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
44 2.14 /proc/<pid>/io - Display the IO accounting fields
46 ------------------------------------------------------------------------------
48 ------------------------------------------------------------------------------
50 0.1 Introduction/Credits
51 ------------------------
53 This documentation is part of a soon (or so we hope) to be released book on
54 the SuSE Linux distribution. As there is no complete documentation for the
55 /proc file system and we've used many freely available sources to write these
56 chapters, it seems only fair to give the work back to the Linux community.
57 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
58 afraid it's still far from complete, but we hope it will be useful. As far as
59 we know, it is the first 'all-in-one' document about the /proc file system. It
60 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
61 SPARC, AXP, etc., features, you probably won't find what you are looking for.
62 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
63 additions and patches are welcome and will be added to this document if you
66 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
67 other people for help compiling this documentation. We'd also like to extend a
68 special thank you to Andi Kleen for documentation, which we relied on heavily
69 to create this document, as well as the additional information he provided.
70 Thanks to everybody else who contributed source or docs to the Linux kernel
71 and helped create a great piece of software... :)
73 If you have any comments, corrections or additions, please don't hesitate to
74 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
77 The latest version of this document is available online at
78 http://skaro.nightcrawler.com/~bb/Docs/Proc as HTML version.
80 If the above direction does not works for you, ypu could try the kernel
81 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
82 comandante@zaralinux.com.
87 We don't guarantee the correctness of this document, and if you come to us
88 complaining about how you screwed up your system because of incorrect
89 documentation, we won't feel responsible...
91 ------------------------------------------------------------------------------
92 CHAPTER 1: COLLECTING SYSTEM INFORMATION
93 ------------------------------------------------------------------------------
95 ------------------------------------------------------------------------------
97 ------------------------------------------------------------------------------
98 * Investigating the properties of the pseudo file system /proc and its
99 ability to provide information on the running Linux system
100 * Examining /proc's structure
101 * Uncovering various information about the kernel and the processes running
103 ------------------------------------------------------------------------------
106 The proc file system acts as an interface to internal data structures in the
107 kernel. It can be used to obtain information about the system and to change
108 certain kernel parameters at runtime (sysctl).
110 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
111 show you how you can use /proc/sys to change settings.
113 1.1 Process-Specific Subdirectories
114 -----------------------------------
116 The directory /proc contains (among other things) one subdirectory for each
117 process running on the system, which is named after the process ID (PID).
119 The link self points to the process reading the file system. Each process
120 subdirectory has the entries listed in Table 1-1.
123 Table 1-1: Process specific entries in /proc
124 ..............................................................................
126 clear_refs Clears page referenced bits shown in smaps output
127 cmdline Command line arguments
128 cpu Current and last cpu in which it was executed (2.4)(smp)
129 cwd Link to the current working directory
130 environ Values of environment variables
131 exe Link to the executable of this process
132 fd Directory, which contains all file descriptors
133 maps Memory maps to executables and library files (2.4)
134 mem Memory held by this process
135 root Link to the root directory of this process
137 statm Process memory status information
138 status Process status in human readable form
139 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan
140 smaps Extension based on maps, the rss size for each mapped file
141 ..............................................................................
143 For example, to get the status information of a process, all you have to do is
144 read the file /proc/PID/status:
146 >cat /proc/self/status
162 SigPnd: 0000000000000000
163 SigBlk: 0000000000000000
164 SigIgn: 0000000000000000
165 SigCgt: 0000000000000000
166 CapInh: 00000000fffffeff
167 CapPrm: 0000000000000000
168 CapEff: 0000000000000000
171 This shows you nearly the same information you would get if you viewed it with
172 the ps command. In fact, ps uses the proc file system to obtain its
173 information. The statm file contains more detailed information about the
174 process memory usage. Its seven fields are explained in Table 1-2. The stat
175 file contains details information about the process itself. Its fields are
176 explained in Table 1-3.
179 Table 1-2: Contents of the statm files (as of 2.6.8-rc3)
180 ..............................................................................
182 size total program size (pages) (same as VmSize in status)
183 resident size of memory portions (pages) (same as VmRSS in status)
184 shared number of pages that are shared (i.e. backed by a file)
185 trs number of pages that are 'code' (not including libs; broken,
186 includes data segment)
187 lrs number of pages of library (always 0 on 2.6)
188 drs number of pages of data/stack (including libs; broken,
189 includes library text)
190 dt number of dirty pages (always 0 on 2.6)
191 ..............................................................................
194 Table 1-3: Contents of the stat files (as of 2.6.22-rc3)
195 ..............................................................................
198 tcomm filename of the executable
199 state state (R is running, S is sleeping, D is sleeping in an
200 uninterruptible wait, Z is zombie, T is traced or stopped)
201 ppid process id of the parent process
202 pgrp pgrp of the process
204 tty_nr tty the process uses
205 tty_pgrp pgrp of the tty
207 min_flt number of minor faults
208 cmin_flt number of minor faults with child's
209 maj_flt number of major faults
210 cmaj_flt number of major faults with child's
211 utime user mode jiffies
212 stime kernel mode jiffies
213 cutime user mode jiffies with child's
214 cstime kernel mode jiffies with child's
215 priority priority level
217 num_threads number of threads
218 start_time time the process started after system boot
219 vsize virtual memory size
220 rss resident set memory size
221 rsslim current limit in bytes on the rss
222 start_code address above which program text can run
223 end_code address below which program text can run
224 start_stack address of the start of the stack
225 esp current value of ESP
226 eip current value of EIP
227 pending bitmap of pending signals (obsolete)
228 blocked bitmap of blocked signals (obsolete)
229 sigign bitmap of ignored signals (obsolete)
230 sigcatch bitmap of catched signals (obsolete)
231 wchan address where process went to sleep
234 exit_signal signal to send to parent thread on exit
235 task_cpu which CPU the task is scheduled on
236 rt_priority realtime priority
237 policy scheduling policy (man sched_setscheduler)
238 blkio_ticks time spent waiting for block IO
239 ..............................................................................
245 Similar to the process entries, the kernel data files give information about
246 the running kernel. The files used to obtain this information are contained in
247 /proc and are listed in Table 1-4. Not all of these will be present in your
248 system. It depends on the kernel configuration and the loaded modules, which
249 files are there, and which are missing.
251 Table 1-4: Kernel info in /proc
252 ..............................................................................
254 apm Advanced power management info
255 buddyinfo Kernel memory allocator information (see text) (2.5)
256 bus Directory containing bus specific information
257 cmdline Kernel command line
258 cpuinfo Info about the CPU
259 devices Available devices (block and character)
260 dma Used DMS channels
261 filesystems Supported filesystems
262 driver Various drivers grouped here, currently rtc (2.4)
263 execdomains Execdomains, related to security (2.4)
264 fb Frame Buffer devices (2.4)
265 fs File system parameters, currently nfs/exports (2.4)
266 ide Directory containing info about the IDE subsystem
267 interrupts Interrupt usage
268 iomem Memory map (2.4)
269 ioports I/O port usage
270 irq Masks for irq to cpu affinity (2.4)(smp?)
271 isapnp ISA PnP (Plug&Play) Info (2.4)
272 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
274 ksyms Kernel symbol table
275 loadavg Load average of last 1, 5 & 15 minutes
279 modules List of loaded modules
280 mounts Mounted filesystems
281 net Networking info (see text)
282 partitions Table of partitions known to the system
283 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
284 decoupled by lspci (2.4)
286 scsi SCSI info (see text)
287 slabinfo Slab pool info
288 stat Overall statistics
289 swaps Swap space utilization
291 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
292 tty Info of tty drivers
294 version Kernel version
295 video bttv info of video resources (2.4)
296 ..............................................................................
