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 DRM Direct Rendering Management support is enabled.
48 DYNAMIC_DEBUG Build in debug messages and enable them at runtime
49 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled
50 EFI EFI Partitioning (GPT) is enabled
51 EIDE EIDE/ATAPI support is enabled.
52 EVM Extended Verification Module
53 FB The frame buffer device is enabled.
54 FTRACE Function tracing enabled.
55 GCOV GCOV profiling is enabled.
56 HW Appropriate hardware is enabled.
57 IA-64 IA-64 architecture is enabled.
58 IMA Integrity measurement architecture is enabled.
59 IOSCHED More than one I/O scheduler is enabled.
60 IP_PNP IP DHCP, BOOTP, or RARP is enabled.
61 IPV6 IPv6 support is enabled.
62 ISAPNP ISA PnP code is enabled.
63 ISDN Appropriate ISDN support is enabled.
64 JOY Appropriate joystick support is enabled.
65 KGDB Kernel debugger support is enabled.
66 KVM Kernel Virtual Machine support is enabled.
67 LIBATA Libata driver is enabled
68 LP Printer support is enabled.
69 LOOP Loopback device support is enabled.
70 M68k M68k architecture is enabled.
71 These options have more detailed description inside of
72 Documentation/m68k/kernel-options.txt.
73 MDA MDA console support is enabled.
74 MIPS MIPS architecture is enabled.
75 MOUSE Appropriate mouse support is enabled.
76 MSI Message Signaled Interrupts (PCI).
77 MTD MTD (Memory Technology Device) support is enabled.
78 NET Appropriate network support is enabled.
79 NUMA NUMA support is enabled.
80 NFS Appropriate NFS support is enabled.
81 OSS OSS sound support is enabled.
82 PV_OPS A paravirtualized kernel is enabled.
83 PARIDE The ParIDE (parallel port IDE) subsystem is enabled.
84 PARISC The PA-RISC architecture is enabled.
85 PCI PCI bus support is enabled.
86 PCIE PCI Express support is enabled.
87 PCMCIA The PCMCIA subsystem is enabled.
88 PNP Plug & Play support is enabled.
89 PPC PowerPC architecture is enabled.
90 PPT Parallel port support is enabled.
91 PS2 Appropriate PS/2 support is enabled.
92 RAM RAM disk support is enabled.
93 S390 S390 architecture is enabled.
94 SCSI Appropriate SCSI support is enabled.
95 A lot of drivers have their options described inside
96 the Documentation/scsi/ sub-directory.
97 SECURITY Different security models are enabled.
98 SELINUX SELinux support is enabled.
99 APPARMOR AppArmor support is enabled.
100 SERIAL Serial support is enabled.
101 SH SuperH architecture is enabled.
102 SMP The kernel is an SMP kernel.
103 SPARC Sparc architecture is enabled.
104 SWSUSP Software suspend (hibernation) is enabled.
105 SUSPEND System suspend states are enabled.
106 TPM TPM drivers are enabled.
107 TS Appropriate touchscreen support is enabled.
108 UMS USB Mass Storage support is enabled.
109 USB USB support is enabled.
110 USBHID USB Human Interface Device support is enabled.
111 V4L Video For Linux support is enabled.
112 VMMIO Driver for memory mapped virtio devices is enabled.
113 VGA The VGA console has been enabled.
114 VT Virtual terminal support is enabled.
115 WDT Watchdog support is enabled.
116 XT IBM PC/XT MFM hard disk support is enabled.
117 X86-32 X86-32, aka i386 architecture is enabled.
118 X86-64 X86-64 architecture is enabled.
119 More X86-64 boot options can be found in
120 Documentation/x86/x86_64/boot-options.txt .
121 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64)
122 XEN Xen support is enabled
124 In addition, the following text indicates that the option:
126 BUGS= Relates to possible processor bugs on the said processor.
127 KNL Is a kernel start-up parameter.
128 BOOT Is a boot loader parameter.
130 Parameters denoted with BOOT are actually interpreted by the boot
131 loader, and have no meaning to the kernel directly.
132 Do not modify the syntax of boot loader parameters without extreme
133 need or coordination with <Documentation/x86/boot.txt>.
135 There are also arch-specific kernel-parameters not documented here.
136 See for example <Documentation/x86/x86_64/boot-options.txt>.
138 Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
139 a trailing = on the name of any parameter states that that parameter will
140 be entered as an environment variable, whereas its absence indicates that
141 it will appear as a kernel argument readable via /proc/cmdline by programs
142 running once the system is up.
144 The number of kernel parameters is not limited, but the length of the
145 complete command line (parameters including spaces etc.) is limited to
146 a fixed number of characters. This limit depends on the architecture
147 and is between 256 and 4096 characters. It is defined in the file
148 ./include/asm/setup.h as COMMAND_LINE_SIZE.
150 Finally, the [KMG] suffix is commonly described after a number of kernel
151 parameter values. These 'K', 'M', and 'G' letters represent the _binary_
152 multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
153 bytes respectively. Such letter suffixes can also be entirely omitted.
157 Advanced Configuration and Power Interface
158 Format: { force | off | strict | noirq | rsdt }
159 force -- enable ACPI if default was off
160 off -- disable ACPI if default was on
161 noirq -- do not use ACPI for IRQ routing
162 strict -- Be less tolerant of platforms that are not
163 strictly ACPI specification compliant.
164 rsdt -- prefer RSDT over (default) XSDT
165 copy_dsdt -- copy DSDT to memory
167 See also Documentation/power/runtime_pm.txt, pci=noacpi
169 acpi_rsdp= [ACPI,EFI,KEXEC]
170 Pass the RSDP address to the kernel, mostly used
171 on machines running EFI runtime service to boot the
172 second kernel for kdump.
174 acpi_apic_instance= [ACPI, IOAPIC]
176 2: use 2nd APIC table, if available
177 1,0: use 1st APIC table
180 acpi_backlight= [HW,ACPI]
181 acpi_backlight=vendor
183 If set to vendor, prefer vendor specific driver
184 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
185 of the ACPI video.ko driver.
187 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
188 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
190 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
191 debug output. Bits in debug_layer correspond to a
192 _COMPONENT in an ACPI source file, e.g.,
193 #define _COMPONENT ACPI_PCI_COMPONENT
194 Bits in debug_level correspond to a level in
195 ACPI_DEBUG_PRINT statements, e.g.,
196 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
197 The debug_level mask defaults to "info". See
198 Documentation/acpi/debug.txt for more information about
199 debug layers and levels.
201 Enable processor driver info messages:
202 acpi.debug_layer=0x20000000
203 Enable PCI/PCI interrupt routing info messages:
204 acpi.debug_layer=0x400000
205 Enable AML "Debug" output, i.e., stores to the Debug
206 object while interpreting AML:
207 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
208 Enable all messages related to ACPI hardware:
209 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
211 Some values produce so much output that the system is
212 unusable. The "log_buf_len" parameter may be useful
213 if you need to capture more output.
215 acpi_irq_balance [HW,ACPI]
216 ACPI will balance active IRQs
219 acpi_irq_nobalance [HW,ACPI]
220 ACPI will not move active IRQs (default)
223 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
224 Format: <irq>,<irq>...
226 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
228 Format: <irq>,<irq>...
230 acpi_no_auto_ssdt [HW,ACPI] Disable automatic loading of SSDT
232 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
233 Format: To spoof as Windows 98: ="Microsoft Windows"
235 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
236 acpi_osi="string1" # add string1 -- only one string
237 acpi_osi="!string2" # remove built-in string2
238 acpi_osi= # disable all strings
241 Override the pmtimer bug detection: force the kernel
242 to assume that this machine's pmtimer latches its value
243 and always returns good values.
245 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
246 Format: { level | edge | high | low }
248 acpi_serialize [HW,ACPI] force serialization of AML methods
250 acpi_skip_timer_override [HW,ACPI]
251 Recognize and ignore IRQ0/pin2 Interrupt Override.
252 For broken nForce2 BIOS resulting in XT-PIC timer.
254 acpi_sleep= [HW,ACPI] Sleep options
255 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
256 old_ordering, nonvs, sci_force_enable }
257 See Documentation/power/video.txt for information on
259 s3_beep is for debugging; it makes the PC's speaker beep
260 as soon as the kernel's real-mode entry point is called.
261 s4_nohwsig prevents ACPI hardware signature from being
262 used during resume from hibernation.
263 old_ordering causes the ACPI 1.0 ordering of the _PTS
264 control method, with respect to putting devices into
265 low power states, to be enforced (the ACPI 2.0 ordering
266 of _PTS is used by default).
267 nonvs prevents the kernel from saving/restoring the
268 ACPI NVS memory during suspend/hibernation and resume.
269 sci_force_enable causes the kernel to set SCI_EN directly
270 on resume from S1/S3 (which is against the ACPI spec,
271 but some broken systems don't work without it).
273 acpi_use_timer_override [HW,ACPI]
274 Use timer override. For some broken Nvidia NF5 boards
275 that require a timer override, but don't have HPET
277 acpi_enforce_resources= [ACPI]
278 { strict | lax | no }
279 Check for resource conflicts between native drivers
280 and ACPI OperationRegions (SystemIO and SystemMemory
281 only). IO ports and memory declared in ACPI might be
282 used by the ACPI subsystem in arbitrary AML code and
283 can interfere with legacy drivers.
284 strict (default): access to resources claimed by ACPI
285 is denied; legacy drivers trying to access reserved
286 resources will fail to bind to device using them.
287 lax: access to resources claimed by ACPI is allowed;
288 legacy drivers trying to access reserved resources
289 will bind successfully but a warning message is logged.
290 no: ACPI OperationRegions are not marked as reserved,
291 no further checks are performed.
293 add_efi_memmap [EFI; X86] Include EFI memory map in
294 kernel's map of available physical RAM.
297 { off | try_unsupported }
298 off: disable AGP support
299 try_unsupported: try to drive unsupported chipsets
300 (may crash computer or cause data corruption)
303 See Documentation/sound/alsa/alsa-parameters.txt
306 Allow the default userspace alignment fault handler
307 behaviour to be specified. Bit 0 enables warnings,
308 bit 1 enables fixups, and bit 2 sends a segfault.
310 align_va_addr= [X86-64]
311 Align virtual addresses by clearing slice [14:12] when
312 allocating a VMA at process creation time. This option
313 gives you up to 3% performance improvement on AMD F15h
314 machines (where it is enabled by default) for a
315 CPU-intensive style benchmark, and it can vary highly in
316 a microbenchmark depending on workload and compiler.
318 32: only for 32-bit processes
319 64: only for 64-bit processes
320 on: enable for both 32- and 64-bit processes
321 off: disable for both 32- and 64-bit processes
323 alloc_snapshot [FTRACE]
324 Allocate the ftrace snapshot buffer on boot up when the
325 main buffer is allocated. This is handy if debugging
326 and you need to use tracing_snapshot() on boot up, and
327 do not want to use tracing_snapshot_alloc() as it needs
328 to be done where GFP_KERNEL allocations are allowed.
330 amd_iommu= [HW,X86-64]
331 Pass parameters to the AMD IOMMU driver in the system.
333 fullflush - enable flushing of IO/TLB entries when
334 they are unmapped. Otherwise they are
335 flushed before they will be reused, which
337 off - do not initialize any AMD IOMMU found in
339 force_isolation - Force device isolation for all
340 devices. The IOMMU driver is not
341 allowed anymore to lift isolation
342 requirements as needed. This option
343 does not override iommu=pt
345 amd_iommu_dump= [HW,X86-64]
346 Enable AMD IOMMU driver option to dump the ACPI table
347 for AMD IOMMU. With this option enabled, AMD IOMMU
348 driver will print ACPI tables for AMD IOMMU during
349 IOMMU initialization.
351 amijoy.map= [HW,JOY] Amiga joystick support
352 Map of devices attached to JOY0DAT and JOY1DAT
354 See also Documentation/input/joystick.txt
356 analog.map= [HW,JOY] Analog joystick and gamepad support
357 Specifies type or capabilities of an analog joystick
358 connected to one of 16 gameports
359 Format: <type1>,<type2>,..<type16>
362 Power management functions (SPARCstation-4/5 + deriv.)
364 Disable APC CPU standby support. SPARCstation-Fox does
365 not play well with APC CPU idle - disable it if you have
366 APC and your system crashes randomly.
368 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
369 Change the output verbosity whilst booting
370 Format: { quiet (default) | verbose | debug }
371 Change the amount of debugging information output
372 when initialising the APIC and IO-APIC components.
375 See Documentation/networking/ipv6.txt.
377 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
378 Limit apic dumping. The parameter defines the maximal
379 number of local apics being dumped. Also it is possible
380 to set it to "all" by meaning -- no limit here.
381 Format: { 1 (default) | 2 | ... | all }.
382 The parameter valid if only apic=debug or
383 apic=verbose is specified.
