4 The following is a consolidated list of the kernel parameters as
5 implemented by the __setup(), core_param() and module_param() macros
6 and sorted into English Dictionary order (defined as ignoring all
7 punctuation and sorting digits before letters in a case insensitive
8 manner), and with descriptions where known.
10 The kernel parses parameters from the kernel command line up to "--";
11 if it doesn't recognize a parameter and it doesn't contain a '.', the
12 parameter gets passed to init: parameters with '=' go into init's
13 environment, others are passed as command line arguments to init.
14 Everything after "--" is passed as an argument to init.
16 Module parameters can be specified in two ways: via the kernel command
17 line with a module name prefix, or via modprobe, e.g.:
19 (kernel command line) usbcore.blinkenlights=1
20 (modprobe command line) modprobe usbcore blinkenlights=1
22 Parameters for modules which are built into the kernel need to be
23 specified on the kernel command line. modprobe looks through the
24 kernel command line (/proc/cmdline) and collects module parameters
25 when it loads a module, so the kernel command line can be used for
28 Hyphens (dashes) and underscores are equivalent in parameter names, so
29 log_buf_len=1M print-fatal-signals=1
30 can also be entered as
31 log-buf-len=1M print_fatal_signals=1
33 Double-quotes can be used to protect spaces in values, e.g.:
34 param="spaces in here"
36 This document may not be entirely up to date and comprehensive. The command
37 "modinfo -p ${modulename}" shows a current list of all parameters of a loadable
38 module. Loadable modules, after being loaded into the running kernel, also
39 reveal their parameters in /sys/module/${modulename}/parameters/. Some of these
40 parameters may be changed at runtime by the command
41 "echo -n ${value} > /sys/module/${modulename}/parameters/${parm}".
43 The parameters listed below are only valid if certain kernel build options were
44 enabled and if respective hardware is present. The text in square brackets at
45 the beginning of each description states the restrictions within which a
46 parameter is applicable:
48 ACPI ACPI support is enabled.
49 AGP AGP (Accelerated Graphics Port) is enabled.
50 ALSA ALSA sound support is enabled.
51 APIC APIC support is enabled.
52 APM Advanced Power Management support is enabled.
53 ARM ARM architecture is enabled.
54 AVR32 AVR32 architecture is enabled.
55 AX25 Appropriate AX.25 support is enabled.
56 BLACKFIN Blackfin architecture is enabled.
57 CLK Common clock infrastructure is enabled.
58 CMA Contiguous Memory Area support is enabled.
59 DRM Direct Rendering Management support is enabled.
60 DYNAMIC_DEBUG Build in debug messages and enable them at runtime
61 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled
62 EFI EFI Partitioning (GPT) is enabled
63 EIDE EIDE/ATAPI support is enabled.
64 EVM Extended Verification Module
65 FB The frame buffer device is enabled.
66 FTRACE Function tracing enabled.
67 GCOV GCOV profiling is enabled.
68 HW Appropriate hardware is enabled.
69 IA-64 IA-64 architecture is enabled.
70 IMA Integrity measurement architecture is enabled.
71 IOSCHED More than one I/O scheduler is enabled.
72 IP_PNP IP DHCP, BOOTP, or RARP is enabled.
73 IPV6 IPv6 support is enabled.
74 ISAPNP ISA PnP code is enabled.
75 ISDN Appropriate ISDN support is enabled.
76 JOY Appropriate joystick support is enabled.
77 KGDB Kernel debugger support is enabled.
78 KVM Kernel Virtual Machine support is enabled.
79 LIBATA Libata driver is enabled
80 LP Printer support is enabled.
81 LOOP Loopback device support is enabled.
82 M68k M68k architecture is enabled.
83 These options have more detailed description inside of
84 Documentation/m68k/kernel-options.txt.
85 MDA MDA console support is enabled.
86 MIPS MIPS architecture is enabled.
87 MOUSE Appropriate mouse support is enabled.
88 MSI Message Signaled Interrupts (PCI).
89 MTD MTD (Memory Technology Device) support is enabled.
90 NET Appropriate network support is enabled.
91 NUMA NUMA support is enabled.
92 NFS Appropriate NFS support is enabled.
93 OSS OSS sound support is enabled.
94 PV_OPS A paravirtualized kernel is enabled.
95 PARIDE The ParIDE (parallel port IDE) subsystem is enabled.
96 PARISC The PA-RISC architecture is enabled.
97 PCI PCI bus support is enabled.
98 PCIE PCI Express support is enabled.
99 PCMCIA The PCMCIA subsystem is enabled.
100 PNP Plug & Play support is enabled.
101 PPC PowerPC architecture is enabled.
102 PPT Parallel port support is enabled.
103 PS2 Appropriate PS/2 support is enabled.
104 RAM RAM disk support is enabled.
105 S390 S390 architecture is enabled.
106 SCSI Appropriate SCSI support is enabled.
107 A lot of drivers have their options described inside
108 the Documentation/scsi/ sub-directory.
109 SECURITY Different security models are enabled.
110 SELINUX SELinux support is enabled.
111 APPARMOR AppArmor support is enabled.
112 SERIAL Serial support is enabled.
113 SH SuperH architecture is enabled.
114 SMP The kernel is an SMP kernel.
115 SPARC Sparc architecture is enabled.
116 SWSUSP Software suspend (hibernation) is enabled.
117 SUSPEND System suspend states are enabled.
118 TPM TPM drivers are enabled.
119 TS Appropriate touchscreen support is enabled.
120 UMS USB Mass Storage support is enabled.
121 USB USB support is enabled.
122 USBHID USB Human Interface Device support is enabled.
123 V4L Video For Linux support is enabled.
124 VMMIO Driver for memory mapped virtio devices is enabled.
125 VGA The VGA console has been enabled.
126 VT Virtual terminal support is enabled.
127 WDT Watchdog support is enabled.
128 XT IBM PC/XT MFM hard disk support is enabled.
129 X86-32 X86-32, aka i386 architecture is enabled.
130 X86-64 X86-64 architecture is enabled.
131 More X86-64 boot options can be found in
132 Documentation/x86/x86_64/boot-options.txt .
133 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64)
134 XEN Xen support is enabled
136 In addition, the following text indicates that the option:
138 BUGS= Relates to possible processor bugs on the said processor.
139 KNL Is a kernel start-up parameter.
140 BOOT Is a boot loader parameter.
142 Parameters denoted with BOOT are actually interpreted by the boot
143 loader, and have no meaning to the kernel directly.
144 Do not modify the syntax of boot loader parameters without extreme
145 need or coordination with <Documentation/x86/boot.txt>.
147 There are also arch-specific kernel-parameters not documented here.
148 See for example <Documentation/x86/x86_64/boot-options.txt>.
150 Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
151 a trailing = on the name of any parameter states that that parameter will
152 be entered as an environment variable, whereas its absence indicates that
153 it will appear as a kernel argument readable via /proc/cmdline by programs
154 running once the system is up.
156 The number of kernel parameters is not limited, but the length of the
157 complete command line (parameters including spaces etc.) is limited to
158 a fixed number of characters. This limit depends on the architecture
159 and is between 256 and 4096 characters. It is defined in the file
160 ./include/asm/setup.h as COMMAND_LINE_SIZE.
162 Finally, the [KMG] suffix is commonly described after a number of kernel
163 parameter values. These 'K', 'M', and 'G' letters represent the _binary_
164 multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
165 bytes respectively. Such letter suffixes can also be entirely omitted.
168 acpi= [HW,ACPI,X86,ARM64]
169 Advanced Configuration and Power Interface
170 Format: { force | off | strict | noirq | rsdt }
171 force -- enable ACPI if default was off
172 off -- disable ACPI if default was on
173 noirq -- do not use ACPI for IRQ routing
174 strict -- Be less tolerant of platforms that are not
175 strictly ACPI specification compliant.
176 rsdt -- prefer RSDT over (default) XSDT
177 copy_dsdt -- copy DSDT to memory
178 For ARM64, ONLY "acpi=off" or "acpi=force" are available
180 See also Documentation/power/runtime_pm.txt, pci=noacpi
182 acpi_rsdp= [ACPI,EFI,KEXEC]
183 Pass the RSDP address to the kernel, mostly used
184 on machines running EFI runtime service to boot the
185 second kernel for kdump.
187 acpi_apic_instance= [ACPI, IOAPIC]
189 2: use 2nd APIC table, if available
190 1,0: use 1st APIC table
193 acpi_backlight= [HW,ACPI]
194 acpi_backlight=vendor
196 If set to vendor, prefer vendor specific driver
197 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
198 of the ACPI video.ko driver.
200 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
201 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
203 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
204 debug output. Bits in debug_layer correspond to a
205 _COMPONENT in an ACPI source file, e.g.,
206 #define _COMPONENT ACPI_PCI_COMPONENT
207 Bits in debug_level correspond to a level in
208 ACPI_DEBUG_PRINT statements, e.g.,
209 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
210 The debug_level mask defaults to "info". See
211 Documentation/acpi/debug.txt for more information about
212 debug layers and levels.
214 Enable processor driver info messages:
215 acpi.debug_layer=0x20000000
216 Enable PCI/PCI interrupt routing info messages:
217 acpi.debug_layer=0x400000
218 Enable AML "Debug" output, i.e., stores to the Debug
219 object while interpreting AML:
220 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
221 Enable all messages related to ACPI hardware:
222 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
224 Some values produce so much output that the system is
225 unusable. The "log_buf_len" parameter may be useful
226 if you need to capture more output.
228 acpi_force_table_verification [HW,ACPI]
229 Enable table checksum verification during early stage.
230 By default, this is disabled due to x86 early mapping
233 acpi_irq_balance [HW,ACPI]
234 ACPI will balance active IRQs
237 acpi_irq_nobalance [HW,ACPI]
238 ACPI will not move active IRQs (default)
241 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
242 Format: <irq>,<irq>...
244 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
246 Format: <irq>,<irq>...
248 acpi_no_auto_serialize [HW,ACPI]
249 Disable auto-serialization of AML methods
250 AML control methods that contain the opcodes to create
251 named objects will be marked as "Serialized" by the
252 auto-serialization feature.
253 This feature is enabled by default.
254 This option allows to turn off the feature.
256 acpi_no_static_ssdt [HW,ACPI]
257 Disable installation of static SSDTs at early boot time
258 By default, SSDTs contained in the RSDT/XSDT will be
259 installed automatically and they will appear under
260 /sys/firmware/acpi/tables.
261 This option turns off this feature.
262 Note that specifying this option does not affect
263 dynamic table installation which will install SSDT
264 tables to /sys/firmware/acpi/tables/dynamic.
266 acpica_no_return_repair [HW, ACPI]
267 Disable AML predefined validation mechanism
268 This mechanism can repair the evaluation result to make
269 the return objects more ACPI specification compliant.
270 This option is useful for developers to identify the
271 root cause of an AML interpreter issue when the issue
272 has something to do with the repair mechanism.
274 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
275 Format: To spoof as Windows 98: ="Microsoft Windows"
277 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
278 of 2 which is mandated by ACPI 6) as the supported ACPI
279 specification revision (when using this switch, it may
280 be necessary to carry out a cold reboot _twice_ in a
281 row to make it take effect on the platform firmware).
283 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
284 acpi_osi="string1" # add string1
285 acpi_osi="!string2" # remove string2
286 acpi_osi=!* # remove all strings
287 acpi_osi=! # disable all built-in OS vendor
289 acpi_osi= # disable all strings
291 'acpi_osi=!' can be used in combination with single or
292 multiple 'acpi_osi="string1"' to support specific OS
293 vendor string(s). Note that such command can only
294 affect the default state of the OS vendor strings, thus
295 it cannot affect the default state of the feature group
296 strings and the current state of the OS vendor strings,
297 specifying it multiple times through kernel command line
298 is meaningless. This command is useful when one do not
299 care about the state of the feature group strings which
300 should be controlled by the OSPM.
302 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
303 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
304 can make '_OSI("Windows 2000")' TRUE.
306 'acpi_osi=' cannot be used in combination with other
307 'acpi_osi=' command lines, the _OSI method will not
308 exist in the ACPI namespace. NOTE that such command can
309 only affect the _OSI support state, thus specifying it
310 multiple times through kernel command line is also
313 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
316 'acpi_osi=!*' can be used in combination with single or
317 multiple 'acpi_osi="string1"' to support specific
318 string(s). Note that such command can affect the
319 current state of both the OS vendor strings and the
320 feature group strings, thus specifying it multiple times
321 through kernel command line is meaningful. But it may
322 still not able to affect the final state of a string if
323 there are quirks related to this string. This command
324 is useful when one want to control the state of the
325 feature group strings to debug BIOS issues related to
328 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
329 '_OSI("Module Device")' FALSE.
330 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
331 '_OSI("Module Device")' TRUE.
332 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
334 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
336 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
337 they all will make '_OSI("Windows 2000")' TRUE.
340 Override the pmtimer bug detection: force the kernel
341 to assume that this machine's pmtimer latches its value
342 and always returns good values.
344 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
345 Format: { level | edge | high | low }
347 acpi_skip_timer_override [HW,ACPI]
348 Recognize and ignore IRQ0/pin2 Interrupt Override.
349 For broken nForce2 BIOS resulting in XT-PIC timer.
351 acpi_sleep= [HW,ACPI] Sleep options
352 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
353 old_ordering, nonvs, sci_force_enable }
354 See Documentation/power/video.txt for information on
356 s3_beep is for debugging; it makes the PC's speaker beep
357 as soon as the kernel's real-mode entry point is called.
358 s4_nohwsig prevents ACPI hardware signature from being
359 used during resume from hibernation.
360 old_ordering causes the ACPI 1.0 ordering of the _PTS
361 control method, with respect to putting devices into
362 low power states, to be enforced (the ACPI 2.0 ordering
363 of _PTS is used by default).
364 nonvs prevents the kernel from saving/restoring the
365 ACPI NVS memory during suspend/hibernation and resume.
366 sci_force_enable causes the kernel to set SCI_EN directly
367 on resume from S1/S3 (which is against the ACPI spec,
368 but some broken systems don't work without it).
370 acpi_use_timer_override [HW,ACPI]
371 Use timer override. For some broken Nvidia NF5 boards
372 that require a timer override, but don't have HPET
374 acpi_enforce_resources= [ACPI]
375 { strict | lax | no }
376 Check for resource conflicts between native drivers
377 and ACPI OperationRegions (SystemIO and SystemMemory
378 only). IO ports and memory declared in ACPI might be
379 used by the ACPI subsystem in arbitrary AML code and
380 can interfere with legacy drivers.
381 strict (default): access to resources claimed by ACPI
382 is denied; legacy drivers trying to access reserved
383 resources will fail to bind to device using them.
384 lax: access to resources claimed by ACPI is allowed;
385 legacy drivers trying to access reserved resources
386 will bind successfully but a warning message is logged.
387 no: ACPI OperationRegions are not marked as reserved,
388 no further checks are performed.
390 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
393 add_efi_memmap [EFI; X86] Include EFI memory map in
394 kernel's map of available physical RAM.
397 { off | try_unsupported }
398 off: disable AGP support
399 try_unsupported: try to drive unsupported chipsets
400 (may crash computer or cause data corruption)
403 See Documentation/sound/alsa/alsa-parameters.txt
406 Allow the default userspace alignment fault handler
407 behaviour to be specified. Bit 0 enables warnings,
408 bit 1 enables fixups, and bit 2 sends a segfault.
410 align_va_addr= [X86-64]
411 Align virtual addresses by clearing slice [14:12] when
412 allocating a VMA at process creation time. This option
413 gives you up to 3% performance improvement on AMD F15h
414 machines (where it is enabled by default) for a
415 CPU-intensive style benchmark, and it can vary highly in
416 a microbenchmark depending on workload and compiler.
418 32: only for 32-bit processes
419 64: only for 64-bit processes
420 on: enable for both 32- and 64-bit processes
421 off: disable for both 32- and 64-bit processes
423 alloc_snapshot [FTRACE]
424 Allocate the ftrace snapshot buffer on boot up when the
425 main buffer is allocated. This is handy if debugging
426 and you need to use tracing_snapshot() on boot up, and
427 do not want to use tracing_snapshot_alloc() as it needs
428 to be done where GFP_KERNEL allocations are allowed.
