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
46 2.16 /proc/<pid>/mountinfo - Information about mounts
47 2.17 /proc/sys/fs/epoll - Configuration options for the epoll interface
49 ------------------------------------------------------------------------------
51 ------------------------------------------------------------------------------
53 0.1 Introduction/Credits
54 ------------------------
56 This documentation is part of a soon (or so we hope) to be released book on
57 the SuSE Linux distribution. As there is no complete documentation for the
58 /proc file system and we've used many freely available sources to write these
59 chapters, it seems only fair to give the work back to the Linux community.
60 This work is based on the 2.2.* kernel version and the upcoming 2.4.*. I'm
61 afraid it's still far from complete, but we hope it will be useful. As far as
62 we know, it is the first 'all-in-one' document about the /proc file system. It
63 is focused on the Intel x86 hardware, so if you are looking for PPC, ARM,
64 SPARC, AXP, etc., features, you probably won't find what you are looking for.
65 It also only covers IPv4 networking, not IPv6 nor other protocols - sorry. But
66 additions and patches are welcome and will be added to this document if you
69 We'd like to thank Alan Cox, Rik van Riel, and Alexey Kuznetsov and a lot of
70 other people for help compiling this documentation. We'd also like to extend a
71 special thank you to Andi Kleen for documentation, which we relied on heavily
72 to create this document, as well as the additional information he provided.
73 Thanks to everybody else who contributed source or docs to the Linux kernel
74 and helped create a great piece of software... :)
76 If you have any comments, corrections or additions, please don't hesitate to
77 contact Bodo Bauer at bb@ricochet.net. We'll be happy to add them to this
80 The latest version of this document is available online at
81 http://skaro.nightcrawler.com/~bb/Docs/Proc as HTML version.
83 If the above direction does not works for you, ypu could try the kernel
84 mailing list at linux-kernel@vger.kernel.org and/or try to reach me at
85 comandante@zaralinux.com.
90 We don't guarantee the correctness of this document, and if you come to us
91 complaining about how you screwed up your system because of incorrect
92 documentation, we won't feel responsible...
94 ------------------------------------------------------------------------------
95 CHAPTER 1: COLLECTING SYSTEM INFORMATION
96 ------------------------------------------------------------------------------
98 ------------------------------------------------------------------------------
100 ------------------------------------------------------------------------------
101 * Investigating the properties of the pseudo file system /proc and its
102 ability to provide information on the running Linux system
103 * Examining /proc's structure
104 * Uncovering various information about the kernel and the processes running
106 ------------------------------------------------------------------------------
109 The proc file system acts as an interface to internal data structures in the
110 kernel. It can be used to obtain information about the system and to change
111 certain kernel parameters at runtime (sysctl).
113 First, we'll take a look at the read-only parts of /proc. In Chapter 2, we
114 show you how you can use /proc/sys to change settings.
116 1.1 Process-Specific Subdirectories
117 -----------------------------------
119 The directory /proc contains (among other things) one subdirectory for each
120 process running on the system, which is named after the process ID (PID).
122 The link self points to the process reading the file system. Each process
123 subdirectory has the entries listed in Table 1-1.
126 Table 1-1: Process specific entries in /proc
127 ..............................................................................
129 clear_refs Clears page referenced bits shown in smaps output
130 cmdline Command line arguments
131 cpu Current and last cpu in which it was executed (2.4)(smp)
132 cwd Link to the current working directory
133 environ Values of environment variables
134 exe Link to the executable of this process
135 fd Directory, which contains all file descriptors
136 maps Memory maps to executables and library files (2.4)
137 mem Memory held by this process
138 root Link to the root directory of this process
140 statm Process memory status information
141 status Process status in human readable form
142 wchan If CONFIG_KALLSYMS is set, a pre-decoded wchan
143 stack Report full stack trace, enable via CONFIG_STACKTRACE
144 smaps Extension based on maps, the rss size for each mapped file
145 ..............................................................................
147 For example, to get the status information of a process, all you have to do is
148 read the file /proc/PID/status:
150 >cat /proc/self/status
166 SigPnd: 0000000000000000
167 SigBlk: 0000000000000000
168 SigIgn: 0000000000000000
169 SigCgt: 0000000000000000
170 CapInh: 00000000fffffeff
171 CapPrm: 0000000000000000
172 CapEff: 0000000000000000
175 This shows you nearly the same information you would get if you viewed it with
176 the ps command. In fact, ps uses the proc file system to obtain its
177 information. The statm file contains more detailed information about the
178 process memory usage. Its seven fields are explained in Table 1-2. The stat
179 file contains details information about the process itself. Its fields are
180 explained in Table 1-3.
183 Table 1-2: Contents of the statm files (as of 2.6.8-rc3)
184 ..............................................................................
186 size total program size (pages) (same as VmSize in status)
187 resident size of memory portions (pages) (same as VmRSS in status)
188 shared number of pages that are shared (i.e. backed by a file)
189 trs number of pages that are 'code' (not including libs; broken,
190 includes data segment)
191 lrs number of pages of library (always 0 on 2.6)
192 drs number of pages of data/stack (including libs; broken,
193 includes library text)
194 dt number of dirty pages (always 0 on 2.6)
195 ..............................................................................
198 Table 1-3: Contents of the stat files (as of 2.6.22-rc3)
199 ..............................................................................
202 tcomm filename of the executable
203 state state (R is running, S is sleeping, D is sleeping in an
204 uninterruptible wait, Z is zombie, T is traced or stopped)
205 ppid process id of the parent process
206 pgrp pgrp of the process
208 tty_nr tty the process uses
209 tty_pgrp pgrp of the tty
211 min_flt number of minor faults
212 cmin_flt number of minor faults with child's
213 maj_flt number of major faults
214 cmaj_flt number of major faults with child's
215 utime user mode jiffies
216 stime kernel mode jiffies
217 cutime user mode jiffies with child's
218 cstime kernel mode jiffies with child's
219 priority priority level
221 num_threads number of threads
222 it_real_value (obsolete, always 0)
223 start_time time the process started after system boot
224 vsize virtual memory size
225 rss resident set memory size
226 rsslim current limit in bytes on the rss
227 start_code address above which program text can run
228 end_code address below which program text can run
229 start_stack address of the start of the stack
230 esp current value of ESP
231 eip current value of EIP
232 pending bitmap of pending signals (obsolete)
233 blocked bitmap of blocked signals (obsolete)
234 sigign bitmap of ignored signals (obsolete)
235 sigcatch bitmap of catched signals (obsolete)
236 wchan address where process went to sleep
239 exit_signal signal to send to parent thread on exit
240 task_cpu which CPU the task is scheduled on
241 rt_priority realtime priority
242 policy scheduling policy (man sched_setscheduler)
243 blkio_ticks time spent waiting for block IO
244 ..............................................................................
250 Similar to the process entries, the kernel data files give information about
251 the running kernel. The files used to obtain this information are contained in
252 /proc and are listed in Table 1-4. Not all of these will be present in your
253 system. It depends on the kernel configuration and the loaded modules, which
254 files are there, and which are missing.
256 Table 1-4: Kernel info in /proc
257 ..............................................................................
259 apm Advanced power management info
260 buddyinfo Kernel memory allocator information (see text) (2.5)
261 bus Directory containing bus specific information
262 cmdline Kernel command line
263 cpuinfo Info about the CPU
264 devices Available devices (block and character)
265 dma Used DMS channels
266 filesystems Supported filesystems
267 driver Various drivers grouped here, currently rtc (2.4)
268 execdomains Execdomains, related to security (2.4)
269 fb Frame Buffer devices (2.4)
270 fs File system parameters, currently nfs/exports (2.4)
271 ide Directory containing info about the IDE subsystem
272 interrupts Interrupt usage
273 iomem Memory map (2.4)
274 ioports I/O port usage
275 irq Masks for irq to cpu affinity (2.4)(smp?)
276 isapnp ISA PnP (Plug&Play) Info (2.4)
277 kcore Kernel core image (can be ELF or A.OUT(deprecated in 2.4))
279 ksyms Kernel symbol table
280 loadavg Load average of last 1, 5 & 15 minutes
284 modules List of loaded modules
285 mounts Mounted filesystems
286 net Networking info (see text)
287 partitions Table of partitions known to the system
288 pci Deprecated info of PCI bus (new way -> /proc/bus/pci/,
289 decoupled by lspci (2.4)
291 scsi SCSI info (see text)
292 slabinfo Slab pool info
293 stat Overall statistics
294 swaps Swap space utilization
296 sysvipc Info of SysVIPC Resources (msg, sem, shm) (2.4)
297 tty Info of tty drivers
299 version Kernel version
300 video bttv info of video resources (2.4)
301 vmallocinfo Show vmalloced areas
302 ..............................................................................
304 You can, for example, check which interrupts are currently in use and what
305 they are used for by looking in the file /proc/interrupts:
307 > cat /proc/interrupts
309 0: 8728810 XT-PIC timer
310 1: 895 XT-PIC keyboard
312 3: 531695 XT-PIC aha152x
313 4: 2014133 XT-PIC serial
314 5: 44401 XT-PIC pcnet_cs
317 12: 182918 XT-PIC PS/2 Mouse
319 14: 1232265 XT-PIC ide0
323 In 2.4.* a couple of lines where added to this file LOC & ERR (this time is the
324 output of a SMP machine):
326 > cat /proc/interrupts
329 0: 1243498 1214548 IO-APIC-edge timer
330 1: 8949 8958 IO-APIC-edge keyboard
331 2: 0 0 XT-PIC cascade
332 5: 11286 10161 IO-APIC-edge soundblaster
333 8: 1 0 IO-APIC-edge rtc
334 9: 27422 27407 IO-APIC-edge 3c503
335 12: 113645 113873 IO-APIC-edge PS/2 Mouse
337 14: 22491 24012 IO-APIC-edge ide0
338 15: 2183 2415 IO-APIC-edge ide1
339 17: 30564 30414 IO-APIC-level eth0
340 18: 177 164 IO-APIC-level bttv
345 NMI is incremented in this case because every timer interrupt generates a NMI
346 (Non Maskable Interrupt) which is used by the NMI Watchdog to detect lockups.
