4 The following is a consolidated list of the kernel parameters as
5 implemented by the __setup(), core_param() and module_param() macros
6 and sorted into English Dictionary order (defined as ignoring all
7 punctuation and sorting digits before letters in a case insensitive
8 manner), and with descriptions where known.
10 The kernel parses parameters from the kernel command line up to "--";
11 if it doesn't recognize a parameter and it doesn't contain a '.', the
12 parameter gets passed to init: parameters with '=' go into init's
13 environment, others are passed as command line arguments to init.
14 Everything after "--" is passed as an argument to init.
16 Module parameters can be specified in two ways: via the kernel command
17 line with a module name prefix, or via modprobe, e.g.:
19 (kernel command line) usbcore.blinkenlights=1
20 (modprobe command line) modprobe usbcore blinkenlights=1
22 Parameters for modules which are built into the kernel need to be
23 specified on the kernel command line. modprobe looks through the
24 kernel command line (/proc/cmdline) and collects module parameters
25 when it loads a module, so the kernel command line can be used for
28 Hyphens (dashes) and underscores are equivalent in parameter names, so
29 log_buf_len=1M print-fatal-signals=1
30 can also be entered as
31 log-buf-len=1M print_fatal_signals=1
33 Double-quotes can be used to protect spaces in values, e.g.:
34 param="spaces in here"
36 This document may not be entirely up to date and comprehensive. The command
37 "modinfo -p ${modulename}" shows a current list of all parameters of a loadable
38 module. Loadable modules, after being loaded into the running kernel, also
39 reveal their parameters in /sys/module/${modulename}/parameters/. Some of these
40 parameters may be changed at runtime by the command
41 "echo -n ${value} > /sys/module/${modulename}/parameters/${parm}".
43 The parameters listed below are only valid if certain kernel build options were
44 enabled and if respective hardware is present. The text in square brackets at
45 the beginning of each description states the restrictions within which a
46 parameter is applicable:
48 ACPI ACPI support is enabled.
49 AGP AGP (Accelerated Graphics Port) is enabled.
50 ALSA ALSA sound support is enabled.
51 APIC APIC support is enabled.
52 APM Advanced Power Management support is enabled.
53 ARM ARM architecture is enabled.
54 AVR32 AVR32 architecture is enabled.
55 AX25 Appropriate AX.25 support is enabled.
56 BLACKFIN Blackfin architecture is enabled.
57 CLK Common clock infrastructure is enabled.
58 CMA Contiguous Memory Area support is enabled.
59 DRM Direct Rendering Management support is enabled.
60 DYNAMIC_DEBUG Build in debug messages and enable them at runtime
61 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled
62 EFI EFI Partitioning (GPT) is enabled
63 EIDE EIDE/ATAPI support is enabled.
64 EVM Extended Verification Module
65 FB The frame buffer device is enabled.
66 FTRACE Function tracing enabled.
67 GCOV GCOV profiling is enabled.
68 HW Appropriate hardware is enabled.
69 IA-64 IA-64 architecture is enabled.
70 IMA Integrity measurement architecture is enabled.
71 IOSCHED More than one I/O scheduler is enabled.
72 IP_PNP IP DHCP, BOOTP, or RARP is enabled.
73 IPV6 IPv6 support is enabled.
74 ISAPNP ISA PnP code is enabled.
75 ISDN Appropriate ISDN support is enabled.
76 JOY Appropriate joystick support is enabled.
77 KGDB Kernel debugger support is enabled.
78 KVM Kernel Virtual Machine support is enabled.
79 LIBATA Libata driver is enabled
80 LP Printer support is enabled.
81 LOOP Loopback device support is enabled.
82 M68k M68k architecture is enabled.
83 These options have more detailed description inside of
84 Documentation/m68k/kernel-options.txt.
85 MDA MDA console support is enabled.
86 MIPS MIPS architecture is enabled.
87 MOUSE Appropriate mouse support is enabled.
88 MSI Message Signaled Interrupts (PCI).
89 MTD MTD (Memory Technology Device) support is enabled.
90 NET Appropriate network support is enabled.
91 NUMA NUMA support is enabled.
92 NFS Appropriate NFS support is enabled.
93 OSS OSS sound support is enabled.
94 PV_OPS A paravirtualized kernel is enabled.
95 PARIDE The ParIDE (parallel port IDE) subsystem is enabled.
96 PARISC The PA-RISC architecture is enabled.
97 PCI PCI bus support is enabled.
98 PCIE PCI Express support is enabled.
99 PCMCIA The PCMCIA subsystem is enabled.
100 PNP Plug & Play support is enabled.
101 PPC PowerPC architecture is enabled.
102 PPT Parallel port support is enabled.
103 PS2 Appropriate PS/2 support is enabled.
104 RAM RAM disk support is enabled.
105 S390 S390 architecture is enabled.
106 SCSI Appropriate SCSI support is enabled.
107 A lot of drivers have their options described inside
108 the Documentation/scsi/ sub-directory.
109 SECURITY Different security models are enabled.
110 SELINUX SELinux support is enabled.
111 APPARMOR AppArmor support is enabled.
112 SERIAL Serial support is enabled.
113 SH SuperH architecture is enabled.
114 SMP The kernel is an SMP kernel.
115 SPARC Sparc architecture is enabled.
116 SWSUSP Software suspend (hibernation) is enabled.
117 SUSPEND System suspend states are enabled.
118 TPM TPM drivers are enabled.
119 TS Appropriate touchscreen support is enabled.
120 UMS USB Mass Storage support is enabled.
121 USB USB support is enabled.
122 USBHID USB Human Interface Device support is enabled.
123 V4L Video For Linux support is enabled.
124 VMMIO Driver for memory mapped virtio devices is enabled.
125 VGA The VGA console has been enabled.
126 VT Virtual terminal support is enabled.
127 WDT Watchdog support is enabled.
128 XT IBM PC/XT MFM hard disk support is enabled.
129 X86-32 X86-32, aka i386 architecture is enabled.
130 X86-64 X86-64 architecture is enabled.
131 More X86-64 boot options can be found in
132 Documentation/x86/x86_64/boot-options.txt .
133 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64)
134 X86_UV SGI UV support is enabled.
135 XEN Xen support is enabled
137 In addition, the following text indicates that the option:
139 BUGS= Relates to possible processor bugs on the said processor.
140 KNL Is a kernel start-up parameter.
141 BOOT Is a boot loader parameter.
143 Parameters denoted with BOOT are actually interpreted by the boot
144 loader, and have no meaning to the kernel directly.
