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 ]
36 - hardlockup_all_cpu_backtrace
38 - hung_task_check_count
39 - hung_task_timeout_secs
43 - kstack_depth_to_print [ X86 only ]
45 - modprobe ==> Documentation/debugging-modules.txt
47 - msg_next_id [ sysv ipc ]
58 - panic_on_stackoverflow
59 - panic_on_unrecovered_nmi
62 - perf_cpu_time_max_percent
64 - perf_event_max_stack
65 - perf_event_max_contexts_per_stack
67 - powersave-nap [ PPC only ]
71 - printk_ratelimit_burst
72 - pty ==> Documentation/filesystems/devpts.txt
74 - real-root-dev ==> Documentation/initrd.txt
75 - reboot-cmd [ SPARC only ]
79 - sem_next_id [ sysv ipc ]
80 - sg-big-buff [ generic SCSI device (sg) ]
81 - shm_next_id [ sysv ipc ]
86 - softlockup_all_cpu_backtrace
88 - stop-a [ SPARC only ]
89 - sysrq ==> Documentation/sysrq.txt
90 - sysctl_writes_strict
93 - unprivileged_bpf_disabled
99 ==============================================================
103 highwater lowwater frequency
105 If BSD-style process accounting is enabled these values control
106 its behaviour. If free space on filesystem where the log lives
107 goes below <lowwater>% accounting suspends. If free space gets
108 above <highwater>% accounting resumes. <Frequency> determines
109 how often do we check the amount of free space (value is in
112 That is, suspend accounting if there left <= 2% free; resume it
113 if we got >=4%; consider information about amount of free space
114 valid for 30 seconds.
116 ==============================================================
122 See Doc*/kernel/power/video.txt, it allows mode of video boot to be
125 ==============================================================
129 This variable has no effect and may be removed in future kernel
130 releases. Reading it always returns 0.
131 Up to Linux 3.17, it enabled/disabled automatic recomputing of msgmni
132 upon memory add/remove or upon ipc namespace creation/removal.
133 Echoing "1" into this file enabled msgmni automatic recomputing.
134 Echoing "0" turned it off. auto_msgmni default value was 1.
137 ==============================================================
141 x86 bootloader identification
143 This gives the bootloader type number as indicated by the bootloader,
144 shifted left by 4, and OR'd with the low four bits of the bootloader
145 version. The reason for this encoding is that this used to match the
146 type_of_loader field in the kernel header; the encoding is kept for
147 backwards compatibility. That is, if the full bootloader type number
148 is 0x15 and the full version number is 0x234, this file will contain
149 the value 340 = 0x154.
151 See the type_of_loader and ext_loader_type fields in
152 Documentation/x86/boot.txt for additional information.
154 ==============================================================
158 x86 bootloader version
160 The complete bootloader version number. In the example above, this
161 file will contain the value 564 = 0x234.
163 See the type_of_loader and ext_loader_ver fields in
164 Documentation/x86/boot.txt for additional information.
166 ==============================================================
170 Controls the kernel's callhome behavior in case of a kernel panic.
172 The s390 hardware allows an operating system to send a notification
173 to a service organization (callhome) in case of an operating system panic.
175 When the value in this file is 0 (which is the default behavior)
176 nothing happens in case of a kernel panic. If this value is set to "1"
177 the complete kernel oops message is send to the IBM customer service
178 organization in case the mainframe the Linux operating system is running
179 on has a service contract with IBM.
181 ==============================================================
185 Highest valid capability of the running kernel. Exports
186 CAP_LAST_CAP from the kernel.
188 ==============================================================
192 core_pattern is used to specify a core dumpfile pattern name.
