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
325 amd_iommu_dump= [HW,X86-64]
326 Enable AMD IOMMU driver option to dump the ACPI table
327 for AMD IOMMU. With this option enabled, AMD IOMMU
328 driver will print ACPI tables for AMD IOMMU during
329 IOMMU initialization.
331 amd_iommu_intr= [HW,X86-64]
332 Specifies one of the following AMD IOMMU interrupt
334 legacy - Use legacy interrupt remapping mode.
335 vapic - Use virtual APIC mode, which allows IOMMU
336 to inject interrupts directly into guest.
337 This mode requires kvm-amd.avic=1.
338 (Default when IOMMU HW support is present.)
340 amijoy.map= [HW,JOY] Amiga joystick support
341 Map of devices attached to JOY0DAT and JOY1DAT
343 See also Documentation/input/joydev/joystick.rst
345 analog.map= [HW,JOY] Analog joystick and gamepad support
346 Specifies type or capabilities of an analog joystick
347 connected to one of 16 gameports
348 Format: <type1>,<type2>,..<type16>
351 Power management functions (SPARCstation-4/5 + deriv.)
353 Disable APC CPU standby support. SPARCstation-Fox does
354 not play well with APC CPU idle - disable it if you have
355 APC and your system crashes randomly.
357 apic= [APIC,X86] Advanced Programmable Interrupt Controller
358 Change the output verbosity while booting
359 Format: { quiet (default) | verbose | debug }
360 Change the amount of debugging information output
361 when initialising the APIC and IO-APIC components.
362 For X86-32, this can also be used to specify an APIC
364 Format: apic=driver_name
365 Examples: apic=bigsmp
367 apic_extnmi= [APIC,X86] External NMI delivery setting
368 Format: { bsp (default) | all | none }
369 bsp: External NMI is delivered only to CPU 0
370 all: External NMIs are broadcast to all CPUs as a
372 none: External NMI is masked for all CPUs. This is
373 useful so that a dump capture kernel won't be
377 See Documentation/networking/ipv6.rst.
379 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
380 Limit apic dumping. The parameter defines the maximal
381 number of local apics being dumped. Also it is possible
382 to set it to "all" by meaning -- no limit here.
383 Format: { 1 (default) | 2 | ... | all }.
384 The parameter valid if only apic=debug or
385 apic=verbose is specified.
386 Example: apic=debug show_lapic=all
388 apm= [APM] Advanced Power Management
389 See header of arch/x86/kernel/apm_32.c.
391 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
392 Format: <io>,<irq>,<nodeID>
394 arm64.nobti [ARM64] Unconditionally disable Branch Target
395 Identification support
397 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
400 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
405 atarimouse= [HW,MOUSE] Atari Mouse
407 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
408 EzKey and similar keyboards
410 atkbd.reset= [HW] Reset keyboard during initialization
412 atkbd.set= [HW] Select keyboard code set
413 Format: <int> (2 = AT (default), 3 = PS/2)
415 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
418 atkbd.softraw= [HW] Choose between synthetic and real raw mode
419 Format: <bool> (0 = real, 1 = synthetic (default))
421 atkbd.softrepeat= [HW]
422 Use software keyboard repeat
424 audit= [KNL] Enable the audit sub-system
425 Format: { "0" | "1" | "off" | "on" }
426 0 | off - kernel audit is disabled and can not be
427 enabled until the next reboot
428 unset - kernel audit is initialized but disabled and
429 will be fully enabled by the userspace auditd.
430 1 | on - kernel audit is initialized and partially
431 enabled, storing at most audit_backlog_limit
432 messages in RAM until it is fully enabled by the
436 audit_backlog_limit= [KNL] Set the audit queue size limit.
437 Format: <int> (must be >=0)
440 bau= [X86_UV] Enable the BAU on SGI UV. The default
441 behavior is to disable the BAU (i.e. bau=0).
442 Format: { "0" | "1" }
445 unset - Disable the BAU.
447 baycom_epp= [HW,AX25]
450 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
452 See header of drivers/net/hamradio/baycom_par.c.
454 baycom_ser_fdx= [HW,AX25]
455 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
456 Format: <io>,<irq>,<mode>[,<baud>]
457 See header of drivers/net/hamradio/baycom_ser_fdx.c.
459 baycom_ser_hdx= [HW,AX25]
460 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
461 Format: <io>,<irq>,<mode>
462 See header of drivers/net/hamradio/baycom_ser_hdx.c.
465 Disable BERT OS support on buggy BIOSes.
467 bgrt_disable [ACPI][X86]
468 Disable BGRT to avoid flickering OEM logo.
470 blkdevparts= Manual partition parsing of block device(s) for
471 embedded devices based on command line input.
472 See Documentation/block/cmdline-partition.rst
474 boot_delay= Milliseconds to delay each printk during boot.
475 Values larger than 10 seconds (10000) are changed to
480 Extended command line options can be added to an initrd
481 and this will cause the kernel to look for it.
483 See Documentation/admin-guide/bootconfig.rst
485 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
486 bttv.radio= Most important insmod options are available as
488 bttv.pll= See Documentation/admin-guide/media/bttv.rst
491 bulk_remove=off [PPC] This parameter disables the use of the pSeries
492 firmware feature for flushing multiple hpte entries
495 c101= [NET] Moxa C101 synchronous serial card
497 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
498 Sometimes CPU hardware bugs make them report the cache
499 size incorrectly. The kernel will attempt work arounds
500 to fix known problems, but for some CPUs it is not
501 possible to determine what the correct size should be.
502 This option provides an override for these situations.
505 [NET] Specifies amount of time (in seconds) that
506 the kernel should wait for a network carrier. By default
507 it waits 120 seconds.
509 ca_keys= [KEYS] This parameter identifies a specific key(s) on
510 the system trusted keyring to be used for certificate
512 format: { id:<keyid> | builtin }
514 cca= [MIPS] Override the kernel pages' cache coherency
515 algorithm. Accepted values range from 0 to 7
516 inclusive. See arch/mips/include/asm/pgtable-bits.h
517 for platform specific values (SB1, Loongson3 and
520 ccw_timeout_log [S390]
521 See Documentation/s390/common_io.rst for details.
523 cgroup_disable= [KNL] Disable a particular controller or optional feature
524 Format: {name of the controller(s) or feature(s) to disable}
525 The effects of cgroup_disable=foo are:
526 - foo isn't auto-mounted if you mount all cgroups in
528 - foo isn't visible as an individually mountable
530 - if foo is an optional feature then the feature is
531 disabled and corresponding cgroup files are not
533 {Currently only "memory" controller deal with this and
534 cut the overhead, others just disable the usage. So
535 only cgroup_disable=memory is actually worthy}
536 Specifying "pressure" disables per-cgroup pressure
537 stall information accounting feature
539 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
540 Format: { { controller | "all" | "named" }
541 [,{ controller | "all" | "named" }...] }
542 Like cgroup_disable, but only applies to cgroup v1;
543 the blacklisted controllers remain available in cgroup2.
544 "all" blacklists all controllers and "named" disables
545 named mounts. Specifying both "all" and "named" disables
548 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
550 nosocket -- Disable socket memory accounting.
551 nokmem -- Disable kernel memory accounting.
553 checkreqprot [SELINUX] Set initial checkreqprot flag value.
554 Format: { "0" | "1" }
555 See security/selinux/Kconfig help text.
556 0 -- check protection applied by kernel (includes
557 any implied execute protection).
558 1 -- check protection requested by application.
559 Default value is set via a kernel config option.
560 Value can be changed at runtime via
561 /sys/fs/selinux/checkreqprot.
562 Setting checkreqprot to 1 is deprecated.
565 See Documentation/s390/common_io.rst for details.
567 clearcpuid=X[,X...] [X86]
568 Disable CPUID feature X for the kernel. See
569 arch/x86/include/asm/cpufeatures.h for the valid bit
570 numbers X. Note the Linux-specific bits are not necessarily
571 stable over kernel options, but the vendor-specific
573 X can also be a string as appearing in the flags: line
574 in /proc/cpuinfo which does not have the above
575 instability issue. However, not all features have names
577 Note that using this option will taint your kernel.
578 Also note that user programs calling CPUID directly
579 or using the feature without checking anything
580 will still see it. This just prevents it from
581 being used by the kernel or shown in /proc/cpuinfo.
582 Also note the kernel might malfunction if you disable
587 Prevents the clock framework from automatically gating
588 clocks that have not been explicitly enabled by a Linux
589 device driver but are enabled in hardware at reset or
590 by the bootloader/firmware. Note that this does not
591 force such clocks to be always-on nor does it reserve
592 those clocks in any way. This parameter is useful for
593 debug and development, but should not be needed on a
594 platform with proper driver support. For more
595 information, see Documentation/driver-api/clk.rst.
597 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
599 Forces specified clocksource (if available) to be used
600 when calculating gettimeofday(). If specified
601 clocksource is not available, it defaults to PIT.
602 Format: { pit | tsc | cyclone | pmtmr }
604 clocksource= Override the default clocksource
606 Override the default clocksource and use the clocksource
607 with the name specified.
608 Some clocksource names to choose from, depending on
610 [all] jiffies (this is the base, fallback clocksource)
612 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
613 pxa_timer,timer3,32k_counter,timer0_1
614 [X86-32] pit,hpet,tsc;
615 scx200_hrt on Geode; cyclone on IBM x440
623 clocksource.arm_arch_timer.evtstrm=
626 Enable/disable the eventstream feature of the ARM
627 architected timer so that code using WFE-based polling
628 loops can be debugged more effectively on production
631 clocksource.max_cswd_read_retries= [KNL]
632 Number of clocksource_watchdog() retries due to
633 external delays before the clock will be marked
634 unstable. Defaults to two retries, that is,
635 three attempts to read the clock under test.
637 clocksource.verify_n_cpus= [KNL]
638 Limit the number of CPUs checked for clocksources
639 marked with CLOCK_SOURCE_VERIFY_PERCPU that
640 are marked unstable due to excessive skew.
641 A negative value says to check all CPUs, while
642 zero says not to check any. Values larger than
643 nr_cpu_ids are silently truncated to nr_cpu_ids.
644 The actual CPUs are chosen randomly, with
645 no replacement if the same CPU is chosen twice.
647 clocksource-wdtest.holdoff= [KNL]
648 Set the time in seconds that the clocksource
649 watchdog test waits before commencing its tests.
650 Defaults to zero when built as a module and to
651 10 seconds when built into the kernel.
653 cma=nn[MG]@[start[MG][-end[MG]]]
655 Sets the size of kernel global memory area for
656 contiguous memory allocations and optionally the
657 placement constraint by the physical address range of
658 memory allocations. A value of 0 disables CMA
659 altogether. For more information, see
660 kernel/dma/contiguous.c
664 Sets the size of kernel per-numa memory area for
665 contiguous memory allocations. A value of 0 disables
666 per-numa CMA altogether. And If this option is not
667 specificed, the default value is 0.
668 With per-numa CMA enabled, DMA users on node nid will
669 first try to allocate buffer from the pernuma area
670 which is located in node nid, if the allocation fails,
671 they will fallback to the global default memory area.
673 cmo_free_hint= [PPC] Format: { yes | no }
674 Specify whether pages are marked as being inactive
675 when they are freed. This is used in CMO environments
676 to determine OS memory pressure for page stealing by
680 coherent_pool=nn[KMG] [ARM,KNL]
681 Sets the size of memory pool for coherent, atomic dma
682 allocations, by default set to 256K.
684 com20020= [HW,NET] ARCnet - COM20020 chipset
686 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
688 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
692 ARCnet - COM90xx chipset (memory-mapped buffers)
693 Format: <io>[,<irq>[,<memstart>]]
695 condev= [HW,S390] console device
698 console= [KNL] Output console device and options.
700 tty<n> Use the virtual console device <n>.
704 Use the specified serial port. The options are of
705 the form "bbbbpnf", where "bbbb" is the baud rate,
706 "p" is parity ("n", "o", or "e"), "n" is number of
707 bits, and "f" is flow control ("r" for RTS or
708 omit it). Default is "9600n8".
710 See Documentation/admin-guide/serial-console.rst for more
712 Documentation/networking/netconsole.rst for an
715 uart[8250],io,<addr>[,options]
716 uart[8250],mmio,<addr>[,options]
717 uart[8250],mmio16,<addr>[,options]
718 uart[8250],mmio32,<addr>[,options]
719 uart[8250],0x<addr>[,options]
720 Start an early, polled-mode console on the 8250/16550
721 UART at the specified I/O port or MMIO address,
722 switching to the matching ttyS device later.
723 MMIO inter-register address stride is either 8-bit
724 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
725 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
726 to be equivalent to 'mmio'. 'options' are specified in
727 the same format described for ttyS above; if unspecified,
728 the h/w is not re-initialized.
730 hvc<n> Use the hypervisor console device <n>. This is for
731 both Xen and PowerPC hypervisors.
734 Use to disable console output, i.e., to have kernel
735 console messages discarded.
736 This must be the only console= parameter used on the
739 If the device connected to the port is not a TTY but a braille
740 device, prepend "brl," before the device type, for instance
742 For now, only VisioBraille is supported.
745 [KNL] Change console messages format
747 By default we print messages on consoles in
748 "[time stamp] text\n" format (time stamp may not be
749 printed, depending on CONFIG_PRINTK_TIME or
750 `printk_time' param).
752 Switch to syslog format: "<%u>[time stamp] text\n"
753 IOW, each message will have a facility and loglevel
754 prefix. The format is similar to one used by syslog()
755 syscall, or to executing "dmesg -S --raw" or to reading
758 consoleblank= [KNL] The console blank (screen saver) timeout in
759 seconds. A value of 0 disables the blank timer.
763 [KNL] Change the default value for
764 /proc/<pid>/coredump_filter.
765 See also Documentation/filesystems/proc.rst.
767 coresight_cpu_debug.enable
770 Enable/disable the CPU sampling based debugging.
771 0: default value, disable debugging
772 1: enable debugging at boot time
774 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
776 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
778 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
779 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
780 Some features depend on CPU0. Known dependencies are:
781 1. Resume from suspend/hibernate depends on CPU0.
782 Suspend/hibernate will fail if CPU0 is offline and you
783 need to online CPU0 before suspend/hibernate.
784 2. PIC interrupts also depend on CPU0. CPU0 can't be
785 removed if a PIC interrupt is detected.
786 It's said poweroff/reboot may depend on CPU0 on some
787 machines although I haven't seen such issues so far
788 after CPU0 is offline on a few tested machines.
789 If the dependencies are under your control, you can
790 turn on cpu0_hotplug.
792 cpuidle.off=1 [CPU_IDLE]
793 disable the cpuidle sub-system
796 [CPU_IDLE] Name of the cpuidle governor to use.
798 cpufreq.off=1 [CPU_FREQ]
799 disable the cpufreq sub-system
801 cpufreq.default_governor=
802 [CPU_FREQ] Name of the default cpufreq governor or
803 policy to use. This governor must be registered in the
804 kernel before the cpufreq driver probes.
807 [X86] Delay for N microsec between assert and de-assert
808 of APIC INIT to start processors. This delay occurs
809 on every CPU online, such as boot, and resume from suspend.
812 crash_kexec_post_notifiers
813 Run kdump after running panic-notifiers and dumping
814 kmsg. This only for the users who doubt kdump always
815 succeeds in any situation.
816 Note that this also increases risks of kdump failure,
817 because some panic notifiers can make the crashed
818 kernel more unstable.
820 crashkernel=size[KMG][@offset[KMG]]
821 [KNL] Using kexec, Linux can switch to a 'crash kernel'
822 upon panic. This parameter reserves the physical
823 memory region [offset, offset + size] for that kernel
824 image. If '@offset' is omitted, then a suitable offset
825 is selected automatically.
826 [KNL, X86-64] Select a region under 4G first, and
827 fall back to reserve region above 4G when '@offset'
828 hasn't been specified.
829 See Documentation/admin-guide/kdump/kdump.rst for further details.
831 crashkernel=range1:size1[,range2:size2,...][@offset]
832 [KNL] Same as above, but depends on the memory
833 in the running system. The syntax of range is
834 start-[end] where start and end are both
835 a memory unit (amount[KMG]). See also
836 Documentation/admin-guide/kdump/kdump.rst for an example.
838 crashkernel=size[KMG],high
839 [KNL, X86-64, ARM64] range could be above 4G. Allow kernel
840 to allocate physical memory region from top, so could
841 be above 4G if system have more than 4G ram installed.
842 Otherwise memory region will be allocated below 4G, if
844 It will be ignored if crashkernel=X is specified.
845 crashkernel=size[KMG],low
846 [KNL, X86-64] range under 4G. When crashkernel=X,high
847 is passed, kernel could allocate physical memory region
848 above 4G, that cause second kernel crash on system
849 that require some amount of low memory, e.g. swiotlb
850 requires at least 64M+32K low memory, also enough extra
851 low memory is needed to make sure DMA buffers for 32-bit
852 devices won't run out. Kernel would try to allocate
853 at least 256M below 4G automatically.
854 This one lets the user specify own low range under 4G
855 for second kernel instead.
856 0: to disable low allocation.
857 It will be ignored when crashkernel=X,high is not used
858 or memory reserved is below 4G.
860 [KNL, ARM64] range in low memory.
861 This one lets the user specify a low range in the
862 DMA zone for the crash dump kernel.
863 It will be ignored when crashkernel=X,high is not used
864 or memory reserved is located in the DMA zones.
867 [KNL] Disable crypto self-tests
872 cs89x0_media= [HW,NET]
873 Format: { rj45 | aui | bnc }
875 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
876 handling. When switched on, additional debug data is
877 printed to the console in case a hanging CPU is
878 detected, and that CPU is pinged again in order to try
879 to resolve the hang situation.
880 0: disable csdlock debugging (default)
881 1: enable basic csdlock debugging (minor impact)
882 ext: enable extended csdlock debugging (more impact,
886 See header of drivers/s390/block/dasd_devmap.c.
888 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
889 (one device per port)
890 Format: <port#>,<type>
891 See also Documentation/input/devices/joystick-parport.rst
893 debug [KNL] Enable kernel debugging (events log level).
896 [KNL] Enable printing [hashed] pointers early in the
897 boot sequence. If enabled, we use a weak hash instead
898 of siphash to hash pointers. Use this option if you are
899 seeing instances of '(___ptrval___)') and need to see a
900 value (hashed pointer) instead. Cryptographically
901 insecure, please do not use on production kernels.
904 [KNL] verbose locking self-tests
906 Print debugging info while doing the locking API
908 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
909 (no extra messages), setting it to -1 (all bits set)
910 will print _a_lot_ more information - normally only
911 useful to lockdep developers.
913 debug_objects [KNL] Enable object debugging
916 [KNL] Disable object debugging
918 debug_guardpage_minorder=
919 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
920 parameter allows control of the order of pages that will
921 be intentionally kept free (and hence protected) by the
922 buddy allocator. Bigger value increase the probability
923 of catching random memory corruption, but reduce the
924 amount of memory for normal system use. The maximum
925 possible value is MAX_ORDER/2. Setting this parameter
926 to 1 or 2 should be enough to identify most random
927 memory corruption problems caused by bugs in kernel or
928 driver code when a CPU writes to (or reads from) a
929 random memory location. Note that there exists a class
930 of memory corruptions problems caused by buggy H/W or
931 F/W or by drivers badly programing DMA (basically when
932 memory is written at bus level and the CPU MMU is
933 bypassed) which are not detectable by
934 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
935 tracking down these problems.
938 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
939 enables the feature at boot time. By default, it is
940 disabled and the system will work mostly the same as a
941 kernel built without CONFIG_DEBUG_PAGEALLOC.
942 Note: to get most of debug_pagealloc error reports, it's
943 useful to also enable the page_owner functionality.
944 on: enable the feature
946 debugfs= [KNL] This parameter enables what is exposed to userspace
947 and debugfs internal clients.
948 Format: { on, no-mount, off }
949 on: All functions are enabled.
951 Filesystem is not registered but kernel clients can
952 access APIs and a crashkernel can be used to read
953 its content. There is nothing to mount.
954 off: Filesystem is not registered and clients
955 get a -EPERM as result when trying to register files
956 or directories within debugfs.
957 This is equivalent of the runtime functionality if
958 debugfs was not enabled in the kernel at all.
959 Default value is set in build-time with a kernel configuration.
961 debugpat [X86] Enable PAT debugging
963 decnet.addr= [HW,NET]
964 Format: <area>[,<node>]
965 See also Documentation/networking/decnet.rst.
968 [HW] The size of the default HugeTLB page. This is
969 the size represented by the legacy /proc/ hugepages
970 APIs. In addition, this is the default hugetlb size
971 used for shmget(), mmap() and mounting hugetlbfs
972 filesystems. If not specified, defaults to the
973 architecture's default huge page size. Huge page
974 sizes are architecture dependent. See also
975 Documentation/admin-guide/mm/hugetlbpage.rst.
978 deferred_probe_timeout=
979 [KNL] Debugging option to set a timeout in seconds for
980 deferred probe to give up waiting on dependencies to
981 probe. Only specific dependencies (subsystems or
982 drivers) that have opted in will be ignored. A timeout
983 of 0 will timeout at the end of initcalls. If the time
984 out hasn't expired, it'll be restarted by each
985 successful driver registration. This option will also
986 dump out devices still on the deferred probe list after
989 delayacct [KNL] Enable per-task delay accounting
991 dell_smm_hwmon.ignore_dmi=
992 [HW] Continue probing hardware even if DMI data
993 indicates that the driver is running on unsupported
996 dell_smm_hwmon.force=
997 [HW] Activate driver even if SMM BIOS signature does
998 not match list of supported models and enable otherwise
999 blacklisted features.
1001 dell_smm_hwmon.power_status=
1002 [HW] Report power status in /proc/i8k
1003 (disabled by default).
1005 dell_smm_hwmon.restricted=
1006 [HW] Allow controlling fans only if SYS_ADMIN
1009 dell_smm_hwmon.fan_mult=
1010 [HW] Factor to multiply fan speed with.
1012 dell_smm_hwmon.fan_max=
1013 [HW] Maximum configurable fan speed.
1016 Format: { on | off | def_only | inf_only | always }
1017 on: s390 zlib hardware support for compression on
1018 level 1 and decompression (default)
1019 off: No s390 zlib hardware support
1020 def_only: s390 zlib hardware support for deflate
1021 only (compression on level 1)
1022 inf_only: s390 zlib hardware support for inflate
1023 only (decompression)
1024 always: Same as 'on' but ignores the selected compression
1025 level always using hardware support (used for debugging)
1027 dhash_entries= [KNL]
1028 Set number of hash buckets for dentry cache.
1030 disable_1tb_segments [PPC]
1031 Disables the use of 1TB hash page table segments. This
1032 causes the kernel to fall back to 256MB segments which
1033 can be useful when debugging issues that require an SLB
1037 Limits the number of kernel SLB entries, and flushes
1038 them frequently to increase the rate of SLB faults
1039 on kernel addresses.
1042 See Documentation/networking/ipv6.rst.
1045 Disable RADIX MMU mode on POWER9
1047 radix_hcall_invalidate=on [PPC/PSERIES]
1048 Disable RADIX GTSE feature and use hcall for TLB
1052 Disable TLBIE instruction. Currently does not work
1053 with KVM, with HASH MMU, or with coherent accelerators.
1055 disable_cpu_apicid= [X86,APIC,SMP]
1057 The number of initial APIC ID for the
1058 corresponding CPU to be disabled at boot,
1059 mostly used for the kdump 2nd kernel to
1060 disable BSP to wake up multiple CPUs without
1061 causing system reset or hang due to sending
1062 INIT from AP to BSP.
1064 disable_ddw [PPC/PSERIES]
1065 Disable Dynamic DMA Window support. Use this
1066 to workaround buggy firmware.
1068 disable_ipv6= [IPV6]
1069 See Documentation/networking/ipv6.rst.
1071 disable_mtrr_cleanup [X86]
1072 The kernel tries to adjust MTRR layout from continuous
1073 to discrete, to make X server driver able to add WB
1074 entry later. This parameter disables that.
1076 disable_mtrr_trim [X86, Intel and AMD only]
1077 By default the kernel will trim any uncacheable
1078 memory out of your available memory pool based on
1079 MTRR settings. This parameter disables that behavior,
1080 possibly causing your machine to run very slowly.
1082 disable_timer_pin_1 [X86]
1083 Disable PIN 1 of APIC timer
1084 Can be useful to work around chipset bugs.
1086 dis_ucode_ldr [X86] Disable the microcode loader.
1088 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1089 this option disables the debugging code at boot.
1091 dma_debug_entries=<number>
1092 This option allows to tune the number of preallocated
1093 entries for DMA-API debugging code. One entry is
1094 required per DMA-API allocation. Use this if the
1095 DMA-API debugging code disables itself because the
1096 architectural default is too low.
1098 dma_debug_driver=<driver_name>
1099 With this option the DMA-API debugging driver
1100 filter feature can be enabled at boot time. Just
1101 pass the driver to filter for as the parameter.
1102 The filter can be disabled or changed to another
1103 driver later using sysfs.
1105 driver_async_probe= [KNL]
1106 List of driver names to be probed asynchronously. *
1107 matches with all driver names. If * is specified, the
1108 rest of the listed driver names are those that will NOT
1110 Format: <driver_name1>,<driver_name2>...
1112 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1113 Broken monitors, graphic adapters, KVMs and EDIDless
1114 panels may send no or incorrect EDID data sets.
1115 This parameter allows to specify an EDID data sets
1116 in the /lib/firmware directory that are used instead.
1117 Generic built-in EDID data sets are used, if one of
1118 edid/1024x768.bin, edid/1280x1024.bin,
1119 edid/1680x1050.bin, or edid/1920x1080.bin is given
1120 and no file with the same name exists. Details and
1121 instructions how to build your own EDID data are
1122 available in Documentation/admin-guide/edid.rst. An EDID
1123 data set will only be used for a particular connector,
1124 if its name and a colon are prepended to the EDID
1125 name. Each connector may use a unique EDID data
1126 set by separating the files with a comma. An EDID
1127 data set with no connector name will be used for
1128 any connectors not explicitly specified.
