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 ]
83 - seccomp/ ==> Documentation/userspace-api/seccomp_filter.rst
85 - sem_next_id [ sysv ipc ]
86 - sg-big-buff [ generic SCSI device (sg) ]
87 - shm_next_id [ sysv ipc ]
92 - softlockup_all_cpu_backtrace
95 - stop-a [ SPARC only ]
96 - sysrq ==> Documentation/admin-guide/sysrq.rst
97 - sysctl_writes_strict
98 - tainted ==> Documentation/admin-guide/tainted-kernels.rst
105 ==============================================================
109 highwater lowwater frequency
111 If BSD-style process accounting is enabled these values control
112 its behaviour. If free space on filesystem where the log lives
113 goes below <lowwater>% accounting suspends. If free space gets
114 above <highwater>% accounting resumes. <Frequency> determines
115 how often do we check the amount of free space (value is in
118 That is, suspend accounting if there left <= 2% free; resume it
119 if we got >=4%; consider information about amount of free space
120 valid for 30 seconds.
122 ==============================================================
128 See Doc*/kernel/power/video.txt, it allows mode of video boot to be
131 ==============================================================
135 This variable has no effect and may be removed in future kernel
136 releases. Reading it always returns 0.
137 Up to Linux 3.17, it enabled/disabled automatic recomputing of msgmni
138 upon memory add/remove or upon ipc namespace creation/removal.
139 Echoing "1" into this file enabled msgmni automatic recomputing.
140 Echoing "0" turned it off. auto_msgmni default value was 1.
143 ==============================================================
147 x86 bootloader identification
149 This gives the bootloader type number as indicated by the bootloader,
150 shifted left by 4, and OR'd with the low four bits of the bootloader
151 version. The reason for this encoding is that this used to match the
152 type_of_loader field in the kernel header; the encoding is kept for
153 backwards compatibility. That is, if the full bootloader type number
154 is 0x15 and the full version number is 0x234, this file will contain
155 the value 340 = 0x154.
157 See the type_of_loader and ext_loader_type fields in
158 Documentation/x86/boot.txt for additional information.
160 ==============================================================
164 x86 bootloader version
166 The complete bootloader version number. In the example above, this
167 file will contain the value 564 = 0x234.
169 See the type_of_loader and ext_loader_ver fields in
170 Documentation/x86/boot.txt for additional information.
172 ==============================================================
176 Controls the kernel's callhome behavior in case of a kernel panic.
178 The s390 hardware allows an operating system to send a notification
179 to a service organization (callhome) in case of an operating system panic.
181 When the value in this file is 0 (which is the default behavior)
182 nothing happens in case of a kernel panic. If this value is set to "1"
183 the complete kernel oops message is send to the IBM customer service
184 organization in case the mainframe the Linux operating system is running
185 on has a service contract with IBM.
187 ==============================================================
191 Highest valid capability of the running kernel. Exports
192 CAP_LAST_CAP from the kernel.
194 ==============================================================
198 core_pattern is used to specify a core dumpfile pattern name.
199 . max length 128 characters; default value is "core"
200 . core_pattern is used as a pattern template for the output filename;
201 certain string patterns (beginning with '%') are substituted with
203 . backward compatibility with core_uses_pid:
204 If core_pattern does not include "%p" (default does not)
205 and core_uses_pid is set, then .PID will be appended to
207 . corename format specifiers:
208 %<NUL> '%' is dropped
211 %P global pid (init PID namespace)
213 %I global tid (init PID namespace)
214 %u uid (in initial user namespace)
215 %g gid (in initial user namespace)
216 %d dump mode, matches PR_SET_DUMPABLE and
217 /proc/sys/fs/suid_dumpable
221 %e executable filename (may be shortened)
223 %<OTHER> both are dropped
224 . If the first character of the pattern is a '|', the kernel will treat
225 the rest of the pattern as a command to run. The core dump will be
226 written to the standard input of that program instead of to a file.
