1 Documentation 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>
5 For general info and legal blurb, please look in README.
7 ==============================================================
9 This file contains documentation for the sysctl files in
10 /proc/sys/kernel/ and is valid for Linux kernel version 2.2.
12 The files in this directory can be used to tune and monitor
13 miscellaneous and general things in the operation of the Linux
14 kernel. Since some of the files _can_ be used to screw up your
15 system, it is advisable to read both documentation and source
16 before actually making adjustments.
18 Currently, these files might (depending on your configuration)
19 show up in /proc/sys/kernel:
24 - bootloader_type [ X86 only ]
25 - bootloader_version [ X86 only ]
26 - callhome [ S390 only ]
37 - kstack_depth_to_print [ X86 only ]
39 - modprobe ==> Documentation/debugging-modules.txt
51 - panic_on_unrecovered_nmi
52 - panic_on_stackoverflow
54 - powersave-nap [ PPC only ]
58 - printk_ratelimit_burst
60 - real-root-dev ==> Documentation/initrd.txt
61 - reboot-cmd [ SPARC only ]
65 - sg-big-buff [ generic SCSI device (sg) ]
71 - stop-a [ SPARC only ]
72 - sysrq ==> Documentation/sysrq.txt
78 ==============================================================
82 highwater lowwater frequency
84 If BSD-style process accounting is enabled these values control
85 its behaviour. If free space on filesystem where the log lives
86 goes below <lowwater>% accounting suspends. If free space gets
87 above <highwater>% accounting resumes. <Frequency> determines
88 how often do we check the amount of free space (value is in
91 That is, suspend accounting if there left <= 2% free; resume it
92 if we got >=4%; consider information about amount of free space
95 ==============================================================
101 See Doc*/kernel/power/video.txt, it allows mode of video boot to be
104 ==============================================================
108 Enables/Disables automatic recomputing of msgmni upon memory add/remove
109 or upon ipc namespace creation/removal (see the msgmni description
110 above). Echoing "1" into this file enables msgmni automatic recomputing.
111 Echoing "0" turns it off. auto_msgmni default value is 1.
114 ==============================================================
118 x86 bootloader identification
120 This gives the bootloader type number as indicated by the bootloader,
121 shifted left by 4, and OR'd with the low four bits of the bootloader
122 version. The reason for this encoding is that this used to match the
123 type_of_loader field in the kernel header; the encoding is kept for
124 backwards compatibility. That is, if the full bootloader type number
125 is 0x15 and the full version number is 0x234, this file will contain
126 the value 340 = 0x154.
128 See the type_of_loader and ext_loader_type fields in
129 Documentation/x86/boot.txt for additional information.
131 ==============================================================
135 x86 bootloader version
137 The complete bootloader version number. In the example above, this
138 file will contain the value 564 = 0x234.
140 See the type_of_loader and ext_loader_ver fields in
141 Documentation/x86/boot.txt for additional information.
143 ==============================================================
147 Controls the kernel's callhome behavior in case of a kernel panic.
149 The s390 hardware allows an operating system to send a notification
150 to a service organization (callhome) in case of an operating system panic.
152 When the value in this file is 0 (which is the default behavior)
153 nothing happens in case of a kernel panic. If this value is set to "1"
154 the complete kernel oops message is send to the IBM customer service
155 organization in case the mainframe the Linux operating system is running
156 on has a service contract with IBM.
158 ==============================================================
162 Highest valid capability of the running kernel. Exports
163 CAP_LAST_CAP from the kernel.
165 ==============================================================
169 core_pattern is used to specify a core dumpfile pattern name.
170 . max length 128 characters; default value is "core"
171 . core_pattern is used as a pattern template for the output filename;
172 certain string patterns (beginning with '%') are substituted with
174 . backward compatibility with core_uses_pid:
175 If core_pattern does not include "%p" (default does not)
176 and core_uses_pid is set, then .PID will be appended to
178 . corename format specifiers:
179 %<NUL> '%' is dropped
187 %e executable filename (may be shortened)
189 %<OTHER> both are dropped
190 . If the first character of the pattern is a '|', the kernel will treat
191 the rest of the pattern as a command to run. The core dump will be
192 written to the standard input of that program instead of to a file.
