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
65 - perf_cpu_time_max_percent
67 - perf_event_max_stack
69 - perf_event_max_contexts_per_stack
71 - powersave-nap [ PPC only ]
75 - printk_ratelimit_burst
76 - pty ==> Documentation/filesystems/devpts.txt
78 - real-root-dev ==> Documentation/admin-guide/initrd.rst
79 - reboot-cmd [ SPARC only ]
82 - seccomp/ ==> Documentation/userspace-api/seccomp_filter.rst
84 - sem_next_id [ sysv ipc ]
85 - sg-big-buff [ generic SCSI device (sg) ]
86 - shm_next_id [ sysv ipc ]
91 - softlockup_all_cpu_backtrace
94 - stop-a [ SPARC only ]
95 - sysrq ==> Documentation/admin-guide/sysrq.rst
96 - sysctl_writes_strict
104 ==============================================================
108 highwater lowwater frequency
110 If BSD-style process accounting is enabled these values control
111 its behaviour. If free space on filesystem where the log lives
112 goes below <lowwater>% accounting suspends. If free space gets
113 above <highwater>% accounting resumes. <Frequency> determines
114 how often do we check the amount of free space (value is in
117 That is, suspend accounting if there left <= 2% free; resume it
118 if we got >=4%; consider information about amount of free space
119 valid for 30 seconds.
121 ==============================================================
127 See Doc*/kernel/power/video.txt, it allows mode of video boot to be
130 ==============================================================
134 This variable has no effect and may be removed in future kernel
135 releases. Reading it always returns 0.
136 Up to Linux 3.17, it enabled/disabled automatic recomputing of msgmni
137 upon memory add/remove or upon ipc namespace creation/removal.
138 Echoing "1" into this file enabled msgmni automatic recomputing.
139 Echoing "0" turned it off. auto_msgmni default value was 1.
142 ==============================================================
146 x86 bootloader identification
148 This gives the bootloader type number as indicated by the bootloader,
149 shifted left by 4, and OR'd with the low four bits of the bootloader
150 version. The reason for this encoding is that this used to match the
151 type_of_loader field in the kernel header; the encoding is kept for
152 backwards compatibility. That is, if the full bootloader type number
153 is 0x15 and the full version number is 0x234, this file will contain
154 the value 340 = 0x154.
156 See the type_of_loader and ext_loader_type fields in
157 Documentation/x86/boot.txt for additional information.
159 ==============================================================
163 x86 bootloader version
165 The complete bootloader version number. In the example above, this
166 file will contain the value 564 = 0x234.
168 See the type_of_loader and ext_loader_ver fields in
169 Documentation/x86/boot.txt for additional information.
171 ==============================================================
175 Controls the kernel's callhome behavior in case of a kernel panic.
177 The s390 hardware allows an operating system to send a notification
178 to a service organization (callhome) in case of an operating system panic.
180 When the value in this file is 0 (which is the default behavior)
181 nothing happens in case of a kernel panic. If this value is set to "1"
182 the complete kernel oops message is send to the IBM customer service
183 organization in case the mainframe the Linux operating system is running
184 on has a service contract with IBM.
186 ==============================================================
190 Highest valid capability of the running kernel. Exports
191 CAP_LAST_CAP from the kernel.
193 ==============================================================
197 core_pattern is used to specify a core dumpfile pattern name.
198 . max length 128 characters; default value is "core"
199 . core_pattern is used as a pattern template for the output filename;
200 certain string patterns (beginning with '%') are substituted with
202 . backward compatibility with core_uses_pid:
203 If core_pattern does not include "%p" (default does not)
204 and core_uses_pid is set, then .PID will be appended to
206 . corename format specifiers:
207 %<NUL> '%' is dropped
210 %P global pid (init PID namespace)
212 %I global tid (init PID namespace)
213 %u uid (in initial user namespace)
214 %g gid (in initial user namespace)
215 %d dump mode, matches PR_SET_DUMPABLE and
216 /proc/sys/fs/suid_dumpable
220 %e executable filename (may be shortened)
222 %<OTHER> both are dropped
223 . If the first character of the pattern is a '|', the kernel will treat
224 the rest of the pattern as a command to run. The core dump will be
225 written to the standard input of that program instead of to a file.
