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
43 ------------------------------------------------------------------------------
45 ------------------------------------------------------------------------------
47 0.1 Introduction/Credits
48 ------------------------
50 This documentation is part of a soon (or so we hope) to be released book on
51 the SuSE Linux distribution. As there is no complete documentation for the
52 /proc file system and we've used many freely available sources to write these
53 chapters, it seems only fair to give the work back to the Linux community.
54 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
55 afraid it's still far from complete, but we hope it will be useful. As far as
56 we know, it is the first 'all-in-one' document about the /proc file system. It
57 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
58 SPARC, AXP, etc., features, you probably won't find what you are looking for.
59 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
60 additions and patches are welcome and will be added to this document if you
63 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
64 other people for help compiling this documentation. We'd also like to extend a
65 special thank you to Andi Kleen for documentation, which we relied on heavily
66 to create this document, as well as the additional information he provided.
67 Thanks to everybody else who contributed source or docs to the Linux kernel
68 and helped create a great piece of software... :)
70 If you have any comments, corrections or additions, please don't hesitate to
71 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
74 The latest version of this document is available online at
75 http://skaro.nightcrawler.com/~bb/Docs/Proc as HTML version.
77 If the above direction does not works for you, ypu could try the kernel
78 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
79 comandante@zaralinux.com.
84 We don't guarantee the correctness of this document, and if you come to us
85 complaining about how you screwed up your system because of incorrect
86 documentation, we won't feel responsible...
88 ------------------------------------------------------------------------------
89 CHAPTER 1: COLLECTING SYSTEM INFORMATION
90 ------------------------------------------------------------------------------
92 ------------------------------------------------------------------------------
94 ------------------------------------------------------------------------------
95 * Investigating the properties of the pseudo file system /proc and its
96 ability to provide information on the running Linux system
97 * Examining /proc's structure
98 * Uncovering various information about the kernel and the processes running
100 ------------------------------------------------------------------------------
103 The proc file system acts as an interface to internal data structures in the
104 kernel. It can be used to obtain information about the system and to change
105 certain kernel parameters at runtime (sysctl).
107 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
108 show you how you can use /proc/sys to change settings.
110 1.1 Process-Specific Subdirectories
111 -----------------------------------
113 The directory /proc contains (among other things) one subdirectory for each
114 process running on the system, which is named after the process ID (PID).
116 The link self points to the process reading the file system. Each process
117 subdirectory has the entries listed in Table 1-1.
120 Table 1-1: Process specific entries in /proc
121 ..............................................................................
123 cmdline Command line arguments
124 cpu Current and last cpu in which it was executed (2.4)(smp)
125 cwd Link to the current working directory
126 environ Values of environment variables
127 exe Link to the executable of this process
128 fd Directory, which contains all file descriptors
129 maps Memory maps to executables and library files (2.4)
130 mem Memory held by this process
131 root Link to the root directory of this process
133 statm Process memory status information
134 status Process status in human readable form
135 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan
136 smaps Extension based on maps, presenting the rss size for each mapped file
137 ..............................................................................
139 For example, to get the status information of a process, all you have to do is
140 read the file /proc/PID/status:
142 >cat /proc/self/status
158 SigPnd: 0000000000000000
159 SigBlk: 0000000000000000
160 SigIgn: 0000000000000000
161 SigCgt: 0000000000000000
162 CapInh: 00000000fffffeff
163 CapPrm: 0000000000000000
164 CapEff: 0000000000000000
167 This shows you nearly the same information you would get if you viewed it with
168 the ps command. In fact, ps uses the proc file system to obtain its
169 information. The statm file contains more detailed information about the
170 process memory usage. Its seven fields are explained in Table 1-2.
173 Table 1-2: Contents of the statm files (as of 2.6.8-rc3)
174 ..............................................................................
176 size total program size (pages) (same as VmSize in status)
177 resident size of memory portions (pages) (same as VmRSS in status)
178 shared number of pages that are shared (i.e. backed by a file)
179 trs number of pages that are 'code' (not including libs; broken,
180 includes data segment)
181 lrs number of pages of library (always 0 on 2.6)
182 drs number of pages of data/stack (including libs; broken,
183 includes library text)
184 dt number of dirty pages (always 0 on 2.6)
185 ..............................................................................
190 Similar to the process entries, the kernel data files give information about
191 the running kernel. The files used to obtain this information are contained in
192 /proc and are listed in Table 1-3. Not all of these will be present in your
193 system. It depends on the kernel configuration and the loaded modules, which
194 files are there, and which are missing.
196 Table 1-3: Kernel info in /proc
197 ..............................................................................
199 apm Advanced power management info
200 buddyinfo Kernel memory allocator information (see text) (2.5)
201 bus Directory containing bus specific information
202 cmdline Kernel command line
203 cpuinfo Info about the CPU
204 devices Available devices (block and character)
205 dma Used DMS channels
206 filesystems Supported filesystems
207 driver Various drivers grouped here, currently rtc (2.4)
208 execdomains Execdomains, related to security (2.4)
209 fb Frame Buffer devices (2.4)
210 fs File system parameters, currently nfs/exports (2.4)
211 ide Directory containing info about the IDE subsystem
212 interrupts Interrupt usage
213 iomem Memory map (2.4)
214 ioports I/O port usage
215 irq Masks for irq to cpu affinity (2.4)(smp?)
216 isapnp ISA PnP (Plug&Play) Info (2.4)
217 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
219 ksyms Kernel symbol table
220 loadavg Load average of last 1, 5 & 15 minutes
224 modules List of loaded modules
225 mounts Mounted filesystems
226 net Networking info (see text)
227 partitions Table of partitions known to the system
228 pci Depreciated info of PCI bus (new way -> /proc/bus/pci/,
229 decoupled by lspci (2.4)
231 scsi SCSI info (see text)
232 slabinfo Slab pool info
233 stat Overall statistics
234 swaps Swap space utilization
236 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
237 tty Info of tty drivers
239 version Kernel version
240 video bttv info of video resources (2.4)
241 ..............................................................................
243 You can, for example, check which interrupts are currently in use and what
244 they are used for by looking in the file /proc/interrupts:
246 > cat /proc/interrupts
248 0: 8728810 XT-PIC timer
249 1: 895 XT-PIC keyboard
251 3: 531695 XT-PIC aha152x
252 4: 2014133 XT-PIC serial
253 5: 44401 XT-PIC pcnet_cs
256 12: 182918 XT-PIC PS/2 Mouse
258 14: 1232265 XT-PIC ide0
262 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
263 output of a SMP machine):
265 > cat /proc/interrupts
268 0: 1243498 1214548 IO-APIC-edge timer
269 1: 8949 8958 IO-APIC-edge keyboard
270 2: 0 0 XT-PIC cascade
271 5: 11286 10161 IO-APIC-edge soundblaster
272 8: 1 0 IO-APIC-edge rtc
273 9: 27422 27407 IO-APIC-edge 3c503
274 12: 113645 113873 IO-APIC-edge PS/2 Mouse
276 14: 22491 24012 IO-APIC-edge ide0
277 15: 2183 2415 IO-APIC-edge ide1
278 17: 30564 30414 IO-APIC-level eth0
279 18: 177 164 IO-APIC-level bttv
284 NMI is incremented in this case because every timer interrupt generates a NMI
285 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
287 LOC is the local interrupt counter of the internal APIC of every CPU.
289 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
290 connects the CPUs in a SMP system. This means that an error has been detected,
291 the IO-APIC automatically retry the transmission, so it should not be a big
292 problem, but you should read the SMP-FAQ.
