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
45 2.15 /proc/<pid>/coredump_filter - Core dump filtering settings
47 ------------------------------------------------------------------------------
49 ------------------------------------------------------------------------------
51 0.1 Introduction/Credits
52 ------------------------
54 This documentation is part of a soon (or so we hope) to be released book on
55 the SuSE Linux distribution. As there is no complete documentation for the
56 /proc file system and we've used many freely available sources to write these
57 chapters, it seems only fair to give the work back to the Linux community.
58 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
59 afraid it's still far from complete, but we hope it will be useful. As far as
60 we know, it is the first 'all-in-one' document about the /proc file system. It
61 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
62 SPARC, AXP, etc., features, you probably won't find what you are looking for.
63 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
64 additions and patches are welcome and will be added to this document if you
67 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
68 other people for help compiling this documentation. We'd also like to extend a
69 special thank you to Andi Kleen for documentation, which we relied on heavily
70 to create this document, as well as the additional information he provided.
71 Thanks to everybody else who contributed source or docs to the Linux kernel
72 and helped create a great piece of software... :)
74 If you have any comments, corrections or additions, please don't hesitate to
75 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
78 The latest version of this document is available online at
79 http://skaro.nightcrawler.com/~bb/Docs/Proc as HTML version.
81 If the above direction does not works for you, ypu could try the kernel
82 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
83 comandante@zaralinux.com.
88 We don't guarantee the correctness of this document, and if you come to us
89 complaining about how you screwed up your system because of incorrect
90 documentation, we won't feel responsible...
92 ------------------------------------------------------------------------------
93 CHAPTER 1: COLLECTING SYSTEM INFORMATION
94 ------------------------------------------------------------------------------
96 ------------------------------------------------------------------------------
98 ------------------------------------------------------------------------------
99 * Investigating the properties of the pseudo file system /proc and its
100 ability to provide information on the running Linux system
101 * Examining /proc's structure
102 * Uncovering various information about the kernel and the processes running
104 ------------------------------------------------------------------------------
107 The proc file system acts as an interface to internal data structures in the
108 kernel. It can be used to obtain information about the system and to change
109 certain kernel parameters at runtime (sysctl).
111 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
112 show you how you can use /proc/sys to change settings.
114 1.1 Process-Specific Subdirectories
115 -----------------------------------
117 The directory /proc contains (among other things) one subdirectory for each
118 process running on the system, which is named after the process ID (PID).
120 The link self points to the process reading the file system. Each process
121 subdirectory has the entries listed in Table 1-1.
124 Table 1-1: Process specific entries in /proc
125 ..............................................................................
127 clear_refs Clears page referenced bits shown in smaps output
128 cmdline Command line arguments
129 cpu Current and last cpu in which it was executed (2.4)(smp)
130 cwd Link to the current working directory
131 environ Values of environment variables
132 exe Link to the executable of this process
133 fd Directory, which contains all file descriptors
134 maps Memory maps to executables and library files (2.4)
135 mem Memory held by this process
136 root Link to the root directory of this process
138 statm Process memory status information
139 status Process status in human readable form
140 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan
141 smaps Extension based on maps, the rss size for each mapped file
142 ..............................................................................
144 For example, to get the status information of a process, all you have to do is
145 read the file /proc/PID/status:
147 >cat /proc/self/status
163 SigPnd: 0000000000000000
164 SigBlk: 0000000000000000
165 SigIgn: 0000000000000000
166 SigCgt: 0000000000000000
167 CapInh: 00000000fffffeff
168 CapPrm: 0000000000000000
169 CapEff: 0000000000000000
172 This shows you nearly the same information you would get if you viewed it with
173 the ps command. In fact, ps uses the proc file system to obtain its
174 information. The statm file contains more detailed information about the
175 process memory usage. Its seven fields are explained in Table 1-2. The stat
176 file contains details information about the process itself. Its fields are
177 explained in Table 1-3.
180 Table 1-2: Contents of the statm files (as of 2.6.8-rc3)
181 ..............................................................................
183 size total program size (pages) (same as VmSize in status)
184 resident size of memory portions (pages) (same as VmRSS in status)
185 shared number of pages that are shared (i.e. backed by a file)
186 trs number of pages that are 'code' (not including libs; broken,
187 includes data segment)
188 lrs number of pages of library (always 0 on 2.6)
189 drs number of pages of data/stack (including libs; broken,
190 includes library text)
191 dt number of dirty pages (always 0 on 2.6)
192 ..............................................................................
195 Table 1-3: Contents of the stat files (as of 2.6.22-rc3)
196 ..............................................................................
199 tcomm filename of the executable
200 state state (R is running, S is sleeping, D is sleeping in an
201 uninterruptible wait, Z is zombie, T is traced or stopped)
202 ppid process id of the parent process
203 pgrp pgrp of the process
205 tty_nr tty the process uses
206 tty_pgrp pgrp of the tty
208 min_flt number of minor faults
209 cmin_flt number of minor faults with child's
210 maj_flt number of major faults
211 cmaj_flt number of major faults with child's
212 utime user mode jiffies
213 stime kernel mode jiffies
214 cutime user mode jiffies with child's
215 cstime kernel mode jiffies with child's
216 priority priority level
218 num_threads number of threads
219 start_time time the process started after system boot
220 vsize virtual memory size
221 rss resident set memory size
222 rsslim current limit in bytes on the rss
223 start_code address above which program text can run
224 end_code address below which program text can run
225 start_stack address of the start of the stack
226 esp current value of ESP
227 eip current value of EIP
228 pending bitmap of pending signals (obsolete)
229 blocked bitmap of blocked signals (obsolete)
230 sigign bitmap of ignored signals (obsolete)
231 sigcatch bitmap of catched signals (obsolete)
232 wchan address where process went to sleep
235 exit_signal signal to send to parent thread on exit
236 task_cpu which CPU the task is scheduled on
237 rt_priority realtime priority
238 policy scheduling policy (man sched_setscheduler)
239 blkio_ticks time spent waiting for block IO
240 ..............................................................................
246 Similar to the process entries, the kernel data files give information about
247 the running kernel. The files used to obtain this information are contained in
248 /proc and are listed in Table 1-4. Not all of these will be present in your
249 system. It depends on the kernel configuration and the loaded modules, which
250 files are there, and which are missing.
252 Table 1-4: Kernel info in /proc
253 ..............................................................................
255 apm Advanced power management info
256 buddyinfo Kernel memory allocator information (see text) (2.5)
257 bus Directory containing bus specific information
258 cmdline Kernel command line
259 cpuinfo Info about the CPU
260 devices Available devices (block and character)
261 dma Used DMS channels
262 filesystems Supported filesystems
263 driver Various drivers grouped here, currently rtc (2.4)
264 execdomains Execdomains, related to security (2.4)
265 fb Frame Buffer devices (2.4)
266 fs File system parameters, currently nfs/exports (2.4)
267 ide Directory containing info about the IDE subsystem
268 interrupts Interrupt usage
269 iomem Memory map (2.4)
270 ioports I/O port usage
271 irq Masks for irq to cpu affinity (2.4)(smp?)
272 isapnp ISA PnP (Plug&Play) Info (2.4)
273 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
275 ksyms Kernel symbol table
276 loadavg Load average of last 1, 5 & 15 minutes
280 modules List of loaded modules
281 mounts Mounted filesystems
282 net Networking info (see text)
283 partitions Table of partitions known to the system
284 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
285 decoupled by lspci (2.4)
287 scsi SCSI info (see text)
288 slabinfo Slab pool info
289 stat Overall statistics
290 swaps Swap space utilization
292 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
293 tty Info of tty drivers
295 version Kernel version
296 video bttv info of video resources (2.4)
297 ..............................................................................
299 You can, for example, check which interrupts are currently in use and what
300 they are used for by looking in the file /proc/interrupts:
302 > cat /proc/interrupts
304 0: 8728810 XT-PIC timer
305 1: 895 XT-PIC keyboard
307 3: 531695 XT-PIC aha152x
308 4: 2014133 XT-PIC serial
309 5: 44401 XT-PIC pcnet_cs
312 12: 182918 XT-PIC PS/2 Mouse
314 14: 1232265 XT-PIC ide0
318 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
319 output of a SMP machine):
321 > cat /proc/interrupts
324 0: 1243498 1214548 IO-APIC-edge timer
325 1: 8949 8958 IO-APIC-edge keyboard
326 2: 0 0 XT-PIC cascade
327 5: 11286 10161 IO-APIC-edge soundblaster
328 8: 1 0 IO-APIC-edge rtc
329 9: 27422 27407 IO-APIC-edge 3c503
330 12: 113645 113873 IO-APIC-edge PS/2 Mouse
332 14: 22491 24012 IO-APIC-edge ide0
333 15: 2183 2415 IO-APIC-edge ide1
334 17: 30564 30414 IO-APIC-level eth0
335 18: 177 164 IO-APIC-level bttv
340 NMI is incremented in this case because every timer interrupt generates a NMI
341 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
343 LOC is the local interrupt counter of the internal APIC of every CPU.
345 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
346 connects the CPUs in a SMP system. This means that an error has been detected,
347 the IO-APIC automatically retry the transmission, so it should not be a big
348 problem, but you should read the SMP-FAQ.