145 Do not modify the syntax of boot loader parameters without extreme
146 need or coordination with <Documentation/x86/boot.txt>.
148 There are also arch-specific kernel-parameters not documented here.
149 See for example <Documentation/x86/x86_64/boot-options.txt>.
151 Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
152 a trailing = on the name of any parameter states that that parameter will
153 be entered as an environment variable, whereas its absence indicates that
154 it will appear as a kernel argument readable via /proc/cmdline by programs
155 running once the system is up.
157 The number of kernel parameters is not limited, but the length of the
158 complete command line (parameters including spaces etc.) is limited to
159 a fixed number of characters. This limit depends on the architecture
160 and is between 256 and 4096 characters. It is defined in the file
161 ./include/asm/setup.h as COMMAND_LINE_SIZE.
163 Finally, the [KMG] suffix is commonly described after a number of kernel
164 parameter values. These 'K', 'M', and 'G' letters represent the _binary_
165 multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
166 bytes respectively. Such letter suffixes can also be entirely omitted.
169 acpi= [HW,ACPI,X86,ARM64]
170 Advanced Configuration and Power Interface
171 Format: { force | on | off | strict | noirq | rsdt |
173 force -- enable ACPI if default was off
174 on -- enable ACPI but allow fallback to DT [arm64]
175 off -- disable ACPI if default was on
176 noirq -- do not use ACPI for IRQ routing
177 strict -- Be less tolerant of platforms that are not
178 strictly ACPI specification compliant.
179 rsdt -- prefer RSDT over (default) XSDT
180 copy_dsdt -- copy DSDT to memory
181 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force"
184 See also Documentation/power/runtime_pm.txt, pci=noacpi
186 acpi_apic_instance= [ACPI, IOAPIC]
188 2: use 2nd APIC table, if available
189 1,0: use 1st APIC table
192 acpi_backlight= [HW,ACPI]
193 acpi_backlight=vendor
195 If set to vendor, prefer vendor specific driver
196 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
197 of the ACPI video.ko driver.
199 acpi_force_32bit_fadt_addr
200 force FADT to use 32 bit addresses rather than the
201 64 bit X_* addresses. Some firmware have broken 64
202 bit addresses for force ACPI ignore these and use
203 the older legacy 32 bit addresses.
205 acpica_no_return_repair [HW, ACPI]
206 Disable AML predefined validation mechanism
207 This mechanism can repair the evaluation result to make
208 the return objects more ACPI specification compliant.
209 This option is useful for developers to identify the
210 root cause of an AML interpreter issue when the issue
211 has something to do with the repair mechanism.
213 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
214 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
216 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
217 debug output. Bits in debug_layer correspond to a
218 _COMPONENT in an ACPI source file, e.g.,
219 #define _COMPONENT ACPI_PCI_COMPONENT
220 Bits in debug_level correspond to a level in
221 ACPI_DEBUG_PRINT statements, e.g.,
222 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
223 The debug_level mask defaults to "info". See
224 Documentation/acpi/debug.txt for more information about
225 debug layers and levels.
227 Enable processor driver info messages:
228 acpi.debug_layer=0x20000000
229 Enable PCI/PCI interrupt routing info messages:
230 acpi.debug_layer=0x400000
231 Enable AML "Debug" output, i.e., stores to the Debug
232 object while interpreting AML:
233 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
234 Enable all messages related to ACPI hardware:
235 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
237 Some values produce so much output that the system is
238 unusable. The "log_buf_len" parameter may be useful
239 if you need to capture more output.
241 acpi_enforce_resources= [ACPI]
242 { strict | lax | no }
243 Check for resource conflicts between native drivers
244 and ACPI OperationRegions (SystemIO and SystemMemory
245 only). IO ports and memory declared in ACPI might be
246 used by the ACPI subsystem in arbitrary AML code and
247 can interfere with legacy drivers.
248 strict (default): access to resources claimed by ACPI
249 is denied; legacy drivers trying to access reserved
250 resources will fail to bind to device using them.
251 lax: access to resources claimed by ACPI is allowed;
252 legacy drivers trying to access reserved resources
253 will bind successfully but a warning message is logged.
254 no: ACPI OperationRegions are not marked as reserved,
255 no further checks are performed.
257 acpi_force_table_verification [HW,ACPI]
258 Enable table checksum verification during early stage.
259 By default, this is disabled due to x86 early mapping
262 acpi_irq_balance [HW,ACPI]
263 ACPI will balance active IRQs
266 acpi_irq_nobalance [HW,ACPI]
267 ACPI will not move active IRQs (default)
270 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
271 Format: <irq>,<irq>...
273 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
275 Format: <irq>,<irq>...
277 acpi_no_auto_serialize [HW,ACPI]
278 Disable auto-serialization of AML methods
279 AML control methods that contain the opcodes to create
280 named objects will be marked as "Serialized" by the
281 auto-serialization feature.
282 This feature is enabled by default.
283 This option allows to turn off the feature.
285 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
288 acpi_no_static_ssdt [HW,ACPI]
289 Disable installation of static SSDTs at early boot time
290 By default, SSDTs contained in the RSDT/XSDT will be
291 installed automatically and they will appear under
292 /sys/firmware/acpi/tables.
293 This option turns off this feature.
294 Note that specifying this option does not affect
295 dynamic table installation which will install SSDT
296 tables to /sys/firmware/acpi/tables/dynamic.
298 acpi_rsdp= [ACPI,EFI,KEXEC]
299 Pass the RSDP address to the kernel, mostly used
300 on machines running EFI runtime service to boot the
301 second kernel for kdump.
