1 acpi= [HW,ACPI,X86,ARM64,RISCV64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
5 force -- enable ACPI if default was off
6 on -- enable ACPI but allow fallback to DT [arm64,riscv64]
7 off -- disable ACPI if default was on
8 noirq -- do not use ACPI for IRQ routing
9 strict -- Be less tolerant of platforms that are not
10 strictly ACPI specification compliant.
11 rsdt -- prefer RSDT over (default) XSDT
12 copy_dsdt -- copy DSDT to memory
13 For ARM64 and RISCV64, ONLY "acpi=off", "acpi=on" or
14 "acpi=force" are available
16 See also Documentation/power/runtime_pm.rst, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
25 { vendor | video | native | none }
26 If set to vendor, prefer vendor-specific driver
27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
28 of the ACPI video.ko driver.
29 If set to video, use the ACPI video.ko driver.
30 If set to native, use the device's native backlight mode.
31 If set to none, disable the ACPI backlight interface.
33 acpi_force_32bit_fadt_addr
34 force FADT to use 32 bit addresses rather than the
35 64 bit X_* addresses. Some firmware have broken 64
36 bit addresses for force ACPI ignore these and use
37 the older legacy 32 bit addresses.
39 acpica_no_return_repair [HW, ACPI]
40 Disable AML predefined validation mechanism
41 This mechanism can repair the evaluation result to make
42 the return objects more ACPI specification compliant.
43 This option is useful for developers to identify the
44 root cause of an AML interpreter issue when the issue
45 has something to do with the repair mechanism.
47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
51 debug output. Bits in debug_layer correspond to a
52 _COMPONENT in an ACPI source file, e.g.,
53 #define _COMPONENT ACPI_EVENTS
54 Bits in debug_level correspond to a level in
55 ACPI_DEBUG_PRINT statements, e.g.,
56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
57 The debug_level mask defaults to "info". See
58 Documentation/firmware-guide/acpi/debug.rst for more information about
59 debug layers and levels.
61 Enable processor driver info messages:
62 acpi.debug_layer=0x20000000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
116 Format: <byte> or <bitmap-list>
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
163 acpi_osi=!! # enable all built-in OS vendor
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig,
229 s4_nohwsig, old_ordering, nonvs,
230 sci_force_enable, nobl }
231 See Documentation/power/video.rst for information on
233 s3_beep is for debugging; it makes the PC's speaker beep
234 as soon as the kernel's real-mode entry point is called.
235 s4_hwsig causes the kernel to check the ACPI hardware
236 signature during resume from hibernation, and gracefully
237 refuse to resume if it has changed. This complies with
238 the ACPI specification but not with reality, since
239 Windows does not do this and many laptops do change it
240 on docking. So the default behaviour is to allow resume
241 and simply warn when the signature changes, unless the
242 s4_hwsig option is enabled.
243 s4_nohwsig prevents ACPI hardware signature from being
244 used (or even warned about) during resume.
245 old_ordering causes the ACPI 1.0 ordering of the _PTS
246 control method, with respect to putting devices into
247 low power states, to be enforced (the ACPI 2.0 ordering
248 of _PTS is used by default).
249 nonvs prevents the kernel from saving/restoring the
250 ACPI NVS memory during suspend/hibernation and resume.
251 sci_force_enable causes the kernel to set SCI_EN directly
252 on resume from S1/S3 (which is against the ACPI spec,
253 but some broken systems don't work without it).
254 nobl causes the internal blacklist of systems known to
255 behave incorrectly in some ways with respect to system
256 suspend and resume to be ignored (use wisely).
258 acpi_use_timer_override [HW,ACPI]
259 Use timer override. For some broken Nvidia NF5 boards
260 that require a timer override, but don't have HPET
262 add_efi_memmap [EFI; X86] Include EFI memory map in
263 kernel's map of available physical RAM.
266 { off | try_unsupported }
267 off: disable AGP support
268 try_unsupported: try to drive unsupported chipsets
269 (may crash computer or cause data corruption)
272 See Documentation/sound/alsa-configuration.rst
275 Allow the default userspace alignment fault handler
276 behaviour to be specified. Bit 0 enables warnings,
277 bit 1 enables fixups, and bit 2 sends a segfault.
279 align_va_addr= [X86-64]
280 Align virtual addresses by clearing slice [14:12] when
281 allocating a VMA at process creation time. This option
282 gives you up to 3% performance improvement on AMD F15h
283 machines (where it is enabled by default) for a
284 CPU-intensive style benchmark, and it can vary highly in
285 a microbenchmark depending on workload and compiler.
287 32: only for 32-bit processes
288 64: only for 64-bit processes
289 on: enable for both 32- and 64-bit processes
290 off: disable for both 32- and 64-bit processes
292 alloc_snapshot [FTRACE]
293 Allocate the ftrace snapshot buffer on boot up when the
294 main buffer is allocated. This is handy if debugging
295 and you need to use tracing_snapshot() on boot up, and
296 do not want to use tracing_snapshot_alloc() as it needs
297 to be done where GFP_KERNEL allocations are allowed.
299 allow_mismatched_32bit_el0 [ARM64]
300 Allow execve() of 32-bit applications and setting of the
301 PER_LINUX32 personality on systems where only a strict
302 subset of the CPUs support 32-bit EL0. When this
303 parameter is present, the set of CPUs supporting 32-bit
304 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
305 and hot-unplug operations may be restricted.
307 See Documentation/arch/arm64/asymmetric-32bit.rst for more
310 amd_iommu= [HW,X86-64]
311 Pass parameters to the AMD IOMMU driver in the system.
313 fullflush - Deprecated, equivalent to iommu.strict=1
314 off - do not initialize any AMD IOMMU found in
316 force_isolation - Force device isolation for all
317 devices. The IOMMU driver is not
318 allowed anymore to lift isolation
319 requirements as needed. This option
320 does not override iommu=pt
321 force_enable - Force enable the IOMMU on platforms known
322 to be buggy with IOMMU enabled. Use this
324 pgtbl_v1 - Use v1 page table for DMA-API (Default).
325 pgtbl_v2 - Use v2 page table for DMA-API.
326 irtcachedis - Disable Interrupt Remapping Table (IRT) caching.
328 amd_iommu_dump= [HW,X86-64]
329 Enable AMD IOMMU driver option to dump the ACPI table
330 for AMD IOMMU. With this option enabled, AMD IOMMU
331 driver will print ACPI tables for AMD IOMMU during
332 IOMMU initialization.
334 amd_iommu_intr= [HW,X86-64]
335 Specifies one of the following AMD IOMMU interrupt
337 legacy - Use legacy interrupt remapping mode.
338 vapic - Use virtual APIC mode, which allows IOMMU
339 to inject interrupts directly into guest.
340 This mode requires kvm-amd.avic=1.
341 (Default when IOMMU HW support is present.)
345 Do not enable amd_pstate as the default
346 scaling driver for the supported processors
348 Use amd_pstate with passive mode as a scaling driver.
349 In this mode autonomous selection is disabled.
350 Driver requests a desired performance level and platform
351 tries to match the same performance level if it is
352 satisfied by guaranteed performance level.
354 Use amd_pstate_epp driver instance as the scaling driver,
355 driver provides a hint to the hardware if software wants
356 to bias toward performance (0x0) or energy efficiency (0xff)
357 to the CPPC firmware. then CPPC power algorithm will
358 calculate the runtime workload and adjust the realtime cores
361 Activate guided autonomous mode. Driver requests minimum and
362 maximum performance level and the platform autonomously
363 selects a performance level in this range and appropriate
364 to the current workload.
366 amijoy.map= [HW,JOY] Amiga joystick support
367 Map of devices attached to JOY0DAT and JOY1DAT
369 See also Documentation/input/joydev/joystick.rst
371 analog.map= [HW,JOY] Analog joystick and gamepad support
372 Specifies type or capabilities of an analog joystick
373 connected to one of 16 gameports
374 Format: <type1>,<type2>,..<type16>
377 Power management functions (SPARCstation-4/5 + deriv.)
379 Disable APC CPU standby support. SPARCstation-Fox does
380 not play well with APC CPU idle - disable it if you have
381 APC and your system crashes randomly.
383 apic= [APIC,X86] Advanced Programmable Interrupt Controller
384 Change the output verbosity while booting
385 Format: { quiet (default) | verbose | debug }
386 Change the amount of debugging information output
387 when initialising the APIC and IO-APIC components.
388 For X86-32, this can also be used to specify an APIC
390 Format: apic=driver_name
391 Examples: apic=bigsmp
393 apic_extnmi= [APIC,X86] External NMI delivery setting
394 Format: { bsp (default) | all | none }
395 bsp: External NMI is delivered only to CPU 0
396 all: External NMIs are broadcast to all CPUs as a
398 none: External NMI is masked for all CPUs. This is
399 useful so that a dump capture kernel won't be
403 See Documentation/networking/ipv6.rst.
405 apm= [APM] Advanced Power Management
406 See header of arch/x86/kernel/apm_32.c.
408 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
409 Format: { "0" | "1" }
410 See security/apparmor/Kconfig help text
413 Default value is set via kernel config option.
415 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
416 Format: <io>,<irq>,<nodeID>
418 arm64.nobti [ARM64] Unconditionally disable Branch Target
419 Identification support
421 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory
422 Set instructions support
424 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
427 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
430 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix
433 arm64.nosve [ARM64] Unconditionally disable Scalable Vector
438 atarimouse= [HW,MOUSE] Atari Mouse
440 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
441 EzKey and similar keyboards
443 atkbd.reset= [HW] Reset keyboard during initialization
445 atkbd.set= [HW] Select keyboard code set
446 Format: <int> (2 = AT (default), 3 = PS/2)
448 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
451 atkbd.softraw= [HW] Choose between synthetic and real raw mode
452 Format: <bool> (0 = real, 1 = synthetic (default))
454 atkbd.softrepeat= [HW]
455 Use software keyboard repeat
457 audit= [KNL] Enable the audit sub-system
458 Format: { "0" | "1" | "off" | "on" }
459 0 | off - kernel audit is disabled and can not be
460 enabled until the next reboot
461 unset - kernel audit is initialized but disabled and
462 will be fully enabled by the userspace auditd.
463 1 | on - kernel audit is initialized and partially
464 enabled, storing at most audit_backlog_limit
465 messages in RAM until it is fully enabled by the
469 audit_backlog_limit= [KNL] Set the audit queue size limit.
470 Format: <int> (must be >=0)
473 bau= [X86_UV] Enable the BAU on SGI UV. The default
474 behavior is to disable the BAU (i.e. bau=0).
475 Format: { "0" | "1" }
478 unset - Disable the BAU.
480 baycom_epp= [HW,AX25]
483 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
485 See header of drivers/net/hamradio/baycom_par.c.
487 baycom_ser_fdx= [HW,AX25]
488 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
489 Format: <io>,<irq>,<mode>[,<baud>]
490 See header of drivers/net/hamradio/baycom_ser_fdx.c.
492 baycom_ser_hdx= [HW,AX25]
493 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
494 Format: <io>,<irq>,<mode>
495 See header of drivers/net/hamradio/baycom_ser_hdx.c.
498 Disable BERT OS support on buggy BIOSes.
500 bgrt_disable [ACPI][X86]
501 Disable BGRT to avoid flickering OEM logo.
503 blkdevparts= Manual partition parsing of block device(s) for
504 embedded devices based on command line input.
505 See Documentation/block/cmdline-partition.rst
507 boot_delay= Milliseconds to delay each printk during boot.
508 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled,
509 and you may also have to specify "lpj=". Boot_delay
510 values larger than 10 seconds (10000) are assumed
511 erroneous and ignored.
515 Extended command line options can be added to an initrd
516 and this will cause the kernel to look for it.
518 See Documentation/admin-guide/bootconfig.rst
520 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
521 bttv.radio= Most important insmod options are available as
523 bttv.pll= See Documentation/admin-guide/media/bttv.rst
526 bulk_remove=off [PPC] This parameter disables the use of the pSeries
527 firmware feature for flushing multiple hpte entries
530 c101= [NET] Moxa C101 synchronous serial card
532 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
533 Sometimes CPU hardware bugs make them report the cache
534 size incorrectly. The kernel will attempt work arounds
535 to fix known problems, but for some CPUs it is not
536 possible to determine what the correct size should be.
537 This option provides an override for these situations.
540 [NET] Specifies amount of time (in seconds) that
541 the kernel should wait for a network carrier. By default
542 it waits 120 seconds.
544 ca_keys= [KEYS] This parameter identifies a specific key(s) on
545 the system trusted keyring to be used for certificate
547 format: { id:<keyid> | builtin }
549 cca= [MIPS] Override the kernel pages' cache coherency
550 algorithm. Accepted values range from 0 to 7
551 inclusive. See arch/mips/include/asm/pgtable-bits.h
552 for platform specific values (SB1, Loongson3 and
555 ccw_timeout_log [S390]
556 See Documentation/arch/s390/common_io.rst for details.
558 cgroup_disable= [KNL] Disable a particular controller or optional feature
559 Format: {name of the controller(s) or feature(s) to disable}
560 The effects of cgroup_disable=foo are:
561 - foo isn't auto-mounted if you mount all cgroups in
563 - foo isn't visible as an individually mountable
565 - if foo is an optional feature then the feature is
566 disabled and corresponding cgroup files are not
568 {Currently only "memory" controller deal with this and
569 cut the overhead, others just disable the usage. So
570 only cgroup_disable=memory is actually worthy}
571 Specifying "pressure" disables per-cgroup pressure
572 stall information accounting feature
574 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
575 Format: { { controller | "all" | "named" }
576 [,{ controller | "all" | "named" }...] }
577 Like cgroup_disable, but only applies to cgroup v1;
578 the blacklisted controllers remain available in cgroup2.
579 "all" blacklists all controllers and "named" disables
580 named mounts. Specifying both "all" and "named" disables
583 cgroup_favordynmods= [KNL] Enable or Disable favordynmods.
584 Format: { "true" | "false" }
585 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS.
587 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
589 nosocket -- Disable socket memory accounting.
590 nokmem -- Disable kernel memory accounting.
591 nobpf -- Disable BPF memory accounting.
593 checkreqprot= [SELINUX] Set initial checkreqprot flag value.
594 Format: { "0" | "1" }
595 See security/selinux/Kconfig help text.
596 0 -- check protection applied by kernel (includes
597 any implied execute protection).
598 1 -- check protection requested by application.
599 Default value is set via a kernel config option.
600 Value can be changed at runtime via
601 /sys/fs/selinux/checkreqprot.
602 Setting checkreqprot to 1 is deprecated.
605 See Documentation/arch/s390/common_io.rst for details.
607 clearcpuid=X[,X...] [X86]
608 Disable CPUID feature X for the kernel. See
609 arch/x86/include/asm/cpufeatures.h for the valid bit
610 numbers X. Note the Linux-specific bits are not necessarily
611 stable over kernel options, but the vendor-specific
613 X can also be a string as appearing in the flags: line
614 in /proc/cpuinfo which does not have the above
615 instability issue. However, not all features have names
617 Note that using this option will taint your kernel.
618 Also note that user programs calling CPUID directly
619 or using the feature without checking anything
620 will still see it. This just prevents it from
621 being used by the kernel or shown in /proc/cpuinfo.
622 Also note the kernel might malfunction if you disable
627 Prevents the clock framework from automatically gating
628 clocks that have not been explicitly enabled by a Linux
629 device driver but are enabled in hardware at reset or
630 by the bootloader/firmware. Note that this does not
631 force such clocks to be always-on nor does it reserve
632 those clocks in any way. This parameter is useful for
633 debug and development, but should not be needed on a
634 platform with proper driver support. For more
635 information, see Documentation/driver-api/clk.rst.
637 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
639 Forces specified clocksource (if available) to be used
640 when calculating gettimeofday(). If specified
641 clocksource is not available, it defaults to PIT.
642 Format: { pit | tsc | cyclone | pmtmr }
644 clocksource= Override the default clocksource
646 Override the default clocksource and use the clocksource
647 with the name specified.
648 Some clocksource names to choose from, depending on
650 [all] jiffies (this is the base, fallback clocksource)
652 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
653 pxa_timer,timer3,32k_counter,timer0_1
654 [X86-32] pit,hpet,tsc;
655 scx200_hrt on Geode; cyclone on IBM x440
663 clocksource.arm_arch_timer.evtstrm=
666 Enable/disable the eventstream feature of the ARM
667 architected timer so that code using WFE-based polling
668 loops can be debugged more effectively on production
671 clocksource.max_cswd_read_retries= [KNL]
672 Number of clocksource_watchdog() retries due to
673 external delays before the clock will be marked
674 unstable. Defaults to two retries, that is,
675 three attempts to read the clock under test.
677 clocksource.verify_n_cpus= [KNL]
678 Limit the number of CPUs checked for clocksources
679 marked with CLOCK_SOURCE_VERIFY_PERCPU that
680 are marked unstable due to excessive skew.
681 A negative value says to check all CPUs, while
682 zero says not to check any. Values larger than
683 nr_cpu_ids are silently truncated to nr_cpu_ids.
684 The actual CPUs are chosen randomly, with
685 no replacement if the same CPU is chosen twice.
687 clocksource-wdtest.holdoff= [KNL]
688 Set the time in seconds that the clocksource
689 watchdog test waits before commencing its tests.
690 Defaults to zero when built as a module and to
691 10 seconds when built into the kernel.
693 cma=nn[MG]@[start[MG][-end[MG]]]
695 Sets the size of kernel global memory area for
696 contiguous memory allocations and optionally the
697 placement constraint by the physical address range of
698 memory allocations. A value of 0 disables CMA
699 altogether. For more information, see
700 kernel/dma/contiguous.c
704 Sets the size of kernel per-numa memory area for
705 contiguous memory allocations. A value of 0 disables
706 per-numa CMA altogether. And If this option is not
707 specified, the default value is 0.
708 With per-numa CMA enabled, DMA users on node nid will
709 first try to allocate buffer from the pernuma area
710 which is located in node nid, if the allocation fails,
711 they will fallback to the global default memory area.
713 numa_cma=<node>:nn[MG][,<node>:nn[MG]]
715 Sets the size of kernel numa memory area for
716 contiguous memory allocations. It will reserve CMA
717 area for the specified node.
719 With numa CMA enabled, DMA users on node nid will
720 first try to allocate buffer from the numa area
721 which is located in node nid, if the allocation fails,
722 they will fallback to the global default memory area.
724 cmo_free_hint= [PPC] Format: { yes | no }
725 Specify whether pages are marked as being inactive
726 when they are freed. This is used in CMO environments
727 to determine OS memory pressure for page stealing by
731 coherent_pool=nn[KMG] [ARM,KNL]
732 Sets the size of memory pool for coherent, atomic dma
733 allocations, by default set to 256K.
735 com20020= [HW,NET] ARCnet - COM20020 chipset
737 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
739 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
743 ARCnet - COM90xx chipset (memory-mapped buffers)
744 Format: <io>[,<irq>[,<memstart>]]
746 condev= [HW,S390] console device
749 con3215_drop= [S390] 3215 console drop mode.
751 When set to true, drop data on the 3215 console when
752 the console buffer is full. In this case the
753 operator using a 3270 terminal emulator (for example
754 x3270) does not have to enter the clear key for the
755 console output to advance and the kernel to continue.
756 This leads to a much faster boot time when a 3270
757 terminal emulator is active. If no 3270 terminal
758 emulator is used, this parameter has no effect.
760 console= [KNL] Output console device and options.
762 tty<n> Use the virtual console device <n>.
766 Use the specified serial port. The options are of
767 the form "bbbbpnf", where "bbbb" is the baud rate,
768 "p" is parity ("n", "o", or "e"), "n" is number of
769 bits, and "f" is flow control ("r" for RTS or
770 omit it). Default is "9600n8".
772 See Documentation/admin-guide/serial-console.rst for more
774 Documentation/networking/netconsole.rst for an
777 uart[8250],io,<addr>[,options]
778 uart[8250],mmio,<addr>[,options]
779 uart[8250],mmio16,<addr>[,options]
780 uart[8250],mmio32,<addr>[,options]
781 uart[8250],0x<addr>[,options]
782 Start an early, polled-mode console on the 8250/16550
783 UART at the specified I/O port or MMIO address,
784 switching to the matching ttyS device later.
785 MMIO inter-register address stride is either 8-bit
786 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
787 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
788 to be equivalent to 'mmio'. 'options' are specified in
789 the same format described for ttyS above; if unspecified,
790 the h/w is not re-initialized.
792 hvc<n> Use the hypervisor console device <n>. This is for
793 both Xen and PowerPC hypervisors.
796 Use to disable console output, i.e., to have kernel
797 console messages discarded.
798 This must be the only console= parameter used on the
801 If the device connected to the port is not a TTY but a braille
802 device, prepend "brl," before the device type, for instance
804 For now, only VisioBraille is supported.
807 [KNL] Change console messages format
809 By default we print messages on consoles in
810 "[time stamp] text\n" format (time stamp may not be
811 printed, depending on CONFIG_PRINTK_TIME or
812 `printk_time' param).
814 Switch to syslog format: "<%u>[time stamp] text\n"
815 IOW, each message will have a facility and loglevel
816 prefix. The format is similar to one used by syslog()
817 syscall, or to executing "dmesg -S --raw" or to reading
820 consoleblank= [KNL] The console blank (screen saver) timeout in
821 seconds. A value of 0 disables the blank timer.
825 [KNL] Change the default value for
826 /proc/<pid>/coredump_filter.
827 See also Documentation/filesystems/proc.rst.
829 coresight_cpu_debug.enable
832 Enable/disable the CPU sampling based debugging.
833 0: default value, disable debugging
834 1: enable debugging at boot time
836 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
838 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
840 cpuidle.off=1 [CPU_IDLE]
841 disable the cpuidle sub-system
844 [CPU_IDLE] Name of the cpuidle governor to use.
846 cpufreq.off=1 [CPU_FREQ]
847 disable the cpufreq sub-system
849 cpufreq.default_governor=
850 [CPU_FREQ] Name of the default cpufreq governor or
851 policy to use. This governor must be registered in the
852 kernel before the cpufreq driver probes.
855 [X86] Delay for N microsec between assert and de-assert
856 of APIC INIT to start processors. This delay occurs
857 on every CPU online, such as boot, and resume from suspend.
861 [SMP] Enable/disable parallel bringup of secondary CPUs
863 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise
864 the parameter has no effect.
866 crash_kexec_post_notifiers
867 Run kdump after running panic-notifiers and dumping
868 kmsg. This only for the users who doubt kdump always
869 succeeds in any situation.
870 Note that this also increases risks of kdump failure,
871 because some panic notifiers can make the crashed
872 kernel more unstable.
874 crashkernel=size[KMG][@offset[KMG]]
875 [KNL] Using kexec, Linux can switch to a 'crash kernel'
876 upon panic. This parameter reserves the physical
877 memory region [offset, offset + size] for that kernel
878 image. If '@offset' is omitted, then a suitable offset
879 is selected automatically.
880 [KNL, X86-64, ARM64, RISCV] Select a region under 4G first, and
881 fall back to reserve region above 4G when '@offset'
882 hasn't been specified.
883 See Documentation/admin-guide/kdump/kdump.rst for further details.
885 crashkernel=range1:size1[,range2:size2,...][@offset]
886 [KNL] Same as above, but depends on the memory
887 in the running system. The syntax of range is
888 start-[end] where start and end are both
889 a memory unit (amount[KMG]). See also
890 Documentation/admin-guide/kdump/kdump.rst for an example.
892 crashkernel=size[KMG],high
893 [KNL, X86-64, ARM64, RISCV] range could be above 4G.
894 Allow kernel to allocate physical memory region from top,
895 so could be above 4G if system have more than 4G ram
896 installed. Otherwise memory region will be allocated
897 below 4G, if available.
898 It will be ignored if crashkernel=X is specified.
899 crashkernel=size[KMG],low
900 [KNL, X86-64, ARM64, RISCV] range under 4G. When crashkernel=X,high
901 is passed, kernel could allocate physical memory region
902 above 4G, that cause second kernel crash on system
903 that require some amount of low memory, e.g. swiotlb
904 requires at least 64M+32K low memory, also enough extra
905 low memory is needed to make sure DMA buffers for 32-bit
906 devices won't run out. Kernel would try to allocate
907 default size of memory below 4G automatically. The default
908 size is platform dependent.
909 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB)
912 This one lets the user specify own low range under 4G
913 for second kernel instead.
914 0: to disable low allocation.
915 It will be ignored when crashkernel=X,high is not used
916 or memory reserved is below 4G.
919 [KNL] Disable crypto self-tests
924 cs89x0_media= [HW,NET]
925 Format: { rj45 | aui | bnc }
927 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU
928 function call handling. When switched on,
929 additional debug data is printed to the console
930 in case a hanging CPU is detected, and that
931 CPU is pinged again in order to try to resolve
932 the hang situation. The default value of this
933 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT
937 See header of drivers/s390/block/dasd_devmap.c.
939 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
940 (one device per port)
941 Format: <port#>,<type>
942 See also Documentation/input/devices/joystick-parport.rst
944 debug [KNL] Enable kernel debugging (events log level).
947 [KNL] Enable printing [hashed] pointers early in the
948 boot sequence. If enabled, we use a weak hash instead
949 of siphash to hash pointers. Use this option if you are
950 seeing instances of '(___ptrval___)') and need to see a
951 value (hashed pointer) instead. Cryptographically
952 insecure, please do not use on production kernels.
955 [KNL] verbose locking self-tests
957 Print debugging info while doing the locking API
959 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
960 (no extra messages), setting it to -1 (all bits set)
961 will print _a_lot_ more information - normally only
962 useful to lockdep developers.
964 debug_objects [KNL] Enable object debugging
966 debug_guardpage_minorder=
967 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
968 parameter allows control of the order of pages that will
969 be intentionally kept free (and hence protected) by the
970 buddy allocator. Bigger value increase the probability
971 of catching random memory corruption, but reduce the
972 amount of memory for normal system use. The maximum
973 possible value is MAX_ORDER/2. Setting this parameter
974 to 1 or 2 should be enough to identify most random
975 memory corruption problems caused by bugs in kernel or
976 driver code when a CPU writes to (or reads from) a
977 random memory location. Note that there exists a class
978 of memory corruptions problems caused by buggy H/W or
979 F/W or by drivers badly programming DMA (basically when
980 memory is written at bus level and the CPU MMU is
981 bypassed) which are not detectable by
982 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
983 tracking down these problems.
986 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
987 enables the feature at boot time. By default, it is
988 disabled and the system will work mostly the same as a
989 kernel built without CONFIG_DEBUG_PAGEALLOC.
990 Note: to get most of debug_pagealloc error reports, it's
991 useful to also enable the page_owner functionality.
992 on: enable the feature
994 debugfs= [KNL] This parameter enables what is exposed to userspace
995 and debugfs internal clients.
996 Format: { on, no-mount, off }
997 on: All functions are enabled.
999 Filesystem is not registered but kernel clients can
1000 access APIs and a crashkernel can be used to read
1001 its content. There is nothing to mount.
1002 off: Filesystem is not registered and clients
1003 get a -EPERM as result when trying to register files
1004 or directories within debugfs.
1005 This is equivalent of the runtime functionality if
1006 debugfs was not enabled in the kernel at all.
1007 Default value is set in build-time with a kernel configuration.
1009 debugpat [X86] Enable PAT debugging
1012 [HW] The size of the default HugeTLB page. This is
1013 the size represented by the legacy /proc/ hugepages
1014 APIs. In addition, this is the default hugetlb size
1015 used for shmget(), mmap() and mounting hugetlbfs
1016 filesystems. If not specified, defaults to the
1017 architecture's default huge page size. Huge page
1018 sizes are architecture dependent. See also
1019 Documentation/admin-guide/mm/hugetlbpage.rst.
1022 deferred_probe_timeout=
1023 [KNL] Debugging option to set a timeout in seconds for
1024 deferred probe to give up waiting on dependencies to
1025 probe. Only specific dependencies (subsystems or
1026 drivers) that have opted in will be ignored. A timeout
1027 of 0 will timeout at the end of initcalls. If the time
1028 out hasn't expired, it'll be restarted by each
1029 successful driver registration. This option will also
1030 dump out devices still on the deferred probe list after
1033 delayacct [KNL] Enable per-task delay accounting
1035 dell_smm_hwmon.ignore_dmi=
1036 [HW] Continue probing hardware even if DMI data
1037 indicates that the driver is running on unsupported
1040 dell_smm_hwmon.force=
1041 [HW] Activate driver even if SMM BIOS signature does
1042 not match list of supported models and enable otherwise
1043 blacklisted features.
1045 dell_smm_hwmon.power_status=
1046 [HW] Report power status in /proc/i8k
1047 (disabled by default).
1049 dell_smm_hwmon.restricted=
1050 [HW] Allow controlling fans only if SYS_ADMIN
1053 dell_smm_hwmon.fan_mult=
1054 [HW] Factor to multiply fan speed with.
1056 dell_smm_hwmon.fan_max=
1057 [HW] Maximum configurable fan speed.
1060 Format: { on | off | def_only | inf_only | always }
1061 on: s390 zlib hardware support for compression on
1062 level 1 and decompression (default)
1063 off: No s390 zlib hardware support
1064 def_only: s390 zlib hardware support for deflate
1065 only (compression on level 1)
1066 inf_only: s390 zlib hardware support for inflate
1067 only (decompression)
1068 always: Same as 'on' but ignores the selected compression
1069 level always using hardware support (used for debugging)
1071 dhash_entries= [KNL]
1072 Set number of hash buckets for dentry cache.
1074 disable_1tb_segments [PPC]
1075 Disables the use of 1TB hash page table segments. This
1076 causes the kernel to fall back to 256MB segments which
1077 can be useful when debugging issues that require an SLB
1081 See Documentation/networking/ipv6.rst.
1084 Disable RADIX MMU mode on POWER9
1087 Disable TLBIE instruction. Currently does not work
1088 with KVM, with HASH MMU, or with coherent accelerators.