1133 Format: {"off" | "known"}
1134 Control how the dt_cpu_ftrs device-tree binding is
1135 used for CPU feature discovery and setup (if it
1137 off: Do not use it, fall back to legacy cpu table.
1138 known: Do not pass through unknown features to guests
1139 or userspace, only those that the kernel is aware of.
1141 dump_apple_properties [X86]
1142 Dump name and content of EFI device properties on
1143 x86 Macs. Useful for driver authors to determine
1144 what data is available or for reverse-engineering.
1146 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1147 <module>.dyndbg[="val"]
1148 Enable debug messages at boot time. See
1149 Documentation/admin-guide/dynamic-debug-howto.rst
1152 nopku [X86] Disable Memory Protection Keys CPU feature found
1155 <module>.async_probe [KNL]
1156 Enable asynchronous probe on this module.
1158 early_ioremap_debug [KNL]
1159 Enable debug messages in early_ioremap support. This
1160 is useful for tracking down temporary early mappings
1161 which are not unmapped.
1163 earlycon= [KNL] Output early console device and options.
1165 When used with no options, the early console is
1166 determined by stdout-path property in device tree's
1167 chosen node or the ACPI SPCR table if supported by
1170 cdns,<addr>[,options]
1171 Start an early, polled-mode console on a Cadence
1172 (xuartps) serial port at the specified address. Only
1173 supported option is baud rate. If baud rate is not
1174 specified, the serial port must already be setup and
1177 uart[8250],io,<addr>[,options]
1178 uart[8250],mmio,<addr>[,options]
1179 uart[8250],mmio32,<addr>[,options]
1180 uart[8250],mmio32be,<addr>[,options]
1181 uart[8250],0x<addr>[,options]
1182 Start an early, polled-mode console on the 8250/16550
1183 UART at the specified I/O port or MMIO address.
1184 MMIO inter-register address stride is either 8-bit
1185 (mmio) or 32-bit (mmio32 or mmio32be).
1186 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1187 to be equivalent to 'mmio'. 'options' are specified
1188 in the same format described for "console=ttyS<n>"; if
1189 unspecified, the h/w is not initialized.
1193 Start an early, polled-mode console on a pl011 serial
1194 port at the specified address. The pl011 serial port
1195 must already be setup and configured. Options are not
1196 yet supported. If 'mmio32' is specified, then only
1197 the driver will use only 32-bit accessors to read/write
1198 the device registers.
1201 Start an early console on a litex serial port at the
1202 specified address. The serial port must already be
1203 setup and configured. Options are not yet supported.
1206 Start an early, polled-mode console on a meson serial
1207 port at the specified address. The serial port must
1208 already be setup and configured. Options are not yet
1212 Start an early, polled-mode console on an msm serial
1213 port at the specified address. The serial port
1214 must already be setup and configured. Options are not
1217 msm_serial_dm,<addr>
1218 Start an early, polled-mode console on an msm serial
1219 dm port at the specified address. The serial port
1220 must already be setup and configured. Options are not
1224 Start an early, polled-mode console on a serial port
1225 of an Actions Semi SoC, such as S500 or S900, at the
1226 specified address. The serial port must already be
1227 setup and configured. Options are not yet supported.
1230 Start an early, polled-mode console on a serial port
1231 of an RDA Micro SoC, such as RDA8810PL, at the
1232 specified address. The serial port must already be
1233 setup and configured. Options are not yet supported.
1236 Use RISC-V SBI (Supervisor Binary Interface) for early
1239 smh Use ARM semihosting calls for early console.
1247 Use early console provided by serial driver available
1248 on Samsung SoCs, requires selecting proper type and
1249 a correct base address of the selected UART port. The
1250 serial port must already be setup and configured.
1251 Options are not yet supported.
1254 Start an early, polled-mode console on a lantiq serial
1255 (lqasc) port at the specified address. The serial port
1256 must already be setup and configured. Options are not
1261 Use early console provided by Freescale LP UART driver
1262 found on Freescale Vybrid and QorIQ LS1021A processors.
1263 A valid base address must be provided, and the serial
1264 port must already be setup and configured.
1268 Start an early, polled-mode, output-only console on the
1269 Freescale i.MX UART at the specified address. The UART
1270 must already be setup and configured.
1273 Start an early, polled-mode console on the
1274 Armada 3700 serial port at the specified
1275 address. The serial port must already be setup
1276 and configured. Options are not yet supported.
1279 Start an early, polled-mode console on a Qualcomm
1280 Generic Interface (GENI) based serial port at the
1281 specified address. The serial port must already be
1282 setup and configured. Options are not yet supported.
1285 Start an early, unaccelerated console on the EFI
1286 memory mapped framebuffer (if available). On cache
1287 coherent non-x86 systems that use system memory for
1288 the framebuffer, pass the 'ram' option so that it is
1289 mapped with the correct attributes.
1292 Use early console provided by Freescale LINFlexD UART
1293 serial driver for NXP S32V234 SoCs. A valid base
1294 address must be provided, and the serial port must
1295 already be setup and configured.
1297 earlyprintk= [X86,SH,ARM,M68k,S390]
1301 earlyprintk=serial[,ttySn[,baudrate]]
1302 earlyprintk=serial[,0x...[,baudrate]]
1303 earlyprintk=ttySn[,baudrate]
1304 earlyprintk=dbgp[debugController#]
1305 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1306 earlyprintk=xdbc[xhciController#]
1308 earlyprintk is useful when the kernel crashes before
1309 the normal console is initialized. It is not enabled by
1310 default because it has some cosmetic problems.
1312 Append ",keep" to not disable it when the real console
1315 Only one of vga, serial, or usb debug port can
1318 Currently only ttyS0 and ttyS1 may be specified by
1319 name. Other I/O ports may be explicitly specified
1320 on some architectures (x86 and arm at least) by
1321 replacing ttySn with an I/O port address, like this:
1322 earlyprintk=serial,0x1008,115200
1323 You can find the port for a given device in
1324 /proc/tty/driver/serial:
1325 2: uart:ST16650V2 port:00001008 irq:18 ...
1327 Interaction with the standard serial driver is not
1330 The VGA output is eventually overwritten by
1333 The xen option can only be used in Xen domains.
1335 The sclp output can only be used on s390.
1337 The optional "force" to "pciserial" enables use of a
1338 PCI device even when its classcode is not of the
1341 edac_report= [HW,EDAC] Control how to report EDAC event
1342 Format: {"on" | "off" | "force"}
1343 on: enable EDAC to report H/W event. May be overridden
1344 by other higher priority error reporting module.
1345 off: disable H/W event reporting through EDAC.
1346 force: enforce the use of EDAC to report H/W event.
1350 Format: {"off" | "on" | "skip[mbr]"}
1353 Format: { "debug", "disable_early_pci_dma",
1354 "nochunk", "noruntime", "nosoftreserve",
1355 "novamap", "no_disable_early_pci_dma" }
1356 debug: enable misc debug output.
1357 disable_early_pci_dma: disable the busmaster bit on all
1358 PCI bridges while in the EFI boot stub.
1359 nochunk: disable reading files in "chunks" in the EFI
1360 boot stub, as chunking can cause problems with some
1361 firmware implementations.
1362 noruntime : disable EFI runtime services support
1363 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1364 attribute may cause the kernel to reserve the
1365 memory range for a memory mapping driver to
1366 claim. Specify efi=nosoftreserve to disable this
1367 reservation and treat the memory by its base type
1368 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1369 novamap: do not call SetVirtualAddressMap().
1370 no_disable_early_pci_dma: Leave the busmaster bit set
1371 on all PCI bridges while in the EFI boot stub
1373 efi_no_storage_paranoia [EFI; X86]
1374 Using this parameter you can use more than 50% of
1375 your efi variable storage. Use this parameter only if
1376 you are really sure that your UEFI does sane gc and
1377 fulfills the spec otherwise your board may brick.
1379 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1380 Add arbitrary attribute to specific memory range by
1381 updating original EFI memory map.
1382 Region of memory which aa attribute is added to is
1385 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1386 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1387 attribute is added to range 0x100000000-0x180000000 and
1388 0x10a0000000-0x1120000000.
1390 If efi_fake_mem=8G@9G:0x40000 is specified, the
1391 EFI_MEMORY_SP(0x40000) attribute is added to
1392 range 0x240000000-0x43fffffff.
1394 Using this parameter you can do debugging of EFI memmap
1395 related features. For example, you can do debugging of
1396 Address Range Mirroring feature even if your box
1397 doesn't support it, or mark specific memory as
1400 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1401 that is to be dynamically loaded by Linux. If there are
1402 multiple variables with the same name but with different
1403 vendor GUIDs, all of them will be loaded. See
1404 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1407 eisa_irq_edge= [PARISC,HW]
1408 See header of drivers/parisc/eisa.c.
1410 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1413 This is designed to be used in conjunction with
1414 the boot argument: earlyprintk=vga
1416 This parameter works in place of the kgdboc parameter
1417 but can only be used if the backing tty is available
1418 very early in the boot process. For early debugging
1419 via a serial port see kgdboc_earlycon instead.
1422 See comment before function elanfreq_setup() in
1423 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1425 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1426 Specifies physical address of start of kernel core
1427 image elf header and optionally the size. Generally
1428 kexec loader will pass this option to capture kernel.
1429 See Documentation/admin-guide/kdump/kdump.rst for details.
1431 enable_mtrr_cleanup [X86]
1432 The kernel tries to adjust MTRR layout from continuous
1433 to discrete, to make X server driver able to add WB
1434 entry later. This parameter enables that.
1436 enable_timer_pin_1 [X86]
1437 Enable PIN 1 of APIC timer
1438 Can be useful to work around chipset bugs
1439 (in particular on some ATI chipsets).
1440 The kernel tries to set a reasonable default.
1442 enforcing [SELINUX] Set initial enforcing status.
1444 See security/selinux/Kconfig help text.
1445 0 -- permissive (log only, no denials).
1446 1 -- enforcing (deny and log).
1448 Value can be changed at runtime via
1449 /sys/fs/selinux/enforce.
1452 Disable Error Record Serialization Table (ERST)
1455 ether= [HW,NET] Ethernet cards parameters
1456 This option is obsoleted by the "netdev=" option, which
1457 has equivalent usage. See its documentation for details.
1461 Permit 'security.evm' to be updated regardless of
1462 current integrity status.
1467 fail_make_request=[KNL]
1468 General fault injection mechanism.
1469 Format: <interval>,<probability>,<space>,<times>
1470 See also Documentation/fault-injection/.
1473 Format: { initns | none }
1474 See Documentation/admin-guide/sysctl/net.rst for
1475 fb_tunnels_only_for_init_ns
1478 See Documentation/admin-guide/blockdev/floppy.rst.
1480 force_pal_cache_flush
1481 [IA-64] Avoid check_sal_cache_flush which may hang on
1482 buggy SAL_CACHE_FLUSH implementations. Using this
1483 parameter will force ia64_sal_cache_flush to call
1484 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1487 Forcefully enable Physical Address Extension (PAE).
1488 Many Pentium M systems disable PAE but may have a
1489 functionally usable PAE implementation.
1490 Warning: use of this parameter will taint the kernel
1491 and may cause unknown problems.
1494 [FTRACE] will set and start the specified tracer
1495 as early as possible in order to facilitate early
1498 ftrace_boot_snapshot
1499 [FTRACE] On boot up, a snapshot will be taken of the
1500 ftrace ring buffer that can be read at:
1501 /sys/kernel/tracing/snapshot.
1502 This is useful if you need tracing information from kernel
1503 boot up that is likely to be overridden by user space
1504 start up functionality.
1506 ftrace_dump_on_oops[=orig_cpu]
1507 [FTRACE] will dump the trace buffers on oops.
1508 If no parameter is passed, ftrace will dump
1509 buffers of all CPUs, but if you pass orig_cpu, it will
1510 dump only the buffer of the CPU that triggered the
1513 ftrace_filter=[function-list]
1514 [FTRACE] Limit the functions traced by the function
1515 tracer at boot up. function-list is a comma-separated
1516 list of functions. This list can be changed at run
1517 time by the set_ftrace_filter file in the debugfs
1520 ftrace_notrace=[function-list]
1521 [FTRACE] Do not trace the functions specified in
1522 function-list. This list can be changed at run time
1523 by the set_ftrace_notrace file in the debugfs
1526 ftrace_graph_filter=[function-list]
1527 [FTRACE] Limit the top level callers functions traced
1528 by the function graph tracer at boot up.
1529 function-list is a comma-separated list of functions
1530 that can be changed at run time by the
1531 set_graph_function file in the debugfs tracing directory.
1533 ftrace_graph_notrace=[function-list]
1534 [FTRACE] Do not trace from the functions specified in
1535 function-list. This list is a comma-separated list of
1536 functions that can be changed at run time by the
1537 set_graph_notrace file in the debugfs tracing directory.
1539 ftrace_graph_max_depth=<uint>
1540 [FTRACE] Used with the function graph tracer. This is
1541 the max depth it will trace into a function. This value
1542 can be changed at run time by the max_graph_depth file
1543 in the tracefs tracing directory. default: 0 (no limit)
1545 fw_devlink= [KNL] Create device links between consumer and supplier
1546 devices by scanning the firmware to infer the
1547 consumer/supplier relationships. This feature is
1548 especially useful when drivers are loaded as modules as
1549 it ensures proper ordering of tasks like device probing
1550 (suppliers first, then consumers), supplier boot state
1551 clean up (only after all consumers have probed),
1552 suspend/resume & runtime PM (consumers first, then
1554 Format: { off | permissive | on | rpm }
1555 off -- Don't create device links from firmware info.
1556 permissive -- Create device links from firmware info
1557 but use it only for ordering boot state clean
1558 up (sync_state() calls).
1559 on -- Create device links from firmware info and use it
1560 to enforce probe and suspend/resume ordering.
1561 rpm -- Like "on", but also use to order runtime PM.
1563 fw_devlink.strict=<bool>
1564 [KNL] Treat all inferred dependencies as mandatory
1565 dependencies. This only applies for fw_devlink=on|rpm.
1569 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1570 support via parallel port (up to 5 devices per port)
1571 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1572 See also Documentation/input/devices/joystick-parport.rst
1576 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1580 gcov_persist= [GCOV] When non-zero (default), profiling data for
1581 kernel modules is saved and remains accessible via
1582 debugfs, even when the module is unloaded/reloaded.
1583 When zero, profiling data is discarded and associated
1584 debugfs files are removed at module unload time.
1586 goldfish [X86] Enable the goldfish android emulator platform.
1587 Don't use this when you are not running on the
1590 gpio-mockup.gpio_mockup_ranges
1591 [HW] Sets the ranges of gpiochip of for this device.
1592 Format: <start1>,<end1>,<start2>,<end2>...
1593 gpio-mockup.gpio_mockup_named_lines
1594 [HW] Let the driver know GPIO lines should be named.
1596 gpt [EFI] Forces disk with valid GPT signature but
1597 invalid Protective MBR to be treated as GPT. If the
1598 primary GPT is corrupted, it enables the backup/alternate
1599 GPT to be used instead.
1601 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1602 the "Enable 0" bit of the configuration register.
1605 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1606 the "Enable 0" bit of the configuration register.
1609 grcan.select= [HW] Select which physical interface to use.
1612 grcan.txsize= [HW] Sets the size of the tx buffer.
1613 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1615 grcan.rxsize= [HW] Sets the size of the rx buffer.
1616 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1620 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1621 hardening is enabled for this boot. Hardened
1622 usercopy checking is used to protect the kernel
1623 from reading or writing beyond known memory
1624 allocation boundaries as a proactive defense
1625 against bounds-checking flaws in the kernel's
1626 copy_to_user()/copy_from_user() interface.
1627 on Perform hardened usercopy checks (default).
1628 off Disable hardened usercopy checks.
1630 hardlockup_all_cpu_backtrace=
1631 [KNL] Should the hard-lockup detector generate
1632 backtraces on all cpus.
1635 hashdist= [KNL,NUMA] Large hashes allocated during boot
1636 are distributed across NUMA nodes. Defaults on
1637 for 64-bit NUMA, off otherwise.
1638 Format: 0 | 1 (for off | on)
1640 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1642 hd= [EIDE] (E)IDE hard drive subsystem geometry
1643 Format: <cyl>,<head>,<sect>
1646 Disable Hardware Error Source Table (HEST) support;
1647 corresponding firmware-first mode error processing
1648 logic will be disabled.
1650 hibernate= [HIBERNATION]
1651 noresume Don't check if there's a hibernation image
1652 present during boot.
1653 nocompress Don't compress/decompress hibernation images.
1654 no Disable hibernation and resume.
1655 protect_image Turn on image protection during restoration
1656 (that will set all pages holding image data
1657 during restoration read-only).
1659 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1660 size of <nn>. This works even on boxes that have no
1661 highmem otherwise. This also works to reduce highmem
1662 size on bigger boxes.
1664 highres= [KNL] Enable/disable high resolution timer mode.
1665 Valid parameters: "on", "off"
1670 hpet= [X86-32,HPET] option to control HPET usage
1671 Format: { enable (default) | disable | force |
1673 disable: disable HPET and use PIT instead
1674 force: allow force enabled of undocumented chips (ICH4,
1676 verbose: show contents of HPET registers during setup
1678 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1679 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1681 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1682 If this follows hugepagesz (below), it specifies
1683 the number of pages of hugepagesz to be allocated.
1684 If this is the first HugeTLB parameter on the command
1685 line, it specifies the number of pages to allocate for
1686 the default huge page size. If using node format, the
1687 number of pages to allocate per-node can be specified.
1688 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1689 Format: <integer> or (node format)
1690 <node>:<integer>[,<node>:<integer>]
1693 [HW] The size of the HugeTLB pages. This is used in
1694 conjunction with hugepages (above) to allocate huge
1695 pages of a specific size at boot. The pair
1696 hugepagesz=X hugepages=Y can be specified once for
1697 each supported huge page size. Huge page sizes are
1698 architecture dependent. See also
1699 Documentation/admin-guide/mm/hugetlbpage.rst.
1702 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1703 of gigantic hugepages. Or using node format, the size
1704 of a CMA area per node can be specified.
1705 Format: nn[KMGTPE] or (node format)
1706 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1708 Reserve a CMA area of given size and allocate gigantic
1709 hugepages using the CMA allocator. If enabled, the
1710 boot-time allocation of gigantic hugepages is skipped.
1712 hugetlb_free_vmemmap=
1713 [KNL] Reguires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1715 Allows heavy hugetlb users to free up some more
1716 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1717 Format: { [oO][Nn]/Y/y/1 | [oO][Ff]/N/n/0 (default) }
1719 [oO][Nn]/Y/y/1: enable the feature
1720 [oO][Ff]/N/n/0: disable the feature
1722 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1725 This is not compatible with memory_hotplug.memmap_on_memory.
1726 If both parameters are enabled, hugetlb_free_vmemmap takes
1727 precedence over memory_hotplug.memmap_on_memory.
1730 [KNL] Should the hung task detector generate panics.
1733 A value of 1 instructs the kernel to panic when a
1734 hung task is detected. The default value is controlled
1735 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1736 option. The value selected by this boot parameter can
1737 be changed later by the kernel.hung_task_panic sysctl.
1739 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1740 terminal devices. Valid values: 0..8
1741 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1742 If specified, z/VM IUCV HVC accepts connections
1743 from listed z/VM user IDs only.
1745 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1746 which allow the hypervisor to 'idle' the
1747 guest on lock contention.
1750 Do not unregister boot console at start. This is only
1751 useful for debugging when something happens in the window
1752 between unregistering the boot console and initializing
1755 i2c_bus= [HW] Override the default board specific I2C bus speed
1756 or register an additional I2C bus that is not
1757 registered from board initialization code.
1761 i8042.debug [HW] Toggle i8042 debug mode
1762 i8042.unmask_kbd_data
1763 [HW] Enable printing of interrupt data from the KBD port
1764 (disabled by default, and as a pre-condition
1765 requires that i8042.debug=1 be enabled)
1766 i8042.direct [HW] Put keyboard port into non-translated mode
1767 i8042.dumbkbd [HW] Pretend that controller can only read data from
1768 keyboard and cannot control its state
1769 (Don't attempt to blink the leds)
1770 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1771 i8042.nokbd [HW] Don't check/create keyboard port
1772 i8042.noloop [HW] Disable the AUX Loopback command while probing
1774 i8042.nomux [HW] Don't check presence of an active multiplexing
1776 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1778 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1779 i8042.reset [HW] Reset the controller during init, cleanup and
1780 suspend-to-ram transitions, only during s2r
1781 transitions, or never reset
1782 Format: { 1 | Y | y | 0 | N | n }
1783 1, Y, y: always reset controller
1784 0, N, n: don't ever reset controller
1785 Default: only on s2r transitions on x86; most other
1786 architectures force reset to be always executed
1787 i8042.unlock [HW] Unlock (ignore) the keylock
1788 i8042.kbdreset [HW] Reset device connected to KBD port
1790 [HW] Allow deferred probing upon i8042 probe errors
1794 i915.invert_brightness=
1795 [DRM] Invert the sense of the variable that is used to
1796 set the brightness of the panel backlight. Normally a
1797 brightness value of 0 indicates backlight switched off,
1798 and the maximum of the brightness value sets the backlight
1799 to maximum brightness. If this parameter is set to 0
1800 (default) and the machine requires it, or this parameter
1801 is set to 1, a brightness value of 0 sets the backlight
1802 to maximum brightness, and the maximum of the brightness
1803 value switches the backlight off.
1804 -1 -- never invert brightness
1805 0 -- machine default
1806 1 -- force brightness inversion
1809 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1813 Format: idle=poll, idle=halt, idle=nomwait
1814 Poll forces a polling idle loop that can slightly
1815 improve the performance of waking up a idle CPU, but
1816 will use a lot of power and make the system run hot.
1818 idle=halt: Halt is forced to be used for CPU idle.
1819 In such case C2/C3 won't be used again.
1820 idle=nomwait: Disable mwait for CPU C-states
1824 Allow force disabling of Shared Virtual Memory (SVA)
1825 support for the idxd driver. By default it is set to
1828 idxd.tc_override= [HW]
1830 Allow override of default traffic class configuration
1831 for the device. By default it is set to false (0).
1833 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1834 Format: { strict | legacy | 2008 | relaxed }
1837 Choose which programs will be accepted for execution
1838 based on the IEEE 754 NaN encoding(s) supported by
1839 the FPU and the NaN encoding requested with the value
1840 of an ELF file header flag individually set by each
1841 binary. Hardware implementations are permitted to
1842 support either or both of the legacy and the 2008 NaN
1845 Available settings are as follows:
1846 strict accept binaries that request a NaN encoding
1847 supported by the FPU
1848 legacy only accept legacy-NaN binaries, if supported
1850 2008 only accept 2008-NaN binaries, if supported
1852 relaxed accept any binaries regardless of whether
1853 supported by the FPU
1855 The FPU emulator is always able to support both NaN
1856 encodings, so if no FPU hardware is present or it has
1857 been disabled with 'nofpu', then the settings of
1858 'legacy' and '2008' strap the emulator accordingly,
1859 'relaxed' straps the emulator for both legacy-NaN and
1860 2008-NaN, whereas 'strict' enables legacy-NaN only on
1861 legacy processors and both NaN encodings on MIPS32 or
1864 The setting for ABS.fmt/NEG.fmt instruction execution
1865 mode generally follows that for the NaN encoding,
1866 except where unsupported by hardware.
1868 ignore_loglevel [KNL]
1869 Ignore loglevel setting - this will print /all/
1870 kernel messages to the console. Useful for debugging.
1871 We also add it as printk module parameter, so users
1872 could change it dynamically, usually by
1873 /sys/module/printk/parameters/ignore_loglevel.
1876 Ignore RLIMIT_DATA setting for data mappings,
1877 print warning at first misuse. Can be changed via
1878 /sys/module/kernel/parameters/ignore_rlimit_data.
1880 ihash_entries= [KNL]
1881 Set number of hash buckets for inode cache.
1883 ima_appraise= [IMA] appraise integrity measurements
1884 Format: { "off" | "enforce" | "fix" | "log" }
1887 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1888 The builtin appraise policy appraises all files
1891 ima_canonical_fmt [IMA]
1892 Use the canonical format for the binary runtime
1893 measurements, instead of host native format.
1896 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1900 The list of supported hash algorithms is defined
1901 in crypto/hash_info.h.
1904 The builtin policies to load during IMA setup.
1905 Format: "tcb | appraise_tcb | secure_boot |
1906 fail_securely | critical_data"
1908 The "tcb" policy measures all programs exec'd, files
1909 mmap'd for exec, and all files opened with the read
1910 mode bit set by either the effective uid (euid=0) or
1913 The "appraise_tcb" policy appraises the integrity of
1914 all files owned by root.
1916 The "secure_boot" policy appraises the integrity
1917 of files (eg. kexec kernel image, kernel modules,
1918 firmware, policy, etc) based on file signatures.
1920 The "fail_securely" policy forces file signature
1921 verification failure also on privileged mounted
1922 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1925 The "critical_data" policy measures kernel integrity
1928 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1929 Load a policy which meets the needs of the Trusted
1930 Computing Base. This means IMA will measure all
1931 programs exec'd, files mmap'd for exec, and all files
1932 opened for read by uid=0.
1935 Select one of defined IMA measurements template formats.
1936 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
1941 [IMA] Define a custom template format.
1942 Format: { "field1|...|fieldN" }
1944 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1945 Format: <min_file_size>
1946 Set the minimal file size for using asynchronous hash.
1947 If left unspecified, ahash usage is disabled.
1949 ahash performance varies for different data sizes on
1950 different crypto accelerators. This option can be used
1951 to achieve the best performance for a particular HW.
1953 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1955 Set hashing buffer size. Default: 4k.
1957 ahash performance varies for different chunk sizes on
1958 different crypto accelerators. This option can be used
1959 to achieve best performance for particular HW.
1963 Run specified binary instead of /sbin/init as init
1966 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1967 for working out where the kernel is dying during
1970 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1971 initcall functions. Useful for debugging built-in
1972 modules and initcalls.
