2 Format: { eager | lazy }
4 By default, unaccepted memory is accepted lazily to
5 avoid prolonged boot times. The lazy option will add
6 some runtime overhead until all memory is eventually
7 accepted. In most cases the overhead is negligible.
8 For some workloads or for debugging purposes
9 accept_memory=eager can be used to accept all memory
12 acpi= [HW,ACPI,X86,ARM64,RISCV64,EARLY]
13 Advanced Configuration and Power Interface
14 Format: { force | on | off | strict | noirq | rsdt |
16 force -- enable ACPI if default was off
17 on -- enable ACPI but allow fallback to DT [arm64,riscv64]
18 off -- disable ACPI if default was on
19 noirq -- do not use ACPI for IRQ routing
20 strict -- Be less tolerant of platforms that are not
21 strictly ACPI specification compliant.
22 rsdt -- prefer RSDT over (default) XSDT
23 copy_dsdt -- copy DSDT to memory
24 For ARM64 and RISCV64, ONLY "acpi=off", "acpi=on" or
25 "acpi=force" are available
27 See also Documentation/power/runtime_pm.rst, pci=noacpi
29 acpi_apic_instance= [ACPI,IOAPIC,EARLY]
31 2: use 2nd APIC table, if available
32 1,0: use 1st APIC table
35 acpi_backlight= [HW,ACPI]
36 { vendor | video | native | none }
37 If set to vendor, prefer vendor-specific driver
38 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
39 of the ACPI video.ko driver.
40 If set to video, use the ACPI video.ko driver.
41 If set to native, use the device's native backlight mode.
42 If set to none, disable the ACPI backlight interface.
44 acpi_force_32bit_fadt_addr [ACPI,EARLY]
45 force FADT to use 32 bit addresses rather than the
46 64 bit X_* addresses. Some firmware have broken 64
47 bit addresses for force ACPI ignore these and use
48 the older legacy 32 bit addresses.
50 acpica_no_return_repair [HW, ACPI]
51 Disable AML predefined validation mechanism
52 This mechanism can repair the evaluation result to make
53 the return objects more ACPI specification compliant.
54 This option is useful for developers to identify the
55 root cause of an AML interpreter issue when the issue
56 has something to do with the repair mechanism.
58 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
59 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
61 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
62 debug output. Bits in debug_layer correspond to a
63 _COMPONENT in an ACPI source file, e.g.,
64 #define _COMPONENT ACPI_EVENTS
65 Bits in debug_level correspond to a level in
66 ACPI_DEBUG_PRINT statements, e.g.,
67 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
68 The debug_level mask defaults to "info". See
69 Documentation/firmware-guide/acpi/debug.rst for more information about
70 debug layers and levels.
72 Enable processor driver info messages:
73 acpi.debug_layer=0x20000000
74 Enable AML "Debug" output, i.e., stores to the Debug
75 object while interpreting AML:
76 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
77 Enable all messages related to ACPI hardware:
78 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
80 Some values produce so much output that the system is
81 unusable. The "log_buf_len" parameter may be useful
82 if you need to capture more output.
84 acpi_enforce_resources= [ACPI]
86 Check for resource conflicts between native drivers
87 and ACPI OperationRegions (SystemIO and SystemMemory
88 only). IO ports and memory declared in ACPI might be
89 used by the ACPI subsystem in arbitrary AML code and
90 can interfere with legacy drivers.
91 strict (default): access to resources claimed by ACPI
92 is denied; legacy drivers trying to access reserved
93 resources will fail to bind to device using them.
94 lax: access to resources claimed by ACPI is allowed;
95 legacy drivers trying to access reserved resources
96 will bind successfully but a warning message is logged.
97 no: ACPI OperationRegions are not marked as reserved,
98 no further checks are performed.
100 acpi_force_table_verification [HW,ACPI,EARLY]
101 Enable table checksum verification during early stage.
102 By default, this is disabled due to x86 early mapping
105 acpi_irq_balance [HW,ACPI]
106 ACPI will balance active IRQs
109 acpi_irq_nobalance [HW,ACPI]
110 ACPI will not move active IRQs (default)
113 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
114 Format: <irq>,<irq>...
116 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
118 Format: <irq>,<irq>...
120 acpi_mask_gpe= [HW,ACPI]
121 Due to the existence of _Lxx/_Exx, some GPEs triggered
122 by unsupported hardware/firmware features can result in
123 GPE floodings that cannot be automatically disabled by
125 This facility can be used to prevent such uncontrolled
127 Format: <byte> or <bitmap-list>
129 acpi_no_auto_serialize [HW,ACPI]
130 Disable auto-serialization of AML methods
131 AML control methods that contain the opcodes to create
132 named objects will be marked as "Serialized" by the
133 auto-serialization feature.
134 This feature is enabled by default.
135 This option allows to turn off the feature.
137 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
140 acpi_no_static_ssdt [HW,ACPI,EARLY]
141 Disable installation of static SSDTs at early boot time
142 By default, SSDTs contained in the RSDT/XSDT will be
143 installed automatically and they will appear under
144 /sys/firmware/acpi/tables.
145 This option turns off this feature.
146 Note that specifying this option does not affect
147 dynamic table installation which will install SSDT
148 tables to /sys/firmware/acpi/tables/dynamic.
150 acpi_no_watchdog [HW,ACPI,WDT]
151 Ignore the ACPI-based watchdog interface (WDAT) and let
152 a native driver control the watchdog device instead.
154 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY]
155 Pass the RSDP address to the kernel, mostly used
156 on machines running EFI runtime service to boot the
157 second kernel for kdump.
159 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
160 Format: To spoof as Windows 98: ="Microsoft Windows"
162 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
163 of 2 which is mandated by ACPI 6) as the supported ACPI
164 specification revision (when using this switch, it may
165 be necessary to carry out a cold reboot _twice_ in a
166 row to make it take effect on the platform firmware).
168 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
169 acpi_osi="string1" # add string1
170 acpi_osi="!string2" # remove string2
171 acpi_osi=!* # remove all strings
172 acpi_osi=! # disable all built-in OS vendor
174 acpi_osi=!! # enable all built-in OS vendor
176 acpi_osi= # disable all strings
178 'acpi_osi=!' can be used in combination with single or
179 multiple 'acpi_osi="string1"' to support specific OS
180 vendor string(s). Note that such command can only
181 affect the default state of the OS vendor strings, thus
182 it cannot affect the default state of the feature group
183 strings and the current state of the OS vendor strings,
184 specifying it multiple times through kernel command line
185 is meaningless. This command is useful when one do not
186 care about the state of the feature group strings which
187 should be controlled by the OSPM.
189 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
190 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
191 can make '_OSI("Windows 2000")' TRUE.
193 'acpi_osi=' cannot be used in combination with other
194 'acpi_osi=' command lines, the _OSI method will not
195 exist in the ACPI namespace. NOTE that such command can
196 only affect the _OSI support state, thus specifying it
197 multiple times through kernel command line is also
200 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
203 'acpi_osi=!*' can be used in combination with single or
204 multiple 'acpi_osi="string1"' to support specific
205 string(s). Note that such command can affect the
206 current state of both the OS vendor strings and the
207 feature group strings, thus specifying it multiple times
208 through kernel command line is meaningful. But it may
209 still not able to affect the final state of a string if
210 there are quirks related to this string. This command
211 is useful when one want to control the state of the
212 feature group strings to debug BIOS issues related to
215 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
216 '_OSI("Module Device")' FALSE.
217 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
218 '_OSI("Module Device")' TRUE.
219 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
221 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
223 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
224 they all will make '_OSI("Windows 2000")' TRUE.
227 Override the pmtimer bug detection: force the kernel
228 to assume that this machine's pmtimer latches its value
229 and always returns good values.
231 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode
232 Format: { level | edge | high | low }
234 acpi_skip_timer_override [HW,ACPI,EARLY]
235 Recognize and ignore IRQ0/pin2 Interrupt Override.
236 For broken nForce2 BIOS resulting in XT-PIC timer.
238 acpi_sleep= [HW,ACPI] Sleep options
239 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig,
240 s4_nohwsig, old_ordering, nonvs,
241 sci_force_enable, nobl }
242 See Documentation/power/video.rst for information on
244 s3_beep is for debugging; it makes the PC's speaker beep
245 as soon as the kernel's real-mode entry point is called.
246 s4_hwsig causes the kernel to check the ACPI hardware
247 signature during resume from hibernation, and gracefully
248 refuse to resume if it has changed. This complies with
249 the ACPI specification but not with reality, since
250 Windows does not do this and many laptops do change it
251 on docking. So the default behaviour is to allow resume
252 and simply warn when the signature changes, unless the
253 s4_hwsig option is enabled.
254 s4_nohwsig prevents ACPI hardware signature from being
255 used (or even warned about) during resume.
256 old_ordering causes the ACPI 1.0 ordering of the _PTS
257 control method, with respect to putting devices into
258 low power states, to be enforced (the ACPI 2.0 ordering
259 of _PTS is used by default).
260 nonvs prevents the kernel from saving/restoring the
261 ACPI NVS memory during suspend/hibernation and resume.
262 sci_force_enable causes the kernel to set SCI_EN directly
263 on resume from S1/S3 (which is against the ACPI spec,
264 but some broken systems don't work without it).
265 nobl causes the internal blacklist of systems known to
266 behave incorrectly in some ways with respect to system
267 suspend and resume to be ignored (use wisely).
269 acpi_use_timer_override [HW,ACPI,EARLY]
270 Use timer override. For some broken Nvidia NF5 boards
271 that require a timer override, but don't have HPET
273 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in
274 kernel's map of available physical RAM.
277 { off | try_unsupported }
278 off: disable AGP support
279 try_unsupported: try to drive unsupported chipsets
280 (may crash computer or cause data corruption)
283 See Documentation/sound/alsa-configuration.rst
286 Allow the default userspace alignment fault handler
287 behaviour to be specified. Bit 0 enables warnings,
288 bit 1 enables fixups, and bit 2 sends a segfault.
290 align_va_addr= [X86-64]
291 Align virtual addresses by clearing slice [14:12] when
292 allocating a VMA at process creation time. This option
293 gives you up to 3% performance improvement on AMD F15h
294 machines (where it is enabled by default) for a
295 CPU-intensive style benchmark, and it can vary highly in
296 a microbenchmark depending on workload and compiler.
298 32: only for 32-bit processes
299 64: only for 64-bit processes
300 on: enable for both 32- and 64-bit processes
301 off: disable for both 32- and 64-bit processes
303 alloc_snapshot [FTRACE]
304 Allocate the ftrace snapshot buffer on boot up when the
305 main buffer is allocated. This is handy if debugging
306 and you need to use tracing_snapshot() on boot up, and
307 do not want to use tracing_snapshot_alloc() as it needs
308 to be done where GFP_KERNEL allocations are allowed.
310 allow_mismatched_32bit_el0 [ARM64,EARLY]
311 Allow execve() of 32-bit applications and setting of the
312 PER_LINUX32 personality on systems where only a strict
313 subset of the CPUs support 32-bit EL0. When this
314 parameter is present, the set of CPUs supporting 32-bit
315 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
316 and hot-unplug operations may be restricted.
318 See Documentation/arch/arm64/asymmetric-32bit.rst for more
321 amd_iommu= [HW,X86-64]
322 Pass parameters to the AMD IOMMU driver in the system.
324 fullflush - Deprecated, equivalent to iommu.strict=1
325 off - do not initialize any AMD IOMMU found in
327 force_isolation - Force device isolation for all
328 devices. The IOMMU driver is not
329 allowed anymore to lift isolation
330 requirements as needed. This option
331 does not override iommu=pt
332 force_enable - Force enable the IOMMU on platforms known
333 to be buggy with IOMMU enabled. Use this
335 pgtbl_v1 - Use v1 page table for DMA-API (Default).
336 pgtbl_v2 - Use v2 page table for DMA-API.
337 irtcachedis - Disable Interrupt Remapping Table (IRT) caching.
339 amd_iommu_dump= [HW,X86-64]
340 Enable AMD IOMMU driver option to dump the ACPI table
341 for AMD IOMMU. With this option enabled, AMD IOMMU
342 driver will print ACPI tables for AMD IOMMU during
343 IOMMU initialization.
345 amd_iommu_intr= [HW,X86-64]
346 Specifies one of the following AMD IOMMU interrupt
348 legacy - Use legacy interrupt remapping mode.
349 vapic - Use virtual APIC mode, which allows IOMMU
350 to inject interrupts directly into guest.
351 This mode requires kvm-amd.avic=1.
352 (Default when IOMMU HW support is present.)
354 amd_pstate= [X86,EARLY]
356 Do not enable amd_pstate as the default
357 scaling driver for the supported processors
359 Use amd_pstate with passive mode as a scaling driver.
360 In this mode autonomous selection is disabled.
361 Driver requests a desired performance level and platform
362 tries to match the same performance level if it is
363 satisfied by guaranteed performance level.
365 Use amd_pstate_epp driver instance as the scaling driver,
366 driver provides a hint to the hardware if software wants
367 to bias toward performance (0x0) or energy efficiency (0xff)
368 to the CPPC firmware. then CPPC power algorithm will
369 calculate the runtime workload and adjust the realtime cores
372 Activate guided autonomous mode. Driver requests minimum and
373 maximum performance level and the platform autonomously
374 selects a performance level in this range and appropriate
375 to the current workload.
377 amijoy.map= [HW,JOY] Amiga joystick support
378 Map of devices attached to JOY0DAT and JOY1DAT
380 See also Documentation/input/joydev/joystick.rst
382 analog.map= [HW,JOY] Analog joystick and gamepad support
383 Specifies type or capabilities of an analog joystick
384 connected to one of 16 gameports
385 Format: <type1>,<type2>,..<type16>
388 Power management functions (SPARCstation-4/5 + deriv.)
390 Disable APC CPU standby support. SPARCstation-Fox does
391 not play well with APC CPU idle - disable it if you have
392 APC and your system crashes randomly.
394 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller
395 Change the output verbosity while booting
396 Format: { quiet (default) | verbose | debug }
397 Change the amount of debugging information output
398 when initialising the APIC and IO-APIC components.
399 For X86-32, this can also be used to specify an APIC
401 Format: apic=driver_name
402 Examples: apic=bigsmp
404 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting
405 Format: { bsp (default) | all | none }
406 bsp: External NMI is delivered only to CPU 0
407 all: External NMIs are broadcast to all CPUs as a
409 none: External NMI is masked for all CPUs. This is
410 useful so that a dump capture kernel won't be
414 See Documentation/networking/ipv6.rst.
416 apm= [APM] Advanced Power Management
417 See header of arch/x86/kernel/apm_32.c.
419 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
420 Format: { "0" | "1" }
421 See security/apparmor/Kconfig help text
424 Default value is set via kernel config option.
426 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
427 Format: <io>,<irq>,<nodeID>
429 arm64.nobti [ARM64] Unconditionally disable Branch Target
430 Identification support
432 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory
433 Set instructions support
435 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
438 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
441 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix
444 arm64.nosve [ARM64] Unconditionally disable Scalable Vector
449 atarimouse= [HW,MOUSE] Atari Mouse
451 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
452 EzKey and similar keyboards
454 atkbd.reset= [HW] Reset keyboard during initialization
456 atkbd.set= [HW] Select keyboard code set
457 Format: <int> (2 = AT (default), 3 = PS/2)
459 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
462 atkbd.softraw= [HW] Choose between synthetic and real raw mode
463 Format: <bool> (0 = real, 1 = synthetic (default))
465 atkbd.softrepeat= [HW]
466 Use software keyboard repeat
468 audit= [KNL] Enable the audit sub-system
469 Format: { "0" | "1" | "off" | "on" }
470 0 | off - kernel audit is disabled and can not be
471 enabled until the next reboot
472 unset - kernel audit is initialized but disabled and
473 will be fully enabled by the userspace auditd.
474 1 | on - kernel audit is initialized and partially
475 enabled, storing at most audit_backlog_limit
476 messages in RAM until it is fully enabled by the
480 audit_backlog_limit= [KNL] Set the audit queue size limit.
481 Format: <int> (must be >=0)
484 bau= [X86_UV] Enable the BAU on SGI UV. The default
485 behavior is to disable the BAU (i.e. bau=0).
486 Format: { "0" | "1" }
489 unset - Disable the BAU.
491 baycom_epp= [HW,AX25]
494 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
496 See header of drivers/net/hamradio/baycom_par.c.
498 baycom_ser_fdx= [HW,AX25]
499 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
500 Format: <io>,<irq>,<mode>[,<baud>]
501 See header of drivers/net/hamradio/baycom_ser_fdx.c.
503 baycom_ser_hdx= [HW,AX25]
504 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
505 Format: <io>,<irq>,<mode>
506 See header of drivers/net/hamradio/baycom_ser_hdx.c.
509 Disable BERT OS support on buggy BIOSes.
511 bgrt_disable [ACPI,X86,EARLY]
512 Disable BGRT to avoid flickering OEM logo.
514 blkdevparts= Manual partition parsing of block device(s) for
515 embedded devices based on command line input.
516 See Documentation/block/cmdline-partition.rst
518 boot_delay= [KNL,EARLY]
519 Milliseconds to delay each printk during boot.
520 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled,
521 and you may also have to specify "lpj=". Boot_delay
522 values larger than 10 seconds (10000) are assumed
523 erroneous and ignored.
526 bootconfig [KNL,EARLY]
527 Extended command line options can be added to an initrd
528 and this will cause the kernel to look for it.
530 See Documentation/admin-guide/bootconfig.rst
532 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
533 bttv.radio= Most important insmod options are available as
535 bttv.pll= See Documentation/admin-guide/media/bttv.rst
538 bulk_remove=off [PPC] This parameter disables the use of the pSeries
539 firmware feature for flushing multiple hpte entries
542 c101= [NET] Moxa C101 synchronous serial card
544 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
545 Sometimes CPU hardware bugs make them report the cache
546 size incorrectly. The kernel will attempt work arounds
547 to fix known problems, but for some CPUs it is not
548 possible to determine what the correct size should be.
549 This option provides an override for these situations.
552 [NET] Specifies amount of time (in seconds) that
553 the kernel should wait for a network carrier. By default
554 it waits 120 seconds.
556 ca_keys= [KEYS] This parameter identifies a specific key(s) on
557 the system trusted keyring to be used for certificate
559 format: { id:<keyid> | builtin }
561 cca= [MIPS,EARLY] Override the kernel pages' cache coherency
562 algorithm. Accepted values range from 0 to 7
563 inclusive. See arch/mips/include/asm/pgtable-bits.h
564 for platform specific values (SB1, Loongson3 and
567 ccw_timeout_log [S390]
568 See Documentation/arch/s390/common_io.rst for details.
570 cgroup_disable= [KNL] Disable a particular controller or optional feature
571 Format: {name of the controller(s) or feature(s) to disable}
572 The effects of cgroup_disable=foo are:
573 - foo isn't auto-mounted if you mount all cgroups in
575 - foo isn't visible as an individually mountable
577 - if foo is an optional feature then the feature is
578 disabled and corresponding cgroup files are not
580 {Currently only "memory" controller deal with this and
581 cut the overhead, others just disable the usage. So
582 only cgroup_disable=memory is actually worthy}
583 Specifying "pressure" disables per-cgroup pressure
584 stall information accounting feature
586 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
587 Format: { { controller | "all" | "named" }
588 [,{ controller | "all" | "named" }...] }
589 Like cgroup_disable, but only applies to cgroup v1;
590 the blacklisted controllers remain available in cgroup2.
591 "all" blacklists all controllers and "named" disables
592 named mounts. Specifying both "all" and "named" disables
595 cgroup_favordynmods= [KNL] Enable or Disable favordynmods.
596 Format: { "true" | "false" }
597 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS.
599 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
601 nosocket -- Disable socket memory accounting.
602 nokmem -- Disable kernel memory accounting.
603 nobpf -- Disable BPF memory accounting.
605 checkreqprot= [SELINUX] Set initial checkreqprot flag value.
606 Format: { "0" | "1" }
607 See security/selinux/Kconfig help text.
608 0 -- check protection applied by kernel (includes
609 any implied execute protection).
610 1 -- check protection requested by application.
611 Default value is set via a kernel config option.
612 Value can be changed at runtime via
613 /sys/fs/selinux/checkreqprot.
614 Setting checkreqprot to 1 is deprecated.
617 See Documentation/arch/s390/common_io.rst for details.
619 clearcpuid=X[,X...] [X86]
620 Disable CPUID feature X for the kernel. See
621 arch/x86/include/asm/cpufeatures.h for the valid bit
622 numbers X. Note the Linux-specific bits are not necessarily
623 stable over kernel options, but the vendor-specific
625 X can also be a string as appearing in the flags: line
626 in /proc/cpuinfo which does not have the above
627 instability issue. However, not all features have names
629 Note that using this option will taint your kernel.
630 Also note that user programs calling CPUID directly
631 or using the feature without checking anything
632 will still see it. This just prevents it from
633 being used by the kernel or shown in /proc/cpuinfo.
634 Also note the kernel might malfunction if you disable
639 Prevents the clock framework from automatically gating
640 clocks that have not been explicitly enabled by a Linux
641 device driver but are enabled in hardware at reset or
642 by the bootloader/firmware. Note that this does not
643 force such clocks to be always-on nor does it reserve
644 those clocks in any way. This parameter is useful for
645 debug and development, but should not be needed on a
646 platform with proper driver support. For more
647 information, see Documentation/driver-api/clk.rst.
649 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
651 Forces specified clocksource (if available) to be used
652 when calculating gettimeofday(). If specified
653 clocksource is not available, it defaults to PIT.
654 Format: { pit | tsc | cyclone | pmtmr }
656 clocksource= Override the default clocksource
658 Override the default clocksource and use the clocksource
659 with the name specified.
660 Some clocksource names to choose from, depending on
662 [all] jiffies (this is the base, fallback clocksource)
664 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
665 pxa_timer,timer3,32k_counter,timer0_1
666 [X86-32] pit,hpet,tsc;
667 scx200_hrt on Geode; cyclone on IBM x440
675 clocksource.arm_arch_timer.evtstrm=
678 Enable/disable the eventstream feature of the ARM
679 architected timer so that code using WFE-based polling
680 loops can be debugged more effectively on production
683 clocksource.verify_n_cpus= [KNL]
684 Limit the number of CPUs checked for clocksources
685 marked with CLOCK_SOURCE_VERIFY_PERCPU that
686 are marked unstable due to excessive skew.
687 A negative value says to check all CPUs, while
688 zero says not to check any. Values larger than
689 nr_cpu_ids are silently truncated to nr_cpu_ids.
690 The actual CPUs are chosen randomly, with
691 no replacement if the same CPU is chosen twice.
693 clocksource-wdtest.holdoff= [KNL]
694 Set the time in seconds that the clocksource
695 watchdog test waits before commencing its tests.
696 Defaults to zero when built as a module and to
697 10 seconds when built into the kernel.
699 cma=nn[MG]@[start[MG][-end[MG]]]
701 Sets the size of kernel global memory area for
702 contiguous memory allocations and optionally the
703 placement constraint by the physical address range of
704 memory allocations. A value of 0 disables CMA
705 altogether. For more information, see
706 kernel/dma/contiguous.c
710 Sets the size of kernel per-numa memory area for
711 contiguous memory allocations. A value of 0 disables
712 per-numa CMA altogether. And If this option is not
713 specified, the default value is 0.
714 With per-numa CMA enabled, DMA users on node nid will
715 first try to allocate buffer from the pernuma area
716 which is located in node nid, if the allocation fails,
717 they will fallback to the global default memory area.
719 numa_cma=<node>:nn[MG][,<node>:nn[MG]]
721 Sets the size of kernel numa memory area for
722 contiguous memory allocations. It will reserve CMA
723 area for the specified node.
725 With numa CMA enabled, DMA users on node nid will
726 first try to allocate buffer from the numa area
727 which is located in node nid, if the allocation fails,
728 they will fallback to the global default memory area.
730 cmo_free_hint= [PPC] Format: { yes | no }
731 Specify whether pages are marked as being inactive
732 when they are freed. This is used in CMO environments
733 to determine OS memory pressure for page stealing by
737 coherent_pool=nn[KMG] [ARM,KNL,EARLY]
738 Sets the size of memory pool for coherent, atomic dma
739 allocations, by default set to 256K.
741 com20020= [HW,NET] ARCnet - COM20020 chipset
743 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
745 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
749 ARCnet - COM90xx chipset (memory-mapped buffers)
750 Format: <io>[,<irq>[,<memstart>]]
752 condev= [HW,S390] console device
755 con3215_drop= [S390,EARLY] 3215 console drop mode.
757 When set to true, drop data on the 3215 console when
758 the console buffer is full. In this case the
759 operator using a 3270 terminal emulator (for example
760 x3270) does not have to enter the clear key for the
761 console output to advance and the kernel to continue.
762 This leads to a much faster boot time when a 3270
763 terminal emulator is active. If no 3270 terminal
764 emulator is used, this parameter has no effect.
766 console= [KNL] Output console device and options.
768 tty<n> Use the virtual console device <n>.
772 Use the specified serial port. The options are of
773 the form "bbbbpnf", where "bbbb" is the baud rate,
774 "p" is parity ("n", "o", or "e"), "n" is number of
775 bits, and "f" is flow control ("r" for RTS or
776 omit it). Default is "9600n8".
778 See Documentation/admin-guide/serial-console.rst for more
780 Documentation/networking/netconsole.rst for an
783 uart[8250],io,<addr>[,options]
784 uart[8250],mmio,<addr>[,options]
785 uart[8250],mmio16,<addr>[,options]
786 uart[8250],mmio32,<addr>[,options]
787 uart[8250],0x<addr>[,options]
788 Start an early, polled-mode console on the 8250/16550
789 UART at the specified I/O port or MMIO address,
790 switching to the matching ttyS device later.
791 MMIO inter-register address stride is either 8-bit
792 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
793 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
794 to be equivalent to 'mmio'. 'options' are specified in
795 the same format described for ttyS above; if unspecified,
796 the h/w is not re-initialized.
798 hvc<n> Use the hypervisor console device <n>. This is for
799 both Xen and PowerPC hypervisors.
802 Use to disable console output, i.e., to have kernel
803 console messages discarded.
804 This must be the only console= parameter used on the
807 If the device connected to the port is not a TTY but a braille
808 device, prepend "brl," before the device type, for instance
810 For now, only VisioBraille is supported.
813 [KNL] Change console messages format
815 By default we print messages on consoles in
816 "[time stamp] text\n" format (time stamp may not be
817 printed, depending on CONFIG_PRINTK_TIME or
818 `printk_time' param).
820 Switch to syslog format: "<%u>[time stamp] text\n"
821 IOW, each message will have a facility and loglevel
822 prefix. The format is similar to one used by syslog()
823 syscall, or to executing "dmesg -S --raw" or to reading
826 consoleblank= [KNL] The console blank (screen saver) timeout in
827 seconds. A value of 0 disables the blank timer.
831 [KNL] Change the default value for
832 /proc/<pid>/coredump_filter.
833 See also Documentation/filesystems/proc.rst.
835 coresight_cpu_debug.enable
838 Enable/disable the CPU sampling based debugging.
839 0: default value, disable debugging
840 1: enable debugging at boot time
842 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
844 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
846 cpuidle.off=1 [CPU_IDLE]
847 disable the cpuidle sub-system
850 [CPU_IDLE] Name of the cpuidle governor to use.
852 cpufreq.off=1 [CPU_FREQ]
853 disable the cpufreq sub-system
855 cpufreq.default_governor=
856 [CPU_FREQ] Name of the default cpufreq governor or
857 policy to use. This governor must be registered in the
858 kernel before the cpufreq driver probes.