228 ==============================================================
232 This sysctl is only applicable when core_pattern is configured to pipe
233 core files to a user space helper (when the first character of
234 core_pattern is a '|', see above). When collecting cores via a pipe
235 to an application, it is occasionally useful for the collecting
236 application to gather data about the crashing process from its
237 /proc/pid directory. In order to do this safely, the kernel must wait
238 for the collecting process to exit, so as not to remove the crashing
239 processes proc files prematurely. This in turn creates the
240 possibility that a misbehaving userspace collecting process can block
241 the reaping of a crashed process simply by never exiting. This sysctl
242 defends against that. It defines how many concurrent crashing
243 processes may be piped to user space applications in parallel. If
244 this value is exceeded, then those crashing processes above that value
245 are noted via the kernel log and their cores are skipped. 0 is a
246 special value, indicating that unlimited processes may be captured in
247 parallel, but that no waiting will take place (i.e. the collecting
248 process is not guaranteed access to /proc/<crashing pid>/). This
251 ==============================================================
255 The default coredump filename is "core". By setting
256 core_uses_pid to 1, the coredump filename becomes core.PID.
257 If core_pattern does not include "%p" (default does not)
258 and core_uses_pid is set, then .PID will be appended to
261 ==============================================================
265 When the value in this file is 0, ctrl-alt-del is trapped and
266 sent to the init(1) program to handle a graceful restart.
267 When, however, the value is > 0, Linux's reaction to a Vulcan
268 Nerve Pinch (tm) will be an immediate reboot, without even
269 syncing its dirty buffers.
271 Note: when a program (like dosemu) has the keyboard in 'raw'
272 mode, the ctrl-alt-del is intercepted by the program before it
273 ever reaches the kernel tty layer, and it's up to the program
274 to decide what to do with it.
276 ==============================================================
280 This toggle indicates whether unprivileged users are prevented
281 from using dmesg(8) to view messages from the kernel's log buffer.
282 When dmesg_restrict is set to (0) there are no restrictions. When
283 dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
286 The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
287 default value of dmesg_restrict.
289 ==============================================================
291 domainname & hostname:
293 These files can be used to set the NIS/YP domainname and the
294 hostname of your box in exactly the same way as the commands
295 domainname and hostname, i.e.:
296 # echo "darkstar" > /proc/sys/kernel/hostname
297 # echo "mydomain" > /proc/sys/kernel/domainname
298 has the same effect as
299 # hostname "darkstar"
300 # domainname "mydomain"
302 Note, however, that the classic darkstar.frop.org has the
303 hostname "darkstar" and DNS (Internet Domain Name Server)
304 domainname "frop.org", not to be confused with the NIS (Network
305 Information Service) or YP (Yellow Pages) domainname. These two
306 domain names are in general different. For a detailed discussion
307 see the hostname(1) man page.
309 ==============================================================
310 hardlockup_all_cpu_backtrace:
312 This value controls the hard lockup detector behavior when a hard
313 lockup condition is detected as to whether or not to gather further
314 debug information. If enabled, arch-specific all-CPU stack dumping
317 0: do nothing. This is the default behavior.
319 1: on detection capture more debug information.
320 ==============================================================
324 This parameter can be used to control whether the kernel panics
325 when a hard lockup is detected.
327 0 - don't panic on hard lockup
328 1 - panic on hard lockup
330 See Documentation/lockup-watchdogs.txt for more information. This can
331 also be set using the nmi_watchdog kernel parameter.
333 ==============================================================
337 Path for the hotplug policy agent.
338 Default value is "/sbin/hotplug".
340 ==============================================================
344 Controls the kernel's behavior when a hung task is detected.
345 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
347 0: continue operation. This is the default behavior.
349 1: panic immediately.
351 ==============================================================
353 hung_task_check_count:
355 The upper bound on the number of tasks that are checked.
356 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
358 ==============================================================
360 hung_task_timeout_secs:
362 When a task in D state did not get scheduled
363 for more than this value report a warning.
364 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
366 0: means infinite timeout - no checking done.
367 Possible values to set are in range {0..LONG_MAX/HZ}.
369 ==============================================================
371 hung_task_check_interval_secs:
373 Hung task check interval. If hung task checking is enabled
374 (see hung_task_timeout_secs), the check is done every
375 hung_task_check_interval_secs seconds.
