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
50 ------------------------------------------------------------------------------
52 ------------------------------------------------------------------------------
54 0.1 Introduction/Credits
55 ------------------------
57 This documentation is part of a soon (or so we hope) to be released book on
58 the SuSE Linux distribution. As there is no complete documentation for the
59 /proc file system and we've used many freely available sources to write these
60 chapters, it seems only fair to give the work back to the Linux community.
61 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
62 afraid it's still far from complete, but we hope it will be useful. As far as
63 we know, it is the first 'all-in-one' document about the /proc file system. It
64 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
65 SPARC, AXP, etc., features, you probably won't find what you are looking for.
66 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
67 additions and patches are welcome and will be added to this document if you
70 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
71 other people for help compiling this documentation. We'd also like to extend a
72 special thank you to Andi Kleen for documentation, which we relied on heavily
73 to create this document, as well as the additional information he provided.
74 Thanks to everybody else who contributed source or docs to the Linux kernel
75 and helped create a great piece of software... :)
77 If you have any comments, corrections or additions, please don't hesitate to
78 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
81 The latest version of this document is available online at
82 http://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html
84 If the above direction does not works for you, you could try the kernel
85 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
86 comandante@zaralinux.com.
91 We don't guarantee the correctness of this document, and if you come to us
92 complaining about how you screwed up your system because of incorrect
93 documentation, we won't feel responsible...
95 ------------------------------------------------------------------------------
96 CHAPTER 1: COLLECTING SYSTEM INFORMATION
97 ------------------------------------------------------------------------------
99 ------------------------------------------------------------------------------
101 ------------------------------------------------------------------------------
102 * Investigating the properties of the pseudo file system /proc and its
103 ability to provide information on the running Linux system
104 * Examining /proc's structure
105 * Uncovering various information about the kernel and the processes running
107 ------------------------------------------------------------------------------
110 The proc file system acts as an interface to internal data structures in the
111 kernel. It can be used to obtain information about the system and to change
112 certain kernel parameters at runtime (sysctl).
114 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
115 show you how you can use /proc/sys to change settings.
117 1.1 Process-Specific Subdirectories
118 -----------------------------------
120 The directory /proc contains (among other things) one subdirectory for each
121 process running on the system, which is named after the process ID (PID).
123 The link self points to the process reading the file system. Each process
124 subdirectory has the entries listed in Table 1-1.
127 Table 1-1: Process specific entries in /proc
128 ..............................................................................
130 clear_refs Clears page referenced bits shown in smaps output
131 cmdline Command line arguments
132 cpu Current and last cpu in which it was executed (2.4)(smp)
133 cwd Link to the current working directory
134 environ Values of environment variables
135 exe Link to the executable of this process
136 fd Directory, which contains all file descriptors
137 maps Memory maps to executables and library files (2.4)
138 mem Memory held by this process
139 root Link to the root directory of this process
141 statm Process memory status information
142 status Process status in human readable form
143 wchan Present with CONFIG_KALLSYMS=y: it shows the kernel function
144 symbol the task is blocked in - or "0" if not blocked.
146 stack Report full stack trace, enable via CONFIG_STACKTRACE
147 smaps a extension based on maps, showing the memory consumption of
148 each mapping and flags associated with it
149 numa_maps an extension based on maps, showing the memory locality and
150 binding policy as well as mem usage (in pages) of each mapping.
151 ..............................................................................
153 For example, to get the status information of a process, all you have to do is
154 read the file /proc/PID/status:
156 >cat /proc/self/status
184 SigPnd: 0000000000000000
185 ShdPnd: 0000000000000000
186 SigBlk: 0000000000000000
187 SigIgn: 0000000000000000
188 SigCgt: 0000000000000000
189 CapInh: 00000000fffffeff
190 CapPrm: 0000000000000000
191 CapEff: 0000000000000000
192 CapBnd: ffffffffffffffff
194 voluntary_ctxt_switches: 0
195 nonvoluntary_ctxt_switches: 1
197 This shows you nearly the same information you would get if you viewed it with
198 the ps command. In fact, ps uses the proc file system to obtain its
199 information. But you get a more detailed view of the process by reading the
200 file /proc/PID/status. It fields are described in table 1-2.
202 The statm file contains more detailed information about the process
203 memory usage. Its seven fields are explained in Table 1-3. The stat file
204 contains details information about the process itself. Its fields are
205 explained in Table 1-4.
207 (for SMP CONFIG users)
208 For making accounting scalable, RSS related information are handled in an
209 asynchronous manner and the value may not be very precise. To see a precise
210 snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table.
211 It's slow but very precise.
213 Table 1-2: Contents of the status files (as of 4.1)
214 ..............................................................................
216 Name filename of the executable
217 State state (R is running, S is sleeping, D is sleeping
218 in an uninterruptible wait, Z is zombie,
219 T is traced or stopped)
221 Ngid NUMA group ID (0 if none)
223 PPid process id of the parent process
224 TracerPid PID of process tracing this process (0 if not)
225 Uid Real, effective, saved set, and file system UIDs
226 Gid Real, effective, saved set, and file system GIDs
227 FDSize number of file descriptor slots currently allocated
228 Groups supplementary group list
229 NStgid descendant namespace thread group ID hierarchy
230 NSpid descendant namespace process ID hierarchy
231 NSpgid descendant namespace process group ID hierarchy
232 NSsid descendant namespace session ID hierarchy
233 VmPeak peak virtual memory size
234 VmSize total program size
235 VmLck locked memory size
236 VmHWM peak resident set size ("high water mark")
237 VmRSS size of memory portions. It contains the three
238 following parts (VmRSS = RssAnon + RssFile + RssShmem)
239 RssAnon size of resident anonymous memory
240 RssFile size of resident file mappings
241 RssShmem size of resident shmem memory (includes SysV shm,
242 mapping of tmpfs and shared anonymous mappings)
243 VmData size of private data segments
244 VmStk size of stack segments
245 VmExe size of text segment
246 VmLib size of shared library code
247 VmPTE size of page table entries
248 VmPMD size of second level page tables
249 VmSwap amount of swap used by anonymous private data
250 (shmem swap usage is not included)
251 HugetlbPages size of hugetlb memory portions
252 Threads number of threads
253 SigQ number of signals queued/max. number for queue
254 SigPnd bitmap of pending signals for the thread
255 ShdPnd bitmap of shared pending signals for the process
256 SigBlk bitmap of blocked signals
257 SigIgn bitmap of ignored signals
258 SigCgt bitmap of caught signals
259 CapInh bitmap of inheritable capabilities
260 CapPrm bitmap of permitted capabilities
261 CapEff bitmap of effective capabilities
262 CapBnd bitmap of capabilities bounding set
263 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...)
264 Cpus_allowed mask of CPUs on which this process may run
265 Cpus_allowed_list Same as previous, but in "list format"
266 Mems_allowed mask of memory nodes allowed to this process
267 Mems_allowed_list Same as previous, but in "list format"
268 voluntary_ctxt_switches number of voluntary context switches
269 nonvoluntary_ctxt_switches number of non voluntary context switches
270 ..............................................................................
272 Table 1-3: Contents of the statm files (as of 2.6.8-rc3)
273 ..............................................................................
275 size total program size (pages) (same as VmSize in status)
276 resident size of memory portions (pages) (same as VmRSS in status)
277 shared number of pages that are shared (i.e. backed by a file, same
278 as RssFile+RssShmem in status)
279 trs number of pages that are 'code' (not including libs; broken,
280 includes data segment)
281 lrs number of pages of library (always 0 on 2.6)
282 drs number of pages of data/stack (including libs; broken,
283 includes library text)
284 dt number of dirty pages (always 0 on 2.6)
285 ..............................................................................
