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.txt, 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]
27 If set to vendor, prefer vendor specific driver
28 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
29 of the ACPI video.ko driver.
31 acpi_force_32bit_fadt_addr
32 force FADT to use 32 bit addresses rather than the
33 64 bit X_* addresses. Some firmware have broken 64
34 bit addresses for force ACPI ignore these and use
35 the older legacy 32 bit addresses.
37 acpica_no_return_repair [HW, ACPI]
38 Disable AML predefined validation mechanism
39 This mechanism can repair the evaluation result to make
40 the return objects more ACPI specification compliant.
41 This option is useful for developers to identify the
42 root cause of an AML interpreter issue when the issue
43 has something to do with the repair mechanism.
45 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
46 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
48 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
49 debug output. Bits in debug_layer correspond to a
50 _COMPONENT in an ACPI source file, e.g.,
51 #define _COMPONENT ACPI_PCI_COMPONENT
52 Bits in debug_level correspond to a level in
53 ACPI_DEBUG_PRINT statements, e.g.,
54 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
55 The debug_level mask defaults to "info". See
56 Documentation/acpi/debug.txt for more information about
57 debug layers and levels.
59 Enable processor driver info messages:
60 acpi.debug_layer=0x20000000
61 Enable PCI/PCI interrupt routing info messages:
62 acpi.debug_layer=0x400000
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
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.txt 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 amd_iommu= [HW,X86-64]
291 Pass parameters to the AMD IOMMU driver in the system.
293 fullflush - enable flushing of IO/TLB entries when
294 they are unmapped. Otherwise they are
295 flushed before they will be reused, which
297 off - do not initialize any AMD IOMMU found in
299 force_isolation - Force device isolation for all
300 devices. The IOMMU driver is not
301 allowed anymore to lift isolation
302 requirements as needed. This option
303 does not override iommu=pt
305 amd_iommu_dump= [HW,X86-64]
306 Enable AMD IOMMU driver option to dump the ACPI table
307 for AMD IOMMU. With this option enabled, AMD IOMMU
308 driver will print ACPI tables for AMD IOMMU during
309 IOMMU initialization.
311 amd_iommu_intr= [HW,X86-64]
312 Specifies one of the following AMD IOMMU interrupt
314 legacy - Use legacy interrupt remapping mode.
315 vapic - Use virtual APIC mode, which allows IOMMU
316 to inject interrupts directly into guest.
317 This mode requires kvm-amd.avic=1.
318 (Default when IOMMU HW support is present.)
320 amijoy.map= [HW,JOY] Amiga joystick support
321 Map of devices attached to JOY0DAT and JOY1DAT
323 See also Documentation/input/joydev/joystick.rst
325 analog.map= [HW,JOY] Analog joystick and gamepad support
326 Specifies type or capabilities of an analog joystick
327 connected to one of 16 gameports
328 Format: <type1>,<type2>,..<type16>
331 Power management functions (SPARCstation-4/5 + deriv.)
333 Disable APC CPU standby support. SPARCstation-Fox does
334 not play well with APC CPU idle - disable it if you have
335 APC and your system crashes randomly.
337 apic= [APIC,X86] Advanced Programmable Interrupt Controller
338 Change the output verbosity whilst booting
339 Format: { quiet (default) | verbose | debug }
340 Change the amount of debugging information output
341 when initialising the APIC and IO-APIC components.
342 For X86-32, this can also be used to specify an APIC
344 Format: apic=driver_name
345 Examples: apic=bigsmp
347 apic_extnmi= [APIC,X86] External NMI delivery setting
348 Format: { bsp (default) | all | none }
349 bsp: External NMI is delivered only to CPU 0
350 all: External NMIs are broadcast to all CPUs as a
352 none: External NMI is masked for all CPUs. This is
353 useful so that a dump capture kernel won't be
357 See Documentation/networking/ipv6.txt.
359 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
360 Limit apic dumping. The parameter defines the maximal
361 number of local apics being dumped. Also it is possible
362 to set it to "all" by meaning -- no limit here.
363 Format: { 1 (default) | 2 | ... | all }.
364 The parameter valid if only apic=debug or
365 apic=verbose is specified.
366 Example: apic=debug show_lapic=all
368 apm= [APM] Advanced Power Management
369 See header of arch/x86/kernel/apm_32.c.
371 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
372 Format: <io>,<irq>,<nodeID>
376 atarimouse= [HW,MOUSE] Atari Mouse
378 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
379 EzKey and similar keyboards
381 atkbd.reset= [HW] Reset keyboard during initialization
383 atkbd.set= [HW] Select keyboard code set
384 Format: <int> (2 = AT (default), 3 = PS/2)
386 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
389 atkbd.softraw= [HW] Choose between synthetic and real raw mode
390 Format: <bool> (0 = real, 1 = synthetic (default))
392 atkbd.softrepeat= [HW]
393 Use software keyboard repeat
395 audit= [KNL] Enable the audit sub-system
396 Format: { "0" | "1" | "off" | "on" }
397 0 | off - kernel audit is disabled and can not be
398 enabled until the next reboot
399 unset - kernel audit is initialized but disabled and
400 will be fully enabled by the userspace auditd.
401 1 | on - kernel audit is initialized and partially
402 enabled, storing at most audit_backlog_limit
403 messages in RAM until it is fully enabled by the
407 audit_backlog_limit= [KNL] Set the audit queue size limit.
408 Format: <int> (must be >=0)
411 bau= [X86_UV] Enable the BAU on SGI UV. The default
412 behavior is to disable the BAU (i.e. bau=0).
413 Format: { "0" | "1" }
416 unset - Disable the BAU.
418 baycom_epp= [HW,AX25]
421 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
423 See header of drivers/net/hamradio/baycom_par.c.
425 baycom_ser_fdx= [HW,AX25]
426 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
427 Format: <io>,<irq>,<mode>[,<baud>]
428 See header of drivers/net/hamradio/baycom_ser_fdx.c.
430 baycom_ser_hdx= [HW,AX25]
431 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
432 Format: <io>,<irq>,<mode>
433 See header of drivers/net/hamradio/baycom_ser_hdx.c.
435 blkdevparts= Manual partition parsing of block device(s) for
436 embedded devices based on command line input.
437 See Documentation/block/cmdline-partition.txt
439 boot_delay= Milliseconds to delay each printk during boot.
440 Values larger than 10 seconds (10000) are changed to
444 bootmem_debug [KNL] Enable bootmem allocator debug messages.
447 Disable BERT OS support on buggy BIOSes.
449 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
450 bttv.radio= Most important insmod options are available as
452 bttv.pll= See Documentation/media/v4l-drivers/bttv.rst
455 bulk_remove=off [PPC] This parameter disables the use of the pSeries
456 firmware feature for flushing multiple hpte entries
459 c101= [NET] Moxa C101 synchronous serial card
461 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
462 Sometimes CPU hardware bugs make them report the cache
463 size incorrectly. The kernel will attempt work arounds
464 to fix known problems, but for some CPUs it is not
465 possible to determine what the correct size should be.
466 This option provides an override for these situations.
468 ca_keys= [KEYS] This parameter identifies a specific key(s) on
469 the system trusted keyring to be used for certificate
471 format: { id:<keyid> | builtin }
473 cca= [MIPS] Override the kernel pages' cache coherency
474 algorithm. Accepted values range from 0 to 7
475 inclusive. See arch/mips/include/asm/pgtable-bits.h
476 for platform specific values (SB1, Loongson3 and
479 ccw_timeout_log [S390]
480 See Documentation/s390/CommonIO for details.
482 cgroup_disable= [KNL] Disable a particular controller
483 Format: {name of the controller(s) to disable}
484 The effects of cgroup_disable=foo are:
485 - foo isn't auto-mounted if you mount all cgroups in
487 - foo isn't visible as an individually mountable
489 {Currently only "memory" controller deal with this and
490 cut the overhead, others just disable the usage. So
491 only cgroup_disable=memory is actually worthy}
493 cgroup_no_v1= [KNL] Disable one, multiple, all cgroup controllers in v1
494 Format: { controller[,controller...] | "all" }
495 Like cgroup_disable, but only applies to cgroup v1;
496 the blacklisted controllers remain available in cgroup2.
498 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
500 nosocket -- Disable socket memory accounting.
501 nokmem -- Disable kernel memory accounting.
503 checkreqprot [SELINUX] Set initial checkreqprot flag value.
504 Format: { "0" | "1" }
505 See security/selinux/Kconfig help text.
506 0 -- check protection applied by kernel (includes
507 any implied execute protection).
508 1 -- check protection requested by application.
509 Default value is set via a kernel config option.
510 Value can be changed at runtime via
511 /selinux/checkreqprot.
514 See Documentation/s390/CommonIO for details.
517 Prevents the clock framework from automatically gating
518 clocks that have not been explicitly enabled by a Linux
519 device driver but are enabled in hardware at reset or
520 by the bootloader/firmware. Note that this does not
521 force such clocks to be always-on nor does it reserve
522 those clocks in any way. This parameter is useful for
523 debug and development, but should not be needed on a
524 platform with proper driver support. For more
525 information, see Documentation/driver-api/clk.rst.
527 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
529 Forces specified clocksource (if available) to be used
530 when calculating gettimeofday(). If specified
531 clocksource is not available, it defaults to PIT.
532 Format: { pit | tsc | cyclone | pmtmr }
534 clocksource= Override the default clocksource
536 Override the default clocksource and use the clocksource
537 with the name specified.
538 Some clocksource names to choose from, depending on
540 [all] jiffies (this is the base, fallback clocksource)
542 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
543 pxa_timer,timer3,32k_counter,timer0_1
544 [X86-32] pit,hpet,tsc;
545 scx200_hrt on Geode; cyclone on IBM x440
553 clocksource.arm_arch_timer.evtstrm=
556 Enable/disable the eventstream feature of the ARM
557 architected timer so that code using WFE-based polling
558 loops can be debugged more effectively on production
561 clocksource.max_cswd_read_retries= [KNL]
562 Number of clocksource_watchdog() retries due to
563 external delays before the clock will be marked
564 unstable. Defaults to three retries, that is,
565 four attempts to read the clock under test.
567 clearcpuid=BITNUM[,BITNUM...] [X86]
568 Disable CPUID feature X for the kernel. See
569 arch/x86/include/asm/cpufeatures.h for the valid bit
570 numbers. Note the Linux specific bits are not necessarily
571 stable over kernel options, but the vendor specific
573 Also note that user programs calling CPUID directly
574 or using the feature without checking anything
575 will still see it. This just prevents it from
576 being used by the kernel or shown in /proc/cpuinfo.
577 Also note the kernel might malfunction if you disable
580 cma=nn[MG]@[start[MG][-end[MG]]]
582 Sets the size of kernel global memory area for
583 contiguous memory allocations and optionally the
584 placement constraint by the physical address range of
585 memory allocations. A value of 0 disables CMA
586 altogether. For more information, see
587 include/linux/dma-contiguous.h
589 cmo_free_hint= [PPC] Format: { yes | no }
590 Specify whether pages are marked as being inactive
591 when they are freed. This is used in CMO environments
592 to determine OS memory pressure for page stealing by
596 coherent_pool=nn[KMG] [ARM,KNL]
597 Sets the size of memory pool for coherent, atomic dma
598 allocations, by default set to 256K.
600 com20020= [HW,NET] ARCnet - COM20020 chipset
602 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
604 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
608 ARCnet - COM90xx chipset (memory-mapped buffers)
609 Format: <io>[,<irq>[,<memstart>]]
611 condev= [HW,S390] console device
614 console= [KNL] Output console device and options.
616 tty<n> Use the virtual console device <n>.
620 Use the specified serial port. The options are of
621 the form "bbbbpnf", where "bbbb" is the baud rate,
622 "p" is parity ("n", "o", or "e"), "n" is number of
623 bits, and "f" is flow control ("r" for RTS or
624 omit it). Default is "9600n8".
626 See Documentation/admin-guide/serial-console.rst for more
628 Documentation/networking/netconsole.txt for an
631 uart[8250],io,<addr>[,options]
632 uart[8250],mmio,<addr>[,options]
633 uart[8250],mmio16,<addr>[,options]
634 uart[8250],mmio32,<addr>[,options]
635 uart[8250],0x<addr>[,options]
636 Start an early, polled-mode console on the 8250/16550
637 UART at the specified I/O port or MMIO address,
638 switching to the matching ttyS device later.
639 MMIO inter-register address stride is either 8-bit
640 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
641 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
642 to be equivalent to 'mmio'. 'options' are specified in
643 the same format described for ttyS above; if unspecified,
644 the h/w is not re-initialized.
646 hvc<n> Use the hypervisor console device <n>. This is for
647 both Xen and PowerPC hypervisors.
649 If the device connected to the port is not a TTY but a braille
650 device, prepend "brl," before the device type, for instance
652 For now, only VisioBraille is supported.
655 [KNL] Change console messages format
657 By default we print messages on consoles in
658 "[time stamp] text\n" format (time stamp may not be
659 printed, depending on CONFIG_PRINTK_TIME or
660 `printk_time' param).
662 Switch to syslog format: "<%u>[time stamp] text\n"
663 IOW, each message will have a facility and loglevel
664 prefix. The format is similar to one used by syslog()
665 syscall, or to executing "dmesg -S --raw" or to reading
668 consoleblank= [KNL] The console blank (screen saver) timeout in
669 seconds. A value of 0 disables the blank timer.
673 [KNL] Change the default value for
674 /proc/<pid>/coredump_filter.
675 See also Documentation/filesystems/proc.txt.
677 coresight_cpu_debug.enable
680 Enable/disable the CPU sampling based debugging.
681 0: default value, disable debugging
682 1: enable debugging at boot time
684 cpuidle.off=1 [CPU_IDLE]
685 disable the cpuidle sub-system
687 cpufreq.off=1 [CPU_FREQ]
688 disable the cpufreq sub-system
691 [X86] Delay for N microsec between assert and de-assert
692 of APIC INIT to start processors. This delay occurs
693 on every CPU online, such as boot, and resume from suspend.
696 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
698 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
700 crashkernel=size[KMG][@offset[KMG]]
701 [KNL] Using kexec, Linux can switch to a 'crash kernel'
702 upon panic. This parameter reserves the physical
703 memory region [offset, offset + size] for that kernel
704 image. If '@offset' is omitted, then a suitable offset
705 is selected automatically. Check
706 Documentation/kdump/kdump.txt for further details.
708 crashkernel=range1:size1[,range2:size2,...][@offset]
709 [KNL] Same as above, but depends on the memory
710 in the running system. The syntax of range is
711 start-[end] where start and end are both
712 a memory unit (amount[KMG]). See also
713 Documentation/kdump/kdump.txt for an example.
715 crashkernel=size[KMG],high
716 [KNL, x86_64] range could be above 4G. Allow kernel
717 to allocate physical memory region from top, so could
718 be above 4G if system have more than 4G ram installed.
719 Otherwise memory region will be allocated below 4G, if
721 It will be ignored if crashkernel=X is specified.
722 crashkernel=size[KMG],low
723 [KNL, x86_64] range under 4G. When crashkernel=X,high
724 is passed, kernel could allocate physical memory region
725 above 4G, that cause second kernel crash on system
726 that require some amount of low memory, e.g. swiotlb
727 requires at least 64M+32K low memory, also enough extra
728 low memory is needed to make sure DMA buffers for 32-bit
729 devices won't run out. Kernel would try to allocate at
730 at least 256M below 4G automatically.
731 This one let user to specify own low range under 4G
732 for second kernel instead.
733 0: to disable low allocation.
734 It will be ignored when crashkernel=X,high is not used
735 or memory reserved is below 4G.
738 [KNL] Disable crypto self-tests
743 cs89x0_media= [HW,NET]
744 Format: { rj45 | aui | bnc }
747 See header of drivers/s390/block/dasd_devmap.c.
749 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
750 (one device per port)
751 Format: <port#>,<type>
752 See also Documentation/input/devices/joystick-parport.rst
754 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
756 Documentation/admin-guide/dynamic-debug-howto.rst for
757 details. Deprecated, see dyndbg.
759 debug [KNL] Enable kernel debugging (events log level).
762 [KNL] Enable printing [hashed] pointers early in the
763 boot sequence. If enabled, we use a weak hash instead
764 of siphash to hash pointers. Use this option if you are
765 seeing instances of '(___ptrval___)') and need to see a
766 value (hashed pointer) instead. Cryptographically
767 insecure, please do not use on production kernels.
770 [KNL] verbose self-tests
772 Print debugging info while doing the locking API
774 We default to 0 (no extra messages), setting it to
775 1 will print _a lot_ more information - normally
776 only useful to kernel developers.
778 debug_objects [KNL] Enable object debugging
781 [KNL] Disable object debugging
783 debug_guardpage_minorder=
784 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
785 parameter allows control of the order of pages that will
786 be intentionally kept free (and hence protected) by the
787 buddy allocator. Bigger value increase the probability
788 of catching random memory corruption, but reduce the
789 amount of memory for normal system use. The maximum
790 possible value is MAX_ORDER/2. Setting this parameter
791 to 1 or 2 should be enough to identify most random
792 memory corruption problems caused by bugs in kernel or
793 driver code when a CPU writes to (or reads from) a
794 random memory location. Note that there exists a class
795 of memory corruptions problems caused by buggy H/W or
796 F/W or by drivers badly programing DMA (basically when
797 memory is written at bus level and the CPU MMU is
798 bypassed) which are not detectable by
799 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
800 tracking down these problems.
803 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
804 parameter enables the feature at boot time. In
805 default, it is disabled. We can avoid allocating huge
806 chunk of memory for debug pagealloc if we don't enable
807 it at boot time and the system will work mostly same
808 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
809 on: enable the feature
811 debugpat [X86] Enable PAT debugging
813 decnet.addr= [HW,NET]
814 Format: <area>[,<node>]
815 See also Documentation/networking/decnet.txt.
818 [same as hugepagesz=] The size of the default
819 HugeTLB page size. This is the size represented by
820 the legacy /proc/ hugepages APIs, used for SHM, and
821 default size when mounting hugetlbfs filesystems.
822 Defaults to the default architecture's huge page size
825 deferred_probe_timeout=
826 [KNL] Debugging option to set a timeout in seconds for
827 deferred probe to give up waiting on dependencies to
828 probe. Only specific dependencies (subsystems or
829 drivers) that have opted in will be ignored. A timeout of 0
830 will timeout at the end of initcalls. This option will also
831 dump out devices still on the deferred probe list after
835 Set number of hash buckets for dentry cache.
837 disable_1tb_segments [PPC]
838 Disables the use of 1TB hash page table segments. This
839 causes the kernel to fall back to 256MB segments which
840 can be useful when debugging issues that require an SLB
844 See Documentation/networking/ipv6.txt.
847 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
848 hardening is enabled for this boot. Hardened
849 usercopy checking is used to protect the kernel
850 from reading or writing beyond known memory
851 allocation boundaries as a proactive defense
852 against bounds-checking flaws in the kernel's
853 copy_to_user()/copy_from_user() interface.
854 on Perform hardened usercopy checks (default).
855 off Disable hardened usercopy checks.
858 Disable RADIX MMU mode on POWER9
860 disable_cpu_apicid= [X86,APIC,SMP]
862 The number of initial APIC ID for the
863 corresponding CPU to be disabled at boot,
864 mostly used for the kdump 2nd kernel to
865 disable BSP to wake up multiple CPUs without
866 causing system reset or hang due to sending
869 disable_ddw [PPC/PSERIES]
870 Disable Dynamic DMA Window support. Use this if
871 to workaround buggy firmware.
874 See Documentation/networking/ipv6.txt.
876 disable_mtrr_cleanup [X86]
877 The kernel tries to adjust MTRR layout from continuous
878 to discrete, to make X server driver able to add WB
879 entry later. This parameter disables that.
881 disable_mtrr_trim [X86, Intel and AMD only]
882 By default the kernel will trim any uncacheable
883 memory out of your available memory pool based on
884 MTRR settings. This parameter disables that behavior,
885 possibly causing your machine to run very slowly.
887 disable_timer_pin_1 [X86]
888 Disable PIN 1 of APIC timer
889 Can be useful to work around chipset bugs.
891 dis_ucode_ldr [X86] Disable the microcode loader.
