4 The following is a consolidated list of the kernel parameters as implemented
5 (mostly) by the __setup() macro and sorted into English Dictionary order
6 (defined as ignoring all punctuation and sorting digits before letters in a
7 case insensitive manner), and with descriptions where known.
9 Module parameters for loadable modules are specified only as the
10 parameter name with optional '=' and value as appropriate, such as:
12 modprobe usbcore blinkenlights=1
14 Module parameters for modules that are built into the kernel image
15 are specified on the kernel command line with the module name plus
16 '.' plus parameter name, with '=' and value if appropriate, such as:
18 usbcore.blinkenlights=1
20 Hyphens (dashes) and underscores are equivalent in parameter names, so
21 log_buf_len=1M print-fatal-signals=1
22 can also be entered as
23 log-buf-len=1M print_fatal_signals=1
26 This document may not be entirely up to date and comprehensive. The command
27 "modinfo -p ${modulename}" shows a current list of all parameters of a loadable
28 module. Loadable modules, after being loaded into the running kernel, also
29 reveal their parameters in /sys/module/${modulename}/parameters/. Some of these
30 parameters may be changed at runtime by the command
31 "echo -n ${value} > /sys/module/${modulename}/parameters/${parm}".
33 The parameters listed below are only valid if certain kernel build options were
34 enabled and if respective hardware is present. The text in square brackets at
35 the beginning of each description states the restrictions within which a
36 parameter is applicable:
38 ACPI ACPI support is enabled.
39 AGP AGP (Accelerated Graphics Port) is enabled.
40 ALSA ALSA sound support is enabled.
41 APIC APIC support is enabled.
42 APM Advanced Power Management support is enabled.
43 ARM ARM architecture is enabled.
44 AVR32 AVR32 architecture is enabled.
45 AX25 Appropriate AX.25 support is enabled.
46 BLACKFIN Blackfin architecture is enabled.
47 CLK Common clock infrastructure is enabled.
48 DRM Direct Rendering Management support is enabled.
49 DYNAMIC_DEBUG Build in debug messages and enable them at runtime
50 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled
51 EFI EFI Partitioning (GPT) is enabled
52 EIDE EIDE/ATAPI support is enabled.
53 EVM Extended Verification Module
54 FB The frame buffer device is enabled.
55 FTRACE Function tracing enabled.
56 GCOV GCOV profiling is enabled.
57 HW Appropriate hardware is enabled.
58 IA-64 IA-64 architecture is enabled.
59 IMA Integrity measurement architecture is enabled.
60 IOSCHED More than one I/O scheduler is enabled.
61 IP_PNP IP DHCP, BOOTP, or RARP is enabled.
62 IPV6 IPv6 support is enabled.
63 ISAPNP ISA PnP code is enabled.
64 ISDN Appropriate ISDN support is enabled.
65 JOY Appropriate joystick support is enabled.
66 KGDB Kernel debugger support is enabled.
67 KVM Kernel Virtual Machine support is enabled.
68 LIBATA Libata driver is enabled
69 LP Printer support is enabled.
70 LOOP Loopback device support is enabled.
71 M68k M68k architecture is enabled.
72 These options have more detailed description inside of
73 Documentation/m68k/kernel-options.txt.
74 MDA MDA console support is enabled.
75 MIPS MIPS architecture is enabled.
76 MOUSE Appropriate mouse support is enabled.
77 MSI Message Signaled Interrupts (PCI).
78 MTD MTD (Memory Technology Device) support is enabled.
79 NET Appropriate network support is enabled.
80 NUMA NUMA support is enabled.
81 NFS Appropriate NFS support is enabled.
82 OSS OSS sound support is enabled.
83 PV_OPS A paravirtualized kernel is enabled.
84 PARIDE The ParIDE (parallel port IDE) subsystem is enabled.
85 PARISC The PA-RISC architecture is enabled.
86 PCI PCI bus support is enabled.
87 PCIE PCI Express support is enabled.
88 PCMCIA The PCMCIA subsystem is enabled.
89 PNP Plug & Play support is enabled.
90 PPC PowerPC architecture is enabled.
91 PPT Parallel port support is enabled.
92 PS2 Appropriate PS/2 support is enabled.
93 RAM RAM disk support is enabled.
94 S390 S390 architecture is enabled.
95 SCSI Appropriate SCSI support is enabled.
96 A lot of drivers have their options described inside
97 the Documentation/scsi/ sub-directory.
98 SECURITY Different security models are enabled.
99 SELINUX SELinux support is enabled.
100 APPARMOR AppArmor support is enabled.
101 SERIAL Serial support is enabled.
102 SH SuperH architecture is enabled.
103 SMP The kernel is an SMP kernel.
104 SPARC Sparc architecture is enabled.
105 SWSUSP Software suspend (hibernation) is enabled.
106 SUSPEND System suspend states are enabled.
107 TPM TPM drivers are enabled.
108 TS Appropriate touchscreen support is enabled.
109 UMS USB Mass Storage support is enabled.
110 USB USB support is enabled.
111 USBHID USB Human Interface Device support is enabled.
112 V4L Video For Linux support is enabled.
113 VMMIO Driver for memory mapped virtio devices is enabled.
114 VGA The VGA console has been enabled.
115 VT Virtual terminal support is enabled.
116 WDT Watchdog support is enabled.
117 XT IBM PC/XT MFM hard disk support is enabled.
118 X86-32 X86-32, aka i386 architecture is enabled.
119 X86-64 X86-64 architecture is enabled.
120 More X86-64 boot options can be found in
121 Documentation/x86/x86_64/boot-options.txt .
122 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64)
123 XEN Xen support is enabled
125 In addition, the following text indicates that the option:
127 BUGS= Relates to possible processor bugs on the said processor.
128 KNL Is a kernel start-up parameter.
129 BOOT Is a boot loader parameter.
131 Parameters denoted with BOOT are actually interpreted by the boot
132 loader, and have no meaning to the kernel directly.
133 Do not modify the syntax of boot loader parameters without extreme
134 need or coordination with <Documentation/x86/boot.txt>.
136 There are also arch-specific kernel-parameters not documented here.
137 See for example <Documentation/x86/x86_64/boot-options.txt>.
139 Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
140 a trailing = on the name of any parameter states that that parameter will
141 be entered as an environment variable, whereas its absence indicates that
142 it will appear as a kernel argument readable via /proc/cmdline by programs
143 running once the system is up.
145 The number of kernel parameters is not limited, but the length of the
146 complete command line (parameters including spaces etc.) is limited to
147 a fixed number of characters. This limit depends on the architecture
148 and is between 256 and 4096 characters. It is defined in the file
149 ./include/asm/setup.h as COMMAND_LINE_SIZE.
151 Finally, the [KMG] suffix is commonly described after a number of kernel
152 parameter values. These 'K', 'M', and 'G' letters represent the _binary_
153 multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
154 bytes respectively. Such letter suffixes can also be entirely omitted.
158 Advanced Configuration and Power Interface
159 Format: { force | off | strict | noirq | rsdt }
160 force -- enable ACPI if default was off
161 off -- disable ACPI if default was on
162 noirq -- do not use ACPI for IRQ routing
163 strict -- Be less tolerant of platforms that are not
164 strictly ACPI specification compliant.
165 rsdt -- prefer RSDT over (default) XSDT
166 copy_dsdt -- copy DSDT to memory
168 See also Documentation/power/runtime_pm.txt, pci=noacpi
170 acpi_rsdp= [ACPI,EFI,KEXEC]
171 Pass the RSDP address to the kernel, mostly used
172 on machines running EFI runtime service to boot the
173 second kernel for kdump.
175 acpi_apic_instance= [ACPI, IOAPIC]
177 2: use 2nd APIC table, if available
178 1,0: use 1st APIC table
181 acpi_backlight= [HW,ACPI]
182 acpi_backlight=vendor
184 If set to vendor, prefer vendor specific driver
185 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
186 of the ACPI video.ko driver.
188 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
189 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
191 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
192 debug output. Bits in debug_layer correspond to a
193 _COMPONENT in an ACPI source file, e.g.,
194 #define _COMPONENT ACPI_PCI_COMPONENT
195 Bits in debug_level correspond to a level in
196 ACPI_DEBUG_PRINT statements, e.g.,
197 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
198 The debug_level mask defaults to "info". See
199 Documentation/acpi/debug.txt for more information about
200 debug layers and levels.
202 Enable processor driver info messages:
203 acpi.debug_layer=0x20000000
204 Enable PCI/PCI interrupt routing info messages:
205 acpi.debug_layer=0x400000
206 Enable AML "Debug" output, i.e., stores to the Debug
207 object while interpreting AML:
208 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
209 Enable all messages related to ACPI hardware:
210 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
212 Some values produce so much output that the system is
213 unusable. The "log_buf_len" parameter may be useful
214 if you need to capture more output.
216 acpi_irq_balance [HW,ACPI]
217 ACPI will balance active IRQs
220 acpi_irq_nobalance [HW,ACPI]
221 ACPI will not move active IRQs (default)
224 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
225 Format: <irq>,<irq>...
227 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
229 Format: <irq>,<irq>...
231 acpi_no_auto_ssdt [HW,ACPI] Disable automatic loading of SSDT
233 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
234 Format: To spoof as Windows 98: ="Microsoft Windows"
236 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
237 acpi_osi="string1" # add string1 -- only one string
238 acpi_osi="!string2" # remove built-in string2
239 acpi_osi= # disable all strings
242 Override the pmtimer bug detection: force the kernel
243 to assume that this machine's pmtimer latches its value
244 and always returns good values.
246 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
247 Format: { level | edge | high | low }
249 acpi_serialize [HW,ACPI] force serialization of AML methods
251 acpi_skip_timer_override [HW,ACPI]
252 Recognize and ignore IRQ0/pin2 Interrupt Override.
253 For broken nForce2 BIOS resulting in XT-PIC timer.
255 acpi_sleep= [HW,ACPI] Sleep options
256 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
257 old_ordering, nonvs, sci_force_enable }
258 See Documentation/power/video.txt for information on
260 s3_beep is for debugging; it makes the PC's speaker beep
261 as soon as the kernel's real-mode entry point is called.
262 s4_nohwsig prevents ACPI hardware signature from being
263 used during resume from hibernation.
264 old_ordering causes the ACPI 1.0 ordering of the _PTS
265 control method, with respect to putting devices into
266 low power states, to be enforced (the ACPI 2.0 ordering
267 of _PTS is used by default).
268 nonvs prevents the kernel from saving/restoring the
269 ACPI NVS memory during suspend/hibernation and resume.
270 sci_force_enable causes the kernel to set SCI_EN directly
271 on resume from S1/S3 (which is against the ACPI spec,
272 but some broken systems don't work without it).
274 acpi_use_timer_override [HW,ACPI]
275 Use timer override. For some broken Nvidia NF5 boards
276 that require a timer override, but don't have HPET
278 acpi_enforce_resources= [ACPI]
279 { strict | lax | no }
280 Check for resource conflicts between native drivers
281 and ACPI OperationRegions (SystemIO and SystemMemory
282 only). IO ports and memory declared in ACPI might be
283 used by the ACPI subsystem in arbitrary AML code and
284 can interfere with legacy drivers.
285 strict (default): access to resources claimed by ACPI
286 is denied; legacy drivers trying to access reserved
287 resources will fail to bind to device using them.
288 lax: access to resources claimed by ACPI is allowed;
289 legacy drivers trying to access reserved resources
290 will bind successfully but a warning message is logged.
291 no: ACPI OperationRegions are not marked as reserved,
292 no further checks are performed.
294 add_efi_memmap [EFI; X86] Include EFI memory map in
295 kernel's map of available physical RAM.
298 { off | try_unsupported }
299 off: disable AGP support
300 try_unsupported: try to drive unsupported chipsets
301 (may crash computer or cause data corruption)
304 See Documentation/sound/alsa/alsa-parameters.txt
307 Allow the default userspace alignment fault handler
308 behaviour to be specified. Bit 0 enables warnings,
309 bit 1 enables fixups, and bit 2 sends a segfault.
311 align_va_addr= [X86-64]
312 Align virtual addresses by clearing slice [14:12] when
313 allocating a VMA at process creation time. This option
314 gives you up to 3% performance improvement on AMD F15h
315 machines (where it is enabled by default) for a
316 CPU-intensive style benchmark, and it can vary highly in
317 a microbenchmark depending on workload and compiler.
319 32: only for 32-bit processes
320 64: only for 64-bit processes
321 on: enable for both 32- and 64-bit processes
322 off: disable for both 32- and 64-bit processes
324 alloc_snapshot [FTRACE]
325 Allocate the ftrace snapshot buffer on boot up when the
326 main buffer is allocated. This is handy if debugging
327 and you need to use tracing_snapshot() on boot up, and
328 do not want to use tracing_snapshot_alloc() as it needs
329 to be done where GFP_KERNEL allocations are allowed.
