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_nohwsig,
229 old_ordering, nonvs, sci_force_enable, nobl }
230 See Documentation/power/video.rst for information on
232 s3_beep is for debugging; it makes the PC's speaker beep
233 as soon as the kernel's real-mode entry point is called.
234 s4_nohwsig prevents ACPI hardware signature from being
235 used during resume from hibernation.
236 old_ordering causes the ACPI 1.0 ordering of the _PTS
237 control method, with respect to putting devices into
238 low power states, to be enforced (the ACPI 2.0 ordering
239 of _PTS is used by default).
240 nonvs prevents the kernel from saving/restoring the
241 ACPI NVS memory during suspend/hibernation and resume.
242 sci_force_enable causes the kernel to set SCI_EN directly
243 on resume from S1/S3 (which is against the ACPI spec,
244 but some broken systems don't work without it).
245 nobl causes the internal blacklist of systems known to
246 behave incorrectly in some ways with respect to system
247 suspend and resume to be ignored (use wisely).
249 acpi_use_timer_override [HW,ACPI]
250 Use timer override. For some broken Nvidia NF5 boards
251 that require a timer override, but don't have HPET
253 add_efi_memmap [EFI; X86] Include EFI memory map in
254 kernel's map of available physical RAM.
257 { off | try_unsupported }
258 off: disable AGP support
259 try_unsupported: try to drive unsupported chipsets
260 (may crash computer or cause data corruption)
263 See Documentation/sound/alsa-configuration.rst
266 Allow the default userspace alignment fault handler
267 behaviour to be specified. Bit 0 enables warnings,
268 bit 1 enables fixups, and bit 2 sends a segfault.
270 align_va_addr= [X86-64]
271 Align virtual addresses by clearing slice [14:12] when
272 allocating a VMA at process creation time. This option
273 gives you up to 3% performance improvement on AMD F15h
274 machines (where it is enabled by default) for a
275 CPU-intensive style benchmark, and it can vary highly in
276 a microbenchmark depending on workload and compiler.
278 32: only for 32-bit processes
279 64: only for 64-bit processes
280 on: enable for both 32- and 64-bit processes
281 off: disable for both 32- and 64-bit processes
283 alloc_snapshot [FTRACE]
284 Allocate the ftrace snapshot buffer on boot up when the
285 main buffer is allocated. This is handy if debugging
286 and you need to use tracing_snapshot() on boot up, and
287 do not want to use tracing_snapshot_alloc() as it needs
288 to be done where GFP_KERNEL allocations are allowed.
290 allow_mismatched_32bit_el0 [ARM64]
291 Allow execve() of 32-bit applications and setting of the
292 PER_LINUX32 personality on systems where only a strict
293 subset of the CPUs support 32-bit EL0. When this
294 parameter is present, the set of CPUs supporting 32-bit
295 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
296 and hot-unplug operations may be restricted.
298 See Documentation/arm64/asymmetric-32bit.rst for more
301 amd_iommu= [HW,X86-64]
302 Pass parameters to the AMD IOMMU driver in the system.
304 fullflush - Deprecated, equivalent to iommu.strict=1
305 off - do not initialize any AMD IOMMU found in
307 force_isolation - Force device isolation for all
308 devices. The IOMMU driver is not
309 allowed anymore to lift isolation
310 requirements as needed. This option
311 does not override iommu=pt
312 force_enable - Force enable the IOMMU on platforms known
313 to be buggy with IOMMU enabled. Use this
316 amd_iommu_dump= [HW,X86-64]
317 Enable AMD IOMMU driver option to dump the ACPI table
318 for AMD IOMMU. With this option enabled, AMD IOMMU
319 driver will print ACPI tables for AMD IOMMU during
320 IOMMU initialization.
322 amd_iommu_intr= [HW,X86-64]
323 Specifies one of the following AMD IOMMU interrupt
325 legacy - Use legacy interrupt remapping mode.
326 vapic - Use virtual APIC mode, which allows IOMMU
327 to inject interrupts directly into guest.
328 This mode requires kvm-amd.avic=1.
329 (Default when IOMMU HW support is present.)
331 amijoy.map= [HW,JOY] Amiga joystick support
332 Map of devices attached to JOY0DAT and JOY1DAT
334 See also Documentation/input/joydev/joystick.rst
336 analog.map= [HW,JOY] Analog joystick and gamepad support
337 Specifies type or capabilities of an analog joystick
338 connected to one of 16 gameports
339 Format: <type1>,<type2>,..<type16>
342 Power management functions (SPARCstation-4/5 + deriv.)
344 Disable APC CPU standby support. SPARCstation-Fox does
345 not play well with APC CPU idle - disable it if you have
346 APC and your system crashes randomly.
348 apic= [APIC,X86] Advanced Programmable Interrupt Controller
349 Change the output verbosity while booting
350 Format: { quiet (default) | verbose | debug }
351 Change the amount of debugging information output
352 when initialising the APIC and IO-APIC components.
353 For X86-32, this can also be used to specify an APIC
355 Format: apic=driver_name
356 Examples: apic=bigsmp
358 apic_extnmi= [APIC,X86] External NMI delivery setting
359 Format: { bsp (default) | all | none }
360 bsp: External NMI is delivered only to CPU 0
361 all: External NMIs are broadcast to all CPUs as a
363 none: External NMI is masked for all CPUs. This is
364 useful so that a dump capture kernel won't be
368 See Documentation/networking/ipv6.rst.
370 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
371 Limit apic dumping. The parameter defines the maximal
372 number of local apics being dumped. Also it is possible
373 to set it to "all" by meaning -- no limit here.
374 Format: { 1 (default) | 2 | ... | all }.
375 The parameter valid if only apic=debug or
376 apic=verbose is specified.
377 Example: apic=debug show_lapic=all
379 apm= [APM] Advanced Power Management
380 See header of arch/x86/kernel/apm_32.c.
382 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
383 Format: <io>,<irq>,<nodeID>
385 arm64.nobti [ARM64] Unconditionally disable Branch Target
386 Identification support
388 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
391 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
396 atarimouse= [HW,MOUSE] Atari Mouse
398 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
399 EzKey and similar keyboards
401 atkbd.reset= [HW] Reset keyboard during initialization
403 atkbd.set= [HW] Select keyboard code set
404 Format: <int> (2 = AT (default), 3 = PS/2)
406 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
409 atkbd.softraw= [HW] Choose between synthetic and real raw mode
410 Format: <bool> (0 = real, 1 = synthetic (default))
412 atkbd.softrepeat= [HW]
413 Use software keyboard repeat
415 audit= [KNL] Enable the audit sub-system
416 Format: { "0" | "1" | "off" | "on" }
417 0 | off - kernel audit is disabled and can not be
418 enabled until the next reboot
419 unset - kernel audit is initialized but disabled and
420 will be fully enabled by the userspace auditd.
421 1 | on - kernel audit is initialized and partially
422 enabled, storing at most audit_backlog_limit
423 messages in RAM until it is fully enabled by the
427 audit_backlog_limit= [KNL] Set the audit queue size limit.
428 Format: <int> (must be >=0)
431 bau= [X86_UV] Enable the BAU on SGI UV. The default
432 behavior is to disable the BAU (i.e. bau=0).
433 Format: { "0" | "1" }
436 unset - Disable the BAU.
438 baycom_epp= [HW,AX25]
441 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
443 See header of drivers/net/hamradio/baycom_par.c.
445 baycom_ser_fdx= [HW,AX25]
446 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
447 Format: <io>,<irq>,<mode>[,<baud>]
448 See header of drivers/net/hamradio/baycom_ser_fdx.c.
450 baycom_ser_hdx= [HW,AX25]
451 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
452 Format: <io>,<irq>,<mode>
453 See header of drivers/net/hamradio/baycom_ser_hdx.c.
455 blkdevparts= Manual partition parsing of block device(s) for
456 embedded devices based on command line input.
457 See Documentation/block/cmdline-partition.rst
459 boot_delay= Milliseconds to delay each printk during boot.
460 Values larger than 10 seconds (10000) are changed to
465 Extended command line options can be added to an initrd
466 and this will cause the kernel to look for it.
468 See Documentation/admin-guide/bootconfig.rst
471 Disable BERT OS support on buggy BIOSes.
473 bgrt_disable [ACPI][X86]
474 Disable BGRT to avoid flickering OEM logo.
476 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
477 bttv.radio= Most important insmod options are available as
479 bttv.pll= See Documentation/admin-guide/media/bttv.rst
482 bulk_remove=off [PPC] This parameter disables the use of the pSeries
483 firmware feature for flushing multiple hpte entries
486 c101= [NET] Moxa C101 synchronous serial card
488 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
489 Sometimes CPU hardware bugs make them report the cache
490 size incorrectly. The kernel will attempt work arounds
491 to fix known problems, but for some CPUs it is not
492 possible to determine what the correct size should be.
493 This option provides an override for these situations.
496 [NET] Specifies amount of time (in seconds) that
497 the kernel should wait for a network carrier. By default
498 it waits 120 seconds.
500 ca_keys= [KEYS] This parameter identifies a specific key(s) on
501 the system trusted keyring to be used for certificate
503 format: { id:<keyid> | builtin }
505 cca= [MIPS] Override the kernel pages' cache coherency
506 algorithm. Accepted values range from 0 to 7
507 inclusive. See arch/mips/include/asm/pgtable-bits.h
508 for platform specific values (SB1, Loongson3 and
511 ccw_timeout_log [S390]
512 See Documentation/s390/common_io.rst for details.
514 cgroup_disable= [KNL] Disable a particular controller or optional feature
515 Format: {name of the controller(s) or feature(s) to disable}
516 The effects of cgroup_disable=foo are:
517 - foo isn't auto-mounted if you mount all cgroups in
519 - foo isn't visible as an individually mountable
521 - if foo is an optional feature then the feature is
522 disabled and corresponding cgroup files are not
524 {Currently only "memory" controller deal with this and
525 cut the overhead, others just disable the usage. So
526 only cgroup_disable=memory is actually worthy}
527 Specifying "pressure" disables per-cgroup pressure
528 stall information accounting feature
530 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
531 Format: { { controller | "all" | "named" }
532 [,{ controller | "all" | "named" }...] }
533 Like cgroup_disable, but only applies to cgroup v1;
534 the blacklisted controllers remain available in cgroup2.
535 "all" blacklists all controllers and "named" disables
536 named mounts. Specifying both "all" and "named" disables
539 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
541 nosocket -- Disable socket memory accounting.
542 nokmem -- Disable kernel memory accounting.
544 checkreqprot [SELINUX] Set initial checkreqprot flag value.
545 Format: { "0" | "1" }
546 See security/selinux/Kconfig help text.
547 0 -- check protection applied by kernel (includes
548 any implied execute protection).
549 1 -- check protection requested by application.
550 Default value is set via a kernel config option.
551 Value can be changed at runtime via
552 /sys/fs/selinux/checkreqprot.
553 Setting checkreqprot to 1 is deprecated.
556 See Documentation/s390/common_io.rst for details.
559 Prevents the clock framework from automatically gating
560 clocks that have not been explicitly enabled by a Linux
561 device driver but are enabled in hardware at reset or
562 by the bootloader/firmware. Note that this does not
563 force such clocks to be always-on nor does it reserve
564 those clocks in any way. This parameter is useful for
565 debug and development, but should not be needed on a
566 platform with proper driver support. For more
567 information, see Documentation/driver-api/clk.rst.
569 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
571 Forces specified clocksource (if available) to be used
572 when calculating gettimeofday(). If specified
573 clocksource is not available, it defaults to PIT.
574 Format: { pit | tsc | cyclone | pmtmr }
576 clocksource= Override the default clocksource
578 Override the default clocksource and use the clocksource
579 with the name specified.
580 Some clocksource names to choose from, depending on
582 [all] jiffies (this is the base, fallback clocksource)
584 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
585 pxa_timer,timer3,32k_counter,timer0_1
586 [X86-32] pit,hpet,tsc;
587 scx200_hrt on Geode; cyclone on IBM x440
595 clocksource.arm_arch_timer.evtstrm=
598 Enable/disable the eventstream feature of the ARM
599 architected timer so that code using WFE-based polling
600 loops can be debugged more effectively on production
603 clocksource.max_cswd_read_retries= [KNL]
604 Number of clocksource_watchdog() retries due to
605 external delays before the clock will be marked
606 unstable. Defaults to three retries, that is,
607 four attempts to read the clock under test.
609 clocksource.verify_n_cpus= [KNL]
610 Limit the number of CPUs checked for clocksources
611 marked with CLOCK_SOURCE_VERIFY_PERCPU that
612 are marked unstable due to excessive skew.
613 A negative value says to check all CPUs, while
614 zero says not to check any. Values larger than
615 nr_cpu_ids are silently truncated to nr_cpu_ids.
616 The actual CPUs are chosen randomly, with
617 no replacement if the same CPU is chosen twice.
619 clocksource-wdtest.holdoff= [KNL]
620 Set the time in seconds that the clocksource
621 watchdog test waits before commencing its tests.
622 Defaults to zero when built as a module and to
623 10 seconds when built into the kernel.
625 clearcpuid=BITNUM[,BITNUM...] [X86]
626 Disable CPUID feature X for the kernel. See
627 arch/x86/include/asm/cpufeatures.h for the valid bit
628 numbers. Note the Linux specific bits are not necessarily
629 stable over kernel options, but the vendor specific
631 Also note that user programs calling CPUID directly
632 or using the feature without checking anything
633 will still see it. This just prevents it from
634 being used by the kernel or shown in /proc/cpuinfo.
635 Also note the kernel might malfunction if you disable
638 cma=nn[MG]@[start[MG][-end[MG]]]
640 Sets the size of kernel global memory area for
641 contiguous memory allocations and optionally the
642 placement constraint by the physical address range of
643 memory allocations. A value of 0 disables CMA
644 altogether. For more information, see
645 kernel/dma/contiguous.c
649 Sets the size of kernel per-numa memory area for
650 contiguous memory allocations. A value of 0 disables
651 per-numa CMA altogether. And If this option is not
652 specificed, the default value is 0.
653 With per-numa CMA enabled, DMA users on node nid will
654 first try to allocate buffer from the pernuma area
655 which is located in node nid, if the allocation fails,
656 they will fallback to the global default memory area.
658 cmo_free_hint= [PPC] Format: { yes | no }
659 Specify whether pages are marked as being inactive
660 when they are freed. This is used in CMO environments
661 to determine OS memory pressure for page stealing by
665 coherent_pool=nn[KMG] [ARM,KNL]
666 Sets the size of memory pool for coherent, atomic dma
667 allocations, by default set to 256K.
669 com20020= [HW,NET] ARCnet - COM20020 chipset
671 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
673 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
677 ARCnet - COM90xx chipset (memory-mapped buffers)
678 Format: <io>[,<irq>[,<memstart>]]
680 condev= [HW,S390] console device
683 console= [KNL] Output console device and options.
685 tty<n> Use the virtual console device <n>.
689 Use the specified serial port. The options are of
690 the form "bbbbpnf", where "bbbb" is the baud rate,
691 "p" is parity ("n", "o", or "e"), "n" is number of
692 bits, and "f" is flow control ("r" for RTS or
693 omit it). Default is "9600n8".
695 See Documentation/admin-guide/serial-console.rst for more
697 Documentation/networking/netconsole.rst for an
700 uart[8250],io,<addr>[,options]
701 uart[8250],mmio,<addr>[,options]
702 uart[8250],mmio16,<addr>[,options]
703 uart[8250],mmio32,<addr>[,options]
704 uart[8250],0x<addr>[,options]
705 Start an early, polled-mode console on the 8250/16550
706 UART at the specified I/O port or MMIO address,
707 switching to the matching ttyS device later.
708 MMIO inter-register address stride is either 8-bit
709 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
710 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
711 to be equivalent to 'mmio'. 'options' are specified in
712 the same format described for ttyS above; if unspecified,
713 the h/w is not re-initialized.
715 hvc<n> Use the hypervisor console device <n>. This is for
716 both Xen and PowerPC hypervisors.
718 If the device connected to the port is not a TTY but a braille
719 device, prepend "brl," before the device type, for instance
721 For now, only VisioBraille is supported.
724 [KNL] Change console messages format
726 By default we print messages on consoles in
727 "[time stamp] text\n" format (time stamp may not be
728 printed, depending on CONFIG_PRINTK_TIME or
729 `printk_time' param).
731 Switch to syslog format: "<%u>[time stamp] text\n"
732 IOW, each message will have a facility and loglevel
733 prefix. The format is similar to one used by syslog()
734 syscall, or to executing "dmesg -S --raw" or to reading
737 consoleblank= [KNL] The console blank (screen saver) timeout in
738 seconds. A value of 0 disables the blank timer.
742 [KNL] Change the default value for
743 /proc/<pid>/coredump_filter.
744 See also Documentation/filesystems/proc.rst.
746 coresight_cpu_debug.enable
749 Enable/disable the CPU sampling based debugging.
750 0: default value, disable debugging
751 1: enable debugging at boot time
753 cpuidle.off=1 [CPU_IDLE]
754 disable the cpuidle sub-system
757 [CPU_IDLE] Name of the cpuidle governor to use.
759 cpufreq.off=1 [CPU_FREQ]
760 disable the cpufreq sub-system
762 cpufreq.default_governor=
763 [CPU_FREQ] Name of the default cpufreq governor or
764 policy to use. This governor must be registered in the
765 kernel before the cpufreq driver probes.
768 [X86] Delay for N microsec between assert and de-assert
769 of APIC INIT to start processors. This delay occurs
770 on every CPU online, such as boot, and resume from suspend.
773 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
775 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
777 crashkernel=size[KMG][@offset[KMG]]
778 [KNL] Using kexec, Linux can switch to a 'crash kernel'
779 upon panic. This parameter reserves the physical
780 memory region [offset, offset + size] for that kernel
781 image. If '@offset' is omitted, then a suitable offset
782 is selected automatically.
783 [KNL, X86-64] Select a region under 4G first, and
784 fall back to reserve region above 4G when '@offset'
785 hasn't been specified.
786 See Documentation/admin-guide/kdump/kdump.rst for further details.
788 crashkernel=range1:size1[,range2:size2,...][@offset]
789 [KNL] Same as above, but depends on the memory
790 in the running system. The syntax of range is
791 start-[end] where start and end are both
792 a memory unit (amount[KMG]). See also
793 Documentation/admin-guide/kdump/kdump.rst for an example.
795 crashkernel=size[KMG],high
796 [KNL, X86-64] range could be above 4G. Allow kernel
797 to allocate physical memory region from top, so could
798 be above 4G if system have more than 4G ram installed.
799 Otherwise memory region will be allocated below 4G, if
801 It will be ignored if crashkernel=X is specified.
802 crashkernel=size[KMG],low
803 [KNL, X86-64] range under 4G. When crashkernel=X,high
804 is passed, kernel could allocate physical memory region
805 above 4G, that cause second kernel crash on system
806 that require some amount of low memory, e.g. swiotlb
807 requires at least 64M+32K low memory, also enough extra
808 low memory is needed to make sure DMA buffers for 32-bit
809 devices won't run out. Kernel would try to allocate at
810 at least 256M below 4G automatically.
811 This one let user to specify own low range under 4G
812 for second kernel instead.
813 0: to disable low allocation.
814 It will be ignored when crashkernel=X,high is not used
815 or memory reserved is below 4G.
818 [KNL] Disable crypto self-tests
823 cs89x0_media= [HW,NET]
824 Format: { rj45 | aui | bnc }
826 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
827 handling. When switched on, additional debug data is
828 printed to the console in case a hanging CPU is
829 detected, and that CPU is pinged again in order to try
830 to resolve the hang situation.
831 0: disable csdlock debugging (default)
832 1: enable basic csdlock debugging (minor impact)
833 ext: enable extended csdlock debugging (more impact,
837 See header of drivers/s390/block/dasd_devmap.c.
839 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
840 (one device per port)
841 Format: <port#>,<type>
842 See also Documentation/input/devices/joystick-parport.rst
844 debug [KNL] Enable kernel debugging (events log level).
847 [KNL] Enable printing [hashed] pointers early in the
848 boot sequence. If enabled, we use a weak hash instead
849 of siphash to hash pointers. Use this option if you are
850 seeing instances of '(___ptrval___)') and need to see a
851 value (hashed pointer) instead. Cryptographically
852 insecure, please do not use on production kernels.
855 [KNL] verbose locking self-tests
857 Print debugging info while doing the locking API
859 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
860 (no extra messages), setting it to -1 (all bits set)
861 will print _a_lot_ more information - normally only
862 useful to lockdep developers.
864 debug_objects [KNL] Enable object debugging
867 [KNL] Disable object debugging
869 debug_guardpage_minorder=
870 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
871 parameter allows control of the order of pages that will
872 be intentionally kept free (and hence protected) by the
873 buddy allocator. Bigger value increase the probability
874 of catching random memory corruption, but reduce the
875 amount of memory for normal system use. The maximum
876 possible value is MAX_ORDER/2. Setting this parameter
877 to 1 or 2 should be enough to identify most random
878 memory corruption problems caused by bugs in kernel or
879 driver code when a CPU writes to (or reads from) a
880 random memory location. Note that there exists a class
881 of memory corruptions problems caused by buggy H/W or
882 F/W or by drivers badly programing DMA (basically when
883 memory is written at bus level and the CPU MMU is
884 bypassed) which are not detectable by
885 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
886 tracking down these problems.
889 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
890 enables the feature at boot time. By default, it is
891 disabled and the system will work mostly the same as a
892 kernel built without CONFIG_DEBUG_PAGEALLOC.
893 Note: to get most of debug_pagealloc error reports, it's
894 useful to also enable the page_owner functionality.
895 on: enable the feature
897 debugfs= [KNL] This parameter enables what is exposed to userspace
898 and debugfs internal clients.
899 Format: { on, no-mount, off }
900 on: All functions are enabled.
902 Filesystem is not registered but kernel clients can
903 access APIs and a crashkernel can be used to read
904 its content. There is nothing to mount.
905 off: Filesystem is not registered and clients
906 get a -EPERM as result when trying to register files
907 or directories within debugfs.
908 This is equivalent of the runtime functionality if
909 debugfs was not enabled in the kernel at all.
910 Default value is set in build-time with a kernel configuration.
912 debugpat [X86] Enable PAT debugging
914 decnet.addr= [HW,NET]
915 Format: <area>[,<node>]
916 See also Documentation/networking/decnet.rst.
919 [HW] The size of the default HugeTLB page. This is
920 the size represented by the legacy /proc/ hugepages
921 APIs. In addition, this is the default hugetlb size
922 used for shmget(), mmap() and mounting hugetlbfs
923 filesystems. If not specified, defaults to the
924 architecture's default huge page size. Huge page
925 sizes are architecture dependent. See also
926 Documentation/admin-guide/mm/hugetlbpage.rst.
929 deferred_probe_timeout=
930 [KNL] Debugging option to set a timeout in seconds for
931 deferred probe to give up waiting on dependencies to
932 probe. Only specific dependencies (subsystems or
933 drivers) that have opted in will be ignored. A timeout of 0
934 will timeout at the end of initcalls. This option will also
935 dump out devices still on the deferred probe list after
939 Format: { on | off | def_only | inf_only | always }
940 on: s390 zlib hardware support for compression on
941 level 1 and decompression (default)
942 off: No s390 zlib hardware support
943 def_only: s390 zlib hardware support for deflate
944 only (compression on level 1)
945 inf_only: s390 zlib hardware support for inflate
947 always: Same as 'on' but ignores the selected compression
948 level always using hardware support (used for debugging)
951 Set number of hash buckets for dentry cache.
953 disable_1tb_segments [PPC]
954 Disables the use of 1TB hash page table segments. This
955 causes the kernel to fall back to 256MB segments which
956 can be useful when debugging issues that require an SLB
960 Limits the number of kernel SLB entries, and flushes
961 them frequently to increase the rate of SLB faults
965 See Documentation/networking/ipv6.rst.
968 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
969 hardening is enabled for this boot. Hardened
970 usercopy checking is used to protect the kernel
971 from reading or writing beyond known memory
972 allocation boundaries as a proactive defense
973 against bounds-checking flaws in the kernel's
974 copy_to_user()/copy_from_user() interface.
975 on Perform hardened usercopy checks (default).
976 off Disable hardened usercopy checks.
979 Disable RADIX MMU mode on POWER9
981 radix_hcall_invalidate=on [PPC/PSERIES]
982 Disable RADIX GTSE feature and use hcall for TLB
986 Disable TLBIE instruction. Currently does not work
987 with KVM, with HASH MMU, or with coherent accelerators.
989 disable_cpu_apicid= [X86,APIC,SMP]
991 The number of initial APIC ID for the
992 corresponding CPU to be disabled at boot,
993 mostly used for the kdump 2nd kernel to
994 disable BSP to wake up multiple CPUs without
995 causing system reset or hang due to sending
998 disable_ddw [PPC/PSERIES]
999 Disable Dynamic DMA Window support. Use this
1000 to workaround buggy firmware.
