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 ------------------------------------------------------------------------------
9 Version 1.3 Kernel version 2.2.12
10 Kernel version 2.4.0-test11-pre4
11 ------------------------------------------------------------------------------
17 0.1 Introduction/Credits
20 1 Collecting System Information
21 1.1 Process-Specific Subdirectories
23 1.3 IDE devices in /proc/ide
24 1.4 Networking info in /proc/net
26 1.6 Parallel port info in /proc/parport
27 1.7 TTY info in /proc/tty
28 1.8 Miscellaneous kernel statistics in /proc/stat
30 2 Modifying System Parameters
31 2.1 /proc/sys/fs - File system data
32 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
33 2.3 /proc/sys/kernel - general kernel parameters
34 2.4 /proc/sys/vm - The virtual memory subsystem
35 2.5 /proc/sys/dev - Device specific parameters
36 2.6 /proc/sys/sunrpc - Remote procedure calls
37 2.7 /proc/sys/net - Networking stuff
38 2.8 /proc/sys/net/ipv4 - IPV4 settings
41 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
42 2.12 /proc/<pid>/oom_adj - Adjust the oom-killer score
43 2.13 /proc/<pid>/oom_score - Display current oom-killer score
45 ------------------------------------------------------------------------------
47 ------------------------------------------------------------------------------
49 0.1 Introduction/Credits
50 ------------------------
52 This documentation is part of a soon (or so we hope) to be released book on
53 the SuSE Linux distribution. As there is no complete documentation for the
54 /proc file system and we've used many freely available sources to write these
55 chapters, it seems only fair to give the work back to the Linux community.
56 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
57 afraid it's still far from complete, but we hope it will be useful. As far as
58 we know, it is the first 'all-in-one' document about the /proc file system. It
59 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
60 SPARC, AXP, etc., features, you probably won't find what you are looking for.
61 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
62 additions and patches are welcome and will be added to this document if you
65 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
66 other people for help compiling this documentation. We'd also like to extend a
67 special thank you to Andi Kleen for documentation, which we relied on heavily
68 to create this document, as well as the additional information he provided.
69 Thanks to everybody else who contributed source or docs to the Linux kernel
70 and helped create a great piece of software... :)
72 If you have any comments, corrections or additions, please don't hesitate to
73 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
76 The latest version of this document is available online at
77 http://skaro.nightcrawler.com/~bb/Docs/Proc as HTML version.
79 If the above direction does not works for you, ypu could try the kernel
80 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
81 comandante@zaralinux.com.
86 We don't guarantee the correctness of this document, and if you come to us
87 complaining about how you screwed up your system because of incorrect
88 documentation, we won't feel responsible...
90 ------------------------------------------------------------------------------
91 CHAPTER 1: COLLECTING SYSTEM INFORMATION
92 ------------------------------------------------------------------------------
94 ------------------------------------------------------------------------------
96 ------------------------------------------------------------------------------
97 * Investigating the properties of the pseudo file system /proc and its
98 ability to provide information on the running Linux system
99 * Examining /proc's structure
100 * Uncovering various information about the kernel and the processes running
102 ------------------------------------------------------------------------------
105 The proc file system acts as an interface to internal data structures in the
106 kernel. It can be used to obtain information about the system and to change
107 certain kernel parameters at runtime (sysctl).
109 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
110 show you how you can use /proc/sys to change settings.
112 1.1 Process-Specific Subdirectories
113 -----------------------------------
115 The directory /proc contains (among other things) one subdirectory for each
116 process running on the system, which is named after the process ID (PID).
118 The link self points to the process reading the file system. Each process
119 subdirectory has the entries listed in Table 1-1.
122 Table 1-1: Process specific entries in /proc
123 ..............................................................................
125 cmdline Command line arguments
126 cpu Current and last cpu in which it was executed (2.4)(smp)
127 cwd Link to the current working directory
128 environ Values of environment variables
129 exe Link to the executable of this process
130 fd Directory, which contains all file descriptors
131 maps Memory maps to executables and library files (2.4)
132 mem Memory held by this process
133 root Link to the root directory of this process
135 statm Process memory status information
136 status Process status in human readable form
137 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan
138 smaps Extension based on maps, presenting the rss size for each mapped file
139 ..............................................................................
141 For example, to get the status information of a process, all you have to do is
142 read the file /proc/PID/status:
144 >cat /proc/self/status
160 SigPnd: 0000000000000000
161 SigBlk: 0000000000000000
162 SigIgn: 0000000000000000
163 SigCgt: 0000000000000000
164 CapInh: 00000000fffffeff
165 CapPrm: 0000000000000000
166 CapEff: 0000000000000000
169 This shows you nearly the same information you would get if you viewed it with
170 the ps command. In fact, ps uses the proc file system to obtain its
171 information. The statm file contains more detailed information about the
172 process memory usage. Its seven fields are explained in Table 1-2.
175 Table 1-2: Contents of the statm files (as of 2.6.8-rc3)
176 ..............................................................................
178 size total program size (pages) (same as VmSize in status)
179 resident size of memory portions (pages) (same as VmRSS in status)
180 shared number of pages that are shared (i.e. backed by a file)
181 trs number of pages that are 'code' (not including libs; broken,
182 includes data segment)
183 lrs number of pages of library (always 0 on 2.6)
184 drs number of pages of data/stack (including libs; broken,
185 includes library text)
186 dt number of dirty pages (always 0 on 2.6)
187 ..............................................................................
192 Similar to the process entries, the kernel data files give information about
193 the running kernel. The files used to obtain this information are contained in
194 /proc and are listed in Table 1-3. Not all of these will be present in your
195 system. It depends on the kernel configuration and the loaded modules, which
196 files are there, and which are missing.
198 Table 1-3: Kernel info in /proc
199 ..............................................................................
201 apm Advanced power management info
202 buddyinfo Kernel memory allocator information (see text) (2.5)
203 bus Directory containing bus specific information
204 cmdline Kernel command line
205 cpuinfo Info about the CPU
206 devices Available devices (block and character)
207 dma Used DMS channels
208 filesystems Supported filesystems
209 driver Various drivers grouped here, currently rtc (2.4)
210 execdomains Execdomains, related to security (2.4)
211 fb Frame Buffer devices (2.4)
212 fs File system parameters, currently nfs/exports (2.4)
213 ide Directory containing info about the IDE subsystem
214 interrupts Interrupt usage
215 iomem Memory map (2.4)
216 ioports I/O port usage
217 irq Masks for irq to cpu affinity (2.4)(smp?)
218 isapnp ISA PnP (Plug&Play) Info (2.4)
219 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
221 ksyms Kernel symbol table
222 loadavg Load average of last 1, 5 & 15 minutes
226 modules List of loaded modules
227 mounts Mounted filesystems
228 net Networking info (see text)
229 partitions Table of partitions known to the system
230 pci Depreciated info of PCI bus (new way -> /proc/bus/pci/,
231 decoupled by lspci (2.4)
233 scsi SCSI info (see text)
234 slabinfo Slab pool info
235 stat Overall statistics
236 swaps Swap space utilization
238 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
239 tty Info of tty drivers
241 version Kernel version
242 video bttv info of video resources (2.4)
243 ..............................................................................