384 Example: apic=debug show_lapic=all
386 apm= [APM] Advanced Power Management
387 See header of arch/x86/kernel/apm_32.c.
389 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
390 Format: <io>,<irq>,<nodeID>
394 atarimouse= [HW,MOUSE] Atari Mouse
396 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
397 EzKey and similar keyboards
399 atkbd.reset= [HW] Reset keyboard during initialization
401 atkbd.set= [HW] Select keyboard code set
402 Format: <int> (2 = AT (default), 3 = PS/2)
404 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
407 atkbd.softraw= [HW] Choose between synthetic and real raw mode
408 Format: <bool> (0 = real, 1 = synthetic (default))
410 atkbd.softrepeat= [HW]
411 Use software keyboard repeat
413 baycom_epp= [HW,AX25]
416 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
418 See header of drivers/net/hamradio/baycom_par.c.
420 baycom_ser_fdx= [HW,AX25]
421 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
422 Format: <io>,<irq>,<mode>[,<baud>]
423 See header of drivers/net/hamradio/baycom_ser_fdx.c.
425 baycom_ser_hdx= [HW,AX25]
426 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
427 Format: <io>,<irq>,<mode>
428 See header of drivers/net/hamradio/baycom_ser_hdx.c.
430 boot_delay= Milliseconds to delay each printk during boot.
431 Values larger than 10 seconds (10000) are changed to
435 bootmem_debug [KNL] Enable bootmem allocator debug messages.
437 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
438 bttv.radio= Most important insmod options are available as
440 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
443 bulk_remove=off [PPC] This parameter disables the use of the pSeries
444 firmware feature for flushing multiple hpte entries
447 c101= [NET] Moxa C101 synchronous serial card
449 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
450 Sometimes CPU hardware bugs make them report the cache
451 size incorrectly. The kernel will attempt work arounds
452 to fix known problems, but for some CPUs it is not
453 possible to determine what the correct size should be.
454 This option provides an override for these situations.
456 ccw_timeout_log [S390]
457 See Documentation/s390/CommonIO for details.
459 cgroup_disable= [KNL] Disable a particular controller
460 Format: {name of the controller(s) to disable}
461 {Currently supported controllers - "memory"}
463 checkreqprot [SELINUX] Set initial checkreqprot flag value.
464 Format: { "0" | "1" }
465 See security/selinux/Kconfig help text.
466 0 -- check protection applied by kernel (includes
467 any implied execute protection).
468 1 -- check protection requested by application.
469 Default value is set via a kernel config option.
470 Value can be changed at runtime via
471 /selinux/checkreqprot.
474 See Documentation/s390/CommonIO for details.
476 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
478 Forces specified clocksource (if available) to be used
479 when calculating gettimeofday(). If specified
480 clocksource is not available, it defaults to PIT.
481 Format: { pit | tsc | cyclone | pmtmr }
483 clocksource= Override the default clocksource
485 Override the default clocksource and use the clocksource
486 with the name specified.
487 Some clocksource names to choose from, depending on
489 [all] jiffies (this is the base, fallback clocksource)
491 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
492 pxa_timer,timer3,32k_counter,timer0_1
494 [X86-32] pit,hpet,tsc;
495 scx200_hrt on Geode; cyclone on IBM x440
503 clearcpuid=BITNUM [X86]
504 Disable CPUID feature X for the kernel. See
505 arch/x86/include/asm/cpufeature.h for the valid bit
506 numbers. Note the Linux specific bits are not necessarily
507 stable over kernel options, but the vendor specific
509 Also note that user programs calling CPUID directly
510 or using the feature without checking anything
511 will still see it. This just prevents it from
512 being used by the kernel or shown in /proc/cpuinfo.
513 Also note the kernel might malfunction if you disable
517 Sets the size of kernel global memory area for contiguous
518 memory allocations. For more information, see
519 include/linux/dma-contiguous.h
521 cmo_free_hint= [PPC] Format: { yes | no }
522 Specify whether pages are marked as being inactive
523 when they are freed. This is used in CMO environments
524 to determine OS memory pressure for page stealing by
528 coherent_pool=nn[KMG] [ARM,KNL]
529 Sets the size of memory pool for coherent, atomic dma
530 allocations, by default set to 256K.
532 code_bytes [X86] How many bytes of object code to print
537 com20020= [HW,NET] ARCnet - COM20020 chipset
539 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
541 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
545 ARCnet - COM90xx chipset (memory-mapped buffers)
546 Format: <io>[,<irq>[,<memstart>]]
548 condev= [HW,S390] console device
551 console= [KNL] Output console device and options.
553 tty<n> Use the virtual console device <n>.
557 Use the specified serial port. The options are of
558 the form "bbbbpnf", where "bbbb" is the baud rate,
559 "p" is parity ("n", "o", or "e"), "n" is number of
560 bits, and "f" is flow control ("r" for RTS or
561 omit it). Default is "9600n8".
563 See Documentation/serial-console.txt for more
565 Documentation/networking/netconsole.txt for an
568 uart[8250],io,<addr>[,options]
569 uart[8250],mmio,<addr>[,options]
570 Start an early, polled-mode console on the 8250/16550
571 UART at the specified I/O port or MMIO address,
572 switching to the matching ttyS device later. The
573 options are the same as for ttyS, above.
574 hvc<n> Use the hypervisor console device <n>. This is for
575 both Xen and PowerPC hypervisors.
577 If the device connected to the port is not a TTY but a braille
578 device, prepend "brl," before the device type, for instance
580 For now, only VisioBraille is supported.
582 consoleblank= [KNL] The console blank (screen saver) timeout in
583 seconds. Defaults to 10*60 = 10mins. A value of 0
584 disables the blank timer.
587 [KNL] Change the default value for
588 /proc/<pid>/coredump_filter.
589 See also Documentation/filesystems/proc.txt.
591 cpuidle.off=1 [CPU_IDLE]
592 disable the cpuidle sub-system
594 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
596 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
598 crashkernel=size[KMG][@offset[KMG]]
599 [KNL] Using kexec, Linux can switch to a 'crash kernel'
600 upon panic. This parameter reserves the physical
601 memory region [offset, offset + size] for that kernel
602 image. If '@offset' is omitted, then a suitable offset
603 is selected automatically. Check
604 Documentation/kdump/kdump.txt for further details.
606 crashkernel_low=size[KMG]
607 [KNL, x86] parts under 4G.
609 crashkernel=range1:size1[,range2:size2,...][@offset]
610 [KNL] Same as above, but depends on the memory
611 in the running system. The syntax of range is
612 start-[end] where start and end are both
613 a memory unit (amount[KMG]). See also
614 Documentation/kdump/kdump.txt for an example.
619 cs89x0_media= [HW,NET]
620 Format: { rj45 | aui | bnc }
623 See header of drivers/s390/block/dasd_devmap.c.
625 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
626 (one device per port)
627 Format: <port#>,<type>
628 See also Documentation/input/joystick-parport.txt
630 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
631 time. See Documentation/dynamic-debug-howto.txt for
632 details. Deprecated, see dyndbg.
634 debug [KNL] Enable kernel debugging (events log level).
637 [KNL] verbose self-tests
639 Print debugging info while doing the locking API
641 We default to 0 (no extra messages), setting it to
642 1 will print _a lot_ more information - normally
643 only useful to kernel developers.
645 debug_objects [KNL] Enable object debugging
648 [KNL] Disable object debugging
650 debug_guardpage_minorder=
651 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
652 parameter allows control of the order of pages that will
653 be intentionally kept free (and hence protected) by the
654 buddy allocator. Bigger value increase the probability
655 of catching random memory corruption, but reduce the
656 amount of memory for normal system use. The maximum
657 possible value is MAX_ORDER/2. Setting this parameter
658 to 1 or 2 should be enough to identify most random
659 memory corruption problems caused by bugs in kernel or
660 driver code when a CPU writes to (or reads from) a
661 random memory location. Note that there exists a class
662 of memory corruptions problems caused by buggy H/W or
663 F/W or by drivers badly programing DMA (basically when
664 memory is written at bus level and the CPU MMU is
665 bypassed) which are not detectable by
666 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
667 tracking down these problems.
669 debugpat [X86] Enable PAT debugging
671 decnet.addr= [HW,NET]
672 Format: <area>[,<node>]
673 See also Documentation/networking/decnet.txt.
676 [same as hugepagesz=] The size of the default
677 HugeTLB page size. This is the size represented by
678 the legacy /proc/ hugepages APIs, used for SHM, and
679 default size when mounting hugetlbfs filesystems.
680 Defaults to the default architecture's huge page size
684 Set number of hash buckets for dentry cache.
687 IO parameters + enable/disable command.
689 digiepca= [HW,SERIAL]
690 See drivers/char/README.epca and
691 Documentation/serial/digiepca.txt.
694 See Documentation/networking/ipv6.txt.
696 disable_ddw [PPC/PSERIES]
697 Disable Dynamic DMA Window support. Use this if
698 to workaround buggy firmware.
701 See Documentation/networking/ipv6.txt.
703 disable_mtrr_cleanup [X86]
704 The kernel tries to adjust MTRR layout from continuous
705 to discrete, to make X server driver able to add WB
706 entry later. This parameter disables that.
708 disable_mtrr_trim [X86, Intel and AMD only]
709 By default the kernel will trim any uncacheable
710 memory out of your available memory pool based on
711 MTRR settings. This parameter disables that behavior,
712 possibly causing your machine to run very slowly.
714 disable_timer_pin_1 [X86]
715 Disable PIN 1 of APIC timer
716 Can be useful to work around chipset bugs.
718 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
719 this option disables the debugging code at boot.
721 dma_debug_entries=<number>
722 This option allows to tune the number of preallocated
723 entries for DMA-API debugging code. One entry is
724 required per DMA-API allocation. Use this if the
725 DMA-API debugging code disables itself because the
726 architectural default is too low.
728 dma_debug_driver=<driver_name>
729 With this option the DMA-API debugging driver
730 filter feature can be enabled at boot time. Just
731 pass the driver to filter for as the parameter.
732 The filter can be disabled or changed to another
733 driver later using sysfs.
735 drm_kms_helper.edid_firmware=[<connector>:]<file>
736 Broken monitors, graphic adapters and KVMs may
737 send no or incorrect EDID data sets. This parameter
738 allows to specify an EDID data set in the
739 /lib/firmware directory that is used instead.
740 Generic built-in EDID data sets are used, if one of
741 edid/1024x768.bin, edid/1280x1024.bin,
742 edid/1680x1050.bin, or edid/1920x1080.bin is given
743 and no file with the same name exists. Details and
744 instructions how to build your own EDID data are
745 available in Documentation/EDID/HOWTO.txt. An EDID
746 data set will only be used for a particular connector,
747 if its name and a colon are prepended to the EDID
752 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
753 module.dyndbg[="val"]
754 Enable debug messages at boot time. See
755 Documentation/dynamic-debug-howto.txt for details.
757 earlycon= [KNL] Output early console device and options.
758 uart[8250],io,<addr>[,options]
759 uart[8250],mmio,<addr>[,options]
760 uart[8250],mmio32,<addr>[,options]
761 Start an early, polled-mode console on the 8250/16550
762 UART at the specified I/O port or MMIO address.
763 MMIO inter-register address stride is either 8-bit
764 (mmio) or 32-bit (mmio32).
765 The options are the same as for ttyS, above.
767 earlyprintk= [X86,SH,BLACKFIN]
770 earlyprintk=serial[,ttySn[,baudrate]]
771 earlyprintk=ttySn[,baudrate]
772 earlyprintk=dbgp[debugController#]
774 Append ",keep" to not disable it when the real console
777 Only vga or serial or usb debug port at a time.
779 Currently only ttyS0 and ttyS1 are supported.
781 Interaction with the standard serial driver is not
784 The VGA output is eventually overwritten by the real
787 The xen output can only be used by Xen PV guests.
789 ekgdboc= [X86,KGDB] Allow early kernel console debugging
792 This is designed to be used in conjunction with
793 the boot argument: earlyprintk=vga
796 Format: {"off" | "on" | "skip[mbr]"}
798 eisa_irq_edge= [PARISC,HW]
799 See header of drivers/parisc/eisa.c.
802 See comment before function elanfreq_setup() in
803 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
806 Format: {"cfq" | "deadline" | "noop"}
807 See Documentation/block/cfq-iosched.txt and
808 Documentation/block/deadline-iosched.txt for details.
810 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
811 Specifies physical address of start of kernel core
812 image elf header and optionally the size. Generally
813 kexec loader will pass this option to capture kernel.
814 See Documentation/kdump/kdump.txt for details.
816 enable_mtrr_cleanup [X86]
817 The kernel tries to adjust MTRR layout from continuous
818 to discrete, to make X server driver able to add WB
819 entry later. This parameter enables that.
821 enable_timer_pin_1 [X86]
822 Enable PIN 1 of APIC timer
823 Can be useful to work around chipset bugs
824 (in particular on some ATI chipsets).
825 The kernel tries to set a reasonable default.
827 enforcing [SELINUX] Set initial enforcing status.
829 See security/selinux/Kconfig help text.
830 0 -- permissive (log only, no denials).
831 1 -- enforcing (deny and log).