1002 disable_ipv6= [IPV6]
1003 See Documentation/networking/ipv6.rst.
1005 disable_mtrr_cleanup [X86]
1006 The kernel tries to adjust MTRR layout from continuous
1007 to discrete, to make X server driver able to add WB
1008 entry later. This parameter disables that.
1010 disable_mtrr_trim [X86, Intel and AMD only]
1011 By default the kernel will trim any uncacheable
1012 memory out of your available memory pool based on
1013 MTRR settings. This parameter disables that behavior,
1014 possibly causing your machine to run very slowly.
1016 disable_timer_pin_1 [X86]
1017 Disable PIN 1 of APIC timer
1018 Can be useful to work around chipset bugs.
1020 dis_ucode_ldr [X86] Disable the microcode loader.
1022 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1023 this option disables the debugging code at boot.
1025 dma_debug_entries=<number>
1026 This option allows to tune the number of preallocated
1027 entries for DMA-API debugging code. One entry is
1028 required per DMA-API allocation. Use this if the
1029 DMA-API debugging code disables itself because the
1030 architectural default is too low.
1032 dma_debug_driver=<driver_name>
1033 With this option the DMA-API debugging driver
1034 filter feature can be enabled at boot time. Just
1035 pass the driver to filter for as the parameter.
1036 The filter can be disabled or changed to another
1037 driver later using sysfs.
1039 driver_async_probe= [KNL]
1040 List of driver names to be probed asynchronously.
1041 Format: <driver_name1>,<driver_name2>...
1043 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1044 Broken monitors, graphic adapters, KVMs and EDIDless
1045 panels may send no or incorrect EDID data sets.
1046 This parameter allows to specify an EDID data sets
1047 in the /lib/firmware directory that are used instead.
1048 Generic built-in EDID data sets are used, if one of
1049 edid/1024x768.bin, edid/1280x1024.bin,
1050 edid/1680x1050.bin, or edid/1920x1080.bin is given
1051 and no file with the same name exists. Details and
1052 instructions how to build your own EDID data are
1053 available in Documentation/admin-guide/edid.rst. An EDID
1054 data set will only be used for a particular connector,
1055 if its name and a colon are prepended to the EDID
1056 name. Each connector may use a unique EDID data
1057 set by separating the files with a comma. An EDID
1058 data set with no connector name will be used for
1059 any connectors not explicitly specified.
1064 Format: {"off" | "known"}
1065 Control how the dt_cpu_ftrs device-tree binding is
1066 used for CPU feature discovery and setup (if it
1068 off: Do not use it, fall back to legacy cpu table.
1069 known: Do not pass through unknown features to guests
1070 or userspace, only those that the kernel is aware of.
1072 dump_apple_properties [X86]
1073 Dump name and content of EFI device properties on
1074 x86 Macs. Useful for driver authors to determine
1075 what data is available or for reverse-engineering.
1077 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1078 <module>.dyndbg[="val"]
1079 Enable debug messages at boot time. See
1080 Documentation/admin-guide/dynamic-debug-howto.rst
1083 nopku [X86] Disable Memory Protection Keys CPU feature found
1086 <module>.async_probe [KNL]
1087 Enable asynchronous probe on this module.
1089 early_ioremap_debug [KNL]
1090 Enable debug messages in early_ioremap support. This
1091 is useful for tracking down temporary early mappings
1092 which are not unmapped.
1094 earlycon= [KNL] Output early console device and options.
1096 When used with no options, the early console is
1097 determined by stdout-path property in device tree's
1098 chosen node or the ACPI SPCR table if supported by
1101 cdns,<addr>[,options]
1102 Start an early, polled-mode console on a Cadence
1103 (xuartps) serial port at the specified address. Only
1104 supported option is baud rate. If baud rate is not
1105 specified, the serial port must already be setup and
1108 uart[8250],io,<addr>[,options]
1109 uart[8250],mmio,<addr>[,options]
1110 uart[8250],mmio32,<addr>[,options]
1111 uart[8250],mmio32be,<addr>[,options]
1112 uart[8250],0x<addr>[,options]
1113 Start an early, polled-mode console on the 8250/16550
1114 UART at the specified I/O port or MMIO address.
1115 MMIO inter-register address stride is either 8-bit
1116 (mmio) or 32-bit (mmio32 or mmio32be).
1117 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1118 to be equivalent to 'mmio'. 'options' are specified
1119 in the same format described for "console=ttyS<n>"; if
1120 unspecified, the h/w is not initialized.
1124 Start an early, polled-mode console on a pl011 serial
1125 port at the specified address. The pl011 serial port
1126 must already be setup and configured. Options are not
1127 yet supported. If 'mmio32' is specified, then only
1128 the driver will use only 32-bit accessors to read/write
1129 the device registers.
1132 Start an early console on a litex serial port at the
1133 specified address. The serial port must already be
1134 setup and configured. Options are not yet supported.
1137 Start an early, polled-mode console on a meson serial
1138 port at the specified address. The serial port must
1139 already be setup and configured. Options are not yet
1143 Start an early, polled-mode console on an msm serial
1144 port at the specified address. The serial port
1145 must already be setup and configured. Options are not
1148 msm_serial_dm,<addr>
1149 Start an early, polled-mode console on an msm serial
1150 dm port at the specified address. The serial port
1151 must already be setup and configured. Options are not
1155 Start an early, polled-mode console on a serial port
1156 of an Actions Semi SoC, such as S500 or S900, at the
1157 specified address. The serial port must already be
1158 setup and configured. Options are not yet supported.
1161 Start an early, polled-mode console on a serial port
1162 of an RDA Micro SoC, such as RDA8810PL, at the
1163 specified address. The serial port must already be
1164 setup and configured. Options are not yet supported.
1167 Use RISC-V SBI (Supervisor Binary Interface) for early
1170 smh Use ARM semihosting calls for early console.
1178 Use early console provided by serial driver available
1179 on Samsung SoCs, requires selecting proper type and
1180 a correct base address of the selected UART port. The
1181 serial port must already be setup and configured.
1182 Options are not yet supported.
1185 Start an early, polled-mode console on a lantiq serial
1186 (lqasc) port at the specified address. The serial port
1187 must already be setup and configured. Options are not
1192 Use early console provided by Freescale LP UART driver
1193 found on Freescale Vybrid and QorIQ LS1021A processors.
1194 A valid base address must be provided, and the serial
1195 port must already be setup and configured.
1199 Start an early, polled-mode, output-only console on the
1200 Freescale i.MX UART at the specified address. The UART
1201 must already be setup and configured.
1204 Start an early, polled-mode console on the
1205 Armada 3700 serial port at the specified
1206 address. The serial port must already be setup
1207 and configured. Options are not yet supported.
1210 Start an early, polled-mode console on a Qualcomm
1211 Generic Interface (GENI) based serial port at the
1212 specified address. The serial port must already be
1213 setup and configured. Options are not yet supported.
1216 Start an early, unaccelerated console on the EFI
1217 memory mapped framebuffer (if available). On cache
1218 coherent non-x86 systems that use system memory for
1219 the framebuffer, pass the 'ram' option so that it is
1220 mapped with the correct attributes.
1223 Use early console provided by Freescale LINFlexD UART
1224 serial driver for NXP S32V234 SoCs. A valid base
1225 address must be provided, and the serial port must
1226 already be setup and configured.
1228 earlyprintk= [X86,SH,ARM,M68k,S390]
1232 earlyprintk=serial[,ttySn[,baudrate]]
1233 earlyprintk=serial[,0x...[,baudrate]]
1234 earlyprintk=ttySn[,baudrate]
1235 earlyprintk=dbgp[debugController#]
1236 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1237 earlyprintk=xdbc[xhciController#]
1239 earlyprintk is useful when the kernel crashes before
1240 the normal console is initialized. It is not enabled by
1241 default because it has some cosmetic problems.
1243 Append ",keep" to not disable it when the real console
1246 Only one of vga, efi, serial, or usb debug port can
1249 Currently only ttyS0 and ttyS1 may be specified by
1250 name. Other I/O ports may be explicitly specified
1251 on some architectures (x86 and arm at least) by
1252 replacing ttySn with an I/O port address, like this:
1253 earlyprintk=serial,0x1008,115200
1254 You can find the port for a given device in
1255 /proc/tty/driver/serial:
1256 2: uart:ST16650V2 port:00001008 irq:18 ...
1258 Interaction with the standard serial driver is not
1261 The VGA and EFI output is eventually overwritten by
1264 The xen option can only be used in Xen domains.
1266 The sclp output can only be used on s390.
1268 The optional "force" to "pciserial" enables use of a
1269 PCI device even when its classcode is not of the
1272 edac_report= [HW,EDAC] Control how to report EDAC event
1273 Format: {"on" | "off" | "force"}
1274 on: enable EDAC to report H/W event. May be overridden
1275 by other higher priority error reporting module.
1276 off: disable H/W event reporting through EDAC.
1277 force: enforce the use of EDAC to report H/W event.
1280 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1283 This is designed to be used in conjunction with
1284 the boot argument: earlyprintk=vga
1286 This parameter works in place of the kgdboc parameter
1287 but can only be used if the backing tty is available
1288 very early in the boot process. For early debugging
1289 via a serial port see kgdboc_earlycon instead.
1292 Format: {"off" | "on" | "skip[mbr]"}
1295 Format: { "debug", "disable_early_pci_dma",
1296 "nochunk", "noruntime", "nosoftreserve",
1297 "novamap", "no_disable_early_pci_dma" }
1298 debug: enable misc debug output.
1299 disable_early_pci_dma: disable the busmaster bit on all
1300 PCI bridges while in the EFI boot stub.
1301 nochunk: disable reading files in "chunks" in the EFI
1302 boot stub, as chunking can cause problems with some
1303 firmware implementations.
1304 noruntime : disable EFI runtime services support
1305 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1306 attribute may cause the kernel to reserve the
1307 memory range for a memory mapping driver to
1308 claim. Specify efi=nosoftreserve to disable this
1309 reservation and treat the memory by its base type
1310 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1311 novamap: do not call SetVirtualAddressMap().
1312 no_disable_early_pci_dma: Leave the busmaster bit set
1313 on all PCI bridges while in the EFI boot stub
1315 efi_no_storage_paranoia [EFI; X86]
1316 Using this parameter you can use more than 50% of
1317 your efi variable storage. Use this parameter only if
1318 you are really sure that your UEFI does sane gc and
1319 fulfills the spec otherwise your board may brick.
1321 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1322 Add arbitrary attribute to specific memory range by
1323 updating original EFI memory map.
1324 Region of memory which aa attribute is added to is
1327 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1328 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1329 attribute is added to range 0x100000000-0x180000000 and
1330 0x10a0000000-0x1120000000.
1332 If efi_fake_mem=8G@9G:0x40000 is specified, the
1333 EFI_MEMORY_SP(0x40000) attribute is added to
1334 range 0x240000000-0x43fffffff.
1336 Using this parameter you can do debugging of EFI memmap
1337 related features. For example, you can do debugging of
1338 Address Range Mirroring feature even if your box
1339 doesn't support it, or mark specific memory as
1342 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1343 that is to be dynamically loaded by Linux. If there are
1344 multiple variables with the same name but with different
1345 vendor GUIDs, all of them will be loaded. See
1346 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1349 eisa_irq_edge= [PARISC,HW]
1350 See header of drivers/parisc/eisa.c.
1353 See comment before function elanfreq_setup() in
1354 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1356 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1357 Specifies physical address of start of kernel core
1358 image elf header and optionally the size. Generally
1359 kexec loader will pass this option to capture kernel.
1360 See Documentation/admin-guide/kdump/kdump.rst for details.
1362 enable_mtrr_cleanup [X86]
1363 The kernel tries to adjust MTRR layout from continuous
1364 to discrete, to make X server driver able to add WB
1365 entry later. This parameter enables that.
1367 enable_timer_pin_1 [X86]
1368 Enable PIN 1 of APIC timer
1369 Can be useful to work around chipset bugs
1370 (in particular on some ATI chipsets).
1371 The kernel tries to set a reasonable default.
1373 enforcing [SELINUX] Set initial enforcing status.
1375 See security/selinux/Kconfig help text.
1376 0 -- permissive (log only, no denials).
1377 1 -- enforcing (deny and log).
1379 Value can be changed at runtime via
1380 /sys/fs/selinux/enforce.
1383 Disable Error Record Serialization Table (ERST)
1386 ether= [HW,NET] Ethernet cards parameters
1387 This option is obsoleted by the "netdev=" option, which
1388 has equivalent usage. See its documentation for details.
1392 Permit 'security.evm' to be updated regardless of
1393 current integrity status.
1398 fail_make_request=[KNL]
1399 General fault injection mechanism.
1400 Format: <interval>,<probability>,<space>,<times>
1401 See also Documentation/fault-injection/.
1404 Format: { initns | none }
1405 See Documentation/admin-guide/sysctl/net.rst for
1406 fb_tunnels_only_for_init_ns
1409 See Documentation/admin-guide/blockdev/floppy.rst.
1411 force_pal_cache_flush
1412 [IA-64] Avoid check_sal_cache_flush which may hang on
1413 buggy SAL_CACHE_FLUSH implementations. Using this
1414 parameter will force ia64_sal_cache_flush to call
1415 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1418 Forcefully enable Physical Address Extension (PAE).
1419 Many Pentium M systems disable PAE but may have a
1420 functionally usable PAE implementation.
1421 Warning: use of this parameter will taint the kernel
1422 and may cause unknown problems.
1425 [FTRACE] will set and start the specified tracer
1426 as early as possible in order to facilitate early
1429 ftrace_dump_on_oops[=orig_cpu]
1430 [FTRACE] will dump the trace buffers on oops.
1431 If no parameter is passed, ftrace will dump
1432 buffers of all CPUs, but if you pass orig_cpu, it will
1433 dump only the buffer of the CPU that triggered the
1436 ftrace_filter=[function-list]
1437 [FTRACE] Limit the functions traced by the function
1438 tracer at boot up. function-list is a comma-separated
1439 list of functions. This list can be changed at run
1440 time by the set_ftrace_filter file in the debugfs
1443 ftrace_notrace=[function-list]
1444 [FTRACE] Do not trace the functions specified in
1445 function-list. This list can be changed at run time
1446 by the set_ftrace_notrace file in the debugfs
1449 ftrace_graph_filter=[function-list]
1450 [FTRACE] Limit the top level callers functions traced
1451 by the function graph tracer at boot up.
1452 function-list is a comma-separated list of functions
1453 that can be changed at run time by the
1454 set_graph_function file in the debugfs tracing directory.
1456 ftrace_graph_notrace=[function-list]
1457 [FTRACE] Do not trace from the functions specified in
1458 function-list. This list is a comma-separated list of
1459 functions that can be changed at run time by the
1460 set_graph_notrace file in the debugfs tracing directory.
1462 ftrace_graph_max_depth=<uint>
1463 [FTRACE] Used with the function graph tracer. This is
1464 the max depth it will trace into a function. This value
1465 can be changed at run time by the max_graph_depth file
1466 in the tracefs tracing directory. default: 0 (no limit)
1468 fw_devlink= [KNL] Create device links between consumer and supplier
1469 devices by scanning the firmware to infer the
1470 consumer/supplier relationships. This feature is
1471 especially useful when drivers are loaded as modules as
1472 it ensures proper ordering of tasks like device probing
1473 (suppliers first, then consumers), supplier boot state
1474 clean up (only after all consumers have probed),
1475 suspend/resume & runtime PM (consumers first, then
1477 Format: { off | permissive | on | rpm }
1478 off -- Don't create device links from firmware info.
1479 permissive -- Create device links from firmware info
1480 but use it only for ordering boot state clean
1481 up (sync_state() calls).
1482 on -- Create device links from firmware info and use it
1483 to enforce probe and suspend/resume ordering.
1484 rpm -- Like "on", but also use to order runtime PM.
1486 fw_devlink.strict=<bool>
1487 [KNL] Treat all inferred dependencies as mandatory
1488 dependencies. This only applies for fw_devlink=on|rpm.
1492 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1493 support via parallel port (up to 5 devices per port)
1494 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1495 See also Documentation/input/devices/joystick-parport.rst
1499 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1503 gcov_persist= [GCOV] When non-zero (default), profiling data for
1504 kernel modules is saved and remains accessible via
1505 debugfs, even when the module is unloaded/reloaded.
1506 When zero, profiling data is discarded and associated
1507 debugfs files are removed at module unload time.
1509 goldfish [X86] Enable the goldfish android emulator platform.
1510 Don't use this when you are not running on the
1513 gpio-mockup.gpio_mockup_ranges
1514 [HW] Sets the ranges of gpiochip of for this device.
1515 Format: <start1>,<end1>,<start2>,<end2>...
1516 gpio-mockup.gpio_mockup_named_lines
1517 [HW] Let the driver know GPIO lines should be named.
1519 gpt [EFI] Forces disk with valid GPT signature but
1520 invalid Protective MBR to be treated as GPT. If the
1521 primary GPT is corrupted, it enables the backup/alternate
1522 GPT to be used instead.
1524 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1525 the "Enable 0" bit of the configuration register.
1528 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1529 the "Enable 0" bit of the configuration register.
1532 grcan.select= [HW] Select which physical interface to use.
1535 grcan.txsize= [HW] Sets the size of the tx buffer.
1536 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1538 grcan.rxsize= [HW] Sets the size of the rx buffer.
1539 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1542 hardlockup_all_cpu_backtrace=
1543 [KNL] Should the hard-lockup detector generate
1544 backtraces on all cpus.
1547 hashdist= [KNL,NUMA] Large hashes allocated during boot
1548 are distributed across NUMA nodes. Defaults on
1549 for 64-bit NUMA, off otherwise.
1550 Format: 0 | 1 (for off | on)
1552 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1554 hd= [EIDE] (E)IDE hard drive subsystem geometry
1555 Format: <cyl>,<head>,<sect>
1558 Disable Hardware Error Source Table (HEST) support;
1559 corresponding firmware-first mode error processing
1560 logic will be disabled.
1562 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1563 size of <nn>. This works even on boxes that have no
1564 highmem otherwise. This also works to reduce highmem
1565 size on bigger boxes.
1567 highres= [KNL] Enable/disable high resolution timer mode.
1568 Valid parameters: "on", "off"
1573 hpet= [X86-32,HPET] option to control HPET usage
1574 Format: { enable (default) | disable | force |
1576 disable: disable HPET and use PIT instead
1577 force: allow force enabled of undocumented chips (ICH4,
1579 verbose: show contents of HPET registers during setup
1581 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1582 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1584 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1585 of gigantic hugepages. Or using node format, the size
1586 of a CMA area per node can be specified.
1587 Format: nn[KMGTPE] or (node format)
1588 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1590 Reserve a CMA area of given size and allocate gigantic
1591 hugepages using the CMA allocator. If enabled, the
1592 boot-time allocation of gigantic hugepages is skipped.
1594 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1595 If this follows hugepagesz (below), it specifies
1596 the number of pages of hugepagesz to be allocated.
1597 If this is the first HugeTLB parameter on the command
1598 line, it specifies the number of pages to allocate for
1599 the default huge page size. If using node format, the
1600 number of pages to allocate per-node can be specified.
1601 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1602 Format: <integer> or (node format)
1603 <node>:<integer>[,<node>:<integer>]
1606 [HW] The size of the HugeTLB pages. This is used in
1607 conjunction with hugepages (above) to allocate huge
1608 pages of a specific size at boot. The pair
1609 hugepagesz=X hugepages=Y can be specified once for
1610 each supported huge page size. Huge page sizes are
1611 architecture dependent. See also
1612 Documentation/admin-guide/mm/hugetlbpage.rst.
1615 hugetlb_free_vmemmap=
1616 [KNL] Reguires CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
1618 Allows heavy hugetlb users to free up some more
1619 memory (6 * PAGE_SIZE for each 2MB hugetlb page).
1620 Format: { on | off (default) }
1622 on: enable the feature
1623 off: disable the feature
1625 Built with CONFIG_HUGETLB_PAGE_FREE_VMEMMAP_DEFAULT_ON=y,
1628 This is not compatible with memory_hotplug.memmap_on_memory.
1629 If both parameters are enabled, hugetlb_free_vmemmap takes
1630 precedence over memory_hotplug.memmap_on_memory.
1633 [KNL] Should the hung task detector generate panics.
1636 A value of 1 instructs the kernel to panic when a
1637 hung task is detected. The default value is controlled
1638 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1639 option. The value selected by this boot parameter can
1640 be changed later by the kernel.hung_task_panic sysctl.
1642 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1643 terminal devices. Valid values: 0..8
1644 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1645 If specified, z/VM IUCV HVC accepts connections
1646 from listed z/VM user IDs only.
1648 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1649 which allow the hypervisor to 'idle' the
1650 guest on lock contention.
1653 Do not unregister boot console at start. This is only
1654 useful for debugging when something happens in the window
1655 between unregistering the boot console and initializing
1658 i2c_bus= [HW] Override the default board specific I2C bus speed
1659 or register an additional I2C bus that is not
1660 registered from board initialization code.
1664 i8042.debug [HW] Toggle i8042 debug mode
1665 i8042.unmask_kbd_data
1666 [HW] Enable printing of interrupt data from the KBD port
1667 (disabled by default, and as a pre-condition
1668 requires that i8042.debug=1 be enabled)
1669 i8042.direct [HW] Put keyboard port into non-translated mode
1670 i8042.dumbkbd [HW] Pretend that controller can only read data from
1671 keyboard and cannot control its state
1672 (Don't attempt to blink the leds)
1673 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1674 i8042.nokbd [HW] Don't check/create keyboard port
1675 i8042.noloop [HW] Disable the AUX Loopback command while probing
1677 i8042.nomux [HW] Don't check presence of an active multiplexing
1679 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1681 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1682 i8042.reset [HW] Reset the controller during init, cleanup and
1683 suspend-to-ram transitions, only during s2r
1684 transitions, or never reset
1685 Format: { 1 | Y | y | 0 | N | n }
1686 1, Y, y: always reset controller
1687 0, N, n: don't ever reset controller
1688 Default: only on s2r transitions on x86; most other
1689 architectures force reset to be always executed
1690 i8042.unlock [HW] Unlock (ignore) the keylock
1691 i8042.kbdreset [HW] Reset device connected to KBD port
1695 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1696 indicates that the driver is running on unsupported
1698 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1699 does not match list of supported models.
1701 [HW] Report power status in /proc/i8k
1702 (disabled by default)
1703 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1706 i915.invert_brightness=
1707 [DRM] Invert the sense of the variable that is used to
1708 set the brightness of the panel backlight. Normally a
1709 brightness value of 0 indicates backlight switched off,
1710 and the maximum of the brightness value sets the backlight
1711 to maximum brightness. If this parameter is set to 0
1712 (default) and the machine requires it, or this parameter
1713 is set to 1, a brightness value of 0 sets the backlight
1714 to maximum brightness, and the maximum of the brightness
1715 value switches the backlight off.
1716 -1 -- never invert brightness
1717 0 -- machine default
1718 1 -- force brightness inversion
1721 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1723 ide-core.nodma= [HW] (E)IDE subsystem
1724 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1725 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1726 .cdrom .chs .ignore_cable are additional options
1727 See Documentation/ide/ide.rst.
1729 ide-generic.probe-mask= [HW] (E)IDE subsystem
1731 Probe mask for legacy ISA IDE ports. Depending on
1732 platform up to 6 ports are supported, enabled by
1733 setting corresponding bits in the mask to 1. The
1734 default value is 0x0, which has a special meaning.
1735 On systems that have PCI, it triggers scanning the
1736 PCI bus for the first and the second port, which
1737 are then probed. On systems without PCI the value
1738 of 0x0 enables probing the two first ports as if it
1741 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1742 Claim all unknown PCI IDE storage controllers.
1745 Format: idle=poll, idle=halt, idle=nomwait
1746 Poll forces a polling idle loop that can slightly
1747 improve the performance of waking up a idle CPU, but
1748 will use a lot of power and make the system run hot.
1750 idle=halt: Halt is forced to be used for CPU idle.
1751 In such case C2/C3 won't be used again.
1752 idle=nomwait: Disable mwait for CPU C-states
1756 Allow force disabling of Shared Virtual Memory (SVA)
1757 support for the idxd driver. By default it is set to
1760 idxd.tc_override= [HW]
1762 Allow override of default traffic class configuration
1763 for the device. By default it is set to false (0).
1765 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1766 Format: { strict | legacy | 2008 | relaxed }
1769 Choose which programs will be accepted for execution
1770 based on the IEEE 754 NaN encoding(s) supported by
1771 the FPU and the NaN encoding requested with the value
1772 of an ELF file header flag individually set by each
1773 binary. Hardware implementations are permitted to
1774 support either or both of the legacy and the 2008 NaN
1777 Available settings are as follows:
1778 strict accept binaries that request a NaN encoding
1779 supported by the FPU
1780 legacy only accept legacy-NaN binaries, if supported
1782 2008 only accept 2008-NaN binaries, if supported
1784 relaxed accept any binaries regardless of whether
1785 supported by the FPU
1787 The FPU emulator is always able to support both NaN
1788 encodings, so if no FPU hardware is present or it has
1789 been disabled with 'nofpu', then the settings of
1790 'legacy' and '2008' strap the emulator accordingly,
1791 'relaxed' straps the emulator for both legacy-NaN and
1792 2008-NaN, whereas 'strict' enables legacy-NaN only on
1793 legacy processors and both NaN encodings on MIPS32 or
1796 The setting for ABS.fmt/NEG.fmt instruction execution
1797 mode generally follows that for the NaN encoding,
1798 except where unsupported by hardware.