331 amd_iommu= [HW,X86-64]
332 Pass parameters to the AMD IOMMU driver in the system.
334 fullflush - enable flushing of IO/TLB entries when
335 they are unmapped. Otherwise they are
336 flushed before they will be reused, which
338 off - do not initialize any AMD IOMMU found in
340 force_isolation - Force device isolation for all
341 devices. The IOMMU driver is not
342 allowed anymore to lift isolation
343 requirements as needed. This option
344 does not override iommu=pt
346 amd_iommu_dump= [HW,X86-64]
347 Enable AMD IOMMU driver option to dump the ACPI table
348 for AMD IOMMU. With this option enabled, AMD IOMMU
349 driver will print ACPI tables for AMD IOMMU during
350 IOMMU initialization.
352 amijoy.map= [HW,JOY] Amiga joystick support
353 Map of devices attached to JOY0DAT and JOY1DAT
355 See also Documentation/input/joystick.txt
357 analog.map= [HW,JOY] Analog joystick and gamepad support
358 Specifies type or capabilities of an analog joystick
359 connected to one of 16 gameports
360 Format: <type1>,<type2>,..<type16>
363 Power management functions (SPARCstation-4/5 + deriv.)
365 Disable APC CPU standby support. SPARCstation-Fox does
366 not play well with APC CPU idle - disable it if you have
367 APC and your system crashes randomly.
369 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
370 Change the output verbosity whilst booting
371 Format: { quiet (default) | verbose | debug }
372 Change the amount of debugging information output
373 when initialising the APIC and IO-APIC components.
376 See Documentation/networking/ipv6.txt.
378 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
379 Limit apic dumping. The parameter defines the maximal
380 number of local apics being dumped. Also it is possible
381 to set it to "all" by meaning -- no limit here.
382 Format: { 1 (default) | 2 | ... | all }.
383 The parameter valid if only apic=debug or
384 apic=verbose is specified.
385 Example: apic=debug show_lapic=all
387 apm= [APM] Advanced Power Management
388 See header of arch/x86/kernel/apm_32.c.
390 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
391 Format: <io>,<irq>,<nodeID>
395 atarimouse= [HW,MOUSE] Atari Mouse
397 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
398 EzKey and similar keyboards
400 atkbd.reset= [HW] Reset keyboard during initialization
402 atkbd.set= [HW] Select keyboard code set
403 Format: <int> (2 = AT (default), 3 = PS/2)
405 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
408 atkbd.softraw= [HW] Choose between synthetic and real raw mode
409 Format: <bool> (0 = real, 1 = synthetic (default))
411 atkbd.softrepeat= [HW]
412 Use software keyboard repeat
414 baycom_epp= [HW,AX25]
417 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
419 See header of drivers/net/hamradio/baycom_par.c.
421 baycom_ser_fdx= [HW,AX25]
422 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
423 Format: <io>,<irq>,<mode>[,<baud>]
424 See header of drivers/net/hamradio/baycom_ser_fdx.c.
426 baycom_ser_hdx= [HW,AX25]
427 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
428 Format: <io>,<irq>,<mode>
429 See header of drivers/net/hamradio/baycom_ser_hdx.c.
431 boot_delay= Milliseconds to delay each printk during boot.
432 Values larger than 10 seconds (10000) are changed to
436 bootmem_debug [KNL] Enable bootmem allocator debug messages.
438 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
439 bttv.radio= Most important insmod options are available as
441 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
444 bulk_remove=off [PPC] This parameter disables the use of the pSeries
445 firmware feature for flushing multiple hpte entries
448 c101= [NET] Moxa C101 synchronous serial card
450 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
451 Sometimes CPU hardware bugs make them report the cache
452 size incorrectly. The kernel will attempt work arounds
453 to fix known problems, but for some CPUs it is not
454 possible to determine what the correct size should be.
455 This option provides an override for these situations.
457 ccw_timeout_log [S390]
458 See Documentation/s390/CommonIO for details.
460 cgroup_disable= [KNL] Disable a particular controller
461 Format: {name of the controller(s) to disable}
462 {Currently supported controllers - "memory"}
464 checkreqprot [SELINUX] Set initial checkreqprot flag value.
465 Format: { "0" | "1" }
466 See security/selinux/Kconfig help text.
467 0 -- check protection applied by kernel (includes
468 any implied execute protection).
469 1 -- check protection requested by application.
470 Default value is set via a kernel config option.
471 Value can be changed at runtime via
472 /selinux/checkreqprot.
475 See Documentation/s390/CommonIO for details.
478 Keep all clocks already enabled by bootloader on,
479 even if no driver has claimed them. This is useful
480 for debug and development, but should not be
481 needed on a platform with proper driver support.
482 For more information, see Documentation/clk.txt.
484 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
486 Forces specified clocksource (if available) to be used
487 when calculating gettimeofday(). If specified
488 clocksource is not available, it defaults to PIT.
489 Format: { pit | tsc | cyclone | pmtmr }
491 clocksource= Override the default clocksource
493 Override the default clocksource and use the clocksource
494 with the name specified.
495 Some clocksource names to choose from, depending on
497 [all] jiffies (this is the base, fallback clocksource)
499 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
500 pxa_timer,timer3,32k_counter,timer0_1
502 [X86-32] pit,hpet,tsc;
503 scx200_hrt on Geode; cyclone on IBM x440
511 clearcpuid=BITNUM [X86]
512 Disable CPUID feature X for the kernel. See
513 arch/x86/include/asm/cpufeature.h for the valid bit
514 numbers. Note the Linux specific bits are not necessarily
515 stable over kernel options, but the vendor specific
517 Also note that user programs calling CPUID directly
518 or using the feature without checking anything
519 will still see it. This just prevents it from
520 being used by the kernel or shown in /proc/cpuinfo.
521 Also note the kernel might malfunction if you disable
525 Sets the size of kernel global memory area for contiguous
526 memory allocations. For more information, see
527 include/linux/dma-contiguous.h
529 cmo_free_hint= [PPC] Format: { yes | no }
530 Specify whether pages are marked as being inactive
531 when they are freed. This is used in CMO environments
532 to determine OS memory pressure for page stealing by
536 coherent_pool=nn[KMG] [ARM,KNL]
537 Sets the size of memory pool for coherent, atomic dma
538 allocations, by default set to 256K.
540 code_bytes [X86] How many bytes of object code to print
545 com20020= [HW,NET] ARCnet - COM20020 chipset
547 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
549 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
553 ARCnet - COM90xx chipset (memory-mapped buffers)
554 Format: <io>[,<irq>[,<memstart>]]
556 condev= [HW,S390] console device
559 console= [KNL] Output console device and options.
561 tty<n> Use the virtual console device <n>.
565 Use the specified serial port. The options are of
566 the form "bbbbpnf", where "bbbb" is the baud rate,
567 "p" is parity ("n", "o", or "e"), "n" is number of
568 bits, and "f" is flow control ("r" for RTS or
569 omit it). Default is "9600n8".
571 See Documentation/serial-console.txt for more
573 Documentation/networking/netconsole.txt for an
576 uart[8250],io,<addr>[,options]
577 uart[8250],mmio,<addr>[,options]
578 Start an early, polled-mode console on the 8250/16550
579 UART at the specified I/O port or MMIO address,
580 switching to the matching ttyS device later. The
581 options are the same as for ttyS, above.
582 hvc<n> Use the hypervisor console device <n>. This is for
583 both Xen and PowerPC hypervisors.
585 If the device connected to the port is not a TTY but a braille
586 device, prepend "brl," before the device type, for instance
588 For now, only VisioBraille is supported.
590 consoleblank= [KNL] The console blank (screen saver) timeout in
591 seconds. Defaults to 10*60 = 10mins. A value of 0
592 disables the blank timer.
595 [KNL] Change the default value for
596 /proc/<pid>/coredump_filter.
597 See also Documentation/filesystems/proc.txt.
599 cpuidle.off=1 [CPU_IDLE]
600 disable the cpuidle sub-system
602 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
604 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
606 crashkernel=size[KMG][@offset[KMG]]
607 [KNL] Using kexec, Linux can switch to a 'crash kernel'
608 upon panic. This parameter reserves the physical
609 memory region [offset, offset + size] for that kernel
610 image. If '@offset' is omitted, then a suitable offset
611 is selected automatically. Check
612 Documentation/kdump/kdump.txt for further details.
614 crashkernel=range1:size1[,range2:size2,...][@offset]
615 [KNL] Same as above, but depends on the memory
616 in the running system. The syntax of range is
617 start-[end] where start and end are both
618 a memory unit (amount[KMG]). See also
619 Documentation/kdump/kdump.txt for an example.
621 crashkernel=size[KMG],high
622 [KNL, x86_64] range could be above 4G. Allow kernel
623 to allocate physical memory region from top, so could
624 be above 4G if system have more than 4G ram installed.
625 Otherwise memory region will be allocated below 4G, if
627 It will be ignored if crashkernel=X is specified.
628 crashkernel=size[KMG],low
629 [KNL, x86_64] range under 4G. When crashkernel=X,high
630 is passed, kernel could allocate physical memory region
631 above 4G, that cause second kernel crash on system
632 that require some amount of low memory, e.g. swiotlb
633 requires at least 64M+32K low memory. Kernel would
634 try to allocate 72M below 4G automatically.
635 This one let user to specify own low range under 4G
636 for second kernel instead.
637 0: to disable low allocation.
638 It will be ignored when crashkernel=X,high is not used
639 or memory reserved is below 4G.
644 cs89x0_media= [HW,NET]
645 Format: { rj45 | aui | bnc }
648 See header of drivers/s390/block/dasd_devmap.c.
650 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
651 (one device per port)
652 Format: <port#>,<type>
653 See also Documentation/input/joystick-parport.txt
655 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
656 time. See Documentation/dynamic-debug-howto.txt for
657 details. Deprecated, see dyndbg.
659 debug [KNL] Enable kernel debugging (events log level).
662 [KNL] verbose self-tests
664 Print debugging info while doing the locking API
666 We default to 0 (no extra messages), setting it to
667 1 will print _a lot_ more information - normally
668 only useful to kernel developers.
670 debug_objects [KNL] Enable object debugging
673 [KNL] Disable object debugging
675 debug_guardpage_minorder=
676 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
677 parameter allows control of the order of pages that will
678 be intentionally kept free (and hence protected) by the
679 buddy allocator. Bigger value increase the probability
680 of catching random memory corruption, but reduce the
681 amount of memory for normal system use. The maximum
682 possible value is MAX_ORDER/2. Setting this parameter
683 to 1 or 2 should be enough to identify most random
684 memory corruption problems caused by bugs in kernel or
685 driver code when a CPU writes to (or reads from) a
686 random memory location. Note that there exists a class
687 of memory corruptions problems caused by buggy H/W or
688 F/W or by drivers badly programing DMA (basically when
689 memory is written at bus level and the CPU MMU is
690 bypassed) which are not detectable by
691 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
692 tracking down these problems.
694 debugpat [X86] Enable PAT debugging
696 decnet.addr= [HW,NET]
697 Format: <area>[,<node>]
698 See also Documentation/networking/decnet.txt.
701 [same as hugepagesz=] The size of the default
702 HugeTLB page size. This is the size represented by
703 the legacy /proc/ hugepages APIs, used for SHM, and
704 default size when mounting hugetlbfs filesystems.
705 Defaults to the default architecture's huge page size
709 Set number of hash buckets for dentry cache.
712 IO parameters + enable/disable command.
714 digiepca= [HW,SERIAL]
715 See drivers/char/README.epca and
716 Documentation/serial/digiepca.txt.
719 See Documentation/networking/ipv6.txt.
721 disable_ddw [PPC/PSERIES]
722 Disable Dynamic DMA Window support. Use this if
723 to workaround buggy firmware.
726 See Documentation/networking/ipv6.txt.
728 disable_mtrr_cleanup [X86]
729 The kernel tries to adjust MTRR layout from continuous
730 to discrete, to make X server driver able to add WB
731 entry later. This parameter disables that.
733 disable_mtrr_trim [X86, Intel and AMD only]
734 By default the kernel will trim any uncacheable
735 memory out of your available memory pool based on
736 MTRR settings. This parameter disables that behavior,
737 possibly causing your machine to run very slowly.
739 disable_timer_pin_1 [X86]
740 Disable PIN 1 of APIC timer
741 Can be useful to work around chipset bugs.
743 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
744 this option disables the debugging code at boot.
746 dma_debug_entries=<number>
747 This option allows to tune the number of preallocated
748 entries for DMA-API debugging code. One entry is
749 required per DMA-API allocation. Use this if the
750 DMA-API debugging code disables itself because the
751 architectural default is too low.
753 dma_debug_driver=<driver_name>
754 With this option the DMA-API debugging driver
755 filter feature can be enabled at boot time. Just
756 pass the driver to filter for as the parameter.
757 The filter can be disabled or changed to another
758 driver later using sysfs.
760 drm_kms_helper.edid_firmware=[<connector>:]<file>
761 Broken monitors, graphic adapters and KVMs may
762 send no or incorrect EDID data sets. This parameter
763 allows to specify an EDID data set in the
764 /lib/firmware directory that is used instead.
765 Generic built-in EDID data sets are used, if one of
766 edid/1024x768.bin, edid/1280x1024.bin,
767 edid/1680x1050.bin, or edid/1920x1080.bin is given
768 and no file with the same name exists. Details and
769 instructions how to build your own EDID data are
770 available in Documentation/EDID/HOWTO.txt. An EDID
771 data set will only be used for a particular connector,
772 if its name and a colon are prepended to the EDID
777 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
778 module.dyndbg[="val"]
779 Enable debug messages at boot time. See
780 Documentation/dynamic-debug-howto.txt for details.
782 earlycon= [KNL] Output early console device and options.
783 uart[8250],io,<addr>[,options]
784 uart[8250],mmio,<addr>[,options]
785 uart[8250],mmio32,<addr>[,options]
786 Start an early, polled-mode console on the 8250/16550
787 UART at the specified I/O port or MMIO address.
788 MMIO inter-register address stride is either 8-bit
789 (mmio) or 32-bit (mmio32).