893 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
894 this option disables the debugging code at boot.
896 dma_debug_entries=<number>
897 This option allows to tune the number of preallocated
898 entries for DMA-API debugging code. One entry is
899 required per DMA-API allocation. Use this if the
900 DMA-API debugging code disables itself because the
901 architectural default is too low.
903 dma_debug_driver=<driver_name>
904 With this option the DMA-API debugging driver
905 filter feature can be enabled at boot time. Just
906 pass the driver to filter for as the parameter.
907 The filter can be disabled or changed to another
908 driver later using sysfs.
910 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
911 Broken monitors, graphic adapters, KVMs and EDIDless
912 panels may send no or incorrect EDID data sets.
913 This parameter allows to specify an EDID data sets
914 in the /lib/firmware directory that are used instead.
915 Generic built-in EDID data sets are used, if one of
916 edid/1024x768.bin, edid/1280x1024.bin,
917 edid/1680x1050.bin, or edid/1920x1080.bin is given
918 and no file with the same name exists. Details and
919 instructions how to build your own EDID data are
920 available in Documentation/EDID/HOWTO.txt. An EDID
921 data set will only be used for a particular connector,
922 if its name and a colon are prepended to the EDID
923 name. Each connector may use a unique EDID data
924 set by separating the files with a comma. An EDID
925 data set with no connector name will be used for
926 any connectors not explicitly specified.
931 Format: {"off" | "known"}
932 Control how the dt_cpu_ftrs device-tree binding is
933 used for CPU feature discovery and setup (if it
935 off: Do not use it, fall back to legacy cpu table.
936 known: Do not pass through unknown features to guests
937 or userspace, only those that the kernel is aware of.
939 dump_apple_properties [X86]
940 Dump name and content of EFI device properties on
941 x86 Macs. Useful for driver authors to determine
942 what data is available or for reverse-engineering.
944 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
945 module.dyndbg[="val"]
946 Enable debug messages at boot time. See
947 Documentation/admin-guide/dynamic-debug-howto.rst
950 nompx [X86] Disables Intel Memory Protection Extensions.
951 See Documentation/x86/intel_mpx.txt for more
952 information about the feature.
954 nopku [X86] Disable Memory Protection Keys CPU feature found
957 module.async_probe [KNL]
958 Enable asynchronous probe on this module.
960 early_ioremap_debug [KNL]
961 Enable debug messages in early_ioremap support. This
962 is useful for tracking down temporary early mappings
963 which are not unmapped.
965 earlycon= [KNL] Output early console device and options.
967 [ARM64] The early console is determined by the
968 stdout-path property in device tree's chosen node,
969 or determined by the ACPI SPCR table.
971 [X86] When used with no options the early console is
972 determined by the ACPI SPCR table.
974 cdns,<addr>[,options]
975 Start an early, polled-mode console on a Cadence
976 (xuartps) serial port at the specified address. Only
977 supported option is baud rate. If baud rate is not
978 specified, the serial port must already be setup and
981 uart[8250],io,<addr>[,options]
982 uart[8250],mmio,<addr>[,options]
983 uart[8250],mmio32,<addr>[,options]
984 uart[8250],mmio32be,<addr>[,options]
985 uart[8250],0x<addr>[,options]
986 Start an early, polled-mode console on the 8250/16550
987 UART at the specified I/O port or MMIO address.
988 MMIO inter-register address stride is either 8-bit
989 (mmio) or 32-bit (mmio32 or mmio32be).
990 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
991 to be equivalent to 'mmio'. 'options' are specified
992 in the same format described for "console=ttyS<n>"; if
993 unspecified, the h/w is not initialized.
997 Start an early, polled-mode console on a pl011 serial
998 port at the specified address. The pl011 serial port
999 must already be setup and configured. Options are not
1000 yet supported. If 'mmio32' is specified, then only
1001 the driver will use only 32-bit accessors to read/write
1002 the device registers.
1005 Start an early, polled-mode console on a meson serial
1006 port at the specified address. The serial port must
1007 already be setup and configured. Options are not yet
1011 Start an early, polled-mode console on an msm serial
1012 port at the specified address. The serial port
1013 must already be setup and configured. Options are not
1016 msm_serial_dm,<addr>
1017 Start an early, polled-mode console on an msm serial
1018 dm port at the specified address. The serial port
1019 must already be setup and configured. Options are not
1023 Start an early, polled-mode console on a serial port
1024 of an Actions Semi SoC, such as S500 or S900, at the
1025 specified address. The serial port must already be
1026 setup and configured. Options are not yet supported.
1028 smh Use ARM semihosting calls for early console.
1036 Use early console provided by serial driver available
1037 on Samsung SoCs, requires selecting proper type and
1038 a correct base address of the selected UART port. The
1039 serial port must already be setup and configured.
1040 Options are not yet supported.
1043 Start an early, polled-mode console on a lantiq serial
1044 (lqasc) port at the specified address. The serial port
1045 must already be setup and configured. Options are not
1050 Use early console provided by Freescale LP UART driver
1051 found on Freescale Vybrid and QorIQ LS1021A processors.
1052 A valid base address must be provided, and the serial
1053 port must already be setup and configured.
1056 Start an early, polled-mode console on the
1057 Armada 3700 serial port at the specified
1058 address. The serial port must already be setup
1059 and configured. Options are not yet supported.
1062 Start an early, polled-mode console on a Qualcomm
1063 Generic Interface (GENI) based serial port at the
1064 specified address. The serial port must already be
1065 setup and configured. Options are not yet supported.
1067 earlyprintk= [X86,SH,ARM,M68k,S390]
1072 earlyprintk=serial[,ttySn[,baudrate]]
1073 earlyprintk=serial[,0x...[,baudrate]]
1074 earlyprintk=ttySn[,baudrate]
1075 earlyprintk=dbgp[debugController#]
1076 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1077 earlyprintk=xdbc[xhciController#]
1079 earlyprintk is useful when the kernel crashes before
1080 the normal console is initialized. It is not enabled by
1081 default because it has some cosmetic problems.
1083 Append ",keep" to not disable it when the real console
1086 Only one of vga, efi, serial, or usb debug port can
1089 Currently only ttyS0 and ttyS1 may be specified by
1090 name. Other I/O ports may be explicitly specified
1091 on some architectures (x86 and arm at least) by
1092 replacing ttySn with an I/O port address, like this:
1093 earlyprintk=serial,0x1008,115200
1094 You can find the port for a given device in
1095 /proc/tty/driver/serial:
1096 2: uart:ST16650V2 port:00001008 irq:18 ...
1098 Interaction with the standard serial driver is not
1101 The VGA and EFI output is eventually overwritten by
1104 The xen output can only be used by Xen PV guests.
1106 The sclp output can only be used on s390.
1108 The optional "force" to "pciserial" enables use of a
1109 PCI device even when its classcode is not of the
1112 edac_report= [HW,EDAC] Control how to report EDAC event
1113 Format: {"on" | "off" | "force"}
1114 on: enable EDAC to report H/W event. May be overridden
1115 by other higher priority error reporting module.
1116 off: disable H/W event reporting through EDAC.
1117 force: enforce the use of EDAC to report H/W event.
1120 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1123 This is designed to be used in conjunction with
1124 the boot argument: earlyprintk=vga
1127 Format: {"off" | "on" | "skip[mbr]"}
1130 Format: { "old_map", "nochunk", "noruntime", "debug" }
1131 old_map [X86-64]: switch to the old ioremap-based EFI
1132 runtime services mapping. 32-bit still uses this one by
1134 nochunk: disable reading files in "chunks" in the EFI
1135 boot stub, as chunking can cause problems with some
1136 firmware implementations.
1137 noruntime : disable EFI runtime services support
1138 debug: enable misc debug output
1140 efi_no_storage_paranoia [EFI; X86]
1141 Using this parameter you can use more than 50% of
1142 your efi variable storage. Use this parameter only if
1143 you are really sure that your UEFI does sane gc and
1144 fulfills the spec otherwise your board may brick.
1146 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1147 Add arbitrary attribute to specific memory range by
1148 updating original EFI memory map.
1149 Region of memory which aa attribute is added to is
1151 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1152 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1153 attribute is added to range 0x100000000-0x180000000 and
1154 0x10a0000000-0x1120000000.
1156 Using this parameter you can do debugging of EFI memmap
1157 related feature. For example, you can do debugging of
1158 Address Range Mirroring feature even if your box
1161 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1162 that is to be dynamically loaded by Linux. If there are
1163 multiple variables with the same name but with different
1164 vendor GUIDs, all of them will be loaded. See
1165 Documentation/acpi/ssdt-overlays.txt for details.
1168 eisa_irq_edge= [PARISC,HW]
1169 See header of drivers/parisc/eisa.c.
1172 See comment before function elanfreq_setup() in
1173 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1176 Format: {"cfq" | "deadline" | "noop"}
1177 See Documentation/block/cfq-iosched.txt and
1178 Documentation/block/deadline-iosched.txt for details.
1180 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1181 Specifies physical address of start of kernel core
1182 image elf header and optionally the size. Generally
1183 kexec loader will pass this option to capture kernel.
1184 See Documentation/kdump/kdump.txt for details.
1186 enable_mtrr_cleanup [X86]
1187 The kernel tries to adjust MTRR layout from continuous
1188 to discrete, to make X server driver able to add WB
1189 entry later. This parameter enables that.
1191 enable_timer_pin_1 [X86]
1192 Enable PIN 1 of APIC timer
1193 Can be useful to work around chipset bugs
1194 (in particular on some ATI chipsets).
1195 The kernel tries to set a reasonable default.
1197 enforcing [SELINUX] Set initial enforcing status.
1199 See security/selinux/Kconfig help text.
1200 0 -- permissive (log only, no denials).
1201 1 -- enforcing (deny and log).
1203 Value can be changed at runtime via /selinux/enforce.
1206 Disable Error Record Serialization Table (ERST)
1209 ether= [HW,NET] Ethernet cards parameters
1210 This option is obsoleted by the "netdev=" option, which
1211 has equivalent usage. See its documentation for details.
1215 Permit 'security.evm' to be updated regardless of
1216 current integrity status.
1220 fail_make_request=[KNL]
1221 General fault injection mechanism.
1222 Format: <interval>,<probability>,<space>,<times>
1223 See also Documentation/fault-injection/.
1226 See Documentation/blockdev/floppy.txt.
1228 force_pal_cache_flush
1229 [IA-64] Avoid check_sal_cache_flush which may hang on
1230 buggy SAL_CACHE_FLUSH implementations. Using this
1231 parameter will force ia64_sal_cache_flush to call
1232 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1235 Forcefully enable Physical Address Extension (PAE).
1236 Many Pentium M systems disable PAE but may have a
1237 functionally usable PAE implementation.
1238 Warning: use of this parameter will taint the kernel
1239 and may cause unknown problems.
1242 [FTRACE] will set and start the specified tracer
1243 as early as possible in order to facilitate early
1246 ftrace_dump_on_oops[=orig_cpu]
1247 [FTRACE] will dump the trace buffers on oops.
1248 If no parameter is passed, ftrace will dump
1249 buffers of all CPUs, but if you pass orig_cpu, it will
1250 dump only the buffer of the CPU that triggered the
1253 ftrace_filter=[function-list]
1254 [FTRACE] Limit the functions traced by the function
1255 tracer at boot up. function-list is a comma separated
1256 list of functions. This list can be changed at run
1257 time by the set_ftrace_filter file in the debugfs
1260 ftrace_notrace=[function-list]
1261 [FTRACE] Do not trace the functions specified in
1262 function-list. This list can be changed at run time
1263 by the set_ftrace_notrace file in the debugfs
1266 ftrace_graph_filter=[function-list]
1267 [FTRACE] Limit the top level callers functions traced
1268 by the function graph tracer at boot up.
1269 function-list is a comma separated list of functions
1270 that can be changed at run time by the
1271 set_graph_function file in the debugfs tracing directory.
1273 ftrace_graph_notrace=[function-list]
1274 [FTRACE] Do not trace from the functions specified in
1275 function-list. This list is a comma separated list of
1276 functions that can be changed at run time by the
1277 set_graph_notrace file in the debugfs tracing directory.
1279 ftrace_graph_max_depth=<uint>
1280 [FTRACE] Used with the function graph tracer. This is
1281 the max depth it will trace into a function. This value
1282 can be changed at run time by the max_graph_depth file
1283 in the tracefs tracing directory. default: 0 (no limit)
1286 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1287 support via parallel port (up to 5 devices per port)
1288 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1289 See also Documentation/input/devices/joystick-parport.rst
1293 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1297 gcov_persist= [GCOV] When non-zero (default), profiling data for
1298 kernel modules is saved and remains accessible via
1299 debugfs, even when the module is unloaded/reloaded.
1300 When zero, profiling data is discarded and associated
1301 debugfs files are removed at module unload time.
1303 goldfish [X86] Enable the goldfish android emulator platform.
1304 Don't use this when you are not running on the
1307 gpt [EFI] Forces disk with valid GPT signature but
1308 invalid Protective MBR to be treated as GPT. If the
1309 primary GPT is corrupted, it enables the backup/alternate
1310 GPT to be used instead.
1312 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1313 the "Enable 0" bit of the configuration register.
1316 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1317 the "Enable 0" bit of the configuration register.
1320 grcan.select= [HW] Select which physical interface to use.
1323 grcan.txsize= [HW] Sets the size of the tx buffer.
1324 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1326 grcan.rxsize= [HW] Sets the size of the rx buffer.
1327 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1330 gpio-mockup.gpio_mockup_ranges
1331 [HW] Sets the ranges of gpiochip of for this device.
1332 Format: <start1>,<end1>,<start2>,<end2>...
1334 hardlockup_all_cpu_backtrace=
1335 [KNL] Should the hard-lockup detector generate
1336 backtraces on all cpus.
1339 hashdist= [KNL,NUMA] Large hashes allocated during boot
1340 are distributed across NUMA nodes. Defaults on
1341 for 64-bit NUMA, off otherwise.
1342 Format: 0 | 1 (for off | on)
1344 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1346 hd= [EIDE] (E)IDE hard drive subsystem geometry
1347 Format: <cyl>,<head>,<sect>
1350 Disable Hardware Error Source Table (HEST) support;
1351 corresponding firmware-first mode error processing
1352 logic will be disabled.
1354 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1355 size of <nn>. This works even on boxes that have no
1356 highmem otherwise. This also works to reduce highmem
1357 size on bigger boxes.
1359 highres= [KNL] Enable/disable high resolution timer mode.
1360 Valid parameters: "on", "off"
1364 See Documentation/isdn/README.HiSax.
1368 hpet= [X86-32,HPET] option to control HPET usage
1369 Format: { enable (default) | disable | force |
1371 disable: disable HPET and use PIT instead
1372 force: allow force enabled of undocumented chips (ICH4,
1374 verbose: show contents of HPET registers during setup
1376 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1377 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1379 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1380 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1381 On x86-64 and powerpc, this option can be specified
1382 multiple times interleaved with hugepages= to reserve
1383 huge pages of different sizes. Valid pages sizes on
1384 x86-64 are 2M (when the CPU supports "pse") and 1G
1385 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1388 [KNL] Should the hung task detector generate panics.
1391 A nonzero value instructs the kernel to panic when a
1392 hung task is detected. The default value is controlled
1393 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1394 option. The value selected by this boot parameter can
1395 be changed later by the kernel.hung_task_panic sysctl.
1397 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1398 terminal devices. Valid values: 0..8
1399 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1400 If specified, z/VM IUCV HVC accepts connections
1401 from listed z/VM user IDs only.
1403 Do not unregister boot console at start. This is only
1404 useful for debugging when something happens in the window
1405 between unregistering the boot console and initializing
1408 i2c_bus= [HW] Override the default board specific I2C bus speed
1409 or register an additional I2C bus that is not
1410 registered from board initialization code.
1414 i8042.debug [HW] Toggle i8042 debug mode
1415 i8042.unmask_kbd_data
1416 [HW] Enable printing of interrupt data from the KBD port
1417 (disabled by default, and as a pre-condition
1418 requires that i8042.debug=1 be enabled)
1419 i8042.direct [HW] Put keyboard port into non-translated mode
1420 i8042.dumbkbd [HW] Pretend that controller can only read data from
1421 keyboard and cannot control its state
1422 (Don't attempt to blink the leds)
1423 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1424 i8042.nokbd [HW] Don't check/create keyboard port
1425 i8042.noloop [HW] Disable the AUX Loopback command while probing
1427 i8042.nomux [HW] Don't check presence of an active multiplexing
1429 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1431 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1432 i8042.reset [HW] Reset the controller during init, cleanup and
1433 suspend-to-ram transitions, only during s2r
1434 transitions, or never reset
1435 Format: { 1 | Y | y | 0 | N | n }
1436 1, Y, y: always reset controller
1437 0, N, n: don't ever reset controller
1438 Default: only on s2r transitions on x86; most other
1439 architectures force reset to be always executed
1440 i8042.unlock [HW] Unlock (ignore) the keylock
1441 i8042.kbdreset [HW] Reset device connected to KBD port
1445 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1446 indicates that the driver is running on unsupported
1448 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1449 does not match list of supported models.
1451 [HW] Report power status in /proc/i8k
1452 (disabled by default)
1453 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1456 i915.invert_brightness=
1457 [DRM] Invert the sense of the variable that is used to
1458 set the brightness of the panel backlight. Normally a
1459 brightness value of 0 indicates backlight switched off,
1460 and the maximum of the brightness value sets the backlight
1461 to maximum brightness. If this parameter is set to 0
1462 (default) and the machine requires it, or this parameter
1463 is set to 1, a brightness value of 0 sets the backlight
1464 to maximum brightness, and the maximum of the brightness
1465 value switches the backlight off.
1466 -1 -- never invert brightness
1467 0 -- machine default
1468 1 -- force brightness inversion
1471 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1473 ide-core.nodma= [HW] (E)IDE subsystem
1474 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1475 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1476 .cdrom .chs .ignore_cable are additional options
1477 See Documentation/ide/ide.txt.
1479 ide-generic.probe-mask= [HW] (E)IDE subsystem
1481 Probe mask for legacy ISA IDE ports. Depending on
1482 platform up to 6 ports are supported, enabled by
1483 setting corresponding bits in the mask to 1. The
1484 default value is 0x0, which has a special meaning.
1485 On systems that have PCI, it triggers scanning the
1486 PCI bus for the first and the second port, which
1487 are then probed. On systems without PCI the value
1488 of 0x0 enables probing the two first ports as if it
1491 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1492 Claim all unknown PCI IDE storage controllers.
1495 Format: idle=poll, idle=halt, idle=nomwait
1496 Poll forces a polling idle loop that can slightly
1497 improve the performance of waking up a idle CPU, but
1498 will use a lot of power and make the system run hot.
1500 idle=halt: Halt is forced to be used for CPU idle.
1501 In such case C2/C3 won't be used again.
1502 idle=nomwait: Disable mwait for CPU C-states
1504 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1505 Format: { strict | legacy | 2008 | relaxed }
1508 Choose which programs will be accepted for execution
1509 based on the IEEE 754 NaN encoding(s) supported by
1510 the FPU and the NaN encoding requested with the value
1511 of an ELF file header flag individually set by each
1512 binary. Hardware implementations are permitted to
1513 support either or both of the legacy and the 2008 NaN
1516 Available settings are as follows:
1517 strict accept binaries that request a NaN encoding
1518 supported by the FPU
1519 legacy only accept legacy-NaN binaries, if supported
1521 2008 only accept 2008-NaN binaries, if supported
1523 relaxed accept any binaries regardless of whether
1524 supported by the FPU
1526 The FPU emulator is always able to support both NaN
1527 encodings, so if no FPU hardware is present or it has
1528 been disabled with 'nofpu', then the settings of
1529 'legacy' and '2008' strap the emulator accordingly,
1530 'relaxed' straps the emulator for both legacy-NaN and
1531 2008-NaN, whereas 'strict' enables legacy-NaN only on
1532 legacy processors and both NaN encodings on MIPS32 or
1535 The setting for ABS.fmt/NEG.fmt instruction execution
1536 mode generally follows that for the NaN encoding,
1537 except where unsupported by hardware.
