1PROC(5) Linux Programmer's Manual PROC(5)
2
6 proc - process information pseudo-filesystem
7
10 The proc filesystem is a pseudo-filesystem which is used as an inter‐
11 face to kernel data structures. It is commonly mounted at /proc. Most
12 of it is read-only, but some files allow kernel variables to be
13 changed.
14 The following outline gives a quick tour through the /proc hierarchy.
16 /proc/[number]
18 There is a numerical subdirectory for each running process; the
19 subdirectory is named by the process ID. Each such subdirectory
20 contains the following pseudo-files and directories.
21 /proc/[number]/auxv (since 2.6.0-test7)
23 This contains the contents of the ELF interpreter information
24 passed to the process at exec time. The format is one unsigned
25 long ID plus one unsigned long value for each entry. The last
26 entry contains two zeros.
27 /proc/[number]/cmdline
29 This holds the complete command line for the process, unless the
30 whole process has been swapped out or the process is a zombie.
31 In either of these latter cases, there is nothing in this file:
32 i.e. a read on this file will return 0 characters. The command
33 line arguments appear in this file as a set of null-separated
34 strings, with a further null byte after the last string.
35 /proc/[number]/cwd
37 This is a symbolic link to the current working directory of the
38 process. To find out the cwd of process 20, for instance, you
39 can do this:
40 cd /proc/20/cwd; /bin/pwd
42 Note that the pwd command is often a shell builtin, and might
44 not work properly. In bash, you may use pwd -P.
45 In a multithreaded process, the contents of this symbolic link
47 are not available if the main thread has already terminated
48 (typically by calling pthread_exit(3).
49 /proc/[number]/environ
51 This file contains the environment for the process. The entries
52 are separated by null bytes ('\0'), and there may be a null
53 bytes at the end. Thus, to print out the environment of process
54 1, you would do:
55 (cat /proc/1/environ; echo) | tr "\000" "\n"
57 (For a reason why one should want to do this, see lilo(8).)
59 /proc/[number]/exe
61 Under Linux 2.2 and later, this file is a symbolic link contain‐
62 ing the actual pathname of the executed command. This symbolic
63 link can be dereferenced normally; attempting to open it will
64 open the executable. You can even type /proc/[number]/exe to
65 run another copy of the same executable as is being run by
66 process [number]. In a multithreaded process, the contents of
67 this symbolic link are not available if the main thread has
68 already terminated (typically by calling pthread_exit(3)).
69 Under Linux 2.0 and earlier /proc/[number]/exe is a pointer to
71 the binary which was executed, and appears as a symbolic link. A
72 readlink(2) call on this file under Linux 2.0 returns a string
73 in the format:
74 [device]:inode
76 For example, [0301]:1502 would be inode 1502 on device major 03
78 (IDE, MFM, etc. drives) minor 01 (first partition on the first
79 drive).
80 find(1) with the -inum option can be used to locate the file.
82 /proc/[number]/fd
84 This is a subdirectory containing one entry for each file which
85 the process has open, named by its file descriptor, and which is
86 a symbolic link to the actual file. Thus, 0 is standard input,
87 1 standard output, 2 standard error, etc.
88 In a multithreaded process, the contents of this directory are
90 not available if the main thread has already terminated (typi‐
91 cally by calling pthread_exit(3)).
92 Programs that will take a filename, but will not take the stan‐
94 dard input, and which write to a file, but will not send their
95 output to standard output, can be effectively foiled this way,
96 assuming that -i is the flag designating an input file and -o is
97 the flag designating an output file:
98 foobar -i /proc/self/fd/0 -o /proc/self/fd/1 ...
100 and you have a working filter.
102 /proc/self/fd/N is approximately the same as /dev/fd/N in some
104 UNIX and UNIX-like systems. Most Linux MAKEDEV scripts symboli‐
105 cally link /dev/fd to /proc/self/fd, in fact.
106 /proc/[number]/maps
108 A file containing the currently mapped memory regions and their
109 access permissions.
110 The format is:
112 address perms offset dev inode pathname
114 08048000-08056000 r-xp 00000000 03:0c 64593 /usr/sbin/gpm
115 08056000-08058000 rw-p 0000d000 03:0c 64593 /usr/sbin/gpm
116 08058000-0805b000 rwxp 00000000 00:00 0
117 40000000-40013000 r-xp 00000000 03:0c 4165 /lib/ld-2.2.4.so
118 40013000-40015000 rw-p 00012000 03:0c 4165 /lib/ld-2.2.4.so
119 4001f000-40135000 r-xp 00000000 03:0c 45494 /lib/libc-2.2.4.so
120 40135000-4013e000 rw-p 00115000 03:0c 45494 /lib/libc-2.2.4.so
121 4013e000-40142000 rw-p 00000000 00:00 0
122 bffff000-c0000000 rwxp 00000000 00:00 0
123 where address is the address space in the process that it occu‐
125 pies, perms is a set of permissions:
126 r = read
128 w = write
129 x = execute
130 s = shared
131 p = private (copy on write)
132 offset is the offset into the file/whatever, dev is the device
134 (major:minor), and inode is the inode on that device. 0 indi‐
135 cates that no inode is associated with the memory region, as the
136 case would be with bss.
137 Under Linux 2.0 there is no field giving pathname.
139 /proc/[number]/mem
141 This file can be used to access the pages of a process's memory
142 through open(2), read(2), and lseek(2).
143 /proc/[number]/root
145 Unix and Linux support the idea of a per-process root of the
146 filesystem, set by the chroot(2) system call. This file is a
147 symbolic link that points to the process's root directory, and
148 behaves as exe, fd/*, etc. do.
149 In a multithreaded process, the contents of this symbolic link
151 are not available if the main thread has already terminated
152 (typically by calling pthread_exit(3)).
153 /proc/[number]/smaps (since Linux 2.6.14)
155 This file shows memory consumption for each of the process's
156 mappings. For each of mappings there is a series of lines as
157 follows:
158 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash
160 Size: 464 kB
161 Rss: 424 kB
162 Shared_Clean: 424 kB
163 Shared_Dirty: 0 kB
164 Private_Clean: 0 kB
165 Private_Dirty: 0 kB
166 The first of these lines shows the same information as is dis‐
168 played for the mapping in /proc/[number]/maps. The remaining
169 lines show the size of the mapping, the amount of the mapping
170 that is currently resident in RAM, the number clean and dirty
171 shared pages in the mapping, and the number clean and dirty pri‐
172 vate pages in the mapping.