833 Value can be changed at runtime via /selinux/enforce.
836 Disable Error Record Serialization Table (ERST)
839 ether= [HW,NET] Ethernet cards parameters
840 This option is obsoleted by the "netdev=" option, which
841 has equivalent usage. See its documentation for details.
845 Permit 'security.evm' to be updated regardless of
846 current integrity status.
850 fail_make_request=[KNL]
851 General fault injection mechanism.
852 Format: <interval>,<probability>,<space>,<times>
853 See also Documentation/fault-injection/.
856 See Documentation/blockdev/floppy.txt.
858 force_pal_cache_flush
859 [IA-64] Avoid check_sal_cache_flush which may hang on
860 buggy SAL_CACHE_FLUSH implementations. Using this
861 parameter will force ia64_sal_cache_flush to call
862 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
865 [FTRACE] will set and start the specified tracer
866 as early as possible in order to facilitate early
869 ftrace_dump_on_oops[=orig_cpu]
870 [FTRACE] will dump the trace buffers on oops.
871 If no parameter is passed, ftrace will dump
872 buffers of all CPUs, but if you pass orig_cpu, it will
873 dump only the buffer of the CPU that triggered the
876 ftrace_filter=[function-list]
877 [FTRACE] Limit the functions traced by the function
878 tracer at boot up. function-list is a comma separated
879 list of functions. This list can be changed at run
880 time by the set_ftrace_filter file in the debugfs
883 ftrace_notrace=[function-list]
884 [FTRACE] Do not trace the functions specified in
885 function-list. This list can be changed at run time
886 by the set_ftrace_notrace file in the debugfs
889 ftrace_graph_filter=[function-list]
890 [FTRACE] Limit the top level callers functions traced
891 by the function graph tracer at boot up.
892 function-list is a comma separated list of functions
893 that can be changed at run time by the
894 set_graph_function file in the debugfs tracing directory.
897 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
898 support via parallel port (up to 5 devices per port)
899 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
900 See also Documentation/input/joystick-parport.txt
904 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
908 gcov_persist= [GCOV] When non-zero (default), profiling data for
909 kernel modules is saved and remains accessible via
910 debugfs, even when the module is unloaded/reloaded.
911 When zero, profiling data is discarded and associated
912 debugfs files are removed at module unload time.
914 gpt [EFI] Forces disk with valid GPT signature but
915 invalid Protective MBR to be treated as GPT.
917 grcan.enable0= [HW] Configuration of physical interface 0. Determines
918 the "Enable 0" bit of the configuration register.
921 grcan.enable1= [HW] Configuration of physical interface 1. Determines
922 the "Enable 0" bit of the configuration register.
925 grcan.select= [HW] Select which physical interface to use.
928 grcan.txsize= [HW] Sets the size of the tx buffer.
929 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
931 grcan.rxsize= [HW] Sets the size of the rx buffer.
932 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
935 hashdist= [KNL,NUMA] Large hashes allocated during boot
936 are distributed across NUMA nodes. Defaults on
937 for 64-bit NUMA, off otherwise.
938 Format: 0 | 1 (for off | on)
940 hcl= [IA-64] SGI's Hardware Graph compatibility layer
942 hd= [EIDE] (E)IDE hard drive subsystem geometry
943 Format: <cyl>,<head>,<sect>
946 Disable Hardware Error Source Table (HEST) support;
947 corresponding firmware-first mode error processing
948 logic will be disabled.
950 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
951 size of <nn>. This works even on boxes that have no
952 highmem otherwise. This also works to reduce highmem
953 size on bigger boxes.
955 highres= [KNL] Enable/disable high resolution timer mode.
956 Valid parameters: "on", "off"
960 See Documentation/isdn/README.HiSax.
964 hpet= [X86-32,HPET] option to control HPET usage
965 Format: { enable (default) | disable | force |
967 disable: disable HPET and use PIT instead
968 force: allow force enabled of undocumented chips (ICH4,
970 verbose: show contents of HPET registers during setup
972 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
973 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
974 On x86-64 and powerpc, this option can be specified
975 multiple times interleaved with hugepages= to reserve
976 huge pages of different sizes. Valid pages sizes on
977 x86-64 are 2M (when the CPU supports "pse") and 1G
978 (when the CPU supports the "pdpe1gb" cpuinfo flag)
979 Note that 1GB pages can only be allocated at boot time
980 using hugepages= and not freed afterwards.
982 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
983 terminal devices. Valid values: 0..8
984 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
985 If specified, z/VM IUCV HVC accepts connections
986 from listed z/VM user IDs only.
988 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
989 hardware thread id mappings.
990 Format: <cpu>:<hwthread>
993 Do not unregister boot console at start. This is only
994 useful for debugging when something happens in the window
995 between unregistering the boot console and initializing
998 i2c_bus= [HW] Override the default board specific I2C bus speed
999 or register an additional I2C bus that is not
1000 registered from board initialization code.
1004 i8042.debug [HW] Toggle i8042 debug mode
1005 i8042.direct [HW] Put keyboard port into non-translated mode
1006 i8042.dumbkbd [HW] Pretend that controller can only read data from
1007 keyboard and cannot control its state
1008 (Don't attempt to blink the leds)
1009 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1010 i8042.nokbd [HW] Don't check/create keyboard port
1011 i8042.noloop [HW] Disable the AUX Loopback command while probing
1013 i8042.nomux [HW] Don't check presence of an active multiplexing
1015 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1017 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1018 i8042.reset [HW] Reset the controller during init and cleanup
1019 i8042.unlock [HW] Unlock (ignore) the keylock
1023 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1024 indicates that the driver is running on unsupported
1026 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1027 does not match list of supported models.
1029 [HW] Report power status in /proc/i8k
1030 (disabled by default)
1031 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1034 i915.invert_brightness=
1035 [DRM] Invert the sense of the variable that is used to
1036 set the brightness of the panel backlight. Normally a
1037 brightness value of 0 indicates backlight switched off,
1038 and the maximum of the brightness value sets the backlight
1039 to maximum brightness. If this parameter is set to 0
1040 (default) and the machine requires it, or this parameter
1041 is set to 1, a brightness value of 0 sets the backlight
1042 to maximum brightness, and the maximum of the brightness
1043 value switches the backlight off.
1044 -1 -- never invert brightness
1045 0 -- machine default
1046 1 -- force brightness inversion
1049 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1051 ide-core.nodma= [HW] (E)IDE subsystem
1052 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1053 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1054 .cdrom .chs .ignore_cable are additional options
1055 See Documentation/ide/ide.txt.
1057 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1058 Claim all unknown PCI IDE storage controllers.
1061 Format: idle=poll, idle=halt, idle=nomwait
1062 Poll forces a polling idle loop that can slightly
1063 improve the performance of waking up a idle CPU, but
1064 will use a lot of power and make the system run hot.
1066 idle=halt: Halt is forced to be used for CPU idle.
1067 In such case C2/C3 won't be used again.
1068 idle=nomwait: Disable mwait for CPU C-states
1070 ignore_loglevel [KNL]
1071 Ignore loglevel setting - this will print /all/
1072 kernel messages to the console. Useful for debugging.
1073 We also add it as printk module parameter, so users
1074 could change it dynamically, usually by
1075 /sys/module/printk/parameters/ignore_loglevel.
1077 ihash_entries= [KNL]
1078 Set number of hash buckets for inode cache.
1080 ima_appraise= [IMA] appraise integrity measurements
1081 Format: { "off" | "enforce" | "fix" }
1084 ima_appraise_tcb [IMA]
1085 The builtin appraise policy appraises all files
1089 Format: { "0" | "1" }
1090 0 -- integrity auditing messages. (Default)
1091 1 -- enable informational integrity auditing messages.
1094 Format: { "sha1" | "md5" }
1098 Load a policy which meets the needs of the Trusted
1099 Computing Base. This means IMA will measure all
1100 programs exec'd, files mmap'd for exec, and all files
1101 opened for read by uid=0.
1105 Run specified binary instead of /sbin/init as init
1108 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1109 for working out where the kernel is dying during
1112 initrd= [BOOT] Specify the location of the initial ramdisk
1114 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1117 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1119 Enable intel iommu driver.
1121 Disable intel iommu driver.
1122 igfx_off [Default Off]
1123 By default, gfx is mapped as normal device. If a gfx
1124 device has a dedicated DMAR unit, the DMAR unit is
1125 bypassed by not enabling DMAR with this option. In
1126 this case, gfx device will use physical address for
1129 With this option iommu will not optimize to look
1130 for io virtual address below 32-bit forcing dual
1131 address cycle on pci bus for cards supporting greater
1132 than 32-bit addressing. The default is to look
1133 for translation below 32-bit and if not available
1134 then look in the higher range.
1135 strict [Default Off]
1136 With this option on every unmap_single operation will
1137 result in a hardware IOTLB flush operation as opposed
1138 to batching them for performance.
1139 sp_off [Default Off]
1140 By default, super page will be supported if Intel IOMMU
1141 has the capability. With this option, super page will
1144 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1145 0 disables intel_idle and fall back on acpi_idle.
1146 1 to 6 specify maximum depth of C-state.
1150 Do not enable intel_pstate as the default
1151 scaling driver for the supported processors
1153 intremap= [X86-64, Intel-IOMMU]
1154 on enable Interrupt Remapping (default)
1155 off disable Interrupt Remapping
1156 nosid disable Source ID checking
1158 BIOS x2APIC opt-out request will be ignored
1160 iomem= Disable strict checking of access to MMIO memory
1161 strict regions from userspace.
1178 io7= [HW] IO7 for Marvel based alpha systems
1179 See comment before marvel_specify_io7 in
1180 arch/alpha/kernel/core_marvel.c.
1182 io_delay= [X86] I/O delay method
1184 Standard port 0x80 based delay
1186 Alternate port 0xed based delay (needed on some systems)
1188 Simple two microseconds delay
1193 See Documentation/filesystems/nfs/nfsroot.txt.
1195 ip2= [HW] Set IO/IRQ pairs for up to 4 IntelliPort boards
1196 See comment before ip2_setup() in
1197 drivers/char/ip2/ip2base.c.
1200 When an interrupt is not handled search all handlers
1201 for it. Intended to get systems with badly broken
1205 When an interrupt is not handled search all handlers
1206 for it. Also check all handlers each timer
1207 interrupt. Intended to get systems with badly broken
1211 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1213 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1215 <cpu number>,...,<cpu number>
1217 <cpu number>-<cpu number>
1218 (must be a positive range in ascending order)
1220 <cpu number>,...,<cpu number>-<cpu number>
1222 This option can be used to specify one or more CPUs
1223 to isolate from the general SMP balancing and scheduling
1224 algorithms. You can move a process onto or off an
1225 "isolated" CPU via the CPU affinity syscalls or cpuset.
1226 <cpu number> begins at 0 and the maximum value is
1227 "number of CPUs in system - 1".
1229 This option is the preferred way to isolate CPUs. The
1230 alternative -- manually setting the CPU mask of all
1231 tasks in the system -- can cause problems and
1232 suboptimal load balancer performance.
1236 js= [HW,JOY] Analog joystick
1237 See Documentation/input/joystick.txt.
1241 kernelcore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1242 specifies the amount of memory usable by the kernel
1243 for non-movable allocations. The requested amount is
1244 spread evenly throughout all nodes in the system. The
1245 remaining memory in each node is used for Movable
1246 pages. In the event, a node is too small to have both
1247 kernelcore and Movable pages, kernelcore pages will
1248 take priority and other nodes will have a larger number
1249 of kernelcore pages. The Movable zone is used for the
1250 allocation of pages that may be reclaimed or moved
1251 by the page migration subsystem. This means that
1252 HugeTLB pages may not be allocated from this zone.
1253 Note that allocations like PTEs-from-HighMem still
1254 use the HighMem zone if it exists, and the Normal
1255 zone if it does not.
1257 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1258 Format: <Controller#>[,poll interval]
1259 The controller # is the number of the ehci usb debug
1260 port as it is probed via PCI. The poll interval is
1261 optional and is the number seconds in between
1262 each poll cycle to the debug port in case you need
1263 the functionality for interrupting the kernel with
1264 gdb or control-c on the dbgp connection. When
1265 not using this parameter you use sysrq-g to break into
1266 the kernel debugger.
1268 kgdboc= [KGDB,HW] kgdb over consoles.
1269 Requires a tty driver that supports console polling,
1270 or a supported polling keyboard driver (non-usb).
1271 Serial only format: <serial_device>[,baud]
1272 keyboard only format: kbd
1273 keyboard and serial format: kbd,<serial_device>[,baud]
1274 Optional Kernel mode setting:
1275 kms, kbd format: kms,kbd
1276 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1278 kgdbwait [KGDB] Stop kernel execution and enter the
1279 kernel debugger at the earliest opportunity.
1281 kmac= [MIPS] korina ethernet MAC address.
1282 Configure the RouterBoard 532 series on-chip
1283 Ethernet adapter MAC address.
1285 kmemleak= [KNL] Boot-time kmemleak enable/disable
1286 Valid arguments: on, off
1289 kstack=N [X86] Print N words from the kernel stack
1292 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1293 Default is 0 (don't ignore, but inject #GP)
1295 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1299 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1300 Default is 1 (enabled)
1302 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1304 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1306 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1307 (virtualized MMU) support on capable Intel chips.