376 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
378 0 (default): means use hung_task_timeout_secs as checking interval.
379 Possible values to set are in range {0..LONG_MAX/HZ}.
381 ==============================================================
385 The maximum number of warnings to report. During a check interval
386 if a hung task is detected, this value is decreased by 1.
387 When this value reaches 0, no more warnings will be reported.
388 This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
390 -1: report an infinite number of warnings.
392 ==============================================================
394 hyperv_record_panic_msg:
396 Controls whether the panic kmsg data should be reported to Hyper-V.
398 0: do not report panic kmsg data.
400 1: report the panic kmsg data. This is the default behavior.
402 ==============================================================
406 A toggle indicating if the kexec_load syscall has been disabled. This
407 value defaults to 0 (false: kexec_load enabled), but can be set to 1
408 (true: kexec_load disabled). Once true, kexec can no longer be used, and
409 the toggle cannot be set back to false. This allows a kexec image to be
410 loaded before disabling the syscall, allowing a system to set up (and
411 later use) an image without it being altered. Generally used together
412 with the "modules_disabled" sysctl.
414 ==============================================================
418 This toggle indicates whether restrictions are placed on
419 exposing kernel addresses via /proc and other interfaces.
421 When kptr_restrict is set to 0 (the default) the address is hashed before
422 printing. (This is the equivalent to %p.)
424 When kptr_restrict is set to (1), kernel pointers printed using the %pK
425 format specifier will be replaced with 0's unless the user has CAP_SYSLOG
426 and effective user and group ids are equal to the real ids. This is
427 because %pK checks are done at read() time rather than open() time, so
428 if permissions are elevated between the open() and the read() (e.g via
429 a setuid binary) then %pK will not leak kernel pointers to unprivileged
430 users. Note, this is a temporary solution only. The correct long-term
431 solution is to do the permission checks at open() time. Consider removing
432 world read permissions from files that use %pK, and using dmesg_restrict
433 to protect against uses of %pK in dmesg(8) if leaking kernel pointer
434 values to unprivileged users is a concern.
436 When kptr_restrict is set to (2), kernel pointers printed using
437 %pK will be replaced with 0's regardless of privileges.
439 ==============================================================
443 This flag controls the L2 cache of G3 processor boards. If
444 0, the cache is disabled. Enabled if nonzero.
446 ==============================================================
450 A toggle value indicating if modules are allowed to be loaded
451 in an otherwise modular kernel. This toggle defaults to off
452 (0), but can be set true (1). Once true, modules can be
453 neither loaded nor unloaded, and the toggle cannot be set back
454 to false. Generally used with the "kexec_load_disabled" toggle.
456 ==============================================================
458 msg_next_id, sem_next_id, and shm_next_id:
460 These three toggles allows to specify desired id for next allocated IPC
461 object: message, semaphore or shared memory respectively.
463 By default they are equal to -1, which means generic allocation logic.
464 Possible values to set are in range {0..INT_MAX}.
467 1) kernel doesn't guarantee, that new object will have desired id. So,
468 it's up to userspace, how to handle an object with "wrong" id.
469 2) Toggle with non-default value will be set back to -1 by kernel after
470 successful IPC object allocation. If an IPC object allocation syscall
471 fails, it is undefined if the value remains unmodified or is reset to -1.
473 ==============================================================
477 This parameter can be used to control the NMI watchdog
478 (i.e. the hard lockup detector) on x86 systems.
480 0 - disable the hard lockup detector
481 1 - enable the hard lockup detector
483 The hard lockup detector monitors each CPU for its ability to respond to
484 timer interrupts. The mechanism utilizes CPU performance counter registers
485 that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
486 while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
488 The NMI watchdog is disabled by default if the kernel is running as a guest
489 in a KVM virtual machine. This default can be overridden by adding
493 to the guest kernel command line (see Documentation/admin-guide/kernel-parameters.rst).
495 ==============================================================
499 Enables/disables automatic page fault based NUMA memory
500 balancing. Memory is moved automatically to nodes
501 that access it often.
503 Enables/disables automatic NUMA memory balancing. On NUMA machines, there
504 is a performance penalty if remote memory is accessed by a CPU. When this
505 feature is enabled the kernel samples what task thread is accessing memory
506 by periodically unmapping pages and later trapping a page fault. At the
507 time of the page fault, it is determined if the data being accessed should
508 be migrated to a local memory node.
510 The unmapping of pages and trapping faults incur additional overhead that
511 ideally is offset by improved memory locality but there is no universal
512 guarantee. If the target workload is already bound to NUMA nodes then this
513 feature should be disabled. Otherwise, if the system overhead from the
514 feature is too high then the rate the kernel samples for NUMA hinting
515 faults may be controlled by the numa_balancing_scan_period_min_ms,
516 numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms,
517 numa_balancing_scan_size_mb, and numa_balancing_settle_count sysctls.
519 ==============================================================
521 numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms,
522 numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb
524 Automatic NUMA balancing scans tasks address space and unmaps pages to
525 detect if pages are properly placed or if the data should be migrated to a
526 memory node local to where the task is running. Every "scan delay" the task
527 scans the next "scan size" number of pages in its address space. When the
528 end of the address space is reached the scanner restarts from the beginning.
530 In combination, the "scan delay" and "scan size" determine the scan rate.
531 When "scan delay" decreases, the scan rate increases. The scan delay and
532 hence the scan rate of every task is adaptive and depends on historical
533 behaviour. If pages are properly placed then the scan delay increases,
534 otherwise the scan delay decreases. The "scan size" is not adaptive but
535 the higher the "scan size", the higher the scan rate.
537 Higher scan rates incur higher system overhead as page faults must be
538 trapped and potentially data must be migrated. However, the higher the scan
539 rate, the more quickly a tasks memory is migrated to a local node if the
540 workload pattern changes and minimises performance impact due to remote
541 memory accesses. These sysctls control the thresholds for scan delays and
542 the number of pages scanned.
544 numa_balancing_scan_period_min_ms is the minimum time in milliseconds to
545 scan a tasks virtual memory. It effectively controls the maximum scanning
548 numa_balancing_scan_delay_ms is the starting "scan delay" used for a task
549 when it initially forks.
551 numa_balancing_scan_period_max_ms is the maximum time in milliseconds to
552 scan a tasks virtual memory. It effectively controls the minimum scanning
555 numa_balancing_scan_size_mb is how many megabytes worth of pages are
556 scanned for a given scan.
558 ==============================================================
560 osrelease, ostype & version:
567 #5 Wed Feb 25 21:49:24 MET 1998
569 The files osrelease and ostype should be clear enough. Version
570 needs a little more clarification however. The '#5' means that
571 this is the fifth kernel built from this source base and the
572 date behind it indicates the time the kernel was built.
573 The only way to tune these values is to rebuild the kernel :-)
575 ==============================================================
577 overflowgid & overflowuid:
579 if your architecture did not always support 32-bit UIDs (i.e. arm,
580 i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
581 applications that use the old 16-bit UID/GID system calls, if the
582 actual UID or GID would exceed 65535.
584 These sysctls allow you to change the value of the fixed UID and GID.
585 The default is 65534.
587 ==============================================================
591 The value in this file represents the number of seconds the kernel
592 waits before rebooting on a panic. When you use the software watchdog,
593 the recommended setting is 60.
595 ==============================================================
599 Controls the kernel's behavior when a CPU receives an NMI caused by
602 0: try to continue operation (default)
604 1: panic immediately. The IO error triggered an NMI. This indicates a
605 serious system condition which could result in IO data corruption.
606 Rather than continuing, panicking might be a better choice. Some
607 servers issue this sort of NMI when the dump button is pushed,
608 and you can use this option to take a crash dump.
610 ==============================================================
614 Controls the kernel's behaviour when an oops or BUG is encountered.
616 0: try to continue operation
618 1: panic immediately. If the `panic' sysctl is also non-zero then the
619 machine will be rebooted.
