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 - boot_reason [ ARM and ARM64 only ]
27 - callhome [ S390 only ]
29 - cold_boot [ ARM and ARM64 only ]
38 - hardlockup_all_cpu_backtrace
40 - hung_task_check_count
41 - hung_task_timeout_secs
45 - kstack_depth_to_print [ X86 only ]
47 - modprobe ==> Documentation/debugging-modules.txt
49 - msg_next_id [ sysv ipc ]
60 - panic_on_stackoverflow
61 - panic_on_unrecovered_nmi
63 - perf_cpu_time_max_percent
66 - powersave-nap [ PPC only ]
70 - printk_ratelimit_burst
72 - real-root-dev ==> Documentation/initrd.txt
73 - reboot-cmd [ SPARC only ]
77 - sem_next_id [ sysv ipc ]
78 - sg-big-buff [ generic SCSI device (sg) ]
79 - shm_next_id [ sysv ipc ]
84 - softlockup_all_cpu_backtrace
86 - stop-a [ SPARC only ]
87 - sysrq ==> Documentation/sysrq.txt
88 - sysctl_writes_strict
96 ==============================================================
100 highwater lowwater frequency
102 If BSD-style process accounting is enabled these values control
103 its behaviour. If free space on filesystem where the log lives
104 goes below <lowwater>% accounting suspends. If free space gets
105 above <highwater>% accounting resumes. <Frequency> determines
106 how often do we check the amount of free space (value is in
109 That is, suspend accounting if there left <= 2% free; resume it
110 if we got >=4%; consider information about amount of free space
111 valid for 30 seconds.
113 ==============================================================
119 See Doc*/kernel/power/video.txt, it allows mode of video boot to be
122 ==============================================================
126 This variable has no effect and may be removed in future kernel
127 releases. Reading it always returns 0.
128 Up to Linux 3.17, it enabled/disabled automatic recomputing of msgmni
129 upon memory add/remove or upon ipc namespace creation/removal.
130 Echoing "1" into this file enabled msgmni automatic recomputing.
131 Echoing "0" turned it off. auto_msgmni default value was 1.
134 ==============================================================
138 x86 bootloader identification
140 This gives the bootloader type number as indicated by the bootloader,
141 shifted left by 4, and OR'd with the low four bits of the bootloader
142 version. The reason for this encoding is that this used to match the
143 type_of_loader field in the kernel header; the encoding is kept for
144 backwards compatibility. That is, if the full bootloader type number
145 is 0x15 and the full version number is 0x234, this file will contain
146 the value 340 = 0x154.
148 See the type_of_loader and ext_loader_type fields in
149 Documentation/x86/boot.txt for additional information.
151 ==============================================================
155 x86 bootloader version
157 The complete bootloader version number. In the example above, this
158 file will contain the value 564 = 0x234.
160 See the type_of_loader and ext_loader_ver fields in
161 Documentation/x86/boot.txt for additional information.
163 ==============================================================
167 ARM and ARM64 -- reason for device boot
169 A single bit will be set in the unsigned integer value to identify the
170 reason the device was booted / powered on. The value will be zero if this
171 feature is not supported on the ARM device being booted.
173 See the power-on-status field definitions in
174 Documentation/arm/msm/boot.txt for Qualcomm's family of devices.
176 ==============================================================
180 Controls the kernel's callhome behavior in case of a kernel panic.
182 The s390 hardware allows an operating system to send a notification
183 to a service organization (callhome) in case of an operating system panic.
185 When the value in this file is 0 (which is the default behavior)
186 nothing happens in case of a kernel panic. If this value is set to "1"
187 the complete kernel oops message is send to the IBM customer service
188 organization in case the mainframe the Linux operating system is running
189 on has a service contract with IBM.
191 ==============================================================
195 Highest valid capability of the running kernel. Exports
196 CAP_LAST_CAP from the kernel.
198 ===============================================================
202 ARM and ARM64 -- indicator for system cold boot
204 A single bit will be set in the unsigned integer value to identify
205 whether the device was booted from a cold or warm state. Zero
206 indicating a warm boot and one indicating a cold boot.
208 ==============================================================
212 core_pattern is used to specify a core dumpfile pattern name.
