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
39 - hung_task_check_count
40 - hung_task_timeout_secs
41 - hung_task_check_interval_secs
43 - hyperv_record_panic_msg
47 - modprobe ==> Documentation/debugging-modules.txt
49 - msg_next_id [ sysv ipc ]
60 - panic_on_stackoverflow
61 - panic_on_unrecovered_nmi
64 - perf_cpu_time_max_percent
66 - perf_event_max_stack
68 - perf_event_max_contexts_per_stack
70 - powersave-nap [ PPC only ]
74 - printk_ratelimit_burst
75 - pty ==> Documentation/filesystems/devpts.txt
77 - real-root-dev ==> Documentation/admin-guide/initrd.rst
78 - reboot-cmd [ SPARC only ]
81 - seccomp/ ==> Documentation/userspace-api/seccomp_filter.rst
83 - sem_next_id [ sysv ipc ]
84 - sg-big-buff [ generic SCSI device (sg) ]
85 - shm_next_id [ sysv ipc ]
90 - softlockup_all_cpu_backtrace
92 - stop-a [ SPARC only ]
93 - sysrq ==> Documentation/admin-guide/sysrq.rst
94 - sysctl_writes_strict
97 - unprivileged_bpf_disabled
103 ==============================================================
107 highwater lowwater frequency
109 If BSD-style process accounting is enabled these values control
110 its behaviour. If free space on filesystem where the log lives
111 goes below <lowwater>% accounting suspends. If free space gets
112 above <highwater>% accounting resumes. <Frequency> determines
113 how often do we check the amount of free space (value is in
116 That is, suspend accounting if there left <= 2% free; resume it
117 if we got >=4%; consider information about amount of free space
118 valid for 30 seconds.
120 ==============================================================
126 See Doc*/kernel/power/video.txt, it allows mode of video boot to be
129 ==============================================================
133 This variable has no effect and may be removed in future kernel
134 releases. Reading it always returns 0.
135 Up to Linux 3.17, it enabled/disabled automatic recomputing of msgmni
136 upon memory add/remove or upon ipc namespace creation/removal.
137 Echoing "1" into this file enabled msgmni automatic recomputing.
138 Echoing "0" turned it off. auto_msgmni default value was 1.
141 ==============================================================
145 x86 bootloader identification
147 This gives the bootloader type number as indicated by the bootloader,
148 shifted left by 4, and OR'd with the low four bits of the bootloader
149 version. The reason for this encoding is that this used to match the
150 type_of_loader field in the kernel header; the encoding is kept for
151 backwards compatibility. That is, if the full bootloader type number
152 is 0x15 and the full version number is 0x234, this file will contain
153 the value 340 = 0x154.
155 See the type_of_loader and ext_loader_type fields in
156 Documentation/x86/boot.txt for additional information.
158 ==============================================================
162 x86 bootloader version
164 The complete bootloader version number. In the example above, this
165 file will contain the value 564 = 0x234.
167 See the type_of_loader and ext_loader_ver fields in
168 Documentation/x86/boot.txt for additional information.
170 ==============================================================
174 Controls the kernel's callhome behavior in case of a kernel panic.
176 The s390 hardware allows an operating system to send a notification
177 to a service organization (callhome) in case of an operating system panic.
179 When the value in this file is 0 (which is the default behavior)
180 nothing happens in case of a kernel panic. If this value is set to "1"
181 the complete kernel oops message is send to the IBM customer service
182 organization in case the mainframe the Linux operating system is running
183 on has a service contract with IBM.
185 ==============================================================
189 Highest valid capability of the running kernel. Exports
190 CAP_LAST_CAP from the kernel.
192 ==============================================================
196 core_pattern is used to specify a core dumpfile pattern name.