1090 disable_cpu_apicid= [X86,APIC,SMP]
1092 The number of initial APIC ID for the
1093 corresponding CPU to be disabled at boot,
1094 mostly used for the kdump 2nd kernel to
1095 disable BSP to wake up multiple CPUs without
1096 causing system reset or hang due to sending
1097 INIT from AP to BSP.
1099 disable_ddw [PPC/PSERIES]
1100 Disable Dynamic DMA Window support. Use this
1101 to workaround buggy firmware.
1103 disable_ipv6= [IPV6]
1104 See Documentation/networking/ipv6.rst.
1106 disable_mtrr_cleanup [X86]
1107 The kernel tries to adjust MTRR layout from continuous
1108 to discrete, to make X server driver able to add WB
1109 entry later. This parameter disables that.
1111 disable_mtrr_trim [X86, Intel and AMD only]
1112 By default the kernel will trim any uncacheable
1113 memory out of your available memory pool based on
1114 MTRR settings. This parameter disables that behavior,
1115 possibly causing your machine to run very slowly.
1117 disable_timer_pin_1 [X86]
1118 Disable PIN 1 of APIC timer
1119 Can be useful to work around chipset bugs.
1121 dis_ucode_ldr [X86] Disable the microcode loader.
1123 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1124 this option disables the debugging code at boot.
1126 dma_debug_entries=<number>
1127 This option allows to tune the number of preallocated
1128 entries for DMA-API debugging code. One entry is
1129 required per DMA-API allocation. Use this if the
1130 DMA-API debugging code disables itself because the
1131 architectural default is too low.
1133 dma_debug_driver=<driver_name>
1134 With this option the DMA-API debugging driver
1135 filter feature can be enabled at boot time. Just
1136 pass the driver to filter for as the parameter.
1137 The filter can be disabled or changed to another
1138 driver later using sysfs.
1140 driver_async_probe= [KNL]
1141 List of driver names to be probed asynchronously. *
1142 matches with all driver names. If * is specified, the
1143 rest of the listed driver names are those that will NOT
1145 Format: <driver_name1>,<driver_name2>...
1147 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1148 Broken monitors, graphic adapters, KVMs and EDIDless
1149 panels may send no or incorrect EDID data sets.
1150 This parameter allows to specify an EDID data sets
1151 in the /lib/firmware directory that are used instead.
1152 Generic built-in EDID data sets are used, if one of
1153 edid/1024x768.bin, edid/1280x1024.bin,
1154 edid/1680x1050.bin, or edid/1920x1080.bin is given
1155 and no file with the same name exists. Details and
1156 instructions how to build your own EDID data are
1157 available in Documentation/admin-guide/edid.rst. An EDID
1158 data set will only be used for a particular connector,
1159 if its name and a colon are prepended to the EDID
1160 name. Each connector may use a unique EDID data
1161 set by separating the files with a comma. An EDID
1162 data set with no connector name will be used for
1163 any connectors not explicitly specified.
1168 Format: {"off" | "known"}
1169 Control how the dt_cpu_ftrs device-tree binding is
1170 used for CPU feature discovery and setup (if it
1172 off: Do not use it, fall back to legacy cpu table.
1173 known: Do not pass through unknown features to guests
1174 or userspace, only those that the kernel is aware of.
1176 dump_apple_properties [X86]
1177 Dump name and content of EFI device properties on
1178 x86 Macs. Useful for driver authors to determine
1179 what data is available or for reverse-engineering.
1181 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1182 <module>.dyndbg[="val"]
1183 Enable debug messages at boot time. See
1184 Documentation/admin-guide/dynamic-debug-howto.rst
1187 early_ioremap_debug [KNL]
1188 Enable debug messages in early_ioremap support. This
1189 is useful for tracking down temporary early mappings
1190 which are not unmapped.
1192 earlycon= [KNL] Output early console device and options.
1194 When used with no options, the early console is
1195 determined by stdout-path property in device tree's
1196 chosen node or the ACPI SPCR table if supported by
1199 cdns,<addr>[,options]
1200 Start an early, polled-mode console on a Cadence
1201 (xuartps) serial port at the specified address. Only
1202 supported option is baud rate. If baud rate is not
1203 specified, the serial port must already be setup and
1206 uart[8250],io,<addr>[,options[,uartclk]]
1207 uart[8250],mmio,<addr>[,options[,uartclk]]
1208 uart[8250],mmio32,<addr>[,options[,uartclk]]
1209 uart[8250],mmio32be,<addr>[,options[,uartclk]]
1210 uart[8250],0x<addr>[,options]
1211 Start an early, polled-mode console on the 8250/16550
1212 UART at the specified I/O port or MMIO address.
1213 MMIO inter-register address stride is either 8-bit
1214 (mmio) or 32-bit (mmio32 or mmio32be).
1215 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1216 to be equivalent to 'mmio'. 'options' are specified
1217 in the same format described for "console=ttyS<n>"; if
1218 unspecified, the h/w is not initialized. 'uartclk' is
1219 the uart clock frequency; if unspecified, it is set
1220 to 'BASE_BAUD' * 16.
1224 Start an early, polled-mode console on a pl011 serial
1225 port at the specified address. The pl011 serial port
1226 must already be setup and configured. Options are not
1227 yet supported. If 'mmio32' is specified, then only
1228 the driver will use only 32-bit accessors to read/write
1229 the device registers.
1232 Start an early console on a litex serial port at the
1233 specified address. The serial port must already be
1234 setup and configured. Options are not yet supported.
1237 Start an early, polled-mode console on a meson serial
1238 port at the specified address. The serial port must
1239 already be setup and configured. Options are not yet
1243 Start an early, polled-mode console on an msm serial
1244 port at the specified address. The serial port
1245 must already be setup and configured. Options are not
1248 msm_serial_dm,<addr>
1249 Start an early, polled-mode console on an msm serial
1250 dm port at the specified address. The serial port
1251 must already be setup and configured. Options are not
1255 Start an early, polled-mode console on a serial port
1256 of an Actions Semi SoC, such as S500 or S900, at the
1257 specified address. The serial port must already be
1258 setup and configured. Options are not yet supported.
1261 Start an early, polled-mode console on a serial port
1262 of an RDA Micro SoC, such as RDA8810PL, at the
1263 specified address. The serial port must already be
1264 setup and configured. Options are not yet supported.
1267 Use RISC-V SBI (Supervisor Binary Interface) for early
1270 smh Use ARM semihosting calls for early console.
1278 Use early console provided by serial driver available
1279 on Samsung SoCs, requires selecting proper type and
1280 a correct base address of the selected UART port. The
1281 serial port must already be setup and configured.
1282 Options are not yet supported.
1285 Start an early, polled-mode console on a lantiq serial
1286 (lqasc) port at the specified address. The serial port
1287 must already be setup and configured. Options are not
1292 Use early console provided by Freescale LP UART driver
1293 found on Freescale Vybrid and QorIQ LS1021A processors.
1294 A valid base address must be provided, and the serial
1295 port must already be setup and configured.
1299 Start an early, polled-mode, output-only console on the
1300 Freescale i.MX UART at the specified address. The UART
1301 must already be setup and configured.
1304 Start an early, polled-mode console on the
1305 Armada 3700 serial port at the specified
1306 address. The serial port must already be setup
1307 and configured. Options are not yet supported.
1310 Start an early, polled-mode console on a Qualcomm
1311 Generic Interface (GENI) based serial port at the
1312 specified address. The serial port must already be
1313 setup and configured. Options are not yet supported.
1316 Start an early, unaccelerated console on the EFI
1317 memory mapped framebuffer (if available). On cache
1318 coherent non-x86 systems that use system memory for
1319 the framebuffer, pass the 'ram' option so that it is
1320 mapped with the correct attributes.
1323 Use early console provided by Freescale LINFlexD UART
1324 serial driver for NXP S32V234 SoCs. A valid base
1325 address must be provided, and the serial port must
1326 already be setup and configured.
1328 earlyprintk= [X86,SH,ARM,M68k,S390]
1332 earlyprintk=serial[,ttySn[,baudrate]]
1333 earlyprintk=serial[,0x...[,baudrate]]
1334 earlyprintk=ttySn[,baudrate]
1335 earlyprintk=dbgp[debugController#]
1336 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1337 earlyprintk=xdbc[xhciController#]
1340 earlyprintk is useful when the kernel crashes before
1341 the normal console is initialized. It is not enabled by
1342 default because it has some cosmetic problems.
1344 Append ",keep" to not disable it when the real console
1347 Only one of vga, serial, or usb debug port can
1350 Currently only ttyS0 and ttyS1 may be specified by
1351 name. Other I/O ports may be explicitly specified
1352 on some architectures (x86 and arm at least) by
1353 replacing ttySn with an I/O port address, like this:
1354 earlyprintk=serial,0x1008,115200
1355 You can find the port for a given device in
1356 /proc/tty/driver/serial:
1357 2: uart:ST16650V2 port:00001008 irq:18 ...
1359 Interaction with the standard serial driver is not
1362 The VGA output is eventually overwritten by
1365 The xen option can only be used in Xen domains.
1367 The sclp output can only be used on s390.
1369 The bios output can only be used on SuperH.
1371 The optional "force" to "pciserial" enables use of a
1372 PCI device even when its classcode is not of the
1375 edac_report= [HW,EDAC] Control how to report EDAC event
1376 Format: {"on" | "off" | "force"}
1377 on: enable EDAC to report H/W event. May be overridden
1378 by other higher priority error reporting module.
1379 off: disable H/W event reporting through EDAC.
1380 force: enforce the use of EDAC to report H/W event.
1384 Format: {"off" | "on" | "skip[mbr]"}
1387 Format: { "debug", "disable_early_pci_dma",
1388 "nochunk", "noruntime", "nosoftreserve",
1389 "novamap", "no_disable_early_pci_dma" }
1390 debug: enable misc debug output.
1391 disable_early_pci_dma: disable the busmaster bit on all
1392 PCI bridges while in the EFI boot stub.
1393 nochunk: disable reading files in "chunks" in the EFI
1394 boot stub, as chunking can cause problems with some
1395 firmware implementations.
1396 noruntime : disable EFI runtime services support
1397 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1398 attribute may cause the kernel to reserve the
1399 memory range for a memory mapping driver to
1400 claim. Specify efi=nosoftreserve to disable this
1401 reservation and treat the memory by its base type
1402 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1403 novamap: do not call SetVirtualAddressMap().
1404 no_disable_early_pci_dma: Leave the busmaster bit set
1405 on all PCI bridges while in the EFI boot stub
1407 efi_no_storage_paranoia [EFI; X86]
1408 Using this parameter you can use more than 50% of
1409 your efi variable storage. Use this parameter only if
1410 you are really sure that your UEFI does sane gc and
1411 fulfills the spec otherwise your board may brick.
1413 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1414 Add arbitrary attribute to specific memory range by
1415 updating original EFI memory map.
1416 Region of memory which aa attribute is added to is
1419 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1420 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1421 attribute is added to range 0x100000000-0x180000000 and
1422 0x10a0000000-0x1120000000.
1424 If efi_fake_mem=8G@9G:0x40000 is specified, the
1425 EFI_MEMORY_SP(0x40000) attribute is added to
1426 range 0x240000000-0x43fffffff.
1428 Using this parameter you can do debugging of EFI memmap
1429 related features. For example, you can do debugging of
1430 Address Range Mirroring feature even if your box
1431 doesn't support it, or mark specific memory as
1434 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1435 that is to be dynamically loaded by Linux. If there are
1436 multiple variables with the same name but with different
1437 vendor GUIDs, all of them will be loaded. See
1438 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1441 eisa_irq_edge= [PARISC,HW]
1442 See header of drivers/parisc/eisa.c.
1444 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1447 This is designed to be used in conjunction with
1448 the boot argument: earlyprintk=vga
1450 This parameter works in place of the kgdboc parameter
1451 but can only be used if the backing tty is available
1452 very early in the boot process. For early debugging
1453 via a serial port see kgdboc_earlycon instead.
1456 See comment before function elanfreq_setup() in
1457 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1459 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390]
1460 Specifies physical address of start of kernel core
1461 image elf header and optionally the size. Generally
1462 kexec loader will pass this option to capture kernel.
1463 See Documentation/admin-guide/kdump/kdump.rst for details.
1465 enable_mtrr_cleanup [X86]
1466 The kernel tries to adjust MTRR layout from continuous
1467 to discrete, to make X server driver able to add WB
1468 entry later. This parameter enables that.
1470 enable_timer_pin_1 [X86]
1471 Enable PIN 1 of APIC timer
1472 Can be useful to work around chipset bugs
1473 (in particular on some ATI chipsets).
1474 The kernel tries to set a reasonable default.
1476 enforcing= [SELINUX] Set initial enforcing status.
1478 See security/selinux/Kconfig help text.
1479 0 -- permissive (log only, no denials).
1480 1 -- enforcing (deny and log).
1482 Value can be changed at runtime via
1483 /sys/fs/selinux/enforce.
1486 Disable Error Record Serialization Table (ERST)
1489 ether= [HW,NET] Ethernet cards parameters
1490 This option is obsoleted by the "netdev=" option, which
1491 has equivalent usage. See its documentation for details.
1495 Permit 'security.evm' to be updated regardless of
1496 current integrity status.
1498 early_page_ext [KNL] Enforces page_ext initialization to earlier
1499 stages so cover more early boot allocations.
1500 Please note that as side effect some optimizations
1501 might be disabled to achieve that (e.g. parallelized
1502 memory initialization is disabled) so the boot process
1503 might take longer, especially on systems with a lot of
1504 memory. Available with CONFIG_PAGE_EXTENSION=y.
1509 fail_make_request=[KNL]
1510 General fault injection mechanism.
1511 Format: <interval>,<probability>,<space>,<times>
1512 See also Documentation/fault-injection/.
1515 Format: { initns | none }
1516 See Documentation/admin-guide/sysctl/net.rst for
1517 fb_tunnels_only_for_init_ns
1520 See Documentation/admin-guide/blockdev/floppy.rst.
1523 Forcefully enable Physical Address Extension (PAE).
1524 Many Pentium M systems disable PAE but may have a
1525 functionally usable PAE implementation.
1526 Warning: use of this parameter will taint the kernel
1527 and may cause unknown problems.
1530 [FTRACE] will set and start the specified tracer
1531 as early as possible in order to facilitate early
1534 ftrace_boot_snapshot
1535 [FTRACE] On boot up, a snapshot will be taken of the
1536 ftrace ring buffer that can be read at:
1537 /sys/kernel/tracing/snapshot.
1538 This is useful if you need tracing information from kernel
1539 boot up that is likely to be overridden by user space
1540 start up functionality.
1542 Optionally, the snapshot can also be defined for a tracing
1543 instance that was created by the trace_instance= command
1546 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo
1548 The above will cause the "foo" tracing instance to trigger
1549 a snapshot at the end of boot up.
1551 ftrace_dump_on_oops[=orig_cpu]
1552 [FTRACE] will dump the trace buffers on oops.
1553 If no parameter is passed, ftrace will dump
1554 buffers of all CPUs, but if you pass orig_cpu, it will
1555 dump only the buffer of the CPU that triggered the
1558 ftrace_filter=[function-list]
1559 [FTRACE] Limit the functions traced by the function
1560 tracer at boot up. function-list is a comma-separated
1561 list of functions. This list can be changed at run
1562 time by the set_ftrace_filter file in the debugfs
1565 ftrace_notrace=[function-list]
1566 [FTRACE] Do not trace the functions specified in
1567 function-list. This list can be changed at run time
1568 by the set_ftrace_notrace file in the debugfs
1571 ftrace_graph_filter=[function-list]
1572 [FTRACE] Limit the top level callers functions traced
1573 by the function graph tracer at boot up.
1574 function-list is a comma-separated list of functions
1575 that can be changed at run time by the
1576 set_graph_function file in the debugfs tracing directory.
1578 ftrace_graph_notrace=[function-list]
1579 [FTRACE] Do not trace from the functions specified in
1580 function-list. This list is a comma-separated list of
1581 functions that can be changed at run time by the
1582 set_graph_notrace file in the debugfs tracing directory.
1584 ftrace_graph_max_depth=<uint>
1585 [FTRACE] Used with the function graph tracer. This is
1586 the max depth it will trace into a function. This value
1587 can be changed at run time by the max_graph_depth file
1588 in the tracefs tracing directory. default: 0 (no limit)
1590 fw_devlink= [KNL] Create device links between consumer and supplier
1591 devices by scanning the firmware to infer the
1592 consumer/supplier relationships. This feature is
1593 especially useful when drivers are loaded as modules as
1594 it ensures proper ordering of tasks like device probing
1595 (suppliers first, then consumers), supplier boot state
1596 clean up (only after all consumers have probed),
1597 suspend/resume & runtime PM (consumers first, then
1599 Format: { off | permissive | on | rpm }
1600 off -- Don't create device links from firmware info.
1601 permissive -- Create device links from firmware info
1602 but use it only for ordering boot state clean
1603 up (sync_state() calls).
1604 on -- Create device links from firmware info and use it
1605 to enforce probe and suspend/resume ordering.
1606 rpm -- Like "on", but also use to order runtime PM.
1608 fw_devlink.strict=<bool>
1609 [KNL] Treat all inferred dependencies as mandatory
1610 dependencies. This only applies for fw_devlink=on|rpm.
1613 fw_devlink.sync_state =
1614 [KNL] When all devices that could probe have finished
1615 probing, this parameter controls what to do with
1616 devices that haven't yet received their sync_state()
1618 Format: { strict | timeout }
1619 strict -- Default. Continue waiting on consumers to
1621 timeout -- Give up waiting on consumers and call
1622 sync_state() on any devices that haven't yet
1623 received their sync_state() calls after
1624 deferred_probe_timeout has expired or by
1625 late_initcall() if !CONFIG_MODULES.
1628 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1629 support via parallel port (up to 5 devices per port)
1630 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1631 See also Documentation/input/devices/joystick-parport.rst
1635 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1639 gather_data_sampling=
1640 [X86,INTEL] Control the Gather Data Sampling (GDS)
1643 Gather Data Sampling is a hardware vulnerability which
1644 allows unprivileged speculative access to data which was
1645 previously stored in vector registers.
1647 This issue is mitigated by default in updated microcode.
1648 The mitigation may have a performance impact but can be
1649 disabled. On systems without the microcode mitigation
1650 disabling AVX serves as a mitigation.
1652 force: Disable AVX to mitigate systems without
1653 microcode mitigation. No effect if the microcode
1654 mitigation is present. Known to cause crashes in
1655 userspace with buggy AVX enumeration.
1657 off: Disable GDS mitigation.
1659 gcov_persist= [GCOV] When non-zero (default), profiling data for
1660 kernel modules is saved and remains accessible via
1661 debugfs, even when the module is unloaded/reloaded.
1662 When zero, profiling data is discarded and associated
1663 debugfs files are removed at module unload time.
1665 goldfish [X86] Enable the goldfish android emulator platform.
1666 Don't use this when you are not running on the
1669 gpio-mockup.gpio_mockup_ranges
1670 [HW] Sets the ranges of gpiochip of for this device.
1671 Format: <start1>,<end1>,<start2>,<end2>...
1672 gpio-mockup.gpio_mockup_named_lines
1673 [HW] Let the driver know GPIO lines should be named.
1675 gpt [EFI] Forces disk with valid GPT signature but
1676 invalid Protective MBR to be treated as GPT. If the
1677 primary GPT is corrupted, it enables the backup/alternate
1678 GPT to be used instead.
1680 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1681 the "Enable 0" bit of the configuration register.
1684 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1685 the "Enable 0" bit of the configuration register.
1688 grcan.select= [HW] Select which physical interface to use.
1691 grcan.txsize= [HW] Sets the size of the tx buffer.
1692 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1694 grcan.rxsize= [HW] Sets the size of the rx buffer.
1695 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1699 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1700 hardening is enabled for this boot. Hardened
1701 usercopy checking is used to protect the kernel
1702 from reading or writing beyond known memory
1703 allocation boundaries as a proactive defense
1704 against bounds-checking flaws in the kernel's
1705 copy_to_user()/copy_from_user() interface.
1706 on Perform hardened usercopy checks (default).
1707 off Disable hardened usercopy checks.
1709 hardlockup_all_cpu_backtrace=
1710 [KNL] Should the hard-lockup detector generate
1711 backtraces on all cpus.
1714 hashdist= [KNL,NUMA] Large hashes allocated during boot
1715 are distributed across NUMA nodes. Defaults on
1716 for 64-bit NUMA, off otherwise.
1717 Format: 0 | 1 (for off | on)
1719 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1721 hd= [EIDE] (E)IDE hard drive subsystem geometry
1722 Format: <cyl>,<head>,<sect>
1725 Disable Hardware Error Source Table (HEST) support;
1726 corresponding firmware-first mode error processing
1727 logic will be disabled.
1729 hibernate= [HIBERNATION]
1730 noresume Don't check if there's a hibernation image
1731 present during boot.
1732 nocompress Don't compress/decompress hibernation images.
1733 no Disable hibernation and resume.
1734 protect_image Turn on image protection during restoration
1735 (that will set all pages holding image data
1736 during restoration read-only).
1738 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1739 size of <nn>. This works even on boxes that have no
1740 highmem otherwise. This also works to reduce highmem
1741 size on bigger boxes.
1743 highres= [KNL] Enable/disable high resolution timer mode.
1744 Valid parameters: "on", "off"
1749 hostname= [KNL] Set the hostname (aka UTS nodename).
1751 This allows setting the system's hostname during early
1752 startup. This sets the name returned by gethostname.
1753 Using this parameter to set the hostname makes it
1754 possible to ensure the hostname is correctly set before
1755 any userspace processes run, avoiding the possibility
1756 that a process may call gethostname before the hostname
1757 has been explicitly set, resulting in the calling
1758 process getting an incorrect result. The string must
1759 not exceed the maximum allowed hostname length (usually
1760 64 characters) and will be truncated otherwise.
1762 hpet= [X86-32,HPET] option to control HPET usage
1763 Format: { enable (default) | disable | force |
1765 disable: disable HPET and use PIT instead
1766 force: allow force enabled of undocumented chips (ICH4,
1768 verbose: show contents of HPET registers during setup
1770 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1771 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1773 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1774 If this follows hugepagesz (below), it specifies
1775 the number of pages of hugepagesz to be allocated.
1776 If this is the first HugeTLB parameter on the command
1777 line, it specifies the number of pages to allocate for
1778 the default huge page size. If using node format, the
1779 number of pages to allocate per-node can be specified.
1780 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1781 Format: <integer> or (node format)
1782 <node>:<integer>[,<node>:<integer>]
1785 [HW] The size of the HugeTLB pages. This is used in
1786 conjunction with hugepages (above) to allocate huge
1787 pages of a specific size at boot. The pair
1788 hugepagesz=X hugepages=Y can be specified once for
1789 each supported huge page size. Huge page sizes are
1790 architecture dependent. See also
1791 Documentation/admin-guide/mm/hugetlbpage.rst.
1794 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1795 of gigantic hugepages. Or using node format, the size
1796 of a CMA area per node can be specified.
1797 Format: nn[KMGTPE] or (node format)
1798 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1800 Reserve a CMA area of given size and allocate gigantic
1801 hugepages using the CMA allocator. If enabled, the
1802 boot-time allocation of gigantic hugepages is skipped.
1804 hugetlb_free_vmemmap=
1805 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1807 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1808 Allows heavy hugetlb users to free up some more
1809 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1810 Format: { on | off (default) }
1815 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1818 Note that the vmemmap pages may be allocated from the added
1819 memory block itself when memory_hotplug.memmap_on_memory is
1820 enabled, those vmemmap pages cannot be optimized even if this
1821 feature is enabled. Other vmemmap pages not allocated from
1822 the added memory block itself do not be affected.
1825 [KNL] Should the hung task detector generate panics.
1828 A value of 1 instructs the kernel to panic when a
1829 hung task is detected. The default value is controlled
1830 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1831 option. The value selected by this boot parameter can
1832 be changed later by the kernel.hung_task_panic sysctl.
1834 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1835 terminal devices. Valid values: 0..8
1836 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1837 If specified, z/VM IUCV HVC accepts connections
1838 from listed z/VM user IDs only.
1840 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1841 which allow the hypervisor to 'idle' the
1842 guest on lock contention.
1844 i2c_bus= [HW] Override the default board specific I2C bus speed
1845 or register an additional I2C bus that is not
1846 registered from board initialization code.
1850 i8042.debug [HW] Toggle i8042 debug mode
1851 i8042.unmask_kbd_data
1852 [HW] Enable printing of interrupt data from the KBD port
1853 (disabled by default, and as a pre-condition
1854 requires that i8042.debug=1 be enabled)
1855 i8042.direct [HW] Put keyboard port into non-translated mode
1856 i8042.dumbkbd [HW] Pretend that controller can only read data from
1857 keyboard and cannot control its state
1858 (Don't attempt to blink the leds)
1859 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1860 i8042.nokbd [HW] Don't check/create keyboard port
1861 i8042.noloop [HW] Disable the AUX Loopback command while probing
1863 i8042.nomux [HW] Don't check presence of an active multiplexing
1865 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1867 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1868 i8042.reset [HW] Reset the controller during init, cleanup and
1869 suspend-to-ram transitions, only during s2r
1870 transitions, or never reset
1871 Format: { 1 | Y | y | 0 | N | n }
1872 1, Y, y: always reset controller
1873 0, N, n: don't ever reset controller
1874 Default: only on s2r transitions on x86; most other
1875 architectures force reset to be always executed
1876 i8042.unlock [HW] Unlock (ignore) the keylock
1877 i8042.kbdreset [HW] Reset device connected to KBD port
1879 [HW] Allow deferred probing upon i8042 probe errors
1883 i915.invert_brightness=
1884 [DRM] Invert the sense of the variable that is used to
1885 set the brightness of the panel backlight. Normally a
1886 brightness value of 0 indicates backlight switched off,
1887 and the maximum of the brightness value sets the backlight
1888 to maximum brightness. If this parameter is set to 0
1889 (default) and the machine requires it, or this parameter
1890 is set to 1, a brightness value of 0 sets the backlight
1891 to maximum brightness, and the maximum of the brightness
1892 value switches the backlight off.
1893 -1 -- never invert brightness
1894 0 -- machine default
1895 1 -- force brightness inversion
1897 ia32_emulation= [X86-64]
1899 When true, allows loading 32-bit programs and executing 32-bit
1900 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at
1901 boot time. When false, unconditionally disables IA32 emulation.
1904 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1908 Format: idle=poll, idle=halt, idle=nomwait
1909 Poll forces a polling idle loop that can slightly
1910 improve the performance of waking up a idle CPU, but
1911 will use a lot of power and make the system run hot.
1913 idle=halt: Halt is forced to be used for CPU idle.
1914 In such case C2/C3 won't be used again.
1915 idle=nomwait: Disable mwait for CPU C-states
1919 Allow force disabling of Shared Virtual Memory (SVA)
1920 support for the idxd driver. By default it is set to
1923 idxd.tc_override= [HW]
1925 Allow override of default traffic class configuration
1926 for the device. By default it is set to false (0).
1928 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1929 Format: { strict | legacy | 2008 | relaxed }
1932 Choose which programs will be accepted for execution
1933 based on the IEEE 754 NaN encoding(s) supported by
1934 the FPU and the NaN encoding requested with the value
1935 of an ELF file header flag individually set by each
1936 binary. Hardware implementations are permitted to
1937 support either or both of the legacy and the 2008 NaN
1940 Available settings are as follows:
1941 strict accept binaries that request a NaN encoding
1942 supported by the FPU
1943 legacy only accept legacy-NaN binaries, if supported
1945 2008 only accept 2008-NaN binaries, if supported
1947 relaxed accept any binaries regardless of whether
1948 supported by the FPU
1950 The FPU emulator is always able to support both NaN
1951 encodings, so if no FPU hardware is present or it has
1952 been disabled with 'nofpu', then the settings of
1953 'legacy' and '2008' strap the emulator accordingly,
1954 'relaxed' straps the emulator for both legacy-NaN and
1955 2008-NaN, whereas 'strict' enables legacy-NaN only on
1956 legacy processors and both NaN encodings on MIPS32 or
1959 The setting for ABS.fmt/NEG.fmt instruction execution
1960 mode generally follows that for the NaN encoding,
1961 except where unsupported by hardware.
1963 ignore_loglevel [KNL]
1964 Ignore loglevel setting - this will print /all/
1965 kernel messages to the console. Useful for debugging.
1966 We also add it as printk module parameter, so users
1967 could change it dynamically, usually by
1968 /sys/module/printk/parameters/ignore_loglevel.
1971 Ignore RLIMIT_DATA setting for data mappings,
1972 print warning at first misuse. Can be changed via
1973 /sys/module/kernel/parameters/ignore_rlimit_data.
1975 ihash_entries= [KNL]
1976 Set number of hash buckets for inode cache.
1978 ima_appraise= [IMA] appraise integrity measurements
1979 Format: { "off" | "enforce" | "fix" | "log" }
1982 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1983 The builtin appraise policy appraises all files
1986 ima_canonical_fmt [IMA]
1987 Use the canonical format for the binary runtime
1988 measurements, instead of host native format.
1991 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1995 The list of supported hash algorithms is defined
1996 in crypto/hash_info.h.
