Loading...
Note: File does not exist in v3.1.
1===================================
2Documentation for /proc/sys/kernel/
3===================================
4
5kernel version 2.2.10
6
7Copyright (c) 1998, 1999, Rik van Riel <riel@nl.linux.org>
8
9Copyright (c) 2009, Shen Feng<shen@cn.fujitsu.com>
10
11For general info and legal blurb, please look in index.rst.
12
13------------------------------------------------------------------------------
14
15This file contains documentation for the sysctl files in
16/proc/sys/kernel/ and is valid for Linux kernel version 2.2.
17
18The files in this directory can be used to tune and monitor
19miscellaneous and general things in the operation of the Linux
20kernel. Since some of the files _can_ be used to screw up your
21system, it is advisable to read both documentation and source
22before actually making adjustments.
23
24Currently, these files might (depending on your configuration)
25show up in /proc/sys/kernel:
26
27- acct
28- acpi_video_flags
29- auto_msgmni
30- bootloader_type [ X86 only ]
31- bootloader_version [ X86 only ]
32- cap_last_cap
33- core_pattern
34- core_pipe_limit
35- core_uses_pid
36- ctrl-alt-del
37- dmesg_restrict
38- domainname
39- hostname
40- hotplug
41- hardlockup_all_cpu_backtrace
42- hardlockup_panic
43- hung_task_panic
44- hung_task_check_count
45- hung_task_timeout_secs
46- hung_task_check_interval_secs
47- hung_task_warnings
48- hyperv_record_panic_msg
49- kexec_load_disabled
50- kptr_restrict
51- l2cr [ PPC only ]
52- modprobe ==> Documentation/debugging-modules.txt
53- modules_disabled
54- msg_next_id [ sysv ipc ]
55- msgmax
56- msgmnb
57- msgmni
58- nmi_watchdog
59- osrelease
60- ostype
61- overflowgid
62- overflowuid
63- panic
64- panic_on_oops
65- panic_on_stackoverflow
66- panic_on_unrecovered_nmi
67- panic_on_warn
68- panic_print
69- panic_on_rcu_stall
70- perf_cpu_time_max_percent
71- perf_event_paranoid
72- perf_event_max_stack
73- perf_event_mlock_kb
74- perf_event_max_contexts_per_stack
75- pid_max
76- powersave-nap [ PPC only ]
77- printk
78- printk_delay
79- printk_ratelimit
80- printk_ratelimit_burst
81- pty ==> Documentation/filesystems/devpts.txt
82- randomize_va_space
83- real-root-dev ==> Documentation/admin-guide/initrd.rst
84- reboot-cmd [ SPARC only ]
85- rtsig-max
86- rtsig-nr
87- sched_energy_aware
88- seccomp/ ==> Documentation/userspace-api/seccomp_filter.rst
89- sem
90- sem_next_id [ sysv ipc ]
91- sg-big-buff [ generic SCSI device (sg) ]
92- shm_next_id [ sysv ipc ]
93- shm_rmid_forced
94- shmall
95- shmmax [ sysv ipc ]
96- shmmni
97- softlockup_all_cpu_backtrace
98- soft_watchdog
99- stack_erasing
100- stop-a [ SPARC only ]
101- sysrq ==> Documentation/admin-guide/sysrq.rst
102- sysctl_writes_strict
103- tainted ==> Documentation/admin-guide/tainted-kernels.rst
104- threads-max
105- unknown_nmi_panic
106- watchdog
107- watchdog_thresh
108- version
109
110
111acct:
112=====
113
114highwater lowwater frequency
115
116If BSD-style process accounting is enabled these values control
117its behaviour. If free space on filesystem where the log lives
118goes below <lowwater>% accounting suspends. If free space gets
119above <highwater>% accounting resumes. <Frequency> determines
120how often do we check the amount of free space (value is in
121seconds). Default:
1224 2 30
123That is, suspend accounting if there left <= 2% free; resume it
124if we got >=4%; consider information about amount of free space
125valid for 30 seconds.
126
127
128acpi_video_flags:
129=================
130
131flags
132
133See Doc*/kernel/power/video.txt, it allows mode of video boot to be
134set during run time.
135
136
137auto_msgmni:
138============
139
140This variable has no effect and may be removed in future kernel
141releases. Reading it always returns 0.
142Up to Linux 3.17, it enabled/disabled automatic recomputing of msgmni
143upon memory add/remove or upon ipc namespace creation/removal.
144Echoing "1" into this file enabled msgmni automatic recomputing.
145Echoing "0" turned it off. auto_msgmni default value was 1.