298 You can, for example, check which interrupts are currently in use and what
299 they are used for by looking in the file /proc/interrupts:
301 > cat /proc/interrupts
303 0: 8728810 XT-PIC timer
304 1: 895 XT-PIC keyboard
306 3: 531695 XT-PIC aha152x
307 4: 2014133 XT-PIC serial
308 5: 44401 XT-PIC pcnet_cs
311 12: 182918 XT-PIC PS/2 Mouse
313 14: 1232265 XT-PIC ide0
317 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
318 output of a SMP machine):
320 > cat /proc/interrupts
323 0: 1243498 1214548 IO-APIC-edge timer
324 1: 8949 8958 IO-APIC-edge keyboard
325 2: 0 0 XT-PIC cascade
326 5: 11286 10161 IO-APIC-edge soundblaster
327 8: 1 0 IO-APIC-edge rtc
328 9: 27422 27407 IO-APIC-edge 3c503
329 12: 113645 113873 IO-APIC-edge PS/2 Mouse
331 14: 22491 24012 IO-APIC-edge ide0
332 15: 2183 2415 IO-APIC-edge ide1
333 17: 30564 30414 IO-APIC-level eth0
334 18: 177 164 IO-APIC-level bttv
339 NMI is incremented in this case because every timer interrupt generates a NMI
340 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
342 LOC is the local interrupt counter of the internal APIC of every CPU.
344 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
345 connects the CPUs in a SMP system. This means that an error has been detected,
346 the IO-APIC automatically retry the transmission, so it should not be a big
347 problem, but you should read the SMP-FAQ.
349 In this context it could be interesting to note the new irq directory in 2.4.
350 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
351 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
352 irq subdir is one subdir for each IRQ, and one file; prof_cpu_mask
356 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
357 1 11 13 15 17 19 3 5 7 9
361 The contents of the prof_cpu_mask file and each smp_affinity file for each IRQ
362 is the same by default:
364 > cat /proc/irq/0/smp_affinity
367 It's a bitmask, in which you can specify which CPUs can handle the IRQ, you can
370 > echo 1 > /proc/irq/prof_cpu_mask
372 This means that only the first CPU will handle the IRQ, but you can also echo 5
373 which means that only the first and fourth CPU can handle the IRQ.
375 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
376 between all the CPUs which are allowed to handle it. As usual the kernel has
377 more info than you and does a better job than you, so the defaults are the
378 best choice for almost everyone.
380 There are three more important subdirectories in /proc: net, scsi, and sys.
381 The general rule is that the contents, or even the existence of these
382 directories, depend on your kernel configuration. If SCSI is not enabled, the
383 directory scsi may not exist. The same is true with the net, which is there
384 only when networking support is present in the running kernel.
386 The slabinfo file gives information about memory usage at the slab level.
387 Linux uses slab pools for memory management above page level in version 2.2.
388 Commonly used objects have their own slab pool (such as network buffers,
389 directory cache, and so on).
391 ..............................................................................
393 > cat /proc/buddyinfo
395 Node 0, zone DMA 0 4 5 4 4 3 ...
396 Node 0, zone Normal 1 0 0 1 101 8 ...
397 Node 0, zone HighMem 2 0 0 1 1 0 ...
399 Memory fragmentation is a problem under some workloads, and buddyinfo is a
400 useful tool for helping diagnose these problems. Buddyinfo will give you a
401 clue as to how big an area you can safely allocate, or why a previous
404 Each column represents the number of pages of a certain order which are
405 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
406 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
407 available in ZONE_NORMAL, etc...
409 ..............................................................................
413 Provides information about distribution and utilization of memory. This
414 varies by architecture and compile options. The following is from a
415 16GB PIII, which has highmem enabled. You may not have all of these fields.
420 MemTotal: 16344972 kB
427 HighTotal: 15597528 kB
428 HighFree: 13629632 kB
437 CommitLimit: 7669796 kB
438 Committed_AS: 100056 kB
440 VmallocTotal: 112216 kB
442 VmallocChunk: 111088 kB
444 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
445 bits and the kernel binary code)
446 MemFree: The sum of LowFree+HighFree
447 Buffers: Relatively temporary storage for raw disk blocks
448 shouldn't get tremendously large (20MB or so)
449 Cached: in-memory cache for files read from the disk (the
450 pagecache). Doesn't include SwapCached
451 SwapCached: Memory that once was swapped out, is swapped back in but
452 still also is in the swapfile (if memory is needed it
453 doesn't need to be swapped out AGAIN because it is already
454 in the swapfile. This saves I/O)
455 Active: Memory that has been used more recently and usually not
456 reclaimed unless absolutely necessary.
457 Inactive: Memory which has been less recently used. It is more
458 eligible to be reclaimed for other purposes
460 HighFree: Highmem is all memory above ~860MB of physical memory
461 Highmem areas are for use by userspace programs, or
462 for the pagecache. The kernel must use tricks to access
463 this memory, making it slower to access than lowmem.
465 LowFree: Lowmem is memory which can be used for everything that
466 highmem can be used for, but it is also available for the
467 kernel's use for its own data structures. Among many
468 other things, it is where everything from the Slab is
469 allocated. Bad things happen when you're out of lowmem.
470 SwapTotal: total amount of swap space available
471 SwapFree: Memory which has been evicted from RAM, and is temporarily
473 Dirty: Memory which is waiting to get written back to the disk
474 Writeback: Memory which is actively being written back to the disk
475 Mapped: files which have been mmaped, such as libraries
476 Slab: in-kernel data structures cache
477 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
478 this is the total amount of memory currently available to
479 be allocated on the system. This limit is only adhered to
480 if strict overcommit accounting is enabled (mode 2 in
481 'vm.overcommit_memory').
482 The CommitLimit is calculated with the following formula:
483 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap
484 For example, on a system with 1G of physical RAM and 7G
485 of swap with a `vm.overcommit_ratio` of 30 it would
486 yield a CommitLimit of 7.3G.
487 For more details, see the memory overcommit documentation
488 in vm/overcommit-accounting.
489 Committed_AS: The amount of memory presently allocated on the system.
490 The committed memory is a sum of all of the memory which
491 has been allocated by processes, even if it has not been
492 "used" by them as of yet. A process which malloc()'s 1G
493 of memory, but only touches 300M of it will only show up
494 as using 300M of memory even if it has the address space
495 allocated for the entire 1G. This 1G is memory which has
496 been "committed" to by the VM and can be used at any time
497 by the allocating application. With strict overcommit
498 enabled on the system (mode 2 in 'vm.overcommit_memory'),
499 allocations which would exceed the CommitLimit (detailed
500 above) will not be permitted. This is useful if one needs
501 to guarantee that processes will not fail due to lack of
502 memory once that memory has been successfully allocated.
503 PageTables: amount of memory dedicated to the lowest level of page
505 VmallocTotal: total size of vmalloc memory area
506 VmallocUsed: amount of vmalloc area which is used
507 VmallocChunk: largest contigious block of vmalloc area which is free
510 1.3 IDE devices in /proc/ide
511 ----------------------------
513 The subdirectory /proc/ide contains information about all IDE devices of which
514 the kernel is aware. There is one subdirectory for each IDE controller, the
515 file drivers and a link for each IDE device, pointing to the device directory
516 in the controller specific subtree.
518 The file drivers contains general information about the drivers used for the
521 > cat /proc/ide/drivers
522 ide-cdrom version 4.53
523 ide-disk version 1.08
525 More detailed information can be found in the controller specific
526 subdirectories. These are named ide0, ide1 and so on. Each of these
527 directories contains the files shown in table 1-5.
530 Table 1-5: IDE controller info in /proc/ide/ide?
531 ..............................................................................
533 channel IDE channel (0 or 1)
534 config Configuration (only for PCI/IDE bridge)
536 model Type/Chipset of IDE controller
537 ..............................................................................
539 Each device connected to a controller has a separate subdirectory in the
540 controllers directory. The files listed in table 1-6 are contained in these
544 Table 1-6: IDE device information
545 ..............................................................................
548 capacity Capacity of the medium (in 512Byte blocks)
549 driver driver and version
550 geometry physical and logical geometry
551 identify device identify block
553 model device identifier
554 settings device setup
555 smart_thresholds IDE disk management thresholds
556 smart_values IDE disk management values
557 ..............................................................................