294 In this context it could be interesting to note the new irq directory in 2.4.
295 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
296 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
297 irq subdir is one subdir for each IRQ, and one file; prof_cpu_mask
301 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
302 1 11 13 15 17 19 3 5 7 9
306 The contents of the prof_cpu_mask file and each smp_affinity file for each IRQ
307 is the same by default:
309 > cat /proc/irq/0/smp_affinity
312 It's a bitmask, in which you can specify which CPUs can handle the IRQ, you can
315 > echo 1 > /proc/irq/prof_cpu_mask
317 This means that only the first CPU will handle the IRQ, but you can also echo 5
318 which means that only the first and fourth CPU can handle the IRQ.
320 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
321 between all the CPUs which are allowed to handle it. As usual the kernel has
322 more info than you and does a better job than you, so the defaults are the
323 best choice for almost everyone.
325 There are three more important subdirectories in /proc: net, scsi, and sys.
326 The general rule is that the contents, or even the existence of these
327 directories, depend on your kernel configuration. If SCSI is not enabled, the
328 directory scsi may not exist. The same is true with the net, which is there
329 only when networking support is present in the running kernel.
331 The slabinfo file gives information about memory usage at the slab level.
332 Linux uses slab pools for memory management above page level in version 2.2.
333 Commonly used objects have their own slab pool (such as network buffers,
334 directory cache, and so on).
336 ..............................................................................
338 > cat /proc/buddyinfo
340 Node 0, zone DMA 0 4 5 4 4 3 ...
341 Node 0, zone Normal 1 0 0 1 101 8 ...
342 Node 0, zone HighMem 2 0 0 1 1 0 ...
344 Memory fragmentation is a problem under some workloads, and buddyinfo is a
345 useful tool for helping diagnose these problems. Buddyinfo will give you a
346 clue as to how big an area you can safely allocate, or why a previous
349 Each column represents the number of pages of a certain order which are
350 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
351 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
352 available in ZONE_NORMAL, etc...
354 ..............................................................................
358 Provides information about distribution and utilization of memory. This
359 varies by architecture and compile options. The following is from a
360 16GB PIII, which has highmem enabled. You may not have all of these fields.
365 MemTotal: 16344972 kB
372 HighTotal: 15597528 kB
373 HighFree: 13629632 kB
382 CommitLimit: 7669796 kB
383 Committed_AS: 100056 kB
385 VmallocTotal: 112216 kB
387 VmallocChunk: 111088 kB
389 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
390 bits and the kernel binary code)
391 MemFree: The sum of LowFree+HighFree
392 Buffers: Relatively temporary storage for raw disk blocks
393 shouldn't get tremendously large (20MB or so)
394 Cached: in-memory cache for files read from the disk (the
395 pagecache). Doesn't include SwapCached
396 SwapCached: Memory that once was swapped out, is swapped back in but
397 still also is in the swapfile (if memory is needed it
398 doesn't need to be swapped out AGAIN because it is already
399 in the swapfile. This saves I/O)
400 Active: Memory that has been used more recently and usually not
401 reclaimed unless absolutely necessary.
402 Inactive: Memory which has been less recently used. It is more
403 eligible to be reclaimed for other purposes
405 HighFree: Highmem is all memory above ~860MB of physical memory
406 Highmem areas are for use by userspace programs, or
407 for the pagecache. The kernel must use tricks to access
408 this memory, making it slower to access than lowmem.
410 LowFree: Lowmem is memory which can be used for everything that
411 highmem can be used for, but it is also availble for the
412 kernel's use for its own data structures. Among many
413 other things, it is where everything from the Slab is
414 allocated. Bad things happen when you're out of lowmem.
415 SwapTotal: total amount of swap space available
416 SwapFree: Memory which has been evicted from RAM, and is temporarily
418 Dirty: Memory which is waiting to get written back to the disk
419 Writeback: Memory which is actively being written back to the disk
420 Mapped: files which have been mmaped, such as libraries
421 Slab: in-kernel data structures cache
422 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
423 this is the total amount of memory currently available to
424 be allocated on the system. This limit is only adhered to
425 if strict overcommit accounting is enabled (mode 2 in
426 'vm.overcommit_memory').
427 The CommitLimit is calculated with the following formula:
428 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap
429 For example, on a system with 1G of physical RAM and 7G
430 of swap with a `vm.overcommit_ratio` of 30 it would
431 yield a CommitLimit of 7.3G.
432 For more details, see the memory overcommit documentation
433 in vm/overcommit-accounting.
434 Committed_AS: The amount of memory presently allocated on the system.
435 The committed memory is a sum of all of the memory which
436 has been allocated by processes, even if it has not been
437 "used" by them as of yet. A process which malloc()'s 1G
438 of memory, but only touches 300M of it will only show up
439 as using 300M of memory even if it has the address space
440 allocated for the entire 1G. This 1G is memory which has
441 been "committed" to by the VM and can be used at any time
442 by the allocating application. With strict overcommit
443 enabled on the system (mode 2 in 'vm.overcommit_memory'),
444 allocations which would exceed the CommitLimit (detailed
445 above) will not be permitted. This is useful if one needs
446 to guarantee that processes will not fail due to lack of
447 memory once that memory has been successfully allocated.
448 PageTables: amount of memory dedicated to the lowest level of page
450 VmallocTotal: total size of vmalloc memory area
451 VmallocUsed: amount of vmalloc area which is used
452 VmallocChunk: largest contigious block of vmalloc area which is free
455 1.3 IDE devices in /proc/ide
456 ----------------------------
458 The subdirectory /proc/ide contains information about all IDE devices of which
459 the kernel is aware. There is one subdirectory for each IDE controller, the
460 file drivers and a link for each IDE device, pointing to the device directory
461 in the controller specific subtree.
463 The file drivers contains general information about the drivers used for the
466 > cat /proc/ide/drivers
467 ide-cdrom version 4.53
468 ide-disk version 1.08
470 More detailed information can be found in the controller specific
471 subdirectories. These are named ide0, ide1 and so on. Each of these
472 directories contains the files shown in table 1-4.
475 Table 1-4: IDE controller info in /proc/ide/ide?
476 ..............................................................................
478 channel IDE channel (0 or 1)
479 config Configuration (only for PCI/IDE bridge)
481 model Type/Chipset of IDE controller
482 ..............................................................................
484 Each device connected to a controller has a separate subdirectory in the
485 controllers directory. The files listed in table 1-5 are contained in these
489 Table 1-5: IDE device information
490 ..............................................................................
493 capacity Capacity of the medium (in 512Byte blocks)
494 driver driver and version
495 geometry physical and logical geometry
496 identify device identify block
498 model device identifier
499 settings device setup
500 smart_thresholds IDE disk management thresholds
501 smart_values IDE disk management values
502 ..............................................................................