146
147
148bootloader_type:
149================
150
151x86 bootloader identification
152
153This gives the bootloader type number as indicated by the bootloader,
154shifted left by 4, and OR'd with the low four bits of the bootloader
155version. The reason for this encoding is that this used to match the
156type_of_loader field in the kernel header; the encoding is kept for
157backwards compatibility. That is, if the full bootloader type number
158is 0x15 and the full version number is 0x234, this file will contain
159the value 340 = 0x154.
160
161See the type_of_loader and ext_loader_type fields in
162Documentation/x86/boot.rst for additional information.
163
164
165bootloader_version:
166===================
167
168x86 bootloader version
169
170The complete bootloader version number. In the example above, this
171file will contain the value 564 = 0x234.
172
173See the type_of_loader and ext_loader_ver fields in
174Documentation/x86/boot.rst for additional information.
175
176
177cap_last_cap:
178=============
179
180Highest valid capability of the running kernel. Exports
181CAP_LAST_CAP from the kernel.
182
183
184core_pattern:
185=============
186
187core_pattern is used to specify a core dumpfile pattern name.
188
189* max length 127 characters; default value is "core"
190* core_pattern is used as a pattern template for the output filename;
191 certain string patterns (beginning with '%') are substituted with
192 their actual values.
193* backward compatibility with core_uses_pid:
194
195 If core_pattern does not include "%p" (default does not)
196 and core_uses_pid is set, then .PID will be appended to
197 the filename.
198
199* corename format specifiers::
200
201 %<NUL> '%' is dropped
202 %% output one '%'
203 %p pid
204 %P global pid (init PID namespace)
205 %i tid
206 %I global tid (init PID namespace)
207 %u uid (in initial user namespace)
208 %g gid (in initial user namespace)
209 %d dump mode, matches PR_SET_DUMPABLE and
210 /proc/sys/fs/suid_dumpable
211 %s signal number
212 %t UNIX time of dump
213 %h hostname
214 %e executable filename (may be shortened)
215 %E executable path
216 %<OTHER> both are dropped
217
218* If the first character of the pattern is a '|', the kernel will treat
219 the rest of the pattern as a command to run. The core dump will be
220 written to the standard input of that program instead of to a file.
221
222
223core_pipe_limit:
224================
225
226This sysctl is only applicable when core_pattern is configured to pipe
227core files to a user space helper (when the first character of
228core_pattern is a '|', see above). When collecting cores via a pipe
229to an application, it is occasionally useful for the collecting
230application to gather data about the crashing process from its
231/proc/pid directory. In order to do this safely, the kernel must wait
232for the collecting process to exit, so as not to remove the crashing
233processes proc files prematurely. This in turn creates the
234possibility that a misbehaving userspace collecting process can block
235the reaping of a crashed process simply by never exiting. This sysctl
236defends against that. It defines how many concurrent crashing
237processes may be piped to user space applications in parallel. If
238this value is exceeded, then those crashing processes above that value
239are noted via the kernel log and their cores are skipped. 0 is a
240special value, indicating that unlimited processes may be captured in
241parallel, but that no waiting will take place (i.e. the collecting
242process is not guaranteed access to /proc/<crashing pid>/). This
243value defaults to 0.
244
245
246core_uses_pid:
247==============
248
249The default coredump filename is "core". By setting
250core_uses_pid to 1, the coredump filename becomes core.PID.
251If core_pattern does not include "%p" (default does not)
252and core_uses_pid is set, then .PID will be appended to
253the filename.
254
255
256ctrl-alt-del:
257=============
258
259When the value in this file is 0, ctrl-alt-del is trapped and
260sent to the init(1) program to handle a graceful restart.
261When, however, the value is > 0, Linux's reaction to a Vulcan
262Nerve Pinch (tm) will be an immediate reboot, without even
263syncing its dirty buffers.
264
265Note:
266 when a program (like dosemu) has the keyboard in 'raw'
267 mode, the ctrl-alt-del is intercepted by the program before it
268 ever reaches the kernel tty layer, and it's up to the program
269 to decide what to do with it.
270
271
272dmesg_restrict:
273===============
274
275This toggle indicates whether unprivileged users are prevented
276from using dmesg(8) to view messages from the kernel's log buffer.
277When dmesg_restrict is set to (0) there are no restrictions. When
278dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use
279dmesg(8).
280
281The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the
282default value of dmesg_restrict.
283
284
285domainname & hostname:
286======================
287
288These files can be used to set the NIS/YP domainname and the
289hostname of your box in exactly the same way as the commands
290domainname and hostname, i.e.::
291
292 # echo "darkstar" > /proc/sys/kernel/hostname
293 # echo "mydomain" > /proc/sys/kernel/domainname
294
295has the same effect as::
296
297 # hostname "darkstar"
298 # domainname "mydomain"
299
300Note, however, that the classic darkstar.frop.org has the
301hostname "darkstar" and DNS (Internet Domain Name Server)
302domainname "frop.org", not to be confused with the NIS (Network
303Information Service) or YP (Yellow Pages) domainname. These two
304domain names are in general different. For a detailed discussion
305see the hostname(1) man page.