559 The most interesting file is settings. This file contains a nice overview of
560 the drive parameters:
562 # cat /proc/ide/ide0/hda/settings
563 name value min max mode
564 ---- ----- --- --- ----
565 bios_cyl 526 0 65535 rw
566 bios_head 255 0 255 rw
568 breada_readahead 4 0 127 rw
570 file_readahead 72 0 2097151 rw
572 keepsettings 0 0 1 rw
573 max_kb_per_request 122 1 127 rw
577 pio_mode write-only 0 255 w
583 1.4 Networking info in /proc/net
584 --------------------------------
586 The subdirectory /proc/net follows the usual pattern. Table 1-6 shows the
587 additional values you get for IP version 6 if you configure the kernel to
588 support this. Table 1-7 lists the files and their meaning.
591 Table 1-6: IPv6 info in /proc/net
592 ..............................................................................
594 udp6 UDP sockets (IPv6)
595 tcp6 TCP sockets (IPv6)
596 raw6 Raw device statistics (IPv6)
597 igmp6 IP multicast addresses, which this host joined (IPv6)
598 if_inet6 List of IPv6 interface addresses
599 ipv6_route Kernel routing table for IPv6
600 rt6_stats Global IPv6 routing tables statistics
601 sockstat6 Socket statistics (IPv6)
602 snmp6 Snmp data (IPv6)
603 ..............................................................................
606 Table 1-7: Network info in /proc/net
607 ..............................................................................
610 dev network devices with statistics
611 dev_mcast the Layer2 multicast groups a device is listening too
612 (interface index, label, number of references, number of bound
614 dev_stat network device status
615 ip_fwchains Firewall chain linkage
616 ip_fwnames Firewall chain names
617 ip_masq Directory containing the masquerading tables
618 ip_masquerade Major masquerading table
619 netstat Network statistics
620 raw raw device statistics
621 route Kernel routing table
622 rpc Directory containing rpc info
623 rt_cache Routing cache
625 sockstat Socket statistics
627 tr_rif Token ring RIF routing table
629 unix UNIX domain sockets
630 wireless Wireless interface data (Wavelan etc)
631 igmp IP multicast addresses, which this host joined
632 psched Global packet scheduler parameters.
633 netlink List of PF_NETLINK sockets
634 ip_mr_vifs List of multicast virtual interfaces
635 ip_mr_cache List of multicast routing cache
636 ..............................................................................
638 You can use this information to see which network devices are available in
639 your system and how much traffic was routed over those devices:
643 face |bytes packets errs drop fifo frame compressed multicast|[...
644 lo: 908188 5596 0 0 0 0 0 0 [...
645 ppp0:15475140 20721 410 0 0 410 0 0 [...
646 eth0: 614530 7085 0 0 0 0 0 1 [...
649 ...] bytes packets errs drop fifo colls carrier compressed
650 ...] 908188 5596 0 0 0 0 0 0
651 ...] 1375103 17405 0 0 0 0 0 0
652 ...] 1703981 5535 0 0 0 3 0 0
654 In addition, each Channel Bond interface has it's own directory. For
655 example, the bond0 device will have a directory called /proc/net/bond0/.
656 It will contain information that is specific to that bond, such as the
657 current slaves of the bond, the link status of the slaves, and how
658 many times the slaves link has failed.
663 If you have a SCSI host adapter in your system, you'll find a subdirectory
664 named after the driver for this adapter in /proc/scsi. You'll also see a list
665 of all recognized SCSI devices in /proc/scsi:
669 Host: scsi0 Channel: 00 Id: 00 Lun: 00
670 Vendor: IBM Model: DGHS09U Rev: 03E0
671 Type: Direct-Access ANSI SCSI revision: 03
672 Host: scsi0 Channel: 00 Id: 06 Lun: 00
673 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
674 Type: CD-ROM ANSI SCSI revision: 02
677 The directory named after the driver has one file for each adapter found in
678 the system. These files contain information about the controller, including
679 the used IRQ and the IO address range. The amount of information shown is
680 dependent on the adapter you use. The example shows the output for an Adaptec
681 AHA-2940 SCSI adapter:
683 > cat /proc/scsi/aic7xxx/0
685 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
687 TCQ Enabled By Default : Disabled
688 AIC7XXX_PROC_STATS : Disabled
689 AIC7XXX_RESET_DELAY : 5
690 Adapter Configuration:
691 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
692 Ultra Wide Controller
693 PCI MMAPed I/O Base: 0xeb001000
694 Adapter SEEPROM Config: SEEPROM found and used.
695 Adaptec SCSI BIOS: Enabled
697 SCBs: Active 0, Max Active 2,
698 Allocated 15, HW 16, Page 255
700 BIOS Control Word: 0x18b6
701 Adapter Control Word: 0x005b
702 Extended Translation: Enabled
703 Disconnect Enable Flags: 0xffff
704 Ultra Enable Flags: 0x0001
705 Tag Queue Enable Flags: 0x0000
706 Ordered Queue Tag Flags: 0x0000
707 Default Tag Queue Depth: 8
708 Tagged Queue By Device array for aic7xxx host instance 0:
709 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
710 Actual queue depth per device for aic7xxx host instance 0:
711 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
714 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
715 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
716 Total transfers 160151 (74577 reads and 85574 writes)
718 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
719 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
720 Total transfers 0 (0 reads and 0 writes)
723 1.6 Parallel port info in /proc/parport
724 ---------------------------------------
726 The directory /proc/parport contains information about the parallel ports of
727 your system. It has one subdirectory for each port, named after the port
730 These directories contain the four files shown in Table 1-8.
733 Table 1-8: Files in /proc/parport
734 ..............................................................................
736 autoprobe Any IEEE-1284 device ID information that has been acquired.
737 devices list of the device drivers using that port. A + will appear by the
738 name of the device currently using the port (it might not appear
740 hardware Parallel port's base address, IRQ line and DMA channel.
741 irq IRQ that parport is using for that port. This is in a separate
742 file to allow you to alter it by writing a new value in (IRQ
744 ..............................................................................
746 1.7 TTY info in /proc/tty
747 -------------------------
749 Information about the available and actually used tty's can be found in the
750 directory /proc/tty.You'll find entries for drivers and line disciplines in
751 this directory, as shown in Table 1-9.
754 Table 1-9: Files in /proc/tty
755 ..............................................................................
757 drivers list of drivers and their usage
758 ldiscs registered line disciplines
759 driver/serial usage statistic and status of single tty lines
760 ..............................................................................
762 To see which tty's are currently in use, you can simply look into the file
765 > cat /proc/tty/drivers
766 pty_slave /dev/pts 136 0-255 pty:slave
767 pty_master /dev/ptm 128 0-255 pty:master
768 pty_slave /dev/ttyp 3 0-255 pty:slave
769 pty_master /dev/pty 2 0-255 pty:master
770 serial /dev/cua 5 64-67 serial:callout
771 serial /dev/ttyS 4 64-67 serial
772 /dev/tty0 /dev/tty0 4 0 system:vtmaster
773 /dev/ptmx /dev/ptmx 5 2 system
774 /dev/console /dev/console 5 1 system:console
775 /dev/tty /dev/tty 5 0 system:/dev/tty
776 unknown /dev/tty 4 1-63 console
779 1.8 Miscellaneous kernel statistics in /proc/stat
780 -------------------------------------------------
782 Various pieces of information about kernel activity are available in the
783 /proc/stat file. All of the numbers reported in this file are aggregates
784 since the system first booted. For a quick look, simply cat the file:
787 cpu 2255 34 2290 22625563 6290 127 456
788 cpu0 1132 34 1441 11311718 3675 127 438
789 cpu1 1123 0 849 11313845 2614 0 18
790 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
797 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
798 lines. These numbers identify the amount of time the CPU has spent performing
799 different kinds of work. Time units are in USER_HZ (typically hundredths of a
800 second). The meanings of the columns are as follows, from left to right:
802 - user: normal processes executing in user mode
803 - nice: niced processes executing in user mode
804 - system: processes executing in kernel mode
805 - idle: twiddling thumbs
806 - iowait: waiting for I/O to complete
807 - irq: servicing interrupts
808 - softirq: servicing softirqs
810 The "intr" line gives counts of interrupts serviced since boot time, for each
811 of the possible system interrupts. The first column is the total of all
812 interrupts serviced; each subsequent column is the total for that particular
815 The "ctxt" line gives the total number of context switches across all CPUs.
817 The "btime" line gives the time at which the system booted, in seconds since
820 The "processes" line gives the number of processes and threads created, which
821 includes (but is not limited to) those created by calls to the fork() and
822 clone() system calls.