350 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
351 /proc/interrupts to display every IRQ vector in use by the system, not
352 just those considered 'most important'. The new vectors are:
354 THR -- interrupt raised when a machine check threshold counter
355 (typically counting ECC corrected errors of memory or cache) exceeds
356 a configurable threshold. Only available on some systems.
358 TRM -- a thermal event interrupt occurs when a temperature threshold
359 has been exceeded for the CPU. This interrupt may also be generated
360 when the temperature drops back to normal.
362 SPU -- a spurious interrupt is some interrupt that was raised then lowered
363 by some IO device before it could be fully processed by the APIC. Hence
364 the APIC sees the interrupt but does not know what device it came from.
365 For this case the APIC will generate the interrupt with a IRQ vector
366 of 0xff. This might also be generated by chipset bugs.
368 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
369 sent from one CPU to another per the needs of the OS. Typically,
370 their statistics are used by kernel developers and interested users to
371 determine the occurance of interrupt of the given type.
373 The above IRQ vectors are displayed only when relevent. For example,
374 the threshold vector does not exist on x86_64 platforms. Others are
375 suppressed when the system is a uniprocessor. As of this writing, only
376 i386 and x86_64 platforms support the new IRQ vector displays.
378 Of some interest is the introduction of the /proc/irq directory to 2.4.
379 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
380 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
381 irq subdir is one subdir for each IRQ, and one file; prof_cpu_mask
385 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
386 1 11 13 15 17 19 3 5 7 9
390 The contents of the prof_cpu_mask file and each smp_affinity file for each IRQ
391 is the same by default:
393 > cat /proc/irq/0/smp_affinity
396 It's a bitmask, in which you can specify which CPUs can handle the IRQ, you can
399 > echo 1 > /proc/irq/prof_cpu_mask
401 This means that only the first CPU will handle the IRQ, but you can also echo 5
402 which means that only the first and fourth CPU can handle the IRQ.
404 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
405 between all the CPUs which are allowed to handle it. As usual the kernel has
406 more info than you and does a better job than you, so the defaults are the
407 best choice for almost everyone.
409 There are three more important subdirectories in /proc: net, scsi, and sys.
410 The general rule is that the contents, or even the existence of these
411 directories, depend on your kernel configuration. If SCSI is not enabled, the
412 directory scsi may not exist. The same is true with the net, which is there
413 only when networking support is present in the running kernel.
415 The slabinfo file gives information about memory usage at the slab level.
416 Linux uses slab pools for memory management above page level in version 2.2.
417 Commonly used objects have their own slab pool (such as network buffers,
418 directory cache, and so on).
420 ..............................................................................
422 > cat /proc/buddyinfo
424 Node 0, zone DMA 0 4 5 4 4 3 ...
425 Node 0, zone Normal 1 0 0 1 101 8 ...
426 Node 0, zone HighMem 2 0 0 1 1 0 ...
428 Memory fragmentation is a problem under some workloads, and buddyinfo is a
429 useful tool for helping diagnose these problems. Buddyinfo will give you a
430 clue as to how big an area you can safely allocate, or why a previous
433 Each column represents the number of pages of a certain order which are
434 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
435 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
436 available in ZONE_NORMAL, etc...
438 ..............................................................................
442 Provides information about distribution and utilization of memory. This
443 varies by architecture and compile options. The following is from a
444 16GB PIII, which has highmem enabled. You may not have all of these fields.
449 MemTotal: 16344972 kB
456 HighTotal: 15597528 kB
457 HighFree: 13629632 kB
466 CommitLimit: 7669796 kB
467 Committed_AS: 100056 kB
469 VmallocTotal: 112216 kB
471 VmallocChunk: 111088 kB
473 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
474 bits and the kernel binary code)
475 MemFree: The sum of LowFree+HighFree
476 Buffers: Relatively temporary storage for raw disk blocks
477 shouldn't get tremendously large (20MB or so)
478 Cached: in-memory cache for files read from the disk (the
479 pagecache). Doesn't include SwapCached
480 SwapCached: Memory that once was swapped out, is swapped back in but
481 still also is in the swapfile (if memory is needed it
482 doesn't need to be swapped out AGAIN because it is already
483 in the swapfile. This saves I/O)
484 Active: Memory that has been used more recently and usually not
485 reclaimed unless absolutely necessary.
486 Inactive: Memory which has been less recently used. It is more
487 eligible to be reclaimed for other purposes
489 HighFree: Highmem is all memory above ~860MB of physical memory
490 Highmem areas are for use by userspace programs, or
491 for the pagecache. The kernel must use tricks to access
492 this memory, making it slower to access than lowmem.
494 LowFree: Lowmem is memory which can be used for everything that
495 highmem can be used for, but it is also available for the
496 kernel's use for its own data structures. Among many
497 other things, it is where everything from the Slab is
498 allocated. Bad things happen when you're out of lowmem.
499 SwapTotal: total amount of swap space available
500 SwapFree: Memory which has been evicted from RAM, and is temporarily
502 Dirty: Memory which is waiting to get written back to the disk
503 Writeback: Memory which is actively being written back to the disk
504 Mapped: files which have been mmaped, such as libraries
505 Slab: in-kernel data structures cache
506 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
507 this is the total amount of memory currently available to
508 be allocated on the system. This limit is only adhered to
509 if strict overcommit accounting is enabled (mode 2 in
510 'vm.overcommit_memory').
511 The CommitLimit is calculated with the following formula:
512 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap
513 For example, on a system with 1G of physical RAM and 7G
514 of swap with a `vm.overcommit_ratio` of 30 it would
515 yield a CommitLimit of 7.3G.
516 For more details, see the memory overcommit documentation
517 in vm/overcommit-accounting.
518 Committed_AS: The amount of memory presently allocated on the system.
519 The committed memory is a sum of all of the memory which
520 has been allocated by processes, even if it has not been
521 "used" by them as of yet. A process which malloc()'s 1G
522 of memory, but only touches 300M of it will only show up
523 as using 300M of memory even if it has the address space
524 allocated for the entire 1G. This 1G is memory which has
525 been "committed" to by the VM and can be used at any time
526 by the allocating application. With strict overcommit
527 enabled on the system (mode 2 in 'vm.overcommit_memory'),
528 allocations which would exceed the CommitLimit (detailed
529 above) will not be permitted. This is useful if one needs
530 to guarantee that processes will not fail due to lack of
531 memory once that memory has been successfully allocated.
532 PageTables: amount of memory dedicated to the lowest level of page
534 VmallocTotal: total size of vmalloc memory area
535 VmallocUsed: amount of vmalloc area which is used
536 VmallocChunk: largest contigious block of vmalloc area which is free
539 1.3 IDE devices in /proc/ide
540 ----------------------------
542 The subdirectory /proc/ide contains information about all IDE devices of which
543 the kernel is aware. There is one subdirectory for each IDE controller, the
544 file drivers and a link for each IDE device, pointing to the device directory
545 in the controller specific subtree.
547 The file drivers contains general information about the drivers used for the
550 > cat /proc/ide/drivers
551 ide-cdrom version 4.53
552 ide-disk version 1.08
554 More detailed information can be found in the controller specific
555 subdirectories. These are named ide0, ide1 and so on. Each of these
556 directories contains the files shown in table 1-5.
559 Table 1-5: IDE controller info in /proc/ide/ide?
560 ..............................................................................
562 channel IDE channel (0 or 1)
563 config Configuration (only for PCI/IDE bridge)
565 model Type/Chipset of IDE controller
566 ..............................................................................
568 Each device connected to a controller has a separate subdirectory in the
569 controllers directory. The files listed in table 1-6 are contained in these
573 Table 1-6: IDE device information
574 ..............................................................................