1974 initramfs_async= [KNL]
1977 This parameter controls whether the initramfs
1978 image is unpacked asynchronously, concurrently
1979 with devices being probed and
1980 initialized. This should normally just work,
1981 but as a debugging aid, one can get the
1982 historical behaviour of the initramfs
1983 unpacking being completed before device_ and
1986 initrd= [BOOT] Specify the location of the initial ramdisk
1988 initrdmem= [KNL] Specify a physical address and size from which to
1989 load the initrd. If an initrd is compiled in or
1990 specified in the bootparams, it takes priority over this
1992 Format: ss[KMG],nn[KMG]
1995 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1998 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2000 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
2002 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2004 init_pkru= [X86] Specify the default memory protection keys rights
2005 register contents for all processes. 0x55555554 by
2006 default (disallow access to all but pkey 0). Can
2007 override in debugfs after boot.
2009 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2012 int_pln_enable [X86] Enable power limit notification interrupt
2014 integrity_audit=[IMA]
2015 Format: { "0" | "1" }
2016 0 -- basic integrity auditing messages. (Default)
2017 1 -- additional integrity auditing messages.
2019 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2021 Enable intel iommu driver.
2023 Disable intel iommu driver.
2024 igfx_off [Default Off]
2025 By default, gfx is mapped as normal device. If a gfx
2026 device has a dedicated DMAR unit, the DMAR unit is
2027 bypassed by not enabling DMAR with this option. In
2028 this case, gfx device will use physical address for
2030 strict [Default Off]
2031 Deprecated, equivalent to iommu.strict=1.
2032 sp_off [Default Off]
2033 By default, super page will be supported if Intel IOMMU
2034 has the capability. With this option, super page will
2037 Enable the Intel IOMMU scalable mode if the hardware
2038 advertises that it has support for the scalable mode
2041 Disallow use of the Intel IOMMU scalable mode.
2042 tboot_noforce [Default Off]
2043 Do not force the Intel IOMMU enabled under tboot.
2044 By default, tboot will force Intel IOMMU on, which
2045 could harm performance of some high-throughput
2046 devices like 40GBit network cards, even if identity
2048 Note that using this option lowers the security
2049 provided by tboot because it makes the system
2050 vulnerable to DMA attacks.
2052 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2053 0 disables intel_idle and fall back on acpi_idle.
2054 1 to 9 specify maximum depth of C-state.
2058 Do not enable intel_pstate as the default
2059 scaling driver for the supported processors
2061 Use intel_pstate as a scaling driver, but configure it
2062 to work with generic cpufreq governors (instead of
2063 enabling its internal governor). This mode cannot be
2064 used along with the hardware-managed P-states (HWP)
2067 Enable intel_pstate on systems that prohibit it by default
2068 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2069 instead of acpi-cpufreq may disable platform features, such
2070 as thermal controls and power capping, that rely on ACPI
2071 P-States information being indicated to OSPM and therefore
2072 should be used with caution. This option does not work with
2073 processors that aren't supported by the intel_pstate driver
2074 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2076 Do not enable hardware P state control (HWP)
2079 Only load intel_pstate on systems which support
2080 hardware P state control (HWP) if available.
2082 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2083 Description Table, specifies preferred power management
2084 profile as "Enterprise Server" or "Performance Server",
2085 then this feature is turned on by default.
2087 Allow per-logical-CPU P-State performance control limits using
2088 cpufreq sysfs interface
2090 intremap= [X86-64, Intel-IOMMU]
2091 on enable Interrupt Remapping (default)
2092 off disable Interrupt Remapping
2093 nosid disable Source ID checking
2095 BIOS x2APIC opt-out request will be ignored
2096 nopost disable Interrupt Posting
2098 iomem= Disable strict checking of access to MMIO memory
2099 strict regions from userspace.
2114 nobypass [PPC/POWERNV]
2115 Disable IOMMU bypass, using IOMMU for PCI devices.
2117 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2118 Format: { "0" | "1" }
2119 0 - Try to allocate a 32-bit DMA address first, before
2120 falling back to the full range if needed.
2121 1 - Allocate directly from the full usable range,
2122 forcing Dual Address Cycle for PCI cards supporting
2123 greater than 32-bit addressing.
2125 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2126 Format: { "0" | "1" }
2128 Request that DMA unmap operations use deferred
2129 invalidation of hardware TLBs, for increased
2130 throughput at the cost of reduced device isolation.
2131 Will fall back to strict mode if not supported by
2132 the relevant IOMMU driver.
2134 DMA unmap operations invalidate IOMMU hardware TLBs
2136 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2137 Note: on x86, strict mode specified via one of the
2138 legacy driver-specific options takes precedence.
2141 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2142 Format: { "0" | "1" }
2143 0 - Use IOMMU translation for DMA.
2144 1 - Bypass the IOMMU for DMA.
2145 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2147 io7= [HW] IO7 for Marvel-based Alpha systems
2148 See comment before marvel_specify_io7 in
2149 arch/alpha/kernel/core_marvel.c.
2151 io_delay= [X86] I/O delay method
2153 Standard port 0x80 based delay
2155 Alternate port 0xed based delay (needed on some systems)
2157 Simple two microseconds delay
2162 See Documentation/admin-guide/nfs/nfsroot.rst.
2164 ipcmni_extend [KNL] Extend the maximum number of unique System V
2165 IPC identifiers from 32,768 to 16,777,216.
2167 irqaffinity= [SMP] Set the default irq affinity mask
2168 The argument is a cpu list, as described above.
2170 irqchip.gicv2_force_probe=
2173 Force the kernel to look for the second 4kB page
2174 of a GICv2 controller even if the memory range
2175 exposed by the device tree is too small.
2177 irqchip.gicv3_nolpi=
2179 Force the kernel to ignore the availability of
2180 LPIs (and by consequence ITSs). Intended for system
2181 that use the kernel as a bootloader, and thus want
2182 to let secondary kernels in charge of setting up
2185 irqchip.gicv3_pseudo_nmi= [ARM64]
2186 Enables support for pseudo-NMIs in the kernel. This
2187 requires the kernel to be built with
2188 CONFIG_ARM64_PSEUDO_NMI.
2191 When an interrupt is not handled search all handlers
2192 for it. Intended to get systems with badly broken
2196 When an interrupt is not handled search all handlers
2197 for it. Also check all handlers each timer
2198 interrupt. Intended to get systems with badly broken
2202 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2204 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2205 [Deprecated - use cpusets instead]
2206 Format: [flag-list,]<cpu-list>
2208 Specify one or more CPUs to isolate from disturbances
2209 specified in the flag list (default: domain):
2212 Disable the tick when a single task runs.
2214 A residual 1Hz tick is offloaded to workqueues, which you
2215 need to affine to housekeeping through the global
2216 workqueue's affinity configured via the
2217 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2218 by using the 'domain' flag described below.
2220 NOTE: by default the global workqueue runs on all CPUs,
2221 so to protect individual CPUs the 'cpumask' file has to
2222 be configured manually after bootup.
2225 Isolate from the general SMP balancing and scheduling
2226 algorithms. Note that performing domain isolation this way
2227 is irreversible: it's not possible to bring back a CPU to
2228 the domains once isolated through isolcpus. It's strongly
2229 advised to use cpusets instead to disable scheduler load
2230 balancing through the "cpuset.sched_load_balance" file.
2231 It offers a much more flexible interface where CPUs can
2232 move in and out of an isolated set anytime.
2234 You can move a process onto or off an "isolated" CPU via
2235 the CPU affinity syscalls or cpuset.
2236 <cpu number> begins at 0 and the maximum value is
2237 "number of CPUs in system - 1".
2241 Isolate from being targeted by managed interrupts
2242 which have an interrupt mask containing isolated
2243 CPUs. The affinity of managed interrupts is
2244 handled by the kernel and cannot be changed via
2245 the /proc/irq/* interfaces.
2247 This isolation is best effort and only effective
2248 if the automatically assigned interrupt mask of a
2249 device queue contains isolated and housekeeping
2250 CPUs. If housekeeping CPUs are online then such
2251 interrupts are directed to the housekeeping CPU
2252 so that IO submitted on the housekeeping CPU
2253 cannot disturb the isolated CPU.
2255 If a queue's affinity mask contains only isolated
2256 CPUs then this parameter has no effect on the
2257 interrupt routing decision, though interrupts are
2258 only delivered when tasks running on those
2259 isolated CPUs submit IO. IO submitted on
2260 housekeeping CPUs has no influence on those
2263 The format of <cpu-list> is described above.
2267 ivrs_ioapic [HW,X86-64]
2268 Provide an override to the IOAPIC-ID<->DEVICE-ID
2269 mapping provided in the IVRS ACPI table. For
2270 example, to map IOAPIC-ID decimal 10 to
2271 PCI device 00:14.0 write the parameter as:
2272 ivrs_ioapic[10]=00:14.0
2274 ivrs_hpet [HW,X86-64]
2275 Provide an override to the HPET-ID<->DEVICE-ID
2276 mapping provided in the IVRS ACPI table. For
2277 example, to map HPET-ID decimal 0 to
2278 PCI device 00:14.0 write the parameter as:
2279 ivrs_hpet[0]=00:14.0
2281 ivrs_acpihid [HW,X86-64]
2282 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2283 mapping provided in the IVRS ACPI table. For
2284 example, to map UART-HID:UID AMD0020:0 to
2285 PCI device 00:14.5 write the parameter as:
2286 ivrs_acpihid[00:14.5]=AMD0020:0
2288 js= [HW,JOY] Analog joystick
2289 See Documentation/input/joydev/joystick.rst.
2292 When CONFIG_RANDOMIZE_BASE is set, this disables
2293 kernel and module base offset ASLR (Address Space
2294 Layout Randomization).
2297 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2298 report on every invalid memory access. Without this
2299 parameter KASAN will print report only for the first
2304 kernelcore= [KNL,X86,IA-64,PPC]
2305 Format: nn[KMGTPE] | nn% | "mirror"
2306 This parameter specifies the amount of memory usable by
2307 the kernel for non-movable allocations. The requested
2308 amount is spread evenly throughout all nodes in the
2309 system as ZONE_NORMAL. The remaining memory is used for
2310 movable memory in its own zone, ZONE_MOVABLE. In the
2311 event, a node is too small to have both ZONE_NORMAL and
2312 ZONE_MOVABLE, kernelcore memory will take priority and
2313 other nodes will have a larger ZONE_MOVABLE.
2315 ZONE_MOVABLE is used for the allocation of pages that
2316 may be reclaimed or moved by the page migration
2317 subsystem. Note that allocations like PTEs-from-HighMem
2318 still use the HighMem zone if it exists, and the Normal
2319 zone if it does not.
2321 It is possible to specify the exact amount of memory in
2322 the form of "nn[KMGTPE]", a percentage of total system
2323 memory in the form of "nn%", or "mirror". If "mirror"
2324 option is specified, mirrored (reliable) memory is used
2325 for non-movable allocations and remaining memory is used
2326 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2327 are exclusive, so you cannot specify multiple forms.
2329 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2330 Format: <Controller#>[,poll interval]
2331 The controller # is the number of the ehci usb debug
2332 port as it is probed via PCI. The poll interval is
2333 optional and is the number seconds in between
2334 each poll cycle to the debug port in case you need
2335 the functionality for interrupting the kernel with
2336 gdb or control-c on the dbgp connection. When
2337 not using this parameter you use sysrq-g to break into
2338 the kernel debugger.
2340 kgdboc= [KGDB,HW] kgdb over consoles.
2341 Requires a tty driver that supports console polling,
2342 or a supported polling keyboard driver (non-usb).
2343 Serial only format: <serial_device>[,baud]
2344 keyboard only format: kbd
2345 keyboard and serial format: kbd,<serial_device>[,baud]
2346 Optional Kernel mode setting:
2347 kms, kbd format: kms,kbd
2348 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2350 kgdboc_earlycon= [KGDB,HW]
2351 If the boot console provides the ability to read
2352 characters and can work in polling mode, you can use
2353 this parameter to tell kgdb to use it as a backend
2354 until the normal console is registered. Intended to
2355 be used together with the kgdboc parameter which
2356 specifies the normal console to transition to.
2358 The name of the early console should be specified
2359 as the value of this parameter. Note that the name of
2360 the early console might be different than the tty
2361 name passed to kgdboc. It's OK to leave the value
2362 blank and the first boot console that implements
2363 read() will be picked.
2365 kgdbwait [KGDB] Stop kernel execution and enter the
2366 kernel debugger at the earliest opportunity.
2368 kmac= [MIPS] Korina ethernet MAC address.
2369 Configure the RouterBoard 532 series on-chip
2370 Ethernet adapter MAC address.
2372 kmemleak= [KNL] Boot-time kmemleak enable/disable
2373 Valid arguments: on, off
2375 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2378 kprobe_event=[probe-list]
2379 [FTRACE] Add kprobe events and enable at boot time.
2380 The probe-list is a semicolon delimited list of probe
2381 definitions. Each definition is same as kprobe_events
2382 interface, but the parameters are comma delimited.
2383 For example, to add a kprobe event on vfs_read with
2384 arg1 and arg2, add to the command line;
2386 kprobe_event=p,vfs_read,$arg1,$arg2
2388 See also Documentation/trace/kprobetrace.rst "Kernel
2389 Boot Parameter" section.
2391 kpti= [ARM64] Control page table isolation of user
2392 and kernel address spaces.
2393 Default: enabled on cores which need mitigation.
2397 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2398 Default is 0 (don't ignore, but inject #GP)
2400 kvm.eager_page_split=
2401 [KVM,X86] Controls whether or not KVM will try to
2402 proactively split all huge pages during dirty logging.
2403 Eager page splitting reduces interruptions to vCPU
2404 execution by eliminating the write-protection faults
2405 and MMU lock contention that would otherwise be
2406 required to split huge pages lazily.
2408 VM workloads that rarely perform writes or that write
2409 only to a small region of VM memory may benefit from
2410 disabling eager page splitting to allow huge pages to
2411 still be used for reads.
2413 The behavior of eager page splitting depends on whether
2414 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2415 disabled, all huge pages in a memslot will be eagerly
2416 split when dirty logging is enabled on that memslot. If
2417 enabled, eager page splitting will be performed during
2418 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2421 Eager page splitting currently only supports splitting
2422 huge pages mapped by the TDP MMU.
2426 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2427 Default is false (don't support).
2430 [KVM] Controls the software workaround for the
2431 X86_BUG_ITLB_MULTIHIT bug.
2432 force : Always deploy workaround.
2433 off : Never deploy workaround.
2434 auto : Deploy workaround based on the presence of
2435 X86_BUG_ITLB_MULTIHIT.
2439 If the software workaround is enabled for the host,
2440 guests do need not to enable it for nested guests.
2442 kvm.nx_huge_pages_recovery_ratio=
2443 [KVM] Controls how many 4KiB pages are periodically zapped
2444 back to huge pages. 0 disables the recovery, otherwise if
2445 the value is N KVM will zap 1/Nth of the 4KiB pages every
2446 period (see below). The default is 60.
2448 kvm.nx_huge_pages_recovery_period_ms=
2449 [KVM] Controls the time period at which KVM zaps 4KiB pages
2450 back to huge pages. If the value is a non-zero N, KVM will
2451 zap a portion (see ratio above) of the pages every N msecs.
2452 If the value is 0 (the default), KVM will pick a period based
2453 on the ratio, such that a page is zapped after 1 hour on average.
2455 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2456 Default is 1 (enabled)
2458 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2460 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2463 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2465 none: Forcefully disable KVM.
2467 nvhe: Standard nVHE-based mode, without support for
2470 protected: nVHE-based mode with support for guests whose
2471 state is kept private from the host.
2472 Not valid if the kernel is running in EL2.
2474 Defaults to VHE/nVHE based on hardware support. Setting
2475 mode to "protected" will disable kexec and hibernation
2478 kvm-arm.vgic_v3_group0_trap=
2479 [KVM,ARM] Trap guest accesses to GICv3 group-0
2482 kvm-arm.vgic_v3_group1_trap=
2483 [KVM,ARM] Trap guest accesses to GICv3 group-1
2486 kvm-arm.vgic_v3_common_trap=
2487 [KVM,ARM] Trap guest accesses to GICv3 common
2490 kvm-arm.vgic_v4_enable=
2491 [KVM,ARM] Allow use of GICv4 for direct injection of
2494 kvm_cma_resv_ratio=n [PPC]
2495 Reserves given percentage from system memory area for
2496 contiguous memory allocation for KVM hash pagetable
2498 By default it reserves 5% of total system memory.
2502 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2503 (virtualized MMU) support on capable Intel chips.
2504 Default is 1 (enabled)
2506 kvm-intel.emulate_invalid_guest_state=
2507 [KVM,Intel] Disable emulation of invalid guest state.
2508 Ignored if kvm-intel.enable_unrestricted_guest=1, as
2509 guest state is never invalid for unrestricted guests.
2510 This param doesn't apply to nested guests (L2), as KVM
2511 never emulates invalid L2 guest state.
2512 Default is 1 (enabled)
2514 kvm-intel.flexpriority=
2515 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2516 Default is 1 (enabled)
2519 [KVM,Intel] Enable VMX nesting (nVMX).
2520 Default is 0 (disabled)
2522 kvm-intel.unrestricted_guest=
2523 [KVM,Intel] Disable unrestricted guest feature
2524 (virtualized real and unpaged mode) on capable
2525 Intel chips. Default is 1 (enabled)
2527 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2530 Valid arguments: never, cond, always
2532 always: L1D cache flush on every VMENTER.
2533 cond: Flush L1D on VMENTER only when the code between
2534 VMEXIT and VMENTER can leak host memory.
2535 never: Disables the mitigation
2537 Default is cond (do L1 cache flush in specific instances)
2539 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2540 feature (tagged TLBs) on capable Intel chips.
2541 Default is 1 (enabled)
2543 l1d_flush= [X86,INTEL]
2544 Control mitigation for L1D based snooping vulnerability.
2546 Certain CPUs are vulnerable to an exploit against CPU
2547 internal buffers which can forward information to a
2548 disclosure gadget under certain conditions.
2550 In vulnerable processors, the speculatively
2551 forwarded data can be used in a cache side channel
2552 attack, to access data to which the attacker does
2553 not have direct access.
2555 This parameter controls the mitigation. The
2558 on - enable the interface for the mitigation
2560 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2563 The kernel PTE inversion protection is unconditionally
2564 enabled and cannot be disabled.
2567 Provides all available mitigations for the
2568 L1TF vulnerability. Disables SMT and
2569 enables all mitigations in the
2570 hypervisors, i.e. unconditional L1D flush.
2572 SMT control and L1D flush control via the
2573 sysfs interface is still possible after
2574 boot. Hypervisors will issue a warning
2575 when the first VM is started in a
2576 potentially insecure configuration,
2577 i.e. SMT enabled or L1D flush disabled.
2580 Same as 'full', but disables SMT and L1D
2581 flush runtime control. Implies the
2582 'nosmt=force' command line option.
2583 (i.e. sysfs control of SMT is disabled.)
2586 Leaves SMT enabled and enables the default
2587 hypervisor mitigation, i.e. conditional
2590 SMT control and L1D flush control via the
2591 sysfs interface is still possible after
2592 boot. Hypervisors will issue a warning
2593 when the first VM is started in a
2594 potentially insecure configuration,
2595 i.e. SMT enabled or L1D flush disabled.
2599 Disables SMT and enables the default
2600 hypervisor mitigation.
2602 SMT control and L1D flush control via the
2603 sysfs interface is still possible after
2604 boot. Hypervisors will issue a warning
2605 when the first VM is started in a
2606 potentially insecure configuration,
2607 i.e. SMT enabled or L1D flush disabled.
2610 Same as 'flush', but hypervisors will not
2611 warn when a VM is started in a potentially
2612 insecure configuration.
2615 Disables hypervisor mitigations and doesn't
2617 It also drops the swap size and available
2618 RAM limit restriction on both hypervisor and
2623 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2629 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2632 lapic= [X86,APIC] Do not use TSC deadline
2633 value for LAPIC timer one-shot implementation. Default
2634 back to the programmable timer unit in the LAPIC.
2635 Format: notscdeadline
2637 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2640 libata.dma= [LIBATA] DMA control
2641 libata.dma=0 Disable all PATA and SATA DMA
2642 libata.dma=1 PATA and SATA Disk DMA only
2643 libata.dma=2 ATAPI (CDROM) DMA only
2644 libata.dma=4 Compact Flash DMA only
2645 Combinations also work, so libata.dma=3 enables DMA
2646 for disks and CDROMs, but not CFs.
2648 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2649 libata.ignore_hpa=0 keep BIOS limits (default)
2650 libata.ignore_hpa=1 ignore limits, using full disk
2652 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2656 libata.force= [LIBATA] Force configurations. The format is a comma-
2657 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2658 PORT and DEVICE are decimal numbers matching port, link
2659 or device. Basically, it matches the ATA ID string
2660 printed on console by libata. If the whole ID part is
2661 omitted, the last PORT and DEVICE values are used. If
2662 ID hasn't been specified yet, the configuration applies
2663 to all ports, links and devices.
2665 If only DEVICE is omitted, the parameter applies to
2666 the port and all links and devices behind it. DEVICE
2667 number of 0 either selects the first device or the
2668 first fan-out link behind PMP device. It does not
2669 select the host link. DEVICE number of 15 selects the
2670 host link and device attached to it.
2672 The VAL specifies the configuration to force. As long
2673 as there is no ambiguity, shortcut notation is allowed.
2674 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2675 The following configurations can be forced.
2677 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2678 Any ID with matching PORT is used.
2680 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2682 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2683 udma[/][16,25,33,44,66,100,133] notation is also
2686 * nohrst, nosrst, norst: suppress hard, soft and both
2689 * rstonce: only attempt one reset during hot-unplug
2692 * [no]dbdelay: Enable or disable the extra 200ms delay
2693 before debouncing a link PHY and device presence
2696 * [no]ncq: Turn on or off NCQ.
2698 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2700 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2702 * [no]trim: Enable or disable (unqueued) TRIM.
2704 * trim_zero: Indicate that TRIM command zeroes data.
2706 * max_trim_128m: Set 128M maximum trim size limit.
2708 * [no]dma: Turn on or off DMA transfers.
2710 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2712 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2713 commands that are not a multiple of 16 bytes.
2715 * [no]dmalog: Enable or disable the use of the
2716 READ LOG DMA EXT command to access logs.
2718 * [no]iddevlog: Enable or disable access to the
2719 identify device data log.
2721 * [no]logdir: Enable or disable access to the general
2722 purpose log directory.
2724 * max_sec_128: Set transfer size limit to 128 sectors.
2726 * max_sec_1024: Set or clear transfer size limit to
2729 * max_sec_lba48: Set or clear transfer size limit to
2732 * [no]lpm: Enable or disable link power management.
2734 * [no]setxfer: Indicate if transfer speed mode setting
2737 * dump_id: Dump IDENTIFY data.
2739 * disable: Disable this device.
2741 If there are multiple matching configurations changing
2742 the same attribute, the last one is used.
2744 load_ramdisk= [RAM] [Deprecated]
2746 lockd.nlm_grace_period=P [NFS] Assign grace period.
2749 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2752 lockd.nlm_timeout=T [NFS] Assign timeout value.
2755 lockd.nlm_udpport=M [NFS] Assign UDP port.
2758 lockdown= [SECURITY]
2759 { integrity | confidentiality }
2760 Enable the kernel lockdown feature. If set to
2761 integrity, kernel features that allow userland to
2762 modify the running kernel are disabled. If set to
2763 confidentiality, kernel features that allow userland
2764 to extract confidential information from the kernel
2767 locktorture.nreaders_stress= [KNL]
2768 Set the number of locking read-acquisition kthreads.
2769 Defaults to being automatically set based on the
2770 number of online CPUs.
2772 locktorture.nwriters_stress= [KNL]
2773 Set the number of locking write-acquisition kthreads.
2775 locktorture.onoff_holdoff= [KNL]
2776 Set time (s) after boot for CPU-hotplug testing.
2778 locktorture.onoff_interval= [KNL]
2779 Set time (s) between CPU-hotplug operations, or
2780 zero to disable CPU-hotplug testing.
2782 locktorture.shuffle_interval= [KNL]
2783 Set task-shuffle interval (jiffies). Shuffling
2784 tasks allows some CPUs to go into dyntick-idle
2785 mode during the locktorture test.
2787 locktorture.shutdown_secs= [KNL]
2788 Set time (s) after boot system shutdown. This
2789 is useful for hands-off automated testing.
2791 locktorture.stat_interval= [KNL]
2792 Time (s) between statistics printk()s.
2794 locktorture.stutter= [KNL]
2795 Time (s) to stutter testing, for example,
2796 specifying five seconds causes the test to run for
2797 five seconds, wait for five seconds, and so on.
2798 This tests the locking primitive's ability to
2799 transition abruptly to and from idle.
2801 locktorture.torture_type= [KNL]
2802 Specify the locking implementation to test.
2804 locktorture.verbose= [KNL]
2805 Enable additional printk() statements.
2807 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2810 loglevel= All Kernel Messages with a loglevel smaller than the
2811 console loglevel will be printed to the console. It can
2812 also be changed with klogd or other programs. The
2813 loglevels are defined as follows:
2815 0 (KERN_EMERG) system is unusable
2816 1 (KERN_ALERT) action must be taken immediately
2817 2 (KERN_CRIT) critical conditions
2818 3 (KERN_ERR) error conditions
2819 4 (KERN_WARNING) warning conditions
2820 5 (KERN_NOTICE) normal but significant condition
2821 6 (KERN_INFO) informational
2822 7 (KERN_DEBUG) debug-level messages
2824 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2825 in bytes. n must be a power of two and greater
2826 than the minimal size. The minimal size is defined
2827 by LOG_BUF_SHIFT kernel config parameter. There is
2828 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2829 that allows to increase the default size depending on
2830 the number of CPUs. See init/Kconfig for more details.
2832 logo.nologo [FB] Disables display of the built-in Linux logo.
2833 This may be used to provide more screen space for
2834 kernel log messages and is useful when debugging
2835 kernel boot problems.