303 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
304 Format: To spoof as Windows 98: ="Microsoft Windows"
306 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
307 of 2 which is mandated by ACPI 6) as the supported ACPI
308 specification revision (when using this switch, it may
309 be necessary to carry out a cold reboot _twice_ in a
310 row to make it take effect on the platform firmware).
312 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
313 acpi_osi="string1" # add string1
314 acpi_osi="!string2" # remove string2
315 acpi_osi=!* # remove all strings
316 acpi_osi=! # disable all built-in OS vendor
318 acpi_osi=!! # enable all built-in OS vendor
320 acpi_osi= # disable all strings
322 'acpi_osi=!' can be used in combination with single or
323 multiple 'acpi_osi="string1"' to support specific OS
324 vendor string(s). Note that such command can only
325 affect the default state of the OS vendor strings, thus
326 it cannot affect the default state of the feature group
327 strings and the current state of the OS vendor strings,
328 specifying it multiple times through kernel command line
329 is meaningless. This command is useful when one do not
330 care about the state of the feature group strings which
331 should be controlled by the OSPM.
333 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
334 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
335 can make '_OSI("Windows 2000")' TRUE.
337 'acpi_osi=' cannot be used in combination with other
338 'acpi_osi=' command lines, the _OSI method will not
339 exist in the ACPI namespace. NOTE that such command can
340 only affect the _OSI support state, thus specifying it
341 multiple times through kernel command line is also
344 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
347 'acpi_osi=!*' can be used in combination with single or
348 multiple 'acpi_osi="string1"' to support specific
349 string(s). Note that such command can affect the
350 current state of both the OS vendor strings and the
351 feature group strings, thus specifying it multiple times
352 through kernel command line is meaningful. But it may
353 still not able to affect the final state of a string if
354 there are quirks related to this string. This command
355 is useful when one want to control the state of the
356 feature group strings to debug BIOS issues related to
359 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
360 '_OSI("Module Device")' FALSE.
361 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
362 '_OSI("Module Device")' TRUE.
363 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
365 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
367 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
368 they all will make '_OSI("Windows 2000")' TRUE.
371 Override the pmtimer bug detection: force the kernel
372 to assume that this machine's pmtimer latches its value
373 and always returns good values.
375 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
376 Format: { level | edge | high | low }
378 acpi_skip_timer_override [HW,ACPI]
379 Recognize and ignore IRQ0/pin2 Interrupt Override.
380 For broken nForce2 BIOS resulting in XT-PIC timer.
382 acpi_sleep= [HW,ACPI] Sleep options
383 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
384 old_ordering, nonvs, sci_force_enable }
385 See Documentation/power/video.txt for information on
387 s3_beep is for debugging; it makes the PC's speaker beep
388 as soon as the kernel's real-mode entry point is called.
389 s4_nohwsig prevents ACPI hardware signature from being
390 used during resume from hibernation.
391 old_ordering causes the ACPI 1.0 ordering of the _PTS
392 control method, with respect to putting devices into
393 low power states, to be enforced (the ACPI 2.0 ordering
394 of _PTS is used by default).
395 nonvs prevents the kernel from saving/restoring the
396 ACPI NVS memory during suspend/hibernation and resume.
397 sci_force_enable causes the kernel to set SCI_EN directly
398 on resume from S1/S3 (which is against the ACPI spec,
399 but some broken systems don't work without it).
401 acpi_use_timer_override [HW,ACPI]
402 Use timer override. For some broken Nvidia NF5 boards
403 that require a timer override, but don't have HPET
405 add_efi_memmap [EFI; X86] Include EFI memory map in
406 kernel's map of available physical RAM.
409 { off | try_unsupported }
410 off: disable AGP support
411 try_unsupported: try to drive unsupported chipsets
412 (may crash computer or cause data corruption)
415 See Documentation/sound/alsa/alsa-parameters.txt
418 Allow the default userspace alignment fault handler
419 behaviour to be specified. Bit 0 enables warnings,
420 bit 1 enables fixups, and bit 2 sends a segfault.
422 align_va_addr= [X86-64]
423 Align virtual addresses by clearing slice [14:12] when
424 allocating a VMA at process creation time. This option
425 gives you up to 3% performance improvement on AMD F15h
426 machines (where it is enabled by default) for a
427 CPU-intensive style benchmark, and it can vary highly in
428 a microbenchmark depending on workload and compiler.
430 32: only for 32-bit processes
431 64: only for 64-bit processes
432 on: enable for both 32- and 64-bit processes
433 off: disable for both 32- and 64-bit processes
435 alloc_snapshot [FTRACE]
436 Allocate the ftrace snapshot buffer on boot up when the
437 main buffer is allocated. This is handy if debugging
438 and you need to use tracing_snapshot() on boot up, and
439 do not want to use tracing_snapshot_alloc() as it needs
440 to be done where GFP_KERNEL allocations are allowed.
442 amd_iommu= [HW,X86-64]
443 Pass parameters to the AMD IOMMU driver in the system.
445 fullflush - enable flushing of IO/TLB entries when
446 they are unmapped. Otherwise they are
447 flushed before they will be reused, which
449 off - do not initialize any AMD IOMMU found in
451 force_isolation - Force device isolation for all
452 devices. The IOMMU driver is not
453 allowed anymore to lift isolation
454 requirements as needed. This option
455 does not override iommu=pt
457 amd_iommu_dump= [HW,X86-64]
458 Enable AMD IOMMU driver option to dump the ACPI table
459 for AMD IOMMU. With this option enabled, AMD IOMMU
460 driver will print ACPI tables for AMD IOMMU during
461 IOMMU initialization.
463 amijoy.map= [HW,JOY] Amiga joystick support
464 Map of devices attached to JOY0DAT and JOY1DAT
466 See also Documentation/input/joystick.txt
468 analog.map= [HW,JOY] Analog joystick and gamepad support
469 Specifies type or capabilities of an analog joystick
470 connected to one of 16 gameports
471 Format: <type1>,<type2>,..<type16>
474 Power management functions (SPARCstation-4/5 + deriv.)
476 Disable APC CPU standby support. SPARCstation-Fox does
477 not play well with APC CPU idle - disable it if you have
478 APC and your system crashes randomly.
480 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
481 Change the output verbosity whilst booting
482 Format: { quiet (default) | verbose | debug }
483 Change the amount of debugging information output
484 when initialising the APIC and IO-APIC components.
