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_rsdp= [ACPI,EFI,KEXEC]
140 Pass the RSDP address to the kernel, mostly used
141 on machines running EFI runtime service to boot the
142 second kernel for kdump.
144 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
145 Format: To spoof as Windows 98: ="Microsoft Windows"
147 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
148 of 2 which is mandated by ACPI 6) as the supported ACPI
149 specification revision (when using this switch, it may
150 be necessary to carry out a cold reboot _twice_ in a
151 row to make it take effect on the platform firmware).
153 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
154 acpi_osi="string1" # add string1
155 acpi_osi="!string2" # remove string2
156 acpi_osi=!* # remove all strings
157 acpi_osi=! # disable all built-in OS vendor
159 acpi_osi=!! # enable all built-in OS vendor
161 acpi_osi= # disable all strings
163 'acpi_osi=!' can be used in combination with single or
164 multiple 'acpi_osi="string1"' to support specific OS
165 vendor string(s). Note that such command can only
166 affect the default state of the OS vendor strings, thus
167 it cannot affect the default state of the feature group
168 strings and the current state of the OS vendor strings,
169 specifying it multiple times through kernel command line
170 is meaningless. This command is useful when one do not
171 care about the state of the feature group strings which
172 should be controlled by the OSPM.
174 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
175 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
176 can make '_OSI("Windows 2000")' TRUE.
178 'acpi_osi=' cannot be used in combination with other
179 'acpi_osi=' command lines, the _OSI method will not
180 exist in the ACPI namespace. NOTE that such command can
181 only affect the _OSI support state, thus specifying it
182 multiple times through kernel command line is also
185 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
188 'acpi_osi=!*' can be used in combination with single or
189 multiple 'acpi_osi="string1"' to support specific
190 string(s). Note that such command can affect the
191 current state of both the OS vendor strings and the
192 feature group strings, thus specifying it multiple times
193 through kernel command line is meaningful. But it may
194 still not able to affect the final state of a string if
195 there are quirks related to this string. This command
196 is useful when one want to control the state of the
197 feature group strings to debug BIOS issues related to
200 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
201 '_OSI("Module Device")' FALSE.
202 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
203 '_OSI("Module Device")' TRUE.
204 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
206 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
208 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
209 they all will make '_OSI("Windows 2000")' TRUE.
212 Override the pmtimer bug detection: force the kernel
213 to assume that this machine's pmtimer latches its value
214 and always returns good values.
216 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
217 Format: { level | edge | high | low }
219 acpi_skip_timer_override [HW,ACPI]
220 Recognize and ignore IRQ0/pin2 Interrupt Override.
221 For broken nForce2 BIOS resulting in XT-PIC timer.
223 acpi_sleep= [HW,ACPI] Sleep options
224 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
225 old_ordering, nonvs, sci_force_enable, nobl }
226 See Documentation/power/video.txt for information on
228 s3_beep is for debugging; it makes the PC's speaker beep
229 as soon as the kernel's real-mode entry point is called.
230 s4_nohwsig prevents ACPI hardware signature from being
231 used during resume from hibernation.
232 old_ordering causes the ACPI 1.0 ordering of the _PTS
233 control method, with respect to putting devices into
234 low power states, to be enforced (the ACPI 2.0 ordering
235 of _PTS is used by default).
236 nonvs prevents the kernel from saving/restoring the
237 ACPI NVS memory during suspend/hibernation and resume.
238 sci_force_enable causes the kernel to set SCI_EN directly
239 on resume from S1/S3 (which is against the ACPI spec,
240 but some broken systems don't work without it).
241 nobl causes the internal blacklist of systems known to
242 behave incorrectly in some ways with respect to system
243 suspend and resume to be ignored (use wisely).
245 acpi_use_timer_override [HW,ACPI]
246 Use timer override. For some broken Nvidia NF5 boards
247 that require a timer override, but don't have HPET
249 add_efi_memmap [EFI; X86] Include EFI memory map in
250 kernel's map of available physical RAM.
253 { off | try_unsupported }
254 off: disable AGP support
255 try_unsupported: try to drive unsupported chipsets
256 (may crash computer or cause data corruption)
259 See Documentation/sound/alsa-configuration.rst
262 Allow the default userspace alignment fault handler
263 behaviour to be specified. Bit 0 enables warnings,
264 bit 1 enables fixups, and bit 2 sends a segfault.
266 align_va_addr= [X86-64]
267 Align virtual addresses by clearing slice [14:12] when
268 allocating a VMA at process creation time. This option
269 gives you up to 3% performance improvement on AMD F15h
270 machines (where it is enabled by default) for a
271 CPU-intensive style benchmark, and it can vary highly in
272 a microbenchmark depending on workload and compiler.
274 32: only for 32-bit processes
275 64: only for 64-bit processes
276 on: enable for both 32- and 64-bit processes
277 off: disable for both 32- and 64-bit processes
279 alloc_snapshot [FTRACE]
280 Allocate the ftrace snapshot buffer on boot up when the
281 main buffer is allocated. This is handy if debugging
282 and you need to use tracing_snapshot() on boot up, and
283 do not want to use tracing_snapshot_alloc() as it needs
284 to be done where GFP_KERNEL allocations are allowed.
286 amd_iommu= [HW,X86-64]
287 Pass parameters to the AMD IOMMU driver in the system.
289 fullflush - enable flushing of IO/TLB entries when
290 they are unmapped. Otherwise they are
291 flushed before they will be reused, which
293 off - do not initialize any AMD IOMMU found in
295 force_isolation - Force device isolation for all
296 devices. The IOMMU driver is not
297 allowed anymore to lift isolation
298 requirements as needed. This option
299 does not override iommu=pt
301 amd_iommu_dump= [HW,X86-64]
302 Enable AMD IOMMU driver option to dump the ACPI table
303 for AMD IOMMU. With this option enabled, AMD IOMMU
304 driver will print ACPI tables for AMD IOMMU during
305 IOMMU initialization.
307 amd_iommu_intr= [HW,X86-64]
308 Specifies one of the following AMD IOMMU interrupt
310 legacy - Use legacy interrupt remapping mode.
311 vapic - Use virtual APIC mode, which allows IOMMU
312 to inject interrupts directly into guest.
313 This mode requires kvm-amd.avic=1.
314 (Default when IOMMU HW support is present.)
316 amijoy.map= [HW,JOY] Amiga joystick support
317 Map of devices attached to JOY0DAT and JOY1DAT
319 See also Documentation/input/joydev/joystick.rst
321 analog.map= [HW,JOY] Analog joystick and gamepad support
322 Specifies type or capabilities of an analog joystick
323 connected to one of 16 gameports
324 Format: <type1>,<type2>,..<type16>
327 Power management functions (SPARCstation-4/5 + deriv.)
329 Disable APC CPU standby support. SPARCstation-Fox does
330 not play well with APC CPU idle - disable it if you have
331 APC and your system crashes randomly.
333 apic= [APIC,X86] Advanced Programmable Interrupt Controller
334 Change the output verbosity while booting
335 Format: { quiet (default) | verbose | debug }
336 Change the amount of debugging information output
337 when initialising the APIC and IO-APIC components.
338 For X86-32, this can also be used to specify an APIC
340 Format: apic=driver_name
341 Examples: apic=bigsmp
343 apic_extnmi= [APIC,X86] External NMI delivery setting
344 Format: { bsp (default) | all | none }
345 bsp: External NMI is delivered only to CPU 0
346 all: External NMIs are broadcast to all CPUs as a
348 none: External NMI is masked for all CPUs. This is
349 useful so that a dump capture kernel won't be
353 See Documentation/networking/ipv6.txt.
355 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
356 Limit apic dumping. The parameter defines the maximal
357 number of local apics being dumped. Also it is possible
358 to set it to "all" by meaning -- no limit here.
359 Format: { 1 (default) | 2 | ... | all }.
360 The parameter valid if only apic=debug or
361 apic=verbose is specified.
362 Example: apic=debug show_lapic=all
364 apm= [APM] Advanced Power Management
365 See header of arch/x86/kernel/apm_32.c.
367 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
368 Format: <io>,<irq>,<nodeID>
372 atarimouse= [HW,MOUSE] Atari Mouse
374 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
375 EzKey and similar keyboards
377 atkbd.reset= [HW] Reset keyboard during initialization
379 atkbd.set= [HW] Select keyboard code set
380 Format: <int> (2 = AT (default), 3 = PS/2)
382 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
385 atkbd.softraw= [HW] Choose between synthetic and real raw mode
386 Format: <bool> (0 = real, 1 = synthetic (default))
388 atkbd.softrepeat= [HW]
389 Use software keyboard repeat
391 audit= [KNL] Enable the audit sub-system
392 Format: { "0" | "1" | "off" | "on" }
393 0 | off - kernel audit is disabled and can not be
394 enabled until the next reboot
395 unset - kernel audit is initialized but disabled and
396 will be fully enabled by the userspace auditd.
397 1 | on - kernel audit is initialized and partially
398 enabled, storing at most audit_backlog_limit
399 messages in RAM until it is fully enabled by the
403 audit_backlog_limit= [KNL] Set the audit queue size limit.
404 Format: <int> (must be >=0)
407 bau= [X86_UV] Enable the BAU on SGI UV. The default
408 behavior is to disable the BAU (i.e. bau=0).
409 Format: { "0" | "1" }
412 unset - Disable the BAU.
414 baycom_epp= [HW,AX25]
417 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
419 See header of drivers/net/hamradio/baycom_par.c.
421 baycom_ser_fdx= [HW,AX25]
422 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
423 Format: <io>,<irq>,<mode>[,<baud>]
424 See header of drivers/net/hamradio/baycom_ser_fdx.c.
426 baycom_ser_hdx= [HW,AX25]
427 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
428 Format: <io>,<irq>,<mode>
429 See header of drivers/net/hamradio/baycom_ser_hdx.c.
431 blkdevparts= Manual partition parsing of block device(s) for
432 embedded devices based on command line input.
433 See Documentation/block/cmdline-partition.txt
435 boot_delay= Milliseconds to delay each printk during boot.
436 Values larger than 10 seconds (10000) are changed to
440 bootmem_debug [KNL] Enable bootmem allocator debug messages.
443 Disable BERT OS support on buggy BIOSes.
445 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
446 bttv.radio= Most important insmod options are available as
448 bttv.pll= See Documentation/media/v4l-drivers/bttv.rst
451 bulk_remove=off [PPC] This parameter disables the use of the pSeries
452 firmware feature for flushing multiple hpte entries
455 c101= [NET] Moxa C101 synchronous serial card
457 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
458 Sometimes CPU hardware bugs make them report the cache
459 size incorrectly. The kernel will attempt work arounds
460 to fix known problems, but for some CPUs it is not
461 possible to determine what the correct size should be.
462 This option provides an override for these situations.
464 ca_keys= [KEYS] This parameter identifies a specific key(s) on
465 the system trusted keyring to be used for certificate
467 format: { id:<keyid> | builtin }
469 cca= [MIPS] Override the kernel pages' cache coherency
470 algorithm. Accepted values range from 0 to 7
471 inclusive. See arch/mips/include/asm/pgtable-bits.h
472 for platform specific values (SB1, Loongson3 and
475 ccw_timeout_log [S390]
476 See Documentation/s390/CommonIO for details.
478 cgroup_disable= [KNL] Disable a particular controller
479 Format: {name of the controller(s) to disable}
480 The effects of cgroup_disable=foo are:
481 - foo isn't auto-mounted if you mount all cgroups in
483 - foo isn't visible as an individually mountable
485 {Currently only "memory" controller deal with this and
486 cut the overhead, others just disable the usage. So
487 only cgroup_disable=memory is actually worthy}
489 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
490 Format: { { controller | "all" | "named" }
491 [,{ controller | "all" | "named" }...] }
492 Like cgroup_disable, but only applies to cgroup v1;
493 the blacklisted controllers remain available in cgroup2.
494 "all" blacklists all controllers and "named" disables
495 named mounts. Specifying both "all" and "named" disables
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 clearcpuid=BITNUM [X86]
562 Disable CPUID feature X for the kernel. See
563 arch/x86/include/asm/cpufeatures.h for the valid bit
564 numbers. Note the Linux specific bits are not necessarily
565 stable over kernel options, but the vendor specific
567 Also note that user programs calling CPUID directly
568 or using the feature without checking anything
569 will still see it. This just prevents it from
570 being used by the kernel or shown in /proc/cpuinfo.
571 Also note the kernel might malfunction if you disable
574 cma=nn[MG]@[start[MG][-end[MG]]]
576 Sets the size of kernel global memory area for
577 contiguous memory allocations and optionally the
578 placement constraint by the physical address range of
579 memory allocations. A value of 0 disables CMA
580 altogether. For more information, see
581 include/linux/dma-contiguous.h
583 cmo_free_hint= [PPC] Format: { yes | no }
584 Specify whether pages are marked as being inactive
585 when they are freed. This is used in CMO environments
586 to determine OS memory pressure for page stealing by
590 coherent_pool=nn[KMG] [ARM,KNL]
591 Sets the size of memory pool for coherent, atomic dma
592 allocations, by default set to 256K.
594 com20020= [HW,NET] ARCnet - COM20020 chipset
596 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
598 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
602 ARCnet - COM90xx chipset (memory-mapped buffers)
603 Format: <io>[,<irq>[,<memstart>]]
605 condev= [HW,S390] console device
608 console= [KNL] Output console device and options.
610 tty<n> Use the virtual console device <n>.
614 Use the specified serial port. The options are of
615 the form "bbbbpnf", where "bbbb" is the baud rate,
616 "p" is parity ("n", "o", or "e"), "n" is number of
617 bits, and "f" is flow control ("r" for RTS or
618 omit it). Default is "9600n8".
620 See Documentation/admin-guide/serial-console.rst for more
622 Documentation/networking/netconsole.txt for an
625 uart[8250],io,<addr>[,options]
626 uart[8250],mmio,<addr>[,options]
627 uart[8250],mmio16,<addr>[,options]
628 uart[8250],mmio32,<addr>[,options]
629 uart[8250],0x<addr>[,options]
630 Start an early, polled-mode console on the 8250/16550
631 UART at the specified I/O port or MMIO address,
632 switching to the matching ttyS device later.
633 MMIO inter-register address stride is either 8-bit
634 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
635 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
636 to be equivalent to 'mmio'. 'options' are specified in
637 the same format described for ttyS above; if unspecified,
638 the h/w is not re-initialized.
640 hvc<n> Use the hypervisor console device <n>. This is for
641 both Xen and PowerPC hypervisors.
643 If the device connected to the port is not a TTY but a braille
644 device, prepend "brl," before the device type, for instance
646 For now, only VisioBraille is supported.
649 [KNL] Change console messages format
651 By default we print messages on consoles in
652 "[time stamp] text\n" format (time stamp may not be
653 printed, depending on CONFIG_PRINTK_TIME or
654 `printk_time' param).
656 Switch to syslog format: "<%u>[time stamp] text\n"
657 IOW, each message will have a facility and loglevel
658 prefix. The format is similar to one used by syslog()
659 syscall, or to executing "dmesg -S --raw" or to reading
662 consoleblank= [KNL] The console blank (screen saver) timeout in
663 seconds. A value of 0 disables the blank timer.
667 [KNL] Change the default value for
668 /proc/<pid>/coredump_filter.
669 See also Documentation/filesystems/proc.txt.
671 coresight_cpu_debug.enable
674 Enable/disable the CPU sampling based debugging.
675 0: default value, disable debugging
676 1: enable debugging at boot time
678 cpuidle.off=1 [CPU_IDLE]
679 disable the cpuidle sub-system
682 [CPU_IDLE] Name of the cpuidle governor to use.
684 cpufreq.off=1 [CPU_FREQ]
685 disable the cpufreq sub-system
688 [X86] Delay for N microsec between assert and de-assert
689 of APIC INIT to start processors. This delay occurs
690 on every CPU online, such as boot, and resume from suspend.
693 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
695 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
697 crashkernel=size[KMG][@offset[KMG]]
698 [KNL] Using kexec, Linux can switch to a 'crash kernel'
699 upon panic. This parameter reserves the physical
700 memory region [offset, offset + size] for that kernel
701 image. If '@offset' is omitted, then a suitable offset
702 is selected automatically. Check
703 Documentation/kdump/kdump.txt for further details.
705 crashkernel=range1:size1[,range2:size2,...][@offset]
706 [KNL] Same as above, but depends on the memory
707 in the running system. The syntax of range is
708 start-[end] where start and end are both
709 a memory unit (amount[KMG]). See also
710 Documentation/kdump/kdump.txt for an example.
712 crashkernel=size[KMG],high
713 [KNL, x86_64] range could be above 4G. Allow kernel
714 to allocate physical memory region from top, so could
715 be above 4G if system have more than 4G ram installed.
716 Otherwise memory region will be allocated below 4G, if
718 It will be ignored if crashkernel=X is specified.
719 crashkernel=size[KMG],low
720 [KNL, x86_64] range under 4G. When crashkernel=X,high
721 is passed, kernel could allocate physical memory region
722 above 4G, that cause second kernel crash on system
723 that require some amount of low memory, e.g. swiotlb
724 requires at least 64M+32K low memory, also enough extra
725 low memory is needed to make sure DMA buffers for 32-bit
726 devices won't run out. Kernel would try to allocate at
727 at least 256M below 4G automatically.
728 This one let user to specify own low range under 4G
729 for second kernel instead.
730 0: to disable low allocation.
731 It will be ignored when crashkernel=X,high is not used
732 or memory reserved is below 4G.
735 [KNL] Disable crypto self-tests
740 cs89x0_media= [HW,NET]
741 Format: { rj45 | aui | bnc }
744 See header of drivers/s390/block/dasd_devmap.c.
746 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
747 (one device per port)
748 Format: <port#>,<type>
749 See also Documentation/input/devices/joystick-parport.rst
751 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
753 Documentation/admin-guide/dynamic-debug-howto.rst for
754 details. Deprecated, see dyndbg.
756 debug [KNL] Enable kernel debugging (events log level).
759 [KNL] Enable printing [hashed] pointers early in the
760 boot sequence. If enabled, we use a weak hash instead
761 of siphash to hash pointers. Use this option if you are
762 seeing instances of '(___ptrval___)') and need to see a
763 value (hashed pointer) instead. Cryptographically
764 insecure, please do not use on production kernels.
767 [KNL] verbose self-tests
769 Print debugging info while doing the locking API
771 We default to 0 (no extra messages), setting it to
772 1 will print _a lot_ more information - normally
773 only useful to kernel developers.
775 debug_objects [KNL] Enable object debugging
778 [KNL] Disable object debugging
780 debug_guardpage_minorder=
781 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
782 parameter allows control of the order of pages that will
783 be intentionally kept free (and hence protected) by the
784 buddy allocator. Bigger value increase the probability
785 of catching random memory corruption, but reduce the
786 amount of memory for normal system use. The maximum
787 possible value is MAX_ORDER/2. Setting this parameter
788 to 1 or 2 should be enough to identify most random
789 memory corruption problems caused by bugs in kernel or
790 driver code when a CPU writes to (or reads from) a
791 random memory location. Note that there exists a class
792 of memory corruptions problems caused by buggy H/W or
793 F/W or by drivers badly programing DMA (basically when
794 memory is written at bus level and the CPU MMU is
795 bypassed) which are not detectable by
796 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
797 tracking down these problems.
800 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
801 parameter enables the feature at boot time. In
802 default, it is disabled. We can avoid allocating huge
803 chunk of memory for debug pagealloc if we don't enable
804 it at boot time and the system will work mostly same
805 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
806 on: enable the feature
808 debugpat [X86] Enable PAT debugging
810 decnet.addr= [HW,NET]
811 Format: <area>[,<node>]
812 See also Documentation/networking/decnet.txt.
815 [same as hugepagesz=] The size of the default
816 HugeTLB page size. This is the size represented by
817 the legacy /proc/ hugepages APIs, used for SHM, and
818 default size when mounting hugetlbfs filesystems.