348 LOC is the local interrupt counter of the internal APIC of every CPU.
350 ERR is incremented in the case of errors in the IO-APIC bus (the bus that
351 connects the CPUs in a SMP system. This means that an error has been detected,
352 the IO-APIC automatically retry the transmission, so it should not be a big
353 problem, but you should read the SMP-FAQ.
355 In 2.6.2* /proc/interrupts was expanded again. This time the goal was for
356 /proc/interrupts to display every IRQ vector in use by the system, not
357 just those considered 'most important'. The new vectors are:
359 THR -- interrupt raised when a machine check threshold counter
360 (typically counting ECC corrected errors of memory or cache) exceeds
361 a configurable threshold. Only available on some systems.
363 TRM -- a thermal event interrupt occurs when a temperature threshold
364 has been exceeded for the CPU. This interrupt may also be generated
365 when the temperature drops back to normal.
367 SPU -- a spurious interrupt is some interrupt that was raised then lowered
368 by some IO device before it could be fully processed by the APIC. Hence
369 the APIC sees the interrupt but does not know what device it came from.
370 For this case the APIC will generate the interrupt with a IRQ vector
371 of 0xff. This might also be generated by chipset bugs.
373 RES, CAL, TLB -- rescheduling, call and TLB flush interrupts are
374 sent from one CPU to another per the needs of the OS. Typically,
375 their statistics are used by kernel developers and interested users to
376 determine the occurance of interrupt of the given type.
378 The above IRQ vectors are displayed only when relevent. For example,
379 the threshold vector does not exist on x86_64 platforms. Others are
380 suppressed when the system is a uniprocessor. As of this writing, only
381 i386 and x86_64 platforms support the new IRQ vector displays.
383 Of some interest is the introduction of the /proc/irq directory to 2.4.
384 It could be used to set IRQ to CPU affinity, this means that you can "hook" an
385 IRQ to only one CPU, or to exclude a CPU of handling IRQs. The contents of the
386 irq subdir is one subdir for each IRQ, and two files; default_smp_affinity and
391 0 10 12 14 16 18 2 4 6 8 prof_cpu_mask
392 1 11 13 15 17 19 3 5 7 9 default_smp_affinity
396 smp_affinity is a bitmask, in which you can specify which CPUs can handle the
397 IRQ, you can set it by doing:
399 > echo 1 > /proc/irq/10/smp_affinity
401 This means that only the first CPU will handle the IRQ, but you can also echo
402 5 which means that only the first and fourth CPU can handle the IRQ.
404 The contents of each smp_affinity file is the same by default:
406 > cat /proc/irq/0/smp_affinity
409 The default_smp_affinity mask applies to all non-active IRQs, which are the
410 IRQs which have not yet been allocated/activated, and hence which lack a
411 /proc/irq/[0-9]* directory.
413 prof_cpu_mask specifies which CPUs are to be profiled by the system wide
414 profiler. Default value is ffffffff (all cpus).
416 The way IRQs are routed is handled by the IO-APIC, and it's Round Robin
417 between all the CPUs which are allowed to handle it. As usual the kernel has
418 more info than you and does a better job than you, so the defaults are the
419 best choice for almost everyone.
421 There are three more important subdirectories in /proc: net, scsi, and sys.
422 The general rule is that the contents, or even the existence of these
423 directories, depend on your kernel configuration. If SCSI is not enabled, the
424 directory scsi may not exist. The same is true with the net, which is there
425 only when networking support is present in the running kernel.
427 The slabinfo file gives information about memory usage at the slab level.
428 Linux uses slab pools for memory management above page level in version 2.2.
429 Commonly used objects have their own slab pool (such as network buffers,
430 directory cache, and so on).
432 ..............................................................................
434 > cat /proc/buddyinfo
436 Node 0, zone DMA 0 4 5 4 4 3 ...
437 Node 0, zone Normal 1 0 0 1 101 8 ...
438 Node 0, zone HighMem 2 0 0 1 1 0 ...
440 Memory fragmentation is a problem under some workloads, and buddyinfo is a
441 useful tool for helping diagnose these problems. Buddyinfo will give you a
442 clue as to how big an area you can safely allocate, or why a previous
445 Each column represents the number of pages of a certain order which are
446 available. In this case, there are 0 chunks of 2^0*PAGE_SIZE available in
447 ZONE_DMA, 4 chunks of 2^1*PAGE_SIZE in ZONE_DMA, 101 chunks of 2^4*PAGE_SIZE
448 available in ZONE_NORMAL, etc...
450 ..............................................................................
454 Provides information about distribution and utilization of memory. This
455 varies by architecture and compile options. The following is from a
456 16GB PIII, which has highmem enabled. You may not have all of these fields.
461 MemTotal: 16344972 kB
468 HighTotal: 15597528 kB
469 HighFree: 13629632 kB
479 SReclaimable: 159856 kB
480 SUnreclaim: 124508 kB
485 CommitLimit: 7669796 kB
486 Committed_AS: 100056 kB
487 VmallocTotal: 112216 kB
489 VmallocChunk: 111088 kB
491 MemTotal: Total usable ram (i.e. physical ram minus a few reserved
492 bits and the kernel binary code)
493 MemFree: The sum of LowFree+HighFree
494 Buffers: Relatively temporary storage for raw disk blocks
495 shouldn't get tremendously large (20MB or so)
496 Cached: in-memory cache for files read from the disk (the
497 pagecache). Doesn't include SwapCached
498 SwapCached: Memory that once was swapped out, is swapped back in but
499 still also is in the swapfile (if memory is needed it
500 doesn't need to be swapped out AGAIN because it is already
501 in the swapfile. This saves I/O)
502 Active: Memory that has been used more recently and usually not
503 reclaimed unless absolutely necessary.
504 Inactive: Memory which has been less recently used. It is more
505 eligible to be reclaimed for other purposes
507 HighFree: Highmem is all memory above ~860MB of physical memory
508 Highmem areas are for use by userspace programs, or
509 for the pagecache. The kernel must use tricks to access
510 this memory, making it slower to access than lowmem.
512 LowFree: Lowmem is memory which can be used for everything that
513 highmem can be used for, but it is also available for the
514 kernel's use for its own data structures. Among many
515 other things, it is where everything from the Slab is
516 allocated. Bad things happen when you're out of lowmem.
517 SwapTotal: total amount of swap space available
518 SwapFree: Memory which has been evicted from RAM, and is temporarily
520 Dirty: Memory which is waiting to get written back to the disk
521 Writeback: Memory which is actively being written back to the disk
522 AnonPages: Non-file backed pages mapped into userspace page tables
523 Mapped: files which have been mmaped, such as libraries
524 Slab: in-kernel data structures cache
525 SReclaimable: Part of Slab, that might be reclaimed, such as caches
526 SUnreclaim: Part of Slab, that cannot be reclaimed on memory pressure
527 PageTables: amount of memory dedicated to the lowest level of page
529 NFS_Unstable: NFS pages sent to the server, but not yet committed to stable
531 Bounce: Memory used for block device "bounce buffers"
532 WritebackTmp: Memory used by FUSE for temporary writeback buffers
533 CommitLimit: Based on the overcommit ratio ('vm.overcommit_ratio'),
534 this is the total amount of memory currently available to
535 be allocated on the system. This limit is only adhered to
536 if strict overcommit accounting is enabled (mode 2 in
537 'vm.overcommit_memory').
538 The CommitLimit is calculated with the following formula:
539 CommitLimit = ('vm.overcommit_ratio' * Physical RAM) + Swap
540 For example, on a system with 1G of physical RAM and 7G
541 of swap with a `vm.overcommit_ratio` of 30 it would
542 yield a CommitLimit of 7.3G.
543 For more details, see the memory overcommit documentation
544 in vm/overcommit-accounting.
545 Committed_AS: The amount of memory presently allocated on the system.
546 The committed memory is a sum of all of the memory which
547 has been allocated by processes, even if it has not been
548 "used" by them as of yet. A process which malloc()'s 1G
549 of memory, but only touches 300M of it will only show up
550 as using 300M of memory even if it has the address space
551 allocated for the entire 1G. This 1G is memory which has
552 been "committed" to by the VM and can be used at any time
553 by the allocating application. With strict overcommit
554 enabled on the system (mode 2 in 'vm.overcommit_memory'),
555 allocations which would exceed the CommitLimit (detailed
556 above) will not be permitted. This is useful if one needs
557 to guarantee that processes will not fail due to lack of
558 memory once that memory has been successfully allocated.
559 VmallocTotal: total size of vmalloc memory area
560 VmallocUsed: amount of vmalloc area which is used
561 VmallocChunk: largest contigious block of vmalloc area which is free
563 ..............................................................................
567 Provides information about vmalloced/vmaped areas. One line per area,
568 containing the virtual address range of the area, size in bytes,
569 caller information of the creator, and optional information depending
570 on the kind of area :
572 pages=nr number of pages
573 phys=addr if a physical address was specified
574 ioremap I/O mapping (ioremap() and friends)
575 vmalloc vmalloc() area
578 vpages buffer for pages pointers was vmalloced (huge area)
579 N<node>=nr (Only on NUMA kernels)
580 Number of pages allocated on memory node <node>
582 > cat /proc/vmallocinfo
583 0xffffc20000000000-0xffffc20000201000 2101248 alloc_large_system_hash+0x204 ...
584 /0x2c0 pages=512 vmalloc N0=128 N1=128 N2=128 N3=128
585 0xffffc20000201000-0xffffc20000302000 1052672 alloc_large_system_hash+0x204 ...
586 /0x2c0 pages=256 vmalloc N0=64 N1=64 N2=64 N3=64
587 0xffffc20000302000-0xffffc20000304000 8192 acpi_tb_verify_table+0x21/0x4f...
588 phys=7fee8000 ioremap
589 0xffffc20000304000-0xffffc20000307000 12288 acpi_tb_verify_table+0x21/0x4f...
590 phys=7fee7000 ioremap
591 0xffffc2000031d000-0xffffc2000031f000 8192 init_vdso_vars+0x112/0x210
592 0xffffc2000031f000-0xffffc2000032b000 49152 cramfs_uncompress_init+0x2e ...