504 The most interesting file is settings. This file contains a nice overview of
505 the drive parameters:
507 # cat /proc/ide/ide0/hda/settings
508 name value min max mode
509 ---- ----- --- --- ----
510 bios_cyl 526 0 65535 rw
511 bios_head 255 0 255 rw
513 breada_readahead 4 0 127 rw
515 file_readahead 72 0 2097151 rw
517 keepsettings 0 0 1 rw
518 max_kb_per_request 122 1 127 rw
522 pio_mode write-only 0 255 w
528 1.4 Networking info in /proc/net
529 --------------------------------
531 The subdirectory /proc/net follows the usual pattern. Table 1-6 shows the
532 additional values you get for IP version 6 if you configure the kernel to
533 support this. Table 1-7 lists the files and their meaning.
536 Table 1-6: IPv6 info in /proc/net
537 ..............................................................................
539 udp6 UDP sockets (IPv6)
540 tcp6 TCP sockets (IPv6)
541 raw6 Raw device statistics (IPv6)
542 igmp6 IP multicast addresses, which this host joined (IPv6)
543 if_inet6 List of IPv6 interface addresses
544 ipv6_route Kernel routing table for IPv6
545 rt6_stats Global IPv6 routing tables statistics
546 sockstat6 Socket statistics (IPv6)
547 snmp6 Snmp data (IPv6)
548 ..............................................................................
551 Table 1-7: Network info in /proc/net
552 ..............................................................................
555 dev network devices with statistics
556 dev_mcast the Layer2 multicast groups a device is listening too
557 (interface index, label, number of references, number of bound
559 dev_stat network device status
560 ip_fwchains Firewall chain linkage
561 ip_fwnames Firewall chain names
562 ip_masq Directory containing the masquerading tables
563 ip_masquerade Major masquerading table
564 netstat Network statistics
565 raw raw device statistics
566 route Kernel routing table
567 rpc Directory containing rpc info
568 rt_cache Routing cache
570 sockstat Socket statistics
572 tr_rif Token ring RIF routing table
574 unix UNIX domain sockets
575 wireless Wireless interface data (Wavelan etc)
576 igmp IP multicast addresses, which this host joined
577 psched Global packet scheduler parameters.
578 netlink List of PF_NETLINK sockets
579 ip_mr_vifs List of multicast virtual interfaces
580 ip_mr_cache List of multicast routing cache
581 ..............................................................................
583 You can use this information to see which network devices are available in
584 your system and how much traffic was routed over those devices:
588 face |bytes packets errs drop fifo frame compressed multicast|[...
589 lo: 908188 5596 0 0 0 0 0 0 [...
590 ppp0:15475140 20721 410 0 0 410 0 0 [...
591 eth0: 614530 7085 0 0 0 0 0 1 [...
594 ...] bytes packets errs drop fifo colls carrier compressed
595 ...] 908188 5596 0 0 0 0 0 0
596 ...] 1375103 17405 0 0 0 0 0 0
597 ...] 1703981 5535 0 0 0 3 0 0
599 In addition, each Channel Bond interface has it's own directory. For
600 example, the bond0 device will have a directory called /proc/net/bond0/.
601 It will contain information that is specific to that bond, such as the
602 current slaves of the bond, the link status of the slaves, and how
603 many times the slaves link has failed.
608 If you have a SCSI host adapter in your system, you'll find a subdirectory
609 named after the driver for this adapter in /proc/scsi. You'll also see a list
610 of all recognized SCSI devices in /proc/scsi:
614 Host: scsi0 Channel: 00 Id: 00 Lun: 00
615 Vendor: IBM Model: DGHS09U Rev: 03E0
616 Type: Direct-Access ANSI SCSI revision: 03
617 Host: scsi0 Channel: 00 Id: 06 Lun: 00
618 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
619 Type: CD-ROM ANSI SCSI revision: 02
622 The directory named after the driver has one file for each adapter found in
623 the system. These files contain information about the controller, including
624 the used IRQ and the IO address range. The amount of information shown is
625 dependent on the adapter you use. The example shows the output for an Adaptec
626 AHA-2940 SCSI adapter:
628 > cat /proc/scsi/aic7xxx/0
630 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
632 TCQ Enabled By Default : Disabled
633 AIC7XXX_PROC_STATS : Disabled
634 AIC7XXX_RESET_DELAY : 5
635 Adapter Configuration:
636 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
637 Ultra Wide Controller
638 PCI MMAPed I/O Base: 0xeb001000
639 Adapter SEEPROM Config: SEEPROM found and used.
640 Adaptec SCSI BIOS: Enabled
642 SCBs: Active 0, Max Active 2,
643 Allocated 15, HW 16, Page 255
645 BIOS Control Word: 0x18b6
646 Adapter Control Word: 0x005b
647 Extended Translation: Enabled
648 Disconnect Enable Flags: 0xffff
649 Ultra Enable Flags: 0x0001
650 Tag Queue Enable Flags: 0x0000
651 Ordered Queue Tag Flags: 0x0000
652 Default Tag Queue Depth: 8
653 Tagged Queue By Device array for aic7xxx host instance 0:
654 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
655 Actual queue depth per device for aic7xxx host instance 0:
656 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
659 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
660 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
661 Total transfers 160151 (74577 reads and 85574 writes)
663 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
664 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
665 Total transfers 0 (0 reads and 0 writes)
668 1.6 Parallel port info in /proc/parport
669 ---------------------------------------
671 The directory /proc/parport contains information about the parallel ports of
672 your system. It has one subdirectory for each port, named after the port
675 These directories contain the four files shown in Table 1-8.
678 Table 1-8: Files in /proc/parport
679 ..............................................................................
681 autoprobe Any IEEE-1284 device ID information that has been acquired.
682 devices list of the device drivers using that port. A + will appear by the
683 name of the device currently using the port (it might not appear
685 hardware Parallel port's base address, IRQ line and DMA channel.
686 irq IRQ that parport is using for that port. This is in a separate
687 file to allow you to alter it by writing a new value in (IRQ
689 ..............................................................................
691 1.7 TTY info in /proc/tty
692 -------------------------
694 Information about the available and actually used tty's can be found in the
695 directory /proc/tty.You'll find entries for drivers and line disciplines in
696 this directory, as shown in Table 1-9.
699 Table 1-9: Files in /proc/tty
700 ..............................................................................
702 drivers list of drivers and their usage
703 ldiscs registered line disciplines
704 driver/serial usage statistic and status of single tty lines
705 ..............................................................................
