1 .\" Copyright (C) 1994, 1995 by Daniel Quinlan (quinlan@yggdrasil.com)
2 .\" and Copyright (C) 2002-2008 Michael Kerrisk <mtk.manpages@gmail.com>
3 .\" with networking additions from Alan Cox (A.Cox@swansea.ac.uk)
4 .\" and scsi additions from Michael Neuffer (neuffer@mail.uni-mainz.de)
5 .\" and sysctl additions from Andries Brouwer (aeb@cwi.nl)
6 .\" and System V IPC (as well as various other) additions from
7 .\" Michael Kerrisk <mtk.manpages@gmail.com>
9 .\" %%%LICENSE_START(GPLv2+_DOC_FULL)
10 .\" This is free documentation; you can redistribute it and/or
11 .\" modify it under the terms of the GNU General Public License as
12 .\" published by the Free Software Foundation; either version 2 of
13 .\" the License, or (at your option) any later version.
15 .\" The GNU General Public License's references to "object code"
16 .\" and "executables" are to be interpreted as the output of any
17 .\" document formatting or typesetting system, including
18 .\" intermediate and printed output.
20 .\" This manual is distributed in the hope that it will be useful,
21 .\" but WITHOUT ANY WARRANTY; without even the implied warranty of
22 .\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23 .\" GNU General Public License for more details.
25 .\" You should have received a copy of the GNU General Public
26 .\" License along with this manual; if not, see
27 .\" <http://www.gnu.org/licenses/>.
30 .\" Modified 1995-05-17 by faith@cs.unc.edu
31 .\" Minor changes by aeb and Marty Leisner (leisner@sdsp.mc.xerox.com).
32 .\" Modified 1996-04-13, 1996-07-22 by aeb@cwi.nl
33 .\" Modified 2001-12-16 by rwhron@earthlink.net
34 .\" Modified 2002-07-13 by jbelton@shaw.ca
35 .\" Modified 2002-07-22, 2003-05-27, 2004-04-06, 2004-05-25
36 .\" by Michael Kerrisk <mtk.manpages@gmail.com>
37 .\" 2004-11-17, mtk -- updated notes on /proc/loadavg
38 .\" 2004-12-01, mtk, rtsig-max and rtsig-nr went away in 2.6.8
39 .\" 2004-12-14, mtk, updated 'statm', and fixed error in order of list
40 .\" 2005-05-12, mtk, updated 'stat'
41 .\" 2005-07-13, mtk, added /proc/sys/fs/mqueue/*
42 .\" 2005-09-16, mtk, Added /proc/sys/fs/suid_dumpable
43 .\" 2005-09-19, mtk, added /proc/zoneinfo
44 .\" 2005-03-01, mtk, moved /proc/sys/fs/mqueue/* material to mq_overview.7.
45 .\" 2008-06-05, mtk, Added /proc/[pid]/oom_score, /proc/[pid]/oom_adj,
46 .\" /proc/[pid]/limits, /proc/[pid]/mountinfo, /proc/[pid]/mountstats,
47 .\" and /proc/[pid]/fdinfo/*.
48 .\" 2008-06-19, mtk, Documented /proc/[pid]/status.
49 .\" 2008-07-15, mtk, added /proc/config.gz
51 .\" FIXME cross check against Documentation/filesystems/proc.txt
52 .\" to see what information could be imported from that file
55 .TH PROC 5 2017-05-03 "Linux" "Linux Programmer's Manual"
57 proc \- process information pseudo-filesystem
61 filesystem is a pseudo-filesystem which provides an interface to
62 kernel data structures.
63 It is commonly mounted at
65 Most of it is read-only, but some files allow kernel variables to be
70 filesystem supports the following mount options:
72 .BR hidepid "=\fIn\fP (since Linux 3.3)"
73 .\" commit 0499680a42141d86417a8fbaa8c8db806bea1201
74 This option controls who can access the information in
79 is one of the following values:
83 Everybody may access all
86 This is the traditional behavior,
87 and the default if this mount option is not specified.
90 Users may not access files and subdirectories inside any
92 directories but their own (the
94 directories themselves remain visible).
95 Sensitive files such as
96 .IR /proc/[pid]/cmdline
98 .IR /proc/[pid]/status
99 are now protected against other users.
100 This makes it impossible to learn whether any user is running a
102 (so long as the program doesn't otherwise reveal itself by its behavior).
103 .\" As an additional bonus, since
104 .\" .IR /proc/[pid]/cmdline
105 .\" is unaccessible for other users,
106 .\" poorly written programs passing sensitive information via
107 .\" program arguments are now protected against local eavesdroppers.
110 As for mode 1, but in addition the
112 directories belonging to other users become invisible.
115 entries can no longer be used to discover the PIDs on the system.
116 This doesn't hide the fact that a process with a specific PID value exists
117 (it can be learned by other means, for example, by "kill -0 $PID"),
118 but it hides a process's UID and GID,
119 which could otherwise be learned by employing
124 This greatly complicates an attacker's task of gathering
125 information about running processes (e.g., discovering whether
126 some daemon is running with elevated privileges,
127 whether another user is running some sensitive program,
128 whether other users are running any program at all, and so on).
131 .BR gid "=\fIgid\fP (since Linux 3.3)"
132 .\" commit 0499680a42141d86417a8fbaa8c8db806bea1201
133 Specifies the ID of a group whose members are authorized to
134 learn process information otherwise prohibited by
136 (i.e., users in this group behave as though
140 This group should be used instead of approaches such as putting
141 nonroot users into the
144 .SS Files and directories
145 The following list describes many of the files and directories under the
151 There is a numerical subdirectory for each running process; the
152 subdirectory is named by the process ID.
156 subdirectory contains the
157 pseudo-files and directories described below.
158 These files are normally owned by the effective user and
159 effective group ID of the process.
160 However, as a security measure, the ownership is made
162 if the process's "dumpable" attribute is set to a value other than 1.
163 This attribute may change for the following reasons:
166 The attribute was explicitly set via the
171 The attribute was reset to the value in the file
172 .IR /proc/sys/fs/suid_dumpable
173 (described below), for the reasons described in
177 Resetting the "dumpable" attribute to 1 reverts the ownership of the
179 files to the process's real UID and real GID.
182 .\" https://lwn.net/Articles/28222/
183 .\" From: Stephen Smalley <sds@epoch.ncsc.mil>
184 .\" To: LKML and others
185 .\" Subject: [RFC][PATCH] Process Attribute API for Security Modules
186 .\" Date: 08 Apr 2003 16:17:52 -0400
188 .\" http://www.nsa.gov/research/_files/selinux/papers/module/x362.shtml
190 The files in this directory provide an API for security modules.
191 The contents of this directory are files that can be read and written
192 in order to set security-related attributes.
193 This directory was added to support SELinux,
194 but the intention was that the API be general enough to support
195 other security modules.
196 For the purpose of explanation,
197 examples of how SELinux uses these files are provided below.
199 This directory is present only if the kernel was configured with
200 .BR CONFIG_SECURITY .
202 .IR /proc/[pid]/attr/current " (since Linux 2.6.0)"
203 The contents of this file represent the current
204 security attributes of the process.
206 In SELinux, this file is used to get the security context of a process.
207 Prior to Linux 2.6.11, this file could not be used to set the security
208 context (a write was always denied), since SELinux limited process security
211 (see the description of
212 .IR /proc/[pid]/attr/exec ,
214 Since Linux 2.6.11, SELinux lifted this restriction and began supporting
215 "set" operations via writes to this node if authorized by policy,
216 although use of this operation is only suitable for applications that are
217 trusted to maintain any desired separation between the old and new security
219 Prior to Linux 2.6.28, SELinux did not allow threads within a
220 multi-threaded process to set their security context via this node
221 as it would yield an inconsistency among the security contexts of the
222 threads sharing the same memory space.
223 Since Linux 2.6.28, SELinux lifted
224 this restriction and began supporting "set" operations for threads within
225 a multithreaded process if the new security context is bounded by the old
226 security context, where the bounded relation is defined in policy and
227 guarantees that the new security context has a subset of the permissions
228 of the old security context.
229 Other security modules may choose to support "set" operations via
232 .IR /proc/[pid]/attr/exec " (since Linux 2.6.0)"
233 This file represents the attributes to assign to the
234 process upon a subsequent
238 this is needed to support role/domain transitions, and
240 is the preferred point to make such transitions because it offers better
241 control over the initialization of the process in the new security label
242 and the inheritance of state.
243 In SELinux, this attribute is reset on
245 so that the new program reverts to the default behavior for any
247 calls that it may make.
248 In SELinux, a process can set
250 .I /proc/[pid]/attr/exec
253 .IR /proc/[pid]/attr/fscreate " (since Linux 2.6.0)"
254 This file represents the attributes to assign to files
255 created by subsequent calls to
262 SELinux employs this file to support creation of a file
263 (using the aforementioned system calls)
265 so that there is no risk of inappropriate access being obtained
266 between the time of creation and the time that attributes are set.
267 In SELinux, this attribute is reset on
269 so that the new program reverts to the default behavior for
270 any file creation calls it may make, but the attribute will persist
271 across multiple file creation calls within a program unless it is
273 In SELinux, a process can set only its own
274 .IR /proc/[pid]/attr/fscreate
277 .IR /proc/[pid]/attr/keycreate " (since Linux 2.6.18)"
278 .\" commit 4eb582cf1fbd7b9e5f466e3718a59c957e75254e
279 .\" /Documentation/keys.txt
280 If a process writes a security context into this file,
281 all subsequently created keys
283 will be labeled with this context.
284 For further information, see the kernel source file
285 .IR Documentation/keys.txt .
287 .IR /proc/[pid]/attr/prev " (since Linux 2.6.0)"
288 This file contains the security context of the process before the last
290 that is, the previous value of
291 .IR /proc/[pid]/attr/current .
293 .IR /proc/[pid]/attr/socketcreate " (since Linux 2.6.18)"
294 .\" commit 42c3e03ef6b298813557cdb997bd6db619cd65a2
295 If a process writes a security context into this file,
296 all subsequently created sockets will be labeled with this context.
298 .IR /proc/[pid]/autogroup " (since Linux 2.6.38)"
299 .\" commit 5091faa449ee0b7d73bc296a93bca9540fc51d0a
303 .IR /proc/[pid]/auxv " (since 2.6.0-test7)"
304 This contains the contents of the ELF interpreter information passed
305 to the process at exec time.
306 The format is one \fIunsigned long\fP ID
307 plus one \fIunsigned long\fP value for each entry.
308 The last entry contains two zeros.
312 Permission to access this file is governed by a ptrace access mode
313 .B PTRACE_MODE_READ_FSCREDS
317 .IR /proc/[pid]/cgroup " (since Linux 2.6.24)"
321 .IR /proc/[pid]/clear_refs " (since Linux 2.6.22)"
322 .\" commit b813e931b4c8235bb42e301096ea97dbdee3e8fe (2.6.22)
323 .\" commit 398499d5f3613c47f2143b8c54a04efb5d7a6da9 (2.6.32)
324 .\" commit 040fa02077de01c7e08fa75be6125e4ca5636011 (3.11)
326 .\" "Clears page referenced bits shown in smaps output"
327 .\" write-only, writable only by the owner of the process
329 This is a write-only file, writable only by owner of the process.
331 The following values may be written to the file:
334 1 (since Linux 2.6.22)
335 .\" Internally: CLEAR_REFS_ALL
336 Reset the PG_Referenced and ACCESSED/YOUNG
337 bits for all the pages associated with the process.
338 (Before kernel 2.6.32, writing any nonzero value to this file
341 2 (since Linux 2.6.32)
342 .\" Internally: CLEAR_REFS_ANON
343 Reset the PG_Referenced and ACCESSED/YOUNG
344 bits for all anonymous pages associated with the process.
346 3 (since Linux 2.6.32)
347 .\" Internally: CLEAR_REFS_MAPPED
348 Reset the PG_Referenced and ACCESSED/YOUNG
349 bits for all file-mapped pages associated with the process.
352 Clearing the PG_Referenced and ACCESSED/YOUNG bits provides a method
353 to measure approximately how much memory a process is using.
354 One first inspects the values in the "Referenced" fields
355 for the VMAs shown in
356 .IR /proc/[pid]/smaps
357 to get an idea of the memory footprint of the
359 One then clears the PG_Referenced and ACCESSED/YOUNG bits
360 and, after some measured time interval,
361 once again inspects the values in the "Referenced" fields
362 to get an idea of the change in memory footprint of the
363 process during the measured interval.
364 If one is interested only in inspecting the selected mapping types,
365 then the value 2 or 3 can be used instead of 1.
367 Further values can be written to affect different properties:
371 Clear the soft-dirty bit for all the pages associated with the process.
372 .\" Internally: CLEAR_REFS_SOFT_DIRTY
373 This is used (in conjunction with
374 .IR /proc/[pid]/pagemap )
375 by the check-point restore system to discover which pages of a process
376 have been dirtied since the file
377 .IR /proc/[pid]/clear_refs
381 .\" Internally: CLEAR_REFS_MM_HIWATER_RSS
382 Reset the peak resident set size ("high water mark") to the process's
383 current resident set size value.
387 .IR /proc/[pid]/clear_refs
388 other than those listed above has no effect.
391 .IR /proc/[pid]/clear_refs
392 file is present only if the
393 .B CONFIG_PROC_PAGE_MONITOR
394 kernel configuration option is enabled.
396 .I /proc/[pid]/cmdline
397 This read-only file holds the complete command line for the process,
398 unless the process is a zombie.
399 .\" In 2.3.26, this also used to be true if the process was swapped out.
400 In the latter case, there is nothing in this file:
401 that is, a read on this file will return 0 characters.
402 The command-line arguments appear in this file as a set of
403 strings separated by null bytes (\(aq\\0\(aq),
404 with a further null byte after the last string.
406 .IR /proc/[pid]/comm " (since Linux 2.6.33)"
407 .\" commit 4614a696bd1c3a9af3a08f0e5874830a85b889d4
408 This file exposes the process's
410 value\(emthat is, the command name associated with the process.
411 Different threads in the same process may have different
413 values, accessible via
414 .IR /proc/[pid]/task/[tid]/comm .
415 A thread may modify its
417 value, or that of any of other thread in the same thread group (see
422 by writing to the file
423 .IR /proc/self/task/[tid]/comm .
426 (16) characters are silently truncated.
428 This file provides a superset of the
433 operations, and is employed by
434 .BR pthread_setname_np (3)
435 when used to rename threads other than the caller.
437 .IR /proc/[pid]/coredump_filter " (since Linux 2.6.23)"
441 .IR /proc/[pid]/cpuset " (since Linux 2.6.12)"
442 .\" and/proc/[pid]/task/[tid]/cpuset
447 This is a symbolic link to the current working directory of the process.
448 To find out the current working directory of process 20,
449 for instance, you can do this:
453 .RB "$" " cd /proc/20/cwd; /bin/pwd"
459 command is often a shell built-in, and might
466 .\" The following was still true as at kernel 2.6.13
467 In a multithreaded process, the contents of this symbolic link
468 are not available if the main thread has already terminated
469 (typically by calling
470 .BR pthread_exit (3)).
472 Permission to dereference or read
474 this symbolic link is governed by a ptrace access mode
475 .B PTRACE_MODE_READ_FSCREDS
479 .I /proc/[pid]/environ
480 This file contains the initial environment that was set
481 when the currently executing program was started via
483 The entries are separated by null bytes (\(aq\\0\(aq),
484 and there may be a null byte at the end.
485 Thus, to print out the environment of process 1, you would do:
489 .RB "$" " strings /proc/1/environ"
495 the process modifies its environment
496 (e.g., by calling functions such as
503 reflect those changes.
505 Furthermore, a process may change the memory location that this file refers via
508 .BR PR_SET_MM_ENV_START .
510 Permission to access this file is governed by a ptrace access mode
511 .B PTRACE_MODE_READ_FSCREDS
516 Under Linux 2.2 and later, this file is a symbolic link
517 containing the actual pathname of the executed command.
518 This symbolic link can be dereferenced normally; attempting to open
519 it will open the executable.
522 to run another copy of the same executable that is being run by
524 If the pathname has been unlinked, the symbolic link will contain the
525 string \(aq(deleted)\(aq appended to the original pathname.
526 .\" The following was still true as at kernel 2.6.13
527 In a multithreaded process, the contents of this symbolic link
528 are not available if the main thread has already terminated
529 (typically by calling
530 .BR pthread_exit (3)).
532 Permission to dereference or read
534 this symbolic link is governed by a ptrace access mode
535 .B PTRACE_MODE_READ_FSCREDS
539 Under Linux 2.0 and earlier,
541 is a pointer to the binary which was executed,
542 and appears as a symbolic link.
545 call on this file under Linux 2.0 returns a string in the format:
549 For example, [0301]:1502 would be inode 1502 on device major 03 (IDE,
550 MFM, etc. drives) minor 01 (first partition on the first drive).
555 option can be used to locate the file.
558 This is a subdirectory containing one entry for each file which the
559 process has open, named by its file descriptor, and which is a
560 symbolic link to the actual file.
561 Thus, 0 is standard input, 1 standard output, 2 standard error, and so on.
563 For file descriptors for pipes and sockets,
564 the entries will be symbolic links whose content is the
565 file type with the inode.
568 call on this file returns a string in the format:
574 will be a socket and its inode is 2248868.
575 For sockets, that inode can be used to find more information
576 in one of the files under
579 For file descriptors that have no corresponding inode
580 (e.g., file descriptors produced by
582 .BR epoll_create (2),
584 .BR inotify_init (2),
585 .BR perf_event_open (2),
587 .BR timerfd_create (2),
589 .BR userfaultfd (2)),
590 the entry will be a symbolic link with contents of the form
592 anon_inode:<file-type>
594 In many cases (but not all), the
596 is surrounded by square brackets.
598 For example, an epoll file descriptor will have a symbolic link
599 whose content is the string
600 .IR "anon_inode:[eventpoll]" .
602 .\"The following was still true as at kernel 2.6.13
603 In a multithreaded process, the contents of this directory
604 are not available if the main thread has already terminated
605 (typically by calling
606 .BR pthread_exit (3)).
608 Programs that take a filename as a command-line argument,
609 but don't take input from standard input if no argument is supplied,
610 and programs that write to a file named as a command-line argument,
611 but don't send their output to standard output
612 if no argument is supplied, can nevertheless be made to use
613 standard input or standard output by using
615 files as command-line arguments.
616 For example, assuming that
618 is the flag designating an input file and
620 is the flag designating an output file:
624 .RB "$" " foobar \-i /proc/self/fd/0 \-o /proc/self/fd/1 ..."
628 and you have a working filter.
629 .\" The following is not true in my tests (MTK):
630 .\" Note that this will not work for
631 .\" programs that seek on their files, as the files in the fd directory
632 .\" are not seekable.
635 is approximately the same as
637 in some UNIX and UNIX-like systems.
638 Most Linux MAKEDEV scripts symbolically link
644 Most systems provide symbolic links
649 which respectively link to the files
656 Thus the example command above could be written as:
660 .RB "$" " foobar \-i /dev/stdin \-o /dev/stdout ..."
664 Permission to dereference or read
666 the symbolic links in this directory is governed by a ptrace access mode
667 .B PTRACE_MODE_READ_FSCREDS
671 .IR /proc/[pid]/fdinfo/ " (since Linux 2.6.22)"
672 This is a subdirectory containing one entry for each file which the
673 process has open, named by its file descriptor.
674 The files in this directory are readable only by the owner of the process.
675 The contents of each file can be read to obtain information
676 about the corresponding file descriptor.
677 The content depends on the type of file referred to by the
678 corresponding file descriptor.
680 For regular files and directories, we see something like:
684 .RB "$" " cat /proc/12015/fdinfo/4"
691 The fields are as follows:
695 This is a decimal number showing the file offset.
698 This is an octal number that displays the
699 file access mode and file status flags (see
701 If the close-on-exec file descriptor flag is set, then
703 will also include the value
707 .\" commit 1117f72ea0217ba0cc19f05adbbd8b9a397f5ab7
708 this field incorrectly displayed the setting of
710 at the time the file was opened,
711 rather than the current setting of the close-on-exec flag.
715 This field, present since Linux 3.15,
716 .\" commit 49d063cb353265c3af701bab215ac438ca7df36d
717 is the ID of the mount point containing this file.
718 See the description of
719 .IR /proc/[pid]/mountinfo .
722 For eventfd file descriptors (see
724 we see (since Linux 3.8)
725 .\" commit cbac5542d48127b546a23d816380a7926eee1c25
726 the following fields:
738 is the current value of the eventfd counter, in hexadecimal.
740 For epoll file descriptors (see
742 we see (since Linux 3.8)
743 .\" commit 138d22b58696c506799f8de759804083ff9effae
744 the following fields:
751 tfd: 9 events: 19 data: 74253d2500000009
752 tfd: 7 events: 19 data: 74253d2500000007
756 Each of the lines beginning
758 describes one of the file descriptors being monitored via
759 the epoll file descriptor (see
764 field is the number of the file descriptor.
767 field is a hexadecimal mask of the events being monitored for this file
771 field is the data value associated with this file descriptor.
773 For signalfd file descriptors (see
775 we see (since Linux 3.8)
776 .\" commit 138d22b58696c506799f8de759804083ff9effae
777 the following fields:
784 sigmask: 0000000000000006
789 is the hexadecimal mask of signals that are accepted via this
790 signalfd file descriptor.
791 (In this example, bits 2 and 3 are set, corresponding to the signals
798 For inotify file descriptors (see
800 we see (since Linux 3.8)
801 the following fields:
808 inotify wd:2 ino:7ef82a sdev:800001 mask:800afff ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:2af87e00220ffd73
809 inotify wd:1 ino:192627 sdev:800001 mask:800afff ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:27261900802dfd73
813 Each of the lines beginning with "inotify" displays information about
814 one file or directory that is being monitored.
815 The fields in this line are as follows:
819 A watch descriptor number (in decimal).
822 The inode number of the target file (in hexadecimal).
825 The ID of the device where the target file resides (in hexadecimal).
828 The mask of events being monitored for the target file (in hexadecimal).
831 If the kernel was built with exportfs support, the path to the target
832 file is exposed as a file handle, via three hexadecimal fields:
838 For fanotify file descriptors (see
840 we see (since Linux 3.8)
841 the following fields:
848 fanotify flags:0 event-flags:88002
849 fanotify ino:19264f sdev:800001 mflags:0 mask:1 ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:4f261900a82dfd73
853 The fourth line displays information defined when the fanotify group
855 .BR fanotify_init (2):
862 .BR fanotify_init (2)
863 (expressed in hexadecimal).
869 .BR fanotify_init (2)
870 (expressed in hexadecimal).
873 Each additional line shown in the file contains information
874 about one of the marks in the fanotify group.
875 Most of these fields are as for inotify, except:
879 The flags associated with the mark
880 (expressed in hexadecimal).
883 The events mask for this mark
884 (expressed in hexadecimal).
887 The mask of events that are ignored for this mark
888 (expressed in hexadecimal).
891 For details on these fields, see
892 .BR fanotify_mark (2).
894 .IR /proc/[pid]/gid_map " (since Linux 3.5)"
896 .BR user_namespaces (7).