790 The options are the same as for ttyS, above.
792 earlyprintk= [X86,SH,BLACKFIN,ARM]
795 earlyprintk=serial[,ttySn[,baudrate]]
796 earlyprintk=serial[,0x...[,baudrate]]
797 earlyprintk=ttySn[,baudrate]
798 earlyprintk=dbgp[debugController#]
800 earlyprintk is useful when the kernel crashes before
801 the normal console is initialized. It is not enabled by
802 default because it has some cosmetic problems.
804 Append ",keep" to not disable it when the real console
807 Only vga or serial or usb debug port at a time.
809 Currently only ttyS0 and ttyS1 may be specified by
810 name. Other I/O ports may be explicitly specified
811 on some architectures (x86 and arm at least) by
812 replacing ttySn with an I/O port address, like this:
813 earlyprintk=serial,0x1008,115200
814 You can find the port for a given device in
815 /proc/tty/driver/serial:
816 2: uart:ST16650V2 port:00001008 irq:18 ...
818 Interaction with the standard serial driver is not
821 The VGA output is eventually overwritten by the real
824 The xen output can only be used by Xen PV guests.
826 ekgdboc= [X86,KGDB] Allow early kernel console debugging
829 This is designed to be used in conjunction with
830 the boot argument: earlyprintk=vga
833 Format: {"off" | "on" | "skip[mbr]"}
835 efi_no_storage_paranoia [EFI; X86]
836 Using this parameter you can use more than 50% of
837 your efi variable storage. Use this parameter only if
838 you are really sure that your UEFI does sane gc and
839 fulfills the spec otherwise your board may brick.
841 eisa_irq_edge= [PARISC,HW]
842 See header of drivers/parisc/eisa.c.
845 See comment before function elanfreq_setup() in
846 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
849 Format: {"cfq" | "deadline" | "noop"}
850 See Documentation/block/cfq-iosched.txt and
851 Documentation/block/deadline-iosched.txt for details.
853 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
854 Specifies physical address of start of kernel core
855 image elf header and optionally the size. Generally
856 kexec loader will pass this option to capture kernel.
857 See Documentation/kdump/kdump.txt for details.
859 enable_mtrr_cleanup [X86]
860 The kernel tries to adjust MTRR layout from continuous
861 to discrete, to make X server driver able to add WB
862 entry later. This parameter enables that.
864 enable_timer_pin_1 [X86]
865 Enable PIN 1 of APIC timer
866 Can be useful to work around chipset bugs
867 (in particular on some ATI chipsets).
868 The kernel tries to set a reasonable default.
870 enforcing [SELINUX] Set initial enforcing status.
872 See security/selinux/Kconfig help text.
873 0 -- permissive (log only, no denials).
874 1 -- enforcing (deny and log).
876 Value can be changed at runtime via /selinux/enforce.
879 Disable Error Record Serialization Table (ERST)
882 ether= [HW,NET] Ethernet cards parameters
883 This option is obsoleted by the "netdev=" option, which
884 has equivalent usage. See its documentation for details.
888 Permit 'security.evm' to be updated regardless of
889 current integrity status.
893 fail_make_request=[KNL]
894 General fault injection mechanism.
895 Format: <interval>,<probability>,<space>,<times>
896 See also Documentation/fault-injection/.
899 See Documentation/blockdev/floppy.txt.
901 force_pal_cache_flush
902 [IA-64] Avoid check_sal_cache_flush which may hang on
903 buggy SAL_CACHE_FLUSH implementations. Using this
904 parameter will force ia64_sal_cache_flush to call
905 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
908 [FTRACE] will set and start the specified tracer
909 as early as possible in order to facilitate early
912 ftrace_dump_on_oops[=orig_cpu]
913 [FTRACE] will dump the trace buffers on oops.
914 If no parameter is passed, ftrace will dump
915 buffers of all CPUs, but if you pass orig_cpu, it will
916 dump only the buffer of the CPU that triggered the
919 ftrace_filter=[function-list]
920 [FTRACE] Limit the functions traced by the function
921 tracer at boot up. function-list is a comma separated
922 list of functions. This list can be changed at run
923 time by the set_ftrace_filter file in the debugfs
926 ftrace_notrace=[function-list]
927 [FTRACE] Do not trace the functions specified in
928 function-list. This list can be changed at run time
929 by the set_ftrace_notrace file in the debugfs
932 ftrace_graph_filter=[function-list]
933 [FTRACE] Limit the top level callers functions traced
934 by the function graph tracer at boot up.
935 function-list is a comma separated list of functions
936 that can be changed at run time by the
937 set_graph_function file in the debugfs tracing directory.
940 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
941 support via parallel port (up to 5 devices per port)
942 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
943 See also Documentation/input/joystick-parport.txt
947 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
951 gcov_persist= [GCOV] When non-zero (default), profiling data for
952 kernel modules is saved and remains accessible via
953 debugfs, even when the module is unloaded/reloaded.
954 When zero, profiling data is discarded and associated
955 debugfs files are removed at module unload time.
957 gpt [EFI] Forces disk with valid GPT signature but
958 invalid Protective MBR to be treated as GPT.
960 grcan.enable0= [HW] Configuration of physical interface 0. Determines
961 the "Enable 0" bit of the configuration register.
964 grcan.enable1= [HW] Configuration of physical interface 1. Determines
965 the "Enable 0" bit of the configuration register.
968 grcan.select= [HW] Select which physical interface to use.
971 grcan.txsize= [HW] Sets the size of the tx buffer.
972 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
974 grcan.rxsize= [HW] Sets the size of the rx buffer.
975 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
978 hashdist= [KNL,NUMA] Large hashes allocated during boot
979 are distributed across NUMA nodes. Defaults on
980 for 64-bit NUMA, off otherwise.
981 Format: 0 | 1 (for off | on)
983 hcl= [IA-64] SGI's Hardware Graph compatibility layer
985 hd= [EIDE] (E)IDE hard drive subsystem geometry
986 Format: <cyl>,<head>,<sect>
989 Disable Hardware Error Source Table (HEST) support;
990 corresponding firmware-first mode error processing
991 logic will be disabled.
993 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
994 size of <nn>. This works even on boxes that have no
995 highmem otherwise. This also works to reduce highmem
996 size on bigger boxes.
998 highres= [KNL] Enable/disable high resolution timer mode.
999 Valid parameters: "on", "off"
1003 See Documentation/isdn/README.HiSax.
1007 hpet= [X86-32,HPET] option to control HPET usage
1008 Format: { enable (default) | disable | force |
1010 disable: disable HPET and use PIT instead
1011 force: allow force enabled of undocumented chips (ICH4,
1013 verbose: show contents of HPET registers during setup
1015 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1016 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1017 On x86-64 and powerpc, this option can be specified
1018 multiple times interleaved with hugepages= to reserve
1019 huge pages of different sizes. Valid pages sizes on
1020 x86-64 are 2M (when the CPU supports "pse") and 1G
1021 (when the CPU supports the "pdpe1gb" cpuinfo flag)
1022 Note that 1GB pages can only be allocated at boot time
1023 using hugepages= and not freed afterwards.
1025 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1026 terminal devices. Valid values: 0..8
1027 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1028 If specified, z/VM IUCV HVC accepts connections
1029 from listed z/VM user IDs only.
1031 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1032 hardware thread id mappings.
1033 Format: <cpu>:<hwthread>
1036 Do not unregister boot console at start. This is only
1037 useful for debugging when something happens in the window
1038 between unregistering the boot console and initializing
1041 i2c_bus= [HW] Override the default board specific I2C bus speed
1042 or register an additional I2C bus that is not
1043 registered from board initialization code.
1047 i8042.debug [HW] Toggle i8042 debug mode
1048 i8042.direct [HW] Put keyboard port into non-translated mode
1049 i8042.dumbkbd [HW] Pretend that controller can only read data from
1050 keyboard and cannot control its state
1051 (Don't attempt to blink the leds)
1052 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1053 i8042.nokbd [HW] Don't check/create keyboard port
1054 i8042.noloop [HW] Disable the AUX Loopback command while probing
1056 i8042.nomux [HW] Don't check presence of an active multiplexing
1058 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1060 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1061 i8042.reset [HW] Reset the controller during init and cleanup
1062 i8042.unlock [HW] Unlock (ignore) the keylock
1066 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1067 indicates that the driver is running on unsupported
1069 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1070 does not match list of supported models.
1072 [HW] Report power status in /proc/i8k
1073 (disabled by default)
1074 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1077 i915.invert_brightness=
1078 [DRM] Invert the sense of the variable that is used to
1079 set the brightness of the panel backlight. Normally a
1080 brightness value of 0 indicates backlight switched off,
1081 and the maximum of the brightness value sets the backlight
1082 to maximum brightness. If this parameter is set to 0
1083 (default) and the machine requires it, or this parameter
1084 is set to 1, a brightness value of 0 sets the backlight
1085 to maximum brightness, and the maximum of the brightness
1086 value switches the backlight off.
1087 -1 -- never invert brightness
1088 0 -- machine default
1089 1 -- force brightness inversion
1092 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1094 ide-core.nodma= [HW] (E)IDE subsystem
1095 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1096 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1097 .cdrom .chs .ignore_cable are additional options
1098 See Documentation/ide/ide.txt.
1100 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1101 Claim all unknown PCI IDE storage controllers.
1104 Format: idle=poll, idle=halt, idle=nomwait
1105 Poll forces a polling idle loop that can slightly
1106 improve the performance of waking up a idle CPU, but
1107 will use a lot of power and make the system run hot.
1109 idle=halt: Halt is forced to be used for CPU idle.
1110 In such case C2/C3 won't be used again.
1111 idle=nomwait: Disable mwait for CPU C-states
1113 ignore_loglevel [KNL]
1114 Ignore loglevel setting - this will print /all/
1115 kernel messages to the console. Useful for debugging.
1116 We also add it as printk module parameter, so users
1117 could change it dynamically, usually by
1118 /sys/module/printk/parameters/ignore_loglevel.
1120 ihash_entries= [KNL]
1121 Set number of hash buckets for inode cache.
1123 ima_appraise= [IMA] appraise integrity measurements
1124 Format: { "off" | "enforce" | "fix" }
1127 ima_appraise_tcb [IMA]
1128 The builtin appraise policy appraises all files
1132 Format: { "0" | "1" }
1133 0 -- integrity auditing messages. (Default)
1134 1 -- enable informational integrity auditing messages.
1137 Format: { "sha1" | "md5" }
1141 Load a policy which meets the needs of the Trusted
1142 Computing Base. This means IMA will measure all
1143 programs exec'd, files mmap'd for exec, and all files
1144 opened for read by uid=0.
1148 Run specified binary instead of /sbin/init as init
1151 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1152 for working out where the kernel is dying during
1155 initrd= [BOOT] Specify the location of the initial ramdisk
1157 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1160 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1162 Enable intel iommu driver.
1164 Disable intel iommu driver.
1165 igfx_off [Default Off]
1166 By default, gfx is mapped as normal device. If a gfx
1167 device has a dedicated DMAR unit, the DMAR unit is
1168 bypassed by not enabling DMAR with this option. In
1169 this case, gfx device will use physical address for
1172 With this option iommu will not optimize to look
1173 for io virtual address below 32-bit forcing dual
1174 address cycle on pci bus for cards supporting greater
1175 than 32-bit addressing. The default is to look
1176 for translation below 32-bit and if not available
1177 then look in the higher range.
1178 strict [Default Off]
1179 With this option on every unmap_single operation will
1180 result in a hardware IOTLB flush operation as opposed
1181 to batching them for performance.
1182 sp_off [Default Off]
1183 By default, super page will be supported if Intel IOMMU
1184 has the capability. With this option, super page will
1187 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1188 0 disables intel_idle and fall back on acpi_idle.
1189 1 to 6 specify maximum depth of C-state.
1193 Do not enable intel_pstate as the default
1194 scaling driver for the supported processors
1196 intremap= [X86-64, Intel-IOMMU]
1197 on enable Interrupt Remapping (default)
1198 off disable Interrupt Remapping
1199 nosid disable Source ID checking
1201 BIOS x2APIC opt-out request will be ignored
1203 iomem= Disable strict checking of access to MMIO memory
1204 strict regions from userspace.
1221 io7= [HW] IO7 for Marvel based alpha systems
1222 See comment before marvel_specify_io7 in
1223 arch/alpha/kernel/core_marvel.c.
1225 io_delay= [X86] I/O delay method
1227 Standard port 0x80 based delay
1229 Alternate port 0xed based delay (needed on some systems)
1231 Simple two microseconds delay
1236 See Documentation/filesystems/nfs/nfsroot.txt.
1238 ip2= [HW] Set IO/IRQ pairs for up to 4 IntelliPort boards
1239 See comment before ip2_setup() in
1240 drivers/char/ip2/ip2base.c.
1243 When an interrupt is not handled search all handlers
1244 for it. Intended to get systems with badly broken
1248 When an interrupt is not handled search all handlers
1249 for it. Also check all handlers each timer
1250 interrupt. Intended to get systems with badly broken
1254 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1256 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1258 <cpu number>,...,<cpu number>
1260 <cpu number>-<cpu number>
1261 (must be a positive range in ascending order)
1263 <cpu number>,...,<cpu number>-<cpu number>
1265 This option can be used to specify one or more CPUs
1266 to isolate from the general SMP balancing and scheduling
1267 algorithms. You can move a process onto or off an
1268 "isolated" CPU via the CPU affinity syscalls or cpuset.
1269 <cpu number> begins at 0 and the maximum value is
1270 "number of CPUs in system - 1".
1272 This option is the preferred way to isolate CPUs. The
1273 alternative -- manually setting the CPU mask of all
1274 tasks in the system -- can cause problems and
1275 suboptimal load balancer performance.