593 /0x80 pages=11 vmalloc N0=3 N1=3 N2=2 N3=3
594 0xffffc2000033a000-0xffffc2000033d000 12288 sys_swapon+0x640/0xac0 ...
596 0xffffc20000347000-0xffffc2000034c000 20480 xt_alloc_table_info+0xfe ...
597 /0x130 [x_tables] pages=4 vmalloc N0=4
598 0xffffffffa0000000-0xffffffffa000f000 61440 sys_init_module+0xc27/0x1d00 ...
599 pages=14 vmalloc N2=14
600 0xffffffffa000f000-0xffffffffa0014000 20480 sys_init_module+0xc27/0x1d00 ...
602 0xffffffffa0014000-0xffffffffa0017000 12288 sys_init_module+0xc27/0x1d00 ...
604 0xffffffffa0017000-0xffffffffa0022000 45056 sys_init_module+0xc27/0x1d00 ...
605 pages=10 vmalloc N0=10
607 1.3 IDE devices in /proc/ide
608 ----------------------------
610 The subdirectory /proc/ide contains information about all IDE devices of which
611 the kernel is aware. There is one subdirectory for each IDE controller, the
612 file drivers and a link for each IDE device, pointing to the device directory
613 in the controller specific subtree.
615 The file drivers contains general information about the drivers used for the
618 > cat /proc/ide/drivers
619 ide-cdrom version 4.53
620 ide-disk version 1.08
622 More detailed information can be found in the controller specific
623 subdirectories. These are named ide0, ide1 and so on. Each of these
624 directories contains the files shown in table 1-5.
627 Table 1-5: IDE controller info in /proc/ide/ide?
628 ..............................................................................
630 channel IDE channel (0 or 1)
631 config Configuration (only for PCI/IDE bridge)
633 model Type/Chipset of IDE controller
634 ..............................................................................
636 Each device connected to a controller has a separate subdirectory in the
637 controllers directory. The files listed in table 1-6 are contained in these
641 Table 1-6: IDE device information
642 ..............................................................................
645 capacity Capacity of the medium (in 512Byte blocks)
646 driver driver and version
647 geometry physical and logical geometry
648 identify device identify block
650 model device identifier
651 settings device setup
652 smart_thresholds IDE disk management thresholds
653 smart_values IDE disk management values
654 ..............................................................................
656 The most interesting file is settings. This file contains a nice overview of
657 the drive parameters:
659 # cat /proc/ide/ide0/hda/settings
660 name value min max mode
661 ---- ----- --- --- ----
662 bios_cyl 526 0 65535 rw
663 bios_head 255 0 255 rw
665 breada_readahead 4 0 127 rw
667 file_readahead 72 0 2097151 rw
669 keepsettings 0 0 1 rw
670 max_kb_per_request 122 1 127 rw
674 pio_mode write-only 0 255 w
680 1.4 Networking info in /proc/net
681 --------------------------------
683 The subdirectory /proc/net follows the usual pattern. Table 1-6 shows the
684 additional values you get for IP version 6 if you configure the kernel to
685 support this. Table 1-7 lists the files and their meaning.
688 Table 1-6: IPv6 info in /proc/net
689 ..............................................................................
691 udp6 UDP sockets (IPv6)
692 tcp6 TCP sockets (IPv6)
693 raw6 Raw device statistics (IPv6)
694 igmp6 IP multicast addresses, which this host joined (IPv6)
695 if_inet6 List of IPv6 interface addresses
696 ipv6_route Kernel routing table for IPv6
697 rt6_stats Global IPv6 routing tables statistics
698 sockstat6 Socket statistics (IPv6)
699 snmp6 Snmp data (IPv6)
700 ..............................................................................
703 Table 1-7: Network info in /proc/net
704 ..............................................................................
707 dev network devices with statistics
708 dev_mcast the Layer2 multicast groups a device is listening too
709 (interface index, label, number of references, number of bound
711 dev_stat network device status
712 ip_fwchains Firewall chain linkage
713 ip_fwnames Firewall chain names
714 ip_masq Directory containing the masquerading tables
715 ip_masquerade Major masquerading table
716 netstat Network statistics
717 raw raw device statistics
718 route Kernel routing table
719 rpc Directory containing rpc info
720 rt_cache Routing cache
722 sockstat Socket statistics
724 tr_rif Token ring RIF routing table
726 unix UNIX domain sockets
727 wireless Wireless interface data (Wavelan etc)
728 igmp IP multicast addresses, which this host joined
729 psched Global packet scheduler parameters.
730 netlink List of PF_NETLINK sockets
731 ip_mr_vifs List of multicast virtual interfaces
732 ip_mr_cache List of multicast routing cache
733 ..............................................................................
735 You can use this information to see which network devices are available in
736 your system and how much traffic was routed over those devices:
740 face |bytes packets errs drop fifo frame compressed multicast|[...
741 lo: 908188 5596 0 0 0 0 0 0 [...
742 ppp0:15475140 20721 410 0 0 410 0 0 [...
743 eth0: 614530 7085 0 0 0 0 0 1 [...
746 ...] bytes packets errs drop fifo colls carrier compressed
747 ...] 908188 5596 0 0 0 0 0 0
748 ...] 1375103 17405 0 0 0 0 0 0
749 ...] 1703981 5535 0 0 0 3 0 0
751 In addition, each Channel Bond interface has it's own directory. For
752 example, the bond0 device will have a directory called /proc/net/bond0/.
753 It will contain information that is specific to that bond, such as the
754 current slaves of the bond, the link status of the slaves, and how
755 many times the slaves link has failed.
760 If you have a SCSI host adapter in your system, you'll find a subdirectory
761 named after the driver for this adapter in /proc/scsi. You'll also see a list
762 of all recognized SCSI devices in /proc/scsi:
766 Host: scsi0 Channel: 00 Id: 00 Lun: 00
767 Vendor: IBM Model: DGHS09U Rev: 03E0
768 Type: Direct-Access ANSI SCSI revision: 03
769 Host: scsi0 Channel: 00 Id: 06 Lun: 00
770 Vendor: PIONEER Model: CD-ROM DR-U06S Rev: 1.04
771 Type: CD-ROM ANSI SCSI revision: 02
774 The directory named after the driver has one file for each adapter found in
775 the system. These files contain information about the controller, including
776 the used IRQ and the IO address range. The amount of information shown is
777 dependent on the adapter you use. The example shows the output for an Adaptec
778 AHA-2940 SCSI adapter:
780 > cat /proc/scsi/aic7xxx/0
782 Adaptec AIC7xxx driver version: 5.1.19/3.2.4
784 TCQ Enabled By Default : Disabled
785 AIC7XXX_PROC_STATS : Disabled
786 AIC7XXX_RESET_DELAY : 5
787 Adapter Configuration:
788 SCSI Adapter: Adaptec AHA-294X Ultra SCSI host adapter
789 Ultra Wide Controller
790 PCI MMAPed I/O Base: 0xeb001000
791 Adapter SEEPROM Config: SEEPROM found and used.
792 Adaptec SCSI BIOS: Enabled
794 SCBs: Active 0, Max Active 2,
795 Allocated 15, HW 16, Page 255
797 BIOS Control Word: 0x18b6
798 Adapter Control Word: 0x005b
799 Extended Translation: Enabled
800 Disconnect Enable Flags: 0xffff
801 Ultra Enable Flags: 0x0001
802 Tag Queue Enable Flags: 0x0000
803 Ordered Queue Tag Flags: 0x0000
804 Default Tag Queue Depth: 8
805 Tagged Queue By Device array for aic7xxx host instance 0:
806 {255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255}
807 Actual queue depth per device for aic7xxx host instance 0:
808 {1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}
811 Device using Wide/Sync transfers at 40.0 MByte/sec, offset 8
812 Transinfo settings: current(12/8/1/0), goal(12/8/1/0), user(12/15/1/0)
813 Total transfers 160151 (74577 reads and 85574 writes)
815 Device using Narrow/Sync transfers at 5.0 MByte/sec, offset 15
816 Transinfo settings: current(50/15/0/0), goal(50/15/0/0), user(50/15/0/0)
817 Total transfers 0 (0 reads and 0 writes)
820 1.6 Parallel port info in /proc/parport
821 ---------------------------------------
823 The directory /proc/parport contains information about the parallel ports of
824 your system. It has one subdirectory for each port, named after the port
827 These directories contain the four files shown in Table 1-8.
830 Table 1-8: Files in /proc/parport
831 ..............................................................................
833 autoprobe Any IEEE-1284 device ID information that has been acquired.
834 devices list of the device drivers using that port. A + will appear by the
835 name of the device currently using the port (it might not appear
837 hardware Parallel port's base address, IRQ line and DMA channel.
838 irq IRQ that parport is using for that port. This is in a separate
839 file to allow you to alter it by writing a new value in (IRQ
841 ..............................................................................
843 1.7 TTY info in /proc/tty
844 -------------------------
846 Information about the available and actually used tty's can be found in the
847 directory /proc/tty.You'll find entries for drivers and line disciplines in
848 this directory, as shown in Table 1-9.
851 Table 1-9: Files in /proc/tty
852 ..............................................................................
854 drivers list of drivers and their usage
855 ldiscs registered line disciplines
856 driver/serial usage statistic and status of single tty lines
857 ..............................................................................
859 To see which tty's are currently in use, you can simply look into the file
862 > cat /proc/tty/drivers
863 pty_slave /dev/pts 136 0-255 pty:slave
864 pty_master /dev/ptm 128 0-255 pty:master
865 pty_slave /dev/ttyp 3 0-255 pty:slave
866 pty_master /dev/pty 2 0-255 pty:master
867 serial /dev/cua 5 64-67 serial:callout
868 serial /dev/ttyS 4 64-67 serial
869 /dev/tty0 /dev/tty0 4 0 system:vtmaster
870 /dev/ptmx /dev/ptmx 5 2 system
871 /dev/console /dev/console 5 1 system:console
872 /dev/tty /dev/tty 5 0 system:/dev/tty
873 unknown /dev/tty 4 1-63 console
876 1.8 Miscellaneous kernel statistics in /proc/stat
877 -------------------------------------------------
879 Various pieces of information about kernel activity are available in the
880 /proc/stat file. All of the numbers reported in this file are aggregates
881 since the system first booted. For a quick look, simply cat the file:
884 cpu 2255 34 2290 22625563 6290 127 456 0
885 cpu0 1132 34 1441 11311718 3675 127 438 0
886 cpu1 1123 0 849 11313845 2614 0 18 0
887 intr 114930548 113199788 3 0 5 263 0 4 [... lots more numbers ...]