193 . max length 128 characters; default value is "core"
194 . core_pattern is used as a pattern template for the output filename;
195 certain string patterns (beginning with '%') are substituted with
197 . backward compatibility with core_uses_pid:
198 If core_pattern does not include "%p" (default does not)
199 and core_uses_pid is set, then .PID will be appended to
201 . corename format specifiers:
202 %<NUL> '%' is dropped
205 %P global pid (init PID namespace)
207 %I global tid (init PID namespace)
208 %u uid (in initial user namespace)
209 %g gid (in initial user namespace)
210 %d dump mode, matches PR_SET_DUMPABLE and
211 /proc/sys/fs/suid_dumpable
215 %e executable filename (may be shortened)
217 %<OTHER> both are dropped
218 . If the first character of the pattern is a '|', the kernel will treat
219 the rest of the pattern as a command to run. The core dump will be
220 written to the standard input of that program instead of to a file.
222 ==============================================================
226 This sysctl is only applicable when core_pattern is configured to pipe
227 core files to a user space helper (when the first character of
228 core_pattern is a '|', see above). When collecting cores via a pipe
229 to an application, it is occasionally useful for the collecting
230 application to gather data about the crashing process from its
231 /proc/pid directory. In order to do this safely, the kernel must wait
232 for the collecting process to exit, so as not to remove the crashing
233 processes proc files prematurely. This in turn creates the
234 possibility that a misbehaving userspace collecting process can block
235 the reaping of a crashed process simply by never exiting. This sysctl
236 defends against that. It defines how many concurrent crashing
237 processes may be piped to user space applications in parallel. If
238 this value is exceeded, then those crashing processes above that value
239 are noted via the kernel log and their cores are skipped. 0 is a
240 special value, indicating that unlimited processes may be captured in
241 parallel, but that no waiting will take place (i.e. the collecting
242 process is not guaranteed access to /proc/<crashing pid>/). This
245 ==============================================================
249 The default coredump filename is "core". By setting
250 core_uses_pid to 1, the coredump filename becomes core.PID.
251 If core_pattern does not include "%p" (default does not)
252 and core_uses_pid is set, then .PID will be appended to
255 ==============================================================
259 When the value in this file is 0, ctrl-alt-del is trapped and
260 sent to the init(1) program to handle a graceful restart.
261 When, however, the value is > 0, Linux's reaction to a Vulcan
262 Nerve Pinch (tm) will be an immediate reboot, without even
263 syncing its dirty buffers.
265 Note: when a program (like dosemu) has the keyboard in 'raw'
266 mode, the ctrl-alt-del is intercepted by the program before it
267 ever reaches the kernel tty layer, and it's up to the program
268 to decide what to do with it.
270 ==============================================================
274 This toggle indicates whether unprivileged users are prevented
275 from using dmesg(8) to view messages from the kernel's log buffer.
276 When dmesg_restrict is set to (0) there are no restrictions. When
277 dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
280 The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
281 default value of dmesg_restrict.
283 ==============================================================
285 domainname & hostname:
287 These files can be used to set the NIS/YP domainname and the
288 hostname of your box in exactly the same way as the commands
289 domainname and hostname, i.e.:
290 # echo "darkstar" > /proc/sys/kernel/hostname
291 # echo "mydomain" > /proc/sys/kernel/domainname
292 has the same effect as
293 # hostname "darkstar"
294 # domainname "mydomain"
296 Note, however, that the classic darkstar.frop.org has the
297 hostname "darkstar" and DNS (Internet Domain Name Server)
298 domainname "frop.org", not to be confused with the NIS (Network
299 Information Service) or YP (Yellow Pages) domainname. These two
300 domain names are in general different. For a detailed discussion
301 see the hostname(1) man page.
303 ==============================================================
304 hardlockup_all_cpu_backtrace:
306 This value controls the hard lockup detector behavior when a hard
307 lockup condition is detected as to whether or not to gather further
308 debug information. If enabled, arch-specific all-CPU stack dumping
311 0: do nothing. This is the default behavior.
313 1: on detection capture more debug information.
314 ==============================================================
318 Path for the hotplug policy agent.
319 Default value is "/sbin/hotplug".
321 ==============================================================
325 Controls the kernel's behavior when a hung task is detected.
326 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
328 0: continue operation. This is the default behavior.