1800 ignore_loglevel [KNL]
1801 Ignore loglevel setting - this will print /all/
1802 kernel messages to the console. Useful for debugging.
1803 We also add it as printk module parameter, so users
1804 could change it dynamically, usually by
1805 /sys/module/printk/parameters/ignore_loglevel.
1808 Ignore RLIMIT_DATA setting for data mappings,
1809 print warning at first misuse. Can be changed via
1810 /sys/module/kernel/parameters/ignore_rlimit_data.
1812 ihash_entries= [KNL]
1813 Set number of hash buckets for inode cache.
1815 ima_appraise= [IMA] appraise integrity measurements
1816 Format: { "off" | "enforce" | "fix" | "log" }
1819 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1820 The builtin appraise policy appraises all files
1823 ima_canonical_fmt [IMA]
1824 Use the canonical format for the binary runtime
1825 measurements, instead of host native format.
1828 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1832 The list of supported hash algorithms is defined
1833 in crypto/hash_info.h.
1836 The builtin policies to load during IMA setup.
1837 Format: "tcb | appraise_tcb | secure_boot |
1838 fail_securely | critical_data"
1840 The "tcb" policy measures all programs exec'd, files
1841 mmap'd for exec, and all files opened with the read
1842 mode bit set by either the effective uid (euid=0) or
1845 The "appraise_tcb" policy appraises the integrity of
1846 all files owned by root.
1848 The "secure_boot" policy appraises the integrity
1849 of files (eg. kexec kernel image, kernel modules,
1850 firmware, policy, etc) based on file signatures.
1852 The "fail_securely" policy forces file signature
1853 verification failure also on privileged mounted
1854 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1857 The "critical_data" policy measures kernel integrity
1860 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1861 Load a policy which meets the needs of the Trusted
1862 Computing Base. This means IMA will measure all
1863 programs exec'd, files mmap'd for exec, and all files
1864 opened for read by uid=0.
1867 Select one of defined IMA measurements template formats.
1868 Formats: { "ima" | "ima-ng" | "ima-sig" }
1872 [IMA] Define a custom template format.
1873 Format: { "field1|...|fieldN" }
1875 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1876 Format: <min_file_size>
1877 Set the minimal file size for using asynchronous hash.
1878 If left unspecified, ahash usage is disabled.
1880 ahash performance varies for different data sizes on
1881 different crypto accelerators. This option can be used
1882 to achieve the best performance for a particular HW.
1884 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1886 Set hashing buffer size. Default: 4k.
1888 ahash performance varies for different chunk sizes on
1889 different crypto accelerators. This option can be used
1890 to achieve best performance for particular HW.
1894 Run specified binary instead of /sbin/init as init
1897 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1898 for working out where the kernel is dying during
1901 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1902 initcall functions. Useful for debugging built-in
1903 modules and initcalls.
1905 initramfs_async= [KNL]
1908 This parameter controls whether the initramfs
1909 image is unpacked asynchronously, concurrently
1910 with devices being probed and
1911 initialized. This should normally just work,
1912 but as a debugging aid, one can get the
1913 historical behaviour of the initramfs
1914 unpacking being completed before device_ and
1917 initrd= [BOOT] Specify the location of the initial ramdisk
1919 initrdmem= [KNL] Specify a physical address and size from which to
1920 load the initrd. If an initrd is compiled in or
1921 specified in the bootparams, it takes priority over this
1923 Format: ss[KMG],nn[KMG]
1926 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1929 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1931 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1933 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1935 init_pkru= [X86] Specify the default memory protection keys rights
1936 register contents for all processes. 0x55555554 by
1937 default (disallow access to all but pkey 0). Can
1938 override in debugfs after boot.
1940 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1943 int_pln_enable [X86] Enable power limit notification interrupt
1945 integrity_audit=[IMA]
1946 Format: { "0" | "1" }
1947 0 -- basic integrity auditing messages. (Default)
1948 1 -- additional integrity auditing messages.
1950 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1952 Enable intel iommu driver.
1954 Disable intel iommu driver.
1955 igfx_off [Default Off]
1956 By default, gfx is mapped as normal device. If a gfx
1957 device has a dedicated DMAR unit, the DMAR unit is
1958 bypassed by not enabling DMAR with this option. In
1959 this case, gfx device will use physical address for
1961 strict [Default Off]
1962 Deprecated, equivalent to iommu.strict=1.
1963 sp_off [Default Off]
1964 By default, super page will be supported if Intel IOMMU
1965 has the capability. With this option, super page will
1968 Enable the Intel IOMMU scalable mode if the hardware
1969 advertises that it has support for the scalable mode
1972 Disallow use of the Intel IOMMU scalable mode.
1973 tboot_noforce [Default Off]
1974 Do not force the Intel IOMMU enabled under tboot.
1975 By default, tboot will force Intel IOMMU on, which
1976 could harm performance of some high-throughput
1977 devices like 40GBit network cards, even if identity
1979 Note that using this option lowers the security
1980 provided by tboot because it makes the system
1981 vulnerable to DMA attacks.
1983 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1984 0 disables intel_idle and fall back on acpi_idle.
1985 1 to 9 specify maximum depth of C-state.
1989 Do not enable intel_pstate as the default
1990 scaling driver for the supported processors
1992 Use intel_pstate as a scaling driver, but configure it
1993 to work with generic cpufreq governors (instead of
1994 enabling its internal governor). This mode cannot be
1995 used along with the hardware-managed P-states (HWP)
1998 Enable intel_pstate on systems that prohibit it by default
1999 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2000 instead of acpi-cpufreq may disable platform features, such
2001 as thermal controls and power capping, that rely on ACPI
2002 P-States information being indicated to OSPM and therefore
2003 should be used with caution. This option does not work with
2004 processors that aren't supported by the intel_pstate driver
2005 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2007 Do not enable hardware P state control (HWP)
2010 Only load intel_pstate on systems which support
2011 hardware P state control (HWP) if available.
2013 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2014 Description Table, specifies preferred power management
2015 profile as "Enterprise Server" or "Performance Server",
2016 then this feature is turned on by default.
2018 Allow per-logical-CPU P-State performance control limits using
2019 cpufreq sysfs interface
2021 intremap= [X86-64, Intel-IOMMU]
2022 on enable Interrupt Remapping (default)
2023 off disable Interrupt Remapping
2024 nosid disable Source ID checking
2026 BIOS x2APIC opt-out request will be ignored
2027 nopost disable Interrupt Posting
2029 iomem= Disable strict checking of access to MMIO memory
2030 strict regions from userspace.
2045 nobypass [PPC/POWERNV]
2046 Disable IOMMU bypass, using IOMMU for PCI devices.
2048 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2049 Format: { "0" | "1" }
2050 0 - Try to allocate a 32-bit DMA address first, before
2051 falling back to the full range if needed.
2052 1 - Allocate directly from the full usable range,
2053 forcing Dual Address Cycle for PCI cards supporting
2054 greater than 32-bit addressing.
2056 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2057 Format: { "0" | "1" }
2059 Request that DMA unmap operations use deferred
2060 invalidation of hardware TLBs, for increased
2061 throughput at the cost of reduced device isolation.
2062 Will fall back to strict mode if not supported by
2063 the relevant IOMMU driver.
2065 DMA unmap operations invalidate IOMMU hardware TLBs
2067 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2068 Note: on x86, strict mode specified via one of the
2069 legacy driver-specific options takes precedence.
2072 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2073 Format: { "0" | "1" }
2074 0 - Use IOMMU translation for DMA.
2075 1 - Bypass the IOMMU for DMA.
2076 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2078 io7= [HW] IO7 for Marvel-based Alpha systems
2079 See comment before marvel_specify_io7 in
2080 arch/alpha/kernel/core_marvel.c.
2082 io_delay= [X86] I/O delay method
2084 Standard port 0x80 based delay
2086 Alternate port 0xed based delay (needed on some systems)
2088 Simple two microseconds delay
2093 See Documentation/admin-guide/nfs/nfsroot.rst.
2095 ipcmni_extend [KNL] Extend the maximum number of unique System V
2096 IPC identifiers from 32,768 to 16,777,216.
2098 irqaffinity= [SMP] Set the default irq affinity mask
2099 The argument is a cpu list, as described above.
2101 irqchip.gicv2_force_probe=
2104 Force the kernel to look for the second 4kB page
2105 of a GICv2 controller even if the memory range
2106 exposed by the device tree is too small.
2108 irqchip.gicv3_nolpi=
2110 Force the kernel to ignore the availability of
2111 LPIs (and by consequence ITSs). Intended for system
2112 that use the kernel as a bootloader, and thus want
2113 to let secondary kernels in charge of setting up
2116 irqchip.gicv3_pseudo_nmi= [ARM64]
2117 Enables support for pseudo-NMIs in the kernel. This
2118 requires the kernel to be built with
2119 CONFIG_ARM64_PSEUDO_NMI.
2122 When an interrupt is not handled search all handlers
2123 for it. Intended to get systems with badly broken
2127 When an interrupt is not handled search all handlers
2128 for it. Also check all handlers each timer
2129 interrupt. Intended to get systems with badly broken
2133 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2135 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2136 [Deprecated - use cpusets instead]
2137 Format: [flag-list,]<cpu-list>
2139 Specify one or more CPUs to isolate from disturbances
2140 specified in the flag list (default: domain):
2143 Disable the tick when a single task runs.
2145 A residual 1Hz tick is offloaded to workqueues, which you
2146 need to affine to housekeeping through the global
2147 workqueue's affinity configured via the
2148 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2149 by using the 'domain' flag described below.
2151 NOTE: by default the global workqueue runs on all CPUs,
2152 so to protect individual CPUs the 'cpumask' file has to
2153 be configured manually after bootup.
2156 Isolate from the general SMP balancing and scheduling
2157 algorithms. Note that performing domain isolation this way
2158 is irreversible: it's not possible to bring back a CPU to
2159 the domains once isolated through isolcpus. It's strongly
2160 advised to use cpusets instead to disable scheduler load
2161 balancing through the "cpuset.sched_load_balance" file.
2162 It offers a much more flexible interface where CPUs can
2163 move in and out of an isolated set anytime.
2165 You can move a process onto or off an "isolated" CPU via
2166 the CPU affinity syscalls or cpuset.
2167 <cpu number> begins at 0 and the maximum value is
2168 "number of CPUs in system - 1".
2172 Isolate from being targeted by managed interrupts
2173 which have an interrupt mask containing isolated
2174 CPUs. The affinity of managed interrupts is
2175 handled by the kernel and cannot be changed via
2176 the /proc/irq/* interfaces.
2178 This isolation is best effort and only effective
2179 if the automatically assigned interrupt mask of a
2180 device queue contains isolated and housekeeping
2181 CPUs. If housekeeping CPUs are online then such
2182 interrupts are directed to the housekeeping CPU
2183 so that IO submitted on the housekeeping CPU
2184 cannot disturb the isolated CPU.
2186 If a queue's affinity mask contains only isolated
2187 CPUs then this parameter has no effect on the
2188 interrupt routing decision, though interrupts are
2189 only delivered when tasks running on those
2190 isolated CPUs submit IO. IO submitted on
2191 housekeeping CPUs has no influence on those
2194 The format of <cpu-list> is described above.
2198 ivrs_ioapic [HW,X86-64]
2199 Provide an override to the IOAPIC-ID<->DEVICE-ID
2200 mapping provided in the IVRS ACPI table. For
2201 example, to map IOAPIC-ID decimal 10 to
2202 PCI device 00:14.0 write the parameter as:
2203 ivrs_ioapic[10]=00:14.0
2205 ivrs_hpet [HW,X86-64]
2206 Provide an override to the HPET-ID<->DEVICE-ID
2207 mapping provided in the IVRS ACPI table. For
2208 example, to map HPET-ID decimal 0 to
2209 PCI device 00:14.0 write the parameter as:
2210 ivrs_hpet[0]=00:14.0
2212 ivrs_acpihid [HW,X86-64]
2213 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2214 mapping provided in the IVRS ACPI table. For
2215 example, to map UART-HID:UID AMD0020:0 to
2216 PCI device 00:14.5 write the parameter as:
2217 ivrs_acpihid[00:14.5]=AMD0020:0
2219 js= [HW,JOY] Analog joystick
2220 See Documentation/input/joydev/joystick.rst.
2223 When CONFIG_RANDOMIZE_BASE is set, this disables
2224 kernel and module base offset ASLR (Address Space
2225 Layout Randomization).
2228 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2229 report on every invalid memory access. Without this
2230 parameter KASAN will print report only for the first
2235 kernelcore= [KNL,X86,IA-64,PPC]
2236 Format: nn[KMGTPE] | nn% | "mirror"
2237 This parameter specifies the amount of memory usable by
2238 the kernel for non-movable allocations. The requested
2239 amount is spread evenly throughout all nodes in the
2240 system as ZONE_NORMAL. The remaining memory is used for
2241 movable memory in its own zone, ZONE_MOVABLE. In the
2242 event, a node is too small to have both ZONE_NORMAL and
2243 ZONE_MOVABLE, kernelcore memory will take priority and
2244 other nodes will have a larger ZONE_MOVABLE.
2246 ZONE_MOVABLE is used for the allocation of pages that
2247 may be reclaimed or moved by the page migration
2248 subsystem. Note that allocations like PTEs-from-HighMem
2249 still use the HighMem zone if it exists, and the Normal
2250 zone if it does not.
2252 It is possible to specify the exact amount of memory in
2253 the form of "nn[KMGTPE]", a percentage of total system
2254 memory in the form of "nn%", or "mirror". If "mirror"
2255 option is specified, mirrored (reliable) memory is used
2256 for non-movable allocations and remaining memory is used
2257 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2258 are exclusive, so you cannot specify multiple forms.
2260 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2261 Format: <Controller#>[,poll interval]
2262 The controller # is the number of the ehci usb debug
2263 port as it is probed via PCI. The poll interval is
2264 optional and is the number seconds in between
2265 each poll cycle to the debug port in case you need
2266 the functionality for interrupting the kernel with
2267 gdb or control-c on the dbgp connection. When
2268 not using this parameter you use sysrq-g to break into
2269 the kernel debugger.
2271 kgdboc= [KGDB,HW] kgdb over consoles.
2272 Requires a tty driver that supports console polling,
2273 or a supported polling keyboard driver (non-usb).
2274 Serial only format: <serial_device>[,baud]
2275 keyboard only format: kbd
2276 keyboard and serial format: kbd,<serial_device>[,baud]
2277 Optional Kernel mode setting:
2278 kms, kbd format: kms,kbd
2279 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2281 kgdboc_earlycon= [KGDB,HW]
2282 If the boot console provides the ability to read
2283 characters and can work in polling mode, you can use
2284 this parameter to tell kgdb to use it as a backend
2285 until the normal console is registered. Intended to
2286 be used together with the kgdboc parameter which
2287 specifies the normal console to transition to.
2289 The name of the early console should be specified
2290 as the value of this parameter. Note that the name of
2291 the early console might be different than the tty
2292 name passed to kgdboc. It's OK to leave the value
2293 blank and the first boot console that implements
2294 read() will be picked.
2296 kgdbwait [KGDB] Stop kernel execution and enter the
2297 kernel debugger at the earliest opportunity.
2299 kmac= [MIPS] Korina ethernet MAC address.
2300 Configure the RouterBoard 532 series on-chip
2301 Ethernet adapter MAC address.
2303 kmemleak= [KNL] Boot-time kmemleak enable/disable
2304 Valid arguments: on, off
2306 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2309 kprobe_event=[probe-list]
2310 [FTRACE] Add kprobe events and enable at boot time.
2311 The probe-list is a semicolon delimited list of probe
2312 definitions. Each definition is same as kprobe_events
2313 interface, but the parameters are comma delimited.
2314 For example, to add a kprobe event on vfs_read with
2315 arg1 and arg2, add to the command line;
2317 kprobe_event=p,vfs_read,$arg1,$arg2
2319 See also Documentation/trace/kprobetrace.rst "Kernel
2320 Boot Parameter" section.
2322 kpti= [ARM64] Control page table isolation of user
2323 and kernel address spaces.
2324 Default: enabled on cores which need mitigation.
2328 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2329 Default is 0 (don't ignore, but inject #GP)
2331 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2332 Default is false (don't support).
2334 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2339 [KVM] Controls the software workaround for the
2340 X86_BUG_ITLB_MULTIHIT bug.
2341 force : Always deploy workaround.
2342 off : Never deploy workaround.
2343 auto : Deploy workaround based on the presence of
2344 X86_BUG_ITLB_MULTIHIT.
2348 If the software workaround is enabled for the host,
2349 guests do need not to enable it for nested guests.
2351 kvm.nx_huge_pages_recovery_ratio=
2352 [KVM] Controls how many 4KiB pages are periodically zapped
2353 back to huge pages. 0 disables the recovery, otherwise if
2354 the value is N KVM will zap 1/Nth of the 4KiB pages every
2355 period (see below). The default is 60.
2357 kvm.nx_huge_pages_recovery_period_ms=
2358 [KVM] Controls the time period at which KVM zaps 4KiB pages
2359 back to huge pages. If the value is a non-zero N, KVM will
2360 zap a portion (see ratio above) of the pages every N msecs.
2361 If the value is 0 (the default), KVM will pick a period based
2362 on the ratio, such that a page is zapped after 1 hour on average.
2364 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2365 Default is 1 (enabled)
2367 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2369 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2372 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2374 none: Forcefully disable KVM.
2376 nvhe: Standard nVHE-based mode, without support for
2379 protected: nVHE-based mode with support for guests whose
2380 state is kept private from the host.
2381 Not valid if the kernel is running in EL2.
2383 Defaults to VHE/nVHE based on hardware support. Setting
2384 mode to "protected" will disable kexec and hibernation
2387 kvm-arm.vgic_v3_group0_trap=
2388 [KVM,ARM] Trap guest accesses to GICv3 group-0
2391 kvm-arm.vgic_v3_group1_trap=
2392 [KVM,ARM] Trap guest accesses to GICv3 group-1
2395 kvm-arm.vgic_v3_common_trap=
2396 [KVM,ARM] Trap guest accesses to GICv3 common
2399 kvm-arm.vgic_v4_enable=
2400 [KVM,ARM] Allow use of GICv4 for direct injection of
2403 kvm_cma_resv_ratio=n [PPC]
2404 Reserves given percentage from system memory area for
2405 contiguous memory allocation for KVM hash pagetable
2407 By default it reserves 5% of total system memory.
2411 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2412 (virtualized MMU) support on capable Intel chips.
2413 Default is 1 (enabled)
2415 kvm-intel.emulate_invalid_guest_state=
2416 [KVM,Intel] Enable emulation of invalid guest states
2417 Default is 0 (disabled)
2419 kvm-intel.flexpriority=
2420 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2421 Default is 1 (enabled)
2424 [KVM,Intel] Enable VMX nesting (nVMX).
2425 Default is 0 (disabled)
2427 kvm-intel.unrestricted_guest=
2428 [KVM,Intel] Disable unrestricted guest feature
2429 (virtualized real and unpaged mode) on capable
2430 Intel chips. Default is 1 (enabled)
2432 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2435 Valid arguments: never, cond, always
2437 always: L1D cache flush on every VMENTER.
2438 cond: Flush L1D on VMENTER only when the code between
2439 VMEXIT and VMENTER can leak host memory.
2440 never: Disables the mitigation
2442 Default is cond (do L1 cache flush in specific instances)
2444 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2445 feature (tagged TLBs) on capable Intel chips.
2446 Default is 1 (enabled)
2448 l1d_flush= [X86,INTEL]
2449 Control mitigation for L1D based snooping vulnerability.
2451 Certain CPUs are vulnerable to an exploit against CPU
2452 internal buffers which can forward information to a
2453 disclosure gadget under certain conditions.
2455 In vulnerable processors, the speculatively
2456 forwarded data can be used in a cache side channel
2457 attack, to access data to which the attacker does
2458 not have direct access.
2460 This parameter controls the mitigation. The
2463 on - enable the interface for the mitigation
2465 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2468 The kernel PTE inversion protection is unconditionally
2469 enabled and cannot be disabled.
2472 Provides all available mitigations for the
2473 L1TF vulnerability. Disables SMT and
2474 enables all mitigations in the
2475 hypervisors, i.e. unconditional L1D flush.
2477 SMT control and L1D flush control via the
2478 sysfs interface is still possible after
2479 boot. Hypervisors will issue a warning
2480 when the first VM is started in a
2481 potentially insecure configuration,
2482 i.e. SMT enabled or L1D flush disabled.
2485 Same as 'full', but disables SMT and L1D
2486 flush runtime control. Implies the
2487 'nosmt=force' command line option.
2488 (i.e. sysfs control of SMT is disabled.)
2491 Leaves SMT enabled and enables the default
2492 hypervisor mitigation, i.e. conditional
2495 SMT control and L1D flush control via the
2496 sysfs interface is still possible after
2497 boot. Hypervisors will issue a warning
2498 when the first VM is started in a
2499 potentially insecure configuration,
2500 i.e. SMT enabled or L1D flush disabled.
2504 Disables SMT and enables the default
2505 hypervisor mitigation.
2507 SMT control and L1D flush control via the
2508 sysfs interface is still possible after
2509 boot. Hypervisors will issue a warning
2510 when the first VM is started in a
2511 potentially insecure configuration,
2512 i.e. SMT enabled or L1D flush disabled.
2515 Same as 'flush', but hypervisors will not
2516 warn when a VM is started in a potentially
2517 insecure configuration.
2520 Disables hypervisor mitigations and doesn't
2522 It also drops the swap size and available
2523 RAM limit restriction on both hypervisor and
2528 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2534 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2537 lapic= [X86,APIC] Do not use TSC deadline
2538 value for LAPIC timer one-shot implementation. Default
2539 back to the programmable timer unit in the LAPIC.
2540 Format: notscdeadline
2542 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2545 libata.dma= [LIBATA] DMA control
2546 libata.dma=0 Disable all PATA and SATA DMA
2547 libata.dma=1 PATA and SATA Disk DMA only
2548 libata.dma=2 ATAPI (CDROM) DMA only
2549 libata.dma=4 Compact Flash DMA only
2550 Combinations also work, so libata.dma=3 enables DMA
2551 for disks and CDROMs, but not CFs.
2553 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2554 libata.ignore_hpa=0 keep BIOS limits (default)
2555 libata.ignore_hpa=1 ignore limits, using full disk
2557 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2561 libata.force= [LIBATA] Force configurations. The format is comma-
2562 separated list of "[ID:]VAL" where ID is
2563 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2564 matching port, link or device. Basically, it matches
2565 the ATA ID string printed on console by libata. If
2566 the whole ID part is omitted, the last PORT and DEVICE
2567 values are used. If ID hasn't been specified yet, the
2568 configuration applies to all ports, links and devices.
2570 If only DEVICE is omitted, the parameter applies to
2571 the port and all links and devices behind it. DEVICE
2572 number of 0 either selects the first device or the
2573 first fan-out link behind PMP device. It does not
2574 select the host link. DEVICE number of 15 selects the
2575 host link and device attached to it.
2577 The VAL specifies the configuration to force. As long
2578 as there's no ambiguity shortcut notation is allowed.
2579 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2580 The following configurations can be forced.
2582 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2583 Any ID with matching PORT is used.
2585 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2587 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2588 udma[/][16,25,33,44,66,100,133] notation is also
2591 * [no]ncq: Turn on or off NCQ.
2593 * [no]ncqtrim: Turn off queued DSM TRIM.
2595 * nohrst, nosrst, norst: suppress hard, soft
2598 * rstonce: only attempt one reset during
2599 hot-unplug link recovery
2601 * dump_id: dump IDENTIFY data.
2603 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2605 * disable: Disable this device.
2607 If there are multiple matching configurations changing
2608 the same attribute, the last one is used.
2610 memblock=debug [KNL] Enable memblock debug messages.
2612 load_ramdisk= [RAM] [Deprecated]
2614 lockd.nlm_grace_period=P [NFS] Assign grace period.
2617 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2620 lockd.nlm_timeout=T [NFS] Assign timeout value.
2623 lockd.nlm_udpport=M [NFS] Assign UDP port.
2626 lockdown= [SECURITY]
2627 { integrity | confidentiality }
2628 Enable the kernel lockdown feature. If set to
2629 integrity, kernel features that allow userland to
2630 modify the running kernel are disabled. If set to
2631 confidentiality, kernel features that allow userland
2632 to extract confidential information from the kernel
2635 locktorture.nreaders_stress= [KNL]
2636 Set the number of locking read-acquisition kthreads.
2637 Defaults to being automatically set based on the
2638 number of online CPUs.
