1 ------------------------------------------------------------------------------
2 T H E /proc F I L E S Y S T E M
3 ------------------------------------------------------------------------------
4 /proc/sys Terrehon Bowden <terrehon@pacbell.net> October 7 1999
5 Bodo Bauer <bb@ricochet.net>
7 2.4.x update Jorge Nerin <comandante@zaralinux.com> November 14 2000
8 move /proc/sys Shen Feng <shen@cn.fujitsu.com> April 1 2009
9 ------------------------------------------------------------------------------
10 Version 1.3 Kernel version 2.2.12
11 Kernel version 2.4.0-test11-pre4
12 ------------------------------------------------------------------------------
13 fixes/update part 1.1 Stefani Seibold <stefani@seibold.net> June 9 2009
19 0.1 Introduction/Credits
22 1 Collecting System Information
23 1.1 Process-Specific Subdirectories
25 1.3 IDE devices in /proc/ide
26 1.4 Networking info in /proc/net
28 1.6 Parallel port info in /proc/parport
29 1.7 TTY info in /proc/tty
30 1.8 Miscellaneous kernel statistics in /proc/stat
31 1.9 Ext4 file system parameters
33 2 Modifying System Parameters
35 3 Per-Process Parameters
36 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj - Adjust the oom-killer
38 3.2 /proc/<pid>/oom_score - Display current oom-killer score
39 3.3 /proc/<pid>/io - Display the IO accounting fields
40 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings
41 3.5 /proc/<pid>/mountinfo - Information about mounts
42 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm
43 3.7 /proc/<pid>/task/<tid>/children - Information about task children
44 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file
45 3.9 /proc/<pid>/map_files - Information about memory mapped files
46 3.10 /proc/<pid>/timerslack_ns - Task timerslack value
51 ------------------------------------------------------------------------------
53 ------------------------------------------------------------------------------
55 0.1 Introduction/Credits
56 ------------------------
58 This documentation is part of a soon (or so we hope) to be released book on
59 the SuSE Linux distribution. As there is no complete documentation for the
60 /proc file system and we've used many freely available sources to write these
61 chapters, it seems only fair to give the work back to the Linux community.
62 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
63 afraid it's still far from complete, but we hope it will be useful. As far as
64 we know, it is the first 'all-in-one' document about the /proc file system. It
65 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
66 SPARC, AXP, etc., features, you probably won't find what you are looking for.
67 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
68 additions and patches are welcome and will be added to this document if you
71 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
72 other people for help compiling this documentation. We'd also like to extend a
73 special thank you to Andi Kleen for documentation, which we relied on heavily
74 to create this document, as well as the additional information he provided.
75 Thanks to everybody else who contributed source or docs to the Linux kernel
76 and helped create a great piece of software... :)
78 If you have any comments, corrections or additions, please don't hesitate to
79 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
82 The latest version of this document is available online at
83 http://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html
85 If the above direction does not works for you, you could try the kernel
86 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
87 comandante@zaralinux.com.
92 We don't guarantee the correctness of this document, and if you come to us
93 complaining about how you screwed up your system because of incorrect
94 documentation, we won't feel responsible...
96 ------------------------------------------------------------------------------
97 CHAPTER 1: COLLECTING SYSTEM INFORMATION
98 ------------------------------------------------------------------------------
100 ------------------------------------------------------------------------------
102 ------------------------------------------------------------------------------
103 * Investigating the properties of the pseudo file system /proc and its
104 ability to provide information on the running Linux system
105 * Examining /proc's structure
106 * Uncovering various information about the kernel and the processes running
108 ------------------------------------------------------------------------------
111 The proc file system acts as an interface to internal data structures in the
112 kernel. It can be used to obtain information about the system and to change
113 certain kernel parameters at runtime (sysctl).
115 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
116 show you how you can use /proc/sys to change settings.
118 1.1 Process-Specific Subdirectories
119 -----------------------------------
121 The directory /proc contains (among other things) one subdirectory for each
122 process running on the system, which is named after the process ID (PID).
124 The link self points to the process reading the file system. Each process
125 subdirectory has the entries listed in Table 1-1.
128 Table 1-1: Process specific entries in /proc
129 ..............................................................................
131 clear_refs Clears page referenced bits shown in smaps output
132 cmdline Command line arguments
133 cpu Current and last cpu in which it was executed (2.4)(smp)
134 cwd Link to the current working directory
135 environ Values of environment variables
136 exe Link to the executable of this process
137 fd Directory, which contains all file descriptors
138 maps Memory maps to executables and library files (2.4)
139 mem Memory held by this process
140 root Link to the root directory of this process
142 statm Process memory status information
143 status Process status in human readable form
144 wchan Present with CONFIG_KALLSYMS=y: it shows the kernel function
145 symbol the task is blocked in - or "0" if not blocked.
147 stack Report full stack trace, enable via CONFIG_STACKTRACE
148 smaps an extension based on maps, showing the memory consumption of
149 each mapping and flags associated with it
150 numa_maps an extension based on maps, showing the memory locality and
151 binding policy as well as mem usage (in pages) of each mapping.
152 ..............................................................................
154 For example, to get the status information of a process, all you have to do is
155 read the file /proc/PID/status:
157 >cat /proc/self/status
185 SigPnd: 0000000000000000
186 ShdPnd: 0000000000000000
187 SigBlk: 0000000000000000
188 SigIgn: 0000000000000000
189 SigCgt: 0000000000000000
190 CapInh: 00000000fffffeff
191 CapPrm: 0000000000000000
192 CapEff: 0000000000000000
193 CapBnd: ffffffffffffffff
196 voluntary_ctxt_switches: 0
197 nonvoluntary_ctxt_switches: 1
199 This shows you nearly the same information you would get if you viewed it with
200 the ps command. In fact, ps uses the proc file system to obtain its
201 information. But you get a more detailed view of the process by reading the
202 file /proc/PID/status. It fields are described in table 1-2.
204 The statm file contains more detailed information about the process
205 memory usage. Its seven fields are explained in Table 1-3. The stat file
206 contains details information about the process itself. Its fields are
207 explained in Table 1-4.
209 (for SMP CONFIG users)
210 For making accounting scalable, RSS related information are handled in an
211 asynchronous manner and the value may not be very precise. To see a precise
212 snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table.
213 It's slow but very precise.
215 Table 1-2: Contents of the status files (as of 4.1)
216 ..............................................................................
218 Name filename of the executable
219 State state (R is running, S is sleeping, D is sleeping
220 in an uninterruptible wait, Z is zombie,
221 T is traced or stopped)
223 Ngid NUMA group ID (0 if none)
225 PPid process id of the parent process
226 TracerPid PID of process tracing this process (0 if not)
227 Uid Real, effective, saved set, and file system UIDs
228 Gid Real, effective, saved set, and file system GIDs
229 Umask file mode creation mask
230 FDSize number of file descriptor slots currently allocated
231 Groups supplementary group list
232 NStgid descendant namespace thread group ID hierarchy
233 NSpid descendant namespace process ID hierarchy
234 NSpgid descendant namespace process group ID hierarchy
235 NSsid descendant namespace session ID hierarchy
236 VmPeak peak virtual memory size
237 VmSize total program size
238 VmLck locked memory size
239 VmHWM peak resident set size ("high water mark")
240 VmRSS size of memory portions. It contains the three
241 following parts (VmRSS = RssAnon + RssFile + RssShmem)
242 RssAnon size of resident anonymous memory
243 RssFile size of resident file mappings
244 RssShmem size of resident shmem memory (includes SysV shm,
245 mapping of tmpfs and shared anonymous mappings)
246 VmData size of private data segments
247 VmStk size of stack segments
248 VmExe size of text segment
249 VmLib size of shared library code
250 VmPTE size of page table entries
251 VmPMD size of second level page tables
252 VmSwap amount of swap used by anonymous private data
253 (shmem swap usage is not included)
254 HugetlbPages size of hugetlb memory portions
255 Threads number of threads
256 SigQ number of signals queued/max. number for queue
257 SigPnd bitmap of pending signals for the thread
258 ShdPnd bitmap of shared pending signals for the process
259 SigBlk bitmap of blocked signals
260 SigIgn bitmap of ignored signals
261 SigCgt bitmap of caught signals
262 CapInh bitmap of inheritable capabilities
263 CapPrm bitmap of permitted capabilities
264 CapEff bitmap of effective capabilities
265 CapBnd bitmap of capabilities bounding set
266 NoNewPrivs no_new_privs, like prctl(PR_GET_NO_NEW_PRIV, ...)
267 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...)