898 .IR /proc/[pid]/io " (since kernel 2.6.20)"
899 .\" commit 7c3ab7381e79dfc7db14a67c6f4f3285664e1ec2
900 This file contains I/O statistics for the process, for example:
904 .RB "#" " cat /proc/3828/io"
910 write_bytes: 323932160
911 cancelled_write_bytes: 0
915 The fields are as follows:
918 .IR rchar ": characters read"
919 The number of bytes which this task has caused to be read from storage.
920 This is simply the sum of bytes which this process passed to
922 and similar system calls.
923 It includes things such as terminal I/O and
924 is unaffected by whether or not actual
925 physical disk I/O was required (the read might have been satisfied from
928 .IR wchar ": characters written"
929 The number of bytes which this task has caused, or shall cause to be written
931 Similar caveats apply here as with
934 .IR syscr ": read syscalls"
935 Attempt to count the number of read I/O operations\(emthat is,
941 .IR syscw ": write syscalls"
942 Attempt to count the number of write I/O operations\(emthat is,
948 .IR read_bytes ": bytes read"
949 Attempt to count the number of bytes which this process really did cause to
950 be fetched from the storage layer.
951 This is accurate for block-backed filesystems.
953 .IR write_bytes ": bytes written"
954 Attempt to count the number of bytes which this process caused to be sent to
957 .IR cancelled_write_bytes :
958 The big inaccuracy here is truncate.
959 If a process writes 1MB to a file and then deletes the file,
960 it will in fact perform no writeout.
961 But it will have been accounted as having caused 1MB of write.
962 In other words: this field represents the number of bytes which this process
963 caused to not happen, by truncating pagecache.
964 A task can cause "negative" I/O too.
965 If this task truncates some dirty pagecache,
966 some I/O which another task has been accounted for
969 will not be happening.
973 In the current implementation, things are a bit racy on 32-bit systems:
974 if process A reads process B's
976 while process B is updating one of these 64-bit counters,
977 process A could see an intermediate result.
979 Permission to access this file is governed by a ptrace access mode
980 .B PTRACE_MODE_READ_FSCREDS
984 .IR /proc/[pid]/limits " (since Linux 2.6.24)"
985 This file displays the soft limit, hard limit, and units of measurement
986 for each of the process's resource limits (see
988 Up to and including Linux 2.6.35,
989 this file is protected to allow reading only by the real UID of the process.
991 .\" commit 3036e7b490bf7878c6dae952eec5fb87b1106589
992 this file is readable by all users on the system.
993 .\" FIXME Describe /proc/[pid]/loginuid
994 .\" Added in 2.6.11; updating requires CAP_AUDIT_CONTROL
995 .\" CONFIG_AUDITSYSCALL
997 .IR /proc/[pid]/map_files/ " (since kernel 3.3)
998 .\" commit 640708a2cff7f81e246243b0073c66e6ece7e53e
999 This subdirectory contains entries corresponding to memory-mapped
1002 Entries are named by memory region start and end
1003 address pair (expressed as hexadecimal numbers),
1004 and are symbolic links to the mapped files themselves.
1005 Here is an example, with the output wrapped and reformatted to fit on an 80-column display:
1009 .RB "#" " ls -l /proc/self/map_files/"
1010 lr\-\-\-\-\-\-\-\-. 1 root root 64 Apr 16 21:31
1011 3252e00000\-3252e20000 \-> /usr/lib64/ld-2.15.so
1016 Although these entries are present for memory regions that were
1019 flag, the way anonymous shared memory (regions created with the
1020 .B MAP_ANON | MAP_SHARED
1022 is implemented in Linux
1023 means that such regions also appear on this directory.
1024 Here is an example where the target file is the deleted
1030 lrw\-\-\-\-\-\-\-. 1 root root 64 Apr 16 21:33
1031 7fc075d2f000\-7fc075e6f000 \-> /dev/zero (deleted)
1035 This directory appears only if the
1036 .B CONFIG_CHECKPOINT_RESTORE
1037 kernel configuration option is enabled.
1039 .RB ( CAP_SYS_ADMIN )
1041 .\" This may change. See the mail thread
1042 .\" "[RFC][PATCH v2] procfs: Always expose /proc/<pid>/map_files/ and make it readable"
1044 is required to view the contents of this directory.
1047 A file containing the currently mapped memory regions and their access
1051 for some further information about memory mappings.
1053 Permission to access this file is governed by a ptrace access mode
1054 .B PTRACE_MODE_READ_FSCREDS
1058 The format of the file is:
1064 .I "address perms offset dev inode pathname"
1065 00400000-00452000 r-xp 00000000 08:02 173521 /usr/bin/dbus-daemon
1066 00651000-00652000 r--p 00051000 08:02 173521 /usr/bin/dbus-daemon
1067 00652000-00655000 rw-p 00052000 08:02 173521 /usr/bin/dbus-daemon
1068 00e03000-00e24000 rw-p 00000000 00:00 0 [heap]
1069 00e24000-011f7000 rw-p 00000000 00:00 0 [heap]
1071 35b1800000-35b1820000 r-xp 00000000 08:02 135522 /usr/lib64/ld-2.15.so
1072 35b1a1f000-35b1a20000 r--p 0001f000 08:02 135522 /usr/lib64/ld-2.15.so
1073 35b1a20000-35b1a21000 rw-p 00020000 08:02 135522 /usr/lib64/ld-2.15.so
1074 35b1a21000-35b1a22000 rw-p 00000000 00:00 0
1075 35b1c00000-35b1dac000 r-xp 00000000 08:02 135870 /usr/lib64/libc-2.15.so
1076 35b1dac000-35b1fac000 ---p 001ac000 08:02 135870 /usr/lib64/libc-2.15.so
1077 35b1fac000-35b1fb0000 r--p 001ac000 08:02 135870 /usr/lib64/libc-2.15.so
1078 35b1fb0000-35b1fb2000 rw-p 001b0000 08:02 135870 /usr/lib64/libc-2.15.so
1080 f2c6ff8c000-7f2c7078c000 rw-p 00000000 00:00 0 [stack:986]
1082 7fffb2c0d000-7fffb2c2e000 rw-p 00000000 00:00 0 [stack]
1083 7fffb2d48000-7fffb2d49000 r-xp 00000000 00:00 0 [vdso]
1089 field is the address space in the process that the mapping occupies.
1092 field is a set of permissions:
1100 p = private (copy on write)
1106 field is the offset into the file/whatever;
1111 is the inode on that device.
1112 0 indicates that no inode is associated with the memory region,
1113 as would be the case with BSS (uninitialized data).
1117 field will usually be the file that is backing the mapping.
1119 you can easily coordinate with the
1121 field by looking at the
1122 Offset field in the ELF program headers
1123 .RI ( "readelf\ \-l" ).
1125 There are additional helpful pseudo-paths:
1129 The initial process's (also known as the main thread's) stack.
1131 .IR [stack:<tid>] " (since Linux 3.4)"
1132 .\" commit b76437579d1344b612cf1851ae610c636cec7db0
1133 A thread's stack (where the
1136 It corresponds to the
1137 .IR /proc/[pid]/task/[tid]/
1141 The virtual dynamically linked shared object.
1153 this is an anonymous mapping as obtained via
1155 There is no easy way to coordinate this back to a process's source,
1156 short of running it through
1161 Under Linux 2.0, there is no field giving pathname.
1164 This file can be used to access the pages of a process's memory through
1170 Permission to access this file is governed by a ptrace access mode
1171 .B PTRACE_MODE_ATTACH_FSCREDS
1175 .IR /proc/[pid]/mountinfo " (since Linux 2.6.26)"
1176 .\" This info adapted from Documentation/filesystems/proc.txt
1177 .\" commit 2d4d4864ac08caff5c204a752bd004eed4f08760
1178 This file contains information about mount points
1179 in the process's mount namespace (see
1180 .BR mount_namespaces (7)).
1181 It supplies various information
1182 (e.g., propagation state, root of mount for bind mounts,
1183 identifier for each mount and its parent) that is missing from the (older)
1184 .IR /proc/[pid]/mounts
1185 file, and fixes various other problems with that file
1186 (e.g., nonextensibility,
1187 failure to distinguish per-mount versus per-superblock options).
1189 The file contains lines of the form:
1193 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 \- ext3 /dev/root rw,errors=continue
1194 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
1198 The numbers in parentheses are labels for the descriptions below:
1202 mount ID: a unique ID for the mount (may be reused after
1206 parent ID: the ID of the parent mount
1207 (or of self for the top of the mount tree).
1210 major:minor: the value of
1212 for files on this filesystem (see
1216 root: the pathname of the directory in the filesystem
1217 which forms the root of this mount.
1220 mount point: the pathname of the mount point relative
1221 to the process's root directory.
1224 mount options: per-mount options.
1227 optional fields: zero or more fields of the form "tag[:value]"; see below.
1230 separator: the end of the optional fields is marked by a single hyphen.
1233 filesystem type: the filesystem type in the form "type[.subtype]".
1236 mount source: filesystem-specific information or "none".
1239 super options: per-superblock options.
1242 Currently, the possible optional fields are
1245 .IR propagate_from ,
1249 .BR mount_namespaces (7)
1250 for a description of these fields.
1251 Parsers should ignore all unrecognized optional fields.
1253 For more information on mount propagation see:
1254 .I Documentation/filesystems/sharedsubtree.txt
1255 in the Linux kernel source tree.
1257 .IR /proc/[pid]/mounts " (since Linux 2.4.19)"
1258 This file lists all the filesystems currently mounted in the
1259 process's mount namespace (see
1260 .BR mount_namespaces (7)).
1261 The format of this file is documented in
1264 Since kernel version 2.6.15, this file is pollable:
1265 after opening the file for reading, a change in this file
1266 (i.e., a filesystem mount or unmount) causes
1268 to mark the file descriptor as having an exceptional condition, and
1272 mark the file as having a priority event
1274 (Before Linux 2.6.30,
1275 a change in this file was indicated by the file descriptor
1276 being marked as readable for
1278 and being marked as having an error condition for
1281 .BR epoll_wait (2).)
1283 .IR /proc/[pid]/mountstats " (since Linux 2.6.17)"
1284 This file exports information (statistics, configuration information)
1285 about the mount points in the process's mount namespace (see
1286 .BR mount_namespaces (7)).
1287 Lines in this file have the form:
1290 device /dev/sda7 mounted on /home with fstype ext3 [statistics]
1291 ( 1 ) ( 2 ) (3 ) (4)
1294 The fields in each line are:
1298 The name of the mounted device
1299 (or "nodevice" if there is no corresponding device).
1302 The mount point within the filesystem tree.
1305 The filesystem type.
1308 Optional statistics and configuration information.
1309 Currently (as at Linux 2.6.26), only NFS filesystems export
1310 information via this field.
1313 This file is readable only by the owner of the process.
1316 /proc/[pid]/net " (since Linux 2.6.25)"
1317 See the description of
1320 .IR /proc/[pid]/ns/ " (since Linux 3.0)"
1321 .\" See commit 6b4e306aa3dc94a0545eb9279475b1ab6209a31f
1322 This is a subdirectory containing one entry for each namespace that
1323 supports being manipulated by
1325 For more information, see
1328 .IR /proc/[pid]/numa_maps " (since Linux 2.6.14)"
1332 .IR /proc/[pid]/oom_adj " (since Linux 2.6.11)"
1333 This file can be used to adjust the score used to select which process
1334 should be killed in an out-of-memory (OOM) situation.
1335 The kernel uses this value for a bit-shift operation of the process's
1338 valid values are in the range \-16 to +15,
1339 plus the special value \-17,
1340 which disables OOM-killing altogether for this process.
1341 A positive score increases the likelihood of this
1342 process being killed by the OOM-killer;
1343 a negative score decreases the likelihood.
1345 The default value for this file is 0;
1346 a new process inherits its parent's
1349 A process must be privileged
1350 .RB ( CAP_SYS_RESOURCE )
1351 to update this file.
1353 Since Linux 2.6.36, use of this file is deprecated in favor of
1354 .IR /proc/[pid]/oom_score_adj .
1356 .IR /proc/[pid]/oom_score " (since Linux 2.6.11)"
1357 .\" See mm/oom_kill.c::badness() in the 2.6.25 sources
1358 This file displays the current score that the kernel gives to
1359 this process for the purpose of selecting a process
1361 A higher score means that the process is more likely to be
1362 selected by the OOM-killer.
1363 The basis for this score is the amount of memory used by the process,
1364 with increases (+) or decreases (\-) for factors including:
1365 .\" See mm/oom_kill.c::badness() in the 2.6.25 sources
1368 whether the process creates a lot of children using
1372 whether the process has been running a long time,
1373 or has used a lot of CPU time (\-);
1375 whether the process has a low nice value (i.e., > 0) (+);
1377 whether the process is privileged (\-); and
1378 .\" More precisely, if it has CAP_SYS_ADMIN or CAP_SYS_RESOURCE
1380 whether the process is making direct hardware access (\-).
1381 .\" More precisely, if it has CAP_SYS_RAWIO
1386 also reflects the adjustment specified by the
1390 setting for the process.
1392 .IR /proc/[pid]/oom_score_adj " (since Linux 2.6.36)"
1393 .\" Text taken from 3.7 Documentation/filesystems/proc.txt
1394 This file can be used to adjust the badness heuristic used to select which
1395 process gets killed in out-of-memory conditions.
1397 The badness heuristic assigns a value to each candidate task ranging from 0
1398 (never kill) to 1000 (always kill) to determine which process is targeted.
1399 The units are roughly a proportion along that range of
1400 allowed memory the process may allocate from,
1401 based on an estimation of its current memory and swap use.
1402 For example, if a task is using all allowed memory,
1403 its badness score will be 1000.
1404 If it is using half of its allowed memory, its score will be 500.
1406 There is an additional factor included in the badness score: root
1407 processes are given 3% extra memory over other tasks.
1409 The amount of "allowed" memory depends on the context
1410 in which the OOM-killer was called.
1411 If it is due to the memory assigned to the allocating task's cpuset
1413 the allowed memory represents the set of mems assigned to that
1416 If it is due to a mempolicy's node(s) being exhausted,
1417 the allowed memory represents the set of mempolicy nodes.
1418 If it is due to a memory limit (or swap limit) being reached,
1419 the allowed memory is that configured limit.
1420 Finally, if it is due to the entire system being out of memory, the
1421 allowed memory represents all allocatable resources.
1425 is added to the badness score before it
1426 is used to determine which task to kill.
1427 Acceptable values range from \-1000
1428 (OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX).
1429 This allows user space to control the preference for OOM-killing,
1430 ranging from always preferring a certain
1431 task or completely disabling it from OOM killing.
1432 The lowest possible value, \-1000, is
1433 equivalent to disabling OOM-killing entirely for that task,
1434 since it will always report a badness score of 0.
1436 Consequently, it is very simple for user space to define
1437 the amount of memory to consider for each task.
1440 value of +500, for example,
1441 is roughly equivalent to allowing the remainder of tasks sharing the
1442 same system, cpuset, mempolicy, or memory controller resources
1443 to use at least 50% more memory.
1444 A value of \-500, on the other hand, would be roughly
1445 equivalent to discounting 50% of the task's
1446 allowed memory from being considered as scoring against the task.
1448 For backward compatibility with previous kernels,
1449 .I /proc/[pid]/oom_adj
1450 can still be used to tune the badness score.
1452 scaled linearly with
1456 .IR /proc/[pid]/oom_score_adj
1458 .IR /proc/[pid]/oom_adj
1459 will change the other with its scaled value.
1461 .IR /proc/[pid]/pagemap " (since Linux 2.6.25)"
1462 This file shows the mapping of each of the process's virtual pages
1463 into physical page frames or swap area.
1464 It contains one 64-bit value for each virtual page,
1465 with the bits set as follows:
1469 If set, the page is present in RAM.
1472 If set, the page is in swap space
1474 61 (since Linux 3.5)
1475 The page is a file-mapped page or a shared anonymous page.
1477 60-56 (since Linux 3.11)
1479 .\" Not quite true; see commit 541c237c0923f567c9c4cabb8a81635baadc713f
1481 55 (since Linux 3.11)
1483 (see the kernel source file
1484 .IR Documentation/vm/soft-dirty.txt ).
1487 If the page is present in RAM (bit 63), then these bits
1488 provide the page frame number, which can be used to index
1489 .IR /proc/kpageflags
1491 .IR /proc/kpagecount .
1492 If the page is present in swap (bit 62),
1493 then bits 4-0 give the swap type, and bits 54-5 encode the swap offset.
1496 Before Linux 3.11, bits 60-55 were
1497 used to encode the base-2 log of the page size.
1500 .IR /proc/[pid]/pagemap
1502 .IR /proc/[pid]/maps
1503 to determine which areas of memory are actually mapped and seek
1504 to skip over unmapped regions.
1507 .IR /proc/[pid]/pagemap
1508 file is present only if the
1509 .B CONFIG_PROC_PAGE_MONITOR
1510 kernel configuration option is enabled.
1512 Permission to access this file is governed by a ptrace access mode
1513 .B PTRACE_MODE_READ_FSCREDS
1517 .IR /proc/[pid]/personality " (since Linux 2.6.28)"
1518 .\" commit 478307230810d7e2a753ed220db9066dfdf88718
1519 This read-only file exposes the process's execution domain, as set by
1520 .BR personality (2).
1521 The value is displayed in hexadecimal notation.
1523 Permission to access this file is governed by a ptrace access mode
1524 .B PTRACE_MODE_ATTACH_FSCREDS
1529 UNIX and Linux support the idea of a per-process root of the
1530 filesystem, set by the
1533 This file is a symbolic link that points to the process's
1534 root directory, and behaves in the same way as
1539 Note however that this file is not merely a symbolic link.
1540 It provides the same view of the filesystem (including namespaces and the
1541 set of per-process mounts) as the process itself.
1542 An example illustrates this point.
1543 In one terminal, we start a shell in new user and mount namespaces,
1544 and in that shell we create some new mount points:
1548 $ \fBPS1='sh1# ' unshare \-Urnm\fP
1549 sh1# \fBmount \-t tmpfs tmpfs /etc\fP # Mount empty tmpfs at /etc
1550 sh1# \fBmount \-\-bind /usr /dev\fP # Mount /usr at /dev
1556 In a second terminal window, in the initial mount namespace,
1557 we look at the contents of the corresponding mounts in
1558 the initial and new namespaces:
1562 $ \fBPS1='sh2# ' sudo sh\fP
1563 sh2# \fBls /etc | wc \-l\fP # In initial NS
1565 sh2# \fBls /proc/27123/root/etc | wc \-l\fP # /etc in other NS
1566 0 # The empty tmpfs dir
1567 sh2# \fBls /dev | wc \-l\fP # In initial NS
1569 sh2# \fBls /proc/27123/root/dev | wc \-l\fP # /dev in other NS
1572 sh2# \fBls /usr | wc \-l\fP # /usr in initial NS
1577 .\" The following was still true as at kernel 2.6.13
1578 In a multithreaded process, the contents of the
1580 symbolic link are not available if the main thread has already terminated
1581 (typically by calling
1582 .BR pthread_exit (3)).
1584 Permission to dereference or read
1586 this symbolic link is governed by a ptrace access mode
1587 .B PTRACE_MODE_READ_FSCREDS
1590 .\" FIXME Describe /proc/[pid]/projid_map
1592 .\" commit f76d207a66c3a53defea67e7d36c3eb1b7d6d61d
1594 .IR /proc/[pid]/seccomp " (Linux 2.6.12 to 2.6.22)"
1595 This file can be used to read and change the process's
1596 secure computing (seccomp) mode setting.
1597 It contains the value 0 if the process is not in seccomp mode,
1598 and 1 if the process is in strict seccomp mode (see
1600 Writing 1 to this file places the process irreversibly in strict seccomp mode.
1601 (Further attempts to write to the file fail with the
1606 this file went away, to be replaced by the
1611 operations (and later by
1616 .IR /proc/[pid]/status ).
1617 .\" FIXME Describe /proc/[pid]/sessionid
1618 .\" commit 1e0bd7550ea9cf474b1ad4c6ff5729a507f75fdc
1619 .\" CONFIG_AUDITSYSCALL
1620 .\" Added in 2.6.25; read-only; only readable by real UID
1622 .\" FIXME Describe /proc/[pid]/sched
1624 .\" CONFIG_SCHED_DEBUG, and additional fields if CONFIG_SCHEDSTATS
1625 .\" Displays various scheduling parameters
1626 .\" This file can be written, to reset stats
1627 .\" The set of fields exposed by this file have changed
1628 .\" significantly over time.
1629 .\" commit 43ae34cb4cd650d1eb4460a8253a8e747ba052ac
1631 .\" FIXME Describe /proc/[pid]/schedstats and
1632 .\" /proc/[pid]/task/[tid]/schedstats
1634 .\" CONFIG_SCHEDSTATS
1636 .IR /proc/[pid]/setgroups " (since Linux 3.19)"
1638 .BR user_namespaces (7).
1640 .IR /proc/[pid]/smaps " (since Linux 2.6.14)"
1641 This file shows memory consumption for each of the process's mappings.
1644 command displays similar information,
1645 in a form that may be easier for parsing.)
1646 For each mapping there is a series of lines such as the following:
1650 00400000-0048a000 r-xp 00000000 fd:03 960637 /bin/bash
1654 Shared_Clean: 452 kB
1661 ShmemHugePages: 0 kB
1662 ShmemPmdMapped: 0 kB
1664 KernelPageSize: 4 kB
1666 KernelPageSize: 4 kB
1670 VmFlags: rd ex mr mw me dw
1674 The first of these lines shows the same information as is displayed
1676 .IR /proc/[pid]/maps .
1677 The following lines show the size of the mapping,
1678 the amount of the mapping that is currently resident in RAM ("Rss"),
1679 the process's proportional share of this mapping ("Pss"),
1680 the number of clean and dirty shared pages in the mapping,
1681 and the number of clean and dirty private pages in the mapping.
1682 "Referenced" indicates the amount of memory currently marked as
1683 referenced or accessed.
1684 "Anonymous" shows the amount of memory
1685 that does not belong to any file.
1686 "Swap" shows how much
1687 would-be-anonymous memory is also used, but out on swap.
1689 The "KernelPageSize" line (available since Linux 2.6.29)
1690 is the page size used by the kernel to back the virtual memory area.
1691 This matches the size used by the MMU in the majority of cases.
1692 However, one counter-example occurs on PPC64 kernels
1693 whereby a kernel using 64kB as a base page size may still use 4kB
1694 pages for the MMU on older processors.
1695 To distinguish the two attributes, the "MMUPageSize" line
1696 (also available since Linux 2.6.29)
1697 reports the page size used by the MMU.
1699 The "Locked" indicates whether the mapping is locked in memory
1702 The "ProtectionKey" line (available since Linux 4.9, on x86 only)
1703 contains the memory protection key (see
1705 associated with the virtual memory area.
1706 This entry is present only if the kernel was built with the
1707 .B CONFIG_X86_INTEL_MEMORY_PROTECTION_KEYS
1708 configuration option.