227 ==============================================================
231 This sysctl is only applicable when core_pattern is configured to pipe
232 core files to a user space helper (when the first character of
233 core_pattern is a '|', see above). When collecting cores via a pipe
234 to an application, it is occasionally useful for the collecting
235 application to gather data about the crashing process from its
236 /proc/pid directory. In order to do this safely, the kernel must wait
237 for the collecting process to exit, so as not to remove the crashing
238 processes proc files prematurely. This in turn creates the
239 possibility that a misbehaving userspace collecting process can block
240 the reaping of a crashed process simply by never exiting. This sysctl
241 defends against that. It defines how many concurrent crashing
242 processes may be piped to user space applications in parallel. If
243 this value is exceeded, then those crashing processes above that value
244 are noted via the kernel log and their cores are skipped. 0 is a
245 special value, indicating that unlimited processes may be captured in
246 parallel, but that no waiting will take place (i.e. the collecting
247 process is not guaranteed access to /proc/<crashing pid>/). This
250 ==============================================================
254 The default coredump filename is "core". By setting
255 core_uses_pid to 1, the coredump filename becomes core.PID.
256 If core_pattern does not include "%p" (default does not)
257 and core_uses_pid is set, then .PID will be appended to
260 ==============================================================
264 When the value in this file is 0, ctrl-alt-del is trapped and
265 sent to the init(1) program to handle a graceful restart.
266 When, however, the value is > 0, Linux's reaction to a Vulcan
267 Nerve Pinch (tm) will be an immediate reboot, without even
268 syncing its dirty buffers.
270 Note: when a program (like dosemu) has the keyboard in 'raw'
271 mode, the ctrl-alt-del is intercepted by the program before it
272 ever reaches the kernel tty layer, and it's up to the program
273 to decide what to do with it.
275 ==============================================================
279 This toggle indicates whether unprivileged users are prevented
280 from using dmesg(8) to view messages from the kernel's log buffer.
281 When dmesg_restrict is set to (0) there are no restrictions. When
282 dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
285 The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
286 default value of dmesg_restrict.
288 ==============================================================
290 domainname & hostname:
292 These files can be used to set the NIS/YP domainname and the
293 hostname of your box in exactly the same way as the commands
294 domainname and hostname, i.e.:
295 # echo "darkstar" > /proc/sys/kernel/hostname
296 # echo "mydomain" > /proc/sys/kernel/domainname
297 has the same effect as
298 # hostname "darkstar"
299 # domainname "mydomain"
301 Note, however, that the classic darkstar.frop.org has the
302 hostname "darkstar" and DNS (Internet Domain Name Server)
303 domainname "frop.org", not to be confused with the NIS (Network
304 Information Service) or YP (Yellow Pages) domainname. These two
305 domain names are in general different. For a detailed discussion
306 see the hostname(1) man page.
308 ==============================================================
309 hardlockup_all_cpu_backtrace:
311 This value controls the hard lockup detector behavior when a hard
312 lockup condition is detected as to whether or not to gather further
313 debug information. If enabled, arch-specific all-CPU stack dumping
316 0: do nothing. This is the default behavior.
318 1: on detection capture more debug information.
319 ==============================================================
323 This parameter can be used to control whether the kernel panics
324 when a hard lockup is detected.
326 0 - don't panic on hard lockup
327 1 - panic on hard lockup
329 See Documentation/lockup-watchdogs.txt for more information. This can
330 also be set using the nmi_watchdog kernel parameter.
332 ==============================================================
336 Path for the hotplug policy agent.
337 Default value is "/sbin/hotplug".
339 ==============================================================
343 Controls the kernel's behavior when a hung task is detected.
344 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
346 0: continue operation. This is the default behavior.
348 1: panic immediately.
350 ==============================================================
352 hung_task_check_count:
354 The upper bound on the number of tasks that are checked.
355 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
357 ==============================================================
359 hung_task_timeout_secs:
361 When a task in D state did not get scheduled
362 for more than this value report a warning.