1539 ignore_loglevel [KNL]
1540 Ignore loglevel setting - this will print /all/
1541 kernel messages to the console. Useful for debugging.
1542 We also add it as printk module parameter, so users
1543 could change it dynamically, usually by
1544 /sys/module/printk/parameters/ignore_loglevel.
1547 Ignore RLIMIT_DATA setting for data mappings,
1548 print warning at first misuse. Can be changed via
1549 /sys/module/kernel/parameters/ignore_rlimit_data.
1551 ihash_entries= [KNL]
1552 Set number of hash buckets for inode cache.
1554 ima_appraise= [IMA] appraise integrity measurements
1555 Format: { "off" | "enforce" | "fix" | "log" }
1558 ima_appraise_tcb [IMA]
1559 The builtin appraise policy appraises all files
1562 ima_canonical_fmt [IMA]
1563 Use the canonical format for the binary runtime
1564 measurements, instead of host native format.
1567 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1571 The list of supported hash algorithms is defined
1572 in crypto/hash_info.h.
1575 The builtin policies to load during IMA setup.
1576 Format: "tcb | appraise_tcb | secure_boot |
1579 The "tcb" policy measures all programs exec'd, files
1580 mmap'd for exec, and all files opened with the read
1581 mode bit set by either the effective uid (euid=0) or
1584 The "appraise_tcb" policy appraises the integrity of
1585 all files owned by root. (This is the equivalent
1586 of ima_appraise_tcb.)
1588 The "secure_boot" policy appraises the integrity
1589 of files (eg. kexec kernel image, kernel modules,
1590 firmware, policy, etc) based on file signatures.
1592 The "fail_securely" policy forces file signature
1593 verification failure also on privileged mounted
1594 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1597 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1598 Load a policy which meets the needs of the Trusted
1599 Computing Base. This means IMA will measure all
1600 programs exec'd, files mmap'd for exec, and all files
1601 opened for read by uid=0.
1604 Select one of defined IMA measurements template formats.
1605 Formats: { "ima" | "ima-ng" | "ima-sig" }
1609 [IMA] Define a custom template format.
1610 Format: { "field1|...|fieldN" }
1612 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1613 Format: <min_file_size>
1614 Set the minimal file size for using asynchronous hash.
1615 If left unspecified, ahash usage is disabled.
1617 ahash performance varies for different data sizes on
1618 different crypto accelerators. This option can be used
1619 to achieve the best performance for a particular HW.
1621 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1623 Set hashing buffer size. Default: 4k.
1625 ahash performance varies for different chunk sizes on
1626 different crypto accelerators. This option can be used
1627 to achieve best performance for particular HW.
1631 Run specified binary instead of /sbin/init as init
1634 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1635 for working out where the kernel is dying during
1638 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1639 initcall functions. Useful for debugging built-in
1640 modules and initcalls.
1642 initrd= [BOOT] Specify the location of the initial ramdisk
1644 init_pkru= [x86] Specify the default memory protection keys rights
1645 register contents for all processes. 0x55555554 by
1646 default (disallow access to all but pkey 0). Can
1647 override in debugfs after boot.
1649 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1652 int_pln_enable [x86] Enable power limit notification interrupt
1654 integrity_audit=[IMA]
1655 Format: { "0" | "1" }
1656 0 -- basic integrity auditing messages. (Default)
1657 1 -- additional integrity auditing messages.
1659 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1661 Enable intel iommu driver.
1663 Disable intel iommu driver.
1664 igfx_off [Default Off]
1665 By default, gfx is mapped as normal device. If a gfx
1666 device has a dedicated DMAR unit, the DMAR unit is
1667 bypassed by not enabling DMAR with this option. In
1668 this case, gfx device will use physical address for
1671 With this option iommu will not optimize to look
1672 for io virtual address below 32-bit forcing dual
1673 address cycle on pci bus for cards supporting greater
1674 than 32-bit addressing. The default is to look
1675 for translation below 32-bit and if not available
1676 then look in the higher range.
1677 strict [Default Off]
1678 With this option on every unmap_single operation will
1679 result in a hardware IOTLB flush operation as opposed
1680 to batching them for performance.
1681 sp_off [Default Off]
1682 By default, super page will be supported if Intel IOMMU
1683 has the capability. With this option, super page will
1685 ecs_off [Default Off]
1686 By default, extended context tables will be supported if
1687 the hardware advertises that it has support both for the
1688 extended tables themselves, and also PASID support. With
1689 this option set, extended tables will not be used even
1690 on hardware which claims to support them.
1691 tboot_noforce [Default Off]
1692 Do not force the Intel IOMMU enabled under tboot.
1693 By default, tboot will force Intel IOMMU on, which
1694 could harm performance of some high-throughput
1695 devices like 40GBit network cards, even if identity
1697 Note that using this option lowers the security
1698 provided by tboot because it makes the system
1699 vulnerable to DMA attacks.
1701 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1702 0 disables intel_idle and fall back on acpi_idle.
1703 1 to 9 specify maximum depth of C-state.
1707 Do not enable intel_pstate as the default
1708 scaling driver for the supported processors
1710 Use intel_pstate as a scaling driver, but configure it
1711 to work with generic cpufreq governors (instead of
1712 enabling its internal governor). This mode cannot be
1713 used along with the hardware-managed P-states (HWP)
1716 Enable intel_pstate on systems that prohibit it by default
1717 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1718 instead of acpi-cpufreq may disable platform features, such
1719 as thermal controls and power capping, that rely on ACPI
1720 P-States information being indicated to OSPM and therefore
1721 should be used with caution. This option does not work with
1722 processors that aren't supported by the intel_pstate driver
1723 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1725 Do not enable hardware P state control (HWP)
1728 Only load intel_pstate on systems which support
1729 hardware P state control (HWP) if available.
1731 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1732 Description Table, specifies preferred power management
1733 profile as "Enterprise Server" or "Performance Server",
1734 then this feature is turned on by default.
1736 Allow per-logical-CPU P-State performance control limits using
1737 cpufreq sysfs interface
1739 intremap= [X86-64, Intel-IOMMU]
1740 on enable Interrupt Remapping (default)
1741 off disable Interrupt Remapping
1742 nosid disable Source ID checking
1744 BIOS x2APIC opt-out request will be ignored
1745 nopost disable Interrupt Posting
1747 iomem= Disable strict checking of access to MMIO memory
1748 strict regions from userspace.
1763 nobypass [PPC/POWERNV]
1764 Disable IOMMU bypass, using IOMMU for PCI devices.
1767 [ARM64] Configure DMA to bypass the IOMMU by default.
1768 Format: { "0" | "1" }
1769 0 - Use IOMMU translation for DMA.
1770 1 - Bypass the IOMMU for DMA.
1771 unset - Use IOMMU translation for DMA.
1773 io7= [HW] IO7 for Marvel based alpha systems
1774 See comment before marvel_specify_io7 in
1775 arch/alpha/kernel/core_marvel.c.
1777 io_delay= [X86] I/O delay method
1779 Standard port 0x80 based delay
1781 Alternate port 0xed based delay (needed on some systems)
1783 Simple two microseconds delay
1788 See Documentation/filesystems/nfs/nfsroot.txt.
1790 irqaffinity= [SMP] Set the default irq affinity mask
1791 The argument is a cpu list, as described above.
1793 irqchip.gicv2_force_probe=
1796 Force the kernel to look for the second 4kB page
1797 of a GICv2 controller even if the memory range
1798 exposed by the device tree is too small.
1800 irqchip.gicv3_nolpi=
1802 Force the kernel to ignore the availability of
1803 LPIs (and by consequence ITSs). Intended for system
1804 that use the kernel as a bootloader, and thus want
1805 to let secondary kernels in charge of setting up
1809 When an interrupt is not handled search all handlers
1810 for it. Intended to get systems with badly broken
1814 When an interrupt is not handled search all handlers
1815 for it. Also check all handlers each timer
1816 interrupt. Intended to get systems with badly broken
1820 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1822 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1823 [Deprecated - use cpusets instead]
1824 Format: [flag-list,]<cpu-list>
1826 Specify one or more CPUs to isolate from disturbances
1827 specified in the flag list (default: domain):
1830 Disable the tick when a single task runs.
1832 A residual 1Hz tick is offloaded to workqueues, which you
1833 need to affine to housekeeping through the global
1834 workqueue's affinity configured via the
1835 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1836 by using the 'domain' flag described below.
1838 NOTE: by default the global workqueue runs on all CPUs,
1839 so to protect individual CPUs the 'cpumask' file has to
1840 be configured manually after bootup.
1843 Isolate from the general SMP balancing and scheduling
1844 algorithms. Note that performing domain isolation this way
1845 is irreversible: it's not possible to bring back a CPU to
1846 the domains once isolated through isolcpus. It's strongly
1847 advised to use cpusets instead to disable scheduler load
1848 balancing through the "cpuset.sched_load_balance" file.
1849 It offers a much more flexible interface where CPUs can
1850 move in and out of an isolated set anytime.
1852 You can move a process onto or off an "isolated" CPU via
1853 the CPU affinity syscalls or cpuset.
1854 <cpu number> begins at 0 and the maximum value is
1855 "number of CPUs in system - 1".
1857 The format of <cpu-list> is described above.
1863 ivrs_ioapic [HW,X86_64]
1864 Provide an override to the IOAPIC-ID<->DEVICE-ID
1865 mapping provided in the IVRS ACPI table. For
1866 example, to map IOAPIC-ID decimal 10 to
1867 PCI device 00:14.0 write the parameter as:
1868 ivrs_ioapic[10]=00:14.0
1870 ivrs_hpet [HW,X86_64]
1871 Provide an override to the HPET-ID<->DEVICE-ID
1872 mapping provided in the IVRS ACPI table. For
1873 example, to map HPET-ID decimal 0 to
1874 PCI device 00:14.0 write the parameter as:
1875 ivrs_hpet[0]=00:14.0
1877 ivrs_acpihid [HW,X86_64]
1878 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1879 mapping provided in the IVRS ACPI table. For
1880 example, to map UART-HID:UID AMD0020:0 to
1881 PCI device 00:14.5 write the parameter as:
1882 ivrs_acpihid[00:14.5]=AMD0020:0
1884 js= [HW,JOY] Analog joystick
1885 See Documentation/input/joydev/joystick.rst.
1888 When CONFIG_RANDOMIZE_BASE is set, this disables
1889 kernel and module base offset ASLR (Address Space
1890 Layout Randomization).
1893 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
1894 report on every invalid memory access. Without this
1895 parameter KASAN will print report only for the first
1900 kernelcore= [KNL,X86,IA-64,PPC]
1901 Format: nn[KMGTPE] | nn% | "mirror"
1902 This parameter specifies the amount of memory usable by
1903 the kernel for non-movable allocations. The requested
1904 amount is spread evenly throughout all nodes in the
1905 system as ZONE_NORMAL. The remaining memory is used for
1906 movable memory in its own zone, ZONE_MOVABLE. In the
1907 event, a node is too small to have both ZONE_NORMAL and
1908 ZONE_MOVABLE, kernelcore memory will take priority and
1909 other nodes will have a larger ZONE_MOVABLE.
1911 ZONE_MOVABLE is used for the allocation of pages that
1912 may be reclaimed or moved by the page migration
1913 subsystem. Note that allocations like PTEs-from-HighMem
1914 still use the HighMem zone if it exists, and the Normal
1915 zone if it does not.
1917 It is possible to specify the exact amount of memory in
1918 the form of "nn[KMGTPE]", a percentage of total system
1919 memory in the form of "nn%", or "mirror". If "mirror"
1920 option is specified, mirrored (reliable) memory is used
1921 for non-movable allocations and remaining memory is used
1922 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
1923 are exclusive, so you cannot specify multiple forms.
1925 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1926 Format: <Controller#>[,poll interval]
1927 The controller # is the number of the ehci usb debug
1928 port as it is probed via PCI. The poll interval is
1929 optional and is the number seconds in between
1930 each poll cycle to the debug port in case you need
1931 the functionality for interrupting the kernel with
1932 gdb or control-c on the dbgp connection. When
1933 not using this parameter you use sysrq-g to break into
1934 the kernel debugger.
1936 kgdboc= [KGDB,HW] kgdb over consoles.
1937 Requires a tty driver that supports console polling,
1938 or a supported polling keyboard driver (non-usb).
1939 Serial only format: <serial_device>[,baud]
1940 keyboard only format: kbd
1941 keyboard and serial format: kbd,<serial_device>[,baud]
1942 Optional Kernel mode setting:
1943 kms, kbd format: kms,kbd
1944 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1946 kgdbwait [KGDB] Stop kernel execution and enter the
1947 kernel debugger at the earliest opportunity.
1949 kmac= [MIPS] korina ethernet MAC address.
1950 Configure the RouterBoard 532 series on-chip
1951 Ethernet adapter MAC address.
1953 kmemleak= [KNL] Boot-time kmemleak enable/disable
1954 Valid arguments: on, off
1956 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1959 kpti= [ARM64] Control page table isolation of user
1960 and kernel address spaces.
1961 Default: enabled on cores which need mitigation.
1965 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1966 Default is 0 (don't ignore, but inject #GP)
1968 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
1969 Default is false (don't support).
1971 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1976 [KVM] Controls the software workaround for the
1977 X86_BUG_ITLB_MULTIHIT bug.
1978 force : Always deploy workaround.
1979 off : Never deploy workaround.
1980 auto : Deploy workaround based on the presence of
1981 X86_BUG_ITLB_MULTIHIT.
1985 If the software workaround is enabled for the host,
1986 guests do need not to enable it for nested guests.
1988 kvm.nx_huge_pages_recovery_ratio=
1989 [KVM] Controls how many 4KiB pages are periodically zapped
1990 back to huge pages. 0 disables the recovery, otherwise if
1991 the value is N KVM will zap 1/Nth of the 4KiB pages every
1992 minute. The default is 60.
1994 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1995 Default is 1 (enabled)
1997 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1999 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2001 kvm-arm.vgic_v3_group0_trap=
2002 [KVM,ARM] Trap guest accesses to GICv3 group-0
2005 kvm-arm.vgic_v3_group1_trap=
2006 [KVM,ARM] Trap guest accesses to GICv3 group-1
2009 kvm-arm.vgic_v3_common_trap=
2010 [KVM,ARM] Trap guest accesses to GICv3 common
2013 kvm-arm.vgic_v4_enable=
2014 [KVM,ARM] Allow use of GICv4 for direct injection of
2017 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2018 (virtualized MMU) support on capable Intel chips.
2019 Default is 1 (enabled)
2021 kvm-intel.emulate_invalid_guest_state=
2022 [KVM,Intel] Enable emulation of invalid guest states
2023 Default is 0 (disabled)
2025 kvm-intel.flexpriority=
2026 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2027 Default is 1 (enabled)
2030 [KVM,Intel] Enable VMX nesting (nVMX).
2031 Default is 0 (disabled)
2033 kvm-intel.unrestricted_guest=
2034 [KVM,Intel] Disable unrestricted guest feature
2035 (virtualized real and unpaged mode) on capable
2036 Intel chips. Default is 1 (enabled)
2038 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2041 Valid arguments: never, cond, always
2043 always: L1D cache flush on every VMENTER.
2044 cond: Flush L1D on VMENTER only when the code between
2045 VMEXIT and VMENTER can leak host memory.
2046 never: Disables the mitigation
2048 Default is cond (do L1 cache flush in specific instances)
2050 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2051 feature (tagged TLBs) on capable Intel chips.
2052 Default is 1 (enabled)
2054 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2057 The kernel PTE inversion protection is unconditionally
2058 enabled and cannot be disabled.
2061 Provides all available mitigations for the
2062 L1TF vulnerability. Disables SMT and
2063 enables all mitigations in the
2064 hypervisors, i.e. unconditional L1D flush.
2066 SMT control and L1D flush control via the
2067 sysfs interface is still possible after
2068 boot. Hypervisors will issue a warning
2069 when the first VM is started in a
2070 potentially insecure configuration,
2071 i.e. SMT enabled or L1D flush disabled.
2074 Same as 'full', but disables SMT and L1D
2075 flush runtime control. Implies the
2076 'nosmt=force' command line option.
2077 (i.e. sysfs control of SMT is disabled.)
2080 Leaves SMT enabled and enables the default
2081 hypervisor mitigation, i.e. conditional
2084 SMT control and L1D flush control via the
2085 sysfs interface is still possible after
2086 boot. Hypervisors will issue a warning
2087 when the first VM is started in a
2088 potentially insecure configuration,
2089 i.e. SMT enabled or L1D flush disabled.
2093 Disables SMT and enables the default
2094 hypervisor mitigation.
2096 SMT control and L1D flush control via the
2097 sysfs interface is still possible after
2098 boot. Hypervisors will issue a warning
2099 when the first VM is started in a
2100 potentially insecure configuration,
2101 i.e. SMT enabled or L1D flush disabled.
2104 Same as 'flush', but hypervisors will not
2105 warn when a VM is started in a potentially
2106 insecure configuration.
2109 Disables hypervisor mitigations and doesn't
2111 It also drops the swap size and available
2112 RAM limit restriction on both hypervisor and
2117 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2123 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2126 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2127 value for LAPIC timer one-shot implementation. Default
2128 back to the programmable timer unit in the LAPIC.
2130 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2133 libata.dma= [LIBATA] DMA control
2134 libata.dma=0 Disable all PATA and SATA DMA
2135 libata.dma=1 PATA and SATA Disk DMA only
2136 libata.dma=2 ATAPI (CDROM) DMA only
2137 libata.dma=4 Compact Flash DMA only
2138 Combinations also work, so libata.dma=3 enables DMA
2139 for disks and CDROMs, but not CFs.
2141 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2142 libata.ignore_hpa=0 keep BIOS limits (default)
2143 libata.ignore_hpa=1 ignore limits, using full disk
2145 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2149 libata.force= [LIBATA] Force configurations. The format is comma
2150 separated list of "[ID:]VAL" where ID is
2151 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2152 matching port, link or device. Basically, it matches
2153 the ATA ID string printed on console by libata. If
2154 the whole ID part is omitted, the last PORT and DEVICE
2155 values are used. If ID hasn't been specified yet, the
2156 configuration applies to all ports, links and devices.
2158 If only DEVICE is omitted, the parameter applies to
2159 the port and all links and devices behind it. DEVICE
2160 number of 0 either selects the first device or the
2161 first fan-out link behind PMP device. It does not
2162 select the host link. DEVICE number of 15 selects the
2163 host link and device attached to it.
2165 The VAL specifies the configuration to force. As long
2166 as there's no ambiguity shortcut notation is allowed.
2167 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2168 The following configurations can be forced.
2170 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2171 Any ID with matching PORT is used.
2173 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2175 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2176 udma[/][16,25,33,44,66,100,133] notation is also
2179 * [no]ncq: Turn on or off NCQ.
2181 * [no]ncqtrim: Turn off queued DSM TRIM.
2183 * nohrst, nosrst, norst: suppress hard, soft
2186 * rstonce: only attempt one reset during
2187 hot-unplug link recovery
2189 * dump_id: dump IDENTIFY data.
2191 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2193 * disable: Disable this device.
2195 If there are multiple matching configurations changing
2196 the same attribute, the last one is used.
2198 memblock=debug [KNL] Enable memblock debug messages.
2200 load_ramdisk= [RAM] List of ramdisks to load from floppy
2201 See Documentation/blockdev/ramdisk.txt.
2203 lockd.nlm_grace_period=P [NFS] Assign grace period.
2206 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2209 lockd.nlm_timeout=T [NFS] Assign timeout value.
2212 lockd.nlm_udpport=M [NFS] Assign UDP port.
2215 locktorture.nreaders_stress= [KNL]
2216 Set the number of locking read-acquisition kthreads.
2217 Defaults to being automatically set based on the
2218 number of online CPUs.
2220 locktorture.nwriters_stress= [KNL]
2221 Set the number of locking write-acquisition kthreads.
2223 locktorture.onoff_holdoff= [KNL]
2224 Set time (s) after boot for CPU-hotplug testing.
2226 locktorture.onoff_interval= [KNL]
2227 Set time (s) between CPU-hotplug operations, or
2228 zero to disable CPU-hotplug testing.