2640 locktorture.nwriters_stress= [KNL]
2641 Set the number of locking write-acquisition kthreads.
2643 locktorture.onoff_holdoff= [KNL]
2644 Set time (s) after boot for CPU-hotplug testing.
2646 locktorture.onoff_interval= [KNL]
2647 Set time (s) between CPU-hotplug operations, or
2648 zero to disable CPU-hotplug testing.
2650 locktorture.shuffle_interval= [KNL]
2651 Set task-shuffle interval (jiffies). Shuffling
2652 tasks allows some CPUs to go into dyntick-idle
2653 mode during the locktorture test.
2655 locktorture.shutdown_secs= [KNL]
2656 Set time (s) after boot system shutdown. This
2657 is useful for hands-off automated testing.
2659 locktorture.stat_interval= [KNL]
2660 Time (s) between statistics printk()s.
2662 locktorture.stutter= [KNL]
2663 Time (s) to stutter testing, for example,
2664 specifying five seconds causes the test to run for
2665 five seconds, wait for five seconds, and so on.
2666 This tests the locking primitive's ability to
2667 transition abruptly to and from idle.
2669 locktorture.torture_type= [KNL]
2670 Specify the locking implementation to test.
2672 locktorture.verbose= [KNL]
2673 Enable additional printk() statements.
2675 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2678 loglevel= All Kernel Messages with a loglevel smaller than the
2679 console loglevel will be printed to the console. It can
2680 also be changed with klogd or other programs. The
2681 loglevels are defined as follows:
2683 0 (KERN_EMERG) system is unusable
2684 1 (KERN_ALERT) action must be taken immediately
2685 2 (KERN_CRIT) critical conditions
2686 3 (KERN_ERR) error conditions
2687 4 (KERN_WARNING) warning conditions
2688 5 (KERN_NOTICE) normal but significant condition
2689 6 (KERN_INFO) informational
2690 7 (KERN_DEBUG) debug-level messages
2692 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2693 in bytes. n must be a power of two and greater
2694 than the minimal size. The minimal size is defined
2695 by LOG_BUF_SHIFT kernel config parameter. There is
2696 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2697 that allows to increase the default size depending on
2698 the number of CPUs. See init/Kconfig for more details.
2700 logo.nologo [FB] Disables display of the built-in Linux logo.
2701 This may be used to provide more screen space for
2702 kernel log messages and is useful when debugging
2703 kernel boot problems.
2705 lp=0 [LP] Specify parallel ports to use, e.g,
2706 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2707 lp=reset first parallel port). 'lp=0' disables the
2708 lp=auto printer driver. 'lp=reset' (which can be
2709 specified in addition to the ports) causes
2710 attached printers to be reset. Using
2711 lp=port1,port2,... specifies the parallel ports
2712 to associate lp devices with, starting with
2713 lp0. A port specification may be 'none' to skip
2714 that lp device, or a parport name such as
2715 'parport0'. Specifying 'lp=auto' instead of a
2716 port specification list means that device IDs
2717 from each port should be examined, to see if
2718 an IEEE 1284-compliant printer is attached; if
2719 so, the driver will manage that printer.
2720 See also header of drivers/char/lp.c.
2723 Sets loops_per_jiffy to given constant, thus avoiding
2724 time-consuming boot-time autodetection (up to 250 ms per
2725 CPU). 0 enables autodetection (default). To determine
2726 the correct value for your kernel, boot with normal
2727 autodetection and see what value is printed. Note that
2728 on SMP systems the preset will be applied to all CPUs,
2729 which is likely to cause problems if your CPUs need
2730 significantly divergent settings. An incorrect value
2731 will cause delays in the kernel to be wrong, leading to
2732 unpredictable I/O errors and other breakage. Although
2733 unlikely, in the extreme case this might damage your
2737 Format: <io>,<irq>,<dma>
2739 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2742 [SECURITY] Choose order of LSM initialization. This
2743 overrides CONFIG_LSM, and the "security=" parameter.
2745 machvec= [IA-64] Force the use of a particular machine-vector
2746 (machvec) in a generic kernel.
2747 Example: machvec=hpzx1
2749 machtype= [Loongson] Share the same kernel image file between
2750 different yeeloong laptops.
2751 Example: machtype=lemote-yeeloong-2f-7inch
2753 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2754 than or equal to this physical address is ignored.
2756 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2757 will bring up during bootup. maxcpus=n : n >= 0 limits
2758 the kernel to bring up 'n' processors. Surely after
2759 bootup you can bring up the other plugged cpu by executing
2760 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2761 only takes effect during system bootup.
2762 While n=0 is a special case, it is equivalent to "nosmp",
2763 which also disables the IO APIC.
2765 max_loop= [LOOP] The number of loop block devices that get
2766 (loop.max_loop) unconditionally pre-created at init time. The default
2767 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2768 of statically allocating a predefined number, loop
2769 devices can be requested on-demand with the
2770 /dev/loop-control interface.
2772 mce [X86-32] Machine Check Exception
2774 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2776 md= [HW] RAID subsystems devices and level
2777 See Documentation/admin-guide/md.rst.
2780 Format: <first>,<last>
2781 Specifies range of consoles to be captured by the MDA.
2784 Control mitigation for the Micro-architectural Data
2785 Sampling (MDS) vulnerability.
2787 Certain CPUs are vulnerable to an exploit against CPU
2788 internal buffers which can forward information to a
2789 disclosure gadget under certain conditions.
2791 In vulnerable processors, the speculatively
2792 forwarded data can be used in a cache side channel
2793 attack, to access data to which the attacker does
2794 not have direct access.
2796 This parameter controls the MDS mitigation. The
2799 full - Enable MDS mitigation on vulnerable CPUs
2800 full,nosmt - Enable MDS mitigation and disable
2801 SMT on vulnerable CPUs
2802 off - Unconditionally disable MDS mitigation
2804 On TAA-affected machines, mds=off can be prevented by
2805 an active TAA mitigation as both vulnerabilities are
2806 mitigated with the same mechanism so in order to disable
2807 this mitigation, you need to specify tsx_async_abort=off
2810 Not specifying this option is equivalent to
2813 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2815 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2816 Amount of memory to be used in cases as follows:
2819 2 when the kernel is not able to see the whole system memory;
2820 3 memory that lies after 'mem=' boundary is excluded from
2821 the hypervisor, then assigned to KVM guests.
2823 [X86] Work as limiting max address. Use together
2824 with memmap= to avoid physical address space collisions.
2825 Without memmap= PCI devices could be placed at addresses
2826 belonging to unused RAM.
2828 Note that this only takes effects during boot time since
2829 in above case 3, memory may need be hot added after boot
2830 if system memory of hypervisor is not sufficient.
2832 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2836 [KNL,SH] Allow user to override the default size for
2837 per-device physically contiguous DMA buffers.
2839 memhp_default_state=online/offline
2840 [KNL] Set the initial state for the memory hotplug
2841 onlining policy. If not specified, the default value is
2842 set according to the
2843 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2845 See Documentation/admin-guide/mm/memory-hotplug.rst.
2847 memmap=exactmap [KNL,X86] Enable setting of an exact
2848 E820 memory map, as specified by the user.
2849 Such memmap=exactmap lines can be constructed based on
2850 BIOS output or other requirements. See the memmap=nn@ss
2853 memmap=nn[KMG]@ss[KMG]
2854 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
2855 Region of memory to be used is from ss to ss+nn.
2856 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2857 which limits max address to nn[KMG].
2858 Multiple different regions can be specified,
2861 memmap=100M@2G,100M#3G,1G!1024G
2863 memmap=nn[KMG]#ss[KMG]
2864 [KNL,ACPI] Mark specific memory as ACPI data.
2865 Region of memory to be marked is from ss to ss+nn.
2867 memmap=nn[KMG]$ss[KMG]
2868 [KNL,ACPI] Mark specific memory as reserved.
2869 Region of memory to be reserved is from ss to ss+nn.
2870 Example: Exclude memory from 0x18690000-0x1869ffff
2871 memmap=64K$0x18690000
2873 memmap=0x10000$0x18690000
2874 Some bootloaders may need an escape character before '$',
2875 like Grub2, otherwise '$' and the following number
2878 memmap=nn[KMG]!ss[KMG]
2879 [KNL,X86] Mark specific memory as protected.
2880 Region of memory to be used, from ss to ss+nn.
2881 The memory region may be marked as e820 type 12 (0xc)
2882 and is NVDIMM or ADR memory.
2884 memmap=<size>%<offset>-<oldtype>+<newtype>
2885 [KNL,ACPI] Convert memory within the specified region
2886 from <oldtype> to <newtype>. If "-<oldtype>" is left
2887 out, the whole region will be marked as <newtype>,
2888 even if previously unavailable. If "+<newtype>" is left
2889 out, matching memory will be removed. Types are
2890 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2891 3 = ACPI, 12 = PRAM.
2893 memory_corruption_check=0/1 [X86]
2894 Some BIOSes seem to corrupt the first 64k of
2895 memory when doing things like suspend/resume.
2896 Setting this option will scan the memory
2897 looking for corruption. Enabling this will
2898 both detect corruption and prevent the kernel
2899 from using the memory being corrupted.
2900 However, its intended as a diagnostic tool; if
2901 repeatable BIOS-originated corruption always
2902 affects the same memory, you can use memmap=
2903 to prevent the kernel from using that memory.
2905 memory_corruption_check_size=size [X86]
2906 By default it checks for corruption in the low
2907 64k, making this memory unavailable for normal
2908 use. Use this parameter to scan for
2909 corruption in more or less memory.
2911 memory_corruption_check_period=seconds [X86]
2912 By default it checks for corruption every 60
2913 seconds. Use this parameter to check at some
2914 other rate. 0 disables periodic checking.
2916 memory_hotplug.memmap_on_memory
2917 [KNL,X86,ARM] Boolean flag to enable this feature.
2918 Format: {on | off (default)}
2919 When enabled, runtime hotplugged memory will
2920 allocate its internal metadata (struct pages)
2921 from the hotadded memory which will allow to
2922 hotadd a lot of memory without requiring
2923 additional memory to do so.
2924 This feature is disabled by default because it
2925 has some implication on large (e.g. GB)
2926 allocations in some configurations (e.g. small
2928 The state of the flag can be read in
2929 /sys/module/memory_hotplug/parameters/memmap_on_memory.
2930 Note that even when enabled, there are a few cases where
2931 the feature is not effective.
2933 This is not compatible with hugetlb_free_vmemmap. If
2934 both parameters are enabled, hugetlb_free_vmemmap takes
2935 precedence over memory_hotplug.memmap_on_memory.
2937 memtest= [KNL,X86,ARM,PPC,RISCV] Enable memtest
2939 default : 0 <disable>
2940 Specifies the number of memtest passes to be
2941 performed. Each pass selects another test
2942 pattern from a given set of patterns. Memtest
2943 fills the memory with this pattern, validates
2944 memory contents and reserves bad memory
2945 regions that are detected.
2947 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2948 Valid arguments: on, off
2949 Default (depends on kernel configuration option):
2950 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2951 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2952 mem_encrypt=on: Activate SME
2953 mem_encrypt=off: Do not activate SME
2955 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2956 for details on when memory encryption can be activated.
2958 mem_sleep_default= [SUSPEND] Default system suspend mode:
2959 s2idle - Suspend-To-Idle
2960 shallow - Power-On Suspend or equivalent (if supported)
2961 deep - Suspend-To-RAM or equivalent (if supported)
2962 See Documentation/admin-guide/pm/sleep-states.rst.
2964 meye.*= [HW] Set MotionEye Camera parameters
2965 See Documentation/admin-guide/media/meye.rst.
2967 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2968 Multi-Function General Purpose Timers on AMD Geode
2971 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2972 the BIOS has incorrectly applied a workaround. TinyBIOS
2973 version 0.98 is known to be affected, 0.99 fixes the
2974 problem by letting the user disable the workaround.
2978 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2979 physical address is ignored.
2981 mini2440= [ARM,HW,KNL]
2982 Format:[0..2][b][c][t]
2984 MINI2440 configuration specification:
2985 0 - The attached screen is the 3.5" TFT
2986 1 - The attached screen is the 7" TFT
2987 2 - The VGA Shield is attached (1024x768)
2988 Leaving out the screen size parameter will not load
2989 the TFT driver, and the framebuffer will be left
2991 b - Enable backlight. The TFT backlight pin will be
2992 linked to the kernel VESA blanking code and a GPIO
2993 LED. This parameter is not necessary when using the
2995 c - Enable the s3c camera interface.
2996 t - Reserved for enabling touchscreen support. The
2997 touchscreen support is not enabled in the mainstream
2998 kernel as of 2.6.30, a preliminary port can be found
2999 in the "bleeding edge" mini2440 support kernel at
3000 https://repo.or.cz/w/linux-2.6/mini2440.git
3003 [X86,PPC,S390,ARM64] Control optional mitigations for
3004 CPU vulnerabilities. This is a set of curated,
3005 arch-independent options, each of which is an
3006 aggregation of existing arch-specific options.
3009 Disable all optional CPU mitigations. This
3010 improves system performance, but it may also
3011 expose users to several CPU vulnerabilities.
3012 Equivalent to: nopti [X86,PPC]
3014 nospectre_v1 [X86,PPC]
3016 nospectre_v2 [X86,PPC,S390,ARM64]
3017 spectre_v2_user=off [X86]
3018 spec_store_bypass_disable=off [X86,PPC]
3019 ssbd=force-off [ARM64]
3022 tsx_async_abort=off [X86]
3023 kvm.nx_huge_pages=off [X86]
3024 no_entry_flush [PPC]
3025 no_uaccess_flush [PPC]
3028 This does not have any effect on
3029 kvm.nx_huge_pages when
3030 kvm.nx_huge_pages=force.
3033 Mitigate all CPU vulnerabilities, but leave SMT
3034 enabled, even if it's vulnerable. This is for
3035 users who don't want to be surprised by SMT
3036 getting disabled across kernel upgrades, or who
3037 have other ways of avoiding SMT-based attacks.
3038 Equivalent to: (default behavior)
3041 Mitigate all CPU vulnerabilities, disabling SMT
3042 if needed. This is for users who always want to
3043 be fully mitigated, even if it means losing SMT.
3044 Equivalent to: l1tf=flush,nosmt [X86]
3045 mds=full,nosmt [X86]
3046 tsx_async_abort=full,nosmt [X86]
3049 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3050 parameter allows control of the logging verbosity for
3051 the additional memory initialisation checks. A value
3052 of 0 disables mminit logging and a level of 4 will
3053 log everything. Information is printed at KERN_DEBUG
3054 so loglevel=8 may also need to be specified.
3057 [KNL] When CONFIG_MODULE_SIG is set, this means that
3058 modules without (valid) signatures will fail to load.
3059 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3060 is always true, so this option does nothing.
3062 module_blacklist= [KNL] Do not load a comma-separated list of
3063 modules. Useful for debugging problem modules.
3066 [MOUSE] Maximum time between finger touching and
3067 leaving touchpad surface for touch to be considered
3068 a tap and be reported as a left button click (for
3069 touchpads working in absolute mode only).
3071 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3072 reporting absolute coordinates, such as tablets
3073 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3074 reporting absolute coordinates, such as tablets
3076 movablecore= [KNL,X86,IA-64,PPC]
3077 Format: nn[KMGTPE] | nn%
3078 This parameter is the complement to kernelcore=, it
3079 specifies the amount of memory used for migratable
3080 allocations. If both kernelcore and movablecore is
3081 specified, then kernelcore will be at *least* the
3082 specified value but may be more. If movablecore on its
3083 own is specified, the administrator must be careful
3084 that the amount of memory usable for all allocations
3087 movable_node [KNL] Boot-time switch to make hotplugable memory
3088 NUMA nodes to be movable. This means that the memory
3089 of such nodes will be usable only for movable
3090 allocations which rules out almost all kernel
3091 allocations. Use with caution!
3093 MTD_Partition= [MTD]
3094 Format: <name>,<region-number>,<size>,<offset>
3096 MTD_Region= [MTD] Format:
3097 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3100 See drivers/mtd/parsers/cmdlinepart.c
3102 multitce=off [PPC] This parameter disables the use of the pSeries
3103 firmware feature for updating multiple TCE entries
3106 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3108 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3110 boundary - index of last SLC block on Flex-OneNAND.
3111 The remaining blocks are configured as MLC blocks.
3112 lock - Configure if Flex-OneNAND boundary should be locked.
3113 Once locked, the boundary cannot be changed.
3114 1 indicates lock status, 0 indicates unlock status.
3117 ARM/S3C2412 JIVE boot control
3119 See arch/arm/mach-s3c/mach-jive.c
3121 mtouchusb.raw_coordinates=
3122 [HW] Make the MicroTouch USB driver use raw coordinates
3123 ('y', default) or cooked coordinates ('n')
3125 mtrr_chunk_size=nn[KMG] [X86]
3126 used for mtrr cleanup. It is largest continuous chunk
3127 that could hold holes aka. UC entries.
3129 mtrr_gran_size=nn[KMG] [X86]
3130 Used for mtrr cleanup. It is granularity of mtrr block.
3132 Large value could prevent small alignment from
3135 mtrr_spare_reg_nr=n [X86]
3137 Range: 0,7 : spare reg number
3139 Used for mtrr cleanup. It is spare mtrr entries number.
3140 Set to 2 or more if your graphical card needs more.
3142 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3144 netdev= [NET] Network devices parameters
3145 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3146 Note that mem_start is often overloaded to mean
3147 something different and driver-specific.
3148 This usage is only documented in each driver source
3152 [NETFILTER] Enable connection tracking flow accounting
3153 0 to disable accounting
3154 1 to enable accounting
3157 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3158 See Documentation/admin-guide/nfs/nfsroot.rst.
3160 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3161 See Documentation/admin-guide/nfs/nfsroot.rst.
3163 nfsrootdebug [NFS] enable nfsroot debugging messages.
3164 See Documentation/admin-guide/nfs/nfsroot.rst.
3166 nfs.callback_nr_threads=
3167 [NFSv4] set the total number of threads that the
3168 NFS client will assign to service NFSv4 callback
3171 nfs.callback_tcpport=
3172 [NFS] set the TCP port on which the NFSv4 callback
3173 channel should listen.
3176 [NFS] sets the pathname to the program which is used
3177 to update the NFS client cache entries.
3179 nfs.cache_getent_timeout=
3180 [NFS] sets the timeout after which an attempt to
3181 update a cache entry is deemed to have failed.
3183 nfs.idmap_cache_timeout=
3184 [NFS] set the maximum lifetime for idmapper cache
3188 [NFS] enable 64-bit inode numbers.
3189 If zero, the NFS client will fake up a 32-bit inode
3190 number for the readdir() and stat() syscalls instead
3191 of returning the full 64-bit number.
3192 The default is to return 64-bit inode numbers.
3194 nfs.max_session_cb_slots=
3195 [NFSv4.1] Sets the maximum number of session
3196 slots the client will assign to the callback
3197 channel. This determines the maximum number of
3198 callbacks the client will process in parallel for
3199 a particular server.
3201 nfs.max_session_slots=
3202 [NFSv4.1] Sets the maximum number of session slots
3203 the client will attempt to negotiate with the server.
3204 This limits the number of simultaneous RPC requests
3205 that the client can send to the NFSv4.1 server.
3206 Note that there is little point in setting this
3207 value higher than the max_tcp_slot_table_limit.
3209 nfs.nfs4_disable_idmapping=
3210 [NFSv4] When set to the default of '1', this option
3211 ensures that both the RPC level authentication
3212 scheme and the NFS level operations agree to use
3213 numeric uids/gids if the mount is using the
3214 'sec=sys' security flavour. In effect it is
3215 disabling idmapping, which can make migration from
3216 legacy NFSv2/v3 systems to NFSv4 easier.
3217 Servers that do not support this mode of operation
3218 will be autodetected by the client, and it will fall
3219 back to using the idmapper.
3220 To turn off this behaviour, set the value to '0'.
3222 [NFS4] Specify an additional fixed unique ident-
3223 ification string that NFSv4 clients can insert into
3224 their nfs_client_id4 string. This is typically a
3225 UUID that is generated at system install time.
3227 nfs.send_implementation_id =
3228 [NFSv4.1] Send client implementation identification
3229 information in exchange_id requests.
3230 If zero, no implementation identification information
3232 The default is to send the implementation identification
3235 nfs.recover_lost_locks =
3236 [NFSv4] Attempt to recover locks that were lost due
3237 to a lease timeout on the server. Please note that
3238 doing this risks data corruption, since there are
3239 no guarantees that the file will remain unchanged
3240 after the locks are lost.
3241 If you want to enable the kernel legacy behaviour of
3242 attempting to recover these locks, then set this
3244 The default parameter value of '0' causes the kernel
3245 not to attempt recovery of lost locks.
3247 nfs4.layoutstats_timer =
3248 [NFSv4.2] Change the rate at which the kernel sends
3249 layoutstats to the pNFS metadata server.
3251 Setting this to value to 0 causes the kernel to use
3252 whatever value is the default set by the layout
3253 driver. A non-zero value sets the minimum interval
3254 in seconds between layoutstats transmissions.
3256 nfsd.nfs4_disable_idmapping=
3257 [NFSv4] When set to the default of '1', the NFSv4
3258 server will return only numeric uids and gids to
3259 clients using auth_sys, and will accept numeric uids
3260 and gids from such clients. This is intended to ease
3261 migration from NFSv2/v3.
3263 nmi_backtrace.backtrace_idle [KNL]
3264 Dump stacks even of idle CPUs in response to an
3265 NMI stack-backtrace request.
3267 nmi_debug= [KNL,SH] Specify one or more actions to take
3268 when a NMI is triggered.
3269 Format: [state][,regs][,debounce][,die]
3271 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3272 Format: [panic,][nopanic,][num]
3274 0 - turn hardlockup detector in nmi_watchdog off
3275 1 - turn hardlockup detector in nmi_watchdog on
3276 When panic is specified, panic when an NMI watchdog
3277 timeout occurs (or 'nopanic' to not panic on an NMI
3278 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3279 To disable both hard and soft lockup detectors,
3280 please see 'nowatchdog'.
3281 This is useful when you use a panic=... timeout and
3282 need the box quickly up again.
3284 These settings can be accessed at runtime via
3285 the nmi_watchdog and hardlockup_panic sysctls.
3287 netpoll.carrier_timeout=
3288 [NET] Specifies amount of time (in seconds) that
3289 netpoll should wait for a carrier. By default netpoll
3292 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3293 emulation library even if a 387 maths coprocessor
3296 no5lvl [X86-64] Disable 5-level paging mode. Forces
3297 kernel to use 4-level paging instead.
3299 nofsgsbase [X86] Disables FSGSBASE instructions.
3302 [HW] Never suspend the console
3303 Disable suspending of consoles during suspend and
3304 hibernate operations. Once disabled, debugging
3305 messages can reach various consoles while the rest
3306 of the system is being put to sleep (ie, while
3307 debugging driver suspend/resume hooks). This may
3308 not work reliably with all consoles, but is known
3309 to work with serial and VGA consoles.
3310 To facilitate more flexible debugging, we also add
3311 console_suspend, a printk module parameter to control
3312 it. Users could use console_suspend (usually
3313 /sys/module/printk/parameters/console_suspend) to
3314 turn on/off it dynamically.
3316 novmcoredd [KNL,KDUMP]
3317 Disable device dump. Device dump allows drivers to
3318 append dump data to vmcore so you can collect driver
3319 specified debug info. Drivers can append the data
3320 without any limit and this data is stored in memory,
3321 so this may cause significant memory stress. Disabling
3322 device dump can help save memory but the driver debug
3323 data will be no longer available. This parameter
3324 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3327 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3328 caches in the slab allocator. Saves per-node memory,
3329 but will impact performance.
3333 noaltinstr [S390] Disables alternative instructions patching
3334 (CPU alternatives feature).
3336 noapic [SMP,APIC] Tells the kernel to not make use of any
3337 IOAPICs that may be present in the system.
3339 noautogroup Disable scheduler automatic task group creation.
3341 nobats [PPC] Do not use BATs for mapping kernel lowmem
3342 on "Classic" PPC cores.
3346 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3348 delayacct [KNL] Enable per-task delay accounting
3350 nodsp [SH] Disable hardware DSP at boot time.
3352 noefi Disable EFI runtime services support.
3354 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3359 On X86-32 available only on PAE configured kernels.
3360 noexec=on: enable non-executable mappings (default)
3361 noexec=off: disable non-executable mappings
3364 Disable SMAP (Supervisor Mode Access Prevention)
3365 even if it is supported by processor.
3368 Disable SMEP (Supervisor Mode Execution Prevention)
3369 even if it is supported by processor.
3372 This affects only 32-bit executables.
3373 noexec32=on: enable non-executable mappings (default)
3374 read doesn't imply executable mappings
3375 noexec32=off: disable non-executable mappings
3376 read implies executable mappings
3378 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3380 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3381 register save and restore. The kernel will only save
3382 legacy floating-point registers on task switch.
3384 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3386 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3388 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3389 Equivalent to smt=1.
3391 [KNL,X86] Disable symmetric multithreading (SMT).
