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 If CONFIG_KALLSYMS is set, a pre-decoded wchan
145 stack Report full stack trace, enable via CONFIG_STACKTRACE
146 smaps a extension based on maps, showing the memory consumption of
147 each mapping and flags associated with it
148 ..............................................................................
150 For example, to get the status information of a process, all you have to do is
151 read the file /proc/PID/status:
153 >cat /proc/self/status
177 SigPnd: 0000000000000000
178 ShdPnd: 0000000000000000
179 SigBlk: 0000000000000000
180 SigIgn: 0000000000000000
181 SigCgt: 0000000000000000
182 CapInh: 00000000fffffeff
183 CapPrm: 0000000000000000
184 CapEff: 0000000000000000
185 CapBnd: ffffffffffffffff
187 voluntary_ctxt_switches: 0
188 nonvoluntary_ctxt_switches: 1
190 This shows you nearly the same information you would get if you viewed it with
191 the ps command. In fact, ps uses the proc file system to obtain its
192 information. But you get a more detailed view of the process by reading the
193 file /proc/PID/status. It fields are described in table 1-2.
195 The statm file contains more detailed information about the process
196 memory usage. Its seven fields are explained in Table 1-3. The stat file
197 contains details information about the process itself. Its fields are
198 explained in Table 1-4.
200 (for SMP CONFIG users)
201 For making accounting scalable, RSS related information are handled in
202 asynchronous manner and the vaule may not be very precise. To see a precise
203 snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table.
204 It's slow but very precise.
206 Table 1-2: Contents of the status files (as of 2.6.30-rc7)
207 ..............................................................................
209 Name filename of the executable
210 State state (R is running, S is sleeping, D is sleeping
211 in an uninterruptible wait, Z is zombie,
212 T is traced or stopped)
215 PPid process id of the parent process
216 TracerPid PID of process tracing this process (0 if not)
217 Uid Real, effective, saved set, and file system UIDs
218 Gid Real, effective, saved set, and file system GIDs
219 FDSize number of file descriptor slots currently allocated
220 Groups supplementary group list
221 VmPeak peak virtual memory size
222 VmSize total program size
223 VmLck locked memory size
224 VmHWM peak resident set size ("high water mark")
225 VmRSS size of memory portions
226 VmData size of data, stack, and text segments
227 VmStk size of data, stack, and text segments
228 VmExe size of text segment
229 VmLib size of shared library code
230 VmPTE size of page table entries
231 VmSwap size of swap usage (the number of referred swapents)
232 Threads number of threads
233 SigQ number of signals queued/max. number for queue
234 SigPnd bitmap of pending signals for the thread
235 ShdPnd bitmap of shared pending signals for the process
236 SigBlk bitmap of blocked signals
237 SigIgn bitmap of ignored signals
238 SigCgt bitmap of caught signals
239 CapInh bitmap of inheritable capabilities
240 CapPrm bitmap of permitted capabilities
241 CapEff bitmap of effective capabilities
242 CapBnd bitmap of capabilities bounding set
243 Seccomp seccomp mode, like prctl(PR_GET_SECCOMP, ...)
244 Cpus_allowed mask of CPUs on which this process may run
245 Cpus_allowed_list Same as previous, but in "list format"
246 Mems_allowed mask of memory nodes allowed to this process
247 Mems_allowed_list Same as previous, but in "list format"
248 voluntary_ctxt_switches number of voluntary context switches
249 nonvoluntary_ctxt_switches number of non voluntary context switches
250 ..............................................................................
252 Table 1-3: Contents of the statm files (as of 2.6.8-rc3)
253 ..............................................................................
255 size total program size (pages) (same as VmSize in status)
256 resident size of memory portions (pages) (same as VmRSS in status)
257 shared number of pages that are shared (i.e. backed by a file)
258 trs number of pages that are 'code' (not including libs; broken,
259 includes data segment)
260 lrs number of pages of library (always 0 on 2.6)
261 drs number of pages of data/stack (including libs; broken,
262 includes library text)
263 dt number of dirty pages (always 0 on 2.6)
264 ..............................................................................
267 Table 1-4: Contents of the stat files (as of 2.6.30-rc7)
268 ..............................................................................
271 tcomm filename of the executable
272 state state (R is running, S is sleeping, D is sleeping in an
273 uninterruptible wait, Z is zombie, T is traced or stopped)
274 ppid process id of the parent process
275 pgrp pgrp of the process
277 tty_nr tty the process uses
278 tty_pgrp pgrp of the tty
280 min_flt number of minor faults
281 cmin_flt number of minor faults with child's
282 maj_flt number of major faults
283 cmaj_flt number of major faults with child's
284 utime user mode jiffies
285 stime kernel mode jiffies
286 cutime user mode jiffies with child's
287 cstime kernel mode jiffies with child's
288 priority priority level
290 num_threads number of threads
291 it_real_value (obsolete, always 0)
292 start_time time the process started after system boot
293 vsize virtual memory size
294 rss resident set memory size
295 rsslim current limit in bytes on the rss
296 start_code address above which program text can run
297 end_code address below which program text can run
298 start_stack address of the start of the main process stack
299 esp current value of ESP
300 eip current value of EIP
301 pending bitmap of pending signals
302 blocked bitmap of blocked signals
303 sigign bitmap of ignored signals
304 sigcatch bitmap of caught signals
305 wchan address where process went to sleep
308 exit_signal signal to send to parent thread on exit
309 task_cpu which CPU the task is scheduled on
310 rt_priority realtime priority
311 policy scheduling policy (man sched_setscheduler)
312 blkio_ticks time spent waiting for block IO
313 gtime guest time of the task in jiffies
314 cgtime guest time of the task children in jiffies
315 start_data address above which program data+bss is placed
316 end_data address below which program data+bss is placed
317 start_brk address above which program heap can be expanded with brk()
318 arg_start address above which program command line is placed
319 arg_end address below which program command line is placed
320 env_start address above which program environment is placed
321 env_end address below which program environment is placed
322 exit_code the thread's exit_code in the form reported by the waitpid system call
323 ..............................................................................
325 The /proc/PID/maps file containing the currently mapped memory regions and
326 their access permissions.
330 address perms offset dev inode pathname
332 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
333 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
334 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
335 a7cb1000-a7cb2000 ---p 00000000 00:00 0
336 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
337 a7eb2000-a7eb3000 ---p 00000000 00:00 0
338 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 [stack:1001]
339 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
340 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
341 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
342 a800b000-a800e000 rw-p 00000000 00:00 0
343 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
344 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
345 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
346 a8024000-a8027000 rw-p 00000000 00:00 0
347 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
348 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
349 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
350 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack]
351 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
353 where "address" is the address space in the process that it occupies, "perms"
354 is a set of permissions:
360 p = private (copy on write)
362 "offset" is the offset into the mapping, "dev" is the device (major:minor), and
363 "inode" is the inode on that device. 0 indicates that no inode is associated
364 with the memory region, as the case would be with BSS (uninitialized data).