486 apic_extnmi= [APIC,X86] External NMI delivery setting
487 Format: { bsp (default) | all | none }
488 bsp: External NMI is delivered only to CPU 0
489 all: External NMIs are broadcast to all CPUs as a
491 none: External NMI is masked for all CPUs. This is
492 useful so that a dump capture kernel won't be
496 See Documentation/networking/ipv6.txt.
498 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
499 Limit apic dumping. The parameter defines the maximal
500 number of local apics being dumped. Also it is possible
501 to set it to "all" by meaning -- no limit here.
502 Format: { 1 (default) | 2 | ... | all }.
503 The parameter valid if only apic=debug or
504 apic=verbose is specified.
505 Example: apic=debug show_lapic=all
507 apm= [APM] Advanced Power Management
508 See header of arch/x86/kernel/apm_32.c.
510 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
511 Format: <io>,<irq>,<nodeID>
515 atarimouse= [HW,MOUSE] Atari Mouse
517 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
518 EzKey and similar keyboards
520 atkbd.reset= [HW] Reset keyboard during initialization
522 atkbd.set= [HW] Select keyboard code set
523 Format: <int> (2 = AT (default), 3 = PS/2)
525 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
528 atkbd.softraw= [HW] Choose between synthetic and real raw mode
529 Format: <bool> (0 = real, 1 = synthetic (default))
531 atkbd.softrepeat= [HW]
532 Use software keyboard repeat
534 audit= [KNL] Enable the audit sub-system
535 Format: { "0" | "1" } (0 = disabled, 1 = enabled)
536 0 - kernel audit is disabled and can not be enabled
537 until the next reboot
538 unset - kernel audit is initialized but disabled and
539 will be fully enabled by the userspace auditd.
540 1 - kernel audit is initialized and partially enabled,
541 storing at most audit_backlog_limit messages in
542 RAM until it is fully enabled by the userspace
546 audit_backlog_limit= [KNL] Set the audit queue size limit.
547 Format: <int> (must be >=0)
550 bau= [X86_UV] Enable the BAU on SGI UV. The default
551 behavior is to disable the BAU (i.e. bau=0).
552 Format: { "0" | "1" }
555 unset - Disable the BAU.
557 baycom_epp= [HW,AX25]
560 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
562 See header of drivers/net/hamradio/baycom_par.c.
564 baycom_ser_fdx= [HW,AX25]
565 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
566 Format: <io>,<irq>,<mode>[,<baud>]
567 See header of drivers/net/hamradio/baycom_ser_fdx.c.
569 baycom_ser_hdx= [HW,AX25]
570 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
571 Format: <io>,<irq>,<mode>
572 See header of drivers/net/hamradio/baycom_ser_hdx.c.
574 blkdevparts= Manual partition parsing of block device(s) for
575 embedded devices based on command line input.
576 See Documentation/block/cmdline-partition.txt
578 boot_delay= Milliseconds to delay each printk during boot.
579 Values larger than 10 seconds (10000) are changed to
583 bootmem_debug [KNL] Enable bootmem allocator debug messages.
586 Disable BERT OS support on buggy BIOSes.
588 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
589 bttv.radio= Most important insmod options are available as
591 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
594 bulk_remove=off [PPC] This parameter disables the use of the pSeries
595 firmware feature for flushing multiple hpte entries
598 c101= [NET] Moxa C101 synchronous serial card
600 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
601 Sometimes CPU hardware bugs make them report the cache
602 size incorrectly. The kernel will attempt work arounds
603 to fix known problems, but for some CPUs it is not
604 possible to determine what the correct size should be.
605 This option provides an override for these situations.
607 ca_keys= [KEYS] This parameter identifies a specific key(s) on
608 the system trusted keyring to be used for certificate
610 format: { id:<keyid> | builtin }
612 cca= [MIPS] Override the kernel pages' cache coherency
613 algorithm. Accepted values range from 0 to 7
614 inclusive. See arch/mips/include/asm/pgtable-bits.h
615 for platform specific values (SB1, Loongson3 and
618 ccw_timeout_log [S390]
619 See Documentation/s390/CommonIO for details.
621 cgroup_disable= [KNL] Disable a particular controller
622 Format: {name of the controller(s) to disable}
623 The effects of cgroup_disable=foo are:
624 - foo isn't auto-mounted if you mount all cgroups in
626 - foo isn't visible as an individually mountable
628 {Currently only "memory" controller deal with this and
629 cut the overhead, others just disable the usage. So
630 only cgroup_disable=memory is actually worthy}
632 cgroup_no_v1= [KNL] Disable one, multiple, all cgroup controllers in v1
633 Format: { controller[,controller...] | "all" }
634 Like cgroup_disable, but only applies to cgroup v1;
635 the blacklisted controllers remain available in cgroup2.
637 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
639 nosocket -- Disable socket memory accounting.
640 nokmem -- Disable kernel memory accounting.
642 checkreqprot [SELINUX] Set initial checkreqprot flag value.
643 Format: { "0" | "1" }
644 See security/selinux/Kconfig help text.
645 0 -- check protection applied by kernel (includes
646 any implied execute protection).
647 1 -- check protection requested by application.
648 Default value is set via a kernel config option.
649 Value can be changed at runtime via
650 /selinux/checkreqprot.
653 See Documentation/s390/CommonIO for details.
656 Prevents the clock framework from automatically gating
657 clocks that have not been explicitly enabled by a Linux
658 device driver but are enabled in hardware at reset or
659 by the bootloader/firmware. Note that this does not
660 force such clocks to be always-on nor does it reserve
661 those clocks in any way. This parameter is useful for
662 debug and development, but should not be needed on a
663 platform with proper driver support. For more
664 information, see Documentation/clk.txt.
666 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
668 Forces specified clocksource (if available) to be used
669 when calculating gettimeofday(). If specified
670 clocksource is not available, it defaults to PIT.