819 Defaults to the default architecture's huge page size
822 deferred_probe_timeout=
823 [KNL] Debugging option to set a timeout in seconds for
824 deferred probe to give up waiting on dependencies to
825 probe. Only specific dependencies (subsystems or
826 drivers) that have opted in will be ignored. A timeout of 0
827 will timeout at the end of initcalls. This option will also
828 dump out devices still on the deferred probe list after
832 Set number of hash buckets for dentry cache.
834 disable_1tb_segments [PPC]
835 Disables the use of 1TB hash page table segments. This
836 causes the kernel to fall back to 256MB segments which
837 can be useful when debugging issues that require an SLB
841 See Documentation/networking/ipv6.txt.
844 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
845 hardening is enabled for this boot. Hardened
846 usercopy checking is used to protect the kernel
847 from reading or writing beyond known memory
848 allocation boundaries as a proactive defense
849 against bounds-checking flaws in the kernel's
850 copy_to_user()/copy_from_user() interface.
851 on Perform hardened usercopy checks (default).
852 off Disable hardened usercopy checks.
855 Disable RADIX MMU mode on POWER9
857 disable_cpu_apicid= [X86,APIC,SMP]
859 The number of initial APIC ID for the
860 corresponding CPU to be disabled at boot,
861 mostly used for the kdump 2nd kernel to
862 disable BSP to wake up multiple CPUs without
863 causing system reset or hang due to sending
866 perf_v4_pmi= [X86,INTEL]
868 Disable Intel PMU counter freezing feature.
869 The feature only exists starting from
870 Arch Perfmon v4 (Skylake and newer).
872 disable_ddw [PPC/PSERIES]
873 Disable Dynamic DMA Window support. Use this if
874 to workaround buggy firmware.
877 See Documentation/networking/ipv6.txt.
879 disable_mtrr_cleanup [X86]
880 The kernel tries to adjust MTRR layout from continuous
881 to discrete, to make X server driver able to add WB
882 entry later. This parameter disables that.
884 disable_mtrr_trim [X86, Intel and AMD only]
885 By default the kernel will trim any uncacheable
886 memory out of your available memory pool based on
887 MTRR settings. This parameter disables that behavior,
888 possibly causing your machine to run very slowly.
890 disable_timer_pin_1 [X86]
891 Disable PIN 1 of APIC timer
892 Can be useful to work around chipset bugs.
894 dis_ucode_ldr [X86] Disable the microcode loader.
896 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
897 this option disables the debugging code at boot.
899 dma_debug_entries=<number>
900 This option allows to tune the number of preallocated
901 entries for DMA-API debugging code. One entry is
902 required per DMA-API allocation. Use this if the
903 DMA-API debugging code disables itself because the
904 architectural default is too low.
906 dma_debug_driver=<driver_name>
907 With this option the DMA-API debugging driver
908 filter feature can be enabled at boot time. Just
909 pass the driver to filter for as the parameter.
910 The filter can be disabled or changed to another
911 driver later using sysfs.
913 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
914 Broken monitors, graphic adapters, KVMs and EDIDless
915 panels may send no or incorrect EDID data sets.
916 This parameter allows to specify an EDID data sets
917 in the /lib/firmware directory that are used instead.
918 Generic built-in EDID data sets are used, if one of
919 edid/1024x768.bin, edid/1280x1024.bin,
920 edid/1680x1050.bin, or edid/1920x1080.bin is given
921 and no file with the same name exists. Details and
922 instructions how to build your own EDID data are
923 available in Documentation/EDID/HOWTO.txt. An EDID
924 data set will only be used for a particular connector,
925 if its name and a colon are prepended to the EDID
926 name. Each connector may use a unique EDID data
927 set by separating the files with a comma. An EDID
928 data set with no connector name will be used for
929 any connectors not explicitly specified.
934 Format: {"off" | "known"}
935 Control how the dt_cpu_ftrs device-tree binding is
936 used for CPU feature discovery and setup (if it
938 off: Do not use it, fall back to legacy cpu table.
939 known: Do not pass through unknown features to guests
940 or userspace, only those that the kernel is aware of.
942 dump_apple_properties [X86]
943 Dump name and content of EFI device properties on
944 x86 Macs. Useful for driver authors to determine
945 what data is available or for reverse-engineering.
947 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
948 module.dyndbg[="val"]
949 Enable debug messages at boot time. See
950 Documentation/admin-guide/dynamic-debug-howto.rst
953 nompx [X86] Disables Intel Memory Protection Extensions.
954 See Documentation/x86/intel_mpx.txt for more
955 information about the feature.
957 nopku [X86] Disable Memory Protection Keys CPU feature found
960 module.async_probe [KNL]
961 Enable asynchronous probe on this module.
963 early_ioremap_debug [KNL]
964 Enable debug messages in early_ioremap support. This
965 is useful for tracking down temporary early mappings
966 which are not unmapped.
968 earlycon= [KNL] Output early console device and options.
970 [ARM64] The early console is determined by the
971 stdout-path property in device tree's chosen node,
972 or determined by the ACPI SPCR table.
974 [X86] When used with no options the early console is
975 determined by the ACPI SPCR table.
977 cdns,<addr>[,options]
978 Start an early, polled-mode console on a Cadence
979 (xuartps) serial port at the specified address. Only
980 supported option is baud rate. If baud rate is not
981 specified, the serial port must already be setup and
984 uart[8250],io,<addr>[,options]
985 uart[8250],mmio,<addr>[,options]
986 uart[8250],mmio32,<addr>[,options]
987 uart[8250],mmio32be,<addr>[,options]
988 uart[8250],0x<addr>[,options]
989 Start an early, polled-mode console on the 8250/16550
990 UART at the specified I/O port or MMIO address.
991 MMIO inter-register address stride is either 8-bit
992 (mmio) or 32-bit (mmio32 or mmio32be).
993 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
994 to be equivalent to 'mmio'. 'options' are specified
995 in the same format described for "console=ttyS<n>"; if
996 unspecified, the h/w is not initialized.
1000 Start an early, polled-mode console on a pl011 serial
1001 port at the specified address. The pl011 serial port
1002 must already be setup and configured. Options are not
1003 yet supported. If 'mmio32' is specified, then only
1004 the driver will use only 32-bit accessors to read/write
1005 the device registers.
1008 Start an early, polled-mode console on a meson serial
1009 port at the specified address. The serial port must
1010 already be setup and configured. Options are not yet
1014 Start an early, polled-mode console on an msm serial
1015 port at the specified address. The serial port
1016 must already be setup and configured. Options are not
1019 msm_serial_dm,<addr>
1020 Start an early, polled-mode console on an msm serial
1021 dm port at the specified address. The serial port
1022 must already be setup and configured. Options are not
1026 Start an early, polled-mode console on a serial port
1027 of an Actions Semi SoC, such as S500 or S900, at the
1028 specified address. The serial port must already be
1029 setup and configured. Options are not yet supported.
1032 Start an early, polled-mode console on a serial port
1033 of an RDA Micro SoC, such as RDA8810PL, at the
1034 specified address. The serial port must already be
1035 setup and configured. Options are not yet supported.
1037 smh Use ARM semihosting calls for early console.
1045 Use early console provided by serial driver available
1046 on Samsung SoCs, requires selecting proper type and
1047 a correct base address of the selected UART port. The
1048 serial port must already be setup and configured.
1049 Options are not yet supported.
1052 Start an early, polled-mode console on a lantiq serial
1053 (lqasc) port at the specified address. The serial port
1054 must already be setup and configured. Options are not
1059 Use early console provided by Freescale LP UART driver
1060 found on Freescale Vybrid and QorIQ LS1021A processors.
1061 A valid base address must be provided, and the serial
1062 port must already be setup and configured.
1065 Start an early, polled-mode console on the
1066 Armada 3700 serial port at the specified
1067 address. The serial port must already be setup
1068 and configured. Options are not yet supported.
1071 Start an early, polled-mode console on a Qualcomm
1072 Generic Interface (GENI) based serial port at the
1073 specified address. The serial port must already be
1074 setup and configured. Options are not yet supported.
1077 Start an early, unaccelerated console on the EFI
1078 memory mapped framebuffer (if available). On cache
1079 coherent non-x86 systems that use system memory for
1080 the framebuffer, pass the 'ram' option so that it is
1081 mapped with the correct attributes.
1083 earlyprintk= [X86,SH,ARM,M68k,S390]
1087 earlyprintk=serial[,ttySn[,baudrate]]
1088 earlyprintk=serial[,0x...[,baudrate]]
1089 earlyprintk=ttySn[,baudrate]
1090 earlyprintk=dbgp[debugController#]
1091 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1092 earlyprintk=xdbc[xhciController#]
1094 earlyprintk is useful when the kernel crashes before
1095 the normal console is initialized. It is not enabled by
1096 default because it has some cosmetic problems.
1098 Append ",keep" to not disable it when the real console
1101 Only one of vga, efi, serial, or usb debug port can
1104 Currently only ttyS0 and ttyS1 may be specified by
1105 name. Other I/O ports may be explicitly specified
1106 on some architectures (x86 and arm at least) by
1107 replacing ttySn with an I/O port address, like this:
1108 earlyprintk=serial,0x1008,115200
1109 You can find the port for a given device in
1110 /proc/tty/driver/serial:
1111 2: uart:ST16650V2 port:00001008 irq:18 ...
1113 Interaction with the standard serial driver is not
1116 The VGA and EFI output is eventually overwritten by
1119 The xen output can only be used by Xen PV guests.
1121 The sclp output can only be used on s390.
1123 The optional "force" to "pciserial" enables use of a
1124 PCI device even when its classcode is not of the
1127 edac_report= [HW,EDAC] Control how to report EDAC event
1128 Format: {"on" | "off" | "force"}
1129 on: enable EDAC to report H/W event. May be overridden
1130 by other higher priority error reporting module.
1131 off: disable H/W event reporting through EDAC.
1132 force: enforce the use of EDAC to report H/W event.
1135 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1138 This is designed to be used in conjunction with
1139 the boot argument: earlyprintk=vga
1142 Format: {"off" | "on" | "skip[mbr]"}
1145 Format: { "old_map", "nochunk", "noruntime", "debug" }
1146 old_map [X86-64]: switch to the old ioremap-based EFI
1147 runtime services mapping. 32-bit still uses this one by
1149 nochunk: disable reading files in "chunks" in the EFI
1150 boot stub, as chunking can cause problems with some
1151 firmware implementations.
1152 noruntime : disable EFI runtime services support
1153 debug: enable misc debug output
1155 efi_no_storage_paranoia [EFI; X86]
1156 Using this parameter you can use more than 50% of
1157 your efi variable storage. Use this parameter only if
1158 you are really sure that your UEFI does sane gc and
1159 fulfills the spec otherwise your board may brick.
1161 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1162 Add arbitrary attribute to specific memory range by
1163 updating original EFI memory map.
1164 Region of memory which aa attribute is added to is
1166 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1167 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1168 attribute is added to range 0x100000000-0x180000000 and
1169 0x10a0000000-0x1120000000.
1171 Using this parameter you can do debugging of EFI memmap
1172 related feature. For example, you can do debugging of
1173 Address Range Mirroring feature even if your box
1176 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1177 that is to be dynamically loaded by Linux. If there are
1178 multiple variables with the same name but with different
1179 vendor GUIDs, all of them will be loaded. See
1180 Documentation/acpi/ssdt-overlays.txt for details.
1183 eisa_irq_edge= [PARISC,HW]
1184 See header of drivers/parisc/eisa.c.
1187 See comment before function elanfreq_setup() in
1188 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1191 Format: {"cfq" | "deadline" | "noop"}
1192 See Documentation/block/cfq-iosched.txt and
1193 Documentation/block/deadline-iosched.txt for details.
1195 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1196 Specifies physical address of start of kernel core
1197 image elf header and optionally the size. Generally
1198 kexec loader will pass this option to capture kernel.
1199 See Documentation/kdump/kdump.txt for details.
1201 enable_mtrr_cleanup [X86]
1202 The kernel tries to adjust MTRR layout from continuous
1203 to discrete, to make X server driver able to add WB
1204 entry later. This parameter enables that.
1206 enable_timer_pin_1 [X86]
1207 Enable PIN 1 of APIC timer
1208 Can be useful to work around chipset bugs
1209 (in particular on some ATI chipsets).
1210 The kernel tries to set a reasonable default.
1212 enforcing [SELINUX] Set initial enforcing status.
1214 See security/selinux/Kconfig help text.
1215 0 -- permissive (log only, no denials).
1216 1 -- enforcing (deny and log).
1218 Value can be changed at runtime via /selinux/enforce.
1221 Disable Error Record Serialization Table (ERST)
1224 ether= [HW,NET] Ethernet cards parameters
1225 This option is obsoleted by the "netdev=" option, which
1226 has equivalent usage. See its documentation for details.
1230 Permit 'security.evm' to be updated regardless of
1231 current integrity status.
1235 fail_make_request=[KNL]
1236 General fault injection mechanism.
1237 Format: <interval>,<probability>,<space>,<times>
1238 See also Documentation/fault-injection/.
1241 See Documentation/blockdev/floppy.txt.
1243 force_pal_cache_flush
1244 [IA-64] Avoid check_sal_cache_flush which may hang on
1245 buggy SAL_CACHE_FLUSH implementations. Using this
1246 parameter will force ia64_sal_cache_flush to call
1247 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1250 Forcefully enable Physical Address Extension (PAE).
1251 Many Pentium M systems disable PAE but may have a
1252 functionally usable PAE implementation.
1253 Warning: use of this parameter will taint the kernel
1254 and may cause unknown problems.
1257 [FTRACE] will set and start the specified tracer
1258 as early as possible in order to facilitate early
1261 ftrace_dump_on_oops[=orig_cpu]
1262 [FTRACE] will dump the trace buffers on oops.
1263 If no parameter is passed, ftrace will dump
1264 buffers of all CPUs, but if you pass orig_cpu, it will
1265 dump only the buffer of the CPU that triggered the
1268 ftrace_filter=[function-list]
1269 [FTRACE] Limit the functions traced by the function
1270 tracer at boot up. function-list is a comma separated
1271 list of functions. This list can be changed at run
1272 time by the set_ftrace_filter file in the debugfs
1275 ftrace_notrace=[function-list]
1276 [FTRACE] Do not trace the functions specified in
1277 function-list. This list can be changed at run time
1278 by the set_ftrace_notrace file in the debugfs
1281 ftrace_graph_filter=[function-list]
1282 [FTRACE] Limit the top level callers functions traced
1283 by the function graph tracer at boot up.
1284 function-list is a comma separated list of functions
1285 that can be changed at run time by the
1286 set_graph_function file in the debugfs tracing directory.
1288 ftrace_graph_notrace=[function-list]
1289 [FTRACE] Do not trace from the functions specified in
1290 function-list. This list is a comma separated list of
1291 functions that can be changed at run time by the
1292 set_graph_notrace file in the debugfs tracing directory.
1294 ftrace_graph_max_depth=<uint>
1295 [FTRACE] Used with the function graph tracer. This is
1296 the max depth it will trace into a function. This value
1297 can be changed at run time by the max_graph_depth file
1298 in the tracefs tracing directory. default: 0 (no limit)
1301 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1302 support via parallel port (up to 5 devices per port)
1303 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1304 See also Documentation/input/devices/joystick-parport.rst
1308 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1312 gcov_persist= [GCOV] When non-zero (default), profiling data for
1313 kernel modules is saved and remains accessible via
1314 debugfs, even when the module is unloaded/reloaded.
1315 When zero, profiling data is discarded and associated
1316 debugfs files are removed at module unload time.
1318 goldfish [X86] Enable the goldfish android emulator platform.
1319 Don't use this when you are not running on the
1322 gpt [EFI] Forces disk with valid GPT signature but
1323 invalid Protective MBR to be treated as GPT. If the
1324 primary GPT is corrupted, it enables the backup/alternate
1325 GPT to be used instead.
1327 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1328 the "Enable 0" bit of the configuration register.
1331 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1332 the "Enable 0" bit of the configuration register.
1335 grcan.select= [HW] Select which physical interface to use.
1338 grcan.txsize= [HW] Sets the size of the tx buffer.
1339 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1341 grcan.rxsize= [HW] Sets the size of the rx buffer.
1342 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1345 gpio-mockup.gpio_mockup_ranges
1346 [HW] Sets the ranges of gpiochip of for this device.
1347 Format: <start1>,<end1>,<start2>,<end2>...
1349 hardlockup_all_cpu_backtrace=
1350 [KNL] Should the hard-lockup detector generate
1351 backtraces on all cpus.
1354 hashdist= [KNL,NUMA] Large hashes allocated during boot
1355 are distributed across NUMA nodes. Defaults on
1356 for 64-bit NUMA, off otherwise.
1357 Format: 0 | 1 (for off | on)
1359 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1361 hd= [EIDE] (E)IDE hard drive subsystem geometry
1362 Format: <cyl>,<head>,<sect>
1365 Disable Hardware Error Source Table (HEST) support;
1366 corresponding firmware-first mode error processing
1367 logic will be disabled.
1369 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1370 size of <nn>. This works even on boxes that have no
1371 highmem otherwise. This also works to reduce highmem
1372 size on bigger boxes.
1374 highres= [KNL] Enable/disable high resolution timer mode.
1375 Valid parameters: "on", "off"
1379 See Documentation/isdn/README.HiSax.
1383 hpet= [X86-32,HPET] option to control HPET usage
1384 Format: { enable (default) | disable | force |
1386 disable: disable HPET and use PIT instead
1387 force: allow force enabled of undocumented chips (ICH4,
1389 verbose: show contents of HPET registers during setup
1391 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1392 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1394 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1395 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1396 On x86-64 and powerpc, this option can be specified
1397 multiple times interleaved with hugepages= to reserve
1398 huge pages of different sizes. Valid pages sizes on
1399 x86-64 are 2M (when the CPU supports "pse") and 1G
1400 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1403 [KNL] Should the hung task detector generate panics.
1406 A nonzero value instructs the kernel to panic when a
1407 hung task is detected. The default value is controlled
1408 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1409 option. The value selected by this boot parameter can
1410 be changed later by the kernel.hung_task_panic sysctl.
1412 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1413 terminal devices. Valid values: 0..8
1414 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1415 If specified, z/VM IUCV HVC accepts connections
1416 from listed z/VM user IDs only.
1418 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1419 which allow the hypervisor to 'idle' the
1420 guest on lock contention.
1423 Do not unregister boot console at start. This is only
1424 useful for debugging when something happens in the window
1425 between unregistering the boot console and initializing
1428 i2c_bus= [HW] Override the default board specific I2C bus speed
1429 or register an additional I2C bus that is not
1430 registered from board initialization code.
1434 i8042.debug [HW] Toggle i8042 debug mode
1435 i8042.unmask_kbd_data
1436 [HW] Enable printing of interrupt data from the KBD port
1437 (disabled by default, and as a pre-condition
1438 requires that i8042.debug=1 be enabled)
1439 i8042.direct [HW] Put keyboard port into non-translated mode
1440 i8042.dumbkbd [HW] Pretend that controller can only read data from
1441 keyboard and cannot control its state
1442 (Don't attempt to blink the leds)
1443 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1444 i8042.nokbd [HW] Don't check/create keyboard port
1445 i8042.noloop [HW] Disable the AUX Loopback command while probing
1447 i8042.nomux [HW] Don't check presence of an active multiplexing
1449 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1451 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1452 i8042.reset [HW] Reset the controller during init, cleanup and
1453 suspend-to-ram transitions, only during s2r
1454 transitions, or never reset
1455 Format: { 1 | Y | y | 0 | N | n }
1456 1, Y, y: always reset controller
1457 0, N, n: don't ever reset controller
1458 Default: only on s2r transitions on x86; most other
1459 architectures force reset to be always executed
1460 i8042.unlock [HW] Unlock (ignore) the keylock
1461 i8042.kbdreset [HW] Reset device connected to KBD port
1465 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1466 indicates that the driver is running on unsupported
1468 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1469 does not match list of supported models.