365 The "pathname" shows the name associated file for this mapping. If the mapping
366 is not associated with a file:
368 [heap] = the heap of the program
369 [stack] = the stack of the main process
370 [stack:1001] = the stack of the thread with tid 1001
371 [vdso] = the "virtual dynamic shared object",
372 the kernel system call handler
374 or if empty, the mapping is anonymous.
376 The /proc/PID/task/TID/maps is a view of the virtual memory from the viewpoint
377 of the individual tasks of a process. In this file you will see a mapping marked
378 as [stack] if that task sees it as a stack. This is a key difference from the
379 content of /proc/PID/maps, where you will see all mappings that are being used
380 as stack by all of those tasks. Hence, for the example above, the task-level
381 map, i.e. /proc/PID/task/TID/maps for thread 1001 will look like this:
383 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
384 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
385 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
386 a7cb1000-a7cb2000 ---p 00000000 00:00 0
387 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
388 a7eb2000-a7eb3000 ---p 00000000 00:00 0
389 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 [stack]
390 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
391 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
392 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
393 a800b000-a800e000 rw-p 00000000 00:00 0
394 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
395 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
396 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
397 a8024000-a8027000 rw-p 00000000 00:00 0
398 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
399 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
400 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
401 aff35000-aff4a000 rw-p 00000000 00:00 0
402 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
404 The /proc/PID/smaps is an extension based on maps, showing the memory
405 consumption for each of the process's mappings. For each of mappings there
406 is a series of lines such as the following:
408 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash
422 VmFlags: rd ex mr mw me de
424 the first of these lines shows the same information as is displayed for the
425 mapping in /proc/PID/maps. The remaining lines show the size of the mapping
426 (size), the amount of the mapping that is currently resident in RAM (RSS), the
427 process' proportional share of this mapping (PSS), the number of clean and
428 dirty private pages in the mapping. Note that even a page which is part of a
429 MAP_SHARED mapping, but has only a single pte mapped, i.e. is currently used
430 by only one process, is accounted as private and not as shared. "Referenced"
431 indicates the amount of memory currently marked as referenced or accessed.
432 "Anonymous" shows the amount of memory that does not belong to any file. Even
433 a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE
434 and a page is modified, the file page is replaced by a private anonymous copy.
435 "Swap" shows how much would-be-anonymous memory is also used, but out on
438 "VmFlags" field deserves a separate description. This member represents the kernel
439 flags associated with the particular virtual memory area in two letter encoded
440 manner. The codes are the following:
449 gd - stack segment growns down
451 dw - disabled write to the mapped file
452 lo - pages are locked in memory
453 io - memory mapped I/O area
454 sr - sequential read advise provided
455 rr - random read advise provided
456 dc - do not copy area on fork
457 de - do not expand area on remapping
458 ac - area is accountable
459 nr - swap space is not reserved for the area
460 ht - area uses huge tlb pages
461 nl - non-linear mapping
462 ar - architecture specific flag
463 dd - do not include area into core dump
466 hg - huge page advise flag
467 nh - no-huge page advise flag
468 mg - mergable advise flag
470 Note that there is no guarantee that every flag and associated mnemonic will
471 be present in all further kernel releases. Things get changed, the flags may
472 be vanished or the reverse -- new added.
474 This file is only present if the CONFIG_MMU kernel configuration option is
477 The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG
478 bits on both physical and virtual pages associated with a process, and the
479 soft-dirty bit on pte (see Documentation/vm/soft-dirty.txt for details).
480 To clear the bits for all the pages associated with the process
481 > echo 1 > /proc/PID/clear_refs
483 To clear the bits for the anonymous pages associated with the process
484 > echo 2 > /proc/PID/clear_refs
486 To clear the bits for the file mapped pages associated with the process
487 > echo 3 > /proc/PID/clear_refs
489 To clear the soft-dirty bit
490 > echo 4 > /proc/PID/clear_refs
492 Any other value written to /proc/PID/clear_refs will have no effect.
494 The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags
495 using /proc/kpageflags and number of times a page is mapped using
496 /proc/kpagecount. For detailed explanation, see Documentation/vm/pagemap.txt.
501 Similar to the process entries, the kernel data files give information about
502 the running kernel. The files used to obtain this information are contained in
503 /proc and are listed in Table 1-5. Not all of these will be present in your
504 system. It depends on the kernel configuration and the loaded modules, which
505 files are there, and which are missing.
507 Table 1-5: Kernel info in /proc
508 ..............................................................................
510 apm Advanced power management info
511 buddyinfo Kernel memory allocator information (see text) (2.5)
512 bus Directory containing bus specific information
513 cmdline Kernel command line
514 cpuinfo Info about the CPU
515 devices Available devices (block and character)
516 dma Used DMS channels
517 filesystems Supported filesystems
518 driver Various drivers grouped here, currently rtc (2.4)
519 execdomains Execdomains, related to security (2.4)
520 fb Frame Buffer devices (2.4)
521 fs File system parameters, currently nfs/exports (2.4)
522 ide Directory containing info about the IDE subsystem
523 interrupts Interrupt usage
524 iomem Memory map (2.4)
525 ioports I/O port usage
526 irq Masks for irq to cpu affinity (2.4)(smp?)
527 isapnp ISA PnP (Plug&Play) Info (2.4)
528 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
530 ksyms Kernel symbol table
531 loadavg Load average of last 1, 5 & 15 minutes
535 modules List of loaded modules
536 mounts Mounted filesystems
537 net Networking info (see text)
538 pagetypeinfo Additional page allocator information (see text) (2.5)
539 partitions Table of partitions known to the system
540 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
541 decoupled by lspci (2.4)
543 scsi SCSI info (see text)
544 slabinfo Slab pool info
545 softirqs softirq usage
546 stat Overall statistics
547 swaps Swap space utilization
549 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
550 tty Info of tty drivers
551 uptime Wall clock since boot, combined idle time of all cpus
552 version Kernel version
553 video bttv info of video resources (2.4)
554 vmallocinfo Show vmalloced areas
555 ..............................................................................
557 You can, for example, check which interrupts are currently in use and what
558 they are used for by looking in the file /proc/interrupts:
560 > cat /proc/interrupts
562 0: 8728810 XT-PIC timer
563 1: 895 XT-PIC keyboard
565 3: 531695 XT-PIC aha152x
566 4: 2014133 XT-PIC serial
567 5: 44401 XT-PIC pcnet_cs
570 12: 182918 XT-PIC PS/2 Mouse
572 14: 1232265 XT-PIC ide0
576 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
577 output of a SMP machine):
579 > cat /proc/interrupts
582 0: 1243498 1214548 IO-APIC-edge timer
583 1: 8949 8958 IO-APIC-edge keyboard
584 2: 0 0 XT-PIC cascade
585 5: 11286 10161 IO-APIC-edge soundblaster
586 8: 1 0 IO-APIC-edge rtc
587 9: 27422 27407 IO-APIC-edge 3c503
588 12: 113645 113873 IO-APIC-edge PS/2 Mouse
590 14: 22491 24012 IO-APIC-edge ide0
591 15: 2183 2415 IO-APIC-edge ide1
592 17: 30564 30414 IO-APIC-level eth0
593 18: 177 164 IO-APIC-level bttv
598 NMI is incremented in this case because every timer interrupt generates a NMI
599 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
601 LOC is the local interrupt counter of the internal APIC of every CPU.
603 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
604 connects the CPUs in a SMP system. This means that an error has been detected,
605 the IO-APIC automatically retry the transmission, so it should not be a big
606 problem, but you should read the SMP-FAQ.