268 Cpus_allowed mask of CPUs on which this process may run
269 Cpus_allowed_list Same as previous, but in "list format"
270 Mems_allowed mask of memory nodes allowed to this process
271 Mems_allowed_list Same as previous, but in "list format"
272 voluntary_ctxt_switches number of voluntary context switches
273 nonvoluntary_ctxt_switches number of non voluntary context switches
274 ..............................................................................
276 Table 1-3: Contents of the statm files (as of 2.6.8-rc3)
277 ..............................................................................
279 size total program size (pages) (same as VmSize in status)
280 resident size of memory portions (pages) (same as VmRSS in status)
281 shared number of pages that are shared (i.e. backed by a file, same
282 as RssFile+RssShmem in status)
283 trs number of pages that are 'code' (not including libs; broken,
284 includes data segment)
285 lrs number of pages of library (always 0 on 2.6)
286 drs number of pages of data/stack (including libs; broken,
287 includes library text)
288 dt number of dirty pages (always 0 on 2.6)
289 ..............................................................................
292 Table 1-4: Contents of the stat files (as of 2.6.30-rc7)
293 ..............................................................................
296 tcomm filename of the executable
297 state state (R is running, S is sleeping, D is sleeping in an
298 uninterruptible wait, Z is zombie, T is traced or stopped)
299 ppid process id of the parent process
300 pgrp pgrp of the process
302 tty_nr tty the process uses
303 tty_pgrp pgrp of the tty
305 min_flt number of minor faults
306 cmin_flt number of minor faults with child's
307 maj_flt number of major faults
308 cmaj_flt number of major faults with child's
309 utime user mode jiffies
310 stime kernel mode jiffies
311 cutime user mode jiffies with child's
312 cstime kernel mode jiffies with child's
313 priority priority level
315 num_threads number of threads
316 it_real_value (obsolete, always 0)
317 start_time time the process started after system boot
318 vsize virtual memory size
319 rss resident set memory size
320 rsslim current limit in bytes on the rss
321 start_code address above which program text can run
322 end_code address below which program text can run
323 start_stack address of the start of the main process stack
324 esp current value of ESP
325 eip current value of EIP
326 pending bitmap of pending signals
327 blocked bitmap of blocked signals
328 sigign bitmap of ignored signals
329 sigcatch bitmap of caught signals
330 0 (place holder, used to be the wchan address, use /proc/PID/wchan instead)
333 exit_signal signal to send to parent thread on exit
334 task_cpu which CPU the task is scheduled on
335 rt_priority realtime priority
336 policy scheduling policy (man sched_setscheduler)
337 blkio_ticks time spent waiting for block IO
338 gtime guest time of the task in jiffies
339 cgtime guest time of the task children in jiffies
340 start_data address above which program data+bss is placed
341 end_data address below which program data+bss is placed
342 start_brk address above which program heap can be expanded with brk()
343 arg_start address above which program command line is placed
344 arg_end address below which program command line is placed
345 env_start address above which program environment is placed
346 env_end address below which program environment is placed
347 exit_code the thread's exit_code in the form reported by the waitpid system call
348 ..............................................................................
350 The /proc/PID/maps file containing the currently mapped memory regions and
351 their access permissions.
355 address perms offset dev inode pathname
357 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
358 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
359 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
360 a7cb1000-a7cb2000 ---p 00000000 00:00 0
361 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
362 a7eb2000-a7eb3000 ---p 00000000 00:00 0
363 a7eb3000-a7ed5000 rw-p 00000000 00:00 0
364 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
365 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
366 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
367 a800b000-a800e000 rw-p 00000000 00:00 0
368 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
369 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
370 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
371 a8024000-a8027000 rw-p 00000000 00:00 0
372 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
373 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
374 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
375 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack]
376 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
378 where "address" is the address space in the process that it occupies, "perms"
379 is a set of permissions:
385 p = private (copy on write)
387 "offset" is the offset into the mapping, "dev" is the device (major:minor), and
388 "inode" is the inode on that device. 0 indicates that no inode is associated
389 with the memory region, as the case would be with BSS (uninitialized data).
390 The "pathname" shows the name associated file for this mapping. If the mapping
391 is not associated with a file:
393 [heap] = the heap of the program
394 [stack] = the stack of the main process
395 [vdso] = the "virtual dynamic shared object",
396 the kernel system call handler
398 or if empty, the mapping is anonymous.
400 The /proc/PID/smaps is an extension based on maps, showing the memory
401 consumption for each of the process's mappings. For each of mappings there
402 is a series of lines such as the following:
404 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash
417 Private_Hugetlb: 0 kB
423 VmFlags: rd ex mr mw me dw
425 the first of these lines shows the same information as is displayed for the
426 mapping in /proc/PID/maps. The remaining lines show the size of the mapping
427 (size), the amount of the mapping that is currently resident in RAM (RSS), the
428 process' proportional share of this mapping (PSS), the number of clean and
429 dirty private pages in the mapping.
431 The "proportional set size" (PSS) of a process is the count of pages it has
432 in memory, where each page is divided by the number of processes sharing it.
433 So if a process has 1000 pages all to itself, and 1000 shared with one other
434 process, its PSS will be 1500.
435 Note that even a page which is part of a MAP_SHARED mapping, but has only
436 a single pte mapped, i.e. is currently used by only one process, is accounted
437 as private and not as shared.
438 "Referenced" indicates the amount of memory currently marked as referenced or
440 "Anonymous" shows the amount of memory that does not belong to any file. Even
441 a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE
442 and a page is modified, the file page is replaced by a private anonymous copy.
443 "AnonHugePages" shows the ammount of memory backed by transparent hugepage.
444 "ShmemPmdMapped" shows the ammount of shared (shmem/tmpfs) memory backed by
446 "Shared_Hugetlb" and "Private_Hugetlb" show the ammounts of memory backed by
447 hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical
448 reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field.
449 "Swap" shows how much would-be-anonymous memory is also used, but out on swap.
450 For shmem mappings, "Swap" includes also the size of the mapped (and not
451 replaced by copy-on-write) part of the underlying shmem object out on swap.
452 "SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this
453 does not take into account swapped out page of underlying shmem objects.
454 "Locked" indicates whether the mapping is locked in memory or not.
456 "VmFlags" field deserves a separate description. This member represents the kernel
457 flags associated with the particular virtual memory area in two letter encoded
458 manner. The codes are the following:
467 gd - stack segment growns down
469 dw - disabled write to the mapped file
470 lo - pages are locked in memory
471 io - memory mapped I/O area
472 sr - sequential read advise provided
473 rr - random read advise provided
474 dc - do not copy area on fork
475 de - do not expand area on remapping
476 ac - area is accountable
477 nr - swap space is not reserved for the area
478 ht - area uses huge tlb pages
479 ar - architecture specific flag
480 dd - do not include area into core dump
483 hg - huge page advise flag
484 nh - no-huge page advise flag
485 mg - mergable advise flag
487 Note that there is no guarantee that every flag and associated mnemonic will
488 be present in all further kernel releases. Things get changed, the flags may
489 be vanished or the reverse -- new added.