288 Table 1-4: Contents of the stat files (as of 2.6.30-rc7)
289 ..............................................................................
292 tcomm filename of the executable
293 state state (R is running, S is sleeping, D is sleeping in an
294 uninterruptible wait, Z is zombie, T is traced or stopped)
295 ppid process id of the parent process
296 pgrp pgrp of the process
298 tty_nr tty the process uses
299 tty_pgrp pgrp of the tty
301 min_flt number of minor faults
302 cmin_flt number of minor faults with child's
303 maj_flt number of major faults
304 cmaj_flt number of major faults with child's
305 utime user mode jiffies
306 stime kernel mode jiffies
307 cutime user mode jiffies with child's
308 cstime kernel mode jiffies with child's
309 priority priority level
311 num_threads number of threads
312 it_real_value (obsolete, always 0)
313 start_time time the process started after system boot
314 vsize virtual memory size
315 rss resident set memory size
316 rsslim current limit in bytes on the rss
317 start_code address above which program text can run
318 end_code address below which program text can run
319 start_stack address of the start of the main process stack
320 esp current value of ESP
321 eip current value of EIP
322 pending bitmap of pending signals
323 blocked bitmap of blocked signals
324 sigign bitmap of ignored signals
325 sigcatch bitmap of caught signals
326 0 (place holder, used to be the wchan address, use /proc/PID/wchan instead)
329 exit_signal signal to send to parent thread on exit
330 task_cpu which CPU the task is scheduled on
331 rt_priority realtime priority
332 policy scheduling policy (man sched_setscheduler)
333 blkio_ticks time spent waiting for block IO
334 gtime guest time of the task in jiffies
335 cgtime guest time of the task children in jiffies
336 start_data address above which program data+bss is placed
337 end_data address below which program data+bss is placed
338 start_brk address above which program heap can be expanded with brk()
339 arg_start address above which program command line is placed
340 arg_end address below which program command line is placed
341 env_start address above which program environment is placed
342 env_end address below which program environment is placed
343 exit_code the thread's exit_code in the form reported by the waitpid system call
344 ..............................................................................
346 The /proc/PID/maps file containing the currently mapped memory regions and
347 their access permissions.
351 address perms offset dev inode pathname
353 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
354 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
355 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
356 a7cb1000-a7cb2000 ---p 00000000 00:00 0
357 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
358 a7eb2000-a7eb3000 ---p 00000000 00:00 0
359 a7eb3000-a7ed5000 rw-p 00000000 00:00 0
360 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
361 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
362 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
363 a800b000-a800e000 rw-p 00000000 00:00 0
364 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
365 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
366 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
367 a8024000-a8027000 rw-p 00000000 00:00 0
368 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
369 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
370 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
371 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack]
372 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
374 where "address" is the address space in the process that it occupies, "perms"
375 is a set of permissions:
381 p = private (copy on write)
383 "offset" is the offset into the mapping, "dev" is the device (major:minor), and
384 "inode" is the inode on that device. 0 indicates that no inode is associated
385 with the memory region, as the case would be with BSS (uninitialized data).
386 The "pathname" shows the name associated file for this mapping. If the mapping
387 is not associated with a file:
389 [heap] = the heap of the program
390 [stack] = the stack of the main process
391 [vdso] = the "virtual dynamic shared object",
392 the kernel system call handler
394 or if empty, the mapping is anonymous.
396 The /proc/PID/task/TID/maps is a view of the virtual memory from the viewpoint
397 of the individual tasks of a process. In this file you will see a mapping marked
398 as [stack] if that task sees it as a stack. Hence, for the example above, the
399 task-level map, i.e. /proc/PID/task/TID/maps for thread 1001 will look like this:
401 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
402 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
403 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
404 a7cb1000-a7cb2000 ---p 00000000 00:00 0
405 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
406 a7eb2000-a7eb3000 ---p 00000000 00:00 0
407 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 [stack]
408 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
409 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
410 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
411 a800b000-a800e000 rw-p 00000000 00:00 0
412 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
413 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
414 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
415 a8024000-a8027000 rw-p 00000000 00:00 0
416 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
417 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
418 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
419 aff35000-aff4a000 rw-p 00000000 00:00 0
420 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
422 The /proc/PID/smaps is an extension based on maps, showing the memory
423 consumption for each of the process's mappings. For each of mappings there
424 is a series of lines such as the following:
426 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash
438 Private_Hugetlb: 0 kB
444 VmFlags: rd ex mr mw me dw
446 the first of these lines shows the same information as is displayed for the
447 mapping in /proc/PID/maps. The remaining lines show the size of the mapping
448 (size), the amount of the mapping that is currently resident in RAM (RSS), the
449 process' proportional share of this mapping (PSS), the number of clean and
450 dirty private pages in the mapping.
452 The "proportional set size" (PSS) of a process is the count of pages it has
453 in memory, where each page is divided by the number of processes sharing it.
454 So if a process has 1000 pages all to itself, and 1000 shared with one other
455 process, its PSS will be 1500.
456 Note that even a page which is part of a MAP_SHARED mapping, but has only
457 a single pte mapped, i.e. is currently used by only one process, is accounted
458 as private and not as shared.
459 "Referenced" indicates the amount of memory currently marked as referenced or
461 "Anonymous" shows the amount of memory that does not belong to any file. Even
462 a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE
463 and a page is modified, the file page is replaced by a private anonymous copy.
464 "AnonHugePages" shows the ammount of memory backed by transparent hugepage.
465 "Shared_Hugetlb" and "Private_Hugetlb" show the ammounts of memory backed by
466 hugetlbfs page which is *not* counted in "RSS" or "PSS" field for historical
467 reasons. And these are not included in {Shared,Private}_{Clean,Dirty} field.
468 "Swap" shows how much would-be-anonymous memory is also used, but out on swap.
469 For shmem mappings, "Swap" includes also the size of the mapped (and not
470 replaced by copy-on-write) part of the underlying shmem object out on swap.
471 "SwapPss" shows proportional swap share of this mapping. Unlike "Swap", this
472 does not take into account swapped out page of underlying shmem objects.
473 "Locked" indicates whether the mapping is locked in memory or not.
475 "VmFlags" field deserves a separate description. This member represents the kernel
476 flags associated with the particular virtual memory area in two letter encoded
477 manner. The codes are the following:
486 gd - stack segment growns down
488 dw - disabled write to the mapped file
489 lo - pages are locked in memory
490 io - memory mapped I/O area
491 sr - sequential read advise provided
492 rr - random read advise provided
493 dc - do not copy area on fork
494 de - do not expand area on remapping
495 ac - area is accountable
496 nr - swap space is not reserved for the area
497 ht - area uses huge tlb pages
498 ar - architecture specific flag
499 dd - do not include area into core dump
502 hg - huge page advise flag
503 nh - no-huge page advise flag
504 mg - mergable advise flag
506 Note that there is no guarantee that every flag and associated mnemonic will
507 be present in all further kernel releases. Things get changed, the flags may
508 be vanished or the reverse -- new added.
510 This file is only present if the CONFIG_MMU kernel configuration option is
513 The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG
514 bits on both physical and virtual pages associated with a process, and the
515 soft-dirty bit on pte (see Documentation/vm/soft-dirty.txt for details).
516 To clear the bits for all the pages associated with the process
517 > echo 1 > /proc/PID/clear_refs
519 To clear the bits for the anonymous pages associated with the process
520 > echo 2 > /proc/PID/clear_refs
522 To clear the bits for the file mapped pages associated with the process
523 > echo 3 > /proc/PID/clear_refs
525 To clear the soft-dirty bit
526 > echo 4 > /proc/PID/clear_refs
528 To reset the peak resident set size ("high water mark") to the process's
530 > echo 5 > /proc/PID/clear_refs
532 Any other value written to /proc/PID/clear_refs will have no effect.
534 The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags
535 using /proc/kpageflags and number of times a page is mapped using
536 /proc/kpagecount. For detailed explanation, see Documentation/vm/pagemap.txt.