1999 The builtin policies to load during IMA setup.
2000 Format: "tcb | appraise_tcb | secure_boot |
2001 fail_securely | critical_data"
2003 The "tcb" policy measures all programs exec'd, files
2004 mmap'd for exec, and all files opened with the read
2005 mode bit set by either the effective uid (euid=0) or
2008 The "appraise_tcb" policy appraises the integrity of
2009 all files owned by root.
2011 The "secure_boot" policy appraises the integrity
2012 of files (eg. kexec kernel image, kernel modules,
2013 firmware, policy, etc) based on file signatures.
2015 The "fail_securely" policy forces file signature
2016 verification failure also on privileged mounted
2017 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
2020 The "critical_data" policy measures kernel integrity
2023 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
2024 Load a policy which meets the needs of the Trusted
2025 Computing Base. This means IMA will measure all
2026 programs exec'd, files mmap'd for exec, and all files
2027 opened for read by uid=0.
2030 Select one of defined IMA measurements template formats.
2031 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
2036 [IMA] Define a custom template format.
2037 Format: { "field1|...|fieldN" }
2039 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
2040 Format: <min_file_size>
2041 Set the minimal file size for using asynchronous hash.
2042 If left unspecified, ahash usage is disabled.
2044 ahash performance varies for different data sizes on
2045 different crypto accelerators. This option can be used
2046 to achieve the best performance for a particular HW.
2048 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
2050 Set hashing buffer size. Default: 4k.
2052 ahash performance varies for different chunk sizes on
2053 different crypto accelerators. This option can be used
2054 to achieve best performance for particular HW.
2058 Run specified binary instead of /sbin/init as init
2061 initcall_debug [KNL] Trace initcalls as they are executed. Useful
2062 for working out where the kernel is dying during
2065 initcall_blacklist= [KNL] Do not execute a comma-separated list of
2066 initcall functions. Useful for debugging built-in
2067 modules and initcalls.
2069 initramfs_async= [KNL]
2072 This parameter controls whether the initramfs
2073 image is unpacked asynchronously, concurrently
2074 with devices being probed and
2075 initialized. This should normally just work,
2076 but as a debugging aid, one can get the
2077 historical behaviour of the initramfs
2078 unpacking being completed before device_ and
2081 initrd= [BOOT] Specify the location of the initial ramdisk
2083 initrdmem= [KNL] Specify a physical address and size from which to
2084 load the initrd. If an initrd is compiled in or
2085 specified in the bootparams, it takes priority over this
2087 Format: ss[KMG],nn[KMG]
2090 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
2093 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2095 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
2097 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2099 init_pkru= [X86] Specify the default memory protection keys rights
2100 register contents for all processes. 0x55555554 by
2101 default (disallow access to all but pkey 0). Can
2102 override in debugfs after boot.
2104 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2107 int_pln_enable [X86] Enable power limit notification interrupt
2109 integrity_audit=[IMA]
2110 Format: { "0" | "1" }
2111 0 -- basic integrity auditing messages. (Default)
2112 1 -- additional integrity auditing messages.
2114 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2116 Enable intel iommu driver.
2118 Disable intel iommu driver.
2119 igfx_off [Default Off]
2120 By default, gfx is mapped as normal device. If a gfx
2121 device has a dedicated DMAR unit, the DMAR unit is
2122 bypassed by not enabling DMAR with this option. In
2123 this case, gfx device will use physical address for
2125 strict [Default Off]
2126 Deprecated, equivalent to iommu.strict=1.
2127 sp_off [Default Off]
2128 By default, super page will be supported if Intel IOMMU
2129 has the capability. With this option, super page will
2132 Enable the Intel IOMMU scalable mode if the hardware
2133 advertises that it has support for the scalable mode
2136 Disallow use of the Intel IOMMU scalable mode.
2137 tboot_noforce [Default Off]
2138 Do not force the Intel IOMMU enabled under tboot.
2139 By default, tboot will force Intel IOMMU on, which
2140 could harm performance of some high-throughput
2141 devices like 40GBit network cards, even if identity
2143 Note that using this option lowers the security
2144 provided by tboot because it makes the system
2145 vulnerable to DMA attacks.
2147 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2148 0 disables intel_idle and fall back on acpi_idle.
2149 1 to 9 specify maximum depth of C-state.
2153 Do not enable intel_pstate as the default
2154 scaling driver for the supported processors
2156 Use intel_pstate driver to bypass the scaling
2157 governors layer of cpufreq and provides it own
2158 algorithms for p-state selection. There are two
2159 P-state selection algorithms provided by
2160 intel_pstate in the active mode: powersave and
2161 performance. The way they both operate depends
2162 on whether or not the hardware managed P-states
2163 (HWP) feature has been enabled in the processor
2164 and possibly on the processor model.
2166 Use intel_pstate as a scaling driver, but configure it
2167 to work with generic cpufreq governors (instead of
2168 enabling its internal governor). This mode cannot be
2169 used along with the hardware-managed P-states (HWP)
2172 Enable intel_pstate on systems that prohibit it by default
2173 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2174 instead of acpi-cpufreq may disable platform features, such
2175 as thermal controls and power capping, that rely on ACPI
2176 P-States information being indicated to OSPM and therefore
2177 should be used with caution. This option does not work with
2178 processors that aren't supported by the intel_pstate driver
2179 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2181 Do not enable hardware P state control (HWP)
2184 Only load intel_pstate on systems which support
2185 hardware P state control (HWP) if available.
2187 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2188 Description Table, specifies preferred power management
2189 profile as "Enterprise Server" or "Performance Server",
2190 then this feature is turned on by default.
2192 Allow per-logical-CPU P-State performance control limits using
2193 cpufreq sysfs interface
2195 intremap= [X86-64, Intel-IOMMU]
2196 on enable Interrupt Remapping (default)
2197 off disable Interrupt Remapping
2198 nosid disable Source ID checking
2200 BIOS x2APIC opt-out request will be ignored
2201 nopost disable Interrupt Posting
2203 iomem= Disable strict checking of access to MMIO memory
2204 strict regions from userspace.
2219 nobypass [PPC/POWERNV]
2220 Disable IOMMU bypass, using IOMMU for PCI devices.
2222 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2223 Format: { "0" | "1" }
2224 0 - Try to allocate a 32-bit DMA address first, before
2225 falling back to the full range if needed.
2226 1 - Allocate directly from the full usable range,
2227 forcing Dual Address Cycle for PCI cards supporting
2228 greater than 32-bit addressing.
2230 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2231 Format: { "0" | "1" }
2233 Request that DMA unmap operations use deferred
2234 invalidation of hardware TLBs, for increased
2235 throughput at the cost of reduced device isolation.
2236 Will fall back to strict mode if not supported by
2237 the relevant IOMMU driver.
2239 DMA unmap operations invalidate IOMMU hardware TLBs
2241 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2242 Note: on x86, strict mode specified via one of the
2243 legacy driver-specific options takes precedence.
2246 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2247 Format: { "0" | "1" }
2248 0 - Use IOMMU translation for DMA.
2249 1 - Bypass the IOMMU for DMA.
2250 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2252 io7= [HW] IO7 for Marvel-based Alpha systems
2253 See comment before marvel_specify_io7 in
2254 arch/alpha/kernel/core_marvel.c.
2256 io_delay= [X86] I/O delay method
2258 Standard port 0x80 based delay
2260 Alternate port 0xed based delay (needed on some systems)
2262 Simple two microseconds delay
2267 See Documentation/admin-guide/nfs/nfsroot.rst.
2269 ipcmni_extend [KNL] Extend the maximum number of unique System V
2270 IPC identifiers from 32,768 to 16,777,216.
2272 irqaffinity= [SMP] Set the default irq affinity mask
2273 The argument is a cpu list, as described above.
2275 irqchip.gicv2_force_probe=
2278 Force the kernel to look for the second 4kB page
2279 of a GICv2 controller even if the memory range
2280 exposed by the device tree is too small.
2282 irqchip.gicv3_nolpi=
2284 Force the kernel to ignore the availability of
2285 LPIs (and by consequence ITSs). Intended for system
2286 that use the kernel as a bootloader, and thus want
2287 to let secondary kernels in charge of setting up
2290 irqchip.gicv3_pseudo_nmi= [ARM64]
2291 Enables support for pseudo-NMIs in the kernel. This
2292 requires the kernel to be built with
2293 CONFIG_ARM64_PSEUDO_NMI.
2296 When an interrupt is not handled search all handlers
2297 for it. Intended to get systems with badly broken
2301 When an interrupt is not handled search all handlers
2302 for it. Also check all handlers each timer
2303 interrupt. Intended to get systems with badly broken
2307 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2309 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2310 [Deprecated - use cpusets instead]
2311 Format: [flag-list,]<cpu-list>
2313 Specify one or more CPUs to isolate from disturbances
2314 specified in the flag list (default: domain):
2317 Disable the tick when a single task runs.
2319 A residual 1Hz tick is offloaded to workqueues, which you
2320 need to affine to housekeeping through the global
2321 workqueue's affinity configured via the
2322 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2323 by using the 'domain' flag described below.
2325 NOTE: by default the global workqueue runs on all CPUs,
2326 so to protect individual CPUs the 'cpumask' file has to
2327 be configured manually after bootup.
2330 Isolate from the general SMP balancing and scheduling
2331 algorithms. Note that performing domain isolation this way
2332 is irreversible: it's not possible to bring back a CPU to
2333 the domains once isolated through isolcpus. It's strongly
2334 advised to use cpusets instead to disable scheduler load
2335 balancing through the "cpuset.sched_load_balance" file.
2336 It offers a much more flexible interface where CPUs can
2337 move in and out of an isolated set anytime.
2339 You can move a process onto or off an "isolated" CPU via
2340 the CPU affinity syscalls or cpuset.
2341 <cpu number> begins at 0 and the maximum value is
2342 "number of CPUs in system - 1".
2346 Isolate from being targeted by managed interrupts
2347 which have an interrupt mask containing isolated
2348 CPUs. The affinity of managed interrupts is
2349 handled by the kernel and cannot be changed via
2350 the /proc/irq/* interfaces.
2352 This isolation is best effort and only effective
2353 if the automatically assigned interrupt mask of a
2354 device queue contains isolated and housekeeping
2355 CPUs. If housekeeping CPUs are online then such
2356 interrupts are directed to the housekeeping CPU
2357 so that IO submitted on the housekeeping CPU
2358 cannot disturb the isolated CPU.
2360 If a queue's affinity mask contains only isolated
2361 CPUs then this parameter has no effect on the
2362 interrupt routing decision, though interrupts are
2363 only delivered when tasks running on those
2364 isolated CPUs submit IO. IO submitted on
2365 housekeeping CPUs has no influence on those
2368 The format of <cpu-list> is described above.
2372 ivrs_ioapic [HW,X86-64]
2373 Provide an override to the IOAPIC-ID<->DEVICE-ID
2374 mapping provided in the IVRS ACPI table.
2375 By default, PCI segment is 0, and can be omitted.
2377 For example, to map IOAPIC-ID decimal 10 to
2378 PCI segment 0x1 and PCI device 00:14.0,
2379 write the parameter as:
2380 ivrs_ioapic=10@0001:00:14.0
2383 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2384 write the parameter as:
2385 ivrs_ioapic[10]=00:14.0
2386 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2387 PCI device 00:14.0 write the parameter as:
2388 ivrs_ioapic[10]=0001:00:14.0
2390 ivrs_hpet [HW,X86-64]
2391 Provide an override to the HPET-ID<->DEVICE-ID
2392 mapping provided in the IVRS ACPI table.
2393 By default, PCI segment is 0, and can be omitted.
2395 For example, to map HPET-ID decimal 10 to
2396 PCI segment 0x1 and PCI device 00:14.0,
2397 write the parameter as:
2398 ivrs_hpet=10@0001:00:14.0
2401 * To map HPET-ID decimal 0 to PCI device 00:14.0
2402 write the parameter as:
2403 ivrs_hpet[0]=00:14.0
2404 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2405 PCI device 00:14.0 write the parameter as:
2406 ivrs_ioapic[10]=0001:00:14.0
2408 ivrs_acpihid [HW,X86-64]
2409 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2410 mapping provided in the IVRS ACPI table.
2411 By default, PCI segment is 0, and can be omitted.
2413 For example, to map UART-HID:UID AMD0020:0 to
2414 PCI segment 0x1 and PCI device ID 00:14.5,
2415 write the parameter as:
2416 ivrs_acpihid=AMD0020:0@0001:00:14.5
2419 * To map UART-HID:UID AMD0020:0 to PCI segment is 0,
2420 PCI device ID 00:14.5, write the parameter as:
2421 ivrs_acpihid[00:14.5]=AMD0020:0
2422 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and
2423 PCI device ID 00:14.5, write the parameter as:
2424 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2426 js= [HW,JOY] Analog joystick
2427 See Documentation/input/joydev/joystick.rst.
2430 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2431 report on every invalid memory access. Without this
2432 parameter KASAN will print report only for the first
2436 Do not unregister boot console at start. This is only
2437 useful for debugging when something happens in the window
2438 between unregistering the boot console and initializing
2443 kernelcore= [KNL,X86,IA-64,PPC]
2444 Format: nn[KMGTPE] | nn% | "mirror"
2445 This parameter specifies the amount of memory usable by
2446 the kernel for non-movable allocations. The requested
2447 amount is spread evenly throughout all nodes in the
2448 system as ZONE_NORMAL. The remaining memory is used for
2449 movable memory in its own zone, ZONE_MOVABLE. In the
2450 event, a node is too small to have both ZONE_NORMAL and
2451 ZONE_MOVABLE, kernelcore memory will take priority and
2452 other nodes will have a larger ZONE_MOVABLE.
2454 ZONE_MOVABLE is used for the allocation of pages that
2455 may be reclaimed or moved by the page migration
2456 subsystem. Note that allocations like PTEs-from-HighMem
2457 still use the HighMem zone if it exists, and the Normal
2458 zone if it does not.
2460 It is possible to specify the exact amount of memory in
2461 the form of "nn[KMGTPE]", a percentage of total system
2462 memory in the form of "nn%", or "mirror". If "mirror"
2463 option is specified, mirrored (reliable) memory is used
2464 for non-movable allocations and remaining memory is used
2465 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2466 are exclusive, so you cannot specify multiple forms.
2468 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2469 Format: <Controller#>[,poll interval]
2470 The controller # is the number of the ehci usb debug
2471 port as it is probed via PCI. The poll interval is
2472 optional and is the number seconds in between
2473 each poll cycle to the debug port in case you need
2474 the functionality for interrupting the kernel with
2475 gdb or control-c on the dbgp connection. When
2476 not using this parameter you use sysrq-g to break into
2477 the kernel debugger.
2479 kgdboc= [KGDB,HW] kgdb over consoles.
2480 Requires a tty driver that supports console polling,
2481 or a supported polling keyboard driver (non-usb).
2482 Serial only format: <serial_device>[,baud]
2483 keyboard only format: kbd
2484 keyboard and serial format: kbd,<serial_device>[,baud]
2485 Optional Kernel mode setting:
2486 kms, kbd format: kms,kbd
2487 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2489 kgdboc_earlycon= [KGDB,HW]
2490 If the boot console provides the ability to read
2491 characters and can work in polling mode, you can use
2492 this parameter to tell kgdb to use it as a backend
2493 until the normal console is registered. Intended to
2494 be used together with the kgdboc parameter which
2495 specifies the normal console to transition to.
2497 The name of the early console should be specified
2498 as the value of this parameter. Note that the name of
2499 the early console might be different than the tty
2500 name passed to kgdboc. It's OK to leave the value
2501 blank and the first boot console that implements
2502 read() will be picked.
2504 kgdbwait [KGDB] Stop kernel execution and enter the
2505 kernel debugger at the earliest opportunity.
2507 kmac= [MIPS] Korina ethernet MAC address.
2508 Configure the RouterBoard 532 series on-chip
2509 Ethernet adapter MAC address.
2511 kmemleak= [KNL] Boot-time kmemleak enable/disable
2512 Valid arguments: on, off
2514 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2517 kprobe_event=[probe-list]
2518 [FTRACE] Add kprobe events and enable at boot time.
2519 The probe-list is a semicolon delimited list of probe
2520 definitions. Each definition is same as kprobe_events
2521 interface, but the parameters are comma delimited.
2522 For example, to add a kprobe event on vfs_read with
2523 arg1 and arg2, add to the command line;
2525 kprobe_event=p,vfs_read,$arg1,$arg2
2527 See also Documentation/trace/kprobetrace.rst "Kernel
2528 Boot Parameter" section.
2530 kpti= [ARM64] Control page table isolation of user
2531 and kernel address spaces.
2532 Default: enabled on cores which need mitigation.
2536 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires
2537 CONFIG_KUNIT to be set to be fully enabled. The
2538 default value can be overridden via
2539 KUNIT_DEFAULT_ENABLED.
2540 Default is 1 (enabled)
2542 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2543 Default is 0 (don't ignore, but inject #GP)
2545 kvm.eager_page_split=
2546 [KVM,X86] Controls whether or not KVM will try to
2547 proactively split all huge pages during dirty logging.
2548 Eager page splitting reduces interruptions to vCPU
2549 execution by eliminating the write-protection faults
2550 and MMU lock contention that would otherwise be
2551 required to split huge pages lazily.
2553 VM workloads that rarely perform writes or that write
2554 only to a small region of VM memory may benefit from
2555 disabling eager page splitting to allow huge pages to
2556 still be used for reads.
2558 The behavior of eager page splitting depends on whether
2559 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2560 disabled, all huge pages in a memslot will be eagerly
2561 split when dirty logging is enabled on that memslot. If
2562 enabled, eager page splitting will be performed during
2563 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2566 Eager page splitting is only supported when kvm.tdp_mmu=Y.
2570 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2571 Default is false (don't support).
2574 [KVM] Controls the software workaround for the
2575 X86_BUG_ITLB_MULTIHIT bug.
2576 force : Always deploy workaround.
2577 off : Never deploy workaround.
2578 auto : Deploy workaround based on the presence of
2579 X86_BUG_ITLB_MULTIHIT.
2583 If the software workaround is enabled for the host,
2584 guests do need not to enable it for nested guests.
2586 kvm.nx_huge_pages_recovery_ratio=
2587 [KVM] Controls how many 4KiB pages are periodically zapped
2588 back to huge pages. 0 disables the recovery, otherwise if
2589 the value is N KVM will zap 1/Nth of the 4KiB pages every
2590 period (see below). The default is 60.
2592 kvm.nx_huge_pages_recovery_period_ms=
2593 [KVM] Controls the time period at which KVM zaps 4KiB pages
2594 back to huge pages. If the value is a non-zero N, KVM will
2595 zap a portion (see ratio above) of the pages every N msecs.
2596 If the value is 0 (the default), KVM will pick a period based
2597 on the ratio, such that a page is zapped after 1 hour on average.
2599 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in
2600 KVM/SVM. Default is 1 (enabled).
2602 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables,
2603 a.k.a. Two-Dimensional Page Tables. Default is 1
2604 (enabled). Disable by KVM if hardware lacks support
2608 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2610 none: Forcefully disable KVM.
2612 nvhe: Standard nVHE-based mode, without support for
2615 protected: nVHE-based mode with support for guests whose
2616 state is kept private from the host.
2618 nested: VHE-based mode with support for nested
2619 virtualization. Requires at least ARMv8.3
2622 Defaults to VHE/nVHE based on hardware support. Setting
2623 mode to "protected" will disable kexec and hibernation
2624 for the host. "nested" is experimental and should be
2625 used with extreme caution.
2627 kvm-arm.vgic_v3_group0_trap=
2628 [KVM,ARM] Trap guest accesses to GICv3 group-0
2631 kvm-arm.vgic_v3_group1_trap=
2632 [KVM,ARM] Trap guest accesses to GICv3 group-1
2635 kvm-arm.vgic_v3_common_trap=
2636 [KVM,ARM] Trap guest accesses to GICv3 common
2639 kvm-arm.vgic_v4_enable=
2640 [KVM,ARM] Allow use of GICv4 for direct injection of
2643 kvm_cma_resv_ratio=n [PPC]
2644 Reserves given percentage from system memory area for
2645 contiguous memory allocation for KVM hash pagetable
2647 By default it reserves 5% of total system memory.
2651 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables,
2652 a.k.a. Two-Dimensional Page Tables. Default is 1
2653 (enabled). Disable by KVM if hardware lacks support
2656 kvm-intel.emulate_invalid_guest_state=
2657 [KVM,Intel] Control whether to emulate invalid guest
2658 state. Ignored if kvm-intel.enable_unrestricted_guest=1,
2659 as guest state is never invalid for unrestricted
2660 guests. This param doesn't apply to nested guests (L2),
2661 as KVM never emulates invalid L2 guest state.
2662 Default is 1 (enabled).
2664 kvm-intel.flexpriority=
2665 [KVM,Intel] Control KVM's use of FlexPriority feature
2666 (TPR shadow). Default is 1 (enabled). Disable by KVM if
2667 hardware lacks support for it.
2670 [KVM,Intel] Control nested virtualization feature in
2671 KVM/VMX. Default is 1 (enabled).
2673 kvm-intel.unrestricted_guest=
2674 [KVM,Intel] Control KVM's use of unrestricted guest
2675 feature (virtualized real and unpaged mode). Default
2676 is 1 (enabled). Disable by KVM if EPT is disabled or
2677 hardware lacks support for it.
2679 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2682 Valid arguments: never, cond, always
2684 always: L1D cache flush on every VMENTER.
2685 cond: Flush L1D on VMENTER only when the code between
2686 VMEXIT and VMENTER can leak host memory.
2687 never: Disables the mitigation
2689 Default is cond (do L1 cache flush in specific instances)
2691 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor
2692 Identification feature (tagged TLBs). Default is 1
2693 (enabled). Disable by KVM if hardware lacks support
2696 l1d_flush= [X86,INTEL]
2697 Control mitigation for L1D based snooping vulnerability.
2699 Certain CPUs are vulnerable to an exploit against CPU
2700 internal buffers which can forward information to a
2701 disclosure gadget under certain conditions.
2703 In vulnerable processors, the speculatively
2704 forwarded data can be used in a cache side channel
2705 attack, to access data to which the attacker does
2706 not have direct access.
2708 This parameter controls the mitigation. The
2711 on - enable the interface for the mitigation
2713 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2716 The kernel PTE inversion protection is unconditionally
2717 enabled and cannot be disabled.
2720 Provides all available mitigations for the
2721 L1TF vulnerability. Disables SMT and
2722 enables all mitigations in the
2723 hypervisors, i.e. unconditional L1D flush.
2725 SMT control and L1D flush control via the
2726 sysfs interface is still possible after
2727 boot. Hypervisors will issue a warning
2728 when the first VM is started in a
2729 potentially insecure configuration,
2730 i.e. SMT enabled or L1D flush disabled.
2733 Same as 'full', but disables SMT and L1D
2734 flush runtime control. Implies the
2735 'nosmt=force' command line option.
2736 (i.e. sysfs control of SMT is disabled.)
2739 Leaves SMT enabled and enables the default
2740 hypervisor mitigation, i.e. conditional
2743 SMT control and L1D flush control via the
2744 sysfs interface is still possible after
2745 boot. Hypervisors will issue a warning
2746 when the first VM is started in a
2747 potentially insecure configuration,
2748 i.e. SMT enabled or L1D flush disabled.
2752 Disables SMT and enables the default
2753 hypervisor mitigation.
2755 SMT control and L1D flush control via the
2756 sysfs interface is still possible after
2757 boot. Hypervisors will issue a warning
2758 when the first VM is started in a
2759 potentially insecure configuration,
2760 i.e. SMT enabled or L1D flush disabled.
2763 Same as 'flush', but hypervisors will not
2764 warn when a VM is started in a potentially
2765 insecure configuration.
2768 Disables hypervisor mitigations and doesn't
2770 It also drops the swap size and available
2771 RAM limit restriction on both hypervisor and
2776 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2782 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2785 lapic= [X86,APIC] Do not use TSC deadline
2786 value for LAPIC timer one-shot implementation. Default
2787 back to the programmable timer unit in the LAPIC.
2788 Format: notscdeadline
2790 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2793 libata.dma= [LIBATA] DMA control
2794 libata.dma=0 Disable all PATA and SATA DMA
2795 libata.dma=1 PATA and SATA Disk DMA only
2796 libata.dma=2 ATAPI (CDROM) DMA only
2797 libata.dma=4 Compact Flash DMA only
2798 Combinations also work, so libata.dma=3 enables DMA
2799 for disks and CDROMs, but not CFs.
2801 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2802 libata.ignore_hpa=0 keep BIOS limits (default)
2803 libata.ignore_hpa=1 ignore limits, using full disk
2805 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2809 libata.force= [LIBATA] Force configurations. The format is a comma-
2810 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2811 PORT and DEVICE are decimal numbers matching port, link
2812 or device. Basically, it matches the ATA ID string
2813 printed on console by libata. If the whole ID part is
2814 omitted, the last PORT and DEVICE values are used. If
2815 ID hasn't been specified yet, the configuration applies
2816 to all ports, links and devices.
2818 If only DEVICE is omitted, the parameter applies to
2819 the port and all links and devices behind it. DEVICE
2820 number of 0 either selects the first device or the
2821 first fan-out link behind PMP device. It does not
2822 select the host link. DEVICE number of 15 selects the
2823 host link and device attached to it.
2825 The VAL specifies the configuration to force. As long
2826 as there is no ambiguity, shortcut notation is allowed.
2827 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2828 The following configurations can be forced.
2830 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2831 Any ID with matching PORT is used.
2833 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2835 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2836 udma[/][16,25,33,44,66,100,133] notation is also
2839 * nohrst, nosrst, norst: suppress hard, soft and both
2842 * rstonce: only attempt one reset during hot-unplug
2845 * [no]dbdelay: Enable or disable the extra 200ms delay
2846 before debouncing a link PHY and device presence
2849 * [no]ncq: Turn on or off NCQ.
2851 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2853 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2855 * [no]trim: Enable or disable (unqueued) TRIM.
2857 * trim_zero: Indicate that TRIM command zeroes data.
2859 * max_trim_128m: Set 128M maximum trim size limit.
2861 * [no]dma: Turn on or off DMA transfers.
2863 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2865 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2866 commands that are not a multiple of 16 bytes.
2868 * [no]dmalog: Enable or disable the use of the
2869 READ LOG DMA EXT command to access logs.
2871 * [no]iddevlog: Enable or disable access to the
2872 identify device data log.
2874 * [no]logdir: Enable or disable access to the general
2875 purpose log directory.
2877 * max_sec_128: Set transfer size limit to 128 sectors.
2879 * max_sec_1024: Set or clear transfer size limit to
2882 * max_sec_lba48: Set or clear transfer size limit to
2885 * [no]lpm: Enable or disable link power management.
2887 * [no]setxfer: Indicate if transfer speed mode setting
2890 * [no]fua: Disable or enable FUA (Force Unit Access)
2891 support for devices supporting this feature.
2893 * dump_id: Dump IDENTIFY data.
2895 * disable: Disable this device.
2897 If there are multiple matching configurations changing
2898 the same attribute, the last one is used.
2900 load_ramdisk= [RAM] [Deprecated]
2902 lockd.nlm_grace_period=P [NFS] Assign grace period.
2905 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2908 lockd.nlm_timeout=T [NFS] Assign timeout value.
2911 lockd.nlm_udpport=M [NFS] Assign UDP port.
2914 lockdown= [SECURITY]
2915 { integrity | confidentiality }
2916 Enable the kernel lockdown feature. If set to
2917 integrity, kernel features that allow userland to
2918 modify the running kernel are disabled. If set to
2919 confidentiality, kernel features that allow userland
2920 to extract confidential information from the kernel
2923 locktorture.acq_writer_lim= [KNL]
2924 Set the time limit in jiffies for a lock
2925 acquisition. Acquisitions exceeding this limit
2926 will result in a splat once they do complete.
2928 locktorture.bind_readers= [KNL]
2929 Specify the list of CPUs to which the readers are
2932 locktorture.bind_writers= [KNL]
2933 Specify the list of CPUs to which the writers are
2936 locktorture.call_rcu_chains= [KNL]
2937 Specify the number of self-propagating call_rcu()
2938 chains to set up. These are used to ensure that
2939 there is a high probability of an RCU grace period
2940 in progress at any given time. Defaults to 0,
2941 which disables these call_rcu() chains.
2943 locktorture.long_hold= [KNL]
2944 Specify the duration in milliseconds for the
2945 occasional long-duration lock hold time. Defaults
2946 to 100 milliseconds. Select 0 to disable.
2948 locktorture.nested_locks= [KNL]
2949 Specify the maximum lock nesting depth that
2950 locktorture is to exercise, up to a limit of 8
2951 (MAX_NESTED_LOCKS). Specify zero to disable.
2952 Note that this parameter is ineffective on types
2953 of locks that do not support nested acquisition.
2955 locktorture.nreaders_stress= [KNL]
2956 Set the number of locking read-acquisition kthreads.
2957 Defaults to being automatically set based on the
2958 number of online CPUs.
2960 locktorture.nwriters_stress= [KNL]
2961 Set the number of locking write-acquisition kthreads.
2963 locktorture.onoff_holdoff= [KNL]
2964 Set time (s) after boot for CPU-hotplug testing.
2966 locktorture.onoff_interval= [KNL]
2967 Set time (s) between CPU-hotplug operations, or
2968 zero to disable CPU-hotplug testing.
2970 locktorture.rt_boost= [KNL]
2971 Do periodic testing of real-time lock priority
2972 boosting. Select 0 to disable, 1 to boost
2973 only rt_mutex, and 2 to boost unconditionally.