197 . max length 128 characters; default value is "core"
198 . core_pattern is used as a pattern template for the output filename;
199 certain string patterns (beginning with '%') are substituted with
201 . backward compatibility with core_uses_pid:
202 If core_pattern does not include "%p" (default does not)
203 and core_uses_pid is set, then .PID will be appended to
205 . corename format specifiers:
206 %<NUL> '%' is dropped
209 %P global pid (init PID namespace)
211 %I global tid (init PID namespace)
212 %u uid (in initial user namespace)
213 %g gid (in initial user namespace)
214 %d dump mode, matches PR_SET_DUMPABLE and
215 /proc/sys/fs/suid_dumpable
219 %e executable filename (may be shortened)
221 %<OTHER> both are dropped
222 . If the first character of the pattern is a '|', the kernel will treat
223 the rest of the pattern as a command to run. The core dump will be
224 written to the standard input of that program instead of to a file.
226 ==============================================================
230 This sysctl is only applicable when core_pattern is configured to pipe
231 core files to a user space helper (when the first character of
232 core_pattern is a '|', see above). When collecting cores via a pipe
233 to an application, it is occasionally useful for the collecting
234 application to gather data about the crashing process from its
235 /proc/pid directory. In order to do this safely, the kernel must wait
236 for the collecting process to exit, so as not to remove the crashing
237 processes proc files prematurely. This in turn creates the
238 possibility that a misbehaving userspace collecting process can block
239 the reaping of a crashed process simply by never exiting. This sysctl
240 defends against that. It defines how many concurrent crashing
241 processes may be piped to user space applications in parallel. If
242 this value is exceeded, then those crashing processes above that value
243 are noted via the kernel log and their cores are skipped. 0 is a
244 special value, indicating that unlimited processes may be captured in
245 parallel, but that no waiting will take place (i.e. the collecting
246 process is not guaranteed access to /proc/<crashing pid>/). This
249 ==============================================================
253 The default coredump filename is "core". By setting
254 core_uses_pid to 1, the coredump filename becomes core.PID.
255 If core_pattern does not include "%p" (default does not)
256 and core_uses_pid is set, then .PID will be appended to
259 ==============================================================
263 When the value in this file is 0, ctrl-alt-del is trapped and
264 sent to the init(1) program to handle a graceful restart.
265 When, however, the value is > 0, Linux's reaction to a Vulcan
266 Nerve Pinch (tm) will be an immediate reboot, without even
267 syncing its dirty buffers.
269 Note: when a program (like dosemu) has the keyboard in 'raw'
270 mode, the ctrl-alt-del is intercepted by the program before it
271 ever reaches the kernel tty layer, and it's up to the program
272 to decide what to do with it.
274 ==============================================================
278 This toggle indicates whether unprivileged users are prevented
279 from using dmesg(8) to view messages from the kernel's log buffer.
280 When dmesg_restrict is set to (0) there are no restrictions. When
281 dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
284 The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
285 default value of dmesg_restrict.
287 ==============================================================
289 domainname & hostname:
291 These files can be used to set the NIS/YP domainname and the
292 hostname of your box in exactly the same way as the commands
293 domainname and hostname, i.e.:
294 # echo "darkstar" > /proc/sys/kernel/hostname
295 # echo "mydomain" > /proc/sys/kernel/domainname
296 has the same effect as
297 # hostname "darkstar"
298 # domainname "mydomain"
300 Note, however, that the classic darkstar.frop.org has the
301 hostname "darkstar" and DNS (Internet Domain Name Server)
302 domainname "frop.org", not to be confused with the NIS (Network
303 Information Service) or YP (Yellow Pages) domainname. These two
304 domain names are in general different. For a detailed discussion
305 see the hostname(1) man page.
307 ==============================================================
308 hardlockup_all_cpu_backtrace:
310 This value controls the hard lockup detector behavior when a hard
311 lockup condition is detected as to whether or not to gather further
312 debug information. If enabled, arch-specific all-CPU stack dumping
315 0: do nothing. This is the default behavior.
317 1: on detection capture more debug information.
318 ==============================================================
322 This parameter can be used to control whether the kernel panics
323 when a hard lockup is detected.
325 0 - don't panic on hard lockup
326 1 - panic on hard lockup
328 See Documentation/lockup-watchdogs.txt for more information. This can
329 also be set using the nmi_watchdog kernel parameter.
331 ==============================================================
335 Path for the hotplug policy agent.
336 Default value is "/sbin/hotplug".
338 ==============================================================
342 Controls the kernel's behavior when a hung task is detected.
343 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
345 0: continue operation. This is the default behavior.