538 The /proc/pid/numa_maps is an extension based on maps, showing the memory
539 locality and binding policy, as well as the memory usage (in pages) of
540 each mapping. The output follows a general format where mapping details get
541 summarized separated by blank spaces, one mapping per each file line:
543 address policy mapping details
545 00400000 default file=/usr/local/bin/app mapped=1 active=0 N3=1 kernelpagesize_kB=4
546 00600000 default file=/usr/local/bin/app anon=1 dirty=1 N3=1 kernelpagesize_kB=4
547 3206000000 default file=/lib64/ld-2.12.so mapped=26 mapmax=6 N0=24 N3=2 kernelpagesize_kB=4
548 320621f000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
549 3206220000 default file=/lib64/ld-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
550 3206221000 default anon=1 dirty=1 N3=1 kernelpagesize_kB=4
551 3206800000 default file=/lib64/libc-2.12.so mapped=59 mapmax=21 active=55 N0=41 N3=18 kernelpagesize_kB=4
552 320698b000 default file=/lib64/libc-2.12.so
553 3206b8a000 default file=/lib64/libc-2.12.so anon=2 dirty=2 N3=2 kernelpagesize_kB=4
554 3206b8e000 default file=/lib64/libc-2.12.so anon=1 dirty=1 N3=1 kernelpagesize_kB=4
555 3206b8f000 default anon=3 dirty=3 active=1 N3=3 kernelpagesize_kB=4
556 7f4dc10a2000 default anon=3 dirty=3 N3=3 kernelpagesize_kB=4
557 7f4dc10b4000 default anon=2 dirty=2 active=1 N3=2 kernelpagesize_kB=4
558 7f4dc1200000 default file=/anon_hugepage\040(deleted) huge anon=1 dirty=1 N3=1 kernelpagesize_kB=2048
559 7fff335f0000 default stack anon=3 dirty=3 N3=3 kernelpagesize_kB=4
560 7fff3369d000 default mapped=1 mapmax=35 active=0 N3=1 kernelpagesize_kB=4
563 "address" is the starting address for the mapping;
564 "policy" reports the NUMA memory policy set for the mapping (see vm/numa_memory_policy.txt);
565 "mapping details" summarizes mapping data such as mapping type, page usage counters,
566 node locality page counters (N0 == node0, N1 == node1, ...) and the kernel page
567 size, in KB, that is backing the mapping up.
572 Similar to the process entries, the kernel data files give information about
573 the running kernel. The files used to obtain this information are contained in
574 /proc and are listed in Table 1-5. Not all of these will be present in your
575 system. It depends on the kernel configuration and the loaded modules, which
576 files are there, and which are missing.
578 Table 1-5: Kernel info in /proc
579 ..............................................................................
581 apm Advanced power management info
582 buddyinfo Kernel memory allocator information (see text) (2.5)
583 bus Directory containing bus specific information
584 cmdline Kernel command line
585 cpuinfo Info about the CPU
586 devices Available devices (block and character)
587 dma Used DMS channels
588 filesystems Supported filesystems
589 driver Various drivers grouped here, currently rtc (2.4)
590 execdomains Execdomains, related to security (2.4)
591 fb Frame Buffer devices (2.4)
592 fs File system parameters, currently nfs/exports (2.4)
593 ide Directory containing info about the IDE subsystem
594 interrupts Interrupt usage
595 iomem Memory map (2.4)
596 ioports I/O port usage
597 irq Masks for irq to cpu affinity (2.4)(smp?)
598 isapnp ISA PnP (Plug&Play) Info (2.4)
599 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
601 ksyms Kernel symbol table
602 loadavg Load average of last 1, 5 & 15 minutes
606 modules List of loaded modules
607 mounts Mounted filesystems
608 net Networking info (see text)
609 pagetypeinfo Additional page allocator information (see text) (2.5)
610 partitions Table of partitions known to the system
611 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
612 decoupled by lspci (2.4)
614 scsi SCSI info (see text)
615 slabinfo Slab pool info
616 softirqs softirq usage
617 stat Overall statistics
618 swaps Swap space utilization
620 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
621 tty Info of tty drivers
622 uptime Wall clock since boot, combined idle time of all cpus
623 version Kernel version
624 video bttv info of video resources (2.4)
625 vmallocinfo Show vmalloced areas
626 ..............................................................................
628 You can, for example, check which interrupts are currently in use and what
629 they are used for by looking in the file /proc/interrupts:
631 > cat /proc/interrupts
633 0: 8728810 XT-PIC timer
634 1: 895 XT-PIC keyboard
636 3: 531695 XT-PIC aha152x
637 4: 2014133 XT-PIC serial
638 5: 44401 XT-PIC pcnet_cs
641 12: 182918 XT-PIC PS/2 Mouse
643 14: 1232265 XT-PIC ide0
647 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
648 output of a SMP machine):
650 > cat /proc/interrupts
653 0: 1243498 1214548 IO-APIC-edge timer
654 1: 8949 8958 IO-APIC-edge keyboard
655 2: 0 0 XT-PIC cascade
656 5: 11286 10161 IO-APIC-edge soundblaster
657 8: 1 0 IO-APIC-edge rtc
658 9: 27422 27407 IO-APIC-edge 3c503
659 12: 113645 113873 IO-APIC-edge PS/2 Mouse
661 14: 22491 24012 IO-APIC-edge ide0
662 15: 2183 2415 IO-APIC-edge ide1
663 17: 30564 30414 IO-APIC-level eth0
664 18: 177 164 IO-APIC-level bttv
669 NMI is incremented in this case because every timer interrupt generates a NMI
670 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
672 LOC is the local interrupt counter of the internal APIC of every CPU.
674 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
675 connects the CPUs in a SMP system. This means that an error has been detected,
676 the IO-APIC automatically retry the transmission, so it should not be a big
677 problem, but you should read the SMP-FAQ.
679 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
680 /proc/interrupts to display every IRQ vector in use by the system, not
681 just those considered 'most important'. The new vectors are:
683 THR -- interrupt raised when a machine check threshold counter
684 (typically counting ECC corrected errors of memory or cache) exceeds
685 a configurable threshold. Only available on some systems.
687 TRM -- a thermal event interrupt occurs when a temperature threshold
688 has been exceeded for the CPU. This interrupt may also be generated
689 when the temperature drops back to normal.
691 SPU -- a spurious interrupt is some interrupt that was raised then lowered
692 by some IO device before it could be fully processed by the APIC. Hence
693 the APIC sees the interrupt but does not know what device it came from.
694 For this case the APIC will generate the interrupt with a IRQ vector
695 of 0xff. This might also be generated by chipset bugs.
697 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
698 sent from one CPU to another per the needs of the OS. Typically,
699 their statistics are used by kernel developers and interested users to
700 determine the occurrence of interrupts of the given type.
702 The above IRQ vectors are displayed only when relevant. For example,
703 the threshold vector does not exist on x86_64 platforms. Others are
704 suppressed when the system is a uniprocessor. As of this writing, only
705 i386 and x86_64 platforms support the new IRQ vector displays.
707 Of some interest is the introduction of the /proc/irq directory to 2.4.
708 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
709 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
710 irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and
715 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
716 1 11 13 15 17 19 3 5 7 9 default_smp_affinity
720 smp_affinity is a bitmask, in which you can specify which CPUs can handle the
721 IRQ, you can set it by doing:
723 > echo 1 > /proc/irq/10/smp_affinity
725 This means that only the first CPU will handle the IRQ, but you can also echo
726 5 which means that only the first and fourth CPU can handle the IRQ.