3392 nosmt=force: Force disable SMT, cannot be undone
3393 via the sysfs control file.
3395 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3396 (bounds check bypass). With this option data leaks are
3397 possible in the system.
3399 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3400 the Spectre variant 2 (indirect branch prediction)
3401 vulnerability. System may allow data leaks with this
3404 nospec_store_bypass_disable
3405 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3408 [PPC] Don't flush the L1-D cache after accessing user data.
3410 noxsave [BUGS=X86] Disables x86 extended register state save
3411 and restore using xsave. The kernel will fallback to
3412 enabling legacy floating-point and sse state.
3414 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3415 register states. The kernel will fall back to use
3416 xsave to save the states. By using this parameter,
3417 performance of saving the states is degraded because
3418 xsave doesn't support modified optimization while
3419 xsaveopt supports it on xsaveopt enabled systems.
3421 noxsaves [X86] Disables xsaves and xrstors used in saving and
3422 restoring x86 extended register state in compacted
3423 form of xsave area. The kernel will fall back to use
3424 xsaveopt and xrstor to save and restore the states
3425 in standard form of xsave area. By using this
3426 parameter, xsave area per process might occupy more
3427 memory on xsaves enabled systems.
3429 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3430 in do_idle() and not use the arch_cpu_idle()
3431 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3432 to be effective. This is useful on platforms where the
3433 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3434 correctly or when doing power measurements to evalute
3435 the impact of the sleep instructions. This is also
3436 useful when using JTAG debugger.
3438 no_file_caps Tells the kernel not to honor file capabilities. The
3439 only way then for a file to be executed with privilege
3440 is to be setuid root or executed by root.
3442 nohalt [IA-64] Tells the kernel not to use the power saving
3443 function PAL_HALT_LIGHT when idle. This increases
3444 power-consumption. On the positive side, it reduces
3445 interrupt wake-up latency, which may improve performance
3446 in certain environments such as networked servers or
3450 Force pointers printed to the console or buffers to be
3451 unhashed. By default, when a pointer is printed via %p
3452 format string, that pointer is "hashed", i.e. obscured
3453 by hashing the pointer value. This is a security feature
3454 that hides actual kernel addresses from unprivileged
3455 users, but it also makes debugging the kernel more
3456 difficult since unequal pointers can no longer be
3457 compared. However, if this command-line option is
3458 specified, then all normal pointers will have their true
3459 value printed. Pointers printed via %pK may still be
3460 hashed. This option should only be specified when
3461 debugging the kernel. Please do not use on production
3464 nohibernate [HIBERNATION] Disable hibernation and resume.
3466 nohz= [KNL] Boottime enable/disable dynamic ticks
3467 Valid arguments: on, off
3470 nohz_full= [KNL,BOOT,SMP,ISOL]
3471 The argument is a cpu list, as described above.
3472 In kernels built with CONFIG_NO_HZ_FULL=y, set
3473 the specified list of CPUs whose tick will be stopped
3474 whenever possible. The boot CPU will be forced outside
3475 the range to maintain the timekeeping. Any CPUs
3476 in this list will have their RCU callbacks offloaded,
3477 just as if they had also been called out in the
3478 rcu_nocbs= boot parameter.
3480 noiotrap [SH] Disables trapped I/O port accesses.
3482 noirqdebug [X86-32] Disables the code which attempts to detect and
3483 disable unhandled interrupt sources.
3485 no_timer_check [X86,APIC] Disables the code which tests for
3486 broken timer IRQ sources.
3488 noisapnp [ISAPNP] Disables ISA PnP code.
3490 noinitrd [RAM] Tells the kernel not to load any configured
3493 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3495 [Deprecated - use intremap=off]
3499 noinvpcid [X86] Disable the INVPCID cpu feature.
3501 nojitter [IA-64] Disables jitter checking for ITC timers.
3503 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3505 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3509 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3510 clock and use the default one.
3512 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3513 accounting. steal time is computed, but won't
3514 influence scheduler behaviour
3516 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3518 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3520 noltlbs [PPC] Do not use large page/tlb entries for kernel
3521 lowmem mapping on PPC40x and PPC8xx
3523 nomca [IA-64] Disable machine check abort handling
3525 nomce [X86-32] Disable Machine Check Exception
3527 nomfgpt [X86-32] Disable Multi-Function General Purpose
3528 Timer usage (for AMD Geode machines).
3530 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3531 shutdown the other cpus. Instead use the REBOOT_VECTOR
3534 nomodule Disable module load
3536 nopat [X86] Disable PAT (page attribute table extension of
3537 pagetables) support.
3539 nopcid [X86-64] Disable the PCID cpu feature.
3541 norandmaps Don't use address space randomization. Equivalent to
3542 echo 0 > /proc/sys/kernel/randomize_va_space
3544 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3545 with UP alternatives
3547 nordrand [X86] Disable kernel use of the RDRAND and
3548 RDSEED instructions even if they are supported
3549 by the processor. RDRAND and RDSEED are still
3550 available to user space applications.
3552 noresume [SWSUSP] Disables resume and restores original swap
3555 no-scroll [VGA] Disables scrollback.
3556 This is required for the Braillex ib80-piezo Braille
3557 reader made by F.H. Papenmeier (Germany).
3561 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3563 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3565 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3566 and disable the IO APIC. legacy for "maxcpus=0".
3568 nosoftlockup [KNL] Disable the soft-lockup detector.
3570 nosync [HW,M68K] Disables sync negotiation for all devices.
3572 nowatchdog [KNL] Disable both lockup detectors, i.e.
3573 soft-lockup and NMI watchdog (hard-lockup).
3577 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3579 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3580 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3581 Some features depend on CPU0. Known dependencies are:
3582 1. Resume from suspend/hibernate depends on CPU0.
3583 Suspend/hibernate will fail if CPU0 is offline and you
3584 need to online CPU0 before suspend/hibernate.
3585 2. PIC interrupts also depend on CPU0. CPU0 can't be
3586 removed if a PIC interrupt is detected.
3587 It's said poweroff/reboot may depend on CPU0 on some
3588 machines although I haven't seen such issues so far
3589 after CPU0 is offline on a few tested machines.
3590 If the dependencies are under your control, you can
3591 turn on cpu0_hotplug.
3593 nps_mtm_hs_ctr= [KNL,ARC]
3594 This parameter sets the maximum duration, in
3595 cycles, each HW thread of the CTOP can run
3596 without interruptions, before HW switches it.
3597 The actual maximum duration is 16 times this
3599 Format: integer between 1 and 255
3602 nptcg= [IA-64] Override max number of concurrent global TLB
3603 purges which is reported from either PAL_VM_SUMMARY or
3606 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3607 could support. nr_cpus=n : n >= 1 limits the kernel to
3608 support 'n' processors. It could be larger than the
3609 number of already plugged CPU during bootup, later in
3610 runtime you can physically add extra cpu until it reaches
3611 n. So during boot up some boot time memory for per-cpu
3612 variables need be pre-allocated for later physical cpu
3615 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3617 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3618 set up a single NUMA node spanning all memory.
3620 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3622 Allowed values are enable and disable
3624 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3625 'node', 'default' can be specified
3626 This can be set from sysctl after boot.
3627 See Documentation/admin-guide/sysctl/vm.rst for details.
3629 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3630 See Documentation/core-api/debugging-via-ohci1394.rst for more
3633 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3634 Rather than timing out after 20 ms if an EC
3635 command is not properly ACKed, override the length
3636 of the timeout. We have interrupts disabled while
3637 waiting for the ACK, so if this is set too high
3638 interrupts *may* be lost!
3640 omap_mux= [OMAP] Override bootloader pin multiplexing.
3641 Format: <mux_mode0.mode_name=value>...
3642 For example, to override I2C bus2:
3643 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3645 oops=panic Always panic on oopses. Default is to just kill the
3646 process, but there is a small probability of
3647 deadlocking the machine.
3648 This will also cause panics on machine check exceptions.
3649 Useful together with panic=30 to trigger a reboot.
3652 [KNL] Boolean flag to control whether the page allocator
3653 should randomize its free lists. The randomization may
3654 be automatically enabled if the kernel detects it is
3655 running on a platform with a direct-mapped memory-side
3656 cache, and this parameter can be used to
3657 override/disable that behavior. The state of the flag
3658 can be read from sysfs at:
3659 /sys/module/page_alloc/parameters/shuffle.
3661 page_owner= [KNL] Boot-time page_owner enabling option.
3662 Storage of the information about who allocated
3663 each page is disabled in default. With this switch,
3665 on: enable the feature
3667 page_poison= [KNL] Boot-time parameter changing the state of
3668 poisoning on the buddy allocator, available with
3669 CONFIG_PAGE_POISONING=y.
3670 off: turn off poisoning (default)
3671 on: turn on poisoning
3673 page_reporting.page_reporting_order=
3674 [KNL] Minimal page reporting order
3676 Adjust the minimal page reporting order. The page
3677 reporting is disabled when it exceeds (MAX_ORDER-1).
3679 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3680 timeout > 0: seconds before rebooting
3681 timeout = 0: wait forever
3682 timeout < 0: reboot immediately
3685 panic_print= Bitmask for printing system info when panic happens.
3686 User can chose combination of the following bits:
3687 bit 0: print all tasks info
3688 bit 1: print system memory info
3689 bit 2: print timer info
3690 bit 3: print locks info if CONFIG_LOCKDEP is on
3691 bit 4: print ftrace buffer
3692 bit 5: print all printk messages in buffer
3694 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3695 Format: <hex>[,nousertaint]
3696 Hexadecimal bitmask representing the set of TAINT flags
3697 that will cause the kernel to panic when add_taint() is
3698 called with any of the flags in this set.
3699 The optional switch "nousertaint" can be utilized to
3700 prevent userspace forced crashes by writing to sysctl
3701 /proc/sys/kernel/tainted any flagset matching with the
3702 bitmask set on panic_on_taint.
3703 See Documentation/admin-guide/tainted-kernels.rst for
3704 extra details on the taint flags that users can pick
3705 to compose the bitmask to assign to panic_on_taint.
3707 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3710 crash_kexec_post_notifiers
3711 Run kdump after running panic-notifiers and dumping
3712 kmsg. This only for the users who doubt kdump always
3713 succeeds in any situation.
3714 Note that this also increases risks of kdump failure,
3715 because some panic notifiers can make the crashed
3716 kernel more unstable.
3718 parkbd.port= [HW] Parallel port number the keyboard adapter is
3719 connected to, default is 0.
3721 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3722 0 for XT, 1 for AT (default is AT).
3725 parport= [HW,PPT] Specify parallel ports. 0 disables.
3726 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3727 Use 'auto' to force the driver to use any
3728 IRQ/DMA settings detected (the default is to
3729 ignore detected IRQ/DMA settings because of
3730 possible conflicts). You can specify the base
3731 address, IRQ, and DMA settings; IRQ and DMA
3732 should be numbers, or 'auto' (for using detected
3733 settings on that particular port), or 'nofifo'
3734 (to avoid using a FIFO even if it is detected).
3735 Parallel ports are assigned in the order they
3736 are specified on the command line, starting
3739 parport_init_mode= [HW,PPT]
3740 Configure VIA parallel port to operate in
3741 a specific mode. This is necessary on Pegasos
3742 computer where firmware has no options for setting
3743 up parallel port mode and sets it to spp.
3744 Currently this function knows 686a and 8231 chips.
3745 Format: [spp|ps2|epp|ecp|ecpepp]
3747 pata_legacy.all= [HW,LIBATA]
3749 Set to non-zero to probe primary and secondary ISA
3750 port ranges on PCI systems where no PCI PATA device
3751 has been found at either range. Disabled by default.
3753 pata_legacy.autospeed= [HW,LIBATA]
3755 Set to non-zero if a chip is present that snoops speed
3756 changes. Disabled by default.
3758 pata_legacy.ht6560a= [HW,LIBATA]
3760 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3761 the secondary channel, or both channels respectively.
3762 Disabled by default.
3764 pata_legacy.ht6560b= [HW,LIBATA]
3766 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3767 the secondary channel, or both channels respectively.
3768 Disabled by default.
3770 pata_legacy.iordy_mask= [HW,LIBATA]
3772 IORDY enable mask. Set individual bits to allow IORDY
3773 for the respective channel. Bit 0 is for the first
3774 legacy channel handled by this driver, bit 1 is for
3775 the second channel, and so on. The sequence will often
3776 correspond to the primary legacy channel, the secondary
3777 legacy channel, and so on, but the handling of a PCI
3778 bus and the use of other driver options may interfere
3779 with the sequence. By default IORDY is allowed across
3782 pata_legacy.opti82c46x= [HW,LIBATA]
3784 Set to 1, 2, or 3 for Opti 82c611A on the primary
3785 channel, the secondary channel, or both channels
3786 respectively. Disabled by default.
3788 pata_legacy.opti82c611a= [HW,LIBATA]
3790 Set to 1, 2, or 3 for Opti 82c465MV on the primary
3791 channel, the secondary channel, or both channels
3792 respectively. Disabled by default.
3794 pata_legacy.pio_mask= [HW,LIBATA]
3796 PIO mode mask for autospeed devices. Set individual
3797 bits to allow the use of the respective PIO modes.
3798 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
3799 All modes allowed by default.
3801 pata_legacy.probe_all= [HW,LIBATA]
3803 Set to non-zero to probe tertiary and further ISA
3804 port ranges on PCI systems. Disabled by default.
3806 pata_legacy.probe_mask= [HW,LIBATA]
3808 Probe mask for legacy ISA PATA ports. Depending on
3809 platform configuration and the use of other driver
3810 options up to 6 legacy ports are supported: 0x1f0,
3811 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
3812 of individual ports can be disabled by setting the
3813 corresponding bits in the mask to 1. Bit 0 is for
3814 the first port in the list above (0x1f0), and so on.
3815 By default all supported ports are probed.
3817 pata_legacy.qdi= [HW,LIBATA]
3819 Set to non-zero to probe QDI controllers. By default
3820 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
3822 pata_legacy.winbond= [HW,LIBATA]
3824 Set to non-zero to probe Winbond controllers. Use
3825 the standard I/O port (0x130) if 1, otherwise the
3826 value given is the I/O port to use (typically 0x1b0).
3827 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
3830 pata_platform.pio_mask= [HW,LIBATA]
3832 Supported PIO mode mask. Set individual bits to allow
3833 the use of the respective PIO modes. Bit 0 is for
3834 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
3838 Halt all CPUs after the first oops has been printed for
3839 the specified number of seconds. This is to be used if
3840 your oopses keep scrolling off the screen.
3845 See header of drivers/block/paride/pcd.c.
3846 See also Documentation/admin-guide/blockdev/paride.rst.
3848 pci=option[,option...] [PCI] various PCI subsystem options.
3850 Some options herein operate on a specific device
3851 or a set of devices (<pci_dev>). These are
3852 specified in one of the following formats:
3854 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3855 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3857 Note: the first format specifies a PCI
3858 bus/device/function address which may change
3859 if new hardware is inserted, if motherboard
3860 firmware changes, or due to changes caused
3861 by other kernel parameters. If the
3862 domain is left unspecified, it is
3863 taken to be zero. Optionally, a path
3864 to a device through multiple device/function
3865 addresses can be specified after the base
3866 address (this is more robust against
3867 renumbering issues). The second format
3868 selects devices using IDs from the
3869 configuration space which may match multiple
3870 devices in the system.
3872 earlydump dump PCI config space before the kernel
3874 off [X86] don't probe for the PCI bus
3875 bios [X86-32] force use of PCI BIOS, don't access
3876 the hardware directly. Use this if your machine
3877 has a non-standard PCI host bridge.
3878 nobios [X86-32] disallow use of PCI BIOS, only direct
3879 hardware access methods are allowed. Use this
3880 if you experience crashes upon bootup and you
3881 suspect they are caused by the BIOS.
3882 conf1 [X86] Force use of PCI Configuration Access
3883 Mechanism 1 (config address in IO port 0xCF8,
3884 data in IO port 0xCFC, both 32-bit).
3885 conf2 [X86] Force use of PCI Configuration Access
3886 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3887 the function, IO port 0xCFA, also 8-bit, sets
3888 bus number. The config space is then accessed
3889 through ports 0xC000-0xCFFF).
3890 See http://wiki.osdev.org/PCI for more info
3891 on the configuration access mechanisms.
3892 noaer [PCIE] If the PCIEAER kernel config parameter is
3893 enabled, this kernel boot option can be used to
3894 disable the use of PCIE advanced error reporting.
3895 nodomains [PCI] Disable support for multiple PCI
3896 root domains (aka PCI segments, in ACPI-speak).
3897 nommconf [X86] Disable use of MMCONFIG for PCI
3899 check_enable_amd_mmconf [X86] check for and enable
3900 properly configured MMIO access to PCI
3901 config space on AMD family 10h CPU
3902 nomsi [MSI] If the PCI_MSI kernel config parameter is
3903 enabled, this kernel boot option can be used to
3904 disable the use of MSI interrupts system-wide.
3905 noioapicquirk [APIC] Disable all boot interrupt quirks.
3906 Safety option to keep boot IRQs enabled. This
3907 should never be necessary.
3908 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3909 primary IO-APIC for bridges that cannot disable
3910 boot IRQs. This fixes a source of spurious IRQs
3911 when the system masks IRQs.
3912 noioapicreroute [APIC] Disable workaround that uses the
3913 boot IRQ equivalent of an IRQ that connects to
3914 a chipset where boot IRQs cannot be disabled.
3915 The opposite of ioapicreroute.
3916 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3917 routing table. These calls are known to be buggy
3918 on several machines and they hang the machine
3919 when used, but on other computers it's the only
3920 way to get the interrupt routing table. Try
3921 this option if the kernel is unable to allocate
3922 IRQs or discover secondary PCI buses on your
3924 rom [X86] Assign address space to expansion ROMs.
3925 Use with caution as certain devices share
3926 address decoders between ROMs and other
3928 norom [X86] Do not assign address space to
3929 expansion ROMs that do not already have
3930 BIOS assigned address ranges.
3931 nobar [X86] Do not assign address space to the
3932 BARs that weren't assigned by the BIOS.
3933 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3934 assigned automatically to PCI devices. You can
3935 make the kernel exclude IRQs of your ISA cards
3937 pirqaddr=0xAAAAA [X86] Specify the physical address
3938 of the PIRQ table (normally generated
3939 by the BIOS) if it is outside the
3940 F0000h-100000h range.
3941 lastbus=N [X86] Scan all buses thru bus #N. Can be
3942 useful if the kernel is unable to find your
3943 secondary buses and you want to tell it
3944 explicitly which ones they are.
3945 assign-busses [X86] Always assign all PCI bus
3946 numbers ourselves, overriding
3947 whatever the firmware may have done.
3948 usepirqmask [X86] Honor the possible IRQ mask stored
3949 in the BIOS $PIR table. This is needed on
3950 some systems with broken BIOSes, notably
3951 some HP Pavilion N5400 and Omnibook XE3
3952 notebooks. This will have no effect if ACPI
3953 IRQ routing is enabled.
3954 noacpi [X86] Do not use ACPI for IRQ routing
3955 or for PCI scanning.
3956 use_crs [X86] Use PCI host bridge window information
3957 from ACPI. On BIOSes from 2008 or later, this
3958 is enabled by default. If you need to use this,
3959 please report a bug.
3960 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3961 If you need to use this, please report a bug.
3962 routeirq Do IRQ routing for all PCI devices.
3963 This is normally done in pci_enable_device(),
3964 so this option is a temporary workaround
3965 for broken drivers that don't call it.
3966 skip_isa_align [X86] do not align io start addr, so can
3967 handle more pci cards
3968 noearly [X86] Don't do any early type 1 scanning.
3969 This might help on some broken boards which
3970 machine check when some devices' config space
3971 is read. But various workarounds are disabled
3972 and some IOMMU drivers will not work.
3973 bfsort Sort PCI devices into breadth-first order.
3974 This sorting is done to get a device
3975 order compatible with older (<= 2.4) kernels.
3976 nobfsort Don't sort PCI devices into breadth-first order.
3977 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3978 tuning and use the BIOS-configured MPS defaults.
3979 pcie_bus_safe Set every device's MPS to the largest value
3980 supported by all devices below the root complex.
3981 pcie_bus_perf Set device MPS to the largest allowable MPS
3982 based on its parent bus. Also set MRRS (Max
3983 Read Request Size) to the largest supported
3984 value (no larger than the MPS that the device
3985 or bus can support) for best performance.
3986 pcie_bus_peer2peer Set every device's MPS to 128B, which
3987 every device is guaranteed to support. This
3988 configuration allows peer-to-peer DMA between
3989 any pair of devices, possibly at the cost of
3990 reduced performance. This also guarantees
3991 that hot-added devices will work.
3992 cbiosize=nn[KMG] The fixed amount of bus space which is
3993 reserved for the CardBus bridge's IO window.
3994 The default value is 256 bytes.
3995 cbmemsize=nn[KMG] The fixed amount of bus space which is
3996 reserved for the CardBus bridge's memory
3997 window. The default value is 64 megabytes.
4000 [<order of align>@]<pci_dev>[; ...]
4001 Specifies alignment and device to reassign
4002 aligned memory resources. How to
4003 specify the device is described above.
4004 If <order of align> is not specified,
4005 PAGE_SIZE is used as alignment.
4006 A PCI-PCI bridge can be specified if resource
4007 windows need to be expanded.
4008 To specify the alignment for several
4009 instances of a device, the PCI vendor,
4010 device, subvendor, and subdevice may be
4011 specified, e.g., 12@pci:8086:9c22:103c:198f
4012 for 4096-byte alignment.
4013 ecrc= Enable/disable PCIe ECRC (transaction layer
4014 end-to-end CRC checking).
4015 bios: Use BIOS/firmware settings. This is the
4019 hpiosize=nn[KMG] The fixed amount of bus space which is
4020 reserved for hotplug bridge's IO window.
4021 Default size is 256 bytes.
4022 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4023 reserved for hotplug bridge's MMIO window.
4024 Default size is 2 megabytes.
4025 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4026 reserved for hotplug bridge's MMIO_PREF window.
4027 Default size is 2 megabytes.
4028 hpmemsize=nn[KMG] The fixed amount of bus space which is
4029 reserved for hotplug bridge's MMIO and
4031 Default size is 2 megabytes.
4032 hpbussize=nn The minimum amount of additional bus numbers
4033 reserved for buses below a hotplug bridge.
4035 realloc= Enable/disable reallocating PCI bridge resources
4036 if allocations done by BIOS are too small to
4037 accommodate resources required by all child
4039 off: Turn realloc off
4041 realloc same as realloc=on
4042 noari do not use PCIe ARI.
4043 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4044 do not use PCIe ATS (and IOMMU device IOTLB).
4045 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4046 only look for one device below a PCIe downstream
4048 big_root_window Try to add a big 64bit memory window to the PCIe
4049 root complex on AMD CPUs. Some GFX hardware
4050 can resize a BAR to allow access to all VRAM.
4051 Adding the window is slightly risky (it may
4052 conflict with unreported devices), so this
4054 disable_acs_redir=<pci_dev>[; ...]
4055 Specify one or more PCI devices (in the format
4056 specified above) separated by semicolons.
4057 Each device specified will have the PCI ACS
4058 redirect capabilities forced off which will
4059 allow P2P traffic between devices through
4060 bridges without forcing it upstream. Note:
4061 this removes isolation between devices and
4062 may put more devices in an IOMMU group.
4063 force_floating [S390] Force usage of floating interrupts.
4064 nomio [S390] Do not use MIO instructions.
4065 norid [S390] ignore the RID field and force use of
4066 one PCI domain per PCI function
4068 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4071 force Enable ASPM even on devices that claim not to support it.
4072 WARNING: Forcing ASPM on may cause system lockups.
4074 pcie_ports= [PCIE] PCIe port services handling:
4075 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4076 even if the platform doesn't give the OS permission to
4077 use them. This may cause conflicts if the platform
4078 also tries to use these services.
4079 dpc-native Use native PCIe service for DPC only. May
4080 cause conflicts if firmware uses AER or DPC.
4081 compat Disable native PCIe services (PME, AER, DPC, PCIe
4084 pcie_port_pm= [PCIE] PCIe port power management handling:
4085 off Disable power management of all PCIe ports
4086 force Forcibly enable power management of all PCIe ports
4088 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4089 nomsi Do not use MSI for native PCIe PME signaling (this makes
4090 all PCIe root ports use INTx for all services).
4092 pcmv= [HW,PCMCIA] BadgePAD 4
4096 Keep all power-domains already enabled by bootloader on,
4097 even if no driver has claimed them. This is useful
4098 for debug and development, but should not be
4099 needed on a platform with proper driver support.
4102 See Documentation/admin-guide/blockdev/paride.rst.
4104 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4107 See arch/parisc/kernel/pdc_chassis.c
4109 percpu_alloc= Select which percpu first chunk allocator to use.
4110 Currently supported values are "embed" and "page".
4111 Archs may support subset or none of the selections.
4112 See comments in mm/percpu.c for details on each
4113 allocator. This parameter is primarily for debugging
4114 and performance comparison.
