4 The following is a consolidated list of the kernel parameters as implemented
5 (mostly) by the __setup() macro and sorted into English Dictionary order
6 (defined as ignoring all punctuation and sorting digits before letters in a
7 case insensitive manner), and with descriptions where known.
9 Module parameters for loadable modules are specified only as the
10 parameter name with optional '=' and value as appropriate, such as:
12 modprobe usbcore blinkenlights=1
14 Module parameters for modules that are built into the kernel image
15 are specified on the kernel command line with the module name plus
16 '.' plus parameter name, with '=' and value if appropriate, such as:
18 usbcore.blinkenlights=1
20 Hyphens (dashes) and underscores are equivalent in parameter names, so
21 log_buf_len=1M print-fatal-signals=1
22 can also be entered as
23 log-buf-len=1M print_fatal_signals=1
26 This document may not be entirely up to date and comprehensive. The command
27 "modinfo -p ${modulename}" shows a current list of all parameters of a loadable
28 module. Loadable modules, after being loaded into the running kernel, also
29 reveal their parameters in /sys/module/${modulename}/parameters/. Some of these
30 parameters may be changed at runtime by the command
31 "echo -n ${value} > /sys/module/${modulename}/parameters/${parm}".
33 The parameters listed below are only valid if certain kernel build options were
34 enabled and if respective hardware is present. The text in square brackets at
35 the beginning of each description states the restrictions within which a
36 parameter is applicable:
38 ACPI ACPI support is enabled.
39 AGP AGP (Accelerated Graphics Port) is enabled.
40 ALSA ALSA sound support is enabled.
41 APIC APIC support is enabled.
42 APM Advanced Power Management support is enabled.
43 ARM ARM architecture is enabled.
44 AVR32 AVR32 architecture is enabled.
45 AX25 Appropriate AX.25 support is enabled.
46 BLACKFIN Blackfin architecture is enabled.
47 CLK Common clock infrastructure is enabled.
48 CMA Contiguous Memory Area support is enabled.
49 DRM Direct Rendering Management support is enabled.
50 DYNAMIC_DEBUG Build in debug messages and enable them at runtime
51 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled
52 EFI EFI Partitioning (GPT) is enabled
53 EIDE EIDE/ATAPI support is enabled.
54 EVM Extended Verification Module
55 FB The frame buffer device is enabled.
56 FTRACE Function tracing enabled.
57 GCOV GCOV profiling is enabled.
58 HW Appropriate hardware is enabled.
59 IA-64 IA-64 architecture is enabled.
60 IMA Integrity measurement architecture is enabled.
61 IOSCHED More than one I/O scheduler is enabled.
62 IP_PNP IP DHCP, BOOTP, or RARP is enabled.
63 IPV6 IPv6 support is enabled.
64 ISAPNP ISA PnP code is enabled.
65 ISDN Appropriate ISDN support is enabled.
66 JOY Appropriate joystick support is enabled.
67 KGDB Kernel debugger support is enabled.
68 KVM Kernel Virtual Machine support is enabled.
69 LIBATA Libata driver is enabled
70 LP Printer support is enabled.
71 LOOP Loopback device support is enabled.
72 M68k M68k architecture is enabled.
73 These options have more detailed description inside of
74 Documentation/m68k/kernel-options.txt.
75 MDA MDA console support is enabled.
76 MIPS MIPS architecture is enabled.
77 MOUSE Appropriate mouse support is enabled.
78 MSI Message Signaled Interrupts (PCI).
79 MTD MTD (Memory Technology Device) support is enabled.
80 NET Appropriate network support is enabled.
81 NUMA NUMA support is enabled.
82 NFS Appropriate NFS support is enabled.
83 OSS OSS sound support is enabled.
84 PV_OPS A paravirtualized kernel is enabled.
85 PARIDE The ParIDE (parallel port IDE) subsystem is enabled.
86 PARISC The PA-RISC architecture is enabled.
87 PCI PCI bus support is enabled.
88 PCIE PCI Express support is enabled.
89 PCMCIA The PCMCIA subsystem is enabled.
90 PNP Plug & Play support is enabled.
91 PPC PowerPC architecture is enabled.
92 PPT Parallel port support is enabled.
93 PS2 Appropriate PS/2 support is enabled.
94 RAM RAM disk support is enabled.
95 S390 S390 architecture is enabled.
96 SCSI Appropriate SCSI support is enabled.
97 A lot of drivers have their options described inside
98 the Documentation/scsi/ sub-directory.
99 SECURITY Different security models are enabled.
100 SELINUX SELinux support is enabled.
101 APPARMOR AppArmor support is enabled.
102 SERIAL Serial support is enabled.
103 SH SuperH architecture is enabled.
104 SMP The kernel is an SMP kernel.
105 SPARC Sparc architecture is enabled.
106 SWSUSP Software suspend (hibernation) is enabled.
107 SUSPEND System suspend states are enabled.
108 TPM TPM drivers are enabled.
109 TS Appropriate touchscreen support is enabled.
110 UMS USB Mass Storage support is enabled.
111 USB USB support is enabled.
112 USBHID USB Human Interface Device support is enabled.
113 V4L Video For Linux support is enabled.
114 VMMIO Driver for memory mapped virtio devices is enabled.
115 VGA The VGA console has been enabled.
116 VT Virtual terminal support is enabled.
117 WDT Watchdog support is enabled.
118 XT IBM PC/XT MFM hard disk support is enabled.
119 X86-32 X86-32, aka i386 architecture is enabled.
120 X86-64 X86-64 architecture is enabled.
121 More X86-64 boot options can be found in
122 Documentation/x86/x86_64/boot-options.txt .
123 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64)
124 XEN Xen support is enabled
126 In addition, the following text indicates that the option:
128 BUGS= Relates to possible processor bugs on the said processor.
129 KNL Is a kernel start-up parameter.
130 BOOT Is a boot loader parameter.
132 Parameters denoted with BOOT are actually interpreted by the boot
133 loader, and have no meaning to the kernel directly.
134 Do not modify the syntax of boot loader parameters without extreme
135 need or coordination with <Documentation/x86/boot.txt>.
137 There are also arch-specific kernel-parameters not documented here.
138 See for example <Documentation/x86/x86_64/boot-options.txt>.
140 Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
141 a trailing = on the name of any parameter states that that parameter will
142 be entered as an environment variable, whereas its absence indicates that
143 it will appear as a kernel argument readable via /proc/cmdline by programs
144 running once the system is up.
146 The number of kernel parameters is not limited, but the length of the
147 complete command line (parameters including spaces etc.) is limited to
148 a fixed number of characters. This limit depends on the architecture
149 and is between 256 and 4096 characters. It is defined in the file
150 ./include/asm/setup.h as COMMAND_LINE_SIZE.
152 Finally, the [KMG] suffix is commonly described after a number of kernel
153 parameter values. These 'K', 'M', and 'G' letters represent the _binary_
154 multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
155 bytes respectively. Such letter suffixes can also be entirely omitted.
159 Advanced Configuration and Power Interface
160 Format: { force | off | strict | noirq | rsdt }
161 force -- enable ACPI if default was off
162 off -- disable ACPI if default was on
163 noirq -- do not use ACPI for IRQ routing
164 strict -- Be less tolerant of platforms that are not
165 strictly ACPI specification compliant.
166 rsdt -- prefer RSDT over (default) XSDT
167 copy_dsdt -- copy DSDT to memory
169 See also Documentation/power/runtime_pm.txt, pci=noacpi
171 acpi_rsdp= [ACPI,EFI,KEXEC]
172 Pass the RSDP address to the kernel, mostly used
173 on machines running EFI runtime service to boot the
174 second kernel for kdump.
176 acpi_apic_instance= [ACPI, IOAPIC]
178 2: use 2nd APIC table, if available
179 1,0: use 1st APIC table
182 acpi_backlight= [HW,ACPI]
183 acpi_backlight=vendor
185 If set to vendor, prefer vendor specific driver
186 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
187 of the ACPI video.ko driver.
189 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
190 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
192 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
193 debug output. Bits in debug_layer correspond to a
194 _COMPONENT in an ACPI source file, e.g.,
195 #define _COMPONENT ACPI_PCI_COMPONENT
196 Bits in debug_level correspond to a level in
197 ACPI_DEBUG_PRINT statements, e.g.,
198 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
199 The debug_level mask defaults to "info". See
200 Documentation/acpi/debug.txt for more information about
201 debug layers and levels.
203 Enable processor driver info messages:
204 acpi.debug_layer=0x20000000
205 Enable PCI/PCI interrupt routing info messages:
206 acpi.debug_layer=0x400000
207 Enable AML "Debug" output, i.e., stores to the Debug
208 object while interpreting AML:
209 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
210 Enable all messages related to ACPI hardware:
211 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
213 Some values produce so much output that the system is
214 unusable. The "log_buf_len" parameter may be useful
215 if you need to capture more output.
217 acpi_irq_balance [HW,ACPI]
218 ACPI will balance active IRQs
221 acpi_irq_nobalance [HW,ACPI]
222 ACPI will not move active IRQs (default)
225 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
226 Format: <irq>,<irq>...
228 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
230 Format: <irq>,<irq>...
232 acpi_no_auto_ssdt [HW,ACPI] Disable automatic loading of SSDT
234 acpica_no_return_repair [HW, ACPI]
235 Disable AML predefined validation mechanism
236 This mechanism can repair the evaluation result to make
237 the return objects more ACPI specification compliant.
238 This option is useful for developers to identify the
239 root cause of an AML interpreter issue when the issue
240 has something to do with the repair mechanism.
242 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
243 Format: To spoof as Windows 98: ="Microsoft Windows"
245 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
246 acpi_osi="string1" # add string1
247 acpi_osi="!string2" # remove string2
248 acpi_osi=!* # remove all strings
249 acpi_osi=! # disable all built-in OS vendor
251 acpi_osi= # disable all strings
253 'acpi_osi=!' can be used in combination with single or
254 multiple 'acpi_osi="string1"' to support specific OS
255 vendor string(s). Note that such command can only
256 affect the default state of the OS vendor strings, thus
257 it cannot affect the default state of the feature group
258 strings and the current state of the OS vendor strings,
259 specifying it multiple times through kernel command line
260 is meaningless. This command is useful when one do not
261 care about the state of the feature group strings which
262 should be controlled by the OSPM.
264 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
265 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
266 can make '_OSI("Windows 2000")' TRUE.
268 'acpi_osi=' cannot be used in combination with other
269 'acpi_osi=' command lines, the _OSI method will not
270 exist in the ACPI namespace. NOTE that such command can
271 only affect the _OSI support state, thus specifying it
272 multiple times through kernel command line is also
275 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
278 'acpi_osi=!*' can be used in combination with single or
279 multiple 'acpi_osi="string1"' to support specific
280 string(s). Note that such command can affect the
281 current state of both the OS vendor strings and the
282 feature group strings, thus specifying it multiple times
283 through kernel command line is meaningful. But it may
284 still not able to affect the final state of a string if
285 there are quirks related to this string. This command
286 is useful when one want to control the state of the
287 feature group strings to debug BIOS issues related to
290 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
291 '_OSI("Module Device")' FALSE.
292 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
293 '_OSI("Module Device")' TRUE.
294 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
296 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
298 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
299 they all will make '_OSI("Windows 2000")' TRUE.
302 Override the pmtimer bug detection: force the kernel
303 to assume that this machine's pmtimer latches its value
304 and always returns good values.
306 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
307 Format: { level | edge | high | low }
309 acpi_skip_timer_override [HW,ACPI]
310 Recognize and ignore IRQ0/pin2 Interrupt Override.
311 For broken nForce2 BIOS resulting in XT-PIC timer.
313 acpi_sleep= [HW,ACPI] Sleep options
314 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
315 old_ordering, nonvs, sci_force_enable }
316 See Documentation/power/video.txt for information on
318 s3_beep is for debugging; it makes the PC's speaker beep
319 as soon as the kernel's real-mode entry point is called.
320 s4_nohwsig prevents ACPI hardware signature from being
321 used during resume from hibernation.
322 old_ordering causes the ACPI 1.0 ordering of the _PTS
323 control method, with respect to putting devices into
324 low power states, to be enforced (the ACPI 2.0 ordering
325 of _PTS is used by default).
326 nonvs prevents the kernel from saving/restoring the
327 ACPI NVS memory during suspend/hibernation and resume.
328 sci_force_enable causes the kernel to set SCI_EN directly
329 on resume from S1/S3 (which is against the ACPI spec,
330 but some broken systems don't work without it).
332 acpi_use_timer_override [HW,ACPI]
333 Use timer override. For some broken Nvidia NF5 boards
334 that require a timer override, but don't have HPET
336 acpi_enforce_resources= [ACPI]
337 { strict | lax | no }
338 Check for resource conflicts between native drivers
339 and ACPI OperationRegions (SystemIO and SystemMemory
340 only). IO ports and memory declared in ACPI might be
341 used by the ACPI subsystem in arbitrary AML code and
342 can interfere with legacy drivers.
343 strict (default): access to resources claimed by ACPI
344 is denied; legacy drivers trying to access reserved
345 resources will fail to bind to device using them.
346 lax: access to resources claimed by ACPI is allowed;
347 legacy drivers trying to access reserved resources
348 will bind successfully but a warning message is logged.
349 no: ACPI OperationRegions are not marked as reserved,
350 no further checks are performed.
352 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
355 add_efi_memmap [EFI; X86] Include EFI memory map in
356 kernel's map of available physical RAM.
359 { off | try_unsupported }
360 off: disable AGP support
361 try_unsupported: try to drive unsupported chipsets
362 (may crash computer or cause data corruption)
365 See Documentation/sound/alsa/alsa-parameters.txt
368 Allow the default userspace alignment fault handler
369 behaviour to be specified. Bit 0 enables warnings,
370 bit 1 enables fixups, and bit 2 sends a segfault.
372 align_va_addr= [X86-64]
373 Align virtual addresses by clearing slice [14:12] when
374 allocating a VMA at process creation time. This option
375 gives you up to 3% performance improvement on AMD F15h
376 machines (where it is enabled by default) for a
377 CPU-intensive style benchmark, and it can vary highly in
378 a microbenchmark depending on workload and compiler.
380 32: only for 32-bit processes
381 64: only for 64-bit processes
382 on: enable for both 32- and 64-bit processes
383 off: disable for both 32- and 64-bit processes
385 alloc_snapshot [FTRACE]
386 Allocate the ftrace snapshot buffer on boot up when the
387 main buffer is allocated. This is handy if debugging
388 and you need to use tracing_snapshot() on boot up, and
389 do not want to use tracing_snapshot_alloc() as it needs
390 to be done where GFP_KERNEL allocations are allowed.