1308 Default is 1 (enabled)
1310 kvm-intel.emulate_invalid_guest_state=
1311 [KVM,Intel] Enable emulation of invalid guest states
1312 Default is 0 (disabled)
1314 kvm-intel.flexpriority=
1315 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1316 Default is 1 (enabled)
1319 [KVM,Intel] Enable VMX nesting (nVMX).
1320 Default is 0 (disabled)
1322 kvm-intel.unrestricted_guest=
1323 [KVM,Intel] Disable unrestricted guest feature
1324 (virtualized real and unpaged mode) on capable
1325 Intel chips. Default is 1 (enabled)
1327 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1328 feature (tagged TLBs) on capable Intel chips.
1329 Default is 1 (enabled)
1335 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1338 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1339 value for LAPIC timer one-shot implementation. Default
1340 back to the programmable timer unit in the LAPIC.
1342 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1345 libata.dma= [LIBATA] DMA control
1346 libata.dma=0 Disable all PATA and SATA DMA
1347 libata.dma=1 PATA and SATA Disk DMA only
1348 libata.dma=2 ATAPI (CDROM) DMA only
1349 libata.dma=4 Compact Flash DMA only
1350 Combinations also work, so libata.dma=3 enables DMA
1351 for disks and CDROMs, but not CFs.
1353 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1354 libata.ignore_hpa=0 keep BIOS limits (default)
1355 libata.ignore_hpa=1 ignore limits, using full disk
1357 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
1361 libata.force= [LIBATA] Force configurations. The format is comma
1362 separated list of "[ID:]VAL" where ID is
1363 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
1364 matching port, link or device. Basically, it matches
1365 the ATA ID string printed on console by libata. If
1366 the whole ID part is omitted, the last PORT and DEVICE
1367 values are used. If ID hasn't been specified yet, the
1368 configuration applies to all ports, links and devices.
1370 If only DEVICE is omitted, the parameter applies to
1371 the port and all links and devices behind it. DEVICE
1372 number of 0 either selects the first device or the
1373 first fan-out link behind PMP device. It does not
1374 select the host link. DEVICE number of 15 selects the
1375 host link and device attached to it.
1377 The VAL specifies the configuration to force. As long
1378 as there's no ambiguity shortcut notation is allowed.
1379 For example, both 1.5 and 1.5G would work for 1.5Gbps.
1380 The following configurations can be forced.
1382 * Cable type: 40c, 80c, short40c, unk, ign or sata.
1383 Any ID with matching PORT is used.
1385 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
1387 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
1388 udma[/][16,25,33,44,66,100,133] notation is also
1391 * [no]ncq: Turn on or off NCQ.
1393 * nohrst, nosrst, norst: suppress hard, soft
1396 * rstonce: only attempt one reset during
1397 hot-unplug link recovery
1399 * dump_id: dump IDENTIFY data.
1401 If there are multiple matching configurations changing
1402 the same attribute, the last one is used.
1404 memblock=debug [KNL] Enable memblock debug messages.
1406 load_ramdisk= [RAM] List of ramdisks to load from floppy
1407 See Documentation/blockdev/ramdisk.txt.
1409 lockd.nlm_grace_period=P [NFS] Assign grace period.
1412 lockd.nlm_tcpport=N [NFS] Assign TCP port.
1415 lockd.nlm_timeout=T [NFS] Assign timeout value.
1418 lockd.nlm_udpport=M [NFS] Assign UDP port.
1421 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
1424 loglevel= All Kernel Messages with a loglevel smaller than the
1425 console loglevel will be printed to the console. It can
1426 also be changed with klogd or other programs. The
1427 loglevels are defined as follows:
1429 0 (KERN_EMERG) system is unusable
1430 1 (KERN_ALERT) action must be taken immediately
1431 2 (KERN_CRIT) critical conditions
1432 3 (KERN_ERR) error conditions
1433 4 (KERN_WARNING) warning conditions
1434 5 (KERN_NOTICE) normal but significant condition
1435 6 (KERN_INFO) informational
1436 7 (KERN_DEBUG) debug-level messages
1438 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
1439 in bytes. n must be a power of two. The default
1440 size is set in the kernel config file.
1442 logo.nologo [FB] Disables display of the built-in Linux logo.
1443 This may be used to provide more screen space for
1444 kernel log messages and is useful when debugging
1445 kernel boot problems.
1447 lp=0 [LP] Specify parallel ports to use, e.g,
1448 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
1449 lp=reset first parallel port). 'lp=0' disables the
1450 lp=auto printer driver. 'lp=reset' (which can be
1451 specified in addition to the ports) causes
1452 attached printers to be reset. Using
1453 lp=port1,port2,... specifies the parallel ports
1454 to associate lp devices with, starting with
1455 lp0. A port specification may be 'none' to skip
1456 that lp device, or a parport name such as
1457 'parport0'. Specifying 'lp=auto' instead of a
1458 port specification list means that device IDs
1459 from each port should be examined, to see if
1460 an IEEE 1284-compliant printer is attached; if
1461 so, the driver will manage that printer.
1462 See also header of drivers/char/lp.c.
1465 Sets loops_per_jiffy to given constant, thus avoiding
1466 time-consuming boot-time autodetection (up to 250 ms per
1467 CPU). 0 enables autodetection (default). To determine
1468 the correct value for your kernel, boot with normal
1469 autodetection and see what value is printed. Note that
1470 on SMP systems the preset will be applied to all CPUs,
1471 which is likely to cause problems if your CPUs need
1472 significantly divergent settings. An incorrect value
1473 will cause delays in the kernel to be wrong, leading to
1474 unpredictable I/O errors and other breakage. Although
1475 unlikely, in the extreme case this might damage your
1479 Format: <io>,<irq>,<dma>
1481 machvec= [IA-64] Force the use of a particular machine-vector
1482 (machvec) in a generic kernel.
1483 Example: machvec=hpzx1_swiotlb
1485 machtype= [Loongson] Share the same kernel image file between different
1487 Example: machtype=lemote-yeeloong-2f-7inch
1489 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
1490 than or equal to this physical address is ignored.
1492 maxcpus= [SMP] Maximum number of processors that an SMP kernel
1493 should make use of. maxcpus=n : n >= 0 limits the
1494 kernel to using 'n' processors. n=0 is a special case,
1495 it is equivalent to "nosmp", which also disables
1498 max_loop= [LOOP] The number of loop block devices that get
1499 (loop.max_loop) unconditionally pre-created at init time. The default
1500 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
1501 of statically allocating a predefined number, loop
1502 devices can be requested on-demand with the
1503 /dev/loop-control interface.
1505 mce [X86-32] Machine Check Exception
1507 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
1509 md= [HW] RAID subsystems devices and level
1510 See Documentation/md.txt.
1513 Format: <first>,<last>
1514 Specifies range of consoles to be captured by the MDA.
1516 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
1517 Amount of memory to be used when the kernel is not able
1518 to see the whole system memory or for test.
1519 [X86] Work as limiting max address. Use together
1520 with memmap= to avoid physical address space collisions.
1521 Without memmap= PCI devices could be placed at addresses
1522 belonging to unused RAM.
1524 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
1528 [KNL,SH] Allow user to override the default size for
1529 per-device physically contiguous DMA buffers.
1531 memmap=exactmap [KNL,X86] Enable setting of an exact
1532 E820 memory map, as specified by the user.
1533 Such memmap=exactmap lines can be constructed based on
1534 BIOS output or other requirements. See the memmap=nn@ss
1537 memmap=nn[KMG]@ss[KMG]
1538 [KNL] Force usage of a specific region of memory
1539 Region of memory to be used, from ss to ss+nn.
1541 memmap=nn[KMG]#ss[KMG]
1542 [KNL,ACPI] Mark specific memory as ACPI data.
1543 Region of memory to be used, from ss to ss+nn.
1545 memmap=nn[KMG]$ss[KMG]
1546 [KNL,ACPI] Mark specific memory as reserved.
1547 Region of memory to be used, from ss to ss+nn.
1548 Example: Exclude memory from 0x18690000-0x1869ffff
1549 memmap=64K$0x18690000
1551 memmap=0x10000$0x18690000
1553 memory_corruption_check=0/1 [X86]
1554 Some BIOSes seem to corrupt the first 64k of
1555 memory when doing things like suspend/resume.
1556 Setting this option will scan the memory
1557 looking for corruption. Enabling this will
1558 both detect corruption and prevent the kernel
1559 from using the memory being corrupted.
1560 However, its intended as a diagnostic tool; if
1561 repeatable BIOS-originated corruption always
1562 affects the same memory, you can use memmap=
1563 to prevent the kernel from using that memory.
1565 memory_corruption_check_size=size [X86]
1566 By default it checks for corruption in the low
1567 64k, making this memory unavailable for normal
1568 use. Use this parameter to scan for
1569 corruption in more or less memory.
1571 memory_corruption_check_period=seconds [X86]
1572 By default it checks for corruption every 60
1573 seconds. Use this parameter to check at some
1574 other rate. 0 disables periodic checking.
1576 memtest= [KNL,X86] Enable memtest
1578 default : 0 <disable>
1579 Specifies the number of memtest passes to be
1580 performed. Each pass selects another test
1581 pattern from a given set of patterns. Memtest
1582 fills the memory with this pattern, validates
1583 memory contents and reserves bad memory
1584 regions that are detected.
1586 meye.*= [HW] Set MotionEye Camera parameters
1587 See Documentation/video4linux/meye.txt.
1589 mfgpt_irq= [IA-32] Specify the IRQ to use for the
1590 Multi-Function General Purpose Timers on AMD Geode
1593 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
1594 the BIOS has incorrectly applied a workaround. TinyBIOS
1595 version 0.98 is known to be affected, 0.99 fixes the
1596 problem by letting the user disable the workaround.
1600 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
1601 physical address is ignored.
1603 mini2440= [ARM,HW,KNL]
1604 Format:[0..2][b][c][t]
1606 MINI2440 configuration specification:
1607 0 - The attached screen is the 3.5" TFT
1608 1 - The attached screen is the 7" TFT
1609 2 - The VGA Shield is attached (1024x768)
1610 Leaving out the screen size parameter will not load
1611 the TFT driver, and the framebuffer will be left
1613 b - Enable backlight. The TFT backlight pin will be
1614 linked to the kernel VESA blanking code and a GPIO
1615 LED. This parameter is not necessary when using the
1617 c - Enable the s3c camera interface.
1618 t - Reserved for enabling touchscreen support. The
1619 touchscreen support is not enabled in the mainstream
1620 kernel as of 2.6.30, a preliminary port can be found
1621 in the "bleeding edge" mini2440 support kernel at
1622 http://repo.or.cz/w/linux-2.6/mini2440.git
1625 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
1626 parameter allows control of the logging verbosity for
1627 the additional memory initialisation checks. A value
1628 of 0 disables mminit logging and a level of 4 will
1629 log everything. Information is printed at KERN_DEBUG
1630 so loglevel=8 may also need to be specified.
1633 [KNL] When CONFIG_MODULE_SIG is set, this means that
1634 modules without (valid) signatures will fail to load.
1635 Note that if CONFIG_MODULE_SIG_ENFORCE is set, that
1636 is always true, so this option does nothing.
1639 [MOUSE] Maximum time between finger touching and
1640 leaving touchpad surface for touch to be considered
1641 a tap and be reported as a left button click (for
1642 touchpads working in absolute mode only).
1644 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
1645 reporting absolute coordinates, such as tablets
1646 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
1647 reporting absolute coordinates, such as tablets
1649 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1650 is similar to kernelcore except it specifies the
1651 amount of memory used for migratable allocations.
1652 If both kernelcore and movablecore is specified,
1653 then kernelcore will be at *least* the specified
1654 value but may be more. If movablecore on its own
1655 is specified, the administrator must be careful
1656 that the amount of memory usable for all allocations
1659 MTD_Partition= [MTD]
1660 Format: <name>,<region-number>,<size>,<offset>
1662 MTD_Region= [MTD] Format:
1663 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
1666 See drivers/mtd/cmdlinepart.c.
1668 multitce=off [PPC] This parameter disables the use of the pSeries
1669 firmware feature for updating multiple TCE entries
1672 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
1674 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
1676 boundary - index of last SLC block on Flex-OneNAND.
1677 The remaining blocks are configured as MLC blocks.
1678 lock - Configure if Flex-OneNAND boundary should be locked.
1679 Once locked, the boundary cannot be changed.
1680 1 indicates lock status, 0 indicates unlock status.
1683 ARM/S3C2412 JIVE boot control
1685 See arch/arm/mach-s3c2412/mach-jive.c
1687 mtouchusb.raw_coordinates=
1688 [HW] Make the MicroTouch USB driver use raw coordinates
1689 ('y', default) or cooked coordinates ('n')
1691 mtrr_chunk_size=nn[KMG] [X86]
1692 used for mtrr cleanup. It is largest continuous chunk
1693 that could hold holes aka. UC entries.