1279 js= [HW,JOY] Analog joystick
1280 See Documentation/input/joystick.txt.
1284 kernelcore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1285 specifies the amount of memory usable by the kernel
1286 for non-movable allocations. The requested amount is
1287 spread evenly throughout all nodes in the system. The
1288 remaining memory in each node is used for Movable
1289 pages. In the event, a node is too small to have both
1290 kernelcore and Movable pages, kernelcore pages will
1291 take priority and other nodes will have a larger number
1292 of kernelcore pages. The Movable zone is used for the
1293 allocation of pages that may be reclaimed or moved
1294 by the page migration subsystem. This means that
1295 HugeTLB pages may not be allocated from this zone.
1296 Note that allocations like PTEs-from-HighMem still
1297 use the HighMem zone if it exists, and the Normal
1298 zone if it does not.
1300 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1301 Format: <Controller#>[,poll interval]
1302 The controller # is the number of the ehci usb debug
1303 port as it is probed via PCI. The poll interval is
1304 optional and is the number seconds in between
1305 each poll cycle to the debug port in case you need
1306 the functionality for interrupting the kernel with
1307 gdb or control-c on the dbgp connection. When
1308 not using this parameter you use sysrq-g to break into
1309 the kernel debugger.
1311 kgdboc= [KGDB,HW] kgdb over consoles.
1312 Requires a tty driver that supports console polling,
1313 or a supported polling keyboard driver (non-usb).
1314 Serial only format: <serial_device>[,baud]
1315 keyboard only format: kbd
1316 keyboard and serial format: kbd,<serial_device>[,baud]
1317 Optional Kernel mode setting:
1318 kms, kbd format: kms,kbd
1319 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1321 kgdbwait [KGDB] Stop kernel execution and enter the
1322 kernel debugger at the earliest opportunity.
1324 kmac= [MIPS] korina ethernet MAC address.
1325 Configure the RouterBoard 532 series on-chip
1326 Ethernet adapter MAC address.
1328 kmemleak= [KNL] Boot-time kmemleak enable/disable
1329 Valid arguments: on, off
1332 kstack=N [X86] Print N words from the kernel stack
1335 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1336 Default is 0 (don't ignore, but inject #GP)
1338 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1342 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1343 Default is 1 (enabled)
1345 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1347 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1349 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1350 (virtualized MMU) support on capable Intel chips.
1351 Default is 1 (enabled)
1353 kvm-intel.emulate_invalid_guest_state=
1354 [KVM,Intel] Enable emulation of invalid guest states
1355 Default is 0 (disabled)
1357 kvm-intel.flexpriority=
1358 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1359 Default is 1 (enabled)
1362 [KVM,Intel] Enable VMX nesting (nVMX).
1363 Default is 0 (disabled)
1365 kvm-intel.unrestricted_guest=
1366 [KVM,Intel] Disable unrestricted guest feature
1367 (virtualized real and unpaged mode) on capable
1368 Intel chips. Default is 1 (enabled)
1370 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1371 feature (tagged TLBs) on capable Intel chips.
1372 Default is 1 (enabled)
1378 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1381 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1382 value for LAPIC timer one-shot implementation. Default
1383 back to the programmable timer unit in the LAPIC.
1385 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1388 libata.dma= [LIBATA] DMA control
1389 libata.dma=0 Disable all PATA and SATA DMA
1390 libata.dma=1 PATA and SATA Disk DMA only
1391 libata.dma=2 ATAPI (CDROM) DMA only
1392 libata.dma=4 Compact Flash DMA only
1393 Combinations also work, so libata.dma=3 enables DMA
1394 for disks and CDROMs, but not CFs.
1396 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1397 libata.ignore_hpa=0 keep BIOS limits (default)
1398 libata.ignore_hpa=1 ignore limits, using full disk
1400 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
1404 libata.force= [LIBATA] Force configurations. The format is comma
1405 separated list of "[ID:]VAL" where ID is
1406 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
1407 matching port, link or device. Basically, it matches
1408 the ATA ID string printed on console by libata. If
1409 the whole ID part is omitted, the last PORT and DEVICE
1410 values are used. If ID hasn't been specified yet, the
1411 configuration applies to all ports, links and devices.
1413 If only DEVICE is omitted, the parameter applies to
1414 the port and all links and devices behind it. DEVICE
1415 number of 0 either selects the first device or the
1416 first fan-out link behind PMP device. It does not
1417 select the host link. DEVICE number of 15 selects the
1418 host link and device attached to it.
1420 The VAL specifies the configuration to force. As long
1421 as there's no ambiguity shortcut notation is allowed.
1422 For example, both 1.5 and 1.5G would work for 1.5Gbps.
1423 The following configurations can be forced.
1425 * Cable type: 40c, 80c, short40c, unk, ign or sata.
1426 Any ID with matching PORT is used.
1428 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
1430 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
1431 udma[/][16,25,33,44,66,100,133] notation is also
1434 * [no]ncq: Turn on or off NCQ.
1436 * nohrst, nosrst, norst: suppress hard, soft
1439 * rstonce: only attempt one reset during
1440 hot-unplug link recovery
1442 * dump_id: dump IDENTIFY data.
1444 If there are multiple matching configurations changing
1445 the same attribute, the last one is used.
1447 memblock=debug [KNL] Enable memblock debug messages.
1449 load_ramdisk= [RAM] List of ramdisks to load from floppy
1450 See Documentation/blockdev/ramdisk.txt.
1452 lockd.nlm_grace_period=P [NFS] Assign grace period.
1455 lockd.nlm_tcpport=N [NFS] Assign TCP port.
1458 lockd.nlm_timeout=T [NFS] Assign timeout value.
1461 lockd.nlm_udpport=M [NFS] Assign UDP port.
1464 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
1467 loglevel= All Kernel Messages with a loglevel smaller than the
1468 console loglevel will be printed to the console. It can
1469 also be changed with klogd or other programs. The
1470 loglevels are defined as follows:
1472 0 (KERN_EMERG) system is unusable
1473 1 (KERN_ALERT) action must be taken immediately
1474 2 (KERN_CRIT) critical conditions
1475 3 (KERN_ERR) error conditions
1476 4 (KERN_WARNING) warning conditions
1477 5 (KERN_NOTICE) normal but significant condition
1478 6 (KERN_INFO) informational
1479 7 (KERN_DEBUG) debug-level messages
1481 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
1482 in bytes. n must be a power of two. The default
1483 size is set in the kernel config file.
1485 logo.nologo [FB] Disables display of the built-in Linux logo.
1486 This may be used to provide more screen space for
1487 kernel log messages and is useful when debugging
1488 kernel boot problems.
1490 lp=0 [LP] Specify parallel ports to use, e.g,
1491 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
1492 lp=reset first parallel port). 'lp=0' disables the
1493 lp=auto printer driver. 'lp=reset' (which can be
1494 specified in addition to the ports) causes
1495 attached printers to be reset. Using
1496 lp=port1,port2,... specifies the parallel ports
1497 to associate lp devices with, starting with
1498 lp0. A port specification may be 'none' to skip
1499 that lp device, or a parport name such as
1500 'parport0'. Specifying 'lp=auto' instead of a
1501 port specification list means that device IDs
1502 from each port should be examined, to see if
1503 an IEEE 1284-compliant printer is attached; if
1504 so, the driver will manage that printer.
1505 See also header of drivers/char/lp.c.
1508 Sets loops_per_jiffy to given constant, thus avoiding
1509 time-consuming boot-time autodetection (up to 250 ms per
1510 CPU). 0 enables autodetection (default). To determine
1511 the correct value for your kernel, boot with normal
1512 autodetection and see what value is printed. Note that
1513 on SMP systems the preset will be applied to all CPUs,
1514 which is likely to cause problems if your CPUs need
1515 significantly divergent settings. An incorrect value
1516 will cause delays in the kernel to be wrong, leading to
1517 unpredictable I/O errors and other breakage. Although
1518 unlikely, in the extreme case this might damage your
1522 Format: <io>,<irq>,<dma>
1524 machvec= [IA-64] Force the use of a particular machine-vector
1525 (machvec) in a generic kernel.
1526 Example: machvec=hpzx1_swiotlb
1528 machtype= [Loongson] Share the same kernel image file between different
1530 Example: machtype=lemote-yeeloong-2f-7inch
1532 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
1533 than or equal to this physical address is ignored.
1535 maxcpus= [SMP] Maximum number of processors that an SMP kernel
1536 should make use of. maxcpus=n : n >= 0 limits the
1537 kernel to using 'n' processors. n=0 is a special case,
1538 it is equivalent to "nosmp", which also disables
1541 max_loop= [LOOP] The number of loop block devices that get
1542 (loop.max_loop) unconditionally pre-created at init time. The default
1543 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
1544 of statically allocating a predefined number, loop
1545 devices can be requested on-demand with the
1546 /dev/loop-control interface.
1548 mce [X86-32] Machine Check Exception
1550 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
1552 md= [HW] RAID subsystems devices and level
1553 See Documentation/md.txt.
1556 Format: <first>,<last>
1557 Specifies range of consoles to be captured by the MDA.
1559 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
1560 Amount of memory to be used when the kernel is not able
1561 to see the whole system memory or for test.
1562 [X86] Work as limiting max address. Use together
1563 with memmap= to avoid physical address space collisions.
1564 Without memmap= PCI devices could be placed at addresses
1565 belonging to unused RAM.
1567 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
1571 [KNL,SH] Allow user to override the default size for
1572 per-device physically contiguous DMA buffers.
1574 memmap=exactmap [KNL,X86] Enable setting of an exact
1575 E820 memory map, as specified by the user.
1576 Such memmap=exactmap lines can be constructed based on
1577 BIOS output or other requirements. See the memmap=nn@ss
1580 memmap=nn[KMG]@ss[KMG]
1581 [KNL] Force usage of a specific region of memory
1582 Region of memory to be used, from ss to ss+nn.
1584 memmap=nn[KMG]#ss[KMG]
1585 [KNL,ACPI] Mark specific memory as ACPI data.
1586 Region of memory to be used, from ss to ss+nn.
1588 memmap=nn[KMG]$ss[KMG]
1589 [KNL,ACPI] Mark specific memory as reserved.
1590 Region of memory to be used, from ss to ss+nn.
1591 Example: Exclude memory from 0x18690000-0x1869ffff
1592 memmap=64K$0x18690000
1594 memmap=0x10000$0x18690000
1596 memory_corruption_check=0/1 [X86]
1597 Some BIOSes seem to corrupt the first 64k of
1598 memory when doing things like suspend/resume.
1599 Setting this option will scan the memory
1600 looking for corruption. Enabling this will
1601 both detect corruption and prevent the kernel
1602 from using the memory being corrupted.
1603 However, its intended as a diagnostic tool; if
1604 repeatable BIOS-originated corruption always
1605 affects the same memory, you can use memmap=
1606 to prevent the kernel from using that memory.
1608 memory_corruption_check_size=size [X86]
1609 By default it checks for corruption in the low
1610 64k, making this memory unavailable for normal
1611 use. Use this parameter to scan for
1612 corruption in more or less memory.
1614 memory_corruption_check_period=seconds [X86]
1615 By default it checks for corruption every 60
1616 seconds. Use this parameter to check at some
1617 other rate. 0 disables periodic checking.
1619 memtest= [KNL,X86] Enable memtest
1621 default : 0 <disable>
1622 Specifies the number of memtest passes to be
1623 performed. Each pass selects another test
1624 pattern from a given set of patterns. Memtest
1625 fills the memory with this pattern, validates
1626 memory contents and reserves bad memory
1627 regions that are detected.
1629 meye.*= [HW] Set MotionEye Camera parameters
1630 See Documentation/video4linux/meye.txt.
1632 mfgpt_irq= [IA-32] Specify the IRQ to use for the
1633 Multi-Function General Purpose Timers on AMD Geode
1636 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
1637 the BIOS has incorrectly applied a workaround. TinyBIOS
1638 version 0.98 is known to be affected, 0.99 fixes the
1639 problem by letting the user disable the workaround.
1643 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
1644 physical address is ignored.
1646 mini2440= [ARM,HW,KNL]
1647 Format:[0..2][b][c][t]
1649 MINI2440 configuration specification:
1650 0 - The attached screen is the 3.5" TFT
1651 1 - The attached screen is the 7" TFT
1652 2 - The VGA Shield is attached (1024x768)
1653 Leaving out the screen size parameter will not load
1654 the TFT driver, and the framebuffer will be left
1656 b - Enable backlight. The TFT backlight pin will be
1657 linked to the kernel VESA blanking code and a GPIO
1658 LED. This parameter is not necessary when using the
1660 c - Enable the s3c camera interface.
1661 t - Reserved for enabling touchscreen support. The
1662 touchscreen support is not enabled in the mainstream
1663 kernel as of 2.6.30, a preliminary port can be found
1664 in the "bleeding edge" mini2440 support kernel at
1665 http://repo.or.cz/w/linux-2.6/mini2440.git
1668 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
1669 parameter allows control of the logging verbosity for
1670 the additional memory initialisation checks. A value
1671 of 0 disables mminit logging and a level of 4 will
1672 log everything. Information is printed at KERN_DEBUG
1673 so loglevel=8 may also need to be specified.
1676 [KNL] When CONFIG_MODULE_SIG is set, this means that
1677 modules without (valid) signatures will fail to load.
1678 Note that if CONFIG_MODULE_SIG_FORCE is set, that
1679 is always true, so this option does nothing.