430 amd_iommu= [HW,X86-64]
431 Pass parameters to the AMD IOMMU driver in the system.
433 fullflush - enable flushing of IO/TLB entries when
434 they are unmapped. Otherwise they are
435 flushed before they will be reused, which
437 off - do not initialize any AMD IOMMU found in
439 force_isolation - Force device isolation for all
440 devices. The IOMMU driver is not
441 allowed anymore to lift isolation
442 requirements as needed. This option
443 does not override iommu=pt
445 amd_iommu_dump= [HW,X86-64]
446 Enable AMD IOMMU driver option to dump the ACPI table
447 for AMD IOMMU. With this option enabled, AMD IOMMU
448 driver will print ACPI tables for AMD IOMMU during
449 IOMMU initialization.
451 amijoy.map= [HW,JOY] Amiga joystick support
452 Map of devices attached to JOY0DAT and JOY1DAT
454 See also Documentation/input/joystick.txt
456 analog.map= [HW,JOY] Analog joystick and gamepad support
457 Specifies type or capabilities of an analog joystick
458 connected to one of 16 gameports
459 Format: <type1>,<type2>,..<type16>
462 Power management functions (SPARCstation-4/5 + deriv.)
464 Disable APC CPU standby support. SPARCstation-Fox does
465 not play well with APC CPU idle - disable it if you have
466 APC and your system crashes randomly.
468 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
469 Change the output verbosity whilst booting
470 Format: { quiet (default) | verbose | debug }
471 Change the amount of debugging information output
472 when initialising the APIC and IO-APIC components.
475 See Documentation/networking/ipv6.txt.
477 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
478 Limit apic dumping. The parameter defines the maximal
479 number of local apics being dumped. Also it is possible
480 to set it to "all" by meaning -- no limit here.
481 Format: { 1 (default) | 2 | ... | all }.
482 The parameter valid if only apic=debug or
483 apic=verbose is specified.
484 Example: apic=debug show_lapic=all
486 apm= [APM] Advanced Power Management
487 See header of arch/x86/kernel/apm_32.c.
489 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
490 Format: <io>,<irq>,<nodeID>
494 atarimouse= [HW,MOUSE] Atari Mouse
496 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
497 EzKey and similar keyboards
499 atkbd.reset= [HW] Reset keyboard during initialization
501 atkbd.set= [HW] Select keyboard code set
502 Format: <int> (2 = AT (default), 3 = PS/2)
504 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
507 atkbd.softraw= [HW] Choose between synthetic and real raw mode
508 Format: <bool> (0 = real, 1 = synthetic (default))
510 atkbd.softrepeat= [HW]
511 Use software keyboard repeat
513 audit= [KNL] Enable the audit sub-system
514 Format: { "0" | "1" } (0 = disabled, 1 = enabled)
515 0 - kernel audit is disabled and can not be enabled
516 until the next reboot
517 unset - kernel audit is initialized but disabled and
518 will be fully enabled by the userspace auditd.
519 1 - kernel audit is initialized and partially enabled,
520 storing at most audit_backlog_limit messages in
521 RAM until it is fully enabled by the userspace
525 audit_backlog_limit= [KNL] Set the audit queue size limit.
526 Format: <int> (must be >=0)
529 baycom_epp= [HW,AX25]
532 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
534 See header of drivers/net/hamradio/baycom_par.c.
536 baycom_ser_fdx= [HW,AX25]
537 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
538 Format: <io>,<irq>,<mode>[,<baud>]
539 See header of drivers/net/hamradio/baycom_ser_fdx.c.
541 baycom_ser_hdx= [HW,AX25]
542 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
543 Format: <io>,<irq>,<mode>
544 See header of drivers/net/hamradio/baycom_ser_hdx.c.
546 blkdevparts= Manual partition parsing of block device(s) for
547 embedded devices based on command line input.
548 See Documentation/block/cmdline-partition.txt
550 boot_delay= Milliseconds to delay each printk during boot.
551 Values larger than 10 seconds (10000) are changed to
555 bootmem_debug [KNL] Enable bootmem allocator debug messages.
557 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
558 bttv.radio= Most important insmod options are available as
560 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
563 bulk_remove=off [PPC] This parameter disables the use of the pSeries
564 firmware feature for flushing multiple hpte entries
567 c101= [NET] Moxa C101 synchronous serial card
569 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
570 Sometimes CPU hardware bugs make them report the cache
571 size incorrectly. The kernel will attempt work arounds
572 to fix known problems, but for some CPUs it is not
573 possible to determine what the correct size should be.
574 This option provides an override for these situations.
576 ca_keys= [KEYS] This parameter identifies a specific key(s) on
577 the system trusted keyring to be used for certificate
579 format: { id:<keyid> | builtin }
581 cca= [MIPS] Override the kernel pages' cache coherency
582 algorithm. Accepted values range from 0 to 7
583 inclusive. See arch/mips/include/asm/pgtable-bits.h
584 for platform specific values (SB1, Loongson3 and
587 ccw_timeout_log [S390]
588 See Documentation/s390/CommonIO for details.
590 cgroup_disable= [KNL] Disable a particular controller
591 Format: {name of the controller(s) to disable}
592 The effects of cgroup_disable=foo are:
593 - foo isn't auto-mounted if you mount all cgroups in
595 - foo isn't visible as an individually mountable
597 {Currently only "memory" controller deal with this and
598 cut the overhead, others just disable the usage. So
599 only cgroup_disable=memory is actually worthy}
601 checkreqprot [SELINUX] Set initial checkreqprot flag value.
602 Format: { "0" | "1" }
603 See security/selinux/Kconfig help text.
604 0 -- check protection applied by kernel (includes
605 any implied execute protection).
606 1 -- check protection requested by application.
607 Default value is set via a kernel config option.
608 Value can be changed at runtime via
609 /selinux/checkreqprot.
612 See Documentation/s390/CommonIO for details.
615 Prevents the clock framework from automatically gating
616 clocks that have not been explicitly enabled by a Linux
617 device driver but are enabled in hardware at reset or
618 by the bootloader/firmware. Note that this does not
619 force such clocks to be always-on nor does it reserve
620 those clocks in any way. This parameter is useful for
621 debug and development, but should not be needed on a
622 platform with proper driver support. For more
623 information, see Documentation/clk.txt.
625 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
627 Forces specified clocksource (if available) to be used
628 when calculating gettimeofday(). If specified
629 clocksource is not available, it defaults to PIT.
630 Format: { pit | tsc | cyclone | pmtmr }
632 clocksource= Override the default clocksource
634 Override the default clocksource and use the clocksource
635 with the name specified.
636 Some clocksource names to choose from, depending on
638 [all] jiffies (this is the base, fallback clocksource)
640 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
641 pxa_timer,timer3,32k_counter,timer0_1
643 [X86-32] pit,hpet,tsc;
644 scx200_hrt on Geode; cyclone on IBM x440
652 clearcpuid=BITNUM [X86]
653 Disable CPUID feature X for the kernel. See
654 arch/x86/include/asm/cpufeature.h for the valid bit
655 numbers. Note the Linux specific bits are not necessarily
656 stable over kernel options, but the vendor specific
658 Also note that user programs calling CPUID directly
659 or using the feature without checking anything
660 will still see it. This just prevents it from
661 being used by the kernel or shown in /proc/cpuinfo.
662 Also note the kernel might malfunction if you disable
665 cma=nn[MG]@[start[MG][-end[MG]]]
667 Sets the size of kernel global memory area for
668 contiguous memory allocations and optionally the
669 placement constraint by the physical address range of
670 memory allocations. A value of 0 disables CMA
671 altogether. For more information, see
672 include/linux/dma-contiguous.h
674 cmo_free_hint= [PPC] Format: { yes | no }
675 Specify whether pages are marked as being inactive
676 when they are freed. This is used in CMO environments
677 to determine OS memory pressure for page stealing by
681 coherent_pool=nn[KMG] [ARM,KNL]
682 Sets the size of memory pool for coherent, atomic dma
683 allocations, by default set to 256K.
685 code_bytes [X86] How many bytes of object code to print
690 com20020= [HW,NET] ARCnet - COM20020 chipset
692 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
694 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
698 ARCnet - COM90xx chipset (memory-mapped buffers)
699 Format: <io>[,<irq>[,<memstart>]]
701 condev= [HW,S390] console device
704 console= [KNL] Output console device and options.
706 tty<n> Use the virtual console device <n>.
710 Use the specified serial port. The options are of
711 the form "bbbbpnf", where "bbbb" is the baud rate,
712 "p" is parity ("n", "o", or "e"), "n" is number of
713 bits, and "f" is flow control ("r" for RTS or
714 omit it). Default is "9600n8".
716 See Documentation/serial-console.txt for more
718 Documentation/networking/netconsole.txt for an
721 uart[8250],io,<addr>[,options]
722 uart[8250],mmio,<addr>[,options]
723 uart[8250],mmio32,<addr>[,options]
724 uart[8250],0x<addr>[,options]
725 Start an early, polled-mode console on the 8250/16550
726 UART at the specified I/O port or MMIO address,
727 switching to the matching ttyS device later.
728 MMIO inter-register address stride is either 8-bit
729 (mmio) or 32-bit (mmio32).
730 If none of [io|mmio|mmio32], <addr> is assumed to be
731 equivalent to 'mmio'. 'options' are specified in the
732 same format described for ttyS above; if unspecified,
733 the h/w is not re-initialized.
735 hvc<n> Use the hypervisor console device <n>. This is for
736 both Xen and PowerPC hypervisors.
738 If the device connected to the port is not a TTY but a braille
739 device, prepend "brl," before the device type, for instance
741 For now, only VisioBraille is supported.
743 consoleblank= [KNL] The console blank (screen saver) timeout in
744 seconds. Defaults to 10*60 = 10mins. A value of 0
745 disables the blank timer.
748 [KNL] Change the default value for
749 /proc/<pid>/coredump_filter.
750 See also Documentation/filesystems/proc.txt.
752 cpuidle.off=1 [CPU_IDLE]
753 disable the cpuidle sub-system
755 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
757 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
759 crashkernel=size[KMG][@offset[KMG]]
760 [KNL] Using kexec, Linux can switch to a 'crash kernel'
761 upon panic. This parameter reserves the physical
762 memory region [offset, offset + size] for that kernel
763 image. If '@offset' is omitted, then a suitable offset
764 is selected automatically. Check
765 Documentation/kdump/kdump.txt for further details.
767 crashkernel=range1:size1[,range2:size2,...][@offset]
768 [KNL] Same as above, but depends on the memory
769 in the running system. The syntax of range is
770 start-[end] where start and end are both
771 a memory unit (amount[KMG]). See also
772 Documentation/kdump/kdump.txt for an example.
774 crashkernel=size[KMG],high
775 [KNL, x86_64] range could be above 4G. Allow kernel
776 to allocate physical memory region from top, so could
777 be above 4G if system have more than 4G ram installed.
778 Otherwise memory region will be allocated below 4G, if
780 It will be ignored if crashkernel=X is specified.
781 crashkernel=size[KMG],low
782 [KNL, x86_64] range under 4G. When crashkernel=X,high
783 is passed, kernel could allocate physical memory region
784 above 4G, that cause second kernel crash on system
785 that require some amount of low memory, e.g. swiotlb
786 requires at least 64M+32K low memory. Kernel would
787 try to allocate 72M below 4G automatically.
788 This one let user to specify own low range under 4G
789 for second kernel instead.
790 0: to disable low allocation.
791 It will be ignored when crashkernel=X,high is not used
792 or memory reserved is below 4G.
797 cs89x0_media= [HW,NET]
798 Format: { rj45 | aui | bnc }
801 See header of drivers/s390/block/dasd_devmap.c.
803 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
804 (one device per port)
805 Format: <port#>,<type>
806 See also Documentation/input/joystick-parport.txt
808 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
809 time. See Documentation/dynamic-debug-howto.txt for
810 details. Deprecated, see dyndbg.
812 debug [KNL] Enable kernel debugging (events log level).
815 [KNL] verbose self-tests
817 Print debugging info while doing the locking API
819 We default to 0 (no extra messages), setting it to
820 1 will print _a lot_ more information - normally
821 only useful to kernel developers.
823 debug_objects [KNL] Enable object debugging
826 [KNL] Disable object debugging
828 debug_guardpage_minorder=
829 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
830 parameter allows control of the order of pages that will
831 be intentionally kept free (and hence protected) by the
832 buddy allocator. Bigger value increase the probability
833 of catching random memory corruption, but reduce the
834 amount of memory for normal system use. The maximum
835 possible value is MAX_ORDER/2. Setting this parameter
836 to 1 or 2 should be enough to identify most random
837 memory corruption problems caused by bugs in kernel or
838 driver code when a CPU writes to (or reads from) a
839 random memory location. Note that there exists a class
840 of memory corruptions problems caused by buggy H/W or
841 F/W or by drivers badly programing DMA (basically when
842 memory is written at bus level and the CPU MMU is
843 bypassed) which are not detectable by
844 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
845 tracking down these problems.
848 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
849 parameter enables the feature at boot time. In
850 default, it is disabled. We can avoid allocating huge
851 chunk of memory for debug pagealloc if we don't enable
852 it at boot time and the system will work mostly same
853 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
854 on: enable the feature
856 debugpat [X86] Enable PAT debugging
858 decnet.addr= [HW,NET]
859 Format: <area>[,<node>]
860 See also Documentation/networking/decnet.txt.
863 [same as hugepagesz=] The size of the default
864 HugeTLB page size. This is the size represented by
865 the legacy /proc/ hugepages APIs, used for SHM, and
866 default size when mounting hugetlbfs filesystems.
867 Defaults to the default architecture's huge page size
871 Set number of hash buckets for dentry cache.
874 See Documentation/networking/ipv6.txt.
876 disable_cpu_apicid= [X86,APIC,SMP]
878 The number of initial APIC ID for the
879 corresponding CPU to be disabled at boot,
880 mostly used for the kdump 2nd kernel to
881 disable BSP to wake up multiple CPUs without
882 causing system reset or hang due to sending
885 disable_ddw [PPC/PSERIES]
886 Disable Dynamic DMA Window support. Use this if
887 to workaround buggy firmware.
890 See Documentation/networking/ipv6.txt.
892 disable_mtrr_cleanup [X86]
893 The kernel tries to adjust MTRR layout from continuous
894 to discrete, to make X server driver able to add WB
895 entry later. This parameter disables that.
897 disable_mtrr_trim [X86, Intel and AMD only]
898 By default the kernel will trim any uncacheable
899 memory out of your available memory pool based on
900 MTRR settings. This parameter disables that behavior,
901 possibly causing your machine to run very slowly.
903 disable_timer_pin_1 [X86]
904 Disable PIN 1 of APIC timer
905 Can be useful to work around chipset bugs.
907 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
908 this option disables the debugging code at boot.
910 dma_debug_entries=<number>
911 This option allows to tune the number of preallocated
912 entries for DMA-API debugging code. One entry is
913 required per DMA-API allocation. Use this if the
914 DMA-API debugging code disables itself because the
915 architectural default is too low.
917 dma_debug_driver=<driver_name>
918 With this option the DMA-API debugging driver
919 filter feature can be enabled at boot time. Just
920 pass the driver to filter for as the parameter.
921 The filter can be disabled or changed to another
922 driver later using sysfs.
924 drm_kms_helper.edid_firmware=[<connector>:]<file>
925 Broken monitors, graphic adapters and KVMs may
926 send no or incorrect EDID data sets. This parameter
927 allows to specify an EDID data set in the
928 /lib/firmware directory that is used instead.
929 Generic built-in EDID data sets are used, if one of
930 edid/1024x768.bin, edid/1280x1024.bin,
931 edid/1680x1050.bin, or edid/1920x1080.bin is given
932 and no file with the same name exists. Details and
933 instructions how to build your own EDID data are
934 available in Documentation/EDID/HOWTO.txt. An EDID
935 data set will only be used for a particular connector,
936 if its name and a colon are prepended to the EDID
941 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
942 module.dyndbg[="val"]
943 Enable debug messages at boot time. See
944 Documentation/dynamic-debug-howto.txt for details.
947 on enable eager fpu restore
948 off disable eager fpu restore
949 auto selects the default scheme, which automatically
950 enables eagerfpu restore for xsaveopt.
952 early_ioremap_debug [KNL]
953 Enable debug messages in early_ioremap support. This
954 is useful for tracking down temporary early mappings
955 which are not unmapped.
957 earlycon= [KNL] Output early console device and options.