491 This file is only present if the CONFIG_MMU kernel configuration option is
494 Note: reading /proc/PID/maps or /proc/PID/smaps is inherently racy (consistent
495 output can be achieved only in the single read call).
496 This typically manifests when doing partial reads of these files while the
497 memory map is being modified. Despite the races, we do provide the following
500 1) The mapped addresses never go backwards, which implies no two
501 regions will ever overlap.
502 2) If there is something at a given vaddr during the entirety of the
503 life of the smaps/maps walk, there will be some output for it.
506 The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG
507 bits on both physical and virtual pages associated with a process, and the
508 soft-dirty bit on pte (see Documentation/vm/soft-dirty.txt for details).
509 To clear the bits for all the pages associated with the process
510 > echo 1 > /proc/PID/clear_refs
512 To clear the bits for the anonymous pages associated with the process
513 > echo 2 > /proc/PID/clear_refs
515 To clear the bits for the file mapped pages associated with the process
516 > echo 3 > /proc/PID/clear_refs
518 To clear the soft-dirty bit
519 > echo 4 > /proc/PID/clear_refs
521 To reset the peak resident set size ("high water mark") to the process's
523 > echo 5 > /proc/PID/clear_refs
525 Any other value written to /proc/PID/clear_refs will have no effect.
527 The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags
528 using /proc/kpageflags and number of times a page is mapped using
529 /proc/kpagecount. For detailed explanation, see Documentation/vm/pagemap.txt.
531 The /proc/pid/numa_maps is an extension based on maps, showing the memory
532 locality and binding policy, as well as the memory usage (in pages) of
533 each mapping. The output follows a general format where mapping details get
534 summarized separated by blank spaces, one mapping per each file line:
536 address policy mapping details
538 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4
539 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4
540 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4
541 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
542 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
543 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4
544 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4
545 320698b000 default file=/lib64/libc-2.12.so
546 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4
547 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
548 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4
549 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4
550 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4
551 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048
552 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4
553 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4
556 "address" is the starting address for the mapping;
557 "policy" reports the NUMA memory policy set for the mapping (see vm/numa_memory_policy.txt);
558 "mapping details" summarizes mapping data such as mapping type, page usage counters,
559 node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page
560 size, in KB, that is backing the mapping up.
565 Similar to the process entries, the kernel data files give information about
566 the running kernel. The files used to obtain this information are contained in
567 /proc and are listed in Table 1-5. Not all of these will be present in your
568 system. It depends on the kernel configuration and the loaded modules, which
569 files are there, and which are missing.
571 Table 1-5: Kernel info in /proc
572 ..............................................................................
574 apm Advanced power management info
575 buddyinfo Kernel memory allocator information (see text) (2.5)
576 bus Directory containing bus specific information
577 cmdline Kernel command line
578 cpuinfo Info about the CPU
579 devices Available devices (block and character)
580 dma Used DMS channels
581 filesystems Supported filesystems
582 driver Various drivers grouped here, currently rtc (2.4)
583 execdomains Execdomains, related to security (2.4)
584 fb Frame Buffer devices (2.4)
585 fs File system parameters, currently nfs/exports (2.4)
586 ide Directory containing info about the IDE subsystem
587 interrupts Interrupt usage
588 iomem Memory map (2.4)
589 ioports I/O port usage
590 irq Masks for irq to cpu affinity (2.4)(smp?)
591 isapnp ISA PnP (Plug&Play) Info (2.4)
592 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
594 ksyms Kernel symbol table
595 loadavg Load average of last 1, 5 & 15 minutes
599 modules List of loaded modules
600 mounts Mounted filesystems
601 net Networking info (see text)
602 pagetypeinfo Additional page allocator information (see text) (2.5)
603 partitions Table of partitions known to the system
604 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
605 decoupled by lspci (2.4)
607 scsi SCSI info (see text)
608 slabinfo Slab pool info
609 softirqs softirq usage
610 stat Overall statistics
611 swaps Swap space utilization
613 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
614 tty Info of tty drivers
615 uptime Wall clock since boot, combined idle time of all cpus
616 version Kernel version
617 video bttv info of video resources (2.4)
618 vmallocinfo Show vmalloced areas
619 ..............................................................................
621 You can, for example, check which interrupts are currently in use and what
622 they are used for by looking in the file /proc/interrupts:
624 > cat /proc/interrupts
626 0: 8728810 XT-PIC timer
627 1: 895 XT-PIC keyboard
629 3: 531695 XT-PIC aha152x
630 4: 2014133 XT-PIC serial
631 5: 44401 XT-PIC pcnet_cs
634 12: 182918 XT-PIC PS/2 Mouse
636 14: 1232265 XT-PIC ide0
640 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
641 output of a SMP machine):
643 > cat /proc/interrupts
646 0: 1243498 1214548 IO-APIC-edge timer
647 1: 8949 8958 IO-APIC-edge keyboard
648 2: 0 0 XT-PIC cascade
649 5: 11286 10161 IO-APIC-edge soundblaster
650 8: 1 0 IO-APIC-edge rtc
651 9: 27422 27407 IO-APIC-edge 3c503
652 12: 113645 113873 IO-APIC-edge PS/2 Mouse
654 14: 22491 24012 IO-APIC-edge ide0
655 15: 2183 2415 IO-APIC-edge ide1
656 17: 30564 30414 IO-APIC-level eth0
657 18: 177 164 IO-APIC-level bttv
662 NMI is incremented in this case because every timer interrupt generates a NMI
663 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
665 LOC is the local interrupt counter of the internal APIC of every CPU.
667 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
668 connects the CPUs in a SMP system. This means that an error has been detected,
669 the IO-APIC automatically retry the transmission, so it should not be a big
670 problem, but you should read the SMP-FAQ.
672 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
673 /proc/interrupts to display every IRQ vector in use by the system, not
674 just those considered 'most important'. The new vectors are:
676 THR -- interrupt raised when a machine check threshold counter
677 (typically counting ECC corrected errors of memory or cache) exceeds
678 a configurable threshold. Only available on some systems.
680 TRM -- a thermal event interrupt occurs when a temperature threshold
681 has been exceeded for the CPU. This interrupt may also be generated
682 when the temperature drops back to normal.
684 SPU -- a spurious interrupt is some interrupt that was raised then lowered
685 by some IO device before it could be fully processed by the APIC. Hence
686 the APIC sees the interrupt but does not know what device it came from.
687 For this case the APIC will generate the interrupt with a IRQ vector
688 of 0xff. This might also be generated by chipset bugs.
690 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
691 sent from one CPU to another per the needs of the OS. Typically,
692 their statistics are used by kernel developers and interested users to
693 determine the occurrence of interrupts of the given type.
695 The above IRQ vectors are displayed only when relevant. For example,
696 the threshold vector does not exist on x86_64 platforms. Others are
697 suppressed when the system is a uniprocessor. As of this writing, only
698 i386 and x86_64 platforms support the new IRQ vector displays.
700 Of some interest is the introduction of the /proc/irq directory to 2.4.
701 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
702 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
703 irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and
708 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
709 1 11 13 15 17 19 3 5 7 9 default_smp_affinity
713 smp_affinity is a bitmask, in which you can specify which CPUs can handle the
714 IRQ, you can set it by doing:
716 > echo 1 > /proc/irq/10/smp_affinity
718 This means that only the first CPU will handle the IRQ, but you can also echo
719 5 which means that only the first and third CPU can handle the IRQ.
721 The contents of each smp_affinity file is the same by default:
723 > cat /proc/irq/0/smp_affinity
726 There is an alternate interface, smp_affinity_list which allows specifying
727 a cpu range instead of a bitmask:
729 > cat /proc/irq/0/smp_affinity_list
732 The default_smp_affinity mask applies to all non-active IRQs, which are the
733 IRQs which have not yet been allocated/activated, and hence which lack a
734 /proc/irq/[0-9]* directory.