306
307
308hardlockup_all_cpu_backtrace:
309=============================
310
311This value controls the hard lockup detector behavior when a hard
312lockup condition is detected as to whether or not to gather further
313debug information. If enabled, arch-specific all-CPU stack dumping
314will be initiated.
315
3160: do nothing. This is the default behavior.
317
3181: on detection capture more debug information.
319
320
321hardlockup_panic:
322=================
323
324This parameter can be used to control whether the kernel panics
325when a hard lockup is detected.
326
327 0 - don't panic on hard lockup
328 1 - panic on hard lockup
329
330See Documentation/admin-guide/lockup-watchdogs.rst for more information. This can
331also be set using the nmi_watchdog kernel parameter.
332
333
334hotplug:
335========
336
337Path for the hotplug policy agent.
338Default value is "/sbin/hotplug".
339
340
341hung_task_panic:
342================
343
344Controls the kernel's behavior when a hung task is detected.
345This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
346
3470: continue operation. This is the default behavior.
348
3491: panic immediately.
350
351
352hung_task_check_count:
353======================
354
355The upper bound on the number of tasks that are checked.
356This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
357
358
359hung_task_timeout_secs:
360=======================
361
362When a task in D state did not get scheduled
363for more than this value report a warning.
364This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
365
3660: means infinite timeout - no checking done.
367
368Possible values to set are in range {0..LONG_MAX/HZ}.
369
370
371hung_task_check_interval_secs:
372==============================
373
374Hung task check interval. If hung task checking is enabled
375(see hung_task_timeout_secs), the check is done every
376hung_task_check_interval_secs seconds.
377This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
378
3790 (default): means use hung_task_timeout_secs as checking interval.
380Possible values to set are in range {0..LONG_MAX/HZ}.
381
382
383hung_task_warnings:
384===================
385
386The maximum number of warnings to report. During a check interval
387if a hung task is detected, this value is decreased by 1.
388When this value reaches 0, no more warnings will be reported.
389This file shows up if CONFIG_DETECT_HUNG_TASK is enabled.
390
391-1: report an infinite number of warnings.
392
393
394hyperv_record_panic_msg:
395========================
396
397Controls whether the panic kmsg data should be reported to Hyper-V.
398
3990: do not report panic kmsg data.
400
4011: report the panic kmsg data. This is the default behavior.
402
403
404kexec_load_disabled:
405====================
406
407A toggle indicating if the kexec_load syscall has been disabled. This
408value defaults to 0 (false: kexec_load enabled), but can be set to 1
409(true: kexec_load disabled). Once true, kexec can no longer be used, and
410the toggle cannot be set back to false. This allows a kexec image to be
411loaded before disabling the syscall, allowing a system to set up (and
412later use) an image without it being altered. Generally used together
413with the "modules_disabled" sysctl.
414
415
416kptr_restrict:
417==============
418
419This toggle indicates whether restrictions are placed on
420exposing kernel addresses via /proc and other interfaces.
421
422When kptr_restrict is set to 0 (the default) the address is hashed before
423printing. (This is the equivalent to %p.)
424
425When kptr_restrict is set to (1), kernel pointers printed using the %pK
426format specifier will be replaced with 0's unless the user has CAP_SYSLOG
427and effective user and group ids are equal to the real ids. This is
428because %pK checks are done at read() time rather than open() time, so
429if permissions are elevated between the open() and the read() (e.g via
430a setuid binary) then %pK will not leak kernel pointers to unprivileged
431users. Note, this is a temporary solution only. The correct long-term
432solution is to do the permission checks at open() time. Consider removing
433world read permissions from files that use %pK, and using dmesg_restrict
434to protect against uses of %pK in dmesg(8) if leaking kernel pointer
435values to unprivileged users is a concern.
436
437When kptr_restrict is set to (2), kernel pointers printed using
438%pK will be replaced with 0's regardless of privileges.
439
440
441l2cr: (PPC only)
442================
443
444This flag controls the L2 cache of G3 processor boards. If
4450, the cache is disabled. Enabled if nonzero.
446
447
448modules_disabled:
449=================
450
451A toggle value indicating if modules are allowed to be loaded
452in an otherwise modular kernel. This toggle defaults to off
453(0), but can be set true (1). Once true, modules can be
454neither loaded nor unloaded, and the toggle cannot be set back
455to false. Generally used with the "kexec_load_disabled" toggle.
