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
47 ------------------------------------------------------------------------------
49 ------------------------------------------------------------------------------
51 0.1 Introduction/Credits
52 ------------------------
54 This documentation is part of a soon (or so we hope) to be released book on
55 the SuSE Linux distribution. As there is no complete documentation for the
56 /proc file system and we've used many freely available sources to write these
57 chapters, it seems only fair to give the work back to the Linux community.
58 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
59 afraid it's still far from complete, but we hope it will be useful. As far as
60 we know, it is the first 'all-in-one' document about the /proc file system. It
61 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
62 SPARC, AXP, etc., features, you probably won't find what you are looking for.
63 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
64 additions and patches are welcome and will be added to this document if you
67 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
68 other people for help compiling this documentation. We'd also like to extend a
69 special thank you to Andi Kleen for documentation, which we relied on heavily
70 to create this document, as well as the additional information he provided.
71 Thanks to everybody else who contributed source or docs to the Linux kernel
72 and helped create a great piece of software... :)
74 If you have any comments, corrections or additions, please don't hesitate to
75 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
78 The latest version of this document is available online at
79 http://tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html
81 If the above direction does not works for you, you could try the kernel
82 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
83 comandante@zaralinux.com.
88 We don't guarantee the correctness of this document, and if you come to us
89 complaining about how you screwed up your system because of incorrect
90 documentation, we won't feel responsible...
92 ------------------------------------------------------------------------------
93 CHAPTER 1: COLLECTING SYSTEM INFORMATION
94 ------------------------------------------------------------------------------
96 ------------------------------------------------------------------------------
98 ------------------------------------------------------------------------------
99 * Investigating the properties of the pseudo file system /proc and its
100 ability to provide information on the running Linux system
101 * Examining /proc's structure
102 * Uncovering various information about the kernel and the processes running
104 ------------------------------------------------------------------------------
107 The proc file system acts as an interface to internal data structures in the
108 kernel. It can be used to obtain information about the system and to change
109 certain kernel parameters at runtime (sysctl).
111 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
112 show you how you can use /proc/sys to change settings.
114 1.1 Process-Specific Subdirectories
115 -----------------------------------
117 The directory /proc contains (among other things) one subdirectory for each
118 process running on the system, which is named after the process ID (PID).
120 The link self points to the process reading the file system. Each process
121 subdirectory has the entries listed in Table 1-1.
124 Table 1-1: Process specific entries in /proc
125 ..............................................................................
127 clear_refs Clears page referenced bits shown in smaps output
128 cmdline Command line arguments
129 cpu Current and last cpu in which it was executed (2.4)(smp)
130 cwd Link to the current working directory
131 environ Values of environment variables
132 exe Link to the executable of this process
133 fd Directory, which contains all file descriptors
134 maps Memory maps to executables and library files (2.4)
135 mem Memory held by this process
136 root Link to the root directory of this process
138 statm Process memory status information
139 status Process status in human readable form
140 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan
142 stack Report full stack trace, enable via CONFIG_STACKTRACE
143 smaps a extension based on maps, showing the memory consumption of
145 ..............................................................................
147 For example, to get the status information of a process, all you have to do is
148 read the file /proc/PID/status:
150 >cat /proc/self/status
174 SigPnd: 0000000000000000
175 ShdPnd: 0000000000000000
176 SigBlk: 0000000000000000
177 SigIgn: 0000000000000000
178 SigCgt: 0000000000000000
179 CapInh: 00000000fffffeff
180 CapPrm: 0000000000000000
181 CapEff: 0000000000000000
182 CapBnd: ffffffffffffffff
183 voluntary_ctxt_switches: 0
184 nonvoluntary_ctxt_switches: 1
186 This shows you nearly the same information you would get if you viewed it with
187 the ps command. In fact, ps uses the proc file system to obtain its
188 information. But you get a more detailed view of the process by reading the
189 file /proc/PID/status. It fields are described in table 1-2.
191 The statm file contains more detailed information about the process
192 memory usage. Its seven fields are explained in Table 1-3. The stat file
193 contains details information about the process itself. Its fields are
194 explained in Table 1-4.
196 (for SMP CONFIG users)
197 For making accounting scalable, RSS related information are handled in
198 asynchronous manner and the vaule may not be very precise. To see a precise
199 snapshot of a moment, you can see /proc/<pid>/smaps file and scan page table.
200 It's slow but very precise.
202 Table 1-2: Contents of the status files (as of 2.6.30-rc7)
203 ..............................................................................
205 Name filename of the executable
206 State state (R is running, S is sleeping, D is sleeping
207 in an uninterruptible wait, Z is zombie,
208 T is traced or stopped)
211 PPid process id of the parent process
212 TracerPid PID of process tracing this process (0 if not)
213 Uid Real, effective, saved set, and file system UIDs
214 Gid Real, effective, saved set, and file system GIDs
215 FDSize number of file descriptor slots currently allocated
216 Groups supplementary group list
217 VmPeak peak virtual memory size
218 VmSize total program size
219 VmLck locked memory size
220 VmHWM peak resident set size ("high water mark")
221 VmRSS size of memory portions
222 VmData size of data, stack, and text segments
223 VmStk size of data, stack, and text segments
224 VmExe size of text segment
225 VmLib size of shared library code
226 VmPTE size of page table entries
227 VmSwap size of swap usage (the number of referred swapents)
228 Threads number of threads
229 SigQ number of signals queued/max. number for queue
230 SigPnd bitmap of pending signals for the thread
231 ShdPnd bitmap of shared pending signals for the process
232 SigBlk bitmap of blocked signals
233 SigIgn bitmap of ignored signals
234 SigCgt bitmap of catched signals
235 CapInh bitmap of inheritable capabilities
236 CapPrm bitmap of permitted capabilities
237 CapEff bitmap of effective capabilities
238 CapBnd bitmap of capabilities bounding set
239 Cpus_allowed mask of CPUs on which this process may run
240 Cpus_allowed_list Same as previous, but in "list format"
241 Mems_allowed mask of memory nodes allowed to this process
242 Mems_allowed_list Same as previous, but in "list format"
243 voluntary_ctxt_switches number of voluntary context switches
244 nonvoluntary_ctxt_switches number of non voluntary context switches
245 ..............................................................................
247 Table 1-3: Contents of the statm files (as of 2.6.8-rc3)
248 ..............................................................................
250 size total program size (pages) (same as VmSize in status)
251 resident size of memory portions (pages) (same as VmRSS in status)
252 shared number of pages that are shared (i.e. backed by a file)
253 trs number of pages that are 'code' (not including libs; broken,
254 includes data segment)
255 lrs number of pages of library (always 0 on 2.6)
256 drs number of pages of data/stack (including libs; broken,
257 includes library text)
258 dt number of dirty pages (always 0 on 2.6)
259 ..............................................................................
