linux/Documentation/sysctl/kernel.txt
<<
>>
Prefs
   1Documentation for /proc/sys/kernel/*    kernel version 2.2.10
   2        (c) 1998, 1999,  Rik van Riel <riel@nl.linux.org>
   3        (c) 2009,        Shen Feng<shen@cn.fujitsu.com>
   4
   5For general info and legal blurb, please look in README.
   6
   7==============================================================
   8
   9This file contains documentation for the sysctl files in
  10/proc/sys/kernel/ and is valid for Linux kernel version 2.2.
  11
  12The files in this directory can be used to tune and monitor
  13miscellaneous and general things in the operation of the Linux
  14kernel. Since some of the files _can_ be used to screw up your
  15system, it is advisable to read both documentation and source
  16before actually making adjustments.
  17
  18Currently, these files might (depending on your configuration)
  19show up in /proc/sys/kernel:
  20
  21- acct
  22- acpi_video_flags
  23- auto_msgmni
  24- bootloader_type            [ X86 only ]
  25- bootloader_version         [ X86 only ]
  26- callhome                   [ S390 only ]
  27- cap_last_cap
  28- core_pattern
  29- core_pipe_limit
  30- core_uses_pid
  31- ctrl-alt-del
  32- dmesg_restrict
  33- domainname
  34- hostname
  35- hotplug
  36- kptr_restrict
  37- kstack_depth_to_print       [ X86 only ]
  38- l2cr                        [ PPC only ]
  39- modprobe                    ==> Documentation/debugging-modules.txt
  40- modules_disabled
  41- msg_next_id                 [ sysv ipc ]
  42- msgmax
  43- msgmnb
  44- msgmni
  45- nmi_watchdog
  46- osrelease
  47- ostype
  48- overflowgid
  49- overflowuid
  50- panic
  51- panic_on_oops
  52- panic_on_unrecovered_nmi
  53- panic_on_stackoverflow
  54- pid_max
  55- powersave-nap               [ PPC only ]
  56- printk
  57- printk_delay
  58- printk_ratelimit
  59- printk_ratelimit_burst
  60- randomize_va_space
  61- real-root-dev               ==> Documentation/initrd.txt
  62- reboot-cmd                  [ SPARC only ]
  63- rtsig-max
  64- rtsig-nr
  65- sem
  66- sem_next_id                 [ sysv ipc ]
  67- sg-big-buff                 [ generic SCSI device (sg) ]
  68- shm_next_id                 [ sysv ipc ]
  69- shm_rmid_forced
  70- shmall
  71- shmmax                      [ sysv ipc ]
  72- shmmni
  73- softlockup_thresh
  74- stop-a                      [ SPARC only ]
  75- sysrq                       ==> Documentation/sysrq.txt
  76- tainted
  77- threads-max
  78- unknown_nmi_panic
  79- version
  80
  81==============================================================
  82
  83acct:
  84
  85highwater lowwater frequency
  86
  87If BSD-style process accounting is enabled these values control
  88its behaviour. If free space on filesystem where the log lives
  89goes below <lowwater>% accounting suspends. If free space gets
  90above <highwater>% accounting resumes. <Frequency> determines
  91how often do we check the amount of free space (value is in
  92seconds). Default:
  934 2 30
  94That is, suspend accounting if there left <= 2% free; resume it
  95if we got >=4%; consider information about amount of free space
  96valid for 30 seconds.
  97
  98==============================================================
  99
 100acpi_video_flags:
 101
 102flags
 103
 104See Doc*/kernel/power/video.txt, it allows mode of video boot to be
 105set during run time.
 106
 107==============================================================
 108
 109auto_msgmni:
 110
 111Enables/Disables automatic recomputing of msgmni upon memory add/remove
 112or upon ipc namespace creation/removal (see the msgmni description
 113above). Echoing "1" into this file enables msgmni automatic recomputing.
 114Echoing "0" turns it off. auto_msgmni default value is 1.
 115
 116
 117==============================================================
 118
 119bootloader_type:
 120
 121x86 bootloader identification
 122
 123This gives the bootloader type number as indicated by the bootloader,
 124shifted left by 4, and OR'd with the low four bits of the bootloader
 125version.  The reason for this encoding is that this used to match the
 126type_of_loader field in the kernel header; the encoding is kept for
 127backwards compatibility.  That is, if the full bootloader type number
 128is 0x15 and the full version number is 0x234, this file will contain
 129the value 340 = 0x154.