2974 Defaults to 2, which might seem to be an
2975 odd choice, but which should be harmless for
2976 non-real-time spinlocks, due to their disabling
2977 of preemption. Note that non-realtime mutexes
2980 locktorture.rt_boost_factor= [KNL]
2981 Number that determines how often and for how
2982 long priority boosting is exercised. This is
2983 scaled down by the number of writers, so that the
2984 number of boosts per unit time remains roughly
2985 constant as the number of writers increases.
2986 On the other hand, the duration of each boost
2987 increases with the number of writers.
2989 locktorture.shuffle_interval= [KNL]
2990 Set task-shuffle interval (jiffies). Shuffling
2991 tasks allows some CPUs to go into dyntick-idle
2992 mode during the locktorture test.
2994 locktorture.shutdown_secs= [KNL]
2995 Set time (s) after boot system shutdown. This
2996 is useful for hands-off automated testing.
2998 locktorture.stat_interval= [KNL]
2999 Time (s) between statistics printk()s.
3001 locktorture.stutter= [KNL]
3002 Time (s) to stutter testing, for example,
3003 specifying five seconds causes the test to run for
3004 five seconds, wait for five seconds, and so on.
3005 This tests the locking primitive's ability to
3006 transition abruptly to and from idle.
3008 locktorture.torture_type= [KNL]
3009 Specify the locking implementation to test.
3011 locktorture.verbose= [KNL]
3012 Enable additional printk() statements.
3014 locktorture.writer_fifo= [KNL]
3015 Run the write-side locktorture kthreads at
3016 sched_set_fifo() real-time priority.
3018 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
3021 loglevel= All Kernel Messages with a loglevel smaller than the
3022 console loglevel will be printed to the console. It can
3023 also be changed with klogd or other programs. The
3024 loglevels are defined as follows:
3026 0 (KERN_EMERG) system is unusable
3027 1 (KERN_ALERT) action must be taken immediately
3028 2 (KERN_CRIT) critical conditions
3029 3 (KERN_ERR) error conditions
3030 4 (KERN_WARNING) warning conditions
3031 5 (KERN_NOTICE) normal but significant condition
3032 6 (KERN_INFO) informational
3033 7 (KERN_DEBUG) debug-level messages
3035 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
3036 in bytes. n must be a power of two and greater
3037 than the minimal size. The minimal size is defined
3038 by LOG_BUF_SHIFT kernel config parameter. There is
3039 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
3040 that allows to increase the default size depending on
3041 the number of CPUs. See init/Kconfig for more details.
3043 logo.nologo [FB] Disables display of the built-in Linux logo.
3044 This may be used to provide more screen space for
3045 kernel log messages and is useful when debugging
3046 kernel boot problems.
3048 lp=0 [LP] Specify parallel ports to use, e.g,
3049 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
3050 lp=reset first parallel port). 'lp=0' disables the
3051 lp=auto printer driver. 'lp=reset' (which can be
3052 specified in addition to the ports) causes
3053 attached printers to be reset. Using
3054 lp=port1,port2,... specifies the parallel ports
3055 to associate lp devices with, starting with
3056 lp0. A port specification may be 'none' to skip
3057 that lp device, or a parport name such as
3058 'parport0'. Specifying 'lp=auto' instead of a
3059 port specification list means that device IDs
3060 from each port should be examined, to see if
3061 an IEEE 1284-compliant printer is attached; if
3062 so, the driver will manage that printer.
3063 See also header of drivers/char/lp.c.
3066 Sets loops_per_jiffy to given constant, thus avoiding
3067 time-consuming boot-time autodetection (up to 250 ms per
3068 CPU). 0 enables autodetection (default). To determine
3069 the correct value for your kernel, boot with normal
3070 autodetection and see what value is printed. Note that
3071 on SMP systems the preset will be applied to all CPUs,
3072 which is likely to cause problems if your CPUs need
3073 significantly divergent settings. An incorrect value
3074 will cause delays in the kernel to be wrong, leading to
3075 unpredictable I/O errors and other breakage. Although
3076 unlikely, in the extreme case this might damage your
3080 Format: <io>,<irq>,<dma>
3082 lsm.debug [SECURITY] Enable LSM initialization debugging output.
3085 [SECURITY] Choose order of LSM initialization. This
3086 overrides CONFIG_LSM, and the "security=" parameter.
3088 machvec= [IA-64] Force the use of a particular machine-vector
3089 (machvec) in a generic kernel.
3090 Example: machvec=hpzx1
3092 machtype= [Loongson] Share the same kernel image file between
3093 different yeeloong laptops.
3094 Example: machtype=lemote-yeeloong-2f-7inch
3096 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
3097 than or equal to this physical address is ignored.
3099 maxcpus= [SMP] Maximum number of processors that an SMP kernel
3100 will bring up during bootup. maxcpus=n : n >= 0 limits
3101 the kernel to bring up 'n' processors. Surely after
3102 bootup you can bring up the other plugged cpu by executing
3103 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
3104 only takes effect during system bootup.
3105 While n=0 is a special case, it is equivalent to "nosmp",
3106 which also disables the IO APIC.
3108 max_loop= [LOOP] The number of loop block devices that get
3109 (loop.max_loop) unconditionally pre-created at init time. The default
3110 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
3111 of statically allocating a predefined number, loop
3112 devices can be requested on-demand with the
3113 /dev/loop-control interface.
3115 mce [X86-32] Machine Check Exception
3117 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
3119 md= [HW] RAID subsystems devices and level
3120 See Documentation/admin-guide/md.rst.
3123 Format: <first>,<last>
3124 Specifies range of consoles to be captured by the MDA.
3127 Control mitigation for the Micro-architectural Data
3128 Sampling (MDS) vulnerability.
3130 Certain CPUs are vulnerable to an exploit against CPU
3131 internal buffers which can forward information to a
3132 disclosure gadget under certain conditions.
3134 In vulnerable processors, the speculatively
3135 forwarded data can be used in a cache side channel
3136 attack, to access data to which the attacker does
3137 not have direct access.
3139 This parameter controls the MDS mitigation. The
3142 full - Enable MDS mitigation on vulnerable CPUs
3143 full,nosmt - Enable MDS mitigation and disable
3144 SMT on vulnerable CPUs
3145 off - Unconditionally disable MDS mitigation
3147 On TAA-affected machines, mds=off can be prevented by
3148 an active TAA mitigation as both vulnerabilities are
3149 mitigated with the same mechanism so in order to disable
3150 this mitigation, you need to specify tsx_async_abort=off
3153 Not specifying this option is equivalent to
3156 For details see: Documentation/admin-guide/hw-vuln/mds.rst
3158 mem=nn[KMG] [HEXAGON] Set the memory size.
3159 Must be specified, otherwise memory size will be 0.
3161 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
3162 Amount of memory to be used in cases as follows:
3165 2 when the kernel is not able to see the whole system memory;
3166 3 memory that lies after 'mem=' boundary is excluded from
3167 the hypervisor, then assigned to KVM guests.
3168 4 to limit the memory available for kdump kernel.
3170 [ARC,MICROBLAZE] - the limit applies only to low memory,
3171 high memory is not affected.
3173 [ARM64] - only limits memory covered by the linear
3174 mapping. The NOMAP regions are not affected.
3176 [X86] Work as limiting max address. Use together
3177 with memmap= to avoid physical address space collisions.
3178 Without memmap= PCI devices could be placed at addresses
3179 belonging to unused RAM.
3181 Note that this only takes effects during boot time since
3182 in above case 3, memory may need be hot added after boot
3183 if system memory of hypervisor is not sufficient.
3186 [ARM,MIPS] - override the memory layout reported by
3188 Define a memory region of size nn[KMG] starting at
3190 Multiple different regions can be specified with
3191 multiple mem= parameters on the command line.
3193 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3196 memblock=debug [KNL] Enable memblock debug messages.
3199 [KNL,SH] Allow user to override the default size for
3200 per-device physically contiguous DMA buffers.
3202 memhp_default_state=online/offline/online_kernel/online_movable
3203 [KNL] Set the initial state for the memory hotplug
3204 onlining policy. If not specified, the default value is
3205 set according to the
3206 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3208 See Documentation/admin-guide/mm/memory-hotplug.rst.
3210 memmap=exactmap [KNL,X86] Enable setting of an exact
3211 E820 memory map, as specified by the user.
3212 Such memmap=exactmap lines can be constructed based on
3213 BIOS output or other requirements. See the memmap=nn@ss
3216 memmap=nn[KMG]@ss[KMG]
3217 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
3218 Region of memory to be used is from ss to ss+nn.
3219 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3220 which limits max address to nn[KMG].
3221 Multiple different regions can be specified,
3224 memmap=100M@2G,100M#3G,1G!1024G
3226 memmap=nn[KMG]#ss[KMG]
3227 [KNL,ACPI] Mark specific memory as ACPI data.
3228 Region of memory to be marked is from ss to ss+nn.
3230 memmap=nn[KMG]$ss[KMG]
3231 [KNL,ACPI] Mark specific memory as reserved.
3232 Region of memory to be reserved is from ss to ss+nn.
3233 Example: Exclude memory from 0x18690000-0x1869ffff
3234 memmap=64K$0x18690000
3236 memmap=0x10000$0x18690000
3237 Some bootloaders may need an escape character before '$',
3238 like Grub2, otherwise '$' and the following number
3241 memmap=nn[KMG]!ss[KMG]
3242 [KNL,X86] Mark specific memory as protected.
3243 Region of memory to be used, from ss to ss+nn.
3244 The memory region may be marked as e820 type 12 (0xc)
3245 and is NVDIMM or ADR memory.
3247 memmap=<size>%<offset>-<oldtype>+<newtype>
3248 [KNL,ACPI] Convert memory within the specified region
3249 from <oldtype> to <newtype>. If "-<oldtype>" is left
3250 out, the whole region will be marked as <newtype>,
3251 even if previously unavailable. If "+<newtype>" is left
3252 out, matching memory will be removed. Types are
3253 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3254 3 = ACPI, 12 = PRAM.
3256 memory_corruption_check=0/1 [X86]
3257 Some BIOSes seem to corrupt the first 64k of
3258 memory when doing things like suspend/resume.
3259 Setting this option will scan the memory
3260 looking for corruption. Enabling this will
3261 both detect corruption and prevent the kernel
3262 from using the memory being corrupted.
3263 However, its intended as a diagnostic tool; if
3264 repeatable BIOS-originated corruption always
3265 affects the same memory, you can use memmap=
3266 to prevent the kernel from using that memory.
3268 memory_corruption_check_size=size [X86]
3269 By default it checks for corruption in the low
3270 64k, making this memory unavailable for normal
3271 use. Use this parameter to scan for
3272 corruption in more or less memory.
3274 memory_corruption_check_period=seconds [X86]
3275 By default it checks for corruption every 60
3276 seconds. Use this parameter to check at some
3277 other rate. 0 disables periodic checking.
3279 memory_hotplug.memmap_on_memory
3280 [KNL,X86,ARM] Boolean flag to enable this feature.
3281 Format: {on | off (default)}
3282 When enabled, runtime hotplugged memory will
3283 allocate its internal metadata (struct pages,
3284 those vmemmap pages cannot be optimized even
3285 if hugetlb_free_vmemmap is enabled) from the
3286 hotadded memory which will allow to hotadd a
3287 lot of memory without requiring additional
3289 This feature is disabled by default because it
3290 has some implication on large (e.g. GB)
3291 allocations in some configurations (e.g. small
3293 The state of the flag can be read in
3294 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3295 Note that even when enabled, there are a few cases where
3296 the feature is not effective.
3298 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3300 default : 0 <disable>
3301 Specifies the number of memtest passes to be
3302 performed. Each pass selects another test
3303 pattern from a given set of patterns. Memtest
3304 fills the memory with this pattern, validates
3305 memory contents and reserves bad memory
3306 regions that are detected.
3308 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3309 Valid arguments: on, off
3310 Default (depends on kernel configuration option):
3311 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3312 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3313 mem_encrypt=on: Activate SME
3314 mem_encrypt=off: Do not activate SME
3316 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3317 for details on when memory encryption can be activated.
3319 mem_sleep_default= [SUSPEND] Default system suspend mode:
3320 s2idle - Suspend-To-Idle
3321 shallow - Power-On Suspend or equivalent (if supported)
3322 deep - Suspend-To-RAM or equivalent (if supported)
3323 See Documentation/admin-guide/pm/sleep-states.rst.
3325 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3326 Multi-Function General Purpose Timers on AMD Geode
3329 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3330 the BIOS has incorrectly applied a workaround. TinyBIOS
3331 version 0.98 is known to be affected, 0.99 fixes the
3332 problem by letting the user disable the workaround.
3336 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3337 physical address is ignored.
3339 mini2440= [ARM,HW,KNL]
3340 Format:[0..2][b][c][t]
3342 MINI2440 configuration specification:
3343 0 - The attached screen is the 3.5" TFT
3344 1 - The attached screen is the 7" TFT
3345 2 - The VGA Shield is attached (1024x768)
3346 Leaving out the screen size parameter will not load
3347 the TFT driver, and the framebuffer will be left
3349 b - Enable backlight. The TFT backlight pin will be
3350 linked to the kernel VESA blanking code and a GPIO
3351 LED. This parameter is not necessary when using the
3353 c - Enable the s3c camera interface.
3354 t - Reserved for enabling touchscreen support. The
3355 touchscreen support is not enabled in the mainstream
3356 kernel as of 2.6.30, a preliminary port can be found
3357 in the "bleeding edge" mini2440 support kernel at
3358 https://repo.or.cz/w/linux-2.6/mini2440.git
3361 [X86,PPC,S390,ARM64] Control optional mitigations for
3362 CPU vulnerabilities. This is a set of curated,
3363 arch-independent options, each of which is an
3364 aggregation of existing arch-specific options.
3367 Disable all optional CPU mitigations. This
3368 improves system performance, but it may also
3369 expose users to several CPU vulnerabilities.
3370 Equivalent to: if nokaslr then kpti=0 [ARM64]
3371 gather_data_sampling=off [X86]
3372 kvm.nx_huge_pages=off [X86]
3375 mmio_stale_data=off [X86]
3376 no_entry_flush [PPC]
3377 no_uaccess_flush [PPC]
3380 nospectre_bhb [ARM64]
3381 nospectre_v1 [X86,PPC]
3382 nospectre_v2 [X86,PPC,S390,ARM64]
3384 spec_store_bypass_disable=off [X86,PPC]
3385 spectre_v2_user=off [X86]
3386 srbds=off [X86,INTEL]
3387 ssbd=force-off [ARM64]
3388 tsx_async_abort=off [X86]
3391 This does not have any effect on
3392 kvm.nx_huge_pages when
3393 kvm.nx_huge_pages=force.
3396 Mitigate all CPU vulnerabilities, but leave SMT
3397 enabled, even if it's vulnerable. This is for
3398 users who don't want to be surprised by SMT
3399 getting disabled across kernel upgrades, or who
3400 have other ways of avoiding SMT-based attacks.
3401 Equivalent to: (default behavior)
3404 Mitigate all CPU vulnerabilities, disabling SMT
3405 if needed. This is for users who always want to
3406 be fully mitigated, even if it means losing SMT.
3407 Equivalent to: l1tf=flush,nosmt [X86]
3408 mds=full,nosmt [X86]
3409 tsx_async_abort=full,nosmt [X86]
3410 mmio_stale_data=full,nosmt [X86]
3411 retbleed=auto,nosmt [X86]
3414 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3415 parameter allows control of the logging verbosity for
3416 the additional memory initialisation checks. A value
3417 of 0 disables mminit logging and a level of 4 will
3418 log everything. Information is printed at KERN_DEBUG
3419 so loglevel=8 may also need to be specified.
3422 [X86,INTEL] Control mitigation for the Processor
3423 MMIO Stale Data vulnerabilities.
3425 Processor MMIO Stale Data is a class of
3426 vulnerabilities that may expose data after an MMIO
3427 operation. Exposed data could originate or end in
3428 the same CPU buffers as affected by MDS and TAA.
3429 Therefore, similar to MDS and TAA, the mitigation
3430 is to clear the affected CPU buffers.
3432 This parameter controls the mitigation. The
3435 full - Enable mitigation on vulnerable CPUs
3437 full,nosmt - Enable mitigation and disable SMT on
3440 off - Unconditionally disable mitigation
3442 On MDS or TAA affected machines,
3443 mmio_stale_data=off can be prevented by an active
3444 MDS or TAA mitigation as these vulnerabilities are
3445 mitigated with the same mechanism so in order to
3446 disable this mitigation, you need to specify
3447 mds=off and tsx_async_abort=off too.
3449 Not specifying this option is equivalent to
3450 mmio_stale_data=full.
3453 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3455 <module>.async_probe[=<bool>] [KNL]
3456 If no <bool> value is specified or if the value
3457 specified is not a valid <bool>, enable asynchronous
3458 probe on this module. Otherwise, enable/disable
3459 asynchronous probe on this module as indicated by the
3460 <bool> value. See also: module.async_probe
3462 module.async_probe=<bool>
3463 [KNL] When set to true, modules will use async probing
3464 by default. To enable/disable async probing for a
3465 specific module, use the module specific control that
3466 is documented under <module>.async_probe. When both
3467 module.async_probe and <module>.async_probe are
3468 specified, <module>.async_probe takes precedence for
3469 the specific module.
3471 module.enable_dups_trace
3472 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set,
3473 this means that duplicate request_module() calls will
3474 trigger a WARN_ON() instead of a pr_warn(). Note that
3475 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s
3476 will always be issued and this option does nothing.
3478 [KNL] When CONFIG_MODULE_SIG is set, this means that
3479 modules without (valid) signatures will fail to load.
3480 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3481 is always true, so this option does nothing.
3483 module_blacklist= [KNL] Do not load a comma-separated list of
3484 modules. Useful for debugging problem modules.
3487 [MOUSE] Maximum time between finger touching and
3488 leaving touchpad surface for touch to be considered
3489 a tap and be reported as a left button click (for
3490 touchpads working in absolute mode only).
3492 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3493 reporting absolute coordinates, such as tablets
3494 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3495 reporting absolute coordinates, such as tablets
3497 movablecore= [KNL,X86,IA-64,PPC]
3498 Format: nn[KMGTPE] | nn%
3499 This parameter is the complement to kernelcore=, it
3500 specifies the amount of memory used for migratable
3501 allocations. If both kernelcore and movablecore is
3502 specified, then kernelcore will be at *least* the
3503 specified value but may be more. If movablecore on its
3504 own is specified, the administrator must be careful
3505 that the amount of memory usable for all allocations
3508 movable_node [KNL] Boot-time switch to make hotplugable memory
3509 NUMA nodes to be movable. This means that the memory
3510 of such nodes will be usable only for movable
3511 allocations which rules out almost all kernel
3512 allocations. Use with caution!
3514 MTD_Partition= [MTD]
3515 Format: <name>,<region-number>,<size>,<offset>
3517 MTD_Region= [MTD] Format:
3518 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3521 See drivers/mtd/parsers/cmdlinepart.c
3524 ARM/S3C2412 JIVE boot control
3526 See arch/arm/mach-s3c/mach-jive.c
3528 mtouchusb.raw_coordinates=
3529 [HW] Make the MicroTouch USB driver use raw coordinates
3530 ('y', default) or cooked coordinates ('n')
3533 Enable printing debug information related to MTRR
3534 registers at boot time.
3536 mtrr_chunk_size=nn[KMG] [X86]
3537 used for mtrr cleanup. It is largest continuous chunk
3538 that could hold holes aka. UC entries.
3540 mtrr_gran_size=nn[KMG] [X86]
3541 Used for mtrr cleanup. It is granularity of mtrr block.
3543 Large value could prevent small alignment from
3546 mtrr_spare_reg_nr=n [X86]
3548 Range: 0,7 : spare reg number
3550 Used for mtrr cleanup. It is spare mtrr entries number.
3551 Set to 2 or more if your graphical card needs more.
3553 multitce=off [PPC] This parameter disables the use of the pSeries
3554 firmware feature for updating multiple TCE entries
3557 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3559 netdev= [NET] Network devices parameters
3560 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3561 Note that mem_start is often overloaded to mean
3562 something different and driver-specific.
3563 This usage is only documented in each driver source
3566 netpoll.carrier_timeout=
3567 [NET] Specifies amount of time (in seconds) that
3568 netpoll should wait for a carrier. By default netpoll
3572 [NETFILTER] Enable connection tracking flow accounting
3573 0 to disable accounting
3574 1 to enable accounting
3578 [NFS] sets the pathname to the program which is used
3579 to update the NFS client cache entries.
3581 nfs.cache_getent_timeout=
3582 [NFS] sets the timeout after which an attempt to
3583 update a cache entry is deemed to have failed.
3585 nfs.callback_nr_threads=
3586 [NFSv4] set the total number of threads that the
3587 NFS client will assign to service NFSv4 callback
3590 nfs.callback_tcpport=
3591 [NFS] set the TCP port on which the NFSv4 callback
3592 channel should listen.
3595 [NFS] enable 64-bit inode numbers.
3596 If zero, the NFS client will fake up a 32-bit inode
3597 number for the readdir() and stat() syscalls instead
3598 of returning the full 64-bit number.
3599 The default is to return 64-bit inode numbers.
3601 nfs.idmap_cache_timeout=
3602 [NFS] set the maximum lifetime for idmapper cache
3605 nfs.max_session_cb_slots=
3606 [NFSv4.1] Sets the maximum number of session
3607 slots the client will assign to the callback
3608 channel. This determines the maximum number of
3609 callbacks the client will process in parallel for
3610 a particular server.
3612 nfs.max_session_slots=
3613 [NFSv4.1] Sets the maximum number of session slots
3614 the client will attempt to negotiate with the server.
3615 This limits the number of simultaneous RPC requests
3616 that the client can send to the NFSv4.1 server.
3617 Note that there is little point in setting this
3618 value higher than the max_tcp_slot_table_limit.
3620 nfs.nfs4_disable_idmapping=
3621 [NFSv4] When set to the default of '1', this option
3622 ensures that both the RPC level authentication
3623 scheme and the NFS level operations agree to use
3624 numeric uids/gids if the mount is using the
3625 'sec=sys' security flavour. In effect it is
3626 disabling idmapping, which can make migration from
3627 legacy NFSv2/v3 systems to NFSv4 easier.
3628 Servers that do not support this mode of operation
3629 will be autodetected by the client, and it will fall
3630 back to using the idmapper.
3631 To turn off this behaviour, set the value to '0'.
3634 [NFS4] Specify an additional fixed unique ident-
3635 ification string that NFSv4 clients can insert into
3636 their nfs_client_id4 string. This is typically a
3637 UUID that is generated at system install time.
3639 nfs.recover_lost_locks=
3640 [NFSv4] Attempt to recover locks that were lost due
3641 to a lease timeout on the server. Please note that
3642 doing this risks data corruption, since there are
3643 no guarantees that the file will remain unchanged
3644 after the locks are lost.
3645 If you want to enable the kernel legacy behaviour of
3646 attempting to recover these locks, then set this
3648 The default parameter value of '0' causes the kernel
3649 not to attempt recovery of lost locks.
3651 nfs.send_implementation_id=
3652 [NFSv4.1] Send client implementation identification
3653 information in exchange_id requests.
3654 If zero, no implementation identification information
3656 The default is to send the implementation identification
3659 nfs4.layoutstats_timer=
3660 [NFSv4.2] Change the rate at which the kernel sends
3661 layoutstats to the pNFS metadata server.
3663 Setting this to value to 0 causes the kernel to use
3664 whatever value is the default set by the layout
3665 driver. A non-zero value sets the minimum interval
3666 in seconds between layoutstats transmissions.
3668 nfsd.inter_copy_offload_enable=
3669 [NFSv4.2] When set to 1, the server will support
3670 server-to-server copies for which this server is
3671 the destination of the copy.
3673 nfsd.nfs4_disable_idmapping=
3674 [NFSv4] When set to the default of '1', the NFSv4
3675 server will return only numeric uids and gids to
3676 clients using auth_sys, and will accept numeric uids
3677 and gids from such clients. This is intended to ease
3678 migration from NFSv2/v3.
3680 nfsd.nfsd4_ssc_umount_timeout=
3681 [NFSv4.2] When used as the destination of a
3682 server-to-server copy, knfsd temporarily mounts
3683 the source server. It caches the mount in case
3684 it will be needed again, and discards it if not
3685 used for the number of milliseconds specified by
3688 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3689 See Documentation/admin-guide/nfs/nfsroot.rst.
3691 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3692 See Documentation/admin-guide/nfs/nfsroot.rst.
3694 nfsrootdebug [NFS] enable nfsroot debugging messages.
3695 See Documentation/admin-guide/nfs/nfsroot.rst.
3697 nmi_backtrace.backtrace_idle [KNL]
3698 Dump stacks even of idle CPUs in response to an
3699 NMI stack-backtrace request.
3701 nmi_debug= [KNL,SH] Specify one or more actions to take
3702 when a NMI is triggered.
3703 Format: [state][,regs][,debounce][,die]
3705 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3706 Format: [panic,][nopanic,][num]
3708 0 - turn hardlockup detector in nmi_watchdog off
3709 1 - turn hardlockup detector in nmi_watchdog on
3710 When panic is specified, panic when an NMI watchdog
3711 timeout occurs (or 'nopanic' to not panic on an NMI
3712 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3713 To disable both hard and soft lockup detectors,
3714 please see 'nowatchdog'.
3715 This is useful when you use a panic=... timeout and
3716 need the box quickly up again.
3718 These settings can be accessed at runtime via
3719 the nmi_watchdog and hardlockup_panic sysctls.
3721 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3722 emulation library even if a 387 maths coprocessor
3725 no4lvl [RISCV] Disable 4-level and 5-level paging modes. Forces
3726 kernel to use 3-level paging instead.
3728 no5lvl [X86-64,RISCV] Disable 5-level paging mode. Forces
3729 kernel to use 4-level paging instead.
3731 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3732 caches in the slab allocator. Saves per-node memory,
3733 but will impact performance.
3737 noaltinstr [S390] Disables alternative instructions patching
3738 (CPU alternatives feature).
3740 noapic [SMP,APIC] Tells the kernel to not make use of any
3741 IOAPICs that may be present in the system.
3743 noautogroup Disable scheduler automatic task group creation.
3748 [HW] Never suspend the console
3749 Disable suspending of consoles during suspend and
3750 hibernate operations. Once disabled, debugging
3751 messages can reach various consoles while the rest
3752 of the system is being put to sleep (ie, while
3753 debugging driver suspend/resume hooks). This may
3754 not work reliably with all consoles, but is known
3755 to work with serial and VGA consoles.
3756 To facilitate more flexible debugging, we also add
3757 console_suspend, a printk module parameter to control
3758 it. Users could use console_suspend (usually
3759 /sys/module/printk/parameters/console_suspend) to
3760 turn on/off it dynamically.
3763 [KNL] Disable object debugging
3765 nodsp [SH] Disable hardware DSP at boot time.
3767 noefi Disable EFI runtime services support.
3769 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3774 This affects only 32-bit executables.
3775 noexec32=on: enable non-executable mappings (default)
3776 read doesn't imply executable mappings
3777 noexec32=off: disable non-executable mappings
3778 read implies executable mappings
3780 no_file_caps Tells the kernel not to honor file capabilities. The
3781 only way then for a file to be executed with privilege
3782 is to be setuid root or executed by root.
3784 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3786 nofsgsbase [X86] Disables FSGSBASE instructions.
3788 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3789 register save and restore. The kernel will only save
3790 legacy floating-point registers on task switch.
3792 nohalt [IA-64] Tells the kernel not to use the power saving
3793 function PAL_HALT_LIGHT when idle. This increases
3794 power-consumption. On the positive side, it reduces
3795 interrupt wake-up latency, which may improve performance
3796 in certain environments such as networked servers or
3800 Force pointers printed to the console or buffers to be
3801 unhashed. By default, when a pointer is printed via %p
3802 format string, that pointer is "hashed", i.e. obscured
3803 by hashing the pointer value. This is a security feature
3804 that hides actual kernel addresses from unprivileged
3805 users, but it also makes debugging the kernel more
3806 difficult since unequal pointers can no longer be
3807 compared. However, if this command-line option is
3808 specified, then all normal pointers will have their true
3809 value printed. This option should only be specified when
3810 debugging the kernel. Please do not use on production
3813 nohibernate [HIBERNATION] Disable hibernation and resume.
3815 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,SH] Forces the kernel to
3816 busy wait in do_idle() and not use the arch_cpu_idle()
3817 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3818 to be effective. This is useful on platforms where the
3819 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3820 correctly or when doing power measurements to evaluate
3821 the impact of the sleep instructions. This is also
3822 useful when using JTAG debugger.
3824 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3826 nohugevmalloc [KNL,X86,PPC,ARM64] Disable kernel huge vmalloc mappings.
3828 nohz= [KNL] Boottime enable/disable dynamic ticks
3829 Valid arguments: on, off
3832 nohz_full= [KNL,BOOT,SMP,ISOL]
3833 The argument is a cpu list, as described above.
3834 In kernels built with CONFIG_NO_HZ_FULL=y, set
3835 the specified list of CPUs whose tick will be stopped
3836 whenever possible. The boot CPU will be forced outside
3837 the range to maintain the timekeeping. Any CPUs
3838 in this list will have their RCU callbacks offloaded,
3839 just as if they had also been called out in the
3840 rcu_nocbs= boot parameter.
3842 Note that this argument takes precedence over
3843 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3845 noinitrd [RAM] Tells the kernel not to load any configured
3848 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3850 [Deprecated - use intremap=off]
3854 noinvpcid [X86] Disable the INVPCID cpu feature.