894 The very first "cpu" line aggregates the numbers in all of the other "cpuN"
895 lines. These numbers identify the amount of time the CPU has spent performing
896 different kinds of work. Time units are in USER_HZ (typically hundredths of a
897 second). The meanings of the columns are as follows, from left to right:
899 - user: normal processes executing in user mode
900 - nice: niced processes executing in user mode
901 - system: processes executing in kernel mode
902 - idle: twiddling thumbs
903 - iowait: waiting for I/O to complete
904 - irq: servicing interrupts
905 - softirq: servicing softirqs
906 - steal: involuntary wait
908 The "intr" line gives counts of interrupts serviced since boot time, for each
909 of the possible system interrupts. The first column is the total of all
910 interrupts serviced; each subsequent column is the total for that particular
913 The "ctxt" line gives the total number of context switches across all CPUs.
915 The "btime" line gives the time at which the system booted, in seconds since
918 The "processes" line gives the number of processes and threads created, which
919 includes (but is not limited to) those created by calls to the fork() and
920 clone() system calls.
922 The "procs_running" line gives the number of processes currently running on
925 The "procs_blocked" line gives the number of processes currently blocked,
926 waiting for I/O to complete.
929 1.9 Ext4 file system parameters
930 ------------------------------
932 Information about mounted ext4 file systems can be found in
933 /proc/fs/ext4. Each mounted filesystem will have a directory in
934 /proc/fs/ext4 based on its device name (i.e., /proc/fs/ext4/hdc or
935 /proc/fs/ext4/dm-0). The files in each per-device directory are shown
936 in Table 1-10, below.
938 Table 1-10: Files in /proc/fs/ext4/<devname>
939 ..............................................................................
941 mb_groups details of multiblock allocator buddy cache of free blocks
942 mb_history multiblock allocation history
943 stats controls whether the multiblock allocator should start
944 collecting statistics, which are shown during the unmount
945 group_prealloc the multiblock allocator will round up allocation
946 requests to a multiple of this tuning parameter if the
947 stripe size is not set in the ext4 superblock
948 max_to_scan The maximum number of extents the multiblock allocator
949 will search to find the best extent
950 min_to_scan The minimum number of extents the multiblock allocator
951 will search to find the best extent
952 order2_req Tuning parameter which controls the minimum size for
953 requests (as a power of 2) where the buddy cache is
955 stream_req Files which have fewer blocks than this tunable
956 parameter will have their blocks allocated out of a
957 block group specific preallocation pool, so that small
958 files are packed closely together. Each large file
959 will have its blocks allocated out of its own unique
961 inode_readahead Tuning parameter which controls the maximum number of
962 inode table blocks that ext4's inode table readahead
963 algorithm will pre-read into the buffer cache
964 ..............................................................................
967 ------------------------------------------------------------------------------
969 ------------------------------------------------------------------------------
970 The /proc file system serves information about the running system. It not only
971 allows access to process data but also allows you to request the kernel status
972 by reading files in the hierarchy.
974 The directory structure of /proc reflects the types of information and makes
975 it easy, if not obvious, where to look for specific data.
976 ------------------------------------------------------------------------------
978 ------------------------------------------------------------------------------
979 CHAPTER 2: MODIFYING SYSTEM PARAMETERS
980 ------------------------------------------------------------------------------
982 ------------------------------------------------------------------------------
984 ------------------------------------------------------------------------------
985 * Modifying kernel parameters by writing into files found in /proc/sys
986 * Exploring the files which modify certain parameters
987 * Review of the /proc/sys file tree
988 ------------------------------------------------------------------------------
991 A very interesting part of /proc is the directory /proc/sys. This is not only
992 a source of information, it also allows you to change parameters within the
993 kernel. Be very careful when attempting this. You can optimize your system,
994 but you can also cause it to crash. Never alter kernel parameters on a
995 production system. Set up a development machine and test to make sure that
996 everything works the way you want it to. You may have no alternative but to
997 reboot the machine once an error has been made.
999 To change a value, simply echo the new value into the file. An example is
1000 given below in the section on the file system data. You need to be root to do
1001 this. You can create your own boot script to perform this every time your
1004 The files in /proc/sys can be used to fine tune and monitor miscellaneous and
1005 general things in the operation of the Linux kernel. Since some of the files
1006 can inadvertently disrupt your system, it is advisable to read both
1007 documentation and source before actually making adjustments. In any case, be
1008 very careful when writing to any of these files. The entries in /proc may
1009 change slightly between the 2.1.* and the 2.2 kernel, so if there is any doubt
1010 review the kernel documentation in the directory /usr/src/linux/Documentation.
1011 This chapter is heavily based on the documentation included in the pre 2.2
1012 kernels, and became part of it in version 2.2.1 of the Linux kernel.
1014 2.1 /proc/sys/fs - File system data
1015 -----------------------------------
1017 This subdirectory contains specific file system, file handle, inode, dentry
1018 and quota information.
1020 Currently, these files are in /proc/sys/fs:
1025 Status of the directory cache. Since directory entries are dynamically
1026 allocated and deallocated, this file indicates the current status. It holds
1027 six values, in which the last two are not used and are always zero. The others
1028 are listed in table 2-1.
1031 Table 2-1: Status files of the directory cache
1032 ..............................................................................
1034 nr_dentry Almost always zero
1035 nr_unused Number of unused cache entries
1037 in seconds after the entry may be reclaimed, when memory is short
1038 want_pages internally
1039 ..............................................................................
1041 dquot-nr and dquot-max
1042 ----------------------
1044 The file dquot-max shows the maximum number of cached disk quota entries.
1046 The file dquot-nr shows the number of allocated disk quota entries and the
1047 number of free disk quota entries.
1049 If the number of available cached disk quotas is very low and you have a large
1050 number of simultaneous system users, you might want to raise the limit.
1052 file-nr and file-max
1053 --------------------
1055 The kernel allocates file handles dynamically, but doesn't free them again at
1058 The value in file-max denotes the maximum number of file handles that the
1059 Linux kernel will allocate. When you get a lot of error messages about running
1060 out of file handles, you might want to raise this limit. The default value is
1061 10% of RAM in kilobytes. To change it, just write the new number into the
1064 # cat /proc/sys/fs/file-max
1066 # echo 8192 > /proc/sys/fs/file-max
1067 # cat /proc/sys/fs/file-max
1071 This method of revision is useful for all customizable parameters of the
1072 kernel - simply echo the new value to the corresponding file.
1074 Historically, the three values in file-nr denoted the number of allocated file
1075 handles, the number of allocated but unused file handles, and the maximum
1076 number of file handles. Linux 2.6 always reports 0 as the number of free file
1077 handles -- this is not an error, it just means that the number of allocated
1078 file handles exactly matches the number of used file handles.
1080 Attempts to allocate more file descriptors than file-max are reported with
1081 printk, look for "VFS: file-max limit <number> reached".
1083 inode-state and inode-nr
1084 ------------------------
1086 The file inode-nr contains the first two items from inode-state, so we'll skip
1089 inode-state contains two actual numbers and five dummy values. The numbers
1090 are nr_inodes and nr_free_inodes (in order of appearance).
1095 Denotes the number of inodes the system has allocated. This number will
1096 grow and shrink dynamically.
1101 Denotes the maximum number of file-handles a process can
1102 allocate. Default value is 1024*1024 (1048576) which should be
1103 enough for most machines. Actual limit depends on RLIMIT_NOFILE
1109 Represents the number of free inodes. Ie. The number of inuse inodes is
1110 (nr_inodes - nr_free_inodes).
1112 aio-nr and aio-max-nr
1113 ---------------------
1115 aio-nr is the running total of the number of events specified on the
1116 io_setup system call for all currently active aio contexts. If aio-nr
1117 reaches aio-max-nr then io_setup will fail with EAGAIN. Note that
1118 raising aio-max-nr does not result in the pre-allocation or re-sizing
1119 of any kernel data structures.
1121 2.2 /proc/sys/fs/binfmt_misc - Miscellaneous binary formats
1122 -----------------------------------------------------------
1124 Besides these files, there is the subdirectory /proc/sys/fs/binfmt_misc. This
1125 handles the kernel support for miscellaneous binary formats.
1127 Binfmt_misc provides the ability to register additional binary formats to the
1128 Kernel without compiling an additional module/kernel. Therefore, binfmt_misc
1129 needs to know magic numbers at the beginning or the filename extension of the
1132 It works by maintaining a linked list of structs that contain a description of
1133 a binary format, including a magic with size (or the filename extension),
1134 offset and mask, and the interpreter name. On request it invokes the given
1135 interpreter with the original program as argument, as binfmt_java and
1136 binfmt_em86 and binfmt_mz do. Since binfmt_misc does not define any default
1137 binary-formats, you have to register an additional binary-format.
1139 There are two general files in binfmt_misc and one file per registered format.
1140 The two general files are register and status.
1142 Registering a new binary format
1143 -------------------------------
1145 To register a new binary format you have to issue the command
1147 echo :name:type:offset:magic:mask:interpreter: > /proc/sys/fs/binfmt_misc/register
1151 with appropriate name (the name for the /proc-dir entry), offset (defaults to
1152 0, if omitted), magic, mask (which can be omitted, defaults to all 0xff) and
1153 last but not least, the interpreter that is to be invoked (for example and
1154 testing /bin/echo). Type can be M for usual magic matching or E for filename
1155 extension matching (give extension in place of magic).