608 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
609 /proc/interrupts to display every IRQ vector in use by the system, not
610 just those considered 'most important'. The new vectors are:
612 THR -- interrupt raised when a machine check threshold counter
613 (typically counting ECC corrected errors of memory or cache) exceeds
614 a configurable threshold. Only available on some systems.
616 TRM -- a thermal event interrupt occurs when a temperature threshold
617 has been exceeded for the CPU. This interrupt may also be generated
618 when the temperature drops back to normal.
620 SPU -- a spurious interrupt is some interrupt that was raised then lowered
621 by some IO device before it could be fully processed by the APIC. Hence
622 the APIC sees the interrupt but does not know what device it came from.
623 For this case the APIC will generate the interrupt with a IRQ vector
624 of 0xff. This might also be generated by chipset bugs.
626 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
627 sent from one CPU to another per the needs of the OS. Typically,
628 their statistics are used by kernel developers and interested users to
629 determine the occurrence of interrupts of the given type.
631 The above IRQ vectors are displayed only when relevant. For example,
632 the threshold vector does not exist on x86_64 platforms. Others are
633 suppressed when the system is a uniprocessor. As of this writing, only
634 i386 and x86_64 platforms support the new IRQ vector displays.
636 Of some interest is the introduction of the /proc/irq directory to 2.4.
637 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
638 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
639 irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and
644 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
645 1 11 13 15 17 19 3 5 7 9 default_smp_affinity
649 smp_affinity is a bitmask, in which you can specify which CPUs can handle the
650 IRQ, you can set it by doing:
652 > echo 1 > /proc/irq/10/smp_affinity
654 This means that only the first CPU will handle the IRQ, but you can also echo
655 5 which means that only the first and fourth CPU can handle the IRQ.
657 The contents of each smp_affinity file is the same by default:
659 > cat /proc/irq/0/smp_affinity
662 There is an alternate interface, smp_affinity_list which allows specifying
663 a cpu range instead of a bitmask:
665 > cat /proc/irq/0/smp_affinity_list
668 The default_smp_affinity mask applies to all non-active IRQs, which are the
669 IRQs which have not yet been allocated/activated, and hence which lack a
670 /proc/irq/[0-9]* directory.
672 The node file on an SMP system shows the node to which the device using the IRQ
673 reports itself as being attached. This hardware locality information does not
674 include information about any possible driver locality preference.
676 prof_cpu_mask specifies which CPUs are to be profiled by the system wide
677 profiler. Default value is ffffffff (all cpus if there are only 32 of them).
679 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
680 between all the CPUs which are allowed to handle it. As usual the kernel has
681 more info than you and does a better job than you, so the defaults are the
682 best choice for almost everyone. [Note this applies only to those IO-APIC's
683 that support "Round Robin" interrupt distribution.]
685 There are three more important subdirectories in /proc: net, scsi, and sys.
686 The general rule is that the contents, or even the existence of these
687 directories, depend on your kernel configuration. If SCSI is not enabled, the
688 directory scsi may not exist. The same is true with the net, which is there
689 only when networking support is present in the running kernel.
691 The slabinfo file gives information about memory usage at the slab level.
692 Linux uses slab pools for memory management above page level in version 2.2.
693 Commonly used objects have their own slab pool (such as network buffers,
694 directory cache, and so on).
696 ..............................................................................
698 > cat /proc/buddyinfo
700 Node 0, zone DMA 0 4 5 4 4 3 ...
701 Node 0, zone Normal 1 0 0 1 101 8 ...
702 Node 0, zone HighMem 2 0 0 1 1 0 ...
704 External fragmentation is a problem under some workloads, and buddyinfo is a
705 useful tool for helping diagnose these problems. Buddyinfo will give you a
706 clue as to how big an area you can safely allocate, or why a previous
709 Each column represents the number of pages of a certain order which are
710 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
711 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
712 available in ZONE_NORMAL, etc...
714 More information relevant to external fragmentation can be found in
717 > cat /proc/pagetypeinfo
721 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
722 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0
723 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0
724 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2
725 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0
726 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0
727 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9
728 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0
729 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452
730 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0
731 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0
733 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate
734 Node 0, zone DMA 2 0 5 1 0
735 Node 0, zone DMA32 41 6 967 2 0
737 Fragmentation avoidance in the kernel works by grouping pages of different
738 migrate types into the same contiguous regions of memory called page blocks.
739 A page block is typically the size of the default hugepage size e.g. 2MB on
740 X86-64. By keeping pages grouped based on their ability to move, the kernel
741 can reclaim pages within a page block to satisfy a high-order allocation.
743 The pagetypinfo begins with information on the size of a page block. It
744 then gives the same type of information as buddyinfo except broken down
745 by migrate-type and finishes with details on how many page blocks of each
748 If min_free_kbytes has been tuned correctly (recommendations made by hugeadm
749 from libhugetlbfs http://sourceforge.net/projects/libhugetlbfs/), one can
750 make an estimate of the likely number of huge pages that can be allocated
751 at a given point in time. All the "Movable" blocks should be allocatable
752 unless memory has been mlock()'d. Some of the Reclaimable blocks should
753 also be allocatable although a lot of filesystem metadata may have to be
754 reclaimed to achieve this.
756 ..............................................................................
760 Provides information about distribution and utilization of memory. This
761 varies by architecture and compile options. The following is from a
762 16GB PIII, which has highmem enabled. You may not have all of these fields.
766 The "Locked" indicates whether the mapping is locked in memory or not.
769 MemTotal: 16344972 kB
771 MemAvailable: 14836172 kB
777 HighTotal: 15597528 kB
778 HighFree: 13629632 kB
788 SReclaimable: 159856 kB
789 SUnreclaim: 124508 kB
794 CommitLimit: 7669796 kB
795 Committed_AS: 100056 kB
796 VmallocTotal: 112216 kB
798 VmallocChunk: 111088 kB
799 AnonHugePages: 49152 kB
801 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
802 bits and the kernel binary code)
803 MemFree: The sum of LowFree+HighFree
804 MemAvailable: An estimate of how much memory is available for starting new
805 applications, without swapping. Calculated from MemFree,
806 SReclaimable, the size of the file LRU lists, and the low
807 watermarks in each zone.
808 The estimate takes into account that the system needs some
809 page cache to function well, and that not all reclaimable
810 slab will be reclaimable, due to items being in use. The
811 impact of those factors will vary from system to system.