392 amd_iommu= [HW,X86-64]
393 Pass parameters to the AMD IOMMU driver in the system.
395 fullflush - enable flushing of IO/TLB entries when
396 they are unmapped. Otherwise they are
397 flushed before they will be reused, which
399 off - do not initialize any AMD IOMMU found in
401 force_isolation - Force device isolation for all
402 devices. The IOMMU driver is not
403 allowed anymore to lift isolation
404 requirements as needed. This option
405 does not override iommu=pt
407 amd_iommu_dump= [HW,X86-64]
408 Enable AMD IOMMU driver option to dump the ACPI table
409 for AMD IOMMU. With this option enabled, AMD IOMMU
410 driver will print ACPI tables for AMD IOMMU during
411 IOMMU initialization.
413 amijoy.map= [HW,JOY] Amiga joystick support
414 Map of devices attached to JOY0DAT and JOY1DAT
416 See also Documentation/input/joystick.txt
418 analog.map= [HW,JOY] Analog joystick and gamepad support
419 Specifies type or capabilities of an analog joystick
420 connected to one of 16 gameports
421 Format: <type1>,<type2>,..<type16>
424 Power management functions (SPARCstation-4/5 + deriv.)
426 Disable APC CPU standby support. SPARCstation-Fox does
427 not play well with APC CPU idle - disable it if you have
428 APC and your system crashes randomly.
430 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
431 Change the output verbosity whilst booting
432 Format: { quiet (default) | verbose | debug }
433 Change the amount of debugging information output
434 when initialising the APIC and IO-APIC components.
437 See Documentation/networking/ipv6.txt.
439 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
440 Limit apic dumping. The parameter defines the maximal
441 number of local apics being dumped. Also it is possible
442 to set it to "all" by meaning -- no limit here.
443 Format: { 1 (default) | 2 | ... | all }.
444 The parameter valid if only apic=debug or
445 apic=verbose is specified.
446 Example: apic=debug show_lapic=all
448 apm= [APM] Advanced Power Management
449 See header of arch/x86/kernel/apm_32.c.
451 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
452 Format: <io>,<irq>,<nodeID>
456 atarimouse= [HW,MOUSE] Atari Mouse
458 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
459 EzKey and similar keyboards
461 atkbd.reset= [HW] Reset keyboard during initialization
463 atkbd.set= [HW] Select keyboard code set
464 Format: <int> (2 = AT (default), 3 = PS/2)
466 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
469 atkbd.softraw= [HW] Choose between synthetic and real raw mode
470 Format: <bool> (0 = real, 1 = synthetic (default))
472 atkbd.softrepeat= [HW]
473 Use software keyboard repeat
475 audit= [KNL] Enable the audit sub-system
476 Format: { "0" | "1" } (0 = disabled, 1 = enabled)
477 0 - kernel audit is disabled and can not be enabled
478 until the next reboot
479 unset - kernel audit is initialized but disabled and
480 will be fully enabled by the userspace auditd.
481 1 - kernel audit is initialized and partially enabled,
482 storing at most audit_backlog_limit messages in
483 RAM until it is fully enabled by the userspace
487 audit_backlog_limit= [KNL] Set the audit queue size limit.
488 Format: <int> (must be >=0)
491 baycom_epp= [HW,AX25]
494 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
496 See header of drivers/net/hamradio/baycom_par.c.
498 baycom_ser_fdx= [HW,AX25]
499 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
500 Format: <io>,<irq>,<mode>[,<baud>]
501 See header of drivers/net/hamradio/baycom_ser_fdx.c.
503 baycom_ser_hdx= [HW,AX25]
504 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
505 Format: <io>,<irq>,<mode>
506 See header of drivers/net/hamradio/baycom_ser_hdx.c.
508 blkdevparts= Manual partition parsing of block device(s) for
509 embedded devices based on command line input.
510 See Documentation/block/cmdline-partition.txt
512 boot_delay= Milliseconds to delay each printk during boot.
513 Values larger than 10 seconds (10000) are changed to
517 bootmem_debug [KNL] Enable bootmem allocator debug messages.
519 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
520 bttv.radio= Most important insmod options are available as
522 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
525 bulk_remove=off [PPC] This parameter disables the use of the pSeries
526 firmware feature for flushing multiple hpte entries
529 c101= [NET] Moxa C101 synchronous serial card
531 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
532 Sometimes CPU hardware bugs make them report the cache
533 size incorrectly. The kernel will attempt work arounds
534 to fix known problems, but for some CPUs it is not
535 possible to determine what the correct size should be.
536 This option provides an override for these situations.
538 ccw_timeout_log [S390]
539 See Documentation/s390/CommonIO for details.
541 cgroup_disable= [KNL] Disable a particular controller
542 Format: {name of the controller(s) to disable}
543 The effects of cgroup_disable=foo are:
544 - foo isn't auto-mounted if you mount all cgroups in
546 - foo isn't visible as an individually mountable
548 {Currently only "memory" controller deal with this and
549 cut the overhead, others just disable the usage. So
550 only cgroup_disable=memory is actually worthy}
552 checkreqprot [SELINUX] Set initial checkreqprot flag value.
553 Format: { "0" | "1" }
554 See security/selinux/Kconfig help text.
555 0 -- check protection applied by kernel (includes
556 any implied execute protection).
557 1 -- check protection requested by application.
558 Default value is set via a kernel config option.
559 Value can be changed at runtime via
560 /selinux/checkreqprot.
563 See Documentation/s390/CommonIO for details.
566 Keep all clocks already enabled by bootloader on,
567 even if no driver has claimed them. This is useful
568 for debug and development, but should not be
569 needed on a platform with proper driver support.
570 For more information, see Documentation/clk.txt.
572 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
574 Forces specified clocksource (if available) to be used
575 when calculating gettimeofday(). If specified
576 clocksource is not available, it defaults to PIT.
577 Format: { pit | tsc | cyclone | pmtmr }
579 clocksource= Override the default clocksource
581 Override the default clocksource and use the clocksource
582 with the name specified.
583 Some clocksource names to choose from, depending on
585 [all] jiffies (this is the base, fallback clocksource)
587 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
588 pxa_timer,timer3,32k_counter,timer0_1
590 [X86-32] pit,hpet,tsc;
591 scx200_hrt on Geode; cyclone on IBM x440
599 clearcpuid=BITNUM [X86]
600 Disable CPUID feature X for the kernel. See
601 arch/x86/include/asm/cpufeature.h for the valid bit
602 numbers. Note the Linux specific bits are not necessarily
603 stable over kernel options, but the vendor specific
605 Also note that user programs calling CPUID directly
606 or using the feature without checking anything
607 will still see it. This just prevents it from
608 being used by the kernel or shown in /proc/cpuinfo.
609 Also note the kernel might malfunction if you disable
613 Sets the size of kernel global memory area for contiguous
614 memory allocations. For more information, see
615 include/linux/dma-contiguous.h
617 cmo_free_hint= [PPC] Format: { yes | no }
618 Specify whether pages are marked as being inactive
619 when they are freed. This is used in CMO environments
620 to determine OS memory pressure for page stealing by
624 coherent_pool=nn[KMG] [ARM,KNL]
625 Sets the size of memory pool for coherent, atomic dma
626 allocations, by default set to 256K.
628 code_bytes [X86] How many bytes of object code to print
633 com20020= [HW,NET] ARCnet - COM20020 chipset
635 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
637 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
641 ARCnet - COM90xx chipset (memory-mapped buffers)
642 Format: <io>[,<irq>[,<memstart>]]
644 condev= [HW,S390] console device
647 console= [KNL] Output console device and options.
649 tty<n> Use the virtual console device <n>.
653 Use the specified serial port. The options are of
654 the form "bbbbpnf", where "bbbb" is the baud rate,
655 "p" is parity ("n", "o", or "e"), "n" is number of
656 bits, and "f" is flow control ("r" for RTS or
657 omit it). Default is "9600n8".
659 See Documentation/serial-console.txt for more
661 Documentation/networking/netconsole.txt for an
664 uart[8250],io,<addr>[,options]
665 uart[8250],mmio,<addr>[,options]
666 Start an early, polled-mode console on the 8250/16550
667 UART at the specified I/O port or MMIO address,
668 switching to the matching ttyS device later. The
669 options are the same as for ttyS, above.
670 hvc<n> Use the hypervisor console device <n>. This is for
671 both Xen and PowerPC hypervisors.
673 If the device connected to the port is not a TTY but a braille
674 device, prepend "brl," before the device type, for instance
676 For now, only VisioBraille is supported.
678 consoleblank= [KNL] The console blank (screen saver) timeout in
679 seconds. Defaults to 10*60 = 10mins. A value of 0
680 disables the blank timer.
683 [KNL] Change the default value for
684 /proc/<pid>/coredump_filter.
685 See also Documentation/filesystems/proc.txt.
687 cpuidle.off=1 [CPU_IDLE]
688 disable the cpuidle sub-system
690 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
692 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
694 crashkernel=size[KMG][@offset[KMG]]
695 [KNL] Using kexec, Linux can switch to a 'crash kernel'
696 upon panic. This parameter reserves the physical
697 memory region [offset, offset + size] for that kernel
698 image. If '@offset' is omitted, then a suitable offset
699 is selected automatically. Check
700 Documentation/kdump/kdump.txt for further details.
702 crashkernel=range1:size1[,range2:size2,...][@offset]
703 [KNL] Same as above, but depends on the memory
704 in the running system. The syntax of range is
705 start-[end] where start and end are both
706 a memory unit (amount[KMG]). See also
707 Documentation/kdump/kdump.txt for an example.
709 crashkernel=size[KMG],high
710 [KNL, x86_64] range could be above 4G. Allow kernel
711 to allocate physical memory region from top, so could
712 be above 4G if system have more than 4G ram installed.
713 Otherwise memory region will be allocated below 4G, if
715 It will be ignored if crashkernel=X is specified.
716 crashkernel=size[KMG],low
717 [KNL, x86_64] range under 4G. When crashkernel=X,high
718 is passed, kernel could allocate physical memory region
719 above 4G, that cause second kernel crash on system
720 that require some amount of low memory, e.g. swiotlb
721 requires at least 64M+32K low memory. Kernel would
722 try to allocate 72M below 4G automatically.
723 This one let user to specify own low range under 4G
724 for second kernel instead.
725 0: to disable low allocation.
726 It will be ignored when crashkernel=X,high is not used
727 or memory reserved is below 4G.
732 cs89x0_media= [HW,NET]
733 Format: { rj45 | aui | bnc }
736 See header of drivers/s390/block/dasd_devmap.c.
738 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
739 (one device per port)
740 Format: <port#>,<type>
741 See also Documentation/input/joystick-parport.txt
743 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
744 time. See Documentation/dynamic-debug-howto.txt for
745 details. Deprecated, see dyndbg.
747 debug [KNL] Enable kernel debugging (events log level).
750 [KNL] verbose self-tests
752 Print debugging info while doing the locking API
754 We default to 0 (no extra messages), setting it to
755 1 will print _a lot_ more information - normally
756 only useful to kernel developers.
758 debug_objects [KNL] Enable object debugging
761 [KNL] Disable object debugging
763 debug_guardpage_minorder=
764 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
765 parameter allows control of the order of pages that will
766 be intentionally kept free (and hence protected) by the
767 buddy allocator. Bigger value increase the probability
768 of catching random memory corruption, but reduce the
769 amount of memory for normal system use. The maximum
770 possible value is MAX_ORDER/2. Setting this parameter
771 to 1 or 2 should be enough to identify most random
772 memory corruption problems caused by bugs in kernel or
773 driver code when a CPU writes to (or reads from) a
774 random memory location. Note that there exists a class
775 of memory corruptions problems caused by buggy H/W or
776 F/W or by drivers badly programing DMA (basically when
777 memory is written at bus level and the CPU MMU is
778 bypassed) which are not detectable by
779 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
780 tracking down these problems.
782 debugpat [X86] Enable PAT debugging
784 decnet.addr= [HW,NET]
785 Format: <area>[,<node>]
786 See also Documentation/networking/decnet.txt.
789 [same as hugepagesz=] The size of the default
790 HugeTLB page size. This is the size represented by
791 the legacy /proc/ hugepages APIs, used for SHM, and
792 default size when mounting hugetlbfs filesystems.
793 Defaults to the default architecture's huge page size
797 Set number of hash buckets for dentry cache.
800 IO parameters + enable/disable command.
802 digiepca= [HW,SERIAL]
803 See drivers/char/README.epca and
804 Documentation/serial/digiepca.txt.
807 See Documentation/networking/ipv6.txt.
809 disable_cpu_apicid= [X86,APIC,SMP]
811 The number of initial APIC ID for the
812 corresponding CPU to be disabled at boot,
813 mostly used for the kdump 2nd kernel to
814 disable BSP to wake up multiple CPUs without
815 causing system reset or hang due to sending
818 disable_ddw [PPC/PSERIES]
819 Disable Dynamic DMA Window support. Use this if
820 to workaround buggy firmware.
823 See Documentation/networking/ipv6.txt.
825 disable_mtrr_cleanup [X86]
826 The kernel tries to adjust MTRR layout from continuous
827 to discrete, to make X server driver able to add WB
828 entry later. This parameter disables that.
830 disable_mtrr_trim [X86, Intel and AMD only]
831 By default the kernel will trim any uncacheable
832 memory out of your available memory pool based on
833 MTRR settings. This parameter disables that behavior,
834 possibly causing your machine to run very slowly.
836 disable_timer_pin_1 [X86]
837 Disable PIN 1 of APIC timer
838 Can be useful to work around chipset bugs.
840 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
841 this option disables the debugging code at boot.
843 dma_debug_entries=<number>
844 This option allows to tune the number of preallocated
845 entries for DMA-API debugging code. One entry is
846 required per DMA-API allocation. Use this if the
847 DMA-API debugging code disables itself because the
848 architectural default is too low.
850 dma_debug_driver=<driver_name>
851 With this option the DMA-API debugging driver
852 filter feature can be enabled at boot time. Just
853 pass the driver to filter for as the parameter.
854 The filter can be disabled or changed to another
855 driver later using sysfs.
857 drm_kms_helper.edid_firmware=[<connector>:]<file>
858 Broken monitors, graphic adapters and KVMs may
859 send no or incorrect EDID data sets. This parameter
860 allows to specify an EDID data set in the
861 /lib/firmware directory that is used instead.