2837 lp=0 [LP] Specify parallel ports to use, e.g,
2838 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2839 lp=reset first parallel port). 'lp=0' disables the
2840 lp=auto printer driver. 'lp=reset' (which can be
2841 specified in addition to the ports) causes
2842 attached printers to be reset. Using
2843 lp=port1,port2,... specifies the parallel ports
2844 to associate lp devices with, starting with
2845 lp0. A port specification may be 'none' to skip
2846 that lp device, or a parport name such as
2847 'parport0'. Specifying 'lp=auto' instead of a
2848 port specification list means that device IDs
2849 from each port should be examined, to see if
2850 an IEEE 1284-compliant printer is attached; if
2851 so, the driver will manage that printer.
2852 See also header of drivers/char/lp.c.
2855 Sets loops_per_jiffy to given constant, thus avoiding
2856 time-consuming boot-time autodetection (up to 250 ms per
2857 CPU). 0 enables autodetection (default). To determine
2858 the correct value for your kernel, boot with normal
2859 autodetection and see what value is printed. Note that
2860 on SMP systems the preset will be applied to all CPUs,
2861 which is likely to cause problems if your CPUs need
2862 significantly divergent settings. An incorrect value
2863 will cause delays in the kernel to be wrong, leading to
2864 unpredictable I/O errors and other breakage. Although
2865 unlikely, in the extreme case this might damage your
2869 Format: <io>,<irq>,<dma>
2871 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2874 [SECURITY] Choose order of LSM initialization. This
2875 overrides CONFIG_LSM, and the "security=" parameter.
2877 machvec= [IA-64] Force the use of a particular machine-vector
2878 (machvec) in a generic kernel.
2879 Example: machvec=hpzx1
2881 machtype= [Loongson] Share the same kernel image file between
2882 different yeeloong laptops.
2883 Example: machtype=lemote-yeeloong-2f-7inch
2885 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
2886 than or equal to this physical address is ignored.
2888 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2889 will bring up during bootup. maxcpus=n : n >= 0 limits
2890 the kernel to bring up 'n' processors. Surely after
2891 bootup you can bring up the other plugged cpu by executing
2892 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2893 only takes effect during system bootup.
2894 While n=0 is a special case, it is equivalent to "nosmp",
2895 which also disables the IO APIC.
2897 max_loop= [LOOP] The number of loop block devices that get
2898 (loop.max_loop) unconditionally pre-created at init time. The default
2899 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2900 of statically allocating a predefined number, loop
2901 devices can be requested on-demand with the
2902 /dev/loop-control interface.
2904 mce [X86-32] Machine Check Exception
2906 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2908 md= [HW] RAID subsystems devices and level
2909 See Documentation/admin-guide/md.rst.
2912 Format: <first>,<last>
2913 Specifies range of consoles to be captured by the MDA.
2916 Control mitigation for the Micro-architectural Data
2917 Sampling (MDS) vulnerability.
2919 Certain CPUs are vulnerable to an exploit against CPU
2920 internal buffers which can forward information to a
2921 disclosure gadget under certain conditions.
2923 In vulnerable processors, the speculatively
2924 forwarded data can be used in a cache side channel
2925 attack, to access data to which the attacker does
2926 not have direct access.
2928 This parameter controls the MDS mitigation. The
2931 full - Enable MDS mitigation on vulnerable CPUs
2932 full,nosmt - Enable MDS mitigation and disable
2933 SMT on vulnerable CPUs
2934 off - Unconditionally disable MDS mitigation
2936 On TAA-affected machines, mds=off can be prevented by
2937 an active TAA mitigation as both vulnerabilities are
2938 mitigated with the same mechanism so in order to disable
2939 this mitigation, you need to specify tsx_async_abort=off
2942 Not specifying this option is equivalent to
2945 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2947 mem=nn[KMG] [HEXAGON] Set the memory size.
2948 Must be specified, otherwise memory size will be 0.
2950 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2951 Amount of memory to be used in cases as follows:
2954 2 when the kernel is not able to see the whole system memory;
2955 3 memory that lies after 'mem=' boundary is excluded from
2956 the hypervisor, then assigned to KVM guests.
2957 4 to limit the memory available for kdump kernel.
2959 [ARC,MICROBLAZE] - the limit applies only to low memory,
2960 high memory is not affected.
2962 [ARM64] - only limits memory covered by the linear
2963 mapping. The NOMAP regions are not affected.
2965 [X86] Work as limiting max address. Use together
2966 with memmap= to avoid physical address space collisions.
2967 Without memmap= PCI devices could be placed at addresses
2968 belonging to unused RAM.
2970 Note that this only takes effects during boot time since
2971 in above case 3, memory may need be hot added after boot
2972 if system memory of hypervisor is not sufficient.
2975 [ARM,MIPS] - override the memory layout reported by
2977 Define a memory region of size nn[KMG] starting at
2979 Multiple different regions can be specified with
2980 multiple mem= parameters on the command line.
2982 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2985 memblock=debug [KNL] Enable memblock debug messages.
2988 [KNL,SH] Allow user to override the default size for
2989 per-device physically contiguous DMA buffers.
2991 memhp_default_state=online/offline
2992 [KNL] Set the initial state for the memory hotplug
2993 onlining policy. If not specified, the default value is
2994 set according to the
2995 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2997 See Documentation/admin-guide/mm/memory-hotplug.rst.
2999 memmap=exactmap [KNL,X86] Enable setting of an exact
3000 E820 memory map, as specified by the user.
3001 Such memmap=exactmap lines can be constructed based on
3002 BIOS output or other requirements. See the memmap=nn@ss
3005 memmap=nn[KMG]@ss[KMG]
3006 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
3007 Region of memory to be used is from ss to ss+nn.
3008 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3009 which limits max address to nn[KMG].
3010 Multiple different regions can be specified,
3013 memmap=100M@2G,100M#3G,1G!1024G
3015 memmap=nn[KMG]#ss[KMG]
3016 [KNL,ACPI] Mark specific memory as ACPI data.
3017 Region of memory to be marked is from ss to ss+nn.
3019 memmap=nn[KMG]$ss[KMG]
3020 [KNL,ACPI] Mark specific memory as reserved.
3021 Region of memory to be reserved is from ss to ss+nn.
3022 Example: Exclude memory from 0x18690000-0x1869ffff
3023 memmap=64K$0x18690000
3025 memmap=0x10000$0x18690000
3026 Some bootloaders may need an escape character before '$',
3027 like Grub2, otherwise '$' and the following number
3030 memmap=nn[KMG]!ss[KMG]
3031 [KNL,X86] Mark specific memory as protected.
3032 Region of memory to be used, from ss to ss+nn.
3033 The memory region may be marked as e820 type 12 (0xc)
3034 and is NVDIMM or ADR memory.
3036 memmap=<size>%<offset>-<oldtype>+<newtype>
3037 [KNL,ACPI] Convert memory within the specified region
3038 from <oldtype> to <newtype>. If "-<oldtype>" is left
3039 out, the whole region will be marked as <newtype>,
3040 even if previously unavailable. If "+<newtype>" is left
3041 out, matching memory will be removed. Types are
3042 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3043 3 = ACPI, 12 = PRAM.
3045 memory_corruption_check=0/1 [X86]
3046 Some BIOSes seem to corrupt the first 64k of
3047 memory when doing things like suspend/resume.
3048 Setting this option will scan the memory
3049 looking for corruption. Enabling this will
3050 both detect corruption and prevent the kernel
3051 from using the memory being corrupted.
3052 However, its intended as a diagnostic tool; if
3053 repeatable BIOS-originated corruption always
3054 affects the same memory, you can use memmap=
3055 to prevent the kernel from using that memory.
3057 memory_corruption_check_size=size [X86]
3058 By default it checks for corruption in the low
3059 64k, making this memory unavailable for normal
3060 use. Use this parameter to scan for
3061 corruption in more or less memory.
3063 memory_corruption_check_period=seconds [X86]
3064 By default it checks for corruption every 60
3065 seconds. Use this parameter to check at some
3066 other rate. 0 disables periodic checking.
3068 memory_hotplug.memmap_on_memory
3069 [KNL,X86,ARM] Boolean flag to enable this feature.
3070 Format: {on | off (default)}
3071 When enabled, runtime hotplugged memory will
3072 allocate its internal metadata (struct pages)
3073 from the hotadded memory which will allow to
3074 hotadd a lot of memory without requiring
3075 additional memory to do so.
3076 This feature is disabled by default because it
3077 has some implication on large (e.g. GB)
3078 allocations in some configurations (e.g. small
3080 The state of the flag can be read in
3081 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3082 Note that even when enabled, there are a few cases where
3083 the feature is not effective.
3085 This is not compatible with hugetlb_free_vmemmap. If
3086 both parameters are enabled, hugetlb_free_vmemmap takes
3087 precedence over memory_hotplug.memmap_on_memory.
3089 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3091 default : 0 <disable>
3092 Specifies the number of memtest passes to be
3093 performed. Each pass selects another test
3094 pattern from a given set of patterns. Memtest
3095 fills the memory with this pattern, validates
3096 memory contents and reserves bad memory
3097 regions that are detected.
3099 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3100 Valid arguments: on, off
3101 Default (depends on kernel configuration option):
3102 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3103 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3104 mem_encrypt=on: Activate SME
3105 mem_encrypt=off: Do not activate SME
3107 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
3108 for details on when memory encryption can be activated.
3110 mem_sleep_default= [SUSPEND] Default system suspend mode:
3111 s2idle - Suspend-To-Idle
3112 shallow - Power-On Suspend or equivalent (if supported)
3113 deep - Suspend-To-RAM or equivalent (if supported)
3114 See Documentation/admin-guide/pm/sleep-states.rst.
3116 meye.*= [HW] Set MotionEye Camera parameters
3117 See Documentation/admin-guide/media/meye.rst.
3119 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3120 Multi-Function General Purpose Timers on AMD Geode
3123 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3124 the BIOS has incorrectly applied a workaround. TinyBIOS
3125 version 0.98 is known to be affected, 0.99 fixes the
3126 problem by letting the user disable the workaround.
3130 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3131 physical address is ignored.
3133 mini2440= [ARM,HW,KNL]
3134 Format:[0..2][b][c][t]
3136 MINI2440 configuration specification:
3137 0 - The attached screen is the 3.5" TFT
3138 1 - The attached screen is the 7" TFT
3139 2 - The VGA Shield is attached (1024x768)
3140 Leaving out the screen size parameter will not load
3141 the TFT driver, and the framebuffer will be left
3143 b - Enable backlight. The TFT backlight pin will be
3144 linked to the kernel VESA blanking code and a GPIO
3145 LED. This parameter is not necessary when using the
3147 c - Enable the s3c camera interface.
3148 t - Reserved for enabling touchscreen support. The
3149 touchscreen support is not enabled in the mainstream
3150 kernel as of 2.6.30, a preliminary port can be found
3151 in the "bleeding edge" mini2440 support kernel at
3152 https://repo.or.cz/w/linux-2.6/mini2440.git
3155 [X86,PPC,S390,ARM64] Control optional mitigations for
3156 CPU vulnerabilities. This is a set of curated,
3157 arch-independent options, each of which is an
3158 aggregation of existing arch-specific options.
3161 Disable all optional CPU mitigations. This
3162 improves system performance, but it may also
3163 expose users to several CPU vulnerabilities.
3164 Equivalent to: nopti [X86,PPC]
3166 nospectre_v1 [X86,PPC]
3168 nospectre_v2 [X86,PPC,S390,ARM64]
3169 spectre_v2_user=off [X86]
3170 spec_store_bypass_disable=off [X86,PPC]
3171 ssbd=force-off [ARM64]
3174 tsx_async_abort=off [X86]
3175 kvm.nx_huge_pages=off [X86]
3176 srbds=off [X86,INTEL]
3177 no_entry_flush [PPC]
3178 no_uaccess_flush [PPC]
3181 This does not have any effect on
3182 kvm.nx_huge_pages when
3183 kvm.nx_huge_pages=force.
3186 Mitigate all CPU vulnerabilities, but leave SMT
3187 enabled, even if it's vulnerable. This is for
3188 users who don't want to be surprised by SMT
3189 getting disabled across kernel upgrades, or who
3190 have other ways of avoiding SMT-based attacks.
3191 Equivalent to: (default behavior)
3194 Mitigate all CPU vulnerabilities, disabling SMT
3195 if needed. This is for users who always want to
3196 be fully mitigated, even if it means losing SMT.
3197 Equivalent to: l1tf=flush,nosmt [X86]
3198 mds=full,nosmt [X86]
3199 tsx_async_abort=full,nosmt [X86]
3202 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3203 parameter allows control of the logging verbosity for
3204 the additional memory initialisation checks. A value
3205 of 0 disables mminit logging and a level of 4 will
3206 log everything. Information is printed at KERN_DEBUG
3207 so loglevel=8 may also need to be specified.
3210 [KNL] When CONFIG_MODULE_SIG is set, this means that
3211 modules without (valid) signatures will fail to load.
3212 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3213 is always true, so this option does nothing.
3215 module_blacklist= [KNL] Do not load a comma-separated list of
3216 modules. Useful for debugging problem modules.
3219 [MOUSE] Maximum time between finger touching and
3220 leaving touchpad surface for touch to be considered
3221 a tap and be reported as a left button click (for
3222 touchpads working in absolute mode only).
3224 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3225 reporting absolute coordinates, such as tablets
3226 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3227 reporting absolute coordinates, such as tablets
3229 movablecore= [KNL,X86,IA-64,PPC]
3230 Format: nn[KMGTPE] | nn%
3231 This parameter is the complement to kernelcore=, it
3232 specifies the amount of memory used for migratable
3233 allocations. If both kernelcore and movablecore is
3234 specified, then kernelcore will be at *least* the
3235 specified value but may be more. If movablecore on its
3236 own is specified, the administrator must be careful
3237 that the amount of memory usable for all allocations
3240 movable_node [KNL] Boot-time switch to make hotplugable memory
3241 NUMA nodes to be movable. This means that the memory
3242 of such nodes will be usable only for movable
3243 allocations which rules out almost all kernel
3244 allocations. Use with caution!
3246 MTD_Partition= [MTD]
3247 Format: <name>,<region-number>,<size>,<offset>
3249 MTD_Region= [MTD] Format:
3250 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3253 See drivers/mtd/parsers/cmdlinepart.c
3256 ARM/S3C2412 JIVE boot control
3258 See arch/arm/mach-s3c/mach-jive.c
3260 mtouchusb.raw_coordinates=
3261 [HW] Make the MicroTouch USB driver use raw coordinates
3262 ('y', default) or cooked coordinates ('n')
3264 mtrr_chunk_size=nn[KMG] [X86]
3265 used for mtrr cleanup. It is largest continuous chunk
3266 that could hold holes aka. UC entries.
3268 mtrr_gran_size=nn[KMG] [X86]
3269 Used for mtrr cleanup. It is granularity of mtrr block.
3271 Large value could prevent small alignment from
3274 mtrr_spare_reg_nr=n [X86]
3276 Range: 0,7 : spare reg number
3278 Used for mtrr cleanup. It is spare mtrr entries number.
3279 Set to 2 or more if your graphical card needs more.
3281 multitce=off [PPC] This parameter disables the use of the pSeries
3282 firmware feature for updating multiple TCE entries
3285 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3287 netdev= [NET] Network devices parameters
3288 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3289 Note that mem_start is often overloaded to mean
3290 something different and driver-specific.
3291 This usage is only documented in each driver source
3294 netpoll.carrier_timeout=
3295 [NET] Specifies amount of time (in seconds) that
3296 netpoll should wait for a carrier. By default netpoll
3300 [NETFILTER] Enable connection tracking flow accounting
3301 0 to disable accounting
3302 1 to enable accounting
3305 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3306 See Documentation/admin-guide/nfs/nfsroot.rst.
3308 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3309 See Documentation/admin-guide/nfs/nfsroot.rst.
3311 nfsrootdebug [NFS] enable nfsroot debugging messages.
3312 See Documentation/admin-guide/nfs/nfsroot.rst.
3314 nfs.callback_nr_threads=
3315 [NFSv4] set the total number of threads that the
3316 NFS client will assign to service NFSv4 callback
3319 nfs.callback_tcpport=
3320 [NFS] set the TCP port on which the NFSv4 callback
3321 channel should listen.
3324 [NFS] sets the pathname to the program which is used
3325 to update the NFS client cache entries.
3327 nfs.cache_getent_timeout=
3328 [NFS] sets the timeout after which an attempt to
3329 update a cache entry is deemed to have failed.
3331 nfs.idmap_cache_timeout=
3332 [NFS] set the maximum lifetime for idmapper cache
3336 [NFS] enable 64-bit inode numbers.
3337 If zero, the NFS client will fake up a 32-bit inode
3338 number for the readdir() and stat() syscalls instead
3339 of returning the full 64-bit number.
3340 The default is to return 64-bit inode numbers.
3342 nfs.max_session_cb_slots=
3343 [NFSv4.1] Sets the maximum number of session
3344 slots the client will assign to the callback
3345 channel. This determines the maximum number of
3346 callbacks the client will process in parallel for
3347 a particular server.
3349 nfs.max_session_slots=
3350 [NFSv4.1] Sets the maximum number of session slots
3351 the client will attempt to negotiate with the server.
3352 This limits the number of simultaneous RPC requests
3353 that the client can send to the NFSv4.1 server.
3354 Note that there is little point in setting this
3355 value higher than the max_tcp_slot_table_limit.
3357 nfs.nfs4_disable_idmapping=
3358 [NFSv4] When set to the default of '1', this option
3359 ensures that both the RPC level authentication
3360 scheme and the NFS level operations agree to use
3361 numeric uids/gids if the mount is using the
3362 'sec=sys' security flavour. In effect it is
3363 disabling idmapping, which can make migration from
3364 legacy NFSv2/v3 systems to NFSv4 easier.
3365 Servers that do not support this mode of operation
3366 will be autodetected by the client, and it will fall
3367 back to using the idmapper.
3368 To turn off this behaviour, set the value to '0'.
3370 [NFS4] Specify an additional fixed unique ident-
3371 ification string that NFSv4 clients can insert into
3372 their nfs_client_id4 string. This is typically a
3373 UUID that is generated at system install time.
3375 nfs.send_implementation_id =
3376 [NFSv4.1] Send client implementation identification
3377 information in exchange_id requests.
3378 If zero, no implementation identification information
3380 The default is to send the implementation identification
3383 nfs.recover_lost_locks =
3384 [NFSv4] Attempt to recover locks that were lost due
3385 to a lease timeout on the server. Please note that
3386 doing this risks data corruption, since there are
3387 no guarantees that the file will remain unchanged
3388 after the locks are lost.
3389 If you want to enable the kernel legacy behaviour of
3390 attempting to recover these locks, then set this
3392 The default parameter value of '0' causes the kernel
3393 not to attempt recovery of lost locks.
3395 nfs4.layoutstats_timer =
3396 [NFSv4.2] Change the rate at which the kernel sends
3397 layoutstats to the pNFS metadata server.
3399 Setting this to value to 0 causes the kernel to use
3400 whatever value is the default set by the layout
3401 driver. A non-zero value sets the minimum interval
3402 in seconds between layoutstats transmissions.
3404 nfsd.inter_copy_offload_enable =
3405 [NFSv4.2] When set to 1, the server will support
3406 server-to-server copies for which this server is
3407 the destination of the copy.
3409 nfsd.nfsd4_ssc_umount_timeout =
3410 [NFSv4.2] When used as the destination of a
3411 server-to-server copy, knfsd temporarily mounts
3412 the source server. It caches the mount in case
3413 it will be needed again, and discards it if not
3414 used for the number of milliseconds specified by
3417 nfsd.nfs4_disable_idmapping=
3418 [NFSv4] When set to the default of '1', the NFSv4
3419 server will return only numeric uids and gids to
3420 clients using auth_sys, and will accept numeric uids
3421 and gids from such clients. This is intended to ease
3422 migration from NFSv2/v3.
3425 nmi_backtrace.backtrace_idle [KNL]
3426 Dump stacks even of idle CPUs in response to an
3427 NMI stack-backtrace request.
3429 nmi_debug= [KNL,SH] Specify one or more actions to take
3430 when a NMI is triggered.
3431 Format: [state][,regs][,debounce][,die]
3433 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3434 Format: [panic,][nopanic,][num]
3436 0 - turn hardlockup detector in nmi_watchdog off
3437 1 - turn hardlockup detector in nmi_watchdog on
3438 When panic is specified, panic when an NMI watchdog
3439 timeout occurs (or 'nopanic' to not panic on an NMI
3440 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3441 To disable both hard and soft lockup detectors,
3442 please see 'nowatchdog'.
3443 This is useful when you use a panic=... timeout and
3444 need the box quickly up again.
3446 These settings can be accessed at runtime via
3447 the nmi_watchdog and hardlockup_panic sysctls.
3449 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3450 emulation library even if a 387 maths coprocessor
3453 no5lvl [X86-64] Disable 5-level paging mode. Forces
3454 kernel to use 4-level paging instead.
3456 nofsgsbase [X86] Disables FSGSBASE instructions.
3459 [HW] Never suspend the console
3460 Disable suspending of consoles during suspend and
3461 hibernate operations. Once disabled, debugging
3462 messages can reach various consoles while the rest
3463 of the system is being put to sleep (ie, while
3464 debugging driver suspend/resume hooks). This may
3465 not work reliably with all consoles, but is known
3466 to work with serial and VGA consoles.
3467 To facilitate more flexible debugging, we also add
3468 console_suspend, a printk module parameter to control
3469 it. Users could use console_suspend (usually
3470 /sys/module/printk/parameters/console_suspend) to
3471 turn on/off it dynamically.
3473 novmcoredd [KNL,KDUMP]
3474 Disable device dump. Device dump allows drivers to
3475 append dump data to vmcore so you can collect driver
3476 specified debug info. Drivers can append the data
3477 without any limit and this data is stored in memory,
3478 so this may cause significant memory stress. Disabling
3479 device dump can help save memory but the driver debug
3480 data will be no longer available. This parameter
3481 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3484 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3485 caches in the slab allocator. Saves per-node memory,
3486 but will impact performance.
3490 noaltinstr [S390] Disables alternative instructions patching
3491 (CPU alternatives feature).
3493 noapic [SMP,APIC] Tells the kernel to not make use of any
3494 IOAPICs that may be present in the system.
3496 noautogroup Disable scheduler automatic task group creation.
3498 nobats [PPC] Do not use BATs for mapping kernel lowmem
3499 on "Classic" PPC cores.
3503 nodsp [SH] Disable hardware DSP at boot time.
3505 noefi Disable EFI runtime services support.
3507 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3512 Disable SMAP (Supervisor Mode Access Prevention)
3513 even if it is supported by processor.
3516 Disable SMEP (Supervisor Mode Execution Prevention)
3517 even if it is supported by processor.
3520 This affects only 32-bit executables.
3521 noexec32=on: enable non-executable mappings (default)
3522 read doesn't imply executable mappings
3523 noexec32=off: disable non-executable mappings
3524 read implies executable mappings
3526 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3528 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3529 register save and restore. The kernel will only save
3530 legacy floating-point registers on task switch.
3532 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3534 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3536 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3537 Equivalent to smt=1.
3539 [KNL,X86] Disable symmetric multithreading (SMT).
3540 nosmt=force: Force disable SMT, cannot be undone
3541 via the sysfs control file.
3543 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3544 (bounds check bypass). With this option data leaks are
3545 possible in the system.
3547 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3548 the Spectre variant 2 (indirect branch prediction)
3549 vulnerability. System may allow data leaks with this
3552 nospec_store_bypass_disable
3553 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3556 [PPC] Don't flush the L1-D cache after accessing user data.
3558 noxsave [BUGS=X86] Disables x86 extended register state save
3559 and restore using xsave. The kernel will fallback to
3560 enabling legacy floating-point and sse state.
3562 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3563 register states. The kernel will fall back to use
3564 xsave to save the states. By using this parameter,
3565 performance of saving the states is degraded because
3566 xsave doesn't support modified optimization while
3567 xsaveopt supports it on xsaveopt enabled systems.
3569 noxsaves [X86] Disables xsaves and xrstors used in saving and
3570 restoring x86 extended register state in compacted
3571 form of xsave area. The kernel will fall back to use
3572 xsaveopt and xrstor to save and restore the states
3573 in standard form of xsave area. By using this
3574 parameter, xsave area per process might occupy more
3575 memory on xsaves enabled systems.
3577 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3578 in do_idle() and not use the arch_cpu_idle()
3579 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3580 to be effective. This is useful on platforms where the
3581 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3582 correctly or when doing power measurements to evalute
3583 the impact of the sleep instructions. This is also
3584 useful when using JTAG debugger.
3586 no_file_caps Tells the kernel not to honor file capabilities. The
3587 only way then for a file to be executed with privilege
3588 is to be setuid root or executed by root.
3590 nohalt [IA-64] Tells the kernel not to use the power saving
3591 function PAL_HALT_LIGHT when idle. This increases
3592 power-consumption. On the positive side, it reduces
3593 interrupt wake-up latency, which may improve performance
3594 in certain environments such as networked servers or
3598 Force pointers printed to the console or buffers to be
3599 unhashed. By default, when a pointer is printed via %p
3600 format string, that pointer is "hashed", i.e. obscured
3601 by hashing the pointer value. This is a security feature
3602 that hides actual kernel addresses from unprivileged
3603 users, but it also makes debugging the kernel more
3604 difficult since unequal pointers can no longer be
3605 compared. However, if this command-line option is
3606 specified, then all normal pointers will have their true
3607 value printed. This option should only be specified when
3608 debugging the kernel. Please do not use on production
3611 nohibernate [HIBERNATION] Disable hibernation and resume.
3613 nohz= [KNL] Boottime enable/disable dynamic ticks
3614 Valid arguments: on, off
3617 nohz_full= [KNL,BOOT,SMP,ISOL]
3618 The argument is a cpu list, as described above.
3619 In kernels built with CONFIG_NO_HZ_FULL=y, set
3620 the specified list of CPUs whose tick will be stopped
3621 whenever possible. The boot CPU will be forced outside
3622 the range to maintain the timekeeping. Any CPUs
3623 in this list will have their RCU callbacks offloaded,
3624 just as if they had also been called out in the
3625 rcu_nocbs= boot parameter.
