1 ===================================
2 Documentation for /proc/sys/kernel/
3 ===================================
5 .. See scripts/check-sysctl-docs to keep this up to date
8 Copyright (c) 1998, 1999, Rik van Riel <riel@nl.linux.org>
10 Copyright (c) 2009, Shen Feng<shen@cn.fujitsu.com>
12 For general info and legal blurb, please look in
13 Documentation/admin-guide/sysctl/index.rst.
15 ------------------------------------------------------------------------------
17 This file contains documentation for the sysctl files in
18 ``/proc/sys/kernel/``.
20 The files in this directory can be used to tune and monitor
21 miscellaneous and general things in the operation of the Linux
22 kernel. Since some of the files *can* be used to screw up your
23 system, it is advisable to read both documentation and source
24 before actually making adjustments.
26 Currently, these files might (depending on your configuration)
27 show up in ``/proc/sys/kernel``:
37 highwater lowwater frequency
39 If BSD-style process accounting is enabled these values control
40 its behaviour. If free space on filesystem where the log lives
41 goes below ``lowwater``\ % accounting suspends. If free space gets
42 above ``highwater``\ % accounting resumes. ``frequency`` determines
43 how often do we check the amount of free space (value is in
50 That is, suspend accounting if free space drops below 2%; resume it
51 if it increases to at least 4%; consider information about amount of
52 free space valid for 30 seconds.
58 See Documentation/power/video.rst. This allows the video resume mode to be set,
59 in a similar fashion to the ``acpi_sleep`` kernel parameter, by
60 combining the following values:
71 The machine hardware name, the same output as ``uname -m``
72 (e.g. ``x86_64`` or ``aarch64``).
77 This variable has no effect and may be removed in future kernel
78 releases. Reading it always returns 0.
79 Up to Linux 3.17, it enabled/disabled automatic recomputing of
81 upon memory add/remove or upon IPC namespace creation/removal.
82 Echoing "1" into this file enabled msgmni automatic recomputing.
83 Echoing "0" turned it off. The default value was 1.
86 bootloader_type (x86 only)
87 ==========================
89 This gives the bootloader type number as indicated by the bootloader,
90 shifted left by 4, and OR'd with the low four bits of the bootloader
91 version. The reason for this encoding is that this used to match the
92 ``type_of_loader`` field in the kernel header; the encoding is kept for
93 backwards compatibility. That is, if the full bootloader type number
94 is 0x15 and the full version number is 0x234, this file will contain
95 the value 340 = 0x154.
97 See the ``type_of_loader`` and ``ext_loader_type`` fields in
98 Documentation/arch/x86/boot.rst for additional information.
101 bootloader_version (x86 only)
102 =============================
104 The complete bootloader version number. In the example above, this
105 file will contain the value 564 = 0x234.
107 See the ``type_of_loader`` and ``ext_loader_ver`` fields in
108 Documentation/arch/x86/boot.rst for additional information.
114 Controls whether the kernel should collect statistics on BPF programs
115 (total time spent running, number of times run...). Enabling
116 statistics causes a slight reduction in performance on each program
117 run. The statistics can be seen using ``bpftool``.
119 = ===================================
120 0 Don't collect statistics (default).
121 1 Collect statistics.
122 = ===================================
128 This is the pid which will be signalled on reboot (notably, by
129 Ctrl-Alt-Delete). Writing a value to this file which doesn't
130 correspond to a running process will result in ``-ESRCH``.
132 See also `ctrl-alt-del`_.
138 Highest valid capability of the running kernel. Exports
139 ``CAP_LAST_CAP`` from the kernel.
147 ``core_pattern`` is used to specify a core dumpfile pattern name.
149 * max length 127 characters; default value is "core"
150 * ``core_pattern`` is used as a pattern template for the output
151 filename; certain string patterns (beginning with '%') are
152 substituted with their actual values.
153 * backward compatibility with ``core_uses_pid``:
155 If ``core_pattern`` does not include "%p" (default does not)
156 and ``core_uses_pid`` is set, then .PID will be appended to
159 * corename format specifiers
161 ======== ==========================================
162 %<NUL> '%' is dropped
165 %P global pid (init PID namespace)
167 %I global tid (init PID namespace)
168 %u uid (in initial user namespace)
169 %g gid (in initial user namespace)
170 %d dump mode, matches ``PR_SET_DUMPABLE`` and
171 ``/proc/sys/fs/suid_dumpable``
175 %e executable filename (may be shortened, could be changed by prctl etc)
176 %f executable filename
178 %c maximum size of core file by resource limit RLIMIT_CORE
179 %C CPU the task ran on
180 %<OTHER> both are dropped
181 ======== ==========================================
183 * If the first character of the pattern is a '|', the kernel will treat
184 the rest of the pattern as a command to run. The core dump will be
185 written to the standard input of that program instead of to a file.
191 This sysctl is only applicable when `core_pattern`_ is configured to
192 pipe core files to a user space helper (when the first character of
193 ``core_pattern`` is a '|', see above).
194 When collecting cores via a pipe to an application, it is occasionally
195 useful for the collecting application to gather data about the
196 crashing process from its ``/proc/pid`` directory.
197 In order to do this safely, the kernel must wait for the collecting
198 process to exit, so as not to remove the crashing processes proc files
200 This in turn creates the possibility that a misbehaving userspace
201 collecting process can block the reaping of a crashed process simply
203 This sysctl defends against that.
204 It defines how many concurrent crashing processes may be piped to user
205 space applications in parallel.
206 If this value is exceeded, then those crashing processes above that
207 value are noted via the kernel log and their cores are skipped.
208 0 is a special value, indicating that unlimited processes may be
209 captured in parallel, but that no waiting will take place (i.e. the
210 collecting process is not guaranteed access to ``/proc/<crashing
212 This value defaults to 0.
218 The default coredump filename is "core". By setting
219 ``core_uses_pid`` to 1, the coredump filename becomes core.PID.
220 If `core_pattern`_ does not include "%p" (default does not)
221 and ``core_uses_pid`` is set, then .PID will be appended to
228 When the value in this file is 0, ctrl-alt-del is trapped and
229 sent to the ``init(1)`` program to handle a graceful restart.