862 Generic built-in EDID data sets are used, if one of
863 edid/1024x768.bin, edid/1280x1024.bin,
864 edid/1680x1050.bin, or edid/1920x1080.bin is given
865 and no file with the same name exists. Details and
866 instructions how to build your own EDID data are
867 available in Documentation/EDID/HOWTO.txt. An EDID
868 data set will only be used for a particular connector,
869 if its name and a colon are prepended to the EDID
874 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
875 module.dyndbg[="val"]
876 Enable debug messages at boot time. See
877 Documentation/dynamic-debug-howto.txt for details.
879 earlycon= [KNL] Output early console device and options.
880 uart[8250],io,<addr>[,options]
881 uart[8250],mmio,<addr>[,options]
882 uart[8250],mmio32,<addr>[,options]
883 Start an early, polled-mode console on the 8250/16550
884 UART at the specified I/O port or MMIO address.
885 MMIO inter-register address stride is either 8-bit
886 (mmio) or 32-bit (mmio32).
887 The options are the same as for ttyS, above.
889 earlyprintk= [X86,SH,BLACKFIN,ARM]
893 earlyprintk=serial[,ttySn[,baudrate]]
894 earlyprintk=serial[,0x...[,baudrate]]
895 earlyprintk=ttySn[,baudrate]
896 earlyprintk=dbgp[debugController#]
898 earlyprintk is useful when the kernel crashes before
899 the normal console is initialized. It is not enabled by
900 default because it has some cosmetic problems.
902 Append ",keep" to not disable it when the real console
905 Only one of vga, efi, serial, or usb debug port can
908 Currently only ttyS0 and ttyS1 may be specified by
909 name. Other I/O ports may be explicitly specified
910 on some architectures (x86 and arm at least) by
911 replacing ttySn with an I/O port address, like this:
912 earlyprintk=serial,0x1008,115200
913 You can find the port for a given device in
914 /proc/tty/driver/serial:
915 2: uart:ST16650V2 port:00001008 irq:18 ...
917 Interaction with the standard serial driver is not
920 The VGA and EFI output is eventually overwritten by
923 The xen output can only be used by Xen PV guests.
925 edac_report= [HW,EDAC] Control how to report EDAC event
926 Format: {"on" | "off" | "force"}
927 on: enable EDAC to report H/W event. May be overridden
928 by other higher priority error reporting module.
929 off: disable H/W event reporting through EDAC.
930 force: enforce the use of EDAC to report H/W event.
933 ekgdboc= [X86,KGDB] Allow early kernel console debugging
936 This is designed to be used in conjunction with
937 the boot argument: earlyprintk=vga
940 Format: {"off" | "on" | "skip[mbr]"}
943 Format: { "old_map" }
944 old_map [X86-64]: switch to the old ioremap-based EFI
945 runtime services mapping. 32-bit still uses this one by
948 efi_no_storage_paranoia [EFI; X86]
949 Using this parameter you can use more than 50% of
950 your efi variable storage. Use this parameter only if
951 you are really sure that your UEFI does sane gc and
952 fulfills the spec otherwise your board may brick.
954 eisa_irq_edge= [PARISC,HW]
955 See header of drivers/parisc/eisa.c.
958 See comment before function elanfreq_setup() in
959 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
962 Format: {"cfq" | "deadline" | "noop"}
963 See Documentation/block/cfq-iosched.txt and
964 Documentation/block/deadline-iosched.txt for details.
966 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
967 Specifies physical address of start of kernel core
968 image elf header and optionally the size. Generally
969 kexec loader will pass this option to capture kernel.
970 See Documentation/kdump/kdump.txt for details.
972 enable_mtrr_cleanup [X86]
973 The kernel tries to adjust MTRR layout from continuous
974 to discrete, to make X server driver able to add WB
975 entry later. This parameter enables that.
977 enable_timer_pin_1 [X86]
978 Enable PIN 1 of APIC timer
979 Can be useful to work around chipset bugs
980 (in particular on some ATI chipsets).
981 The kernel tries to set a reasonable default.
983 enforcing [SELINUX] Set initial enforcing status.
985 See security/selinux/Kconfig help text.
986 0 -- permissive (log only, no denials).
987 1 -- enforcing (deny and log).
989 Value can be changed at runtime via /selinux/enforce.
992 Disable Error Record Serialization Table (ERST)
995 ether= [HW,NET] Ethernet cards parameters
996 This option is obsoleted by the "netdev=" option, which
997 has equivalent usage. See its documentation for details.
1001 Permit 'security.evm' to be updated regardless of
1002 current integrity status.
1006 fail_make_request=[KNL]
1007 General fault injection mechanism.
1008 Format: <interval>,<probability>,<space>,<times>
1009 See also Documentation/fault-injection/.
1012 See Documentation/blockdev/floppy.txt.
1014 force_pal_cache_flush
1015 [IA-64] Avoid check_sal_cache_flush which may hang on
1016 buggy SAL_CACHE_FLUSH implementations. Using this
1017 parameter will force ia64_sal_cache_flush to call
1018 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1021 [FTRACE] will set and start the specified tracer
1022 as early as possible in order to facilitate early
1025 ftrace_dump_on_oops[=orig_cpu]
1026 [FTRACE] will dump the trace buffers on oops.
1027 If no parameter is passed, ftrace will dump
1028 buffers of all CPUs, but if you pass orig_cpu, it will
1029 dump only the buffer of the CPU that triggered the
1032 ftrace_filter=[function-list]
1033 [FTRACE] Limit the functions traced by the function
1034 tracer at boot up. function-list is a comma separated
1035 list of functions. This list can be changed at run
1036 time by the set_ftrace_filter file in the debugfs
1039 ftrace_notrace=[function-list]
1040 [FTRACE] Do not trace the functions specified in
1041 function-list. This list can be changed at run time
1042 by the set_ftrace_notrace file in the debugfs
1045 ftrace_graph_filter=[function-list]
1046 [FTRACE] Limit the top level callers functions traced
1047 by the function graph tracer at boot up.
1048 function-list is a comma separated list of functions
1049 that can be changed at run time by the
1050 set_graph_function file in the debugfs tracing directory.
1053 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1054 support via parallel port (up to 5 devices per port)
1055 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1056 See also Documentation/input/joystick-parport.txt
1060 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1064 gcov_persist= [GCOV] When non-zero (default), profiling data for
1065 kernel modules is saved and remains accessible via
1066 debugfs, even when the module is unloaded/reloaded.
1067 When zero, profiling data is discarded and associated
1068 debugfs files are removed at module unload time.
1070 gpt [EFI] Forces disk with valid GPT signature but
1071 invalid Protective MBR to be treated as GPT. If the
1072 primary GPT is corrupted, it enables the backup/alternate
1073 GPT to be used instead.
1075 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1076 the "Enable 0" bit of the configuration register.
1079 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1080 the "Enable 0" bit of the configuration register.
1083 grcan.select= [HW] Select which physical interface to use.
1086 grcan.txsize= [HW] Sets the size of the tx buffer.
1087 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1089 grcan.rxsize= [HW] Sets the size of the rx buffer.
1090 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1093 hashdist= [KNL,NUMA] Large hashes allocated during boot
1094 are distributed across NUMA nodes. Defaults on
1095 for 64-bit NUMA, off otherwise.
1096 Format: 0 | 1 (for off | on)
1098 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1100 hd= [EIDE] (E)IDE hard drive subsystem geometry
1101 Format: <cyl>,<head>,<sect>
1104 Disable Hardware Error Source Table (HEST) support;
1105 corresponding firmware-first mode error processing
1106 logic will be disabled.
1108 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1109 size of <nn>. This works even on boxes that have no
1110 highmem otherwise. This also works to reduce highmem
1111 size on bigger boxes.
1113 highres= [KNL] Enable/disable high resolution timer mode.
1114 Valid parameters: "on", "off"
1118 See Documentation/isdn/README.HiSax.
1122 hpet= [X86-32,HPET] option to control HPET usage
1123 Format: { enable (default) | disable | force |
1125 disable: disable HPET and use PIT instead
1126 force: allow force enabled of undocumented chips (ICH4,
1128 verbose: show contents of HPET registers during setup
1130 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1131 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1133 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1134 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1135 On x86-64 and powerpc, this option can be specified
1136 multiple times interleaved with hugepages= to reserve
1137 huge pages of different sizes. Valid pages sizes on
1138 x86-64 are 2M (when the CPU supports "pse") and 1G
1139 (when the CPU supports the "pdpe1gb" cpuinfo flag)
1140 Note that 1GB pages can only be allocated at boot time
1141 using hugepages= and not freed afterwards.
1143 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1144 terminal devices. Valid values: 0..8
1145 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1146 If specified, z/VM IUCV HVC accepts connections
1147 from listed z/VM user IDs only.
1149 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1150 hardware thread id mappings.
1151 Format: <cpu>:<hwthread>
1154 Do not unregister boot console at start. This is only
1155 useful for debugging when something happens in the window
1156 between unregistering the boot console and initializing
1159 i2c_bus= [HW] Override the default board specific I2C bus speed
1160 or register an additional I2C bus that is not
1161 registered from board initialization code.
1165 i8042.debug [HW] Toggle i8042 debug mode
1166 i8042.direct [HW] Put keyboard port into non-translated mode
1167 i8042.dumbkbd [HW] Pretend that controller can only read data from
1168 keyboard and cannot control its state
1169 (Don't attempt to blink the leds)
1170 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1171 i8042.nokbd [HW] Don't check/create keyboard port
1172 i8042.noloop [HW] Disable the AUX Loopback command while probing
1174 i8042.nomux [HW] Don't check presence of an active multiplexing
1176 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1178 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1179 i8042.reset [HW] Reset the controller during init and cleanup
1180 i8042.unlock [HW] Unlock (ignore) the keylock
1184 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1185 indicates that the driver is running on unsupported
1187 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1188 does not match list of supported models.
1190 [HW] Report power status in /proc/i8k
1191 (disabled by default)
1192 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1195 i915.invert_brightness=
1196 [DRM] Invert the sense of the variable that is used to
1197 set the brightness of the panel backlight. Normally a
1198 brightness value of 0 indicates backlight switched off,
1199 and the maximum of the brightness value sets the backlight
1200 to maximum brightness. If this parameter is set to 0
1201 (default) and the machine requires it, or this parameter
1202 is set to 1, a brightness value of 0 sets the backlight
1203 to maximum brightness, and the maximum of the brightness
1204 value switches the backlight off.
1205 -1 -- never invert brightness
1206 0 -- machine default
1207 1 -- force brightness inversion
1210 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1212 ide-core.nodma= [HW] (E)IDE subsystem
1213 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1214 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1215 .cdrom .chs .ignore_cable are additional options
1216 See Documentation/ide/ide.txt.
1218 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1219 Claim all unknown PCI IDE storage controllers.
1222 Format: idle=poll, idle=halt, idle=nomwait
1223 Poll forces a polling idle loop that can slightly
1224 improve the performance of waking up a idle CPU, but
1225 will use a lot of power and make the system run hot.
1227 idle=halt: Halt is forced to be used for CPU idle.
1228 In such case C2/C3 won't be used again.
1229 idle=nomwait: Disable mwait for CPU C-states
1231 ignore_loglevel [KNL]
1232 Ignore loglevel setting - this will print /all/
1233 kernel messages to the console. Useful for debugging.
1234 We also add it as printk module parameter, so users
1235 could change it dynamically, usually by
1236 /sys/module/printk/parameters/ignore_loglevel.
1238 ihash_entries= [KNL]
1239 Set number of hash buckets for inode cache.
1241 ima_appraise= [IMA] appraise integrity measurements
1242 Format: { "off" | "enforce" | "fix" }
1245 ima_appraise_tcb [IMA]
1246 The builtin appraise policy appraises all files
1250 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1254 The list of supported hash algorithms is defined
1255 in crypto/hash_info.h.
1258 Load a policy which meets the needs of the Trusted
1259 Computing Base. This means IMA will measure all
1260 programs exec'd, files mmap'd for exec, and all files
1261 opened for read by uid=0.
1264 Select one of defined IMA measurements template formats.
1265 Formats: { "ima" | "ima-ng" }
1270 Run specified binary instead of /sbin/init as init
1273 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1274 for working out where the kernel is dying during
1277 initrd= [BOOT] Specify the location of the initial ramdisk
1279 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1282 int_pln_enable [x86] Enable power limit notification interrupt
1284 integrity_audit=[IMA]
1285 Format: { "0" | "1" }
1286 0 -- basic integrity auditing messages. (Default)
1287 1 -- additional integrity auditing messages.
1289 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1291 Enable intel iommu driver.
1293 Disable intel iommu driver.
1294 igfx_off [Default Off]
1295 By default, gfx is mapped as normal device. If a gfx
1296 device has a dedicated DMAR unit, the DMAR unit is
1297 bypassed by not enabling DMAR with this option. In
1298 this case, gfx device will use physical address for
1301 With this option iommu will not optimize to look
1302 for io virtual address below 32-bit forcing dual
1303 address cycle on pci bus for cards supporting greater
1304 than 32-bit addressing. The default is to look
1305 for translation below 32-bit and if not available
1306 then look in the higher range.
1307 strict [Default Off]
1308 With this option on every unmap_single operation will
1309 result in a hardware IOTLB flush operation as opposed
1310 to batching them for performance.
1311 sp_off [Default Off]
1312 By default, super page will be supported if Intel IOMMU
1313 has the capability. With this option, super page will
1316 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1317 0 disables intel_idle and fall back on acpi_idle.
1318 1 to 6 specify maximum depth of C-state.
1322 Do not enable intel_pstate as the default
1323 scaling driver for the supported processors
1325 intremap= [X86-64, Intel-IOMMU]
1326 on enable Interrupt Remapping (default)
1327 off disable Interrupt Remapping
1328 nosid disable Source ID checking
1330 BIOS x2APIC opt-out request will be ignored
1332 iomem= Disable strict checking of access to MMIO memory
1333 strict regions from userspace.
1350 io7= [HW] IO7 for Marvel based alpha systems
1351 See comment before marvel_specify_io7 in
1352 arch/alpha/kernel/core_marvel.c.
1354 io_delay= [X86] I/O delay method
1356 Standard port 0x80 based delay
1358 Alternate port 0xed based delay (needed on some systems)
1360 Simple two microseconds delay
1365 See Documentation/filesystems/nfs/nfsroot.txt.
1367 ip2= [HW] Set IO/IRQ pairs for up to 4 IntelliPort boards
1368 See comment before ip2_setup() in
1369 drivers/char/ip2/ip2base.c.