330 1: panic immediately.
332 ==============================================================
334 hung_task_check_count:
336 The upper bound on the number of tasks that are checked.
337 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
339 ==============================================================
341 hung_task_timeout_secs:
343 Check interval. When a task in D state did not get scheduled
344 for more than this value report a warning.
345 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
347 0: means infinite timeout - no checking done.
348 Possible values to set are in range {0..LONG_MAX/HZ}.
350 ==============================================================
354 The maximum number of warnings to report. During a check interval
355 if a hung task is detected, this value is decreased by 1.
356 When this value reaches 0, no more warnings will be reported.
357 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
359 -1: report an infinite number of warnings.
361 ==============================================================
365 A toggle indicating if the kexec_load syscall has been disabled. This
366 value defaults to 0 (false: kexec_load enabled), but can be set to 1
367 (true: kexec_load disabled). Once true, kexec can no longer be used, and
368 the toggle cannot be set back to false. This allows a kexec image to be
369 loaded before disabling the syscall, allowing a system to set up (and
370 later use) an image without it being altered. Generally used together
371 with the "modules_disabled" sysctl.
373 ==============================================================
377 This toggle indicates whether restrictions are placed on
378 exposing kernel addresses via /proc and other interfaces.
380 When kptr_restrict is set to (0), the default, there are no restrictions.
382 When kptr_restrict is set to (1), kernel pointers printed using the %pK
383 format specifier will be replaced with 0's unless the user has CAP_SYSLOG
384 and effective user and group ids are equal to the real ids. This is
385 because %pK checks are done at read() time rather than open() time, so
386 if permissions are elevated between the open() and the read() (e.g via
387 a setuid binary) then %pK will not leak kernel pointers to unprivileged
388 users. Note, this is a temporary solution only. The correct long-term
389 solution is to do the permission checks at open() time. Consider removing
390 world read permissions from files that use %pK, and using dmesg_restrict
391 to protect against uses of %pK in dmesg(8) if leaking kernel pointer
392 values to unprivileged users is a concern.
394 When kptr_restrict is set to (2), kernel pointers printed using
395 %pK will be replaced with 0's regardless of privileges.
397 ==============================================================
399 kstack_depth_to_print: (X86 only)
401 Controls the number of words to print when dumping the raw
404 ==============================================================
408 This flag controls the L2 cache of G3 processor boards. If
409 0, the cache is disabled. Enabled if nonzero.
411 ==============================================================
415 A toggle value indicating if modules are allowed to be loaded
416 in an otherwise modular kernel. This toggle defaults to off
417 (0), but can be set true (1). Once true, modules can be
418 neither loaded nor unloaded, and the toggle cannot be set back
419 to false. Generally used with the "kexec_load_disabled" toggle.
421 ==============================================================
423 msg_next_id, sem_next_id, and shm_next_id:
425 These three toggles allows to specify desired id for next allocated IPC
426 object: message, semaphore or shared memory respectively.
428 By default they are equal to -1, which means generic allocation logic.
429 Possible values to set are in range {0..INT_MAX}.
432 1) kernel doesn't guarantee, that new object will have desired id. So,
433 it's up to userspace, how to handle an object with "wrong" id.
434 2) Toggle with non-default value will be set back to -1 by kernel after
435 successful IPC object allocation.
437 ==============================================================
441 This parameter can be used to control the NMI watchdog
442 (i.e. the hard lockup detector) on x86 systems.
444 0 - disable the hard lockup detector
445 1 - enable the hard lockup detector
447 The hard lockup detector monitors each CPU for its ability to respond to
448 timer interrupts. The mechanism utilizes CPU performance counter registers
449 that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
450 while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
452 The NMI watchdog is disabled by default if the kernel is running as a guest
453 in a KVM virtual machine. This default can be overridden by adding
457 to the guest kernel command line (see Documentation/kernel-parameters.txt).