262 Table 1-4: Contents of the stat files (as of 2.6.30-rc7)
263 ..............................................................................
266 tcomm filename of the executable
267 state state (R is running, S is sleeping, D is sleeping in an
268 uninterruptible wait, Z is zombie, T is traced or stopped)
269 ppid process id of the parent process
270 pgrp pgrp of the process
272 tty_nr tty the process uses
273 tty_pgrp pgrp of the tty
275 min_flt number of minor faults
276 cmin_flt number of minor faults with child's
277 maj_flt number of major faults
278 cmaj_flt number of major faults with child's
279 utime user mode jiffies
280 stime kernel mode jiffies
281 cutime user mode jiffies with child's
282 cstime kernel mode jiffies with child's
283 priority priority level
285 num_threads number of threads
286 it_real_value (obsolete, always 0)
287 start_time time the process started after system boot
288 vsize virtual memory size
289 rss resident set memory size
290 rsslim current limit in bytes on the rss
291 start_code address above which program text can run
292 end_code address below which program text can run
293 start_stack address of the start of the main process stack
294 esp current value of ESP
295 eip current value of EIP
296 pending bitmap of pending signals
297 blocked bitmap of blocked signals
298 sigign bitmap of ignored signals
299 sigcatch bitmap of catched signals
300 wchan address where process went to sleep
303 exit_signal signal to send to parent thread on exit
304 task_cpu which CPU the task is scheduled on
305 rt_priority realtime priority
306 policy scheduling policy (man sched_setscheduler)
307 blkio_ticks time spent waiting for block IO
308 gtime guest time of the task in jiffies
309 cgtime guest time of the task children in jiffies
310 start_data address above which program data+bss is placed
311 end_data address below which program data+bss is placed
312 start_brk address above which program heap can be expanded with brk()
313 ..............................................................................
315 The /proc/PID/maps file containing the currently mapped memory regions and
316 their access permissions.
320 address perms offset dev inode pathname
322 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
323 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
324 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
325 a7cb1000-a7cb2000 ---p 00000000 00:00 0
326 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
327 a7eb2000-a7eb3000 ---p 00000000 00:00 0
328 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 [stack:1001]
329 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
330 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
331 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
332 a800b000-a800e000 rw-p 00000000 00:00 0
333 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
334 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
335 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
336 a8024000-a8027000 rw-p 00000000 00:00 0
337 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
338 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
339 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
340 aff35000-aff4a000 rw-p 00000000 00:00 0 [stack]
341 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
343 where "address" is the address space in the process that it occupies, "perms"
344 is a set of permissions:
350 p = private (copy on write)
352 "offset" is the offset into the mapping, "dev" is the device (major:minor), and
353 "inode" is the inode on that device. 0 indicates that no inode is associated
354 with the memory region, as the case would be with BSS (uninitialized data).
355 The "pathname" shows the name associated file for this mapping. If the mapping
356 is not associated with a file:
358 [heap] = the heap of the program
359 [stack] = the stack of the main process
360 [stack:1001] = the stack of the thread with tid 1001
361 [vdso] = the "virtual dynamic shared object",
362 the kernel system call handler
364 or if empty, the mapping is anonymous.
366 The /proc/PID/task/TID/maps is a view of the virtual memory from the viewpoint
367 of the individual tasks of a process. In this file you will see a mapping marked
368 as [stack] if that task sees it as a stack. This is a key difference from the
369 content of /proc/PID/maps, where you will see all mappings that are being used
370 as stack by all of those tasks. Hence, for the example above, the task-level
371 map, i.e. /proc/PID/task/TID/maps for thread 1001 will look like this:
373 08048000-08049000 r-xp 00000000 03:00 8312 /opt/test
374 08049000-0804a000 rw-p 00001000 03:00 8312 /opt/test
375 0804a000-0806b000 rw-p 00000000 00:00 0 [heap]
376 a7cb1000-a7cb2000 ---p 00000000 00:00 0
377 a7cb2000-a7eb2000 rw-p 00000000 00:00 0
378 a7eb2000-a7eb3000 ---p 00000000 00:00 0
379 a7eb3000-a7ed5000 rw-p 00000000 00:00 0 [stack]
380 a7ed5000-a8008000 r-xp 00000000 03:00 4222 /lib/libc.so.6
381 a8008000-a800a000 r--p 00133000 03:00 4222 /lib/libc.so.6
382 a800a000-a800b000 rw-p 00135000 03:00 4222 /lib/libc.so.6
383 a800b000-a800e000 rw-p 00000000 00:00 0
384 a800e000-a8022000 r-xp 00000000 03:00 14462 /lib/libpthread.so.0
385 a8022000-a8023000 r--p 00013000 03:00 14462 /lib/libpthread.so.0
386 a8023000-a8024000 rw-p 00014000 03:00 14462 /lib/libpthread.so.0
387 a8024000-a8027000 rw-p 00000000 00:00 0
388 a8027000-a8043000 r-xp 00000000 03:00 8317 /lib/ld-linux.so.2
389 a8043000-a8044000 r--p 0001b000 03:00 8317 /lib/ld-linux.so.2
390 a8044000-a8045000 rw-p 0001c000 03:00 8317 /lib/ld-linux.so.2
391 aff35000-aff4a000 rw-p 00000000 00:00 0
392 ffffe000-fffff000 r-xp 00000000 00:00 0 [vdso]
394 The /proc/PID/smaps is an extension based on maps, showing the memory
395 consumption for each of the process's mappings. For each of mappings there
396 is a series of lines such as the following:
398 08048000-080bc000 r-xp 00000000 03:02 13130 /bin/bash
413 The first of these lines shows the same information as is displayed for the
414 mapping in /proc/PID/maps. The remaining lines show the size of the mapping
415 (size), the amount of the mapping that is currently resident in RAM (RSS), the
416 process' proportional share of this mapping (PSS), the number of clean and
417 dirty private pages in the mapping. Note that even a page which is part of a
418 MAP_SHARED mapping, but has only a single pte mapped, i.e. is currently used
419 by only one process, is accounted as private and not as shared. "Referenced"
420 indicates the amount of memory currently marked as referenced or accessed.
421 "Anonymous" shows the amount of memory that does not belong to any file. Even
422 a mapping associated with a file may contain anonymous pages: when MAP_PRIVATE
423 and a page is modified, the file page is replaced by a private anonymous copy.
424 "Swap" shows how much would-be-anonymous memory is also used, but out on
427 This file is only present if the CONFIG_MMU kernel configuration option is
430 The /proc/PID/clear_refs is used to reset the PG_Referenced and ACCESSED/YOUNG
431 bits on both physical and virtual pages associated with a process.