230 When, however, the value is > 0, Linux's reaction to a Vulcan
231 Nerve Pinch (tm) will be an immediate reboot, without even
232 syncing its dirty buffers.
235 when a program (like dosemu) has the keyboard in 'raw'
236 mode, the ctrl-alt-del is intercepted by the program before it
237 ever reaches the kernel tty layer, and it's up to the program
238 to decide what to do with it.
244 This toggle indicates whether unprivileged users are prevented
245 from using ``dmesg(8)`` to view messages from the kernel's log
247 When ``dmesg_restrict`` is set to 0 there are no restrictions.
248 When ``dmesg_restrict`` is set to 1, users must have
249 ``CAP_SYSLOG`` to use ``dmesg(8)``.
251 The kernel config option ``CONFIG_SECURITY_DMESG_RESTRICT`` sets the
252 default value of ``dmesg_restrict``.
255 domainname & hostname
256 =====================
258 These files can be used to set the NIS/YP domainname and the
259 hostname of your box in exactly the same way as the commands
260 domainname and hostname, i.e.::
262 # echo "darkstar" > /proc/sys/kernel/hostname
263 # echo "mydomain" > /proc/sys/kernel/domainname
265 has the same effect as::
267 # hostname "darkstar"
268 # domainname "mydomain"
270 Note, however, that the classic darkstar.frop.org has the
271 hostname "darkstar" and DNS (Internet Domain Name Server)
272 domainname "frop.org", not to be confused with the NIS (Network
273 Information Service) or YP (Yellow Pages) domainname. These two
274 domain names are in general different. For a detailed discussion
275 see the ``hostname(1)`` man page.
281 See Documentation/driver-api/firmware/fallback-mechanisms.rst.
283 The entries in this directory allow the firmware loader helper
284 fallback to be controlled:
286 * ``force_sysfs_fallback``, when set to 1, forces the use of the
288 * ``ignore_sysfs_fallback``, when set to 1, ignores any fallback.
294 Determines whether ``ftrace_dump()`` should be called on an oops (or
295 kernel panic). This will output the contents of the ftrace buffers to
296 the console. This is very useful for capturing traces that lead to
297 crashes and outputting them to a serial console.
299 ======================= ===========================================
300 0 Disabled (default).
301 1 Dump buffers of all CPUs.
302 2(orig_cpu) Dump the buffer of the CPU that triggered the
304 <instance> Dump the specific instance buffer on all CPUs.
305 <instance>=2(orig_cpu) Dump the specific instance buffer on the CPU
306 that triggered the oops.
307 ======================= ===========================================
309 Multiple instance dump is also supported, and instances are separated
310 by commas. If global buffer also needs to be dumped, please specify
311 the dump mode (1/2/orig_cpu) first for global buffer.
313 So for example to dump "foo" and "bar" instance buffer on all CPUs,
316 echo "foo,bar" > /proc/sys/kernel/ftrace_dump_on_oops
318 To dump global buffer and "foo" instance buffer on all
319 CPUs along with the "bar" instance buffer on CPU that triggered the
322 echo "1,foo,bar=2" > /proc/sys/kernel/ftrace_dump_on_oops
324 ftrace_enabled, stack_tracer_enabled
325 ====================================
327 See Documentation/trace/ftrace.rst.
330 hardlockup_all_cpu_backtrace
331 ============================
333 This value controls the hard lockup detector behavior when a hard
334 lockup condition is detected as to whether or not to gather further
335 debug information. If enabled, arch-specific all-CPU stack dumping
338 = ============================================
339 0 Do nothing. This is the default behavior.
340 1 On detection capture more debug information.
341 = ============================================
347 This parameter can be used to control whether the kernel panics
348 when a hard lockup is detected.
350 = ===========================
351 0 Don't panic on hard lockup.
352 1 Panic on hard lockup.
353 = ===========================
355 See Documentation/admin-guide/lockup-watchdogs.rst for more information.
356 This can also be set using the nmi_watchdog kernel parameter.
362 Path for the hotplug policy agent.
363 Default value is ``CONFIG_UEVENT_HELPER_PATH``, which in turn defaults
366 This file only exists when ``CONFIG_UEVENT_HELPER`` is enabled. Most
367 modern systems rely exclusively on the netlink-based uevent source and
371 hung_task_all_cpu_backtrace
372 ===========================
374 If this option is set, the kernel will send an NMI to all CPUs to dump
375 their backtraces when a hung task is detected. This file shows up if
376 CONFIG_DETECT_HUNG_TASK and CONFIG_SMP are enabled.
378 0: Won't show all CPUs backtraces when a hung task is detected.
379 This is the default behavior.
381 1: Will non-maskably interrupt all CPUs and dump their backtraces when
382 a hung task is detected.
388 Controls the kernel's behavior when a hung task is detected.
389 This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
391 = =================================================
392 0 Continue operation. This is the default behavior.
394 = =================================================
397 hung_task_check_count
398 =====================
400 The upper bound on the number of tasks that are checked.
401 This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
404 hung_task_timeout_secs
405 ======================
407 When a task in D state did not get scheduled
408 for more than this value report a warning.
409 This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
411 0 means infinite timeout, no checking is done.
413 Possible values to set are in range {0:``LONG_MAX``/``HZ``}.
416 hung_task_check_interval_secs
417 =============================
419 Hung task check interval. If hung task checking is enabled
420 (see `hung_task_timeout_secs`_), the check is done every
421 ``hung_task_check_interval_secs`` seconds.
422 This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
424 0 (default) means use ``hung_task_timeout_secs`` as checking
427 Possible values to set are in range {0:``LONG_MAX``/``HZ``}.
433 The maximum number of warnings to report. During a check interval
434 if a hung task is detected, this value is decreased by 1.
435 When this value reaches 0, no more warnings will be reported.
436 This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
438 -1: report an infinite number of warnings.
441 hyperv_record_panic_msg
442 =======================
444 Controls whether the panic kmsg data should be reported to Hyper-V.
446 = =========================================================
447 0 Do not report panic kmsg data.
448 1 Report the panic kmsg data. This is the default behavior.
449 = =========================================================
452 ignore-unaligned-usertrap
453 =========================
455 On architectures where unaligned accesses cause traps, and where this
456 feature is supported (``CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN``;
457 currently, ``arc`` and ``loongarch``), controls whether all
458 unaligned traps are logged.
