1Documentation 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> 4 5For general info and legal blurb, please look in README. 6 7============================================================== 8 9This file contains documentation for the sysctl files in 10/proc/sys/fs/ and is valid for Linux kernel version 2.2. 11 12The files in this directory can be used to tune and monitor 13miscellaneous and general things in the operation of the Linux 14kernel. Since some of the files _can_ be used to screw up your 15system, it is advisable to read both documentation and source 16before actually making adjustments. 17 181. /proc/sys/fs 19---------------------------------------------------------- 20 21Currently, these files are in /proc/sys/fs: 22- aio-max-nr 23- aio-nr 24- dentry-state 25- dquot-max 26- dquot-nr 27- file-max 28- file-nr 29- inode-max 30- inode-nr 31- inode-state 32- nr_open 33- overflowuid 34- overflowgid 35- pipe-user-pages-hard 36- pipe-user-pages-soft 37- protected_hardlinks 38- protected_symlinks 39- suid_dumpable 40- super-max 41- super-nr 42 43============================================================== 44 45aio-nr & aio-max-nr: 46 47aio-nr is the running total of the number of events specified on the 48io_setup system call for all currently active aio contexts. If aio-nr 49reaches aio-max-nr then io_setup will fail with EAGAIN. Note that 50raising aio-max-nr does not result in the pre-allocation or re-sizing 51of any kernel data structures. 52 53============================================================== 54 55dentry-state: 56 57From linux/fs/dentry.c: 58-------------------------------------------------------------- 59struct { 60 int nr_dentry; 61 int nr_unused; 62 int age_limit; /* age in seconds */ 63 int want_pages; /* pages requested by system */ 64 int dummy[2]; 65} dentry_stat = {0, 0, 45, 0,}; 66-------------------------------------------------------------- 67 68Dentries are dynamically allocated and deallocated, and 69nr_dentry seems to be 0 all the time. Hence it's safe to 70assume that only nr_unused, age_limit and want_pages are 71used. Nr_unused seems to be exactly what its name says. 72Age_limit is the age in seconds after which dcache entries 73can be reclaimed when memory is short and want_pages is 74nonzero when shrink_dcache_pages() has been called and the 75dcache isn't pruned yet. 76 77============================================================== 78 79dquot-max & dquot-nr: 80 81The file dquot-max shows the maximum number of cached disk 82quota entries. 83 84The file dquot-nr shows the number of allocated disk quota 85entries and the number of free disk quota entries. 86 87If the number of free cached disk quotas is very low and 88you have some awesome number of simultaneous system users, 89you might want to raise the limit. 90 91============================================================== 92 93file-max & file-nr: 94 95The value in file-max denotes the maximum number of file- 96handles that the Linux kernel will allocate. When you get lots 97of error messages about running out of file handles, you might 98want to increase this limit. 99 100Historically,the kernel was able to allocate file handles 101dynamically, but not to free them again. The three values in 102file-nr denote the number of allocated file handles, the number 103of allocated but unused file handles, and the maximum number of 104file handles. Linux 2.6 always reports 0 as the number of free 105file handles -- this is not an error, it just means that the 106number of allocated file handles exactly matches the number of 107used file handles. 108 109Attempts to allocate more file descriptors than file-max are 110reported with printk, look for "VFS: file-max limit <number> 111reached". 112============================================================== 113 114nr_open: 115 116This denotes the maximum number of file-handles a process can 117allocate. Default value is 1024*1024 (1048576) which should be 118enough for most machines. Actual limit depends on RLIMIT_NOFILE 119resource limit. 120 121============================================================== 122 123inode-max, inode-nr & inode-state: 124 125As with file handles, the kernel allocates the inode structures 126dynamically, but can't free them yet. 127 128The value in inode-max denotes the maximum number of inode 129handlers. This value should be 3-4 times larger than the value 130in file-max, since stdin, stdout and network sockets also 131need an inode struct to handle them. When you regularly run 132out of inodes, you need to increase this value. 133 134The file inode-nr contains the first two items from 135inode-state, so we'll skip to that file... 136 137Inode-state contains three actual numbers and four dummies. 138The actual numbers are, in order of appearance, nr_inodes, 139nr_free_inodes and preshrink. 140 141Nr_inodes stands for the number of inodes the system has 142allocated, this can be slightly more than inode-max because 143Linux allocates them one pageful at a time. 144 145Nr_free_inodes represents the number of free inodes (?) and 146preshrink is nonzero when the nr_inodes > inode-max and the 147system needs to prune the inode list instead of allocating 148more. 149 150============================================================== 151 152overflowgid & overflowuid: 153 154Some filesystems only support 16-bit UIDs and GIDs, although in Linux 155UIDs and GIDs are 32 bits. When one of these filesystems is mounted 156with writes enabled, any UID or GID that would exceed 65535 is translated 157to a fixed value before being written to disk. 158 159These sysctls allow you to change the value of the fixed UID and GID. 160The default is 65534. 161 162============================================================== 163 164pipe-user-pages-hard: 165 166Maximum total number of pages a non-privileged user may allocate for pipes. 167Once this limit is reached, no new pipes may be allocated until usage goes 168below the limit again. When set to 0, no limit is applied, which is the default 169setting. 170 171============================================================== 172 173pipe-user-pages-soft: 174 175Maximum total number of pages a non-privileged user may allocate for pipes 176before the pipe size gets limited to a single page. Once this limit is reached, 177new pipes will be limited to a single page in size for this user in order to 178limit total memory usage, and trying to increase them using fcntl() will be 179denied until usage goes below the limit again. The default value allows to 180allocate up to 1024 pipes at their default size. When set to 0, no limit is 181applied. 