1710 The "VmFlags" line (available since Linux 3.8)
1711 represents the kernel flags associated with the virtual memory area,
1712 encoded using the following two-letter codes:
1722 gd - stack segment grows down
1724 dw - disabled write to the mapped file
1725 lo - pages are locked in memory
1726 io - memory mapped I/O area
1727 sr - sequential read advise provided
1728 rr - random read advise provided
1729 dc - do not copy area on fork
1730 de - do not expand area on remapping
1731 ac - area is accountable
1732 nr - swap space is not reserved for the area
1733 ht - area uses huge tlb pages
1734 nl - non-linear mapping
1735 ar - architecture specific flag
1736 dd - do not include area into core dump
1737 sd - soft-dirty flag
1739 hg - huge page advise flag
1740 nh - no-huge page advise flag
1741 mg - mergeable advise flag
1743 "ProtectionKey" field contains the memory protection key (see
1745 associated with the virtual memory area.
1746 Present only if the kernel was built with the
1747 .B CONFIG_X86_INTEL_MEMORY_PROTECTION_KEYS
1748 configuration option. (since Linux 4.6)
1751 .IR /proc/[pid]/smaps
1752 file is present only if the
1753 .B CONFIG_PROC_PAGE_MONITOR
1754 kernel configuration option is enabled.
1756 .IR /proc/[pid]/stack " (since Linux 2.6.29)"
1757 .\" 2ec220e27f5040aec1e88901c1b6ea3d135787ad
1758 This file provides a symbolic trace of the function calls in this
1759 process's kernel stack.
1760 This file is provided only if the kernel was built with the
1761 .B CONFIG_STACKTRACE
1762 configuration option.
1764 Permission to access this file is governed by a ptrace access mode
1765 .B PTRACE_MODE_ATTACH_FSCREDS
1770 Status information about the process.
1773 It is defined in the kernel source file
1774 .IR fs/proc/array.c "."
1776 The fields, in order, with their proper
1778 format specifiers, are listed below.
1779 Whether or not certain of these fields display valid information is governed by
1780 a ptrace access mode
1781 .BR PTRACE_MODE_READ_FSCREDS "\ |\ " PTRACE_MODE_NOAUDIT
1784 If the check denies access, then the field value is displayed as 0.
1785 The affected fields are indicated with the marking [PT].
1794 The filename of the executable, in parentheses.
1795 This is visible whether or not the executable is swapped out.
1797 (3) \fIstate\fP \ %c
1798 One of the following characters, indicating process state:
1803 Sleeping in an interruptible wait
1805 Waiting in uninterruptible
1810 Stopped (on a signal) or (before Linux 2.6.33) trace stopped
1812 .\" commit 44d90df6b757c59651ddd55f1a84f28132b50d29
1813 Tracing stop (Linux 2.6.33 onward)
1815 Paging (only before Linux 2.6.0)
1817 Dead (from Linux 2.6.0 onward)
1819 .\" commit 44d90df6b757c59651ddd55f1a84f28132b50d29
1820 Dead (Linux 2.6.33 to
1821 .\" commit 74e37200de8e9c4e09b70c21c3f13c2071e77457
1824 .\" commit 44d90df6b757c59651ddd55f1a84f28132b50d29
1825 Wakekill (Linux 2.6.33 to
1826 .\" commit 74e37200de8e9c4e09b70c21c3f13c2071e77457
1829 .\" commit 44d90df6b757c59651ddd55f1a84f28132b50d29
1830 Waking (Linux 2.6.33 to
1831 .\" commit 74e37200de8e9c4e09b70c21c3f13c2071e77457
1834 .\" commit f2530dc71cf0822f90bb63ea4600caaef33a66bb
1835 Parked (Linux 3.9 to
1836 .\" commit 74e37200de8e9c4e09b70c21c3f13c2071e77457
1841 The PID of the parent of this process.
1844 The process group ID of the process.
1846 (6) \fIsession\fP \ %d
1847 The session ID of the process.
1849 (7) \fItty_nr\fP \ %d
1850 The controlling terminal of the process.
1851 (The minor device number is contained in the combination of bits
1852 31 to 20 and 7 to 0;
1853 the major device number is in bits 15 to 8.)
1855 (8) \fItpgid\fP \ %d
1856 .\" This field and following, up to and including wchan added 0.99.1
1857 The ID of the foreground process group of the controlling
1858 terminal of the process.
1860 (9) \fIflags\fP \ %u
1861 The kernel flags word of the process.
1863 see the PF_* defines in the Linux kernel source file
1864 .IR include/linux/sched.h .
1865 Details depend on the kernel version.
1867 The format for this field was %lu before Linux 2.6.
1869 (10) \fIminflt\fP \ %lu
1870 The number of minor faults the process has made which have not
1871 required loading a memory page from disk.
1873 (11) \fIcminflt\fP \ %lu
1874 The number of minor faults that the process's
1875 waited-for children have made.
1877 (12) \fImajflt\fP \ %lu
1878 The number of major faults the process has made which have
1879 required loading a memory page from disk.
1881 (13) \fIcmajflt\fP \ %lu
1882 The number of major faults that the process's
1883 waited-for children have made.
1885 (14) \fIutime\fP \ %lu
1886 Amount of time that this process has been scheduled in user mode,
1887 measured in clock ticks (divide by
1888 .IR sysconf(_SC_CLK_TCK) ).
1889 This includes guest time, \fIguest_time\fP
1890 (time spent running a virtual CPU, see below),
1891 so that applications that are not aware of the guest time field
1892 do not lose that time from their calculations.
1894 (15) \fIstime\fP \ %lu
1895 Amount of time that this process has been scheduled in kernel mode,
1896 measured in clock ticks (divide by
1897 .IR sysconf(_SC_CLK_TCK) ).
1899 (16) \fIcutime\fP \ %ld
1900 Amount of time that this process's
1901 waited-for children have been scheduled in user mode,
1902 measured in clock ticks (divide by
1903 .IR sysconf(_SC_CLK_TCK) ).
1906 This includes guest time, \fIcguest_time\fP
1907 (time spent running a virtual CPU, see below).
1909 (17) \fIcstime\fP \ %ld
1910 Amount of time that this process's
1911 waited-for children have been scheduled in kernel mode,
1912 measured in clock ticks (divide by
1913 .IR sysconf(_SC_CLK_TCK) ).
1915 (18) \fIpriority\fP \ %ld
1916 (Explanation for Linux 2.6)
1917 For processes running a real-time scheduling policy
1920 .BR sched_setscheduler (2)),
1921 this is the negated scheduling priority, minus one;
1922 that is, a number in the range \-2 to \-100,
1923 corresponding to real-time priorities 1 to 99.
1924 For processes running under a non-real-time scheduling policy,
1925 this is the raw nice value
1926 .RB ( setpriority (2))
1927 as represented in the kernel.
1928 The kernel stores nice values as numbers
1929 in the range 0 (high) to 39 (low),
1930 corresponding to the user-visible nice range of \-20 to 19.
1932 Before Linux 2.6, this was a scaled value based on
1933 the scheduler weighting given to this process.
1934 .\" And back in kernel 1.2 days things were different again.
1936 (19) \fInice\fP \ %ld
1938 .BR setpriority (2)),
1939 a value in the range 19 (low priority) to \-20 (high priority).
1940 .\" Back in kernel 1.2 days things were different.
1942 .\" \fIcounter\fP %ld
1943 .\" The current maximum size in jiffies of the process's next timeslice,
1944 .\" or what is currently left of its current timeslice, if it is the
1945 .\" currently running process.
1947 .\" \fItimeout\fP %u
1948 .\" The time in jiffies of the process's next timeout.
1949 .\" timeout was removed sometime around 2.1/2.2
1951 (20) \fInum_threads\fP \ %ld
1952 Number of threads in this process (since Linux 2.6).
1953 Before kernel 2.6, this field was hard coded to 0 as a placeholder
1954 for an earlier removed field.
1956 (21) \fIitrealvalue\fP \ %ld
1957 The time in jiffies before the next
1959 is sent to the process due to an interval timer.
1960 Since kernel 2.6.17, this field is no longer maintained,
1961 and is hard coded as 0.
1963 (22) \fIstarttime\fP \ %llu
1964 The time the process started after system boot.
1965 In kernels before Linux 2.6, this value was expressed in jiffies.
1966 Since Linux 2.6, the value is expressed in clock ticks (divide by
1967 .IR sysconf(_SC_CLK_TCK) ).
1969 The format for this field was %lu before Linux 2.6.
1971 (23) \fIvsize\fP \ %lu
1972 Virtual memory size in bytes.
1974 (24) \fIrss\fP \ %ld
1975 Resident Set Size: number of pages the process has in real memory.
1976 This is just the pages which
1977 count toward text, data, or stack space.
1978 This does not include pages
1979 which have not been demand-loaded in, or which are swapped out.
1981 (25) \fIrsslim\fP \ %lu
1982 Current soft limit in bytes on the rss of the process;
1983 see the description of
1988 (26) \fIstartcode\fP \ %lu \ [PT]
1989 The address above which program text can run.
1991 (27) \fIendcode\fP \ %lu \ [PT]
1992 The address below which program text can run.
1994 (28) \fIstartstack\fP \ %lu \ [PT]
1995 The address of the start (i.e., bottom) of the stack.
1997 (29) \fIkstkesp\fP \ %lu \ [PT]
1998 The current value of ESP (stack pointer), as found in the
1999 kernel stack page for the process.
2001 (30) \fIkstkeip\fP \ %lu \ [PT]
2002 The current EIP (instruction pointer).
2004 (31) \fIsignal\fP \ %lu
2005 The bitmap of pending signals, displayed as a decimal number.
2006 Obsolete, because it does not provide information on real-time signals; use
2007 .I /proc/[pid]/status
2010 (32) \fIblocked\fP \ %lu
2011 The bitmap of blocked signals, displayed as a decimal number.
2012 Obsolete, because it does not provide information on real-time signals; use
2013 .I /proc/[pid]/status
2016 (33) \fIsigignore\fP \ %lu
2017 The bitmap of ignored signals, displayed as a decimal number.
2018 Obsolete, because it does not provide information on real-time signals; use
2019 .I /proc/[pid]/status
2022 (34) \fIsigcatch\fP \ %lu
2023 The bitmap of caught signals, displayed as a decimal number.
2024 Obsolete, because it does not provide information on real-time signals; use
2025 .I /proc/[pid]/status
2028 (35) \fIwchan\fP \ %lu \ [PT]
2029 This is the "channel" in which the process is waiting.
2030 It is the address of a location in the kernel where the process is sleeping.
2031 The corresponding symbolic name can be found in
2032 .IR /proc/[pid]/wchan .
2034 (36) \fInswap\fP \ %lu
2035 .\" nswap was added in 2.0
2036 Number of pages swapped (not maintained).
2038 (37) \fIcnswap\fP \ %lu
2039 .\" cnswap was added in 2.0
2040 Cumulative \fInswap\fP for child processes (not maintained).
2042 (38) \fIexit_signal\fP \ %d \ (since Linux 2.1.22)
2043 Signal to be sent to parent when we die.
2045 (39) \fIprocessor\fP \ %d \ (since Linux 2.2.8)
2046 CPU number last executed on.
2048 (40) \fIrt_priority\fP \ %u \ (since Linux 2.5.19)
2049 Real-time scheduling priority, a number in the range 1 to 99 for
2050 processes scheduled under a real-time policy,
2051 or 0, for non-real-time processes (see
2052 .BR sched_setscheduler (2)).
2054 (41) \fIpolicy\fP \ %u \ (since Linux 2.5.19)
2055 Scheduling policy (see
2056 .BR sched_setscheduler (2)).
2057 Decode using the SCHED_* constants in
2060 The format for this field was %lu before Linux 2.6.22.
2062 (42) \fIdelayacct_blkio_ticks\fP \ %llu \ (since Linux 2.6.18)
2063 Aggregated block I/O delays, measured in clock ticks (centiseconds).
2065 (43) \fIguest_time\fP \ %lu \ (since Linux 2.6.24)
2066 Guest time of the process (time spent running a virtual CPU
2067 for a guest operating system), measured in clock ticks (divide by
2068 .IR sysconf(_SC_CLK_TCK) ).
2070 (44) \fIcguest_time\fP \ %ld \ (since Linux 2.6.24)
2071 Guest time of the process's children, measured in clock ticks (divide by
2072 .IR sysconf(_SC_CLK_TCK) ).
2074 (45) \fIstart_data\fP \ %lu \ (since Linux 3.3) \ [PT]
2075 .\" commit b3f7f573a20081910e34e99cbc91831f4f02f1ff
2076 Address above which program initialized and
2077 uninitialized (BSS) data are placed.
2079 (46) \fIend_data\fP \ %lu \ (since Linux 3.3) \ [PT]
2080 .\" commit b3f7f573a20081910e34e99cbc91831f4f02f1ff
2081 Address below which program initialized and
2082 uninitialized (BSS) data are placed.
2084 (47) \fIstart_brk\fP \ %lu \ (since Linux 3.3) \ [PT]
2085 .\" commit b3f7f573a20081910e34e99cbc91831f4f02f1ff
2086 Address above which program heap can be expanded with
2089 (48) \fIarg_start\fP \ %lu \ (since Linux 3.5) \ [PT]
2090 .\" commit 5b172087f99189416d5f47fd7ab5e6fb762a9ba3
2091 Address above which program command-line arguments
2095 (49) \fIarg_end\fP \ %lu \ (since Linux 3.5) \ [PT]
2096 .\" commit 5b172087f99189416d5f47fd7ab5e6fb762a9ba3
2097 Address below program command-line arguments
2101 (50) \fIenv_start\fP \ %lu \ (since Linux 3.5) \ [PT]
2102 .\" commit 5b172087f99189416d5f47fd7ab5e6fb762a9ba3
2103 Address above which program environment is placed.
2105 (51) \fIenv_end\fP \ %lu \ (since Linux 3.5) \ [PT]
2106 .\" commit 5b172087f99189416d5f47fd7ab5e6fb762a9ba3
2107 Address below which program environment is placed.
2109 (52) \fIexit_code\fP \ %d \ (since Linux 3.5) \ [PT]
2110 .\" commit 5b172087f99189416d5f47fd7ab5e6fb762a9ba3
2111 The thread's exit status in the form reported by
2115 .I /proc/[pid]/statm
2116 Provides information about memory usage, measured in pages.
2121 size (1) total program size
2122 (same as VmSize in \fI/proc/[pid]/status\fP)
2123 resident (2) resident set size
2124 (same as VmRSS in \fI/proc/[pid]/status\fP)
2125 shared (3) number of resident shared pages (i.e., backed by a file)
2126 (same as RssFile+RssShmem in \fI/proc/[pid]/status\fP)
2127 text (4) text (code)
2128 .\" (not including libs; broken, includes data segment)
2129 lib (5) library (unused since Linux 2.6; always 0)
2130 data (6) data + stack
2131 .\" (including libs; broken, includes library text)
2132 dt (7) dirty pages (unused since Linux 2.6; always 0)
2136 .I /proc/[pid]/status
2137 Provides much of the information in
2140 .I /proc/[pid]/statm
2141 in a format that's easier for humans to parse.
2146 .RB "$" " cat /proc/$$/status"
2155 Uid: 1000 1000 1000 1000
2156 Gid: 100 100 100 100
2179 HugetlbPages: 0 kB # 4.4
2182 SigPnd: 0000000000000000
2183 ShdPnd: 0000000000000000
2184 SigBlk: 0000000000010000
2185 SigIgn: 0000000000384004
2186 SigCgt: 000000004b813efb
2187 CapInh: 0000000000000000
2188 CapPrm: 0000000000000000
2189 CapEff: 0000000000000000
2190 CapBnd: ffffffffffffffff
2191 CapAmb: 0000000000000000
2194 Cpus_allowed: 00000001
2195 Cpus_allowed_list: 0
2197 Mems_allowed_list: 0
2198 voluntary_ctxt_switches: 150
2199 nonvoluntary_ctxt_switches: 545
2203 The fields are as follows:
2207 Command run by this process.
2210 Process umask, expressed in octal with a leading zero; see
2215 Current state of the process.
2227 Thread group ID (i.e., Process ID).
2230 NUMA group ID (0 if none; since Linux 3.13).
2237 PID of parent process.
2240 PID of process tracing this process (0 if not being traced).
2243 Real, effective, saved set, and filesystem UIDs (GIDs).
2246 Number of file descriptor slots currently allocated.
2249 Supplementary group list.
2252 Thread group ID (i.e., PID) in each of the PID namespaces of which
2255 The leftmost entry shows the value with respect to the PID namespace
2256 of the reading process,
2257 followed by the value in successively nested inner namespaces.
2258 .\" commit e4bc33245124db69b74a6d853ac76c2976f472d5
2262 Thread ID in each of the PID namespaces of which
2265 The fields are ordered as for
2270 Process group ID in each of the PID namespaces of which
2273 The fields are ordered as for
2278 descendant namespace session ID hierarchy
2279 Session ID in each of the PID namespaces of which
2282 The fields are ordered as for
2287 Peak virtual memory size.
2290 Virtual memory size.
2293 Locked memory size (see
2298 .\" commit bc3e53f682d93df677dbd5006a404722b3adfe18
2300 These are pages that can't be moved because something needs to
2301 directly access physical memory.
2304 Peak resident set size ("high water mark").
2308 Note that the value here is the sum of
2315 Size of resident anonymous memory.
2316 .\" commit bf9683d6990589390b5178dafe8fd06808869293
2320 Size of resident file mappings.
2321 .\" commit bf9683d6990589390b5178dafe8fd06808869293
2325 Size of resident shared memory (includes System V shared memory,
2328 and shared anonymous mappings).
2329 .\" commit bf9683d6990589390b5178dafe8fd06808869293
2332 .IR VmData ", " VmStk ", " VmExe :
2333 Size of data, stack, and text segments.
2336 Shared library code size.
2339 Page table entries size (since Linux 2.6.10).
2342 .\" commit dc6c9a35b66b520cf67e05d8ca60ebecad3b0479
2343 Size of second-level page tables (since Linux 4.0).
2346 .\" commit b084d4353ff99d824d3bc5a5c2c22c70b1fba722
2347 Swapped-out virtual memory size by anonymous private pages;
2348 shmem swap usage is not included (since Linux 2.6.34).
2351 Size of hugetlb memory portions.
2352 .\" commit 5d317b2b6536592a9b51fe65faed43d65ca9158e
2356 Number of threads in process containing this thread.
2359 This field contains two slash-separated numbers that relate to
2360 queued signals for the real user ID of this process.
2361 The first of these is the number of currently queued
2362 signals for this real user ID, and the second is the
2363 resource limit on the number of queued signals for this process
2364 (see the description of
2365 .BR RLIMIT_SIGPENDING
2369 .IR SigPnd ", " ShdPnd :
2370 Number of signals pending for thread and for process as a whole (see
2375 .IR SigBlk ", " SigIgn ", " SigCgt :
2376 Masks indicating signals being blocked, ignored, and caught (see
2379 .IR CapInh ", " CapPrm ", " CapEff :
2380 Masks of capabilities enabled in inheritable, permitted, and effective sets
2382 .BR capabilities (7)).
2385 Capability Bounding set
2386 (since Linux 2.6.26, see
2387 .BR capabilities (7)).
2390 Ambient capability set
2391 (since Linux 4.3, see
2392 .BR capabilities (7)).
2395 .\" commit af884cd4a5ae62fcf5e321fecf0ec1014730353d
2399 (since Linux 4.10, see
2403 .\" commit 2f4b3bf6b2318cfaa177ec5a802f4d8d6afbd816
2404 Seccomp mode of the process
2405 (since Linux 3.8, see
2408 .BR SECCOMP_MODE_DISABLED ;
2410 .BR SECCOMP_MODE_STRICT ;
2412 .BR SECCOMP_MODE_FILTER .
2413 This field is provided only if the kernel was built with the
2415 kernel configuration option enabled.
2418 Mask of CPUs on which this process may run
2419 (since Linux 2.6.24, see
2422 .IR Cpus_allowed_list :
2423 Same as previous, but in "list format"
2424 (since Linux 2.6.26, see
2428 Mask of memory nodes allowed to this process
2429 (since Linux 2.6.24, see
2432 .IR Mems_allowed_list :
2433 Same as previous, but in "list format"
2434 (since Linux 2.6.26, see
2437 .IR voluntary_ctxt_switches ", " nonvoluntary_ctxt_switches :
2438 Number of voluntary and involuntary context switches (since Linux 2.6.23).
2441 .IR /proc/[pid]/syscall " (since Linux 2.6.27)"
2442 .\" commit ebcb67341fee34061430f3367f2e507e52ee051b
2443 This file exposes the system call number and argument registers for the
2444 system call currently being executed by the process,
2445 followed by the values of the stack pointer and program counter registers.
2446 The values of all six argument registers are exposed,
2447 although most system calls use fewer registers.
2449 If the process is blocked, but not in a system call,
2450 then the file displays \-1 in place of the system call number,
2451 followed by just the values of the stack pointer and program counter.
2452 If process is not blocked, then the file contains just the string "running".
2454 This file is present only if the kernel was configured with
2455 .BR CONFIG_HAVE_ARCH_TRACEHOOK .
2457 Permission to access this file is governed by a ptrace access mode
2458 .B PTRACE_MODE_ATTACH_FSCREDS
2462 .IR /proc/[pid]/task " (since Linux 2.6.0-test6)"
2463 This is a directory that contains one subdirectory
2464 for each thread in the process.
2465 The name of each subdirectory is the numerical thread ID
2469 Within each of these subdirectories, there is a set of
2470 files with the same names and contents as under the
2473 For attributes that are shared by all threads, the contents for
2474 each of the files under the
2476 subdirectories will be the same as in the corresponding
2480 (e.g., in a multithreaded process, all of the
2482 files will have the same value as the
2484 file in the parent directory, since all of the threads in a process
2485 share a working directory).
2486 For attributes that are distinct for each thread,
2487 the corresponding files under
2489 may have different values (e.g., various fields in each of the
2490 .I task/[tid]/status
2491 files may be different for each thread),
2492 .\" in particular: "children" :/
2493 or they might not exist in
2496 .\" The following was still true as at kernel 2.6.13
2497 In a multithreaded process, the contents of the
2499 directory are not available if the main thread has already terminated
2500 (typically by calling
2501 .BR pthread_exit (3)).
2504 .IR /proc/[pid]/task/[tid]/children " (since Linux 3.5)"
2505 .\" commit 818411616baf46ceba0cff6f05af3a9b294734f7
2506 A space-separated list of child tasks of this task.
2507 Each child task is represented by its TID.
2509 .\" see comments in get_children_pid() in fs/proc/array.c
2510 This option is intended for use by the checkpoint-restore (CRIU) system,
2511 and reliably provides a list of children only if all of the child processes
2512 are stopped or frozen.
2513 It does not work properly if children of the target task exit while
2514 the file is being read!
2515 Exiting children may cause non-exiting children to be omitted from the list.
2516 This makes this interface even more unreliable than classic PID-based
2517 approaches if the inspected task and its children aren't frozen,
2518 and most code should probably not use this interface.
2520 Until Linux 4.2, the presence of this file was governed by the
2521 .B CONFIG_CHECKPOINT_RESTORE
2522 kernel configuration option.
