Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | Documentation for /proc/sys/kernel/* kernel version 2.2.10 |
2 | (c) 1998, 1999, Rik van Riel <riel@nl.linux.org> | |
760df93e | 3 | (c) 2009, Shen Feng<shen@cn.fujitsu.com> |
1da177e4 LT |
4 | |
5 | For general info and legal blurb, please look in README. | |
6 | ||
7 | ============================================================== | |
8 | ||
9 | This file contains documentation for the sysctl files in | |
10 | /proc/sys/kernel/ and is valid for Linux kernel version 2.2. | |
11 | ||
12 | The files in this directory can be used to tune and monitor | |
13 | miscellaneous and general things in the operation of the Linux | |
14 | kernel. Since some of the files _can_ be used to screw up your | |
15 | system, it is advisable to read both documentation and source | |
16 | before actually making adjustments. | |
17 | ||
18 | Currently, these files might (depending on your configuration) | |
19 | show up in /proc/sys/kernel: | |
807094c0 | 20 | |
1da177e4 | 21 | - acct |
807094c0 BP |
22 | - acpi_video_flags |
23 | - auto_msgmni | |
d75757ab PA |
24 | - bootloader_type [ X86 only ] |
25 | - bootloader_version [ X86 only ] | |
c114728a | 26 | - callhome [ S390 only ] |
73efc039 | 27 | - cap_last_cap |
1da177e4 | 28 | - core_pattern |
a293980c | 29 | - core_pipe_limit |
1da177e4 LT |
30 | - core_uses_pid |
31 | - ctrl-alt-del | |
eaf06b24 | 32 | - dmesg_restrict |
1da177e4 LT |
33 | - domainname |
34 | - hostname | |
35 | - hotplug | |
270750db AT |
36 | - hung_task_panic |
37 | - hung_task_check_count | |
38 | - hung_task_timeout_secs | |
39 | - hung_task_warnings | |
7984754b | 40 | - kexec_load_disabled |
455cd5ab | 41 | - kptr_restrict |
0741f4d2 | 42 | - kstack_depth_to_print [ X86 only ] |
1da177e4 | 43 | - l2cr [ PPC only ] |
ac76cff2 | 44 | - modprobe ==> Documentation/debugging-modules.txt |
3d43321b | 45 | - modules_disabled |
03f59566 | 46 | - msg_next_id [ sysv ipc ] |
1da177e4 LT |
47 | - msgmax |
48 | - msgmnb | |
49 | - msgmni | |
760df93e | 50 | - nmi_watchdog |
1da177e4 LT |
51 | - osrelease |
52 | - ostype | |
53 | - overflowgid | |
54 | - overflowuid | |
55 | - panic | |
807094c0 BP |
56 | - panic_on_oops |
57 | - panic_on_unrecovered_nmi | |
55af7796 | 58 | - panic_on_stackoverflow |
1da177e4 LT |
59 | - pid_max |
60 | - powersave-nap [ PPC only ] | |
61 | - printk | |
807094c0 BP |
62 | - printk_delay |
63 | - printk_ratelimit | |
64 | - printk_ratelimit_burst | |
1ec7fd50 | 65 | - randomize_va_space |
1da177e4 LT |
66 | - real-root-dev ==> Documentation/initrd.txt |
67 | - reboot-cmd [ SPARC only ] | |
68 | - rtsig-max | |
69 | - rtsig-nr | |
70 | - sem | |
03f59566 | 71 | - sem_next_id [ sysv ipc ] |
1da177e4 | 72 | - sg-big-buff [ generic SCSI device (sg) ] |
03f59566 | 73 | - shm_next_id [ sysv ipc ] |
b34a6b1d | 74 | - shm_rmid_forced |
1da177e4 LT |
75 | - shmall |
76 | - shmmax [ sysv ipc ] | |
77 | - shmmni | |
ed235875 | 78 | - softlockup_all_cpu_backtrace |
1da177e4 LT |
79 | - stop-a [ SPARC only ] |
80 | - sysrq ==> Documentation/sysrq.txt | |
f4aacea2 | 81 | - sysctl_writes_strict |
1da177e4 LT |
82 | - tainted |
83 | - threads-max | |
760df93e | 84 | - unknown_nmi_panic |
08825c90 | 85 | - watchdog_thresh |
1da177e4 LT |
86 | - version |
87 | ||
88 | ============================================================== | |
89 | ||
90 | acct: | |
91 | ||
92 | highwater lowwater frequency | |
93 | ||
94 | If BSD-style process accounting is enabled these values control | |
95 | its behaviour. If free space on filesystem where the log lives | |
96 | goes below <lowwater>% accounting suspends. If free space gets | |
97 | above <highwater>% accounting resumes. <Frequency> determines | |
98 | how often do we check the amount of free space (value is in | |
99 | seconds). Default: | |
100 | 4 2 30 | |
101 | That is, suspend accounting if there left <= 2% free; resume it | |
102 | if we got >=4%; consider information about amount of free space | |
103 | valid for 30 seconds. | |
104 | ||
807094c0 BP |
105 | ============================================================== |
106 | ||
107 | acpi_video_flags: | |
108 | ||
109 | flags | |
110 | ||
111 | See Doc*/kernel/power/video.txt, it allows mode of video boot to be | |
112 | set during run time. | |
113 | ||
114 | ============================================================== | |
115 | ||
116 | auto_msgmni: | |
117 | ||
118 | Enables/Disables automatic recomputing of msgmni upon memory add/remove | |
119 | or upon ipc namespace creation/removal (see the msgmni description | |
120 | above). Echoing "1" into this file enables msgmni automatic recomputing. | |
121 | Echoing "0" turns it off. auto_msgmni default value is 1. | |
122 | ||
123 | ||
1da177e4 LT |
124 | ============================================================== |
125 | ||
d75757ab PA |
126 | bootloader_type: |
127 | ||
128 | x86 bootloader identification | |
129 | ||
130 | This gives the bootloader type number as indicated by the bootloader, | |