960 Start an early, polled-mode console on a cadence serial
961 port at the specified address. The cadence serial port
962 must already be setup and configured. Options are not
965 uart[8250],io,<addr>[,options]
966 uart[8250],mmio,<addr>[,options]
967 uart[8250],mmio32,<addr>[,options]
968 uart[8250],0x<addr>[,options]
969 Start an early, polled-mode console on the 8250/16550
970 UART at the specified I/O port or MMIO address.
971 MMIO inter-register address stride is either 8-bit
972 (mmio) or 32-bit (mmio32).
973 If none of [io|mmio|mmio32], <addr> is assumed to be
974 equivalent to 'mmio'. 'options' are specified in the
975 same format described for "console=ttyS<n>"; if
976 unspecified, the h/w is not initialized.
979 Start an early, polled-mode console on a pl011 serial
980 port at the specified address. The pl011 serial port
981 must already be setup and configured. Options are not
985 Start an early, polled-mode console on an msm serial
986 port at the specified address. The serial port
987 must already be setup and configured. Options are not
991 Start an early, polled-mode console on an msm serial
992 dm port at the specified address. The serial port
993 must already be setup and configured. Options are not
996 smh Use ARM semihosting calls for early console.
1004 Use early console provided by serial driver available
1005 on Samsung SoCs, requires selecting proper type and
1006 a correct base address of the selected UART port. The
1007 serial port must already be setup and configured.
1008 Options are not yet supported.
1010 earlyprintk= [X86,SH,BLACKFIN,ARM,M68k]
1014 earlyprintk=serial[,ttySn[,baudrate]]
1015 earlyprintk=serial[,0x...[,baudrate]]
1016 earlyprintk=ttySn[,baudrate]
1017 earlyprintk=dbgp[debugController#]
1019 earlyprintk is useful when the kernel crashes before
1020 the normal console is initialized. It is not enabled by
1021 default because it has some cosmetic problems.
1023 Append ",keep" to not disable it when the real console
1026 Only one of vga, efi, serial, or usb debug port can
1029 Currently only ttyS0 and ttyS1 may be specified by
1030 name. Other I/O ports may be explicitly specified
1031 on some architectures (x86 and arm at least) by
1032 replacing ttySn with an I/O port address, like this:
1033 earlyprintk=serial,0x1008,115200
1034 You can find the port for a given device in
1035 /proc/tty/driver/serial:
1036 2: uart:ST16650V2 port:00001008 irq:18 ...
1038 Interaction with the standard serial driver is not
1041 The VGA and EFI output is eventually overwritten by
1044 The xen output can only be used by Xen PV guests.
1046 edac_report= [HW,EDAC] Control how to report EDAC event
1047 Format: {"on" | "off" | "force"}
1048 on: enable EDAC to report H/W event. May be overridden
1049 by other higher priority error reporting module.
1050 off: disable H/W event reporting through EDAC.
1051 force: enforce the use of EDAC to report H/W event.
1054 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1057 This is designed to be used in conjunction with
1058 the boot argument: earlyprintk=vga
1061 Format: {"off" | "on" | "skip[mbr]"}
1064 Format: { "old_map", "nochunk", "noruntime", "debug" }
1065 old_map [X86-64]: switch to the old ioremap-based EFI
1066 runtime services mapping. 32-bit still uses this one by
1068 nochunk: disable reading files in "chunks" in the EFI
1069 boot stub, as chunking can cause problems with some
1070 firmware implementations.
1071 noruntime : disable EFI runtime services support
1072 debug: enable misc debug output
1074 efi_no_storage_paranoia [EFI; X86]
1075 Using this parameter you can use more than 50% of
1076 your efi variable storage. Use this parameter only if
1077 you are really sure that your UEFI does sane gc and
1078 fulfills the spec otherwise your board may brick.
1080 eisa_irq_edge= [PARISC,HW]
1081 See header of drivers/parisc/eisa.c.
1084 See comment before function elanfreq_setup() in
1085 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1088 Format: {"cfq" | "deadline" | "noop"}
1089 See Documentation/block/cfq-iosched.txt and
1090 Documentation/block/deadline-iosched.txt for details.
1092 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1093 Specifies physical address of start of kernel core
1094 image elf header and optionally the size. Generally
1095 kexec loader will pass this option to capture kernel.
1096 See Documentation/kdump/kdump.txt for details.
1098 enable_mtrr_cleanup [X86]
1099 The kernel tries to adjust MTRR layout from continuous
1100 to discrete, to make X server driver able to add WB
1101 entry later. This parameter enables that.
1103 enable_timer_pin_1 [X86]
1104 Enable PIN 1 of APIC timer
1105 Can be useful to work around chipset bugs
1106 (in particular on some ATI chipsets).
1107 The kernel tries to set a reasonable default.
1109 enforcing [SELINUX] Set initial enforcing status.
1111 See security/selinux/Kconfig help text.
1112 0 -- permissive (log only, no denials).
1113 1 -- enforcing (deny and log).
1115 Value can be changed at runtime via /selinux/enforce.
1118 Disable Error Record Serialization Table (ERST)
1121 ether= [HW,NET] Ethernet cards parameters
1122 This option is obsoleted by the "netdev=" option, which
1123 has equivalent usage. See its documentation for details.
1127 Permit 'security.evm' to be updated regardless of
1128 current integrity status.
1132 fail_make_request=[KNL]
1133 General fault injection mechanism.
1134 Format: <interval>,<probability>,<space>,<times>
1135 See also Documentation/fault-injection/.
1138 See Documentation/blockdev/floppy.txt.
1140 force_pal_cache_flush
1141 [IA-64] Avoid check_sal_cache_flush which may hang on
1142 buggy SAL_CACHE_FLUSH implementations. Using this
1143 parameter will force ia64_sal_cache_flush to call
1144 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1147 Forcefully enable Physical Address Extension (PAE).
1148 Many Pentium M systems disable PAE but may have a
1149 functionally usable PAE implementation.
1150 Warning: use of this parameter will taint the kernel
1151 and may cause unknown problems.
1154 [FTRACE] will set and start the specified tracer
1155 as early as possible in order to facilitate early
1158 ftrace_dump_on_oops[=orig_cpu]
1159 [FTRACE] will dump the trace buffers on oops.
1160 If no parameter is passed, ftrace will dump
1161 buffers of all CPUs, but if you pass orig_cpu, it will
1162 dump only the buffer of the CPU that triggered the
1165 ftrace_filter=[function-list]
1166 [FTRACE] Limit the functions traced by the function
1167 tracer at boot up. function-list is a comma separated
1168 list of functions. This list can be changed at run
1169 time by the set_ftrace_filter file in the debugfs
1172 ftrace_notrace=[function-list]
1173 [FTRACE] Do not trace the functions specified in
1174 function-list. This list can be changed at run time
1175 by the set_ftrace_notrace file in the debugfs
1178 ftrace_graph_filter=[function-list]
1179 [FTRACE] Limit the top level callers functions traced
1180 by the function graph tracer at boot up.
1181 function-list is a comma separated list of functions
1182 that can be changed at run time by the
1183 set_graph_function file in the debugfs tracing directory.
1185 ftrace_graph_notrace=[function-list]
1186 [FTRACE] Do not trace from the functions specified in
1187 function-list. This list is a comma separated list of
1188 functions that can be changed at run time by the
1189 set_graph_notrace file in the debugfs tracing directory.
1192 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1193 support via parallel port (up to 5 devices per port)
1194 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1195 See also Documentation/input/joystick-parport.txt
1199 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1203 gcov_persist= [GCOV] When non-zero (default), profiling data for
1204 kernel modules is saved and remains accessible via
1205 debugfs, even when the module is unloaded/reloaded.
1206 When zero, profiling data is discarded and associated
1207 debugfs files are removed at module unload time.
1209 gpt [EFI] Forces disk with valid GPT signature but
1210 invalid Protective MBR to be treated as GPT. If the
1211 primary GPT is corrupted, it enables the backup/alternate
1212 GPT to be used instead.
1214 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1215 the "Enable 0" bit of the configuration register.
1218 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1219 the "Enable 0" bit of the configuration register.
1222 grcan.select= [HW] Select which physical interface to use.
1225 grcan.txsize= [HW] Sets the size of the tx buffer.
1226 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1228 grcan.rxsize= [HW] Sets the size of the rx buffer.
1229 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1232 hashdist= [KNL,NUMA] Large hashes allocated during boot
1233 are distributed across NUMA nodes. Defaults on
1234 for 64-bit NUMA, off otherwise.
1235 Format: 0 | 1 (for off | on)
1237 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1239 hd= [EIDE] (E)IDE hard drive subsystem geometry
1240 Format: <cyl>,<head>,<sect>
1243 Disable Hardware Error Source Table (HEST) support;
1244 corresponding firmware-first mode error processing
1245 logic will be disabled.
1247 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1248 size of <nn>. This works even on boxes that have no
1249 highmem otherwise. This also works to reduce highmem
1250 size on bigger boxes.
1252 highres= [KNL] Enable/disable high resolution timer mode.
1253 Valid parameters: "on", "off"
1257 See Documentation/isdn/README.HiSax.
1261 hpet= [X86-32,HPET] option to control HPET usage
1262 Format: { enable (default) | disable | force |
1264 disable: disable HPET and use PIT instead
1265 force: allow force enabled of undocumented chips (ICH4,
1267 verbose: show contents of HPET registers during setup
1269 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1270 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1272 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1273 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1274 On x86-64 and powerpc, this option can be specified
1275 multiple times interleaved with hugepages= to reserve
1276 huge pages of different sizes. Valid pages sizes on
1277 x86-64 are 2M (when the CPU supports "pse") and 1G
1278 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1280 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1281 terminal devices. Valid values: 0..8
1282 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1283 If specified, z/VM IUCV HVC accepts connections
1284 from listed z/VM user IDs only.
1286 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1287 hardware thread id mappings.
1288 Format: <cpu>:<hwthread>
1291 Do not unregister boot console at start. This is only
1292 useful for debugging when something happens in the window
1293 between unregistering the boot console and initializing
1296 i2c_bus= [HW] Override the default board specific I2C bus speed
1297 or register an additional I2C bus that is not
1298 registered from board initialization code.
1302 i8042.debug [HW] Toggle i8042 debug mode
1303 i8042.direct [HW] Put keyboard port into non-translated mode
1304 i8042.dumbkbd [HW] Pretend that controller can only read data from
1305 keyboard and cannot control its state
1306 (Don't attempt to blink the leds)
1307 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1308 i8042.nokbd [HW] Don't check/create keyboard port
1309 i8042.noloop [HW] Disable the AUX Loopback command while probing
1311 i8042.nomux [HW] Don't check presence of an active multiplexing
1313 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1315 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1316 i8042.reset [HW] Reset the controller during init and cleanup
1317 i8042.unlock [HW] Unlock (ignore) the keylock
1318 i8042.kbdreset [HW] Reset device connected to KBD port
1322 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1323 indicates that the driver is running on unsupported
1325 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1326 does not match list of supported models.
1328 [HW] Report power status in /proc/i8k
1329 (disabled by default)
1330 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1333 i915.invert_brightness=
1334 [DRM] Invert the sense of the variable that is used to
1335 set the brightness of the panel backlight. Normally a
1336 brightness value of 0 indicates backlight switched off,
1337 and the maximum of the brightness value sets the backlight
1338 to maximum brightness. If this parameter is set to 0
1339 (default) and the machine requires it, or this parameter
1340 is set to 1, a brightness value of 0 sets the backlight
1341 to maximum brightness, and the maximum of the brightness
1342 value switches the backlight off.
1343 -1 -- never invert brightness
1344 0 -- machine default
1345 1 -- force brightness inversion
1348 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1350 ide-core.nodma= [HW] (E)IDE subsystem
1351 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1352 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1353 .cdrom .chs .ignore_cable are additional options
1354 See Documentation/ide/ide.txt.
1356 ide-generic.probe-mask= [HW] (E)IDE subsystem
1358 Probe mask for legacy ISA IDE ports. Depending on
1359 platform up to 6 ports are supported, enabled by
1360 setting corresponding bits in the mask to 1. The
1361 default value is 0x0, which has a special meaning.
1362 On systems that have PCI, it triggers scanning the
1363 PCI bus for the first and the second port, which
1364 are then probed. On systems without PCI the value
1365 of 0x0 enables probing the two first ports as if it
1368 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1369 Claim all unknown PCI IDE storage controllers.
1372 Format: idle=poll, idle=halt, idle=nomwait
1373 Poll forces a polling idle loop that can slightly
1374 improve the performance of waking up a idle CPU, but
1375 will use a lot of power and make the system run hot.
1377 idle=halt: Halt is forced to be used for CPU idle.
1378 In such case C2/C3 won't be used again.
1379 idle=nomwait: Disable mwait for CPU C-states
1381 ignore_loglevel [KNL]
1382 Ignore loglevel setting - this will print /all/
1383 kernel messages to the console. Useful for debugging.
1384 We also add it as printk module parameter, so users
1385 could change it dynamically, usually by
1386 /sys/module/printk/parameters/ignore_loglevel.
1388 ihash_entries= [KNL]
1389 Set number of hash buckets for inode cache.
1391 ima_appraise= [IMA] appraise integrity measurements
1392 Format: { "off" | "enforce" | "fix" | "log" }
1395 ima_appraise_tcb [IMA]
1396 The builtin appraise policy appraises all files
1400 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1404 The list of supported hash algorithms is defined
1405 in crypto/hash_info.h.
1408 Load a policy which meets the needs of the Trusted
1409 Computing Base. This means IMA will measure all
1410 programs exec'd, files mmap'd for exec, and all files
1411 opened for read by uid=0.
1414 Select one of defined IMA measurements template formats.
1415 Formats: { "ima" | "ima-ng" }
1419 [IMA] Define a custom template format.
1420 Format: { "field1|...|fieldN" }
1422 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1423 Format: <min_file_size>
1424 Set the minimal file size for using asynchronous hash.
1425 If left unspecified, ahash usage is disabled.
1427 ahash performance varies for different data sizes on
1428 different crypto accelerators. This option can be used
1429 to achieve the best performance for a particular HW.
1431 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1433 Set hashing buffer size. Default: 4k.
1435 ahash performance varies for different chunk sizes on
1436 different crypto accelerators. This option can be used
1437 to achieve best performance for particular HW.
1441 Run specified binary instead of /sbin/init as init
1444 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1445 for working out where the kernel is dying during
1448 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1449 initcall functions. Useful for debugging built-in
1450 modules and initcalls.
1452 initrd= [BOOT] Specify the location of the initial ramdisk
1454 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1457 int_pln_enable [x86] Enable power limit notification interrupt
1459 integrity_audit=[IMA]
1460 Format: { "0" | "1" }
1461 0 -- basic integrity auditing messages. (Default)
1462 1 -- additional integrity auditing messages.
1464 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1466 Enable intel iommu driver.
1468 Disable intel iommu driver.
1469 igfx_off [Default Off]
1470 By default, gfx is mapped as normal device. If a gfx
1471 device has a dedicated DMAR unit, the DMAR unit is
1472 bypassed by not enabling DMAR with this option. In
1473 this case, gfx device will use physical address for
1476 With this option iommu will not optimize to look
1477 for io virtual address below 32-bit forcing dual
1478 address cycle on pci bus for cards supporting greater
1479 than 32-bit addressing. The default is to look
1480 for translation below 32-bit and if not available
1481 then look in the higher range.
1482 strict [Default Off]
1483 With this option on every unmap_single operation will
1484 result in a hardware IOTLB flush operation as opposed
1485 to batching them for performance.
1486 sp_off [Default Off]
1487 By default, super page will be supported if Intel IOMMU
1488 has the capability. With this option, super page will
1491 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1492 0 disables intel_idle and fall back on acpi_idle.
1493 1 to 6 specify maximum depth of C-state.
1497 Do not enable intel_pstate as the default
1498 scaling driver for the supported processors
1500 Enable intel_pstate on systems that prohibit it by default
1501 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1502 instead of acpi-cpufreq may disable platform features, such
1503 as thermal controls and power capping, that rely on ACPI
1504 P-States information being indicated to OSPM and therefore
1505 should be used with caution. This option does not work with
1506 processors that aren't supported by the intel_pstate driver
1507 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1509 Do not enable hardware P state control (HWP)
1512 Only load intel_pstate on systems which support
1513 hardware P state control (HWP) if available.
1515 intremap= [X86-64, Intel-IOMMU]
1516 on enable Interrupt Remapping (default)
1517 off disable Interrupt Remapping
1518 nosid disable Source ID checking
1520 BIOS x2APIC opt-out request will be ignored
1522 iomem= Disable strict checking of access to MMIO memory
1523 strict regions from userspace.