2230 locktorture.shuffle_interval= [KNL]
2231 Set task-shuffle interval (jiffies). Shuffling
2232 tasks allows some CPUs to go into dyntick-idle
2233 mode during the locktorture test.
2235 locktorture.shutdown_secs= [KNL]
2236 Set time (s) after boot system shutdown. This
2237 is useful for hands-off automated testing.
2239 locktorture.stat_interval= [KNL]
2240 Time (s) between statistics printk()s.
2242 locktorture.stutter= [KNL]
2243 Time (s) to stutter testing, for example,
2244 specifying five seconds causes the test to run for
2245 five seconds, wait for five seconds, and so on.
2246 This tests the locking primitive's ability to
2247 transition abruptly to and from idle.
2249 locktorture.torture_type= [KNL]
2250 Specify the locking implementation to test.
2252 locktorture.verbose= [KNL]
2253 Enable additional printk() statements.
2255 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2258 loglevel= All Kernel Messages with a loglevel smaller than the
2259 console loglevel will be printed to the console. It can
2260 also be changed with klogd or other programs. The
2261 loglevels are defined as follows:
2263 0 (KERN_EMERG) system is unusable
2264 1 (KERN_ALERT) action must be taken immediately
2265 2 (KERN_CRIT) critical conditions
2266 3 (KERN_ERR) error conditions
2267 4 (KERN_WARNING) warning conditions
2268 5 (KERN_NOTICE) normal but significant condition
2269 6 (KERN_INFO) informational
2270 7 (KERN_DEBUG) debug-level messages
2272 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2273 in bytes. n must be a power of two and greater
2274 than the minimal size. The minimal size is defined
2275 by LOG_BUF_SHIFT kernel config parameter. There is
2276 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2277 that allows to increase the default size depending on
2278 the number of CPUs. See init/Kconfig for more details.
2280 logo.nologo [FB] Disables display of the built-in Linux logo.
2281 This may be used to provide more screen space for
2282 kernel log messages and is useful when debugging
2283 kernel boot problems.
2285 lp=0 [LP] Specify parallel ports to use, e.g,
2286 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2287 lp=reset first parallel port). 'lp=0' disables the
2288 lp=auto printer driver. 'lp=reset' (which can be
2289 specified in addition to the ports) causes
2290 attached printers to be reset. Using
2291 lp=port1,port2,... specifies the parallel ports
2292 to associate lp devices with, starting with
2293 lp0. A port specification may be 'none' to skip
2294 that lp device, or a parport name such as
2295 'parport0'. Specifying 'lp=auto' instead of a
2296 port specification list means that device IDs
2297 from each port should be examined, to see if
2298 an IEEE 1284-compliant printer is attached; if
2299 so, the driver will manage that printer.
2300 See also header of drivers/char/lp.c.
2303 Sets loops_per_jiffy to given constant, thus avoiding
2304 time-consuming boot-time autodetection (up to 250 ms per
2305 CPU). 0 enables autodetection (default). To determine
2306 the correct value for your kernel, boot with normal
2307 autodetection and see what value is printed. Note that
2308 on SMP systems the preset will be applied to all CPUs,
2309 which is likely to cause problems if your CPUs need
2310 significantly divergent settings. An incorrect value
2311 will cause delays in the kernel to be wrong, leading to
2312 unpredictable I/O errors and other breakage. Although
2313 unlikely, in the extreme case this might damage your
2317 Format: <io>,<irq>,<dma>
2319 machvec= [IA-64] Force the use of a particular machine-vector
2320 (machvec) in a generic kernel.
2321 Example: machvec=hpzx1_swiotlb
2323 machtype= [Loongson] Share the same kernel image file between different
2325 Example: machtype=lemote-yeeloong-2f-7inch
2327 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2328 than or equal to this physical address is ignored.
2330 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2331 will bring up during bootup. maxcpus=n : n >= 0 limits
2332 the kernel to bring up 'n' processors. Surely after
2333 bootup you can bring up the other plugged cpu by executing
2334 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2335 only takes effect during system bootup.
2336 While n=0 is a special case, it is equivalent to "nosmp",
2337 which also disables the IO APIC.
2339 max_loop= [LOOP] The number of loop block devices that get
2340 (loop.max_loop) unconditionally pre-created at init time. The default
2341 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2342 of statically allocating a predefined number, loop
2343 devices can be requested on-demand with the
2344 /dev/loop-control interface.
2346 mce [X86-32] Machine Check Exception
2348 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2350 md= [HW] RAID subsystems devices and level
2351 See Documentation/admin-guide/md.rst.
2354 Format: <first>,<last>
2355 Specifies range of consoles to be captured by the MDA.
2358 Control mitigation for the Micro-architectural Data
2359 Sampling (MDS) vulnerability.
2361 Certain CPUs are vulnerable to an exploit against CPU
2362 internal buffers which can forward information to a
2363 disclosure gadget under certain conditions.
2365 In vulnerable processors, the speculatively
2366 forwarded data can be used in a cache side channel
2367 attack, to access data to which the attacker does
2368 not have direct access.
2370 This parameter controls the MDS mitigation. The
2373 full - Enable MDS mitigation on vulnerable CPUs
2374 full,nosmt - Enable MDS mitigation and disable
2375 SMT on vulnerable CPUs
2376 off - Unconditionally disable MDS mitigation
2378 On TAA-affected machines, mds=off can be prevented by
2379 an active TAA mitigation as both vulnerabilities are
2380 mitigated with the same mechanism so in order to disable
2381 this mitigation, you need to specify tsx_async_abort=off
2384 Not specifying this option is equivalent to
2387 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2389 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2390 Amount of memory to be used when the kernel is not able
2391 to see the whole system memory or for test.
2392 [X86] Work as limiting max address. Use together
2393 with memmap= to avoid physical address space collisions.
2394 Without memmap= PCI devices could be placed at addresses
2395 belonging to unused RAM.
2397 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2401 [KNL,SH] Allow user to override the default size for
2402 per-device physically contiguous DMA buffers.
2404 memhp_default_state=online/offline
2405 [KNL] Set the initial state for the memory hotplug
2406 onlining policy. If not specified, the default value is
2407 set according to the
2408 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2410 See Documentation/memory-hotplug.txt.
2412 memmap=exactmap [KNL,X86] Enable setting of an exact
2413 E820 memory map, as specified by the user.
2414 Such memmap=exactmap lines can be constructed based on
2415 BIOS output or other requirements. See the memmap=nn@ss
2418 memmap=nn[KMG]@ss[KMG]
2419 [KNL] Force usage of a specific region of memory.
2420 Region of memory to be used is from ss to ss+nn.
2421 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2422 which limits max address to nn[KMG].
2423 Multiple different regions can be specified,
2426 memmap=100M@2G,100M#3G,1G!1024G
2428 memmap=nn[KMG]#ss[KMG]
2429 [KNL,ACPI] Mark specific memory as ACPI data.
2430 Region of memory to be marked is from ss to ss+nn.
2432 memmap=nn[KMG]$ss[KMG]
2433 [KNL,ACPI] Mark specific memory as reserved.
2434 Region of memory to be reserved is from ss to ss+nn.
2435 Example: Exclude memory from 0x18690000-0x1869ffff
2436 memmap=64K$0x18690000
2438 memmap=0x10000$0x18690000
2439 Some bootloaders may need an escape character before '$',
2440 like Grub2, otherwise '$' and the following number
2443 memmap=nn[KMG]!ss[KMG]
2444 [KNL,X86] Mark specific memory as protected.
2445 Region of memory to be used, from ss to ss+nn.
2446 The memory region may be marked as e820 type 12 (0xc)
2447 and is NVDIMM or ADR memory.
2449 memmap=<size>%<offset>-<oldtype>+<newtype>
2450 [KNL,ACPI] Convert memory within the specified region
2451 from <oldtype> to <newtype>. If "-<oldtype>" is left
2452 out, the whole region will be marked as <newtype>,
2453 even if previously unavailable. If "+<newtype>" is left
2454 out, matching memory will be removed. Types are
2455 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2456 3 = ACPI, 12 = PRAM.
2458 memory_corruption_check=0/1 [X86]
2459 Some BIOSes seem to corrupt the first 64k of
2460 memory when doing things like suspend/resume.
2461 Setting this option will scan the memory
2462 looking for corruption. Enabling this will
2463 both detect corruption and prevent the kernel
2464 from using the memory being corrupted.
2465 However, its intended as a diagnostic tool; if
2466 repeatable BIOS-originated corruption always
2467 affects the same memory, you can use memmap=
2468 to prevent the kernel from using that memory.
2470 memory_corruption_check_size=size [X86]
2471 By default it checks for corruption in the low
2472 64k, making this memory unavailable for normal
2473 use. Use this parameter to scan for
2474 corruption in more or less memory.
2476 memory_corruption_check_period=seconds [X86]
2477 By default it checks for corruption every 60
2478 seconds. Use this parameter to check at some
2479 other rate. 0 disables periodic checking.
2481 memtest= [KNL,X86,ARM] Enable memtest
2483 default : 0 <disable>
2484 Specifies the number of memtest passes to be
2485 performed. Each pass selects another test
2486 pattern from a given set of patterns. Memtest
2487 fills the memory with this pattern, validates
2488 memory contents and reserves bad memory
2489 regions that are detected.
2491 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2492 Valid arguments: on, off
2493 Default (depends on kernel configuration option):
2494 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2495 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2496 mem_encrypt=on: Activate SME
2497 mem_encrypt=off: Do not activate SME
2499 Refer to Documentation/x86/amd-memory-encryption.txt
2500 for details on when memory encryption can be activated.
2502 mem_sleep_default= [SUSPEND] Default system suspend mode:
2503 s2idle - Suspend-To-Idle
2504 shallow - Power-On Suspend or equivalent (if supported)
2505 deep - Suspend-To-RAM or equivalent (if supported)
2506 See Documentation/admin-guide/pm/sleep-states.rst.
2508 meye.*= [HW] Set MotionEye Camera parameters
2509 See Documentation/media/v4l-drivers/meye.rst.
2511 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2512 Multi-Function General Purpose Timers on AMD Geode
2515 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2516 the BIOS has incorrectly applied a workaround. TinyBIOS
2517 version 0.98 is known to be affected, 0.99 fixes the
2518 problem by letting the user disable the workaround.
2522 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2523 physical address is ignored.
2525 mini2440= [ARM,HW,KNL]
2526 Format:[0..2][b][c][t]
2528 MINI2440 configuration specification:
2529 0 - The attached screen is the 3.5" TFT
2530 1 - The attached screen is the 7" TFT
2531 2 - The VGA Shield is attached (1024x768)
2532 Leaving out the screen size parameter will not load
2533 the TFT driver, and the framebuffer will be left
2535 b - Enable backlight. The TFT backlight pin will be
2536 linked to the kernel VESA blanking code and a GPIO
2537 LED. This parameter is not necessary when using the
2539 c - Enable the s3c camera interface.
2540 t - Reserved for enabling touchscreen support. The
2541 touchscreen support is not enabled in the mainstream
2542 kernel as of 2.6.30, a preliminary port can be found
2543 in the "bleeding edge" mini2440 support kernel at
2544 http://repo.or.cz/w/linux-2.6/mini2440.git
2547 [X86,PPC,S390,ARM64] Control optional mitigations for
2548 CPU vulnerabilities. This is a set of curated,
2549 arch-independent options, each of which is an
2550 aggregation of existing arch-specific options.
2553 Disable all optional CPU mitigations. This
2554 improves system performance, but it may also
2555 expose users to several CPU vulnerabilities.
2556 Equivalent to: nopti [X86,PPC]
2561 nospectre_v2 [X86,PPC,S390,ARM64]
2562 spectre_v2_user=off [X86]
2563 spec_store_bypass_disable=off [X86,PPC]
2564 ssbd=force-off [ARM64]
2567 tsx_async_abort=off [X86]
2568 kvm.nx_huge_pages=off [X86]
2569 no_entry_flush [PPC]
2570 no_uaccess_flush [PPC]
2573 This does not have any effect on
2574 kvm.nx_huge_pages when
2575 kvm.nx_huge_pages=force.
2578 Mitigate all CPU vulnerabilities, but leave SMT
2579 enabled, even if it's vulnerable. This is for
2580 users who don't want to be surprised by SMT
2581 getting disabled across kernel upgrades, or who
2582 have other ways of avoiding SMT-based attacks.
2583 Equivalent to: (default behavior)
2586 Mitigate all CPU vulnerabilities, disabling SMT
2587 if needed. This is for users who always want to
2588 be fully mitigated, even if it means losing SMT.
2589 Equivalent to: l1tf=flush,nosmt [X86]
2590 mds=full,nosmt [X86]
2591 tsx_async_abort=full,nosmt [X86]
2594 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2595 parameter allows control of the logging verbosity for
2596 the additional memory initialisation checks. A value
2597 of 0 disables mminit logging and a level of 4 will
2598 log everything. Information is printed at KERN_DEBUG
2599 so loglevel=8 may also need to be specified.
2602 [KNL] When CONFIG_MODULE_SIG is set, this means that
2603 modules without (valid) signatures will fail to load.
2604 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2605 is always true, so this option does nothing.
2607 module_blacklist= [KNL] Do not load a comma-separated list of
2608 modules. Useful for debugging problem modules.
2611 [MOUSE] Maximum time between finger touching and
2612 leaving touchpad surface for touch to be considered
2613 a tap and be reported as a left button click (for
2614 touchpads working in absolute mode only).
2616 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2617 reporting absolute coordinates, such as tablets
2618 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2619 reporting absolute coordinates, such as tablets
2621 movablecore= [KNL,X86,IA-64,PPC]
2622 Format: nn[KMGTPE] | nn%
2623 This parameter is the complement to kernelcore=, it
2624 specifies the amount of memory used for migratable
2625 allocations. If both kernelcore and movablecore is
2626 specified, then kernelcore will be at *least* the
2627 specified value but may be more. If movablecore on its
2628 own is specified, the administrator must be careful
2629 that the amount of memory usable for all allocations
2632 movable_node [KNL] Boot-time switch to make hotplugable memory
2633 NUMA nodes to be movable. This means that the memory
2634 of such nodes will be usable only for movable
2635 allocations which rules out almost all kernel
2636 allocations. Use with caution!
2638 MTD_Partition= [MTD]
2639 Format: <name>,<region-number>,<size>,<offset>
2641 MTD_Region= [MTD] Format:
2642 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2645 See drivers/mtd/cmdlinepart.c.
2647 multitce=off [PPC] This parameter disables the use of the pSeries
2648 firmware feature for updating multiple TCE entries
2651 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2653 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2655 boundary - index of last SLC block on Flex-OneNAND.
2656 The remaining blocks are configured as MLC blocks.
2657 lock - Configure if Flex-OneNAND boundary should be locked.
2658 Once locked, the boundary cannot be changed.
2659 1 indicates lock status, 0 indicates unlock status.
2662 ARM/S3C2412 JIVE boot control
2664 See arch/arm/mach-s3c2412/mach-jive.c
2666 mtouchusb.raw_coordinates=
2667 [HW] Make the MicroTouch USB driver use raw coordinates
2668 ('y', default) or cooked coordinates ('n')
2670 mtrr_chunk_size=nn[KMG] [X86]
2671 used for mtrr cleanup. It is largest continuous chunk
2672 that could hold holes aka. UC entries.
2674 mtrr_gran_size=nn[KMG] [X86]
2675 Used for mtrr cleanup. It is granularity of mtrr block.
2677 Large value could prevent small alignment from
2680 mtrr_spare_reg_nr=n [X86]
2682 Range: 0,7 : spare reg number
2684 Used for mtrr cleanup. It is spare mtrr entries number.
2685 Set to 2 or more if your graphical card needs more.
2687 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2689 netdev= [NET] Network devices parameters
2690 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2691 Note that mem_start is often overloaded to mean
2692 something different and driver-specific.
2693 This usage is only documented in each driver source
2697 [NETFILTER] Enable connection tracking flow accounting
2698 0 to disable accounting
2699 1 to enable accounting
2702 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2703 See Documentation/filesystems/nfs/nfsroot.txt.
2705 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2706 See Documentation/filesystems/nfs/nfsroot.txt.
2708 nfsrootdebug [NFS] enable nfsroot debugging messages.
2709 See Documentation/filesystems/nfs/nfsroot.txt.
2711 nfs.callback_nr_threads=
2712 [NFSv4] set the total number of threads that the
2713 NFS client will assign to service NFSv4 callback
2716 nfs.callback_tcpport=
2717 [NFS] set the TCP port on which the NFSv4 callback
2718 channel should listen.
2721 [NFS] sets the pathname to the program which is used
2722 to update the NFS client cache entries.
2724 nfs.cache_getent_timeout=
2725 [NFS] sets the timeout after which an attempt to
2726 update a cache entry is deemed to have failed.
2728 nfs.idmap_cache_timeout=
2729 [NFS] set the maximum lifetime for idmapper cache
2733 [NFS] enable 64-bit inode numbers.
2734 If zero, the NFS client will fake up a 32-bit inode
2735 number for the readdir() and stat() syscalls instead
2736 of returning the full 64-bit number.
2737 The default is to return 64-bit inode numbers.
2739 nfs.max_session_cb_slots=
2740 [NFSv4.1] Sets the maximum number of session
2741 slots the client will assign to the callback
2742 channel. This determines the maximum number of
2743 callbacks the client will process in parallel for
2744 a particular server.
2746 nfs.max_session_slots=
2747 [NFSv4.1] Sets the maximum number of session slots
2748 the client will attempt to negotiate with the server.
2749 This limits the number of simultaneous RPC requests
2750 that the client can send to the NFSv4.1 server.
2751 Note that there is little point in setting this
2752 value higher than the max_tcp_slot_table_limit.
2754 nfs.nfs4_disable_idmapping=
2755 [NFSv4] When set to the default of '1', this option
2756 ensures that both the RPC level authentication
2757 scheme and the NFS level operations agree to use
2758 numeric uids/gids if the mount is using the
2759 'sec=sys' security flavour. In effect it is
2760 disabling idmapping, which can make migration from
2761 legacy NFSv2/v3 systems to NFSv4 easier.
2762 Servers that do not support this mode of operation
2763 will be autodetected by the client, and it will fall
2764 back to using the idmapper.
2765 To turn off this behaviour, set the value to '0'.
2767 [NFS4] Specify an additional fixed unique ident-
2768 ification string that NFSv4 clients can insert into
2769 their nfs_client_id4 string. This is typically a
2770 UUID that is generated at system install time.
2772 nfs.send_implementation_id =
2773 [NFSv4.1] Send client implementation identification
2774 information in exchange_id requests.
2775 If zero, no implementation identification information
2777 The default is to send the implementation identification
2780 nfs.recover_lost_locks =
2781 [NFSv4] Attempt to recover locks that were lost due
2782 to a lease timeout on the server. Please note that
2783 doing this risks data corruption, since there are
2784 no guarantees that the file will remain unchanged
2785 after the locks are lost.
2786 If you want to enable the kernel legacy behaviour of
2787 attempting to recover these locks, then set this
2789 The default parameter value of '0' causes the kernel
2790 not to attempt recovery of lost locks.
2792 nfs4.layoutstats_timer =
2793 [NFSv4.2] Change the rate at which the kernel sends
2794 layoutstats to the pNFS metadata server.
2796 Setting this to value to 0 causes the kernel to use
2797 whatever value is the default set by the layout
2798 driver. A non-zero value sets the minimum interval
2799 in seconds between layoutstats transmissions.
2801 nfsd.nfs4_disable_idmapping=
2802 [NFSv4] When set to the default of '1', the NFSv4
2803 server will return only numeric uids and gids to
2804 clients using auth_sys, and will accept numeric uids
2805 and gids from such clients. This is intended to ease
2806 migration from NFSv2/v3.
2808 nmi_debug= [KNL,SH] Specify one or more actions to take
2809 when a NMI is triggered.
2810 Format: [state][,regs][,debounce][,die]
2812 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2813 Format: [panic,][nopanic,][num]
2815 0 - turn hardlockup detector in nmi_watchdog off
2816 1 - turn hardlockup detector in nmi_watchdog on
2817 When panic is specified, panic when an NMI watchdog
2818 timeout occurs (or 'nopanic' to override the opposite
2819 default). To disable both hard and soft lockup detectors,
2820 please see 'nowatchdog'.