460 = =============================================================
461 0 Log all unaligned accesses.
462 1 Only warn the first time a process traps. This is the default
464 = =============================================================
466 See also `unaligned-trap`_.
471 Prevents all processes from creating new io_uring instances. Enabling this
472 shrinks the kernel's attack surface.
474 = ======================================================================
475 0 All processes can create io_uring instances as normal. This is the
477 1 io_uring creation is disabled (io_uring_setup() will fail with
478 -EPERM) for unprivileged processes not in the io_uring_group group.
479 Existing io_uring instances can still be used. See the
480 documentation for io_uring_group for more information.
481 2 io_uring creation is disabled for all processes. io_uring_setup()
482 always fails with -EPERM. Existing io_uring instances can still be
484 = ======================================================================
490 When io_uring_disabled is set to 1, a process must either be
491 privileged (CAP_SYS_ADMIN) or be in the io_uring_group group in order
492 to create an io_uring instance. If io_uring_group is set to -1 (the
493 default), only processes with the CAP_SYS_ADMIN capability may create
500 A toggle indicating if the syscalls ``kexec_load`` and
501 ``kexec_file_load`` have been disabled.
502 This value defaults to 0 (false: ``kexec_*load`` enabled), but can be
503 set to 1 (true: ``kexec_*load`` disabled).
504 Once true, kexec can no longer be used, and the toggle cannot be set
506 This allows a kexec image to be loaded before disabling the syscall,
507 allowing a system to set up (and later use) an image without it being
509 Generally used together with the `modules_disabled`_ sysctl.
511 kexec_load_limit_panic
512 ======================
514 This parameter specifies a limit to the number of times the syscalls
515 ``kexec_load`` and ``kexec_file_load`` can be called with a crash
516 image. It can only be set with a more restrictive value than the
519 == ======================================================
520 -1 Unlimited calls to kexec. This is the default setting.
521 N Number of calls left.
522 == ======================================================
524 kexec_load_limit_reboot
525 =======================
527 Similar functionality as ``kexec_load_limit_panic``, but for a normal
533 This toggle indicates whether restrictions are placed on
534 exposing kernel addresses via ``/proc`` and other interfaces.
536 When ``kptr_restrict`` is set to 0 (the default) the address is hashed
538 (This is the equivalent to %p.)
540 When ``kptr_restrict`` is set to 1, kernel pointers printed using the
541 %pK format specifier will be replaced with 0s unless the user has
542 ``CAP_SYSLOG`` and effective user and group ids are equal to the real
544 This is because %pK checks are done at read() time rather than open()
545 time, so if permissions are elevated between the open() and the read()
546 (e.g via a setuid binary) then %pK will not leak kernel pointers to
548 Note, this is a temporary solution only.
549 The correct long-term solution is to do the permission checks at
551 Consider removing world read permissions from files that use %pK, and
552 using `dmesg_restrict`_ to protect against uses of %pK in ``dmesg(8)``
553 if leaking kernel pointer values to unprivileged users is a concern.
555 When ``kptr_restrict`` is set to 2, kernel pointers printed using
556 %pK will be replaced with 0s regardless of privileges.
562 The full path to the usermode helper for autoloading kernel modules,
563 by default ``CONFIG_MODPROBE_PATH``, which in turn defaults to
564 "/sbin/modprobe". This binary is executed when the kernel requests a
565 module. For example, if userspace passes an unknown filesystem type
566 to mount(), then the kernel will automatically request the
567 corresponding filesystem module by executing this usermode helper.
568 This usermode helper should insert the needed module into the kernel.
570 This sysctl only affects module autoloading. It has no effect on the
571 ability to explicitly insert modules.
573 This sysctl can be used to debug module loading requests::
575 echo '#! /bin/sh' > /tmp/modprobe
576 echo 'echo "$@" >> /tmp/modprobe.log' >> /tmp/modprobe
577 echo 'exec /sbin/modprobe "$@"' >> /tmp/modprobe
578 chmod a+x /tmp/modprobe
579 echo /tmp/modprobe > /proc/sys/kernel/modprobe
581 Alternatively, if this sysctl is set to the empty string, then module
582 autoloading is completely disabled. The kernel will not try to
583 execute a usermode helper at all, nor will it call the
584 kernel_module_request LSM hook.
586 If CONFIG_STATIC_USERMODEHELPER=y is set in the kernel configuration,
587 then the configured static usermode helper overrides this sysctl,
588 except that the empty string is still accepted to completely disable
589 module autoloading as described above.
594 A toggle value indicating if modules are allowed to be loaded
595 in an otherwise modular kernel. This toggle defaults to off
596 (0), but can be set true (1). Once true, modules can be
597 neither loaded nor unloaded, and the toggle cannot be set back
598 to false. Generally used with the `kexec_load_disabled`_ toggle.
603 msgmax, msgmnb, and msgmni
604 ==========================
606 ``msgmax`` is the maximum size of an IPC message, in bytes. 8192 by
607 default (``MSGMAX``).
609 ``msgmnb`` is the maximum size of an IPC queue, in bytes. 16384 by
610 default (``MSGMNB``).
612 ``msgmni`` is the maximum number of IPC queues. 32000 by default
615 All of these parameters are set per ipc namespace. The maximum number of bytes
616 in POSIX message queues is limited by ``RLIMIT_MSGQUEUE``. This limit is
617 respected hierarchically in the each user namespace.
619 msg_next_id, sem_next_id, and shm_next_id (System V IPC)
620 ========================================================
622 These three toggles allows to specify desired id for next allocated IPC
623 object: message, semaphore or shared memory respectively.
625 By default they are equal to -1, which means generic allocation logic.
626 Possible values to set are in range {0:``INT_MAX``}.
629 1) kernel doesn't guarantee, that new object will have desired id. So,
630 it's up to userspace, how to handle an object with "wrong" id.
631 2) Toggle with non-default value will be set back to -1 by kernel after
632 successful IPC object allocation. If an IPC object allocation syscall
633 fails, it is undefined if the value remains unmodified or is reset to -1.
639 Maximum number of supplementary groups, _i.e._ the maximum size which
640 ``setgroups`` will accept. Exports ``NGROUPS_MAX`` from the kernel.