173 This file is only present if the CONFIG_MMU kernel configuration
175 option is enabled.
176 /proc/[number]/stat
178 Status information about the process. This is used by ps(1).
179 It is defined in /usr/src/linux/fs/proc/array.c.
180 The fields, in order, with their proper scanf(3) format speci‐
182 fiers, are:
183 pid %d The process ID.
185 comm %s
187 The filename of the executable, in parentheses. This is
188 visible whether or not the executable is swapped out.
189 state %c
191 One character from the string "RSDZTW" where R is run‐
192 ning, S is sleeping in an interruptible wait, D is wait‐
193 ing in uninterruptible disk sleep, Z is zombie, T is
194 traced or stopped (on a signal), and W is paging.
195 ppid %d
197 The PID of the parent.
198 pgrp %d
200 The process group ID of the process.
201 session %d
203 The session ID of the process.
204 tty_nr %d
206 The tty the process uses.
207 tpgid %d
209 The process group ID of the process which currently owns
210 the tty that the process is connected to.
211 flags %lu
213 The kernel flags word of the process. For bit meanings,
214 see the PF_* defines in <linux/sched.h>. Details depend
215 on the kernel version.
216 minflt %lu
218 The number of minor faults the process has made which
219 have not required loading a memory page from disk.
220 cminflt %lu
222 The number of minor faults that the process's waited-for
223 children have made.
224 majflt %lu
226 The number of major faults the process has made which
227 have required loading a memory page from disk.
228 cmajflt %lu
230 The number of major faults that the process's waited-for
231 children have made.
232 utime %lu
234 The number of jiffies that this process has been sched‐
235 uled in user mode.
236 stime %lu
238 The number of jiffies that this process has been sched‐
239 uled in kernel mode.
240 cutime %ld
242 The number of jiffies that this process's waited-for
243 children have been scheduled in user mode. (See also
244 times(2).)
245 cstime %ld
247 The number of jiffies that this process's waited-for
248 children have been scheduled in kernel mode.
249 priority %ld
251 The standard nice value, plus fifteen. The value is
252 never negative in the kernel.
253 nice %ld
255 The nice value ranges from 19 (nicest) to -19 (not nice
256 to others).
257 0 %ld This value is hard coded to 0 as a placeholder for a
259 removed field.
260 itrealvalue %ld
262 The time in jiffies before the next SIGALRM is sent to
263 the process due to an interval timer.
264 starttime %lu
266 The time in jiffies the process started after system
267 boot.
268 vsize %lu
270 Virtual memory size in bytes.
271 rss %ld
273 Resident Set Size: number of pages the process has in
274 real memory, minus 3 for administrative purposes. This is
275 just the pages which count towards text, data, or stack
276 space. This does not include pages which have not been
277 demand-loaded in, or which are swapped out.
278 rlim %lu
280 Current limit in bytes on the rss of the process (usually
281 4294967295 on i386).
282 startcode %lu
284 The address above which program text can run.
285 endcode %lu
287 The address below which program text can run.
288 startstack %lu
290 The address of the start of the stack.
291 kstkesp %lu
293 The current value of esp (stack pointer), as found in the
294 kernel stack page for the process.
295 kstkeip %lu
297 The current EIP (instruction pointer).
298 signal %lu
300 The bitmap of pending signals.
301 blocked %lu
303 The bitmap of blocked signals.
304 sigignore %lu
306 The bitmap of ignored signals.
307 sigcatch %lu
309 The bitmap of caught signals.
310 wchan %lu
312 This is the "channel" in which the process is waiting.
313 It is the address of a system call, and can be looked up
314 in a namelist if you need a textual name. (If you have
315 an up-to-date /etc/psdatabase, then try ps -l to see the
316 WCHAN field in action.)
317 nswap %lu
319 Number of pages swapped (not maintained).
320 cnswap %lu
322 Cumulative nswap for child processes (not maintained).
323 exit_signal %d
325 Signal to be sent to parent when we die.
326 processor %d
328 CPU number last executed on.
329 rt_priority %lu (since kernel 2.5.19)
331 Real-time scheduling priority (see sched_setsched‐
332 uler(2)).
333 policy %lu (since kernel 2.5.19)
335 Scheduling policy (see sched_setscheduler(2)).
336 delayacct_blkio_ticks (since kernel 2.6.18)
338 Aggregated block I/O delays (measured in clock ticks
339 (centiseconds)).
340 /proc/[number]/statm
342 Provides information about memory status in pages. The columns
343 are:
344 size total program size
345 resident resident set size
346 share shared pages
347 text text (code)
348 lib library
349 data data/stack
350 dt dirty pages (unused in Linux 2.6)
351 /proc/[number]/status
353 Provides much of the information in /proc/[number]/stat and
354 /proc/[number]/statm in a format that's easier for humans to
355 parse.
356 /proc/[number]/task (since kernel 2.6.0-test6)
358 This is a directory that contains one subdirectory for each
359 thread in the process. The name of each subdirectory is the
360 numerical thread ID of the thread (see gettid(2)). Within each
361 of these subdirectories, there is a set of files with the same
362 names and contents as under the /proc/[number] directories. For
363 attributes that are shared by all threads, the contents for each
364 of the files under the task/[thread-ID] subdirectories will be
365 the same as in the corresponding file in the parent /proc/[num‐
366 ber] directory (e.g., in a multithreaded process, all of the
367 task/[thread-ID]/cwd files will have the same value as the
368 /proc/[number]/cwd file in the parent directory, since all of
369 the threads in a process share a working directory). For
370 attributes that are distinct for each thread, the corresponding
371 files under task/[thread-ID] may have different values (e.g.,
372 various fields in each of the task/[thread-ID]/status files may
373 be different for each thread).
374 In a multithreaded process, the contents of the /proc/[num‐
376 ber]/task directory are not available if the main thread has
377 already terminated (typically by calling pthread_exit(3)).
