admin-guide/sysctl/kernel.rst

5 .. See scripts/check-sysctl-docs to keep this up to date
13 Documentation/admin-guide/sysctl/index.rst.
15 ------------------------------------------------------------------------------
39 If BSD-style process accounting is enabled these values control
71 The machine hardware name, the same output as ``uname -m``
77 This variable has no effect and may be removed in future kernel
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`_.
170 	%d		dump mode, matches ``PR_SET_DUMPABLE`` and
173 	%t		UNIX time of dump
184   the rest of the pattern as a command to run.  The core dump will be
209 captured in parallel, but that no waiting will take place (i.e. the
225 ctrl-alt-del
228 When the value in this file is 0, ctrl-alt-del is trapped and
236   mode, the ctrl-alt-del is intercepted by the program before it
247 When ``dmesg_restrict`` is set to 0 there are no restrictions.
281 See Documentation/driver-api/firmware/fallback-mechanisms.rst.
294 Determines whether ``ftrace_dump()`` should be called on an oops (or
301 1 Dump buffers of all CPUs.
302 2 Dump the buffer of the CPU that triggered the oops.
317 debug information. If enabled, arch-specific all-CPU stack dumping
337 See Documentation/admin-guide/lockup-watchdogs.rst for more information.
349 modern systems rely exclusively on the netlink-based uevent source and
356 If this option is set, the kernel will send an NMI to all CPUs to dump
363 1: Will non-maskably interrupt all CPUs and dump their backtraces when
393 0 means infinite timeout, no checking is done.
417 When this value reaches 0, no more warnings will be reported.
420 -1: report an infinite number of warnings.
426 Controls whether the panic kmsg data should be reported to Hyper-V.
434 ignore-unaligned-usertrap
448 See also `unaligned-trap`_.
460   -EPERM) for unprivileged processes not in the io_uring_group group.
464   always fails with -EPERM. Existing io_uring instances can still be
474 to create an io_uring instance.  If io_uring_group is set to -1 (the
486 Once true, kexec can no longer be used, and the toggle cannot be set
502 -1 Unlimited calls to kexec. This is the default setting.
531 The correct long-term solution is to do the permission checks at
552 This sysctl only affects module autoloading.  It has no effect on the
604 By default they are equal to -1, which means generic allocation logic.
610   2) Toggle with non-default value will be set back to -1 by kernel after
612      fails, it is undefined if the value remains unmodified or is reset to -1.
636 that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
645 Documentation/admin-guide/kernel-parameters.rst).
656 A value of 0 means no change. The default value is 200 meaning the NMI
682 ideally is offset by improved memory locality but there is no universal
696 promotion throughput.  The per-node max promotion throughput in MB/s
697 will be limited to be no more than the set value.
705 If this option is set, the kernel will send an NMI to all CPUs to dump
706 their backtraces when an oops event occurs. It should be used as a last
711 0: Won't show all CPUs backtraces when an oops is detected.
714 1: Will non-maskably interrupt all CPUs and dump their backtraces when
715 an oops event is detected.
744 The only way to tune these values is to rebuild the kernel :-)
750 if your architecture did not always support 32-bit UIDs (i.e. arm,
752 applications that use the old 16-bit UID/GID system calls, if the
784   servers issue this sort of NMI when the dump button is pushed,
785   and you can use this option to take a crash dump.
792 Controls the kernel's behaviour when an oops or BUG is encountered.
796 1 Panic immediately.  If the `panic` sysctl is also non-zero then the
877 When ``panic_on_rcu_stall`` is set to 0, this value is has no effect.
895       sampling rate no matter how CPU time it takes.
897 1-100 Attempt to throttle perf's sample rate to this
920  -1  Allow use of (almost) all events by all users.
941 '``perf record -g``' or '``perf trace --call-graph fp``'.
943 This can only be done when no events are in use that have callchains
944 enabled, otherwise writing to this file will return ``-EBUSY``.
952 Control size of per-cpu ring buffer not counted against mlock limit.
962 instance, when using '``perf record -g``' or '``perf trace --call-graph fp``'.
964 This can only be done when no events are in use that have callchains
965 enabled, otherwise writing to this file will return ``-EBUSY``.
1016 powersave-nap (PPC only)
1019 If set, Linux-PPC will use the 'nap' mode of powersaving,
1052 Value from 0 - 10000 is allowed.
1088 a one-time setting until next reboot: once set, it cannot be changed by
1117   writable for compatibility purposes, but writing to it has no effect
1123   writing to it has no effect on any RNG behavior.
1140     loaded to random addresses.  Also for PIE-linked binaries, the
1150     start of the brk area is randomized.  There are however no known
1151     non-legacy applications that would be broken this way, so for most
1160 real-root-dev
1163 See Documentation/admin-guide/initrd.rst.
1166 reboot-cmd (SPARC only)
1182 this value to 0. On Non-EAS platforms, write operation fails and
1189 Documentation/accounting/delay-accounting.rst. Enabling this feature incurs
1258 See Documentation/userspace-api/seccomp_filter.rst.
1261 sg-big-buff
1316 feature will do you no good unless you also configure your resource
1331   -1   Legacy per-write sysctl value handling, with no printk warnings.
1397 On x86, each "split lock" imposes a system-wide performance penalty. On larger
1399 denials of service to well-behaved and potentially more important users.
1410 0 Disable the mitigation mode - just warns the split lock on kernel log
1412 1 Enable the mitigation mode (this is the default) - penalizes the split
1435 stop-a (SPARC only)
1438 Controls Stop-A:
1441 0 Stop-A has no effect.
1442 1 Stop-A breaks to the PROM (default).
1445 Stop-A is always enabled on a panic, so that the user can return to
1452 See Documentation/admin-guide/sysrq.rst.
1458 Non-zero if the kernel has been tainted. Numeric values, which can be
1459 ORed together. The letters are seen in "Tainted" line of Oops reports.
1469    128  `(D)`  kernel died recently, i.e. there was an OOPS or BUG
1474   4096  `(O)`  externally-built ("out-of-tree") module was loaded
1482 See Documentation/admin-guide/tainted-kernels.rst for more information.
1489   See Documentation/admin-guide/kernel-parameters.rst for more details on
1493 threads-max
1503 The minimum value that can be written to ``threads-max`` is 1.
1505 The maximum value that can be written to ``threads-max`` is given by the
1508 If a value outside of this range is written to ``threads-max`` an
1535 See Documentation/admin-guide/kernel-parameters.rst and
1536 Documentation/trace/boottime-trace.rst.
1539 unaligned-trap
1552 See also `ignore-unaligned-usertrap`_.
1559 value is non-zero, unknown NMI is trapped and then panic occurs. At
1571 will return ``-EPERM``. Once set to 1, this can't be cleared from the
1630 to re-enable cores that by default were not running the watchdog,
1637   echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask