1Documentation for /proc/sys/kernel/* kernel version 2.2.10 2 (c) 1998, 1999, Rik van Riel <riel@nl.linux.org> 3 4For general info and legal blurb, please look in README. 5 6============================================================== 7 8This file contains documentation for the sysctl files in 9/proc/sys/kernel/ and is valid for Linux kernel version 2.2. 10 11The files in this directory can be used to tune and monitor 12miscellaneous and general things in the operation of the Linux 13kernel. Since some of the files _can_ be used to screw up your 14system, it is advisable to read both documentation and source 15before actually making adjustments. 16 17Currently, these files might (depending on your configuration) 18show up in /proc/sys/kernel: 19- acpi_video_flags 20- acct 21- core_pattern 22- core_uses_pid 23- ctrl-alt-del 24- dentry-state 25- domainname 26- hostname 27- hotplug 28- java-appletviewer [ binfmt_java, obsolete ] 29- java-interpreter [ binfmt_java, obsolete ] 30- kstack_depth_to_print [ X86 only ] 31- l2cr [ PPC only ] 32- modprobe ==> Documentation/debugging-modules.txt 33- msgmax 34- msgmnb 35- msgmni 36- osrelease 37- ostype 38- overflowgid 39- overflowuid 40- panic 41- pid_max 42- powersave-nap [ PPC only ] 43- printk 44- randomize_va_space 45- real-root-dev ==> Documentation/initrd.txt 46- reboot-cmd [ SPARC only ] 47- rtsig-max 48- rtsig-nr 49- sem 50- sg-big-buff [ generic SCSI device (sg) ] 51- shmall 52- shmmax [ sysv ipc ] 53- shmmni 54- stop-a [ SPARC only ] 55- sysrq ==> Documentation/sysrq.txt 56- tainted 57- threads-max 58- version 59 60============================================================== 61 62acpi_video_flags: 63 64flags 65 66See Doc*/kernel/power/video.txt, it allows mode of video boot to be 67set during run time. 68 69============================================================== 70 71acct: 72 73highwater lowwater frequency 74 75If BSD-style process accounting is enabled these values control 76its behaviour. If free space on filesystem where the log lives 77goes below <lowwater>% accounting suspends. If free space gets 78above <highwater>% accounting resumes. <Frequency> determines 79how often do we check the amount of free space (value is in 80seconds). Default: 814 2 30 82That is, suspend accounting if there left <= 2% free; resume it 83if we got >=4%; consider information about amount of free space 84valid for 30 seconds. 85 86============================================================== 87 88core_pattern: 89 90core_pattern is used to specify a core dumpfile pattern name. 91. max length 128 characters; default value is "core" 92. core_pattern is used as a pattern template for the output filename; 93 certain string patterns (beginning with '%') are substituted with 94 their actual values. 95. backward compatibility with core_uses_pid: 96 If core_pattern does not include "%p" (default does not) 97 and core_uses_pid is set, then .PID will be appended to 98 the filename. 99. corename format specifiers: 100 %<NUL> '%' is dropped 101 %% output one '%' 102 %p pid 103 %u uid 104 %g gid 105 %s signal number 106 %t UNIX time of dump 107 %h hostname 108 %e executable filename 109 %<OTHER> both are dropped 110. If the first character of the pattern is a '|', the kernel will treat 111 the rest of the pattern as a command to run. The core dump will be 112 written to the standard input of that program instead of to a file. 113 114============================================================== 115 116core_uses_pid: 117 118The default coredump filename is "core". By setting 119core_uses_pid to 1, the coredump filename becomes core.PID. 120If core_pattern does not include "%p" (default does not) 121and core_uses_pid is set, then .PID will be appended to 122the filename. 123 124============================================================== 125 126ctrl-alt-del: 127 128When the value in this file is 0, ctrl-alt-del is trapped and 129sent to the init(1) program to handle a graceful restart. 130When, however, the value is > 0, Linux's reaction to a Vulcan 131Nerve Pinch (tm) will be an immediate reboot, without even 132syncing its dirty buffers. 