 130
 131See the type_of_loader and ext_loader_type fields in
 132Documentation/x86/boot.txt for additional information.
 133
 134==============================================================
 135
 136bootloader_version:
 137
 138x86 bootloader version
 139
 140The complete bootloader version number.  In the example above, this
 141file will contain the value 564 = 0x234.
 142
 143See the type_of_loader and ext_loader_ver fields in
 144Documentation/x86/boot.txt for additional information.
 145
 146==============================================================
 147
 148callhome:
 149
 150Controls the kernel's callhome behavior in case of a kernel panic.
 151
 152The s390 hardware allows an operating system to send a notification
 153to a service organization (callhome) in case of an operating system panic.
 154
 155When the value in this file is 0 (which is the default behavior)
 156nothing happens in case of a kernel panic. If this value is set to "1"
 157the complete kernel oops message is send to the IBM customer service
 158organization in case the mainframe the Linux operating system is running
 159on has a service contract with IBM.
 160
 161==============================================================
 162
 163cap_last_cap
 164
 165Highest valid capability of the running kernel.  Exports
 166CAP_LAST_CAP from the kernel.
 167
 168==============================================================
 169
 170core_pattern:
 171
 172core_pattern is used to specify a core dumpfile pattern name.
 173. max length 128 characters; default value is "core"
 174. core_pattern is used as a pattern template for the output filename;
 175  certain string patterns (beginning with '%') are substituted with
 176  their actual values.
 177. backward compatibility with core_uses_pid:
 178        If core_pattern does not include "%p" (default does not)
 179        and core_uses_pid is set, then .PID will be appended to
 180        the filename.
 181. corename format specifiers:
 182        %<NUL>  '%' is dropped
 183        %%      output one '%'
 184        %p      pid
 185        %u      uid
 186        %g      gid
 187        %d      dump mode, matches PR_SET_DUMPABLE and
 188                /proc/sys/fs/suid_dumpable
 189        %s      signal number
 190        %t      UNIX time of dump
 191        %h      hostname
 192        %e      executable filename (may be shortened)
 193        %E      executable path
 194        %<OTHER> both are dropped
 195. If the first character of the pattern is a '|', the kernel will treat
 196  the rest of the pattern as a command to run.  The core dump will be
 197  written to the standard input of that program instead of to a file.
 198
 199==============================================================
 200
 201core_pipe_limit:
 202
 203This sysctl is only applicable when core_pattern is configured to pipe
 204core files to a user space helper (when the first character of
 205core_pattern is a '|', see above).  When collecting cores via a pipe
 206to an application, it is occasionally useful for the collecting
 207application to gather data about the crashing process from its
 208/proc/pid directory.  In order to do this safely, the kernel must wait
 209for the collecting process to exit, so as not to remove the crashing
 210processes proc files prematurely.  This in turn creates the
 211possibility that a misbehaving userspace collecting process can block
 212the reaping of a crashed process simply by never exiting.  This sysctl
 213defends against that.  It defines how many concurrent crashing
 214processes may be piped to user space applications in parallel.  If
 215this value is exceeded, then those crashing processes above that value
 216are noted via the kernel log and their cores are skipped.  0 is a
 217special value, indicating that unlimited processes may be captured in
 218parallel, but that no waiting will take place (i.e. the collecting
 219process is not guaranteed access to /proc/<crashing pid>/).  This
 220value defaults to 0.
 221
 222==============================================================
 223
 224core_uses_pid:
 225
 226The default coredump filename is "core".  By setting
 227core_uses_pid to 1, the coredump filename becomes core.PID.
 228If core_pattern does not include "%p" (default does not)
 229and core_uses_pid is set, then .PID will be appended to
 230the filename.
 231
 232==============================================================
 233
 234ctrl-alt-del:
 235
 236When the value in this file is 0, ctrl-alt-del is trapped and
 237sent to the init(1) program to handle a graceful restart.
 238When, however, the value is > 0, Linux's reaction to a Vulcan
 239Nerve Pinch (tm) will be an immediate reboot, without even
 240syncing its dirty buffers.
 241
 242Note: when a program (like dosemu) has the keyboard in 'raw'
 243mode, the ctrl-alt-del is intercepted by the program before it
 244ever reaches the kernel tty layer, and it's up to the program
 245to decide what to do with it.
 246
 247==============================================================
 248
 249dmesg_restrict:
 250
 251This toggle indicates whether unprivileged users are prevented
 252from using dmesg(8) to view messages from the kernel's log buffer.
