1 acpi= [HW,ACPI,X86,ARM64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
5 force -- enable ACPI if default was off
6 on -- enable ACPI but allow fallback to DT [arm64]
7 off -- disable ACPI if default was on
8 noirq -- do not use ACPI for IRQ routing
9 strict -- Be less tolerant of platforms that are not
10 strictly ACPI specification compliant.
11 rsdt -- prefer RSDT over (default) XSDT
12 copy_dsdt -- copy DSDT to memory
13 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force"
16 See also Documentation/power/runtime_pm.rst, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
25 { vendor | video | native | none }
26 If set to vendor, prefer vendor-specific driver
27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
28 of the ACPI video.ko driver.
29 If set to video, use the ACPI video.ko driver.
30 If set to native, use the device's native backlight mode.
31 If set to none, disable the ACPI backlight interface.
33 acpi_force_32bit_fadt_addr
34 force FADT to use 32 bit addresses rather than the
35 64 bit X_* addresses. Some firmware have broken 64
36 bit addresses for force ACPI ignore these and use
37 the older legacy 32 bit addresses.
39 acpica_no_return_repair [HW, ACPI]
40 Disable AML predefined validation mechanism
41 This mechanism can repair the evaluation result to make
42 the return objects more ACPI specification compliant.
43 This option is useful for developers to identify the
44 root cause of an AML interpreter issue when the issue
45 has something to do with the repair mechanism.
47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
51 debug output. Bits in debug_layer correspond to a
52 _COMPONENT in an ACPI source file, e.g.,
53 #define _COMPONENT ACPI_EVENTS
54 Bits in debug_level correspond to a level in
55 ACPI_DEBUG_PRINT statements, e.g.,
56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
57 The debug_level mask defaults to "info". See
58 Documentation/firmware-guide/acpi/debug.rst for more information about
59 debug layers and levels.
61 Enable processor driver info messages:
62 acpi.debug_layer=0x20000000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
116 Format: <byte> or <bitmap-list>
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
163 acpi_osi=!! # enable all built-in OS vendor
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig,
229 s4_nohwsig, old_ordering, nonvs,
230 sci_force_enable, nobl }
231 See Documentation/power/video.rst for information on
233 s3_beep is for debugging; it makes the PC's speaker beep
234 as soon as the kernel's real-mode entry point is called.
235 s4_hwsig causes the kernel to check the ACPI hardware
236 signature during resume from hibernation, and gracefully
237 refuse to resume if it has changed. This complies with
238 the ACPI specification but not with reality, since
239 Windows does not do this and many laptops do change it
240 on docking. So the default behaviour is to allow resume
241 and simply warn when the signature changes, unless the
242 s4_hwsig option is enabled.
243 s4_nohwsig prevents ACPI hardware signature from being
244 used (or even warned about) during resume.
245 old_ordering causes the ACPI 1.0 ordering of the _PTS
246 control method, with respect to putting devices into
247 low power states, to be enforced (the ACPI 2.0 ordering
248 of _PTS is used by default).
249 nonvs prevents the kernel from saving/restoring the
250 ACPI NVS memory during suspend/hibernation and resume.
251 sci_force_enable causes the kernel to set SCI_EN directly
252 on resume from S1/S3 (which is against the ACPI spec,
253 but some broken systems don't work without it).
254 nobl causes the internal blacklist of systems known to
255 behave incorrectly in some ways with respect to system
256 suspend and resume to be ignored (use wisely).
258 acpi_use_timer_override [HW,ACPI]
259 Use timer override. For some broken Nvidia NF5 boards
260 that require a timer override, but don't have HPET
262 add_efi_memmap [EFI; X86] Include EFI memory map in
263 kernel's map of available physical RAM.
266 { off | try_unsupported }
267 off: disable AGP support
268 try_unsupported: try to drive unsupported chipsets
269 (may crash computer or cause data corruption)
272 See Documentation/sound/alsa-configuration.rst
275 Allow the default userspace alignment fault handler
276 behaviour to be specified. Bit 0 enables warnings,
277 bit 1 enables fixups, and bit 2 sends a segfault.
279 align_va_addr= [X86-64]
280 Align virtual addresses by clearing slice [14:12] when
281 allocating a VMA at process creation time. This option
282 gives you up to 3% performance improvement on AMD F15h
283 machines (where it is enabled by default) for a
284 CPU-intensive style benchmark, and it can vary highly in
285 a microbenchmark depending on workload and compiler.
287 32: only for 32-bit processes
288 64: only for 64-bit processes
289 on: enable for both 32- and 64-bit processes
290 off: disable for both 32- and 64-bit processes
292 alloc_snapshot [FTRACE]
293 Allocate the ftrace snapshot buffer on boot up when the
294 main buffer is allocated. This is handy if debugging
295 and you need to use tracing_snapshot() on boot up, and
296 do not want to use tracing_snapshot_alloc() as it needs
297 to be done where GFP_KERNEL allocations are allowed.
299 allow_mismatched_32bit_el0 [ARM64]
300 Allow execve() of 32-bit applications and setting of the
301 PER_LINUX32 personality on systems where only a strict
302 subset of the CPUs support 32-bit EL0. When this
303 parameter is present, the set of CPUs supporting 32-bit
304 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
305 and hot-unplug operations may be restricted.
307 See Documentation/arm64/asymmetric-32bit.rst for more
310 amd_iommu= [HW,X86-64]
311 Pass parameters to the AMD IOMMU driver in the system.
313 fullflush - Deprecated, equivalent to iommu.strict=1
314 off - do not initialize any AMD IOMMU found in
316 force_isolation - Force device isolation for all
317 devices. The IOMMU driver is not
318 allowed anymore to lift isolation
319 requirements as needed. This option
320 does not override iommu=pt
321 force_enable - Force enable the IOMMU on platforms known
322 to be buggy with IOMMU enabled. Use this
324 pgtbl_v1 - Use v1 page table for DMA-API (Default).
325 pgtbl_v2 - Use v2 page table for DMA-API.
327 amd_iommu_dump= [HW,X86-64]
328 Enable AMD IOMMU driver option to dump the ACPI table
329 for AMD IOMMU. With this option enabled, AMD IOMMU
330 driver will print ACPI tables for AMD IOMMU during
331 IOMMU initialization.
333 amd_iommu_intr= [HW,X86-64]
334 Specifies one of the following AMD IOMMU interrupt
336 legacy - Use legacy interrupt remapping mode.
337 vapic - Use virtual APIC mode, which allows IOMMU
338 to inject interrupts directly into guest.
339 This mode requires kvm-amd.avic=1.
340 (Default when IOMMU HW support is present.)
344 Do not enable amd_pstate as the default
345 scaling driver for the supported processors
347 Use amd_pstate with passive mode as a scaling driver.
348 In this mode autonomous selection is disabled.
349 Driver requests a desired performance level and platform
350 tries to match the same performance level if it is
351 satisfied by guaranteed performance level.
353 Use amd_pstate_epp driver instance as the scaling driver,
354 driver provides a hint to the hardware if software wants
355 to bias toward performance (0x0) or energy efficiency (0xff)
356 to the CPPC firmware. then CPPC power algorithm will
357 calculate the runtime workload and adjust the realtime cores
360 Activate guided autonomous mode. Driver requests minimum and
361 maximum performance level and the platform autonomously
362 selects a performance level in this range and appropriate
363 to the current workload.
365 amijoy.map= [HW,JOY] Amiga joystick support
366 Map of devices attached to JOY0DAT and JOY1DAT
368 See also Documentation/input/joydev/joystick.rst
370 analog.map= [HW,JOY] Analog joystick and gamepad support
371 Specifies type or capabilities of an analog joystick
372 connected to one of 16 gameports
373 Format: <type1>,<type2>,..<type16>
376 Power management functions (SPARCstation-4/5 + deriv.)
378 Disable APC CPU standby support. SPARCstation-Fox does
379 not play well with APC CPU idle - disable it if you have
380 APC and your system crashes randomly.
382 apic= [APIC,X86] Advanced Programmable Interrupt Controller
383 Change the output verbosity while booting
384 Format: { quiet (default) | verbose | debug }
385 Change the amount of debugging information output
386 when initialising the APIC and IO-APIC components.
387 For X86-32, this can also be used to specify an APIC
389 Format: apic=driver_name
390 Examples: apic=bigsmp
392 apic_extnmi= [APIC,X86] External NMI delivery setting
393 Format: { bsp (default) | all | none }
394 bsp: External NMI is delivered only to CPU 0
395 all: External NMIs are broadcast to all CPUs as a
397 none: External NMI is masked for all CPUs. This is
398 useful so that a dump capture kernel won't be
402 See Documentation/networking/ipv6.rst.
404 apm= [APM] Advanced Power Management
405 See header of arch/x86/kernel/apm_32.c.
407 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
408 Format: { "0" | "1" }
409 See security/apparmor/Kconfig help text
412 Default value is set via kernel config option.
414 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
415 Format: <io>,<irq>,<nodeID>
417 arm64.nobti [ARM64] Unconditionally disable Branch Target
418 Identification support
420 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
423 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
426 arm64.nosve [ARM64] Unconditionally disable Scalable Vector
429 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix
434 atarimouse= [HW,MOUSE] Atari Mouse
436 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
437 EzKey and similar keyboards
439 atkbd.reset= [HW] Reset keyboard during initialization
441 atkbd.set= [HW] Select keyboard code set
442 Format: <int> (2 = AT (default), 3 = PS/2)
444 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
447 atkbd.softraw= [HW] Choose between synthetic and real raw mode
448 Format: <bool> (0 = real, 1 = synthetic (default))
450 atkbd.softrepeat= [HW]
451 Use software keyboard repeat
453 audit= [KNL] Enable the audit sub-system
454 Format: { "0" | "1" | "off" | "on" }
455 0 | off - kernel audit is disabled and can not be
456 enabled until the next reboot
457 unset - kernel audit is initialized but disabled and
458 will be fully enabled by the userspace auditd.
459 1 | on - kernel audit is initialized and partially
460 enabled, storing at most audit_backlog_limit
461 messages in RAM until it is fully enabled by the
465 audit_backlog_limit= [KNL] Set the audit queue size limit.
466 Format: <int> (must be >=0)
469 bau= [X86_UV] Enable the BAU on SGI UV. The default
470 behavior is to disable the BAU (i.e. bau=0).
471 Format: { "0" | "1" }
474 unset - Disable the BAU.
476 baycom_epp= [HW,AX25]
479 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
481 See header of drivers/net/hamradio/baycom_par.c.
483 baycom_ser_fdx= [HW,AX25]
484 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
485 Format: <io>,<irq>,<mode>[,<baud>]
486 See header of drivers/net/hamradio/baycom_ser_fdx.c.
488 baycom_ser_hdx= [HW,AX25]
489 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
490 Format: <io>,<irq>,<mode>
491 See header of drivers/net/hamradio/baycom_ser_hdx.c.
494 Disable BERT OS support on buggy BIOSes.
496 bgrt_disable [ACPI][X86]
497 Disable BGRT to avoid flickering OEM logo.
499 blkdevparts= Manual partition parsing of block device(s) for
500 embedded devices based on command line input.
501 See Documentation/block/cmdline-partition.rst
503 boot_delay= Milliseconds to delay each printk during boot.
504 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled,
505 and you may also have to specify "lpj=". Boot_delay
506 values larger than 10 seconds (10000) are assumed
507 erroneous and ignored.
511 Extended command line options can be added to an initrd
512 and this will cause the kernel to look for it.
514 See Documentation/admin-guide/bootconfig.rst
516 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
517 bttv.radio= Most important insmod options are available as
519 bttv.pll= See Documentation/admin-guide/media/bttv.rst
522 bulk_remove=off [PPC] This parameter disables the use of the pSeries
523 firmware feature for flushing multiple hpte entries
526 c101= [NET] Moxa C101 synchronous serial card
528 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
529 Sometimes CPU hardware bugs make them report the cache
530 size incorrectly. The kernel will attempt work arounds
531 to fix known problems, but for some CPUs it is not
532 possible to determine what the correct size should be.
533 This option provides an override for these situations.
536 [NET] Specifies amount of time (in seconds) that
537 the kernel should wait for a network carrier. By default
538 it waits 120 seconds.
540 ca_keys= [KEYS] This parameter identifies a specific key(s) on
541 the system trusted keyring to be used for certificate
543 format: { id:<keyid> | builtin }
545 cca= [MIPS] Override the kernel pages' cache coherency
546 algorithm. Accepted values range from 0 to 7
547 inclusive. See arch/mips/include/asm/pgtable-bits.h
548 for platform specific values (SB1, Loongson3 and
551 ccw_timeout_log [S390]
552 See Documentation/s390/common_io.rst for details.
554 cgroup_disable= [KNL] Disable a particular controller or optional feature
555 Format: {name of the controller(s) or feature(s) to disable}
556 The effects of cgroup_disable=foo are:
557 - foo isn't auto-mounted if you mount all cgroups in
559 - foo isn't visible as an individually mountable
561 - if foo is an optional feature then the feature is
562 disabled and corresponding cgroup files are not
564 {Currently only "memory" controller deal with this and
565 cut the overhead, others just disable the usage. So
566 only cgroup_disable=memory is actually worthy}
567 Specifying "pressure" disables per-cgroup pressure
568 stall information accounting feature
570 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
571 Format: { { controller | "all" | "named" }
572 [,{ controller | "all" | "named" }...] }
573 Like cgroup_disable, but only applies to cgroup v1;
574 the blacklisted controllers remain available in cgroup2.
575 "all" blacklists all controllers and "named" disables
576 named mounts. Specifying both "all" and "named" disables
579 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
581 nosocket -- Disable socket memory accounting.
582 nokmem -- Disable kernel memory accounting.
583 nobpf -- Disable BPF memory accounting.
585 checkreqprot= [SELINUX] Set initial checkreqprot flag value.
586 Format: { "0" | "1" }
587 See security/selinux/Kconfig help text.
588 0 -- check protection applied by kernel (includes
589 any implied execute protection).
590 1 -- check protection requested by application.
591 Default value is set via a kernel config option.
592 Value can be changed at runtime via
593 /sys/fs/selinux/checkreqprot.
594 Setting checkreqprot to 1 is deprecated.
597 See Documentation/s390/common_io.rst for details.
599 clearcpuid=X[,X...] [X86]
600 Disable CPUID feature X for the kernel. See
601 arch/x86/include/asm/cpufeatures.h for the valid bit
602 numbers X. Note the Linux-specific bits are not necessarily
603 stable over kernel options, but the vendor-specific
605 X can also be a string as appearing in the flags: line
606 in /proc/cpuinfo which does not have the above
607 instability issue. However, not all features have names
609 Note that using this option will taint your kernel.
610 Also note that user programs calling CPUID directly
611 or using the feature without checking anything
612 will still see it. This just prevents it from
613 being used by the kernel or shown in /proc/cpuinfo.
614 Also note the kernel might malfunction if you disable
619 Prevents the clock framework from automatically gating
620 clocks that have not been explicitly enabled by a Linux
621 device driver but are enabled in hardware at reset or
622 by the bootloader/firmware. Note that this does not
623 force such clocks to be always-on nor does it reserve
624 those clocks in any way. This parameter is useful for
625 debug and development, but should not be needed on a
626 platform with proper driver support. For more
627 information, see Documentation/driver-api/clk.rst.
629 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
631 Forces specified clocksource (if available) to be used
632 when calculating gettimeofday(). If specified
633 clocksource is not available, it defaults to PIT.
634 Format: { pit | tsc | cyclone | pmtmr }
636 clocksource= Override the default clocksource
638 Override the default clocksource and use the clocksource
639 with the name specified.
640 Some clocksource names to choose from, depending on
642 [all] jiffies (this is the base, fallback clocksource)
644 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
645 pxa_timer,timer3,32k_counter,timer0_1
646 [X86-32] pit,hpet,tsc;
647 scx200_hrt on Geode; cyclone on IBM x440
655 clocksource.arm_arch_timer.evtstrm=
658 Enable/disable the eventstream feature of the ARM
659 architected timer so that code using WFE-based polling
660 loops can be debugged more effectively on production
663 clocksource.max_cswd_read_retries= [KNL]
664 Number of clocksource_watchdog() retries due to
665 external delays before the clock will be marked
666 unstable. Defaults to two retries, that is,
667 three attempts to read the clock under test.
669 clocksource.verify_n_cpus= [KNL]
670 Limit the number of CPUs checked for clocksources
671 marked with CLOCK_SOURCE_VERIFY_PERCPU that
672 are marked unstable due to excessive skew.
673 A negative value says to check all CPUs, while
674 zero says not to check any. Values larger than
675 nr_cpu_ids are silently truncated to nr_cpu_ids.
676 The actual CPUs are chosen randomly, with
677 no replacement if the same CPU is chosen twice.
679 clocksource-wdtest.holdoff= [KNL]
680 Set the time in seconds that the clocksource
681 watchdog test waits before commencing its tests.
682 Defaults to zero when built as a module and to
683 10 seconds when built into the kernel.
685 cma=nn[MG]@[start[MG][-end[MG]]]
687 Sets the size of kernel global memory area for
688 contiguous memory allocations and optionally the
689 placement constraint by the physical address range of
690 memory allocations. A value of 0 disables CMA
691 altogether. For more information, see
692 kernel/dma/contiguous.c
696 Sets the size of kernel per-numa memory area for
697 contiguous memory allocations. A value of 0 disables
698 per-numa CMA altogether. And If this option is not
699 specified, the default value is 0.
700 With per-numa CMA enabled, DMA users on node nid will
701 first try to allocate buffer from the pernuma area
702 which is located in node nid, if the allocation fails,
703 they will fallback to the global default memory area.
705 cmo_free_hint= [PPC] Format: { yes | no }
706 Specify whether pages are marked as being inactive
707 when they are freed. This is used in CMO environments
708 to determine OS memory pressure for page stealing by
712 coherent_pool=nn[KMG] [ARM,KNL]
713 Sets the size of memory pool for coherent, atomic dma
714 allocations, by default set to 256K.
716 com20020= [HW,NET] ARCnet - COM20020 chipset
718 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
720 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
724 ARCnet - COM90xx chipset (memory-mapped buffers)
725 Format: <io>[,<irq>[,<memstart>]]
727 condev= [HW,S390] console device
730 con3215_drop= [S390] 3215 console drop mode.
732 When set to true, drop data on the 3215 console when
733 the console buffer is full. In this case the
734 operator using a 3270 terminal emulator (for example
735 x3270) does not have to enter the clear key for the
736 console output to advance and the kernel to continue.
737 This leads to a much faster boot time when a 3270
738 terminal emulator is active. If no 3270 terminal
739 emulator is used, this parameter has no effect.
741 console= [KNL] Output console device and options.
743 tty<n> Use the virtual console device <n>.
747 Use the specified serial port. The options are of
748 the form "bbbbpnf", where "bbbb" is the baud rate,
749 "p" is parity ("n", "o", or "e"), "n" is number of
750 bits, and "f" is flow control ("r" for RTS or
751 omit it). Default is "9600n8".
753 See Documentation/admin-guide/serial-console.rst for more
755 Documentation/networking/netconsole.rst for an
758 uart[8250],io,<addr>[,options]
759 uart[8250],mmio,<addr>[,options]
760 uart[8250],mmio16,<addr>[,options]
761 uart[8250],mmio32,<addr>[,options]
762 uart[8250],0x<addr>[,options]
763 Start an early, polled-mode console on the 8250/16550
764 UART at the specified I/O port or MMIO address,
765 switching to the matching ttyS device later.
766 MMIO inter-register address stride is either 8-bit
767 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
768 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
769 to be equivalent to 'mmio'. 'options' are specified in
770 the same format described for ttyS above; if unspecified,
771 the h/w is not re-initialized.
773 hvc<n> Use the hypervisor console device <n>. This is for
774 both Xen and PowerPC hypervisors.
777 Use to disable console output, i.e., to have kernel
778 console messages discarded.
779 This must be the only console= parameter used on the
782 If the device connected to the port is not a TTY but a braille
783 device, prepend "brl," before the device type, for instance
785 For now, only VisioBraille is supported.
788 [KNL] Change console messages format
790 By default we print messages on consoles in
791 "[time stamp] text\n" format (time stamp may not be
792 printed, depending on CONFIG_PRINTK_TIME or
793 `printk_time' param).
795 Switch to syslog format: "<%u>[time stamp] text\n"
796 IOW, each message will have a facility and loglevel
797 prefix. The format is similar to one used by syslog()
798 syscall, or to executing "dmesg -S --raw" or to reading
801 consoleblank= [KNL] The console blank (screen saver) timeout in
802 seconds. A value of 0 disables the blank timer.
806 [KNL] Change the default value for
807 /proc/<pid>/coredump_filter.
808 See also Documentation/filesystems/proc.rst.
810 coresight_cpu_debug.enable
813 Enable/disable the CPU sampling based debugging.
814 0: default value, disable debugging
815 1: enable debugging at boot time
817 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
819 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
821 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
822 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
823 Some features depend on CPU0. Known dependencies are:
824 1. Resume from suspend/hibernate depends on CPU0.
825 Suspend/hibernate will fail if CPU0 is offline and you
826 need to online CPU0 before suspend/hibernate.
827 2. PIC interrupts also depend on CPU0. CPU0 can't be
828 removed if a PIC interrupt is detected.
829 It's said poweroff/reboot may depend on CPU0 on some
830 machines although I haven't seen such issues so far
831 after CPU0 is offline on a few tested machines.
832 If the dependencies are under your control, you can
833 turn on cpu0_hotplug.
835 cpuidle.off=1 [CPU_IDLE]
836 disable the cpuidle sub-system
839 [CPU_IDLE] Name of the cpuidle governor to use.
841 cpufreq.off=1 [CPU_FREQ]
842 disable the cpufreq sub-system
844 cpufreq.default_governor=
845 [CPU_FREQ] Name of the default cpufreq governor or
846 policy to use. This governor must be registered in the
847 kernel before the cpufreq driver probes.
850 [X86] Delay for N microsec between assert and de-assert
851 of APIC INIT to start processors. This delay occurs
852 on every CPU online, such as boot, and resume from suspend.
855 crash_kexec_post_notifiers
856 Run kdump after running panic-notifiers and dumping
857 kmsg. This only for the users who doubt kdump always
858 succeeds in any situation.
859 Note that this also increases risks of kdump failure,
860 because some panic notifiers can make the crashed
861 kernel more unstable.
863 crashkernel=size[KMG][@offset[KMG]]
864 [KNL] Using kexec, Linux can switch to a 'crash kernel'
865 upon panic. This parameter reserves the physical
866 memory region [offset, offset + size] for that kernel
867 image. If '@offset' is omitted, then a suitable offset
868 is selected automatically.
869 [KNL, X86-64, ARM64] Select a region under 4G first, and
870 fall back to reserve region above 4G when '@offset'
871 hasn't been specified.
872 See Documentation/admin-guide/kdump/kdump.rst for further details.
874 crashkernel=range1:size1[,range2:size2,...][@offset]
875 [KNL] Same as above, but depends on the memory
876 in the running system. The syntax of range is
877 start-[end] where start and end are both
878 a memory unit (amount[KMG]). See also
879 Documentation/admin-guide/kdump/kdump.rst for an example.
881 crashkernel=size[KMG],high
882 [KNL, X86-64, ARM64] range could be above 4G. Allow kernel
883 to allocate physical memory region from top, so could
884 be above 4G if system have more than 4G ram installed.
885 Otherwise memory region will be allocated below 4G, if
887 It will be ignored if crashkernel=X is specified.
888 crashkernel=size[KMG],low
889 [KNL, X86-64, ARM64] range under 4G. When crashkernel=X,high
890 is passed, kernel could allocate physical memory region
891 above 4G, that cause second kernel crash on system
892 that require some amount of low memory, e.g. swiotlb
893 requires at least 64M+32K low memory, also enough extra
894 low memory is needed to make sure DMA buffers for 32-bit
895 devices won't run out. Kernel would try to allocate
896 default size of memory below 4G automatically. The default
897 size is platform dependent.
898 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB)
900 This one lets the user specify own low range under 4G
901 for second kernel instead.
902 0: to disable low allocation.
903 It will be ignored when crashkernel=X,high is not used
904 or memory reserved is below 4G.
907 [KNL] Disable crypto self-tests
912 cs89x0_media= [HW,NET]
913 Format: { rj45 | aui | bnc }
915 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
916 handling. When switched on, additional debug data is
917 printed to the console in case a hanging CPU is
918 detected, and that CPU is pinged again in order to try
919 to resolve the hang situation.
920 0: disable csdlock debugging (default)
921 1: enable basic csdlock debugging (minor impact)
922 ext: enable extended csdlock debugging (more impact,
926 See header of drivers/s390/block/dasd_devmap.c.
928 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
929 (one device per port)
930 Format: <port#>,<type>
931 See also Documentation/input/devices/joystick-parport.rst
933 debug [KNL] Enable kernel debugging (events log level).
936 [KNL] Enable printing [hashed] pointers early in the
937 boot sequence. If enabled, we use a weak hash instead
938 of siphash to hash pointers. Use this option if you are
939 seeing instances of '(___ptrval___)') and need to see a
940 value (hashed pointer) instead. Cryptographically
941 insecure, please do not use on production kernels.
944 [KNL] verbose locking self-tests
946 Print debugging info while doing the locking API
948 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
949 (no extra messages), setting it to -1 (all bits set)
950 will print _a_lot_ more information - normally only
951 useful to lockdep developers.
953 debug_objects [KNL] Enable object debugging
955 debug_guardpage_minorder=
956 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
957 parameter allows control of the order of pages that will
958 be intentionally kept free (and hence protected) by the
959 buddy allocator. Bigger value increase the probability
960 of catching random memory corruption, but reduce the
961 amount of memory for normal system use. The maximum
962 possible value is MAX_ORDER/2. Setting this parameter
963 to 1 or 2 should be enough to identify most random
964 memory corruption problems caused by bugs in kernel or
965 driver code when a CPU writes to (or reads from) a
966 random memory location. Note that there exists a class
967 of memory corruptions problems caused by buggy H/W or
968 F/W or by drivers badly programming DMA (basically when
969 memory is written at bus level and the CPU MMU is
970 bypassed) which are not detectable by
971 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
972 tracking down these problems.
975 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
976 enables the feature at boot time. By default, it is
977 disabled and the system will work mostly the same as a
978 kernel built without CONFIG_DEBUG_PAGEALLOC.
979 Note: to get most of debug_pagealloc error reports, it's
980 useful to also enable the page_owner functionality.
981 on: enable the feature
983 debugfs= [KNL] This parameter enables what is exposed to userspace
984 and debugfs internal clients.
985 Format: { on, no-mount, off }
986 on: All functions are enabled.
988 Filesystem is not registered but kernel clients can
989 access APIs and a crashkernel can be used to read
990 its content. There is nothing to mount.
991 off: Filesystem is not registered and clients
992 get a -EPERM as result when trying to register files
993 or directories within debugfs.
994 This is equivalent of the runtime functionality if
995 debugfs was not enabled in the kernel at all.
996 Default value is set in build-time with a kernel configuration.
998 debugpat [X86] Enable PAT debugging
1001 [HW] The size of the default HugeTLB page. This is
1002 the size represented by the legacy /proc/ hugepages
1003 APIs. In addition, this is the default hugetlb size
1004 used for shmget(), mmap() and mounting hugetlbfs
1005 filesystems. If not specified, defaults to the
1006 architecture's default huge page size. Huge page
1007 sizes are architecture dependent. See also
1008 Documentation/admin-guide/mm/hugetlbpage.rst.
1011 deferred_probe_timeout=
1012 [KNL] Debugging option to set a timeout in seconds for
1013 deferred probe to give up waiting on dependencies to
1014 probe. Only specific dependencies (subsystems or
1015 drivers) that have opted in will be ignored. A timeout
1016 of 0 will timeout at the end of initcalls. If the time
1017 out hasn't expired, it'll be restarted by each
1018 successful driver registration. This option will also
1019 dump out devices still on the deferred probe list after
1022 delayacct [KNL] Enable per-task delay accounting
1024 dell_smm_hwmon.ignore_dmi=
1025 [HW] Continue probing hardware even if DMI data
1026 indicates that the driver is running on unsupported
1029 dell_smm_hwmon.force=
1030 [HW] Activate driver even if SMM BIOS signature does
1031 not match list of supported models and enable otherwise
1032 blacklisted features.