647 This parameter can be used to control the NMI watchdog
648 (i.e. the hard lockup detector) on x86 systems.
650 = =================================
651 0 Disable the hard lockup detector.
652 1 Enable the hard lockup detector.
653 = =================================
655 The hard lockup detector monitors each CPU for its ability to respond to
656 timer interrupts. The mechanism utilizes CPU performance counter registers
657 that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
658 while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
660 The NMI watchdog is disabled by default if the kernel is running as a guest
661 in a KVM virtual machine. This default can be overridden by adding::
665 to the guest kernel command line (see
666 Documentation/admin-guide/kernel-parameters.rst).
669 nmi_wd_lpm_factor (PPC only)
670 ============================
672 Factor to apply to the NMI watchdog timeout (only when ``nmi_watchdog`` is
673 set to 1). This factor represents the percentage added to
674 ``watchdog_thresh`` when calculating the NMI watchdog timeout during an
675 LPM. The soft lockup timeout is not impacted.
677 A value of 0 means no change. The default value is 200 meaning the NMI
678 watchdog is set to 30s (based on ``watchdog_thresh`` equal to 10).
684 Enables/disables and configures automatic page fault based NUMA memory
685 balancing. Memory is moved automatically to nodes that access it often.
686 The value to set can be the result of ORing the following:
688 = =================================
689 0 NUMA_BALANCING_DISABLED
690 1 NUMA_BALANCING_NORMAL
691 2 NUMA_BALANCING_MEMORY_TIERING
692 = =================================
694 Or NUMA_BALANCING_NORMAL to optimize page placement among different
695 NUMA nodes to reduce remote accessing. On NUMA machines, there is a
696 performance penalty if remote memory is accessed by a CPU. When this
697 feature is enabled the kernel samples what task thread is accessing
698 memory by periodically unmapping pages and later trapping a page
699 fault. At the time of the page fault, it is determined if the data
700 being accessed should be migrated to a local memory node.
702 The unmapping of pages and trapping faults incur additional overhead that
703 ideally is offset by improved memory locality but there is no universal
704 guarantee. If the target workload is already bound to NUMA nodes then this
705 feature should be disabled.
707 Or NUMA_BALANCING_MEMORY_TIERING to optimize page placement among
708 different types of memory (represented as different NUMA nodes) to
709 place the hot pages in the fast memory. This is implemented based on
710 unmapping and page fault too.
712 numa_balancing_promote_rate_limit_MBps
713 ======================================
715 Too high promotion/demotion throughput between different memory types
716 may hurt application latency. This can be used to rate limit the
717 promotion throughput. The per-node max promotion throughput in MB/s
718 will be limited to be no more than the set value.
720 A rule of thumb is to set this to less than 1/10 of the PMEM node
723 oops_all_cpu_backtrace
724 ======================
726 If this option is set, the kernel will send an NMI to all CPUs to dump
727 their backtraces when an oops event occurs. It should be used as a last
728 resort in case a panic cannot be triggered (to protect VMs running, for
729 example) or kdump can't be collected. This file shows up if CONFIG_SMP
732 0: Won't show all CPUs backtraces when an oops is detected.
733 This is the default behavior.
735 1: Will non-maskably interrupt all CPUs and dump their backtraces when
736 an oops event is detected.
742 Number of kernel oopses after which the kernel should panic when
743 ``panic_on_oops`` is not set. Setting this to 0 disables checking
744 the count. Setting this to 1 has the same effect as setting
745 ``panic_on_oops=1``. The default value is 10000.
748 osrelease, ostype & version
749 ===========================
758 #5 Wed Feb 25 21:49:24 MET 1998
760 The files ``osrelease`` and ``ostype`` should be clear enough.
762 needs a little more clarification however. The '#5' means that
763 this is the fifth kernel built from this source base and the
764 date behind it indicates the time the kernel was built.
765 The only way to tune these values is to rebuild the kernel :-)
768 overflowgid & overflowuid
769 =========================
771 if your architecture did not always support 32-bit UIDs (i.e. arm,
772 i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
773 applications that use the old 16-bit UID/GID system calls, if the
774 actual UID or GID would exceed 65535.
776 These sysctls allow you to change the value of the fixed UID and GID.
777 The default is 65534.
783 The value in this file determines the behaviour of the kernel on a
786 * if zero, the kernel will loop forever;
787 * if negative, the kernel will reboot immediately;
788 * if positive, the kernel will reboot after the corresponding number
791 When you use the software watchdog, the recommended setting is 60.
797 Controls the kernel's behavior when a CPU receives an NMI caused by
800 = ==================================================================
801 0 Try to continue operation (default).
802 1 Panic immediately. The IO error triggered an NMI. This indicates a
803 serious system condition which could result in IO data corruption.
804 Rather than continuing, panicking might be a better choice. Some
805 servers issue this sort of NMI when the dump button is pushed,
806 and you can use this option to take a crash dump.
807 = ==================================================================
813 Controls the kernel's behaviour when an oops or BUG is encountered.
815 = ===================================================================
816 0 Try to continue operation.
817 1 Panic immediately. If the `panic` sysctl is also non-zero then the
818 machine will be rebooted.
819 = ===================================================================
822 panic_on_stackoverflow
823 ======================
825 Controls the kernel's behavior when detecting the overflows of
826 kernel, IRQ and exception stacks except a user stack.
827 This file shows up if ``CONFIG_DEBUG_STACKOVERFLOW`` is enabled.
829 = ==========================
830 0 Try to continue operation.
832 = ==========================
835 panic_on_unrecovered_nmi
836 ========================
838 The default Linux behaviour on an NMI of either memory or unknown is
839 to continue operation. For many environments such as scientific
840 computing it is preferable that the box is taken out and the error
841 dealt with than an uncorrected parity/ECC error get propagated.
843 A small number of systems do generate NMIs for bizarre random reasons
844 such as power management so the default is off. That sysctl works like
845 the existing panic controls already in that directory.
851 Calls panic() in the WARN() path when set to 1. This is useful to avoid
852 a kernel rebuild when attempting to kdump at the location of a WARN().
854 = ================================================
855 0 Only WARN(), default behaviour.
856 1 Call panic() after printing out WARN() location.