456
457
458msg_next_id, sem_next_id, and shm_next_id:
459==========================================
460
461These three toggles allows to specify desired id for next allocated IPC
462object: message, semaphore or shared memory respectively.
463
464By default they are equal to -1, which means generic allocation logic.
465Possible values to set are in range {0..INT_MAX}.
466
467Notes:
468 1) kernel doesn't guarantee, that new object will have desired id. So,
469 it's up to userspace, how to handle an object with "wrong" id.
470 2) Toggle with non-default value will be set back to -1 by kernel after
471 successful IPC object allocation. If an IPC object allocation syscall
472 fails, it is undefined if the value remains unmodified or is reset to -1.
473
474
475nmi_watchdog:
476=============
477
478This parameter can be used to control the NMI watchdog
479(i.e. the hard lockup detector) on x86 systems.
480
4810 - disable the hard lockup detector
482
4831 - enable the hard lockup detector
484
485The hard lockup detector monitors each CPU for its ability to respond to
486timer interrupts. The mechanism utilizes CPU performance counter registers
487that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
488while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
489
490The NMI watchdog is disabled by default if the kernel is running as a guest
491in a KVM virtual machine. This default can be overridden by adding::
492
493 nmi_watchdog=1
494
495to the guest kernel command line (see Documentation/admin-guide/kernel-parameters.rst).
496
497
498numa_balancing:
499===============
500
501Enables/disables automatic page fault based NUMA memory
502balancing. Memory is moved automatically to nodes
503that access it often.
504
505Enables/disables automatic NUMA memory balancing. On NUMA machines, there
506is a performance penalty if remote memory is accessed by a CPU. When this
507feature is enabled the kernel samples what task thread is accessing memory
508by periodically unmapping pages and later trapping a page fault. At the
509time of the page fault, it is determined if the data being accessed should
510be migrated to a local memory node.
511
512The unmapping of pages and trapping faults incur additional overhead that
513ideally is offset by improved memory locality but there is no universal
514guarantee. If the target workload is already bound to NUMA nodes then this
515feature should be disabled. Otherwise, if the system overhead from the
516feature is too high then the rate the kernel samples for NUMA hinting
517faults may be controlled by the numa_balancing_scan_period_min_ms,
518numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms,
519numa_balancing_scan_size_mb, and numa_balancing_settle_count sysctls.
520
521numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb
522===============================================================================================================================
523
524
525Automatic NUMA balancing scans tasks address space and unmaps pages to
526detect if pages are properly placed or if the data should be migrated to a
527memory node local to where the task is running. Every "scan delay" the task
528scans the next "scan size" number of pages in its address space. When the
529end of the address space is reached the scanner restarts from the beginning.
530
531In combination, the "scan delay" and "scan size" determine the scan rate.
532When "scan delay" decreases, the scan rate increases. The scan delay and
533hence the scan rate of every task is adaptive and depends on historical
534behaviour. If pages are properly placed then the scan delay increases,
535otherwise the scan delay decreases. The "scan size" is not adaptive but
536the higher the "scan size", the higher the scan rate.
537
538Higher scan rates incur higher system overhead as page faults must be
539trapped and potentially data must be migrated. However, the higher the scan
540rate, the more quickly a tasks memory is migrated to a local node if the
541workload pattern changes and minimises performance impact due to remote
542memory accesses. These sysctls control the thresholds for scan delays and
543the number of pages scanned.
544
545numa_balancing_scan_period_min_ms is the minimum time in milliseconds to
546scan a tasks virtual memory. It effectively controls the maximum scanning
547rate for each task.
548
549numa_balancing_scan_delay_ms is the starting "scan delay" used for a task
550when it initially forks.
551
552numa_balancing_scan_period_max_ms is the maximum time in milliseconds to
553scan a tasks virtual memory. It effectively controls the minimum scanning
554rate for each task.
555
556numa_balancing_scan_size_mb is how many megabytes worth of pages are
557scanned for a given scan.
558
559
560osrelease, ostype & version:
561============================
562
563::
564
565 # cat osrelease
566 2.1.88
567 # cat ostype
568 Linux
569 # cat version
570 #5 Wed Feb 25 21:49:24 MET 1998
571
572The files osrelease and ostype should be clear enough. Version
573needs a little more clarification however. The '#5' means that
574this is the fifth kernel built from this source base and the
575date behind it indicates the time the kernel was built.
576The only way to tune these values is to rebuild the kernel :-)
577
578
579overflowgid & overflowuid:
580==========================
581
582if your architecture did not always support 32-bit UIDs (i.e. arm,
583i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
584applications that use the old 16-bit UID/GID system calls, if the
585actual UID or GID would exceed 65535.