363 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
365 0: means infinite timeout - no checking done.
366 Possible values to set are in range {0..LONG_MAX/HZ}.
368 ==============================================================
370 hung_task_check_interval_secs:
372 Hung task check interval. If hung task checking is enabled
373 (see hung_task_timeout_secs), the check is done every
374 hung_task_check_interval_secs seconds.
375 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
377 0 (default): means use hung_task_timeout_secs as checking interval.
378 Possible values to set are in range {0..LONG_MAX/HZ}.
380 ==============================================================
384 The maximum number of warnings to report. During a check interval
385 if a hung task is detected, this value is decreased by 1.
386 When this value reaches 0, no more warnings will be reported.
387 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
389 -1: report an infinite number of warnings.
391 ==============================================================
393 hyperv_record_panic_msg:
395 Controls whether the panic kmsg data should be reported to Hyper-V.
397 0: do not report panic kmsg data.
399 1: report the panic kmsg data. This is the default behavior.
401 ==============================================================
405 A toggle indicating if the kexec_load syscall has been disabled. This
406 value defaults to 0 (false: kexec_load enabled), but can be set to 1
407 (true: kexec_load disabled). Once true, kexec can no longer be used, and
408 the toggle cannot be set back to false. This allows a kexec image to be
409 loaded before disabling the syscall, allowing a system to set up (and
410 later use) an image without it being altered. Generally used together
411 with the "modules_disabled" sysctl.
413 ==============================================================
417 This toggle indicates whether restrictions are placed on
418 exposing kernel addresses via /proc and other interfaces.
420 When kptr_restrict is set to 0 (the default) the address is hashed before
421 printing. (This is the equivalent to %p.)
423 When kptr_restrict is set to (1), kernel pointers printed using the %pK
424 format specifier will be replaced with 0's unless the user has CAP_SYSLOG
425 and effective user and group ids are equal to the real ids. This is
426 because %pK checks are done at read() time rather than open() time, so
427 if permissions are elevated between the open() and the read() (e.g via
428 a setuid binary) then %pK will not leak kernel pointers to unprivileged
429 users. Note, this is a temporary solution only. The correct long-term
430 solution is to do the permission checks at open() time. Consider removing
431 world read permissions from files that use %pK, and using dmesg_restrict
432 to protect against uses of %pK in dmesg(8) if leaking kernel pointer
433 values to unprivileged users is a concern.
435 When kptr_restrict is set to (2), kernel pointers printed using
436 %pK will be replaced with 0's regardless of privileges.
438 ==============================================================
442 This flag controls the L2 cache of G3 processor boards. If
443 0, the cache is disabled. Enabled if nonzero.
445 ==============================================================
449 A toggle value indicating if modules are allowed to be loaded
450 in an otherwise modular kernel. This toggle defaults to off
451 (0), but can be set true (1). Once true, modules can be
452 neither loaded nor unloaded, and the toggle cannot be set back
453 to false. Generally used with the "kexec_load_disabled" toggle.
455 ==============================================================
457 msg_next_id, sem_next_id, and shm_next_id:
459 These three toggles allows to specify desired id for next allocated IPC
460 object: message, semaphore or shared memory respectively.
462 By default they are equal to -1, which means generic allocation logic.
463 Possible values to set are in range {0..INT_MAX}.
466 1) kernel doesn't guarantee, that new object will have desired id. So,
467 it's up to userspace, how to handle an object with "wrong" id.
468 2) Toggle with non-default value will be set back to -1 by kernel after
469 successful IPC object allocation. If an IPC object allocation syscall
470 fails, it is undefined if the value remains unmodified or is reset to -1.
472 ==============================================================
476 This parameter can be used to control the NMI watchdog
477 (i.e. the hard lockup detector) on x86 systems.
479 0 - disable the hard lockup detector
480 1 - enable the hard lockup detector
482 The hard lockup detector monitors each CPU for its ability to respond to
483 timer interrupts. The mechanism utilizes CPU performance counter registers
484 that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
485 while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
487 The NMI watchdog is disabled by default if the kernel is running as a guest
488 in a KVM virtual machine. This default can be overridden by adding
492 to the guest kernel command line (see Documentation/admin-guide/kernel-parameters.rst).