4117 See Documentation/admin-guide/blockdev/paride.rst.
4120 See Documentation/admin-guide/blockdev/paride.rst.
4122 pirq= [SMP,APIC] Manual mp-table setup
4123 See Documentation/x86/i386/IO-APIC.rst.
4125 plip= [PPT,NET] Parallel port network link
4126 Format: { parport<nr> | timid | 0 }
4127 See also Documentation/admin-guide/parport.rst.
4129 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4130 Override pmtimer IOPort with a hex value.
4133 pm_debug_messages [SUSPEND,KNL]
4134 Enable suspend/resume debug messages during boot up.
4137 Enable PNP debug messages (depends on the
4138 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4139 via /sys/module/pnp/parameters/debug. We always show
4140 current resource usage; turning this on also shows
4141 possible settings and some assignment information.
4147 { on | off | curr | res | no-curr | no-res }
4150 [ISAPNP] Exclude IRQs for the autoconfiguration
4153 [ISAPNP] Exclude DMAs for the autoconfiguration
4155 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4156 Ranges are in pairs (I/O port base and size).
4159 [ISAPNP] Exclude memory regions for the
4161 Ranges are in pairs (memory base and size).
4163 ports= [IP_VS_FTP] IPVS ftp helper module
4165 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4167 Format: <port>,<port>....
4169 powersave=off [PPC] This option disables power saving features.
4170 It specifically disables cpuidle and sets the
4171 platform machine description specific power_save
4172 function to NULL. On Idle the CPU just reduces
4175 ppc_strict_facility_enable
4176 [PPC] This option catches any kernel floating point,
4177 Altivec, VSX and SPE outside of regions specifically
4178 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4179 There is some performance impact when enabling this.
4183 Disable Hardware Transactional Memory
4186 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4187 none - Limited to cond_resched() calls
4188 voluntary - Limited to cond_resched() and might_sleep() calls
4189 full - Any section that isn't explicitly preempt disabled
4190 can be preempted anytime.
4192 print-fatal-signals=
4193 [KNL] debug: print fatal signals
4195 If enabled, warn about various signal handling
4196 related application anomalies: too many signals,
4197 too many POSIX.1 timers, fatal signals causing a
4200 If you hit the warning due to signal overflow,
4201 you might want to try "ulimit -i unlimited".
4205 printk.always_kmsg_dump=
4206 Trigger kmsg_dump for cases other than kernel oops or
4208 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4211 printk.console_no_auto_verbose=
4212 Disable console loglevel raise on oops, panic
4213 or lockdep-detected issues (only if lock debug is on).
4214 With an exception to setups with low baudrate on
4215 serial console, keeping this 0 is a good choice
4216 in order to provide more debug information.
4218 default: 0 (auto_verbose is enabled)
4220 printk.devkmsg={on,off,ratelimit}
4221 Control writing to /dev/kmsg.
4222 on - unlimited logging to /dev/kmsg from userspace
4223 off - logging to /dev/kmsg disabled
4224 ratelimit - ratelimit the logging
4227 printk.time= Show timing data prefixed to each printk message line
4228 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4230 processor.max_cstate= [HW,ACPI]
4231 Limit processor to maximum C-state
4232 max_cstate=9 overrides any DMI blacklist limit.
4234 processor.nocst [HW,ACPI]
4235 Ignore the _CST method to determine C-states,
4236 instead using the legacy FADT method
4238 profile= [KNL] Enable kernel profiling via /proc/profile
4239 Format: [<profiletype>,]<number>
4240 Param: <profiletype>: "schedule", "sleep", or "kvm"
4241 [defaults to kernel profiling]
4242 Param: "schedule" - profile schedule points.
4243 Param: "sleep" - profile D-state sleeping (millisecs).
4244 Requires CONFIG_SCHEDSTATS
4245 Param: "kvm" - profile VM exits.
4246 Param: <number> - step/bucket size as a power of 2 for
4247 statistical time based profiling.
4249 prompt_ramdisk= [RAM] [Deprecated]
4251 prot_virt= [S390] enable hosting protected virtual machines
4252 isolated from the hypervisor (if hardware supports
4256 psi= [KNL] Enable or disable pressure stall information
4260 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4261 probe for; one of (bare|imps|exps|lifebook|any).
4262 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4264 psmouse.resetafter= [HW,MOUSE]
4265 Try to reset the device after so many bad packets
4268 [HW,MOUSE] Set desired mouse resolution, in dpi.
4269 psmouse.smartscroll=
4270 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4271 0 = disabled, 1 = enabled (default).
4273 pstore.backend= Specify the name of the pstore backend to use
4276 See Documentation/admin-guide/blockdev/paride.rst.
4278 pti= [X86-64] Control Page Table Isolation of user and
4279 kernel address spaces. Disabling this feature
4280 removes hardening, but improves performance of
4281 system calls and interrupts.
4283 on - unconditionally enable
4284 off - unconditionally disable
4285 auto - kernel detects whether your CPU model is
4286 vulnerable to issues that PTI mitigates
4288 Not specifying this option is equivalent to pti=auto.
4291 Equivalent to pti=off
4294 [KNL] Number of legacy pty's. Overwrites compiled-in
4297 quiet [KNL] Disable most log messages
4302 See Documentation/admin-guide/md.rst.
4304 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4305 See Documentation/admin-guide/blockdev/ramdisk.rst.
4307 ramdisk_start= [RAM] RAM disk image start address
4309 random.trust_cpu={on,off}
4310 [KNL] Enable or disable trusting the use of the
4311 CPU's random number generator (if available) to
4312 fully seed the kernel's CRNG. Default is controlled
4313 by CONFIG_RANDOM_TRUST_CPU.
4315 randomize_kstack_offset=
4316 [KNL] Enable or disable kernel stack offset
4317 randomization, which provides roughly 5 bits of
4318 entropy, frustrating memory corruption attacks
4319 that depend on stack address determinism or
4320 cross-syscall address exposures. This is only
4321 available on architectures that have defined
4322 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4323 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4324 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4326 ras=option[,option,...] [KNL] RAS-specific options
4329 Disable the Correctable Errors Collector,
4330 see CONFIG_RAS_CEC help text.
4333 The argument is a cpu list, as described above.
4335 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
4336 the specified list of CPUs to be no-callback CPUs.
4337 Invocation of these CPUs' RCU callbacks will be
4338 offloaded to "rcuox/N" kthreads created for that
4339 purpose, where "x" is "p" for RCU-preempt, and
4340 "s" for RCU-sched, and "N" is the CPU number.
4341 This reduces OS jitter on the offloaded CPUs,
4342 which can be useful for HPC and real-time
4343 workloads. It can also improve energy efficiency
4344 for asymmetric multiprocessors.
4347 Rather than requiring that offloaded CPUs
4348 (specified by rcu_nocbs= above) explicitly
4349 awaken the corresponding "rcuoN" kthreads,
4350 make these kthreads poll for callbacks.
4351 This improves the real-time response for the
4352 offloaded CPUs by relieving them of the need to
4353 wake up the corresponding kthread, but degrades
4354 energy efficiency by requiring that the kthreads
4355 periodically wake up to do the polling.
4357 rcutree.blimit= [KNL]
4358 Set maximum number of finished RCU callbacks to
4359 process in one batch.
4361 rcutree.dump_tree= [KNL]
4362 Dump the structure of the rcu_node combining tree
4363 out at early boot. This is used for diagnostic
4364 purposes, to verify correct tree setup.
4366 rcutree.gp_cleanup_delay= [KNL]
4367 Set the number of jiffies to delay each step of
4368 RCU grace-period cleanup.
4370 rcutree.gp_init_delay= [KNL]
4371 Set the number of jiffies to delay each step of
4372 RCU grace-period initialization.
4374 rcutree.gp_preinit_delay= [KNL]
4375 Set the number of jiffies to delay each step of
4376 RCU grace-period pre-initialization, that is,
4377 the propagation of recent CPU-hotplug changes up
4378 the rcu_node combining tree.
4380 rcutree.use_softirq= [KNL]
4381 If set to zero, move all RCU_SOFTIRQ processing to
4382 per-CPU rcuc kthreads. Defaults to a non-zero
4383 value, meaning that RCU_SOFTIRQ is used by default.
4384 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4386 But note that CONFIG_PREEMPT_RT=y kernels disable
4387 this kernel boot parameter, forcibly setting it
4390 rcutree.rcu_fanout_exact= [KNL]
4391 Disable autobalancing of the rcu_node combining
4392 tree. This is used by rcutorture, and might
4393 possibly be useful for architectures having high
4394 cache-to-cache transfer latencies.
4396 rcutree.rcu_fanout_leaf= [KNL]
4397 Change the number of CPUs assigned to each
4398 leaf rcu_node structure. Useful for very
4399 large systems, which will choose the value 64,
4400 and for NUMA systems with large remote-access
4401 latencies, which will choose a value aligned
4402 with the appropriate hardware boundaries.
4404 rcutree.rcu_min_cached_objs= [KNL]
4405 Minimum number of objects which are cached and
4406 maintained per one CPU. Object size is equal
4407 to PAGE_SIZE. The cache allows to reduce the
4408 pressure to page allocator, also it makes the
4409 whole algorithm to behave better in low memory
4412 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4413 Set the page-cache refill delay (in milliseconds)
4414 in response to low-memory conditions. The range
4415 of permitted values is in the range 0:100000.
4417 rcutree.jiffies_till_first_fqs= [KNL]
4418 Set delay from grace-period initialization to
4419 first attempt to force quiescent states.
4420 Units are jiffies, minimum value is zero,
4421 and maximum value is HZ.
4423 rcutree.jiffies_till_next_fqs= [KNL]
4424 Set delay between subsequent attempts to force
4425 quiescent states. Units are jiffies, minimum
4426 value is one, and maximum value is HZ.
4428 rcutree.jiffies_till_sched_qs= [KNL]
4429 Set required age in jiffies for a
4430 given grace period before RCU starts
4431 soliciting quiescent-state help from
4432 rcu_note_context_switch() and cond_resched().
4433 If not specified, the kernel will calculate
4434 a value based on the most recent settings
4435 of rcutree.jiffies_till_first_fqs
4436 and rcutree.jiffies_till_next_fqs.
4437 This calculated value may be viewed in
4438 rcutree.jiffies_to_sched_qs. Any attempt to set
4439 rcutree.jiffies_to_sched_qs will be cheerfully
4442 rcutree.kthread_prio= [KNL,BOOT]
4443 Set the SCHED_FIFO priority of the RCU per-CPU
4444 kthreads (rcuc/N). This value is also used for
4445 the priority of the RCU boost threads (rcub/N)
4446 and for the RCU grace-period kthreads (rcu_bh,
4447 rcu_preempt, and rcu_sched). If RCU_BOOST is
4448 set, valid values are 1-99 and the default is 1
4449 (the least-favored priority). Otherwise, when
4450 RCU_BOOST is not set, valid values are 0-99 and
4451 the default is zero (non-realtime operation).
4453 rcutree.rcu_nocb_gp_stride= [KNL]
4454 Set the number of NOCB callback kthreads in
4455 each group, which defaults to the square root
4456 of the number of CPUs. Larger numbers reduce
4457 the wakeup overhead on the global grace-period
4458 kthread, but increases that same overhead on
4459 each group's NOCB grace-period kthread.
4461 rcutree.qhimark= [KNL]
4462 Set threshold of queued RCU callbacks beyond which
4463 batch limiting is disabled.
4465 rcutree.qlowmark= [KNL]
4466 Set threshold of queued RCU callbacks below which
4467 batch limiting is re-enabled.
4469 rcutree.qovld= [KNL]
4470 Set threshold of queued RCU callbacks beyond which
4471 RCU's force-quiescent-state scan will aggressively
4472 enlist help from cond_resched() and sched IPIs to
4473 help CPUs more quickly reach quiescent states.
4474 Set to less than zero to make this be set based
4475 on rcutree.qhimark at boot time and to zero to
4476 disable more aggressive help enlistment.
4478 rcutree.rcu_idle_gp_delay= [KNL]
4479 Set wakeup interval for idle CPUs that have
4480 RCU callbacks (RCU_FAST_NO_HZ=y).
4482 rcutree.rcu_kick_kthreads= [KNL]
4483 Cause the grace-period kthread to get an extra
4484 wake_up() if it sleeps three times longer than
4485 it should at force-quiescent-state time.
4486 This wake_up() will be accompanied by a
4487 WARN_ONCE() splat and an ftrace_dump().
4489 rcutree.rcu_unlock_delay= [KNL]
4490 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4491 this specifies an rcu_read_unlock()-time delay
4492 in microseconds. This defaults to zero.
4493 Larger delays increase the probability of
4494 catching RCU pointer leaks, that is, buggy use
4495 of RCU-protected pointers after the relevant
4496 rcu_read_unlock() has completed.
4498 rcutree.sysrq_rcu= [KNL]
4499 Commandeer a sysrq key to dump out Tree RCU's
4500 rcu_node tree with an eye towards determining
4501 why a new grace period has not yet started.
4503 rcuscale.gp_async= [KNL]
4504 Measure performance of asynchronous
4505 grace-period primitives such as call_rcu().
4507 rcuscale.gp_async_max= [KNL]
4508 Specify the maximum number of outstanding
4509 callbacks per writer thread. When a writer
4510 thread exceeds this limit, it invokes the
4511 corresponding flavor of rcu_barrier() to allow
4512 previously posted callbacks to drain.
4514 rcuscale.gp_exp= [KNL]
4515 Measure performance of expedited synchronous
4516 grace-period primitives.
4518 rcuscale.holdoff= [KNL]
4519 Set test-start holdoff period. The purpose of
4520 this parameter is to delay the start of the
4521 test until boot completes in order to avoid
4524 rcuscale.kfree_rcu_test= [KNL]
4525 Set to measure performance of kfree_rcu() flooding.
4527 rcuscale.kfree_rcu_test_double= [KNL]
4528 Test the double-argument variant of kfree_rcu().
4529 If this parameter has the same value as
4530 rcuscale.kfree_rcu_test_single, both the single-
4531 and double-argument variants are tested.
4533 rcuscale.kfree_rcu_test_single= [KNL]
4534 Test the single-argument variant of kfree_rcu().
4535 If this parameter has the same value as
4536 rcuscale.kfree_rcu_test_double, both the single-
4537 and double-argument variants are tested.
4539 rcuscale.kfree_nthreads= [KNL]
4540 The number of threads running loops of kfree_rcu().
4542 rcuscale.kfree_alloc_num= [KNL]
4543 Number of allocations and frees done in an iteration.
4545 rcuscale.kfree_loops= [KNL]
4546 Number of loops doing rcuscale.kfree_alloc_num number
4547 of allocations and frees.
4549 rcuscale.nreaders= [KNL]
4550 Set number of RCU readers. The value -1 selects
4551 N, where N is the number of CPUs. A value
4552 "n" less than -1 selects N-n+1, where N is again
4553 the number of CPUs. For example, -2 selects N
4554 (the number of CPUs), -3 selects N+1, and so on.
4555 A value of "n" less than or equal to -N selects
4558 rcuscale.nwriters= [KNL]
4559 Set number of RCU writers. The values operate
4560 the same as for rcuscale.nreaders.
4561 N, where N is the number of CPUs
4563 rcuscale.perf_type= [KNL]
4564 Specify the RCU implementation to test.
4566 rcuscale.shutdown= [KNL]
4567 Shut the system down after performance tests
4568 complete. This is useful for hands-off automated
4571 rcuscale.verbose= [KNL]
4572 Enable additional printk() statements.
4574 rcuscale.writer_holdoff= [KNL]
4575 Write-side holdoff between grace periods,
4576 in microseconds. The default of zero says
4579 rcutorture.fqs_duration= [KNL]
4580 Set duration of force_quiescent_state bursts
4583 rcutorture.fqs_holdoff= [KNL]
4584 Set holdoff time within force_quiescent_state bursts
4587 rcutorture.fqs_stutter= [KNL]
4588 Set wait time between force_quiescent_state bursts
4591 rcutorture.fwd_progress= [KNL]
4592 Enable RCU grace-period forward-progress testing
4593 for the types of RCU supporting this notion.
4595 rcutorture.fwd_progress_div= [KNL]
4596 Specify the fraction of a CPU-stall-warning
4597 period to do tight-loop forward-progress testing.
4599 rcutorture.fwd_progress_holdoff= [KNL]
4600 Number of seconds to wait between successive
4601 forward-progress tests.
4603 rcutorture.fwd_progress_need_resched= [KNL]
4604 Enclose cond_resched() calls within checks for
4605 need_resched() during tight-loop forward-progress
4608 rcutorture.gp_cond= [KNL]
4609 Use conditional/asynchronous update-side
4610 primitives, if available.
4612 rcutorture.gp_exp= [KNL]
4613 Use expedited update-side primitives, if available.
4615 rcutorture.gp_normal= [KNL]
4616 Use normal (non-expedited) asynchronous
4617 update-side primitives, if available.
4619 rcutorture.gp_sync= [KNL]
4620 Use normal (non-expedited) synchronous
4621 update-side primitives, if available. If all
4622 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4623 rcutorture.gp_normal=, and rcutorture.gp_sync=
4624 are zero, rcutorture acts as if is interpreted
4625 they are all non-zero.
4627 rcutorture.irqreader= [KNL]
4628 Run RCU readers from irq handlers, or, more
4629 accurately, from a timer handler. Not all RCU
4630 flavors take kindly to this sort of thing.
4632 rcutorture.leakpointer= [KNL]
4633 Leak an RCU-protected pointer out of the reader.
4634 This can of course result in splats, and is
4635 intended to test the ability of things like
4636 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4639 rcutorture.n_barrier_cbs= [KNL]
4640 Set callbacks/threads for rcu_barrier() testing.
4642 rcutorture.nfakewriters= [KNL]
4643 Set number of concurrent RCU writers. These just
4644 stress RCU, they don't participate in the actual
4645 test, hence the "fake".
4647 rcutorture.nocbs_nthreads= [KNL]
4648 Set number of RCU callback-offload togglers.
4649 Zero (the default) disables toggling.
4651 rcutorture.nocbs_toggle= [KNL]
4652 Set the delay in milliseconds between successive
4653 callback-offload toggling attempts.
4655 rcutorture.nreaders= [KNL]
4656 Set number of RCU readers. The value -1 selects
4657 N-1, where N is the number of CPUs. A value
4658 "n" less than -1 selects N-n-2, where N is again
4659 the number of CPUs. For example, -2 selects N
4660 (the number of CPUs), -3 selects N+1, and so on.
4662 rcutorture.object_debug= [KNL]
4663 Enable debug-object double-call_rcu() testing.
4665 rcutorture.onoff_holdoff= [KNL]
4666 Set time (s) after boot for CPU-hotplug testing.
4668 rcutorture.onoff_interval= [KNL]
4669 Set time (jiffies) between CPU-hotplug operations,
4670 or zero to disable CPU-hotplug testing.
4672 rcutorture.read_exit= [KNL]
4673 Set the number of read-then-exit kthreads used
4674 to test the interaction of RCU updaters and
4675 task-exit processing.
4677 rcutorture.read_exit_burst= [KNL]
4678 The number of times in a given read-then-exit
4679 episode that a set of read-then-exit kthreads
4682 rcutorture.read_exit_delay= [KNL]
4683 The delay, in seconds, between successive
4684 read-then-exit testing episodes.
4686 rcutorture.shuffle_interval= [KNL]
4687 Set task-shuffle interval (s). Shuffling tasks
4688 allows some CPUs to go into dyntick-idle mode
4689 during the rcutorture test.
4691 rcutorture.shutdown_secs= [KNL]
4692 Set time (s) after boot system shutdown. This
4693 is useful for hands-off automated testing.
4695 rcutorture.stall_cpu= [KNL]
4696 Duration of CPU stall (s) to test RCU CPU stall
4697 warnings, zero to disable.
4699 rcutorture.stall_cpu_block= [KNL]
4700 Sleep while stalling if set. This will result
4701 in warnings from preemptible RCU in addition
4702 to any other stall-related activity.
4704 rcutorture.stall_cpu_holdoff= [KNL]
4705 Time to wait (s) after boot before inducing stall.
4707 rcutorture.stall_cpu_irqsoff= [KNL]
4708 Disable interrupts while stalling if set.
4710 rcutorture.stall_gp_kthread= [KNL]
4711 Duration (s) of forced sleep within RCU
4712 grace-period kthread to test RCU CPU stall
4713 warnings, zero to disable. If both stall_cpu
4714 and stall_gp_kthread are specified, the
4715 kthread is starved first, then the CPU.
4717 rcutorture.stat_interval= [KNL]
4718 Time (s) between statistics printk()s.
4720 rcutorture.stutter= [KNL]
4721 Time (s) to stutter testing, for example, specifying
4722 five seconds causes the test to run for five seconds,
4723 wait for five seconds, and so on. This tests RCU's
4724 ability to transition abruptly to and from idle.
4726 rcutorture.test_boost= [KNL]
4727 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4728 "Maybe" means test if the RCU implementation
4729 under test support RCU priority boosting.
4731 rcutorture.test_boost_duration= [KNL]
4732 Duration (s) of each individual boost test.
4734 rcutorture.test_boost_interval= [KNL]
4735 Interval (s) between each boost test.
4737 rcutorture.test_no_idle_hz= [KNL]
4738 Test RCU's dyntick-idle handling. See also the
4739 rcutorture.shuffle_interval parameter.
4741 rcutorture.torture_type= [KNL]
4742 Specify the RCU implementation to test.
4744 rcutorture.verbose= [KNL]
4745 Enable additional printk() statements.
4747 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4748 Dump ftrace buffer after reporting RCU CPU
4751 rcupdate.rcu_cpu_stall_suppress= [KNL]
4752 Suppress RCU CPU stall warning messages.
4754 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4755 Suppress RCU CPU stall warning messages and
4756 rcutorture writer stall warnings that occur
4757 during early boot, that is, during the time
4758 before the init task is spawned.
4760 rcupdate.rcu_cpu_stall_timeout= [KNL]
4761 Set timeout for RCU CPU stall warning messages.
4763 rcupdate.rcu_expedited= [KNL]
4764 Use expedited grace-period primitives, for
4765 example, synchronize_rcu_expedited() instead
4766 of synchronize_rcu(). This reduces latency,
4767 but can increase CPU utilization, degrade
4768 real-time latency, and degrade energy efficiency.
4769 No effect on CONFIG_TINY_RCU kernels.
4771 rcupdate.rcu_normal= [KNL]
4772 Use only normal grace-period primitives,
4773 for example, synchronize_rcu() instead of
4774 synchronize_rcu_expedited(). This improves
4775 real-time latency, CPU utilization, and
4776 energy efficiency, but can expose users to
4777 increased grace-period latency. This parameter
4778 overrides rcupdate.rcu_expedited. No effect on
4779 CONFIG_TINY_RCU kernels.
4781 rcupdate.rcu_normal_after_boot= [KNL]
4782 Once boot has completed (that is, after
4783 rcu_end_inkernel_boot() has been invoked), use
4784 only normal grace-period primitives. No effect
4785 on CONFIG_TINY_RCU kernels.
4787 But note that CONFIG_PREEMPT_RT=y kernels enables
4788 this kernel boot parameter, forcibly setting
4789 it to the value one, that is, converting any
4790 post-boot attempt at an expedited RCU grace
4791 period to instead use normal non-expedited
4792 grace-period processing.
4794 rcupdate.rcu_task_ipi_delay= [KNL]
4795 Set time in jiffies during which RCU tasks will
4796 avoid sending IPIs, starting with the beginning
4797 of a given grace period. Setting a large
4798 number avoids disturbing real-time workloads,
4799 but lengthens grace periods.
4801 rcupdate.rcu_task_stall_timeout= [KNL]
4802 Set timeout in jiffies for RCU task stall warning
4803 messages. Disable with a value less than or equal
4806 rcupdate.rcu_self_test= [KNL]
4807 Run the RCU early boot self tests
4811 Run specified binary instead of /init from the ramdisk,
4812 used for early userspace startup. See initrd.
4815 force - Override the decision by the kernel to hide the
4816 advertisement of RDRAND support (this affects
4817 certain AMD processors because of buggy BIOS
4818 support, specifically around the suspend/resume
4822 Turn on/off individual RDT features. List is:
4823 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4825 E.g. to turn on cmt and turn off mba use:
4829 Format (x86 or x86_64):
4830 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
4832 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4834 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4835 (prefix with 'panic_' to set mode for panic
4837 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4838 reboot_force is either force or not specified,
4839 reboot_cpu is s[mp]#### with #### being the processor
4840 to be used for rebooting.
4842 refscale.holdoff= [KNL]
4843 Set test-start holdoff period. The purpose of
4844 this parameter is to delay the start of the
4845 test until boot completes in order to avoid
4848 refscale.loops= [KNL]
4849 Set the number of loops over the synchronization
4850 primitive under test. Increasing this number
4851 reduces noise due to loop start/end overhead,
4852 but the default has already reduced the per-pass
4853 noise to a handful of picoseconds on ca. 2020
4856 refscale.nreaders= [KNL]
4857 Set number of readers. The default value of -1
4858 selects N, where N is roughly 75% of the number
4859 of CPUs. A value of zero is an interesting choice.