621 ==============================================================
623 panic_on_stackoverflow:
625 Controls the kernel's behavior when detecting the overflows of
626 kernel, IRQ and exception stacks except a user stack.
627 This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
629 0: try to continue operation.
631 1: panic immediately.
633 ==============================================================
635 panic_on_unrecovered_nmi:
637 The default Linux behaviour on an NMI of either memory or unknown is
638 to continue operation. For many environments such as scientific
639 computing it is preferable that the box is taken out and the error
640 dealt with than an uncorrected parity/ECC error get propagated.
642 A small number of systems do generate NMI's for bizarre random reasons
643 such as power management so the default is off. That sysctl works like
644 the existing panic controls already in that directory.
646 ==============================================================
650 Calls panic() in the WARN() path when set to 1. This is useful to avoid
651 a kernel rebuild when attempting to kdump at the location of a WARN().
653 0: only WARN(), default behaviour.
655 1: call panic() after printing out WARN() location.
657 ==============================================================
661 Bitmask for printing system info when panic happens. User can chose
662 combination of the following bits:
664 bit 0: print all tasks info
665 bit 1: print system memory info
666 bit 2: print timer info
667 bit 3: print locks info if CONFIG_LOCKDEP is on
668 bit 4: print ftrace buffer
670 So for example to print tasks and memory info on panic, user can:
671 echo 3 > /proc/sys/kernel/panic_print
673 ==============================================================
677 When set to 1, calls panic() after RCU stall detection messages. This
678 is useful to define the root cause of RCU stalls using a vmcore.
680 0: do not panic() when RCU stall takes place, default behavior.
682 1: panic() after printing RCU stall messages.
684 ==============================================================
686 perf_cpu_time_max_percent:
688 Hints to the kernel how much CPU time it should be allowed to
689 use to handle perf sampling events. If the perf subsystem
690 is informed that its samples are exceeding this limit, it
691 will drop its sampling frequency to attempt to reduce its CPU
694 Some perf sampling happens in NMIs. If these samples
695 unexpectedly take too long to execute, the NMIs can become
696 stacked up next to each other so much that nothing else is
699 0: disable the mechanism. Do not monitor or correct perf's
700 sampling rate no matter how CPU time it takes.
702 1-100: attempt to throttle perf's sample rate to this
703 percentage of CPU. Note: the kernel calculates an
704 "expected" length of each sample event. 100 here means
705 100% of that expected length. Even if this is set to
706 100, you may still see sample throttling if this
707 length is exceeded. Set to 0 if you truly do not care
708 how much CPU is consumed.
710 ==============================================================
714 Controls use of the performance events system by unprivileged
715 users (without CAP_SYS_ADMIN). The default value is 2.
717 -1: Allow use of (almost) all events by all users
718 Ignore mlock limit after perf_event_mlock_kb without CAP_IPC_LOCK
719 >=0: Disallow ftrace function tracepoint by users without CAP_SYS_ADMIN
720 Disallow raw tracepoint access by users without CAP_SYS_ADMIN
721 >=1: Disallow CPU event access by users without CAP_SYS_ADMIN
722 >=2: Disallow kernel profiling by users without CAP_SYS_ADMIN
724 ==============================================================
726 perf_event_max_stack:
728 Controls maximum number of stack frames to copy for (attr.sample_type &
729 PERF_SAMPLE_CALLCHAIN) configured events, for instance, when using
730 'perf record -g' or 'perf trace --call-graph fp'.
732 This can only be done when no events are in use that have callchains
733 enabled, otherwise writing to this file will return -EBUSY.
735 The default value is 127.
737 ==============================================================
741 Control size of per-cpu ring buffer not counted agains mlock limit.
743 The default value is 512 + 1 page
745 ==============================================================
747 perf_event_max_contexts_per_stack:
749 Controls maximum number of stack frame context entries for
750 (attr.sample_type & PERF_SAMPLE_CALLCHAIN) configured events, for
751 instance, when using 'perf record -g' or 'perf trace --call-graph fp'.
753 This can only be done when no events are in use that have callchains
754 enabled, otherwise writing to this file will return -EBUSY.