577 capacity Capacity of the medium (in 512Byte blocks)
578 driver driver and version
579 geometry physical and logical geometry
580 identify device identify block
582 model device identifier
583 settings device setup
584 smart_thresholds IDE disk management thresholds
585 smart_values IDE disk management values
586 ..............................................................................
588 The most interesting file is settings. This file contains a nice overview of
589 the drive parameters:
591 # cat /proc/ide/ide0/hda/settings
592 name value min max mode
593 ---- ----- --- --- ----
594 bios_cyl 526 0 65535 rw
595 bios_head 255 0 255 rw
597 breada_readahead 4 0 127 rw
599 file_readahead 72 0 2097151 rw
601 keepsettings 0 0 1 rw
602 max_kb_per_request 122 1 127 rw
606 pio_mode write-only 0 255 w
612 1.4 Networking info in /proc/net
613 --------------------------------
615 The subdirectory /proc/net follows the usual pattern. Table 1-6 shows the
616 additional values you get for IP version 6 if you configure the kernel to
617 support this. Table 1-7 lists the files and their meaning.
620 Table 1-6: IPv6 info in /proc/net
621 ..............................................................................
623 udp6 UDP sockets (IPv6)
624 tcp6 TCP sockets (IPv6)
625 raw6 Raw device statistics (IPv6)
626 igmp6 IP multicast addresses, which this host joined (IPv6)
627 if_inet6 List of IPv6 interface addresses
628 ipv6_route Kernel routing table for IPv6
629 rt6_stats Global IPv6 routing tables statistics
630 sockstat6 Socket statistics (IPv6)
631 snmp6 Snmp data (IPv6)
632 ..............................................................................
635 Table 1-7: Network info in /proc/net
636 ..............................................................................
639 dev network devices with statistics
640 dev_mcast the Layer2 multicast groups a device is listening too
641 (interface index, label, number of references, number of bound
643 dev_stat network device status
644 ip_fwchains Firewall chain linkage
645 ip_fwnames Firewall chain names
646 ip_masq Directory containing the masquerading tables
647 ip_masquerade Major masquerading table
648 netstat Network statistics
649 raw raw device statistics
650 route Kernel routing table
651 rpc Directory containing rpc info
652 rt_cache Routing cache
654 sockstat Socket statistics
656 tr_rif Token ring RIF routing table
658 unix UNIX domain sockets
659 wireless Wireless interface data (Wavelan etc)
660 igmp IP multicast addresses, which this host joined
661 psched Global packet scheduler parameters.
662 netlink List of PF_NETLINK sockets
663 ip_mr_vifs List of multicast virtual interfaces
664 ip_mr_cache List of multicast routing cache
665 ..............................................................................
667 You can use this information to see which network devices are available in
668 your system and how much traffic was routed over those devices:
672 face |bytes packets errs drop fifo frame compressed multicast|[...
673 lo: 908188 5596 0 0 0 0 0 0 [...
674 ppp0:15475140 20721 410 0 0 410 0 0 [...
675 eth0: 614530 7085 0 0 0 0 0 1 [...
678 ...] bytes packets errs drop fifo colls carrier compressed
679 ...] 908188 5596 0 0 0 0 0 0
680 ...] 1375103 17405 0 0 0 0 0 0
681 ...] 1703981 5535 0 0 0 3 0 0
683 In addition, each Channel Bond interface has it's own directory. For
684 example, the bond0 device will have a directory called /proc/net/bond0/.
685 It will contain information that is specific to that bond, such as the
686 current slaves of the bond, the link status of the slaves, and how
687 many times the slaves link has failed.
692 If you have a SCSI host adapter in your system, you'll find a subdirectory
693 named after the driver for this adapter in /proc/scsi. You'll also see a list
694 of all recognized SCSI devices in /proc/scsi:
698 Host: scsi0 Channel: 00 Id: 00 Lun: 00
699 Vendor: IBM Model: DGHS09U Rev: 03E0
700 Type: Direct-Access ANSI SCSI revision: 03
701 Host: scsi0 Channel: 00 Id: 06 Lun: 00
702 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
703 Type: CD-ROM ANSI SCSI revision: 02
706 The directory named after the driver has one file for each adapter found in
707 the system. These files contain information about the controller, including
708 the used IRQ and the IO address range. The amount of information shown is
709 dependent on the adapter you use. The example shows the output for an Adaptec
710 AHA-2940 SCSI adapter:
712 > cat /proc/scsi/aic7xxx/0
714 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
716 TCQ Enabled By Default : Disabled
717 AIC7XXX_PROC_STATS : Disabled
718 AIC7XXX_RESET_DELAY : 5
719 Adapter Configuration:
720 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
721 Ultra Wide Controller
722 PCI MMAPed I/O Base: 0xeb001000
723 Adapter SEEPROM Config: SEEPROM found and used.
724 Adaptec SCSI BIOS: Enabled
726 SCBs: Active 0, Max Active 2,
727 Allocated 15, HW 16, Page 255
729 BIOS Control Word: 0x18b6
730 Adapter Control Word: 0x005b
731 Extended Translation: Enabled
732 Disconnect Enable Flags: 0xffff
733 Ultra Enable Flags: 0x0001
734 Tag Queue Enable Flags: 0x0000
735 Ordered Queue Tag Flags: 0x0000
736 Default Tag Queue Depth: 8
737 Tagged Queue By Device array for aic7xxx host instance 0:
738 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
739 Actual queue depth per device for aic7xxx host instance 0:
740 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
743 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
744 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
745 Total transfers 160151 (74577 reads and 85574 writes)
747 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
748 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
749 Total transfers 0 (0 reads and 0 writes)
752 1.6 Parallel port info in /proc/parport
753 ---------------------------------------
755 The directory /proc/parport contains information about the parallel ports of
756 your system. It has one subdirectory for each port, named after the port
759 These directories contain the four files shown in Table 1-8.
762 Table 1-8: Files in /proc/parport
763 ..............................................................................
765 autoprobe Any IEEE-1284 device ID information that has been acquired.
766 devices list of the device drivers using that port. A + will appear by the
767 name of the device currently using the port (it might not appear
769 hardware Parallel port's base address, IRQ line and DMA channel.
770 irq IRQ that parport is using for that port. This is in a separate
771 file to allow you to alter it by writing a new value in (IRQ
773 ..............................................................................
775 1.7 TTY info in /proc/tty
776 -------------------------
778 Information about the available and actually used tty's can be found in the
779 directory /proc/tty.You'll find entries for drivers and line disciplines in
780 this directory, as shown in Table 1-9.
783 Table 1-9: Files in /proc/tty
784 ..............................................................................
786 drivers list of drivers and their usage
787 ldiscs registered line disciplines
788 driver/serial usage statistic and status of single tty lines
789 ..............................................................................
791 To see which tty's are currently in use, you can simply look into the file
794 > cat /proc/tty/drivers
795 pty_slave /dev/pts 136 0-255 pty:slave
796 pty_master /dev/ptm 128 0-255 pty:master
797 pty_slave /dev/ttyp 3 0-255 pty:slave
798 pty_master /dev/pty 2 0-255 pty:master
799 serial /dev/cua 5 64-67 serial:callout
800 serial /dev/ttyS 4 64-67 serial
801 /dev/tty0 /dev/tty0 4 0 system:vtmaster
802 /dev/ptmx /dev/ptmx 5 2 system
803 /dev/console /dev/console 5 1 system:console
804 /dev/tty /dev/tty 5 0 system:/dev/tty
805 unknown /dev/tty 4 1-63 console
808 1.8 Miscellaneous kernel statistics in /proc/stat
809 -------------------------------------------------
811 Various pieces of information about kernel activity are available in the
812 /proc/stat file. All of the numbers reported in this file are aggregates
813 since the system first booted. For a quick look, simply cat the file:
816 cpu 2255 34 2290 22625563 6290 127 456 0
817 cpu0 1132 34 1441 11311718 3675 127 438 0
818 cpu1 1123 0 849 11313845 2614 0 18 0
819 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
826 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
827 lines. These numbers identify the amount of time the CPU has spent performing
828 different kinds of work. Time units are in USER_HZ (typically hundredths of a
829 second). The meanings of the columns are as follows, from left to right:
831 - user: normal processes executing in user mode
832 - nice: niced processes executing in user mode
833 - system: processes executing in kernel mode
834 - idle: twiddling thumbs
835 - iowait: waiting for I/O to complete
836 - irq: servicing interrupts
837 - softirq: servicing softirqs
838 - steal: involuntary wait
840 The "intr" line gives counts of interrupts serviced since boot time, for each
841 of the possible system interrupts. The first column is the total of all
842 interrupts serviced; each subsequent column is the total for that particular
845 The "ctxt" line gives the total number of context switches across all CPUs.
847 The "btime" line gives the time at which the system booted, in seconds since
850 The "processes" line gives the number of processes and threads created, which
851 includes (but is not limited to) those created by calls to the fork() and
852 clone() system calls.