 253When dmesg_restrict is set to (0) there are no restrictions. When
 254dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
 255dmesg(8).
 256
 257The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
 258default value of dmesg_restrict.
 259
 260==============================================================
 261
 262domainname & hostname:
 263
 264These files can be used to set the NIS/YP domainname and the
 265hostname of your box in exactly the same way as the commands
 266domainname and hostname, i.e.:
 267# echo "darkstar" > /proc/sys/kernel/hostname
 268# echo "mydomain" > /proc/sys/kernel/domainname
 269has the same effect as
 270# hostname "darkstar"
 271# domainname "mydomain"
 272
 273Note, however, that the classic darkstar.frop.org has the
 274hostname "darkstar" and DNS (Internet Domain Name Server)
 275domainname "frop.org", not to be confused with the NIS (Network
 276Information Service) or YP (Yellow Pages) domainname. These two
 277domain names are in general different. For a detailed discussion
 278see the hostname(1) man page.
 279
 280==============================================================
 281
 282hotplug:
 283
 284Path for the hotplug policy agent.
 285Default value is "/sbin/hotplug".
 286
 287==============================================================
 288
 289kptr_restrict:
 290
 291This toggle indicates whether restrictions are placed on
 292exposing kernel addresses via /proc and other interfaces.  When
 293kptr_restrict is set to (0), there are no restrictions.  When
 294kptr_restrict is set to (1), the default, kernel pointers
 295printed using the %pK format specifier will be replaced with 0's
 296unless the user has CAP_SYSLOG.  When kptr_restrict is set to
 297(2), kernel pointers printed using %pK will be replaced with 0's
 298regardless of privileges.
 299
 300==============================================================
 301
 302kstack_depth_to_print: (X86 only)
 303
 304Controls the number of words to print when dumping the raw
 305kernel stack.
 306
 307==============================================================
 308
 309l2cr: (PPC only)
 310
 311This flag controls the L2 cache of G3 processor boards. If
 3120, the cache is disabled. Enabled if nonzero.
 313
 314==============================================================
 315
 316modules_disabled:
 317
 318A toggle value indicating if modules are allowed to be loaded
 319in an otherwise modular kernel.  This toggle defaults to off
 320(0), but can be set true (1).  Once true, modules can be
 321neither loaded nor unloaded, and the toggle cannot be set back
 322to false.
 323
 324==============================================================
 325
 326msg_next_id, sem_next_id, and shm_next_id:
 327
 328These three toggles allows to specify desired id for next allocated IPC
 329object: message, semaphore or shared memory respectively.
 330
 331By default they are equal to -1, which means generic allocation logic.
 332Possible values to set are in range {0..INT_MAX}.
 333
 334Notes:
 3351) kernel doesn't guarantee, that new object will have desired id. So,
 336it's up to userspace, how to handle an object with "wrong" id.
 3372) Toggle with non-default value will be set back to -1 by kernel after
 338successful IPC object allocation.
 339
 340==============================================================
 341
 342nmi_watchdog:
 343
 344Enables/Disables the NMI watchdog on x86 systems. When the value is
 345non-zero the NMI watchdog is enabled and will continuously test all
 346online cpus to determine whether or not they are still functioning
 347properly. Currently, passing "nmi_watchdog=" parameter at boot time is
 348required for this function to work.
 349
 350If LAPIC NMI watchdog method is in use (nmi_watchdog=2 kernel
 351parameter), the NMI watchdog shares registers with oprofile. By
 352disabling the NMI watchdog, oprofile may have more registers to
 353utilize.
 354
 355==============================================================
 356
 357osrelease, ostype & version:
 358
 359# cat osrelease
 3602.1.88
 361# cat ostype
 362Linux
 363# cat version
 364#5 Wed Feb 25 21:49:24 MET 1998
 365
 366The files osrelease and ostype should be clear enough. Version
 367needs a little more clarification however. The '#5' means that
 368this is the fifth kernel built from this source base and the
 369date behind it indicates the time the kernel was built.
 370The only way to tune these values is to rebuild the kernel :-)
 371
 372==============================================================
 373
 374overflowgid & overflowuid:
 375
 376if your architecture did not always support 32-bit UIDs (i.e. arm,
 377i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
 378applications that use the old 16-bit UID/GID system calls, if the
 379actual UID or GID would exceed 65535.
 380
 381These sysctls allow you to change the value of the fixed UID and GID.
 382The default is 65534.
 383
 384==============================================================
 385
 386panic:
 387
 388The value in this file represents the number of seconds the kernel
 389waits before rebooting on a panic. When you use the software watchdog,
 390the recommended setting is 60.