756 The default value is 8.
758 ==============================================================
762 PID allocation wrap value. When the kernel's next PID value
763 reaches this value, it wraps back to a minimum PID value.
764 PIDs of value pid_max or larger are not allocated.
766 ==============================================================
770 The last pid allocated in the current (the one task using this sysctl
771 lives in) pid namespace. When selecting a pid for a next task on fork
772 kernel tries to allocate a number starting from this one.
774 ==============================================================
776 powersave-nap: (PPC only)
778 If set, Linux-PPC will use the 'nap' mode of powersaving,
779 otherwise the 'doze' mode will be used.
781 ==============================================================
785 The four values in printk denote: console_loglevel,
786 default_message_loglevel, minimum_console_loglevel and
787 default_console_loglevel respectively.
789 These values influence printk() behavior when printing or
790 logging error messages. See 'man 2 syslog' for more info on
791 the different loglevels.
793 - console_loglevel: messages with a higher priority than
794 this will be printed to the console
795 - default_message_loglevel: messages without an explicit priority
796 will be printed with this priority
797 - minimum_console_loglevel: minimum (highest) value to which
798 console_loglevel can be set
799 - default_console_loglevel: default value for console_loglevel
801 ==============================================================
805 Delay each printk message in printk_delay milliseconds
807 Value from 0 - 10000 is allowed.
809 ==============================================================
813 Some warning messages are rate limited. printk_ratelimit specifies
814 the minimum length of time between these messages (in jiffies), by
815 default we allow one every 5 seconds.
817 A value of 0 will disable rate limiting.
819 ==============================================================
821 printk_ratelimit_burst:
823 While long term we enforce one message per printk_ratelimit
824 seconds, we do allow a burst of messages to pass through.
825 printk_ratelimit_burst specifies the number of messages we can
826 send before ratelimiting kicks in.
828 ==============================================================
832 Control the logging to /dev/kmsg from userspace:
834 ratelimit: default, ratelimited
835 on: unlimited logging to /dev/kmsg from userspace
836 off: logging to /dev/kmsg disabled
838 The kernel command line parameter printk.devkmsg= overrides this and is
839 a one-time setting until next reboot: once set, it cannot be changed by
840 this sysctl interface anymore.
842 ==============================================================
846 This option can be used to select the type of process address
847 space randomization that is used in the system, for architectures
848 that support this feature.
850 0 - Turn the process address space randomization off. This is the
851 default for architectures that do not support this feature anyways,
852 and kernels that are booted with the "norandmaps" parameter.
854 1 - Make the addresses of mmap base, stack and VDSO page randomized.
855 This, among other things, implies that shared libraries will be
856 loaded to random addresses. Also for PIE-linked binaries, the
857 location of code start is randomized. This is the default if the
858 CONFIG_COMPAT_BRK option is enabled.
860 2 - Additionally enable heap randomization. This is the default if
861 CONFIG_COMPAT_BRK is disabled.
863 There are a few legacy applications out there (such as some ancient
864 versions of libc.so.5 from 1996) that assume that brk area starts
865 just after the end of the code+bss. These applications break when
866 start of the brk area is randomized. There are however no known
867 non-legacy applications that would be broken this way, so for most
868 systems it is safe to choose full randomization.
870 Systems with ancient and/or broken binaries should be configured
871 with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
872 address space randomization.
874 ==============================================================
876 reboot-cmd: (Sparc only)
878 ??? This seems to be a way to give an argument to the Sparc
879 ROM/Flash boot loader. Maybe to tell it what to do after
882 ==============================================================
884 rtsig-max & rtsig-nr:
886 The file rtsig-max can be used to tune the maximum number
887 of POSIX realtime (queued) signals that can be outstanding
890 rtsig-nr shows the number of RT signals currently queued.
892 ==============================================================
896 Enables/disables Energy Aware Scheduling (EAS). EAS starts
897 automatically on platforms where it can run (that is,
898 platforms with asymmetric CPU topologies and having an Energy
899 Model available). If your platform happens to meet the
900 requirements for EAS but you do not want to use it, change
903 ==============================================================
907 Enables/disables scheduler statistics. Enabling this feature
908 incurs a small amount of overhead in the scheduler but is
909 useful for debugging and performance tuning.