2524 .\" commit 2e13ba54a2682eea24918b87ad3edf70c2cf085b
2525 it is governed by the
2526 .B CONFIG_PROC_CHILDREN
2529 .IR /proc/[pid]/timers " (since Linux 3.10)"
2530 .\" commit 5ed67f05f66c41e39880a6d61358438a25f9fee5
2531 .\" commit 48f6a7a511ef8823fdff39afee0320092d43a8a0
2532 A list of the POSIX timers for this process.
2533 Each timer is listed with a line that starts with the string "ID:".
2539 signal: 60/00007fff86e452a8
2540 notify: signal/pid.2634
2543 signal: 60/00007fff86e452a8
2544 notify: signal/pid.2634
2549 The lines shown for each timer have the following meanings:
2553 The ID for this timer.
2554 This is not the same as the timer ID returned by
2555 .BR timer_create (2);
2556 rather, it is the same kernel-internal ID that is available via the
2564 This is the signal number that this timer uses to deliver notifications
2565 followed by a slash, and then the
2567 value supplied to the signal handler.
2568 Valid only for timers that notify via a signal.
2571 The part before the slash specifies the mechanism
2572 that this timer uses to deliver notifications,
2573 and is one of "thread", "signal", or "none".
2574 Immediately following the slash is either the string "tid" for timers
2577 notification, or "pid" for timers that notify by other mechanisms.
2578 Following the "." is the PID of the process
2579 (or the kernel thread ID of the thread) that will be delivered
2580 a signal if the timer delivers notifications via a signal.
2583 This field identifies the clock that the timer uses for measuring time.
2584 For most clocks, this is a number that matches one of the user-space
2586 constants exposed via
2588 .B CLOCK_PROCESS_CPUTIME_ID
2589 timers display with a value of -6
2591 .B CLOCK_THREAD_CPUTIME_ID
2592 timers display with a value of -2
2596 This file is available only when the kernel was configured with
2597 .BR CONFIG_CHECKPOINT_RESTORE .
2599 .IR /proc/[pid]/timerslack_ns " (since Linux 4.6)"
2600 .\" commit da8b44d5a9f8bf26da637b7336508ca534d6b319
2601 .\" commit 5de23d435e88996b1efe0e2cebe242074ce67c9e
2602 This file exposes the process's "current" timer slack value,
2603 expressed in nanoseconds.
2604 The file is writable,
2605 allowing the process's timer slack value to be changed.
2606 Writing 0 to this file resets the "current" timer slack to the
2607 "default" timer slack value.
2608 For further details, see the discussion of
2609 .BR PR_SET_TIMERSLACK
2613 Permission to access this file is governed by a ptrace access mode
2614 .B PTRACE_MODE_ATTACH_FSCREDS
2618 .IR /proc/[pid]/uid_map ", " /proc/[pid]/gid_map " (since Linux 3.5)"
2620 .BR user_namespaces (7).
2622 .IR /proc/[pid]/wchan " (since Linux 2.6.0)"
2623 The symbolic name corresponding to the location
2624 in the kernel where the process is sleeping.
2626 Permission to access this file is governed by a ptrace access mode
2627 .B PTRACE_MODE_READ_FSCREDS
2632 Advanced power management version and battery information when
2634 is defined at kernel compilation time.
2637 This file contains information which is used for diagnosing memory
2638 fragmentation issues.
2639 Each line starts with the identification of the node and the name
2640 of the zone which together identify a memory region
2642 followed by the count of available chunks of a certain order in
2643 which these zones are split.
2644 The size in bytes of a certain order is given by the formula:
2646 (2^order)\ *\ PAGE_SIZE
2648 The binary buddy allocator algorithm inside the kernel will split
2649 one chunk into two chunks of a smaller order (thus with half the
2650 size) or combine two contiguous chunks into one larger chunk of
2651 a higher order (thus with double the size) to satisfy allocation
2652 requests and to counter memory fragmentation.
2653 The order matches the column number, when starting to count at zero.
2655 For example on a x86_64 system:
2659 Node 0, zone DMA 1 1 1 0 2 1 1 0 1 1 3
2660 Node 0, zone DMA32 65 47 4 81 52 28 13 10 5 1 404
2661 Node 0, zone Normal 216 55 189 101 84 38 37 27 5 3 587
2665 In this example, there is one node containing three zones and there
2666 are 11 different chunk sizes.
2667 If the page size is 4 kilobytes, then the first zone called
2669 (on x86 the first 16 megabyte of memory) has 1 chunk of 4 kilobytes
2670 (order 0) available and has 3 chunks of 4 megabytes (order 10) available.
2672 If the memory is heavily fragmented, the counters for higher
2673 order chunks will be zero and allocation of large contiguous areas
2676 Further information about the zones can be found in
2677 .IR /proc/zoneinfo .
2680 Contains subdirectories for installed busses.
2683 Subdirectory for PCMCIA devices when
2685 is set at kernel compilation time.
2687 .I /proc/bus/pccard/drivers
2690 Contains various bus subdirectories and pseudo-files containing
2691 information about PCI busses, installed devices, and device
2693 Some of these files are not ASCII.
2695 .I /proc/bus/pci/devices
2696 Information about PCI devices.
2697 They may be accessed through
2702 .IR /proc/cgroups " (since Linux 2.6.24)"
2707 Arguments passed to the Linux kernel at boot time.
2708 Often done via a boot manager such as
2713 .IR /proc/config.gz " (since Linux 2.6)"
2714 This file exposes the configuration options that were used
2715 to build the currently running kernel,
2716 in the same format as they would be shown in the
2718 file that resulted when configuring the kernel (using
2719 .IR "make xconfig" ,
2722 The file contents are compressed; view or search them using
2726 As long as no changes have been made to the following file,
2729 are the same as those provided by :
2733 cat /lib/modules/$(uname \-r)/build/.config
2738 is provided only if the kernel is configured with
2739 .BR CONFIG_IKCONFIG_PROC .
2742 A list of the ciphers provided by the kernel crypto API.
2743 For details, see the kernel
2744 .I "Linux Kernel Crypto API"
2745 documentation available under the kernel source directory
2746 .IR Documentation/DocBook .
2747 (That documentation can be built using a command such as
2749 in the root directory of the kernel source tree.)
2752 This is a collection of CPU and system architecture dependent items,
2753 for each supported architecture a different list.
2754 Two common entries are \fIprocessor\fP which gives CPU number and
2755 \fIbogomips\fP; a system constant that is calculated
2756 during kernel initialization.
2757 SMP machines have information for
2761 command gathers its information from this file.
2764 Text listing of major numbers and device groups.
2765 This can be used by MAKEDEV scripts for consistency with the kernel.
2767 .IR /proc/diskstats " (since Linux 2.5.69)"
2768 This file contains disk I/O statistics for each disk device.
2769 See the Linux kernel source file
2770 .I Documentation/iostats.txt
2771 for further information.
2774 This is a list of the registered \fIISA\fP DMA (direct memory access)
2780 .I /proc/execdomains
2781 List of the execution domains (ABI personalities).
2784 Frame buffer information when
2786 is defined during kernel compilation.
2788 .I /proc/filesystems
2789 A text listing of the filesystems which are supported by the kernel,
2790 namely filesystems which were compiled into the kernel or whose kernel
2791 modules are currently loaded.
2793 .BR filesystems (5).)
2794 If a filesystem is marked with "nodev",
2795 this means that it does not require a block device to be mounted
2796 (e.g., virtual filesystem, network filesystem).
2798 Incidentally, this file may be used by
2800 when no filesystem is specified and it didn't manage to determine the
2802 Then filesystems contained in this file are tried
2803 (excepted those that are marked with "nodev").
2806 .\" FIXME Much more needs to be said about /proc/fs
2808 Contains subdirectories that in turn contain files
2809 with information about (certain) mounted filesystems.
2813 exists on systems with the IDE bus.
2814 There are directories for each IDE channel and attached device.
2819 cache buffer size in KB
2820 capacity number of sectors
2821 driver driver version
2822 geometry physical and logical geometry
2823 identify in hexadecimal
2825 model manufacturer's model number
2826 settings drive settings
2827 smart_thresholds in hexadecimal
2828 smart_values in hexadecimal
2834 utility provides access to this information in a friendly format.
2837 This is used to record the number of interrupts per CPU per IO device.
2839 for the i386 and x86_64 architectures, at least, this also includes
2840 interrupts internal to the system (that is, not associated with a device
2841 as such), such as NMI (nonmaskable interrupt), LOC (local timer interrupt),
2842 and for SMP systems, TLB (TLB flush interrupt), RES (rescheduling
2843 interrupt), CAL (remote function call interrupt), and possibly others.
2844 Very easy to read formatting, done in ASCII.
2847 I/O memory map in Linux 2.4.
2850 This is a list of currently registered Input-Output port regions that
2853 .IR /proc/kallsyms " (since Linux 2.5.71)"
2854 This holds the kernel exported symbol definitions used by the
2856 tools to dynamically link and bind loadable modules.
2857 In Linux 2.5.47 and earlier, a similar file with slightly different syntax
2862 This file represents the physical memory of the system and is stored
2863 in the ELF core file format.
2864 With this pseudo-file, and an unstripped
2866 .RI ( /usr/src/linux/vmlinux )
2867 binary, GDB can be used to
2868 examine the current state of any kernel data structures.
2870 The total length of the file is the size of physical memory (RAM) plus
2873 .IR /proc/keys " (since Linux 2.6.10)"
2877 .IR /proc/key-users " (since Linux 2.6.10)"
2882 This file can be used instead of the
2884 system call to read kernel messages.
2885 A process must have superuser
2886 privileges to read this file, and only one process should read this
2888 This file should not be read if a syslog process is running
2891 system call facility to log kernel messages.
2893 Information in this file is retrieved with the
2897 .IR /proc/kpagecount " (since Linux 2.6.25)"
2898 This file contains a 64-bit count of the number of
2899 times each physical page frame is mapped,
2900 indexed by page frame number (see the discussion of
2901 .IR /proc/[pid]/pagemap ).
2904 .IR /proc/kpagecount
2905 file is present only if the
2906 .B CONFIG_PROC_PAGE_MONITOR
2907 kernel configuration option is enabled.
2909 .IR /proc/kpageflags " (since Linux 2.6.25)"
2910 This file contains 64-bit masks corresponding to each physical page frame;
2911 it is indexed by page frame number (see the discussion of
2912 .IR /proc/[pid]/pagemap ).
2913 The bits are as follows:
2926 11 - KPF_MMAP (since Linux 2.6.31)
2927 12 - KPF_ANON (since Linux 2.6.31)
2928 13 - KPF_SWAPCACHE (since Linux 2.6.31)
2929 14 - KPF_SWAPBACKED (since Linux 2.6.31)
2930 15 - KPF_COMPOUND_HEAD (since Linux 2.6.31)
2931 16 - KPF_COMPOUND_TAIL (since Linux 2.6.31)
2932 16 - KPF_HUGE (since Linux 2.6.31)
2933 18 - KPF_UNEVICTABLE (since Linux 2.6.31)
2934 19 - KPF_HWPOISON (since Linux 2.6.31)
2935 20 - KPF_NOPAGE (since Linux 2.6.31)
2936 21 - KPF_KSM (since Linux 2.6.32)
2937 22 - KPF_THP (since Linux 3.4)
2939 For further details on the meanings of these bits,
2940 see the kernel source file
2941 .IR Documentation/vm/pagemap.txt .
2942 Before kernel 2.6.29,
2943 .\" commit ad3bdefe877afb47480418fdb05ecd42842de65e
2944 .\" commit e07a4b9217d1e97d2f3a62b6b070efdc61212110
2950 did not report correctly.
2953 .IR /proc/kpageflags
2954 file is present only if the
2955 .B CONFIG_PROC_PAGE_MONITOR
2956 kernel configuration option is enabled.
2958 .IR /proc/ksyms " (Linux 1.1.23-2.5.47)"
2960 .IR /proc/kallsyms .
2963 The first three fields in this file are load average figures
2964 giving the number of jobs in the run queue (state R)
2965 or waiting for disk I/O (state D) averaged over 1, 5, and 15 minutes.
2966 They are the same as the load average numbers given by
2969 The fourth field consists of two numbers separated by a slash (/).
2970 The first of these is the number of currently runnable kernel
2971 scheduling entities (processes, threads).
2972 The value after the slash is the number of kernel scheduling entities
2973 that currently exist on the system.
2974 The fifth field is the PID of the process that was most
2975 recently created on the system.
2978 This file shows current file locks
2979 .RB ( flock "(2) and " fcntl (2))
2984 command provides a bit more information about each lock.
2986 .IR /proc/malloc " (only up to and including Linux 2.2)"
2987 .\" It looks like this only ever did something back in 1.0 days
2988 This file is present only if
2989 .B CONFIG_DEBUG_MALLOC
2990 was defined during compilation.
2993 This file reports statistics about memory usage on the system.
2996 to report the amount of free and used memory (both physical and swap)
2997 on the system as well as the shared memory and buffers used by the
2999 Each line of the file consists of a parameter name, followed by a colon,
3000 the value of the parameter, and an option unit of measurement (e.g., "kB").
3001 The list below describes the parameter names and
3002 the format specifier required to read the field value.
3003 Except as noted below,
3004 all of the fields have been present since at least Linux 2.6.0.
3005 Some fields are displayed only if the kernel was configured
3006 with various options; those dependencies are noted in the list.
3010 Total usable RAM (i.e., physical RAM minus a few reserved
3011 bits and the kernel binary code).
3015 .IR LowFree + HighFree .
3017 .IR MemAvailable " %lu (since Linux 3.14)"
3018 An estimate of how much memory is available for starting new
3019 applications, without swapping.
3022 Relatively temporary storage for raw disk blocks that
3023 shouldn't get tremendously large (20MB or so).
3026 In-memory cache for files read from the disk (the page cache).
3030 .IR SwapCached " %lu"
3031 Memory that once was swapped out, is swapped back in but
3032 still also is in the swap file.
3033 (If memory pressure is high, these pages
3034 don't need to be swapped out again because they are already
3039 Memory that has been used more recently and usually not
3040 reclaimed unless absolutely necessary.
3043 Memory which has been less recently used.
3044 It is more eligible to be reclaimed for other purposes.
3046 .IR Active(anon) " %lu (since Linux 2.6.28)"
3049 .IR Inactive(anon) " %lu (since Linux 2.6.28)"
3052 .IR Active(file) " %lu (since Linux 2.6.28)"
3055 .IR Inactive(file) " %lu (since Linux 2.6.28)"
3058 .IR Unevictable " %lu (since Linux 2.6.28)"
3059 (From Linux 2.6.28 to 2.6.30,
3060 \fBCONFIG_UNEVICTABLE_LRU\fP was required.)
3063 .IR Mlocked " %lu (since Linux 2.6.28)"
3064 (From Linux 2.6.28 to 2.6.30,
3065 \fBCONFIG_UNEVICTABLE_LRU\fP was required.)
3068 .IR HighTotal " %lu"
3069 (Starting with Linux 2.6.19, \fBCONFIG_HIGHMEM\fP is required.)
3070 Total amount of highmem.
3071 Highmem is all memory above ~860MB of physical memory.
3072 Highmem areas are for use by user-space programs,
3073 or for the page cache.
3074 The kernel must use tricks to access
3075 this memory, making it slower to access than lowmem.
3078 (Starting with Linux 2.6.19, \fBCONFIG_HIGHMEM\fP is required.)
3079 Amount of free highmem.
3082 (Starting with Linux 2.6.19, \fBCONFIG_HIGHMEM\fP is required.)
3083 Total amount of lowmem.
3084 Lowmem is memory which can be used for everything that
3085 highmem can be used for, but it is also available for the
3086 kernel's use for its own data structures.
3087 Among many other things,
3088 it is where everything from
3091 Bad things happen when you're out of lowmem.
3094 (Starting with Linux 2.6.19, \fBCONFIG_HIGHMEM\fP is required.)
3095 Amount of free lowmem.
3097 .IR MmapCopy " %lu (since Linux 2.6.29)"
3102 .IR SwapTotal " %lu"
3103 Total amount of swap space available.
3106 Amount of swap space that is currently unused.
3109 Memory which is waiting to get written back to the disk.
3111 .IR Writeback " %lu"
3112 Memory which is actively being written back to the disk.
3114 .IR AnonPages " %lu (since Linux 2.6.18)"
3115 Non-file backed pages mapped into user-space page tables.
3118 Files which have been mapped into memory (with
3122 .IR Shmem " %lu (since Linux 2.6.32)"
3123 Amount of memory consumed in
3128 In-kernel data structures cache.
3132 .IR SReclaimable " %lu (since Linux 2.6.19)"
3135 that might be reclaimed, such as caches.
3137 .IR SUnreclaim " %lu (since Linux 2.6.19)"
3140 that cannot be reclaimed on memory pressure.
3142 .IR KernelStack " %lu (since Linux 2.6.32)"
3143 Amount of memory allocated to kernel stacks.
3145 .IR PageTables " %lu (since Linux 2.6.18)"
3146 Amount of memory dedicated to the lowest level of page tables.
3148 .IR Quicklists " %lu (since Linux 2.6.27)"
3149 (\fBCONFIG_QUICKLIST\fP is required.)
3152 .IR NFS_Unstable " %lu (since Linux 2.6.18)"
3153 NFS pages sent to the server, but not yet committed to stable storage.
3155 .IR Bounce " %lu (since Linux 2.6.18)"
3156 Memory used for block device "bounce buffers".
3158 .IR WritebackTmp " %lu (since Linux 2.6.26)"
3159 Memory used by FUSE for temporary writeback buffers.
3161 .IR CommitLimit " %lu (since Linux 2.6.10)"
3162 This is the total amount of memory currently available to
3163 be allocated on the system, expressed in kilobytes.
3164 This limit is adhered to
3165 only if strict overcommit accounting is enabled (mode 2 in
3166 .IR /proc/sys/vm/overcommit_memory ).
3167 The limit is calculated according to the formula described under
3168 .IR /proc/sys/vm/overcommit_memory .
3169 For further details, see the kernel source file
3170 .IR Documentation/vm/overcommit-accounting .
3172 .IR Committed_AS " %lu"
3173 The amount of memory presently allocated on the system.
3174 The committed memory is a sum of all of the memory which
3175 has been allocated by processes, even if it has not been
3176 "used" by them as of yet.
3177 A process which allocates 1GB of memory (using
3179 or similar), but touches only 300MB of that memory will show up
3180 as using only 300MB of memory even if it has the address space
3181 allocated for the entire 1GB.
3183 This 1GB is memory which has been "committed" to by the VM
3184 and can be used at any time by the allocating application.
3185 With strict overcommit enabled on the system (mode 2 in
3186 .IR /proc/sys/vm/overcommit_memory ),
3187 allocations which would exceed the
3189 will not be permitted.
3190 This is useful if one needs to guarantee that processes will not
3191 fail due to lack of memory once that memory has been successfully allocated.
3193 .IR VmallocTotal " %lu"
3194 Total size of vmalloc memory area.
3196 .IR VmallocUsed " %lu"
3197 Amount of vmalloc area which is used.
3199 .IR VmallocChunk " %lu"
3200 Largest contiguous block of vmalloc area which is free.
3202 .IR HardwareCorrupted " %lu (since Linux 2.6.32)"
3203 (\fBCONFIG_MEMORY_FAILURE\fP is required.)
3206 .IR AnonHugePages " %lu (since Linux 2.6.38)"
3207 (\fBCONFIG_TRANSPARENT_HUGEPAGE\fP is required.)
3208 Non-file backed huge pages mapped into user-space page tables.
3210 .IR ShmemHugePages " %lu (since Linux 4.8)"
3211 (\fBCONFIG_TRANSPARENT_HUGEPAGE\fP is required.)
3212 Memory used by shared memory (shmem) and
3214 allocated with huge pages
3216 .IR ShmemPmdMapped " %lu (since Linux 4.8)"
3217 (\fBCONFIG_TRANSPARENT_HUGEPAGE\fP is required.)
3218 Shared memory mapped into user space with huge pages.
3220 .IR CmaTotal " %lu (since Linux 3.1)"
3221 Total CMA (Contiguous Memory Allocator) pages.
3222 (\fBCONFIG_CMA\fP is required.)
3224 .IR CmaFree " %lu (since Linux 3.1)"
3225 Free CMA (Contiguous Memory Allocator) pages.
3226 (\fBCONFIG_CMA\fP is required.)
3228 .IR HugePages_Total " %lu"
3229 (\fBCONFIG_HUGETLB_PAGE\fP is required.)
3230 The size of the pool of huge pages.
3232 .IR HugePages_Free " %lu"
3233 (\fBCONFIG_HUGETLB_PAGE\fP is required.)
3234 The number of huge pages in the pool that are not yet allocated.
3236 .IR HugePages_Rsvd " %lu (since Linux 2.6.17)"
3237 (\fBCONFIG_HUGETLB_PAGE\fP is required.)
3238 This is the number of huge pages for
3239 which a commitment to allocate from the pool has been made,
3240 but no allocation has yet been made.
3241 These reserved huge pages
3242 guarantee that an application will be able to allocate a
3243 huge page from the pool of huge pages at fault time.
3245 .IR HugePages_Surp " %lu (since Linux 2.6.24)"
3246 (\fBCONFIG_HUGETLB_PAGE\fP is required.)
3247 This is the number of huge pages in
3248 the pool above the value in
3249 .IR /proc/sys/vm/nr_hugepages .
3250 The maximum number of surplus huge pages is controlled by
3251 .IR /proc/sys/vm/nr_overcommit_hugepages .
3253 .IR Hugepagesize " %lu"
3254 (\fBCONFIG_HUGETLB_PAGE\fP is required.)
3255 The size of huge pages.
3257 .IR DirectMap4k " %lu (since Linux 2.6.27)"
3258 Number of bytes of RAM linearly mapped by kernel in 4kB pages.
3261 .IR DirectMap4M " %lu (since Linux 2.6.27)"
3262 Number of bytes of RAM linearly mapped by kernel in 4MB pages.
3269 .IR DirectMap2M " %lu (since Linux 2.6.27)"
3270 Number of bytes of RAM linearly mapped by kernel in 2MB pages.
3277 .IR DirectMap1G " %lu (since Linux 2.6.27)"
3281 .B CONFIG_X86_DIRECT_GBPAGES
3286 A text list of the modules that have been loaded by the system.
3291 Before kernel 2.4.19, this file was a list
3292 of all the filesystems currently mounted on the system.
3293 With the introduction of per-process mount namespaces in Linux 2.4.19 (see
3294 .BR mount_namespaces (7)),
3295 this file became a link to
3296 .IR /proc/self/mounts ,
3297 which lists the mount points of the process's own mount namespace.
3298 The format of this file is documented in
3302 Memory Type Range Registers.
3303 See the Linux kernel source file
3304 .I Documentation/mtrr.txt
3308 This directory contains various files and subdirectories containing
3309 information about the networking layer.
3310 The files contain ASCII structures and are,
3311 therefore, readable with
3313 However, the standard
3315 suite provides much cleaner access to these files.
3317 With the advent of network namespaces,
3318 various information relating to the network stack is virtualized (see
3319 .BR namespaces (7)).
3320 Thus, since Linux 2.6.25,
3321 .\" commit e9720acd728a46cb40daa52c99a979f7c4ff195c
3323 is a symbolic link to the directory
3324 .IR /proc/self/net ,
3325 which contains the same files and directories as listed below.
3326 However, these files and directories now expose information
3327 for the network namespace of which the process is a member.
3330 This holds an ASCII readable dump of the kernel ARP table used for
3331 address resolutions.