432 To clear the bits for all the pages associated with the process
433 > echo 1 > /proc/PID/clear_refs
435 To clear the bits for the anonymous pages associated with the process
436 > echo 2 > /proc/PID/clear_refs
438 To clear the bits for the file mapped pages associated with the process
439 > echo 3 > /proc/PID/clear_refs
440 Any other value written to /proc/PID/clear_refs will have no effect.
442 The /proc/pid/pagemap gives the PFN, which can be used to find the pageflags
443 using /proc/kpageflags and number of times a page is mapped using
444 /proc/kpagecount. For detailed explanation, see Documentation/vm/pagemap.txt.
449 Similar to the process entries, the kernel data files give information about
450 the running kernel. The files used to obtain this information are contained in
451 /proc and are listed in Table 1-5. Not all of these will be present in your
452 system. It depends on the kernel configuration and the loaded modules, which
453 files are there, and which are missing.
455 Table 1-5: Kernel info in /proc
456 ..............................................................................
458 apm Advanced power management info
459 buddyinfo Kernel memory allocator information (see text) (2.5)
460 bus Directory containing bus specific information
461 cmdline Kernel command line
462 cpuinfo Info about the CPU
463 devices Available devices (block and character)
464 dma Used DMS channels
465 filesystems Supported filesystems
466 driver Various drivers grouped here, currently rtc (2.4)
467 execdomains Execdomains, related to security (2.4)
468 fb Frame Buffer devices (2.4)
469 fs File system parameters, currently nfs/exports (2.4)
470 ide Directory containing info about the IDE subsystem
471 interrupts Interrupt usage
472 iomem Memory map (2.4)
473 ioports I/O port usage
474 irq Masks for irq to cpu affinity (2.4)(smp?)
475 isapnp ISA PnP (Plug&Play) Info (2.4)
476 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
478 ksyms Kernel symbol table
479 loadavg Load average of last 1, 5 & 15 minutes
483 modules List of loaded modules
484 mounts Mounted filesystems
485 net Networking info (see text)
486 pagetypeinfo Additional page allocator information (see text) (2.5)
487 partitions Table of partitions known to the system
488 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
489 decoupled by lspci (2.4)
491 scsi SCSI info (see text)
492 slabinfo Slab pool info
493 softirqs softirq usage
494 stat Overall statistics
495 swaps Swap space utilization
497 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
498 tty Info of tty drivers
500 version Kernel version
501 video bttv info of video resources (2.4)
502 vmallocinfo Show vmalloced areas
503 ..............................................................................
505 You can, for example, check which interrupts are currently in use and what
506 they are used for by looking in the file /proc/interrupts:
508 > cat /proc/interrupts
510 0: 8728810 XT-PIC timer
511 1: 895 XT-PIC keyboard
513 3: 531695 XT-PIC aha152x
514 4: 2014133 XT-PIC serial
515 5: 44401 XT-PIC pcnet_cs
518 12: 182918 XT-PIC PS/2 Mouse
520 14: 1232265 XT-PIC ide0
524 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
525 output of a SMP machine):
527 > cat /proc/interrupts
530 0: 1243498 1214548 IO-APIC-edge timer
531 1: 8949 8958 IO-APIC-edge keyboard
532 2: 0 0 XT-PIC cascade
533 5: 11286 10161 IO-APIC-edge soundblaster
534 8: 1 0 IO-APIC-edge rtc
535 9: 27422 27407 IO-APIC-edge 3c503
536 12: 113645 113873 IO-APIC-edge PS/2 Mouse
538 14: 22491 24012 IO-APIC-edge ide0
539 15: 2183 2415 IO-APIC-edge ide1
540 17: 30564 30414 IO-APIC-level eth0
541 18: 177 164 IO-APIC-level bttv
546 NMI is incremented in this case because every timer interrupt generates a NMI
547 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
549 LOC is the local interrupt counter of the internal APIC of every CPU.
551 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
552 connects the CPUs in a SMP system. This means that an error has been detected,
553 the IO-APIC automatically retry the transmission, so it should not be a big
554 problem, but you should read the SMP-FAQ.
556 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
557 /proc/interrupts to display every IRQ vector in use by the system, not
558 just those considered 'most important'. The new vectors are:
560 THR -- interrupt raised when a machine check threshold counter
561 (typically counting ECC corrected errors of memory or cache) exceeds
562 a configurable threshold. Only available on some systems.
564 TRM -- a thermal event interrupt occurs when a temperature threshold
565 has been exceeded for the CPU. This interrupt may also be generated
566 when the temperature drops back to normal.
568 SPU -- a spurious interrupt is some interrupt that was raised then lowered
569 by some IO device before it could be fully processed by the APIC. Hence
570 the APIC sees the interrupt but does not know what device it came from.
571 For this case the APIC will generate the interrupt with a IRQ vector
572 of 0xff. This might also be generated by chipset bugs.
574 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
575 sent from one CPU to another per the needs of the OS. Typically,
576 their statistics are used by kernel developers and interested users to
577 determine the occurrence of interrupts of the given type.
579 The above IRQ vectors are displayed only when relevant. For example,
580 the threshold vector does not exist on x86_64 platforms. Others are
581 suppressed when the system is a uniprocessor. As of this writing, only
582 i386 and x86_64 platforms support the new IRQ vector displays.
584 Of some interest is the introduction of the /proc/irq directory to 2.4.
585 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
586 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
587 irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and
592 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
593 1 11 13 15 17 19 3 5 7 9 default_smp_affinity
597 smp_affinity is a bitmask, in which you can specify which CPUs can handle the
598 IRQ, you can set it by doing:
600 > echo 1 > /proc/irq/10/smp_affinity
602 This means that only the first CPU will handle the IRQ, but you can also echo
603 5 which means that only the first and fourth CPU can handle the IRQ.
605 The contents of each smp_affinity file is the same by default:
607 > cat /proc/irq/0/smp_affinity
610 There is an alternate interface, smp_affinity_list which allows specifying
611 a cpu range instead of a bitmask:
613 > cat /proc/irq/0/smp_affinity_list
616 The default_smp_affinity mask applies to all non-active IRQs, which are the
617 IRQs which have not yet been allocated/activated, and hence which lack a
618 /proc/irq/[0-9]* directory.
620 The node file on an SMP system shows the node to which the device using the IRQ
621 reports itself as being attached. This hardware locality information does not
622 include information about any possible driver locality preference.
624 prof_cpu_mask specifies which CPUs are to be profiled by the system wide
625 profiler. Default value is ffffffff (all cpus if there are only 32 of them).
627 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
628 between all the CPUs which are allowed to handle it. As usual the kernel has
629 more info than you and does a better job than you, so the defaults are the
630 best choice for almost everyone. [Note this applies only to those IO-APIC's
631 that support "Round Robin" interrupt distribution.]
633 There are three more important subdirectories in /proc: net, scsi, and sys.
634 The general rule is that the contents, or even the existence of these
635 directories, depend on your kernel configuration. If SCSI is not enabled, the
636 directory scsi may not exist. The same is true with the net, which is there
637 only when networking support is present in the running kernel.