245 You can, for example, check which interrupts are currently in use and what
246 they are used for by looking in the file /proc/interrupts:
248 > cat /proc/interrupts
250 0: 8728810 XT-PIC timer
251 1: 895 XT-PIC keyboard
253 3: 531695 XT-PIC aha152x
254 4: 2014133 XT-PIC serial
255 5: 44401 XT-PIC pcnet_cs
258 12: 182918 XT-PIC PS/2 Mouse
260 14: 1232265 XT-PIC ide0
264 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
265 output of a SMP machine):
267 > cat /proc/interrupts
270 0: 1243498 1214548 IO-APIC-edge timer
271 1: 8949 8958 IO-APIC-edge keyboard
272 2: 0 0 XT-PIC cascade
273 5: 11286 10161 IO-APIC-edge soundblaster
274 8: 1 0 IO-APIC-edge rtc
275 9: 27422 27407 IO-APIC-edge 3c503
276 12: 113645 113873 IO-APIC-edge PS/2 Mouse
278 14: 22491 24012 IO-APIC-edge ide0
279 15: 2183 2415 IO-APIC-edge ide1
280 17: 30564 30414 IO-APIC-level eth0
281 18: 177 164 IO-APIC-level bttv
286 NMI is incremented in this case because every timer interrupt generates a NMI
287 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
289 LOC is the local interrupt counter of the internal APIC of every CPU.
291 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
292 connects the CPUs in a SMP system. This means that an error has been detected,
293 the IO-APIC automatically retry the transmission, so it should not be a big
294 problem, but you should read the SMP-FAQ.
296 In this context it could be interesting to note the new irq directory in 2.4.
297 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
298 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
299 irq subdir is one subdir for each IRQ, and one file; prof_cpu_mask
303 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
304 1 11 13 15 17 19 3 5 7 9
308 The contents of the prof_cpu_mask file and each smp_affinity file for each IRQ
309 is the same by default:
311 > cat /proc/irq/0/smp_affinity
314 It's a bitmask, in which you can specify which CPUs can handle the IRQ, you can
317 > echo 1 > /proc/irq/prof_cpu_mask
319 This means that only the first CPU will handle the IRQ, but you can also echo 5
320 which means that only the first and fourth CPU can handle the IRQ.
322 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
323 between all the CPUs which are allowed to handle it. As usual the kernel has
324 more info than you and does a better job than you, so the defaults are the
325 best choice for almost everyone.
327 There are three more important subdirectories in /proc: net, scsi, and sys.
328 The general rule is that the contents, or even the existence of these
329 directories, depend on your kernel configuration. If SCSI is not enabled, the
330 directory scsi may not exist. The same is true with the net, which is there
331 only when networking support is present in the running kernel.
333 The slabinfo file gives information about memory usage at the slab level.
334 Linux uses slab pools for memory management above page level in version 2.2.
335 Commonly used objects have their own slab pool (such as network buffers,
336 directory cache, and so on).
338 ..............................................................................
340 > cat /proc/buddyinfo
342 Node 0, zone DMA 0 4 5 4 4 3 ...
343 Node 0, zone Normal 1 0 0 1 101 8 ...
344 Node 0, zone HighMem 2 0 0 1 1 0 ...
346 Memory fragmentation is a problem under some workloads, and buddyinfo is a
347 useful tool for helping diagnose these problems. Buddyinfo will give you a
348 clue as to how big an area you can safely allocate, or why a previous
351 Each column represents the number of pages of a certain order which are
352 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
353 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
354 available in ZONE_NORMAL, etc...
356 ..............................................................................
360 Provides information about distribution and utilization of memory. This
361 varies by architecture and compile options. The following is from a
362 16GB PIII, which has highmem enabled. You may not have all of these fields.
367 MemTotal: 16344972 kB
374 HighTotal: 15597528 kB
375 HighFree: 13629632 kB
384 CommitLimit: 7669796 kB
385 Committed_AS: 100056 kB
387 VmallocTotal: 112216 kB
389 VmallocChunk: 111088 kB
391 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
392 bits and the kernel binary code)
393 MemFree: The sum of LowFree+HighFree
394 Buffers: Relatively temporary storage for raw disk blocks
395 shouldn't get tremendously large (20MB or so)
396 Cached: in-memory cache for files read from the disk (the
397 pagecache). Doesn't include SwapCached
398 SwapCached: Memory that once was swapped out, is swapped back in but
399 still also is in the swapfile (if memory is needed it
400 doesn't need to be swapped out AGAIN because it is already
401 in the swapfile. This saves I/O)
402 Active: Memory that has been used more recently and usually not
403 reclaimed unless absolutely necessary.
404 Inactive: Memory which has been less recently used. It is more
405 eligible to be reclaimed for other purposes
407 HighFree: Highmem is all memory above ~860MB of physical memory
408 Highmem areas are for use by userspace programs, or
409 for the pagecache. The kernel must use tricks to access
410 this memory, making it slower to access than lowmem.
412 LowFree: Lowmem is memory which can be used for everything that
413 highmem can be used for, but it is also available for the
414 kernel's use for its own data structures. Among many
415 other things, it is where everything from the Slab is
416 allocated. Bad things happen when you're out of lowmem.
417 SwapTotal: total amount of swap space available
418 SwapFree: Memory which has been evicted from RAM, and is temporarily
420 Dirty: Memory which is waiting to get written back to the disk
421 Writeback: Memory which is actively being written back to the disk
422 Mapped: files which have been mmaped, such as libraries
423 Slab: in-kernel data structures cache
424 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
425 this is the total amount of memory currently available to
426 be allocated on the system. This limit is only adhered to
427 if strict overcommit accounting is enabled (mode 2 in
428 'vm.overcommit_memory').
429 The CommitLimit is calculated with the following formula:
430 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap
431 For example, on a system with 1G of physical RAM and 7G
432 of swap with a `vm.overcommit_ratio` of 30 it would
433 yield a CommitLimit of 7.3G.
434 For more details, see the memory overcommit documentation
435 in vm/overcommit-accounting.
436 Committed_AS: The amount of memory presently allocated on the system.
437 The committed memory is a sum of all of the memory which
438 has been allocated by processes, even if it has not been
439 "used" by them as of yet. A process which malloc()'s 1G
440 of memory, but only touches 300M of it will only show up
441 as using 300M of memory even if it has the address space
442 allocated for the entire 1G. This 1G is memory which has
443 been "committed" to by the VM and can be used at any time
444 by the allocating application. With strict overcommit
445 enabled on the system (mode 2 in 'vm.overcommit_memory'),
446 allocations which would exceed the CommitLimit (detailed
447 above) will not be permitted. This is useful if one needs
448 to guarantee that processes will not fail due to lack of
449 memory once that memory has been successfully allocated.
450 PageTables: amount of memory dedicated to the lowest level of page
452 VmallocTotal: total size of vmalloc memory area
453 VmallocUsed: amount of vmalloc area which is used
454 VmallocChunk: largest contigious block of vmalloc area which is free
457 1.3 IDE devices in /proc/ide
458 ----------------------------
460 The subdirectory /proc/ide contains information about all IDE devices of which
461 the kernel is aware. There is one subdirectory for each IDE controller, the
462 file drivers and a link for each IDE device, pointing to the device directory
463 in the controller specific subtree.
465 The file drivers contains general information about the drivers used for the
468 > cat /proc/ide/drivers
469 ide-cdrom version 4.53
470 ide-disk version 1.08
472 More detailed information can be found in the controller specific
473 subdirectories. These are named ide0, ide1 and so on. Each of these
474 directories contains the files shown in table 1-4.
477 Table 1-4: IDE controller info in /proc/ide/ide?
478 ..............................................................................
480 channel IDE channel (0 or 1)
481 config Configuration (only for PCI/IDE bridge)
483 model Type/Chipset of IDE controller
484 ..............................................................................