671 Format: { pit | tsc | cyclone | pmtmr }
673 clocksource= Override the default clocksource
675 Override the default clocksource and use the clocksource
676 with the name specified.
677 Some clocksource names to choose from, depending on
679 [all] jiffies (this is the base, fallback clocksource)
681 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
682 pxa_timer,timer3,32k_counter,timer0_1
684 [X86-32] pit,hpet,tsc;
685 scx200_hrt on Geode; cyclone on IBM x440
693 clearcpuid=BITNUM [X86]
694 Disable CPUID feature X for the kernel. See
695 arch/x86/include/asm/cpufeatures.h for the valid bit
696 numbers. Note the Linux specific bits are not necessarily
697 stable over kernel options, but the vendor specific
699 Also note that user programs calling CPUID directly
700 or using the feature without checking anything
701 will still see it. This just prevents it from
702 being used by the kernel or shown in /proc/cpuinfo.
703 Also note the kernel might malfunction if you disable
706 cma=nn[MG]@[start[MG][-end[MG]]]
708 Sets the size of kernel global memory area for
709 contiguous memory allocations and optionally the
710 placement constraint by the physical address range of
711 memory allocations. A value of 0 disables CMA
712 altogether. For more information, see
713 include/linux/dma-contiguous.h
715 cmo_free_hint= [PPC] Format: { yes | no }
716 Specify whether pages are marked as being inactive
717 when they are freed. This is used in CMO environments
718 to determine OS memory pressure for page stealing by
722 coherent_pool=nn[KMG] [ARM,KNL]
723 Sets the size of memory pool for coherent, atomic dma
724 allocations, by default set to 256K.
726 code_bytes [X86] How many bytes of object code to print
731 com20020= [HW,NET] ARCnet - COM20020 chipset
733 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
735 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
739 ARCnet - COM90xx chipset (memory-mapped buffers)
740 Format: <io>[,<irq>[,<memstart>]]
742 condev= [HW,S390] console device
745 console= [KNL] Output console device and options.
747 tty<n> Use the virtual console device <n>.
751 Use the specified serial port. The options are of
752 the form "bbbbpnf", where "bbbb" is the baud rate,
753 "p" is parity ("n", "o", or "e"), "n" is number of
754 bits, and "f" is flow control ("r" for RTS or
755 omit it). Default is "9600n8".
757 See Documentation/serial-console.txt for more
759 Documentation/networking/netconsole.txt for an
762 uart[8250],io,<addr>[,options]
763 uart[8250],mmio,<addr>[,options]
764 uart[8250],mmio16,<addr>[,options]
765 uart[8250],mmio32,<addr>[,options]
766 uart[8250],0x<addr>[,options]
767 Start an early, polled-mode console on the 8250/16550
768 UART at the specified I/O port or MMIO address,
769 switching to the matching ttyS device later.
770 MMIO inter-register address stride is either 8-bit
771 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
772 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
773 to be equivalent to 'mmio'. 'options' are specified in
774 the same format described for ttyS above; if unspecified,
775 the h/w is not re-initialized.
777 hvc<n> Use the hypervisor console device <n>. This is for
778 both Xen and PowerPC hypervisors.
780 If the device connected to the port is not a TTY but a braille
781 device, prepend "brl," before the device type, for instance
783 For now, only VisioBraille is supported.
785 consoleblank= [KNL] The console blank (screen saver) timeout in
786 seconds. Defaults to 10*60 = 10mins. A value of 0
787 disables the blank timer.
790 [KNL] Change the default value for
791 /proc/<pid>/coredump_filter.
792 See also Documentation/filesystems/proc.txt.
794 cpuidle.off=1 [CPU_IDLE]
795 disable the cpuidle sub-system
798 [X86] Delay for N microsec between assert and de-assert
799 of APIC INIT to start processors. This delay occurs
800 on every CPU online, such as boot, and resume from suspend.
803 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
805 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
807 crashkernel=size[KMG][@offset[KMG]]
808 [KNL] Using kexec, Linux can switch to a 'crash kernel'
809 upon panic. This parameter reserves the physical
810 memory region [offset, offset + size] for that kernel
811 image. If '@offset' is omitted, then a suitable offset
812 is selected automatically. Check
813 Documentation/kdump/kdump.txt for further details.
815 crashkernel=range1:size1[,range2:size2,...][@offset]
816 [KNL] Same as above, but depends on the memory
817 in the running system. The syntax of range is
818 start-[end] where start and end are both
819 a memory unit (amount[KMG]). See also
820 Documentation/kdump/kdump.txt for an example.
822 crashkernel=size[KMG],high
823 [KNL, x86_64] range could be above 4G. Allow kernel
824 to allocate physical memory region from top, so could
825 be above 4G if system have more than 4G ram installed.
826 Otherwise memory region will be allocated below 4G, if
828 It will be ignored if crashkernel=X is specified.
829 crashkernel=size[KMG],low
830 [KNL, x86_64] range under 4G. When crashkernel=X,high
831 is passed, kernel could allocate physical memory region
832 above 4G, that cause second kernel crash on system
833 that require some amount of low memory, e.g. swiotlb
834 requires at least 64M+32K low memory, also enough extra
835 low memory is needed to make sure DMA buffers for 32-bit
836 devices won't run out. Kernel would try to allocate at
837 at least 256M below 4G automatically.
838 This one let user to specify own low range under 4G
839 for second kernel instead.
840 0: to disable low allocation.
841 It will be ignored when crashkernel=X,high is not used
842 or memory reserved is below 4G.
845 [KNL] Disable crypto self-tests
850 cs89x0_media= [HW,NET]
851 Format: { rj45 | aui | bnc }
854 See header of drivers/s390/block/dasd_devmap.c.
856 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
857 (one device per port)
858 Format: <port#>,<type>
859 See also Documentation/input/joystick-parport.txt
861 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
862 time. See Documentation/dynamic-debug-howto.txt for
863 details. Deprecated, see dyndbg.
865 debug [KNL] Enable kernel debugging (events log level).
