| 1 | ===================================
|
| 2 | Documentation for /proc/sys/kernel/
|
| 3 | ===================================
|
| 4 |
|
| 5 | |
| 6 |
|
| 7 |
|
| 8 | Copyright (c) 1998, 1999, Rik van Riel <[email protected]>
|
| 9 |
|
| 10 | Copyright (c) 2009, Shen Feng<[email protected]>
|
| 11 |
|
| 12 | For general info and legal blurb, please look in
|
| 13 | Documentation/admin-guide/sysctl/index.rst.
|
| 14 |
|
| 15 | ------------------------------------------------------------------------------
|
| 16 |
|
| 17 | This file contains documentation for the sysctl files in
|
| 18 | ``/proc/sys/kernel/``.
|
| 19 |
|
| 20 | The files in this directory can be used to tune and monitor
|
| 21 | miscellaneous and general things in the operation of the Linux
|
| 22 | kernel. Since some of the files *can* be used to screw up your
|
| 23 | system, it is advisable to read both documentation and source
|
| 24 | before actually making adjustments.
|
| 25 |
|
| 26 | Currently, these files might (depending on your configuration)
|
| 27 | show up in ``/proc/sys/kernel``:
|
| 28 |
|
| 29 | .. contents:: :local:
|
| 30 |
|
| 31 |
|
| 32 | acct
|
| 33 | ====
|
| 34 |
|
| 35 | ::
|
| 36 |
|
| 37 | highwater lowwater frequency
|
| 38 |
|
| 39 | If BSD-style process accounting is enabled these values control
|
| 40 | its behaviour. If free space on filesystem where the log lives
|
| 41 | goes below ``lowwater``\ % accounting suspends. If free space gets
|
| 42 | above ``highwater``\ % accounting resumes. ``frequency`` determines
|
| 43 | how often do we check the amount of free space (value is in
|
| 44 | seconds). Default:
|
| 45 |
|
| 46 | ::
|
| 47 |
|
| 48 | 4 2 30
|
| 49 |
|
| 50 | That is, suspend accounting if free space drops below 2%; resume it
|
| 51 | if it increases to at least 4%; consider information about amount of
|
| 52 | free space valid for 30 seconds.
|
| 53 |
|
| 54 |
|
| 55 | acpi_video_flags
|
| 56 | ================
|
| 57 |
|
| 58 | See Documentation/power/video.rst. This allows the video resume mode to be set,
|
| 59 | in a similar fashion to the ``acpi_sleep`` kernel parameter, by
|
| 60 | combining the following values:
|
| 61 |
|
| 62 | = =======
|
| 63 | 1 s3_bios
|
| 64 | 2 s3_mode
|
| 65 | 4 s3_beep
|
| 66 | = =======
|
| 67 |
|
| 68 | arch
|
| 69 | ====
|
| 70 |
|
| 71 | The machine hardware name, the same output as ``uname -m``
|
| 72 | (e.g. ``x86_64`` or ``aarch64``).
|
| 73 |
|
| 74 | auto_msgmni
|
| 75 | ===========
|
| 76 |
|
| 77 | This variable has no effect and may be removed in future kernel
|
| 78 | releases. Reading it always returns 0.
|
| 79 | Up to Linux 3.17, it enabled/disabled automatic recomputing of
|
| 80 | `msgmni`_
|
| 81 | upon memory add/remove or upon IPC namespace creation/removal.
|
| 82 | Echoing "1" into this file enabled msgmni automatic recomputing.
|
| 83 | Echoing "0" turned it off. The default value was 1.
|
| 84 |
|
| 85 |
|
| 86 | bootloader_type (x86 only)
|
| 87 | ==========================
|
| 88 |
|
| 89 | This gives the bootloader type number as indicated by the bootloader,
|
| 90 | shifted left by 4, and OR'd with the low four bits of the bootloader
|
| 91 | version. The reason for this encoding is that this used to match the
|
| 92 | ``type_of_loader`` field in the kernel header; the encoding is kept for
|
| 93 | backwards compatibility. That is, if the full bootloader type number
|
| 94 | is 0x15 and the full version number is 0x234, this file will contain
|
| 95 | the value 340 = 0x154.
|
| 96 |
|
| 97 | See the ``type_of_loader`` and ``ext_loader_type`` fields in
|
| 98 | Documentation/arch/x86/boot.rst for additional information.
|
| 99 |
|
| 100 |
|
| 101 | bootloader_version (x86 only)
|
| 102 | =============================
|
| 103 |
|
| 104 | The complete bootloader version number. In the example above, this
|
| 105 | file will contain the value 564 = 0x234.
|
| 106 |
|
| 107 | See the ``type_of_loader`` and ``ext_loader_ver`` fields in
|
| 108 | Documentation/arch/x86/boot.rst for additional information.
|
| 109 |
|
| 110 |
|
| 111 | bpf_stats_enabled
|
| 112 | =================
|
| 113 |
|
| 114 | Controls whether the kernel should collect statistics on BPF programs
|
| 115 | (total time spent running, number of times run...). Enabling
|
| 116 | statistics causes a slight reduction in performance on each program
|
| 117 | run. The statistics can be seen using ``bpftool``.
|
| 118 |
|
| 119 | = ===================================
|
| 120 | 0 Don't collect statistics (default).
|
| 121 | 1 Collect statistics.
|
| 122 | = ===================================
|
| 123 |
|
| 124 |
|
| 125 | cad_pid
|
| 126 | =======
|
| 127 |
|
| 128 | This is the pid which will be signalled on reboot (notably, by
|
| 129 | Ctrl-Alt-Delete). Writing a value to this file which doesn't
|
| 130 | correspond to a running process will result in ``-ESRCH``.
|
| 131 |
|
| 132 | See also `ctrl-alt-del`_.
|
| 133 |
|
| 134 |
|
| 135 | cap_last_cap
|
| 136 | ============
|
| 137 |
|
| 138 | Highest valid capability of the running kernel. Exports
|
| 139 | ``CAP_LAST_CAP`` from the kernel.
|
| 140 |
|
| 141 |
|
| 142 | .. _core_pattern:
|
| 143 |
|
| 144 | core_pattern
|
| 145 | ============
|
| 146 |
|
| 147 | ``core_pattern`` is used to specify a core dumpfile pattern name.
|
| 148 |
|
| 149 | * max length 127 characters; default value is "core"
|
| 150 | * ``core_pattern`` is used as a pattern template for the output
|
| 151 | filename; certain string patterns (beginning with '%') are
|
| 152 | substituted with their actual values.
|
| 153 | * backward compatibility with ``core_uses_pid``:
|
| 154 |
|
| 155 | If ``core_pattern`` does not include "%p" (default does not)
|
| 156 | and ``core_uses_pid`` is set, then .PID will be appended to
|
| 157 | the filename.
|
| 158 |
|
| 159 | * corename format specifiers
|
| 160 |
|
| 161 | ======== ==========================================
|
| 162 | %<NUL> '%' is dropped
|
| 163 | %% output one '%'
|
| 164 | %p pid
|
| 165 | %P global pid (init PID namespace)
|
| 166 | %i tid
|
| 167 | %I global tid (init PID namespace)
|
| 168 | %u uid (in initial user namespace)
|
| 169 | %g gid (in initial user namespace)
|
| 170 | %d dump mode, matches ``PR_SET_DUMPABLE`` and
|
| 171 | ``/proc/sys/fs/suid_dumpable``
|
| 172 | %s signal number
|
| 173 | %t UNIX time of dump
|
| 174 | %h hostname
|
| 175 | %e executable filename (may be shortened, could be changed by prctl etc)
|
| 176 | %f executable filename
|
| 177 | %E executable path
|
| 178 | %c maximum size of core file by resource limit RLIMIT_CORE
|
| 179 | %C CPU the task ran on
|
| 180 | %F pidfd number
|
| 181 | %<OTHER> both are dropped
|
| 182 | ======== ==========================================
|
| 183 |
|
| 184 | * If the first character of the pattern is a '|', the kernel will treat
|
| 185 | the rest of the pattern as a command to run. The core dump will be
|
| 186 | written to the standard input of that program instead of to a file.
|
| 187 |
|
| 188 |
|
| 189 | core_pipe_limit
|
| 190 | ===============
|
| 191 |
|
| 192 | This sysctl is only applicable when `core_pattern`_ is configured to
|
| 193 | pipe core files to a user space helper (when the first character of
|
| 194 | ``core_pattern`` is a '|', see above).
|
| 195 | When collecting cores via a pipe to an application, it is occasionally
|
| 196 | useful for the collecting application to gather data about the
|
| 197 | crashing process from its ``/proc/pid`` directory.
|
| 198 | In order to do this safely, the kernel must wait for the collecting
|
| 199 | process to exit, so as not to remove the crashing processes proc files
|
| 200 | prematurely.
|
| 201 | This in turn creates the possibility that a misbehaving userspace
|
| 202 | collecting process can block the reaping of a crashed process simply
|
| 203 | by never exiting.
|
| 204 | This sysctl defends against that.
|
| 205 | It defines how many concurrent crashing processes may be piped to user
|
| 206 | space applications in parallel.
|
| 207 | If this value is exceeded, then those crashing processes above that
|
| 208 | value are noted via the kernel log and their cores are skipped.
|
| 209 | 0 is a special value, indicating that unlimited processes may be
|
| 210 | captured in parallel, but that no waiting will take place (i.e. the
|
| 211 | collecting process is not guaranteed access to ``/proc/<crashing
|
| 212 | pid>/``).
|
| 213 | This value defaults to 0.
|
| 214 |
|
| 215 |
|
| 216 | core_sort_vma
|
| 217 | =============
|
| 218 |
|
| 219 | The default coredump writes VMAs in address order. By setting
|
| 220 | ``core_sort_vma`` to 1, VMAs will be written from smallest size
|
| 221 | to largest size. This is known to break at least elfutils, but
|
| 222 | can be handy when dealing with very large (and truncated)
|
| 223 | coredumps where the more useful debugging details are included
|
| 224 | in the smaller VMAs.
|
| 225 |
|
| 226 |
|
| 227 | core_uses_pid
|
| 228 | =============
|
| 229 |
|
| 230 | The default coredump filename is "core". By setting
|
| 231 | ``core_uses_pid`` to 1, the coredump filename becomes core.PID.
|
| 232 | If `core_pattern`_ does not include "%p" (default does not)
|
| 233 | and ``core_uses_pid`` is set, then .PID will be appended to
|
| 234 | the filename.