378 /proc/apm
380 Advanced power management version and battery information when
381 CONFIG_APM is defined at kernel compilation time.
382 /proc/bus
384 Contains subdirectories for installed busses.
385 /proc/bus/pccard
387 Subdirectory for pcmcia devices when CONFIG_PCMCIA is set at
388 kernel compilation time.
389 /proc/bus/pccard/drivers
391 /proc/bus/pci
393 Contains various bus subdirectories and pseudo-files containing
394 information about pci busses, installed devices, and device
395 drivers. Some of these files are not ASCII.
396 /proc/bus/pci/devices
398 Information about pci devices. They may be accessed through
399 lspci(8) and setpci(8).
400 /proc/cmdline
402 Arguments passed to the Linux kernel at boot time. Often done
403 via a boot manager such as lilo(1).
404 /proc/cpuinfo
406 This is a collection of CPU and system architecture dependent
407 items, for each supported architecture a different list. Two
408 common entries are processor which gives CPU number and
409 bogomips; a system constant that is calculated during kernel
410 initialization. SMP machines have information for each CPU.
411 /proc/devices
413 Text listing of major numbers and device groups. This can be
414 used by MAKEDEV scripts for consistency with the kernel.
415 /proc/diskstats (since Linux 2.5.69)
417 This file contains disk I/O statistics for each disk device.
418 See the kernel source file Documentation/iostats.txt for further
419 information.
420 /proc/dma
422 This is a list of the registered ISA DMA (direct memory access)
423 channels in use.
424 /proc/driver
426 Empty subdirectory.
427 /proc/execdomains
429 List of the execution domains (ABI personalities).
430 /proc/fb
432 Frame buffer information when CONFIG_FB is defined during kernel
433 compilation.
434 /proc/filesystems
436 A text listing of the filesystems which were compiled into the
437 kernel. Incidentally, this is used by mount(1) to cycle through
438 different filesystems when none is specified.
439 /proc/fs
441 Empty subdirectory.
442 /proc/ide
444 This directory exists on systems with the ide bus. There are
445 directories for each ide channel and attached device. Files
446 include:
447 cache buffer size in KB
449 capacity number of sectors
450 driver driver version
451 geometry physical and logical geometry
452 identify in hexadecimal
453 media media type
454 model manufacturer's model number
455 settings drive settings
456 smart_thresholds in hexadecimal
457 smart_values in hexadecimal
458 The hdparm(8) utility provides access to this information in a
460 friendly format.
461 /proc/interrupts
463 This is used to record the number of interrupts per each IRQ on
464 (at least) the i386 architecture. Very easy to read formatting,
465 done in ASCII.
466 /proc/iomem
468 I/O memory map in Linux 2.4.
469 /proc/ioports
471 This is a list of currently registered Input-Output port regions
472 that are in use.
473 /proc/kallsyms (since Linux 2.5.71)
475 This holds the kernel exported symbol definitions used by the
476 modules(X) tools to dynamically link and bind loadable modules.
477 In Linux 2.5.47 and earlier, a similar file with slightly dif‐
478 ferent syntax was named ksyms.
479 /proc/kcore
481 This file represents the physical memory of the system and is
482 stored in the ELF core file format. With this pseudo-file, and
483 an unstripped kernel (/usr/src/linux/vmlinux) binary, GDB can be
484 used to examine the current state of any kernel data structures.
485 The total length of the file is the size of physical memory
487 (RAM) plus 4KB.
488 /proc/kmsg
490 This file can be used instead of the syslog(2) system call to
491 read kernel messages. A process must have superuser privileges
492 to read this file, and only one process should read this file.
493 This file should not be read if a syslog process is running
494 which uses the syslog(2) system call facility to log kernel mes‐
495 sages.
496 Information in this file is retrieved with the dmesg(8) program.
498 /proc/ksyms (Linux 1.1.23-2.5.47)
500 See /proc/kallsyms.
501 /proc/loadavg
503 The first three fields in this file are load average figures
504 giving the number of jobs in the run queue (state R) or waiting
505 for disk I/O (state D) averaged over 1, 5, and 15 minutes. They
506 are the same as the load average numbers given by uptime(1) and
507 other programs. The fourth field consists of two numbers sepa‐
508 rated by a slash (/). The first of these is the number of cur‐
509 rently executing kernel scheduling entities (processes,
510 threads); this will be less than or equal to the number of CPUs.
511 The value after the slash is the number of kernel scheduling
512 entities that currently exist on the system. The fifth field is
513 the PID of the process that was most recently created on the
514 system.
515 /proc/locks
517 This file shows current file locks (flock(2) and fcntl(2)) and
518 leases (fcntl(2)).
519 /proc/malloc
521 This file is only present if CONFIGDEBUGMALLOC was defined dur‐
522 ing compilation.
523 /proc/meminfo
525 This is used by free(1) to report the amount of free and used
526 memory (both physical and swap) on the system as well as the
527 shared memory and buffers used by the kernel.
528 It is in the same format as free(1), except in bytes rather than
530 KB.
531 /proc/mounts
533 This is a list of all the file systems currently mounted on the
534 system. The format of this file is documented in fstab(5).
535 Since kernel version 2.6.15, this file is pollable: after open‐
536 ing the file for reading, a change in this file (i.e., a file
537 system mount or unmount) causes select(2) to mark the file
538 descriptor as readable, and poll(2) and epoll_wait(2) mark the
539 file as having an error condition.
540 /proc/modules
542 A text list of the modules that have been loaded by the system.
543 See also lsmod(8).
544 /proc/mtrr
546 Memory Type Range Registers. See /usr/src/linux/Documenta‐
547 tion/mtrr.txt for details.
548 /proc/net
550 various net pseudo-files, all of which give the status of some
551 part of the networking layer. These files contain ASCII struc‐
552 tures and are, therefore, readable with cat. However, the stan‐
553 dard netstat(8) suite provides much cleaner access to these
554 files.
555 /proc/net/arp
557 This holds an ASCII readable dump of the kernel ARP table used
558 for address resolutions. It will show both dynamically learned
559 and pre-programmed ARP entries. The format is:
560 IP address HW type Flags HW address Mask Device
562 192.168.0.50 0x1 0x2 00:50:BF:25:68:F3 * eth0
563 192.168.0.250 0x1 0xc 00:00:00:00:00:00 * eth0
564 Here 'IP address' is the IPv4 address of the machine and the 'HW
566 type' is the hardware type of the address from RFC 826. The
567 flags are the internal flags of the ARP structure (as defined in
568 /usr/include/linux/if_arp.h) and the 'HW address' is the data
569 link layer mapping for that IP address if it is known.