1034 dell_smm_hwmon.power_status=
1035 [HW] Report power status in /proc/i8k
1036 (disabled by default).
1038 dell_smm_hwmon.restricted=
1039 [HW] Allow controlling fans only if SYS_ADMIN
1042 dell_smm_hwmon.fan_mult=
1043 [HW] Factor to multiply fan speed with.
1045 dell_smm_hwmon.fan_max=
1046 [HW] Maximum configurable fan speed.
1049 Format: { on | off | def_only | inf_only | always }
1050 on: s390 zlib hardware support for compression on
1051 level 1 and decompression (default)
1052 off: No s390 zlib hardware support
1053 def_only: s390 zlib hardware support for deflate
1054 only (compression on level 1)
1055 inf_only: s390 zlib hardware support for inflate
1056 only (decompression)
1057 always: Same as 'on' but ignores the selected compression
1058 level always using hardware support (used for debugging)
1060 dhash_entries= [KNL]
1061 Set number of hash buckets for dentry cache.
1063 disable_1tb_segments [PPC]
1064 Disables the use of 1TB hash page table segments. This
1065 causes the kernel to fall back to 256MB segments which
1066 can be useful when debugging issues that require an SLB
1070 See Documentation/networking/ipv6.rst.
1073 Disable RADIX MMU mode on POWER9
1076 Disable TLBIE instruction. Currently does not work
1077 with KVM, with HASH MMU, or with coherent accelerators.
1079 disable_cpu_apicid= [X86,APIC,SMP]
1081 The number of initial APIC ID for the
1082 corresponding CPU to be disabled at boot,
1083 mostly used for the kdump 2nd kernel to
1084 disable BSP to wake up multiple CPUs without
1085 causing system reset or hang due to sending
1086 INIT from AP to BSP.
1088 disable_ddw [PPC/PSERIES]
1089 Disable Dynamic DMA Window support. Use this
1090 to workaround buggy firmware.
1092 disable_ipv6= [IPV6]
1093 See Documentation/networking/ipv6.rst.
1095 disable_mtrr_cleanup [X86]
1096 The kernel tries to adjust MTRR layout from continuous
1097 to discrete, to make X server driver able to add WB
1098 entry later. This parameter disables that.
1100 disable_mtrr_trim [X86, Intel and AMD only]
1101 By default the kernel will trim any uncacheable
1102 memory out of your available memory pool based on
1103 MTRR settings. This parameter disables that behavior,
1104 possibly causing your machine to run very slowly.
1106 disable_timer_pin_1 [X86]
1107 Disable PIN 1 of APIC timer
1108 Can be useful to work around chipset bugs.
1110 dis_ucode_ldr [X86] Disable the microcode loader.
1112 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1113 this option disables the debugging code at boot.
1115 dma_debug_entries=<number>
1116 This option allows to tune the number of preallocated
1117 entries for DMA-API debugging code. One entry is
1118 required per DMA-API allocation. Use this if the
1119 DMA-API debugging code disables itself because the
1120 architectural default is too low.
1122 dma_debug_driver=<driver_name>
1123 With this option the DMA-API debugging driver
1124 filter feature can be enabled at boot time. Just
1125 pass the driver to filter for as the parameter.
1126 The filter can be disabled or changed to another
1127 driver later using sysfs.
1129 driver_async_probe= [KNL]
1130 List of driver names to be probed asynchronously. *
1131 matches with all driver names. If * is specified, the
1132 rest of the listed driver names are those that will NOT
1134 Format: <driver_name1>,<driver_name2>...
1136 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1137 Broken monitors, graphic adapters, KVMs and EDIDless
1138 panels may send no or incorrect EDID data sets.
1139 This parameter allows to specify an EDID data sets
1140 in the /lib/firmware directory that are used instead.
1141 Generic built-in EDID data sets are used, if one of
1142 edid/1024x768.bin, edid/1280x1024.bin,
1143 edid/1680x1050.bin, or edid/1920x1080.bin is given
1144 and no file with the same name exists. Details and
1145 instructions how to build your own EDID data are
1146 available in Documentation/admin-guide/edid.rst. An EDID
1147 data set will only be used for a particular connector,
1148 if its name and a colon are prepended to the EDID
1149 name. Each connector may use a unique EDID data
1150 set by separating the files with a comma. An EDID
1151 data set with no connector name will be used for
1152 any connectors not explicitly specified.
1157 Format: {"off" | "known"}
1158 Control how the dt_cpu_ftrs device-tree binding is
1159 used for CPU feature discovery and setup (if it
1161 off: Do not use it, fall back to legacy cpu table.
1162 known: Do not pass through unknown features to guests
1163 or userspace, only those that the kernel is aware of.
1165 dump_apple_properties [X86]
1166 Dump name and content of EFI device properties on
1167 x86 Macs. Useful for driver authors to determine
1168 what data is available or for reverse-engineering.
1170 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1171 <module>.dyndbg[="val"]
1172 Enable debug messages at boot time. See
1173 Documentation/admin-guide/dynamic-debug-howto.rst
1176 early_ioremap_debug [KNL]
1177 Enable debug messages in early_ioremap support. This
1178 is useful for tracking down temporary early mappings
1179 which are not unmapped.
1181 earlycon= [KNL] Output early console device and options.
1183 When used with no options, the early console is
1184 determined by stdout-path property in device tree's
1185 chosen node or the ACPI SPCR table if supported by
1188 cdns,<addr>[,options]
1189 Start an early, polled-mode console on a Cadence
1190 (xuartps) serial port at the specified address. Only
1191 supported option is baud rate. If baud rate is not
1192 specified, the serial port must already be setup and
1195 uart[8250],io,<addr>[,options[,uartclk]]
1196 uart[8250],mmio,<addr>[,options[,uartclk]]
1197 uart[8250],mmio32,<addr>[,options[,uartclk]]
1198 uart[8250],mmio32be,<addr>[,options[,uartclk]]
1199 uart[8250],0x<addr>[,options]
1200 Start an early, polled-mode console on the 8250/16550
1201 UART at the specified I/O port or MMIO address.
1202 MMIO inter-register address stride is either 8-bit
1203 (mmio) or 32-bit (mmio32 or mmio32be).
1204 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1205 to be equivalent to 'mmio'. 'options' are specified
1206 in the same format described for "console=ttyS<n>"; if
1207 unspecified, the h/w is not initialized. 'uartclk' is
1208 the uart clock frequency; if unspecified, it is set
1209 to 'BASE_BAUD' * 16.
1213 Start an early, polled-mode console on a pl011 serial
1214 port at the specified address. The pl011 serial port
1215 must already be setup and configured. Options are not
1216 yet supported. If 'mmio32' is specified, then only
1217 the driver will use only 32-bit accessors to read/write
1218 the device registers.
1221 Start an early console on a litex serial port at the
1222 specified address. The serial port must already be
1223 setup and configured. Options are not yet supported.
1226 Start an early, polled-mode console on a meson serial
1227 port at the specified address. The serial port must
1228 already be setup and configured. Options are not yet
1232 Start an early, polled-mode console on an msm serial
1233 port at the specified address. The serial port
1234 must already be setup and configured. Options are not
1237 msm_serial_dm,<addr>
1238 Start an early, polled-mode console on an msm serial
1239 dm port at the specified address. The serial port
1240 must already be setup and configured. Options are not
1244 Start an early, polled-mode console on a serial port
1245 of an Actions Semi SoC, such as S500 or S900, at the
1246 specified address. The serial port must already be
1247 setup and configured. Options are not yet supported.
1250 Start an early, polled-mode console on a serial port
1251 of an RDA Micro SoC, such as RDA8810PL, at the
1252 specified address. The serial port must already be
1253 setup and configured. Options are not yet supported.
1256 Use RISC-V SBI (Supervisor Binary Interface) for early
1259 smh Use ARM semihosting calls for early console.
1267 Use early console provided by serial driver available
1268 on Samsung SoCs, requires selecting proper type and
1269 a correct base address of the selected UART port. The
1270 serial port must already be setup and configured.
1271 Options are not yet supported.
1274 Start an early, polled-mode console on a lantiq serial
1275 (lqasc) port at the specified address. The serial port
1276 must already be setup and configured. Options are not
1281 Use early console provided by Freescale LP UART driver
1282 found on Freescale Vybrid and QorIQ LS1021A processors.
1283 A valid base address must be provided, and the serial
1284 port must already be setup and configured.
1288 Start an early, polled-mode, output-only console on the
1289 Freescale i.MX UART at the specified address. The UART
1290 must already be setup and configured.
1293 Start an early, polled-mode console on the
1294 Armada 3700 serial port at the specified
1295 address. The serial port must already be setup
1296 and configured. Options are not yet supported.
1299 Start an early, polled-mode console on a Qualcomm
1300 Generic Interface (GENI) based serial port at the
1301 specified address. The serial port must already be
1302 setup and configured. Options are not yet supported.
1305 Start an early, unaccelerated console on the EFI
1306 memory mapped framebuffer (if available). On cache
1307 coherent non-x86 systems that use system memory for
1308 the framebuffer, pass the 'ram' option so that it is
1309 mapped with the correct attributes.
1312 Use early console provided by Freescale LINFlexD UART
1313 serial driver for NXP S32V234 SoCs. A valid base
1314 address must be provided, and the serial port must
1315 already be setup and configured.
1317 earlyprintk= [X86,SH,ARM,M68k,S390]
1321 earlyprintk=serial[,ttySn[,baudrate]]
1322 earlyprintk=serial[,0x...[,baudrate]]
1323 earlyprintk=ttySn[,baudrate]
1324 earlyprintk=dbgp[debugController#]
1325 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1326 earlyprintk=xdbc[xhciController#]
1328 earlyprintk is useful when the kernel crashes before
1329 the normal console is initialized. It is not enabled by
1330 default because it has some cosmetic problems.
1332 Append ",keep" to not disable it when the real console
1335 Only one of vga, serial, or usb debug port can
1338 Currently only ttyS0 and ttyS1 may be specified by
1339 name. Other I/O ports may be explicitly specified
1340 on some architectures (x86 and arm at least) by
1341 replacing ttySn with an I/O port address, like this:
1342 earlyprintk=serial,0x1008,115200
1343 You can find the port for a given device in
1344 /proc/tty/driver/serial:
1345 2: uart:ST16650V2 port:00001008 irq:18 ...
1347 Interaction with the standard serial driver is not
1350 The VGA output is eventually overwritten by
1353 The xen option can only be used in Xen domains.
1355 The sclp output can only be used on s390.
1357 The optional "force" to "pciserial" enables use of a
1358 PCI device even when its classcode is not of the
1361 edac_report= [HW,EDAC] Control how to report EDAC event
1362 Format: {"on" | "off" | "force"}
1363 on: enable EDAC to report H/W event. May be overridden
1364 by other higher priority error reporting module.
1365 off: disable H/W event reporting through EDAC.
1366 force: enforce the use of EDAC to report H/W event.
1370 Format: {"off" | "on" | "skip[mbr]"}
1373 Format: { "debug", "disable_early_pci_dma",
1374 "nochunk", "noruntime", "nosoftreserve",
1375 "novamap", "no_disable_early_pci_dma" }
1376 debug: enable misc debug output.
1377 disable_early_pci_dma: disable the busmaster bit on all
1378 PCI bridges while in the EFI boot stub.
1379 nochunk: disable reading files in "chunks" in the EFI
1380 boot stub, as chunking can cause problems with some
1381 firmware implementations.
1382 noruntime : disable EFI runtime services support
1383 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1384 attribute may cause the kernel to reserve the
1385 memory range for a memory mapping driver to
1386 claim. Specify efi=nosoftreserve to disable this
1387 reservation and treat the memory by its base type
1388 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1389 novamap: do not call SetVirtualAddressMap().
1390 no_disable_early_pci_dma: Leave the busmaster bit set
1391 on all PCI bridges while in the EFI boot stub
1393 efi_no_storage_paranoia [EFI; X86]
1394 Using this parameter you can use more than 50% of
1395 your efi variable storage. Use this parameter only if
1396 you are really sure that your UEFI does sane gc and
1397 fulfills the spec otherwise your board may brick.
1399 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1400 Add arbitrary attribute to specific memory range by
1401 updating original EFI memory map.
1402 Region of memory which aa attribute is added to is
1405 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1406 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1407 attribute is added to range 0x100000000-0x180000000 and
1408 0x10a0000000-0x1120000000.
1410 If efi_fake_mem=8G@9G:0x40000 is specified, the
1411 EFI_MEMORY_SP(0x40000) attribute is added to
1412 range 0x240000000-0x43fffffff.
1414 Using this parameter you can do debugging of EFI memmap
1415 related features. For example, you can do debugging of
1416 Address Range Mirroring feature even if your box
1417 doesn't support it, or mark specific memory as
1420 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1421 that is to be dynamically loaded by Linux. If there are
1422 multiple variables with the same name but with different
1423 vendor GUIDs, all of them will be loaded. See
1424 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1427 eisa_irq_edge= [PARISC,HW]
1428 See header of drivers/parisc/eisa.c.
1430 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1433 This is designed to be used in conjunction with
1434 the boot argument: earlyprintk=vga
1436 This parameter works in place of the kgdboc parameter
1437 but can only be used if the backing tty is available
1438 very early in the boot process. For early debugging
1439 via a serial port see kgdboc_earlycon instead.
1442 See comment before function elanfreq_setup() in
1443 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1445 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1446 Specifies physical address of start of kernel core
1447 image elf header and optionally the size. Generally
1448 kexec loader will pass this option to capture kernel.
1449 See Documentation/admin-guide/kdump/kdump.rst for details.
1451 enable_mtrr_cleanup [X86]
1452 The kernel tries to adjust MTRR layout from continuous
1453 to discrete, to make X server driver able to add WB
1454 entry later. This parameter enables that.
1456 enable_timer_pin_1 [X86]
1457 Enable PIN 1 of APIC timer
1458 Can be useful to work around chipset bugs
1459 (in particular on some ATI chipsets).
1460 The kernel tries to set a reasonable default.
1462 enforcing= [SELINUX] Set initial enforcing status.
1464 See security/selinux/Kconfig help text.
1465 0 -- permissive (log only, no denials).
1466 1 -- enforcing (deny and log).
1468 Value can be changed at runtime via
1469 /sys/fs/selinux/enforce.
1472 Disable Error Record Serialization Table (ERST)
1475 ether= [HW,NET] Ethernet cards parameters
1476 This option is obsoleted by the "netdev=" option, which
1477 has equivalent usage. See its documentation for details.
1481 Permit 'security.evm' to be updated regardless of
1482 current integrity status.
1484 early_page_ext [KNL] Enforces page_ext initialization to earlier
1485 stages so cover more early boot allocations.
1486 Please note that as side effect some optimizations
1487 might be disabled to achieve that (e.g. parallelized
1488 memory initialization is disabled) so the boot process
1489 might take longer, especially on systems with a lot of
1490 memory. Available with CONFIG_PAGE_EXTENSION=y.
1495 fail_make_request=[KNL]
1496 General fault injection mechanism.
1497 Format: <interval>,<probability>,<space>,<times>
1498 See also Documentation/fault-injection/.
1501 Format: { initns | none }
1502 See Documentation/admin-guide/sysctl/net.rst for
1503 fb_tunnels_only_for_init_ns
1506 See Documentation/admin-guide/blockdev/floppy.rst.
1508 force_pal_cache_flush
1509 [IA-64] Avoid check_sal_cache_flush which may hang on
1510 buggy SAL_CACHE_FLUSH implementations. Using this
1511 parameter will force ia64_sal_cache_flush to call
1512 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1515 Forcefully enable Physical Address Extension (PAE).
1516 Many Pentium M systems disable PAE but may have a
1517 functionally usable PAE implementation.
1518 Warning: use of this parameter will taint the kernel
1519 and may cause unknown problems.
1522 [FTRACE] will set and start the specified tracer
1523 as early as possible in order to facilitate early
1526 ftrace_boot_snapshot
1527 [FTRACE] On boot up, a snapshot will be taken of the
1528 ftrace ring buffer that can be read at:
1529 /sys/kernel/tracing/snapshot.
1530 This is useful if you need tracing information from kernel
1531 boot up that is likely to be overridden by user space
1532 start up functionality.
1534 Optionally, the snapshot can also be defined for a tracing
1535 instance that was created by the trace_instance= command
1538 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo
1540 The above will cause the "foo" tracing instance to trigger
1541 a snapshot at the end of boot up.
1543 ftrace_dump_on_oops[=orig_cpu]
1544 [FTRACE] will dump the trace buffers on oops.
1545 If no parameter is passed, ftrace will dump
1546 buffers of all CPUs, but if you pass orig_cpu, it will
1547 dump only the buffer of the CPU that triggered the
1550 ftrace_filter=[function-list]
1551 [FTRACE] Limit the functions traced by the function
1552 tracer at boot up. function-list is a comma-separated
1553 list of functions. This list can be changed at run
1554 time by the set_ftrace_filter file in the debugfs
1557 ftrace_notrace=[function-list]
1558 [FTRACE] Do not trace the functions specified in
1559 function-list. This list can be changed at run time
1560 by the set_ftrace_notrace file in the debugfs
1563 ftrace_graph_filter=[function-list]
1564 [FTRACE] Limit the top level callers functions traced
1565 by the function graph tracer at boot up.
1566 function-list is a comma-separated list of functions
1567 that can be changed at run time by the
1568 set_graph_function file in the debugfs tracing directory.
1570 ftrace_graph_notrace=[function-list]
1571 [FTRACE] Do not trace from the functions specified in
1572 function-list. This list is a comma-separated list of
1573 functions that can be changed at run time by the
1574 set_graph_notrace file in the debugfs tracing directory.
1576 ftrace_graph_max_depth=<uint>
1577 [FTRACE] Used with the function graph tracer. This is
1578 the max depth it will trace into a function. This value
1579 can be changed at run time by the max_graph_depth file
1580 in the tracefs tracing directory. default: 0 (no limit)
1582 fw_devlink= [KNL] Create device links between consumer and supplier
1583 devices by scanning the firmware to infer the
1584 consumer/supplier relationships. This feature is
1585 especially useful when drivers are loaded as modules as
1586 it ensures proper ordering of tasks like device probing
1587 (suppliers first, then consumers), supplier boot state
1588 clean up (only after all consumers have probed),
1589 suspend/resume & runtime PM (consumers first, then
1591 Format: { off | permissive | on | rpm }
1592 off -- Don't create device links from firmware info.
1593 permissive -- Create device links from firmware info
1594 but use it only for ordering boot state clean
1595 up (sync_state() calls).
1596 on -- Create device links from firmware info and use it
1597 to enforce probe and suspend/resume ordering.
1598 rpm -- Like "on", but also use to order runtime PM.
1600 fw_devlink.strict=<bool>
1601 [KNL] Treat all inferred dependencies as mandatory
1602 dependencies. This only applies for fw_devlink=on|rpm.
1606 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1607 support via parallel port (up to 5 devices per port)
1608 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1609 See also Documentation/input/devices/joystick-parport.rst
1613 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1617 gcov_persist= [GCOV] When non-zero (default), profiling data for
1618 kernel modules is saved and remains accessible via
1619 debugfs, even when the module is unloaded/reloaded.
1620 When zero, profiling data is discarded and associated
1621 debugfs files are removed at module unload time.
1623 goldfish [X86] Enable the goldfish android emulator platform.
1624 Don't use this when you are not running on the
1627 gpio-mockup.gpio_mockup_ranges
1628 [HW] Sets the ranges of gpiochip of for this device.
1629 Format: <start1>,<end1>,<start2>,<end2>...
1630 gpio-mockup.gpio_mockup_named_lines
1631 [HW] Let the driver know GPIO lines should be named.
1633 gpt [EFI] Forces disk with valid GPT signature but
1634 invalid Protective MBR to be treated as GPT. If the
1635 primary GPT is corrupted, it enables the backup/alternate
1636 GPT to be used instead.
1638 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1639 the "Enable 0" bit of the configuration register.
1642 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1643 the "Enable 0" bit of the configuration register.
1646 grcan.select= [HW] Select which physical interface to use.
1649 grcan.txsize= [HW] Sets the size of the tx buffer.
1650 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1652 grcan.rxsize= [HW] Sets the size of the rx buffer.
1653 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1657 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1658 hardening is enabled for this boot. Hardened
1659 usercopy checking is used to protect the kernel
1660 from reading or writing beyond known memory
1661 allocation boundaries as a proactive defense
1662 against bounds-checking flaws in the kernel's
1663 copy_to_user()/copy_from_user() interface.
1664 on Perform hardened usercopy checks (default).
1665 off Disable hardened usercopy checks.
1667 hardlockup_all_cpu_backtrace=
1668 [KNL] Should the hard-lockup detector generate
1669 backtraces on all cpus.
1672 hashdist= [KNL,NUMA] Large hashes allocated during boot
1673 are distributed across NUMA nodes. Defaults on
1674 for 64-bit NUMA, off otherwise.
1675 Format: 0 | 1 (for off | on)
1677 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1679 hd= [EIDE] (E)IDE hard drive subsystem geometry
1680 Format: <cyl>,<head>,<sect>
1683 Disable Hardware Error Source Table (HEST) support;
1684 corresponding firmware-first mode error processing
1685 logic will be disabled.
1687 hibernate= [HIBERNATION]
1688 noresume Don't check if there's a hibernation image
1689 present during boot.
1690 nocompress Don't compress/decompress hibernation images.
1691 no Disable hibernation and resume.
1692 protect_image Turn on image protection during restoration
1693 (that will set all pages holding image data
1694 during restoration read-only).
1696 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1697 size of <nn>. This works even on boxes that have no
1698 highmem otherwise. This also works to reduce highmem
1699 size on bigger boxes.
1701 highres= [KNL] Enable/disable high resolution timer mode.
1702 Valid parameters: "on", "off"
1707 hostname= [KNL] Set the hostname (aka UTS nodename).
1709 This allows setting the system's hostname during early
1710 startup. This sets the name returned by gethostname.
1711 Using this parameter to set the hostname makes it
1712 possible to ensure the hostname is correctly set before
1713 any userspace processes run, avoiding the possibility
1714 that a process may call gethostname before the hostname
1715 has been explicitly set, resulting in the calling
1716 process getting an incorrect result. The string must
1717 not exceed the maximum allowed hostname length (usually
1718 64 characters) and will be truncated otherwise.
1720 hpet= [X86-32,HPET] option to control HPET usage
1721 Format: { enable (default) | disable | force |
1723 disable: disable HPET and use PIT instead
1724 force: allow force enabled of undocumented chips (ICH4,
1726 verbose: show contents of HPET registers during setup
1728 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1729 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1731 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1732 If this follows hugepagesz (below), it specifies
1733 the number of pages of hugepagesz to be allocated.
1734 If this is the first HugeTLB parameter on the command
1735 line, it specifies the number of pages to allocate for
1736 the default huge page size. If using node format, the
1737 number of pages to allocate per-node can be specified.
1738 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1739 Format: <integer> or (node format)
1740 <node>:<integer>[,<node>:<integer>]
1743 [HW] The size of the HugeTLB pages. This is used in
1744 conjunction with hugepages (above) to allocate huge
1745 pages of a specific size at boot. The pair
1746 hugepagesz=X hugepages=Y can be specified once for
1747 each supported huge page size. Huge page sizes are
1748 architecture dependent. See also
1749 Documentation/admin-guide/mm/hugetlbpage.rst.
1752 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1753 of gigantic hugepages. Or using node format, the size
1754 of a CMA area per node can be specified.
1755 Format: nn[KMGTPE] or (node format)
1756 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1758 Reserve a CMA area of given size and allocate gigantic
1759 hugepages using the CMA allocator. If enabled, the
1760 boot-time allocation of gigantic hugepages is skipped.
1762 hugetlb_free_vmemmap=
1763 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1765 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1766 Allows heavy hugetlb users to free up some more
1767 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1768 Format: { on | off (default) }
1773 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1776 Note that the vmemmap pages may be allocated from the added
1777 memory block itself when memory_hotplug.memmap_on_memory is
1778 enabled, those vmemmap pages cannot be optimized even if this
1779 feature is enabled. Other vmemmap pages not allocated from
1780 the added memory block itself do not be affected.
1783 [KNL] Should the hung task detector generate panics.
1786 A value of 1 instructs the kernel to panic when a
1787 hung task is detected. The default value is controlled
1788 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1789 option. The value selected by this boot parameter can
1790 be changed later by the kernel.hung_task_panic sysctl.
1792 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1793 terminal devices. Valid values: 0..8
1794 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1795 If specified, z/VM IUCV HVC accepts connections
1796 from listed z/VM user IDs only.
1798 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1799 which allow the hypervisor to 'idle' the
1800 guest on lock contention.
1802 i2c_bus= [HW] Override the default board specific I2C bus speed
1803 or register an additional I2C bus that is not
1804 registered from board initialization code.
1808 i8042.debug [HW] Toggle i8042 debug mode
1809 i8042.unmask_kbd_data
1810 [HW] Enable printing of interrupt data from the KBD port
1811 (disabled by default, and as a pre-condition
1812 requires that i8042.debug=1 be enabled)
1813 i8042.direct [HW] Put keyboard port into non-translated mode
1814 i8042.dumbkbd [HW] Pretend that controller can only read data from
1815 keyboard and cannot control its state
1816 (Don't attempt to blink the leds)
1817 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1818 i8042.nokbd [HW] Don't check/create keyboard port
1819 i8042.noloop [HW] Disable the AUX Loopback command while probing
1821 i8042.nomux [HW] Don't check presence of an active multiplexing
1823 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1825 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1826 i8042.reset [HW] Reset the controller during init, cleanup and
1827 suspend-to-ram transitions, only during s2r
1828 transitions, or never reset
1829 Format: { 1 | Y | y | 0 | N | n }
1830 1, Y, y: always reset controller
1831 0, N, n: don't ever reset controller
1832 Default: only on s2r transitions on x86; most other
1833 architectures force reset to be always executed
1834 i8042.unlock [HW] Unlock (ignore) the keylock
1835 i8042.kbdreset [HW] Reset device connected to KBD port
1837 [HW] Allow deferred probing upon i8042 probe errors
1841 i915.invert_brightness=
1842 [DRM] Invert the sense of the variable that is used to
1843 set the brightness of the panel backlight. Normally a
1844 brightness value of 0 indicates backlight switched off,
1845 and the maximum of the brightness value sets the backlight
1846 to maximum brightness. If this parameter is set to 0
1847 (default) and the machine requires it, or this parameter
1848 is set to 1, a brightness value of 0 sets the backlight
1849 to maximum brightness, and the maximum of the brightness
1850 value switches the backlight off.
1851 -1 -- never invert brightness
1852 0 -- machine default
1853 1 -- force brightness inversion
1856 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1860 Format: idle=poll, idle=halt, idle=nomwait
1861 Poll forces a polling idle loop that can slightly
1862 improve the performance of waking up a idle CPU, but
1863 will use a lot of power and make the system run hot.
1865 idle=halt: Halt is forced to be used for CPU idle.
1866 In such case C2/C3 won't be used again.
1867 idle=nomwait: Disable mwait for CPU C-states
1871 Allow force disabling of Shared Virtual Memory (SVA)
1872 support for the idxd driver. By default it is set to
1875 idxd.tc_override= [HW]
1877 Allow override of default traffic class configuration
1878 for the device. By default it is set to false (0).
1880 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1881 Format: { strict | legacy | 2008 | relaxed }
1884 Choose which programs will be accepted for execution
1885 based on the IEEE 754 NaN encoding(s) supported by
1886 the FPU and the NaN encoding requested with the value
1887 of an ELF file header flag individually set by each
1888 binary. Hardware implementations are permitted to
1889 support either or both of the legacy and the 2008 NaN
1892 Available settings are as follows:
1893 strict accept binaries that request a NaN encoding
1894 supported by the FPU
1895 legacy only accept legacy-NaN binaries, if supported
1897 2008 only accept 2008-NaN binaries, if supported
1899 relaxed accept any binaries regardless of whether
1900 supported by the FPU
1902 The FPU emulator is always able to support both NaN
1903 encodings, so if no FPU hardware is present or it has
1904 been disabled with 'nofpu', then the settings of
1905 'legacy' and '2008' strap the emulator accordingly,
1906 'relaxed' straps the emulator for both legacy-NaN and
1907 2008-NaN, whereas 'strict' enables legacy-NaN only on
1908 legacy processors and both NaN encodings on MIPS32 or
1911 The setting for ABS.fmt/NEG.fmt instruction execution
1912 mode generally follows that for the NaN encoding,
1913 except where unsupported by hardware.