639 The slabinfo file gives information about memory usage at the slab level.
640 Linux uses slab pools for memory management above page level in version 2.2.
641 Commonly used objects have their own slab pool (such as network buffers,
642 directory cache, and so on).
644 ..............................................................................
646 > cat /proc/buddyinfo
648 Node 0, zone DMA 0 4 5 4 4 3 ...
649 Node 0, zone Normal 1 0 0 1 101 8 ...
650 Node 0, zone HighMem 2 0 0 1 1 0 ...
652 External fragmentation is a problem under some workloads, and buddyinfo is a
653 useful tool for helping diagnose these problems. Buddyinfo will give you a
654 clue as to how big an area you can safely allocate, or why a previous
657 Each column represents the number of pages of a certain order which are
658 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
659 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
660 available in ZONE_NORMAL, etc...
662 More information relevant to external fragmentation can be found in
665 > cat /proc/pagetypeinfo
669 Free pages count per migrate type at order 0 1 2 3 4 5 6 7 8 9 10
670 Node 0, zone DMA, type Unmovable 0 0 0 1 1 1 1 1 1 1 0
671 Node 0, zone DMA, type Reclaimable 0 0 0 0 0 0 0 0 0 0 0
672 Node 0, zone DMA, type Movable 1 1 2 1 2 1 1 0 1 0 2
673 Node 0, zone DMA, type Reserve 0 0 0 0 0 0 0 0 0 1 0
674 Node 0, zone DMA, type Isolate 0 0 0 0 0 0 0 0 0 0 0
675 Node 0, zone DMA32, type Unmovable 103 54 77 1 1 1 11 8 7 1 9
676 Node 0, zone DMA32, type Reclaimable 0 0 2 1 0 0 0 0 1 0 0
677 Node 0, zone DMA32, type Movable 169 152 113 91 77 54 39 13 6 1 452
678 Node 0, zone DMA32, type Reserve 1 2 2 2 2 0 1 1 1 1 0
679 Node 0, zone DMA32, type Isolate 0 0 0 0 0 0 0 0 0 0 0
681 Number of blocks type Unmovable Reclaimable Movable Reserve Isolate
682 Node 0, zone DMA 2 0 5 1 0
683 Node 0, zone DMA32 41 6 967 2 0
685 Fragmentation avoidance in the kernel works by grouping pages of different
686 migrate types into the same contiguous regions of memory called page blocks.
687 A page block is typically the size of the default hugepage size e.g. 2MB on
688 X86-64. By keeping pages grouped based on their ability to move, the kernel
689 can reclaim pages within a page block to satisfy a high-order allocation.
691 The pagetypinfo begins with information on the size of a page block. It
692 then gives the same type of information as buddyinfo except broken down
693 by migrate-type and finishes with details on how many page blocks of each
696 If min_free_kbytes has been tuned correctly (recommendations made by hugeadm
697 from libhugetlbfs http://sourceforge.net/projects/libhugetlbfs/), one can
698 make an estimate of the likely number of huge pages that can be allocated
699 at a given point in time. All the "Movable" blocks should be allocatable
700 unless memory has been mlock()'d. Some of the Reclaimable blocks should
701 also be allocatable although a lot of filesystem metadata may have to be
702 reclaimed to achieve this.
704 ..............................................................................
708 Provides information about distribution and utilization of memory. This
709 varies by architecture and compile options. The following is from a
710 16GB PIII, which has highmem enabled. You may not have all of these fields.
714 The "Locked" indicates whether the mapping is locked in memory or not.
717 MemTotal: 16344972 kB
724 HighTotal: 15597528 kB
725 HighFree: 13629632 kB
735 SReclaimable: 159856 kB
736 SUnreclaim: 124508 kB
741 CommitLimit: 7669796 kB
742 Committed_AS: 100056 kB
743 VmallocTotal: 112216 kB
745 VmallocChunk: 111088 kB
747 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
748 bits and the kernel binary code)
749 MemFree: The sum of LowFree+HighFree
750 Buffers: Relatively temporary storage for raw disk blocks
751 shouldn't get tremendously large (20MB or so)
752 Cached: in-memory cache for files read from the disk (the
753 pagecache). Doesn't include SwapCached
754 SwapCached: Memory that once was swapped out, is swapped back in but
755 still also is in the swapfile (if memory is needed it
756 doesn't need to be swapped out AGAIN because it is already
757 in the swapfile. This saves I/O)
758 Active: Memory that has been used more recently and usually not
759 reclaimed unless absolutely necessary.
760 Inactive: Memory which has been less recently used. It is more
761 eligible to be reclaimed for other purposes
763 HighFree: Highmem is all memory above ~860MB of physical memory
764 Highmem areas are for use by userspace programs, or
765 for the pagecache. The kernel must use tricks to access
766 this memory, making it slower to access than lowmem.
768 LowFree: Lowmem is memory which can be used for everything that
769 highmem can be used for, but it is also available for the
770 kernel's use for its own data structures. Among many
771 other things, it is where everything from the Slab is
772 allocated. Bad things happen when you're out of lowmem.
773 SwapTotal: total amount of swap space available
774 SwapFree: Memory which has been evicted from RAM, and is temporarily
776 Dirty: Memory which is waiting to get written back to the disk
777 Writeback: Memory which is actively being written back to the disk
778 AnonPages: Non-file backed pages mapped into userspace page tables
779 Mapped: files which have been mmaped, such as libraries
780 Slab: in-kernel data structures cache
781 SReclaimable: Part of Slab, that might be reclaimed, such as caches
782 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure
783 PageTables: amount of memory dedicated to the lowest level of page
785 NFS_Unstable: NFS pages sent to the server, but not yet committed to stable
787 Bounce: Memory used for block device "bounce buffers"
788 WritebackTmp: Memory used by FUSE for temporary writeback buffers
789 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
790 this is the total amount of memory currently available to
791 be allocated on the system. This limit is only adhered to
792 if strict overcommit accounting is enabled (mode 2 in
793 'vm.overcommit_memory').
794 The CommitLimit is calculated with the following formula:
795 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap
796 For example, on a system with 1G of physical RAM and 7G
797 of swap with a `vm.overcommit_ratio` of 30 it would
798 yield a CommitLimit of 7.3G.
799 For more details, see the memory overcommit documentation
800 in vm/overcommit-accounting.
801 Committed_AS: The amount of memory presently allocated on the system.
802 The committed memory is a sum of all of the memory which
803 has been allocated by processes, even if it has not been
804 "used" by them as of yet. A process which malloc()'s 1G
805 of memory, but only touches 300M of it will only show up
806 as using 300M of memory even if it has the address space
807 allocated for the entire 1G. This 1G is memory which has
808 been "committed" to by the VM and can be used at any time
809 by the allocating application. With strict overcommit
810 enabled on the system (mode 2 in 'vm.overcommit_memory'),
811 allocations which would exceed the CommitLimit (detailed
812 above) will not be permitted. This is useful if one needs
813 to guarantee that processes will not fail due to lack of
814 memory once that memory has been successfully allocated.