812 Buffers: Relatively temporary storage for raw disk blocks
813 shouldn't get tremendously large (20MB or so)
814 Cached: in-memory cache for files read from the disk (the
815 pagecache). Doesn't include SwapCached
816 SwapCached: Memory that once was swapped out, is swapped back in but
817 still also is in the swapfile (if memory is needed it
818 doesn't need to be swapped out AGAIN because it is already
819 in the swapfile. This saves I/O)
820 Active: Memory that has been used more recently and usually not
821 reclaimed unless absolutely necessary.
822 Inactive: Memory which has been less recently used. It is more
823 eligible to be reclaimed for other purposes
825 HighFree: Highmem is all memory above ~860MB of physical memory
826 Highmem areas are for use by userspace programs, or
827 for the pagecache. The kernel must use tricks to access
828 this memory, making it slower to access than lowmem.
830 LowFree: Lowmem is memory which can be used for everything that
831 highmem can be used for, but it is also available for the
832 kernel's use for its own data structures. Among many
833 other things, it is where everything from the Slab is
834 allocated. Bad things happen when you're out of lowmem.
835 SwapTotal: total amount of swap space available
836 SwapFree: Memory which has been evicted from RAM, and is temporarily
838 Dirty: Memory which is waiting to get written back to the disk
839 Writeback: Memory which is actively being written back to the disk
840 AnonPages: Non-file backed pages mapped into userspace page tables
841 AnonHugePages: Non-file backed huge pages mapped into userspace page tables
842 Mapped: files which have been mmaped, such as libraries
843 Slab: in-kernel data structures cache
844 SReclaimable: Part of Slab, that might be reclaimed, such as caches
845 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure
846 PageTables: amount of memory dedicated to the lowest level of page
848 NFS_Unstable: NFS pages sent to the server, but not yet committed to stable
850 Bounce: Memory used for block device "bounce buffers"
851 WritebackTmp: Memory used by FUSE for temporary writeback buffers
852 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
853 this is the total amount of memory currently available to
854 be allocated on the system. This limit is only adhered to
855 if strict overcommit accounting is enabled (mode 2 in
856 'vm.overcommit_memory').
857 The CommitLimit is calculated with the following formula:
858 CommitLimit = ([total RAM pages] - [total huge TLB pages]) *
859 overcommit_ratio / 100 + [total swap pages]
860 For example, on a system with 1G of physical RAM and 7G
861 of swap with a `vm.overcommit_ratio` of 30 it would
862 yield a CommitLimit of 7.3G.
863 For more details, see the memory overcommit documentation
864 in vm/overcommit-accounting.
865 Committed_AS: The amount of memory presently allocated on the system.
866 The committed memory is a sum of all of the memory which
867 has been allocated by processes, even if it has not been
868 "used" by them as of yet. A process which malloc()'s 1G
869 of memory, but only touches 300M of it will show up as
870 using 1G. This 1G is memory which has been "committed" to
871 by the VM and can be used at any time by the allocating
872 application. With strict overcommit enabled on the system
873 (mode 2 in 'vm.overcommit_memory'),allocations which would
874 exceed the CommitLimit (detailed above) will not be permitted.
875 This is useful if one needs to guarantee that processes will
876 not fail due to lack of memory once that memory has been
877 successfully allocated.
878 VmallocTotal: total size of vmalloc memory area
879 VmallocUsed: amount of vmalloc area which is used
880 VmallocChunk: largest contiguous block of vmalloc area which is free
882 ..............................................................................
886 Provides information about vmalloced/vmaped areas. One line per area,
887 containing the virtual address range of the area, size in bytes,
888 caller information of the creator, and optional information depending
889 on the kind of area :
891 pages=nr number of pages
892 phys=addr if a physical address was specified
893 ioremap I/O mapping (ioremap() and friends)
894 vmalloc vmalloc() area
897 vpages buffer for pages pointers was vmalloced (huge area)
898 N<node>=nr (Only on NUMA kernels)
899 Number of pages allocated on memory node <node>
901 > cat /proc/vmallocinfo
902 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ...
903 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128
904 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ...
905 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64
906 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f...
907 phys=7fee8000 ioremap
908 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f...
909 phys=7fee7000 ioremap
910 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210
911 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ...
912 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3
913 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ...
915 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ...
916 /0x130 [x_tables] pages=4 vmalloc N0=4
917 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ...
918 pages=14 vmalloc N2=14
919 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ...
921 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ...
923 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ...
924 pages=10 vmalloc N0=10
926 ..............................................................................
930 Provides counts of softirq handlers serviced since boot time, for each cpu.
935 TIMER: 27166 27120 27097 27034
940 SCHED: 27035 26983 26971 26746
942 RCU: 1678 1769 2178 2250
945 1.3 IDE devices in /proc/ide
946 ----------------------------
948 The subdirectory /proc/ide contains information about all IDE devices of which
949 the kernel is aware. There is one subdirectory for each IDE controller, the
950 file drivers and a link for each IDE device, pointing to the device directory
951 in the controller specific subtree.
953 The file drivers contains general information about the drivers used for the
956 > cat /proc/ide/drivers
957 ide-cdrom version 4.53
958 ide-disk version 1.08
960 More detailed information can be found in the controller specific
961 subdirectories. These are named ide0, ide1 and so on. Each of these
962 directories contains the files shown in table 1-6.
965 Table 1-6: IDE controller info in /proc/ide/ide?
966 ..............................................................................
968 channel IDE channel (0 or 1)
969 config Configuration (only for PCI/IDE bridge)
971 model Type/Chipset of IDE controller
972 ..............................................................................
974 Each device connected to a controller has a separate subdirectory in the
975 controllers directory. The files listed in table 1-7 are contained in these
979 Table 1-7: IDE device information
980 ..............................................................................
983 capacity Capacity of the medium (in 512Byte blocks)
984 driver driver and version
985 geometry physical and logical geometry
986 identify device identify block
988 model device identifier
989 settings device setup
990 smart_thresholds IDE disk management thresholds
991 smart_values IDE disk management values
992 ..............................................................................
994 The most interesting file is settings. This file contains a nice overview of
995 the drive parameters:
997 # cat /proc/ide/ide0/hda/settings
998 name value min max mode
999 ---- ----- --- --- ----
1000 bios_cyl 526 0 65535 rw
1001 bios_head 255 0 255 rw
1002 bios_sect 63 0 63 rw
1003 breada_readahead 4 0 127 rw
1005 file_readahead 72 0 2097151 rw
1007 keepsettings 0 0 1 rw
1008 max_kb_per_request 122 1 127 rw
1012 pio_mode write-only 0 255 w
1018 1.4 Networking info in /proc/net
1019 --------------------------------
1021 The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the
1022 additional values you get for IP version 6 if you configure the kernel to
1023 support this. Table 1-9 lists the files and their meaning.
1026 Table 1-8: IPv6 info in /proc/net
1027 ..............................................................................