1682 [MOUSE] Maximum time between finger touching and
1683 leaving touchpad surface for touch to be considered
1684 a tap and be reported as a left button click (for
1685 touchpads working in absolute mode only).
1687 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
1688 reporting absolute coordinates, such as tablets
1689 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
1690 reporting absolute coordinates, such as tablets
1692 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
1693 is similar to kernelcore except it specifies the
1694 amount of memory used for migratable allocations.
1695 If both kernelcore and movablecore is specified,
1696 then kernelcore will be at *least* the specified
1697 value but may be more. If movablecore on its own
1698 is specified, the administrator must be careful
1699 that the amount of memory usable for all allocations
1702 MTD_Partition= [MTD]
1703 Format: <name>,<region-number>,<size>,<offset>
1705 MTD_Region= [MTD] Format:
1706 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
1709 See drivers/mtd/cmdlinepart.c.
1711 multitce=off [PPC] This parameter disables the use of the pSeries
1712 firmware feature for updating multiple TCE entries
1715 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
1717 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
1719 boundary - index of last SLC block on Flex-OneNAND.
1720 The remaining blocks are configured as MLC blocks.
1721 lock - Configure if Flex-OneNAND boundary should be locked.
1722 Once locked, the boundary cannot be changed.
1723 1 indicates lock status, 0 indicates unlock status.
1726 ARM/S3C2412 JIVE boot control
1728 See arch/arm/mach-s3c2412/mach-jive.c
1730 mtouchusb.raw_coordinates=
1731 [HW] Make the MicroTouch USB driver use raw coordinates
1732 ('y', default) or cooked coordinates ('n')
1734 mtrr_chunk_size=nn[KMG] [X86]
1735 used for mtrr cleanup. It is largest continuous chunk
1736 that could hold holes aka. UC entries.
1738 mtrr_gran_size=nn[KMG] [X86]
1739 Used for mtrr cleanup. It is granularity of mtrr block.
1741 Large value could prevent small alignment from
1744 mtrr_spare_reg_nr=n [X86]
1746 Range: 0,7 : spare reg number
1748 Used for mtrr cleanup. It is spare mtrr entries number.
1749 Set to 2 or more if your graphical card needs more.
1751 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
1753 netdev= [NET] Network devices parameters
1754 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
1755 Note that mem_start is often overloaded to mean
1756 something different and driver-specific.
1757 This usage is only documented in each driver source
1761 [NETFILTER] Enable connection tracking flow accounting
1762 0 to disable accounting
1763 1 to enable accounting
1766 nfsaddrs= [NFS] Deprecated. Use ip= instead.
1767 See Documentation/filesystems/nfs/nfsroot.txt.
1769 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
1770 See Documentation/filesystems/nfs/nfsroot.txt.
1772 nfsrootdebug [NFS] enable nfsroot debugging messages.
1773 See Documentation/filesystems/nfs/nfsroot.txt.
1775 nfs.callback_tcpport=
1776 [NFS] set the TCP port on which the NFSv4 callback
1777 channel should listen.
1780 [NFS] sets the pathname to the program which is used
1781 to update the NFS client cache entries.
1783 nfs.cache_getent_timeout=
1784 [NFS] sets the timeout after which an attempt to
1785 update a cache entry is deemed to have failed.
1787 nfs.idmap_cache_timeout=
1788 [NFS] set the maximum lifetime for idmapper cache
1792 [NFS] enable 64-bit inode numbers.
1793 If zero, the NFS client will fake up a 32-bit inode
1794 number for the readdir() and stat() syscalls instead
1795 of returning the full 64-bit number.
1796 The default is to return 64-bit inode numbers.
1798 nfs.max_session_slots=
1799 [NFSv4.1] Sets the maximum number of session slots
1800 the client will attempt to negotiate with the server.
1801 This limits the number of simultaneous RPC requests
1802 that the client can send to the NFSv4.1 server.
1803 Note that there is little point in setting this
1804 value higher than the max_tcp_slot_table_limit.
1806 nfs.nfs4_disable_idmapping=
1807 [NFSv4] When set to the default of '1', this option
1808 ensures that both the RPC level authentication
1809 scheme and the NFS level operations agree to use
1810 numeric uids/gids if the mount is using the
1811 'sec=sys' security flavour. In effect it is
1812 disabling idmapping, which can make migration from
1813 legacy NFSv2/v3 systems to NFSv4 easier.
1814 Servers that do not support this mode of operation
1815 will be autodetected by the client, and it will fall
1816 back to using the idmapper.
1817 To turn off this behaviour, set the value to '0'.
1819 [NFS4] Specify an additional fixed unique ident-
1820 ification string that NFSv4 clients can insert into
1821 their nfs_client_id4 string. This is typically a
1822 UUID that is generated at system install time.
1824 nfs.send_implementation_id =
1825 [NFSv4.1] Send client implementation identification
1826 information in exchange_id requests.
1827 If zero, no implementation identification information
1829 The default is to send the implementation identification
1832 nfsd.nfs4_disable_idmapping=
1833 [NFSv4] When set to the default of '1', the NFSv4
1834 server will return only numeric uids and gids to
1835 clients using auth_sys, and will accept numeric uids
1836 and gids from such clients. This is intended to ease
1837 migration from NFSv2/v3.
1839 objlayoutdriver.osd_login_prog=
1840 [NFS] [OBJLAYOUT] sets the pathname to the program which
1841 is used to automatically discover and login into new
1842 osd-targets. Please see:
1843 Documentation/filesystems/pnfs.txt for more explanations
1845 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
1846 when a NMI is triggered.
1847 Format: [state][,regs][,debounce][,die]
1849 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
1850 Format: [panic,][nopanic,][num]
1852 0 - turn nmi_watchdog off
1853 When panic is specified, panic when an NMI watchdog
1854 timeout occurs (or 'nopanic' to override the opposite
1856 This is useful when you use a panic=... timeout and
1857 need the box quickly up again.
1859 netpoll.carrier_timeout=
1860 [NET] Specifies amount of time (in seconds) that
1861 netpoll should wait for a carrier. By default netpoll
1864 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
1865 emulation library even if a 387 maths coprocessor
1869 [HW] Never suspend the console
1870 Disable suspending of consoles during suspend and
1871 hibernate operations. Once disabled, debugging
1872 messages can reach various consoles while the rest
1873 of the system is being put to sleep (ie, while
1874 debugging driver suspend/resume hooks). This may
1875 not work reliably with all consoles, but is known
1876 to work with serial and VGA consoles.
1877 To facilitate more flexible debugging, we also add
1878 console_suspend, a printk module parameter to control
1879 it. Users could use console_suspend (usually
1880 /sys/module/printk/parameters/console_suspend) to
1881 turn on/off it dynamically.
1883 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
1884 caches in the slab allocator. Saves per-node memory,
1885 but will impact performance.
1889 noapic [SMP,APIC] Tells the kernel to not make use of any
1890 IOAPICs that may be present in the system.
1892 noautogroup Disable scheduler automatic task group creation.
1894 nobats [PPC] Do not use BATs for mapping kernel lowmem
1895 on "Classic" PPC cores.
1899 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
1901 nodelayacct [KNL] Disable per-task delay accounting
1903 nodisconnect [HW,SCSI,M68K] Disables SCSI disconnects.
1905 nodsp [SH] Disable hardware DSP at boot time.
1907 noefi [X86] Disable EFI runtime services support.
1912 On X86-32 available only on PAE configured kernels.
1913 noexec=on: enable non-executable mappings (default)
1914 noexec=off: disable non-executable mappings
1917 Disable SMAP (Supervisor Mode Access Prevention)
1918 even if it is supported by processor.
1921 Disable SMEP (Supervisor Mode Execution Prevention)
1922 even if it is supported by processor.
1925 This affects only 32-bit executables.
1926 noexec32=on: enable non-executable mappings (default)
1927 read doesn't imply executable mappings
1928 noexec32=off: disable non-executable mappings
1929 read implies executable mappings
1931 nofpu [SH] Disable hardware FPU at boot time.
1933 nofxsr [BUGS=X86-32] Disables x86 floating point extended
1934 register save and restore. The kernel will only save
1935 legacy floating-point registers on task switch.
1937 noxsave [BUGS=X86] Disables x86 extended register state save
1938 and restore using xsave. The kernel will fallback to
1939 enabling legacy floating-point and sse state.
1942 on enable eager fpu restore
1943 off disable eager fpu restore
1944 auto selects the default scheme, which automatically
1945 enables eagerfpu restore for xsaveopt.
1947 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
1948 wfi(ARM) instruction doesn't work correctly and not to
1949 use it. This is also useful when using JTAG debugger.
1951 no_file_caps Tells the kernel not to honor file capabilities. The
1952 only way then for a file to be executed with privilege
1953 is to be setuid root or executed by root.
1955 nohalt [IA-64] Tells the kernel not to use the power saving
1956 function PAL_HALT_LIGHT when idle. This increases
1957 power-consumption. On the positive side, it reduces
1958 interrupt wake-up latency, which may improve performance
1959 in certain environments such as networked servers or
1962 nohz= [KNL] Boottime enable/disable dynamic ticks
1963 Valid arguments: on, off
1966 noiotrap [SH] Disables trapped I/O port accesses.
1968 noirqdebug [X86-32] Disables the code which attempts to detect and
1969 disable unhandled interrupt sources.
1971 no_timer_check [X86,APIC] Disables the code which tests for
1972 broken timer IRQ sources.
1974 noisapnp [ISAPNP] Disables ISA PnP code.
1976 noinitrd [RAM] Tells the kernel not to load any configured
1979 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
1981 [Deprecated - use intremap=off]
1985 nojitter [IA-64] Disables jitter checking for ITC timers.
1987 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
1989 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
1992 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
1993 steal time is computed, but won't influence scheduler
1996 nolapic [X86-32,APIC] Do not enable or use the local APIC.
1998 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2000 noltlbs [PPC] Do not use large page/tlb entries for kernel
2001 lowmem mapping on PPC40x.
2003 nomca [IA-64] Disable machine check abort handling
2005 nomce [X86-32] Machine Check Exception
2007 nomfgpt [X86-32] Disable Multi-Function General Purpose
2008 Timer usage (for AMD Geode machines).
2010 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2011 shutdown the other cpus. Instead use the REBOOT_VECTOR
2014 nomodule Disable module load
2016 nopat [X86] Disable PAT (page attribute table extension of
2017 pagetables) support.
2019 norandmaps Don't use address space randomization. Equivalent to
2020 echo 0 > /proc/sys/kernel/randomize_va_space
2022 noreplace-paravirt [X86,IA-64,PV_OPS] Don't patch paravirt_ops
2024 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2025 with UP alternatives
2027 nordrand [X86] Disable the direct use of the RDRAND
2028 instruction even if it is supported by the
2029 processor. RDRAND is still available to user
2032 noresume [SWSUSP] Disables resume and restores original swap
2035 no-scroll [VGA] Disables scrollback.
2036 This is required for the Braillex ib80-piezo Braille
2037 reader made by F.H. Papenmeier (Germany).
2041 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2043 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2044 and disable the IO APIC. legacy for "maxcpus=0".
2046 nosoftlockup [KNL] Disable the soft-lockup detector.
2048 nosync [HW,M68K] Disables sync negotiation for all devices.
2050 notsc [BUGS=X86-32] Disable Time Stamp Counter
2052 nousb [USB] Disable the USB subsystem
2054 nowatchdog [KNL] Disable the lockup detector (NMI watchdog).
2058 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2060 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2061 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2062 Some features depend on CPU0. Known dependencies are:
2063 1. Resume from suspend/hibernate depends on CPU0.
2064 Suspend/hibernate will fail if CPU0 is offline and you
2065 need to online CPU0 before suspend/hibernate.
2066 2. PIC interrupts also depend on CPU0. CPU0 can't be
2067 removed if a PIC interrupt is detected.
2068 It's said poweroff/reboot may depend on CPU0 on some
2069 machines although I haven't seen such issues so far
2070 after CPU0 is offline on a few tested machines.
2071 If the dependencies are under your control, you can
2072 turn on cpu0_hotplug.
2074 nptcg= [IA-64] Override max number of concurrent global TLB
2075 purges which is reported from either PAL_VM_SUMMARY or
2078 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2079 could support. nr_cpus=n : n >= 1 limits the kernel to
2080 supporting 'n' processors. Later in runtime you can not
2081 use hotplug cpu feature to put more cpu back to online.
2082 just like you compile the kernel NR_CPUS=n
2084 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2086 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2087 Allowed values are enable and disable
2089 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2090 one of ['zone', 'node', 'default'] can be specified
2091 This can be set from sysctl after boot.
2092 See Documentation/sysctl/vm.txt for details.
2094 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2095 See Documentation/debugging-via-ohci1394.txt for more
2098 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2099 Rather than timing out after 20 ms if an EC
2100 command is not properly ACKed, override the length
2101 of the timeout. We have interrupts disabled while
2102 waiting for the ACK, so if this is set too high
2103 interrupts *may* be lost!
2105 omap_mux= [OMAP] Override bootloader pin multiplexing.
2106 Format: <mux_mode0.mode_name=value>...
2107 For example, to override I2C bus2:
2108 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2110 oprofile.timer= [HW]
2111 Use timer interrupt instead of performance counters
2113 oprofile.cpu_type= Force an oprofile cpu type
2114 This might be useful if you have an older oprofile
2115 userland or if you want common events.
2116 Format: { arch_perfmon }
2117 arch_perfmon: [X86] Force use of architectural
2118 perfmon on Intel CPUs instead of the
2119 CPU specific event set.
2120 timer: [X86] Force use of architectural NMI
2121 timer mode (see also oprofile.timer
2122 for generic hr timer mode)
2123 [s390] Force legacy basic mode sampling
2124 (report cpu_type "timer")
2126 oops=panic Always panic on oopses. Default is to just kill the
2127 process, but there is a small probability of
2128 deadlocking the machine.