868 [KNL] verbose self-tests
870 Print debugging info while doing the locking API
872 We default to 0 (no extra messages), setting it to
873 1 will print _a lot_ more information - normally
874 only useful to kernel developers.
876 debug_objects [KNL] Enable object debugging
879 [KNL] Disable object debugging
881 debug_guardpage_minorder=
882 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
883 parameter allows control of the order of pages that will
884 be intentionally kept free (and hence protected) by the
885 buddy allocator. Bigger value increase the probability
886 of catching random memory corruption, but reduce the
887 amount of memory for normal system use. The maximum
888 possible value is MAX_ORDER/2. Setting this parameter
889 to 1 or 2 should be enough to identify most random
890 memory corruption problems caused by bugs in kernel or
891 driver code when a CPU writes to (or reads from) a
892 random memory location. Note that there exists a class
893 of memory corruptions problems caused by buggy H/W or
894 F/W or by drivers badly programing DMA (basically when
895 memory is written at bus level and the CPU MMU is
896 bypassed) which are not detectable by
897 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
898 tracking down these problems.
901 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
902 parameter enables the feature at boot time. In
903 default, it is disabled. We can avoid allocating huge
904 chunk of memory for debug pagealloc if we don't enable
905 it at boot time and the system will work mostly same
906 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
907 on: enable the feature
909 debugpat [X86] Enable PAT debugging
911 decnet.addr= [HW,NET]
912 Format: <area>[,<node>]
913 See also Documentation/networking/decnet.txt.
916 [same as hugepagesz=] The size of the default
917 HugeTLB page size. This is the size represented by
918 the legacy /proc/ hugepages APIs, used for SHM, and
919 default size when mounting hugetlbfs filesystems.
920 Defaults to the default architecture's huge page size
924 Set number of hash buckets for dentry cache.
927 See Documentation/networking/ipv6.txt.
929 disable_cpu_apicid= [X86,APIC,SMP]
931 The number of initial APIC ID for the
932 corresponding CPU to be disabled at boot,
933 mostly used for the kdump 2nd kernel to
934 disable BSP to wake up multiple CPUs without
935 causing system reset or hang due to sending
938 disable_ddw [PPC/PSERIES]
939 Disable Dynamic DMA Window support. Use this if
940 to workaround buggy firmware.
943 See Documentation/networking/ipv6.txt.
945 disable_mtrr_cleanup [X86]
946 The kernel tries to adjust MTRR layout from continuous
947 to discrete, to make X server driver able to add WB
948 entry later. This parameter disables that.
950 disable_mtrr_trim [X86, Intel and AMD only]
951 By default the kernel will trim any uncacheable
952 memory out of your available memory pool based on
953 MTRR settings. This parameter disables that behavior,
954 possibly causing your machine to run very slowly.
956 disable_timer_pin_1 [X86]
957 Disable PIN 1 of APIC timer
958 Can be useful to work around chipset bugs.
960 dis_ucode_ldr [X86] Disable the microcode loader.
962 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
963 this option disables the debugging code at boot.
965 dma_debug_entries=<number>
966 This option allows to tune the number of preallocated
967 entries for DMA-API debugging code. One entry is
968 required per DMA-API allocation. Use this if the
969 DMA-API debugging code disables itself because the
970 architectural default is too low.
972 dma_debug_driver=<driver_name>
973 With this option the DMA-API debugging driver
974 filter feature can be enabled at boot time. Just
975 pass the driver to filter for as the parameter.
976 The filter can be disabled or changed to another
977 driver later using sysfs.
979 drm_kms_helper.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
980 Broken monitors, graphic adapters, KVMs and EDIDless
981 panels may send no or incorrect EDID data sets.
982 This parameter allows to specify an EDID data sets
983 in the /lib/firmware directory that are used instead.
984 Generic built-in EDID data sets are used, if one of
985 edid/1024x768.bin, edid/1280x1024.bin,
986 edid/1680x1050.bin, or edid/1920x1080.bin is given
987 and no file with the same name exists. Details and
988 instructions how to build your own EDID data are
989 available in Documentation/EDID/HOWTO.txt. An EDID
990 data set will only be used for a particular connector,
991 if its name and a colon are prepended to the EDID
992 name. Each connector may use a unique EDID data
993 set by separating the files with a comma. An EDID
994 data set with no connector name will be used for
995 any connectors not explicitly specified.
999 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1000 module.dyndbg[="val"]
1001 Enable debug messages at boot time. See
1002 Documentation/dynamic-debug-howto.txt for details.
1004 nompx [X86] Disables Intel Memory Protection Extensions.
1005 See Documentation/x86/intel_mpx.txt for more
1006 information about the feature.
1008 nopku [X86] Disable Memory Protection Keys CPU feature found
1012 on enable eager fpu restore
1013 off disable eager fpu restore
1014 auto selects the default scheme, which automatically
1015 enables eagerfpu restore for xsaveopt.
1017 module.async_probe [KNL]
1018 Enable asynchronous probe on this module.
1020 early_ioremap_debug [KNL]
1021 Enable debug messages in early_ioremap support. This
1022 is useful for tracking down temporary early mappings
1023 which are not unmapped.
1025 earlycon= [KNL] Output early console device and options.
1027 When used with no options, the early console is
1028 determined by the stdout-path property in device
1032 Start an early, polled-mode console on a cadence serial
1033 port at the specified address. The cadence serial port
1034 must already be setup and configured. Options are not
1037 uart[8250],io,<addr>[,options]
1038 uart[8250],mmio,<addr>[,options]
1039 uart[8250],mmio32,<addr>[,options]
1040 uart[8250],mmio32be,<addr>[,options]
1041 uart[8250],0x<addr>[,options]
1042 Start an early, polled-mode console on the 8250/16550
1043 UART at the specified I/O port or MMIO address.
1044 MMIO inter-register address stride is either 8-bit
1045 (mmio) or 32-bit (mmio32 or mmio32be).
1046 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1047 to be equivalent to 'mmio'. 'options' are specified
1048 in the same format described for "console=ttyS<n>"; if
1049 unspecified, the h/w is not initialized.