1538 nobypass [PPC/POWERNV]
1539 Disable IOMMU bypass, using IOMMU for PCI devices.
1542 io7= [HW] IO7 for Marvel based alpha systems
1543 See comment before marvel_specify_io7 in
1544 arch/alpha/kernel/core_marvel.c.
1546 io_delay= [X86] I/O delay method
1548 Standard port 0x80 based delay
1550 Alternate port 0xed based delay (needed on some systems)
1552 Simple two microseconds delay
1557 See Documentation/filesystems/nfs/nfsroot.txt.
1560 When an interrupt is not handled search all handlers
1561 for it. Intended to get systems with badly broken
1565 When an interrupt is not handled search all handlers
1566 for it. Also check all handlers each timer
1567 interrupt. Intended to get systems with badly broken
1571 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1573 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1575 <cpu number>,...,<cpu number>
1577 <cpu number>-<cpu number>
1578 (must be a positive range in ascending order)
1580 <cpu number>,...,<cpu number>-<cpu number>
1582 This option can be used to specify one or more CPUs
1583 to isolate from the general SMP balancing and scheduling
1584 algorithms. You can move a process onto or off an
1585 "isolated" CPU via the CPU affinity syscalls or cpuset.
1586 <cpu number> begins at 0 and the maximum value is
1587 "number of CPUs in system - 1".
1589 This option is the preferred way to isolate CPUs. The
1590 alternative -- manually setting the CPU mask of all
1591 tasks in the system -- can cause problems and
1592 suboptimal load balancer performance.
1596 ivrs_ioapic [HW,X86_64]
1597 Provide an override to the IOAPIC-ID<->DEVICE-ID
1598 mapping provided in the IVRS ACPI table. For
1599 example, to map IOAPIC-ID decimal 10 to
1600 PCI device 00:14.0 write the parameter as:
1601 ivrs_ioapic[10]=00:14.0
1603 ivrs_hpet [HW,X86_64]
1604 Provide an override to the HPET-ID<->DEVICE-ID
1605 mapping provided in the IVRS ACPI table. For
1606 example, to map HPET-ID decimal 0 to
1607 PCI device 00:14.0 write the parameter as:
1608 ivrs_hpet[0]=00:14.0
1610 js= [HW,JOY] Analog joystick
1611 See Documentation/input/joystick.txt.
1614 Enable/disable kernel and module base offset ASLR
1615 (Address Space Layout Randomization) if built into
1616 the kernel. When CONFIG_HIBERNATION is selected,
1617 kASLR is disabled by default. When kASLR is enabled,
1618 hibernation will be disabled.
1622 kernelcore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1623 specifies the amount of memory usable by the kernel
1624 for non-movable allocations. The requested amount is
1625 spread evenly throughout all nodes in the system. The
1626 remaining memory in each node is used for Movable
1627 pages. In the event, a node is too small to have both
1628 kernelcore and Movable pages, kernelcore pages will
1629 take priority and other nodes will have a larger number
1630 of Movable pages. The Movable zone is used for the
1631 allocation of pages that may be reclaimed or moved
1632 by the page migration subsystem. This means that
1633 HugeTLB pages may not be allocated from this zone.
1634 Note that allocations like PTEs-from-HighMem still
1635 use the HighMem zone if it exists, and the Normal
1636 zone if it does not.
1638 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1639 Format: <Controller#>[,poll interval]
1640 The controller # is the number of the ehci usb debug
1641 port as it is probed via PCI. The poll interval is
1642 optional and is the number seconds in between
1643 each poll cycle to the debug port in case you need
1644 the functionality for interrupting the kernel with
1645 gdb or control-c on the dbgp connection. When
1646 not using this parameter you use sysrq-g to break into
1647 the kernel debugger.
1649 kgdboc= [KGDB,HW] kgdb over consoles.
1650 Requires a tty driver that supports console polling,
1651 or a supported polling keyboard driver (non-usb).
1652 Serial only format: <serial_device>[,baud]
1653 keyboard only format: kbd
1654 keyboard and serial format: kbd,<serial_device>[,baud]
1655 Optional Kernel mode setting:
1656 kms, kbd format: kms,kbd
1657 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1659 kgdbwait [KGDB] Stop kernel execution and enter the
1660 kernel debugger at the earliest opportunity.
1662 kmac= [MIPS] korina ethernet MAC address.
1663 Configure the RouterBoard 532 series on-chip
1664 Ethernet adapter MAC address.
1666 kmemleak= [KNL] Boot-time kmemleak enable/disable
1667 Valid arguments: on, off
1669 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1672 kmemcheck= [X86] Boot-time kmemcheck enable/disable/one-shot mode
1673 Valid arguments: 0, 1, 2
1674 kmemcheck=0 (disabled)
1675 kmemcheck=1 (enabled)
1676 kmemcheck=2 (one-shot mode)
1677 Default: 2 (one-shot mode)
1679 kstack=N [X86] Print N words from the kernel stack
1682 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1683 Default is 0 (don't ignore, but inject #GP)
1685 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1689 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1690 Default is 1 (enabled)
1692 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1694 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1696 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1697 (virtualized MMU) support on capable Intel chips.
1698 Default is 1 (enabled)
1700 kvm-intel.emulate_invalid_guest_state=
1701 [KVM,Intel] Enable emulation of invalid guest states
1702 Default is 0 (disabled)
1704 kvm-intel.flexpriority=
1705 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1706 Default is 1 (enabled)
1709 [KVM,Intel] Enable VMX nesting (nVMX).
1710 Default is 0 (disabled)
1712 kvm-intel.unrestricted_guest=
1713 [KVM,Intel] Disable unrestricted guest feature
1714 (virtualized real and unpaged mode) on capable
1715 Intel chips. Default is 1 (enabled)
1717 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1718 feature (tagged TLBs) on capable Intel chips.
1719 Default is 1 (enabled)
1725 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1728 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1729 value for LAPIC timer one-shot implementation. Default
1730 back to the programmable timer unit in the LAPIC.
1732 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1735 libata.dma= [LIBATA] DMA control
1736 libata.dma=0 Disable all PATA and SATA DMA
1737 libata.dma=1 PATA and SATA Disk DMA only
1738 libata.dma=2 ATAPI (CDROM) DMA only
1739 libata.dma=4 Compact Flash DMA only
1740 Combinations also work, so libata.dma=3 enables DMA
1741 for disks and CDROMs, but not CFs.
1743 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1744 libata.ignore_hpa=0 keep BIOS limits (default)
1745 libata.ignore_hpa=1 ignore limits, using full disk
1747 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
1751 libata.force= [LIBATA] Force configurations. The format is comma
1752 separated list of "[ID:]VAL" where ID is
1753 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
1754 matching port, link or device. Basically, it matches
1755 the ATA ID string printed on console by libata. If
1756 the whole ID part is omitted, the last PORT and DEVICE
1757 values are used. If ID hasn't been specified yet, the
1758 configuration applies to all ports, links and devices.
1760 If only DEVICE is omitted, the parameter applies to
1761 the port and all links and devices behind it. DEVICE
1762 number of 0 either selects the first device or the
1763 first fan-out link behind PMP device. It does not
1764 select the host link. DEVICE number of 15 selects the
1765 host link and device attached to it.
1767 The VAL specifies the configuration to force. As long
1768 as there's no ambiguity shortcut notation is allowed.
1769 For example, both 1.5 and 1.5G would work for 1.5Gbps.
1770 The following configurations can be forced.
1772 * Cable type: 40c, 80c, short40c, unk, ign or sata.
1773 Any ID with matching PORT is used.
1775 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
1777 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
1778 udma[/][16,25,33,44,66,100,133] notation is also
1781 * [no]ncq: Turn on or off NCQ.
1783 * nohrst, nosrst, norst: suppress hard, soft
1786 * rstonce: only attempt one reset during
1787 hot-unplug link recovery
1789 * dump_id: dump IDENTIFY data.
1791 * atapi_dmadir: Enable ATAPI DMADIR bridge support
1793 * disable: Disable this device.
1795 If there are multiple matching configurations changing
1796 the same attribute, the last one is used.
1798 memblock=debug [KNL] Enable memblock debug messages.
1800 load_ramdisk= [RAM] List of ramdisks to load from floppy
1801 See Documentation/blockdev/ramdisk.txt.
1803 lockd.nlm_grace_period=P [NFS] Assign grace period.
1806 lockd.nlm_tcpport=N [NFS] Assign TCP port.
1809 lockd.nlm_timeout=T [NFS] Assign timeout value.
1812 lockd.nlm_udpport=M [NFS] Assign UDP port.
1815 locktorture.nreaders_stress= [KNL]
1816 Set the number of locking read-acquisition kthreads.
1817 Defaults to being automatically set based on the
1818 number of online CPUs.
1820 locktorture.nwriters_stress= [KNL]
1821 Set the number of locking write-acquisition kthreads.
1823 locktorture.onoff_holdoff= [KNL]
1824 Set time (s) after boot for CPU-hotplug testing.
1826 locktorture.onoff_interval= [KNL]
1827 Set time (s) between CPU-hotplug operations, or
1828 zero to disable CPU-hotplug testing.
1830 locktorture.shuffle_interval= [KNL]
1831 Set task-shuffle interval (jiffies). Shuffling
1832 tasks allows some CPUs to go into dyntick-idle
1833 mode during the locktorture test.
1835 locktorture.shutdown_secs= [KNL]
1836 Set time (s) after boot system shutdown. This
1837 is useful for hands-off automated testing.
1839 locktorture.stat_interval= [KNL]
1840 Time (s) between statistics printk()s.
1842 locktorture.stutter= [KNL]
1843 Time (s) to stutter testing, for example,
1844 specifying five seconds causes the test to run for
1845 five seconds, wait for five seconds, and so on.
1846 This tests the locking primitive's ability to
1847 transition abruptly to and from idle.
1849 locktorture.torture_runnable= [BOOT]
1850 Start locktorture running at boot time.
1852 locktorture.torture_type= [KNL]
1853 Specify the locking implementation to test.
1855 locktorture.verbose= [KNL]
1856 Enable additional printk() statements.
1858 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
1861 loglevel= All Kernel Messages with a loglevel smaller than the
1862 console loglevel will be printed to the console. It can
1863 also be changed with klogd or other programs. The
1864 loglevels are defined as follows:
1866 0 (KERN_EMERG) system is unusable
1867 1 (KERN_ALERT) action must be taken immediately
1868 2 (KERN_CRIT) critical conditions
1869 3 (KERN_ERR) error conditions
1870 4 (KERN_WARNING) warning conditions
1871 5 (KERN_NOTICE) normal but significant condition
1872 6 (KERN_INFO) informational
1873 7 (KERN_DEBUG) debug-level messages
1875 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
1876 in bytes. n must be a power of two and greater
1877 than the minimal size. The minimal size is defined
1878 by LOG_BUF_SHIFT kernel config parameter. There is
1879 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
1880 that allows to increase the default size depending on
1881 the number of CPUs. See init/Kconfig for more details.
1883 logo.nologo [FB] Disables display of the built-in Linux logo.
1884 This may be used to provide more screen space for
1885 kernel log messages and is useful when debugging
1886 kernel boot problems.
1888 lp=0 [LP] Specify parallel ports to use, e.g,
1889 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
1890 lp=reset first parallel port). 'lp=0' disables the
1891 lp=auto printer driver. 'lp=reset' (which can be
1892 specified in addition to the ports) causes
1893 attached printers to be reset. Using
1894 lp=port1,port2,... specifies the parallel ports
1895 to associate lp devices with, starting with
1896 lp0. A port specification may be 'none' to skip
1897 that lp device, or a parport name such as
1898 'parport0'. Specifying 'lp=auto' instead of a
1899 port specification list means that device IDs
1900 from each port should be examined, to see if
1901 an IEEE 1284-compliant printer is attached; if
1902 so, the driver will manage that printer.
1903 See also header of drivers/char/lp.c.
1906 Sets loops_per_jiffy to given constant, thus avoiding
1907 time-consuming boot-time autodetection (up to 250 ms per
1908 CPU). 0 enables autodetection (default). To determine
1909 the correct value for your kernel, boot with normal
1910 autodetection and see what value is printed. Note that
1911 on SMP systems the preset will be applied to all CPUs,
1912 which is likely to cause problems if your CPUs need
1913 significantly divergent settings. An incorrect value
1914 will cause delays in the kernel to be wrong, leading to
1915 unpredictable I/O errors and other breakage. Although
1916 unlikely, in the extreme case this might damage your
1920 Format: <io>,<irq>,<dma>
1922 machvec= [IA-64] Force the use of a particular machine-vector
1923 (machvec) in a generic kernel.
1924 Example: machvec=hpzx1_swiotlb
1926 machtype= [Loongson] Share the same kernel image file between different
1928 Example: machtype=lemote-yeeloong-2f-7inch
1930 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
1931 than or equal to this physical address is ignored.
1933 maxcpus= [SMP] Maximum number of processors that an SMP kernel
1934 should make use of. maxcpus=n : n >= 0 limits the
1935 kernel to using 'n' processors. n=0 is a special case,
1936 it is equivalent to "nosmp", which also disables
1939 max_loop= [LOOP] The number of loop block devices that get
1940 (loop.max_loop) unconditionally pre-created at init time. The default
1941 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
1942 of statically allocating a predefined number, loop
1943 devices can be requested on-demand with the
1944 /dev/loop-control interface.
1946 mce [X86-32] Machine Check Exception
1948 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
1950 md= [HW] RAID subsystems devices and level
1951 See Documentation/md.txt.
1954 Format: <first>,<last>
1955 Specifies range of consoles to be captured by the MDA.
1957 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
1958 Amount of memory to be used when the kernel is not able
1959 to see the whole system memory or for test.
1960 [X86] Work as limiting max address. Use together
1961 with memmap= to avoid physical address space collisions.
1962 Without memmap= PCI devices could be placed at addresses
1963 belonging to unused RAM.
1965 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
1969 [KNL,SH] Allow user to override the default size for
1970 per-device physically contiguous DMA buffers.
1972 memmap=exactmap [KNL,X86] Enable setting of an exact
1973 E820 memory map, as specified by the user.
1974 Such memmap=exactmap lines can be constructed based on
1975 BIOS output or other requirements. See the memmap=nn@ss
1978 memmap=nn[KMG]@ss[KMG]
1979 [KNL] Force usage of a specific region of memory.
1980 Region of memory to be used is from ss to ss+nn.
1982 memmap=nn[KMG]#ss[KMG]
1983 [KNL,ACPI] Mark specific memory as ACPI data.
1984 Region of memory to be marked is from ss to ss+nn.
1986 memmap=nn[KMG]$ss[KMG]
1987 [KNL,ACPI] Mark specific memory as reserved.
1988 Region of memory to be reserved is from ss to ss+nn.
1989 Example: Exclude memory from 0x18690000-0x1869ffff
1990 memmap=64K$0x18690000
1992 memmap=0x10000$0x18690000
1994 memmap=nn[KMG]!ss[KMG]
1995 [KNL,X86] Mark specific memory as protected.
1996 Region of memory to be used, from ss to ss+nn.
1997 The memory region may be marked as e820 type 12 (0xc)
1998 and is NVDIMM or ADR memory.
2000 memory_corruption_check=0/1 [X86]
2001 Some BIOSes seem to corrupt the first 64k of
2002 memory when doing things like suspend/resume.
2003 Setting this option will scan the memory
2004 looking for corruption. Enabling this will
2005 both detect corruption and prevent the kernel
2006 from using the memory being corrupted.
2007 However, its intended as a diagnostic tool; if
2008 repeatable BIOS-originated corruption always
2009 affects the same memory, you can use memmap=
2010 to prevent the kernel from using that memory.
2012 memory_corruption_check_size=size [X86]
2013 By default it checks for corruption in the low
2014 64k, making this memory unavailable for normal
2015 use. Use this parameter to scan for
2016 corruption in more or less memory.
2018 memory_corruption_check_period=seconds [X86]
2019 By default it checks for corruption every 60
2020 seconds. Use this parameter to check at some
2021 other rate. 0 disables periodic checking.
2023 memtest= [KNL,X86,ARM] Enable memtest
2025 default : 0 <disable>
2026 Specifies the number of memtest passes to be
2027 performed. Each pass selects another test
2028 pattern from a given set of patterns. Memtest
2029 fills the memory with this pattern, validates
2030 memory contents and reserves bad memory
2031 regions that are detected.