854 The "procs_running" line gives the number of processes currently running on
857 The "procs_blocked" line gives the number of processes currently blocked,
858 waiting for I/O to complete.
860 1.9 Ext4 file system parameters
861 ------------------------------
862 Ext4 file system have one directory per partition under /proc/fs/ext4/
863 # ls /proc/fs/ext4/hdc/
864 group_prealloc max_to_scan mb_groups mb_history min_to_scan order2_req
868 This file gives the details of mutiblock allocator buddy cache of free blocks
871 Multiblock allocation history.
874 This file indicate whether the multiblock allocator should start collecting
875 statistics. The statistics are shown during unmount
878 The multiblock allocator normalize the block allocation request to
879 group_prealloc filesystem blocks if we don't have strip value set.
880 The stripe value can be specified at mount time or during mke2fs.
883 How long multiblock allocator can look for a best extent (in found extents)
886 How long multiblock allocator must look for a best extent
889 Multiblock allocator use 2^N search using buddies only for requests greater
890 than or equal to order2_req. The request size is specfied in file system
891 blocks. A value of 2 indicate only if the requests are greater than or equal
895 Files smaller than stream_req are served by the stream allocator, whose
896 purpose is to pack requests as close each to other as possible to
897 produce smooth I/O traffic. Avalue of 16 indicate that file smaller than 16
898 filesystem block size will use group based preallocation.
900 ------------------------------------------------------------------------------
902 ------------------------------------------------------------------------------
903 The /proc file system serves information about the running system. It not only
904 allows access to process data but also allows you to request the kernel status
905 by reading files in the hierarchy.
907 The directory structure of /proc reflects the types of information and makes
908 it easy, if not obvious, where to look for specific data.
909 ------------------------------------------------------------------------------
911 ------------------------------------------------------------------------------
912 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
913 ------------------------------------------------------------------------------
915 ------------------------------------------------------------------------------
917 ------------------------------------------------------------------------------
918 * Modifying kernel parameters by writing into files found in /proc/sys
919 * Exploring the files which modify certain parameters
920 * Review of the /proc/sys file tree
921 ------------------------------------------------------------------------------
924 A very interesting part of /proc is the directory /proc/sys. This is not only
925 a source of information, it also allows you to change parameters within the
926 kernel. Be very careful when attempting this. You can optimize your system,
927 but you can also cause it to crash. Never alter kernel parameters on a
928 production system. Set up a development machine and test to make sure that
929 everything works the way you want it to. You may have no alternative but to
930 reboot the machine once an error has been made.
932 To change a value, simply echo the new value into the file. An example is
933 given below in the section on the file system data. You need to be root to do
934 this. You can create your own boot script to perform this every time your
937 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
938 general things in the operation of the Linux kernel. Since some of the files
939 can inadvertently disrupt your system, it is advisable to read both
940 documentation and source before actually making adjustments. In any case, be
941 very careful when writing to any of these files. The entries in /proc may
942 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
943 review the kernel documentation in the directory /usr/src/linux/Documentation.
944 This chapter is heavily based on the documentation included in the pre 2.2
945 kernels, and became part of it in version 2.2.1 of the Linux kernel.
947 2.1 /proc/sys/fs - File system data
948 -----------------------------------
950 This subdirectory contains specific file system, file handle, inode, dentry
951 and quota information.
953 Currently, these files are in /proc/sys/fs:
958 Status of the directory cache. Since directory entries are dynamically
959 allocated and deallocated, this file indicates the current status. It holds
960 six values, in which the last two are not used and are always zero. The others
961 are listed in table 2-1.
964 Table 2-1: Status files of the directory cache
965 ..............................................................................
967 nr_dentry Almost always zero
968 nr_unused Number of unused cache entries
970 in seconds after the entry may be reclaimed, when memory is short
971 want_pages internally
972 ..............................................................................
974 dquot-nr and dquot-max
975 ----------------------
977 The file dquot-max shows the maximum number of cached disk quota entries.
979 The file dquot-nr shows the number of allocated disk quota entries and the
980 number of free disk quota entries.
982 If the number of available cached disk quotas is very low and you have a large
983 number of simultaneous system users, you might want to raise the limit.
988 The kernel allocates file handles dynamically, but doesn't free them again at
991 The value in file-max denotes the maximum number of file handles that the
992 Linux kernel will allocate. When you get a lot of error messages about running
993 out of file handles, you might want to raise this limit. The default value is
994 10% of RAM in kilobytes. To change it, just write the new number into the
997 # cat /proc/sys/fs/file-max
999 # echo 8192 > /proc/sys/fs/file-max
1000 # cat /proc/sys/fs/file-max
1004 This method of revision is useful for all customizable parameters of the
1005 kernel - simply echo the new value to the corresponding file.
1007 Historically, the three values in file-nr denoted the number of allocated file
1008 handles, the number of allocated but unused file handles, and the maximum
1009 number of file handles. Linux 2.6 always reports 0 as the number of free file
1010 handles -- this is not an error, it just means that the number of allocated
1011 file handles exactly matches the number of used file handles.
1013 Attempts to allocate more file descriptors than file-max are reported with
1014 printk, look for "VFS: file-max limit <number> reached".
1016 inode-state and inode-nr
1017 ------------------------
1019 The file inode-nr contains the first two items from inode-state, so we'll skip
1022 inode-state contains two actual numbers and five dummy values. The numbers
1023 are nr_inodes and nr_free_inodes (in order of appearance).
1028 Denotes the number of inodes the system has allocated. This number will
1029 grow and shrink dynamically.
1034 Represents the number of free inodes. Ie. The number of inuse inodes is
1035 (nr_inodes - nr_free_inodes).
1037 aio-nr and aio-max-nr
1038 ---------------------
1040 aio-nr is the running total of the number of events specified on the
1041 io_setup system call for all currently active aio contexts. If aio-nr
1042 reaches aio-max-nr then io_setup will fail with EAGAIN. Note that
1043 raising aio-max-nr does not result in the pre-allocation or re-sizing
1044 of any kernel data structures.
1046 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
1047 -----------------------------------------------------------
1049 Besides these files, there is the subdirectory /proc/sys/fs/binfmt_misc. This
1050 handles the kernel support for miscellaneous binary formats.
1052 Binfmt_misc provides the ability to register additional binary formats to the
1053 Kernel without compiling an additional module/kernel. Therefore, binfmt_misc
1054 needs to know magic numbers at the beginning or the filename extension of the
1057 It works by maintaining a linked list of structs that contain a description of
1058 a binary format, including a magic with size (or the filename extension),
1059 offset and mask, and the interpreter name. On request it invokes the given
1060 interpreter with the original program as argument, as binfmt_java and
1061 binfmt_em86 and binfmt_mz do. Since binfmt_misc does not define any default
1062 binary-formats, you have to register an additional binary-format.
1064 There are two general files in binfmt_misc and one file per registered format.
1065 The two general files are register and status.
1067 Registering a new binary format
1068 -------------------------------
1070 To register a new binary format you have to issue the command
1072 echo :name:type:offset:magic:mask:interpreter: > /proc/sys/fs/binfmt_misc/register
1076 with appropriate name (the name for the /proc-dir entry), offset (defaults to
1077 0, if omitted), magic, mask (which can be omitted, defaults to all 0xff) and
1078 last but not least, the interpreter that is to be invoked (for example and
1079 testing /bin/echo). Type can be M for usual magic matching or E for filename
1080 extension matching (give extension in place of magic).
1082 Check or reset the status of the binary format handler
1083 ------------------------------------------------------
1085 If you do a cat on the file /proc/sys/fs/binfmt_misc/status, you will get the
1086 current status (enabled/disabled) of binfmt_misc. Change the status by echoing
1087 0 (disables) or 1 (enables) or -1 (caution: this clears all previously
1088 registered binary formats) to status. For example echo 0 > status to disable
1089 binfmt_misc (temporarily).
1091 Status of a single handler
1092 --------------------------
1094 Each registered handler has an entry in /proc/sys/fs/binfmt_misc. These files
1095 perform the same function as status, but their scope is limited to the actual
1096 binary format. By cating this file, you also receive all related information
1097 about the interpreter/magic of the binfmt.