131 | shifted left by 4, and OR'd with the low four bits of the bootloader | |
132 | version. The reason for this encoding is that this used to match the | |
133 | type_of_loader field in the kernel header; the encoding is kept for | |
134 | backwards compatibility. That is, if the full bootloader type number | |
135 | is 0x15 and the full version number is 0x234, this file will contain | |
136 | the value 340 = 0x154. | |
137 | ||
138 | See the type_of_loader and ext_loader_type fields in | |
139 | Documentation/x86/boot.txt for additional information. | |
140 | ||
141 | ============================================================== | |
142 | ||
143 | bootloader_version: | |
144 | ||
145 | x86 bootloader version | |
146 | ||
147 | The complete bootloader version number. In the example above, this | |
148 | file will contain the value 564 = 0x234. | |
149 | ||
150 | See the type_of_loader and ext_loader_ver fields in | |
151 | Documentation/x86/boot.txt for additional information. | |
152 | ||
153 | ============================================================== | |
154 | ||
c114728a HJP |
155 | callhome: |
156 | ||
157 | Controls the kernel's callhome behavior in case of a kernel panic. | |
158 | ||
159 | The s390 hardware allows an operating system to send a notification | |
160 | to a service organization (callhome) in case of an operating system panic. | |
161 | ||
162 | When the value in this file is 0 (which is the default behavior) | |
163 | nothing happens in case of a kernel panic. If this value is set to "1" | |
164 | the complete kernel oops message is send to the IBM customer service | |
165 | organization in case the mainframe the Linux operating system is running | |
166 | on has a service contract with IBM. | |
167 | ||
168 | ============================================================== | |
169 | ||
73efc039 DB |
170 | cap_last_cap |
171 | ||
172 | Highest valid capability of the running kernel. Exports | |
173 | CAP_LAST_CAP from the kernel. | |
174 | ||
175 | ============================================================== | |
176 | ||
1da177e4 LT |
177 | core_pattern: |
178 | ||
179 | core_pattern is used to specify a core dumpfile pattern name. | |
cd081041 | 180 | . max length 128 characters; default value is "core" |
1da177e4 LT |
181 | . core_pattern is used as a pattern template for the output filename; |
182 | certain string patterns (beginning with '%') are substituted with | |
183 | their actual values. | |
184 | . backward compatibility with core_uses_pid: | |
185 | If core_pattern does not include "%p" (default does not) | |
186 | and core_uses_pid is set, then .PID will be appended to | |
187 | the filename. | |
188 | . corename format specifiers: | |
189 | %<NUL> '%' is dropped | |
190 | %% output one '%' | |
191 | %p pid | |
65aafb1e | 192 | %P global pid (init PID namespace) |
1da177e4 LT |
193 | %u uid |
194 | %g gid | |
12a2b4b2 ON |
195 | %d dump mode, matches PR_SET_DUMPABLE and |
196 | /proc/sys/fs/suid_dumpable | |
1da177e4 LT |
197 | %s signal number |
198 | %t UNIX time of dump | |
199 | %h hostname | |
57cc083a JS |
200 | %e executable filename (may be shortened) |
201 | %E executable path | |
1da177e4 | 202 | %<OTHER> both are dropped |
cd081041 MU |
203 | . If the first character of the pattern is a '|', the kernel will treat |
204 | the rest of the pattern as a command to run. The core dump will be | |
205 | written to the standard input of that program instead of to a file. | |
1da177e4 LT |
206 | |
207 | ============================================================== | |
208 | ||
a293980c NH |
209 | core_pipe_limit: |
210 | ||
807094c0 BP |
211 | This sysctl is only applicable when core_pattern is configured to pipe |
212 | core files to a user space helper (when the first character of | |
213 | core_pattern is a '|', see above). When collecting cores via a pipe | |
214 | to an application, it is occasionally useful for the collecting | |
215 | application to gather data about the crashing process from its | |
216 | /proc/pid directory. In order to do this safely, the kernel must wait | |
217 | for the collecting process to exit, so as not to remove the crashing | |
218 | processes proc files prematurely. This in turn creates the | |
219 | possibility that a misbehaving userspace collecting process can block | |
220 | the reaping of a crashed process simply by never exiting. This sysctl | |
221 | defends against that. It defines how many concurrent crashing | |
222 | processes may be piped to user space applications in parallel. If | |
223 | this value is exceeded, then those crashing processes above that value | |
224 | are noted via the kernel log and their cores are skipped. 0 is a | |
225 | special value, indicating that unlimited processes may be captured in | |
226 | parallel, but that no waiting will take place (i.e. the collecting | |
227 | process is not guaranteed access to /proc/<crashing pid>/). This | |
228 | value defaults to 0. | |
a293980c NH |