815 VmallocTotal: total size of vmalloc memory area
816 VmallocUsed: amount of vmalloc area which is used
817 VmallocChunk: largest contiguous block of vmalloc area which is free
819 ..............................................................................
823 Provides information about vmalloced/vmaped areas. One line per area,
824 containing the virtual address range of the area, size in bytes,
825 caller information of the creator, and optional information depending
826 on the kind of area :
828 pages=nr number of pages
829 phys=addr if a physical address was specified
830 ioremap I/O mapping (ioremap() and friends)
831 vmalloc vmalloc() area
834 vpages buffer for pages pointers was vmalloced (huge area)
835 N<node>=nr (Only on NUMA kernels)
836 Number of pages allocated on memory node <node>
838 > cat /proc/vmallocinfo
839 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ...
840 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128
841 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ...
842 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64
843 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f...
844 phys=7fee8000 ioremap
845 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f...
846 phys=7fee7000 ioremap
847 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210
848 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ...
849 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3
850 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ...
852 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ...
853 /0x130 [x_tables] pages=4 vmalloc N0=4
854 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ...
855 pages=14 vmalloc N2=14
856 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ...
858 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ...
860 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ...
861 pages=10 vmalloc N0=10
863 ..............................................................................
867 Provides counts of softirq handlers serviced since boot time, for each cpu.
872 TIMER: 27166 27120 27097 27034
877 SCHED: 27035 26983 26971 26746
879 RCU: 1678 1769 2178 2250
882 1.3 IDE devices in /proc/ide
883 ----------------------------
885 The subdirectory /proc/ide contains information about all IDE devices of which
886 the kernel is aware. There is one subdirectory for each IDE controller, the
887 file drivers and a link for each IDE device, pointing to the device directory
888 in the controller specific subtree.
890 The file drivers contains general information about the drivers used for the
893 > cat /proc/ide/drivers
894 ide-cdrom version 4.53
895 ide-disk version 1.08
897 More detailed information can be found in the controller specific
898 subdirectories. These are named ide0, ide1 and so on. Each of these
899 directories contains the files shown in table 1-6.
902 Table 1-6: IDE controller info in /proc/ide/ide?
903 ..............................................................................
905 channel IDE channel (0 or 1)
906 config Configuration (only for PCI/IDE bridge)
908 model Type/Chipset of IDE controller
909 ..............................................................................
911 Each device connected to a controller has a separate subdirectory in the
912 controllers directory. The files listed in table 1-7 are contained in these
916 Table 1-7: IDE device information
917 ..............................................................................
920 capacity Capacity of the medium (in 512Byte blocks)
921 driver driver and version
922 geometry physical and logical geometry
923 identify device identify block
925 model device identifier
926 settings device setup
927 smart_thresholds IDE disk management thresholds
928 smart_values IDE disk management values
929 ..............................................................................
931 The most interesting file is settings. This file contains a nice overview of
932 the drive parameters:
934 # cat /proc/ide/ide0/hda/settings
935 name value min max mode
936 ---- ----- --- --- ----
937 bios_cyl 526 0 65535 rw
938 bios_head 255 0 255 rw
940 breada_readahead 4 0 127 rw
942 file_readahead 72 0 2097151 rw
944 keepsettings 0 0 1 rw
945 max_kb_per_request 122 1 127 rw
949 pio_mode write-only 0 255 w
955 1.4 Networking info in /proc/net
956 --------------------------------
958 The subdirectory /proc/net follows the usual pattern. Table 1-8 shows the
959 additional values you get for IP version 6 if you configure the kernel to
960 support this. Table 1-9 lists the files and their meaning.
963 Table 1-8: IPv6 info in /proc/net
964 ..............................................................................
966 udp6 UDP sockets (IPv6)
967 tcp6 TCP sockets (IPv6)
968 raw6 Raw device statistics (IPv6)
969 igmp6 IP multicast addresses, which this host joined (IPv6)
970 if_inet6 List of IPv6 interface addresses
971 ipv6_route Kernel routing table for IPv6
972 rt6_stats Global IPv6 routing tables statistics
973 sockstat6 Socket statistics (IPv6)
974 snmp6 Snmp data (IPv6)
975 ..............................................................................
978 Table 1-9: Network info in /proc/net
979 ..............................................................................
982 dev network devices with statistics
983 dev_mcast the Layer2 multicast groups a device is listening too
984 (interface index, label, number of references, number of bound
986 dev_stat network device status
987 ip_fwchains Firewall chain linkage
988 ip_fwnames Firewall chain names
989 ip_masq Directory containing the masquerading tables
990 ip_masquerade Major masquerading table
991 netstat Network statistics
992 raw raw device statistics
993 route Kernel routing table
994 rpc Directory containing rpc info
995 rt_cache Routing cache
997 sockstat Socket statistics
999 tr_rif Token ring RIF routing table
1001 unix UNIX domain sockets
1002 wireless Wireless interface data (Wavelan etc)
1003 igmp IP multicast addresses, which this host joined
1004 psched Global packet scheduler parameters.
1005 netlink List of PF_NETLINK sockets
1006 ip_mr_vifs List of multicast virtual interfaces
1007 ip_mr_cache List of multicast routing cache
1008 ..............................................................................
1010 You can use this information to see which network devices are available in
1011 your system and how much traffic was routed over those devices:
1014 Inter-|Receive |[...
1015 face |bytes packets errs drop fifo frame compressed multicast|[...
1016 lo: 908188 5596 0 0 0 0 0 0 [...
1017 ppp0:15475140 20721 410 0 0 410 0 0 [...
1018 eth0: 614530 7085 0 0 0 0 0 1 [...
1021 ...] bytes packets errs drop fifo colls carrier compressed
1022 ...] 908188 5596 0 0 0 0 0 0
1023 ...] 1375103 17405 0 0 0 0 0 0
1024 ...] 1703981 5535 0 0 0 3 0 0
1026 In addition, each Channel Bond interface has its own directory. For
1027 example, the bond0 device will have a directory called /proc/net/bond0/.
1028 It will contain information that is specific to that bond, such as the
1029 current slaves of the bond, the link status of the slaves, and how
1030 many times the slaves link has failed.