486 Each device connected to a controller has a separate subdirectory in the
487 controllers directory. The files listed in table 1-5 are contained in these
491 Table 1-5: IDE device information
492 ..............................................................................
495 capacity Capacity of the medium (in 512Byte blocks)
496 driver driver and version
497 geometry physical and logical geometry
498 identify device identify block
500 model device identifier
501 settings device setup
502 smart_thresholds IDE disk management thresholds
503 smart_values IDE disk management values
504 ..............................................................................
506 The most interesting file is settings. This file contains a nice overview of
507 the drive parameters:
509 # cat /proc/ide/ide0/hda/settings
510 name value min max mode
511 ---- ----- --- --- ----
512 bios_cyl 526 0 65535 rw
513 bios_head 255 0 255 rw
515 breada_readahead 4 0 127 rw
517 file_readahead 72 0 2097151 rw
519 keepsettings 0 0 1 rw
520 max_kb_per_request 122 1 127 rw
524 pio_mode write-only 0 255 w
530 1.4 Networking info in /proc/net
531 --------------------------------
533 The subdirectory /proc/net follows the usual pattern. Table 1-6 shows the
534 additional values you get for IP version 6 if you configure the kernel to
535 support this. Table 1-7 lists the files and their meaning.
538 Table 1-6: IPv6 info in /proc/net
539 ..............................................................................
541 udp6 UDP sockets (IPv6)
542 tcp6 TCP sockets (IPv6)
543 raw6 Raw device statistics (IPv6)
544 igmp6 IP multicast addresses, which this host joined (IPv6)
545 if_inet6 List of IPv6 interface addresses
546 ipv6_route Kernel routing table for IPv6
547 rt6_stats Global IPv6 routing tables statistics
548 sockstat6 Socket statistics (IPv6)
549 snmp6 Snmp data (IPv6)
550 ..............................................................................
553 Table 1-7: Network info in /proc/net
554 ..............................................................................
557 dev network devices with statistics
558 dev_mcast the Layer2 multicast groups a device is listening too
559 (interface index, label, number of references, number of bound
561 dev_stat network device status
562 ip_fwchains Firewall chain linkage
563 ip_fwnames Firewall chain names
564 ip_masq Directory containing the masquerading tables
565 ip_masquerade Major masquerading table
566 netstat Network statistics
567 raw raw device statistics
568 route Kernel routing table
569 rpc Directory containing rpc info
570 rt_cache Routing cache
572 sockstat Socket statistics
574 tr_rif Token ring RIF routing table
576 unix UNIX domain sockets
577 wireless Wireless interface data (Wavelan etc)
578 igmp IP multicast addresses, which this host joined
579 psched Global packet scheduler parameters.
580 netlink List of PF_NETLINK sockets
581 ip_mr_vifs List of multicast virtual interfaces
582 ip_mr_cache List of multicast routing cache
583 ..............................................................................
585 You can use this information to see which network devices are available in
586 your system and how much traffic was routed over those devices:
590 face |bytes packets errs drop fifo frame compressed multicast|[...
591 lo: 908188 5596 0 0 0 0 0 0 [...
592 ppp0:15475140 20721 410 0 0 410 0 0 [...
593 eth0: 614530 7085 0 0 0 0 0 1 [...
596 ...] bytes packets errs drop fifo colls carrier compressed
597 ...] 908188 5596 0 0 0 0 0 0
598 ...] 1375103 17405 0 0 0 0 0 0
599 ...] 1703981 5535 0 0 0 3 0 0
601 In addition, each Channel Bond interface has it's own directory. For
602 example, the bond0 device will have a directory called /proc/net/bond0/.
603 It will contain information that is specific to that bond, such as the
604 current slaves of the bond, the link status of the slaves, and how
605 many times the slaves link has failed.
610 If you have a SCSI host adapter in your system, you'll find a subdirectory
611 named after the driver for this adapter in /proc/scsi. You'll also see a list
612 of all recognized SCSI devices in /proc/scsi:
616 Host: scsi0 Channel: 00 Id: 00 Lun: 00
617 Vendor: IBM Model: DGHS09U Rev: 03E0
618 Type: Direct-Access ANSI SCSI revision: 03
619 Host: scsi0 Channel: 00 Id: 06 Lun: 00
620 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
621 Type: CD-ROM ANSI SCSI revision: 02
624 The directory named after the driver has one file for each adapter found in
625 the system. These files contain information about the controller, including
626 the used IRQ and the IO address range. The amount of information shown is
627 dependent on the adapter you use. The example shows the output for an Adaptec
628 AHA-2940 SCSI adapter:
630 > cat /proc/scsi/aic7xxx/0
632 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
634 TCQ Enabled By Default : Disabled
635 AIC7XXX_PROC_STATS : Disabled
636 AIC7XXX_RESET_DELAY : 5
637 Adapter Configuration:
638 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
639 Ultra Wide Controller
640 PCI MMAPed I/O Base: 0xeb001000
641 Adapter SEEPROM Config: SEEPROM found and used.
642 Adaptec SCSI BIOS: Enabled
644 SCBs: Active 0, Max Active 2,
645 Allocated 15, HW 16, Page 255
647 BIOS Control Word: 0x18b6
648 Adapter Control Word: 0x005b
649 Extended Translation: Enabled
650 Disconnect Enable Flags: 0xffff
651 Ultra Enable Flags: 0x0001
652 Tag Queue Enable Flags: 0x0000
653 Ordered Queue Tag Flags: 0x0000
654 Default Tag Queue Depth: 8
655 Tagged Queue By Device array for aic7xxx host instance 0:
656 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
657 Actual queue depth per device for aic7xxx host instance 0:
658 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
661 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
662 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
663 Total transfers 160151 (74577 reads and 85574 writes)
665 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
666 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
667 Total transfers 0 (0 reads and 0 writes)
670 1.6 Parallel port info in /proc/parport
671 ---------------------------------------
673 The directory /proc/parport contains information about the parallel ports of
674 your system. It has one subdirectory for each port, named after the port
677 These directories contain the four files shown in Table 1-8.
680 Table 1-8: Files in /proc/parport
681 ..............................................................................
683 autoprobe Any IEEE-1284 device ID information that has been acquired.
684 devices list of the device drivers using that port. A + will appear by the
685 name of the device currently using the port (it might not appear
687 hardware Parallel port's base address, IRQ line and DMA channel.
688 irq IRQ that parport is using for that port. This is in a separate
689 file to allow you to alter it by writing a new value in (IRQ
691 ..............................................................................
693 1.7 TTY info in /proc/tty
694 -------------------------
696 Information about the available and actually used tty's can be found in the
697 directory /proc/tty.You'll find entries for drivers and line disciplines in
698 this directory, as shown in Table 1-9.
701 Table 1-9: Files in /proc/tty
702 ..............................................................................
704 drivers list of drivers and their usage
705 ldiscs registered line disciplines
706 driver/serial usage statistic and status of single tty lines
707 ..............................................................................