3856 noiotrap [SH] Disables trapped I/O port accesses.
3858 noirqdebug [X86-32] Disables the code which attempts to detect and
3859 disable unhandled interrupt sources.
3861 noisapnp [ISAPNP] Disables ISA PnP code.
3863 nojitter [IA-64] Disables jitter checking for ITC timers.
3866 When CONFIG_RANDOMIZE_BASE is set, this disables
3867 kernel and module base offset ASLR (Address Space
3868 Layout Randomization).
3870 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3873 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3875 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3877 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3879 nomca [IA-64] Disable machine check abort handling
3881 nomce [X86-32] Disable Machine Check Exception
3883 nomfgpt [X86-32] Disable Multi-Function General Purpose
3884 Timer usage (for AMD Geode machines).
3886 nomodeset Disable kernel modesetting. Most systems' firmware
3887 sets up a display mode and provides framebuffer memory
3888 for output. With nomodeset, DRM and fbdev drivers will
3889 not load if they could possibly displace the pre-
3890 initialized output. Only the system framebuffer will
3891 be available for use. The respective drivers will not
3892 perform display-mode changes or accelerated rendering.
3894 Useful as error fallback, or for testing and debugging.
3896 nomodule Disable module load
3898 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3899 shutdown the other cpus. Instead use the REBOOT_VECTOR
3902 nopat [X86] Disable PAT (page attribute table extension of
3903 pagetables) support.
3905 nopcid [X86-64] Disable the PCID cpu feature.
3907 nopku [X86] Disable Memory Protection Keys CPU feature found
3911 Equivalent to pti=off
3913 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
3914 Disables the PV optimizations forcing the guest to run
3915 as generic guest with no PV drivers. Currently support
3916 XEN HVM, KVM, HYPER_V and VMWARE guest.
3918 nopvspin [X86,XEN,KVM]
3919 Disables the qspinlock slow path using PV optimizations
3920 which allow the hypervisor to 'idle' the guest on lock
3923 norandmaps Don't use address space randomization. Equivalent to
3924 echo 0 > /proc/sys/kernel/randomize_va_space
3926 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3927 with UP alternatives
3929 noresume [SWSUSP] Disables resume and restores original swap
3934 no-scroll [VGA] Disables scrollback.
3935 This is required for the Braillex ib80-piezo Braille
3936 reader made by F.H. Papenmeier (Germany).
3938 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3941 Disable SMAP (Supervisor Mode Access Prevention)
3942 even if it is supported by processor.
3945 Disable SMEP (Supervisor Mode Execution Prevention)
3946 even if it is supported by processor.
3948 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3949 and disable the IO APIC. legacy for "maxcpus=0".
3951 nosmt [KNL,MIPS,PPC,S390] Disable symmetric multithreading (SMT).
3952 Equivalent to smt=1.
3954 [KNL,X86,PPC] Disable symmetric multithreading (SMT).
3955 nosmt=force: Force disable SMT, cannot be undone
3956 via the sysfs control file.
3958 nosoftlockup [KNL] Disable the soft-lockup detector.
3960 nospec_store_bypass_disable
3961 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3963 nospectre_bhb [ARM64] Disable all mitigations for Spectre-BHB (branch
3964 history injection) vulnerability. System may allow data leaks
3967 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3968 (bounds check bypass). With this option data leaks are
3969 possible in the system.
3971 nospectre_v2 [X86,PPC_E500,ARM64] Disable all mitigations for
3972 the Spectre variant 2 (indirect branch prediction)
3973 vulnerability. System may allow data leaks with this
3976 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES] Disable paravirtualized
3977 steal time accounting. steal time is computed, but
3978 won't influence scheduler behaviour
3980 nosync [HW,M68K] Disables sync negotiation for all devices.
3982 no_timer_check [X86,APIC] Disables the code which tests for
3983 broken timer IRQ sources.
3986 [PPC] Don't flush the L1-D cache after accessing user data.
3988 novmcoredd [KNL,KDUMP]
3989 Disable device dump. Device dump allows drivers to
3990 append dump data to vmcore so you can collect driver
3991 specified debug info. Drivers can append the data
3992 without any limit and this data is stored in memory,
3993 so this may cause significant memory stress. Disabling
3994 device dump can help save memory but the driver debug
3995 data will be no longer available. This parameter
3996 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
4000 [X86,PV_OPS] Disable paravirtualized VMware scheduler
4001 clock and use the default one.
4003 nowatchdog [KNL] Disable both lockup detectors, i.e.
4004 soft-lockup and NMI watchdog (hard-lockup).
4008 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
4010 NOTE: this parameter will be ignored on systems with the
4011 LEGACY_XAPIC_DISABLED bit set in the
4012 IA32_XAPIC_DISABLE_STATUS MSR.
4014 noxsave [BUGS=X86] Disables x86 extended register state save
4015 and restore using xsave. The kernel will fallback to
4016 enabling legacy floating-point and sse state.
4018 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
4019 register states. The kernel will fall back to use
4020 xsave to save the states. By using this parameter,
4021 performance of saving the states is degraded because
4022 xsave doesn't support modified optimization while
4023 xsaveopt supports it on xsaveopt enabled systems.
4025 noxsaves [X86] Disables xsaves and xrstors used in saving and
4026 restoring x86 extended register state in compacted
4027 form of xsave area. The kernel will fall back to use
4028 xsaveopt and xrstor to save and restore the states
4029 in standard form of xsave area. By using this
4030 parameter, xsave area per process might occupy more
4031 memory on xsaves enabled systems.
4033 nps_mtm_hs_ctr= [KNL,ARC]
4034 This parameter sets the maximum duration, in
4035 cycles, each HW thread of the CTOP can run
4036 without interruptions, before HW switches it.
4037 The actual maximum duration is 16 times this
4039 Format: integer between 1 and 255
4042 nptcg= [IA-64] Override max number of concurrent global TLB
4043 purges which is reported from either PAL_VM_SUMMARY or
4046 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
4047 could support. nr_cpus=n : n >= 1 limits the kernel to
4048 support 'n' processors. It could be larger than the
4049 number of already plugged CPU during bootup, later in
4050 runtime you can physically add extra cpu until it reaches
4051 n. So during boot up some boot time memory for per-cpu
4052 variables need be pre-allocated for later physical cpu
4055 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
4057 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
4058 set up a single NUMA node spanning all memory.
4060 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
4062 Allowed values are enable and disable
4064 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
4065 'node', 'default' can be specified
4066 This can be set from sysctl after boot.
4067 See Documentation/admin-guide/sysctl/vm.rst for details.
4069 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
4070 See Documentation/core-api/debugging-via-ohci1394.rst for more
4073 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
4074 Rather than timing out after 20 ms if an EC
4075 command is not properly ACKed, override the length
4076 of the timeout. We have interrupts disabled while
4077 waiting for the ACK, so if this is set too high
4078 interrupts *may* be lost!
4080 omap_mux= [OMAP] Override bootloader pin multiplexing.
4081 Format: <mux_mode0.mode_name=value>...
4082 For example, to override I2C bus2:
4083 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
4085 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
4087 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
4089 boundary - index of last SLC block on Flex-OneNAND.
4090 The remaining blocks are configured as MLC blocks.
4091 lock - Configure if Flex-OneNAND boundary should be locked.
4092 Once locked, the boundary cannot be changed.
4093 1 indicates lock status, 0 indicates unlock status.
4095 oops=panic Always panic on oopses. Default is to just kill the
4096 process, but there is a small probability of
4097 deadlocking the machine.
4098 This will also cause panics on machine check exceptions.
4099 Useful together with panic=30 to trigger a reboot.
4102 [KNL] Boolean flag to control whether the page allocator
4103 should randomize its free lists. The randomization may
4104 be automatically enabled if the kernel detects it is
4105 running on a platform with a direct-mapped memory-side
4106 cache, and this parameter can be used to
4107 override/disable that behavior. The state of the flag
4108 can be read from sysfs at:
4109 /sys/module/page_alloc/parameters/shuffle.
4111 page_owner= [KNL] Boot-time page_owner enabling option.
4112 Storage of the information about who allocated
4113 each page is disabled in default. With this switch,
4115 on: enable the feature
4117 page_poison= [KNL] Boot-time parameter changing the state of
4118 poisoning on the buddy allocator, available with
4119 CONFIG_PAGE_POISONING=y.
4120 off: turn off poisoning (default)
4121 on: turn on poisoning
4123 page_reporting.page_reporting_order=
4124 [KNL] Minimal page reporting order
4126 Adjust the minimal page reporting order. The page
4127 reporting is disabled when it exceeds MAX_ORDER.
4129 panic= [KNL] Kernel behaviour on panic: delay <timeout>
4130 timeout > 0: seconds before rebooting
4131 timeout = 0: wait forever
4132 timeout < 0: reboot immediately
4135 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
4136 Format: <hex>[,nousertaint]
4137 Hexadecimal bitmask representing the set of TAINT flags
4138 that will cause the kernel to panic when add_taint() is
4139 called with any of the flags in this set.
4140 The optional switch "nousertaint" can be utilized to
4141 prevent userspace forced crashes by writing to sysctl
4142 /proc/sys/kernel/tainted any flagset matching with the
4143 bitmask set on panic_on_taint.
4144 See Documentation/admin-guide/tainted-kernels.rst for
4145 extra details on the taint flags that users can pick
4146 to compose the bitmask to assign to panic_on_taint.
4148 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump
4151 panic_print= Bitmask for printing system info when panic happens.
4152 User can chose combination of the following bits:
4153 bit 0: print all tasks info
4154 bit 1: print system memory info
4155 bit 2: print timer info
4156 bit 3: print locks info if CONFIG_LOCKDEP is on
4157 bit 4: print ftrace buffer
4158 bit 5: print all printk messages in buffer
4159 bit 6: print all CPUs backtrace (if available in the arch)
4160 *Be aware* that this option may print a _lot_ of lines,
4161 so there are risks of losing older messages in the log.
4162 Use this option carefully, maybe worth to setup a
4163 bigger log buffer with "log_buf_len" along with this.
4165 parkbd.port= [HW] Parallel port number the keyboard adapter is
4166 connected to, default is 0.
4168 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
4169 0 for XT, 1 for AT (default is AT).
4172 parport= [HW,PPT] Specify parallel ports. 0 disables.
4173 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
4174 Use 'auto' to force the driver to use any
4175 IRQ/DMA settings detected (the default is to
4176 ignore detected IRQ/DMA settings because of
4177 possible conflicts). You can specify the base
4178 address, IRQ, and DMA settings; IRQ and DMA
4179 should be numbers, or 'auto' (for using detected
4180 settings on that particular port), or 'nofifo'
4181 (to avoid using a FIFO even if it is detected).
4182 Parallel ports are assigned in the order they
4183 are specified on the command line, starting
4186 parport_init_mode= [HW,PPT]
4187 Configure VIA parallel port to operate in
4188 a specific mode. This is necessary on Pegasos
4189 computer where firmware has no options for setting
4190 up parallel port mode and sets it to spp.
4191 Currently this function knows 686a and 8231 chips.
4192 Format: [spp|ps2|epp|ecp|ecpepp]
4194 pata_legacy.all= [HW,LIBATA]
4196 Set to non-zero to probe primary and secondary ISA
4197 port ranges on PCI systems where no PCI PATA device
4198 has been found at either range. Disabled by default.
4200 pata_legacy.autospeed= [HW,LIBATA]
4202 Set to non-zero if a chip is present that snoops speed
4203 changes. Disabled by default.
4205 pata_legacy.ht6560a= [HW,LIBATA]
4207 Set to 1, 2, or 3 for HT 6560A on the primary channel,
4208 the secondary channel, or both channels respectively.
4209 Disabled by default.
4211 pata_legacy.ht6560b= [HW,LIBATA]
4213 Set to 1, 2, or 3 for HT 6560B on the primary channel,
4214 the secondary channel, or both channels respectively.
4215 Disabled by default.
4217 pata_legacy.iordy_mask= [HW,LIBATA]
4219 IORDY enable mask. Set individual bits to allow IORDY
4220 for the respective channel. Bit 0 is for the first
4221 legacy channel handled by this driver, bit 1 is for
4222 the second channel, and so on. The sequence will often
4223 correspond to the primary legacy channel, the secondary
4224 legacy channel, and so on, but the handling of a PCI
4225 bus and the use of other driver options may interfere
4226 with the sequence. By default IORDY is allowed across
4229 pata_legacy.opti82c46x= [HW,LIBATA]
4231 Set to 1, 2, or 3 for Opti 82c611A on the primary
4232 channel, the secondary channel, or both channels
4233 respectively. Disabled by default.
4235 pata_legacy.opti82c611a= [HW,LIBATA]
4237 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4238 channel, the secondary channel, or both channels
4239 respectively. Disabled by default.
4241 pata_legacy.pio_mask= [HW,LIBATA]
4243 PIO mode mask for autospeed devices. Set individual
4244 bits to allow the use of the respective PIO modes.
4245 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4246 All modes allowed by default.
4248 pata_legacy.probe_all= [HW,LIBATA]
4250 Set to non-zero to probe tertiary and further ISA
4251 port ranges on PCI systems. Disabled by default.
4253 pata_legacy.probe_mask= [HW,LIBATA]
4255 Probe mask for legacy ISA PATA ports. Depending on
4256 platform configuration and the use of other driver
4257 options up to 6 legacy ports are supported: 0x1f0,
4258 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4259 of individual ports can be disabled by setting the
4260 corresponding bits in the mask to 1. Bit 0 is for
4261 the first port in the list above (0x1f0), and so on.
4262 By default all supported ports are probed.
4264 pata_legacy.qdi= [HW,LIBATA]
4266 Set to non-zero to probe QDI controllers. By default
4267 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4269 pata_legacy.winbond= [HW,LIBATA]
4271 Set to non-zero to probe Winbond controllers. Use
4272 the standard I/O port (0x130) if 1, otherwise the
4273 value given is the I/O port to use (typically 0x1b0).
4274 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4277 pata_platform.pio_mask= [HW,LIBATA]
4279 Supported PIO mode mask. Set individual bits to allow
4280 the use of the respective PIO modes. Bit 0 is for
4281 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4285 Halt all CPUs after the first oops has been printed for
4286 the specified number of seconds. This is to be used if
4287 your oopses keep scrolling off the screen.
4291 pci=option[,option...] [PCI] various PCI subsystem options.
4293 Some options herein operate on a specific device
4294 or a set of devices (<pci_dev>). These are
4295 specified in one of the following formats:
4297 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4298 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4300 Note: the first format specifies a PCI
4301 bus/device/function address which may change
4302 if new hardware is inserted, if motherboard
4303 firmware changes, or due to changes caused
4304 by other kernel parameters. If the
4305 domain is left unspecified, it is
4306 taken to be zero. Optionally, a path
4307 to a device through multiple device/function
4308 addresses can be specified after the base
4309 address (this is more robust against
4310 renumbering issues). The second format
4311 selects devices using IDs from the
4312 configuration space which may match multiple
4313 devices in the system.
4315 earlydump dump PCI config space before the kernel
4317 off [X86] don't probe for the PCI bus
4318 bios [X86-32] force use of PCI BIOS, don't access
4319 the hardware directly. Use this if your machine
4320 has a non-standard PCI host bridge.
4321 nobios [X86-32] disallow use of PCI BIOS, only direct
4322 hardware access methods are allowed. Use this
4323 if you experience crashes upon bootup and you
4324 suspect they are caused by the BIOS.
4325 conf1 [X86] Force use of PCI Configuration Access
4326 Mechanism 1 (config address in IO port 0xCF8,
4327 data in IO port 0xCFC, both 32-bit).
4328 conf2 [X86] Force use of PCI Configuration Access
4329 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4330 the function, IO port 0xCFA, also 8-bit, sets
4331 bus number. The config space is then accessed
4332 through ports 0xC000-0xCFFF).
4333 See http://wiki.osdev.org/PCI for more info
4334 on the configuration access mechanisms.
4335 noaer [PCIE] If the PCIEAER kernel config parameter is
4336 enabled, this kernel boot option can be used to
4337 disable the use of PCIE advanced error reporting.
4338 nodomains [PCI] Disable support for multiple PCI
4339 root domains (aka PCI segments, in ACPI-speak).
4340 nommconf [X86] Disable use of MMCONFIG for PCI
4342 check_enable_amd_mmconf [X86] check for and enable
4343 properly configured MMIO access to PCI
4344 config space on AMD family 10h CPU
4345 nomsi [MSI] If the PCI_MSI kernel config parameter is
4346 enabled, this kernel boot option can be used to
4347 disable the use of MSI interrupts system-wide.
4348 noioapicquirk [APIC] Disable all boot interrupt quirks.
4349 Safety option to keep boot IRQs enabled. This
4350 should never be necessary.
4351 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4352 primary IO-APIC for bridges that cannot disable
4353 boot IRQs. This fixes a source of spurious IRQs
4354 when the system masks IRQs.
4355 noioapicreroute [APIC] Disable workaround that uses the
4356 boot IRQ equivalent of an IRQ that connects to
4357 a chipset where boot IRQs cannot be disabled.
4358 The opposite of ioapicreroute.
4359 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4360 routing table. These calls are known to be buggy
4361 on several machines and they hang the machine
4362 when used, but on other computers it's the only
4363 way to get the interrupt routing table. Try
4364 this option if the kernel is unable to allocate
4365 IRQs or discover secondary PCI buses on your
4367 rom [X86] Assign address space to expansion ROMs.
4368 Use with caution as certain devices share
4369 address decoders between ROMs and other
4371 norom [X86] Do not assign address space to
4372 expansion ROMs that do not already have
4373 BIOS assigned address ranges.
4374 nobar [X86] Do not assign address space to the
4375 BARs that weren't assigned by the BIOS.
4376 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4377 assigned automatically to PCI devices. You can
4378 make the kernel exclude IRQs of your ISA cards
4380 pirqaddr=0xAAAAA [X86] Specify the physical address
4381 of the PIRQ table (normally generated
4382 by the BIOS) if it is outside the
4383 F0000h-100000h range.
4384 lastbus=N [X86] Scan all buses thru bus #N. Can be
4385 useful if the kernel is unable to find your
4386 secondary buses and you want to tell it
4387 explicitly which ones they are.
4388 assign-busses [X86] Always assign all PCI bus
4389 numbers ourselves, overriding
4390 whatever the firmware may have done.
4391 usepirqmask [X86] Honor the possible IRQ mask stored
4392 in the BIOS $PIR table. This is needed on
4393 some systems with broken BIOSes, notably
4394 some HP Pavilion N5400 and Omnibook XE3
4395 notebooks. This will have no effect if ACPI
4396 IRQ routing is enabled.
4397 noacpi [X86] Do not use ACPI for IRQ routing
4398 or for PCI scanning.
4399 use_crs [X86] Use PCI host bridge window information
4400 from ACPI. On BIOSes from 2008 or later, this
4401 is enabled by default. If you need to use this,
4402 please report a bug.
4403 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4404 If you need to use this, please report a bug.
4405 use_e820 [X86] Use E820 reservations to exclude parts of
4406 PCI host bridge windows. This is a workaround
4407 for BIOS defects in host bridge _CRS methods.
4408 If you need to use this, please report a bug to
4409 <linux-pci@vger.kernel.org>.
4410 no_e820 [X86] Ignore E820 reservations for PCI host
4411 bridge windows. This is the default on modern
4412 hardware. If you need to use this, please report
4413 a bug to <linux-pci@vger.kernel.org>.
4414 routeirq Do IRQ routing for all PCI devices.
4415 This is normally done in pci_enable_device(),
4416 so this option is a temporary workaround
4417 for broken drivers that don't call it.
4418 skip_isa_align [X86] do not align io start addr, so can
4419 handle more pci cards
4420 noearly [X86] Don't do any early type 1 scanning.
4421 This might help on some broken boards which
4422 machine check when some devices' config space
4423 is read. But various workarounds are disabled
4424 and some IOMMU drivers will not work.
4425 bfsort Sort PCI devices into breadth-first order.
4426 This sorting is done to get a device
4427 order compatible with older (<= 2.4) kernels.
4428 nobfsort Don't sort PCI devices into breadth-first order.
4429 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4430 tuning and use the BIOS-configured MPS defaults.
4431 pcie_bus_safe Set every device's MPS to the largest value
4432 supported by all devices below the root complex.
4433 pcie_bus_perf Set device MPS to the largest allowable MPS
4434 based on its parent bus. Also set MRRS (Max
4435 Read Request Size) to the largest supported
4436 value (no larger than the MPS that the device
4437 or bus can support) for best performance.
4438 pcie_bus_peer2peer Set every device's MPS to 128B, which
4439 every device is guaranteed to support. This
4440 configuration allows peer-to-peer DMA between
4441 any pair of devices, possibly at the cost of
4442 reduced performance. This also guarantees
4443 that hot-added devices will work.
4444 cbiosize=nn[KMG] The fixed amount of bus space which is
4445 reserved for the CardBus bridge's IO window.
4446 The default value is 256 bytes.
4447 cbmemsize=nn[KMG] The fixed amount of bus space which is
4448 reserved for the CardBus bridge's memory
4449 window. The default value is 64 megabytes.
4452 [<order of align>@]<pci_dev>[; ...]
4453 Specifies alignment and device to reassign
4454 aligned memory resources. How to
4455 specify the device is described above.
4456 If <order of align> is not specified,
4457 PAGE_SIZE is used as alignment.
4458 A PCI-PCI bridge can be specified if resource
4459 windows need to be expanded.
4460 To specify the alignment for several
4461 instances of a device, the PCI vendor,
4462 device, subvendor, and subdevice may be
4463 specified, e.g., 12@pci:8086:9c22:103c:198f
4464 for 4096-byte alignment.
4465 ecrc= Enable/disable PCIe ECRC (transaction layer
4466 end-to-end CRC checking). Only effective if
4467 OS has native AER control (either granted by
4468 ACPI _OSC or forced via "pcie_ports=native")
4469 bios: Use BIOS/firmware settings. This is the
4473 hpiosize=nn[KMG] The fixed amount of bus space which is
4474 reserved for hotplug bridge's IO window.
4475 Default size is 256 bytes.
4476 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4477 reserved for hotplug bridge's MMIO window.
4478 Default size is 2 megabytes.
4479 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4480 reserved for hotplug bridge's MMIO_PREF window.
4481 Default size is 2 megabytes.
4482 hpmemsize=nn[KMG] The fixed amount of bus space which is
4483 reserved for hotplug bridge's MMIO and
4485 Default size is 2 megabytes.
4486 hpbussize=nn The minimum amount of additional bus numbers
4487 reserved for buses below a hotplug bridge.
4489 realloc= Enable/disable reallocating PCI bridge resources
4490 if allocations done by BIOS are too small to
4491 accommodate resources required by all child
4493 off: Turn realloc off
4495 realloc same as realloc=on
4496 noari do not use PCIe ARI.
4497 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4498 do not use PCIe ATS (and IOMMU device IOTLB).
4499 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4500 only look for one device below a PCIe downstream
4502 big_root_window Try to add a big 64bit memory window to the PCIe
4503 root complex on AMD CPUs. Some GFX hardware
4504 can resize a BAR to allow access to all VRAM.
4505 Adding the window is slightly risky (it may
4506 conflict with unreported devices), so this
4508 disable_acs_redir=<pci_dev>[; ...]
4509 Specify one or more PCI devices (in the format
4510 specified above) separated by semicolons.
4511 Each device specified will have the PCI ACS
4512 redirect capabilities forced off which will
4513 allow P2P traffic between devices through
4514 bridges without forcing it upstream. Note:
4515 this removes isolation between devices and
4516 may put more devices in an IOMMU group.
4517 force_floating [S390] Force usage of floating interrupts.
4518 nomio [S390] Do not use MIO instructions.
4519 norid [S390] ignore the RID field and force use of
4520 one PCI domain per PCI function
4522 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4525 force Enable ASPM even on devices that claim not to support it.
4526 WARNING: Forcing ASPM on may cause system lockups.
4528 pcie_ports= [PCIE] PCIe port services handling:
4529 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4530 even if the platform doesn't give the OS permission to
4531 use them. This may cause conflicts if the platform
4532 also tries to use these services.
4533 dpc-native Use native PCIe service for DPC only. May
4534 cause conflicts if firmware uses AER or DPC.
4535 compat Disable native PCIe services (PME, AER, DPC, PCIe
4538 pcie_port_pm= [PCIE] PCIe port power management handling:
4539 off Disable power management of all PCIe ports
4540 force Forcibly enable power management of all PCIe ports
4542 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4543 nomsi Do not use MSI for native PCIe PME signaling (this makes
4544 all PCIe root ports use INTx for all services).
4546 pcmv= [HW,PCMCIA] BadgePAD 4
4550 Keep all power-domains already enabled by bootloader on,
4551 even if no driver has claimed them. This is useful
4552 for debug and development, but should not be
4553 needed on a platform with proper driver support.
4555 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4558 See arch/parisc/kernel/pdc_chassis.c
4560 percpu_alloc= Select which percpu first chunk allocator to use.
4561 Currently supported values are "embed" and "page".
4562 Archs may support subset or none of the selections.
4563 See comments in mm/percpu.c for details on each
4564 allocator. This parameter is primarily for debugging
4565 and performance comparison.
4567 pirq= [SMP,APIC] Manual mp-table setup
4568 See Documentation/arch/x86/i386/IO-APIC.rst.
4570 plip= [PPT,NET] Parallel port network link
4571 Format: { parport<nr> | timid | 0 }
4572 See also Documentation/admin-guide/parport.rst.
4574 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4575 Override pmtimer IOPort with a hex value.
4578 pmu_override= [PPC] Override the PMU.
4579 This option takes over the PMU facility, so it is no
4580 longer usable by perf. Setting this option starts the
4581 PMU counters by setting MMCR0 to 0 (the FC bit is
4582 cleared). If a number is given, then MMCR1 is set to
4583 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4586 pm_debug_messages [SUSPEND,KNL]
4587 Enable suspend/resume debug messages during boot up.
4590 Enable PNP debug messages (depends on the
4591 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4592 via /sys/module/pnp/parameters/debug. We always show
4593 current resource usage; turning this on also shows
4594 possible settings and some assignment information.
4600 { on | off | curr | res | no-curr | no-res }
4603 [ISAPNP] Exclude IRQs for the autoconfiguration
4606 [ISAPNP] Exclude DMAs for the autoconfiguration
4608 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4609 Ranges are in pairs (I/O port base and size).
4612 [ISAPNP] Exclude memory regions for the
4614 Ranges are in pairs (memory base and size).
4616 ports= [IP_VS_FTP] IPVS ftp helper module
4618 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4620 Format: <port>,<port>....
4622 powersave=off [PPC] This option disables power saving features.
4623 It specifically disables cpuidle and sets the
4624 platform machine description specific power_save
4625 function to NULL. On Idle the CPU just reduces
4628 ppc_strict_facility_enable
4629 [PPC] This option catches any kernel floating point,
4630 Altivec, VSX and SPE outside of regions specifically
4631 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4632 There is some performance impact when enabling this.
4636 Disable Hardware Transactional Memory
4639 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4640 none - Limited to cond_resched() calls
4641 voluntary - Limited to cond_resched() and might_sleep() calls
4642 full - Any section that isn't explicitly preempt disabled
4643 can be preempted anytime.
4645 print-fatal-signals=
4646 [KNL] debug: print fatal signals
4648 If enabled, warn about various signal handling
4649 related application anomalies: too many signals,
4650 too many POSIX.1 timers, fatal signals causing a
4653 If you hit the warning due to signal overflow,
4654 you might want to try "ulimit -i unlimited".
4658 printk.always_kmsg_dump=
4659 Trigger kmsg_dump for cases other than kernel oops or
4661 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4664 printk.console_no_auto_verbose=
4665 Disable console loglevel raise on oops, panic
4666 or lockdep-detected issues (only if lock debug is on).
4667 With an exception to setups with low baudrate on
4668 serial console, keeping this 0 is a good choice
4669 in order to provide more debug information.
4671 default: 0 (auto_verbose is enabled)
4673 printk.devkmsg={on,off,ratelimit}
4674 Control writing to /dev/kmsg.
4675 on - unlimited logging to /dev/kmsg from userspace
4676 off - logging to /dev/kmsg disabled
4677 ratelimit - ratelimit the logging
4680 printk.time= Show timing data prefixed to each printk message line
4681 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4683 processor.max_cstate= [HW,ACPI]
4684 Limit processor to maximum C-state
4685 max_cstate=9 overrides any DMI blacklist limit.
4687 processor.nocst [HW,ACPI]
4688 Ignore the _CST method to determine C-states,
4689 instead using the legacy FADT method
4691 profile= [KNL] Enable kernel profiling via /proc/profile
4692 Format: [<profiletype>,]<number>
4693 Param: <profiletype>: "schedule", "sleep", or "kvm"
4694 [defaults to kernel profiling]
4695 Param: "schedule" - profile schedule points.
4696 Param: "sleep" - profile D-state sleeping (millisecs).
4697 Requires CONFIG_SCHEDSTATS
4698 Param: "kvm" - profile VM exits.
4699 Param: <number> - step/bucket size as a power of 2 for
4700 statistical time based profiling.
4702 prompt_ramdisk= [RAM] [Deprecated]
4704 prot_virt= [S390] enable hosting protected virtual machines
4705 isolated from the hypervisor (if hardware supports
4709 psi= [KNL] Enable or disable pressure stall information
4713 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4714 probe for; one of (bare|imps|exps|lifebook|any).
4715 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4717 psmouse.resetafter= [HW,MOUSE]
4718 Try to reset the device after so many bad packets
4721 [HW,MOUSE] Set desired mouse resolution, in dpi.