1471 [HW] Report power status in /proc/i8k
1472 (disabled by default)
1473 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1476 i915.invert_brightness=
1477 [DRM] Invert the sense of the variable that is used to
1478 set the brightness of the panel backlight. Normally a
1479 brightness value of 0 indicates backlight switched off,
1480 and the maximum of the brightness value sets the backlight
1481 to maximum brightness. If this parameter is set to 0
1482 (default) and the machine requires it, or this parameter
1483 is set to 1, a brightness value of 0 sets the backlight
1484 to maximum brightness, and the maximum of the brightness
1485 value switches the backlight off.
1486 -1 -- never invert brightness
1487 0 -- machine default
1488 1 -- force brightness inversion
1491 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1493 ide-core.nodma= [HW] (E)IDE subsystem
1494 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1495 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1496 .cdrom .chs .ignore_cable are additional options
1497 See Documentation/ide/ide.txt.
1499 ide-generic.probe-mask= [HW] (E)IDE subsystem
1501 Probe mask for legacy ISA IDE ports. Depending on
1502 platform up to 6 ports are supported, enabled by
1503 setting corresponding bits in the mask to 1. The
1504 default value is 0x0, which has a special meaning.
1505 On systems that have PCI, it triggers scanning the
1506 PCI bus for the first and the second port, which
1507 are then probed. On systems without PCI the value
1508 of 0x0 enables probing the two first ports as if it
1511 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1512 Claim all unknown PCI IDE storage controllers.
1515 Format: idle=poll, idle=halt, idle=nomwait
1516 Poll forces a polling idle loop that can slightly
1517 improve the performance of waking up a idle CPU, but
1518 will use a lot of power and make the system run hot.
1520 idle=halt: Halt is forced to be used for CPU idle.
1521 In such case C2/C3 won't be used again.
1522 idle=nomwait: Disable mwait for CPU C-states
1524 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1525 Format: { strict | legacy | 2008 | relaxed }
1528 Choose which programs will be accepted for execution
1529 based on the IEEE 754 NaN encoding(s) supported by
1530 the FPU and the NaN encoding requested with the value
1531 of an ELF file header flag individually set by each
1532 binary. Hardware implementations are permitted to
1533 support either or both of the legacy and the 2008 NaN
1536 Available settings are as follows:
1537 strict accept binaries that request a NaN encoding
1538 supported by the FPU
1539 legacy only accept legacy-NaN binaries, if supported
1541 2008 only accept 2008-NaN binaries, if supported
1543 relaxed accept any binaries regardless of whether
1544 supported by the FPU
1546 The FPU emulator is always able to support both NaN
1547 encodings, so if no FPU hardware is present or it has
1548 been disabled with 'nofpu', then the settings of
1549 'legacy' and '2008' strap the emulator accordingly,
1550 'relaxed' straps the emulator for both legacy-NaN and
1551 2008-NaN, whereas 'strict' enables legacy-NaN only on
1552 legacy processors and both NaN encodings on MIPS32 or
1555 The setting for ABS.fmt/NEG.fmt instruction execution
1556 mode generally follows that for the NaN encoding,
1557 except where unsupported by hardware.
1559 ignore_loglevel [KNL]
1560 Ignore loglevel setting - this will print /all/
1561 kernel messages to the console. Useful for debugging.
1562 We also add it as printk module parameter, so users
1563 could change it dynamically, usually by
1564 /sys/module/printk/parameters/ignore_loglevel.
1567 Ignore RLIMIT_DATA setting for data mappings,
1568 print warning at first misuse. Can be changed via
1569 /sys/module/kernel/parameters/ignore_rlimit_data.
1571 ihash_entries= [KNL]
1572 Set number of hash buckets for inode cache.
1574 ima_appraise= [IMA] appraise integrity measurements
1575 Format: { "off" | "enforce" | "fix" | "log" }
1578 ima_appraise_tcb [IMA]
1579 The builtin appraise policy appraises all files
1582 ima_canonical_fmt [IMA]
1583 Use the canonical format for the binary runtime
1584 measurements, instead of host native format.
1587 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1591 The list of supported hash algorithms is defined
1592 in crypto/hash_info.h.
1595 The builtin policies to load during IMA setup.
1596 Format: "tcb | appraise_tcb | secure_boot |
1599 The "tcb" policy measures all programs exec'd, files
1600 mmap'd for exec, and all files opened with the read
1601 mode bit set by either the effective uid (euid=0) or
1604 The "appraise_tcb" policy appraises the integrity of
1605 all files owned by root. (This is the equivalent
1606 of ima_appraise_tcb.)
1608 The "secure_boot" policy appraises the integrity
1609 of files (eg. kexec kernel image, kernel modules,
1610 firmware, policy, etc) based on file signatures.
1612 The "fail_securely" policy forces file signature
1613 verification failure also on privileged mounted
1614 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1617 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1618 Load a policy which meets the needs of the Trusted
1619 Computing Base. This means IMA will measure all
1620 programs exec'd, files mmap'd for exec, and all files
1621 opened for read by uid=0.
1624 Select one of defined IMA measurements template formats.
1625 Formats: { "ima" | "ima-ng" | "ima-sig" }
1629 [IMA] Define a custom template format.
1630 Format: { "field1|...|fieldN" }
1632 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1633 Format: <min_file_size>
1634 Set the minimal file size for using asynchronous hash.
1635 If left unspecified, ahash usage is disabled.
1637 ahash performance varies for different data sizes on
1638 different crypto accelerators. This option can be used
1639 to achieve the best performance for a particular HW.
1641 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1643 Set hashing buffer size. Default: 4k.
1645 ahash performance varies for different chunk sizes on
1646 different crypto accelerators. This option can be used
1647 to achieve best performance for particular HW.
1651 Run specified binary instead of /sbin/init as init
1654 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1655 for working out where the kernel is dying during
1658 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1659 initcall functions. Useful for debugging built-in
1660 modules and initcalls.
1662 initrd= [BOOT] Specify the location of the initial ramdisk
1664 init_pkru= [x86] Specify the default memory protection keys rights
1665 register contents for all processes. 0x55555554 by
1666 default (disallow access to all but pkey 0). Can
1667 override in debugfs after boot.
1669 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1672 int_pln_enable [x86] Enable power limit notification interrupt
1674 integrity_audit=[IMA]
1675 Format: { "0" | "1" }
1676 0 -- basic integrity auditing messages. (Default)
1677 1 -- additional integrity auditing messages.
1679 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1681 Enable intel iommu driver.
1683 Disable intel iommu driver.
1684 igfx_off [Default Off]
1685 By default, gfx is mapped as normal device. If a gfx
1686 device has a dedicated DMAR unit, the DMAR unit is
1687 bypassed by not enabling DMAR with this option. In
1688 this case, gfx device will use physical address for
1691 With this option iommu will not optimize to look
1692 for io virtual address below 32-bit forcing dual
1693 address cycle on pci bus for cards supporting greater
1694 than 32-bit addressing. The default is to look
1695 for translation below 32-bit and if not available
1696 then look in the higher range.
1697 strict [Default Off]
1698 With this option on every unmap_single operation will
1699 result in a hardware IOTLB flush operation as opposed
1700 to batching them for performance.
1701 sp_off [Default Off]
1702 By default, super page will be supported if Intel IOMMU
1703 has the capability. With this option, super page will
1705 sm_off [Default Off]
1706 By default, scalable mode will be supported if the
1707 hardware advertises that it has support for the scalable
1708 mode translation. With this option set, scalable mode
1709 will not be used even on hardware which claims to support
1711 tboot_noforce [Default Off]
1712 Do not force the Intel IOMMU enabled under tboot.
1713 By default, tboot will force Intel IOMMU on, which
1714 could harm performance of some high-throughput
1715 devices like 40GBit network cards, even if identity
1717 Note that using this option lowers the security
1718 provided by tboot because it makes the system
1719 vulnerable to DMA attacks.
1721 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1722 0 disables intel_idle and fall back on acpi_idle.
1723 1 to 9 specify maximum depth of C-state.
1727 Do not enable intel_pstate as the default
1728 scaling driver for the supported processors
1730 Use intel_pstate as a scaling driver, but configure it
1731 to work with generic cpufreq governors (instead of
1732 enabling its internal governor). This mode cannot be
1733 used along with the hardware-managed P-states (HWP)
1736 Enable intel_pstate on systems that prohibit it by default
1737 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1738 instead of acpi-cpufreq may disable platform features, such
1739 as thermal controls and power capping, that rely on ACPI
1740 P-States information being indicated to OSPM and therefore
1741 should be used with caution. This option does not work with
1742 processors that aren't supported by the intel_pstate driver
1743 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1745 Do not enable hardware P state control (HWP)
1748 Only load intel_pstate on systems which support
1749 hardware P state control (HWP) if available.
1751 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1752 Description Table, specifies preferred power management
1753 profile as "Enterprise Server" or "Performance Server",
1754 then this feature is turned on by default.
1756 Allow per-logical-CPU P-State performance control limits using
1757 cpufreq sysfs interface
1759 intremap= [X86-64, Intel-IOMMU]
1760 on enable Interrupt Remapping (default)
1761 off disable Interrupt Remapping
1762 nosid disable Source ID checking
1764 BIOS x2APIC opt-out request will be ignored
1765 nopost disable Interrupt Posting
1767 iomem= Disable strict checking of access to MMIO memory
1768 strict regions from userspace.
1783 nobypass [PPC/POWERNV]
1784 Disable IOMMU bypass, using IOMMU for PCI devices.
1786 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1787 Format: { "0" | "1" }
1789 Request that DMA unmap operations use deferred
1790 invalidation of hardware TLBs, for increased
1791 throughput at the cost of reduced device isolation.
1792 Will fall back to strict mode if not supported by
1793 the relevant IOMMU driver.
1794 1 - Strict mode (default).
1795 DMA unmap operations invalidate IOMMU hardware TLBs
1799 [ARM64] Configure DMA to bypass the IOMMU by default.
1800 Format: { "0" | "1" }
1801 0 - Use IOMMU translation for DMA.
1802 1 - Bypass the IOMMU for DMA.
1803 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
1805 io7= [HW] IO7 for Marvel based alpha systems
1806 See comment before marvel_specify_io7 in
1807 arch/alpha/kernel/core_marvel.c.
1809 io_delay= [X86] I/O delay method
1811 Standard port 0x80 based delay
1813 Alternate port 0xed based delay (needed on some systems)
1815 Simple two microseconds delay
1820 See Documentation/filesystems/nfs/nfsroot.txt.
1822 irqaffinity= [SMP] Set the default irq affinity mask
1823 The argument is a cpu list, as described above.
1825 irqchip.gicv2_force_probe=
1828 Force the kernel to look for the second 4kB page
1829 of a GICv2 controller even if the memory range
1830 exposed by the device tree is too small.
1832 irqchip.gicv3_nolpi=
1834 Force the kernel to ignore the availability of
1835 LPIs (and by consequence ITSs). Intended for system
1836 that use the kernel as a bootloader, and thus want
1837 to let secondary kernels in charge of setting up
1841 When an interrupt is not handled search all handlers
1842 for it. Intended to get systems with badly broken
1846 When an interrupt is not handled search all handlers
1847 for it. Also check all handlers each timer
1848 interrupt. Intended to get systems with badly broken
1852 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1854 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1855 [Deprecated - use cpusets instead]
1856 Format: [flag-list,]<cpu-list>
1858 Specify one or more CPUs to isolate from disturbances
1859 specified in the flag list (default: domain):
1862 Disable the tick when a single task runs.
1864 A residual 1Hz tick is offloaded to workqueues, which you
1865 need to affine to housekeeping through the global
1866 workqueue's affinity configured via the
1867 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1868 by using the 'domain' flag described below.
1870 NOTE: by default the global workqueue runs on all CPUs,
1871 so to protect individual CPUs the 'cpumask' file has to
1872 be configured manually after bootup.
1875 Isolate from the general SMP balancing and scheduling
1876 algorithms. Note that performing domain isolation this way
1877 is irreversible: it's not possible to bring back a CPU to
1878 the domains once isolated through isolcpus. It's strongly
1879 advised to use cpusets instead to disable scheduler load
1880 balancing through the "cpuset.sched_load_balance" file.
1881 It offers a much more flexible interface where CPUs can
1882 move in and out of an isolated set anytime.
1884 You can move a process onto or off an "isolated" CPU via
1885 the CPU affinity syscalls or cpuset.
1886 <cpu number> begins at 0 and the maximum value is
1887 "number of CPUs in system - 1".
1889 The format of <cpu-list> is described above.
1895 ivrs_ioapic [HW,X86_64]
1896 Provide an override to the IOAPIC-ID<->DEVICE-ID
1897 mapping provided in the IVRS ACPI table. For
1898 example, to map IOAPIC-ID decimal 10 to
1899 PCI device 00:14.0 write the parameter as:
1900 ivrs_ioapic[10]=00:14.0
1902 ivrs_hpet [HW,X86_64]
1903 Provide an override to the HPET-ID<->DEVICE-ID
1904 mapping provided in the IVRS ACPI table. For
1905 example, to map HPET-ID decimal 0 to
1906 PCI device 00:14.0 write the parameter as:
1907 ivrs_hpet[0]=00:14.0
1909 ivrs_acpihid [HW,X86_64]
1910 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1911 mapping provided in the IVRS ACPI table. For
1912 example, to map UART-HID:UID AMD0020:0 to
1913 PCI device 00:14.5 write the parameter as:
1914 ivrs_acpihid[00:14.5]=AMD0020:0
1916 js= [HW,JOY] Analog joystick
1917 See Documentation/input/joydev/joystick.rst.
1920 When CONFIG_RANDOMIZE_BASE is set, this disables
1921 kernel and module base offset ASLR (Address Space
1922 Layout Randomization).
1925 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
1926 report on every invalid memory access. Without this
1927 parameter KASAN will print report only for the first
1932 kernelcore= [KNL,X86,IA-64,PPC]
1933 Format: nn[KMGTPE] | nn% | "mirror"
1934 This parameter specifies the amount of memory usable by
1935 the kernel for non-movable allocations. The requested
1936 amount is spread evenly throughout all nodes in the
1937 system as ZONE_NORMAL. The remaining memory is used for
1938 movable memory in its own zone, ZONE_MOVABLE. In the
1939 event, a node is too small to have both ZONE_NORMAL and
1940 ZONE_MOVABLE, kernelcore memory will take priority and
1941 other nodes will have a larger ZONE_MOVABLE.
1943 ZONE_MOVABLE is used for the allocation of pages that
1944 may be reclaimed or moved by the page migration
1945 subsystem. Note that allocations like PTEs-from-HighMem
1946 still use the HighMem zone if it exists, and the Normal
1947 zone if it does not.
1949 It is possible to specify the exact amount of memory in
1950 the form of "nn[KMGTPE]", a percentage of total system
1951 memory in the form of "nn%", or "mirror". If "mirror"
1952 option is specified, mirrored (reliable) memory is used
1953 for non-movable allocations and remaining memory is used
1954 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
1955 are exclusive, so you cannot specify multiple forms.
1957 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1958 Format: <Controller#>[,poll interval]
1959 The controller # is the number of the ehci usb debug
1960 port as it is probed via PCI. The poll interval is
1961 optional and is the number seconds in between
1962 each poll cycle to the debug port in case you need
1963 the functionality for interrupting the kernel with
1964 gdb or control-c on the dbgp connection. When
1965 not using this parameter you use sysrq-g to break into
1966 the kernel debugger.
1968 kgdboc= [KGDB,HW] kgdb over consoles.
1969 Requires a tty driver that supports console polling,
1970 or a supported polling keyboard driver (non-usb).
1971 Serial only format: <serial_device>[,baud]
1972 keyboard only format: kbd
1973 keyboard and serial format: kbd,<serial_device>[,baud]
1974 Optional Kernel mode setting:
1975 kms, kbd format: kms,kbd
1976 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1978 kgdbwait [KGDB] Stop kernel execution and enter the
1979 kernel debugger at the earliest opportunity.
1981 kmac= [MIPS] korina ethernet MAC address.
1982 Configure the RouterBoard 532 series on-chip
1983 Ethernet adapter MAC address.
1985 kmemleak= [KNL] Boot-time kmemleak enable/disable
1986 Valid arguments: on, off
1988 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1991 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1992 Default is 0 (don't ignore, but inject #GP)
1994 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
1995 Default is false (don't support).
1997 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2001 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2002 Default is 1 (enabled)
2004 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2006 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2008 kvm-arm.vgic_v3_group0_trap=
2009 [KVM,ARM] Trap guest accesses to GICv3 group-0
2012 kvm-arm.vgic_v3_group1_trap=
2013 [KVM,ARM] Trap guest accesses to GICv3 group-1
2016 kvm-arm.vgic_v3_common_trap=
2017 [KVM,ARM] Trap guest accesses to GICv3 common
2020 kvm-arm.vgic_v4_enable=
2021 [KVM,ARM] Allow use of GICv4 for direct injection of
2024 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2025 (virtualized MMU) support on capable Intel chips.
2026 Default is 1 (enabled)
2028 kvm-intel.emulate_invalid_guest_state=
2029 [KVM,Intel] Enable emulation of invalid guest states
2030 Default is 0 (disabled)
2032 kvm-intel.flexpriority=
2033 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2034 Default is 1 (enabled)
2037 [KVM,Intel] Enable VMX nesting (nVMX).
2038 Default is 0 (disabled)
2040 kvm-intel.unrestricted_guest=
2041 [KVM,Intel] Disable unrestricted guest feature
2042 (virtualized real and unpaged mode) on capable
2043 Intel chips. Default is 1 (enabled)
2045 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2048 Valid arguments: never, cond, always
2050 always: L1D cache flush on every VMENTER.
2051 cond: Flush L1D on VMENTER only when the code between
2052 VMEXIT and VMENTER can leak host memory.
2053 never: Disables the mitigation
2055 Default is cond (do L1 cache flush in specific instances)
2057 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2058 feature (tagged TLBs) on capable Intel chips.
2059 Default is 1 (enabled)
2061 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2064 The kernel PTE inversion protection is unconditionally
2065 enabled and cannot be disabled.
2068 Provides all available mitigations for the
2069 L1TF vulnerability. Disables SMT and
2070 enables all mitigations in the
2071 hypervisors, i.e. unconditional L1D flush.
2073 SMT control and L1D flush control via the
2074 sysfs interface is still possible after
2075 boot. Hypervisors will issue a warning
2076 when the first VM is started in a
2077 potentially insecure configuration,
2078 i.e. SMT enabled or L1D flush disabled.
2081 Same as 'full', but disables SMT and L1D
2082 flush runtime control. Implies the
2083 'nosmt=force' command line option.
2084 (i.e. sysfs control of SMT is disabled.)
2087 Leaves SMT enabled and enables the default
2088 hypervisor mitigation, i.e. conditional
2091 SMT control and L1D flush control via the
2092 sysfs interface is still possible after
2093 boot. Hypervisors will issue a warning
2094 when the first VM is started in a
2095 potentially insecure configuration,
2096 i.e. SMT enabled or L1D flush disabled.
2100 Disables SMT and enables the default
2101 hypervisor mitigation.
2103 SMT control and L1D flush control via the
2104 sysfs interface is still possible after
2105 boot. Hypervisors will issue a warning
2106 when the first VM is started in a
2107 potentially insecure configuration,
2108 i.e. SMT enabled or L1D flush disabled.
2111 Same as 'flush', but hypervisors will not
2112 warn when a VM is started in a potentially
2113 insecure configuration.
