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 a 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
182 SigPnd: 0000000000000000
183 ShdPnd: 0000000000000000
184 SigBlk: 0000000000000000
185 SigIgn: 0000000000000000
186 SigCgt: 0000000000000000
187 CapInh: 00000000fffffeff
188 CapPrm: 0000000000000000
189 CapEff: 0000000000000000
190 CapBnd: ffffffffffffffff
192 voluntary_ctxt_switches: 0
193 nonvoluntary_ctxt_switches: 1
195 This shows you nearly the same information you would get if you viewed it with
196 the ps command. In fact, ps uses the proc file system to obtain its
197 information. But you get a more detailed view of the process by reading the
198 file /proc/PID/status. It fields are described in table 1-2.
200 The statm file contains more detailed information about the process
201 memory usage. Its seven fields are explained in Table 1-3. The stat file
202 contains details information about the process itself. Its fields are
203 explained in Table 1-4.
205 (for SMP CONFIG users)
206 For making accounting scalable, RSS related information are handled in an
207 asynchronous manner and the value may not be very precise. To see a precise
208 snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table.
209 It's slow but very precise.
211 Table 1-2: Contents of the status files (as of 4.1)
212 ..............................................................................
214 Name filename of the executable
215 State state (R is running, S is sleeping, D is sleeping
216 in an uninterruptible wait, Z is zombie,
217 T is traced or stopped)
219 Ngid NUMA group ID (0 if none)
221 PPid process id of the parent process
222 TracerPid PID of process tracing this process (0 if not)
223 Uid Real, effective, saved set, and file system UIDs
224 Gid Real, effective, saved set, and file system GIDs
225 FDSize number of file descriptor slots currently allocated
226 Groups supplementary group list
227 NStgid descendant namespace thread group ID hierarchy
228 NSpid descendant namespace process ID hierarchy
229 NSpgid descendant namespace process group ID hierarchy
230 NSsid descendant namespace session ID hierarchy
231 VmPeak peak virtual memory size
232 VmSize total program size
233 VmLck locked memory size
234 VmHWM peak resident set size ("high water mark")
235 VmRSS size of memory portions
236 VmData size of data, stack, and text segments
237 VmStk size of data, stack, and text segments
238 VmExe size of text segment
239 VmLib size of shared library code
240 VmPTE size of page table entries
241 VmPMD size of second level page tables
242 VmSwap size of swap usage (the number of referred swapents)
243 HugetlbPages size of hugetlb memory portions
244 Threads number of threads
245 SigQ number of signals queued/max. number for queue
246 SigPnd bitmap of pending signals for the thread
247 ShdPnd bitmap of shared pending signals for the process
248 SigBlk bitmap of blocked signals
249 SigIgn bitmap of ignored signals
250 SigCgt bitmap of caught signals
251 CapInh bitmap of inheritable capabilities
252 CapPrm bitmap of permitted capabilities
253 CapEff bitmap of effective capabilities
254 CapBnd bitmap of capabilities bounding set
255 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...)
256 Cpus_allowed mask of CPUs on which this process may run
257 Cpus_allowed_list Same as previous, but in "list format"
258 Mems_allowed mask of memory nodes allowed to this process
259 Mems_allowed_list Same as previous, but in "list format"
260 voluntary_ctxt_switches number of voluntary context switches
261 nonvoluntary_ctxt_switches number of non voluntary context switches
262 ..............................................................................
264 Table 1-3: Contents of the statm files (as of 2.6.8-rc3)
265 ..............................................................................
267 size total program size (pages) (same as VmSize in status)
268 resident size of memory portions (pages) (same as VmRSS in status)
269 shared number of pages that are shared (i.e. backed by a file)
270 trs number of pages that are 'code' (not including libs; broken,
271 includes data segment)
272 lrs number of pages of library (always 0 on 2.6)
273 drs number of pages of data/stack (including libs; broken,
274 includes library text)
275 dt number of dirty pages (always 0 on 2.6)
276 ..............................................................................
279 Table 1-4: Contents of the stat files (as of 2.6.30-rc7)
280 ..............................................................................
283 tcomm filename of the executable
284 state state (R is running, S is sleeping, D is sleeping in an
285 uninterruptible wait, Z is zombie, T is traced or stopped)
286 ppid process id of the parent process
287 pgrp pgrp of the process
289 tty_nr tty the process uses
290 tty_pgrp pgrp of the tty
292 min_flt number of minor faults
293 cmin_flt number of minor faults with child's
294 maj_flt number of major faults
295 cmaj_flt number of major faults with child's
296 utime user mode jiffies
297 stime kernel mode jiffies
298 cutime user mode jiffies with child's
299 cstime kernel mode jiffies with child's
300 priority priority level
302 num_threads number of threads
303 it_real_value (obsolete, always 0)
304 start_time time the process started after system boot
305 vsize virtual memory size
306 rss resident set memory size
307 rsslim current limit in bytes on the rss
308 start_code address above which program text can run
309 end_code address below which program text can run
310 start_stack address of the start of the main process stack
311 esp current value of ESP
312 eip current value of EIP
313 pending bitmap of pending signals
314 blocked bitmap of blocked signals
315 sigign bitmap of ignored signals
316 sigcatch bitmap of caught signals
317 0 (place holder, used to be the wchan address, use /proc/PID/wchan instead)
320 exit_signal signal to send to parent thread on exit
321 task_cpu which CPU the task is scheduled on
322 rt_priority realtime priority
323 policy scheduling policy (man sched_setscheduler)
324 blkio_ticks time spent waiting for block IO
325 gtime guest time of the task in jiffies
326 cgtime guest time of the task children in jiffies
327 start_data address above which program data+bss is placed
328 end_data address below which program data+bss is placed
329 start_brk address above which program heap can be expanded with brk()
330 arg_start address above which program command line is placed
331 arg_end address below which program command line is placed
332 env_start address above which program environment is placed
333 env_end address below which program environment is placed
334 exit_code the thread's exit_code in the form reported by the waitpid system call
335 ..............................................................................
337 The /proc/PID/maps file containing the currently mapped memory regions and
338 their access permissions.
342 address perms offset dev inode pathname
344 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
345 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
346 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
347 a7cb1000-a7cb2000 ---p 00000000 00:00 0
348 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
349 a7eb2000-a7eb3000 ---p 00000000 00:00 0
350 a7eb3000-a7ed5000 rw-p 00000000 00:00 0
351 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
352 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
353 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
354 a800b000-a800e000 rw-p 00000000 00:00 0
355 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
356 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
357 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
358 a8024000-a8027000 rw-p 00000000 00:00 0
359 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
360 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
361 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
362 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack]
363 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
365 where "address" is the address space in the process that it occupies, "perms"
366 is a set of permissions:
372 p = private (copy on write)
374 "offset" is the offset into the mapping, "dev" is the device (major:minor), and
375 "inode" is the inode on that device. 0 indicates that no inode is associated
376 with the memory region, as the case would be with BSS (uninitialized data).