861 [X86,EARLY] Delay for N microsec between assert and de-assert
862 of APIC INIT to start processors. This delay occurs
863 on every CPU online, such as boot, and resume from suspend.
867 [SMP] Enable/disable parallel bringup of secondary CPUs
869 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise
870 the parameter has no effect.
872 crash_kexec_post_notifiers
873 Run kdump after running panic-notifiers and dumping
874 kmsg. This only for the users who doubt kdump always
875 succeeds in any situation.
876 Note that this also increases risks of kdump failure,
877 because some panic notifiers can make the crashed
878 kernel more unstable.
880 crashkernel=size[KMG][@offset[KMG]]
881 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel'
882 upon panic. This parameter reserves the physical
883 memory region [offset, offset + size] for that kernel
884 image. If '@offset' is omitted, then a suitable offset
885 is selected automatically.
886 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region
887 under 4G first, and fall back to reserve region above
888 4G when '@offset' hasn't been specified.
889 See Documentation/admin-guide/kdump/kdump.rst for further details.
891 crashkernel=range1:size1[,range2:size2,...][@offset]
892 [KNL] Same as above, but depends on the memory
893 in the running system. The syntax of range is
894 start-[end] where start and end are both
895 a memory unit (amount[KMG]). See also
896 Documentation/admin-guide/kdump/kdump.rst for an example.
898 crashkernel=size[KMG],high
899 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be
901 Allow kernel to allocate physical memory region from top,
902 so could be above 4G if system have more than 4G ram
903 installed. Otherwise memory region will be allocated
904 below 4G, if available.
905 It will be ignored if crashkernel=X is specified.
906 crashkernel=size[KMG],low
907 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G.
908 When crashkernel=X,high is passed, kernel could allocate
909 physical memory region above 4G, that cause second kernel
910 crash on system that require some amount of low memory,
911 e.g. swiotlb requires at least 64M+32K low memory, also
912 enough extra low memory is needed to make sure DMA buffers
913 for 32-bit devices won't run out. Kernel would try to allocate
914 default size of memory below 4G automatically. The default
915 size is platform dependent.
916 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB)
919 --> loongarch: 128MiB
920 This one lets the user specify own low range under 4G
921 for second kernel instead.
922 0: to disable low allocation.
923 It will be ignored when crashkernel=X,high is not used
924 or memory reserved is below 4G.
927 [KNL] Disable crypto self-tests
932 cs89x0_media= [HW,NET]
933 Format: { rj45 | aui | bnc }
935 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU
936 function call handling. When switched on,
937 additional debug data is printed to the console
938 in case a hanging CPU is detected, and that
939 CPU is pinged again in order to try to resolve
940 the hang situation. The default value of this
941 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT
945 See header of drivers/s390/block/dasd_devmap.c.
947 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
948 (one device per port)
949 Format: <port#>,<type>
950 See also Documentation/input/devices/joystick-parport.rst
952 debug [KNL,EARLY] Enable kernel debugging (events log level).
955 [KNL,EARLY] Enable printing [hashed] pointers early in the
956 boot sequence. If enabled, we use a weak hash instead
957 of siphash to hash pointers. Use this option if you are
958 seeing instances of '(___ptrval___)') and need to see a
959 value (hashed pointer) instead. Cryptographically
960 insecure, please do not use on production kernels.
963 [KNL] verbose locking self-tests
965 Print debugging info while doing the locking API
967 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
968 (no extra messages), setting it to -1 (all bits set)
969 will print _a_lot_ more information - normally only
970 useful to lockdep developers.
972 debug_objects [KNL,EARLY] Enable object debugging
974 debug_guardpage_minorder=
975 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this
976 parameter allows control of the order of pages that will
977 be intentionally kept free (and hence protected) by the
978 buddy allocator. Bigger value increase the probability
979 of catching random memory corruption, but reduce the
980 amount of memory for normal system use. The maximum
981 possible value is MAX_PAGE_ORDER/2. Setting this
982 parameter to 1 or 2 should be enough to identify most
983 random memory corruption problems caused by bugs in
984 kernel or driver code when a CPU writes to (or reads
985 from) a random memory location. Note that there exists
986 a class of memory corruptions problems caused by buggy
987 H/W or F/W or by drivers badly programming DMA
988 (basically when memory is written at bus level and the
989 CPU MMU is bypassed) which are not detectable by
990 CONFIG_DEBUG_PAGEALLOC, hence this option will not
991 help tracking down these problems.
994 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
995 enables the feature at boot time. By default, it is
996 disabled and the system will work mostly the same as a
997 kernel built without CONFIG_DEBUG_PAGEALLOC.
998 Note: to get most of debug_pagealloc error reports, it's
999 useful to also enable the page_owner functionality.
1000 on: enable the feature
1002 debugfs= [KNL,EARLY] This parameter enables what is exposed to
1003 userspace and debugfs internal clients.
1004 Format: { on, no-mount, off }
1005 on: All functions are enabled.
1007 Filesystem is not registered but kernel clients can
1008 access APIs and a crashkernel can be used to read
1009 its content. There is nothing to mount.
1010 off: Filesystem is not registered and clients
1011 get a -EPERM as result when trying to register files
1012 or directories within debugfs.
1013 This is equivalent of the runtime functionality if
1014 debugfs was not enabled in the kernel at all.
1015 Default value is set in build-time with a kernel configuration.
1017 debugpat [X86] Enable PAT debugging
1020 [HW] The size of the default HugeTLB page. This is
1021 the size represented by the legacy /proc/ hugepages
1022 APIs. In addition, this is the default hugetlb size
1023 used for shmget(), mmap() and mounting hugetlbfs
1024 filesystems. If not specified, defaults to the
1025 architecture's default huge page size. Huge page
1026 sizes are architecture dependent. See also
1027 Documentation/admin-guide/mm/hugetlbpage.rst.
1030 deferred_probe_timeout=
1031 [KNL] Debugging option to set a timeout in seconds for
1032 deferred probe to give up waiting on dependencies to
1033 probe. Only specific dependencies (subsystems or
1034 drivers) that have opted in will be ignored. A timeout
1035 of 0 will timeout at the end of initcalls. If the time
1036 out hasn't expired, it'll be restarted by each
1037 successful driver registration. This option will also
1038 dump out devices still on the deferred probe list after
1041 delayacct [KNL] Enable per-task delay accounting
1043 dell_smm_hwmon.ignore_dmi=
1044 [HW] Continue probing hardware even if DMI data
1045 indicates that the driver is running on unsupported
1048 dell_smm_hwmon.force=
1049 [HW] Activate driver even if SMM BIOS signature does
1050 not match list of supported models and enable otherwise
1051 blacklisted features.
1053 dell_smm_hwmon.power_status=
1054 [HW] Report power status in /proc/i8k
1055 (disabled by default).
1057 dell_smm_hwmon.restricted=
1058 [HW] Allow controlling fans only if SYS_ADMIN
1061 dell_smm_hwmon.fan_mult=
1062 [HW] Factor to multiply fan speed with.
1064 dell_smm_hwmon.fan_max=
1065 [HW] Maximum configurable fan speed.
1068 Format: { on | off | def_only | inf_only | always }
1069 on: s390 zlib hardware support for compression on
1070 level 1 and decompression (default)
1071 off: No s390 zlib hardware support
1072 def_only: s390 zlib hardware support for deflate
1073 only (compression on level 1)
1074 inf_only: s390 zlib hardware support for inflate
1075 only (decompression)
1076 always: Same as 'on' but ignores the selected compression
1077 level always using hardware support (used for debugging)
1079 dhash_entries= [KNL]
1080 Set number of hash buckets for dentry cache.
1082 disable_1tb_segments [PPC,EARLY]
1083 Disables the use of 1TB hash page table segments. This
1084 causes the kernel to fall back to 256MB segments which
1085 can be useful when debugging issues that require an SLB
1089 See Documentation/networking/ipv6.rst.
1091 disable_radix [PPC,EARLY]
1092 Disable RADIX MMU mode on POWER9
1095 Disable TLBIE instruction. Currently does not work
1096 with KVM, with HASH MMU, or with coherent accelerators.
1098 disable_ddw [PPC/PSERIES,EARLY]
1099 Disable Dynamic DMA Window support. Use this
1100 to workaround buggy firmware.
1102 disable_ipv6= [IPV6]
1103 See Documentation/networking/ipv6.rst.
1105 disable_mtrr_cleanup [X86,EARLY]
1106 The kernel tries to adjust MTRR layout from continuous
1107 to discrete, to make X server driver able to add WB
1108 entry later. This parameter disables that.
1110 disable_mtrr_trim [X86, Intel and AMD only,EARLY]
1111 By default the kernel will trim any uncacheable
1112 memory out of your available memory pool based on
1113 MTRR settings. This parameter disables that behavior,
1114 possibly causing your machine to run very slowly.
1116 disable_timer_pin_1 [X86,EARLY]
1117 Disable PIN 1 of APIC timer
1118 Can be useful to work around chipset bugs.
1120 dis_ucode_ldr [X86] Disable the microcode loader.
1122 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1123 this option disables the debugging code at boot.
1125 dma_debug_entries=<number>
1126 This option allows to tune the number of preallocated
1127 entries for DMA-API debugging code. One entry is
1128 required per DMA-API allocation. Use this if the
1129 DMA-API debugging code disables itself because the
1130 architectural default is too low.
1132 dma_debug_driver=<driver_name>
1133 With this option the DMA-API debugging driver
1134 filter feature can be enabled at boot time. Just
1135 pass the driver to filter for as the parameter.
1136 The filter can be disabled or changed to another
1137 driver later using sysfs.
1139 driver_async_probe= [KNL]
1140 List of driver names to be probed asynchronously. *
1141 matches with all driver names. If * is specified, the
1142 rest of the listed driver names are those that will NOT
1144 Format: <driver_name1>,<driver_name2>...
1146 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1147 Broken monitors, graphic adapters, KVMs and EDIDless
1148 panels may send no or incorrect EDID data sets.
1149 This parameter allows to specify an EDID data sets
1150 in the /lib/firmware directory that are used instead.
1151 Generic built-in EDID data sets are used, if one of
1152 edid/1024x768.bin, edid/1280x1024.bin,
1153 edid/1680x1050.bin, or edid/1920x1080.bin is given
1154 and no file with the same name exists. Details and
1155 instructions how to build your own EDID data are
1156 available in Documentation/admin-guide/edid.rst. An EDID
1157 data set will only be used for a particular connector,
1158 if its name and a colon are prepended to the EDID
1159 name. Each connector may use a unique EDID data
1160 set by separating the files with a comma. An EDID
1161 data set with no connector name will be used for
1162 any connectors not explicitly specified.
1166 dt_cpu_ftrs= [PPC,EARLY]
1167 Format: {"off" | "known"}
1168 Control how the dt_cpu_ftrs device-tree binding is
1169 used for CPU feature discovery and setup (if it
1171 off: Do not use it, fall back to legacy cpu table.
1172 known: Do not pass through unknown features to guests
1173 or userspace, only those that the kernel is aware of.
1175 dump_apple_properties [X86]
1176 Dump name and content of EFI device properties on
1177 x86 Macs. Useful for driver authors to determine
1178 what data is available or for reverse-engineering.
1180 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1181 <module>.dyndbg[="val"]
1182 Enable debug messages at boot time. See
1183 Documentation/admin-guide/dynamic-debug-howto.rst
1186 early_ioremap_debug [KNL,EARLY]
1187 Enable debug messages in early_ioremap support. This
1188 is useful for tracking down temporary early mappings
1189 which are not unmapped.
1191 earlycon= [KNL,EARLY] Output early console device and options.
1193 When used with no options, the early console is
1194 determined by stdout-path property in device tree's
1195 chosen node or the ACPI SPCR table if supported by
1198 cdns,<addr>[,options]
1199 Start an early, polled-mode console on a Cadence
1200 (xuartps) serial port at the specified address. Only
1201 supported option is baud rate. If baud rate is not
1202 specified, the serial port must already be setup and
1205 uart[8250],io,<addr>[,options[,uartclk]]
1206 uart[8250],mmio,<addr>[,options[,uartclk]]
1207 uart[8250],mmio32,<addr>[,options[,uartclk]]
1208 uart[8250],mmio32be,<addr>[,options[,uartclk]]
1209 uart[8250],0x<addr>[,options]
1210 Start an early, polled-mode console on the 8250/16550
1211 UART at the specified I/O port or MMIO address.
1212 MMIO inter-register address stride is either 8-bit
1213 (mmio) or 32-bit (mmio32 or mmio32be).
1214 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1215 to be equivalent to 'mmio'. 'options' are specified
1216 in the same format described for "console=ttyS<n>"; if
1217 unspecified, the h/w is not initialized. 'uartclk' is
1218 the uart clock frequency; if unspecified, it is set
1219 to 'BASE_BAUD' * 16.
1223 Start an early, polled-mode console on a pl011 serial
1224 port at the specified address. The pl011 serial port
1225 must already be setup and configured. Options are not
1226 yet supported. If 'mmio32' is specified, then only
1227 the driver will use only 32-bit accessors to read/write
1228 the device registers.
1231 Start an early console on a litex serial port at the
1232 specified address. The serial port must already be
1233 setup and configured. Options are not yet supported.
1236 Start an early, polled-mode console on a meson serial
1237 port at the specified address. The serial port must
1238 already be setup and configured. Options are not yet
1242 Start an early, polled-mode console on an msm serial
1243 port at the specified address. The serial port
1244 must already be setup and configured. Options are not
1247 msm_serial_dm,<addr>
1248 Start an early, polled-mode console on an msm serial
1249 dm port at the specified address. The serial port
1250 must already be setup and configured. Options are not
1254 Start an early, polled-mode console on a serial port
1255 of an Actions Semi SoC, such as S500 or S900, at the
1256 specified address. The serial port must already be
1257 setup and configured. Options are not yet supported.
1260 Start an early, polled-mode console on a serial port
1261 of an RDA Micro SoC, such as RDA8810PL, at the
1262 specified address. The serial port must already be
1263 setup and configured. Options are not yet supported.
1266 Use RISC-V SBI (Supervisor Binary Interface) for early
1269 smh Use ARM semihosting calls for early console.
1277 Use early console provided by serial driver available
1278 on Samsung SoCs, requires selecting proper type and
1279 a correct base address of the selected UART port. The
1280 serial port must already be setup and configured.
1281 Options are not yet supported.
1284 Start an early, polled-mode console on a lantiq serial
1285 (lqasc) port at the specified address. The serial port
1286 must already be setup and configured. Options are not
1291 Use early console provided by Freescale LP UART driver
1292 found on Freescale Vybrid and QorIQ LS1021A processors.
1293 A valid base address must be provided, and the serial
1294 port must already be setup and configured.
1298 Start an early, polled-mode, output-only console on the
1299 Freescale i.MX UART at the specified address. The UART
1300 must already be setup and configured.
1303 Start an early, polled-mode console on the
1304 Armada 3700 serial port at the specified
1305 address. The serial port must already be setup
1306 and configured. Options are not yet supported.
1309 Start an early, polled-mode console on a Qualcomm
1310 Generic Interface (GENI) based serial port at the
1311 specified address. The serial port must already be
1312 setup and configured. Options are not yet supported.
1315 Start an early, unaccelerated console on the EFI
1316 memory mapped framebuffer (if available). On cache
1317 coherent non-x86 systems that use system memory for
1318 the framebuffer, pass the 'ram' option so that it is
1319 mapped with the correct attributes.
1322 Use early console provided by Freescale LINFlexD UART
1323 serial driver for NXP S32V234 SoCs. A valid base
1324 address must be provided, and the serial port must
1325 already be setup and configured.
1327 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY]
1331 earlyprintk=serial[,ttySn[,baudrate]]
1332 earlyprintk=serial[,0x...[,baudrate]]
1333 earlyprintk=ttySn[,baudrate]
1334 earlyprintk=dbgp[debugController#]
1335 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1336 earlyprintk=xdbc[xhciController#]
1339 earlyprintk is useful when the kernel crashes before
1340 the normal console is initialized. It is not enabled by
1341 default because it has some cosmetic problems.
1343 Append ",keep" to not disable it when the real console
1346 Only one of vga, serial, or usb debug port can
1349 Currently only ttyS0 and ttyS1 may be specified by
1350 name. Other I/O ports may be explicitly specified
1351 on some architectures (x86 and arm at least) by
1352 replacing ttySn with an I/O port address, like this:
1353 earlyprintk=serial,0x1008,115200
1354 You can find the port for a given device in
1355 /proc/tty/driver/serial:
1356 2: uart:ST16650V2 port:00001008 irq:18 ...
1358 Interaction with the standard serial driver is not
1361 The VGA output is eventually overwritten by
1364 The xen option can only be used in Xen domains.
1366 The sclp output can only be used on s390.
1368 The bios output can only be used on SuperH.
1370 The optional "force" to "pciserial" enables use of a
1371 PCI device even when its classcode is not of the
1374 edac_report= [HW,EDAC] Control how to report EDAC event
1375 Format: {"on" | "off" | "force"}
1376 on: enable EDAC to report H/W event. May be overridden
1377 by other higher priority error reporting module.
1378 off: disable H/W event reporting through EDAC.
1379 force: enforce the use of EDAC to report H/W event.
1383 Format: {"off" | "on" | "skip[mbr]"}
1386 Format: { "debug", "disable_early_pci_dma",
1387 "nochunk", "noruntime", "nosoftreserve",
1388 "novamap", "no_disable_early_pci_dma" }
1389 debug: enable misc debug output.
1390 disable_early_pci_dma: disable the busmaster bit on all
1391 PCI bridges while in the EFI boot stub.
1392 nochunk: disable reading files in "chunks" in the EFI
1393 boot stub, as chunking can cause problems with some
1394 firmware implementations.
1395 noruntime : disable EFI runtime services support
1396 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1397 attribute may cause the kernel to reserve the
1398 memory range for a memory mapping driver to
1399 claim. Specify efi=nosoftreserve to disable this
1400 reservation and treat the memory by its base type
1401 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1402 novamap: do not call SetVirtualAddressMap().
1403 no_disable_early_pci_dma: Leave the busmaster bit set
1404 on all PCI bridges while in the EFI boot stub
1406 efi_no_storage_paranoia [EFI,X86,EARLY]
1407 Using this parameter you can use more than 50% of
1408 your efi variable storage. Use this parameter only if
1409 you are really sure that your UEFI does sane gc and
1410 fulfills the spec otherwise your board may brick.
1412 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI,X86,EARLY]
1413 Add arbitrary attribute to specific memory range by
1414 updating original EFI memory map.
1415 Region of memory which aa attribute is added to is
1418 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1419 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1420 attribute is added to range 0x100000000-0x180000000 and
1421 0x10a0000000-0x1120000000.
1423 If efi_fake_mem=8G@9G:0x40000 is specified, the
1424 EFI_MEMORY_SP(0x40000) attribute is added to
1425 range 0x240000000-0x43fffffff.
1427 Using this parameter you can do debugging of EFI memmap
1428 related features. For example, you can do debugging of
1429 Address Range Mirroring feature even if your box
1430 doesn't support it, or mark specific memory as
1433 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1434 that is to be dynamically loaded by Linux. If there are
1435 multiple variables with the same name but with different
1436 vendor GUIDs, all of them will be loaded. See
1437 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1440 eisa_irq_edge= [PARISC,HW]
1441 See header of drivers/parisc/eisa.c.
1443 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging
1446 This is designed to be used in conjunction with
1447 the boot argument: earlyprintk=vga
1449 This parameter works in place of the kgdboc parameter
1450 but can only be used if the backing tty is available
1451 very early in the boot process. For early debugging
1452 via a serial port see kgdboc_earlycon instead.
1455 See comment before function elanfreq_setup() in
1456 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1458 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY]
1459 Specifies physical address of start of kernel core
1460 image elf header and optionally the size. Generally
1461 kexec loader will pass this option to capture kernel.
1462 See Documentation/admin-guide/kdump/kdump.rst for details.
1464 enable_mtrr_cleanup [X86,EARLY]
1465 The kernel tries to adjust MTRR layout from continuous
1466 to discrete, to make X server driver able to add WB
1467 entry later. This parameter enables that.
1469 enable_timer_pin_1 [X86]
1470 Enable PIN 1 of APIC timer
1471 Can be useful to work around chipset bugs
1472 (in particular on some ATI chipsets).
1473 The kernel tries to set a reasonable default.
1475 enforcing= [SELINUX] Set initial enforcing status.
1477 See security/selinux/Kconfig help text.
1478 0 -- permissive (log only, no denials).
1479 1 -- enforcing (deny and log).
1481 Value can be changed at runtime via
1482 /sys/fs/selinux/enforce.
1485 Disable Error Record Serialization Table (ERST)
1488 ether= [HW,NET] Ethernet cards parameters
1489 This option is obsoleted by the "netdev=" option, which
1490 has equivalent usage. See its documentation for details.
1494 Permit 'security.evm' to be updated regardless of
1495 current integrity status.
1497 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier
1498 stages so cover more early boot allocations.
1499 Please note that as side effect some optimizations
1500 might be disabled to achieve that (e.g. parallelized
1501 memory initialization is disabled) so the boot process
1502 might take longer, especially on systems with a lot of
1503 memory. Available with CONFIG_PAGE_EXTENSION=y.
1508 fail_make_request=[KNL]
1509 General fault injection mechanism.
1510 Format: <interval>,<probability>,<space>,<times>
1511 See also Documentation/fault-injection/.
1514 Format: { initns | none }
1515 See Documentation/admin-guide/sysctl/net.rst for
1516 fb_tunnels_only_for_init_ns
1519 See Documentation/admin-guide/blockdev/floppy.rst.
1522 Forcefully enable Physical Address Extension (PAE).
1523 Many Pentium M systems disable PAE but may have a
1524 functionally usable PAE implementation.
1525 Warning: use of this parameter will taint the kernel
1526 and may cause unknown problems.
1529 [FTRACE] will set and start the specified tracer
1530 as early as possible in order to facilitate early
1533 ftrace_boot_snapshot
1534 [FTRACE] On boot up, a snapshot will be taken of the
1535 ftrace ring buffer that can be read at:
1536 /sys/kernel/tracing/snapshot.
1537 This is useful if you need tracing information from kernel
1538 boot up that is likely to be overridden by user space
1539 start up functionality.
1541 Optionally, the snapshot can also be defined for a tracing
1542 instance that was created by the trace_instance= command
1545 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo
1547 The above will cause the "foo" tracing instance to trigger
1548 a snapshot at the end of boot up.
1550 ftrace_dump_on_oops[=orig_cpu]
1551 [FTRACE] will dump the trace buffers on oops.
1552 If no parameter is passed, ftrace will dump
1553 buffers of all CPUs, but if you pass orig_cpu, it will
1554 dump only the buffer of the CPU that triggered the
1557 ftrace_filter=[function-list]
1558 [FTRACE] Limit the functions traced by the function
1559 tracer at boot up. function-list is a comma-separated
1560 list of functions. This list can be changed at run
1561 time by the set_ftrace_filter file in the debugfs
1564 ftrace_notrace=[function-list]
1565 [FTRACE] Do not trace the functions specified in
1566 function-list. This list can be changed at run time
1567 by the set_ftrace_notrace file in the debugfs
1570 ftrace_graph_filter=[function-list]
1571 [FTRACE] Limit the top level callers functions traced
1572 by the function graph tracer at boot up.
1573 function-list is a comma-separated list of functions
1574 that can be changed at run time by the
1575 set_graph_function file in the debugfs tracing directory.
1577 ftrace_graph_notrace=[function-list]
1578 [FTRACE] Do not trace from the functions specified in
1579 function-list. This list is a comma-separated list of
1580 functions that can be changed at run time by the
1581 set_graph_notrace file in the debugfs tracing directory.
1583 ftrace_graph_max_depth=<uint>
1584 [FTRACE] Used with the function graph tracer. This is
1585 the max depth it will trace into a function. This value
1586 can be changed at run time by the max_graph_depth file
1587 in the tracefs tracing directory. default: 0 (no limit)
1589 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier
1590 devices by scanning the firmware to infer the
1591 consumer/supplier relationships. This feature is
1592 especially useful when drivers are loaded as modules as
1593 it ensures proper ordering of tasks like device probing
1594 (suppliers first, then consumers), supplier boot state
1595 clean up (only after all consumers have probed),
1596 suspend/resume & runtime PM (consumers first, then
1598 Format: { off | permissive | on | rpm }
1599 off -- Don't create device links from firmware info.
1600 permissive -- Create device links from firmware info
1601 but use it only for ordering boot state clean
1602 up (sync_state() calls).
1603 on -- Create device links from firmware info and use it
1604 to enforce probe and suspend/resume ordering.
1605 rpm -- Like "on", but also use to order runtime PM.
1607 fw_devlink.strict=<bool>
1608 [KNL,EARLY] Treat all inferred dependencies as mandatory
1609 dependencies. This only applies for fw_devlink=on|rpm.
1612 fw_devlink.sync_state =
1613 [KNL,EARLY] When all devices that could probe have finished
1614 probing, this parameter controls what to do with
1615 devices that haven't yet received their sync_state()
1617 Format: { strict | timeout }
1618 strict -- Default. Continue waiting on consumers to
1620 timeout -- Give up waiting on consumers and call
1621 sync_state() on any devices that haven't yet
1622 received their sync_state() calls after
1623 deferred_probe_timeout has expired or by
1624 late_initcall() if !CONFIG_MODULES.
1627 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1628 support via parallel port (up to 5 devices per port)
1629 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1630 See also Documentation/input/devices/joystick-parport.rst
1634 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART
1638 gather_data_sampling=
1639 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS)
1642 Gather Data Sampling is a hardware vulnerability which
1643 allows unprivileged speculative access to data which was
1644 previously stored in vector registers.
1646 This issue is mitigated by default in updated microcode.
1647 The mitigation may have a performance impact but can be
1648 disabled. On systems without the microcode mitigation
1649 disabling AVX serves as a mitigation.
1651 force: Disable AVX to mitigate systems without
1652 microcode mitigation. No effect if the microcode
1653 mitigation is present. Known to cause crashes in
1654 userspace with buggy AVX enumeration.
1656 off: Disable GDS mitigation.
1658 gcov_persist= [GCOV] When non-zero (default), profiling data for
1659 kernel modules is saved and remains accessible via
1660 debugfs, even when the module is unloaded/reloaded.
1661 When zero, profiling data is discarded and associated
1662 debugfs files are removed at module unload time.
1664 goldfish [X86] Enable the goldfish android emulator platform.
1665 Don't use this when you are not running on the
1668 gpio-mockup.gpio_mockup_ranges
1669 [HW] Sets the ranges of gpiochip of for this device.
1670 Format: <start1>,<end1>,<start2>,<end2>...
1671 gpio-mockup.gpio_mockup_named_lines
1672 [HW] Let the driver know GPIO lines should be named.
1674 gpt [EFI] Forces disk with valid GPT signature but
1675 invalid Protective MBR to be treated as GPT. If the
1676 primary GPT is corrupted, it enables the backup/alternate
1677 GPT to be used instead.
1679 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1680 the "Enable 0" bit of the configuration register.
1683 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1684 the "Enable 0" bit of the configuration register.
1687 grcan.select= [HW] Select which physical interface to use.
1690 grcan.txsize= [HW] Sets the size of the tx buffer.
1691 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1693 grcan.rxsize= [HW] Sets the size of the rx buffer.
1694 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1698 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1699 hardening is enabled for this boot. Hardened
1700 usercopy checking is used to protect the kernel
1701 from reading or writing beyond known memory
1702 allocation boundaries as a proactive defense
1703 against bounds-checking flaws in the kernel's
1704 copy_to_user()/copy_from_user() interface.
1705 on Perform hardened usercopy checks (default).