459 ==============================================================
463 Enables/disables automatic page fault based NUMA memory
464 balancing. Memory is moved automatically to nodes
465 that access it often.
467 Enables/disables automatic NUMA memory balancing. On NUMA machines, there
468 is a performance penalty if remote memory is accessed by a CPU. When this
469 feature is enabled the kernel samples what task thread is accessing memory
470 by periodically unmapping pages and later trapping a page fault. At the
471 time of the page fault, it is determined if the data being accessed should
472 be migrated to a local memory node.
474 The unmapping of pages and trapping faults incur additional overhead that
475 ideally is offset by improved memory locality but there is no universal
476 guarantee. If the target workload is already bound to NUMA nodes then this
477 feature should be disabled. Otherwise, if the system overhead from the
478 feature is too high then the rate the kernel samples for NUMA hinting
479 faults may be controlled by the numa_balancing_scan_period_min_ms,
480 numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms,
481 numa_balancing_scan_size_mb, and numa_balancing_settle_count sysctls.
483 ==============================================================
485 numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms,
486 numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb
488 Automatic NUMA balancing scans tasks address space and unmaps pages to
489 detect if pages are properly placed or if the data should be migrated to a
490 memory node local to where the task is running. Every "scan delay" the task
491 scans the next "scan size" number of pages in its address space. When the
492 end of the address space is reached the scanner restarts from the beginning.
494 In combination, the "scan delay" and "scan size" determine the scan rate.
495 When "scan delay" decreases, the scan rate increases. The scan delay and
496 hence the scan rate of every task is adaptive and depends on historical
497 behaviour. If pages are properly placed then the scan delay increases,
498 otherwise the scan delay decreases. The "scan size" is not adaptive but
499 the higher the "scan size", the higher the scan rate.
501 Higher scan rates incur higher system overhead as page faults must be
502 trapped and potentially data must be migrated. However, the higher the scan
503 rate, the more quickly a tasks memory is migrated to a local node if the
504 workload pattern changes and minimises performance impact due to remote
505 memory accesses. These sysctls control the thresholds for scan delays and
506 the number of pages scanned.
508 numa_balancing_scan_period_min_ms is the minimum time in milliseconds to
509 scan a tasks virtual memory. It effectively controls the maximum scanning
512 numa_balancing_scan_delay_ms is the starting "scan delay" used for a task
513 when it initially forks.
515 numa_balancing_scan_period_max_ms is the maximum time in milliseconds to
516 scan a tasks virtual memory. It effectively controls the minimum scanning
519 numa_balancing_scan_size_mb is how many megabytes worth of pages are
520 scanned for a given scan.
522 ==============================================================
524 osrelease, ostype & version:
531 #5 Wed Feb 25 21:49:24 MET 1998
533 The files osrelease and ostype should be clear enough. Version
534 needs a little more clarification however. The '#5' means that
535 this is the fifth kernel built from this source base and the
536 date behind it indicates the time the kernel was built.
537 The only way to tune these values is to rebuild the kernel :-)
539 ==============================================================
541 overflowgid & overflowuid:
543 if your architecture did not always support 32-bit UIDs (i.e. arm,
544 i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
545 applications that use the old 16-bit UID/GID system calls, if the
546 actual UID or GID would exceed 65535.
548 These sysctls allow you to change the value of the fixed UID and GID.
549 The default is 65534.
551 ==============================================================
555 The value in this file represents the number of seconds the kernel
556 waits before rebooting on a panic. When you use the software watchdog,
557 the recommended setting is 60.
559 ==============================================================
563 Controls the kernel's behavior when a CPU receives an NMI caused by
566 0: try to continue operation (default)
568 1: panic immediately. The IO error triggered an NMI. This indicates a
569 serious system condition which could result in IO data corruption.
570 Rather than continuing, panicking might be a better choice. Some
571 servers issue this sort of NMI when the dump button is pushed,
572 and you can use this option to take a crash dump.