1099 Example usage of binfmt_misc (emulate binfmt_java)
1100 --------------------------------------------------
1102 cd /proc/sys/fs/binfmt_misc
1103 echo ':Java:M::\xca\xfe\xba\xbe::/usr/local/java/bin/javawrapper:' > register
1104 echo ':HTML:E::html::/usr/local/java/bin/appletviewer:' > register
1105 echo ':Applet:M::<!--applet::/usr/local/java/bin/appletviewer:' > register
1106 echo ':DEXE:M::\x0eDEX::/usr/bin/dosexec:' > register
1109 These four lines add support for Java executables and Java applets (like
1110 binfmt_java, additionally recognizing the .html extension with no need to put
1111 <!--applet> to every applet file). You have to install the JDK and the
1112 shell-script /usr/local/java/bin/javawrapper too. It works around the
1113 brokenness of the Java filename handling. To add a Java binary, just create a
1114 link to the class-file somewhere in the path.
1116 2.3 /proc/sys/kernel - general kernel parameters
1117 ------------------------------------------------
1119 This directory reflects general kernel behaviors. As I've said before, the
1120 contents depend on your configuration. Here you'll find the most important
1121 files, along with descriptions of what they mean and how to use them.
1126 The file contains three values; highwater, lowwater, and frequency.
1128 It exists only when BSD-style process accounting is enabled. These values
1129 control its behavior. If the free space on the file system where the log lives
1130 goes below lowwater percentage, accounting suspends. If it goes above
1131 highwater percentage, accounting resumes. Frequency determines how often you
1132 check the amount of free space (value is in seconds). Default settings are: 4,
1133 2, and 30. That is, suspend accounting if there is less than 2 percent free;
1134 resume it if we have a value of 3 or more percent; consider information about
1135 the amount of free space valid for 30 seconds
1140 The file contains a single value denoting the limit on the argv array size
1141 for execve (in KiB). This limit is only applied when system call auditing for
1142 execve is enabled, otherwise the value is ignored.
1147 When the value in this file is 0, ctrl-alt-del is trapped and sent to the init
1148 program to handle a graceful restart. However, when the value is greater that
1149 zero, Linux's reaction to this key combination will be an immediate reboot,
1150 without syncing its dirty buffers.
1153 When a program (like dosemu) has the keyboard in raw mode, the
1154 ctrl-alt-del is intercepted by the program before it ever reaches the
1155 kernel tty layer, and it is up to the program to decide what to do with
1158 domainname and hostname
1159 -----------------------
1161 These files can be controlled to set the NIS domainname and hostname of your
1162 box. For the classic darkstar.frop.org a simple:
1164 # echo "darkstar" > /proc/sys/kernel/hostname
1165 # echo "frop.org" > /proc/sys/kernel/domainname
1168 would suffice to set your hostname and NIS domainname.
1170 osrelease, ostype and version
1171 -----------------------------
1173 The names make it pretty obvious what these fields contain:
1175 > cat /proc/sys/kernel/osrelease
1178 > cat /proc/sys/kernel/ostype
1181 > cat /proc/sys/kernel/version
1182 #4 Fri Oct 1 12:41:14 PDT 1999
1185 The files osrelease and ostype should be clear enough. Version needs a little
1186 more clarification. The #4 means that this is the 4th kernel built from this
1187 source base and the date after it indicates the time the kernel was built. The
1188 only way to tune these values is to rebuild the kernel.
1193 The value in this file represents the number of seconds the kernel waits
1194 before rebooting on a panic. When you use the software watchdog, the
1195 recommended setting is 60. If set to 0, the auto reboot after a kernel panic
1196 is disabled, which is the default setting.
1201 The four values in printk denote
1203 * default_message_loglevel,
1204 * minimum_console_loglevel and
1205 * default_console_loglevel
1208 These values influence printk() behavior when printing or logging error
1209 messages, which come from inside the kernel. See syslog(2) for more
1210 information on the different log levels.
1215 Messages with a higher priority than this will be printed to the console.
1217 default_message_level
1218 ---------------------
1220 Messages without an explicit priority will be printed with this priority.
1222 minimum_console_loglevel
1223 ------------------------
1225 Minimum (highest) value to which the console_loglevel can be set.
1227 default_console_loglevel
1228 ------------------------
1230 Default value for console_loglevel.
1235 This file shows the size of the generic SCSI (sg) buffer. At this point, you
1236 can't tune it yet, but you can change it at compile time by editing
1237 include/scsi/sg.h and changing the value of SG_BIG_BUFF.
1239 If you use a scanner with SANE (Scanner Access Now Easy) you might want to set
1240 this to a higher value. Refer to the SANE documentation on this issue.
1245 The location where the modprobe binary is located. The kernel uses this
1246 program to load modules on demand.
1251 The value in this file affects behavior of handling NMI. When the value is
1252 non-zero, unknown NMI is trapped and then panic occurs. At that time, kernel
1253 debugging information is displayed on console.
1255 NMI switch that most IA32 servers have fires unknown NMI up, for example.
1256 If a system hangs up, try pressing the NMI switch.
1261 Enables/Disables the NMI watchdog on x86 systems. When the value is non-zero
1262 the NMI watchdog is enabled and will continuously test all online cpus to
1263 determine whether or not they are still functioning properly.
1265 Because the NMI watchdog shares registers with oprofile, by disabling the NMI
1266 watchdog, oprofile may have more registers to utilize.
1271 Enables/Disables the protection of the per-process proc entries "maps" and
1272 "smaps". When enabled, the contents of these files are visible only to
1273 readers that are allowed to ptrace() the given process.
1276 2.4 /proc/sys/vm - The virtual memory subsystem
1277 -----------------------------------------------
1279 The files in this directory can be used to tune the operation of the virtual
1280 memory (VM) subsystem of the Linux kernel.
1285 Controls the tendency of the kernel to reclaim the memory which is used for
1286 caching of directory and inode objects.
1288 At the default value of vfs_cache_pressure=100 the kernel will attempt to
1289 reclaim dentries and inodes at a "fair" rate with respect to pagecache and
1290 swapcache reclaim. Decreasing vfs_cache_pressure causes the kernel to prefer
1291 to retain dentry and inode caches. Increasing vfs_cache_pressure beyond 100
1292 causes the kernel to prefer to reclaim dentries and inodes.
1294 dirty_background_ratio
1295 ----------------------
1297 Contains, as a percentage of total system memory, the number of pages at which
1298 the pdflush background writeback daemon will start writing out dirty data.
1303 Contains, as a percentage of total system memory, the number of pages at which
1304 a process which is generating disk writes will itself start writing out dirty
1307 dirty_writeback_centisecs
1308 -------------------------
1310 The pdflush writeback daemons will periodically wake up and write `old' data
1311 out to disk. This tunable expresses the interval between those wakeups, in
1312 100'ths of a second.
1314 Setting this to zero disables periodic writeback altogether.
1316 dirty_expire_centisecs
1317 ----------------------
1319 This tunable is used to define when dirty data is old enough to be eligible
1320 for writeout by the pdflush daemons. It is expressed in 100'ths of a second.
1321 Data which has been dirty in-memory for longer than this interval will be
1322 written out next time a pdflush daemon wakes up.
1327 If non-zero, this sysctl disables the new 32-bit mmap mmap layout - the kernel
1328 will use the legacy (2.4) layout for all processes.
1330 lower_zone_protection
1331 ---------------------
1333 For some specialised workloads on highmem machines it is dangerous for
1334 the kernel to allow process memory to be allocated from the "lowmem"
1335 zone. This is because that memory could then be pinned via the mlock()
1336 system call, or by unavailability of swapspace.
1338 And on large highmem machines this lack of reclaimable lowmem memory
1341 So the Linux page allocator has a mechanism which prevents allocations
1342 which _could_ use highmem from using too much lowmem. This means that
1343 a certain amount of lowmem is defended from the possibility of being
1344 captured into pinned user memory.
1346 (The same argument applies to the old 16 megabyte ISA DMA region. This
1347 mechanism will also defend that region from allocations which could use
1350 The `lower_zone_protection' tunable determines how aggressive the kernel is
1351 in defending these lower zones. The default value is zero - no
1354 If you have a machine which uses highmem or ISA DMA and your
1355 applications are using mlock(), or if you are running with no swap then
1356 you probably should increase the lower_zone_protection setting.
1358 The units of this tunable are fairly vague. It is approximately equal
1359 to "megabytes," so setting lower_zone_protection=100 will protect around 100
1360 megabytes of the lowmem zone from user allocations. It will also make
1361 those 100 megabytes unavailable for use by applications and by
1362 pagecache, so there is a cost.