586
587These sysctls allow you to change the value of the fixed UID and GID.
588The default is 65534.
589
590
591panic:
592======
593
594The value in this file represents the number of seconds the kernel
595waits before rebooting on a panic. When you use the software watchdog,
596the recommended setting is 60.
597
598
599panic_on_io_nmi:
600================
601
602Controls the kernel's behavior when a CPU receives an NMI caused by
603an IO error.
604
6050: try to continue operation (default)
606
6071: panic immediately. The IO error triggered an NMI. This indicates a
608 serious system condition which could result in IO data corruption.
609 Rather than continuing, panicking might be a better choice. Some
610 servers issue this sort of NMI when the dump button is pushed,
611 and you can use this option to take a crash dump.
612
613
614panic_on_oops:
615==============
616
617Controls the kernel's behaviour when an oops or BUG is encountered.
618
6190: try to continue operation
620
6211: panic immediately. If the `panic` sysctl is also non-zero then the
622 machine will be rebooted.
623
624
625panic_on_stackoverflow:
626=======================
627
628Controls the kernel's behavior when detecting the overflows of
629kernel, IRQ and exception stacks except a user stack.
630This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled.
631
6320: try to continue operation.
633
6341: panic immediately.
635
636
637panic_on_unrecovered_nmi:
638=========================
639
640The default Linux behaviour on an NMI of either memory or unknown is
641to continue operation. For many environments such as scientific
642computing it is preferable that the box is taken out and the error
643dealt with than an uncorrected parity/ECC error get propagated.
644
645A small number of systems do generate NMI's for bizarre random reasons
646such as power management so the default is off. That sysctl works like
647the existing panic controls already in that directory.
648
649
650panic_on_warn:
651==============
652
653Calls panic() in the WARN() path when set to 1. This is useful to avoid
654a kernel rebuild when attempting to kdump at the location of a WARN().
655
6560: only WARN(), default behaviour.
657
6581: call panic() after printing out WARN() location.
659
660
661panic_print:
662============
663
664Bitmask for printing system info when panic happens. User can chose
665combination of the following bits:
666
667===== ========================================
668bit 0 print all tasks info
669bit 1 print system memory info
670bit 2 print timer info
671bit 3 print locks info if CONFIG_LOCKDEP is on
672bit 4 print ftrace buffer
673===== ========================================
674
675So for example to print tasks and memory info on panic, user can::
676
677 echo 3 > /proc/sys/kernel/panic_print
678
679
680panic_on_rcu_stall:
681===================
682
683When set to 1, calls panic() after RCU stall detection messages. This
684is useful to define the root cause of RCU stalls using a vmcore.
685
6860: do not panic() when RCU stall takes place, default behavior.
687
6881: panic() after printing RCU stall messages.
689
690
691perf_cpu_time_max_percent:
692==========================
693
694Hints to the kernel how much CPU time it should be allowed to
695use to handle perf sampling events. If the perf subsystem
696is informed that its samples are exceeding this limit, it
697will drop its sampling frequency to attempt to reduce its CPU
698usage.
699
700Some perf sampling happens in NMIs. If these samples
701unexpectedly take too long to execute, the NMIs can become
702stacked up next to each other so much that nothing else is
703allowed to execute.
704
7050:
706 disable the mechanism. Do not monitor or correct perf's
707 sampling rate no matter how CPU time it takes.
708
7091-100:
710 attempt to throttle perf's sample rate to this
711 percentage of CPU. Note: the kernel calculates an
712 "expected" length of each sample event. 100 here means
713 100% of that expected length. Even if this is set to
714 100, you may still see sample throttling if this
715 length is exceeded. Set to 0 if you truly do not care
716 how much CPU is consumed.
717
718
719perf_event_paranoid:
720====================
721
722Controls use of the performance events system by unprivileged
723users (without CAP_SYS_ADMIN). The default value is 2.
724
725=== ==================================================================
726 -1 Allow use of (almost) all events by all users
727
728 Ignore mlock limit after perf_event_mlock_kb without CAP_IPC_LOCK
729
730>=0 Disallow ftrace function tracepoint by users without CAP_SYS_ADMIN
731
732 Disallow raw tracepoint access by users without CAP_SYS_ADMIN
733
734>=1 Disallow CPU event access by users without CAP_SYS_ADMIN
735
736>=2 Disallow kernel profiling by users without CAP_SYS_ADMIN
737=== ==================================================================
738
739
740perf_event_max_stack:
741=====================
742
743Controls maximum number of stack frames to copy for (attr.sample_type &
744PERF_SAMPLE_CALLCHAIN) configured events, for instance, when using
745'perf record -g' or 'perf trace --call-graph fp'.