857 = ================================================
863 Bitmask for printing system info when panic happens. User can chose
864 combination of the following bits:
866 ===== ============================================
867 bit 0 print all tasks info
868 bit 1 print system memory info
869 bit 2 print timer info
870 bit 3 print locks info if ``CONFIG_LOCKDEP`` is on
871 bit 4 print ftrace buffer
872 bit 5 print all printk messages in buffer
873 bit 6 print all CPUs backtrace (if available in the arch)
874 bit 7 print only tasks in uninterruptible (blocked) state
875 ===== ============================================
877 So for example to print tasks and memory info on panic, user can::
879 echo 3 > /proc/sys/kernel/panic_print
885 When set to 1, calls panic() after RCU stall detection messages. This
886 is useful to define the root cause of RCU stalls using a vmcore.
888 = ============================================================
889 0 Do not panic() when RCU stall takes place, default behavior.
890 1 panic() after printing RCU stall messages.
891 = ============================================================
893 max_rcu_stall_to_panic
894 ======================
896 When ``panic_on_rcu_stall`` is set to 1, this value determines the
897 number of times that RCU can stall before panic() is called.
899 When ``panic_on_rcu_stall`` is set to 0, this value is has no effect.
901 perf_cpu_time_max_percent
902 =========================
904 Hints to the kernel how much CPU time it should be allowed to
905 use to handle perf sampling events. If the perf subsystem
906 is informed that its samples are exceeding this limit, it
907 will drop its sampling frequency to attempt to reduce its CPU
910 Some perf sampling happens in NMIs. If these samples
911 unexpectedly take too long to execute, the NMIs can become
912 stacked up next to each other so much that nothing else is
915 ===== ========================================================
916 0 Disable the mechanism. Do not monitor or correct perf's
917 sampling rate no matter how CPU time it takes.
919 1-100 Attempt to throttle perf's sample rate to this
920 percentage of CPU. Note: the kernel calculates an
921 "expected" length of each sample event. 100 here means
922 100% of that expected length. Even if this is set to
923 100, you may still see sample throttling if this
924 length is exceeded. Set to 0 if you truly do not care
925 how much CPU is consumed.
926 ===== ========================================================
932 Controls use of the performance events system by unprivileged
933 users (without CAP_PERFMON). The default value is 2.
935 For backward compatibility reasons access to system performance
936 monitoring and observability remains open for CAP_SYS_ADMIN
937 privileged processes but CAP_SYS_ADMIN usage for secure system
938 performance monitoring and observability operations is discouraged
939 with respect to CAP_PERFMON use cases.
941 === ==================================================================
942 -1 Allow use of (almost) all events by all users.
944 Ignore mlock limit after perf_event_mlock_kb without
947 >=0 Disallow ftrace function tracepoint by users without
950 Disallow raw tracepoint access by users without ``CAP_PERFMON``.
952 >=1 Disallow CPU event access by users without ``CAP_PERFMON``.
954 >=2 Disallow kernel profiling by users without ``CAP_PERFMON``.
955 === ==================================================================
961 Controls maximum number of stack frames to copy for (``attr.sample_type &
962 PERF_SAMPLE_CALLCHAIN``) configured events, for instance, when using
963 '``perf record -g``' or '``perf trace --call-graph fp``'.
965 This can only be done when no events are in use that have callchains
966 enabled, otherwise writing to this file will return ``-EBUSY``.
968 The default value is 127.
974 Control size of per-cpu ring buffer not counted against mlock limit.
976 The default value is 512 + 1 page
979 perf_event_max_contexts_per_stack
980 =================================
982 Controls maximum number of stack frame context entries for
983 (``attr.sample_type & PERF_SAMPLE_CALLCHAIN``) configured events, for
984 instance, when using '``perf record -g``' or '``perf trace --call-graph fp``'.
986 This can only be done when no events are in use that have callchains
987 enabled, otherwise writing to this file will return ``-EBUSY``.
989 The default value is 8.
992 perf_user_access (arm64 and riscv only)
993 =======================================
995 Controls user space access for reading perf event counters.
1000 The default value is 0 (access disabled).
1002 When set to 1, user space can read performance monitor counter registers
1005 See Documentation/arch/arm64/perf.rst for more information.
1010 When set to 0, user space access is disabled.
1012 The default value is 1, user space can read performance monitor counter
1013 registers through perf, any direct access without perf intervention will trigger
1014 an illegal instruction.
1016 When set to 2, which enables legacy mode (user space has direct access to cycle
1017 and insret CSRs only). Note that this legacy value is deprecated and will be
1018 removed once all user space applications are fixed.
1020 Note that the time CSR is always directly accessible to all modes.
1025 PID allocation wrap value. When the kernel's next PID value
1026 reaches this value, it wraps back to a minimum PID value.
1027 PIDs of value ``pid_max`` or larger are not allocated.
1033 The last pid allocated in the current (the one task using this sysctl
1034 lives in) pid namespace. When selecting a pid for a next task on fork
1035 kernel tries to allocate a number starting from this one.
1038 powersave-nap (PPC only)
1039 ========================
1041 If set, Linux-PPC will use the 'nap' mode of powersaving,
1042 otherwise the 'doze' mode will be used.
1045 ==============================================================
1050 The four values in printk denote: ``console_loglevel``,
1051 ``default_message_loglevel``, ``minimum_console_loglevel`` and
1052 ``default_console_loglevel`` respectively.
1054 These values influence printk() behavior when printing or
1055 logging error messages. See '``man 2 syslog``' for more info on
1056 the different loglevels.
1058 ======================== =====================================
1059 console_loglevel messages with a higher priority than
1060 this will be printed to the console
1061 default_message_loglevel messages without an explicit priority
1062 will be printed with this priority
1063 minimum_console_loglevel minimum (highest) value to which
1064 console_loglevel can be set
1065 default_console_loglevel default value for console_loglevel
1066 ======================== =====================================
1072 Delay each printk message in ``printk_delay`` milliseconds
1074 Value from 0 - 10000 is allowed.
1080 Some warning messages are rate limited. ``printk_ratelimit`` specifies
1081 the minimum length of time between these messages (in seconds).
1082 The default value is 5 seconds.
1084 A value of 0 will disable rate limiting.
1087 printk_ratelimit_burst
1088 ======================
1090 While long term we enforce one message per `printk_ratelimit`_
1091 seconds, we do allow a burst of messages to pass through.
