1 Documentation for /proc/sys/fs/* kernel version 2.2.10
2 (c) 1998, 1999, Rik van Riel <riel@nl.linux.org>
3 (c) 2009, Shen Feng<shen@cn.fujitsu.com>
5 For general info and legal blurb, please look in README.
7 ==============================================================
9 This file contains documentation for the sysctl files in
10 /proc/sys/fs/ and is valid for Linux kernel version 2.2.
12 The files in this directory can be used to tune and monitor
13 miscellaneous and general things in the operation of the Linux
14 kernel. Since some of the files _can_ be used to screw up your
15 system, it is advisable to read both documentation and source
16 before actually making adjustments.
19 ----------------------------------------------------------
21 Currently, these files are in /proc/sys/fs:
39 ==============================================================
43 aio-nr is the running total of the number of events specified on the
44 io_setup system call for all currently active aio contexts. If aio-nr
45 reaches aio-max-nr then io_setup will fail with EAGAIN. Note that
46 raising aio-max-nr does not result in the pre-allocation or re-sizing
47 of any kernel data structures.
49 ==============================================================
53 From linux/fs/dentry.c:
54 --------------------------------------------------------------
58 int age_limit; /* age in seconds */
59 int want_pages; /* pages requested by system */
61 } dentry_stat = {0, 0, 45, 0,};
62 --------------------------------------------------------------
64 Dentries are dynamically allocated and deallocated, and
65 nr_dentry seems to be 0 all the time. Hence it's safe to
66 assume that only nr_unused, age_limit and want_pages are
67 used. Nr_unused seems to be exactly what its name says.
68 Age_limit is the age in seconds after which dcache entries
69 can be reclaimed when memory is short and want_pages is
70 nonzero when shrink_dcache_pages() has been called and the
71 dcache isn't pruned yet.
73 ==============================================================
77 The file dquot-max shows the maximum number of cached disk
80 The file dquot-nr shows the number of allocated disk quota
81 entries and the number of free disk quota entries.
83 If the number of free cached disk quotas is very low and
84 you have some awesome number of simultaneous system users,
85 you might want to raise the limit.
87 ==============================================================
91 The kernel allocates file handles dynamically, but as yet it
92 doesn't free them again.
94 The value in file-max denotes the maximum number of file-
95 handles that the Linux kernel will allocate. When you get lots
96 of error messages about running out of file handles, you might
97 want to increase this limit.
99 The three values in file-nr denote the number of allocated
100 file handles, the number of unused file handles and the maximum
101 number of file handles. When the allocated file handles come
102 close to the maximum, but the number of unused file handles is
103 significantly greater than 0, you've encountered a peak in your
104 usage of file handles and you don't need to increase the maximum.
106 ==============================================================
110 This denotes the maximum number of file-handles a process can
111 allocate. Default value is 1024*1024 (1048576) which should be
112 enough for most machines. Actual limit depends on RLIMIT_NOFILE
115 ==============================================================
117 inode-max, inode-nr & inode-state:
119 As with file handles, the kernel allocates the inode structures
120 dynamically, but can't free them yet.
122 The value in inode-max denotes the maximum number of inode
123 handlers. This value should be 3-4 times larger than the value
124 in file-max, since stdin, stdout and network sockets also
125 need an inode struct to handle them. When you regularly run
126 out of inodes, you need to increase this value.
128 The file inode-nr contains the first two items from
129 inode-state, so we'll skip to that file...
131 Inode-state contains three actual numbers and four dummies.
132 The actual numbers are, in order of appearance, nr_inodes,
133 nr_free_inodes and preshrink.
135 Nr_inodes stands for the number of inodes the system has
136 allocated, this can be slightly more than inode-max because
137 Linux allocates them one pageful at a time.
139 Nr_free_inodes represents the number of free inodes (?) and
140 preshrink is nonzero when the nr_inodes > inode-max and the
141 system needs to prune the inode list instead of allocating
144 ==============================================================
146 overflowgid & overflowuid:
148 Some filesystems only support 16-bit UIDs and GIDs, although in Linux
149 UIDs and GIDs are 32 bits. When one of these filesystems is mounted
150 with writes enabled, any UID or GID that would exceed 65535 is translated
151 to a fixed value before being written to disk.
153 These sysctls allow you to change the value of the fixed UID and GID.
154 The default is 65534.
156 ==============================================================
160 This value can be used to query and set the core dump mode for setuid
161 or otherwise protected/tainted binaries. The modes are
163 0 - (default) - traditional behaviour. Any process which has changed
164 privilege levels or is execute only will not be dumped
165 1 - (debug) - all processes dump core when possible. The core dump is
166 owned by the current user and no security is applied. This is
167 intended for system debugging situations only. Ptrace is unchecked.
168 2 - (suidsafe) - any binary which normally would not be dumped is dumped
169 readable by root only. This allows the end user to remove
170 such a dump but not access it directly. For security reasons
171 core dumps in this mode will not overwrite one another or
172 other files. This mode is appropriate when administrators are
173 attempting to debug problems in a normal environment.
175 ==============================================================
177 super-max & super-nr:
179 These numbers control the maximum number of superblocks, and
180 thus the maximum number of mounted filesystems the kernel
181 can have. You only need to increase super-max if you need to
182 mount more filesystems than the current value in super-max
185 ==============================================================
189 aio-nr shows the current system-wide number of asynchronous io
190 requests. aio-max-nr allows you to change the maximum value
193 ==============================================================
196 2. /proc/sys/fs/binfmt_misc
197 ----------------------------------------------------------
199 Documentation for the files in /proc/sys/fs/binfmt_misc is
200 in Documentation/binfmt_misc.txt.
203 3. /proc/sys/fs/mqueue - POSIX message queues filesystem
204 ----------------------------------------------------------
206 The "mqueue" filesystem provides the necessary kernel features to enable the
207 creation of a user space library that implements the POSIX message queues
208 API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System
209 Interfaces specification.)
211 The "mqueue" filesystem contains values for determining/setting the amount of
212 resources used by the file system.
214 /proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the
215 maximum number of message queues allowed on the system.
217 /proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the
218 maximum number of messages in a queue value. In fact it is the limiting value
219 for another (user) limit which is set in mq_open invocation. This attribute of
220 a queue must be less or equal then msg_max.
222 /proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the
223 maximum message size value (it is every message queue's attribute set during
227 4. /proc/sys/fs/epoll - Configuration options for the epoll interface
228 --------------------------------------------------------
230 This directory contains configuration options for the epoll(7) interface.
235 This is the maximum number of epoll file descriptors that a single user can
236 have open at a given time. The default value is 128, and should be enough
242 Every epoll file descriptor can store a number of files to be monitored
243 for event readiness. Each one of these monitored files constitutes a "watch".
244 This configuration option sets the maximum number of "watches" that are
245 allowed for each user.
246 Each "watch" costs roughly 90 bytes on a 32bit kernel, and roughly 160 bytes
248 The current default value for max_user_watches is the 1/32 of the available
249 low memory, divided for the "watch" cost in bytes.