|
| 235 |
|
| 236 |
|
| 237 | ctrl-alt-del
|
| 238 | ============
|
| 239 |
|
| 240 | When the value in this file is 0, ctrl-alt-del is trapped and
|
| 241 | sent to the ``init(1)`` program to handle a graceful restart.
|
| 242 | When, however, the value is > 0, Linux's reaction to a Vulcan
|
| 243 | Nerve Pinch (tm) will be an immediate reboot, without even
|
| 244 | syncing its dirty buffers.
|
| 245 |
|
| 246 | Note:
|
| 247 | when a program (like dosemu) has the keyboard in 'raw'
|
| 248 | mode, the ctrl-alt-del is intercepted by the program before it
|
| 249 | ever reaches the kernel tty layer, and it's up to the program
|
| 250 | to decide what to do with it.
|
| 251 |
|
| 252 |
|
| 253 | dmesg_restrict
|
| 254 | ==============
|
| 255 |
|
| 256 | This toggle indicates whether unprivileged users are prevented
|
| 257 | from using ``dmesg(8)`` to view messages from the kernel's log
|
| 258 | buffer.
|
| 259 | When ``dmesg_restrict`` is set to 0 there are no restrictions.
|
| 260 | When ``dmesg_restrict`` is set to 1, users must have
|
| 261 | ``CAP_SYSLOG`` to use ``dmesg(8)``.
|
| 262 |
|
| 263 | The kernel config option ``CONFIG_SECURITY_DMESG_RESTRICT`` sets the
|
| 264 | default value of ``dmesg_restrict``.
|
| 265 |
|
| 266 |
|
| 267 | domainname & hostname
|
| 268 | =====================
|
| 269 |
|
| 270 | These files can be used to set the NIS/YP domainname and the
|
| 271 | hostname of your box in exactly the same way as the commands
|
| 272 | domainname and hostname, i.e.::
|
| 273 |
|
| 274 | # echo "darkstar" > /proc/sys/kernel/hostname
|
| 275 | # echo "mydomain" > /proc/sys/kernel/domainname
|
| 276 |
|
| 277 | has the same effect as::
|
| 278 |
|
| 279 | # hostname "darkstar"
|
| 280 | # domainname "mydomain"
|
| 281 |
|
| 282 | Note, however, that the classic darkstar.frop.org has the
|
| 283 | hostname "darkstar" and DNS (Internet Domain Name Server)
|
| 284 | domainname "frop.org", not to be confused with the NIS (Network
|
| 285 | Information Service) or YP (Yellow Pages) domainname. These two
|
| 286 | domain names are in general different. For a detailed discussion
|
| 287 | see the ``hostname(1)`` man page.
|
| 288 |
|
| 289 |
|
| 290 | firmware_config
|
| 291 | ===============
|
| 292 |
|
| 293 | See Documentation/driver-api/firmware/fallback-mechanisms.rst.
|
| 294 |
|
| 295 | The entries in this directory allow the firmware loader helper
|
| 296 | fallback to be controlled:
|
| 297 |
|
| 298 | * ``force_sysfs_fallback``, when set to 1, forces the use of the
|
| 299 | fallback;
|
| 300 | * ``ignore_sysfs_fallback``, when set to 1, ignores any fallback.
|
| 301 |
|
| 302 |
|
| 303 | ftrace_dump_on_oops
|
| 304 | ===================
|
| 305 |
|
| 306 | Determines whether ``ftrace_dump()`` should be called on an oops (or
|
| 307 | kernel panic). This will output the contents of the ftrace buffers to
|
| 308 | the console. This is very useful for capturing traces that lead to
|
| 309 | crashes and outputting them to a serial console.
|
| 310 |
|
| 311 | ======================= ===========================================
|
| 312 | 0 Disabled (default).
|
| 313 | 1 Dump buffers of all CPUs.
|
| 314 | 2(orig_cpu) Dump the buffer of the CPU that triggered the
|
| 315 | oops.
|
| 316 | <instance> Dump the specific instance buffer on all CPUs.
|
| 317 | <instance>=2(orig_cpu) Dump the specific instance buffer on the CPU
|
| 318 | that triggered the oops.
|
| 319 | ======================= ===========================================
|
| 320 |
|
| 321 | Multiple instance dump is also supported, and instances are separated
|
| 322 | by commas. If global buffer also needs to be dumped, please specify
|
| 323 | the dump mode (1/2/orig_cpu) first for global buffer.
|
| 324 |
|
| 325 | So for example to dump "foo" and "bar" instance buffer on all CPUs,
|
| 326 | user can::
|
| 327 |
|
| 328 | echo "foo,bar" > /proc/sys/kernel/ftrace_dump_on_oops
|
| 329 |
|
| 330 | To dump global buffer and "foo" instance buffer on all
|
| 331 | CPUs along with the "bar" instance buffer on CPU that triggered the
|
| 332 | oops, user can::
|
| 333 |
|
| 334 | echo "1,foo,bar=2" > /proc/sys/kernel/ftrace_dump_on_oops
|
| 335 |
|
| 336 | ftrace_enabled, stack_tracer_enabled
|
| 337 | ====================================
|
| 338 |
|
| 339 | See Documentation/trace/ftrace.rst.
|
| 340 |
|
| 341 |
|
| 342 | hardlockup_all_cpu_backtrace
|
| 343 | ============================
|
| 344 |
|
| 345 | This value controls the hard lockup detector behavior when a hard
|
| 346 | lockup condition is detected as to whether or not to gather further
|
| 347 | debug information. If enabled, arch-specific all-CPU stack dumping
|
| 348 | will be initiated.
|
| 349 |
|
| 350 | = ============================================
|
| 351 | 0 Do nothing. This is the default behavior.
|
| 352 | 1 On detection capture more debug information.
|
| 353 | = ============================================
|
| 354 |
|
| 355 |
|
| 356 | hardlockup_panic
|
| 357 | ================
|
| 358 |
|
| 359 | This parameter can be used to control whether the kernel panics
|
| 360 | when a hard lockup is detected.
|
| 361 |
|
| 362 | = ===========================
|
| 363 | 0 Don't panic on hard lockup.
|
| 364 | 1 Panic on hard lockup.
|
| 365 | = ===========================
|
| 366 |
|
| 367 | See Documentation/admin-guide/lockup-watchdogs.rst for more information.
|
| 368 | This can also be set using the nmi_watchdog kernel parameter.
|
| 369 |
|
| 370 |
|
| 371 | hotplug
|
| 372 | =======
|
| 373 |
|
| 374 | Path for the hotplug policy agent.
|
| 375 | Default value is ``CONFIG_UEVENT_HELPER_PATH``, which in turn defaults
|
| 376 | to the empty string.
|
| 377 |
|
| 378 | This file only exists when ``CONFIG_UEVENT_HELPER`` is enabled. Most
|
| 379 | modern systems rely exclusively on the netlink-based uevent source and
|
| 380 | don't need this.
|
| 381 |
|
| 382 |
|
| 383 | hung_task_all_cpu_backtrace
|
| 384 | ===========================
|
| 385 |
|
| 386 | If this option is set, the kernel will send an NMI to all CPUs to dump
|
| 387 | their backtraces when a hung task is detected. This file shows up if
|
| 388 | CONFIG_DETECT_HUNG_TASK and CONFIG_SMP are enabled.
|
| 389 |
|
| 390 | 0: Won't show all CPUs backtraces when a hung task is detected.
|
| 391 | This is the default behavior.
|
| 392 |
|
| 393 | 1: Will non-maskably interrupt all CPUs and dump their backtraces when
|
| 394 | a hung task is detected.
|
| 395 |
|
| 396 |
|
| 397 | hung_task_panic
|
| 398 | ===============
|
| 399 |
|
| 400 | When set to a non-zero value, a kernel panic will be triggered if the
|
| 401 | number of hung tasks found during a single scan reaches this value.
|
| 402 | This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
|
| 403 |
|
| 404 | = =======================================================
|
| 405 | 0 Continue operation. This is the default behavior.
|
| 406 | N Panic when N hung tasks are found during a single scan.
|
| 407 | = =======================================================
|
| 408 |
|
| 409 |
|
| 410 | hung_task_check_count
|
| 411 | =====================
|
| 412 |
|
| 413 | The upper bound on the number of tasks that are checked.
|
| 414 | This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
|
| 415 |
|
| 416 |
|
| 417 | hung_task_detect_count
|
| 418 | ======================
|
| 419 |
|
| 420 | Indicates the total number of tasks that have been detected as hung since
|
| 421 | the system boot or since the counter was reset. The counter is zeroed when
|
| 422 | a value of 0 is written.
|
| 423 |
|
| 424 | This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
|
| 425 |
|
| 426 | hung_task_sys_info
|
| 427 | ==================
|
| 428 | A comma separated list of extra system information to be dumped when
|
| 429 | hung task is detected, for example, "tasks,mem,timers,locks,...".
|
| 430 | Refer 'panic_sys_info' section below for more details.
|
| 431 |
|
| 432 | hung_task_timeout_secs
|
| 433 | ======================
|
| 434 |
|
| 435 | When a task in D state did not get scheduled
|
| 436 | for more than this value report a warning.
|
| 437 | This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
|
| 438 |
|
| 439 | 0 means infinite timeout, no checking is done.
|
| 440 |
|
| 441 | Possible values to set are in range {0:``LONG_MAX``/``HZ``}.
|
| 442 |
|
| 443 |
|
| 444 | hung_task_check_interval_secs
|
| 445 | =============================
|
| 446 |
|
| 447 | Hung task check interval. If hung task checking is enabled
|
| 448 | (see `hung_task_timeout_secs`_), the check is done every
|
| 449 | ``hung_task_check_interval_secs`` seconds.
|
| 450 | This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
|
| 451 |
|
| 452 | 0 (default) means use ``hung_task_timeout_secs`` as checking
|
| 453 | interval.
|
| 454 |
|
| 455 | Possible values to set are in range {0:``LONG_MAX``/``HZ``}.
|
| 456 |
|
| 457 |
|
| 458 | hung_task_warnings
|
| 459 | ==================
|
| 460 |
|
| 461 | The maximum number of warnings to report. During a check interval
|
| 462 | if a hung task is detected, this value is decreased by 1.
|
| 463 | When this value reaches 0, no more warnings will be reported.
|
| 464 | This file shows up if ``CONFIG_DETECT_HUNG_TASK`` is enabled.