1706 off Disable hardened usercopy checks.
1708 hardlockup_all_cpu_backtrace=
1709 [KNL] Should the hard-lockup detector generate
1710 backtraces on all cpus.
1713 hashdist= [KNL,NUMA] Large hashes allocated during boot
1714 are distributed across NUMA nodes. Defaults on
1715 for 64-bit NUMA, off otherwise.
1716 Format: 0 | 1 (for off | on)
1718 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1720 hd= [EIDE] (E)IDE hard drive subsystem geometry
1721 Format: <cyl>,<head>,<sect>
1724 Disable Hardware Error Source Table (HEST) support;
1725 corresponding firmware-first mode error processing
1726 logic will be disabled.
1728 hibernate= [HIBERNATION]
1729 noresume Don't check if there's a hibernation image
1730 present during boot.
1731 nocompress Don't compress/decompress hibernation images.
1732 no Disable hibernation and resume.
1733 protect_image Turn on image protection during restoration
1734 (that will set all pages holding image data
1735 during restoration read-only).
1737 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact
1738 size of <nn>. This works even on boxes that have no
1739 highmem otherwise. This also works to reduce highmem
1740 size on bigger boxes.
1742 highres= [KNL] Enable/disable high resolution timer mode.
1743 Valid parameters: "on", "off"
1748 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename).
1750 This allows setting the system's hostname during early
1751 startup. This sets the name returned by gethostname.
1752 Using this parameter to set the hostname makes it
1753 possible to ensure the hostname is correctly set before
1754 any userspace processes run, avoiding the possibility
1755 that a process may call gethostname before the hostname
1756 has been explicitly set, resulting in the calling
1757 process getting an incorrect result. The string must
1758 not exceed the maximum allowed hostname length (usually
1759 64 characters) and will be truncated otherwise.
1761 hpet= [X86-32,HPET] option to control HPET usage
1762 Format: { enable (default) | disable | force |
1764 disable: disable HPET and use PIT instead
1765 force: allow force enabled of undocumented chips (ICH4,
1767 verbose: show contents of HPET registers during setup
1769 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1770 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1772 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1773 If this follows hugepagesz (below), it specifies
1774 the number of pages of hugepagesz to be allocated.
1775 If this is the first HugeTLB parameter on the command
1776 line, it specifies the number of pages to allocate for
1777 the default huge page size. If using node format, the
1778 number of pages to allocate per-node can be specified.
1779 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1780 Format: <integer> or (node format)
1781 <node>:<integer>[,<node>:<integer>]
1784 [HW] The size of the HugeTLB pages. This is used in
1785 conjunction with hugepages (above) to allocate huge
1786 pages of a specific size at boot. The pair
1787 hugepagesz=X hugepages=Y can be specified once for
1788 each supported huge page size. Huge page sizes are
1789 architecture dependent. See also
1790 Documentation/admin-guide/mm/hugetlbpage.rst.
1793 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation
1794 of gigantic hugepages. Or using node format, the size
1795 of a CMA area per node can be specified.
1796 Format: nn[KMGTPE] or (node format)
1797 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1799 Reserve a CMA area of given size and allocate gigantic
1800 hugepages using the CMA allocator. If enabled, the
1801 boot-time allocation of gigantic hugepages is skipped.
1803 hugetlb_free_vmemmap=
1804 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1806 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1807 Allows heavy hugetlb users to free up some more
1808 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1809 Format: { on | off (default) }
1814 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1817 Note that the vmemmap pages may be allocated from the added
1818 memory block itself when memory_hotplug.memmap_on_memory is
1819 enabled, those vmemmap pages cannot be optimized even if this
1820 feature is enabled. Other vmemmap pages not allocated from
1821 the added memory block itself do not be affected.
1824 [KNL] Should the hung task detector generate panics.
1827 A value of 1 instructs the kernel to panic when a
1828 hung task is detected. The default value is controlled
1829 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1830 option. The value selected by this boot parameter can
1831 be changed later by the kernel.hung_task_panic sysctl.
1833 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1834 terminal devices. Valid values: 0..8
1835 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1836 If specified, z/VM IUCV HVC accepts connections
1837 from listed z/VM user IDs only.
1839 hv_nopvspin [X86,HYPER_V,EARLY]
1840 Disables the paravirt spinlock optimizations
1841 which allow the hypervisor to 'idle' the guest
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,EARLY]
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,EARLY] Specify the location of the initial ramdisk
2083 initrdmem= [KNL,EARLY] 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,EARLY] Fill newly allocated pages and heap objects with
2093 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2095 init_on_free= [MM,EARLY] 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.
2151 intel_pstate= [X86,EARLY]
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,EARLY]
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,EARLY] 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,S390,EARLY] 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,EARLY] 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,EARLY] 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,EARLY] 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,EARLY]
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
2435 keep_bootcon [KNL,EARLY]
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
2441 keepinitrd [HW,ARM] See retain_initrd.
2443 kernelcore= [KNL,X86,IA-64,PPC,EARLY]
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,EARLY] 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,EARLY]
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,EARLY] 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,EARLY] 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,EARLY] Control page table isolation of
2531 user 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,EARLY] Select one of KVM/arm64's modes of
2611 none: Forcefully disable KVM.
2613 nvhe: Standard nVHE-based mode, without support for
2616 protected: nVHE-based mode with support for guests whose
2617 state is kept private from the host.
2619 nested: VHE-based mode with support for nested
2620 virtualization. Requires at least ARMv8.3
2623 Defaults to VHE/nVHE based on hardware support. Setting
2624 mode to "protected" will disable kexec and hibernation
2625 for the host. "nested" is experimental and should be
2626 used with extreme caution.
2628 kvm-arm.vgic_v3_group0_trap=
2629 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0
2632 kvm-arm.vgic_v3_group1_trap=
2633 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1
2636 kvm-arm.vgic_v3_common_trap=
2637 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common
2640 kvm-arm.vgic_v4_enable=
2641 [KVM,ARM,EARLY] Allow use of GICv4 for direct
2644 kvm_cma_resv_ratio=n [PPC,EARLY]
2645 Reserves given percentage from system memory area for
2646 contiguous memory allocation for KVM hash pagetable
2648 By default it reserves 5% of total system memory.
2652 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables,
2653 a.k.a. Two-Dimensional Page Tables. Default is 1
2654 (enabled). Disable by KVM if hardware lacks support
2657 kvm-intel.emulate_invalid_guest_state=
2658 [KVM,Intel] Control whether to emulate invalid guest
2659 state. Ignored if kvm-intel.enable_unrestricted_guest=1,
2660 as guest state is never invalid for unrestricted
2661 guests. This param doesn't apply to nested guests (L2),
2662 as KVM never emulates invalid L2 guest state.
2663 Default is 1 (enabled).
2665 kvm-intel.flexpriority=
2666 [KVM,Intel] Control KVM's use of FlexPriority feature
2667 (TPR shadow). Default is 1 (enabled). Disable by KVM if
2668 hardware lacks support for it.
2671 [KVM,Intel] Control nested virtualization feature in
2672 KVM/VMX. Default is 1 (enabled).
2674 kvm-intel.unrestricted_guest=
2675 [KVM,Intel] Control KVM's use of unrestricted guest
2676 feature (virtualized real and unpaged mode). Default
2677 is 1 (enabled). Disable by KVM if EPT is disabled or
2678 hardware lacks support for it.
2680 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2683 Valid arguments: never, cond, always
2685 always: L1D cache flush on every VMENTER.
2686 cond: Flush L1D on VMENTER only when the code between
2687 VMEXIT and VMENTER can leak host memory.
2688 never: Disables the mitigation
2690 Default is cond (do L1 cache flush in specific instances)
2692 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor
2693 Identification feature (tagged TLBs). Default is 1
2694 (enabled). Disable by KVM if hardware lacks support
2697 l1d_flush= [X86,INTEL,EARLY]
2698 Control mitigation for L1D based snooping vulnerability.
2700 Certain CPUs are vulnerable to an exploit against CPU
2701 internal buffers which can forward information to a
2702 disclosure gadget under certain conditions.
2704 In vulnerable processors, the speculatively
2705 forwarded data can be used in a cache side channel
2706 attack, to access data to which the attacker does
2707 not have direct access.
2709 This parameter controls the mitigation. The
2712 on - enable the interface for the mitigation
2714 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on
2717 The kernel PTE inversion protection is unconditionally
2718 enabled and cannot be disabled.
2721 Provides all available mitigations for the
2722 L1TF vulnerability. Disables SMT and
2723 enables all mitigations in the
2724 hypervisors, i.e. unconditional L1D flush.
2726 SMT control and L1D flush control via the
2727 sysfs interface is still possible after
2728 boot. Hypervisors will issue a warning
2729 when the first VM is started in a
2730 potentially insecure configuration,
2731 i.e. SMT enabled or L1D flush disabled.
2734 Same as 'full', but disables SMT and L1D
2735 flush runtime control. Implies the
2736 'nosmt=force' command line option.
2737 (i.e. sysfs control of SMT is disabled.)
2740 Leaves SMT enabled and enables the default
2741 hypervisor mitigation, i.e. conditional
2744 SMT control and L1D flush control via the
2745 sysfs interface is still possible after
2746 boot. Hypervisors will issue a warning
2747 when the first VM is started in a
2748 potentially insecure configuration,
2749 i.e. SMT enabled or L1D flush disabled.
2753 Disables SMT and enables the default
2754 hypervisor mitigation.
2756 SMT control and L1D flush control via the
2757 sysfs interface is still possible after
2758 boot. Hypervisors will issue a warning
2759 when the first VM is started in a
2760 potentially insecure configuration,
2761 i.e. SMT enabled or L1D flush disabled.
2764 Same as 'flush', but hypervisors will not
2765 warn when a VM is started in a potentially
2766 insecure configuration.
2769 Disables hypervisor mitigations and doesn't
2771 It also drops the swap size and available
2772 RAM limit restriction on both hypervisor and
2777 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2783 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS
2786 lapic= [X86,APIC] Do not use TSC deadline
2787 value for LAPIC timer one-shot implementation. Default
2788 back to the programmable timer unit in the LAPIC.
2789 Format: notscdeadline
2791 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer
2794 libata.dma= [LIBATA] DMA control
2795 libata.dma=0 Disable all PATA and SATA DMA
2796 libata.dma=1 PATA and SATA Disk DMA only
2797 libata.dma=2 ATAPI (CDROM) DMA only
2798 libata.dma=4 Compact Flash DMA only
2799 Combinations also work, so libata.dma=3 enables DMA
2800 for disks and CDROMs, but not CFs.
2802 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2803 libata.ignore_hpa=0 keep BIOS limits (default)
2804 libata.ignore_hpa=1 ignore limits, using full disk
2806 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2810 libata.force= [LIBATA] Force configurations. The format is a comma-
2811 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2812 PORT and DEVICE are decimal numbers matching port, link
2813 or device. Basically, it matches the ATA ID string
2814 printed on console by libata. If the whole ID part is
2815 omitted, the last PORT and DEVICE values are used. If
2816 ID hasn't been specified yet, the configuration applies
2817 to all ports, links and devices.
2819 If only DEVICE is omitted, the parameter applies to
2820 the port and all links and devices behind it. DEVICE
2821 number of 0 either selects the first device or the
2822 first fan-out link behind PMP device. It does not
2823 select the host link. DEVICE number of 15 selects the
2824 host link and device attached to it.
2826 The VAL specifies the configuration to force. As long
2827 as there is no ambiguity, shortcut notation is allowed.
2828 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2829 The following configurations can be forced.
2831 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2832 Any ID with matching PORT is used.
2834 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2836 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2837 udma[/][16,25,33,44,66,100,133] notation is also
2840 * nohrst, nosrst, norst: suppress hard, soft and both
2843 * rstonce: only attempt one reset during hot-unplug
2846 * [no]dbdelay: Enable or disable the extra 200ms delay
2847 before debouncing a link PHY and device presence
2850 * [no]ncq: Turn on or off NCQ.
2852 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2854 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2856 * [no]trim: Enable or disable (unqueued) TRIM.
2858 * trim_zero: Indicate that TRIM command zeroes data.
2860 * max_trim_128m: Set 128M maximum trim size limit.
2862 * [no]dma: Turn on or off DMA transfers.
2864 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2866 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2867 commands that are not a multiple of 16 bytes.
2869 * [no]dmalog: Enable or disable the use of the
2870 READ LOG DMA EXT command to access logs.
2872 * [no]iddevlog: Enable or disable access to the
2873 identify device data log.
2875 * [no]logdir: Enable or disable access to the general
2876 purpose log directory.
2878 * max_sec_128: Set transfer size limit to 128 sectors.
2880 * max_sec_1024: Set or clear transfer size limit to
2883 * max_sec_lba48: Set or clear transfer size limit to
2886 * [no]lpm: Enable or disable link power management.
2888 * [no]setxfer: Indicate if transfer speed mode setting
2891 * [no]fua: Disable or enable FUA (Force Unit Access)
2892 support for devices supporting this feature.
2894 * dump_id: Dump IDENTIFY data.
2896 * disable: Disable this device.
2898 If there are multiple matching configurations changing
2899 the same attribute, the last one is used.
2901 load_ramdisk= [RAM] [Deprecated]
2903 lockd.nlm_grace_period=P [NFS] Assign grace period.
2906 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2909 lockd.nlm_timeout=T [NFS] Assign timeout value.
2912 lockd.nlm_udpport=M [NFS] Assign UDP port.
2915 lockdown= [SECURITY,EARLY]
2916 { integrity | confidentiality }
2917 Enable the kernel lockdown feature. If set to
2918 integrity, kernel features that allow userland to
2919 modify the running kernel are disabled. If set to
2920 confidentiality, kernel features that allow userland
2921 to extract confidential information from the kernel
2924 locktorture.acq_writer_lim= [KNL]
2925 Set the time limit in jiffies for a lock
2926 acquisition. Acquisitions exceeding this limit
2927 will result in a splat once they do complete.
2929 locktorture.bind_readers= [KNL]
2930 Specify the list of CPUs to which the readers are
2933 locktorture.bind_writers= [KNL]
2934 Specify the list of CPUs to which the writers are
2937 locktorture.call_rcu_chains= [KNL]
2938 Specify the number of self-propagating call_rcu()
2939 chains to set up. These are used to ensure that
2940 there is a high probability of an RCU grace period
2941 in progress at any given time. Defaults to 0,
2942 which disables these call_rcu() chains.
2944 locktorture.long_hold= [KNL]
2945 Specify the duration in milliseconds for the
2946 occasional long-duration lock hold time. Defaults
2947 to 100 milliseconds. Select 0 to disable.
2949 locktorture.nested_locks= [KNL]
2950 Specify the maximum lock nesting depth that
2951 locktorture is to exercise, up to a limit of 8
2952 (MAX_NESTED_LOCKS). Specify zero to disable.
2953 Note that this parameter is ineffective on types
2954 of locks that do not support nested acquisition.
2956 locktorture.nreaders_stress= [KNL]
2957 Set the number of locking read-acquisition kthreads.
2958 Defaults to being automatically set based on the
2959 number of online CPUs.
2961 locktorture.nwriters_stress= [KNL]
2962 Set the number of locking write-acquisition kthreads.
2964 locktorture.onoff_holdoff= [KNL]
2965 Set time (s) after boot for CPU-hotplug testing.
2967 locktorture.onoff_interval= [KNL]
2968 Set time (s) between CPU-hotplug operations, or
2969 zero to disable CPU-hotplug testing.
2971 locktorture.rt_boost= [KNL]
2972 Do periodic testing of real-time lock priority
2973 boosting. Select 0 to disable, 1 to boost
2974 only rt_mutex, and 2 to boost unconditionally.
2975 Defaults to 2, which might seem to be an
2976 odd choice, but which should be harmless for
2977 non-real-time spinlocks, due to their disabling
2978 of preemption. Note that non-realtime mutexes
2981 locktorture.rt_boost_factor= [KNL]
2982 Number that determines how often and for how
2983 long priority boosting is exercised. This is
2984 scaled down by the number of writers, so that the
2985 number of boosts per unit time remains roughly
2986 constant as the number of writers increases.
2987 On the other hand, the duration of each boost
2988 increases with the number of writers.
2990 locktorture.shuffle_interval= [KNL]
2991 Set task-shuffle interval (jiffies). Shuffling
2992 tasks allows some CPUs to go into dyntick-idle
2993 mode during the locktorture test.
2995 locktorture.shutdown_secs= [KNL]
2996 Set time (s) after boot system shutdown. This
2997 is useful for hands-off automated testing.
2999 locktorture.stat_interval= [KNL]
3000 Time (s) between statistics printk()s.
3002 locktorture.stutter= [KNL]
3003 Time (s) to stutter testing, for example,
3004 specifying five seconds causes the test to run for
3005 five seconds, wait for five seconds, and so on.
3006 This tests the locking primitive's ability to
3007 transition abruptly to and from idle.
3009 locktorture.torture_type= [KNL]
3010 Specify the locking implementation to test.
3012 locktorture.verbose= [KNL]
3013 Enable additional printk() statements.
3015 locktorture.writer_fifo= [KNL]
3016 Run the write-side locktorture kthreads at
3017 sched_set_fifo() real-time priority.
3019 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
3022 loglevel= [KNL,EARLY]
3023 All Kernel Messages with a loglevel smaller than the
3024 console loglevel will be printed to the console. It can
3025 also be changed with klogd or other programs. The
3026 loglevels are defined as follows:
3028 0 (KERN_EMERG) system is unusable
3029 1 (KERN_ALERT) action must be taken immediately
3030 2 (KERN_CRIT) critical conditions
3031 3 (KERN_ERR) error conditions
3032 4 (KERN_WARNING) warning conditions
3033 5 (KERN_NOTICE) normal but significant condition
3034 6 (KERN_INFO) informational
3035 7 (KERN_DEBUG) debug-level messages
3037 log_buf_len=n[KMG] [KNL,EARLY]
3038 Sets the size of the printk ring buffer, in bytes.
3039 n must be a power of two and greater than the
3040 minimal size. The minimal size is defined by
3041 LOG_BUF_SHIFT kernel config parameter. There
3042 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config
3043 parameter that allows to increase the default size
3044 depending on the number of CPUs. See init/Kconfig
3047 logo.nologo [FB] Disables display of the built-in Linux logo.
3048 This may be used to provide more screen space for
3049 kernel log messages and is useful when debugging
3050 kernel boot problems.
3052 lp=0 [LP] Specify parallel ports to use, e.g,
3053 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
3054 lp=reset first parallel port). 'lp=0' disables the
3055 lp=auto printer driver. 'lp=reset' (which can be
3056 specified in addition to the ports) causes
3057 attached printers to be reset. Using
3058 lp=port1,port2,... specifies the parallel ports
3059 to associate lp devices with, starting with
3060 lp0. A port specification may be 'none' to skip
3061 that lp device, or a parport name such as
3062 'parport0'. Specifying 'lp=auto' instead of a
3063 port specification list means that device IDs
3064 from each port should be examined, to see if
3065 an IEEE 1284-compliant printer is attached; if
3066 so, the driver will manage that printer.
3067 See also header of drivers/char/lp.c.
3070 Sets loops_per_jiffy to given constant, thus avoiding
3071 time-consuming boot-time autodetection (up to 250 ms per
3072 CPU). 0 enables autodetection (default). To determine
3073 the correct value for your kernel, boot with normal
3074 autodetection and see what value is printed. Note that
3075 on SMP systems the preset will be applied to all CPUs,
3076 which is likely to cause problems if your CPUs need
3077 significantly divergent settings. An incorrect value
3078 will cause delays in the kernel to be wrong, leading to
3079 unpredictable I/O errors and other breakage. Although
3080 unlikely, in the extreme case this might damage your
3084 Format: <io>,<irq>,<dma>
3086 lsm.debug [SECURITY] Enable LSM initialization debugging output.
3089 [SECURITY] Choose order of LSM initialization. This
3090 overrides CONFIG_LSM, and the "security=" parameter.
3092 machvec= [IA-64] Force the use of a particular machine-vector
3093 (machvec) in a generic kernel.
3094 Example: machvec=hpzx1
3096 machtype= [Loongson] Share the same kernel image file between
3097 different yeeloong laptops.
3098 Example: machtype=lemote-yeeloong-2f-7inch
3100 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
3101 than or equal to this physical address is ignored.
3103 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel
3104 will bring up during bootup. maxcpus=n : n >= 0 limits
3105 the kernel to bring up 'n' processors. Surely after
3106 bootup you can bring up the other plugged cpu by executing
3107 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
3108 only takes effect during system bootup.
3109 While n=0 is a special case, it is equivalent to "nosmp",
3110 which also disables the IO APIC.
3112 max_loop= [LOOP] The number of loop block devices that get
3113 (loop.max_loop) unconditionally pre-created at init time. The default
3114 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
3115 of statically allocating a predefined number, loop
3116 devices can be requested on-demand with the
3117 /dev/loop-control interface.
3119 mce [X86-32] Machine Check Exception
3121 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
3123 md= [HW] RAID subsystems devices and level
3124 See Documentation/admin-guide/md.rst.
3127 Format: <first>,<last>
3128 Specifies range of consoles to be captured by the MDA.
3130 mds= [X86,INTEL,EARLY]
3131 Control mitigation for the Micro-architectural Data
3132 Sampling (MDS) vulnerability.
3134 Certain CPUs are vulnerable to an exploit against CPU
3135 internal buffers which can forward information to a
3136 disclosure gadget under certain conditions.
3138 In vulnerable processors, the speculatively
3139 forwarded data can be used in a cache side channel
3140 attack, to access data to which the attacker does
3141 not have direct access.
3143 This parameter controls the MDS mitigation. The
3146 full - Enable MDS mitigation on vulnerable CPUs
3147 full,nosmt - Enable MDS mitigation and disable
3148 SMT on vulnerable CPUs
3149 off - Unconditionally disable MDS mitigation
3151 On TAA-affected machines, mds=off can be prevented by
3152 an active TAA mitigation as both vulnerabilities are
3153 mitigated with the same mechanism so in order to disable
3154 this mitigation, you need to specify tsx_async_abort=off
3157 Not specifying this option is equivalent to
3160 For details see: Documentation/admin-guide/hw-vuln/mds.rst
3162 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size.
3163 Must be specified, otherwise memory size will be 0.
3165 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount
3166 of memory Amount of memory to be used in cases
3170 2 when the kernel is not able to see the whole system memory;
3171 3 memory that lies after 'mem=' boundary is excluded from
3172 the hypervisor, then assigned to KVM guests.
3173 4 to limit the memory available for kdump kernel.
3175 [ARC,MICROBLAZE] - the limit applies only to low memory,
3176 high memory is not affected.
3178 [ARM64] - only limits memory covered by the linear
3179 mapping. The NOMAP regions are not affected.
3181 [X86] Work as limiting max address. Use together
3182 with memmap= to avoid physical address space collisions.
3183 Without memmap= PCI devices could be placed at addresses
3184 belonging to unused RAM.
3186 Note that this only takes effects during boot time since
3187 in above case 3, memory may need be hot added after boot
3188 if system memory of hypervisor is not sufficient.
3191 [ARM,MIPS,EARLY] - override the memory layout
3192 reported by firmware.
3193 Define a memory region of size nn[KMG] starting at
3195 Multiple different regions can be specified with
3196 multiple mem= parameters on the command line.
3198 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3201 memblock=debug [KNL,EARLY] Enable memblock debug messages.
3204 [KNL,SH] Allow user to override the default size for
3205 per-device physically contiguous DMA buffers.
3207 memhp_default_state=online/offline/online_kernel/online_movable
3208 [KNL] Set the initial state for the memory hotplug
3209 onlining policy. If not specified, the default value is
3210 set according to the
3211 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3213 See Documentation/admin-guide/mm/memory-hotplug.rst.
3215 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact
3216 E820 memory map, as specified by the user.
3217 Such memmap=exactmap lines can be constructed based on
3218 BIOS output or other requirements. See the memmap=nn@ss
3221 memmap=nn[KMG]@ss[KMG]
3222 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory.
3223 Region of memory to be used is from ss to ss+nn.
3224 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3225 which limits max address to nn[KMG].
3226 Multiple different regions can be specified,
3229 memmap=100M@2G,100M#3G,1G!1024G
3231 memmap=nn[KMG]#ss[KMG]
3232 [KNL,ACPI,EARLY] Mark specific memory as ACPI data.
3233 Region of memory to be marked is from ss to ss+nn.
3235 memmap=nn[KMG]$ss[KMG]
3236 [KNL,ACPI,EARLY] Mark specific memory as reserved.
3237 Region of memory to be reserved is from ss to ss+nn.
3238 Example: Exclude memory from 0x18690000-0x1869ffff
3239 memmap=64K$0x18690000
3241 memmap=0x10000$0x18690000
3242 Some bootloaders may need an escape character before '$',
3243 like Grub2, otherwise '$' and the following number
3246 memmap=nn[KMG]!ss[KMG,EARLY]
3247 [KNL,X86] Mark specific memory as protected.
3248 Region of memory to be used, from ss to ss+nn.
3249 The memory region may be marked as e820 type 12 (0xc)
3250 and is NVDIMM or ADR memory.
3252 memmap=<size>%<offset>-<oldtype>+<newtype>
3253 [KNL,ACPI,EARLY] Convert memory within the specified region
3254 from <oldtype> to <newtype>. If "-<oldtype>" is left
3255 out, the whole region will be marked as <newtype>,
3256 even if previously unavailable. If "+<newtype>" is left
3257 out, matching memory will be removed. Types are
3258 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3259 3 = ACPI, 12 = PRAM.
3261 memory_corruption_check=0/1 [X86,EARLY]
3262 Some BIOSes seem to corrupt the first 64k of
3263 memory when doing things like suspend/resume.
3264 Setting this option will scan the memory
3265 looking for corruption. Enabling this will
3266 both detect corruption and prevent the kernel
3267 from using the memory being corrupted.
3268 However, its intended as a diagnostic tool; if
3269 repeatable BIOS-originated corruption always
3270 affects the same memory, you can use memmap=
3271 to prevent the kernel from using that memory.
3273 memory_corruption_check_size=size [X86,EARLY]
3274 By default it checks for corruption in the low
3275 64k, making this memory unavailable for normal
3276 use. Use this parameter to scan for
3277 corruption in more or less memory.
3279 memory_corruption_check_period=seconds [X86,EARLY]
3280 By default it checks for corruption every 60
3281 seconds. Use this parameter to check at some
3282 other rate. 0 disables periodic checking.
3284 memory_hotplug.memmap_on_memory
3285 [KNL,X86,ARM] Boolean flag to enable this feature.
3286 Format: {on | off (default)}
3287 When enabled, runtime hotplugged memory will
3288 allocate its internal metadata (struct pages,
3289 those vmemmap pages cannot be optimized even
3290 if hugetlb_free_vmemmap is enabled) from the
3291 hotadded memory which will allow to hotadd a
3292 lot of memory without requiring additional
3294 This feature is disabled by default because it
3295 has some implication on large (e.g. GB)
3296 allocations in some configurations (e.g. small
3298 The state of the flag can be read in
3299 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3300 Note that even when enabled, there are a few cases where
3301 the feature is not effective.
3303 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest
3305 default : 0 <disable>
3306 Specifies the number of memtest passes to be
3307 performed. Each pass selects another test
3308 pattern from a given set of patterns. Memtest
3309 fills the memory with this pattern, validates
3310 memory contents and reserves bad memory
3311 regions that are detected.
3313 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3314 Valid arguments: on, off
3315 Default (depends on kernel configuration option):
3316 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3317 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3318 mem_encrypt=on: Activate SME
3319 mem_encrypt=off: Do not activate SME
3321 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3322 for details on when memory encryption can be activated.