736 The node file on an SMP system shows the node to which the device using the IRQ
737 reports itself as being attached. This hardware locality information does not
738 include information about any possible driver locality preference.
740 prof_cpu_mask specifies which CPUs are to be profiled by the system wide
741 profiler. Default value is ffffffff (all cpus if there are only 32 of them).
743 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
744 between all the CPUs which are allowed to handle it. As usual the kernel has
745 more info than you and does a better job than you, so the defaults are the
746 best choice for almost everyone. [Note this applies only to those IO-APIC's
747 that support "Round Robin" interrupt distribution.]
749 There are three more important subdirectories in /proc: net, scsi, and sys.
750 The general rule is that the contents, or even the existence of these
751 directories, depend on your kernel configuration. If SCSI is not enabled, the
752 directory scsi may not exist. The same is true with the net, which is there
753 only when networking support is present in the running kernel.
755 The slabinfo file gives information about memory usage at the slab level.
756 Linux uses slab pools for memory management above page level in version 2.2.
757 Commonly used objects have their own slab pool (such as network buffers,
758 directory cache, and so on).
760 ..............................................................................
762 > cat /proc/buddyinfo
764 Node 0, zone DMA 0 4 5 4 4 3 ...
765 Node 0, zone Normal 1 0 0 1 101 8 ...
766 Node 0, zone HighMem 2 0 0 1 1 0 ...
768 External fragmentation is a problem under some workloads, and buddyinfo is a
769 useful tool for helping diagnose these problems. Buddyinfo will give you a
770 clue as to how big an area you can safely allocate, or why a previous
773 Each column represents the number of pages of a certain order which are
774 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
775 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
776 available in ZONE_NORMAL, etc...
778 More information relevant to external fragmentation can be found in
781 > cat /proc/pagetypeinfo
785 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
786 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0
787 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0
788 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2
789 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0
790 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0
791 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9
792 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0
793 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452
794 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0
795 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0
797 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate
798 Node 0, zone DMA 2 0 5 1 0
799 Node 0, zone DMA32 41 6 967 2 0
801 Fragmentation avoidance in the kernel works by grouping pages of different
802 migrate types into the same contiguous regions of memory called page blocks.
803 A page block is typically the size of the default hugepage size e.g. 2MB on
804 X86-64. By keeping pages grouped based on their ability to move, the kernel
805 can reclaim pages within a page block to satisfy a high-order allocation.
807 The pagetypinfo begins with information on the size of a page block. It
808 then gives the same type of information as buddyinfo except broken down
809 by migrate-type and finishes with details on how many page blocks of each
812 If min_free_kbytes has been tuned correctly (recommendations made by hugeadm
813 from libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can
814 make an estimate of the likely number of huge pages that can be allocated
815 at a given point in time. All the "Movable" blocks should be allocatable
816 unless memory has been mlock()'d. Some of the Reclaimable blocks should
817 also be allocatable although a lot of filesystem metadata may have to be
818 reclaimed to achieve this.
820 ..............................................................................
824 Provides information about distribution and utilization of memory. This
825 varies by architecture and compile options. The following is from a
826 16GB PIII, which has highmem enabled. You may not have all of these fields.
830 MemTotal: 16344972 kB
832 MemAvailable: 14836172 kB
838 HighTotal: 15597528 kB
839 HighFree: 13629632 kB
850 SReclaimable: 159856 kB
851 SUnreclaim: 124508 kB
856 CommitLimit: 7669796 kB
857 Committed_AS: 100056 kB
858 VmallocTotal: 112216 kB
860 VmallocChunk: 111088 kB
861 AnonHugePages: 49152 kB
866 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
867 bits and the kernel binary code)
868 MemFree: The sum of LowFree+HighFree
869 MemAvailable: An estimate of how much memory is available for starting new
870 applications, without swapping. Calculated from MemFree,
871 SReclaimable, the size of the file LRU lists, and the low
872 watermarks in each zone.
873 The estimate takes into account that the system needs some
874 page cache to function well, and that not all reclaimable
875 slab will be reclaimable, due to items being in use. The
876 impact of those factors will vary from system to system.
877 Buffers: Relatively temporary storage for raw disk blocks
878 shouldn't get tremendously large (20MB or so)
879 Cached: in-memory cache for files read from the disk (the
880 pagecache). Doesn't include SwapCached
881 SwapCached: Memory that once was swapped out, is swapped back in but
882 still also is in the swapfile (if memory is needed it
883 doesn't need to be swapped out AGAIN because it is already
884 in the swapfile. This saves I/O)
885 Active: Memory that has been used more recently and usually not
886 reclaimed unless absolutely necessary.
887 Inactive: Memory which has been less recently used. It is more
888 eligible to be reclaimed for other purposes
890 HighFree: Highmem is all memory above ~860MB of physical memory
891 Highmem areas are for use by userspace programs, or
892 for the pagecache. The kernel must use tricks to access
893 this memory, making it slower to access than lowmem.
895 LowFree: Lowmem is memory which can be used for everything that
896 highmem can be used for, but it is also available for the
897 kernel's use for its own data structures. Among many
898 other things, it is where everything from the Slab is
899 allocated. Bad things happen when you're out of lowmem.
900 SwapTotal: total amount of swap space available
901 SwapFree: Memory which has been evicted from RAM, and is temporarily
903 Dirty: Memory which is waiting to get written back to the disk
904 Writeback: Memory which is actively being written back to the disk
905 AnonPages: Non-file backed pages mapped into userspace page tables
906 AnonHugePages: Non-file backed huge pages mapped into userspace page tables
907 Mapped: files which have been mmaped, such as libraries
908 Shmem: Total memory used by shared memory (shmem) and tmpfs
909 ShmemHugePages: Memory used by shared memory (shmem) and tmpfs allocated
911 ShmemPmdMapped: Shared memory mapped into userspace with huge pages
912 Slab: in-kernel data structures cache
913 SReclaimable: Part of Slab, that might be reclaimed, such as caches
914 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure
915 PageTables: amount of memory dedicated to the lowest level of page
917 NFS_Unstable: NFS pages sent to the server, but not yet committed to stable
919 Bounce: Memory used for block device "bounce buffers"
920 WritebackTmp: Memory used by FUSE for temporary writeback buffers
921 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
922 this is the total amount of memory currently available to
923 be allocated on the system. This limit is only adhered to
924 if strict overcommit accounting is enabled (mode 2 in
925 'vm.overcommit_memory').
926 The CommitLimit is calculated with the following formula:
927 CommitLimit = ([total RAM pages] - [total huge TLB pages]) *
928 overcommit_ratio / 100 + [total swap pages]
929 For example, on a system with 1G of physical RAM and 7G
930 of swap with a `vm.overcommit_ratio` of 30 it would
931 yield a CommitLimit of 7.3G.
932 For more details, see the memory overcommit documentation
933 in vm/overcommit-accounting.
934 Committed_AS: The amount of memory presently allocated on the system.
935 The committed memory is a sum of all of the memory which
936 has been allocated by processes, even if it has not been
937 "used" by them as of yet. A process which malloc()'s 1G
938 of memory, but only touches 300M of it will show up as
939 using 1G. This 1G is memory which has been "committed" to
940 by the VM and can be used at any time by the allocating
941 application. With strict overcommit enabled on the system
942 (mode 2 in 'vm.overcommit_memory'),allocations which would
943 exceed the CommitLimit (detailed above) will not be permitted.
944 This is useful if one needs to guarantee that processes will
945 not fail due to lack of memory once that memory has been
946 successfully allocated.
947 VmallocTotal: total size of vmalloc memory area
948 VmallocUsed: amount of vmalloc area which is used
949 VmallocChunk: largest contiguous block of vmalloc area which is free
951 ..............................................................................