|
| 465 |
|
| 466 | -1: report an infinite number of warnings.
|
| 467 |
|
| 468 |
|
| 469 | hyperv_record_panic_msg
|
| 470 | =======================
|
| 471 |
|
| 472 | Controls whether the panic kmsg data should be reported to Hyper-V.
|
| 473 |
|
| 474 | = =========================================================
|
| 475 | 0 Do not report panic kmsg data.
|
| 476 | 1 Report the panic kmsg data. This is the default behavior.
|
| 477 | = =========================================================
|
| 478 |
|
| 479 |
|
| 480 | ignore-unaligned-usertrap
|
| 481 | =========================
|
| 482 |
|
| 483 | On architectures where unaligned accesses cause traps, and where this
|
| 484 | feature is supported (``CONFIG_SYSCTL_ARCH_UNALIGN_NO_WARN``;
|
| 485 | currently, ``arc``, ``parisc`` and ``loongarch``), controls whether all
|
| 486 | unaligned traps are logged.
|
| 487 |
|
| 488 | = =============================================================
|
| 489 | 0 Log all unaligned accesses.
|
| 490 | 1 Only warn the first time a process traps. This is the default
|
| 491 | setting.
|
| 492 | = =============================================================
|
| 493 |
|
| 494 | See also `unaligned-trap`_.
|
| 495 |
|
| 496 | io_uring_disabled
|
| 497 | =================
|
| 498 |
|
| 499 | Prevents all processes from creating new io_uring instances. Enabling this
|
| 500 | shrinks the kernel's attack surface.
|
| 501 |
|
| 502 | = ======================================================================
|
| 503 | 0 All processes can create io_uring instances as normal. This is the
|
| 504 | default setting.
|
| 505 | 1 io_uring creation is disabled (io_uring_setup() will fail with
|
| 506 | -EPERM) for unprivileged processes not in the io_uring_group group.
|
| 507 | Existing io_uring instances can still be used. See the
|
| 508 | documentation for io_uring_group for more information.
|
| 509 | 2 io_uring creation is disabled for all processes. io_uring_setup()
|
| 510 | always fails with -EPERM. Existing io_uring instances can still be
|
| 511 | used.
|
| 512 | = ======================================================================
|
| 513 |
|
| 514 |
|
| 515 | io_uring_group
|
| 516 | ==============
|
| 517 |
|
| 518 | When io_uring_disabled is set to 1, a process must either be
|
| 519 | privileged (CAP_SYS_ADMIN) or be in the io_uring_group group in order
|
| 520 | to create an io_uring instance. If io_uring_group is set to -1 (the
|
| 521 | default), only processes with the CAP_SYS_ADMIN capability may create
|
| 522 | io_uring instances.
|
| 523 |
|
| 524 |
|
| 525 | kernel_sys_info
|
| 526 | ===============
|
| 527 | A comma separated list of extra system information to be dumped when
|
| 528 | soft/hard lockup is detected, for example, "tasks,mem,timers,locks,...".
|
| 529 | Refer 'panic_sys_info' section below for more details.
|
| 530 |
|
| 531 | It serves as the default kernel control knob, which will take effect
|
| 532 | when a kernel module calls sys_info() with parameter==0.
|
| 533 |
|
| 534 | kexec_load_disabled
|
| 535 | ===================
|
| 536 |
|
| 537 | A toggle indicating if the syscalls ``kexec_load`` and
|
| 538 | ``kexec_file_load`` have been disabled.
|
| 539 | This value defaults to 0 (false: ``kexec_*load`` enabled), but can be
|
| 540 | set to 1 (true: ``kexec_*load`` disabled).
|
| 541 | Once true, kexec can no longer be used, and the toggle cannot be set
|
| 542 | back to false.
|
| 543 | This allows a kexec image to be loaded before disabling the syscall,
|
| 544 | allowing a system to set up (and later use) an image without it being
|
| 545 | altered.
|
| 546 | Generally used together with the `modules_disabled`_ sysctl.
|
| 547 |
|
| 548 | kexec_load_limit_panic
|
| 549 | ======================
|
| 550 |
|
| 551 | This parameter specifies a limit to the number of times the syscalls
|
| 552 | ``kexec_load`` and ``kexec_file_load`` can be called with a crash
|
| 553 | image. It can only be set with a more restrictive value than the
|
| 554 | current one.
|
| 555 |
|
| 556 | == ======================================================
|
| 557 | -1 Unlimited calls to kexec. This is the default setting.
|
| 558 | N Number of calls left.
|
| 559 | == ======================================================
|
| 560 |
|
| 561 | kexec_load_limit_reboot
|
| 562 | =======================
|
| 563 |
|
| 564 | Similar functionality as ``kexec_load_limit_panic``, but for a normal
|
| 565 | image.
|
| 566 |
|
| 567 | kptr_restrict
|
| 568 | =============
|
| 569 |
|
| 570 | This toggle indicates whether restrictions are placed on
|
| 571 | exposing kernel addresses via ``/proc`` and other interfaces.
|
| 572 |
|
| 573 | When ``kptr_restrict`` is set to 0 (the default) the address is hashed
|
| 574 | before printing.
|
| 575 | (This is the equivalent to %p.)
|
| 576 |
|
| 577 | When ``kptr_restrict`` is set to 1, kernel pointers printed using the
|
| 578 | %pK format specifier will be replaced with 0s unless the user has
|
| 579 | ``CAP_SYSLOG`` and effective user and group ids are equal to the real
|
| 580 | ids.
|
| 581 | This is because %pK checks are done at read() time rather than open()
|
| 582 | time, so if permissions are elevated between the open() and the read()
|
| 583 | (e.g via a setuid binary) then %pK will not leak kernel pointers to
|
| 584 | unprivileged users.
|
| 585 | Note, this is a temporary solution only.
|
| 586 | The correct long-term solution is to do the permission checks at
|
| 587 | open() time.
|
| 588 | Consider removing world read permissions from files that use %pK, and
|
| 589 | using `dmesg_restrict`_ to protect against uses of %pK in ``dmesg(8)``
|
| 590 | if leaking kernel pointer values to unprivileged users is a concern.
|
| 591 |
|
| 592 | When ``kptr_restrict`` is set to 2, kernel pointers printed using
|
| 593 | %pK will be replaced with 0s regardless of privileges.
|
| 594 |
|
| 595 | For disabling these security restrictions early at boot time (and once
|
| 596 | for all), use the ``hash_pointers`` boot parameter instead.
|
| 597 |
|
| 598 | softlockup_sys_info & hardlockup_sys_info
|
| 599 | =========================================
|
| 600 | A comma separated list of extra system information to be dumped when
|
| 601 | soft/hard lockup is detected, for example, "tasks,mem,timers,locks,...".
|
| 602 | Refer 'panic_sys_info' section below for more details.
|
| 603 |
|
| 604 | modprobe
|
| 605 | ========
|
| 606 |
|
| 607 | The full path to the usermode helper for autoloading kernel modules,
|
| 608 | by default ``CONFIG_MODPROBE_PATH``, which in turn defaults to
|
| 609 | "/sbin/modprobe". This binary is executed when the kernel requests a
|
| 610 | module. For example, if userspace passes an unknown filesystem type
|
| 611 | to mount(), then the kernel will automatically request the
|
| 612 | corresponding filesystem module by executing this usermode helper.
|
| 613 | This usermode helper should insert the needed module into the kernel.
|
| 614 |
|
| 615 | This sysctl only affects module autoloading. It has no effect on the
|
| 616 | ability to explicitly insert modules.
|
| 617 |
|
| 618 | This sysctl can be used to debug module loading requests::
|
| 619 |
|
| 620 | echo '#! /bin/sh' > /tmp/modprobe
|
| 621 | echo 'echo "$@" >> /tmp/modprobe.log' >> /tmp/modprobe
|
| 622 | echo 'exec /sbin/modprobe "$@"' >> /tmp/modprobe
|
| 623 | chmod a+x /tmp/modprobe
|
| 624 | echo /tmp/modprobe > /proc/sys/kernel/modprobe
|
| 625 |
|
| 626 | Alternatively, if this sysctl is set to the empty string, then module
|
| 627 | autoloading is completely disabled. The kernel will not try to
|
| 628 | execute a usermode helper at all, nor will it call the
|
| 629 | kernel_module_request LSM hook.
|
| 630 |
|
| 631 | If CONFIG_STATIC_USERMODEHELPER=y is set in the kernel configuration,
|
| 632 | then the configured static usermode helper overrides this sysctl,
|
| 633 | except that the empty string is still accepted to completely disable
|
| 634 | module autoloading as described above.
|
| 635 |
|
| 636 | modules_disabled
|
| 637 | ================
|
| 638 |
|
| 639 | A toggle value indicating if modules are allowed to be loaded
|
| 640 | in an otherwise modular kernel. This toggle defaults to off
|
| 641 | (0), but can be set true (1). Once true, modules can be
|
| 642 | neither loaded nor unloaded, and the toggle cannot be set back
|
| 643 | to false. Generally used with the `kexec_load_disabled`_ toggle.
|
| 644 |
|
| 645 |
|
| 646 | .. _msgmni:
|
| 647 |
|
| 648 | msgmax, msgmnb, and msgmni
|
| 649 | ==========================
|
| 650 |
|
| 651 | ``msgmax`` is the maximum size of an IPC message, in bytes. 8192 by
|
| 652 | default (``MSGMAX``).
|
| 653 |
|
| 654 | ``msgmnb`` is the maximum size of an IPC queue, in bytes. 16384 by
|
| 655 | default (``MSGMNB``).
|
| 656 |
|
| 657 | ``msgmni`` is the maximum number of IPC queues. 32000 by default
|
| 658 | (``MSGMNI``).
|
| 659 |
|
| 660 | All of these parameters are set per ipc namespace. The maximum number of bytes
|
| 661 | in POSIX message queues is limited by ``RLIMIT_MSGQUEUE``. This limit is
|
| 662 | respected hierarchically in the each user namespace.
|
| 663 |
|
| 664 | msg_next_id, sem_next_id, and shm_next_id (System V IPC)
|
| 665 | ========================================================
|
| 666 |
|
| 667 | These three toggles allows to specify desired id for next allocated IPC
|
| 668 | object: message, semaphore or shared memory respectively.
|
| 669 |
|
| 670 | By default they are equal to -1, which means generic allocation logic.
|
| 671 | Possible values to set are in range {0:``INT_MAX``}.