1364 The effects of this tunable may be observed by monitoring
1365 /proc/meminfo:LowFree. Write a single huge file and observe the point
1366 at which LowFree ceases to fall.
1368 A reasonable value for lower_zone_protection is 100.
1373 page-cluster controls the number of pages which are written to swap in
1374 a single attempt. The swap I/O size.
1376 It is a logarithmic value - setting it to zero means "1 page", setting
1377 it to 1 means "2 pages", setting it to 2 means "4 pages", etc.
1379 The default value is three (eight pages at a time). There may be some
1380 small benefits in tuning this to a different value if your workload is
1386 Controls overcommit of system memory, possibly allowing processes
1387 to allocate (but not use) more memory than is actually available.
1390 0 - Heuristic overcommit handling. Obvious overcommits of
1391 address space are refused. Used for a typical system. It
1392 ensures a seriously wild allocation fails while allowing
1393 overcommit to reduce swap usage. root is allowed to
1394 allocate slightly more memory in this mode. This is the
1397 1 - Always overcommit. Appropriate for some scientific
1400 2 - Don't overcommit. The total address space commit
1401 for the system is not permitted to exceed swap plus a
1402 configurable percentage (default is 50) of physical RAM.
1403 Depending on the percentage you use, in most situations
1404 this means a process will not be killed while attempting
1405 to use already-allocated memory but will receive errors
1406 on memory allocation as appropriate.
1411 Percentage of physical memory size to include in overcommit calculations
1414 Memory allocation limit = swapspace + physmem * (overcommit_ratio / 100)
1416 swapspace = total size of all swap areas
1417 physmem = size of physical memory in system
1419 nr_hugepages and hugetlb_shm_group
1420 ----------------------------------
1422 nr_hugepages configures number of hugetlb page reserved for the system.
1424 hugetlb_shm_group contains group id that is allowed to create SysV shared
1425 memory segment using hugetlb page.
1427 hugepages_treat_as_movable
1428 --------------------------
1430 This parameter is only useful when kernelcore= is specified at boot time to
1431 create ZONE_MOVABLE for pages that may be reclaimed or migrated. Huge pages
1432 are not movable so are not normally allocated from ZONE_MOVABLE. A non-zero
1433 value written to hugepages_treat_as_movable allows huge pages to be allocated
1436 Once enabled, the ZONE_MOVABLE is treated as an area of memory the huge
1437 pages pool can easily grow or shrink within. Assuming that applications are
1438 not running that mlock() a lot of memory, it is likely the huge pages pool
1439 can grow to the size of ZONE_MOVABLE by repeatedly entering the desired value
1440 into nr_hugepages and triggering page reclaim.
1445 laptop_mode is a knob that controls "laptop mode". All the things that are
1446 controlled by this knob are discussed in Documentation/laptop-mode.txt.
1451 block_dump enables block I/O debugging when set to a nonzero value. More
1452 information on block I/O debugging is in Documentation/laptop-mode.txt.
1457 This file contains valid hold time of swap out protection token. The Linux
1458 VM has token based thrashing control mechanism and uses the token to prevent
1459 unnecessary page faults in thrashing situation. The unit of the value is
1460 second. The value would be useful to tune thrashing behavior.
1465 Writing to this will cause the kernel to drop clean caches, dentries and
1466 inodes from memory, causing that memory to become free.
1469 echo 1 > /proc/sys/vm/drop_caches
1470 To free dentries and inodes:
1471 echo 2 > /proc/sys/vm/drop_caches
1472 To free pagecache, dentries and inodes:
1473 echo 3 > /proc/sys/vm/drop_caches
1475 As this is a non-destructive operation and dirty objects are not freeable, the
1476 user should run `sync' first.
1479 2.5 /proc/sys/dev - Device specific parameters
1480 ----------------------------------------------
1482 Currently there is only support for CDROM drives, and for those, there is only
1483 one read-only file containing information about the CD-ROM drives attached to
1486 >cat /proc/sys/dev/cdrom/info
1487 CD-ROM information, Id: cdrom.c 2.55 1999/04/25
1491 drive # of slots: 1 0
1495 Can change speed: 1 1
1496 Can select disk: 0 1
1497 Can read multisession: 1 1
1499 Reports media changed: 1 1
1503 You see two drives, sr0 and hdb, along with a list of their features.
1505 2.6 /proc/sys/sunrpc - Remote procedure calls
1506 ---------------------------------------------
1508 This directory contains four files, which enable or disable debugging for the
1509 RPC functions NFS, NFS-daemon, RPC and NLM. The default values are 0. They can
1510 be set to one to turn debugging on. (The default value is 0 for each)
1512 2.7 /proc/sys/net - Networking stuff
1513 ------------------------------------
1515 The interface to the networking parts of the kernel is located in
1516 /proc/sys/net. Table 2-3 shows all possible subdirectories. You may see only
1517 some of them, depending on your kernel's configuration.
1520 Table 2-3: Subdirectories in /proc/sys/net
1521 ..............................................................................
1522 Directory Content Directory Content
1523 core General parameter appletalk Appletalk protocol
1524 unix Unix domain sockets netrom NET/ROM
1525 802 E802 protocol ax25 AX25
1526 ethernet Ethernet protocol rose X.25 PLP layer
1527 ipv4 IP version 4 x25 X.25 protocol
1528 ipx IPX token-ring IBM token ring
1529 bridge Bridging decnet DEC net
1531 ..............................................................................
1533 We will concentrate on IP networking here. Since AX15, X.25, and DEC Net are
1534 only minor players in the Linux world, we'll skip them in this chapter. You'll
1535 find some short info on Appletalk and IPX further on in this chapter. Review
1536 the online documentation and the kernel source to get a detailed view of the
1537 parameters for those protocols. In this section we'll discuss the
1538 subdirectories printed in bold letters in the table above. As default values
1539 are suitable for most needs, there is no need to change these values.
1541 /proc/sys/net/core - Network core options
1542 -----------------------------------------
1547 The default setting of the socket receive buffer in bytes.
1552 The maximum receive socket buffer size in bytes.
1557 The default setting (in bytes) of the socket send buffer.
1562 The maximum send socket buffer size in bytes.
1564 message_burst and message_cost
1565 ------------------------------
1567 These parameters are used to limit the warning messages written to the kernel
1568 log from the networking code. They enforce a rate limit to make a
1569 denial-of-service attack impossible. A higher message_cost factor, results in
1570 fewer messages that will be written. Message_burst controls when messages will
1571 be dropped. The default settings limit warning messages to one every five
1577 This controls console messages from the networking stack that can occur because
1578 of problems on the network like duplicate address or bad checksums. Normally,
1579 this should be enabled, but if the problem persists the messages can be
1586 Maximum number of packets, queued on the INPUT side, when the interface
1587 receives packets faster than kernel can process them.
1592 Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
1593 of struct cmsghdr structures with appended data.
1595 /proc/sys/net/unix - Parameters for Unix domain sockets
1596 -------------------------------------------------------
1598 There are only two files in this subdirectory. They control the delays for
1599 deleting and destroying socket descriptors.
1601 2.8 /proc/sys/net/ipv4 - IPV4 settings
1602 --------------------------------------
1604 IP version 4 is still the most used protocol in Unix networking. It will be
1605 replaced by IP version 6 in the next couple of years, but for the moment it's
1606 the de facto standard for the internet and is used in most networking
1607 environments around the world. Because of the importance of this protocol,
1608 we'll have a deeper look into the subtree controlling the behavior of the IPv4
1609 subsystem of the Linux kernel.
1611 Let's start with the entries in /proc/sys/net/ipv4.
1616 icmp_echo_ignore_all and icmp_echo_ignore_broadcasts
1617 ----------------------------------------------------
1619 Turn on (1) or off (0), if the kernel should ignore all ICMP ECHO requests, or
1620 just those to broadcast and multicast addresses.
1622 Please note that if you accept ICMP echo requests with a broadcast/multi\-cast
1623 destination address your network may be used as an exploder for denial of
1624 service packet flooding attacks to other hosts.
1626 icmp_destunreach_rate, icmp_echoreply_rate, icmp_paramprob_rate and icmp_timeexeed_rate
1627 ---------------------------------------------------------------------------------------
1629 Sets limits for sending ICMP packets to specific targets. A value of zero
1630 disables all limiting. Any positive value sets the maximum package rate in
1631 hundredth of a second (on Intel systems).