955 Provides information about vmalloced/vmaped areas. One line per area,
956 containing the virtual address range of the area, size in bytes,
957 caller information of the creator, and optional information depending
958 on the kind of area :
960 pages=nr number of pages
961 phys=addr if a physical address was specified
962 ioremap I/O mapping (ioremap() and friends)
963 vmalloc vmalloc() area
966 vpages buffer for pages pointers was vmalloced (huge area)
967 N<node>=nr (Only on NUMA kernels)
968 Number of pages allocated on memory node <node>
970 > cat /proc/vmallocinfo
971 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ...
972 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128
973 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ...
974 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64
975 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f...
976 phys=7fee8000 ioremap
977 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f...
978 phys=7fee7000 ioremap
979 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210
980 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ...
981 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3
982 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ...
984 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ...
985 /0x130 [x_tables] pages=4 vmalloc N0=4
986 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ...
987 pages=14 vmalloc N2=14
988 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ...
990 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ...
992 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ...
993 pages=10 vmalloc N0=10
995 ..............................................................................
999 Provides counts of softirq handlers serviced since boot time, for each cpu.
1001 > cat /proc/softirqs
1004 TIMER: 27166 27120 27097 27034
1009 SCHED: 27035 26983 26971 26746
1011 RCU: 1678 1769 2178 2250
1014 1.3 IDE devices in /proc/ide
1015 ----------------------------
1017 The subdirectory /proc/ide contains information about all IDE devices of which
1018 the kernel is aware. There is one subdirectory for each IDE controller, the
1019 file drivers and a link for each IDE device, pointing to the device directory
1020 in the controller specific subtree.
1022 The file drivers contains general information about the drivers used for the
1025 > cat /proc/ide/drivers
1026 ide-cdrom version 4.53
1027 ide-disk version 1.08
1029 More detailed information can be found in the controller specific
1030 subdirectories. These are named ide0, ide1 and so on. Each of these
1031 directories contains the files shown in table 1-6.
1034 Table 1-6: IDE controller info in /proc/ide/ide?
1035 ..............................................................................
1037 channel IDE channel (0 or 1)
1038 config Configuration (only for PCI/IDE bridge)
1040 model Type/Chipset of IDE controller
1041 ..............................................................................
1043 Each device connected to a controller has a separate subdirectory in the
1044 controllers directory. The files listed in table 1-7 are contained in these
1048 Table 1-7: IDE device information
1049 ..............................................................................
1052 capacity Capacity of the medium (in 512Byte blocks)
1053 driver driver and version
1054 geometry physical and logical geometry
1055 identify device identify block
1057 model device identifier
1058 settings device setup
1059 smart_thresholds IDE disk management thresholds
1060 smart_values IDE disk management values
1061 ..............................................................................
1063 The most interesting file is settings. This file contains a nice overview of
1064 the drive parameters:
1066 # cat /proc/ide/ide0/hda/settings
1067 name value min max mode
1068 ---- ----- --- --- ----
1069 bios_cyl 526 0 65535 rw
1070 bios_head 255 0 255 rw
1071 bios_sect 63 0 63 rw
1072 breada_readahead 4 0 127 rw
1074 file_readahead 72 0 2097151 rw
1076 keepsettings 0 0 1 rw
1077 max_kb_per_request 122 1 127 rw
1081 pio_mode write-only 0 255 w
1087 1.4 Networking info in /proc/net
1088 --------------------------------
1090 The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the
1091 additional values you get for IP version 6 if you configure the kernel to
1092 support this. Table 1-9 lists the files and their meaning.
1095 Table 1-8: IPv6 info in /proc/net
1096 ..............................................................................
1098 udp6 UDP sockets (IPv6)
1099 tcp6 TCP sockets (IPv6)
1100 raw6 Raw device statistics (IPv6)
1101 igmp6 IP multicast addresses, which this host joined (IPv6)
1102 if_inet6 List of IPv6 interface addresses
1103 ipv6_route Kernel routing table for IPv6
1104 rt6_stats Global IPv6 routing tables statistics
1105 sockstat6 Socket statistics (IPv6)
1106 snmp6 Snmp data (IPv6)
1107 ..............................................................................
1110 Table 1-9: Network info in /proc/net
1111 ..............................................................................
1113 arp Kernel ARP table
1114 dev network devices with statistics
1115 dev_mcast the Layer2 multicast groups a device is listening too
1116 (interface index, label, number of references, number of bound
1118 dev_stat network device status
1119 ip_fwchains Firewall chain linkage
1120 ip_fwnames Firewall chain names
1121 ip_masq Directory containing the masquerading tables
1122 ip_masquerade Major masquerading table
1123 netstat Network statistics
1124 raw raw device statistics
1125 route Kernel routing table
1126 rpc Directory containing rpc info
1127 rt_cache Routing cache
1129 sockstat Socket statistics
1132 unix UNIX domain sockets
1133 wireless Wireless interface data (Wavelan etc)
1134 igmp IP multicast addresses, which this host joined
1135 psched Global packet scheduler parameters.
1136 netlink List of PF_NETLINK sockets
1137 ip_mr_vifs List of multicast virtual interfaces
1138 ip_mr_cache List of multicast routing cache
1139 ..............................................................................
1141 You can use this information to see which network devices are available in
1142 your system and how much traffic was routed over those devices:
1145 Inter-|Receive |[...
1146 face |bytes packets errs drop fifo frame compressed multicast|[...
1147 lo: 908188 5596 0 0 0 0 0 0 [...
1148 ppp0:15475140 20721 410 0 0 410 0 0 [...
1149 eth0: 614530 7085 0 0 0 0 0 1 [...
1152 ...] bytes packets errs drop fifo colls carrier compressed
1153 ...] 908188 5596 0 0 0 0 0 0
1154 ...] 1375103 17405 0 0 0 0 0 0
1155 ...] 1703981 5535 0 0 0 3 0 0
1157 In addition, each Channel Bond interface has its own directory. For
1158 example, the bond0 device will have a directory called /proc/net/bond0/.
1159 It will contain information that is specific to that bond, such as the
1160 current slaves of the bond, the link status of the slaves, and how
1161 many times the slaves link has failed.
1166 If you have a SCSI host adapter in your system, you'll find a subdirectory
1167 named after the driver for this adapter in /proc/scsi. You'll also see a list
1168 of all recognized SCSI devices in /proc/scsi:
1170 >cat /proc/scsi/scsi
1172 Host: scsi0 Channel: 00 Id: 00 Lun: 00
1173 Vendor: IBM Model: DGHS09U Rev: 03E0
1174 Type: Direct-Access ANSI SCSI revision: 03
1175 Host: scsi0 Channel: 00 Id: 06 Lun: 00
1176 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
1177 Type: CD-ROM ANSI SCSI revision: 02
1180 The directory named after the driver has one file for each adapter found in
1181 the system. These files contain information about the controller, including
1182 the used IRQ and the IO address range. The amount of information shown is
1183 dependent on the adapter you use. The example shows the output for an Adaptec
1184 AHA-2940 SCSI adapter:
1186 > cat /proc/scsi/aic7xxx/0
1188 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
1190 TCQ Enabled By Default : Disabled
1191 AIC7XXX_PROC_STATS : Disabled
1192 AIC7XXX_RESET_DELAY : 5
1193 Adapter Configuration:
1194 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
1195 Ultra Wide Controller
1196 PCI MMAPed I/O Base: 0xeb001000
1197 Adapter SEEPROM Config: SEEPROM found and used.
1198 Adaptec SCSI BIOS: Enabled
1200 SCBs: Active 0, Max Active 2,
1201 Allocated 15, HW 16, Page 255
1203 BIOS Control Word: 0x18b6
1204 Adapter Control Word: 0x005b
1205 Extended Translation: Enabled
1206 Disconnect Enable Flags: 0xffff
1207 Ultra Enable Flags: 0x0001
1208 Tag Queue Enable Flags: 0x0000
1209 Ordered Queue Tag Flags: 0x0000
1210 Default Tag Queue Depth: 8
1211 Tagged Queue By Device array for aic7xxx host instance 0:
1212 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
1213 Actual queue depth per device for aic7xxx host instance 0:
1214 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
1217 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
1218 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
1219 Total transfers 160151 (74577 reads and 85574 writes)
1221 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
1222 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
1223 Total transfers 0 (0 reads and 0 writes)
1226 1.6 Parallel port info in /proc/parport
1227 ---------------------------------------
1229 The directory /proc/parport contains information about the parallel ports of
1230 your system. It has one subdirectory for each port, named after the port
1233 These directories contain the four files shown in Table 1-10.
1236 Table 1-10: Files in /proc/parport
1237 ..............................................................................