1695 mtrr_gran_size=nn[KMG] [X86]
1696 Used for mtrr cleanup. It is granularity of mtrr block.
1698 Large value could prevent small alignment from
1701 mtrr_spare_reg_nr=n [X86]
1703 Range: 0,7 : spare reg number
1705 Used for mtrr cleanup. It is spare mtrr entries number.
1706 Set to 2 or more if your graphical card needs more.
1708 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
1710 netdev= [NET] Network devices parameters
1711 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
1712 Note that mem_start is often overloaded to mean
1713 something different and driver-specific.
1714 This usage is only documented in each driver source
1718 [NETFILTER] Enable connection tracking flow accounting
1719 0 to disable accounting
1720 1 to enable accounting
1723 nfsaddrs= [NFS] Deprecated. Use ip= instead.
1724 See Documentation/filesystems/nfs/nfsroot.txt.
1726 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
1727 See Documentation/filesystems/nfs/nfsroot.txt.
1729 nfsrootdebug [NFS] enable nfsroot debugging messages.
1730 See Documentation/filesystems/nfs/nfsroot.txt.
1732 nfs.callback_tcpport=
1733 [NFS] set the TCP port on which the NFSv4 callback
1734 channel should listen.
1737 [NFS] sets the pathname to the program which is used
1738 to update the NFS client cache entries.
1740 nfs.cache_getent_timeout=
1741 [NFS] sets the timeout after which an attempt to
1742 update a cache entry is deemed to have failed.
1744 nfs.idmap_cache_timeout=
1745 [NFS] set the maximum lifetime for idmapper cache
1749 [NFS] enable 64-bit inode numbers.
1750 If zero, the NFS client will fake up a 32-bit inode
1751 number for the readdir() and stat() syscalls instead
1752 of returning the full 64-bit number.
1753 The default is to return 64-bit inode numbers.
1755 nfs.max_session_slots=
1756 [NFSv4.1] Sets the maximum number of session slots
1757 the client will attempt to negotiate with the server.
1758 This limits the number of simultaneous RPC requests
1759 that the client can send to the NFSv4.1 server.
1760 Note that there is little point in setting this
1761 value higher than the max_tcp_slot_table_limit.
1763 nfs.nfs4_disable_idmapping=
1764 [NFSv4] When set to the default of '1', this option
1765 ensures that both the RPC level authentication
1766 scheme and the NFS level operations agree to use
1767 numeric uids/gids if the mount is using the
1768 'sec=sys' security flavour. In effect it is
1769 disabling idmapping, which can make migration from
1770 legacy NFSv2/v3 systems to NFSv4 easier.
1771 Servers that do not support this mode of operation
1772 will be autodetected by the client, and it will fall
1773 back to using the idmapper.
1774 To turn off this behaviour, set the value to '0'.
1776 [NFS4] Specify an additional fixed unique ident-
1777 ification string that NFSv4 clients can insert into
1778 their nfs_client_id4 string. This is typically a
1779 UUID that is generated at system install time.
1781 nfs.send_implementation_id =
1782 [NFSv4.1] Send client implementation identification
1783 information in exchange_id requests.
1784 If zero, no implementation identification information
1786 The default is to send the implementation identification
1789 nfsd.nfs4_disable_idmapping=
1790 [NFSv4] When set to the default of '1', the NFSv4
1791 server will return only numeric uids and gids to
1792 clients using auth_sys, and will accept numeric uids
1793 and gids from such clients. This is intended to ease
1794 migration from NFSv2/v3.
1796 objlayoutdriver.osd_login_prog=
1797 [NFS] [OBJLAYOUT] sets the pathname to the program which
1798 is used to automatically discover and login into new
1799 osd-targets. Please see:
1800 Documentation/filesystems/pnfs.txt for more explanations
1802 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
1803 when a NMI is triggered.
1804 Format: [state][,regs][,debounce][,die]
1806 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
1807 Format: [panic,][nopanic,][num]
1809 0 - turn nmi_watchdog off
1810 When panic is specified, panic when an NMI watchdog
1811 timeout occurs (or 'nopanic' to override the opposite
1813 This is useful when you use a panic=... timeout and
1814 need the box quickly up again.
1816 netpoll.carrier_timeout=
1817 [NET] Specifies amount of time (in seconds) that
1818 netpoll should wait for a carrier. By default netpoll
1821 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
1822 emulation library even if a 387 maths coprocessor
1826 [HW] Never suspend the console
1827 Disable suspending of consoles during suspend and
1828 hibernate operations. Once disabled, debugging
1829 messages can reach various consoles while the rest
1830 of the system is being put to sleep (ie, while
1831 debugging driver suspend/resume hooks). This may
1832 not work reliably with all consoles, but is known
1833 to work with serial and VGA consoles.
1834 To facilitate more flexible debugging, we also add
1835 console_suspend, a printk module parameter to control
1836 it. Users could use console_suspend (usually
1837 /sys/module/printk/parameters/console_suspend) to
1838 turn on/off it dynamically.
1840 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
1841 caches in the slab allocator. Saves per-node memory,
1842 but will impact performance.
1846 noapic [SMP,APIC] Tells the kernel to not make use of any
1847 IOAPICs that may be present in the system.
1849 noautogroup Disable scheduler automatic task group creation.
1851 nobats [PPC] Do not use BATs for mapping kernel lowmem
1852 on "Classic" PPC cores.
1856 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
1858 nodelayacct [KNL] Disable per-task delay accounting
1860 nodisconnect [HW,SCSI,M68K] Disables SCSI disconnects.
1862 nodsp [SH] Disable hardware DSP at boot time.
1864 noefi [X86] Disable EFI runtime services support.
1869 On X86-32 available only on PAE configured kernels.
1870 noexec=on: enable non-executable mappings (default)
1871 noexec=off: disable non-executable mappings
1874 Disable SMAP (Supervisor Mode Access Prevention)
1875 even if it is supported by processor.
1878 Disable SMEP (Supervisor Mode Execution Prevention)
1879 even if it is supported by processor.
1882 This affects only 32-bit executables.
1883 noexec32=on: enable non-executable mappings (default)
1884 read doesn't imply executable mappings
1885 noexec32=off: disable non-executable mappings
1886 read implies executable mappings
1888 nofpu [SH] Disable hardware FPU at boot time.
1890 nofxsr [BUGS=X86-32] Disables x86 floating point extended
1891 register save and restore. The kernel will only save
1892 legacy floating-point registers on task switch.
1894 noxsave [BUGS=X86] Disables x86 extended register state save
1895 and restore using xsave. The kernel will fallback to
1896 enabling legacy floating-point and sse state.
1899 on enable eager fpu restore
1900 off disable eager fpu restore
1901 auto selects the default scheme, which automatically
1902 enables eagerfpu restore for xsaveopt.
1904 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
1905 wfi(ARM) instruction doesn't work correctly and not to
1906 use it. This is also useful when using JTAG debugger.
1908 no_file_caps Tells the kernel not to honor file capabilities. The
1909 only way then for a file to be executed with privilege
1910 is to be setuid root or executed by root.
1912 nohalt [IA-64] Tells the kernel not to use the power saving
1913 function PAL_HALT_LIGHT when idle. This increases
1914 power-consumption. On the positive side, it reduces
1915 interrupt wake-up latency, which may improve performance
1916 in certain environments such as networked servers or
1919 nohz= [KNL] Boottime enable/disable dynamic ticks
1920 Valid arguments: on, off
1923 noiotrap [SH] Disables trapped I/O port accesses.
1925 noirqdebug [X86-32] Disables the code which attempts to detect and
1926 disable unhandled interrupt sources.
1928 no_timer_check [X86,APIC] Disables the code which tests for
1929 broken timer IRQ sources.
1931 noisapnp [ISAPNP] Disables ISA PnP code.
1933 noinitrd [RAM] Tells the kernel not to load any configured
1936 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
1938 [Deprecated - use intremap=off]
1942 nojitter [IA-64] Disables jitter checking for ITC timers.
1944 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
1946 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
1949 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
1950 steal time is computed, but won't influence scheduler
1953 nolapic [X86-32,APIC] Do not enable or use the local APIC.
1955 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
1957 noltlbs [PPC] Do not use large page/tlb entries for kernel
1958 lowmem mapping on PPC40x.
1960 nomca [IA-64] Disable machine check abort handling
1962 nomce [X86-32] Machine Check Exception
1964 nomfgpt [X86-32] Disable Multi-Function General Purpose
1965 Timer usage (for AMD Geode machines).
1967 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
1968 shutdown the other cpus. Instead use the REBOOT_VECTOR
1971 nomodule Disable module load
1973 nopat [X86] Disable PAT (page attribute table extension of
1974 pagetables) support.
1976 norandmaps Don't use address space randomization. Equivalent to
1977 echo 0 > /proc/sys/kernel/randomize_va_space
1979 noreplace-paravirt [X86,IA-64,PV_OPS] Don't patch paravirt_ops
1981 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
1982 with UP alternatives
1984 noresidual [PPC] Don't use residual data on PReP machines.
1986 nordrand [X86] Disable the direct use of the RDRAND
1987 instruction even if it is supported by the
1988 processor. RDRAND is still available to user
1991 noresume [SWSUSP] Disables resume and restores original swap
1994 no-scroll [VGA] Disables scrollback.
1995 This is required for the Braillex ib80-piezo Braille
1996 reader made by F.H. Papenmeier (Germany).
2000 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2002 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2003 and disable the IO APIC. legacy for "maxcpus=0".
2005 nosoftlockup [KNL] Disable the soft-lockup detector.
2007 nosync [HW,M68K] Disables sync negotiation for all devices.
2009 notsc [BUGS=X86-32] Disable Time Stamp Counter
2011 nousb [USB] Disable the USB subsystem
2013 nowatchdog [KNL] Disable the lockup detector (NMI watchdog).
2017 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2019 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2020 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2021 Some features depend on CPU0. Known dependencies are:
2022 1. Resume from suspend/hibernate depends on CPU0.
2023 Suspend/hibernate will fail if CPU0 is offline and you
2024 need to online CPU0 before suspend/hibernate.
2025 2. PIC interrupts also depend on CPU0. CPU0 can't be
2026 removed if a PIC interrupt is detected.
2027 It's said poweroff/reboot may depend on CPU0 on some
2028 machines although I haven't seen such issues so far
2029 after CPU0 is offline on a few tested machines.
2030 If the dependencies are under your control, you can
2031 turn on cpu0_hotplug.
2033 nptcg= [IA-64] Override max number of concurrent global TLB
2034 purges which is reported from either PAL_VM_SUMMARY or
2037 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2038 could support. nr_cpus=n : n >= 1 limits the kernel to
2039 supporting 'n' processors. Later in runtime you can not
2040 use hotplug cpu feature to put more cpu back to online.
2041 just like you compile the kernel NR_CPUS=n
2043 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2045 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2046 Allowed values are enable and disable
2048 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2049 one of ['zone', 'node', 'default'] can be specified
2050 This can be set from sysctl after boot.
2051 See Documentation/sysctl/vm.txt for details.
2053 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2054 See Documentation/debugging-via-ohci1394.txt for more
2057 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2058 Rather than timing out after 20 ms if an EC
2059 command is not properly ACKed, override the length
2060 of the timeout. We have interrupts disabled while
2061 waiting for the ACK, so if this is set too high
2062 interrupts *may* be lost!
2064 omap_mux= [OMAP] Override bootloader pin multiplexing.
2065 Format: <mux_mode0.mode_name=value>...
2066 For example, to override I2C bus2:
2067 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2069 oprofile.timer= [HW]
2070 Use timer interrupt instead of performance counters
2072 oprofile.cpu_type= Force an oprofile cpu type
2073 This might be useful if you have an older oprofile
2074 userland or if you want common events.
2075 Format: { arch_perfmon }
2076 arch_perfmon: [X86] Force use of architectural
2077 perfmon on Intel CPUs instead of the
2078 CPU specific event set.
2079 timer: [X86] Force use of architectural NMI
2080 timer mode (see also oprofile.timer
2081 for generic hr timer mode)
2082 [s390] Force legacy basic mode sampling
2083 (report cpu_type "timer")
2085 oops=panic Always panic on oopses. Default is to just kill the
2086 process, but there is a small probability of
2087 deadlocking the machine.
2088 This will also cause panics on machine check exceptions.
2089 Useful together with panic=30 to trigger a reboot.
2092 See Documentation/sound/oss/oss-parameters.txt
2094 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2095 timeout > 0: seconds before rebooting
2096 timeout = 0: wait forever
2097 timeout < 0: reboot immediately
2100 parkbd.port= [HW] Parallel port number the keyboard adapter is
2101 connected to, default is 0.
2103 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2104 0 for XT, 1 for AT (default is AT).
2107 parport= [HW,PPT] Specify parallel ports. 0 disables.
2108 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2109 Use 'auto' to force the driver to use any
2110 IRQ/DMA settings detected (the default is to
2111 ignore detected IRQ/DMA settings because of
2112 possible conflicts). You can specify the base
2113 address, IRQ, and DMA settings; IRQ and DMA
2114 should be numbers, or 'auto' (for using detected
2115 settings on that particular port), or 'nofifo'
2116 (to avoid using a FIFO even if it is detected).