1029 udp6 UDP sockets (IPv6)
1030 tcp6 TCP sockets (IPv6)
1031 raw6 Raw device statistics (IPv6)
1032 igmp6 IP multicast addresses, which this host joined (IPv6)
1033 if_inet6 List of IPv6 interface addresses
1034 ipv6_route Kernel routing table for IPv6
1035 rt6_stats Global IPv6 routing tables statistics
1036 sockstat6 Socket statistics (IPv6)
1037 snmp6 Snmp data (IPv6)
1038 ..............................................................................
1041 Table 1-9: Network info in /proc/net
1042 ..............................................................................
1044 arp Kernel ARP table
1045 dev network devices with statistics
1046 dev_mcast the Layer2 multicast groups a device is listening too
1047 (interface index, label, number of references, number of bound
1049 dev_stat network device status
1050 ip_fwchains Firewall chain linkage
1051 ip_fwnames Firewall chain names
1052 ip_masq Directory containing the masquerading tables
1053 ip_masquerade Major masquerading table
1054 netstat Network statistics
1055 raw raw device statistics
1056 route Kernel routing table
1057 rpc Directory containing rpc info
1058 rt_cache Routing cache
1060 sockstat Socket statistics
1063 unix UNIX domain sockets
1064 wireless Wireless interface data (Wavelan etc)
1065 igmp IP multicast addresses, which this host joined
1066 psched Global packet scheduler parameters.
1067 netlink List of PF_NETLINK sockets
1068 ip_mr_vifs List of multicast virtual interfaces
1069 ip_mr_cache List of multicast routing cache
1070 ..............................................................................
1072 You can use this information to see which network devices are available in
1073 your system and how much traffic was routed over those devices:
1076 Inter-|Receive |[...
1077 face |bytes packets errs drop fifo frame compressed multicast|[...
1078 lo: 908188 5596 0 0 0 0 0 0 [...
1079 ppp0:15475140 20721 410 0 0 410 0 0 [...
1080 eth0: 614530 7085 0 0 0 0 0 1 [...
1083 ...] bytes packets errs drop fifo colls carrier compressed
1084 ...] 908188 5596 0 0 0 0 0 0
1085 ...] 1375103 17405 0 0 0 0 0 0
1086 ...] 1703981 5535 0 0 0 3 0 0
1088 In addition, each Channel Bond interface has its own directory. For
1089 example, the bond0 device will have a directory called /proc/net/bond0/.
1090 It will contain information that is specific to that bond, such as the
1091 current slaves of the bond, the link status of the slaves, and how
1092 many times the slaves link has failed.
1097 If you have a SCSI host adapter in your system, you'll find a subdirectory
1098 named after the driver for this adapter in /proc/scsi. You'll also see a list
1099 of all recognized SCSI devices in /proc/scsi:
1101 >cat /proc/scsi/scsi
1103 Host: scsi0 Channel: 00 Id: 00 Lun: 00
1104 Vendor: IBM Model: DGHS09U Rev: 03E0
1105 Type: Direct-Access ANSI SCSI revision: 03
1106 Host: scsi0 Channel: 00 Id: 06 Lun: 00
1107 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
1108 Type: CD-ROM ANSI SCSI revision: 02
1111 The directory named after the driver has one file for each adapter found in
1112 the system. These files contain information about the controller, including
1113 the used IRQ and the IO address range. The amount of information shown is
1114 dependent on the adapter you use. The example shows the output for an Adaptec
1115 AHA-2940 SCSI adapter:
1117 > cat /proc/scsi/aic7xxx/0
1119 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
1121 TCQ Enabled By Default : Disabled
1122 AIC7XXX_PROC_STATS : Disabled
1123 AIC7XXX_RESET_DELAY : 5
1124 Adapter Configuration:
1125 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
1126 Ultra Wide Controller
1127 PCI MMAPed I/O Base: 0xeb001000
1128 Adapter SEEPROM Config: SEEPROM found and used.
1129 Adaptec SCSI BIOS: Enabled
1131 SCBs: Active 0, Max Active 2,
1132 Allocated 15, HW 16, Page 255
1134 BIOS Control Word: 0x18b6
1135 Adapter Control Word: 0x005b
1136 Extended Translation: Enabled
1137 Disconnect Enable Flags: 0xffff
1138 Ultra Enable Flags: 0x0001
1139 Tag Queue Enable Flags: 0x0000
1140 Ordered Queue Tag Flags: 0x0000
1141 Default Tag Queue Depth: 8
1142 Tagged Queue By Device array for aic7xxx host instance 0:
1143 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
1144 Actual queue depth per device for aic7xxx host instance 0:
1145 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
1148 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
1149 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
1150 Total transfers 160151 (74577 reads and 85574 writes)
1152 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
1153 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
1154 Total transfers 0 (0 reads and 0 writes)
1157 1.6 Parallel port info in /proc/parport
1158 ---------------------------------------
1160 The directory /proc/parport contains information about the parallel ports of
1161 your system. It has one subdirectory for each port, named after the port
1164 These directories contain the four files shown in Table 1-10.
1167 Table 1-10: Files in /proc/parport
1168 ..............................................................................
1170 autoprobe Any IEEE-1284 device ID information that has been acquired.
1171 devices list of the device drivers using that port. A + will appear by the
1172 name of the device currently using the port (it might not appear
1174 hardware Parallel port's base address, IRQ line and DMA channel.
1175 irq IRQ that parport is using for that port. This is in a separate
1176 file to allow you to alter it by writing a new value in (IRQ
1178 ..............................................................................
1180 1.7 TTY info in /proc/tty
1181 -------------------------
1183 Information about the available and actually used tty's can be found in the
1184 directory /proc/tty.You'll find entries for drivers and line disciplines in
1185 this directory, as shown in Table 1-11.
1188 Table 1-11: Files in /proc/tty
1189 ..............................................................................
1191 drivers list of drivers and their usage
1192 ldiscs registered line disciplines
1193 driver/serial usage statistic and status of single tty lines
1194 ..............................................................................
1196 To see which tty's are currently in use, you can simply look into the file
1199 > cat /proc/tty/drivers
1200 pty_slave /dev/pts 136 0-255 pty:slave
1201 pty_master /dev/ptm 128 0-255 pty:master
1202 pty_slave /dev/ttyp 3 0-255 pty:slave
1203 pty_master /dev/pty 2 0-255 pty:master
1204 serial /dev/cua 5 64-67 serial:callout
1205 serial /dev/ttyS 4 64-67 serial
1206 /dev/tty0 /dev/tty0 4 0 system:vtmaster
1207 /dev/ptmx /dev/ptmx 5 2 system
1208 /dev/console /dev/console 5 1 system:console
1209 /dev/tty /dev/tty 5 0 system:/dev/tty
1210 unknown /dev/tty 4 1-63 console
1213 1.8 Miscellaneous kernel statistics in /proc/stat
1214 -------------------------------------------------
1216 Various pieces of information about kernel activity are available in the
1217 /proc/stat file. All of the numbers reported in this file are aggregates
1218 since the system first booted. For a quick look, simply cat the file:
1221 cpu 2255 34 2290 22625563 6290 127 456 0 0
1222 cpu0 1132 34 1441 11311718 3675 127 438 0 0
1223 cpu1 1123 0 849 11313845 2614 0 18 0 0
1224 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
1230 softirq 183433 0 21755 12 39 1137 231 21459 2263
1232 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
1233 lines. These numbers identify the amount of time the CPU has spent performing
1234 different kinds of work. Time units are in USER_HZ (typically hundredths of a
1235 second). The meanings of the columns are as follows, from left to right:
1237 - user: normal processes executing in user mode
1238 - nice: niced processes executing in user mode
1239 - system: processes executing in kernel mode
1240 - idle: twiddling thumbs
1241 - iowait: waiting for I/O to complete
1242 - irq: servicing interrupts
1243 - softirq: servicing softirqs
1244 - steal: involuntary wait
1245 - guest: running a normal guest
1246 - guest_nice: running a niced guest
1248 The "intr" line gives counts of interrupts serviced since boot time, for each
1249 of the possible system interrupts. The first column is the total of all
1250 interrupts serviced including unnumbered architecture specific interrupts;
1251 each subsequent column is the total for that particular numbered interrupt.