1239 autoprobe Any IEEE-1284 device ID information that has been acquired.
1240 devices list of the device drivers using that port. A + will appear by the
1241 name of the device currently using the port (it might not appear
1243 hardware Parallel port's base address, IRQ line and DMA channel.
1244 irq IRQ that parport is using for that port. This is in a separate
1245 file to allow you to alter it by writing a new value in (IRQ
1247 ..............................................................................
1249 1.7 TTY info in /proc/tty
1250 -------------------------
1252 Information about the available and actually used tty's can be found in the
1253 directory /proc/tty.You'll find entries for drivers and line disciplines in
1254 this directory, as shown in Table 1-11.
1257 Table 1-11: Files in /proc/tty
1258 ..............................................................................
1260 drivers list of drivers and their usage
1261 ldiscs registered line disciplines
1262 driver/serial usage statistic and status of single tty lines
1263 ..............................................................................
1265 To see which tty's are currently in use, you can simply look into the file
1268 > cat /proc/tty/drivers
1269 pty_slave /dev/pts 136 0-255 pty:slave
1270 pty_master /dev/ptm 128 0-255 pty:master
1271 pty_slave /dev/ttyp 3 0-255 pty:slave
1272 pty_master /dev/pty 2 0-255 pty:master
1273 serial /dev/cua 5 64-67 serial:callout
1274 serial /dev/ttyS 4 64-67 serial
1275 /dev/tty0 /dev/tty0 4 0 system:vtmaster
1276 /dev/ptmx /dev/ptmx 5 2 system
1277 /dev/console /dev/console 5 1 system:console
1278 /dev/tty /dev/tty 5 0 system:/dev/tty
1279 unknown /dev/tty 4 1-63 console
1282 1.8 Miscellaneous kernel statistics in /proc/stat
1283 -------------------------------------------------
1285 Various pieces of information about kernel activity are available in the
1286 /proc/stat file. All of the numbers reported in this file are aggregates
1287 since the system first booted. For a quick look, simply cat the file:
1290 cpu 2255 34 2290 22625563 6290 127 456 0 0 0
1291 cpu0 1132 34 1441 11311718 3675 127 438 0 0 0
1292 cpu1 1123 0 849 11313845 2614 0 18 0 0 0
1293 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
1299 softirq 183433 0 21755 12 39 1137 231 21459 2263
1301 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
1302 lines. These numbers identify the amount of time the CPU has spent performing
1303 different kinds of work. Time units are in USER_HZ (typically hundredths of a
1304 second). The meanings of the columns are as follows, from left to right:
1306 - user: normal processes executing in user mode
1307 - nice: niced processes executing in user mode
1308 - system: processes executing in kernel mode
1309 - idle: twiddling thumbs
1310 - iowait: In a word, iowait stands for waiting for I/O to complete. But there
1311 are several problems:
1312 1. Cpu will not wait for I/O to complete, iowait is the time that a task is
1313 waiting for I/O to complete. When cpu goes into idle state for
1314 outstanding task io, another task will be scheduled on this CPU.
1315 2. In a multi-core CPU, the task waiting for I/O to complete is not running
1316 on any CPU, so the iowait of each CPU is difficult to calculate.
1317 3. The value of iowait field in /proc/stat will decrease in certain
1319 So, the iowait is not reliable by reading from /proc/stat.
1320 - irq: servicing interrupts
1321 - softirq: servicing softirqs
1322 - steal: involuntary wait
1323 - guest: running a normal guest
1324 - guest_nice: running a niced guest
1326 The "intr" line gives counts of interrupts serviced since boot time, for each
1327 of the possible system interrupts. The first column is the total of all
1328 interrupts serviced including unnumbered architecture specific interrupts;
1329 each subsequent column is the total for that particular numbered interrupt.
1330 Unnumbered interrupts are not shown, only summed into the total.
1332 The "ctxt" line gives the total number of context switches across all CPUs.
1334 The "btime" line gives the time at which the system booted, in seconds since
1337 The "processes" line gives the number of processes and threads created, which
1338 includes (but is not limited to) those created by calls to the fork() and
1339 clone() system calls.
1341 The "procs_running" line gives the total number of threads that are
1342 running or ready to run (i.e., the total number of runnable threads).
1344 The "procs_blocked" line gives the number of processes currently blocked,
1345 waiting for I/O to complete.
1347 The "softirq" line gives counts of softirqs serviced since boot time, for each
1348 of the possible system softirqs. The first column is the total of all
1349 softirqs serviced; each subsequent column is the total for that particular
1353 1.9 Ext4 file system parameters
1354 -------------------------------
1356 Information about mounted ext4 file systems can be found in
1357 /proc/fs/ext4. Each mounted filesystem will have a directory in
1358 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
1359 /proc/fs/ext4/dm-0). The files in each per-device directory are shown
1360 in Table 1-12, below.
1362 Table 1-12: Files in /proc/fs/ext4/<devname>
1363 ..............................................................................
1365 mb_groups details of multiblock allocator buddy cache of free blocks
1366 ..............................................................................
1370 Shows registered system console lines.
1372 To see which character device lines are currently used for the system console
1373 /dev/console, you may simply look into the file /proc/consoles:
1375 > cat /proc/consoles
1381 device name of the device
1382 operations R = can do read operations
1383 W = can do write operations
1385 flags E = it is enabled
1386 C = it is preferred console
1387 B = it is primary boot console
1388 p = it is used for printk buffer
1389 b = it is not a TTY but a Braille device
1390 a = it is safe to use when cpu is offline
1391 major:minor major and minor number of the device separated by a colon
1393 ------------------------------------------------------------------------------
1395 ------------------------------------------------------------------------------
1396 The /proc file system serves information about the running system. It not only
1397 allows access to process data but also allows you to request the kernel status
1398 by reading files in the hierarchy.
1400 The directory structure of /proc reflects the types of information and makes
1401 it easy, if not obvious, where to look for specific data.
1402 ------------------------------------------------------------------------------
1404 ------------------------------------------------------------------------------
1405 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
1406 ------------------------------------------------------------------------------
1408 ------------------------------------------------------------------------------
1410 ------------------------------------------------------------------------------
1411 * Modifying kernel parameters by writing into files found in /proc/sys
1412 * Exploring the files which modify certain parameters
1413 * Review of the /proc/sys file tree
1414 ------------------------------------------------------------------------------
1417 A very interesting part of /proc is the directory /proc/sys. This is not only
1418 a source of information, it also allows you to change parameters within the
1419 kernel. Be very careful when attempting this. You can optimize your system,
1420 but you can also cause it to crash. Never alter kernel parameters on a
1421 production system. Set up a development machine and test to make sure that
1422 everything works the way you want it to. You may have no alternative but to
1423 reboot the machine once an error has been made.
1425 To change a value, simply echo the new value into the file. An example is
1426 given below in the section on the file system data. You need to be root to do
1427 this. You can create your own boot script to perform this every time your
1430 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
1431 general things in the operation of the Linux kernel. Since some of the files
1432 can inadvertently disrupt your system, it is advisable to read both
1433 documentation and source before actually making adjustments. In any case, be
1434 very careful when writing to any of these files. The entries in /proc may
1435 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
1436 review the kernel documentation in the directory /usr/src/linux/Documentation.