377 The "pathname" shows the name associated file for this mapping. If the mapping
378 is not associated with a file:
380 [heap] = the heap of the program
381 [stack] = the stack of the main process
382 [vdso] = the "virtual dynamic shared object",
383 the kernel system call handler
384 [anon:<name>] = an anonymous mapping that has been
387 or if empty, the mapping is anonymous.
389 The /proc/PID/smaps is an extension based on maps, showing the memory
390 consumption for each of the process's mappings. For each of mappings there
391 is a series of lines such as the following:
393 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash
405 Private_Hugetlb: 0 kB
411 VmFlags: rd ex mr mw me dw
412 Name: name from userspace
414 the first of these lines shows the same information as is displayed for the
415 mapping in /proc/PID/maps. The remaining lines show the size of the mapping
416 (size), the amount of the mapping that is currently resident in RAM (RSS), the
417 process' proportional share of this mapping (PSS), the number of clean and
418 dirty private pages in the mapping.
420 The "proportional set size" (PSS) of a process is the count of pages it has
421 in memory, where each page is divided by the number of processes sharing it.
422 So if a process has 1000 pages all to itself, and 1000 shared with one other
423 process, its PSS will be 1500.
424 Note that even a page which is part of a MAP_SHARED mapping, but has only
425 a single pte mapped, i.e. is currently used by only one process, is accounted
426 as private and not as shared.
427 "Referenced" indicates the amount of memory currently marked as referenced or
429 "Anonymous" shows the amount of memory that does not belong to any file. Even
430 a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE
431 and a page is modified, the file page is replaced by a private anonymous copy.
432 "AnonHugePages" shows the ammount of memory backed by transparent hugepage.
433 "Shared_Hugetlb" and "Private_Hugetlb" show the ammounts of memory backed by
434 hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical
435 reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field.
436 "Swap" shows how much would-be-anonymous memory is also used, but out on swap.
437 "SwapPss" shows proportional swap share of this mapping.
438 "Locked" indicates whether the mapping is locked in memory or not.
440 "VmFlags" field deserves a separate description. This member represents the kernel
441 flags associated with the particular virtual memory area in two letter encoded
442 manner. The codes are the following:
451 gd - stack segment growns down
453 dw - disabled write to the mapped file
454 lo - pages are locked in memory
455 io - memory mapped I/O area
456 sr - sequential read advise provided
457 rr - random read advise provided
458 dc - do not copy area on fork
459 de - do not expand area on remapping
460 ac - area is accountable
461 nr - swap space is not reserved for the area
462 ht - area uses huge tlb pages
463 ar - architecture specific flag
464 dd - do not include area into core dump
467 hg - huge page advise flag
468 nh - no-huge page advise flag
469 mg - mergable advise flag
471 Note that there is no guarantee that every flag and associated mnemonic will
472 be present in all further kernel releases. Things get changed, the flags may
473 be vanished or the reverse -- new added.
475 The "Name" field will only be present on a mapping that has been named by
476 userspace, and will show the name passed in by userspace.
478 This file is only present if the CONFIG_MMU kernel configuration option is
481 The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG
482 bits on both physical and virtual pages associated with a process, and the
483 soft-dirty bit on pte (see Documentation/vm/soft-dirty.txt for details).
484 To clear the bits for all the pages associated with the process
485 > echo 1 > /proc/PID/clear_refs
487 To clear the bits for the anonymous pages associated with the process
488 > echo 2 > /proc/PID/clear_refs
490 To clear the bits for the file mapped pages associated with the process
491 > echo 3 > /proc/PID/clear_refs
493 To clear the soft-dirty bit
494 > echo 4 > /proc/PID/clear_refs
496 To reset the peak resident set size ("high water mark") to the process's
498 > echo 5 > /proc/PID/clear_refs
500 Any other value written to /proc/PID/clear_refs will have no effect.
502 The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags
503 using /proc/kpageflags and number of times a page is mapped using
504 /proc/kpagecount. For detailed explanation, see Documentation/vm/pagemap.txt.
506 The /proc/pid/numa_maps is an extension based on maps, showing the memory
507 locality and binding policy, as well as the memory usage (in pages) of
508 each mapping. The output follows a general format where mapping details get
509 summarized separated by blank spaces, one mapping per each file line:
511 address policy mapping details
513 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4
514 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4
515 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4
516 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
517 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
518 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4
519 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4
520 320698b000 default file=/lib64/libc-2.12.so
521 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4
522 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
523 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4
524 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4
525 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4
526 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048
527 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4
528 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4
531 "address" is the starting address for the mapping;
532 "policy" reports the NUMA memory policy set for the mapping (see vm/numa_memory_policy.txt);
533 "mapping details" summarizes mapping data such as mapping type, page usage counters,
534 node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page
535 size, in KB, that is backing the mapping up.
540 Similar to the process entries, the kernel data files give information about
541 the running kernel. The files used to obtain this information are contained in
542 /proc and are listed in Table 1-5. Not all of these will be present in your
543 system. It depends on the kernel configuration and the loaded modules, which
544 files are there, and which are missing.
546 Table 1-5: Kernel info in /proc
547 ..............................................................................
549 apm Advanced power management info
550 buddyinfo Kernel memory allocator information (see text) (2.5)
551 bus Directory containing bus specific information
552 cmdline Kernel command line
553 cpuinfo Info about the CPU
554 devices Available devices (block and character)
555 dma Used DMS channels
556 filesystems Supported filesystems
557 driver Various drivers grouped here, currently rtc (2.4)
558 execdomains Execdomains, related to security (2.4)
559 fb Frame Buffer devices (2.4)
560 fs File system parameters, currently nfs/exports (2.4)
561 ide Directory containing info about the IDE subsystem
562 interrupts Interrupt usage
563 iomem Memory map (2.4)
564 ioports I/O port usage
565 irq Masks for irq to cpu affinity (2.4)(smp?)
566 isapnp ISA PnP (Plug&Play) Info (2.4)
567 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
569 ksyms Kernel symbol table
570 loadavg Load average of last 1, 5 & 15 minutes
574 modules List of loaded modules
575 mounts Mounted filesystems
576 net Networking info (see text)
577 pagetypeinfo Additional page allocator information (see text) (2.5)
578 partitions Table of partitions known to the system
579 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
580 decoupled by lspci (2.4)
582 scsi SCSI info (see text)
583 slabinfo Slab pool info
584 softirqs softirq usage
585 stat Overall statistics
586 swaps Swap space utilization
588 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
589 tty Info of tty drivers
590 uptime Wall clock since boot, combined idle time of all cpus
591 version Kernel version
592 video bttv info of video resources (2.4)
593 vmallocinfo Show vmalloced areas
594 ..............................................................................