2821 This is useful when you use a panic=... timeout and
2822 need the box quickly up again.
2824 These settings can be accessed at runtime via
2825 the nmi_watchdog and hardlockup_panic sysctls.
2827 netpoll.carrier_timeout=
2828 [NET] Specifies amount of time (in seconds) that
2829 netpoll should wait for a carrier. By default netpoll
2832 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2833 emulation library even if a 387 maths coprocessor
2836 no5lvl [X86-64] Disable 5-level paging mode. Forces
2837 kernel to use 4-level paging instead.
2840 [HW] Never suspend the console
2841 Disable suspending of consoles during suspend and
2842 hibernate operations. Once disabled, debugging
2843 messages can reach various consoles while the rest
2844 of the system is being put to sleep (ie, while
2845 debugging driver suspend/resume hooks). This may
2846 not work reliably with all consoles, but is known
2847 to work with serial and VGA consoles.
2848 To facilitate more flexible debugging, we also add
2849 console_suspend, a printk module parameter to control
2850 it. Users could use console_suspend (usually
2851 /sys/module/printk/parameters/console_suspend) to
2852 turn on/off it dynamically.
2854 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2855 caches in the slab allocator. Saves per-node memory,
2856 but will impact performance.
2860 noaltinstr [S390] Disables alternative instructions patching
2861 (CPU alternatives feature).
2863 noapic [SMP,APIC] Tells the kernel to not make use of any
2864 IOAPICs that may be present in the system.
2866 noautogroup Disable scheduler automatic task group creation.
2868 nobats [PPC] Do not use BATs for mapping kernel lowmem
2869 on "Classic" PPC cores.
2873 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2875 nodelayacct [KNL] Disable per-task delay accounting
2877 nodsp [SH] Disable hardware DSP at boot time.
2879 noefi Disable EFI runtime services support.
2881 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
2886 On X86-32 available only on PAE configured kernels.
2887 noexec=on: enable non-executable mappings (default)
2888 noexec=off: disable non-executable mappings
2891 Disable SMAP (Supervisor Mode Access Prevention)
2892 even if it is supported by processor.
2895 Disable SMEP (Supervisor Mode Execution Prevention)
2896 even if it is supported by processor.
2899 This affects only 32-bit executables.
2900 noexec32=on: enable non-executable mappings (default)
2901 read doesn't imply executable mappings
2902 noexec32=off: disable non-executable mappings
2903 read implies executable mappings
2905 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2907 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2908 register save and restore. The kernel will only save
2909 legacy floating-point registers on task switch.
2911 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2913 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2914 Equivalent to smt=1.
2916 [KNL,x86] Disable symmetric multithreading (SMT).
2917 nosmt=force: Force disable SMT, cannot be undone
2918 via the sysfs control file.
2920 nospectre_v1 [X66, PPC] Disable mitigations for Spectre Variant 1
2921 (bounds check bypass). With this option data leaks
2922 are possible in the system.
2924 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
2925 the Spectre variant 2 (indirect branch prediction)
2926 vulnerability. System may allow data leaks with this
2929 nospec_store_bypass_disable
2930 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
2933 [PPC] Don't flush the L1-D cache after accessing user data.
2935 noxsave [BUGS=X86] Disables x86 extended register state save
2936 and restore using xsave. The kernel will fallback to
2937 enabling legacy floating-point and sse state.
2939 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2940 register states. The kernel will fall back to use
2941 xsave to save the states. By using this parameter,
2942 performance of saving the states is degraded because
2943 xsave doesn't support modified optimization while
2944 xsaveopt supports it on xsaveopt enabled systems.
2946 noxsaves [X86] Disables xsaves and xrstors used in saving and
2947 restoring x86 extended register state in compacted
2948 form of xsave area. The kernel will fall back to use
2949 xsaveopt and xrstor to save and restore the states
2950 in standard form of xsave area. By using this
2951 parameter, xsave area per process might occupy more
2952 memory on xsaves enabled systems.
2954 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2955 wfi(ARM) instruction doesn't work correctly and not to
2956 use it. This is also useful when using JTAG debugger.
2958 no_file_caps Tells the kernel not to honor file capabilities. The
2959 only way then for a file to be executed with privilege
2960 is to be setuid root or executed by root.
2962 nohalt [IA-64] Tells the kernel not to use the power saving
2963 function PAL_HALT_LIGHT when idle. This increases
2964 power-consumption. On the positive side, it reduces
2965 interrupt wake-up latency, which may improve performance
2966 in certain environments such as networked servers or
2969 nohibernate [HIBERNATION] Disable hibernation and resume.
2971 nohz= [KNL] Boottime enable/disable dynamic ticks
2972 Valid arguments: on, off
2975 nohz_full= [KNL,BOOT,SMP,ISOL]
2976 The argument is a cpu list, as described above.
2977 In kernels built with CONFIG_NO_HZ_FULL=y, set
2978 the specified list of CPUs whose tick will be stopped
2979 whenever possible. The boot CPU will be forced outside
2980 the range to maintain the timekeeping. Any CPUs
2981 in this list will have their RCU callbacks offloaded,
2982 just as if they had also been called out in the
2983 rcu_nocbs= boot parameter.
2985 noiotrap [SH] Disables trapped I/O port accesses.
2987 noirqdebug [X86-32] Disables the code which attempts to detect and
2988 disable unhandled interrupt sources.
2990 no_timer_check [X86,APIC] Disables the code which tests for
2991 broken timer IRQ sources.
2993 noisapnp [ISAPNP] Disables ISA PnP code.
2995 noinitrd [RAM] Tells the kernel not to load any configured
2998 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3000 [Deprecated - use intremap=off]
3004 noinvpcid [X86] Disable the INVPCID cpu feature.
3006 nojitter [IA-64] Disables jitter checking for ITC timers.
3008 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3010 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3014 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3015 clock and use the default one.
3017 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
3018 steal time is computed, but won't influence scheduler
3021 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3023 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3025 noltlbs [PPC] Do not use large page/tlb entries for kernel
3026 lowmem mapping on PPC40x and PPC8xx
3028 nomca [IA-64] Disable machine check abort handling
3030 nomce [X86-32] Disable Machine Check Exception
3032 nomfgpt [X86-32] Disable Multi-Function General Purpose
3033 Timer usage (for AMD Geode machines).
3035 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3036 shutdown the other cpus. Instead use the REBOOT_VECTOR
3039 nomodule Disable module load
3041 nopat [X86] Disable PAT (page attribute table extension of
3042 pagetables) support.
3044 nopcid [X86-64] Disable the PCID cpu feature.
3046 norandmaps Don't use address space randomization. Equivalent to
3047 echo 0 > /proc/sys/kernel/randomize_va_space
3049 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3050 with UP alternatives
3052 nordrand [X86] Disable kernel use of the RDRAND and
3053 RDSEED instructions even if they are supported
3054 by the processor. RDRAND and RDSEED are still
3055 available to user space applications.
3057 noresume [SWSUSP] Disables resume and restores original swap
3060 no-scroll [VGA] Disables scrollback.
3061 This is required for the Braillex ib80-piezo Braille
3062 reader made by F.H. Papenmeier (Germany).
3066 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3068 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3069 and disable the IO APIC. legacy for "maxcpus=0".
3071 nosoftlockup [KNL] Disable the soft-lockup detector.
3073 nosync [HW,M68K] Disables sync negotiation for all devices.
3075 nowatchdog [KNL] Disable both lockup detectors, i.e.
3076 soft-lockup and NMI watchdog (hard-lockup).
3080 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3082 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3083 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3084 Some features depend on CPU0. Known dependencies are:
3085 1. Resume from suspend/hibernate depends on CPU0.
3086 Suspend/hibernate will fail if CPU0 is offline and you
3087 need to online CPU0 before suspend/hibernate.
3088 2. PIC interrupts also depend on CPU0. CPU0 can't be
3089 removed if a PIC interrupt is detected.
3090 It's said poweroff/reboot may depend on CPU0 on some
3091 machines although I haven't seen such issues so far
3092 after CPU0 is offline on a few tested machines.
3093 If the dependencies are under your control, you can
3094 turn on cpu0_hotplug.
3096 nps_mtm_hs_ctr= [KNL,ARC]
3097 This parameter sets the maximum duration, in
3098 cycles, each HW thread of the CTOP can run
3099 without interruptions, before HW switches it.
3100 The actual maximum duration is 16 times this
3102 Format: integer between 1 and 255
3105 nptcg= [IA-64] Override max number of concurrent global TLB
3106 purges which is reported from either PAL_VM_SUMMARY or
3109 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3110 could support. nr_cpus=n : n >= 1 limits the kernel to
3111 support 'n' processors. It could be larger than the
3112 number of already plugged CPU during bootup, later in
3113 runtime you can physically add extra cpu until it reaches
3114 n. So during boot up some boot time memory for per-cpu
3115 variables need be pre-allocated for later physical cpu
3118 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3120 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3121 Allowed values are enable and disable
3123 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3124 'node', 'default' can be specified
3125 This can be set from sysctl after boot.
3126 See Documentation/sysctl/vm.txt for details.
3128 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3129 See Documentation/debugging-via-ohci1394.txt for more
3132 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3133 Rather than timing out after 20 ms if an EC
3134 command is not properly ACKed, override the length
3135 of the timeout. We have interrupts disabled while
3136 waiting for the ACK, so if this is set too high
3137 interrupts *may* be lost!
3139 omap_mux= [OMAP] Override bootloader pin multiplexing.
3140 Format: <mux_mode0.mode_name=value>...
3141 For example, to override I2C bus2:
3142 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3144 oprofile.timer= [HW]
3145 Use timer interrupt instead of performance counters
3147 oprofile.cpu_type= Force an oprofile cpu type
3148 This might be useful if you have an older oprofile
3149 userland or if you want common events.
3150 Format: { arch_perfmon }
3151 arch_perfmon: [X86] Force use of architectural
3152 perfmon on Intel CPUs instead of the
3153 CPU specific event set.
3154 timer: [X86] Force use of architectural NMI
3155 timer mode (see also oprofile.timer
3156 for generic hr timer mode)
3158 oops=panic Always panic on oopses. Default is to just kill the
3159 process, but there is a small probability of
3160 deadlocking the machine.
3161 This will also cause panics on machine check exceptions.
3162 Useful together with panic=30 to trigger a reboot.
3164 page_owner= [KNL] Boot-time page_owner enabling option.
3165 Storage of the information about who allocated
3166 each page is disabled in default. With this switch,
3168 on: enable the feature
3170 page_poison= [KNL] Boot-time parameter changing the state of
3171 poisoning on the buddy allocator, available with
3172 CONFIG_PAGE_POISONING=y.
3173 off: turn off poisoning (default)
3174 on: turn on poisoning
3176 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3177 timeout > 0: seconds before rebooting
3178 timeout = 0: wait forever
3179 timeout < 0: reboot immediately
3182 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3185 crash_kexec_post_notifiers
3186 Run kdump after running panic-notifiers and dumping
3187 kmsg. This only for the users who doubt kdump always
3188 succeeds in any situation.
3189 Note that this also increases risks of kdump failure,
3190 because some panic notifiers can make the crashed
3191 kernel more unstable.
3193 parkbd.port= [HW] Parallel port number the keyboard adapter is
3194 connected to, default is 0.
3196 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3197 0 for XT, 1 for AT (default is AT).
3200 parport= [HW,PPT] Specify parallel ports. 0 disables.
3201 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3202 Use 'auto' to force the driver to use any
3203 IRQ/DMA settings detected (the default is to
3204 ignore detected IRQ/DMA settings because of
3205 possible conflicts). You can specify the base
3206 address, IRQ, and DMA settings; IRQ and DMA
3207 should be numbers, or 'auto' (for using detected
3208 settings on that particular port), or 'nofifo'
3209 (to avoid using a FIFO even if it is detected).
3210 Parallel ports are assigned in the order they
3211 are specified on the command line, starting
3214 parport_init_mode= [HW,PPT]
3215 Configure VIA parallel port to operate in
3216 a specific mode. This is necessary on Pegasos
3217 computer where firmware has no options for setting
3218 up parallel port mode and sets it to spp.
3219 Currently this function knows 686a and 8231 chips.
3220 Format: [spp|ps2|epp|ecp|ecpepp]
3223 Halt all CPUs after the first oops has been printed for
3224 the specified number of seconds. This is to be used if
3225 your oopses keep scrolling off the screen.
3230 See header of drivers/block/paride/pcd.c.
3231 See also Documentation/blockdev/paride.txt.
3233 pci=option[,option...] [PCI] various PCI subsystem options.
3235 Some options herein operate on a specific device
3236 or a set of devices (<pci_dev>). These are
3237 specified in one of the following formats:
3239 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3240 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3242 Note: the first format specifies a PCI
3243 bus/device/function address which may change
3244 if new hardware is inserted, if motherboard
3245 firmware changes, or due to changes caused
3246 by other kernel parameters. If the
3247 domain is left unspecified, it is
3248 taken to be zero. Optionally, a path
3249 to a device through multiple device/function
3250 addresses can be specified after the base
3251 address (this is more robust against
3252 renumbering issues). The second format
3253 selects devices using IDs from the
3254 configuration space which may match multiple
3255 devices in the system.
3257 earlydump dump PCI config space before the kernel
3259 off [X86] don't probe for the PCI bus
3260 bios [X86-32] force use of PCI BIOS, don't access
3261 the hardware directly. Use this if your machine
3262 has a non-standard PCI host bridge.
3263 nobios [X86-32] disallow use of PCI BIOS, only direct
3264 hardware access methods are allowed. Use this
3265 if you experience crashes upon bootup and you
3266 suspect they are caused by the BIOS.
3267 conf1 [X86] Force use of PCI Configuration Access
3268 Mechanism 1 (config address in IO port 0xCF8,
3269 data in IO port 0xCFC, both 32-bit).
3270 conf2 [X86] Force use of PCI Configuration Access
3271 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3272 the function, IO port 0xCFA, also 8-bit, sets
3273 bus number. The config space is then accessed
3274 through ports 0xC000-0xCFFF).
3275 See http://wiki.osdev.org/PCI for more info
3276 on the configuration access mechanisms.
3277 noaer [PCIE] If the PCIEAER kernel config parameter is
3278 enabled, this kernel boot option can be used to
3279 disable the use of PCIE advanced error reporting.
3280 nodomains [PCI] Disable support for multiple PCI
3281 root domains (aka PCI segments, in ACPI-speak).
3282 nommconf [X86] Disable use of MMCONFIG for PCI
3284 check_enable_amd_mmconf [X86] check for and enable
3285 properly configured MMIO access to PCI
3286 config space on AMD family 10h CPU
3287 nomsi [MSI] If the PCI_MSI kernel config parameter is
3288 enabled, this kernel boot option can be used to
3289 disable the use of MSI interrupts system-wide.
3290 noioapicquirk [APIC] Disable all boot interrupt quirks.
3291 Safety option to keep boot IRQs enabled. This
3292 should never be necessary.
3293 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3294 primary IO-APIC for bridges that cannot disable
3295 boot IRQs. This fixes a source of spurious IRQs
3296 when the system masks IRQs.
3297 noioapicreroute [APIC] Disable workaround that uses the
3298 boot IRQ equivalent of an IRQ that connects to
3299 a chipset where boot IRQs cannot be disabled.
3300 The opposite of ioapicreroute.
3301 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3302 routing table. These calls are known to be buggy
3303 on several machines and they hang the machine
3304 when used, but on other computers it's the only
3305 way to get the interrupt routing table. Try
3306 this option if the kernel is unable to allocate
3307 IRQs or discover secondary PCI buses on your
3309 rom [X86] Assign address space to expansion ROMs.
3310 Use with caution as certain devices share
3311 address decoders between ROMs and other
3313 norom [X86] Do not assign address space to
3314 expansion ROMs that do not already have
3315 BIOS assigned address ranges.
3316 nobar [X86] Do not assign address space to the
3317 BARs that weren't assigned by the BIOS.
3318 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3319 assigned automatically to PCI devices. You can
3320 make the kernel exclude IRQs of your ISA cards
3322 pirqaddr=0xAAAAA [X86] Specify the physical address
3323 of the PIRQ table (normally generated
3324 by the BIOS) if it is outside the
3325 F0000h-100000h range.
3326 lastbus=N [X86] Scan all buses thru bus #N. Can be
3327 useful if the kernel is unable to find your
3328 secondary buses and you want to tell it
3329 explicitly which ones they are.
3330 assign-busses [X86] Always assign all PCI bus
3331 numbers ourselves, overriding
3332 whatever the firmware may have done.
3333 usepirqmask [X86] Honor the possible IRQ mask stored
3334 in the BIOS $PIR table. This is needed on
3335 some systems with broken BIOSes, notably
3336 some HP Pavilion N5400 and Omnibook XE3
3337 notebooks. This will have no effect if ACPI
3338 IRQ routing is enabled.
3339 noacpi [X86] Do not use ACPI for IRQ routing
3340 or for PCI scanning.
3341 use_crs [X86] Use PCI host bridge window information
3342 from ACPI. On BIOSes from 2008 or later, this
3343 is enabled by default. If you need to use this,
3344 please report a bug.
3345 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3346 If you need to use this, please report a bug.
3347 routeirq Do IRQ routing for all PCI devices.
3348 This is normally done in pci_enable_device(),
3349 so this option is a temporary workaround
3350 for broken drivers that don't call it.
3351 skip_isa_align [X86] do not align io start addr, so can
3352 handle more pci cards
3353 noearly [X86] Don't do any early type 1 scanning.
3354 This might help on some broken boards which
3355 machine check when some devices' config space
3356 is read. But various workarounds are disabled
3357 and some IOMMU drivers will not work.
3358 bfsort Sort PCI devices into breadth-first order.
3359 This sorting is done to get a device
3360 order compatible with older (<= 2.4) kernels.
3361 nobfsort Don't sort PCI devices into breadth-first order.
3362 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3363 tuning and use the BIOS-configured MPS defaults.
3364 pcie_bus_safe Set every device's MPS to the largest value
3365 supported by all devices below the root complex.
3366 pcie_bus_perf Set device MPS to the largest allowable MPS
3367 based on its parent bus. Also set MRRS (Max
3368 Read Request Size) to the largest supported
3369 value (no larger than the MPS that the device
3370 or bus can support) for best performance.
3371 pcie_bus_peer2peer Set every device's MPS to 128B, which
3372 every device is guaranteed to support. This
3373 configuration allows peer-to-peer DMA between
3374 any pair of devices, possibly at the cost of
3375 reduced performance. This also guarantees
3376 that hot-added devices will work.
3377 cbiosize=nn[KMG] The fixed amount of bus space which is
3378 reserved for the CardBus bridge's IO window.
3379 The default value is 256 bytes.
3380 cbmemsize=nn[KMG] The fixed amount of bus space which is
3381 reserved for the CardBus bridge's memory
3382 window. The default value is 64 megabytes.
3385 [<order of align>@]<pci_dev>[; ...]
3386 Specifies alignment and device to reassign
3387 aligned memory resources. How to
3388 specify the device is described above.
3389 If <order of align> is not specified,
3390 PAGE_SIZE is used as alignment.
3391 PCI-PCI bridge can be specified, if resource
3392 windows need to be expanded.
3393 To specify the alignment for several
3394 instances of a device, the PCI vendor,
3395 device, subvendor, and subdevice may be
3396 specified, e.g., 4096@pci:8086:9c22:103c:198f
3397 ecrc= Enable/disable PCIe ECRC (transaction layer
3398 end-to-end CRC checking).
3399 bios: Use BIOS/firmware settings. This is the
3403 hpiosize=nn[KMG] The fixed amount of bus space which is
3404 reserved for hotplug bridge's IO window.
3405 Default size is 256 bytes.
3406 hpmemsize=nn[KMG] The fixed amount of bus space which is
3407 reserved for hotplug bridge's memory window.
3408 Default size is 2 megabytes.
3409 hpbussize=nn The minimum amount of additional bus numbers
3410 reserved for buses below a hotplug bridge.
3412 realloc= Enable/disable reallocating PCI bridge resources
3413 if allocations done by BIOS are too small to
3414 accommodate resources required by all child
3416 off: Turn realloc off
3418 realloc same as realloc=on
3419 noari do not use PCIe ARI.
3420 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3421 do not use PCIe ATS (and IOMMU device IOTLB).
3422 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3423 only look for one device below a PCIe downstream
3425 big_root_window Try to add a big 64bit memory window to the PCIe
3426 root complex on AMD CPUs. Some GFX hardware
3427 can resize a BAR to allow access to all VRAM.
3428 Adding the window is slightly risky (it may
3429 conflict with unreported devices), so this
3431 disable_acs_redir=<pci_dev>[; ...]
3432 Specify one or more PCI devices (in the format
3433 specified above) separated by semicolons.
3434 Each device specified will have the PCI ACS
3435 redirect capabilities forced off which will
3436 allow P2P traffic between devices through
3437 bridges without forcing it upstream. Note:
3438 this removes isolation between devices and
3439 may put more devices in an IOMMU group.
3441 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3444 force Enable ASPM even on devices that claim not to support it.
3445 WARNING: Forcing ASPM on may cause system lockups.
3447 pcie_ports= [PCIE] PCIe port services handling:
3448 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3449 even if the platform doesn't give the OS permission to
3450 use them. This may cause conflicts if the platform
3451 also tries to use these services.
3452 compat Disable native PCIe services (PME, AER, DPC, PCIe
3455 pcie_port_pm= [PCIE] PCIe port power management handling:
3456 off Disable power management of all PCIe ports
3457 force Forcibly enable power management of all PCIe ports
3459 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3460 nomsi Do not use MSI for native PCIe PME signaling (this makes
3461 all PCIe root ports use INTx for all services).
3463 pcmv= [HW,PCMCIA] BadgePAD 4
3467 Keep all power-domains already enabled by bootloader on,
3468 even if no driver has claimed them. This is useful
3469 for debug and development, but should not be
3470 needed on a platform with proper driver support.
3473 See Documentation/blockdev/paride.txt.
3475 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3478 See arch/parisc/kernel/pdc_chassis.c
3480 percpu_alloc= Select which percpu first chunk allocator to use.
3481 Currently supported values are "embed" and "page".
3482 Archs may support subset or none of the selections.
3483 See comments in mm/percpu.c for details on each
3484 allocator. This parameter is primarily for debugging
3485 and performance comparison.
3488 See Documentation/blockdev/paride.txt.
3491 See Documentation/blockdev/paride.txt.
3493 pirq= [SMP,APIC] Manual mp-table setup
3494 See Documentation/x86/i386/IO-APIC.txt.
3496 plip= [PPT,NET] Parallel port network link
3497 Format: { parport<nr> | timid | 0 }
3498 See also Documentation/admin-guide/parport.rst.
3500 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3501 Override pmtimer IOPort with a hex value.
3505 Enable PNP debug messages (depends on the
3506 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3507 via /sys/module/pnp/parameters/debug. We always show
3508 current resource usage; turning this on also shows
3509 possible settings and some assignment information.
3515 { on | off | curr | res | no-curr | no-res }
3518 [ISAPNP] Exclude IRQs for the autoconfiguration
3521 [ISAPNP] Exclude DMAs for the autoconfiguration
3523 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3524 Ranges are in pairs (I/O port base and size).
3527 [ISAPNP] Exclude memory regions for the
3529 Ranges are in pairs (memory base and size).
3531 ports= [IP_VS_FTP] IPVS ftp helper module
3533 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3535 Format: <port>,<port>....
3537 powersave=off [PPC] This option disables power saving features.
3538 It specifically disables cpuidle and sets the
3539 platform machine description specific power_save
3540 function to NULL. On Idle the CPU just reduces
3543 ppc_strict_facility_enable
3544 [PPC] This option catches any kernel floating point,
3545 Altivec, VSX and SPE outside of regions specifically
3546 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3547 There is some performance impact when enabling this.
3551 Disable Hardware Transactional Memory
3553 print-fatal-signals=
3554 [KNL] debug: print fatal signals
3556 If enabled, warn about various signal handling
3557 related application anomalies: too many signals,
3558 too many POSIX.1 timers, fatal signals causing a
3561 If you hit the warning due to signal overflow,
3562 you might want to try "ulimit -i unlimited".
3566 printk.always_kmsg_dump=
3567 Trigger kmsg_dump for cases other than kernel oops or
3569 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3572 printk.devkmsg={on,off,ratelimit}
3573 Control writing to /dev/kmsg.
3574 on - unlimited logging to /dev/kmsg from userspace
3575 off - logging to /dev/kmsg disabled
3576 ratelimit - ratelimit the logging
3579 printk.time= Show timing data prefixed to each printk message line
3580 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3582 processor.max_cstate= [HW,ACPI]
3583 Limit processor to maximum C-state
3584 max_cstate=9 overrides any DMI blacklist limit.
3586 processor.nocst [HW,ACPI]
3587 Ignore the _CST method to determine C-states,
3588 instead using the legacy FADT method
3590 profile= [KNL] Enable kernel profiling via /proc/profile
3591 Format: [<profiletype>,]<number>
3592 Param: <profiletype>: "schedule", "sleep", or "kvm"
3593 [defaults to kernel profiling]
3594 Param: "schedule" - profile schedule points.
3595 Param: "sleep" - profile D-state sleeping (millisecs).
3596 Requires CONFIG_SCHEDSTATS
3597 Param: "kvm" - profile VM exits.
3598 Param: <number> - step/bucket size as a power of 2 for
3599 statistical time based profiling.
3601 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3603 See Documentation/blockdev/ramdisk.txt.
3605 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3606 probe for; one of (bare|imps|exps|lifebook|any).
3607 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3609 psmouse.resetafter= [HW,MOUSE]
3610 Try to reset the device after so many bad packets
3613 [HW,MOUSE] Set desired mouse resolution, in dpi.
3614 psmouse.smartscroll=
3615 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3616 0 = disabled, 1 = enabled (default).
3618 pstore.backend= Specify the name of the pstore backend to use
3621 See Documentation/blockdev/paride.txt.
3623 pti= [X86_64] Control Page Table Isolation of user and
3624 kernel address spaces. Disabling this feature
3625 removes hardening, but improves performance of
3626 system calls and interrupts.
3628 on - unconditionally enable
3629 off - unconditionally disable
3630 auto - kernel detects whether your CPU model is
3631 vulnerable to issues that PTI mitigates
3633 Not specifying this option is equivalent to pti=auto.
3636 Equivalent to pti=off
3639 [KNL] Number of legacy pty's. Overwrites compiled-in
3642 quiet [KNL] Disable most log messages
3647 See Documentation/admin-guide/md.rst.
3649 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3650 See Documentation/blockdev/ramdisk.txt.
3652 random.trust_cpu={on,off}
3653 [KNL] Enable or disable trusting the use of the
3654 CPU's random number generator (if available) to
3655 fully seed the kernel's CRNG. Default is controlled
3656 by CONFIG_RANDOM_TRUST_CPU.
3658 ras=option[,option,...] [KNL] RAS-specific options
3661 Disable the Correctable Errors Collector,
3662 see CONFIG_RAS_CEC help text.
3665 The argument is a cpu list, as described above.
3667 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3668 the specified list of CPUs to be no-callback CPUs.
3669 Invocation of these CPUs' RCU callbacks will
3670 be offloaded to "rcuox/N" kthreads created for
3671 that purpose, where "x" is "b" for RCU-bh, "p"
3672 for RCU-preempt, and "s" for RCU-sched, and "N"
3673 is the CPU number. This reduces OS jitter on the
3674 offloaded CPUs, which can be useful for HPC and
3675 real-time workloads. It can also improve energy
3676 efficiency for asymmetric multiprocessors.
3679 Rather than requiring that offloaded CPUs
3680 (specified by rcu_nocbs= above) explicitly
3681 awaken the corresponding "rcuoN" kthreads,
3682 make these kthreads poll for callbacks.
3683 This improves the real-time response for the
3684 offloaded CPUs by relieving them of the need to
3685 wake up the corresponding kthread, but degrades
3686 energy efficiency by requiring that the kthreads
3687 periodically wake up to do the polling.
3689 rcutree.blimit= [KNL]
3690 Set maximum number of finished RCU callbacks to
3691 process in one batch.
3693 rcutree.dump_tree= [KNL]
3694 Dump the structure of the rcu_node combining tree
3695 out at early boot. This is used for diagnostic
3696 purposes, to verify correct tree setup.
3698 rcutree.gp_cleanup_delay= [KNL]
3699 Set the number of jiffies to delay each step of
3700 RCU grace-period cleanup.
3702 rcutree.gp_init_delay= [KNL]
3703 Set the number of jiffies to delay each step of
3704 RCU grace-period initialization.
3706 rcutree.gp_preinit_delay= [KNL]
3707 Set the number of jiffies to delay each step of
3708 RCU grace-period pre-initialization, that is,
3709 the propagation of recent CPU-hotplug changes up
3710 the rcu_node combining tree.
3712 rcutree.rcu_fanout_exact= [KNL]
3713 Disable autobalancing of the rcu_node combining
3714 tree. This is used by rcutorture, and might
3715 possibly be useful for architectures having high
3716 cache-to-cache transfer latencies.
3718 rcutree.rcu_fanout_leaf= [KNL]
3719 Change the number of CPUs assigned to each
3720 leaf rcu_node structure. Useful for very
3721 large systems, which will choose the value 64,
3722 and for NUMA systems with large remote-access
3723 latencies, which will choose a value aligned
3724 with the appropriate hardware boundaries.
3726 rcutree.jiffies_till_sched_qs= [KNL]
3727 Set required age in jiffies for a
3728 given grace period before RCU starts
3729 soliciting quiescent-state help from
3730 rcu_note_context_switch().
3732 rcutree.jiffies_till_first_fqs= [KNL]
3733 Set delay from grace-period initialization to
3734 first attempt to force quiescent states.
3735 Units are jiffies, minimum value is zero,
3736 and maximum value is HZ.
3738 rcutree.jiffies_till_next_fqs= [KNL]
3739 Set delay between subsequent attempts to force
3740 quiescent states. Units are jiffies, minimum
3741 value is one, and maximum value is HZ.
3743 rcutree.kthread_prio= [KNL,BOOT]
3744 Set the SCHED_FIFO priority of the RCU per-CPU
3745 kthreads (rcuc/N). This value is also used for
3746 the priority of the RCU boost threads (rcub/N)
3747 and for the RCU grace-period kthreads (rcu_bh,
3748 rcu_preempt, and rcu_sched). If RCU_BOOST is
3749 set, valid values are 1-99 and the default is 1
3750 (the least-favored priority). Otherwise, when
3751 RCU_BOOST is not set, valid values are 0-99 and
3752 the default is zero (non-realtime operation).
3754 rcutree.rcu_nocb_leader_stride= [KNL]
3755 Set the number of NOCB kthread groups, which
3756 defaults to the square root of the number of
3757 CPUs. Larger numbers reduces the wakeup overhead
3758 on the per-CPU grace-period kthreads, but increases
3759 that same overhead on each group's leader.
3761 rcutree.qhimark= [KNL]
3762 Set threshold of queued RCU callbacks beyond which
3763 batch limiting is disabled.
3765 rcutree.qlowmark= [KNL]
3766 Set threshold of queued RCU callbacks below which
3767 batch limiting is re-enabled.
3769 rcutree.rcu_idle_gp_delay= [KNL]
3770 Set wakeup interval for idle CPUs that have
3771 RCU callbacks (RCU_FAST_NO_HZ=y).
3773 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3774 Set wakeup interval for idle CPUs that have
3775 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3776 Lazy RCU callbacks are those which RCU can
3777 prove do nothing more than free memory.
3779 rcutree.rcu_kick_kthreads= [KNL]
3780 Cause the grace-period kthread to get an extra
3781 wake_up() if it sleeps three times longer than
3782 it should at force-quiescent-state time.
3783 This wake_up() will be accompanied by a
3784 WARN_ONCE() splat and an ftrace_dump().
3786 rcuperf.gp_async= [KNL]
3787 Measure performance of asynchronous
3788 grace-period primitives such as call_rcu().
3790 rcuperf.gp_async_max= [KNL]
3791 Specify the maximum number of outstanding
3792 callbacks per writer thread. When a writer
3793 thread exceeds this limit, it invokes the
3794 corresponding flavor of rcu_barrier() to allow
3795 previously posted callbacks to drain.
3797 rcuperf.gp_exp= [KNL]
3798 Measure performance of expedited synchronous
3799 grace-period primitives.
3801 rcuperf.holdoff= [KNL]
3802 Set test-start holdoff period. The purpose of
3803 this parameter is to delay the start of the
3804 test until boot completes in order to avoid
3807 rcuperf.nreaders= [KNL]
3808 Set number of RCU readers. The value -1 selects
3809 N, where N is the number of CPUs. A value
3810 "n" less than -1 selects N-n+1, where N is again
3811 the number of CPUs. For example, -2 selects N
3812 (the number of CPUs), -3 selects N+1, and so on.
3813 A value of "n" less than or equal to -N selects
3816 rcuperf.nwriters= [KNL]
3817 Set number of RCU writers. The values operate
3818 the same as for rcuperf.nreaders.
3819 N, where N is the number of CPUs
3821 rcuperf.perf_type= [KNL]
3822 Specify the RCU implementation to test.
3824 rcuperf.shutdown= [KNL]
3825 Shut the system down after performance tests
3826 complete. This is useful for hands-off automated
3829 rcuperf.verbose= [KNL]
3830 Enable additional printk() statements.
3832 rcuperf.writer_holdoff= [KNL]
3833 Write-side holdoff between grace periods,
3834 in microseconds. The default of zero says
3837 rcutorture.cbflood_inter_holdoff= [KNL]
3838 Set holdoff time (jiffies) between successive
3839 callback-flood tests.
3841 rcutorture.cbflood_intra_holdoff= [KNL]
3842 Set holdoff time (jiffies) between successive
3843 bursts of callbacks within a given callback-flood
3846 rcutorture.cbflood_n_burst= [KNL]
3847 Set the number of bursts making up a given
3848 callback-flood test. Set this to zero to
3849 disable callback-flood testing.
3851 rcutorture.cbflood_n_per_burst= [KNL]
3852 Set the number of callbacks to be registered
3853 in a given burst of a callback-flood test.
3855 rcutorture.fqs_duration= [KNL]
3856 Set duration of force_quiescent_state bursts
3859 rcutorture.fqs_holdoff= [KNL]
3860 Set holdoff time within force_quiescent_state bursts
3863 rcutorture.fqs_stutter= [KNL]
3864 Set wait time between force_quiescent_state bursts
3867 rcutorture.gp_cond= [KNL]
3868 Use conditional/asynchronous update-side
3869 primitives, if available.
3871 rcutorture.gp_exp= [KNL]
3872 Use expedited update-side primitives, if available.
3874 rcutorture.gp_normal= [KNL]
3875 Use normal (non-expedited) asynchronous
3876 update-side primitives, if available.
3878 rcutorture.gp_sync= [KNL]
3879 Use normal (non-expedited) synchronous
3880 update-side primitives, if available. If all
3881 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3882 rcutorture.gp_normal=, and rcutorture.gp_sync=
3883 are zero, rcutorture acts as if is interpreted
3884 they are all non-zero.
3886 rcutorture.n_barrier_cbs= [KNL]
3887 Set callbacks/threads for rcu_barrier() testing.
3889 rcutorture.nfakewriters= [KNL]
3890 Set number of concurrent RCU writers. These just
3891 stress RCU, they don't participate in the actual
3892 test, hence the "fake".
3894 rcutorture.nreaders= [KNL]
3895 Set number of RCU readers. The value -1 selects
3896 N-1, where N is the number of CPUs. A value
3897 "n" less than -1 selects N-n-2, where N is again
3898 the number of CPUs. For example, -2 selects N
3899 (the number of CPUs), -3 selects N+1, and so on.
3901 rcutorture.object_debug= [KNL]
3902 Enable debug-object double-call_rcu() testing.
3904 rcutorture.onoff_holdoff= [KNL]
3905 Set time (s) after boot for CPU-hotplug testing.
3907 rcutorture.onoff_interval= [KNL]
3908 Set time (jiffies) between CPU-hotplug operations,
3909 or zero to disable CPU-hotplug testing.
3911 rcutorture.shuffle_interval= [KNL]
3912 Set task-shuffle interval (s). Shuffling tasks
3913 allows some CPUs to go into dyntick-idle mode
3914 during the rcutorture test.
3916 rcutorture.shutdown_secs= [KNL]
3917 Set time (s) after boot system shutdown. This
3918 is useful for hands-off automated testing.
3920 rcutorture.stall_cpu= [KNL]
3921 Duration of CPU stall (s) to test RCU CPU stall
3922 warnings, zero to disable.
3924 rcutorture.stall_cpu_holdoff= [KNL]
3925 Time to wait (s) after boot before inducing stall.
3927 rcutorture.stall_cpu_irqsoff= [KNL]
3928 Disable interrupts while stalling if set.
3930 rcutorture.stat_interval= [KNL]
3931 Time (s) between statistics printk()s.
3933 rcutorture.stutter= [KNL]
3934 Time (s) to stutter testing, for example, specifying
3935 five seconds causes the test to run for five seconds,
3936 wait for five seconds, and so on. This tests RCU's
3937 ability to transition abruptly to and from idle.
3939 rcutorture.test_boost= [KNL]
3940 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3941 "Maybe" means test if the RCU implementation
3942 under test support RCU priority boosting.
3944 rcutorture.test_boost_duration= [KNL]
3945 Duration (s) of each individual boost test.
3947 rcutorture.test_boost_interval= [KNL]
3948 Interval (s) between each boost test.
3950 rcutorture.test_no_idle_hz= [KNL]
3951 Test RCU's dyntick-idle handling. See also the
3952 rcutorture.shuffle_interval parameter.
3954 rcutorture.torture_type= [KNL]
3955 Specify the RCU implementation to test.
3957 rcutorture.verbose= [KNL]
3958 Enable additional printk() statements.
3960 rcupdate.rcu_cpu_stall_suppress= [KNL]
3961 Suppress RCU CPU stall warning messages.
3963 rcupdate.rcu_cpu_stall_timeout= [KNL]
3964 Set timeout for RCU CPU stall warning messages.
3966 rcupdate.rcu_expedited= [KNL]
3967 Use expedited grace-period primitives, for
3968 example, synchronize_rcu_expedited() instead
3969 of synchronize_rcu(). This reduces latency,
3970 but can increase CPU utilization, degrade
3971 real-time latency, and degrade energy efficiency.
3972 No effect on CONFIG_TINY_RCU kernels.
3974 rcupdate.rcu_normal= [KNL]
3975 Use only normal grace-period primitives,
3976 for example, synchronize_rcu() instead of
3977 synchronize_rcu_expedited(). This improves
3978 real-time latency, CPU utilization, and
3979 energy efficiency, but can expose users to
3980 increased grace-period latency. This parameter
3981 overrides rcupdate.rcu_expedited. No effect on
3982 CONFIG_TINY_RCU kernels.
3984 rcupdate.rcu_normal_after_boot= [KNL]
3985 Once boot has completed (that is, after
3986 rcu_end_inkernel_boot() has been invoked), use
3987 only normal grace-period primitives. No effect
3988 on CONFIG_TINY_RCU kernels.
3990 rcupdate.rcu_task_stall_timeout= [KNL]
3991 Set timeout in jiffies for RCU task stall warning
3992 messages. Disable with a value less than or equal
3995 rcupdate.rcu_self_test= [KNL]
3996 Run the RCU early boot self tests
3998 rcupdate.rcu_self_test_bh= [KNL]
3999 Run the RCU bh early boot self tests
4001 rcupdate.rcu_self_test_sched= [KNL]
4002 Run the RCU sched early boot self tests
4006 Run specified binary instead of /init from the ramdisk,
4007 used for early userspace startup. See initrd.