|
| 672 |
|
| 673 | Notes:
|
| 674 | 1) kernel doesn't guarantee, that new object will have desired id. So,
|
| 675 | it's up to userspace, how to handle an object with "wrong" id.
|
| 676 | 2) Toggle with non-default value will be set back to -1 by kernel after
|
| 677 | successful IPC object allocation. If an IPC object allocation syscall
|
| 678 | fails, it is undefined if the value remains unmodified or is reset to -1.
|
| 679 |
|
| 680 |
|
| 681 | ngroups_max
|
| 682 | ===========
|
| 683 |
|
| 684 | Maximum number of supplementary groups, _i.e._ the maximum size which
|
| 685 | ``setgroups`` will accept. Exports ``NGROUPS_MAX`` from the kernel.
|
| 686 |
|
| 687 |
|
| 688 |
|
| 689 | nmi_watchdog
|
| 690 | ============
|
| 691 |
|
| 692 | This parameter can be used to control the NMI watchdog
|
| 693 | (i.e. the hard lockup detector) on x86 systems.
|
| 694 |
|
| 695 | = =================================
|
| 696 | 0 Disable the hard lockup detector.
|
| 697 | 1 Enable the hard lockup detector.
|
| 698 | = =================================
|
| 699 |
|
| 700 | The hard lockup detector monitors each CPU for its ability to respond to
|
| 701 | timer interrupts. The mechanism utilizes CPU performance counter registers
|
| 702 | that are programmed to generate Non-Maskable Interrupts (NMIs) periodically
|
| 703 | while a CPU is busy. Hence, the alternative name 'NMI watchdog'.
|
| 704 |
|
| 705 | The NMI watchdog is disabled by default if the kernel is running as a guest
|
| 706 | in a KVM virtual machine. This default can be overridden by adding::
|
| 707 |
|
| 708 | nmi_watchdog=1
|
| 709 |
|
| 710 | to the guest kernel command line (see
|
| 711 | Documentation/admin-guide/kernel-parameters.rst).
|
| 712 |
|
| 713 |
|
| 714 | nmi_wd_lpm_factor (PPC only)
|
| 715 | ============================
|
| 716 |
|
| 717 | Factor to apply to the NMI watchdog timeout (only when ``nmi_watchdog`` is
|
| 718 | set to 1). This factor represents the percentage added to
|
| 719 | ``watchdog_thresh`` when calculating the NMI watchdog timeout during an
|
| 720 | LPM. The soft lockup timeout is not impacted.
|
| 721 |
|
| 722 | A value of 0 means no change. The default value is 200 meaning the NMI
|
| 723 | watchdog is set to 30s (based on ``watchdog_thresh`` equal to 10).
|
| 724 |
|
| 725 |
|
| 726 | numa_balancing
|
| 727 | ==============
|
| 728 |
|
| 729 | Enables/disables and configures automatic page fault based NUMA memory
|
| 730 | balancing. Memory is moved automatically to nodes that access it often.
|
| 731 | The value to set can be the result of ORing the following:
|
| 732 |
|
| 733 | = =================================
|
| 734 | 0 NUMA_BALANCING_DISABLED
|
| 735 | 1 NUMA_BALANCING_NORMAL
|
| 736 | 2 NUMA_BALANCING_MEMORY_TIERING
|
| 737 | = =================================
|
| 738 |
|
| 739 | Or NUMA_BALANCING_NORMAL to optimize page placement among different
|
| 740 | NUMA nodes to reduce remote accessing. On NUMA machines, there is a
|
| 741 | performance penalty if remote memory is accessed by a CPU. When this
|
| 742 | feature is enabled the kernel samples what task thread is accessing
|
| 743 | memory by periodically unmapping pages and later trapping a page
|
| 744 | fault. At the time of the page fault, it is determined if the data
|
| 745 | being accessed should be migrated to a local memory node.
|
| 746 |
|
| 747 | The unmapping of pages and trapping faults incur additional overhead that
|
| 748 | ideally is offset by improved memory locality but there is no universal
|
| 749 | guarantee. If the target workload is already bound to NUMA nodes then this
|
| 750 | feature should be disabled.
|
| 751 |
|
| 752 | Or NUMA_BALANCING_MEMORY_TIERING to optimize page placement among
|
| 753 | different types of memory (represented as different NUMA nodes) to
|
| 754 | place the hot pages in the fast memory. This is implemented based on
|
| 755 | unmapping and page fault too.
|
| 756 |
|
| 757 | numa_balancing_promote_rate_limit_MBps
|
| 758 | ======================================
|
| 759 |
|
| 760 | Too high promotion/demotion throughput between different memory types
|
| 761 | may hurt application latency. This can be used to rate limit the
|
| 762 | promotion throughput. The per-node max promotion throughput in MB/s
|
| 763 | will be limited to be no more than the set value.
|
| 764 |
|
| 765 | A rule of thumb is to set this to less than 1/10 of the PMEM node
|
| 766 | write bandwidth.
|
| 767 |
|
| 768 | oops_all_cpu_backtrace
|
| 769 | ======================
|
| 770 |
|
| 771 | If this option is set, the kernel will send an NMI to all CPUs to dump
|
| 772 | their backtraces when an oops event occurs. It should be used as a last
|
| 773 | resort in case a panic cannot be triggered (to protect VMs running, for
|
| 774 | example) or kdump can't be collected. This file shows up if CONFIG_SMP
|
| 775 | is enabled.
|
| 776 |
|
| 777 | 0: Won't show all CPUs backtraces when an oops is detected.
|
| 778 | This is the default behavior.
|
| 779 |
|
| 780 | 1: Will non-maskably interrupt all CPUs and dump their backtraces when
|
| 781 | an oops event is detected.
|
| 782 |
|
| 783 |
|
| 784 | oops_limit
|
| 785 | ==========
|
| 786 |
|
| 787 | Number of kernel oopses after which the kernel should panic when
|
| 788 | ``panic_on_oops`` is not set. Setting this to 0 disables checking
|
| 789 | the count. Setting this to 1 has the same effect as setting
|
| 790 | ``panic_on_oops=1``. The default value is 10000.
|
| 791 |
|
| 792 |
|
| 793 | osrelease, ostype & version
|
| 794 | ===========================
|
| 795 |
|
| 796 | ::
|
| 797 |
|
| 798 | # cat osrelease
|
| 799 | 2.1.88
|
| 800 | # cat ostype
|
| 801 | Linux
|
| 802 | # cat version
|
| 803 | #5 Wed Feb 25 21:49:24 MET 1998
|
| 804 |
|
| 805 | The files ``osrelease`` and ``ostype`` should be clear enough.
|
| 806 | ``version``
|
| 807 | needs a little more clarification however. The '#5' means that
|
| 808 | this is the fifth kernel built from this source base and the
|
| 809 | date behind it indicates the time the kernel was built.
|
| 810 | The only way to tune these values is to rebuild the kernel :-)
|
| 811 |
|
| 812 |
|
| 813 | overflowgid & overflowuid
|
| 814 | =========================
|
| 815 |
|
| 816 | if your architecture did not always support 32-bit UIDs (i.e. arm,
|
| 817 | i386, m68k, sh, and sparc32), a fixed UID and GID will be returned to
|
| 818 | applications that use the old 16-bit UID/GID system calls, if the
|
| 819 | actual UID or GID would exceed 65535.
|
| 820 |
|
| 821 | These sysctls allow you to change the value of the fixed UID and GID.
|
| 822 | The default is 65534.
|
| 823 |
|
| 824 |
|
| 825 | panic
|
| 826 | =====
|
| 827 |
|
| 828 | The value in this file determines the behaviour of the kernel on a
|
| 829 | panic:
|
| 830 |
|
| 831 | * if zero, the kernel will loop forever;
|
| 832 | * if negative, the kernel will reboot immediately;
|
| 833 | * if positive, the kernel will reboot after the corresponding number
|
| 834 | of seconds.
|
| 835 |
|
| 836 | When you use the software watchdog, the recommended setting is 60.
|
| 837 |
|
| 838 |
|
| 839 | panic_on_io_nmi
|
| 840 | ===============
|
| 841 |
|
| 842 | Controls the kernel's behavior when a CPU receives an NMI caused by
|
| 843 | an IO error.
|
| 844 |
|
| 845 | = ==================================================================
|
| 846 | 0 Try to continue operation (default).
|
| 847 | 1 Panic immediately. The IO error triggered an NMI. This indicates a
|
| 848 | serious system condition which could result in IO data corruption.
|
| 849 | Rather than continuing, panicking might be a better choice. Some
|
| 850 | servers issue this sort of NMI when the dump button is pushed,
|
| 851 | and you can use this option to take a crash dump.
|
| 852 | = ==================================================================
|
| 853 |
|
| 854 |
|
| 855 | panic_on_oops
|
| 856 | =============
|
| 857 |
|
| 858 | Controls the kernel's behaviour when an oops or BUG is encountered.
|
| 859 |
|
| 860 | = ===================================================================
|
| 861 | 0 Try to continue operation.
|
| 862 | 1 Panic immediately. If the `panic` sysctl is also non-zero then the
|
| 863 | machine will be rebooted.
|
| 864 | = ===================================================================
|
| 865 |
|
| 866 |
|
| 867 | panic_on_stackoverflow
|
| 868 | ======================
|
| 869 |
|
| 870 | Controls the kernel's behavior when detecting the overflows of
|
| 871 | kernel, IRQ and exception stacks except a user stack.
|
| 872 | This file shows up if ``CONFIG_DEBUG_STACKOVERFLOW`` is enabled.
|
| 873 |
|
| 874 | = ==========================
|
| 875 | 0 Try to continue operation.
|
| 876 | 1 Panic immediately.
|
| 877 | = ==========================
|
| 878 |
|
| 879 |
|
| 880 | panic_on_unrecovered_nmi
|
| 881 | ========================
|
| 882 |
|
| 883 | The default Linux behaviour on an NMI of either memory or unknown is
|
| 884 | to continue operation. For many environments such as scientific
|
| 885 | computing it is preferable that the box is taken out and the error
|
| 886 | dealt with than an uncorrected parity/ECC error get propagated.
|
| 887 |
|
| 888 | A small number of systems do generate NMIs for bizarre random reasons
|
| 889 | such as power management so the default is off. That sysctl works like
|
| 890 | the existing panic controls already in that directory.
|
| 891 |
|
| 892 |
|
| 893 | panic_on_warn
|
| 894 | =============
|
| 895 |
|
| 896 | Calls panic() in the WARN() path when set to 1. This is useful to avoid
|
| 897 | a kernel rebuild when attempting to kdump at the location of a WARN().