2116 Disables hypervisor mitigations and doesn't
2118 It also drops the swap size and available
2119 RAM limit restriction on both hypervisor and
2124 For details see: Documentation/admin-guide/l1tf.rst
2130 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2133 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2134 value for LAPIC timer one-shot implementation. Default
2135 back to the programmable timer unit in the LAPIC.
2137 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2140 libata.dma= [LIBATA] DMA control
2141 libata.dma=0 Disable all PATA and SATA DMA
2142 libata.dma=1 PATA and SATA Disk DMA only
2143 libata.dma=2 ATAPI (CDROM) DMA only
2144 libata.dma=4 Compact Flash DMA only
2145 Combinations also work, so libata.dma=3 enables DMA
2146 for disks and CDROMs, but not CFs.
2148 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2149 libata.ignore_hpa=0 keep BIOS limits (default)
2150 libata.ignore_hpa=1 ignore limits, using full disk
2152 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2156 libata.force= [LIBATA] Force configurations. The format is comma
2157 separated list of "[ID:]VAL" where ID is
2158 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2159 matching port, link or device. Basically, it matches
2160 the ATA ID string printed on console by libata. If
2161 the whole ID part is omitted, the last PORT and DEVICE
2162 values are used. If ID hasn't been specified yet, the
2163 configuration applies to all ports, links and devices.
2165 If only DEVICE is omitted, the parameter applies to
2166 the port and all links and devices behind it. DEVICE
2167 number of 0 either selects the first device or the
2168 first fan-out link behind PMP device. It does not
2169 select the host link. DEVICE number of 15 selects the
2170 host link and device attached to it.
2172 The VAL specifies the configuration to force. As long
2173 as there's no ambiguity shortcut notation is allowed.
2174 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2175 The following configurations can be forced.
2177 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2178 Any ID with matching PORT is used.
2180 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2182 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2183 udma[/][16,25,33,44,66,100,133] notation is also
2186 * [no]ncq: Turn on or off NCQ.
2188 * [no]ncqtrim: Turn off queued DSM TRIM.
2190 * nohrst, nosrst, norst: suppress hard, soft
2193 * rstonce: only attempt one reset during
2194 hot-unplug link recovery
2196 * dump_id: dump IDENTIFY data.
2198 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2200 * disable: Disable this device.
2202 If there are multiple matching configurations changing
2203 the same attribute, the last one is used.
2205 memblock=debug [KNL] Enable memblock debug messages.
2207 load_ramdisk= [RAM] List of ramdisks to load from floppy
2208 See Documentation/blockdev/ramdisk.txt.
2210 lockd.nlm_grace_period=P [NFS] Assign grace period.
2213 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2216 lockd.nlm_timeout=T [NFS] Assign timeout value.
2219 lockd.nlm_udpport=M [NFS] Assign UDP port.
2222 locktorture.nreaders_stress= [KNL]
2223 Set the number of locking read-acquisition kthreads.
2224 Defaults to being automatically set based on the
2225 number of online CPUs.
2227 locktorture.nwriters_stress= [KNL]
2228 Set the number of locking write-acquisition kthreads.
2230 locktorture.onoff_holdoff= [KNL]
2231 Set time (s) after boot for CPU-hotplug testing.
2233 locktorture.onoff_interval= [KNL]
2234 Set time (s) between CPU-hotplug operations, or
2235 zero to disable CPU-hotplug testing.
2237 locktorture.shuffle_interval= [KNL]
2238 Set task-shuffle interval (jiffies). Shuffling
2239 tasks allows some CPUs to go into dyntick-idle
2240 mode during the locktorture test.
2242 locktorture.shutdown_secs= [KNL]
2243 Set time (s) after boot system shutdown. This
2244 is useful for hands-off automated testing.
2246 locktorture.stat_interval= [KNL]
2247 Time (s) between statistics printk()s.
2249 locktorture.stutter= [KNL]
2250 Time (s) to stutter testing, for example,
2251 specifying five seconds causes the test to run for
2252 five seconds, wait for five seconds, and so on.
2253 This tests the locking primitive's ability to
2254 transition abruptly to and from idle.
2256 locktorture.torture_type= [KNL]
2257 Specify the locking implementation to test.
2259 locktorture.verbose= [KNL]
2260 Enable additional printk() statements.
2262 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2265 loglevel= All Kernel Messages with a loglevel smaller than the
2266 console loglevel will be printed to the console. It can
2267 also be changed with klogd or other programs. The
2268 loglevels are defined as follows:
2270 0 (KERN_EMERG) system is unusable
2271 1 (KERN_ALERT) action must be taken immediately
2272 2 (KERN_CRIT) critical conditions
2273 3 (KERN_ERR) error conditions
2274 4 (KERN_WARNING) warning conditions
2275 5 (KERN_NOTICE) normal but significant condition
2276 6 (KERN_INFO) informational
2277 7 (KERN_DEBUG) debug-level messages
2279 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2280 in bytes. n must be a power of two and greater
2281 than the minimal size. The minimal size is defined
2282 by LOG_BUF_SHIFT kernel config parameter. There is
2283 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2284 that allows to increase the default size depending on
2285 the number of CPUs. See init/Kconfig for more details.
2287 logo.nologo [FB] Disables display of the built-in Linux logo.
2288 This may be used to provide more screen space for
2289 kernel log messages and is useful when debugging
2290 kernel boot problems.
2292 lp=0 [LP] Specify parallel ports to use, e.g,
2293 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2294 lp=reset first parallel port). 'lp=0' disables the
2295 lp=auto printer driver. 'lp=reset' (which can be
2296 specified in addition to the ports) causes
2297 attached printers to be reset. Using
2298 lp=port1,port2,... specifies the parallel ports
2299 to associate lp devices with, starting with
2300 lp0. A port specification may be 'none' to skip
2301 that lp device, or a parport name such as
2302 'parport0'. Specifying 'lp=auto' instead of a
2303 port specification list means that device IDs
2304 from each port should be examined, to see if
2305 an IEEE 1284-compliant printer is attached; if
2306 so, the driver will manage that printer.
2307 See also header of drivers/char/lp.c.
2310 Sets loops_per_jiffy to given constant, thus avoiding
2311 time-consuming boot-time autodetection (up to 250 ms per
2312 CPU). 0 enables autodetection (default). To determine
2313 the correct value for your kernel, boot with normal
2314 autodetection and see what value is printed. Note that
2315 on SMP systems the preset will be applied to all CPUs,
2316 which is likely to cause problems if your CPUs need
2317 significantly divergent settings. An incorrect value
2318 will cause delays in the kernel to be wrong, leading to
2319 unpredictable I/O errors and other breakage. Although
2320 unlikely, in the extreme case this might damage your
2324 Format: <io>,<irq>,<dma>
2326 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2328 machvec= [IA-64] Force the use of a particular machine-vector
2329 (machvec) in a generic kernel.
2330 Example: machvec=hpzx1_swiotlb
2332 machtype= [Loongson] Share the same kernel image file between different
2334 Example: machtype=lemote-yeeloong-2f-7inch
2336 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2337 than or equal to this physical address is ignored.
2339 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2340 will bring up during bootup. maxcpus=n : n >= 0 limits
2341 the kernel to bring up 'n' processors. Surely after
2342 bootup you can bring up the other plugged cpu by executing
2343 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2344 only takes effect during system bootup.
2345 While n=0 is a special case, it is equivalent to "nosmp",
2346 which also disables the IO APIC.
2348 max_loop= [LOOP] The number of loop block devices that get
2349 (loop.max_loop) unconditionally pre-created at init time. The default
2350 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2351 of statically allocating a predefined number, loop
2352 devices can be requested on-demand with the
2353 /dev/loop-control interface.
2355 mce [X86-32] Machine Check Exception
2357 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2359 md= [HW] RAID subsystems devices and level
2360 See Documentation/admin-guide/md.rst.
2363 Format: <first>,<last>
2364 Specifies range of consoles to be captured by the MDA.
2366 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2367 Amount of memory to be used when the kernel is not able
2368 to see the whole system memory or for test.
2369 [X86] Work as limiting max address. Use together
2370 with memmap= to avoid physical address space collisions.
2371 Without memmap= PCI devices could be placed at addresses
2372 belonging to unused RAM.
2374 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2378 [KNL,SH] Allow user to override the default size for
2379 per-device physically contiguous DMA buffers.
2381 memhp_default_state=online/offline
2382 [KNL] Set the initial state for the memory hotplug
2383 onlining policy. If not specified, the default value is
2384 set according to the
2385 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2387 See Documentation/memory-hotplug.txt.
2389 memmap=exactmap [KNL,X86] Enable setting of an exact
2390 E820 memory map, as specified by the user.
2391 Such memmap=exactmap lines can be constructed based on
2392 BIOS output or other requirements. See the memmap=nn@ss
2395 memmap=nn[KMG]@ss[KMG]
2396 [KNL] Force usage of a specific region of memory.
2397 Region of memory to be used is from ss to ss+nn.
2398 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2399 which limits max address to nn[KMG].
2400 Multiple different regions can be specified,
2403 memmap=100M@2G,100M#3G,1G!1024G
2405 memmap=nn[KMG]#ss[KMG]
2406 [KNL,ACPI] Mark specific memory as ACPI data.
2407 Region of memory to be marked is from ss to ss+nn.
2409 memmap=nn[KMG]$ss[KMG]
2410 [KNL,ACPI] Mark specific memory as reserved.
2411 Region of memory to be reserved is from ss to ss+nn.
2412 Example: Exclude memory from 0x18690000-0x1869ffff
2413 memmap=64K$0x18690000
2415 memmap=0x10000$0x18690000
2416 Some bootloaders may need an escape character before '$',
2417 like Grub2, otherwise '$' and the following number
2420 memmap=nn[KMG]!ss[KMG]
2421 [KNL,X86] Mark specific memory as protected.
2422 Region of memory to be used, from ss to ss+nn.
2423 The memory region may be marked as e820 type 12 (0xc)
2424 and is NVDIMM or ADR memory.
2426 memmap=<size>%<offset>-<oldtype>+<newtype>
2427 [KNL,ACPI] Convert memory within the specified region
2428 from <oldtype> to <newtype>. If "-<oldtype>" is left
2429 out, the whole region will be marked as <newtype>,
2430 even if previously unavailable. If "+<newtype>" is left
2431 out, matching memory will be removed. Types are
2432 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2433 3 = ACPI, 12 = PRAM.
2435 memory_corruption_check=0/1 [X86]
2436 Some BIOSes seem to corrupt the first 64k of
2437 memory when doing things like suspend/resume.
2438 Setting this option will scan the memory
2439 looking for corruption. Enabling this will
2440 both detect corruption and prevent the kernel
2441 from using the memory being corrupted.
2442 However, its intended as a diagnostic tool; if
2443 repeatable BIOS-originated corruption always
2444 affects the same memory, you can use memmap=
2445 to prevent the kernel from using that memory.
2447 memory_corruption_check_size=size [X86]
2448 By default it checks for corruption in the low
2449 64k, making this memory unavailable for normal
2450 use. Use this parameter to scan for
2451 corruption in more or less memory.
2453 memory_corruption_check_period=seconds [X86]
2454 By default it checks for corruption every 60
2455 seconds. Use this parameter to check at some
2456 other rate. 0 disables periodic checking.
2458 memtest= [KNL,X86,ARM,PPC] Enable memtest
2460 default : 0 <disable>
2461 Specifies the number of memtest passes to be
2462 performed. Each pass selects another test
2463 pattern from a given set of patterns. Memtest
2464 fills the memory with this pattern, validates
2465 memory contents and reserves bad memory
2466 regions that are detected.
2468 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2469 Valid arguments: on, off
2470 Default (depends on kernel configuration option):
2471 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2472 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2473 mem_encrypt=on: Activate SME
2474 mem_encrypt=off: Do not activate SME
2476 Refer to Documentation/x86/amd-memory-encryption.txt
2477 for details on when memory encryption can be activated.
2479 mem_sleep_default= [SUSPEND] Default system suspend mode:
2480 s2idle - Suspend-To-Idle
2481 shallow - Power-On Suspend or equivalent (if supported)
2482 deep - Suspend-To-RAM or equivalent (if supported)
2483 See Documentation/admin-guide/pm/sleep-states.rst.
2485 meye.*= [HW] Set MotionEye Camera parameters
2486 See Documentation/media/v4l-drivers/meye.rst.
2488 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2489 Multi-Function General Purpose Timers on AMD Geode
2492 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2493 the BIOS has incorrectly applied a workaround. TinyBIOS
2494 version 0.98 is known to be affected, 0.99 fixes the
2495 problem by letting the user disable the workaround.
2499 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2500 physical address is ignored.
2502 mini2440= [ARM,HW,KNL]
2503 Format:[0..2][b][c][t]
2505 MINI2440 configuration specification:
2506 0 - The attached screen is the 3.5" TFT
2507 1 - The attached screen is the 7" TFT
2508 2 - The VGA Shield is attached (1024x768)
2509 Leaving out the screen size parameter will not load
2510 the TFT driver, and the framebuffer will be left
2512 b - Enable backlight. The TFT backlight pin will be
2513 linked to the kernel VESA blanking code and a GPIO
2514 LED. This parameter is not necessary when using the
2516 c - Enable the s3c camera interface.
2517 t - Reserved for enabling touchscreen support. The
2518 touchscreen support is not enabled in the mainstream
2519 kernel as of 2.6.30, a preliminary port can be found
2520 in the "bleeding edge" mini2440 support kernel at
2521 http://repo.or.cz/w/linux-2.6/mini2440.git
2524 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2525 parameter allows control of the logging verbosity for
2526 the additional memory initialisation checks. A value
2527 of 0 disables mminit logging and a level of 4 will
2528 log everything. Information is printed at KERN_DEBUG
2529 so loglevel=8 may also need to be specified.
2532 [KNL] When CONFIG_MODULE_SIG is set, this means that
2533 modules without (valid) signatures will fail to load.
2534 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2535 is always true, so this option does nothing.
2537 module_blacklist= [KNL] Do not load a comma-separated list of
2538 modules. Useful for debugging problem modules.
2541 [MOUSE] Maximum time between finger touching and
2542 leaving touchpad surface for touch to be considered
2543 a tap and be reported as a left button click (for
2544 touchpads working in absolute mode only).
2546 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2547 reporting absolute coordinates, such as tablets
2548 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2549 reporting absolute coordinates, such as tablets
2551 movablecore= [KNL,X86,IA-64,PPC]
2552 Format: nn[KMGTPE] | nn%
2553 This parameter is the complement to kernelcore=, it
2554 specifies the amount of memory used for migratable
2555 allocations. If both kernelcore and movablecore is
2556 specified, then kernelcore will be at *least* the
2557 specified value but may be more. If movablecore on its
2558 own is specified, the administrator must be careful
2559 that the amount of memory usable for all allocations
2562 movable_node [KNL] Boot-time switch to make hotplugable memory
2563 NUMA nodes to be movable. This means that the memory
2564 of such nodes will be usable only for movable
2565 allocations which rules out almost all kernel
2566 allocations. Use with caution!
2568 MTD_Partition= [MTD]
2569 Format: <name>,<region-number>,<size>,<offset>
2571 MTD_Region= [MTD] Format:
2572 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2575 See drivers/mtd/cmdlinepart.c.
2577 multitce=off [PPC] This parameter disables the use of the pSeries
2578 firmware feature for updating multiple TCE entries
2581 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2583 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2585 boundary - index of last SLC block on Flex-OneNAND.
2586 The remaining blocks are configured as MLC blocks.
2587 lock - Configure if Flex-OneNAND boundary should be locked.
2588 Once locked, the boundary cannot be changed.
2589 1 indicates lock status, 0 indicates unlock status.
2592 ARM/S3C2412 JIVE boot control
2594 See arch/arm/mach-s3c2412/mach-jive.c
2596 mtouchusb.raw_coordinates=
2597 [HW] Make the MicroTouch USB driver use raw coordinates
2598 ('y', default) or cooked coordinates ('n')
2600 mtrr_chunk_size=nn[KMG] [X86]
2601 used for mtrr cleanup. It is largest continuous chunk
2602 that could hold holes aka. UC entries.
2604 mtrr_gran_size=nn[KMG] [X86]
2605 Used for mtrr cleanup. It is granularity of mtrr block.
2607 Large value could prevent small alignment from
2610 mtrr_spare_reg_nr=n [X86]
2612 Range: 0,7 : spare reg number
2614 Used for mtrr cleanup. It is spare mtrr entries number.
2615 Set to 2 or more if your graphical card needs more.
2617 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2619 netdev= [NET] Network devices parameters
2620 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2621 Note that mem_start is often overloaded to mean
2622 something different and driver-specific.
2623 This usage is only documented in each driver source
2627 [NETFILTER] Enable connection tracking flow accounting
2628 0 to disable accounting
2629 1 to enable accounting
2632 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2633 See Documentation/filesystems/nfs/nfsroot.txt.
2635 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2636 See Documentation/filesystems/nfs/nfsroot.txt.
2638 nfsrootdebug [NFS] enable nfsroot debugging messages.
2639 See Documentation/filesystems/nfs/nfsroot.txt.
2641 nfs.callback_nr_threads=
2642 [NFSv4] set the total number of threads that the
2643 NFS client will assign to service NFSv4 callback
2646 nfs.callback_tcpport=
2647 [NFS] set the TCP port on which the NFSv4 callback
2648 channel should listen.
2651 [NFS] sets the pathname to the program which is used
2652 to update the NFS client cache entries.
2654 nfs.cache_getent_timeout=
2655 [NFS] sets the timeout after which an attempt to
2656 update a cache entry is deemed to have failed.
2658 nfs.idmap_cache_timeout=
2659 [NFS] set the maximum lifetime for idmapper cache
2663 [NFS] enable 64-bit inode numbers.
2664 If zero, the NFS client will fake up a 32-bit inode
2665 number for the readdir() and stat() syscalls instead
2666 of returning the full 64-bit number.
2667 The default is to return 64-bit inode numbers.
2669 nfs.max_session_cb_slots=
2670 [NFSv4.1] Sets the maximum number of session
2671 slots the client will assign to the callback
2672 channel. This determines the maximum number of
2673 callbacks the client will process in parallel for
2674 a particular server.
2676 nfs.max_session_slots=
2677 [NFSv4.1] Sets the maximum number of session slots
2678 the client will attempt to negotiate with the server.
2679 This limits the number of simultaneous RPC requests
2680 that the client can send to the NFSv4.1 server.
2681 Note that there is little point in setting this
2682 value higher than the max_tcp_slot_table_limit.
2684 nfs.nfs4_disable_idmapping=
2685 [NFSv4] When set to the default of '1', this option
2686 ensures that both the RPC level authentication
2687 scheme and the NFS level operations agree to use
2688 numeric uids/gids if the mount is using the
2689 'sec=sys' security flavour. In effect it is
2690 disabling idmapping, which can make migration from
2691 legacy NFSv2/v3 systems to NFSv4 easier.
2692 Servers that do not support this mode of operation
2693 will be autodetected by the client, and it will fall
2694 back to using the idmapper.
2695 To turn off this behaviour, set the value to '0'.
2697 [NFS4] Specify an additional fixed unique ident-
2698 ification string that NFSv4 clients can insert into
2699 their nfs_client_id4 string. This is typically a
2700 UUID that is generated at system install time.
2702 nfs.send_implementation_id =
2703 [NFSv4.1] Send client implementation identification
2704 information in exchange_id requests.
2705 If zero, no implementation identification information
2707 The default is to send the implementation identification
2710 nfs.recover_lost_locks =
2711 [NFSv4] Attempt to recover locks that were lost due
2712 to a lease timeout on the server. Please note that
2713 doing this risks data corruption, since there are
2714 no guarantees that the file will remain unchanged
2715 after the locks are lost.
2716 If you want to enable the kernel legacy behaviour of
2717 attempting to recover these locks, then set this
2719 The default parameter value of '0' causes the kernel
2720 not to attempt recovery of lost locks.