1035 If you have a SCSI host adapter in your system, you'll find a subdirectory
1036 named after the driver for this adapter in /proc/scsi. You'll also see a list
1037 of all recognized SCSI devices in /proc/scsi:
1039 >cat /proc/scsi/scsi
1041 Host: scsi0 Channel: 00 Id: 00 Lun: 00
1042 Vendor: IBM Model: DGHS09U Rev: 03E0
1043 Type: Direct-Access ANSI SCSI revision: 03
1044 Host: scsi0 Channel: 00 Id: 06 Lun: 00
1045 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
1046 Type: CD-ROM ANSI SCSI revision: 02
1049 The directory named after the driver has one file for each adapter found in
1050 the system. These files contain information about the controller, including
1051 the used IRQ and the IO address range. The amount of information shown is
1052 dependent on the adapter you use. The example shows the output for an Adaptec
1053 AHA-2940 SCSI adapter:
1055 > cat /proc/scsi/aic7xxx/0
1057 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
1059 TCQ Enabled By Default : Disabled
1060 AIC7XXX_PROC_STATS : Disabled
1061 AIC7XXX_RESET_DELAY : 5
1062 Adapter Configuration:
1063 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
1064 Ultra Wide Controller
1065 PCI MMAPed I/O Base: 0xeb001000
1066 Adapter SEEPROM Config: SEEPROM found and used.
1067 Adaptec SCSI BIOS: Enabled
1069 SCBs: Active 0, Max Active 2,
1070 Allocated 15, HW 16, Page 255
1072 BIOS Control Word: 0x18b6
1073 Adapter Control Word: 0x005b
1074 Extended Translation: Enabled
1075 Disconnect Enable Flags: 0xffff
1076 Ultra Enable Flags: 0x0001
1077 Tag Queue Enable Flags: 0x0000
1078 Ordered Queue Tag Flags: 0x0000
1079 Default Tag Queue Depth: 8
1080 Tagged Queue By Device array for aic7xxx host instance 0:
1081 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
1082 Actual queue depth per device for aic7xxx host instance 0:
1083 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
1086 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
1087 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
1088 Total transfers 160151 (74577 reads and 85574 writes)
1090 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
1091 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
1092 Total transfers 0 (0 reads and 0 writes)
1095 1.6 Parallel port info in /proc/parport
1096 ---------------------------------------
1098 The directory /proc/parport contains information about the parallel ports of
1099 your system. It has one subdirectory for each port, named after the port
1102 These directories contain the four files shown in Table 1-10.
1105 Table 1-10: Files in /proc/parport
1106 ..............................................................................
1108 autoprobe Any IEEE-1284 device ID information that has been acquired.
1109 devices list of the device drivers using that port. A + will appear by the
1110 name of the device currently using the port (it might not appear
1112 hardware Parallel port's base address, IRQ line and DMA channel.
1113 irq IRQ that parport is using for that port. This is in a separate
1114 file to allow you to alter it by writing a new value in (IRQ
1116 ..............................................................................
1118 1.7 TTY info in /proc/tty
1119 -------------------------
1121 Information about the available and actually used tty's can be found in the
1122 directory /proc/tty.You'll find entries for drivers and line disciplines in
1123 this directory, as shown in Table 1-11.
1126 Table 1-11: Files in /proc/tty
1127 ..............................................................................
1129 drivers list of drivers and their usage
1130 ldiscs registered line disciplines
1131 driver/serial usage statistic and status of single tty lines
1132 ..............................................................................
1134 To see which tty's are currently in use, you can simply look into the file
1137 > cat /proc/tty/drivers
1138 pty_slave /dev/pts 136 0-255 pty:slave
1139 pty_master /dev/ptm 128 0-255 pty:master
1140 pty_slave /dev/ttyp 3 0-255 pty:slave
1141 pty_master /dev/pty 2 0-255 pty:master
1142 serial /dev/cua 5 64-67 serial:callout
1143 serial /dev/ttyS 4 64-67 serial
1144 /dev/tty0 /dev/tty0 4 0 system:vtmaster
1145 /dev/ptmx /dev/ptmx 5 2 system
1146 /dev/console /dev/console 5 1 system:console
1147 /dev/tty /dev/tty 5 0 system:/dev/tty
1148 unknown /dev/tty 4 1-63 console
1151 1.8 Miscellaneous kernel statistics in /proc/stat
1152 -------------------------------------------------
1154 Various pieces of information about kernel activity are available in the
1155 /proc/stat file. All of the numbers reported in this file are aggregates
1156 since the system first booted. For a quick look, simply cat the file:
1159 cpu 2255 34 2290 22625563 6290 127 456 0 0
1160 cpu0 1132 34 1441 11311718 3675 127 438 0 0
1161 cpu1 1123 0 849 11313845 2614 0 18 0 0
1162 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
1168 softirq 183433 0 21755 12 39 1137 231 21459 2263
1170 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
1171 lines. These numbers identify the amount of time the CPU has spent performing
1172 different kinds of work. Time units are in USER_HZ (typically hundredths of a
1173 second). The meanings of the columns are as follows, from left to right:
1175 - user: normal processes executing in user mode
1176 - nice: niced processes executing in user mode
1177 - system: processes executing in kernel mode
1178 - idle: twiddling thumbs
1179 - iowait: waiting for I/O to complete
1180 - irq: servicing interrupts
1181 - softirq: servicing softirqs
1182 - steal: involuntary wait
1183 - guest: running a normal guest
1184 - guest_nice: running a niced guest
1186 The "intr" line gives counts of interrupts serviced since boot time, for each
1187 of the possible system interrupts. The first column is the total of all
1188 interrupts serviced; each subsequent column is the total for that particular
1191 The "ctxt" line gives the total number of context switches across all CPUs.
1193 The "btime" line gives the time at which the system booted, in seconds since
1196 The "processes" line gives the number of processes and threads created, which
1197 includes (but is not limited to) those created by calls to the fork() and
1198 clone() system calls.
1200 The "procs_running" line gives the total number of threads that are
1201 running or ready to run (i.e., the total number of runnable threads).
1203 The "procs_blocked" line gives the number of processes currently blocked,
1204 waiting for I/O to complete.
1206 The "softirq" line gives counts of softirqs serviced since boot time, for each
1207 of the possible system softirqs. The first column is the total of all
1208 softirqs serviced; each subsequent column is the total for that particular
1212 1.9 Ext4 file system parameters
1213 ------------------------------
1215 Information about mounted ext4 file systems can be found in
1216 /proc/fs/ext4. Each mounted filesystem will have a directory in
1217 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
1218 /proc/fs/ext4/dm-0). The files in each per-device directory are shown
1219 in Table 1-12, below.
1221 Table 1-12: Files in /proc/fs/ext4/<devname>
1222 ..............................................................................
1224 mb_groups details of multiblock allocator buddy cache of free blocks
1225 ..............................................................................
1229 Shows registered system console lines.
1231 To see which character device lines are currently used for the system console
1232 /dev/console, you may simply look into the file /proc/consoles:
1234 > cat /proc/consoles
1240 device name of the device
1241 operations R = can do read operations
1242 W = can do write operations
1244 flags E = it is enabled
1245 C = it is preferred console
1246 B = it is primary boot console
1247 p = it is used for printk buffer
1248 b = it is not a TTY but a Braille device
1249 a = it is safe to use when cpu is offline
1250 major:minor major and minor number of the device separated by a colon
1252 ------------------------------------------------------------------------------
1254 ------------------------------------------------------------------------------
1255 The /proc file system serves information about the running system. It not only
1256 allows access to process data but also allows you to request the kernel status
1257 by reading files in the hierarchy.