709 To see which tty's are currently in use, you can simply look into the file
712 > cat /proc/tty/drivers
713 pty_slave /dev/pts 136 0-255 pty:slave
714 pty_master /dev/ptm 128 0-255 pty:master
715 pty_slave /dev/ttyp 3 0-255 pty:slave
716 pty_master /dev/pty 2 0-255 pty:master
717 serial /dev/cua 5 64-67 serial:callout
718 serial /dev/ttyS 4 64-67 serial
719 /dev/tty0 /dev/tty0 4 0 system:vtmaster
720 /dev/ptmx /dev/ptmx 5 2 system
721 /dev/console /dev/console 5 1 system:console
722 /dev/tty /dev/tty 5 0 system:/dev/tty
723 unknown /dev/tty 4 1-63 console
726 1.8 Miscellaneous kernel statistics in /proc/stat
727 -------------------------------------------------
729 Various pieces of information about kernel activity are available in the
730 /proc/stat file. All of the numbers reported in this file are aggregates
731 since the system first booted. For a quick look, simply cat the file:
734 cpu 2255 34 2290 22625563 6290 127 456
735 cpu0 1132 34 1441 11311718 3675 127 438
736 cpu1 1123 0 849 11313845 2614 0 18
737 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
744 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
745 lines. These numbers identify the amount of time the CPU has spent performing
746 different kinds of work. Time units are in USER_HZ (typically hundredths of a
747 second). The meanings of the columns are as follows, from left to right:
749 - user: normal processes executing in user mode
750 - nice: niced processes executing in user mode
751 - system: processes executing in kernel mode
752 - idle: twiddling thumbs
753 - iowait: waiting for I/O to complete
754 - irq: servicing interrupts
755 - softirq: servicing softirqs
757 The "intr" line gives counts of interrupts serviced since boot time, for each
758 of the possible system interrupts. The first column is the total of all
759 interrupts serviced; each subsequent column is the total for that particular
762 The "ctxt" line gives the total number of context switches across all CPUs.
764 The "btime" line gives the time at which the system booted, in seconds since
767 The "processes" line gives the number of processes and threads created, which
768 includes (but is not limited to) those created by calls to the fork() and
769 clone() system calls.
771 The "procs_running" line gives the number of processes currently running on
774 The "procs_blocked" line gives the number of processes currently blocked,
775 waiting for I/O to complete.
778 ------------------------------------------------------------------------------
780 ------------------------------------------------------------------------------
781 The /proc file system serves information about the running system. It not only
782 allows access to process data but also allows you to request the kernel status
783 by reading files in the hierarchy.
785 The directory structure of /proc reflects the types of information and makes
786 it easy, if not obvious, where to look for specific data.
787 ------------------------------------------------------------------------------
789 ------------------------------------------------------------------------------
790 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
791 ------------------------------------------------------------------------------
793 ------------------------------------------------------------------------------
795 ------------------------------------------------------------------------------
796 * Modifying kernel parameters by writing into files found in /proc/sys
797 * Exploring the files which modify certain parameters
798 * Review of the /proc/sys file tree
799 ------------------------------------------------------------------------------
802 A very interesting part of /proc is the directory /proc/sys. This is not only
803 a source of information, it also allows you to change parameters within the
804 kernel. Be very careful when attempting this. You can optimize your system,
805 but you can also cause it to crash. Never alter kernel parameters on a
806 production system. Set up a development machine and test to make sure that
807 everything works the way you want it to. You may have no alternative but to
808 reboot the machine once an error has been made.
810 To change a value, simply echo the new value into the file. An example is
811 given below in the section on the file system data. You need to be root to do
812 this. You can create your own boot script to perform this every time your
815 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
816 general things in the operation of the Linux kernel. Since some of the files
817 can inadvertently disrupt your system, it is advisable to read both
818 documentation and source before actually making adjustments. In any case, be
819 very careful when writing to any of these files. The entries in /proc may
820 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
821 review the kernel documentation in the directory /usr/src/linux/Documentation.
822 This chapter is heavily based on the documentation included in the pre 2.2
823 kernels, and became part of it in version 2.2.1 of the Linux kernel.
825 2.1 /proc/sys/fs - File system data
826 -----------------------------------
828 This subdirectory contains specific file system, file handle, inode, dentry
829 and quota information.
831 Currently, these files are in /proc/sys/fs:
836 Status of the directory cache. Since directory entries are dynamically
837 allocated and deallocated, this file indicates the current status. It holds
838 six values, in which the last two are not used and are always zero. The others
839 are listed in table 2-1.
842 Table 2-1: Status files of the directory cache
843 ..............................................................................
845 nr_dentry Almost always zero
846 nr_unused Number of unused cache entries
848 in seconds after the entry may be reclaimed, when memory is short
849 want_pages internally
850 ..............................................................................
852 dquot-nr and dquot-max
853 ----------------------
855 The file dquot-max shows the maximum number of cached disk quota entries.
857 The file dquot-nr shows the number of allocated disk quota entries and the
858 number of free disk quota entries.
860 If the number of available cached disk quotas is very low and you have a large
861 number of simultaneous system users, you might want to raise the limit.
866 The kernel allocates file handles dynamically, but doesn't free them again at
869 The value in file-max denotes the maximum number of file handles that the
870 Linux kernel will allocate. When you get a lot of error messages about running
871 out of file handles, you might want to raise this limit. The default value is
872 10% of RAM in kilobytes. To change it, just write the new number into the
875 # cat /proc/sys/fs/file-max
877 # echo 8192 > /proc/sys/fs/file-max
878 # cat /proc/sys/fs/file-max
882 This method of revision is useful for all customizable parameters of the
883 kernel - simply echo the new value to the corresponding file.
885 Historically, the three values in file-nr denoted the number of allocated file
886 handles, the number of allocated but unused file handles, and the maximum
887 number of file handles. Linux 2.6 always reports 0 as the number of free file
888 handles -- this is not an error, it just means that the number of allocated
889 file handles exactly matches the number of used file handles.
891 Attempts to allocate more file descriptors than file-max are reported with
892 printk, look for "VFS: file-max limit <number> reached".
894 inode-state and inode-nr
895 ------------------------
897 The file inode-nr contains the first two items from inode-state, so we'll skip
900 inode-state contains two actual numbers and five dummy values. The numbers
901 are nr_inodes and nr_free_inodes (in order of appearance).
906 Denotes the number of inodes the system has allocated. This number will
907 grow and shrink dynamically.
912 Represents the number of free inodes. Ie. The number of inuse inodes is
913 (nr_inodes - nr_free_inodes).
915 aio-nr and aio-max-nr
916 ---------------------
918 aio-nr is the running total of the number of events specified on the
919 io_setup system call for all currently active aio contexts. If aio-nr
920 reaches aio-max-nr then io_setup will fail with EAGAIN. Note that
921 raising aio-max-nr does not result in the pre-allocation or re-sizing
922 of any kernel data structures.
924 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
925 -----------------------------------------------------------
927 Besides these files, there is the subdirectory /proc/sys/fs/binfmt_misc. This
928 handles the kernel support for miscellaneous binary formats.
930 Binfmt_misc provides the ability to register additional binary formats to the
931 Kernel without compiling an additional module/kernel. Therefore, binfmt_misc
932 needs to know magic numbers at the beginning or the filename extension of the
935 It works by maintaining a linked list of structs that contain a description of
936 a binary format, including a magic with size (or the filename extension),
937 offset and mask, and the interpreter name. On request it invokes the given
938 interpreter with the original program as argument, as binfmt_java and
939 binfmt_em86 and binfmt_mz do. Since binfmt_misc does not define any default
940 binary-formats, you have to register an additional binary-format.
942 There are two general files in binfmt_misc and one file per registered format.
943 The two general files are register and status.
945 Registering a new binary format
946 -------------------------------
948 To register a new binary format you have to issue the command
950 echo :name:type:offset:magic:mask:interpreter: > /proc/sys/fs/binfmt_misc/register
954 with appropriate name (the name for the /proc-dir entry), offset (defaults to
955 0, if omitted), magic, mask (which can be omitted, defaults to all 0xff) and
956 last but not least, the interpreter that is to be invoked (for example and
957 testing /bin/echo). Type can be M for usual magic matching or E for filename
958 extension matching (give extension in place of magic).