2722 nfs4.layoutstats_timer =
2723 [NFSv4.2] Change the rate at which the kernel sends
2724 layoutstats to the pNFS metadata server.
2726 Setting this to value to 0 causes the kernel to use
2727 whatever value is the default set by the layout
2728 driver. A non-zero value sets the minimum interval
2729 in seconds between layoutstats transmissions.
2731 nfsd.nfs4_disable_idmapping=
2732 [NFSv4] When set to the default of '1', the NFSv4
2733 server will return only numeric uids and gids to
2734 clients using auth_sys, and will accept numeric uids
2735 and gids from such clients. This is intended to ease
2736 migration from NFSv2/v3.
2738 nmi_debug= [KNL,SH] Specify one or more actions to take
2739 when a NMI is triggered.
2740 Format: [state][,regs][,debounce][,die]
2742 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2743 Format: [panic,][nopanic,][num]
2745 0 - turn hardlockup detector in nmi_watchdog off
2746 1 - turn hardlockup detector in nmi_watchdog on
2747 When panic is specified, panic when an NMI watchdog
2748 timeout occurs (or 'nopanic' to override the opposite
2749 default). To disable both hard and soft lockup detectors,
2750 please see 'nowatchdog'.
2751 This is useful when you use a panic=... timeout and
2752 need the box quickly up again.
2754 These settings can be accessed at runtime via
2755 the nmi_watchdog and hardlockup_panic sysctls.
2757 netpoll.carrier_timeout=
2758 [NET] Specifies amount of time (in seconds) that
2759 netpoll should wait for a carrier. By default netpoll
2762 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2763 emulation library even if a 387 maths coprocessor
2766 no5lvl [X86-64] Disable 5-level paging mode. Forces
2767 kernel to use 4-level paging instead.
2770 [HW] Never suspend the console
2771 Disable suspending of consoles during suspend and
2772 hibernate operations. Once disabled, debugging
2773 messages can reach various consoles while the rest
2774 of the system is being put to sleep (ie, while
2775 debugging driver suspend/resume hooks). This may
2776 not work reliably with all consoles, but is known
2777 to work with serial and VGA consoles.
2778 To facilitate more flexible debugging, we also add
2779 console_suspend, a printk module parameter to control
2780 it. Users could use console_suspend (usually
2781 /sys/module/printk/parameters/console_suspend) to
2782 turn on/off it dynamically.
2784 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2785 caches in the slab allocator. Saves per-node memory,
2786 but will impact performance.
2790 noaltinstr [S390] Disables alternative instructions patching
2791 (CPU alternatives feature).
2793 noapic [SMP,APIC] Tells the kernel to not make use of any
2794 IOAPICs that may be present in the system.
2796 noautogroup Disable scheduler automatic task group creation.
2798 nobats [PPC] Do not use BATs for mapping kernel lowmem
2799 on "Classic" PPC cores.
2803 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2805 nodelayacct [KNL] Disable per-task delay accounting
2807 nodsp [SH] Disable hardware DSP at boot time.
2809 noefi Disable EFI runtime services support.
2814 On X86-32 available only on PAE configured kernels.
2815 noexec=on: enable non-executable mappings (default)
2816 noexec=off: disable non-executable mappings
2819 Disable SMAP (Supervisor Mode Access Prevention)
2820 even if it is supported by processor.
2823 Disable SMEP (Supervisor Mode Execution Prevention)
2824 even if it is supported by processor.
2827 This affects only 32-bit executables.
2828 noexec32=on: enable non-executable mappings (default)
2829 read doesn't imply executable mappings
2830 noexec32=off: disable non-executable mappings
2831 read implies executable mappings
2833 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2835 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2836 register save and restore. The kernel will only save
2837 legacy floating-point registers on task switch.
2839 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2841 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2842 Equivalent to smt=1.
2844 [KNL,x86] Disable symmetric multithreading (SMT).
2845 nosmt=force: Force disable SMT, cannot be undone
2846 via the sysfs control file.
2848 nospectre_v1 [PPC] Disable mitigations for Spectre Variant 1 (bounds
2849 check bypass). With this option data leaks are possible
2852 nospectre_v2 [X86,PPC_FSL_BOOK3E] Disable all mitigations for the Spectre variant 2
2853 (indirect branch prediction) vulnerability. System may
2854 allow data leaks with this option, which is equivalent
2857 nospec_store_bypass_disable
2858 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
2860 noxsave [BUGS=X86] Disables x86 extended register state save
2861 and restore using xsave. The kernel will fallback to
2862 enabling legacy floating-point and sse state.
2864 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2865 register states. The kernel will fall back to use
2866 xsave to save the states. By using this parameter,
2867 performance of saving the states is degraded because
2868 xsave doesn't support modified optimization while
2869 xsaveopt supports it on xsaveopt enabled systems.
2871 noxsaves [X86] Disables xsaves and xrstors used in saving and
2872 restoring x86 extended register state in compacted
2873 form of xsave area. The kernel will fall back to use
2874 xsaveopt and xrstor to save and restore the states
2875 in standard form of xsave area. By using this
2876 parameter, xsave area per process might occupy more
2877 memory on xsaves enabled systems.
2879 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2880 wfi(ARM) instruction doesn't work correctly and not to
2881 use it. This is also useful when using JTAG debugger.
2883 no_file_caps Tells the kernel not to honor file capabilities. The
2884 only way then for a file to be executed with privilege
2885 is to be setuid root or executed by root.
2887 nohalt [IA-64] Tells the kernel not to use the power saving
2888 function PAL_HALT_LIGHT when idle. This increases
2889 power-consumption. On the positive side, it reduces
2890 interrupt wake-up latency, which may improve performance
2891 in certain environments such as networked servers or
2894 nohibernate [HIBERNATION] Disable hibernation and resume.
2896 nohz= [KNL] Boottime enable/disable dynamic ticks
2897 Valid arguments: on, off
2900 nohz_full= [KNL,BOOT,SMP,ISOL]
2901 The argument is a cpu list, as described above.
2902 In kernels built with CONFIG_NO_HZ_FULL=y, set
2903 the specified list of CPUs whose tick will be stopped
2904 whenever possible. The boot CPU will be forced outside
2905 the range to maintain the timekeeping. Any CPUs
2906 in this list will have their RCU callbacks offloaded,
2907 just as if they had also been called out in the
2908 rcu_nocbs= boot parameter.
2910 noiotrap [SH] Disables trapped I/O port accesses.
2912 noirqdebug [X86-32] Disables the code which attempts to detect and
2913 disable unhandled interrupt sources.
2915 no_timer_check [X86,APIC] Disables the code which tests for
2916 broken timer IRQ sources.
2918 noisapnp [ISAPNP] Disables ISA PnP code.
2920 noinitrd [RAM] Tells the kernel not to load any configured
2923 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2925 [Deprecated - use intremap=off]
2929 noinvpcid [X86] Disable the INVPCID cpu feature.
2931 nojitter [IA-64] Disables jitter checking for ITC timers.
2933 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2935 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2939 [X86,PV_OPS] Disable paravirtualized VMware scheduler
2940 clock and use the default one.
2942 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2943 steal time is computed, but won't influence scheduler
2946 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2948 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2950 noltlbs [PPC] Do not use large page/tlb entries for kernel
2951 lowmem mapping on PPC40x and PPC8xx
2953 nomca [IA-64] Disable machine check abort handling
2955 nomce [X86-32] Disable Machine Check Exception
2957 nomfgpt [X86-32] Disable Multi-Function General Purpose
2958 Timer usage (for AMD Geode machines).
2960 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2961 shutdown the other cpus. Instead use the REBOOT_VECTOR
2964 nomodule Disable module load
2966 nopat [X86] Disable PAT (page attribute table extension of
2967 pagetables) support.
2969 nopcid [X86-64] Disable the PCID cpu feature.
2971 norandmaps Don't use address space randomization. Equivalent to
2972 echo 0 > /proc/sys/kernel/randomize_va_space
2974 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2975 with UP alternatives
2977 nordrand [X86] Disable kernel use of the RDRAND and
2978 RDSEED instructions even if they are supported
2979 by the processor. RDRAND and RDSEED are still
2980 available to user space applications.
2982 noresume [SWSUSP] Disables resume and restores original swap
2985 no-scroll [VGA] Disables scrollback.
2986 This is required for the Braillex ib80-piezo Braille
2987 reader made by F.H. Papenmeier (Germany).
2991 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2993 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2994 and disable the IO APIC. legacy for "maxcpus=0".
2996 nosoftlockup [KNL] Disable the soft-lockup detector.
2998 nosync [HW,M68K] Disables sync negotiation for all devices.
3000 nowatchdog [KNL] Disable both lockup detectors, i.e.
3001 soft-lockup and NMI watchdog (hard-lockup).
3005 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3007 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3008 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3009 Some features depend on CPU0. Known dependencies are:
3010 1. Resume from suspend/hibernate depends on CPU0.
3011 Suspend/hibernate will fail if CPU0 is offline and you
3012 need to online CPU0 before suspend/hibernate.
3013 2. PIC interrupts also depend on CPU0. CPU0 can't be
3014 removed if a PIC interrupt is detected.
3015 It's said poweroff/reboot may depend on CPU0 on some
3016 machines although I haven't seen such issues so far
3017 after CPU0 is offline on a few tested machines.
3018 If the dependencies are under your control, you can
3019 turn on cpu0_hotplug.
3021 nps_mtm_hs_ctr= [KNL,ARC]
3022 This parameter sets the maximum duration, in
3023 cycles, each HW thread of the CTOP can run
3024 without interruptions, before HW switches it.
3025 The actual maximum duration is 16 times this
3027 Format: integer between 1 and 255
3030 nptcg= [IA-64] Override max number of concurrent global TLB
3031 purges which is reported from either PAL_VM_SUMMARY or
3034 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3035 could support. nr_cpus=n : n >= 1 limits the kernel to
3036 support 'n' processors. It could be larger than the
3037 number of already plugged CPU during bootup, later in
3038 runtime you can physically add extra cpu until it reaches
3039 n. So during boot up some boot time memory for per-cpu
3040 variables need be pre-allocated for later physical cpu
3043 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3045 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3046 Allowed values are enable and disable
3048 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3049 'node', 'default' can be specified
3050 This can be set from sysctl after boot.
3051 See Documentation/sysctl/vm.txt for details.
3053 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3054 See Documentation/debugging-via-ohci1394.txt for more
3057 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3058 Rather than timing out after 20 ms if an EC
3059 command is not properly ACKed, override the length
3060 of the timeout. We have interrupts disabled while
3061 waiting for the ACK, so if this is set too high
3062 interrupts *may* be lost!
3064 omap_mux= [OMAP] Override bootloader pin multiplexing.
3065 Format: <mux_mode0.mode_name=value>...
3066 For example, to override I2C bus2:
3067 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3069 oprofile.timer= [HW]
3070 Use timer interrupt instead of performance counters
3072 oprofile.cpu_type= Force an oprofile cpu type
3073 This might be useful if you have an older oprofile
3074 userland or if you want common events.
3075 Format: { arch_perfmon }
3076 arch_perfmon: [X86] Force use of architectural
3077 perfmon on Intel CPUs instead of the
3078 CPU specific event set.
3079 timer: [X86] Force use of architectural NMI
3080 timer mode (see also oprofile.timer
3081 for generic hr timer mode)
3083 oops=panic Always panic on oopses. Default is to just kill the
3084 process, but there is a small probability of
3085 deadlocking the machine.
3086 This will also cause panics on machine check exceptions.
3087 Useful together with panic=30 to trigger a reboot.
3089 page_owner= [KNL] Boot-time page_owner enabling option.
3090 Storage of the information about who allocated
3091 each page is disabled in default. With this switch,
3093 on: enable the feature
3095 page_poison= [KNL] Boot-time parameter changing the state of
3096 poisoning on the buddy allocator, available with
3097 CONFIG_PAGE_POISONING=y.
3098 off: turn off poisoning (default)
3099 on: turn on poisoning
3101 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3102 timeout > 0: seconds before rebooting
3103 timeout = 0: wait forever
3104 timeout < 0: reboot immediately
3107 panic_print= Bitmask for printing system info when panic happens.
3108 User can chose combination of the following bits:
3109 bit 0: print all tasks info
3110 bit 1: print system memory info
3111 bit 2: print timer info
3112 bit 3: print locks info if CONFIG_LOCKDEP is on
3113 bit 4: print ftrace buffer
3115 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3118 crash_kexec_post_notifiers
3119 Run kdump after running panic-notifiers and dumping
3120 kmsg. This only for the users who doubt kdump always
3121 succeeds in any situation.
3122 Note that this also increases risks of kdump failure,
3123 because some panic notifiers can make the crashed
3124 kernel more unstable.
3126 parkbd.port= [HW] Parallel port number the keyboard adapter is
3127 connected to, default is 0.
3129 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3130 0 for XT, 1 for AT (default is AT).
3133 parport= [HW,PPT] Specify parallel ports. 0 disables.
3134 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3135 Use 'auto' to force the driver to use any
3136 IRQ/DMA settings detected (the default is to
3137 ignore detected IRQ/DMA settings because of
3138 possible conflicts). You can specify the base
3139 address, IRQ, and DMA settings; IRQ and DMA
3140 should be numbers, or 'auto' (for using detected
3141 settings on that particular port), or 'nofifo'
3142 (to avoid using a FIFO even if it is detected).
3143 Parallel ports are assigned in the order they
3144 are specified on the command line, starting
3147 parport_init_mode= [HW,PPT]
3148 Configure VIA parallel port to operate in
3149 a specific mode. This is necessary on Pegasos
3150 computer where firmware has no options for setting
3151 up parallel port mode and sets it to spp.
3152 Currently this function knows 686a and 8231 chips.
3153 Format: [spp|ps2|epp|ecp|ecpepp]
3156 Halt all CPUs after the first oops has been printed for
3157 the specified number of seconds. This is to be used if
3158 your oopses keep scrolling off the screen.
3163 See header of drivers/block/paride/pcd.c.
3164 See also Documentation/blockdev/paride.txt.
3166 pci=option[,option...] [PCI] various PCI subsystem options.
3168 Some options herein operate on a specific device
3169 or a set of devices (<pci_dev>). These are
3170 specified in one of the following formats:
3172 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3173 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3175 Note: the first format specifies a PCI
3176 bus/device/function address which may change
3177 if new hardware is inserted, if motherboard
3178 firmware changes, or due to changes caused
3179 by other kernel parameters. If the
3180 domain is left unspecified, it is
3181 taken to be zero. Optionally, a path
3182 to a device through multiple device/function
3183 addresses can be specified after the base
3184 address (this is more robust against
3185 renumbering issues). The second format
3186 selects devices using IDs from the
3187 configuration space which may match multiple
3188 devices in the system.
3190 earlydump dump PCI config space before the kernel
3192 off [X86] don't probe for the PCI bus
3193 bios [X86-32] force use of PCI BIOS, don't access
3194 the hardware directly. Use this if your machine
3195 has a non-standard PCI host bridge.
3196 nobios [X86-32] disallow use of PCI BIOS, only direct
3197 hardware access methods are allowed. Use this
3198 if you experience crashes upon bootup and you
3199 suspect they are caused by the BIOS.
3200 conf1 [X86] Force use of PCI Configuration Access
3201 Mechanism 1 (config address in IO port 0xCF8,
3202 data in IO port 0xCFC, both 32-bit).
3203 conf2 [X86] Force use of PCI Configuration Access
3204 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3205 the function, IO port 0xCFA, also 8-bit, sets
3206 bus number. The config space is then accessed
3207 through ports 0xC000-0xCFFF).
3208 See http://wiki.osdev.org/PCI for more info
3209 on the configuration access mechanisms.
3210 noaer [PCIE] If the PCIEAER kernel config parameter is
3211 enabled, this kernel boot option can be used to
3212 disable the use of PCIE advanced error reporting.
3213 nodomains [PCI] Disable support for multiple PCI
3214 root domains (aka PCI segments, in ACPI-speak).
3215 nommconf [X86] Disable use of MMCONFIG for PCI
3217 check_enable_amd_mmconf [X86] check for and enable
3218 properly configured MMIO access to PCI
3219 config space on AMD family 10h CPU
3220 nomsi [MSI] If the PCI_MSI kernel config parameter is
3221 enabled, this kernel boot option can be used to
3222 disable the use of MSI interrupts system-wide.
3223 noioapicquirk [APIC] Disable all boot interrupt quirks.
3224 Safety option to keep boot IRQs enabled. This
3225 should never be necessary.
3226 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3227 primary IO-APIC for bridges that cannot disable
3228 boot IRQs. This fixes a source of spurious IRQs
3229 when the system masks IRQs.
3230 noioapicreroute [APIC] Disable workaround that uses the
3231 boot IRQ equivalent of an IRQ that connects to
3232 a chipset where boot IRQs cannot be disabled.
3233 The opposite of ioapicreroute.
3234 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3235 routing table. These calls are known to be buggy
3236 on several machines and they hang the machine
3237 when used, but on other computers it's the only
3238 way to get the interrupt routing table. Try
3239 this option if the kernel is unable to allocate
3240 IRQs or discover secondary PCI buses on your
3242 rom [X86] Assign address space to expansion ROMs.
3243 Use with caution as certain devices share
3244 address decoders between ROMs and other
3246 norom [X86] Do not assign address space to
3247 expansion ROMs that do not already have
3248 BIOS assigned address ranges.
3249 nobar [X86] Do not assign address space to the
3250 BARs that weren't assigned by the BIOS.
3251 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3252 assigned automatically to PCI devices. You can
3253 make the kernel exclude IRQs of your ISA cards
3255 pirqaddr=0xAAAAA [X86] Specify the physical address
3256 of the PIRQ table (normally generated
3257 by the BIOS) if it is outside the
3258 F0000h-100000h range.
3259 lastbus=N [X86] Scan all buses thru bus #N. Can be
3260 useful if the kernel is unable to find your
3261 secondary buses and you want to tell it
3262 explicitly which ones they are.
3263 assign-busses [X86] Always assign all PCI bus
3264 numbers ourselves, overriding
3265 whatever the firmware may have done.
3266 usepirqmask [X86] Honor the possible IRQ mask stored
3267 in the BIOS $PIR table. This is needed on
3268 some systems with broken BIOSes, notably
3269 some HP Pavilion N5400 and Omnibook XE3
3270 notebooks. This will have no effect if ACPI
3271 IRQ routing is enabled.
3272 noacpi [X86] Do not use ACPI for IRQ routing
3273 or for PCI scanning.
3274 use_crs [X86] Use PCI host bridge window information
3275 from ACPI. On BIOSes from 2008 or later, this
3276 is enabled by default. If you need to use this,
3277 please report a bug.
3278 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3279 If you need to use this, please report a bug.
3280 routeirq Do IRQ routing for all PCI devices.
3281 This is normally done in pci_enable_device(),
3282 so this option is a temporary workaround
3283 for broken drivers that don't call it.
3284 skip_isa_align [X86] do not align io start addr, so can
3285 handle more pci cards
3286 noearly [X86] Don't do any early type 1 scanning.
3287 This might help on some broken boards which
3288 machine check when some devices' config space
3289 is read. But various workarounds are disabled
3290 and some IOMMU drivers will not work.
3291 bfsort Sort PCI devices into breadth-first order.
3292 This sorting is done to get a device
3293 order compatible with older (<= 2.4) kernels.
3294 nobfsort Don't sort PCI devices into breadth-first order.
3295 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3296 tuning and use the BIOS-configured MPS defaults.
3297 pcie_bus_safe Set every device's MPS to the largest value
3298 supported by all devices below the root complex.
3299 pcie_bus_perf Set device MPS to the largest allowable MPS
3300 based on its parent bus. Also set MRRS (Max
3301 Read Request Size) to the largest supported
3302 value (no larger than the MPS that the device
3303 or bus can support) for best performance.