1915 ignore_loglevel [KNL]
1916 Ignore loglevel setting - this will print /all/
1917 kernel messages to the console. Useful for debugging.
1918 We also add it as printk module parameter, so users
1919 could change it dynamically, usually by
1920 /sys/module/printk/parameters/ignore_loglevel.
1923 Ignore RLIMIT_DATA setting for data mappings,
1924 print warning at first misuse. Can be changed via
1925 /sys/module/kernel/parameters/ignore_rlimit_data.
1927 ihash_entries= [KNL]
1928 Set number of hash buckets for inode cache.
1930 ima_appraise= [IMA] appraise integrity measurements
1931 Format: { "off" | "enforce" | "fix" | "log" }
1934 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1935 The builtin appraise policy appraises all files
1938 ima_canonical_fmt [IMA]
1939 Use the canonical format for the binary runtime
1940 measurements, instead of host native format.
1943 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1947 The list of supported hash algorithms is defined
1948 in crypto/hash_info.h.
1951 The builtin policies to load during IMA setup.
1952 Format: "tcb | appraise_tcb | secure_boot |
1953 fail_securely | critical_data"
1955 The "tcb" policy measures all programs exec'd, files
1956 mmap'd for exec, and all files opened with the read
1957 mode bit set by either the effective uid (euid=0) or
1960 The "appraise_tcb" policy appraises the integrity of
1961 all files owned by root.
1963 The "secure_boot" policy appraises the integrity
1964 of files (eg. kexec kernel image, kernel modules,
1965 firmware, policy, etc) based on file signatures.
1967 The "fail_securely" policy forces file signature
1968 verification failure also on privileged mounted
1969 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1972 The "critical_data" policy measures kernel integrity
1975 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1976 Load a policy which meets the needs of the Trusted
1977 Computing Base. This means IMA will measure all
1978 programs exec'd, files mmap'd for exec, and all files
1979 opened for read by uid=0.
1982 Select one of defined IMA measurements template formats.
1983 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
1988 [IMA] Define a custom template format.
1989 Format: { "field1|...|fieldN" }
1991 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1992 Format: <min_file_size>
1993 Set the minimal file size for using asynchronous hash.
1994 If left unspecified, ahash usage is disabled.
1996 ahash performance varies for different data sizes on
1997 different crypto accelerators. This option can be used
1998 to achieve the best performance for a particular HW.
2000 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
2002 Set hashing buffer size. Default: 4k.
2004 ahash performance varies for different chunk sizes on
2005 different crypto accelerators. This option can be used
2006 to achieve best performance for particular HW.
2010 Run specified binary instead of /sbin/init as init
2013 initcall_debug [KNL] Trace initcalls as they are executed. Useful
2014 for working out where the kernel is dying during
2017 initcall_blacklist= [KNL] Do not execute a comma-separated list of
2018 initcall functions. Useful for debugging built-in
2019 modules and initcalls.
2021 initramfs_async= [KNL]
2024 This parameter controls whether the initramfs
2025 image is unpacked asynchronously, concurrently
2026 with devices being probed and
2027 initialized. This should normally just work,
2028 but as a debugging aid, one can get the
2029 historical behaviour of the initramfs
2030 unpacking being completed before device_ and
2033 initrd= [BOOT] Specify the location of the initial ramdisk
2035 initrdmem= [KNL] Specify a physical address and size from which to
2036 load the initrd. If an initrd is compiled in or
2037 specified in the bootparams, it takes priority over this
2039 Format: ss[KMG],nn[KMG]
2042 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
2045 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2047 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
2049 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2051 init_pkru= [X86] Specify the default memory protection keys rights
2052 register contents for all processes. 0x55555554 by
2053 default (disallow access to all but pkey 0). Can
2054 override in debugfs after boot.
2056 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2059 int_pln_enable [X86] Enable power limit notification interrupt
2061 integrity_audit=[IMA]
2062 Format: { "0" | "1" }
2063 0 -- basic integrity auditing messages. (Default)
2064 1 -- additional integrity auditing messages.
2066 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2068 Enable intel iommu driver.
2070 Disable intel iommu driver.
2071 igfx_off [Default Off]
2072 By default, gfx is mapped as normal device. If a gfx
2073 device has a dedicated DMAR unit, the DMAR unit is
2074 bypassed by not enabling DMAR with this option. In
2075 this case, gfx device will use physical address for
2077 strict [Default Off]
2078 Deprecated, equivalent to iommu.strict=1.
2079 sp_off [Default Off]
2080 By default, super page will be supported if Intel IOMMU
2081 has the capability. With this option, super page will
2084 Enable the Intel IOMMU scalable mode if the hardware
2085 advertises that it has support for the scalable mode
2088 Disallow use of the Intel IOMMU scalable mode.
2089 tboot_noforce [Default Off]
2090 Do not force the Intel IOMMU enabled under tboot.
2091 By default, tboot will force Intel IOMMU on, which
2092 could harm performance of some high-throughput
2093 devices like 40GBit network cards, even if identity
2095 Note that using this option lowers the security
2096 provided by tboot because it makes the system
2097 vulnerable to DMA attacks.
2099 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2100 0 disables intel_idle and fall back on acpi_idle.
2101 1 to 9 specify maximum depth of C-state.
2105 Do not enable intel_pstate as the default
2106 scaling driver for the supported processors
2108 Use intel_pstate as a scaling driver, but configure it
2109 to work with generic cpufreq governors (instead of
2110 enabling its internal governor). This mode cannot be
2111 used along with the hardware-managed P-states (HWP)
2114 Enable intel_pstate on systems that prohibit it by default
2115 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2116 instead of acpi-cpufreq may disable platform features, such
2117 as thermal controls and power capping, that rely on ACPI
2118 P-States information being indicated to OSPM and therefore
2119 should be used with caution. This option does not work with
2120 processors that aren't supported by the intel_pstate driver
2121 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2123 Do not enable hardware P state control (HWP)
2126 Only load intel_pstate on systems which support
2127 hardware P state control (HWP) if available.
2129 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2130 Description Table, specifies preferred power management
2131 profile as "Enterprise Server" or "Performance Server",
2132 then this feature is turned on by default.
2134 Allow per-logical-CPU P-State performance control limits using
2135 cpufreq sysfs interface
2137 intremap= [X86-64, Intel-IOMMU]
2138 on enable Interrupt Remapping (default)
2139 off disable Interrupt Remapping
2140 nosid disable Source ID checking
2142 BIOS x2APIC opt-out request will be ignored
2143 nopost disable Interrupt Posting
2145 iomem= Disable strict checking of access to MMIO memory
2146 strict regions from userspace.
2161 nobypass [PPC/POWERNV]
2162 Disable IOMMU bypass, using IOMMU for PCI devices.
2164 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2165 Format: { "0" | "1" }
2166 0 - Try to allocate a 32-bit DMA address first, before
2167 falling back to the full range if needed.
2168 1 - Allocate directly from the full usable range,
2169 forcing Dual Address Cycle for PCI cards supporting
2170 greater than 32-bit addressing.
2172 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2173 Format: { "0" | "1" }
2175 Request that DMA unmap operations use deferred
2176 invalidation of hardware TLBs, for increased
2177 throughput at the cost of reduced device isolation.
2178 Will fall back to strict mode if not supported by
2179 the relevant IOMMU driver.
2181 DMA unmap operations invalidate IOMMU hardware TLBs
2183 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2184 Note: on x86, strict mode specified via one of the
2185 legacy driver-specific options takes precedence.
2188 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2189 Format: { "0" | "1" }
2190 0 - Use IOMMU translation for DMA.
2191 1 - Bypass the IOMMU for DMA.
2192 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2194 io7= [HW] IO7 for Marvel-based Alpha systems
2195 See comment before marvel_specify_io7 in
2196 arch/alpha/kernel/core_marvel.c.
2198 io_delay= [X86] I/O delay method
2200 Standard port 0x80 based delay
2202 Alternate port 0xed based delay (needed on some systems)
2204 Simple two microseconds delay
2209 See Documentation/admin-guide/nfs/nfsroot.rst.
2211 ipcmni_extend [KNL] Extend the maximum number of unique System V
2212 IPC identifiers from 32,768 to 16,777,216.
2214 irqaffinity= [SMP] Set the default irq affinity mask
2215 The argument is a cpu list, as described above.
2217 irqchip.gicv2_force_probe=
2220 Force the kernel to look for the second 4kB page
2221 of a GICv2 controller even if the memory range
2222 exposed by the device tree is too small.
2224 irqchip.gicv3_nolpi=
2226 Force the kernel to ignore the availability of
2227 LPIs (and by consequence ITSs). Intended for system
2228 that use the kernel as a bootloader, and thus want
2229 to let secondary kernels in charge of setting up
2232 irqchip.gicv3_pseudo_nmi= [ARM64]
2233 Enables support for pseudo-NMIs in the kernel. This
2234 requires the kernel to be built with
2235 CONFIG_ARM64_PSEUDO_NMI.
2238 When an interrupt is not handled search all handlers
2239 for it. Intended to get systems with badly broken
2243 When an interrupt is not handled search all handlers
2244 for it. Also check all handlers each timer
2245 interrupt. Intended to get systems with badly broken
2249 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2251 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2252 [Deprecated - use cpusets instead]
2253 Format: [flag-list,]<cpu-list>
2255 Specify one or more CPUs to isolate from disturbances
2256 specified in the flag list (default: domain):
2259 Disable the tick when a single task runs.
2261 A residual 1Hz tick is offloaded to workqueues, which you
2262 need to affine to housekeeping through the global
2263 workqueue's affinity configured via the
2264 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2265 by using the 'domain' flag described below.
2267 NOTE: by default the global workqueue runs on all CPUs,
2268 so to protect individual CPUs the 'cpumask' file has to
2269 be configured manually after bootup.
2272 Isolate from the general SMP balancing and scheduling
2273 algorithms. Note that performing domain isolation this way
2274 is irreversible: it's not possible to bring back a CPU to
2275 the domains once isolated through isolcpus. It's strongly
2276 advised to use cpusets instead to disable scheduler load
2277 balancing through the "cpuset.sched_load_balance" file.
2278 It offers a much more flexible interface where CPUs can
2279 move in and out of an isolated set anytime.
2281 You can move a process onto or off an "isolated" CPU via
2282 the CPU affinity syscalls or cpuset.
2283 <cpu number> begins at 0 and the maximum value is
2284 "number of CPUs in system - 1".
2288 Isolate from being targeted by managed interrupts
2289 which have an interrupt mask containing isolated
2290 CPUs. The affinity of managed interrupts is
2291 handled by the kernel and cannot be changed via
2292 the /proc/irq/* interfaces.
2294 This isolation is best effort and only effective
2295 if the automatically assigned interrupt mask of a
2296 device queue contains isolated and housekeeping
2297 CPUs. If housekeeping CPUs are online then such
2298 interrupts are directed to the housekeeping CPU
2299 so that IO submitted on the housekeeping CPU
2300 cannot disturb the isolated CPU.
2302 If a queue's affinity mask contains only isolated
2303 CPUs then this parameter has no effect on the
2304 interrupt routing decision, though interrupts are
2305 only delivered when tasks running on those
2306 isolated CPUs submit IO. IO submitted on
2307 housekeeping CPUs has no influence on those
2310 The format of <cpu-list> is described above.
2314 ivrs_ioapic [HW,X86-64]
2315 Provide an override to the IOAPIC-ID<->DEVICE-ID
2316 mapping provided in the IVRS ACPI table.
2317 By default, PCI segment is 0, and can be omitted.
2319 For example, to map IOAPIC-ID decimal 10 to
2320 PCI segment 0x1 and PCI device 00:14.0,
2321 write the parameter as:
2322 ivrs_ioapic=10@0001:00:14.0
2325 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2326 write the parameter as:
2327 ivrs_ioapic[10]=00:14.0
2328 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2329 PCI device 00:14.0 write the parameter as:
2330 ivrs_ioapic[10]=0001:00:14.0
2332 ivrs_hpet [HW,X86-64]
2333 Provide an override to the HPET-ID<->DEVICE-ID
2334 mapping provided in the IVRS ACPI table.
2335 By default, PCI segment is 0, and can be omitted.
2337 For example, to map HPET-ID decimal 10 to
2338 PCI segment 0x1 and PCI device 00:14.0,
2339 write the parameter as:
2340 ivrs_hpet=10@0001:00:14.0
2343 * To map HPET-ID decimal 0 to PCI device 00:14.0
2344 write the parameter as:
2345 ivrs_hpet[0]=00:14.0
2346 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2347 PCI device 00:14.0 write the parameter as:
2348 ivrs_ioapic[10]=0001:00:14.0
2350 ivrs_acpihid [HW,X86-64]
2351 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2352 mapping provided in the IVRS ACPI table.
2353 By default, PCI segment is 0, and can be omitted.
2355 For example, to map UART-HID:UID AMD0020:0 to
2356 PCI segment 0x1 and PCI device ID 00:14.5,
2357 write the parameter as:
2358 ivrs_acpihid=AMD0020:0@0001:00:14.5
2361 * To map UART-HID:UID AMD0020:0 to PCI segment is 0,
2362 PCI device ID 00:14.5, write the parameter as:
2363 ivrs_acpihid[00:14.5]=AMD0020:0
2364 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and
2365 PCI device ID 00:14.5, write the parameter as:
2366 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2368 js= [HW,JOY] Analog joystick
2369 See Documentation/input/joydev/joystick.rst.
2372 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2373 report on every invalid memory access. Without this
2374 parameter KASAN will print report only for the first
2378 Do not unregister boot console at start. This is only
2379 useful for debugging when something happens in the window
2380 between unregistering the boot console and initializing
2385 kernelcore= [KNL,X86,IA-64,PPC]
2386 Format: nn[KMGTPE] | nn% | "mirror"
2387 This parameter specifies the amount of memory usable by
2388 the kernel for non-movable allocations. The requested
2389 amount is spread evenly throughout all nodes in the
2390 system as ZONE_NORMAL. The remaining memory is used for
2391 movable memory in its own zone, ZONE_MOVABLE. In the
2392 event, a node is too small to have both ZONE_NORMAL and
2393 ZONE_MOVABLE, kernelcore memory will take priority and
2394 other nodes will have a larger ZONE_MOVABLE.
2396 ZONE_MOVABLE is used for the allocation of pages that
2397 may be reclaimed or moved by the page migration
2398 subsystem. Note that allocations like PTEs-from-HighMem
2399 still use the HighMem zone if it exists, and the Normal
2400 zone if it does not.
2402 It is possible to specify the exact amount of memory in
2403 the form of "nn[KMGTPE]", a percentage of total system
2404 memory in the form of "nn%", or "mirror". If "mirror"
2405 option is specified, mirrored (reliable) memory is used
2406 for non-movable allocations and remaining memory is used
2407 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2408 are exclusive, so you cannot specify multiple forms.
2410 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2411 Format: <Controller#>[,poll interval]
2412 The controller # is the number of the ehci usb debug
2413 port as it is probed via PCI. The poll interval is
2414 optional and is the number seconds in between
2415 each poll cycle to the debug port in case you need
2416 the functionality for interrupting the kernel with
2417 gdb or control-c on the dbgp connection. When
2418 not using this parameter you use sysrq-g to break into
2419 the kernel debugger.
2421 kgdboc= [KGDB,HW] kgdb over consoles.
2422 Requires a tty driver that supports console polling,
2423 or a supported polling keyboard driver (non-usb).
2424 Serial only format: <serial_device>[,baud]
2425 keyboard only format: kbd
2426 keyboard and serial format: kbd,<serial_device>[,baud]
2427 Optional Kernel mode setting:
2428 kms, kbd format: kms,kbd
2429 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2431 kgdboc_earlycon= [KGDB,HW]
2432 If the boot console provides the ability to read
2433 characters and can work in polling mode, you can use
2434 this parameter to tell kgdb to use it as a backend
2435 until the normal console is registered. Intended to
2436 be used together with the kgdboc parameter which
2437 specifies the normal console to transition to.
2439 The name of the early console should be specified
2440 as the value of this parameter. Note that the name of
2441 the early console might be different than the tty
2442 name passed to kgdboc. It's OK to leave the value
2443 blank and the first boot console that implements
2444 read() will be picked.
2446 kgdbwait [KGDB] Stop kernel execution and enter the
2447 kernel debugger at the earliest opportunity.
2449 kmac= [MIPS] Korina ethernet MAC address.
2450 Configure the RouterBoard 532 series on-chip
2451 Ethernet adapter MAC address.
2453 kmemleak= [KNL] Boot-time kmemleak enable/disable
2454 Valid arguments: on, off
2456 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2459 kprobe_event=[probe-list]
2460 [FTRACE] Add kprobe events and enable at boot time.
2461 The probe-list is a semicolon delimited list of probe
2462 definitions. Each definition is same as kprobe_events
2463 interface, but the parameters are comma delimited.
2464 For example, to add a kprobe event on vfs_read with
2465 arg1 and arg2, add to the command line;
2467 kprobe_event=p,vfs_read,$arg1,$arg2
2469 See also Documentation/trace/kprobetrace.rst "Kernel
2470 Boot Parameter" section.
2472 kpti= [ARM64] Control page table isolation of user
2473 and kernel address spaces.
2474 Default: enabled on cores which need mitigation.
2478 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires
2479 CONFIG_KUNIT to be set to be fully enabled. The
2480 default value can be overridden via
2481 KUNIT_DEFAULT_ENABLED.
2482 Default is 1 (enabled)
2484 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2485 Default is 0 (don't ignore, but inject #GP)
2487 kvm.eager_page_split=
2488 [KVM,X86] Controls whether or not KVM will try to
2489 proactively split all huge pages during dirty logging.
2490 Eager page splitting reduces interruptions to vCPU
2491 execution by eliminating the write-protection faults
2492 and MMU lock contention that would otherwise be
2493 required to split huge pages lazily.
2495 VM workloads that rarely perform writes or that write
2496 only to a small region of VM memory may benefit from
2497 disabling eager page splitting to allow huge pages to
2498 still be used for reads.
2500 The behavior of eager page splitting depends on whether
2501 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2502 disabled, all huge pages in a memslot will be eagerly
2503 split when dirty logging is enabled on that memslot. If
2504 enabled, eager page splitting will be performed during
2505 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2508 Eager page splitting is only supported when kvm.tdp_mmu=Y.
2512 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2513 Default is false (don't support).
2516 [KVM] Controls the software workaround for the
2517 X86_BUG_ITLB_MULTIHIT bug.
2518 force : Always deploy workaround.
2519 off : Never deploy workaround.
2520 auto : Deploy workaround based on the presence of
2521 X86_BUG_ITLB_MULTIHIT.
2525 If the software workaround is enabled for the host,
2526 guests do need not to enable it for nested guests.
2528 kvm.nx_huge_pages_recovery_ratio=
2529 [KVM] Controls how many 4KiB pages are periodically zapped
2530 back to huge pages. 0 disables the recovery, otherwise if
2531 the value is N KVM will zap 1/Nth of the 4KiB pages every
2532 period (see below). The default is 60.
2534 kvm.nx_huge_pages_recovery_period_ms=
2535 [KVM] Controls the time period at which KVM zaps 4KiB pages
2536 back to huge pages. If the value is a non-zero N, KVM will
2537 zap a portion (see ratio above) of the pages every N msecs.
2538 If the value is 0 (the default), KVM will pick a period based
2539 on the ratio, such that a page is zapped after 1 hour on average.
2541 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2542 Default is 1 (enabled)
2544 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2546 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2549 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2551 none: Forcefully disable KVM.
2553 nvhe: Standard nVHE-based mode, without support for
2556 protected: nVHE-based mode with support for guests whose
2557 state is kept private from the host.
2559 nested: VHE-based mode with support for nested
2560 virtualization. Requires at least ARMv8.3
2563 Defaults to VHE/nVHE based on hardware support. Setting
2564 mode to "protected" will disable kexec and hibernation
2565 for the host. "nested" is experimental and should be
2566 used with extreme caution.
2568 kvm-arm.vgic_v3_group0_trap=
2569 [KVM,ARM] Trap guest accesses to GICv3 group-0
2572 kvm-arm.vgic_v3_group1_trap=
2573 [KVM,ARM] Trap guest accesses to GICv3 group-1
2576 kvm-arm.vgic_v3_common_trap=
2577 [KVM,ARM] Trap guest accesses to GICv3 common
2580 kvm-arm.vgic_v4_enable=
2581 [KVM,ARM] Allow use of GICv4 for direct injection of
2584 kvm_cma_resv_ratio=n [PPC]
2585 Reserves given percentage from system memory area for
2586 contiguous memory allocation for KVM hash pagetable
2588 By default it reserves 5% of total system memory.
2592 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2593 (virtualized MMU) support on capable Intel chips.
2594 Default is 1 (enabled)
2596 kvm-intel.emulate_invalid_guest_state=
2597 [KVM,Intel] Disable emulation of invalid guest state.
2598 Ignored if kvm-intel.enable_unrestricted_guest=1, as
2599 guest state is never invalid for unrestricted guests.
2600 This param doesn't apply to nested guests (L2), as KVM
2601 never emulates invalid L2 guest state.
2602 Default is 1 (enabled)
2604 kvm-intel.flexpriority=
2605 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2606 Default is 1 (enabled)
2609 [KVM,Intel] Enable VMX nesting (nVMX).
2610 Default is 0 (disabled)
2612 kvm-intel.unrestricted_guest=
2613 [KVM,Intel] Disable unrestricted guest feature
2614 (virtualized real and unpaged mode) on capable
2615 Intel chips. Default is 1 (enabled)
2617 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2620 Valid arguments: never, cond, always
2622 always: L1D cache flush on every VMENTER.
2623 cond: Flush L1D on VMENTER only when the code between
2624 VMEXIT and VMENTER can leak host memory.
2625 never: Disables the mitigation
2627 Default is cond (do L1 cache flush in specific instances)
2629 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2630 feature (tagged TLBs) on capable Intel chips.
2631 Default is 1 (enabled)
2633 l1d_flush= [X86,INTEL]
2634 Control mitigation for L1D based snooping vulnerability.
2636 Certain CPUs are vulnerable to an exploit against CPU
2637 internal buffers which can forward information to a
2638 disclosure gadget under certain conditions.
2640 In vulnerable processors, the speculatively
2641 forwarded data can be used in a cache side channel
2642 attack, to access data to which the attacker does
2643 not have direct access.
2645 This parameter controls the mitigation. The
2648 on - enable the interface for the mitigation
2650 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2653 The kernel PTE inversion protection is unconditionally
2654 enabled and cannot be disabled.
2657 Provides all available mitigations for the
2658 L1TF vulnerability. Disables SMT and
2659 enables all mitigations in the
2660 hypervisors, i.e. unconditional L1D flush.
2662 SMT control and L1D flush control via the
2663 sysfs interface is still possible after
2664 boot. Hypervisors will issue a warning
2665 when the first VM is started in a
2666 potentially insecure configuration,
2667 i.e. SMT enabled or L1D flush disabled.
2670 Same as 'full', but disables SMT and L1D
2671 flush runtime control. Implies the
2672 'nosmt=force' command line option.
2673 (i.e. sysfs control of SMT is disabled.)
2676 Leaves SMT enabled and enables the default
2677 hypervisor mitigation, i.e. conditional
2680 SMT control and L1D flush control via the
2681 sysfs interface is still possible after
2682 boot. Hypervisors will issue a warning
2683 when the first VM is started in a
2684 potentially insecure configuration,
2685 i.e. SMT enabled or L1D flush disabled.
2689 Disables SMT and enables the default
2690 hypervisor mitigation.
2692 SMT control and L1D flush control via the
2693 sysfs interface is still possible after
2694 boot. Hypervisors will issue a warning
2695 when the first VM is started in a
2696 potentially insecure configuration,
2697 i.e. SMT enabled or L1D flush disabled.
2700 Same as 'flush', but hypervisors will not
2701 warn when a VM is started in a potentially
2702 insecure configuration.
2705 Disables hypervisor mitigations and doesn't
2707 It also drops the swap size and available
2708 RAM limit restriction on both hypervisor and
2713 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2719 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2722 lapic= [X86,APIC] Do not use TSC deadline
2723 value for LAPIC timer one-shot implementation. Default
2724 back to the programmable timer unit in the LAPIC.
2725 Format: notscdeadline
2727 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2730 libata.dma= [LIBATA] DMA control
2731 libata.dma=0 Disable all PATA and SATA DMA
2732 libata.dma=1 PATA and SATA Disk DMA only
2733 libata.dma=2 ATAPI (CDROM) DMA only
2734 libata.dma=4 Compact Flash DMA only
2735 Combinations also work, so libata.dma=3 enables DMA
2736 for disks and CDROMs, but not CFs.
2738 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2739 libata.ignore_hpa=0 keep BIOS limits (default)
2740 libata.ignore_hpa=1 ignore limits, using full disk
2742 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2746 libata.force= [LIBATA] Force configurations. The format is a comma-
2747 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2748 PORT and DEVICE are decimal numbers matching port, link
2749 or device. Basically, it matches the ATA ID string
2750 printed on console by libata. If the whole ID part is
2751 omitted, the last PORT and DEVICE values are used. If
2752 ID hasn't been specified yet, the configuration applies
2753 to all ports, links and devices.
2755 If only DEVICE is omitted, the parameter applies to
2756 the port and all links and devices behind it. DEVICE
2757 number of 0 either selects the first device or the
2758 first fan-out link behind PMP device. It does not
2759 select the host link. DEVICE number of 15 selects the
2760 host link and device attached to it.
2762 The VAL specifies the configuration to force. As long
2763 as there is no ambiguity, shortcut notation is allowed.
2764 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2765 The following configurations can be forced.
2767 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2768 Any ID with matching PORT is used.
2770 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2772 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2773 udma[/][16,25,33,44,66,100,133] notation is also
2776 * nohrst, nosrst, norst: suppress hard, soft and both
2779 * rstonce: only attempt one reset during hot-unplug
2782 * [no]dbdelay: Enable or disable the extra 200ms delay
2783 before debouncing a link PHY and device presence
2786 * [no]ncq: Turn on or off NCQ.
2788 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2790 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2792 * [no]trim: Enable or disable (unqueued) TRIM.
2794 * trim_zero: Indicate that TRIM command zeroes data.
2796 * max_trim_128m: Set 128M maximum trim size limit.
2798 * [no]dma: Turn on or off DMA transfers.
2800 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2802 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2803 commands that are not a multiple of 16 bytes.
2805 * [no]dmalog: Enable or disable the use of the
2806 READ LOG DMA EXT command to access logs.
2808 * [no]iddevlog: Enable or disable access to the
2809 identify device data log.
2811 * [no]logdir: Enable or disable access to the general
2812 purpose log directory.
2814 * max_sec_128: Set transfer size limit to 128 sectors.
2816 * max_sec_1024: Set or clear transfer size limit to
2819 * max_sec_lba48: Set or clear transfer size limit to
2822 * [no]lpm: Enable or disable link power management.
2824 * [no]setxfer: Indicate if transfer speed mode setting
2827 * [no]fua: Disable or enable FUA (Force Unit Access)
2828 support for devices supporting this feature.
2830 * dump_id: Dump IDENTIFY data.
2832 * disable: Disable this device.
2834 If there are multiple matching configurations changing
2835 the same attribute, the last one is used.
2837 load_ramdisk= [RAM] [Deprecated]
2839 lockd.nlm_grace_period=P [NFS] Assign grace period.
2842 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2845 lockd.nlm_timeout=T [NFS] Assign timeout value.
2848 lockd.nlm_udpport=M [NFS] Assign UDP port.
2851 lockdown= [SECURITY]
2852 { integrity | confidentiality }
2853 Enable the kernel lockdown feature. If set to
2854 integrity, kernel features that allow userland to
2855 modify the running kernel are disabled. If set to
2856 confidentiality, kernel features that allow userland
2857 to extract confidential information from the kernel
2860 locktorture.nreaders_stress= [KNL]
2861 Set the number of locking read-acquisition kthreads.
2862 Defaults to being automatically set based on the
2863 number of online CPUs.
