB
242 RAM/SWAP in 10240 inodes and it is only access 171 RAM/SWAP in 10240 inodes and it is only accessible by root.
243 172
244 173
245 :Author: 174 :Author:
246 Christoph Rohland <cr@sap.com>, 1.12.01 175 Christoph Rohland <cr@sap.com>, 1.12.01
247 :Updated: 176 :Updated:
248 Hugh Dickins, 4 June 2007 177 Hugh Dickins, 4 June 2007
249 :Updated: 178 :Updated:
250 KOSAKI Motohiro, 16 Mar 2010 179 KOSAKI Motohiro, 16 Mar 2010
251 :Updated: 180 :Updated:
252 Chris Down, 13 July 2020 181 Chris Down, 13 July 2020
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.