1259 The directory structure of /proc reflects the types of information and makes
1260 it easy, if not obvious, where to look for specific data.
1261 ------------------------------------------------------------------------------
1263 ------------------------------------------------------------------------------
1264 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
1265 ------------------------------------------------------------------------------
1267 ------------------------------------------------------------------------------
1269 ------------------------------------------------------------------------------
1270 * Modifying kernel parameters by writing into files found in /proc/sys
1271 * Exploring the files which modify certain parameters
1272 * Review of the /proc/sys file tree
1273 ------------------------------------------------------------------------------
1276 A very interesting part of /proc is the directory /proc/sys. This is not only
1277 a source of information, it also allows you to change parameters within the
1278 kernel. Be very careful when attempting this. You can optimize your system,
1279 but you can also cause it to crash. Never alter kernel parameters on a
1280 production system. Set up a development machine and test to make sure that
1281 everything works the way you want it to. You may have no alternative but to
1282 reboot the machine once an error has been made.
1284 To change a value, simply echo the new value into the file. An example is
1285 given below in the section on the file system data. You need to be root to do
1286 this. You can create your own boot script to perform this every time your
1289 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
1290 general things in the operation of the Linux kernel. Since some of the files
1291 can inadvertently disrupt your system, it is advisable to read both
1292 documentation and source before actually making adjustments. In any case, be
1293 very careful when writing to any of these files. The entries in /proc may
1294 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
1295 review the kernel documentation in the directory /usr/src/linux/Documentation.
1296 This chapter is heavily based on the documentation included in the pre 2.2
1297 kernels, and became part of it in version 2.2.1 of the Linux kernel.
1299 Please see: Documentation/sysctl/ directory for descriptions of these
1302 ------------------------------------------------------------------------------
1304 ------------------------------------------------------------------------------
1305 Certain aspects of kernel behavior can be modified at runtime, without the
1306 need to recompile the kernel, or even to reboot the system. The files in the
1307 /proc/sys tree can not only be read, but also modified. You can use the echo
1308 command to write value into these files, thereby changing the default settings
1310 ------------------------------------------------------------------------------
1312 ------------------------------------------------------------------------------
1313 CHAPTER 3: PER-PROCESS PARAMETERS
1314 ------------------------------------------------------------------------------
1316 3.1 /proc/<pid>/oom_adj & /proc/<pid>/oom_score_adj- Adjust the oom-killer score
1317 --------------------------------------------------------------------------------
1319 These file can be used to adjust the badness heuristic used to select which
1320 process gets killed in out of memory conditions.
1322 The badness heuristic assigns a value to each candidate task ranging from 0
1323 (never kill) to 1000 (always kill) to determine which process is targeted. The
1324 units are roughly a proportion along that range of allowed memory the process
1325 may allocate from based on an estimation of its current memory and swap use.
1326 For example, if a task is using all allowed memory, its badness score will be
1327 1000. If it is using half of its allowed memory, its score will be 500.
1329 There is an additional factor included in the badness score: root
1330 processes are given 3% extra memory over other tasks.
1332 The amount of "allowed" memory depends on the context in which the oom killer
1333 was called. If it is due to the memory assigned to the allocating task's cpuset
1334 being exhausted, the allowed memory represents the set of mems assigned to that
1335 cpuset. If it is due to a mempolicy's node(s) being exhausted, the allowed
1336 memory represents the set of mempolicy nodes. If it is due to a memory
1337 limit (or swap limit) being reached, the allowed memory is that configured
1338 limit. Finally, if it is due to the entire system being out of memory, the
1339 allowed memory represents all allocatable resources.
1341 The value of /proc/<pid>/oom_score_adj is added to the badness score before it
1342 is used to determine which task to kill. Acceptable values range from -1000
1343 (OOM_SCORE_ADJ_MIN) to +1000 (OOM_SCORE_ADJ_MAX). This allows userspace to
1344 polarize the preference for oom killing either by always preferring a certain
1345 task or completely disabling it. The lowest possible value, -1000, is
1346 equivalent to disabling oom killing entirely for that task since it will always
1347 report a badness score of 0.
1349 Consequently, it is very simple for userspace to define the amount of memory to
1350 consider for each task. Setting a /proc/<pid>/oom_score_adj value of +500, for
1351 example, is roughly equivalent to allowing the remainder of tasks sharing the
1352 same system, cpuset, mempolicy, or memory controller resources to use at least
1353 50% more memory. A value of -500, on the other hand, would be roughly
1354 equivalent to discounting 50% of the task's allowed memory from being considered
1355 as scoring against the task.
1357 For backwards compatibility with previous kernels, /proc/<pid>/oom_adj may also
1358 be used to tune the badness score. Its acceptable values range from -16
1359 (OOM_ADJUST_MIN) to +15 (OOM_ADJUST_MAX) and a special value of -17
1360 (OOM_DISABLE) to disable oom killing entirely for that task. Its value is
1361 scaled linearly with /proc/<pid>/oom_score_adj.
1363 Writing to /proc/<pid>/oom_score_adj or /proc/<pid>/oom_adj will change the
1364 other with its scaled value.
1366 The value of /proc/<pid>/oom_score_adj may be reduced no lower than the last
1367 value set by a CAP_SYS_RESOURCE process. To reduce the value any lower
1368 requires CAP_SYS_RESOURCE.
1370 NOTICE: /proc/<pid>/oom_adj is deprecated and will be removed, please see
1371 Documentation/feature-removal-schedule.txt.
1373 Caveat: when a parent task is selected, the oom killer will sacrifice any first
1374 generation children with separate address spaces instead, if possible. This
1375 avoids servers and important system daemons from being killed and loses the
1376 minimal amount of work.
1379 3.2 /proc/<pid>/oom_score - Display current oom-killer score
1380 -------------------------------------------------------------
1382 This file can be used to check the current score used by the oom-killer is for
1383 any given <pid>. Use it together with /proc/<pid>/oom_adj to tune which
1384 process should be killed in an out-of-memory situation.
1387 3.3 /proc/<pid>/io - Display the IO accounting fields
1388 -------------------------------------------------------
1390 This file contains IO statistics for each running process
1395 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
1398 test:/tmp # cat /proc/3828/io
1404 write_bytes: 323932160
1405 cancelled_write_bytes: 0
1414 I/O counter: chars read
1415 The number of bytes which this task has caused to be read from storage. This
1416 is simply the sum of bytes which this process passed to read() and pread().
1417 It includes things like tty IO and it is unaffected by whether or not actual
1418 physical disk IO was required (the read might have been satisfied from
1425 I/O counter: chars written
1426 The number of bytes which this task has caused, or shall cause to be written
1427 to disk. Similar caveats apply here as with rchar.