570 /proc/net/dev
572 The dev pseudo-file contains network device status information.
573 This gives the number of received and sent packets, the number
574 of errors and collisions and other basic statistics. These are
575 used by the ifconfig(8) program to report device status. The
576 format is:
577 Inter-| Receive | Transmit
579 face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed
580 lo: 2776770 11307 0 0 0 0 0 0 2776770 11307 0 0 0 0 0 0
581 eth0: 1215645 2751 0 0 0 0 0 0 1782404 4324 0 0 0 427 0 0
582 ppp0: 1622270 5552 1 0 0 0 0 0 354130 5669 0 0 0 0 0 0
583 tap0: 7714 81 0 0 0 0 0 0 7714 81 0 0 0 0 0 0
584 /proc/net/dev_mcast
586 Defined in /usr/src/linux/net/core/dev_mcast.c:
587 indx interface_name dmi_u dmi_g dmi_address
588 2 eth0 1 0 01005e000001
589 3 eth1 1 0 01005e000001
590 4 eth2 1 0 01005e000001
591 /proc/net/igmp
593 Internet Group Management Protocol. Defined in
594 /usr/src/linux/net/core/igmp.c.
595 /proc/net/rarp
597 This file uses the same format as the arp file and contains the
598 current reverse mapping database used to provide rarp(8) reverse
599 address lookup services. If RARP is not configured into the ker‐
600 nel, this file will not be present.
601 /proc/net/raw
603 Holds a dump of the RAW socket table. Much of the information is
604 not of use apart from debugging. The 'sl' value is the kernel
605 hash slot for the socket, the 'local address' is the local
606 address and protocol number pair."St" is the internal status of
607 the socket. The "tx_queue" and "rx_queue" are the outgoing and
608 incoming data queue in terms of kernel memory usage. The "tr",
609 "tm->when", and "rexmits" fields are not used by RAW. The "uid"
610 field holds the effective UID of the creator of the socket.
611 /proc/net/snmp
613 This file holds the ASCII data needed for the IP, ICMP, TCP, and
614 UDP management information bases for an snmp agent.
615 /proc/net/tcp
617 Holds a dump of the TCP socket table. Much of the information is
618 not of use apart from debugging. The "sl" value is the kernel
619 hash slot for the socket, the "local address" is the local
620 address and port number pair. The "remote address" is the
621 remote address and port number pair (if connected). 'St' is the
622 internal status of the socket. The 'tx_queue' and 'rx_queue'
623 are the outgoing and incoming data queue in terms of kernel mem‐
624 ory usage. The "tr", "tm->when", and "rexmits" fields hold
625 internal information of the kernel socket state and are only
626 useful for debugging. The "uid" field holds the effective UID
627 of the creator of the socket.
628 /proc/net/udp
630 Holds a dump of the UDP socket table. Much of the information is
631 not of use apart from debugging. The "sl" value is the kernel
632 hash slot for the socket, the "local address" is the local
633 address and port number pair. The "remote address" is the
634 remote address and port number pair (if connected). "St" is the
635 internal status of the socket. The "tx_queue" and "rx_queue"
636 are the outgoing and incoming data queue in terms of kernel mem‐
637 ory usage. The "tr", "tm->when", and "rexmits" fields are not
638 used by UDP. The "uid" field holds the effective UID of the
639 creator of the socket. The format is:
640 sl local_address rem_address st tx_queue rx_queue tr rexmits tm->when uid
642 1: 01642C89:0201 0C642C89:03FF 01 00000000:00000001 01:000071BA 00000000 0
643 1: 00000000:0801 00000000:0000 0A 00000000:00000000 00:00000000 6F000100 0
644 1: 00000000:0201 00000000:0000 0A 00000000:00000000 00:00000000 00000000 0
645 /proc/net/unix
647 Lists the UNIX domain sockets present within the system and
648 their status. The format is:
649 Num RefCount Protocol Flags Type St Path
650 0: 00000002 00000000 00000000 0001 03
651 1: 00000001 00000000 00010000 0001 01 /dev/printer
652 Here 'Num' is the kernel table slot number, 'RefCount' is the
654 number of users of the socket, 'Protocol' is currently always 0,
655 'Flags' represent the internal kernel flags holding the status
656 of the socket. Currently, type is always '1' (Unix domain data‐
657 gram sockets are not yet supported in the kernel). 'St' is the
658 internal state of the socket and Path is the bound path (if any)
659 of the socket.
660 /proc/partitions
662 Contains major and minor numbers of each partition as well as
663 number of blocks and partition name.
664 /proc/pci
666 This is a listing of all PCI devices found during kernel ini‐
667 tialization and their configuration.
668 /proc/scsi
670 A directory with the scsi mid-level pseudo-file and various SCSI
671 lowlevel driver directories, which contain a file for each SCSI
672 host in this system, all of which give the status of some part
673 of the SCSI IO subsystem. These files contain ASCII structures
674 and are, therefore, readable with cat.
675 You can also write to some of the files to reconfigure the sub‐
677 system or switch certain features on or off.
678 /proc/scsi/scsi
680 This is a listing of all SCSI devices known to the kernel. The
681 listing is similar to the one seen during bootup. scsi cur‐
682 rently supports only the add-single-device command which allows
683 root to add a hotplugged device to the list of known devices.
684 An echo 'scsi add-single-device 1 0 5 0' > /proc/scsi/scsi will
686 cause host scsi1 to scan on SCSI channel 0 for a device on ID 5
687 LUN 0. If there is already a device known on this address or the
688 address is invalid, an error will be returned.
689 /proc/scsi/[drivername]
691 [drivername] can currently be NCR53c7xx, aha152x, aha1542,
692 aha1740, aic7xxx, buslogic, eata_dma, eata_pio, fdomain, in2000,
693 pas16, qlogic, scsi_debug, seagate, t128, u15-24f, ultrastore,
694 or wd7000. These directories show up for all drivers that reg‐
695 istered at least one SCSI HBA. Every directory contains one file
696 per registered host. Every host-file is named after the number
697 the host was assigned during initialization.