194 ==============================================================
198 This sysctl is only applicable when core_pattern is configured to pipe
199 core files to a user space helper (when the first character of
200 core_pattern is a '|', see above). When collecting cores via a pipe
201 to an application, it is occasionally useful for the collecting
202 application to gather data about the crashing process from its
203 /proc/pid directory. In order to do this safely, the kernel must wait
204 for the collecting process to exit, so as not to remove the crashing
205 processes proc files prematurely. This in turn creates the
206 possibility that a misbehaving userspace collecting process can block
207 the reaping of a crashed process simply by never exiting. This sysctl
208 defends against that. It defines how many concurrent crashing
209 processes may be piped to user space applications in parallel. If
210 this value is exceeded, then those crashing processes above that value
211 are noted via the kernel log and their cores are skipped. 0 is a
212 special value, indicating that unlimited processes may be captured in
213 parallel, but that no waiting will take place (i.e. the collecting
214 process is not guaranteed access to /proc/<crashing pid>/). This
217 ==============================================================
221 The default coredump filename is "core". By setting
222 core_uses_pid to 1, the coredump filename becomes core.PID.
223 If core_pattern does not include "%p" (default does not)
224 and core_uses_pid is set, then .PID will be appended to
227 ==============================================================
231 When the value in this file is 0, ctrl-alt-del is trapped and
232 sent to the init(1) program to handle a graceful restart.
233 When, however, the value is > 0, Linux's reaction to a Vulcan
234 Nerve Pinch (tm) will be an immediate reboot, without even
235 syncing its dirty buffers.
237 Note: when a program (like dosemu) has the keyboard in 'raw'
238 mode, the ctrl-alt-del is intercepted by the program before it
239 ever reaches the kernel tty layer, and it's up to the program
240 to decide what to do with it.
242 ==============================================================
246 This toggle indicates whether unprivileged users are prevented
247 from using dmesg(8) to view messages from the kernel's log buffer.
248 When dmesg_restrict is set to (0) there are no restrictions. When
249 dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
252 The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
253 default value of dmesg_restrict.
255 ==============================================================
257 domainname & hostname:
259 These files can be used to set the NIS/YP domainname and the
260 hostname of your box in exactly the same way as the commands
261 domainname and hostname, i.e.:
262 # echo "darkstar" > /proc/sys/kernel/hostname
263 # echo "mydomain" > /proc/sys/kernel/domainname
264 has the same effect as
265 # hostname "darkstar"
266 # domainname "mydomain"
268 Note, however, that the classic darkstar.frop.org has the
269 hostname "darkstar" and DNS (Internet Domain Name Server)
270 domainname "frop.org", not to be confused with the NIS (Network
271 Information Service) or YP (Yellow Pages) domainname. These two
272 domain names are in general different. For a detailed discussion
273 see the hostname(1) man page.
275 ==============================================================
279 Path for the hotplug policy agent.
280 Default value is "/sbin/hotplug".
282 ==============================================================
286 This toggle indicates whether restrictions are placed on
287 exposing kernel addresses via /proc and other interfaces. When
288 kptr_restrict is set to (0), there are no restrictions. When
289 kptr_restrict is set to (1), the default, kernel pointers
290 printed using the %pK format specifier will be replaced with 0's
291 unless the user has CAP_SYSLOG. When kptr_restrict is set to
292 (2), kernel pointers printed using %pK will be replaced with 0's
293 regardless of privileges.
295 ==============================================================
297 kstack_depth_to_print: (X86 only)
299 Controls the number of words to print when dumping the raw
302 ==============================================================
306 This flag controls the L2 cache of G3 processor boards. If
307 0, the cache is disabled. Enabled if nonzero.
309 ==============================================================
313 A toggle value indicating if modules are allowed to be loaded
314 in an otherwise modular kernel. This toggle defaults to off
315 (0), but can be set true (1). Once true, modules can be
316 neither loaded nor unloaded, and the toggle cannot be set back
319 ==============================================================
323 Enables/Disables the NMI watchdog on x86 systems. When the value is
324 non-zero the NMI watchdog is enabled and will continuously test all
325 online cpus to determine whether or not they are still functioning
326 properly. Currently, passing "nmi_watchdog=" parameter at boot time is
327 required for this function to work.