1157 Check or reset the status of the binary format handler
1158 ------------------------------------------------------
1160 If you do a cat on the file /proc/sys/fs/binfmt_misc/status, you will get the
1161 current status (enabled/disabled) of binfmt_misc. Change the status by echoing
1162 0 (disables) or 1 (enables) or -1 (caution: this clears all previously
1163 registered binary formats) to status. For example echo 0 > status to disable
1164 binfmt_misc (temporarily).
1166 Status of a single handler
1167 --------------------------
1169 Each registered handler has an entry in /proc/sys/fs/binfmt_misc. These files
1170 perform the same function as status, but their scope is limited to the actual
1171 binary format. By cating this file, you also receive all related information
1172 about the interpreter/magic of the binfmt.
1174 Example usage of binfmt_misc (emulate binfmt_java)
1175 --------------------------------------------------
1177 cd /proc/sys/fs/binfmt_misc
1178 echo ':Java:M::\xca\xfe\xba\xbe::/usr/local/java/bin/javawrapper:' > register
1179 echo ':HTML:E::html::/usr/local/java/bin/appletviewer:' > register
1180 echo ':Applet:M::<!--applet::/usr/local/java/bin/appletviewer:' > register
1181 echo ':DEXE:M::\x0eDEX::/usr/bin/dosexec:' > register
1184 These four lines add support for Java executables and Java applets (like
1185 binfmt_java, additionally recognizing the .html extension with no need to put
1186 <!--applet> to every applet file). You have to install the JDK and the
1187 shell-script /usr/local/java/bin/javawrapper too. It works around the
1188 brokenness of the Java filename handling. To add a Java binary, just create a
1189 link to the class-file somewhere in the path.
1191 2.3 /proc/sys/kernel - general kernel parameters
1192 ------------------------------------------------
1194 This directory reflects general kernel behaviors. As I've said before, the
1195 contents depend on your configuration. Here you'll find the most important
1196 files, along with descriptions of what they mean and how to use them.
1201 The file contains three values; highwater, lowwater, and frequency.
1203 It exists only when BSD-style process accounting is enabled. These values
1204 control its behavior. If the free space on the file system where the log lives
1205 goes below lowwater percentage, accounting suspends. If it goes above
1206 highwater percentage, accounting resumes. Frequency determines how often you
1207 check the amount of free space (value is in seconds). Default settings are: 4,
1208 2, and 30. That is, suspend accounting if there is less than 2 percent free;
1209 resume it if we have a value of 3 or more percent; consider information about
1210 the amount of free space valid for 30 seconds
1215 When the value in this file is 0, ctrl-alt-del is trapped and sent to the init
1216 program to handle a graceful restart. However, when the value is greater that
1217 zero, Linux's reaction to this key combination will be an immediate reboot,
1218 without syncing its dirty buffers.
1221 When a program (like dosemu) has the keyboard in raw mode, the
1222 ctrl-alt-del is intercepted by the program before it ever reaches the
1223 kernel tty layer, and it is up to the program to decide what to do with
1226 domainname and hostname
1227 -----------------------
1229 These files can be controlled to set the NIS domainname and hostname of your
1230 box. For the classic darkstar.frop.org a simple:
1232 # echo "darkstar" > /proc/sys/kernel/hostname
1233 # echo "frop.org" > /proc/sys/kernel/domainname
1236 would suffice to set your hostname and NIS domainname.
1238 osrelease, ostype and version
1239 -----------------------------
1241 The names make it pretty obvious what these fields contain:
1243 > cat /proc/sys/kernel/osrelease
1246 > cat /proc/sys/kernel/ostype
1249 > cat /proc/sys/kernel/version
1250 #4 Fri Oct 1 12:41:14 PDT 1999
1253 The files osrelease and ostype should be clear enough. Version needs a little
1254 more clarification. The #4 means that this is the 4th kernel built from this
1255 source base and the date after it indicates the time the kernel was built. The
1256 only way to tune these values is to rebuild the kernel.
1261 The value in this file represents the number of seconds the kernel waits
1262 before rebooting on a panic. When you use the software watchdog, the
1263 recommended setting is 60. If set to 0, the auto reboot after a kernel panic
1264 is disabled, which is the default setting.
1269 The four values in printk denote
1271 * default_message_loglevel,
1272 * minimum_console_loglevel and
1273 * default_console_loglevel
1276 These values influence printk() behavior when printing or logging error
1277 messages, which come from inside the kernel. See syslog(2) for more
1278 information on the different log levels.
1283 Messages with a higher priority than this will be printed to the console.
1285 default_message_level
1286 ---------------------
1288 Messages without an explicit priority will be printed with this priority.
1290 minimum_console_loglevel
1291 ------------------------
1293 Minimum (highest) value to which the console_loglevel can be set.
1295 default_console_loglevel
1296 ------------------------
1298 Default value for console_loglevel.
1303 This file shows the size of the generic SCSI (sg) buffer. At this point, you
1304 can't tune it yet, but you can change it at compile time by editing
1305 include/scsi/sg.h and changing the value of SG_BIG_BUFF.
1307 If you use a scanner with SANE (Scanner Access Now Easy) you might want to set
1308 this to a higher value. Refer to the SANE documentation on this issue.
1313 The location where the modprobe binary is located. The kernel uses this
1314 program to load modules on demand.
1319 The value in this file affects behavior of handling NMI. When the value is
1320 non-zero, unknown NMI is trapped and then panic occurs. At that time, kernel
1321 debugging information is displayed on console.
1323 NMI switch that most IA32 servers have fires unknown NMI up, for example.
1324 If a system hangs up, try pressing the NMI switch.
1326 panic_on_unrecovered_nmi
1327 ------------------------
1329 The default Linux behaviour on an NMI of either memory or unknown is to continue
1330 operation. For many environments such as scientific computing it is preferable
1331 that the box is taken out and the error dealt with than an uncorrected
1332 parity/ECC error get propogated.
1334 A small number of systems do generate NMI's for bizarre random reasons such as
1335 power management so the default is off. That sysctl works like the existing
1336 panic controls already in that directory.
1341 Enables/Disables the NMI watchdog on x86 systems. When the value is non-zero
1342 the NMI watchdog is enabled and will continuously test all online cpus to
1343 determine whether or not they are still functioning properly. Currently,
1344 passing "nmi_watchdog=" parameter at boot time is required for this function
1347 If LAPIC NMI watchdog method is in use (nmi_watchdog=2 kernel parameter), the
1348 NMI watchdog shares registers with oprofile. By disabling the NMI watchdog,
1349 oprofile may have more registers to utilize.
1354 Maximum number of message queue ids on the system.
1355 This value scales to the amount of lowmem. It is automatically recomputed
1356 upon memory add/remove or ipc namespace creation/removal.
1357 When a value is written into this file, msgmni's value becomes fixed, i.e. it
1358 is not recomputed anymore when one of the above events occurs.
1359 Use auto_msgmni to change this behavior.
1364 Enables/Disables automatic recomputing of msgmni upon memory add/remove or
1365 upon ipc namespace creation/removal (see the msgmni description above).
1366 Echoing "1" into this file enables msgmni automatic recomputing.
1367 Echoing "0" turns it off.
1368 auto_msgmni default value is 1.
1371 2.4 /proc/sys/vm - The virtual memory subsystem
1372 -----------------------------------------------
1374 The files in this directory can be used to tune the operation of the virtual
1375 memory (VM) subsystem of the Linux kernel.
1380 Controls the tendency of the kernel to reclaim the memory which is used for
1381 caching of directory and inode objects.
1383 At the default value of vfs_cache_pressure=100 the kernel will attempt to
1384 reclaim dentries and inodes at a "fair" rate with respect to pagecache and
1385 swapcache reclaim. Decreasing vfs_cache_pressure causes the kernel to prefer
1386 to retain dentry and inode caches. Increasing vfs_cache_pressure beyond 100
1387 causes the kernel to prefer to reclaim dentries and inodes.
1389 dirty_background_ratio
1390 ----------------------
1392 Contains, as a percentage of the dirtyable system memory (free pages + mapped
1393 pages + file cache, not including locked pages and HugePages), the number of
1394 pages at which the pdflush background writeback daemon will start writing out
1400 Contains, as a percentage of the dirtyable system memory (free pages + mapped
1401 pages + file cache, not including locked pages and HugePages), the number of
1402 pages at which a process which is generating disk writes will itself start
1403 writing out dirty data.
1405 dirty_writeback_centisecs
1406 -------------------------
1408 The pdflush writeback daemons will periodically wake up and write `old' data
1409 out to disk. This tunable expresses the interval between those wakeups, in
1410 100'ths of a second.
1412 Setting this to zero disables periodic writeback altogether.
1414 dirty_expire_centisecs
1415 ----------------------
1417 This tunable is used to define when dirty data is old enough to be eligible
1418 for writeout by the pdflush daemons. It is expressed in 100'ths of a second.
1419 Data which has been dirty in-memory for longer than this interval will be
1420 written out next time a pdflush daemon wakes up.
1422 highmem_is_dirtyable
1423 --------------------
1425 Only present if CONFIG_HIGHMEM is set.
1427 This defaults to 0 (false), meaning that the ratios set above are calculated
1428 as a percentage of lowmem only. This protects against excessive scanning
1429 in page reclaim, swapping and general VM distress.
1431 Setting this to 1 can be useful on 32 bit machines where you want to make
1432 random changes within an MMAPed file that is larger than your available
1433 lowmem without causing large quantities of random IO. Is is safe if the
1434 behavior of all programs running on the machine is known and memory will
1435 not be otherwise stressed.
1440 If non-zero, this sysctl disables the new 32-bit mmap mmap layout - the kernel
1441 will use the legacy (2.4) layout for all processes.
1443 lowmem_reserve_ratio
1444 ---------------------
1446 For some specialised workloads on highmem machines it is dangerous for
1447 the kernel to allow process memory to be allocated from the "lowmem"
1448 zone. This is because that memory could then be pinned via the mlock()
1449 system call, or by unavailability of swapspace.
1451 And on large highmem machines this lack of reclaimable lowmem memory
1454 So the Linux page allocator has a mechanism which prevents allocations
1455 which _could_ use highmem from using too much lowmem. This means that
1456 a certain amount of lowmem is defended from the possibility of being
1457 captured into pinned user memory.