213 . max length 128 characters; default value is "core"
214 . core_pattern is used as a pattern template for the output filename;
215 certain string patterns (beginning with '%') are substituted with
217 . backward compatibility with core_uses_pid:
218 If core_pattern does not include "%p" (default does not)
219 and core_uses_pid is set, then .PID will be appended to
221 . corename format specifiers:
222 %<NUL> '%' is dropped
225 %P global pid (init PID namespace)
227 %I global tid (init PID namespace)
228 %u uid (in initial user namespace)
229 %g gid (in initial user namespace)
230 %d dump mode, matches PR_SET_DUMPABLE and
231 /proc/sys/fs/suid_dumpable
235 %e executable filename (may be shortened)
237 %<OTHER> both are dropped
238 . If the first character of the pattern is a '|', the kernel will treat
239 the rest of the pattern as a command to run. The core dump will be
240 written to the standard input of that program instead of to a file.
242 ==============================================================
246 This sysctl is only applicable when core_pattern is configured to pipe
247 core files to a user space helper (when the first character of
248 core_pattern is a '|', see above). When collecting cores via a pipe
249 to an application, it is occasionally useful for the collecting
250 application to gather data about the crashing process from its
251 /proc/pid directory. In order to do this safely, the kernel must wait
252 for the collecting process to exit, so as not to remove the crashing
253 processes proc files prematurely. This in turn creates the
254 possibility that a misbehaving userspace collecting process can block
255 the reaping of a crashed process simply by never exiting. This sysctl
256 defends against that. It defines how many concurrent crashing
257 processes may be piped to user space applications in parallel. If
258 this value is exceeded, then those crashing processes above that value
259 are noted via the kernel log and their cores are skipped. 0 is a
260 special value, indicating that unlimited processes may be captured in
261 parallel, but that no waiting will take place (i.e. the collecting
262 process is not guaranteed access to /proc/<crashing pid>/). This
265 ==============================================================
269 The default coredump filename is "core". By setting
270 core_uses_pid to 1, the coredump filename becomes core.PID.
271 If core_pattern does not include "%p" (default does not)
272 and core_uses_pid is set, then .PID will be appended to
275 ==============================================================
279 When the value in this file is 0, ctrl-alt-del is trapped and
280 sent to the init(1) program to handle a graceful restart.
281 When, however, the value is > 0, Linux's reaction to a Vulcan
282 Nerve Pinch (tm) will be an immediate reboot, without even
283 syncing its dirty buffers.
285 Note: when a program (like dosemu) has the keyboard in 'raw'
286 mode, the ctrl-alt-del is intercepted by the program before it
287 ever reaches the kernel tty layer, and it's up to the program
288 to decide what to do with it.
290 ==============================================================
294 This toggle indicates whether unprivileged users are prevented
295 from using dmesg(8) to view messages from the kernel's log buffer.
296 When dmesg_restrict is set to (0) there are no restrictions. When
297 dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
300 The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
301 default value of dmesg_restrict.
303 ==============================================================
305 domainname & hostname:
307 These files can be used to set the NIS/YP domainname and the
308 hostname of your box in exactly the same way as the commands
309 domainname and hostname, i.e.:
310 # echo "darkstar" > /proc/sys/kernel/hostname
311 # echo "mydomain" > /proc/sys/kernel/domainname
312 has the same effect as
313 # hostname "darkstar"
314 # domainname "mydomain"
316 Note, however, that the classic darkstar.frop.org has the
317 hostname "darkstar" and DNS (Internet Domain Name Server)
318 domainname "frop.org", not to be confused with the NIS (Network
319 Information Service) or YP (Yellow Pages) domainname. These two
320 domain names are in general different. For a detailed discussion
321 see the hostname(1) man page.
323 ==============================================================
324 hardlockup_all_cpu_backtrace:
326 This value controls the hard lockup detector behavior when a hard
327 lockup condition is detected as to whether or not to gather further
328 debug information. If enabled, arch-specific all-CPU stack dumping
331 0: do nothing. This is the default behavior.
333 1: on detection capture more debug information.