229 | |
230 | ============================================================== | |
231 | ||
1da177e4 LT |
232 | core_uses_pid: |
233 | ||
234 | The default coredump filename is "core". By setting | |
235 | core_uses_pid to 1, the coredump filename becomes core.PID. | |
236 | If core_pattern does not include "%p" (default does not) | |
237 | and core_uses_pid is set, then .PID will be appended to | |
238 | the filename. | |
239 | ||
240 | ============================================================== | |
241 | ||
242 | ctrl-alt-del: | |
243 | ||
244 | When the value in this file is 0, ctrl-alt-del is trapped and | |
245 | sent to the init(1) program to handle a graceful restart. | |
246 | When, however, the value is > 0, Linux's reaction to a Vulcan | |
247 | Nerve Pinch (tm) will be an immediate reboot, without even | |
248 | syncing its dirty buffers. | |
249 | ||
250 | Note: when a program (like dosemu) has the keyboard in 'raw' | |
251 | mode, the ctrl-alt-del is intercepted by the program before it | |
252 | ever reaches the kernel tty layer, and it's up to the program | |
253 | to decide what to do with it. | |
254 | ||
255 | ============================================================== | |
256 | ||
eaf06b24 DR |
257 | dmesg_restrict: |
258 | ||
807094c0 BP |
259 | This toggle indicates whether unprivileged users are prevented |
260 | from using dmesg(8) to view messages from the kernel's log buffer. | |
261 | When dmesg_restrict is set to (0) there are no restrictions. When | |
38ef4c2e | 262 | dmesg_restrict is set set to (1), users must have CAP_SYSLOG to use |
eaf06b24 DR |
263 | dmesg(8). |
264 | ||
807094c0 BP |
265 | The kernel config option CONFIG_SECURITY_DMESG_RESTRICT sets the |
266 | default value of dmesg_restrict. | |
eaf06b24 DR |
267 | |
268 | ============================================================== | |
269 | ||
1da177e4 LT |
270 | domainname & hostname: |
271 | ||
272 | These files can be used to set the NIS/YP domainname and the | |
273 | hostname of your box in exactly the same way as the commands | |
274 | domainname and hostname, i.e.: | |
275 | # echo "darkstar" > /proc/sys/kernel/hostname | |
276 | # echo "mydomain" > /proc/sys/kernel/domainname | |
277 | has the same effect as | |
278 | # hostname "darkstar" | |
279 | # domainname "mydomain" | |
280 | ||
281 | Note, however, that the classic darkstar.frop.org has the | |
282 | hostname "darkstar" and DNS (Internet Domain Name Server) | |
283 | domainname "frop.org", not to be confused with the NIS (Network | |
284 | Information Service) or YP (Yellow Pages) domainname. These two | |
285 | domain names are in general different. For a detailed discussion | |
286 | see the hostname(1) man page. | |
287 | ||
288 | ============================================================== | |
289 | ||
290 | hotplug: | |
291 | ||
292 | Path for the hotplug policy agent. | |
293 | Default value is "/sbin/hotplug". | |
294 | ||
295 | ============================================================== | |
296 | ||
270750db AT |
297 | hung_task_panic: |
298 | ||
299 | Controls the kernel's behavior when a hung task is detected. | |
300 | This file shows up if CONFIG_DETECT_HUNG_TASK is enabled. | |
301 | ||
302 | 0: continue operation. This is the default behavior. | |
303 | ||
304 | 1: panic immediately. | |
305 | ||
306 | ============================================================== | |
307 | ||
308 | hung_task_check_count: | |
309 | ||
310 | The upper bound on the number of tasks that are checked. | |
311 | This file shows up if CONFIG_DETECT_HUNG_TASK is enabled. | |
312 | ||
313 | ============================================================== | |
314 | ||
315 | hung_task_timeout_secs: | |
316 | ||
317 | Check interval. When a task in D state did not get scheduled | |
318 | for more than this value report a warning. | |
319 | This file shows up if CONFIG_DETECT_HUNG_TASK is enabled. | |
320 | ||
321 | 0: means infinite timeout - no checking done. | |
80df2847 | 322 | Possible values to set are in range {0..LONG_MAX/HZ}. |
270750db AT |
323 | |
324 | ============================================================== | |
325 | ||
70e0ac5f | 326 | hung_task_warnings: |
270750db AT |
327 | |
328 | The maximum number of warnings to report. During a check interval | |
70e0ac5f AT |
329 | if a hung task is detected, this value is decreased by 1. |
330 | When this value reaches 0, no more warnings will be reported. | |
270750db AT |
331 | This file shows up if CONFIG_DETECT_HUNG_TASK is enabled. |
332 | ||
333 | -1: report an infinite number of warnings. | |
334 | ||
335 | ============================================================== | |
336 | ||
7984754b KC |
337 | kexec_load_disabled: |
338 | ||
339 | A toggle indicating if the kexec_load syscall has been disabled. This | |
340 | value defaults to 0 (false: kexec_load enabled), but can be set to 1 | |
341 | (true: kexec_load disabled). Once true, kexec can no longer be used, and | |