494 ==============================================================
498 Enables/disables automatic page fault based NUMA memory
499 balancing. Memory is moved automatically to nodes
500 that access it often.
502 Enables/disables automatic NUMA memory balancing. On NUMA machines, there
503 is a performance penalty if remote memory is accessed by a CPU. When this
504 feature is enabled the kernel samples what task thread is accessing memory
505 by periodically unmapping pages and later trapping a page fault. At the
506 time of the page fault, it is determined if the data being accessed should
507 be migrated to a local memory node.
509 The unmapping of pages and trapping faults incur additional overhead that
510 ideally is offset by improved memory locality but there is no universal
511 guarantee. If the target workload is already bound to NUMA nodes then this
512 feature should be disabled. Otherwise, if the system overhead from the
513 feature is too high then the rate the kernel samples for NUMA hinting
514 faults may be controlled by the numa_balancing_scan_period_min_ms,
515 numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms,
516 numa_balancing_scan_size_mb, and numa_balancing_settle_count sysctls.
518 ==============================================================
520 numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms,
521 numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb
523 Automatic NUMA balancing scans tasks address space and unmaps pages to
524 detect if pages are properly placed or if the data should be migrated to a
525 memory node local to where the task is running. Every "scan delay" the task
526 scans the next "scan size" number of pages in its address space. When the
527 end of the address space is reached the scanner restarts from the beginning.
529 In combination, the "scan delay" and "scan size" determine the scan rate.
530 When "scan delay" decreases, the scan rate increases. The scan delay and
531 hence the scan rate of every task is adaptive and depends on historical
532 behaviour. If pages are properly placed then the scan delay increases,
533 otherwise the scan delay decreases. The "scan size" is not adaptive but
534 the higher the "scan size", the higher the scan rate.
536 Higher scan rates incur higher system overhead as page faults must be
537 trapped and potentially data must be migrated. However, the higher the scan
538 rate, the more quickly a tasks memory is migrated to a local node if the
539 workload pattern changes and minimises performance impact due to remote
540 memory accesses. These sysctls control the thresholds for scan delays and
541 the number of pages scanned.
543 numa_balancing_scan_period_min_ms is the minimum time in milliseconds to
544 scan a tasks virtual memory. It effectively controls the maximum scanning
547 numa_balancing_scan_delay_ms is the starting "scan delay" used for a task
548 when it initially forks.
550 numa_balancing_scan_period_max_ms is the maximum time in milliseconds to
551 scan a tasks virtual memory. It effectively controls the minimum scanning
554 numa_balancing_scan_size_mb is how many megabytes worth of pages are
555 scanned for a given scan.
557 ==============================================================
559 osrelease, ostype & version:
566 #5 Wed Feb 25 21:49:24 MET 1998
568 The files osrelease and ostype should be clear enough. Version
569 needs a little more clarification however. The '#5' means that
570 this is the fifth kernel built from this source base and the
571 date behind it indicates the time the kernel was built.
572 The only way to tune these values is to rebuild the kernel :-)
574 ==============================================================
576 overflowgid & overflowuid:
578 if your architecture did not always support 32-bit UIDs (i.e. arm,
579 i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
580 applications that use the old 16-bit UID/GID system calls, if the
581 actual UID or GID would exceed 65535.
583 These sysctls allow you to change the value of the fixed UID and GID.
584 The default is 65534.
586 ==============================================================
590 The value in this file represents the number of seconds the kernel
591 waits before rebooting on a panic. When you use the software watchdog,
592 the recommended setting is 60.
594 ==============================================================
598 Controls the kernel's behavior when a CPU receives an NMI caused by
601 0: try to continue operation (default)
603 1: panic immediately. The IO error triggered an NMI. This indicates a
604 serious system condition which could result in IO data corruption.
605 Rather than continuing, panicking might be a better choice. Some
606 servers issue this sort of NMI when the dump button is pushed,
607 and you can use this option to take a crash dump.