2865 locktorture.nwriters_stress= [KNL]
2866 Set the number of locking write-acquisition kthreads.
2868 locktorture.onoff_holdoff= [KNL]
2869 Set time (s) after boot for CPU-hotplug testing.
2871 locktorture.onoff_interval= [KNL]
2872 Set time (s) between CPU-hotplug operations, or
2873 zero to disable CPU-hotplug testing.
2875 locktorture.shuffle_interval= [KNL]
2876 Set task-shuffle interval (jiffies). Shuffling
2877 tasks allows some CPUs to go into dyntick-idle
2878 mode during the locktorture test.
2880 locktorture.shutdown_secs= [KNL]
2881 Set time (s) after boot system shutdown. This
2882 is useful for hands-off automated testing.
2884 locktorture.stat_interval= [KNL]
2885 Time (s) between statistics printk()s.
2887 locktorture.stutter= [KNL]
2888 Time (s) to stutter testing, for example,
2889 specifying five seconds causes the test to run for
2890 five seconds, wait for five seconds, and so on.
2891 This tests the locking primitive's ability to
2892 transition abruptly to and from idle.
2894 locktorture.torture_type= [KNL]
2895 Specify the locking implementation to test.
2897 locktorture.verbose= [KNL]
2898 Enable additional printk() statements.
2900 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2903 loglevel= All Kernel Messages with a loglevel smaller than the
2904 console loglevel will be printed to the console. It can
2905 also be changed with klogd or other programs. The
2906 loglevels are defined as follows:
2908 0 (KERN_EMERG) system is unusable
2909 1 (KERN_ALERT) action must be taken immediately
2910 2 (KERN_CRIT) critical conditions
2911 3 (KERN_ERR) error conditions
2912 4 (KERN_WARNING) warning conditions
2913 5 (KERN_NOTICE) normal but significant condition
2914 6 (KERN_INFO) informational
2915 7 (KERN_DEBUG) debug-level messages
2917 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2918 in bytes. n must be a power of two and greater
2919 than the minimal size. The minimal size is defined
2920 by LOG_BUF_SHIFT kernel config parameter. There is
2921 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2922 that allows to increase the default size depending on
2923 the number of CPUs. See init/Kconfig for more details.
2925 logo.nologo [FB] Disables display of the built-in Linux logo.
2926 This may be used to provide more screen space for
2927 kernel log messages and is useful when debugging
2928 kernel boot problems.
2930 lp=0 [LP] Specify parallel ports to use, e.g,
2931 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2932 lp=reset first parallel port). 'lp=0' disables the
2933 lp=auto printer driver. 'lp=reset' (which can be
2934 specified in addition to the ports) causes
2935 attached printers to be reset. Using
2936 lp=port1,port2,... specifies the parallel ports
2937 to associate lp devices with, starting with
2938 lp0. A port specification may be 'none' to skip
2939 that lp device, or a parport name such as
2940 'parport0'. Specifying 'lp=auto' instead of a
2941 port specification list means that device IDs
2942 from each port should be examined, to see if
2943 an IEEE 1284-compliant printer is attached; if
2944 so, the driver will manage that printer.
2945 See also header of drivers/char/lp.c.
2948 Sets loops_per_jiffy to given constant, thus avoiding
2949 time-consuming boot-time autodetection (up to 250 ms per
2950 CPU). 0 enables autodetection (default). To determine
2951 the correct value for your kernel, boot with normal
2952 autodetection and see what value is printed. Note that
2953 on SMP systems the preset will be applied to all CPUs,
2954 which is likely to cause problems if your CPUs need
2955 significantly divergent settings. An incorrect value
2956 will cause delays in the kernel to be wrong, leading to
2957 unpredictable I/O errors and other breakage. Although
2958 unlikely, in the extreme case this might damage your
2962 Format: <io>,<irq>,<dma>
2964 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2967 [SECURITY] Choose order of LSM initialization. This
2968 overrides CONFIG_LSM, and the "security=" parameter.
2970 machvec= [IA-64] Force the use of a particular machine-vector
2971 (machvec) in a generic kernel.
2972 Example: machvec=hpzx1
2974 machtype= [Loongson] Share the same kernel image file between
2975 different yeeloong laptops.
2976 Example: machtype=lemote-yeeloong-2f-7inch
2978 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
2979 than or equal to this physical address is ignored.
2981 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2982 will bring up during bootup. maxcpus=n : n >= 0 limits
2983 the kernel to bring up 'n' processors. Surely after
2984 bootup you can bring up the other plugged cpu by executing
2985 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2986 only takes effect during system bootup.
2987 While n=0 is a special case, it is equivalent to "nosmp",
2988 which also disables the IO APIC.
2990 max_loop= [LOOP] The number of loop block devices that get
2991 (loop.max_loop) unconditionally pre-created at init time. The default
2992 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2993 of statically allocating a predefined number, loop
2994 devices can be requested on-demand with the
2995 /dev/loop-control interface.
2997 mce [X86-32] Machine Check Exception
2999 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
3001 md= [HW] RAID subsystems devices and level
3002 See Documentation/admin-guide/md.rst.
3005 Format: <first>,<last>
3006 Specifies range of consoles to be captured by the MDA.
3009 Control mitigation for the Micro-architectural Data
3010 Sampling (MDS) vulnerability.
3012 Certain CPUs are vulnerable to an exploit against CPU
3013 internal buffers which can forward information to a
3014 disclosure gadget under certain conditions.
3016 In vulnerable processors, the speculatively
3017 forwarded data can be used in a cache side channel
3018 attack, to access data to which the attacker does
3019 not have direct access.
3021 This parameter controls the MDS mitigation. The
3024 full - Enable MDS mitigation on vulnerable CPUs
3025 full,nosmt - Enable MDS mitigation and disable
3026 SMT on vulnerable CPUs
3027 off - Unconditionally disable MDS mitigation
3029 On TAA-affected machines, mds=off can be prevented by
3030 an active TAA mitigation as both vulnerabilities are
3031 mitigated with the same mechanism so in order to disable
3032 this mitigation, you need to specify tsx_async_abort=off
3035 Not specifying this option is equivalent to
3038 For details see: Documentation/admin-guide/hw-vuln/mds.rst
3040 mem=nn[KMG] [HEXAGON] Set the memory size.
3041 Must be specified, otherwise memory size will be 0.
3043 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
3044 Amount of memory to be used in cases as follows:
3047 2 when the kernel is not able to see the whole system memory;
3048 3 memory that lies after 'mem=' boundary is excluded from
3049 the hypervisor, then assigned to KVM guests.
3050 4 to limit the memory available for kdump kernel.
3052 [ARC,MICROBLAZE] - the limit applies only to low memory,
3053 high memory is not affected.
3055 [ARM64] - only limits memory covered by the linear
3056 mapping. The NOMAP regions are not affected.
3058 [X86] Work as limiting max address. Use together
3059 with memmap= to avoid physical address space collisions.
3060 Without memmap= PCI devices could be placed at addresses
3061 belonging to unused RAM.
3063 Note that this only takes effects during boot time since
3064 in above case 3, memory may need be hot added after boot
3065 if system memory of hypervisor is not sufficient.
3068 [ARM,MIPS] - override the memory layout reported by
3070 Define a memory region of size nn[KMG] starting at
3072 Multiple different regions can be specified with
3073 multiple mem= parameters on the command line.
3075 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3078 memblock=debug [KNL] Enable memblock debug messages.
3081 [KNL,SH] Allow user to override the default size for
3082 per-device physically contiguous DMA buffers.
3084 memhp_default_state=online/offline
3085 [KNL] Set the initial state for the memory hotplug
3086 onlining policy. If not specified, the default value is
3087 set according to the
3088 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3090 See Documentation/admin-guide/mm/memory-hotplug.rst.
3092 memmap=exactmap [KNL,X86] Enable setting of an exact
3093 E820 memory map, as specified by the user.
3094 Such memmap=exactmap lines can be constructed based on
3095 BIOS output or other requirements. See the memmap=nn@ss
3098 memmap=nn[KMG]@ss[KMG]
3099 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
3100 Region of memory to be used is from ss to ss+nn.
3101 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3102 which limits max address to nn[KMG].
3103 Multiple different regions can be specified,
3106 memmap=100M@2G,100M#3G,1G!1024G
3108 memmap=nn[KMG]#ss[KMG]
3109 [KNL,ACPI] Mark specific memory as ACPI data.
3110 Region of memory to be marked is from ss to ss+nn.
3112 memmap=nn[KMG]$ss[KMG]
3113 [KNL,ACPI] Mark specific memory as reserved.
3114 Region of memory to be reserved is from ss to ss+nn.
3115 Example: Exclude memory from 0x18690000-0x1869ffff
3116 memmap=64K$0x18690000
3118 memmap=0x10000$0x18690000
3119 Some bootloaders may need an escape character before '$',
3120 like Grub2, otherwise '$' and the following number
3123 memmap=nn[KMG]!ss[KMG]
3124 [KNL,X86] Mark specific memory as protected.
3125 Region of memory to be used, from ss to ss+nn.
3126 The memory region may be marked as e820 type 12 (0xc)
3127 and is NVDIMM or ADR memory.
3129 memmap=<size>%<offset>-<oldtype>+<newtype>
3130 [KNL,ACPI] Convert memory within the specified region
3131 from <oldtype> to <newtype>. If "-<oldtype>" is left
3132 out, the whole region will be marked as <newtype>,
3133 even if previously unavailable. If "+<newtype>" is left
3134 out, matching memory will be removed. Types are
3135 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3136 3 = ACPI, 12 = PRAM.
3138 memory_corruption_check=0/1 [X86]
3139 Some BIOSes seem to corrupt the first 64k of
3140 memory when doing things like suspend/resume.
3141 Setting this option will scan the memory
3142 looking for corruption. Enabling this will
3143 both detect corruption and prevent the kernel
3144 from using the memory being corrupted.
3145 However, its intended as a diagnostic tool; if
3146 repeatable BIOS-originated corruption always
3147 affects the same memory, you can use memmap=
3148 to prevent the kernel from using that memory.
3150 memory_corruption_check_size=size [X86]
3151 By default it checks for corruption in the low
3152 64k, making this memory unavailable for normal
3153 use. Use this parameter to scan for
3154 corruption in more or less memory.
3156 memory_corruption_check_period=seconds [X86]
3157 By default it checks for corruption every 60
3158 seconds. Use this parameter to check at some
3159 other rate. 0 disables periodic checking.
3161 memory_hotplug.memmap_on_memory
3162 [KNL,X86,ARM] Boolean flag to enable this feature.
3163 Format: {on | off (default)}
3164 When enabled, runtime hotplugged memory will
3165 allocate its internal metadata (struct pages,
3166 those vmemmap pages cannot be optimized even
3167 if hugetlb_free_vmemmap is enabled) from the
3168 hotadded memory which will allow to hotadd a
3169 lot of memory without requiring additional
3171 This feature is disabled by default because it
3172 has some implication on large (e.g. GB)
3173 allocations in some configurations (e.g. small
3175 The state of the flag can be read in
3176 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3177 Note that even when enabled, there are a few cases where
3178 the feature is not effective.
3180 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3182 default : 0 <disable>
3183 Specifies the number of memtest passes to be
3184 performed. Each pass selects another test
3185 pattern from a given set of patterns. Memtest
3186 fills the memory with this pattern, validates
3187 memory contents and reserves bad memory
3188 regions that are detected.
3190 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3191 Valid arguments: on, off
3192 Default (depends on kernel configuration option):
3193 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3194 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3195 mem_encrypt=on: Activate SME
3196 mem_encrypt=off: Do not activate SME
3198 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3199 for details on when memory encryption can be activated.
3201 mem_sleep_default= [SUSPEND] Default system suspend mode:
3202 s2idle - Suspend-To-Idle
3203 shallow - Power-On Suspend or equivalent (if supported)
3204 deep - Suspend-To-RAM or equivalent (if supported)
3205 See Documentation/admin-guide/pm/sleep-states.rst.
3207 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3208 Multi-Function General Purpose Timers on AMD Geode
3211 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3212 the BIOS has incorrectly applied a workaround. TinyBIOS
3213 version 0.98 is known to be affected, 0.99 fixes the
3214 problem by letting the user disable the workaround.
3218 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3219 physical address is ignored.
3221 mini2440= [ARM,HW,KNL]
3222 Format:[0..2][b][c][t]
3224 MINI2440 configuration specification:
3225 0 - The attached screen is the 3.5" TFT
3226 1 - The attached screen is the 7" TFT
3227 2 - The VGA Shield is attached (1024x768)
3228 Leaving out the screen size parameter will not load
3229 the TFT driver, and the framebuffer will be left
3231 b - Enable backlight. The TFT backlight pin will be
3232 linked to the kernel VESA blanking code and a GPIO
3233 LED. This parameter is not necessary when using the
3235 c - Enable the s3c camera interface.
3236 t - Reserved for enabling touchscreen support. The
3237 touchscreen support is not enabled in the mainstream
3238 kernel as of 2.6.30, a preliminary port can be found
3239 in the "bleeding edge" mini2440 support kernel at
3240 https://repo.or.cz/w/linux-2.6/mini2440.git
3243 [X86,PPC,S390,ARM64] Control optional mitigations for
3244 CPU vulnerabilities. This is a set of curated,
3245 arch-independent options, each of which is an
3246 aggregation of existing arch-specific options.
3249 Disable all optional CPU mitigations. This
3250 improves system performance, but it may also
3251 expose users to several CPU vulnerabilities.
3252 Equivalent to: nopti [X86,PPC]
3253 if nokaslr then kpti=0 [ARM64]
3254 nospectre_v1 [X86,PPC]
3256 nospectre_v2 [X86,PPC,S390,ARM64]
3257 spectre_v2_user=off [X86]
3258 spec_store_bypass_disable=off [X86,PPC]
3259 ssbd=force-off [ARM64]
3260 nospectre_bhb [ARM64]
3263 tsx_async_abort=off [X86]
3264 kvm.nx_huge_pages=off [X86]
3265 srbds=off [X86,INTEL]
3266 no_entry_flush [PPC]
3267 no_uaccess_flush [PPC]
3268 mmio_stale_data=off [X86]
3272 This does not have any effect on
3273 kvm.nx_huge_pages when
3274 kvm.nx_huge_pages=force.
3277 Mitigate all CPU vulnerabilities, but leave SMT
3278 enabled, even if it's vulnerable. This is for
3279 users who don't want to be surprised by SMT
3280 getting disabled across kernel upgrades, or who
3281 have other ways of avoiding SMT-based attacks.
3282 Equivalent to: (default behavior)
3285 Mitigate all CPU vulnerabilities, disabling SMT
3286 if needed. This is for users who always want to
3287 be fully mitigated, even if it means losing SMT.
3288 Equivalent to: l1tf=flush,nosmt [X86]
3289 mds=full,nosmt [X86]
3290 tsx_async_abort=full,nosmt [X86]
3291 mmio_stale_data=full,nosmt [X86]
3292 retbleed=auto,nosmt [X86]
3295 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3296 parameter allows control of the logging verbosity for
3297 the additional memory initialisation checks. A value
3298 of 0 disables mminit logging and a level of 4 will
3299 log everything. Information is printed at KERN_DEBUG
3300 so loglevel=8 may also need to be specified.
3303 [X86,INTEL] Control mitigation for the Processor
3304 MMIO Stale Data vulnerabilities.
3306 Processor MMIO Stale Data is a class of
3307 vulnerabilities that may expose data after an MMIO
3308 operation. Exposed data could originate or end in
3309 the same CPU buffers as affected by MDS and TAA.
3310 Therefore, similar to MDS and TAA, the mitigation
3311 is to clear the affected CPU buffers.
3313 This parameter controls the mitigation. The
3316 full - Enable mitigation on vulnerable CPUs
3318 full,nosmt - Enable mitigation and disable SMT on
3321 off - Unconditionally disable mitigation
3323 On MDS or TAA affected machines,
3324 mmio_stale_data=off can be prevented by an active
3325 MDS or TAA mitigation as these vulnerabilities are
3326 mitigated with the same mechanism so in order to
3327 disable this mitigation, you need to specify
3328 mds=off and tsx_async_abort=off too.
3330 Not specifying this option is equivalent to
3331 mmio_stale_data=full.
3334 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3336 <module>.async_probe[=<bool>] [KNL]
3337 If no <bool> value is specified or if the value
3338 specified is not a valid <bool>, enable asynchronous
3339 probe on this module. Otherwise, enable/disable
3340 asynchronous probe on this module as indicated by the
3341 <bool> value. See also: module.async_probe
3343 module.async_probe=<bool>
3344 [KNL] When set to true, modules will use async probing
3345 by default. To enable/disable async probing for a
3346 specific module, use the module specific control that
3347 is documented under <module>.async_probe. When both
3348 module.async_probe and <module>.async_probe are
3349 specified, <module>.async_probe takes precedence for
3350 the specific module.
3353 [KNL] When CONFIG_MODULE_SIG is set, this means that
3354 modules without (valid) signatures will fail to load.
3355 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3356 is always true, so this option does nothing.
3358 module_blacklist= [KNL] Do not load a comma-separated list of
3359 modules. Useful for debugging problem modules.
3362 [MOUSE] Maximum time between finger touching and
3363 leaving touchpad surface for touch to be considered
3364 a tap and be reported as a left button click (for
3365 touchpads working in absolute mode only).
3367 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3368 reporting absolute coordinates, such as tablets
3369 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3370 reporting absolute coordinates, such as tablets
3372 movablecore= [KNL,X86,IA-64,PPC]
3373 Format: nn[KMGTPE] | nn%
3374 This parameter is the complement to kernelcore=, it
3375 specifies the amount of memory used for migratable
3376 allocations. If both kernelcore and movablecore is
3377 specified, then kernelcore will be at *least* the
3378 specified value but may be more. If movablecore on its
3379 own is specified, the administrator must be careful
3380 that the amount of memory usable for all allocations
3383 movable_node [KNL] Boot-time switch to make hotplugable memory
3384 NUMA nodes to be movable. This means that the memory
3385 of such nodes will be usable only for movable
3386 allocations which rules out almost all kernel
3387 allocations. Use with caution!
3389 MTD_Partition= [MTD]
3390 Format: <name>,<region-number>,<size>,<offset>
3392 MTD_Region= [MTD] Format:
3393 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3396 See drivers/mtd/parsers/cmdlinepart.c
3399 ARM/S3C2412 JIVE boot control
3401 See arch/arm/mach-s3c/mach-jive.c
3403 mtouchusb.raw_coordinates=
3404 [HW] Make the MicroTouch USB driver use raw coordinates
3405 ('y', default) or cooked coordinates ('n')
3407 mtrr_chunk_size=nn[KMG] [X86]
3408 used for mtrr cleanup. It is largest continuous chunk
3409 that could hold holes aka. UC entries.
3411 mtrr_gran_size=nn[KMG] [X86]
3412 Used for mtrr cleanup. It is granularity of mtrr block.
3414 Large value could prevent small alignment from
3417 mtrr_spare_reg_nr=n [X86]
3419 Range: 0,7 : spare reg number
3421 Used for mtrr cleanup. It is spare mtrr entries number.
3422 Set to 2 or more if your graphical card needs more.
3424 multitce=off [PPC] This parameter disables the use of the pSeries
3425 firmware feature for updating multiple TCE entries
3428 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3430 netdev= [NET] Network devices parameters
3431 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3432 Note that mem_start is often overloaded to mean
3433 something different and driver-specific.
3434 This usage is only documented in each driver source
3437 netpoll.carrier_timeout=
3438 [NET] Specifies amount of time (in seconds) that
3439 netpoll should wait for a carrier. By default netpoll
3443 [NETFILTER] Enable connection tracking flow accounting
3444 0 to disable accounting
3445 1 to enable accounting
3449 [NFS] sets the pathname to the program which is used
3450 to update the NFS client cache entries.
3452 nfs.cache_getent_timeout=
3453 [NFS] sets the timeout after which an attempt to
3454 update a cache entry is deemed to have failed.
3456 nfs.callback_nr_threads=
3457 [NFSv4] set the total number of threads that the
3458 NFS client will assign to service NFSv4 callback
3461 nfs.callback_tcpport=
3462 [NFS] set the TCP port on which the NFSv4 callback
3463 channel should listen.
3466 [NFS] enable 64-bit inode numbers.
3467 If zero, the NFS client will fake up a 32-bit inode
3468 number for the readdir() and stat() syscalls instead
3469 of returning the full 64-bit number.
3470 The default is to return 64-bit inode numbers.
3472 nfs.idmap_cache_timeout=
3473 [NFS] set the maximum lifetime for idmapper cache
3476 nfs.max_session_cb_slots=
3477 [NFSv4.1] Sets the maximum number of session
3478 slots the client will assign to the callback
3479 channel. This determines the maximum number of
3480 callbacks the client will process in parallel for
3481 a particular server.
3483 nfs.max_session_slots=
3484 [NFSv4.1] Sets the maximum number of session slots
3485 the client will attempt to negotiate with the server.
3486 This limits the number of simultaneous RPC requests
3487 that the client can send to the NFSv4.1 server.
3488 Note that there is little point in setting this
3489 value higher than the max_tcp_slot_table_limit.
3491 nfs.nfs4_disable_idmapping=
3492 [NFSv4] When set to the default of '1', this option
3493 ensures that both the RPC level authentication
3494 scheme and the NFS level operations agree to use
3495 numeric uids/gids if the mount is using the
3496 'sec=sys' security flavour. In effect it is
3497 disabling idmapping, which can make migration from
3498 legacy NFSv2/v3 systems to NFSv4 easier.
3499 Servers that do not support this mode of operation
3500 will be autodetected by the client, and it will fall
3501 back to using the idmapper.
3502 To turn off this behaviour, set the value to '0'.
3505 [NFS4] Specify an additional fixed unique ident-
3506 ification string that NFSv4 clients can insert into
3507 their nfs_client_id4 string. This is typically a
3508 UUID that is generated at system install time.
3510 nfs.recover_lost_locks=
3511 [NFSv4] Attempt to recover locks that were lost due
3512 to a lease timeout on the server. Please note that
3513 doing this risks data corruption, since there are
3514 no guarantees that the file will remain unchanged
3515 after the locks are lost.
3516 If you want to enable the kernel legacy behaviour of
3517 attempting to recover these locks, then set this
3519 The default parameter value of '0' causes the kernel
3520 not to attempt recovery of lost locks.
3522 nfs.send_implementation_id=
3523 [NFSv4.1] Send client implementation identification
3524 information in exchange_id requests.
3525 If zero, no implementation identification information
3527 The default is to send the implementation identification
3530 nfs4.layoutstats_timer=
3531 [NFSv4.2] Change the rate at which the kernel sends
3532 layoutstats to the pNFS metadata server.
3534 Setting this to value to 0 causes the kernel to use
3535 whatever value is the default set by the layout
3536 driver. A non-zero value sets the minimum interval
3537 in seconds between layoutstats transmissions.
3539 nfsd.inter_copy_offload_enable=
3540 [NFSv4.2] When set to 1, the server will support
3541 server-to-server copies for which this server is
3542 the destination of the copy.
3544 nfsd.nfs4_disable_idmapping=
3545 [NFSv4] When set to the default of '1', the NFSv4
3546 server will return only numeric uids and gids to
3547 clients using auth_sys, and will accept numeric uids
3548 and gids from such clients. This is intended to ease
3549 migration from NFSv2/v3.
3551 nfsd.nfsd4_ssc_umount_timeout=
3552 [NFSv4.2] When used as the destination of a
3553 server-to-server copy, knfsd temporarily mounts
3554 the source server. It caches the mount in case
3555 it will be needed again, and discards it if not
3556 used for the number of milliseconds specified by
3559 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3560 See Documentation/admin-guide/nfs/nfsroot.rst.
3562 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3563 See Documentation/admin-guide/nfs/nfsroot.rst.
3565 nfsrootdebug [NFS] enable nfsroot debugging messages.
3566 See Documentation/admin-guide/nfs/nfsroot.rst.
3568 nmi_backtrace.backtrace_idle [KNL]
3569 Dump stacks even of idle CPUs in response to an
3570 NMI stack-backtrace request.
3572 nmi_debug= [KNL,SH] Specify one or more actions to take
3573 when a NMI is triggered.
3574 Format: [state][,regs][,debounce][,die]
3576 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3577 Format: [panic,][nopanic,][num]
3579 0 - turn hardlockup detector in nmi_watchdog off
3580 1 - turn hardlockup detector in nmi_watchdog on
3581 When panic is specified, panic when an NMI watchdog
3582 timeout occurs (or 'nopanic' to not panic on an NMI
3583 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3584 To disable both hard and soft lockup detectors,
3585 please see 'nowatchdog'.
3586 This is useful when you use a panic=... timeout and
3587 need the box quickly up again.
3589 These settings can be accessed at runtime via
3590 the nmi_watchdog and hardlockup_panic sysctls.
3592 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3593 emulation library even if a 387 maths coprocessor
3596 no5lvl [X86-64] Disable 5-level paging mode. Forces
3597 kernel to use 4-level paging instead.
3599 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3600 caches in the slab allocator. Saves per-node memory,
3601 but will impact performance.
3605 noaltinstr [S390] Disables alternative instructions patching
3606 (CPU alternatives feature).
3608 noapic [SMP,APIC] Tells the kernel to not make use of any
3609 IOAPICs that may be present in the system.
3611 noautogroup Disable scheduler automatic task group creation.
3616 [HW] Never suspend the console
3617 Disable suspending of consoles during suspend and
3618 hibernate operations. Once disabled, debugging
3619 messages can reach various consoles while the rest
3620 of the system is being put to sleep (ie, while
3621 debugging driver suspend/resume hooks). This may
3622 not work reliably with all consoles, but is known
3623 to work with serial and VGA consoles.
3624 To facilitate more flexible debugging, we also add
3625 console_suspend, a printk module parameter to control
3626 it. Users could use console_suspend (usually
3627 /sys/module/printk/parameters/console_suspend) to
3628 turn on/off it dynamically.
3631 [KNL] Disable object debugging
3633 nodsp [SH] Disable hardware DSP at boot time.
3635 noefi Disable EFI runtime services support.
3637 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3642 This affects only 32-bit executables.
3643 noexec32=on: enable non-executable mappings (default)
3644 read doesn't imply executable mappings
3645 noexec32=off: disable non-executable mappings
3646 read implies executable mappings
3648 no_file_caps Tells the kernel not to honor file capabilities. The
3649 only way then for a file to be executed with privilege
3650 is to be setuid root or executed by root.
3652 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3654 nofsgsbase [X86] Disables FSGSBASE instructions.
3656 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3657 register save and restore. The kernel will only save
3658 legacy floating-point registers on task switch.
3660 nohalt [IA-64] Tells the kernel not to use the power saving
3661 function PAL_HALT_LIGHT when idle. This increases
3662 power-consumption. On the positive side, it reduces
3663 interrupt wake-up latency, which may improve performance
3664 in certain environments such as networked servers or
3668 Force pointers printed to the console or buffers to be
3669 unhashed. By default, when a pointer is printed via %p
3670 format string, that pointer is "hashed", i.e. obscured
3671 by hashing the pointer value. This is a security feature
3672 that hides actual kernel addresses from unprivileged
3673 users, but it also makes debugging the kernel more
3674 difficult since unequal pointers can no longer be
3675 compared. However, if this command-line option is
3676 specified, then all normal pointers will have their true
3677 value printed. This option should only be specified when
3678 debugging the kernel. Please do not use on production
3681 nohibernate [HIBERNATION] Disable hibernation and resume.
3683 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3684 in do_idle() and not use the arch_cpu_idle()
3685 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3686 to be effective. This is useful on platforms where the
3687 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3688 correctly or when doing power measurements to evaluate
3689 the impact of the sleep instructions. This is also
3690 useful when using JTAG debugger.
3692 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3694 nohugevmalloc [KNL,X86,PPC,ARM64] Disable kernel huge vmalloc mappings.
3696 nohz= [KNL] Boottime enable/disable dynamic ticks
3697 Valid arguments: on, off
3700 nohz_full= [KNL,BOOT,SMP,ISOL]
3701 The argument is a cpu list, as described above.
3702 In kernels built with CONFIG_NO_HZ_FULL=y, set
3703 the specified list of CPUs whose tick will be stopped
3704 whenever possible. The boot CPU will be forced outside
3705 the range to maintain the timekeeping. Any CPUs
3706 in this list will have their RCU callbacks offloaded,
3707 just as if they had also been called out in the
3708 rcu_nocbs= boot parameter.