182 183============================================================== 184 185protected_hardlinks: 186 187A long-standing class of security issues is the hardlink-based 188time-of-check-time-of-use race, most commonly seen in world-writable 189directories like /tmp. The common method of exploitation of this flaw 190is to cross privilege boundaries when following a given hardlink (i.e. a 191root process follows a hardlink created by another user). Additionally, 192on systems without separated partitions, this stops unauthorized users 193from "pinning" vulnerable setuid/setgid files against being upgraded by 194the administrator, or linking to special files. 195 196When set to "0", hardlink creation behavior is unrestricted. 197 198When set to "1" hardlinks cannot be created by users if they do not 199already own the source file, or do not have read/write access to it. 200 201This protection is based on the restrictions in Openwall and grsecurity. 202 203============================================================== 204 205protected_symlinks: 206 207A long-standing class of security issues is the symlink-based 208time-of-check-time-of-use race, most commonly seen in world-writable 209directories like /tmp. The common method of exploitation of this flaw 210is to cross privilege boundaries when following a given symlink (i.e. a 211root process follows a symlink belonging to another user). For a likely 212incomplete list of hundreds of examples across the years, please see: 213http://cve.mitre.org/cgi-bin/cvekey.cgi?keyword=/tmp 214 215When set to "0", symlink following behavior is unrestricted. 216 217When set to "1" symlinks are permitted to be followed only when outside 218a sticky world-writable directory, or when the uid of the symlink and 219follower match, or when the directory owner matches the symlink's owner. 220 221This protection is based on the restrictions in Openwall and grsecurity. 222 223============================================================== 224 225suid_dumpable: 226 227This value can be used to query and set the core dump mode for setuid 228or otherwise protected/tainted binaries. The modes are 229 2300 - (default) - traditional behaviour. Any process which has changed 231 privilege levels or is execute only will not be dumped. 2321 - (debug) - all processes dump core when possible. The core dump is 233 owned by the current user and no security is applied. This is 234 intended for system debugging situations only. Ptrace is unchecked. 235 This is insecure as it allows regular users to examine the memory 236 contents of privileged processes. 2372 - (suidsafe) - any binary which normally would not be dumped is dumped 238 anyway, but only if the "core_pattern" kernel sysctl is set to 239 either a pipe handler or a fully qualified path. (For more details 240 on this limitation, see CVE-2006-2451.) This mode is appropriate 241 when administrators are attempting to debug problems in a normal 242 environment, and either have a core dump pipe handler that knows 243 to treat privileged core dumps with care, or specific directory 244 defined for catching core dumps. If a core dump happens without 245 a pipe handler or fully qualifid path, a message will be emitted 246 to syslog warning about the lack of a correct setting. 247 248============================================================== 249 250super-max & super-nr: 251 252These numbers control the maximum number of superblocks, and 253thus the maximum number of mounted filesystems the kernel 254can have. You only need to increase super-max if you need to 255mount more filesystems than the current value in super-max 256allows you to. 257 258============================================================== 259 260aio-nr & aio-max-nr: 261 262aio-nr shows the current system-wide number of asynchronous io 263requests. aio-max-nr allows you to change the maximum value 264aio-nr can grow to. 265 266============================================================== 267 268 2692. /proc/sys/fs/binfmt_misc 270---------------------------------------------------------- 271 272Documentation for the files in /proc/sys/fs/binfmt_misc is 273in Documentation/binfmt_misc.txt. 274 275 2763. /proc/sys/fs/mqueue - POSIX message queues filesystem 277---------------------------------------------------------- 278 279The "mqueue" filesystem provides the necessary kernel features to enable the 280creation of a user space library that implements the POSIX message queues 281API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System 282Interfaces specification.) 283 284The "mqueue" filesystem contains values for determining/setting the amount of 285resources used by the file system. 286 287/proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the 288maximum number of message queues allowed on the system. 289 290/proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the 291maximum number of messages in a queue value. In fact it is the limiting value 292for another (user) limit which is set in mq_open invocation. This attribute of 293a queue must be less or equal then msg_max. 294 295/proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the 296maximum message size value (it is every message queue's attribute set during 297its creation). 298 299/proc/sys/fs/mqueue/msg_default is a read/write file for setting/getting the 300default number of messages in a queue value if attr parameter of mq_open(2) is 301NULL. If it exceed msg_max, the default value is initialized msg_max. 302 303/proc/sys/fs/mqueue/msgsize_default is a read/write file for setting/getting 304the default message size value if attr parameter of mq_open(2) is NULL. If it 305exceed msgsize_max, the default value is initialized msgsize_max. 306 3074. /proc/sys/fs/epoll - Configuration options for the epoll interface 308-------------------------------------------------------- 309 310This directory contains configuration options for the epoll(7) interface. 311 312max_user_watches 313---------------- 314 315Every epoll file descriptor can store a number of files to be monitored 316for event readiness. Each one of these monitored files constitutes a "watch". 317This configuration option sets the maximum number of "watches" that are 318allowed for each user. 319Each "watch" costs roughly 90 bytes on a 32bit kernel, and roughly 160 bytes 320on a 64bit one. 321The current default value for max_user_watches is the 1/32 of the available 322low memory, divided for the "watch" cost in bytes. 323 324