609 ==============================================================
613 Controls the kernel's behaviour when an oops or BUG is encountered.
615 0: try to continue operation
617 1: panic immediately. If the `panic' sysctl is also non-zero then the
618 machine will be rebooted.
620 ==============================================================
622 panic_on_stackoverflow:
624 Controls the kernel's behavior when detecting the overflows of
625 kernel, IRQ and exception stacks except a user stack.
626 This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
628 0: try to continue operation.
630 1: panic immediately.
632 ==============================================================
634 panic_on_unrecovered_nmi:
636 The default Linux behaviour on an NMI of either memory or unknown is
637 to continue operation. For many environments such as scientific
638 computing it is preferable that the box is taken out and the error
639 dealt with than an uncorrected parity/ECC error get propagated.
641 A small number of systems do generate NMI's for bizarre random reasons
642 such as power management so the default is off. That sysctl works like
643 the existing panic controls already in that directory.
645 ==============================================================
649 Calls panic() in the WARN() path when set to 1. This is useful to avoid
650 a kernel rebuild when attempting to kdump at the location of a WARN().
652 0: only WARN(), default behaviour.
654 1: call panic() after printing out WARN() location.
656 ==============================================================
660 Bitmask for printing system info when panic happens. User can chose
661 combination of the following bits:
663 bit 0: print all tasks info
664 bit 1: print system memory info
665 bit 2: print timer info
666 bit 3: print locks info if CONFIG_LOCKDEP is on
667 bit 4: print ftrace buffer
669 So for example to print tasks and memory info on panic, user can:
670 echo 3 > /proc/sys/kernel/panic_print
672 ==============================================================
676 When set to 1, calls panic() after RCU stall detection messages. This
677 is useful to define the root cause of RCU stalls using a vmcore.
679 0: do not panic() when RCU stall takes place, default behavior.
681 1: panic() after printing RCU stall messages.
683 ==============================================================
685 perf_cpu_time_max_percent:
687 Hints to the kernel how much CPU time it should be allowed to
688 use to handle perf sampling events. If the perf subsystem
689 is informed that its samples are exceeding this limit, it
690 will drop its sampling frequency to attempt to reduce its CPU
693 Some perf sampling happens in NMIs. If these samples
694 unexpectedly take too long to execute, the NMIs can become
695 stacked up next to each other so much that nothing else is
698 0: disable the mechanism. Do not monitor or correct perf's
699 sampling rate no matter how CPU time it takes.
701 1-100: attempt to throttle perf's sample rate to this
702 percentage of CPU. Note: the kernel calculates an
703 "expected" length of each sample event. 100 here means
704 100% of that expected length. Even if this is set to
705 100, you may still see sample throttling if this
706 length is exceeded. Set to 0 if you truly do not care
707 how much CPU is consumed.
709 ==============================================================
713 Controls use of the performance events system by unprivileged
714 users (without CAP_SYS_ADMIN). The default value is 2.
716 -1: Allow use of (almost) all events by all users
717 Ignore mlock limit after perf_event_mlock_kb without CAP_IPC_LOCK
718 >=0: Disallow ftrace function tracepoint by users without CAP_SYS_ADMIN
719 Disallow raw tracepoint access by users without CAP_SYS_ADMIN
720 >=1: Disallow CPU event access by users without CAP_SYS_ADMIN
721 >=2: Disallow kernel profiling by users without CAP_SYS_ADMIN
723 ==============================================================
725 perf_event_max_stack:
727 Controls maximum number of stack frames to copy for (attr.sample_type &
728 PERF_SAMPLE_CALLCHAIN) configured events, for instance, when using
729 'perf record -g' or 'perf trace --call-graph fp'.
731 This can only be done when no events are in use that have callchains
732 enabled, otherwise writing to this file will return -EBUSY.
734 The default value is 127.
736 ==============================================================
740 Control size of per-cpu ring buffer not counted agains mlock limit.
742 The default value is 512 + 1 page
744 ==============================================================
746 perf_event_max_contexts_per_stack:
748 Controls maximum number of stack frame context entries for
749 (attr.sample_type & PERF_SAMPLE_CALLCHAIN) configured events, for
750 instance, when using 'perf record -g' or 'perf trace --call-graph fp'.