4861 refscale.nruns= [KNL]
4862 Set number of runs, each of which is dumped onto
4865 refscale.readdelay= [KNL]
4866 Set the read-side critical-section duration,
4867 measured in microseconds.
4869 refscale.scale_type= [KNL]
4870 Specify the read-protection implementation to test.
4872 refscale.shutdown= [KNL]
4873 Shut down the system at the end of the performance
4874 test. This defaults to 1 (shut it down) when
4875 refscale is built into the kernel and to 0 (leave
4876 it running) when refscale is built as a module.
4878 refscale.verbose= [KNL]
4879 Enable additional printk() statements.
4881 refscale.verbose_batched= [KNL]
4882 Batch the additional printk() statements. If zero
4883 (the default) or negative, print everything. Otherwise,
4884 print every Nth verbose statement, where N is the value
4888 [KNL, SMP] Set scheduler's default relax_domain_level.
4889 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4891 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4892 Format: <base1>,<size1>[,<base2>,<size2>,...]
4893 Reserve I/O ports or memory so the kernel won't use
4894 them. If <base> is less than 0x10000, the region
4895 is assumed to be I/O ports; otherwise it is memory.
4897 reservetop= [X86-32]
4899 Reserves a hole at the top of the kernel virtual
4902 reset_devices [KNL] Force drivers to reset the underlying device
4903 during initialization.
4906 Specify the partition device for software suspend
4908 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4910 resume_offset= [SWSUSP]
4911 Specify the offset from the beginning of the partition
4912 given by "resume=" at which the swap header is located,
4913 in <PAGE_SIZE> units (needed only for swap files).
4914 See Documentation/power/swsusp-and-swap-files.rst
4916 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4917 read the resume files
4919 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4920 Useful for devices that are detected asynchronously
4921 (e.g. USB and MMC devices).
4923 hibernate= [HIBERNATION]
4924 noresume Don't check if there's a hibernation image
4925 present during boot.
4926 nocompress Don't compress/decompress hibernation images.
4927 no Disable hibernation and resume.
4928 protect_image Turn on image protection during restoration
4929 (that will set all pages holding image data
4930 during restoration read-only).
4932 retain_initrd [RAM] Keep initrd memory after extraction
4934 rfkill.default_state=
4935 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4936 etc. communication is blocked by default.
4939 rfkill.master_switch_mode=
4940 0 The "airplane mode" button does nothing.
4941 1 The "airplane mode" button toggles between everything
4942 blocked and the previous configuration.
4943 2 The "airplane mode" button toggles between everything
4944 blocked and everything unblocked.
4946 rhash_entries= [KNL,NET]
4947 Set number of hash buckets for route cache
4950 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4953 ro [KNL] Mount root device read-only on boot
4956 on Mark read-only kernel memory as read-only (default).
4957 off Leave read-only kernel memory writable for debugging.
4960 Enable the uart passthrough on the designated usb port
4961 on Rockchip SoCs. When active, the signals of the
4962 debug-uart get routed to the D+ and D- pins of the usb
4963 port and the regular usb controller gets disabled.
4965 root= [KNL] Root filesystem
4966 See name_to_dev_t comment in init/do_mounts.c.
4968 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4969 mount the root filesystem
4971 rootflags= [KNL] Set root filesystem mount option string
4973 rootfstype= [KNL] Set root filesystem type
4975 rootwait [KNL] Wait (indefinitely) for root device to show up.
4976 Useful for devices that are detected asynchronously
4977 (e.g. USB and MMC devices).
4979 rproc_mem=nn[KMG][@address]
4980 [KNL,ARM,CMA] Remoteproc physical memory block.
4981 Memory area to be used by remote processor image,
4984 rw [KNL] Mount root device read-write on boot
4986 S [KNL] Run init in single mode
4988 s390_iommu= [HW,S390]
4989 Set s390 IOTLB flushing mode
4991 With strict flushing every unmap operation will result in
4992 an IOTLB flush. Default is lazy flushing before reuse,
4995 s390_iommu_aperture= [KNL,S390]
4996 Specifies the size of the per device DMA address space
4997 accessible through the DMA and IOMMU APIs as a decimal
4998 factor of the size of main memory.
4999 The default is 1 meaning that one can concurrently use
5000 as many DMA addresses as physical memory is installed,
5001 if supported by hardware, and thus map all of memory
5002 once. With a value of 2 one can map all of memory twice
5003 and so on. As a special case a factor of 0 imposes no
5004 restrictions other than those given by hardware at the
5005 cost of significant additional memory use for tables.
5008 See drivers/net/irda/sa1100_ir.c.
5010 sched_verbose [KNL] Enables verbose scheduler debug messages.
5012 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5013 Allowed values are enable and disable. This feature
5014 incurs a small amount of overhead in the scheduler
5015 but is useful for debugging and performance tuning.
5017 sched_thermal_decay_shift=
5018 [KNL, SMP] Set a decay shift for scheduler thermal
5019 pressure signal. Thermal pressure signal follows the
5020 default decay period of other scheduler pelt
5021 signals(usually 32 ms but configurable). Setting
5022 sched_thermal_decay_shift will left shift the decay
5023 period for the thermal pressure signal by the shift
5025 i.e. with the default pelt decay period of 32 ms
5026 sched_thermal_decay_shift thermal pressure decay pr
5030 Format: integer between 0 and 10
5033 scftorture.holdoff= [KNL]
5034 Number of seconds to hold off before starting
5035 test. Defaults to zero for module insertion and
5036 to 10 seconds for built-in smp_call_function()
5039 scftorture.longwait= [KNL]
5040 Request ridiculously long waits randomly selected
5041 up to the chosen limit in seconds. Zero (the
5042 default) disables this feature. Please note
5043 that requesting even small non-zero numbers of
5044 seconds can result in RCU CPU stall warnings,
5045 softlockup complaints, and so on.
5047 scftorture.nthreads= [KNL]
5048 Number of kthreads to spawn to invoke the
5049 smp_call_function() family of functions.
5050 The default of -1 specifies a number of kthreads
5051 equal to the number of CPUs.
5053 scftorture.onoff_holdoff= [KNL]
5054 Number seconds to wait after the start of the
5055 test before initiating CPU-hotplug operations.
5057 scftorture.onoff_interval= [KNL]
5058 Number seconds to wait between successive
5059 CPU-hotplug operations. Specifying zero (which
5060 is the default) disables CPU-hotplug operations.
5062 scftorture.shutdown_secs= [KNL]
5063 The number of seconds following the start of the
5064 test after which to shut down the system. The
5065 default of zero avoids shutting down the system.
5066 Non-zero values are useful for automated tests.
5068 scftorture.stat_interval= [KNL]
5069 The number of seconds between outputting the
5070 current test statistics to the console. A value
5071 of zero disables statistics output.
5073 scftorture.stutter_cpus= [KNL]
5074 The number of jiffies to wait between each change
5075 to the set of CPUs under test.
5077 scftorture.use_cpus_read_lock= [KNL]
5078 Use use_cpus_read_lock() instead of the default
5079 preempt_disable() to disable CPU hotplug
5080 while invoking one of the smp_call_function*()
5083 scftorture.verbose= [KNL]
5084 Enable additional printk() statements.
5086 scftorture.weight_single= [KNL]
5087 The probability weighting to use for the
5088 smp_call_function_single() function with a zero
5089 "wait" parameter. A value of -1 selects the
5090 default if all other weights are -1. However,
5091 if at least one weight has some other value, a
5092 value of -1 will instead select a weight of zero.
5094 scftorture.weight_single_wait= [KNL]
5095 The probability weighting to use for the
5096 smp_call_function_single() function with a
5097 non-zero "wait" parameter. See weight_single.
5099 scftorture.weight_many= [KNL]
5100 The probability weighting to use for the
5101 smp_call_function_many() function with a zero
5102 "wait" parameter. See weight_single.
5103 Note well that setting a high probability for
5104 this weighting can place serious IPI load
5107 scftorture.weight_many_wait= [KNL]
5108 The probability weighting to use for the
5109 smp_call_function_many() function with a
5110 non-zero "wait" parameter. See weight_single
5113 scftorture.weight_all= [KNL]
5114 The probability weighting to use for the
5115 smp_call_function_all() function with a zero
5116 "wait" parameter. See weight_single and
5119 scftorture.weight_all_wait= [KNL]
5120 The probability weighting to use for the
5121 smp_call_function_all() function with a
5122 non-zero "wait" parameter. See weight_single
5125 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5126 xtime_lock contention on larger systems, and/or RCU lock
5127 contention on all systems with CONFIG_MAXSMP set.
5128 Format: { "0" | "1" }
5129 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5131 Note: increases power consumption, thus should only be
5132 enabled if running jitter sensitive (HPC/RT) workloads.
5134 security= [SECURITY] Choose a legacy "major" security module to
5135 enable at boot. This has been deprecated by the
5138 selinux= [SELINUX] Disable or enable SELinux at boot time.
5139 Format: { "0" | "1" }
5140 See security/selinux/Kconfig help text.
5145 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5146 Format: { "0" | "1" }
5147 See security/apparmor/Kconfig help text
5150 Default value is set via kernel config option.
5152 serialnumber [BUGS=X86-32]
5155 Maximal number of shapers.
5163 Enable merging of slabs with similar size when the
5164 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5167 Disable merging of slabs with similar size. May be
5168 necessary if there is some reason to distinguish
5169 allocs to different slabs, especially in hardened
5170 environments where the risk of heap overflows and
5171 layout control by attackers can usually be
5172 frustrated by disabling merging. This will reduce
5173 most of the exposure of a heap attack to a single
5174 cache (risks via metadata attacks are mostly
5175 unchanged). Debug options disable merging on their
5177 For more information see Documentation/vm/slub.rst.
5179 slab_max_order= [MM, SLAB]
5180 Determines the maximum allowed order for slabs.
5181 A high setting may cause OOMs due to memory
5182 fragmentation. Defaults to 1 for systems with
5183 more than 32MB of RAM, 0 otherwise.
5185 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5186 Enabling slub_debug allows one to determine the
5187 culprit if slab objects become corrupted. Enabling
5188 slub_debug can create guard zones around objects and
5189 may poison objects when not in use. Also tracks the
5190 last alloc / free. For more information see
5191 Documentation/vm/slub.rst.
5193 slub_max_order= [MM, SLUB]
5194 Determines the maximum allowed order for slabs.
5195 A high setting may cause OOMs due to memory
5196 fragmentation. For more information see
5197 Documentation/vm/slub.rst.
5199 slub_min_objects= [MM, SLUB]
5200 The minimum number of objects per slab. SLUB will
5201 increase the slab order up to slub_max_order to
5202 generate a sufficiently large slab able to contain
5203 the number of objects indicated. The higher the number
5204 of objects the smaller the overhead of tracking slabs
5205 and the less frequently locks need to be acquired.
5206 For more information see Documentation/vm/slub.rst.
5208 slub_min_order= [MM, SLUB]
5209 Determines the minimum page order for slabs. Must be
5210 lower than slub_max_order.
5211 For more information see Documentation/vm/slub.rst.
5213 slub_merge [MM, SLUB]
5214 Same with slab_merge.
5216 slub_nomerge [MM, SLUB]
5217 Same with slab_nomerge. This is supported for legacy.
5218 See slab_nomerge for more information.
5221 Format: <io1>[,<io2>[,...,<io8>]]
5223 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5224 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5225 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5226 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5227 smsc-ircc2.ircc_irq= [HW] IRQ line
5228 smsc-ircc2.ircc_dma= [HW] DMA channel
5229 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5230 0: Toshiba Satellite 1800 (GP data pin select)
5231 1: Fast pin select (default)
5234 smt [KNL,S390] Set the maximum number of threads (logical
5235 CPUs) to use per physical CPU on systems capable of
5236 symmetric multithreading (SMT). Will be capped to the
5237 actual hardware limit.
5239 Default: -1 (no limit)
5242 [KNL] Should the soft-lockup detector generate panics.
5245 A value of 1 instructs the soft-lockup detector
5246 to panic the machine when a soft-lockup occurs. It is
5247 also controlled by the kernel.softlockup_panic sysctl
5248 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5249 respective build-time switch to that functionality.
5251 softlockup_all_cpu_backtrace=
5252 [KNL] Should the soft-lockup detector generate
5253 backtraces on all cpus.
5256 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5257 See Documentation/admin-guide/laptops/sonypi.rst
5259 spectre_v2= [X86] Control mitigation of Spectre variant 2
5260 (indirect branch speculation) vulnerability.
5261 The default operation protects the kernel from
5264 on - unconditionally enable, implies
5266 off - unconditionally disable, implies
5268 auto - kernel detects whether your CPU model is
5271 Selecting 'on' will, and 'auto' may, choose a
5272 mitigation method at run time according to the
5273 CPU, the available microcode, the setting of the
5274 CONFIG_RETPOLINE configuration option, and the
5275 compiler with which the kernel was built.
5277 Selecting 'on' will also enable the mitigation
5278 against user space to user space task attacks.
5280 Selecting 'off' will disable both the kernel and
5281 the user space protections.
5283 Specific mitigations can also be selected manually:
5285 retpoline - replace indirect branches
5286 retpoline,generic - google's original retpoline
5287 retpoline,amd - AMD-specific minimal thunk
5289 Not specifying this option is equivalent to
5293 [X86] Control mitigation of Spectre variant 2
5294 (indirect branch speculation) vulnerability between
5297 on - Unconditionally enable mitigations. Is
5298 enforced by spectre_v2=on
5300 off - Unconditionally disable mitigations. Is
5301 enforced by spectre_v2=off
5303 prctl - Indirect branch speculation is enabled,
5304 but mitigation can be enabled via prctl
5305 per thread. The mitigation control state
5306 is inherited on fork.
5309 - Like "prctl" above, but only STIBP is
5310 controlled per thread. IBPB is issued
5311 always when switching between different user
5315 - Same as "prctl" above, but all seccomp
5316 threads will enable the mitigation unless
5317 they explicitly opt out.
5320 - Like "seccomp" above, but only STIBP is
5321 controlled per thread. IBPB is issued
5322 always when switching between different
5323 user space processes.
5325 auto - Kernel selects the mitigation depending on
5326 the available CPU features and vulnerability.
5328 Default mitigation: "prctl"
5330 Not specifying this option is equivalent to
5331 spectre_v2_user=auto.
5333 spec_store_bypass_disable=
5334 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5335 (Speculative Store Bypass vulnerability)
5337 Certain CPUs are vulnerable to an exploit against a
5338 a common industry wide performance optimization known
5339 as "Speculative Store Bypass" in which recent stores
5340 to the same memory location may not be observed by
5341 later loads during speculative execution. The idea
5342 is that such stores are unlikely and that they can
5343 be detected prior to instruction retirement at the
5344 end of a particular speculation execution window.
5346 In vulnerable processors, the speculatively forwarded
5347 store can be used in a cache side channel attack, for
5348 example to read memory to which the attacker does not
5349 directly have access (e.g. inside sandboxed code).
5351 This parameter controls whether the Speculative Store
5352 Bypass optimization is used.
5354 On x86 the options are:
5356 on - Unconditionally disable Speculative Store Bypass
5357 off - Unconditionally enable Speculative Store Bypass
5358 auto - Kernel detects whether the CPU model contains an
5359 implementation of Speculative Store Bypass and
5360 picks the most appropriate mitigation. If the
5361 CPU is not vulnerable, "off" is selected. If the
5362 CPU is vulnerable the default mitigation is
5363 architecture and Kconfig dependent. See below.
5364 prctl - Control Speculative Store Bypass per thread
5365 via prctl. Speculative Store Bypass is enabled
5366 for a process by default. The state of the control
5367 is inherited on fork.
5368 seccomp - Same as "prctl" above, but all seccomp threads
5369 will disable SSB unless they explicitly opt out.
5371 Default mitigations:
5374 On powerpc the options are:
5376 on,auto - On Power8 and Power9 insert a store-forwarding
5377 barrier on kernel entry and exit. On Power7
5378 perform a software flush on kernel entry and
5382 Not specifying this option is equivalent to
5383 spec_store_bypass_disable=auto.
5385 spia_io_base= [HW,MTD]
5391 [X86] Enable split lock detection or bus lock detection
5393 When enabled (and if hardware support is present), atomic
5394 instructions that access data across cache line
5395 boundaries will result in an alignment check exception
5396 for split lock detection or a debug exception for
5401 warn - the kernel will emit rate-limited warnings
5402 about applications triggering the #AC
5403 exception or the #DB exception. This mode is
5404 the default on CPUs that support split lock
5405 detection or bus lock detection. Default
5406 behavior is by #AC if both features are
5407 enabled in hardware.
5409 fatal - the kernel will send SIGBUS to applications
5410 that trigger the #AC exception or the #DB
5411 exception. Default behavior is by #AC if
5412 both features are enabled in hardware.
5415 Set system wide rate limit to N bus locks
5416 per second for bus lock detection.
5419 N/A for split lock detection.
5422 If an #AC exception is hit in the kernel or in
5423 firmware (i.e. not while executing in user mode)
5424 the kernel will oops in either "warn" or "fatal"
5427 #DB exception for bus lock is triggered only when
5431 Control the Special Register Buffer Data Sampling
5434 Certain CPUs are vulnerable to an MDS-like
5435 exploit which can leak bits from the random
5438 By default, this issue is mitigated by
5439 microcode. However, the microcode fix can cause
5440 the RDRAND and RDSEED instructions to become
5441 much slower. Among other effects, this will
5442 result in reduced throughput from /dev/urandom.
5444 The microcode mitigation can be disabled with
5445 the following option:
5447 off: Disable mitigation and remove
5448 performance impact to RDRAND and RDSEED
5450 srcutree.counter_wrap_check [KNL]
5451 Specifies how frequently to check for
5452 grace-period sequence counter wrap for the
5453 srcu_data structure's ->srcu_gp_seq_needed field.
5454 The greater the number of bits set in this kernel
5455 parameter, the less frequently counter wrap will
5456 be checked for. Note that the bottom two bits
5459 srcutree.exp_holdoff [KNL]
5460 Specifies how many nanoseconds must elapse
5461 since the end of the last SRCU grace period for
5462 a given srcu_struct until the next normal SRCU
5463 grace period will be considered for automatic
5464 expediting. Set to zero to disable automatic
5468 Speculative Store Bypass Disable control
5470 On CPUs that are vulnerable to the Speculative
5471 Store Bypass vulnerability and offer a
5472 firmware based mitigation, this parameter
5473 indicates how the mitigation should be used:
5475 force-on: Unconditionally enable mitigation for
5476 for both kernel and userspace
5477 force-off: Unconditionally disable mitigation for
5478 for both kernel and userspace
5479 kernel: Always enable mitigation in the
5480 kernel, and offer a prctl interface
5481 to allow userspace to register its
5482 interest in being mitigated too.
5484 stack_guard_gap= [MM]
5485 override the default stack gap protection. The value
5486 is in page units and it defines how many pages prior
5487 to (for stacks growing down) resp. after (for stacks
5488 growing up) the main stack are reserved for no other
5489 mapping. Default value is 256 pages.
5491 stack_depot_disable= [KNL]
5492 Setting this to true through kernel command line will
5493 disable the stack depot thereby saving the static memory
5494 consumed by the stack hash table. By default this is set
5498 Enabled the stack tracer on boot up.
5500 stacktrace_filter=[function-list]
5501 [FTRACE] Limit the functions that the stack tracer
5502 will trace at boot up. function-list is a comma-separated
5503 list of functions. This list can be changed at run
5504 time by the stack_trace_filter file in the debugfs
5505 tracing directory. Note, this enables stack tracing
5506 and the stacktrace above is not needed.
5510 Set the STI (builtin display/keyboard on the HP-PARISC
5511 machines) console (graphic card) which should be used
5512 as the initial boot-console.
5513 See also comment in drivers/video/console/sticore.c.
5516 See comment in drivers/video/console/sticore.c.
5519 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5524 Enable or disable strict sigaltstack size checks
5525 against the required signal frame size which
5526 depends on the supported FPU features. This can
5527 be used to filter out binaries which have
5528 not yet been made aware of AT_MINSIGSTKSZ.
5530 sunrpc.min_resvport=
5531 sunrpc.max_resvport=
5533 SunRPC servers often require that client requests
5534 originate from a privileged port (i.e. a port in the
5535 range 0 < portnr < 1024).
5536 An administrator who wishes to reserve some of these
5537 ports for other uses may adjust the range that the
5538 kernel's sunrpc client considers to be privileged
5539 using these two parameters to set the minimum and
5540 maximum port values.
5542 sunrpc.svc_rpc_per_connection_limit=
5544 Limit the number of requests that the server will
5545 process in parallel from a single connection.
5546 The default value is 0 (no limit).
5550 Control how the NFS server code allocates CPUs to
5551 service thread pools. Depending on how many NICs
5552 you have and where their interrupts are bound, this
5553 option will affect which CPUs will do NFS serving.
5554 Note: this parameter cannot be changed while the
5555 NFS server is running.
5557 auto the server chooses an appropriate mode
5558 automatically using heuristics
5559 global a single global pool contains all CPUs
5560 percpu one pool for each CPU
5561 pernode one pool for each NUMA node (equivalent
5562 to global on non-NUMA machines)
5564 sunrpc.tcp_slot_table_entries=
5565 sunrpc.udp_slot_table_entries=
5567 Sets the upper limit on the number of simultaneous
5568 RPC calls that can be sent from the client to a
5569 server. Increasing these values may allow you to
5570 improve throughput, but will also increase the
5571 amount of memory reserved for use by the client.
5573 suspend.pm_test_delay=
5575 Sets the number of seconds to remain in a suspend test
5576 mode before resuming the system (see
5577 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5578 is set. Default value is 5.
5581 Format: { on | off | y | n | 1 | 0 }
5582 This parameter controls use of the Protected
5583 Execution Facility on pSeries.
5586 [KNL] Enable accounting of swap in memory resource
5587 controller if no parameter or 1 is given or disable
5588 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5590 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5591 Format: { <int> | force | noforce }
5592 <int> -- Number of I/O TLB slabs
5593 force -- force using of bounce buffers even if they
5594 wouldn't be automatically used by the kernel
5595 noforce -- Never use bounce buffers (for debugging)
5600 Set a sysctl parameter, right before loading the init
5601 process, as if the value was written to the respective
5602 /proc/sys/... file. Both '.' and '/' are recognized as
5603 separators. Unrecognized parameters and invalid values
5604 are reported in the kernel log. Sysctls registered
5605 later by a loaded module cannot be set this way.
5606 Example: sysctl.vm.swappiness=40
5608 sysfs.deprecated=0|1 [KNL]
5609 Enable/disable old style sysfs layout for old udev
5610 on older distributions. When this option is enabled
5611 very new udev will not work anymore. When this option
5612 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5613 in older udev will not work anymore.
5614 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5615 the kernel configuration.
5617 sysrq_always_enabled
5619 Ignore sysrq setting - this boot parameter will
5620 neutralize any effect of /proc/sys/kernel/sysrq.
5621 Useful for debugging.
5623 tcpmhash_entries= [KNL,NET]
5624 Set the number of tcp_metrics_hash slots.
5625 Default value is 8192 or 16384 depending on total
5626 ram pages. This is used to specify the TCP metrics
5627 cache size. See Documentation/networking/ip-sysctl.rst
5628 "tcp_no_metrics_save" section for more details.
5632 test_suspend= [SUSPEND][,N]
5633 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5634 standby suspend) or "freeze" (for suspend type freeze)
5635 as the system sleep state during system startup with
5636 the optional capability to repeat N number of times.
5637 The system is woken from this state using a
5638 wakeup-capable RTC alarm.
5640 thash_entries= [KNL,NET]
5641 Set number of hash buckets for TCP connection
5643 thermal.act= [HW,ACPI]
5644 -1: disable all active trip points in all thermal zones
5645 <degrees C>: override all lowest active trip points
5647 thermal.crt= [HW,ACPI]
5648 -1: disable all critical trip points in all thermal zones
5649 <degrees C>: override all critical trip points
5651 thermal.nocrt= [HW,ACPI]
5652 Set to disable actions on ACPI thermal zone
5653 critical and hot trip points.
5655 thermal.off= [HW,ACPI]
5656 1: disable ACPI thermal control
5658 thermal.psv= [HW,ACPI]
5659 -1: disable all passive trip points
5660 <degrees C>: override all passive trip points to this
5663 thermal.tzp= [HW,ACPI]
5664 Specify global default ACPI thermal zone polling rate
5665 <deci-seconds>: poll all this frequency
5666 0: no polling (default)
5669 Force threading of all interrupt handlers except those
5670 marked explicitly IRQF_NO_THREAD.
5674 Specify if the kernel should make use of the cpu
5675 topology information if the hardware supports this.
5676 The scheduler will make use of this information and
5677 e.g. base its process migration decisions on it.