1437 This chapter is heavily based on the documentation included in the pre 2.2
1438 kernels, and became part of it in version 2.2.1 of the Linux kernel.
1440 Please see: Documentation/sysctl/ directory for descriptions of these
1443 ------------------------------------------------------------------------------
1445 ------------------------------------------------------------------------------
1446 Certain aspects of kernel behavior can be modified at runtime, without the
1447 need to recompile the kernel, or even to reboot the system. The files in the
1448 /proc/sys tree can not only be read, but also modified. You can use the echo
1449 command to write value into these files, thereby changing the default settings
1451 ------------------------------------------------------------------------------
1453 ------------------------------------------------------------------------------
1454 CHAPTER 3: PER-PROCESS PARAMETERS
1455 ------------------------------------------------------------------------------
1457 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score
1458 --------------------------------------------------------------------------------
1460 These file can be used to adjust the badness heuristic used to select which
1461 process gets killed in out of memory conditions.
1463 The badness heuristic assigns a value to each candidate task ranging from 0
1464 (never kill) to 1000 (always kill) to determine which process is targeted. The
1465 units are roughly a proportion along that range of allowed memory the process
1466 may allocate from based on an estimation of its current memory and swap use.
1467 For example, if a task is using all allowed memory, its badness score will be
1468 1000. If it is using half of its allowed memory, its score will be 500.
1470 There is an additional factor included in the badness score: the current memory
1471 and swap usage is discounted by 3% for root processes.
1473 The amount of "allowed" memory depends on the context in which the oom killer
1474 was called. If it is due to the memory assigned to the allocating task's cpuset
1475 being exhausted, the allowed memory represents the set of mems assigned to that
1476 cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed
1477 memory represents the set of mempolicy nodes. If it is due to a memory
1478 limit (or swap limit) being reached, the allowed memory is that configured
1479 limit. Finally, if it is due to the entire system being out of memory, the
1480 allowed memory represents all allocatable resources.
1482 The value of /proc/<pid>/oom_score_adj is added to the badness score before it
1483 is used to determine which task to kill. Acceptable values range from -1000
1484 (OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to
1485 polarize the preference for oom killing either by always preferring a certain
1486 task or completely disabling it. The lowest possible value, -1000, is
1487 equivalent to disabling oom killing entirely for that task since it will always
1488 report a badness score of 0.
1490 Consequently, it is very simple for userspace to define the amount of memory to
1491 consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for
1492 example, is roughly equivalent to allowing the remainder of tasks sharing the
1493 same system, cpuset, mempolicy, or memory controller resources to use at least
1494 50% more memory. A value of -500, on the other hand, would be roughly
1495 equivalent to discounting 50% of the task's allowed memory from being considered
1496 as scoring against the task.
1498 For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also
1499 be used to tune the badness score. Its acceptable values range from -16
1500 (OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17
1501 (OOM_DISABLE) to disable oom killing entirely for that task. Its value is
1502 scaled linearly with /proc/<pid>/oom_score_adj.
1504 The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last
1505 value set by a CAP_SYS_RESOURCE process. To reduce the value any lower
1506 requires CAP_SYS_RESOURCE.
1508 Caveat: when a parent task is selected, the oom killer will sacrifice any first
1509 generation children with separate address spaces instead, if possible. This
1510 avoids servers and important system daemons from being killed and loses the
1511 minimal amount of work.
1514 3.2 /proc/<pid>/oom_score - Display current oom-killer score
1515 -------------------------------------------------------------
1517 This file can be used to check the current score used by the oom-killer is for
1518 any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which
1519 process should be killed in an out-of-memory situation.
1522 3.3 /proc/<pid>/io - Display the IO accounting fields
1523 -------------------------------------------------------
1525 This file contains IO statistics for each running process
1530 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
1533 test:/tmp # cat /proc/3828/io
1539 write_bytes: 323932160
1540 cancelled_write_bytes: 0
1549 I/O counter: chars read
1550 The number of bytes which this task has caused to be read from storage. This
1551 is simply the sum of bytes which this process passed to read() and pread().
1552 It includes things like tty IO and it is unaffected by whether or not actual
1553 physical disk IO was required (the read might have been satisfied from
1560 I/O counter: chars written
1561 The number of bytes which this task has caused, or shall cause to be written
1562 to disk. Similar caveats apply here as with rchar.
1568 I/O counter: read syscalls
1569 Attempt to count the number of read I/O operations, i.e. syscalls like read()
1576 I/O counter: write syscalls
1577 Attempt to count the number of write I/O operations, i.e. syscalls like
1578 write() and pwrite().
1584 I/O counter: bytes read
1585 Attempt to count the number of bytes which this process really did cause to
1586 be fetched from the storage layer. Done at the submit_bio() level, so it is
1587 accurate for block-backed filesystems. <please add status regarding NFS and
1588 CIFS at a later time>
1594 I/O counter: bytes written
1595 Attempt to count the number of bytes which this process caused to be sent to
1596 the storage layer. This is done at page-dirtying time.
1599 cancelled_write_bytes
1600 ---------------------
1602 The big inaccuracy here is truncate. If a process writes 1MB to a file and
1603 then deletes the file, it will in fact perform no writeout. But it will have
1604 been accounted as having caused 1MB of write.
1605 In other words: The number of bytes which this process caused to not happen,
1606 by truncating pagecache. A task can cause "negative" IO too. If this task
1607 truncates some dirty pagecache, some IO which another task has been accounted
1608 for (in its write_bytes) will not be happening. We _could_ just subtract that
1609 from the truncating task's write_bytes, but there is information loss in doing
1616 At its current implementation state, this is a bit racy on 32-bit machines: if
1617 process A reads process B's /proc/pid/io while process B is updating one of
1618 those 64-bit counters, process A could see an intermediate result.
1621 More information about this can be found within the taskstats documentation in
1622 Documentation/accounting.
1624 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings
1625 ---------------------------------------------------------------
1626 When a process is dumped, all anonymous memory is written to a core file as
1627 long as the size of the core file isn't limited. But sometimes we don't want
1628 to dump some memory segments, for example, huge shared memory or DAX.
1629 Conversely, sometimes we want to save file-backed memory segments into a core
1630 file, not only the individual files.
1632 /proc/<pid>/coredump_filter allows you to customize which memory segments
1633 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
1634 of memory types. If a bit of the bitmask is set, memory segments of the
1635 corresponding memory type are dumped, otherwise they are not dumped.
1637 The following 9 memory types are supported:
1638 - (bit 0) anonymous private memory
1639 - (bit 1) anonymous shared memory
1640 - (bit 2) file-backed private memory
1641 - (bit 3) file-backed shared memory
1642 - (bit 4) ELF header pages in file-backed private memory areas (it is
1643 effective only if the bit 2 is cleared)
1644 - (bit 5) hugetlb private memory
1645 - (bit 6) hugetlb shared memory
1646 - (bit 7) DAX private memory
1647 - (bit 8) DAX shared memory
1649 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
1650 are always dumped regardless of the bitmask status.
1652 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is
1653 only affected by bit 5-6, and DAX is only affected by bits 7-8.
1655 The default value of coredump_filter is 0x33; this means all anonymous memory
1656 segments, ELF header pages and hugetlb private memory are dumped.
1658 If you don't want to dump all shared memory segments attached to pid 1234,
1659 write 0x31 to the process's proc file.
1661 $ echo 0x31 > /proc/1234/coredump_filter
1663 When a new process is created, the process inherits the bitmask status from its
1664 parent. It is useful to set up coredump_filter before the program runs.