574 ==============================================================
578 Controls the kernel's behaviour when an oops or BUG is encountered.
580 0: try to continue operation
582 1: panic immediately. If the `panic' sysctl is also non-zero then the
583 machine will be rebooted.
585 ==============================================================
587 panic_on_stackoverflow:
589 Controls the kernel's behavior when detecting the overflows of
590 kernel, IRQ and exception stacks except a user stack.
591 This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
593 0: try to continue operation.
595 1: panic immediately.
597 ==============================================================
599 panic_on_unrecovered_nmi:
601 The default Linux behaviour on an NMI of either memory or unknown is
602 to continue operation. For many environments such as scientific
603 computing it is preferable that the box is taken out and the error
604 dealt with than an uncorrected parity/ECC error get propagated.
606 A small number of systems do generate NMI's for bizarre random reasons
607 such as power management so the default is off. That sysctl works like
608 the existing panic controls already in that directory.
610 ==============================================================
614 Calls panic() in the WARN() path when set to 1. This is useful to avoid
615 a kernel rebuild when attempting to kdump at the location of a WARN().
617 0: only WARN(), default behaviour.
619 1: call panic() after printing out WARN() location.
621 ==============================================================
625 When set to 1, calls panic() after RCU stall detection messages. This
626 is useful to define the root cause of RCU stalls using a vmcore.
628 0: do not panic() when RCU stall takes place, default behavior.
630 1: panic() after printing RCU stall messages.
632 ==============================================================
634 perf_cpu_time_max_percent:
636 Hints to the kernel how much CPU time it should be allowed to
637 use to handle perf sampling events. If the perf subsystem
638 is informed that its samples are exceeding this limit, it
639 will drop its sampling frequency to attempt to reduce its CPU
642 Some perf sampling happens in NMIs. If these samples
643 unexpectedly take too long to execute, the NMIs can become
644 stacked up next to each other so much that nothing else is
647 0: disable the mechanism. Do not monitor or correct perf's
648 sampling rate no matter how CPU time it takes.
650 1-100: attempt to throttle perf's sample rate to this
651 percentage of CPU. Note: the kernel calculates an
652 "expected" length of each sample event. 100 here means
653 100% of that expected length. Even if this is set to
654 100, you may still see sample throttling if this
655 length is exceeded. Set to 0 if you truly do not care
656 how much CPU is consumed.
658 ==============================================================
662 Controls use of the performance events system by unprivileged
663 users (without CAP_SYS_ADMIN). The default value is 2.
665 -1: Allow use of (almost) all events by all users
666 >=0: Disallow raw tracepoint access by users without CAP_IOC_LOCK
667 >=1: Disallow CPU event access by users without CAP_SYS_ADMIN
668 >=2: Disallow kernel profiling by users without CAP_SYS_ADMIN
670 ==============================================================
672 perf_event_max_stack:
674 Controls maximum number of stack frames to copy for (attr.sample_type &
675 PERF_SAMPLE_CALLCHAIN) configured events, for instance, when using
676 'perf record -g' or 'perf trace --call-graph fp'.
678 This can only be done when no events are in use that have callchains
679 enabled, otherwise writing to this file will return -EBUSY.
681 The default value is 127.
683 ==============================================================
685 perf_event_max_contexts_per_stack:
687 Controls maximum number of stack frame context entries for
688 (attr.sample_type & PERF_SAMPLE_CALLCHAIN) configured events, for
689 instance, when using 'perf record -g' or 'perf trace --call-graph fp'.
691 This can only be done when no events are in use that have callchains
692 enabled, otherwise writing to this file will return -EBUSY.
694 The default value is 8.
696 ==============================================================
700 PID allocation wrap value. When the kernel's next PID value
701 reaches this value, it wraps back to a minimum PID value.