707 To see which tty's are currently in use, you can simply look into the file
710 > cat /proc/tty/drivers
711 pty_slave /dev/pts 136 0-255 pty:slave
712 pty_master /dev/ptm 128 0-255 pty:master
713 pty_slave /dev/ttyp 3 0-255 pty:slave
714 pty_master /dev/pty 2 0-255 pty:master
715 serial /dev/cua 5 64-67 serial:callout
716 serial /dev/ttyS 4 64-67 serial
717 /dev/tty0 /dev/tty0 4 0 system:vtmaster
718 /dev/ptmx /dev/ptmx 5 2 system
719 /dev/console /dev/console 5 1 system:console
720 /dev/tty /dev/tty 5 0 system:/dev/tty
721 unknown /dev/tty 4 1-63 console
724 1.8 Miscellaneous kernel statistics in /proc/stat
725 -------------------------------------------------
727 Various pieces of information about kernel activity are available in the
728 /proc/stat file. All of the numbers reported in this file are aggregates
729 since the system first booted. For a quick look, simply cat the file:
732 cpu 2255 34 2290 22625563 6290 127 456
733 cpu0 1132 34 1441 11311718 3675 127 438
734 cpu1 1123 0 849 11313845 2614 0 18
735 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
742 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
743 lines. These numbers identify the amount of time the CPU has spent performing
744 different kinds of work. Time units are in USER_HZ (typically hundredths of a
745 second). The meanings of the columns are as follows, from left to right:
747 - user: normal processes executing in user mode
748 - nice: niced processes executing in user mode
749 - system: processes executing in kernel mode
750 - idle: twiddling thumbs
751 - iowait: waiting for I/O to complete
752 - irq: servicing interrupts
753 - softirq: servicing softirqs
755 The "intr" line gives counts of interrupts serviced since boot time, for each
756 of the possible system interrupts. The first column is the total of all
757 interrupts serviced; each subsequent column is the total for that particular
760 The "ctxt" line gives the total number of context switches across all CPUs.
762 The "btime" line gives the time at which the system booted, in seconds since
765 The "processes" line gives the number of processes and threads created, which
766 includes (but is not limited to) those created by calls to the fork() and
767 clone() system calls.
769 The "procs_running" line gives the number of processes currently running on
772 The "procs_blocked" line gives the number of processes currently blocked,
773 waiting for I/O to complete.
776 ------------------------------------------------------------------------------
778 ------------------------------------------------------------------------------
779 The /proc file system serves information about the running system. It not only
780 allows access to process data but also allows you to request the kernel status
781 by reading files in the hierarchy.
783 The directory structure of /proc reflects the types of information and makes
784 it easy, if not obvious, where to look for specific data.
785 ------------------------------------------------------------------------------
787 ------------------------------------------------------------------------------
788 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
789 ------------------------------------------------------------------------------
791 ------------------------------------------------------------------------------
793 ------------------------------------------------------------------------------
794 * Modifying kernel parameters by writing into files found in /proc/sys
795 * Exploring the files which modify certain parameters
796 * Review of the /proc/sys file tree
797 ------------------------------------------------------------------------------
800 A very interesting part of /proc is the directory /proc/sys. This is not only
801 a source of information, it also allows you to change parameters within the
802 kernel. Be very careful when attempting this. You can optimize your system,
803 but you can also cause it to crash. Never alter kernel parameters on a
804 production system. Set up a development machine and test to make sure that
805 everything works the way you want it to. You may have no alternative but to
806 reboot the machine once an error has been made.
808 To change a value, simply echo the new value into the file. An example is
809 given below in the section on the file system data. You need to be root to do
810 this. You can create your own boot script to perform this every time your
813 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
814 general things in the operation of the Linux kernel. Since some of the files
815 can inadvertently disrupt your system, it is advisable to read both
816 documentation and source before actually making adjustments. In any case, be
817 very careful when writing to any of these files. The entries in /proc may
818 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
819 review the kernel documentation in the directory /usr/src/linux/Documentation.
820 This chapter is heavily based on the documentation included in the pre 2.2
821 kernels, and became part of it in version 2.2.1 of the Linux kernel.
823 2.1 /proc/sys/fs - File system data
824 -----------------------------------
826 This subdirectory contains specific file system, file handle, inode, dentry
827 and quota information.
829 Currently, these files are in /proc/sys/fs:
834 Status of the directory cache. Since directory entries are dynamically
835 allocated and deallocated, this file indicates the current status. It holds
836 six values, in which the last two are not used and are always zero. The others
837 are listed in table 2-1.
840 Table 2-1: Status files of the directory cache
841 ..............................................................................
843 nr_dentry Almost always zero
844 nr_unused Number of unused cache entries
846 in seconds after the entry may be reclaimed, when memory is short
847 want_pages internally
848 ..............................................................................
850 dquot-nr and dquot-max
851 ----------------------
853 The file dquot-max shows the maximum number of cached disk quota entries.
855 The file dquot-nr shows the number of allocated disk quota entries and the
856 number of free disk quota entries.
858 If the number of available cached disk quotas is very low and you have a large
859 number of simultaneous system users, you might want to raise the limit.
864 The kernel allocates file handles dynamically, but doesn't free them again at
867 The value in file-max denotes the maximum number of file handles that the
868 Linux kernel will allocate. When you get a lot of error messages about running
869 out of file handles, you might want to raise this limit. The default value is
870 10% of RAM in kilobytes. To change it, just write the new number into the
873 # cat /proc/sys/fs/file-max
875 # echo 8192 > /proc/sys/fs/file-max
876 # cat /proc/sys/fs/file-max
880 This method of revision is useful for all customizable parameters of the
881 kernel - simply echo the new value to the corresponding file.
883 Historically, the three values in file-nr denoted the number of allocated file
884 handles, the number of allocated but unused file handles, and the maximum
885 number of file handles. Linux 2.6 always reports 0 as the number of free file
886 handles -- this is not an error, it just means that the number of allocated
887 file handles exactly matches the number of used file handles.
889 Attempts to allocate more file descriptors than file-max are reported with
890 printk, look for "VFS: file-max limit <number> reached".
892 inode-state and inode-nr
893 ------------------------
895 The file inode-nr contains the first two items from inode-state, so we'll skip
898 inode-state contains two actual numbers and five dummy values. The numbers
899 are nr_inodes and nr_free_inodes (in order of appearance).
904 Denotes the number of inodes the system has allocated. This number will
905 grow and shrink dynamically.
910 Represents the number of free inodes. Ie. The number of inuse inodes is
911 (nr_inodes - nr_free_inodes).
913 aio-nr and aio-max-nr
914 ---------------------
916 aio-nr is the running total of the number of events specified on the
917 io_setup system call for all currently active aio contexts. If aio-nr
918 reaches aio-max-nr then io_setup will fail with EAGAIN. Note that
919 raising aio-max-nr does not result in the pre-allocation or re-sizing
920 of any kernel data structures.
922 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
923 -----------------------------------------------------------
925 Besides these files, there is the subdirectory /proc/sys/fs/binfmt_misc. This
926 handles the kernel support for miscellaneous binary formats.
928 Binfmt_misc provides the ability to register additional binary formats to the
929 Kernel without compiling an additional module/kernel. Therefore, binfmt_misc
930 needs to know magic numbers at the beginning or the filename extension of the
933 It works by maintaining a linked list of structs that contain a description of
934 a binary format, including a magic with size (or the filename extension),
935 offset and mask, and the interpreter name. On request it invokes the given
936 interpreter with the original program as argument, as binfmt_java and
937 binfmt_em86 and binfmt_mz do. Since binfmt_misc does not define any default
938 binary-formats, you have to register an additional binary-format.
940 There are two general files in binfmt_misc and one file per registered format.
941 The two general files are register and status.
943 Registering a new binary format
944 -------------------------------
946 To register a new binary format you have to issue the command
948 echo :name:type:offset:magic:mask:interpreter: > /proc/sys/fs/binfmt_misc/register
952 with appropriate name (the name for the /proc-dir entry), offset (defaults to
953 0, if omitted), magic, mask (which can be omitted, defaults to all 0xff) and
954 last but not least, the interpreter that is to be invoked (for example and
955 testing /bin/echo). Type can be M for usual magic matching or E for filename
956 extension matching (give extension in place of magic).