960 Check or reset the status of the binary format handler
961 ------------------------------------------------------
963 If you do a cat on the file /proc/sys/fs/binfmt_misc/status, you will get the
964 current status (enabled/disabled) of binfmt_misc. Change the status by echoing
965 0 (disables) or 1 (enables) or -1 (caution: this clears all previously
966 registered binary formats) to status. For example echo 0 > status to disable
967 binfmt_misc (temporarily).
969 Status of a single handler
970 --------------------------
972 Each registered handler has an entry in /proc/sys/fs/binfmt_misc. These files
973 perform the same function as status, but their scope is limited to the actual
974 binary format. By cating this file, you also receive all related information
975 about the interpreter/magic of the binfmt.
977 Example usage of binfmt_misc (emulate binfmt_java)
978 --------------------------------------------------
980 cd /proc/sys/fs/binfmt_misc
981 echo ':Java:M::\xca\xfe\xba\xbe::/usr/local/java/bin/javawrapper:' > register
982 echo ':HTML:E::html::/usr/local/java/bin/appletviewer:' > register
983 echo ':Applet:M::<!--applet::/usr/local/java/bin/appletviewer:' > register
984 echo ':DEXE:M::\x0eDEX::/usr/bin/dosexec:' > register
987 These four lines add support for Java executables and Java applets (like
988 binfmt_java, additionally recognizing the .html extension with no need to put
989 <!--applet> to every applet file). You have to install the JDK and the
990 shell-script /usr/local/java/bin/javawrapper too. It works around the
991 brokenness of the Java filename handling. To add a Java binary, just create a
992 link to the class-file somewhere in the path.
994 2.3 /proc/sys/kernel - general kernel parameters
995 ------------------------------------------------
997 This directory reflects general kernel behaviors. As I've said before, the
998 contents depend on your configuration. Here you'll find the most important
999 files, along with descriptions of what they mean and how to use them.
1004 The file contains three values; highwater, lowwater, and frequency.
1006 It exists only when BSD-style process accounting is enabled. These values
1007 control its behavior. If the free space on the file system where the log lives
1008 goes below lowwater percentage, accounting suspends. If it goes above
1009 highwater percentage, accounting resumes. Frequency determines how often you
1010 check the amount of free space (value is in seconds). Default settings are: 4,
1011 2, and 30. That is, suspend accounting if there is less than 2 percent free;
1012 resume it if we have a value of 3 or more percent; consider information about
1013 the amount of free space valid for 30 seconds
1018 When the value in this file is 0, ctrl-alt-del is trapped and sent to the init
1019 program to handle a graceful restart. However, when the value is greater that
1020 zero, Linux's reaction to this key combination will be an immediate reboot,
1021 without syncing its dirty buffers.
1024 When a program (like dosemu) has the keyboard in raw mode, the
1025 ctrl-alt-del is intercepted by the program before it ever reaches the
1026 kernel tty layer, and it is up to the program to decide what to do with
1029 domainname and hostname
1030 -----------------------
1032 These files can be controlled to set the NIS domainname and hostname of your
1033 box. For the classic darkstar.frop.org a simple:
1035 # echo "darkstar" > /proc/sys/kernel/hostname
1036 # echo "frop.org" > /proc/sys/kernel/domainname
1039 would suffice to set your hostname and NIS domainname.
1041 osrelease, ostype and version
1042 -----------------------------
1044 The names make it pretty obvious what these fields contain:
1046 > cat /proc/sys/kernel/osrelease
1049 > cat /proc/sys/kernel/ostype
1052 > cat /proc/sys/kernel/version
1053 #4 Fri Oct 1 12:41:14 PDT 1999
1056 The files osrelease and ostype should be clear enough. Version needs a little
1057 more clarification. The #4 means that this is the 4th kernel built from this
1058 source base and the date after it indicates the time the kernel was built. The
1059 only way to tune these values is to rebuild the kernel.
1064 The value in this file represents the number of seconds the kernel waits
1065 before rebooting on a panic. When you use the software watchdog, the
1066 recommended setting is 60. If set to 0, the auto reboot after a kernel panic
1067 is disabled, which is the default setting.
1072 The four values in printk denote
1074 * default_message_loglevel,
1075 * minimum_console_loglevel and
1076 * default_console_loglevel
1079 These values influence printk() behavior when printing or logging error
1080 messages, which come from inside the kernel. See syslog(2) for more
1081 information on the different log levels.
1086 Messages with a higher priority than this will be printed to the console.
1088 default_message_level
1089 ---------------------
1091 Messages without an explicit priority will be printed with this priority.
1093 minimum_console_loglevel
1094 ------------------------
1096 Minimum (highest) value to which the console_loglevel can be set.
1098 default_console_loglevel
1099 ------------------------
1101 Default value for console_loglevel.
1106 This file shows the size of the generic SCSI (sg) buffer. At this point, you
1107 can't tune it yet, but you can change it at compile time by editing
1108 include/scsi/sg.h and changing the value of SG_BIG_BUFF.
1110 If you use a scanner with SANE (Scanner Access Now Easy) you might want to set
1111 this to a higher value. Refer to the SANE documentation on this issue.
1116 The location where the modprobe binary is located. The kernel uses this
1117 program to load modules on demand.
1122 The value in this file affects behavior of handling NMI. When the value is
1123 non-zero, unknown NMI is trapped and then panic occurs. At that time, kernel
1124 debugging information is displayed on console.
1126 NMI switch that most IA32 servers have fires unknown NMI up, for example.
1127 If a system hangs up, try pressing the NMI switch.
1132 Enables/Disables the NMI watchdog on x86 systems. When the value is non-zero
1133 the NMI watchdog is enabled and will continuously test all online cpus to
1134 determine whether or not they are still functioning properly.
1136 Because the NMI watchdog shares registers with oprofile, by disabling the NMI
1137 watchdog, oprofile may have more registers to utilize.
1140 2.4 /proc/sys/vm - The virtual memory subsystem
1141 -----------------------------------------------
1143 The files in this directory can be used to tune the operation of the virtual
1144 memory (VM) subsystem of the Linux kernel.
1149 Controls the tendency of the kernel to reclaim the memory which is used for
1150 caching of directory and inode objects.
1152 At the default value of vfs_cache_pressure=100 the kernel will attempt to
1153 reclaim dentries and inodes at a "fair" rate with respect to pagecache and
1154 swapcache reclaim. Decreasing vfs_cache_pressure causes the kernel to prefer
1155 to retain dentry and inode caches. Increasing vfs_cache_pressure beyond 100
1156 causes the kernel to prefer to reclaim dentries and inodes.
1158 dirty_background_ratio
1159 ----------------------
1161 Contains, as a percentage of total system memory, the number of pages at which
1162 the pdflush background writeback daemon will start writing out dirty data.
1167 Contains, as a percentage of total system memory, the number of pages at which
1168 a process which is generating disk writes will itself start writing out dirty
1171 dirty_writeback_centisecs
1172 -------------------------
1174 The pdflush writeback daemons will periodically wake up and write `old' data
1175 out to disk. This tunable expresses the interval between those wakeups, in
1176 100'ths of a second.
1178 Setting this to zero disables periodic writeback altogether.
1180 dirty_expire_centisecs
1181 ----------------------
1183 This tunable is used to define when dirty data is old enough to be eligible
1184 for writeout by the pdflush daemons. It is expressed in 100'ths of a second.