329 If LAPIC NMI watchdog method is in use (nmi_watchdog=2 kernel
330 parameter), the NMI watchdog shares registers with oprofile. By
331 disabling the NMI watchdog, oprofile may have more registers to
334 ==============================================================
336 osrelease, ostype & version:
343 #5 Wed Feb 25 21:49:24 MET 1998
345 The files osrelease and ostype should be clear enough. Version
346 needs a little more clarification however. The '#5' means that
347 this is the fifth kernel built from this source base and the
348 date behind it indicates the time the kernel was built.
349 The only way to tune these values is to rebuild the kernel :-)
351 ==============================================================
353 overflowgid & overflowuid:
355 if your architecture did not always support 32-bit UIDs (i.e. arm,
356 i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
357 applications that use the old 16-bit UID/GID system calls, if the
358 actual UID or GID would exceed 65535.
360 These sysctls allow you to change the value of the fixed UID and GID.
361 The default is 65534.
363 ==============================================================
367 The value in this file represents the number of seconds the kernel
368 waits before rebooting on a panic. When you use the software watchdog,
369 the recommended setting is 60.
371 ==============================================================
373 panic_on_unrecovered_nmi:
375 The default Linux behaviour on an NMI of either memory or unknown is
376 to continue operation. For many environments such as scientific
377 computing it is preferable that the box is taken out and the error
378 dealt with than an uncorrected parity/ECC error get propagated.
380 A small number of systems do generate NMI's for bizarre random reasons
381 such as power management so the default is off. That sysctl works like
382 the existing panic controls already in that directory.
384 ==============================================================
388 Controls the kernel's behaviour when an oops or BUG is encountered.
390 0: try to continue operation
392 1: panic immediately. If the `panic' sysctl is also non-zero then the
393 machine will be rebooted.
395 ==============================================================
397 panic_on_stackoverflow:
399 Controls the kernel's behavior when detecting the overflows of
400 kernel, IRQ and exception stacks except a user stack.
401 This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
403 0: try to continue operation.
405 1: panic immediately.
407 ==============================================================
412 PID allocation wrap value. When the kernel's next PID value
413 reaches this value, it wraps back to a minimum PID value.
414 PIDs of value pid_max or larger are not allocated.
416 ==============================================================
418 powersave-nap: (PPC only)
420 If set, Linux-PPC will use the 'nap' mode of powersaving,
421 otherwise the 'doze' mode will be used.
423 ==============================================================
427 The four values in printk denote: console_loglevel,
428 default_message_loglevel, minimum_console_loglevel and
429 default_console_loglevel respectively.
431 These values influence printk() behavior when printing or
432 logging error messages. See 'man 2 syslog' for more info on
433 the different loglevels.
435 - console_loglevel: messages with a higher priority than
436 this will be printed to the console
437 - default_message_loglevel: messages without an explicit priority
438 will be printed with this priority
439 - minimum_console_loglevel: minimum (highest) value to which
440 console_loglevel can be set
441 - default_console_loglevel: default value for console_loglevel
443 ==============================================================
447 Delay each printk message in printk_delay milliseconds
449 Value from 0 - 10000 is allowed.
451 ==============================================================
455 Some warning messages are rate limited. printk_ratelimit specifies
456 the minimum length of time between these messages (in jiffies), by
457 default we allow one every 5 seconds.
459 A value of 0 will disable rate limiting.
461 ==============================================================
463 printk_ratelimit_burst:
465 While long term we enforce one message per printk_ratelimit
466 seconds, we do allow a burst of messages to pass through.
467 printk_ratelimit_burst specifies the number of messages we can
468 send before ratelimiting kicks in.
470 ==============================================================
474 This option can be used to select the type of process address
475 space randomization that is used in the system, for architectures
476 that support this feature.
478 0 - Turn the process address space randomization off. This is the
479 default for architectures that do not support this feature anyways,
480 and kernels that are booted with the "norandmaps" parameter.