824 The "procs_running" line gives the number of processes currently running on
827 The "procs_blocked" line gives the number of processes currently blocked,
828 waiting for I/O to complete.
831 ------------------------------------------------------------------------------
833 ------------------------------------------------------------------------------
834 The /proc file system serves information about the running system. It not only
835 allows access to process data but also allows you to request the kernel status
836 by reading files in the hierarchy.
838 The directory structure of /proc reflects the types of information and makes
839 it easy, if not obvious, where to look for specific data.
840 ------------------------------------------------------------------------------
842 ------------------------------------------------------------------------------
843 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
844 ------------------------------------------------------------------------------
846 ------------------------------------------------------------------------------
848 ------------------------------------------------------------------------------
849 * Modifying kernel parameters by writing into files found in /proc/sys
850 * Exploring the files which modify certain parameters
851 * Review of the /proc/sys file tree
852 ------------------------------------------------------------------------------
855 A very interesting part of /proc is the directory /proc/sys. This is not only
856 a source of information, it also allows you to change parameters within the
857 kernel. Be very careful when attempting this. You can optimize your system,
858 but you can also cause it to crash. Never alter kernel parameters on a
859 production system. Set up a development machine and test to make sure that
860 everything works the way you want it to. You may have no alternative but to
861 reboot the machine once an error has been made.
863 To change a value, simply echo the new value into the file. An example is
864 given below in the section on the file system data. You need to be root to do
865 this. You can create your own boot script to perform this every time your
868 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
869 general things in the operation of the Linux kernel. Since some of the files
870 can inadvertently disrupt your system, it is advisable to read both
871 documentation and source before actually making adjustments. In any case, be
872 very careful when writing to any of these files. The entries in /proc may
873 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
874 review the kernel documentation in the directory /usr/src/linux/Documentation.
875 This chapter is heavily based on the documentation included in the pre 2.2
876 kernels, and became part of it in version 2.2.1 of the Linux kernel.
878 2.1 /proc/sys/fs - File system data
879 -----------------------------------
881 This subdirectory contains specific file system, file handle, inode, dentry
882 and quota information.
884 Currently, these files are in /proc/sys/fs:
889 Status of the directory cache. Since directory entries are dynamically
890 allocated and deallocated, this file indicates the current status. It holds
891 six values, in which the last two are not used and are always zero. The others
892 are listed in table 2-1.
895 Table 2-1: Status files of the directory cache
896 ..............................................................................
898 nr_dentry Almost always zero
899 nr_unused Number of unused cache entries
901 in seconds after the entry may be reclaimed, when memory is short
902 want_pages internally
903 ..............................................................................
905 dquot-nr and dquot-max
906 ----------------------
908 The file dquot-max shows the maximum number of cached disk quota entries.
910 The file dquot-nr shows the number of allocated disk quota entries and the
911 number of free disk quota entries.
913 If the number of available cached disk quotas is very low and you have a large
914 number of simultaneous system users, you might want to raise the limit.
919 The kernel allocates file handles dynamically, but doesn't free them again at
922 The value in file-max denotes the maximum number of file handles that the
923 Linux kernel will allocate. When you get a lot of error messages about running
924 out of file handles, you might want to raise this limit. The default value is
925 10% of RAM in kilobytes. To change it, just write the new number into the
928 # cat /proc/sys/fs/file-max
930 # echo 8192 > /proc/sys/fs/file-max
931 # cat /proc/sys/fs/file-max
935 This method of revision is useful for all customizable parameters of the
936 kernel - simply echo the new value to the corresponding file.
938 Historically, the three values in file-nr denoted the number of allocated file
939 handles, the number of allocated but unused file handles, and the maximum
940 number of file handles. Linux 2.6 always reports 0 as the number of free file
941 handles -- this is not an error, it just means that the number of allocated
942 file handles exactly matches the number of used file handles.
944 Attempts to allocate more file descriptors than file-max are reported with
945 printk, look for "VFS: file-max limit <number> reached".
947 inode-state and inode-nr
948 ------------------------
950 The file inode-nr contains the first two items from inode-state, so we'll skip
953 inode-state contains two actual numbers and five dummy values. The numbers
954 are nr_inodes and nr_free_inodes (in order of appearance).
959 Denotes the number of inodes the system has allocated. This number will
960 grow and shrink dynamically.
965 Represents the number of free inodes. Ie. The number of inuse inodes is
966 (nr_inodes - nr_free_inodes).
968 aio-nr and aio-max-nr
969 ---------------------
971 aio-nr is the running total of the number of events specified on the
972 io_setup system call for all currently active aio contexts. If aio-nr
973 reaches aio-max-nr then io_setup will fail with EAGAIN. Note that
974 raising aio-max-nr does not result in the pre-allocation or re-sizing
975 of any kernel data structures.
977 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
978 -----------------------------------------------------------
980 Besides these files, there is the subdirectory /proc/sys/fs/binfmt_misc. This
981 handles the kernel support for miscellaneous binary formats.
983 Binfmt_misc provides the ability to register additional binary formats to the
984 Kernel without compiling an additional module/kernel. Therefore, binfmt_misc
985 needs to know magic numbers at the beginning or the filename extension of the
988 It works by maintaining a linked list of structs that contain a description of
989 a binary format, including a magic with size (or the filename extension),
990 offset and mask, and the interpreter name. On request it invokes the given
991 interpreter with the original program as argument, as binfmt_java and
992 binfmt_em86 and binfmt_mz do. Since binfmt_misc does not define any default
993 binary-formats, you have to register an additional binary-format.
995 There are two general files in binfmt_misc and one file per registered format.
996 The two general files are register and status.
998 Registering a new binary format
999 -------------------------------
1001 To register a new binary format you have to issue the command
1003 echo :name:type:offset:magic:mask:interpreter: > /proc/sys/fs/binfmt_misc/register
1007 with appropriate name (the name for the /proc-dir entry), offset (defaults to
1008 0, if omitted), magic, mask (which can be omitted, defaults to all 0xff) and
1009 last but not least, the interpreter that is to be invoked (for example and
1010 testing /bin/echo). Type can be M for usual magic matching or E for filename
1011 extension matching (give extension in place of magic).
1013 Check or reset the status of the binary format handler
1014 ------------------------------------------------------
1016 If you do a cat on the file /proc/sys/fs/binfmt_misc/status, you will get the
1017 current status (enabled/disabled) of binfmt_misc. Change the status by echoing
1018 0 (disables) or 1 (enables) or -1 (caution: this clears all previously
1019 registered binary formats) to status. For example echo 0 > status to disable
1020 binfmt_misc (temporarily).
1022 Status of a single handler
1023 --------------------------
1025 Each registered handler has an entry in /proc/sys/fs/binfmt_misc. These files
1026 perform the same function as status, but their scope is limited to the actual
1027 binary format. By cating this file, you also receive all related information
1028 about the interpreter/magic of the binfmt.
1030 Example usage of binfmt_misc (emulate binfmt_java)
1031 --------------------------------------------------
1033 cd /proc/sys/fs/binfmt_misc
1034 echo ':Java:M::\xca\xfe\xba\xbe::/usr/local/java/bin/javawrapper:' > register
1035 echo ':HTML:E::html::/usr/local/java/bin/appletviewer:' > register
1036 echo ':Applet:M::<!--applet::/usr/local/java/bin/appletviewer:' > register
1037 echo ':DEXE:M::\x0eDEX::/usr/bin/dosexec:' > register
1040 These four lines add support for Java executables and Java applets (like
1041 binfmt_java, additionally recognizing the .html extension with no need to put
1042 <!--applet> to every applet file). You have to install the JDK and the
1043 shell-script /usr/local/java/bin/javawrapper too. It works around the
1044 brokenness of the Java filename handling. To add a Java binary, just create a
1045 link to the class-file somewhere in the path.