1372 When an interrupt is not handled search all handlers
1373 for it. Intended to get systems with badly broken
1377 When an interrupt is not handled search all handlers
1378 for it. Also check all handlers each timer
1379 interrupt. Intended to get systems with badly broken
1383 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1385 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1387 <cpu number>,...,<cpu number>
1389 <cpu number>-<cpu number>
1390 (must be a positive range in ascending order)
1392 <cpu number>,...,<cpu number>-<cpu number>
1394 This option can be used to specify one or more CPUs
1395 to isolate from the general SMP balancing and scheduling
1396 algorithms. You can move a process onto or off an
1397 "isolated" CPU via the CPU affinity syscalls or cpuset.
1398 <cpu number> begins at 0 and the maximum value is
1399 "number of CPUs in system - 1".
1401 This option is the preferred way to isolate CPUs. The
1402 alternative -- manually setting the CPU mask of all
1403 tasks in the system -- can cause problems and
1404 suboptimal load balancer performance.
1408 ivrs_ioapic [HW,X86_64]
1409 Provide an override to the IOAPIC-ID<->DEVICE-ID
1410 mapping provided in the IVRS ACPI table. For
1411 example, to map IOAPIC-ID decimal 10 to
1412 PCI device 00:14.0 write the parameter as:
1413 ivrs_ioapic[10]=00:14.0
1415 ivrs_hpet [HW,X86_64]
1416 Provide an override to the HPET-ID<->DEVICE-ID
1417 mapping provided in the IVRS ACPI table. For
1418 example, to map HPET-ID decimal 0 to
1419 PCI device 00:14.0 write the parameter as:
1420 ivrs_hpet[0]=00:14.0
1422 js= [HW,JOY] Analog joystick
1423 See Documentation/input/joystick.txt.
1427 kernelcore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1428 specifies the amount of memory usable by the kernel
1429 for non-movable allocations. The requested amount is
1430 spread evenly throughout all nodes in the system. The
1431 remaining memory in each node is used for Movable
1432 pages. In the event, a node is too small to have both
1433 kernelcore and Movable pages, kernelcore pages will
1434 take priority and other nodes will have a larger number
1435 of Movable pages. The Movable zone is used for the
1436 allocation of pages that may be reclaimed or moved
1437 by the page migration subsystem. This means that
1438 HugeTLB pages may not be allocated from this zone.
1439 Note that allocations like PTEs-from-HighMem still
1440 use the HighMem zone if it exists, and the Normal
1441 zone if it does not.
1443 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1444 Format: <Controller#>[,poll interval]
1445 The controller # is the number of the ehci usb debug
1446 port as it is probed via PCI. The poll interval is
1447 optional and is the number seconds in between
1448 each poll cycle to the debug port in case you need
1449 the functionality for interrupting the kernel with
1450 gdb or control-c on the dbgp connection. When
1451 not using this parameter you use sysrq-g to break into
1452 the kernel debugger.
1454 kgdboc= [KGDB,HW] kgdb over consoles.
1455 Requires a tty driver that supports console polling,
1456 or a supported polling keyboard driver (non-usb).
1457 Serial only format: <serial_device>[,baud]
1458 keyboard only format: kbd
1459 keyboard and serial format: kbd,<serial_device>[,baud]
1460 Optional Kernel mode setting:
1461 kms, kbd format: kms,kbd
1462 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1464 kgdbwait [KGDB] Stop kernel execution and enter the
1465 kernel debugger at the earliest opportunity.
1467 kmac= [MIPS] korina ethernet MAC address.
1468 Configure the RouterBoard 532 series on-chip
1469 Ethernet adapter MAC address.
1471 kmemleak= [KNL] Boot-time kmemleak enable/disable
1472 Valid arguments: on, off
1475 kmemcheck= [X86] Boot-time kmemcheck enable/disable/one-shot mode
1476 Valid arguments: 0, 1, 2
1477 kmemcheck=0 (disabled)
1478 kmemcheck=1 (enabled)
1479 kmemcheck=2 (one-shot mode)
1480 Default: 2 (one-shot mode)
1482 kstack=N [X86] Print N words from the kernel stack
1485 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1486 Default is 0 (don't ignore, but inject #GP)
1488 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1492 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1493 Default is 1 (enabled)
1495 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1497 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1499 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1500 (virtualized MMU) support on capable Intel chips.
1501 Default is 1 (enabled)
1503 kvm-intel.emulate_invalid_guest_state=
1504 [KVM,Intel] Enable emulation of invalid guest states
1505 Default is 0 (disabled)
1507 kvm-intel.flexpriority=
1508 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1509 Default is 1 (enabled)
1512 [KVM,Intel] Enable VMX nesting (nVMX).
1513 Default is 0 (disabled)
1515 kvm-intel.unrestricted_guest=
1516 [KVM,Intel] Disable unrestricted guest feature
1517 (virtualized real and unpaged mode) on capable
1518 Intel chips. Default is 1 (enabled)
1520 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1521 feature (tagged TLBs) on capable Intel chips.
1522 Default is 1 (enabled)
1528 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1531 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1532 value for LAPIC timer one-shot implementation. Default
1533 back to the programmable timer unit in the LAPIC.
1535 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1538 libata.dma= [LIBATA] DMA control
1539 libata.dma=0 Disable all PATA and SATA DMA
1540 libata.dma=1 PATA and SATA Disk DMA only
1541 libata.dma=2 ATAPI (CDROM) DMA only
1542 libata.dma=4 Compact Flash DMA only
1543 Combinations also work, so libata.dma=3 enables DMA
1544 for disks and CDROMs, but not CFs.
1546 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1547 libata.ignore_hpa=0 keep BIOS limits (default)
1548 libata.ignore_hpa=1 ignore limits, using full disk
1550 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
1554 libata.force= [LIBATA] Force configurations. The format is comma
1555 separated list of "[ID:]VAL" where ID is
1556 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
1557 matching port, link or device. Basically, it matches
1558 the ATA ID string printed on console by libata. If
1559 the whole ID part is omitted, the last PORT and DEVICE
1560 values are used. If ID hasn't been specified yet, the
1561 configuration applies to all ports, links and devices.
1563 If only DEVICE is omitted, the parameter applies to
1564 the port and all links and devices behind it. DEVICE
1565 number of 0 either selects the first device or the
1566 first fan-out link behind PMP device. It does not
1567 select the host link. DEVICE number of 15 selects the
1568 host link and device attached to it.
1570 The VAL specifies the configuration to force. As long
1571 as there's no ambiguity shortcut notation is allowed.
1572 For example, both 1.5 and 1.5G would work for 1.5Gbps.
1573 The following configurations can be forced.
1575 * Cable type: 40c, 80c, short40c, unk, ign or sata.
1576 Any ID with matching PORT is used.
1578 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
1580 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
1581 udma[/][16,25,33,44,66,100,133] notation is also
1584 * [no]ncq: Turn on or off NCQ.
1586 * nohrst, nosrst, norst: suppress hard, soft
1589 * rstonce: only attempt one reset during
1590 hot-unplug link recovery
1592 * dump_id: dump IDENTIFY data.
1594 * atapi_dmadir: Enable ATAPI DMADIR bridge support
1596 * disable: Disable this device.
1598 If there are multiple matching configurations changing
1599 the same attribute, the last one is used.
1601 memblock=debug [KNL] Enable memblock debug messages.
1603 load_ramdisk= [RAM] List of ramdisks to load from floppy
1604 See Documentation/blockdev/ramdisk.txt.
1606 lockd.nlm_grace_period=P [NFS] Assign grace period.
1609 lockd.nlm_tcpport=N [NFS] Assign TCP port.
1612 lockd.nlm_timeout=T [NFS] Assign timeout value.
1615 lockd.nlm_udpport=M [NFS] Assign UDP port.
1618 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
1621 loglevel= All Kernel Messages with a loglevel smaller than the
1622 console loglevel will be printed to the console. It can
1623 also be changed with klogd or other programs. The
1624 loglevels are defined as follows:
1626 0 (KERN_EMERG) system is unusable
1627 1 (KERN_ALERT) action must be taken immediately
1628 2 (KERN_CRIT) critical conditions
1629 3 (KERN_ERR) error conditions
1630 4 (KERN_WARNING) warning conditions
1631 5 (KERN_NOTICE) normal but significant condition
1632 6 (KERN_INFO) informational
1633 7 (KERN_DEBUG) debug-level messages
1635 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
1636 in bytes. n must be a power of two. The default
1637 size is set in the kernel config file.
1639 logo.nologo [FB] Disables display of the built-in Linux logo.
1640 This may be used to provide more screen space for
1641 kernel log messages and is useful when debugging
1642 kernel boot problems.
1644 lp=0 [LP] Specify parallel ports to use, e.g,
1645 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
1646 lp=reset first parallel port). 'lp=0' disables the
1647 lp=auto printer driver. 'lp=reset' (which can be
1648 specified in addition to the ports) causes
1649 attached printers to be reset. Using
1650 lp=port1,port2,... specifies the parallel ports
1651 to associate lp devices with, starting with
1652 lp0. A port specification may be 'none' to skip
1653 that lp device, or a parport name such as
1654 'parport0'. Specifying 'lp=auto' instead of a
1655 port specification list means that device IDs
1656 from each port should be examined, to see if
1657 an IEEE 1284-compliant printer is attached; if
1658 so, the driver will manage that printer.
1659 See also header of drivers/char/lp.c.
1662 Sets loops_per_jiffy to given constant, thus avoiding
1663 time-consuming boot-time autodetection (up to 250 ms per
1664 CPU). 0 enables autodetection (default). To determine
1665 the correct value for your kernel, boot with normal
1666 autodetection and see what value is printed. Note that
1667 on SMP systems the preset will be applied to all CPUs,
1668 which is likely to cause problems if your CPUs need
1669 significantly divergent settings. An incorrect value
1670 will cause delays in the kernel to be wrong, leading to
1671 unpredictable I/O errors and other breakage. Although
1672 unlikely, in the extreme case this might damage your
1676 Format: <io>,<irq>,<dma>
1678 machvec= [IA-64] Force the use of a particular machine-vector
1679 (machvec) in a generic kernel.
1680 Example: machvec=hpzx1_swiotlb
1682 machtype= [Loongson] Share the same kernel image file between different
1684 Example: machtype=lemote-yeeloong-2f-7inch
1686 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
1687 than or equal to this physical address is ignored.
1689 maxcpus= [SMP] Maximum number of processors that an SMP kernel
1690 should make use of. maxcpus=n : n >= 0 limits the
1691 kernel to using 'n' processors. n=0 is a special case,
1692 it is equivalent to "nosmp", which also disables
1695 max_loop= [LOOP] The number of loop block devices that get
1696 (loop.max_loop) unconditionally pre-created at init time. The default
1697 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
1698 of statically allocating a predefined number, loop
1699 devices can be requested on-demand with the
1700 /dev/loop-control interface.
1702 mce [X86-32] Machine Check Exception
1704 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
1706 md= [HW] RAID subsystems devices and level
1707 See Documentation/md.txt.
1710 Format: <first>,<last>
1711 Specifies range of consoles to be captured by the MDA.
1713 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
1714 Amount of memory to be used when the kernel is not able
1715 to see the whole system memory or for test.
1716 [X86] Work as limiting max address. Use together
1717 with memmap= to avoid physical address space collisions.
1718 Without memmap= PCI devices could be placed at addresses
1719 belonging to unused RAM.
1721 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
1725 [KNL,SH] Allow user to override the default size for
1726 per-device physically contiguous DMA buffers.
1728 memmap=exactmap [KNL,X86] Enable setting of an exact
1729 E820 memory map, as specified by the user.
1730 Such memmap=exactmap lines can be constructed based on
1731 BIOS output or other requirements. See the memmap=nn@ss
1734 memmap=nn[KMG]@ss[KMG]
1735 [KNL] Force usage of a specific region of memory.
1736 Region of memory to be used is from ss to ss+nn.
1738 memmap=nn[KMG]#ss[KMG]
1739 [KNL,ACPI] Mark specific memory as ACPI data.
1740 Region of memory to be marked is from ss to ss+nn.
1742 memmap=nn[KMG]$ss[KMG]
1743 [KNL,ACPI] Mark specific memory as reserved.
1744 Region of memory to be reserved is from ss to ss+nn.
1745 Example: Exclude memory from 0x18690000-0x1869ffff
1746 memmap=64K$0x18690000
1748 memmap=0x10000$0x18690000
1750 memory_corruption_check=0/1 [X86]
1751 Some BIOSes seem to corrupt the first 64k of
1752 memory when doing things like suspend/resume.
1753 Setting this option will scan the memory
1754 looking for corruption. Enabling this will
1755 both detect corruption and prevent the kernel
1756 from using the memory being corrupted.
1757 However, its intended as a diagnostic tool; if
1758 repeatable BIOS-originated corruption always
1759 affects the same memory, you can use memmap=
1760 to prevent the kernel from using that memory.
1762 memory_corruption_check_size=size [X86]
1763 By default it checks for corruption in the low
1764 64k, making this memory unavailable for normal
1765 use. Use this parameter to scan for
1766 corruption in more or less memory.
1768 memory_corruption_check_period=seconds [X86]
1769 By default it checks for corruption every 60
1770 seconds. Use this parameter to check at some
1771 other rate. 0 disables periodic checking.
1773 memtest= [KNL,X86] Enable memtest
1775 default : 0 <disable>
1776 Specifies the number of memtest passes to be
1777 performed. Each pass selects another test
1778 pattern from a given set of patterns. Memtest
1779 fills the memory with this pattern, validates
1780 memory contents and reserves bad memory
1781 regions that are detected.
1783 meye.*= [HW] Set MotionEye Camera parameters
1784 See Documentation/video4linux/meye.txt.
1786 mfgpt_irq= [IA-32] Specify the IRQ to use for the
1787 Multi-Function General Purpose Timers on AMD Geode
1790 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
1791 the BIOS has incorrectly applied a workaround. TinyBIOS
1792 version 0.98 is known to be affected, 0.99 fixes the
1793 problem by letting the user disable the workaround.
1797 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
1798 physical address is ignored.
1800 mini2440= [ARM,HW,KNL]
1801 Format:[0..2][b][c][t]
1803 MINI2440 configuration specification:
1804 0 - The attached screen is the 3.5" TFT
1805 1 - The attached screen is the 7" TFT
1806 2 - The VGA Shield is attached (1024x768)
1807 Leaving out the screen size parameter will not load
1808 the TFT driver, and the framebuffer will be left
1810 b - Enable backlight. The TFT backlight pin will be
1811 linked to the kernel VESA blanking code and a GPIO
1812 LED. This parameter is not necessary when using the
1814 c - Enable the s3c camera interface.