2033 meye.*= [HW] Set MotionEye Camera parameters
2034 See Documentation/video4linux/meye.txt.
2036 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2037 Multi-Function General Purpose Timers on AMD Geode
2040 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2041 the BIOS has incorrectly applied a workaround. TinyBIOS
2042 version 0.98 is known to be affected, 0.99 fixes the
2043 problem by letting the user disable the workaround.
2047 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2048 physical address is ignored.
2050 mini2440= [ARM,HW,KNL]
2051 Format:[0..2][b][c][t]
2053 MINI2440 configuration specification:
2054 0 - The attached screen is the 3.5" TFT
2055 1 - The attached screen is the 7" TFT
2056 2 - The VGA Shield is attached (1024x768)
2057 Leaving out the screen size parameter will not load
2058 the TFT driver, and the framebuffer will be left
2060 b - Enable backlight. The TFT backlight pin will be
2061 linked to the kernel VESA blanking code and a GPIO
2062 LED. This parameter is not necessary when using the
2064 c - Enable the s3c camera interface.
2065 t - Reserved for enabling touchscreen support. The
2066 touchscreen support is not enabled in the mainstream
2067 kernel as of 2.6.30, a preliminary port can be found
2068 in the "bleeding edge" mini2440 support kernel at
2069 http://repo.or.cz/w/linux-2.6/mini2440.git
2072 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2073 parameter allows control of the logging verbosity for
2074 the additional memory initialisation checks. A value
2075 of 0 disables mminit logging and a level of 4 will
2076 log everything. Information is printed at KERN_DEBUG
2077 so loglevel=8 may also need to be specified.
2080 [KNL] When CONFIG_MODULE_SIG is set, this means that
2081 modules without (valid) signatures will fail to load.
2082 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2083 is always true, so this option does nothing.
2086 [MOUSE] Maximum time between finger touching and
2087 leaving touchpad surface for touch to be considered
2088 a tap and be reported as a left button click (for
2089 touchpads working in absolute mode only).
2091 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2092 reporting absolute coordinates, such as tablets
2093 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2094 reporting absolute coordinates, such as tablets
2096 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
2097 is similar to kernelcore except it specifies the
2098 amount of memory used for migratable allocations.
2099 If both kernelcore and movablecore is specified,
2100 then kernelcore will be at *least* the specified
2101 value but may be more. If movablecore on its own
2102 is specified, the administrator must be careful
2103 that the amount of memory usable for all allocations
2106 movable_node [KNL,X86] Boot-time switch to enable the effects
2107 of CONFIG_MOVABLE_NODE=y. See mm/Kconfig for details.
2109 MTD_Partition= [MTD]
2110 Format: <name>,<region-number>,<size>,<offset>
2112 MTD_Region= [MTD] Format:
2113 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2116 See drivers/mtd/cmdlinepart.c.
2118 multitce=off [PPC] This parameter disables the use of the pSeries
2119 firmware feature for updating multiple TCE entries
2122 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2124 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2126 boundary - index of last SLC block on Flex-OneNAND.
2127 The remaining blocks are configured as MLC blocks.
2128 lock - Configure if Flex-OneNAND boundary should be locked.
2129 Once locked, the boundary cannot be changed.
2130 1 indicates lock status, 0 indicates unlock status.
2133 ARM/S3C2412 JIVE boot control
2135 See arch/arm/mach-s3c2412/mach-jive.c
2137 mtouchusb.raw_coordinates=
2138 [HW] Make the MicroTouch USB driver use raw coordinates
2139 ('y', default) or cooked coordinates ('n')
2141 mtrr_chunk_size=nn[KMG] [X86]
2142 used for mtrr cleanup. It is largest continuous chunk
2143 that could hold holes aka. UC entries.
2145 mtrr_gran_size=nn[KMG] [X86]
2146 Used for mtrr cleanup. It is granularity of mtrr block.
2148 Large value could prevent small alignment from
2151 mtrr_spare_reg_nr=n [X86]
2153 Range: 0,7 : spare reg number
2155 Used for mtrr cleanup. It is spare mtrr entries number.
2156 Set to 2 or more if your graphical card needs more.
2158 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2160 netdev= [NET] Network devices parameters
2161 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2162 Note that mem_start is often overloaded to mean
2163 something different and driver-specific.
2164 This usage is only documented in each driver source
2168 [NETFILTER] Enable connection tracking flow accounting
2169 0 to disable accounting
2170 1 to enable accounting
2173 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2174 See Documentation/filesystems/nfs/nfsroot.txt.
2176 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2177 See Documentation/filesystems/nfs/nfsroot.txt.
2179 nfsrootdebug [NFS] enable nfsroot debugging messages.
2180 See Documentation/filesystems/nfs/nfsroot.txt.
2182 nfs.callback_tcpport=
2183 [NFS] set the TCP port on which the NFSv4 callback
2184 channel should listen.
2187 [NFS] sets the pathname to the program which is used
2188 to update the NFS client cache entries.
2190 nfs.cache_getent_timeout=
2191 [NFS] sets the timeout after which an attempt to
2192 update a cache entry is deemed to have failed.
2194 nfs.idmap_cache_timeout=
2195 [NFS] set the maximum lifetime for idmapper cache
2199 [NFS] enable 64-bit inode numbers.
2200 If zero, the NFS client will fake up a 32-bit inode
2201 number for the readdir() and stat() syscalls instead
2202 of returning the full 64-bit number.
2203 The default is to return 64-bit inode numbers.
2205 nfs.max_session_slots=
2206 [NFSv4.1] Sets the maximum number of session slots
2207 the client will attempt to negotiate with the server.
2208 This limits the number of simultaneous RPC requests
2209 that the client can send to the NFSv4.1 server.
2210 Note that there is little point in setting this
2211 value higher than the max_tcp_slot_table_limit.
2213 nfs.nfs4_disable_idmapping=
2214 [NFSv4] When set to the default of '1', this option
2215 ensures that both the RPC level authentication
2216 scheme and the NFS level operations agree to use
2217 numeric uids/gids if the mount is using the
2218 'sec=sys' security flavour. In effect it is
2219 disabling idmapping, which can make migration from
2220 legacy NFSv2/v3 systems to NFSv4 easier.
2221 Servers that do not support this mode of operation
2222 will be autodetected by the client, and it will fall
2223 back to using the idmapper.
2224 To turn off this behaviour, set the value to '0'.
2226 [NFS4] Specify an additional fixed unique ident-
2227 ification string that NFSv4 clients can insert into
2228 their nfs_client_id4 string. This is typically a
2229 UUID that is generated at system install time.
2231 nfs.send_implementation_id =
2232 [NFSv4.1] Send client implementation identification
2233 information in exchange_id requests.
2234 If zero, no implementation identification information
2236 The default is to send the implementation identification
2239 nfs.recover_lost_locks =
2240 [NFSv4] Attempt to recover locks that were lost due
2241 to a lease timeout on the server. Please note that
2242 doing this risks data corruption, since there are
2243 no guarantees that the file will remain unchanged
2244 after the locks are lost.
2245 If you want to enable the kernel legacy behaviour of
2246 attempting to recover these locks, then set this
2248 The default parameter value of '0' causes the kernel
2249 not to attempt recovery of lost locks.
2251 nfsd.nfs4_disable_idmapping=
2252 [NFSv4] When set to the default of '1', the NFSv4
2253 server will return only numeric uids and gids to
2254 clients using auth_sys, and will accept numeric uids
2255 and gids from such clients. This is intended to ease
2256 migration from NFSv2/v3.
2258 objlayoutdriver.osd_login_prog=
2259 [NFS] [OBJLAYOUT] sets the pathname to the program which
2260 is used to automatically discover and login into new
2261 osd-targets. Please see:
2262 Documentation/filesystems/pnfs.txt for more explanations
2264 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
2265 when a NMI is triggered.
2266 Format: [state][,regs][,debounce][,die]
2268 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2269 Format: [panic,][nopanic,][num]
2271 0 - turn nmi_watchdog off
2272 1 - turn nmi_watchdog on
2273 When panic is specified, panic when an NMI watchdog
2274 timeout occurs (or 'nopanic' to override the opposite
2276 This is useful when you use a panic=... timeout and
2277 need the box quickly up again.
2279 netpoll.carrier_timeout=
2280 [NET] Specifies amount of time (in seconds) that
2281 netpoll should wait for a carrier. By default netpoll
2284 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2285 emulation library even if a 387 maths coprocessor
2289 [HW] Never suspend the console
2290 Disable suspending of consoles during suspend and
2291 hibernate operations. Once disabled, debugging
2292 messages can reach various consoles while the rest
2293 of the system is being put to sleep (ie, while
2294 debugging driver suspend/resume hooks). This may
2295 not work reliably with all consoles, but is known
2296 to work with serial and VGA consoles.
2297 To facilitate more flexible debugging, we also add
2298 console_suspend, a printk module parameter to control
2299 it. Users could use console_suspend (usually
2300 /sys/module/printk/parameters/console_suspend) to
2301 turn on/off it dynamically.
2303 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2304 caches in the slab allocator. Saves per-node memory,
2305 but will impact performance.
2309 noapic [SMP,APIC] Tells the kernel to not make use of any
2310 IOAPICs that may be present in the system.
2312 noautogroup Disable scheduler automatic task group creation.
2314 nobats [PPC] Do not use BATs for mapping kernel lowmem
2315 on "Classic" PPC cores.
2319 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2321 nodelayacct [KNL] Disable per-task delay accounting
2323 nodisconnect [HW,SCSI,M68K] Disables SCSI disconnects.
2325 nodsp [SH] Disable hardware DSP at boot time.
2327 noefi Disable EFI runtime services support.
2332 On X86-32 available only on PAE configured kernels.
2333 noexec=on: enable non-executable mappings (default)
2334 noexec=off: disable non-executable mappings
2337 Disable SMAP (Supervisor Mode Access Prevention)
2338 even if it is supported by processor.
2341 Disable SMEP (Supervisor Mode Execution Prevention)
2342 even if it is supported by processor.
2345 This affects only 32-bit executables.
2346 noexec32=on: enable non-executable mappings (default)
2347 read doesn't imply executable mappings
2348 noexec32=off: disable non-executable mappings
2349 read implies executable mappings
2351 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2353 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2354 register save and restore. The kernel will only save
2355 legacy floating-point registers on task switch.
2357 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2359 noxsave [BUGS=X86] Disables x86 extended register state save
2360 and restore using xsave. The kernel will fallback to
2361 enabling legacy floating-point and sse state.
2363 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2364 register states. The kernel will fall back to use
2365 xsave to save the states. By using this parameter,
2366 performance of saving the states is degraded because
2367 xsave doesn't support modified optimization while
2368 xsaveopt supports it on xsaveopt enabled systems.
2370 noxsaves [X86] Disables xsaves and xrstors used in saving and
2371 restoring x86 extended register state in compacted
2372 form of xsave area. The kernel will fall back to use
2373 xsaveopt and xrstor to save and restore the states
2374 in standard form of xsave area. By using this
2375 parameter, xsave area per process might occupy more
2376 memory on xsaves enabled systems.
2378 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2379 wfi(ARM) instruction doesn't work correctly and not to
2380 use it. This is also useful when using JTAG debugger.
2382 no_file_caps Tells the kernel not to honor file capabilities. The
2383 only way then for a file to be executed with privilege
2384 is to be setuid root or executed by root.
2386 nohalt [IA-64] Tells the kernel not to use the power saving
2387 function PAL_HALT_LIGHT when idle. This increases
2388 power-consumption. On the positive side, it reduces
2389 interrupt wake-up latency, which may improve performance
2390 in certain environments such as networked servers or
2393 nohibernate [HIBERNATION] Disable hibernation and resume.
2395 nohz= [KNL] Boottime enable/disable dynamic ticks
2396 Valid arguments: on, off
2399 nohz_full= [KNL,BOOT]
2400 In kernels built with CONFIG_NO_HZ_FULL=y, set
2401 the specified list of CPUs whose tick will be stopped
2402 whenever possible. The boot CPU will be forced outside
2403 the range to maintain the timekeeping.
2404 The CPUs in this range must also be included in the
2407 noiotrap [SH] Disables trapped I/O port accesses.
2409 noirqdebug [X86-32] Disables the code which attempts to detect and
2410 disable unhandled interrupt sources.
2412 no_timer_check [X86,APIC] Disables the code which tests for
2413 broken timer IRQ sources.
2415 noisapnp [ISAPNP] Disables ISA PnP code.
2417 noinitrd [RAM] Tells the kernel not to load any configured
2420 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2422 [Deprecated - use intremap=off]
2426 nojitter [IA-64] Disables jitter checking for ITC timers.
2428 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2430 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2433 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2434 steal time is computed, but won't influence scheduler
2437 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2439 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2441 noltlbs [PPC] Do not use large page/tlb entries for kernel
2442 lowmem mapping on PPC40x.
2444 nomca [IA-64] Disable machine check abort handling
2446 nomce [X86-32] Machine Check Exception
2448 nomfgpt [X86-32] Disable Multi-Function General Purpose
2449 Timer usage (for AMD Geode machines).
2451 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2452 shutdown the other cpus. Instead use the REBOOT_VECTOR
2455 nomodule Disable module load
2457 nopat [X86] Disable PAT (page attribute table extension of
2458 pagetables) support.
2460 norandmaps Don't use address space randomization. Equivalent to
2461 echo 0 > /proc/sys/kernel/randomize_va_space
2463 noreplace-paravirt [X86,IA-64,PV_OPS] Don't patch paravirt_ops
2465 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2466 with UP alternatives
2468 nordrand [X86] Disable kernel use of the RDRAND and
2469 RDSEED instructions even if they are supported
2470 by the processor. RDRAND and RDSEED are still
2471 available to user space applications.
2473 noresume [SWSUSP] Disables resume and restores original swap
2476 no-scroll [VGA] Disables scrollback.
2477 This is required for the Braillex ib80-piezo Braille
2478 reader made by F.H. Papenmeier (Germany).
2482 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2484 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2485 and disable the IO APIC. legacy for "maxcpus=0".
2487 nosoftlockup [KNL] Disable the soft-lockup detector.
2489 nosync [HW,M68K] Disables sync negotiation for all devices.
2491 notsc [BUGS=X86-32] Disable Time Stamp Counter
2493 nousb [USB] Disable the USB subsystem
2495 nowatchdog [KNL] Disable both lockup detectors, i.e.
2496 soft-lockup and NMI watchdog (hard-lockup).
2500 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2502 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2503 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2504 Some features depend on CPU0. Known dependencies are:
2505 1. Resume from suspend/hibernate depends on CPU0.
2506 Suspend/hibernate will fail if CPU0 is offline and you
2507 need to online CPU0 before suspend/hibernate.
2508 2. PIC interrupts also depend on CPU0. CPU0 can't be
2509 removed if a PIC interrupt is detected.
2510 It's said poweroff/reboot may depend on CPU0 on some
2511 machines although I haven't seen such issues so far
2512 after CPU0 is offline on a few tested machines.
2513 If the dependencies are under your control, you can
2514 turn on cpu0_hotplug.
2516 nptcg= [IA-64] Override max number of concurrent global TLB
2517 purges which is reported from either PAL_VM_SUMMARY or
2520 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2521 could support. nr_cpus=n : n >= 1 limits the kernel to
2522 supporting 'n' processors. Later in runtime you can not
2523 use hotplug cpu feature to put more cpu back to online.
2524 just like you compile the kernel NR_CPUS=n
2526 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2528 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2529 Allowed values are enable and disable
2531 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2532 one of ['zone', 'node', 'default'] can be specified
2533 This can be set from sysctl after boot.
2534 See Documentation/sysctl/vm.txt for details.
2536 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2537 See Documentation/debugging-via-ohci1394.txt for more
2540 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2541 Rather than timing out after 20 ms if an EC
2542 command is not properly ACKed, override the length
2543 of the timeout. We have interrupts disabled while
2544 waiting for the ACK, so if this is set too high
2545 interrupts *may* be lost!
2547 omap_mux= [OMAP] Override bootloader pin multiplexing.
2548 Format: <mux_mode0.mode_name=value>...
2549 For example, to override I2C bus2:
2550 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2552 oprofile.timer= [HW]
2553 Use timer interrupt instead of performance counters
2555 oprofile.cpu_type= Force an oprofile cpu type
2556 This might be useful if you have an older oprofile
2557 userland or if you want common events.