1667 $ echo 0x7 > /proc/self/coredump_filter
1670 3.5 /proc/<pid>/mountinfo - Information about mounts
1671 --------------------------------------------------------
1673 This file contains lines of the form:
1675 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
1676 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
1678 (1) mount ID: unique identifier of the mount (may be reused after umount)
1679 (2) parent ID: ID of parent (or of self for the top of the mount tree)
1680 (3) major:minor: value of st_dev for files on filesystem
1681 (4) root: root of the mount within the filesystem
1682 (5) mount point: mount point relative to the process's root
1683 (6) mount options: per mount options
1684 (7) optional fields: zero or more fields of the form "tag[:value]"
1685 (8) separator: marks the end of the optional fields
1686 (9) filesystem type: name of filesystem of the form "type[.subtype]"
1687 (10) mount source: filesystem specific information or "none"
1688 (11) super options: per super block options
1690 Parsers should ignore all unrecognised optional fields. Currently the
1691 possible optional fields are:
1693 shared:X mount is shared in peer group X
1694 master:X mount is slave to peer group X
1695 propagate_from:X mount is slave and receives propagation from peer group X (*)
1696 unbindable mount is unbindable
1698 (*) X is the closest dominant peer group under the process's root. If
1699 X is the immediate master of the mount, or if there's no dominant peer
1700 group under the same root, then only the "master:X" field is present
1701 and not the "propagate_from:X" field.
1703 For more information on mount propagation see:
1705 Documentation/filesystems/sharedsubtree.txt
1708 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm
1709 --------------------------------------------------------
1710 These files provide a method to access a tasks comm value. It also allows for
1711 a task to set its own or one of its thread siblings comm value. The comm value
1712 is limited in size compared to the cmdline value, so writing anything longer
1713 then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated
1717 3.7 /proc/<pid>/task/<tid>/children - Information about task children
1718 -------------------------------------------------------------------------
1719 This file provides a fast way to retrieve first level children pids
1720 of a task pointed by <pid>/<tid> pair. The format is a space separated
1723 Note the "first level" here -- if a child has own children they will
1724 not be listed here, one needs to read /proc/<children-pid>/task/<tid>/children
1725 to obtain the descendants.
1727 Since this interface is intended to be fast and cheap it doesn't
1728 guarantee to provide precise results and some children might be
1729 skipped, especially if they've exited right after we printed their
1730 pids, so one need to either stop or freeze processes being inspected
1731 if precise results are needed.
1734 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file
1735 ---------------------------------------------------------------
1736 This file provides information associated with an opened file. The regular
1737 files have at least three fields -- 'pos', 'flags' and mnt_id. The 'pos'
1738 represents the current offset of the opened file in decimal form [see lseek(2)
1739 for details], 'flags' denotes the octal O_xxx mask the file has been
1740 created with [see open(2) for details] and 'mnt_id' represents mount ID of
1741 the file system containing the opened file [see 3.5 /proc/<pid>/mountinfo
1750 All locks associated with a file descriptor are shown in its fdinfo too.
1752 lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF
1754 The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags
1755 pair provide additional information particular to the objects they represent.
1764 where 'eventfd-count' is hex value of a counter.
1771 sigmask: 0000000000000200
1773 where 'sigmask' is hex value of the signal mask associated
1781 tfd: 5 events: 1d data: ffffffffffffffff
1783 where 'tfd' is a target file descriptor number in decimal form,
1784 'events' is events mask being watched and the 'data' is data
1785 associated with a target [see epoll(7) for more details].
1789 For inotify files the format is the following
1793 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d
1795 where 'wd' is a watch descriptor in decimal form, ie a target file
1796 descriptor number, 'ino' and 'sdev' are inode and device where the
1797 target file resides and the 'mask' is the mask of events, all in hex
1798 form [see inotify(7) for more details].
1800 If the kernel was built with exportfs support, the path to the target
1801 file is encoded as a file handle. The file handle is provided by three
1802 fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex
1805 If the kernel is built without exportfs support the file handle won't be
1808 If there is no inotify mark attached yet the 'inotify' line will be omitted.
1810 For fanotify files the format is
1815 fanotify flags:10 event-flags:0
1816 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003
1817 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4
1819 where fanotify 'flags' and 'event-flags' are values used in fanotify_init
1820 call, 'mnt_id' is the mount point identifier, 'mflags' is the value of
1821 flags associated with mark which are tracked separately from events
1822 mask. 'ino', 'sdev' are target inode and device, 'mask' is the events
1823 mask and 'ignored_mask' is the mask of events which are to be ignored.
1824 All in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask'
1825 does provide information about flags and mask used in fanotify_mark
1826 call [see fsnotify manpage for details].
1828 While the first three lines are mandatory and always printed, the rest is
1829 optional and may be omitted if no marks created yet.
1840 it_value: (0, 49406829)
1843 where 'clockid' is the clock type and 'ticks' is the number of the timer expirations
1844 that have occurred [see timerfd_create(2) for details]. 'settime flags' are
1845 flags in octal form been used to setup the timer [see timerfd_settime(2) for
1846 details]. 'it_value' is remaining time until the timer exiration.
1847 'it_interval' is the interval for the timer. Note the timer might be set up
1848 with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value'
1849 still exhibits timer's remaining time.
1851 3.9 /proc/<pid>/map_files - Information about memory mapped files
1852 ---------------------------------------------------------------------
1853 This directory contains symbolic links which represent memory mapped files
1854 the process is maintaining. Example output:
1856 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so
1857 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so
1858 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so
1860 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1
1861 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls
1863 The name of a link represents the virtual memory bounds of a mapping, i.e.
1864 vm_area_struct::vm_start-vm_area_struct::vm_end.
1866 The main purpose of the map_files is to retrieve a set of memory mapped
1867 files in a fast way instead of parsing /proc/<pid>/maps or
1868 /proc/<pid>/smaps, both of which contain many more records. At the same
1869 time one can open(2) mappings from the listings of two processes and
1870 comparing their inode numbers to figure out which anonymous memory areas
1871 are actually shared.
1873 3.10 /proc/<pid>/timerslack_ns - Task timerslack value
1874 ---------------------------------------------------------
1875 This file provides the value of the task's timerslack value in nanoseconds.
1876 This value specifies a amount of time that normal timers may be deferred
1877 in order to coalesce timers and avoid unnecessary wakeups.
1879 This allows a task's interactivity vs power consumption trade off to be
1882 Writing 0 to the file will set the tasks timerslack to the default value.
1884 Valid values are from 0 - ULLONG_MAX
1886 An application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level
1887 permissions on the task specified to change its timerslack_ns value.
1890 ------------------------------------------------------------------------------
1892 ------------------------------------------------------------------------------
1895 ---------------------
1897 The following mount options are supported:
1899 hidepid= Set /proc/<pid>/ access mode.
1900 gid= Set the group authorized to learn processes information.
1902 hidepid=0 means classic mode - everybody may access all /proc/<pid>/ directories
1905 hidepid=1 means users may not access any /proc/<pid>/ directories but their
1906 own. Sensitive files like cmdline, sched*, status are now protected against
1907 other users. This makes it impossible to learn whether any user runs
1908 specific program (given the program doesn't reveal itself by its behaviour).
1909 As an additional bonus, as /proc/<pid>/cmdline is unaccessible for other users,
1910 poorly written programs passing sensitive information via program arguments are
1911 now protected against local eavesdroppers.
1913 hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be fully invisible to other
1914 users. It doesn't mean that it hides a fact whether a process with a specific
1915 pid value exists (it can be learned by other means, e.g. by "kill -0 $PID"),
1916 but it hides process' uid and gid, which may be learned by stat()'ing
1917 /proc/<pid>/ otherwise. It greatly complicates an intruder's task of gathering
1918 information about running processes, whether some daemon runs with elevated
1919 privileges, whether other user runs some sensitive program, whether other users
1920 run any program at all, etc.
1922 gid= defines a group authorized to learn processes information otherwise
1923 prohibited by hidepid=. If you use some daemon like identd which needs to learn
1924 information about processes information, just add identd to this group.
projects / linux-2.6-block.git / blob