596 You can, for example, check which interrupts are currently in use and what
597 they are used for by looking in the file /proc/interrupts:
599 > cat /proc/interrupts
601 0: 8728810 XT-PIC timer
602 1: 895 XT-PIC keyboard
604 3: 531695 XT-PIC aha152x
605 4: 2014133 XT-PIC serial
606 5: 44401 XT-PIC pcnet_cs
609 12: 182918 XT-PIC PS/2 Mouse
611 14: 1232265 XT-PIC ide0
615 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
616 output of a SMP machine):
618 > cat /proc/interrupts
621 0: 1243498 1214548 IO-APIC-edge timer
622 1: 8949 8958 IO-APIC-edge keyboard
623 2: 0 0 XT-PIC cascade
624 5: 11286 10161 IO-APIC-edge soundblaster
625 8: 1 0 IO-APIC-edge rtc
626 9: 27422 27407 IO-APIC-edge 3c503
627 12: 113645 113873 IO-APIC-edge PS/2 Mouse
629 14: 22491 24012 IO-APIC-edge ide0
630 15: 2183 2415 IO-APIC-edge ide1
631 17: 30564 30414 IO-APIC-level eth0
632 18: 177 164 IO-APIC-level bttv
637 NMI is incremented in this case because every timer interrupt generates a NMI
638 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
640 LOC is the local interrupt counter of the internal APIC of every CPU.
642 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
643 connects the CPUs in a SMP system. This means that an error has been detected,
644 the IO-APIC automatically retry the transmission, so it should not be a big
645 problem, but you should read the SMP-FAQ.
647 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
648 /proc/interrupts to display every IRQ vector in use by the system, not
649 just those considered 'most important'. The new vectors are:
651 THR -- interrupt raised when a machine check threshold counter
652 (typically counting ECC corrected errors of memory or cache) exceeds
653 a configurable threshold. Only available on some systems.
655 TRM -- a thermal event interrupt occurs when a temperature threshold
656 has been exceeded for the CPU. This interrupt may also be generated
657 when the temperature drops back to normal.
659 SPU -- a spurious interrupt is some interrupt that was raised then lowered
660 by some IO device before it could be fully processed by the APIC. Hence
661 the APIC sees the interrupt but does not know what device it came from.
662 For this case the APIC will generate the interrupt with a IRQ vector
663 of 0xff. This might also be generated by chipset bugs.
665 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
666 sent from one CPU to another per the needs of the OS. Typically,
667 their statistics are used by kernel developers and interested users to
668 determine the occurrence of interrupts of the given type.
670 The above IRQ vectors are displayed only when relevant. For example,
671 the threshold vector does not exist on x86_64 platforms. Others are
672 suppressed when the system is a uniprocessor. As of this writing, only
673 i386 and x86_64 platforms support the new IRQ vector displays.
675 Of some interest is the introduction of the /proc/irq directory to 2.4.
676 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
677 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
678 irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and
683 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
684 1 11 13 15 17 19 3 5 7 9 default_smp_affinity
688 smp_affinity is a bitmask, in which you can specify which CPUs can handle the
689 IRQ, you can set it by doing:
691 > echo 1 > /proc/irq/10/smp_affinity
693 This means that only the first CPU will handle the IRQ, but you can also echo
694 5 which means that only the first and fourth CPU can handle the IRQ.
696 The contents of each smp_affinity file is the same by default:
698 > cat /proc/irq/0/smp_affinity
701 There is an alternate interface, smp_affinity_list which allows specifying
702 a cpu range instead of a bitmask:
704 > cat /proc/irq/0/smp_affinity_list
707 The default_smp_affinity mask applies to all non-active IRQs, which are the
708 IRQs which have not yet been allocated/activated, and hence which lack a
709 /proc/irq/[0-9]* directory.
711 The node file on an SMP system shows the node to which the device using the IRQ
712 reports itself as being attached. This hardware locality information does not
713 include information about any possible driver locality preference.
715 prof_cpu_mask specifies which CPUs are to be profiled by the system wide
716 profiler. Default value is ffffffff (all cpus if there are only 32 of them).
718 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
719 between all the CPUs which are allowed to handle it. As usual the kernel has
720 more info than you and does a better job than you, so the defaults are the
721 best choice for almost everyone. [Note this applies only to those IO-APIC's
722 that support "Round Robin" interrupt distribution.]
724 There are three more important subdirectories in /proc: net, scsi, and sys.
725 The general rule is that the contents, or even the existence of these
726 directories, depend on your kernel configuration. If SCSI is not enabled, the
727 directory scsi may not exist. The same is true with the net, which is there
728 only when networking support is present in the running kernel.
730 The slabinfo file gives information about memory usage at the slab level.
731 Linux uses slab pools for memory management above page level in version 2.2.
732 Commonly used objects have their own slab pool (such as network buffers,
733 directory cache, and so on).
735 ..............................................................................
737 > cat /proc/buddyinfo
739 Node 0, zone DMA 0 4 5 4 4 3 ...
740 Node 0, zone Normal 1 0 0 1 101 8 ...
741 Node 0, zone HighMem 2 0 0 1 1 0 ...
743 External fragmentation is a problem under some workloads, and buddyinfo is a
744 useful tool for helping diagnose these problems. Buddyinfo will give you a
745 clue as to how big an area you can safely allocate, or why a previous
748 Each column represents the number of pages of a certain order which are
749 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
750 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
751 available in ZONE_NORMAL, etc...
753 More information relevant to external fragmentation can be found in
756 > cat /proc/pagetypeinfo
760 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
761 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0
762 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0
763 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2
764 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0
765 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0
766 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9
767 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0
768 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452
769 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0
770 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0
772 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate
773 Node 0, zone DMA 2 0 5 1 0
774 Node 0, zone DMA32 41 6 967 2 0
776 Fragmentation avoidance in the kernel works by grouping pages of different
777 migrate types into the same contiguous regions of memory called page blocks.
778 A page block is typically the size of the default hugepage size e.g. 2MB on
779 X86-64. By keeping pages grouped based on their ability to move, the kernel
780 can reclaim pages within a page block to satisfy a high-order allocation.
782 The pagetypinfo begins with information on the size of a page block. It
783 then gives the same type of information as buddyinfo except broken down
784 by migrate-type and finishes with details on how many page blocks of each
787 If min_free_kbytes has been tuned correctly (recommendations made by hugeadm
788 from libhugetlbfs http://sourceforge.net/projects/libhugetlbfs/), one can
789 make an estimate of the likely number of huge pages that can be allocated
790 at a given point in time. All the "Movable" blocks should be allocatable
791 unless memory has been mlock()'d. Some of the Reclaimable blocks should
792 also be allocatable although a lot of filesystem metadata may have to be
793 reclaimed to achieve this.
795 ..............................................................................
799 Provides information about distribution and utilization of memory. This
800 varies by architecture and compile options. The following is from a
801 16GB PIII, which has highmem enabled. You may not have all of these fields.
805 MemTotal: 16344972 kB
807 MemAvailable: 14836172 kB
813 HighTotal: 15597528 kB
814 HighFree: 13629632 kB
824 SReclaimable: 159856 kB
825 SUnreclaim: 124508 kB
830 CommitLimit: 7669796 kB
831 Committed_AS: 100056 kB
832 VmallocTotal: 112216 kB
834 VmallocChunk: 111088 kB
835 AnonHugePages: 49152 kB
837 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
838 bits and the kernel binary code)
839 MemFree: The sum of LowFree+HighFree
840 MemAvailable: An estimate of how much memory is available for starting new
841 applications, without swapping. Calculated from MemFree,
842 SReclaimable, the size of the file LRU lists, and the low
843 watermarks in each zone.