1185 Data which has been dirty in-memory for longer than this interval will be
1186 written out next time a pdflush daemon wakes up.
1191 If non-zero, this sysctl disables the new 32-bit mmap mmap layout - the kernel
1192 will use the legacy (2.4) layout for all processes.
1194 lower_zone_protection
1195 ---------------------
1197 For some specialised workloads on highmem machines it is dangerous for
1198 the kernel to allow process memory to be allocated from the "lowmem"
1199 zone. This is because that memory could then be pinned via the mlock()
1200 system call, or by unavailability of swapspace.
1202 And on large highmem machines this lack of reclaimable lowmem memory
1205 So the Linux page allocator has a mechanism which prevents allocations
1206 which _could_ use highmem from using too much lowmem. This means that
1207 a certain amount of lowmem is defended from the possibility of being
1208 captured into pinned user memory.
1210 (The same argument applies to the old 16 megabyte ISA DMA region. This
1211 mechanism will also defend that region from allocations which could use
1214 The `lower_zone_protection' tunable determines how aggressive the kernel is
1215 in defending these lower zones. The default value is zero - no
1218 If you have a machine which uses highmem or ISA DMA and your
1219 applications are using mlock(), or if you are running with no swap then
1220 you probably should increase the lower_zone_protection setting.
1222 The units of this tunable are fairly vague. It is approximately equal
1223 to "megabytes". So setting lower_zone_protection=100 will protect around 100
1224 megabytes of the lowmem zone from user allocations. It will also make
1225 those 100 megabytes unavaliable for use by applications and by
1226 pagecache, so there is a cost.
1228 The effects of this tunable may be observed by monitoring
1229 /proc/meminfo:LowFree. Write a single huge file and observe the point
1230 at which LowFree ceases to fall.
1232 A reasonable value for lower_zone_protection is 100.
1237 page-cluster controls the number of pages which are written to swap in
1238 a single attempt. The swap I/O size.
1240 It is a logarithmic value - setting it to zero means "1 page", setting
1241 it to 1 means "2 pages", setting it to 2 means "4 pages", etc.
1243 The default value is three (eight pages at a time). There may be some
1244 small benefits in tuning this to a different value if your workload is
1250 Controls overcommit of system memory, possibly allowing processes
1251 to allocate (but not use) more memory than is actually available.
1254 0 - Heuristic overcommit handling. Obvious overcommits of
1255 address space are refused. Used for a typical system. It
1256 ensures a seriously wild allocation fails while allowing
1257 overcommit to reduce swap usage. root is allowed to
1258 allocate slightly more memory in this mode. This is the
1261 1 - Always overcommit. Appropriate for some scientific
1264 2 - Don't overcommit. The total address space commit
1265 for the system is not permitted to exceed swap plus a
1266 configurable percentage (default is 50) of physical RAM.
1267 Depending on the percentage you use, in most situations
1268 this means a process will not be killed while attempting
1269 to use already-allocated memory but will receive errors
1270 on memory allocation as appropriate.
1275 Percentage of physical memory size to include in overcommit calculations
1278 Memory allocation limit = swapspace + physmem * (overcommit_ratio / 100)
1280 swapspace = total size of all swap areas
1281 physmem = size of physical memory in system
1283 nr_hugepages and hugetlb_shm_group
1284 ----------------------------------
1286 nr_hugepages configures number of hugetlb page reserved for the system.
1288 hugetlb_shm_group contains group id that is allowed to create SysV shared
1289 memory segment using hugetlb page.
1294 laptop_mode is a knob that controls "laptop mode". All the things that are
1295 controlled by this knob are discussed in Documentation/laptop-mode.txt.
1300 block_dump enables block I/O debugging when set to a nonzero value. More
1301 information on block I/O debugging is in Documentation/laptop-mode.txt.
1306 This file contains valid hold time of swap out protection token. The Linux
1307 VM has token based thrashing control mechanism and uses the token to prevent
1308 unnecessary page faults in thrashing situation. The unit of the value is
1309 second. The value would be useful to tune thrashing behavior.
1314 Writing to this will cause the kernel to drop clean caches, dentries and
1315 inodes from memory, causing that memory to become free.
1318 echo 1 > /proc/sys/vm/drop_caches
1319 To free dentries and inodes:
1320 echo 2 > /proc/sys/vm/drop_caches
1321 To free pagecache, dentries and inodes:
1322 echo 3 > /proc/sys/vm/drop_caches
1324 As this is a non-destructive operation and dirty objects are not freeable, the
1325 user should run `sync' first.
1328 2.5 /proc/sys/dev - Device specific parameters
1329 ----------------------------------------------
1331 Currently there is only support for CDROM drives, and for those, there is only
1332 one read-only file containing information about the CD-ROM drives attached to
1335 >cat /proc/sys/dev/cdrom/info
1336 CD-ROM information, Id: cdrom.c 2.55 1999/04/25
1340 drive # of slots: 1 0
1344 Can change speed: 1 1
1345 Can select disk: 0 1
1346 Can read multisession: 1 1
1348 Reports media changed: 1 1
1352 You see two drives, sr0 and hdb, along with a list of their features.
1354 2.6 /proc/sys/sunrpc - Remote procedure calls
1355 ---------------------------------------------
1357 This directory contains four files, which enable or disable debugging for the
1358 RPC functions NFS, NFS-daemon, RPC and NLM. The default values are 0. They can
1359 be set to one to turn debugging on. (The default value is 0 for each)
1361 2.7 /proc/sys/net - Networking stuff
1362 ------------------------------------
1364 The interface to the networking parts of the kernel is located in
1365 /proc/sys/net. Table 2-3 shows all possible subdirectories. You may see only
1366 some of them, depending on your kernel's configuration.
1369 Table 2-3: Subdirectories in /proc/sys/net
1370 ..............................................................................
1371 Directory Content Directory Content
1372 core General parameter appletalk Appletalk protocol
1373 unix Unix domain sockets netrom NET/ROM
1374 802 E802 protocol ax25 AX25
1375 ethernet Ethernet protocol rose X.25 PLP layer
1376 ipv4 IP version 4 x25 X.25 protocol
1377 ipx IPX token-ring IBM token ring
1378 bridge Bridging decnet DEC net
1380 ..............................................................................
1382 We will concentrate on IP networking here. Since AX15, X.25, and DEC Net are
1383 only minor players in the Linux world, we'll skip them in this chapter. You'll
1384 find some short info on Appletalk and IPX further on in this chapter. Review
1385 the online documentation and the kernel source to get a detailed view of the
1386 parameters for those protocols. In this section we'll discuss the
1387 subdirectories printed in bold letters in the table above. As default values
1388 are suitable for most needs, there is no need to change these values.
1390 /proc/sys/net/core - Network core options
1391 -----------------------------------------
1396 The default setting of the socket receive buffer in bytes.
1401 The maximum receive socket buffer size in bytes.
1406 The default setting (in bytes) of the socket send buffer.
1411 The maximum send socket buffer size in bytes.
1413 message_burst and message_cost
1414 ------------------------------
1416 These parameters are used to limit the warning messages written to the kernel
1417 log from the networking code. They enforce a rate limit to make a
1418 denial-of-service attack impossible. A higher message_cost factor, results in
1419 fewer messages that will be written. Message_burst controls when messages will
1420 be dropped. The default settings limit warning messages to one every five
1426 Maximum number of packets, queued on the INPUT side, when the interface
1427 receives packets faster than kernel can process them.