334 ==============================================================
338 Path for the hotplug policy agent.
339 Default value is "/sbin/hotplug".
341 ==============================================================
345 Controls the kernel's behavior when a hung task is detected.
346 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
348 0: continue operation. This is the default behavior.
350 1: panic immediately.
352 ==============================================================
354 hung_task_check_count:
356 The upper bound on the number of tasks that are checked.
357 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
359 ==============================================================
361 hung_task_timeout_secs:
363 Check interval. When a task in D state did not get scheduled
364 for more than this value report a warning.
365 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
367 0: means infinite timeout - no checking done.
368 Possible values to set are in range {0..LONG_MAX/HZ}.
370 ==============================================================
374 The maximum number of warnings to report. During a check interval
375 if a hung task is detected, this value is decreased by 1.
376 When this value reaches 0, no more warnings will be reported.
377 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
379 -1: report an infinite number of warnings.
381 ==============================================================
385 A toggle indicating if the kexec_load syscall has been disabled. This
386 value defaults to 0 (false: kexec_load enabled), but can be set to 1
387 (true: kexec_load disabled). Once true, kexec can no longer be used, and
388 the toggle cannot be set back to false. This allows a kexec image to be
389 loaded before disabling the syscall, allowing a system to set up (and
390 later use) an image without it being altered. Generally used together
391 with the "modules_disabled" sysctl.
393 ==============================================================
397 This toggle indicates whether restrictions are placed on
398 exposing kernel addresses via /proc and other interfaces.
400 When kptr_restrict is set to (0), the default, there are no restrictions.
402 When kptr_restrict is set to (1), kernel pointers printed using the %pK
403 format specifier will be replaced with 0's unless the user has CAP_SYSLOG
404 and effective user and group ids are equal to the real ids. This is
405 because %pK checks are done at read() time rather than open() time, so
406 if permissions are elevated between the open() and the read() (e.g via
407 a setuid binary) then %pK will not leak kernel pointers to unprivileged
408 users. Note, this is a temporary solution only. The correct long-term
409 solution is to do the permission checks at open() time. Consider removing
410 world read permissions from files that use %pK, and using dmesg_restrict
411 to protect against uses of %pK in dmesg(8) if leaking kernel pointer
412 values to unprivileged users is a concern.
414 When kptr_restrict is set to (2), kernel pointers printed using
415 %pK will be replaced with 0's regardless of privileges.
417 ==============================================================
419 kstack_depth_to_print: (X86 only)
421 Controls the number of words to print when dumping the raw
424 ==============================================================
428 This flag controls the L2 cache of G3 processor boards. If
429 0, the cache is disabled. Enabled if nonzero.
431 ==============================================================
435 A toggle value indicating if modules are allowed to be loaded
436 in an otherwise modular kernel. This toggle defaults to off
437 (0), but can be set true (1). Once true, modules can be
438 neither loaded nor unloaded, and the toggle cannot be set back
439 to false. Generally used with the "kexec_load_disabled" toggle.
441 ==============================================================
443 msg_next_id, sem_next_id, and shm_next_id:
445 These three toggles allows to specify desired id for next allocated IPC
446 object: message, semaphore or shared memory respectively.
448 By default they are equal to -1, which means generic allocation logic.
449 Possible values to set are in range {0..INT_MAX}.
452 1) kernel doesn't guarantee, that new object will have desired id. So,
453 it's up to userspace, how to handle an object with "wrong" id.
454 2) Toggle with non-default value will be set back to -1 by kernel after
455 successful IPC object allocation.
457 ==============================================================
461 This parameter can be used to control the NMI watchdog
462 (i.e. the hard lockup detector) on x86 systems.
464 0 - disable the hard lockup detector
465 1 - enable the hard lockup detector
467 The hard lockup detector monitors each CPU for its ability to respond to
468 timer interrupts. The mechanism utilizes CPU performance counter registers
469 that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
470 while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
472 The NMI watchdog is disabled by default if the kernel is running as a guest
473 in a KVM virtual machine. This default can be overridden by adding
477 to the guest kernel command line (see Documentation/kernel-parameters.txt).
479 ==============================================================
483 Enables/disables automatic page fault based NUMA memory
484 balancing. Memory is moved automatically to nodes
485 that access it often.