3332 It will show both dynamically learned and preprogrammed ARP entries.
3338 IP address HW type Flags HW address Mask Device
3339 192.168.0.50 0x1 0x2 00:50:BF:25:68:F3 * eth0
3340 192.168.0.250 0x1 0xc 00:00:00:00:00:00 * eth0
3345 Here "IP address" is the IPv4 address of the machine and the "HW type"
3346 is the hardware type of the address from RFC\ 826.
3347 The flags are the internal
3348 flags of the ARP structure (as defined in
3349 .IR /usr/include/linux/if_arp.h )
3351 the "HW address" is the data link layer mapping for that IP address if
3355 The dev pseudo-file contains network device status information.
3357 the number of received and sent packets, the number of errors and
3359 and other basic statistics.
3360 These are used by the
3362 program to report device status.
3368 Inter-| Receive | Transmit
3369 face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed
3370 lo: 2776770 11307 0 0 0 0 0 0 2776770 11307 0 0 0 0 0 0
3371 eth0: 1215645 2751 0 0 0 0 0 0 1782404 4324 0 0 0 427 0 0
3372 ppp0: 1622270 5552 1 0 0 0 0 0 354130 5669 0 0 0 0 0 0
3373 tap0: 7714 81 0 0 0 0 0 0 7714 81 0 0 0 0 0 0
3378 .\" .I /proc/net/ipx
3381 .\" .I /proc/net/ipx_route
3384 .I /proc/net/dev_mcast
3386 .IR /usr/src/linux/net/core/dev_mcast.c :
3389 indx interface_name dmi_u dmi_g dmi_address
3390 2 eth0 1 0 01005e000001
3391 3 eth1 1 0 01005e000001
3392 4 eth2 1 0 01005e000001
3397 Internet Group Management Protocol.
3399 .IR /usr/src/linux/net/core/igmp.c .
3402 This file uses the same format as the
3404 file and contains the current reverse mapping database used to provide
3406 reverse address lookup services.
3407 If RARP is not configured into the
3409 this file will not be present.
3412 Holds a dump of the RAW socket table.
3413 Much of the information is not of
3415 apart from debugging.
3416 The "sl" value is the kernel hash slot for the
3418 the "local_address" is the local address and protocol number pair.
3420 the internal status of the socket.
3421 The "tx_queue" and "rx_queue" are the
3422 outgoing and incoming data queue in terms of kernel memory usage.
3423 The "tr", "tm\->when", and "rexmits" fields are not used by RAW.
3425 field holds the effective UID of the creator of the socket.
3427 .\" .I /proc/net/route
3428 .\" No information, but looks similar to
3432 This file holds the ASCII data needed for the IP, ICMP, TCP, and UDP
3434 information bases for an SNMP agent.
3437 Holds a dump of the TCP socket table.
3438 Much of the information is not
3439 of use apart from debugging.
3440 The "sl" value is the kernel hash slot
3441 for the socket, the "local_address" is the local address and port number pair.
3442 The "rem_address" is the remote address and port number pair
3444 \&"St" is the internal status of the socket.
3445 The "tx_queue" and "rx_queue" are the
3446 outgoing and incoming data queue in terms of kernel memory usage.
3447 The "tr", "tm\->when", and "rexmits" fields hold internal information of
3448 the kernel socket state and are useful only for debugging.
3450 field holds the effective UID of the creator of the socket.
3453 Holds a dump of the UDP socket table.
3454 Much of the information is not of
3455 use apart from debugging.
3456 The "sl" value is the kernel hash slot for the
3457 socket, the "local_address" is the local address and port number pair.
3458 The "rem_address" is the remote address and port number pair
3460 "St" is the internal status of the socket.
3461 The "tx_queue" and "rx_queue" are the outgoing and incoming data queue
3462 in terms of kernel memory usage.
3463 The "tr", "tm\->when", and "rexmits" fields
3464 are not used by UDP.
3466 field holds the effective UID of the creator of the socket.
3472 sl local_address rem_address st tx_queue rx_queue tr rexmits tm\->when uid
3473 1: 01642C89:0201 0C642C89:03FF 01 00000000:00000001 01:000071BA 00000000 0
3474 1: 00000000:0801 00000000:0000 0A 00000000:00000000 00:00000000 6F000100 0
3475 1: 00000000:0201 00000000:0000 0A 00000000:00000000 00:00000000 00000000 0
3481 Lists the UNIX domain sockets present within the system and their
3487 Num RefCount Protocol Flags Type St Path
3488 0: 00000002 00000000 00000000 0001 03
3489 1: 00000001 00000000 00010000 0001 01 /dev/printer
3494 The fields are as follows:
3498 the kernel table slot number.
3501 the number of users of the socket.
3507 the internal kernel flags holding the status of the socket.
3513 sockets, this is 0001; for
3515 sockets, it is 0002; and for
3517 sockets, it is 0005.
3520 the internal state of the socket.
3523 the bound path (if any) of the socket.
3524 Sockets in the abstract namespace are included in the list,
3525 and are shown with a
3527 that commences with the character '@'.
3530 .I /proc/net/netfilter/nfnetlink_queue
3531 This file contains information about netfilter user-space queueing, if used.
3532 Each line represents a queue.
3533 Queues that have not been subscribed to
3534 by user space are not shown.
3537 1 4207 0 2 65535 0 0 0 1
3538 (1) (2) (3)(4) (5) (6) (7) (8)
3541 The fields in each line are:
3545 The ID of the queue.
3546 This matches what is specified in the
3549 .B \-\-queue\-balance
3554 .BR iptables-extensions (8)
3555 for more information.
3558 The netlink port ID subscribed to the queue.
3561 The number of packets currently queued and waiting to be processed by
3565 The copy mode of the queue.
3566 It is either 1 (metadata only) or 2
3567 (also copy payload data to user space).
3570 Copy range; that is, how many bytes of packet payload should be copied to
3575 Number of packets that had to be dropped by the kernel because
3576 too many packets are already waiting for user space to send back the mandatory
3577 accept/drop verdicts.
3581 Number of packets that were dropped within the netlink
3583 Such drops usually happen when the corresponding socket buffer is
3584 full; that is, user space is not able to read messages fast enough.
3588 Every queued packet is associated with a (32-bit)
3589 monotonically-increasing sequence number.
3590 This shows the ID of the most recent packet queued.
3593 The last number exists only for compatibility reasons and is always 1.
3596 Contains the major and minor numbers of each partition as well as the number
3597 of 1024-byte blocks and the partition name.
3600 This is a listing of all PCI devices found during kernel initialization
3601 and their configuration.
3603 This file has been deprecated in favor of a new
3606 .RI ( /proc/bus/pci ).
3607 It became optional in Linux 2.2 (available with
3608 .B CONFIG_PCI_OLD_PROC
3609 set at kernel compilation).
3610 It became once more nonoptionally enabled in Linux 2.4.
3611 Next, it was deprecated in Linux 2.6 (still available with
3612 .B CONFIG_PCI_LEGACY_PROC
3613 set), and finally removed altogether since Linux 2.6.17.
3614 .\" FIXME Document /proc/sched_debug (since Linux 2.6.23)
3615 .\" See also /proc/[pid]/sched
3617 .IR /proc/profile " (since Linux 2.4)"
3618 This file is present only if the kernel was booted with the
3620 command-line option.
3621 It exposes kernel profiling information in a binary format for use by
3622 .BR readprofile (1).
3623 Writing (e.g., an empty string) to this file resets the profiling counters;
3624 on some architectures,
3625 writing a binary integer "profiling multiplier" of size
3627 sets the profiling interrupt frequency.
3630 A directory with the
3632 mid-level pseudo-file and various SCSI low-level
3634 which contain a file for each SCSI host in this system, all of
3635 which give the status of some part of the SCSI IO subsystem.
3636 These files contain ASCII structures and are, therefore, readable with
3639 You can also write to some of the files to reconfigure the subsystem or
3640 switch certain features on or off.
3643 This is a listing of all SCSI devices known to the kernel.
3644 The listing is similar to the one seen during bootup.
3645 scsi currently supports only the \fIadd-single-device\fP command which
3646 allows root to add a hotplugged device to the list of known devices.
3652 echo \(aqscsi add-single-device 1 0 5 0\(aq > /proc/scsi/scsi
3657 host scsi1 to scan on SCSI channel 0 for a device on ID 5 LUN 0.
3659 is already a device known on this address or the address is invalid, an
3660 error will be returned.
3662 .I /proc/scsi/[drivername]
3663 \fI[drivername]\fP can currently be NCR53c7xx, aha152x, aha1542, aha1740,
3664 aic7xxx, buslogic, eata_dma, eata_pio, fdomain, in2000, pas16, qlogic,
3665 scsi_debug, seagate, t128, u15-24f, ultrastore, or wd7000.
3666 These directories show up for all drivers that registered at least one
3668 Every directory contains one file per registered host.
3669 Every host-file is named after the number the host was assigned during
3672 Reading these files will usually show driver and host configuration,
3673 statistics, and so on.
3675 Writing to these files allows different things on different hosts.
3676 For example, with the \fIlatency\fP and \fInolatency\fP commands,
3677 root can switch on and off command latency measurement code in the
3679 With the \fIlockup\fP and \fIunlock\fP commands,
3680 root can control bus lockups simulated by the scsi_debug driver.
3683 This directory refers to the process accessing the
3686 and is identical to the
3688 directory named by the process ID of the same process.
3691 Information about kernel caches.
3697 kernel/system statistics.
3698 Varies with architecture.
3703 \fIcpu 3357 0 4313 1362393\fP
3704 The amount of time, measured in units of
3705 USER_HZ (1/100ths of a second on most architectures, use
3706 .IR sysconf(_SC_CLK_TCK)
3707 to obtain the right value),
3708 .\" 1024 on Alpha and ia64
3709 that the system spent in various states:
3713 (1) Time spent in user mode.
3716 (2) Time spent in user mode with low priority (nice).
3719 (3) Time spent in system mode.
3722 (4) Time spent in the idle task.
3723 .\" FIXME . Actually, the following info about the /proc/stat 'cpu' field
3724 .\" does not seem to be quite right (at least in 2.6.12 or 3.6):
3725 .\" the idle time in /proc/uptime does not quite match this value
3726 This value should be USER_HZ times the
3731 .IR iowait " (since Linux 2.5.41)"
3732 (5) Time waiting for I/O to complete.
3734 .IR irq " (since Linux 2.6.0-test4)"
3735 (6) Time servicing interrupts.
3737 .IR softirq " (since Linux 2.6.0-test4)"
3738 (7) Time servicing softirqs.
3740 .IR steal " (since Linux 2.6.11)"
3741 (8) Stolen time, which is the time spent in other operating systems when
3742 running in a virtualized environment
3744 .IR guest " (since Linux 2.6.24)"
3745 (9) Time spent running a virtual CPU for guest
3746 operating systems under the control of the Linux kernel.
3747 .\" See Changelog entry for 5e84cfde51cf303d368fcb48f22059f37b3872de
3749 .IR guest_nice " (since Linux 2.6.33)"
3750 .\" commit ce0e7b28fb75cb003cfc8d0238613aaf1c55e797
3751 (10) Time spent running a niced guest (virtual CPU for guest
3752 operating systems under the control of the Linux kernel).
3755 \fIpage 5741 1808\fP
3756 The number of pages the system paged in and the number that were paged
3760 The number of swap pages that have been brought in and out.
3762 .\" FIXME . The following is not the full picture for the 'intr' of
3763 .\" /proc/stat on 2.6:
3765 This line shows counts of interrupts serviced since boot time,
3766 for each of the possible system interrupts.
3767 The first column is the total of all interrupts serviced
3768 including unnumbered architecture specific interrupts;
3769 each subsequent column is the total for that particular numbered interrupt.
3770 Unnumbered interrupts are not shown, only summed into the total.
3772 \fIdisk_io: (2,0):(31,30,5764,1,2) (3,0):\fP...
3773 (major,disk_idx):(noinfo, read_io_ops, blks_read, write_io_ops, blks_written)
3778 The number of context switches that the system underwent.
3780 \fIbtime 769041601\fP
3781 boot time, in seconds since the Epoch, 1970-01-01 00:00:00 +0000 (UTC).
3783 \fIprocesses 86031\fP
3784 Number of forks since boot.
3786 \fIprocs_running 6\fP
3787 Number of processes in runnable state.
3788 (Linux 2.5.45 onward.)
3790 \fIprocs_blocked 2\fP
3791 Number of processes blocked waiting for I/O to complete.
3792 (Linux 2.5.45 onward.)
3801 This directory (present since 1.3.57) contains a number of files
3802 and subdirectories corresponding to kernel variables.
3803 These variables can be read and sometimes modified using
3804 the \fI/proc\fP filesystem, and the (deprecated)
3808 String values may be terminated by either \(aq\\0\(aq or \(aq\\n\(aq.
3810 Integer and long values may be written either in decimal or in
3811 hexadecimal notation (e.g. 0x3FFF).
3812 When writing multiple integer or long values, these may be separated
3813 by any of the following whitespace characters:
3814 \(aq\ \(aq, \(aq\\t\(aq, or \(aq\\n\(aq.
3815 Using other separators leads to the error
3818 .IR /proc/sys/abi " (since Linux 2.4.10)"
3819 This directory may contain files with application binary information.
3820 .\" On some systems, it is not present.
3821 See the Linux kernel source file
3822 .I Documentation/sysctl/abi.txt
3823 for more information.
3826 This directory may be empty.
3829 This directory contains device-specific information (e.g.,
3830 .IR dev/cdrom/info ).
3832 some systems, it may be empty.
3835 This directory contains the files and subdirectories for kernel variables
3836 related to filesystems.
3838 .I /proc/sys/fs/binfmt_misc
3839 Documentation for files in this directory can be found
3840 in the Linux kernel sources in
3841 .IR Documentation/binfmt_misc.txt .
3843 .IR /proc/sys/fs/dentry-state " (since Linux 2.2)"
3844 This file contains information about the status of the
3845 directory cache (dcache).
3846 The file contains six numbers,
3847 .IR nr_dentry ", " nr_unused ", " age_limit " (age in seconds), "
3849 (pages requested by system) and two dummy values.
3853 is the number of allocated dentries (dcache entries).
3854 This field is unused in Linux 2.2.
3857 is the number of unused dentries.
3860 .\" looks like this is unused in kernels 2.2 to 2.6
3861 is the age in seconds after which dcache entries
3862 can be reclaimed when memory is short.
3865 .\" looks like this is unused in kernels 2.2 to 2.6
3866 is nonzero when the kernel has called shrink_dcache_pages() and the
3867 dcache isn't pruned yet.
3870 .I /proc/sys/fs/dir-notify-enable
3871 This file can be used to disable or enable the
3873 interface described in
3875 on a system-wide basis.
3876 A value of 0 in this file disables the interface,
3877 and a value of 1 enables it.
3879 .I /proc/sys/fs/dquot-max
3880 This file shows the maximum number of cached disk quota entries.
3881 On some (2.4) systems, it is not present.
3882 If the number of free cached disk quota entries is very low and
3883 you have some awesome number of simultaneous system users,
3884 you might want to raise the limit.
3886 .I /proc/sys/fs/dquot-nr
3887 This file shows the number of allocated disk quota
3888 entries and the number of free disk quota entries.
3890 .IR /proc/sys/fs/epoll " (since Linux 2.6.28)"
3891 This directory contains the file
3892 .IR max_user_watches ,
3893 which can be used to limit the amount of kernel memory consumed by the
3896 For further details, see
3899 .I /proc/sys/fs/file-max
3901 a system-wide limit on the number of open files for all processes.
3902 System calls that fail when encountering this limit fail with the error
3906 which can be used by a process to set the per-process limit,
3908 on the number of files it may open.)
3910 of error messages in the kernel log about running out of file handles
3911 (look for "VFS: file-max limit <number> reached"),
3912 try increasing this value:
3918 echo 100000 > /proc/sys/fs/file-max
3922 Privileged processes
3923 .RB ( CAP_SYS_ADMIN )
3928 .I /proc/sys/fs/file-nr
3929 This (read-only) file contains three numbers:
3930 the number of allocated file handles
3931 (i.e., the number of files presently opened);
3932 the number of free file handles;
3933 and the maximum number of file handles (i.e., the same value as
3934 .IR /proc/sys/fs/file-max ).
3935 If the number of allocated file handles is close to the
3936 maximum, you should consider increasing the maximum.
3938 the kernel allocated file handles dynamically,
3939 but it didn't free them again.
3940 Instead the free file handles were kept in a list for reallocation;
3941 the "free file handles" value indicates the size of that list.
3942 A large number of free file handles indicates that there was
3943 a past peak in the usage of open file handles.
3944 Since Linux 2.6, the kernel does deallocate freed file handles,
3945 and the "free file handles" value is always zero.
3947 .IR /proc/sys/fs/inode-max " (only present until Linux 2.2)"
3948 This file contains the maximum number of in-memory inodes.
3949 This value should be 3-4 times larger
3952 since \fIstdin\fP, \fIstdout\fP
3953 and network sockets also need an inode to handle them.
3954 When you regularly run out of inodes, you need to increase this value.
3956 Starting with Linux 2.4,
3957 there is no longer a static limit on the number of inodes,
3958 and this file is removed.
3960 .I /proc/sys/fs/inode-nr
3961 This file contains the first two values from
3964 .I /proc/sys/fs/inode-state
3966 contains seven numbers:
3968 .IR nr_free_inodes ,
3970 and four dummy values (always zero).
3973 is the number of inodes the system has allocated.
3974 .\" This can be slightly more than
3976 .\" because Linux allocates them one page full at a time.
3978 represents the number of free inodes.
3985 and the system needs to prune the inode list instead of allocating more;
3986 since Linux 2.4, this field is a dummy value (always zero).
3988 .IR /proc/sys/fs/inotify " (since Linux 2.6.13)"
3989 This directory contains files
3990 .IR max_queued_events ", " max_user_instances ", and " max_user_watches ,
3991 that can be used to limit the amount of kernel memory consumed by the
3994 For further details, see
3997 .I /proc/sys/fs/lease-break-time
3998 This file specifies the grace period that the kernel grants to a process
3999 holding a file lease
4001 after it has sent a signal to that process notifying it
4002 that another process is waiting to open the file.
4003 If the lease holder does not remove or downgrade the lease within
4004 this grace period, the kernel forcibly breaks the lease.
4006 .I /proc/sys/fs/leases-enable
4007 This file can be used to enable or disable file leases
4009 on a system-wide basis.
4010 If this file contains the value 0, leases are disabled.
4011 A nonzero value enables leases.
4013 .IR /proc/sys/fs/mount-max " (since Linux 4.9)"
4014 .\" commit d29216842a85c7970c536108e093963f02714498
4015 The value in this file specifies the maximum number of mounts that may exist
4016 in a mount namespace.
4017 The default value in this file is 100,000.
4019 .IR /proc/sys/fs/mqueue " (since Linux 2.6.6)"
4020 This directory contains files
4021 .IR msg_max ", " msgsize_max ", and " queues_max ,
4022 controlling the resources used by POSIX message queues.
4027 .IR /proc/sys/fs/nr_open " (since Linux 2.6.25)
4028 .\" commit 9cfe015aa424b3c003baba3841a60dd9b5ad319b
4029 This file imposes ceiling on the value to which the
4031 resource limit can be raised (see
4033 This ceiling is enforced for both unprivileged and privileged process.
4034 The default value in this file is 1048576.
4035 (Before Linux 2.6.25, the ceiling for
4037 was hard-coded to the same value.)
4039 .IR /proc/sys/fs/overflowgid " and " /proc/sys/fs/overflowuid
4041 allow you to change the value of the fixed UID and GID.
4042 The default is 65534.
4043 Some filesystems support only 16-bit UIDs and GIDs, although in Linux
4044 UIDs and GIDs are 32 bits.
4045 When one of these filesystems is mounted
4046 with writes enabled, any UID or GID that would exceed 65535 is translated
4047 to the overflow value before being written to disk.
4049 .IR /proc/sys/fs/pipe-max-size " (since Linux 2.6.35)"
4053 .IR /proc/sys/fs/pipe-user-pages-hard " (since Linux 4.5)"
4057 .IR /proc/sys/fs/pipe-user-pages-soft " (since Linux 4.5)"
4061 .IR /proc/sys/fs/protected_hardlinks " (since Linux 3.6)"
4062 .\" commit 800179c9b8a1e796e441674776d11cd4c05d61d7
4063 When the value in this file is 0,
4064 no restrictions are placed on the creation of hard links
4065 (i.e., this is the historical behavior before Linux 3.6).
4066 When the value in this file is 1,
4067 a hard link can be created to a target file
4068 only if one of the following conditions is true:
4071 The calling process has the
4073 capability in its user namespace
4074 and the file UID has a mapping in the namespace.
4076 The filesystem UID of the process creating the link matches
4077 the owner (UID) of the target file
4079 .BR credentials (7),
4080 a process's filesystem UID is normally the same as its effective UID).
4082 All of the following conditions are true:
4085 the target is a regular file;
4087 the target file does not have its set-user-ID mode bit enabled;
4089 the target file does not have both its set-group-ID and
4090 group-executable mode bits enabled; and
4092 the caller has permission to read and write the target file
4093 (either via the file's permissions mask or because it has
4094 suitable capabilities).
4098 The default value in this file is 0.
4099 Setting the value to 1
4100 prevents a longstanding class of security issues caused by
4101 hard-link-based time-of-check, time-of-use races,
4102 most commonly seen in world-writable directories such as
4104 The common method of exploiting this flaw
4105 is to cross privilege boundaries when following a given hard link
4106 (i.e., a root process follows a hard link created by another user).
4107 Additionally, on systems without separated partitions,
4108 this stops unauthorized users from "pinning" vulnerable set-user-ID and
4109 set-group-ID files against being upgraded by
4110 the administrator, or linking to special files.
4112 .IR /proc/sys/fs/protected_symlinks " (since Linux 3.6)"
4113 .\" commit 800179c9b8a1e796e441674776d11cd4c05d61d7
4114 When the value in this file is 0,
4115 no restrictions are placed on following symbolic links
4116 (i.e., this is the historical behavior before Linux 3.6).
4117 When the value in this file is 1, symbolic links are followed only
4118 in the following circumstances:
4121 the filesystem UID of the process following the link matches
4122 the owner (UID) of the symbolic link
4124 .BR credentials (7),
4125 a process's filesystem UID is normally the same as its effective UID);
4127 the link is not in a sticky world-writable directory; or
4129 the symbolic link and its parent directory have the same owner (UID)
4132 A system call that fails to follow a symbolic link
4133 because of the above restrictions returns the error
4138 The default value in this file is 0.
4139 Setting the value to 1 avoids a longstanding class of security issues
4140 based on time-of-check, time-of-use races when accessing symbolic links.
4142 .IR /proc/sys/fs/suid_dumpable " (since Linux 2.6.13)"
4143 .\" The following is based on text from Documentation/sysctl/kernel.txt
4144 The value in this file is assigned to a process's "dumpable" flag
4145 in the circumstances described in
4148 the value in this file determines whether core dump files are
4149 produced for set-user-ID or otherwise protected/tainted binaries.
4150 The "dumpable" setting also affects the ownership of files in a process's
4152 directory, as described above.