3304 pcie_bus_peer2peer Set every device's MPS to 128B, which
3305 every device is guaranteed to support. This
3306 configuration allows peer-to-peer DMA between
3307 any pair of devices, possibly at the cost of
3308 reduced performance. This also guarantees
3309 that hot-added devices will work.
3310 cbiosize=nn[KMG] The fixed amount of bus space which is
3311 reserved for the CardBus bridge's IO window.
3312 The default value is 256 bytes.
3313 cbmemsize=nn[KMG] The fixed amount of bus space which is
3314 reserved for the CardBus bridge's memory
3315 window. The default value is 64 megabytes.
3318 [<order of align>@]<pci_dev>[; ...]
3319 Specifies alignment and device to reassign
3320 aligned memory resources. How to
3321 specify the device is described above.
3322 If <order of align> is not specified,
3323 PAGE_SIZE is used as alignment.
3324 PCI-PCI bridge can be specified, if resource
3325 windows need to be expanded.
3326 To specify the alignment for several
3327 instances of a device, the PCI vendor,
3328 device, subvendor, and subdevice may be
3329 specified, e.g., 4096@pci:8086:9c22:103c:198f
3330 ecrc= Enable/disable PCIe ECRC (transaction layer
3331 end-to-end CRC checking).
3332 bios: Use BIOS/firmware settings. This is the
3336 hpiosize=nn[KMG] The fixed amount of bus space which is
3337 reserved for hotplug bridge's IO window.
3338 Default size is 256 bytes.
3339 hpmemsize=nn[KMG] The fixed amount of bus space which is
3340 reserved for hotplug bridge's memory window.
3341 Default size is 2 megabytes.
3342 hpbussize=nn The minimum amount of additional bus numbers
3343 reserved for buses below a hotplug bridge.
3345 realloc= Enable/disable reallocating PCI bridge resources
3346 if allocations done by BIOS are too small to
3347 accommodate resources required by all child
3349 off: Turn realloc off
3351 realloc same as realloc=on
3352 noari do not use PCIe ARI.
3353 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3354 do not use PCIe ATS (and IOMMU device IOTLB).
3355 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3356 only look for one device below a PCIe downstream
3358 big_root_window Try to add a big 64bit memory window to the PCIe
3359 root complex on AMD CPUs. Some GFX hardware
3360 can resize a BAR to allow access to all VRAM.
3361 Adding the window is slightly risky (it may
3362 conflict with unreported devices), so this
3364 disable_acs_redir=<pci_dev>[; ...]
3365 Specify one or more PCI devices (in the format
3366 specified above) separated by semicolons.
3367 Each device specified will have the PCI ACS
3368 redirect capabilities forced off which will
3369 allow P2P traffic between devices through
3370 bridges without forcing it upstream. Note:
3371 this removes isolation between devices and
3372 may put more devices in an IOMMU group.
3374 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3377 force Enable ASPM even on devices that claim not to support it.
3378 WARNING: Forcing ASPM on may cause system lockups.
3380 pcie_ports= [PCIE] PCIe port services handling:
3381 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3382 even if the platform doesn't give the OS permission to
3383 use them. This may cause conflicts if the platform
3384 also tries to use these services.
3385 compat Disable native PCIe services (PME, AER, DPC, PCIe
3388 pcie_port_pm= [PCIE] PCIe port power management handling:
3389 off Disable power management of all PCIe ports
3390 force Forcibly enable power management of all PCIe ports
3392 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3393 nomsi Do not use MSI for native PCIe PME signaling (this makes
3394 all PCIe root ports use INTx for all services).
3396 pcmv= [HW,PCMCIA] BadgePAD 4
3400 Keep all power-domains already enabled by bootloader on,
3401 even if no driver has claimed them. This is useful
3402 for debug and development, but should not be
3403 needed on a platform with proper driver support.
3406 See Documentation/blockdev/paride.txt.
3408 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3411 See arch/parisc/kernel/pdc_chassis.c
3413 percpu_alloc= Select which percpu first chunk allocator to use.
3414 Currently supported values are "embed" and "page".
3415 Archs may support subset or none of the selections.
3416 See comments in mm/percpu.c for details on each
3417 allocator. This parameter is primarily for debugging
3418 and performance comparison.
3421 See Documentation/blockdev/paride.txt.
3424 See Documentation/blockdev/paride.txt.
3426 pirq= [SMP,APIC] Manual mp-table setup
3427 See Documentation/x86/i386/IO-APIC.txt.
3429 plip= [PPT,NET] Parallel port network link
3430 Format: { parport<nr> | timid | 0 }
3431 See also Documentation/admin-guide/parport.rst.
3433 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3434 Override pmtimer IOPort with a hex value.
3438 Enable PNP debug messages (depends on the
3439 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3440 via /sys/module/pnp/parameters/debug. We always show
3441 current resource usage; turning this on also shows
3442 possible settings and some assignment information.
3448 { on | off | curr | res | no-curr | no-res }
3451 [ISAPNP] Exclude IRQs for the autoconfiguration
3454 [ISAPNP] Exclude DMAs for the autoconfiguration
3456 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3457 Ranges are in pairs (I/O port base and size).
3460 [ISAPNP] Exclude memory regions for the
3462 Ranges are in pairs (memory base and size).
3464 ports= [IP_VS_FTP] IPVS ftp helper module
3466 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3468 Format: <port>,<port>....
3470 powersave=off [PPC] This option disables power saving features.
3471 It specifically disables cpuidle and sets the
3472 platform machine description specific power_save
3473 function to NULL. On Idle the CPU just reduces
3476 ppc_strict_facility_enable
3477 [PPC] This option catches any kernel floating point,
3478 Altivec, VSX and SPE outside of regions specifically
3479 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3480 There is some performance impact when enabling this.
3484 Disable Hardware Transactional Memory
3486 print-fatal-signals=
3487 [KNL] debug: print fatal signals
3489 If enabled, warn about various signal handling
3490 related application anomalies: too many signals,
3491 too many POSIX.1 timers, fatal signals causing a
3494 If you hit the warning due to signal overflow,
3495 you might want to try "ulimit -i unlimited".
3499 printk.always_kmsg_dump=
3500 Trigger kmsg_dump for cases other than kernel oops or
3502 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3505 printk.devkmsg={on,off,ratelimit}
3506 Control writing to /dev/kmsg.
3507 on - unlimited logging to /dev/kmsg from userspace
3508 off - logging to /dev/kmsg disabled
3509 ratelimit - ratelimit the logging
3512 printk.time= Show timing data prefixed to each printk message line
3513 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3515 processor.max_cstate= [HW,ACPI]
3516 Limit processor to maximum C-state
3517 max_cstate=9 overrides any DMI blacklist limit.
3519 processor.nocst [HW,ACPI]
3520 Ignore the _CST method to determine C-states,
3521 instead using the legacy FADT method
3523 profile= [KNL] Enable kernel profiling via /proc/profile
3524 Format: [<profiletype>,]<number>
3525 Param: <profiletype>: "schedule", "sleep", or "kvm"
3526 [defaults to kernel profiling]
3527 Param: "schedule" - profile schedule points.
3528 Param: "sleep" - profile D-state sleeping (millisecs).
3529 Requires CONFIG_SCHEDSTATS
3530 Param: "kvm" - profile VM exits.
3531 Param: <number> - step/bucket size as a power of 2 for
3532 statistical time based profiling.
3534 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3536 See Documentation/blockdev/ramdisk.txt.
3538 psi= [KNL] Enable or disable pressure stall information
3542 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3543 probe for; one of (bare|imps|exps|lifebook|any).
3544 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3546 psmouse.resetafter= [HW,MOUSE]
3547 Try to reset the device after so many bad packets
3550 [HW,MOUSE] Set desired mouse resolution, in dpi.
3551 psmouse.smartscroll=
3552 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3553 0 = disabled, 1 = enabled (default).
3555 pstore.backend= Specify the name of the pstore backend to use
3558 See Documentation/blockdev/paride.txt.
3560 pti= [X86_64] Control Page Table Isolation of user and
3561 kernel address spaces. Disabling this feature
3562 removes hardening, but improves performance of
3563 system calls and interrupts.
3565 on - unconditionally enable
3566 off - unconditionally disable
3567 auto - kernel detects whether your CPU model is
3568 vulnerable to issues that PTI mitigates
3570 Not specifying this option is equivalent to pti=auto.
3573 Equivalent to pti=off
3576 [KNL] Number of legacy pty's. Overwrites compiled-in
3579 quiet [KNL] Disable most log messages
3584 See Documentation/admin-guide/md.rst.
3586 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3587 See Documentation/blockdev/ramdisk.txt.
3589 random.trust_cpu={on,off}
3590 [KNL] Enable or disable trusting the use of the
3591 CPU's random number generator (if available) to
3592 fully seed the kernel's CRNG. Default is controlled
3593 by CONFIG_RANDOM_TRUST_CPU.
3595 ras=option[,option,...] [KNL] RAS-specific options
3598 Disable the Correctable Errors Collector,
3599 see CONFIG_RAS_CEC help text.
3602 The argument is a cpu list, as described above.
3604 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3605 the specified list of CPUs to be no-callback CPUs.
3606 Invocation of these CPUs' RCU callbacks will be
3607 offloaded to "rcuox/N" kthreads created for that
3608 purpose, where "x" is "p" for RCU-preempt, and
3609 "s" for RCU-sched, and "N" is the CPU number.
3610 This reduces OS jitter on the offloaded CPUs,
3611 which can be useful for HPC and real-time
3612 workloads. It can also improve energy efficiency
3613 for asymmetric multiprocessors.
3616 Rather than requiring that offloaded CPUs
3617 (specified by rcu_nocbs= above) explicitly
3618 awaken the corresponding "rcuoN" kthreads,
3619 make these kthreads poll for callbacks.
3620 This improves the real-time response for the
3621 offloaded CPUs by relieving them of the need to
3622 wake up the corresponding kthread, but degrades
3623 energy efficiency by requiring that the kthreads
3624 periodically wake up to do the polling.
3626 rcutree.blimit= [KNL]
3627 Set maximum number of finished RCU callbacks to
3628 process in one batch.
3630 rcutree.dump_tree= [KNL]
3631 Dump the structure of the rcu_node combining tree
3632 out at early boot. This is used for diagnostic
3633 purposes, to verify correct tree setup.
3635 rcutree.gp_cleanup_delay= [KNL]
3636 Set the number of jiffies to delay each step of
3637 RCU grace-period cleanup.
3639 rcutree.gp_init_delay= [KNL]
3640 Set the number of jiffies to delay each step of
3641 RCU grace-period initialization.
3643 rcutree.gp_preinit_delay= [KNL]
3644 Set the number of jiffies to delay each step of
3645 RCU grace-period pre-initialization, that is,
3646 the propagation of recent CPU-hotplug changes up
3647 the rcu_node combining tree.
3649 rcutree.rcu_fanout_exact= [KNL]
3650 Disable autobalancing of the rcu_node combining
3651 tree. This is used by rcutorture, and might
3652 possibly be useful for architectures having high
3653 cache-to-cache transfer latencies.
3655 rcutree.rcu_fanout_leaf= [KNL]
3656 Change the number of CPUs assigned to each
3657 leaf rcu_node structure. Useful for very
3658 large systems, which will choose the value 64,
3659 and for NUMA systems with large remote-access
3660 latencies, which will choose a value aligned
3661 with the appropriate hardware boundaries.
3663 rcutree.jiffies_till_sched_qs= [KNL]
3664 Set required age in jiffies for a
3665 given grace period before RCU starts
3666 soliciting quiescent-state help from
3667 rcu_note_context_switch(). If not specified, the
3668 kernel will calculate a value based on the most
3669 recent settings of rcutree.jiffies_till_first_fqs
3670 and rcutree.jiffies_till_next_fqs.
3671 This calculated value may be viewed in
3672 rcutree.jiffies_to_sched_qs. Any attempt to
3673 set rcutree.jiffies_to_sched_qs will be
3674 cheerfully overwritten.
3676 rcutree.jiffies_till_first_fqs= [KNL]
3677 Set delay from grace-period initialization to
3678 first attempt to force quiescent states.
3679 Units are jiffies, minimum value is zero,
3680 and maximum value is HZ.
3682 rcutree.jiffies_till_next_fqs= [KNL]
3683 Set delay between subsequent attempts to force
3684 quiescent states. Units are jiffies, minimum
3685 value is one, and maximum value is HZ.
3687 rcutree.kthread_prio= [KNL,BOOT]
3688 Set the SCHED_FIFO priority of the RCU per-CPU
3689 kthreads (rcuc/N). This value is also used for
3690 the priority of the RCU boost threads (rcub/N)
3691 and for the RCU grace-period kthreads (rcu_bh,
3692 rcu_preempt, and rcu_sched). If RCU_BOOST is
3693 set, valid values are 1-99 and the default is 1
3694 (the least-favored priority). Otherwise, when
3695 RCU_BOOST is not set, valid values are 0-99 and
3696 the default is zero (non-realtime operation).
3698 rcutree.rcu_nocb_leader_stride= [KNL]
3699 Set the number of NOCB kthread groups, which
3700 defaults to the square root of the number of
3701 CPUs. Larger numbers reduces the wakeup overhead
3702 on the per-CPU grace-period kthreads, but increases
3703 that same overhead on each group's leader.
3705 rcutree.qhimark= [KNL]
3706 Set threshold of queued RCU callbacks beyond which
3707 batch limiting is disabled.
3709 rcutree.qlowmark= [KNL]
3710 Set threshold of queued RCU callbacks below which
3711 batch limiting is re-enabled.
3713 rcutree.rcu_idle_gp_delay= [KNL]
3714 Set wakeup interval for idle CPUs that have
3715 RCU callbacks (RCU_FAST_NO_HZ=y).
3717 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3718 Set wakeup interval for idle CPUs that have
3719 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3720 Lazy RCU callbacks are those which RCU can
3721 prove do nothing more than free memory.
3723 rcutree.rcu_kick_kthreads= [KNL]
3724 Cause the grace-period kthread to get an extra
3725 wake_up() if it sleeps three times longer than
3726 it should at force-quiescent-state time.
3727 This wake_up() will be accompanied by a
3728 WARN_ONCE() splat and an ftrace_dump().
3730 rcuperf.gp_async= [KNL]
3731 Measure performance of asynchronous
3732 grace-period primitives such as call_rcu().
3734 rcuperf.gp_async_max= [KNL]
3735 Specify the maximum number of outstanding
3736 callbacks per writer thread. When a writer
3737 thread exceeds this limit, it invokes the
3738 corresponding flavor of rcu_barrier() to allow
3739 previously posted callbacks to drain.
3741 rcuperf.gp_exp= [KNL]
3742 Measure performance of expedited synchronous
3743 grace-period primitives.
3745 rcuperf.holdoff= [KNL]
3746 Set test-start holdoff period. The purpose of
3747 this parameter is to delay the start of the
3748 test until boot completes in order to avoid
3751 rcuperf.nreaders= [KNL]
3752 Set number of RCU readers. The value -1 selects
3753 N, where N is the number of CPUs. A value
3754 "n" less than -1 selects N-n+1, where N is again
3755 the number of CPUs. For example, -2 selects N
3756 (the number of CPUs), -3 selects N+1, and so on.
3757 A value of "n" less than or equal to -N selects
3760 rcuperf.nwriters= [KNL]
3761 Set number of RCU writers. The values operate
3762 the same as for rcuperf.nreaders.
3763 N, where N is the number of CPUs
3765 rcuperf.perf_type= [KNL]
3766 Specify the RCU implementation to test.
3768 rcuperf.shutdown= [KNL]
3769 Shut the system down after performance tests
3770 complete. This is useful for hands-off automated
3773 rcuperf.verbose= [KNL]
3774 Enable additional printk() statements.
3776 rcuperf.writer_holdoff= [KNL]
3777 Write-side holdoff between grace periods,
3778 in microseconds. The default of zero says
3781 rcutorture.fqs_duration= [KNL]
3782 Set duration of force_quiescent_state bursts
3785 rcutorture.fqs_holdoff= [KNL]
3786 Set holdoff time within force_quiescent_state bursts
3789 rcutorture.fqs_stutter= [KNL]
3790 Set wait time between force_quiescent_state bursts
3793 rcutorture.fwd_progress= [KNL]
3794 Enable RCU grace-period forward-progress testing
3795 for the types of RCU supporting this notion.
3797 rcutorture.fwd_progress_div= [KNL]
3798 Specify the fraction of a CPU-stall-warning
3799 period to do tight-loop forward-progress testing.
3801 rcutorture.fwd_progress_holdoff= [KNL]
3802 Number of seconds to wait between successive
3803 forward-progress tests.
3805 rcutorture.fwd_progress_need_resched= [KNL]
3806 Enclose cond_resched() calls within checks for
3807 need_resched() during tight-loop forward-progress
3810 rcutorture.gp_cond= [KNL]
3811 Use conditional/asynchronous update-side
3812 primitives, if available.
3814 rcutorture.gp_exp= [KNL]
3815 Use expedited update-side primitives, if available.
3817 rcutorture.gp_normal= [KNL]
3818 Use normal (non-expedited) asynchronous
3819 update-side primitives, if available.
3821 rcutorture.gp_sync= [KNL]
3822 Use normal (non-expedited) synchronous
3823 update-side primitives, if available. If all
3824 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3825 rcutorture.gp_normal=, and rcutorture.gp_sync=
3826 are zero, rcutorture acts as if is interpreted
3827 they are all non-zero.
3829 rcutorture.n_barrier_cbs= [KNL]
3830 Set callbacks/threads for rcu_barrier() testing.
3832 rcutorture.nfakewriters= [KNL]
3833 Set number of concurrent RCU writers. These just
3834 stress RCU, they don't participate in the actual
3835 test, hence the "fake".
3837 rcutorture.nreaders= [KNL]
3838 Set number of RCU readers. The value -1 selects
3839 N-1, where N is the number of CPUs. A value
3840 "n" less than -1 selects N-n-2, where N is again
3841 the number of CPUs. For example, -2 selects N
3842 (the number of CPUs), -3 selects N+1, and so on.
3844 rcutorture.object_debug= [KNL]
3845 Enable debug-object double-call_rcu() testing.
3847 rcutorture.onoff_holdoff= [KNL]
3848 Set time (s) after boot for CPU-hotplug testing.
3850 rcutorture.onoff_interval= [KNL]
3851 Set time (jiffies) between CPU-hotplug operations,
3852 or zero to disable CPU-hotplug testing.
3854 rcutorture.shuffle_interval= [KNL]
3855 Set task-shuffle interval (s). Shuffling tasks
3856 allows some CPUs to go into dyntick-idle mode
3857 during the rcutorture test.
3859 rcutorture.shutdown_secs= [KNL]
3860 Set time (s) after boot system shutdown. This
3861 is useful for hands-off automated testing.
3863 rcutorture.stall_cpu= [KNL]
3864 Duration of CPU stall (s) to test RCU CPU stall
3865 warnings, zero to disable.
3867 rcutorture.stall_cpu_holdoff= [KNL]
3868 Time to wait (s) after boot before inducing stall.
3870 rcutorture.stall_cpu_irqsoff= [KNL]
3871 Disable interrupts while stalling if set.
3873 rcutorture.stat_interval= [KNL]
3874 Time (s) between statistics printk()s.
3876 rcutorture.stutter= [KNL]
3877 Time (s) to stutter testing, for example, specifying
3878 five seconds causes the test to run for five seconds,
3879 wait for five seconds, and so on. This tests RCU's
3880 ability to transition abruptly to and from idle.
3882 rcutorture.test_boost= [KNL]
3883 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3884 "Maybe" means test if the RCU implementation
3885 under test support RCU priority boosting.