1815 t - Reserved for enabling touchscreen support. The
1816 touchscreen support is not enabled in the mainstream
1817 kernel as of 2.6.30, a preliminary port can be found
1818 in the "bleeding edge" mini2440 support kernel at
1819 http://repo.or.cz/w/linux-2.6/mini2440.git
1822 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
1823 parameter allows control of the logging verbosity for
1824 the additional memory initialisation checks. A value
1825 of 0 disables mminit logging and a level of 4 will
1826 log everything. Information is printed at KERN_DEBUG
1827 so loglevel=8 may also need to be specified.
1830 [KNL] When CONFIG_MODULE_SIG is set, this means that
1831 modules without (valid) signatures will fail to load.
1832 Note that if CONFIG_MODULE_SIG_FORCE is set, that
1833 is always true, so this option does nothing.
1836 [MOUSE] Maximum time between finger touching and
1837 leaving touchpad surface for touch to be considered
1838 a tap and be reported as a left button click (for
1839 touchpads working in absolute mode only).
1841 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
1842 reporting absolute coordinates, such as tablets
1843 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
1844 reporting absolute coordinates, such as tablets
1846 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1847 is similar to kernelcore except it specifies the
1848 amount of memory used for migratable allocations.
1849 If both kernelcore and movablecore is specified,
1850 then kernelcore will be at *least* the specified
1851 value but may be more. If movablecore on its own
1852 is specified, the administrator must be careful
1853 that the amount of memory usable for all allocations
1856 movable_node [KNL,X86] Boot-time switch to enable the effects
1857 of CONFIG_MOVABLE_NODE=y. See mm/Kconfig for details.
1859 MTD_Partition= [MTD]
1860 Format: <name>,<region-number>,<size>,<offset>
1862 MTD_Region= [MTD] Format:
1863 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
1866 See drivers/mtd/cmdlinepart.c.
1868 multitce=off [PPC] This parameter disables the use of the pSeries
1869 firmware feature for updating multiple TCE entries
1872 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
1874 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
1876 boundary - index of last SLC block on Flex-OneNAND.
1877 The remaining blocks are configured as MLC blocks.
1878 lock - Configure if Flex-OneNAND boundary should be locked.
1879 Once locked, the boundary cannot be changed.
1880 1 indicates lock status, 0 indicates unlock status.
1883 ARM/S3C2412 JIVE boot control
1885 See arch/arm/mach-s3c2412/mach-jive.c
1887 mtouchusb.raw_coordinates=
1888 [HW] Make the MicroTouch USB driver use raw coordinates
1889 ('y', default) or cooked coordinates ('n')
1891 mtrr_chunk_size=nn[KMG] [X86]
1892 used for mtrr cleanup. It is largest continuous chunk
1893 that could hold holes aka. UC entries.
1895 mtrr_gran_size=nn[KMG] [X86]
1896 Used for mtrr cleanup. It is granularity of mtrr block.
1898 Large value could prevent small alignment from
1901 mtrr_spare_reg_nr=n [X86]
1903 Range: 0,7 : spare reg number
1905 Used for mtrr cleanup. It is spare mtrr entries number.
1906 Set to 2 or more if your graphical card needs more.
1908 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
1910 netdev= [NET] Network devices parameters
1911 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
1912 Note that mem_start is often overloaded to mean
1913 something different and driver-specific.
1914 This usage is only documented in each driver source
1918 [NETFILTER] Enable connection tracking flow accounting
1919 0 to disable accounting
1920 1 to enable accounting
1923 nfsaddrs= [NFS] Deprecated. Use ip= instead.
1924 See Documentation/filesystems/nfs/nfsroot.txt.
1926 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
1927 See Documentation/filesystems/nfs/nfsroot.txt.
1929 nfsrootdebug [NFS] enable nfsroot debugging messages.
1930 See Documentation/filesystems/nfs/nfsroot.txt.
1932 nfs.callback_tcpport=
1933 [NFS] set the TCP port on which the NFSv4 callback
1934 channel should listen.
1937 [NFS] sets the pathname to the program which is used
1938 to update the NFS client cache entries.
1940 nfs.cache_getent_timeout=
1941 [NFS] sets the timeout after which an attempt to
1942 update a cache entry is deemed to have failed.
1944 nfs.idmap_cache_timeout=
1945 [NFS] set the maximum lifetime for idmapper cache
1949 [NFS] enable 64-bit inode numbers.
1950 If zero, the NFS client will fake up a 32-bit inode
1951 number for the readdir() and stat() syscalls instead
1952 of returning the full 64-bit number.
1953 The default is to return 64-bit inode numbers.
1955 nfs.max_session_slots=
1956 [NFSv4.1] Sets the maximum number of session slots
1957 the client will attempt to negotiate with the server.
1958 This limits the number of simultaneous RPC requests
1959 that the client can send to the NFSv4.1 server.
1960 Note that there is little point in setting this
1961 value higher than the max_tcp_slot_table_limit.
1963 nfs.nfs4_disable_idmapping=
1964 [NFSv4] When set to the default of '1', this option
1965 ensures that both the RPC level authentication
1966 scheme and the NFS level operations agree to use
1967 numeric uids/gids if the mount is using the
1968 'sec=sys' security flavour. In effect it is
1969 disabling idmapping, which can make migration from
1970 legacy NFSv2/v3 systems to NFSv4 easier.
1971 Servers that do not support this mode of operation
1972 will be autodetected by the client, and it will fall
1973 back to using the idmapper.
1974 To turn off this behaviour, set the value to '0'.
1976 [NFS4] Specify an additional fixed unique ident-
1977 ification string that NFSv4 clients can insert into
1978 their nfs_client_id4 string. This is typically a
1979 UUID that is generated at system install time.
1981 nfs.send_implementation_id =
1982 [NFSv4.1] Send client implementation identification
1983 information in exchange_id requests.
1984 If zero, no implementation identification information
1986 The default is to send the implementation identification
1989 nfs.recover_lost_locks =
1990 [NFSv4] Attempt to recover locks that were lost due
1991 to a lease timeout on the server. Please note that
1992 doing this risks data corruption, since there are
1993 no guarantees that the file will remain unchanged
1994 after the locks are lost.
1995 If you want to enable the kernel legacy behaviour of
1996 attempting to recover these locks, then set this
1998 The default parameter value of '0' causes the kernel
1999 not to attempt recovery of lost locks.
2001 nfsd.nfs4_disable_idmapping=
2002 [NFSv4] When set to the default of '1', the NFSv4
2003 server will return only numeric uids and gids to
2004 clients using auth_sys, and will accept numeric uids
2005 and gids from such clients. This is intended to ease
2006 migration from NFSv2/v3.
2008 objlayoutdriver.osd_login_prog=
2009 [NFS] [OBJLAYOUT] sets the pathname to the program which
2010 is used to automatically discover and login into new
2011 osd-targets. Please see:
2012 Documentation/filesystems/pnfs.txt for more explanations
2014 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
2015 when a NMI is triggered.
2016 Format: [state][,regs][,debounce][,die]
2018 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2019 Format: [panic,][nopanic,][num]
2021 0 - turn nmi_watchdog off
2022 When panic is specified, panic when an NMI watchdog
2023 timeout occurs (or 'nopanic' to override the opposite
2025 This is useful when you use a panic=... timeout and
2026 need the box quickly up again.
2028 netpoll.carrier_timeout=
2029 [NET] Specifies amount of time (in seconds) that
2030 netpoll should wait for a carrier. By default netpoll
2033 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2034 emulation library even if a 387 maths coprocessor
2038 [HW] Never suspend the console
2039 Disable suspending of consoles during suspend and
2040 hibernate operations. Once disabled, debugging
2041 messages can reach various consoles while the rest
2042 of the system is being put to sleep (ie, while
2043 debugging driver suspend/resume hooks). This may
2044 not work reliably with all consoles, but is known
2045 to work with serial and VGA consoles.
2046 To facilitate more flexible debugging, we also add
2047 console_suspend, a printk module parameter to control
2048 it. Users could use console_suspend (usually
2049 /sys/module/printk/parameters/console_suspend) to
2050 turn on/off it dynamically.
2052 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2053 caches in the slab allocator. Saves per-node memory,
2054 but will impact performance.
2058 noapic [SMP,APIC] Tells the kernel to not make use of any
2059 IOAPICs that may be present in the system.
2062 Disable kernel base offset ASLR (Address Space
2063 Layout Randomization) if built into the kernel.
2065 noautogroup Disable scheduler automatic task group creation.
2067 nobats [PPC] Do not use BATs for mapping kernel lowmem
2068 on "Classic" PPC cores.
2072 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2074 nodelayacct [KNL] Disable per-task delay accounting
2076 nodisconnect [HW,SCSI,M68K] Disables SCSI disconnects.
2078 nodsp [SH] Disable hardware DSP at boot time.
2080 noefi [X86] Disable EFI runtime services support.
2085 On X86-32 available only on PAE configured kernels.
2086 noexec=on: enable non-executable mappings (default)
2087 noexec=off: disable non-executable mappings
2090 Disable SMAP (Supervisor Mode Access Prevention)
2091 even if it is supported by processor.
2094 Disable SMEP (Supervisor Mode Execution Prevention)
2095 even if it is supported by processor.
2098 This affects only 32-bit executables.
2099 noexec32=on: enable non-executable mappings (default)
2100 read doesn't imply executable mappings
2101 noexec32=off: disable non-executable mappings
2102 read implies executable mappings
2104 nofpu [SH] Disable hardware FPU at boot time.
2106 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2107 register save and restore. The kernel will only save
2108 legacy floating-point registers on task switch.
2110 noxsave [BUGS=X86] Disables x86 extended register state save
2111 and restore using xsave. The kernel will fallback to
2112 enabling legacy floating-point and sse state.
2115 on enable eager fpu restore
2116 off disable eager fpu restore
2117 auto selects the default scheme, which automatically
2118 enables eagerfpu restore for xsaveopt.
2120 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2121 wfi(ARM) instruction doesn't work correctly and not to
2122 use it. This is also useful when using JTAG debugger.
2124 no_file_caps Tells the kernel not to honor file capabilities. The
2125 only way then for a file to be executed with privilege
2126 is to be setuid root or executed by root.
2128 nohalt [IA-64] Tells the kernel not to use the power saving
2129 function PAL_HALT_LIGHT when idle. This increases
2130 power-consumption. On the positive side, it reduces
2131 interrupt wake-up latency, which may improve performance
2132 in certain environments such as networked servers or
2135 nohz= [KNL] Boottime enable/disable dynamic ticks
2136 Valid arguments: on, off
2139 nohz_full= [KNL,BOOT]
2140 In kernels built with CONFIG_NO_HZ_FULL=y, set
2141 the specified list of CPUs whose tick will be stopped
2142 whenever possible. The boot CPU will be forced outside
2143 the range to maintain the timekeeping.
2144 The CPUs in this range must also be included in the
2147 noiotrap [SH] Disables trapped I/O port accesses.
2149 noirqdebug [X86-32] Disables the code which attempts to detect and
2150 disable unhandled interrupt sources.
2152 no_timer_check [X86,APIC] Disables the code which tests for
2153 broken timer IRQ sources.
2155 noisapnp [ISAPNP] Disables ISA PnP code.
2157 noinitrd [RAM] Tells the kernel not to load any configured
2160 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2162 [Deprecated - use intremap=off]
2166 nojitter [IA-64] Disables jitter checking for ITC timers.
2168 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2170 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2173 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2174 steal time is computed, but won't influence scheduler
2177 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2179 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2181 noltlbs [PPC] Do not use large page/tlb entries for kernel
2182 lowmem mapping on PPC40x.
2184 nomca [IA-64] Disable machine check abort handling
2186 nomce [X86-32] Machine Check Exception
2188 nomfgpt [X86-32] Disable Multi-Function General Purpose
2189 Timer usage (for AMD Geode machines).
2191 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2192 shutdown the other cpus. Instead use the REBOOT_VECTOR
2195 nomodule Disable module load
2197 nopat [X86] Disable PAT (page attribute table extension of
2198 pagetables) support.
2200 norandmaps Don't use address space randomization. Equivalent to
2201 echo 0 > /proc/sys/kernel/randomize_va_space
2203 noreplace-paravirt [X86,IA-64,PV_OPS] Don't patch paravirt_ops
2205 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2206 with UP alternatives
2208 nordrand [X86] Disable the direct use of the RDRAND
2209 instruction even if it is supported by the
2210 processor. RDRAND is still available to user
2213 noresume [SWSUSP] Disables resume and restores original swap
2216 no-scroll [VGA] Disables scrollback.
2217 This is required for the Braillex ib80-piezo Braille
2218 reader made by F.H. Papenmeier (Germany).
2222 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2224 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2225 and disable the IO APIC. legacy for "maxcpus=0".
2227 nosoftlockup [KNL] Disable the soft-lockup detector.
2229 nosync [HW,M68K] Disables sync negotiation for all devices.
2231 notsc [BUGS=X86-32] Disable Time Stamp Counter
2233 nousb [USB] Disable the USB subsystem
2235 nowatchdog [KNL] Disable the lockup detector (NMI watchdog).
2239 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2241 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2242 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2243 Some features depend on CPU0. Known dependencies are:
2244 1. Resume from suspend/hibernate depends on CPU0.
2245 Suspend/hibernate will fail if CPU0 is offline and you
2246 need to online CPU0 before suspend/hibernate.
2247 2. PIC interrupts also depend on CPU0. CPU0 can't be
2248 removed if a PIC interrupt is detected.
2249 It's said poweroff/reboot may depend on CPU0 on some
2250 machines although I haven't seen such issues so far
2251 after CPU0 is offline on a few tested machines.
2252 If the dependencies are under your control, you can
2253 turn on cpu0_hotplug.
2255 nptcg= [IA-64] Override max number of concurrent global TLB
2256 purges which is reported from either PAL_VM_SUMMARY or
2259 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2260 could support. nr_cpus=n : n >= 1 limits the kernel to
2261 supporting 'n' processors. Later in runtime you can not
2262 use hotplug cpu feature to put more cpu back to online.
2263 just like you compile the kernel NR_CPUS=n
2265 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2267 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2268 Allowed values are enable and disable
2270 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2271 one of ['zone', 'node', 'default'] can be specified
2272 This can be set from sysctl after boot.
2273 See Documentation/sysctl/vm.txt for details.
2275 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2276 See Documentation/debugging-via-ohci1394.txt for more
2279 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2280 Rather than timing out after 20 ms if an EC
2281 command is not properly ACKed, override the length
2282 of the timeout. We have interrupts disabled while
2283 waiting for the ACK, so if this is set too high
2284 interrupts *may* be lost!
2286 omap_mux= [OMAP] Override bootloader pin multiplexing.