1433 I/O counter: read syscalls
1434 Attempt to count the number of read I/O operations, i.e. syscalls like read()
1441 I/O counter: write syscalls
1442 Attempt to count the number of write I/O operations, i.e. syscalls like
1443 write() and pwrite().
1449 I/O counter: bytes read
1450 Attempt to count the number of bytes which this process really did cause to
1451 be fetched from the storage layer. Done at the submit_bio() level, so it is
1452 accurate for block-backed filesystems. <please add status regarding NFS and
1453 CIFS at a later time>
1459 I/O counter: bytes written
1460 Attempt to count the number of bytes which this process caused to be sent to
1461 the storage layer. This is done at page-dirtying time.
1464 cancelled_write_bytes
1465 ---------------------
1467 The big inaccuracy here is truncate. If a process writes 1MB to a file and
1468 then deletes the file, it will in fact perform no writeout. But it will have
1469 been accounted as having caused 1MB of write.
1470 In other words: The number of bytes which this process caused to not happen,
1471 by truncating pagecache. A task can cause "negative" IO too. If this task
1472 truncates some dirty pagecache, some IO which another task has been accounted
1473 for (in its write_bytes) will not be happening. We _could_ just subtract that
1474 from the truncating task's write_bytes, but there is information loss in doing
1481 At its current implementation state, this is a bit racy on 32-bit machines: if
1482 process A reads process B's /proc/pid/io while process B is updating one of
1483 those 64-bit counters, process A could see an intermediate result.
1486 More information about this can be found within the taskstats documentation in
1487 Documentation/accounting.
1489 3.4 /proc/<pid>/coredump_filter - Core dump filtering settings
1490 ---------------------------------------------------------------
1491 When a process is dumped, all anonymous memory is written to a core file as
1492 long as the size of the core file isn't limited. But sometimes we don't want
1493 to dump some memory segments, for example, huge shared memory. Conversely,
1494 sometimes we want to save file-backed memory segments into a core file, not
1495 only the individual files.
1497 /proc/<pid>/coredump_filter allows you to customize which memory segments
1498 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
1499 of memory types. If a bit of the bitmask is set, memory segments of the
1500 corresponding memory type are dumped, otherwise they are not dumped.
1502 The following 7 memory types are supported:
1503 - (bit 0) anonymous private memory
1504 - (bit 1) anonymous shared memory
1505 - (bit 2) file-backed private memory
1506 - (bit 3) file-backed shared memory
1507 - (bit 4) ELF header pages in file-backed private memory areas (it is
1508 effective only if the bit 2 is cleared)
1509 - (bit 5) hugetlb private memory
1510 - (bit 6) hugetlb shared memory
1512 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
1513 are always dumped regardless of the bitmask status.
1515 Note bit 0-4 doesn't effect any hugetlb memory. hugetlb memory are only
1516 effected by bit 5-6.
1518 Default value of coredump_filter is 0x23; this means all anonymous memory
1519 segments and hugetlb private memory are dumped.
1521 If you don't want to dump all shared memory segments attached to pid 1234,
1522 write 0x21 to the process's proc file.
1524 $ echo 0x21 > /proc/1234/coredump_filter
1526 When a new process is created, the process inherits the bitmask status from its
1527 parent. It is useful to set up coredump_filter before the program runs.
1530 $ echo 0x7 > /proc/self/coredump_filter
1533 3.5 /proc/<pid>/mountinfo - Information about mounts
1534 --------------------------------------------------------
1536 This file contains lines of the form:
1538 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
1539 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
1541 (1) mount ID: unique identifier of the mount (may be reused after umount)
1542 (2) parent ID: ID of parent (or of self for the top of the mount tree)
1543 (3) major:minor: value of st_dev for files on filesystem
1544 (4) root: root of the mount within the filesystem
1545 (5) mount point: mount point relative to the process's root
1546 (6) mount options: per mount options
1547 (7) optional fields: zero or more fields of the form "tag[:value]"
1548 (8) separator: marks the end of the optional fields
1549 (9) filesystem type: name of filesystem of the form "type[.subtype]"
1550 (10) mount source: filesystem specific information or "none"
1551 (11) super options: per super block options
1553 Parsers should ignore all unrecognised optional fields. Currently the
1554 possible optional fields are:
1556 shared:X mount is shared in peer group X
1557 master:X mount is slave to peer group X
1558 propagate_from:X mount is slave and receives propagation from peer group X (*)
1559 unbindable mount is unbindable
1561 (*) X is the closest dominant peer group under the process's root. If
1562 X is the immediate master of the mount, or if there's no dominant peer
1563 group under the same root, then only the "master:X" field is present
1564 and not the "propagate_from:X" field.
1566 For more information on mount propagation see:
1568 Documentation/filesystems/sharedsubtree.txt
1571 3.6 /proc/<pid>/comm & /proc/<pid>/task/<tid>/comm
1572 --------------------------------------------------------
1573 These files provide a method to access a tasks comm value. It also allows for
1574 a task to set its own or one of its thread siblings comm value. The comm value
1575 is limited in size compared to the cmdline value, so writing anything longer
1576 then the kernel's TASK_COMM_LEN (currently 16 chars) will result in a truncated
1580 ------------------------------------------------------------------------------
1582 ------------------------------------------------------------------------------
1585 ---------------------
1587 The following mount options are supported:
1589 hidepid= Set /proc/<pid>/ access mode.
1590 gid= Set the group authorized to learn processes information.
1592 hidepid=0 means classic mode - everybody may access all /proc/<pid>/ directories
1595 hidepid=1 means users may not access any /proc/<pid>/ directories but their
1596 own. Sensitive files like cmdline, sched*, status are now protected against
1597 other users. This makes it impossible to learn whether any user runs
1598 specific program (given the program doesn't reveal itself by its behaviour).
1599 As an additional bonus, as /proc/<pid>/cmdline is unaccessible for other users,
1600 poorly written programs passing sensitive information via program arguments are
1601 now protected against local eavesdroppers.
1603 hidepid=2 means hidepid=1 plus all /proc/<pid>/ will be fully invisible to other
1604 users. It doesn't mean that it hides a fact whether a process with a specific
1605 pid value exists (it can be learned by other means, e.g. by "kill -0 $PID"),
1606 but it hides process' uid and gid, which may be learned by stat()'ing
1607 /proc/<pid>/ otherwise. It greatly complicates an intruder's task of gathering
1608 information about running processes, whether some daemon runs with elevated
1609 privileges, whether other user runs some sensitive program, whether other users
1610 run any program at all, etc.
1612 gid= defines a group authorized to learn processes information otherwise
1613 prohibited by hidepid=. If you use some daemon like identd which needs to learn
1614 information about processes information, just add identd to this group.
projects / deliverable / linux.git / blob