4010 force - Override the decision by the kernel to hide the
4011 advertisement of RDRAND support (this affects
4012 certain AMD processors because of buggy BIOS
4013 support, specifically around the suspend/resume
4017 Turn on/off individual RDT features. List is:
4018 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4020 E.g. to turn on cmt and turn off mba use:
4024 Format (x86 or x86_64):
4025 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4027 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4029 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
4030 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4031 reboot_force is either force or not specified,
4032 reboot_cpu is s[mp]#### with #### being the processor
4033 to be used for rebooting.
4036 [KNL, SMP] Set scheduler's default relax_domain_level.
4037 See Documentation/cgroup-v1/cpusets.txt.
4039 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4040 Format: <base1>,<size1>[,<base2>,<size2>,...]
4041 Reserve I/O ports or memory so the kernel won't use
4042 them. If <base> is less than 0x10000, the region
4043 is assumed to be I/O ports; otherwise it is memory.
4045 reservetop= [X86-32]
4047 Reserves a hole at the top of the kernel virtual
4052 Set the amount of memory to reserve for BIOS at
4053 the bottom of the address space.
4055 reset_devices [KNL] Force drivers to reset the underlying device
4056 during initialization.
4059 Specify the partition device for software suspend
4061 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4063 resume_offset= [SWSUSP]
4064 Specify the offset from the beginning of the partition
4065 given by "resume=" at which the swap header is located,
4066 in <PAGE_SIZE> units (needed only for swap files).
4067 See Documentation/power/swsusp-and-swap-files.txt
4069 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4070 read the resume files
4072 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4073 Useful for devices that are detected asynchronously
4074 (e.g. USB and MMC devices).
4076 hibernate= [HIBERNATION]
4077 noresume Don't check if there's a hibernation image
4078 present during boot.
4079 nocompress Don't compress/decompress hibernation images.
4080 no Disable hibernation and resume.
4081 protect_image Turn on image protection during restoration
4082 (that will set all pages holding image data
4083 during restoration read-only).
4085 retain_initrd [RAM] Keep initrd memory after extraction
4087 rfkill.default_state=
4088 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4089 etc. communication is blocked by default.
4092 rfkill.master_switch_mode=
4093 0 The "airplane mode" button does nothing.
4094 1 The "airplane mode" button toggles between everything
4095 blocked and the previous configuration.
4096 2 The "airplane mode" button toggles between everything
4097 blocked and everything unblocked.
4099 rhash_entries= [KNL,NET]
4100 Set number of hash buckets for route cache
4103 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4106 ro [KNL] Mount root device read-only on boot
4109 on Mark read-only kernel memory as read-only (default).
4110 off Leave read-only kernel memory writable for debugging.
4113 Enable the uart passthrough on the designated usb port
4114 on Rockchip SoCs. When active, the signals of the
4115 debug-uart get routed to the D+ and D- pins of the usb
4116 port and the regular usb controller gets disabled.
4118 root= [KNL] Root filesystem
4119 See name_to_dev_t comment in init/do_mounts.c.
4121 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4122 mount the root filesystem
4124 rootflags= [KNL] Set root filesystem mount option string
4126 rootfstype= [KNL] Set root filesystem type
4128 rootwait [KNL] Wait (indefinitely) for root device to show up.
4129 Useful for devices that are detected asynchronously
4130 (e.g. USB and MMC devices).
4132 rproc_mem=nn[KMG][@address]
4133 [KNL,ARM,CMA] Remoteproc physical memory block.
4134 Memory area to be used by remote processor image,
4137 rw [KNL] Mount root device read-write on boot
4139 S [KNL] Run init in single mode
4141 s390_iommu= [HW,S390]
4142 Set s390 IOTLB flushing mode
4144 With strict flushing every unmap operation will result in
4145 an IOTLB flush. Default is lazy flushing before reuse,
4149 See drivers/net/irda/sa1100_ir.c.
4151 sbni= [NET] Granch SBNI12 leased line adapter
4153 sched_debug [KNL] Enables verbose scheduler debug messages.
4155 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4156 Allowed values are enable and disable. This feature
4157 incurs a small amount of overhead in the scheduler
4158 but is useful for debugging and performance tuning.
4160 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4161 xtime_lock contention on larger systems, and/or RCU lock
4162 contention on all systems with CONFIG_MAXSMP set.
4163 Format: { "0" | "1" }
4164 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4166 Note: increases power consumption, thus should only be
4167 enabled if running jitter sensitive (HPC/RT) workloads.
4169 security= [SECURITY] Choose a security module to enable at boot.
4170 If this boot parameter is not specified, only the first
4171 security module asking for security registration will be
4172 loaded. An invalid security module name will be treated
4173 as if no module has been chosen.
4175 selinux= [SELINUX] Disable or enable SELinux at boot time.
4176 Format: { "0" | "1" }
4177 See security/selinux/Kconfig help text.
4180 Default value is set via kernel config option.
4181 If enabled at boot time, /selinux/disable can be used
4182 later to disable prior to initial policy load.
4184 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4185 Format: { "0" | "1" }
4186 See security/apparmor/Kconfig help text
4189 Default value is set via kernel config option.
4191 serialnumber [BUGS=X86-32]
4194 Maximal number of shapers.
4202 Disable merging of slabs with similar size. May be
4203 necessary if there is some reason to distinguish
4204 allocs to different slabs, especially in hardened
4205 environments where the risk of heap overflows and
4206 layout control by attackers can usually be
4207 frustrated by disabling merging. This will reduce
4208 most of the exposure of a heap attack to a single
4209 cache (risks via metadata attacks are mostly
4210 unchanged). Debug options disable merging on their
4212 For more information see Documentation/vm/slub.rst.
4214 slab_max_order= [MM, SLAB]
4215 Determines the maximum allowed order for slabs.
4216 A high setting may cause OOMs due to memory
4217 fragmentation. Defaults to 1 for systems with
4218 more than 32MB of RAM, 0 otherwise.
4220 slub_debug[=options[,slabs]] [MM, SLUB]
4221 Enabling slub_debug allows one to determine the
4222 culprit if slab objects become corrupted. Enabling
4223 slub_debug can create guard zones around objects and
4224 may poison objects when not in use. Also tracks the
4225 last alloc / free. For more information see
4226 Documentation/vm/slub.rst.
4228 slub_memcg_sysfs= [MM, SLUB]
4229 Determines whether to enable sysfs directories for
4230 memory cgroup sub-caches. 1 to enable, 0 to disable.
4231 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4232 Enabling this can lead to a very high number of debug
4233 directories and files being created under
4236 slub_max_order= [MM, SLUB]
4237 Determines the maximum allowed order for slabs.
4238 A high setting may cause OOMs due to memory
4239 fragmentation. For more information see
4240 Documentation/vm/slub.rst.
4242 slub_min_objects= [MM, SLUB]
4243 The minimum number of objects per slab. SLUB will
4244 increase the slab order up to slub_max_order to
4245 generate a sufficiently large slab able to contain
4246 the number of objects indicated. The higher the number
4247 of objects the smaller the overhead of tracking slabs
4248 and the less frequently locks need to be acquired.
4249 For more information see Documentation/vm/slub.rst.
4251 slub_min_order= [MM, SLUB]
4252 Determines the minimum page order for slabs. Must be
4253 lower than slub_max_order.
4254 For more information see Documentation/vm/slub.rst.
4256 slub_nomerge [MM, SLUB]
4257 Same with slab_nomerge. This is supported for legacy.
4258 See slab_nomerge for more information.
4261 Format: <io1>[,<io2>[,...,<io8>]]
4263 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4264 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4265 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4266 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4267 smsc-ircc2.ircc_irq= [HW] IRQ line
4268 smsc-ircc2.ircc_dma= [HW] DMA channel
4269 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4270 0: Toshiba Satellite 1800 (GP data pin select)
4271 1: Fast pin select (default)
4274 smt [KNL,S390] Set the maximum number of threads (logical
4275 CPUs) to use per physical CPU on systems capable of
4276 symmetric multithreading (SMT). Will be capped to the
4277 actual hardware limit.
4279 Default: -1 (no limit)
4282 [KNL] Should the soft-lockup detector generate panics.
4285 A nonzero value instructs the soft-lockup detector
4286 to panic the machine when a soft-lockup occurs. This
4287 is also controlled by CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC
4288 which is the respective build-time switch to that
4291 softlockup_all_cpu_backtrace=
4292 [KNL] Should the soft-lockup detector generate
4293 backtraces on all cpus.
4296 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4297 See Documentation/laptops/sonypi.txt
4299 spectre_v2= [X86] Control mitigation of Spectre variant 2
4300 (indirect branch speculation) vulnerability.
4301 The default operation protects the kernel from
4304 on - unconditionally enable, implies
4306 off - unconditionally disable, implies
4308 auto - kernel detects whether your CPU model is
4311 Selecting 'on' will, and 'auto' may, choose a
4312 mitigation method at run time according to the
4313 CPU, the available microcode, the setting of the
4314 CONFIG_RETPOLINE configuration option, and the
4315 compiler with which the kernel was built.
4317 Selecting 'on' will also enable the mitigation
4318 against user space to user space task attacks.
4320 Selecting 'off' will disable both the kernel and
4321 the user space protections.
4323 Specific mitigations can also be selected manually:
4325 retpoline - replace indirect branches
4326 retpoline,generic - google's original retpoline
4327 retpoline,amd - AMD-specific minimal thunk
4329 Not specifying this option is equivalent to
4333 [X86] Control mitigation of Spectre variant 2
4334 (indirect branch speculation) vulnerability between
4337 on - Unconditionally enable mitigations. Is
4338 enforced by spectre_v2=on
4340 off - Unconditionally disable mitigations. Is
4341 enforced by spectre_v2=off
4343 prctl - Indirect branch speculation is enabled,
4344 but mitigation can be enabled via prctl
4345 per thread. The mitigation control state
4346 is inherited on fork.
4349 - Like "prctl" above, but only STIBP is
4350 controlled per thread. IBPB is issued
4351 always when switching between different user
4355 - Same as "prctl" above, but all seccomp
4356 threads will enable the mitigation unless
4357 they explicitly opt out.
4360 - Like "seccomp" above, but only STIBP is
4361 controlled per thread. IBPB is issued
4362 always when switching between different
4363 user space processes.
4365 auto - Kernel selects the mitigation depending on
4366 the available CPU features and vulnerability.
4369 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4371 Not specifying this option is equivalent to
4372 spectre_v2_user=auto.
4374 spec_store_bypass_disable=
4375 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4376 (Speculative Store Bypass vulnerability)
4378 Certain CPUs are vulnerable to an exploit against a
4379 a common industry wide performance optimization known
4380 as "Speculative Store Bypass" in which recent stores
4381 to the same memory location may not be observed by
4382 later loads during speculative execution. The idea
4383 is that such stores are unlikely and that they can
4384 be detected prior to instruction retirement at the
4385 end of a particular speculation execution window.
4387 In vulnerable processors, the speculatively forwarded
4388 store can be used in a cache side channel attack, for
4389 example to read memory to which the attacker does not
4390 directly have access (e.g. inside sandboxed code).
4392 This parameter controls whether the Speculative Store
4393 Bypass optimization is used.
4395 On x86 the options are:
4397 on - Unconditionally disable Speculative Store Bypass
4398 off - Unconditionally enable Speculative Store Bypass
4399 auto - Kernel detects whether the CPU model contains an
4400 implementation of Speculative Store Bypass and
4401 picks the most appropriate mitigation. If the
4402 CPU is not vulnerable, "off" is selected. If the
4403 CPU is vulnerable the default mitigation is
4404 architecture and Kconfig dependent. See below.
4405 prctl - Control Speculative Store Bypass per thread
4406 via prctl. Speculative Store Bypass is enabled
4407 for a process by default. The state of the control
4408 is inherited on fork.
4409 seccomp - Same as "prctl" above, but all seccomp threads
4410 will disable SSB unless they explicitly opt out.
4412 Default mitigations:
4413 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4415 On powerpc the options are:
4417 on,auto - On Power8 and Power9 insert a store-forwarding
4418 barrier on kernel entry and exit. On Power7
4419 perform a software flush on kernel entry and
4423 Not specifying this option is equivalent to
4424 spec_store_bypass_disable=auto.
4426 spia_io_base= [HW,MTD]
4432 Control the Special Register Buffer Data Sampling
4435 Certain CPUs are vulnerable to an MDS-like
4436 exploit which can leak bits from the random
4439 By default, this issue is mitigated by
4440 microcode. However, the microcode fix can cause
4441 the RDRAND and RDSEED instructions to become
4442 much slower. Among other effects, this will
4443 result in reduced throughput from /dev/urandom.
4445 The microcode mitigation can be disabled with
4446 the following option:
4448 off: Disable mitigation and remove
4449 performance impact to RDRAND and RDSEED
4451 srcutree.counter_wrap_check [KNL]
4452 Specifies how frequently to check for
4453 grace-period sequence counter wrap for the
4454 srcu_data structure's ->srcu_gp_seq_needed field.
4455 The greater the number of bits set in this kernel
4456 parameter, the less frequently counter wrap will
4457 be checked for. Note that the bottom two bits
4460 srcutree.exp_holdoff [KNL]
4461 Specifies how many nanoseconds must elapse
4462 since the end of the last SRCU grace period for
4463 a given srcu_struct until the next normal SRCU
4464 grace period will be considered for automatic
4465 expediting. Set to zero to disable automatic
4469 Speculative Store Bypass Disable control
4471 On CPUs that are vulnerable to the Speculative
4472 Store Bypass vulnerability and offer a
4473 firmware based mitigation, this parameter
4474 indicates how the mitigation should be used:
4476 force-on: Unconditionally enable mitigation for
4477 for both kernel and userspace
4478 force-off: Unconditionally disable mitigation for
4479 for both kernel and userspace
4480 kernel: Always enable mitigation in the
4481 kernel, and offer a prctl interface
4482 to allow userspace to register its
4483 interest in being mitigated too.
4485 stack_guard_gap= [MM]
4486 override the default stack gap protection. The value
4487 is in page units and it defines how many pages prior
4488 to (for stacks growing down) resp. after (for stacks
4489 growing up) the main stack are reserved for no other
4490 mapping. Default value is 256 pages.
4493 Enabled the stack tracer on boot up.
4495 stacktrace_filter=[function-list]
4496 [FTRACE] Limit the functions that the stack tracer
4497 will trace at boot up. function-list is a comma separated
4498 list of functions. This list can be changed at run
4499 time by the stack_trace_filter file in the debugfs
4500 tracing directory. Note, this enables stack tracing
4501 and the stacktrace above is not needed.
4505 Set the STI (builtin display/keyboard on the HP-PARISC
4506 machines) console (graphic card) which should be used
4507 as the initial boot-console.
4508 See also comment in drivers/video/console/sticore.c.
4511 See comment in drivers/video/console/sticore.c.
4514 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4516 sunrpc.min_resvport=
4517 sunrpc.max_resvport=
4519 SunRPC servers often require that client requests
4520 originate from a privileged port (i.e. a port in the
4521 range 0 < portnr < 1024).
4522 An administrator who wishes to reserve some of these
4523 ports for other uses may adjust the range that the
4524 kernel's sunrpc client considers to be privileged
4525 using these two parameters to set the minimum and
4526 maximum port values.
4528 sunrpc.svc_rpc_per_connection_limit=
4530 Limit the number of requests that the server will
4531 process in parallel from a single connection.
4532 The default value is 0 (no limit).
4536 Control how the NFS server code allocates CPUs to
4537 service thread pools. Depending on how many NICs
4538 you have and where their interrupts are bound, this
4539 option will affect which CPUs will do NFS serving.
4540 Note: this parameter cannot be changed while the
4541 NFS server is running.
4543 auto the server chooses an appropriate mode
4544 automatically using heuristics
4545 global a single global pool contains all CPUs
4546 percpu one pool for each CPU
4547 pernode one pool for each NUMA node (equivalent
4548 to global on non-NUMA machines)
4550 sunrpc.tcp_slot_table_entries=
4551 sunrpc.udp_slot_table_entries=
4553 Sets the upper limit on the number of simultaneous
4554 RPC calls that can be sent from the client to a
4555 server. Increasing these values may allow you to
4556 improve throughput, but will also increase the
4557 amount of memory reserved for use by the client.
4559 suspend.pm_test_delay=
4561 Sets the number of seconds to remain in a suspend test
4562 mode before resuming the system (see
4563 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4564 is set. Default value is 5.
4567 [KNL] Enable accounting of swap in memory resource
4568 controller if no parameter or 1 is given or disable
4569 it if 0 is given (See Documentation/cgroup-v1/memory.txt)
4571 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4572 Format: { <int> | force | noforce }
4573 <int> -- Number of I/O TLB slabs
4574 force -- force using of bounce buffers even if they
4575 wouldn't be automatically used by the kernel
4576 noforce -- Never use bounce buffers (for debugging)
4580 sysfs.deprecated=0|1 [KNL]
4581 Enable/disable old style sysfs layout for old udev
4582 on older distributions. When this option is enabled
4583 very new udev will not work anymore. When this option
4584 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4585 in older udev will not work anymore.
4586 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4587 the kernel configuration.
4589 sysrq_always_enabled
4591 Ignore sysrq setting - this boot parameter will
4592 neutralize any effect of /proc/sys/kernel/sysrq.
4593 Useful for debugging.
4595 tcpmhash_entries= [KNL,NET]
4596 Set the number of tcp_metrics_hash slots.
4597 Default value is 8192 or 16384 depending on total
4598 ram pages. This is used to specify the TCP metrics
4599 cache size. See Documentation/networking/ip-sysctl.txt
4600 "tcp_no_metrics_save" section for more details.
4604 test_suspend= [SUSPEND][,N]
4605 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4606 standby suspend) or "freeze" (for suspend type freeze)
4607 as the system sleep state during system startup with
4608 the optional capability to repeat N number of times.
4609 The system is woken from this state using a
4610 wakeup-capable RTC alarm.
4612 thash_entries= [KNL,NET]
4613 Set number of hash buckets for TCP connection
4615 thermal.act= [HW,ACPI]
4616 -1: disable all active trip points in all thermal zones
4617 <degrees C>: override all lowest active trip points
4619 thermal.crt= [HW,ACPI]
4620 -1: disable all critical trip points in all thermal zones
4621 <degrees C>: override all critical trip points
4623 thermal.nocrt= [HW,ACPI]
4624 Set to disable actions on ACPI thermal zone
4625 critical and hot trip points.
4627 thermal.off= [HW,ACPI]
4628 1: disable ACPI thermal control
4630 thermal.psv= [HW,ACPI]
4631 -1: disable all passive trip points
4632 <degrees C>: override all passive trip points to this
4635 thermal.tzp= [HW,ACPI]
4636 Specify global default ACPI thermal zone polling rate
4637 <deci-seconds>: poll all this frequency
4638 0: no polling (default)
4641 Force threading of all interrupt handlers except those
4642 marked explicitly IRQF_NO_THREAD.
4645 Enable the Transcendent memory driver if built-in.
4647 tmem.cleancache=0|1 [KNL, XEN]
4648 Default is on (1). Disable the usage of the cleancache
4649 API to send anonymous pages to the hypervisor.
4651 tmem.frontswap=0|1 [KNL, XEN]
4652 Default is on (1). Disable the usage of the frontswap
4653 API to send swap pages to the hypervisor. If disabled
4654 the selfballooning and selfshrinking are force disabled.
4656 tmem.selfballooning=0|1 [KNL, XEN]
4657 Default is on (1). Disable the driving of swap pages
4660 tmem.selfshrinking=0|1 [KNL, XEN]
4661 Default is on (1). Partial swapoff that immediately
4662 transfers pages from Xen hypervisor back to the
4663 kernel based on different criteria.
4667 Specify if the kernel should make use of the cpu
4668 topology information if the hardware supports this.
4669 The scheduler will make use of this information and
4670 e.g. base its process migration decisions on it.
4673 topology_updates= [KNL, PPC, NUMA]
4675 Specify if the kernel should ignore (off)
4676 topology updates sent by the hypervisor to this
4681 tpm_suspend_pcr=[HW,TPM]
4682 Format: integer pcr id
4683 Specify that at suspend time, the tpm driver
4684 should extend the specified pcr with zeros,
4685 as a workaround for some chips which fail to
4686 flush the last written pcr on TPM_SaveState.