844 The estimate takes into account that the system needs some
845 page cache to function well, and that not all reclaimable
846 slab will be reclaimable, due to items being in use. The
847 impact of those factors will vary from system to system.
848 Buffers: Relatively temporary storage for raw disk blocks
849 shouldn't get tremendously large (20MB or so)
850 Cached: in-memory cache for files read from the disk (the
851 pagecache). Doesn't include SwapCached
852 SwapCached: Memory that once was swapped out, is swapped back in but
853 still also is in the swapfile (if memory is needed it
854 doesn't need to be swapped out AGAIN because it is already
855 in the swapfile. This saves I/O)
856 Active: Memory that has been used more recently and usually not
857 reclaimed unless absolutely necessary.
858 Inactive: Memory which has been less recently used. It is more
859 eligible to be reclaimed for other purposes
861 HighFree: Highmem is all memory above ~860MB of physical memory
862 Highmem areas are for use by userspace programs, or
863 for the pagecache. The kernel must use tricks to access
864 this memory, making it slower to access than lowmem.
866 LowFree: Lowmem is memory which can be used for everything that
867 highmem can be used for, but it is also available for the
868 kernel's use for its own data structures. Among many
869 other things, it is where everything from the Slab is
870 allocated. Bad things happen when you're out of lowmem.
871 SwapTotal: total amount of swap space available
872 SwapFree: Memory which has been evicted from RAM, and is temporarily
874 Dirty: Memory which is waiting to get written back to the disk
875 Writeback: Memory which is actively being written back to the disk
876 AnonPages: Non-file backed pages mapped into userspace page tables
877 AnonHugePages: Non-file backed huge pages mapped into userspace page tables
878 Mapped: files which have been mmaped, such as libraries
879 Slab: in-kernel data structures cache
880 SReclaimable: Part of Slab, that might be reclaimed, such as caches
881 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure
882 PageTables: amount of memory dedicated to the lowest level of page
884 NFS_Unstable: NFS pages sent to the server, but not yet committed to stable
886 Bounce: Memory used for block device "bounce buffers"
887 WritebackTmp: Memory used by FUSE for temporary writeback buffers
888 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
889 this is the total amount of memory currently available to
890 be allocated on the system. This limit is only adhered to
891 if strict overcommit accounting is enabled (mode 2 in
892 'vm.overcommit_memory').
893 The CommitLimit is calculated with the following formula:
894 CommitLimit = ([total RAM pages] - [total huge TLB pages]) *
895 overcommit_ratio / 100 + [total swap pages]
896 For example, on a system with 1G of physical RAM and 7G
897 of swap with a `vm.overcommit_ratio` of 30 it would
898 yield a CommitLimit of 7.3G.
899 For more details, see the memory overcommit documentation
900 in vm/overcommit-accounting.
901 Committed_AS: The amount of memory presently allocated on the system.
902 The committed memory is a sum of all of the memory which
903 has been allocated by processes, even if it has not been
904 "used" by them as of yet. A process which malloc()'s 1G
905 of memory, but only touches 300M of it will show up as
906 using 1G. This 1G is memory which has been "committed" to
907 by the VM and can be used at any time by the allocating
908 application. With strict overcommit enabled on the system
909 (mode 2 in 'vm.overcommit_memory'),allocations which would
910 exceed the CommitLimit (detailed above) will not be permitted.
911 This is useful if one needs to guarantee that processes will
912 not fail due to lack of memory once that memory has been
913 successfully allocated.
914 VmallocTotal: total size of vmalloc memory area
915 VmallocUsed: amount of vmalloc area which is used
916 VmallocChunk: largest contiguous block of vmalloc area which is free
918 ..............................................................................
922 Provides information about vmalloced/vmaped areas. One line per area,
923 containing the virtual address range of the area, size in bytes,
924 caller information of the creator, and optional information depending
925 on the kind of area :
927 pages=nr number of pages
928 phys=addr if a physical address was specified
929 ioremap I/O mapping (ioremap() and friends)
930 vmalloc vmalloc() area
933 vpages buffer for pages pointers was vmalloced (huge area)
934 N<node>=nr (Only on NUMA kernels)
935 Number of pages allocated on memory node <node>
937 > cat /proc/vmallocinfo
938 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ...
939 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128
940 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ...
941 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64
942 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f...
943 phys=7fee8000 ioremap
944 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f...
945 phys=7fee7000 ioremap
946 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210
947 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ...
948 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3
949 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ...
951 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ...
952 /0x130 [x_tables] pages=4 vmalloc N0=4
953 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ...
954 pages=14 vmalloc N2=14
955 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ...
957 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ...
959 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ...
960 pages=10 vmalloc N0=10
962 ..............................................................................
966 Provides counts of softirq handlers serviced since boot time, for each cpu.
971 TIMER: 27166 27120 27097 27034
976 SCHED: 27035 26983 26971 26746
978 RCU: 1678 1769 2178 2250
981 1.3 IDE devices in /proc/ide
982 ----------------------------
984 The subdirectory /proc/ide contains information about all IDE devices of which
985 the kernel is aware. There is one subdirectory for each IDE controller, the
986 file drivers and a link for each IDE device, pointing to the device directory
987 in the controller specific subtree.
989 The file drivers contains general information about the drivers used for the
992 > cat /proc/ide/drivers
993 ide-cdrom version 4.53
994 ide-disk version 1.08
996 More detailed information can be found in the controller specific
997 subdirectories. These are named ide0, ide1 and so on. Each of these
998 directories contains the files shown in table 1-6.
1001 Table 1-6: IDE controller info in /proc/ide/ide?
1002 ..............................................................................
1004 channel IDE channel (0 or 1)
1005 config Configuration (only for PCI/IDE bridge)
1007 model Type/Chipset of IDE controller
1008 ..............................................................................
1010 Each device connected to a controller has a separate subdirectory in the
1011 controllers directory. The files listed in table 1-7 are contained in these
1015 Table 1-7: IDE device information
1016 ..............................................................................
1019 capacity Capacity of the medium (in 512Byte blocks)
1020 driver driver and version
1021 geometry physical and logical geometry
1022 identify device identify block
1024 model device identifier
1025 settings device setup
1026 smart_thresholds IDE disk management thresholds
1027 smart_values IDE disk management values
1028 ..............................................................................
1030 The most interesting file is settings. This file contains a nice overview of
1031 the drive parameters:
1033 # cat /proc/ide/ide0/hda/settings
1034 name value min max mode
1035 ---- ----- --- --- ----
1036 bios_cyl 526 0 65535 rw
1037 bios_head 255 0 255 rw
1038 bios_sect 63 0 63 rw
1039 breada_readahead 4 0 127 rw
1041 file_readahead 72 0 2097151 rw
1043 keepsettings 0 0 1 rw
1044 max_kb_per_request 122 1 127 rw
1048 pio_mode write-only 0 255 w
1054 1.4 Networking info in /proc/net
1055 --------------------------------
1057 The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the
1058 additional values you get for IP version 6 if you configure the kernel to
1059 support this. Table 1-9 lists the files and their meaning.