911 ==============================================================
915 This file shows the size of the generic SCSI (sg) buffer.
916 You can't tune it just yet, but you could change it on
917 compile time by editing include/scsi/sg.h and changing
918 the value of SG_BIG_BUFF.
920 There shouldn't be any reason to change this value. If
921 you can come up with one, you probably know what you
924 ==============================================================
928 This parameter sets the total amount of shared memory pages that
929 can be used system wide. Hence, SHMALL should always be at least
930 ceil(shmmax/PAGE_SIZE).
932 If you are not sure what the default PAGE_SIZE is on your Linux
933 system, you can run the following command:
937 ==============================================================
941 This value can be used to query and set the run time limit
942 on the maximum shared memory segment size that can be created.
943 Shared memory segments up to 1Gb are now supported in the
944 kernel. This value defaults to SHMMAX.
946 ==============================================================
950 Linux lets you set resource limits, including how much memory one
951 process can consume, via setrlimit(2). Unfortunately, shared memory
952 segments are allowed to exist without association with any process, and
953 thus might not be counted against any resource limits. If enabled,
954 shared memory segments are automatically destroyed when their attach
955 count becomes zero after a detach or a process termination. It will
956 also destroy segments that were created, but never attached to, on exit
957 from the process. The only use left for IPC_RMID is to immediately
958 destroy an unattached segment. Of course, this breaks the way things are
959 defined, so some applications might stop working. Note that this
960 feature will do you no good unless you also configure your resource
961 limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't
964 Note that if you change this from 0 to 1, already created segments
965 without users and with a dead originative process will be destroyed.
967 ==============================================================
969 sysctl_writes_strict:
971 Control how file position affects the behavior of updating sysctl values
972 via the /proc/sys interface:
974 -1 - Legacy per-write sysctl value handling, with no printk warnings.
975 Each write syscall must fully contain the sysctl value to be
976 written, and multiple writes on the same sysctl file descriptor
977 will rewrite the sysctl value, regardless of file position.
978 0 - Same behavior as above, but warn about processes that perform writes
979 to a sysctl file descriptor when the file position is not 0.
980 1 - (default) Respect file position when writing sysctl strings. Multiple
981 writes will append to the sysctl value buffer. Anything past the max
982 length of the sysctl value buffer will be ignored. Writes to numeric
983 sysctl entries must always be at file position 0 and the value must
984 be fully contained in the buffer sent in the write syscall.
986 ==============================================================
988 softlockup_all_cpu_backtrace:
990 This value controls the soft lockup detector thread's behavior
991 when a soft lockup condition is detected as to whether or not
992 to gather further debug information. If enabled, each cpu will
993 be issued an NMI and instructed to capture stack trace.
995 This feature is only applicable for architectures which support
998 0: do nothing. This is the default behavior.
1000 1: on detection capture more debug information.
1002 ==============================================================
1006 This parameter can be used to control the soft lockup detector.
1008 0 - disable the soft lockup detector
1009 1 - enable the soft lockup detector
1011 The soft lockup detector monitors CPUs for threads that are hogging the CPUs
1012 without rescheduling voluntarily, and thus prevent the 'watchdog/N' threads
1013 from running. The mechanism depends on the CPUs ability to respond to timer
1014 interrupts which are needed for the 'watchdog/N' threads to be woken up by
1015 the watchdog timer function, otherwise the NMI watchdog - if enabled - can
1016 detect a hard lockup condition.
1018 ==============================================================
1022 This parameter can be used to control kernel stack erasing at the end
1023 of syscalls for kernels built with CONFIG_GCC_PLUGIN_STACKLEAK.
1025 That erasing reduces the information which kernel stack leak bugs
1026 can reveal and blocks some uninitialized stack variable attacks.
1027 The tradeoff is the performance impact: on a single CPU system kernel
1028 compilation sees a 1% slowdown, other systems and workloads may vary.