958 Check or reset the status of the binary format handler
959 ------------------------------------------------------
961 If you do a cat on the file /proc/sys/fs/binfmt_misc/status, you will get the
962 current status (enabled/disabled) of binfmt_misc. Change the status by echoing
963 0 (disables) or 1 (enables) or -1 (caution: this clears all previously
964 registered binary formats) to status. For example echo 0 > status to disable
965 binfmt_misc (temporarily).
967 Status of a single handler
968 --------------------------
970 Each registered handler has an entry in /proc/sys/fs/binfmt_misc. These files
971 perform the same function as status, but their scope is limited to the actual
972 binary format. By cating this file, you also receive all related information
973 about the interpreter/magic of the binfmt.
975 Example usage of binfmt_misc (emulate binfmt_java)
976 --------------------------------------------------
978 cd /proc/sys/fs/binfmt_misc
979 echo ':Java:M::\xca\xfe\xba\xbe::/usr/local/java/bin/javawrapper:' > register
980 echo ':HTML:E::html::/usr/local/java/bin/appletviewer:' > register
981 echo ':Applet:M::<!--applet::/usr/local/java/bin/appletviewer:' > register
982 echo ':DEXE:M::\x0eDEX::/usr/bin/dosexec:' > register
985 These four lines add support for Java executables and Java applets (like
986 binfmt_java, additionally recognizing the .html extension with no need to put
987 <!--applet> to every applet file). You have to install the JDK and the
988 shell-script /usr/local/java/bin/javawrapper too. It works around the
989 brokenness of the Java filename handling. To add a Java binary, just create a
990 link to the class-file somewhere in the path.
992 2.3 /proc/sys/kernel - general kernel parameters
993 ------------------------------------------------
995 This directory reflects general kernel behaviors. As I've said before, the
996 contents depend on your configuration. Here you'll find the most important
997 files, along with descriptions of what they mean and how to use them.
1002 The file contains three values; highwater, lowwater, and frequency.
1004 It exists only when BSD-style process accounting is enabled. These values
1005 control its behavior. If the free space on the file system where the log lives
1006 goes below lowwater percentage, accounting suspends. If it goes above
1007 highwater percentage, accounting resumes. Frequency determines how often you
1008 check the amount of free space (value is in seconds). Default settings are: 4,
1009 2, and 30. That is, suspend accounting if there is less than 2 percent free;
1010 resume it if we have a value of 3 or more percent; consider information about
1011 the amount of free space valid for 30 seconds
1016 When the value in this file is 0, ctrl-alt-del is trapped and sent to the init
1017 program to handle a graceful restart. However, when the value is greater that
1018 zero, Linux's reaction to this key combination will be an immediate reboot,
1019 without syncing its dirty buffers.
1022 When a program (like dosemu) has the keyboard in raw mode, the
1023 ctrl-alt-del is intercepted by the program before it ever reaches the
1024 kernel tty layer, and it is up to the program to decide what to do with
1027 domainname and hostname
1028 -----------------------
1030 These files can be controlled to set the NIS domainname and hostname of your
1031 box. For the classic darkstar.frop.org a simple:
1033 # echo "darkstar" > /proc/sys/kernel/hostname
1034 # echo "frop.org" > /proc/sys/kernel/domainname
1037 would suffice to set your hostname and NIS domainname.
1039 osrelease, ostype and version
1040 -----------------------------
1042 The names make it pretty obvious what these fields contain:
1044 > cat /proc/sys/kernel/osrelease
1047 > cat /proc/sys/kernel/ostype
1050 > cat /proc/sys/kernel/version
1051 #4 Fri Oct 1 12:41:14 PDT 1999
1054 The files osrelease and ostype should be clear enough. Version needs a little
1055 more clarification. The #4 means that this is the 4th kernel built from this
1056 source base and the date after it indicates the time the kernel was built. The
1057 only way to tune these values is to rebuild the kernel.
1062 The value in this file represents the number of seconds the kernel waits
1063 before rebooting on a panic. When you use the software watchdog, the
1064 recommended setting is 60. If set to 0, the auto reboot after a kernel panic
1065 is disabled, which is the default setting.
1070 The four values in printk denote
1072 * default_message_loglevel,
1073 * minimum_console_loglevel and
1074 * default_console_loglevel
1077 These values influence printk() behavior when printing or logging error
1078 messages, which come from inside the kernel. See syslog(2) for more
1079 information on the different log levels.
1084 Messages with a higher priority than this will be printed to the console.
1086 default_message_level
1087 ---------------------
1089 Messages without an explicit priority will be printed with this priority.
1091 minimum_console_loglevel
1092 ------------------------
1094 Minimum (highest) value to which the console_loglevel can be set.
1096 default_console_loglevel
1097 ------------------------
1099 Default value for console_loglevel.
1104 This file shows the size of the generic SCSI (sg) buffer. At this point, you
1105 can't tune it yet, but you can change it at compile time by editing
1106 include/scsi/sg.h and changing the value of SG_BIG_BUFF.
1108 If you use a scanner with SANE (Scanner Access Now Easy) you might want to set
1109 this to a higher value. Refer to the SANE documentation on this issue.
1114 The location where the modprobe binary is located. The kernel uses this
1115 program to load modules on demand.
1120 The value in this file affects behavior of handling NMI. When the value is
1121 non-zero, unknown NMI is trapped and then panic occurs. At that time, kernel
1122 debugging information is displayed on console.
1124 NMI switch that most IA32 servers have fires unknown NMI up, for example.
1125 If a system hangs up, try pressing the NMI switch.
1130 Enables/Disables the NMI watchdog on x86 systems. When the value is non-zero
1131 the NMI watchdog is enabled and will continuously test all online cpus to
1132 determine whether or not they are still functioning properly.
1134 Because the NMI watchdog shares registers with oprofile, by disabling the NMI
1135 watchdog, oprofile may have more registers to utilize.
1138 2.4 /proc/sys/vm - The virtual memory subsystem
1139 -----------------------------------------------
1141 The files in this directory can be used to tune the operation of the virtual
1142 memory (VM) subsystem of the Linux kernel.
1147 Controls the tendency of the kernel to reclaim the memory which is used for
1148 caching of directory and inode objects.
1150 At the default value of vfs_cache_pressure=100 the kernel will attempt to
1151 reclaim dentries and inodes at a "fair" rate with respect to pagecache and
1152 swapcache reclaim. Decreasing vfs_cache_pressure causes the kernel to prefer
1153 to retain dentry and inode caches. Increasing vfs_cache_pressure beyond 100
1154 causes the kernel to prefer to reclaim dentries and inodes.
1156 dirty_background_ratio
1157 ----------------------
1159 Contains, as a percentage of total system memory, the number of pages at which
1160 the pdflush background writeback daemon will start writing out dirty data.
1165 Contains, as a percentage of total system memory, the number of pages at which
1166 a process which is generating disk writes will itself start writing out dirty
1169 dirty_writeback_centisecs
1170 -------------------------
1172 The pdflush writeback daemons will periodically wake up and write `old' data
1173 out to disk. This tunable expresses the interval between those wakeups, in
1174 100'ths of a second.