3887 rcutorture.test_boost_duration= [KNL]
3888 Duration (s) of each individual boost test.
3890 rcutorture.test_boost_interval= [KNL]
3891 Interval (s) between each boost test.
3893 rcutorture.test_no_idle_hz= [KNL]
3894 Test RCU's dyntick-idle handling. See also the
3895 rcutorture.shuffle_interval parameter.
3897 rcutorture.torture_type= [KNL]
3898 Specify the RCU implementation to test.
3900 rcutorture.verbose= [KNL]
3901 Enable additional printk() statements.
3903 rcupdate.rcu_cpu_stall_suppress= [KNL]
3904 Suppress RCU CPU stall warning messages.
3906 rcupdate.rcu_cpu_stall_timeout= [KNL]
3907 Set timeout for RCU CPU stall warning messages.
3909 rcupdate.rcu_expedited= [KNL]
3910 Use expedited grace-period primitives, for
3911 example, synchronize_rcu_expedited() instead
3912 of synchronize_rcu(). This reduces latency,
3913 but can increase CPU utilization, degrade
3914 real-time latency, and degrade energy efficiency.
3915 No effect on CONFIG_TINY_RCU kernels.
3917 rcupdate.rcu_normal= [KNL]
3918 Use only normal grace-period primitives,
3919 for example, synchronize_rcu() instead of
3920 synchronize_rcu_expedited(). This improves
3921 real-time latency, CPU utilization, and
3922 energy efficiency, but can expose users to
3923 increased grace-period latency. This parameter
3924 overrides rcupdate.rcu_expedited. No effect on
3925 CONFIG_TINY_RCU kernels.
3927 rcupdate.rcu_normal_after_boot= [KNL]
3928 Once boot has completed (that is, after
3929 rcu_end_inkernel_boot() has been invoked), use
3930 only normal grace-period primitives. No effect
3931 on CONFIG_TINY_RCU kernels.
3933 rcupdate.rcu_task_stall_timeout= [KNL]
3934 Set timeout in jiffies for RCU task stall warning
3935 messages. Disable with a value less than or equal
3938 rcupdate.rcu_self_test= [KNL]
3939 Run the RCU early boot self tests
3943 Run specified binary instead of /init from the ramdisk,
3944 used for early userspace startup. See initrd.
3947 Turn on/off individual RDT features. List is:
3948 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
3950 E.g. to turn on cmt and turn off mba use:
3954 Format (x86 or x86_64):
3955 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3957 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3959 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3960 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3961 reboot_force is either force or not specified,
3962 reboot_cpu is s[mp]#### with #### being the processor
3963 to be used for rebooting.
3966 [KNL, SMP] Set scheduler's default relax_domain_level.
3967 See Documentation/cgroup-v1/cpusets.txt.
3969 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
3970 Format: <base1>,<size1>[,<base2>,<size2>,...]
3971 Reserve I/O ports or memory so the kernel won't use
3972 them. If <base> is less than 0x10000, the region
3973 is assumed to be I/O ports; otherwise it is memory.
3975 reservetop= [X86-32]
3977 Reserves a hole at the top of the kernel virtual
3982 Set the amount of memory to reserve for BIOS at
3983 the bottom of the address space.
3985 reset_devices [KNL] Force drivers to reset the underlying device
3986 during initialization.
3989 Specify the partition device for software suspend
3991 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3993 resume_offset= [SWSUSP]
3994 Specify the offset from the beginning of the partition
3995 given by "resume=" at which the swap header is located,
3996 in <PAGE_SIZE> units (needed only for swap files).
3997 See Documentation/power/swsusp-and-swap-files.txt
3999 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4000 read the resume files
4002 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4003 Useful for devices that are detected asynchronously
4004 (e.g. USB and MMC devices).
4006 hibernate= [HIBERNATION]
4007 noresume Don't check if there's a hibernation image
4008 present during boot.
4009 nocompress Don't compress/decompress hibernation images.
4010 no Disable hibernation and resume.
4011 protect_image Turn on image protection during restoration
4012 (that will set all pages holding image data
4013 during restoration read-only).
4015 retain_initrd [RAM] Keep initrd memory after extraction
4017 rfkill.default_state=
4018 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4019 etc. communication is blocked by default.
4022 rfkill.master_switch_mode=
4023 0 The "airplane mode" button does nothing.
4024 1 The "airplane mode" button toggles between everything
4025 blocked and the previous configuration.
4026 2 The "airplane mode" button toggles between everything
4027 blocked and everything unblocked.
4029 rhash_entries= [KNL,NET]
4030 Set number of hash buckets for route cache
4033 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4036 ro [KNL] Mount root device read-only on boot
4039 on Mark read-only kernel memory as read-only (default).
4040 off Leave read-only kernel memory writable for debugging.
4043 Enable the uart passthrough on the designated usb port
4044 on Rockchip SoCs. When active, the signals of the
4045 debug-uart get routed to the D+ and D- pins of the usb
4046 port and the regular usb controller gets disabled.
4048 root= [KNL] Root filesystem
4049 See name_to_dev_t comment in init/do_mounts.c.
4051 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4052 mount the root filesystem
4054 rootflags= [KNL] Set root filesystem mount option string
4056 rootfstype= [KNL] Set root filesystem type
4058 rootwait [KNL] Wait (indefinitely) for root device to show up.
4059 Useful for devices that are detected asynchronously
4060 (e.g. USB and MMC devices).
4062 rproc_mem=nn[KMG][@address]
4063 [KNL,ARM,CMA] Remoteproc physical memory block.
4064 Memory area to be used by remote processor image,
4067 rw [KNL] Mount root device read-write on boot
4069 S [KNL] Run init in single mode
4071 s390_iommu= [HW,S390]
4072 Set s390 IOTLB flushing mode
4074 With strict flushing every unmap operation will result in
4075 an IOTLB flush. Default is lazy flushing before reuse,
4079 See drivers/net/irda/sa1100_ir.c.
4081 sbni= [NET] Granch SBNI12 leased line adapter
4083 sched_debug [KNL] Enables verbose scheduler debug messages.
4085 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4086 Allowed values are enable and disable. This feature
4087 incurs a small amount of overhead in the scheduler
4088 but is useful for debugging and performance tuning.
4090 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4091 xtime_lock contention on larger systems, and/or RCU lock
4092 contention on all systems with CONFIG_MAXSMP set.
4093 Format: { "0" | "1" }
4094 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4096 Note: increases power consumption, thus should only be
4097 enabled if running jitter sensitive (HPC/RT) workloads.
4099 security= [SECURITY] Choose a security module to enable at boot.
4100 If this boot parameter is not specified, only the first
4101 security module asking for security registration will be
4102 loaded. An invalid security module name will be treated
4103 as if no module has been chosen.
4105 selinux= [SELINUX] Disable or enable SELinux at boot time.
4106 Format: { "0" | "1" }
4107 See security/selinux/Kconfig help text.
4110 Default value is set via kernel config option.
4111 If enabled at boot time, /selinux/disable can be used
4112 later to disable prior to initial policy load.
4114 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4115 Format: { "0" | "1" }
4116 See security/apparmor/Kconfig help text
4119 Default value is set via kernel config option.
4121 serialnumber [BUGS=X86-32]
4124 Maximal number of shapers.
4132 Disable merging of slabs with similar size. May be
4133 necessary if there is some reason to distinguish
4134 allocs to different slabs, especially in hardened
4135 environments where the risk of heap overflows and
4136 layout control by attackers can usually be
4137 frustrated by disabling merging. This will reduce
4138 most of the exposure of a heap attack to a single
4139 cache (risks via metadata attacks are mostly
4140 unchanged). Debug options disable merging on their
4142 For more information see Documentation/vm/slub.rst.
4144 slab_max_order= [MM, SLAB]
4145 Determines the maximum allowed order for slabs.
4146 A high setting may cause OOMs due to memory
4147 fragmentation. Defaults to 1 for systems with
4148 more than 32MB of RAM, 0 otherwise.
4150 slub_debug[=options[,slabs]] [MM, SLUB]
4151 Enabling slub_debug allows one to determine the
4152 culprit if slab objects become corrupted. Enabling
4153 slub_debug can create guard zones around objects and
4154 may poison objects when not in use. Also tracks the
4155 last alloc / free. For more information see
4156 Documentation/vm/slub.rst.
4158 slub_memcg_sysfs= [MM, SLUB]
4159 Determines whether to enable sysfs directories for
4160 memory cgroup sub-caches. 1 to enable, 0 to disable.
4161 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4162 Enabling this can lead to a very high number of debug
4163 directories and files being created under
4166 slub_max_order= [MM, SLUB]
4167 Determines the maximum allowed order for slabs.
4168 A high setting may cause OOMs due to memory
4169 fragmentation. For more information see
4170 Documentation/vm/slub.rst.
4172 slub_min_objects= [MM, SLUB]
4173 The minimum number of objects per slab. SLUB will
4174 increase the slab order up to slub_max_order to
4175 generate a sufficiently large slab able to contain
4176 the number of objects indicated. The higher the number
4177 of objects the smaller the overhead of tracking slabs
4178 and the less frequently locks need to be acquired.
4179 For more information see Documentation/vm/slub.rst.
4181 slub_min_order= [MM, SLUB]
4182 Determines the minimum page order for slabs. Must be
4183 lower than slub_max_order.
4184 For more information see Documentation/vm/slub.rst.
4186 slub_nomerge [MM, SLUB]
4187 Same with slab_nomerge. This is supported for legacy.
4188 See slab_nomerge for more information.
4191 Format: <io1>[,<io2>[,...,<io8>]]
4193 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4194 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4195 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4196 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4197 smsc-ircc2.ircc_irq= [HW] IRQ line
4198 smsc-ircc2.ircc_dma= [HW] DMA channel
4199 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4200 0: Toshiba Satellite 1800 (GP data pin select)
4201 1: Fast pin select (default)
4204 smt [KNL,S390] Set the maximum number of threads (logical
4205 CPUs) to use per physical CPU on systems capable of
4206 symmetric multithreading (SMT). Will be capped to the
4207 actual hardware limit.
4209 Default: -1 (no limit)
4212 [KNL] Should the soft-lockup detector generate panics.
4215 A nonzero value instructs the soft-lockup detector
4216 to panic the machine when a soft-lockup occurs. This
4217 is also controlled by CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC
4218 which is the respective build-time switch to that
4221 softlockup_all_cpu_backtrace=
4222 [KNL] Should the soft-lockup detector generate
4223 backtraces on all cpus.
4226 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4227 See Documentation/laptops/sonypi.txt
4229 spectre_v2= [X86] Control mitigation of Spectre variant 2
4230 (indirect branch speculation) vulnerability.
4231 The default operation protects the kernel from
4234 on - unconditionally enable, implies
4236 off - unconditionally disable, implies
4238 auto - kernel detects whether your CPU model is
4241 Selecting 'on' will, and 'auto' may, choose a
4242 mitigation method at run time according to the
4243 CPU, the available microcode, the setting of the
4244 CONFIG_RETPOLINE configuration option, and the
4245 compiler with which the kernel was built.
4247 Selecting 'on' will also enable the mitigation
4248 against user space to user space task attacks.
4250 Selecting 'off' will disable both the kernel and
4251 the user space protections.
4253 Specific mitigations can also be selected manually:
4255 retpoline - replace indirect branches
4256 retpoline,generic - google's original retpoline
4257 retpoline,amd - AMD-specific minimal thunk
4259 Not specifying this option is equivalent to
4263 [X86] Control mitigation of Spectre variant 2
4264 (indirect branch speculation) vulnerability between
4267 on - Unconditionally enable mitigations. Is
4268 enforced by spectre_v2=on
4270 off - Unconditionally disable mitigations. Is
4271 enforced by spectre_v2=off
4273 prctl - Indirect branch speculation is enabled,
4274 but mitigation can be enabled via prctl
4275 per thread. The mitigation control state
4276 is inherited on fork.
4279 - Like "prctl" above, but only STIBP is
4280 controlled per thread. IBPB is issued
4281 always when switching between different user
4285 - Same as "prctl" above, but all seccomp
4286 threads will enable the mitigation unless
4287 they explicitly opt out.
4290 - Like "seccomp" above, but only STIBP is
4291 controlled per thread. IBPB is issued
4292 always when switching between different
4293 user space processes.
4295 auto - Kernel selects the mitigation depending on
4296 the available CPU features and vulnerability.
4299 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4301 Not specifying this option is equivalent to
4302 spectre_v2_user=auto.
4304 spec_store_bypass_disable=
4305 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4306 (Speculative Store Bypass vulnerability)
4308 Certain CPUs are vulnerable to an exploit against a
4309 a common industry wide performance optimization known
4310 as "Speculative Store Bypass" in which recent stores
4311 to the same memory location may not be observed by
4312 later loads during speculative execution. The idea
4313 is that such stores are unlikely and that they can
4314 be detected prior to instruction retirement at the
4315 end of a particular speculation execution window.
4317 In vulnerable processors, the speculatively forwarded
4318 store can be used in a cache side channel attack, for
4319 example to read memory to which the attacker does not
4320 directly have access (e.g. inside sandboxed code).
4322 This parameter controls whether the Speculative Store
4323 Bypass optimization is used.
4325 On x86 the options are:
4327 on - Unconditionally disable Speculative Store Bypass
4328 off - Unconditionally enable Speculative Store Bypass
4329 auto - Kernel detects whether the CPU model contains an
4330 implementation of Speculative Store Bypass and
4331 picks the most appropriate mitigation. If the
4332 CPU is not vulnerable, "off" is selected. If the
4333 CPU is vulnerable the default mitigation is
4334 architecture and Kconfig dependent. See below.
4335 prctl - Control Speculative Store Bypass per thread
4336 via prctl. Speculative Store Bypass is enabled
4337 for a process by default. The state of the control
4338 is inherited on fork.
4339 seccomp - Same as "prctl" above, but all seccomp threads
4340 will disable SSB unless they explicitly opt out.
4342 Default mitigations:
4343 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4345 On powerpc the options are:
4347 on,auto - On Power8 and Power9 insert a store-forwarding
4348 barrier on kernel entry and exit. On Power7
4349 perform a software flush on kernel entry and
4353 Not specifying this option is equivalent to
4354 spec_store_bypass_disable=auto.
4356 spia_io_base= [HW,MTD]
4361 srcutree.counter_wrap_check [KNL]
4362 Specifies how frequently to check for
4363 grace-period sequence counter wrap for the
4364 srcu_data structure's ->srcu_gp_seq_needed field.
4365 The greater the number of bits set in this kernel
4366 parameter, the less frequently counter wrap will
4367 be checked for. Note that the bottom two bits
4370 srcutree.exp_holdoff [KNL]
4371 Specifies how many nanoseconds must elapse
4372 since the end of the last SRCU grace period for
4373 a given srcu_struct until the next normal SRCU
4374 grace period will be considered for automatic
4375 expediting. Set to zero to disable automatic
4379 Speculative Store Bypass Disable control
4381 On CPUs that are vulnerable to the Speculative
4382 Store Bypass vulnerability and offer a
4383 firmware based mitigation, this parameter
4384 indicates how the mitigation should be used:
4386 force-on: Unconditionally enable mitigation for
4387 for both kernel and userspace
4388 force-off: Unconditionally disable mitigation for
4389 for both kernel and userspace
4390 kernel: Always enable mitigation in the
4391 kernel, and offer a prctl interface
4392 to allow userspace to register its
4393 interest in being mitigated too.
4395 stack_guard_gap= [MM]
4396 override the default stack gap protection. The value
4397 is in page units and it defines how many pages prior
4398 to (for stacks growing down) resp. after (for stacks
4399 growing up) the main stack are reserved for no other
4400 mapping. Default value is 256 pages.
4403 Enabled the stack tracer on boot up.
4405 stacktrace_filter=[function-list]
4406 [FTRACE] Limit the functions that the stack tracer
4407 will trace at boot up. function-list is a comma separated
4408 list of functions. This list can be changed at run
4409 time by the stack_trace_filter file in the debugfs
4410 tracing directory. Note, this enables stack tracing
4411 and the stacktrace above is not needed.
4415 Set the STI (builtin display/keyboard on the HP-PARISC
4416 machines) console (graphic card) which should be used
4417 as the initial boot-console.
4418 See also comment in drivers/video/console/sticore.c.
4421 See comment in drivers/video/console/sticore.c.
4424 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4426 sunrpc.min_resvport=
4427 sunrpc.max_resvport=
4429 SunRPC servers often require that client requests
4430 originate from a privileged port (i.e. a port in the
4431 range 0 < portnr < 1024).
4432 An administrator who wishes to reserve some of these
4433 ports for other uses may adjust the range that the
4434 kernel's sunrpc client considers to be privileged
4435 using these two parameters to set the minimum and
4436 maximum port values.
4438 sunrpc.svc_rpc_per_connection_limit=
4440 Limit the number of requests that the server will
4441 process in parallel from a single connection.
4442 The default value is 0 (no limit).
4446 Control how the NFS server code allocates CPUs to
4447 service thread pools. Depending on how many NICs
4448 you have and where their interrupts are bound, this
4449 option will affect which CPUs will do NFS serving.
4450 Note: this parameter cannot be changed while the
4451 NFS server is running.
4453 auto the server chooses an appropriate mode
4454 automatically using heuristics
4455 global a single global pool contains all CPUs
4456 percpu one pool for each CPU
4457 pernode one pool for each NUMA node (equivalent
4458 to global on non-NUMA machines)
4460 sunrpc.tcp_slot_table_entries=
4461 sunrpc.udp_slot_table_entries=
4463 Sets the upper limit on the number of simultaneous
4464 RPC calls that can be sent from the client to a
4465 server. Increasing these values may allow you to
4466 improve throughput, but will also increase the
4467 amount of memory reserved for use by the client.
4469 suspend.pm_test_delay=
4471 Sets the number of seconds to remain in a suspend test
4472 mode before resuming the system (see
4473 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4474 is set. Default value is 5.
4477 [KNL] Enable accounting of swap in memory resource
4478 controller if no parameter or 1 is given or disable
4479 it if 0 is given (See Documentation/cgroup-v1/memory.txt)
4481 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4482 Format: { <int> | force | noforce }
4483 <int> -- Number of I/O TLB slabs
4484 force -- force using of bounce buffers even if they
4485 wouldn't be automatically used by the kernel
4486 noforce -- Never use bounce buffers (for debugging)
4490 sysfs.deprecated=0|1 [KNL]
4491 Enable/disable old style sysfs layout for old udev
4492 on older distributions. When this option is enabled
4493 very new udev will not work anymore. When this option
4494 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4495 in older udev will not work anymore.
4496 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4497 the kernel configuration.
4499 sysrq_always_enabled
4501 Ignore sysrq setting - this boot parameter will
4502 neutralize any effect of /proc/sys/kernel/sysrq.
4503 Useful for debugging.
4505 tcpmhash_entries= [KNL,NET]
4506 Set the number of tcp_metrics_hash slots.
4507 Default value is 8192 or 16384 depending on total
4508 ram pages. This is used to specify the TCP metrics
4509 cache size. See Documentation/networking/ip-sysctl.txt
4510 "tcp_no_metrics_save" section for more details.
4514 test_suspend= [SUSPEND][,N]
4515 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4516 standby suspend) or "freeze" (for suspend type freeze)
4517 as the system sleep state during system startup with
4518 the optional capability to repeat N number of times.
4519 The system is woken from this state using a
4520 wakeup-capable RTC alarm.
4522 thash_entries= [KNL,NET]
4523 Set number of hash buckets for TCP connection
4525 thermal.act= [HW,ACPI]
4526 -1: disable all active trip points in all thermal zones
4527 <degrees C>: override all lowest active trip points
4529 thermal.crt= [HW,ACPI]
4530 -1: disable all critical trip points in all thermal zones
4531 <degrees C>: override all critical trip points
4533 thermal.nocrt= [HW,ACPI]
4534 Set to disable actions on ACPI thermal zone
4535 critical and hot trip points.