|
| 898 |
|
| 899 | = ================================================
|
| 900 | 0 Only WARN(), default behaviour.
|
| 901 | 1 Call panic() after printing out WARN() location.
|
| 902 | = ================================================
|
| 903 |
|
| 904 |
|
| 905 | panic_print
|
| 906 | ===========
|
| 907 |
|
| 908 | Bitmask for printing system info when panic happens. User can chose
|
| 909 | combination of the following bits:
|
| 910 |
|
| 911 | ===== ============================================
|
| 912 | bit 0 print all tasks info
|
| 913 | bit 1 print system memory info
|
| 914 | bit 2 print timer info
|
| 915 | bit 3 print locks info if ``CONFIG_LOCKDEP`` is on
|
| 916 | bit 4 print ftrace buffer
|
| 917 | bit 5 replay all kernel messages on consoles at the end of panic
|
| 918 | bit 6 print all CPUs backtrace (if available in the arch)
|
| 919 | bit 7 print only tasks in uninterruptible (blocked) state
|
| 920 | ===== ============================================
|
| 921 |
|
| 922 | So for example to print tasks and memory info on panic, user can::
|
| 923 |
|
| 924 | echo 3 > /proc/sys/kernel/panic_print
|
| 925 |
|
| 926 |
|
| 927 | panic_sys_info
|
| 928 | ==============
|
| 929 |
|
| 930 | A comma separated list of extra information to be dumped on panic,
|
| 931 | for example, "tasks,mem,timers,...". It is a human readable alternative
|
| 932 | to 'panic_print'. Possible values are:
|
| 933 |
|
| 934 | ============= ===================================================
|
| 935 | tasks print all tasks info
|
| 936 | mem print system memory info
|
| 937 | timers print timers info
|
| 938 | locks print locks info if CONFIG_LOCKDEP is on
|
| 939 | ftrace print ftrace buffer
|
| 940 | all_bt print all CPUs backtrace (if available in the arch)
|
| 941 | blocked_tasks print only tasks in uninterruptible (blocked) state
|
| 942 | ============= ===================================================
|
| 943 |
|
| 944 |
|
| 945 | panic_on_rcu_stall
|
| 946 | ==================
|
| 947 |
|
| 948 | When set to 1, calls panic() after RCU stall detection messages. This
|
| 949 | is useful to define the root cause of RCU stalls using a vmcore.
|
| 950 |
|
| 951 | = ============================================================
|
| 952 | 0 Do not panic() when RCU stall takes place, default behavior.
|
| 953 | 1 panic() after printing RCU stall messages.
|
| 954 | = ============================================================
|
| 955 |
|
| 956 | max_rcu_stall_to_panic
|
| 957 | ======================
|
| 958 |
|
| 959 | When ``panic_on_rcu_stall`` is set to 1, this value determines the
|
| 960 | number of times that RCU can stall before panic() is called.
|
| 961 |
|
| 962 | When ``panic_on_rcu_stall`` is set to 0, this value is has no effect.
|
| 963 |
|
| 964 | perf_cpu_time_max_percent
|
| 965 | =========================
|
| 966 |
|
| 967 | Hints to the kernel how much CPU time it should be allowed to
|
| 968 | use to handle perf sampling events. If the perf subsystem
|
| 969 | is informed that its samples are exceeding this limit, it
|
| 970 | will drop its sampling frequency to attempt to reduce its CPU
|
| 971 | usage.
|
| 972 |
|
| 973 | Some perf sampling happens in NMIs. If these samples
|
| 974 | unexpectedly take too long to execute, the NMIs can become
|
| 975 | stacked up next to each other so much that nothing else is
|
| 976 | allowed to execute.
|
| 977 |
|
| 978 | ===== ========================================================
|
| 979 | 0 Disable the mechanism. Do not monitor or correct perf's
|
| 980 | sampling rate no matter how CPU time it takes.
|
| 981 |
|
| 982 | 1-100 Attempt to throttle perf's sample rate to this
|
| 983 | percentage of CPU. Note: the kernel calculates an
|
| 984 | "expected" length of each sample event. 100 here means
|
| 985 | 100% of that expected length. Even if this is set to
|
| 986 | 100, you may still see sample throttling if this
|
| 987 | length is exceeded. Set to 0 if you truly do not care
|
| 988 | how much CPU is consumed.
|
| 989 | ===== ========================================================
|
| 990 |
|
| 991 |
|
| 992 | perf_event_paranoid
|
| 993 | ===================
|
| 994 |
|
| 995 | Controls use of the performance events system by unprivileged
|
| 996 | users (without CAP_PERFMON). The default value is 2.
|
| 997 |
|
| 998 | For backward compatibility reasons access to system performance
|
| 999 | monitoring and observability remains open for CAP_SYS_ADMIN
|
| 1000 | privileged processes but CAP_SYS_ADMIN usage for secure system
|
| 1001 | performance monitoring and observability operations is discouraged
|
| 1002 | with respect to CAP_PERFMON use cases.
|
| 1003 |
|
| 1004 | === ==================================================================
|
| 1005 | -1 Allow use of (almost) all events by all users.
|
| 1006 |
|
| 1007 | Ignore mlock limit after perf_event_mlock_kb without
|
| 1008 | ``CAP_IPC_LOCK``.
|
| 1009 |
|
| 1010 | >=0 Disallow ftrace function tracepoint by users without
|
| 1011 | ``CAP_PERFMON``.
|
| 1012 |
|
| 1013 | Disallow raw tracepoint access by users without ``CAP_PERFMON``.
|
| 1014 |
|
| 1015 | >=1 Disallow CPU event access by users without ``CAP_PERFMON``.
|
| 1016 |
|
| 1017 | >=2 Disallow kernel profiling by users without ``CAP_PERFMON``.
|
| 1018 | === ==================================================================
|
| 1019 |
|
| 1020 |
|
| 1021 | perf_event_max_stack
|
| 1022 | ====================
|
| 1023 |
|
| 1024 | Controls maximum number of stack frames to copy for (``attr.sample_type &
|
| 1025 | PERF_SAMPLE_CALLCHAIN``) configured events, for instance, when using
|
| 1026 | '``perf record -g``' or '``perf trace --call-graph fp``'.
|
| 1027 |
|
| 1028 | This can only be done when no events are in use that have callchains
|
| 1029 | enabled, otherwise writing to this file will return ``-EBUSY``.
|
| 1030 |
|
| 1031 | The default value is 127.
|
| 1032 |
|
| 1033 |
|
| 1034 | perf_event_mlock_kb
|
| 1035 | ===================
|
| 1036 |
|
| 1037 | Control size of per-cpu ring buffer not counted against mlock limit.
|
| 1038 |
|
| 1039 | The default value is 512 + 1 page
|
| 1040 |
|
| 1041 |
|
| 1042 | perf_event_max_contexts_per_stack
|
| 1043 | =================================
|
| 1044 |
|
| 1045 | Controls maximum number of stack frame context entries for
|
| 1046 | (``attr.sample_type & PERF_SAMPLE_CALLCHAIN``) configured events, for
|
| 1047 | instance, when using '``perf record -g``' or '``perf trace --call-graph fp``'.
|
| 1048 |
|
| 1049 | This can only be done when no events are in use that have callchains
|
| 1050 | enabled, otherwise writing to this file will return ``-EBUSY``.
|
| 1051 |
|
| 1052 | The default value is 8.
|
| 1053 |
|
| 1054 |
|
| 1055 | perf_user_access (arm64 and riscv only)
|
| 1056 | =======================================
|
| 1057 |
|
| 1058 | Controls user space access for reading perf event counters.
|
| 1059 |
|
| 1060 | * for arm64
|
| 1061 | The default value is 0 (access disabled).
|
| 1062 |
|
| 1063 | When set to 1, user space can read performance monitor counter registers
|
| 1064 | directly.
|
| 1065 |
|
| 1066 | See Documentation/arch/arm64/perf.rst for more information.
|
| 1067 |
|
| 1068 | * for riscv
|
| 1069 | When set to 0, user space access is disabled.
|
| 1070 |
|
| 1071 | The default value is 1, user space can read performance monitor counter
|
| 1072 | registers through perf, any direct access without perf intervention will trigger
|
| 1073 | an illegal instruction.
|
| 1074 |
|
| 1075 | When set to 2, which enables legacy mode (user space has direct access to cycle
|
| 1076 | and insret CSRs only). Note that this legacy value is deprecated and will be
|
| 1077 | removed once all user space applications are fixed.
|
| 1078 |
|
| 1079 | Note that the time CSR is always directly accessible to all modes.
|
| 1080 |
|
| 1081 | pid_max
|
| 1082 | =======
|
| 1083 |
|
| 1084 | PID allocation wrap value. When the kernel's next PID value
|
| 1085 | reaches this value, it wraps back to a minimum PID value.
|
| 1086 | PIDs of value ``pid_max`` or larger are not allocated.
|
| 1087 |
|
| 1088 |
|
| 1089 | ns_last_pid
|
| 1090 | ===========
|
| 1091 |
|
| 1092 | The last pid allocated in the current (the one task using this sysctl
|
| 1093 | lives in) pid namespace. When selecting a pid for a next task on fork
|
| 1094 | kernel tries to allocate a number starting from this one.
|
| 1095 |
|
| 1096 |
|
| 1097 | powersave-nap (PPC only)
|
| 1098 | ========================
|
| 1099 |
|
| 1100 | If set, Linux-PPC will use the 'nap' mode of powersaving,
|
| 1101 | otherwise the 'doze' mode will be used.
|
| 1102 |
|
| 1103 |
|
| 1104 | ==============================================================
|
| 1105 |
|
| 1106 | printk
|
| 1107 | ======
|
| 1108 |
|
| 1109 | The four values in printk denote: ``console_loglevel``,
|
| 1110 | ``default_message_loglevel``, ``minimum_console_loglevel`` and
|
| 1111 | ``default_console_loglevel`` respectively.
|
| 1112 |
|
| 1113 | These values influence printk() behavior when printing or
|
| 1114 | logging error messages. See '``man 2 syslog``' for more info on
|
| 1115 | the different loglevels.