728 The contents of each smp_affinity file is the same by default:
730 > cat /proc/irq/0/smp_affinity
733 There is an alternate interface, smp_affinity_list which allows specifying
734 a cpu range instead of a bitmask:
736 > cat /proc/irq/0/smp_affinity_list
739 The default_smp_affinity mask applies to all non-active IRQs, which are the
740 IRQs which have not yet been allocated/activated, and hence which lack a
741 /proc/irq/[0-9]* directory.
743 The node file on an SMP system shows the node to which the device using the IRQ
744 reports itself as being attached. This hardware locality information does not
745 include information about any possible driver locality preference.
747 prof_cpu_mask specifies which CPUs are to be profiled by the system wide
748 profiler. Default value is ffffffff (all cpus if there are only 32 of them).
750 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
751 between all the CPUs which are allowed to handle it. As usual the kernel has
752 more info than you and does a better job than you, so the defaults are the
753 best choice for almost everyone. [Note this applies only to those IO-APIC's
754 that support "Round Robin" interrupt distribution.]
756 There are three more important subdirectories in /proc: net, scsi, and sys.
757 The general rule is that the contents, or even the existence of these
758 directories, depend on your kernel configuration. If SCSI is not enabled, the
759 directory scsi may not exist. The same is true with the net, which is there
760 only when networking support is present in the running kernel.
762 The slabinfo file gives information about memory usage at the slab level.
763 Linux uses slab pools for memory management above page level in version 2.2.
764 Commonly used objects have their own slab pool (such as network buffers,
765 directory cache, and so on).
767 ..............................................................................
769 > cat /proc/buddyinfo
771 Node 0, zone DMA 0 4 5 4 4 3 ...
772 Node 0, zone Normal 1 0 0 1 101 8 ...
773 Node 0, zone HighMem 2 0 0 1 1 0 ...
775 External fragmentation is a problem under some workloads, and buddyinfo is a
776 useful tool for helping diagnose these problems. Buddyinfo will give you a
777 clue as to how big an area you can safely allocate, or why a previous
780 Each column represents the number of pages of a certain order which are
781 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
782 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
783 available in ZONE_NORMAL, etc...
785 More information relevant to external fragmentation can be found in
788 > cat /proc/pagetypeinfo
792 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
793 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0
794 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0
795 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2
796 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0
797 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0
798 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9
799 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0
800 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452
801 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0
802 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0
804 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate
805 Node 0, zone DMA 2 0 5 1 0
806 Node 0, zone DMA32 41 6 967 2 0
808 Fragmentation avoidance in the kernel works by grouping pages of different
809 migrate types into the same contiguous regions of memory called page blocks.
810 A page block is typically the size of the default hugepage size e.g. 2MB on
811 X86-64. By keeping pages grouped based on their ability to move, the kernel
812 can reclaim pages within a page block to satisfy a high-order allocation.
814 The pagetypinfo begins with information on the size of a page block. It
815 then gives the same type of information as buddyinfo except broken down
816 by migrate-type and finishes with details on how many page blocks of each
819 If min_free_kbytes has been tuned correctly (recommendations made by hugeadm
820 from libhugetlbfs https://github.com/libhugetlbfs/libhugetlbfs/), one can
821 make an estimate of the likely number of huge pages that can be allocated
822 at a given point in time. All the "Movable" blocks should be allocatable
823 unless memory has been mlock()'d. Some of the Reclaimable blocks should
824 also be allocatable although a lot of filesystem metadata may have to be
825 reclaimed to achieve this.
827 ..............................................................................
831 Provides information about distribution and utilization of memory. This
832 varies by architecture and compile options. The following is from a
833 16GB PIII, which has highmem enabled. You may not have all of these fields.
837 MemTotal: 16344972 kB
839 MemAvailable: 14836172 kB
845 HighTotal: 15597528 kB
846 HighFree: 13629632 kB
857 SReclaimable: 159856 kB
858 SUnreclaim: 124508 kB
863 CommitLimit: 7669796 kB
864 Committed_AS: 100056 kB
865 VmallocTotal: 112216 kB
867 VmallocChunk: 111088 kB
868 AnonHugePages: 49152 kB
870 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
871 bits and the kernel binary code)
872 MemFree: The sum of LowFree+HighFree
873 MemAvailable: An estimate of how much memory is available for starting new
874 applications, without swapping. Calculated from MemFree,
875 SReclaimable, the size of the file LRU lists, and the low
876 watermarks in each zone.
877 The estimate takes into account that the system needs some
878 page cache to function well, and that not all reclaimable
879 slab will be reclaimable, due to items being in use. The
880 impact of those factors will vary from system to system.
881 Buffers: Relatively temporary storage for raw disk blocks
882 shouldn't get tremendously large (20MB or so)
883 Cached: in-memory cache for files read from the disk (the
884 pagecache). Doesn't include SwapCached
885 SwapCached: Memory that once was swapped out, is swapped back in but
886 still also is in the swapfile (if memory is needed it
887 doesn't need to be swapped out AGAIN because it is already
888 in the swapfile. This saves I/O)
889 Active: Memory that has been used more recently and usually not
890 reclaimed unless absolutely necessary.
891 Inactive: Memory which has been less recently used. It is more
892 eligible to be reclaimed for other purposes
894 HighFree: Highmem is all memory above ~860MB of physical memory
895 Highmem areas are for use by userspace programs, or
896 for the pagecache. The kernel must use tricks to access
897 this memory, making it slower to access than lowmem.
899 LowFree: Lowmem is memory which can be used for everything that
900 highmem can be used for, but it is also available for the
901 kernel's use for its own data structures. Among many
902 other things, it is where everything from the Slab is
903 allocated. Bad things happen when you're out of lowmem.
904 SwapTotal: total amount of swap space available
905 SwapFree: Memory which has been evicted from RAM, and is temporarily
907 Dirty: Memory which is waiting to get written back to the disk
908 Writeback: Memory which is actively being written back to the disk
909 AnonPages: Non-file backed pages mapped into userspace page tables
910 AnonHugePages: Non-file backed huge pages mapped into userspace page tables
911 Mapped: files which have been mmaped, such as libraries
912 Shmem: Total memory used by shared memory (shmem) and tmpfs
913 Slab: in-kernel data structures cache
914 SReclaimable: Part of Slab, that might be reclaimed, such as caches
915 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure
916 PageTables: amount of memory dedicated to the lowest level of page
918 NFS_Unstable: NFS pages sent to the server, but not yet committed to stable
920 Bounce: Memory used for block device "bounce buffers"
921 WritebackTmp: Memory used by FUSE for temporary writeback buffers
922 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
923 this is the total amount of memory currently available to
924 be allocated on the system. This limit is only adhered to
925 if strict overcommit accounting is enabled (mode 2 in
926 'vm.overcommit_memory').
927 The CommitLimit is calculated with the following formula:
928 CommitLimit = ([total RAM pages] - [total huge TLB pages]) *
929 overcommit_ratio / 100 + [total swap pages]
930 For example, on a system with 1G of physical RAM and 7G
931 of swap with a `vm.overcommit_ratio` of 30 it would
932 yield a CommitLimit of 7.3G.
933 For more details, see the memory overcommit documentation
934 in vm/overcommit-accounting.
935 Committed_AS: The amount of memory presently allocated on the system.
936 The committed memory is a sum of all of the memory which
937 has been allocated by processes, even if it has not been
938 "used" by them as of yet. A process which malloc()'s 1G
939 of memory, but only touches 300M of it will show up as
940 using 1G. This 1G is memory which has been "committed" to
941 by the VM and can be used at any time by the allocating
942 application. With strict overcommit enabled on the system
943 (mode 2 in 'vm.overcommit_memory'),allocations which would
944 exceed the CommitLimit (detailed above) will not be permitted.