1030 0: kernel stack erasing is disabled, STACKLEAK_METRICS are not updated.
1032 1: kernel stack erasing is enabled (default), it is performed before
1033 returning to the userspace at the end of syscalls.
1034 ==============================================================
1038 Non-zero if the kernel has been tainted. Numeric values, which can be
1039 ORed together. The letters are seen in "Tainted" line of Oops reports.
1041 1 (P): proprietary module was loaded
1042 2 (F): module was force loaded
1043 4 (S): SMP kernel oops on an officially SMP incapable processor
1044 8 (R): module was force unloaded
1045 16 (M): processor reported a Machine Check Exception (MCE)
1046 32 (B): bad page referenced or some unexpected page flags
1047 64 (U): taint requested by userspace application
1048 128 (D): kernel died recently, i.e. there was an OOPS or BUG
1049 256 (A): an ACPI table was overridden by user
1050 512 (W): kernel issued warning
1051 1024 (C): staging driver was loaded
1052 2048 (I): workaround for bug in platform firmware applied
1053 4096 (O): externally-built ("out-of-tree") module was loaded
1054 8192 (E): unsigned module was loaded
1055 16384 (L): soft lockup occurred
1056 32768 (K): kernel has been live patched
1057 65536 (X): Auxiliary taint, defined and used by for distros
1058 131072 (T): The kernel was built with the struct randomization plugin
1060 See Documentation/admin-guide/tainted-kernels.rst for more information.
1062 ==============================================================
1066 This value controls the maximum number of threads that can be created
1069 During initialization the kernel sets this value such that even if the
1070 maximum number of threads is created, the thread structures occupy only
1071 a part (1/8th) of the available RAM pages.
1073 The minimum value that can be written to threads-max is 20.
1074 The maximum value that can be written to threads-max is given by the
1075 constant FUTEX_TID_MASK (0x3fffffff).
1076 If a value outside of this range is written to threads-max an error
1079 The value written is checked against the available RAM pages. If the
1080 thread structures would occupy too much (more than 1/8th) of the
1081 available RAM pages threads-max is reduced accordingly.
1083 ==============================================================
1087 The value in this file affects behavior of handling NMI. When the
1088 value is non-zero, unknown NMI is trapped and then panic occurs. At
1089 that time, kernel debugging information is displayed on console.
1091 NMI switch that most IA32 servers have fires unknown NMI up, for
1092 example. If a system hangs up, try pressing the NMI switch.
1094 ==============================================================
1098 This parameter can be used to disable or enable the soft lockup detector
1099 _and_ the NMI watchdog (i.e. the hard lockup detector) at the same time.
1101 0 - disable both lockup detectors
1102 1 - enable both lockup detectors
1104 The soft lockup detector and the NMI watchdog can also be disabled or
1105 enabled individually, using the soft_watchdog and nmi_watchdog parameters.
1106 If the watchdog parameter is read, for example by executing
1108 cat /proc/sys/kernel/watchdog
1110 the output of this command (0 or 1) shows the logical OR of soft_watchdog
1113 ==============================================================
1117 This value can be used to control on which cpus the watchdog may run.
1118 The default cpumask is all possible cores, but if NO_HZ_FULL is
1119 enabled in the kernel config, and cores are specified with the
1120 nohz_full= boot argument, those cores are excluded by default.
1121 Offline cores can be included in this mask, and if the core is later
1122 brought online, the watchdog will be started based on the mask value.
1124 Typically this value would only be touched in the nohz_full case
1125 to re-enable cores that by default were not running the watchdog,
1126 if a kernel lockup was suspected on those cores.
1128 The argument value is the standard cpulist format for cpumasks,
1129 so for example to enable the watchdog on cores 0, 2, 3, and 4 you
1132 echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
1134 ==============================================================
1138 This value can be used to control the frequency of hrtimer and NMI
1139 events and the soft and hard lockup thresholds. The default threshold
1142 The softlockup threshold is (2 * watchdog_thresh). Setting this
1143 tunable to zero will disable lockup detection altogether.
1145 ==============================================================