347 1: panic immediately.
349 ==============================================================
351 hung_task_check_count:
353 The upper bound on the number of tasks that are checked.
354 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
356 ==============================================================
358 hung_task_timeout_secs:
360 When a task in D state did not get scheduled
361 for more than this value report a warning.
362 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
364 0: means infinite timeout - no checking done.
365 Possible values to set are in range {0..LONG_MAX/HZ}.
367 ==============================================================
369 hung_task_check_interval_secs:
371 Hung task check interval. If hung task checking is enabled
372 (see hung_task_timeout_secs), the check is done every
373 hung_task_check_interval_secs seconds.
374 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
376 0 (default): means use hung_task_timeout_secs as checking interval.
377 Possible values to set are in range {0..LONG_MAX/HZ}.
379 ==============================================================
383 The maximum number of warnings to report. During a check interval
384 if a hung task is detected, this value is decreased by 1.
385 When this value reaches 0, no more warnings will be reported.
386 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
388 -1: report an infinite number of warnings.
390 ==============================================================
392 hyperv_record_panic_msg:
394 Controls whether the panic kmsg data should be reported to Hyper-V.
396 0: do not report panic kmsg data.
398 1: report the panic kmsg data. This is the default behavior.
400 ==============================================================
404 A toggle indicating if the kexec_load syscall has been disabled. This
405 value defaults to 0 (false: kexec_load enabled), but can be set to 1
406 (true: kexec_load disabled). Once true, kexec can no longer be used, and
407 the toggle cannot be set back to false. This allows a kexec image to be
408 loaded before disabling the syscall, allowing a system to set up (and
409 later use) an image without it being altered. Generally used together
410 with the "modules_disabled" sysctl.
412 ==============================================================
416 This toggle indicates whether restrictions are placed on
417 exposing kernel addresses via /proc and other interfaces.
419 When kptr_restrict is set to 0 (the default) the address is hashed before
420 printing. (This is the equivalent to %p.)
422 When kptr_restrict is set to (1), kernel pointers printed using the %pK
423 format specifier will be replaced with 0's unless the user has CAP_SYSLOG
424 and effective user and group ids are equal to the real ids. This is
425 because %pK checks are done at read() time rather than open() time, so
426 if permissions are elevated between the open() and the read() (e.g via
427 a setuid binary) then %pK will not leak kernel pointers to unprivileged
428 users. Note, this is a temporary solution only. The correct long-term
429 solution is to do the permission checks at open() time. Consider removing
430 world read permissions from files that use %pK, and using dmesg_restrict
431 to protect against uses of %pK in dmesg(8) if leaking kernel pointer
432 values to unprivileged users is a concern.
434 When kptr_restrict is set to (2), kernel pointers printed using
435 %pK will be replaced with 0's regardless of privileges.
437 ==============================================================
441 This flag controls the L2 cache of G3 processor boards. If
442 0, the cache is disabled. Enabled if nonzero.
444 ==============================================================
448 A toggle value indicating if modules are allowed to be loaded
449 in an otherwise modular kernel. This toggle defaults to off
450 (0), but can be set true (1). Once true, modules can be
451 neither loaded nor unloaded, and the toggle cannot be set back
452 to false. Generally used with the "kexec_load_disabled" toggle.
454 ==============================================================
456 msg_next_id, sem_next_id, and shm_next_id:
458 These three toggles allows to specify desired id for next allocated IPC
459 object: message, semaphore or shared memory respectively.
461 By default they are equal to -1, which means generic allocation logic.
462 Possible values to set are in range {0..INT_MAX}.
465 1) kernel doesn't guarantee, that new object will have desired id. So,
466 it's up to userspace, how to handle an object with "wrong" id.
467 2) Toggle with non-default value will be set back to -1 by kernel after
468 successful IPC object allocation. If an IPC object allocation syscall
469 fails, it is undefined if the value remains unmodified or is reset to -1.
471 ==============================================================
475 This parameter can be used to control the NMI watchdog
476 (i.e. the hard lockup detector) on x86 systems.