3710 Note that this argument takes precedence over
3711 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3713 noinitrd [RAM] Tells the kernel not to load any configured
3716 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3718 [Deprecated - use intremap=off]
3722 noinvpcid [X86] Disable the INVPCID cpu feature.
3724 noiotrap [SH] Disables trapped I/O port accesses.
3726 noirqdebug [X86-32] Disables the code which attempts to detect and
3727 disable unhandled interrupt sources.
3729 noisapnp [ISAPNP] Disables ISA PnP code.
3731 nojitter [IA-64] Disables jitter checking for ITC timers.
3734 When CONFIG_RANDOMIZE_BASE is set, this disables
3735 kernel and module base offset ASLR (Address Space
3736 Layout Randomization).
3738 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3741 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3743 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3745 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3747 nomca [IA-64] Disable machine check abort handling
3749 nomce [X86-32] Disable Machine Check Exception
3751 nomfgpt [X86-32] Disable Multi-Function General Purpose
3752 Timer usage (for AMD Geode machines).
3754 nomodeset Disable kernel modesetting. Most systems' firmware
3755 sets up a display mode and provides framebuffer memory
3756 for output. With nomodeset, DRM and fbdev drivers will
3757 not load if they could possibly displace the pre-
3758 initialized output. Only the system framebuffer will
3759 be available for use. The respective drivers will not
3760 perform display-mode changes or accelerated rendering.
3762 Useful as error fallback, or for testing and debugging.
3764 nomodule Disable module load
3766 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3767 shutdown the other cpus. Instead use the REBOOT_VECTOR
3770 nopat [X86] Disable PAT (page attribute table extension of
3771 pagetables) support.
3773 nopcid [X86-64] Disable the PCID cpu feature.
3775 nopku [X86] Disable Memory Protection Keys CPU feature found
3779 Equivalent to pti=off
3781 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
3782 Disables the PV optimizations forcing the guest to run
3783 as generic guest with no PV drivers. Currently support
3784 XEN HVM, KVM, HYPER_V and VMWARE guest.
3786 nopvspin [X86,XEN,KVM]
3787 Disables the qspinlock slow path using PV optimizations
3788 which allow the hypervisor to 'idle' the guest on lock
3791 norandmaps Don't use address space randomization. Equivalent to
3792 echo 0 > /proc/sys/kernel/randomize_va_space
3794 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3795 with UP alternatives
3797 noresume [SWSUSP] Disables resume and restores original swap
3802 no-scroll [VGA] Disables scrollback.
3803 This is required for the Braillex ib80-piezo Braille
3804 reader made by F.H. Papenmeier (Germany).
3806 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3809 Disable SMAP (Supervisor Mode Access Prevention)
3810 even if it is supported by processor.
3813 Disable SMEP (Supervisor Mode Execution Prevention)
3814 even if it is supported by processor.
3816 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3817 and disable the IO APIC. legacy for "maxcpus=0".
3819 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3820 Equivalent to smt=1.
3822 [KNL,X86] Disable symmetric multithreading (SMT).
3823 nosmt=force: Force disable SMT, cannot be undone
3824 via the sysfs control file.
3826 nosoftlockup [KNL] Disable the soft-lockup detector.
3828 nospec_store_bypass_disable
3829 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3831 nospectre_bhb [ARM64] Disable all mitigations for Spectre-BHB (branch
3832 history injection) vulnerability. System may allow data leaks
3835 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3836 (bounds check bypass). With this option data leaks are
3837 possible in the system.
3839 nospectre_v2 [X86,PPC_E500,ARM64] Disable all mitigations for
3840 the Spectre variant 2 (indirect branch prediction)
3841 vulnerability. System may allow data leaks with this
3844 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES] Disable paravirtualized
3845 steal time accounting. steal time is computed, but
3846 won't influence scheduler behaviour
3848 nosync [HW,M68K] Disables sync negotiation for all devices.
3850 no_timer_check [X86,APIC] Disables the code which tests for
3851 broken timer IRQ sources.
3854 [PPC] Don't flush the L1-D cache after accessing user data.
3856 novmcoredd [KNL,KDUMP]
3857 Disable device dump. Device dump allows drivers to
3858 append dump data to vmcore so you can collect driver
3859 specified debug info. Drivers can append the data
3860 without any limit and this data is stored in memory,
3861 so this may cause significant memory stress. Disabling
3862 device dump can help save memory but the driver debug
3863 data will be no longer available. This parameter
3864 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3868 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3869 clock and use the default one.
3871 nowatchdog [KNL] Disable both lockup detectors, i.e.
3872 soft-lockup and NMI watchdog (hard-lockup).
3876 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3878 NOTE: this parameter will be ignored on systems with the
3879 LEGACY_XAPIC_DISABLED bit set in the
3880 IA32_XAPIC_DISABLE_STATUS MSR.
3882 noxsave [BUGS=X86] Disables x86 extended register state save
3883 and restore using xsave. The kernel will fallback to
3884 enabling legacy floating-point and sse state.
3886 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3887 register states. The kernel will fall back to use
3888 xsave to save the states. By using this parameter,
3889 performance of saving the states is degraded because
3890 xsave doesn't support modified optimization while
3891 xsaveopt supports it on xsaveopt enabled systems.
3893 noxsaves [X86] Disables xsaves and xrstors used in saving and
3894 restoring x86 extended register state in compacted
3895 form of xsave area. The kernel will fall back to use
3896 xsaveopt and xrstor to save and restore the states
3897 in standard form of xsave area. By using this
3898 parameter, xsave area per process might occupy more
3899 memory on xsaves enabled systems.
3901 nps_mtm_hs_ctr= [KNL,ARC]
3902 This parameter sets the maximum duration, in
3903 cycles, each HW thread of the CTOP can run
3904 without interruptions, before HW switches it.
3905 The actual maximum duration is 16 times this
3907 Format: integer between 1 and 255
3910 nptcg= [IA-64] Override max number of concurrent global TLB
3911 purges which is reported from either PAL_VM_SUMMARY or
3914 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3915 could support. nr_cpus=n : n >= 1 limits the kernel to
3916 support 'n' processors. It could be larger than the
3917 number of already plugged CPU during bootup, later in
3918 runtime you can physically add extra cpu until it reaches
3919 n. So during boot up some boot time memory for per-cpu
3920 variables need be pre-allocated for later physical cpu
3923 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3925 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3926 set up a single NUMA node spanning all memory.
3928 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3930 Allowed values are enable and disable
3932 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3933 'node', 'default' can be specified
3934 This can be set from sysctl after boot.
3935 See Documentation/admin-guide/sysctl/vm.rst for details.
3937 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3938 See Documentation/core-api/debugging-via-ohci1394.rst for more
3941 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3942 Rather than timing out after 20 ms if an EC
3943 command is not properly ACKed, override the length
3944 of the timeout. We have interrupts disabled while
3945 waiting for the ACK, so if this is set too high
3946 interrupts *may* be lost!
3948 omap_mux= [OMAP] Override bootloader pin multiplexing.
3949 Format: <mux_mode0.mode_name=value>...
3950 For example, to override I2C bus2:
3951 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3953 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3955 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3957 boundary - index of last SLC block on Flex-OneNAND.
3958 The remaining blocks are configured as MLC blocks.
3959 lock - Configure if Flex-OneNAND boundary should be locked.
3960 Once locked, the boundary cannot be changed.
3961 1 indicates lock status, 0 indicates unlock status.
3963 oops=panic Always panic on oopses. Default is to just kill the
3964 process, but there is a small probability of
3965 deadlocking the machine.
3966 This will also cause panics on machine check exceptions.
3967 Useful together with panic=30 to trigger a reboot.
3970 [KNL] Boolean flag to control whether the page allocator
3971 should randomize its free lists. The randomization may
3972 be automatically enabled if the kernel detects it is
3973 running on a platform with a direct-mapped memory-side
3974 cache, and this parameter can be used to
3975 override/disable that behavior. The state of the flag
3976 can be read from sysfs at:
3977 /sys/module/page_alloc/parameters/shuffle.
3979 page_owner= [KNL] Boot-time page_owner enabling option.
3980 Storage of the information about who allocated
3981 each page is disabled in default. With this switch,
3983 on: enable the feature
3985 page_poison= [KNL] Boot-time parameter changing the state of
3986 poisoning on the buddy allocator, available with
3987 CONFIG_PAGE_POISONING=y.
3988 off: turn off poisoning (default)
3989 on: turn on poisoning
3991 page_reporting.page_reporting_order=
3992 [KNL] Minimal page reporting order
3994 Adjust the minimal page reporting order. The page
3995 reporting is disabled when it exceeds (MAX_ORDER-1).
3997 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3998 timeout > 0: seconds before rebooting
3999 timeout = 0: wait forever
4000 timeout < 0: reboot immediately
4003 panic_print= Bitmask for printing system info when panic happens.
4004 User can chose combination of the following bits:
4005 bit 0: print all tasks info
4006 bit 1: print system memory info
4007 bit 2: print timer info
4008 bit 3: print locks info if CONFIG_LOCKDEP is on
4009 bit 4: print ftrace buffer
4010 bit 5: print all printk messages in buffer
4011 bit 6: print all CPUs backtrace (if available in the arch)
4012 *Be aware* that this option may print a _lot_ of lines,
4013 so there are risks of losing older messages in the log.
4014 Use this option carefully, maybe worth to setup a
4015 bigger log buffer with "log_buf_len" along with this.
4017 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
4018 Format: <hex>[,nousertaint]
4019 Hexadecimal bitmask representing the set of TAINT flags
4020 that will cause the kernel to panic when add_taint() is
4021 called with any of the flags in this set.
4022 The optional switch "nousertaint" can be utilized to
4023 prevent userspace forced crashes by writing to sysctl
4024 /proc/sys/kernel/tainted any flagset matching with the
4025 bitmask set on panic_on_taint.
4026 See Documentation/admin-guide/tainted-kernels.rst for
4027 extra details on the taint flags that users can pick
4028 to compose the bitmask to assign to panic_on_taint.
4030 panic_on_warn panic() instead of WARN(). Useful to cause kdump
4033 parkbd.port= [HW] Parallel port number the keyboard adapter is
4034 connected to, default is 0.
4036 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
4037 0 for XT, 1 for AT (default is AT).
4040 parport= [HW,PPT] Specify parallel ports. 0 disables.
4041 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
4042 Use 'auto' to force the driver to use any
4043 IRQ/DMA settings detected (the default is to
4044 ignore detected IRQ/DMA settings because of
4045 possible conflicts). You can specify the base
4046 address, IRQ, and DMA settings; IRQ and DMA
4047 should be numbers, or 'auto' (for using detected
4048 settings on that particular port), or 'nofifo'
4049 (to avoid using a FIFO even if it is detected).
4050 Parallel ports are assigned in the order they
4051 are specified on the command line, starting
4054 parport_init_mode= [HW,PPT]
4055 Configure VIA parallel port to operate in
4056 a specific mode. This is necessary on Pegasos
4057 computer where firmware has no options for setting
4058 up parallel port mode and sets it to spp.
4059 Currently this function knows 686a and 8231 chips.
4060 Format: [spp|ps2|epp|ecp|ecpepp]
4062 pata_legacy.all= [HW,LIBATA]
4064 Set to non-zero to probe primary and secondary ISA
4065 port ranges on PCI systems where no PCI PATA device
4066 has been found at either range. Disabled by default.
4068 pata_legacy.autospeed= [HW,LIBATA]
4070 Set to non-zero if a chip is present that snoops speed
4071 changes. Disabled by default.
4073 pata_legacy.ht6560a= [HW,LIBATA]
4075 Set to 1, 2, or 3 for HT 6560A on the primary channel,
4076 the secondary channel, or both channels respectively.
4077 Disabled by default.
4079 pata_legacy.ht6560b= [HW,LIBATA]
4081 Set to 1, 2, or 3 for HT 6560B on the primary channel,
4082 the secondary channel, or both channels respectively.
4083 Disabled by default.
4085 pata_legacy.iordy_mask= [HW,LIBATA]
4087 IORDY enable mask. Set individual bits to allow IORDY
4088 for the respective channel. Bit 0 is for the first
4089 legacy channel handled by this driver, bit 1 is for
4090 the second channel, and so on. The sequence will often
4091 correspond to the primary legacy channel, the secondary
4092 legacy channel, and so on, but the handling of a PCI
4093 bus and the use of other driver options may interfere
4094 with the sequence. By default IORDY is allowed across
4097 pata_legacy.opti82c46x= [HW,LIBATA]
4099 Set to 1, 2, or 3 for Opti 82c611A on the primary
4100 channel, the secondary channel, or both channels
4101 respectively. Disabled by default.
4103 pata_legacy.opti82c611a= [HW,LIBATA]
4105 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4106 channel, the secondary channel, or both channels
4107 respectively. Disabled by default.
4109 pata_legacy.pio_mask= [HW,LIBATA]
4111 PIO mode mask for autospeed devices. Set individual
4112 bits to allow the use of the respective PIO modes.
4113 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4114 All modes allowed by default.
4116 pata_legacy.probe_all= [HW,LIBATA]
4118 Set to non-zero to probe tertiary and further ISA
4119 port ranges on PCI systems. Disabled by default.
4121 pata_legacy.probe_mask= [HW,LIBATA]
4123 Probe mask for legacy ISA PATA ports. Depending on
4124 platform configuration and the use of other driver
4125 options up to 6 legacy ports are supported: 0x1f0,
4126 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4127 of individual ports can be disabled by setting the
4128 corresponding bits in the mask to 1. Bit 0 is for
4129 the first port in the list above (0x1f0), and so on.
4130 By default all supported ports are probed.
4132 pata_legacy.qdi= [HW,LIBATA]
4134 Set to non-zero to probe QDI controllers. By default
4135 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4137 pata_legacy.winbond= [HW,LIBATA]
4139 Set to non-zero to probe Winbond controllers. Use
4140 the standard I/O port (0x130) if 1, otherwise the
4141 value given is the I/O port to use (typically 0x1b0).
4142 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4145 pata_platform.pio_mask= [HW,LIBATA]
4147 Supported PIO mode mask. Set individual bits to allow
4148 the use of the respective PIO modes. Bit 0 is for
4149 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4153 Halt all CPUs after the first oops has been printed for
4154 the specified number of seconds. This is to be used if
4155 your oopses keep scrolling off the screen.
4159 pci=option[,option...] [PCI] various PCI subsystem options.
4161 Some options herein operate on a specific device
4162 or a set of devices (<pci_dev>). These are
4163 specified in one of the following formats:
4165 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4166 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4168 Note: the first format specifies a PCI
4169 bus/device/function address which may change
4170 if new hardware is inserted, if motherboard
4171 firmware changes, or due to changes caused
4172 by other kernel parameters. If the
4173 domain is left unspecified, it is
4174 taken to be zero. Optionally, a path
4175 to a device through multiple device/function
4176 addresses can be specified after the base
4177 address (this is more robust against
4178 renumbering issues). The second format
4179 selects devices using IDs from the
4180 configuration space which may match multiple
4181 devices in the system.
4183 earlydump dump PCI config space before the kernel
4185 off [X86] don't probe for the PCI bus
4186 bios [X86-32] force use of PCI BIOS, don't access
4187 the hardware directly. Use this if your machine
4188 has a non-standard PCI host bridge.
4189 nobios [X86-32] disallow use of PCI BIOS, only direct
4190 hardware access methods are allowed. Use this
4191 if you experience crashes upon bootup and you
4192 suspect they are caused by the BIOS.
4193 conf1 [X86] Force use of PCI Configuration Access
4194 Mechanism 1 (config address in IO port 0xCF8,
4195 data in IO port 0xCFC, both 32-bit).
4196 conf2 [X86] Force use of PCI Configuration Access
4197 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4198 the function, IO port 0xCFA, also 8-bit, sets
4199 bus number. The config space is then accessed
4200 through ports 0xC000-0xCFFF).
4201 See http://wiki.osdev.org/PCI for more info
4202 on the configuration access mechanisms.
4203 noaer [PCIE] If the PCIEAER kernel config parameter is
4204 enabled, this kernel boot option can be used to
4205 disable the use of PCIE advanced error reporting.
4206 nodomains [PCI] Disable support for multiple PCI
4207 root domains (aka PCI segments, in ACPI-speak).
4208 nommconf [X86] Disable use of MMCONFIG for PCI
4210 check_enable_amd_mmconf [X86] check for and enable
4211 properly configured MMIO access to PCI
4212 config space on AMD family 10h CPU
4213 nomsi [MSI] If the PCI_MSI kernel config parameter is
4214 enabled, this kernel boot option can be used to
4215 disable the use of MSI interrupts system-wide.
4216 noioapicquirk [APIC] Disable all boot interrupt quirks.
4217 Safety option to keep boot IRQs enabled. This
4218 should never be necessary.
4219 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4220 primary IO-APIC for bridges that cannot disable
4221 boot IRQs. This fixes a source of spurious IRQs
4222 when the system masks IRQs.
4223 noioapicreroute [APIC] Disable workaround that uses the
4224 boot IRQ equivalent of an IRQ that connects to
4225 a chipset where boot IRQs cannot be disabled.
4226 The opposite of ioapicreroute.
4227 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4228 routing table. These calls are known to be buggy
4229 on several machines and they hang the machine
4230 when used, but on other computers it's the only
4231 way to get the interrupt routing table. Try
4232 this option if the kernel is unable to allocate
4233 IRQs or discover secondary PCI buses on your
4235 rom [X86] Assign address space to expansion ROMs.
4236 Use with caution as certain devices share
4237 address decoders between ROMs and other
4239 norom [X86] Do not assign address space to
4240 expansion ROMs that do not already have
4241 BIOS assigned address ranges.
4242 nobar [X86] Do not assign address space to the
4243 BARs that weren't assigned by the BIOS.
4244 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4245 assigned automatically to PCI devices. You can
4246 make the kernel exclude IRQs of your ISA cards
4248 pirqaddr=0xAAAAA [X86] Specify the physical address
4249 of the PIRQ table (normally generated
4250 by the BIOS) if it is outside the
4251 F0000h-100000h range.
4252 lastbus=N [X86] Scan all buses thru bus #N. Can be
4253 useful if the kernel is unable to find your
4254 secondary buses and you want to tell it
4255 explicitly which ones they are.
4256 assign-busses [X86] Always assign all PCI bus
4257 numbers ourselves, overriding
4258 whatever the firmware may have done.
4259 usepirqmask [X86] Honor the possible IRQ mask stored
4260 in the BIOS $PIR table. This is needed on
4261 some systems with broken BIOSes, notably
4262 some HP Pavilion N5400 and Omnibook XE3
4263 notebooks. This will have no effect if ACPI
4264 IRQ routing is enabled.
4265 noacpi [X86] Do not use ACPI for IRQ routing
4266 or for PCI scanning.
4267 use_crs [X86] Use PCI host bridge window information
4268 from ACPI. On BIOSes from 2008 or later, this
4269 is enabled by default. If you need to use this,
4270 please report a bug.
4271 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4272 If you need to use this, please report a bug.
4273 use_e820 [X86] Use E820 reservations to exclude parts of
4274 PCI host bridge windows. This is a workaround
4275 for BIOS defects in host bridge _CRS methods.
4276 If you need to use this, please report a bug to
4277 <linux-pci@vger.kernel.org>.
4278 no_e820 [X86] Ignore E820 reservations for PCI host
4279 bridge windows. This is the default on modern
4280 hardware. If you need to use this, please report
4281 a bug to <linux-pci@vger.kernel.org>.
4282 routeirq Do IRQ routing for all PCI devices.
4283 This is normally done in pci_enable_device(),
4284 so this option is a temporary workaround
4285 for broken drivers that don't call it.
4286 skip_isa_align [X86] do not align io start addr, so can
4287 handle more pci cards
4288 noearly [X86] Don't do any early type 1 scanning.
4289 This might help on some broken boards which
4290 machine check when some devices' config space
4291 is read. But various workarounds are disabled
4292 and some IOMMU drivers will not work.
4293 bfsort Sort PCI devices into breadth-first order.
4294 This sorting is done to get a device
4295 order compatible with older (<= 2.4) kernels.
4296 nobfsort Don't sort PCI devices into breadth-first order.
4297 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4298 tuning and use the BIOS-configured MPS defaults.
4299 pcie_bus_safe Set every device's MPS to the largest value
4300 supported by all devices below the root complex.
4301 pcie_bus_perf Set device MPS to the largest allowable MPS
4302 based on its parent bus. Also set MRRS (Max
4303 Read Request Size) to the largest supported
4304 value (no larger than the MPS that the device
4305 or bus can support) for best performance.
4306 pcie_bus_peer2peer Set every device's MPS to 128B, which
4307 every device is guaranteed to support. This
4308 configuration allows peer-to-peer DMA between
4309 any pair of devices, possibly at the cost of
4310 reduced performance. This also guarantees
4311 that hot-added devices will work.
4312 cbiosize=nn[KMG] The fixed amount of bus space which is
4313 reserved for the CardBus bridge's IO window.
4314 The default value is 256 bytes.
4315 cbmemsize=nn[KMG] The fixed amount of bus space which is
4316 reserved for the CardBus bridge's memory
4317 window. The default value is 64 megabytes.
4320 [<order of align>@]<pci_dev>[; ...]
4321 Specifies alignment and device to reassign
4322 aligned memory resources. How to
4323 specify the device is described above.
4324 If <order of align> is not specified,
4325 PAGE_SIZE is used as alignment.
4326 A PCI-PCI bridge can be specified if resource
4327 windows need to be expanded.
4328 To specify the alignment for several
4329 instances of a device, the PCI vendor,
4330 device, subvendor, and subdevice may be
4331 specified, e.g., 12@pci:8086:9c22:103c:198f
4332 for 4096-byte alignment.
4333 ecrc= Enable/disable PCIe ECRC (transaction layer
4334 end-to-end CRC checking). Only effective if
4335 OS has native AER control (either granted by
4336 ACPI _OSC or forced via "pcie_ports=native")
4337 bios: Use BIOS/firmware settings. This is the
4341 hpiosize=nn[KMG] The fixed amount of bus space which is
4342 reserved for hotplug bridge's IO window.
4343 Default size is 256 bytes.
4344 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4345 reserved for hotplug bridge's MMIO window.
4346 Default size is 2 megabytes.
4347 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4348 reserved for hotplug bridge's MMIO_PREF window.
4349 Default size is 2 megabytes.
4350 hpmemsize=nn[KMG] The fixed amount of bus space which is
4351 reserved for hotplug bridge's MMIO and
4353 Default size is 2 megabytes.
4354 hpbussize=nn The minimum amount of additional bus numbers
4355 reserved for buses below a hotplug bridge.
4357 realloc= Enable/disable reallocating PCI bridge resources
4358 if allocations done by BIOS are too small to
4359 accommodate resources required by all child
4361 off: Turn realloc off
4363 realloc same as realloc=on
4364 noari do not use PCIe ARI.
4365 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4366 do not use PCIe ATS (and IOMMU device IOTLB).
4367 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4368 only look for one device below a PCIe downstream
4370 big_root_window Try to add a big 64bit memory window to the PCIe
4371 root complex on AMD CPUs. Some GFX hardware
4372 can resize a BAR to allow access to all VRAM.
4373 Adding the window is slightly risky (it may
4374 conflict with unreported devices), so this
4376 disable_acs_redir=<pci_dev>[; ...]
4377 Specify one or more PCI devices (in the format
4378 specified above) separated by semicolons.
4379 Each device specified will have the PCI ACS
4380 redirect capabilities forced off which will
4381 allow P2P traffic between devices through
4382 bridges without forcing it upstream. Note:
4383 this removes isolation between devices and
4384 may put more devices in an IOMMU group.
4385 force_floating [S390] Force usage of floating interrupts.
4386 nomio [S390] Do not use MIO instructions.
4387 norid [S390] ignore the RID field and force use of
4388 one PCI domain per PCI function
4390 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4393 force Enable ASPM even on devices that claim not to support it.
4394 WARNING: Forcing ASPM on may cause system lockups.
4396 pcie_ports= [PCIE] PCIe port services handling:
4397 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4398 even if the platform doesn't give the OS permission to
4399 use them. This may cause conflicts if the platform
4400 also tries to use these services.
4401 dpc-native Use native PCIe service for DPC only. May
4402 cause conflicts if firmware uses AER or DPC.
4403 compat Disable native PCIe services (PME, AER, DPC, PCIe
4406 pcie_port_pm= [PCIE] PCIe port power management handling:
4407 off Disable power management of all PCIe ports
4408 force Forcibly enable power management of all PCIe ports
4410 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4411 nomsi Do not use MSI for native PCIe PME signaling (this makes
4412 all PCIe root ports use INTx for all services).
4414 pcmv= [HW,PCMCIA] BadgePAD 4
4418 Keep all power-domains already enabled by bootloader on,
4419 even if no driver has claimed them. This is useful
4420 for debug and development, but should not be
4421 needed on a platform with proper driver support.
4423 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4426 See arch/parisc/kernel/pdc_chassis.c
4428 percpu_alloc= Select which percpu first chunk allocator to use.
4429 Currently supported values are "embed" and "page".
4430 Archs may support subset or none of the selections.
4431 See comments in mm/percpu.c for details on each
4432 allocator. This parameter is primarily for debugging
4433 and performance comparison.
4435 pirq= [SMP,APIC] Manual mp-table setup
4436 See Documentation/arch/x86/i386/IO-APIC.rst.
4438 plip= [PPT,NET] Parallel port network link
4439 Format: { parport<nr> | timid | 0 }
4440 See also Documentation/admin-guide/parport.rst.
4442 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4443 Override pmtimer IOPort with a hex value.
4446 pmu_override= [PPC] Override the PMU.
4447 This option takes over the PMU facility, so it is no
4448 longer usable by perf. Setting this option starts the
4449 PMU counters by setting MMCR0 to 0 (the FC bit is
4450 cleared). If a number is given, then MMCR1 is set to
4451 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4454 pm_debug_messages [SUSPEND,KNL]
4455 Enable suspend/resume debug messages during boot up.
4458 Enable PNP debug messages (depends on the
4459 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4460 via /sys/module/pnp/parameters/debug. We always show
4461 current resource usage; turning this on also shows
4462 possible settings and some assignment information.
4468 { on | off | curr | res | no-curr | no-res }
4471 [ISAPNP] Exclude IRQs for the autoconfiguration
4474 [ISAPNP] Exclude DMAs for the autoconfiguration
4476 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4477 Ranges are in pairs (I/O port base and size).
4480 [ISAPNP] Exclude memory regions for the
4482 Ranges are in pairs (memory base and size).
4484 ports= [IP_VS_FTP] IPVS ftp helper module
4486 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4488 Format: <port>,<port>....
4490 powersave=off [PPC] This option disables power saving features.
4491 It specifically disables cpuidle and sets the
4492 platform machine description specific power_save
4493 function to NULL. On Idle the CPU just reduces
4496 ppc_strict_facility_enable
4497 [PPC] This option catches any kernel floating point,
4498 Altivec, VSX and SPE outside of regions specifically
4499 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4500 There is some performance impact when enabling this.
4504 Disable Hardware Transactional Memory
4507 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4508 none - Limited to cond_resched() calls
4509 voluntary - Limited to cond_resched() and might_sleep() calls
4510 full - Any section that isn't explicitly preempt disabled
4511 can be preempted anytime.
4513 print-fatal-signals=
4514 [KNL] debug: print fatal signals
4516 If enabled, warn about various signal handling
4517 related application anomalies: too many signals,
4518 too many POSIX.1 timers, fatal signals causing a
4521 If you hit the warning due to signal overflow,
4522 you might want to try "ulimit -i unlimited".
4526 printk.always_kmsg_dump=
4527 Trigger kmsg_dump for cases other than kernel oops or
4529 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4532 printk.console_no_auto_verbose=
4533 Disable console loglevel raise on oops, panic
4534 or lockdep-detected issues (only if lock debug is on).
4535 With an exception to setups with low baudrate on
4536 serial console, keeping this 0 is a good choice
4537 in order to provide more debug information.