1252 Unnumbered interrupts are not shown, only summed into the total.
1254 The "ctxt" line gives the total number of context switches across all CPUs.
1256 The "btime" line gives the time at which the system booted, in seconds since
1259 The "processes" line gives the number of processes and threads created, which
1260 includes (but is not limited to) those created by calls to the fork() and
1261 clone() system calls.
1263 The "procs_running" line gives the total number of threads that are
1264 running or ready to run (i.e., the total number of runnable threads).
1266 The "procs_blocked" line gives the number of processes currently blocked,
1267 waiting for I/O to complete.
1269 The "softirq" line gives counts of softirqs serviced since boot time, for each
1270 of the possible system softirqs. The first column is the total of all
1271 softirqs serviced; each subsequent column is the total for that particular
1275 1.9 Ext4 file system parameters
1276 -------------------------------
1278 Information about mounted ext4 file systems can be found in
1279 /proc/fs/ext4. Each mounted filesystem will have a directory in
1280 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
1281 /proc/fs/ext4/dm-0). The files in each per-device directory are shown
1282 in Table 1-12, below.
1284 Table 1-12: Files in /proc/fs/ext4/<devname>
1285 ..............................................................................
1287 mb_groups details of multiblock allocator buddy cache of free blocks
1288 ..............................................................................
1292 Shows registered system console lines.
1294 To see which character device lines are currently used for the system console
1295 /dev/console, you may simply look into the file /proc/consoles:
1297 > cat /proc/consoles
1303 device name of the device
1304 operations R = can do read operations
1305 W = can do write operations
1307 flags E = it is enabled
1308 C = it is preferred console
1309 B = it is primary boot console
1310 p = it is used for printk buffer
1311 b = it is not a TTY but a Braille device
1312 a = it is safe to use when cpu is offline
1313 major:minor major and minor number of the device separated by a colon
1315 ------------------------------------------------------------------------------
1317 ------------------------------------------------------------------------------
1318 The /proc file system serves information about the running system. It not only
1319 allows access to process data but also allows you to request the kernel status
1320 by reading files in the hierarchy.
1322 The directory structure of /proc reflects the types of information and makes
1323 it easy, if not obvious, where to look for specific data.
1324 ------------------------------------------------------------------------------
1326 ------------------------------------------------------------------------------
1327 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
1328 ------------------------------------------------------------------------------
1330 ------------------------------------------------------------------------------
1332 ------------------------------------------------------------------------------
1333 * Modifying kernel parameters by writing into files found in /proc/sys
1334 * Exploring the files which modify certain parameters
1335 * Review of the /proc/sys file tree
1336 ------------------------------------------------------------------------------
1339 A very interesting part of /proc is the directory /proc/sys. This is not only
1340 a source of information, it also allows you to change parameters within the
1341 kernel. Be very careful when attempting this. You can optimize your system,
1342 but you can also cause it to crash. Never alter kernel parameters on a
1343 production system. Set up a development machine and test to make sure that
1344 everything works the way you want it to. You may have no alternative but to
1345 reboot the machine once an error has been made.
1347 To change a value, simply echo the new value into the file. An example is
1348 given below in the section on the file system data. You need to be root to do
1349 this. You can create your own boot script to perform this every time your
1352 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
1353 general things in the operation of the Linux kernel. Since some of the files
1354 can inadvertently disrupt your system, it is advisable to read both
1355 documentation and source before actually making adjustments. In any case, be
1356 very careful when writing to any of these files. The entries in /proc may
1357 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
1358 review the kernel documentation in the directory /usr/src/linux/Documentation.
1359 This chapter is heavily based on the documentation included in the pre 2.2
1360 kernels, and became part of it in version 2.2.1 of the Linux kernel.
1362 Please see: Documentation/sysctl/ directory for descriptions of these
1365 ------------------------------------------------------------------------------
1367 ------------------------------------------------------------------------------
1368 Certain aspects of kernel behavior can be modified at runtime, without the
1369 need to recompile the kernel, or even to reboot the system. The files in the
1370 /proc/sys tree can not only be read, but also modified. You can use the echo
1371 command to write value into these files, thereby changing the default settings
1373 ------------------------------------------------------------------------------
1375 ------------------------------------------------------------------------------
1376 CHAPTER 3: PER-PROCESS PARAMETERS
1377 ------------------------------------------------------------------------------
1379 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score
1380 --------------------------------------------------------------------------------
1382 These file can be used to adjust the badness heuristic used to select which
1383 process gets killed in out of memory conditions.
1385 The badness heuristic assigns a value to each candidate task ranging from 0
1386 (never kill) to 1000 (always kill) to determine which process is targeted. The
1387 units are roughly a proportion along that range of allowed memory the process
1388 may allocate from based on an estimation of its current memory and swap use.
1389 For example, if a task is using all allowed memory, its badness score will be
1390 1000. If it is using half of its allowed memory, its score will be 500.
1392 There is an additional factor included in the badness score: the current memory
1393 and swap usage is discounted by 3% for root processes.
1395 The amount of "allowed" memory depends on the context in which the oom killer
1396 was called. If it is due to the memory assigned to the allocating task's cpuset
1397 being exhausted, the allowed memory represents the set of mems assigned to that
1398 cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed
1399 memory represents the set of mempolicy nodes. If it is due to a memory
1400 limit (or swap limit) being reached, the allowed memory is that configured
1401 limit. Finally, if it is due to the entire system being out of memory, the
1402 allowed memory represents all allocatable resources.
1404 The value of /proc/<pid>/oom_score_adj is added to the badness score before it
1405 is used to determine which task to kill. Acceptable values range from -1000
1406 (OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to
1407 polarize the preference for oom killing either by always preferring a certain
1408 task or completely disabling it. The lowest possible value, -1000, is
1409 equivalent to disabling oom killing entirely for that task since it will always
1410 report a badness score of 0.