3627 noiotrap [SH] Disables trapped I/O port accesses.
3629 noirqdebug [X86-32] Disables the code which attempts to detect and
3630 disable unhandled interrupt sources.
3632 no_timer_check [X86,APIC] Disables the code which tests for
3633 broken timer IRQ sources.
3635 noisapnp [ISAPNP] Disables ISA PnP code.
3637 noinitrd [RAM] Tells the kernel not to load any configured
3640 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3642 [Deprecated - use intremap=off]
3646 noinvpcid [X86] Disable the INVPCID cpu feature.
3648 nojitter [IA-64] Disables jitter checking for ITC timers.
3650 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3652 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3656 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3657 clock and use the default one.
3659 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3660 accounting. steal time is computed, but won't
3661 influence scheduler behaviour
3663 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3665 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3667 noltlbs [PPC] Do not use large page/tlb entries for kernel
3668 lowmem mapping on PPC40x and PPC8xx
3670 nomca [IA-64] Disable machine check abort handling
3672 nomce [X86-32] Disable Machine Check Exception
3674 nomfgpt [X86-32] Disable Multi-Function General Purpose
3675 Timer usage (for AMD Geode machines).
3677 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3678 shutdown the other cpus. Instead use the REBOOT_VECTOR
3681 nomodeset Disable kernel modesetting. DRM drivers will not perform
3682 display-mode changes or accelerated rendering. Only the
3683 system framebuffer will be available for use if this was
3684 set-up by the firmware or boot loader.
3686 Useful as fallback, or for testing and debugging.
3688 nomodule Disable module load
3690 nopat [X86] Disable PAT (page attribute table extension of
3691 pagetables) support.
3693 nopcid [X86-64] Disable the PCID cpu feature.
3695 norandmaps Don't use address space randomization. Equivalent to
3696 echo 0 > /proc/sys/kernel/randomize_va_space
3698 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3699 with UP alternatives
3701 nordrand [X86] Disable kernel use of the RDRAND and
3702 RDSEED instructions even if they are supported
3703 by the processor. RDRAND and RDSEED are still
3704 available to user space applications.
3706 noresume [SWSUSP] Disables resume and restores original swap
3709 no-scroll [VGA] Disables scrollback.
3710 This is required for the Braillex ib80-piezo Braille
3711 reader made by F.H. Papenmeier (Germany).
3715 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3717 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3718 and disable the IO APIC. legacy for "maxcpus=0".
3720 nosoftlockup [KNL] Disable the soft-lockup detector.
3722 nosync [HW,M68K] Disables sync negotiation for all devices.
3724 nowatchdog [KNL] Disable both lockup detectors, i.e.
3725 soft-lockup and NMI watchdog (hard-lockup).
3729 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3731 nps_mtm_hs_ctr= [KNL,ARC]
3732 This parameter sets the maximum duration, in
3733 cycles, each HW thread of the CTOP can run
3734 without interruptions, before HW switches it.
3735 The actual maximum duration is 16 times this
3737 Format: integer between 1 and 255
3740 nptcg= [IA-64] Override max number of concurrent global TLB
3741 purges which is reported from either PAL_VM_SUMMARY or
3744 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3745 could support. nr_cpus=n : n >= 1 limits the kernel to
3746 support 'n' processors. It could be larger than the
3747 number of already plugged CPU during bootup, later in
3748 runtime you can physically add extra cpu until it reaches
3749 n. So during boot up some boot time memory for per-cpu
3750 variables need be pre-allocated for later physical cpu
3753 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3755 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3756 set up a single NUMA node spanning all memory.
3758 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3760 Allowed values are enable and disable
3762 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3763 'node', 'default' can be specified
3764 This can be set from sysctl after boot.
3765 See Documentation/admin-guide/sysctl/vm.rst for details.
3767 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3768 See Documentation/core-api/debugging-via-ohci1394.rst for more
3771 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3772 Rather than timing out after 20 ms if an EC
3773 command is not properly ACKed, override the length
3774 of the timeout. We have interrupts disabled while
3775 waiting for the ACK, so if this is set too high
3776 interrupts *may* be lost!
3778 omap_mux= [OMAP] Override bootloader pin multiplexing.
3779 Format: <mux_mode0.mode_name=value>...
3780 For example, to override I2C bus2:
3781 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3783 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3785 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3787 boundary - index of last SLC block on Flex-OneNAND.
3788 The remaining blocks are configured as MLC blocks.
3789 lock - Configure if Flex-OneNAND boundary should be locked.
3790 Once locked, the boundary cannot be changed.
3791 1 indicates lock status, 0 indicates unlock status.
3793 oops=panic Always panic on oopses. Default is to just kill the
3794 process, but there is a small probability of
3795 deadlocking the machine.
3796 This will also cause panics on machine check exceptions.
3797 Useful together with panic=30 to trigger a reboot.
3800 [KNL] Boolean flag to control whether the page allocator
3801 should randomize its free lists. The randomization may
3802 be automatically enabled if the kernel detects it is
3803 running on a platform with a direct-mapped memory-side
3804 cache, and this parameter can be used to
3805 override/disable that behavior. The state of the flag
3806 can be read from sysfs at:
3807 /sys/module/page_alloc/parameters/shuffle.
3809 page_owner= [KNL] Boot-time page_owner enabling option.
3810 Storage of the information about who allocated
3811 each page is disabled in default. With this switch,
3813 on: enable the feature
3815 page_poison= [KNL] Boot-time parameter changing the state of
3816 poisoning on the buddy allocator, available with
3817 CONFIG_PAGE_POISONING=y.
3818 off: turn off poisoning (default)
3819 on: turn on poisoning
3821 page_reporting.page_reporting_order=
3822 [KNL] Minimal page reporting order
3824 Adjust the minimal page reporting order. The page
3825 reporting is disabled when it exceeds (MAX_ORDER-1).
3827 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3828 timeout > 0: seconds before rebooting
3829 timeout = 0: wait forever
3830 timeout < 0: reboot immediately
3833 panic_print= Bitmask for printing system info when panic happens.
3834 User can chose combination of the following bits:
3835 bit 0: print all tasks info
3836 bit 1: print system memory info
3837 bit 2: print timer info
3838 bit 3: print locks info if CONFIG_LOCKDEP is on
3839 bit 4: print ftrace buffer
3840 bit 5: print all printk messages in buffer
3841 bit 6: print all CPUs backtrace (if available in the arch)
3842 *Be aware* that this option may print a _lot_ of lines,
3843 so there are risks of losing older messages in the log.
3844 Use this option carefully, maybe worth to setup a
3845 bigger log buffer with "log_buf_len" along with this.
3847 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3848 Format: <hex>[,nousertaint]
3849 Hexadecimal bitmask representing the set of TAINT flags
3850 that will cause the kernel to panic when add_taint() is
3851 called with any of the flags in this set.
3852 The optional switch "nousertaint" can be utilized to
3853 prevent userspace forced crashes by writing to sysctl
3854 /proc/sys/kernel/tainted any flagset matching with the
3855 bitmask set on panic_on_taint.
3856 See Documentation/admin-guide/tainted-kernels.rst for
3857 extra details on the taint flags that users can pick
3858 to compose the bitmask to assign to panic_on_taint.
3860 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3863 parkbd.port= [HW] Parallel port number the keyboard adapter is
3864 connected to, default is 0.
3866 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3867 0 for XT, 1 for AT (default is AT).
3870 parport= [HW,PPT] Specify parallel ports. 0 disables.
3871 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3872 Use 'auto' to force the driver to use any
3873 IRQ/DMA settings detected (the default is to
3874 ignore detected IRQ/DMA settings because of
3875 possible conflicts). You can specify the base
3876 address, IRQ, and DMA settings; IRQ and DMA
3877 should be numbers, or 'auto' (for using detected
3878 settings on that particular port), or 'nofifo'
3879 (to avoid using a FIFO even if it is detected).
3880 Parallel ports are assigned in the order they
3881 are specified on the command line, starting
3884 parport_init_mode= [HW,PPT]
3885 Configure VIA parallel port to operate in
3886 a specific mode. This is necessary on Pegasos
3887 computer where firmware has no options for setting
3888 up parallel port mode and sets it to spp.
3889 Currently this function knows 686a and 8231 chips.
3890 Format: [spp|ps2|epp|ecp|ecpepp]
3892 pata_legacy.all= [HW,LIBATA]
3894 Set to non-zero to probe primary and secondary ISA
3895 port ranges on PCI systems where no PCI PATA device
3896 has been found at either range. Disabled by default.
3898 pata_legacy.autospeed= [HW,LIBATA]
3900 Set to non-zero if a chip is present that snoops speed
3901 changes. Disabled by default.
3903 pata_legacy.ht6560a= [HW,LIBATA]
3905 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3906 the secondary channel, or both channels respectively.
3907 Disabled by default.
3909 pata_legacy.ht6560b= [HW,LIBATA]
3911 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3912 the secondary channel, or both channels respectively.
3913 Disabled by default.
3915 pata_legacy.iordy_mask= [HW,LIBATA]
3917 IORDY enable mask. Set individual bits to allow IORDY
3918 for the respective channel. Bit 0 is for the first
3919 legacy channel handled by this driver, bit 1 is for
3920 the second channel, and so on. The sequence will often
3921 correspond to the primary legacy channel, the secondary
3922 legacy channel, and so on, but the handling of a PCI
3923 bus and the use of other driver options may interfere
3924 with the sequence. By default IORDY is allowed across
3927 pata_legacy.opti82c46x= [HW,LIBATA]
3929 Set to 1, 2, or 3 for Opti 82c611A on the primary
3930 channel, the secondary channel, or both channels
3931 respectively. Disabled by default.
3933 pata_legacy.opti82c611a= [HW,LIBATA]
3935 Set to 1, 2, or 3 for Opti 82c465MV on the primary
3936 channel, the secondary channel, or both channels
3937 respectively. Disabled by default.
3939 pata_legacy.pio_mask= [HW,LIBATA]
3941 PIO mode mask for autospeed devices. Set individual
3942 bits to allow the use of the respective PIO modes.
3943 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
3944 All modes allowed by default.
3946 pata_legacy.probe_all= [HW,LIBATA]
3948 Set to non-zero to probe tertiary and further ISA
3949 port ranges on PCI systems. Disabled by default.
3951 pata_legacy.probe_mask= [HW,LIBATA]
3953 Probe mask for legacy ISA PATA ports. Depending on
3954 platform configuration and the use of other driver
3955 options up to 6 legacy ports are supported: 0x1f0,
3956 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
3957 of individual ports can be disabled by setting the
3958 corresponding bits in the mask to 1. Bit 0 is for
3959 the first port in the list above (0x1f0), and so on.
3960 By default all supported ports are probed.
3962 pata_legacy.qdi= [HW,LIBATA]
3964 Set to non-zero to probe QDI controllers. By default
3965 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
3967 pata_legacy.winbond= [HW,LIBATA]
3969 Set to non-zero to probe Winbond controllers. Use
3970 the standard I/O port (0x130) if 1, otherwise the
3971 value given is the I/O port to use (typically 0x1b0).
3972 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
3975 pata_platform.pio_mask= [HW,LIBATA]
3977 Supported PIO mode mask. Set individual bits to allow
3978 the use of the respective PIO modes. Bit 0 is for
3979 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
3983 Halt all CPUs after the first oops has been printed for
3984 the specified number of seconds. This is to be used if
3985 your oopses keep scrolling off the screen.
3990 See header of drivers/block/paride/pcd.c.
3991 See also Documentation/admin-guide/blockdev/paride.rst.
3993 pci=option[,option...] [PCI] various PCI subsystem options.
3995 Some options herein operate on a specific device
3996 or a set of devices (<pci_dev>). These are
3997 specified in one of the following formats:
3999 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4000 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4002 Note: the first format specifies a PCI
4003 bus/device/function address which may change
4004 if new hardware is inserted, if motherboard
4005 firmware changes, or due to changes caused
4006 by other kernel parameters. If the
4007 domain is left unspecified, it is
4008 taken to be zero. Optionally, a path
4009 to a device through multiple device/function
4010 addresses can be specified after the base
4011 address (this is more robust against
4012 renumbering issues). The second format
4013 selects devices using IDs from the
4014 configuration space which may match multiple
4015 devices in the system.
4017 earlydump dump PCI config space before the kernel
4019 off [X86] don't probe for the PCI bus
4020 bios [X86-32] force use of PCI BIOS, don't access
4021 the hardware directly. Use this if your machine
4022 has a non-standard PCI host bridge.
4023 nobios [X86-32] disallow use of PCI BIOS, only direct
4024 hardware access methods are allowed. Use this
4025 if you experience crashes upon bootup and you
4026 suspect they are caused by the BIOS.
4027 conf1 [X86] Force use of PCI Configuration Access
4028 Mechanism 1 (config address in IO port 0xCF8,
4029 data in IO port 0xCFC, both 32-bit).
4030 conf2 [X86] Force use of PCI Configuration Access
4031 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4032 the function, IO port 0xCFA, also 8-bit, sets
4033 bus number. The config space is then accessed
4034 through ports 0xC000-0xCFFF).
4035 See http://wiki.osdev.org/PCI for more info
4036 on the configuration access mechanisms.
4037 noaer [PCIE] If the PCIEAER kernel config parameter is
4038 enabled, this kernel boot option can be used to
4039 disable the use of PCIE advanced error reporting.
4040 nodomains [PCI] Disable support for multiple PCI
4041 root domains (aka PCI segments, in ACPI-speak).
4042 nommconf [X86] Disable use of MMCONFIG for PCI
4044 check_enable_amd_mmconf [X86] check for and enable
4045 properly configured MMIO access to PCI
4046 config space on AMD family 10h CPU
4047 nomsi [MSI] If the PCI_MSI kernel config parameter is
4048 enabled, this kernel boot option can be used to
4049 disable the use of MSI interrupts system-wide.
4050 noioapicquirk [APIC] Disable all boot interrupt quirks.
4051 Safety option to keep boot IRQs enabled. This
4052 should never be necessary.
4053 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4054 primary IO-APIC for bridges that cannot disable
4055 boot IRQs. This fixes a source of spurious IRQs
4056 when the system masks IRQs.
4057 noioapicreroute [APIC] Disable workaround that uses the
4058 boot IRQ equivalent of an IRQ that connects to
4059 a chipset where boot IRQs cannot be disabled.
4060 The opposite of ioapicreroute.
4061 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4062 routing table. These calls are known to be buggy
4063 on several machines and they hang the machine
4064 when used, but on other computers it's the only
4065 way to get the interrupt routing table. Try
4066 this option if the kernel is unable to allocate
4067 IRQs or discover secondary PCI buses on your
4069 rom [X86] Assign address space to expansion ROMs.
4070 Use with caution as certain devices share
4071 address decoders between ROMs and other
4073 norom [X86] Do not assign address space to
4074 expansion ROMs that do not already have
4075 BIOS assigned address ranges.
4076 nobar [X86] Do not assign address space to the
4077 BARs that weren't assigned by the BIOS.
4078 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4079 assigned automatically to PCI devices. You can
4080 make the kernel exclude IRQs of your ISA cards
4082 pirqaddr=0xAAAAA [X86] Specify the physical address
4083 of the PIRQ table (normally generated
4084 by the BIOS) if it is outside the
4085 F0000h-100000h range.
4086 lastbus=N [X86] Scan all buses thru bus #N. Can be
4087 useful if the kernel is unable to find your
4088 secondary buses and you want to tell it
4089 explicitly which ones they are.
4090 assign-busses [X86] Always assign all PCI bus
4091 numbers ourselves, overriding
4092 whatever the firmware may have done.
4093 usepirqmask [X86] Honor the possible IRQ mask stored
4094 in the BIOS $PIR table. This is needed on
4095 some systems with broken BIOSes, notably
4096 some HP Pavilion N5400 and Omnibook XE3
4097 notebooks. This will have no effect if ACPI
4098 IRQ routing is enabled.
4099 noacpi [X86] Do not use ACPI for IRQ routing
4100 or for PCI scanning.
4101 use_crs [X86] Use PCI host bridge window information
4102 from ACPI. On BIOSes from 2008 or later, this
4103 is enabled by default. If you need to use this,
4104 please report a bug.
4105 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4106 If you need to use this, please report a bug.
4107 use_e820 [X86] Use E820 reservations to exclude parts of
4108 PCI host bridge windows. This is a workaround
4109 for BIOS defects in host bridge _CRS methods.
4110 If you need to use this, please report a bug to
4111 <linux-pci@vger.kernel.org>.
4112 no_e820 [X86] Ignore E820 reservations for PCI host
4113 bridge windows. This is the default on modern
4114 hardware. If you need to use this, please report
4115 a bug to <linux-pci@vger.kernel.org>.
4116 routeirq Do IRQ routing for all PCI devices.
4117 This is normally done in pci_enable_device(),
4118 so this option is a temporary workaround
4119 for broken drivers that don't call it.
4120 skip_isa_align [X86] do not align io start addr, so can
4121 handle more pci cards
4122 noearly [X86] Don't do any early type 1 scanning.
4123 This might help on some broken boards which
4124 machine check when some devices' config space
4125 is read. But various workarounds are disabled
4126 and some IOMMU drivers will not work.
4127 bfsort Sort PCI devices into breadth-first order.
4128 This sorting is done to get a device
4129 order compatible with older (<= 2.4) kernels.
4130 nobfsort Don't sort PCI devices into breadth-first order.
4131 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4132 tuning and use the BIOS-configured MPS defaults.
4133 pcie_bus_safe Set every device's MPS to the largest value
4134 supported by all devices below the root complex.
4135 pcie_bus_perf Set device MPS to the largest allowable MPS
4136 based on its parent bus. Also set MRRS (Max
4137 Read Request Size) to the largest supported
4138 value (no larger than the MPS that the device
4139 or bus can support) for best performance.
4140 pcie_bus_peer2peer Set every device's MPS to 128B, which
4141 every device is guaranteed to support. This
4142 configuration allows peer-to-peer DMA between
4143 any pair of devices, possibly at the cost of
4144 reduced performance. This also guarantees
4145 that hot-added devices will work.
4146 cbiosize=nn[KMG] The fixed amount of bus space which is
4147 reserved for the CardBus bridge's IO window.
4148 The default value is 256 bytes.
4149 cbmemsize=nn[KMG] The fixed amount of bus space which is
4150 reserved for the CardBus bridge's memory
4151 window. The default value is 64 megabytes.
4154 [<order of align>@]<pci_dev>[; ...]
4155 Specifies alignment and device to reassign
4156 aligned memory resources. How to
4157 specify the device is described above.
4158 If <order of align> is not specified,
4159 PAGE_SIZE is used as alignment.
4160 A PCI-PCI bridge can be specified if resource
4161 windows need to be expanded.
4162 To specify the alignment for several
4163 instances of a device, the PCI vendor,
4164 device, subvendor, and subdevice may be
4165 specified, e.g., 12@pci:8086:9c22:103c:198f
4166 for 4096-byte alignment.
4167 ecrc= Enable/disable PCIe ECRC (transaction layer
4168 end-to-end CRC checking).
4169 bios: Use BIOS/firmware settings. This is the
4173 hpiosize=nn[KMG] The fixed amount of bus space which is
4174 reserved for hotplug bridge's IO window.
4175 Default size is 256 bytes.
4176 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4177 reserved for hotplug bridge's MMIO window.
4178 Default size is 2 megabytes.
4179 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4180 reserved for hotplug bridge's MMIO_PREF window.
4181 Default size is 2 megabytes.
4182 hpmemsize=nn[KMG] The fixed amount of bus space which is
4183 reserved for hotplug bridge's MMIO and
4185 Default size is 2 megabytes.
4186 hpbussize=nn The minimum amount of additional bus numbers
4187 reserved for buses below a hotplug bridge.
4189 realloc= Enable/disable reallocating PCI bridge resources
4190 if allocations done by BIOS are too small to
4191 accommodate resources required by all child
4193 off: Turn realloc off
4195 realloc same as realloc=on
4196 noari do not use PCIe ARI.
4197 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4198 do not use PCIe ATS (and IOMMU device IOTLB).
4199 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4200 only look for one device below a PCIe downstream
4202 big_root_window Try to add a big 64bit memory window to the PCIe
4203 root complex on AMD CPUs. Some GFX hardware
4204 can resize a BAR to allow access to all VRAM.
4205 Adding the window is slightly risky (it may
4206 conflict with unreported devices), so this
4208 disable_acs_redir=<pci_dev>[; ...]
4209 Specify one or more PCI devices (in the format
4210 specified above) separated by semicolons.
4211 Each device specified will have the PCI ACS
4212 redirect capabilities forced off which will
4213 allow P2P traffic between devices through
4214 bridges without forcing it upstream. Note:
4215 this removes isolation between devices and
4216 may put more devices in an IOMMU group.
4217 force_floating [S390] Force usage of floating interrupts.
4218 nomio [S390] Do not use MIO instructions.
4219 norid [S390] ignore the RID field and force use of
4220 one PCI domain per PCI function
4222 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4225 force Enable ASPM even on devices that claim not to support it.
4226 WARNING: Forcing ASPM on may cause system lockups.
4228 pcie_ports= [PCIE] PCIe port services handling:
4229 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4230 even if the platform doesn't give the OS permission to
4231 use them. This may cause conflicts if the platform
4232 also tries to use these services.
4233 dpc-native Use native PCIe service for DPC only. May
4234 cause conflicts if firmware uses AER or DPC.
4235 compat Disable native PCIe services (PME, AER, DPC, PCIe
4238 pcie_port_pm= [PCIE] PCIe port power management handling:
4239 off Disable power management of all PCIe ports
4240 force Forcibly enable power management of all PCIe ports
4242 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4243 nomsi Do not use MSI for native PCIe PME signaling (this makes
4244 all PCIe root ports use INTx for all services).
4246 pcmv= [HW,PCMCIA] BadgePAD 4
4250 Keep all power-domains already enabled by bootloader on,
4251 even if no driver has claimed them. This is useful
4252 for debug and development, but should not be
4253 needed on a platform with proper driver support.
4256 See Documentation/admin-guide/blockdev/paride.rst.
4258 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4261 See arch/parisc/kernel/pdc_chassis.c
4263 percpu_alloc= Select which percpu first chunk allocator to use.
4264 Currently supported values are "embed" and "page".
4265 Archs may support subset or none of the selections.
4266 See comments in mm/percpu.c for details on each
4267 allocator. This parameter is primarily for debugging
4268 and performance comparison.
4271 See Documentation/admin-guide/blockdev/paride.rst.
4274 See Documentation/admin-guide/blockdev/paride.rst.
4276 pirq= [SMP,APIC] Manual mp-table setup
4277 See Documentation/x86/i386/IO-APIC.rst.
4279 plip= [PPT,NET] Parallel port network link
4280 Format: { parport<nr> | timid | 0 }
4281 See also Documentation/admin-guide/parport.rst.
4283 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4284 Override pmtimer IOPort with a hex value.
4287 pmu_override= [PPC] Override the PMU.
4288 This option takes over the PMU facility, so it is no
4289 longer usable by perf. Setting this option starts the
4290 PMU counters by setting MMCR0 to 0 (the FC bit is
4291 cleared). If a number is given, then MMCR1 is set to
4292 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4295 pm_debug_messages [SUSPEND,KNL]
4296 Enable suspend/resume debug messages during boot up.
4299 Enable PNP debug messages (depends on the
4300 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4301 via /sys/module/pnp/parameters/debug. We always show
4302 current resource usage; turning this on also shows
4303 possible settings and some assignment information.
4309 { on | off | curr | res | no-curr | no-res }
4312 [ISAPNP] Exclude IRQs for the autoconfiguration
4315 [ISAPNP] Exclude DMAs for the autoconfiguration
4317 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4318 Ranges are in pairs (I/O port base and size).
4321 [ISAPNP] Exclude memory regions for the
4323 Ranges are in pairs (memory base and size).
4325 ports= [IP_VS_FTP] IPVS ftp helper module
4327 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4329 Format: <port>,<port>....
4331 powersave=off [PPC] This option disables power saving features.
4332 It specifically disables cpuidle and sets the
4333 platform machine description specific power_save
4334 function to NULL. On Idle the CPU just reduces
4337 ppc_strict_facility_enable
4338 [PPC] This option catches any kernel floating point,
4339 Altivec, VSX and SPE outside of regions specifically
4340 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4341 There is some performance impact when enabling this.
4345 Disable Hardware Transactional Memory
4348 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4349 none - Limited to cond_resched() calls
4350 voluntary - Limited to cond_resched() and might_sleep() calls
4351 full - Any section that isn't explicitly preempt disabled
4352 can be preempted anytime.
4354 print-fatal-signals=
4355 [KNL] debug: print fatal signals
4357 If enabled, warn about various signal handling
4358 related application anomalies: too many signals,
4359 too many POSIX.1 timers, fatal signals causing a
4362 If you hit the warning due to signal overflow,
4363 you might want to try "ulimit -i unlimited".
4367 printk.always_kmsg_dump=
4368 Trigger kmsg_dump for cases other than kernel oops or
4370 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4373 printk.console_no_auto_verbose=
4374 Disable console loglevel raise on oops, panic
4375 or lockdep-detected issues (only if lock debug is on).
4376 With an exception to setups with low baudrate on
4377 serial console, keeping this 0 is a good choice
4378 in order to provide more debug information.
4380 default: 0 (auto_verbose is enabled)
4382 printk.devkmsg={on,off,ratelimit}
4383 Control writing to /dev/kmsg.
4384 on - unlimited logging to /dev/kmsg from userspace
4385 off - logging to /dev/kmsg disabled
4386 ratelimit - ratelimit the logging
4389 printk.time= Show timing data prefixed to each printk message line
4390 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4392 processor.max_cstate= [HW,ACPI]
4393 Limit processor to maximum C-state
4394 max_cstate=9 overrides any DMI blacklist limit.