752 This can only be done when no events are in use that have callchains
753 enabled, otherwise writing to this file will return -EBUSY.
755 The default value is 8.
757 ==============================================================
761 PID allocation wrap value. When the kernel's next PID value
762 reaches this value, it wraps back to a minimum PID value.
763 PIDs of value pid_max or larger are not allocated.
765 ==============================================================
769 The last pid allocated in the current (the one task using this sysctl
770 lives in) pid namespace. When selecting a pid for a next task on fork
771 kernel tries to allocate a number starting from this one.
773 ==============================================================
775 powersave-nap: (PPC only)
777 If set, Linux-PPC will use the 'nap' mode of powersaving,
778 otherwise the 'doze' mode will be used.
780 ==============================================================
784 The four values in printk denote: console_loglevel,
785 default_message_loglevel, minimum_console_loglevel and
786 default_console_loglevel respectively.
788 These values influence printk() behavior when printing or
789 logging error messages. See 'man 2 syslog' for more info on
790 the different loglevels.
792 - console_loglevel: messages with a higher priority than
793 this will be printed to the console
794 - default_message_loglevel: messages without an explicit priority
795 will be printed with this priority
796 - minimum_console_loglevel: minimum (highest) value to which
797 console_loglevel can be set
798 - default_console_loglevel: default value for console_loglevel
800 ==============================================================
804 Delay each printk message in printk_delay milliseconds
806 Value from 0 - 10000 is allowed.
808 ==============================================================
812 Some warning messages are rate limited. printk_ratelimit specifies
813 the minimum length of time between these messages (in jiffies), by
814 default we allow one every 5 seconds.
816 A value of 0 will disable rate limiting.
818 ==============================================================
820 printk_ratelimit_burst:
822 While long term we enforce one message per printk_ratelimit
823 seconds, we do allow a burst of messages to pass through.
824 printk_ratelimit_burst specifies the number of messages we can
825 send before ratelimiting kicks in.
827 ==============================================================
831 Control the logging to /dev/kmsg from userspace:
833 ratelimit: default, ratelimited
834 on: unlimited logging to /dev/kmsg from userspace
835 off: logging to /dev/kmsg disabled
837 The kernel command line parameter printk.devkmsg= overrides this and is
838 a one-time setting until next reboot: once set, it cannot be changed by
839 this sysctl interface anymore.
841 ==============================================================
845 This option can be used to select the type of process address
846 space randomization that is used in the system, for architectures
847 that support this feature.
849 0 - Turn the process address space randomization off. This is the
850 default for architectures that do not support this feature anyways,
851 and kernels that are booted with the "norandmaps" parameter.
853 1 - Make the addresses of mmap base, stack and VDSO page randomized.
854 This, among other things, implies that shared libraries will be
855 loaded to random addresses. Also for PIE-linked binaries, the
856 location of code start is randomized. This is the default if the
857 CONFIG_COMPAT_BRK option is enabled.
859 2 - Additionally enable heap randomization. This is the default if
860 CONFIG_COMPAT_BRK is disabled.
862 There are a few legacy applications out there (such as some ancient
863 versions of libc.so.5 from 1996) that assume that brk area starts
864 just after the end of the code+bss. These applications break when
865 start of the brk area is randomized. There are however no known
866 non-legacy applications that would be broken this way, so for most
867 systems it is safe to choose full randomization.
869 Systems with ancient and/or broken binaries should be configured
870 with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
871 address space randomization.
873 ==============================================================
875 reboot-cmd: (Sparc only)
877 ??? This seems to be a way to give an argument to the Sparc
878 ROM/Flash boot loader. Maybe to tell it what to do after
881 ==============================================================
883 rtsig-max & rtsig-nr:
885 The file rtsig-max can be used to tune the maximum number
886 of POSIX realtime (queued) signals that can be outstanding
889 rtsig-nr shows the number of RT signals currently queued.
891 ==============================================================
895 Enables/disables scheduler statistics. Enabling this feature
896 incurs a small amount of overhead in the scheduler but is
897 useful for debugging and performance tuning.