4539 default: 0 (auto_verbose is enabled)
4541 printk.devkmsg={on,off,ratelimit}
4542 Control writing to /dev/kmsg.
4543 on - unlimited logging to /dev/kmsg from userspace
4544 off - logging to /dev/kmsg disabled
4545 ratelimit - ratelimit the logging
4548 printk.time= Show timing data prefixed to each printk message line
4549 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4551 processor.max_cstate= [HW,ACPI]
4552 Limit processor to maximum C-state
4553 max_cstate=9 overrides any DMI blacklist limit.
4555 processor.nocst [HW,ACPI]
4556 Ignore the _CST method to determine C-states,
4557 instead using the legacy FADT method
4559 profile= [KNL] Enable kernel profiling via /proc/profile
4560 Format: [<profiletype>,]<number>
4561 Param: <profiletype>: "schedule", "sleep", or "kvm"
4562 [defaults to kernel profiling]
4563 Param: "schedule" - profile schedule points.
4564 Param: "sleep" - profile D-state sleeping (millisecs).
4565 Requires CONFIG_SCHEDSTATS
4566 Param: "kvm" - profile VM exits.
4567 Param: <number> - step/bucket size as a power of 2 for
4568 statistical time based profiling.
4570 prompt_ramdisk= [RAM] [Deprecated]
4572 prot_virt= [S390] enable hosting protected virtual machines
4573 isolated from the hypervisor (if hardware supports
4577 psi= [KNL] Enable or disable pressure stall information
4581 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4582 probe for; one of (bare|imps|exps|lifebook|any).
4583 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4585 psmouse.resetafter= [HW,MOUSE]
4586 Try to reset the device after so many bad packets
4589 [HW,MOUSE] Set desired mouse resolution, in dpi.
4590 psmouse.smartscroll=
4591 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4592 0 = disabled, 1 = enabled (default).
4594 pstore.backend= Specify the name of the pstore backend to use
4596 pti= [X86-64] Control Page Table Isolation of user and
4597 kernel address spaces. Disabling this feature
4598 removes hardening, but improves performance of
4599 system calls and interrupts.
4601 on - unconditionally enable
4602 off - unconditionally disable
4603 auto - kernel detects whether your CPU model is
4604 vulnerable to issues that PTI mitigates
4606 Not specifying this option is equivalent to pti=auto.
4609 [KNL] Number of legacy pty's. Overwrites compiled-in
4612 quiet [KNL] Disable most log messages
4616 radix_hcall_invalidate=on [PPC/PSERIES]
4617 Disable RADIX GTSE feature and use hcall for TLB
4621 See Documentation/admin-guide/md.rst.
4623 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4624 See Documentation/admin-guide/blockdev/ramdisk.rst.
4626 ramdisk_start= [RAM] RAM disk image start address
4628 random.trust_cpu=off
4629 [KNL] Disable trusting the use of the CPU's
4630 random number generator (if available) to
4631 initialize the kernel's RNG.
4633 random.trust_bootloader=off
4634 [KNL] Disable trusting the use of the a seed
4635 passed by the bootloader (if available) to
4636 initialize the kernel's RNG.
4638 randomize_kstack_offset=
4639 [KNL] Enable or disable kernel stack offset
4640 randomization, which provides roughly 5 bits of
4641 entropy, frustrating memory corruption attacks
4642 that depend on stack address determinism or
4643 cross-syscall address exposures. This is only
4644 available on architectures that have defined
4645 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4646 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4647 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4649 ras=option[,option,...] [KNL] RAS-specific options
4652 Disable the Correctable Errors Collector,
4653 see CONFIG_RAS_CEC help text.
4655 rcu_nocbs[=cpu-list]
4656 [KNL] The optional argument is a cpu list,
4659 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4660 enable the no-callback CPU mode, which prevents
4661 such CPUs' callbacks from being invoked in
4662 softirq context. Invocation of such CPUs' RCU
4663 callbacks will instead be offloaded to "rcuox/N"
4664 kthreads created for that purpose, where "x" is
4665 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4666 for the kthreads that mediate grace periods; and
4667 "N" is the CPU number. This reduces OS jitter on
4668 the offloaded CPUs, which can be useful for HPC
4669 and real-time workloads. It can also improve
4670 energy efficiency for asymmetric multiprocessors.
4672 If a cpulist is passed as an argument, the specified
4673 list of CPUs is set to no-callback mode from boot.
4675 Otherwise, if the '=' sign and the cpulist
4676 arguments are omitted, no CPU will be set to
4677 no-callback mode from boot but the mode may be
4678 toggled at runtime via cpusets.
4680 Note that this argument takes precedence over
4681 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4684 Rather than requiring that offloaded CPUs
4685 (specified by rcu_nocbs= above) explicitly
4686 awaken the corresponding "rcuoN" kthreads,
4687 make these kthreads poll for callbacks.
4688 This improves the real-time response for the
4689 offloaded CPUs by relieving them of the need to
4690 wake up the corresponding kthread, but degrades
4691 energy efficiency by requiring that the kthreads
4692 periodically wake up to do the polling.
4694 rcutree.blimit= [KNL]
4695 Set maximum number of finished RCU callbacks to
4696 process in one batch.
4698 rcutree.dump_tree= [KNL]
4699 Dump the structure of the rcu_node combining tree
4700 out at early boot. This is used for diagnostic
4701 purposes, to verify correct tree setup.
4703 rcutree.gp_cleanup_delay= [KNL]
4704 Set the number of jiffies to delay each step of
4705 RCU grace-period cleanup.
4707 rcutree.gp_init_delay= [KNL]
4708 Set the number of jiffies to delay each step of
4709 RCU grace-period initialization.
4711 rcutree.gp_preinit_delay= [KNL]
4712 Set the number of jiffies to delay each step of
4713 RCU grace-period pre-initialization, that is,
4714 the propagation of recent CPU-hotplug changes up
4715 the rcu_node combining tree.
4717 rcutree.use_softirq= [KNL]
4718 If set to zero, move all RCU_SOFTIRQ processing to
4719 per-CPU rcuc kthreads. Defaults to a non-zero
4720 value, meaning that RCU_SOFTIRQ is used by default.
4721 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4723 But note that CONFIG_PREEMPT_RT=y kernels disable
4724 this kernel boot parameter, forcibly setting it
4727 rcutree.rcu_fanout_exact= [KNL]
4728 Disable autobalancing of the rcu_node combining
4729 tree. This is used by rcutorture, and might
4730 possibly be useful for architectures having high
4731 cache-to-cache transfer latencies.
4733 rcutree.rcu_fanout_leaf= [KNL]
4734 Change the number of CPUs assigned to each
4735 leaf rcu_node structure. Useful for very
4736 large systems, which will choose the value 64,
4737 and for NUMA systems with large remote-access
4738 latencies, which will choose a value aligned
4739 with the appropriate hardware boundaries.
4741 rcutree.rcu_min_cached_objs= [KNL]
4742 Minimum number of objects which are cached and
4743 maintained per one CPU. Object size is equal
4744 to PAGE_SIZE. The cache allows to reduce the
4745 pressure to page allocator, also it makes the
4746 whole algorithm to behave better in low memory
4749 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4750 Set the page-cache refill delay (in milliseconds)
4751 in response to low-memory conditions. The range
4752 of permitted values is in the range 0:100000.
4754 rcutree.jiffies_till_first_fqs= [KNL]
4755 Set delay from grace-period initialization to
4756 first attempt to force quiescent states.
4757 Units are jiffies, minimum value is zero,
4758 and maximum value is HZ.
4760 rcutree.jiffies_till_next_fqs= [KNL]
4761 Set delay between subsequent attempts to force
4762 quiescent states. Units are jiffies, minimum
4763 value is one, and maximum value is HZ.
4765 rcutree.jiffies_till_sched_qs= [KNL]
4766 Set required age in jiffies for a
4767 given grace period before RCU starts
4768 soliciting quiescent-state help from
4769 rcu_note_context_switch() and cond_resched().
4770 If not specified, the kernel will calculate
4771 a value based on the most recent settings
4772 of rcutree.jiffies_till_first_fqs
4773 and rcutree.jiffies_till_next_fqs.
4774 This calculated value may be viewed in
4775 rcutree.jiffies_to_sched_qs. Any attempt to set
4776 rcutree.jiffies_to_sched_qs will be cheerfully
4779 rcutree.kthread_prio= [KNL,BOOT]
4780 Set the SCHED_FIFO priority of the RCU per-CPU
4781 kthreads (rcuc/N). This value is also used for
4782 the priority of the RCU boost threads (rcub/N)
4783 and for the RCU grace-period kthreads (rcu_bh,
4784 rcu_preempt, and rcu_sched). If RCU_BOOST is
4785 set, valid values are 1-99 and the default is 1
4786 (the least-favored priority). Otherwise, when
4787 RCU_BOOST is not set, valid values are 0-99 and
4788 the default is zero (non-realtime operation).
4789 When RCU_NOCB_CPU is set, also adjust the
4790 priority of NOCB callback kthreads.
4792 rcutree.rcu_divisor= [KNL]
4793 Set the shift-right count to use to compute
4794 the callback-invocation batch limit bl from
4795 the number of callbacks queued on this CPU.
4796 The result will be bounded below by the value of
4797 the rcutree.blimit kernel parameter. Every bl
4798 callbacks, the softirq handler will exit in
4799 order to allow the CPU to do other work.
4801 Please note that this callback-invocation batch
4802 limit applies only to non-offloaded callback
4803 invocation. Offloaded callbacks are instead
4804 invoked in the context of an rcuoc kthread, which
4805 scheduler will preempt as it does any other task.
4807 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4808 On callback-offloaded (rcu_nocbs) CPUs,
4809 RCU reduces the lock contention that would
4810 otherwise be caused by callback floods through
4811 use of the ->nocb_bypass list. However, in the
4812 common non-flooded case, RCU queues directly to
4813 the main ->cblist in order to avoid the extra
4814 overhead of the ->nocb_bypass list and its lock.
4815 But if there are too many callbacks queued during
4816 a single jiffy, RCU pre-queues the callbacks into
4817 the ->nocb_bypass queue. The definition of "too
4818 many" is supplied by this kernel boot parameter.
4820 rcutree.rcu_nocb_gp_stride= [KNL]
4821 Set the number of NOCB callback kthreads in
4822 each group, which defaults to the square root
4823 of the number of CPUs. Larger numbers reduce
4824 the wakeup overhead on the global grace-period
4825 kthread, but increases that same overhead on
4826 each group's NOCB grace-period kthread.
4828 rcutree.qhimark= [KNL]
4829 Set threshold of queued RCU callbacks beyond which
4830 batch limiting is disabled.
4832 rcutree.qlowmark= [KNL]
4833 Set threshold of queued RCU callbacks below which
4834 batch limiting is re-enabled.
4836 rcutree.qovld= [KNL]
4837 Set threshold of queued RCU callbacks beyond which
4838 RCU's force-quiescent-state scan will aggressively
4839 enlist help from cond_resched() and sched IPIs to
4840 help CPUs more quickly reach quiescent states.
4841 Set to less than zero to make this be set based
4842 on rcutree.qhimark at boot time and to zero to
4843 disable more aggressive help enlistment.
4845 rcutree.rcu_kick_kthreads= [KNL]
4846 Cause the grace-period kthread to get an extra
4847 wake_up() if it sleeps three times longer than
4848 it should at force-quiescent-state time.
4849 This wake_up() will be accompanied by a
4850 WARN_ONCE() splat and an ftrace_dump().
4852 rcutree.rcu_unlock_delay= [KNL]
4853 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4854 this specifies an rcu_read_unlock()-time delay
4855 in microseconds. This defaults to zero.
4856 Larger delays increase the probability of
4857 catching RCU pointer leaks, that is, buggy use
4858 of RCU-protected pointers after the relevant
4859 rcu_read_unlock() has completed.
4861 rcutree.sysrq_rcu= [KNL]
4862 Commandeer a sysrq key to dump out Tree RCU's
4863 rcu_node tree with an eye towards determining
4864 why a new grace period has not yet started.
4866 rcuscale.gp_async= [KNL]
4867 Measure performance of asynchronous
4868 grace-period primitives such as call_rcu().
4870 rcuscale.gp_async_max= [KNL]
4871 Specify the maximum number of outstanding
4872 callbacks per writer thread. When a writer
4873 thread exceeds this limit, it invokes the
4874 corresponding flavor of rcu_barrier() to allow
4875 previously posted callbacks to drain.
4877 rcuscale.gp_exp= [KNL]
4878 Measure performance of expedited synchronous
4879 grace-period primitives.
4881 rcuscale.holdoff= [KNL]
4882 Set test-start holdoff period. The purpose of
4883 this parameter is to delay the start of the
4884 test until boot completes in order to avoid
4887 rcuscale.kfree_rcu_test= [KNL]
4888 Set to measure performance of kfree_rcu() flooding.
4890 rcuscale.kfree_rcu_test_double= [KNL]
4891 Test the double-argument variant of kfree_rcu().
4892 If this parameter has the same value as
4893 rcuscale.kfree_rcu_test_single, both the single-
4894 and double-argument variants are tested.
4896 rcuscale.kfree_rcu_test_single= [KNL]
4897 Test the single-argument variant of kfree_rcu().
4898 If this parameter has the same value as
4899 rcuscale.kfree_rcu_test_double, both the single-
4900 and double-argument variants are tested.
4902 rcuscale.kfree_nthreads= [KNL]
4903 The number of threads running loops of kfree_rcu().
4905 rcuscale.kfree_alloc_num= [KNL]
4906 Number of allocations and frees done in an iteration.
4908 rcuscale.kfree_loops= [KNL]
4909 Number of loops doing rcuscale.kfree_alloc_num number
4910 of allocations and frees.
4912 rcuscale.nreaders= [KNL]
4913 Set number of RCU readers. The value -1 selects
4914 N, where N is the number of CPUs. A value
4915 "n" less than -1 selects N-n+1, where N is again
4916 the number of CPUs. For example, -2 selects N
4917 (the number of CPUs), -3 selects N+1, and so on.
4918 A value of "n" less than or equal to -N selects
4921 rcuscale.nwriters= [KNL]
4922 Set number of RCU writers. The values operate
4923 the same as for rcuscale.nreaders.
4924 N, where N is the number of CPUs
4926 rcuscale.perf_type= [KNL]
4927 Specify the RCU implementation to test.
4929 rcuscale.shutdown= [KNL]
4930 Shut the system down after performance tests
4931 complete. This is useful for hands-off automated
4934 rcuscale.verbose= [KNL]
4935 Enable additional printk() statements.
4937 rcuscale.writer_holdoff= [KNL]
4938 Write-side holdoff between grace periods,
4939 in microseconds. The default of zero says
4942 rcutorture.fqs_duration= [KNL]
4943 Set duration of force_quiescent_state bursts
4946 rcutorture.fqs_holdoff= [KNL]
4947 Set holdoff time within force_quiescent_state bursts
4950 rcutorture.fqs_stutter= [KNL]
4951 Set wait time between force_quiescent_state bursts
4954 rcutorture.fwd_progress= [KNL]
4955 Specifies the number of kthreads to be used
4956 for RCU grace-period forward-progress testing
4957 for the types of RCU supporting this notion.
4958 Defaults to 1 kthread, values less than zero or
4959 greater than the number of CPUs cause the number
4962 rcutorture.fwd_progress_div= [KNL]
4963 Specify the fraction of a CPU-stall-warning
4964 period to do tight-loop forward-progress testing.
4966 rcutorture.fwd_progress_holdoff= [KNL]
4967 Number of seconds to wait between successive
4968 forward-progress tests.
4970 rcutorture.fwd_progress_need_resched= [KNL]
4971 Enclose cond_resched() calls within checks for
4972 need_resched() during tight-loop forward-progress
4975 rcutorture.gp_cond= [KNL]
4976 Use conditional/asynchronous update-side
4977 primitives, if available.
4979 rcutorture.gp_exp= [KNL]
4980 Use expedited update-side primitives, if available.
4982 rcutorture.gp_normal= [KNL]
4983 Use normal (non-expedited) asynchronous
4984 update-side primitives, if available.
4986 rcutorture.gp_sync= [KNL]
4987 Use normal (non-expedited) synchronous
4988 update-side primitives, if available. If all
4989 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4990 rcutorture.gp_normal=, and rcutorture.gp_sync=
4991 are zero, rcutorture acts as if is interpreted
4992 they are all non-zero.
4994 rcutorture.irqreader= [KNL]
4995 Run RCU readers from irq handlers, or, more
4996 accurately, from a timer handler. Not all RCU
4997 flavors take kindly to this sort of thing.
4999 rcutorture.leakpointer= [KNL]
5000 Leak an RCU-protected pointer out of the reader.
5001 This can of course result in splats, and is
5002 intended to test the ability of things like
5003 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
5006 rcutorture.n_barrier_cbs= [KNL]
5007 Set callbacks/threads for rcu_barrier() testing.
5009 rcutorture.nfakewriters= [KNL]
5010 Set number of concurrent RCU writers. These just
5011 stress RCU, they don't participate in the actual
5012 test, hence the "fake".
5014 rcutorture.nocbs_nthreads= [KNL]
5015 Set number of RCU callback-offload togglers.
5016 Zero (the default) disables toggling.
5018 rcutorture.nocbs_toggle= [KNL]
5019 Set the delay in milliseconds between successive
5020 callback-offload toggling attempts.
5022 rcutorture.nreaders= [KNL]
5023 Set number of RCU readers. The value -1 selects
5024 N-1, where N is the number of CPUs. A value
5025 "n" less than -1 selects N-n-2, where N is again
5026 the number of CPUs. For example, -2 selects N
5027 (the number of CPUs), -3 selects N+1, and so on.
5029 rcutorture.object_debug= [KNL]
5030 Enable debug-object double-call_rcu() testing.
5032 rcutorture.onoff_holdoff= [KNL]
5033 Set time (s) after boot for CPU-hotplug testing.
5035 rcutorture.onoff_interval= [KNL]
5036 Set time (jiffies) between CPU-hotplug operations,
5037 or zero to disable CPU-hotplug testing.
5039 rcutorture.read_exit= [KNL]
5040 Set the number of read-then-exit kthreads used
5041 to test the interaction of RCU updaters and
5042 task-exit processing.
5044 rcutorture.read_exit_burst= [KNL]
5045 The number of times in a given read-then-exit
5046 episode that a set of read-then-exit kthreads
5049 rcutorture.read_exit_delay= [KNL]
5050 The delay, in seconds, between successive
5051 read-then-exit testing episodes.
5053 rcutorture.shuffle_interval= [KNL]
5054 Set task-shuffle interval (s). Shuffling tasks
5055 allows some CPUs to go into dyntick-idle mode
5056 during the rcutorture test.
5058 rcutorture.shutdown_secs= [KNL]
5059 Set time (s) after boot system shutdown. This
5060 is useful for hands-off automated testing.
5062 rcutorture.stall_cpu= [KNL]
5063 Duration of CPU stall (s) to test RCU CPU stall
5064 warnings, zero to disable.
5066 rcutorture.stall_cpu_block= [KNL]
5067 Sleep while stalling if set. This will result
5068 in warnings from preemptible RCU in addition
5069 to any other stall-related activity.
5071 rcutorture.stall_cpu_holdoff= [KNL]
5072 Time to wait (s) after boot before inducing stall.
5074 rcutorture.stall_cpu_irqsoff= [KNL]
5075 Disable interrupts while stalling if set.
5077 rcutorture.stall_gp_kthread= [KNL]
5078 Duration (s) of forced sleep within RCU
5079 grace-period kthread to test RCU CPU stall
5080 warnings, zero to disable. If both stall_cpu
5081 and stall_gp_kthread are specified, the
5082 kthread is starved first, then the CPU.
5084 rcutorture.stat_interval= [KNL]
5085 Time (s) between statistics printk()s.
5087 rcutorture.stutter= [KNL]
5088 Time (s) to stutter testing, for example, specifying
5089 five seconds causes the test to run for five seconds,
5090 wait for five seconds, and so on. This tests RCU's
5091 ability to transition abruptly to and from idle.
5093 rcutorture.test_boost= [KNL]
5094 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5095 "Maybe" means test if the RCU implementation
5096 under test support RCU priority boosting.
5098 rcutorture.test_boost_duration= [KNL]
5099 Duration (s) of each individual boost test.
5101 rcutorture.test_boost_interval= [KNL]
5102 Interval (s) between each boost test.
5104 rcutorture.test_no_idle_hz= [KNL]
5105 Test RCU's dyntick-idle handling. See also the
5106 rcutorture.shuffle_interval parameter.
5108 rcutorture.torture_type= [KNL]
5109 Specify the RCU implementation to test.
5111 rcutorture.verbose= [KNL]
5112 Enable additional printk() statements.
5114 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5115 Dump ftrace buffer after reporting RCU CPU
5118 rcupdate.rcu_cpu_stall_suppress= [KNL]
5119 Suppress RCU CPU stall warning messages.
5121 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5122 Suppress RCU CPU stall warning messages and
5123 rcutorture writer stall warnings that occur
5124 during early boot, that is, during the time
5125 before the init task is spawned.
5127 rcupdate.rcu_cpu_stall_timeout= [KNL]
5128 Set timeout for RCU CPU stall warning messages.
5129 The value is in seconds and the maximum allowed
5130 value is 300 seconds.
5132 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5133 Set timeout for expedited RCU CPU stall warning
5134 messages. The value is in milliseconds
5135 and the maximum allowed value is 21000
5136 milliseconds. Please note that this value is
5137 adjusted to an arch timer tick resolution.
5138 Setting this to zero causes the value from
5139 rcupdate.rcu_cpu_stall_timeout to be used (after
5140 conversion from seconds to milliseconds).
5142 rcupdate.rcu_cpu_stall_cputime= [KNL]
5143 Provide statistics on the cputime and count of
5144 interrupts and tasks during the sampling period. For
5145 multiple continuous RCU stalls, all sampling periods
5146 begin at half of the first RCU stall timeout.
5148 rcupdate.rcu_exp_stall_task_details= [KNL]
5149 Print stack dumps of any tasks blocking the
5150 current expedited RCU grace period during an
5151 expedited RCU CPU stall warning.
5153 rcupdate.rcu_expedited= [KNL]
5154 Use expedited grace-period primitives, for
5155 example, synchronize_rcu_expedited() instead
5156 of synchronize_rcu(). This reduces latency,
5157 but can increase CPU utilization, degrade
5158 real-time latency, and degrade energy efficiency.
5159 No effect on CONFIG_TINY_RCU kernels.
5161 rcupdate.rcu_normal= [KNL]
5162 Use only normal grace-period primitives,
5163 for example, synchronize_rcu() instead of
5164 synchronize_rcu_expedited(). This improves
5165 real-time latency, CPU utilization, and
5166 energy efficiency, but can expose users to
5167 increased grace-period latency. This parameter
5168 overrides rcupdate.rcu_expedited. No effect on
5169 CONFIG_TINY_RCU kernels.
5171 rcupdate.rcu_normal_after_boot= [KNL]
5172 Once boot has completed (that is, after
5173 rcu_end_inkernel_boot() has been invoked), use
5174 only normal grace-period primitives. No effect
5175 on CONFIG_TINY_RCU kernels.
5177 But note that CONFIG_PREEMPT_RT=y kernels enables
5178 this kernel boot parameter, forcibly setting
5179 it to the value one, that is, converting any
5180 post-boot attempt at an expedited RCU grace
5181 period to instead use normal non-expedited
5182 grace-period processing.
5184 rcupdate.rcu_task_collapse_lim= [KNL]
5185 Set the maximum number of callbacks present
5186 at the beginning of a grace period that allows
5187 the RCU Tasks flavors to collapse back to using
5188 a single callback queue. This switching only
5189 occurs when rcupdate.rcu_task_enqueue_lim is
5190 set to the default value of -1.
5192 rcupdate.rcu_task_contend_lim= [KNL]
5193 Set the minimum number of callback-queuing-time
5194 lock-contention events per jiffy required to
5195 cause the RCU Tasks flavors to switch to per-CPU
5196 callback queuing. This switching only occurs
5197 when rcupdate.rcu_task_enqueue_lim is set to
5198 the default value of -1.
5200 rcupdate.rcu_task_enqueue_lim= [KNL]
5201 Set the number of callback queues to use for the
5202 RCU Tasks family of RCU flavors. The default
5203 of -1 allows this to be automatically (and
5204 dynamically) adjusted. This parameter is intended
5207 rcupdate.rcu_task_ipi_delay= [KNL]
5208 Set time in jiffies during which RCU tasks will
5209 avoid sending IPIs, starting with the beginning
5210 of a given grace period. Setting a large
5211 number avoids disturbing real-time workloads,
5212 but lengthens grace periods.
5214 rcupdate.rcu_task_stall_info= [KNL]
5215 Set initial timeout in jiffies for RCU task stall
5216 informational messages, which give some indication
5217 of the problem for those not patient enough to
5218 wait for ten minutes. Informational messages are
5219 only printed prior to the stall-warning message
5220 for a given grace period. Disable with a value
5221 less than or equal to zero. Defaults to ten
5222 seconds. A change in value does not take effect
5223 until the beginning of the next grace period.
5225 rcupdate.rcu_task_stall_info_mult= [KNL]
5226 Multiplier for time interval between successive
5227 RCU task stall informational messages for a given
5228 RCU tasks grace period. This value is clamped
5229 to one through ten, inclusive. It defaults to
5230 the value three, so that the first informational
5231 message is printed 10 seconds into the grace
5232 period, the second at 40 seconds, the third at
5233 160 seconds, and then the stall warning at 600
5234 seconds would prevent a fourth at 640 seconds.
5236 rcupdate.rcu_task_stall_timeout= [KNL]
5237 Set timeout in jiffies for RCU task stall
5238 warning messages. Disable with a value less
5239 than or equal to zero. Defaults to ten minutes.
5240 A change in value does not take effect until
5241 the beginning of the next grace period.
5243 rcupdate.rcu_self_test= [KNL]
5244 Run the RCU early boot self tests
5248 Run specified binary instead of /init from the ramdisk,
5249 used for early userspace startup. See initrd.
5252 force - Override the decision by the kernel to hide the
5253 advertisement of RDRAND support (this affects
5254 certain AMD processors because of buggy BIOS
5255 support, specifically around the suspend/resume
5259 Turn on/off individual RDT features. List is:
5260 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5262 E.g. to turn on cmt and turn off mba use:
5266 Format (x86 or x86_64):
5267 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5269 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5271 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5272 (prefix with 'panic_' to set mode for panic
5274 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5275 reboot_force is either force or not specified,
5276 reboot_cpu is s[mp]#### with #### being the processor
5277 to be used for rebooting.
5279 refscale.holdoff= [KNL]
5280 Set test-start holdoff period. The purpose of
5281 this parameter is to delay the start of the
5282 test until boot completes in order to avoid
5285 refscale.loops= [KNL]
5286 Set the number of loops over the synchronization
5287 primitive under test. Increasing this number
5288 reduces noise due to loop start/end overhead,
5289 but the default has already reduced the per-pass
5290 noise to a handful of picoseconds on ca. 2020
5293 refscale.nreaders= [KNL]
5294 Set number of readers. The default value of -1
5295 selects N, where N is roughly 75% of the number
5296 of CPUs. A value of zero is an interesting choice.
5298 refscale.nruns= [KNL]
5299 Set number of runs, each of which is dumped onto
5302 refscale.readdelay= [KNL]
5303 Set the read-side critical-section duration,
5304 measured in microseconds.
5306 refscale.scale_type= [KNL]
5307 Specify the read-protection implementation to test.
5309 refscale.shutdown= [KNL]
5310 Shut down the system at the end of the performance
5311 test. This defaults to 1 (shut it down) when
5312 refscale is built into the kernel and to 0 (leave
5313 it running) when refscale is built as a module.
5315 refscale.verbose= [KNL]
5316 Enable additional printk() statements.
5318 refscale.verbose_batched= [KNL]
5319 Batch the additional printk() statements. If zero
5320 (the default) or negative, print everything. Otherwise,
5321 print every Nth verbose statement, where N is the value
5325 [KNL, SMP] Set scheduler's default relax_domain_level.
5326 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5328 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5329 Format: <base1>,<size1>[,<base2>,<size2>,...]
5330 Reserve I/O ports or memory so the kernel won't use
5331 them. If <base> is less than 0x10000, the region
5332 is assumed to be I/O ports; otherwise it is memory.