1053 Start an early, polled-mode console on a pl011 serial
1054 port at the specified address. The pl011 serial port
1055 must already be setup and configured. Options are not
1056 yet supported. If 'mmio32' is specified, then only
1057 the driver will use only 32-bit accessors to read/write
1058 the device registers.
1061 Start an early, polled-mode console on a meson serial
1062 port at the specified address. The serial port must
1063 already be setup and configured. Options are not yet
1067 Start an early, polled-mode console on an msm serial
1068 port at the specified address. The serial port
1069 must already be setup and configured. Options are not
1072 msm_serial_dm,<addr>
1073 Start an early, polled-mode console on an msm serial
1074 dm port at the specified address. The serial port
1075 must already be setup and configured. Options are not
1078 smh Use ARM semihosting calls for early console.
1086 Use early console provided by serial driver available
1087 on Samsung SoCs, requires selecting proper type and
1088 a correct base address of the selected UART port. The
1089 serial port must already be setup and configured.
1090 Options are not yet supported.
1094 Use early console provided by Freescale LP UART driver
1095 found on Freescale Vybrid and QorIQ LS1021A processors.
1096 A valid base address must be provided, and the serial
1097 port must already be setup and configured.
1099 armada3700_uart,<addr>
1100 Start an early, polled-mode console on the
1101 Armada 3700 serial port at the specified
1102 address. The serial port must already be setup
1103 and configured. Options are not yet supported.
1105 earlyprintk= [X86,SH,BLACKFIN,ARM,M68k]
1109 earlyprintk=serial[,ttySn[,baudrate]]
1110 earlyprintk=serial[,0x...[,baudrate]]
1111 earlyprintk=ttySn[,baudrate]
1112 earlyprintk=dbgp[debugController#]
1113 earlyprintk=pciserial,bus:device.function[,baudrate]
1115 earlyprintk is useful when the kernel crashes before
1116 the normal console is initialized. It is not enabled by
1117 default because it has some cosmetic problems.
1119 Append ",keep" to not disable it when the real console
1122 Only one of vga, efi, serial, or usb debug port can
1125 Currently only ttyS0 and ttyS1 may be specified by
1126 name. Other I/O ports may be explicitly specified
1127 on some architectures (x86 and arm at least) by
1128 replacing ttySn with an I/O port address, like this:
1129 earlyprintk=serial,0x1008,115200
1130 You can find the port for a given device in
1131 /proc/tty/driver/serial:
1132 2: uart:ST16650V2 port:00001008 irq:18 ...
1134 Interaction with the standard serial driver is not
1137 The VGA and EFI output is eventually overwritten by
1140 The xen output can only be used by Xen PV guests.
1142 edac_report= [HW,EDAC] Control how to report EDAC event
1143 Format: {"on" | "off" | "force"}
1144 on: enable EDAC to report H/W event. May be overridden
1145 by other higher priority error reporting module.
1146 off: disable H/W event reporting through EDAC.
1147 force: enforce the use of EDAC to report H/W event.
1150 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1153 This is designed to be used in conjunction with
1154 the boot argument: earlyprintk=vga
1157 Format: {"off" | "on" | "skip[mbr]"}
1160 Format: { "old_map", "nochunk", "noruntime", "debug" }
1161 old_map [X86-64]: switch to the old ioremap-based EFI
1162 runtime services mapping. 32-bit still uses this one by
1164 nochunk: disable reading files in "chunks" in the EFI
1165 boot stub, as chunking can cause problems with some
1166 firmware implementations.
1167 noruntime : disable EFI runtime services support
1168 debug: enable misc debug output
1170 efi_no_storage_paranoia [EFI; X86]
1171 Using this parameter you can use more than 50% of
1172 your efi variable storage. Use this parameter only if
1173 you are really sure that your UEFI does sane gc and
1174 fulfills the spec otherwise your board may brick.
1176 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1177 Add arbitrary attribute to specific memory range by
1178 updating original EFI memory map.
1179 Region of memory which aa attribute is added to is
1181 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1182 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1183 attribute is added to range 0x100000000-0x180000000 and
1184 0x10a0000000-0x1120000000.
1186 Using this parameter you can do debugging of EFI memmap
1187 related feature. For example, you can do debugging of
1188 Address Range Mirroring feature even if your box
1191 eisa_irq_edge= [PARISC,HW]
1192 See header of drivers/parisc/eisa.c.
1195 See comment before function elanfreq_setup() in
1196 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1199 Format: {"cfq" | "deadline" | "noop"}
1200 See Documentation/block/cfq-iosched.txt and
1201 Documentation/block/deadline-iosched.txt for details.
1203 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1204 Specifies physical address of start of kernel core
1205 image elf header and optionally the size. Generally
1206 kexec loader will pass this option to capture kernel.
1207 See Documentation/kdump/kdump.txt for details.
1209 enable_mtrr_cleanup [X86]
1210 The kernel tries to adjust MTRR layout from continuous
1211 to discrete, to make X server driver able to add WB
1212 entry later. This parameter enables that.
1214 enable_timer_pin_1 [X86]
1215 Enable PIN 1 of APIC timer
1216 Can be useful to work around chipset bugs
1217 (in particular on some ATI chipsets).
1218 The kernel tries to set a reasonable default.
1220 enforcing [SELINUX] Set initial enforcing status.
1222 See security/selinux/Kconfig help text.
1223 0 -- permissive (log only, no denials).
1224 1 -- enforcing (deny and log).
1226 Value can be changed at runtime via /selinux/enforce.
1229 Disable Error Record Serialization Table (ERST)
1232 ether= [HW,NET] Ethernet cards parameters
1233 This option is obsoleted by the "netdev=" option, which
1234 has equivalent usage. See its documentation for details.
1238 Permit 'security.evm' to be updated regardless of
1239 current integrity status.
1243 fail_make_request=[KNL]
1244 General fault injection mechanism.
1245 Format: <interval>,<probability>,<space>,<times>
1246 See also Documentation/fault-injection/.
1249 See Documentation/blockdev/floppy.txt.
1251 force_pal_cache_flush
1252 [IA-64] Avoid check_sal_cache_flush which may hang on
1253 buggy SAL_CACHE_FLUSH implementations. Using this
1254 parameter will force ia64_sal_cache_flush to call
1255 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1258 Forcefully enable Physical Address Extension (PAE).