133 134Note: when a program (like dosemu) has the keyboard in 'raw' 135mode, the ctrl-alt-del is intercepted by the program before it 136ever reaches the kernel tty layer, and it's up to the program 137to decide what to do with it. 138 139============================================================== 140 141domainname & hostname: 142 143These files can be used to set the NIS/YP domainname and the 144hostname of your box in exactly the same way as the commands 145domainname and hostname, i.e.: 146# echo "darkstar" > /proc/sys/kernel/hostname 147# echo "mydomain" > /proc/sys/kernel/domainname 148has the same effect as 149# hostname "darkstar" 150# domainname "mydomain" 151 152Note, however, that the classic darkstar.frop.org has the 153hostname "darkstar" and DNS (Internet Domain Name Server) 154domainname "frop.org", not to be confused with the NIS (Network 155Information Service) or YP (Yellow Pages) domainname. These two 156domain names are in general different. For a detailed discussion 157see the hostname(1) man page. 158 159============================================================== 160 161hotplug: 162 163Path for the hotplug policy agent. 164Default value is "/sbin/hotplug". 165 166============================================================== 167 168l2cr: (PPC only) 169 170This flag controls the L2 cache of G3 processor boards. If 1710, the cache is disabled. Enabled if nonzero. 172 173============================================================== 174 175kstack_depth_to_print: (X86 only) 176 177Controls the number of words to print when dumping the raw 178kernel stack. 179 180============================================================== 181 182osrelease, ostype & version: 183 184# cat osrelease 1852.1.88 186# cat ostype 187Linux 188# cat version 189#5 Wed Feb 25 21:49:24 MET 1998 190 191The files osrelease and ostype should be clear enough. Version 192needs a little more clarification however. The '#5' means that 193this is the fifth kernel built from this source base and the 194date behind it indicates the time the kernel was built. 195The only way to tune these values is to rebuild the kernel :-) 196 197============================================================== 198 199overflowgid & overflowuid: 200 201if your architecture did not always support 32-bit UIDs (i.e. arm, i386, 202m68k, sh, and sparc32), a fixed UID and GID will be returned to 203applications that use the old 16-bit UID/GID system calls, if the actual 204UID or GID would exceed 65535. 205 206These sysctls allow you to change the value of the fixed UID and GID. 207The default is 65534. 208 209============================================================== 210 211panic: 212 213The value in this file represents the number of seconds the 214kernel waits before rebooting on a panic. When you use the 215software watchdog, the recommended setting is 60. 216 217============================================================== 218 219panic_on_oops: 220 221Controls the kernel's behaviour when an oops or BUG is encountered. 222 2230: try to continue operation 224 2251: panic immediately. If the `panic' sysctl is also non-zero then the 226 machine will be rebooted. 227 228============================================================== 229 230pid_max: 231 232PID allocation wrap value. When the kernel's next PID value 233reaches this value, it wraps back to a minimum PID value. 234PIDs of value pid_max or larger are not allocated. 235 236============================================================== 237 238powersave-nap: (PPC only) 239 240If set, Linux-PPC will use the 'nap' mode of powersaving, 241otherwise the 'doze' mode will be used. 242 243============================================================== 244 245printk: 246 247The four values in printk denote: console_loglevel, 248default_message_loglevel, minimum_console_loglevel and 249default_console_loglevel respectively. 250 251These values influence printk() behavior when printing or 252logging error messages. See 'man 2 syslog' for more info on 253the different loglevels. 254 255- console_loglevel: messages with a higher priority than 256 this will be printed to the console 257- default_message_level: messages without an explicit priority 258 will be printed with this priority 259- minimum_console_loglevel: minimum (highest) value to which 260 console_loglevel can be set 261- default_console_loglevel: default value for console_loglevel 262 263============================================================== 264 265printk_ratelimit: 266 267Some warning messages are rate limited. printk_ratelimit specifies 268the minimum length of time between these messages (in jiffies), by 269default we allow one every 5 seconds. 270 271A value of 0 will disable rate limiting. 