5680 topology_updates= [KNL, PPC, NUMA]
5682 Specify if the kernel should ignore (off)
5683 topology updates sent by the hypervisor to this
5686 torture.disable_onoff_at_boot= [KNL]
5687 Prevent the CPU-hotplug component of torturing
5688 until after init has spawned.
5690 torture.ftrace_dump_at_shutdown= [KNL]
5691 Dump the ftrace buffer at torture-test shutdown,
5692 even if there were no errors. This can be a
5693 very costly operation when many torture tests
5694 are running concurrently, especially on systems
5695 with rotating-rust storage.
5697 torture.verbose_sleep_frequency= [KNL]
5698 Specifies how many verbose printk()s should be
5699 emitted between each sleep. The default of zero
5700 disables verbose-printk() sleeping.
5702 torture.verbose_sleep_duration= [KNL]
5703 Duration of each verbose-printk() sleep in jiffies.
5707 tpm_suspend_pcr=[HW,TPM]
5708 Format: integer pcr id
5709 Specify that at suspend time, the tpm driver
5710 should extend the specified pcr with zeros,
5711 as a workaround for some chips which fail to
5712 flush the last written pcr on TPM_SaveState.
5713 This will guarantee that all the other pcrs
5716 trace_buf_size=nn[KMG]
5717 [FTRACE] will set tracing buffer size on each cpu.
5719 trace_event=[event-list]
5720 [FTRACE] Set and start specified trace events in order
5721 to facilitate early boot debugging. The event-list is a
5722 comma-separated list of trace events to enable. See
5723 also Documentation/trace/events.rst
5725 trace_options=[option-list]
5726 [FTRACE] Enable or disable tracer options at boot.
5727 The option-list is a comma delimited list of options
5728 that can be enabled or disabled just as if you were
5729 to echo the option name into
5731 /sys/kernel/debug/tracing/trace_options
5733 For example, to enable stacktrace option (to dump the
5734 stack trace of each event), add to the command line:
5736 trace_options=stacktrace
5738 See also Documentation/trace/ftrace.rst "trace options"
5742 Have the tracepoints sent to printk as well as the
5743 tracing ring buffer. This is useful for early boot up
5744 where the system hangs or reboots and does not give the
5745 option for reading the tracing buffer or performing a
5746 ftrace_dump_on_oops.
5748 To turn off having tracepoints sent to printk,
5749 echo 0 > /proc/sys/kernel/tracepoint_printk
5750 Note, echoing 1 into this file without the
5751 tracepoint_printk kernel cmdline option has no effect.
5753 The tp_printk_stop_on_boot (see below) can also be used
5754 to stop the printing of events to console at
5759 Having tracepoints sent to printk() and activating high
5760 frequency tracepoints such as irq or sched, can cause
5761 the system to live lock.
5763 tp_printk_stop_on_boot[FTRACE]
5764 When tp_printk (above) is set, it can cause a lot of noise
5765 on the console. It may be useful to only include the
5766 printing of events during boot up, as user space may
5767 make the system inoperable.
5769 This command line option will stop the printing of events
5770 to console at the late_initcall_sync() time frame.
5773 [FTRACE] enable this option to disable tracing when a
5774 warning is hit. This turns off "tracing_on". Tracing can
5775 be enabled again by echoing '1' into the "tracing_on"
5776 file located in /sys/kernel/debug/tracing/
5778 This option is useful, as it disables the trace before
5779 the WARNING dump is called, which prevents the trace to
5780 be filled with content caused by the warning output.
5782 This option can also be set at run time via the sysctl
5783 option: kernel/traceoff_on_warning
5785 transparent_hugepage=
5787 Format: [always|madvise|never]
5788 Can be used to control the default behavior of the system
5789 with respect to transparent hugepages.
5790 See Documentation/admin-guide/mm/transhuge.rst
5793 trusted.source= [KEYS]
5795 This parameter identifies the trust source as a backend
5796 for trusted keys implementation. Supported trust
5800 If not specified then it defaults to iterating through
5801 the trust source list starting with TPM and assigns the
5802 first trust source as a backend which is initialized
5803 successfully during iteration.
5805 tsc= Disable clocksource stability checks for TSC.
5807 [x86] reliable: mark tsc clocksource as reliable, this
5808 disables clocksource verification at runtime, as well
5809 as the stability checks done at bootup. Used to enable
5810 high-resolution timer mode on older hardware, and in
5811 virtualized environment.
5812 [x86] noirqtime: Do not use TSC to do irq accounting.
5813 Used to run time disable IRQ_TIME_ACCOUNTING on any
5814 platforms where RDTSC is slow and this accounting
5816 [x86] unstable: mark the TSC clocksource as unstable, this
5817 marks the TSC unconditionally unstable at bootup and
5818 avoids any further wobbles once the TSC watchdog notices.
5819 [x86] nowatchdog: disable clocksource watchdog. Used
5820 in situations with strict latency requirements (where
5821 interruptions from clocksource watchdog are not
5824 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5825 value instead. Useful when the early TSC frequency discovery
5826 procedure is not reliable, such as on overclocked systems
5827 with CPUID.16h support and partial CPUID.15h support.
5828 Format: <unsigned int>
5830 tsx= [X86] Control Transactional Synchronization
5831 Extensions (TSX) feature in Intel processors that
5832 support TSX control.
5834 This parameter controls the TSX feature. The options are:
5836 on - Enable TSX on the system. Although there are
5837 mitigations for all known security vulnerabilities,
5838 TSX has been known to be an accelerator for
5839 several previous speculation-related CVEs, and
5840 so there may be unknown security risks associated
5841 with leaving it enabled.
5843 off - Disable TSX on the system. (Note that this
5844 option takes effect only on newer CPUs which are
5845 not vulnerable to MDS, i.e., have
5846 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5847 the new IA32_TSX_CTRL MSR through a microcode
5848 update. This new MSR allows for the reliable
5849 deactivation of the TSX functionality.)
5851 auto - Disable TSX if X86_BUG_TAA is present,
5852 otherwise enable TSX on the system.
5854 Not specifying this option is equivalent to tsx=off.
5856 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5859 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5860 Abort (TAA) vulnerability.
5862 Similar to Micro-architectural Data Sampling (MDS)
5863 certain CPUs that support Transactional
5864 Synchronization Extensions (TSX) are vulnerable to an
5865 exploit against CPU internal buffers which can forward
5866 information to a disclosure gadget under certain
5869 In vulnerable processors, the speculatively forwarded
5870 data can be used in a cache side channel attack, to
5871 access data to which the attacker does not have direct
5874 This parameter controls the TAA mitigation. The
5877 full - Enable TAA mitigation on vulnerable CPUs
5880 full,nosmt - Enable TAA mitigation and disable SMT on
5881 vulnerable CPUs. If TSX is disabled, SMT
5882 is not disabled because CPU is not
5883 vulnerable to cross-thread TAA attacks.
5884 off - Unconditionally disable TAA mitigation
5886 On MDS-affected machines, tsx_async_abort=off can be
5887 prevented by an active MDS mitigation as both vulnerabilities
5888 are mitigated with the same mechanism so in order to disable
5889 this mitigation, you need to specify mds=off too.
5891 Not specifying this option is equivalent to
5892 tsx_async_abort=full. On CPUs which are MDS affected
5893 and deploy MDS mitigation, TAA mitigation is not
5894 required and doesn't provide any additional
5898 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5900 turbografx.map[2|3]= [HW,JOY]
5901 TurboGraFX parallel port interface
5903 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5904 See also Documentation/input/devices/joystick-parport.rst
5906 udbg-immortal [PPC] When debugging early kernel crashes that
5907 happen after console_init() and before a proper
5908 console driver takes over, this boot options might
5909 help "seeing" what's going on.
5911 uhash_entries= [KNL,NET]
5912 Set number of hash buckets for UDP/UDP-Lite connections
5915 [USB] Ignore overcurrent events (default N).
5916 Some badly-designed motherboards generate lots of
5917 bogus events, for ports that aren't wired to
5918 anything. Set this parameter to avoid log spamming.
5919 Note that genuine overcurrent events won't be
5923 [X86] Cause panic on unknown NMI.
5925 usbcore.authorized_default=
5926 [USB] Default USB device authorization:
5927 (default -1 = authorized except for wireless USB,
5928 0 = not authorized, 1 = authorized, 2 = authorized
5929 if device connected to internal port)
5931 usbcore.autosuspend=
5932 [USB] The autosuspend time delay (in seconds) used
5933 for newly-detected USB devices (default 2). This
5934 is the time required before an idle device will be
5935 autosuspended. Devices for which the delay is set
5936 to a negative value won't be autosuspended at all.
5938 usbcore.usbfs_snoop=
5939 [USB] Set to log all usbfs traffic (default 0 = off).
5941 usbcore.usbfs_snoop_max=
5942 [USB] Maximum number of bytes to snoop in each URB
5945 usbcore.blinkenlights=
5946 [USB] Set to cycle leds on hubs (default 0 = off).
5948 usbcore.old_scheme_first=
5949 [USB] Start with the old device initialization
5950 scheme (default 0 = off).
5952 usbcore.usbfs_memory_mb=
5953 [USB] Memory limit (in MB) for buffers allocated by
5954 usbfs (default = 16, 0 = max = 2047).
5956 usbcore.use_both_schemes=
5957 [USB] Try the other device initialization scheme
5958 if the first one fails (default 1 = enabled).
5960 usbcore.initial_descriptor_timeout=
5961 [USB] Specifies timeout for the initial 64-byte
5962 USB_REQ_GET_DESCRIPTOR request in milliseconds
5963 (default 5000 = 5.0 seconds).
5965 usbcore.nousb [USB] Disable the USB subsystem
5968 [USB] A list of quirk entries to augment the built-in
5969 usb core quirk list. List entries are separated by
5970 commas. Each entry has the form
5971 VendorID:ProductID:Flags. The IDs are 4-digit hex
5972 numbers and Flags is a set of letters. Each letter
5973 will change the built-in quirk; setting it if it is
5974 clear and clearing it if it is set. The letters have
5975 the following meanings:
5976 a = USB_QUIRK_STRING_FETCH_255 (string
5977 descriptors must not be fetched using
5979 b = USB_QUIRK_RESET_RESUME (device can't resume
5980 correctly so reset it instead);
5981 c = USB_QUIRK_NO_SET_INTF (device can't handle
5982 Set-Interface requests);
5983 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5984 handle its Configuration or Interface
5986 e = USB_QUIRK_RESET (device can't be reset
5987 (e.g morph devices), don't use reset);
5988 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5989 more interface descriptions than the
5990 bNumInterfaces count, and can't handle
5991 talking to these interfaces);
5992 g = USB_QUIRK_DELAY_INIT (device needs a pause
5993 during initialization, after we read
5994 the device descriptor);
5995 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5996 high speed and super speed interrupt
5997 endpoints, the USB 2.0 and USB 3.0 spec
5998 require the interval in microframes (1
5999 microframe = 125 microseconds) to be
6000 calculated as interval = 2 ^
6002 Devices with this quirk report their
6003 bInterval as the result of this
6004 calculation instead of the exponent
6005 variable used in the calculation);
6006 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6007 handle device_qualifier descriptor
6009 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6010 generates spurious wakeup, ignore
6011 remote wakeup capability);
6012 k = USB_QUIRK_NO_LPM (device can't handle Link
6014 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6015 (Device reports its bInterval as linear
6016 frames instead of the USB 2.0
6018 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6019 to be disconnected before suspend to
6020 prevent spurious wakeup);
6021 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6022 pause after every control message);
6023 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6024 delay after resetting its port);
6025 Example: quirks=0781:5580:bk,0a5c:5834:gij
6028 [USBHID] The interval which mice are to be polled at.
6031 [USBHID] The interval which joysticks are to be polled at.
6034 [USBHID] The interval which keyboards are to be polled at.
6036 usb-storage.delay_use=
6037 [UMS] The delay in seconds before a new device is
6038 scanned for Logical Units (default 1).
6041 [UMS] A list of quirks entries to supplement or
6042 override the built-in unusual_devs list. List
6043 entries are separated by commas. Each entry has
6044 the form VID:PID:Flags where VID and PID are Vendor
6045 and Product ID values (4-digit hex numbers) and
6046 Flags is a set of characters, each corresponding
6047 to a common usb-storage quirk flag as follows:
6048 a = SANE_SENSE (collect more than 18 bytes
6049 of sense data, not on uas);
6050 b = BAD_SENSE (don't collect more than 18
6051 bytes of sense data, not on uas);
6052 c = FIX_CAPACITY (decrease the reported
6053 device capacity by one sector);
6054 d = NO_READ_DISC_INFO (don't use
6055 READ_DISC_INFO command, not on uas);
6056 e = NO_READ_CAPACITY_16 (don't use
6057 READ_CAPACITY_16 command);
6058 f = NO_REPORT_OPCODES (don't use report opcodes
6060 g = MAX_SECTORS_240 (don't transfer more than
6061 240 sectors at a time, uas only);
6062 h = CAPACITY_HEURISTICS (decrease the
6063 reported device capacity by one
6064 sector if the number is odd);
6065 i = IGNORE_DEVICE (don't bind to this
6067 j = NO_REPORT_LUNS (don't use report luns
6069 k = NO_SAME (do not use WRITE_SAME, uas only)
6070 l = NOT_LOCKABLE (don't try to lock and
6071 unlock ejectable media, not on uas);
6072 m = MAX_SECTORS_64 (don't transfer more
6073 than 64 sectors = 32 KB at a time,
6075 n = INITIAL_READ10 (force a retry of the
6076 initial READ(10) command, not on uas);
6077 o = CAPACITY_OK (accept the capacity
6078 reported by the device, not on uas);
6079 p = WRITE_CACHE (the device cache is ON
6080 by default, not on uas);
6081 r = IGNORE_RESIDUE (the device reports
6082 bogus residue values, not on uas);
6083 s = SINGLE_LUN (the device has only one
6085 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6086 commands, uas only);
6087 u = IGNORE_UAS (don't bind to the uas driver);
6088 w = NO_WP_DETECT (don't test whether the
6089 medium is write-protected).
6090 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6091 even if the device claims no cache,
6093 Example: quirks=0419:aaf5:rl,0421:0433:rc
6095 user_debug= [KNL,ARM]
6097 See arch/arm/Kconfig.debug help text.
6098 1 - undefined instruction events
6100 4 - invalid data aborts
6103 Example: user_debug=31
6106 [X86] Flags controlling user PTE allocations.
6108 nohigh = do not allocate PTE pages in
6109 HIGHMEM regardless of setting
6113 On X86_32, this is an alias for vdso32=. Otherwise:
6115 vdso=1: enable VDSO (the default)
6116 vdso=0: disable VDSO mapping
6118 vdso32= [X86] Control the 32-bit vDSO
6119 vdso32=1: enable 32-bit VDSO
6120 vdso32=0 or vdso32=2: disable 32-bit VDSO
6122 See the help text for CONFIG_COMPAT_VDSO for more
6123 details. If CONFIG_COMPAT_VDSO is set, the default is
6124 vdso32=0; otherwise, the default is vdso32=1.
6126 For compatibility with older kernels, vdso32=2 is an
6129 Try vdso32=0 if you encounter an error that says:
6130 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6133 vector=percpu: enable percpu vector domain
6135 video= [FB] Frame buffer configuration
6136 See Documentation/fb/modedb.rst.
6138 video.brightness_switch_enabled= [0,1]
6139 If set to 1, on receiving an ACPI notify event
6140 generated by hotkey, video driver will adjust brightness
6141 level and then send out the event to user space through
6142 the allocated input device; If set to 0, video driver
6143 will only send out the event without touching backlight
6148 [VMMIO] Memory mapped virtio (platform) device.
6150 <size>@<baseaddr>:<irq>[:<id>]
6152 <size> := size (can use standard suffixes
6154 <baseaddr> := physical base address
6155 <irq> := interrupt number (as passed to
6157 <id> := (optional) platform device id
6159 virtio_mmio.device=1K@0x100b0000:48:7
6161 Can be used multiple times for multiple devices.
6163 vga= [BOOT,X86-32] Select a particular video mode
6164 See Documentation/x86/boot.rst and
6165 Documentation/admin-guide/svga.rst.
6166 Use vga=ask for menu.
6167 This is actually a boot loader parameter; the value is
6168 passed to the kernel using a special protocol.
6170 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6171 May slow down system boot speed, especially when
6172 enabled on systems with a large amount of memory.
6173 All options are enabled by default, and this
6174 interface is meant to allow for selectively
6175 enabling or disabling specific virtual memory
6178 Available options are:
6179 P Enable page structure init time poisoning
6180 - Disable all of the above options
6182 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6183 size of <nn>. This can be used to increase the
6184 minimum size (128MB on x86). It can also be used to
6185 decrease the size and leave more room for directly
6188 vmcp_cma=nn[MG] [KNL,S390]
6189 Sets the memory size reserved for contiguous memory
6190 allocations for the vmcp device driver.
6192 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6195 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6198 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6202 Controls the behavior of vsyscalls (i.e. calls to
6203 fixed addresses of 0xffffffffff600x00 from legacy
6204 code). Most statically-linked binaries and older
6205 versions of glibc use these calls. Because these
6206 functions are at fixed addresses, they make nice
6207 targets for exploits that can control RIP.
6209 emulate [default] Vsyscalls turn into traps and are
6210 emulated reasonably safely. The vsyscall
6213 xonly Vsyscalls turn into traps and are
6214 emulated reasonably safely. The vsyscall
6215 page is not readable.
6217 none Vsyscalls don't work at all. This makes
6218 them quite hard to use for exploits but
6219 might break your system.
6221 vt.color= [VT] Default text color.
6222 Format: 0xYX, X = foreground, Y = background.
6223 Default: 0x07 = light gray on black.
6225 vt.cur_default= [VT] Default cursor shape.
6226 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6227 the parameters of the <Esc>[?A;B;Cc escape sequence;
6228 see VGA-softcursor.txt. Default: 2 = underline.
6230 vt.default_blu= [VT]
6231 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6232 Change the default blue palette of the console.
6233 This is a 16-member array composed of values
6236 vt.default_grn= [VT]
6237 Format: <green0>,<green1>,<green2>,...,<green15>
6238 Change the default green palette of the console.
6239 This is a 16-member array composed of values
6242 vt.default_red= [VT]
6243 Format: <red0>,<red1>,<red2>,...,<red15>
6244 Change the default red palette of the console.
6245 This is a 16-member array composed of values
6251 Set system-wide default UTF-8 mode for all tty's.
6252 Default is 1, i.e. UTF-8 mode is enabled for all
6253 newly opened terminals.
6255 vt.global_cursor_default=
6258 Set system-wide default for whether a cursor
6259 is shown on new VTs. Default is -1,
6260 i.e. cursors will be created by default unless
6261 overridden by individual drivers. 0 will hide
6262 cursors, 1 will display them.
6264 vt.italic= [VT] Default color for italic text; 0-15.
6267 vt.underline= [VT] Default color for underlined text; 0-15.
6270 watchdog timers [HW,WDT] For information on watchdog timers,
6271 see Documentation/watchdog/watchdog-parameters.rst
6272 or other driver-specific files in the
6273 Documentation/watchdog/ directory.
6277 Set the hard lockup detector stall duration
6278 threshold in seconds. The soft lockup detector
6279 threshold is set to twice the value. A value of 0
6280 disables both lockup detectors. Default is 10
6283 workqueue.watchdog_thresh=
6284 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6285 warn stall conditions and dump internal state to
6286 help debugging. 0 disables workqueue stall
6287 detection; otherwise, it's the stall threshold
6288 duration in seconds. The default value is 30 and
6289 it can be updated at runtime by writing to the
6290 corresponding sysfs file.
6292 workqueue.disable_numa
6293 By default, all work items queued to unbound
6294 workqueues are affine to the NUMA nodes they're
6295 issued on, which results in better behavior in
6296 general. If NUMA affinity needs to be disabled for
6297 whatever reason, this option can be used. Note
6298 that this also can be controlled per-workqueue for
6299 workqueues visible under /sys/bus/workqueue/.
6301 workqueue.power_efficient
6302 Per-cpu workqueues are generally preferred because
6303 they show better performance thanks to cache
6304 locality; unfortunately, per-cpu workqueues tend to
6305 be more power hungry than unbound workqueues.
6307 Enabling this makes the per-cpu workqueues which
6308 were observed to contribute significantly to power
6309 consumption unbound, leading to measurably lower
6310 power usage at the cost of small performance
6313 The default value of this parameter is determined by
6314 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6316 workqueue.debug_force_rr_cpu
6317 Workqueue used to implicitly guarantee that work
6318 items queued without explicit CPU specified are put
6319 on the local CPU. This guarantee is no longer true
6320 and while local CPU is still preferred work items
6321 may be put on foreign CPUs. This debug option
6322 forces round-robin CPU selection to flush out
6323 usages which depend on the now broken guarantee.
6324 When enabled, memory and cache locality will be
6327 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6328 default x2apic cluster mode on platforms
6331 xen_512gb_limit [KNL,X86-64,XEN]
6332 Restricts the kernel running paravirtualized under Xen
6333 to use only up to 512 GB of RAM. The reason to do so is
6334 crash analysis tools and Xen tools for doing domain
6335 save/restore/migration must be enabled to handle larger
6338 xen_emul_unplug= [HW,X86,XEN]
6339 Unplug Xen emulated devices
6340 Format: [unplug0,][unplug1]
6341 ide-disks -- unplug primary master IDE devices
6342 aux-ide-disks -- unplug non-primary-master IDE devices
6343 nics -- unplug network devices
6344 all -- unplug all emulated devices (NICs and IDE disks)
6345 unnecessary -- unplugging emulated devices is
6346 unnecessary even if the host did not respond to
6348 never -- do not unplug even if version check succeeds
6350 xen_legacy_crash [X86,XEN]
6351 Crash from Xen panic notifier, without executing late
6352 panic() code such as dumping handler.
6354 xen_nopvspin [X86,XEN]
6355 Disables the qspinlock slowpath using Xen PV optimizations.
6356 This parameter is obsoleted by "nopvspin" parameter, which
6357 has equivalent effect for XEN platform.
6360 Disables the PV optimizations forcing the HVM guest to
6361 run as generic HVM guest with no PV drivers.
6362 This option is obsoleted by the "nopv" option, which
6363 has equivalent effect for XEN platform.
6365 xen_no_vector_callback
6366 [KNL,X86,XEN] Disable the vector callback for Xen
6367 event channel interrupts.
6369 xen_scrub_pages= [XEN]
6370 Boolean option to control scrubbing pages before giving them back
6371 to Xen, for use by other domains. Can be also changed at runtime
6372 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6373 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6375 xen_timer_slop= [X86-64,XEN]
6376 Set the timer slop (in nanoseconds) for the virtual Xen
6377 timers (default is 100000). This adjusts the minimum
6378 delta of virtualized Xen timers, where lower values
6379 improve timer resolution at the expense of processing
6380 more timer interrupts.
6382 xen.event_eoi_delay= [XEN]
6383 How long to delay EOI handling in case of event
6384 storms (jiffies). Default is 10.
6386 xen.event_loop_timeout= [XEN]
6387 After which time (jiffies) the event handling loop
6388 should start to delay EOI handling. Default is 2.
6390 xen.fifo_events= [XEN]
6391 Boolean parameter to disable using fifo event handling
6392 even if available. Normally fifo event handling is
6393 preferred over the 2-level event handling, as it is
6394 fairer and the number of possible event channels is
6395 much higher. Default is on (use fifo events).
6397 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6398 Disables the PV optimizations forcing the guest to run
6399 as generic guest with no PV drivers. Currently support
6400 XEN HVM, KVM, HYPER_V and VMWARE guest.
6402 nopvspin [X86,XEN,KVM]
6403 Disables the qspinlock slow path using PV optimizations
6404 which allow the hypervisor to 'idle' the guest on lock
6407 xirc2ps_cs= [NET,PCMCIA]
6409 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6412 By default on POWER9 and above, the kernel will
6413 natively use the XIVE interrupt controller. This option
6414 allows the fallback firmware mode to be used:
6416 off Fallback to firmware control of XIVE interrupt
6417 controller on both pseries and powernv
6418 platforms. Only useful on POWER9 and above.
6420 xhci-hcd.quirks [USB,KNL]
6421 A hex value specifying bitmask with supplemental xhci
6422 host controller quirks. Meaning of each bit can be
6423 consulted in header drivers/usb/host/xhci.h.
6426 Format: { early | on | rw | ro | off }
6427 Controls if xmon debugger is enabled. Default is off.
6428 Passing only "xmon" is equivalent to "xmon=early".
6429 early Call xmon as early as possible on boot; xmon
6430 debugger is called from setup_arch().
6431 on xmon debugger hooks will be installed so xmon
6432 is only called on a kernel crash. Default mode,
6433 i.e. either "ro" or "rw" mode, is controlled
6434 with CONFIG_XMON_DEFAULT_RO_MODE.
6435 rw xmon debugger hooks will be installed so xmon
6436 is called only on a kernel crash, mode is write,
6437 meaning SPR registers, memory and, other data
6438 can be written using xmon commands.
6439 ro same as "rw" option above but SPR registers,
6440 memory, and other data can't be written using
6442 off xmon is disabled.