1459 (The same argument applies to the old 16 megabyte ISA DMA region. This
1460 mechanism will also defend that region from allocations which could use
1463 The `lowmem_reserve_ratio' tunable determines how aggressive the kernel is
1464 in defending these lower zones.
1466 If you have a machine which uses highmem or ISA DMA and your
1467 applications are using mlock(), or if you are running with no swap then
1468 you probably should change the lowmem_reserve_ratio setting.
1470 The lowmem_reserve_ratio is an array. You can see them by reading this file.
1472 % cat /proc/sys/vm/lowmem_reserve_ratio
1475 Note: # of this elements is one fewer than number of zones. Because the highest
1476 zone's value is not necessary for following calculation.
1478 But, these values are not used directly. The kernel calculates # of protection
1479 pages for each zones from them. These are shown as array of protection pages
1480 in /proc/zoneinfo like followings. (This is an example of x86-64 box).
1481 Each zone has an array of protection pages like this.
1492 protection: (0, 2004, 2004, 2004)
1493 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1498 These protections are added to score to judge whether this zone should be used
1499 for page allocation or should be reclaimed.
1501 In this example, if normal pages (index=2) are required to this DMA zone and
1502 pages_high is used for watermark, the kernel judges this zone should not be
1503 used because pages_free(1355) is smaller than watermark + protection[2]
1504 (4 + 2004 = 2008). If this protection value is 0, this zone would be used for
1505 normal page requirement. If requirement is DMA zone(index=0), protection[0]
1508 zone[i]'s protection[j] is calculated by following expression.
1511 zone[i]->protection[j]
1512 = (total sums of present_pages from zone[i+1] to zone[j] on the node)
1513 / lowmem_reserve_ratio[i];
1515 (should not be protected. = 0;
1517 (not necessary, but looks 0)
1519 The default values of lowmem_reserve_ratio[i] are
1520 256 (if zone[i] means DMA or DMA32 zone)
1522 As above expression, they are reciprocal number of ratio.
1523 256 means 1/256. # of protection pages becomes about "0.39%" of total present
1524 pages of higher zones on the node.
1526 If you would like to protect more pages, smaller values are effective.
1527 The minimum value is 1 (1/1 -> 100%).
1532 page-cluster controls the number of pages which are written to swap in
1533 a single attempt. The swap I/O size.
1535 It is a logarithmic value - setting it to zero means "1 page", setting
1536 it to 1 means "2 pages", setting it to 2 means "4 pages", etc.
1538 The default value is three (eight pages at a time). There may be some
1539 small benefits in tuning this to a different value if your workload is
1545 Controls overcommit of system memory, possibly allowing processes
1546 to allocate (but not use) more memory than is actually available.
1549 0 - Heuristic overcommit handling. Obvious overcommits of
1550 address space are refused. Used for a typical system. It
1551 ensures a seriously wild allocation fails while allowing
1552 overcommit to reduce swap usage. root is allowed to
1553 allocate slightly more memory in this mode. This is the
1556 1 - Always overcommit. Appropriate for some scientific
1559 2 - Don't overcommit. The total address space commit
1560 for the system is not permitted to exceed swap plus a
1561 configurable percentage (default is 50) of physical RAM.
1562 Depending on the percentage you use, in most situations
1563 this means a process will not be killed while attempting
1564 to use already-allocated memory but will receive errors
1565 on memory allocation as appropriate.
1570 Percentage of physical memory size to include in overcommit calculations
1573 Memory allocation limit = swapspace + physmem * (overcommit_ratio / 100)
1575 swapspace = total size of all swap areas
1576 physmem = size of physical memory in system
1578 nr_hugepages and hugetlb_shm_group
1579 ----------------------------------
1581 nr_hugepages configures number of hugetlb page reserved for the system.
1583 hugetlb_shm_group contains group id that is allowed to create SysV shared
1584 memory segment using hugetlb page.
1586 hugepages_treat_as_movable
1587 --------------------------
1589 This parameter is only useful when kernelcore= is specified at boot time to
1590 create ZONE_MOVABLE for pages that may be reclaimed or migrated. Huge pages
1591 are not movable so are not normally allocated from ZONE_MOVABLE. A non-zero
1592 value written to hugepages_treat_as_movable allows huge pages to be allocated
1595 Once enabled, the ZONE_MOVABLE is treated as an area of memory the huge
1596 pages pool can easily grow or shrink within. Assuming that applications are
1597 not running that mlock() a lot of memory, it is likely the huge pages pool
1598 can grow to the size of ZONE_MOVABLE by repeatedly entering the desired value
1599 into nr_hugepages and triggering page reclaim.
1604 laptop_mode is a knob that controls "laptop mode". All the things that are
1605 controlled by this knob are discussed in Documentation/laptops/laptop-mode.txt.
1610 block_dump enables block I/O debugging when set to a nonzero value. More
1611 information on block I/O debugging is in Documentation/laptops/laptop-mode.txt.
1616 This file contains valid hold time of swap out protection token. The Linux
1617 VM has token based thrashing control mechanism and uses the token to prevent
1618 unnecessary page faults in thrashing situation. The unit of the value is
1619 second. The value would be useful to tune thrashing behavior.
1624 Writing to this will cause the kernel to drop clean caches, dentries and
1625 inodes from memory, causing that memory to become free.
1628 echo 1 > /proc/sys/vm/drop_caches
1629 To free dentries and inodes:
1630 echo 2 > /proc/sys/vm/drop_caches
1631 To free pagecache, dentries and inodes:
1632 echo 3 > /proc/sys/vm/drop_caches
1634 As this is a non-destructive operation and dirty objects are not freeable, the
1635 user should run `sync' first.
1638 2.5 /proc/sys/dev - Device specific parameters
1639 ----------------------------------------------
1641 Currently there is only support for CDROM drives, and for those, there is only
1642 one read-only file containing information about the CD-ROM drives attached to
1645 >cat /proc/sys/dev/cdrom/info
1646 CD-ROM information, Id: cdrom.c 2.55 1999/04/25
1650 drive # of slots: 1 0
1654 Can change speed: 1 1
1655 Can select disk: 0 1
1656 Can read multisession: 1 1
1658 Reports media changed: 1 1
1662 You see two drives, sr0 and hdb, along with a list of their features.
1664 2.6 /proc/sys/sunrpc - Remote procedure calls
1665 ---------------------------------------------
1667 This directory contains four files, which enable or disable debugging for the
1668 RPC functions NFS, NFS-daemon, RPC and NLM. The default values are 0. They can
1669 be set to one to turn debugging on. (The default value is 0 for each)
1671 2.7 /proc/sys/net - Networking stuff
1672 ------------------------------------
1674 The interface to the networking parts of the kernel is located in
1675 /proc/sys/net. Table 2-3 shows all possible subdirectories. You may see only
1676 some of them, depending on your kernel's configuration.
1679 Table 2-3: Subdirectories in /proc/sys/net
1680 ..............................................................................
1681 Directory Content Directory Content
1682 core General parameter appletalk Appletalk protocol
1683 unix Unix domain sockets netrom NET/ROM
1684 802 E802 protocol ax25 AX25
1685 ethernet Ethernet protocol rose X.25 PLP layer
1686 ipv4 IP version 4 x25 X.25 protocol
1687 ipx IPX token-ring IBM token ring
1688 bridge Bridging decnet DEC net
1690 ..............................................................................
1692 We will concentrate on IP networking here. Since AX15, X.25, and DEC Net are
1693 only minor players in the Linux world, we'll skip them in this chapter. You'll
1694 find some short info on Appletalk and IPX further on in this chapter. Review
1695 the online documentation and the kernel source to get a detailed view of the
1696 parameters for those protocols. In this section we'll discuss the
1697 subdirectories printed in bold letters in the table above. As default values
1698 are suitable for most needs, there is no need to change these values.
1700 /proc/sys/net/core - Network core options
1701 -----------------------------------------
1706 The default setting of the socket receive buffer in bytes.
1711 The maximum receive socket buffer size in bytes.
1716 The default setting (in bytes) of the socket send buffer.
1721 The maximum send socket buffer size in bytes.
1723 message_burst and message_cost
1724 ------------------------------
1726 These parameters are used to limit the warning messages written to the kernel
1727 log from the networking code. They enforce a rate limit to make a
1728 denial-of-service attack impossible. A higher message_cost factor, results in
1729 fewer messages that will be written. Message_burst controls when messages will
1730 be dropped. The default settings limit warning messages to one every five
1736 This controls console messages from the networking stack that can occur because
1737 of problems on the network like duplicate address or bad checksums. Normally,
1738 this should be enabled, but if the problem persists the messages can be
1745 Maximum number of packets, queued on the INPUT side, when the interface
1746 receives packets faster than kernel can process them.
1751 Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
1752 of struct cmsghdr structures with appended data.
1754 /proc/sys/net/unix - Parameters for Unix domain sockets
1755 -------------------------------------------------------
1757 There are only two files in this subdirectory. They control the delays for
1758 deleting and destroying socket descriptors.
1760 2.8 /proc/sys/net/ipv4 - IPV4 settings
1761 --------------------------------------
1763 IP version 4 is still the most used protocol in Unix networking. It will be
1764 replaced by IP version 6 in the next couple of years, but for the moment it's
1765 the de facto standard for the internet and is used in most networking
1766 environments around the world. Because of the importance of this protocol,
1767 we'll have a deeper look into the subtree controlling the behavior of the IPv4
1768 subsystem of the Linux kernel.
1770 Let's start with the entries in /proc/sys/net/ipv4.
1775 icmp_echo_ignore_all and icmp_echo_ignore_broadcasts
1776 ----------------------------------------------------
1778 Turn on (1) or off (0), if the kernel should ignore all ICMP ECHO requests, or
1779 just those to broadcast and multicast addresses.
1781 Please note that if you accept ICMP echo requests with a broadcast/multi\-cast
1782 destination address your network may be used as an exploder for denial of
1783 service packet flooding attacks to other hosts.