342 | the toggle cannot be set back to false. This allows a kexec image to be | |
343 | loaded before disabling the syscall, allowing a system to set up (and | |
344 | later use) an image without it being altered. Generally used together | |
345 | with the "modules_disabled" sysctl. | |
346 | ||
347 | ============================================================== | |
348 | ||
455cd5ab DR |
349 | kptr_restrict: |
350 | ||
351 | This toggle indicates whether restrictions are placed on | |
312b4e22 RM |
352 | exposing kernel addresses via /proc and other interfaces. |
353 | ||
354 | When kptr_restrict is set to (0), the default, there are no restrictions. | |
355 | ||
356 | When kptr_restrict is set to (1), kernel pointers printed using the %pK | |
357 | format specifier will be replaced with 0's unless the user has CAP_SYSLOG | |
358 | and effective user and group ids are equal to the real ids. This is | |
359 | because %pK checks are done at read() time rather than open() time, so | |
360 | if permissions are elevated between the open() and the read() (e.g via | |
361 | a setuid binary) then %pK will not leak kernel pointers to unprivileged | |
362 | users. Note, this is a temporary solution only. The correct long-term | |
363 | solution is to do the permission checks at open() time. Consider removing | |
364 | world read permissions from files that use %pK, and using dmesg_restrict | |
365 | to protect against uses of %pK in dmesg(8) if leaking kernel pointer | |
366 | values to unprivileged users is a concern. | |
367 | ||
368 | When kptr_restrict is set to (2), kernel pointers printed using | |
369 | %pK will be replaced with 0's regardless of privileges. | |
455cd5ab DR |
370 | |
371 | ============================================================== | |
372 | ||
0741f4d2 CE |
373 | kstack_depth_to_print: (X86 only) |
374 | ||
375 | Controls the number of words to print when dumping the raw | |
376 | kernel stack. | |
377 | ||
378 | ============================================================== | |
379 | ||
807094c0 BP |
380 | l2cr: (PPC only) |
381 | ||
382 | This flag controls the L2 cache of G3 processor boards. If | |
383 | 0, the cache is disabled. Enabled if nonzero. | |
384 | ||
385 | ============================================================== | |
386 | ||
3d43321b KC |
387 | modules_disabled: |
388 | ||
389 | A toggle value indicating if modules are allowed to be loaded | |
390 | in an otherwise modular kernel. This toggle defaults to off | |
391 | (0), but can be set true (1). Once true, modules can be | |
392 | neither loaded nor unloaded, and the toggle cannot be set back | |
7984754b | 393 | to false. Generally used with the "kexec_load_disabled" toggle. |
3d43321b KC |
394 | |
395 | ============================================================== | |
396 | ||
03f59566 SK |
397 | msg_next_id, sem_next_id, and shm_next_id: |
398 | ||
399 | These three toggles allows to specify desired id for next allocated IPC | |
400 | object: message, semaphore or shared memory respectively. | |
401 | ||
402 | By default they are equal to -1, which means generic allocation logic. | |
403 | Possible values to set are in range {0..INT_MAX}. | |
404 | ||
405 | Notes: | |
406 | 1) kernel doesn't guarantee, that new object will have desired id. So, | |
407 | it's up to userspace, how to handle an object with "wrong" id. | |
408 | 2) Toggle with non-default value will be set back to -1 by kernel after | |
409 | successful IPC object allocation. | |
410 | ||
411 | ============================================================== | |
412 | ||
807094c0 BP |
413 | nmi_watchdog: |
414 | ||
415 | Enables/Disables the NMI watchdog on x86 systems. When the value is | |
416 | non-zero the NMI watchdog is enabled and will continuously test all | |
417 | online cpus to determine whether or not they are still functioning | |
418 | properly. Currently, passing "nmi_watchdog=" parameter at boot time is | |
419 | required for this function to work. | |
420 | ||
421 | If LAPIC NMI watchdog method is in use (nmi_watchdog=2 kernel | |
422 | parameter), the NMI watchdog shares registers with oprofile. By | |
423 | disabling the NMI watchdog, oprofile may have more registers to | |
424 | utilize. | |
425 | ||
426 | ============================================================== | |
427 | ||
10fc05d0 MG |
428 | numa_balancing |
429 | ||
430 | Enables/disables automatic page fault based NUMA memory | |
431 | balancing. Memory is moved automatically to nodes | |
432 | that access it often. | |
433 | ||
434 | Enables/disables automatic NUMA memory balancing. On NUMA machines, there | |
435 | is a performance penalty if remote memory is accessed by a CPU. When this | |
436 | feature is enabled the kernel samples what task thread is accessing memory | |
437 | by periodically unmapping pages and later trapping a page fault. At the | |
438 | time of the page fault, it is determined if the data being accessed should | |
439 | be migrated to a local memory node. | |
440 | ||