2117 Parallel ports are assigned in the order they
2118 are specified on the command line, starting
2121 parport_init_mode= [HW,PPT]
2122 Configure VIA parallel port to operate in
2123 a specific mode. This is necessary on Pegasos
2124 computer where firmware has no options for setting
2125 up parallel port mode and sets it to spp.
2126 Currently this function knows 686a and 8231 chips.
2127 Format: [spp|ps2|epp|ecp|ecpepp]
2130 Halt all CPUs after the first oops has been printed for
2131 the specified number of seconds. This is to be used if
2132 your oopses keep scrolling off the screen.
2137 See header of drivers/block/paride/pcd.c.
2138 See also Documentation/blockdev/paride.txt.
2140 pci=option[,option...] [PCI] various PCI subsystem options:
2141 earlydump [X86] dump PCI config space before the kernel
2143 off [X86] don't probe for the PCI bus
2144 bios [X86-32] force use of PCI BIOS, don't access
2145 the hardware directly. Use this if your machine
2146 has a non-standard PCI host bridge.
2147 nobios [X86-32] disallow use of PCI BIOS, only direct
2148 hardware access methods are allowed. Use this
2149 if you experience crashes upon bootup and you
2150 suspect they are caused by the BIOS.
2151 conf1 [X86] Force use of PCI Configuration
2153 conf2 [X86] Force use of PCI Configuration
2155 noaer [PCIE] If the PCIEAER kernel config parameter is
2156 enabled, this kernel boot option can be used to
2157 disable the use of PCIE advanced error reporting.
2158 nodomains [PCI] Disable support for multiple PCI
2159 root domains (aka PCI segments, in ACPI-speak).
2160 nommconf [X86] Disable use of MMCONFIG for PCI
2162 check_enable_amd_mmconf [X86] check for and enable
2163 properly configured MMIO access to PCI
2164 config space on AMD family 10h CPU
2165 nomsi [MSI] If the PCI_MSI kernel config parameter is
2166 enabled, this kernel boot option can be used to
2167 disable the use of MSI interrupts system-wide.
2168 noioapicquirk [APIC] Disable all boot interrupt quirks.
2169 Safety option to keep boot IRQs enabled. This
2170 should never be necessary.
2171 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
2172 primary IO-APIC for bridges that cannot disable
2173 boot IRQs. This fixes a source of spurious IRQs
2174 when the system masks IRQs.
2175 noioapicreroute [APIC] Disable workaround that uses the
2176 boot IRQ equivalent of an IRQ that connects to
2177 a chipset where boot IRQs cannot be disabled.
2178 The opposite of ioapicreroute.
2179 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
2180 routing table. These calls are known to be buggy
2181 on several machines and they hang the machine
2182 when used, but on other computers it's the only
2183 way to get the interrupt routing table. Try
2184 this option if the kernel is unable to allocate
2185 IRQs or discover secondary PCI buses on your
2187 rom [X86] Assign address space to expansion ROMs.
2188 Use with caution as certain devices share
2189 address decoders between ROMs and other
2191 norom [X86] Do not assign address space to
2192 expansion ROMs that do not already have
2193 BIOS assigned address ranges.
2194 nobar [X86] Do not assign address space to the
2195 BARs that weren't assigned by the BIOS.
2196 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
2197 assigned automatically to PCI devices. You can
2198 make the kernel exclude IRQs of your ISA cards
2200 pirqaddr=0xAAAAA [X86] Specify the physical address
2201 of the PIRQ table (normally generated
2202 by the BIOS) if it is outside the
2203 F0000h-100000h range.
2204 lastbus=N [X86] Scan all buses thru bus #N. Can be
2205 useful if the kernel is unable to find your
2206 secondary buses and you want to tell it
2207 explicitly which ones they are.
2208 assign-busses [X86] Always assign all PCI bus
2209 numbers ourselves, overriding
2210 whatever the firmware may have done.
2211 usepirqmask [X86] Honor the possible IRQ mask stored
2212 in the BIOS $PIR table. This is needed on
2213 some systems with broken BIOSes, notably
2214 some HP Pavilion N5400 and Omnibook XE3
2215 notebooks. This will have no effect if ACPI
2216 IRQ routing is enabled.
2217 noacpi [X86] Do not use ACPI for IRQ routing
2218 or for PCI scanning.
2219 use_crs [X86] Use PCI host bridge window information
2220 from ACPI. On BIOSes from 2008 or later, this
2221 is enabled by default. If you need to use this,
2222 please report a bug.
2223 nocrs [X86] Ignore PCI host bridge windows from ACPI.
2224 If you need to use this, please report a bug.
2225 routeirq Do IRQ routing for all PCI devices.
2226 This is normally done in pci_enable_device(),
2227 so this option is a temporary workaround
2228 for broken drivers that don't call it.
2229 skip_isa_align [X86] do not align io start addr, so can
2230 handle more pci cards
2231 firmware [ARM] Do not re-enumerate the bus but instead
2232 just use the configuration from the
2233 bootloader. This is currently used on
2234 IXP2000 systems where the bus has to be
2235 configured a certain way for adjunct CPUs.
2236 noearly [X86] Don't do any early type 1 scanning.
2237 This might help on some broken boards which
2238 machine check when some devices' config space
2239 is read. But various workarounds are disabled
2240 and some IOMMU drivers will not work.
2241 bfsort Sort PCI devices into breadth-first order.
2242 This sorting is done to get a device
2243 order compatible with older (<= 2.4) kernels.
2244 nobfsort Don't sort PCI devices into breadth-first order.
2245 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
2246 tuning and use the BIOS-configured MPS defaults.
2247 pcie_bus_safe Set every device's MPS to the largest value
2248 supported by all devices below the root complex.
2249 pcie_bus_perf Set device MPS to the largest allowable MPS
2250 based on its parent bus. Also set MRRS (Max
2251 Read Request Size) to the largest supported
2252 value (no larger than the MPS that the device
2253 or bus can support) for best performance.
2254 pcie_bus_peer2peer Set every device's MPS to 128B, which
2255 every device is guaranteed to support. This
2256 configuration allows peer-to-peer DMA between
2257 any pair of devices, possibly at the cost of
2258 reduced performance. This also guarantees
2259 that hot-added devices will work.
2260 cbiosize=nn[KMG] The fixed amount of bus space which is
2261 reserved for the CardBus bridge's IO window.
2262 The default value is 256 bytes.
2263 cbmemsize=nn[KMG] The fixed amount of bus space which is
2264 reserved for the CardBus bridge's memory
2265 window. The default value is 64 megabytes.
2268 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
2269 Specifies alignment and device to reassign
2270 aligned memory resources.
2271 If <order of align> is not specified,
2272 PAGE_SIZE is used as alignment.
2273 PCI-PCI bridge can be specified, if resource
2274 windows need to be expanded.
2275 ecrc= Enable/disable PCIe ECRC (transaction layer
2276 end-to-end CRC checking).
2277 bios: Use BIOS/firmware settings. This is the
2281 hpiosize=nn[KMG] The fixed amount of bus space which is
2282 reserved for hotplug bridge's IO window.
2283 Default size is 256 bytes.
2284 hpmemsize=nn[KMG] The fixed amount of bus space which is
2285 reserved for hotplug bridge's memory window.
2286 Default size is 2 megabytes.
2287 realloc= Enable/disable reallocating PCI bridge resources
2288 if allocations done by BIOS are too small to
2289 accommodate resources required by all child
2291 off: Turn realloc off
2293 realloc same as realloc=on
2294 noari do not use PCIe ARI.
2295 pcie_scan_all Scan all possible PCIe devices. Otherwise we
2296 only look for one device below a PCIe downstream
2299 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
2302 force Enable ASPM even on devices that claim not to support it.
2303 WARNING: Forcing ASPM on may cause system lockups.
2305 pcie_hp= [PCIE] PCI Express Hotplug driver options:
2306 nomsi Do not use MSI for PCI Express Native Hotplug (this
2307 makes all PCIe ports use INTx for hotplug services).
2309 pcie_ports= [PCIE] PCIe ports handling:
2310 auto Ask the BIOS whether or not to use native PCIe services
2311 associated with PCIe ports (PME, hot-plug, AER). Use
2312 them only if that is allowed by the BIOS.
2313 native Use native PCIe services associated with PCIe ports
2315 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
2318 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
2319 nomsi Do not use MSI for native PCIe PME signaling (this makes
2320 all PCIe root ports use INTx for all services).
2322 pcmv= [HW,PCMCIA] BadgePAD 4
2325 See Documentation/blockdev/paride.txt.
2327 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
2330 See arch/parisc/kernel/pdc_chassis.c
2332 percpu_alloc= Select which percpu first chunk allocator to use.
2333 Currently supported values are "embed" and "page".
2334 Archs may support subset or none of the selections.
2335 See comments in mm/percpu.c for details on each
2336 allocator. This parameter is primarily for debugging
2337 and performance comparison.
2340 See Documentation/blockdev/paride.txt.
2343 See Documentation/blockdev/paride.txt.
2345 pirq= [SMP,APIC] Manual mp-table setup
2346 See Documentation/x86/i386/IO-APIC.txt.
2348 plip= [PPT,NET] Parallel port network link
2349 Format: { parport<nr> | timid | 0 }
2350 See also Documentation/parport.txt.
2352 pmtmr= [X86] Manual setup of pmtmr I/O Port.
2353 Override pmtimer IOPort with a hex value.
2357 Enable PNP debug messages (depends on the
2358 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
2359 via /sys/module/pnp/parameters/debug. We always show
2360 current resource usage; turning this on also shows
2361 possible settings and some assignment information.
2367 { on | off | curr | res | no-curr | no-res }
2370 [ISAPNP] Exclude IRQs for the autoconfiguration
2373 [ISAPNP] Exclude DMAs for the autoconfiguration
2375 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
2376 Ranges are in pairs (I/O port base and size).
2379 [ISAPNP] Exclude memory regions for the
2381 Ranges are in pairs (memory base and size).
2383 ports= [IP_VS_FTP] IPVS ftp helper module
2385 Up to 8 (IP_VS_APP_MAX_PORTS) ports
2387 Format: <port>,<port>....
2389 print-fatal-signals=
2390 [KNL] debug: print fatal signals
2392 If enabled, warn about various signal handling
2393 related application anomalies: too many signals,
2394 too many POSIX.1 timers, fatal signals causing a
2397 If you hit the warning due to signal overflow,
2398 you might want to try "ulimit -i unlimited".
2402 printk.always_kmsg_dump=
2403 Trigger kmsg_dump for cases other than kernel oops or
2405 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2408 printk.time= Show timing data prefixed to each printk message line
2409 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2411 processor.max_cstate= [HW,ACPI]
2412 Limit processor to maximum C-state
2413 max_cstate=9 overrides any DMI blacklist limit.
2415 processor.nocst [HW,ACPI]
2416 Ignore the _CST method to determine C-states,
2417 instead using the legacy FADT method
2419 profile= [KNL] Enable kernel profiling via /proc/profile
2420 Format: [schedule,]<number>
2421 Param: "schedule" - profile schedule points.
2422 Param: <number> - step/bucket size as a power of 2 for
2423 statistical time based profiling.
2424 Param: "sleep" - profile D-state sleeping (millisecs).
2425 Requires CONFIG_SCHEDSTATS
2426 Param: "kvm" - profile VM exits.
2428 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
2430 See Documentation/blockdev/ramdisk.txt.
2432 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
2433 probe for; one of (bare|imps|exps|lifebook|any).
2434 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
2436 psmouse.resetafter= [HW,MOUSE]
2437 Try to reset the device after so many bad packets
2440 [HW,MOUSE] Set desired mouse resolution, in dpi.
2441 psmouse.smartscroll=
2442 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
2443 0 = disabled, 1 = enabled (default).
2445 pstore.backend= Specify the name of the pstore backend to use
2448 See Documentation/blockdev/paride.txt.
2451 [KNL] Number of legacy pty's. Overwrites compiled-in
2454 quiet [KNL] Disable most log messages
2459 See Documentation/md.txt.
2461 ramdisk_blocksize= [RAM]
2462 See Documentation/blockdev/ramdisk.txt.
2464 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
2465 See Documentation/blockdev/ramdisk.txt.
2467 rcu_nocbs= [KNL,BOOT]
2468 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
2469 the specified list of CPUs to be no-callback CPUs.
2470 Invocation of these CPUs' RCU callbacks will
2471 be offloaded to "rcuox/N" kthreads created for
2472 that purpose, where "x" is "b" for RCU-bh, "p"
2473 for RCU-preempt, and "s" for RCU-sched, and "N"
2474 is the CPU number. This reduces OS jitter on the
2475 offloaded CPUs, which can be useful for HPC and
2477 real-time workloads. It can also improve energy
2478 efficiency for asymmetric multiprocessors.
2480 rcu_nocb_poll [KNL,BOOT]
2481 Rather than requiring that offloaded CPUs
2482 (specified by rcu_nocbs= above) explicitly
2483 awaken the corresponding "rcuoN" kthreads,
2484 make these kthreads poll for callbacks.