4396 processor.nocst [HW,ACPI]
4397 Ignore the _CST method to determine C-states,
4398 instead using the legacy FADT method
4400 profile= [KNL] Enable kernel profiling via /proc/profile
4401 Format: [<profiletype>,]<number>
4402 Param: <profiletype>: "schedule", "sleep", or "kvm"
4403 [defaults to kernel profiling]
4404 Param: "schedule" - profile schedule points.
4405 Param: "sleep" - profile D-state sleeping (millisecs).
4406 Requires CONFIG_SCHEDSTATS
4407 Param: "kvm" - profile VM exits.
4408 Param: <number> - step/bucket size as a power of 2 for
4409 statistical time based profiling.
4411 prompt_ramdisk= [RAM] [Deprecated]
4413 prot_virt= [S390] enable hosting protected virtual machines
4414 isolated from the hypervisor (if hardware supports
4418 psi= [KNL] Enable or disable pressure stall information
4422 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4423 probe for; one of (bare|imps|exps|lifebook|any).
4424 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4426 psmouse.resetafter= [HW,MOUSE]
4427 Try to reset the device after so many bad packets
4430 [HW,MOUSE] Set desired mouse resolution, in dpi.
4431 psmouse.smartscroll=
4432 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4433 0 = disabled, 1 = enabled (default).
4435 pstore.backend= Specify the name of the pstore backend to use
4438 See Documentation/admin-guide/blockdev/paride.rst.
4440 pti= [X86-64] Control Page Table Isolation of user and
4441 kernel address spaces. Disabling this feature
4442 removes hardening, but improves performance of
4443 system calls and interrupts.
4445 on - unconditionally enable
4446 off - unconditionally disable
4447 auto - kernel detects whether your CPU model is
4448 vulnerable to issues that PTI mitigates
4450 Not specifying this option is equivalent to pti=auto.
4453 Equivalent to pti=off
4456 [KNL] Number of legacy pty's. Overwrites compiled-in
4459 quiet [KNL] Disable most log messages
4464 See Documentation/admin-guide/md.rst.
4466 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4467 See Documentation/admin-guide/blockdev/ramdisk.rst.
4469 ramdisk_start= [RAM] RAM disk image start address
4471 random.trust_cpu={on,off}
4472 [KNL] Enable or disable trusting the use of the
4473 CPU's random number generator (if available) to
4474 fully seed the kernel's CRNG. Default is controlled
4475 by CONFIG_RANDOM_TRUST_CPU.
4477 random.trust_bootloader={on,off}
4478 [KNL] Enable or disable trusting the use of a
4479 seed passed by the bootloader (if available) to
4480 fully seed the kernel's CRNG. Default is controlled
4481 by CONFIG_RANDOM_TRUST_BOOTLOADER.
4483 randomize_kstack_offset=
4484 [KNL] Enable or disable kernel stack offset
4485 randomization, which provides roughly 5 bits of
4486 entropy, frustrating memory corruption attacks
4487 that depend on stack address determinism or
4488 cross-syscall address exposures. This is only
4489 available on architectures that have defined
4490 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4491 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4492 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4494 ras=option[,option,...] [KNL] RAS-specific options
4497 Disable the Correctable Errors Collector,
4498 see CONFIG_RAS_CEC help text.
4500 rcu_nocbs[=cpu-list]
4501 [KNL] The optional argument is a cpu list,
4504 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4505 enable the no-callback CPU mode, which prevents
4506 such CPUs' callbacks from being invoked in
4507 softirq context. Invocation of such CPUs' RCU
4508 callbacks will instead be offloaded to "rcuox/N"
4509 kthreads created for that purpose, where "x" is
4510 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4511 for the kthreads that mediate grace periods; and
4512 "N" is the CPU number. This reduces OS jitter on
4513 the offloaded CPUs, which can be useful for HPC
4514 and real-time workloads. It can also improve
4515 energy efficiency for asymmetric multiprocessors.
4517 If a cpulist is passed as an argument, the specified
4518 list of CPUs is set to no-callback mode from boot.
4520 Otherwise, if the '=' sign and the cpulist
4521 arguments are omitted, no CPU will be set to
4522 no-callback mode from boot but the mode may be
4523 toggled at runtime via cpusets.
4526 Rather than requiring that offloaded CPUs
4527 (specified by rcu_nocbs= above) explicitly
4528 awaken the corresponding "rcuoN" kthreads,
4529 make these kthreads poll for callbacks.
4530 This improves the real-time response for the
4531 offloaded CPUs by relieving them of the need to
4532 wake up the corresponding kthread, but degrades
4533 energy efficiency by requiring that the kthreads
4534 periodically wake up to do the polling.
4536 rcutree.blimit= [KNL]
4537 Set maximum number of finished RCU callbacks to
4538 process in one batch.
4540 rcutree.dump_tree= [KNL]
4541 Dump the structure of the rcu_node combining tree
4542 out at early boot. This is used for diagnostic
4543 purposes, to verify correct tree setup.
4545 rcutree.gp_cleanup_delay= [KNL]
4546 Set the number of jiffies to delay each step of
4547 RCU grace-period cleanup.
4549 rcutree.gp_init_delay= [KNL]
4550 Set the number of jiffies to delay each step of
4551 RCU grace-period initialization.
4553 rcutree.gp_preinit_delay= [KNL]
4554 Set the number of jiffies to delay each step of
4555 RCU grace-period pre-initialization, that is,
4556 the propagation of recent CPU-hotplug changes up
4557 the rcu_node combining tree.
4559 rcutree.use_softirq= [KNL]
4560 If set to zero, move all RCU_SOFTIRQ processing to
4561 per-CPU rcuc kthreads. Defaults to a non-zero
4562 value, meaning that RCU_SOFTIRQ is used by default.
4563 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4565 But note that CONFIG_PREEMPT_RT=y kernels disable
4566 this kernel boot parameter, forcibly setting it
4569 rcutree.rcu_fanout_exact= [KNL]
4570 Disable autobalancing of the rcu_node combining
4571 tree. This is used by rcutorture, and might
4572 possibly be useful for architectures having high
4573 cache-to-cache transfer latencies.
4575 rcutree.rcu_fanout_leaf= [KNL]
4576 Change the number of CPUs assigned to each
4577 leaf rcu_node structure. Useful for very
4578 large systems, which will choose the value 64,
4579 and for NUMA systems with large remote-access
4580 latencies, which will choose a value aligned
4581 with the appropriate hardware boundaries.
4583 rcutree.rcu_min_cached_objs= [KNL]
4584 Minimum number of objects which are cached and
4585 maintained per one CPU. Object size is equal
4586 to PAGE_SIZE. The cache allows to reduce the
4587 pressure to page allocator, also it makes the
4588 whole algorithm to behave better in low memory
4591 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4592 Set the page-cache refill delay (in milliseconds)
4593 in response to low-memory conditions. The range
4594 of permitted values is in the range 0:100000.
4596 rcutree.jiffies_till_first_fqs= [KNL]
4597 Set delay from grace-period initialization to
4598 first attempt to force quiescent states.
4599 Units are jiffies, minimum value is zero,
4600 and maximum value is HZ.
4602 rcutree.jiffies_till_next_fqs= [KNL]
4603 Set delay between subsequent attempts to force
4604 quiescent states. Units are jiffies, minimum
4605 value is one, and maximum value is HZ.
4607 rcutree.jiffies_till_sched_qs= [KNL]
4608 Set required age in jiffies for a
4609 given grace period before RCU starts
4610 soliciting quiescent-state help from
4611 rcu_note_context_switch() and cond_resched().
4612 If not specified, the kernel will calculate
4613 a value based on the most recent settings
4614 of rcutree.jiffies_till_first_fqs
4615 and rcutree.jiffies_till_next_fqs.
4616 This calculated value may be viewed in
4617 rcutree.jiffies_to_sched_qs. Any attempt to set
4618 rcutree.jiffies_to_sched_qs will be cheerfully
4621 rcutree.kthread_prio= [KNL,BOOT]
4622 Set the SCHED_FIFO priority of the RCU per-CPU
4623 kthreads (rcuc/N). This value is also used for
4624 the priority of the RCU boost threads (rcub/N)
4625 and for the RCU grace-period kthreads (rcu_bh,
4626 rcu_preempt, and rcu_sched). If RCU_BOOST is
4627 set, valid values are 1-99 and the default is 1
4628 (the least-favored priority). Otherwise, when
4629 RCU_BOOST is not set, valid values are 0-99 and
4630 the default is zero (non-realtime operation).
4631 When RCU_NOCB_CPU is set, also adjust the
4632 priority of NOCB callback kthreads.
4634 rcutree.rcu_nocb_gp_stride= [KNL]
4635 Set the number of NOCB callback kthreads in
4636 each group, which defaults to the square root
4637 of the number of CPUs. Larger numbers reduce
4638 the wakeup overhead on the global grace-period
4639 kthread, but increases that same overhead on
4640 each group's NOCB grace-period kthread.
4642 rcutree.qhimark= [KNL]
4643 Set threshold of queued RCU callbacks beyond which
4644 batch limiting is disabled.
4646 rcutree.qlowmark= [KNL]
4647 Set threshold of queued RCU callbacks below which
4648 batch limiting is re-enabled.
4650 rcutree.qovld= [KNL]
4651 Set threshold of queued RCU callbacks beyond which
4652 RCU's force-quiescent-state scan will aggressively
4653 enlist help from cond_resched() and sched IPIs to
4654 help CPUs more quickly reach quiescent states.
4655 Set to less than zero to make this be set based
4656 on rcutree.qhimark at boot time and to zero to
4657 disable more aggressive help enlistment.
4659 rcutree.rcu_kick_kthreads= [KNL]
4660 Cause the grace-period kthread to get an extra
4661 wake_up() if it sleeps three times longer than
4662 it should at force-quiescent-state time.
4663 This wake_up() will be accompanied by a
4664 WARN_ONCE() splat and an ftrace_dump().
4666 rcutree.rcu_unlock_delay= [KNL]
4667 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4668 this specifies an rcu_read_unlock()-time delay
4669 in microseconds. This defaults to zero.
4670 Larger delays increase the probability of
4671 catching RCU pointer leaks, that is, buggy use
4672 of RCU-protected pointers after the relevant
4673 rcu_read_unlock() has completed.
4675 rcutree.sysrq_rcu= [KNL]
4676 Commandeer a sysrq key to dump out Tree RCU's
4677 rcu_node tree with an eye towards determining
4678 why a new grace period has not yet started.
4680 rcuscale.gp_async= [KNL]
4681 Measure performance of asynchronous
4682 grace-period primitives such as call_rcu().
4684 rcuscale.gp_async_max= [KNL]
4685 Specify the maximum number of outstanding
4686 callbacks per writer thread. When a writer
4687 thread exceeds this limit, it invokes the
4688 corresponding flavor of rcu_barrier() to allow
4689 previously posted callbacks to drain.
4691 rcuscale.gp_exp= [KNL]
4692 Measure performance of expedited synchronous
4693 grace-period primitives.
4695 rcuscale.holdoff= [KNL]
4696 Set test-start holdoff period. The purpose of
4697 this parameter is to delay the start of the
4698 test until boot completes in order to avoid
4701 rcuscale.kfree_rcu_test= [KNL]
4702 Set to measure performance of kfree_rcu() flooding.
4704 rcuscale.kfree_rcu_test_double= [KNL]
4705 Test the double-argument variant of kfree_rcu().
4706 If this parameter has the same value as
4707 rcuscale.kfree_rcu_test_single, both the single-
4708 and double-argument variants are tested.
4710 rcuscale.kfree_rcu_test_single= [KNL]
4711 Test the single-argument variant of kfree_rcu().
4712 If this parameter has the same value as
4713 rcuscale.kfree_rcu_test_double, both the single-
4714 and double-argument variants are tested.
4716 rcuscale.kfree_nthreads= [KNL]
4717 The number of threads running loops of kfree_rcu().
4719 rcuscale.kfree_alloc_num= [KNL]
4720 Number of allocations and frees done in an iteration.
4722 rcuscale.kfree_loops= [KNL]
4723 Number of loops doing rcuscale.kfree_alloc_num number
4724 of allocations and frees.
4726 rcuscale.nreaders= [KNL]
4727 Set number of RCU readers. The value -1 selects
4728 N, where N is the number of CPUs. A value
4729 "n" less than -1 selects N-n+1, where N is again
4730 the number of CPUs. For example, -2 selects N
4731 (the number of CPUs), -3 selects N+1, and so on.
4732 A value of "n" less than or equal to -N selects
4735 rcuscale.nwriters= [KNL]
4736 Set number of RCU writers. The values operate
4737 the same as for rcuscale.nreaders.
4738 N, where N is the number of CPUs
4740 rcuscale.perf_type= [KNL]
4741 Specify the RCU implementation to test.
4743 rcuscale.shutdown= [KNL]
4744 Shut the system down after performance tests
4745 complete. This is useful for hands-off automated
4748 rcuscale.verbose= [KNL]
4749 Enable additional printk() statements.
4751 rcuscale.writer_holdoff= [KNL]
4752 Write-side holdoff between grace periods,
4753 in microseconds. The default of zero says
4756 rcutorture.fqs_duration= [KNL]
4757 Set duration of force_quiescent_state bursts
4760 rcutorture.fqs_holdoff= [KNL]
4761 Set holdoff time within force_quiescent_state bursts
4764 rcutorture.fqs_stutter= [KNL]
4765 Set wait time between force_quiescent_state bursts
4768 rcutorture.fwd_progress= [KNL]
4769 Specifies the number of kthreads to be used
4770 for RCU grace-period forward-progress testing
4771 for the types of RCU supporting this notion.
4772 Defaults to 1 kthread, values less than zero or
4773 greater than the number of CPUs cause the number
4776 rcutorture.fwd_progress_div= [KNL]
4777 Specify the fraction of a CPU-stall-warning
4778 period to do tight-loop forward-progress testing.
4780 rcutorture.fwd_progress_holdoff= [KNL]
4781 Number of seconds to wait between successive
4782 forward-progress tests.
4784 rcutorture.fwd_progress_need_resched= [KNL]
4785 Enclose cond_resched() calls within checks for
4786 need_resched() during tight-loop forward-progress
4789 rcutorture.gp_cond= [KNL]
4790 Use conditional/asynchronous update-side
4791 primitives, if available.
4793 rcutorture.gp_exp= [KNL]
4794 Use expedited update-side primitives, if available.
4796 rcutorture.gp_normal= [KNL]
4797 Use normal (non-expedited) asynchronous
4798 update-side primitives, if available.
4800 rcutorture.gp_sync= [KNL]
4801 Use normal (non-expedited) synchronous
4802 update-side primitives, if available. If all
4803 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4804 rcutorture.gp_normal=, and rcutorture.gp_sync=
4805 are zero, rcutorture acts as if is interpreted
4806 they are all non-zero.
4808 rcutorture.irqreader= [KNL]
4809 Run RCU readers from irq handlers, or, more
4810 accurately, from a timer handler. Not all RCU
4811 flavors take kindly to this sort of thing.
4813 rcutorture.leakpointer= [KNL]
4814 Leak an RCU-protected pointer out of the reader.
4815 This can of course result in splats, and is
4816 intended to test the ability of things like
4817 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4820 rcutorture.n_barrier_cbs= [KNL]
4821 Set callbacks/threads for rcu_barrier() testing.
4823 rcutorture.nfakewriters= [KNL]
4824 Set number of concurrent RCU writers. These just
4825 stress RCU, they don't participate in the actual
4826 test, hence the "fake".
4828 rcutorture.nocbs_nthreads= [KNL]
4829 Set number of RCU callback-offload togglers.
4830 Zero (the default) disables toggling.
4832 rcutorture.nocbs_toggle= [KNL]
4833 Set the delay in milliseconds between successive
4834 callback-offload toggling attempts.
4836 rcutorture.nreaders= [KNL]
4837 Set number of RCU readers. The value -1 selects
4838 N-1, where N is the number of CPUs. A value
4839 "n" less than -1 selects N-n-2, where N is again
4840 the number of CPUs. For example, -2 selects N
4841 (the number of CPUs), -3 selects N+1, and so on.
4843 rcutorture.object_debug= [KNL]
4844 Enable debug-object double-call_rcu() testing.
4846 rcutorture.onoff_holdoff= [KNL]
4847 Set time (s) after boot for CPU-hotplug testing.
4849 rcutorture.onoff_interval= [KNL]
4850 Set time (jiffies) between CPU-hotplug operations,
4851 or zero to disable CPU-hotplug testing.
4853 rcutorture.read_exit= [KNL]
4854 Set the number of read-then-exit kthreads used
4855 to test the interaction of RCU updaters and
4856 task-exit processing.
4858 rcutorture.read_exit_burst= [KNL]
4859 The number of times in a given read-then-exit
4860 episode that a set of read-then-exit kthreads
4863 rcutorture.read_exit_delay= [KNL]
4864 The delay, in seconds, between successive
4865 read-then-exit testing episodes.
4867 rcutorture.shuffle_interval= [KNL]
4868 Set task-shuffle interval (s). Shuffling tasks
4869 allows some CPUs to go into dyntick-idle mode
4870 during the rcutorture test.
4872 rcutorture.shutdown_secs= [KNL]
4873 Set time (s) after boot system shutdown. This
4874 is useful for hands-off automated testing.
4876 rcutorture.stall_cpu= [KNL]
4877 Duration of CPU stall (s) to test RCU CPU stall
4878 warnings, zero to disable.
4880 rcutorture.stall_cpu_block= [KNL]
4881 Sleep while stalling if set. This will result
4882 in warnings from preemptible RCU in addition
4883 to any other stall-related activity.
4885 rcutorture.stall_cpu_holdoff= [KNL]
4886 Time to wait (s) after boot before inducing stall.
4888 rcutorture.stall_cpu_irqsoff= [KNL]
4889 Disable interrupts while stalling if set.
4891 rcutorture.stall_gp_kthread= [KNL]
4892 Duration (s) of forced sleep within RCU
4893 grace-period kthread to test RCU CPU stall
4894 warnings, zero to disable. If both stall_cpu
4895 and stall_gp_kthread are specified, the
4896 kthread is starved first, then the CPU.
4898 rcutorture.stat_interval= [KNL]
4899 Time (s) between statistics printk()s.
4901 rcutorture.stutter= [KNL]
4902 Time (s) to stutter testing, for example, specifying
4903 five seconds causes the test to run for five seconds,
4904 wait for five seconds, and so on. This tests RCU's
4905 ability to transition abruptly to and from idle.
4907 rcutorture.test_boost= [KNL]
4908 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4909 "Maybe" means test if the RCU implementation
4910 under test support RCU priority boosting.
4912 rcutorture.test_boost_duration= [KNL]
4913 Duration (s) of each individual boost test.
4915 rcutorture.test_boost_interval= [KNL]
4916 Interval (s) between each boost test.
4918 rcutorture.test_no_idle_hz= [KNL]
4919 Test RCU's dyntick-idle handling. See also the
4920 rcutorture.shuffle_interval parameter.
4922 rcutorture.torture_type= [KNL]
4923 Specify the RCU implementation to test.
4925 rcutorture.verbose= [KNL]
4926 Enable additional printk() statements.
4928 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4929 Dump ftrace buffer after reporting RCU CPU
4932 rcupdate.rcu_cpu_stall_suppress= [KNL]
4933 Suppress RCU CPU stall warning messages.
4935 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4936 Suppress RCU CPU stall warning messages and
4937 rcutorture writer stall warnings that occur
4938 during early boot, that is, during the time
4939 before the init task is spawned.
4941 rcupdate.rcu_cpu_stall_timeout= [KNL]
4942 Set timeout for RCU CPU stall warning messages.
4943 The value is in seconds and the maximum allowed
4944 value is 300 seconds.
4946 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
4947 Set timeout for expedited RCU CPU stall warning
4948 messages. The value is in milliseconds
4949 and the maximum allowed value is 21000
4950 milliseconds. Please note that this value is
4951 adjusted to an arch timer tick resolution.
4952 Setting this to zero causes the value from
4953 rcupdate.rcu_cpu_stall_timeout to be used (after
4954 conversion from seconds to milliseconds).
4956 rcupdate.rcu_expedited= [KNL]
4957 Use expedited grace-period primitives, for
4958 example, synchronize_rcu_expedited() instead
4959 of synchronize_rcu(). This reduces latency,
4960 but can increase CPU utilization, degrade
4961 real-time latency, and degrade energy efficiency.
4962 No effect on CONFIG_TINY_RCU kernels.
4964 rcupdate.rcu_normal= [KNL]
4965 Use only normal grace-period primitives,
4966 for example, synchronize_rcu() instead of
4967 synchronize_rcu_expedited(). This improves
4968 real-time latency, CPU utilization, and
4969 energy efficiency, but can expose users to
4970 increased grace-period latency. This parameter
4971 overrides rcupdate.rcu_expedited. No effect on
4972 CONFIG_TINY_RCU kernels.
4974 rcupdate.rcu_normal_after_boot= [KNL]
4975 Once boot has completed (that is, after
4976 rcu_end_inkernel_boot() has been invoked), use
4977 only normal grace-period primitives. No effect
4978 on CONFIG_TINY_RCU kernels.
4980 But note that CONFIG_PREEMPT_RT=y kernels enables
4981 this kernel boot parameter, forcibly setting
4982 it to the value one, that is, converting any
4983 post-boot attempt at an expedited RCU grace
4984 period to instead use normal non-expedited
4985 grace-period processing.
4987 rcupdate.rcu_task_collapse_lim= [KNL]
4988 Set the maximum number of callbacks present
4989 at the beginning of a grace period that allows
4990 the RCU Tasks flavors to collapse back to using
4991 a single callback queue. This switching only
4992 occurs when rcupdate.rcu_task_enqueue_lim is
4993 set to the default value of -1.
4995 rcupdate.rcu_task_contend_lim= [KNL]
4996 Set the minimum number of callback-queuing-time
4997 lock-contention events per jiffy required to
4998 cause the RCU Tasks flavors to switch to per-CPU
4999 callback queuing. This switching only occurs
5000 when rcupdate.rcu_task_enqueue_lim is set to
5001 the default value of -1.
5003 rcupdate.rcu_task_enqueue_lim= [KNL]
5004 Set the number of callback queues to use for the
5005 RCU Tasks family of RCU flavors. The default
5006 of -1 allows this to be automatically (and
5007 dynamically) adjusted. This parameter is intended
5010 rcupdate.rcu_task_ipi_delay= [KNL]
5011 Set time in jiffies during which RCU tasks will
5012 avoid sending IPIs, starting with the beginning
5013 of a given grace period. Setting a large
5014 number avoids disturbing real-time workloads,
5015 but lengthens grace periods.
5017 rcupdate.rcu_task_stall_info= [KNL]
5018 Set initial timeout in jiffies for RCU task stall
5019 informational messages, which give some indication
5020 of the problem for those not patient enough to
5021 wait for ten minutes. Informational messages are
5022 only printed prior to the stall-warning message
5023 for a given grace period. Disable with a value
5024 less than or equal to zero. Defaults to ten
5025 seconds. A change in value does not take effect
5026 until the beginning of the next grace period.
5028 rcupdate.rcu_task_stall_info_mult= [KNL]
5029 Multiplier for time interval between successive
5030 RCU task stall informational messages for a given
5031 RCU tasks grace period. This value is clamped
5032 to one through ten, inclusive. It defaults to
5033 the value three, so that the first informational
5034 message is printed 10 seconds into the grace
5035 period, the second at 40 seconds, the third at
5036 160 seconds, and then the stall warning at 600
5037 seconds would prevent a fourth at 640 seconds.
5039 rcupdate.rcu_task_stall_timeout= [KNL]
5040 Set timeout in jiffies for RCU task stall
5041 warning messages. Disable with a value less
5042 than or equal to zero. Defaults to ten minutes.
5043 A change in value does not take effect until
5044 the beginning of the next grace period.
5046 rcupdate.rcu_self_test= [KNL]
5047 Run the RCU early boot self tests
5051 Run specified binary instead of /init from the ramdisk,
5052 used for early userspace startup. See initrd.
5055 force - Override the decision by the kernel to hide the
5056 advertisement of RDRAND support (this affects
5057 certain AMD processors because of buggy BIOS
5058 support, specifically around the suspend/resume
5062 Turn on/off individual RDT features. List is:
5063 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5065 E.g. to turn on cmt and turn off mba use:
5069 Format (x86 or x86_64):
5070 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5072 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5074 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5075 (prefix with 'panic_' to set mode for panic
5077 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5078 reboot_force is either force or not specified,
5079 reboot_cpu is s[mp]#### with #### being the processor
5080 to be used for rebooting.
5082 refscale.holdoff= [KNL]
5083 Set test-start holdoff period. The purpose of
5084 this parameter is to delay the start of the
5085 test until boot completes in order to avoid
5088 refscale.loops= [KNL]
5089 Set the number of loops over the synchronization
5090 primitive under test. Increasing this number
5091 reduces noise due to loop start/end overhead,
5092 but the default has already reduced the per-pass
5093 noise to a handful of picoseconds on ca. 2020
5096 refscale.nreaders= [KNL]
5097 Set number of readers. The default value of -1
5098 selects N, where N is roughly 75% of the number
5099 of CPUs. A value of zero is an interesting choice.
5101 refscale.nruns= [KNL]
5102 Set number of runs, each of which is dumped onto
5105 refscale.readdelay= [KNL]
5106 Set the read-side critical-section duration,
5107 measured in microseconds.
5109 refscale.scale_type= [KNL]
5110 Specify the read-protection implementation to test.
5112 refscale.shutdown= [KNL]
5113 Shut down the system at the end of the performance
5114 test. This defaults to 1 (shut it down) when
5115 refscale is built into the kernel and to 0 (leave
5116 it running) when refscale is built as a module.
5118 refscale.verbose= [KNL]
5119 Enable additional printk() statements.
5121 refscale.verbose_batched= [KNL]
5122 Batch the additional printk() statements. If zero
5123 (the default) or negative, print everything. Otherwise,
5124 print every Nth verbose statement, where N is the value
5128 [KNL, SMP] Set scheduler's default relax_domain_level.
5129 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5131 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5132 Format: <base1>,<size1>[,<base2>,<size2>,...]
5133 Reserve I/O ports or memory so the kernel won't use
5134 them. If <base> is less than 0x10000, the region
5135 is assumed to be I/O ports; otherwise it is memory.