441 | The unmapping of pages and trapping faults incur additional overhead that | |
442 | ideally is offset by improved memory locality but there is no universal | |
443 | guarantee. If the target workload is already bound to NUMA nodes then this | |
444 | feature should be disabled. Otherwise, if the system overhead from the | |
445 | feature is too high then the rate the kernel samples for NUMA hinting | |
446 | faults may be controlled by the numa_balancing_scan_period_min_ms, | |
930aa174 | 447 | numa_balancing_scan_delay_ms, numa_balancing_scan_period_max_ms, |
52bf84aa | 448 | numa_balancing_scan_size_mb, and numa_balancing_settle_count sysctls. |
10fc05d0 MG |
449 | |
450 | ============================================================== | |
451 | ||
452 | numa_balancing_scan_period_min_ms, numa_balancing_scan_delay_ms, | |
930aa174 | 453 | numa_balancing_scan_period_max_ms, numa_balancing_scan_size_mb |
10fc05d0 MG |
454 | |
455 | Automatic NUMA balancing scans tasks address space and unmaps pages to | |
456 | detect if pages are properly placed or if the data should be migrated to a | |
457 | memory node local to where the task is running. Every "scan delay" the task | |
458 | scans the next "scan size" number of pages in its address space. When the | |
459 | end of the address space is reached the scanner restarts from the beginning. | |
460 | ||
461 | In combination, the "scan delay" and "scan size" determine the scan rate. | |
462 | When "scan delay" decreases, the scan rate increases. The scan delay and | |
463 | hence the scan rate of every task is adaptive and depends on historical | |
464 | behaviour. If pages are properly placed then the scan delay increases, | |
465 | otherwise the scan delay decreases. The "scan size" is not adaptive but | |
466 | the higher the "scan size", the higher the scan rate. | |
467 | ||
468 | Higher scan rates incur higher system overhead as page faults must be | |
469 | trapped and potentially data must be migrated. However, the higher the scan | |
470 | rate, the more quickly a tasks memory is migrated to a local node if the | |
471 | workload pattern changes and minimises performance impact due to remote | |
472 | memory accesses. These sysctls control the thresholds for scan delays and | |
473 | the number of pages scanned. | |
474 | ||
598f0ec0 MG |
475 | numa_balancing_scan_period_min_ms is the minimum time in milliseconds to |
476 | scan a tasks virtual memory. It effectively controls the maximum scanning | |
477 | rate for each task. | |
10fc05d0 MG |
478 | |
479 | numa_balancing_scan_delay_ms is the starting "scan delay" used for a task | |
480 | when it initially forks. | |
481 | ||
598f0ec0 MG |
482 | numa_balancing_scan_period_max_ms is the maximum time in milliseconds to |
483 | scan a tasks virtual memory. It effectively controls the minimum scanning | |
484 | rate for each task. | |
10fc05d0 MG |
485 | |
486 | numa_balancing_scan_size_mb is how many megabytes worth of pages are | |
487 | scanned for a given scan. | |
488 | ||
10fc05d0 MG |
489 | ============================================================== |
490 | ||
1da177e4 LT |
491 | osrelease, ostype & version: |
492 | ||
493 | # cat osrelease | |
494 | 2.1.88 | |
495 | # cat ostype | |
496 | Linux | |
497 | # cat version | |
498 | #5 Wed Feb 25 21:49:24 MET 1998 | |
499 | ||
500 | The files osrelease and ostype should be clear enough. Version | |
501 | needs a little more clarification however. The '#5' means that | |
502 | this is the fifth kernel built from this source base and the | |
503 | date behind it indicates the time the kernel was built. | |
504 | The only way to tune these values is to rebuild the kernel :-) | |
505 | ||
506 | ============================================================== | |
507 | ||
508 | overflowgid & overflowuid: | |
509 | ||
807094c0 BP |
510 | if your architecture did not always support 32-bit UIDs (i.e. arm, |
511 | i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to | |
512 | applications that use the old 16-bit UID/GID system calls, if the | |
513 | actual UID or GID would exceed 65535. | |
1da177e4 LT |
514 | |
515 | These sysctls allow you to change the value of the fixed UID and GID. | |
516 | The default is 65534. | |
517 | ||
518 | ============================================================== | |
519 | ||
520 | panic: | |
521 | ||
807094c0 BP |
522 | The value in this file represents the number of seconds the kernel |
523 | waits before rebooting on a panic. When you use the software watchdog, | |
524 | the recommended setting is 60. | |
525 | ||
526 | ============================================================== | |
527 | ||
528 | panic_on_unrecovered_nmi: | |
529 | ||
530 | The default Linux behaviour on an NMI of either memory or unknown is | |
531 | to continue operation. For many environments such as scientific | |
532 | computing it is preferable that the box is taken out and the error | |
533 | dealt with than an uncorrected parity/ECC error get propagated. | |