2485 This improves the real-time response for the
2486 offloaded CPUs by relieving them of the need to
2487 wake up the corresponding kthread, but degrades
2488 energy efficiency by requiring that the kthreads
2489 periodically wake up to do the polling.
2491 rcutree.blimit= [KNL,BOOT]
2492 Set maximum number of finished RCU callbacks to process
2495 rcutree.fanout_leaf= [KNL,BOOT]
2496 Increase the number of CPUs assigned to each
2497 leaf rcu_node structure. Useful for very large
2500 rcutree.jiffies_till_first_fqs= [KNL,BOOT]
2501 Set delay from grace-period initialization to
2502 first attempt to force quiescent states.
2503 Units are jiffies, minimum value is zero,
2504 and maximum value is HZ.
2506 rcutree.jiffies_till_next_fqs= [KNL,BOOT]
2507 Set delay between subsequent attempts to force
2508 quiescent states. Units are jiffies, minimum
2509 value is one, and maximum value is HZ.
2511 rcutree.qhimark= [KNL,BOOT]
2512 Set threshold of queued
2513 RCU callbacks over which batch limiting is disabled.
2515 rcutree.qlowmark= [KNL,BOOT]
2516 Set threshold of queued RCU callbacks below which
2517 batch limiting is re-enabled.
2519 rcutree.rcu_cpu_stall_suppress= [KNL,BOOT]
2520 Suppress RCU CPU stall warning messages.
2522 rcutree.rcu_cpu_stall_timeout= [KNL,BOOT]
2523 Set timeout for RCU CPU stall warning messages.
2525 rcutree.rcu_idle_gp_delay= [KNL,BOOT]
2526 Set wakeup interval for idle CPUs that have
2527 RCU callbacks (RCU_FAST_NO_HZ=y).
2529 rcutree.rcu_idle_lazy_gp_delay= [KNL,BOOT]
2530 Set wakeup interval for idle CPUs that have
2531 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
2532 Lazy RCU callbacks are those which RCU can
2533 prove do nothing more than free memory.
2535 rcutorture.fqs_duration= [KNL,BOOT]
2536 Set duration of force_quiescent_state bursts.
2538 rcutorture.fqs_holdoff= [KNL,BOOT]
2539 Set holdoff time within force_quiescent_state bursts.
2541 rcutorture.fqs_stutter= [KNL,BOOT]
2542 Set wait time between force_quiescent_state bursts.
2544 rcutorture.irqreader= [KNL,BOOT]
2545 Test RCU readers from irq handlers.
2547 rcutorture.n_barrier_cbs= [KNL,BOOT]
2548 Set callbacks/threads for rcu_barrier() testing.
2550 rcutorture.nfakewriters= [KNL,BOOT]
2551 Set number of concurrent RCU writers. These just
2552 stress RCU, they don't participate in the actual
2553 test, hence the "fake".
2555 rcutorture.nreaders= [KNL,BOOT]
2556 Set number of RCU readers.
2558 rcutorture.onoff_holdoff= [KNL,BOOT]
2559 Set time (s) after boot for CPU-hotplug testing.
2561 rcutorture.onoff_interval= [KNL,BOOT]
2562 Set time (s) between CPU-hotplug operations, or
2563 zero to disable CPU-hotplug testing.
2565 rcutorture.shuffle_interval= [KNL,BOOT]
2566 Set task-shuffle interval (s). Shuffling tasks
2567 allows some CPUs to go into dyntick-idle mode
2568 during the rcutorture test.
2570 rcutorture.shutdown_secs= [KNL,BOOT]
2571 Set time (s) after boot system shutdown. This
2572 is useful for hands-off automated testing.
2574 rcutorture.stall_cpu= [KNL,BOOT]
2575 Duration of CPU stall (s) to test RCU CPU stall
2576 warnings, zero to disable.
2578 rcutorture.stall_cpu_holdoff= [KNL,BOOT]
2579 Time to wait (s) after boot before inducing stall.
2581 rcutorture.stat_interval= [KNL,BOOT]
2582 Time (s) between statistics printk()s.
2584 rcutorture.stutter= [KNL,BOOT]
2585 Time (s) to stutter testing, for example, specifying
2586 five seconds causes the test to run for five seconds,
2587 wait for five seconds, and so on. This tests RCU's
2588 ability to transition abruptly to and from idle.
2590 rcutorture.test_boost= [KNL,BOOT]
2591 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
2592 "Maybe" means test if the RCU implementation
2593 under test support RCU priority boosting.
2595 rcutorture.test_boost_duration= [KNL,BOOT]
2596 Duration (s) of each individual boost test.
2598 rcutorture.test_boost_interval= [KNL,BOOT]
2599 Interval (s) between each boost test.
2601 rcutorture.test_no_idle_hz= [KNL,BOOT]
2602 Test RCU's dyntick-idle handling. See also the
2603 rcutorture.shuffle_interval parameter.
2605 rcutorture.torture_type= [KNL,BOOT]
2606 Specify the RCU implementation to test.
2608 rcutorture.verbose= [KNL,BOOT]
2609 Enable additional printk() statements.
2613 Run specified binary instead of /init from the ramdisk,
2614 used for early userspace startup. See initrd.
2616 reboot= [BUGS=X86-32,BUGS=ARM,BUGS=IA-64] Rebooting mode
2617 Format: <reboot_mode>[,<reboot_mode2>[,...]]
2618 See arch/*/kernel/reboot.c or arch/*/kernel/process.c
2621 [KNL, SMP] Set scheduler's default relax_domain_level.
2622 See Documentation/cgroups/cpusets.txt.
2624 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
2626 reservetop= [X86-32]
2628 Reserves a hole at the top of the kernel virtual
2633 Set the amount of memory to reserve for BIOS at
2634 the bottom of the address space.
2636 reset_devices [KNL] Force drivers to reset the underlying device
2637 during initialization.
2640 Specify the partition device for software suspend
2642 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
2644 resume_offset= [SWSUSP]
2645 Specify the offset from the beginning of the partition
2646 given by "resume=" at which the swap header is located,
2647 in <PAGE_SIZE> units (needed only for swap files).
2648 See Documentation/power/swsusp-and-swap-files.txt
2650 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
2651 read the resume files
2653 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
2654 Useful for devices that are detected asynchronously
2655 (e.g. USB and MMC devices).
2657 hibernate= [HIBERNATION]
2658 noresume Don't check if there's a hibernation image
2659 present during boot.
2660 nocompress Don't compress/decompress hibernation images.
2662 retain_initrd [RAM] Keep initrd memory after extraction
2664 rhash_entries= [KNL,NET]
2665 Set number of hash buckets for route cache
2667 riscom8= [HW,SERIAL]
2668 Format: <io_board1>[,<io_board2>[,...<io_boardN>]]
2670 ro [KNL] Mount root device read-only on boot
2672 root= [KNL] Root filesystem
2673 See name_to_dev_t comment in init/do_mounts.c.
2675 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
2676 mount the root filesystem
2678 rootflags= [KNL] Set root filesystem mount option string
2680 rootfstype= [KNL] Set root filesystem type
2682 rootwait [KNL] Wait (indefinitely) for root device to show up.
2683 Useful for devices that are detected asynchronously
2684 (e.g. USB and MMC devices).
2686 rw [KNL] Mount root device read-write on boot
2688 S [KNL] Run init in single mode
2691 See drivers/net/irda/sa1100_ir.c.
2693 sbni= [NET] Granch SBNI12 leased line adapter
2695 sched_debug [KNL] Enables verbose scheduler debug messages.
2697 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
2698 xtime_lock contention on larger systems, and/or RCU lock
2699 contention on all systems with CONFIG_MAXSMP set.
2700 Format: { "0" | "1" }
2701 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
2703 Note: increases power consumption, thus should only be
2704 enabled if running jitter sensitive (HPC/RT) workloads.
2706 security= [SECURITY] Choose a security module to enable at boot.
2707 If this boot parameter is not specified, only the first
2708 security module asking for security registration will be
2709 loaded. An invalid security module name will be treated
2710 as if no module has been chosen.
2712 selinux= [SELINUX] Disable or enable SELinux at boot time.
2713 Format: { "0" | "1" }
2714 See security/selinux/Kconfig help text.
2717 Default value is set via kernel config option.
2718 If enabled at boot time, /selinux/disable can be used
2719 later to disable prior to initial policy load.
2721 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
2722 Format: { "0" | "1" }
2723 See security/apparmor/Kconfig help text
2726 Default value is set via kernel config option.
2728 serialnumber [BUGS=X86-32]
2731 Maximal number of shapers.
2733 show_msr= [x86] show boot-time MSR settings
2734 Format: { <integer> }
2735 Show boot-time (BIOS-initialized) MSR settings.
2736 The parameter means the number of CPUs to show,
2737 for example 1 means boot CPU only.
2744 slab_max_order= [MM, SLAB]
2745 Determines the maximum allowed order for slabs.
2746 A high setting may cause OOMs due to memory
2747 fragmentation. Defaults to 1 for systems with
2748 more than 32MB of RAM, 0 otherwise.
2750 slub_debug[=options[,slabs]] [MM, SLUB]
2751 Enabling slub_debug allows one to determine the
2752 culprit if slab objects become corrupted. Enabling
2753 slub_debug can create guard zones around objects and
2754 may poison objects when not in use. Also tracks the
2755 last alloc / free. For more information see
2756 Documentation/vm/slub.txt.
2758 slub_max_order= [MM, SLUB]
2759 Determines the maximum allowed order for slabs.
2760 A high setting may cause OOMs due to memory
2761 fragmentation. For more information see
2762 Documentation/vm/slub.txt.
2764 slub_min_objects= [MM, SLUB]
2765 The minimum number of objects per slab. SLUB will
2766 increase the slab order up to slub_max_order to
2767 generate a sufficiently large slab able to contain
2768 the number of objects indicated. The higher the number
2769 of objects the smaller the overhead of tracking slabs
2770 and the less frequently locks need to be acquired.
2771 For more information see Documentation/vm/slub.txt.
2773 slub_min_order= [MM, SLUB]
2774 Determines the minimum page order for slabs. Must be
2775 lower than slub_max_order.
2776 For more information see Documentation/vm/slub.txt.
2778 slub_nomerge [MM, SLUB]
2779 Disable merging of slabs with similar size. May be
2780 necessary if there is some reason to distinguish
2781 allocs to different slabs. Debug options disable
2782 merging on their own.
2783 For more information see Documentation/vm/slub.txt.
2786 Format: <io1>[,<io2>[,...,<io8>]]
2788 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
2789 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
2790 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
2791 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
2792 smsc-ircc2.ircc_irq= [HW] IRQ line
2793 smsc-ircc2.ircc_dma= [HW] DMA channel
2794 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
2795 0: Toshiba Satellite 1800 (GP data pin select)
2796 1: Fast pin select (default)
2800 [KNL] Should the soft-lockup detector generate panics.
2803 sonypi.*= [HW] Sony Programmable I/O Control Device driver
2804 See Documentation/laptops/sonypi.txt
2806 specialix= [HW,SERIAL] Specialix multi-serial port adapter
2807 See Documentation/serial/specialix.txt.
2809 spia_io_base= [HW,MTD]
2815 Enabled the stack tracer on boot up.
2817 stacktrace_filter=[function-list]
2818 [FTRACE] Limit the functions that the stack tracer
2819 will trace at boot up. function-list is a comma separated
2820 list of functions. This list can be changed at run
2821 time by the stack_trace_filter file in the debugfs
2822 tracing directory. Note, this enables stack tracing
2823 and the stacktrace above is not needed.
2827 Set the STI (builtin display/keyboard on the HP-PARISC
2828 machines) console (graphic card) which should be used
2829 as the initial boot-console.
2830 See also comment in drivers/video/console/sticore.c.
2833 See comment in drivers/video/console/sticore.c.
2836 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
2838 sunrpc.min_resvport=
2839 sunrpc.max_resvport=
2841 SunRPC servers often require that client requests
2842 originate from a privileged port (i.e. a port in the
2843 range 0 < portnr < 1024).
2844 An administrator who wishes to reserve some of these
2845 ports for other uses may adjust the range that the
2846 kernel's sunrpc client considers to be privileged
2847 using these two parameters to set the minimum and
2848 maximum port values.
2852 Control how the NFS server code allocates CPUs to
2853 service thread pools. Depending on how many NICs
2854 you have and where their interrupts are bound, this
2855 option will affect which CPUs will do NFS serving.
2856 Note: this parameter cannot be changed while the
2857 NFS server is running.
2859 auto the server chooses an appropriate mode
2860 automatically using heuristics
2861 global a single global pool contains all CPUs
2862 percpu one pool for each CPU
2863 pernode one pool for each NUMA node (equivalent
2864 to global on non-NUMA machines)
2866 sunrpc.tcp_slot_table_entries=
2867 sunrpc.udp_slot_table_entries=
2869 Sets the upper limit on the number of simultaneous
2870 RPC calls that can be sent from the client to a
2871 server. Increasing these values may allow you to
2872 improve throughput, but will also increase the
2873 amount of memory reserved for use by the client.