478 0 - disable the hard lockup detector
479 1 - enable the hard lockup detector
481 The hard lockup detector monitors each CPU for its ability to respond to
482 timer interrupts. The mechanism utilizes CPU performance counter registers
483 that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
484 while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
486 The NMI watchdog is disabled by default if the kernel is running as a guest
487 in a KVM virtual machine. This default can be overridden by adding
491 to the guest kernel command line (see Documentation/admin-guide/kernel-parameters.rst).
493 ==============================================================
497 Enables/disables automatic page fault based NUMA memory
498 balancing. Memory is moved automatically to nodes
499 that access it often.
501 Enables/disables automatic NUMA memory balancing. On NUMA machines, there
502 is a performance penalty if remote memory is accessed by a CPU. When this
503 feature is enabled the kernel samples what task thread is accessing memory
504 by periodically unmapping pages and later trapping a page fault. At the
505 time of the page fault, it is determined if the data being accessed should
506 be migrated to a local memory node.
508 The unmapping of pages and trapping faults incur additional overhead that
509 ideally is offset by improved memory locality but there is no universal
510 guarantee. If the target workload is already bound to NUMA nodes then this
511 feature should be disabled. Otherwise, if the system overhead from the
512 feature is too high then the rate the kernel samples for NUMA hinting
513 faults may be controlled by the numa_balancing_scan_period_min_ms,
514 numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms,
515 numa_balancing_scan_size_mb, and numa_balancing_settle_count sysctls.
517 ==============================================================
519 numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms,
520 numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb
522 Automatic NUMA balancing scans tasks address space and unmaps pages to
523 detect if pages are properly placed or if the data should be migrated to a
524 memory node local to where the task is running. Every "scan delay" the task
525 scans the next "scan size" number of pages in its address space. When the
526 end of the address space is reached the scanner restarts from the beginning.
528 In combination, the "scan delay" and "scan size" determine the scan rate.
529 When "scan delay" decreases, the scan rate increases. The scan delay and
530 hence the scan rate of every task is adaptive and depends on historical
531 behaviour. If pages are properly placed then the scan delay increases,
532 otherwise the scan delay decreases. The "scan size" is not adaptive but
533 the higher the "scan size", the higher the scan rate.
535 Higher scan rates incur higher system overhead as page faults must be
536 trapped and potentially data must be migrated. However, the higher the scan
537 rate, the more quickly a tasks memory is migrated to a local node if the
538 workload pattern changes and minimises performance impact due to remote
539 memory accesses. These sysctls control the thresholds for scan delays and
540 the number of pages scanned.
542 numa_balancing_scan_period_min_ms is the minimum time in milliseconds to
543 scan a tasks virtual memory. It effectively controls the maximum scanning
546 numa_balancing_scan_delay_ms is the starting "scan delay" used for a task
547 when it initially forks.
549 numa_balancing_scan_period_max_ms is the maximum time in milliseconds to
550 scan a tasks virtual memory. It effectively controls the minimum scanning
553 numa_balancing_scan_size_mb is how many megabytes worth of pages are
554 scanned for a given scan.
556 ==============================================================
558 osrelease, ostype & version:
565 #5 Wed Feb 25 21:49:24 MET 1998
567 The files osrelease and ostype should be clear enough. Version
568 needs a little more clarification however. The '#5' means that
569 this is the fifth kernel built from this source base and the
570 date behind it indicates the time the kernel was built.
571 The only way to tune these values is to rebuild the kernel :-)
573 ==============================================================
575 overflowgid & overflowuid:
577 if your architecture did not always support 32-bit UIDs (i.e. arm,
578 i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
579 applications that use the old 16-bit UID/GID system calls, if the
580 actual UID or GID would exceed 65535.
582 These sysctls allow you to change the value of the fixed UID and GID.
583 The default is 65534.
585 ==============================================================
589 The value in this file represents the number of seconds the kernel
590 waits before rebooting on a panic. When you use the software watchdog,
591 the recommended setting is 60.
593 ==============================================================
597 Controls the kernel's behavior when a CPU receives an NMI caused by
600 0: try to continue operation (default)
602 1: panic immediately. The IO error triggered an NMI. This indicates a
603 serious system condition which could result in IO data corruption.