702 PIDs of value pid_max or larger are not allocated.
704 ==============================================================
708 The last pid allocated in the current (the one task using this sysctl
709 lives in) pid namespace. When selecting a pid for a next task on fork
710 kernel tries to allocate a number starting from this one.
712 ==============================================================
714 powersave-nap: (PPC only)
716 If set, Linux-PPC will use the 'nap' mode of powersaving,
717 otherwise the 'doze' mode will be used.
719 ==============================================================
723 The four values in printk denote: console_loglevel,
724 default_message_loglevel, minimum_console_loglevel and
725 default_console_loglevel respectively.
727 These values influence printk() behavior when printing or
728 logging error messages. See 'man 2 syslog' for more info on
729 the different loglevels.
731 - console_loglevel: messages with a higher priority than
732 this will be printed to the console
733 - default_message_loglevel: messages without an explicit priority
734 will be printed with this priority
735 - minimum_console_loglevel: minimum (highest) value to which
736 console_loglevel can be set
737 - default_console_loglevel: default value for console_loglevel
739 ==============================================================
743 Delay each printk message in printk_delay milliseconds
745 Value from 0 - 10000 is allowed.
747 ==============================================================
751 Some warning messages are rate limited. printk_ratelimit specifies
752 the minimum length of time between these messages (in jiffies), by
753 default we allow one every 5 seconds.
755 A value of 0 will disable rate limiting.
757 ==============================================================
759 printk_ratelimit_burst:
761 While long term we enforce one message per printk_ratelimit
762 seconds, we do allow a burst of messages to pass through.
763 printk_ratelimit_burst specifies the number of messages we can
764 send before ratelimiting kicks in.
766 ==============================================================
770 Control the logging to /dev/kmsg from userspace:
772 ratelimit: default, ratelimited
773 on: unlimited logging to /dev/kmsg from userspace
774 off: logging to /dev/kmsg disabled
776 The kernel command line parameter printk.devkmsg= overrides this and is
777 a one-time setting until next reboot: once set, it cannot be changed by
778 this sysctl interface anymore.
780 ==============================================================
784 This option can be used to select the type of process address
785 space randomization that is used in the system, for architectures
786 that support this feature.
788 0 - Turn the process address space randomization off. This is the
789 default for architectures that do not support this feature anyways,
790 and kernels that are booted with the "norandmaps" parameter.
792 1 - Make the addresses of mmap base, stack and VDSO page randomized.
793 This, among other things, implies that shared libraries will be
794 loaded to random addresses. Also for PIE-linked binaries, the
795 location of code start is randomized. This is the default if the
796 CONFIG_COMPAT_BRK option is enabled.
798 2 - Additionally enable heap randomization. This is the default if
799 CONFIG_COMPAT_BRK is disabled.
801 There are a few legacy applications out there (such as some ancient
802 versions of libc.so.5 from 1996) that assume that brk area starts
803 just after the end of the code+bss. These applications break when
804 start of the brk area is randomized. There are however no known
805 non-legacy applications that would be broken this way, so for most
806 systems it is safe to choose full randomization.
808 Systems with ancient and/or broken binaries should be configured
809 with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
810 address space randomization.
812 ==============================================================
814 reboot-cmd: (Sparc only)
816 ??? This seems to be a way to give an argument to the Sparc
817 ROM/Flash boot loader. Maybe to tell it what to do after
820 ==============================================================
822 rtsig-max & rtsig-nr:
824 The file rtsig-max can be used to tune the maximum number
825 of POSIX realtime (queued) signals that can be outstanding
828 rtsig-nr shows the number of RT signals currently queued.
830 ==============================================================
834 Enables/disables scheduler statistics. Enabling this feature
835 incurs a small amount of overhead in the scheduler but is
836 useful for debugging and performance tuning.
838 ==============================================================
842 This file shows the size of the generic SCSI (sg) buffer.
843 You can't tune it just yet, but you could change it on
844 compile time by editing include/scsi/sg.h and changing
845 the value of SG_BIG_BUFF.