1412 Consequently, it is very simple for userspace to define the amount of memory to
1413 consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for
1414 example, is roughly equivalent to allowing the remainder of tasks sharing the
1415 same system, cpuset, mempolicy, or memory controller resources to use at least
1416 50% more memory. A value of -500, on the other hand, would be roughly
1417 equivalent to discounting 50% of the task's allowed memory from being considered
1418 as scoring against the task.
1420 For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also
1421 be used to tune the badness score. Its acceptable values range from -16
1422 (OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17
1423 (OOM_DISABLE) to disable oom killing entirely for that task. Its value is
1424 scaled linearly with /proc/<pid>/oom_score_adj.
1426 The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last
1427 value set by a CAP_SYS_RESOURCE process. To reduce the value any lower
1428 requires CAP_SYS_RESOURCE.
1430 Caveat: when a parent task is selected, the oom killer will sacrifice any first
1431 generation children with separate address spaces instead, if possible. This
1432 avoids servers and important system daemons from being killed and loses the
1433 minimal amount of work.
1436 3.2 /proc/<pid>/oom_score - Display current oom-killer score
1437 -------------------------------------------------------------
1439 This file can be used to check the current score used by the oom-killer is for
1440 any given <pid>. Use it together with /proc/<pid>/oom_score_adj to tune which
1441 process should be killed in an out-of-memory situation.
1444 3.3 /proc/<pid>/io - Display the IO accounting fields
1445 -------------------------------------------------------
1447 This file contains IO statistics for each running process
1452 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
1455 test:/tmp # cat /proc/3828/io
1461 write_bytes: 323932160
1462 cancelled_write_bytes: 0
1471 I/O counter: chars read
1472 The number of bytes which this task has caused to be read from storage. This
1473 is simply the sum of bytes which this process passed to read() and pread().
1474 It includes things like tty IO and it is unaffected by whether or not actual
1475 physical disk IO was required (the read might have been satisfied from
1482 I/O counter: chars written
1483 The number of bytes which this task has caused, or shall cause to be written
1484 to disk. Similar caveats apply here as with rchar.
1490 I/O counter: read syscalls
1491 Attempt to count the number of read I/O operations, i.e. syscalls like read()
1498 I/O counter: write syscalls
1499 Attempt to count the number of write I/O operations, i.e. syscalls like
1500 write() and pwrite().
1506 I/O counter: bytes read
1507 Attempt to count the number of bytes which this process really did cause to
1508 be fetched from the storage layer. Done at the submit_bio() level, so it is
1509 accurate for block-backed filesystems. <please add status regarding NFS and
1510 CIFS at a later time>
1516 I/O counter: bytes written
1517 Attempt to count the number of bytes which this process caused to be sent to
1518 the storage layer. This is done at page-dirtying time.
1521 cancelled_write_bytes
1522 ---------------------
1524 The big inaccuracy here is truncate. If a process writes 1MB to a file and
1525 then deletes the file, it will in fact perform no writeout. But it will have
1526 been accounted as having caused 1MB of write.
1527 In other words: The number of bytes which this process caused to not happen,
1528 by truncating pagecache. A task can cause "negative" IO too. If this task
1529 truncates some dirty pagecache, some IO which another task has been accounted
1530 for (in its write_bytes) will not be happening. We _could_ just subtract that
1531 from the truncating task's write_bytes, but there is information loss in doing
1538 At its current implementation state, this is a bit racy on 32-bit machines: if
1539 process A reads process B's /proc/pid/io while process B is updating one of
1540 those 64-bit counters, process A could see an intermediate result.
1543 More information about this can be found within the taskstats documentation in
1544 Documentation/accounting.
1546 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings
1547 ---------------------------------------------------------------
1548 When a process is dumped, all anonymous memory is written to a core file as
1549 long as the size of the core file isn't limited. But sometimes we don't want
1550 to dump some memory segments, for example, huge shared memory. Conversely,
1551 sometimes we want to save file-backed memory segments into a core file, not
1552 only the individual files.
1554 /proc/<pid>/coredump_filter allows you to customize which memory segments
1555 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
1556 of memory types. If a bit of the bitmask is set, memory segments of the
1557 corresponding memory type are dumped, otherwise they are not dumped.
1559 The following 7 memory types are supported:
1560 - (bit 0) anonymous private memory
1561 - (bit 1) anonymous shared memory
1562 - (bit 2) file-backed private memory
1563 - (bit 3) file-backed shared memory
1564 - (bit 4) ELF header pages in file-backed private memory areas (it is
1565 effective only if the bit 2 is cleared)
1566 - (bit 5) hugetlb private memory
1567 - (bit 6) hugetlb shared memory
1569 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
1570 are always dumped regardless of the bitmask status.
1572 Note bit 0-4 doesn't effect any hugetlb memory. hugetlb memory are only
1573 effected by bit 5-6.
1575 Default value of coredump_filter is 0x23; this means all anonymous memory
1576 segments and hugetlb private memory are dumped.
1578 If you don't want to dump all shared memory segments attached to pid 1234,
1579 write 0x21 to the process's proc file.
1581 $ echo 0x21 > /proc/1234/coredump_filter
1583 When a new process is created, the process inherits the bitmask status from its
1584 parent. It is useful to set up coredump_filter before the program runs.
1587 $ echo 0x7 > /proc/self/coredump_filter
1590 3.5 /proc/<pid>/mountinfo - Information about mounts
1591 --------------------------------------------------------
1593 This file contains lines of the form:
1595 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
1596 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
1598 (1) mount ID: unique identifier of the mount (may be reused after umount)
1599 (2) parent ID: ID of parent (or of self for the top of the mount tree)
1600 (3) major:minor: value of st_dev for files on filesystem
1601 (4) root: root of the mount within the filesystem
1602 (5) mount point: mount point relative to the process's root
1603 (6) mount options: per mount options
1604 (7) optional fields: zero or more fields of the form "tag[:value]"
1605 (8) separator: marks the end of the optional fields
1606 (9) filesystem type: name of filesystem of the form "type[.subtype]"
1607 (10) mount source: filesystem specific information or "none"
1608 (11) super options: per super block options
1610 Parsers should ignore all unrecognised optional fields. Currently the
1611 possible optional fields are:
1613 shared:X mount is shared in peer group X
1614 master:X mount is slave to peer group X
1615 propagate_from:X mount is slave and receives propagation from peer group X (*)
1616 unbindable mount is unbindable
1618 (*) X is the closest dominant peer group under the process's root. If
1619 X is the immediate master of the mount, or if there's no dominant peer
1620 group under the same root, then only the "master:X" field is present
1621 and not the "propagate_from:X" field.
1623 For more information on mount propagation see:
1625 Documentation/filesystems/sharedsubtree.txt
1628 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm
1629 --------------------------------------------------------
1630 These files provide a method to access a tasks comm value. It also allows for
1631 a task to set its own or one of its thread siblings comm value. The comm value
1632 is limited in size compared to the cmdline value, so writing anything longer
1633 then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated
1637 3.7 /proc/<pid>/task/<tid>/children - Information about task children
1638 -------------------------------------------------------------------------
1639 This file provides a fast way to retrieve first level children pids
1640 of a task pointed by <pid>/<tid> pair. The format is a space separated
1643 Note the "first level" here -- if a child has own children they will
1644 not be listed here, one needs to read /proc/<children-pid>/task/<tid>/children
1645 to obtain the descendants.