604 Rather than continuing, panicking might be a better choice. Some
605 servers issue this sort of NMI when the dump button is pushed,
606 and you can use this option to take a crash dump.
608 ==============================================================
612 Controls the kernel's behaviour when an oops or BUG is encountered.
614 0: try to continue operation
616 1: panic immediately. If the `panic' sysctl is also non-zero then the
617 machine will be rebooted.
619 ==============================================================
621 panic_on_stackoverflow:
623 Controls the kernel's behavior when detecting the overflows of
624 kernel, IRQ and exception stacks except a user stack.
625 This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
627 0: try to continue operation.
629 1: panic immediately.
631 ==============================================================
633 panic_on_unrecovered_nmi:
635 The default Linux behaviour on an NMI of either memory or unknown is
636 to continue operation. For many environments such as scientific
637 computing it is preferable that the box is taken out and the error
638 dealt with than an uncorrected parity/ECC error get propagated.
640 A small number of systems do generate NMI's for bizarre random reasons
641 such as power management so the default is off. That sysctl works like
642 the existing panic controls already in that directory.
644 ==============================================================
648 Calls panic() in the WARN() path when set to 1. This is useful to avoid
649 a kernel rebuild when attempting to kdump at the location of a WARN().
651 0: only WARN(), default behaviour.
653 1: call panic() after printing out WARN() location.
655 ==============================================================
659 When set to 1, calls panic() after RCU stall detection messages. This
660 is useful to define the root cause of RCU stalls using a vmcore.
662 0: do not panic() when RCU stall takes place, default behavior.
664 1: panic() after printing RCU stall messages.
666 ==============================================================
668 perf_cpu_time_max_percent:
670 Hints to the kernel how much CPU time it should be allowed to
671 use to handle perf sampling events. If the perf subsystem
672 is informed that its samples are exceeding this limit, it
673 will drop its sampling frequency to attempt to reduce its CPU
676 Some perf sampling happens in NMIs. If these samples
677 unexpectedly take too long to execute, the NMIs can become
678 stacked up next to each other so much that nothing else is
681 0: disable the mechanism. Do not monitor or correct perf's
682 sampling rate no matter how CPU time it takes.
684 1-100: attempt to throttle perf's sample rate to this
685 percentage of CPU. Note: the kernel calculates an
686 "expected" length of each sample event. 100 here means
687 100% of that expected length. Even if this is set to
688 100, you may still see sample throttling if this
689 length is exceeded. Set to 0 if you truly do not care
690 how much CPU is consumed.
692 ==============================================================
696 Controls use of the performance events system by unprivileged
697 users (without CAP_SYS_ADMIN). The default value is 2.
699 -1: Allow use of (almost) all events by all users
700 Ignore mlock limit after perf_event_mlock_kb without CAP_IPC_LOCK
701 >=0: Disallow ftrace function tracepoint by users without CAP_SYS_ADMIN
702 Disallow raw tracepoint access by users without CAP_SYS_ADMIN
703 >=1: Disallow CPU event access by users without CAP_SYS_ADMIN
704 >=2: Disallow kernel profiling by users without CAP_SYS_ADMIN
706 ==============================================================
708 perf_event_max_stack:
710 Controls maximum number of stack frames to copy for (attr.sample_type &
711 PERF_SAMPLE_CALLCHAIN) configured events, for instance, when using
712 'perf record -g' or 'perf trace --call-graph fp'.
714 This can only be done when no events are in use that have callchains
715 enabled, otherwise writing to this file will return -EBUSY.
717 The default value is 127.
719 ==============================================================
723 Control size of per-cpu ring buffer not counted agains mlock limit.
725 The default value is 512 + 1 page
727 ==============================================================
729 perf_event_max_contexts_per_stack:
731 Controls maximum number of stack frame context entries for
732 (attr.sample_type & PERF_SAMPLE_CALLCHAIN) configured events, for
733 instance, when using 'perf record -g' or 'perf trace --call-graph fp'.
735 This can only be done when no events are in use that have callchains
736 enabled, otherwise writing to this file will return -EBUSY.