|
| 1116 |
|
| 1117 | ======================== =====================================
|
| 1118 | console_loglevel messages with a higher priority than
|
| 1119 | this will be printed to the console
|
| 1120 | default_message_loglevel messages without an explicit priority
|
| 1121 | will be printed with this priority
|
| 1122 | minimum_console_loglevel minimum (highest) value to which
|
| 1123 | console_loglevel can be set
|
| 1124 | default_console_loglevel default value for console_loglevel
|
| 1125 | ======================== =====================================
|
| 1126 |
|
| 1127 |
|
| 1128 | printk_delay
|
| 1129 | ============
|
| 1130 |
|
| 1131 | Delay each printk message in ``printk_delay`` milliseconds
|
| 1132 |
|
| 1133 | Value from 0 - 10000 is allowed.
|
| 1134 |
|
| 1135 |
|
| 1136 | printk_ratelimit
|
| 1137 | ================
|
| 1138 |
|
| 1139 | Some warning messages are rate limited. ``printk_ratelimit`` specifies
|
| 1140 | the minimum length of time between these messages (in seconds).
|
| 1141 | The default value is 5 seconds.
|
| 1142 |
|
| 1143 | A value of 0 will disable rate limiting.
|
| 1144 |
|
| 1145 |
|
| 1146 | printk_ratelimit_burst
|
| 1147 | ======================
|
| 1148 |
|
| 1149 | While long term we enforce one message per `printk_ratelimit`_
|
| 1150 | seconds, we do allow a burst of messages to pass through.
|
| 1151 | ``printk_ratelimit_burst`` specifies the number of messages we can
|
| 1152 | send before ratelimiting kicks in. After `printk_ratelimit`_ seconds
|
| 1153 | have elapsed, another burst of messages may be sent.
|
| 1154 |
|
| 1155 | The default value is 10 messages.
|
| 1156 |
|
| 1157 |
|
| 1158 | printk_devkmsg
|
| 1159 | ==============
|
| 1160 |
|
| 1161 | Control the logging to ``/dev/kmsg`` from userspace:
|
| 1162 |
|
| 1163 | ========= =============================================
|
| 1164 | ratelimit default, ratelimited
|
| 1165 | on unlimited logging to /dev/kmsg from userspace
|
| 1166 | off logging to /dev/kmsg disabled
|
| 1167 | ========= =============================================
|
| 1168 |
|
| 1169 | The kernel command line parameter ``printk.devkmsg=`` overrides this and is
|
| 1170 | a one-time setting until next reboot: once set, it cannot be changed by
|
| 1171 | this sysctl interface anymore.
|
| 1172 |
|
| 1173 | ==============================================================
|
| 1174 |
|
| 1175 |
|
| 1176 | pty
|
| 1177 | ===
|
| 1178 |
|
| 1179 | See Documentation/filesystems/devpts.rst.
|
| 1180 |
|
| 1181 |
|
| 1182 | random
|
| 1183 | ======
|
| 1184 |
|
| 1185 | This is a directory, with the following entries:
|
| 1186 |
|
| 1187 | * ``boot_id``: a UUID generated the first time this is retrieved, and
|
| 1188 | unvarying after that;
|
| 1189 |
|
| 1190 | * ``uuid``: a UUID generated every time this is retrieved (this can
|
| 1191 | thus be used to generate UUIDs at will);
|
| 1192 |
|
| 1193 | * ``entropy_avail``: the pool's entropy count, in bits;
|
| 1194 |
|
| 1195 | * ``poolsize``: the entropy pool size, in bits;
|
| 1196 |
|
| 1197 | * ``urandom_min_reseed_secs``: obsolete (used to determine the minimum
|
| 1198 | number of seconds between urandom pool reseeding). This file is
|
| 1199 | writable for compatibility purposes, but writing to it has no effect
|
| 1200 | on any RNG behavior;
|
| 1201 |
|
| 1202 | * ``write_wakeup_threshold``: when the entropy count drops below this
|
| 1203 | (as a number of bits), processes waiting to write to ``/dev/random``
|
| 1204 | are woken up. This file is writable for compatibility purposes, but
|
| 1205 | writing to it has no effect on any RNG behavior.
|
| 1206 |
|
| 1207 |
|
| 1208 | randomize_va_space
|
| 1209 | ==================
|
| 1210 |
|
| 1211 | This option can be used to select the type of process address
|
| 1212 | space randomization that is used in the system, for architectures
|
| 1213 | that support this feature.
|
| 1214 |
|
| 1215 | == ===========================================================================
|
| 1216 | 0 Turn the process address space randomization off. This is the
|
| 1217 | default for architectures that do not support this feature anyways,
|
| 1218 | and kernels that are booted with the "norandmaps" parameter.
|
| 1219 |
|
| 1220 | 1 Make the addresses of mmap base, stack and VDSO page randomized.
|
| 1221 | This, among other things, implies that shared libraries will be
|
| 1222 | loaded to random addresses. Also for PIE-linked binaries, the
|
| 1223 | location of code start is randomized. This is the default if the
|
| 1224 | ``CONFIG_COMPAT_BRK`` option is enabled.
|
| 1225 |
|
| 1226 | 2 Additionally enable heap randomization. This is the default if
|
| 1227 | ``CONFIG_COMPAT_BRK`` is disabled.
|
| 1228 |
|
| 1229 | There are a few legacy applications out there (such as some ancient
|
| 1230 | versions of libc.so.5 from 1996) that assume that brk area starts
|
| 1231 | just after the end of the code+bss. These applications break when
|
| 1232 | start of the brk area is randomized. There are however no known
|
| 1233 | non-legacy applications that would be broken this way, so for most
|
| 1234 | systems it is safe to choose full randomization.
|
| 1235 |
|
| 1236 | Systems with ancient and/or broken binaries should be configured
|
| 1237 | with ``CONFIG_COMPAT_BRK`` enabled, which excludes the heap from process
|
| 1238 | address space randomization.
|
| 1239 | == ===========================================================================
|
| 1240 |
|
| 1241 |
|
| 1242 | reboot-cmd (SPARC only)
|
| 1243 | =======================
|
| 1244 |
|
| 1245 | ??? This seems to be a way to give an argument to the Sparc
|
| 1246 | ROM/Flash boot loader. Maybe to tell it what to do after
|
| 1247 | rebooting. ???
|
| 1248 |
|
| 1249 |
|
| 1250 | sched_energy_aware
|
| 1251 | ==================
|
| 1252 |
|
| 1253 | Enables/disables Energy Aware Scheduling (EAS). EAS starts
|
| 1254 | automatically on platforms where it can run (that is,
|
| 1255 | platforms with asymmetric CPU topologies and having an Energy
|
| 1256 | Model available). If your platform happens to meet the
|
| 1257 | requirements for EAS but you do not want to use it, change
|
| 1258 | this value to 0. On Non-EAS platforms, write operation fails and
|
| 1259 | read doesn't return anything.
|
| 1260 |
|
| 1261 | task_delayacct
|
| 1262 | ===============
|
| 1263 |
|
| 1264 | Enables/disables task delay accounting (see
|
| 1265 | Documentation/accounting/delay-accounting.rst. Enabling this feature incurs
|
| 1266 | a small amount of overhead in the scheduler but is useful for debugging
|
| 1267 | and performance tuning. It is required by some tools such as iotop.
|
| 1268 |
|
| 1269 | sched_schedstats
|
| 1270 | ================
|
| 1271 |
|
| 1272 | Enables/disables scheduler statistics. Enabling this feature
|
| 1273 | incurs a small amount of overhead in the scheduler but is
|
| 1274 | useful for debugging and performance tuning.
|
| 1275 |
|
| 1276 | sched_util_clamp_min
|
| 1277 | ====================
|
| 1278 |
|
| 1279 | Max allowed *minimum* utilization.
|
| 1280 |
|
| 1281 | Default value is 1024, which is the maximum possible value.
|
| 1282 |
|
| 1283 | It means that any requested uclamp.min value cannot be greater than
|
| 1284 | sched_util_clamp_min, i.e., it is restricted to the range
|
| 1285 | [0:sched_util_clamp_min].
|
| 1286 |
|
| 1287 | sched_util_clamp_max
|
| 1288 | ====================
|
| 1289 |
|
| 1290 | Max allowed *maximum* utilization.
|
| 1291 |
|
| 1292 | Default value is 1024, which is the maximum possible value.
|
| 1293 |
|
| 1294 | It means that any requested uclamp.max value cannot be greater than
|
| 1295 | sched_util_clamp_max, i.e., it is restricted to the range
|
| 1296 | [0:sched_util_clamp_max].
|
| 1297 |
|
| 1298 | sched_util_clamp_min_rt_default
|
| 1299 | ===============================
|
| 1300 |
|
| 1301 | By default Linux is tuned for performance. Which means that RT tasks always run
|
| 1302 | at the highest frequency and most capable (highest capacity) CPU (in
|
| 1303 | heterogeneous systems).
|
| 1304 |
|
| 1305 | Uclamp achieves this by setting the requested uclamp.min of all RT tasks to
|
| 1306 | 1024 by default, which effectively boosts the tasks to run at the highest
|
| 1307 | frequency and biases them to run on the biggest CPU.
|
| 1308 |
|
| 1309 | This knob allows admins to change the default behavior when uclamp is being
|
| 1310 | used. In battery powered devices particularly, running at the maximum
|
| 1311 | capacity and frequency will increase energy consumption and shorten the battery
|
| 1312 | life.
|
| 1313 |
|
| 1314 | This knob is only effective for RT tasks which the user hasn't modified their
|
| 1315 | requested uclamp.min value via sched_setattr() syscall.
|
| 1316 |
|
| 1317 | This knob will not escape the range constraint imposed by sched_util_clamp_min
|
| 1318 | defined above.
|
| 1319 |
|
| 1320 | For example if
|
| 1321 |
|
| 1322 | sched_util_clamp_min_rt_default = 800
|
| 1323 | sched_util_clamp_min = 600
|
| 1324 |
|
| 1325 | Then the boost will be clamped to 600 because 800 is outside of the permissible
|
| 1326 | range of [0:600]. This could happen for instance if a powersave mode will
|
| 1327 | restrict all boosts temporarily by modifying sched_util_clamp_min. As soon as
|
| 1328 | this restriction is lifted, the requested sched_util_clamp_min_rt_default
|
| 1329 | will take effect.
|
| 1330 |
|
| 1331 | seccomp
|
| 1332 | =======
|
| 1333 |
|
| 1334 | See Documentation/userspace-api/seccomp_filter.rst.