487 Enables/disables automatic NUMA memory balancing. On NUMA machines, there
488 is a performance penalty if remote memory is accessed by a CPU. When this
489 feature is enabled the kernel samples what task thread is accessing memory
490 by periodically unmapping pages and later trapping a page fault. At the
491 time of the page fault, it is determined if the data being accessed should
492 be migrated to a local memory node.
494 The unmapping of pages and trapping faults incur additional overhead that
495 ideally is offset by improved memory locality but there is no universal
496 guarantee. If the target workload is already bound to NUMA nodes then this
497 feature should be disabled. Otherwise, if the system overhead from the
498 feature is too high then the rate the kernel samples for NUMA hinting
499 faults may be controlled by the numa_balancing_scan_period_min_ms,
500 numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms,
501 numa_balancing_scan_size_mb, and numa_balancing_settle_count sysctls.
503 ==============================================================
505 numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms,
506 numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb
508 Automatic NUMA balancing scans tasks address space and unmaps pages to
509 detect if pages are properly placed or if the data should be migrated to a
510 memory node local to where the task is running. Every "scan delay" the task
511 scans the next "scan size" number of pages in its address space. When the
512 end of the address space is reached the scanner restarts from the beginning.
514 In combination, the "scan delay" and "scan size" determine the scan rate.
515 When "scan delay" decreases, the scan rate increases. The scan delay and
516 hence the scan rate of every task is adaptive and depends on historical
517 behaviour. If pages are properly placed then the scan delay increases,
518 otherwise the scan delay decreases. The "scan size" is not adaptive but
519 the higher the "scan size", the higher the scan rate.
521 Higher scan rates incur higher system overhead as page faults must be
522 trapped and potentially data must be migrated. However, the higher the scan
523 rate, the more quickly a tasks memory is migrated to a local node if the
524 workload pattern changes and minimises performance impact due to remote
525 memory accesses. These sysctls control the thresholds for scan delays and
526 the number of pages scanned.
528 numa_balancing_scan_period_min_ms is the minimum time in milliseconds to
529 scan a tasks virtual memory. It effectively controls the maximum scanning
532 numa_balancing_scan_delay_ms is the starting "scan delay" used for a task
533 when it initially forks.
535 numa_balancing_scan_period_max_ms is the maximum time in milliseconds to
536 scan a tasks virtual memory. It effectively controls the minimum scanning
539 numa_balancing_scan_size_mb is how many megabytes worth of pages are
540 scanned for a given scan.
542 ==============================================================
544 osrelease, ostype & version:
551 #5 Wed Feb 25 21:49:24 MET 1998
553 The files osrelease and ostype should be clear enough. Version
554 needs a little more clarification however. The '#5' means that
555 this is the fifth kernel built from this source base and the
556 date behind it indicates the time the kernel was built.
557 The only way to tune these values is to rebuild the kernel :-)
559 ==============================================================
561 overflowgid & overflowuid:
563 if your architecture did not always support 32-bit UIDs (i.e. arm,
564 i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
565 applications that use the old 16-bit UID/GID system calls, if the
566 actual UID or GID would exceed 65535.
568 These sysctls allow you to change the value of the fixed UID and GID.
569 The default is 65534.
571 ==============================================================
575 The value in this file represents the number of seconds the kernel
576 waits before rebooting on a panic. When you use the software watchdog,
577 the recommended setting is 60.
579 ==============================================================
583 Controls the kernel's behaviour when an oops or BUG is encountered.
585 0: try to continue operation
587 1: panic immediately. If the `panic' sysctl is also non-zero then the
588 machine will be rebooted.
590 ==============================================================
592 panic_on_stackoverflow:
594 Controls the kernel's behavior when detecting the overflows of
595 kernel, IRQ and exception stacks except a user stack.
596 This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
598 0: try to continue operation.
600 1: panic immediately.
602 ==============================================================
606 Calls panic() in the WARN() path when set to 1. This is useful to avoid
607 a kernel rebuild when attempting to kdump at the location of a WARN().
609 0: only WARN(), default behaviour.
611 1: call panic() after printing out WARN() location.