899 ==============================================================
903 This file shows the size of the generic SCSI (sg) buffer.
904 You can't tune it just yet, but you could change it on
905 compile time by editing include/scsi/sg.h and changing
906 the value of SG_BIG_BUFF.
908 There shouldn't be any reason to change this value. If
909 you can come up with one, you probably know what you
912 ==============================================================
916 This parameter sets the total amount of shared memory pages that
917 can be used system wide. Hence, SHMALL should always be at least
918 ceil(shmmax/PAGE_SIZE).
920 If you are not sure what the default PAGE_SIZE is on your Linux
921 system, you can run the following command:
925 ==============================================================
929 This value can be used to query and set the run time limit
930 on the maximum shared memory segment size that can be created.
931 Shared memory segments up to 1Gb are now supported in the
932 kernel. This value defaults to SHMMAX.
934 ==============================================================
938 Linux lets you set resource limits, including how much memory one
939 process can consume, via setrlimit(2). Unfortunately, shared memory
940 segments are allowed to exist without association with any process, and
941 thus might not be counted against any resource limits. If enabled,
942 shared memory segments are automatically destroyed when their attach
943 count becomes zero after a detach or a process termination. It will
944 also destroy segments that were created, but never attached to, on exit
945 from the process. The only use left for IPC_RMID is to immediately
946 destroy an unattached segment. Of course, this breaks the way things are
947 defined, so some applications might stop working. Note that this
948 feature will do you no good unless you also configure your resource
949 limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't
952 Note that if you change this from 0 to 1, already created segments
953 without users and with a dead originative process will be destroyed.
955 ==============================================================
957 sysctl_writes_strict:
959 Control how file position affects the behavior of updating sysctl values
960 via the /proc/sys interface:
962 -1 - Legacy per-write sysctl value handling, with no printk warnings.
963 Each write syscall must fully contain the sysctl value to be
964 written, and multiple writes on the same sysctl file descriptor
965 will rewrite the sysctl value, regardless of file position.
966 0 - Same behavior as above, but warn about processes that perform writes
967 to a sysctl file descriptor when the file position is not 0.
968 1 - (default) Respect file position when writing sysctl strings. Multiple
969 writes will append to the sysctl value buffer. Anything past the max
970 length of the sysctl value buffer will be ignored. Writes to numeric
971 sysctl entries must always be at file position 0 and the value must
972 be fully contained in the buffer sent in the write syscall.
974 ==============================================================
976 softlockup_all_cpu_backtrace:
978 This value controls the soft lockup detector thread's behavior
979 when a soft lockup condition is detected as to whether or not
980 to gather further debug information. If enabled, each cpu will
981 be issued an NMI and instructed to capture stack trace.
983 This feature is only applicable for architectures which support
986 0: do nothing. This is the default behavior.
988 1: on detection capture more debug information.
990 ==============================================================
994 This parameter can be used to control the soft lockup detector.
996 0 - disable the soft lockup detector
997 1 - enable the soft lockup detector
999 The soft lockup detector monitors CPUs for threads that are hogging the CPUs
1000 without rescheduling voluntarily, and thus prevent the 'watchdog/N' threads
1001 from running. The mechanism depends on the CPUs ability to respond to timer
1002 interrupts which are needed for the 'watchdog/N' threads to be woken up by
1003 the watchdog timer function, otherwise the NMI watchdog - if enabled - can
1004 detect a hard lockup condition.
1006 ==============================================================
1010 This parameter can be used to control kernel stack erasing at the end
1011 of syscalls for kernels built with CONFIG_GCC_PLUGIN_STACKLEAK.
1013 That erasing reduces the information which kernel stack leak bugs
1014 can reveal and blocks some uninitialized stack variable attacks.
1015 The tradeoff is the performance impact: on a single CPU system kernel
1016 compilation sees a 1% slowdown, other systems and workloads may vary.
1018 0: kernel stack erasing is disabled, STACKLEAK_METRICS are not updated.
1020 1: kernel stack erasing is enabled (default), it is performed before
1021 returning to the userspace at the end of syscalls.