746
747This can only be done when no events are in use that have callchains
748enabled, otherwise writing to this file will return -EBUSY.
749
750The default value is 127.
751
752
753perf_event_mlock_kb:
754====================
755
756Control size of per-cpu ring buffer not counted agains mlock limit.
757
758The default value is 512 + 1 page
759
760
761perf_event_max_contexts_per_stack:
762==================================
763
764Controls maximum number of stack frame context entries for
765(attr.sample_type & PERF_SAMPLE_CALLCHAIN) configured events, for
766instance, when using 'perf record -g' or 'perf trace --call-graph fp'.
767
768This can only be done when no events are in use that have callchains
769enabled, otherwise writing to this file will return -EBUSY.
770
771The default value is 8.
772
773
774pid_max:
775========
776
777PID allocation wrap value. When the kernel's next PID value
778reaches this value, it wraps back to a minimum PID value.
779PIDs of value pid_max or larger are not allocated.
780
781
782ns_last_pid:
783============
784
785The last pid allocated in the current (the one task using this sysctl
786lives in) pid namespace. When selecting a pid for a next task on fork
787kernel tries to allocate a number starting from this one.
788
789
790powersave-nap: (PPC only)
791=========================
792
793If set, Linux-PPC will use the 'nap' mode of powersaving,
794otherwise the 'doze' mode will be used.
795
796==============================================================
797
798printk:
799=======
800
801The four values in printk denote: console_loglevel,
802default_message_loglevel, minimum_console_loglevel and
803default_console_loglevel respectively.
804
805These values influence printk() behavior when printing or
806logging error messages. See 'man 2 syslog' for more info on
807the different loglevels.
808
809- console_loglevel:
810 messages with a higher priority than
811 this will be printed to the console
812- default_message_loglevel:
813 messages without an explicit priority
814 will be printed with this priority
815- minimum_console_loglevel:
816 minimum (highest) value to which
817 console_loglevel can be set
818- default_console_loglevel:
819 default value for console_loglevel
820
821
822printk_delay:
823=============
824
825Delay each printk message in printk_delay milliseconds
826
827Value from 0 - 10000 is allowed.
828
829
830printk_ratelimit:
831=================
832
833Some warning messages are rate limited. printk_ratelimit specifies
834the minimum length of time between these messages (in jiffies), by
835default we allow one every 5 seconds.
836
837A value of 0 will disable rate limiting.
838
839
840printk_ratelimit_burst:
841=======================
842
843While long term we enforce one message per printk_ratelimit
844seconds, we do allow a burst of messages to pass through.
845printk_ratelimit_burst specifies the number of messages we can
846send before ratelimiting kicks in.
847
848
849printk_devkmsg:
850===============
851
852Control the logging to /dev/kmsg from userspace:
853
854ratelimit:
855 default, ratelimited
856
857on: unlimited logging to /dev/kmsg from userspace
858
859off: logging to /dev/kmsg disabled
860
861The kernel command line parameter printk.devkmsg= overrides this and is
862a one-time setting until next reboot: once set, it cannot be changed by
863this sysctl interface anymore.
864
865
866randomize_va_space:
867===================
868
869This option can be used to select the type of process address
870space randomization that is used in the system, for architectures
871that support this feature.
872
873== ===========================================================================
8740 Turn the process address space randomization off. This is the
875 default for architectures that do not support this feature anyways,
876 and kernels that are booted with the "norandmaps" parameter.
877
8781 Make the addresses of mmap base, stack and VDSO page randomized.
879 This, among other things, implies that shared libraries will be
880 loaded to random addresses. Also for PIE-linked binaries, the
881 location of code start is randomized. This is the default if the
882 CONFIG_COMPAT_BRK option is enabled.
883
8842 Additionally enable heap randomization. This is the default if
885 CONFIG_COMPAT_BRK is disabled.
886
887 There are a few legacy applications out there (such as some ancient
888 versions of libc.so.5 from 1996) that assume that brk area starts
889 just after the end of the code+bss. These applications break when
890 start of the brk area is randomized. There are however no known
891 non-legacy applications that would be broken this way, so for most
892 systems it is safe to choose full randomization.
893
894 Systems with ancient and/or broken binaries should be configured
895 with CONFIG_COMPAT_BRK enabled, which excludes the heap from process
896 address space randomization.
897== ===========================================================================
898
899
900reboot-cmd: (Sparc only)
901========================
902
903??? This seems to be a way to give an argument to the Sparc
904ROM/Flash boot loader. Maybe to tell it what to do after
905rebooting. ???
906
907
908rtsig-max & rtsig-nr:
909=====================
910
911The file rtsig-max can be used to tune the maximum number
912of POSIX realtime (queued) signals that can be outstanding
913in the system.