3324 mem_sleep_default= [SUSPEND] Default system suspend mode:
3325 s2idle - Suspend-To-Idle
3326 shallow - Power-On Suspend or equivalent (if supported)
3327 deep - Suspend-To-RAM or equivalent (if supported)
3328 See Documentation/admin-guide/pm/sleep-states.rst.
3330 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3331 Multi-Function General Purpose Timers on AMD Geode
3334 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3335 the BIOS has incorrectly applied a workaround. TinyBIOS
3336 version 0.98 is known to be affected, 0.99 fixes the
3337 problem by letting the user disable the workaround.
3341 microcode.force_minrev= [X86]
3343 Enable or disable the microcode minimal revision
3344 enforcement for the runtime microcode loader.
3346 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3347 physical address is ignored.
3349 mini2440= [ARM,HW,KNL]
3350 Format:[0..2][b][c][t]
3352 MINI2440 configuration specification:
3353 0 - The attached screen is the 3.5" TFT
3354 1 - The attached screen is the 7" TFT
3355 2 - The VGA Shield is attached (1024x768)
3356 Leaving out the screen size parameter will not load
3357 the TFT driver, and the framebuffer will be left
3359 b - Enable backlight. The TFT backlight pin will be
3360 linked to the kernel VESA blanking code and a GPIO
3361 LED. This parameter is not necessary when using the
3363 c - Enable the s3c camera interface.
3364 t - Reserved for enabling touchscreen support. The
3365 touchscreen support is not enabled in the mainstream
3366 kernel as of 2.6.30, a preliminary port can be found
3367 in the "bleeding edge" mini2440 support kernel at
3368 https://repo.or.cz/w/linux-2.6/mini2440.git
3371 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for
3372 CPU vulnerabilities. This is a set of curated,
3373 arch-independent options, each of which is an
3374 aggregation of existing arch-specific options.
3377 Disable all optional CPU mitigations. This
3378 improves system performance, but it may also
3379 expose users to several CPU vulnerabilities.
3380 Equivalent to: if nokaslr then kpti=0 [ARM64]
3381 gather_data_sampling=off [X86]
3382 kvm.nx_huge_pages=off [X86]
3385 mmio_stale_data=off [X86]
3386 no_entry_flush [PPC]
3387 no_uaccess_flush [PPC]
3390 nospectre_bhb [ARM64]
3391 nospectre_v1 [X86,PPC]
3392 nospectre_v2 [X86,PPC,S390,ARM64]
3394 spec_store_bypass_disable=off [X86,PPC]
3395 spectre_v2_user=off [X86]
3396 srbds=off [X86,INTEL]
3397 ssbd=force-off [ARM64]
3398 tsx_async_abort=off [X86]
3401 This does not have any effect on
3402 kvm.nx_huge_pages when
3403 kvm.nx_huge_pages=force.
3406 Mitigate all CPU vulnerabilities, but leave SMT
3407 enabled, even if it's vulnerable. This is for
3408 users who don't want to be surprised by SMT
3409 getting disabled across kernel upgrades, or who
3410 have other ways of avoiding SMT-based attacks.
3411 Equivalent to: (default behavior)
3414 Mitigate all CPU vulnerabilities, disabling SMT
3415 if needed. This is for users who always want to
3416 be fully mitigated, even if it means losing SMT.
3417 Equivalent to: l1tf=flush,nosmt [X86]
3418 mds=full,nosmt [X86]
3419 tsx_async_abort=full,nosmt [X86]
3420 mmio_stale_data=full,nosmt [X86]
3421 retbleed=auto,nosmt [X86]
3424 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this
3425 parameter allows control of the logging verbosity for
3426 the additional memory initialisation checks. A value
3427 of 0 disables mminit logging and a level of 4 will
3428 log everything. Information is printed at KERN_DEBUG
3429 so loglevel=8 may also need to be specified.
3432 [X86,INTEL,EARLY] Control mitigation for the Processor
3433 MMIO Stale Data vulnerabilities.
3435 Processor MMIO Stale Data is a class of
3436 vulnerabilities that may expose data after an MMIO
3437 operation. Exposed data could originate or end in
3438 the same CPU buffers as affected by MDS and TAA.
3439 Therefore, similar to MDS and TAA, the mitigation
3440 is to clear the affected CPU buffers.
3442 This parameter controls the mitigation. The
3445 full - Enable mitigation on vulnerable CPUs
3447 full,nosmt - Enable mitigation and disable SMT on
3450 off - Unconditionally disable mitigation
3452 On MDS or TAA affected machines,
3453 mmio_stale_data=off can be prevented by an active
3454 MDS or TAA mitigation as these vulnerabilities are
3455 mitigated with the same mechanism so in order to
3456 disable this mitigation, you need to specify
3457 mds=off and tsx_async_abort=off too.
3459 Not specifying this option is equivalent to
3460 mmio_stale_data=full.
3463 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3465 <module>.async_probe[=<bool>] [KNL]
3466 If no <bool> value is specified or if the value
3467 specified is not a valid <bool>, enable asynchronous
3468 probe on this module. Otherwise, enable/disable
3469 asynchronous probe on this module as indicated by the
3470 <bool> value. See also: module.async_probe
3472 module.async_probe=<bool>
3473 [KNL] When set to true, modules will use async probing
3474 by default. To enable/disable async probing for a
3475 specific module, use the module specific control that
3476 is documented under <module>.async_probe. When both
3477 module.async_probe and <module>.async_probe are
3478 specified, <module>.async_probe takes precedence for
3479 the specific module.
3481 module.enable_dups_trace
3482 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set,
3483 this means that duplicate request_module() calls will
3484 trigger a WARN_ON() instead of a pr_warn(). Note that
3485 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s
3486 will always be issued and this option does nothing.
3488 [KNL] When CONFIG_MODULE_SIG is set, this means that
3489 modules without (valid) signatures will fail to load.
3490 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3491 is always true, so this option does nothing.
3493 module_blacklist= [KNL] Do not load a comma-separated list of
3494 modules. Useful for debugging problem modules.
3497 [MOUSE] Maximum time between finger touching and
3498 leaving touchpad surface for touch to be considered
3499 a tap and be reported as a left button click (for
3500 touchpads working in absolute mode only).
3502 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3503 reporting absolute coordinates, such as tablets
3504 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3505 reporting absolute coordinates, such as tablets
3507 movablecore= [KNL,X86,IA-64,PPC,EARLY]
3508 Format: nn[KMGTPE] | nn%
3509 This parameter is the complement to kernelcore=, it
3510 specifies the amount of memory used for migratable
3511 allocations. If both kernelcore and movablecore is
3512 specified, then kernelcore will be at *least* the
3513 specified value but may be more. If movablecore on its
3514 own is specified, the administrator must be careful
3515 that the amount of memory usable for all allocations
3518 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory
3519 NUMA nodes to be movable. This means that the memory
3520 of such nodes will be usable only for movable
3521 allocations which rules out almost all kernel
3522 allocations. Use with caution!
3524 MTD_Partition= [MTD]
3525 Format: <name>,<region-number>,<size>,<offset>
3527 MTD_Region= [MTD] Format:
3528 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3531 See drivers/mtd/parsers/cmdlinepart.c
3534 ARM/S3C2412 JIVE boot control
3536 See arch/arm/mach-s3c/mach-jive.c
3538 mtouchusb.raw_coordinates=
3539 [HW] Make the MicroTouch USB driver use raw coordinates
3540 ('y', default) or cooked coordinates ('n')
3542 mtrr=debug [X86,EARLY]
3543 Enable printing debug information related to MTRR
3544 registers at boot time.
3546 mtrr_chunk_size=nn[KMG,X86,EARLY]
3547 used for mtrr cleanup. It is largest continuous chunk
3548 that could hold holes aka. UC entries.
3550 mtrr_gran_size=nn[KMG,X86,EARLY]
3551 Used for mtrr cleanup. It is granularity of mtrr block.
3553 Large value could prevent small alignment from
3556 mtrr_spare_reg_nr=n [X86,EARLY]
3558 Range: 0,7 : spare reg number
3560 Used for mtrr cleanup. It is spare mtrr entries number.
3561 Set to 2 or more if your graphical card needs more.
3563 multitce=off [PPC] This parameter disables the use of the pSeries
3564 firmware feature for updating multiple TCE entries
3567 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3569 netdev= [NET] Network devices parameters
3570 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3571 Note that mem_start is often overloaded to mean
3572 something different and driver-specific.
3573 This usage is only documented in each driver source
3576 netpoll.carrier_timeout=
3577 [NET] Specifies amount of time (in seconds) that
3578 netpoll should wait for a carrier. By default netpoll
3582 [NETFILTER] Enable connection tracking flow accounting
3583 0 to disable accounting
3584 1 to enable accounting
3588 [NFS] sets the pathname to the program which is used
3589 to update the NFS client cache entries.
3591 nfs.cache_getent_timeout=
3592 [NFS] sets the timeout after which an attempt to
3593 update a cache entry is deemed to have failed.
3595 nfs.callback_nr_threads=
3596 [NFSv4] set the total number of threads that the
3597 NFS client will assign to service NFSv4 callback
3600 nfs.callback_tcpport=
3601 [NFS] set the TCP port on which the NFSv4 callback
3602 channel should listen.
3605 [NFS] specifies the number of times the NFSv4 client
3606 retries the request before returning an EAGAIN error,
3607 after a reply of NFS4ERR_DELAY from the server.
3608 Only applies if the softerr mount option is enabled,
3609 and the specified value is >= 0.
3612 [NFS] enable 64-bit inode numbers.
3613 If zero, the NFS client will fake up a 32-bit inode
3614 number for the readdir() and stat() syscalls instead
3615 of returning the full 64-bit number.
3616 The default is to return 64-bit inode numbers.
3618 nfs.idmap_cache_timeout=
3619 [NFS] set the maximum lifetime for idmapper cache
3622 nfs.max_session_cb_slots=
3623 [NFSv4.1] Sets the maximum number of session
3624 slots the client will assign to the callback
3625 channel. This determines the maximum number of
3626 callbacks the client will process in parallel for
3627 a particular server.
3629 nfs.max_session_slots=
3630 [NFSv4.1] Sets the maximum number of session slots
3631 the client will attempt to negotiate with the server.
3632 This limits the number of simultaneous RPC requests
3633 that the client can send to the NFSv4.1 server.
3634 Note that there is little point in setting this
3635 value higher than the max_tcp_slot_table_limit.
3637 nfs.nfs4_disable_idmapping=
3638 [NFSv4] When set to the default of '1', this option
3639 ensures that both the RPC level authentication
3640 scheme and the NFS level operations agree to use
3641 numeric uids/gids if the mount is using the
3642 'sec=sys' security flavour. In effect it is
3643 disabling idmapping, which can make migration from
3644 legacy NFSv2/v3 systems to NFSv4 easier.
3645 Servers that do not support this mode of operation
3646 will be autodetected by the client, and it will fall
3647 back to using the idmapper.
3648 To turn off this behaviour, set the value to '0'.
3651 [NFS4] Specify an additional fixed unique ident-
3652 ification string that NFSv4 clients can insert into
3653 their nfs_client_id4 string. This is typically a
3654 UUID that is generated at system install time.
3656 nfs.recover_lost_locks=
3657 [NFSv4] Attempt to recover locks that were lost due
3658 to a lease timeout on the server. Please note that
3659 doing this risks data corruption, since there are
3660 no guarantees that the file will remain unchanged
3661 after the locks are lost.
3662 If you want to enable the kernel legacy behaviour of
3663 attempting to recover these locks, then set this
3665 The default parameter value of '0' causes the kernel
3666 not to attempt recovery of lost locks.
3668 nfs.send_implementation_id=
3669 [NFSv4.1] Send client implementation identification
3670 information in exchange_id requests.
3671 If zero, no implementation identification information
3673 The default is to send the implementation identification
3676 nfs4.layoutstats_timer=
3677 [NFSv4.2] Change the rate at which the kernel sends
3678 layoutstats to the pNFS metadata server.
3680 Setting this to value to 0 causes the kernel to use
3681 whatever value is the default set by the layout
3682 driver. A non-zero value sets the minimum interval
3683 in seconds between layoutstats transmissions.
3685 nfsd.inter_copy_offload_enable=
3686 [NFSv4.2] When set to 1, the server will support
3687 server-to-server copies for which this server is
3688 the destination of the copy.
3690 nfsd.nfs4_disable_idmapping=
3691 [NFSv4] When set to the default of '1', the NFSv4
3692 server will return only numeric uids and gids to
3693 clients using auth_sys, and will accept numeric uids
3694 and gids from such clients. This is intended to ease
3695 migration from NFSv2/v3.
3697 nfsd.nfsd4_ssc_umount_timeout=
3698 [NFSv4.2] When used as the destination of a
3699 server-to-server copy, knfsd temporarily mounts
3700 the source server. It caches the mount in case
3701 it will be needed again, and discards it if not
3702 used for the number of milliseconds specified by
3705 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3706 See Documentation/admin-guide/nfs/nfsroot.rst.
3708 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3709 See Documentation/admin-guide/nfs/nfsroot.rst.
3711 nfsrootdebug [NFS] enable nfsroot debugging messages.
3712 See Documentation/admin-guide/nfs/nfsroot.rst.
3714 nmi_backtrace.backtrace_idle [KNL]
3715 Dump stacks even of idle CPUs in response to an
3716 NMI stack-backtrace request.
3718 nmi_debug= [KNL,SH] Specify one or more actions to take
3719 when a NMI is triggered.
3720 Format: [state][,regs][,debounce][,die]
3722 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3723 Format: [panic,][nopanic,][num]
3725 0 - turn hardlockup detector in nmi_watchdog off
3726 1 - turn hardlockup detector in nmi_watchdog on
3727 When panic is specified, panic when an NMI watchdog
3728 timeout occurs (or 'nopanic' to not panic on an NMI
3729 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3730 To disable both hard and soft lockup detectors,
3731 please see 'nowatchdog'.
3732 This is useful when you use a panic=... timeout and
3733 need the box quickly up again.
3735 These settings can be accessed at runtime via
3736 the nmi_watchdog and hardlockup_panic sysctls.
3738 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3739 emulation library even if a 387 maths coprocessor
3742 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes.
3743 Forces kernel to use 3-level paging instead.
3745 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces
3746 kernel to use 4-level paging instead.
3748 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3749 caches in the slab allocator. Saves per-node memory,
3750 but will impact performance.
3754 noaltinstr [S390,EARLY] Disables alternative instructions
3755 patching (CPU alternatives feature).
3757 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any
3758 IOAPICs that may be present in the system.
3760 noautogroup Disable scheduler automatic task group creation.
3765 [HW] Never suspend the console
3766 Disable suspending of consoles during suspend and
3767 hibernate operations. Once disabled, debugging
3768 messages can reach various consoles while the rest
3769 of the system is being put to sleep (ie, while
3770 debugging driver suspend/resume hooks). This may
3771 not work reliably with all consoles, but is known
3772 to work with serial and VGA consoles.
3773 To facilitate more flexible debugging, we also add
3774 console_suspend, a printk module parameter to control
3775 it. Users could use console_suspend (usually
3776 /sys/module/printk/parameters/console_suspend) to
3777 turn on/off it dynamically.
3780 [KNL,EARLY] Disable object debugging
3782 nodsp [SH] Disable hardware DSP at boot time.
3784 noefi [EFI,EARLY] Disable EFI runtime services support.
3786 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel.
3791 This affects only 32-bit executables.
3792 noexec32=on: enable non-executable mappings (default)
3793 read doesn't imply executable mappings
3794 noexec32=off: disable non-executable mappings
3795 read implies executable mappings
3797 no_file_caps Tells the kernel not to honor file capabilities. The
3798 only way then for a file to be executed with privilege
3799 is to be setuid root or executed by root.
3801 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3803 nofsgsbase [X86] Disables FSGSBASE instructions.
3805 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3806 register save and restore. The kernel will only save
3807 legacy floating-point registers on task switch.
3809 nohalt [IA-64] Tells the kernel not to use the power saving
3810 function PAL_HALT_LIGHT when idle. This increases
3811 power-consumption. On the positive side, it reduces
3812 interrupt wake-up latency, which may improve performance
3813 in certain environments such as networked servers or
3818 Force pointers printed to the console or buffers to be
3819 unhashed. By default, when a pointer is printed via %p
3820 format string, that pointer is "hashed", i.e. obscured
3821 by hashing the pointer value. This is a security feature
3822 that hides actual kernel addresses from unprivileged
3823 users, but it also makes debugging the kernel more
3824 difficult since unequal pointers can no longer be
3825 compared. However, if this command-line option is
3826 specified, then all normal pointers will have their true
3827 value printed. This option should only be specified when
3828 debugging the kernel. Please do not use on production
3831 nohibernate [HIBERNATION] Disable hibernation and resume.
3833 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,SH] Forces the kernel to
3834 busy wait in do_idle() and not use the arch_cpu_idle()
3835 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3836 to be effective. This is useful on platforms where the
3837 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3838 correctly or when doing power measurements to evaluate
3839 the impact of the sleep instructions. This is also
3840 useful when using JTAG debugger.
3842 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings.
3844 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings.
3846 nohz= [KNL] Boottime enable/disable dynamic ticks
3847 Valid arguments: on, off
3850 nohz_full= [KNL,BOOT,SMP,ISOL]
3851 The argument is a cpu list, as described above.
3852 In kernels built with CONFIG_NO_HZ_FULL=y, set
3853 the specified list of CPUs whose tick will be stopped
3854 whenever possible. The boot CPU will be forced outside
3855 the range to maintain the timekeeping. Any CPUs
3856 in this list will have their RCU callbacks offloaded,
3857 just as if they had also been called out in the
3858 rcu_nocbs= boot parameter.
3860 Note that this argument takes precedence over
3861 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3863 noinitrd [RAM] Tells the kernel not to load any configured
3866 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt
3868 [Deprecated - use intremap=off]
3872 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature.
3874 noiotrap [SH] Disables trapped I/O port accesses.
3876 noirqdebug [X86-32] Disables the code which attempts to detect and
3877 disable unhandled interrupt sources.
3879 noisapnp [ISAPNP] Disables ISA PnP code.
3881 nojitter [IA-64] Disables jitter checking for ITC timers.
3884 When CONFIG_RANDOMIZE_BASE is set, this disables
3885 kernel and module base offset ASLR (Address Space
3886 Layout Randomization).
3888 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page
3891 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver
3893 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC.
3895 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer.
3897 nomca [IA-64] Disable machine check abort handling
3899 nomce [X86-32] Disable Machine Check Exception
3901 nomfgpt [X86-32] Disable Multi-Function General Purpose
3902 Timer usage (for AMD Geode machines).
3904 nomodeset Disable kernel modesetting. Most systems' firmware
3905 sets up a display mode and provides framebuffer memory
3906 for output. With nomodeset, DRM and fbdev drivers will
3907 not load if they could possibly displace the pre-
3908 initialized output. Only the system framebuffer will
3909 be available for use. The respective drivers will not
3910 perform display-mode changes or accelerated rendering.
3912 Useful as error fallback, or for testing and debugging.
3914 nomodule Disable module load
3916 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3917 shutdown the other cpus. Instead use the REBOOT_VECTOR
3920 nopat [X86,EARLY] Disable PAT (page attribute table extension of
3921 pagetables) support.
3923 nopcid [X86-64,EARLY] Disable the PCID cpu feature.
3925 nopku [X86] Disable Memory Protection Keys CPU feature found
3928 nopti [X86-64,EARLY]
3929 Equivalent to pti=off
3931 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY]
3932 Disables the PV optimizations forcing the guest to run
3933 as generic guest with no PV drivers. Currently support
3934 XEN HVM, KVM, HYPER_V and VMWARE guest.
3936 nopvspin [X86,XEN,KVM,EARLY]
3937 Disables the qspinlock slow path using PV optimizations
3938 which allow the hypervisor to 'idle' the guest on lock
3941 norandmaps Don't use address space randomization. Equivalent to
3942 echo 0 > /proc/sys/kernel/randomize_va_space
3944 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3945 with UP alternatives
3947 noresume [SWSUSP] Disables resume and restores original swap
3952 no-scroll [VGA] Disables scrollback.
3953 This is required for the Braillex ib80-piezo Braille
3954 reader made by F.H. Papenmeier (Germany).
3956 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support.
3959 Disable SMAP (Supervisor Mode Access Prevention)
3960 even if it is supported by processor.
3962 nosmep [PPC64s,EARLY]
3963 Disable SMEP (Supervisor Mode Execution Prevention)
3964 even if it is supported by processor.
3966 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel,
3967 and disable the IO APIC. legacy for "maxcpus=0".
3969 nosmt [KNL,MIPS,PPC,S390,EARLY] Disable symmetric multithreading (SMT).
3970 Equivalent to smt=1.
3972 [KNL,X86,PPC] Disable symmetric multithreading (SMT).
3973 nosmt=force: Force disable SMT, cannot be undone
3974 via the sysfs control file.
3976 nosoftlockup [KNL] Disable the soft-lockup detector.
3978 nospec_store_bypass_disable
3979 [HW,EARLY] Disable all mitigations for the Speculative
3980 Store Bypass vulnerability
3982 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch
3983 history injection) vulnerability. System may allow data leaks
3986 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1
3987 (bounds check bypass). With this option data leaks are
3988 possible in the system.
3990 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations
3991 for the Spectre variant 2 (indirect branch
3992 prediction) vulnerability. System may allow data
3993 leaks with this option.
3995 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,EARLY] Disable
3996 paravirtualized steal time accounting. steal time is
3997 computed, but won't influence scheduler behaviour
3999 nosync [HW,M68K] Disables sync negotiation for all devices.
4001 no_timer_check [X86,APIC] Disables the code which tests for
4002 broken timer IRQ sources.
4005 [PPC,EARLY] Don't flush the L1-D cache after accessing user data.
4007 novmcoredd [KNL,KDUMP]
4008 Disable device dump. Device dump allows drivers to
4009 append dump data to vmcore so you can collect driver
4010 specified debug info. Drivers can append the data
4011 without any limit and this data is stored in memory,
4012 so this may cause significant memory stress. Disabling
4013 device dump can help save memory but the driver debug
4014 data will be no longer available. This parameter
4015 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
4019 [X86,PV_OPS,EARLY] Disable paravirtualized VMware
4020 scheduler clock and use the default one.
4022 nowatchdog [KNL] Disable both lockup detectors, i.e.
4023 soft-lockup and NMI watchdog (hard-lockup).
4027 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode.
4029 NOTE: this parameter will be ignored on systems with the
4030 LEGACY_XAPIC_DISABLED bit set in the
4031 IA32_XAPIC_DISABLE_STATUS MSR.
4033 noxsave [BUGS=X86] Disables x86 extended register state save
4034 and restore using xsave. The kernel will fallback to
4035 enabling legacy floating-point and sse state.
4037 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
4038 register states. The kernel will fall back to use
4039 xsave to save the states. By using this parameter,
4040 performance of saving the states is degraded because
4041 xsave doesn't support modified optimization while
4042 xsaveopt supports it on xsaveopt enabled systems.
4044 noxsaves [X86] Disables xsaves and xrstors used in saving and
4045 restoring x86 extended register state in compacted
4046 form of xsave area. The kernel will fall back to use
4047 xsaveopt and xrstor to save and restore the states
4048 in standard form of xsave area. By using this
4049 parameter, xsave area per process might occupy more
4050 memory on xsaves enabled systems.
4052 nps_mtm_hs_ctr= [KNL,ARC]
4053 This parameter sets the maximum duration, in
4054 cycles, each HW thread of the CTOP can run
4055 without interruptions, before HW switches it.
4056 The actual maximum duration is 16 times this
4058 Format: integer between 1 and 255
4061 nptcg= [IA-64] Override max number of concurrent global TLB
4062 purges which is reported from either PAL_VM_SUMMARY or
4065 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel
4066 could support. nr_cpus=n : n >= 1 limits the kernel to
4067 support 'n' processors. It could be larger than the
4068 number of already plugged CPU during bootup, later in
4069 runtime you can physically add extra cpu until it reaches
4070 n. So during boot up some boot time memory for per-cpu
4071 variables need be pre-allocated for later physical cpu
4074 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
4076 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY]
4077 Disable NUMA, Only set up a single NUMA node
4078 spanning all memory.
4080 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
4082 Allowed values are enable and disable
4084 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
4085 'node', 'default' can be specified
4086 This can be set from sysctl after boot.
4087 See Documentation/admin-guide/sysctl/vm.rst for details.
4089 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver.
4090 See Documentation/core-api/debugging-via-ohci1394.rst for more
4093 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
4094 Rather than timing out after 20 ms if an EC
4095 command is not properly ACKed, override the length
4096 of the timeout. We have interrupts disabled while
4097 waiting for the ACK, so if this is set too high
4098 interrupts *may* be lost!
4100 omap_mux= [OMAP] Override bootloader pin multiplexing.
4101 Format: <mux_mode0.mode_name=value>...
4102 For example, to override I2C bus2:
4103 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
4105 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
4107 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
4109 boundary - index of last SLC block on Flex-OneNAND.
4110 The remaining blocks are configured as MLC blocks.
4111 lock - Configure if Flex-OneNAND boundary should be locked.
4112 Once locked, the boundary cannot be changed.
4113 1 indicates lock status, 0 indicates unlock status.
4115 oops=panic [KNL,EARLY]
4116 Always panic on oopses. Default is to just kill the
4117 process, but there is a small probability of
4118 deadlocking the machine.
4119 This will also cause panics on machine check exceptions.
4120 Useful together with panic=30 to trigger a reboot.
4123 [KNL] Boolean flag to control whether the page allocator
4124 should randomize its free lists. The randomization may
4125 be automatically enabled if the kernel detects it is
4126 running on a platform with a direct-mapped memory-side
4127 cache, and this parameter can be used to
4128 override/disable that behavior. The state of the flag
4129 can be read from sysfs at:
4130 /sys/module/page_alloc/parameters/shuffle.
4132 page_owner= [KNL,EARLY] Boot-time page_owner enabling option.
4133 Storage of the information about who allocated
4134 each page is disabled in default. With this switch,
4136 on: enable the feature
4138 page_poison= [KNL,EARLY] Boot-time parameter changing the state of
4139 poisoning on the buddy allocator, available with
4140 CONFIG_PAGE_POISONING=y.
4141 off: turn off poisoning (default)
4142 on: turn on poisoning
4144 page_reporting.page_reporting_order=
4145 [KNL] Minimal page reporting order
4147 Adjust the minimal page reporting order. The page
4148 reporting is disabled when it exceeds MAX_PAGE_ORDER.
4150 panic= [KNL] Kernel behaviour on panic: delay <timeout>
4151 timeout > 0: seconds before rebooting
4152 timeout = 0: wait forever
4153 timeout < 0: reboot immediately
4156 panic_on_taint= [KNL,EARLY]
4157 Bitmask for conditionally calling panic() in add_taint()
4158 Format: <hex>[,nousertaint]
4159 Hexadecimal bitmask representing the set of TAINT flags
4160 that will cause the kernel to panic when add_taint() is
4161 called with any of the flags in this set.
4162 The optional switch "nousertaint" can be utilized to
4163 prevent userspace forced crashes by writing to sysctl
4164 /proc/sys/kernel/tainted any flagset matching with the
4165 bitmask set on panic_on_taint.
4166 See Documentation/admin-guide/tainted-kernels.rst for
4167 extra details on the taint flags that users can pick
4168 to compose the bitmask to assign to panic_on_taint.
4170 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump
4173 panic_print= Bitmask for printing system info when panic happens.