1062 Table 1-8: IPv6 info in /proc/net
1063 ..............................................................................
1065 udp6 UDP sockets (IPv6)
1066 tcp6 TCP sockets (IPv6)
1067 raw6 Raw device statistics (IPv6)
1068 igmp6 IP multicast addresses, which this host joined (IPv6)
1069 if_inet6 List of IPv6 interface addresses
1070 ipv6_route Kernel routing table for IPv6
1071 rt6_stats Global IPv6 routing tables statistics
1072 sockstat6 Socket statistics (IPv6)
1073 snmp6 Snmp data (IPv6)
1074 ..............................................................................
1077 Table 1-9: Network info in /proc/net
1078 ..............................................................................
1080 arp Kernel ARP table
1081 dev network devices with statistics
1082 dev_mcast the Layer2 multicast groups a device is listening too
1083 (interface index, label, number of references, number of bound
1085 dev_stat network device status
1086 ip_fwchains Firewall chain linkage
1087 ip_fwnames Firewall chain names
1088 ip_masq Directory containing the masquerading tables
1089 ip_masquerade Major masquerading table
1090 netstat Network statistics
1091 raw raw device statistics
1092 route Kernel routing table
1093 rpc Directory containing rpc info
1094 rt_cache Routing cache
1096 sockstat Socket statistics
1099 unix UNIX domain sockets
1100 wireless Wireless interface data (Wavelan etc)
1101 igmp IP multicast addresses, which this host joined
1102 psched Global packet scheduler parameters.
1103 netlink List of PF_NETLINK sockets
1104 ip_mr_vifs List of multicast virtual interfaces
1105 ip_mr_cache List of multicast routing cache
1106 ..............................................................................
1108 You can use this information to see which network devices are available in
1109 your system and how much traffic was routed over those devices:
1112 Inter-|Receive |[...
1113 face |bytes packets errs drop fifo frame compressed multicast|[...
1114 lo: 908188 5596 0 0 0 0 0 0 [...
1115 ppp0:15475140 20721 410 0 0 410 0 0 [...
1116 eth0: 614530 7085 0 0 0 0 0 1 [...
1119 ...] bytes packets errs drop fifo colls carrier compressed
1120 ...] 908188 5596 0 0 0 0 0 0
1121 ...] 1375103 17405 0 0 0 0 0 0
1122 ...] 1703981 5535 0 0 0 3 0 0
1124 In addition, each Channel Bond interface has its own directory. For
1125 example, the bond0 device will have a directory called /proc/net/bond0/.
1126 It will contain information that is specific to that bond, such as the
1127 current slaves of the bond, the link status of the slaves, and how
1128 many times the slaves link has failed.
1133 If you have a SCSI host adapter in your system, you'll find a subdirectory
1134 named after the driver for this adapter in /proc/scsi. You'll also see a list
1135 of all recognized SCSI devices in /proc/scsi:
1137 >cat /proc/scsi/scsi
1139 Host: scsi0 Channel: 00 Id: 00 Lun: 00
1140 Vendor: IBM Model: DGHS09U Rev: 03E0
1141 Type: Direct-Access ANSI SCSI revision: 03
1142 Host: scsi0 Channel: 00 Id: 06 Lun: 00
1143 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
1144 Type: CD-ROM ANSI SCSI revision: 02
1147 The directory named after the driver has one file for each adapter found in
1148 the system. These files contain information about the controller, including
1149 the used IRQ and the IO address range. The amount of information shown is
1150 dependent on the adapter you use. The example shows the output for an Adaptec
1151 AHA-2940 SCSI adapter:
1153 > cat /proc/scsi/aic7xxx/0
1155 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
1157 TCQ Enabled By Default : Disabled
1158 AIC7XXX_PROC_STATS : Disabled
1159 AIC7XXX_RESET_DELAY : 5
1160 Adapter Configuration:
1161 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
1162 Ultra Wide Controller
1163 PCI MMAPed I/O Base: 0xeb001000
1164 Adapter SEEPROM Config: SEEPROM found and used.
1165 Adaptec SCSI BIOS: Enabled
1167 SCBs: Active 0, Max Active 2,
1168 Allocated 15, HW 16, Page 255
1170 BIOS Control Word: 0x18b6
1171 Adapter Control Word: 0x005b
1172 Extended Translation: Enabled
1173 Disconnect Enable Flags: 0xffff
1174 Ultra Enable Flags: 0x0001
1175 Tag Queue Enable Flags: 0x0000
1176 Ordered Queue Tag Flags: 0x0000
1177 Default Tag Queue Depth: 8
1178 Tagged Queue By Device array for aic7xxx host instance 0:
1179 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
1180 Actual queue depth per device for aic7xxx host instance 0:
1181 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
1184 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
1185 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
1186 Total transfers 160151 (74577 reads and 85574 writes)
1188 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
1189 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
1190 Total transfers 0 (0 reads and 0 writes)
1193 1.6 Parallel port info in /proc/parport
1194 ---------------------------------------
1196 The directory /proc/parport contains information about the parallel ports of
1197 your system. It has one subdirectory for each port, named after the port
1200 These directories contain the four files shown in Table 1-10.
1203 Table 1-10: Files in /proc/parport
1204 ..............................................................................
1206 autoprobe Any IEEE-1284 device ID information that has been acquired.
1207 devices list of the device drivers using that port. A + will appear by the
1208 name of the device currently using the port (it might not appear
1210 hardware Parallel port's base address, IRQ line and DMA channel.
1211 irq IRQ that parport is using for that port. This is in a separate
1212 file to allow you to alter it by writing a new value in (IRQ
1214 ..............................................................................
1216 1.7 TTY info in /proc/tty
1217 -------------------------
1219 Information about the available and actually used tty's can be found in the
1220 directory /proc/tty.You'll find entries for drivers and line disciplines in
1221 this directory, as shown in Table 1-11.
1224 Table 1-11: Files in /proc/tty
1225 ..............................................................................
1227 drivers list of drivers and their usage
1228 ldiscs registered line disciplines
1229 driver/serial usage statistic and status of single tty lines
1230 ..............................................................................