5334 reservetop= [X86-32]
5336 Reserves a hole at the top of the kernel virtual
5339 reset_devices [KNL] Force drivers to reset the underlying device
5340 during initialization.
5343 Specify the partition device for software suspend
5345 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5347 resume_offset= [SWSUSP]
5348 Specify the offset from the beginning of the partition
5349 given by "resume=" at which the swap header is located,
5350 in <PAGE_SIZE> units (needed only for swap files).
5351 See Documentation/power/swsusp-and-swap-files.rst
5353 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5354 read the resume files
5356 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5357 Useful for devices that are detected asynchronously
5358 (e.g. USB and MMC devices).
5360 retain_initrd [RAM] Keep initrd memory after extraction
5362 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5363 Speculative Code Execution with Return Instructions)
5366 AMD-based UNRET and IBPB mitigations alone do not stop
5367 sibling threads from influencing the predictions of other
5368 sibling threads. For that reason, STIBP is used on pro-
5369 cessors that support it, and mitigate SMT on processors
5373 auto - automatically select a migitation
5374 auto,nosmt - automatically select a mitigation,
5375 disabling SMT if necessary for
5376 the full mitigation (only on Zen1
5377 and older without STIBP).
5378 ibpb - On AMD, mitigate short speculation
5379 windows on basic block boundaries too.
5380 Safe, highest perf impact. It also
5381 enables STIBP if present. Not suitable
5383 ibpb,nosmt - Like "ibpb" above but will disable SMT
5384 when STIBP is not available. This is
5385 the alternative for systems which do not
5387 unret - Force enable untrained return thunks,
5388 only effective on AMD f15h-f17h based
5390 unret,nosmt - Like unret, but will disable SMT when STIBP
5391 is not available. This is the alternative for
5392 systems which do not have STIBP.
5394 Selecting 'auto' will choose a mitigation method at run
5395 time according to the CPU.
5397 Not specifying this option is equivalent to retbleed=auto.
5399 rfkill.default_state=
5400 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5401 etc. communication is blocked by default.
5404 rfkill.master_switch_mode=
5405 0 The "airplane mode" button does nothing.
5406 1 The "airplane mode" button toggles between everything
5407 blocked and the previous configuration.
5408 2 The "airplane mode" button toggles between everything
5409 blocked and everything unblocked.
5411 rhash_entries= [KNL,NET]
5412 Set number of hash buckets for route cache
5415 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5418 ro [KNL] Mount root device read-only on boot
5421 on Mark read-only kernel memory as read-only (default).
5422 off Leave read-only kernel memory writable for debugging.
5423 full Mark read-only kernel memory and aliases as read-only
5427 Enable the uart passthrough on the designated usb port
5428 on Rockchip SoCs. When active, the signals of the
5429 debug-uart get routed to the D+ and D- pins of the usb
5430 port and the regular usb controller gets disabled.
5432 root= [KNL] Root filesystem
5433 See name_to_dev_t comment in init/do_mounts.c.
5435 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5436 mount the root filesystem
5438 rootflags= [KNL] Set root filesystem mount option string
5440 rootfstype= [KNL] Set root filesystem type
5442 rootwait [KNL] Wait (indefinitely) for root device to show up.
5443 Useful for devices that are detected asynchronously
5444 (e.g. USB and MMC devices).
5446 rproc_mem=nn[KMG][@address]
5447 [KNL,ARM,CMA] Remoteproc physical memory block.
5448 Memory area to be used by remote processor image,
5451 rw [KNL] Mount root device read-write on boot
5453 S [KNL] Run init in single mode
5455 s390_iommu= [HW,S390]
5456 Set s390 IOTLB flushing mode
5458 With strict flushing every unmap operation will result in
5459 an IOTLB flush. Default is lazy flushing before reuse,
5462 s390_iommu_aperture= [KNL,S390]
5463 Specifies the size of the per device DMA address space
5464 accessible through the DMA and IOMMU APIs as a decimal
5465 factor of the size of main memory.
5466 The default is 1 meaning that one can concurrently use
5467 as many DMA addresses as physical memory is installed,
5468 if supported by hardware, and thus map all of memory
5469 once. With a value of 2 one can map all of memory twice
5470 and so on. As a special case a factor of 0 imposes no
5471 restrictions other than those given by hardware at the
5472 cost of significant additional memory use for tables.
5475 See drivers/net/irda/sa1100_ir.c.
5477 sched_verbose [KNL] Enables verbose scheduler debug messages.
5479 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5480 Allowed values are enable and disable. This feature
5481 incurs a small amount of overhead in the scheduler
5482 but is useful for debugging and performance tuning.
5484 sched_thermal_decay_shift=
5485 [KNL, SMP] Set a decay shift for scheduler thermal
5486 pressure signal. Thermal pressure signal follows the
5487 default decay period of other scheduler pelt
5488 signals(usually 32 ms but configurable). Setting
5489 sched_thermal_decay_shift will left shift the decay
5490 period for the thermal pressure signal by the shift
5492 i.e. with the default pelt decay period of 32 ms
5493 sched_thermal_decay_shift thermal pressure decay pr
5497 Format: integer between 0 and 10
5500 scftorture.holdoff= [KNL]
5501 Number of seconds to hold off before starting
5502 test. Defaults to zero for module insertion and
5503 to 10 seconds for built-in smp_call_function()
5506 scftorture.longwait= [KNL]
5507 Request ridiculously long waits randomly selected
5508 up to the chosen limit in seconds. Zero (the
5509 default) disables this feature. Please note
5510 that requesting even small non-zero numbers of
5511 seconds can result in RCU CPU stall warnings,
5512 softlockup complaints, and so on.
5514 scftorture.nthreads= [KNL]
5515 Number of kthreads to spawn to invoke the
5516 smp_call_function() family of functions.
5517 The default of -1 specifies a number of kthreads
5518 equal to the number of CPUs.
5520 scftorture.onoff_holdoff= [KNL]
5521 Number seconds to wait after the start of the
5522 test before initiating CPU-hotplug operations.
5524 scftorture.onoff_interval= [KNL]
5525 Number seconds to wait between successive
5526 CPU-hotplug operations. Specifying zero (which
5527 is the default) disables CPU-hotplug operations.
5529 scftorture.shutdown_secs= [KNL]
5530 The number of seconds following the start of the
5531 test after which to shut down the system. The
5532 default of zero avoids shutting down the system.
5533 Non-zero values are useful for automated tests.
5535 scftorture.stat_interval= [KNL]
5536 The number of seconds between outputting the
5537 current test statistics to the console. A value
5538 of zero disables statistics output.
5540 scftorture.stutter_cpus= [KNL]
5541 The number of jiffies to wait between each change
5542 to the set of CPUs under test.
5544 scftorture.use_cpus_read_lock= [KNL]
5545 Use use_cpus_read_lock() instead of the default
5546 preempt_disable() to disable CPU hotplug
5547 while invoking one of the smp_call_function*()
5550 scftorture.verbose= [KNL]
5551 Enable additional printk() statements.
5553 scftorture.weight_single= [KNL]
5554 The probability weighting to use for the
5555 smp_call_function_single() function with a zero
5556 "wait" parameter. A value of -1 selects the
5557 default if all other weights are -1. However,
5558 if at least one weight has some other value, a
5559 value of -1 will instead select a weight of zero.
5561 scftorture.weight_single_wait= [KNL]
5562 The probability weighting to use for the
5563 smp_call_function_single() function with a
5564 non-zero "wait" parameter. See weight_single.
5566 scftorture.weight_many= [KNL]
5567 The probability weighting to use for the
5568 smp_call_function_many() function with a zero
5569 "wait" parameter. See weight_single.
5570 Note well that setting a high probability for
5571 this weighting can place serious IPI load
5574 scftorture.weight_many_wait= [KNL]
5575 The probability weighting to use for the
5576 smp_call_function_many() function with a
5577 non-zero "wait" parameter. See weight_single
5580 scftorture.weight_all= [KNL]
5581 The probability weighting to use for the
5582 smp_call_function_all() function with a zero
5583 "wait" parameter. See weight_single and
5586 scftorture.weight_all_wait= [KNL]
5587 The probability weighting to use for the
5588 smp_call_function_all() function with a
5589 non-zero "wait" parameter. See weight_single
5592 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5593 xtime_lock contention on larger systems, and/or RCU lock
5594 contention on all systems with CONFIG_MAXSMP set.
5595 Format: { "0" | "1" }
5596 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5598 Note: increases power consumption, thus should only be
5599 enabled if running jitter sensitive (HPC/RT) workloads.
5601 security= [SECURITY] Choose a legacy "major" security module to
5602 enable at boot. This has been deprecated by the
5605 selinux= [SELINUX] Disable or enable SELinux at boot time.
5606 Format: { "0" | "1" }
5607 See security/selinux/Kconfig help text.
5612 serialnumber [BUGS=X86-32]
5614 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
5617 Maximal number of shapers.
5619 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
5620 Limit apic dumping. The parameter defines the maximal
5621 number of local apics being dumped. Also it is possible
5622 to set it to "all" by meaning -- no limit here.
5623 Format: { 1 (default) | 2 | ... | all }.
5624 The parameter valid if only apic=debug or
5625 apic=verbose is specified.
5626 Example: apic=debug show_lapic=all
5634 Enable merging of slabs with similar size when the
5635 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5638 Disable merging of slabs with similar size. May be
5639 necessary if there is some reason to distinguish
5640 allocs to different slabs, especially in hardened
5641 environments where the risk of heap overflows and
5642 layout control by attackers can usually be
5643 frustrated by disabling merging. This will reduce
5644 most of the exposure of a heap attack to a single
5645 cache (risks via metadata attacks are mostly
5646 unchanged). Debug options disable merging on their
5648 For more information see Documentation/mm/slub.rst.
5650 slab_max_order= [MM, SLAB]
5651 Determines the maximum allowed order for slabs.
5652 A high setting may cause OOMs due to memory
5653 fragmentation. Defaults to 1 for systems with
5654 more than 32MB of RAM, 0 otherwise.
5656 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5657 Enabling slub_debug allows one to determine the
5658 culprit if slab objects become corrupted. Enabling
5659 slub_debug can create guard zones around objects and
5660 may poison objects when not in use. Also tracks the
5661 last alloc / free. For more information see
5662 Documentation/mm/slub.rst.
5664 slub_max_order= [MM, SLUB]
5665 Determines the maximum allowed order for slabs.
5666 A high setting may cause OOMs due to memory
5667 fragmentation. For more information see
5668 Documentation/mm/slub.rst.
5670 slub_min_objects= [MM, SLUB]
5671 The minimum number of objects per slab. SLUB will
5672 increase the slab order up to slub_max_order to
5673 generate a sufficiently large slab able to contain
5674 the number of objects indicated. The higher the number
5675 of objects the smaller the overhead of tracking slabs
5676 and the less frequently locks need to be acquired.
5677 For more information see Documentation/mm/slub.rst.
5679 slub_min_order= [MM, SLUB]
5680 Determines the minimum page order for slabs. Must be
5681 lower than slub_max_order.
5682 For more information see Documentation/mm/slub.rst.
5684 slub_merge [MM, SLUB]
5685 Same with slab_merge.
5687 slub_nomerge [MM, SLUB]
5688 Same with slab_nomerge. This is supported for legacy.
5689 See slab_nomerge for more information.
5692 Format: <io1>[,<io2>[,...,<io8>]]
5694 smp.csd_lock_timeout= [KNL]
5695 Specify the period of time in milliseconds
5696 that smp_call_function() and friends will wait
5697 for a CPU to release the CSD lock. This is
5698 useful when diagnosing bugs involving CPUs
5699 disabling interrupts for extended periods
5700 of time. Defaults to 5,000 milliseconds, and
5701 setting a value of zero disables this feature.
5702 This feature may be more efficiently disabled
5703 using the csdlock_debug- kernel parameter.
5705 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5706 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5707 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5708 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5709 smsc-ircc2.ircc_irq= [HW] IRQ line
5710 smsc-ircc2.ircc_dma= [HW] DMA channel
5711 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5712 0: Toshiba Satellite 1800 (GP data pin select)
5713 1: Fast pin select (default)
5716 smt= [KNL,S390] Set the maximum number of threads (logical
5717 CPUs) to use per physical CPU on systems capable of
5718 symmetric multithreading (SMT). Will be capped to the
5719 actual hardware limit.
5721 Default: -1 (no limit)
5724 [KNL] Should the soft-lockup detector generate panics.
5727 A value of 1 instructs the soft-lockup detector
5728 to panic the machine when a soft-lockup occurs. It is
5729 also controlled by the kernel.softlockup_panic sysctl
5730 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5731 respective build-time switch to that functionality.
5733 softlockup_all_cpu_backtrace=
5734 [KNL] Should the soft-lockup detector generate
5735 backtraces on all cpus.
5738 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5739 See Documentation/admin-guide/laptops/sonypi.rst
5741 spectre_v2= [X86] Control mitigation of Spectre variant 2
5742 (indirect branch speculation) vulnerability.
5743 The default operation protects the kernel from
5746 on - unconditionally enable, implies
5748 off - unconditionally disable, implies
5750 auto - kernel detects whether your CPU model is
5753 Selecting 'on' will, and 'auto' may, choose a
5754 mitigation method at run time according to the
5755 CPU, the available microcode, the setting of the
5756 CONFIG_RETPOLINE configuration option, and the
5757 compiler with which the kernel was built.
5759 Selecting 'on' will also enable the mitigation
5760 against user space to user space task attacks.
5762 Selecting 'off' will disable both the kernel and
5763 the user space protections.
5765 Specific mitigations can also be selected manually:
5767 retpoline - replace indirect branches
5768 retpoline,generic - Retpolines
5769 retpoline,lfence - LFENCE; indirect branch
5770 retpoline,amd - alias for retpoline,lfence
5771 eibrs - Enhanced/Auto IBRS
5772 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines
5773 eibrs,lfence - Enhanced/Auto IBRS + LFENCE
5774 ibrs - use IBRS to protect kernel
5776 Not specifying this option is equivalent to
5780 [X86] Control mitigation of Spectre variant 2
5781 (indirect branch speculation) vulnerability between
5784 on - Unconditionally enable mitigations. Is
5785 enforced by spectre_v2=on
5787 off - Unconditionally disable mitigations. Is
5788 enforced by spectre_v2=off
5790 prctl - Indirect branch speculation is enabled,
5791 but mitigation can be enabled via prctl
5792 per thread. The mitigation control state
5793 is inherited on fork.
5796 - Like "prctl" above, but only STIBP is
5797 controlled per thread. IBPB is issued
5798 always when switching between different user
5802 - Same as "prctl" above, but all seccomp
5803 threads will enable the mitigation unless
5804 they explicitly opt out.
5807 - Like "seccomp" above, but only STIBP is
5808 controlled per thread. IBPB is issued
5809 always when switching between different
5810 user space processes.
5812 auto - Kernel selects the mitigation depending on
5813 the available CPU features and vulnerability.
5815 Default mitigation: "prctl"
5817 Not specifying this option is equivalent to
5818 spectre_v2_user=auto.
5820 spec_store_bypass_disable=
5821 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5822 (Speculative Store Bypass vulnerability)
5824 Certain CPUs are vulnerable to an exploit against a
5825 a common industry wide performance optimization known
5826 as "Speculative Store Bypass" in which recent stores
5827 to the same memory location may not be observed by
5828 later loads during speculative execution. The idea
5829 is that such stores are unlikely and that they can
5830 be detected prior to instruction retirement at the
5831 end of a particular speculation execution window.
5833 In vulnerable processors, the speculatively forwarded
5834 store can be used in a cache side channel attack, for
5835 example to read memory to which the attacker does not
5836 directly have access (e.g. inside sandboxed code).
5838 This parameter controls whether the Speculative Store
5839 Bypass optimization is used.
5841 On x86 the options are:
5843 on - Unconditionally disable Speculative Store Bypass
5844 off - Unconditionally enable Speculative Store Bypass
5845 auto - Kernel detects whether the CPU model contains an
5846 implementation of Speculative Store Bypass and
5847 picks the most appropriate mitigation. If the
5848 CPU is not vulnerable, "off" is selected. If the
5849 CPU is vulnerable the default mitigation is
5850 architecture and Kconfig dependent. See below.
5851 prctl - Control Speculative Store Bypass per thread
5852 via prctl. Speculative Store Bypass is enabled
5853 for a process by default. The state of the control
5854 is inherited on fork.
5855 seccomp - Same as "prctl" above, but all seccomp threads
5856 will disable SSB unless they explicitly opt out.
5858 Default mitigations:
5861 On powerpc the options are:
5863 on,auto - On Power8 and Power9 insert a store-forwarding
5864 barrier on kernel entry and exit. On Power7
5865 perform a software flush on kernel entry and
5869 Not specifying this option is equivalent to
5870 spec_store_bypass_disable=auto.
5872 spia_io_base= [HW,MTD]
5878 [X86] Enable split lock detection or bus lock detection
5880 When enabled (and if hardware support is present), atomic
5881 instructions that access data across cache line
5882 boundaries will result in an alignment check exception
5883 for split lock detection or a debug exception for
5888 warn - the kernel will emit rate-limited warnings
5889 about applications triggering the #AC
5890 exception or the #DB exception. This mode is
5891 the default on CPUs that support split lock
5892 detection or bus lock detection. Default
5893 behavior is by #AC if both features are
5894 enabled in hardware.
5896 fatal - the kernel will send SIGBUS to applications
5897 that trigger the #AC exception or the #DB
5898 exception. Default behavior is by #AC if
5899 both features are enabled in hardware.
5902 Set system wide rate limit to N bus locks
5903 per second for bus lock detection.
5906 N/A for split lock detection.
5909 If an #AC exception is hit in the kernel or in
5910 firmware (i.e. not while executing in user mode)
5911 the kernel will oops in either "warn" or "fatal"
5914 #DB exception for bus lock is triggered only when
5918 Control the Special Register Buffer Data Sampling
5921 Certain CPUs are vulnerable to an MDS-like
5922 exploit which can leak bits from the random
5925 By default, this issue is mitigated by
5926 microcode. However, the microcode fix can cause
5927 the RDRAND and RDSEED instructions to become
5928 much slower. Among other effects, this will
5929 result in reduced throughput from /dev/urandom.
5931 The microcode mitigation can be disabled with
5932 the following option:
5934 off: Disable mitigation and remove
5935 performance impact to RDRAND and RDSEED
5937 srcutree.big_cpu_lim [KNL]
5938 Specifies the number of CPUs constituting a
5939 large system, such that srcu_struct structures
5940 should immediately allocate an srcu_node array.
5941 This kernel-boot parameter defaults to 128,
5942 but takes effect only when the low-order four
5943 bits of srcutree.convert_to_big is equal to 3
5946 srcutree.convert_to_big [KNL]
5947 Specifies under what conditions an SRCU tree
5948 srcu_struct structure will be converted to big
5949 form, that is, with an rcu_node tree:
5952 1: At init_srcu_struct() time.
5953 2: When rcutorture decides to.
5954 3: Decide at boot time (default).
5955 0x1X: Above plus if high contention.
5957 Either way, the srcu_node tree will be sized based
5958 on the actual runtime number of CPUs (nr_cpu_ids)
5959 instead of the compile-time CONFIG_NR_CPUS.
5961 srcutree.counter_wrap_check [KNL]
5962 Specifies how frequently to check for
5963 grace-period sequence counter wrap for the
5964 srcu_data structure's ->srcu_gp_seq_needed field.
5965 The greater the number of bits set in this kernel
5966 parameter, the less frequently counter wrap will
5967 be checked for. Note that the bottom two bits
5970 srcutree.exp_holdoff [KNL]
5971 Specifies how many nanoseconds must elapse
5972 since the end of the last SRCU grace period for
5973 a given srcu_struct until the next normal SRCU
5974 grace period will be considered for automatic
5975 expediting. Set to zero to disable automatic
5978 srcutree.srcu_max_nodelay [KNL]
5979 Specifies the number of no-delay instances
5980 per jiffy for which the SRCU grace period
5981 worker thread will be rescheduled with zero
5982 delay. Beyond this limit, worker thread will
5983 be rescheduled with a sleep delay of one jiffy.
5985 srcutree.srcu_max_nodelay_phase [KNL]
5986 Specifies the per-grace-period phase, number of
5987 non-sleeping polls of readers. Beyond this limit,
5988 grace period worker thread will be rescheduled
5989 with a sleep delay of one jiffy, between each
5990 rescan of the readers, for a grace period phase.
5992 srcutree.srcu_retry_check_delay [KNL]
5993 Specifies number of microseconds of non-sleeping
5994 delay between each non-sleeping poll of readers.
5996 srcutree.small_contention_lim [KNL]
5997 Specifies the number of update-side contention
5998 events per jiffy will be tolerated before
5999 initiating a conversion of an srcu_struct
6000 structure to big form. Note that the value of
6001 srcutree.convert_to_big must have the 0x10 bit
6002 set for contention-based conversions to occur.
6005 Speculative Store Bypass Disable control
6007 On CPUs that are vulnerable to the Speculative
6008 Store Bypass vulnerability and offer a
6009 firmware based mitigation, this parameter
6010 indicates how the mitigation should be used:
6012 force-on: Unconditionally enable mitigation for
6013 for both kernel and userspace
6014 force-off: Unconditionally disable mitigation for
6015 for both kernel and userspace
6016 kernel: Always enable mitigation in the
6017 kernel, and offer a prctl interface
6018 to allow userspace to register its
6019 interest in being mitigated too.
6021 stack_guard_gap= [MM]
6022 override the default stack gap protection. The value
6023 is in page units and it defines how many pages prior
6024 to (for stacks growing down) resp. after (for stacks
6025 growing up) the main stack are reserved for no other
6026 mapping. Default value is 256 pages.
6028 stack_depot_disable= [KNL]
6029 Setting this to true through kernel command line will
6030 disable the stack depot thereby saving the static memory
6031 consumed by the stack hash table. By default this is set
6035 Enabled the stack tracer on boot up.
6037 stacktrace_filter=[function-list]
6038 [FTRACE] Limit the functions that the stack tracer
6039 will trace at boot up. function-list is a comma-separated
6040 list of functions. This list can be changed at run
6041 time by the stack_trace_filter file in the debugfs
6042 tracing directory. Note, this enables stack tracing
6043 and the stacktrace above is not needed.
6047 Set the STI (builtin display/keyboard on the HP-PARISC
6048 machines) console (graphic card) which should be used
6049 as the initial boot-console.
6050 See also comment in drivers/video/console/sticore.c.
6053 See comment in drivers/video/console/sticore.c.
6056 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
6061 Enable or disable strict sigaltstack size checks
6062 against the required signal frame size which
6063 depends on the supported FPU features. This can
6064 be used to filter out binaries which have
6065 not yet been made aware of AT_MINSIGSTKSZ.
6068 Limits the number of kernel HPT entries in the hash
6069 page table to increase the rate of hash page table
6070 faults on kernel addresses.
6073 Limits the number of kernel SLB entries, and flushes
6074 them frequently to increase the rate of SLB faults
6075 on kernel addresses.
6077 sunrpc.min_resvport=
6078 sunrpc.max_resvport=
6080 SunRPC servers often require that client requests
6081 originate from a privileged port (i.e. a port in the
6082 range 0 < portnr < 1024).
6083 An administrator who wishes to reserve some of these
6084 ports for other uses may adjust the range that the
6085 kernel's sunrpc client considers to be privileged
6086 using these two parameters to set the minimum and
6087 maximum port values.
6089 sunrpc.svc_rpc_per_connection_limit=
6091 Limit the number of requests that the server will
6092 process in parallel from a single connection.
6093 The default value is 0 (no limit).
6097 Control how the NFS server code allocates CPUs to
6098 service thread pools. Depending on how many NICs
6099 you have and where their interrupts are bound, this
6100 option will affect which CPUs will do NFS serving.
6101 Note: this parameter cannot be changed while the
6102 NFS server is running.
6104 auto the server chooses an appropriate mode
6105 automatically using heuristics
6106 global a single global pool contains all CPUs
6107 percpu one pool for each CPU
6108 pernode one pool for each NUMA node (equivalent
6109 to global on non-NUMA machines)
6111 sunrpc.tcp_slot_table_entries=
6112 sunrpc.udp_slot_table_entries=
6114 Sets the upper limit on the number of simultaneous
6115 RPC calls that can be sent from the client to a
6116 server. Increasing these values may allow you to
6117 improve throughput, but will also increase the
6118 amount of memory reserved for use by the client.
6120 suspend.pm_test_delay=
6122 Sets the number of seconds to remain in a suspend test
6123 mode before resuming the system (see
6124 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6125 is set. Default value is 5.
6128 Format: { on | off | y | n | 1 | 0 }
6129 This parameter controls use of the Protected
6130 Execution Facility on pSeries.
6132 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
6133 Format: { <int> [,<int>] | force | noforce }
6134 <int> -- Number of I/O TLB slabs
6135 <int> -- Second integer after comma. Number of swiotlb
6136 areas with their own lock. Will be rounded up
6138 force -- force using of bounce buffers even if they
6139 wouldn't be automatically used by the kernel
6140 noforce -- Never use bounce buffers (for debugging)
6145 Set a sysctl parameter, right before loading the init
6146 process, as if the value was written to the respective
6147 /proc/sys/... file. Both '.' and '/' are recognized as
6148 separators. Unrecognized parameters and invalid values
6149 are reported in the kernel log. Sysctls registered
6150 later by a loaded module cannot be set this way.
6151 Example: sysctl.vm.swappiness=40
6153 sysfs.deprecated=0|1 [KNL]
6154 Enable/disable old style sysfs layout for old udev
6155 on older distributions. When this option is enabled
6156 very new udev will not work anymore. When this option
6157 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
6158 in older udev will not work anymore.
6159 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
6160 the kernel configuration.
6162 sysrq_always_enabled
6164 Ignore sysrq setting - this boot parameter will
6165 neutralize any effect of /proc/sys/kernel/sysrq.
6166 Useful for debugging.
6168 tcpmhash_entries= [KNL,NET]
6169 Set the number of tcp_metrics_hash slots.
6170 Default value is 8192 or 16384 depending on total
6171 ram pages. This is used to specify the TCP metrics
6172 cache size. See Documentation/networking/ip-sysctl.rst
6173 "tcp_no_metrics_save" section for more details.
6177 test_suspend= [SUSPEND]
6178 Format: { "mem" | "standby" | "freeze" }[,N]
6179 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6180 standby suspend) or "freeze" (for suspend type freeze)
6181 as the system sleep state during system startup with
6182 the optional capability to repeat N number of times.
6183 The system is woken from this state using a
6184 wakeup-capable RTC alarm.
6186 thash_entries= [KNL,NET]
6187 Set number of hash buckets for TCP connection
6189 thermal.act= [HW,ACPI]
6190 -1: disable all active trip points in all thermal zones
6191 <degrees C>: override all lowest active trip points
6193 thermal.crt= [HW,ACPI]
6194 -1: disable all critical trip points in all thermal zones
6195 <degrees C>: override all critical trip points
6197 thermal.nocrt= [HW,ACPI]
6198 Set to disable actions on ACPI thermal zone
6199 critical and hot trip points.
6201 thermal.off= [HW,ACPI]
6202 1: disable ACPI thermal control
6204 thermal.psv= [HW,ACPI]
6205 -1: disable all passive trip points
6206 <degrees C>: override all passive trip points to this
6209 thermal.tzp= [HW,ACPI]
6210 Specify global default ACPI thermal zone polling rate
6211 <deci-seconds>: poll all this frequency
6212 0: no polling (default)
6215 Force threading of all interrupt handlers except those
6216 marked explicitly IRQF_NO_THREAD.
6220 Specify if the kernel should make use of the cpu
6221 topology information if the hardware supports this.
6222 The scheduler will make use of this information and
6223 e.g. base its process migration decisions on it.
6226 topology_updates= [KNL, PPC, NUMA]
6228 Specify if the kernel should ignore (off)
6229 topology updates sent by the hypervisor to this
6232 torture.disable_onoff_at_boot= [KNL]
6233 Prevent the CPU-hotplug component of torturing
6234 until after init has spawned.
6236 torture.ftrace_dump_at_shutdown= [KNL]
6237 Dump the ftrace buffer at torture-test shutdown,
6238 even if there were no errors. This can be a
6239 very costly operation when many torture tests
6240 are running concurrently, especially on systems
6241 with rotating-rust storage.