2558 Format: { arch_perfmon }
2559 arch_perfmon: [X86] Force use of architectural
2560 perfmon on Intel CPUs instead of the
2561 CPU specific event set.
2562 timer: [X86] Force use of architectural NMI
2563 timer mode (see also oprofile.timer
2564 for generic hr timer mode)
2565 [s390] Force legacy basic mode sampling
2566 (report cpu_type "timer")
2568 oops=panic Always panic on oopses. Default is to just kill the
2569 process, but there is a small probability of
2570 deadlocking the machine.
2571 This will also cause panics on machine check exceptions.
2572 Useful together with panic=30 to trigger a reboot.
2575 See Documentation/sound/oss/oss-parameters.txt
2577 page_owner= [KNL] Boot-time page_owner enabling option.
2578 Storage of the information about who allocated
2579 each page is disabled in default. With this switch,
2581 on: enable the feature
2583 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2584 timeout > 0: seconds before rebooting
2585 timeout = 0: wait forever
2586 timeout < 0: reboot immediately
2589 panic_on_warn panic() instead of WARN(). Useful to cause kdump
2592 crash_kexec_post_notifiers
2593 Run kdump after running panic-notifiers and dumping
2594 kmsg. This only for the users who doubt kdump always
2595 succeeds in any situation.
2596 Note that this also increases risks of kdump failure,
2597 because some panic notifiers can make the crashed
2598 kernel more unstable.
2600 parkbd.port= [HW] Parallel port number the keyboard adapter is
2601 connected to, default is 0.
2603 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2604 0 for XT, 1 for AT (default is AT).
2607 parport= [HW,PPT] Specify parallel ports. 0 disables.
2608 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2609 Use 'auto' to force the driver to use any
2610 IRQ/DMA settings detected (the default is to
2611 ignore detected IRQ/DMA settings because of
2612 possible conflicts). You can specify the base
2613 address, IRQ, and DMA settings; IRQ and DMA
2614 should be numbers, or 'auto' (for using detected
2615 settings on that particular port), or 'nofifo'
2616 (to avoid using a FIFO even if it is detected).
2617 Parallel ports are assigned in the order they
2618 are specified on the command line, starting
2621 parport_init_mode= [HW,PPT]
2622 Configure VIA parallel port to operate in
2623 a specific mode. This is necessary on Pegasos
2624 computer where firmware has no options for setting
2625 up parallel port mode and sets it to spp.
2626 Currently this function knows 686a and 8231 chips.
2627 Format: [spp|ps2|epp|ecp|ecpepp]
2630 Halt all CPUs after the first oops has been printed for
2631 the specified number of seconds. This is to be used if
2632 your oopses keep scrolling off the screen.
2637 See header of drivers/block/paride/pcd.c.
2638 See also Documentation/blockdev/paride.txt.
2640 pci=option[,option...] [PCI] various PCI subsystem options:
2641 earlydump [X86] dump PCI config space before the kernel
2643 off [X86] don't probe for the PCI bus
2644 bios [X86-32] force use of PCI BIOS, don't access
2645 the hardware directly. Use this if your machine
2646 has a non-standard PCI host bridge.
2647 nobios [X86-32] disallow use of PCI BIOS, only direct
2648 hardware access methods are allowed. Use this
2649 if you experience crashes upon bootup and you
2650 suspect they are caused by the BIOS.
2651 conf1 [X86] Force use of PCI Configuration
2653 conf2 [X86] Force use of PCI Configuration
2655 noaer [PCIE] If the PCIEAER kernel config parameter is
2656 enabled, this kernel boot option can be used to
2657 disable the use of PCIE advanced error reporting.
2658 nodomains [PCI] Disable support for multiple PCI
2659 root domains (aka PCI segments, in ACPI-speak).
2660 nommconf [X86] Disable use of MMCONFIG for PCI
2662 check_enable_amd_mmconf [X86] check for and enable
2663 properly configured MMIO access to PCI
2664 config space on AMD family 10h CPU
2665 nomsi [MSI] If the PCI_MSI kernel config parameter is
2666 enabled, this kernel boot option can be used to
2667 disable the use of MSI interrupts system-wide.
2668 noioapicquirk [APIC] Disable all boot interrupt quirks.
2669 Safety option to keep boot IRQs enabled. This
2670 should never be necessary.
2671 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
2672 primary IO-APIC for bridges that cannot disable
2673 boot IRQs. This fixes a source of spurious IRQs
2674 when the system masks IRQs.
2675 noioapicreroute [APIC] Disable workaround that uses the
2676 boot IRQ equivalent of an IRQ that connects to
2677 a chipset where boot IRQs cannot be disabled.
2678 The opposite of ioapicreroute.
2679 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
2680 routing table. These calls are known to be buggy
2681 on several machines and they hang the machine
2682 when used, but on other computers it's the only
2683 way to get the interrupt routing table. Try
2684 this option if the kernel is unable to allocate
2685 IRQs or discover secondary PCI buses on your
2687 rom [X86] Assign address space to expansion ROMs.
2688 Use with caution as certain devices share
2689 address decoders between ROMs and other
2691 norom [X86] Do not assign address space to
2692 expansion ROMs that do not already have
2693 BIOS assigned address ranges.
2694 nobar [X86] Do not assign address space to the
2695 BARs that weren't assigned by the BIOS.
2696 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
2697 assigned automatically to PCI devices. You can
2698 make the kernel exclude IRQs of your ISA cards
2700 pirqaddr=0xAAAAA [X86] Specify the physical address
2701 of the PIRQ table (normally generated
2702 by the BIOS) if it is outside the
2703 F0000h-100000h range.
2704 lastbus=N [X86] Scan all buses thru bus #N. Can be
2705 useful if the kernel is unable to find your
2706 secondary buses and you want to tell it
2707 explicitly which ones they are.
2708 assign-busses [X86] Always assign all PCI bus
2709 numbers ourselves, overriding
2710 whatever the firmware may have done.
2711 usepirqmask [X86] Honor the possible IRQ mask stored
2712 in the BIOS $PIR table. This is needed on
2713 some systems with broken BIOSes, notably
2714 some HP Pavilion N5400 and Omnibook XE3
2715 notebooks. This will have no effect if ACPI
2716 IRQ routing is enabled.
2717 noacpi [X86] Do not use ACPI for IRQ routing
2718 or for PCI scanning.
2719 use_crs [X86] Use PCI host bridge window information
2720 from ACPI. On BIOSes from 2008 or later, this
2721 is enabled by default. If you need to use this,
2722 please report a bug.
2723 nocrs [X86] Ignore PCI host bridge windows from ACPI.
2724 If you need to use this, please report a bug.
2725 routeirq Do IRQ routing for all PCI devices.
2726 This is normally done in pci_enable_device(),
2727 so this option is a temporary workaround
2728 for broken drivers that don't call it.
2729 skip_isa_align [X86] do not align io start addr, so can
2730 handle more pci cards
2731 firmware [ARM] Do not re-enumerate the bus but instead
2732 just use the configuration from the
2733 bootloader. This is currently used on
2734 IXP2000 systems where the bus has to be
2735 configured a certain way for adjunct CPUs.
2736 noearly [X86] Don't do any early type 1 scanning.
2737 This might help on some broken boards which
2738 machine check when some devices' config space
2739 is read. But various workarounds are disabled
2740 and some IOMMU drivers will not work.
2741 bfsort Sort PCI devices into breadth-first order.
2742 This sorting is done to get a device
2743 order compatible with older (<= 2.4) kernels.
2744 nobfsort Don't sort PCI devices into breadth-first order.
2745 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
2746 tuning and use the BIOS-configured MPS defaults.
2747 pcie_bus_safe Set every device's MPS to the largest value
2748 supported by all devices below the root complex.
2749 pcie_bus_perf Set device MPS to the largest allowable MPS
2750 based on its parent bus. Also set MRRS (Max
2751 Read Request Size) to the largest supported
2752 value (no larger than the MPS that the device
2753 or bus can support) for best performance.
2754 pcie_bus_peer2peer Set every device's MPS to 128B, which
2755 every device is guaranteed to support. This
2756 configuration allows peer-to-peer DMA between
2757 any pair of devices, possibly at the cost of
2758 reduced performance. This also guarantees
2759 that hot-added devices will work.
2760 cbiosize=nn[KMG] The fixed amount of bus space which is
2761 reserved for the CardBus bridge's IO window.
2762 The default value is 256 bytes.
2763 cbmemsize=nn[KMG] The fixed amount of bus space which is
2764 reserved for the CardBus bridge's memory
2765 window. The default value is 64 megabytes.
2768 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
2769 Specifies alignment and device to reassign
2770 aligned memory resources.
2771 If <order of align> is not specified,
2772 PAGE_SIZE is used as alignment.
2773 PCI-PCI bridge can be specified, if resource
2774 windows need to be expanded.
2775 ecrc= Enable/disable PCIe ECRC (transaction layer
2776 end-to-end CRC checking).
2777 bios: Use BIOS/firmware settings. This is the
2781 hpiosize=nn[KMG] The fixed amount of bus space which is
2782 reserved for hotplug bridge's IO window.
2783 Default size is 256 bytes.
2784 hpmemsize=nn[KMG] The fixed amount of bus space which is
2785 reserved for hotplug bridge's memory window.
2786 Default size is 2 megabytes.
2787 realloc= Enable/disable reallocating PCI bridge resources
2788 if allocations done by BIOS are too small to
2789 accommodate resources required by all child
2791 off: Turn realloc off
2793 realloc same as realloc=on
2794 noari do not use PCIe ARI.
2795 pcie_scan_all Scan all possible PCIe devices. Otherwise we
2796 only look for one device below a PCIe downstream
2799 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
2802 force Enable ASPM even on devices that claim not to support it.
2803 WARNING: Forcing ASPM on may cause system lockups.
2805 pcie_hp= [PCIE] PCI Express Hotplug driver options:
2806 nomsi Do not use MSI for PCI Express Native Hotplug (this
2807 makes all PCIe ports use INTx for hotplug services).
2809 pcie_ports= [PCIE] PCIe ports handling:
2810 auto Ask the BIOS whether or not to use native PCIe services
2811 associated with PCIe ports (PME, hot-plug, AER). Use
2812 them only if that is allowed by the BIOS.
2813 native Use native PCIe services associated with PCIe ports
2815 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
2818 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
2819 nomsi Do not use MSI for native PCIe PME signaling (this makes
2820 all PCIe root ports use INTx for all services).
2822 pcmv= [HW,PCMCIA] BadgePAD 4
2826 Keep all power-domains already enabled by bootloader on,
2827 even if no driver has claimed them. This is useful
2828 for debug and development, but should not be
2829 needed on a platform with proper driver support.
2832 See Documentation/blockdev/paride.txt.
2834 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
2837 See arch/parisc/kernel/pdc_chassis.c
2839 percpu_alloc= Select which percpu first chunk allocator to use.
2840 Currently supported values are "embed" and "page".
2841 Archs may support subset or none of the selections.
2842 See comments in mm/percpu.c for details on each
2843 allocator. This parameter is primarily for debugging
2844 and performance comparison.
2847 See Documentation/blockdev/paride.txt.
2850 See Documentation/blockdev/paride.txt.
2852 pirq= [SMP,APIC] Manual mp-table setup
2853 See Documentation/x86/i386/IO-APIC.txt.
2855 plip= [PPT,NET] Parallel port network link
2856 Format: { parport<nr> | timid | 0 }
2857 See also Documentation/parport.txt.
2859 pmtmr= [X86] Manual setup of pmtmr I/O Port.
2860 Override pmtimer IOPort with a hex value.
2864 Enable PNP debug messages (depends on the
2865 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
2866 via /sys/module/pnp/parameters/debug. We always show
2867 current resource usage; turning this on also shows
2868 possible settings and some assignment information.
2874 { on | off | curr | res | no-curr | no-res }
2877 [ISAPNP] Exclude IRQs for the autoconfiguration
2880 [ISAPNP] Exclude DMAs for the autoconfiguration
2882 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
2883 Ranges are in pairs (I/O port base and size).
2886 [ISAPNP] Exclude memory regions for the
2888 Ranges are in pairs (memory base and size).
2890 ports= [IP_VS_FTP] IPVS ftp helper module
2892 Up to 8 (IP_VS_APP_MAX_PORTS) ports
2894 Format: <port>,<port>....
2896 print-fatal-signals=
2897 [KNL] debug: print fatal signals
2899 If enabled, warn about various signal handling
2900 related application anomalies: too many signals,
2901 too many POSIX.1 timers, fatal signals causing a
2904 If you hit the warning due to signal overflow,
2905 you might want to try "ulimit -i unlimited".
2909 printk.always_kmsg_dump=
2910 Trigger kmsg_dump for cases other than kernel oops or
2912 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2915 printk.time= Show timing data prefixed to each printk message line
2916 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2918 processor.max_cstate= [HW,ACPI]
2919 Limit processor to maximum C-state
2920 max_cstate=9 overrides any DMI blacklist limit.
2922 processor.nocst [HW,ACPI]
2923 Ignore the _CST method to determine C-states,
2924 instead using the legacy FADT method
2926 profile= [KNL] Enable kernel profiling via /proc/profile
2927 Format: [schedule,]<number>
2928 Param: "schedule" - profile schedule points.
2929 Param: <number> - step/bucket size as a power of 2 for
2930 statistical time based profiling.
2931 Param: "sleep" - profile D-state sleeping (millisecs).
2932 Requires CONFIG_SCHEDSTATS
2933 Param: "kvm" - profile VM exits.
2935 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
2937 See Documentation/blockdev/ramdisk.txt.
2939 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
2940 probe for; one of (bare|imps|exps|lifebook|any).
2941 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
2943 psmouse.resetafter= [HW,MOUSE]
2944 Try to reset the device after so many bad packets
2947 [HW,MOUSE] Set desired mouse resolution, in dpi.
2948 psmouse.smartscroll=
2949 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
2950 0 = disabled, 1 = enabled (default).
2952 pstore.backend= Specify the name of the pstore backend to use
2955 See Documentation/blockdev/paride.txt.
2958 [KNL] Number of legacy pty's. Overwrites compiled-in
2961 quiet [KNL] Disable most log messages
2966 See Documentation/md.txt.
2968 ramdisk_blocksize= [RAM]
2969 See Documentation/blockdev/ramdisk.txt.
2971 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
2972 See Documentation/blockdev/ramdisk.txt.
2975 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
2976 the specified list of CPUs to be no-callback CPUs.
2977 Invocation of these CPUs' RCU callbacks will
2978 be offloaded to "rcuox/N" kthreads created for
2979 that purpose, where "x" is "b" for RCU-bh, "p"
2980 for RCU-preempt, and "s" for RCU-sched, and "N"
2981 is the CPU number. This reduces OS jitter on the
2982 offloaded CPUs, which can be useful for HPC and
2983 real-time workloads. It can also improve energy
2984 efficiency for asymmetric multiprocessors.
2987 Rather than requiring that offloaded CPUs
2988 (specified by rcu_nocbs= above) explicitly
2989 awaken the corresponding "rcuoN" kthreads,
2990 make these kthreads poll for callbacks.
2991 This improves the real-time response for the
2992 offloaded CPUs by relieving them of the need to
2993 wake up the corresponding kthread, but degrades
2994 energy efficiency by requiring that the kthreads
2995 periodically wake up to do the polling.
2997 rcutree.blimit= [KNL]
2998 Set maximum number of finished RCU callbacks to
2999 process in one batch.
3001 rcutree.gp_init_delay= [KNL]
3002 Set the number of jiffies to delay each step of
3003 RCU grace-period initialization. This only has
3004 effect when CONFIG_RCU_TORTURE_TEST_SLOW_INIT is
3007 rcutree.rcu_fanout_leaf= [KNL]
3008 Increase the number of CPUs assigned to each
3009 leaf rcu_node structure. Useful for very large
3012 rcutree.jiffies_till_sched_qs= [KNL]
3013 Set required age in jiffies for a
3014 given grace period before RCU starts
3015 soliciting quiescent-state help from
3016 rcu_note_context_switch().
3018 rcutree.jiffies_till_first_fqs= [KNL]
3019 Set delay from grace-period initialization to
3020 first attempt to force quiescent states.
3021 Units are jiffies, minimum value is zero,
3022 and maximum value is HZ.
3024 rcutree.jiffies_till_next_fqs= [KNL]
3025 Set delay between subsequent attempts to force
3026 quiescent states. Units are jiffies, minimum
3027 value is one, and maximum value is HZ.