1232 To see which tty's are currently in use, you can simply look into the file
1235 > cat /proc/tty/drivers
1236 pty_slave /dev/pts 136 0-255 pty:slave
1237 pty_master /dev/ptm 128 0-255 pty:master
1238 pty_slave /dev/ttyp 3 0-255 pty:slave
1239 pty_master /dev/pty 2 0-255 pty:master
1240 serial /dev/cua 5 64-67 serial:callout
1241 serial /dev/ttyS 4 64-67 serial
1242 /dev/tty0 /dev/tty0 4 0 system:vtmaster
1243 /dev/ptmx /dev/ptmx 5 2 system
1244 /dev/console /dev/console 5 1 system:console
1245 /dev/tty /dev/tty 5 0 system:/dev/tty
1246 unknown /dev/tty 4 1-63 console
1249 1.8 Miscellaneous kernel statistics in /proc/stat
1250 -------------------------------------------------
1252 Various pieces of information about kernel activity are available in the
1253 /proc/stat file. All of the numbers reported in this file are aggregates
1254 since the system first booted. For a quick look, simply cat the file:
1257 cpu 2255 34 2290 22625563 6290 127 456 0 0 0
1258 cpu0 1132 34 1441 11311718 3675 127 438 0 0 0
1259 cpu1 1123 0 849 11313845 2614 0 18 0 0 0
1260 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
1266 softirq 183433 0 21755 12 39 1137 231 21459 2263
1268 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
1269 lines. These numbers identify the amount of time the CPU has spent performing
1270 different kinds of work. Time units are in USER_HZ (typically hundredths of a
1271 second). The meanings of the columns are as follows, from left to right:
1273 - user: normal processes executing in user mode
1274 - nice: niced processes executing in user mode
1275 - system: processes executing in kernel mode
1276 - idle: twiddling thumbs
1277 - iowait: waiting for I/O to complete
1278 - irq: servicing interrupts
1279 - softirq: servicing softirqs
1280 - steal: involuntary wait
1281 - guest: running a normal guest
1282 - guest_nice: running a niced guest
1284 The "intr" line gives counts of interrupts serviced since boot time, for each
1285 of the possible system interrupts. The first column is the total of all
1286 interrupts serviced including unnumbered architecture specific interrupts;
1287 each subsequent column is the total for that particular numbered interrupt.
1288 Unnumbered interrupts are not shown, only summed into the total.
1290 The "ctxt" line gives the total number of context switches across all CPUs.
1292 The "btime" line gives the time at which the system booted, in seconds since
1295 The "processes" line gives the number of processes and threads created, which
1296 includes (but is not limited to) those created by calls to the fork() and
1297 clone() system calls.
1299 The "procs_running" line gives the total number of threads that are
1300 running or ready to run (i.e., the total number of runnable threads).
1302 The "procs_blocked" line gives the number of processes currently blocked,
1303 waiting for I/O to complete.
1305 The "softirq" line gives counts of softirqs serviced since boot time, for each
1306 of the possible system softirqs. The first column is the total of all
1307 softirqs serviced; each subsequent column is the total for that particular
1311 1.9 Ext4 file system parameters
1312 -------------------------------
1314 Information about mounted ext4 file systems can be found in
1315 /proc/fs/ext4. Each mounted filesystem will have a directory in
1316 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
1317 /proc/fs/ext4/dm-0). The files in each per-device directory are shown
1318 in Table 1-12, below.
1320 Table 1-12: Files in /proc/fs/ext4/<devname>
1321 ..............................................................................
1323 mb_groups details of multiblock allocator buddy cache of free blocks
1324 ..............................................................................
1328 Shows registered system console lines.
1330 To see which character device lines are currently used for the system console
1331 /dev/console, you may simply look into the file /proc/consoles:
1333 > cat /proc/consoles
1339 device name of the device
1340 operations R = can do read operations
1341 W = can do write operations
1343 flags E = it is enabled
1344 C = it is preferred console
1345 B = it is primary boot console
1346 p = it is used for printk buffer
1347 b = it is not a TTY but a Braille device
1348 a = it is safe to use when cpu is offline
1349 major:minor major and minor number of the device separated by a colon
1351 ------------------------------------------------------------------------------
1353 ------------------------------------------------------------------------------
1354 The /proc file system serves information about the running system. It not only
1355 allows access to process data but also allows you to request the kernel status
1356 by reading files in the hierarchy.
1358 The directory structure of /proc reflects the types of information and makes
1359 it easy, if not obvious, where to look for specific data.
1360 ------------------------------------------------------------------------------
1362 ------------------------------------------------------------------------------
1363 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
1364 ------------------------------------------------------------------------------
1366 ------------------------------------------------------------------------------
1368 ------------------------------------------------------------------------------
1369 * Modifying kernel parameters by writing into files found in /proc/sys
1370 * Exploring the files which modify certain parameters
1371 * Review of the /proc/sys file tree
1372 ------------------------------------------------------------------------------
1375 A very interesting part of /proc is the directory /proc/sys. This is not only
1376 a source of information, it also allows you to change parameters within the
1377 kernel. Be very careful when attempting this. You can optimize your system,
1378 but you can also cause it to crash. Never alter kernel parameters on a
1379 production system. Set up a development machine and test to make sure that
1380 everything works the way you want it to. You may have no alternative but to
1381 reboot the machine once an error has been made.
1383 To change a value, simply echo the new value into the file. An example is
1384 given below in the section on the file system data. You need to be root to do
1385 this. You can create your own boot script to perform this every time your
1388 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
1389 general things in the operation of the Linux kernel. Since some of the files
1390 can inadvertently disrupt your system, it is advisable to read both
1391 documentation and source before actually making adjustments. In any case, be
1392 very careful when writing to any of these files. The entries in /proc may
1393 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
1394 review the kernel documentation in the directory /usr/src/linux/Documentation.
1395 This chapter is heavily based on the documentation included in the pre 2.2
1396 kernels, and became part of it in version 2.2.1 of the Linux kernel.
1398 Please see: Documentation/sysctl/ directory for descriptions of these
1401 ------------------------------------------------------------------------------
1403 ------------------------------------------------------------------------------
1404 Certain aspects of kernel behavior can be modified at runtime, without the
1405 need to recompile the kernel, or even to reboot the system. The files in the
1406 /proc/sys tree can not only be read, but also modified. You can use the echo
1407 command to write value into these files, thereby changing the default settings
1409 ------------------------------------------------------------------------------
1411 ------------------------------------------------------------------------------
1412 CHAPTER 3: PER-PROCESS PARAMETERS
1413 ------------------------------------------------------------------------------
1415 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score
1416 --------------------------------------------------------------------------------
1418 These file can be used to adjust the badness heuristic used to select which
1419 process gets killed in out of memory conditions.
1421 The badness heuristic assigns a value to each candidate task ranging from 0
1422 (never kill) to 1000 (always kill) to determine which process is targeted. The
1423 units are roughly a proportion along that range of allowed memory the process
1424 may allocate from based on an estimation of its current memory and swap use.
1425 For example, if a task is using all allowed memory, its badness score will be
1426 1000. If it is using half of its allowed memory, its score will be 500.
1428 There is an additional factor included in the badness score: the current memory
1429 and swap usage is discounted by 3% for root processes.
1431 The amount of "allowed" memory depends on the context in which the oom killer
1432 was called. If it is due to the memory assigned to the allocating task's cpuset
1433 being exhausted, the allowed memory represents the set of mems assigned to that
1434 cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed
1435 memory represents the set of mempolicy nodes. If it is due to a memory
1436 limit (or swap limit) being reached, the allowed memory is that configured
1437 limit. Finally, if it is due to the entire system being out of memory, the
1438 allowed memory represents all allocatable resources.