2287 Format: <mux_mode0.mode_name=value>...
2288 For example, to override I2C bus2:
2289 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2291 oprofile.timer= [HW]
2292 Use timer interrupt instead of performance counters
2294 oprofile.cpu_type= Force an oprofile cpu type
2295 This might be useful if you have an older oprofile
2296 userland or if you want common events.
2297 Format: { arch_perfmon }
2298 arch_perfmon: [X86] Force use of architectural
2299 perfmon on Intel CPUs instead of the
2300 CPU specific event set.
2301 timer: [X86] Force use of architectural NMI
2302 timer mode (see also oprofile.timer
2303 for generic hr timer mode)
2304 [s390] Force legacy basic mode sampling
2305 (report cpu_type "timer")
2307 oops=panic Always panic on oopses. Default is to just kill the
2308 process, but there is a small probability of
2309 deadlocking the machine.
2310 This will also cause panics on machine check exceptions.
2311 Useful together with panic=30 to trigger a reboot.
2314 See Documentation/sound/oss/oss-parameters.txt
2316 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2317 timeout > 0: seconds before rebooting
2318 timeout = 0: wait forever
2319 timeout < 0: reboot immediately
2322 parkbd.port= [HW] Parallel port number the keyboard adapter is
2323 connected to, default is 0.
2325 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2326 0 for XT, 1 for AT (default is AT).
2329 parport= [HW,PPT] Specify parallel ports. 0 disables.
2330 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2331 Use 'auto' to force the driver to use any
2332 IRQ/DMA settings detected (the default is to
2333 ignore detected IRQ/DMA settings because of
2334 possible conflicts). You can specify the base
2335 address, IRQ, and DMA settings; IRQ and DMA
2336 should be numbers, or 'auto' (for using detected
2337 settings on that particular port), or 'nofifo'
2338 (to avoid using a FIFO even if it is detected).
2339 Parallel ports are assigned in the order they
2340 are specified on the command line, starting
2343 parport_init_mode= [HW,PPT]
2344 Configure VIA parallel port to operate in
2345 a specific mode. This is necessary on Pegasos
2346 computer where firmware has no options for setting
2347 up parallel port mode and sets it to spp.
2348 Currently this function knows 686a and 8231 chips.
2349 Format: [spp|ps2|epp|ecp|ecpepp]
2352 Halt all CPUs after the first oops has been printed for
2353 the specified number of seconds. This is to be used if
2354 your oopses keep scrolling off the screen.
2359 See header of drivers/block/paride/pcd.c.
2360 See also Documentation/blockdev/paride.txt.
2362 pci=option[,option...] [PCI] various PCI subsystem options:
2363 earlydump [X86] dump PCI config space before the kernel
2365 off [X86] don't probe for the PCI bus
2366 bios [X86-32] force use of PCI BIOS, don't access
2367 the hardware directly. Use this if your machine
2368 has a non-standard PCI host bridge.
2369 nobios [X86-32] disallow use of PCI BIOS, only direct
2370 hardware access methods are allowed. Use this
2371 if you experience crashes upon bootup and you
2372 suspect they are caused by the BIOS.
2373 conf1 [X86] Force use of PCI Configuration
2375 conf2 [X86] Force use of PCI Configuration
2377 noaer [PCIE] If the PCIEAER kernel config parameter is
2378 enabled, this kernel boot option can be used to
2379 disable the use of PCIE advanced error reporting.
2380 nodomains [PCI] Disable support for multiple PCI
2381 root domains (aka PCI segments, in ACPI-speak).
2382 nommconf [X86] Disable use of MMCONFIG for PCI
2384 check_enable_amd_mmconf [X86] check for and enable
2385 properly configured MMIO access to PCI
2386 config space on AMD family 10h CPU
2387 nomsi [MSI] If the PCI_MSI kernel config parameter is
2388 enabled, this kernel boot option can be used to
2389 disable the use of MSI interrupts system-wide.
2390 noioapicquirk [APIC] Disable all boot interrupt quirks.
2391 Safety option to keep boot IRQs enabled. This
2392 should never be necessary.
2393 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
2394 primary IO-APIC for bridges that cannot disable
2395 boot IRQs. This fixes a source of spurious IRQs
2396 when the system masks IRQs.
2397 noioapicreroute [APIC] Disable workaround that uses the
2398 boot IRQ equivalent of an IRQ that connects to
2399 a chipset where boot IRQs cannot be disabled.
2400 The opposite of ioapicreroute.
2401 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
2402 routing table. These calls are known to be buggy
2403 on several machines and they hang the machine
2404 when used, but on other computers it's the only
2405 way to get the interrupt routing table. Try
2406 this option if the kernel is unable to allocate
2407 IRQs or discover secondary PCI buses on your
2409 rom [X86] Assign address space to expansion ROMs.
2410 Use with caution as certain devices share
2411 address decoders between ROMs and other
2413 norom [X86] Do not assign address space to
2414 expansion ROMs that do not already have
2415 BIOS assigned address ranges.
2416 nobar [X86] Do not assign address space to the
2417 BARs that weren't assigned by the BIOS.
2418 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
2419 assigned automatically to PCI devices. You can
2420 make the kernel exclude IRQs of your ISA cards
2422 pirqaddr=0xAAAAA [X86] Specify the physical address
2423 of the PIRQ table (normally generated
2424 by the BIOS) if it is outside the
2425 F0000h-100000h range.
2426 lastbus=N [X86] Scan all buses thru bus #N. Can be
2427 useful if the kernel is unable to find your
2428 secondary buses and you want to tell it
2429 explicitly which ones they are.
2430 assign-busses [X86] Always assign all PCI bus
2431 numbers ourselves, overriding
2432 whatever the firmware may have done.
2433 usepirqmask [X86] Honor the possible IRQ mask stored
2434 in the BIOS $PIR table. This is needed on
2435 some systems with broken BIOSes, notably
2436 some HP Pavilion N5400 and Omnibook XE3
2437 notebooks. This will have no effect if ACPI
2438 IRQ routing is enabled.
2439 noacpi [X86] Do not use ACPI for IRQ routing
2440 or for PCI scanning.
2441 use_crs [X86] Use PCI host bridge window information
2442 from ACPI. On BIOSes from 2008 or later, this
2443 is enabled by default. If you need to use this,
2444 please report a bug.
2445 nocrs [X86] Ignore PCI host bridge windows from ACPI.
2446 If you need to use this, please report a bug.
2447 routeirq Do IRQ routing for all PCI devices.
2448 This is normally done in pci_enable_device(),
2449 so this option is a temporary workaround
2450 for broken drivers that don't call it.
2451 skip_isa_align [X86] do not align io start addr, so can
2452 handle more pci cards
2453 firmware [ARM] Do not re-enumerate the bus but instead
2454 just use the configuration from the
2455 bootloader. This is currently used on
2456 IXP2000 systems where the bus has to be
2457 configured a certain way for adjunct CPUs.
2458 noearly [X86] Don't do any early type 1 scanning.
2459 This might help on some broken boards which
2460 machine check when some devices' config space
2461 is read. But various workarounds are disabled
2462 and some IOMMU drivers will not work.
2463 bfsort Sort PCI devices into breadth-first order.
2464 This sorting is done to get a device
2465 order compatible with older (<= 2.4) kernels.
2466 nobfsort Don't sort PCI devices into breadth-first order.
2467 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
2468 tuning and use the BIOS-configured MPS defaults.
2469 pcie_bus_safe Set every device's MPS to the largest value
2470 supported by all devices below the root complex.
2471 pcie_bus_perf Set device MPS to the largest allowable MPS
2472 based on its parent bus. Also set MRRS (Max
2473 Read Request Size) to the largest supported
2474 value (no larger than the MPS that the device
2475 or bus can support) for best performance.
2476 pcie_bus_peer2peer Set every device's MPS to 128B, which
2477 every device is guaranteed to support. This
2478 configuration allows peer-to-peer DMA between
2479 any pair of devices, possibly at the cost of
2480 reduced performance. This also guarantees
2481 that hot-added devices will work.
2482 cbiosize=nn[KMG] The fixed amount of bus space which is
2483 reserved for the CardBus bridge's IO window.
2484 The default value is 256 bytes.
2485 cbmemsize=nn[KMG] The fixed amount of bus space which is
2486 reserved for the CardBus bridge's memory
2487 window. The default value is 64 megabytes.
2490 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
2491 Specifies alignment and device to reassign
2492 aligned memory resources.
2493 If <order of align> is not specified,
2494 PAGE_SIZE is used as alignment.
2495 PCI-PCI bridge can be specified, if resource
2496 windows need to be expanded.
2497 ecrc= Enable/disable PCIe ECRC (transaction layer
2498 end-to-end CRC checking).
2499 bios: Use BIOS/firmware settings. This is the
2503 hpiosize=nn[KMG] The fixed amount of bus space which is
2504 reserved for hotplug bridge's IO window.
2505 Default size is 256 bytes.
2506 hpmemsize=nn[KMG] The fixed amount of bus space which is
2507 reserved for hotplug bridge's memory window.
2508 Default size is 2 megabytes.
2509 realloc= Enable/disable reallocating PCI bridge resources
2510 if allocations done by BIOS are too small to
2511 accommodate resources required by all child
2513 off: Turn realloc off
2515 realloc same as realloc=on
2516 noari do not use PCIe ARI.
2517 pcie_scan_all Scan all possible PCIe devices. Otherwise we
2518 only look for one device below a PCIe downstream
2521 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
2524 force Enable ASPM even on devices that claim not to support it.
2525 WARNING: Forcing ASPM on may cause system lockups.
2527 pcie_hp= [PCIE] PCI Express Hotplug driver options:
2528 nomsi Do not use MSI for PCI Express Native Hotplug (this
2529 makes all PCIe ports use INTx for hotplug services).
2531 pcie_ports= [PCIE] PCIe ports handling:
2532 auto Ask the BIOS whether or not to use native PCIe services
2533 associated with PCIe ports (PME, hot-plug, AER). Use
2534 them only if that is allowed by the BIOS.
2535 native Use native PCIe services associated with PCIe ports
2537 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
2540 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
2541 nomsi Do not use MSI for native PCIe PME signaling (this makes
2542 all PCIe root ports use INTx for all services).
2544 pcmv= [HW,PCMCIA] BadgePAD 4
2547 See Documentation/blockdev/paride.txt.
2549 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
2552 See arch/parisc/kernel/pdc_chassis.c
2554 percpu_alloc= Select which percpu first chunk allocator to use.
2555 Currently supported values are "embed" and "page".
2556 Archs may support subset or none of the selections.
2557 See comments in mm/percpu.c for details on each
2558 allocator. This parameter is primarily for debugging
2559 and performance comparison.
2562 See Documentation/blockdev/paride.txt.
2565 See Documentation/blockdev/paride.txt.
2567 pirq= [SMP,APIC] Manual mp-table setup
2568 See Documentation/x86/i386/IO-APIC.txt.
2570 plip= [PPT,NET] Parallel port network link
2571 Format: { parport<nr> | timid | 0 }
2572 See also Documentation/parport.txt.
2574 pmtmr= [X86] Manual setup of pmtmr I/O Port.
2575 Override pmtimer IOPort with a hex value.
2579 Enable PNP debug messages (depends on the
2580 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
2581 via /sys/module/pnp/parameters/debug. We always show
2582 current resource usage; turning this on also shows
2583 possible settings and some assignment information.
2589 { on | off | curr | res | no-curr | no-res }
2592 [ISAPNP] Exclude IRQs for the autoconfiguration
2595 [ISAPNP] Exclude DMAs for the autoconfiguration
2597 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
2598 Ranges are in pairs (I/O port base and size).
2601 [ISAPNP] Exclude memory regions for the
2603 Ranges are in pairs (memory base and size).
2605 ports= [IP_VS_FTP] IPVS ftp helper module
2607 Up to 8 (IP_VS_APP_MAX_PORTS) ports
2609 Format: <port>,<port>....
2611 print-fatal-signals=
2612 [KNL] debug: print fatal signals
2614 If enabled, warn about various signal handling
2615 related application anomalies: too many signals,
2616 too many POSIX.1 timers, fatal signals causing a
2619 If you hit the warning due to signal overflow,
2620 you might want to try "ulimit -i unlimited".
2624 printk.always_kmsg_dump=
2625 Trigger kmsg_dump for cases other than kernel oops or
2627 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2630 printk.time= Show timing data prefixed to each printk message line
2631 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2633 processor.max_cstate= [HW,ACPI]
2634 Limit processor to maximum C-state
2635 max_cstate=9 overrides any DMI blacklist limit.
2637 processor.nocst [HW,ACPI]
2638 Ignore the _CST method to determine C-states,
2639 instead using the legacy FADT method
2641 profile= [KNL] Enable kernel profiling via /proc/profile
2642 Format: [schedule,]<number>
2643 Param: "schedule" - profile schedule points.
2644 Param: <number> - step/bucket size as a power of 2 for
2645 statistical time based profiling.
2646 Param: "sleep" - profile D-state sleeping (millisecs).
2647 Requires CONFIG_SCHEDSTATS
2648 Param: "kvm" - profile VM exits.
2650 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
2652 See Documentation/blockdev/ramdisk.txt.
2654 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
2655 probe for; one of (bare|imps|exps|lifebook|any).
2656 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
2658 psmouse.resetafter= [HW,MOUSE]
2659 Try to reset the device after so many bad packets
2662 [HW,MOUSE] Set desired mouse resolution, in dpi.
2663 psmouse.smartscroll=
2664 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
2665 0 = disabled, 1 = enabled (default).
2667 pstore.backend= Specify the name of the pstore backend to use
2670 See Documentation/blockdev/paride.txt.
2673 [KNL] Number of legacy pty's. Overwrites compiled-in
2676 quiet [KNL] Disable most log messages
2681 See Documentation/md.txt.
2683 ramdisk_blocksize= [RAM]
2684 See Documentation/blockdev/ramdisk.txt.
2686 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
2687 See Documentation/blockdev/ramdisk.txt.
2690 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
2691 the specified list of CPUs to be no-callback CPUs.
2692 Invocation of these CPUs' RCU callbacks will
2693 be offloaded to "rcuox/N" kthreads created for
2694 that purpose, where "x" is "b" for RCU-bh, "p"
2695 for RCU-preempt, and "s" for RCU-sched, and "N"
2696 is the CPU number. This reduces OS jitter on the
2697 offloaded CPUs, which can be useful for HPC and
2698 real-time workloads. It can also improve energy
2699 efficiency for asymmetric multiprocessors.
2702 Rather than requiring that offloaded CPUs
2703 (specified by rcu_nocbs= above) explicitly
2704 awaken the corresponding "rcuoN" kthreads,
2705 make these kthreads poll for callbacks.