5137 reservetop= [X86-32]
5139 Reserves a hole at the top of the kernel virtual
5142 reset_devices [KNL] Force drivers to reset the underlying device
5143 during initialization.
5146 Specify the partition device for software suspend
5148 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5150 resume_offset= [SWSUSP]
5151 Specify the offset from the beginning of the partition
5152 given by "resume=" at which the swap header is located,
5153 in <PAGE_SIZE> units (needed only for swap files).
5154 See Documentation/power/swsusp-and-swap-files.rst
5156 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5157 read the resume files
5159 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5160 Useful for devices that are detected asynchronously
5161 (e.g. USB and MMC devices).
5163 retain_initrd [RAM] Keep initrd memory after extraction
5165 rfkill.default_state=
5166 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5167 etc. communication is blocked by default.
5170 rfkill.master_switch_mode=
5171 0 The "airplane mode" button does nothing.
5172 1 The "airplane mode" button toggles between everything
5173 blocked and the previous configuration.
5174 2 The "airplane mode" button toggles between everything
5175 blocked and everything unblocked.
5177 rhash_entries= [KNL,NET]
5178 Set number of hash buckets for route cache
5181 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5184 ro [KNL] Mount root device read-only on boot
5187 on Mark read-only kernel memory as read-only (default).
5188 off Leave read-only kernel memory writable for debugging.
5191 Enable the uart passthrough on the designated usb port
5192 on Rockchip SoCs. When active, the signals of the
5193 debug-uart get routed to the D+ and D- pins of the usb
5194 port and the regular usb controller gets disabled.
5196 root= [KNL] Root filesystem
5197 See name_to_dev_t comment in init/do_mounts.c.
5199 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5200 mount the root filesystem
5202 rootflags= [KNL] Set root filesystem mount option string
5204 rootfstype= [KNL] Set root filesystem type
5206 rootwait [KNL] Wait (indefinitely) for root device to show up.
5207 Useful for devices that are detected asynchronously
5208 (e.g. USB and MMC devices).
5210 rproc_mem=nn[KMG][@address]
5211 [KNL,ARM,CMA] Remoteproc physical memory block.
5212 Memory area to be used by remote processor image,
5215 rw [KNL] Mount root device read-write on boot
5217 S [KNL] Run init in single mode
5219 s390_iommu= [HW,S390]
5220 Set s390 IOTLB flushing mode
5222 With strict flushing every unmap operation will result in
5223 an IOTLB flush. Default is lazy flushing before reuse,
5226 s390_iommu_aperture= [KNL,S390]
5227 Specifies the size of the per device DMA address space
5228 accessible through the DMA and IOMMU APIs as a decimal
5229 factor of the size of main memory.
5230 The default is 1 meaning that one can concurrently use
5231 as many DMA addresses as physical memory is installed,
5232 if supported by hardware, and thus map all of memory
5233 once. With a value of 2 one can map all of memory twice
5234 and so on. As a special case a factor of 0 imposes no
5235 restrictions other than those given by hardware at the
5236 cost of significant additional memory use for tables.
5239 See drivers/net/irda/sa1100_ir.c.
5241 sched_verbose [KNL] Enables verbose scheduler debug messages.
5243 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5244 Allowed values are enable and disable. This feature
5245 incurs a small amount of overhead in the scheduler
5246 but is useful for debugging and performance tuning.
5248 sched_thermal_decay_shift=
5249 [KNL, SMP] Set a decay shift for scheduler thermal
5250 pressure signal. Thermal pressure signal follows the
5251 default decay period of other scheduler pelt
5252 signals(usually 32 ms but configurable). Setting
5253 sched_thermal_decay_shift will left shift the decay
5254 period for the thermal pressure signal by the shift
5256 i.e. with the default pelt decay period of 32 ms
5257 sched_thermal_decay_shift thermal pressure decay pr
5261 Format: integer between 0 and 10
5264 scftorture.holdoff= [KNL]
5265 Number of seconds to hold off before starting
5266 test. Defaults to zero for module insertion and
5267 to 10 seconds for built-in smp_call_function()
5270 scftorture.longwait= [KNL]
5271 Request ridiculously long waits randomly selected
5272 up to the chosen limit in seconds. Zero (the
5273 default) disables this feature. Please note
5274 that requesting even small non-zero numbers of
5275 seconds can result in RCU CPU stall warnings,
5276 softlockup complaints, and so on.
5278 scftorture.nthreads= [KNL]
5279 Number of kthreads to spawn to invoke the
5280 smp_call_function() family of functions.
5281 The default of -1 specifies a number of kthreads
5282 equal to the number of CPUs.
5284 scftorture.onoff_holdoff= [KNL]
5285 Number seconds to wait after the start of the
5286 test before initiating CPU-hotplug operations.
5288 scftorture.onoff_interval= [KNL]
5289 Number seconds to wait between successive
5290 CPU-hotplug operations. Specifying zero (which
5291 is the default) disables CPU-hotplug operations.
5293 scftorture.shutdown_secs= [KNL]
5294 The number of seconds following the start of the
5295 test after which to shut down the system. The
5296 default of zero avoids shutting down the system.
5297 Non-zero values are useful for automated tests.
5299 scftorture.stat_interval= [KNL]
5300 The number of seconds between outputting the
5301 current test statistics to the console. A value
5302 of zero disables statistics output.
5304 scftorture.stutter_cpus= [KNL]
5305 The number of jiffies to wait between each change
5306 to the set of CPUs under test.
5308 scftorture.use_cpus_read_lock= [KNL]
5309 Use use_cpus_read_lock() instead of the default
5310 preempt_disable() to disable CPU hotplug
5311 while invoking one of the smp_call_function*()
5314 scftorture.verbose= [KNL]
5315 Enable additional printk() statements.
5317 scftorture.weight_single= [KNL]
5318 The probability weighting to use for the
5319 smp_call_function_single() function with a zero
5320 "wait" parameter. A value of -1 selects the
5321 default if all other weights are -1. However,
5322 if at least one weight has some other value, a
5323 value of -1 will instead select a weight of zero.
5325 scftorture.weight_single_wait= [KNL]
5326 The probability weighting to use for the
5327 smp_call_function_single() function with a
5328 non-zero "wait" parameter. See weight_single.
5330 scftorture.weight_many= [KNL]
5331 The probability weighting to use for the
5332 smp_call_function_many() function with a zero
5333 "wait" parameter. See weight_single.
5334 Note well that setting a high probability for
5335 this weighting can place serious IPI load
5338 scftorture.weight_many_wait= [KNL]
5339 The probability weighting to use for the
5340 smp_call_function_many() function with a
5341 non-zero "wait" parameter. See weight_single
5344 scftorture.weight_all= [KNL]
5345 The probability weighting to use for the
5346 smp_call_function_all() function with a zero
5347 "wait" parameter. See weight_single and
5350 scftorture.weight_all_wait= [KNL]
5351 The probability weighting to use for the
5352 smp_call_function_all() function with a
5353 non-zero "wait" parameter. See weight_single
5356 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5357 xtime_lock contention on larger systems, and/or RCU lock
5358 contention on all systems with CONFIG_MAXSMP set.
5359 Format: { "0" | "1" }
5360 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5362 Note: increases power consumption, thus should only be
5363 enabled if running jitter sensitive (HPC/RT) workloads.
5365 security= [SECURITY] Choose a legacy "major" security module to
5366 enable at boot. This has been deprecated by the
5369 selinux= [SELINUX] Disable or enable SELinux at boot time.
5370 Format: { "0" | "1" }
5371 See security/selinux/Kconfig help text.
5376 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5377 Format: { "0" | "1" }
5378 See security/apparmor/Kconfig help text
5381 Default value is set via kernel config option.
5383 serialnumber [BUGS=X86-32]
5385 sev=option[,option...] [X86-64] See Documentation/x86/x86_64/boot-options.rst
5388 Maximal number of shapers.
5396 Enable merging of slabs with similar size when the
5397 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5400 Disable merging of slabs with similar size. May be
5401 necessary if there is some reason to distinguish
5402 allocs to different slabs, especially in hardened
5403 environments where the risk of heap overflows and
5404 layout control by attackers can usually be
5405 frustrated by disabling merging. This will reduce
5406 most of the exposure of a heap attack to a single
5407 cache (risks via metadata attacks are mostly
5408 unchanged). Debug options disable merging on their
5410 For more information see Documentation/vm/slub.rst.
5412 slab_max_order= [MM, SLAB]
5413 Determines the maximum allowed order for slabs.
5414 A high setting may cause OOMs due to memory
5415 fragmentation. Defaults to 1 for systems with
5416 more than 32MB of RAM, 0 otherwise.
5418 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5419 Enabling slub_debug allows one to determine the
5420 culprit if slab objects become corrupted. Enabling
5421 slub_debug can create guard zones around objects and
5422 may poison objects when not in use. Also tracks the
5423 last alloc / free. For more information see
5424 Documentation/vm/slub.rst.
5426 slub_max_order= [MM, SLUB]
5427 Determines the maximum allowed order for slabs.
5428 A high setting may cause OOMs due to memory
5429 fragmentation. For more information see
5430 Documentation/vm/slub.rst.
5432 slub_min_objects= [MM, SLUB]
5433 The minimum number of objects per slab. SLUB will
5434 increase the slab order up to slub_max_order to
5435 generate a sufficiently large slab able to contain
5436 the number of objects indicated. The higher the number
5437 of objects the smaller the overhead of tracking slabs
5438 and the less frequently locks need to be acquired.
5439 For more information see Documentation/vm/slub.rst.
5441 slub_min_order= [MM, SLUB]
5442 Determines the minimum page order for slabs. Must be
5443 lower than slub_max_order.
5444 For more information see Documentation/vm/slub.rst.
5446 slub_merge [MM, SLUB]
5447 Same with slab_merge.
5449 slub_nomerge [MM, SLUB]
5450 Same with slab_nomerge. This is supported for legacy.
5451 See slab_nomerge for more information.
5454 Format: <io1>[,<io2>[,...,<io8>]]
5456 smp.csd_lock_timeout= [KNL]
5457 Specify the period of time in milliseconds
5458 that smp_call_function() and friends will wait
5459 for a CPU to release the CSD lock. This is
5460 useful when diagnosing bugs involving CPUs
5461 disabling interrupts for extended periods
5462 of time. Defaults to 5,000 milliseconds, and
5463 setting a value of zero disables this feature.
5464 This feature may be more efficiently disabled
5465 using the csdlock_debug- kernel parameter.
5467 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5468 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5469 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5470 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5471 smsc-ircc2.ircc_irq= [HW] IRQ line
5472 smsc-ircc2.ircc_dma= [HW] DMA channel
5473 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5474 0: Toshiba Satellite 1800 (GP data pin select)
5475 1: Fast pin select (default)
5478 smt= [KNL,S390] Set the maximum number of threads (logical
5479 CPUs) to use per physical CPU on systems capable of
5480 symmetric multithreading (SMT). Will be capped to the
5481 actual hardware limit.
5483 Default: -1 (no limit)
5486 [KNL] Should the soft-lockup detector generate panics.
5489 A value of 1 instructs the soft-lockup detector
5490 to panic the machine when a soft-lockup occurs. It is
5491 also controlled by the kernel.softlockup_panic sysctl
5492 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5493 respective build-time switch to that functionality.
5495 softlockup_all_cpu_backtrace=
5496 [KNL] Should the soft-lockup detector generate
5497 backtraces on all cpus.
5500 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5501 See Documentation/admin-guide/laptops/sonypi.rst
5503 spectre_v2= [X86] Control mitigation of Spectre variant 2
5504 (indirect branch speculation) vulnerability.
5505 The default operation protects the kernel from
5508 on - unconditionally enable, implies
5510 off - unconditionally disable, implies
5512 auto - kernel detects whether your CPU model is
5515 Selecting 'on' will, and 'auto' may, choose a
5516 mitigation method at run time according to the
5517 CPU, the available microcode, the setting of the
5518 CONFIG_RETPOLINE configuration option, and the
5519 compiler with which the kernel was built.
5521 Selecting 'on' will also enable the mitigation
5522 against user space to user space task attacks.
5524 Selecting 'off' will disable both the kernel and
5525 the user space protections.
5527 Specific mitigations can also be selected manually:
5529 retpoline - replace indirect branches
5530 retpoline,generic - Retpolines
5531 retpoline,lfence - LFENCE; indirect branch
5532 retpoline,amd - alias for retpoline,lfence
5533 eibrs - enhanced IBRS
5534 eibrs,retpoline - enhanced IBRS + Retpolines
5535 eibrs,lfence - enhanced IBRS + LFENCE
5537 Not specifying this option is equivalent to
5541 [X86] Control mitigation of Spectre variant 2
5542 (indirect branch speculation) vulnerability between
5545 on - Unconditionally enable mitigations. Is
5546 enforced by spectre_v2=on
5548 off - Unconditionally disable mitigations. Is
5549 enforced by spectre_v2=off
5551 prctl - Indirect branch speculation is enabled,
5552 but mitigation can be enabled via prctl
5553 per thread. The mitigation control state
5554 is inherited on fork.
5557 - Like "prctl" above, but only STIBP is
5558 controlled per thread. IBPB is issued
5559 always when switching between different user
5563 - Same as "prctl" above, but all seccomp
5564 threads will enable the mitigation unless
5565 they explicitly opt out.
5568 - Like "seccomp" above, but only STIBP is
5569 controlled per thread. IBPB is issued
5570 always when switching between different
5571 user space processes.
5573 auto - Kernel selects the mitigation depending on
5574 the available CPU features and vulnerability.
5576 Default mitigation: "prctl"
5578 Not specifying this option is equivalent to
5579 spectre_v2_user=auto.
5581 spec_store_bypass_disable=
5582 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5583 (Speculative Store Bypass vulnerability)
5585 Certain CPUs are vulnerable to an exploit against a
5586 a common industry wide performance optimization known
5587 as "Speculative Store Bypass" in which recent stores
5588 to the same memory location may not be observed by
5589 later loads during speculative execution. The idea
5590 is that such stores are unlikely and that they can
5591 be detected prior to instruction retirement at the
5592 end of a particular speculation execution window.
5594 In vulnerable processors, the speculatively forwarded
5595 store can be used in a cache side channel attack, for
5596 example to read memory to which the attacker does not
5597 directly have access (e.g. inside sandboxed code).
5599 This parameter controls whether the Speculative Store
5600 Bypass optimization is used.
5602 On x86 the options are:
5604 on - Unconditionally disable Speculative Store Bypass
5605 off - Unconditionally enable Speculative Store Bypass
5606 auto - Kernel detects whether the CPU model contains an
5607 implementation of Speculative Store Bypass and
5608 picks the most appropriate mitigation. If the
5609 CPU is not vulnerable, "off" is selected. If the
5610 CPU is vulnerable the default mitigation is
5611 architecture and Kconfig dependent. See below.
5612 prctl - Control Speculative Store Bypass per thread
5613 via prctl. Speculative Store Bypass is enabled
5614 for a process by default. The state of the control
5615 is inherited on fork.
5616 seccomp - Same as "prctl" above, but all seccomp threads
5617 will disable SSB unless they explicitly opt out.
5619 Default mitigations:
5622 On powerpc the options are:
5624 on,auto - On Power8 and Power9 insert a store-forwarding
5625 barrier on kernel entry and exit. On Power7
5626 perform a software flush on kernel entry and
5630 Not specifying this option is equivalent to
5631 spec_store_bypass_disable=auto.
5633 spia_io_base= [HW,MTD]
5639 [X86] Enable split lock detection or bus lock detection
5641 When enabled (and if hardware support is present), atomic
5642 instructions that access data across cache line
5643 boundaries will result in an alignment check exception
5644 for split lock detection or a debug exception for
5649 warn - the kernel will emit rate-limited warnings
5650 about applications triggering the #AC
5651 exception or the #DB exception. This mode is
5652 the default on CPUs that support split lock
5653 detection or bus lock detection. Default
5654 behavior is by #AC if both features are
5655 enabled in hardware.
5657 fatal - the kernel will send SIGBUS to applications
5658 that trigger the #AC exception or the #DB
5659 exception. Default behavior is by #AC if
5660 both features are enabled in hardware.
5663 Set system wide rate limit to N bus locks
5664 per second for bus lock detection.
5667 N/A for split lock detection.
5670 If an #AC exception is hit in the kernel or in
5671 firmware (i.e. not while executing in user mode)
5672 the kernel will oops in either "warn" or "fatal"
5675 #DB exception for bus lock is triggered only when
5679 Control the Special Register Buffer Data Sampling
5682 Certain CPUs are vulnerable to an MDS-like
5683 exploit which can leak bits from the random
5686 By default, this issue is mitigated by
5687 microcode. However, the microcode fix can cause
5688 the RDRAND and RDSEED instructions to become
5689 much slower. Among other effects, this will
5690 result in reduced throughput from /dev/urandom.
5692 The microcode mitigation can be disabled with
5693 the following option:
5695 off: Disable mitigation and remove
5696 performance impact to RDRAND and RDSEED
5698 srcutree.big_cpu_lim [KNL]
5699 Specifies the number of CPUs constituting a
5700 large system, such that srcu_struct structures
5701 should immediately allocate an srcu_node array.
5702 This kernel-boot parameter defaults to 128,
5703 but takes effect only when the low-order four
5704 bits of srcutree.convert_to_big is equal to 3
5707 srcutree.convert_to_big [KNL]
5708 Specifies under what conditions an SRCU tree
5709 srcu_struct structure will be converted to big
5710 form, that is, with an rcu_node tree:
5713 1: At init_srcu_struct() time.
5714 2: When rcutorture decides to.
5715 3: Decide at boot time (default).
5716 0x1X: Above plus if high contention.
5718 Either way, the srcu_node tree will be sized based
5719 on the actual runtime number of CPUs (nr_cpu_ids)
5720 instead of the compile-time CONFIG_NR_CPUS.
5722 srcutree.counter_wrap_check [KNL]
5723 Specifies how frequently to check for
5724 grace-period sequence counter wrap for the
5725 srcu_data structure's ->srcu_gp_seq_needed field.
5726 The greater the number of bits set in this kernel
5727 parameter, the less frequently counter wrap will
5728 be checked for. Note that the bottom two bits
5731 srcutree.exp_holdoff [KNL]
5732 Specifies how many nanoseconds must elapse
5733 since the end of the last SRCU grace period for
5734 a given srcu_struct until the next normal SRCU
5735 grace period will be considered for automatic
5736 expediting. Set to zero to disable automatic
5739 srcutree.small_contention_lim [KNL]
5740 Specifies the number of update-side contention
5741 events per jiffy will be tolerated before
5742 initiating a conversion of an srcu_struct
5743 structure to big form. Note that the value of
5744 srcutree.convert_to_big must have the 0x10 bit
5745 set for contention-based conversions to occur.
5748 Speculative Store Bypass Disable control
5750 On CPUs that are vulnerable to the Speculative
5751 Store Bypass vulnerability and offer a
5752 firmware based mitigation, this parameter
5753 indicates how the mitigation should be used:
5755 force-on: Unconditionally enable mitigation for
5756 for both kernel and userspace
5757 force-off: Unconditionally disable mitigation for
5758 for both kernel and userspace
5759 kernel: Always enable mitigation in the
5760 kernel, and offer a prctl interface
5761 to allow userspace to register its
5762 interest in being mitigated too.
5764 stack_guard_gap= [MM]
5765 override the default stack gap protection. The value
5766 is in page units and it defines how many pages prior
5767 to (for stacks growing down) resp. after (for stacks
5768 growing up) the main stack are reserved for no other
5769 mapping. Default value is 256 pages.
5771 stack_depot_disable= [KNL]
5772 Setting this to true through kernel command line will
5773 disable the stack depot thereby saving the static memory
5774 consumed by the stack hash table. By default this is set
5778 Enabled the stack tracer on boot up.
5780 stacktrace_filter=[function-list]
5781 [FTRACE] Limit the functions that the stack tracer
5782 will trace at boot up. function-list is a comma-separated
5783 list of functions. This list can be changed at run
5784 time by the stack_trace_filter file in the debugfs
5785 tracing directory. Note, this enables stack tracing
5786 and the stacktrace above is not needed.
5790 Set the STI (builtin display/keyboard on the HP-PARISC
5791 machines) console (graphic card) which should be used
5792 as the initial boot-console.
5793 See also comment in drivers/video/console/sticore.c.
5796 See comment in drivers/video/console/sticore.c.
5799 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5804 Enable or disable strict sigaltstack size checks
5805 against the required signal frame size which
5806 depends on the supported FPU features. This can
5807 be used to filter out binaries which have
5808 not yet been made aware of AT_MINSIGSTKSZ.
5810 sunrpc.min_resvport=
5811 sunrpc.max_resvport=
5813 SunRPC servers often require that client requests
5814 originate from a privileged port (i.e. a port in the
5815 range 0 < portnr < 1024).
5816 An administrator who wishes to reserve some of these
5817 ports for other uses may adjust the range that the
5818 kernel's sunrpc client considers to be privileged
5819 using these two parameters to set the minimum and
5820 maximum port values.
5822 sunrpc.svc_rpc_per_connection_limit=
5824 Limit the number of requests that the server will
5825 process in parallel from a single connection.
5826 The default value is 0 (no limit).
5830 Control how the NFS server code allocates CPUs to
5831 service thread pools. Depending on how many NICs
5832 you have and where their interrupts are bound, this
5833 option will affect which CPUs will do NFS serving.
5834 Note: this parameter cannot be changed while the
5835 NFS server is running.
5837 auto the server chooses an appropriate mode
5838 automatically using heuristics
5839 global a single global pool contains all CPUs
5840 percpu one pool for each CPU
5841 pernode one pool for each NUMA node (equivalent
5842 to global on non-NUMA machines)
5844 sunrpc.tcp_slot_table_entries=
5845 sunrpc.udp_slot_table_entries=
5847 Sets the upper limit on the number of simultaneous
5848 RPC calls that can be sent from the client to a
5849 server. Increasing these values may allow you to
5850 improve throughput, but will also increase the
5851 amount of memory reserved for use by the client.
5853 suspend.pm_test_delay=
5855 Sets the number of seconds to remain in a suspend test
5856 mode before resuming the system (see
5857 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5858 is set. Default value is 5.
5861 Format: { on | off | y | n | 1 | 0 }
5862 This parameter controls use of the Protected
5863 Execution Facility on pSeries.
5867 Enable accounting of swap in memory resource
5868 controller if no parameter or 1 is given or disable
5869 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5871 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5872 Format: { <int> | force | noforce }
5873 <int> -- Number of I/O TLB slabs
5874 force -- force using of bounce buffers even if they
5875 wouldn't be automatically used by the kernel
5876 noforce -- Never use bounce buffers (for debugging)
5881 Set a sysctl parameter, right before loading the init
5882 process, as if the value was written to the respective
5883 /proc/sys/... file. Both '.' and '/' are recognized as
5884 separators. Unrecognized parameters and invalid values
5885 are reported in the kernel log. Sysctls registered
5886 later by a loaded module cannot be set this way.
5887 Example: sysctl.vm.swappiness=40
5889 sysfs.deprecated=0|1 [KNL]
5890 Enable/disable old style sysfs layout for old udev
5891 on older distributions. When this option is enabled
5892 very new udev will not work anymore. When this option
5893 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5894 in older udev will not work anymore.
5895 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5896 the kernel configuration.
5898 sysrq_always_enabled
5900 Ignore sysrq setting - this boot parameter will
5901 neutralize any effect of /proc/sys/kernel/sysrq.
5902 Useful for debugging.
5904 tcpmhash_entries= [KNL,NET]
5905 Set the number of tcp_metrics_hash slots.
5906 Default value is 8192 or 16384 depending on total
5907 ram pages. This is used to specify the TCP metrics
5908 cache size. See Documentation/networking/ip-sysctl.rst
5909 "tcp_no_metrics_save" section for more details.
5913 test_suspend= [SUSPEND]
5914 Format: { "mem" | "standby" | "freeze" }[,N]
5915 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5916 standby suspend) or "freeze" (for suspend type freeze)
5917 as the system sleep state during system startup with
5918 the optional capability to repeat N number of times.
5919 The system is woken from this state using a
5920 wakeup-capable RTC alarm.
5922 thash_entries= [KNL,NET]
5923 Set number of hash buckets for TCP connection
5925 thermal.act= [HW,ACPI]
5926 -1: disable all active trip points in all thermal zones
5927 <degrees C>: override all lowest active trip points
5929 thermal.crt= [HW,ACPI]
5930 -1: disable all critical trip points in all thermal zones
5931 <degrees C>: override all critical trip points
5933 thermal.nocrt= [HW,ACPI]
5934 Set to disable actions on ACPI thermal zone
5935 critical and hot trip points.
5937 thermal.off= [HW,ACPI]
5938 1: disable ACPI thermal control
5940 thermal.psv= [HW,ACPI]
5941 -1: disable all passive trip points
5942 <degrees C>: override all passive trip points to this
5945 thermal.tzp= [HW,ACPI]
5946 Specify global default ACPI thermal zone polling rate
5947 <deci-seconds>: poll all this frequency
5948 0: no polling (default)
5951 Force threading of all interrupt handlers except those
5952 marked explicitly IRQF_NO_THREAD.
5956 Specify if the kernel should make use of the cpu
5957 topology information if the hardware supports this.
5958 The scheduler will make use of this information and
5959 e.g. base its process migration decisions on it.
5962 topology_updates= [KNL, PPC, NUMA]
5964 Specify if the kernel should ignore (off)
5965 topology updates sent by the hypervisor to this
5968 torture.disable_onoff_at_boot= [KNL]
5969 Prevent the CPU-hotplug component of torturing
5970 until after init has spawned.
5972 torture.ftrace_dump_at_shutdown= [KNL]
5973 Dump the ftrace buffer at torture-test shutdown,
5974 even if there were no errors. This can be a
5975 very costly operation when many torture tests
5976 are running concurrently, especially on systems
5977 with rotating-rust storage.
5979 torture.verbose_sleep_frequency= [KNL]
5980 Specifies how many verbose printk()s should be
5981 emitted between each sleep. The default of zero
5982 disables verbose-printk() sleeping.