482 1 - Make the addresses of mmap base, stack and VDSO page randomized.
483 This, among other things, implies that shared libraries will be
484 loaded to random addresses. Also for PIE-linked binaries, the
485 location of code start is randomized. This is the default if the
486 CONFIG_COMPAT_BRK option is enabled.
488 2 - Additionally enable heap randomization. This is the default if
489 CONFIG_COMPAT_BRK is disabled.
491 There are a few legacy applications out there (such as some ancient
492 versions of libc.so.5 from 1996) that assume that brk area starts
493 just after the end of the code+bss. These applications break when
494 start of the brk area is randomized. There are however no known
495 non-legacy applications that would be broken this way, so for most
496 systems it is safe to choose full randomization.
498 Systems with ancient and/or broken binaries should be configured
499 with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
500 address space randomization.
502 ==============================================================
504 reboot-cmd: (Sparc only)
506 ??? This seems to be a way to give an argument to the Sparc
507 ROM/Flash boot loader. Maybe to tell it what to do after
510 ==============================================================
512 rtsig-max & rtsig-nr:
514 The file rtsig-max can be used to tune the maximum number
515 of POSIX realtime (queued) signals that can be outstanding
518 rtsig-nr shows the number of RT signals currently queued.
520 ==============================================================
524 This file shows the size of the generic SCSI (sg) buffer.
525 You can't tune it just yet, but you could change it on
526 compile time by editing include/scsi/sg.h and changing
527 the value of SG_BIG_BUFF.
529 There shouldn't be any reason to change this value. If
530 you can come up with one, you probably know what you
533 ==============================================================
537 This value can be used to query and set the run time limit
538 on the maximum shared memory segment size that can be created.
539 Shared memory segments up to 1Gb are now supported in the
540 kernel. This value defaults to SHMMAX.
542 ==============================================================
546 Linux lets you set resource limits, including how much memory one
547 process can consume, via setrlimit(2). Unfortunately, shared memory
548 segments are allowed to exist without association with any process, and
549 thus might not be counted against any resource limits. If enabled,
550 shared memory segments are automatically destroyed when their attach
551 count becomes zero after a detach or a process termination. It will
552 also destroy segments that were created, but never attached to, on exit
553 from the process. The only use left for IPC_RMID is to immediately
554 destroy an unattached segment. Of course, this breaks the way things are
555 defined, so some applications might stop working. Note that this
556 feature will do you no good unless you also configure your resource
557 limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't
560 Note that if you change this from 0 to 1, already created segments
561 without users and with a dead originative process will be destroyed.
563 ==============================================================
567 This value can be used to lower the softlockup tolerance threshold. The
568 default threshold is 60 seconds. If a cpu is locked up for 60 seconds,
569 the kernel complains. Valid values are 1-60 seconds. Setting this
570 tunable to zero will disable the softlockup detection altogether.
572 ==============================================================
576 Non-zero if the kernel has been tainted. Numeric values, which
577 can be ORed together:
579 1 - A module with a non-GPL license has been loaded, this
580 includes modules with no license.
581 Set by modutils >= 2.4.9 and module-init-tools.
582 2 - A module was force loaded by insmod -f.
583 Set by modutils >= 2.4.9 and module-init-tools.
584 4 - Unsafe SMP processors: SMP with CPUs not designed for SMP.
585 8 - A module was forcibly unloaded from the system by rmmod -f.
586 16 - A hardware machine check error occurred on the system.
587 32 - A bad page was discovered on the system.
588 64 - The user has asked that the system be marked "tainted". This
589 could be because they are running software that directly modifies
590 the hardware, or for other reasons.
591 128 - The system has died.
592 256 - The ACPI DSDT has been overridden with one supplied by the user
593 instead of using the one provided by the hardware.
594 512 - A kernel warning has occurred.
595 1024 - A module from drivers/staging was loaded.
597 ==============================================================
601 The value in this file affects behavior of handling NMI. When the
602 value is non-zero, unknown NMI is trapped and then panic occurs. At
603 that time, kernel debugging information is displayed on console.
605 NMI switch that most IA32 servers have fires unknown NMI up, for
606 example. If a system hangs up, try pressing the NMI switch.