|
| 1335 |
|
| 1336 |
|
| 1337 | shmall
|
| 1338 | ======
|
| 1339 |
|
| 1340 | This parameter sets the total amount of shared memory pages that can be used
|
| 1341 | inside ipc namespace. The shared memory pages counting occurs for each ipc
|
| 1342 | namespace separately and is not inherited. Hence, ``shmall`` should always be at
|
| 1343 | least ``ceil(shmmax/PAGE_SIZE)``.
|
| 1344 |
|
| 1345 | If you are not sure what the default ``PAGE_SIZE`` is on your Linux
|
| 1346 | system, you can run the following command::
|
| 1347 |
|
| 1348 | # getconf PAGE_SIZE
|
| 1349 |
|
| 1350 | To reduce or disable the ability to allocate shared memory, you must create a
|
| 1351 | new ipc namespace, set this parameter to the required value and prohibit the
|
| 1352 | creation of a new ipc namespace in the current user namespace or cgroups can
|
| 1353 | be used.
|
| 1354 |
|
| 1355 | shmmax
|
| 1356 | ======
|
| 1357 |
|
| 1358 | This value can be used to query and set the run time limit
|
| 1359 | on the maximum shared memory segment size that can be created.
|
| 1360 | Shared memory segments up to 1Gb are now supported in the
|
| 1361 | kernel. This value defaults to ``SHMMAX``.
|
| 1362 |
|
| 1363 |
|
| 1364 | shmmni
|
| 1365 | ======
|
| 1366 |
|
| 1367 | This value determines the maximum number of shared memory segments.
|
| 1368 | 4096 by default (``SHMMNI``).
|
| 1369 |
|
| 1370 |
|
| 1371 | shm_rmid_forced
|
| 1372 | ===============
|
| 1373 |
|
| 1374 | Linux lets you set resource limits, including how much memory one
|
| 1375 | process can consume, via ``setrlimit(2)``. Unfortunately, shared memory
|
| 1376 | segments are allowed to exist without association with any process, and
|
| 1377 | thus might not be counted against any resource limits. If enabled,
|
| 1378 | shared memory segments are automatically destroyed when their attach
|
| 1379 | count becomes zero after a detach or a process termination. It will
|
| 1380 | also destroy segments that were created, but never attached to, on exit
|
| 1381 | from the process. The only use left for ``IPC_RMID`` is to immediately
|
| 1382 | destroy an unattached segment. Of course, this breaks the way things are
|
| 1383 | defined, so some applications might stop working. Note that this
|
| 1384 | feature will do you no good unless you also configure your resource
|
| 1385 | limits (in particular, ``RLIMIT_AS`` and ``RLIMIT_NPROC``). Most systems don't
|
| 1386 | need this.
|
| 1387 |
|
| 1388 | Note that if you change this from 0 to 1, already created segments
|
| 1389 | without users and with a dead originative process will be destroyed.
|
| 1390 |
|
| 1391 |
|
| 1392 | syscall_user_dispatch
|
| 1393 | =====================
|
| 1394 |
|
| 1395 | Controls whether userspace may arm Syscall User Dispatch via
|
| 1396 | ``prctl(PR_SET_SYSCALL_USER_DISPATCH, ...)`` or the
|
| 1397 | ``PTRACE_SET_SYSCALL_USER_DISPATCH_CONFIG`` ptrace request:
|
| 1398 |
|
| 1399 | == ===================================================================
|
| 1400 | 0 Arming syscall user dispatch is denied with ``-EPERM``. Tasks that
|
| 1401 | already armed it keep it, and disabling it is always permitted.
|
| 1402 | 1 (default) Arming syscall user dispatch is permitted.
|
| 1403 | == ===================================================================
|
| 1404 |
|
| 1405 | Only present when the kernel is built with ``CONFIG_SYSCALL_USER_DISPATCH``
|
| 1406 | and ``CONFIG_SYSCTL``.
|
| 1407 |
|
| 1408 |
|
| 1409 | sysctl_writes_strict
|
| 1410 | ====================
|
| 1411 |
|
| 1412 | Control how file position affects the behavior of updating sysctl values
|
| 1413 | via the ``/proc/sys`` interface:
|
| 1414 |
|
| 1415 | == ======================================================================
|
| 1416 | -1 Legacy per-write sysctl value handling, with no printk warnings.
|
| 1417 | Each write syscall must fully contain the sysctl value to be
|
| 1418 | written, and multiple writes on the same sysctl file descriptor
|
| 1419 | will rewrite the sysctl value, regardless of file position.
|
| 1420 | 0 Same behavior as above, but warn about processes that perform writes
|
| 1421 | to a sysctl file descriptor when the file position is not 0.
|
| 1422 | 1 (default) Respect file position when writing sysctl strings. Multiple
|
| 1423 | writes will append to the sysctl value buffer. Anything past the max
|
| 1424 | length of the sysctl value buffer will be ignored. Writes to numeric
|
| 1425 | sysctl entries must always be at file position 0 and the value must
|
| 1426 | be fully contained in the buffer sent in the write syscall.
|
| 1427 | == ======================================================================
|
| 1428 |
|
| 1429 |
|
| 1430 | softlockup_all_cpu_backtrace
|
| 1431 | ============================
|
| 1432 |
|
| 1433 | This value controls the soft lockup detector thread's behavior
|
| 1434 | when a soft lockup condition is detected as to whether or not
|
| 1435 | to gather further debug information. If enabled, each cpu will
|
| 1436 | be issued an NMI and instructed to capture stack trace.
|
| 1437 |
|
| 1438 | This feature is only applicable for architectures which support
|
| 1439 | NMI.
|
| 1440 |
|
| 1441 | = ============================================
|
| 1442 | 0 Do nothing. This is the default behavior.
|
| 1443 | 1 On detection capture more debug information.
|
| 1444 | = ============================================
|
| 1445 |
|
| 1446 |
|
| 1447 | softlockup_panic
|
| 1448 | =================
|
| 1449 |
|
| 1450 | This parameter can be used to control whether the kernel panics
|
| 1451 | when a soft lockup is detected.
|
| 1452 |
|
| 1453 | = ============================================
|
| 1454 | 0 Don't panic on soft lockup.
|
| 1455 | 1 Panic on soft lockup.
|
| 1456 | = ============================================
|
| 1457 |
|
| 1458 | This can also be set using the softlockup_panic kernel parameter.
|
| 1459 |
|
| 1460 |
|
| 1461 | soft_watchdog
|
| 1462 | =============
|
| 1463 |
|
| 1464 | This parameter can be used to control the soft lockup detector.
|
| 1465 |
|
| 1466 | = =================================
|
| 1467 | 0 Disable the soft lockup detector.
|
| 1468 | 1 Enable the soft lockup detector.
|
| 1469 | = =================================
|
| 1470 |
|
| 1471 | The soft lockup detector monitors CPUs for threads that are hogging the CPUs
|
| 1472 | without rescheduling voluntarily, and thus prevent the 'migration/N' threads
|
| 1473 | from running, causing the watchdog work fail to execute. The mechanism depends
|
| 1474 | on the CPUs ability to respond to timer interrupts which are needed for the
|
| 1475 | watchdog work to be queued by the watchdog timer function, otherwise the NMI
|
| 1476 | watchdog — if enabled — can detect a hard lockup condition.
|
| 1477 |
|
| 1478 |
|
| 1479 | split_lock_mitigate (x86 only)
|
| 1480 | ==============================
|
| 1481 |
|
| 1482 | On x86, each "split lock" imposes a system-wide performance penalty. On larger
|
| 1483 | systems, large numbers of split locks from unprivileged users can result in
|
| 1484 | denials of service to well-behaved and potentially more important users.
|
| 1485 |
|
| 1486 | The kernel mitigates these bad users by detecting split locks and imposing
|
| 1487 | penalties: forcing them to wait and only allowing one core to execute split
|
| 1488 | locks at a time.
|
| 1489 |
|
| 1490 | These mitigations can make those bad applications unbearably slow. Setting
|
| 1491 | split_lock_mitigate=0 may restore some application performance, but will also
|
| 1492 | increase system exposure to denial of service attacks from split lock users.
|
| 1493 |
|
| 1494 | = ===================================================================
|
| 1495 | 0 Disable the mitigation mode - just warns the split lock on kernel log
|
| 1496 | and exposes the system to denials of service from the split lockers.
|
| 1497 | 1 Enable the mitigation mode (this is the default) - penalizes the split
|
| 1498 | lockers with intentional performance degradation.
|
| 1499 | = ===================================================================
|
| 1500 |
|
| 1501 |
|
| 1502 | stack_erasing
|
| 1503 | =============
|
| 1504 |
|
| 1505 | This parameter can be used to control kernel stack erasing at the end
|
| 1506 | of syscalls for kernels built with ``CONFIG_KSTACK_ERASE``.
|
| 1507 |
|
| 1508 | That erasing reduces the information which kernel stack leak bugs
|
| 1509 | can reveal and blocks some uninitialized stack variable attacks.
|
| 1510 | The tradeoff is the performance impact: on a single CPU system kernel
|
| 1511 | compilation sees a 1% slowdown, other systems and workloads may vary.
|
| 1512 |
|
| 1513 | = ====================================================================
|
| 1514 | 0 Kernel stack erasing is disabled, KSTACK_ERASE_METRICS are not updated.
|
| 1515 | 1 Kernel stack erasing is enabled (default), it is performed before
|
| 1516 | returning to the userspace at the end of syscalls.
|
| 1517 | = ====================================================================
|
| 1518 |
|
| 1519 |
|
| 1520 | stop-a (SPARC only)
|
| 1521 | ===================
|
| 1522 |
|
| 1523 | Controls Stop-A:
|
| 1524 |
|
| 1525 | = ====================================
|
| 1526 | 0 Stop-A has no effect.
|
| 1527 | 1 Stop-A breaks to the PROM (default).
|
| 1528 | = ====================================
|
| 1529 |
|
| 1530 | Stop-A is always enabled on a panic, so that the user can return to
|
| 1531 | the boot PROM.
|
| 1532 |
|
| 1533 |
|
| 1534 | sysrq
|
| 1535 | =====
|
| 1536 |
|
| 1537 | See Documentation/admin-guide/sysrq.rst.
|
| 1538 |
|
| 1539 |
|
| 1540 | tainted
|
| 1541 | =======
|
| 1542 |
|
| 1543 | Non-zero if the kernel has been tainted. Numeric values, which can be
|
| 1544 | ORed together. The letters are seen in "Tainted" line of Oops reports.