1092 ``printk_ratelimit_burst`` specifies the number of messages we can
1093 send before ratelimiting kicks in.
1095 The default value is 10 messages.
1101 Control the logging to ``/dev/kmsg`` from userspace:
1103 ========= =============================================
1104 ratelimit default, ratelimited
1105 on unlimited logging to /dev/kmsg from userspace
1106 off logging to /dev/kmsg disabled
1107 ========= =============================================
1109 The kernel command line parameter ``printk.devkmsg=`` overrides this and is
1110 a one-time setting until next reboot: once set, it cannot be changed by
1111 this sysctl interface anymore.
1113 ==============================================================
1119 See Documentation/filesystems/devpts.rst.
1125 This is a directory, with the following entries:
1127 * ``boot_id``: a UUID generated the first time this is retrieved, and
1128 unvarying after that;
1130 * ``uuid``: a UUID generated every time this is retrieved (this can
1131 thus be used to generate UUIDs at will);
1133 * ``entropy_avail``: the pool's entropy count, in bits;
1135 * ``poolsize``: the entropy pool size, in bits;
1137 * ``urandom_min_reseed_secs``: obsolete (used to determine the minimum
1138 number of seconds between urandom pool reseeding). This file is
1139 writable for compatibility purposes, but writing to it has no effect
1140 on any RNG behavior;
1142 * ``write_wakeup_threshold``: when the entropy count drops below this
1143 (as a number of bits), processes waiting to write to ``/dev/random``
1144 are woken up. This file is writable for compatibility purposes, but
1145 writing to it has no effect on any RNG behavior.
1151 This option can be used to select the type of process address
1152 space randomization that is used in the system, for architectures
1153 that support this feature.
1155 == ===========================================================================
1156 0 Turn the process address space randomization off. This is the
1157 default for architectures that do not support this feature anyways,
1158 and kernels that are booted with the "norandmaps" parameter.
1160 1 Make the addresses of mmap base, stack and VDSO page randomized.
1161 This, among other things, implies that shared libraries will be
1162 loaded to random addresses. Also for PIE-linked binaries, the
1163 location of code start is randomized. This is the default if the
1164 ``CONFIG_COMPAT_BRK`` option is enabled.
1166 2 Additionally enable heap randomization. This is the default if
1167 ``CONFIG_COMPAT_BRK`` is disabled.
1169 There are a few legacy applications out there (such as some ancient
1170 versions of libc.so.5 from 1996) that assume that brk area starts
1171 just after the end of the code+bss. These applications break when
1172 start of the brk area is randomized. There are however no known
1173 non-legacy applications that would be broken this way, so for most
1174 systems it is safe to choose full randomization.
1176 Systems with ancient and/or broken binaries should be configured
1177 with ``CONFIG_COMPAT_BRK`` enabled, which excludes the heap from process
1178 address space randomization.
1179 == ===========================================================================
1185 See Documentation/admin-guide/initrd.rst.
1188 reboot-cmd (SPARC only)
1189 =======================
1191 ??? This seems to be a way to give an argument to the Sparc
1192 ROM/Flash boot loader. Maybe to tell it what to do after
1199 Enables/disables Energy Aware Scheduling (EAS). EAS starts
1200 automatically on platforms where it can run (that is,
1201 platforms with asymmetric CPU topologies and having an Energy
1202 Model available). If your platform happens to meet the
1203 requirements for EAS but you do not want to use it, change
1204 this value to 0. On Non-EAS platforms, write operation fails and
1205 read doesn't return anything.
1210 Enables/disables task delay accounting (see
1211 Documentation/accounting/delay-accounting.rst. Enabling this feature incurs
1212 a small amount of overhead in the scheduler but is useful for debugging
1213 and performance tuning. It is required by some tools such as iotop.
1218 Enables/disables scheduler statistics. Enabling this feature
1219 incurs a small amount of overhead in the scheduler but is
1220 useful for debugging and performance tuning.
1222 sched_util_clamp_min
1223 ====================
1225 Max allowed *minimum* utilization.
1227 Default value is 1024, which is the maximum possible value.
1229 It means that any requested uclamp.min value cannot be greater than
1230 sched_util_clamp_min, i.e., it is restricted to the range
1231 [0:sched_util_clamp_min].
1233 sched_util_clamp_max
1234 ====================
1236 Max allowed *maximum* utilization.
1238 Default value is 1024, which is the maximum possible value.
1240 It means that any requested uclamp.max value cannot be greater than
1241 sched_util_clamp_max, i.e., it is restricted to the range
1242 [0:sched_util_clamp_max].
1244 sched_util_clamp_min_rt_default
1245 ===============================
1247 By default Linux is tuned for performance. Which means that RT tasks always run
1248 at the highest frequency and most capable (highest capacity) CPU (in
1249 heterogeneous systems).
1251 Uclamp achieves this by setting the requested uclamp.min of all RT tasks to
1252 1024 by default, which effectively boosts the tasks to run at the highest
1253 frequency and biases them to run on the biggest CPU.
1255 This knob allows admins to change the default behavior when uclamp is being
1256 used. In battery powered devices particularly, running at the maximum
1257 capacity and frequency will increase energy consumption and shorten the battery
1260 This knob is only effective for RT tasks which the user hasn't modified their
1261 requested uclamp.min value via sched_setattr() syscall.
1263 This knob will not escape the range constraint imposed by sched_util_clamp_min
1268 sched_util_clamp_min_rt_default = 800
1269 sched_util_clamp_min = 600
1271 Then the boost will be clamped to 600 because 800 is outside of the permissible
1272 range of [0:600]. This could happen for instance if a powersave mode will
1273 restrict all boosts temporarily by modifying sched_util_clamp_min. As soon as
1274 this restriction is lifted, the requested sched_util_clamp_min_rt_default
1280 See Documentation/userspace-api/seccomp_filter.rst.
1286 This file shows the size of the generic SCSI (sg) buffer.
1287 You can't tune it just yet, but you could change it on
1288 compile time by editing ``include/scsi/sg.h`` and changing
1289 the value of ``SG_BIG_BUFF``.