1440 The value of /proc/<pid>/oom_score_adj is added to the badness score before it
1441 is used to determine which task to kill. Acceptable values range from -1000
1442 (OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to
1443 polarize the preference for oom killing either by always preferring a certain
1444 task or completely disabling it. The lowest possible value, -1000, is
1445 equivalent to disabling oom killing entirely for that task since it will always
1446 report a badness score of 0.
1448 Consequently, it is very simple for userspace to define the amount of memory to
1449 consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for
1450 example, is roughly equivalent to allowing the remainder of tasks sharing the
1451 same system, cpuset, mempolicy, or memory controller resources to use at least
1452 50% more memory. A value of -500, on the other hand, would be roughly
1453 equivalent to discounting 50% of the task's allowed memory from being considered
1454 as scoring against the task.
1456 For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also
1457 be used to tune the badness score. Its acceptable values range from -16
1458 (OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17
1459 (OOM_DISABLE) to disable oom killing entirely for that task. Its value is
1460 scaled linearly with /proc/<pid>/oom_score_adj.
1462 The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last
1463 value set by a CAP_SYS_RESOURCE process. To reduce the value any lower
1464 requires CAP_SYS_RESOURCE.
1466 Caveat: when a parent task is selected, the oom killer will sacrifice any first
1467 generation children with separate address spaces instead, if possible. This
1468 avoids servers and important system daemons from being killed and loses the
1469 minimal amount of work.
1472 3.2 /proc/<pid>/oom_score - Display current oom-killer score
1473 -------------------------------------------------------------
1475 This file can be used to check the current score used by the oom-killer is for
1476 any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which
1477 process should be killed in an out-of-memory situation.
1480 3.3 /proc/<pid>/io - Display the IO accounting fields
1481 -------------------------------------------------------
1483 This file contains IO statistics for each running process
1488 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
1491 test:/tmp # cat /proc/3828/io
1497 write_bytes: 323932160
1498 cancelled_write_bytes: 0
1507 I/O counter: chars read
1508 The number of bytes which this task has caused to be read from storage. This
1509 is simply the sum of bytes which this process passed to read() and pread().
1510 It includes things like tty IO and it is unaffected by whether or not actual
1511 physical disk IO was required (the read might have been satisfied from
1518 I/O counter: chars written
1519 The number of bytes which this task has caused, or shall cause to be written
1520 to disk. Similar caveats apply here as with rchar.
1526 I/O counter: read syscalls
1527 Attempt to count the number of read I/O operations, i.e. syscalls like read()
1534 I/O counter: write syscalls
1535 Attempt to count the number of write I/O operations, i.e. syscalls like
1536 write() and pwrite().
1542 I/O counter: bytes read
1543 Attempt to count the number of bytes which this process really did cause to
1544 be fetched from the storage layer. Done at the submit_bio() level, so it is
1545 accurate for block-backed filesystems. <please add status regarding NFS and
1546 CIFS at a later time>
1552 I/O counter: bytes written
1553 Attempt to count the number of bytes which this process caused to be sent to
1554 the storage layer. This is done at page-dirtying time.
1557 cancelled_write_bytes
1558 ---------------------
1560 The big inaccuracy here is truncate. If a process writes 1MB to a file and
1561 then deletes the file, it will in fact perform no writeout. But it will have
1562 been accounted as having caused 1MB of write.
1563 In other words: The number of bytes which this process caused to not happen,
1564 by truncating pagecache. A task can cause "negative" IO too. If this task
1565 truncates some dirty pagecache, some IO which another task has been accounted
1566 for (in its write_bytes) will not be happening. We _could_ just subtract that
1567 from the truncating task's write_bytes, but there is information loss in doing
1574 At its current implementation state, this is a bit racy on 32-bit machines: if
1575 process A reads process B's /proc/pid/io while process B is updating one of
1576 those 64-bit counters, process A could see an intermediate result.
1579 More information about this can be found within the taskstats documentation in
1580 Documentation/accounting.
1582 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings
1583 ---------------------------------------------------------------
1584 When a process is dumped, all anonymous memory is written to a core file as
1585 long as the size of the core file isn't limited. But sometimes we don't want
1586 to dump some memory segments, for example, huge shared memory or DAX.
1587 Conversely, sometimes we want to save file-backed memory segments into a core
1588 file, not only the individual files.
1590 /proc/<pid>/coredump_filter allows you to customize which memory segments
1591 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
1592 of memory types. If a bit of the bitmask is set, memory segments of the
1593 corresponding memory type are dumped, otherwise they are not dumped.
1595 The following 9 memory types are supported:
1596 - (bit 0) anonymous private memory
1597 - (bit 1) anonymous shared memory
1598 - (bit 2) file-backed private memory
1599 - (bit 3) file-backed shared memory
1600 - (bit 4) ELF header pages in file-backed private memory areas (it is
1601 effective only if the bit 2 is cleared)
1602 - (bit 5) hugetlb private memory
1603 - (bit 6) hugetlb shared memory
1604 - (bit 7) DAX private memory
1605 - (bit 8) DAX shared memory
1607 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
1608 are always dumped regardless of the bitmask status.
1610 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is
1611 only affected by bit 5-6, and DAX is only affected by bits 7-8.
1613 The default value of coredump_filter is 0x33; this means all anonymous memory
1614 segments, ELF header pages and hugetlb private memory are dumped.
1616 If you don't want to dump all shared memory segments attached to pid 1234,
1617 write 0x31 to the process's proc file.
1619 $ echo 0x31 > /proc/1234/coredump_filter
1621 When a new process is created, the process inherits the bitmask status from its
1622 parent. It is useful to set up coredump_filter before the program runs.
1625 $ echo 0x7 > /proc/self/coredump_filter
1628 3.5 /proc/<pid>/mountinfo - Information about mounts
1629 --------------------------------------------------------
1631 This file contains lines of the form:
1633 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
1634 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
1636 (1) mount ID: unique identifier of the mount (may be reused after umount)
1637 (2) parent ID: ID of parent (or of self for the top of the mount tree)
1638 (3) major:minor: value of st_dev for files on filesystem
1639 (4) root: root of the mount within the filesystem
1640 (5) mount point: mount point relative to the process's root
1641 (6) mount options: per mount options
1642 (7) optional fields: zero or more fields of the form "tag[:value]"
1643 (8) separator: marks the end of the optional fields
1644 (9) filesystem type: name of filesystem of the form "type[.subtype]"
1645 (10) mount source: filesystem specific information or "none"
1646 (11) super options: per super block options
1648 Parsers should ignore all unrecognised optional fields. Currently the
1649 possible optional fields are:
1651 shared:X mount is shared in peer group X
1652 master:X mount is slave to peer group X
1653 propagate_from:X mount is slave and receives propagation from peer group X (*)
1654 unbindable mount is unbindable
1656 (*) X is the closest dominant peer group under the process's root. If
1657 X is the immediate master of the mount, or if there's no dominant peer
1658 group under the same root, then only the "master:X" field is present
1659 and not the "propagate_from:X" field.