1047 2.3 /proc/sys/kernel - general kernel parameters
1048 ------------------------------------------------
1050 This directory reflects general kernel behaviors. As I've said before, the
1051 contents depend on your configuration. Here you'll find the most important
1052 files, along with descriptions of what they mean and how to use them.
1057 The file contains three values; highwater, lowwater, and frequency.
1059 It exists only when BSD-style process accounting is enabled. These values
1060 control its behavior. If the free space on the file system where the log lives
1061 goes below lowwater percentage, accounting suspends. If it goes above
1062 highwater percentage, accounting resumes. Frequency determines how often you
1063 check the amount of free space (value is in seconds). Default settings are: 4,
1064 2, and 30. That is, suspend accounting if there is less than 2 percent free;
1065 resume it if we have a value of 3 or more percent; consider information about
1066 the amount of free space valid for 30 seconds
1071 When the value in this file is 0, ctrl-alt-del is trapped and sent to the init
1072 program to handle a graceful restart. However, when the value is greater that
1073 zero, Linux's reaction to this key combination will be an immediate reboot,
1074 without syncing its dirty buffers.
1077 When a program (like dosemu) has the keyboard in raw mode, the
1078 ctrl-alt-del is intercepted by the program before it ever reaches the
1079 kernel tty layer, and it is up to the program to decide what to do with
1082 domainname and hostname
1083 -----------------------
1085 These files can be controlled to set the NIS domainname and hostname of your
1086 box. For the classic darkstar.frop.org a simple:
1088 # echo "darkstar" > /proc/sys/kernel/hostname
1089 # echo "frop.org" > /proc/sys/kernel/domainname
1092 would suffice to set your hostname and NIS domainname.
1094 osrelease, ostype and version
1095 -----------------------------
1097 The names make it pretty obvious what these fields contain:
1099 > cat /proc/sys/kernel/osrelease
1102 > cat /proc/sys/kernel/ostype
1105 > cat /proc/sys/kernel/version
1106 #4 Fri Oct 1 12:41:14 PDT 1999
1109 The files osrelease and ostype should be clear enough. Version needs a little
1110 more clarification. The #4 means that this is the 4th kernel built from this
1111 source base and the date after it indicates the time the kernel was built. The
1112 only way to tune these values is to rebuild the kernel.
1117 The value in this file represents the number of seconds the kernel waits
1118 before rebooting on a panic. When you use the software watchdog, the
1119 recommended setting is 60. If set to 0, the auto reboot after a kernel panic
1120 is disabled, which is the default setting.
1125 The four values in printk denote
1127 * default_message_loglevel,
1128 * minimum_console_loglevel and
1129 * default_console_loglevel
1132 These values influence printk() behavior when printing or logging error
1133 messages, which come from inside the kernel. See syslog(2) for more
1134 information on the different log levels.
1139 Messages with a higher priority than this will be printed to the console.
1141 default_message_level
1142 ---------------------
1144 Messages without an explicit priority will be printed with this priority.
1146 minimum_console_loglevel
1147 ------------------------
1149 Minimum (highest) value to which the console_loglevel can be set.
1151 default_console_loglevel
1152 ------------------------
1154 Default value for console_loglevel.
1159 This file shows the size of the generic SCSI (sg) buffer. At this point, you
1160 can't tune it yet, but you can change it at compile time by editing
1161 include/scsi/sg.h and changing the value of SG_BIG_BUFF.
1163 If you use a scanner with SANE (Scanner Access Now Easy) you might want to set
1164 this to a higher value. Refer to the SANE documentation on this issue.
1169 The location where the modprobe binary is located. The kernel uses this
1170 program to load modules on demand.
1175 The value in this file affects behavior of handling NMI. When the value is
1176 non-zero, unknown NMI is trapped and then panic occurs. At that time, kernel
1177 debugging information is displayed on console.
1179 NMI switch that most IA32 servers have fires unknown NMI up, for example.
1180 If a system hangs up, try pressing the NMI switch.
1185 Enables/Disables the NMI watchdog on x86 systems. When the value is non-zero
1186 the NMI watchdog is enabled and will continuously test all online cpus to
1187 determine whether or not they are still functioning properly.
1189 Because the NMI watchdog shares registers with oprofile, by disabling the NMI
1190 watchdog, oprofile may have more registers to utilize.
1195 Enables/Disables the protection of the per-process proc entries "maps" and
1196 "smaps". When enabled, the contents of these files are visible only to
1197 readers that are allowed to ptrace() the given process.
1200 2.4 /proc/sys/vm - The virtual memory subsystem
1201 -----------------------------------------------
1203 The files in this directory can be used to tune the operation of the virtual
1204 memory (VM) subsystem of the Linux kernel.
1209 Controls the tendency of the kernel to reclaim the memory which is used for
1210 caching of directory and inode objects.
1212 At the default value of vfs_cache_pressure=100 the kernel will attempt to
1213 reclaim dentries and inodes at a "fair" rate with respect to pagecache and
1214 swapcache reclaim. Decreasing vfs_cache_pressure causes the kernel to prefer
1215 to retain dentry and inode caches. Increasing vfs_cache_pressure beyond 100
1216 causes the kernel to prefer to reclaim dentries and inodes.
1218 dirty_background_ratio
1219 ----------------------
1221 Contains, as a percentage of total system memory, the number of pages at which
1222 the pdflush background writeback daemon will start writing out dirty data.
1227 Contains, as a percentage of total system memory, the number of pages at which
1228 a process which is generating disk writes will itself start writing out dirty
1231 dirty_writeback_centisecs
1232 -------------------------
1234 The pdflush writeback daemons will periodically wake up and write `old' data
1235 out to disk. This tunable expresses the interval between those wakeups, in
1236 100'ths of a second.
1238 Setting this to zero disables periodic writeback altogether.
1240 dirty_expire_centisecs
1241 ----------------------
1243 This tunable is used to define when dirty data is old enough to be eligible
1244 for writeout by the pdflush daemons. It is expressed in 100'ths of a second.
1245 Data which has been dirty in-memory for longer than this interval will be
1246 written out next time a pdflush daemon wakes up.
1251 If non-zero, this sysctl disables the new 32-bit mmap mmap layout - the kernel
1252 will use the legacy (2.4) layout for all processes.
1254 lower_zone_protection
1255 ---------------------
1257 For some specialised workloads on highmem machines it is dangerous for
1258 the kernel to allow process memory to be allocated from the "lowmem"
1259 zone. This is because that memory could then be pinned via the mlock()
1260 system call, or by unavailability of swapspace.
1262 And on large highmem machines this lack of reclaimable lowmem memory
1265 So the Linux page allocator has a mechanism which prevents allocations
1266 which _could_ use highmem from using too much lowmem. This means that
1267 a certain amount of lowmem is defended from the possibility of being
1268 captured into pinned user memory.
1270 (The same argument applies to the old 16 megabyte ISA DMA region. This
1271 mechanism will also defend that region from allocations which could use
1274 The `lower_zone_protection' tunable determines how aggressive the kernel is
1275 in defending these lower zones. The default value is zero - no
1278 If you have a machine which uses highmem or ISA DMA and your
1279 applications are using mlock(), or if you are running with no swap then
1280 you probably should increase the lower_zone_protection setting.
1282 The units of this tunable are fairly vague. It is approximately equal
1283 to "megabytes," so setting lower_zone_protection=100 will protect around 100
1284 megabytes of the lowmem zone from user allocations. It will also make
1285 those 100 megabytes unavailable for use by applications and by
1286 pagecache, so there is a cost.