847 There shouldn't be any reason to change this value. If
848 you can come up with one, you probably know what you
851 ==============================================================
855 This parameter sets the total amount of shared memory pages that
856 can be used system wide. Hence, SHMALL should always be at least
857 ceil(shmmax/PAGE_SIZE).
859 If you are not sure what the default PAGE_SIZE is on your Linux
860 system, you can run the following command:
864 ==============================================================
868 This value can be used to query and set the run time limit
869 on the maximum shared memory segment size that can be created.
870 Shared memory segments up to 1Gb are now supported in the
871 kernel. This value defaults to SHMMAX.
873 ==============================================================
877 Linux lets you set resource limits, including how much memory one
878 process can consume, via setrlimit(2). Unfortunately, shared memory
879 segments are allowed to exist without association with any process, and
880 thus might not be counted against any resource limits. If enabled,
881 shared memory segments are automatically destroyed when their attach
882 count becomes zero after a detach or a process termination. It will
883 also destroy segments that were created, but never attached to, on exit
884 from the process. The only use left for IPC_RMID is to immediately
885 destroy an unattached segment. Of course, this breaks the way things are
886 defined, so some applications might stop working. Note that this
887 feature will do you no good unless you also configure your resource
888 limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't
891 Note that if you change this from 0 to 1, already created segments
892 without users and with a dead originative process will be destroyed.
894 ==============================================================
896 sysctl_writes_strict:
898 Control how file position affects the behavior of updating sysctl values
899 via the /proc/sys interface:
901 -1 - Legacy per-write sysctl value handling, with no printk warnings.
902 Each write syscall must fully contain the sysctl value to be
903 written, and multiple writes on the same sysctl file descriptor
904 will rewrite the sysctl value, regardless of file position.
905 0 - Same behavior as above, but warn about processes that perform writes
906 to a sysctl file descriptor when the file position is not 0.
907 1 - (default) Respect file position when writing sysctl strings. Multiple
908 writes will append to the sysctl value buffer. Anything past the max
909 length of the sysctl value buffer will be ignored. Writes to numeric
910 sysctl entries must always be at file position 0 and the value must
911 be fully contained in the buffer sent in the write syscall.
913 ==============================================================
915 softlockup_all_cpu_backtrace:
917 This value controls the soft lockup detector thread's behavior
918 when a soft lockup condition is detected as to whether or not
919 to gather further debug information. If enabled, each cpu will
920 be issued an NMI and instructed to capture stack trace.
922 This feature is only applicable for architectures which support
925 0: do nothing. This is the default behavior.
927 1: on detection capture more debug information.
929 ==============================================================
933 This parameter can be used to control the soft lockup detector.
935 0 - disable the soft lockup detector
936 1 - enable the soft lockup detector
938 The soft lockup detector monitors CPUs for threads that are hogging the CPUs
939 without rescheduling voluntarily, and thus prevent the 'watchdog/N' threads
940 from running. The mechanism depends on the CPUs ability to respond to timer
941 interrupts which are needed for the 'watchdog/N' threads to be woken up by
942 the watchdog timer function, otherwise the NMI watchdog - if enabled - can
943 detect a hard lockup condition.
945 ==============================================================
949 Non-zero if the kernel has been tainted. Numeric values, which
950 can be ORed together:
952 1 - A module with a non-GPL license has been loaded, this
953 includes modules with no license.
954 Set by modutils >= 2.4.9 and module-init-tools.
955 2 - A module was force loaded by insmod -f.
956 Set by modutils >= 2.4.9 and module-init-tools.
957 4 - Unsafe SMP processors: SMP with CPUs not designed for SMP.
958 8 - A module was forcibly unloaded from the system by rmmod -f.
959 16 - A hardware machine check error occurred on the system.
960 32 - A bad page was discovered on the system.
961 64 - The user has asked that the system be marked "tainted". This
962 could be because they are running software that directly modifies
963 the hardware, or for other reasons.