4722 psmouse.smartscroll=
4723 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4724 0 = disabled, 1 = enabled (default).
4726 pstore.backend= Specify the name of the pstore backend to use
4728 pti= [X86-64] Control Page Table Isolation of user and
4729 kernel address spaces. Disabling this feature
4730 removes hardening, but improves performance of
4731 system calls and interrupts.
4733 on - unconditionally enable
4734 off - unconditionally disable
4735 auto - kernel detects whether your CPU model is
4736 vulnerable to issues that PTI mitigates
4738 Not specifying this option is equivalent to pti=auto.
4741 [KNL] Number of legacy pty's. Overwrites compiled-in
4744 quiet [KNL] Disable most log messages
4748 radix_hcall_invalidate=on [PPC/PSERIES]
4749 Disable RADIX GTSE feature and use hcall for TLB
4753 See Documentation/admin-guide/md.rst.
4755 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4756 See Documentation/admin-guide/blockdev/ramdisk.rst.
4758 ramdisk_start= [RAM] RAM disk image start address
4760 random.trust_cpu=off
4761 [KNL] Disable trusting the use of the CPU's
4762 random number generator (if available) to
4763 initialize the kernel's RNG.
4765 random.trust_bootloader=off
4766 [KNL] Disable trusting the use of the a seed
4767 passed by the bootloader (if available) to
4768 initialize the kernel's RNG.
4770 randomize_kstack_offset=
4771 [KNL] Enable or disable kernel stack offset
4772 randomization, which provides roughly 5 bits of
4773 entropy, frustrating memory corruption attacks
4774 that depend on stack address determinism or
4775 cross-syscall address exposures. This is only
4776 available on architectures that have defined
4777 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4778 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4779 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4781 ras=option[,option,...] [KNL] RAS-specific options
4784 Disable the Correctable Errors Collector,
4785 see CONFIG_RAS_CEC help text.
4787 rcu_nocbs[=cpu-list]
4788 [KNL] The optional argument is a cpu list,
4791 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4792 enable the no-callback CPU mode, which prevents
4793 such CPUs' callbacks from being invoked in
4794 softirq context. Invocation of such CPUs' RCU
4795 callbacks will instead be offloaded to "rcuox/N"
4796 kthreads created for that purpose, where "x" is
4797 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4798 for the kthreads that mediate grace periods; and
4799 "N" is the CPU number. This reduces OS jitter on
4800 the offloaded CPUs, which can be useful for HPC
4801 and real-time workloads. It can also improve
4802 energy efficiency for asymmetric multiprocessors.
4804 If a cpulist is passed as an argument, the specified
4805 list of CPUs is set to no-callback mode from boot.
4807 Otherwise, if the '=' sign and the cpulist
4808 arguments are omitted, no CPU will be set to
4809 no-callback mode from boot but the mode may be
4810 toggled at runtime via cpusets.
4812 Note that this argument takes precedence over
4813 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4816 Rather than requiring that offloaded CPUs
4817 (specified by rcu_nocbs= above) explicitly
4818 awaken the corresponding "rcuoN" kthreads,
4819 make these kthreads poll for callbacks.
4820 This improves the real-time response for the
4821 offloaded CPUs by relieving them of the need to
4822 wake up the corresponding kthread, but degrades
4823 energy efficiency by requiring that the kthreads
4824 periodically wake up to do the polling.
4826 rcutree.blimit= [KNL]
4827 Set maximum number of finished RCU callbacks to
4828 process in one batch.
4830 rcutree.do_rcu_barrier= [KNL]
4831 Request a call to rcu_barrier(). This is
4832 throttled so that userspace tests can safely
4833 hammer on the sysfs variable if they so choose.
4834 If triggered before the RCU grace-period machinery
4835 is fully active, this will error out with EAGAIN.
4837 rcutree.dump_tree= [KNL]
4838 Dump the structure of the rcu_node combining tree
4839 out at early boot. This is used for diagnostic
4840 purposes, to verify correct tree setup.
4842 rcutree.gp_cleanup_delay= [KNL]
4843 Set the number of jiffies to delay each step of
4844 RCU grace-period cleanup.
4846 rcutree.gp_init_delay= [KNL]
4847 Set the number of jiffies to delay each step of
4848 RCU grace-period initialization.
4850 rcutree.gp_preinit_delay= [KNL]
4851 Set the number of jiffies to delay each step of
4852 RCU grace-period pre-initialization, that is,
4853 the propagation of recent CPU-hotplug changes up
4854 the rcu_node combining tree.
4856 rcutree.jiffies_till_first_fqs= [KNL]
4857 Set delay from grace-period initialization to
4858 first attempt to force quiescent states.
4859 Units are jiffies, minimum value is zero,
4860 and maximum value is HZ.
4862 rcutree.jiffies_till_next_fqs= [KNL]
4863 Set delay between subsequent attempts to force
4864 quiescent states. Units are jiffies, minimum
4865 value is one, and maximum value is HZ.
4867 rcutree.jiffies_till_sched_qs= [KNL]
4868 Set required age in jiffies for a
4869 given grace period before RCU starts
4870 soliciting quiescent-state help from
4871 rcu_note_context_switch() and cond_resched().
4872 If not specified, the kernel will calculate
4873 a value based on the most recent settings
4874 of rcutree.jiffies_till_first_fqs
4875 and rcutree.jiffies_till_next_fqs.
4876 This calculated value may be viewed in
4877 rcutree.jiffies_to_sched_qs. Any attempt to set
4878 rcutree.jiffies_to_sched_qs will be cheerfully
4881 rcutree.kthread_prio= [KNL,BOOT]
4882 Set the SCHED_FIFO priority of the RCU per-CPU
4883 kthreads (rcuc/N). This value is also used for
4884 the priority of the RCU boost threads (rcub/N)
4885 and for the RCU grace-period kthreads (rcu_bh,
4886 rcu_preempt, and rcu_sched). If RCU_BOOST is
4887 set, valid values are 1-99 and the default is 1
4888 (the least-favored priority). Otherwise, when
4889 RCU_BOOST is not set, valid values are 0-99 and
4890 the default is zero (non-realtime operation).
4891 When RCU_NOCB_CPU is set, also adjust the
4892 priority of NOCB callback kthreads.
4894 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4895 On callback-offloaded (rcu_nocbs) CPUs,
4896 RCU reduces the lock contention that would
4897 otherwise be caused by callback floods through
4898 use of the ->nocb_bypass list. However, in the
4899 common non-flooded case, RCU queues directly to
4900 the main ->cblist in order to avoid the extra
4901 overhead of the ->nocb_bypass list and its lock.
4902 But if there are too many callbacks queued during
4903 a single jiffy, RCU pre-queues the callbacks into
4904 the ->nocb_bypass queue. The definition of "too
4905 many" is supplied by this kernel boot parameter.
4907 rcutree.qhimark= [KNL]
4908 Set threshold of queued RCU callbacks beyond which
4909 batch limiting is disabled.
4911 rcutree.qlowmark= [KNL]
4912 Set threshold of queued RCU callbacks below which
4913 batch limiting is re-enabled.
4915 rcutree.qovld= [KNL]
4916 Set threshold of queued RCU callbacks beyond which
4917 RCU's force-quiescent-state scan will aggressively
4918 enlist help from cond_resched() and sched IPIs to
4919 help CPUs more quickly reach quiescent states.
4920 Set to less than zero to make this be set based
4921 on rcutree.qhimark at boot time and to zero to
4922 disable more aggressive help enlistment.
4924 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4925 Set the page-cache refill delay (in milliseconds)
4926 in response to low-memory conditions. The range
4927 of permitted values is in the range 0:100000.
4929 rcutree.rcu_divisor= [KNL]
4930 Set the shift-right count to use to compute
4931 the callback-invocation batch limit bl from
4932 the number of callbacks queued on this CPU.
4933 The result will be bounded below by the value of
4934 the rcutree.blimit kernel parameter. Every bl
4935 callbacks, the softirq handler will exit in
4936 order to allow the CPU to do other work.
4938 Please note that this callback-invocation batch
4939 limit applies only to non-offloaded callback
4940 invocation. Offloaded callbacks are instead
4941 invoked in the context of an rcuoc kthread, which
4942 scheduler will preempt as it does any other task.
4944 rcutree.rcu_fanout_exact= [KNL]
4945 Disable autobalancing of the rcu_node combining
4946 tree. This is used by rcutorture, and might
4947 possibly be useful for architectures having high
4948 cache-to-cache transfer latencies.
4950 rcutree.rcu_fanout_leaf= [KNL]
4951 Change the number of CPUs assigned to each
4952 leaf rcu_node structure. Useful for very
4953 large systems, which will choose the value 64,
4954 and for NUMA systems with large remote-access
4955 latencies, which will choose a value aligned
4956 with the appropriate hardware boundaries.
4958 rcutree.rcu_min_cached_objs= [KNL]
4959 Minimum number of objects which are cached and
4960 maintained per one CPU. Object size is equal
4961 to PAGE_SIZE. The cache allows to reduce the
4962 pressure to page allocator, also it makes the
4963 whole algorithm to behave better in low memory
4966 rcutree.rcu_nocb_gp_stride= [KNL]
4967 Set the number of NOCB callback kthreads in
4968 each group, which defaults to the square root
4969 of the number of CPUs. Larger numbers reduce
4970 the wakeup overhead on the global grace-period
4971 kthread, but increases that same overhead on
4972 each group's NOCB grace-period kthread.
4974 rcutree.rcu_kick_kthreads= [KNL]
4975 Cause the grace-period kthread to get an extra
4976 wake_up() if it sleeps three times longer than
4977 it should at force-quiescent-state time.
4978 This wake_up() will be accompanied by a
4979 WARN_ONCE() splat and an ftrace_dump().
4981 rcutree.rcu_resched_ns= [KNL]
4982 Limit the time spend invoking a batch of RCU
4983 callbacks to the specified number of nanoseconds.
4984 By default, this limit is checked only once
4985 every 32 callbacks in order to limit the pain
4986 inflicted by local_clock() overhead.
4988 rcutree.rcu_unlock_delay= [KNL]
4989 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4990 this specifies an rcu_read_unlock()-time delay
4991 in microseconds. This defaults to zero.
4992 Larger delays increase the probability of
4993 catching RCU pointer leaks, that is, buggy use
4994 of RCU-protected pointers after the relevant
4995 rcu_read_unlock() has completed.
4997 rcutree.sysrq_rcu= [KNL]
4998 Commandeer a sysrq key to dump out Tree RCU's
4999 rcu_node tree with an eye towards determining
5000 why a new grace period has not yet started.
5002 rcutree.use_softirq= [KNL]
5003 If set to zero, move all RCU_SOFTIRQ processing to
5004 per-CPU rcuc kthreads. Defaults to a non-zero
5005 value, meaning that RCU_SOFTIRQ is used by default.
5006 Specify rcutree.use_softirq=0 to use rcuc kthreads.
5008 But note that CONFIG_PREEMPT_RT=y kernels disable
5009 this kernel boot parameter, forcibly setting it
5012 rcuscale.gp_async= [KNL]
5013 Measure performance of asynchronous
5014 grace-period primitives such as call_rcu().
5016 rcuscale.gp_async_max= [KNL]
5017 Specify the maximum number of outstanding
5018 callbacks per writer thread. When a writer
5019 thread exceeds this limit, it invokes the
5020 corresponding flavor of rcu_barrier() to allow
5021 previously posted callbacks to drain.
5023 rcuscale.gp_exp= [KNL]
5024 Measure performance of expedited synchronous
5025 grace-period primitives.
5027 rcuscale.holdoff= [KNL]
5028 Set test-start holdoff period. The purpose of
5029 this parameter is to delay the start of the
5030 test until boot completes in order to avoid
5033 rcuscale.kfree_by_call_rcu= [KNL]
5034 In kernels built with CONFIG_RCU_LAZY=y, test
5035 call_rcu() instead of kfree_rcu().
5037 rcuscale.kfree_mult= [KNL]
5038 Instead of allocating an object of size kfree_obj,
5039 allocate one of kfree_mult * sizeof(kfree_obj).
5042 rcuscale.kfree_rcu_test= [KNL]
5043 Set to measure performance of kfree_rcu() flooding.
5045 rcuscale.kfree_rcu_test_double= [KNL]
5046 Test the double-argument variant of kfree_rcu().
5047 If this parameter has the same value as
5048 rcuscale.kfree_rcu_test_single, both the single-
5049 and double-argument variants are tested.
5051 rcuscale.kfree_rcu_test_single= [KNL]
5052 Test the single-argument variant of kfree_rcu().
5053 If this parameter has the same value as
5054 rcuscale.kfree_rcu_test_double, both the single-
5055 and double-argument variants are tested.
5057 rcuscale.kfree_nthreads= [KNL]
5058 The number of threads running loops of kfree_rcu().
5060 rcuscale.kfree_alloc_num= [KNL]
5061 Number of allocations and frees done in an iteration.
5063 rcuscale.kfree_loops= [KNL]
5064 Number of loops doing rcuscale.kfree_alloc_num number
5065 of allocations and frees.
5067 rcuscale.minruntime= [KNL]
5068 Set the minimum test run time in seconds. This
5069 does not affect the data-collection interval,
5070 but instead allows better measurement of things
5071 like CPU consumption.
5073 rcuscale.nreaders= [KNL]
5074 Set number of RCU readers. The value -1 selects
5075 N, where N is the number of CPUs. A value
5076 "n" less than -1 selects N-n+1, where N is again
5077 the number of CPUs. For example, -2 selects N
5078 (the number of CPUs), -3 selects N+1, and so on.
5079 A value of "n" less than or equal to -N selects
5082 rcuscale.nwriters= [KNL]
5083 Set number of RCU writers. The values operate
5084 the same as for rcuscale.nreaders.
5085 N, where N is the number of CPUs
5087 rcuscale.scale_type= [KNL]
5088 Specify the RCU implementation to test.
5090 rcuscale.shutdown= [KNL]
5091 Shut the system down after performance tests
5092 complete. This is useful for hands-off automated
5095 rcuscale.verbose= [KNL]
5096 Enable additional printk() statements.
5098 rcuscale.writer_holdoff= [KNL]
5099 Write-side holdoff between grace periods,
5100 in microseconds. The default of zero says
5103 rcuscale.writer_holdoff_jiffies= [KNL]
5104 Additional write-side holdoff between grace
5105 periods, but in jiffies. The default of zero
5108 rcutorture.fqs_duration= [KNL]
5109 Set duration of force_quiescent_state bursts
5112 rcutorture.fqs_holdoff= [KNL]
5113 Set holdoff time within force_quiescent_state bursts
5116 rcutorture.fqs_stutter= [KNL]
5117 Set wait time between force_quiescent_state bursts
5120 rcutorture.fwd_progress= [KNL]
5121 Specifies the number of kthreads to be used
5122 for RCU grace-period forward-progress testing
5123 for the types of RCU supporting this notion.
5124 Defaults to 1 kthread, values less than zero or
5125 greater than the number of CPUs cause the number
5128 rcutorture.fwd_progress_div= [KNL]
5129 Specify the fraction of a CPU-stall-warning
5130 period to do tight-loop forward-progress testing.
5132 rcutorture.fwd_progress_holdoff= [KNL]
5133 Number of seconds to wait between successive
5134 forward-progress tests.
5136 rcutorture.fwd_progress_need_resched= [KNL]
5137 Enclose cond_resched() calls within checks for
5138 need_resched() during tight-loop forward-progress
5141 rcutorture.gp_cond= [KNL]
5142 Use conditional/asynchronous update-side
5143 primitives, if available.
5145 rcutorture.gp_exp= [KNL]
5146 Use expedited update-side primitives, if available.
5148 rcutorture.gp_normal= [KNL]
5149 Use normal (non-expedited) asynchronous
5150 update-side primitives, if available.
5152 rcutorture.gp_sync= [KNL]
5153 Use normal (non-expedited) synchronous
5154 update-side primitives, if available. If all
5155 of rcutorture.gp_cond=, rcutorture.gp_exp=,
5156 rcutorture.gp_normal=, and rcutorture.gp_sync=
5157 are zero, rcutorture acts as if is interpreted
5158 they are all non-zero.
5160 rcutorture.irqreader= [KNL]
5161 Run RCU readers from irq handlers, or, more
5162 accurately, from a timer handler. Not all RCU
5163 flavors take kindly to this sort of thing.
5165 rcutorture.leakpointer= [KNL]
5166 Leak an RCU-protected pointer out of the reader.
5167 This can of course result in splats, and is
5168 intended to test the ability of things like
5169 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
5172 rcutorture.n_barrier_cbs= [KNL]
5173 Set callbacks/threads for rcu_barrier() testing.
5175 rcutorture.nfakewriters= [KNL]
5176 Set number of concurrent RCU writers. These just
5177 stress RCU, they don't participate in the actual
5178 test, hence the "fake".
5180 rcutorture.nocbs_nthreads= [KNL]
5181 Set number of RCU callback-offload togglers.
5182 Zero (the default) disables toggling.
5184 rcutorture.nocbs_toggle= [KNL]
5185 Set the delay in milliseconds between successive
5186 callback-offload toggling attempts.
5188 rcutorture.nreaders= [KNL]
5189 Set number of RCU readers. The value -1 selects
5190 N-1, where N is the number of CPUs. A value
5191 "n" less than -1 selects N-n-2, where N is again
5192 the number of CPUs. For example, -2 selects N
5193 (the number of CPUs), -3 selects N+1, and so on.
5195 rcutorture.object_debug= [KNL]
5196 Enable debug-object double-call_rcu() testing.
5198 rcutorture.onoff_holdoff= [KNL]
5199 Set time (s) after boot for CPU-hotplug testing.
5201 rcutorture.onoff_interval= [KNL]
5202 Set time (jiffies) between CPU-hotplug operations,
5203 or zero to disable CPU-hotplug testing.
5205 rcutorture.read_exit= [KNL]
5206 Set the number of read-then-exit kthreads used
5207 to test the interaction of RCU updaters and
5208 task-exit processing.
5210 rcutorture.read_exit_burst= [KNL]
5211 The number of times in a given read-then-exit
5212 episode that a set of read-then-exit kthreads
5215 rcutorture.read_exit_delay= [KNL]
5216 The delay, in seconds, between successive
5217 read-then-exit testing episodes.
5219 rcutorture.shuffle_interval= [KNL]
5220 Set task-shuffle interval (s). Shuffling tasks
5221 allows some CPUs to go into dyntick-idle mode
5222 during the rcutorture test.
5224 rcutorture.shutdown_secs= [KNL]
5225 Set time (s) after boot system shutdown. This
5226 is useful for hands-off automated testing.
5228 rcutorture.stall_cpu= [KNL]
5229 Duration of CPU stall (s) to test RCU CPU stall
5230 warnings, zero to disable.
5232 rcutorture.stall_cpu_block= [KNL]
5233 Sleep while stalling if set. This will result
5234 in warnings from preemptible RCU in addition to
5235 any other stall-related activity. Note that
5236 in kernels built with CONFIG_PREEMPTION=n and
5237 CONFIG_PREEMPT_COUNT=y, this parameter will
5238 cause the CPU to pass through a quiescent state.
5239 Given CONFIG_PREEMPTION=n, this will suppress
5240 RCU CPU stall warnings, but will instead result
5241 in scheduling-while-atomic splats.
5243 Use of this module parameter results in splats.
5246 rcutorture.stall_cpu_holdoff= [KNL]
5247 Time to wait (s) after boot before inducing stall.
5249 rcutorture.stall_cpu_irqsoff= [KNL]
5250 Disable interrupts while stalling if set.
5252 rcutorture.stall_gp_kthread= [KNL]
5253 Duration (s) of forced sleep within RCU
5254 grace-period kthread to test RCU CPU stall
5255 warnings, zero to disable. If both stall_cpu
5256 and stall_gp_kthread are specified, the
5257 kthread is starved first, then the CPU.
5259 rcutorture.stat_interval= [KNL]
5260 Time (s) between statistics printk()s.
5262 rcutorture.stutter= [KNL]
5263 Time (s) to stutter testing, for example, specifying
5264 five seconds causes the test to run for five seconds,
5265 wait for five seconds, and so on. This tests RCU's
5266 ability to transition abruptly to and from idle.
5268 rcutorture.test_boost= [KNL]
5269 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5270 "Maybe" means test if the RCU implementation
5271 under test support RCU priority boosting.
5273 rcutorture.test_boost_duration= [KNL]
5274 Duration (s) of each individual boost test.
5276 rcutorture.test_boost_interval= [KNL]
5277 Interval (s) between each boost test.
5279 rcutorture.test_no_idle_hz= [KNL]
5280 Test RCU's dyntick-idle handling. See also the
5281 rcutorture.shuffle_interval parameter.
5283 rcutorture.torture_type= [KNL]
5284 Specify the RCU implementation to test.
5286 rcutorture.verbose= [KNL]
5287 Enable additional printk() statements.
5289 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5290 Dump ftrace buffer after reporting RCU CPU
5293 rcupdate.rcu_cpu_stall_suppress= [KNL]
5294 Suppress RCU CPU stall warning messages.
5296 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5297 Suppress RCU CPU stall warning messages and
5298 rcutorture writer stall warnings that occur
5299 during early boot, that is, during the time
5300 before the init task is spawned.
5302 rcupdate.rcu_cpu_stall_timeout= [KNL]
5303 Set timeout for RCU CPU stall warning messages.
5304 The value is in seconds and the maximum allowed
5305 value is 300 seconds.
5307 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5308 Set timeout for expedited RCU CPU stall warning
5309 messages. The value is in milliseconds
5310 and the maximum allowed value is 21000
5311 milliseconds. Please note that this value is
5312 adjusted to an arch timer tick resolution.
5313 Setting this to zero causes the value from
5314 rcupdate.rcu_cpu_stall_timeout to be used (after
5315 conversion from seconds to milliseconds).
5317 rcupdate.rcu_cpu_stall_cputime= [KNL]
5318 Provide statistics on the cputime and count of
5319 interrupts and tasks during the sampling period. For
5320 multiple continuous RCU stalls, all sampling periods
5321 begin at half of the first RCU stall timeout.
5323 rcupdate.rcu_exp_stall_task_details= [KNL]
5324 Print stack dumps of any tasks blocking the
5325 current expedited RCU grace period during an
5326 expedited RCU CPU stall warning.
5328 rcupdate.rcu_expedited= [KNL]
5329 Use expedited grace-period primitives, for
5330 example, synchronize_rcu_expedited() instead
5331 of synchronize_rcu(). This reduces latency,
5332 but can increase CPU utilization, degrade
5333 real-time latency, and degrade energy efficiency.
5334 No effect on CONFIG_TINY_RCU kernels.
5336 rcupdate.rcu_normal= [KNL]
5337 Use only normal grace-period primitives,
5338 for example, synchronize_rcu() instead of
5339 synchronize_rcu_expedited(). This improves
5340 real-time latency, CPU utilization, and
5341 energy efficiency, but can expose users to
5342 increased grace-period latency. This parameter
5343 overrides rcupdate.rcu_expedited. No effect on
5344 CONFIG_TINY_RCU kernels.
5346 rcupdate.rcu_normal_after_boot= [KNL]
5347 Once boot has completed (that is, after
5348 rcu_end_inkernel_boot() has been invoked), use
5349 only normal grace-period primitives. No effect
5350 on CONFIG_TINY_RCU kernels.
5352 But note that CONFIG_PREEMPT_RT=y kernels enables
5353 this kernel boot parameter, forcibly setting
5354 it to the value one, that is, converting any
5355 post-boot attempt at an expedited RCU grace
5356 period to instead use normal non-expedited
5357 grace-period processing.
5359 rcupdate.rcu_task_collapse_lim= [KNL]
5360 Set the maximum number of callbacks present
5361 at the beginning of a grace period that allows
5362 the RCU Tasks flavors to collapse back to using
5363 a single callback queue. This switching only
5364 occurs when rcupdate.rcu_task_enqueue_lim is
5365 set to the default value of -1.
5367 rcupdate.rcu_task_contend_lim= [KNL]
5368 Set the minimum number of callback-queuing-time
5369 lock-contention events per jiffy required to
5370 cause the RCU Tasks flavors to switch to per-CPU
5371 callback queuing. This switching only occurs
5372 when rcupdate.rcu_task_enqueue_lim is set to
5373 the default value of -1.
5375 rcupdate.rcu_task_enqueue_lim= [KNL]
5376 Set the number of callback queues to use for the
5377 RCU Tasks family of RCU flavors. The default
5378 of -1 allows this to be automatically (and
5379 dynamically) adjusted. This parameter is intended
5382 rcupdate.rcu_task_ipi_delay= [KNL]
5383 Set time in jiffies during which RCU tasks will
5384 avoid sending IPIs, starting with the beginning
5385 of a given grace period. Setting a large
5386 number avoids disturbing real-time workloads,
5387 but lengthens grace periods.
5389 rcupdate.rcu_task_lazy_lim= [KNL]
5390 Number of callbacks on a given CPU that will
5391 cancel laziness on that CPU. Use -1 to disable
5392 cancellation of laziness, but be advised that
5393 doing so increases the danger of OOM due to
5396 rcupdate.rcu_task_stall_info= [KNL]
5397 Set initial timeout in jiffies for RCU task stall
5398 informational messages, which give some indication
5399 of the problem for those not patient enough to
5400 wait for ten minutes. Informational messages are
5401 only printed prior to the stall-warning message
5402 for a given grace period. Disable with a value
5403 less than or equal to zero. Defaults to ten
5404 seconds. A change in value does not take effect
5405 until the beginning of the next grace period.
5407 rcupdate.rcu_task_stall_info_mult= [KNL]
5408 Multiplier for time interval between successive
5409 RCU task stall informational messages for a given
5410 RCU tasks grace period. This value is clamped
5411 to one through ten, inclusive. It defaults to
5412 the value three, so that the first informational
5413 message is printed 10 seconds into the grace
5414 period, the second at 40 seconds, the third at
5415 160 seconds, and then the stall warning at 600
5416 seconds would prevent a fourth at 640 seconds.
5418 rcupdate.rcu_task_stall_timeout= [KNL]
5419 Set timeout in jiffies for RCU task stall
5420 warning messages. Disable with a value less
5421 than or equal to zero. Defaults to ten minutes.
5422 A change in value does not take effect until
5423 the beginning of the next grace period.
5425 rcupdate.rcu_tasks_lazy_ms= [KNL]
5426 Set timeout in milliseconds RCU Tasks asynchronous
5427 callback batching for call_rcu_tasks().
5428 A negative value will take the default. A value
5429 of zero will disable batching. Batching is
5430 always disabled for synchronize_rcu_tasks().
5432 rcupdate.rcu_tasks_rude_lazy_ms= [KNL]
5433 Set timeout in milliseconds RCU Tasks
5434 Rude asynchronous callback batching for
5435 call_rcu_tasks_rude(). A negative value
5436 will take the default. A value of zero will
5437 disable batching. Batching is always disabled
5438 for synchronize_rcu_tasks_rude().
5440 rcupdate.rcu_tasks_trace_lazy_ms= [KNL]
5441 Set timeout in milliseconds RCU Tasks
5442 Trace asynchronous callback batching for
5443 call_rcu_tasks_trace(). A negative value
5444 will take the default. A value of zero will
5445 disable batching. Batching is always disabled
5446 for synchronize_rcu_tasks_trace().
5448 rcupdate.rcu_self_test= [KNL]
5449 Run the RCU early boot self tests
5453 Run specified binary instead of /init from the ramdisk,
5454 used for early userspace startup. See initrd.
5457 force - Override the decision by the kernel to hide the
5458 advertisement of RDRAND support (this affects
5459 certain AMD processors because of buggy BIOS
5460 support, specifically around the suspend/resume
5464 Turn on/off individual RDT features. List is:
5465 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5467 E.g. to turn on cmt and turn off mba use:
5471 Format (x86 or x86_64):
5472 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5474 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5476 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5477 (prefix with 'panic_' to set mode for panic
5479 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5480 reboot_force is either force or not specified,
5481 reboot_cpu is s[mp]#### with #### being the processor
5482 to be used for rebooting.
5484 refscale.holdoff= [KNL]
5485 Set test-start holdoff period. The purpose of
5486 this parameter is to delay the start of the
5487 test until boot completes in order to avoid
5490 refscale.lookup_instances= [KNL]
5491 Number of data elements to use for the forms of
5492 SLAB_TYPESAFE_BY_RCU testing. A negative number
5493 is negated and multiplied by nr_cpu_ids, while
5494 zero specifies nr_cpu_ids.
5496 refscale.loops= [KNL]
5497 Set the number of loops over the synchronization
5498 primitive under test. Increasing this number
5499 reduces noise due to loop start/end overhead,
5500 but the default has already reduced the per-pass
5501 noise to a handful of picoseconds on ca. 2020
5504 refscale.nreaders= [KNL]
5505 Set number of readers. The default value of -1
5506 selects N, where N is roughly 75% of the number
5507 of CPUs. A value of zero is an interesting choice.
5509 refscale.nruns= [KNL]
5510 Set number of runs, each of which is dumped onto
5513 refscale.readdelay= [KNL]
5514 Set the read-side critical-section duration,
5515 measured in microseconds.
5517 refscale.scale_type= [KNL]
5518 Specify the read-protection implementation to test.
5520 refscale.shutdown= [KNL]
5521 Shut down the system at the end of the performance
5522 test. This defaults to 1 (shut it down) when
5523 refscale is built into the kernel and to 0 (leave
5524 it running) when refscale is built as a module.