1288 The effects of this tunable may be observed by monitoring
1289 /proc/meminfo:LowFree. Write a single huge file and observe the point
1290 at which LowFree ceases to fall.
1292 A reasonable value for lower_zone_protection is 100.
1297 page-cluster controls the number of pages which are written to swap in
1298 a single attempt. The swap I/O size.
1300 It is a logarithmic value - setting it to zero means "1 page", setting
1301 it to 1 means "2 pages", setting it to 2 means "4 pages", etc.
1303 The default value is three (eight pages at a time). There may be some
1304 small benefits in tuning this to a different value if your workload is
1310 Controls overcommit of system memory, possibly allowing processes
1311 to allocate (but not use) more memory than is actually available.
1314 0 - Heuristic overcommit handling. Obvious overcommits of
1315 address space are refused. Used for a typical system. It
1316 ensures a seriously wild allocation fails while allowing
1317 overcommit to reduce swap usage. root is allowed to
1318 allocate slightly more memory in this mode. This is the
1321 1 - Always overcommit. Appropriate for some scientific
1324 2 - Don't overcommit. The total address space commit
1325 for the system is not permitted to exceed swap plus a
1326 configurable percentage (default is 50) of physical RAM.
1327 Depending on the percentage you use, in most situations
1328 this means a process will not be killed while attempting
1329 to use already-allocated memory but will receive errors
1330 on memory allocation as appropriate.
1335 Percentage of physical memory size to include in overcommit calculations
1338 Memory allocation limit = swapspace + physmem * (overcommit_ratio / 100)
1340 swapspace = total size of all swap areas
1341 physmem = size of physical memory in system
1343 nr_hugepages and hugetlb_shm_group
1344 ----------------------------------
1346 nr_hugepages configures number of hugetlb page reserved for the system.
1348 hugetlb_shm_group contains group id that is allowed to create SysV shared
1349 memory segment using hugetlb page.
1351 hugepages_treat_as_movable
1352 --------------------------
1354 This parameter is only useful when kernelcore= is specified at boot time to
1355 create ZONE_MOVABLE for pages that may be reclaimed or migrated. Huge pages
1356 are not movable so are not normally allocated from ZONE_MOVABLE. A non-zero
1357 value written to hugepages_treat_as_movable allows huge pages to be allocated
1360 Once enabled, the ZONE_MOVABLE is treated as an area of memory the huge
1361 pages pool can easily grow or shrink within. Assuming that applications are
1362 not running that mlock() a lot of memory, it is likely the huge pages pool
1363 can grow to the size of ZONE_MOVABLE by repeatedly entering the desired value
1364 into nr_hugepages and triggering page reclaim.
1369 laptop_mode is a knob that controls "laptop mode". All the things that are
1370 controlled by this knob are discussed in Documentation/laptop-mode.txt.
1375 block_dump enables block I/O debugging when set to a nonzero value. More
1376 information on block I/O debugging is in Documentation/laptop-mode.txt.
1381 This file contains valid hold time of swap out protection token. The Linux
1382 VM has token based thrashing control mechanism and uses the token to prevent
1383 unnecessary page faults in thrashing situation. The unit of the value is
1384 second. The value would be useful to tune thrashing behavior.
1389 Writing to this will cause the kernel to drop clean caches, dentries and
1390 inodes from memory, causing that memory to become free.
1393 echo 1 > /proc/sys/vm/drop_caches
1394 To free dentries and inodes:
1395 echo 2 > /proc/sys/vm/drop_caches
1396 To free pagecache, dentries and inodes:
1397 echo 3 > /proc/sys/vm/drop_caches
1399 As this is a non-destructive operation and dirty objects are not freeable, the
1400 user should run `sync' first.
1403 2.5 /proc/sys/dev - Device specific parameters
1404 ----------------------------------------------
1406 Currently there is only support for CDROM drives, and for those, there is only
1407 one read-only file containing information about the CD-ROM drives attached to
1410 >cat /proc/sys/dev/cdrom/info
1411 CD-ROM information, Id: cdrom.c 2.55 1999/04/25
1415 drive # of slots: 1 0
1419 Can change speed: 1 1
1420 Can select disk: 0 1
1421 Can read multisession: 1 1
1423 Reports media changed: 1 1
1427 You see two drives, sr0 and hdb, along with a list of their features.
1429 2.6 /proc/sys/sunrpc - Remote procedure calls
1430 ---------------------------------------------
1432 This directory contains four files, which enable or disable debugging for the
1433 RPC functions NFS, NFS-daemon, RPC and NLM. The default values are 0. They can
1434 be set to one to turn debugging on. (The default value is 0 for each)
1436 2.7 /proc/sys/net - Networking stuff
1437 ------------------------------------
1439 The interface to the networking parts of the kernel is located in
1440 /proc/sys/net. Table 2-3 shows all possible subdirectories. You may see only
1441 some of them, depending on your kernel's configuration.
1444 Table 2-3: Subdirectories in /proc/sys/net
1445 ..............................................................................
1446 Directory Content Directory Content
1447 core General parameter appletalk Appletalk protocol
1448 unix Unix domain sockets netrom NET/ROM
1449 802 E802 protocol ax25 AX25
1450 ethernet Ethernet protocol rose X.25 PLP layer
1451 ipv4 IP version 4 x25 X.25 protocol
1452 ipx IPX token-ring IBM token ring
1453 bridge Bridging decnet DEC net
1455 ..............................................................................
1457 We will concentrate on IP networking here. Since AX15, X.25, and DEC Net are
1458 only minor players in the Linux world, we'll skip them in this chapter. You'll
1459 find some short info on Appletalk and IPX further on in this chapter. Review
1460 the online documentation and the kernel source to get a detailed view of the
1461 parameters for those protocols. In this section we'll discuss the
1462 subdirectories printed in bold letters in the table above. As default values
1463 are suitable for most needs, there is no need to change these values.
1465 /proc/sys/net/core - Network core options
1466 -----------------------------------------
1471 The default setting of the socket receive buffer in bytes.
1476 The maximum receive socket buffer size in bytes.
1481 The default setting (in bytes) of the socket send buffer.
1486 The maximum send socket buffer size in bytes.
1488 message_burst and message_cost
1489 ------------------------------
1491 These parameters are used to limit the warning messages written to the kernel
1492 log from the networking code. They enforce a rate limit to make a
1493 denial-of-service attack impossible. A higher message_cost factor, results in
1494 fewer messages that will be written. Message_burst controls when messages will
1495 be dropped. The default settings limit warning messages to one every five
1501 This controls console messages from the networking stack that can occur because
1502 of problems on the network like duplicate address or bad checksums. Normally,
1503 this should be enabled, but if the problem persists the messages can be
1510 Maximum number of packets, queued on the INPUT side, when the interface
1511 receives packets faster than kernel can process them.
1516 Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
1517 of struct cmsghdr structures with appended data.
1519 /proc/sys/net/unix - Parameters for Unix domain sockets
1520 -------------------------------------------------------
1522 There are only two files in this subdirectory. They control the delays for
1523 deleting and destroying socket descriptors.
1525 2.8 /proc/sys/net/ipv4 - IPV4 settings
1526 --------------------------------------
1528 IP version 4 is still the most used protocol in Unix networking. It will be
1529 replaced by IP version 6 in the next couple of years, but for the moment it's
1530 the de facto standard for the internet and is used in most networking
1531 environments around the world. Because of the importance of this protocol,
1532 we'll have a deeper look into the subtree controlling the behavior of the IPv4
1533 subsystem of the Linux kernel.
1535 Let's start with the entries in /proc/sys/net/ipv4.
1540 icmp_echo_ignore_all and icmp_echo_ignore_broadcasts
1541 ----------------------------------------------------
1543 Turn on (1) or off (0), if the kernel should ignore all ICMP ECHO requests, or
1544 just those to broadcast and multicast addresses.
1546 Please note that if you accept ICMP echo requests with a broadcast/multi\-cast
1547 destination address your network may be used as an exploder for denial of
1548 service packet flooding attacks to other hosts.