964 128 - The system has died.
965 256 - The ACPI DSDT has been overridden with one supplied by the user
966 instead of using the one provided by the hardware.
967 512 - A kernel warning has occurred.
968 1024 - A module from drivers/staging was loaded.
969 2048 - The system is working around a severe firmware bug.
970 4096 - An out-of-tree module has been loaded.
971 8192 - An unsigned module has been loaded in a kernel supporting module
973 16384 - A soft lockup has previously occurred on the system.
974 32768 - The kernel has been live patched.
976 ==============================================================
980 This value controls the maximum number of threads that can be created
983 During initialization the kernel sets this value such that even if the
984 maximum number of threads is created, the thread structures occupy only
985 a part (1/8th) of the available RAM pages.
987 The minimum value that can be written to threads-max is 20.
988 The maximum value that can be written to threads-max is given by the
989 constant FUTEX_TID_MASK (0x3fffffff).
990 If a value outside of this range is written to threads-max an error
993 The value written is checked against the available RAM pages. If the
994 thread structures would occupy too much (more than 1/8th) of the
995 available RAM pages threads-max is reduced accordingly.
997 ==============================================================
999 unprivileged_bpf_disabled:
1001 Writing 1 to this entry will disable unprivileged calls to bpf();
1002 once disabled, calling bpf() without CAP_SYS_ADMIN will return
1003 -EPERM. Once set to 1, this can't be cleared from the running kernel
1006 Writing 2 to this entry will also disable unprivileged calls to bpf(),
1007 however, an admin can still change this setting later on, if needed, by
1008 writing 0 or 1 to this entry.
1010 If BPF_UNPRIV_DEFAULT_OFF is enabled in the kernel config, then this
1011 entry will default to 2 instead of 0.
1013 0 - Unprivileged calls to bpf() are enabled
1014 1 - Unprivileged calls to bpf() are disabled without recovery
1015 2 - Unprivileged calls to bpf() are disabled
1017 ==============================================================
1021 The value in this file affects behavior of handling NMI. When the
1022 value is non-zero, unknown NMI is trapped and then panic occurs. At
1023 that time, kernel debugging information is displayed on console.
1025 NMI switch that most IA32 servers have fires unknown NMI up, for
1026 example. If a system hangs up, try pressing the NMI switch.
1028 ==============================================================
1032 This parameter can be used to disable or enable the soft lockup detector
1033 _and_ the NMI watchdog (i.e. the hard lockup detector) at the same time.
1035 0 - disable both lockup detectors
1036 1 - enable both lockup detectors
1038 The soft lockup detector and the NMI watchdog can also be disabled or
1039 enabled individually, using the soft_watchdog and nmi_watchdog parameters.
1040 If the watchdog parameter is read, for example by executing
1042 cat /proc/sys/kernel/watchdog
1044 the output of this command (0 or 1) shows the logical OR of soft_watchdog
1047 ==============================================================
1051 This value can be used to control on which cpus the watchdog may run.
1052 The default cpumask is all possible cores, but if NO_HZ_FULL is
1053 enabled in the kernel config, and cores are specified with the
1054 nohz_full= boot argument, those cores are excluded by default.
1055 Offline cores can be included in this mask, and if the core is later
1056 brought online, the watchdog will be started based on the mask value.
1058 Typically this value would only be touched in the nohz_full case
1059 to re-enable cores that by default were not running the watchdog,
1060 if a kernel lockup was suspected on those cores.
1062 The argument value is the standard cpulist format for cpumasks,
1063 so for example to enable the watchdog on cores 0, 2, 3, and 4 you
1066 echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
1068 ==============================================================
1072 This value can be used to control the frequency of hrtimer and NMI
1073 events and the soft and hard lockup thresholds. The default threshold
1076 The softlockup threshold is (2 * watchdog_thresh). Setting this
1077 tunable to zero will disable lockup detection altogether.
1079 ==============================================================