613 ==============================================================
615 perf_cpu_time_max_percent:
617 Hints to the kernel how much CPU time it should be allowed to
618 use to handle perf sampling events. If the perf subsystem
619 is informed that its samples are exceeding this limit, it
620 will drop its sampling frequency to attempt to reduce its CPU
623 Some perf sampling happens in NMIs. If these samples
624 unexpectedly take too long to execute, the NMIs can become
625 stacked up next to each other so much that nothing else is
628 0: disable the mechanism. Do not monitor or correct perf's
629 sampling rate no matter how CPU time it takes.
631 1-100: attempt to throttle perf's sample rate to this
632 percentage of CPU. Note: the kernel calculates an
633 "expected" length of each sample event. 100 here means
634 100% of that expected length. Even if this is set to
635 100, you may still see sample throttling if this
636 length is exceeded. Set to 0 if you truly do not care
637 how much CPU is consumed.
639 ==============================================================
641 panic_on_unrecovered_nmi:
643 The default Linux behaviour on an NMI of either memory or unknown is
644 to continue operation. For many environments such as scientific
645 computing it is preferable that the box is taken out and the error
646 dealt with than an uncorrected parity/ECC error get propagated.
648 A small number of systems do generate NMI's for bizarre random reasons
649 such as power management so the default is off. That sysctl works like
650 the existing panic controls already in that directory.
652 ==============================================================
656 Controls use of the performance events system by unprivileged
657 users (without CAP_SYS_ADMIN). The default value is 3 if
658 CONFIG_SECURITY_PERF_EVENTS_RESTRICT is set, or 1 otherwise.
660 -1: Allow use of (almost) all events by all users
661 >=0: Disallow raw tracepoint access by users without CAP_IOC_LOCK
662 >=1: Disallow CPU event access by users without CAP_SYS_ADMIN
663 >=2: Disallow kernel profiling by users without CAP_SYS_ADMIN
664 >=3: Disallow all event access by users without CAP_SYS_ADMIN
666 ==============================================================
670 PID allocation wrap value. When the kernel's next PID value
671 reaches this value, it wraps back to a minimum PID value.
672 PIDs of value pid_max or larger are not allocated.
674 ==============================================================
678 The last pid allocated in the current (the one task using this sysctl
679 lives in) pid namespace. When selecting a pid for a next task on fork
680 kernel tries to allocate a number starting from this one.
682 ==============================================================
684 powersave-nap: (PPC only)
686 If set, Linux-PPC will use the 'nap' mode of powersaving,
687 otherwise the 'doze' mode will be used.
689 ==============================================================
693 The four values in printk denote: console_loglevel,
694 default_message_loglevel, minimum_console_loglevel and
695 default_console_loglevel respectively.
697 These values influence printk() behavior when printing or
698 logging error messages. See 'man 2 syslog' for more info on
699 the different loglevels.
701 - console_loglevel: messages with a higher priority than
702 this will be printed to the console
703 - default_message_loglevel: messages without an explicit priority
704 will be printed with this priority
705 - minimum_console_loglevel: minimum (highest) value to which
706 console_loglevel can be set
707 - default_console_loglevel: default value for console_loglevel
709 ==============================================================
713 Delay each printk message in printk_delay milliseconds
715 Value from 0 - 10000 is allowed.
717 ==============================================================
721 Some warning messages are rate limited. printk_ratelimit specifies
722 the minimum length of time between these messages (in jiffies), by
723 default we allow one every 5 seconds.
725 A value of 0 will disable rate limiting.
727 ==============================================================
729 printk_ratelimit_burst:
731 While long term we enforce one message per printk_ratelimit
732 seconds, we do allow a burst of messages to pass through.
733 printk_ratelimit_burst specifies the number of messages we can
734 send before ratelimiting kicks in.
736 ==============================================================
740 This option can be used to select the type of process address
741 space randomization that is used in the system, for architectures
742 that support this feature.
744 0 - Turn the process address space randomization off. This is the
745 default for architectures that do not support this feature anyways,
746 and kernels that are booted with the "norandmaps" parameter.
748 1 - Make the addresses of mmap base, stack and VDSO page randomized.