1291 There shouldn't be any reason to change this value. If
1292 you can come up with one, you probably know what you
1299 This parameter sets the total amount of shared memory pages that can be used
1300 inside ipc namespace. The shared memory pages counting occurs for each ipc
1301 namespace separately and is not inherited. Hence, ``shmall`` should always be at
1302 least ``ceil(shmmax/PAGE_SIZE)``.
1304 If you are not sure what the default ``PAGE_SIZE`` is on your Linux
1305 system, you can run the following command::
1309 To reduce or disable the ability to allocate shared memory, you must create a
1310 new ipc namespace, set this parameter to the required value and prohibit the
1311 creation of a new ipc namespace in the current user namespace or cgroups can
1317 This value can be used to query and set the run time limit
1318 on the maximum shared memory segment size that can be created.
1319 Shared memory segments up to 1Gb are now supported in the
1320 kernel. This value defaults to ``SHMMAX``.
1326 This value determines the maximum number of shared memory segments.
1327 4096 by default (``SHMMNI``).
1333 Linux lets you set resource limits, including how much memory one
1334 process can consume, via ``setrlimit(2)``. Unfortunately, shared memory
1335 segments are allowed to exist without association with any process, and
1336 thus might not be counted against any resource limits. If enabled,
1337 shared memory segments are automatically destroyed when their attach
1338 count becomes zero after a detach or a process termination. It will
1339 also destroy segments that were created, but never attached to, on exit
1340 from the process. The only use left for ``IPC_RMID`` is to immediately
1341 destroy an unattached segment. Of course, this breaks the way things are
1342 defined, so some applications might stop working. Note that this
1343 feature will do you no good unless you also configure your resource
1344 limits (in particular, ``RLIMIT_AS`` and ``RLIMIT_NPROC``). Most systems don't
1347 Note that if you change this from 0 to 1, already created segments
1348 without users and with a dead originative process will be destroyed.
1351 sysctl_writes_strict
1352 ====================
1354 Control how file position affects the behavior of updating sysctl values
1355 via the ``/proc/sys`` interface:
1357 == ======================================================================
1358 -1 Legacy per-write sysctl value handling, with no printk warnings.
1359 Each write syscall must fully contain the sysctl value to be
1360 written, and multiple writes on the same sysctl file descriptor
1361 will rewrite the sysctl value, regardless of file position.
1362 0 Same behavior as above, but warn about processes that perform writes
1363 to a sysctl file descriptor when the file position is not 0.
1364 1 (default) Respect file position when writing sysctl strings. Multiple
1365 writes will append to the sysctl value buffer. Anything past the max
1366 length of the sysctl value buffer will be ignored. Writes to numeric
1367 sysctl entries must always be at file position 0 and the value must
1368 be fully contained in the buffer sent in the write syscall.
1369 == ======================================================================
1372 softlockup_all_cpu_backtrace
1373 ============================
1375 This value controls the soft lockup detector thread's behavior
1376 when a soft lockup condition is detected as to whether or not
1377 to gather further debug information. If enabled, each cpu will
1378 be issued an NMI and instructed to capture stack trace.
1380 This feature is only applicable for architectures which support
1383 = ============================================
1384 0 Do nothing. This is the default behavior.
1385 1 On detection capture more debug information.
1386 = ============================================
1392 This parameter can be used to control whether the kernel panics
1393 when a soft lockup is detected.
1395 = ============================================
1396 0 Don't panic on soft lockup.
1397 1 Panic on soft lockup.
1398 = ============================================
1400 This can also be set using the softlockup_panic kernel parameter.
1406 This parameter can be used to control the soft lockup detector.
1408 = =================================
1409 0 Disable the soft lockup detector.
1410 1 Enable the soft lockup detector.
1411 = =================================
1413 The soft lockup detector monitors CPUs for threads that are hogging the CPUs
1414 without rescheduling voluntarily, and thus prevent the 'migration/N' threads
1415 from running, causing the watchdog work fail to execute. The mechanism depends
1416 on the CPUs ability to respond to timer interrupts which are needed for the
1417 watchdog work to be queued by the watchdog timer function, otherwise the NMI
1418 watchdog — if enabled — can detect a hard lockup condition.
1421 split_lock_mitigate (x86 only)
1422 ==============================
1424 On x86, each "split lock" imposes a system-wide performance penalty. On larger
1425 systems, large numbers of split locks from unprivileged users can result in
1426 denials of service to well-behaved and potentially more important users.
1428 The kernel mitigates these bad users by detecting split locks and imposing
1429 penalties: forcing them to wait and only allowing one core to execute split
1432 These mitigations can make those bad applications unbearably slow. Setting
1433 split_lock_mitigate=0 may restore some application performance, but will also
1434 increase system exposure to denial of service attacks from split lock users.
1436 = ===================================================================
1437 0 Disable the mitigation mode - just warns the split lock on kernel log
1438 and exposes the system to denials of service from the split lockers.
1439 1 Enable the mitigation mode (this is the default) - penalizes the split
1440 lockers with intentional performance degradation.
1441 = ===================================================================
1447 This parameter can be used to control kernel stack erasing at the end
1448 of syscalls for kernels built with ``CONFIG_GCC_PLUGIN_STACKLEAK``.
1450 That erasing reduces the information which kernel stack leak bugs
1451 can reveal and blocks some uninitialized stack variable attacks.
1452 The tradeoff is the performance impact: on a single CPU system kernel
1453 compilation sees a 1% slowdown, other systems and workloads may vary.
1455 = ====================================================================
1456 0 Kernel stack erasing is disabled, STACKLEAK_METRICS are not updated.
1457 1 Kernel stack erasing is enabled (default), it is performed before
1458 returning to the userspace at the end of syscalls.
1459 = ====================================================================
1467 = ====================================
1468 0 Stop-A has no effect.
1469 1 Stop-A breaks to the PROM (default).