4154 Three different integer values can be specified:
4158 .\" In kernel source: SUID_DUMP_DISABLE
4159 This provides the traditional (pre-Linux 2.6.13) behavior.
4160 A core dump will not be produced for a process which has
4161 changed credentials (by calling
4164 or similar, or by executing a set-user-ID or set-group-ID program)
4165 or whose binary does not have read permission enabled.
4168 .\" In kernel source: SUID_DUMP_USER
4169 All processes dump core when possible.
4170 (Reasons why a process might nevertheless not dump core are described in
4172 The core dump is owned by the filesystem user ID of the dumping process
4173 and no security is applied.
4174 This is intended for system debugging situations only:
4175 this mode is insecure because it allows unprivileged users to
4176 examine the memory contents of privileged processes.
4178 \fI2\ ("suidsafe")\fP
4179 .\" In kernel source: SUID_DUMP_ROOT
4180 Any binary which normally would not be dumped (see "0" above)
4181 is dumped readable by root only.
4182 This allows the user to remove the core dump file but not to read it.
4183 For security reasons core dumps in this mode will not overwrite one
4184 another or other files.
4185 This mode is appropriate when administrators are
4186 attempting to debug problems in a normal environment.
4188 Additionally, since Linux 3.6,
4189 .\" 9520628e8ceb69fa9a4aee6b57f22675d9e1b709
4190 .I /proc/sys/kernel/core_pattern
4191 must either be an absolute pathname
4192 or a pipe command, as detailed in
4194 Warnings will be written to the kernel log if
4196 does not follow these rules, and no core dump will be produced.
4197 .\" 54b501992dd2a839e94e76aa392c392b55080ce8
4200 For details of the effect of a process's "dumpable" setting
4201 on ptrace access mode checking, see
4204 .I /proc/sys/fs/super-max
4206 controls the maximum number of superblocks, and
4207 thus the maximum number of mounted filesystems the kernel
4209 You need increase only
4211 if you need to mount more filesystems than the current value in
4215 .I /proc/sys/fs/super-nr
4217 contains the number of filesystems currently mounted.
4220 This directory contains files controlling a range of kernel parameters,
4223 .I /proc/sys/kernel/acct
4225 contains three numbers:
4230 If BSD-style process accounting is enabled, these values control
4232 If free space on filesystem where the log lives goes below
4234 percent, accounting suspends.
4235 If free space gets above
4237 percent, accounting resumes.
4240 how often the kernel checks the amount of free space (value is in
4242 Default values are 4, 2 and 30.
4243 That is, suspend accounting if 2% or less space is free; resume it
4244 if 4% or more space is free; consider information about amount of free space
4245 valid for 30 seconds.
4247 .IR /proc/sys/kernel/auto_msgmni " (Linux 2.6.27 to 3.18)"
4248 .\" commit 9eefe520c814f6f62c5d36a2ddcd3fb99dfdb30e (introduces feature)
4249 .\" commit 0050ee059f7fc86b1df2527aaa14ed5dc72f9973 (rendered redundant)
4250 From Linux 2.6.27 to 3.18,
4251 this file was used to control recomputing of the value in
4252 .IR /proc/sys/kernel/msgmni
4253 upon the addition or removal of memory or upon IPC namespace creation/removal.
4254 Echoing "1" into this file enabled
4256 automatic recomputing (and triggered a recomputation of
4258 based on the current amount of available memory and number of IPC namespaces).
4259 Echoing "0" disabled automatic recomputing.
4260 (Automatic recomputing was also disabled if a value was explicitly assigned to
4261 .IR /proc/sys/kernel/msgmni .)
4262 The default value in
4266 Since Linux 3.19, the content of this file has no effect (because
4268 .\" FIXME Must document the 3.19 'msgmni' changes.
4269 defaults to near the maximum value possible),
4270 and reads from this file always return the value "0".
4272 .IR /proc/sys/kernel/cap_last_cap " (since Linux 3.2)"
4274 .BR capabilities (7).
4276 .IR /proc/sys/kernel/cap-bound " (from Linux 2.2 to 2.6.24)"
4277 This file holds the value of the kernel
4278 .I "capability bounding set"
4279 (expressed as a signed decimal number).
4280 This set is ANDed against the capabilities permitted to a process
4283 Starting with Linux 2.6.25,
4284 the system-wide capability bounding set disappeared,
4285 and was replaced by a per-thread bounding set; see
4286 .BR capabilities (7).
4288 .I /proc/sys/kernel/core_pattern
4292 .I /proc/sys/kernel/core_pipe_limit
4296 .I /proc/sys/kernel/core_uses_pid
4300 .I /proc/sys/kernel/ctrl-alt-del
4302 controls the handling of Ctrl-Alt-Del from the keyboard.
4303 When the value in this file is 0, Ctrl-Alt-Del is trapped and
4306 program to handle a graceful restart.
4307 When the value is greater than zero, Linux's reaction to a Vulcan
4308 Nerve Pinch (tm) will be an immediate reboot, without even
4309 syncing its dirty buffers.
4310 Note: when a program (like dosemu) has the keyboard in "raw"
4311 mode, the ctrl-alt-del is intercepted by the program before it
4312 ever reaches the kernel tty layer, and it's up to the program
4313 to decide what to do with it.
4315 .IR /proc/sys/kernel/dmesg_restrict " (since Linux 2.6.37)"
4316 The value in this file determines who can see kernel syslog contents.
4317 A value of 0 in this file imposes no restrictions.
4318 If the value is 1, only privileged users can read the kernel syslog.
4323 .\" commit 620f6e8e855d6d447688a5f67a4e176944a084e8
4326 capability may change the value in this file.
4328 .IR /proc/sys/kernel/domainname " and " /proc/sys/kernel/hostname
4329 can be used to set the NIS/YP domainname and the
4330 hostname of your box in exactly the same way as the commands
4338 .RB "#" " echo \(aqdarkstar\(aq > /proc/sys/kernel/hostname"
4339 .RB "#" " echo \(aqmydomain\(aq > /proc/sys/kernel/domainname"
4343 has the same effect as
4347 .RB "#" " hostname \(aqdarkstar\(aq"
4348 .RB "#" " domainname \(aqmydomain\(aq"
4352 Note, however, that the classic darkstar.frop.org has the
4353 hostname "darkstar" and DNS (Internet Domain Name Server)
4354 domainname "frop.org", not to be confused with the NIS (Network
4355 Information Service) or YP (Yellow Pages) domainname.
4357 domain names are in general different.
4358 For a detailed discussion
4363 .I /proc/sys/kernel/hotplug
4365 contains the path for the hotplug policy agent.
4366 The default value in this file is
4369 .I /proc/sys/kernel/htab-reclaim
4370 (PowerPC only) If this file is set to a nonzero value,
4373 .IR Documentation/powerpc/ppc_htab.txt )
4375 each time the system hits the idle loop.
4377 .IR /proc/sys/kernel/keys/*
4378 This directory contains various files that define parameters and limits
4379 for the key-management facility.
4380 These files are described in
4383 .IR /proc/sys/kernel/kptr_restrict " (since Linux 2.6.38)"
4384 .\" 455cd5ab305c90ffc422dd2e0fb634730942b257
4385 The value in this file determines whether kernel addresses are exposed via
4387 files and other interfaces.
4388 A value of 0 in this file imposes no restrictions.
4389 If the value is 1, kernel pointers printed using the
4391 format specifier will be replaced with zeros unless the user has the
4394 If the value is 2, kernel pointers printed using the
4396 format specifier will be replaced with zeros regardless
4397 of the user's capabilities.
4398 The initial default value for this file was 1,
4399 but the default was changed
4400 .\" commit 411f05f123cbd7f8aa1edcae86970755a6e2a9d9
4401 to 0 in Linux 2.6.39.
4403 .\" commit 620f6e8e855d6d447688a5f67a4e176944a084e8
4406 capability can change the value in this file.
4408 .I /proc/sys/kernel/l2cr
4409 (PowerPC only) This file
4410 contains a flag that controls the L2 cache of G3 processor
4412 If 0, the cache is disabled.
4415 .I /proc/sys/kernel/modprobe
4416 This file contains the path for the kernel module loader.
4417 The default value is
4418 .IR /sbin/modprobe .
4419 The file is present only if the kernel is built with the
4422 in Linux 2.6.26 and earlier)
4424 It is described by the Linux kernel source file
4425 .I Documentation/kmod.txt
4426 (present only in kernel 2.4 and earlier).
4428 .IR /proc/sys/kernel/modules_disabled " (since Linux 2.6.31)"
4429 .\" 3d43321b7015387cfebbe26436d0e9d299162ea1
4430 .\" From Documentation/sysctl/kernel.txt
4431 A toggle value indicating if modules are allowed to be loaded
4432 in an otherwise modular kernel.
4433 This toggle defaults to off (0), but can be set true (1).
4434 Once true, modules can be neither loaded nor unloaded,
4435 and the toggle cannot be set back to false.
4436 The file is present only if the kernel is built with the
4440 .IR /proc/sys/kernel/msgmax " (since Linux 2.2)"
4442 a system-wide limit specifying the maximum number of bytes in
4443 a single message written on a System V message queue.
4445 .IR /proc/sys/kernel/msgmni " (since Linux 2.4)"
4446 This file defines the system-wide limit on the number of
4447 message queue identifiers.
4449 .IR /proc/sys/kernel/auto_msgmni .
4451 .IR /proc/sys/kernel/msgmnb " (since Linux 2.2)"
4452 This file defines a system-wide parameter used to initialize the
4454 setting for subsequently created message queues.
4457 setting specifies the maximum number of bytes that may be written to the
4460 .IR /proc/sys/kernel/ngroups_max " (since Linux 2.6.4)"
4461 This is a read-only file that displays the upper limit on the
4462 number of a process's group memberships.
4464 .IR /proc/sys/kernel/ostype " and " /proc/sys/kernel/osrelease
4469 .IR /proc/sys/kernel/overflowgid " and " /proc/sys/kernel/overflowuid
4470 These files duplicate the files
4471 .I /proc/sys/fs/overflowgid
4473 .IR /proc/sys/fs/overflowuid .
4475 .I /proc/sys/kernel/panic
4476 This file gives read/write access to the kernel variable
4478 If this is zero, the kernel will loop on a panic; if nonzero,
4479 it indicates that the kernel should autoreboot after this number
4482 software watchdog device driver, the recommended setting is 60.
4484 .IR /proc/sys/kernel/panic_on_oops " (since Linux 2.5.68)"
4485 This file controls the kernel's behavior when an oops
4486 or BUG is encountered.
4487 If this file contains 0, then the system
4488 tries to continue operation.
4489 If it contains 1, then the system
4490 delays a few seconds (to give klogd time to record the oops output)
4493 .I /proc/sys/kernel/panic
4494 file is also nonzero, then the machine will be rebooted.
4496 .IR /proc/sys/kernel/pid_max " (since Linux 2.5.34)"
4497 This file specifies the value at which PIDs wrap around
4498 (i.e., the value in this file is one greater than the maximum PID).
4499 PIDs greater than this value are not allocated;
4500 thus, the value in this file also acts as a system-wide limit
4501 on the total number of processes and threads.
4502 The default value for this file, 32768,
4503 results in the same range of PIDs as on earlier kernels.
4504 On 32-bit platforms, 32768 is the maximum value for
4508 can be set to any value up to 2^22
4509 .RB ( PID_MAX_LIMIT ,
4510 approximately 4 million).
4511 .\" Prior to 2.6.10, pid_max could also be raised above 32768 on 32-bit
4512 .\" platforms, but this broke /proc/[pid]
4513 .\" See http://marc.theaimsgroup.com/?l=linux-kernel&m=109513010926152&w=2
4515 .IR /proc/sys/kernel/powersave-nap " (PowerPC only)"
4516 This file contains a flag.
4517 If set, Linux-PPC will use the "nap" mode of
4519 otherwise the "doze" mode will be used.
4521 .I /proc/sys/kernel/printk
4525 .IR /proc/sys/kernel/pty " (since Linux 2.6.4)"
4526 This directory contains two files relating to the number of UNIX 98
4527 pseudoterminals (see
4531 .I /proc/sys/kernel/pty/max
4532 This file defines the maximum number of pseudoterminals.
4533 .\" FIXME Document /proc/sys/kernel/pty/reserve
4534 .\" New in Linux 3.3
4535 .\" commit e9aba5158a80098447ff207a452a3418ae7ee386
4537 .I /proc/sys/kernel/pty/nr
4539 indicates how many pseudoterminals are currently in use.
4541 .I /proc/sys/kernel/random
4543 contains various parameters controlling the operation of the file
4547 for further information.
4549 .IR /proc/sys/kernel/random/uuid " (since Linux 2.4)"
4550 Each read from this read-only file returns a randomly generated 128-bit UUID,
4551 as a string in the standard UUID format.
4553 .IR /proc/sys/kernel/randomize_va_space " (since Linux 2.6.12)"
4554 .\" Some further details can be found in Documentation/sysctl/kernel.txt
4555 Select the address space layout randomization (ASLR) policy for the system
4556 (on architectures that support ASLR).
4557 Three values are supported for this file:
4561 This is the default for architectures that don't support ASLR,
4562 and when the kernel is booted with the
4566 Make the addresses of
4568 allocations, the stack, and the VDSO page randomized.
4569 Among other things, this means that shared libraries will be
4570 loaded at randomized addresses.
4571 The text segment of PIE-linked binaries will also be loaded
4572 at a randomized address.
4573 This value is the default if the kernel was configured with
4574 .BR CONFIG_COMPAT_BRK .
4576 (Since Linux 2.6.25)
4577 .\" commit c1d171a002942ea2d93b4fbd0c9583c56fce0772
4578 Also support heap randomization.
4579 This value is the default if the kernel was not configured with
4580 .BR CONFIG_COMPAT_BRK .
4583 .I /proc/sys/kernel/real-root-dev
4584 This file is documented in the Linux kernel source file
4585 .IR Documentation/initrd.txt .
4587 .IR /proc/sys/kernel/reboot-cmd " (Sparc only) "
4588 This file seems to be a way to give an argument to the SPARC
4589 ROM/Flash boot loader.
4590 Maybe to tell it what to do after
4593 .I /proc/sys/kernel/rtsig-max
4594 (Only in kernels up to and including 2.6.7; see
4596 This file can be used to tune the maximum number
4597 of POSIX real-time (queued) signals that can be outstanding
4600 .I /proc/sys/kernel/rtsig-nr
4601 (Only in kernels up to and including 2.6.7.)
4602 This file shows the number of POSIX real-time signals currently queued.
4604 .IR /proc/[pid]/sched_autogroup_enabled " (since Linux 2.6.38)"
4605 .\" commit 5091faa449ee0b7d73bc296a93bca9540fc51d0a
4609 .IR /proc/sys/kernel/sched_child_runs_first " (since Linux 2.6.23)"
4610 If this file contains the value zero, then, after a
4612 the parent is first scheduled on the CPU.
4613 If the file contains a nonzero value,
4614 then the child is scheduled first on the CPU.
4615 (Of course, on a multiprocessor system,
4616 the parent and the child might both immediately be scheduled on a CPU.)
4618 .IR /proc/sys/kernel/sched_rr_timeslice_ms " (since Linux 3.9)"
4620 .BR sched_rr_get_interval (2).
4622 .IR /proc/sys/kernel/sched_rt_period_us " (since Linux 2.6.25)"
4626 .IR /proc/sys/kernel/sched_rt_runtime_us " (since Linux 2.6.25)"
4630 .IR /proc/sys/kernel/sem " (since Linux 2.4)"
4631 This file contains 4 numbers defining limits for System V IPC semaphores.
4632 These fields are, in order:
4635 The maximum semaphores per semaphore set.
4637 A system-wide limit on the number of semaphores in all semaphore sets.
4639 The maximum number of operations that may be specified in a
4643 A system-wide limit on the maximum number of semaphore identifiers.
4646 .I /proc/sys/kernel/sg-big-buff
4648 shows the size of the generic SCSI device (sg) buffer.
4649 You can't tune it just yet, but you could change it at
4650 compile time by editing
4651 .I include/scsi/sg.h
4655 However, there shouldn't be any reason to change this value.
4657 .IR /proc/sys/kernel/shm_rmid_forced " (since Linux 3.1)"
4658 .\" commit b34a6b1da371ed8af1221459a18c67970f7e3d53
4659 .\" See also Documentation/sysctl/kernel.txt
4660 If this file is set to 1, all System V shared memory segments will
4661 be marked for destruction as soon as the number of attached processes
4663 in other words, it is no longer possible to create shared memory segments
4664 that exist independently of any attached process.
4666 The effect is as though a
4669 is performed on all existing segments as well as all segments
4670 created in the future (until this file is reset to 0).
4671 Note that existing segments that are attached to no process will be
4672 immediately destroyed when this file is set to 1.
4673 Setting this option will also destroy segments that were created,
4675 upon termination of the process that created the segment with
4678 Setting this file to 1 provides a way of ensuring that
4679 all System V shared memory segments are counted against the
4680 resource usage and resource limits (see the description of
4684 of at least one process.
4686 Because setting this file to 1 produces behavior that is nonstandard
4687 and could also break existing applications,
4688 the default value in this file is 0.
4689 Set this file to 1 only if you have a good understanding
4690 of the semantics of the applications using
4691 System V shared memory on your system.
4693 .IR /proc/sys/kernel/shmall " (since Linux 2.2)"
4695 contains the system-wide limit on the total number of pages of
4696 System V shared memory.
4698 .IR /proc/sys/kernel/shmmax " (since Linux 2.2)"
4700 can be used to query and set the run-time limit
4701 on the maximum (System V IPC) shared memory segment size that can be
4703 Shared memory segments up to 1GB are now supported in the
4705 This value defaults to
4708 .IR /proc/sys/kernel/shmmni " (since Linux 2.4)"
4710 specifies the system-wide maximum number of System V shared memory
4711 segments that can be created.
4713 .IR /proc/sys/kernel/sysctl_writes_strict " (since Linux 3.16)"
4714 .\" commit f88083005ab319abba5d0b2e4e997558245493c8
4715 .\" commit 2ca9bb456ada8bcbdc8f77f8fc78207653bbaa92
4716 .\" commit f4aacea2f5d1a5f7e3154e967d70cf3f711bcd61
4717 .\" commit 24fe831c17ab8149413874f2fd4e5c8a41fcd294
4718 The value in this file determines how the file offset affects
4719 the behavior of updating entries in files under
4721 The file has three possible values:
4725 This provides legacy handling, with no printk warnings.
4728 must fully contain the value to be written,
4729 and multiple writes on the same file descriptor
4730 will overwrite the entire value, regardless of the file position.
4733 (default) This provides the same behavior as for \-1,
4734 but printk warnings are written for processes that
4735 perform writes when the file offset is not 0.
4738 Respect the file offset when writing strings into
4741 Multiple writes will
4743 to the value buffer.
4744 Anything written beyond the maximum length
4745 of the value buffer will be ignored.
4748 entries must always be at file offset 0 and the value must be
4749 fully contained in the buffer provided to
4752 .\" With /proc/sys/kernel/sysctl_writes_strict==1, writes at an
4753 .\" offset other than 0 do not generate an error. Instead, the
4754 .\" write() succeeds, but the file is left unmodified.
4755 .\" This is surprising. The behavior may change in the future.
4756 .\" See thread.gmane.org/gmane.linux.man/9197
4757 .\" From: Michael Kerrisk (man-pages <mtk.manpages@...>
4758 .\" Subject: sysctl_writes_strict documentation + an oddity?
4759 .\" Newsgroups: gmane.linux.man, gmane.linux.kernel
4760 .\" Date: 2015-05-09 08:54:11 GMT
4763 .I /proc/sys/kernel/sysrq
4764 This file controls the functions allowed to be invoked by the SysRq key.
4766 the file contains 1 meaning that every possible SysRq request is allowed
4767 (in older kernel versions, SysRq was disabled by default,
4768 and you were required to specifically enable it at run-time,
4769 but this is not the case any more).
4770 Possible values in this file are:
4774 Disable sysrq completely
4777 Enable all functions of sysrq
4780 Bit mask of allowed sysrq functions, as follows:
4785 Enable control of console logging level
4788 Enable control of keyboard (SAK, unraw)
4791 Enable debugging dumps of processes etc.
4797 Enable remount read-only
4800 Enable signaling of processes (term, kill, oom-kill)
4803 Allow reboot/poweroff
4806 Allow nicing of all real-time tasks
4811 This file is present only if the
4812 .B CONFIG_MAGIC_SYSRQ
4813 kernel configuration option is enabled.
4814 For further details see the Linux kernel source file
4815 .IR Documentation/sysrq.txt .
4817 .I /proc/sys/kernel/version
4818 This file contains a string such as:
4820 #5 Wed Feb 25 21:49:24 MET 1998
4823 this is the fifth kernel built from this source base and the
4824 date following it indicates the time the kernel was built.
4826 .IR /proc/sys/kernel/threads-max " (since Linux 2.3.11)"
4827 .\" The following is based on Documentation/sysctl/kernel.txt
4828 This file specifies the system-wide limit on the number of
4829 threads (tasks) that can be created on the system.
4832 .\" commit 230633d109e35b0a24277498e773edeb79b4a331
4833 the value that can be written to
4836 The minimum value that can be written is 20.
4837 The maximum value that can be written is given by the
4841 If a value outside of this range is written to
4847 The value written is checked against the available RAM pages.
4848 If the thread structures would occupy too much (more than 1/8th)
4849 of the available RAM pages,
4851 is reduced accordingly.
4853 .IR /proc/sys/kernel/yama/ptrace_scope " (since Linux 3.5)"
4857 .IR /proc/sys/kernel/zero-paged " (PowerPC only) "
4860 When enabled (nonzero), Linux-PPC will pre-zero pages in
4861 the idle loop, possibly speeding up get_free_pages.
4864 This directory contains networking stuff.
4865 Explanations for some of the files under this directory can be found in
4870 .B /proc/sys/net/core/bpf_jit_enable
4874 .I /proc/sys/net/core/somaxconn
4875 This file defines a ceiling value for the
4881 manual page for details.
4884 This directory may be empty.
4887 This directory supports Sun remote procedure call for network filesystem
4889 On some systems, it is not present.
4891 .IR /proc/sys/user " (since Linux 4.9)"
4896 This directory contains files for memory management tuning, buffer and
4899 .IR /proc/sys/vm/admin_reserve_kbytes " (since Linux 3.10)"
4900 .\" commit 4eeab4f5580d11bffedc697684b91b0bca0d5009
4901 This file defines the amount of free memory (in KiB) on the system that
4902 that should be reserved for users with the capability
4905 The default value in this file is the minimum of [3% of free pages, 8MiB]
4907 The default is intended to provide enough for the superuser
4908 to log in and kill a process, if necessary,
4909 under the default overcommit 'guess' mode (i.e., 0 in
4910 .IR /proc/sys/vm/overcommit_memory ).
4912 Systems running in "overcommit never" mode (i.e., 2 in
4913 .IR /proc/sys/vm/overcommit_memory )
4914 should increase the value in this file to account
4915 for the full virtual memory size of the programs used to recover (e.g.,
4920 Otherwise, the superuser may not be able to log in to recover the system.
4921 For example, on x86_64 a suitable value is 131072 (128MiB reserved).
4923 Changing the value in this file takes effect whenever
4924 an application requests memory.