738 The default value is 8.
740 ==============================================================
744 PID allocation wrap value. When the kernel's next PID value
745 reaches this value, it wraps back to a minimum PID value.
746 PIDs of value pid_max or larger are not allocated.
748 ==============================================================
752 The last pid allocated in the current (the one task using this sysctl
753 lives in) pid namespace. When selecting a pid for a next task on fork
754 kernel tries to allocate a number starting from this one.
756 ==============================================================
758 powersave-nap: (PPC only)
760 If set, Linux-PPC will use the 'nap' mode of powersaving,
761 otherwise the 'doze' mode will be used.
763 ==============================================================
767 The four values in printk denote: console_loglevel,
768 default_message_loglevel, minimum_console_loglevel and
769 default_console_loglevel respectively.
771 These values influence printk() behavior when printing or
772 logging error messages. See 'man 2 syslog' for more info on
773 the different loglevels.
775 - console_loglevel: messages with a higher priority than
776 this will be printed to the console
777 - default_message_loglevel: messages without an explicit priority
778 will be printed with this priority
779 - minimum_console_loglevel: minimum (highest) value to which
780 console_loglevel can be set
781 - default_console_loglevel: default value for console_loglevel
783 ==============================================================
787 Delay each printk message in printk_delay milliseconds
789 Value from 0 - 10000 is allowed.
791 ==============================================================
795 Some warning messages are rate limited. printk_ratelimit specifies
796 the minimum length of time between these messages (in jiffies), by
797 default we allow one every 5 seconds.
799 A value of 0 will disable rate limiting.
801 ==============================================================
803 printk_ratelimit_burst:
805 While long term we enforce one message per printk_ratelimit
806 seconds, we do allow a burst of messages to pass through.
807 printk_ratelimit_burst specifies the number of messages we can
808 send before ratelimiting kicks in.
810 ==============================================================
814 Control the logging to /dev/kmsg from userspace:
816 ratelimit: default, ratelimited
817 on: unlimited logging to /dev/kmsg from userspace
818 off: logging to /dev/kmsg disabled
820 The kernel command line parameter printk.devkmsg= overrides this and is
821 a one-time setting until next reboot: once set, it cannot be changed by
822 this sysctl interface anymore.
824 ==============================================================
828 This option can be used to select the type of process address
829 space randomization that is used in the system, for architectures
830 that support this feature.
832 0 - Turn the process address space randomization off. This is the
833 default for architectures that do not support this feature anyways,
834 and kernels that are booted with the "norandmaps" parameter.
836 1 - Make the addresses of mmap base, stack and VDSO page randomized.
837 This, among other things, implies that shared libraries will be
838 loaded to random addresses. Also for PIE-linked binaries, the
839 location of code start is randomized. This is the default if the
840 CONFIG_COMPAT_BRK option is enabled.
842 2 - Additionally enable heap randomization. This is the default if
843 CONFIG_COMPAT_BRK is disabled.
845 There are a few legacy applications out there (such as some ancient
846 versions of libc.so.5 from 1996) that assume that brk area starts
847 just after the end of the code+bss. These applications break when
848 start of the brk area is randomized. There are however no known
849 non-legacy applications that would be broken this way, so for most
850 systems it is safe to choose full randomization.
852 Systems with ancient and/or broken binaries should be configured
853 with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
854 address space randomization.
856 ==============================================================
858 reboot-cmd: (Sparc only)
860 ??? This seems to be a way to give an argument to the Sparc
861 ROM/Flash boot loader. Maybe to tell it what to do after
864 ==============================================================
866 rtsig-max & rtsig-nr:
868 The file rtsig-max can be used to tune the maximum number
869 of POSIX realtime (queued) signals that can be outstanding
872 rtsig-nr shows the number of RT signals currently queued.
874 ==============================================================
878 Enables/disables scheduler statistics. Enabling this feature
879 incurs a small amount of overhead in the scheduler but is
880 useful for debugging and performance tuning.
882 ==============================================================
886 This file shows the size of the generic SCSI (sg) buffer.
887 You can't tune it just yet, but you could change it on
888 compile time by editing include/scsi/sg.h and changing
889 the value of SG_BIG_BUFF.