1470 = ====================================
1472 Stop-A is always enabled on a panic, so that the user can return to
1479 See Documentation/admin-guide/sysrq.rst.
1485 Non-zero if the kernel has been tainted. Numeric values, which can be
1486 ORed together. The letters are seen in "Tainted" line of Oops reports.
1488 ====== ===== ==============================================================
1489 1 `(P)` proprietary module was loaded
1490 2 `(F)` module was force loaded
1491 4 `(S)` kernel running on an out of specification system
1492 8 `(R)` module was force unloaded
1493 16 `(M)` processor reported a Machine Check Exception (MCE)
1494 32 `(B)` bad page referenced or some unexpected page flags
1495 64 `(U)` taint requested by userspace application
1496 128 `(D)` kernel died recently, i.e. there was an OOPS or BUG
1497 256 `(A)` an ACPI table was overridden by user
1498 512 `(W)` kernel issued warning
1499 1024 `(C)` staging driver was loaded
1500 2048 `(I)` workaround for bug in platform firmware applied
1501 4096 `(O)` externally-built ("out-of-tree") module was loaded
1502 8192 `(E)` unsigned module was loaded
1503 16384 `(L)` soft lockup occurred
1504 32768 `(K)` kernel has been live patched
1505 65536 `(X)` Auxiliary taint, defined and used by for distros
1506 131072 `(T)` The kernel was built with the struct randomization plugin
1507 ====== ===== ==============================================================
1509 See Documentation/admin-guide/tainted-kernels.rst for more information.
1512 writes to this sysctl interface will fail with ``EINVAL`` if the kernel is
1513 booted with the command line option ``panic_on_taint=<bitmask>,nousertaint``
1514 and any of the ORed together values being written to ``tainted`` match with
1515 the bitmask declared on panic_on_taint.
1516 See Documentation/admin-guide/kernel-parameters.rst for more details on
1517 that particular kernel command line option and its optional
1518 ``nousertaint`` switch.
1523 This value controls the maximum number of threads that can be created
1526 During initialization the kernel sets this value such that even if the
1527 maximum number of threads is created, the thread structures occupy only
1528 a part (1/8th) of the available RAM pages.
1530 The minimum value that can be written to ``threads-max`` is 1.
1532 The maximum value that can be written to ``threads-max`` is given by the
1533 constant ``FUTEX_TID_MASK`` (0x3fffffff).
1535 If a value outside of this range is written to ``threads-max`` an
1536 ``EINVAL`` error occurs.
1542 When set, disables tracing (see Documentation/trace/ftrace.rst) when a
1549 When tracepoints are sent to printk() (enabled by the ``tp_printk``
1550 boot parameter), this entry provides runtime control::
1552 echo 0 > /proc/sys/kernel/tracepoint_printk
1554 will stop tracepoints from being sent to printk(), and::
1556 echo 1 > /proc/sys/kernel/tracepoint_printk
1558 will send them to printk() again.
1560 This only works if the kernel was booted with ``tp_printk`` enabled.
1562 See Documentation/admin-guide/kernel-parameters.rst and
1563 Documentation/trace/boottime-trace.rst.
1569 On architectures where unaligned accesses cause traps, and where this
1570 feature is supported (``CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW``; currently,
1571 ``arc``, ``parisc`` and ``loongarch``), controls whether unaligned traps
1572 are caught and emulated (instead of failing).
1574 = ========================================================
1575 0 Do not emulate unaligned accesses.
1576 1 Emulate unaligned accesses. This is the default setting.
1577 = ========================================================
1579 See also `ignore-unaligned-usertrap`_.
1585 The value in this file affects behavior of handling NMI. When the
1586 value is non-zero, unknown NMI is trapped and then panic occurs. At
1587 that time, kernel debugging information is displayed on console.
1589 NMI switch that most IA32 servers have fires unknown NMI up, for
1590 example. If a system hangs up, try pressing the NMI switch.
1593 unprivileged_bpf_disabled
1594 =========================
1596 Writing 1 to this entry will disable unprivileged calls to ``bpf()``;
1597 once disabled, calling ``bpf()`` without ``CAP_SYS_ADMIN`` or ``CAP_BPF``
1598 will return ``-EPERM``. Once set to 1, this can't be cleared from the
1599 running kernel anymore.
1601 Writing 2 to this entry will also disable unprivileged calls to ``bpf()``,
1602 however, an admin can still change this setting later on, if needed, by
1603 writing 0 or 1 to this entry.
1605 If ``BPF_UNPRIV_DEFAULT_OFF`` is enabled in the kernel config, then this
1606 entry will default to 2 instead of 0.
1608 = =============================================================
1609 0 Unprivileged calls to ``bpf()`` are enabled
1610 1 Unprivileged calls to ``bpf()`` are disabled without recovery
1611 2 Unprivileged calls to ``bpf()`` are disabled
1612 = =============================================================
1618 Number of kernel warnings after which the kernel should panic when
1619 ``panic_on_warn`` is not set. Setting this to 0 disables checking
1620 the warning count. Setting this to 1 has the same effect as setting
1621 ``panic_on_warn=1``. The default value is 0.
1627 This parameter can be used to disable or enable the soft lockup detector
1628 *and* the NMI watchdog (i.e. the hard lockup detector) at the same time.
1630 = ==============================
1631 0 Disable both lockup detectors.
1632 1 Enable both lockup detectors.
1633 = ==============================
1635 The soft lockup detector and the NMI watchdog can also be disabled or
1636 enabled individually, using the ``soft_watchdog`` and ``nmi_watchdog``
1638 If the ``watchdog`` parameter is read, for example by executing::
1640 cat /proc/sys/kernel/watchdog
1642 the output of this command (0 or 1) shows the logical OR of
1643 ``soft_watchdog`` and ``nmi_watchdog``.
1649 This value can be used to control on which cpus the watchdog may run.
1650 The default cpumask is all possible cores, but if ``NO_HZ_FULL`` is
1651 enabled in the kernel config, and cores are specified with the
1652 ``nohz_full=`` boot argument, those cores are excluded by default.
1653 Offline cores can be included in this mask, and if the core is later
1654 brought online, the watchdog will be started based on the mask value.
1656 Typically this value would only be touched in the ``nohz_full`` case
1657 to re-enable cores that by default were not running the watchdog,
1658 if a kernel lockup was suspected on those cores.
1660 The argument value is the standard cpulist format for cpumasks,
1661 so for example to enable the watchdog on cores 0, 2, 3, and 4 you
1664 echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
1670 This value can be used to control the frequency of hrtimer and NMI
1671 events and the soft and hard lockup thresholds. The default threshold
1674 The softlockup threshold is (``2 * watchdog_thresh``). Setting this
1675 tunable to zero will disable lockup detection altogether.