4174 User can chose combination of the following bits:
4175 bit 0: print all tasks info
4176 bit 1: print system memory info
4177 bit 2: print timer info
4178 bit 3: print locks info if CONFIG_LOCKDEP is on
4179 bit 4: print ftrace buffer
4180 bit 5: print all printk messages in buffer
4181 bit 6: print all CPUs backtrace (if available in the arch)
4182 *Be aware* that this option may print a _lot_ of lines,
4183 so there are risks of losing older messages in the log.
4184 Use this option carefully, maybe worth to setup a
4185 bigger log buffer with "log_buf_len" along with this.
4187 parkbd.port= [HW] Parallel port number the keyboard adapter is
4188 connected to, default is 0.
4190 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
4191 0 for XT, 1 for AT (default is AT).
4194 parport= [HW,PPT] Specify parallel ports. 0 disables.
4195 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
4196 Use 'auto' to force the driver to use any
4197 IRQ/DMA settings detected (the default is to
4198 ignore detected IRQ/DMA settings because of
4199 possible conflicts). You can specify the base
4200 address, IRQ, and DMA settings; IRQ and DMA
4201 should be numbers, or 'auto' (for using detected
4202 settings on that particular port), or 'nofifo'
4203 (to avoid using a FIFO even if it is detected).
4204 Parallel ports are assigned in the order they
4205 are specified on the command line, starting
4208 parport_init_mode= [HW,PPT]
4209 Configure VIA parallel port to operate in
4210 a specific mode. This is necessary on Pegasos
4211 computer where firmware has no options for setting
4212 up parallel port mode and sets it to spp.
4213 Currently this function knows 686a and 8231 chips.
4214 Format: [spp|ps2|epp|ecp|ecpepp]
4216 pata_legacy.all= [HW,LIBATA]
4218 Set to non-zero to probe primary and secondary ISA
4219 port ranges on PCI systems where no PCI PATA device
4220 has been found at either range. Disabled by default.
4222 pata_legacy.autospeed= [HW,LIBATA]
4224 Set to non-zero if a chip is present that snoops speed
4225 changes. Disabled by default.
4227 pata_legacy.ht6560a= [HW,LIBATA]
4229 Set to 1, 2, or 3 for HT 6560A on the primary channel,
4230 the secondary channel, or both channels respectively.
4231 Disabled by default.
4233 pata_legacy.ht6560b= [HW,LIBATA]
4235 Set to 1, 2, or 3 for HT 6560B on the primary channel,
4236 the secondary channel, or both channels respectively.
4237 Disabled by default.
4239 pata_legacy.iordy_mask= [HW,LIBATA]
4241 IORDY enable mask. Set individual bits to allow IORDY
4242 for the respective channel. Bit 0 is for the first
4243 legacy channel handled by this driver, bit 1 is for
4244 the second channel, and so on. The sequence will often
4245 correspond to the primary legacy channel, the secondary
4246 legacy channel, and so on, but the handling of a PCI
4247 bus and the use of other driver options may interfere
4248 with the sequence. By default IORDY is allowed across
4251 pata_legacy.opti82c46x= [HW,LIBATA]
4253 Set to 1, 2, or 3 for Opti 82c611A on the primary
4254 channel, the secondary channel, or both channels
4255 respectively. Disabled by default.
4257 pata_legacy.opti82c611a= [HW,LIBATA]
4259 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4260 channel, the secondary channel, or both channels
4261 respectively. Disabled by default.
4263 pata_legacy.pio_mask= [HW,LIBATA]
4265 PIO mode mask for autospeed devices. Set individual
4266 bits to allow the use of the respective PIO modes.
4267 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4268 All modes allowed by default.
4270 pata_legacy.probe_all= [HW,LIBATA]
4272 Set to non-zero to probe tertiary and further ISA
4273 port ranges on PCI systems. Disabled by default.
4275 pata_legacy.probe_mask= [HW,LIBATA]
4277 Probe mask for legacy ISA PATA ports. Depending on
4278 platform configuration and the use of other driver
4279 options up to 6 legacy ports are supported: 0x1f0,
4280 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4281 of individual ports can be disabled by setting the
4282 corresponding bits in the mask to 1. Bit 0 is for
4283 the first port in the list above (0x1f0), and so on.
4284 By default all supported ports are probed.
4286 pata_legacy.qdi= [HW,LIBATA]
4288 Set to non-zero to probe QDI controllers. By default
4289 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4291 pata_legacy.winbond= [HW,LIBATA]
4293 Set to non-zero to probe Winbond controllers. Use
4294 the standard I/O port (0x130) if 1, otherwise the
4295 value given is the I/O port to use (typically 0x1b0).
4296 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4299 pata_platform.pio_mask= [HW,LIBATA]
4301 Supported PIO mode mask. Set individual bits to allow
4302 the use of the respective PIO modes. Bit 0 is for
4303 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4307 Halt all CPUs after the first oops has been printed for
4308 the specified number of seconds. This is to be used if
4309 your oopses keep scrolling off the screen.
4313 pci=option[,option...] [PCI,EARLY] various PCI subsystem options.
4315 Some options herein operate on a specific device
4316 or a set of devices (<pci_dev>). These are
4317 specified in one of the following formats:
4319 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4320 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4322 Note: the first format specifies a PCI
4323 bus/device/function address which may change
4324 if new hardware is inserted, if motherboard
4325 firmware changes, or due to changes caused
4326 by other kernel parameters. If the
4327 domain is left unspecified, it is
4328 taken to be zero. Optionally, a path
4329 to a device through multiple device/function
4330 addresses can be specified after the base
4331 address (this is more robust against
4332 renumbering issues). The second format
4333 selects devices using IDs from the
4334 configuration space which may match multiple
4335 devices in the system.
4337 earlydump dump PCI config space before the kernel
4339 off [X86] don't probe for the PCI bus
4340 bios [X86-32] force use of PCI BIOS, don't access
4341 the hardware directly. Use this if your machine
4342 has a non-standard PCI host bridge.
4343 nobios [X86-32] disallow use of PCI BIOS, only direct
4344 hardware access methods are allowed. Use this
4345 if you experience crashes upon bootup and you
4346 suspect they are caused by the BIOS.
4347 conf1 [X86] Force use of PCI Configuration Access
4348 Mechanism 1 (config address in IO port 0xCF8,
4349 data in IO port 0xCFC, both 32-bit).
4350 conf2 [X86] Force use of PCI Configuration Access
4351 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4352 the function, IO port 0xCFA, also 8-bit, sets
4353 bus number. The config space is then accessed
4354 through ports 0xC000-0xCFFF).
4355 See http://wiki.osdev.org/PCI for more info
4356 on the configuration access mechanisms.
4357 noaer [PCIE] If the PCIEAER kernel config parameter is
4358 enabled, this kernel boot option can be used to
4359 disable the use of PCIE advanced error reporting.
4360 nodomains [PCI] Disable support for multiple PCI
4361 root domains (aka PCI segments, in ACPI-speak).
4362 nommconf [X86] Disable use of MMCONFIG for PCI
4364 check_enable_amd_mmconf [X86] check for and enable
4365 properly configured MMIO access to PCI
4366 config space on AMD family 10h CPU
4367 nomsi [MSI] If the PCI_MSI kernel config parameter is
4368 enabled, this kernel boot option can be used to
4369 disable the use of MSI interrupts system-wide.
4370 noioapicquirk [APIC] Disable all boot interrupt quirks.
4371 Safety option to keep boot IRQs enabled. This
4372 should never be necessary.
4373 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4374 primary IO-APIC for bridges that cannot disable
4375 boot IRQs. This fixes a source of spurious IRQs
4376 when the system masks IRQs.
4377 noioapicreroute [APIC] Disable workaround that uses the
4378 boot IRQ equivalent of an IRQ that connects to
4379 a chipset where boot IRQs cannot be disabled.
4380 The opposite of ioapicreroute.
4381 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4382 routing table. These calls are known to be buggy
4383 on several machines and they hang the machine
4384 when used, but on other computers it's the only
4385 way to get the interrupt routing table. Try
4386 this option if the kernel is unable to allocate
4387 IRQs or discover secondary PCI buses on your
4389 rom [X86] Assign address space to expansion ROMs.
4390 Use with caution as certain devices share
4391 address decoders between ROMs and other
4393 norom [X86] Do not assign address space to
4394 expansion ROMs that do not already have
4395 BIOS assigned address ranges.
4396 nobar [X86] Do not assign address space to the
4397 BARs that weren't assigned by the BIOS.
4398 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4399 assigned automatically to PCI devices. You can
4400 make the kernel exclude IRQs of your ISA cards
4402 pirqaddr=0xAAAAA [X86] Specify the physical address
4403 of the PIRQ table (normally generated
4404 by the BIOS) if it is outside the
4405 F0000h-100000h range.
4406 lastbus=N [X86] Scan all buses thru bus #N. Can be
4407 useful if the kernel is unable to find your
4408 secondary buses and you want to tell it
4409 explicitly which ones they are.
4410 assign-busses [X86] Always assign all PCI bus
4411 numbers ourselves, overriding
4412 whatever the firmware may have done.
4413 usepirqmask [X86] Honor the possible IRQ mask stored
4414 in the BIOS $PIR table. This is needed on
4415 some systems with broken BIOSes, notably
4416 some HP Pavilion N5400 and Omnibook XE3
4417 notebooks. This will have no effect if ACPI
4418 IRQ routing is enabled.
4419 noacpi [X86] Do not use ACPI for IRQ routing
4420 or for PCI scanning.
4421 use_crs [X86] Use PCI host bridge window information
4422 from ACPI. On BIOSes from 2008 or later, this
4423 is enabled by default. If you need to use this,
4424 please report a bug.
4425 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4426 If you need to use this, please report a bug.
4427 use_e820 [X86] Use E820 reservations to exclude parts of
4428 PCI host bridge windows. This is a workaround
4429 for BIOS defects in host bridge _CRS methods.
4430 If you need to use this, please report a bug to
4431 <linux-pci@vger.kernel.org>.
4432 no_e820 [X86] Ignore E820 reservations for PCI host
4433 bridge windows. This is the default on modern
4434 hardware. If you need to use this, please report
4435 a bug to <linux-pci@vger.kernel.org>.
4436 routeirq Do IRQ routing for all PCI devices.
4437 This is normally done in pci_enable_device(),
4438 so this option is a temporary workaround
4439 for broken drivers that don't call it.
4440 skip_isa_align [X86] do not align io start addr, so can
4441 handle more pci cards
4442 noearly [X86] Don't do any early type 1 scanning.
4443 This might help on some broken boards which
4444 machine check when some devices' config space
4445 is read. But various workarounds are disabled
4446 and some IOMMU drivers will not work.
4447 bfsort Sort PCI devices into breadth-first order.
4448 This sorting is done to get a device
4449 order compatible with older (<= 2.4) kernels.
4450 nobfsort Don't sort PCI devices into breadth-first order.
4451 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4452 tuning and use the BIOS-configured MPS defaults.
4453 pcie_bus_safe Set every device's MPS to the largest value
4454 supported by all devices below the root complex.
4455 pcie_bus_perf Set device MPS to the largest allowable MPS
4456 based on its parent bus. Also set MRRS (Max
4457 Read Request Size) to the largest supported
4458 value (no larger than the MPS that the device
4459 or bus can support) for best performance.
4460 pcie_bus_peer2peer Set every device's MPS to 128B, which
4461 every device is guaranteed to support. This
4462 configuration allows peer-to-peer DMA between
4463 any pair of devices, possibly at the cost of
4464 reduced performance. This also guarantees
4465 that hot-added devices will work.
4466 cbiosize=nn[KMG] The fixed amount of bus space which is
4467 reserved for the CardBus bridge's IO window.
4468 The default value is 256 bytes.
4469 cbmemsize=nn[KMG] The fixed amount of bus space which is
4470 reserved for the CardBus bridge's memory
4471 window. The default value is 64 megabytes.
4474 [<order of align>@]<pci_dev>[; ...]
4475 Specifies alignment and device to reassign
4476 aligned memory resources. How to
4477 specify the device is described above.
4478 If <order of align> is not specified,
4479 PAGE_SIZE is used as alignment.
4480 A PCI-PCI bridge can be specified if resource
4481 windows need to be expanded.
4482 To specify the alignment for several
4483 instances of a device, the PCI vendor,
4484 device, subvendor, and subdevice may be
4485 specified, e.g., 12@pci:8086:9c22:103c:198f
4486 for 4096-byte alignment.
4487 ecrc= Enable/disable PCIe ECRC (transaction layer
4488 end-to-end CRC checking). Only effective if
4489 OS has native AER control (either granted by
4490 ACPI _OSC or forced via "pcie_ports=native")
4491 bios: Use BIOS/firmware settings. This is the
4495 hpiosize=nn[KMG] The fixed amount of bus space which is
4496 reserved for hotplug bridge's IO window.
4497 Default size is 256 bytes.
4498 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4499 reserved for hotplug bridge's MMIO window.
4500 Default size is 2 megabytes.
4501 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4502 reserved for hotplug bridge's MMIO_PREF window.
4503 Default size is 2 megabytes.
4504 hpmemsize=nn[KMG] The fixed amount of bus space which is
4505 reserved for hotplug bridge's MMIO and
4507 Default size is 2 megabytes.
4508 hpbussize=nn The minimum amount of additional bus numbers
4509 reserved for buses below a hotplug bridge.
4511 realloc= Enable/disable reallocating PCI bridge resources
4512 if allocations done by BIOS are too small to
4513 accommodate resources required by all child
4515 off: Turn realloc off
4517 realloc same as realloc=on
4518 noari do not use PCIe ARI.
4519 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4520 do not use PCIe ATS (and IOMMU device IOTLB).
4521 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4522 only look for one device below a PCIe downstream
4524 big_root_window Try to add a big 64bit memory window to the PCIe
4525 root complex on AMD CPUs. Some GFX hardware
4526 can resize a BAR to allow access to all VRAM.
4527 Adding the window is slightly risky (it may
4528 conflict with unreported devices), so this
4530 disable_acs_redir=<pci_dev>[; ...]
4531 Specify one or more PCI devices (in the format
4532 specified above) separated by semicolons.
4533 Each device specified will have the PCI ACS
4534 redirect capabilities forced off which will
4535 allow P2P traffic between devices through
4536 bridges without forcing it upstream. Note:
4537 this removes isolation between devices and
4538 may put more devices in an IOMMU group.
4539 force_floating [S390] Force usage of floating interrupts.
4540 nomio [S390] Do not use MIO instructions.
4541 norid [S390] ignore the RID field and force use of
4542 one PCI domain per PCI function
4544 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4547 force Enable ASPM even on devices that claim not to support it.
4548 WARNING: Forcing ASPM on may cause system lockups.
4550 pcie_ports= [PCIE] PCIe port services handling:
4551 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4552 even if the platform doesn't give the OS permission to
4553 use them. This may cause conflicts if the platform
4554 also tries to use these services.
4555 dpc-native Use native PCIe service for DPC only. May
4556 cause conflicts if firmware uses AER or DPC.
4557 compat Disable native PCIe services (PME, AER, DPC, PCIe
4560 pcie_port_pm= [PCIE] PCIe port power management handling:
4561 off Disable power management of all PCIe ports
4562 force Forcibly enable power management of all PCIe ports
4564 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4565 nomsi Do not use MSI for native PCIe PME signaling (this makes
4566 all PCIe root ports use INTx for all services).
4568 pcmv= [HW,PCMCIA] BadgePAD 4
4572 Keep all power-domains already enabled by bootloader on,
4573 even if no driver has claimed them. This is useful
4574 for debug and development, but should not be
4575 needed on a platform with proper driver support.
4577 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4580 See arch/parisc/kernel/pdc_chassis.c
4582 percpu_alloc= [MM,EARLY]
4583 Select which percpu first chunk allocator to use.
4584 Currently supported values are "embed" and "page".
4585 Archs may support subset or none of the selections.
4586 See comments in mm/percpu.c for details on each
4587 allocator. This parameter is primarily for debugging
4588 and performance comparison.
4590 pirq= [SMP,APIC] Manual mp-table setup
4591 See Documentation/arch/x86/i386/IO-APIC.rst.
4593 plip= [PPT,NET] Parallel port network link
4594 Format: { parport<nr> | timid | 0 }
4595 See also Documentation/admin-guide/parport.rst.
4597 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4598 Override pmtimer IOPort with a hex value.
4601 pmu_override= [PPC] Override the PMU.
4602 This option takes over the PMU facility, so it is no
4603 longer usable by perf. Setting this option starts the
4604 PMU counters by setting MMCR0 to 0 (the FC bit is
4605 cleared). If a number is given, then MMCR1 is set to
4606 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4609 pm_debug_messages [SUSPEND,KNL]
4610 Enable suspend/resume debug messages during boot up.
4613 Enable PNP debug messages (depends on the
4614 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4615 via /sys/module/pnp/parameters/debug. We always show
4616 current resource usage; turning this on also shows
4617 possible settings and some assignment information.
4623 { on | off | curr | res | no-curr | no-res }
4626 [ISAPNP] Exclude IRQs for the autoconfiguration
4629 [ISAPNP] Exclude DMAs for the autoconfiguration
4631 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4632 Ranges are in pairs (I/O port base and size).
4635 [ISAPNP] Exclude memory regions for the
4637 Ranges are in pairs (memory base and size).
4639 ports= [IP_VS_FTP] IPVS ftp helper module
4641 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4643 Format: <port>,<port>....
4645 powersave=off [PPC] This option disables power saving features.
4646 It specifically disables cpuidle and sets the
4647 platform machine description specific power_save
4648 function to NULL. On Idle the CPU just reduces
4651 ppc_strict_facility_enable
4652 [PPC,ENABLE] This option catches any kernel floating point,
4653 Altivec, VSX and SPE outside of regions specifically
4654 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4655 There is some performance impact when enabling this.
4659 Disable Hardware Transactional Memory
4662 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4663 none - Limited to cond_resched() calls
4664 voluntary - Limited to cond_resched() and might_sleep() calls
4665 full - Any section that isn't explicitly preempt disabled
4666 can be preempted anytime.
4668 print-fatal-signals=
4669 [KNL] debug: print fatal signals
4671 If enabled, warn about various signal handling
4672 related application anomalies: too many signals,
4673 too many POSIX.1 timers, fatal signals causing a
4676 If you hit the warning due to signal overflow,
4677 you might want to try "ulimit -i unlimited".
4681 printk.always_kmsg_dump=
4682 Trigger kmsg_dump for cases other than kernel oops or
4684 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4687 printk.console_no_auto_verbose=
4688 Disable console loglevel raise on oops, panic
4689 or lockdep-detected issues (only if lock debug is on).
4690 With an exception to setups with low baudrate on
4691 serial console, keeping this 0 is a good choice
4692 in order to provide more debug information.
4694 default: 0 (auto_verbose is enabled)
4696 printk.devkmsg={on,off,ratelimit}
4697 Control writing to /dev/kmsg.
4698 on - unlimited logging to /dev/kmsg from userspace
4699 off - logging to /dev/kmsg disabled
4700 ratelimit - ratelimit the logging
4703 printk.time= Show timing data prefixed to each printk message line
4704 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4706 processor.max_cstate= [HW,ACPI]
4707 Limit processor to maximum C-state
4708 max_cstate=9 overrides any DMI blacklist limit.
4710 processor.nocst [HW,ACPI]
4711 Ignore the _CST method to determine C-states,
4712 instead using the legacy FADT method
4714 profile= [KNL] Enable kernel profiling via /proc/profile
4715 Format: [<profiletype>,]<number>
4716 Param: <profiletype>: "schedule", "sleep", or "kvm"
4717 [defaults to kernel profiling]
4718 Param: "schedule" - profile schedule points.
4719 Param: "sleep" - profile D-state sleeping (millisecs).
4720 Requires CONFIG_SCHEDSTATS
4721 Param: "kvm" - profile VM exits.
4722 Param: <number> - step/bucket size as a power of 2 for
4723 statistical time based profiling.
4725 prompt_ramdisk= [RAM] [Deprecated]
4727 prot_virt= [S390] enable hosting protected virtual machines
4728 isolated from the hypervisor (if hardware supports
4732 psi= [KNL] Enable or disable pressure stall information
4736 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4737 probe for; one of (bare|imps|exps|lifebook|any).
4738 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4740 psmouse.resetafter= [HW,MOUSE]
4741 Try to reset the device after so many bad packets
4744 [HW,MOUSE] Set desired mouse resolution, in dpi.
4745 psmouse.smartscroll=
4746 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4747 0 = disabled, 1 = enabled (default).
4749 pstore.backend= Specify the name of the pstore backend to use
4751 pti= [X86-64] Control Page Table Isolation of user and
4752 kernel address spaces. Disabling this feature
4753 removes hardening, but improves performance of
4754 system calls and interrupts.
4756 on - unconditionally enable
4757 off - unconditionally disable
4758 auto - kernel detects whether your CPU model is
4759 vulnerable to issues that PTI mitigates
4761 Not specifying this option is equivalent to pti=auto.
4764 [KNL] Number of legacy pty's. Overwrites compiled-in
4767 quiet [KNL,EARLY] Disable most log messages
4771 radix_hcall_invalidate=on [PPC/PSERIES]
4772 Disable RADIX GTSE feature and use hcall for TLB
4776 See Documentation/admin-guide/md.rst.
4778 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4779 See Documentation/admin-guide/blockdev/ramdisk.rst.
4781 ramdisk_start= [RAM] RAM disk image start address
4783 random.trust_cpu=off
4784 [KNL,EARLY] Disable trusting the use of the CPU's
4785 random number generator (if available) to
4786 initialize the kernel's RNG.
4788 random.trust_bootloader=off
4789 [KNL,EARLY] Disable trusting the use of the a seed
4790 passed by the bootloader (if available) to
4791 initialize the kernel's RNG.
4793 randomize_kstack_offset=
4794 [KNL,EARLY] Enable or disable kernel stack offset
4795 randomization, which provides roughly 5 bits of
4796 entropy, frustrating memory corruption attacks
4797 that depend on stack address determinism or
4798 cross-syscall address exposures. This is only
4799 available on architectures that have defined
4800 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4801 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4802 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4804 ras=option[,option,...] [KNL] RAS-specific options
4807 Disable the Correctable Errors Collector,
4808 see CONFIG_RAS_CEC help text.
4810 rcu_nocbs[=cpu-list]
4811 [KNL] The optional argument is a cpu list,
4814 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4815 enable the no-callback CPU mode, which prevents
4816 such CPUs' callbacks from being invoked in
4817 softirq context. Invocation of such CPUs' RCU
4818 callbacks will instead be offloaded to "rcuox/N"
4819 kthreads created for that purpose, where "x" is
4820 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4821 for the kthreads that mediate grace periods; and
4822 "N" is the CPU number. This reduces OS jitter on
4823 the offloaded CPUs, which can be useful for HPC
4824 and real-time workloads. It can also improve
4825 energy efficiency for asymmetric multiprocessors.
4827 If a cpulist is passed as an argument, the specified
4828 list of CPUs is set to no-callback mode from boot.
4830 Otherwise, if the '=' sign and the cpulist
4831 arguments are omitted, no CPU will be set to
4832 no-callback mode from boot but the mode may be
4833 toggled at runtime via cpusets.
4835 Note that this argument takes precedence over
4836 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4839 Rather than requiring that offloaded CPUs
4840 (specified by rcu_nocbs= above) explicitly
4841 awaken the corresponding "rcuoN" kthreads,
4842 make these kthreads poll for callbacks.
4843 This improves the real-time response for the
4844 offloaded CPUs by relieving them of the need to
4845 wake up the corresponding kthread, but degrades
4846 energy efficiency by requiring that the kthreads
4847 periodically wake up to do the polling.
4849 rcutree.blimit= [KNL]
4850 Set maximum number of finished RCU callbacks to
4851 process in one batch.
4853 rcutree.do_rcu_barrier= [KNL]
4854 Request a call to rcu_barrier(). This is
4855 throttled so that userspace tests can safely
4856 hammer on the sysfs variable if they so choose.
4857 If triggered before the RCU grace-period machinery
4858 is fully active, this will error out with EAGAIN.
4860 rcutree.dump_tree= [KNL]
4861 Dump the structure of the rcu_node combining tree
4862 out at early boot. This is used for diagnostic
4863 purposes, to verify correct tree setup.
4865 rcutree.gp_cleanup_delay= [KNL]
4866 Set the number of jiffies to delay each step of
4867 RCU grace-period cleanup.
4869 rcutree.gp_init_delay= [KNL]
4870 Set the number of jiffies to delay each step of
4871 RCU grace-period initialization.
4873 rcutree.gp_preinit_delay= [KNL]
4874 Set the number of jiffies to delay each step of
4875 RCU grace-period pre-initialization, that is,
4876 the propagation of recent CPU-hotplug changes up
4877 the rcu_node combining tree.
4879 rcutree.jiffies_till_first_fqs= [KNL]
4880 Set delay from grace-period initialization to
4881 first attempt to force quiescent states.
4882 Units are jiffies, minimum value is zero,
4883 and maximum value is HZ.
4885 rcutree.jiffies_till_next_fqs= [KNL]
4886 Set delay between subsequent attempts to force
4887 quiescent states. Units are jiffies, minimum
4888 value is one, and maximum value is HZ.
4890 rcutree.jiffies_till_sched_qs= [KNL]
4891 Set required age in jiffies for a
4892 given grace period before RCU starts
4893 soliciting quiescent-state help from
4894 rcu_note_context_switch() and cond_resched().
4895 If not specified, the kernel will calculate
4896 a value based on the most recent settings
4897 of rcutree.jiffies_till_first_fqs
4898 and rcutree.jiffies_till_next_fqs.
4899 This calculated value may be viewed in
4900 rcutree.jiffies_to_sched_qs. Any attempt to set
4901 rcutree.jiffies_to_sched_qs will be cheerfully
4904 rcutree.kthread_prio= [KNL,BOOT]
4905 Set the SCHED_FIFO priority of the RCU per-CPU
4906 kthreads (rcuc/N). This value is also used for
4907 the priority of the RCU boost threads (rcub/N)
4908 and for the RCU grace-period kthreads (rcu_bh,
4909 rcu_preempt, and rcu_sched). If RCU_BOOST is
4910 set, valid values are 1-99 and the default is 1
4911 (the least-favored priority). Otherwise, when
4912 RCU_BOOST is not set, valid values are 0-99 and
4913 the default is zero (non-realtime operation).
4914 When RCU_NOCB_CPU is set, also adjust the
4915 priority of NOCB callback kthreads.
4917 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4918 On callback-offloaded (rcu_nocbs) CPUs,
4919 RCU reduces the lock contention that would
4920 otherwise be caused by callback floods through
4921 use of the ->nocb_bypass list. However, in the
4922 common non-flooded case, RCU queues directly to
4923 the main ->cblist in order to avoid the extra
4924 overhead of the ->nocb_bypass list and its lock.
4925 But if there are too many callbacks queued during
4926 a single jiffy, RCU pre-queues the callbacks into
4927 the ->nocb_bypass queue. The definition of "too
4928 many" is supplied by this kernel boot parameter.
4930 rcutree.qhimark= [KNL]
4931 Set threshold of queued RCU callbacks beyond which
4932 batch limiting is disabled.
4934 rcutree.qlowmark= [KNL]
4935 Set threshold of queued RCU callbacks below which
4936 batch limiting is re-enabled.
4938 rcutree.qovld= [KNL]
4939 Set threshold of queued RCU callbacks beyond which
4940 RCU's force-quiescent-state scan will aggressively
4941 enlist help from cond_resched() and sched IPIs to
4942 help CPUs more quickly reach quiescent states.
4943 Set to less than zero to make this be set based
4944 on rcutree.qhimark at boot time and to zero to
4945 disable more aggressive help enlistment.