1176 Setting this to zero disables periodic writeback altogether.
1178 dirty_expire_centisecs
1179 ----------------------
1181 This tunable is used to define when dirty data is old enough to be eligible
1182 for writeout by the pdflush daemons. It is expressed in 100'ths of a second.
1183 Data which has been dirty in-memory for longer than this interval will be
1184 written out next time a pdflush daemon wakes up.
1189 If non-zero, this sysctl disables the new 32-bit mmap mmap layout - the kernel
1190 will use the legacy (2.4) layout for all processes.
1192 lower_zone_protection
1193 ---------------------
1195 For some specialised workloads on highmem machines it is dangerous for
1196 the kernel to allow process memory to be allocated from the "lowmem"
1197 zone. This is because that memory could then be pinned via the mlock()
1198 system call, or by unavailability of swapspace.
1200 And on large highmem machines this lack of reclaimable lowmem memory
1203 So the Linux page allocator has a mechanism which prevents allocations
1204 which _could_ use highmem from using too much lowmem. This means that
1205 a certain amount of lowmem is defended from the possibility of being
1206 captured into pinned user memory.
1208 (The same argument applies to the old 16 megabyte ISA DMA region. This
1209 mechanism will also defend that region from allocations which could use
1212 The `lower_zone_protection' tunable determines how aggressive the kernel is
1213 in defending these lower zones. The default value is zero - no
1216 If you have a machine which uses highmem or ISA DMA and your
1217 applications are using mlock(), or if you are running with no swap then
1218 you probably should increase the lower_zone_protection setting.
1220 The units of this tunable are fairly vague. It is approximately equal
1221 to "megabytes". So setting lower_zone_protection=100 will protect around 100
1222 megabytes of the lowmem zone from user allocations. It will also make
1223 those 100 megabytes unavaliable for use by applications and by
1224 pagecache, so there is a cost.
1226 The effects of this tunable may be observed by monitoring
1227 /proc/meminfo:LowFree. Write a single huge file and observe the point
1228 at which LowFree ceases to fall.
1230 A reasonable value for lower_zone_protection is 100.
1235 page-cluster controls the number of pages which are written to swap in
1236 a single attempt. The swap I/O size.
1238 It is a logarithmic value - setting it to zero means "1 page", setting
1239 it to 1 means "2 pages", setting it to 2 means "4 pages", etc.
1241 The default value is three (eight pages at a time). There may be some
1242 small benefits in tuning this to a different value if your workload is
1248 Controls overcommit of system memory, possibly allowing processes
1249 to allocate (but not use) more memory than is actually available.
1252 0 - Heuristic overcommit handling. Obvious overcommits of
1253 address space are refused. Used for a typical system. It
1254 ensures a seriously wild allocation fails while allowing
1255 overcommit to reduce swap usage. root is allowed to
1256 allocate slighly more memory in this mode. This is the
1259 1 - Always overcommit. Appropriate for some scientific
1262 2 - Don't overcommit. The total address space commit
1263 for the system is not permitted to exceed swap plus a
1264 configurable percentage (default is 50) of physical RAM.
1265 Depending on the percentage you use, in most situations
1266 this means a process will not be killed while attempting
1267 to use already-allocated memory but will receive errors
1268 on memory allocation as appropriate.
1273 Percentage of physical memory size to include in overcommit calculations
1276 Memory allocation limit = swapspace + physmem * (overcommit_ratio / 100)
1278 swapspace = total size of all swap areas
1279 physmem = size of physical memory in system
1281 nr_hugepages and hugetlb_shm_group
1282 ----------------------------------
1284 nr_hugepages configures number of hugetlb page reserved for the system.
1286 hugetlb_shm_group contains group id that is allowed to create SysV shared
1287 memory segment using hugetlb page.
1292 laptop_mode is a knob that controls "laptop mode". All the things that are
1293 controlled by this knob are discussed in Documentation/laptop-mode.txt.
1298 block_dump enables block I/O debugging when set to a nonzero value. More
1299 information on block I/O debugging is in Documentation/laptop-mode.txt.
1304 This file contains valid hold time of swap out protection token. The Linux
1305 VM has token based thrashing control mechanism and uses the token to prevent
1306 unnecessary page faults in thrashing situation. The unit of the value is
1307 second. The value would be useful to tune thrashing behavior.
1312 Writing to this will cause the kernel to drop clean caches, dentries and
1313 inodes from memory, causing that memory to become free.
1316 echo 1 > /proc/sys/vm/drop_caches
1317 To free dentries and inodes:
1318 echo 2 > /proc/sys/vm/drop_caches
1319 To free pagecache, dentries and inodes:
1320 echo 3 > /proc/sys/vm/drop_caches
1322 As this is a non-destructive operation and dirty objects are not freeable, the
1323 user should run `sync' first.
1326 2.5 /proc/sys/dev - Device specific parameters
1327 ----------------------------------------------
1329 Currently there is only support for CDROM drives, and for those, there is only
1330 one read-only file containing information about the CD-ROM drives attached to
1333 >cat /proc/sys/dev/cdrom/info
1334 CD-ROM information, Id: cdrom.c 2.55 1999/04/25
1338 drive # of slots: 1 0
1342 Can change speed: 1 1
1343 Can select disk: 0 1
1344 Can read multisession: 1 1
1346 Reports media changed: 1 1
1350 You see two drives, sr0 and hdb, along with a list of their features.
1352 2.6 /proc/sys/sunrpc - Remote procedure calls
1353 ---------------------------------------------
1355 This directory contains four files, which enable or disable debugging for the
1356 RPC functions NFS, NFS-daemon, RPC and NLM. The default values are 0. They can
1357 be set to one to turn debugging on. (The default value is 0 for each)
1359 2.7 /proc/sys/net - Networking stuff
1360 ------------------------------------
1362 The interface to the networking parts of the kernel is located in
1363 /proc/sys/net. Table 2-3 shows all possible subdirectories. You may see only
1364 some of them, depending on your kernel's configuration.
1367 Table 2-3: Subdirectories in /proc/sys/net
1368 ..............................................................................
1369 Directory Content Directory Content
1370 core General parameter appletalk Appletalk protocol
1371 unix Unix domain sockets netrom NET/ROM
1372 802 E802 protocol ax25 AX25
1373 ethernet Ethernet protocol rose X.25 PLP layer
1374 ipv4 IP version 4 x25 X.25 protocol
1375 ipx IPX token-ring IBM token ring
1376 bridge Bridging decnet DEC net
1378 ..............................................................................
1380 We will concentrate on IP networking here. Since AX15, X.25, and DEC Net are
1381 only minor players in the Linux world, we'll skip them in this chapter. You'll
1382 find some short info on Appletalk and IPX further on in this chapter. Review
1383 the online documentation and the kernel source to get a detailed view of the
1384 parameters for those protocols. In this section we'll discuss the
1385 subdirectories printed in bold letters in the table above. As default values
1386 are suitable for most needs, there is no need to change these values.
1388 /proc/sys/net/core - Network core options
1389 -----------------------------------------
1394 The default setting of the socket receive buffer in bytes.
1399 The maximum receive socket buffer size in bytes.
1404 The default setting (in bytes) of the socket send buffer.
1409 The maximum send socket buffer size in bytes.
1411 message_burst and message_cost
1412 ------------------------------
1414 These parameters are used to limit the warning messages written to the kernel
1415 log from the networking code. They enforce a rate limit to make a
1416 denial-of-service attack impossible. A higher message_cost factor, results in
1417 fewer messages that will be written. Message_burst controls when messages will
1418 be dropped. The default settings limit warning messages to one every five
1424 Maximum number of packets, queued on the INPUT side, when the interface
1425 receives packets faster than kernel can process them.