1639 This file contains the number one if the host received its IP configuration by
1640 RARP, BOOTP, DHCP or a similar mechanism. Otherwise it is zero.
1645 TTL (Time To Live) for IPv4 interfaces. This is simply the maximum number of
1646 hops a packet may travel.
1651 Enable dynamic socket address rewriting on interface address change. This is
1652 useful for dialup interface with changing IP addresses.
1657 Enable or disable forwarding of IP packages between interfaces. Changing this
1658 value resets all other parameters to their default values. They differ if the
1659 kernel is configured as host or router.
1664 Range of ports used by TCP and UDP to choose the local port. Contains two
1665 numbers, the first number is the lowest port, the second number the highest
1666 local port. Default is 1024-4999. Should be changed to 32768-61000 for
1672 Global switch to turn path MTU discovery off. It can also be set on a per
1673 socket basis by the applications or on a per route basis.
1678 Enable/disable debugging of IP masquerading.
1680 IP fragmentation settings
1681 -------------------------
1683 ipfrag_high_trash and ipfrag_low_trash
1684 --------------------------------------
1686 Maximum memory used to reassemble IP fragments. When ipfrag_high_thresh bytes
1687 of memory is allocated for this purpose, the fragment handler will toss
1688 packets until ipfrag_low_thresh is reached.
1693 Time in seconds to keep an IP fragment in memory.
1701 This file controls the use of the ECN bit in the IPv4 headers. This is a new
1702 feature about Explicit Congestion Notification, but some routers and firewalls
1703 block traffic that has this bit set, so it could be necessary to echo 0 to
1704 /proc/sys/net/ipv4/tcp_ecn if you want to talk to these sites. For more info
1705 you could read RFC2481.
1707 tcp_retrans_collapse
1708 --------------------
1710 Bug-to-bug compatibility with some broken printers. On retransmit, try to send
1711 larger packets to work around bugs in certain TCP stacks. Can be turned off by
1714 tcp_keepalive_probes
1715 --------------------
1717 Number of keep alive probes TCP sends out, until it decides that the
1718 connection is broken.
1723 How often TCP sends out keep alive messages, when keep alive is enabled. The
1729 Number of times initial SYNs for a TCP connection attempt will be
1730 retransmitted. Should not be higher than 255. This is only the timeout for
1731 outgoing connections, for incoming connections the number of retransmits is
1732 defined by tcp_retries1.
1737 Enable select acknowledgments after RFC2018.
1742 Enable timestamps as defined in RFC1323.
1747 Enable the strict RFC793 interpretation of the TCP urgent pointer field. The
1748 default is to use the BSD compatible interpretation of the urgent pointer
1749 pointing to the first byte after the urgent data. The RFC793 interpretation is
1750 to have it point to the last byte of urgent data. Enabling this option may
1751 lead to interoperability problems. Disabled by default.
1756 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES. Send out
1757 syncookies when the syn backlog queue of a socket overflows. This is to ward
1758 off the common 'syn flood attack'. Disabled by default.
1760 Note that the concept of a socket backlog is abandoned. This means the peer
1761 may not receive reliable error messages from an over loaded server with
1767 Enable window scaling as defined in RFC1323.
1772 The length of time in seconds it takes to receive a final FIN before the
1773 socket is always closed. This is strictly a violation of the TCP
1774 specification, but required to prevent denial-of-service attacks.
1779 Indicates how many keep alive probes are sent per slow timer run. Should not
1780 be set too high to prevent bursts.
1785 Length of the per socket backlog queue. Since Linux 2.2 the backlog specified
1786 in listen(2) only specifies the length of the backlog queue of already
1787 established sockets. When more connection requests arrive Linux starts to drop
1788 packets. When syncookies are enabled the packets are still answered and the
1789 maximum queue is effectively ignored.
1794 Defines how often an answer to a TCP connection request is retransmitted
1800 Defines how often a TCP packet is retransmitted before giving up.
1802 Interface specific settings
1803 ---------------------------
1805 In the directory /proc/sys/net/ipv4/conf you'll find one subdirectory for each
1806 interface the system knows about and one directory calls all. Changes in the
1807 all subdirectory affect all interfaces, whereas changes in the other
1808 subdirectories affect only one interface. All directories have the same
1814 This switch decides if the kernel accepts ICMP redirect messages or not. The
1815 default is 'yes' if the kernel is configured for a regular host and 'no' for a
1816 router configuration.
1821 Should source routed packages be accepted or declined. The default is
1822 dependent on the kernel configuration. It's 'yes' for routers and 'no' for
1828 Accept packets with source address 0.b.c.d with destinations not to this host
1829 as local ones. It is supposed that a BOOTP relay daemon will catch and forward
1832 The default is 0, since this feature is not implemented yet (kernel version
1838 Enable or disable IP forwarding on this interface.
1843 Log packets with source addresses with no known route to kernel log.
1848 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE and a
1849 multicast routing daemon is required.
1854 Does (1) or does not (0) perform proxy ARP.
1859 Integer value determines if a source validation should be made. 1 means yes, 0
1860 means no. Disabled by default, but local/broadcast address spoofing is always
1863 If you set this to 1 on a router that is the only connection for a network to
1864 the net, it will prevent spoofing attacks against your internal networks
1865 (external addresses can still be spoofed), without the need for additional
1871 Accept ICMP redirect messages only for gateways, listed in default gateway
1872 list. Enabled by default.
1877 If it is not set the kernel does not assume that different subnets on this
1878 device can communicate directly. Default setting is 'yes'.
1883 Determines whether to send ICMP redirects to other hosts.
1888 The directory /proc/sys/net/ipv4/route contains several file to control
1891 error_burst and error_cost
1892 --------------------------
1894 These parameters are used to limit how many ICMP destination unreachable to
1895 send from the host in question. ICMP destination unreachable messages are
1896 sent when we cannot reach the next hop while trying to transmit a packet.
1897 It will also print some error messages to kernel logs if someone is ignoring
1898 our ICMP redirects. The higher the error_cost factor is, the fewer
1899 destination unreachable and error messages will be let through. Error_burst
1900 controls when destination unreachable messages and error messages will be
1901 dropped. The default settings limit warning messages to five every second.
1906 Writing to this file results in a flush of the routing cache.
1908 gc_elasticity, gc_interval, gc_min_interval_ms, gc_timeout, gc_thresh
1909 ---------------------------------------------------------------------
1911 Values to control the frequency and behavior of the garbage collection
1912 algorithm for the routing cache. gc_min_interval is deprecated and replaced
1913 by gc_min_interval_ms.
1919 Maximum size of the routing cache. Old entries will be purged once the cache
1920 reached has this size.
1922 max_delay, min_delay
1923 --------------------
1925 Delays for flushing the routing cache.
1927 redirect_load, redirect_number
1928 ------------------------------
1930 Factors which determine if more ICPM redirects should be sent to a specific
1931 host. No redirects will be sent once the load limit or the maximum number of
1932 redirects has been reached.
1937 Timeout for redirects. After this period redirects will be sent again, even if
1938 this has been stopped, because the load or number limit has been reached.
1940 Network Neighbor handling
1941 -------------------------
1943 Settings about how to handle connections with direct neighbors (nodes attached
1944 to the same link) can be found in the directory /proc/sys/net/ipv4/neigh.
1946 As we saw it in the conf directory, there is a default subdirectory which
1947 holds the default values, and one directory for each interface. The contents
1948 of the directories are identical, with the single exception that the default
1949 settings contain additional options to set garbage collection parameters.
1951 In the interface directories you'll find the following entries:
1953 base_reachable_time, base_reachable_time_ms
1954 -------------------------------------------
1956 A base value used for computing the random reachable time value as specified
1959 Expression of base_reachable_time, which is deprecated, is in seconds.
1960 Expression of base_reachable_time_ms is in milliseconds.
1962 retrans_time, retrans_time_ms
1963 -----------------------------
1965 The time between retransmitted Neighbor Solicitation messages.
1966 Used for address resolution and to determine if a neighbor is
1969 Expression of retrans_time, which is deprecated, is in 1/100 seconds (for
1970 IPv4) or in jiffies (for IPv6).
1971 Expression of retrans_time_ms is in milliseconds.
1976 Maximum queue length for a pending arp request - the number of packets which
1977 are accepted from other layers while the ARP address is still resolved.