1647 Since this interface is intended to be fast and cheap it doesn't
1648 guarantee to provide precise results and some children might be
1649 skipped, especially if they've exited right after we printed their
1650 pids, so one need to either stop or freeze processes being inspected
1651 if precise results are needed.
1654 3.8 /proc/<pid>/fdinfo/<fd> - Information about opened file
1655 ---------------------------------------------------------------
1656 This file provides information associated with an opened file. The regular
1657 files have at least three fields -- 'pos', 'flags' and mnt_id. The 'pos'
1658 represents the current offset of the opened file in decimal form [see lseek(2)
1659 for details], 'flags' denotes the octal O_xxx mask the file has been
1660 created with [see open(2) for details] and 'mnt_id' represents mount ID of
1661 the file system containing the opened file [see 3.5 /proc/<pid>/mountinfo
1670 The files such as eventfd, fsnotify, signalfd, epoll among the regular pos/flags
1671 pair provide additional information particular to the objects they represent.
1680 where 'eventfd-count' is hex value of a counter.
1687 sigmask: 0000000000000200
1689 where 'sigmask' is hex value of the signal mask associated
1697 tfd: 5 events: 1d data: ffffffffffffffff
1699 where 'tfd' is a target file descriptor number in decimal form,
1700 'events' is events mask being watched and the 'data' is data
1701 associated with a target [see epoll(7) for more details].
1705 For inotify files the format is the following
1709 inotify wd:3 ino:9e7e sdev:800013 mask:800afce ignored_mask:0 fhandle-bytes:8 fhandle-type:1 f_handle:7e9e0000640d1b6d
1711 where 'wd' is a watch descriptor in decimal form, ie a target file
1712 descriptor number, 'ino' and 'sdev' are inode and device where the
1713 target file resides and the 'mask' is the mask of events, all in hex
1714 form [see inotify(7) for more details].
1716 If the kernel was built with exportfs support, the path to the target
1717 file is encoded as a file handle. The file handle is provided by three
1718 fields 'fhandle-bytes', 'fhandle-type' and 'f_handle', all in hex
1721 If the kernel is built without exportfs support the file handle won't be
1724 If there is no inotify mark attached yet the 'inotify' line will be omitted.
1726 For fanotify files the format is
1731 fanotify flags:10 event-flags:0
1732 fanotify mnt_id:12 mflags:40 mask:38 ignored_mask:40000003
1733 fanotify ino:4f969 sdev:800013 mflags:0 mask:3b ignored_mask:40000000 fhandle-bytes:8 fhandle-type:1 f_handle:69f90400c275b5b4
1735 where fanotify 'flags' and 'event-flags' are values used in fanotify_init
1736 call, 'mnt_id' is the mount point identifier, 'mflags' is the value of
1737 flags associated with mark which are tracked separately from events
1738 mask. 'ino', 'sdev' are target inode and device, 'mask' is the events
1739 mask and 'ignored_mask' is the mask of events which are to be ignored.
1740 All in hex format. Incorporation of 'mflags', 'mask' and 'ignored_mask'
1741 does provide information about flags and mask used in fanotify_mark
1742 call [see fsnotify manpage for details].
1744 While the first three lines are mandatory and always printed, the rest is
1745 optional and may be omitted if no marks created yet.
1756 it_value: (0, 49406829)
1759 where 'clockid' is the clock type and 'ticks' is the number of the timer expirations
1760 that have occurred [see timerfd_create(2) for details]. 'settime flags' are
1761 flags in octal form been used to setup the timer [see timerfd_settime(2) for
1762 details]. 'it_value' is remaining time until the timer exiration.
1763 'it_interval' is the interval for the timer. Note the timer might be set up
1764 with TIMER_ABSTIME option which will be shown in 'settime flags', but 'it_value'
1765 still exhibits timer's remaining time.
1767 3.9 /proc/<pid>/map_files - Information about memory mapped files
1768 ---------------------------------------------------------------------
1769 This directory contains symbolic links which represent memory mapped files
1770 the process is maintaining. Example output:
1772 | lr-------- 1 root root 64 Jan 27 11:24 333c600000-333c620000 -> /usr/lib64/ld-2.18.so
1773 | lr-------- 1 root root 64 Jan 27 11:24 333c81f000-333c820000 -> /usr/lib64/ld-2.18.so
1774 | lr-------- 1 root root 64 Jan 27 11:24 333c820000-333c821000 -> /usr/lib64/ld-2.18.so
1776 | lr-------- 1 root root 64 Jan 27 11:24 35d0421000-35d0422000 -> /usr/lib64/libselinux.so.1
1777 | lr-------- 1 root root 64 Jan 27 11:24 400000-41a000 -> /usr/bin/ls
1779 The name of a link represents the virtual memory bounds of a mapping, i.e.
1780 vm_area_struct::vm_start-vm_area_struct::vm_end.
1782 The main purpose of the map_files is to retrieve a set of memory mapped
1783 files in a fast way instead of parsing /proc/<pid>/maps or
1784 /proc/<pid>/smaps, both of which contain many more records. At the same
1785 time one can open(2) mappings from the listings of two processes and
1786 comparing their inode numbers to figure out which anonymous memory areas
1787 are actually shared.
1789 ------------------------------------------------------------------------------
1791 ------------------------------------------------------------------------------
1794 ---------------------
1796 The following mount options are supported:
1798 hidepid= Set /proc/<pid>/ access mode.
1799 gid= Set the group authorized to learn processes information.
1801 hidepid=0 means classic mode - everybody may access all /proc/<pid>/ directories
1804 hidepid=1 means users may not access any /proc/<pid>/ directories but their
1805 own. Sensitive files like cmdline, sched*, status are now protected against
1806 other users. This makes it impossible to learn whether any user runs
1807 specific program (given the program doesn't reveal itself by its behaviour).
1808 As an additional bonus, as /proc/<pid>/cmdline is unaccessible for other users,
1809 poorly written programs passing sensitive information via program arguments are
1810 now protected against local eavesdroppers.
1812 hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be fully invisible to other
1813 users. It doesn't mean that it hides a fact whether a process with a specific
1814 pid value exists (it can be learned by other means, e.g. by "kill -0 $PID"),
1815 but it hides process' uid and gid, which may be learned by stat()'ing
1816 /proc/<pid>/ otherwise. It greatly complicates an intruder's task of gathering
1817 information about running processes, whether some daemon runs with elevated
1818 privileges, whether other user runs some sensitive program, whether other users
1819 run any program at all, etc.
1821 gid= defines a group authorized to learn processes information otherwise
1822 prohibited by hidepid=. If you use some daemon like identd which needs to learn
1823 information about processes information, just add identd to this group.