 391
 392==============================================================
 393
 394panic_on_unrecovered_nmi:
 395
 396The default Linux behaviour on an NMI of either memory or unknown is
 397to continue operation. For many environments such as scientific
 398computing it is preferable that the box is taken out and the error
 399dealt with than an uncorrected parity/ECC error get propagated.
 400
 401A small number of systems do generate NMI's for bizarre random reasons
 402such as power management so the default is off. That sysctl works like
 403the existing panic controls already in that directory.
 404
 405==============================================================
 406
 407panic_on_oops:
 408
 409Controls the kernel's behaviour when an oops or BUG is encountered.
 410
 4110: try to continue operation
 412
 4131: panic immediately.  If the `panic' sysctl is also non-zero then the
 414   machine will be rebooted.
 415
 416==============================================================
 417
 418panic_on_stackoverflow:
 419
 420Controls the kernel's behavior when detecting the overflows of
 421kernel, IRQ and exception stacks except a user stack.
 422This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
 423
 4240: try to continue operation.
 425
 4261: panic immediately.
 427
 428==============================================================
 429
 430
 431pid_max:
 432
 433PID allocation wrap value.  When the kernel's next PID value
 434reaches this value, it wraps back to a minimum PID value.
 435PIDs of value pid_max or larger are not allocated.
 436
 437==============================================================
 438
 439ns_last_pid:
 440
 441The last pid allocated in the current (the one task using this sysctl
 442lives in) pid namespace. When selecting a pid for a next task on fork
 443kernel tries to allocate a number starting from this one.
 444
 445==============================================================
 446
 447powersave-nap: (PPC only)
 448
 449If set, Linux-PPC will use the 'nap' mode of powersaving,
 450otherwise the 'doze' mode will be used.
 451
 452==============================================================
 453
 454printk:
 455
 456The four values in printk denote: console_loglevel,
 457default_message_loglevel, minimum_console_loglevel and
 458default_console_loglevel respectively.
 459
 460These values influence printk() behavior when printing or
 461logging error messages. See 'man 2 syslog' for more info on
 462the different loglevels.
 463
 464- console_loglevel: messages with a higher priority than
 465  this will be printed to the console
 466- default_message_loglevel: messages without an explicit priority
 467  will be printed with this priority
 468- minimum_console_loglevel: minimum (highest) value to which
 469  console_loglevel can be set
 470- default_console_loglevel: default value for console_loglevel
 471
 472==============================================================
 473
 474printk_delay:
 475
 476Delay each printk message in printk_delay milliseconds
 477
 478Value from 0 - 10000 is allowed.
 479
 480==============================================================
 481
 482printk_ratelimit:
 483
 484Some warning messages are rate limited. printk_ratelimit specifies
 485the minimum length of time between these messages (in jiffies), by
 486default we allow one every 5 seconds.
 487
 488A value of 0 will disable rate limiting.
 489
 490==============================================================
 491
 492printk_ratelimit_burst:
 493
 494While long term we enforce one message per printk_ratelimit
 495seconds, we do allow a burst of messages to pass through.
 496printk_ratelimit_burst specifies the number of messages we can
 497send before ratelimiting kicks in.
 498
 499==============================================================
 500
 501randomize_va_space:
 502
 503This option can be used to select the type of process address
 504space randomization that is used in the system, for architectures
 505that support this feature.
 506
 5070 - Turn the process address space randomization off.  This is the
 508    default for architectures that do not support this feature anyways,
 509    and kernels that are booted with the "norandmaps" parameter.
 510
 5111 - Make the addresses of mmap base, stack and VDSO page randomized.
 512    This, among other things, implies that shared libraries will be
 513    loaded to random addresses.  Also for PIE-linked binaries, the
 514    location of code start is randomized.  This is the default if the
 515    CONFIG_COMPAT_BRK option is enabled.
 516
 5172 - Additionally enable heap randomization.  This is the default if
 518    CONFIG_COMPAT_BRK is disabled.
 519
 520    There are a few legacy applications out there (such as some ancient
 521    versions of libc.so.5 from 1996) that assume that brk area starts
 522    just after the end of the code+bss.  These applications break when
 523    start of the brk area is randomized.  There are however no known
 524    non-legacy applications that would be broken this way, so for most
 525    systems it is safe to choose full randomization.
 526
 527    Systems with ancient and/or broken binaries should be configured
 528    with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
 529    address space randomization.