272 273============================================================== 274 275printk_ratelimit_burst: 276 277While long term we enforce one message per printk_ratelimit 278seconds, we do allow a burst of messages to pass through. 279printk_ratelimit_burst specifies the number of messages we can 280send before ratelimiting kicks in. 281 282============================================================== 283 284randomize-va-space: 285 286This option can be used to select the type of process address 287space randomization that is used in the system, for architectures 288that support this feature. 289 2900 - Turn the process address space randomization off by default. 291 2921 - Make the addresses of mmap base, stack and VDSO page randomized. 293 This, among other things, implies that shared libraries will be 294 loaded to random addresses. Also for PIE-linked binaries, the location 295 of code start is randomized. 296 297 With heap randomization, the situation is a little bit more 298 complicated. 299 There a few legacy applications out there (such as some ancient 300 versions of libc.so.5 from 1996) that assume that brk area starts 301 just after the end of the code+bss. These applications break when 302 start of the brk area is randomized. There are however no known 303 non-legacy applications that would be broken this way, so for most 304 systems it is safe to choose full randomization. However there is 305 a CONFIG_COMPAT_BRK option for systems with ancient and/or broken 306 binaries, that makes heap non-randomized, but keeps all other 307 parts of process address space randomized if randomize_va_space 308 sysctl is turned on. 309 310============================================================== 311 312reboot-cmd: (Sparc only) 313 314??? This seems to be a way to give an argument to the Sparc 315ROM/Flash boot loader. Maybe to tell it what to do after 316rebooting. ??? 317 318============================================================== 319 320rtsig-max & rtsig-nr: 321 322The file rtsig-max can be used to tune the maximum number 323of POSIX realtime (queued) signals that can be outstanding 324in the system. 325 326rtsig-nr shows the number of RT signals currently queued. 327 328============================================================== 329 330sg-big-buff: 331 332This file shows the size of the generic SCSI (sg) buffer. 333You can't tune it just yet, but you could change it on 334compile time by editing include/scsi/sg.h and changing 335the value of SG_BIG_BUFF. 336 337There shouldn't be any reason to change this value. If 338you can come up with one, you probably know what you 339are doing anyway :) 340 341============================================================== 342 343shmmax: 344 345This value can be used to query and set the run time limit 346on the maximum shared memory segment size that can be created. 347Shared memory segments up to 1Gb are now supported in the 348kernel. This value defaults to SHMMAX. 349 350============================================================== 351 352softlockup_thresh: 353 354This value can be used to lower the softlockup tolerance threshold. The 355default threshold is 60 seconds. If a cpu is locked up for 60 seconds, 356the kernel complains. Valid values are 1-60 seconds. Setting this 357tunable to zero will disable the softlockup detection altogether. 358 359============================================================== 360 361tainted: 362 363Non-zero if the kernel has been tainted. Numeric values, which 364can be ORed together: 365 366 1 - A module with a non-GPL license has been loaded, this 367 includes modules with no license. 368 Set by modutils >= 2.4.9 and module-init-tools. 369 2 - A module was force loaded by insmod -f. 370 Set by modutils >= 2.4.9 and module-init-tools. 371 4 - Unsafe SMP processors: SMP with CPUs not designed for SMP. 372 8 - A module was forcibly unloaded from the system by rmmod -f. 373 16 - A hardware machine check error occurred on the system. 374 32 - A bad page was discovered on the system. 375 64 - The user has asked that the system be marked "tainted". This 376 could be because they are running software that directly modifies 377 the hardware, or for other reasons. 378 128 - The system has died. 379 256 - The ACPI DSDT has been overridden with one supplied by the user 380 instead of using the one provided by the hardware. 381 512 - A kernel warning has occurred. 3821024 - A module from drivers/staging was loaded. 383 384