945 This is useful if one needs to guarantee that processes will
946 not fail due to lack of memory once that memory has been
947 successfully allocated.
948 VmallocTotal: total size of vmalloc memory area
949 VmallocUsed: amount of vmalloc area which is used
950 VmallocChunk: largest contiguous block of vmalloc area which is free
952 ..............................................................................
956 Provides information about vmalloced/vmaped areas. One line per area,
957 containing the virtual address range of the area, size in bytes,
958 caller information of the creator, and optional information depending
959 on the kind of area :
961 pages=nr number of pages
962 phys=addr if a physical address was specified
963 ioremap I/O mapping (ioremap() and friends)
964 vmalloc vmalloc() area
967 vpages buffer for pages pointers was vmalloced (huge area)
968 N<node>=nr (Only on NUMA kernels)
969 Number of pages allocated on memory node <node>
971 > cat /proc/vmallocinfo
972 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ...
973 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128
974 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ...
975 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64
976 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f...
977 phys=7fee8000 ioremap
978 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f...
979 phys=7fee7000 ioremap
980 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210
981 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ...
982 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3
983 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ...
985 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ...
986 /0x130 [x_tables] pages=4 vmalloc N0=4
987 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ...
988 pages=14 vmalloc N2=14
989 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ...
991 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ...
993 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ...
994 pages=10 vmalloc N0=10
996 ..............................................................................
1000 Provides counts of softirq handlers serviced since boot time, for each cpu.
1002 > cat /proc/softirqs
1005 TIMER: 27166 27120 27097 27034
1010 SCHED: 27035 26983 26971 26746
1012 RCU: 1678 1769 2178 2250
1015 1.3 IDE devices in /proc/ide
1016 ----------------------------
1018 The subdirectory /proc/ide contains information about all IDE devices of which
1019 the kernel is aware. There is one subdirectory for each IDE controller, the
1020 file drivers and a link for each IDE device, pointing to the device directory
1021 in the controller specific subtree.
1023 The file drivers contains general information about the drivers used for the
1026 > cat /proc/ide/drivers
1027 ide-cdrom version 4.53
1028 ide-disk version 1.08
1030 More detailed information can be found in the controller specific
1031 subdirectories. These are named ide0, ide1 and so on. Each of these
1032 directories contains the files shown in table 1-6.
1035 Table 1-6: IDE controller info in /proc/ide/ide?
1036 ..............................................................................
1038 channel IDE channel (0 or 1)
1039 config Configuration (only for PCI/IDE bridge)
1041 model Type/Chipset of IDE controller
1042 ..............................................................................
1044 Each device connected to a controller has a separate subdirectory in the
1045 controllers directory. The files listed in table 1-7 are contained in these
1049 Table 1-7: IDE device information
1050 ..............................................................................
1053 capacity Capacity of the medium (in 512Byte blocks)
1054 driver driver and version
1055 geometry physical and logical geometry
1056 identify device identify block
1058 model device identifier
1059 settings device setup
1060 smart_thresholds IDE disk management thresholds
1061 smart_values IDE disk management values
1062 ..............................................................................
1064 The most interesting file is settings. This file contains a nice overview of
1065 the drive parameters:
1067 # cat /proc/ide/ide0/hda/settings
1068 name value min max mode
1069 ---- ----- --- --- ----
1070 bios_cyl 526 0 65535 rw
1071 bios_head 255 0 255 rw
1072 bios_sect 63 0 63 rw
1073 breada_readahead 4 0 127 rw
1075 file_readahead 72 0 2097151 rw
1077 keepsettings 0 0 1 rw
1078 max_kb_per_request 122 1 127 rw
1082 pio_mode write-only 0 255 w
1088 1.4 Networking info in /proc/net
1089 --------------------------------
1091 The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the
1092 additional values you get for IP version 6 if you configure the kernel to
1093 support this. Table 1-9 lists the files and their meaning.
1096 Table 1-8: IPv6 info in /proc/net
1097 ..............................................................................
1099 udp6 UDP sockets (IPv6)
1100 tcp6 TCP sockets (IPv6)
1101 raw6 Raw device statistics (IPv6)
1102 igmp6 IP multicast addresses, which this host joined (IPv6)
1103 if_inet6 List of IPv6 interface addresses
1104 ipv6_route Kernel routing table for IPv6
1105 rt6_stats Global IPv6 routing tables statistics
1106 sockstat6 Socket statistics (IPv6)
1107 snmp6 Snmp data (IPv6)
1108 ..............................................................................
1111 Table 1-9: Network info in /proc/net
1112 ..............................................................................
1114 arp Kernel ARP table
1115 dev network devices with statistics
1116 dev_mcast the Layer2 multicast groups a device is listening too
1117 (interface index, label, number of references, number of bound
1119 dev_stat network device status
1120 ip_fwchains Firewall chain linkage
1121 ip_fwnames Firewall chain names
1122 ip_masq Directory containing the masquerading tables
1123 ip_masquerade Major masquerading table
1124 netstat Network statistics
1125 raw raw device statistics
1126 route Kernel routing table
1127 rpc Directory containing rpc info
1128 rt_cache Routing cache
1130 sockstat Socket statistics
1133 unix UNIX domain sockets
1134 wireless Wireless interface data (Wavelan etc)
1135 igmp IP multicast addresses, which this host joined
1136 psched Global packet scheduler parameters.
1137 netlink List of PF_NETLINK sockets
1138 ip_mr_vifs List of multicast virtual interfaces
1139 ip_mr_cache List of multicast routing cache
1140 ..............................................................................
1142 You can use this information to see which network devices are available in
1143 your system and how much traffic was routed over those devices:
1146 Inter-|Receive |[...
1147 face |bytes packets errs drop fifo frame compressed multicast|[...
1148 lo: 908188 5596 0 0 0 0 0 0 [...
1149 ppp0:15475140 20721 410 0 0 410 0 0 [...
1150 eth0: 614530 7085 0 0 0 0 0 1 [...
1153 ...] bytes packets errs drop fifo colls carrier compressed
1154 ...] 908188 5596 0 0 0 0 0 0
1155 ...] 1375103 17405 0 0 0 0 0 0
1156 ...] 1703981 5535 0 0 0 3 0 0
1158 In addition, each Channel Bond interface has its own directory. For
1159 example, the bond0 device will have a directory called /proc/net/bond0/.
1160 It will contain information that is specific to that bond, such as the
1161 current slaves of the bond, the link status of the slaves, and how
1162 many times the slaves link has failed.
1167 If you have a SCSI host adapter in your system, you'll find a subdirectory
1168 named after the driver for this adapter in /proc/scsi. You'll also see a list
1169 of all recognized SCSI devices in /proc/scsi:
1171 >cat /proc/scsi/scsi
1173 Host: scsi0 Channel: 00 Id: 00 Lun: 00
1174 Vendor: IBM Model: DGHS09U Rev: 03E0
1175 Type: Direct-Access ANSI SCSI revision: 03
1176 Host: scsi0 Channel: 00 Id: 06 Lun: 00
1177 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
1178 Type: CD-ROM ANSI SCSI revision: 02
1181 The directory named after the driver has one file for each adapter found in
1182 the system. These files contain information about the controller, including
1183 the used IRQ and the IO address range. The amount of information shown is
1184 dependent on the adapter you use. The example shows the output for an Adaptec