534 | ||
535 | A small number of systems do generate NMI's for bizarre random reasons | |
536 | such as power management so the default is off. That sysctl works like | |
537 | the existing panic controls already in that directory. | |
1da177e4 LT |
538 | |
539 | ============================================================== | |
540 | ||
541 | panic_on_oops: | |
542 | ||
543 | Controls the kernel's behaviour when an oops or BUG is encountered. | |
544 | ||
545 | 0: try to continue operation | |
546 | ||
a982ac06 | 547 | 1: panic immediately. If the `panic' sysctl is also non-zero then the |
8b23d04d | 548 | machine will be rebooted. |
1da177e4 LT |
549 | |
550 | ============================================================== | |
551 | ||
55af7796 MH |
552 | panic_on_stackoverflow: |
553 | ||
554 | Controls the kernel's behavior when detecting the overflows of | |
555 | kernel, IRQ and exception stacks except a user stack. | |
556 | This file shows up if CONFIG_DEBUG_STACKOVERFLOW is enabled. | |
557 | ||
558 | 0: try to continue operation. | |
559 | ||
560 | 1: panic immediately. | |
561 | ||
562 | ============================================================== | |
563 | ||
14c63f17 DH |
564 | perf_cpu_time_max_percent: |
565 | ||
566 | Hints to the kernel how much CPU time it should be allowed to | |
567 | use to handle perf sampling events. If the perf subsystem | |
568 | is informed that its samples are exceeding this limit, it | |
569 | will drop its sampling frequency to attempt to reduce its CPU | |
570 | usage. | |
571 | ||
572 | Some perf sampling happens in NMIs. If these samples | |
573 | unexpectedly take too long to execute, the NMIs can become | |
574 | stacked up next to each other so much that nothing else is | |
575 | allowed to execute. | |
576 | ||
577 | 0: disable the mechanism. Do not monitor or correct perf's | |
578 | sampling rate no matter how CPU time it takes. | |
579 | ||
580 | 1-100: attempt to throttle perf's sample rate to this | |
581 | percentage of CPU. Note: the kernel calculates an | |
582 | "expected" length of each sample event. 100 here means | |
583 | 100% of that expected length. Even if this is set to | |
584 | 100, you may still see sample throttling if this | |
585 | length is exceeded. Set to 0 if you truly do not care | |
586 | how much CPU is consumed. | |
587 | ||
588 | ============================================================== | |
589 | ||
55af7796 | 590 | |
1da177e4 LT |
591 | pid_max: |
592 | ||
beb7dd86 | 593 | PID allocation wrap value. When the kernel's next PID value |
1da177e4 LT |
594 | reaches this value, it wraps back to a minimum PID value. |
595 | PIDs of value pid_max or larger are not allocated. | |
596 | ||
597 | ============================================================== | |
598 | ||
b8f566b0 PE |
599 | ns_last_pid: |
600 | ||
601 | The last pid allocated in the current (the one task using this sysctl | |
602 | lives in) pid namespace. When selecting a pid for a next task on fork | |
603 | kernel tries to allocate a number starting from this one. | |
604 | ||
605 | ============================================================== | |
606 | ||
1da177e4 LT |
607 | powersave-nap: (PPC only) |
608 | ||
609 | If set, Linux-PPC will use the 'nap' mode of powersaving, | |
610 | otherwise the 'doze' mode will be used. | |
611 | ||
612 | ============================================================== | |
613 | ||
614 | printk: | |
615 | ||
616 | The four values in printk denote: console_loglevel, | |
617 | default_message_loglevel, minimum_console_loglevel and | |
618 | default_console_loglevel respectively. | |
619 | ||
620 | These values influence printk() behavior when printing or | |
621 | logging error messages. See 'man 2 syslog' for more info on | |
622 | the different loglevels. | |
623 | ||
624 | - console_loglevel: messages with a higher priority than | |
625 | this will be printed to the console | |
87889e15 | 626 | - default_message_loglevel: messages without an explicit priority |
1da177e4 LT |
627 | will be printed with this priority |
628 | - minimum_console_loglevel: minimum (highest) value to which | |
629 | console_loglevel can be set | |
630 | - default_console_loglevel: default value for console_loglevel | |
631 | ||
632 | ============================================================== | |
633 | ||
807094c0 BP |
634 | printk_delay: |
635 | ||
636 | Delay each printk message in printk_delay milliseconds | |
637 | ||
638 | Value from 0 - 10000 is allowed. | |
639 | ||
640 | ============================================================== | |
641 | ||
1da177e4 LT |
642 | printk_ratelimit: |
643 | ||
644 | Some warning messages are rate limited. printk_ratelimit specifies | |
645 | the minimum length of time between these messages (in jiffies), by | |
646 | default we allow one every 5 seconds. | |
647 | ||
648 | A value of 0 will disable rate limiting. | |
649 | ||
650 | ============================================================== | |