749 This, among other things, implies that shared libraries will be
750 loaded to random addresses. Also for PIE-linked binaries, the
751 location of code start is randomized. This is the default if the
752 CONFIG_COMPAT_BRK option is enabled.
754 2 - Additionally enable heap randomization. This is the default if
755 CONFIG_COMPAT_BRK is disabled.
757 There are a few legacy applications out there (such as some ancient
758 versions of libc.so.5 from 1996) that assume that brk area starts
759 just after the end of the code+bss. These applications break when
760 start of the brk area is randomized. There are however no known
761 non-legacy applications that would be broken this way, so for most
762 systems it is safe to choose full randomization.
764 Systems with ancient and/or broken binaries should be configured
765 with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
766 address space randomization.
768 ==============================================================
770 reboot-cmd: (Sparc only)
772 ??? This seems to be a way to give an argument to the Sparc
773 ROM/Flash boot loader. Maybe to tell it what to do after
776 ==============================================================
778 rtsig-max & rtsig-nr:
780 The file rtsig-max can be used to tune the maximum number
781 of POSIX realtime (queued) signals that can be outstanding
784 rtsig-nr shows the number of RT signals currently queued.
786 ==============================================================
790 This file shows the size of the generic SCSI (sg) buffer.
791 You can't tune it just yet, but you could change it on
792 compile time by editing include/scsi/sg.h and changing
793 the value of SG_BIG_BUFF.
795 There shouldn't be any reason to change this value. If
796 you can come up with one, you probably know what you
799 ==============================================================
803 This parameter sets the total amount of shared memory pages that
804 can be used system wide. Hence, SHMALL should always be at least
805 ceil(shmmax/PAGE_SIZE).
807 If you are not sure what the default PAGE_SIZE is on your Linux
808 system, you can run the following command:
812 ==============================================================
816 This value can be used to query and set the run time limit
817 on the maximum shared memory segment size that can be created.
818 Shared memory segments up to 1Gb are now supported in the
819 kernel. This value defaults to SHMMAX.
821 ==============================================================
825 Linux lets you set resource limits, including how much memory one
826 process can consume, via setrlimit(2). Unfortunately, shared memory
827 segments are allowed to exist without association with any process, and
828 thus might not be counted against any resource limits. If enabled,
829 shared memory segments are automatically destroyed when their attach
830 count becomes zero after a detach or a process termination. It will
831 also destroy segments that were created, but never attached to, on exit
832 from the process. The only use left for IPC_RMID is to immediately
833 destroy an unattached segment. Of course, this breaks the way things are
834 defined, so some applications might stop working. Note that this
835 feature will do you no good unless you also configure your resource
836 limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't
839 Note that if you change this from 0 to 1, already created segments
840 without users and with a dead originative process will be destroyed.
842 ==============================================================
844 sysctl_writes_strict:
846 Control how file position affects the behavior of updating sysctl values
847 via the /proc/sys interface:
849 -1 - Legacy per-write sysctl value handling, with no printk warnings.
850 Each write syscall must fully contain the sysctl value to be
851 written, and multiple writes on the same sysctl file descriptor
852 will rewrite the sysctl value, regardless of file position.
853 0 - Same behavior as above, but warn about processes that perform writes
854 to a sysctl file descriptor when the file position is not 0.
855 1 - (default) Respect file position when writing sysctl strings. Multiple
856 writes will append to the sysctl value buffer. Anything past the max
857 length of the sysctl value buffer will be ignored. Writes to numeric
858 sysctl entries must always be at file position 0 and the value must
859 be fully contained in the buffer sent in the write syscall.
861 ==============================================================
863 softlockup_all_cpu_backtrace:
865 This value controls the soft lockup detector thread's behavior
866 when a soft lockup condition is detected as to whether or not
867 to gather further debug information. If enabled, each cpu will
868 be issued an NMI and instructed to capture stack trace.
870 This feature is only applicable for architectures which support
873 0: do nothing. This is the default behavior.
875 1: on detection capture more debug information.