1661 For more information on mount propagation see:
1663 Documentation/filesystems/sharedsubtree.txt
1666 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm
1667 --------------------------------------------------------
1668 These files provide a method to access a tasks comm value. It also allows for
1669 a task to set its own or one of its thread siblings comm value. The comm value
1670 is limited in size compared to the cmdline value, so writing anything longer
1671 then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated
1675 3.7 /proc/<pid>/task/<tid>/children - Information about task children
1676 -------------------------------------------------------------------------
1677 This file provides a fast way to retrieve first level children pids
1678 of a task pointed by <pid>/<tid> pair. The format is a space separated
1681 Note the "first level" here -- if a child has own children they will
1682 not be listed here, one needs to read /proc/<children-pid>/task/<tid>/children
1683 to obtain the descendants.
1685 Since this interface is intended to be fast and cheap it doesn't
1686 guarantee to provide precise results and some children might be
1687 skipped, especially if they've exited right after we printed their
1688 pids, so one need to either stop or freeze processes being inspected
1689 if precise results are needed.
1692 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file
1693 ---------------------------------------------------------------
1694 This file provides information associated with an opened file. The regular
1695 files have at least three fields -- 'pos', 'flags' and mnt_id. The 'pos'
1696 represents the current offset of the opened file in decimal form [see lseek(2)
1697 for details], 'flags' denotes the octal O_xxx mask the file has been
1698 created with [see open(2) for details] and 'mnt_id' represents mount ID of
1699 the file system containing the opened file [see 3.5 /proc/<pid>/mountinfo
1708 All locks associated with a file descriptor are shown in its fdinfo too.
1710 lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF
1712 The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags
1713 pair provide additional information particular to the objects they represent.
1722 where 'eventfd-count' is hex value of a counter.
1729 sigmask: 0000000000000200
1731 where 'sigmask' is hex value of the signal mask associated
1739 tfd: 5 events: 1d data: ffffffffffffffff
1741 where 'tfd' is a target file descriptor number in decimal form,
1742 'events' is events mask being watched and the 'data' is data
1743 associated with a target [see epoll(7) for more details].
1747 For inotify files the format is the following
1751 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d
1753 where 'wd' is a watch descriptor in decimal form, ie a target file
1754 descriptor number, 'ino' and 'sdev' are inode and device where the
1755 target file resides and the 'mask' is the mask of events, all in hex
1756 form [see inotify(7) for more details].
1758 If the kernel was built with exportfs support, the path to the target
1759 file is encoded as a file handle. The file handle is provided by three
1760 fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex
1763 If the kernel is built without exportfs support the file handle won't be
1766 If there is no inotify mark attached yet the 'inotify' line will be omitted.
1768 For fanotify files the format is
1773 fanotify flags:10 event-flags:0
1774 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003
1775 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4
1777 where fanotify 'flags' and 'event-flags' are values used in fanotify_init
1778 call, 'mnt_id' is the mount point identifier, 'mflags' is the value of
1779 flags associated with mark which are tracked separately from events
1780 mask. 'ino', 'sdev' are target inode and device, 'mask' is the events
1781 mask and 'ignored_mask' is the mask of events which are to be ignored.
1782 All in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask'
1783 does provide information about flags and mask used in fanotify_mark
1784 call [see fsnotify manpage for details].
1786 While the first three lines are mandatory and always printed, the rest is
1787 optional and may be omitted if no marks created yet.
1798 it_value: (0, 49406829)
1801 where 'clockid' is the clock type and 'ticks' is the number of the timer expirations
1802 that have occurred [see timerfd_create(2) for details]. 'settime flags' are
1803 flags in octal form been used to setup the timer [see timerfd_settime(2) for
1804 details]. 'it_value' is remaining time until the timer exiration.
1805 'it_interval' is the interval for the timer. Note the timer might be set up
1806 with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value'
1807 still exhibits timer's remaining time.
1809 3.9 /proc/<pid>/map_files - Information about memory mapped files
1810 ---------------------------------------------------------------------
1811 This directory contains symbolic links which represent memory mapped files
1812 the process is maintaining. Example output:
1814 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so
1815 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so
1816 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so
1818 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1
1819 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls
1821 The name of a link represents the virtual memory bounds of a mapping, i.e.
1822 vm_area_struct::vm_start-vm_area_struct::vm_end.
1824 The main purpose of the map_files is to retrieve a set of memory mapped
1825 files in a fast way instead of parsing /proc/<pid>/maps or
1826 /proc/<pid>/smaps, both of which contain many more records. At the same
1827 time one can open(2) mappings from the listings of two processes and
1828 comparing their inode numbers to figure out which anonymous memory areas
1829 are actually shared.
1831 3.10 /proc/<pid>/timerslack_ns - Task timerslack value
1832 ---------------------------------------------------------
1833 This file provides the value of the task's timerslack value in nanoseconds.
1834 This value specifies a amount of time that normal timers may be deferred
1835 in order to coalesce timers and avoid unnecessary wakeups.
1837 This allows a task's interactivity vs power consumption trade off to be
1840 Writing 0 to the file will set the tasks timerslack to the default value.
1842 Valid values are from 0 - ULLONG_MAX
1844 An application setting the value must have PTRACE_MODE_ATTACH_FSCREDS level
1845 permissions on the task specified to change its timerslack_ns value.
1848 ------------------------------------------------------------------------------
1850 ------------------------------------------------------------------------------
1853 ---------------------
1855 The following mount options are supported:
1857 hidepid= Set /proc/<pid>/ access mode.
1858 gid= Set the group authorized to learn processes information.
1860 hidepid=0 means classic mode - everybody may access all /proc/<pid>/ directories
1863 hidepid=1 means users may not access any /proc/<pid>/ directories but their
1864 own. Sensitive files like cmdline, sched*, status are now protected against
1865 other users. This makes it impossible to learn whether any user runs
1866 specific program (given the program doesn't reveal itself by its behaviour).
1867 As an additional bonus, as /proc/<pid>/cmdline is unaccessible for other users,
1868 poorly written programs passing sensitive information via program arguments are
1869 now protected against local eavesdroppers.
1871 hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be fully invisible to other
1872 users. It doesn't mean that it hides a fact whether a process with a specific
1873 pid value exists (it can be learned by other means, e.g. by "kill -0 $PID"),
1874 but it hides process' uid and gid, which may be learned by stat()'ing
1875 /proc/<pid>/ otherwise. It greatly complicates an intruder's task of gathering
1876 information about running processes, whether some daemon runs with elevated
1877 privileges, whether other user runs some sensitive program, whether other users
1878 run any program at all, etc.
1880 gid= defines a group authorized to learn processes information otherwise
1881 prohibited by hidepid=. If you use some daemon like identd which needs to learn
1882 information about processes information, just add identd to this group.