5526 refscale.verbose= [KNL]
5527 Enable additional printk() statements.
5529 refscale.verbose_batched= [KNL]
5530 Batch the additional printk() statements. If zero
5531 (the default) or negative, print everything. Otherwise,
5532 print every Nth verbose statement, where N is the value
5536 [KNL, SMP] Set scheduler's default relax_domain_level.
5537 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5539 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5540 Format: <base1>,<size1>[,<base2>,<size2>,...]
5541 Reserve I/O ports or memory so the kernel won't use
5542 them. If <base> is less than 0x10000, the region
5543 is assumed to be I/O ports; otherwise it is memory.
5545 reservetop= [X86-32]
5547 Reserves a hole at the top of the kernel virtual
5550 reset_devices [KNL] Force drivers to reset the underlying device
5551 during initialization.
5554 Specify the partition device for software suspend
5556 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5558 resume_offset= [SWSUSP]
5559 Specify the offset from the beginning of the partition
5560 given by "resume=" at which the swap header is located,
5561 in <PAGE_SIZE> units (needed only for swap files).
5562 See Documentation/power/swsusp-and-swap-files.rst
5564 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5565 read the resume files
5567 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5568 Useful for devices that are detected asynchronously
5569 (e.g. USB and MMC devices).
5571 retain_initrd [RAM] Keep initrd memory after extraction
5573 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5574 Speculative Code Execution with Return Instructions)
5577 AMD-based UNRET and IBPB mitigations alone do not stop
5578 sibling threads from influencing the predictions of other
5579 sibling threads. For that reason, STIBP is used on pro-
5580 cessors that support it, and mitigate SMT on processors
5584 auto - automatically select a migitation
5585 auto,nosmt - automatically select a mitigation,
5586 disabling SMT if necessary for
5587 the full mitigation (only on Zen1
5588 and older without STIBP).
5589 ibpb - On AMD, mitigate short speculation
5590 windows on basic block boundaries too.
5591 Safe, highest perf impact. It also
5592 enables STIBP if present. Not suitable
5594 ibpb,nosmt - Like "ibpb" above but will disable SMT
5595 when STIBP is not available. This is
5596 the alternative for systems which do not
5598 unret - Force enable untrained return thunks,
5599 only effective on AMD f15h-f17h based
5601 unret,nosmt - Like unret, but will disable SMT when STIBP
5602 is not available. This is the alternative for
5603 systems which do not have STIBP.
5605 Selecting 'auto' will choose a mitigation method at run
5606 time according to the CPU.
5608 Not specifying this option is equivalent to retbleed=auto.
5610 rfkill.default_state=
5611 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5612 etc. communication is blocked by default.
5615 rfkill.master_switch_mode=
5616 0 The "airplane mode" button does nothing.
5617 1 The "airplane mode" button toggles between everything
5618 blocked and the previous configuration.
5619 2 The "airplane mode" button toggles between everything
5620 blocked and everything unblocked.
5622 rhash_entries= [KNL,NET]
5623 Set number of hash buckets for route cache
5626 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5629 riscv_isa_fallback [RISCV]
5630 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit
5631 falling back to detecting extension support by parsing
5632 "riscv,isa" property on devicetree systems when the
5633 replacement properties are not found. See the Kconfig
5634 entry for RISCV_ISA_FALLBACK.
5636 ro [KNL] Mount root device read-only on boot
5639 on Mark read-only kernel memory as read-only (default).
5640 off Leave read-only kernel memory writable for debugging.
5641 full Mark read-only kernel memory and aliases as read-only
5645 Enable the uart passthrough on the designated usb port
5646 on Rockchip SoCs. When active, the signals of the
5647 debug-uart get routed to the D+ and D- pins of the usb
5648 port and the regular usb controller gets disabled.
5650 root= [KNL] Root filesystem
5651 Usually this a a block device specifier of some kind,
5652 see the early_lookup_bdev comment in
5653 block/early-lookup.c for details.
5654 Alternatively this can be "ram" for the legacy initial
5655 ramdisk, "nfs" and "cifs" for root on a network file
5656 system, or "mtd" and "ubi" for mounting from raw flash.
5658 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5659 mount the root filesystem
5661 rootflags= [KNL] Set root filesystem mount option string
5663 rootfstype= [KNL] Set root filesystem type
5665 rootwait [KNL] Wait (indefinitely) for root device to show up.
5666 Useful for devices that are detected asynchronously
5667 (e.g. USB and MMC devices).
5669 rootwait= [KNL] Maximum time (in seconds) to wait for root device
5670 to show up before attempting to mount the root
5673 rproc_mem=nn[KMG][@address]
5674 [KNL,ARM,CMA] Remoteproc physical memory block.
5675 Memory area to be used by remote processor image,
5678 rw [KNL] Mount root device read-write on boot
5680 S [KNL] Run init in single mode
5682 s390_iommu= [HW,S390]
5683 Set s390 IOTLB flushing mode
5685 With strict flushing every unmap operation will result in
5686 an IOTLB flush. Default is lazy flushing before reuse,
5689 s390_iommu_aperture= [KNL,S390]
5690 Specifies the size of the per device DMA address space
5691 accessible through the DMA and IOMMU APIs as a decimal
5692 factor of the size of main memory.
5693 The default is 1 meaning that one can concurrently use
5694 as many DMA addresses as physical memory is installed,
5695 if supported by hardware, and thus map all of memory
5696 once. With a value of 2 one can map all of memory twice
5697 and so on. As a special case a factor of 0 imposes no
5698 restrictions other than those given by hardware at the
5699 cost of significant additional memory use for tables.
5702 See drivers/net/irda/sa1100_ir.c.
5704 sched_verbose [KNL] Enables verbose scheduler debug messages.
5706 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5707 Allowed values are enable and disable. This feature
5708 incurs a small amount of overhead in the scheduler
5709 but is useful for debugging and performance tuning.
5711 sched_thermal_decay_shift=
5712 [KNL, SMP] Set a decay shift for scheduler thermal
5713 pressure signal. Thermal pressure signal follows the
5714 default decay period of other scheduler pelt
5715 signals(usually 32 ms but configurable). Setting
5716 sched_thermal_decay_shift will left shift the decay
5717 period for the thermal pressure signal by the shift
5719 i.e. with the default pelt decay period of 32 ms
5720 sched_thermal_decay_shift thermal pressure decay pr
5724 Format: integer between 0 and 10
5727 scftorture.holdoff= [KNL]
5728 Number of seconds to hold off before starting
5729 test. Defaults to zero for module insertion and
5730 to 10 seconds for built-in smp_call_function()
5733 scftorture.longwait= [KNL]
5734 Request ridiculously long waits randomly selected
5735 up to the chosen limit in seconds. Zero (the
5736 default) disables this feature. Please note
5737 that requesting even small non-zero numbers of
5738 seconds can result in RCU CPU stall warnings,
5739 softlockup complaints, and so on.
5741 scftorture.nthreads= [KNL]
5742 Number of kthreads to spawn to invoke the
5743 smp_call_function() family of functions.
5744 The default of -1 specifies a number of kthreads
5745 equal to the number of CPUs.
5747 scftorture.onoff_holdoff= [KNL]
5748 Number seconds to wait after the start of the
5749 test before initiating CPU-hotplug operations.
5751 scftorture.onoff_interval= [KNL]
5752 Number seconds to wait between successive
5753 CPU-hotplug operations. Specifying zero (which
5754 is the default) disables CPU-hotplug operations.
5756 scftorture.shutdown_secs= [KNL]
5757 The number of seconds following the start of the
5758 test after which to shut down the system. The
5759 default of zero avoids shutting down the system.
5760 Non-zero values are useful for automated tests.
5762 scftorture.stat_interval= [KNL]
5763 The number of seconds between outputting the
5764 current test statistics to the console. A value
5765 of zero disables statistics output.
5767 scftorture.stutter_cpus= [KNL]
5768 The number of jiffies to wait between each change
5769 to the set of CPUs under test.
5771 scftorture.use_cpus_read_lock= [KNL]
5772 Use use_cpus_read_lock() instead of the default
5773 preempt_disable() to disable CPU hotplug
5774 while invoking one of the smp_call_function*()
5777 scftorture.verbose= [KNL]
5778 Enable additional printk() statements.
5780 scftorture.weight_single= [KNL]
5781 The probability weighting to use for the
5782 smp_call_function_single() function with a zero
5783 "wait" parameter. A value of -1 selects the
5784 default if all other weights are -1. However,
5785 if at least one weight has some other value, a
5786 value of -1 will instead select a weight of zero.
5788 scftorture.weight_single_wait= [KNL]
5789 The probability weighting to use for the
5790 smp_call_function_single() function with a
5791 non-zero "wait" parameter. See weight_single.
5793 scftorture.weight_many= [KNL]
5794 The probability weighting to use for the
5795 smp_call_function_many() function with a zero
5796 "wait" parameter. See weight_single.
5797 Note well that setting a high probability for
5798 this weighting can place serious IPI load
5801 scftorture.weight_many_wait= [KNL]
5802 The probability weighting to use for the
5803 smp_call_function_many() function with a
5804 non-zero "wait" parameter. See weight_single
5807 scftorture.weight_all= [KNL]
5808 The probability weighting to use for the
5809 smp_call_function_all() function with a zero
5810 "wait" parameter. See weight_single and
5813 scftorture.weight_all_wait= [KNL]
5814 The probability weighting to use for the
5815 smp_call_function_all() function with a
5816 non-zero "wait" parameter. See weight_single
5819 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5820 xtime_lock contention on larger systems, and/or RCU lock
5821 contention on all systems with CONFIG_MAXSMP set.
5822 Format: { "0" | "1" }
5823 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5825 Note: increases power consumption, thus should only be
5826 enabled if running jitter sensitive (HPC/RT) workloads.
5828 security= [SECURITY] Choose a legacy "major" security module to
5829 enable at boot. This has been deprecated by the
5832 selinux= [SELINUX] Disable or enable SELinux at boot time.
5833 Format: { "0" | "1" }
5834 See security/selinux/Kconfig help text.
5839 serialnumber [BUGS=X86-32]
5841 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
5844 Maximal number of shapers.
5846 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
5847 Limit apic dumping. The parameter defines the maximal
5848 number of local apics being dumped. Also it is possible
5849 to set it to "all" by meaning -- no limit here.
5850 Format: { 1 (default) | 2 | ... | all }.
5851 The parameter valid if only apic=debug or
5852 apic=verbose is specified.
5853 Example: apic=debug show_lapic=all
5861 Enable merging of slabs with similar size when the
5862 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5865 Disable merging of slabs with similar size. May be
5866 necessary if there is some reason to distinguish
5867 allocs to different slabs, especially in hardened
5868 environments where the risk of heap overflows and
5869 layout control by attackers can usually be
5870 frustrated by disabling merging. This will reduce
5871 most of the exposure of a heap attack to a single
5872 cache (risks via metadata attacks are mostly
5873 unchanged). Debug options disable merging on their
5875 For more information see Documentation/mm/slub.rst.
5877 slab_max_order= [MM, SLAB]
5878 Determines the maximum allowed order for slabs.
5879 A high setting may cause OOMs due to memory
5880 fragmentation. Defaults to 1 for systems with
5881 more than 32MB of RAM, 0 otherwise.
5883 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5884 Enabling slub_debug allows one to determine the
5885 culprit if slab objects become corrupted. Enabling
5886 slub_debug can create guard zones around objects and
5887 may poison objects when not in use. Also tracks the
5888 last alloc / free. For more information see
5889 Documentation/mm/slub.rst.
5891 slub_max_order= [MM, SLUB]
5892 Determines the maximum allowed order for slabs.
5893 A high setting may cause OOMs due to memory
5894 fragmentation. For more information see
5895 Documentation/mm/slub.rst.
5897 slub_min_objects= [MM, SLUB]
5898 The minimum number of objects per slab. SLUB will
5899 increase the slab order up to slub_max_order to
5900 generate a sufficiently large slab able to contain
5901 the number of objects indicated. The higher the number
5902 of objects the smaller the overhead of tracking slabs
5903 and the less frequently locks need to be acquired.
5904 For more information see Documentation/mm/slub.rst.
5906 slub_min_order= [MM, SLUB]
5907 Determines the minimum page order for slabs. Must be
5908 lower than slub_max_order.
5909 For more information see Documentation/mm/slub.rst.
5911 slub_merge [MM, SLUB]
5912 Same with slab_merge.
5914 slub_nomerge [MM, SLUB]
5915 Same with slab_nomerge. This is supported for legacy.
5916 See slab_nomerge for more information.
5919 Format: <io1>[,<io2>[,...,<io8>]]
5921 smp.csd_lock_timeout= [KNL]
5922 Specify the period of time in milliseconds
5923 that smp_call_function() and friends will wait
5924 for a CPU to release the CSD lock. This is
5925 useful when diagnosing bugs involving CPUs
5926 disabling interrupts for extended periods
5927 of time. Defaults to 5,000 milliseconds, and
5928 setting a value of zero disables this feature.
5929 This feature may be more efficiently disabled
5930 using the csdlock_debug- kernel parameter.
5932 smp.panic_on_ipistall= [KNL]
5933 If a csd_lock_timeout extends for more than
5934 the specified number of milliseconds, panic the
5935 system. By default, let CSD-lock acquisition
5936 take as long as they take. Specifying 300,000
5937 for this value provides a 5-minute timeout.
5939 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5940 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5941 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5942 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5943 smsc-ircc2.ircc_irq= [HW] IRQ line
5944 smsc-ircc2.ircc_dma= [HW] DMA channel
5945 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5946 0: Toshiba Satellite 1800 (GP data pin select)
5947 1: Fast pin select (default)
5950 smt= [KNL,MIPS,S390] Set the maximum number of threads (logical
5951 CPUs) to use per physical CPU on systems capable of
5952 symmetric multithreading (SMT). Will be capped to the
5953 actual hardware limit.
5955 Default: -1 (no limit)
5958 [KNL] Should the soft-lockup detector generate panics.
5961 A value of 1 instructs the soft-lockup detector
5962 to panic the machine when a soft-lockup occurs. It is
5963 also controlled by the kernel.softlockup_panic sysctl
5964 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5965 respective build-time switch to that functionality.
5967 softlockup_all_cpu_backtrace=
5968 [KNL] Should the soft-lockup detector generate
5969 backtraces on all cpus.
5972 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5973 See Documentation/admin-guide/laptops/sonypi.rst
5975 spectre_v2= [X86] Control mitigation of Spectre variant 2
5976 (indirect branch speculation) vulnerability.
5977 The default operation protects the kernel from
5980 on - unconditionally enable, implies
5982 off - unconditionally disable, implies
5984 auto - kernel detects whether your CPU model is
5987 Selecting 'on' will, and 'auto' may, choose a
5988 mitigation method at run time according to the
5989 CPU, the available microcode, the setting of the
5990 CONFIG_RETPOLINE configuration option, and the
5991 compiler with which the kernel was built.
5993 Selecting 'on' will also enable the mitigation
5994 against user space to user space task attacks.
5996 Selecting 'off' will disable both the kernel and
5997 the user space protections.
5999 Specific mitigations can also be selected manually:
6001 retpoline - replace indirect branches
6002 retpoline,generic - Retpolines
6003 retpoline,lfence - LFENCE; indirect branch
6004 retpoline,amd - alias for retpoline,lfence
6005 eibrs - Enhanced/Auto IBRS
6006 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines
6007 eibrs,lfence - Enhanced/Auto IBRS + LFENCE
6008 ibrs - use IBRS to protect kernel
6010 Not specifying this option is equivalent to
6014 [X86] Control mitigation of Spectre variant 2
6015 (indirect branch speculation) vulnerability between
6018 on - Unconditionally enable mitigations. Is
6019 enforced by spectre_v2=on
6021 off - Unconditionally disable mitigations. Is
6022 enforced by spectre_v2=off
6024 prctl - Indirect branch speculation is enabled,
6025 but mitigation can be enabled via prctl
6026 per thread. The mitigation control state
6027 is inherited on fork.
6030 - Like "prctl" above, but only STIBP is
6031 controlled per thread. IBPB is issued
6032 always when switching between different user
6036 - Same as "prctl" above, but all seccomp
6037 threads will enable the mitigation unless
6038 they explicitly opt out.
6041 - Like "seccomp" above, but only STIBP is
6042 controlled per thread. IBPB is issued
6043 always when switching between different
6044 user space processes.
6046 auto - Kernel selects the mitigation depending on
6047 the available CPU features and vulnerability.
6049 Default mitigation: "prctl"
6051 Not specifying this option is equivalent to
6052 spectre_v2_user=auto.
6054 spec_rstack_overflow=
6055 [X86] Control RAS overflow mitigation on AMD Zen CPUs
6057 off - Disable mitigation
6058 microcode - Enable microcode mitigation only
6059 safe-ret - Enable sw-only safe RET mitigation (default)
6060 ibpb - Enable mitigation by issuing IBPB on
6062 ibpb-vmexit - Issue IBPB only on VMEXIT
6063 (cloud-specific mitigation)
6065 spec_store_bypass_disable=
6066 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
6067 (Speculative Store Bypass vulnerability)
6069 Certain CPUs are vulnerable to an exploit against a
6070 a common industry wide performance optimization known
6071 as "Speculative Store Bypass" in which recent stores
6072 to the same memory location may not be observed by
6073 later loads during speculative execution. The idea
6074 is that such stores are unlikely and that they can
6075 be detected prior to instruction retirement at the
6076 end of a particular speculation execution window.
6078 In vulnerable processors, the speculatively forwarded
6079 store can be used in a cache side channel attack, for
6080 example to read memory to which the attacker does not
6081 directly have access (e.g. inside sandboxed code).
6083 This parameter controls whether the Speculative Store
6084 Bypass optimization is used.
6086 On x86 the options are:
6088 on - Unconditionally disable Speculative Store Bypass
6089 off - Unconditionally enable Speculative Store Bypass
6090 auto - Kernel detects whether the CPU model contains an
6091 implementation of Speculative Store Bypass and
6092 picks the most appropriate mitigation. If the
6093 CPU is not vulnerable, "off" is selected. If the
6094 CPU is vulnerable the default mitigation is
6095 architecture and Kconfig dependent. See below.
6096 prctl - Control Speculative Store Bypass per thread
6097 via prctl. Speculative Store Bypass is enabled
6098 for a process by default. The state of the control
6099 is inherited on fork.
6100 seccomp - Same as "prctl" above, but all seccomp threads
6101 will disable SSB unless they explicitly opt out.
6103 Default mitigations:
6106 On powerpc the options are:
6108 on,auto - On Power8 and Power9 insert a store-forwarding
6109 barrier on kernel entry and exit. On Power7
6110 perform a software flush on kernel entry and
6114 Not specifying this option is equivalent to
6115 spec_store_bypass_disable=auto.
6117 spia_io_base= [HW,MTD]
6123 [X86] Enable split lock detection or bus lock detection
6125 When enabled (and if hardware support is present), atomic
6126 instructions that access data across cache line
6127 boundaries will result in an alignment check exception
6128 for split lock detection or a debug exception for
6133 warn - the kernel will emit rate-limited warnings
6134 about applications triggering the #AC
6135 exception or the #DB exception. This mode is
6136 the default on CPUs that support split lock
6137 detection or bus lock detection. Default
6138 behavior is by #AC if both features are
6139 enabled in hardware.
6141 fatal - the kernel will send SIGBUS to applications
6142 that trigger the #AC exception or the #DB
6143 exception. Default behavior is by #AC if
6144 both features are enabled in hardware.
6147 Set system wide rate limit to N bus locks
6148 per second for bus lock detection.
6151 N/A for split lock detection.
6154 If an #AC exception is hit in the kernel or in
6155 firmware (i.e. not while executing in user mode)
6156 the kernel will oops in either "warn" or "fatal"
6159 #DB exception for bus lock is triggered only when
6163 Control the Special Register Buffer Data Sampling
6166 Certain CPUs are vulnerable to an MDS-like
6167 exploit which can leak bits from the random
6170 By default, this issue is mitigated by
6171 microcode. However, the microcode fix can cause
6172 the RDRAND and RDSEED instructions to become
6173 much slower. Among other effects, this will
6174 result in reduced throughput from /dev/urandom.
6176 The microcode mitigation can be disabled with
6177 the following option:
6179 off: Disable mitigation and remove
6180 performance impact to RDRAND and RDSEED
6182 srcutree.big_cpu_lim [KNL]
6183 Specifies the number of CPUs constituting a
6184 large system, such that srcu_struct structures
6185 should immediately allocate an srcu_node array.
6186 This kernel-boot parameter defaults to 128,
6187 but takes effect only when the low-order four
6188 bits of srcutree.convert_to_big is equal to 3
6191 srcutree.convert_to_big [KNL]
6192 Specifies under what conditions an SRCU tree
6193 srcu_struct structure will be converted to big
6194 form, that is, with an rcu_node tree:
6197 1: At init_srcu_struct() time.
6198 2: When rcutorture decides to.
6199 3: Decide at boot time (default).
6200 0x1X: Above plus if high contention.
6202 Either way, the srcu_node tree will be sized based
6203 on the actual runtime number of CPUs (nr_cpu_ids)
6204 instead of the compile-time CONFIG_NR_CPUS.
6206 srcutree.counter_wrap_check [KNL]
6207 Specifies how frequently to check for
6208 grace-period sequence counter wrap for the
6209 srcu_data structure's ->srcu_gp_seq_needed field.
6210 The greater the number of bits set in this kernel
6211 parameter, the less frequently counter wrap will
6212 be checked for. Note that the bottom two bits
6215 srcutree.exp_holdoff [KNL]
6216 Specifies how many nanoseconds must elapse
6217 since the end of the last SRCU grace period for
6218 a given srcu_struct until the next normal SRCU
6219 grace period will be considered for automatic
6220 expediting. Set to zero to disable automatic
6223 srcutree.srcu_max_nodelay [KNL]
6224 Specifies the number of no-delay instances
6225 per jiffy for which the SRCU grace period
6226 worker thread will be rescheduled with zero
6227 delay. Beyond this limit, worker thread will
6228 be rescheduled with a sleep delay of one jiffy.
6230 srcutree.srcu_max_nodelay_phase [KNL]
6231 Specifies the per-grace-period phase, number of
6232 non-sleeping polls of readers. Beyond this limit,
6233 grace period worker thread will be rescheduled
6234 with a sleep delay of one jiffy, between each
6235 rescan of the readers, for a grace period phase.
6237 srcutree.srcu_retry_check_delay [KNL]
6238 Specifies number of microseconds of non-sleeping
6239 delay between each non-sleeping poll of readers.
6241 srcutree.small_contention_lim [KNL]
6242 Specifies the number of update-side contention
6243 events per jiffy will be tolerated before
6244 initiating a conversion of an srcu_struct
6245 structure to big form. Note that the value of
6246 srcutree.convert_to_big must have the 0x10 bit
6247 set for contention-based conversions to occur.
6250 Speculative Store Bypass Disable control
6252 On CPUs that are vulnerable to the Speculative
6253 Store Bypass vulnerability and offer a
6254 firmware based mitigation, this parameter
6255 indicates how the mitigation should be used:
6257 force-on: Unconditionally enable mitigation for
6258 for both kernel and userspace
6259 force-off: Unconditionally disable mitigation for
6260 for both kernel and userspace
6261 kernel: Always enable mitigation in the
6262 kernel, and offer a prctl interface
6263 to allow userspace to register its
6264 interest in being mitigated too.
6266 stack_guard_gap= [MM]
6267 override the default stack gap protection. The value
6268 is in page units and it defines how many pages prior
6269 to (for stacks growing down) resp. after (for stacks
6270 growing up) the main stack are reserved for no other
6271 mapping. Default value is 256 pages.
6273 stack_depot_disable= [KNL]
6274 Setting this to true through kernel command line will
6275 disable the stack depot thereby saving the static memory
6276 consumed by the stack hash table. By default this is set
6280 Enabled the stack tracer on boot up.
6282 stacktrace_filter=[function-list]
6283 [FTRACE] Limit the functions that the stack tracer
6284 will trace at boot up. function-list is a comma-separated
6285 list of functions. This list can be changed at run
6286 time by the stack_trace_filter file in the debugfs
6287 tracing directory. Note, this enables stack tracing
6288 and the stacktrace above is not needed.
6292 Set the STI (builtin display/keyboard on the HP-PARISC
6293 machines) console (graphic card) which should be used
6294 as the initial boot-console.
6295 See also comment in drivers/video/console/sticore.c.
6298 See comment in drivers/video/console/sticore.c.
6301 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
6306 Enable or disable strict sigaltstack size checks
6307 against the required signal frame size which
6308 depends on the supported FPU features. This can
6309 be used to filter out binaries which have
6310 not yet been made aware of AT_MINSIGSTKSZ.
6313 Limits the number of kernel HPT entries in the hash
6314 page table to increase the rate of hash page table
6315 faults on kernel addresses.
6318 Limits the number of kernel SLB entries, and flushes
6319 them frequently to increase the rate of SLB faults
6320 on kernel addresses.
6322 sunrpc.min_resvport=
6323 sunrpc.max_resvport=
6325 SunRPC servers often require that client requests
6326 originate from a privileged port (i.e. a port in the
6327 range 0 < portnr < 1024).
6328 An administrator who wishes to reserve some of these
6329 ports for other uses may adjust the range that the
6330 kernel's sunrpc client considers to be privileged
6331 using these two parameters to set the minimum and
6332 maximum port values.
6334 sunrpc.svc_rpc_per_connection_limit=
6336 Limit the number of requests that the server will
6337 process in parallel from a single connection.
6338 The default value is 0 (no limit).
6342 Control how the NFS server code allocates CPUs to
6343 service thread pools. Depending on how many NICs
6344 you have and where their interrupts are bound, this
6345 option will affect which CPUs will do NFS serving.
6346 Note: this parameter cannot be changed while the
6347 NFS server is running.
6349 auto the server chooses an appropriate mode
6350 automatically using heuristics
6351 global a single global pool contains all CPUs
6352 percpu one pool for each CPU
6353 pernode one pool for each NUMA node (equivalent
6354 to global on non-NUMA machines)
6356 sunrpc.tcp_slot_table_entries=
6357 sunrpc.udp_slot_table_entries=
6359 Sets the upper limit on the number of simultaneous
6360 RPC calls that can be sent from the client to a
6361 server. Increasing these values may allow you to
6362 improve throughput, but will also increase the
6363 amount of memory reserved for use by the client.
6365 suspend.pm_test_delay=
6367 Sets the number of seconds to remain in a suspend test
6368 mode before resuming the system (see
6369 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6370 is set. Default value is 5.
6373 Format: { on | off | y | n | 1 | 0 }
6374 This parameter controls use of the Protected
6375 Execution Facility on pSeries.
6377 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
6378 Format: { <int> [,<int>] | force | noforce }
6379 <int> -- Number of I/O TLB slabs
6380 <int> -- Second integer after comma. Number of swiotlb
6381 areas with their own lock. Will be rounded up
6383 force -- force using of bounce buffers even if they
6384 wouldn't be automatically used by the kernel
6385 noforce -- Never use bounce buffers (for debugging)
6390 Set a sysctl parameter, right before loading the init
6391 process, as if the value was written to the respective
6392 /proc/sys/... file. Both '.' and '/' are recognized as
6393 separators. Unrecognized parameters and invalid values
6394 are reported in the kernel log. Sysctls registered
6395 later by a loaded module cannot be set this way.
6396 Example: sysctl.vm.swappiness=40
6398 sysrq_always_enabled
6400 Ignore sysrq setting - this boot parameter will
6401 neutralize any effect of /proc/sys/kernel/sysrq.
6402 Useful for debugging.
6404 tcpmhash_entries= [KNL,NET]
6405 Set the number of tcp_metrics_hash slots.
6406 Default value is 8192 or 16384 depending on total
6407 ram pages. This is used to specify the TCP metrics
6408 cache size. See Documentation/networking/ip-sysctl.rst
6409 "tcp_no_metrics_save" section for more details.
6413 test_suspend= [SUSPEND]
6414 Format: { "mem" | "standby" | "freeze" }[,N]
6415 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6416 standby suspend) or "freeze" (for suspend type freeze)
6417 as the system sleep state during system startup with
6418 the optional capability to repeat N number of times.
6419 The system is woken from this state using a
6420 wakeup-capable RTC alarm.
6422 thash_entries= [KNL,NET]
6423 Set number of hash buckets for TCP connection
6425 thermal.act= [HW,ACPI]
6426 -1: disable all active trip points in all thermal zones
6427 <degrees C>: override all lowest active trip points
6429 thermal.crt= [HW,ACPI]
6430 -1: disable all critical trip points in all thermal zones
6431 <degrees C>: override all critical trip points
6433 thermal.off= [HW,ACPI]
6434 1: disable ACPI thermal control
6436 thermal.psv= [HW,ACPI]
6437 -1: disable all passive trip points
6438 <degrees C>: override all passive trip points to this
6441 thermal.tzp= [HW,ACPI]
6442 Specify global default ACPI thermal zone polling rate
6443 <deci-seconds>: poll all this frequency
6444 0: no polling (default)
6447 Force threading of all interrupt handlers except those
6448 marked explicitly IRQF_NO_THREAD.
6452 Specify if the kernel should make use of the cpu
6453 topology information if the hardware supports this.
6454 The scheduler will make use of this information and
6455 e.g. base its process migration decisions on it.
6458 topology_updates= [KNL, PPC, NUMA]
6460 Specify if the kernel should ignore (off)
6461 topology updates sent by the hypervisor to this
6464 torture.disable_onoff_at_boot= [KNL]
6465 Prevent the CPU-hotplug component of torturing
6466 until after init has spawned.
6468 torture.ftrace_dump_at_shutdown= [KNL]
6469 Dump the ftrace buffer at torture-test shutdown,
6470 even if there were no errors. This can be a
6471 very costly operation when many torture tests
6472 are running concurrently, especially on systems
6473 with rotating-rust storage.