2129 This will also cause panics on machine check exceptions.
2130 Useful together with panic=30 to trigger a reboot.
2133 See Documentation/sound/oss/oss-parameters.txt
2135 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2136 timeout > 0: seconds before rebooting
2137 timeout = 0: wait forever
2138 timeout < 0: reboot immediately
2141 parkbd.port= [HW] Parallel port number the keyboard adapter is
2142 connected to, default is 0.
2144 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2145 0 for XT, 1 for AT (default is AT).
2148 parport= [HW,PPT] Specify parallel ports. 0 disables.
2149 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2150 Use 'auto' to force the driver to use any
2151 IRQ/DMA settings detected (the default is to
2152 ignore detected IRQ/DMA settings because of
2153 possible conflicts). You can specify the base
2154 address, IRQ, and DMA settings; IRQ and DMA
2155 should be numbers, or 'auto' (for using detected
2156 settings on that particular port), or 'nofifo'
2157 (to avoid using a FIFO even if it is detected).
2158 Parallel ports are assigned in the order they
2159 are specified on the command line, starting
2162 parport_init_mode= [HW,PPT]
2163 Configure VIA parallel port to operate in
2164 a specific mode. This is necessary on Pegasos
2165 computer where firmware has no options for setting
2166 up parallel port mode and sets it to spp.
2167 Currently this function knows 686a and 8231 chips.
2168 Format: [spp|ps2|epp|ecp|ecpepp]
2171 Halt all CPUs after the first oops has been printed for
2172 the specified number of seconds. This is to be used if
2173 your oopses keep scrolling off the screen.
2178 See header of drivers/block/paride/pcd.c.
2179 See also Documentation/blockdev/paride.txt.
2181 pci=option[,option...] [PCI] various PCI subsystem options:
2182 earlydump [X86] dump PCI config space before the kernel
2184 off [X86] don't probe for the PCI bus
2185 bios [X86-32] force use of PCI BIOS, don't access
2186 the hardware directly. Use this if your machine
2187 has a non-standard PCI host bridge.
2188 nobios [X86-32] disallow use of PCI BIOS, only direct
2189 hardware access methods are allowed. Use this
2190 if you experience crashes upon bootup and you
2191 suspect they are caused by the BIOS.
2192 conf1 [X86] Force use of PCI Configuration
2194 conf2 [X86] Force use of PCI Configuration
2196 noaer [PCIE] If the PCIEAER kernel config parameter is
2197 enabled, this kernel boot option can be used to
2198 disable the use of PCIE advanced error reporting.
2199 nodomains [PCI] Disable support for multiple PCI
2200 root domains (aka PCI segments, in ACPI-speak).
2201 nommconf [X86] Disable use of MMCONFIG for PCI
2203 check_enable_amd_mmconf [X86] check for and enable
2204 properly configured MMIO access to PCI
2205 config space on AMD family 10h CPU
2206 nomsi [MSI] If the PCI_MSI kernel config parameter is
2207 enabled, this kernel boot option can be used to
2208 disable the use of MSI interrupts system-wide.
2209 noioapicquirk [APIC] Disable all boot interrupt quirks.
2210 Safety option to keep boot IRQs enabled. This
2211 should never be necessary.
2212 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
2213 primary IO-APIC for bridges that cannot disable
2214 boot IRQs. This fixes a source of spurious IRQs
2215 when the system masks IRQs.
2216 noioapicreroute [APIC] Disable workaround that uses the
2217 boot IRQ equivalent of an IRQ that connects to
2218 a chipset where boot IRQs cannot be disabled.
2219 The opposite of ioapicreroute.
2220 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
2221 routing table. These calls are known to be buggy
2222 on several machines and they hang the machine
2223 when used, but on other computers it's the only
2224 way to get the interrupt routing table. Try
2225 this option if the kernel is unable to allocate
2226 IRQs or discover secondary PCI buses on your
2228 rom [X86] Assign address space to expansion ROMs.
2229 Use with caution as certain devices share
2230 address decoders between ROMs and other
2232 norom [X86] Do not assign address space to
2233 expansion ROMs that do not already have
2234 BIOS assigned address ranges.
2235 nobar [X86] Do not assign address space to the
2236 BARs that weren't assigned by the BIOS.
2237 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
2238 assigned automatically to PCI devices. You can
2239 make the kernel exclude IRQs of your ISA cards
2241 pirqaddr=0xAAAAA [X86] Specify the physical address
2242 of the PIRQ table (normally generated
2243 by the BIOS) if it is outside the
2244 F0000h-100000h range.
2245 lastbus=N [X86] Scan all buses thru bus #N. Can be
2246 useful if the kernel is unable to find your
2247 secondary buses and you want to tell it
2248 explicitly which ones they are.
2249 assign-busses [X86] Always assign all PCI bus
2250 numbers ourselves, overriding
2251 whatever the firmware may have done.
2252 usepirqmask [X86] Honor the possible IRQ mask stored
2253 in the BIOS $PIR table. This is needed on
2254 some systems with broken BIOSes, notably
2255 some HP Pavilion N5400 and Omnibook XE3
2256 notebooks. This will have no effect if ACPI
2257 IRQ routing is enabled.
2258 noacpi [X86] Do not use ACPI for IRQ routing
2259 or for PCI scanning.
2260 use_crs [X86] Use PCI host bridge window information
2261 from ACPI. On BIOSes from 2008 or later, this
2262 is enabled by default. If you need to use this,
2263 please report a bug.
2264 nocrs [X86] Ignore PCI host bridge windows from ACPI.
2265 If you need to use this, please report a bug.
2266 routeirq Do IRQ routing for all PCI devices.
2267 This is normally done in pci_enable_device(),
2268 so this option is a temporary workaround
2269 for broken drivers that don't call it.
2270 skip_isa_align [X86] do not align io start addr, so can
2271 handle more pci cards
2272 firmware [ARM] Do not re-enumerate the bus but instead
2273 just use the configuration from the
2274 bootloader. This is currently used on
2275 IXP2000 systems where the bus has to be
2276 configured a certain way for adjunct CPUs.
2277 noearly [X86] Don't do any early type 1 scanning.
2278 This might help on some broken boards which
2279 machine check when some devices' config space
2280 is read. But various workarounds are disabled
2281 and some IOMMU drivers will not work.
2282 bfsort Sort PCI devices into breadth-first order.
2283 This sorting is done to get a device
2284 order compatible with older (<= 2.4) kernels.
2285 nobfsort Don't sort PCI devices into breadth-first order.
2286 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
2287 tuning and use the BIOS-configured MPS defaults.
2288 pcie_bus_safe Set every device's MPS to the largest value
2289 supported by all devices below the root complex.
2290 pcie_bus_perf Set device MPS to the largest allowable MPS
2291 based on its parent bus. Also set MRRS (Max
2292 Read Request Size) to the largest supported
2293 value (no larger than the MPS that the device
2294 or bus can support) for best performance.
2295 pcie_bus_peer2peer Set every device's MPS to 128B, which
2296 every device is guaranteed to support. This
2297 configuration allows peer-to-peer DMA between
2298 any pair of devices, possibly at the cost of
2299 reduced performance. This also guarantees
2300 that hot-added devices will work.
2301 cbiosize=nn[KMG] The fixed amount of bus space which is
2302 reserved for the CardBus bridge's IO window.
2303 The default value is 256 bytes.
2304 cbmemsize=nn[KMG] The fixed amount of bus space which is
2305 reserved for the CardBus bridge's memory
2306 window. The default value is 64 megabytes.
2309 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
2310 Specifies alignment and device to reassign
2311 aligned memory resources.
2312 If <order of align> is not specified,
2313 PAGE_SIZE is used as alignment.
2314 PCI-PCI bridge can be specified, if resource
2315 windows need to be expanded.
2316 ecrc= Enable/disable PCIe ECRC (transaction layer
2317 end-to-end CRC checking).
2318 bios: Use BIOS/firmware settings. This is the
2322 hpiosize=nn[KMG] The fixed amount of bus space which is
2323 reserved for hotplug bridge's IO window.
2324 Default size is 256 bytes.
2325 hpmemsize=nn[KMG] The fixed amount of bus space which is
2326 reserved for hotplug bridge's memory window.
2327 Default size is 2 megabytes.
2328 realloc= Enable/disable reallocating PCI bridge resources
2329 if allocations done by BIOS are too small to
2330 accommodate resources required by all child
2332 off: Turn realloc off
2334 realloc same as realloc=on
2335 noari do not use PCIe ARI.
2336 pcie_scan_all Scan all possible PCIe devices. Otherwise we
2337 only look for one device below a PCIe downstream
2340 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
2343 force Enable ASPM even on devices that claim not to support it.
2344 WARNING: Forcing ASPM on may cause system lockups.
2346 pcie_hp= [PCIE] PCI Express Hotplug driver options:
2347 nomsi Do not use MSI for PCI Express Native Hotplug (this
2348 makes all PCIe ports use INTx for hotplug services).
2350 pcie_ports= [PCIE] PCIe ports handling:
2351 auto Ask the BIOS whether or not to use native PCIe services
2352 associated with PCIe ports (PME, hot-plug, AER). Use
2353 them only if that is allowed by the BIOS.
2354 native Use native PCIe services associated with PCIe ports
2356 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
2359 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
2360 nomsi Do not use MSI for native PCIe PME signaling (this makes
2361 all PCIe root ports use INTx for all services).
2363 pcmv= [HW,PCMCIA] BadgePAD 4
2366 See Documentation/blockdev/paride.txt.
2368 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
2371 See arch/parisc/kernel/pdc_chassis.c
2373 percpu_alloc= Select which percpu first chunk allocator to use.
2374 Currently supported values are "embed" and "page".
2375 Archs may support subset or none of the selections.
2376 See comments in mm/percpu.c for details on each
2377 allocator. This parameter is primarily for debugging
2378 and performance comparison.
2381 See Documentation/blockdev/paride.txt.
2384 See Documentation/blockdev/paride.txt.
2386 pirq= [SMP,APIC] Manual mp-table setup
2387 See Documentation/x86/i386/IO-APIC.txt.
2389 plip= [PPT,NET] Parallel port network link
2390 Format: { parport<nr> | timid | 0 }
2391 See also Documentation/parport.txt.
2393 pmtmr= [X86] Manual setup of pmtmr I/O Port.
2394 Override pmtimer IOPort with a hex value.
2398 Enable PNP debug messages (depends on the
2399 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
2400 via /sys/module/pnp/parameters/debug. We always show
2401 current resource usage; turning this on also shows
2402 possible settings and some assignment information.
2408 { on | off | curr | res | no-curr | no-res }
2411 [ISAPNP] Exclude IRQs for the autoconfiguration
2414 [ISAPNP] Exclude DMAs for the autoconfiguration
2416 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
2417 Ranges are in pairs (I/O port base and size).
2420 [ISAPNP] Exclude memory regions for the
2422 Ranges are in pairs (memory base and size).
2424 ports= [IP_VS_FTP] IPVS ftp helper module
2426 Up to 8 (IP_VS_APP_MAX_PORTS) ports
2428 Format: <port>,<port>....
2430 print-fatal-signals=
2431 [KNL] debug: print fatal signals
2433 If enabled, warn about various signal handling
2434 related application anomalies: too many signals,
2435 too many POSIX.1 timers, fatal signals causing a
2438 If you hit the warning due to signal overflow,
2439 you might want to try "ulimit -i unlimited".
2443 printk.always_kmsg_dump=
2444 Trigger kmsg_dump for cases other than kernel oops or
2446 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2449 printk.time= Show timing data prefixed to each printk message line
2450 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
2452 processor.max_cstate= [HW,ACPI]
2453 Limit processor to maximum C-state
2454 max_cstate=9 overrides any DMI blacklist limit.
2456 processor.nocst [HW,ACPI]
2457 Ignore the _CST method to determine C-states,
2458 instead using the legacy FADT method
2460 profile= [KNL] Enable kernel profiling via /proc/profile
2461 Format: [schedule,]<number>
2462 Param: "schedule" - profile schedule points.
2463 Param: <number> - step/bucket size as a power of 2 for
2464 statistical time based profiling.
2465 Param: "sleep" - profile D-state sleeping (millisecs).
2466 Requires CONFIG_SCHEDSTATS
2467 Param: "kvm" - profile VM exits.
2469 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
2471 See Documentation/blockdev/ramdisk.txt.
2473 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
2474 probe for; one of (bare|imps|exps|lifebook|any).
2475 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
2477 psmouse.resetafter= [HW,MOUSE]
2478 Try to reset the device after so many bad packets
2481 [HW,MOUSE] Set desired mouse resolution, in dpi.
2482 psmouse.smartscroll=
2483 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
2484 0 = disabled, 1 = enabled (default).
2486 pstore.backend= Specify the name of the pstore backend to use
2489 See Documentation/blockdev/paride.txt.
2492 [KNL] Number of legacy pty's. Overwrites compiled-in
2495 quiet [KNL] Disable most log messages
2500 See Documentation/md.txt.
2502 ramdisk_blocksize= [RAM]
2503 See Documentation/blockdev/ramdisk.txt.
2505 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
2506 See Documentation/blockdev/ramdisk.txt.
2508 rcu_nocbs= [KNL,BOOT]
2509 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
2510 the specified list of CPUs to be no-callback CPUs.