1430 Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
1431 of struct cmsghdr structures with appended data.
1433 /proc/sys/net/unix - Parameters for Unix domain sockets
1434 -------------------------------------------------------
1436 There are only two files in this subdirectory. They control the delays for
1437 deleting and destroying socket descriptors.
1439 2.8 /proc/sys/net/ipv4 - IPV4 settings
1440 --------------------------------------
1442 IP version 4 is still the most used protocol in Unix networking. It will be
1443 replaced by IP version 6 in the next couple of years, but for the moment it's
1444 the de facto standard for the internet and is used in most networking
1445 environments around the world. Because of the importance of this protocol,
1446 we'll have a deeper look into the subtree controlling the behavior of the IPv4
1447 subsystem of the Linux kernel.
1449 Let's start with the entries in /proc/sys/net/ipv4.
1454 icmp_echo_ignore_all and icmp_echo_ignore_broadcasts
1455 ----------------------------------------------------
1457 Turn on (1) or off (0), if the kernel should ignore all ICMP ECHO requests, or
1458 just those to broadcast and multicast addresses.
1460 Please note that if you accept ICMP echo requests with a broadcast/multi\-cast
1461 destination address your network may be used as an exploder for denial of
1462 service packet flooding attacks to other hosts.
1464 icmp_destunreach_rate, icmp_echoreply_rate, icmp_paramprob_rate and icmp_timeexeed_rate
1465 ---------------------------------------------------------------------------------------
1467 Sets limits for sending ICMP packets to specific targets. A value of zero
1468 disables all limiting. Any positive value sets the maximum package rate in
1469 hundredth of a second (on Intel systems).
1477 This file contains the number one if the host received its IP configuration by
1478 RARP, BOOTP, DHCP or a similar mechanism. Otherwise it is zero.
1483 TTL (Time To Live) for IPv4 interfaces. This is simply the maximum number of
1484 hops a packet may travel.
1489 Enable dynamic socket address rewriting on interface address change. This is
1490 useful for dialup interface with changing IP addresses.
1495 Enable or disable forwarding of IP packages between interfaces. Changing this
1496 value resets all other parameters to their default values. They differ if the
1497 kernel is configured as host or router.
1502 Range of ports used by TCP and UDP to choose the local port. Contains two
1503 numbers, the first number is the lowest port, the second number the highest
1504 local port. Default is 1024-4999. Should be changed to 32768-61000 for
1510 Global switch to turn path MTU discovery off. It can also be set on a per
1511 socket basis by the applications or on a per route basis.
1516 Enable/disable debugging of IP masquerading.
1518 IP fragmentation settings
1519 -------------------------
1521 ipfrag_high_trash and ipfrag_low_trash
1522 --------------------------------------
1524 Maximum memory used to reassemble IP fragments. When ipfrag_high_thresh bytes
1525 of memory is allocated for this purpose, the fragment handler will toss
1526 packets until ipfrag_low_thresh is reached.
1531 Time in seconds to keep an IP fragment in memory.
1539 This file controls the use of the ECN bit in the IPv4 headers, this is a new
1540 feature about Explicit Congestion Notification, but some routers and firewalls
1541 block trafic that has this bit set, so it could be necessary to echo 0 to
1542 /proc/sys/net/ipv4/tcp_ecn, if you want to talk to this sites. For more info
1543 you could read RFC2481.
1545 tcp_retrans_collapse
1546 --------------------
1548 Bug-to-bug compatibility with some broken printers. On retransmit, try to send
1549 larger packets to work around bugs in certain TCP stacks. Can be turned off by
1552 tcp_keepalive_probes
1553 --------------------
1555 Number of keep alive probes TCP sends out, until it decides that the
1556 connection is broken.
1561 How often TCP sends out keep alive messages, when keep alive is enabled. The
1567 Number of times initial SYNs for a TCP connection attempt will be
1568 retransmitted. Should not be higher than 255. This is only the timeout for
1569 outgoing connections, for incoming connections the number of retransmits is
1570 defined by tcp_retries1.
1575 Enable select acknowledgments after RFC2018.
1580 Enable timestamps as defined in RFC1323.
1585 Enable the strict RFC793 interpretation of the TCP urgent pointer field. The
1586 default is to use the BSD compatible interpretation of the urgent pointer
1587 pointing to the first byte after the urgent data. The RFC793 interpretation is
1588 to have it point to the last byte of urgent data. Enabling this option may
1589 lead to interoperatibility problems. Disabled by default.
1594 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES. Send out
1595 syncookies when the syn backlog queue of a socket overflows. This is to ward
1596 off the common 'syn flood attack'. Disabled by default.
1598 Note that the concept of a socket backlog is abandoned. This means the peer
1599 may not receive reliable error messages from an over loaded server with
1605 Enable window scaling as defined in RFC1323.
1610 The length of time in seconds it takes to receive a final FIN before the
1611 socket is always closed. This is strictly a violation of the TCP
1612 specification, but required to prevent denial-of-service attacks.
1617 Indicates how many keep alive probes are sent per slow timer run. Should not
1618 be set too high to prevent bursts.
1623 Length of the per socket backlog queue. Since Linux 2.2 the backlog specified
1624 in listen(2) only specifies the length of the backlog queue of already
1625 established sockets. When more connection requests arrive Linux starts to drop
1626 packets. When syncookies are enabled the packets are still answered and the
1627 maximum queue is effectively ignored.
1632 Defines how often an answer to a TCP connection request is retransmitted
1638 Defines how often a TCP packet is retransmitted before giving up.
1640 Interface specific settings
1641 ---------------------------
1643 In the directory /proc/sys/net/ipv4/conf you'll find one subdirectory for each
1644 interface the system knows about and one directory calls all. Changes in the
1645 all subdirectory affect all interfaces, whereas changes in the other
1646 subdirectories affect only one interface. All directories have the same
1652 This switch decides if the kernel accepts ICMP redirect messages or not. The
1653 default is 'yes' if the kernel is configured for a regular host and 'no' for a
1654 router configuration.
1659 Should source routed packages be accepted or declined. The default is
1660 dependent on the kernel configuration. It's 'yes' for routers and 'no' for
1666 Accept packets with source address 0.b.c.d with destinations not to this host
1667 as local ones. It is supposed that a BOOTP relay daemon will catch and forward
1670 The default is 0, since this feature is not implemented yet (kernel version
1676 Enable or disable IP forwarding on this interface.
1681 Log packets with source addresses with no known route to kernel log.
1686 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE and a
1687 multicast routing daemon is required.
1692 Does (1) or does not (0) perform proxy ARP.
1697 Integer value determines if a source validation should be made. 1 means yes, 0
1698 means no. Disabled by default, but local/broadcast address spoofing is always
1701 If you set this to 1 on a router that is the only connection for a network to
1702 the net, it will prevent spoofing attacks against your internal networks
1703 (external addresses can still be spoofed), without the need for additional
1709 Accept ICMP redirect messages only for gateways, listed in default gateway
1710 list. Enabled by default.