2706 This improves the real-time response for the
2707 offloaded CPUs by relieving them of the need to
2708 wake up the corresponding kthread, but degrades
2709 energy efficiency by requiring that the kthreads
2710 periodically wake up to do the polling.
2712 rcutree.blimit= [KNL]
2713 Set maximum number of finished RCU callbacks to
2714 process in one batch.
2716 rcutree.rcu_fanout_leaf= [KNL]
2717 Increase the number of CPUs assigned to each
2718 leaf rcu_node structure. Useful for very large
2721 rcutree.jiffies_till_first_fqs= [KNL]
2722 Set delay from grace-period initialization to
2723 first attempt to force quiescent states.
2724 Units are jiffies, minimum value is zero,
2725 and maximum value is HZ.
2727 rcutree.jiffies_till_next_fqs= [KNL]
2728 Set delay between subsequent attempts to force
2729 quiescent states. Units are jiffies, minimum
2730 value is one, and maximum value is HZ.
2732 rcutree.qhimark= [KNL]
2733 Set threshold of queued RCU callbacks beyond which
2734 batch limiting is disabled.
2736 rcutree.qlowmark= [KNL]
2737 Set threshold of queued RCU callbacks below which
2738 batch limiting is re-enabled.
2740 rcutree.rcu_idle_gp_delay= [KNL]
2741 Set wakeup interval for idle CPUs that have
2742 RCU callbacks (RCU_FAST_NO_HZ=y).
2744 rcutree.rcu_idle_lazy_gp_delay= [KNL]
2745 Set wakeup interval for idle CPUs that have
2746 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
2747 Lazy RCU callbacks are those which RCU can
2748 prove do nothing more than free memory.
2750 rcutorture.fqs_duration= [KNL]
2751 Set duration of force_quiescent_state bursts.
2753 rcutorture.fqs_holdoff= [KNL]
2754 Set holdoff time within force_quiescent_state bursts.
2756 rcutorture.fqs_stutter= [KNL]
2757 Set wait time between force_quiescent_state bursts.
2759 rcutorture.gp_exp= [KNL]
2760 Use expedited update-side primitives.
2762 rcutorture.gp_normal= [KNL]
2763 Use normal (non-expedited) update-side primitives.
2764 If both gp_exp and gp_normal are set, do both.
2765 If neither gp_exp nor gp_normal are set, still
2768 rcutorture.n_barrier_cbs= [KNL]
2769 Set callbacks/threads for rcu_barrier() testing.
2771 rcutorture.nfakewriters= [KNL]
2772 Set number of concurrent RCU writers. These just
2773 stress RCU, they don't participate in the actual
2774 test, hence the "fake".
2776 rcutorture.nreaders= [KNL]
2777 Set number of RCU readers.
2779 rcutorture.object_debug= [KNL]
2780 Enable debug-object double-call_rcu() testing.
2782 rcutorture.onoff_holdoff= [KNL]
2783 Set time (s) after boot for CPU-hotplug testing.
2785 rcutorture.onoff_interval= [KNL]
2786 Set time (s) between CPU-hotplug operations, or
2787 zero to disable CPU-hotplug testing.
2789 rcutorture.rcutorture_runnable= [BOOT]
2790 Start rcutorture running at boot time.
2792 rcutorture.shuffle_interval= [KNL]
2793 Set task-shuffle interval (s). Shuffling tasks
2794 allows some CPUs to go into dyntick-idle mode
2795 during the rcutorture test.
2797 rcutorture.shutdown_secs= [KNL]
2798 Set time (s) after boot system shutdown. This
2799 is useful for hands-off automated testing.
2801 rcutorture.stall_cpu= [KNL]
2802 Duration of CPU stall (s) to test RCU CPU stall
2803 warnings, zero to disable.
2805 rcutorture.stall_cpu_holdoff= [KNL]
2806 Time to wait (s) after boot before inducing stall.
2808 rcutorture.stat_interval= [KNL]
2809 Time (s) between statistics printk()s.
2811 rcutorture.stutter= [KNL]
2812 Time (s) to stutter testing, for example, specifying
2813 five seconds causes the test to run for five seconds,
2814 wait for five seconds, and so on. This tests RCU's
2815 ability to transition abruptly to and from idle.
2817 rcutorture.test_boost= [KNL]
2818 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
2819 "Maybe" means test if the RCU implementation
2820 under test support RCU priority boosting.
2822 rcutorture.test_boost_duration= [KNL]
2823 Duration (s) of each individual boost test.
2825 rcutorture.test_boost_interval= [KNL]
2826 Interval (s) between each boost test.
2828 rcutorture.test_no_idle_hz= [KNL]
2829 Test RCU's dyntick-idle handling. See also the
2830 rcutorture.shuffle_interval parameter.
2832 rcutorture.torture_type= [KNL]
2833 Specify the RCU implementation to test.
2835 rcutorture.verbose= [KNL]
2836 Enable additional printk() statements.
2838 rcupdate.rcu_expedited= [KNL]
2839 Use expedited grace-period primitives, for
2840 example, synchronize_rcu_expedited() instead
2841 of synchronize_rcu(). This reduces latency,
2842 but can increase CPU utilization, degrade
2843 real-time latency, and degrade energy efficiency.
2845 rcupdate.rcu_cpu_stall_suppress= [KNL]
2846 Suppress RCU CPU stall warning messages.
2848 rcupdate.rcu_cpu_stall_timeout= [KNL]
2849 Set timeout for RCU CPU stall warning messages.
2853 Run specified binary instead of /init from the ramdisk,
2854 used for early userspace startup. See initrd.
2857 Format (x86 or x86_64):
2858 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
2860 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
2862 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
2863 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
2864 reboot_force is either force or not specified,
2865 reboot_cpu is s[mp]#### with #### being the processor
2866 to be used for rebooting.
2869 [KNL, SMP] Set scheduler's default relax_domain_level.
2870 See Documentation/cgroups/cpusets.txt.
2872 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
2874 reservetop= [X86-32]
2876 Reserves a hole at the top of the kernel virtual
2881 Set the amount of memory to reserve for BIOS at
2882 the bottom of the address space.
2884 reset_devices [KNL] Force drivers to reset the underlying device
2885 during initialization.
2888 Specify the partition device for software suspend
2890 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
2892 resume_offset= [SWSUSP]
2893 Specify the offset from the beginning of the partition
2894 given by "resume=" at which the swap header is located,
2895 in <PAGE_SIZE> units (needed only for swap files).
2896 See Documentation/power/swsusp-and-swap-files.txt
2898 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
2899 read the resume files
2901 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
2902 Useful for devices that are detected asynchronously
2903 (e.g. USB and MMC devices).
2905 hibernate= [HIBERNATION]
2906 noresume Don't check if there's a hibernation image
2907 present during boot.
2908 nocompress Don't compress/decompress hibernation images.
2910 retain_initrd [RAM] Keep initrd memory after extraction
2912 rhash_entries= [KNL,NET]
2913 Set number of hash buckets for route cache
2915 riscom8= [HW,SERIAL]
2916 Format: <io_board1>[,<io_board2>[,...<io_boardN>]]
2918 ro [KNL] Mount root device read-only on boot
2920 root= [KNL] Root filesystem
2921 See name_to_dev_t comment in init/do_mounts.c.
2923 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
2924 mount the root filesystem
2926 rootflags= [KNL] Set root filesystem mount option string
2928 rootfstype= [KNL] Set root filesystem type
2930 rootwait [KNL] Wait (indefinitely) for root device to show up.
2931 Useful for devices that are detected asynchronously
2932 (e.g. USB and MMC devices).
2934 rproc_mem=nn[KMG][@address]
2935 [KNL,ARM,CMA] Remoteproc physical memory block.
2936 Memory area to be used by remote processor image,
2939 rw [KNL] Mount root device read-write on boot
2941 S [KNL] Run init in single mode
2944 See drivers/net/irda/sa1100_ir.c.
2946 sbni= [NET] Granch SBNI12 leased line adapter
2948 sched_debug [KNL] Enables verbose scheduler debug messages.
2950 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
2951 xtime_lock contention on larger systems, and/or RCU lock
2952 contention on all systems with CONFIG_MAXSMP set.
2953 Format: { "0" | "1" }
2954 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
2956 Note: increases power consumption, thus should only be
2957 enabled if running jitter sensitive (HPC/RT) workloads.
2959 security= [SECURITY] Choose a security module to enable at boot.
2960 If this boot parameter is not specified, only the first
2961 security module asking for security registration will be
2962 loaded. An invalid security module name will be treated
2963 as if no module has been chosen.
2965 selinux= [SELINUX] Disable or enable SELinux at boot time.
2966 Format: { "0" | "1" }
2967 See security/selinux/Kconfig help text.
2970 Default value is set via kernel config option.
2971 If enabled at boot time, /selinux/disable can be used
2972 later to disable prior to initial policy load.
2974 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
2975 Format: { "0" | "1" }
2976 See security/apparmor/Kconfig help text
2979 Default value is set via kernel config option.
2981 serialnumber [BUGS=X86-32]
2984 Maximal number of shapers.
2986 show_msr= [x86] show boot-time MSR settings
2987 Format: { <integer> }
2988 Show boot-time (BIOS-initialized) MSR settings.
2989 The parameter means the number of CPUs to show,
2990 for example 1 means boot CPU only.
2997 slab_max_order= [MM, SLAB]
2998 Determines the maximum allowed order for slabs.
2999 A high setting may cause OOMs due to memory
3000 fragmentation. Defaults to 1 for systems with
3001 more than 32MB of RAM, 0 otherwise.
3003 slub_debug[=options[,slabs]] [MM, SLUB]
3004 Enabling slub_debug allows one to determine the
3005 culprit if slab objects become corrupted. Enabling
3006 slub_debug can create guard zones around objects and
3007 may poison objects when not in use. Also tracks the
3008 last alloc / free. For more information see
3009 Documentation/vm/slub.txt.
3011 slub_max_order= [MM, SLUB]
3012 Determines the maximum allowed order for slabs.
3013 A high setting may cause OOMs due to memory
3014 fragmentation. For more information see
3015 Documentation/vm/slub.txt.
3017 slub_min_objects= [MM, SLUB]
3018 The minimum number of objects per slab. SLUB will
3019 increase the slab order up to slub_max_order to
3020 generate a sufficiently large slab able to contain
3021 the number of objects indicated. The higher the number
3022 of objects the smaller the overhead of tracking slabs
3023 and the less frequently locks need to be acquired.
3024 For more information see Documentation/vm/slub.txt.
3026 slub_min_order= [MM, SLUB]
3027 Determines the minimum page order for slabs. Must be
3028 lower than slub_max_order.
3029 For more information see Documentation/vm/slub.txt.
3031 slub_nomerge [MM, SLUB]
3032 Disable merging of slabs with similar size. May be
3033 necessary if there is some reason to distinguish
3034 allocs to different slabs. Debug options disable
3035 merging on their own.
3036 For more information see Documentation/vm/slub.txt.
3039 Format: <io1>[,<io2>[,...,<io8>]]
3041 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
3042 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
3043 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
3044 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
3045 smsc-ircc2.ircc_irq= [HW] IRQ line
3046 smsc-ircc2.ircc_dma= [HW] DMA channel
3047 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
3048 0: Toshiba Satellite 1800 (GP data pin select)
3049 1: Fast pin select (default)
3053 [KNL] Should the soft-lockup detector generate panics.
3056 sonypi.*= [HW] Sony Programmable I/O Control Device driver
3057 See Documentation/laptops/sonypi.txt
3059 specialix= [HW,SERIAL] Specialix multi-serial port adapter
3060 See Documentation/serial/specialix.txt.
3062 spia_io_base= [HW,MTD]
3068 Enabled the stack tracer on boot up.
3070 stacktrace_filter=[function-list]
3071 [FTRACE] Limit the functions that the stack tracer
3072 will trace at boot up. function-list is a comma separated
3073 list of functions. This list can be changed at run
3074 time by the stack_trace_filter file in the debugfs
3075 tracing directory. Note, this enables stack tracing
3076 and the stacktrace above is not needed.
3080 Set the STI (builtin display/keyboard on the HP-PARISC
3081 machines) console (graphic card) which should be used
3082 as the initial boot-console.
3083 See also comment in drivers/video/console/sticore.c.
3086 See comment in drivers/video/console/sticore.c.
3089 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
3091 sunrpc.min_resvport=
3092 sunrpc.max_resvport=
3094 SunRPC servers often require that client requests
3095 originate from a privileged port (i.e. a port in the
3096 range 0 < portnr < 1024).
3097 An administrator who wishes to reserve some of these
3098 ports for other uses may adjust the range that the
3099 kernel's sunrpc client considers to be privileged
3100 using these two parameters to set the minimum and
3101 maximum port values.
3105 Control how the NFS server code allocates CPUs to
3106 service thread pools. Depending on how many NICs
3107 you have and where their interrupts are bound, this
3108 option will affect which CPUs will do NFS serving.
3109 Note: this parameter cannot be changed while the
3110 NFS server is running.
3112 auto the server chooses an appropriate mode
3113 automatically using heuristics
3114 global a single global pool contains all CPUs
3115 percpu one pool for each CPU
3116 pernode one pool for each NUMA node (equivalent
3117 to global on non-NUMA machines)
3119 sunrpc.tcp_slot_table_entries=
3120 sunrpc.udp_slot_table_entries=
3122 Sets the upper limit on the number of simultaneous
3123 RPC calls that can be sent from the client to a
3124 server. Increasing these values may allow you to
3125 improve throughput, but will also increase the
3126 amount of memory reserved for use by the client.
3129 [KNL] Enable accounting of swap in memory resource
3130 controller if no parameter or 1 is given or disable
3131 it if 0 is given (See Documentation/cgroups/memory.txt)
3133 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
3134 Format: { <int> | force }
3135 <int> -- Number of I/O TLB slabs
3136 force -- force using of bounce buffers even if they
3137 wouldn't be automatically used by the kernel
3141 sysfs.deprecated=0|1 [KNL]
3142 Enable/disable old style sysfs layout for old udev
3143 on older distributions. When this option is enabled
3144 very new udev will not work anymore. When this option
3145 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
3146 in older udev will not work anymore.
3147 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
3148 the kernel configuration.
3150 sysrq_always_enabled
3152 Ignore sysrq setting - this boot parameter will
3153 neutralize any effect of /proc/sys/kernel/sysrq.
3154 Useful for debugging.
3158 test_suspend= [SUSPEND]
3159 Specify "mem" (for Suspend-to-RAM) or "standby" (for
3160 standby suspend) as the system sleep state to briefly
3161 enter during system startup. The system is woken from
3162 this state using a wakeup-capable RTC alarm.