3029 rcutree.kthread_prio= [KNL,BOOT]
3030 Set the SCHED_FIFO priority of the RCU per-CPU
3031 kthreads (rcuc/N). This value is also used for
3032 the priority of the RCU boost threads (rcub/N)
3033 and for the RCU grace-period kthreads (rcu_bh,
3034 rcu_preempt, and rcu_sched). If RCU_BOOST is
3035 set, valid values are 1-99 and the default is 1
3036 (the least-favored priority). Otherwise, when
3037 RCU_BOOST is not set, valid values are 0-99 and
3038 the default is zero (non-realtime operation).
3040 rcutree.rcu_nocb_leader_stride= [KNL]
3041 Set the number of NOCB kthread groups, which
3042 defaults to the square root of the number of
3043 CPUs. Larger numbers reduces the wakeup overhead
3044 on the per-CPU grace-period kthreads, but increases
3045 that same overhead on each group's leader.
3047 rcutree.qhimark= [KNL]
3048 Set threshold of queued RCU callbacks beyond which
3049 batch limiting is disabled.
3051 rcutree.qlowmark= [KNL]
3052 Set threshold of queued RCU callbacks below which
3053 batch limiting is re-enabled.
3055 rcutree.rcu_idle_gp_delay= [KNL]
3056 Set wakeup interval for idle CPUs that have
3057 RCU callbacks (RCU_FAST_NO_HZ=y).
3059 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3060 Set wakeup interval for idle CPUs that have
3061 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3062 Lazy RCU callbacks are those which RCU can
3063 prove do nothing more than free memory.
3065 rcutorture.cbflood_inter_holdoff= [KNL]
3066 Set holdoff time (jiffies) between successive
3067 callback-flood tests.
3069 rcutorture.cbflood_intra_holdoff= [KNL]
3070 Set holdoff time (jiffies) between successive
3071 bursts of callbacks within a given callback-flood
3074 rcutorture.cbflood_n_burst= [KNL]
3075 Set the number of bursts making up a given
3076 callback-flood test. Set this to zero to
3077 disable callback-flood testing.
3079 rcutorture.cbflood_n_per_burst= [KNL]
3080 Set the number of callbacks to be registered
3081 in a given burst of a callback-flood test.
3083 rcutorture.fqs_duration= [KNL]
3084 Set duration of force_quiescent_state bursts.
3086 rcutorture.fqs_holdoff= [KNL]
3087 Set holdoff time within force_quiescent_state bursts.
3089 rcutorture.fqs_stutter= [KNL]
3090 Set wait time between force_quiescent_state bursts.
3092 rcutorture.gp_exp= [KNL]
3093 Use expedited update-side primitives.
3095 rcutorture.gp_normal= [KNL]
3096 Use normal (non-expedited) update-side primitives.
3097 If both gp_exp and gp_normal are set, do both.
3098 If neither gp_exp nor gp_normal are set, still
3101 rcutorture.n_barrier_cbs= [KNL]
3102 Set callbacks/threads for rcu_barrier() testing.
3104 rcutorture.nfakewriters= [KNL]
3105 Set number of concurrent RCU writers. These just
3106 stress RCU, they don't participate in the actual
3107 test, hence the "fake".
3109 rcutorture.nreaders= [KNL]
3110 Set number of RCU readers.
3112 rcutorture.object_debug= [KNL]
3113 Enable debug-object double-call_rcu() testing.
3115 rcutorture.onoff_holdoff= [KNL]
3116 Set time (s) after boot for CPU-hotplug testing.
3118 rcutorture.onoff_interval= [KNL]
3119 Set time (s) between CPU-hotplug operations, or
3120 zero to disable CPU-hotplug testing.
3122 rcutorture.torture_runnable= [BOOT]
3123 Start rcutorture running at boot time.
3125 rcutorture.shuffle_interval= [KNL]
3126 Set task-shuffle interval (s). Shuffling tasks
3127 allows some CPUs to go into dyntick-idle mode
3128 during the rcutorture test.
3130 rcutorture.shutdown_secs= [KNL]
3131 Set time (s) after boot system shutdown. This
3132 is useful for hands-off automated testing.
3134 rcutorture.stall_cpu= [KNL]
3135 Duration of CPU stall (s) to test RCU CPU stall
3136 warnings, zero to disable.
3138 rcutorture.stall_cpu_holdoff= [KNL]
3139 Time to wait (s) after boot before inducing stall.
3141 rcutorture.stat_interval= [KNL]
3142 Time (s) between statistics printk()s.
3144 rcutorture.stutter= [KNL]
3145 Time (s) to stutter testing, for example, specifying
3146 five seconds causes the test to run for five seconds,
3147 wait for five seconds, and so on. This tests RCU's
3148 ability to transition abruptly to and from idle.
3150 rcutorture.test_boost= [KNL]
3151 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3152 "Maybe" means test if the RCU implementation
3153 under test support RCU priority boosting.
3155 rcutorture.test_boost_duration= [KNL]
3156 Duration (s) of each individual boost test.
3158 rcutorture.test_boost_interval= [KNL]
3159 Interval (s) between each boost test.
3161 rcutorture.test_no_idle_hz= [KNL]
3162 Test RCU's dyntick-idle handling. See also the
3163 rcutorture.shuffle_interval parameter.
3165 rcutorture.torture_type= [KNL]
3166 Specify the RCU implementation to test.
3168 rcutorture.verbose= [KNL]
3169 Enable additional printk() statements.
3171 rcupdate.rcu_expedited= [KNL]
3172 Use expedited grace-period primitives, for
3173 example, synchronize_rcu_expedited() instead
3174 of synchronize_rcu(). This reduces latency,
3175 but can increase CPU utilization, degrade
3176 real-time latency, and degrade energy efficiency.
3178 rcupdate.rcu_cpu_stall_suppress= [KNL]
3179 Suppress RCU CPU stall warning messages.
3181 rcupdate.rcu_cpu_stall_timeout= [KNL]
3182 Set timeout for RCU CPU stall warning messages.
3184 rcupdate.rcu_task_stall_timeout= [KNL]
3185 Set timeout in jiffies for RCU task stall warning
3186 messages. Disable with a value less than or equal
3189 rcupdate.rcu_self_test= [KNL]
3190 Run the RCU early boot self tests
3192 rcupdate.rcu_self_test_bh= [KNL]
3193 Run the RCU bh early boot self tests
3195 rcupdate.rcu_self_test_sched= [KNL]
3196 Run the RCU sched early boot self tests
3200 Run specified binary instead of /init from the ramdisk,
3201 used for early userspace startup. See initrd.
3204 Format (x86 or x86_64):
3205 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3207 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3209 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3210 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3211 reboot_force is either force or not specified,
3212 reboot_cpu is s[mp]#### with #### being the processor
3213 to be used for rebooting.
3216 [KNL, SMP] Set scheduler's default relax_domain_level.
3217 See Documentation/cgroups/cpusets.txt.
3219 relative_sleep_states=
3220 [SUSPEND] Use sleep state labeling where the deepest
3221 state available other than hibernation is always "mem".
3222 Format: { "0" | "1" }
3223 0 -- Traditional sleep state labels.
3224 1 -- Relative sleep state labels.
3226 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
3228 reservetop= [X86-32]
3230 Reserves a hole at the top of the kernel virtual
3235 Set the amount of memory to reserve for BIOS at
3236 the bottom of the address space.
3238 reset_devices [KNL] Force drivers to reset the underlying device
3239 during initialization.
3242 Specify the partition device for software suspend
3244 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3246 resume_offset= [SWSUSP]
3247 Specify the offset from the beginning of the partition
3248 given by "resume=" at which the swap header is located,
3249 in <PAGE_SIZE> units (needed only for swap files).
3250 See Documentation/power/swsusp-and-swap-files.txt
3252 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
3253 read the resume files
3255 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
3256 Useful for devices that are detected asynchronously
3257 (e.g. USB and MMC devices).
3259 hibernate= [HIBERNATION]
3260 noresume Don't check if there's a hibernation image
3261 present during boot.
3262 nocompress Don't compress/decompress hibernation images.
3263 no Disable hibernation and resume.
3265 retain_initrd [RAM] Keep initrd memory after extraction
3267 rfkill.default_state=
3268 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
3269 etc. communication is blocked by default.
3272 rfkill.master_switch_mode=
3273 0 The "airplane mode" button does nothing.
3274 1 The "airplane mode" button toggles between everything
3275 blocked and the previous configuration.
3276 2 The "airplane mode" button toggles between everything
3277 blocked and everything unblocked.
3279 rhash_entries= [KNL,NET]
3280 Set number of hash buckets for route cache
3282 ro [KNL] Mount root device read-only on boot
3284 root= [KNL] Root filesystem
3285 See name_to_dev_t comment in init/do_mounts.c.
3287 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
3288 mount the root filesystem
3290 rootflags= [KNL] Set root filesystem mount option string
3292 rootfstype= [KNL] Set root filesystem type
3294 rootwait [KNL] Wait (indefinitely) for root device to show up.
3295 Useful for devices that are detected asynchronously
3296 (e.g. USB and MMC devices).
3298 rproc_mem=nn[KMG][@address]
3299 [KNL,ARM,CMA] Remoteproc physical memory block.
3300 Memory area to be used by remote processor image,
3303 rw [KNL] Mount root device read-write on boot
3305 S [KNL] Run init in single mode
3307 s390_iommu= [HW,S390]
3308 Set s390 IOTLB flushing mode
3310 With strict flushing every unmap operation will result in
3311 an IOTLB flush. Default is lazy flushing before reuse,
3315 See drivers/net/irda/sa1100_ir.c.
3317 sbni= [NET] Granch SBNI12 leased line adapter
3319 sched_debug [KNL] Enables verbose scheduler debug messages.
3321 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
3322 xtime_lock contention on larger systems, and/or RCU lock
3323 contention on all systems with CONFIG_MAXSMP set.
3324 Format: { "0" | "1" }
3325 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
3327 Note: increases power consumption, thus should only be
3328 enabled if running jitter sensitive (HPC/RT) workloads.
3330 security= [SECURITY] Choose a security module to enable at boot.
3331 If this boot parameter is not specified, only the first
3332 security module asking for security registration will be
3333 loaded. An invalid security module name will be treated
3334 as if no module has been chosen.
3336 selinux= [SELINUX] Disable or enable SELinux at boot time.
3337 Format: { "0" | "1" }
3338 See security/selinux/Kconfig help text.
3341 Default value is set via kernel config option.
3342 If enabled at boot time, /selinux/disable can be used
3343 later to disable prior to initial policy load.
3345 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
3346 Format: { "0" | "1" }
3347 See security/apparmor/Kconfig help text
3350 Default value is set via kernel config option.
3352 serialnumber [BUGS=X86-32]
3355 Maximal number of shapers.
3357 show_msr= [x86] show boot-time MSR settings
3358 Format: { <integer> }
3359 Show boot-time (BIOS-initialized) MSR settings.
3360 The parameter means the number of CPUs to show,
3361 for example 1 means boot CPU only.
3369 Disable merging of slabs with similar size. May be
3370 necessary if there is some reason to distinguish
3371 allocs to different slabs. Debug options disable
3372 merging on their own.
3373 For more information see Documentation/vm/slub.txt.
3375 slab_max_order= [MM, SLAB]
3376 Determines the maximum allowed order for slabs.
3377 A high setting may cause OOMs due to memory
3378 fragmentation. Defaults to 1 for systems with
3379 more than 32MB of RAM, 0 otherwise.
3381 slub_debug[=options[,slabs]] [MM, SLUB]
3382 Enabling slub_debug allows one to determine the
3383 culprit if slab objects become corrupted. Enabling
3384 slub_debug can create guard zones around objects and
3385 may poison objects when not in use. Also tracks the
3386 last alloc / free. For more information see
3387 Documentation/vm/slub.txt.
3389 slub_max_order= [MM, SLUB]
3390 Determines the maximum allowed order for slabs.
3391 A high setting may cause OOMs due to memory
3392 fragmentation. For more information see
3393 Documentation/vm/slub.txt.
3395 slub_min_objects= [MM, SLUB]
3396 The minimum number of objects per slab. SLUB will
3397 increase the slab order up to slub_max_order to
3398 generate a sufficiently large slab able to contain
3399 the number of objects indicated. The higher the number
3400 of objects the smaller the overhead of tracking slabs
3401 and the less frequently locks need to be acquired.
3402 For more information see Documentation/vm/slub.txt.
3404 slub_min_order= [MM, SLUB]
3405 Determines the minimum page order for slabs. Must be
3406 lower than slub_max_order.
3407 For more information see Documentation/vm/slub.txt.
3409 slub_nomerge [MM, SLUB]
3410 Same with slab_nomerge. This is supported for legacy.
3411 See slab_nomerge for more information.
3414 Format: <io1>[,<io2>[,...,<io8>]]
3416 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
3417 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
3418 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
3419 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
3420 smsc-ircc2.ircc_irq= [HW] IRQ line
3421 smsc-ircc2.ircc_dma= [HW] DMA channel
3422 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
3423 0: Toshiba Satellite 1800 (GP data pin select)
3424 1: Fast pin select (default)
3428 [KNL] Should the soft-lockup detector generate panics.
3431 softlockup_all_cpu_backtrace=
3432 [KNL] Should the soft-lockup detector generate
3433 backtraces on all cpus.
3436 sonypi.*= [HW] Sony Programmable I/O Control Device driver
3437 See Documentation/laptops/sonypi.txt
3439 spia_io_base= [HW,MTD]
3445 Enabled the stack tracer on boot up.
3447 stacktrace_filter=[function-list]
3448 [FTRACE] Limit the functions that the stack tracer
3449 will trace at boot up. function-list is a comma separated
3450 list of functions. This list can be changed at run
3451 time by the stack_trace_filter file in the debugfs
3452 tracing directory. Note, this enables stack tracing
3453 and the stacktrace above is not needed.
3457 Set the STI (builtin display/keyboard on the HP-PARISC
3458 machines) console (graphic card) which should be used
3459 as the initial boot-console.
3460 See also comment in drivers/video/console/sticore.c.
3463 See comment in drivers/video/console/sticore.c.
3466 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
3468 sunrpc.min_resvport=
3469 sunrpc.max_resvport=
3471 SunRPC servers often require that client requests
3472 originate from a privileged port (i.e. a port in the
3473 range 0 < portnr < 1024).
3474 An administrator who wishes to reserve some of these
3475 ports for other uses may adjust the range that the
3476 kernel's sunrpc client considers to be privileged
3477 using these two parameters to set the minimum and
3478 maximum port values.
3482 Control how the NFS server code allocates CPUs to
3483 service thread pools. Depending on how many NICs
3484 you have and where their interrupts are bound, this
3485 option will affect which CPUs will do NFS serving.
3486 Note: this parameter cannot be changed while the
3487 NFS server is running.
3489 auto the server chooses an appropriate mode
3490 automatically using heuristics
3491 global a single global pool contains all CPUs
3492 percpu one pool for each CPU
3493 pernode one pool for each NUMA node (equivalent
3494 to global on non-NUMA machines)
3496 sunrpc.tcp_slot_table_entries=
3497 sunrpc.udp_slot_table_entries=
3499 Sets the upper limit on the number of simultaneous
3500 RPC calls that can be sent from the client to a
3501 server. Increasing these values may allow you to
3502 improve throughput, but will also increase the
3503 amount of memory reserved for use by the client.
3505 suspend.pm_test_delay=
3507 Sets the number of seconds to remain in a suspend test
3508 mode before resuming the system (see
3509 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
3510 is set. Default value is 5.
3513 [KNL] Enable accounting of swap in memory resource
3514 controller if no parameter or 1 is given or disable
3515 it if 0 is given (See Documentation/cgroups/memory.txt)
3517 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
3518 Format: { <int> | force }
3519 <int> -- Number of I/O TLB slabs
3520 force -- force using of bounce buffers even if they
3521 wouldn't be automatically used by the kernel
3525 sysfs.deprecated=0|1 [KNL]
3526 Enable/disable old style sysfs layout for old udev
3527 on older distributions. When this option is enabled
3528 very new udev will not work anymore. When this option
3529 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
3530 in older udev will not work anymore.
3531 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
3532 the kernel configuration.
3534 sysrq_always_enabled
3536 Ignore sysrq setting - this boot parameter will
3537 neutralize any effect of /proc/sys/kernel/sysrq.
3538 Useful for debugging.
3540 tcpmhash_entries= [KNL,NET]
3541 Set the number of tcp_metrics_hash slots.
3542 Default value is 8192 or 16384 depending on total
3543 ram pages. This is used to specify the TCP metrics
3544 cache size. See Documentation/networking/ip-sysctl.txt
3545 "tcp_no_metrics_save" section for more details.