651 | ||
652 | printk_ratelimit_burst: | |
653 | ||
654 | While long term we enforce one message per printk_ratelimit | |
655 | seconds, we do allow a burst of messages to pass through. | |
656 | printk_ratelimit_burst specifies the number of messages we can | |
657 | send before ratelimiting kicks in. | |
658 | ||
659 | ============================================================== | |
660 | ||
807094c0 | 661 | randomize_va_space: |
1ec7fd50 JK |
662 | |
663 | This option can be used to select the type of process address | |
664 | space randomization that is used in the system, for architectures | |
665 | that support this feature. | |
666 | ||
b7f5ab6f HS |
667 | 0 - Turn the process address space randomization off. This is the |
668 | default for architectures that do not support this feature anyways, | |
669 | and kernels that are booted with the "norandmaps" parameter. | |
1ec7fd50 JK |
670 | |
671 | 1 - Make the addresses of mmap base, stack and VDSO page randomized. | |
672 | This, among other things, implies that shared libraries will be | |
b7f5ab6f HS |
673 | loaded to random addresses. Also for PIE-linked binaries, the |
674 | location of code start is randomized. This is the default if the | |
675 | CONFIG_COMPAT_BRK option is enabled. | |
1ec7fd50 | 676 | |
b7f5ab6f HS |
677 | 2 - Additionally enable heap randomization. This is the default if |
678 | CONFIG_COMPAT_BRK is disabled. | |
679 | ||
680 | There are a few legacy applications out there (such as some ancient | |
1ec7fd50 | 681 | versions of libc.so.5 from 1996) that assume that brk area starts |
b7f5ab6f HS |
682 | just after the end of the code+bss. These applications break when |
683 | start of the brk area is randomized. There are however no known | |
1ec7fd50 | 684 | non-legacy applications that would be broken this way, so for most |
b7f5ab6f HS |
685 | systems it is safe to choose full randomization. |
686 | ||
687 | Systems with ancient and/or broken binaries should be configured | |
688 | with CONFIG_COMPAT_BRK enabled, which excludes the heap from process | |
689 | address space randomization. | |
1ec7fd50 JK |
690 | |
691 | ============================================================== | |
692 | ||
1da177e4 LT |
693 | reboot-cmd: (Sparc only) |
694 | ||
695 | ??? This seems to be a way to give an argument to the Sparc | |
696 | ROM/Flash boot loader. Maybe to tell it what to do after | |
697 | rebooting. ??? | |
698 | ||
699 | ============================================================== | |
700 | ||
701 | rtsig-max & rtsig-nr: | |
702 | ||
703 | The file rtsig-max can be used to tune the maximum number | |
704 | of POSIX realtime (queued) signals that can be outstanding | |
705 | in the system. | |
706 | ||
707 | rtsig-nr shows the number of RT signals currently queued. | |
708 | ||
709 | ============================================================== | |
710 | ||
711 | sg-big-buff: | |
712 | ||
713 | This file shows the size of the generic SCSI (sg) buffer. | |
714 | You can't tune it just yet, but you could change it on | |
715 | compile time by editing include/scsi/sg.h and changing | |
716 | the value of SG_BIG_BUFF. | |
717 | ||
718 | There shouldn't be any reason to change this value. If | |
719 | you can come up with one, you probably know what you | |
720 | are doing anyway :) | |
721 | ||
722 | ============================================================== | |
723 | ||
358e419f CALP |
724 | shmall: |
725 | ||
726 | This parameter sets the total amount of shared memory pages that | |
727 | can be used system wide. Hence, SHMALL should always be at least | |
728 | ceil(shmmax/PAGE_SIZE). | |
729 | ||
730 | If you are not sure what the default PAGE_SIZE is on your Linux | |
731 | system, you can run the following command: | |
732 | ||
733 | # getconf PAGE_SIZE | |
734 | ||
735 | ============================================================== | |
736 | ||
807094c0 | 737 | shmmax: |
1da177e4 LT |
738 | |
739 | This value can be used to query and set the run time limit | |
740 | on the maximum shared memory segment size that can be created. | |
807094c0 | 741 | Shared memory segments up to 1Gb are now supported in the |
1da177e4 LT |
742 | kernel. This value defaults to SHMMAX. |
743 | ||
744 | ============================================================== | |
745 | ||
b34a6b1d VK |
746 | shm_rmid_forced: |
747 | ||
748 | Linux lets you set resource limits, including how much memory one | |
749 | process can consume, via setrlimit(2). Unfortunately, shared memory | |
750 | segments are allowed to exist without association with any process, and | |
751 | thus might not be counted against any resource limits. If enabled, | |
752 | shared memory segments are automatically destroyed when their attach | |
753 | count becomes zero after a detach or a process termination. It will | |
754 | also destroy segments that were created, but never attached to, on exit | |
755 | from the process. The only use left for IPC_RMID is to immediately | |