4926 .IR /proc/sys/vm/compact_memory " (since Linux 2.6.35)"
4927 When 1 is written to this file, all zones are compacted such that free
4928 memory is available in contiguous blocks where possible.
4929 The effect of this action can be seen by examining
4930 .IR /proc/buddyinfo .
4932 Present only if the kernel was configured with
4933 .BR CONFIG_COMPACTION .
4935 .IR /proc/sys/vm/drop_caches " (since Linux 2.6.16)"
4936 Writing to this file causes the kernel to drop clean caches, dentries, and
4937 inodes from memory, causing that memory to become free.
4938 This can be useful for memory management testing and
4939 performing reproducible filesystem benchmarks.
4940 Because writing to this file causes the benefits of caching to be lost,
4941 it can degrade overall system performance.
4943 To free pagecache, use:
4945 echo 1 > /proc/sys/vm/drop_caches
4947 To free dentries and inodes, use:
4949 echo 2 > /proc/sys/vm/drop_caches
4951 To free pagecache, dentries and inodes, use:
4953 echo 3 > /proc/sys/vm/drop_caches
4955 Because writing to this file is a nondestructive operation and dirty objects
4956 are not freeable, the
4961 .IR /proc/sys/vm/legacy_va_layout " (since Linux 2.6.9)"
4962 .\" The following is from Documentation/filesystems/proc.txt
4963 If nonzero, this disables the new 32-bit memory-mapping layout;
4964 the kernel will use the legacy (2.4) layout for all processes.
4966 .IR /proc/sys/vm/memory_failure_early_kill " (since Linux 2.6.32)"
4967 .\" The following is based on the text in Documentation/sysctl/vm.txt
4968 Control how to kill processes when an uncorrected memory error
4969 (typically a 2-bit error in a memory module)
4970 that cannot be handled by the kernel
4971 is detected in the background by hardware.
4972 In some cases (like the page still having a valid copy on disk),
4973 the kernel will handle the failure
4974 transparently without affecting any applications.
4975 But if there is no other up-to-date copy of the data,
4976 it will kill processes to prevent any data corruptions from propagating.
4978 The file has one of the following values:
4981 Kill all processes that have the corrupted-and-not-reloadable page mapped
4982 as soon as the corruption is detected.
4983 Note that this is not supported for a few types of pages,
4984 such as kernel internally
4985 allocated data or the swap cache, but works for the majority of user pages.
4987 Unmap the corrupted page from all processes and kill a process
4988 only if it tries to access the page.
4991 The kill is performed using a
4997 Processes can handle this if they want to; see
5001 This feature is active only on architectures/platforms with advanced machine
5002 check handling and depends on the hardware capabilities.
5004 Applications can override the
5005 .I memory_failure_early_kill
5006 setting individually with the
5011 Present only if the kernel was configured with
5012 .BR CONFIG_MEMORY_FAILURE .
5014 .IR /proc/sys/vm/memory_failure_recovery " (since Linux 2.6.32)"
5015 .\" The following is based on the text in Documentation/sysctl/vm.txt
5016 Enable memory failure recovery (when supported by the platform)
5021 Always panic on a memory failure.
5024 Present only if the kernel was configured with
5025 .BR CONFIG_MEMORY_FAILURE .
5027 .IR /proc/sys/vm/oom_dump_tasks " (since Linux 2.6.25)"
5028 .\" The following is from Documentation/sysctl/vm.txt
5029 Enables a system-wide task dump (excluding kernel threads) to be
5030 produced when the kernel performs an OOM-killing.
5031 The dump includes the following information
5032 for each task (thread, process):
5033 thread ID, real user ID, thread group ID (process ID),
5034 virtual memory size, resident set size,
5035 the CPU that the task is scheduled on,
5036 oom_adj score (see the description of
5037 .IR /proc/[pid]/oom_adj ),
5039 This is helpful to determine why the OOM-killer was invoked
5040 and to identify the rogue task that caused it.
5042 If this contains the value zero, this information is suppressed.
5043 On very large systems with thousands of tasks,
5044 it may not be feasible to dump the memory state information for each one.
5045 Such systems should not be forced to incur a performance penalty in
5046 OOM situations when the information may not be desired.
5048 If this is set to nonzero, this information is shown whenever the
5049 OOM-killer actually kills a memory-hogging task.
5051 The default value is 0.
5053 .IR /proc/sys/vm/oom_kill_allocating_task " (since Linux 2.6.24)"
5054 .\" The following is from Documentation/sysctl/vm.txt
5055 This enables or disables killing the OOM-triggering task in
5056 out-of-memory situations.
5058 If this is set to zero, the OOM-killer will scan through the entire
5059 tasklist and select a task based on heuristics to kill.
5060 This normally selects a rogue memory-hogging task that
5061 frees up a large amount of memory when killed.
5063 If this is set to nonzero, the OOM-killer simply kills the task that
5064 triggered the out-of-memory condition.
5065 This avoids a possibly expensive tasklist scan.
5068 .I /proc/sys/vm/panic_on_oom
5069 is nonzero, it takes precedence over whatever value is used in
5070 .IR /proc/sys/vm/oom_kill_allocating_task .
5072 The default value is 0.
5074 .IR /proc/sys/vm/overcommit_kbytes " (since Linux 3.14)"
5075 .\" commit 49f0ce5f92321cdcf741e35f385669a421013cb7
5076 This writable file provides an alternative to
5077 .IR /proc/sys/vm/overcommit_ratio
5081 .IR /proc/sys/vm/overcommit_memory
5083 It allows the amount of memory overcommitting to be specified as
5084 an absolute value (in kB),
5085 rather than as a percentage, as is done with
5086 .IR overcommit_ratio .
5087 This allows for finer-grained control of
5089 on systems with extremely large memory sizes.
5092 .IR overcommit_kbytes
5094 .IR overcommit_ratio
5097 .IR overcommit_kbytes
5098 has a nonzero value, then it is used to calculate
5101 .IR overcommit_ratio
5103 Writing a value to either of these files causes the
5104 value in the other file to be set to zero.
5106 .I /proc/sys/vm/overcommit_memory
5107 This file contains the kernel virtual memory accounting mode.
5111 0: heuristic overcommit (this is the default)
5113 1: always overcommit, never check
5115 2: always check, never overcommit
5122 are not checked, and the default check is very weak,
5123 leading to the risk of getting a process "OOM-killed".
5125 In mode 1, the kernel pretends there is always enough memory,
5126 until memory actually runs out.
5127 One use case for this mode is scientific computing applications
5128 that employ large sparse arrays.
5129 In Linux kernel versions before 2.6.0, any nonzero value implies mode 1.
5131 In mode 2 (available since Linux 2.6), the total virtual address space
5132 that can be allocated
5138 CommitLimit = (total_RAM - total_huge_TLB) *
5139 overcommit_ratio / 100 + total_swap
5145 is the total amount of RAM on the system;
5148 is the amount of memory set aside for huge pages;
5152 .IR /proc/sys/vm/overcommit_ratio ;
5156 is the amount of swap space.
5159 For example, on a system with 16GB of physical RAM, 16GB
5160 of swap, no space dedicated to huge pages, and an
5162 of 50, this formula yields a
5166 Since Linux 3.14, if the value in
5167 .I /proc/sys/vm/overcommit_kbytes
5170 is instead calculated as:
5172 CommitLimit = overcommit_kbytes + total_swap
5174 See also the description of
5175 .IR /proc/sys/vm/admiin_reserve_kbytes
5177 .IR /proc/sys/vm/user_reserve_kbytes .
5179 .IR /proc/sys/vm/overcommit_ratio " (since Linux 2.6.0)"
5180 This writable file defines a percentage by which memory
5181 can be overcommitted.
5182 The default value in the file is 50.
5183 See the description of
5184 .IR /proc/sys/vm/overcommit_memory .
5186 .IR /proc/sys/vm/panic_on_oom " (since Linux 2.6.18)"
5187 .\" The following is adapted from Documentation/sysctl/vm.txt
5188 This enables or disables a kernel panic in
5189 an out-of-memory situation.
5191 If this file is set to the value 0,
5192 the kernel's OOM-killer will kill some rogue process.
5193 Usually, the OOM-killer is able to kill a rogue process and the
5194 system will survive.
5196 If this file is set to the value 1,
5197 then the kernel normally panics when out-of-memory happens.
5198 However, if a process limits allocations to certain nodes
5199 using memory policies
5204 and those nodes reach memory exhaustion status,
5205 one process may be killed by the OOM-killer.
5206 No panic occurs in this case:
5207 because other nodes' memory may be free,
5208 this means the system as a whole may not have reached
5209 an out-of-memory situation yet.
5211 If this file is set to the value 2,
5212 the kernel always panics when an out-of-memory condition occurs.
5214 The default value is 0.
5215 1 and 2 are for failover of clustering.
5216 Select either according to your policy of failover.
5218 .IR /proc/sys/vm/swappiness
5219 .\" The following is from Documentation/sysctl/vm.txt
5220 The value in this file controls how aggressively the kernel will swap
5222 Higher values increase aggressiveness, lower values
5223 decrease aggressiveness.
5224 The default value is 60.
5226 .IR /proc/sys/vm/user_reserve_kbytes " (since Linux 3.10)"
5227 .\" commit c9b1d0981fcce3d9976d7b7a56e4e0503bc610dd
5228 Specifies an amount of memory (in KiB) to reserve for user processes,
5229 This is intended to prevent a user from starting a single memory hogging
5230 process, such that they cannot recover (kill the hog).
5231 The value in this file has an effect only when
5232 .IR /proc/sys/vm/overcommit_memory
5233 is set to 2 ("overcommit never" mode).
5234 In this case, the system reserves an amount of memory that is the minimum
5235 of [3% of current process size,
5236 .IR user_reserve_kbytes ].
5238 The default value in this file is the minimum of [3% of free pages, 128MiB]
5241 If the value in this file is set to zero,
5242 then a user will be allowed to allocate all free memory with a single process
5243 (minus the amount reserved by
5244 .IR /proc/sys/vm/admin_reserve_kbytes ).
5245 Any subsequent attempts to execute a command will result in
5246 "fork: Cannot allocate memory".
5248 Changing the value in this file takes effect whenever
5249 an application requests memory.
5251 .IR /proc/sysrq-trigger " (since Linux 2.4.21)"
5252 Writing a character to this file triggers the same SysRq function as
5253 typing ALT-SysRq-<character> (see the description of
5254 .IR /proc/sys/kernel/sysrq ).
5255 This file is normally writable only by
5257 For further details see the Linux kernel source file
5258 .IR Documentation/sysrq.txt .
5261 Subdirectory containing the pseudo-files
5262 .IR msg ", " sem " and " shm "."
5263 These files list the System V Interprocess Communication (IPC) objects
5264 (respectively: message queues, semaphores, and shared memory)
5265 that currently exist on the system,
5266 providing similar information to that available via
5268 These files have headers and are formatted (one IPC object per line)
5269 for easy understanding.
5271 provides further background on the information shown by these files.
5273 .IR /proc/thread-self " (since Linux 3.17)"
5274 .\" commit 0097875bd41528922fb3bb5f348c53f17e00e2fd
5275 This directory refers to the thread accessing the
5278 and is identical to the
5279 .I /proc/self/task/[tid]
5280 directory named by the process thread ID
5284 .IR /proc/timer_list " (since Linux 2.6.21)"
5285 .\" commit 289f480af87e45f7a6de6ba9b4c061c2e259fe98
5286 This read-only file exposes a list of all currently pending
5287 (high-resolution) timers,
5288 all clock-event sources, and their parameters in a human-readable form.
5290 .IR /proc/timer_stats " (since Linux 2.6.21)"
5291 .\" commit 82f67cd9fca8c8762c15ba7ed0d5747588c1e221
5292 .\" Date: Fri Feb 16 01:28:13 2007 -0800
5293 .\" Text largely derived from Documentation/timers/timer_stats.txt
5294 This is a debugging facility to make timer (ab)use in a Linux
5295 system visible to kernel and user-space developers.
5296 It can be used by kernel and user-space developers to verify that
5297 their code does not make undue use of timers.
5298 The goal is to avoid unnecessary wakeups,
5299 thereby optimizing power consumption.
5301 If enabled in the kernel
5302 .RB ( CONFIG_TIMER_STATS ),
5304 it has almost zero runtime overhead and a relatively small
5305 data-structure overhead.
5306 Even if collection is enabled at runtime, overhead is low:
5307 all the locking is per-CPU and lookup is hashed.
5310 .I /proc/timer_stats
5311 file is used both to control sampling facility and to read out the
5312 sampled information.
5314 The timer_stats functionality is inactive on bootup.
5315 A sampling period can be started using the following command:
5317 # echo 1 > /proc/timer_stats
5319 The following command stops a sampling period:
5321 # echo 0 > /proc/timer_stats
5323 The statistics can be retrieved by:
5325 $ cat /proc/timer_stats
5327 While sampling is enabled, each readout from
5328 .I /proc/timer_stats
5330 newly updated statistics.
5331 Once sampling is disabled, the sampled information
5332 is kept until a new sample period is started.
5333 This allows multiple readouts.
5336 .IR /proc/timer_stats :
5340 .RB $ " cat /proc/timer_stats"
5341 Timer Stats Version: v0.3
5342 Sample period: 1.764 s
5344 255, 0 swapper/3 hrtimer_start_range_ns (tick_sched_timer)
5345 71, 0 swapper/1 hrtimer_start_range_ns (tick_sched_timer)
5346 58, 0 swapper/0 hrtimer_start_range_ns (tick_sched_timer)
5347 4, 1694 gnome-shell mod_delayed_work_on (delayed_work_timer_fn)
5348 17, 7 rcu_sched rcu_gp_kthread (process_timeout)
5350 1, 4911 kworker/u16:0 mod_delayed_work_on (delayed_work_timer_fn)
5351 1D, 2522 kworker/0:0 queue_delayed_work_on (delayed_work_timer_fn)
5352 1029 total events, 583.333 events/sec
5356 The output columns are:
5359 a count of the number of events,
5360 optionally (since Linux 2.6.23) followed by the letter \(aqD\(aq
5361 .\" commit c5c061b8f9726bc2c25e19dec227933a13d1e6b7 deferrable timers
5362 if this is a deferrable timer;
5364 the PID of the process that initialized the timer;
5366 the name of the process that initialized the timer;
5368 the function where the timer was initialized; and
5371 the callback function that is associated with the timer.
5375 Subdirectory containing the pseudo-files and subdirectories for
5376 tty drivers and line disciplines.
5379 This file contains two numbers: the uptime of the system (seconds),
5380 and the amount of time spent in idle process (seconds).
5383 This string identifies the kernel version that is currently running.
5384 It includes the contents of
5385 .IR /proc/sys/kernel/ostype ,
5386 .I /proc/sys/kernel/osrelease
5388 .IR /proc/sys/kernel/version .
5393 Linux version 1.0.9 (quinlan@phaze) #1 Sat May 14 01:51:54 EDT 1994
5397 .\" FIXME 2.6.13 seems to have /proc/vmcore implemented; document this
5398 .\" See Documentation/kdump/kdump.txt
5399 .\" commit 666bfddbe8b8fd4fd44617d6c55193d5ac7edb29
5400 .\" Needs CONFIG_VMCORE
5403 .IR /proc/vmstat " (since Linux 2.6.0)"
5404 This file displays various virtual memory statistics.
5405 Each line of this file contains a single name-value pair,
5406 delimited by white space.
5407 Some files are present only if the kernel was configured with
5409 (In some cases, the options required for particular files have changed
5410 across kernel versions, so they are not listed here.
5411 Details can be found by consulting the kernel source code.)
5412 The following fields may be present:
5413 .\" FIXME We need explanations for each of the following fields...
5416 .IR nr_free_pages " (since Linux 2.6.31)"
5417 .\" commit d23ad42324cc4378132e51f2fc5c9ba6cbe75182
5419 .IR nr_alloc_batch " (since Linux 3.12)"
5420 .\" commit 81c0a2bb515fd4daae8cab64352877480792b515
5422 .IR nr_inactive_anon " (since Linux 2.6.28)"
5423 .\" commit 4f98a2fee8acdb4ac84545df98cccecfd130f8db
5425 .IR nr_active_anon " (since Linux 2.6.28)"
5426 .\" commit 4f98a2fee8acdb4ac84545df98cccecfd130f8db
5428 .IR nr_inactive_file " (since Linux 2.6.28)"
5429 .\" commit 4f98a2fee8acdb4ac84545df98cccecfd130f8db
5431 .IR nr_active_file " (since Linux 2.6.28)"
5432 .\" commit 4f98a2fee8acdb4ac84545df98cccecfd130f8db
5434 .IR nr_unevictable " (since Linux 2.6.28)"
5435 .\" commit 7b854121eb3e5ba0241882ff939e2c485228c9c5
5437 .IR nr_mlock " (since Linux 2.6.28)"
5438 .\" commit 5344b7e648980cc2ca613ec03a56a8222ff48820
5440 .IR nr_anon_pages " (since Linux 2.6.18)"
5441 .\" commit f3dbd34460ff54962d3e3244b6bcb7f5295356e6
5443 .IR nr_mapped " (since Linux 2.6.0)"
5445 .IR nr_file_pages " (since Linux 2.6.18)"
5446 .\" commit 347ce434d57da80fd5809c0c836f206a50999c26
5448 .IR nr_dirty " (since Linux 2.6.0)"
5450 .IR nr_writeback " (since Linux 2.6.0)"
5452 .IR nr_slab_reclaimable " (since Linux 2.6.19)"
5453 .\" commit 972d1a7b140569084439a81265a0f15b74e924e0
5454 .\" Linux 2.6.0 had nr_slab
5456 .IR nr_slab_unreclaimable " (since Linux 2.6.19)"
5457 .\" commit 972d1a7b140569084439a81265a0f15b74e924e0
5459 .IR nr_page_table_pages " (since Linux 2.6.0)"
5461 .IR nr_kernel_stack " (since Linux 2.6.32)"
5462 .\" commit c6a7f5728a1db45d30df55a01adc130b4ab0327c
5463 Amount of memory allocated to kernel stacks.
5465 .IR nr_unstable " (since Linux 2.6.0)"
5467 .IR nr_bounce " (since Linux 2.6.12)"
5468 .\" commit edfbe2b0038723e5699ab22695ccd62b5542a5c1
5470 .IR nr_vmscan_write " (since Linux 2.6.19)"
5471 .\" commit e129b5c23c2b471d47f1c5d2b8b193fc2034af43
5473 .IR nr_vmscan_immediate_reclaim " (since Linux 3.2)"
5474 .\" commit 49ea7eb65e7c5060807fb9312b1ad4c3eab82e2c
5476 .IR nr_writeback_temp " (since Linux 2.6.26)"
5477 .\" commit fc3ba692a4d19019387c5acaea63131f9eab05dd
5479 .IR nr_isolated_anon " (since Linux 2.6.32)"
5480 .\" commit a731286de62294b63d8ceb3c5914ac52cc17e690
5482 .IR nr_isolated_file " (since Linux 2.6.32)"
5483 .\" commit a731286de62294b63d8ceb3c5914ac52cc17e690
5485 .IR nr_shmem " (since Linux 2.6.32)"
5486 .\" commit 4b02108ac1b3354a22b0d83c684797692efdc395
5487 Pages used by shmem and
5490 .IR nr_dirtied " (since Linux 2.6.37)"
5491 .\" commit ea941f0e2a8c02ae876cd73deb4e1557248f258c
5493 .IR nr_written " (since Linux 2.6.37)"
5494 .\" commit ea941f0e2a8c02ae876cd73deb4e1557248f258c
5496 .IR nr_pages_scanned " (since Linux 3.17)"
5497 .\" commit 0d5d823ab4e608ec7b52ac4410de4cb74bbe0edd
5499 .IR numa_hit " (since Linux 2.6.18)"
5500 .\" commit ca889e6c45e0b112cb2ca9d35afc66297519b5d5
5501 .\" Present only if the kernel was configured with
5502 .\" .BR CONFIG_NUMA .
5504 .IR numa_miss " (since Linux 2.6.18)"
5505 .\" commit ca889e6c45e0b112cb2ca9d35afc66297519b5d5
5506 .\" Present only if the kernel was configured with
5507 .\" .BR CONFIG_NUMA .
5509 .IR numa_foreign " (since Linux 2.6.18)"
5510 .\" commit ca889e6c45e0b112cb2ca9d35afc66297519b5d5
5511 .\" Present only if the kernel was configured with
5512 .\" .BR CONFIG_NUMA .
5514 .IR numa_interleave " (since Linux 2.6.18)"
5515 .\" commit ca889e6c45e0b112cb2ca9d35afc66297519b5d5
5516 .\" Present only if the kernel was configured with
5517 .\" .BR CONFIG_NUMA .
5519 .IR numa_local " (since Linux 2.6.18)"
5520 .\" commit ca889e6c45e0b112cb2ca9d35afc66297519b5d5
5521 .\" Present only if the kernel was configured with
5522 .\" .BR CONFIG_NUMA .
5524 .IR numa_other " (since Linux 2.6.18)"
5525 .\" commit ca889e6c45e0b112cb2ca9d35afc66297519b5d5
5526 .\" Present only if the kernel was configured with
5527 .\" .BR CONFIG_NUMA .
5529 .IR workingset_refault " (since Linux 3.15)"
5530 .\" commit a528910e12ec7ee203095eb1711468a66b9b60b0
5531 .\" Present only if the kernel was configured with
5532 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5534 .IR workingset_activate " (since Linux 3.15)"
5535 .\" commit a528910e12ec7ee203095eb1711468a66b9b60b0
5536 .\" Present only if the kernel was configured with
5537 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5539 .IR workingset_nodereclaim " (since Linux 3.15)"
5540 .\" commit 449dd6984d0e47643c04c807f609dd56d48d5bcc
5541 .\" Present only if the kernel was configured with
5542 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5544 .IR nr_anon_transparent_hugepages " (since Linux 2.6.38)"
5545 .\" Present only if the kernel was configured with
5546 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5548 .IR nr_free_cma " (since Linux 3.7)"
5549 .\" commit d1ce749a0db12202b711d1aba1d29e823034648d
5550 Number of free CMA (Contiguous Memory Allocator) pages.
5551 .\" Present only if the kernel was configured with
5552 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5554 .IR nr_dirty_threshold " (since Linux 2.6.37)"
5555 .\" commit 79da826aee6a10902ef411bc65864bd02102fa83
5556 .\" Present only if the kernel was configured with
5557 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5559 .IR nr_dirty_background_threshold " (since Linux 2.6.37)"
5560 .\" commit 79da826aee6a10902ef411bc65864bd02102fa83
5561 .\" Present only if the kernel was configured with
5562 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5564 .IR pgpgin " (since Linux 2.6.0)"
5565 .\" Present only if the kernel was configured with
5566 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5568 .IR pgpgout " (since Linux 2.6.0)"
5569 .\" Present only if the kernel was configured with
5570 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5572 .IR pswpin " (since Linux 2.6.0)"
5573 .\" Present only if the kernel was configured with
5574 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5576 .IR pswpout " (since Linux 2.6.0)"
5577 .\" Present only if the kernel was configured with
5578 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5580 .IR pgalloc_dma " (since Linux 2.6.5)"
5581 .\" Linux 2.6.0 had pgalloc
5582 .\" Present only if the kernel was configured with
5583 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5585 .IR pgalloc_dma32 " (since Linux 2.6.16)"
5586 .\" commit 9328b8faae922e52073785ed6c1eaa8565648a0e
5587 .\" Present only if the kernel was configured with
5588 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5590 .IR pgalloc_normal " (since Linux 2.6.5)"
5591 .\" Present only if the kernel was configured with
5592 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5594 .IR pgalloc_high " (since Linux 2.6.5)"
5595 .\" Present only if the kernel was configured with
5596 .\" .BR CONFIG_VM_EVENT_COUNTERS
5598 .\" .BR CONFIG_HIGHMEM .