877 ==============================================================
881 This parameter can be used to control the soft lockup detector.
883 0 - disable the soft lockup detector
884 1 - enable the soft lockup detector
886 The soft lockup detector monitors CPUs for threads that are hogging the CPUs
887 without rescheduling voluntarily, and thus prevent the 'watchdog/N' threads
888 from running. The mechanism depends on the CPUs ability to respond to timer
889 interrupts which are needed for the 'watchdog/N' threads to be woken up by
890 the watchdog timer function, otherwise the NMI watchdog - if enabled - can
891 detect a hard lockup condition.
893 ==============================================================
897 Non-zero if the kernel has been tainted. Numeric values, which
898 can be ORed together:
900 1 - A module with a non-GPL license has been loaded, this
901 includes modules with no license.
902 Set by modutils >= 2.4.9 and module-init-tools.
903 2 - A module was force loaded by insmod -f.
904 Set by modutils >= 2.4.9 and module-init-tools.
905 4 - Unsafe SMP processors: SMP with CPUs not designed for SMP.
906 8 - A module was forcibly unloaded from the system by rmmod -f.
907 16 - A hardware machine check error occurred on the system.
908 32 - A bad page was discovered on the system.
909 64 - The user has asked that the system be marked "tainted". This
910 could be because they are running software that directly modifies
911 the hardware, or for other reasons.
912 128 - The system has died.
913 256 - The ACPI DSDT has been overridden with one supplied by the user
914 instead of using the one provided by the hardware.
915 512 - A kernel warning has occurred.
916 1024 - A module from drivers/staging was loaded.
917 2048 - The system is working around a severe firmware bug.
918 4096 - An out-of-tree module has been loaded.
919 8192 - An unsigned module has been loaded in a kernel supporting module
921 16384 - A soft lockup has previously occurred on the system.
922 32768 - The kernel has been live patched.
924 ==============================================================
928 This value controls the maximum number of threads that can be created
931 During initialization the kernel sets this value such that even if the
932 maximum number of threads is created, the thread structures occupy only
933 a part (1/8th) of the available RAM pages.
935 The minimum value that can be written to threads-max is 20.
936 The maximum value that can be written to threads-max is given by the
937 constant FUTEX_TID_MASK (0x3fffffff).
938 If a value outside of this range is written to threads-max an error
941 The value written is checked against the available RAM pages. If the
942 thread structures would occupy too much (more than 1/8th) of the
943 available RAM pages threads-max is reduced accordingly.
945 ==============================================================
949 The value in this file affects behavior of handling NMI. When the
950 value is non-zero, unknown NMI is trapped and then panic occurs. At
951 that time, kernel debugging information is displayed on console.
953 NMI switch that most IA32 servers have fires unknown NMI up, for
954 example. If a system hangs up, try pressing the NMI switch.
956 ==============================================================
960 This parameter can be used to disable or enable the soft lockup detector
961 _and_ the NMI watchdog (i.e. the hard lockup detector) at the same time.
963 0 - disable both lockup detectors
964 1 - enable both lockup detectors
966 The soft lockup detector and the NMI watchdog can also be disabled or
967 enabled individually, using the soft_watchdog and nmi_watchdog parameters.
968 If the watchdog parameter is read, for example by executing
970 cat /proc/sys/kernel/watchdog
972 the output of this command (0 or 1) shows the logical OR of soft_watchdog
975 ==============================================================
979 This value can be used to control on which cpus the watchdog may run.
980 The default cpumask is all possible cores, but if NO_HZ_FULL is
981 enabled in the kernel config, and cores are specified with the
982 nohz_full= boot argument, those cores are excluded by default.
983 Offline cores can be included in this mask, and if the core is later
984 brought online, the watchdog will be started based on the mask value.
986 Typically this value would only be touched in the nohz_full case
987 to re-enable cores that by default were not running the watchdog,
988 if a kernel lockup was suspected on those cores.
990 The argument value is the standard cpulist format for cpumasks,
991 so for example to enable the watchdog on cores 0, 2, 3, and 4 you
994 echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
996 ==============================================================
1000 This value can be used to control the frequency of hrtimer and NMI
1001 events and the soft and hard lockup thresholds. The default threshold
1004 The softlockup threshold is (2 * watchdog_thresh). Setting this
1005 tunable to zero will disable lockup detection altogether.
1007 ==============================================================