914
915rtsig-nr shows the number of RT signals currently queued.
916
917
918sched_energy_aware:
919===================
920
921Enables/disables Energy Aware Scheduling (EAS). EAS starts
922automatically on platforms where it can run (that is,
923platforms with asymmetric CPU topologies and having an Energy
924Model available). If your platform happens to meet the
925requirements for EAS but you do not want to use it, change
926this value to 0.
927
928
929sched_schedstats:
930=================
931
932Enables/disables scheduler statistics. Enabling this feature
933incurs a small amount of overhead in the scheduler but is
934useful for debugging and performance tuning.
935
936
937sg-big-buff:
938============
939
940This file shows the size of the generic SCSI (sg) buffer.
941You can't tune it just yet, but you could change it on
942compile time by editing include/scsi/sg.h and changing
943the value of SG_BIG_BUFF.
944
945There shouldn't be any reason to change this value. If
946you can come up with one, you probably know what you
947are doing anyway :)
948
949
950shmall:
951=======
952
953This parameter sets the total amount of shared memory pages that
954can be used system wide. Hence, SHMALL should always be at least
955ceil(shmmax/PAGE_SIZE).
956
957If you are not sure what the default PAGE_SIZE is on your Linux
958system, you can run the following command:
959
960 # getconf PAGE_SIZE
961
962
963shmmax:
964=======
965
966This value can be used to query and set the run time limit
967on the maximum shared memory segment size that can be created.
968Shared memory segments up to 1Gb are now supported in the
969kernel. This value defaults to SHMMAX.
970
971
972shm_rmid_forced:
973================
974
975Linux lets you set resource limits, including how much memory one
976process can consume, via setrlimit(2). Unfortunately, shared memory
977segments are allowed to exist without association with any process, and
978thus might not be counted against any resource limits. If enabled,
979shared memory segments are automatically destroyed when their attach
980count becomes zero after a detach or a process termination. It will
981also destroy segments that were created, but never attached to, on exit
982from the process. The only use left for IPC_RMID is to immediately
983destroy an unattached segment. Of course, this breaks the way things are
984defined, so some applications might stop working. Note that this
985feature will do you no good unless you also configure your resource
986limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't
987need this.
988
989Note that if you change this from 0 to 1, already created segments
990without users and with a dead originative process will be destroyed.
991
992
993sysctl_writes_strict:
994=====================
995
996Control how file position affects the behavior of updating sysctl values
997via the /proc/sys interface:
998
999 == ======================================================================
1000 -1 Legacy per-write sysctl value handling, with no printk warnings.
1001 Each write syscall must fully contain the sysctl value to be
1002 written, and multiple writes on the same sysctl file descriptor
1003 will rewrite the sysctl value, regardless of file position.
1004 0 Same behavior as above, but warn about processes that perform writes
1005 to a sysctl file descriptor when the file position is not 0.
1006 1 (default) Respect file position when writing sysctl strings. Multiple
1007 writes will append to the sysctl value buffer. Anything past the max
1008 length of the sysctl value buffer will be ignored. Writes to numeric
1009 sysctl entries must always be at file position 0 and the value must
1010 be fully contained in the buffer sent in the write syscall.
1011 == ======================================================================
1012
1013
1014softlockup_all_cpu_backtrace:
1015=============================
1016
1017This value controls the soft lockup detector thread's behavior
1018when a soft lockup condition is detected as to whether or not
1019to gather further debug information. If enabled, each cpu will
1020be issued an NMI and instructed to capture stack trace.
1021
1022This feature is only applicable for architectures which support
1023NMI.
1024
10250: do nothing. This is the default behavior.
1026
10271: on detection capture more debug information.
1028
1029
1030soft_watchdog:
1031==============
1032
1033This parameter can be used to control the soft lockup detector.
1034
1035 0 - disable the soft lockup detector
1036
1037 1 - enable the soft lockup detector
1038
1039The soft lockup detector monitors CPUs for threads that are hogging the CPUs
1040without rescheduling voluntarily, and thus prevent the 'watchdog/N' threads
1041from running. The mechanism depends on the CPUs ability to respond to timer
1042interrupts which are needed for the 'watchdog/N' threads to be woken up by
1043the watchdog timer function, otherwise the NMI watchdog - if enabled - can
1044detect a hard lockup condition.
1045
1046
1047stack_erasing:
1048==============
1049
1050This parameter can be used to control kernel stack erasing at the end
1051of syscalls for kernels built with CONFIG_GCC_PLUGIN_STACKLEAK.
1052
1053That erasing reduces the information which kernel stack leak bugs
1054can reveal and blocks some uninitialized stack variable attacks.