698 Reading these files will usually show driver and host configura‐
700 tion, statistics etc.
701 Writing to these files allows different things on different
703 hosts. For example, with the latency and nolatency commands,
704 root can switch on and off command latency measurement code in
705 the eata_dma driver. With the lockup and unlock commands, root
706 can control bus lockups simulated by the scsi_debug driver.
707 /proc/self
709 This directory refers to the process accessing the /proc
710 filesystem, and is identical to the /proc directory named by the
711 process ID of the same process.
712 /proc/slabinfo
714 Information about kernel caches. The columns are:
715 cache-name
716 num-active-objs
717 total-objs
718 object-size
719 num-active-slabs
720 total-slabs
721 num-pages-per-slab
722 See slabinfo(5) for details.
723 /proc/stat
725 kernel/system statistics. Varies with architecture. Common
726 entries include:
727 cpu 3357 0 4313 1362393
729 The amount of time, measured in units of USER_HZ
730 (1/100ths of a second on most architectures), that the
731 system spent in user mode, user mode with low priority
732 (nice), system mode, and the idle task, respectively.
733 The last value should be USER_HZ times the second entry
734 in the uptime pseudo-file.
735 In Linux 2.6 this line includes three additional columns:
737 iowait - time waiting for I/O to complete (since 2.5.41);
738 irq - time servicing interrupts (since 2.6.0-test4);
739 softirq - time servicing softirqs (since 2.6.0-test4).
740 page 5741 1808
742 The number of pages the system paged in and the number
743 that were paged out (from disk).
744 swap 1 0
746 The number of swap pages that have been brought in and
747 out.
748 intr 1462898
750 This line shows counts of interrupts serviced since boot
751 time, for each of the possible system interrupts. The
752 first column is the total of all interrupts serviced;
753 each subsequent column is the total for a particular
754 interrupt.
755 disk_io: (2,0):(31,30,5764,1,2) (3,0):...
757 (major,minor):(noinfo, read_io_ops, blks_read,
758 write_io_ops, blks_written)
759 (Linux 2.4 only)
760 ctxt 115315
762 The number of context switches that the system underwent.
763 btime 769041601
765 boot time, in seconds since the epoch (January 1, 1970).
766 processes 86031
768 Number of forks since boot.
769 procs_running 6
771 Number of processes in runnable state. (Linux 2.5.45
772 onwards.)
773 procs_blocked 2
775 Number of processes blocked waiting for I/O to complete.
776 (Linux 2.5.45 onwards.)
777 /proc/swaps
779 Swap areas in use. See also swapon(8).
780 /proc/sys
782 This directory (present since 1.3.57) contains a number of files
783 and subdirectories corresponding to kernel variables. These
784 variables can be read and sometimes modified using the proc file
785 system, and the sysctl(2) system call. Presently, there are sub‐
786 directories abi, debug, dev, fs, kernel, net, proc, rxrpc, sun‐
787 rpc and vm that each contain more files and subdirectories.
788 /proc/sys/abi
790 This directory may contain files with application binary infor‐
791 mation. On some systems, it is not present.
792 /proc/sys/debug
794 This directory may be empty.
795 /proc/sys/dev
797 This directory contains device specific information (eg
798 dev/cdrom/info). On some systems, it may be empty.
799 /proc/sys/fs
801 This contains the subdirectories binfmt_misc, inotify, and
802 mqueue, and files dentry-state, dir-notify-enable, dquot-nr,
803 file-max, file-nr, inode-max, inode-nr, inode-state, lease-
804 break-time, leases-enable, overflowgid, overflowuid,
805 suid_dumpable, super-max, and super-nr.
806 /proc/sys/fs/binfmt_misc
808 Documentation for files in this directory can be found in the
809 kernel sources in Documentation/binfmt_misc.txt.
810 /proc/sys/fs/dentry-state
812 This file contains six numbers, nr_dentry, nr_unused, age_limit
813 (age in seconds), want_pages (pages requested by system) and two
814 dummy values. nr_dentry seems to be 0 all the time. nr_unused
815 seems to be the number of unused dentries. age_limit is the age
816 in seconds after which dcache entries can be reclaimed when mem‐
817 ory is short and want_pages is non-zero when the kernel has
818 called shrink_dcache_pages() and the dcache isn't pruned yet.
819 /proc/sys/fs/dir-notify-enable
821 This file can be used to disable or enable the dnotify interface
822 described in fcntl(2) on a system-wide basis. A value of 0 in
823 this file disables the interface, and a value of 1 enables it.
824 /proc/sys/fs/dquot-max
826 This file shows the maximum number of cached disk quota entries.
827 On some (2.4) systems, it is not present. If the number of free
828 cached disk quota entries is very low and you have some awesome
829 number of simultaneous system users, you might want to raise the
830 limit.
831 /proc/sys/fs/dquot-nr
833 This file shows the number of allocated disk quota entries and
834 the number of free disk quota entries.
835 /proc/sys/fs/file-max
837 This file defines a system-wide limit on the number of open
838 files for all processes. (See also setrlimit(2), which can be
839 used by a process to set the per-process limit, RLIMIT_NOFILE,
840 on the number of files it may open.) If you get lots of error
841 messages about running out of file handles, try increasing this
842 value:
843 echo 100000 > /proc/sys/fs/file-max
845 The kernel constant NR_OPEN imposes an upper limit on the value
847 that may be placed in file-max.
848 If you increase /proc/sys/fs/file-max, be sure to increase
850 /proc/sys/fs/inode-max to 3-4 times the new value of
851 /proc/sys/fs/file-max, or you will run out of inodes.
852 /proc/sys/fs/file-nr
854 This (read-only) file gives the number of files presently
855 opened. It contains three numbers: The number of allocated file
856 handles, the number of free file handles and the maximum number
857 of file handles. The kernel allocates file handles dynamically,
858 but it doesn't free them again. If the number of allocated
859 files is close to the
860 maximum, you should consider increasing the maximum. When the
862 number of free file handles is large, you've encountered a peak
863 in your usage of file handles and you probably don't need to
864 increase the maximum.