|
| 1545 |
|
| 1546 | ====== ===== ==============================================================
|
| 1547 | 1 `(P)` proprietary module was loaded
|
| 1548 | 2 `(F)` module was force loaded
|
| 1549 | 4 `(S)` kernel running on an out of specification system
|
| 1550 | 8 `(R)` module was force unloaded
|
| 1551 | 16 `(M)` processor reported a Machine Check Exception (MCE)
|
| 1552 | 32 `(B)` bad page referenced or some unexpected page flags
|
| 1553 | 64 `(U)` taint requested by userspace application
|
| 1554 | 128 `(D)` kernel died recently, i.e. there was an OOPS or BUG
|
| 1555 | 256 `(A)` an ACPI table was overridden by user
|
| 1556 | 512 `(W)` kernel issued warning
|
| 1557 | 1024 `(C)` staging driver was loaded
|
| 1558 | 2048 `(I)` workaround for bug in platform firmware applied
|
| 1559 | 4096 `(O)` externally-built ("out-of-tree") module was loaded
|
| 1560 | 8192 `(E)` unsigned module was loaded
|
| 1561 | 16384 `(L)` soft lockup occurred
|
| 1562 | 32768 `(K)` kernel has been live patched
|
| 1563 | 65536 `(X)` Auxiliary taint, defined and used by for distros
|
| 1564 | 131072 `(T)` The kernel was built with the struct randomization plugin
|
| 1565 | ====== ===== ==============================================================
|
| 1566 |
|
| 1567 | See Documentation/admin-guide/tainted-kernels.rst for more information.
|
| 1568 |
|
| 1569 | Note:
|
| 1570 | writes to this sysctl interface will fail with ``EINVAL`` if the kernel is
|
| 1571 | booted with the command line option ``panic_on_taint=<bitmask>,nousertaint``
|
| 1572 | and any of the ORed together values being written to ``tainted`` match with
|
| 1573 | the bitmask declared on panic_on_taint.
|
| 1574 | See Documentation/admin-guide/kernel-parameters.rst for more details on
|
| 1575 | that particular kernel command line option and its optional
|
| 1576 | ``nousertaint`` switch.
|
| 1577 |
|
| 1578 | threads-max
|
| 1579 | ===========
|
| 1580 |
|
| 1581 | This value controls the maximum number of threads that can be created
|
| 1582 | using ``fork()``.
|
| 1583 |
|
| 1584 | During initialization the kernel sets this value such that even if the
|
| 1585 | maximum number of threads is created, the thread structures occupy only
|
| 1586 | a part (1/8th) of the available RAM pages.
|
| 1587 |
|
| 1588 | The minimum value that can be written to ``threads-max`` is 1.
|
| 1589 |
|
| 1590 | The maximum value that can be written to ``threads-max`` is given by the
|
| 1591 | constant ``FUTEX_TID_MASK`` (0x3fffffff).
|
| 1592 |
|
| 1593 | If a value outside of this range is written to ``threads-max`` an
|
| 1594 | ``EINVAL`` error occurs.
|
| 1595 |
|
| 1596 | timer_migration
|
| 1597 | ===============
|
| 1598 |
|
| 1599 | When set to a non-zero value, attempt to migrate high-resolution timers from
|
| 1600 | nohz isolated (nohz_full) to housekeeping CPUs.
|
| 1601 | See Documentation/admin-guide/cpu-isolation.rst
|
| 1602 | and Documentation/admin-guide/kernel-parameters.rst.
|
| 1603 |
|
| 1604 | Default is set (1).
|
| 1605 |
|
| 1606 | traceoff_on_warning
|
| 1607 | ===================
|
| 1608 |
|
| 1609 | When set, disables tracing (see Documentation/trace/ftrace.rst) when a
|
| 1610 | ``WARN()`` is hit.
|
| 1611 |
|
| 1612 |
|
| 1613 | tracepoint_printk
|
| 1614 | =================
|
| 1615 |
|
| 1616 | When tracepoints are sent to printk() (enabled by the ``tp_printk``
|
| 1617 | boot parameter), this entry provides runtime control::
|
| 1618 |
|
| 1619 | echo 0 > /proc/sys/kernel/tracepoint_printk
|
| 1620 |
|
| 1621 | will stop tracepoints from being sent to printk(), and::
|
| 1622 |
|
| 1623 | echo 1 > /proc/sys/kernel/tracepoint_printk
|
| 1624 |
|
| 1625 | will send them to printk() again.
|
| 1626 |
|
| 1627 | This only works if the kernel was booted with ``tp_printk`` enabled.
|
| 1628 |
|
| 1629 | See Documentation/admin-guide/kernel-parameters.rst and
|
| 1630 | Documentation/trace/boottime-trace.rst.
|
| 1631 |
|
| 1632 |
|
| 1633 | unaligned-trap
|
| 1634 | ==============
|
| 1635 |
|
| 1636 | On architectures where unaligned accesses cause traps, and where this
|
| 1637 | feature is supported (``CONFIG_SYSCTL_ARCH_UNALIGN_ALLOW``; currently,
|
| 1638 | ``arc``, ``parisc`` and ``loongarch``), controls whether unaligned traps
|
| 1639 | are caught and emulated (instead of failing).
|
| 1640 |
|
| 1641 | = ========================================================
|
| 1642 | 0 Do not emulate unaligned accesses.
|
| 1643 | 1 Emulate unaligned accesses. This is the default setting.
|
| 1644 | = ========================================================
|
| 1645 |
|
| 1646 | See also `ignore-unaligned-usertrap`_.
|
| 1647 |
|
| 1648 |
|
| 1649 | unknown_nmi_panic
|
| 1650 | =================
|
| 1651 |
|
| 1652 | The value in this file affects behavior of handling NMI. When the
|
| 1653 | value is non-zero, unknown NMI is trapped and then panic occurs. At
|
| 1654 | that time, kernel debugging information is displayed on console.
|
| 1655 |
|
| 1656 | NMI switch that most IA32 servers have fires unknown NMI up, for
|
| 1657 | example. If a system hangs up, try pressing the NMI switch.
|
| 1658 |
|
| 1659 |
|
| 1660 | unprivileged_bpf_disabled
|
| 1661 | =========================
|
| 1662 |
|
| 1663 | Writing 1 to this entry will disable unprivileged calls to ``bpf()``;
|
| 1664 | once disabled, calling ``bpf()`` without ``CAP_SYS_ADMIN`` or ``CAP_BPF``
|
| 1665 | will return ``-EPERM``. Once set to 1, this can't be cleared from the
|
| 1666 | running kernel anymore.
|
| 1667 |
|
| 1668 | Writing 2 to this entry will also disable unprivileged calls to ``bpf()``,
|
| 1669 | however, an admin can still change this setting later on, if needed, by
|
| 1670 | writing 0 or 1 to this entry.
|
| 1671 |
|
| 1672 | If ``BPF_UNPRIV_DEFAULT_OFF`` is enabled in the kernel config, then this
|
| 1673 | entry will default to 2 instead of 0.
|
| 1674 |
|
| 1675 | = =============================================================
|
| 1676 | 0 Unprivileged calls to ``bpf()`` are enabled
|
| 1677 | 1 Unprivileged calls to ``bpf()`` are disabled without recovery
|
| 1678 | 2 Unprivileged calls to ``bpf()`` are disabled
|
| 1679 | = =============================================================
|
| 1680 |
|
| 1681 |
|
| 1682 | warn_limit
|
| 1683 | ==========
|
| 1684 |
|
| 1685 | Number of kernel warnings after which the kernel should panic when
|
| 1686 | ``panic_on_warn`` is not set. Setting this to 0 disables checking
|
| 1687 | the warning count. Setting this to 1 has the same effect as setting
|
| 1688 | ``panic_on_warn=1``. The default value is 0.
|
| 1689 |
|
| 1690 |
|
| 1691 | watchdog
|
| 1692 | ========
|
| 1693 |
|
| 1694 | This parameter can be used to disable or enable the soft lockup detector
|
| 1695 | *and* the NMI watchdog (i.e. the hard lockup detector) at the same time.
|
| 1696 |
|
| 1697 | = ==============================
|
| 1698 | 0 Disable both lockup detectors.
|
| 1699 | 1 Enable both lockup detectors.
|
| 1700 | = ==============================
|
| 1701 |
|
| 1702 | The soft lockup detector and the NMI watchdog can also be disabled or
|
| 1703 | enabled individually, using the ``soft_watchdog`` and ``nmi_watchdog``
|
| 1704 | parameters.
|
| 1705 | If the ``watchdog`` parameter is read, for example by executing::
|
| 1706 |
|
| 1707 | cat /proc/sys/kernel/watchdog
|
| 1708 |
|
| 1709 | the output of this command (0 or 1) shows the logical OR of
|
| 1710 | ``soft_watchdog`` and ``nmi_watchdog``.
|
| 1711 |
|
| 1712 |
|
| 1713 | watchdog_cpumask
|
| 1714 | ================
|
| 1715 |
|
| 1716 | This value can be used to control on which cpus the watchdog may run.
|
| 1717 | The default cpumask is all possible cores, but if ``NO_HZ_FULL`` is
|
| 1718 | enabled in the kernel config, and cores are specified with the
|
| 1719 | ``nohz_full=`` boot argument, those cores are excluded by default.
|
| 1720 | Offline cores can be included in this mask, and if the core is later
|
| 1721 | brought online, the watchdog will be started based on the mask value.
|
| 1722 |
|
| 1723 | Typically this value would only be touched in the ``nohz_full`` case
|
| 1724 | to re-enable cores that by default were not running the watchdog,
|
| 1725 | if a kernel lockup was suspected on those cores.
|
| 1726 |
|
| 1727 | The argument value is the standard cpulist format for cpumasks,
|
| 1728 | so for example to enable the watchdog on cores 0, 2, 3, and 4 you
|
| 1729 | might say::
|
| 1730 |
|
| 1731 | echo 0,2-4 > /proc/sys/kernel/watchdog_cpumask
|
| 1732 |
|
| 1733 |
|
| 1734 | watchdog_thresh
|
| 1735 | ===============
|
| 1736 |
|
| 1737 | This value can be used to control the frequency of hrtimer and NMI
|
| 1738 | events and the soft and hard lockup thresholds. The default threshold
|
| 1739 | is 10 seconds.
|
| 1740 |
|
| 1741 | The softlockup threshold is (``2 * watchdog_thresh``). Setting this
|
| 1742 | tunable to zero will disable lockup detection altogether.
|