1432 Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
1433 of struct cmsghdr structures with appended data.
1435 /proc/sys/net/unix - Parameters for Unix domain sockets
1436 -------------------------------------------------------
1438 There are only two files in this subdirectory. They control the delays for
1439 deleting and destroying socket descriptors.
1441 2.8 /proc/sys/net/ipv4 - IPV4 settings
1442 --------------------------------------
1444 IP version 4 is still the most used protocol in Unix networking. It will be
1445 replaced by IP version 6 in the next couple of years, but for the moment it's
1446 the de facto standard for the internet and is used in most networking
1447 environments around the world. Because of the importance of this protocol,
1448 we'll have a deeper look into the subtree controlling the behavior of the IPv4
1449 subsystem of the Linux kernel.
1451 Let's start with the entries in /proc/sys/net/ipv4.
1456 icmp_echo_ignore_all and icmp_echo_ignore_broadcasts
1457 ----------------------------------------------------
1459 Turn on (1) or off (0), if the kernel should ignore all ICMP ECHO requests, or
1460 just those to broadcast and multicast addresses.
1462 Please note that if you accept ICMP echo requests with a broadcast/multi\-cast
1463 destination address your network may be used as an exploder for denial of
1464 service packet flooding attacks to other hosts.
1466 icmp_destunreach_rate, icmp_echoreply_rate, icmp_paramprob_rate and icmp_timeexeed_rate
1467 ---------------------------------------------------------------------------------------
1469 Sets limits for sending ICMP packets to specific targets. A value of zero
1470 disables all limiting. Any positive value sets the maximum package rate in
1471 hundredth of a second (on Intel systems).
1479 This file contains the number one if the host received its IP configuration by
1480 RARP, BOOTP, DHCP or a similar mechanism. Otherwise it is zero.
1485 TTL (Time To Live) for IPv4 interfaces. This is simply the maximum number of
1486 hops a packet may travel.
1491 Enable dynamic socket address rewriting on interface address change. This is
1492 useful for dialup interface with changing IP addresses.
1497 Enable or disable forwarding of IP packages between interfaces. Changing this
1498 value resets all other parameters to their default values. They differ if the
1499 kernel is configured as host or router.
1504 Range of ports used by TCP and UDP to choose the local port. Contains two
1505 numbers, the first number is the lowest port, the second number the highest
1506 local port. Default is 1024-4999. Should be changed to 32768-61000 for
1512 Global switch to turn path MTU discovery off. It can also be set on a per
1513 socket basis by the applications or on a per route basis.
1518 Enable/disable debugging of IP masquerading.
1520 IP fragmentation settings
1521 -------------------------
1523 ipfrag_high_trash and ipfrag_low_trash
1524 --------------------------------------
1526 Maximum memory used to reassemble IP fragments. When ipfrag_high_thresh bytes
1527 of memory is allocated for this purpose, the fragment handler will toss
1528 packets until ipfrag_low_thresh is reached.
1533 Time in seconds to keep an IP fragment in memory.
1541 This file controls the use of the ECN bit in the IPv4 headers, this is a new
1542 feature about Explicit Congestion Notification, but some routers and firewalls
1543 block trafic that has this bit set, so it could be necessary to echo 0 to
1544 /proc/sys/net/ipv4/tcp_ecn, if you want to talk to this sites. For more info
1545 you could read RFC2481.
1547 tcp_retrans_collapse
1548 --------------------
1550 Bug-to-bug compatibility with some broken printers. On retransmit, try to send
1551 larger packets to work around bugs in certain TCP stacks. Can be turned off by
1554 tcp_keepalive_probes
1555 --------------------
1557 Number of keep alive probes TCP sends out, until it decides that the
1558 connection is broken.
1563 How often TCP sends out keep alive messages, when keep alive is enabled. The
1569 Number of times initial SYNs for a TCP connection attempt will be
1570 retransmitted. Should not be higher than 255. This is only the timeout for
1571 outgoing connections, for incoming connections the number of retransmits is
1572 defined by tcp_retries1.
1577 Enable select acknowledgments after RFC2018.
1582 Enable timestamps as defined in RFC1323.
1587 Enable the strict RFC793 interpretation of the TCP urgent pointer field. The
1588 default is to use the BSD compatible interpretation of the urgent pointer
1589 pointing to the first byte after the urgent data. The RFC793 interpretation is
1590 to have it point to the last byte of urgent data. Enabling this option may
1591 lead to interoperability problems. Disabled by default.
1596 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES. Send out
1597 syncookies when the syn backlog queue of a socket overflows. This is to ward
1598 off the common 'syn flood attack'. Disabled by default.
1600 Note that the concept of a socket backlog is abandoned. This means the peer
1601 may not receive reliable error messages from an over loaded server with
1607 Enable window scaling as defined in RFC1323.
1612 The length of time in seconds it takes to receive a final FIN before the
1613 socket is always closed. This is strictly a violation of the TCP
1614 specification, but required to prevent denial-of-service attacks.
1619 Indicates how many keep alive probes are sent per slow timer run. Should not
1620 be set too high to prevent bursts.
1625 Length of the per socket backlog queue. Since Linux 2.2 the backlog specified
1626 in listen(2) only specifies the length of the backlog queue of already
1627 established sockets. When more connection requests arrive Linux starts to drop
1628 packets. When syncookies are enabled the packets are still answered and the
1629 maximum queue is effectively ignored.
1634 Defines how often an answer to a TCP connection request is retransmitted
1640 Defines how often a TCP packet is retransmitted before giving up.
1642 Interface specific settings
1643 ---------------------------
1645 In the directory /proc/sys/net/ipv4/conf you'll find one subdirectory for each
1646 interface the system knows about and one directory calls all. Changes in the
1647 all subdirectory affect all interfaces, whereas changes in the other
1648 subdirectories affect only one interface. All directories have the same
1654 This switch decides if the kernel accepts ICMP redirect messages or not. The
1655 default is 'yes' if the kernel is configured for a regular host and 'no' for a
1656 router configuration.
1661 Should source routed packages be accepted or declined. The default is
1662 dependent on the kernel configuration. It's 'yes' for routers and 'no' for
1668 Accept packets with source address 0.b.c.d with destinations not to this host
1669 as local ones. It is supposed that a BOOTP relay daemon will catch and forward
1672 The default is 0, since this feature is not implemented yet (kernel version
1678 Enable or disable IP forwarding on this interface.
1683 Log packets with source addresses with no known route to kernel log.
1688 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE and a
1689 multicast routing daemon is required.
1694 Does (1) or does not (0) perform proxy ARP.
1699 Integer value determines if a source validation should be made. 1 means yes, 0
1700 means no. Disabled by default, but local/broadcast address spoofing is always
1703 If you set this to 1 on a router that is the only connection for a network to
1704 the net, it will prevent spoofing attacks against your internal networks
1705 (external addresses can still be spoofed), without the need for additional
1711 Accept ICMP redirect messages only for gateways, listed in default gateway
1712 list. Enabled by default.
1717 If it is not set the kernel does not assume that different subnets on this
1718 device can communicate directly. Default setting is 'yes'.
1723 Determines whether to send ICMP redirects to other hosts.
1728 The directory /proc/sys/net/ipv4/route contains several file to control
1731 error_burst and error_cost
1732 --------------------------
1734 These parameters are used to limit how many ICMP destination unreachable to
1735 send from the host in question. ICMP destination unreachable messages are
1736 sent when we cannot reach the next hop while trying to transmit a packet.