4947 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4948 Set the page-cache refill delay (in milliseconds)
4949 in response to low-memory conditions. The range
4950 of permitted values is in the range 0:100000.
4952 rcutree.rcu_divisor= [KNL]
4953 Set the shift-right count to use to compute
4954 the callback-invocation batch limit bl from
4955 the number of callbacks queued on this CPU.
4956 The result will be bounded below by the value of
4957 the rcutree.blimit kernel parameter. Every bl
4958 callbacks, the softirq handler will exit in
4959 order to allow the CPU to do other work.
4961 Please note that this callback-invocation batch
4962 limit applies only to non-offloaded callback
4963 invocation. Offloaded callbacks are instead
4964 invoked in the context of an rcuoc kthread, which
4965 scheduler will preempt as it does any other task.
4967 rcutree.rcu_fanout_exact= [KNL]
4968 Disable autobalancing of the rcu_node combining
4969 tree. This is used by rcutorture, and might
4970 possibly be useful for architectures having high
4971 cache-to-cache transfer latencies.
4973 rcutree.rcu_fanout_leaf= [KNL]
4974 Change the number of CPUs assigned to each
4975 leaf rcu_node structure. Useful for very
4976 large systems, which will choose the value 64,
4977 and for NUMA systems with large remote-access
4978 latencies, which will choose a value aligned
4979 with the appropriate hardware boundaries.
4981 rcutree.rcu_min_cached_objs= [KNL]
4982 Minimum number of objects which are cached and
4983 maintained per one CPU. Object size is equal
4984 to PAGE_SIZE. The cache allows to reduce the
4985 pressure to page allocator, also it makes the
4986 whole algorithm to behave better in low memory
4989 rcutree.rcu_nocb_gp_stride= [KNL]
4990 Set the number of NOCB callback kthreads in
4991 each group, which defaults to the square root
4992 of the number of CPUs. Larger numbers reduce
4993 the wakeup overhead on the global grace-period
4994 kthread, but increases that same overhead on
4995 each group's NOCB grace-period kthread.
4997 rcutree.rcu_kick_kthreads= [KNL]
4998 Cause the grace-period kthread to get an extra
4999 wake_up() if it sleeps three times longer than
5000 it should at force-quiescent-state time.
5001 This wake_up() will be accompanied by a
5002 WARN_ONCE() splat and an ftrace_dump().
5004 rcutree.rcu_resched_ns= [KNL]
5005 Limit the time spend invoking a batch of RCU
5006 callbacks to the specified number of nanoseconds.
5007 By default, this limit is checked only once
5008 every 32 callbacks in order to limit the pain
5009 inflicted by local_clock() overhead.
5011 rcutree.rcu_unlock_delay= [KNL]
5012 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
5013 this specifies an rcu_read_unlock()-time delay
5014 in microseconds. This defaults to zero.
5015 Larger delays increase the probability of
5016 catching RCU pointer leaks, that is, buggy use
5017 of RCU-protected pointers after the relevant
5018 rcu_read_unlock() has completed.
5020 rcutree.sysrq_rcu= [KNL]
5021 Commandeer a sysrq key to dump out Tree RCU's
5022 rcu_node tree with an eye towards determining
5023 why a new grace period has not yet started.
5025 rcutree.use_softirq= [KNL]
5026 If set to zero, move all RCU_SOFTIRQ processing to
5027 per-CPU rcuc kthreads. Defaults to a non-zero
5028 value, meaning that RCU_SOFTIRQ is used by default.
5029 Specify rcutree.use_softirq=0 to use rcuc kthreads.
5031 But note that CONFIG_PREEMPT_RT=y kernels disable
5032 this kernel boot parameter, forcibly setting it
5035 rcutree.enable_rcu_lazy= [KNL]
5036 To save power, batch RCU callbacks and flush after
5037 delay, memory pressure or callback list growing too
5040 rcuscale.gp_async= [KNL]
5041 Measure performance of asynchronous
5042 grace-period primitives such as call_rcu().
5044 rcuscale.gp_async_max= [KNL]
5045 Specify the maximum number of outstanding
5046 callbacks per writer thread. When a writer
5047 thread exceeds this limit, it invokes the
5048 corresponding flavor of rcu_barrier() to allow
5049 previously posted callbacks to drain.
5051 rcuscale.gp_exp= [KNL]
5052 Measure performance of expedited synchronous
5053 grace-period primitives.
5055 rcuscale.holdoff= [KNL]
5056 Set test-start holdoff period. The purpose of
5057 this parameter is to delay the start of the
5058 test until boot completes in order to avoid
5061 rcuscale.kfree_by_call_rcu= [KNL]
5062 In kernels built with CONFIG_RCU_LAZY=y, test
5063 call_rcu() instead of kfree_rcu().
5065 rcuscale.kfree_mult= [KNL]
5066 Instead of allocating an object of size kfree_obj,
5067 allocate one of kfree_mult * sizeof(kfree_obj).
5070 rcuscale.kfree_rcu_test= [KNL]
5071 Set to measure performance of kfree_rcu() flooding.
5073 rcuscale.kfree_rcu_test_double= [KNL]
5074 Test the double-argument variant of kfree_rcu().
5075 If this parameter has the same value as
5076 rcuscale.kfree_rcu_test_single, both the single-
5077 and double-argument variants are tested.
5079 rcuscale.kfree_rcu_test_single= [KNL]
5080 Test the single-argument variant of kfree_rcu().
5081 If this parameter has the same value as
5082 rcuscale.kfree_rcu_test_double, both the single-
5083 and double-argument variants are tested.
5085 rcuscale.kfree_nthreads= [KNL]
5086 The number of threads running loops of kfree_rcu().
5088 rcuscale.kfree_alloc_num= [KNL]
5089 Number of allocations and frees done in an iteration.
5091 rcuscale.kfree_loops= [KNL]
5092 Number of loops doing rcuscale.kfree_alloc_num number
5093 of allocations and frees.
5095 rcuscale.minruntime= [KNL]
5096 Set the minimum test run time in seconds. This
5097 does not affect the data-collection interval,
5098 but instead allows better measurement of things
5099 like CPU consumption.
5101 rcuscale.nreaders= [KNL]
5102 Set number of RCU readers. The value -1 selects
5103 N, where N is the number of CPUs. A value
5104 "n" less than -1 selects N-n+1, where N is again
5105 the number of CPUs. For example, -2 selects N
5106 (the number of CPUs), -3 selects N+1, and so on.
5107 A value of "n" less than or equal to -N selects
5110 rcuscale.nwriters= [KNL]
5111 Set number of RCU writers. The values operate
5112 the same as for rcuscale.nreaders.
5113 N, where N is the number of CPUs
5115 rcuscale.scale_type= [KNL]
5116 Specify the RCU implementation to test.
5118 rcuscale.shutdown= [KNL]
5119 Shut the system down after performance tests
5120 complete. This is useful for hands-off automated
5123 rcuscale.verbose= [KNL]
5124 Enable additional printk() statements.
5126 rcuscale.writer_holdoff= [KNL]
5127 Write-side holdoff between grace periods,
5128 in microseconds. The default of zero says
5131 rcuscale.writer_holdoff_jiffies= [KNL]
5132 Additional write-side holdoff between grace
5133 periods, but in jiffies. The default of zero
5136 rcutorture.fqs_duration= [KNL]
5137 Set duration of force_quiescent_state bursts
5140 rcutorture.fqs_holdoff= [KNL]
5141 Set holdoff time within force_quiescent_state bursts
5144 rcutorture.fqs_stutter= [KNL]
5145 Set wait time between force_quiescent_state bursts
5148 rcutorture.fwd_progress= [KNL]
5149 Specifies the number of kthreads to be used
5150 for RCU grace-period forward-progress testing
5151 for the types of RCU supporting this notion.
5152 Defaults to 1 kthread, values less than zero or
5153 greater than the number of CPUs cause the number
5156 rcutorture.fwd_progress_div= [KNL]
5157 Specify the fraction of a CPU-stall-warning
5158 period to do tight-loop forward-progress testing.
5160 rcutorture.fwd_progress_holdoff= [KNL]
5161 Number of seconds to wait between successive
5162 forward-progress tests.
5164 rcutorture.fwd_progress_need_resched= [KNL]
5165 Enclose cond_resched() calls within checks for
5166 need_resched() during tight-loop forward-progress
5169 rcutorture.gp_cond= [KNL]
5170 Use conditional/asynchronous update-side
5171 primitives, if available.
5173 rcutorture.gp_exp= [KNL]
5174 Use expedited update-side primitives, if available.
5176 rcutorture.gp_normal= [KNL]
5177 Use normal (non-expedited) asynchronous
5178 update-side primitives, if available.
5180 rcutorture.gp_sync= [KNL]
5181 Use normal (non-expedited) synchronous
5182 update-side primitives, if available. If all
5183 of rcutorture.gp_cond=, rcutorture.gp_exp=,
5184 rcutorture.gp_normal=, and rcutorture.gp_sync=
5185 are zero, rcutorture acts as if is interpreted
5186 they are all non-zero.
5188 rcutorture.irqreader= [KNL]
5189 Run RCU readers from irq handlers, or, more
5190 accurately, from a timer handler. Not all RCU
5191 flavors take kindly to this sort of thing.
5193 rcutorture.leakpointer= [KNL]
5194 Leak an RCU-protected pointer out of the reader.
5195 This can of course result in splats, and is
5196 intended to test the ability of things like
5197 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
5200 rcutorture.n_barrier_cbs= [KNL]
5201 Set callbacks/threads for rcu_barrier() testing.
5203 rcutorture.nfakewriters= [KNL]
5204 Set number of concurrent RCU writers. These just
5205 stress RCU, they don't participate in the actual
5206 test, hence the "fake".
5208 rcutorture.nocbs_nthreads= [KNL]
5209 Set number of RCU callback-offload togglers.
5210 Zero (the default) disables toggling.
5212 rcutorture.nocbs_toggle= [KNL]
5213 Set the delay in milliseconds between successive
5214 callback-offload toggling attempts.
5216 rcutorture.nreaders= [KNL]
5217 Set number of RCU readers. The value -1 selects
5218 N-1, where N is the number of CPUs. A value
5219 "n" less than -1 selects N-n-2, where N is again
5220 the number of CPUs. For example, -2 selects N
5221 (the number of CPUs), -3 selects N+1, and so on.
5223 rcutorture.object_debug= [KNL]
5224 Enable debug-object double-call_rcu() testing.
5226 rcutorture.onoff_holdoff= [KNL]
5227 Set time (s) after boot for CPU-hotplug testing.
5229 rcutorture.onoff_interval= [KNL]
5230 Set time (jiffies) between CPU-hotplug operations,
5231 or zero to disable CPU-hotplug testing.
5233 rcutorture.read_exit= [KNL]
5234 Set the number of read-then-exit kthreads used
5235 to test the interaction of RCU updaters and
5236 task-exit processing.
5238 rcutorture.read_exit_burst= [KNL]
5239 The number of times in a given read-then-exit
5240 episode that a set of read-then-exit kthreads
5243 rcutorture.read_exit_delay= [KNL]
5244 The delay, in seconds, between successive
5245 read-then-exit testing episodes.
5247 rcutorture.shuffle_interval= [KNL]
5248 Set task-shuffle interval (s). Shuffling tasks
5249 allows some CPUs to go into dyntick-idle mode
5250 during the rcutorture test.
5252 rcutorture.shutdown_secs= [KNL]
5253 Set time (s) after boot system shutdown. This
5254 is useful for hands-off automated testing.
5256 rcutorture.stall_cpu= [KNL]
5257 Duration of CPU stall (s) to test RCU CPU stall
5258 warnings, zero to disable.
5260 rcutorture.stall_cpu_block= [KNL]
5261 Sleep while stalling if set. This will result
5262 in warnings from preemptible RCU in addition to
5263 any other stall-related activity. Note that
5264 in kernels built with CONFIG_PREEMPTION=n and
5265 CONFIG_PREEMPT_COUNT=y, this parameter will
5266 cause the CPU to pass through a quiescent state.
5267 Given CONFIG_PREEMPTION=n, this will suppress
5268 RCU CPU stall warnings, but will instead result
5269 in scheduling-while-atomic splats.
5271 Use of this module parameter results in splats.
5274 rcutorture.stall_cpu_holdoff= [KNL]
5275 Time to wait (s) after boot before inducing stall.
5277 rcutorture.stall_cpu_irqsoff= [KNL]
5278 Disable interrupts while stalling if set.
5280 rcutorture.stall_gp_kthread= [KNL]
5281 Duration (s) of forced sleep within RCU
5282 grace-period kthread to test RCU CPU stall
5283 warnings, zero to disable. If both stall_cpu
5284 and stall_gp_kthread are specified, the
5285 kthread is starved first, then the CPU.
5287 rcutorture.stat_interval= [KNL]
5288 Time (s) between statistics printk()s.
5290 rcutorture.stutter= [KNL]
5291 Time (s) to stutter testing, for example, specifying
5292 five seconds causes the test to run for five seconds,
5293 wait for five seconds, and so on. This tests RCU's
5294 ability to transition abruptly to and from idle.
5296 rcutorture.test_boost= [KNL]
5297 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5298 "Maybe" means test if the RCU implementation
5299 under test support RCU priority boosting.
5301 rcutorture.test_boost_duration= [KNL]
5302 Duration (s) of each individual boost test.
5304 rcutorture.test_boost_interval= [KNL]
5305 Interval (s) between each boost test.
5307 rcutorture.test_no_idle_hz= [KNL]
5308 Test RCU's dyntick-idle handling. See also the
5309 rcutorture.shuffle_interval parameter.
5311 rcutorture.torture_type= [KNL]
5312 Specify the RCU implementation to test.
5314 rcutorture.verbose= [KNL]
5315 Enable additional printk() statements.
5317 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5318 Dump ftrace buffer after reporting RCU CPU
5321 rcupdate.rcu_cpu_stall_notifiers= [KNL]
5322 Provide RCU CPU stall notifiers, but see the
5323 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig
5324 option's help text. TL;DR: You almost certainly
5325 do not want rcupdate.rcu_cpu_stall_notifiers.
5327 rcupdate.rcu_cpu_stall_suppress= [KNL]
5328 Suppress RCU CPU stall warning messages.
5330 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5331 Suppress RCU CPU stall warning messages and
5332 rcutorture writer stall warnings that occur
5333 during early boot, that is, during the time
5334 before the init task is spawned.
5336 rcupdate.rcu_cpu_stall_timeout= [KNL]
5337 Set timeout for RCU CPU stall warning messages.
5338 The value is in seconds and the maximum allowed
5339 value is 300 seconds.
5341 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5342 Set timeout for expedited RCU CPU stall warning
5343 messages. The value is in milliseconds
5344 and the maximum allowed value is 21000
5345 milliseconds. Please note that this value is
5346 adjusted to an arch timer tick resolution.
5347 Setting this to zero causes the value from
5348 rcupdate.rcu_cpu_stall_timeout to be used (after
5349 conversion from seconds to milliseconds).
5351 rcupdate.rcu_cpu_stall_cputime= [KNL]
5352 Provide statistics on the cputime and count of
5353 interrupts and tasks during the sampling period. For
5354 multiple continuous RCU stalls, all sampling periods
5355 begin at half of the first RCU stall timeout.
5357 rcupdate.rcu_exp_stall_task_details= [KNL]
5358 Print stack dumps of any tasks blocking the
5359 current expedited RCU grace period during an
5360 expedited RCU CPU stall warning.
5362 rcupdate.rcu_expedited= [KNL]
5363 Use expedited grace-period primitives, for
5364 example, synchronize_rcu_expedited() instead
5365 of synchronize_rcu(). This reduces latency,
5366 but can increase CPU utilization, degrade
5367 real-time latency, and degrade energy efficiency.
5368 No effect on CONFIG_TINY_RCU kernels.
5370 rcupdate.rcu_normal= [KNL]
5371 Use only normal grace-period primitives,
5372 for example, synchronize_rcu() instead of
5373 synchronize_rcu_expedited(). This improves
5374 real-time latency, CPU utilization, and
5375 energy efficiency, but can expose users to
5376 increased grace-period latency. This parameter
5377 overrides rcupdate.rcu_expedited. No effect on
5378 CONFIG_TINY_RCU kernels.
5380 rcupdate.rcu_normal_after_boot= [KNL]
5381 Once boot has completed (that is, after
5382 rcu_end_inkernel_boot() has been invoked), use
5383 only normal grace-period primitives. No effect
5384 on CONFIG_TINY_RCU kernels.
5386 But note that CONFIG_PREEMPT_RT=y kernels enables
5387 this kernel boot parameter, forcibly setting
5388 it to the value one, that is, converting any
5389 post-boot attempt at an expedited RCU grace
5390 period to instead use normal non-expedited
5391 grace-period processing.
5393 rcupdate.rcu_task_collapse_lim= [KNL]
5394 Set the maximum number of callbacks present
5395 at the beginning of a grace period that allows
5396 the RCU Tasks flavors to collapse back to using
5397 a single callback queue. This switching only
5398 occurs when rcupdate.rcu_task_enqueue_lim is
5399 set to the default value of -1.
5401 rcupdate.rcu_task_contend_lim= [KNL]
5402 Set the minimum number of callback-queuing-time
5403 lock-contention events per jiffy required to
5404 cause the RCU Tasks flavors to switch to per-CPU
5405 callback queuing. This switching only occurs
5406 when rcupdate.rcu_task_enqueue_lim is set to
5407 the default value of -1.
5409 rcupdate.rcu_task_enqueue_lim= [KNL]
5410 Set the number of callback queues to use for the
5411 RCU Tasks family of RCU flavors. The default
5412 of -1 allows this to be automatically (and
5413 dynamically) adjusted. This parameter is intended
5416 rcupdate.rcu_task_ipi_delay= [KNL]
5417 Set time in jiffies during which RCU tasks will
5418 avoid sending IPIs, starting with the beginning
5419 of a given grace period. Setting a large
5420 number avoids disturbing real-time workloads,
5421 but lengthens grace periods.
5423 rcupdate.rcu_task_lazy_lim= [KNL]
5424 Number of callbacks on a given CPU that will
5425 cancel laziness on that CPU. Use -1 to disable
5426 cancellation of laziness, but be advised that
5427 doing so increases the danger of OOM due to
5430 rcupdate.rcu_task_stall_info= [KNL]
5431 Set initial timeout in jiffies for RCU task stall
5432 informational messages, which give some indication
5433 of the problem for those not patient enough to
5434 wait for ten minutes. Informational messages are
5435 only printed prior to the stall-warning message
5436 for a given grace period. Disable with a value
5437 less than or equal to zero. Defaults to ten
5438 seconds. A change in value does not take effect
5439 until the beginning of the next grace period.
5441 rcupdate.rcu_task_stall_info_mult= [KNL]
5442 Multiplier for time interval between successive
5443 RCU task stall informational messages for a given
5444 RCU tasks grace period. This value is clamped
5445 to one through ten, inclusive. It defaults to
5446 the value three, so that the first informational
5447 message is printed 10 seconds into the grace
5448 period, the second at 40 seconds, the third at
5449 160 seconds, and then the stall warning at 600
5450 seconds would prevent a fourth at 640 seconds.
5452 rcupdate.rcu_task_stall_timeout= [KNL]
5453 Set timeout in jiffies for RCU task stall
5454 warning messages. Disable with a value less
5455 than or equal to zero. Defaults to ten minutes.
5456 A change in value does not take effect until
5457 the beginning of the next grace period.
5459 rcupdate.rcu_tasks_lazy_ms= [KNL]
5460 Set timeout in milliseconds RCU Tasks asynchronous
5461 callback batching for call_rcu_tasks().
5462 A negative value will take the default. A value
5463 of zero will disable batching. Batching is
5464 always disabled for synchronize_rcu_tasks().
5466 rcupdate.rcu_tasks_rude_lazy_ms= [KNL]
5467 Set timeout in milliseconds RCU Tasks
5468 Rude asynchronous callback batching for
5469 call_rcu_tasks_rude(). A negative value
5470 will take the default. A value of zero will
5471 disable batching. Batching is always disabled
5472 for synchronize_rcu_tasks_rude().
5474 rcupdate.rcu_tasks_trace_lazy_ms= [KNL]
5475 Set timeout in milliseconds RCU Tasks
5476 Trace asynchronous callback batching for
5477 call_rcu_tasks_trace(). A negative value
5478 will take the default. A value of zero will
5479 disable batching. Batching is always disabled
5480 for synchronize_rcu_tasks_trace().
5482 rcupdate.rcu_self_test= [KNL]
5483 Run the RCU early boot self tests
5487 Run specified binary instead of /init from the ramdisk,
5488 used for early userspace startup. See initrd.
5491 force - Override the decision by the kernel to hide the
5492 advertisement of RDRAND support (this affects
5493 certain AMD processors because of buggy BIOS
5494 support, specifically around the suspend/resume
5498 Turn on/off individual RDT features. List is:
5499 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5501 E.g. to turn on cmt and turn off mba use:
5505 Format (x86 or x86_64):
5506 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5508 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5510 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5511 (prefix with 'panic_' to set mode for panic
5513 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5514 reboot_force is either force or not specified,
5515 reboot_cpu is s[mp]#### with #### being the processor
5516 to be used for rebooting.
5518 refscale.holdoff= [KNL]
5519 Set test-start holdoff period. The purpose of
5520 this parameter is to delay the start of the
5521 test until boot completes in order to avoid
5524 refscale.lookup_instances= [KNL]
5525 Number of data elements to use for the forms of
5526 SLAB_TYPESAFE_BY_RCU testing. A negative number
5527 is negated and multiplied by nr_cpu_ids, while
5528 zero specifies nr_cpu_ids.
5530 refscale.loops= [KNL]
5531 Set the number of loops over the synchronization
5532 primitive under test. Increasing this number
5533 reduces noise due to loop start/end overhead,
5534 but the default has already reduced the per-pass
5535 noise to a handful of picoseconds on ca. 2020
5538 refscale.nreaders= [KNL]
5539 Set number of readers. The default value of -1
5540 selects N, where N is roughly 75% of the number
5541 of CPUs. A value of zero is an interesting choice.
5543 refscale.nruns= [KNL]
5544 Set number of runs, each of which is dumped onto
5547 refscale.readdelay= [KNL]
5548 Set the read-side critical-section duration,
5549 measured in microseconds.
5551 refscale.scale_type= [KNL]
5552 Specify the read-protection implementation to test.
5554 refscale.shutdown= [KNL]
5555 Shut down the system at the end of the performance
5556 test. This defaults to 1 (shut it down) when
5557 refscale is built into the kernel and to 0 (leave
5558 it running) when refscale is built as a module.
5560 refscale.verbose= [KNL]
5561 Enable additional printk() statements.
5563 refscale.verbose_batched= [KNL]
5564 Batch the additional printk() statements. If zero
5565 (the default) or negative, print everything. Otherwise,
5566 print every Nth verbose statement, where N is the value
5569 regulator_ignore_unused
5571 Prevents regulator framework from disabling regulators
5572 that are unused, due no driver claiming them. This may
5573 be useful for debug and development, but should not be
5574 needed on a platform with proper driver support.
5577 [KNL, SMP] Set scheduler's default relax_domain_level.
5578 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5580 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5581 Format: <base1>,<size1>[,<base2>,<size2>,...]
5582 Reserve I/O ports or memory so the kernel won't use
5583 them. If <base> is less than 0x10000, the region
5584 is assumed to be I/O ports; otherwise it is memory.
5586 reservetop= [X86-32,EARLY]
5588 Reserves a hole at the top of the kernel virtual
5591 reset_devices [KNL] Force drivers to reset the underlying device
5592 during initialization.
5595 Specify the partition device for software suspend
5597 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5599 resume_offset= [SWSUSP]
5600 Specify the offset from the beginning of the partition
5601 given by "resume=" at which the swap header is located,
5602 in <PAGE_SIZE> units (needed only for swap files).
5603 See Documentation/power/swsusp-and-swap-files.rst
5605 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5606 read the resume files
5608 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5609 Useful for devices that are detected asynchronously
5610 (e.g. USB and MMC devices).
5612 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will
5613 be accessible via /sys/firmware/initrd.
5615 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5616 Speculative Code Execution with Return Instructions)
5619 AMD-based UNRET and IBPB mitigations alone do not stop
5620 sibling threads from influencing the predictions of other
5621 sibling threads. For that reason, STIBP is used on pro-
5622 cessors that support it, and mitigate SMT on processors
5626 auto - automatically select a migitation
5627 auto,nosmt - automatically select a mitigation,
5628 disabling SMT if necessary for
5629 the full mitigation (only on Zen1
5630 and older without STIBP).
5631 ibpb - On AMD, mitigate short speculation
5632 windows on basic block boundaries too.
5633 Safe, highest perf impact. It also
5634 enables STIBP if present. Not suitable
5636 ibpb,nosmt - Like "ibpb" above but will disable SMT
5637 when STIBP is not available. This is
5638 the alternative for systems which do not
5640 unret - Force enable untrained return thunks,
5641 only effective on AMD f15h-f17h based
5643 unret,nosmt - Like unret, but will disable SMT when STIBP
5644 is not available. This is the alternative for
5645 systems which do not have STIBP.
5647 Selecting 'auto' will choose a mitigation method at run
5648 time according to the CPU.
5650 Not specifying this option is equivalent to retbleed=auto.
5652 rfkill.default_state=
5653 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5654 etc. communication is blocked by default.
5657 rfkill.master_switch_mode=
5658 0 The "airplane mode" button does nothing.
5659 1 The "airplane mode" button toggles between everything
5660 blocked and the previous configuration.
5661 2 The "airplane mode" button toggles between everything
5662 blocked and everything unblocked.
5664 rhash_entries= [KNL,NET]
5665 Set number of hash buckets for route cache
5668 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5671 riscv_isa_fallback [RISCV,EARLY]
5672 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit
5673 falling back to detecting extension support by parsing
5674 "riscv,isa" property on devicetree systems when the
5675 replacement properties are not found. See the Kconfig
5676 entry for RISCV_ISA_FALLBACK.
5678 ro [KNL] Mount root device read-only on boot
5681 on Mark read-only kernel memory as read-only (default).
5682 off Leave read-only kernel memory writable for debugging.
5683 full Mark read-only kernel memory and aliases as read-only
5688 Enable the uart passthrough on the designated usb port
5689 on Rockchip SoCs. When active, the signals of the
5690 debug-uart get routed to the D+ and D- pins of the usb
5691 port and the regular usb controller gets disabled.
5693 root= [KNL] Root filesystem
5694 Usually this a a block device specifier of some kind,
5695 see the early_lookup_bdev comment in
5696 block/early-lookup.c for details.
5697 Alternatively this can be "ram" for the legacy initial
5698 ramdisk, "nfs" and "cifs" for root on a network file
5699 system, or "mtd" and "ubi" for mounting from raw flash.
5701 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5702 mount the root filesystem
5704 rootflags= [KNL] Set root filesystem mount option string
5706 rootfstype= [KNL] Set root filesystem type
5708 rootwait [KNL] Wait (indefinitely) for root device to show up.
5709 Useful for devices that are detected asynchronously
5710 (e.g. USB and MMC devices).
5712 rootwait= [KNL] Maximum time (in seconds) to wait for root device
5713 to show up before attempting to mount the root
5716 rproc_mem=nn[KMG][@address]
5717 [KNL,ARM,CMA] Remoteproc physical memory block.
5718 Memory area to be used by remote processor image,
5721 rw [KNL] Mount root device read-write on boot
5723 S [KNL] Run init in single mode
5725 s390_iommu= [HW,S390]
5726 Set s390 IOTLB flushing mode
5728 With strict flushing every unmap operation will result
5729 in an IOTLB flush. Default is lazy flushing before
5730 reuse, which is faster. Deprecated, equivalent to
5733 s390_iommu_aperture= [KNL,S390]
5734 Specifies the size of the per device DMA address space
5735 accessible through the DMA and IOMMU APIs as a decimal
5736 factor of the size of main memory.