1185 AHA-2940 SCSI adapter:
1187 > cat /proc/scsi/aic7xxx/0
1189 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
1191 TCQ Enabled By Default : Disabled
1192 AIC7XXX_PROC_STATS : Disabled
1193 AIC7XXX_RESET_DELAY : 5
1194 Adapter Configuration:
1195 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
1196 Ultra Wide Controller
1197 PCI MMAPed I/O Base: 0xeb001000
1198 Adapter SEEPROM Config: SEEPROM found and used.
1199 Adaptec SCSI BIOS: Enabled
1201 SCBs: Active 0, Max Active 2,
1202 Allocated 15, HW 16, Page 255
1204 BIOS Control Word: 0x18b6
1205 Adapter Control Word: 0x005b
1206 Extended Translation: Enabled
1207 Disconnect Enable Flags: 0xffff
1208 Ultra Enable Flags: 0x0001
1209 Tag Queue Enable Flags: 0x0000
1210 Ordered Queue Tag Flags: 0x0000
1211 Default Tag Queue Depth: 8
1212 Tagged Queue By Device array for aic7xxx host instance 0:
1213 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
1214 Actual queue depth per device for aic7xxx host instance 0:
1215 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
1218 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
1219 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
1220 Total transfers 160151 (74577 reads and 85574 writes)
1222 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
1223 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
1224 Total transfers 0 (0 reads and 0 writes)
1227 1.6 Parallel port info in /proc/parport
1228 ---------------------------------------
1230 The directory /proc/parport contains information about the parallel ports of
1231 your system. It has one subdirectory for each port, named after the port
1234 These directories contain the four files shown in Table 1-10.
1237 Table 1-10: Files in /proc/parport
1238 ..............................................................................
1240 autoprobe Any IEEE-1284 device ID information that has been acquired.
1241 devices list of the device drivers using that port. A + will appear by the
1242 name of the device currently using the port (it might not appear
1244 hardware Parallel port's base address, IRQ line and DMA channel.
1245 irq IRQ that parport is using for that port. This is in a separate
1246 file to allow you to alter it by writing a new value in (IRQ
1248 ..............................................................................
1250 1.7 TTY info in /proc/tty
1251 -------------------------
1253 Information about the available and actually used tty's can be found in the
1254 directory /proc/tty.You'll find entries for drivers and line disciplines in
1255 this directory, as shown in Table 1-11.
1258 Table 1-11: Files in /proc/tty
1259 ..............................................................................
1261 drivers list of drivers and their usage
1262 ldiscs registered line disciplines
1263 driver/serial usage statistic and status of single tty lines
1264 ..............................................................................
1266 To see which tty's are currently in use, you can simply look into the file
1269 > cat /proc/tty/drivers
1270 pty_slave /dev/pts 136 0-255 pty:slave
1271 pty_master /dev/ptm 128 0-255 pty:master
1272 pty_slave /dev/ttyp 3 0-255 pty:slave
1273 pty_master /dev/pty 2 0-255 pty:master
1274 serial /dev/cua 5 64-67 serial:callout
1275 serial /dev/ttyS 4 64-67 serial
1276 /dev/tty0 /dev/tty0 4 0 system:vtmaster
1277 /dev/ptmx /dev/ptmx 5 2 system
1278 /dev/console /dev/console 5 1 system:console
1279 /dev/tty /dev/tty 5 0 system:/dev/tty
1280 unknown /dev/tty 4 1-63 console
1283 1.8 Miscellaneous kernel statistics in /proc/stat
1284 -------------------------------------------------
1286 Various pieces of information about kernel activity are available in the
1287 /proc/stat file. All of the numbers reported in this file are aggregates
1288 since the system first booted. For a quick look, simply cat the file:
1291 cpu 2255 34 2290 22625563 6290 127 456 0 0 0
1292 cpu0 1132 34 1441 11311718 3675 127 438 0 0 0
1293 cpu1 1123 0 849 11313845 2614 0 18 0 0 0
1294 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
1300 softirq 183433 0 21755 12 39 1137 231 21459 2263
1302 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
1303 lines. These numbers identify the amount of time the CPU has spent performing
1304 different kinds of work. Time units are in USER_HZ (typically hundredths of a
1305 second). The meanings of the columns are as follows, from left to right:
1307 - user: normal processes executing in user mode
1308 - nice: niced processes executing in user mode
1309 - system: processes executing in kernel mode
1310 - idle: twiddling thumbs
1311 - iowait: waiting for I/O to complete
1312 - irq: servicing interrupts
1313 - softirq: servicing softirqs
1314 - steal: involuntary wait
1315 - guest: running a normal guest
1316 - guest_nice: running a niced guest
1318 The "intr" line gives counts of interrupts serviced since boot time, for each
1319 of the possible system interrupts. The first column is the total of all
1320 interrupts serviced including unnumbered architecture specific interrupts;
1321 each subsequent column is the total for that particular numbered interrupt.
1322 Unnumbered interrupts are not shown, only summed into the total.
1324 The "ctxt" line gives the total number of context switches across all CPUs.
1326 The "btime" line gives the time at which the system booted, in seconds since
1329 The "processes" line gives the number of processes and threads created, which
1330 includes (but is not limited to) those created by calls to the fork() and
1331 clone() system calls.
1333 The "procs_running" line gives the total number of threads that are
1334 running or ready to run (i.e., the total number of runnable threads).
1336 The "procs_blocked" line gives the number of processes currently blocked,
1337 waiting for I/O to complete.
1339 The "softirq" line gives counts of softirqs serviced since boot time, for each
1340 of the possible system softirqs. The first column is the total of all
1341 softirqs serviced; each subsequent column is the total for that particular
1345 1.9 Ext4 file system parameters
1346 -------------------------------
1348 Information about mounted ext4 file systems can be found in
1349 /proc/fs/ext4. Each mounted filesystem will have a directory in
1350 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
1351 /proc/fs/ext4/dm-0). The files in each per-device directory are shown
1352 in Table 1-12, below.
1354 Table 1-12: Files in /proc/fs/ext4/<devname>
1355 ..............................................................................
1357 mb_groups details of multiblock allocator buddy cache of free blocks
1358 ..............................................................................
1362 Shows registered system console lines.
1364 To see which character device lines are currently used for the system console
1365 /dev/console, you may simply look into the file /proc/consoles:
1367 > cat /proc/consoles
1373 device name of the device
1374 operations R = can do read operations
1375 W = can do write operations
1377 flags E = it is enabled
1378 C = it is preferred console
1379 B = it is primary boot console
1380 p = it is used for printk buffer
1381 b = it is not a TTY but a Braille device
1382 a = it is safe to use when cpu is offline
1383 major:minor major and minor number of the device separated by a colon
1385 ------------------------------------------------------------------------------
1387 ------------------------------------------------------------------------------
1388 The /proc file system serves information about the running system. It not only
1389 allows access to process data but also allows you to request the kernel status
1390 by reading files in the hierarchy.
1392 The directory structure of /proc reflects the types of information and makes
1393 it easy, if not obvious, where to look for specific data.
1394 ------------------------------------------------------------------------------
1396 ------------------------------------------------------------------------------
1397 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
1398 ------------------------------------------------------------------------------
1400 ------------------------------------------------------------------------------
1402 ------------------------------------------------------------------------------
1403 * Modifying kernel parameters by writing into files found in /proc/sys
1404 * Exploring the files which modify certain parameters
1405 * Review of the /proc/sys file tree
1406 ------------------------------------------------------------------------------
1409 A very interesting part of /proc is the directory /proc/sys. This is not only
1410 a source of information, it also allows you to change parameters within the
1411 kernel. Be very careful when attempting this. You can optimize your system,
1412 but you can also cause it to crash. Never alter kernel parameters on a
1413 production system. Set up a development machine and test to make sure that
1414 everything works the way you want it to. You may have no alternative but to
1415 reboot the machine once an error has been made.
1417 To change a value, simply echo the new value into the file. An example is
1418 given below in the section on the file system data. You need to be root to do
1419 this. You can create your own boot script to perform this every time your
1422 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
1423 general things in the operation of the Linux kernel. Since some of the files
1424 can inadvertently disrupt your system, it is advisable to read both
1425 documentation and source before actually making adjustments. In any case, be
1426 very careful when writing to any of these files. The entries in /proc may
1427 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
1428 review the kernel documentation in the directory /usr/src/linux/Documentation.