1023 ==============================================================
1027 Non-zero if the kernel has been tainted. Numeric values, which can be
1028 ORed together. The letters are seen in "Tainted" line of Oops reports.
1030 1 (P): A module with a non-GPL license has been loaded, this
1031 includes modules with no license.
1032 Set by modutils >= 2.4.9 and module-init-tools.
1033 2 (F): A module was force loaded by insmod -f.
1034 Set by modutils >= 2.4.9 and module-init-tools.
1035 4 (S): Unsafe SMP processors: SMP with CPUs not designed for SMP.
1036 8 (R): A module was forcibly unloaded from the system by rmmod -f.
1037 16 (M): A hardware machine check error occurred on the system.
1038 32 (B): A bad page was discovered on the system.
1039 64 (U): The user has asked that the system be marked "tainted". This
1040 could be because they are running software that directly modifies
1041 the hardware, or for other reasons.
1042 128 (D): The system has died.
1043 256 (A): The ACPI DSDT has been overridden with one supplied by the user
1044 instead of using the one provided by the hardware.
1045 512 (W): A kernel warning has occurred.
1046 1024 (C): A module from drivers/staging was loaded.
1047 2048 (I): The system is working around a severe firmware bug.
1048 4096 (O): An out-of-tree module has been loaded.
1049 8192 (E): An unsigned module has been loaded in a kernel supporting module
1051 16384 (L): A soft lockup has previously occurred on the system.
1052 32768 (K): The kernel has been live patched.
1053 65536 (X): Auxiliary taint, defined and used by for distros.
1054 131072 (T): The kernel was built with the struct randomization plugin.
1056 ==============================================================
1060 This value controls the maximum number of threads that can be created
1063 During initialization the kernel sets this value such that even if the
1064 maximum number of threads is created, the thread structures occupy only
1065 a part (1/8th) of the available RAM pages.
1067 The minimum value that can be written to threads-max is 20.
1068 The maximum value that can be written to threads-max is given by the
1069 constant FUTEX_TID_MASK (0x3fffffff).
1070 If a value outside of this range is written to threads-max an error
1073 The value written is checked against the available RAM pages. If the
1074 thread structures would occupy too much (more than 1/8th) of the
1075 available RAM pages threads-max is reduced accordingly.
1077 ==============================================================
1081 The value in this file affects behavior of handling NMI. When the
1082 value is non-zero, unknown NMI is trapped and then panic occurs. At
1083 that time, kernel debugging information is displayed on console.
1085 NMI switch that most IA32 servers have fires unknown NMI up, for
1086 example. If a system hangs up, try pressing the NMI switch.
1088 ==============================================================
1092 This parameter can be used to disable or enable the soft lockup detector
1093 _and_ the NMI watchdog (i.e. the hard lockup detector) at the same time.
1095 0 - disable both lockup detectors
1096 1 - enable both lockup detectors
1098 The soft lockup detector and the NMI watchdog can also be disabled or
1099 enabled individually, using the soft_watchdog and nmi_watchdog parameters.
1100 If the watchdog parameter is read, for example by executing
1102 cat /proc/sys/kernel/watchdog
1104 the output of this command (0 or 1) shows the logical OR of soft_watchdog
1107 ==============================================================
1111 This value can be used to control on which cpus the watchdog may run.
1112 The default cpumask is all possible cores, but if NO_HZ_FULL is
1113 enabled in the kernel config, and cores are specified with the
1114 nohz_full= boot argument, those cores are excluded by default.
1115 Offline cores can be included in this mask, and if the core is later
1116 brought online, the watchdog will be started based on the mask value.
1118 Typically this value would only be touched in the nohz_full case
1119 to re-enable cores that by default were not running the watchdog,
1120 if a kernel lockup was suspected on those cores.
1122 The argument value is the standard cpulist format for cpumasks,
1123 so for example to enable the watchdog on cores 0, 2, 3, and 4 you
1126 echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
1128 ==============================================================
1132 This value can be used to control the frequency of hrtimer and NMI
1133 events and the soft and hard lockup thresholds. The default threshold
1136 The softlockup threshold is (2 * watchdog_thresh). Setting this
1137 tunable to zero will disable lockup detection altogether.
1139 ==============================================================