3549 test_suspend= [SUSPEND][,N]
3550 Specify "mem" (for Suspend-to-RAM) or "standby" (for
3551 standby suspend) or "freeze" (for suspend type freeze)
3552 as the system sleep state during system startup with
3553 the optional capability to repeat N number of times.
3554 The system is woken from this state using a
3555 wakeup-capable RTC alarm.
3557 thash_entries= [KNL,NET]
3558 Set number of hash buckets for TCP connection
3560 thermal.act= [HW,ACPI]
3561 -1: disable all active trip points in all thermal zones
3562 <degrees C>: override all lowest active trip points
3564 thermal.crt= [HW,ACPI]
3565 -1: disable all critical trip points in all thermal zones
3566 <degrees C>: override all critical trip points
3568 thermal.nocrt= [HW,ACPI]
3569 Set to disable actions on ACPI thermal zone
3570 critical and hot trip points.
3572 thermal.off= [HW,ACPI]
3573 1: disable ACPI thermal control
3575 thermal.psv= [HW,ACPI]
3576 -1: disable all passive trip points
3577 <degrees C>: override all passive trip points to this
3580 thermal.tzp= [HW,ACPI]
3581 Specify global default ACPI thermal zone polling rate
3582 <deci-seconds>: poll all this frequency
3583 0: no polling (default)
3586 Force threading of all interrupt handlers except those
3587 marked explicitly IRQF_NO_THREAD.
3590 Enable the Transcendent memory driver if built-in.
3592 tmem.cleancache=0|1 [KNL, XEN]
3593 Default is on (1). Disable the usage of the cleancache
3594 API to send anonymous pages to the hypervisor.
3596 tmem.frontswap=0|1 [KNL, XEN]
3597 Default is on (1). Disable the usage of the frontswap
3598 API to send swap pages to the hypervisor. If disabled
3599 the selfballooning and selfshrinking are force disabled.
3601 tmem.selfballooning=0|1 [KNL, XEN]
3602 Default is on (1). Disable the driving of swap pages
3605 tmem.selfshrinking=0|1 [KNL, XEN]
3606 Default is on (1). Partial swapoff that immediately
3607 transfers pages from Xen hypervisor back to the
3608 kernel based on different criteria.
3612 Specify if the kernel should make use of the cpu
3613 topology information if the hardware supports this.
3614 The scheduler will make use of this information and
3615 e.g. base its process migration decisions on it.
3618 topology_updates= [KNL, PPC, NUMA]
3620 Specify if the kernel should ignore (off)
3621 topology updates sent by the hypervisor to this
3626 tpm_suspend_pcr=[HW,TPM]
3627 Format: integer pcr id
3628 Specify that at suspend time, the tpm driver
3629 should extend the specified pcr with zeros,
3630 as a workaround for some chips which fail to
3631 flush the last written pcr on TPM_SaveState.
3632 This will guarantee that all the other pcrs
3635 trace_buf_size=nn[KMG]
3636 [FTRACE] will set tracing buffer size on each cpu.
3638 trace_event=[event-list]
3639 [FTRACE] Set and start specified trace events in order
3640 to facilitate early boot debugging.
3641 See also Documentation/trace/events.txt
3643 trace_options=[option-list]
3644 [FTRACE] Enable or disable tracer options at boot.
3645 The option-list is a comma delimited list of options
3646 that can be enabled or disabled just as if you were
3647 to echo the option name into
3649 /sys/kernel/debug/tracing/trace_options
3651 For example, to enable stacktrace option (to dump the
3652 stack trace of each event), add to the command line:
3654 trace_options=stacktrace
3656 See also Documentation/trace/ftrace.txt "trace options"
3660 Have the tracepoints sent to printk as well as the
3661 tracing ring buffer. This is useful for early boot up
3662 where the system hangs or reboots and does not give the
3663 option for reading the tracing buffer or performing a
3664 ftrace_dump_on_oops.
3666 To turn off having tracepoints sent to printk,
3667 echo 0 > /proc/sys/kernel/tracepoint_printk
3668 Note, echoing 1 into this file without the
3669 tracepoint_printk kernel cmdline option has no effect.
3673 Having tracepoints sent to printk() and activating high
3674 frequency tracepoints such as irq or sched, can cause
3675 the system to live lock.
3678 [FTRACE] enable this option to disable tracing when a
3679 warning is hit. This turns off "tracing_on". Tracing can
3680 be enabled again by echoing '1' into the "tracing_on"
3681 file located in /sys/kernel/debug/tracing/
3683 This option is useful, as it disables the trace before
3684 the WARNING dump is called, which prevents the trace to
3685 be filled with content caused by the warning output.
3687 This option can also be set at run time via the sysctl
3688 option: kernel/traceoff_on_warning
3690 transparent_hugepage=
3692 Format: [always|madvise|never]
3693 Can be used to control the default behavior of the system
3694 with respect to transparent hugepages.
3695 See Documentation/vm/transhuge.txt for more details.
3697 tsc= Disable clocksource stability checks for TSC.
3699 [x86] reliable: mark tsc clocksource as reliable, this
3700 disables clocksource verification at runtime, as well
3701 as the stability checks done at bootup. Used to enable
3702 high-resolution timer mode on older hardware, and in
3703 virtualized environment.
3704 [x86] noirqtime: Do not use TSC to do irq accounting.
3705 Used to run time disable IRQ_TIME_ACCOUNTING on any
3706 platforms where RDTSC is slow and this accounting
3709 turbografx.map[2|3]= [HW,JOY]
3710 TurboGraFX parallel port interface
3712 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
3713 See also Documentation/input/joystick-parport.txt
3715 udbg-immortal [PPC] When debugging early kernel crashes that
3716 happen after console_init() and before a proper
3717 console driver takes over, this boot options might
3718 help "seeing" what's going on.
3720 uhash_entries= [KNL,NET]
3721 Set number of hash buckets for UDP/UDP-Lite connections
3724 [USB] Ignore overcurrent events (default N).
3725 Some badly-designed motherboards generate lots of
3726 bogus events, for ports that aren't wired to
3727 anything. Set this parameter to avoid log spamming.
3728 Note that genuine overcurrent events won't be
3732 [X86] Cause panic on unknown NMI.
3734 usbcore.authorized_default=
3735 [USB] Default USB device authorization:
3736 (default -1 = authorized except for wireless USB,
3737 0 = not authorized, 1 = authorized)
3739 usbcore.autosuspend=
3740 [USB] The autosuspend time delay (in seconds) used
3741 for newly-detected USB devices (default 2). This
3742 is the time required before an idle device will be
3743 autosuspended. Devices for which the delay is set
3744 to a negative value won't be autosuspended at all.
3746 usbcore.usbfs_snoop=
3747 [USB] Set to log all usbfs traffic (default 0 = off).
3749 usbcore.blinkenlights=
3750 [USB] Set to cycle leds on hubs (default 0 = off).
3752 usbcore.old_scheme_first=
3753 [USB] Start with the old device initialization
3754 scheme (default 0 = off).
3756 usbcore.usbfs_memory_mb=
3757 [USB] Memory limit (in MB) for buffers allocated by
3758 usbfs (default = 16, 0 = max = 2047).
3760 usbcore.use_both_schemes=
3761 [USB] Try the other device initialization scheme
3762 if the first one fails (default 1 = enabled).
3764 usbcore.initial_descriptor_timeout=
3765 [USB] Specifies timeout for the initial 64-byte
3766 USB_REQ_GET_DESCRIPTOR request in milliseconds
3767 (default 5000 = 5.0 seconds).
3770 [USBHID] The interval which mice are to be polled at.
3772 usb-storage.delay_use=
3773 [UMS] The delay in seconds before a new device is
3774 scanned for Logical Units (default 1).
3777 [UMS] A list of quirks entries to supplement or
3778 override the built-in unusual_devs list. List
3779 entries are separated by commas. Each entry has
3780 the form VID:PID:Flags where VID and PID are Vendor
3781 and Product ID values (4-digit hex numbers) and
3782 Flags is a set of characters, each corresponding
3783 to a common usb-storage quirk flag as follows:
3784 a = SANE_SENSE (collect more than 18 bytes
3786 b = BAD_SENSE (don't collect more than 18
3787 bytes of sense data);
3788 c = FIX_CAPACITY (decrease the reported
3789 device capacity by one sector);
3790 d = NO_READ_DISC_INFO (don't use
3791 READ_DISC_INFO command);
3792 e = NO_READ_CAPACITY_16 (don't use
3793 READ_CAPACITY_16 command);
3794 f = NO_REPORT_OPCODES (don't use report opcodes
3796 g = MAX_SECTORS_240 (don't transfer more than
3797 240 sectors at a time, uas only);
3798 h = CAPACITY_HEURISTICS (decrease the
3799 reported device capacity by one
3800 sector if the number is odd);
3801 i = IGNORE_DEVICE (don't bind to this
3803 l = NOT_LOCKABLE (don't try to lock and
3804 unlock ejectable media);
3805 m = MAX_SECTORS_64 (don't transfer more
3806 than 64 sectors = 32 KB at a time);
3807 n = INITIAL_READ10 (force a retry of the
3808 initial READ(10) command);
3809 o = CAPACITY_OK (accept the capacity
3810 reported by the device);
3811 p = WRITE_CACHE (the device cache is ON
3813 r = IGNORE_RESIDUE (the device reports
3814 bogus residue values);
3815 s = SINGLE_LUN (the device has only one
3817 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
3818 commands, uas only);
3819 u = IGNORE_UAS (don't bind to the uas driver);
3820 w = NO_WP_DETECT (don't test whether the
3821 medium is write-protected).
3822 Example: quirks=0419:aaf5:rl,0421:0433:rc
3824 user_debug= [KNL,ARM]
3826 See arch/arm/Kconfig.debug help text.
3827 1 - undefined instruction events
3829 4 - invalid data aborts
3832 Example: user_debug=31
3835 [X86] Flags controlling user PTE allocations.
3837 nohigh = do not allocate PTE pages in
3838 HIGHMEM regardless of setting
3842 On X86_32, this is an alias for vdso32=. Otherwise:
3844 vdso=1: enable VDSO (the default)
3845 vdso=0: disable VDSO mapping
3847 vdso32= [X86] Control the 32-bit vDSO
3848 vdso32=1: enable 32-bit VDSO
3849 vdso32=0 or vdso32=2: disable 32-bit VDSO
3851 See the help text for CONFIG_COMPAT_VDSO for more
3852 details. If CONFIG_COMPAT_VDSO is set, the default is
3853 vdso32=0; otherwise, the default is vdso32=1.
3855 For compatibility with older kernels, vdso32=2 is an
3858 Try vdso32=0 if you encounter an error that says:
3859 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
3862 vector=percpu: enable percpu vector domain
3864 video= [FB] Frame buffer configuration
3865 See Documentation/fb/modedb.txt.
3867 video.brightness_switch_enabled= [0,1]
3868 If set to 1, on receiving an ACPI notify event
3869 generated by hotkey, video driver will adjust brightness
3870 level and then send out the event to user space through
3871 the allocated input device; If set to 0, video driver
3872 will only send out the event without touching backlight
3877 [VMMIO] Memory mapped virtio (platform) device.
3879 <size>@<baseaddr>:<irq>[:<id>]
3881 <size> := size (can use standard suffixes
3883 <baseaddr> := physical base address
3884 <irq> := interrupt number (as passed to
3886 <id> := (optional) platform device id
3888 virtio_mmio.device=1K@0x100b0000:48:7
3890 Can be used multiple times for multiple devices.
3892 vga= [BOOT,X86-32] Select a particular video mode
3893 See Documentation/x86/boot.txt and
3894 Documentation/svga.txt.
3895 Use vga=ask for menu.
3896 This is actually a boot loader parameter; the value is
3897 passed to the kernel using a special protocol.
3899 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
3900 size of <nn>. This can be used to increase the
3901 minimum size (128MB on x86). It can also be used to
3902 decrease the size and leave more room for directly
3905 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
3908 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
3911 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
3915 Controls the behavior of vsyscalls (i.e. calls to
3916 fixed addresses of 0xffffffffff600x00 from legacy
3917 code). Most statically-linked binaries and older
3918 versions of glibc use these calls. Because these
3919 functions are at fixed addresses, they make nice
3920 targets for exploits that can control RIP.
3922 emulate [default] Vsyscalls turn into traps and are
3923 emulated reasonably safely.
3925 native Vsyscalls are native syscall instructions.
3926 This is a little bit faster than trapping
3927 and makes a few dynamic recompilers work
3928 better than they would in emulation mode.
3929 It also makes exploits much easier to write.
3931 none Vsyscalls don't work at all. This makes
3932 them quite hard to use for exploits but
3933 might break your system.
3935 vt.color= [VT] Default text color.
3936 Format: 0xYX, X = foreground, Y = background.
3937 Default: 0x07 = light gray on black.
3939 vt.cur_default= [VT] Default cursor shape.
3940 Format: 0xCCBBAA, where AA, BB, and CC are the same as
3941 the parameters of the <Esc>[?A;B;Cc escape sequence;
3942 see VGA-softcursor.txt. Default: 2 = underline.
3944 vt.default_blu= [VT]
3945 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
3946 Change the default blue palette of the console.
3947 This is a 16-member array composed of values
3950 vt.default_grn= [VT]
3951 Format: <green0>,<green1>,<green2>,...,<green15>
3952 Change the default green palette of the console.
3953 This is a 16-member array composed of values
3956 vt.default_red= [VT]
3957 Format: <red0>,<red1>,<red2>,...,<red15>
3958 Change the default red palette of the console.
3959 This is a 16-member array composed of values
3965 Set system-wide default UTF-8 mode for all tty's.
3966 Default is 1, i.e. UTF-8 mode is enabled for all
3967 newly opened terminals.
3969 vt.global_cursor_default=
3972 Set system-wide default for whether a cursor
3973 is shown on new VTs. Default is -1,
3974 i.e. cursors will be created by default unless
3975 overridden by individual drivers. 0 will hide
3976 cursors, 1 will display them.
3978 vt.italic= [VT] Default color for italic text; 0-15.
3981 vt.underline= [VT] Default color for underlined text; 0-15.
3984 watchdog timers [HW,WDT] For information on watchdog timers,
3985 see Documentation/watchdog/watchdog-parameters.txt
3986 or other driver-specific files in the
3987 Documentation/watchdog/ directory.
3989 workqueue.disable_numa
3990 By default, all work items queued to unbound
3991 workqueues are affine to the NUMA nodes they're
3992 issued on, which results in better behavior in
3993 general. If NUMA affinity needs to be disabled for
3994 whatever reason, this option can be used. Note
3995 that this also can be controlled per-workqueue for
3996 workqueues visible under /sys/bus/workqueue/.
3998 workqueue.power_efficient
3999 Per-cpu workqueues are generally preferred because
4000 they show better performance thanks to cache
4001 locality; unfortunately, per-cpu workqueues tend to
4002 be more power hungry than unbound workqueues.
4004 Enabling this makes the per-cpu workqueues which
4005 were observed to contribute significantly to power
4006 consumption unbound, leading to measurably lower
4007 power usage at the cost of small performance
4010 The default value of this parameter is determined by
4011 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
4013 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
4014 default x2apic cluster mode on platforms
4017 x86_intel_mid_timer= [X86-32,APBT]
4018 Choose timer option for x86 Intel MID platform.
4019 Two valid options are apbt timer only and lapic timer
4020 plus one apbt timer for broadcast timer.
4021 x86_intel_mid_timer=apbt_only | lapic_and_apbt
4023 xen_emul_unplug= [HW,X86,XEN]
4024 Unplug Xen emulated devices
4025 Format: [unplug0,][unplug1]
4026 ide-disks -- unplug primary master IDE devices
4027 aux-ide-disks -- unplug non-primary-master IDE devices
4028 nics -- unplug network devices
4029 all -- unplug all emulated devices (NICs and IDE disks)
4030 unnecessary -- unplugging emulated devices is
4031 unnecessary even if the host did not respond to
4033 never -- do not unplug even if version check succeeds
4035 xen_nopvspin [X86,XEN]
4036 Disables the ticketlock slowpath using Xen PV
4040 Disables the PV optimizations forcing the HVM guest to
4041 run as generic HVM guest with no PV drivers.
4043 xirc2ps_cs= [NET,PCMCIA]
4045 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
4047 ______________________________________________________________________
4051 Add more DRM drivers.