5737 The default is 1 meaning that one can concurrently use
5738 as many DMA addresses as physical memory is installed,
5739 if supported by hardware, and thus map all of memory
5740 once. With a value of 2 one can map all of memory twice
5741 and so on. As a special case a factor of 0 imposes no
5742 restrictions other than those given by hardware at the
5743 cost of significant additional memory use for tables.
5746 See drivers/net/irda/sa1100_ir.c.
5748 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages.
5750 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5751 Allowed values are enable and disable. This feature
5752 incurs a small amount of overhead in the scheduler
5753 but is useful for debugging and performance tuning.
5755 sched_thermal_decay_shift=
5756 [KNL, SMP] Set a decay shift for scheduler thermal
5757 pressure signal. Thermal pressure signal follows the
5758 default decay period of other scheduler pelt
5759 signals(usually 32 ms but configurable). Setting
5760 sched_thermal_decay_shift will left shift the decay
5761 period for the thermal pressure signal by the shift
5763 i.e. with the default pelt decay period of 32 ms
5764 sched_thermal_decay_shift thermal pressure decay pr
5768 Format: integer between 0 and 10
5771 scftorture.holdoff= [KNL]
5772 Number of seconds to hold off before starting
5773 test. Defaults to zero for module insertion and
5774 to 10 seconds for built-in smp_call_function()
5777 scftorture.longwait= [KNL]
5778 Request ridiculously long waits randomly selected
5779 up to the chosen limit in seconds. Zero (the
5780 default) disables this feature. Please note
5781 that requesting even small non-zero numbers of
5782 seconds can result in RCU CPU stall warnings,
5783 softlockup complaints, and so on.
5785 scftorture.nthreads= [KNL]
5786 Number of kthreads to spawn to invoke the
5787 smp_call_function() family of functions.
5788 The default of -1 specifies a number of kthreads
5789 equal to the number of CPUs.
5791 scftorture.onoff_holdoff= [KNL]
5792 Number seconds to wait after the start of the
5793 test before initiating CPU-hotplug operations.
5795 scftorture.onoff_interval= [KNL]
5796 Number seconds to wait between successive
5797 CPU-hotplug operations. Specifying zero (which
5798 is the default) disables CPU-hotplug operations.
5800 scftorture.shutdown_secs= [KNL]
5801 The number of seconds following the start of the
5802 test after which to shut down the system. The
5803 default of zero avoids shutting down the system.
5804 Non-zero values are useful for automated tests.
5806 scftorture.stat_interval= [KNL]
5807 The number of seconds between outputting the
5808 current test statistics to the console. A value
5809 of zero disables statistics output.
5811 scftorture.stutter_cpus= [KNL]
5812 The number of jiffies to wait between each change
5813 to the set of CPUs under test.
5815 scftorture.use_cpus_read_lock= [KNL]
5816 Use use_cpus_read_lock() instead of the default
5817 preempt_disable() to disable CPU hotplug
5818 while invoking one of the smp_call_function*()
5821 scftorture.verbose= [KNL]
5822 Enable additional printk() statements.
5824 scftorture.weight_single= [KNL]
5825 The probability weighting to use for the
5826 smp_call_function_single() function with a zero
5827 "wait" parameter. A value of -1 selects the
5828 default if all other weights are -1. However,
5829 if at least one weight has some other value, a
5830 value of -1 will instead select a weight of zero.
5832 scftorture.weight_single_wait= [KNL]
5833 The probability weighting to use for the
5834 smp_call_function_single() function with a
5835 non-zero "wait" parameter. See weight_single.
5837 scftorture.weight_many= [KNL]
5838 The probability weighting to use for the
5839 smp_call_function_many() function with a zero
5840 "wait" parameter. See weight_single.
5841 Note well that setting a high probability for
5842 this weighting can place serious IPI load
5845 scftorture.weight_many_wait= [KNL]
5846 The probability weighting to use for the
5847 smp_call_function_many() function with a
5848 non-zero "wait" parameter. See weight_single
5851 scftorture.weight_all= [KNL]
5852 The probability weighting to use for the
5853 smp_call_function_all() function with a zero
5854 "wait" parameter. See weight_single and
5857 scftorture.weight_all_wait= [KNL]
5858 The probability weighting to use for the
5859 smp_call_function_all() function with a
5860 non-zero "wait" parameter. See weight_single
5863 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate
5864 xtime_lock contention on larger systems, and/or RCU lock
5865 contention on all systems with CONFIG_MAXSMP set.
5866 Format: { "0" | "1" }
5867 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5869 Note: increases power consumption, thus should only be
5870 enabled if running jitter sensitive (HPC/RT) workloads.
5872 security= [SECURITY] Choose a legacy "major" security module to
5873 enable at boot. This has been deprecated by the
5876 selinux= [SELINUX] Disable or enable SELinux at boot time.
5877 Format: { "0" | "1" }
5878 See security/selinux/Kconfig help text.
5883 serialnumber [BUGS=X86-32]
5885 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
5888 Maximal number of shapers.
5890 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
5891 Limit apic dumping. The parameter defines the maximal
5892 number of local apics being dumped. Also it is possible
5893 to set it to "all" by meaning -- no limit here.
5894 Format: { 1 (default) | 2 | ... | all }.
5895 The parameter valid if only apic=debug or
5896 apic=verbose is specified.
5897 Example: apic=debug show_lapic=all
5905 Enable merging of slabs with similar size when the
5906 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5909 Disable merging of slabs with similar size. May be
5910 necessary if there is some reason to distinguish
5911 allocs to different slabs, especially in hardened
5912 environments where the risk of heap overflows and
5913 layout control by attackers can usually be
5914 frustrated by disabling merging. This will reduce
5915 most of the exposure of a heap attack to a single
5916 cache (risks via metadata attacks are mostly
5917 unchanged). Debug options disable merging on their
5919 For more information see Documentation/mm/slub.rst.
5921 slab_max_order= [MM, SLAB]
5922 Determines the maximum allowed order for slabs.
5923 A high setting may cause OOMs due to memory
5924 fragmentation. Defaults to 1 for systems with
5925 more than 32MB of RAM, 0 otherwise.
5927 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5928 Enabling slub_debug allows one to determine the
5929 culprit if slab objects become corrupted. Enabling
5930 slub_debug can create guard zones around objects and
5931 may poison objects when not in use. Also tracks the
5932 last alloc / free. For more information see
5933 Documentation/mm/slub.rst.
5935 slub_max_order= [MM, SLUB]
5936 Determines the maximum allowed order for slabs.
5937 A high setting may cause OOMs due to memory
5938 fragmentation. For more information see
5939 Documentation/mm/slub.rst.
5941 slub_min_objects= [MM, SLUB]
5942 The minimum number of objects per slab. SLUB will
5943 increase the slab order up to slub_max_order to
5944 generate a sufficiently large slab able to contain
5945 the number of objects indicated. The higher the number
5946 of objects the smaller the overhead of tracking slabs
5947 and the less frequently locks need to be acquired.
5948 For more information see Documentation/mm/slub.rst.
5950 slub_min_order= [MM, SLUB]
5951 Determines the minimum page order for slabs. Must be
5952 lower than slub_max_order.
5953 For more information see Documentation/mm/slub.rst.
5955 slub_merge [MM, SLUB]
5956 Same with slab_merge.
5958 slub_nomerge [MM, SLUB]
5959 Same with slab_nomerge. This is supported for legacy.
5960 See slab_nomerge for more information.
5963 Format: <io1>[,<io2>[,...,<io8>]]
5965 smp.csd_lock_timeout= [KNL]
5966 Specify the period of time in milliseconds
5967 that smp_call_function() and friends will wait
5968 for a CPU to release the CSD lock. This is
5969 useful when diagnosing bugs involving CPUs
5970 disabling interrupts for extended periods
5971 of time. Defaults to 5,000 milliseconds, and
5972 setting a value of zero disables this feature.
5973 This feature may be more efficiently disabled
5974 using the csdlock_debug- kernel parameter.
5976 smp.panic_on_ipistall= [KNL]
5977 If a csd_lock_timeout extends for more than
5978 the specified number of milliseconds, panic the
5979 system. By default, let CSD-lock acquisition
5980 take as long as they take. Specifying 300,000
5981 for this value provides a 5-minute timeout.
5983 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5984 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5985 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5986 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5987 smsc-ircc2.ircc_irq= [HW] IRQ line
5988 smsc-ircc2.ircc_dma= [HW] DMA channel
5989 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5990 0: Toshiba Satellite 1800 (GP data pin select)
5991 1: Fast pin select (default)
5994 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads
5995 (logical CPUs) to use per physical CPU on systems
5996 capable of symmetric multithreading (SMT). Will
5997 be capped to the actual hardware limit.
5999 Default: -1 (no limit)
6002 [KNL] Should the soft-lockup detector generate panics.
6005 A value of 1 instructs the soft-lockup detector
6006 to panic the machine when a soft-lockup occurs. It is
6007 also controlled by the kernel.softlockup_panic sysctl
6008 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
6009 respective build-time switch to that functionality.
6011 softlockup_all_cpu_backtrace=
6012 [KNL] Should the soft-lockup detector generate
6013 backtraces on all cpus.
6016 sonypi.*= [HW] Sony Programmable I/O Control Device driver
6017 See Documentation/admin-guide/laptops/sonypi.rst
6019 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2
6020 (indirect branch speculation) vulnerability.
6021 The default operation protects the kernel from
6024 on - unconditionally enable, implies
6026 off - unconditionally disable, implies
6028 auto - kernel detects whether your CPU model is
6031 Selecting 'on' will, and 'auto' may, choose a
6032 mitigation method at run time according to the
6033 CPU, the available microcode, the setting of the
6034 CONFIG_RETPOLINE configuration option, and the
6035 compiler with which the kernel was built.
6037 Selecting 'on' will also enable the mitigation
6038 against user space to user space task attacks.
6040 Selecting 'off' will disable both the kernel and
6041 the user space protections.
6043 Specific mitigations can also be selected manually:
6045 retpoline - replace indirect branches
6046 retpoline,generic - Retpolines
6047 retpoline,lfence - LFENCE; indirect branch
6048 retpoline,amd - alias for retpoline,lfence
6049 eibrs - Enhanced/Auto IBRS
6050 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines
6051 eibrs,lfence - Enhanced/Auto IBRS + LFENCE
6052 ibrs - use IBRS to protect kernel
6054 Not specifying this option is equivalent to
6058 [X86] Control mitigation of Spectre variant 2
6059 (indirect branch speculation) vulnerability between
6062 on - Unconditionally enable mitigations. Is
6063 enforced by spectre_v2=on
6065 off - Unconditionally disable mitigations. Is
6066 enforced by spectre_v2=off
6068 prctl - Indirect branch speculation is enabled,
6069 but mitigation can be enabled via prctl
6070 per thread. The mitigation control state
6071 is inherited on fork.
6074 - Like "prctl" above, but only STIBP is
6075 controlled per thread. IBPB is issued
6076 always when switching between different user
6080 - Same as "prctl" above, but all seccomp
6081 threads will enable the mitigation unless
6082 they explicitly opt out.
6085 - Like "seccomp" above, but only STIBP is
6086 controlled per thread. IBPB is issued
6087 always when switching between different
6088 user space processes.
6090 auto - Kernel selects the mitigation depending on
6091 the available CPU features and vulnerability.
6093 Default mitigation: "prctl"
6095 Not specifying this option is equivalent to
6096 spectre_v2_user=auto.
6098 spec_rstack_overflow=
6099 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs
6101 off - Disable mitigation
6102 microcode - Enable microcode mitigation only
6103 safe-ret - Enable sw-only safe RET mitigation (default)
6104 ibpb - Enable mitigation by issuing IBPB on
6106 ibpb-vmexit - Issue IBPB only on VMEXIT
6107 (cloud-specific mitigation)
6109 spec_store_bypass_disable=
6110 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation
6111 (Speculative Store Bypass vulnerability)
6113 Certain CPUs are vulnerable to an exploit against a
6114 a common industry wide performance optimization known
6115 as "Speculative Store Bypass" in which recent stores
6116 to the same memory location may not be observed by
6117 later loads during speculative execution. The idea
6118 is that such stores are unlikely and that they can
6119 be detected prior to instruction retirement at the
6120 end of a particular speculation execution window.
6122 In vulnerable processors, the speculatively forwarded
6123 store can be used in a cache side channel attack, for
6124 example to read memory to which the attacker does not
6125 directly have access (e.g. inside sandboxed code).
6127 This parameter controls whether the Speculative Store
6128 Bypass optimization is used.
6130 On x86 the options are:
6132 on - Unconditionally disable Speculative Store Bypass
6133 off - Unconditionally enable Speculative Store Bypass
6134 auto - Kernel detects whether the CPU model contains an
6135 implementation of Speculative Store Bypass and
6136 picks the most appropriate mitigation. If the
6137 CPU is not vulnerable, "off" is selected. If the
6138 CPU is vulnerable the default mitigation is
6139 architecture and Kconfig dependent. See below.
6140 prctl - Control Speculative Store Bypass per thread
6141 via prctl. Speculative Store Bypass is enabled
6142 for a process by default. The state of the control
6143 is inherited on fork.
6144 seccomp - Same as "prctl" above, but all seccomp threads
6145 will disable SSB unless they explicitly opt out.
6147 Default mitigations:
6150 On powerpc the options are:
6152 on,auto - On Power8 and Power9 insert a store-forwarding
6153 barrier on kernel entry and exit. On Power7
6154 perform a software flush on kernel entry and
6158 Not specifying this option is equivalent to
6159 spec_store_bypass_disable=auto.
6161 spia_io_base= [HW,MTD]
6167 [X86] Enable split lock detection or bus lock detection
6169 When enabled (and if hardware support is present), atomic
6170 instructions that access data across cache line
6171 boundaries will result in an alignment check exception
6172 for split lock detection or a debug exception for
6177 warn - the kernel will emit rate-limited warnings
6178 about applications triggering the #AC
6179 exception or the #DB exception. This mode is
6180 the default on CPUs that support split lock
6181 detection or bus lock detection. Default
6182 behavior is by #AC if both features are
6183 enabled in hardware.
6185 fatal - the kernel will send SIGBUS to applications
6186 that trigger the #AC exception or the #DB
6187 exception. Default behavior is by #AC if
6188 both features are enabled in hardware.
6191 Set system wide rate limit to N bus locks
6192 per second for bus lock detection.
6195 N/A for split lock detection.
6198 If an #AC exception is hit in the kernel or in
6199 firmware (i.e. not while executing in user mode)
6200 the kernel will oops in either "warn" or "fatal"
6203 #DB exception for bus lock is triggered only when
6206 srbds= [X86,INTEL,EARLY]
6207 Control the Special Register Buffer Data Sampling
6210 Certain CPUs are vulnerable to an MDS-like
6211 exploit which can leak bits from the random
6214 By default, this issue is mitigated by
6215 microcode. However, the microcode fix can cause
6216 the RDRAND and RDSEED instructions to become
6217 much slower. Among other effects, this will
6218 result in reduced throughput from /dev/urandom.
6220 The microcode mitigation can be disabled with
6221 the following option:
6223 off: Disable mitigation and remove
6224 performance impact to RDRAND and RDSEED
6226 srcutree.big_cpu_lim [KNL]
6227 Specifies the number of CPUs constituting a
6228 large system, such that srcu_struct structures
6229 should immediately allocate an srcu_node array.
6230 This kernel-boot parameter defaults to 128,
6231 but takes effect only when the low-order four
6232 bits of srcutree.convert_to_big is equal to 3
6235 srcutree.convert_to_big [KNL]
6236 Specifies under what conditions an SRCU tree
6237 srcu_struct structure will be converted to big
6238 form, that is, with an rcu_node tree:
6241 1: At init_srcu_struct() time.
6242 2: When rcutorture decides to.
6243 3: Decide at boot time (default).
6244 0x1X: Above plus if high contention.
6246 Either way, the srcu_node tree will be sized based
6247 on the actual runtime number of CPUs (nr_cpu_ids)
6248 instead of the compile-time CONFIG_NR_CPUS.
6250 srcutree.counter_wrap_check [KNL]
6251 Specifies how frequently to check for
6252 grace-period sequence counter wrap for the
6253 srcu_data structure's ->srcu_gp_seq_needed field.
6254 The greater the number of bits set in this kernel
6255 parameter, the less frequently counter wrap will
6256 be checked for. Note that the bottom two bits
6259 srcutree.exp_holdoff [KNL]
6260 Specifies how many nanoseconds must elapse
6261 since the end of the last SRCU grace period for
6262 a given srcu_struct until the next normal SRCU
6263 grace period will be considered for automatic
6264 expediting. Set to zero to disable automatic
6267 srcutree.srcu_max_nodelay [KNL]
6268 Specifies the number of no-delay instances
6269 per jiffy for which the SRCU grace period
6270 worker thread will be rescheduled with zero
6271 delay. Beyond this limit, worker thread will
6272 be rescheduled with a sleep delay of one jiffy.
6274 srcutree.srcu_max_nodelay_phase [KNL]
6275 Specifies the per-grace-period phase, number of
6276 non-sleeping polls of readers. Beyond this limit,
6277 grace period worker thread will be rescheduled
6278 with a sleep delay of one jiffy, between each
6279 rescan of the readers, for a grace period phase.
6281 srcutree.srcu_retry_check_delay [KNL]
6282 Specifies number of microseconds of non-sleeping
6283 delay between each non-sleeping poll of readers.
6285 srcutree.small_contention_lim [KNL]
6286 Specifies the number of update-side contention
6287 events per jiffy will be tolerated before
6288 initiating a conversion of an srcu_struct
6289 structure to big form. Note that the value of
6290 srcutree.convert_to_big must have the 0x10 bit
6291 set for contention-based conversions to occur.
6293 ssbd= [ARM64,HW,EARLY]
6294 Speculative Store Bypass Disable control
6296 On CPUs that are vulnerable to the Speculative
6297 Store Bypass vulnerability and offer a
6298 firmware based mitigation, this parameter
6299 indicates how the mitigation should be used:
6301 force-on: Unconditionally enable mitigation for
6302 for both kernel and userspace
6303 force-off: Unconditionally disable mitigation for
6304 for both kernel and userspace
6305 kernel: Always enable mitigation in the
6306 kernel, and offer a prctl interface
6307 to allow userspace to register its
6308 interest in being mitigated too.
6310 stack_guard_gap= [MM]
6311 override the default stack gap protection. The value
6312 is in page units and it defines how many pages prior
6313 to (for stacks growing down) resp. after (for stacks
6314 growing up) the main stack are reserved for no other
6315 mapping. Default value is 256 pages.
6317 stack_depot_disable= [KNL,EARLY]
6318 Setting this to true through kernel command line will
6319 disable the stack depot thereby saving the static memory
6320 consumed by the stack hash table. By default this is set
6324 Enabled the stack tracer on boot up.
6326 stacktrace_filter=[function-list]
6327 [FTRACE] Limit the functions that the stack tracer
6328 will trace at boot up. function-list is a comma-separated
6329 list of functions. This list can be changed at run
6330 time by the stack_trace_filter file in the debugfs
6331 tracing directory. Note, this enables stack tracing
6332 and the stacktrace above is not needed.
6336 Set the STI (builtin display/keyboard on the HP-PARISC
6337 machines) console (graphic card) which should be used
6338 as the initial boot-console.
6339 See also comment in drivers/video/console/sticore.c.
6342 See comment in drivers/video/console/sticore.c.
6345 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
6350 Enable or disable strict sigaltstack size checks
6351 against the required signal frame size which
6352 depends on the supported FPU features. This can
6353 be used to filter out binaries which have
6354 not yet been made aware of AT_MINSIGSTKSZ.
6356 stress_hpt [PPC,EARLY]
6357 Limits the number of kernel HPT entries in the hash
6358 page table to increase the rate of hash page table
6359 faults on kernel addresses.
6361 stress_slb [PPC,EARLY]
6362 Limits the number of kernel SLB entries, and flushes
6363 them frequently to increase the rate of SLB faults
6364 on kernel addresses.
6366 sunrpc.min_resvport=
6367 sunrpc.max_resvport=
6369 SunRPC servers often require that client requests
6370 originate from a privileged port (i.e. a port in the
6371 range 0 < portnr < 1024).
6372 An administrator who wishes to reserve some of these
6373 ports for other uses may adjust the range that the
6374 kernel's sunrpc client considers to be privileged
6375 using these two parameters to set the minimum and
6376 maximum port values.
6378 sunrpc.svc_rpc_per_connection_limit=
6380 Limit the number of requests that the server will
6381 process in parallel from a single connection.
6382 The default value is 0 (no limit).
6386 Control how the NFS server code allocates CPUs to
6387 service thread pools. Depending on how many NICs
6388 you have and where their interrupts are bound, this
6389 option will affect which CPUs will do NFS serving.
6390 Note: this parameter cannot be changed while the
6391 NFS server is running.
6393 auto the server chooses an appropriate mode
6394 automatically using heuristics
6395 global a single global pool contains all CPUs
6396 percpu one pool for each CPU
6397 pernode one pool for each NUMA node (equivalent
6398 to global on non-NUMA machines)
6400 sunrpc.tcp_slot_table_entries=
6401 sunrpc.udp_slot_table_entries=
6403 Sets the upper limit on the number of simultaneous
6404 RPC calls that can be sent from the client to a
6405 server. Increasing these values may allow you to
6406 improve throughput, but will also increase the
6407 amount of memory reserved for use by the client.
6409 suspend.pm_test_delay=
6411 Sets the number of seconds to remain in a suspend test
6412 mode before resuming the system (see
6413 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6414 is set. Default value is 5.
6417 Format: { on | off | y | n | 1 | 0 }
6418 This parameter controls use of the Protected
6419 Execution Facility on pSeries.
6421 swiotlb= [ARM,IA-64,PPC,MIPS,X86,EARLY]
6422 Format: { <int> [,<int>] | force | noforce }
6423 <int> -- Number of I/O TLB slabs
6424 <int> -- Second integer after comma. Number of swiotlb
6425 areas with their own lock. Will be rounded up
6427 force -- force using of bounce buffers even if they
6428 wouldn't be automatically used by the kernel
6429 noforce -- Never use bounce buffers (for debugging)
6431 switches= [HW,M68k,EARLY]
6434 Set a sysctl parameter, right before loading the init
6435 process, as if the value was written to the respective
6436 /proc/sys/... file. Both '.' and '/' are recognized as
6437 separators. Unrecognized parameters and invalid values
6438 are reported in the kernel log. Sysctls registered
6439 later by a loaded module cannot be set this way.
6440 Example: sysctl.vm.swappiness=40
6442 sysrq_always_enabled
6444 Ignore sysrq setting - this boot parameter will
6445 neutralize any effect of /proc/sys/kernel/sysrq.
6446 Useful for debugging.
6448 tcpmhash_entries= [KNL,NET]
6449 Set the number of tcp_metrics_hash slots.
6450 Default value is 8192 or 16384 depending on total
6451 ram pages. This is used to specify the TCP metrics
6452 cache size. See Documentation/networking/ip-sysctl.rst
6453 "tcp_no_metrics_save" section for more details.
6457 test_suspend= [SUSPEND]
6458 Format: { "mem" | "standby" | "freeze" }[,N]
6459 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6460 standby suspend) or "freeze" (for suspend type freeze)
6461 as the system sleep state during system startup with
6462 the optional capability to repeat N number of times.
6463 The system is woken from this state using a
6464 wakeup-capable RTC alarm.
6466 thash_entries= [KNL,NET]
6467 Set number of hash buckets for TCP connection
6469 thermal.act= [HW,ACPI]
6470 -1: disable all active trip points in all thermal zones
6471 <degrees C>: override all lowest active trip points
6473 thermal.crt= [HW,ACPI]
6474 -1: disable all critical trip points in all thermal zones
6475 <degrees C>: override all critical trip points
6477 thermal.off= [HW,ACPI]
6478 1: disable ACPI thermal control
6480 thermal.psv= [HW,ACPI]
6481 -1: disable all passive trip points
6482 <degrees C>: override all passive trip points to this
6485 thermal.tzp= [HW,ACPI]
6486 Specify global default ACPI thermal zone polling rate
6487 <deci-seconds>: poll all this frequency
6488 0: no polling (default)
6490 threadirqs [KNL,EARLY]
6491 Force threading of all interrupt handlers except those
6492 marked explicitly IRQF_NO_THREAD.
6494 topology= [S390,EARLY]
6496 Specify if the kernel should make use of the cpu
6497 topology information if the hardware supports this.
6498 The scheduler will make use of this information and
6499 e.g. base its process migration decisions on it.
6502 topology_updates= [KNL, PPC, NUMA]
6504 Specify if the kernel should ignore (off)
6505 topology updates sent by the hypervisor to this
6508 torture.disable_onoff_at_boot= [KNL]
6509 Prevent the CPU-hotplug component of torturing
6510 until after init has spawned.
6512 torture.ftrace_dump_at_shutdown= [KNL]
6513 Dump the ftrace buffer at torture-test shutdown,
6514 even if there were no errors. This can be a
6515 very costly operation when many torture tests
6516 are running concurrently, especially on systems
6517 with rotating-rust storage.
6519 torture.verbose_sleep_frequency= [KNL]
6520 Specifies how many verbose printk()s should be
6521 emitted between each sleep. The default of zero
6522 disables verbose-printk() sleeping.
6524 torture.verbose_sleep_duration= [KNL]
6525 Duration of each verbose-printk() sleep in jiffies.
6529 tpm_suspend_pcr=[HW,TPM]
6530 Format: integer pcr id
6531 Specify that at suspend time, the tpm driver
6532 should extend the specified pcr with zeros,
6533 as a workaround for some chips which fail to
6534 flush the last written pcr on TPM_SaveState.
6535 This will guarantee that all the other pcrs
6538 tpm_tis.interrupts= [HW,TPM]
6539 Enable interrupts for the MMIO based physical layer
6540 for the FIFO interface. By default it is set to false
6541 (0). For more information about TPM hardware interfaces
6542 defined by Trusted Computing Group (TCG) see
6543 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/
6546 Have the tracepoints sent to printk as well as the
6547 tracing ring buffer. This is useful for early boot up
6548 where the system hangs or reboots and does not give the
6549 option for reading the tracing buffer or performing a
6550 ftrace_dump_on_oops.
6552 To turn off having tracepoints sent to printk,
6553 echo 0 > /proc/sys/kernel/tracepoint_printk
6554 Note, echoing 1 into this file without the
6555 tracepoint_printk kernel cmdline option has no effect.
6557 The tp_printk_stop_on_boot (see below) can also be used
6558 to stop the printing of events to console at
6563 Having tracepoints sent to printk() and activating high
6564 frequency tracepoints such as irq or sched, can cause
6565 the system to live lock.
6567 tp_printk_stop_on_boot [FTRACE]
6568 When tp_printk (above) is set, it can cause a lot of noise
6569 on the console. It may be useful to only include the
6570 printing of events during boot up, as user space may
6571 make the system inoperable.
6573 This command line option will stop the printing of events
6574 to console at the late_initcall_sync() time frame.
6576 trace_buf_size=nn[KMG]
6577 [FTRACE] will set tracing buffer size on each cpu.
6579 trace_clock= [FTRACE] Set the clock used for tracing events
6581 local - Use the per CPU time stamp counter
6582 (converted into nanoseconds). Fast, but
6583 depending on the architecture, may not be
6584 in sync between CPUs.
6585 global - Event time stamps are synchronize across
6586 CPUs. May be slower than the local clock,
6587 but better for some race conditions.