865 /proc/sys/fs/inode-max
867 This file contains the maximum number of in-memory inodes. On
868 some (2.4) systems, it may not be present. This value should be
869 3-4 times larger than the value in file-max, since stdin, stdout
870 and network sockets also need an inode to handle them. When you
871 regularly run out of inodes, you need to increase this value.
872 /proc/sys/fs/inode-nr
874 This file contains the first two values from inode-state.
875 /proc/sys/fs/inode-state
877 This file contains seven numbers: nr_inodes, nr_free_inodes,
878 preshrink and four dummy values. nr_inodes is the number of
879 inodes the system has allocated. This can be slightly more than
880 inode-max because Linux allocates them one page full at a time.
881 nr_free_inodes represents the number of free inodes. preshrink
882 is non-zero when the nr_inodes > inode-max and the system needs
883 to prune the inode list instead of allocating more.
884 /proc/sys/fs/inotify (since Linux 2.6.13)
886 This directory contains files max_queued_events,
887 max_user_instances, and max_user_watches, that can be used to
888 limit the amount of kernel memory consumed by the inotify inter‐
889 face. For further details, see inotify(7).
890 /proc/sys/fs/lease-break-time
892 This file specifies the grace period that the kernel grants to a
893 process holding a file lease (fcntl(2)) after it has sent a sig‐
894 nal to that process notifying it that another process is waiting
895 to open the file. If the lease holder does not remove or down‐
896 grade the lease within this grace period, the kernel forcibly
897 breaks the lease.
898 /proc/sys/fs/leases-enable
900 This file can be used to enable or disable file leases
901 (fcntl(2)) on a system-wide basis. If this file contains the
902 value 0, leases are disabled. A non-zero value enables leases.
903 /proc/sys/fs/mqueue (since Linux 2.6.6)
905 This directory contains files msg_max, msgsize_max, and
906 queues_max, controlling the resources used by POSIX message
907 queues. See mq_overview(7) for details.
908 /proc/sys/fs/overflowgid and /proc/sys/fs/overflowuid
910 These files allow you to change the value of the fixed UID and
911 GID. The default is 65534. Some filesystems only support
912 16-bit UIDs and GIDs, although in Linux UIDs and GIDs are 32
913 bits. When one of these filesystems is mounted with writes
914 enabled, any UID or GID that would exceed 65535 is translated to
915 the overflow value before being written to disk.
916 /proc/sys/fs/suid_dumpable (since Linux 2.6.13)
918 The value in this file determines whether core dump files are
919 produced for set-user-ID or otherwise protected/tainted bina‐
920 ries. Three different integer values can be specified:
921 0 (default) This provides the traditional (pre-Linux 2.6.13) be‐
923 haviour. A core dump will not be produced for a process which
924 has changed credentials (by calling seteuid(2), setgid(2), or
925 similar, or by executing a set-user-ID or set-group-ID program)
926 or whose binary does not have read permission enabled.
927 1 ("debug") All processes dump core when possible. The core
929 dump is owned by the file system user ID of the dumping process
930 and no security is applied. This is intended for system debug‐
931 ging situations only. Ptrace is unchecked.
932 2 ("suidsafe") Any binary which normally would not be dumped
934 (see "0" above) is dumped readable by root only. This allows
935 the user to remove the core dump file but not to read it. For
936 security reasons core dumps in this mode will not overwrite one
937 another or other files. This mode is appropriate when adminis‐
938 trators are attempting to debug problems in a normal environ‐
939 ment.
940 /proc/sys/fs/super-max
942 This file controls the maximum number of superblocks, and thus
943 the maximum number of mounted filesystems the kernel can have.
944 You only need to increase super-max if you need to mount more
945 filesystems than the current value in super-max allows you to.
946 /proc/sys/fs/super-nr
948 This file contains the number of filesystems currently mounted.
949 /proc/sys/kernel
951 This directory contains files acct, cad_pid, cap-bound,
952 core_pattern, core_uses_pid, ctrl-alt-del, dentry-state, domain‐
953 name, hotplug, hostname, htab-reclaim (PowerPC only), java-
954 appletviewer (binfmt_java, obsolete), java-interpreter
955 (binfmt_java, obsolete), l2cr (PowerPC only), modprobe, msgmax,
956 msgmnb, msgmni, osrelease, ostype, overflowgid, overflowuid,
957 panic, panic_on_oops, pid_max, powersave-nap (PowerPC only),
958 printk, pty, random, real-root-dev, reboot-cmd (SPARC only),
959 rtsig-max, rtsig-nr, sem, sg-big-buff, shmall, shmmax, shmmni,
960 sysrq, tainted, threads-max, version, and zero-paged (PowerPC
961 only).
962 /proc/sys/kernel/acct
964 This file contains three numbers: highwater, lowwater and fre‐
965 quency. If BSD-style process accounting is enabled these values
966 control its behaviour. If free space on filesystem where the log
967 lives goes below lowwater percent accounting suspends. If free
968 space gets above highwater percent accounting resumes. Fre‐
969 quency determines how often the kernel checks the amount of free
970 space (value is in seconds). Default values are 4, 2 and 30.
971 That is, suspend accounting if <= 2% of space is free; resume it
972 if >= 4% of space is free; consider information about amount of
973 free space valid for 30 seconds.
974 /proc/sys/kernel/cap-bound
976 This file holds the value of the kernel capability bounding set
977 (expressed as a signed decimal number). This set is ANDed
978 against the capabilities permitted to a process during exec().
979 /proc/sys/kernel/core_pattern
981 See core(5). /proc/sys/kernel/core_uses_pid See core(5).
982 /proc/sys/kernel/ctrl-alt-del
984 This file controls the handling of Ctrl-Alt-Del from the key‐
985 board. When the value in this file is 0, Ctrl-Alt-Del is
986 trapped and sent to the init(1) program to handle a graceful
987 restart. When the value is > 0, Linux's reaction to a Vulcan
988 Nerve Pinch (tm) will be an immediate reboot, without even sync‐
989 ing its dirty buffers. Note: when a program (like dosemu) has
990 the keyboard in 'raw' mode, the ctrl-alt-del is intercepted by
991 the program before it ever reaches the kernel tty layer, and
992 it's up to the program to decide what to do with it.