1982 Maximum for random delay of answers to neighbor solicitation messages in
1983 jiffies (1/100 sec). Not yet implemented (Linux does not have anycast support
1989 Maximum number of retries for unicast solicitation.
1994 Maximum number of retries for multicast solicitation.
1996 delay_first_probe_time
1997 ----------------------
1999 Delay for the first time probe if the neighbor is reachable. (see
2005 An ARP/neighbor entry is only replaced with a new one if the old is at least
2006 locktime old. This prevents ARP cache thrashing.
2011 Maximum time (real time is random [0..proxytime]) before answering to an ARP
2012 request for which we have an proxy ARP entry. In some cases, this is used to
2013 prevent network flooding.
2018 Maximum queue length of the delayed proxy arp timer. (see proxy_delay).
2023 Determines the number of requests to send to the user level ARP daemon. Use 0
2029 Determines how often to check for stale ARP entries. After an ARP entry is
2030 stale it will be resolved again (which is useful when an IP address migrates
2031 to another machine). When ucast_solicit is greater than 0 it first tries to
2032 send an ARP packet directly to the known host When that fails and
2033 mcast_solicit is greater than 0, an ARP request is broadcasted.
2038 The /proc/sys/net/appletalk directory holds the Appletalk configuration data
2039 when Appletalk is loaded. The configurable parameters are:
2044 The amount of time we keep an ARP entry before expiring it. Used to age out
2050 The amount of time we will spend trying to resolve an Appletalk address.
2052 aarp-retransmit-limit
2053 ---------------------
2055 The number of times we will retransmit a query before giving up.
2060 Controls the rate at which expires are checked.
2062 The directory /proc/net/appletalk holds the list of active Appletalk sockets
2065 The fields indicate the DDP type, the local address (in network:node format)
2066 the remote address, the size of the transmit pending queue, the size of the
2067 received queue (bytes waiting for applications to read) the state and the uid
2070 /proc/net/atalk_iface lists all the interfaces configured for appletalk.It
2071 shows the name of the interface, its Appletalk address, the network range on
2072 that address (or network number for phase 1 networks), and the status of the
2075 /proc/net/atalk_route lists each known network route. It lists the target
2076 (network) that the route leads to, the router (may be directly connected), the
2077 route flags, and the device the route is using.
2082 The IPX protocol has no tunable values in proc/sys/net.
2084 The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
2085 socket giving the local and remote addresses in Novell format (that is
2086 network:node:port). In accordance with the strange Novell tradition,
2087 everything but the port is in hex. Not_Connected is displayed for sockets that
2088 are not tied to a specific remote address. The Tx and Rx queue sizes indicate
2089 the number of bytes pending for transmission and reception. The state
2090 indicates the state the socket is in and the uid is the owning uid of the
2093 The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
2094 it gives the network number, the node number, and indicates if the network is
2095 the primary network. It also indicates which device it is bound to (or
2096 Internal for internal networks) and the Frame Type if appropriate. Linux
2097 supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
2100 The /proc/net/ipx_route table holds a list of IPX routes. For each route it
2101 gives the destination network, the router node (or Directly) and the network
2102 address of the router (or Connected) for internal networks.
2104 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
2105 ----------------------------------------------------------
2107 The "mqueue" filesystem provides the necessary kernel features to enable the
2108 creation of a user space library that implements the POSIX message queues
2109 API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System
2110 Interfaces specification.)
2112 The "mqueue" filesystem contains values for determining/setting the amount of
2113 resources used by the file system.
2115 /proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the
2116 maximum number of message queues allowed on the system.
2118 /proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the
2119 maximum number of messages in a queue value. In fact it is the limiting value
2120 for another (user) limit which is set in mq_open invocation. This attribute of
2121 a queue must be less or equal then msg_max.
2123 /proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the
2124 maximum message size value (it is every message queue's attribute set during
2127 2.12 /proc/<pid>/oom_adj - Adjust the oom-killer score
2128 ------------------------------------------------------
2130 This file can be used to adjust the score used to select which processes
2131 should be killed in an out-of-memory situation. Giving it a high score will
2132 increase the likelihood of this process being killed by the oom-killer. Valid
2133 values are in the range -16 to +15, plus the special value -17, which disables
2134 oom-killing altogether for this process.
2136 2.13 /proc/<pid>/oom_score - Display current oom-killer score
2137 -------------------------------------------------------------
2139 ------------------------------------------------------------------------------
2140 This file can be used to check the current score used by the oom-killer is for
2141 any given <pid>. Use it together with /proc/<pid>/oom_adj to tune which
2142 process should be killed in an out-of-memory situation.
2144 ------------------------------------------------------------------------------
2146 ------------------------------------------------------------------------------
2147 Certain aspects of kernel behavior can be modified at runtime, without the
2148 need to recompile the kernel, or even to reboot the system. The files in the
2149 /proc/sys tree can not only be read, but also modified. You can use the echo
2150 command to write value into these files, thereby changing the default settings
2152 ------------------------------------------------------------------------------
2154 2.14 /proc/<pid>/io - Display the IO accounting fields
2155 -------------------------------------------------------
2157 This file contains IO statistics for each running process
2162 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
2165 test:/tmp # cat /proc/3828/io
2171 write_bytes: 323932160
2172 cancelled_write_bytes: 0
2181 I/O counter: chars read
2182 The number of bytes which this task has caused to be read from storage. This
2183 is simply the sum of bytes which this process passed to read() and pread().
2184 It includes things like tty IO and it is unaffected by whether or not actual
2185 physical disk IO was required (the read might have been satisfied from
2192 I/O counter: chars written
2193 The number of bytes which this task has caused, or shall cause to be written
2194 to disk. Similar caveats apply here as with rchar.
2200 I/O counter: read syscalls
2201 Attempt to count the number of read I/O operations, i.e. syscalls like read()
2208 I/O counter: write syscalls
2209 Attempt to count the number of write I/O operations, i.e. syscalls like
2210 write() and pwrite().
2216 I/O counter: bytes read
2217 Attempt to count the number of bytes which this process really did cause to
2218 be fetched from the storage layer. Done at the submit_bio() level, so it is
2219 accurate for block-backed filesystems. <please add status regarding NFS and
2220 CIFS at a later time>
2226 I/O counter: bytes written
2227 Attempt to count the number of bytes which this process caused to be sent to
2228 the storage layer. This is done at page-dirtying time.
2231 cancelled_write_bytes
2232 ---------------------
2234 The big inaccuracy here is truncate. If a process writes 1MB to a file and
2235 then deletes the file, it will in fact perform no writeout. But it will have
2236 been accounted as having caused 1MB of write.
2237 In other words: The number of bytes which this process caused to not happen,
2238 by truncating pagecache. A task can cause "negative" IO too. If this task
2239 truncates some dirty pagecache, some IO which another task has been accounted
2240 for (in it's write_bytes) will not be happening. We _could_ just subtract that
2241 from the truncating task's write_bytes, but there is information loss in doing
2248 At its current implementation state, this is a bit racy on 32-bit machines: if
2249 process A reads process B's /proc/pid/io while process B is updating one of
2250 those 64-bit counters, process A could see an intermediate result.
2253 More information about this can be found within the taskstats documentation in
2254 Documentation/accounting.
2256 2.15 /proc/<pid>/coredump_filter - Core dump filtering settings
2257 ---------------------------------------------------------------
2258 When a process is dumped, all anonymous memory is written to a core file as
2259 long as the size of the core file isn't limited. But sometimes we don't want
2260 to dump some memory segments, for example, huge shared memory. Conversely,
2261 sometimes we want to save file-backed memory segments into a core file, not
2262 only the individual files.
2264 /proc/<pid>/coredump_filter allows you to customize which memory segments
2265 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
2266 of memory types. If a bit of the bitmask is set, memory segments of the
2267 corresponding memory type are dumped, otherwise they are not dumped.
2269 The following 4 memory types are supported:
2270 - (bit 0) anonymous private memory
2271 - (bit 1) anonymous shared memory
2272 - (bit 2) file-backed private memory
2273 - (bit 3) file-backed shared memory
2275 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
2276 are always dumped regardless of the bitmask status.
2278 Default value of coredump_filter is 0x3; this means all anonymous memory
2279 segments are dumped.
2281 If you don't want to dump all shared memory segments attached to pid 1234,
2282 write 1 to the process's proc file.
2284 $ echo 0x1 > /proc/1234/coredump_filter
2286 When a new process is created, the process inherits the bitmask status from its
2287 parent. It is useful to set up coredump_filter before the program runs.
2290 $ echo 0x7 > /proc/self/coredump_filter
2293 ------------------------------------------------------------------------------
projects / deliverable / linux.git / blob