4537 thermal.off= [HW,ACPI]
4538 1: disable ACPI thermal control
4540 thermal.psv= [HW,ACPI]
4541 -1: disable all passive trip points
4542 <degrees C>: override all passive trip points to this
4545 thermal.tzp= [HW,ACPI]
4546 Specify global default ACPI thermal zone polling rate
4547 <deci-seconds>: poll all this frequency
4548 0: no polling (default)
4551 Force threading of all interrupt handlers except those
4552 marked explicitly IRQF_NO_THREAD.
4555 Enable the Transcendent memory driver if built-in.
4557 tmem.cleancache=0|1 [KNL, XEN]
4558 Default is on (1). Disable the usage of the cleancache
4559 API to send anonymous pages to the hypervisor.
4561 tmem.frontswap=0|1 [KNL, XEN]
4562 Default is on (1). Disable the usage of the frontswap
4563 API to send swap pages to the hypervisor. If disabled
4564 the selfballooning and selfshrinking are force disabled.
4566 tmem.selfballooning=0|1 [KNL, XEN]
4567 Default is on (1). Disable the driving of swap pages
4570 tmem.selfshrinking=0|1 [KNL, XEN]
4571 Default is on (1). Partial swapoff that immediately
4572 transfers pages from Xen hypervisor back to the
4573 kernel based on different criteria.
4577 Specify if the kernel should make use of the cpu
4578 topology information if the hardware supports this.
4579 The scheduler will make use of this information and
4580 e.g. base its process migration decisions on it.
4583 topology_updates= [KNL, PPC, NUMA]
4585 Specify if the kernel should ignore (off)
4586 topology updates sent by the hypervisor to this
4591 tpm_suspend_pcr=[HW,TPM]
4592 Format: integer pcr id
4593 Specify that at suspend time, the tpm driver
4594 should extend the specified pcr with zeros,
4595 as a workaround for some chips which fail to
4596 flush the last written pcr on TPM_SaveState.
4597 This will guarantee that all the other pcrs
4600 trace_buf_size=nn[KMG]
4601 [FTRACE] will set tracing buffer size on each cpu.
4603 trace_event=[event-list]
4604 [FTRACE] Set and start specified trace events in order
4605 to facilitate early boot debugging. The event-list is a
4606 comma separated list of trace events to enable. See
4607 also Documentation/trace/events.rst
4609 trace_options=[option-list]
4610 [FTRACE] Enable or disable tracer options at boot.
4611 The option-list is a comma delimited list of options
4612 that can be enabled or disabled just as if you were
4613 to echo the option name into
4615 /sys/kernel/debug/tracing/trace_options
4617 For example, to enable stacktrace option (to dump the
4618 stack trace of each event), add to the command line:
4620 trace_options=stacktrace
4622 See also Documentation/trace/ftrace.rst "trace options"
4626 Have the tracepoints sent to printk as well as the
4627 tracing ring buffer. This is useful for early boot up
4628 where the system hangs or reboots and does not give the
4629 option for reading the tracing buffer or performing a
4630 ftrace_dump_on_oops.
4632 To turn off having tracepoints sent to printk,
4633 echo 0 > /proc/sys/kernel/tracepoint_printk
4634 Note, echoing 1 into this file without the
4635 tracepoint_printk kernel cmdline option has no effect.
4639 Having tracepoints sent to printk() and activating high
4640 frequency tracepoints such as irq or sched, can cause
4641 the system to live lock.
4644 [FTRACE] enable this option to disable tracing when a
4645 warning is hit. This turns off "tracing_on". Tracing can
4646 be enabled again by echoing '1' into the "tracing_on"
4647 file located in /sys/kernel/debug/tracing/
4649 This option is useful, as it disables the trace before
4650 the WARNING dump is called, which prevents the trace to
4651 be filled with content caused by the warning output.
4653 This option can also be set at run time via the sysctl
4654 option: kernel/traceoff_on_warning
4656 transparent_hugepage=
4658 Format: [always|madvise|never]
4659 Can be used to control the default behavior of the system
4660 with respect to transparent hugepages.
4661 See Documentation/admin-guide/mm/transhuge.rst
4664 tsc= Disable clocksource stability checks for TSC.
4666 [x86] reliable: mark tsc clocksource as reliable, this
4667 disables clocksource verification at runtime, as well
4668 as the stability checks done at bootup. Used to enable
4669 high-resolution timer mode on older hardware, and in
4670 virtualized environment.
4671 [x86] noirqtime: Do not use TSC to do irq accounting.
4672 Used to run time disable IRQ_TIME_ACCOUNTING on any
4673 platforms where RDTSC is slow and this accounting
4675 [x86] unstable: mark the TSC clocksource as unstable, this
4676 marks the TSC unconditionally unstable at bootup and
4677 avoids any further wobbles once the TSC watchdog notices.
4679 turbografx.map[2|3]= [HW,JOY]
4680 TurboGraFX parallel port interface
4682 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4683 See also Documentation/input/devices/joystick-parport.rst
4685 udbg-immortal [PPC] When debugging early kernel crashes that
4686 happen after console_init() and before a proper
4687 console driver takes over, this boot options might
4688 help "seeing" what's going on.
4690 uhash_entries= [KNL,NET]
4691 Set number of hash buckets for UDP/UDP-Lite connections
4694 [USB] Ignore overcurrent events (default N).
4695 Some badly-designed motherboards generate lots of
4696 bogus events, for ports that aren't wired to
4697 anything. Set this parameter to avoid log spamming.
4698 Note that genuine overcurrent events won't be
4702 [X86] Cause panic on unknown NMI.
4704 usbcore.authorized_default=
4705 [USB] Default USB device authorization:
4706 (default -1 = authorized except for wireless USB,
4707 0 = not authorized, 1 = authorized)
4709 usbcore.autosuspend=
4710 [USB] The autosuspend time delay (in seconds) used
4711 for newly-detected USB devices (default 2). This
4712 is the time required before an idle device will be
4713 autosuspended. Devices for which the delay is set
4714 to a negative value won't be autosuspended at all.
4716 usbcore.usbfs_snoop=
4717 [USB] Set to log all usbfs traffic (default 0 = off).
4719 usbcore.usbfs_snoop_max=
4720 [USB] Maximum number of bytes to snoop in each URB
4723 usbcore.blinkenlights=
4724 [USB] Set to cycle leds on hubs (default 0 = off).
4726 usbcore.old_scheme_first=
4727 [USB] Start with the old device initialization
4728 scheme, applies only to low and full-speed devices
4731 usbcore.usbfs_memory_mb=
4732 [USB] Memory limit (in MB) for buffers allocated by
4733 usbfs (default = 16, 0 = max = 2047).
4735 usbcore.use_both_schemes=
4736 [USB] Try the other device initialization scheme
4737 if the first one fails (default 1 = enabled).
4739 usbcore.initial_descriptor_timeout=
4740 [USB] Specifies timeout for the initial 64-byte
4741 USB_REQ_GET_DESCRIPTOR request in milliseconds
4742 (default 5000 = 5.0 seconds).
4744 usbcore.nousb [USB] Disable the USB subsystem
4747 [USB] A list of quirk entries to augment the built-in
4748 usb core quirk list. List entries are separated by
4749 commas. Each entry has the form
4750 VendorID:ProductID:Flags. The IDs are 4-digit hex
4751 numbers and Flags is a set of letters. Each letter
4752 will change the built-in quirk; setting it if it is
4753 clear and clearing it if it is set. The letters have
4754 the following meanings:
4755 a = USB_QUIRK_STRING_FETCH_255 (string
4756 descriptors must not be fetched using
4758 b = USB_QUIRK_RESET_RESUME (device can't resume
4759 correctly so reset it instead);
4760 c = USB_QUIRK_NO_SET_INTF (device can't handle
4761 Set-Interface requests);
4762 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
4763 handle its Configuration or Interface
4765 e = USB_QUIRK_RESET (device can't be reset
4766 (e.g morph devices), don't use reset);
4767 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
4768 more interface descriptions than the
4769 bNumInterfaces count, and can't handle
4770 talking to these interfaces);
4771 g = USB_QUIRK_DELAY_INIT (device needs a pause
4772 during initialization, after we read
4773 the device descriptor);
4774 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
4775 high speed and super speed interrupt
4776 endpoints, the USB 2.0 and USB 3.0 spec
4777 require the interval in microframes (1
4778 microframe = 125 microseconds) to be
4779 calculated as interval = 2 ^
4781 Devices with this quirk report their
4782 bInterval as the result of this
4783 calculation instead of the exponent
4784 variable used in the calculation);
4785 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
4786 handle device_qualifier descriptor
4788 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
4789 generates spurious wakeup, ignore
4790 remote wakeup capability);
4791 k = USB_QUIRK_NO_LPM (device can't handle Link
4793 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
4794 (Device reports its bInterval as linear
4795 frames instead of the USB 2.0
4797 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
4798 to be disconnected before suspend to
4799 prevent spurious wakeup);
4800 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
4801 pause after every control message);
4802 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
4803 delay after resetting its port);
4804 Example: quirks=0781:5580:bk,0a5c:5834:gij
4807 [USBHID] The interval which mice are to be polled at.
4810 [USBHID] The interval which joysticks are to be polled at.
4813 [USBHID] The interval which keyboards are to be polled at.
4815 usb-storage.delay_use=
4816 [UMS] The delay in seconds before a new device is
4817 scanned for Logical Units (default 1).
4820 [UMS] A list of quirks entries to supplement or
4821 override the built-in unusual_devs list. List
4822 entries are separated by commas. Each entry has
4823 the form VID:PID:Flags where VID and PID are Vendor
4824 and Product ID values (4-digit hex numbers) and
4825 Flags is a set of characters, each corresponding
4826 to a common usb-storage quirk flag as follows:
4827 a = SANE_SENSE (collect more than 18 bytes
4829 b = BAD_SENSE (don't collect more than 18
4830 bytes of sense data);
4831 c = FIX_CAPACITY (decrease the reported
4832 device capacity by one sector);
4833 d = NO_READ_DISC_INFO (don't use
4834 READ_DISC_INFO command);
4835 e = NO_READ_CAPACITY_16 (don't use
4836 READ_CAPACITY_16 command);
4837 f = NO_REPORT_OPCODES (don't use report opcodes
4839 g = MAX_SECTORS_240 (don't transfer more than
4840 240 sectors at a time, uas only);
4841 h = CAPACITY_HEURISTICS (decrease the
4842 reported device capacity by one
4843 sector if the number is odd);
4844 i = IGNORE_DEVICE (don't bind to this
4846 j = NO_REPORT_LUNS (don't use report luns
4848 l = NOT_LOCKABLE (don't try to lock and
4849 unlock ejectable media);
4850 m = MAX_SECTORS_64 (don't transfer more
4851 than 64 sectors = 32 KB at a time);
4852 n = INITIAL_READ10 (force a retry of the
4853 initial READ(10) command);
4854 o = CAPACITY_OK (accept the capacity
4855 reported by the device);
4856 p = WRITE_CACHE (the device cache is ON
4858 r = IGNORE_RESIDUE (the device reports
4859 bogus residue values);
4860 s = SINGLE_LUN (the device has only one
4862 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
4863 commands, uas only);
4864 u = IGNORE_UAS (don't bind to the uas driver);
4865 w = NO_WP_DETECT (don't test whether the
4866 medium is write-protected).
4867 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
4868 even if the device claims no cache)
4869 Example: quirks=0419:aaf5:rl,0421:0433:rc
4871 user_debug= [KNL,ARM]
4873 See arch/arm/Kconfig.debug help text.
4874 1 - undefined instruction events
4876 4 - invalid data aborts
4879 Example: user_debug=31
4882 [X86] Flags controlling user PTE allocations.
4884 nohigh = do not allocate PTE pages in
4885 HIGHMEM regardless of setting
4889 On X86_32, this is an alias for vdso32=. Otherwise:
4891 vdso=1: enable VDSO (the default)
4892 vdso=0: disable VDSO mapping
4894 vdso32= [X86] Control the 32-bit vDSO
4895 vdso32=1: enable 32-bit VDSO
4896 vdso32=0 or vdso32=2: disable 32-bit VDSO
4898 See the help text for CONFIG_COMPAT_VDSO for more
4899 details. If CONFIG_COMPAT_VDSO is set, the default is
4900 vdso32=0; otherwise, the default is vdso32=1.
4902 For compatibility with older kernels, vdso32=2 is an
4905 Try vdso32=0 if you encounter an error that says:
4906 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
4909 vector=percpu: enable percpu vector domain
4911 video= [FB] Frame buffer configuration
4912 See Documentation/fb/modedb.txt.
4914 video.brightness_switch_enabled= [0,1]
4915 If set to 1, on receiving an ACPI notify event
4916 generated by hotkey, video driver will adjust brightness
4917 level and then send out the event to user space through
4918 the allocated input device; If set to 0, video driver
4919 will only send out the event without touching backlight
4924 [VMMIO] Memory mapped virtio (platform) device.
4926 <size>@<baseaddr>:<irq>[:<id>]
4928 <size> := size (can use standard suffixes
4930 <baseaddr> := physical base address
4931 <irq> := interrupt number (as passed to
4933 <id> := (optional) platform device id
4935 virtio_mmio.device=1K@0x100b0000:48:7
4937 Can be used multiple times for multiple devices.
4939 vga= [BOOT,X86-32] Select a particular video mode
4940 See Documentation/x86/boot.txt and
4941 Documentation/svga.txt.
4942 Use vga=ask for menu.
4943 This is actually a boot loader parameter; the value is
4944 passed to the kernel using a special protocol.
4946 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
4947 May slow down system boot speed, especially when
4948 enabled on systems with a large amount of memory.
4949 All options are enabled by default, and this
4950 interface is meant to allow for selectively
4951 enabling or disabling specific virtual memory
4954 Available options are:
4955 P Enable page structure init time poisoning
4956 - Disable all of the above options
4958 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
4959 size of <nn>. This can be used to increase the
4960 minimum size (128MB on x86). It can also be used to
4961 decrease the size and leave more room for directly
4964 vmcp_cma=nn[MG] [KNL,S390]
4965 Sets the memory size reserved for contiguous memory
4966 allocations for the vmcp device driver.
4968 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
4971 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
4974 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
4978 Controls the behavior of vsyscalls (i.e. calls to
4979 fixed addresses of 0xffffffffff600x00 from legacy
4980 code). Most statically-linked binaries and older
4981 versions of glibc use these calls. Because these
4982 functions are at fixed addresses, they make nice
4983 targets for exploits that can control RIP.
4985 emulate [default] Vsyscalls turn into traps and are
4986 emulated reasonably safely.
4988 native Vsyscalls are native syscall instructions.
4989 This is a little bit faster than trapping
4990 and makes a few dynamic recompilers work
4991 better than they would in emulation mode.
4992 It also makes exploits much easier to write.
4994 none Vsyscalls don't work at all. This makes
4995 them quite hard to use for exploits but
4996 might break your system.
4998 vt.color= [VT] Default text color.
4999 Format: 0xYX, X = foreground, Y = background.
5000 Default: 0x07 = light gray on black.
5002 vt.cur_default= [VT] Default cursor shape.
5003 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5004 the parameters of the <Esc>[?A;B;Cc escape sequence;
5005 see VGA-softcursor.txt. Default: 2 = underline.
5007 vt.default_blu= [VT]
5008 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5009 Change the default blue palette of the console.
5010 This is a 16-member array composed of values
5013 vt.default_grn= [VT]
5014 Format: <green0>,<green1>,<green2>,...,<green15>
5015 Change the default green palette of the console.
5016 This is a 16-member array composed of values
5019 vt.default_red= [VT]
5020 Format: <red0>,<red1>,<red2>,...,<red15>
5021 Change the default red palette of the console.
5022 This is a 16-member array composed of values
5028 Set system-wide default UTF-8 mode for all tty's.
5029 Default is 1, i.e. UTF-8 mode is enabled for all
5030 newly opened terminals.
5032 vt.global_cursor_default=
5035 Set system-wide default for whether a cursor
5036 is shown on new VTs. Default is -1,
5037 i.e. cursors will be created by default unless
5038 overridden by individual drivers. 0 will hide
5039 cursors, 1 will display them.
5041 vt.italic= [VT] Default color for italic text; 0-15.
5044 vt.underline= [VT] Default color for underlined text; 0-15.
5047 watchdog timers [HW,WDT] For information on watchdog timers,
5048 see Documentation/watchdog/watchdog-parameters.txt
5049 or other driver-specific files in the
5050 Documentation/watchdog/ directory.
5052 workqueue.watchdog_thresh=
5053 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5054 warn stall conditions and dump internal state to
5055 help debugging. 0 disables workqueue stall
5056 detection; otherwise, it's the stall threshold
5057 duration in seconds. The default value is 30 and
5058 it can be updated at runtime by writing to the
5059 corresponding sysfs file.
5061 workqueue.disable_numa
5062 By default, all work items queued to unbound
5063 workqueues are affine to the NUMA nodes they're
5064 issued on, which results in better behavior in
5065 general. If NUMA affinity needs to be disabled for
5066 whatever reason, this option can be used. Note
5067 that this also can be controlled per-workqueue for
5068 workqueues visible under /sys/bus/workqueue/.
5070 workqueue.power_efficient
5071 Per-cpu workqueues are generally preferred because
5072 they show better performance thanks to cache
5073 locality; unfortunately, per-cpu workqueues tend to
5074 be more power hungry than unbound workqueues.
5076 Enabling this makes the per-cpu workqueues which
5077 were observed to contribute significantly to power
5078 consumption unbound, leading to measurably lower
5079 power usage at the cost of small performance
5082 The default value of this parameter is determined by
5083 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5085 workqueue.debug_force_rr_cpu
5086 Workqueue used to implicitly guarantee that work
5087 items queued without explicit CPU specified are put
5088 on the local CPU. This guarantee is no longer true
5089 and while local CPU is still preferred work items
5090 may be put on foreign CPUs. This debug option
5091 forces round-robin CPU selection to flush out
5092 usages which depend on the now broken guarantee.
5093 When enabled, memory and cache locality will be
5096 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5097 default x2apic cluster mode on platforms
5100 x86_intel_mid_timer= [X86-32,APBT]
5101 Choose timer option for x86 Intel MID platform.
5102 Two valid options are apbt timer only and lapic timer
5103 plus one apbt timer for broadcast timer.
5104 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5106 xen_512gb_limit [KNL,X86-64,XEN]
5107 Restricts the kernel running paravirtualized under Xen
5108 to use only up to 512 GB of RAM. The reason to do so is
5109 crash analysis tools and Xen tools for doing domain
5110 save/restore/migration must be enabled to handle larger
5113 xen_emul_unplug= [HW,X86,XEN]
5114 Unplug Xen emulated devices
5115 Format: [unplug0,][unplug1]
5116 ide-disks -- unplug primary master IDE devices
5117 aux-ide-disks -- unplug non-primary-master IDE devices
5118 nics -- unplug network devices
5119 all -- unplug all emulated devices (NICs and IDE disks)
5120 unnecessary -- unplugging emulated devices is
5121 unnecessary even if the host did not respond to
5123 never -- do not unplug even if version check succeeds
5125 xen_nopvspin [X86,XEN]
5126 Disables the ticketlock slowpath using Xen PV
5130 Disables the PV optimizations forcing the HVM guest to
5131 run as generic HVM guest with no PV drivers.
5133 xen_scrub_pages= [XEN]
5134 Boolean option to control scrubbing pages before giving them back
5135 to Xen, for use by other domains. Can be also changed at runtime
5136 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5137 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5139 xirc2ps_cs= [NET,PCMCIA]
5141 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5143 xhci-hcd.quirks [USB,KNL]
5144 A hex value specifying bitmask with supplemental xhci
5145 host controller quirks. Meaning of each bit can be
5146 consulted in header drivers/usb/host/xhci.h.