993 /proc/sys/kernel/hotplug
995 This file contains the path for the hotplug policy agent. The
996 default value in this file "/sbin/hotplug".
997 /proc/sys/kernel/domainname and /proc/sys/kernel/hostname
999 can be used to set the NIS/YP domainname and the hostname of
1000 your box in exactly the same way as the commands domainname and
1001 hostname, i.e.:
1002 # echo "darkstar" > /proc/sys/kernel/hostname
1004 # echo "mydomain" > /proc/sys/kernel/domainname
1005 has the same effect as
1007 # hostname "darkstar"
1009 # domainname "mydomain"
1010 Note, however, that the classic darkstar.frop.org has the host‐
1012 name "darkstar" and DNS (Internet Domain Name Server) domainname
1013 "frop.org", not to be confused with the NIS (Network Information
1014 Service) or YP (Yellow Pages) domainname. These two domain names
1015 are in general different. For a detailed discussion see the
1016 hostname(1) man page.
1017 /proc/sys/kernel/htab-reclaim
1019 (PowerPC only) If this file is set to a non-zero value, the Pow‐
1020 erPC htab (see kernel file Documentation/powerpc/ppc_htab.txt)
1021 is pruned each time the system hits the idle loop.
1022 /proc/sys/kernel/l2cr
1024 (PowerPC only) This file contains a flag that controls the L2
1025 cache of G3 processor boards. If 0, the cache is disabled.
1026 Enabled if non-zero.
1027 /proc/sys/kernel/modprobe
1029 This file is described by the kernel source file Documenta‐
1030 tion/kmod.txt.
1031 /proc/sys/kernel/msgmax
1033 This file defines a system-wide limit specifying the maximum
1034 number of bytes in a single message written on a System V mes‐
1035 sage queue.
1036 /proc/sys/kernel/msgmni
1038 This file defines the system-wide limit on the number of message
1039 queue identifiers. (This file is only present in Linux 2.4
1040 onwards.)
1041 /proc/sys/kernel/msgmnb
1043 This file defines a system-wide parameter used to initialise the
1044 msg_qbytes setting for subsequently created message queues. The
1045 msg_qbytes setting specifies the maximum number of bytes that
1046 may be written to the message queue.
1047 /proc/sys/kernel/ostype and /proc/sys/kernel/osrelease
1049 These files give substrings of /proc/version.
1050 /proc/sys/kernel/overflowgid and /proc/sys/kernel/overflowuid
1052 These files duplicate the files /proc/sys/fs/overflowgid and
1053 /proc/sys/fs/overflowuid.
1054 /proc/sys/kernel/panic
1056 gives read/write access to the kernel variable panic_timeout.
1057 If this is zero, the kernel will loop on a panic; if non-zero it
1058 indicates that the kernel should autoreboot after this number of
1059 seconds. When you use the software watchdog device driver, the
1060 recommended setting is 60.
1061 /proc/sys/kernel/panic_on_oops
1063 This file (new in Linux 2.5) controls the kernel's behaviour
1064 when an oops or BUG is encountered. If this file contains 0,
1065 then the system tries to continue operation. If it contains 1,
1066 then the system delays a few seconds (to give klogd time to
1067 record the oops output) and then panics. If the /proc/sys/ker‐
1068 nel/panic file is also non-zero then the machine will be
1069 rebooted.
1070 /proc/sys/kernel/pid_max
1072 This file (new in Linux 2.5) specifies the value at which PIDs
1073 wrap around (i.e., the value in this file is one greater than
1074 the maximum PID). The default value for this file, 32768,
1075 results in the same range of PIDs as on earlier kernels. On
1076 32-bit platforms, 32768 is the maximum value for pid_max. On
1077 64-bit systems, pid_max can be set to any value up to 2^22
1078 (PID_MAX_LIMIT, approximately 4 million).
1079 /proc/sys/kernel/powersave-nap (PowerPC only)
1081 This file contains a flag. If set, Linux-PPC will use the 'nap'
1082 mode of powersaving, otherwise the 'doze' mode will be used.
1083 /proc/sys/kernel/printk
1085 The four values in this file are console_loglevel, default_mes‐
1086 sage_loglevel, minimum_console_level and default_con‐
1087 sole_loglevel. These values influence printk() behavior when
1088 printing or logging error messages. See syslog(2) for more info
1089 on the different loglevels. Messages with a higher priority
1090 than console_loglevel will be printed to the console. Messages
1091 without an explicit priority will be printed with priority
1092 default_message_level. minimum_console_loglevel is the minimum
1093 (highest) value to which console_loglevel can be set.
1094 default_console_loglevel is the default value for con‐
1095 sole_loglevel.
1096 /proc/sys/kernel/pty (since Linux 2.6.4)
1098 This directory contains two files relating to the number of Unix
1099 98 pseudo-terminals (see pts(4)) on the system.
1100 /proc/sys/kernel/pty/max
1102 This file defines the maximum number of pseudo-terminals.
1103 /proc/sys/kernel/pty/nr
1105 This read-only file indicates how many pseudo-terminals are cur‐
1106 rently in use.
1107 /proc/sys/kernel/random
1109 This directory contains various parameters controlling the oper‐
1110 ation of the file /dev/random. See random(4) for further infor‐
1111 mation.
1112 /proc/sys/kernel/real-root-dev
1114 This file is documented in the kernel source file Documenta‐
1115 tion/initrd.txt.
1116 /proc/sys/kernel/reboot-cmd (Sparc only)
1118 This file seems to be a way to give an argument to the SPARC
1119 ROM/Flash boot loader. Maybe to tell it what to do after reboot‐
1120 ing?
1121 /proc/sys/kernel/rtsig-max
1123 (Only in kernels up to and including 2.6.7; see setrlimit(2))
1124 This file can be used to tune the maximum number of POSIX real‐
1125 time (queued) signals that can be outstanding in the system.
1126 /proc/sys/kernel/rtsig-nr
1128 (Only in kernels up to and including 2.6.7.) This file shows
1129 the number POSIX realtime signals currently queued.
1130 /proc/sys/kernel/sem (since Linux 2.4)
1132 This file contains 4 numbers defining limits for System V IPC
1133 semaphores. These fields are, in order:
1134 SEMMSL The maximum semaphores per semaphore set.
1136 SEMMNS A system-wide limit on the number of semaphores in all
1138 semaphore sets.