1429 This chapter is heavily based on the documentation included in the pre 2.2
1430 kernels, and became part of it in version 2.2.1 of the Linux kernel.
1432 Please see: Documentation/sysctl/ directory for descriptions of these
1435 ------------------------------------------------------------------------------
1437 ------------------------------------------------------------------------------
1438 Certain aspects of kernel behavior can be modified at runtime, without the
1439 need to recompile the kernel, or even to reboot the system. The files in the
1440 /proc/sys tree can not only be read, but also modified. You can use the echo
1441 command to write value into these files, thereby changing the default settings
1443 ------------------------------------------------------------------------------
1445 ------------------------------------------------------------------------------
1446 CHAPTER 3: PER-PROCESS PARAMETERS
1447 ------------------------------------------------------------------------------
1449 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score
1450 --------------------------------------------------------------------------------
1452 These file can be used to adjust the badness heuristic used to select which
1453 process gets killed in out of memory conditions.
1455 The badness heuristic assigns a value to each candidate task ranging from 0
1456 (never kill) to 1000 (always kill) to determine which process is targeted. The
1457 units are roughly a proportion along that range of allowed memory the process
1458 may allocate from based on an estimation of its current memory and swap use.
1459 For example, if a task is using all allowed memory, its badness score will be
1460 1000. If it is using half of its allowed memory, its score will be 500.
1462 There is an additional factor included in the badness score: the current memory
1463 and swap usage is discounted by 3% for root processes.
1465 The amount of "allowed" memory depends on the context in which the oom killer
1466 was called. If it is due to the memory assigned to the allocating task's cpuset
1467 being exhausted, the allowed memory represents the set of mems assigned to that
1468 cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed
1469 memory represents the set of mempolicy nodes. If it is due to a memory
1470 limit (or swap limit) being reached, the allowed memory is that configured
1471 limit. Finally, if it is due to the entire system being out of memory, the
1472 allowed memory represents all allocatable resources.
1474 The value of /proc/<pid>/oom_score_adj is added to the badness score before it
1475 is used to determine which task to kill. Acceptable values range from -1000
1476 (OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to
1477 polarize the preference for oom killing either by always preferring a certain
1478 task or completely disabling it. The lowest possible value, -1000, is
1479 equivalent to disabling oom killing entirely for that task since it will always
1480 report a badness score of 0.
1482 Consequently, it is very simple for userspace to define the amount of memory to
1483 consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for
1484 example, is roughly equivalent to allowing the remainder of tasks sharing the
1485 same system, cpuset, mempolicy, or memory controller resources to use at least
1486 50% more memory. A value of -500, on the other hand, would be roughly
1487 equivalent to discounting 50% of the task's allowed memory from being considered
1488 as scoring against the task.
1490 For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also
1491 be used to tune the badness score. Its acceptable values range from -16
1492 (OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17
1493 (OOM_DISABLE) to disable oom killing entirely for that task. Its value is
1494 scaled linearly with /proc/<pid>/oom_score_adj.
1496 The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last
1497 value set by a CAP_SYS_RESOURCE process. To reduce the value any lower
1498 requires CAP_SYS_RESOURCE.
1500 Caveat: when a parent task is selected, the oom killer will sacrifice any first
1501 generation children with separate address spaces instead, if possible. This
1502 avoids servers and important system daemons from being killed and loses the
1503 minimal amount of work.
1506 3.2 /proc/<pid>/oom_score - Display current oom-killer score
1507 -------------------------------------------------------------
1509 This file can be used to check the current score used by the oom-killer is for
1510 any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which
1511 process should be killed in an out-of-memory situation.
1514 3.3 /proc/<pid>/io - Display the IO accounting fields
1515 -------------------------------------------------------
1517 This file contains IO statistics for each running process
1522 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
1525 test:/tmp # cat /proc/3828/io
1531 write_bytes: 323932160
1532 cancelled_write_bytes: 0
1541 I/O counter: chars read
1542 The number of bytes which this task has caused to be read from storage. This
1543 is simply the sum of bytes which this process passed to read() and pread().
1544 It includes things like tty IO and it is unaffected by whether or not actual
1545 physical disk IO was required (the read might have been satisfied from
1552 I/O counter: chars written
1553 The number of bytes which this task has caused, or shall cause to be written
1554 to disk. Similar caveats apply here as with rchar.
1560 I/O counter: read syscalls
1561 Attempt to count the number of read I/O operations, i.e. syscalls like read()
1568 I/O counter: write syscalls
1569 Attempt to count the number of write I/O operations, i.e. syscalls like
1570 write() and pwrite().
1576 I/O counter: bytes read
1577 Attempt to count the number of bytes which this process really did cause to
1578 be fetched from the storage layer. Done at the submit_bio() level, so it is
1579 accurate for block-backed filesystems. <please add status regarding NFS and
1580 CIFS at a later time>
1586 I/O counter: bytes written
1587 Attempt to count the number of bytes which this process caused to be sent to
1588 the storage layer. This is done at page-dirtying time.
1591 cancelled_write_bytes
1592 ---------------------
1594 The big inaccuracy here is truncate. If a process writes 1MB to a file and
1595 then deletes the file, it will in fact perform no writeout. But it will have
1596 been accounted as having caused 1MB of write.
1597 In other words: The number of bytes which this process caused to not happen,
1598 by truncating pagecache. A task can cause "negative" IO too. If this task
1599 truncates some dirty pagecache, some IO which another task has been accounted
1600 for (in its write_bytes) will not be happening. We _could_ just subtract that
1601 from the truncating task's write_bytes, but there is information loss in doing
1608 At its current implementation state, this is a bit racy on 32-bit machines: if
1609 process A reads process B's /proc/pid/io while process B is updating one of
1610 those 64-bit counters, process A could see an intermediate result.
1613 More information about this can be found within the taskstats documentation in
1614 Documentation/accounting.
1616 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings
1617 ---------------------------------------------------------------
1618 When a process is dumped, all anonymous memory is written to a core file as
1619 long as the size of the core file isn't limited. But sometimes we don't want
1620 to dump some memory segments, for example, huge shared memory or DAX.
1621 Conversely, sometimes we want to save file-backed memory segments into a core
1622 file, not only the individual files.
1624 /proc/<pid>/coredump_filter allows you to customize which memory segments
1625 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
1626 of memory types. If a bit of the bitmask is set, memory segments of the
1627 corresponding memory type are dumped, otherwise they are not dumped.
1629 The following 9 memory types are supported:
1630 - (bit 0) anonymous private memory
1631 - (bit 1) anonymous shared memory
1632 - (bit 2) file-backed private memory
1633 - (bit 3) file-backed shared memory
1634 - (bit 4) ELF header pages in file-backed private memory areas (it is
1635 effective only if the bit 2 is cleared)
1636 - (bit 5) hugetlb private memory
1637 - (bit 6) hugetlb shared memory
1638 - (bit 7) DAX private memory
1639 - (bit 8) DAX shared memory
1641 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
1642 are always dumped regardless of the bitmask status.
1644 Note that bits 0-4 don't affect hugetlb or DAX memory. hugetlb memory is
1645 only affected by bit 5-6, and DAX is only affected by bits 7-8.
1647 The default value of coredump_filter is 0x33; this means all anonymous memory
1648 segments, ELF header pages and hugetlb private memory are dumped.
1650 If you don't want to dump all shared memory segments attached to pid 1234,
1651 write 0x31 to the process's proc file.