1055The tradeoff is the performance impact: on a single CPU system kernel
1056compilation sees a 1% slowdown, other systems and workloads may vary.
1057
1058 0: kernel stack erasing is disabled, STACKLEAK_METRICS are not updated.
1059
1060 1: kernel stack erasing is enabled (default), it is performed before
1061 returning to the userspace at the end of syscalls.
1062
1063
1064tainted
1065=======
1066
1067Non-zero if the kernel has been tainted. Numeric values, which can be
1068ORed together. The letters are seen in "Tainted" line of Oops reports.
1069
1070====== ===== ==============================================================
1071 1 `(P)` proprietary module was loaded
1072 2 `(F)` module was force loaded
1073 4 `(S)` SMP kernel oops on an officially SMP incapable processor
1074 8 `(R)` module was force unloaded
1075 16 `(M)` processor reported a Machine Check Exception (MCE)
1076 32 `(B)` bad page referenced or some unexpected page flags
1077 64 `(U)` taint requested by userspace application
1078 128 `(D)` kernel died recently, i.e. there was an OOPS or BUG
1079 256 `(A)` an ACPI table was overridden by user
1080 512 `(W)` kernel issued warning
1081 1024 `(C)` staging driver was loaded
1082 2048 `(I)` workaround for bug in platform firmware applied
1083 4096 `(O)` externally-built ("out-of-tree") module was loaded
1084 8192 `(E)` unsigned module was loaded
1085 16384 `(L)` soft lockup occurred
1086 32768 `(K)` kernel has been live patched
1087 65536 `(X)` Auxiliary taint, defined and used by for distros
1088131072 `(T)` The kernel was built with the struct randomization plugin
1089====== ===== ==============================================================
1090
1091See Documentation/admin-guide/tainted-kernels.rst for more information.
1092
1093
1094threads-max:
1095============
1096
1097This value controls the maximum number of threads that can be created
1098using fork().
1099
1100During initialization the kernel sets this value such that even if the
1101maximum number of threads is created, the thread structures occupy only
1102a part (1/8th) of the available RAM pages.
1103
1104The minimum value that can be written to threads-max is 20.
1105
1106The maximum value that can be written to threads-max is given by the
1107constant FUTEX_TID_MASK (0x3fffffff).
1108
1109If a value outside of this range is written to threads-max an error
1110EINVAL occurs.
1111
1112The value written is checked against the available RAM pages. If the
1113thread structures would occupy too much (more than 1/8th) of the
1114available RAM pages threads-max is reduced accordingly.
1115
1116
1117unknown_nmi_panic:
1118==================
1119
1120The value in this file affects behavior of handling NMI. When the
1121value is non-zero, unknown NMI is trapped and then panic occurs. At
1122that time, kernel debugging information is displayed on console.
1123
1124NMI switch that most IA32 servers have fires unknown NMI up, for
1125example. If a system hangs up, try pressing the NMI switch.
1126
1127
1128watchdog:
1129=========
1130
1131This parameter can be used to disable or enable the soft lockup detector
1132_and_ the NMI watchdog (i.e. the hard lockup detector) at the same time.
1133
1134 0 - disable both lockup detectors
1135
1136 1 - enable both lockup detectors
1137
1138The soft lockup detector and the NMI watchdog can also be disabled or
1139enabled individually, using the soft_watchdog and nmi_watchdog parameters.
1140If the watchdog parameter is read, for example by executing::
1141
1142 cat /proc/sys/kernel/watchdog
1143
1144the output of this command (0 or 1) shows the logical OR of soft_watchdog
1145and nmi_watchdog.
1146
1147
1148watchdog_cpumask:
1149=================
1150
1151This value can be used to control on which cpus the watchdog may run.
1152The default cpumask is all possible cores, but if NO_HZ_FULL is
1153enabled in the kernel config, and cores are specified with the
1154nohz_full= boot argument, those cores are excluded by default.
1155Offline cores can be included in this mask, and if the core is later
1156brought online, the watchdog will be started based on the mask value.
1157
1158Typically this value would only be touched in the nohz_full case
1159to re-enable cores that by default were not running the watchdog,
1160if a kernel lockup was suspected on those cores.
1161
1162The argument value is the standard cpulist format for cpumasks,
1163so for example to enable the watchdog on cores 0, 2, 3, and 4 you
1164might say::
1165
1166 echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
1167
1168
1169watchdog_thresh:
1170================
1171
1172This value can be used to control the frequency of hrtimer and NMI
1173events and the soft and hard lockup thresholds. The default threshold
1174is 10 seconds.
1175
1176The softlockup threshold is (2 * watchdog_thresh). Setting this
1177tunable to zero will disable lockup detection altogether.