2876 [KNL] Enable accounting of swap in memory resource
2877 controller if no parameter or 1 is given or disable
2878 it if 0 is given (See Documentation/cgroups/memory.txt)
2880 swiotlb= [IA-64] Number of I/O TLB slabs
2884 sysfs.deprecated=0|1 [KNL]
2885 Enable/disable old style sysfs layout for old udev
2886 on older distributions. When this option is enabled
2887 very new udev will not work anymore. When this option
2888 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
2889 in older udev will not work anymore.
2890 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
2891 the kernel configuration.
2893 sysrq_always_enabled
2895 Ignore sysrq setting - this boot parameter will
2896 neutralize any effect of /proc/sys/kernel/sysrq.
2897 Useful for debugging.
2901 test_suspend= [SUSPEND]
2902 Specify "mem" (for Suspend-to-RAM) or "standby" (for
2903 standby suspend) as the system sleep state to briefly
2904 enter during system startup. The system is woken from
2905 this state using a wakeup-capable RTC alarm.
2907 thash_entries= [KNL,NET]
2908 Set number of hash buckets for TCP connection
2910 thermal.act= [HW,ACPI]
2911 -1: disable all active trip points in all thermal zones
2912 <degrees C>: override all lowest active trip points
2914 thermal.crt= [HW,ACPI]
2915 -1: disable all critical trip points in all thermal zones
2916 <degrees C>: override all critical trip points
2918 thermal.nocrt= [HW,ACPI]
2919 Set to disable actions on ACPI thermal zone
2920 critical and hot trip points.
2922 thermal.off= [HW,ACPI]
2923 1: disable ACPI thermal control
2925 thermal.psv= [HW,ACPI]
2926 -1: disable all passive trip points
2927 <degrees C>: override all passive trip points to this
2930 thermal.tzp= [HW,ACPI]
2931 Specify global default ACPI thermal zone polling rate
2932 <deci-seconds>: poll all this frequency
2933 0: no polling (default)
2936 Force threading of all interrupt handlers except those
2937 marked explicitly IRQF_NO_THREAD.
2941 Specify if the kernel should make use of the cpu
2942 topology information if the hardware supports this.
2943 The scheduler will make use of this information and
2944 e.g. base its process migration decisions on it.
2949 tpm_suspend_pcr=[HW,TPM]
2950 Format: integer pcr id
2951 Specify that at suspend time, the tpm driver
2952 should extend the specified pcr with zeros,
2953 as a workaround for some chips which fail to
2954 flush the last written pcr on TPM_SaveState.
2955 This will guarantee that all the other pcrs
2958 trace_buf_size=nn[KMG]
2959 [FTRACE] will set tracing buffer size.
2961 trace_event=[event-list]
2962 [FTRACE] Set and start specified trace events in order
2963 to facilitate early boot debugging.
2964 See also Documentation/trace/events.txt
2966 trace_options=[option-list]
2967 [FTRACE] Enable or disable tracer options at boot.
2968 The option-list is a comma delimited list of options
2969 that can be enabled or disabled just as if you were
2970 to echo the option name into
2972 /sys/kernel/debug/tracing/trace_options
2974 For example, to enable stacktrace option (to dump the
2975 stack trace of each event), add to the command line:
2977 trace_options=stacktrace
2979 See also Documentation/trace/ftrace.txt "trace options"
2982 transparent_hugepage=
2984 Format: [always|madvise|never]
2985 Can be used to control the default behavior of the system
2986 with respect to transparent hugepages.
2987 See Documentation/vm/transhuge.txt for more details.
2989 tsc= Disable clocksource stability checks for TSC.
2991 [x86] reliable: mark tsc clocksource as reliable, this
2992 disables clocksource verification at runtime, as well
2993 as the stability checks done at bootup. Used to enable
2994 high-resolution timer mode on older hardware, and in
2995 virtualized environment.
2996 [x86] noirqtime: Do not use TSC to do irq accounting.
2997 Used to run time disable IRQ_TIME_ACCOUNTING on any
2998 platforms where RDTSC is slow and this accounting
3001 turbografx.map[2|3]= [HW,JOY]
3002 TurboGraFX parallel port interface
3004 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
3005 See also Documentation/input/joystick-parport.txt
3007 udbg-immortal [PPC] When debugging early kernel crashes that
3008 happen after console_init() and before a proper
3009 console driver takes over, this boot options might
3010 help "seeing" what's going on.
3012 uhash_entries= [KNL,NET]
3013 Set number of hash buckets for UDP/UDP-Lite connections
3016 [USB] Ignore overcurrent events (default N).
3017 Some badly-designed motherboards generate lots of
3018 bogus events, for ports that aren't wired to
3019 anything. Set this parameter to avoid log spamming.
3020 Note that genuine overcurrent events won't be
3024 [X86] Cause panic on unknown NMI.
3026 usbcore.authorized_default=
3027 [USB] Default USB device authorization:
3028 (default -1 = authorized except for wireless USB,
3029 0 = not authorized, 1 = authorized)
3031 usbcore.autosuspend=
3032 [USB] The autosuspend time delay (in seconds) used
3033 for newly-detected USB devices (default 2). This
3034 is the time required before an idle device will be
3035 autosuspended. Devices for which the delay is set
3036 to a negative value won't be autosuspended at all.
3038 usbcore.usbfs_snoop=
3039 [USB] Set to log all usbfs traffic (default 0 = off).
3041 usbcore.blinkenlights=
3042 [USB] Set to cycle leds on hubs (default 0 = off).
3044 usbcore.old_scheme_first=
3045 [USB] Start with the old device initialization
3046 scheme (default 0 = off).
3048 usbcore.usbfs_memory_mb=
3049 [USB] Memory limit (in MB) for buffers allocated by
3050 usbfs (default = 16, 0 = max = 2047).
3052 usbcore.use_both_schemes=
3053 [USB] Try the other device initialization scheme
3054 if the first one fails (default 1 = enabled).
3056 usbcore.initial_descriptor_timeout=
3057 [USB] Specifies timeout for the initial 64-byte
3058 USB_REQ_GET_DESCRIPTOR request in milliseconds
3059 (default 5000 = 5.0 seconds).
3062 [USBHID] The interval which mice are to be polled at.
3064 usb-storage.delay_use=
3065 [UMS] The delay in seconds before a new device is
3066 scanned for Logical Units (default 5).
3069 [UMS] A list of quirks entries to supplement or
3070 override the built-in unusual_devs list. List
3071 entries are separated by commas. Each entry has
3072 the form VID:PID:Flags where VID and PID are Vendor
3073 and Product ID values (4-digit hex numbers) and
3074 Flags is a set of characters, each corresponding
3075 to a common usb-storage quirk flag as follows:
3076 a = SANE_SENSE (collect more than 18 bytes
3078 b = BAD_SENSE (don't collect more than 18
3079 bytes of sense data);
3080 c = FIX_CAPACITY (decrease the reported
3081 device capacity by one sector);
3082 d = NO_READ_DISC_INFO (don't use
3083 READ_DISC_INFO command);
3084 e = NO_READ_CAPACITY_16 (don't use
3085 READ_CAPACITY_16 command);
3086 h = CAPACITY_HEURISTICS (decrease the
3087 reported device capacity by one
3088 sector if the number is odd);
3089 i = IGNORE_DEVICE (don't bind to this
3091 l = NOT_LOCKABLE (don't try to lock and
3092 unlock ejectable media);
3093 m = MAX_SECTORS_64 (don't transfer more
3094 than 64 sectors = 32 KB at a time);
3095 n = INITIAL_READ10 (force a retry of the
3096 initial READ(10) command);
3097 o = CAPACITY_OK (accept the capacity
3098 reported by the device);
3099 p = WRITE_CACHE (the device cache is ON
3101 r = IGNORE_RESIDUE (the device reports
3102 bogus residue values);
3103 s = SINGLE_LUN (the device has only one
3105 w = NO_WP_DETECT (don't test whether the
3106 medium is write-protected).
3107 Example: quirks=0419:aaf5:rl,0421:0433:rc
3109 user_debug= [KNL,ARM]
3111 See arch/arm/Kconfig.debug help text.
3112 1 - undefined instruction events
3114 4 - invalid data aborts
3117 Example: user_debug=31
3120 [X86] Flags controlling user PTE allocations.
3122 nohigh = do not allocate PTE pages in
3123 HIGHMEM regardless of setting
3127 vdso=2: enable compat VDSO (default with COMPAT_VDSO)
3128 vdso=1: enable VDSO (default)
3129 vdso=0: disable VDSO mapping
3132 vdso32=2: enable compat VDSO (default with COMPAT_VDSO)
3133 vdso32=1: enable 32-bit VDSO (default)
3134 vdso32=0: disable 32-bit VDSO mapping
3137 vector=percpu: enable percpu vector domain
3139 video= [FB] Frame buffer configuration
3140 See Documentation/fb/modedb.txt.
3143 [VMMIO] Memory mapped virtio (platform) device.
3145 <size>@<baseaddr>:<irq>[:<id>]
3147 <size> := size (can use standard suffixes
3149 <baseaddr> := physical base address
3150 <irq> := interrupt number (as passed to
3152 <id> := (optional) platform device id
3154 virtio_mmio.device=1K@0x100b0000:48:7
3156 Can be used multiple times for multiple devices.
3158 vga= [BOOT,X86-32] Select a particular video mode
3159 See Documentation/x86/boot.txt and
3160 Documentation/svga.txt.
3161 Use vga=ask for menu.
3162 This is actually a boot loader parameter; the value is
3163 passed to the kernel using a special protocol.
3165 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
3166 size of <nn>. This can be used to increase the
3167 minimum size (128MB on x86). It can also be used to
3168 decrease the size and leave more room for directly
3171 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
3174 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
3177 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
3181 Controls the behavior of vsyscalls (i.e. calls to
3182 fixed addresses of 0xffffffffff600x00 from legacy
3183 code). Most statically-linked binaries and older
3184 versions of glibc use these calls. Because these
3185 functions are at fixed addresses, they make nice
3186 targets for exploits that can control RIP.
3188 emulate [default] Vsyscalls turn into traps and are
3189 emulated reasonably safely.
3191 native Vsyscalls are native syscall instructions.
3192 This is a little bit faster than trapping
3193 and makes a few dynamic recompilers work
3194 better than they would in emulation mode.
3195 It also makes exploits much easier to write.
3197 none Vsyscalls don't work at all. This makes
3198 them quite hard to use for exploits but
3199 might break your system.
3201 vt.cur_default= [VT] Default cursor shape.
3202 Format: 0xCCBBAA, where AA, BB, and CC are the same as
3203 the parameters of the <Esc>[?A;B;Cc escape sequence;
3204 see VGA-softcursor.txt. Default: 2 = underline.
3206 vt.default_blu= [VT]
3207 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
3208 Change the default blue palette of the console.
3209 This is a 16-member array composed of values
3212 vt.default_grn= [VT]
3213 Format: <green0>,<green1>,<green2>,...,<green15>
3214 Change the default green palette of the console.
3215 This is a 16-member array composed of values
3218 vt.default_red= [VT]
3219 Format: <red0>,<red1>,<red2>,...,<red15>
3220 Change the default red palette of the console.
3221 This is a 16-member array composed of values
3227 Set system-wide default UTF-8 mode for all tty's.
3228 Default is 1, i.e. UTF-8 mode is enabled for all
3229 newly opened terminals.
3231 vt.global_cursor_default=
3234 Set system-wide default for whether a cursor
3235 is shown on new VTs. Default is -1,
3236 i.e. cursors will be created by default unless
3237 overridden by individual drivers. 0 will hide
3238 cursors, 1 will display them.
3240 watchdog timers [HW,WDT] For information on watchdog timers,
3241 see Documentation/watchdog/watchdog-parameters.txt
3242 or other driver-specific files in the
3243 Documentation/watchdog/ directory.
3245 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
3246 default x2apic cluster mode on platforms
3249 x86_mrst_timer= [X86-32,APBT]
3250 Choose timer option for x86 Moorestown MID platform.
3251 Two valid options are apbt timer only and lapic timer
3252 plus one apbt timer for broadcast timer.
3253 x86_mrst_timer=apbt_only | lapic_and_apbt
3255 xd= [HW,XT] Original XT pre-IDE (RLL encoded) disks.
3256 xd_geo= See header of drivers/block/xd.c.
3258 xen_emul_unplug= [HW,X86,XEN]
3259 Unplug Xen emulated devices
3260 Format: [unplug0,][unplug1]
3261 ide-disks -- unplug primary master IDE devices
3262 aux-ide-disks -- unplug non-primary-master IDE devices
3263 nics -- unplug network devices
3264 all -- unplug all emulated devices (NICs and IDE disks)
3265 unnecessary -- unplugging emulated devices is
3266 unnecessary even if the host did not respond to
3268 never -- do not unplug even if version check succeeds
3270 xirc2ps_cs= [NET,PCMCIA]
3272 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
3274 ______________________________________________________________________
3278 Add more DRM drivers.