 530
 531==============================================================
 532
 533reboot-cmd: (Sparc only)
 534
 535??? This seems to be a way to give an argument to the Sparc
 536ROM/Flash boot loader. Maybe to tell it what to do after
 537rebooting. ???
 538
 539==============================================================
 540
 541rtsig-max & rtsig-nr:
 542
 543The file rtsig-max can be used to tune the maximum number
 544of POSIX realtime (queued) signals that can be outstanding
 545in the system.
 546
 547rtsig-nr shows the number of RT signals currently queued.
 548
 549==============================================================
 550
 551sg-big-buff:
 552
 553This file shows the size of the generic SCSI (sg) buffer.
 554You can't tune it just yet, but you could change it on
 555compile time by editing include/scsi/sg.h and changing
 556the value of SG_BIG_BUFF.
 557
 558There shouldn't be any reason to change this value. If
 559you can come up with one, you probably know what you
 560are doing anyway :)
 561
 562==============================================================
 563
 564shmall:
 565
 566This parameter sets the total amount of shared memory pages that
 567can be used system wide. Hence, SHMALL should always be at least
 568ceil(shmmax/PAGE_SIZE).
 569
 570If you are not sure what the default PAGE_SIZE is on your Linux
 571system, you can run the following command:
 572
 573# getconf PAGE_SIZE
 574
 575==============================================================
 576
 577shmmax:
 578
 579This value can be used to query and set the run time limit
 580on the maximum shared memory segment size that can be created.
 581Shared memory segments up to 1Gb are now supported in the
 582kernel.  This value defaults to SHMMAX.
 583
 584==============================================================
 585
 586shm_rmid_forced:
 587
 588Linux lets you set resource limits, including how much memory one
 589process can consume, via setrlimit(2).  Unfortunately, shared memory
 590segments are allowed to exist without association with any process, and
 591thus might not be counted against any resource limits.  If enabled,
 592shared memory segments are automatically destroyed when their attach
 593count becomes zero after a detach or a process termination.  It will
 594also destroy segments that were created, but never attached to, on exit
 595from the process.  The only use left for IPC_RMID is to immediately
 596destroy an unattached segment.  Of course, this breaks the way things are
 597defined, so some applications might stop working.  Note that this
 598feature will do you no good unless you also configure your resource
 599limits (in particular, RLIMIT_AS and RLIMIT_NPROC).  Most systems don't
 600need this.
 601
 602Note that if you change this from 0 to 1, already created segments
 603without users and with a dead originative process will be destroyed.
 604
 605==============================================================
 606
 607softlockup_thresh:
 608
 609This value can be used to lower the softlockup tolerance threshold.  The
 610default threshold is 60 seconds.  If a cpu is locked up for 60 seconds,
 611the kernel complains.  Valid values are 1-60 seconds.  Setting this
 612tunable to zero will disable the softlockup detection altogether.
 613
 614==============================================================
 615
 616tainted:
 617
 618Non-zero if the kernel has been tainted.  Numeric values, which
 619can be ORed together:
 620
 621   1 - A module with a non-GPL license has been loaded, this
 622       includes modules with no license.
 623       Set by modutils >= 2.4.9 and module-init-tools.
 624   2 - A module was force loaded by insmod -f.
 625       Set by modutils >= 2.4.9 and module-init-tools.
 626   4 - Unsafe SMP processors: SMP with CPUs not designed for SMP.
 627   8 - A module was forcibly unloaded from the system by rmmod -f.
 628  16 - A hardware machine check error occurred on the system.
 629  32 - A bad page was discovered on the system.
 630  64 - The user has asked that the system be marked "tainted".  This
 631       could be because they are running software that directly modifies
 632       the hardware, or for other reasons.
 633 128 - The system has died.
 634 256 - The ACPI DSDT has been overridden with one supplied by the user
 635        instead of using the one provided by the hardware.
 636 512 - A kernel warning has occurred.
 6371024 - A module from drivers/staging was loaded.
 6382048 - The system is working around a severe firmware bug.
 6394096 - An out-of-tree module has been loaded.
 640
 641==============================================================
 642
 643unknown_nmi_panic:
 644
 645The value in this file affects behavior of handling NMI. When the
 646value is non-zero, unknown NMI is trapped and then panic occurs. At
 647that time, kernel debugging information is displayed on console.
 648
 649NMI switch that most IA32 servers have fires unknown NMI up, for
 650example.  If a system hangs up, try pressing the NMI switch.
 651
lxr.linux.no kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.