891 There shouldn't be any reason to change this value. If
892 you can come up with one, you probably know what you
895 ==============================================================
899 This parameter sets the total amount of shared memory pages that
900 can be used system wide. Hence, SHMALL should always be at least
901 ceil(shmmax/PAGE_SIZE).
903 If you are not sure what the default PAGE_SIZE is on your Linux
904 system, you can run the following command:
908 ==============================================================
912 This value can be used to query and set the run time limit
913 on the maximum shared memory segment size that can be created.
914 Shared memory segments up to 1Gb are now supported in the
915 kernel. This value defaults to SHMMAX.
917 ==============================================================
921 Linux lets you set resource limits, including how much memory one
922 process can consume, via setrlimit(2). Unfortunately, shared memory
923 segments are allowed to exist without association with any process, and
924 thus might not be counted against any resource limits. If enabled,
925 shared memory segments are automatically destroyed when their attach
926 count becomes zero after a detach or a process termination. It will
927 also destroy segments that were created, but never attached to, on exit
928 from the process. The only use left for IPC_RMID is to immediately
929 destroy an unattached segment. Of course, this breaks the way things are
930 defined, so some applications might stop working. Note that this
931 feature will do you no good unless you also configure your resource
932 limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't
935 Note that if you change this from 0 to 1, already created segments
936 without users and with a dead originative process will be destroyed.
938 ==============================================================
940 sysctl_writes_strict:
942 Control how file position affects the behavior of updating sysctl values
943 via the /proc/sys interface:
945 -1 - Legacy per-write sysctl value handling, with no printk warnings.
946 Each write syscall must fully contain the sysctl value to be
947 written, and multiple writes on the same sysctl file descriptor
948 will rewrite the sysctl value, regardless of file position.
949 0 - Same behavior as above, but warn about processes that perform writes
950 to a sysctl file descriptor when the file position is not 0.
951 1 - (default) Respect file position when writing sysctl strings. Multiple
952 writes will append to the sysctl value buffer. Anything past the max
953 length of the sysctl value buffer will be ignored. Writes to numeric
954 sysctl entries must always be at file position 0 and the value must
955 be fully contained in the buffer sent in the write syscall.
957 ==============================================================
959 softlockup_all_cpu_backtrace:
961 This value controls the soft lockup detector thread's behavior
962 when a soft lockup condition is detected as to whether or not
963 to gather further debug information. If enabled, each cpu will
964 be issued an NMI and instructed to capture stack trace.
966 This feature is only applicable for architectures which support
969 0: do nothing. This is the default behavior.
971 1: on detection capture more debug information.
973 ==============================================================
977 This parameter can be used to control the soft lockup detector.
979 0 - disable the soft lockup detector
980 1 - enable the soft lockup detector
982 The soft lockup detector monitors CPUs for threads that are hogging the CPUs
983 without rescheduling voluntarily, and thus prevent the 'watchdog/N' threads
984 from running. The mechanism depends on the CPUs ability to respond to timer
985 interrupts which are needed for the 'watchdog/N' threads to be woken up by
986 the watchdog timer function, otherwise the NMI watchdog - if enabled - can
987 detect a hard lockup condition.
989 ==============================================================
993 Non-zero if the kernel has been tainted. Numeric values, which can be
994 ORed together. The letters are seen in "Tainted" line of Oops reports.
996 1 (P): A module with a non-GPL license has been loaded, this
997 includes modules with no license.
998 Set by modutils >= 2.4.9 and module-init-tools.
999 2 (F): A module was force loaded by insmod -f.
1000 Set by modutils >= 2.4.9 and module-init-tools.
1001 4 (S): Unsafe SMP processors: SMP with CPUs not designed for SMP.
1002 8 (R): A module was forcibly unloaded from the system by rmmod -f.
1003 16 (M): A hardware machine check error occurred on the system.
1004 32 (B): A bad page was discovered on the system.
1005 64 (U): The user has asked that the system be marked "tainted". This
1006 could be because they are running software that directly modifies
1007 the hardware, or for other reasons.