6475 torture.verbose_sleep_frequency= [KNL]
6476 Specifies how many verbose printk()s should be
6477 emitted between each sleep. The default of zero
6478 disables verbose-printk() sleeping.
6480 torture.verbose_sleep_duration= [KNL]
6481 Duration of each verbose-printk() sleep in jiffies.
6485 tpm_suspend_pcr=[HW,TPM]
6486 Format: integer pcr id
6487 Specify that at suspend time, the tpm driver
6488 should extend the specified pcr with zeros,
6489 as a workaround for some chips which fail to
6490 flush the last written pcr on TPM_SaveState.
6491 This will guarantee that all the other pcrs
6494 tpm_tis.interrupts= [HW,TPM]
6495 Enable interrupts for the MMIO based physical layer
6496 for the FIFO interface. By default it is set to false
6497 (0). For more information about TPM hardware interfaces
6498 defined by Trusted Computing Group (TCG) see
6499 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/
6502 Have the tracepoints sent to printk as well as the
6503 tracing ring buffer. This is useful for early boot up
6504 where the system hangs or reboots and does not give the
6505 option for reading the tracing buffer or performing a
6506 ftrace_dump_on_oops.
6508 To turn off having tracepoints sent to printk,
6509 echo 0 > /proc/sys/kernel/tracepoint_printk
6510 Note, echoing 1 into this file without the
6511 tracepoint_printk kernel cmdline option has no effect.
6513 The tp_printk_stop_on_boot (see below) can also be used
6514 to stop the printing of events to console at
6519 Having tracepoints sent to printk() and activating high
6520 frequency tracepoints such as irq or sched, can cause
6521 the system to live lock.
6523 tp_printk_stop_on_boot [FTRACE]
6524 When tp_printk (above) is set, it can cause a lot of noise
6525 on the console. It may be useful to only include the
6526 printing of events during boot up, as user space may
6527 make the system inoperable.
6529 This command line option will stop the printing of events
6530 to console at the late_initcall_sync() time frame.
6532 trace_buf_size=nn[KMG]
6533 [FTRACE] will set tracing buffer size on each cpu.
6535 trace_clock= [FTRACE] Set the clock used for tracing events
6537 local - Use the per CPU time stamp counter
6538 (converted into nanoseconds). Fast, but
6539 depending on the architecture, may not be
6540 in sync between CPUs.
6541 global - Event time stamps are synchronize across
6542 CPUs. May be slower than the local clock,
6543 but better for some race conditions.
6544 counter - Simple counting of events (1, 2, ..)
6545 note, some counts may be skipped due to the
6546 infrastructure grabbing the clock more than
6548 uptime - Use jiffies as the time stamp.
6549 perf - Use the same clock that perf uses.
6550 mono - Use ktime_get_mono_fast_ns() for time stamps.
6551 mono_raw - Use ktime_get_raw_fast_ns() for time
6553 boot - Use ktime_get_boot_fast_ns() for time stamps.
6554 Architectures may add more clocks. See
6555 Documentation/trace/ftrace.rst for more details.
6557 trace_event=[event-list]
6558 [FTRACE] Set and start specified trace events in order
6559 to facilitate early boot debugging. The event-list is a
6560 comma-separated list of trace events to enable. See
6561 also Documentation/trace/events.rst
6563 trace_instance=[instance-info]
6564 [FTRACE] Create a ring buffer instance early in boot up.
6565 This will be listed in:
6567 /sys/kernel/tracing/instances
6569 Events can be enabled at the time the instance is created
6572 trace_instance=<name>,<system1>:<event1>,<system2>:<event2>
6574 Note, the "<system*>:" portion is optional if the event is
6577 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall
6579 will enable the "sched_switch" event (note, the "sched:" is optional, and
6580 the same thing would happen if it was left off). The irq_handler_entry
6581 event, and all events under the "initcall" system.
6583 trace_options=[option-list]
6584 [FTRACE] Enable or disable tracer options at boot.
6585 The option-list is a comma delimited list of options
6586 that can be enabled or disabled just as if you were
6587 to echo the option name into
6589 /sys/kernel/tracing/trace_options
6591 For example, to enable stacktrace option (to dump the
6592 stack trace of each event), add to the command line:
6594 trace_options=stacktrace
6596 See also Documentation/trace/ftrace.rst "trace options"
6599 trace_trigger=[trigger-list]
6600 [FTRACE] Add a event trigger on specific events.
6601 Set a trigger on top of a specific event, with an optional
6604 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..."
6605 Where more than one trigger may be specified that are comma deliminated.
6609 trace_trigger="sched_switch.stacktrace if prev_state == 2"
6611 The above will enable the "stacktrace" trigger on the "sched_switch"
6612 event but only trigger it if the "prev_state" of the "sched_switch"
6613 event is "2" (TASK_UNINTERUPTIBLE).
6615 See also "Event triggers" in Documentation/trace/events.rst
6619 [FTRACE] enable this option to disable tracing when a
6620 warning is hit. This turns off "tracing_on". Tracing can
6621 be enabled again by echoing '1' into the "tracing_on"
6622 file located in /sys/kernel/tracing/
6624 This option is useful, as it disables the trace before
6625 the WARNING dump is called, which prevents the trace to
6626 be filled with content caused by the warning output.
6628 This option can also be set at run time via the sysctl
6629 option: kernel/traceoff_on_warning
6631 transparent_hugepage=
6633 Format: [always|madvise|never]
6634 Can be used to control the default behavior of the system
6635 with respect to transparent hugepages.
6636 See Documentation/admin-guide/mm/transhuge.rst
6639 trusted.source= [KEYS]
6641 This parameter identifies the trust source as a backend
6642 for trusted keys implementation. Supported trust
6647 If not specified then it defaults to iterating through
6648 the trust source list starting with TPM and assigns the
6649 first trust source as a backend which is initialized
6650 successfully during iteration.
6654 The RNG used to generate key material for trusted keys.
6657 - the same value as trusted.source: "tpm" or "tee"
6659 If not specified, "default" is used. In this case,
6660 the RNG's choice is left to each individual trust source.
6662 tsc= Disable clocksource stability checks for TSC.
6664 [x86] reliable: mark tsc clocksource as reliable, this
6665 disables clocksource verification at runtime, as well
6666 as the stability checks done at bootup. Used to enable
6667 high-resolution timer mode on older hardware, and in
6668 virtualized environment.
6669 [x86] noirqtime: Do not use TSC to do irq accounting.
6670 Used to run time disable IRQ_TIME_ACCOUNTING on any
6671 platforms where RDTSC is slow and this accounting
6673 [x86] unstable: mark the TSC clocksource as unstable, this
6674 marks the TSC unconditionally unstable at bootup and
6675 avoids any further wobbles once the TSC watchdog notices.
6676 [x86] nowatchdog: disable clocksource watchdog. Used
6677 in situations with strict latency requirements (where
6678 interruptions from clocksource watchdog are not
6680 [x86] recalibrate: force recalibration against a HW timer
6681 (HPET or PM timer) on systems whose TSC frequency was
6682 obtained from HW or FW using either an MSR or CPUID(0x15).
6683 Warn if the difference is more than 500 ppm.
6684 [x86] watchdog: Use TSC as the watchdog clocksource with
6685 which to check other HW timers (HPET or PM timer), but
6686 only on systems where TSC has been deemed trustworthy.
6687 This will be suppressed by an earlier tsc=nowatchdog and
6688 can be overridden by a later tsc=nowatchdog. A console
6689 message will flag any such suppression or overriding.
6691 tsc_early_khz= [X86] Skip early TSC calibration and use the given
6692 value instead. Useful when the early TSC frequency discovery
6693 procedure is not reliable, such as on overclocked systems
6694 with CPUID.16h support and partial CPUID.15h support.
6695 Format: <unsigned int>
6697 tsx= [X86] Control Transactional Synchronization
6698 Extensions (TSX) feature in Intel processors that
6699 support TSX control.
6701 This parameter controls the TSX feature. The options are:
6703 on - Enable TSX on the system. Although there are
6704 mitigations for all known security vulnerabilities,
6705 TSX has been known to be an accelerator for
6706 several previous speculation-related CVEs, and
6707 so there may be unknown security risks associated
6708 with leaving it enabled.
6710 off - Disable TSX on the system. (Note that this
6711 option takes effect only on newer CPUs which are
6712 not vulnerable to MDS, i.e., have
6713 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6714 the new IA32_TSX_CTRL MSR through a microcode
6715 update. This new MSR allows for the reliable
6716 deactivation of the TSX functionality.)
6718 auto - Disable TSX if X86_BUG_TAA is present,
6719 otherwise enable TSX on the system.
6721 Not specifying this option is equivalent to tsx=off.
6723 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6726 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6727 Abort (TAA) vulnerability.
6729 Similar to Micro-architectural Data Sampling (MDS)
6730 certain CPUs that support Transactional
6731 Synchronization Extensions (TSX) are vulnerable to an
6732 exploit against CPU internal buffers which can forward
6733 information to a disclosure gadget under certain
6736 In vulnerable processors, the speculatively forwarded
6737 data can be used in a cache side channel attack, to
6738 access data to which the attacker does not have direct
6741 This parameter controls the TAA mitigation. The
6744 full - Enable TAA mitigation on vulnerable CPUs
6747 full,nosmt - Enable TAA mitigation and disable SMT on
6748 vulnerable CPUs. If TSX is disabled, SMT
6749 is not disabled because CPU is not
6750 vulnerable to cross-thread TAA attacks.
6751 off - Unconditionally disable TAA mitigation
6753 On MDS-affected machines, tsx_async_abort=off can be
6754 prevented by an active MDS mitigation as both vulnerabilities
6755 are mitigated with the same mechanism so in order to disable
6756 this mitigation, you need to specify mds=off too.
6758 Not specifying this option is equivalent to
6759 tsx_async_abort=full. On CPUs which are MDS affected
6760 and deploy MDS mitigation, TAA mitigation is not
6761 required and doesn't provide any additional
6765 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6767 turbografx.map[2|3]= [HW,JOY]
6768 TurboGraFX parallel port interface
6770 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6771 See also Documentation/input/devices/joystick-parport.rst
6773 udbg-immortal [PPC] When debugging early kernel crashes that
6774 happen after console_init() and before a proper
6775 console driver takes over, this boot options might
6776 help "seeing" what's going on.
6778 uhash_entries= [KNL,NET]
6779 Set number of hash buckets for UDP/UDP-Lite connections
6782 [USB] Ignore overcurrent events (default N).
6783 Some badly-designed motherboards generate lots of
6784 bogus events, for ports that aren't wired to
6785 anything. Set this parameter to avoid log spamming.
6786 Note that genuine overcurrent events won't be
6790 [X86] Cause panic on unknown NMI.
6792 unwind_debug [X86-64]
6793 Enable unwinder debug output. This can be
6794 useful for debugging certain unwinder error
6795 conditions, including corrupt stacks and
6796 bad/missing unwinder metadata.
6798 usbcore.authorized_default=
6799 [USB] Default USB device authorization:
6800 (default -1 = authorized (same as 1),
6801 0 = not authorized, 1 = authorized, 2 = authorized
6802 if device connected to internal port)
6804 usbcore.autosuspend=
6805 [USB] The autosuspend time delay (in seconds) used
6806 for newly-detected USB devices (default 2). This
6807 is the time required before an idle device will be
6808 autosuspended. Devices for which the delay is set
6809 to a negative value won't be autosuspended at all.
6811 usbcore.usbfs_snoop=
6812 [USB] Set to log all usbfs traffic (default 0 = off).
6814 usbcore.usbfs_snoop_max=
6815 [USB] Maximum number of bytes to snoop in each URB
6818 usbcore.blinkenlights=
6819 [USB] Set to cycle leds on hubs (default 0 = off).
6821 usbcore.old_scheme_first=
6822 [USB] Start with the old device initialization
6823 scheme (default 0 = off).
6825 usbcore.usbfs_memory_mb=
6826 [USB] Memory limit (in MB) for buffers allocated by
6827 usbfs (default = 16, 0 = max = 2047).
6829 usbcore.use_both_schemes=
6830 [USB] Try the other device initialization scheme
6831 if the first one fails (default 1 = enabled).
6833 usbcore.initial_descriptor_timeout=
6834 [USB] Specifies timeout for the initial 64-byte
6835 USB_REQ_GET_DESCRIPTOR request in milliseconds
6836 (default 5000 = 5.0 seconds).
6838 usbcore.nousb [USB] Disable the USB subsystem
6841 [USB] A list of quirk entries to augment the built-in
6842 usb core quirk list. List entries are separated by
6843 commas. Each entry has the form
6844 VendorID:ProductID:Flags. The IDs are 4-digit hex
6845 numbers and Flags is a set of letters. Each letter
6846 will change the built-in quirk; setting it if it is
6847 clear and clearing it if it is set. The letters have
6848 the following meanings:
6849 a = USB_QUIRK_STRING_FETCH_255 (string
6850 descriptors must not be fetched using
6852 b = USB_QUIRK_RESET_RESUME (device can't resume
6853 correctly so reset it instead);
6854 c = USB_QUIRK_NO_SET_INTF (device can't handle
6855 Set-Interface requests);
6856 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6857 handle its Configuration or Interface
6859 e = USB_QUIRK_RESET (device can't be reset
6860 (e.g morph devices), don't use reset);
6861 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6862 more interface descriptions than the
6863 bNumInterfaces count, and can't handle
6864 talking to these interfaces);
6865 g = USB_QUIRK_DELAY_INIT (device needs a pause
6866 during initialization, after we read
6867 the device descriptor);
6868 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6869 high speed and super speed interrupt
6870 endpoints, the USB 2.0 and USB 3.0 spec
6871 require the interval in microframes (1
6872 microframe = 125 microseconds) to be
6873 calculated as interval = 2 ^
6875 Devices with this quirk report their
6876 bInterval as the result of this
6877 calculation instead of the exponent
6878 variable used in the calculation);
6879 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6880 handle device_qualifier descriptor
6882 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6883 generates spurious wakeup, ignore
6884 remote wakeup capability);
6885 k = USB_QUIRK_NO_LPM (device can't handle Link
6887 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6888 (Device reports its bInterval as linear
6889 frames instead of the USB 2.0
6891 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6892 to be disconnected before suspend to
6893 prevent spurious wakeup);
6894 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6895 pause after every control message);
6896 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6897 delay after resetting its port);
6898 Example: quirks=0781:5580:bk,0a5c:5834:gij
6901 [USBHID] The interval which mice are to be polled at.
6904 [USBHID] The interval which joysticks are to be polled at.
6907 [USBHID] The interval which keyboards are to be polled at.
6909 usb-storage.delay_use=
6910 [UMS] The delay in seconds before a new device is
6911 scanned for Logical Units (default 1).
6914 [UMS] A list of quirks entries to supplement or
6915 override the built-in unusual_devs list. List
6916 entries are separated by commas. Each entry has
6917 the form VID:PID:Flags where VID and PID are Vendor
6918 and Product ID values (4-digit hex numbers) and
6919 Flags is a set of characters, each corresponding
6920 to a common usb-storage quirk flag as follows:
6921 a = SANE_SENSE (collect more than 18 bytes
6922 of sense data, not on uas);
6923 b = BAD_SENSE (don't collect more than 18
6924 bytes of sense data, not on uas);
6925 c = FIX_CAPACITY (decrease the reported
6926 device capacity by one sector);
6927 d = NO_READ_DISC_INFO (don't use
6928 READ_DISC_INFO command, not on uas);
6929 e = NO_READ_CAPACITY_16 (don't use
6930 READ_CAPACITY_16 command);
6931 f = NO_REPORT_OPCODES (don't use report opcodes
6933 g = MAX_SECTORS_240 (don't transfer more than
6934 240 sectors at a time, uas only);
6935 h = CAPACITY_HEURISTICS (decrease the
6936 reported device capacity by one
6937 sector if the number is odd);
6938 i = IGNORE_DEVICE (don't bind to this
6940 j = NO_REPORT_LUNS (don't use report luns
6942 k = NO_SAME (do not use WRITE_SAME, uas only)
6943 l = NOT_LOCKABLE (don't try to lock and
6944 unlock ejectable media, not on uas);
6945 m = MAX_SECTORS_64 (don't transfer more
6946 than 64 sectors = 32 KB at a time,
6948 n = INITIAL_READ10 (force a retry of the
6949 initial READ(10) command, not on uas);
6950 o = CAPACITY_OK (accept the capacity
6951 reported by the device, not on uas);
6952 p = WRITE_CACHE (the device cache is ON
6953 by default, not on uas);
6954 r = IGNORE_RESIDUE (the device reports
6955 bogus residue values, not on uas);
6956 s = SINGLE_LUN (the device has only one
6958 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6959 commands, uas only);
6960 u = IGNORE_UAS (don't bind to the uas driver);
6961 w = NO_WP_DETECT (don't test whether the
6962 medium is write-protected).
6963 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6964 even if the device claims no cache,
6966 Example: quirks=0419:aaf5:rl,0421:0433:rc
6968 user_debug= [KNL,ARM]
6970 See arch/arm/Kconfig.debug help text.
6971 1 - undefined instruction events
6973 4 - invalid data aborts
6976 Example: user_debug=31
6979 [X86] Flags controlling user PTE allocations.
6981 nohigh = do not allocate PTE pages in
6982 HIGHMEM regardless of setting
6985 vdso= [X86,SH,SPARC]
6986 On X86_32, this is an alias for vdso32=. Otherwise:
6988 vdso=1: enable VDSO (the default)
6989 vdso=0: disable VDSO mapping
6991 vdso32= [X86] Control the 32-bit vDSO
6992 vdso32=1: enable 32-bit VDSO
6993 vdso32=0 or vdso32=2: disable 32-bit VDSO
6995 See the help text for CONFIG_COMPAT_VDSO for more
6996 details. If CONFIG_COMPAT_VDSO is set, the default is
6997 vdso32=0; otherwise, the default is vdso32=1.
6999 For compatibility with older kernels, vdso32=2 is an
7002 Try vdso32=0 if you encounter an error that says:
7003 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
7006 vector=percpu: enable percpu vector domain
7008 video= [FB] Frame buffer configuration
7009 See Documentation/fb/modedb.rst.
7011 video.brightness_switch_enabled= [ACPI]
7013 If set to 1, on receiving an ACPI notify event
7014 generated by hotkey, video driver will adjust brightness
7015 level and then send out the event to user space through
7016 the allocated input device. If set to 0, video driver
7017 will only send out the event without touching backlight
7022 [VMMIO] Memory mapped virtio (platform) device.
7024 <size>@<baseaddr>:<irq>[:<id>]
7026 <size> := size (can use standard suffixes
7028 <baseaddr> := physical base address
7029 <irq> := interrupt number (as passed to
7031 <id> := (optional) platform device id
7033 virtio_mmio.device=1K@0x100b0000:48:7
7035 Can be used multiple times for multiple devices.
7037 vga= [BOOT,X86-32] Select a particular video mode
7038 See Documentation/arch/x86/boot.rst and
7039 Documentation/admin-guide/svga.rst.
7040 Use vga=ask for menu.
7041 This is actually a boot loader parameter; the value is
7042 passed to the kernel using a special protocol.
7044 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
7045 May slow down system boot speed, especially when
7046 enabled on systems with a large amount of memory.
7047 All options are enabled by default, and this
7048 interface is meant to allow for selectively
7049 enabling or disabling specific virtual memory
7052 Available options are:
7053 P Enable page structure init time poisoning
7054 - Disable all of the above options
7056 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
7057 size of <nn>. This can be used to increase the
7058 minimum size (128MB on x86). It can also be used to
7059 decrease the size and leave more room for directly
7062 vmcp_cma=nn[MG] [KNL,S390]
7063 Sets the memory size reserved for contiguous memory
7064 allocations for the vmcp device driver.
7066 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
7069 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
7072 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
7076 Controls the behavior of vsyscalls (i.e. calls to
7077 fixed addresses of 0xffffffffff600x00 from legacy
7078 code). Most statically-linked binaries and older
7079 versions of glibc use these calls. Because these
7080 functions are at fixed addresses, they make nice
7081 targets for exploits that can control RIP.
7083 emulate Vsyscalls turn into traps and are emulated
7084 reasonably safely. The vsyscall page is
7087 xonly [default] Vsyscalls turn into traps and are
7088 emulated reasonably safely. The vsyscall
7089 page is not readable.
7091 none Vsyscalls don't work at all. This makes
7092 them quite hard to use for exploits but
7093 might break your system.
7095 vt.color= [VT] Default text color.
7096 Format: 0xYX, X = foreground, Y = background.
7097 Default: 0x07 = light gray on black.
7099 vt.cur_default= [VT] Default cursor shape.
7100 Format: 0xCCBBAA, where AA, BB, and CC are the same as
7101 the parameters of the <Esc>[?A;B;Cc escape sequence;
7102 see VGA-softcursor.txt. Default: 2 = underline.
7104 vt.default_blu= [VT]
7105 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
7106 Change the default blue palette of the console.
7107 This is a 16-member array composed of values
7110 vt.default_grn= [VT]
7111 Format: <green0>,<green1>,<green2>,...,<green15>
7112 Change the default green palette of the console.
7113 This is a 16-member array composed of values
7116 vt.default_red= [VT]
7117 Format: <red0>,<red1>,<red2>,...,<red15>
7118 Change the default red palette of the console.
7119 This is a 16-member array composed of values
7125 Set system-wide default UTF-8 mode for all tty's.
7126 Default is 1, i.e. UTF-8 mode is enabled for all
7127 newly opened terminals.
7129 vt.global_cursor_default=
7132 Set system-wide default for whether a cursor
7133 is shown on new VTs. Default is -1,
7134 i.e. cursors will be created by default unless
7135 overridden by individual drivers. 0 will hide
7136 cursors, 1 will display them.
7138 vt.italic= [VT] Default color for italic text; 0-15.
7141 vt.underline= [VT] Default color for underlined text; 0-15.
7144 watchdog timers [HW,WDT] For information on watchdog timers,
7145 see Documentation/watchdog/watchdog-parameters.rst
7146 or other driver-specific files in the
7147 Documentation/watchdog/ directory.
7151 Set the hard lockup detector stall duration
7152 threshold in seconds. The soft lockup detector
7153 threshold is set to twice the value. A value of 0
7154 disables both lockup detectors. Default is 10
7157 workqueue.unbound_cpus=
7158 [KNL,SMP] Specify to constrain one or some CPUs
7159 to use in unbound workqueues.
7161 By default, all online CPUs are available for
7164 workqueue.watchdog_thresh=
7165 If CONFIG_WQ_WATCHDOG is configured, workqueue can
7166 warn stall conditions and dump internal state to
7167 help debugging. 0 disables workqueue stall
7168 detection; otherwise, it's the stall threshold
7169 duration in seconds. The default value is 30 and
7170 it can be updated at runtime by writing to the
7171 corresponding sysfs file.
7173 workqueue.cpu_intensive_thresh_us=
7174 Per-cpu work items which run for longer than this
7175 threshold are automatically considered CPU intensive
7176 and excluded from concurrency management to prevent
7177 them from noticeably delaying other per-cpu work
7178 items. Default is 10000 (10ms).
7180 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel
7181 will report the work functions which violate this
7182 threshold repeatedly. They are likely good
7183 candidates for using WQ_UNBOUND workqueues instead.
7185 workqueue.power_efficient
7186 Per-cpu workqueues are generally preferred because
7187 they show better performance thanks to cache
7188 locality; unfortunately, per-cpu workqueues tend to
7189 be more power hungry than unbound workqueues.
7191 Enabling this makes the per-cpu workqueues which
7192 were observed to contribute significantly to power
7193 consumption unbound, leading to measurably lower
7194 power usage at the cost of small performance
7197 The default value of this parameter is determined by
7198 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
7200 workqueue.default_affinity_scope=
7201 Select the default affinity scope to use for unbound
7202 workqueues. Can be one of "cpu", "smt", "cache",
7203 "numa" and "system". Default is "cache". For more
7204 information, see the Affinity Scopes section in
7205 Documentation/core-api/workqueue.rst.
7207 This can be changed after boot by writing to the
7208 matching /sys/module/workqueue/parameters file. All
7209 workqueues with the "default" affinity scope will be
7210 updated accordignly.
7212 workqueue.debug_force_rr_cpu
7213 Workqueue used to implicitly guarantee that work
7214 items queued without explicit CPU specified are put
7215 on the local CPU. This guarantee is no longer true
7216 and while local CPU is still preferred work items
7217 may be put on foreign CPUs. This debug option
7218 forces round-robin CPU selection to flush out
7219 usages which depend on the now broken guarantee.
7220 When enabled, memory and cache locality will be
7223 writecombine= [LOONGARCH] Control the MAT (Memory Access Type) of
7226 on - Enable writecombine, use WUC for ioremap_wc()
7227 off - Disable writecombine, use SUC for ioremap_wc()
7229 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
7230 default x2apic cluster mode on platforms
7233 xen_512gb_limit [KNL,X86-64,XEN]
7234 Restricts the kernel running paravirtualized under Xen
7235 to use only up to 512 GB of RAM. The reason to do so is
7236 crash analysis tools and Xen tools for doing domain
7237 save/restore/migration must be enabled to handle larger
7240 xen_emul_unplug= [HW,X86,XEN]
7241 Unplug Xen emulated devices
7242 Format: [unplug0,][unplug1]
7243 ide-disks -- unplug primary master IDE devices
7244 aux-ide-disks -- unplug non-primary-master IDE devices
7245 nics -- unplug network devices
7246 all -- unplug all emulated devices (NICs and IDE disks)
7247 unnecessary -- unplugging emulated devices is
7248 unnecessary even if the host did not respond to
7250 never -- do not unplug even if version check succeeds
7252 xen_legacy_crash [X86,XEN]
7253 Crash from Xen panic notifier, without executing late
7254 panic() code such as dumping handler.
7256 xen_msr_safe= [X86,XEN]
7258 Select whether to always use non-faulting (safe) MSR
7259 access functions when running as Xen PV guest. The
7260 default value is controlled by CONFIG_XEN_PV_MSR_SAFE.
7262 xen_nopvspin [X86,XEN]
7263 Disables the qspinlock slowpath using Xen PV optimizations.
7264 This parameter is obsoleted by "nopvspin" parameter, which
7265 has equivalent effect for XEN platform.
7268 Disables the PV optimizations forcing the HVM guest to
7269 run as generic HVM guest with no PV drivers.
7270 This option is obsoleted by the "nopv" option, which
7271 has equivalent effect for XEN platform.
7273 xen_no_vector_callback
7274 [KNL,X86,XEN] Disable the vector callback for Xen
7275 event channel interrupts.
7277 xen_scrub_pages= [XEN]
7278 Boolean option to control scrubbing pages before giving them back
7279 to Xen, for use by other domains. Can be also changed at runtime
7280 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
7281 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
7283 xen_timer_slop= [X86-64,XEN]
7284 Set the timer slop (in nanoseconds) for the virtual Xen
7285 timers (default is 100000). This adjusts the minimum
7286 delta of virtualized Xen timers, where lower values
7287 improve timer resolution at the expense of processing
7288 more timer interrupts.
7290 xen.balloon_boot_timeout= [XEN]
7291 The time (in seconds) to wait before giving up to boot
7292 in case initial ballooning fails to free enough memory.
7293 Applies only when running as HVM or PVH guest and
7294 started with less memory configured than allowed at
7295 max. Default is 180.
7297 xen.event_eoi_delay= [XEN]
7298 How long to delay EOI handling in case of event
7299 storms (jiffies). Default is 10.
7301 xen.event_loop_timeout= [XEN]
7302 After which time (jiffies) the event handling loop
7303 should start to delay EOI handling. Default is 2.
7305 xen.fifo_events= [XEN]
7306 Boolean parameter to disable using fifo event handling
7307 even if available. Normally fifo event handling is
7308 preferred over the 2-level event handling, as it is
7309 fairer and the number of possible event channels is
7310 much higher. Default is on (use fifo events).
7312 xirc2ps_cs= [NET,PCMCIA]
7314 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
7317 By default on POWER9 and above, the kernel will
7318 natively use the XIVE interrupt controller. This option
7319 allows the fallback firmware mode to be used:
7321 off Fallback to firmware control of XIVE interrupt
7322 controller on both pseries and powernv
7323 platforms. Only useful on POWER9 and above.
7325 xive.store-eoi=off [PPC]
7326 By default on POWER10 and above, the kernel will use
7327 stores for EOI handling when the XIVE interrupt mode
7328 is active. This option allows the XIVE driver to use
7329 loads instead, as on POWER9.
7331 xhci-hcd.quirks [USB,KNL]
7332 A hex value specifying bitmask with supplemental xhci
7333 host controller quirks. Meaning of each bit can be
7334 consulted in header drivers/usb/host/xhci.h.
7337 Format: { early | on | rw | ro | off }
7338 Controls if xmon debugger is enabled. Default is off.
7339 Passing only "xmon" is equivalent to "xmon=early".
7340 early Call xmon as early as possible on boot; xmon
7341 debugger is called from setup_arch().
7342 on xmon debugger hooks will be installed so xmon
7343 is only called on a kernel crash. Default mode,
7344 i.e. either "ro" or "rw" mode, is controlled
7345 with CONFIG_XMON_DEFAULT_RO_MODE.
7346 rw xmon debugger hooks will be installed so xmon
7347 is called only on a kernel crash, mode is write,
7348 meaning SPR registers, memory and, other data
7349 can be written using xmon commands.
7350 ro same as "rw" option above but SPR registers,
7351 memory, and other data can't be written using
7353 off xmon is disabled.