2511 Invocation of these CPUs' RCU callbacks will
2512 be offloaded to "rcuox/N" kthreads created for
2513 that purpose, where "x" is "b" for RCU-bh, "p"
2514 for RCU-preempt, and "s" for RCU-sched, and "N"
2515 is the CPU number. This reduces OS jitter on the
2516 offloaded CPUs, which can be useful for HPC and
2518 real-time workloads. It can also improve energy
2519 efficiency for asymmetric multiprocessors.
2521 rcu_nocb_poll [KNL,BOOT]
2522 Rather than requiring that offloaded CPUs
2523 (specified by rcu_nocbs= above) explicitly
2524 awaken the corresponding "rcuoN" kthreads,
2525 make these kthreads poll for callbacks.
2526 This improves the real-time response for the
2527 offloaded CPUs by relieving them of the need to
2528 wake up the corresponding kthread, but degrades
2529 energy efficiency by requiring that the kthreads
2530 periodically wake up to do the polling.
2532 rcutree.blimit= [KNL,BOOT]
2533 Set maximum number of finished RCU callbacks to process
2536 rcutree.fanout_leaf= [KNL,BOOT]
2537 Increase the number of CPUs assigned to each
2538 leaf rcu_node structure. Useful for very large
2541 rcutree.jiffies_till_first_fqs= [KNL,BOOT]
2542 Set delay from grace-period initialization to
2543 first attempt to force quiescent states.
2544 Units are jiffies, minimum value is zero,
2545 and maximum value is HZ.
2547 rcutree.jiffies_till_next_fqs= [KNL,BOOT]
2548 Set delay between subsequent attempts to force
2549 quiescent states. Units are jiffies, minimum
2550 value is one, and maximum value is HZ.
2552 rcutree.qhimark= [KNL,BOOT]
2553 Set threshold of queued
2554 RCU callbacks over which batch limiting is disabled.
2556 rcutree.qlowmark= [KNL,BOOT]
2557 Set threshold of queued RCU callbacks below which
2558 batch limiting is re-enabled.
2560 rcutree.rcu_cpu_stall_suppress= [KNL,BOOT]
2561 Suppress RCU CPU stall warning messages.
2563 rcutree.rcu_cpu_stall_timeout= [KNL,BOOT]
2564 Set timeout for RCU CPU stall warning messages.
2566 rcutree.rcu_idle_gp_delay= [KNL,BOOT]
2567 Set wakeup interval for idle CPUs that have
2568 RCU callbacks (RCU_FAST_NO_HZ=y).
2570 rcutree.rcu_idle_lazy_gp_delay= [KNL,BOOT]
2571 Set wakeup interval for idle CPUs that have
2572 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
2573 Lazy RCU callbacks are those which RCU can
2574 prove do nothing more than free memory.
2576 rcutorture.fqs_duration= [KNL,BOOT]
2577 Set duration of force_quiescent_state bursts.
2579 rcutorture.fqs_holdoff= [KNL,BOOT]
2580 Set holdoff time within force_quiescent_state bursts.
2582 rcutorture.fqs_stutter= [KNL,BOOT]
2583 Set wait time between force_quiescent_state bursts.
2585 rcutorture.irqreader= [KNL,BOOT]
2586 Test RCU readers from irq handlers.
2588 rcutorture.n_barrier_cbs= [KNL,BOOT]
2589 Set callbacks/threads for rcu_barrier() testing.
2591 rcutorture.nfakewriters= [KNL,BOOT]
2592 Set number of concurrent RCU writers. These just
2593 stress RCU, they don't participate in the actual
2594 test, hence the "fake".
2596 rcutorture.nreaders= [KNL,BOOT]
2597 Set number of RCU readers.
2599 rcutorture.onoff_holdoff= [KNL,BOOT]
2600 Set time (s) after boot for CPU-hotplug testing.
2602 rcutorture.onoff_interval= [KNL,BOOT]
2603 Set time (s) between CPU-hotplug operations, or
2604 zero to disable CPU-hotplug testing.
2606 rcutorture.shuffle_interval= [KNL,BOOT]
2607 Set task-shuffle interval (s). Shuffling tasks
2608 allows some CPUs to go into dyntick-idle mode
2609 during the rcutorture test.
2611 rcutorture.shutdown_secs= [KNL,BOOT]
2612 Set time (s) after boot system shutdown. This
2613 is useful for hands-off automated testing.
2615 rcutorture.stall_cpu= [KNL,BOOT]
2616 Duration of CPU stall (s) to test RCU CPU stall
2617 warnings, zero to disable.
2619 rcutorture.stall_cpu_holdoff= [KNL,BOOT]
2620 Time to wait (s) after boot before inducing stall.
2622 rcutorture.stat_interval= [KNL,BOOT]
2623 Time (s) between statistics printk()s.
2625 rcutorture.stutter= [KNL,BOOT]
2626 Time (s) to stutter testing, for example, specifying
2627 five seconds causes the test to run for five seconds,
2628 wait for five seconds, and so on. This tests RCU's
2629 ability to transition abruptly to and from idle.
2631 rcutorture.test_boost= [KNL,BOOT]
2632 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
2633 "Maybe" means test if the RCU implementation
2634 under test support RCU priority boosting.
2636 rcutorture.test_boost_duration= [KNL,BOOT]
2637 Duration (s) of each individual boost test.
2639 rcutorture.test_boost_interval= [KNL,BOOT]
2640 Interval (s) between each boost test.
2642 rcutorture.test_no_idle_hz= [KNL,BOOT]
2643 Test RCU's dyntick-idle handling. See also the
2644 rcutorture.shuffle_interval parameter.
2646 rcutorture.torture_type= [KNL,BOOT]
2647 Specify the RCU implementation to test.
2649 rcutorture.verbose= [KNL,BOOT]
2650 Enable additional printk() statements.
2654 Run specified binary instead of /init from the ramdisk,
2655 used for early userspace startup. See initrd.
2657 reboot= [BUGS=X86-32,BUGS=ARM,BUGS=IA-64] Rebooting mode
2658 Format: <reboot_mode>[,<reboot_mode2>[,...]]
2659 See arch/*/kernel/reboot.c or arch/*/kernel/process.c
2662 [KNL, SMP] Set scheduler's default relax_domain_level.
2663 See Documentation/cgroups/cpusets.txt.
2665 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
2667 reservetop= [X86-32]
2669 Reserves a hole at the top of the kernel virtual
2674 Set the amount of memory to reserve for BIOS at
2675 the bottom of the address space.
2677 reset_devices [KNL] Force drivers to reset the underlying device
2678 during initialization.
2681 Specify the partition device for software suspend
2683 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
2685 resume_offset= [SWSUSP]
2686 Specify the offset from the beginning of the partition
2687 given by "resume=" at which the swap header is located,
2688 in <PAGE_SIZE> units (needed only for swap files).
2689 See Documentation/power/swsusp-and-swap-files.txt
2691 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
2692 read the resume files
2694 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
2695 Useful for devices that are detected asynchronously
2696 (e.g. USB and MMC devices).
2698 hibernate= [HIBERNATION]
2699 noresume Don't check if there's a hibernation image
2700 present during boot.
2701 nocompress Don't compress/decompress hibernation images.
2703 retain_initrd [RAM] Keep initrd memory after extraction
2705 rhash_entries= [KNL,NET]
2706 Set number of hash buckets for route cache
2708 riscom8= [HW,SERIAL]
2709 Format: <io_board1>[,<io_board2>[,...<io_boardN>]]
2711 ro [KNL] Mount root device read-only on boot
2713 root= [KNL] Root filesystem
2714 See name_to_dev_t comment in init/do_mounts.c.
2716 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
2717 mount the root filesystem
2719 rootflags= [KNL] Set root filesystem mount option string
2721 rootfstype= [KNL] Set root filesystem type
2723 rootwait [KNL] Wait (indefinitely) for root device to show up.
2724 Useful for devices that are detected asynchronously
2725 (e.g. USB and MMC devices).
2727 rw [KNL] Mount root device read-write on boot
2729 S [KNL] Run init in single mode
2732 See drivers/net/irda/sa1100_ir.c.
2734 sbni= [NET] Granch SBNI12 leased line adapter
2736 sched_debug [KNL] Enables verbose scheduler debug messages.
2738 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
2739 xtime_lock contention on larger systems, and/or RCU lock
2740 contention on all systems with CONFIG_MAXSMP set.
2741 Format: { "0" | "1" }
2742 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
2744 Note: increases power consumption, thus should only be
2745 enabled if running jitter sensitive (HPC/RT) workloads.
2747 security= [SECURITY] Choose a security module to enable at boot.
2748 If this boot parameter is not specified, only the first
2749 security module asking for security registration will be
2750 loaded. An invalid security module name will be treated
2751 as if no module has been chosen.
2753 selinux= [SELINUX] Disable or enable SELinux at boot time.
2754 Format: { "0" | "1" }
2755 See security/selinux/Kconfig help text.
2758 Default value is set via kernel config option.
2759 If enabled at boot time, /selinux/disable can be used
2760 later to disable prior to initial policy load.
2762 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
2763 Format: { "0" | "1" }
2764 See security/apparmor/Kconfig help text
2767 Default value is set via kernel config option.
2769 serialnumber [BUGS=X86-32]
2772 Maximal number of shapers.
2774 show_msr= [x86] show boot-time MSR settings
2775 Format: { <integer> }
2776 Show boot-time (BIOS-initialized) MSR settings.
2777 The parameter means the number of CPUs to show,
2778 for example 1 means boot CPU only.
2785 slab_max_order= [MM, SLAB]
2786 Determines the maximum allowed order for slabs.
2787 A high setting may cause OOMs due to memory
2788 fragmentation. Defaults to 1 for systems with
2789 more than 32MB of RAM, 0 otherwise.
2791 slub_debug[=options[,slabs]] [MM, SLUB]
2792 Enabling slub_debug allows one to determine the
2793 culprit if slab objects become corrupted. Enabling
2794 slub_debug can create guard zones around objects and
2795 may poison objects when not in use. Also tracks the
2796 last alloc / free. For more information see
2797 Documentation/vm/slub.txt.
2799 slub_max_order= [MM, SLUB]
2800 Determines the maximum allowed order for slabs.
2801 A high setting may cause OOMs due to memory
2802 fragmentation. For more information see
2803 Documentation/vm/slub.txt.
2805 slub_min_objects= [MM, SLUB]
2806 The minimum number of objects per slab. SLUB will
2807 increase the slab order up to slub_max_order to
2808 generate a sufficiently large slab able to contain
2809 the number of objects indicated. The higher the number
2810 of objects the smaller the overhead of tracking slabs
2811 and the less frequently locks need to be acquired.
2812 For more information see Documentation/vm/slub.txt.
2814 slub_min_order= [MM, SLUB]
2815 Determines the minimum page order for slabs. Must be
2816 lower than slub_max_order.
2817 For more information see Documentation/vm/slub.txt.
2819 slub_nomerge [MM, SLUB]
2820 Disable merging of slabs with similar size. May be
2821 necessary if there is some reason to distinguish
2822 allocs to different slabs. Debug options disable
2823 merging on their own.
2824 For more information see Documentation/vm/slub.txt.
2827 Format: <io1>[,<io2>[,...,<io8>]]
2829 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
2830 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
2831 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
2832 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
2833 smsc-ircc2.ircc_irq= [HW] IRQ line
2834 smsc-ircc2.ircc_dma= [HW] DMA channel
2835 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
2836 0: Toshiba Satellite 1800 (GP data pin select)
2837 1: Fast pin select (default)
2841 [KNL] Should the soft-lockup detector generate panics.
2844 sonypi.*= [HW] Sony Programmable I/O Control Device driver
2845 See Documentation/laptops/sonypi.txt
2847 specialix= [HW,SERIAL] Specialix multi-serial port adapter
2848 See Documentation/serial/specialix.txt.
2850 spia_io_base= [HW,MTD]
2856 Enabled the stack tracer on boot up.
2858 stacktrace_filter=[function-list]
2859 [FTRACE] Limit the functions that the stack tracer
2860 will trace at boot up. function-list is a comma separated
2861 list of functions. This list can be changed at run
2862 time by the stack_trace_filter file in the debugfs
2863 tracing directory. Note, this enables stack tracing
2864 and the stacktrace above is not needed.
2868 Set the STI (builtin display/keyboard on the HP-PARISC
2869 machines) console (graphic card) which should be used
2870 as the initial boot-console.
2871 See also comment in drivers/video/console/sticore.c.
2874 See comment in drivers/video/console/sticore.c.
2877 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
2879 sunrpc.min_resvport=
2880 sunrpc.max_resvport=
2882 SunRPC servers often require that client requests
2883 originate from a privileged port (i.e. a port in the
2884 range 0 < portnr < 1024).
2885 An administrator who wishes to reserve some of these
2886 ports for other uses may adjust the range that the
2887 kernel's sunrpc client considers to be privileged
2888 using these two parameters to set the minimum and
2889 maximum port values.
2893 Control how the NFS server code allocates CPUs to
2894 service thread pools. Depending on how many NICs
2895 you have and where their interrupts are bound, this
2896 option will affect which CPUs will do NFS serving.
2897 Note: this parameter cannot be changed while the
2898 NFS server is running.
2900 auto the server chooses an appropriate mode
2901 automatically using heuristics
2902 global a single global pool contains all CPUs
2903 percpu one pool for each CPU
2904 pernode one pool for each NUMA node (equivalent
2905 to global on non-NUMA machines)
2907 sunrpc.tcp_slot_table_entries=
2908 sunrpc.udp_slot_table_entries=
2910 Sets the upper limit on the number of simultaneous
2911 RPC calls that can be sent from the client to a
2912 server. Increasing these values may allow you to
2913 improve throughput, but will also increase the
2914 amount of memory reserved for use by the client.