6243 torture.verbose_sleep_frequency= [KNL]
6244 Specifies how many verbose printk()s should be
6245 emitted between each sleep. The default of zero
6246 disables verbose-printk() sleeping.
6248 torture.verbose_sleep_duration= [KNL]
6249 Duration of each verbose-printk() sleep in jiffies.
6253 tpm_suspend_pcr=[HW,TPM]
6254 Format: integer pcr id
6255 Specify that at suspend time, the tpm driver
6256 should extend the specified pcr with zeros,
6257 as a workaround for some chips which fail to
6258 flush the last written pcr on TPM_SaveState.
6259 This will guarantee that all the other pcrs
6263 Have the tracepoints sent to printk as well as the
6264 tracing ring buffer. This is useful for early boot up
6265 where the system hangs or reboots and does not give the
6266 option for reading the tracing buffer or performing a
6267 ftrace_dump_on_oops.
6269 To turn off having tracepoints sent to printk,
6270 echo 0 > /proc/sys/kernel/tracepoint_printk
6271 Note, echoing 1 into this file without the
6272 tracepoint_printk kernel cmdline option has no effect.
6274 The tp_printk_stop_on_boot (see below) can also be used
6275 to stop the printing of events to console at
6280 Having tracepoints sent to printk() and activating high
6281 frequency tracepoints such as irq or sched, can cause
6282 the system to live lock.
6284 tp_printk_stop_on_boot [FTRACE]
6285 When tp_printk (above) is set, it can cause a lot of noise
6286 on the console. It may be useful to only include the
6287 printing of events during boot up, as user space may
6288 make the system inoperable.
6290 This command line option will stop the printing of events
6291 to console at the late_initcall_sync() time frame.
6293 trace_buf_size=nn[KMG]
6294 [FTRACE] will set tracing buffer size on each cpu.
6296 trace_clock= [FTRACE] Set the clock used for tracing events
6298 local - Use the per CPU time stamp counter
6299 (converted into nanoseconds). Fast, but
6300 depending on the architecture, may not be
6301 in sync between CPUs.
6302 global - Event time stamps are synchronize across
6303 CPUs. May be slower than the local clock,
6304 but better for some race conditions.
6305 counter - Simple counting of events (1, 2, ..)
6306 note, some counts may be skipped due to the
6307 infrastructure grabbing the clock more than
6309 uptime - Use jiffies as the time stamp.
6310 perf - Use the same clock that perf uses.
6311 mono - Use ktime_get_mono_fast_ns() for time stamps.
6312 mono_raw - Use ktime_get_raw_fast_ns() for time
6314 boot - Use ktime_get_boot_fast_ns() for time stamps.
6315 Architectures may add more clocks. See
6316 Documentation/trace/ftrace.rst for more details.
6318 trace_event=[event-list]
6319 [FTRACE] Set and start specified trace events in order
6320 to facilitate early boot debugging. The event-list is a
6321 comma-separated list of trace events to enable. See
6322 also Documentation/trace/events.rst
6324 trace_instance=[instance-info]
6325 [FTRACE] Create a ring buffer instance early in boot up.
6326 This will be listed in:
6328 /sys/kernel/tracing/instances
6330 Events can be enabled at the time the instance is created
6333 trace_instance=<name>,<system1>:<event1>,<system2>:<event2>
6335 Note, the "<system*>:" portion is optional if the event is
6338 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall
6340 will enable the "sched_switch" event (note, the "sched:" is optional, and
6341 the same thing would happen if it was left off). The irq_handler_entry
6342 event, and all events under the "initcall" system.
6344 trace_options=[option-list]
6345 [FTRACE] Enable or disable tracer options at boot.
6346 The option-list is a comma delimited list of options
6347 that can be enabled or disabled just as if you were
6348 to echo the option name into
6350 /sys/kernel/tracing/trace_options
6352 For example, to enable stacktrace option (to dump the
6353 stack trace of each event), add to the command line:
6355 trace_options=stacktrace
6357 See also Documentation/trace/ftrace.rst "trace options"
6360 trace_trigger=[trigger-list]
6361 [FTRACE] Add a event trigger on specific events.
6362 Set a trigger on top of a specific event, with an optional
6365 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..."
6366 Where more than one trigger may be specified that are comma deliminated.
6370 trace_trigger="sched_switch.stacktrace if prev_state == 2"
6372 The above will enable the "stacktrace" trigger on the "sched_switch"
6373 event but only trigger it if the "prev_state" of the "sched_switch"
6374 event is "2" (TASK_UNINTERUPTIBLE).
6376 See also "Event triggers" in Documentation/trace/events.rst
6380 [FTRACE] enable this option to disable tracing when a
6381 warning is hit. This turns off "tracing_on". Tracing can
6382 be enabled again by echoing '1' into the "tracing_on"
6383 file located in /sys/kernel/tracing/
6385 This option is useful, as it disables the trace before
6386 the WARNING dump is called, which prevents the trace to
6387 be filled with content caused by the warning output.
6389 This option can also be set at run time via the sysctl
6390 option: kernel/traceoff_on_warning
6392 transparent_hugepage=
6394 Format: [always|madvise|never]
6395 Can be used to control the default behavior of the system
6396 with respect to transparent hugepages.
6397 See Documentation/admin-guide/mm/transhuge.rst
6400 trusted.source= [KEYS]
6402 This parameter identifies the trust source as a backend
6403 for trusted keys implementation. Supported trust
6408 If not specified then it defaults to iterating through
6409 the trust source list starting with TPM and assigns the
6410 first trust source as a backend which is initialized
6411 successfully during iteration.
6415 The RNG used to generate key material for trusted keys.
6418 - the same value as trusted.source: "tpm" or "tee"
6420 If not specified, "default" is used. In this case,
6421 the RNG's choice is left to each individual trust source.
6423 tsc= Disable clocksource stability checks for TSC.
6425 [x86] reliable: mark tsc clocksource as reliable, this
6426 disables clocksource verification at runtime, as well
6427 as the stability checks done at bootup. Used to enable
6428 high-resolution timer mode on older hardware, and in
6429 virtualized environment.
6430 [x86] noirqtime: Do not use TSC to do irq accounting.
6431 Used to run time disable IRQ_TIME_ACCOUNTING on any
6432 platforms where RDTSC is slow and this accounting
6434 [x86] unstable: mark the TSC clocksource as unstable, this
6435 marks the TSC unconditionally unstable at bootup and
6436 avoids any further wobbles once the TSC watchdog notices.
6437 [x86] nowatchdog: disable clocksource watchdog. Used
6438 in situations with strict latency requirements (where
6439 interruptions from clocksource watchdog are not
6441 [x86] recalibrate: force recalibration against a HW timer
6442 (HPET or PM timer) on systems whose TSC frequency was
6443 obtained from HW or FW using either an MSR or CPUID(0x15).
6444 Warn if the difference is more than 500 ppm.
6445 [x86] watchdog: Use TSC as the watchdog clocksource with
6446 which to check other HW timers (HPET or PM timer), but
6447 only on systems where TSC has been deemed trustworthy.
6448 This will be suppressed by an earlier tsc=nowatchdog and
6449 can be overridden by a later tsc=nowatchdog. A console
6450 message will flag any such suppression or overriding.
6452 tsc_early_khz= [X86] Skip early TSC calibration and use the given
6453 value instead. Useful when the early TSC frequency discovery
6454 procedure is not reliable, such as on overclocked systems
6455 with CPUID.16h support and partial CPUID.15h support.
6456 Format: <unsigned int>
6458 tsx= [X86] Control Transactional Synchronization
6459 Extensions (TSX) feature in Intel processors that
6460 support TSX control.
6462 This parameter controls the TSX feature. The options are:
6464 on - Enable TSX on the system. Although there are
6465 mitigations for all known security vulnerabilities,
6466 TSX has been known to be an accelerator for
6467 several previous speculation-related CVEs, and
6468 so there may be unknown security risks associated
6469 with leaving it enabled.
6471 off - Disable TSX on the system. (Note that this
6472 option takes effect only on newer CPUs which are
6473 not vulnerable to MDS, i.e., have
6474 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6475 the new IA32_TSX_CTRL MSR through a microcode
6476 update. This new MSR allows for the reliable
6477 deactivation of the TSX functionality.)
6479 auto - Disable TSX if X86_BUG_TAA is present,
6480 otherwise enable TSX on the system.
6482 Not specifying this option is equivalent to tsx=off.
6484 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6487 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6488 Abort (TAA) vulnerability.
6490 Similar to Micro-architectural Data Sampling (MDS)
6491 certain CPUs that support Transactional
6492 Synchronization Extensions (TSX) are vulnerable to an
6493 exploit against CPU internal buffers which can forward
6494 information to a disclosure gadget under certain
6497 In vulnerable processors, the speculatively forwarded
6498 data can be used in a cache side channel attack, to
6499 access data to which the attacker does not have direct
6502 This parameter controls the TAA mitigation. The
6505 full - Enable TAA mitigation on vulnerable CPUs
6508 full,nosmt - Enable TAA mitigation and disable SMT on
6509 vulnerable CPUs. If TSX is disabled, SMT
6510 is not disabled because CPU is not
6511 vulnerable to cross-thread TAA attacks.
6512 off - Unconditionally disable TAA mitigation
6514 On MDS-affected machines, tsx_async_abort=off can be
6515 prevented by an active MDS mitigation as both vulnerabilities
6516 are mitigated with the same mechanism so in order to disable
6517 this mitigation, you need to specify mds=off too.
6519 Not specifying this option is equivalent to
6520 tsx_async_abort=full. On CPUs which are MDS affected
6521 and deploy MDS mitigation, TAA mitigation is not
6522 required and doesn't provide any additional
6526 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6528 turbografx.map[2|3]= [HW,JOY]
6529 TurboGraFX parallel port interface
6531 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6532 See also Documentation/input/devices/joystick-parport.rst
6534 udbg-immortal [PPC] When debugging early kernel crashes that
6535 happen after console_init() and before a proper
6536 console driver takes over, this boot options might
6537 help "seeing" what's going on.
6539 uhash_entries= [KNL,NET]
6540 Set number of hash buckets for UDP/UDP-Lite connections
6543 [USB] Ignore overcurrent events (default N).
6544 Some badly-designed motherboards generate lots of
6545 bogus events, for ports that aren't wired to
6546 anything. Set this parameter to avoid log spamming.
6547 Note that genuine overcurrent events won't be
6551 [X86] Cause panic on unknown NMI.
6553 usbcore.authorized_default=
6554 [USB] Default USB device authorization:
6555 (default -1 = authorized except for wireless USB,
6556 0 = not authorized, 1 = authorized, 2 = authorized
6557 if device connected to internal port)
6559 usbcore.autosuspend=
6560 [USB] The autosuspend time delay (in seconds) used
6561 for newly-detected USB devices (default 2). This
6562 is the time required before an idle device will be
6563 autosuspended. Devices for which the delay is set
6564 to a negative value won't be autosuspended at all.
6566 usbcore.usbfs_snoop=
6567 [USB] Set to log all usbfs traffic (default 0 = off).
6569 usbcore.usbfs_snoop_max=
6570 [USB] Maximum number of bytes to snoop in each URB
6573 usbcore.blinkenlights=
6574 [USB] Set to cycle leds on hubs (default 0 = off).
6576 usbcore.old_scheme_first=
6577 [USB] Start with the old device initialization
6578 scheme (default 0 = off).
6580 usbcore.usbfs_memory_mb=
6581 [USB] Memory limit (in MB) for buffers allocated by
6582 usbfs (default = 16, 0 = max = 2047).
6584 usbcore.use_both_schemes=
6585 [USB] Try the other device initialization scheme
6586 if the first one fails (default 1 = enabled).
6588 usbcore.initial_descriptor_timeout=
6589 [USB] Specifies timeout for the initial 64-byte
6590 USB_REQ_GET_DESCRIPTOR request in milliseconds
6591 (default 5000 = 5.0 seconds).
6593 usbcore.nousb [USB] Disable the USB subsystem
6596 [USB] A list of quirk entries to augment the built-in
6597 usb core quirk list. List entries are separated by
6598 commas. Each entry has the form
6599 VendorID:ProductID:Flags. The IDs are 4-digit hex
6600 numbers and Flags is a set of letters. Each letter
6601 will change the built-in quirk; setting it if it is
6602 clear and clearing it if it is set. The letters have
6603 the following meanings:
6604 a = USB_QUIRK_STRING_FETCH_255 (string
6605 descriptors must not be fetched using
6607 b = USB_QUIRK_RESET_RESUME (device can't resume
6608 correctly so reset it instead);
6609 c = USB_QUIRK_NO_SET_INTF (device can't handle
6610 Set-Interface requests);
6611 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6612 handle its Configuration or Interface
6614 e = USB_QUIRK_RESET (device can't be reset
6615 (e.g morph devices), don't use reset);
6616 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6617 more interface descriptions than the
6618 bNumInterfaces count, and can't handle
6619 talking to these interfaces);
6620 g = USB_QUIRK_DELAY_INIT (device needs a pause
6621 during initialization, after we read
6622 the device descriptor);
6623 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6624 high speed and super speed interrupt
6625 endpoints, the USB 2.0 and USB 3.0 spec
6626 require the interval in microframes (1
6627 microframe = 125 microseconds) to be
6628 calculated as interval = 2 ^
6630 Devices with this quirk report their
6631 bInterval as the result of this
6632 calculation instead of the exponent
6633 variable used in the calculation);
6634 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6635 handle device_qualifier descriptor
6637 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6638 generates spurious wakeup, ignore
6639 remote wakeup capability);
6640 k = USB_QUIRK_NO_LPM (device can't handle Link
6642 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6643 (Device reports its bInterval as linear
6644 frames instead of the USB 2.0
6646 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6647 to be disconnected before suspend to
6648 prevent spurious wakeup);
6649 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6650 pause after every control message);
6651 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6652 delay after resetting its port);
6653 Example: quirks=0781:5580:bk,0a5c:5834:gij
6656 [USBHID] The interval which mice are to be polled at.
6659 [USBHID] The interval which joysticks are to be polled at.
6662 [USBHID] The interval which keyboards are to be polled at.
6664 usb-storage.delay_use=
6665 [UMS] The delay in seconds before a new device is
6666 scanned for Logical Units (default 1).
6669 [UMS] A list of quirks entries to supplement or
6670 override the built-in unusual_devs list. List
6671 entries are separated by commas. Each entry has
6672 the form VID:PID:Flags where VID and PID are Vendor
6673 and Product ID values (4-digit hex numbers) and
6674 Flags is a set of characters, each corresponding
6675 to a common usb-storage quirk flag as follows:
6676 a = SANE_SENSE (collect more than 18 bytes
6677 of sense data, not on uas);
6678 b = BAD_SENSE (don't collect more than 18
6679 bytes of sense data, not on uas);
6680 c = FIX_CAPACITY (decrease the reported
6681 device capacity by one sector);
6682 d = NO_READ_DISC_INFO (don't use
6683 READ_DISC_INFO command, not on uas);
6684 e = NO_READ_CAPACITY_16 (don't use
6685 READ_CAPACITY_16 command);
6686 f = NO_REPORT_OPCODES (don't use report opcodes
6688 g = MAX_SECTORS_240 (don't transfer more than
6689 240 sectors at a time, uas only);
6690 h = CAPACITY_HEURISTICS (decrease the
6691 reported device capacity by one
6692 sector if the number is odd);
6693 i = IGNORE_DEVICE (don't bind to this
6695 j = NO_REPORT_LUNS (don't use report luns
6697 k = NO_SAME (do not use WRITE_SAME, uas only)
6698 l = NOT_LOCKABLE (don't try to lock and
6699 unlock ejectable media, not on uas);
6700 m = MAX_SECTORS_64 (don't transfer more
6701 than 64 sectors = 32 KB at a time,
6703 n = INITIAL_READ10 (force a retry of the
6704 initial READ(10) command, not on uas);
6705 o = CAPACITY_OK (accept the capacity
6706 reported by the device, not on uas);
6707 p = WRITE_CACHE (the device cache is ON
6708 by default, not on uas);
6709 r = IGNORE_RESIDUE (the device reports
6710 bogus residue values, not on uas);
6711 s = SINGLE_LUN (the device has only one
6713 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6714 commands, uas only);
6715 u = IGNORE_UAS (don't bind to the uas driver);
6716 w = NO_WP_DETECT (don't test whether the
6717 medium is write-protected).
6718 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6719 even if the device claims no cache,
6721 Example: quirks=0419:aaf5:rl,0421:0433:rc
6723 user_debug= [KNL,ARM]
6725 See arch/arm/Kconfig.debug help text.
6726 1 - undefined instruction events
6728 4 - invalid data aborts
6731 Example: user_debug=31
6734 [X86] Flags controlling user PTE allocations.
6736 nohigh = do not allocate PTE pages in
6737 HIGHMEM regardless of setting
6740 vdso= [X86,SH,SPARC]
6741 On X86_32, this is an alias for vdso32=. Otherwise:
6743 vdso=1: enable VDSO (the default)
6744 vdso=0: disable VDSO mapping
6746 vdso32= [X86] Control the 32-bit vDSO
6747 vdso32=1: enable 32-bit VDSO
6748 vdso32=0 or vdso32=2: disable 32-bit VDSO
6750 See the help text for CONFIG_COMPAT_VDSO for more
6751 details. If CONFIG_COMPAT_VDSO is set, the default is
6752 vdso32=0; otherwise, the default is vdso32=1.
6754 For compatibility with older kernels, vdso32=2 is an
6757 Try vdso32=0 if you encounter an error that says:
6758 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6761 vector=percpu: enable percpu vector domain
6763 video= [FB] Frame buffer configuration
6764 See Documentation/fb/modedb.rst.
6766 video.brightness_switch_enabled= [ACPI]
6768 If set to 1, on receiving an ACPI notify event
6769 generated by hotkey, video driver will adjust brightness
6770 level and then send out the event to user space through
6771 the allocated input device. If set to 0, video driver
6772 will only send out the event without touching backlight
6777 [VMMIO] Memory mapped virtio (platform) device.
6779 <size>@<baseaddr>:<irq>[:<id>]
6781 <size> := size (can use standard suffixes
6783 <baseaddr> := physical base address
6784 <irq> := interrupt number (as passed to
6786 <id> := (optional) platform device id
6788 virtio_mmio.device=1K@0x100b0000:48:7
6790 Can be used multiple times for multiple devices.
6792 vga= [BOOT,X86-32] Select a particular video mode
6793 See Documentation/arch/x86/boot.rst and
6794 Documentation/admin-guide/svga.rst.
6795 Use vga=ask for menu.
6796 This is actually a boot loader parameter; the value is
6797 passed to the kernel using a special protocol.
6799 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6800 May slow down system boot speed, especially when
6801 enabled on systems with a large amount of memory.
6802 All options are enabled by default, and this
6803 interface is meant to allow for selectively
6804 enabling or disabling specific virtual memory
6807 Available options are:
6808 P Enable page structure init time poisoning
6809 - Disable all of the above options
6811 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6812 size of <nn>. This can be used to increase the
6813 minimum size (128MB on x86). It can also be used to
6814 decrease the size and leave more room for directly
6817 vmcp_cma=nn[MG] [KNL,S390]
6818 Sets the memory size reserved for contiguous memory
6819 allocations for the vmcp device driver.
6821 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6824 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6827 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6831 Controls the behavior of vsyscalls (i.e. calls to
6832 fixed addresses of 0xffffffffff600x00 from legacy
6833 code). Most statically-linked binaries and older
6834 versions of glibc use these calls. Because these
6835 functions are at fixed addresses, they make nice
6836 targets for exploits that can control RIP.
6838 emulate Vsyscalls turn into traps and are emulated
6839 reasonably safely. The vsyscall page is
6842 xonly [default] Vsyscalls turn into traps and are
6843 emulated reasonably safely. The vsyscall
6844 page is not readable.
6846 none Vsyscalls don't work at all. This makes
6847 them quite hard to use for exploits but
6848 might break your system.
6850 vt.color= [VT] Default text color.
6851 Format: 0xYX, X = foreground, Y = background.
6852 Default: 0x07 = light gray on black.
6854 vt.cur_default= [VT] Default cursor shape.
6855 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6856 the parameters of the <Esc>[?A;B;Cc escape sequence;
6857 see VGA-softcursor.txt. Default: 2 = underline.
6859 vt.default_blu= [VT]
6860 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6861 Change the default blue palette of the console.
6862 This is a 16-member array composed of values
6865 vt.default_grn= [VT]
6866 Format: <green0>,<green1>,<green2>,...,<green15>
6867 Change the default green palette of the console.
6868 This is a 16-member array composed of values
6871 vt.default_red= [VT]
6872 Format: <red0>,<red1>,<red2>,...,<red15>
6873 Change the default red palette of the console.
6874 This is a 16-member array composed of values
6880 Set system-wide default UTF-8 mode for all tty's.
6881 Default is 1, i.e. UTF-8 mode is enabled for all
6882 newly opened terminals.
6884 vt.global_cursor_default=
6887 Set system-wide default for whether a cursor
6888 is shown on new VTs. Default is -1,
6889 i.e. cursors will be created by default unless
6890 overridden by individual drivers. 0 will hide
6891 cursors, 1 will display them.
6893 vt.italic= [VT] Default color for italic text; 0-15.
6896 vt.underline= [VT] Default color for underlined text; 0-15.
6899 watchdog timers [HW,WDT] For information on watchdog timers,
6900 see Documentation/watchdog/watchdog-parameters.rst
6901 or other driver-specific files in the
6902 Documentation/watchdog/ directory.
6906 Set the hard lockup detector stall duration
6907 threshold in seconds. The soft lockup detector
6908 threshold is set to twice the value. A value of 0
6909 disables both lockup detectors. Default is 10
6912 workqueue.watchdog_thresh=
6913 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6914 warn stall conditions and dump internal state to
6915 help debugging. 0 disables workqueue stall
6916 detection; otherwise, it's the stall threshold
6917 duration in seconds. The default value is 30 and
6918 it can be updated at runtime by writing to the
6919 corresponding sysfs file.
6921 workqueue.disable_numa
6922 By default, all work items queued to unbound
6923 workqueues are affine to the NUMA nodes they're
6924 issued on, which results in better behavior in
6925 general. If NUMA affinity needs to be disabled for
6926 whatever reason, this option can be used. Note
6927 that this also can be controlled per-workqueue for
6928 workqueues visible under /sys/bus/workqueue/.
6930 workqueue.power_efficient
6931 Per-cpu workqueues are generally preferred because
6932 they show better performance thanks to cache
6933 locality; unfortunately, per-cpu workqueues tend to
6934 be more power hungry than unbound workqueues.
6936 Enabling this makes the per-cpu workqueues which
6937 were observed to contribute significantly to power
6938 consumption unbound, leading to measurably lower
6939 power usage at the cost of small performance
6942 The default value of this parameter is determined by
6943 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6945 workqueue.debug_force_rr_cpu
6946 Workqueue used to implicitly guarantee that work
6947 items queued without explicit CPU specified are put
6948 on the local CPU. This guarantee is no longer true
6949 and while local CPU is still preferred work items
6950 may be put on foreign CPUs. This debug option
6951 forces round-robin CPU selection to flush out
6952 usages which depend on the now broken guarantee.
6953 When enabled, memory and cache locality will be
6956 writecombine= [LOONGARCH] Control the MAT (Memory Access Type) of
6959 on - Enable writecombine, use WUC for ioremap_wc()
6960 off - Disable writecombine, use SUC for ioremap_wc()
6962 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6963 default x2apic cluster mode on platforms
6966 xen_512gb_limit [KNL,X86-64,XEN]
6967 Restricts the kernel running paravirtualized under Xen
6968 to use only up to 512 GB of RAM. The reason to do so is
6969 crash analysis tools and Xen tools for doing domain
6970 save/restore/migration must be enabled to handle larger
6973 xen_emul_unplug= [HW,X86,XEN]
6974 Unplug Xen emulated devices
6975 Format: [unplug0,][unplug1]
6976 ide-disks -- unplug primary master IDE devices
6977 aux-ide-disks -- unplug non-primary-master IDE devices
6978 nics -- unplug network devices
6979 all -- unplug all emulated devices (NICs and IDE disks)
6980 unnecessary -- unplugging emulated devices is
6981 unnecessary even if the host did not respond to
6983 never -- do not unplug even if version check succeeds
6985 xen_legacy_crash [X86,XEN]
6986 Crash from Xen panic notifier, without executing late
6987 panic() code such as dumping handler.
6989 xen_msr_safe= [X86,XEN]
6991 Select whether to always use non-faulting (safe) MSR
6992 access functions when running as Xen PV guest. The
6993 default value is controlled by CONFIG_XEN_PV_MSR_SAFE.
6995 xen_nopvspin [X86,XEN]
6996 Disables the qspinlock slowpath using Xen PV optimizations.
6997 This parameter is obsoleted by "nopvspin" parameter, which
6998 has equivalent effect for XEN platform.
7001 Disables the PV optimizations forcing the HVM guest to
7002 run as generic HVM guest with no PV drivers.
7003 This option is obsoleted by the "nopv" option, which
7004 has equivalent effect for XEN platform.
7006 xen_no_vector_callback
7007 [KNL,X86,XEN] Disable the vector callback for Xen
7008 event channel interrupts.
7010 xen_scrub_pages= [XEN]
7011 Boolean option to control scrubbing pages before giving them back
7012 to Xen, for use by other domains. Can be also changed at runtime
7013 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
7014 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
7016 xen_timer_slop= [X86-64,XEN]
7017 Set the timer slop (in nanoseconds) for the virtual Xen
7018 timers (default is 100000). This adjusts the minimum
7019 delta of virtualized Xen timers, where lower values
7020 improve timer resolution at the expense of processing
7021 more timer interrupts.
7023 xen.balloon_boot_timeout= [XEN]
7024 The time (in seconds) to wait before giving up to boot
7025 in case initial ballooning fails to free enough memory.
7026 Applies only when running as HVM or PVH guest and
7027 started with less memory configured than allowed at
7028 max. Default is 180.
7030 xen.event_eoi_delay= [XEN]
7031 How long to delay EOI handling in case of event
7032 storms (jiffies). Default is 10.
7034 xen.event_loop_timeout= [XEN]
7035 After which time (jiffies) the event handling loop
7036 should start to delay EOI handling. Default is 2.
7038 xen.fifo_events= [XEN]
7039 Boolean parameter to disable using fifo event handling
7040 even if available. Normally fifo event handling is
7041 preferred over the 2-level event handling, as it is
7042 fairer and the number of possible event channels is
7043 much higher. Default is on (use fifo events).
7045 xirc2ps_cs= [NET,PCMCIA]
7047 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
7050 By default on POWER9 and above, the kernel will
7051 natively use the XIVE interrupt controller. This option
7052 allows the fallback firmware mode to be used:
7054 off Fallback to firmware control of XIVE interrupt
7055 controller on both pseries and powernv
7056 platforms. Only useful on POWER9 and above.
7058 xive.store-eoi=off [PPC]
7059 By default on POWER10 and above, the kernel will use
7060 stores for EOI handling when the XIVE interrupt mode
7061 is active. This option allows the XIVE driver to use
7062 loads instead, as on POWER9.
7064 xhci-hcd.quirks [USB,KNL]
7065 A hex value specifying bitmask with supplemental xhci
7066 host controller quirks. Meaning of each bit can be
7067 consulted in header drivers/usb/host/xhci.h.
7070 Format: { early | on | rw | ro | off }
7071 Controls if xmon debugger is enabled. Default is off.
7072 Passing only "xmon" is equivalent to "xmon=early".
7073 early Call xmon as early as possible on boot; xmon
7074 debugger is called from setup_arch().
7075 on xmon debugger hooks will be installed so xmon
7076 is only called on a kernel crash. Default mode,
7077 i.e. either "ro" or "rw" mode, is controlled
7078 with CONFIG_XMON_DEFAULT_RO_MODE.
7079 rw xmon debugger hooks will be installed so xmon
7080 is called only on a kernel crash, mode is write,
7081 meaning SPR registers, memory and, other data
7082 can be written using xmon commands.
7083 ro same as "rw" option above but SPR registers,
7084 memory, and other data can't be written using
7086 off xmon is disabled.
projects / linux-block.git / blob