1785 icmp_destunreach_rate, icmp_echoreply_rate, icmp_paramprob_rate and icmp_timeexeed_rate
1786 ---------------------------------------------------------------------------------------
1788 Sets limits for sending ICMP packets to specific targets. A value of zero
1789 disables all limiting. Any positive value sets the maximum package rate in
1790 hundredth of a second (on Intel systems).
1798 This file contains the number one if the host received its IP configuration by
1799 RARP, BOOTP, DHCP or a similar mechanism. Otherwise it is zero.
1804 TTL (Time To Live) for IPv4 interfaces. This is simply the maximum number of
1805 hops a packet may travel.
1810 Enable dynamic socket address rewriting on interface address change. This is
1811 useful for dialup interface with changing IP addresses.
1816 Enable or disable forwarding of IP packages between interfaces. Changing this
1817 value resets all other parameters to their default values. They differ if the
1818 kernel is configured as host or router.
1823 Range of ports used by TCP and UDP to choose the local port. Contains two
1824 numbers, the first number is the lowest port, the second number the highest
1825 local port. Default is 1024-4999. Should be changed to 32768-61000 for
1831 Global switch to turn path MTU discovery off. It can also be set on a per
1832 socket basis by the applications or on a per route basis.
1837 Enable/disable debugging of IP masquerading.
1839 IP fragmentation settings
1840 -------------------------
1842 ipfrag_high_trash and ipfrag_low_trash
1843 --------------------------------------
1845 Maximum memory used to reassemble IP fragments. When ipfrag_high_thresh bytes
1846 of memory is allocated for this purpose, the fragment handler will toss
1847 packets until ipfrag_low_thresh is reached.
1852 Time in seconds to keep an IP fragment in memory.
1860 This file controls the use of the ECN bit in the IPv4 headers. This is a new
1861 feature about Explicit Congestion Notification, but some routers and firewalls
1862 block traffic that has this bit set, so it could be necessary to echo 0 to
1863 /proc/sys/net/ipv4/tcp_ecn if you want to talk to these sites. For more info
1864 you could read RFC2481.
1866 tcp_retrans_collapse
1867 --------------------
1869 Bug-to-bug compatibility with some broken printers. On retransmit, try to send
1870 larger packets to work around bugs in certain TCP stacks. Can be turned off by
1873 tcp_keepalive_probes
1874 --------------------
1876 Number of keep alive probes TCP sends out, until it decides that the
1877 connection is broken.
1882 How often TCP sends out keep alive messages, when keep alive is enabled. The
1888 Number of times initial SYNs for a TCP connection attempt will be
1889 retransmitted. Should not be higher than 255. This is only the timeout for
1890 outgoing connections, for incoming connections the number of retransmits is
1891 defined by tcp_retries1.
1896 Enable select acknowledgments after RFC2018.
1901 Enable timestamps as defined in RFC1323.
1906 Enable the strict RFC793 interpretation of the TCP urgent pointer field. The
1907 default is to use the BSD compatible interpretation of the urgent pointer
1908 pointing to the first byte after the urgent data. The RFC793 interpretation is
1909 to have it point to the last byte of urgent data. Enabling this option may
1910 lead to interoperability problems. Disabled by default.
1915 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES. Send out
1916 syncookies when the syn backlog queue of a socket overflows. This is to ward
1917 off the common 'syn flood attack'. Disabled by default.
1919 Note that the concept of a socket backlog is abandoned. This means the peer
1920 may not receive reliable error messages from an over loaded server with
1926 Enable window scaling as defined in RFC1323.
1931 The length of time in seconds it takes to receive a final FIN before the
1932 socket is always closed. This is strictly a violation of the TCP
1933 specification, but required to prevent denial-of-service attacks.
1938 Indicates how many keep alive probes are sent per slow timer run. Should not
1939 be set too high to prevent bursts.
1944 Length of the per socket backlog queue. Since Linux 2.2 the backlog specified
1945 in listen(2) only specifies the length of the backlog queue of already
1946 established sockets. When more connection requests arrive Linux starts to drop
1947 packets. When syncookies are enabled the packets are still answered and the
1948 maximum queue is effectively ignored.
1953 Defines how often an answer to a TCP connection request is retransmitted
1959 Defines how often a TCP packet is retransmitted before giving up.
1961 Interface specific settings
1962 ---------------------------
1964 In the directory /proc/sys/net/ipv4/conf you'll find one subdirectory for each
1965 interface the system knows about and one directory calls all. Changes in the
1966 all subdirectory affect all interfaces, whereas changes in the other
1967 subdirectories affect only one interface. All directories have the same
1973 This switch decides if the kernel accepts ICMP redirect messages or not. The
1974 default is 'yes' if the kernel is configured for a regular host and 'no' for a
1975 router configuration.
1980 Should source routed packages be accepted or declined. The default is
1981 dependent on the kernel configuration. It's 'yes' for routers and 'no' for
1987 Accept packets with source address 0.b.c.d with destinations not to this host
1988 as local ones. It is supposed that a BOOTP relay daemon will catch and forward
1991 The default is 0, since this feature is not implemented yet (kernel version
1997 Enable or disable IP forwarding on this interface.
2002 Log packets with source addresses with no known route to kernel log.
2007 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE and a
2008 multicast routing daemon is required.
2013 Does (1) or does not (0) perform proxy ARP.
2018 Integer value determines if a source validation should be made. 1 means yes, 0
2019 means no. Disabled by default, but local/broadcast address spoofing is always
2022 If you set this to 1 on a router that is the only connection for a network to
2023 the net, it will prevent spoofing attacks against your internal networks
2024 (external addresses can still be spoofed), without the need for additional
2030 Accept ICMP redirect messages only for gateways, listed in default gateway
2031 list. Enabled by default.
2036 If it is not set the kernel does not assume that different subnets on this
2037 device can communicate directly. Default setting is 'yes'.
2042 Determines whether to send ICMP redirects to other hosts.
2047 The directory /proc/sys/net/ipv4/route contains several file to control
2050 error_burst and error_cost
2051 --------------------------
2053 These parameters are used to limit how many ICMP destination unreachable to
2054 send from the host in question. ICMP destination unreachable messages are
2055 sent when we cannot reach the next hop while trying to transmit a packet.
2056 It will also print some error messages to kernel logs if someone is ignoring
2057 our ICMP redirects. The higher the error_cost factor is, the fewer
2058 destination unreachable and error messages will be let through. Error_burst
2059 controls when destination unreachable messages and error messages will be
2060 dropped. The default settings limit warning messages to five every second.
2065 Writing to this file results in a flush of the routing cache.
2067 gc_elasticity, gc_interval, gc_min_interval_ms, gc_timeout, gc_thresh
2068 ---------------------------------------------------------------------
2070 Values to control the frequency and behavior of the garbage collection
2071 algorithm for the routing cache. gc_min_interval is deprecated and replaced
2072 by gc_min_interval_ms.
2078 Maximum size of the routing cache. Old entries will be purged once the cache
2079 reached has this size.
2081 redirect_load, redirect_number
2082 ------------------------------
2084 Factors which determine if more ICPM redirects should be sent to a specific
2085 host. No redirects will be sent once the load limit or the maximum number of
2086 redirects has been reached.
2091 Timeout for redirects. After this period redirects will be sent again, even if
2092 this has been stopped, because the load or number limit has been reached.
2094 Network Neighbor handling
2095 -------------------------
2097 Settings about how to handle connections with direct neighbors (nodes attached
2098 to the same link) can be found in the directory /proc/sys/net/ipv4/neigh.
2100 As we saw it in the conf directory, there is a default subdirectory which
2101 holds the default values, and one directory for each interface. The contents
2102 of the directories are identical, with the single exception that the default
2103 settings contain additional options to set garbage collection parameters.
2105 In the interface directories you'll find the following entries:
2107 base_reachable_time, base_reachable_time_ms
2108 -------------------------------------------
2110 A base value used for computing the random reachable time value as specified
2113 Expression of base_reachable_time, which is deprecated, is in seconds.
2114 Expression of base_reachable_time_ms is in milliseconds.
2116 retrans_time, retrans_time_ms
2117 -----------------------------
2119 The time between retransmitted Neighbor Solicitation messages.
2120 Used for address resolution and to determine if a neighbor is
2123 Expression of retrans_time, which is deprecated, is in 1/100 seconds (for
2124 IPv4) or in jiffies (for IPv6).
2125 Expression of retrans_time_ms is in milliseconds.
2130 Maximum queue length for a pending arp request - the number of packets which
2131 are accepted from other layers while the ARP address is still resolved.
2136 Maximum for random delay of answers to neighbor solicitation messages in
2137 jiffies (1/100 sec). Not yet implemented (Linux does not have anycast support
2143 Maximum number of retries for unicast solicitation.
2148 Maximum number of retries for multicast solicitation.
2150 delay_first_probe_time
2151 ----------------------
2153 Delay for the first time probe if the neighbor is reachable. (see
2159 An ARP/neighbor entry is only replaced with a new one if the old is at least
2160 locktime old. This prevents ARP cache thrashing.
2165 Maximum time (real time is random [0..proxytime]) before answering to an ARP
2166 request for which we have an proxy ARP entry. In some cases, this is used to
2167 prevent network flooding.
2172 Maximum queue length of the delayed proxy arp timer. (see proxy_delay).
2177 Determines the number of requests to send to the user level ARP daemon. Use 0
2183 Determines how often to check for stale ARP entries. After an ARP entry is
2184 stale it will be resolved again (which is useful when an IP address migrates
2185 to another machine). When ucast_solicit is greater than 0 it first tries to
2186 send an ARP packet directly to the known host When that fails and
2187 mcast_solicit is greater than 0, an ARP request is broadcasted.
2192 The /proc/sys/net/appletalk directory holds the Appletalk configuration data
2193 when Appletalk is loaded. The configurable parameters are:
2198 The amount of time we keep an ARP entry before expiring it. Used to age out
2204 The amount of time we will spend trying to resolve an Appletalk address.
2206 aarp-retransmit-limit
2207 ---------------------
2209 The number of times we will retransmit a query before giving up.
2214 Controls the rate at which expires are checked.
2216 The directory /proc/net/appletalk holds the list of active Appletalk sockets
2219 The fields indicate the DDP type, the local address (in network:node format)
2220 the remote address, the size of the transmit pending queue, the size of the
2221 received queue (bytes waiting for applications to read) the state and the uid
2224 /proc/net/atalk_iface lists all the interfaces configured for appletalk.It
2225 shows the name of the interface, its Appletalk address, the network range on
2226 that address (or network number for phase 1 networks), and the status of the
2229 /proc/net/atalk_route lists each known network route. It lists the target
2230 (network) that the route leads to, the router (may be directly connected), the
2231 route flags, and the device the route is using.