1550 icmp_destunreach_rate, icmp_echoreply_rate, icmp_paramprob_rate and icmp_timeexeed_rate
1551 ---------------------------------------------------------------------------------------
1553 Sets limits for sending ICMP packets to specific targets. A value of zero
1554 disables all limiting. Any positive value sets the maximum package rate in
1555 hundredth of a second (on Intel systems).
1563 This file contains the number one if the host received its IP configuration by
1564 RARP, BOOTP, DHCP or a similar mechanism. Otherwise it is zero.
1569 TTL (Time To Live) for IPv4 interfaces. This is simply the maximum number of
1570 hops a packet may travel.
1575 Enable dynamic socket address rewriting on interface address change. This is
1576 useful for dialup interface with changing IP addresses.
1581 Enable or disable forwarding of IP packages between interfaces. Changing this
1582 value resets all other parameters to their default values. They differ if the
1583 kernel is configured as host or router.
1588 Range of ports used by TCP and UDP to choose the local port. Contains two
1589 numbers, the first number is the lowest port, the second number the highest
1590 local port. Default is 1024-4999. Should be changed to 32768-61000 for
1596 Global switch to turn path MTU discovery off. It can also be set on a per
1597 socket basis by the applications or on a per route basis.
1602 Enable/disable debugging of IP masquerading.
1604 IP fragmentation settings
1605 -------------------------
1607 ipfrag_high_trash and ipfrag_low_trash
1608 --------------------------------------
1610 Maximum memory used to reassemble IP fragments. When ipfrag_high_thresh bytes
1611 of memory is allocated for this purpose, the fragment handler will toss
1612 packets until ipfrag_low_thresh is reached.
1617 Time in seconds to keep an IP fragment in memory.
1625 This file controls the use of the ECN bit in the IPv4 headers. This is a new
1626 feature about Explicit Congestion Notification, but some routers and firewalls
1627 block traffic that has this bit set, so it could be necessary to echo 0 to
1628 /proc/sys/net/ipv4/tcp_ecn if you want to talk to these sites. For more info
1629 you could read RFC2481.
1631 tcp_retrans_collapse
1632 --------------------
1634 Bug-to-bug compatibility with some broken printers. On retransmit, try to send
1635 larger packets to work around bugs in certain TCP stacks. Can be turned off by
1638 tcp_keepalive_probes
1639 --------------------
1641 Number of keep alive probes TCP sends out, until it decides that the
1642 connection is broken.
1647 How often TCP sends out keep alive messages, when keep alive is enabled. The
1653 Number of times initial SYNs for a TCP connection attempt will be
1654 retransmitted. Should not be higher than 255. This is only the timeout for
1655 outgoing connections, for incoming connections the number of retransmits is
1656 defined by tcp_retries1.
1661 Enable select acknowledgments after RFC2018.
1666 Enable timestamps as defined in RFC1323.
1671 Enable the strict RFC793 interpretation of the TCP urgent pointer field. The
1672 default is to use the BSD compatible interpretation of the urgent pointer
1673 pointing to the first byte after the urgent data. The RFC793 interpretation is
1674 to have it point to the last byte of urgent data. Enabling this option may
1675 lead to interoperability problems. Disabled by default.
1680 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES. Send out
1681 syncookies when the syn backlog queue of a socket overflows. This is to ward
1682 off the common 'syn flood attack'. Disabled by default.
1684 Note that the concept of a socket backlog is abandoned. This means the peer
1685 may not receive reliable error messages from an over loaded server with
1691 Enable window scaling as defined in RFC1323.
1696 The length of time in seconds it takes to receive a final FIN before the
1697 socket is always closed. This is strictly a violation of the TCP
1698 specification, but required to prevent denial-of-service attacks.
1703 Indicates how many keep alive probes are sent per slow timer run. Should not
1704 be set too high to prevent bursts.
1709 Length of the per socket backlog queue. Since Linux 2.2 the backlog specified
1710 in listen(2) only specifies the length of the backlog queue of already
1711 established sockets. When more connection requests arrive Linux starts to drop
1712 packets. When syncookies are enabled the packets are still answered and the
1713 maximum queue is effectively ignored.
1718 Defines how often an answer to a TCP connection request is retransmitted
1724 Defines how often a TCP packet is retransmitted before giving up.
1726 Interface specific settings
1727 ---------------------------
1729 In the directory /proc/sys/net/ipv4/conf you'll find one subdirectory for each
1730 interface the system knows about and one directory calls all. Changes in the
1731 all subdirectory affect all interfaces, whereas changes in the other
1732 subdirectories affect only one interface. All directories have the same
1738 This switch decides if the kernel accepts ICMP redirect messages or not. The
1739 default is 'yes' if the kernel is configured for a regular host and 'no' for a
1740 router configuration.
1745 Should source routed packages be accepted or declined. The default is
1746 dependent on the kernel configuration. It's 'yes' for routers and 'no' for
1752 Accept packets with source address 0.b.c.d with destinations not to this host
1753 as local ones. It is supposed that a BOOTP relay daemon will catch and forward
1756 The default is 0, since this feature is not implemented yet (kernel version
1762 Enable or disable IP forwarding on this interface.
1767 Log packets with source addresses with no known route to kernel log.
1772 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE and a
1773 multicast routing daemon is required.
1778 Does (1) or does not (0) perform proxy ARP.
1783 Integer value determines if a source validation should be made. 1 means yes, 0
1784 means no. Disabled by default, but local/broadcast address spoofing is always
1787 If you set this to 1 on a router that is the only connection for a network to
1788 the net, it will prevent spoofing attacks against your internal networks
1789 (external addresses can still be spoofed), without the need for additional
1795 Accept ICMP redirect messages only for gateways, listed in default gateway
1796 list. Enabled by default.
1801 If it is not set the kernel does not assume that different subnets on this
1802 device can communicate directly. Default setting is 'yes'.
1807 Determines whether to send ICMP redirects to other hosts.
1812 The directory /proc/sys/net/ipv4/route contains several file to control
1815 error_burst and error_cost
1816 --------------------------
1818 These parameters are used to limit how many ICMP destination unreachable to
1819 send from the host in question. ICMP destination unreachable messages are
1820 sent when we cannot reach the next hop while trying to transmit a packet.
1821 It will also print some error messages to kernel logs if someone is ignoring
1822 our ICMP redirects. The higher the error_cost factor is, the fewer
1823 destination unreachable and error messages will be let through. Error_burst
1824 controls when destination unreachable messages and error messages will be
1825 dropped. The default settings limit warning messages to five every second.
1830 Writing to this file results in a flush of the routing cache.
1832 gc_elasticity, gc_interval, gc_min_interval_ms, gc_timeout, gc_thresh
1833 ---------------------------------------------------------------------
1835 Values to control the frequency and behavior of the garbage collection
1836 algorithm for the routing cache. gc_min_interval is deprecated and replaced
1837 by gc_min_interval_ms.
1843 Maximum size of the routing cache. Old entries will be purged once the cache
1844 reached has this size.
1846 max_delay, min_delay
1847 --------------------
1849 Delays for flushing the routing cache.
1851 redirect_load, redirect_number
1852 ------------------------------
1854 Factors which determine if more ICPM redirects should be sent to a specific
1855 host. No redirects will be sent once the load limit or the maximum number of
1856 redirects has been reached.
1861 Timeout for redirects. After this period redirects will be sent again, even if
1862 this has been stopped, because the load or number limit has been reached.
1864 Network Neighbor handling
1865 -------------------------
1867 Settings about how to handle connections with direct neighbors (nodes attached
1868 to the same link) can be found in the directory /proc/sys/net/ipv4/neigh.
1870 As we saw it in the conf directory, there is a default subdirectory which
1871 holds the default values, and one directory for each interface. The contents
1872 of the directories are identical, with the single exception that the default
1873 settings contain additional options to set garbage collection parameters.