1653 $ echo 0x31 > /proc/1234/coredump_filter
1655 When a new process is created, the process inherits the bitmask status from its
1656 parent. It is useful to set up coredump_filter before the program runs.
1659 $ echo 0x7 > /proc/self/coredump_filter
1662 3.5 /proc/<pid>/mountinfo - Information about mounts
1663 --------------------------------------------------------
1665 This file contains lines of the form:
1667 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
1668 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
1670 (1) mount ID: unique identifier of the mount (may be reused after umount)
1671 (2) parent ID: ID of parent (or of self for the top of the mount tree)
1672 (3) major:minor: value of st_dev for files on filesystem
1673 (4) root: root of the mount within the filesystem
1674 (5) mount point: mount point relative to the process's root
1675 (6) mount options: per mount options
1676 (7) optional fields: zero or more fields of the form "tag[:value]"
1677 (8) separator: marks the end of the optional fields
1678 (9) filesystem type: name of filesystem of the form "type[.subtype]"
1679 (10) mount source: filesystem specific information or "none"
1680 (11) super options: per super block options
1682 Parsers should ignore all unrecognised optional fields. Currently the
1683 possible optional fields are:
1685 shared:X mount is shared in peer group X
1686 master:X mount is slave to peer group X
1687 propagate_from:X mount is slave and receives propagation from peer group X (*)
1688 unbindable mount is unbindable
1690 (*) X is the closest dominant peer group under the process's root. If
1691 X is the immediate master of the mount, or if there's no dominant peer
1692 group under the same root, then only the "master:X" field is present
1693 and not the "propagate_from:X" field.
1695 For more information on mount propagation see:
1697 Documentation/filesystems/sharedsubtree.txt
1700 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm
1701 --------------------------------------------------------
1702 These files provide a method to access a tasks comm value. It also allows for
1703 a task to set its own or one of its thread siblings comm value. The comm value
1704 is limited in size compared to the cmdline value, so writing anything longer
1705 then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated
1709 3.7 /proc/<pid>/task/<tid>/children - Information about task children
1710 -------------------------------------------------------------------------
1711 This file provides a fast way to retrieve first level children pids
1712 of a task pointed by <pid>/<tid> pair. The format is a space separated
1715 Note the "first level" here -- if a child has own children they will
1716 not be listed here, one needs to read /proc/<children-pid>/task/<tid>/children
1717 to obtain the descendants.
1719 Since this interface is intended to be fast and cheap it doesn't
1720 guarantee to provide precise results and some children might be
1721 skipped, especially if they've exited right after we printed their
1722 pids, so one need to either stop or freeze processes being inspected
1723 if precise results are needed.
1726 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file
1727 ---------------------------------------------------------------
1728 This file provides information associated with an opened file. The regular
1729 files have at least three fields -- 'pos', 'flags' and mnt_id. The 'pos'
1730 represents the current offset of the opened file in decimal form [see lseek(2)
1731 for details], 'flags' denotes the octal O_xxx mask the file has been
1732 created with [see open(2) for details] and 'mnt_id' represents mount ID of
1733 the file system containing the opened file [see 3.5 /proc/<pid>/mountinfo
1742 All locks associated with a file descriptor are shown in its fdinfo too.
1744 lock: 1: FLOCK ADVISORY WRITE 359 00:13:11691 0 EOF
1746 The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags
1747 pair provide additional information particular to the objects they represent.
1756 where 'eventfd-count' is hex value of a counter.
1763 sigmask: 0000000000000200
1765 where 'sigmask' is hex value of the signal mask associated
1773 tfd: 5 events: 1d data: ffffffffffffffff
1775 where 'tfd' is a target file descriptor number in decimal form,
1776 'events' is events mask being watched and the 'data' is data
1777 associated with a target [see epoll(7) for more details].
1781 For inotify files the format is the following
1785 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d
1787 where 'wd' is a watch descriptor in decimal form, ie a target file
1788 descriptor number, 'ino' and 'sdev' are inode and device where the
1789 target file resides and the 'mask' is the mask of events, all in hex
1790 form [see inotify(7) for more details].
1792 If the kernel was built with exportfs support, the path to the target
1793 file is encoded as a file handle. The file handle is provided by three
1794 fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex
1797 If the kernel is built without exportfs support the file handle won't be
1800 If there is no inotify mark attached yet the 'inotify' line will be omitted.
1802 For fanotify files the format is
1807 fanotify flags:10 event-flags:0
1808 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003
1809 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4
1811 where fanotify 'flags' and 'event-flags' are values used in fanotify_init
1812 call, 'mnt_id' is the mount point identifier, 'mflags' is the value of
1813 flags associated with mark which are tracked separately from events
1814 mask. 'ino', 'sdev' are target inode and device, 'mask' is the events
1815 mask and 'ignored_mask' is the mask of events which are to be ignored.
1816 All in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask'
1817 does provide information about flags and mask used in fanotify_mark
1818 call [see fsnotify manpage for details].
1820 While the first three lines are mandatory and always printed, the rest is
1821 optional and may be omitted if no marks created yet.
1832 it_value: (0, 49406829)
1835 where 'clockid' is the clock type and 'ticks' is the number of the timer expirations
1836 that have occurred [see timerfd_create(2) for details]. 'settime flags' are
1837 flags in octal form been used to setup the timer [see timerfd_settime(2) for
1838 details]. 'it_value' is remaining time until the timer exiration.
1839 'it_interval' is the interval for the timer. Note the timer might be set up
1840 with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value'
1841 still exhibits timer's remaining time.
1843 3.9 /proc/<pid>/map_files - Information about memory mapped files
1844 ---------------------------------------------------------------------
1845 This directory contains symbolic links which represent memory mapped files
1846 the process is maintaining. Example output:
1848 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so
1849 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so
1850 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so
1852 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1
1853 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls
1855 The name of a link represents the virtual memory bounds of a mapping, i.e.
1856 vm_area_struct::vm_start-vm_area_struct::vm_end.
1858 The main purpose of the map_files is to retrieve a set of memory mapped
1859 files in a fast way instead of parsing /proc/<pid>/maps or
1860 /proc/<pid>/smaps, both of which contain many more records. At the same
1861 time one can open(2) mappings from the listings of two processes and
1862 comparing their inode numbers to figure out which anonymous memory areas
1863 are actually shared.
1865 ------------------------------------------------------------------------------
1867 ------------------------------------------------------------------------------
1870 ---------------------
1872 The following mount options are supported:
1874 hidepid= Set /proc/<pid>/ access mode.
1875 gid= Set the group authorized to learn processes information.
1877 hidepid=0 means classic mode - everybody may access all /proc/<pid>/ directories
1880 hidepid=1 means users may not access any /proc/<pid>/ directories but their
1881 own. Sensitive files like cmdline, sched*, status are now protected against
1882 other users. This makes it impossible to learn whether any user runs
1883 specific program (given the program doesn't reveal itself by its behaviour).
1884 As an additional bonus, as /proc/<pid>/cmdline is unaccessible for other users,
1885 poorly written programs passing sensitive information via program arguments are
1886 now protected against local eavesdroppers.
1888 hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be fully invisible to other
1889 users. It doesn't mean that it hides a fact whether a process with a specific
1890 pid value exists (it can be learned by other means, e.g. by "kill -0 $PID"),
1891 but it hides process' uid and gid, which may be learned by stat()'ing
1892 /proc/<pid>/ otherwise. It greatly complicates an intruder's task of gathering
1893 information about running processes, whether some daemon runs with elevated
1894 privileges, whether other user runs some sensitive program, whether other users
1895 run any program at all, etc.
1897 gid= defines a group authorized to learn processes information otherwise
1898 prohibited by hidepid=. If you use some daemon like identd which needs to learn
1899 information about processes information, just add identd to this group.
projects / deliverable / linux.git / blob