2236 The IPX protocol has no tunable values in proc/sys/net.
2238 The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
2239 socket giving the local and remote addresses in Novell format (that is
2240 network:node:port). In accordance with the strange Novell tradition,
2241 everything but the port is in hex. Not_Connected is displayed for sockets that
2242 are not tied to a specific remote address. The Tx and Rx queue sizes indicate
2243 the number of bytes pending for transmission and reception. The state
2244 indicates the state the socket is in and the uid is the owning uid of the
2247 The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
2248 it gives the network number, the node number, and indicates if the network is
2249 the primary network. It also indicates which device it is bound to (or
2250 Internal for internal networks) and the Frame Type if appropriate. Linux
2251 supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
2254 The /proc/net/ipx_route table holds a list of IPX routes. For each route it
2255 gives the destination network, the router node (or Directly) and the network
2256 address of the router (or Connected) for internal networks.
2258 2.11 /proc/sys/fs/mqueue - POSIX message queues filesystem
2259 ----------------------------------------------------------
2261 The "mqueue" filesystem provides the necessary kernel features to enable the
2262 creation of a user space library that implements the POSIX message queues
2263 API (as noted by the MSG tag in the POSIX 1003.1-2001 version of the System
2264 Interfaces specification.)
2266 The "mqueue" filesystem contains values for determining/setting the amount of
2267 resources used by the file system.
2269 /proc/sys/fs/mqueue/queues_max is a read/write file for setting/getting the
2270 maximum number of message queues allowed on the system.
2272 /proc/sys/fs/mqueue/msg_max is a read/write file for setting/getting the
2273 maximum number of messages in a queue value. In fact it is the limiting value
2274 for another (user) limit which is set in mq_open invocation. This attribute of
2275 a queue must be less or equal then msg_max.
2277 /proc/sys/fs/mqueue/msgsize_max is a read/write file for setting/getting the
2278 maximum message size value (it is every message queue's attribute set during
2281 2.12 /proc/<pid>/oom_adj - Adjust the oom-killer score
2282 ------------------------------------------------------
2284 This file can be used to adjust the score used to select which processes
2285 should be killed in an out-of-memory situation. Giving it a high score will
2286 increase the likelihood of this process being killed by the oom-killer. Valid
2287 values are in the range -16 to +15, plus the special value -17, which disables
2288 oom-killing altogether for this process.
2290 2.13 /proc/<pid>/oom_score - Display current oom-killer score
2291 -------------------------------------------------------------
2293 ------------------------------------------------------------------------------
2294 This file can be used to check the current score used by the oom-killer is for
2295 any given <pid>. Use it together with /proc/<pid>/oom_adj to tune which
2296 process should be killed in an out-of-memory situation.
2298 ------------------------------------------------------------------------------
2300 ------------------------------------------------------------------------------
2301 Certain aspects of kernel behavior can be modified at runtime, without the
2302 need to recompile the kernel, or even to reboot the system. The files in the
2303 /proc/sys tree can not only be read, but also modified. You can use the echo
2304 command to write value into these files, thereby changing the default settings
2306 ------------------------------------------------------------------------------
2308 2.14 /proc/<pid>/io - Display the IO accounting fields
2309 -------------------------------------------------------
2311 This file contains IO statistics for each running process
2316 test:/tmp # dd if=/dev/zero of=/tmp/test.dat &
2319 test:/tmp # cat /proc/3828/io
2325 write_bytes: 323932160
2326 cancelled_write_bytes: 0
2335 I/O counter: chars read
2336 The number of bytes which this task has caused to be read from storage. This
2337 is simply the sum of bytes which this process passed to read() and pread().
2338 It includes things like tty IO and it is unaffected by whether or not actual
2339 physical disk IO was required (the read might have been satisfied from
2346 I/O counter: chars written
2347 The number of bytes which this task has caused, or shall cause to be written
2348 to disk. Similar caveats apply here as with rchar.
2354 I/O counter: read syscalls
2355 Attempt to count the number of read I/O operations, i.e. syscalls like read()
2362 I/O counter: write syscalls
2363 Attempt to count the number of write I/O operations, i.e. syscalls like
2364 write() and pwrite().
2370 I/O counter: bytes read
2371 Attempt to count the number of bytes which this process really did cause to
2372 be fetched from the storage layer. Done at the submit_bio() level, so it is
2373 accurate for block-backed filesystems. <please add status regarding NFS and
2374 CIFS at a later time>
2380 I/O counter: bytes written
2381 Attempt to count the number of bytes which this process caused to be sent to
2382 the storage layer. This is done at page-dirtying time.
2385 cancelled_write_bytes
2386 ---------------------
2388 The big inaccuracy here is truncate. If a process writes 1MB to a file and
2389 then deletes the file, it will in fact perform no writeout. But it will have
2390 been accounted as having caused 1MB of write.
2391 In other words: The number of bytes which this process caused to not happen,
2392 by truncating pagecache. A task can cause "negative" IO too. If this task
2393 truncates some dirty pagecache, some IO which another task has been accounted
2394 for (in it's write_bytes) will not be happening. We _could_ just subtract that
2395 from the truncating task's write_bytes, but there is information loss in doing
2402 At its current implementation state, this is a bit racy on 32-bit machines: if
2403 process A reads process B's /proc/pid/io while process B is updating one of
2404 those 64-bit counters, process A could see an intermediate result.
2407 More information about this can be found within the taskstats documentation in
2408 Documentation/accounting.
2410 2.15 /proc/<pid>/coredump_filter - Core dump filtering settings
2411 ---------------------------------------------------------------
2412 When a process is dumped, all anonymous memory is written to a core file as
2413 long as the size of the core file isn't limited. But sometimes we don't want
2414 to dump some memory segments, for example, huge shared memory. Conversely,
2415 sometimes we want to save file-backed memory segments into a core file, not
2416 only the individual files.
2418 /proc/<pid>/coredump_filter allows you to customize which memory segments
2419 will be dumped when the <pid> process is dumped. coredump_filter is a bitmask
2420 of memory types. If a bit of the bitmask is set, memory segments of the
2421 corresponding memory type are dumped, otherwise they are not dumped.
2423 The following 7 memory types are supported:
2424 - (bit 0) anonymous private memory
2425 - (bit 1) anonymous shared memory
2426 - (bit 2) file-backed private memory
2427 - (bit 3) file-backed shared memory
2428 - (bit 4) ELF header pages in file-backed private memory areas (it is
2429 effective only if the bit 2 is cleared)
2430 - (bit 5) hugetlb private memory
2431 - (bit 6) hugetlb shared memory
2433 Note that MMIO pages such as frame buffer are never dumped and vDSO pages
2434 are always dumped regardless of the bitmask status.
2436 Note bit 0-4 doesn't effect any hugetlb memory. hugetlb memory are only
2437 effected by bit 5-6.
2439 Default value of coredump_filter is 0x23; this means all anonymous memory
2440 segments and hugetlb private memory are dumped.
2442 If you don't want to dump all shared memory segments attached to pid 1234,
2443 write 0x21 to the process's proc file.
2445 $ echo 0x21 > /proc/1234/coredump_filter
2447 When a new process is created, the process inherits the bitmask status from its
2448 parent. It is useful to set up coredump_filter before the program runs.
2451 $ echo 0x7 > /proc/self/coredump_filter
2454 2.16 /proc/<pid>/mountinfo - Information about mounts
2455 --------------------------------------------------------
2457 This file contains lines of the form:
2459 36 35 98:0 /mnt1 /mnt2 rw,noatime master:1 - ext3 /dev/root rw,errors=continue
2460 (1)(2)(3) (4) (5) (6) (7) (8) (9) (10) (11)
2462 (1) mount ID: unique identifier of the mount (may be reused after umount)
2463 (2) parent ID: ID of parent (or of self for the top of the mount tree)
2464 (3) major:minor: value of st_dev for files on filesystem
2465 (4) root: root of the mount within the filesystem
2466 (5) mount point: mount point relative to the process's root
2467 (6) mount options: per mount options
2468 (7) optional fields: zero or more fields of the form "tag[:value]"
2469 (8) separator: marks the end of the optional fields
2470 (9) filesystem type: name of filesystem of the form "type[.subtype]"
2471 (10) mount source: filesystem specific information or "none"
2472 (11) super options: per super block options
2474 Parsers should ignore all unrecognised optional fields. Currently the
2475 possible optional fields are:
2477 shared:X mount is shared in peer group X
2478 master:X mount is slave to peer group X
2479 propagate_from:X mount is slave and receives propagation from peer group X (*)
2480 unbindable mount is unbindable
2482 (*) X is the closest dominant peer group under the process's root. If
2483 X is the immediate master of the mount, or if there's no dominant peer
2484 group under the same root, then only the "master:X" field is present
2485 and not the "propagate_from:X" field.
2487 For more information on mount propagation see:
2489 Documentation/filesystems/sharedsubtree.txt
2491 2.17 /proc/sys/fs/epoll - Configuration options for the epoll interface
2492 --------------------------------------------------------
2494 This directory contains configuration options for the epoll(7) interface.
2499 This is the maximum number of epoll file descriptors that a single user can
2500 have open at a given time. The default value is 128, and should be enough
2506 Every epoll file descriptor can store a number of files to be monitored
2507 for event readiness. Each one of these monitored files constitutes a "watch".
2508 This configuration option sets the maximum number of "watches" that are
2509 allowed for each user.
2510 Each "watch" costs roughly 90 bytes on a 32bit kernel, and roughly 160 bytes
2512 The current default value for max_user_watches is the 1/32 of the available
2513 low memory, divided for the "watch" cost in bytes.
2516 ------------------------------------------------------------------------------