1875 In the interface directories you'll find the following entries:
1877 base_reachable_time, base_reachable_time_ms
1878 -------------------------------------------
1880 A base value used for computing the random reachable time value as specified
1883 Expression of base_reachable_time, which is deprecated, is in seconds.
1884 Expression of base_reachable_time_ms is in milliseconds.
1886 retrans_time, retrans_time_ms
1887 -----------------------------
1889 The time between retransmitted Neighbor Solicitation messages.
1890 Used for address resolution and to determine if a neighbor is
1893 Expression of retrans_time, which is deprecated, is in 1/100 seconds (for
1894 IPv4) or in jiffies (for IPv6).
1895 Expression of retrans_time_ms is in milliseconds.
1900 Maximum queue length for a pending arp request - the number of packets which
1901 are accepted from other layers while the ARP address is still resolved.
1906 Maximum for random delay of answers to neighbor solicitation messages in
1907 jiffies (1/100 sec). Not yet implemented (Linux does not have anycast support
1913 Maximum number of retries for unicast solicitation.
1918 Maximum number of retries for multicast solicitation.
1920 delay_first_probe_time
1921 ----------------------
1923 Delay for the first time probe if the neighbor is reachable. (see
1929 An ARP/neighbor entry is only replaced with a new one if the old is at least
1930 locktime old. This prevents ARP cache thrashing.
1935 Maximum time (real time is random [0..proxytime]) before answering to an ARP
1936 request for which we have an proxy ARP entry. In some cases, this is used to
1937 prevent network flooding.
1942 Maximum queue length of the delayed proxy arp timer. (see proxy_delay).
1947 Determines the number of requests to send to the user level ARP daemon. Use 0
1953 Determines how often to check for stale ARP entries. After an ARP entry is
1954 stale it will be resolved again (which is useful when an IP address migrates
1955 to another machine). When ucast_solicit is greater than 0 it first tries to
1956 send an ARP packet directly to the known host When that fails and
1957 mcast_solicit is greater than 0, an ARP request is broadcasted.
1962 The /proc/sys/net/appletalk directory holds the Appletalk configuration data
1963 when Appletalk is loaded. The configurable parameters are:
1968 The amount of time we keep an ARP entry before expiring it. Used to age out
1974 The amount of time we will spend trying to resolve an Appletalk address.
1976 aarp-retransmit-limit
1977 ---------------------
1979 The number of times we will retransmit a query before giving up.
1984 Controls the rate at which expires are checked.
1986 The directory /proc/net/appletalk holds the list of active Appletalk sockets
1989 The fields indicate the DDP type, the local address (in network:node format)
1990 the remote address, the size of the transmit pending queue, the size of the
1991 received queue (bytes waiting for applications to read) the state and the uid
1994 /proc/net/atalk_iface lists all the interfaces configured for appletalk.It
1995 shows the name of the interface, its Appletalk address, the network range on
1996 that address (or network number for phase 1 networks), and the status of the
1999 /proc/net/atalk_route lists each known network route. It lists the target
2000 (network) that the route leads to, the router (may be directly connected), the
2001 route flags, and the device the route is using.
2006 The IPX protocol has no tunable values in proc/sys/net.
2008 The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
2009 socket giving the local and remote addresses in Novell format (that is
2010 network:node:port). In accordance with the strange Novell tradition,
2011 everything but the port is in hex. Not_Connected is displayed for sockets that
2012 are not tied to a specific remote address. The Tx and Rx queue sizes indicate
2013 the number of bytes pending for transmission and reception. The state
2014 indicates the state the socket is in and the uid is the owning uid of the
2017 The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
2018 it gives the network number, the node number, and indicates if the network is
2019 the primary network. It also indicates which device it is bound to (or
2020 Internal for internal networks) and the Frame Type if appropriate. Linux
2021 supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
2024 The /proc/net/ipx_route table holds a list of IPX routes. For each route it
2025 gives the destination network, the router node (or Directly) and the network
2026 address of the router (or Connected) for internal networks.
2028 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
2029 ----------------------------------------------------------
2031 The "mqueue" filesystem provides the necessary kernel features to enable the
2032 creation of a user space library that implements the POSIX message queues
2033 API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System
2034 Interfaces specification.)
2036 The "mqueue" filesystem contains values for determining/setting the amount of
2037 resources used by the file system.
2039 /proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the
2040 maximum number of message queues allowed on the system.
2042 /proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the
2043 maximum number of messages in a queue value. In fact it is the limiting value
2044 for another (user) limit which is set in mq_open invocation. This attribute of
2045 a queue must be less or equal then msg_max.
2047 /proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the
2048 maximum message size value (it is every message queue's attribute set during
2051 2.12 /proc/<pid>/oom_adj - Adjust the oom-killer score
2052 ------------------------------------------------------
2054 This file can be used to adjust the score used to select which processes
2055 should be killed in an out-of-memory situation. Giving it a high score will
2056 increase the likelihood of this process being killed by the oom-killer. Valid
2057 values are in the range -16 to +15, plus the special value -17, which disables
2058 oom-killing altogether for this process.
2060 2.13 /proc/<pid>/oom_score - Display current oom-killer score
2061 -------------------------------------------------------------
2063 ------------------------------------------------------------------------------
2064 This file can be used to check the current score used by the oom-killer is for
2065 any given <pid>. Use it together with /proc/<pid>/oom_adj to tune which
2066 process should be killed in an out-of-memory situation.
2068 ------------------------------------------------------------------------------
2070 ------------------------------------------------------------------------------
2071 Certain aspects of kernel behavior can be modified at runtime, without the
2072 need to recompile the kernel, or even to reboot the system. The files in the
2073 /proc/sys tree can not only be read, but also modified. You can use the echo
2074 command to write value into these files, thereby changing the default settings
2076 ------------------------------------------------------------------------------
2078 2.14 /proc/<pid>/io - Display the IO accounting fields
2079 -------------------------------------------------------
2081 This file contains IO statistics for each running process
2086 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
2089 test:/tmp # cat /proc/3828/io
2095 write_bytes: 323932160
2096 cancelled_write_bytes: 0
2105 I/O counter: chars read
2106 The number of bytes which this task has caused to be read from storage. This
2107 is simply the sum of bytes which this process passed to read() and pread().
2108 It includes things like tty IO and it is unaffected by whether or not actual
2109 physical disk IO was required (the read might have been satisfied from
2116 I/O counter: chars written
2117 The number of bytes which this task has caused, or shall cause to be written
2118 to disk. Similar caveats apply here as with rchar.
2124 I/O counter: read syscalls
2125 Attempt to count the number of read I/O operations, i.e. syscalls like read()
2132 I/O counter: write syscalls
2133 Attempt to count the number of write I/O operations, i.e. syscalls like
2134 write() and pwrite().
2140 I/O counter: bytes read
2141 Attempt to count the number of bytes which this process really did cause to
2142 be fetched from the storage layer. Done at the submit_bio() level, so it is
2143 accurate for block-backed filesystems. <please add status regarding NFS and
2144 CIFS at a later time>
2150 I/O counter: bytes written
2151 Attempt to count the number of bytes which this process caused to be sent to
2152 the storage layer. This is done at page-dirtying time.
2155 cancelled_write_bytes
2156 ---------------------
2158 The big inaccuracy here is truncate. If a process writes 1MB to a file and
2159 then deletes the file, it will in fact perform no writeout. But it will have
2160 been accounted as having caused 1MB of write.
2161 In other words: The number of bytes which this process caused to not happen,
2162 by truncating pagecache. A task can cause "negative" IO too. If this task
2163 truncates some dirty pagecache, some IO which another task has been accounted
2164 for (in it's write_bytes) will not be happening. We _could_ just subtract that
2165 from the truncating task's write_bytes, but there is information loss in doing
2172 At its current implementation state, this is a bit racy on 32-bit machines: if
2173 process A reads process B's /proc/pid/io while process B is updating one of
2174 those 64-bit counters, process A could see an intermediate result.
2177 More information about this can be found within the taskstats documentation in
2178 Documentation/accounting.
2180 ------------------------------------------------------------------------------
git.openpandora.org / pandora-kernel.git / blob