1715 If it is not set the kernel does not assume that different subnets on this
1716 device can communicate directly. Default setting is 'yes'.
1721 Determines whether to send ICMP redirects to other hosts.
1726 The directory /proc/sys/net/ipv4/route contains several file to control
1729 error_burst and error_cost
1730 --------------------------
1732 These parameters are used to limit how many ICMP destination unreachable to
1733 send from the host in question. ICMP destination unreachable messages are
1734 sent when we can not reach the next hop, while trying to transmit a packet.
1735 It will also print some error messages to kernel logs if someone is ignoring
1736 our ICMP redirects. The higher the error_cost factor is, the fewer
1737 destination unreachable and error messages will be let through. Error_burst
1738 controls when destination unreachable messages and error messages will be
1739 dropped. The default settings limit warning messages to five every second.
1744 Writing to this file results in a flush of the routing cache.
1746 gc_elasticity, gc_interval, gc_min_interval_ms, gc_timeout, gc_thresh
1747 ---------------------------------------------------------------------
1749 Values to control the frequency and behavior of the garbage collection
1750 algorithm for the routing cache. gc_min_interval is deprecated and replaced
1751 by gc_min_interval_ms.
1757 Maximum size of the routing cache. Old entries will be purged once the cache
1758 reached has this size.
1760 max_delay, min_delay
1761 --------------------
1763 Delays for flushing the routing cache.
1765 redirect_load, redirect_number
1766 ------------------------------
1768 Factors which determine if more ICPM redirects should be sent to a specific
1769 host. No redirects will be sent once the load limit or the maximum number of
1770 redirects has been reached.
1775 Timeout for redirects. After this period redirects will be sent again, even if
1776 this has been stopped, because the load or number limit has been reached.
1778 Network Neighbor handling
1779 -------------------------
1781 Settings about how to handle connections with direct neighbors (nodes attached
1782 to the same link) can be found in the directory /proc/sys/net/ipv4/neigh.
1784 As we saw it in the conf directory, there is a default subdirectory which
1785 holds the default values, and one directory for each interface. The contents
1786 of the directories are identical, with the single exception that the default
1787 settings contain additional options to set garbage collection parameters.
1789 In the interface directories you'll find the following entries:
1791 base_reachable_time, base_reachable_time_ms
1792 -------------------------------------------
1794 A base value used for computing the random reachable time value as specified
1797 Expression of base_reachable_time, which is deprecated, is in seconds.
1798 Expression of base_reachable_time_ms is in milliseconds.
1800 retrans_time, retrans_time_ms
1801 -----------------------------
1803 The time between retransmitted Neighbor Solicitation messages.
1804 Used for address resolution and to determine if a neighbor is
1807 Expression of retrans_time, which is deprecated, is in 1/100 seconds (for
1808 IPv4) or in jiffies (for IPv6).
1809 Expression of retrans_time_ms is in milliseconds.
1814 Maximum queue length for a pending arp request - the number of packets which
1815 are accepted from other layers while the ARP address is still resolved.
1820 Maximum for random delay of answers to neighbor solicitation messages in
1821 jiffies (1/100 sec). Not yet implemented (Linux does not have anycast support
1827 Maximum number of retries for unicast solicitation.
1832 Maximum number of retries for multicast solicitation.
1834 delay_first_probe_time
1835 ----------------------
1837 Delay for the first time probe if the neighbor is reachable. (see
1843 An ARP/neighbor entry is only replaced with a new one if the old is at least
1844 locktime old. This prevents ARP cache thrashing.
1849 Maximum time (real time is random [0..proxytime]) before answering to an ARP
1850 request for which we have an proxy ARP entry. In some cases, this is used to
1851 prevent network flooding.
1856 Maximum queue length of the delayed proxy arp timer. (see proxy_delay).
1861 Determines the number of requests to send to the user level ARP daemon. Use 0
1867 Determines how often to check for stale ARP entries. After an ARP entry is
1868 stale it will be resolved again (which is useful when an IP address migrates
1869 to another machine). When ucast_solicit is greater than 0 it first tries to
1870 send an ARP packet directly to the known host When that fails and
1871 mcast_solicit is greater than 0, an ARP request is broadcasted.
1876 The /proc/sys/net/appletalk directory holds the Appletalk configuration data
1877 when Appletalk is loaded. The configurable parameters are:
1882 The amount of time we keep an ARP entry before expiring it. Used to age out
1888 The amount of time we will spend trying to resolve an Appletalk address.
1890 aarp-retransmit-limit
1891 ---------------------
1893 The number of times we will retransmit a query before giving up.
1898 Controls the rate at which expires are checked.
1900 The directory /proc/net/appletalk holds the list of active Appletalk sockets
1903 The fields indicate the DDP type, the local address (in network:node format)
1904 the remote address, the size of the transmit pending queue, the size of the
1905 received queue (bytes waiting for applications to read) the state and the uid
1908 /proc/net/atalk_iface lists all the interfaces configured for appletalk.It
1909 shows the name of the interface, its Appletalk address, the network range on
1910 that address (or network number for phase 1 networks), and the status of the
1913 /proc/net/atalk_route lists each known network route. It lists the target
1914 (network) that the route leads to, the router (may be directly connected), the
1915 route flags, and the device the route is using.
1920 The IPX protocol has no tunable values in proc/sys/net.
1922 The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
1923 socket giving the local and remote addresses in Novell format (that is
1924 network:node:port). In accordance with the strange Novell tradition,
1925 everything but the port is in hex. Not_Connected is displayed for sockets that
1926 are not tied to a specific remote address. The Tx and Rx queue sizes indicate
1927 the number of bytes pending for transmission and reception. The state
1928 indicates the state the socket is in and the uid is the owning uid of the
1931 The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
1932 it gives the network number, the node number, and indicates if the network is
1933 the primary network. It also indicates which device it is bound to (or
1934 Internal for internal networks) and the Frame Type if appropriate. Linux
1935 supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
1938 The /proc/net/ipx_route table holds a list of IPX routes. For each route it
1939 gives the destination network, the router node (or Directly) and the network
1940 address of the router (or Connected) for internal networks.
1942 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
1943 ----------------------------------------------------------
1945 The "mqueue" filesystem provides the necessary kernel features to enable the
1946 creation of a user space library that implements the POSIX message queues
1947 API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System
1948 Interfaces specification.)
1950 The "mqueue" filesystem contains values for determining/setting the amount of
1951 resources used by the file system.
1953 /proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the
1954 maximum number of message queues allowed on the system.
1956 /proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the
1957 maximum number of messages in a queue value. In fact it is the limiting value
1958 for another (user) limit which is set in mq_open invocation. This attribute of
1959 a queue must be less or equal then msg_max.
1961 /proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the
1962 maximum message size value (it is every message queue's attribute set during
1966 ------------------------------------------------------------------------------
1968 ------------------------------------------------------------------------------
1969 Certain aspects of kernel behavior can be modified at runtime, without the
1970 need to recompile the kernel, or even to reboot the system. The files in the
1971 /proc/sys tree can not only be read, but also modified. You can use the echo
1972 command to write value into these files, thereby changing the default settings
1974 ------------------------------------------------------------------------------