1139 SEMOPM The maximum number of operations that may be specified
1141 in a semop(2) call.
1142 SEMMNI A system-wide limit on the maximum number of semaphore
1144 identifiers.
1145 /proc/sys/kernel/sg-big-buff
1147 This file shows the size of the generic SCSI device (sg) buffer.
1148 You can't tune it just yet, but you could change it on compile
1149 time by editing include/scsi/sg.h and changing the value of
1150 SG_BIG_BUFF. However, there shouldn't be any reason to change
1151 this value.
1152 /proc/sys/kernel/shmall
1154 This file contains the system-wide limit on the total number of
1155 pages of System V shared memory.
1156 /proc/sys/kernel/shmmax
1158 This file can be used to query and set the run time limit on the
1159 maximum (System V IPC) shared memory segment size that can be
1160 created. Shared memory segments up to 1Gb are now supported in
1161 the kernel. This value defaults to SHMMAX.
1162 /proc/sys/kernel/shmmni
1164 (available in Linux 2.4 and onwards) This file specifies the
1165 system-wide maximum number of System V shared memory segments
1166 that can be created.
1167 /proc/sys/kernel/version
1169 contains a string like:
1170 #5 Wed Feb 25 21:49:24 MET 1998.TP
1172 The '#5' means that this is the fifth kernel built from this
1174 source base and the date behind it indicates the time the kernel
1175 was built.
1176 /proc/sys/kernel/zero-paged (PowerPC only)
1178 This file contains a flag. When enabled (non-zero), Linux-PPC
1179 will pre-zero pages in the idle loop, possibly speeding up
1180 get_free_pages.
1181 /proc/sys/net
1183 This directory contains networking stuff. Explanations for some
1184 of the files under this directory can be found in tcp(7) and
1185 ip(7).
1186 /proc/sys/proc
1188 This directory may be empty.
1189 /proc/sys/sunrpc
1191 This directory supports Sun remote procedure call for network
1192 file system (NFS). On some systems, it is not present.
1193 /proc/sys/vm
1195 This directory contains files for memory management tuning, buf‐
1196 fer and cache management.
1197 /proc/sys/vm/drop_caches (since Linux 2.6.16)
1199 Writing to this file causes the kernel to drop clean caches,
1200 dentries and inodes from memory, causing that memory to become
1201 free.
1202 To free pagecache, use echo 1 > /proc/sys/vm/drop_caches; to
1204 free dentries and inodes, use echo 2 > /proc/sys/vm/drop_caches;
1205 to free pagecache, dentries and inodes, use echo 3 >
1206 /proc/sys/vm/drop_caches.
1207 Because this is a non-destructive operation and dirty objects
1209 are not freeable, the user should run sync(8) first.
1210 /proc/sys/vm/legacy_va_layout (since Linux 2.6.9)
1212 If non-zero, this disable the new 32-bit memory-mapping layout;
1213 the kernel will use the legacy (2.4) layout for all processes.
1214 /proc/sys/vm/overcommit_memory
1216 This file contains the kernel virtual memory accounting mode.
1217 Values are:
1218 0: heuristic overcommit (this is the default)
1219 1: always overcommit, never check
1220 2: always check, never overcommit
1221 In mode 0, calls of mmap(2) with MAP_NORESERVE set are not
1222 checked, and the default check is very weak, leading to the risk
1223 of getting a process "OOM-killed". Under Linux 2.4 any non-zero
1224 value implies mode 1. In mode 2 (available since Linux 2.6),
1225 the total virtual address space on the system is limited to (SS
1226 + RAM*(r/100)), where SS is the size of the swap space, and RAM
1227 is the size of the physical memory, and r is the contents of the
1228 file /proc/sys/vm/overcommit_ratio.
1229 /proc/sys/vm/overcommit_ratio
1231 See the description of /proc/sys/vm/overcommit_memory.
1232 /proc/sysvipc
1234 Subdirectory containing the pseudo-files msg, sem and shm.
1235 These files list the System V Interprocess Communication (IPC)
1236 objects (respectively: message queues, semaphores, and shared
1237 memory) that currently exist on the system, providing similar
1238 information to that available via ipcs(1). These files have
1239 headers and are formatted (one IPC object per line) for easy
1240 understanding. svipc(7) provides further background on the
1241 information shown by these files.
1242 /proc/tty
1244 Subdirectory containing the pseudo-files and subdirectories for
1245 tty drivers and line disciplines.
1246 /proc/uptime
1248 This file contains two numbers: the uptime of the system (sec‐
1249 onds), and the amount of time spent in idle process (seconds).
1250 /proc/version
1252 This string identifies the kernel version that is currently run‐
1253 ning. It includes the contents of /proc/sys/ostype,
1254 /proc/sys/osrelease and /proc/sys/version. For example:
1255 Linux version 1.0.9 (quinlan@phaze) #1 Sat May 14 01:51:54 EDT 1994
1256 /proc/vmstat (since Linux 2.6)
1258 This file displays various virtual memory statistics.
1259 /proc/zoneinfo (since Linux 2.6.13)
1262 This file display information about memory zones. This is use‐
1263 ful for analysing virtual memory behaviour.
1264
1266 cat(1), find(1), free(1), mount(1), ps(1), tr(1), uptime(1), chroot(2),
1267 mmap(2), readlink(2), syslog(2), slabinfo(5), hier(7), arp(8),
1268 dmesg(8), hdparm(8), ifconfig(8), init(8), lsmod(8), lspci(8), net‐
1269 stat(8), procinfo(8), route(8)
1270 /usr/src/linux/Documentation/filesystems/proc.txt
1271
1273 Note that many strings (i.e., the environment and command line) are in
1274 the internal format, with sub-fields terminated by null bytes ('\0'),
1275 so you may find that things are more readable if you use od -c or tr
1276 "\000" "\n" to read them. Alternatively, echo `cat <file>` works well.
1277 This manual page is incomplete, possibly inaccurate, and is the kind of
1279 thing that needs to be updated very often.
1280
1282 The material on /proc/sys/fs and /proc/sys/kernel is closely based on
1283 kernel source documentation files written by Rik van Riel.
1284 2005-05-12 PROC(5)