6588 counter - Simple counting of events (1, 2, ..)
6589 note, some counts may be skipped due to the
6590 infrastructure grabbing the clock more than
6592 uptime - Use jiffies as the time stamp.
6593 perf - Use the same clock that perf uses.
6594 mono - Use ktime_get_mono_fast_ns() for time stamps.
6595 mono_raw - Use ktime_get_raw_fast_ns() for time
6597 boot - Use ktime_get_boot_fast_ns() for time stamps.
6598 Architectures may add more clocks. See
6599 Documentation/trace/ftrace.rst for more details.
6601 trace_event=[event-list]
6602 [FTRACE] Set and start specified trace events in order
6603 to facilitate early boot debugging. The event-list is a
6604 comma-separated list of trace events to enable. See
6605 also Documentation/trace/events.rst
6607 trace_instance=[instance-info]
6608 [FTRACE] Create a ring buffer instance early in boot up.
6609 This will be listed in:
6611 /sys/kernel/tracing/instances
6613 Events can be enabled at the time the instance is created
6616 trace_instance=<name>,<system1>:<event1>,<system2>:<event2>
6618 Note, the "<system*>:" portion is optional if the event is
6621 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall
6623 will enable the "sched_switch" event (note, the "sched:" is optional, and
6624 the same thing would happen if it was left off). The irq_handler_entry
6625 event, and all events under the "initcall" system.
6627 trace_options=[option-list]
6628 [FTRACE] Enable or disable tracer options at boot.
6629 The option-list is a comma delimited list of options
6630 that can be enabled or disabled just as if you were
6631 to echo the option name into
6633 /sys/kernel/tracing/trace_options
6635 For example, to enable stacktrace option (to dump the
6636 stack trace of each event), add to the command line:
6638 trace_options=stacktrace
6640 See also Documentation/trace/ftrace.rst "trace options"
6643 trace_trigger=[trigger-list]
6644 [FTRACE] Add a event trigger on specific events.
6645 Set a trigger on top of a specific event, with an optional
6648 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..."
6649 Where more than one trigger may be specified that are comma deliminated.
6653 trace_trigger="sched_switch.stacktrace if prev_state == 2"
6655 The above will enable the "stacktrace" trigger on the "sched_switch"
6656 event but only trigger it if the "prev_state" of the "sched_switch"
6657 event is "2" (TASK_UNINTERUPTIBLE).
6659 See also "Event triggers" in Documentation/trace/events.rst
6663 [FTRACE] enable this option to disable tracing when a
6664 warning is hit. This turns off "tracing_on". Tracing can
6665 be enabled again by echoing '1' into the "tracing_on"
6666 file located in /sys/kernel/tracing/
6668 This option is useful, as it disables the trace before
6669 the WARNING dump is called, which prevents the trace to
6670 be filled with content caused by the warning output.
6672 This option can also be set at run time via the sysctl
6673 option: kernel/traceoff_on_warning
6675 transparent_hugepage=
6677 Format: [always|madvise|never]
6678 Can be used to control the default behavior of the system
6679 with respect to transparent hugepages.
6680 See Documentation/admin-guide/mm/transhuge.rst
6683 trusted.source= [KEYS]
6685 This parameter identifies the trust source as a backend
6686 for trusted keys implementation. Supported trust
6691 If not specified then it defaults to iterating through
6692 the trust source list starting with TPM and assigns the
6693 first trust source as a backend which is initialized
6694 successfully during iteration.
6698 The RNG used to generate key material for trusted keys.
6701 - the same value as trusted.source: "tpm" or "tee"
6703 If not specified, "default" is used. In this case,
6704 the RNG's choice is left to each individual trust source.
6706 tsc= Disable clocksource stability checks for TSC.
6708 [x86] reliable: mark tsc clocksource as reliable, this
6709 disables clocksource verification at runtime, as well
6710 as the stability checks done at bootup. Used to enable
6711 high-resolution timer mode on older hardware, and in
6712 virtualized environment.
6713 [x86] noirqtime: Do not use TSC to do irq accounting.
6714 Used to run time disable IRQ_TIME_ACCOUNTING on any
6715 platforms where RDTSC is slow and this accounting
6717 [x86] unstable: mark the TSC clocksource as unstable, this
6718 marks the TSC unconditionally unstable at bootup and
6719 avoids any further wobbles once the TSC watchdog notices.
6720 [x86] nowatchdog: disable clocksource watchdog. Used
6721 in situations with strict latency requirements (where
6722 interruptions from clocksource watchdog are not
6724 [x86] recalibrate: force recalibration against a HW timer
6725 (HPET or PM timer) on systems whose TSC frequency was
6726 obtained from HW or FW using either an MSR or CPUID(0x15).
6727 Warn if the difference is more than 500 ppm.
6728 [x86] watchdog: Use TSC as the watchdog clocksource with
6729 which to check other HW timers (HPET or PM timer), but
6730 only on systems where TSC has been deemed trustworthy.
6731 This will be suppressed by an earlier tsc=nowatchdog and
6732 can be overridden by a later tsc=nowatchdog. A console
6733 message will flag any such suppression or overriding.
6735 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given
6736 value instead. Useful when the early TSC frequency discovery
6737 procedure is not reliable, such as on overclocked systems
6738 with CPUID.16h support and partial CPUID.15h support.
6739 Format: <unsigned int>
6741 tsx= [X86] Control Transactional Synchronization
6742 Extensions (TSX) feature in Intel processors that
6743 support TSX control.
6745 This parameter controls the TSX feature. The options are:
6747 on - Enable TSX on the system. Although there are
6748 mitigations for all known security vulnerabilities,
6749 TSX has been known to be an accelerator for
6750 several previous speculation-related CVEs, and
6751 so there may be unknown security risks associated
6752 with leaving it enabled.
6754 off - Disable TSX on the system. (Note that this
6755 option takes effect only on newer CPUs which are
6756 not vulnerable to MDS, i.e., have
6757 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6758 the new IA32_TSX_CTRL MSR through a microcode
6759 update. This new MSR allows for the reliable
6760 deactivation of the TSX functionality.)
6762 auto - Disable TSX if X86_BUG_TAA is present,
6763 otherwise enable TSX on the system.
6765 Not specifying this option is equivalent to tsx=off.
6767 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6770 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async
6771 Abort (TAA) vulnerability.
6773 Similar to Micro-architectural Data Sampling (MDS)
6774 certain CPUs that support Transactional
6775 Synchronization Extensions (TSX) are vulnerable to an
6776 exploit against CPU internal buffers which can forward
6777 information to a disclosure gadget under certain
6780 In vulnerable processors, the speculatively forwarded
6781 data can be used in a cache side channel attack, to
6782 access data to which the attacker does not have direct
6785 This parameter controls the TAA mitigation. The
6788 full - Enable TAA mitigation on vulnerable CPUs
6791 full,nosmt - Enable TAA mitigation and disable SMT on
6792 vulnerable CPUs. If TSX is disabled, SMT
6793 is not disabled because CPU is not
6794 vulnerable to cross-thread TAA attacks.
6795 off - Unconditionally disable TAA mitigation
6797 On MDS-affected machines, tsx_async_abort=off can be
6798 prevented by an active MDS mitigation as both vulnerabilities
6799 are mitigated with the same mechanism so in order to disable
6800 this mitigation, you need to specify mds=off too.
6802 Not specifying this option is equivalent to
6803 tsx_async_abort=full. On CPUs which are MDS affected
6804 and deploy MDS mitigation, TAA mitigation is not
6805 required and doesn't provide any additional
6809 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6811 turbografx.map[2|3]= [HW,JOY]
6812 TurboGraFX parallel port interface
6814 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6815 See also Documentation/input/devices/joystick-parport.rst
6817 udbg-immortal [PPC] When debugging early kernel crashes that
6818 happen after console_init() and before a proper
6819 console driver takes over, this boot options might
6820 help "seeing" what's going on.
6822 uhash_entries= [KNL,NET]
6823 Set number of hash buckets for UDP/UDP-Lite connections
6826 [USB] Ignore overcurrent events (default N).
6827 Some badly-designed motherboards generate lots of
6828 bogus events, for ports that aren't wired to
6829 anything. Set this parameter to avoid log spamming.
6830 Note that genuine overcurrent events won't be
6834 [X86] Cause panic on unknown NMI.
6836 unwind_debug [X86-64,EARLY]
6837 Enable unwinder debug output. This can be
6838 useful for debugging certain unwinder error
6839 conditions, including corrupt stacks and
6840 bad/missing unwinder metadata.
6842 usbcore.authorized_default=
6843 [USB] Default USB device authorization:
6844 (default -1 = authorized (same as 1),
6845 0 = not authorized, 1 = authorized, 2 = authorized
6846 if device connected to internal port)
6848 usbcore.autosuspend=
6849 [USB] The autosuspend time delay (in seconds) used
6850 for newly-detected USB devices (default 2). This
6851 is the time required before an idle device will be
6852 autosuspended. Devices for which the delay is set
6853 to a negative value won't be autosuspended at all.
6855 usbcore.usbfs_snoop=
6856 [USB] Set to log all usbfs traffic (default 0 = off).
6858 usbcore.usbfs_snoop_max=
6859 [USB] Maximum number of bytes to snoop in each URB
6862 usbcore.blinkenlights=
6863 [USB] Set to cycle leds on hubs (default 0 = off).
6865 usbcore.old_scheme_first=
6866 [USB] Start with the old device initialization
6867 scheme (default 0 = off).
6869 usbcore.usbfs_memory_mb=
6870 [USB] Memory limit (in MB) for buffers allocated by
6871 usbfs (default = 16, 0 = max = 2047).
6873 usbcore.use_both_schemes=
6874 [USB] Try the other device initialization scheme
6875 if the first one fails (default 1 = enabled).
6877 usbcore.initial_descriptor_timeout=
6878 [USB] Specifies timeout for the initial 64-byte
6879 USB_REQ_GET_DESCRIPTOR request in milliseconds
6880 (default 5000 = 5.0 seconds).
6882 usbcore.nousb [USB] Disable the USB subsystem
6885 [USB] A list of quirk entries to augment the built-in
6886 usb core quirk list. List entries are separated by
6887 commas. Each entry has the form
6888 VendorID:ProductID:Flags. The IDs are 4-digit hex
6889 numbers and Flags is a set of letters. Each letter
6890 will change the built-in quirk; setting it if it is
6891 clear and clearing it if it is set. The letters have
6892 the following meanings:
6893 a = USB_QUIRK_STRING_FETCH_255 (string
6894 descriptors must not be fetched using
6896 b = USB_QUIRK_RESET_RESUME (device can't resume
6897 correctly so reset it instead);
6898 c = USB_QUIRK_NO_SET_INTF (device can't handle
6899 Set-Interface requests);
6900 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6901 handle its Configuration or Interface
6903 e = USB_QUIRK_RESET (device can't be reset
6904 (e.g morph devices), don't use reset);
6905 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6906 more interface descriptions than the
6907 bNumInterfaces count, and can't handle
6908 talking to these interfaces);
6909 g = USB_QUIRK_DELAY_INIT (device needs a pause
6910 during initialization, after we read
6911 the device descriptor);
6912 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6913 high speed and super speed interrupt
6914 endpoints, the USB 2.0 and USB 3.0 spec
6915 require the interval in microframes (1
6916 microframe = 125 microseconds) to be
6917 calculated as interval = 2 ^
6919 Devices with this quirk report their
6920 bInterval as the result of this
6921 calculation instead of the exponent
6922 variable used in the calculation);
6923 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6924 handle device_qualifier descriptor
6926 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6927 generates spurious wakeup, ignore
6928 remote wakeup capability);
6929 k = USB_QUIRK_NO_LPM (device can't handle Link
6931 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6932 (Device reports its bInterval as linear
6933 frames instead of the USB 2.0
6935 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6936 to be disconnected before suspend to
6937 prevent spurious wakeup);
6938 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6939 pause after every control message);
6940 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6941 delay after resetting its port);
6942 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT
6943 (Reduce timeout of the SET_ADDRESS
6944 request from 5000 ms to 500 ms);
6945 Example: quirks=0781:5580:bk,0a5c:5834:gij
6948 [USBHID] The interval which mice are to be polled at.
6951 [USBHID] The interval which joysticks are to be polled at.
6954 [USBHID] The interval which keyboards are to be polled at.
6956 usb-storage.delay_use=
6957 [UMS] The delay in seconds before a new device is
6958 scanned for Logical Units (default 1).
6961 [UMS] A list of quirks entries to supplement or
6962 override the built-in unusual_devs list. List
6963 entries are separated by commas. Each entry has
6964 the form VID:PID:Flags where VID and PID are Vendor
6965 and Product ID values (4-digit hex numbers) and
6966 Flags is a set of characters, each corresponding
6967 to a common usb-storage quirk flag as follows:
6968 a = SANE_SENSE (collect more than 18 bytes
6969 of sense data, not on uas);
6970 b = BAD_SENSE (don't collect more than 18
6971 bytes of sense data, not on uas);
6972 c = FIX_CAPACITY (decrease the reported
6973 device capacity by one sector);
6974 d = NO_READ_DISC_INFO (don't use
6975 READ_DISC_INFO command, not on uas);
6976 e = NO_READ_CAPACITY_16 (don't use
6977 READ_CAPACITY_16 command);
6978 f = NO_REPORT_OPCODES (don't use report opcodes
6980 g = MAX_SECTORS_240 (don't transfer more than
6981 240 sectors at a time, uas only);
6982 h = CAPACITY_HEURISTICS (decrease the
6983 reported device capacity by one
6984 sector if the number is odd);
6985 i = IGNORE_DEVICE (don't bind to this
6987 j = NO_REPORT_LUNS (don't use report luns
6989 k = NO_SAME (do not use WRITE_SAME, uas only)
6990 l = NOT_LOCKABLE (don't try to lock and
6991 unlock ejectable media, not on uas);
6992 m = MAX_SECTORS_64 (don't transfer more
6993 than 64 sectors = 32 KB at a time,
6995 n = INITIAL_READ10 (force a retry of the
6996 initial READ(10) command, not on uas);
6997 o = CAPACITY_OK (accept the capacity
6998 reported by the device, not on uas);
6999 p = WRITE_CACHE (the device cache is ON
7000 by default, not on uas);
7001 r = IGNORE_RESIDUE (the device reports
7002 bogus residue values, not on uas);
7003 s = SINGLE_LUN (the device has only one
7005 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
7006 commands, uas only);
7007 u = IGNORE_UAS (don't bind to the uas driver);
7008 w = NO_WP_DETECT (don't test whether the
7009 medium is write-protected).
7010 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
7011 even if the device claims no cache,
7013 Example: quirks=0419:aaf5:rl,0421:0433:rc
7015 user_debug= [KNL,ARM]
7017 See arch/arm/Kconfig.debug help text.
7018 1 - undefined instruction events
7020 4 - invalid data aborts
7023 Example: user_debug=31
7026 [X86,EARLY] Flags controlling user PTE allocations.
7028 nohigh = do not allocate PTE pages in
7029 HIGHMEM regardless of setting
7032 vdso= [X86,SH,SPARC]
7033 On X86_32, this is an alias for vdso32=. Otherwise:
7035 vdso=1: enable VDSO (the default)
7036 vdso=0: disable VDSO mapping
7038 vdso32= [X86] Control the 32-bit vDSO
7039 vdso32=1: enable 32-bit VDSO
7040 vdso32=0 or vdso32=2: disable 32-bit VDSO
7042 See the help text for CONFIG_COMPAT_VDSO for more
7043 details. If CONFIG_COMPAT_VDSO is set, the default is
7044 vdso32=0; otherwise, the default is vdso32=1.
7046 For compatibility with older kernels, vdso32=2 is an
7049 Try vdso32=0 if you encounter an error that says:
7050 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
7053 vector=percpu: enable percpu vector domain
7055 video= [FB,EARLY] Frame buffer configuration
7056 See Documentation/fb/modedb.rst.
7058 video.brightness_switch_enabled= [ACPI]
7060 If set to 1, on receiving an ACPI notify event
7061 generated by hotkey, video driver will adjust brightness
7062 level and then send out the event to user space through
7063 the allocated input device. If set to 0, video driver
7064 will only send out the event without touching backlight
7069 [VMMIO] Memory mapped virtio (platform) device.
7071 <size>@<baseaddr>:<irq>[:<id>]
7073 <size> := size (can use standard suffixes
7075 <baseaddr> := physical base address
7076 <irq> := interrupt number (as passed to
7078 <id> := (optional) platform device id
7080 virtio_mmio.device=1K@0x100b0000:48:7
7082 Can be used multiple times for multiple devices.
7084 vga= [BOOT,X86-32] Select a particular video mode
7085 See Documentation/arch/x86/boot.rst and
7086 Documentation/admin-guide/svga.rst.
7087 Use vga=ask for menu.
7088 This is actually a boot loader parameter; the value is
7089 passed to the kernel using a special protocol.
7091 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
7092 May slow down system boot speed, especially when
7093 enabled on systems with a large amount of memory.
7094 All options are enabled by default, and this
7095 interface is meant to allow for selectively
7096 enabling or disabling specific virtual memory
7099 Available options are:
7100 P Enable page structure init time poisoning
7101 - Disable all of the above options
7103 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an
7104 exact size of <nn>. This can be used to increase
7105 the minimum size (128MB on x86). It can also be
7106 used to decrease the size and leave more room
7107 for directly mapped kernel RAM.
7109 vmcp_cma=nn[MG] [KNL,S390,EARLY]
7110 Sets the memory size reserved for contiguous memory
7111 allocations for the vmcp device driver.
7113 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
7116 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
7119 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
7122 vsyscall= [X86-64,EARLY]
7123 Controls the behavior of vsyscalls (i.e. calls to
7124 fixed addresses of 0xffffffffff600x00 from legacy
7125 code). Most statically-linked binaries and older
7126 versions of glibc use these calls. Because these
7127 functions are at fixed addresses, they make nice
7128 targets for exploits that can control RIP.
7130 emulate Vsyscalls turn into traps and are emulated
7131 reasonably safely. The vsyscall page is
7134 xonly [default] Vsyscalls turn into traps and are
7135 emulated reasonably safely. The vsyscall
7136 page is not readable.
7138 none Vsyscalls don't work at all. This makes
7139 them quite hard to use for exploits but
7140 might break your system.
7142 vt.color= [VT] Default text color.
7143 Format: 0xYX, X = foreground, Y = background.
7144 Default: 0x07 = light gray on black.
7146 vt.cur_default= [VT] Default cursor shape.
7147 Format: 0xCCBBAA, where AA, BB, and CC are the same as
7148 the parameters of the <Esc>[?A;B;Cc escape sequence;
7149 see VGA-softcursor.txt. Default: 2 = underline.
7151 vt.default_blu= [VT]
7152 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
7153 Change the default blue palette of the console.
7154 This is a 16-member array composed of values
7157 vt.default_grn= [VT]
7158 Format: <green0>,<green1>,<green2>,...,<green15>
7159 Change the default green palette of the console.
7160 This is a 16-member array composed of values
7163 vt.default_red= [VT]
7164 Format: <red0>,<red1>,<red2>,...,<red15>
7165 Change the default red palette of the console.
7166 This is a 16-member array composed of values
7172 Set system-wide default UTF-8 mode for all tty's.
7173 Default is 1, i.e. UTF-8 mode is enabled for all
7174 newly opened terminals.
7176 vt.global_cursor_default=
7179 Set system-wide default for whether a cursor
7180 is shown on new VTs. Default is -1,
7181 i.e. cursors will be created by default unless
7182 overridden by individual drivers. 0 will hide
7183 cursors, 1 will display them.
7185 vt.italic= [VT] Default color for italic text; 0-15.
7188 vt.underline= [VT] Default color for underlined text; 0-15.
7191 watchdog timers [HW,WDT] For information on watchdog timers,
7192 see Documentation/watchdog/watchdog-parameters.rst
7193 or other driver-specific files in the
7194 Documentation/watchdog/ directory.
7198 Set the hard lockup detector stall duration
7199 threshold in seconds. The soft lockup detector
7200 threshold is set to twice the value. A value of 0
7201 disables both lockup detectors. Default is 10
7204 workqueue.unbound_cpus=
7205 [KNL,SMP] Specify to constrain one or some CPUs
7206 to use in unbound workqueues.
7208 By default, all online CPUs are available for
7211 workqueue.watchdog_thresh=
7212 If CONFIG_WQ_WATCHDOG is configured, workqueue can
7213 warn stall conditions and dump internal state to
7214 help debugging. 0 disables workqueue stall
7215 detection; otherwise, it's the stall threshold
7216 duration in seconds. The default value is 30 and
7217 it can be updated at runtime by writing to the
7218 corresponding sysfs file.
7220 workqueue.cpu_intensive_thresh_us=
7221 Per-cpu work items which run for longer than this
7222 threshold are automatically considered CPU intensive
7223 and excluded from concurrency management to prevent
7224 them from noticeably delaying other per-cpu work
7225 items. Default is 10000 (10ms).
7227 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel
7228 will report the work functions which violate this
7229 threshold repeatedly. They are likely good
7230 candidates for using WQ_UNBOUND workqueues instead.
7232 workqueue.cpu_intensive_warning_thresh=<uint>
7233 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel
7234 will report the work functions which violate the
7235 intensive_threshold_us repeatedly. In order to prevent
7236 spurious warnings, start printing only after a work
7237 function has violated this threshold number of times.
7239 The default is 4 times. 0 disables the warning.
7241 workqueue.power_efficient
7242 Per-cpu workqueues are generally preferred because
7243 they show better performance thanks to cache
7244 locality; unfortunately, per-cpu workqueues tend to
7245 be more power hungry than unbound workqueues.
7247 Enabling this makes the per-cpu workqueues which
7248 were observed to contribute significantly to power
7249 consumption unbound, leading to measurably lower
7250 power usage at the cost of small performance
7253 The default value of this parameter is determined by
7254 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
7256 workqueue.default_affinity_scope=
7257 Select the default affinity scope to use for unbound
7258 workqueues. Can be one of "cpu", "smt", "cache",
7259 "numa" and "system". Default is "cache". For more
7260 information, see the Affinity Scopes section in
7261 Documentation/core-api/workqueue.rst.
7263 This can be changed after boot by writing to the
7264 matching /sys/module/workqueue/parameters file. All
7265 workqueues with the "default" affinity scope will be
7266 updated accordignly.
7268 workqueue.debug_force_rr_cpu
7269 Workqueue used to implicitly guarantee that work
7270 items queued without explicit CPU specified are put
7271 on the local CPU. This guarantee is no longer true
7272 and while local CPU is still preferred work items
7273 may be put on foreign CPUs. This debug option
7274 forces round-robin CPU selection to flush out
7275 usages which depend on the now broken guarantee.
7276 When enabled, memory and cache locality will be
7279 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access
7280 Type) of ioremap_wc().
7282 on - Enable writecombine, use WUC for ioremap_wc()
7283 off - Disable writecombine, use SUC for ioremap_wc()
7285 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of
7286 default x2apic cluster mode on platforms
7289 xen_512gb_limit [KNL,X86-64,XEN]
7290 Restricts the kernel running paravirtualized under Xen
7291 to use only up to 512 GB of RAM. The reason to do so is
7292 crash analysis tools and Xen tools for doing domain
7293 save/restore/migration must be enabled to handle larger
7296 xen_emul_unplug= [HW,X86,XEN,EARLY]
7297 Unplug Xen emulated devices
7298 Format: [unplug0,][unplug1]
7299 ide-disks -- unplug primary master IDE devices
7300 aux-ide-disks -- unplug non-primary-master IDE devices
7301 nics -- unplug network devices
7302 all -- unplug all emulated devices (NICs and IDE disks)
7303 unnecessary -- unplugging emulated devices is
7304 unnecessary even if the host did not respond to
7306 never -- do not unplug even if version check succeeds
7308 xen_legacy_crash [X86,XEN,EARLY]
7309 Crash from Xen panic notifier, without executing late
7310 panic() code such as dumping handler.
7312 xen_msr_safe= [X86,XEN,EARLY]
7314 Select whether to always use non-faulting (safe) MSR
7315 access functions when running as Xen PV guest. The
7316 default value is controlled by CONFIG_XEN_PV_MSR_SAFE.
7318 xen_nopvspin [X86,XEN,EARLY]
7319 Disables the qspinlock slowpath using Xen PV optimizations.
7320 This parameter is obsoleted by "nopvspin" parameter, which
7321 has equivalent effect for XEN platform.
7324 Disables the PV optimizations forcing the HVM guest to
7325 run as generic HVM guest with no PV drivers.
7326 This option is obsoleted by the "nopv" option, which
7327 has equivalent effect for XEN platform.
7329 xen_no_vector_callback
7330 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen
7331 event channel interrupts.
7333 xen_scrub_pages= [XEN]
7334 Boolean option to control scrubbing pages before giving them back
7335 to Xen, for use by other domains. Can be also changed at runtime
7336 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
7337 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
7339 xen_timer_slop= [X86-64,XEN,EARLY]
7340 Set the timer slop (in nanoseconds) for the virtual Xen
7341 timers (default is 100000). This adjusts the minimum
7342 delta of virtualized Xen timers, where lower values
7343 improve timer resolution at the expense of processing
7344 more timer interrupts.
7346 xen.balloon_boot_timeout= [XEN]
7347 The time (in seconds) to wait before giving up to boot
7348 in case initial ballooning fails to free enough memory.
7349 Applies only when running as HVM or PVH guest and
7350 started with less memory configured than allowed at
7351 max. Default is 180.
7353 xen.event_eoi_delay= [XEN]
7354 How long to delay EOI handling in case of event
7355 storms (jiffies). Default is 10.
7357 xen.event_loop_timeout= [XEN]
7358 After which time (jiffies) the event handling loop
7359 should start to delay EOI handling. Default is 2.
7361 xen.fifo_events= [XEN]
7362 Boolean parameter to disable using fifo event handling
7363 even if available. Normally fifo event handling is
7364 preferred over the 2-level event handling, as it is
7365 fairer and the number of possible event channels is
7366 much higher. Default is on (use fifo events).
7368 xirc2ps_cs= [NET,PCMCIA]
7370 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
7373 By default on POWER9 and above, the kernel will
7374 natively use the XIVE interrupt controller. This option
7375 allows the fallback firmware mode to be used:
7377 off Fallback to firmware control of XIVE interrupt
7378 controller on both pseries and powernv
7379 platforms. Only useful on POWER9 and above.
7381 xive.store-eoi=off [PPC]
7382 By default on POWER10 and above, the kernel will use
7383 stores for EOI handling when the XIVE interrupt mode
7384 is active. This option allows the XIVE driver to use
7385 loads instead, as on POWER9.
7387 xhci-hcd.quirks [USB,KNL]
7388 A hex value specifying bitmask with supplemental xhci
7389 host controller quirks. Meaning of each bit can be
7390 consulted in header drivers/usb/host/xhci.h.
7393 Format: { early | on | rw | ro | off }
7394 Controls if xmon debugger is enabled. Default is off.
7395 Passing only "xmon" is equivalent to "xmon=early".
7396 early Call xmon as early as possible on boot; xmon
7397 debugger is called from setup_arch().
7398 on xmon debugger hooks will be installed so xmon
7399 is only called on a kernel crash. Default mode,
7400 i.e. either "ro" or "rw" mode, is controlled
7401 with CONFIG_XMON_DEFAULT_RO_MODE.
7402 rw xmon debugger hooks will be installed so xmon
7403 is called only on a kernel crash, mode is write,
7404 meaning SPR registers, memory and, other data
7405 can be written using xmon commands.
7406 ro same as "rw" option above but SPR registers,
7407 memory, and other data can't be written using
7409 off xmon is disabled.