5600 .IR pgalloc_movable " (since Linux 2.6.23)"
5601 .\" commit 2a1e274acf0b1c192face19a4be7c12d4503eaaf
5602 .\" Present only if the kernel was configured with
5603 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5605 .IR pgfree " (since Linux 2.6.0)"
5606 .\" Present only if the kernel was configured with
5607 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5609 .IR pgactivate " (since Linux 2.6.0)"
5610 .\" Present only if the kernel was configured with
5611 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5613 .IR pgdeactivate " (since Linux 2.6.0)"
5614 .\" Present only if the kernel was configured with
5615 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5617 .IR pgfault " (since Linux 2.6.0)"
5618 .\" Present only if the kernel was configured with
5619 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5621 .IR pgmajfault " (since Linux 2.6.0)"
5622 .\" Present only if the kernel was configured with
5623 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5625 .IR pgrefill_dma " (since Linux 2.6.5)"
5626 .\" Linux 2.6.0 had pgrefill
5627 .\" Present only if the kernel was configured with
5628 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5630 .IR pgrefill_dma32 " (since Linux 2.6.16)"
5631 .\" commit 9328b8faae922e52073785ed6c1eaa8565648a0e
5632 .\" Present only if the kernel was configured with
5633 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5635 .IR pgrefill_normal " (since Linux 2.6.5)"
5636 .\" Present only if the kernel was configured with
5637 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5639 .IR pgrefill_high " (since Linux 2.6.5)"
5640 .\" Present only if the kernel was configured with
5641 .\" .BR CONFIG_VM_EVENT_COUNTERS
5643 .\" .BR CONFIG_HIGHMEM .
5645 .IR pgrefill_movable " (since Linux 2.6.23)"
5646 .\" commit 2a1e274acf0b1c192face19a4be7c12d4503eaaf
5647 .\" Present only if the kernel was configured with
5648 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5649 .\" Formerly there were
5654 .\" These were split out into pgsteal_kswapd* and pgsteal_direct*
5655 .\" in commit 904249aa68010c8e223263c922fcbb840a3f42e4
5657 .IR pgsteal_kswapd_dma " (since Linux 3.4)"
5658 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5659 .\" Linux 2.6.0 had pgsteal
5660 .\" Present only if the kernel was configured with
5661 .\" .\" .BR CONFIG_VM_EVENT_COUNTERS .
5663 .IR pgsteal_kswapd_dma32 " (since Linux 3.4)"
5664 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5665 .\" commit 9328b8faae922e52073785ed6c1eaa8565648a0e
5666 .\" Present only if the kernel was configured with
5667 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5669 .IR pgsteal_kswapd_normal " (since Linux 3.4)"
5670 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5671 .\" Present only if the kernel was configured with
5672 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5674 .IR pgsteal_kswapd_high " (since Linux 3.4)"
5675 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5676 .\" Present only if the kernel was configured with
5677 .\" .BR CONFIG_VM_EVENT_COUNTERS
5679 .\" .BR CONFIG_HIGHMEM .
5681 .IR pgsteal_kswapd_movable " (since Linux 3.4)"
5682 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5683 .\" Present only if the kernel was configured with
5684 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5686 .IR pgsteal_direct_dma
5687 .\" Present only if the kernel was configured with
5688 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5690 .IR pgsteal_direct_dma32 " (since Linux 3.4)"
5691 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5692 .\" Present only if the kernel was configured with
5693 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5695 .IR pgsteal_direct_normal " (since Linux 3.4)"
5696 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5697 .\" Present only if the kernel was configured with
5698 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5700 .IR pgsteal_direct_high " (since Linux 3.4)"
5701 .\" commit 904249aa68010c8e223263c922fcbb840a3f42e4
5702 .\" Present only if the kernel was configured with
5703 .\" .BR CONFIG_VM_EVENT_COUNTERS
5705 .\" .BR CONFIG_HIGHMEM .
5707 .IR pgsteal_direct_movable " (since Linux 2.6.23)"
5708 .\" commit 2a1e274acf0b1c192face19a4be7c12d4503eaaf
5709 .\" Present only if the kernel was configured with
5710 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5712 .IR pgscan_kswapd_dma
5713 .\" Linux 2.6.0 had pgscan
5714 .\" Present only if the kernel was configured with
5715 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5717 .IR pgscan_kswapd_dma32 " (since Linux 2.6.16)"
5718 .\" commit 9328b8faae922e52073785ed6c1eaa8565648a0e
5719 .\" Present only if the kernel was configured with
5720 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5722 .IR pgscan_kswapd_normal " (since Linux 2.6.5)"
5723 .\" Present only if the kernel was configured with
5724 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5726 .IR pgscan_kswapd_high
5727 .\" Present only if the kernel was configured with
5728 .\" .BR CONFIG_VM_EVENT_COUNTERS
5730 .\" .BR CONFIG_HIGHMEM .
5732 .IR pgscan_kswapd_movable " (since Linux 2.6.23)"
5733 .\" commit 2a1e274acf0b1c192face19a4be7c12d4503eaaf
5734 .\" Present only if the kernel was configured with
5735 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5737 .IR pgscan_direct_dma
5738 .\" Present only if the kernel was configured with
5739 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5741 .IR pgscan_direct_dma32 " (since Linux 2.6.16)"
5742 .\" commit 9328b8faae922e52073785ed6c1eaa8565648a0e
5743 .\" Present only if the kernel was configured with
5744 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5746 .IR pgscan_direct_normal
5747 .\" Present only if the kernel was configured with
5748 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5750 .IR pgscan_direct_high
5751 .\" Present only if the kernel was configured with
5752 .\" .BR CONFIG_VM_EVENT_COUNTERS
5754 .\" .BR CONFIG_HIGHMEM .
5756 .IR pgscan_direct_movable " (since Linux 2.6.23)"
5757 .\" commit 2a1e274acf0b1c192face19a4be7c12d4503eaaf
5758 .\" Present only if the kernel was configured with
5759 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5761 .IR pgscan_direct_throttle " (since Linux 3.6)"
5762 .\" commit 68243e76ee343d63c6cf76978588a885951e2818
5763 .\" Present only if the kernel was configured with
5764 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5766 .IR zone_reclaim_failed " (since linux 2.6.31)"
5767 .\" commit 24cf72518c79cdcda486ed26074ff8151291cf65
5768 .\" Present only if the kernel was configured with
5769 .\" .BR CONFIG_VM_EVENT_COUNTERS
5771 .\" .BR CONFIG_NUMA .
5773 .IR pginodesteal " (since linux 2.6.0)"
5774 .\" Present only if the kernel was configured with
5775 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5777 .IR slabs_scanned " (since linux 2.6.5)"
5778 .\" Present only if the kernel was configured with
5779 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5781 .IR kswapd_inodesteal " (since linux 2.6.0)"
5782 .\" Present only if the kernel was configured with
5783 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5785 .IR kswapd_low_wmark_hit_quickly " (since 2.6.33)"
5786 .\" commit bb3ab596832b920c703d1aea1ce76d69c0f71fb7
5787 .\" Present only if the kernel was configured with
5788 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5790 .IR kswapd_high_wmark_hit_quickly " (since 2.6.33)"
5791 .\" commit bb3ab596832b920c703d1aea1ce76d69c0f71fb7
5792 .\" Present only if the kernel was configured with
5793 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5795 .IR pageoutrun " (since Linux 2.6.0)"
5796 .\" Present only if the kernel was configured with
5797 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5799 .IR allocstall " (since Linux 2.6.0)"
5800 .\" Present only if the kernel was configured with
5801 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5803 .IR pgrotated " (since Linux 2.6.0)"
5804 .\" Present only if the kernel was configured with
5805 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5807 .IR drop_pagecache " (since Linux 3.15)"
5808 .\" commit 5509a5d27b971a90b940e148ca9ca53312e4fa7a
5809 .\" Present only if the kernel was configured with
5810 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5812 .IR drop_slab " (since Linux 3.15)"
5813 .\" commit 5509a5d27b971a90b940e148ca9ca53312e4fa7a
5814 .\" Present only if the kernel was configured with
5815 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5817 .IR numa_pte_updates " (since Linux 3.8)"
5818 .\" commit 03c5a6e16322c997bf8f264851bfa3f532ad515f
5819 .\" Present only if the kernel was configured with
5820 .\" .BR CONFIG_VM_EVENT_COUNTERS
5822 .\" .BR CONFIG_NUMA_BALANCING .
5824 .IR numa_huge_pte_updates " (since Linux 3.13)"
5825 .\" commit 72403b4a0fbdf433c1fe0127e49864658f6f6468
5826 .\" Present only if the kernel was configured with
5827 .\" .BR CONFIG_VM_EVENT_COUNTERS
5829 .\" .BR CONFIG_NUMA_BALANCING .
5831 .IR numa_hint_faults " (since Linux 3.8)"
5832 .\" commit 03c5a6e16322c997bf8f264851bfa3f532ad515f
5833 .\" Present only if the kernel was configured with
5834 .\" .BR CONFIG_VM_EVENT_COUNTERS
5836 .\" .BR CONFIG_NUMA_BALANCING .
5838 .IR numa_hint_faults_local " (since Linux 3.8)"
5839 .\" commit 03c5a6e16322c997bf8f264851bfa3f532ad515f
5840 .\" Present only if the kernel was configured with
5841 .\" .BR CONFIG_VM_EVENT_COUNTERS
5843 .\" .BR CONFIG_NUMA_BALANCING .
5845 .IR numa_pages_migrated " (since Linux 3.8)"
5846 .\" commit 03c5a6e16322c997bf8f264851bfa3f532ad515f
5847 .\" Present only if the kernel was configured with
5848 .\" .BR CONFIG_VM_EVENT_COUNTERS
5850 .\" .BR CONFIG_NUMA_BALANCING
5852 .\" .BR CONFIG_NUMA_BALANCING .
5854 .IR pgmigrate_success " (since Linux 3.8)"
5855 .\" commit 5647bc293ab15f66a7b1cda850c5e9d162a6c7c2
5856 .\" Present only if the kernel was configured with
5857 .\" .BR CONFIG_VM_EVENT_COUNTERS
5859 .\" .BR CONFIG_MIGRATION .
5861 .IR pgmigrate_fail " (since Linux 3.8)"
5862 .\" commit 5647bc293ab15f66a7b1cda850c5e9d162a6c7c2
5863 .\" Present only if the kernel was configured with
5864 .\" .BR CONFIG_VM_EVENT_COUNTERS
5866 .\" .BR CONFIG_MIGRATION .
5868 .IR compact_migrate_scanned " (since Linux 3.8)"
5869 .\" commit 397487db696cae0b026a474a5cd66f4e372995e6
5870 .\" Linux 3.8 dropped compact_blocks_moved, compact_pages_moved, and
5871 .\" compact_pagemigrate_failed
5872 .\" Present only if the kernel was configured with
5873 .\" .BR CONFIG_VM_EVENT_COUNTERS
5875 .\" .BR CONFIG_COMPACTION .
5877 .IR compact_free_scanned " (since Linux 3.8)"
5878 .\" commit 397487db696cae0b026a474a5cd66f4e372995e6
5879 .\" Present only if the kernel was configured with
5880 .\" .BR CONFIG_VM_EVENT_COUNTERS
5882 .\" .BR CONFIG_COMPACTION .
5884 .IR compact_isolated " (since Linux 3.8)"
5885 .\" commit 397487db696cae0b026a474a5cd66f4e372995e6
5886 .\" Present only if the kernel was configured with
5887 .\" .BR CONFIG_VM_EVENT_COUNTERS
5889 .\" .BR CONFIG_COMPACTION .
5891 .IR compact_stall " (since Linux 2.6.35)"
5892 .\" commit 56de7263fcf3eb10c8dcdf8d59a9cec831795f3f
5893 See the kernel source file
5894 .IR Documentation/vm/transhuge.txt .
5895 .\" Present only if the kernel was configured with
5896 .\" .BR CONFIG_VM_EVENT_COUNTERS
5898 .\" .BR CONFIG_COMPACTION .
5900 .IR compact_fail " (since Linux 2.6.35)"
5901 .\" commit 56de7263fcf3eb10c8dcdf8d59a9cec831795f3f
5902 See the kernel source file
5903 .IR Documentation/vm/transhuge.txt .
5904 .\" Present only if the kernel was configured with
5905 .\" .BR CONFIG_VM_EVENT_COUNTERS
5907 .\" .BR CONFIG_COMPACTION .
5909 .IR compact_success " (since Linux 2.6.35)"
5910 .\" commit 56de7263fcf3eb10c8dcdf8d59a9cec831795f3f
5911 See the kernel source file
5912 .IR Documentation/vm/transhuge.txt .
5913 .\" Present only if the kernel was configured with
5914 .\" .BR CONFIG_VM_EVENT_COUNTERS
5916 .\" .BR CONFIG_COMPACTION .
5918 .IR htlb_buddy_alloc_success " (since Linux 2.6.26)"
5919 .\" commit 3b1163006332302117b1b2acf226d4014ff46525
5920 .\" Present only if the kernel was configured with
5921 .\" .BR CONFIG_VM_EVENT_COUNTERS
5923 .\" .BR CONFIG_HUGETLB_PAGE .
5925 .IR htlb_buddy_alloc_fail " (since Linux 2.6.26)"
5926 .\" commit 3b1163006332302117b1b2acf226d4014ff46525
5927 .\" Present only if the kernel was configured with
5928 .\" .BR CONFIG_VM_EVENT_COUNTERS
5930 .\" .BR CONFIG_HUGETLB_PAGE .
5932 .IR unevictable_pgs_culled " (since Linux 2.6.28)"
5933 .\" commit bbfd28eee9fbd73e780b19beb3dc562befbb94fa
5934 .\" Present only if the kernel was configured with
5935 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5937 .IR unevictable_pgs_scanned " (since Linux 2.6.28)"
5938 .\" commit bbfd28eee9fbd73e780b19beb3dc562befbb94fa
5939 .\" Present only if the kernel was configured with
5940 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5942 .IR unevictable_pgs_rescued " (since Linux 2.6.28)"
5943 .\" commit bbfd28eee9fbd73e780b19beb3dc562befbb94fa
5944 .\" Present only if the kernel was configured with
5945 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5947 .IR unevictable_pgs_mlocked " (since Linux 2.6.28)"
5948 .\" commit 5344b7e648980cc2ca613ec03a56a8222ff48820
5949 .\" Present only if the kernel was configured with
5950 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5952 .IR unevictable_pgs_munlocked " (since Linux 2.6.28)"
5953 .\" commit 5344b7e648980cc2ca613ec03a56a8222ff48820
5954 .\" Present only if the kernel was configured with
5955 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5957 .IR unevictable_pgs_cleared " (since Linux 2.6.28)"
5958 .\" commit 5344b7e648980cc2ca613ec03a56a8222ff48820
5959 .\" Present only if the kernel was configured with
5960 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5962 .IR unevictable_pgs_stranded " (since Linux 2.6.28)"
5963 .\" commit 5344b7e648980cc2ca613ec03a56a8222ff48820
5964 .\" Present only if the kernel was configured with
5965 .\" .BR CONFIG_VM_EVENT_COUNTERS .
5966 .\" Linux 3.7 removed unevictable_pgs_mlockfreed
5968 .IR thp_fault_alloc " (since Linux 2.6.39)"
5969 .\" commit 81ab4201fb7d91d6b0cd9ad5b4b16776e4bed145
5970 See the kernel source file
5971 .IR Documentation/vm/transhuge.txt .
5972 .\" Present only if the kernel was configured with
5973 .\" .BR CONFIG_VM_EVENT_COUNTERS
5975 .\" .BR CONFIG_TRANSPARENT_HUGEPAGE .
5977 .IR thp_fault_fallback " (since Linux 2.6.39)"
5978 .\" commit 81ab4201fb7d91d6b0cd9ad5b4b16776e4bed145
5979 See the kernel source file
5980 .IR Documentation/vm/transhuge.txt .
5981 .\" Present only if the kernel was configured with
5982 .\" .BR CONFIG_VM_EVENT_COUNTERS
5984 .\" .BR CONFIG_TRANSPARENT_HUGEPAGE .
5986 .IR thp_collapse_alloc " (since Linux 2.6.39)"
5987 .\" commit 81ab4201fb7d91d6b0cd9ad5b4b16776e4bed145
5988 See the kernel source file
5989 .IR Documentation/vm/transhuge.txt .
5990 .\" Present only if the kernel was configured with
5991 .\" .BR CONFIG_VM_EVENT_COUNTERS
5993 .\" .BR CONFIG_TRANSPARENT_HUGEPAGE .
5995 .IR thp_collapse_alloc_failed " (since Linux 2.6.39)"
5996 .\" commit 81ab4201fb7d91d6b0cd9ad5b4b16776e4bed145
5997 See the kernel source file
5998 .IR Documentation/vm/transhuge.txt .
5999 .\" Present only if the kernel was configured with
6000 .\" .BR CONFIG_VM_EVENT_COUNTERS
6002 .\" .BR CONFIG_TRANSPARENT_HUGEPAGE .
6004 .IR thp_split " (since Linux 2.6.39)"
6005 .\" commit 81ab4201fb7d91d6b0cd9ad5b4b16776e4bed145
6006 See the kernel source file
6007 .IR Documentation/vm/transhuge.txt .
6008 .\" Present only if the kernel was configured with
6009 .\" .BR CONFIG_VM_EVENT_COUNTERS
6011 .\" .BR CONFIG_TRANSPARENT_HUGEPAGE .
6013 .IR thp_zero_page_alloc " (since Linux 3.8)"
6014 .\" commit d8a8e1f0da3d29d7268b3300c96a059d63901b76
6015 See the kernel source file
6016 .IR Documentation/vm/transhuge.txt .
6017 .\" Present only if the kernel was configured with
6018 .\" .BR CONFIG_VM_EVENT_COUNTERS
6020 .\" .BR CONFIG_TRANSPARENT_HUGEPAGE .
6022 .IR thp_zero_page_alloc_failed " (since Linux 3.8)"
6023 .\" commit d8a8e1f0da3d29d7268b3300c96a059d63901b76
6024 See the kernel source file
6025 .IR Documentation/vm/transhuge.txt .
6026 .\" Present only if the kernel was configured with
6027 .\" .BR CONFIG_VM_EVENT_COUNTERS
6029 .\" .BR CONFIG_TRANSPARENT_HUGEPAGE .
6031 .IR balloon_inflate " (since Linux 3.18)"
6032 .\" commit 09316c09dde33aae14f34489d9e3d243ec0d5938
6033 .\" Present only if the kernel was configured with
6034 .\" .BR CONFIG_VM_EVENT_COUNTERS
6036 .\" .BR CONFIG_MEMORY_BALLOON .
6038 .IR balloon_deflate " (since Linux 3.18)"
6039 .\" commit 09316c09dde33aae14f34489d9e3d243ec0d5938
6040 .\" Present only if the kernel was configured with
6041 .\" .BR CONFIG_VM_EVENT_COUNTERS
6043 .\" .BR CONFIG_MEMORY_BALLOON .
6045 .IR balloon_migrate " (since Linux 3.18)"
6046 .\" commit 09316c09dde33aae14f34489d9e3d243ec0d5938
6047 .\" Present only if the kernel was configured with
6048 .\" .BR CONFIG_VM_EVENT_COUNTERS ,
6049 .\" .BR CONFIG_MEMORY_BALLOON ,
6051 .\" .BR CONFIG_BALLOON_COMPACTION .
6053 .IR nr_tlb_remote_flush " (since Linux 3.12)"
6054 .\" commit 9824cf9753ecbe8f5b47aa9b2f218207defea211
6055 .\" Present only if the kernel was configured with
6056 .\" .BR CONFIG_DEBUG_TLBFLUSH
6058 .\" .BR CONFIG_SMP .
6060 .IR nr_tlb_remote_flush_received " (since Linux 3.12)"
6061 .\" commit 9824cf9753ecbe8f5b47aa9b2f218207defea211
6062 .\" Present only if the kernel was configured with
6063 .\" .BR CONFIG_DEBUG_TLBFLUSH
6065 .\" .BR CONFIG_SMP .
6067 .IR nr_tlb_local_flush_all " (since Linux 3.12)"
6068 .\" commit 9824cf9753ecbe8f5b47aa9b2f218207defea211
6069 .\" Present only if the kernel was configured with
6070 .\" .BR CONFIG_DEBUG_TLBFLUSH .
6072 .IR nr_tlb_local_flush_one " (since Linux 3.12)"
6073 .\" commit 9824cf9753ecbe8f5b47aa9b2f218207defea211
6074 .\" Present only if the kernel was configured with
6075 .\" .BR CONFIG_DEBUG_TLBFLUSH .
6077 .IR vmacache_find_calls " (since Linux 3.16)"
6078 .\" commit 4f115147ff802267d0aa41e361c5aa5bd933d896
6079 .\" Present only if the kernel was configured with
6080 .\" .BR CONFIG_DEBUG_VM_VMACACHE .
6082 .IR vmacache_find_hits " (since Linux 3.16)"
6083 .\" commit 4f115147ff802267d0aa41e361c5aa5bd933d896
6084 .\" Present only if the kernel was configured with
6085 .\" .BR CONFIG_DEBUG_VM_VMACACHE .
6087 .IR vmacache_full_flushes " (since Linux 3.19)"
6088 .\" commit f5f302e21257ebb0c074bbafc37606c26d28cc3d
6089 .\" Present only if the kernel was configured with
6090 .\" .BR CONFIG_DEBUG_VM_VMACACHE .
6093 .IR /proc/zoneinfo " (since Linux 2.6.13)"
6094 This file display information about memory zones.
6095 This is useful for analyzing virtual memory behavior.
6096 .\" FIXME more should be said about /proc/zoneinfo
6098 Many strings (i.e., the environment and command line) are in
6099 the internal format, with subfields terminated by null bytes (\(aq\\0\(aq),
6101 may find that things are more readable if you use \fIod \-c\fP or \fItr
6102 "\\000" "\\n"\fP to read them.
6103 Alternatively, \fIecho \`cat <file>\`\fP works well.
6105 This manual page is incomplete, possibly inaccurate, and is the kind
6106 of thing that needs to be updated very often.
6107 .\" .SH ACKNOWLEDGEMENTS
6108 .\" The material on /proc/sys/fs and /proc/sys/kernel is closely based on
6109 .\" kernel source documentation files written by Rik van Riel.
6138 The Linux kernel source files:
6139 .IR Documentation/filesystems/proc.txt
6140 .IR Documentation/sysctl/fs.txt ,
6141 .IR Documentation/sysctl/kernel.txt ,
6142 .IR Documentation/sysctl/net.txt ,
6144 .IR Documentation/sysctl/vm.txt .