1737 It will also print some error messages to kernel logs if someone is ignoring
1738 our ICMP redirects. The higher the error_cost factor is, the fewer
1739 destination unreachable and error messages will be let through. Error_burst
1740 controls when destination unreachable messages and error messages will be
1741 dropped. The default settings limit warning messages to five every second.
1746 Writing to this file results in a flush of the routing cache.
1748 gc_elasticity, gc_interval, gc_min_interval_ms, gc_timeout, gc_thresh
1749 ---------------------------------------------------------------------
1751 Values to control the frequency and behavior of the garbage collection
1752 algorithm for the routing cache. gc_min_interval is deprecated and replaced
1753 by gc_min_interval_ms.
1759 Maximum size of the routing cache. Old entries will be purged once the cache
1760 reached has this size.
1762 max_delay, min_delay
1763 --------------------
1765 Delays for flushing the routing cache.
1767 redirect_load, redirect_number
1768 ------------------------------
1770 Factors which determine if more ICPM redirects should be sent to a specific
1771 host. No redirects will be sent once the load limit or the maximum number of
1772 redirects has been reached.
1777 Timeout for redirects. After this period redirects will be sent again, even if
1778 this has been stopped, because the load or number limit has been reached.
1780 Network Neighbor handling
1781 -------------------------
1783 Settings about how to handle connections with direct neighbors (nodes attached
1784 to the same link) can be found in the directory /proc/sys/net/ipv4/neigh.
1786 As we saw it in the conf directory, there is a default subdirectory which
1787 holds the default values, and one directory for each interface. The contents
1788 of the directories are identical, with the single exception that the default
1789 settings contain additional options to set garbage collection parameters.
1791 In the interface directories you'll find the following entries:
1793 base_reachable_time, base_reachable_time_ms
1794 -------------------------------------------
1796 A base value used for computing the random reachable time value as specified
1799 Expression of base_reachable_time, which is deprecated, is in seconds.
1800 Expression of base_reachable_time_ms is in milliseconds.
1802 retrans_time, retrans_time_ms
1803 -----------------------------
1805 The time between retransmitted Neighbor Solicitation messages.
1806 Used for address resolution and to determine if a neighbor is
1809 Expression of retrans_time, which is deprecated, is in 1/100 seconds (for
1810 IPv4) or in jiffies (for IPv6).
1811 Expression of retrans_time_ms is in milliseconds.
1816 Maximum queue length for a pending arp request - the number of packets which
1817 are accepted from other layers while the ARP address is still resolved.
1822 Maximum for random delay of answers to neighbor solicitation messages in
1823 jiffies (1/100 sec). Not yet implemented (Linux does not have anycast support
1829 Maximum number of retries for unicast solicitation.
1834 Maximum number of retries for multicast solicitation.
1836 delay_first_probe_time
1837 ----------------------
1839 Delay for the first time probe if the neighbor is reachable. (see
1845 An ARP/neighbor entry is only replaced with a new one if the old is at least
1846 locktime old. This prevents ARP cache thrashing.
1851 Maximum time (real time is random [0..proxytime]) before answering to an ARP
1852 request for which we have an proxy ARP entry. In some cases, this is used to
1853 prevent network flooding.
1858 Maximum queue length of the delayed proxy arp timer. (see proxy_delay).
1863 Determines the number of requests to send to the user level ARP daemon. Use 0
1869 Determines how often to check for stale ARP entries. After an ARP entry is
1870 stale it will be resolved again (which is useful when an IP address migrates
1871 to another machine). When ucast_solicit is greater than 0 it first tries to
1872 send an ARP packet directly to the known host When that fails and
1873 mcast_solicit is greater than 0, an ARP request is broadcasted.
1878 The /proc/sys/net/appletalk directory holds the Appletalk configuration data
1879 when Appletalk is loaded. The configurable parameters are:
1884 The amount of time we keep an ARP entry before expiring it. Used to age out
1890 The amount of time we will spend trying to resolve an Appletalk address.
1892 aarp-retransmit-limit
1893 ---------------------
1895 The number of times we will retransmit a query before giving up.
1900 Controls the rate at which expires are checked.
1902 The directory /proc/net/appletalk holds the list of active Appletalk sockets
1905 The fields indicate the DDP type, the local address (in network:node format)
1906 the remote address, the size of the transmit pending queue, the size of the
1907 received queue (bytes waiting for applications to read) the state and the uid
1910 /proc/net/atalk_iface lists all the interfaces configured for appletalk.It
1911 shows the name of the interface, its Appletalk address, the network range on
1912 that address (or network number for phase 1 networks), and the status of the
1915 /proc/net/atalk_route lists each known network route. It lists the target
1916 (network) that the route leads to, the router (may be directly connected), the
1917 route flags, and the device the route is using.
1922 The IPX protocol has no tunable values in proc/sys/net.
1924 The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
1925 socket giving the local and remote addresses in Novell format (that is
1926 network:node:port). In accordance with the strange Novell tradition,
1927 everything but the port is in hex. Not_Connected is displayed for sockets that
1928 are not tied to a specific remote address. The Tx and Rx queue sizes indicate
1929 the number of bytes pending for transmission and reception. The state
1930 indicates the state the socket is in and the uid is the owning uid of the
1933 The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
1934 it gives the network number, the node number, and indicates if the network is
1935 the primary network. It also indicates which device it is bound to (or
1936 Internal for internal networks) and the Frame Type if appropriate. Linux
1937 supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
1940 The /proc/net/ipx_route table holds a list of IPX routes. For each route it
1941 gives the destination network, the router node (or Directly) and the network
1942 address of the router (or Connected) for internal networks.
1944 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
1945 ----------------------------------------------------------
1947 The "mqueue" filesystem provides the necessary kernel features to enable the
1948 creation of a user space library that implements the POSIX message queues
1949 API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System
1950 Interfaces specification.)
1952 The "mqueue" filesystem contains values for determining/setting the amount of
1953 resources used by the file system.
1955 /proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the
1956 maximum number of message queues allowed on the system.
1958 /proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the
1959 maximum number of messages in a queue value. In fact it is the limiting value
1960 for another (user) limit which is set in mq_open invocation. This attribute of
1961 a queue must be less or equal then msg_max.
1963 /proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the
1964 maximum message size value (it is every message queue's attribute set during
1967 2.12 /proc/<pid>/oom_adj - Adjust the oom-killer score
1968 ------------------------------------------------------
1970 This file can be used to adjust the score used to select which processes
1971 should be killed in an out-of-memory situation. Giving it a high score will
1972 increase the likelihood of this process being killed by the oom-killer. Valid
1973 values are in the range -16 to +15, plus the special value -17, which disables
1974 oom-killing altogether for this process.
1976 2.13 /proc/<pid>/oom_score - Display current oom-killer score
1977 -------------------------------------------------------------
1979 ------------------------------------------------------------------------------
1980 This file can be used to check the current score used by the oom-killer is for
1981 any given <pid>. Use it together with /proc/<pid>/oom_adj to tune which
1982 process should be killed in an out-of-memory situation.
1984 ------------------------------------------------------------------------------
1986 ------------------------------------------------------------------------------
1987 Certain aspects of kernel behavior can be modified at runtime, without the
1988 need to recompile the kernel, or even to reboot the system. The files in the
1989 /proc/sys tree can not only be read, but also modified. You can use the echo
1990 command to write value into these files, thereby changing the default settings
1992 ------------------------------------------------------------------------------
git.samba.org / sfrench / cifs-2.6.git / blob