5984 torture.verbose_sleep_duration= [KNL]
5985 Duration of each verbose-printk() sleep in jiffies.
5989 tpm_suspend_pcr=[HW,TPM]
5990 Format: integer pcr id
5991 Specify that at suspend time, the tpm driver
5992 should extend the specified pcr with zeros,
5993 as a workaround for some chips which fail to
5994 flush the last written pcr on TPM_SaveState.
5995 This will guarantee that all the other pcrs
5999 Have the tracepoints sent to printk as well as the
6000 tracing ring buffer. This is useful for early boot up
6001 where the system hangs or reboots and does not give the
6002 option for reading the tracing buffer or performing a
6003 ftrace_dump_on_oops.
6005 To turn off having tracepoints sent to printk,
6006 echo 0 > /proc/sys/kernel/tracepoint_printk
6007 Note, echoing 1 into this file without the
6008 tracepoint_printk kernel cmdline option has no effect.
6010 The tp_printk_stop_on_boot (see below) can also be used
6011 to stop the printing of events to console at
6016 Having tracepoints sent to printk() and activating high
6017 frequency tracepoints such as irq or sched, can cause
6018 the system to live lock.
6020 tp_printk_stop_on_boot [FTRACE]
6021 When tp_printk (above) is set, it can cause a lot of noise
6022 on the console. It may be useful to only include the
6023 printing of events during boot up, as user space may
6024 make the system inoperable.
6026 This command line option will stop the printing of events
6027 to console at the late_initcall_sync() time frame.
6029 trace_buf_size=nn[KMG]
6030 [FTRACE] will set tracing buffer size on each cpu.
6032 trace_clock= [FTRACE] Set the clock used for tracing events
6034 local - Use the per CPU time stamp counter
6035 (converted into nanoseconds). Fast, but
6036 depending on the architecture, may not be
6037 in sync between CPUs.
6038 global - Event time stamps are synchronize across
6039 CPUs. May be slower than the local clock,
6040 but better for some race conditions.
6041 counter - Simple counting of events (1, 2, ..)
6042 note, some counts may be skipped due to the
6043 infrastructure grabbing the clock more than
6045 uptime - Use jiffies as the time stamp.
6046 perf - Use the same clock that perf uses.
6047 mono - Use ktime_get_mono_fast_ns() for time stamps.
6048 mono_raw - Use ktime_get_raw_fast_ns() for time
6050 boot - Use ktime_get_boot_fast_ns() for time stamps.
6051 Architectures may add more clocks. See
6052 Documentation/trace/ftrace.rst for more details.
6054 trace_event=[event-list]
6055 [FTRACE] Set and start specified trace events in order
6056 to facilitate early boot debugging. The event-list is a
6057 comma-separated list of trace events to enable. See
6058 also Documentation/trace/events.rst
6060 trace_options=[option-list]
6061 [FTRACE] Enable or disable tracer options at boot.
6062 The option-list is a comma delimited list of options
6063 that can be enabled or disabled just as if you were
6064 to echo the option name into
6066 /sys/kernel/debug/tracing/trace_options
6068 For example, to enable stacktrace option (to dump the
6069 stack trace of each event), add to the command line:
6071 trace_options=stacktrace
6073 See also Documentation/trace/ftrace.rst "trace options"
6077 [FTRACE] enable this option to disable tracing when a
6078 warning is hit. This turns off "tracing_on". Tracing can
6079 be enabled again by echoing '1' into the "tracing_on"
6080 file located in /sys/kernel/debug/tracing/
6082 This option is useful, as it disables the trace before
6083 the WARNING dump is called, which prevents the trace to
6084 be filled with content caused by the warning output.
6086 This option can also be set at run time via the sysctl
6087 option: kernel/traceoff_on_warning
6089 transparent_hugepage=
6091 Format: [always|madvise|never]
6092 Can be used to control the default behavior of the system
6093 with respect to transparent hugepages.
6094 See Documentation/admin-guide/mm/transhuge.rst
6097 trusted.source= [KEYS]
6099 This parameter identifies the trust source as a backend
6100 for trusted keys implementation. Supported trust
6105 If not specified then it defaults to iterating through
6106 the trust source list starting with TPM and assigns the
6107 first trust source as a backend which is initialized
6108 successfully during iteration.
6112 The RNG used to generate key material for trusted keys.
6115 - the same value as trusted.source: "tpm" or "tee"
6117 If not specified, "default" is used. In this case,
6118 the RNG's choice is left to each individual trust source.
6120 tsc= Disable clocksource stability checks for TSC.
6122 [x86] reliable: mark tsc clocksource as reliable, this
6123 disables clocksource verification at runtime, as well
6124 as the stability checks done at bootup. Used to enable
6125 high-resolution timer mode on older hardware, and in
6126 virtualized environment.
6127 [x86] noirqtime: Do not use TSC to do irq accounting.
6128 Used to run time disable IRQ_TIME_ACCOUNTING on any
6129 platforms where RDTSC is slow and this accounting
6131 [x86] unstable: mark the TSC clocksource as unstable, this
6132 marks the TSC unconditionally unstable at bootup and
6133 avoids any further wobbles once the TSC watchdog notices.
6134 [x86] nowatchdog: disable clocksource watchdog. Used
6135 in situations with strict latency requirements (where
6136 interruptions from clocksource watchdog are not
6139 tsc_early_khz= [X86] Skip early TSC calibration and use the given
6140 value instead. Useful when the early TSC frequency discovery
6141 procedure is not reliable, such as on overclocked systems
6142 with CPUID.16h support and partial CPUID.15h support.
6143 Format: <unsigned int>
6145 tsx= [X86] Control Transactional Synchronization
6146 Extensions (TSX) feature in Intel processors that
6147 support TSX control.
6149 This parameter controls the TSX feature. The options are:
6151 on - Enable TSX on the system. Although there are
6152 mitigations for all known security vulnerabilities,
6153 TSX has been known to be an accelerator for
6154 several previous speculation-related CVEs, and
6155 so there may be unknown security risks associated
6156 with leaving it enabled.
6158 off - Disable TSX on the system. (Note that this
6159 option takes effect only on newer CPUs which are
6160 not vulnerable to MDS, i.e., have
6161 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6162 the new IA32_TSX_CTRL MSR through a microcode
6163 update. This new MSR allows for the reliable
6164 deactivation of the TSX functionality.)
6166 auto - Disable TSX if X86_BUG_TAA is present,
6167 otherwise enable TSX on the system.
6169 Not specifying this option is equivalent to tsx=off.
6171 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6174 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6175 Abort (TAA) vulnerability.
6177 Similar to Micro-architectural Data Sampling (MDS)
6178 certain CPUs that support Transactional
6179 Synchronization Extensions (TSX) are vulnerable to an
6180 exploit against CPU internal buffers which can forward
6181 information to a disclosure gadget under certain
6184 In vulnerable processors, the speculatively forwarded
6185 data can be used in a cache side channel attack, to
6186 access data to which the attacker does not have direct
6189 This parameter controls the TAA mitigation. The
6192 full - Enable TAA mitigation on vulnerable CPUs
6195 full,nosmt - Enable TAA mitigation and disable SMT on
6196 vulnerable CPUs. If TSX is disabled, SMT
6197 is not disabled because CPU is not
6198 vulnerable to cross-thread TAA attacks.
6199 off - Unconditionally disable TAA mitigation
6201 On MDS-affected machines, tsx_async_abort=off can be
6202 prevented by an active MDS mitigation as both vulnerabilities
6203 are mitigated with the same mechanism so in order to disable
6204 this mitigation, you need to specify mds=off too.
6206 Not specifying this option is equivalent to
6207 tsx_async_abort=full. On CPUs which are MDS affected
6208 and deploy MDS mitigation, TAA mitigation is not
6209 required and doesn't provide any additional
6213 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6215 turbografx.map[2|3]= [HW,JOY]
6216 TurboGraFX parallel port interface
6218 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6219 See also Documentation/input/devices/joystick-parport.rst
6221 udbg-immortal [PPC] When debugging early kernel crashes that
6222 happen after console_init() and before a proper
6223 console driver takes over, this boot options might
6224 help "seeing" what's going on.
6226 uhash_entries= [KNL,NET]
6227 Set number of hash buckets for UDP/UDP-Lite connections
6230 [USB] Ignore overcurrent events (default N).
6231 Some badly-designed motherboards generate lots of
6232 bogus events, for ports that aren't wired to
6233 anything. Set this parameter to avoid log spamming.
6234 Note that genuine overcurrent events won't be
6238 [X86] Cause panic on unknown NMI.
6240 usbcore.authorized_default=
6241 [USB] Default USB device authorization:
6242 (default -1 = authorized except for wireless USB,
6243 0 = not authorized, 1 = authorized, 2 = authorized
6244 if device connected to internal port)
6246 usbcore.autosuspend=
6247 [USB] The autosuspend time delay (in seconds) used
6248 for newly-detected USB devices (default 2). This
6249 is the time required before an idle device will be
6250 autosuspended. Devices for which the delay is set
6251 to a negative value won't be autosuspended at all.
6253 usbcore.usbfs_snoop=
6254 [USB] Set to log all usbfs traffic (default 0 = off).
6256 usbcore.usbfs_snoop_max=
6257 [USB] Maximum number of bytes to snoop in each URB
6260 usbcore.blinkenlights=
6261 [USB] Set to cycle leds on hubs (default 0 = off).
6263 usbcore.old_scheme_first=
6264 [USB] Start with the old device initialization
6265 scheme (default 0 = off).
6267 usbcore.usbfs_memory_mb=
6268 [USB] Memory limit (in MB) for buffers allocated by
6269 usbfs (default = 16, 0 = max = 2047).
6271 usbcore.use_both_schemes=
6272 [USB] Try the other device initialization scheme
6273 if the first one fails (default 1 = enabled).
6275 usbcore.initial_descriptor_timeout=
6276 [USB] Specifies timeout for the initial 64-byte
6277 USB_REQ_GET_DESCRIPTOR request in milliseconds
6278 (default 5000 = 5.0 seconds).
6280 usbcore.nousb [USB] Disable the USB subsystem
6283 [USB] A list of quirk entries to augment the built-in
6284 usb core quirk list. List entries are separated by
6285 commas. Each entry has the form
6286 VendorID:ProductID:Flags. The IDs are 4-digit hex
6287 numbers and Flags is a set of letters. Each letter
6288 will change the built-in quirk; setting it if it is
6289 clear and clearing it if it is set. The letters have
6290 the following meanings:
6291 a = USB_QUIRK_STRING_FETCH_255 (string
6292 descriptors must not be fetched using
6294 b = USB_QUIRK_RESET_RESUME (device can't resume
6295 correctly so reset it instead);
6296 c = USB_QUIRK_NO_SET_INTF (device can't handle
6297 Set-Interface requests);
6298 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6299 handle its Configuration or Interface
6301 e = USB_QUIRK_RESET (device can't be reset
6302 (e.g morph devices), don't use reset);
6303 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6304 more interface descriptions than the
6305 bNumInterfaces count, and can't handle
6306 talking to these interfaces);
6307 g = USB_QUIRK_DELAY_INIT (device needs a pause
6308 during initialization, after we read
6309 the device descriptor);
6310 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6311 high speed and super speed interrupt
6312 endpoints, the USB 2.0 and USB 3.0 spec
6313 require the interval in microframes (1
6314 microframe = 125 microseconds) to be
6315 calculated as interval = 2 ^
6317 Devices with this quirk report their
6318 bInterval as the result of this
6319 calculation instead of the exponent
6320 variable used in the calculation);
6321 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6322 handle device_qualifier descriptor
6324 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6325 generates spurious wakeup, ignore
6326 remote wakeup capability);
6327 k = USB_QUIRK_NO_LPM (device can't handle Link
6329 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6330 (Device reports its bInterval as linear
6331 frames instead of the USB 2.0
6333 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6334 to be disconnected before suspend to
6335 prevent spurious wakeup);
6336 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6337 pause after every control message);
6338 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6339 delay after resetting its port);
6340 Example: quirks=0781:5580:bk,0a5c:5834:gij
6343 [USBHID] The interval which mice are to be polled at.
6346 [USBHID] The interval which joysticks are to be polled at.
6349 [USBHID] The interval which keyboards are to be polled at.
6351 usb-storage.delay_use=
6352 [UMS] The delay in seconds before a new device is
6353 scanned for Logical Units (default 1).
6356 [UMS] A list of quirks entries to supplement or
6357 override the built-in unusual_devs list. List
6358 entries are separated by commas. Each entry has
6359 the form VID:PID:Flags where VID and PID are Vendor
6360 and Product ID values (4-digit hex numbers) and
6361 Flags is a set of characters, each corresponding
6362 to a common usb-storage quirk flag as follows:
6363 a = SANE_SENSE (collect more than 18 bytes
6364 of sense data, not on uas);
6365 b = BAD_SENSE (don't collect more than 18
6366 bytes of sense data, not on uas);
6367 c = FIX_CAPACITY (decrease the reported
6368 device capacity by one sector);
6369 d = NO_READ_DISC_INFO (don't use
6370 READ_DISC_INFO command, not on uas);
6371 e = NO_READ_CAPACITY_16 (don't use
6372 READ_CAPACITY_16 command);
6373 f = NO_REPORT_OPCODES (don't use report opcodes
6375 g = MAX_SECTORS_240 (don't transfer more than
6376 240 sectors at a time, uas only);
6377 h = CAPACITY_HEURISTICS (decrease the
6378 reported device capacity by one
6379 sector if the number is odd);
6380 i = IGNORE_DEVICE (don't bind to this
6382 j = NO_REPORT_LUNS (don't use report luns
6384 k = NO_SAME (do not use WRITE_SAME, uas only)
6385 l = NOT_LOCKABLE (don't try to lock and
6386 unlock ejectable media, not on uas);
6387 m = MAX_SECTORS_64 (don't transfer more
6388 than 64 sectors = 32 KB at a time,
6390 n = INITIAL_READ10 (force a retry of the
6391 initial READ(10) command, not on uas);
6392 o = CAPACITY_OK (accept the capacity
6393 reported by the device, not on uas);
6394 p = WRITE_CACHE (the device cache is ON
6395 by default, not on uas);
6396 r = IGNORE_RESIDUE (the device reports
6397 bogus residue values, not on uas);
6398 s = SINGLE_LUN (the device has only one
6400 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6401 commands, uas only);
6402 u = IGNORE_UAS (don't bind to the uas driver);
6403 w = NO_WP_DETECT (don't test whether the
6404 medium is write-protected).
6405 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6406 even if the device claims no cache,
6408 Example: quirks=0419:aaf5:rl,0421:0433:rc
6410 user_debug= [KNL,ARM]
6412 See arch/arm/Kconfig.debug help text.
6413 1 - undefined instruction events
6415 4 - invalid data aborts
6418 Example: user_debug=31
6421 [X86] Flags controlling user PTE allocations.
6423 nohigh = do not allocate PTE pages in
6424 HIGHMEM regardless of setting
6427 vdso= [X86,SH,SPARC]
6428 On X86_32, this is an alias for vdso32=. Otherwise:
6430 vdso=1: enable VDSO (the default)
6431 vdso=0: disable VDSO mapping
6433 vdso32= [X86] Control the 32-bit vDSO
6434 vdso32=1: enable 32-bit VDSO
6435 vdso32=0 or vdso32=2: disable 32-bit VDSO
6437 See the help text for CONFIG_COMPAT_VDSO for more
6438 details. If CONFIG_COMPAT_VDSO is set, the default is
6439 vdso32=0; otherwise, the default is vdso32=1.
6441 For compatibility with older kernels, vdso32=2 is an
6444 Try vdso32=0 if you encounter an error that says:
6445 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6448 vector=percpu: enable percpu vector domain
6450 video= [FB] Frame buffer configuration
6451 See Documentation/fb/modedb.rst.
6453 video.brightness_switch_enabled= [ACPI]
6455 If set to 1, on receiving an ACPI notify event
6456 generated by hotkey, video driver will adjust brightness
6457 level and then send out the event to user space through
6458 the allocated input device. If set to 0, video driver
6459 will only send out the event without touching backlight
6464 [VMMIO] Memory mapped virtio (platform) device.
6466 <size>@<baseaddr>:<irq>[:<id>]
6468 <size> := size (can use standard suffixes
6470 <baseaddr> := physical base address
6471 <irq> := interrupt number (as passed to
6473 <id> := (optional) platform device id
6475 virtio_mmio.device=1K@0x100b0000:48:7
6477 Can be used multiple times for multiple devices.
6479 vga= [BOOT,X86-32] Select a particular video mode
6480 See Documentation/x86/boot.rst and
6481 Documentation/admin-guide/svga.rst.
6482 Use vga=ask for menu.
6483 This is actually a boot loader parameter; the value is
6484 passed to the kernel using a special protocol.
6486 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6487 May slow down system boot speed, especially when
6488 enabled on systems with a large amount of memory.
6489 All options are enabled by default, and this
6490 interface is meant to allow for selectively
6491 enabling or disabling specific virtual memory
6494 Available options are:
6495 P Enable page structure init time poisoning
6496 - Disable all of the above options
6498 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6499 size of <nn>. This can be used to increase the
6500 minimum size (128MB on x86). It can also be used to
6501 decrease the size and leave more room for directly
6504 vmcp_cma=nn[MG] [KNL,S390]
6505 Sets the memory size reserved for contiguous memory
6506 allocations for the vmcp device driver.
6508 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6511 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6514 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6518 Controls the behavior of vsyscalls (i.e. calls to
6519 fixed addresses of 0xffffffffff600x00 from legacy
6520 code). Most statically-linked binaries and older
6521 versions of glibc use these calls. Because these
6522 functions are at fixed addresses, they make nice
6523 targets for exploits that can control RIP.
6525 emulate [default] Vsyscalls turn into traps and are
6526 emulated reasonably safely. The vsyscall
6529 xonly Vsyscalls turn into traps and are
6530 emulated reasonably safely. The vsyscall
6531 page is not readable.
6533 none Vsyscalls don't work at all. This makes
6534 them quite hard to use for exploits but
6535 might break your system.
6537 vt.color= [VT] Default text color.
6538 Format: 0xYX, X = foreground, Y = background.
6539 Default: 0x07 = light gray on black.
6541 vt.cur_default= [VT] Default cursor shape.
6542 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6543 the parameters of the <Esc>[?A;B;Cc escape sequence;
6544 see VGA-softcursor.txt. Default: 2 = underline.
6546 vt.default_blu= [VT]
6547 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6548 Change the default blue palette of the console.
6549 This is a 16-member array composed of values
6552 vt.default_grn= [VT]
6553 Format: <green0>,<green1>,<green2>,...,<green15>
6554 Change the default green palette of the console.
6555 This is a 16-member array composed of values
6558 vt.default_red= [VT]
6559 Format: <red0>,<red1>,<red2>,...,<red15>
6560 Change the default red palette of the console.
6561 This is a 16-member array composed of values
6567 Set system-wide default UTF-8 mode for all tty's.
6568 Default is 1, i.e. UTF-8 mode is enabled for all
6569 newly opened terminals.
6571 vt.global_cursor_default=
6574 Set system-wide default for whether a cursor
6575 is shown on new VTs. Default is -1,
6576 i.e. cursors will be created by default unless
6577 overridden by individual drivers. 0 will hide
6578 cursors, 1 will display them.
6580 vt.italic= [VT] Default color for italic text; 0-15.
6583 vt.underline= [VT] Default color for underlined text; 0-15.
6586 watchdog timers [HW,WDT] For information on watchdog timers,
6587 see Documentation/watchdog/watchdog-parameters.rst
6588 or other driver-specific files in the
6589 Documentation/watchdog/ directory.
6593 Set the hard lockup detector stall duration
6594 threshold in seconds. The soft lockup detector
6595 threshold is set to twice the value. A value of 0
6596 disables both lockup detectors. Default is 10
6599 workqueue.watchdog_thresh=
6600 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6601 warn stall conditions and dump internal state to
6602 help debugging. 0 disables workqueue stall
6603 detection; otherwise, it's the stall threshold
6604 duration in seconds. The default value is 30 and
6605 it can be updated at runtime by writing to the
6606 corresponding sysfs file.
6608 workqueue.disable_numa
6609 By default, all work items queued to unbound
6610 workqueues are affine to the NUMA nodes they're
6611 issued on, which results in better behavior in
6612 general. If NUMA affinity needs to be disabled for
6613 whatever reason, this option can be used. Note
6614 that this also can be controlled per-workqueue for
6615 workqueues visible under /sys/bus/workqueue/.
6617 workqueue.power_efficient
6618 Per-cpu workqueues are generally preferred because
6619 they show better performance thanks to cache
6620 locality; unfortunately, per-cpu workqueues tend to
6621 be more power hungry than unbound workqueues.
6623 Enabling this makes the per-cpu workqueues which
6624 were observed to contribute significantly to power
6625 consumption unbound, leading to measurably lower
6626 power usage at the cost of small performance
6629 The default value of this parameter is determined by
6630 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6632 workqueue.debug_force_rr_cpu
6633 Workqueue used to implicitly guarantee that work
6634 items queued without explicit CPU specified are put
6635 on the local CPU. This guarantee is no longer true
6636 and while local CPU is still preferred work items
6637 may be put on foreign CPUs. This debug option
6638 forces round-robin CPU selection to flush out
6639 usages which depend on the now broken guarantee.
6640 When enabled, memory and cache locality will be
6643 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6644 default x2apic cluster mode on platforms
6647 xen_512gb_limit [KNL,X86-64,XEN]
6648 Restricts the kernel running paravirtualized under Xen
6649 to use only up to 512 GB of RAM. The reason to do so is
6650 crash analysis tools and Xen tools for doing domain
6651 save/restore/migration must be enabled to handle larger
6654 xen_emul_unplug= [HW,X86,XEN]
6655 Unplug Xen emulated devices
6656 Format: [unplug0,][unplug1]
6657 ide-disks -- unplug primary master IDE devices
6658 aux-ide-disks -- unplug non-primary-master IDE devices
6659 nics -- unplug network devices
6660 all -- unplug all emulated devices (NICs and IDE disks)
6661 unnecessary -- unplugging emulated devices is
6662 unnecessary even if the host did not respond to
6664 never -- do not unplug even if version check succeeds
6666 xen_legacy_crash [X86,XEN]
6667 Crash from Xen panic notifier, without executing late
6668 panic() code such as dumping handler.
6670 xen_nopvspin [X86,XEN]
6671 Disables the qspinlock slowpath using Xen PV optimizations.
6672 This parameter is obsoleted by "nopvspin" parameter, which
6673 has equivalent effect for XEN platform.
6676 Disables the PV optimizations forcing the HVM guest to
6677 run as generic HVM guest with no PV drivers.
6678 This option is obsoleted by the "nopv" option, which
6679 has equivalent effect for XEN platform.
6681 xen_no_vector_callback
6682 [KNL,X86,XEN] Disable the vector callback for Xen
6683 event channel interrupts.
6685 xen_scrub_pages= [XEN]
6686 Boolean option to control scrubbing pages before giving them back
6687 to Xen, for use by other domains. Can be also changed at runtime
6688 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6689 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6691 xen_timer_slop= [X86-64,XEN]
6692 Set the timer slop (in nanoseconds) for the virtual Xen
6693 timers (default is 100000). This adjusts the minimum
6694 delta of virtualized Xen timers, where lower values
6695 improve timer resolution at the expense of processing
6696 more timer interrupts.
6698 xen.balloon_boot_timeout= [XEN]
6699 The time (in seconds) to wait before giving up to boot
6700 in case initial ballooning fails to free enough memory.
6701 Applies only when running as HVM or PVH guest and
6702 started with less memory configured than allowed at
6703 max. Default is 180.
6705 xen.event_eoi_delay= [XEN]
6706 How long to delay EOI handling in case of event
6707 storms (jiffies). Default is 10.
6709 xen.event_loop_timeout= [XEN]
6710 After which time (jiffies) the event handling loop
6711 should start to delay EOI handling. Default is 2.
6713 xen.fifo_events= [XEN]
6714 Boolean parameter to disable using fifo event handling
6715 even if available. Normally fifo event handling is
6716 preferred over the 2-level event handling, as it is
6717 fairer and the number of possible event channels is
6718 much higher. Default is on (use fifo events).
6720 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6721 Disables the PV optimizations forcing the guest to run
6722 as generic guest with no PV drivers. Currently support
6723 XEN HVM, KVM, HYPER_V and VMWARE guest.
6725 nopvspin [X86,XEN,KVM]
6726 Disables the qspinlock slow path using PV optimizations
6727 which allow the hypervisor to 'idle' the guest on lock
6730 xirc2ps_cs= [NET,PCMCIA]
6732 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6735 By default on POWER9 and above, the kernel will
6736 natively use the XIVE interrupt controller. This option
6737 allows the fallback firmware mode to be used:
6739 off Fallback to firmware control of XIVE interrupt
6740 controller on both pseries and powernv
6741 platforms. Only useful on POWER9 and above.
6743 xive.store-eoi=off [PPC]
6744 By default on POWER10 and above, the kernel will use
6745 stores for EOI handling when the XIVE interrupt mode
6746 is active. This option allows the XIVE driver to use
6747 loads instead, as on POWER9.
6749 xhci-hcd.quirks [USB,KNL]
6750 A hex value specifying bitmask with supplemental xhci
6751 host controller quirks. Meaning of each bit can be
6752 consulted in header drivers/usb/host/xhci.h.
6755 Format: { early | on | rw | ro | off }
6756 Controls if xmon debugger is enabled. Default is off.
6757 Passing only "xmon" is equivalent to "xmon=early".
6758 early Call xmon as early as possible on boot; xmon
6759 debugger is called from setup_arch().
6760 on xmon debugger hooks will be installed so xmon
6761 is only called on a kernel crash. Default mode,
6762 i.e. either "ro" or "rw" mode, is controlled
6763 with CONFIG_XMON_DEFAULT_RO_MODE.
6764 rw xmon debugger hooks will be installed so xmon
6765 is called only on a kernel crash, mode is write,
6766 meaning SPR registers, memory and, other data
6767 can be written using xmon commands.
6768 ro same as "rw" option above but SPR registers,
6769 memory, and other data can't be written using
6771 off xmon is disabled.