2917 [KNL] Enable accounting of swap in memory resource
2918 controller if no parameter or 1 is given or disable
2919 it if 0 is given (See Documentation/cgroups/memory.txt)
2921 swiotlb= [IA-64] Number of I/O TLB slabs
2925 sysfs.deprecated=0|1 [KNL]
2926 Enable/disable old style sysfs layout for old udev
2927 on older distributions. When this option is enabled
2928 very new udev will not work anymore. When this option
2929 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
2930 in older udev will not work anymore.
2931 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
2932 the kernel configuration.
2934 sysrq_always_enabled
2936 Ignore sysrq setting - this boot parameter will
2937 neutralize any effect of /proc/sys/kernel/sysrq.
2938 Useful for debugging.
2942 test_suspend= [SUSPEND]
2943 Specify "mem" (for Suspend-to-RAM) or "standby" (for
2944 standby suspend) as the system sleep state to briefly
2945 enter during system startup. The system is woken from
2946 this state using a wakeup-capable RTC alarm.
2948 thash_entries= [KNL,NET]
2949 Set number of hash buckets for TCP connection
2951 thermal.act= [HW,ACPI]
2952 -1: disable all active trip points in all thermal zones
2953 <degrees C>: override all lowest active trip points
2955 thermal.crt= [HW,ACPI]
2956 -1: disable all critical trip points in all thermal zones
2957 <degrees C>: override all critical trip points
2959 thermal.nocrt= [HW,ACPI]
2960 Set to disable actions on ACPI thermal zone
2961 critical and hot trip points.
2963 thermal.off= [HW,ACPI]
2964 1: disable ACPI thermal control
2966 thermal.psv= [HW,ACPI]
2967 -1: disable all passive trip points
2968 <degrees C>: override all passive trip points to this
2971 thermal.tzp= [HW,ACPI]
2972 Specify global default ACPI thermal zone polling rate
2973 <deci-seconds>: poll all this frequency
2974 0: no polling (default)
2977 Force threading of all interrupt handlers except those
2978 marked explicitly IRQF_NO_THREAD.
2982 Specify if the kernel should make use of the cpu
2983 topology information if the hardware supports this.
2984 The scheduler will make use of this information and
2985 e.g. base its process migration decisions on it.
2990 tpm_suspend_pcr=[HW,TPM]
2991 Format: integer pcr id
2992 Specify that at suspend time, the tpm driver
2993 should extend the specified pcr with zeros,
2994 as a workaround for some chips which fail to
2995 flush the last written pcr on TPM_SaveState.
2996 This will guarantee that all the other pcrs
2999 trace_buf_size=nn[KMG]
3000 [FTRACE] will set tracing buffer size.
3002 trace_event=[event-list]
3003 [FTRACE] Set and start specified trace events in order
3004 to facilitate early boot debugging.
3005 See also Documentation/trace/events.txt
3007 trace_options=[option-list]
3008 [FTRACE] Enable or disable tracer options at boot.
3009 The option-list is a comma delimited list of options
3010 that can be enabled or disabled just as if you were
3011 to echo the option name into
3013 /sys/kernel/debug/tracing/trace_options
3015 For example, to enable stacktrace option (to dump the
3016 stack trace of each event), add to the command line:
3018 trace_options=stacktrace
3020 See also Documentation/trace/ftrace.txt "trace options"
3023 transparent_hugepage=
3025 Format: [always|madvise|never]
3026 Can be used to control the default behavior of the system
3027 with respect to transparent hugepages.
3028 See Documentation/vm/transhuge.txt for more details.
3030 tsc= Disable clocksource stability checks for TSC.
3032 [x86] reliable: mark tsc clocksource as reliable, this
3033 disables clocksource verification at runtime, as well
3034 as the stability checks done at bootup. Used to enable
3035 high-resolution timer mode on older hardware, and in
3036 virtualized environment.
3037 [x86] noirqtime: Do not use TSC to do irq accounting.
3038 Used to run time disable IRQ_TIME_ACCOUNTING on any
3039 platforms where RDTSC is slow and this accounting
3042 turbografx.map[2|3]= [HW,JOY]
3043 TurboGraFX parallel port interface
3045 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
3046 See also Documentation/input/joystick-parport.txt
3048 udbg-immortal [PPC] When debugging early kernel crashes that
3049 happen after console_init() and before a proper
3050 console driver takes over, this boot options might
3051 help "seeing" what's going on.
3053 uhash_entries= [KNL,NET]
3054 Set number of hash buckets for UDP/UDP-Lite connections
3057 [USB] Ignore overcurrent events (default N).
3058 Some badly-designed motherboards generate lots of
3059 bogus events, for ports that aren't wired to
3060 anything. Set this parameter to avoid log spamming.
3061 Note that genuine overcurrent events won't be
3065 [X86] Cause panic on unknown NMI.
3067 usbcore.authorized_default=
3068 [USB] Default USB device authorization:
3069 (default -1 = authorized except for wireless USB,
3070 0 = not authorized, 1 = authorized)
3072 usbcore.autosuspend=
3073 [USB] The autosuspend time delay (in seconds) used
3074 for newly-detected USB devices (default 2). This
3075 is the time required before an idle device will be
3076 autosuspended. Devices for which the delay is set
3077 to a negative value won't be autosuspended at all.
3079 usbcore.usbfs_snoop=
3080 [USB] Set to log all usbfs traffic (default 0 = off).
3082 usbcore.blinkenlights=
3083 [USB] Set to cycle leds on hubs (default 0 = off).
3085 usbcore.old_scheme_first=
3086 [USB] Start with the old device initialization
3087 scheme (default 0 = off).
3089 usbcore.usbfs_memory_mb=
3090 [USB] Memory limit (in MB) for buffers allocated by
3091 usbfs (default = 16, 0 = max = 2047).
3093 usbcore.use_both_schemes=
3094 [USB] Try the other device initialization scheme
3095 if the first one fails (default 1 = enabled).
3097 usbcore.initial_descriptor_timeout=
3098 [USB] Specifies timeout for the initial 64-byte
3099 USB_REQ_GET_DESCRIPTOR request in milliseconds
3100 (default 5000 = 5.0 seconds).
3103 [USBHID] The interval which mice are to be polled at.
3105 usb-storage.delay_use=
3106 [UMS] The delay in seconds before a new device is
3107 scanned for Logical Units (default 5).
3110 [UMS] A list of quirks entries to supplement or
3111 override the built-in unusual_devs list. List
3112 entries are separated by commas. Each entry has
3113 the form VID:PID:Flags where VID and PID are Vendor
3114 and Product ID values (4-digit hex numbers) and
3115 Flags is a set of characters, each corresponding
3116 to a common usb-storage quirk flag as follows:
3117 a = SANE_SENSE (collect more than 18 bytes
3119 b = BAD_SENSE (don't collect more than 18
3120 bytes of sense data);
3121 c = FIX_CAPACITY (decrease the reported
3122 device capacity by one sector);
3123 d = NO_READ_DISC_INFO (don't use
3124 READ_DISC_INFO command);
3125 e = NO_READ_CAPACITY_16 (don't use
3126 READ_CAPACITY_16 command);
3127 h = CAPACITY_HEURISTICS (decrease the
3128 reported device capacity by one
3129 sector if the number is odd);
3130 i = IGNORE_DEVICE (don't bind to this
3132 l = NOT_LOCKABLE (don't try to lock and
3133 unlock ejectable media);
3134 m = MAX_SECTORS_64 (don't transfer more
3135 than 64 sectors = 32 KB at a time);
3136 n = INITIAL_READ10 (force a retry of the
3137 initial READ(10) command);
3138 o = CAPACITY_OK (accept the capacity
3139 reported by the device);
3140 p = WRITE_CACHE (the device cache is ON
3142 r = IGNORE_RESIDUE (the device reports
3143 bogus residue values);
3144 s = SINGLE_LUN (the device has only one
3146 w = NO_WP_DETECT (don't test whether the
3147 medium is write-protected).
3148 Example: quirks=0419:aaf5:rl,0421:0433:rc
3150 user_debug= [KNL,ARM]
3152 See arch/arm/Kconfig.debug help text.
3153 1 - undefined instruction events
3155 4 - invalid data aborts
3158 Example: user_debug=31
3161 [X86] Flags controlling user PTE allocations.
3163 nohigh = do not allocate PTE pages in
3164 HIGHMEM regardless of setting
3168 vdso=2: enable compat VDSO (default with COMPAT_VDSO)
3169 vdso=1: enable VDSO (default)
3170 vdso=0: disable VDSO mapping
3173 vdso32=2: enable compat VDSO (default with COMPAT_VDSO)
3174 vdso32=1: enable 32-bit VDSO (default)
3175 vdso32=0: disable 32-bit VDSO mapping
3178 vector=percpu: enable percpu vector domain
3180 video= [FB] Frame buffer configuration
3181 See Documentation/fb/modedb.txt.
3184 [VMMIO] Memory mapped virtio (platform) device.
3186 <size>@<baseaddr>:<irq>[:<id>]
3188 <size> := size (can use standard suffixes
3190 <baseaddr> := physical base address
3191 <irq> := interrupt number (as passed to
3193 <id> := (optional) platform device id
3195 virtio_mmio.device=1K@0x100b0000:48:7
3197 Can be used multiple times for multiple devices.
3199 vga= [BOOT,X86-32] Select a particular video mode
3200 See Documentation/x86/boot.txt and
3201 Documentation/svga.txt.
3202 Use vga=ask for menu.
3203 This is actually a boot loader parameter; the value is
3204 passed to the kernel using a special protocol.
3206 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
3207 size of <nn>. This can be used to increase the
3208 minimum size (128MB on x86). It can also be used to
3209 decrease the size and leave more room for directly
3212 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
3215 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
3218 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
3222 Controls the behavior of vsyscalls (i.e. calls to
3223 fixed addresses of 0xffffffffff600x00 from legacy
3224 code). Most statically-linked binaries and older
3225 versions of glibc use these calls. Because these
3226 functions are at fixed addresses, they make nice
3227 targets for exploits that can control RIP.
3229 emulate [default] Vsyscalls turn into traps and are
3230 emulated reasonably safely.
3232 native Vsyscalls are native syscall instructions.
3233 This is a little bit faster than trapping
3234 and makes a few dynamic recompilers work
3235 better than they would in emulation mode.
3236 It also makes exploits much easier to write.
3238 none Vsyscalls don't work at all. This makes
3239 them quite hard to use for exploits but
3240 might break your system.
3242 vt.cur_default= [VT] Default cursor shape.
3243 Format: 0xCCBBAA, where AA, BB, and CC are the same as
3244 the parameters of the <Esc>[?A;B;Cc escape sequence;
3245 see VGA-softcursor.txt. Default: 2 = underline.
3247 vt.default_blu= [VT]
3248 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
3249 Change the default blue palette of the console.
3250 This is a 16-member array composed of values
3253 vt.default_grn= [VT]
3254 Format: <green0>,<green1>,<green2>,...,<green15>
3255 Change the default green palette of the console.
3256 This is a 16-member array composed of values
3259 vt.default_red= [VT]
3260 Format: <red0>,<red1>,<red2>,...,<red15>
3261 Change the default red palette of the console.
3262 This is a 16-member array composed of values
3268 Set system-wide default UTF-8 mode for all tty's.
3269 Default is 1, i.e. UTF-8 mode is enabled for all
3270 newly opened terminals.
3272 vt.global_cursor_default=
3275 Set system-wide default for whether a cursor
3276 is shown on new VTs. Default is -1,
3277 i.e. cursors will be created by default unless
3278 overridden by individual drivers. 0 will hide
3279 cursors, 1 will display them.
3281 watchdog timers [HW,WDT] For information on watchdog timers,
3282 see Documentation/watchdog/watchdog-parameters.txt
3283 or other driver-specific files in the
3284 Documentation/watchdog/ directory.
3286 workqueue.disable_numa
3287 By default, all work items queued to unbound
3288 workqueues are affine to the NUMA nodes they're
3289 issued on, which results in better behavior in
3290 general. If NUMA affinity needs to be disabled for
3291 whatever reason, this option can be used. Note
3292 that this also can be controlled per-workqueue for
3293 workqueues visible under /sys/bus/workqueue/.
3295 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
3296 default x2apic cluster mode on platforms
3299 x86_mrst_timer= [X86-32,APBT]
3300 Choose timer option for x86 Moorestown MID platform.
3301 Two valid options are apbt timer only and lapic timer
3302 plus one apbt timer for broadcast timer.
3303 x86_mrst_timer=apbt_only | lapic_and_apbt
3305 xd= [HW,XT] Original XT pre-IDE (RLL encoded) disks.
3306 xd_geo= See header of drivers/block/xd.c.
3308 xen_emul_unplug= [HW,X86,XEN]
3309 Unplug Xen emulated devices
3310 Format: [unplug0,][unplug1]
3311 ide-disks -- unplug primary master IDE devices
3312 aux-ide-disks -- unplug non-primary-master IDE devices
3313 nics -- unplug network devices
3314 all -- unplug all emulated devices (NICs and IDE disks)
3315 unnecessary -- unplugging emulated devices is
3316 unnecessary even if the host did not respond to
3318 never -- do not unplug even if version check succeeds
3320 xirc2ps_cs= [NET,PCMCIA]
3322 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
3324 ______________________________________________________________________
3328 Add more DRM drivers.
projects / cascardo / linux.git / blob