3164 thash_entries= [KNL,NET]
3165 Set number of hash buckets for TCP connection
3167 thermal.act= [HW,ACPI]
3168 -1: disable all active trip points in all thermal zones
3169 <degrees C>: override all lowest active trip points
3171 thermal.crt= [HW,ACPI]
3172 -1: disable all critical trip points in all thermal zones
3173 <degrees C>: override all critical trip points
3175 thermal.nocrt= [HW,ACPI]
3176 Set to disable actions on ACPI thermal zone
3177 critical and hot trip points.
3179 thermal.off= [HW,ACPI]
3180 1: disable ACPI thermal control
3182 thermal.psv= [HW,ACPI]
3183 -1: disable all passive trip points
3184 <degrees C>: override all passive trip points to this
3187 thermal.tzp= [HW,ACPI]
3188 Specify global default ACPI thermal zone polling rate
3189 <deci-seconds>: poll all this frequency
3190 0: no polling (default)
3193 Force threading of all interrupt handlers except those
3194 marked explicitly IRQF_NO_THREAD.
3197 Enable the Transcendent memory driver if built-in.
3199 tmem.cleancache=0|1 [KNL, XEN]
3200 Default is on (1). Disable the usage of the cleancache
3201 API to send anonymous pages to the hypervisor.
3203 tmem.frontswap=0|1 [KNL, XEN]
3204 Default is on (1). Disable the usage of the frontswap
3205 API to send swap pages to the hypervisor. If disabled
3206 the selfballooning and selfshrinking are force disabled.
3208 tmem.selfballooning=0|1 [KNL, XEN]
3209 Default is on (1). Disable the driving of swap pages
3212 tmem.selfshrinking=0|1 [KNL, XEN]
3213 Default is on (1). Partial swapoff that immediately
3214 transfers pages from Xen hypervisor back to the
3215 kernel based on different criteria.
3219 Specify if the kernel should make use of the cpu
3220 topology information if the hardware supports this.
3221 The scheduler will make use of this information and
3222 e.g. base its process migration decisions on it.
3227 tpm_suspend_pcr=[HW,TPM]
3228 Format: integer pcr id
3229 Specify that at suspend time, the tpm driver
3230 should extend the specified pcr with zeros,
3231 as a workaround for some chips which fail to
3232 flush the last written pcr on TPM_SaveState.
3233 This will guarantee that all the other pcrs
3236 trace_buf_size=nn[KMG]
3237 [FTRACE] will set tracing buffer size.
3239 trace_event=[event-list]
3240 [FTRACE] Set and start specified trace events in order
3241 to facilitate early boot debugging.
3242 See also Documentation/trace/events.txt
3244 trace_options=[option-list]
3245 [FTRACE] Enable or disable tracer options at boot.
3246 The option-list is a comma delimited list of options
3247 that can be enabled or disabled just as if you were
3248 to echo the option name into
3250 /sys/kernel/debug/tracing/trace_options
3252 For example, to enable stacktrace option (to dump the
3253 stack trace of each event), add to the command line:
3255 trace_options=stacktrace
3257 See also Documentation/trace/ftrace.txt "trace options"
3261 [FTRACE] enable this option to disable tracing when a
3262 warning is hit. This turns off "tracing_on". Tracing can
3263 be enabled again by echoing '1' into the "tracing_on"
3264 file located in /sys/kernel/debug/tracing/
3266 This option is useful, as it disables the trace before
3267 the WARNING dump is called, which prevents the trace to
3268 be filled with content caused by the warning output.
3270 This option can also be set at run time via the sysctl
3271 option: kernel/traceoff_on_warning
3273 transparent_hugepage=
3275 Format: [always|madvise|never]
3276 Can be used to control the default behavior of the system
3277 with respect to transparent hugepages.
3278 See Documentation/vm/transhuge.txt for more details.
3280 tsc= Disable clocksource stability checks for TSC.
3282 [x86] reliable: mark tsc clocksource as reliable, this
3283 disables clocksource verification at runtime, as well
3284 as the stability checks done at bootup. Used to enable
3285 high-resolution timer mode on older hardware, and in
3286 virtualized environment.
3287 [x86] noirqtime: Do not use TSC to do irq accounting.
3288 Used to run time disable IRQ_TIME_ACCOUNTING on any
3289 platforms where RDTSC is slow and this accounting
3292 turbografx.map[2|3]= [HW,JOY]
3293 TurboGraFX parallel port interface
3295 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
3296 See also Documentation/input/joystick-parport.txt
3298 udbg-immortal [PPC] When debugging early kernel crashes that
3299 happen after console_init() and before a proper
3300 console driver takes over, this boot options might
3301 help "seeing" what's going on.
3303 uhash_entries= [KNL,NET]
3304 Set number of hash buckets for UDP/UDP-Lite connections
3307 [USB] Ignore overcurrent events (default N).
3308 Some badly-designed motherboards generate lots of
3309 bogus events, for ports that aren't wired to
3310 anything. Set this parameter to avoid log spamming.
3311 Note that genuine overcurrent events won't be
3315 [X86] Cause panic on unknown NMI.
3317 usbcore.authorized_default=
3318 [USB] Default USB device authorization:
3319 (default -1 = authorized except for wireless USB,
3320 0 = not authorized, 1 = authorized)
3322 usbcore.autosuspend=
3323 [USB] The autosuspend time delay (in seconds) used
3324 for newly-detected USB devices (default 2). This
3325 is the time required before an idle device will be
3326 autosuspended. Devices for which the delay is set
3327 to a negative value won't be autosuspended at all.
3329 usbcore.usbfs_snoop=
3330 [USB] Set to log all usbfs traffic (default 0 = off).
3332 usbcore.blinkenlights=
3333 [USB] Set to cycle leds on hubs (default 0 = off).
3335 usbcore.old_scheme_first=
3336 [USB] Start with the old device initialization
3337 scheme (default 0 = off).
3339 usbcore.usbfs_memory_mb=
3340 [USB] Memory limit (in MB) for buffers allocated by
3341 usbfs (default = 16, 0 = max = 2047).
3343 usbcore.use_both_schemes=
3344 [USB] Try the other device initialization scheme
3345 if the first one fails (default 1 = enabled).
3347 usbcore.initial_descriptor_timeout=
3348 [USB] Specifies timeout for the initial 64-byte
3349 USB_REQ_GET_DESCRIPTOR request in milliseconds
3350 (default 5000 = 5.0 seconds).
3353 [USBHID] The interval which mice are to be polled at.
3355 usb-storage.delay_use=
3356 [UMS] The delay in seconds before a new device is
3357 scanned for Logical Units (default 5).
3360 [UMS] A list of quirks entries to supplement or
3361 override the built-in unusual_devs list. List
3362 entries are separated by commas. Each entry has
3363 the form VID:PID:Flags where VID and PID are Vendor
3364 and Product ID values (4-digit hex numbers) and
3365 Flags is a set of characters, each corresponding
3366 to a common usb-storage quirk flag as follows:
3367 a = SANE_SENSE (collect more than 18 bytes
3369 b = BAD_SENSE (don't collect more than 18
3370 bytes of sense data);
3371 c = FIX_CAPACITY (decrease the reported
3372 device capacity by one sector);
3373 d = NO_READ_DISC_INFO (don't use
3374 READ_DISC_INFO command);
3375 e = NO_READ_CAPACITY_16 (don't use
3376 READ_CAPACITY_16 command);
3377 h = CAPACITY_HEURISTICS (decrease the
3378 reported device capacity by one
3379 sector if the number is odd);
3380 i = IGNORE_DEVICE (don't bind to this
3382 l = NOT_LOCKABLE (don't try to lock and
3383 unlock ejectable media);
3384 m = MAX_SECTORS_64 (don't transfer more
3385 than 64 sectors = 32 KB at a time);
3386 n = INITIAL_READ10 (force a retry of the
3387 initial READ(10) command);
3388 o = CAPACITY_OK (accept the capacity
3389 reported by the device);
3390 p = WRITE_CACHE (the device cache is ON
3392 r = IGNORE_RESIDUE (the device reports
3393 bogus residue values);
3394 s = SINGLE_LUN (the device has only one
3396 w = NO_WP_DETECT (don't test whether the
3397 medium is write-protected).
3398 Example: quirks=0419:aaf5:rl,0421:0433:rc
3400 user_debug= [KNL,ARM]
3402 See arch/arm/Kconfig.debug help text.
3403 1 - undefined instruction events
3405 4 - invalid data aborts
3408 Example: user_debug=31
3411 [X86] Flags controlling user PTE allocations.
3413 nohigh = do not allocate PTE pages in
3414 HIGHMEM regardless of setting
3418 vdso=2: enable compat VDSO (default with COMPAT_VDSO)
3419 vdso=1: enable VDSO (default)
3420 vdso=0: disable VDSO mapping
3423 vdso32=2: enable compat VDSO (default with COMPAT_VDSO)
3424 vdso32=1: enable 32-bit VDSO (default)
3425 vdso32=0: disable 32-bit VDSO mapping
3428 vector=percpu: enable percpu vector domain
3430 video= [FB] Frame buffer configuration
3431 See Documentation/fb/modedb.txt.
3433 video.brightness_switch_enabled= [0,1]
3434 If set to 1, on receiving an ACPI notify event
3435 generated by hotkey, video driver will adjust brightness
3436 level and then send out the event to user space through
3437 the allocated input device; If set to 0, video driver
3438 will only send out the event without touching backlight
3443 [VMMIO] Memory mapped virtio (platform) device.
3445 <size>@<baseaddr>:<irq>[:<id>]
3447 <size> := size (can use standard suffixes
3449 <baseaddr> := physical base address
3450 <irq> := interrupt number (as passed to
3452 <id> := (optional) platform device id
3454 virtio_mmio.device=1K@0x100b0000:48:7
3456 Can be used multiple times for multiple devices.
3458 vga= [BOOT,X86-32] Select a particular video mode
3459 See Documentation/x86/boot.txt and
3460 Documentation/svga.txt.
3461 Use vga=ask for menu.
3462 This is actually a boot loader parameter; the value is
3463 passed to the kernel using a special protocol.
3465 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
3466 size of <nn>. This can be used to increase the
3467 minimum size (128MB on x86). It can also be used to
3468 decrease the size and leave more room for directly
3471 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
3474 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
3477 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
3481 Controls the behavior of vsyscalls (i.e. calls to
3482 fixed addresses of 0xffffffffff600x00 from legacy
3483 code). Most statically-linked binaries and older
3484 versions of glibc use these calls. Because these
3485 functions are at fixed addresses, they make nice
3486 targets for exploits that can control RIP.
3488 emulate [default] Vsyscalls turn into traps and are
3489 emulated reasonably safely.
3491 native Vsyscalls are native syscall instructions.
3492 This is a little bit faster than trapping
3493 and makes a few dynamic recompilers work
3494 better than they would in emulation mode.
3495 It also makes exploits much easier to write.
3497 none Vsyscalls don't work at all. This makes
3498 them quite hard to use for exploits but
3499 might break your system.
3501 vt.color= [VT] Default text color.
3502 Format: 0xYX, X = foreground, Y = background.
3503 Default: 0x07 = light gray on black.
3505 vt.cur_default= [VT] Default cursor shape.
3506 Format: 0xCCBBAA, where AA, BB, and CC are the same as
3507 the parameters of the <Esc>[?A;B;Cc escape sequence;
3508 see VGA-softcursor.txt. Default: 2 = underline.
3510 vt.default_blu= [VT]
3511 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
3512 Change the default blue palette of the console.
3513 This is a 16-member array composed of values
3516 vt.default_grn= [VT]
3517 Format: <green0>,<green1>,<green2>,...,<green15>
3518 Change the default green palette of the console.
3519 This is a 16-member array composed of values
3522 vt.default_red= [VT]
3523 Format: <red0>,<red1>,<red2>,...,<red15>
3524 Change the default red palette of the console.
3525 This is a 16-member array composed of values
3531 Set system-wide default UTF-8 mode for all tty's.
3532 Default is 1, i.e. UTF-8 mode is enabled for all
3533 newly opened terminals.
3535 vt.global_cursor_default=
3538 Set system-wide default for whether a cursor
3539 is shown on new VTs. Default is -1,
3540 i.e. cursors will be created by default unless
3541 overridden by individual drivers. 0 will hide
3542 cursors, 1 will display them.
3544 vt.italic= [VT] Default color for italic text; 0-15.
3547 vt.underline= [VT] Default color for underlined text; 0-15.
3550 watchdog timers [HW,WDT] For information on watchdog timers,
3551 see Documentation/watchdog/watchdog-parameters.txt
3552 or other driver-specific files in the
3553 Documentation/watchdog/ directory.
3555 workqueue.disable_numa
3556 By default, all work items queued to unbound
3557 workqueues are affine to the NUMA nodes they're
3558 issued on, which results in better behavior in
3559 general. If NUMA affinity needs to be disabled for
3560 whatever reason, this option can be used. Note
3561 that this also can be controlled per-workqueue for
3562 workqueues visible under /sys/bus/workqueue/.
3564 workqueue.power_efficient
3565 Per-cpu workqueues are generally preferred because
3566 they show better performance thanks to cache
3567 locality; unfortunately, per-cpu workqueues tend to
3568 be more power hungry than unbound workqueues.
3570 Enabling this makes the per-cpu workqueues which
3571 were observed to contribute significantly to power
3572 consumption unbound, leading to measurably lower
3573 power usage at the cost of small performance
3576 The default value of this parameter is determined by
3577 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
3579 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
3580 default x2apic cluster mode on platforms
3583 x86_intel_mid_timer= [X86-32,APBT]
3584 Choose timer option for x86 Intel MID platform.
3585 Two valid options are apbt timer only and lapic timer
3586 plus one apbt timer for broadcast timer.
3587 x86_intel_mid_timer=apbt_only | lapic_and_apbt
3589 xen_emul_unplug= [HW,X86,XEN]
3590 Unplug Xen emulated devices
3591 Format: [unplug0,][unplug1]
3592 ide-disks -- unplug primary master IDE devices
3593 aux-ide-disks -- unplug non-primary-master IDE devices
3594 nics -- unplug network devices
3595 all -- unplug all emulated devices (NICs and IDE disks)
3596 unnecessary -- unplugging emulated devices is
3597 unnecessary even if the host did not respond to
3599 never -- do not unplug even if version check succeeds
3601 xen_nopvspin [X86,XEN]
3602 Disables the ticketlock slowpath using Xen PV
3605 xirc2ps_cs= [NET,PCMCIA]
3607 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
3609 ______________________________________________________________________
3613 Add more DRM drivers.