1008 128 (D): The system has died.
1009 256 (A): The ACPI DSDT has been overridden with one supplied by the user
1010 instead of using the one provided by the hardware.
1011 512 (W): A kernel warning has occurred.
1012 1024 (C): A module from drivers/staging was loaded.
1013 2048 (I): The system is working around a severe firmware bug.
1014 4096 (O): An out-of-tree module has been loaded.
1015 8192 (E): An unsigned module has been loaded in a kernel supporting module
1017 16384 (L): A soft lockup has previously occurred on the system.
1018 32768 (K): The kernel has been live patched.
1019 65536 (X): Auxiliary taint, defined and used by for distros.
1020 131072 (T): The kernel was built with the struct randomization plugin.
1022 ==============================================================
1026 This value controls the maximum number of threads that can be created
1029 During initialization the kernel sets this value such that even if the
1030 maximum number of threads is created, the thread structures occupy only
1031 a part (1/8th) of the available RAM pages.
1033 The minimum value that can be written to threads-max is 20.
1034 The maximum value that can be written to threads-max is given by the
1035 constant FUTEX_TID_MASK (0x3fffffff).
1036 If a value outside of this range is written to threads-max an error
1039 The value written is checked against the available RAM pages. If the
1040 thread structures would occupy too much (more than 1/8th) of the
1041 available RAM pages threads-max is reduced accordingly.
1043 ==============================================================
1045 unprivileged_bpf_disabled:
1047 Writing 1 to this entry will disable unprivileged calls to bpf();
1048 once disabled, calling bpf() without CAP_SYS_ADMIN will return
1049 -EPERM. Once set to 1, this can't be cleared from the running kernel
1052 Writing 2 to this entry will also disable unprivileged calls to bpf(),
1053 however, an admin can still change this setting later on, if needed, by
1054 writing 0 or 1 to this entry.
1056 If BPF_UNPRIV_DEFAULT_OFF is enabled in the kernel config, then this
1057 entry will default to 2 instead of 0.
1059 0 - Unprivileged calls to bpf() are enabled
1060 1 - Unprivileged calls to bpf() are disabled without recovery
1061 2 - Unprivileged calls to bpf() are disabled
1063 ==============================================================
1067 The value in this file affects behavior of handling NMI. When the
1068 value is non-zero, unknown NMI is trapped and then panic occurs. At
1069 that time, kernel debugging information is displayed on console.
1071 NMI switch that most IA32 servers have fires unknown NMI up, for
1072 example. If a system hangs up, try pressing the NMI switch.
1074 ==============================================================
1078 This parameter can be used to disable or enable the soft lockup detector
1079 _and_ the NMI watchdog (i.e. the hard lockup detector) at the same time.
1081 0 - disable both lockup detectors
1082 1 - enable both lockup detectors
1084 The soft lockup detector and the NMI watchdog can also be disabled or
1085 enabled individually, using the soft_watchdog and nmi_watchdog parameters.
1086 If the watchdog parameter is read, for example by executing
1088 cat /proc/sys/kernel/watchdog
1090 the output of this command (0 or 1) shows the logical OR of soft_watchdog
1093 ==============================================================
1097 This value can be used to control on which cpus the watchdog may run.
1098 The default cpumask is all possible cores, but if NO_HZ_FULL is
1099 enabled in the kernel config, and cores are specified with the
1100 nohz_full= boot argument, those cores are excluded by default.
1101 Offline cores can be included in this mask, and if the core is later
1102 brought online, the watchdog will be started based on the mask value.
1104 Typically this value would only be touched in the nohz_full case
1105 to re-enable cores that by default were not running the watchdog,
1106 if a kernel lockup was suspected on those cores.
1108 The argument value is the standard cpulist format for cpumasks,
1109 so for example to enable the watchdog on cores 0, 2, 3, and 4 you
1112 echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
1114 ==============================================================
1118 This value can be used to control the frequency of hrtimer and NMI
1119 events and the soft and hard lockup thresholds. The default threshold
1122 The softlockup threshold is (2 * watchdog_thresh). Setting this
1123 tunable to zero will disable lockup detection altogether.
1125 ==============================================================