756 | destroy an unattached segment. Of course, this breaks the way things are | |
757 | defined, so some applications might stop working. Note that this | |
758 | feature will do you no good unless you also configure your resource | |
759 | limits (in particular, RLIMIT_AS and RLIMIT_NPROC). Most systems don't | |
760 | need this. | |
761 | ||
762 | Note that if you change this from 0 to 1, already created segments | |
763 | without users and with a dead originative process will be destroyed. | |
764 | ||
765 | ============================================================== | |
766 | ||
f4aacea2 KC |
767 | sysctl_writes_strict: |
768 | ||
769 | Control how file position affects the behavior of updating sysctl values | |
770 | via the /proc/sys interface: | |
771 | ||
772 | -1 - Legacy per-write sysctl value handling, with no printk warnings. | |
773 | Each write syscall must fully contain the sysctl value to be | |
774 | written, and multiple writes on the same sysctl file descriptor | |
775 | will rewrite the sysctl value, regardless of file position. | |
776 | 0 - (default) Same behavior as above, but warn about processes that | |
777 | perform writes to a sysctl file descriptor when the file position | |
778 | is not 0. | |
779 | 1 - Respect file position when writing sysctl strings. Multiple writes | |
780 | will append to the sysctl value buffer. Anything past the max length | |
781 | of the sysctl value buffer will be ignored. Writes to numeric sysctl | |
782 | entries must always be at file position 0 and the value must be | |
783 | fully contained in the buffer sent in the write syscall. | |
784 | ||
785 | ============================================================== | |
786 | ||
ed235875 AT |
787 | softlockup_all_cpu_backtrace: |
788 | ||
789 | This value controls the soft lockup detector thread's behavior | |
790 | when a soft lockup condition is detected as to whether or not | |
791 | to gather further debug information. If enabled, each cpu will | |
792 | be issued an NMI and instructed to capture stack trace. | |
793 | ||
794 | This feature is only applicable for architectures which support | |
795 | NMI. | |
796 | ||
797 | 0: do nothing. This is the default behavior. | |
798 | ||
799 | 1: on detection capture more debug information. | |
800 | ||
801 | ============================================================== | |
802 | ||
807094c0 | 803 | tainted: |
1da177e4 LT |
804 | |
805 | Non-zero if the kernel has been tainted. Numeric values, which | |
806 | can be ORed together: | |
807 | ||
bb20698d GKH |
808 | 1 - A module with a non-GPL license has been loaded, this |
809 | includes modules with no license. | |
810 | Set by modutils >= 2.4.9 and module-init-tools. | |
811 | 2 - A module was force loaded by insmod -f. | |
812 | Set by modutils >= 2.4.9 and module-init-tools. | |
813 | 4 - Unsafe SMP processors: SMP with CPUs not designed for SMP. | |
814 | 8 - A module was forcibly unloaded from the system by rmmod -f. | |
815 | 16 - A hardware machine check error occurred on the system. | |
816 | 32 - A bad page was discovered on the system. | |
817 | 64 - The user has asked that the system be marked "tainted". This | |
818 | could be because they are running software that directly modifies | |
819 | the hardware, or for other reasons. | |
820 | 128 - The system has died. | |
821 | 256 - The ACPI DSDT has been overridden with one supplied by the user | |
822 | instead of using the one provided by the hardware. | |
823 | 512 - A kernel warning has occurred. | |
824 | 1024 - A module from drivers/staging was loaded. | |
f5fe184b LF |
825 | 2048 - The system is working around a severe firmware bug. |
826 | 4096 - An out-of-tree module has been loaded. | |
66cc69e3 MD |
827 | 8192 - An unsigned module has been loaded in a kernel supporting module |
828 | signature. | |
1da177e4 | 829 | |
760df93e SF |
830 | ============================================================== |
831 | ||
760df93e SF |
832 | unknown_nmi_panic: |
833 | ||
807094c0 BP |
834 | The value in this file affects behavior of handling NMI. When the |
835 | value is non-zero, unknown NMI is trapped and then panic occurs. At | |
836 | that time, kernel debugging information is displayed on console. | |
760df93e | 837 | |
807094c0 BP |
838 | NMI switch that most IA32 servers have fires unknown NMI up, for |
839 | example. If a system hangs up, try pressing the NMI switch. | |
08825c90 LZ |
840 | |
841 | ============================================================== | |
842 | ||
843 | watchdog_thresh: | |
844 | ||
845 | This value can be used to control the frequency of hrtimer and NMI | |
846 | events and the soft and hard lockup thresholds. The default threshold | |
847 | is 10 seconds. | |
848 | ||
849 | The softlockup threshold is (2 * watchdog_thresh). Setting this | |
850 | tunable to zero will disable lockup detection altogether. | |
851 | ||
852 | ============================================================== |