4687 This will guarantee that all the other pcrs
4690 trace_buf_size=nn[KMG]
4691 [FTRACE] will set tracing buffer size on each cpu.
4693 trace_event=[event-list]
4694 [FTRACE] Set and start specified trace events in order
4695 to facilitate early boot debugging. The event-list is a
4696 comma separated list of trace events to enable. See
4697 also Documentation/trace/events.rst
4699 trace_options=[option-list]
4700 [FTRACE] Enable or disable tracer options at boot.
4701 The option-list is a comma delimited list of options
4702 that can be enabled or disabled just as if you were
4703 to echo the option name into
4705 /sys/kernel/debug/tracing/trace_options
4707 For example, to enable stacktrace option (to dump the
4708 stack trace of each event), add to the command line:
4710 trace_options=stacktrace
4712 See also Documentation/trace/ftrace.rst "trace options"
4716 Have the tracepoints sent to printk as well as the
4717 tracing ring buffer. This is useful for early boot up
4718 where the system hangs or reboots and does not give the
4719 option for reading the tracing buffer or performing a
4720 ftrace_dump_on_oops.
4722 To turn off having tracepoints sent to printk,
4723 echo 0 > /proc/sys/kernel/tracepoint_printk
4724 Note, echoing 1 into this file without the
4725 tracepoint_printk kernel cmdline option has no effect.
4729 Having tracepoints sent to printk() and activating high
4730 frequency tracepoints such as irq or sched, can cause
4731 the system to live lock.
4734 [FTRACE] enable this option to disable tracing when a
4735 warning is hit. This turns off "tracing_on". Tracing can
4736 be enabled again by echoing '1' into the "tracing_on"
4737 file located in /sys/kernel/debug/tracing/
4739 This option is useful, as it disables the trace before
4740 the WARNING dump is called, which prevents the trace to
4741 be filled with content caused by the warning output.
4743 This option can also be set at run time via the sysctl
4744 option: kernel/traceoff_on_warning
4746 transparent_hugepage=
4748 Format: [always|madvise|never]
4749 Can be used to control the default behavior of the system
4750 with respect to transparent hugepages.
4751 See Documentation/admin-guide/mm/transhuge.rst
4754 tsc= Disable clocksource stability checks for TSC.
4756 [x86] reliable: mark tsc clocksource as reliable, this
4757 disables clocksource verification at runtime, as well
4758 as the stability checks done at bootup. Used to enable
4759 high-resolution timer mode on older hardware, and in
4760 virtualized environment.
4761 [x86] noirqtime: Do not use TSC to do irq accounting.
4762 Used to run time disable IRQ_TIME_ACCOUNTING on any
4763 platforms where RDTSC is slow and this accounting
4765 [x86] unstable: mark the TSC clocksource as unstable, this
4766 marks the TSC unconditionally unstable at bootup and
4767 avoids any further wobbles once the TSC watchdog notices.
4769 tsx= [X86] Control Transactional Synchronization
4770 Extensions (TSX) feature in Intel processors that
4771 support TSX control.
4773 This parameter controls the TSX feature. The options are:
4775 on - Enable TSX on the system. Although there are
4776 mitigations for all known security vulnerabilities,
4777 TSX has been known to be an accelerator for
4778 several previous speculation-related CVEs, and
4779 so there may be unknown security risks associated
4780 with leaving it enabled.
4782 off - Disable TSX on the system. (Note that this
4783 option takes effect only on newer CPUs which are
4784 not vulnerable to MDS, i.e., have
4785 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
4786 the new IA32_TSX_CTRL MSR through a microcode
4787 update. This new MSR allows for the reliable
4788 deactivation of the TSX functionality.)
4790 auto - Disable TSX if X86_BUG_TAA is present,
4791 otherwise enable TSX on the system.
4793 Not specifying this option is equivalent to tsx=off.
4795 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
4798 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
4799 Abort (TAA) vulnerability.
4801 Similar to Micro-architectural Data Sampling (MDS)
4802 certain CPUs that support Transactional
4803 Synchronization Extensions (TSX) are vulnerable to an
4804 exploit against CPU internal buffers which can forward
4805 information to a disclosure gadget under certain
4808 In vulnerable processors, the speculatively forwarded
4809 data can be used in a cache side channel attack, to
4810 access data to which the attacker does not have direct
4813 This parameter controls the TAA mitigation. The
4816 full - Enable TAA mitigation on vulnerable CPUs
4819 full,nosmt - Enable TAA mitigation and disable SMT on
4820 vulnerable CPUs. If TSX is disabled, SMT
4821 is not disabled because CPU is not
4822 vulnerable to cross-thread TAA attacks.
4823 off - Unconditionally disable TAA mitigation
4825 On MDS-affected machines, tsx_async_abort=off can be
4826 prevented by an active MDS mitigation as both vulnerabilities
4827 are mitigated with the same mechanism so in order to disable
4828 this mitigation, you need to specify mds=off too.
4830 Not specifying this option is equivalent to
4831 tsx_async_abort=full. On CPUs which are MDS affected
4832 and deploy MDS mitigation, TAA mitigation is not
4833 required and doesn't provide any additional
4837 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
4839 turbografx.map[2|3]= [HW,JOY]
4840 TurboGraFX parallel port interface
4842 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4843 See also Documentation/input/devices/joystick-parport.rst
4845 udbg-immortal [PPC] When debugging early kernel crashes that
4846 happen after console_init() and before a proper
4847 console driver takes over, this boot options might
4848 help "seeing" what's going on.
4850 uhash_entries= [KNL,NET]
4851 Set number of hash buckets for UDP/UDP-Lite connections
4854 [USB] Ignore overcurrent events (default N).
4855 Some badly-designed motherboards generate lots of
4856 bogus events, for ports that aren't wired to
4857 anything. Set this parameter to avoid log spamming.
4858 Note that genuine overcurrent events won't be
4862 [X86] Cause panic on unknown NMI.
4864 usbcore.authorized_default=
4865 [USB] Default USB device authorization:
4866 (default -1 = authorized except for wireless USB,
4867 0 = not authorized, 1 = authorized)
4869 usbcore.autosuspend=
4870 [USB] The autosuspend time delay (in seconds) used
4871 for newly-detected USB devices (default 2). This
4872 is the time required before an idle device will be
4873 autosuspended. Devices for which the delay is set
4874 to a negative value won't be autosuspended at all.
4876 usbcore.usbfs_snoop=
4877 [USB] Set to log all usbfs traffic (default 0 = off).
4879 usbcore.usbfs_snoop_max=
4880 [USB] Maximum number of bytes to snoop in each URB
4883 usbcore.blinkenlights=
4884 [USB] Set to cycle leds on hubs (default 0 = off).
4886 usbcore.old_scheme_first=
4887 [USB] Start with the old device initialization
4888 scheme (default 0 = off).
4890 usbcore.usbfs_memory_mb=
4891 [USB] Memory limit (in MB) for buffers allocated by
4892 usbfs (default = 16, 0 = max = 2047).
4894 usbcore.use_both_schemes=
4895 [USB] Try the other device initialization scheme
4896 if the first one fails (default 1 = enabled).
4898 usbcore.initial_descriptor_timeout=
4899 [USB] Specifies timeout for the initial 64-byte
4900 USB_REQ_GET_DESCRIPTOR request in milliseconds
4901 (default 5000 = 5.0 seconds).
4903 usbcore.nousb [USB] Disable the USB subsystem
4906 [USB] A list of quirk entries to augment the built-in
4907 usb core quirk list. List entries are separated by
4908 commas. Each entry has the form
4909 VendorID:ProductID:Flags. The IDs are 4-digit hex
4910 numbers and Flags is a set of letters. Each letter
4911 will change the built-in quirk; setting it if it is
4912 clear and clearing it if it is set. The letters have
4913 the following meanings:
4914 a = USB_QUIRK_STRING_FETCH_255 (string
4915 descriptors must not be fetched using
4917 b = USB_QUIRK_RESET_RESUME (device can't resume
4918 correctly so reset it instead);
4919 c = USB_QUIRK_NO_SET_INTF (device can't handle
4920 Set-Interface requests);
4921 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
4922 handle its Configuration or Interface
4924 e = USB_QUIRK_RESET (device can't be reset
4925 (e.g morph devices), don't use reset);
4926 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
4927 more interface descriptions than the
4928 bNumInterfaces count, and can't handle
4929 talking to these interfaces);
4930 g = USB_QUIRK_DELAY_INIT (device needs a pause
4931 during initialization, after we read
4932 the device descriptor);
4933 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
4934 high speed and super speed interrupt
4935 endpoints, the USB 2.0 and USB 3.0 spec
4936 require the interval in microframes (1
4937 microframe = 125 microseconds) to be
4938 calculated as interval = 2 ^
4940 Devices with this quirk report their
4941 bInterval as the result of this
4942 calculation instead of the exponent
4943 variable used in the calculation);
4944 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
4945 handle device_qualifier descriptor
4947 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
4948 generates spurious wakeup, ignore
4949 remote wakeup capability);
4950 k = USB_QUIRK_NO_LPM (device can't handle Link
4952 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
4953 (Device reports its bInterval as linear
4954 frames instead of the USB 2.0
4956 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
4957 to be disconnected before suspend to
4958 prevent spurious wakeup);
4959 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
4960 pause after every control message);
4961 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
4962 delay after resetting its port);
4963 Example: quirks=0781:5580:bk,0a5c:5834:gij
4966 [USBHID] The interval which mice are to be polled at.
4969 [USBHID] The interval which joysticks are to be polled at.
4972 [USBHID] The interval which keyboards are to be polled at.
4974 usb-storage.delay_use=
4975 [UMS] The delay in seconds before a new device is
4976 scanned for Logical Units (default 1).
4979 [UMS] A list of quirks entries to supplement or
4980 override the built-in unusual_devs list. List
4981 entries are separated by commas. Each entry has
4982 the form VID:PID:Flags where VID and PID are Vendor
4983 and Product ID values (4-digit hex numbers) and
4984 Flags is a set of characters, each corresponding
4985 to a common usb-storage quirk flag as follows:
4986 a = SANE_SENSE (collect more than 18 bytes
4987 of sense data, not on uas);
4988 b = BAD_SENSE (don't collect more than 18
4989 bytes of sense data, not on uas);
4990 c = FIX_CAPACITY (decrease the reported
4991 device capacity by one sector);
4992 d = NO_READ_DISC_INFO (don't use
4993 READ_DISC_INFO command, not on uas);
4994 e = NO_READ_CAPACITY_16 (don't use
4995 READ_CAPACITY_16 command);
4996 f = NO_REPORT_OPCODES (don't use report opcodes
4998 g = MAX_SECTORS_240 (don't transfer more than
4999 240 sectors at a time, uas only);
5000 h = CAPACITY_HEURISTICS (decrease the
5001 reported device capacity by one
5002 sector if the number is odd);
5003 i = IGNORE_DEVICE (don't bind to this
5005 j = NO_REPORT_LUNS (don't use report luns
5007 k = NO_SAME (do not use WRITE_SAME, uas only)
5008 l = NOT_LOCKABLE (don't try to lock and
5009 unlock ejectable media, not on uas);
5010 m = MAX_SECTORS_64 (don't transfer more
5011 than 64 sectors = 32 KB at a time,
5013 n = INITIAL_READ10 (force a retry of the
5014 initial READ(10) command, not on uas);
5015 o = CAPACITY_OK (accept the capacity
5016 reported by the device, not on uas);
5017 p = WRITE_CACHE (the device cache is ON
5018 by default, not on uas);
5019 r = IGNORE_RESIDUE (the device reports
5020 bogus residue values, not on uas);
5021 s = SINGLE_LUN (the device has only one
5023 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5024 commands, uas only);
5025 u = IGNORE_UAS (don't bind to the uas driver);
5026 w = NO_WP_DETECT (don't test whether the
5027 medium is write-protected).
5028 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5029 even if the device claims no cache,
5031 Example: quirks=0419:aaf5:rl,0421:0433:rc
5033 user_debug= [KNL,ARM]
5035 See arch/arm/Kconfig.debug help text.
5036 1 - undefined instruction events
5038 4 - invalid data aborts
5041 Example: user_debug=31
5044 [X86] Flags controlling user PTE allocations.
5046 nohigh = do not allocate PTE pages in
5047 HIGHMEM regardless of setting
5051 On X86_32, this is an alias for vdso32=. Otherwise:
5053 vdso=1: enable VDSO (the default)
5054 vdso=0: disable VDSO mapping
5056 vdso32= [X86] Control the 32-bit vDSO
5057 vdso32=1: enable 32-bit VDSO
5058 vdso32=0 or vdso32=2: disable 32-bit VDSO
5060 See the help text for CONFIG_COMPAT_VDSO for more
5061 details. If CONFIG_COMPAT_VDSO is set, the default is
5062 vdso32=0; otherwise, the default is vdso32=1.
5064 For compatibility with older kernels, vdso32=2 is an
5067 Try vdso32=0 if you encounter an error that says:
5068 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5071 vector=percpu: enable percpu vector domain
5073 video= [FB] Frame buffer configuration
5074 See Documentation/fb/modedb.txt.
5076 video.brightness_switch_enabled= [0,1]
5077 If set to 1, on receiving an ACPI notify event
5078 generated by hotkey, video driver will adjust brightness
5079 level and then send out the event to user space through
5080 the allocated input device; If set to 0, video driver
5081 will only send out the event without touching backlight
5086 [VMMIO] Memory mapped virtio (platform) device.
5088 <size>@<baseaddr>:<irq>[:<id>]
5090 <size> := size (can use standard suffixes
5092 <baseaddr> := physical base address
5093 <irq> := interrupt number (as passed to
5095 <id> := (optional) platform device id
5097 virtio_mmio.device=1K@0x100b0000:48:7
5099 Can be used multiple times for multiple devices.
5101 vga= [BOOT,X86-32] Select a particular video mode
5102 See Documentation/x86/boot.txt and
5103 Documentation/svga.txt.
5104 Use vga=ask for menu.
5105 This is actually a boot loader parameter; the value is
5106 passed to the kernel using a special protocol.
5108 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5109 size of <nn>. This can be used to increase the
5110 minimum size (128MB on x86). It can also be used to
5111 decrease the size and leave more room for directly
5114 vmcp_cma=nn[MG] [KNL,S390]
5115 Sets the memory size reserved for contiguous memory
5116 allocations for the vmcp device driver.
5118 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5121 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5124 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5128 Controls the behavior of vsyscalls (i.e. calls to
5129 fixed addresses of 0xffffffffff600x00 from legacy
5130 code). Most statically-linked binaries and older
5131 versions of glibc use these calls. Because these
5132 functions are at fixed addresses, they make nice
5133 targets for exploits that can control RIP.
5135 emulate [default] Vsyscalls turn into traps and are
5136 emulated reasonably safely.
5138 none Vsyscalls don't work at all. This makes
5139 them quite hard to use for exploits but
5140 might break your system.
5142 vt.color= [VT] Default text color.
5143 Format: 0xYX, X = foreground, Y = background.
5144 Default: 0x07 = light gray on black.
5146 vt.cur_default= [VT] Default cursor shape.
5147 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5148 the parameters of the <Esc>[?A;B;Cc escape sequence;
5149 see VGA-softcursor.txt. Default: 2 = underline.
5151 vt.default_blu= [VT]
5152 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5153 Change the default blue palette of the console.
5154 This is a 16-member array composed of values
5157 vt.default_grn= [VT]
5158 Format: <green0>,<green1>,<green2>,...,<green15>
5159 Change the default green palette of the console.
5160 This is a 16-member array composed of values
5163 vt.default_red= [VT]
5164 Format: <red0>,<red1>,<red2>,...,<red15>
5165 Change the default red palette of the console.
5166 This is a 16-member array composed of values
5172 Set system-wide default UTF-8 mode for all tty's.
5173 Default is 1, i.e. UTF-8 mode is enabled for all
5174 newly opened terminals.
5176 vt.global_cursor_default=
5179 Set system-wide default for whether a cursor
5180 is shown on new VTs. Default is -1,
5181 i.e. cursors will be created by default unless
5182 overridden by individual drivers. 0 will hide
5183 cursors, 1 will display them.
5185 vt.italic= [VT] Default color for italic text; 0-15.
5188 vt.underline= [VT] Default color for underlined text; 0-15.
5191 watchdog timers [HW,WDT] For information on watchdog timers,
5192 see Documentation/watchdog/watchdog-parameters.txt
5193 or other driver-specific files in the
5194 Documentation/watchdog/ directory.
5196 workqueue.watchdog_thresh=
5197 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5198 warn stall conditions and dump internal state to
5199 help debugging. 0 disables workqueue stall
5200 detection; otherwise, it's the stall threshold
5201 duration in seconds. The default value is 30 and
5202 it can be updated at runtime by writing to the
5203 corresponding sysfs file.
5205 workqueue.disable_numa
5206 By default, all work items queued to unbound
5207 workqueues are affine to the NUMA nodes they're
5208 issued on, which results in better behavior in
5209 general. If NUMA affinity needs to be disabled for
5210 whatever reason, this option can be used. Note
5211 that this also can be controlled per-workqueue for
5212 workqueues visible under /sys/bus/workqueue/.
5214 workqueue.power_efficient
5215 Per-cpu workqueues are generally preferred because
5216 they show better performance thanks to cache
5217 locality; unfortunately, per-cpu workqueues tend to
5218 be more power hungry than unbound workqueues.
5220 Enabling this makes the per-cpu workqueues which
5221 were observed to contribute significantly to power
5222 consumption unbound, leading to measurably lower
5223 power usage at the cost of small performance
5226 The default value of this parameter is determined by
5227 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5229 workqueue.debug_force_rr_cpu
5230 Workqueue used to implicitly guarantee that work
5231 items queued without explicit CPU specified are put
5232 on the local CPU. This guarantee is no longer true
5233 and while local CPU is still preferred work items
5234 may be put on foreign CPUs. This debug option
5235 forces round-robin CPU selection to flush out
5236 usages which depend on the now broken guarantee.
5237 When enabled, memory and cache locality will be
5240 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5241 default x2apic cluster mode on platforms
5244 x86_intel_mid_timer= [X86-32,APBT]
5245 Choose timer option for x86 Intel MID platform.
5246 Two valid options are apbt timer only and lapic timer
5247 plus one apbt timer for broadcast timer.
5248 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5250 xen_512gb_limit [KNL,X86-64,XEN]
5251 Restricts the kernel running paravirtualized under Xen
5252 to use only up to 512 GB of RAM. The reason to do so is
5253 crash analysis tools and Xen tools for doing domain
5254 save/restore/migration must be enabled to handle larger
5257 xen_emul_unplug= [HW,X86,XEN]
5258 Unplug Xen emulated devices
5259 Format: [unplug0,][unplug1]
5260 ide-disks -- unplug primary master IDE devices
5261 aux-ide-disks -- unplug non-primary-master IDE devices
5262 nics -- unplug network devices
5263 all -- unplug all emulated devices (NICs and IDE disks)
5264 unnecessary -- unplugging emulated devices is
5265 unnecessary even if the host did not respond to
5267 never -- do not unplug even if version check succeeds
5269 xen_legacy_crash [X86,XEN]
5270 Crash from Xen panic notifier, without executing late
5271 panic() code such as dumping handler.
5273 xen_nopvspin [X86,XEN]
5274 Disables the ticketlock slowpath using Xen PV
5278 Disables the PV optimizations forcing the HVM guest to
5279 run as generic HVM guest with no PV drivers.
5281 xen_scrub_pages= [XEN]
5282 Boolean option to control scrubbing pages before giving them back
5283 to Xen, for use by other domains. Can be also changed at runtime
5284 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5285 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5287 xen.event_eoi_delay= [XEN]
5288 How long to delay EOI handling in case of event
5289 storms (jiffies). Default is 10.
5291 xen.event_loop_timeout= [XEN]
5292 After which time (jiffies) the event handling loop
5293 should start to delay EOI handling. Default is 2.
5295 xirc2ps_cs= [NET,PCMCIA]
5297 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5299 xhci-hcd.quirks [USB,KNL]
5300 A hex value specifying bitmask with supplemental xhci
5301 host controller quirks. Meaning of each bit can be
5302 consulted in header drivers/usb/host/xhci.h.