1259 Many Pentium M systems disable PAE but may have a
1260 functionally usable PAE implementation.
1261 Warning: use of this parameter will taint the kernel
1262 and may cause unknown problems.
1265 [FTRACE] will set and start the specified tracer
1266 as early as possible in order to facilitate early
1269 ftrace_dump_on_oops[=orig_cpu]
1270 [FTRACE] will dump the trace buffers on oops.
1271 If no parameter is passed, ftrace will dump
1272 buffers of all CPUs, but if you pass orig_cpu, it will
1273 dump only the buffer of the CPU that triggered the
1276 ftrace_filter=[function-list]
1277 [FTRACE] Limit the functions traced by the function
1278 tracer at boot up. function-list is a comma separated
1279 list of functions. This list can be changed at run
1280 time by the set_ftrace_filter file in the debugfs
1283 ftrace_notrace=[function-list]
1284 [FTRACE] Do not trace the functions specified in
1285 function-list. This list can be changed at run time
1286 by the set_ftrace_notrace file in the debugfs
1289 ftrace_graph_filter=[function-list]
1290 [FTRACE] Limit the top level callers functions traced
1291 by the function graph tracer at boot up.
1292 function-list is a comma separated list of functions
1293 that can be changed at run time by the
1294 set_graph_function file in the debugfs tracing directory.
1296 ftrace_graph_notrace=[function-list]
1297 [FTRACE] Do not trace from the functions specified in
1298 function-list. This list is a comma separated list of
1299 functions that can be changed at run time by the
1300 set_graph_notrace file in the debugfs tracing directory.
1303 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1304 support via parallel port (up to 5 devices per port)
1305 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1306 See also Documentation/input/joystick-parport.txt
1310 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1314 gcov_persist= [GCOV] When non-zero (default), profiling data for
1315 kernel modules is saved and remains accessible via
1316 debugfs, even when the module is unloaded/reloaded.
1317 When zero, profiling data is discarded and associated
1318 debugfs files are removed at module unload time.
1320 gpt [EFI] Forces disk with valid GPT signature but
1321 invalid Protective MBR to be treated as GPT. If the
1322 primary GPT is corrupted, it enables the backup/alternate
1323 GPT to be used instead.
1325 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1326 the "Enable 0" bit of the configuration register.
1329 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1330 the "Enable 0" bit of the configuration register.
1333 grcan.select= [HW] Select which physical interface to use.
1336 grcan.txsize= [HW] Sets the size of the tx buffer.
1337 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1339 grcan.rxsize= [HW] Sets the size of the rx buffer.
1340 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1343 hardlockup_all_cpu_backtrace=
1344 [KNL] Should the hard-lockup detector generate
1345 backtraces on all cpus.
1348 hashdist= [KNL,NUMA] Large hashes allocated during boot
1349 are distributed across NUMA nodes. Defaults on
1350 for 64-bit NUMA, off otherwise.
1351 Format: 0 | 1 (for off | on)
1353 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1355 hd= [EIDE] (E)IDE hard drive subsystem geometry
1356 Format: <cyl>,<head>,<sect>
1359 Disable Hardware Error Source Table (HEST) support;
1360 corresponding firmware-first mode error processing
1361 logic will be disabled.
1363 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1364 size of <nn>. This works even on boxes that have no
1365 highmem otherwise. This also works to reduce highmem
1366 size on bigger boxes.
1368 highres= [KNL] Enable/disable high resolution timer mode.
1369 Valid parameters: "on", "off"
1373 See Documentation/isdn/README.HiSax.
1377 hpet= [X86-32,HPET] option to control HPET usage
1378 Format: { enable (default) | disable | force |
1380 disable: disable HPET and use PIT instead
1381 force: allow force enabled of undocumented chips (ICH4,
1383 verbose: show contents of HPET registers during setup
1385 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1386 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1388 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1389 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1390 On x86-64 and powerpc, this option can be specified
1391 multiple times interleaved with hugepages= to reserve
1392 huge pages of different sizes. Valid pages sizes on
1393 x86-64 are 2M (when the CPU supports "pse") and 1G
1394 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1396 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1397 terminal devices. Valid values: 0..8
1398 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1399 If specified, z/VM IUCV HVC accepts connections
1400 from listed z/VM user IDs only.
1402 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1403 hardware thread id mappings.
1404 Format: <cpu>:<hwthread>
1407 Do not unregister boot console at start. This is only
1408 useful for debugging when something happens in the window
1409 between unregistering the boot console and initializing
1412 i2c_bus= [HW] Override the default board specific I2C bus speed
1413 or register an additional I2C bus that is not
1414 registered from board initialization code.
1418 i8042.debug [HW] Toggle i8042 debug mode
1419 i8042.unmask_kbd_data
1420 [HW] Enable printing of interrupt data from the KBD port
1421 (disabled by default, and as a pre-condition
1422 requires that i8042.debug=1 be enabled)
1423 i8042.direct [HW] Put keyboard port into non-translated mode
1424 i8042.dumbkbd [HW] Pretend that controller can only read data from
1425 keyboard and cannot control its state
1426 (Don't attempt to blink the leds)
1427 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1428 i8042.nokbd [HW] Don't check/create keyboard port
1429 i8042.noloop [HW] Disable the AUX Loopback command while probing
1431 i8042.nomux [HW] Don't check presence of an active multiplexing
1433 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1435 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1436 i8042.reset [HW] Reset the controller during init and cleanup
1437 i8042.unlock [HW] Unlock (ignore) the keylock
1438 i8042.kbdreset [HW] Reset device connected to KBD port
1442 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1443 indicates that the driver is running on unsupported
1445 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1446 does not match list of supported models.
1448 [HW] Report power status in /proc/i8k
1449 (disabled by default)
1450 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1453 i915.invert_brightness=
1454 [DRM] Invert the sense of the variable that is used to
1455 set the brightness of the panel backlight. Normally a
1456 brightness value of 0 indicates backlight switched off,
1457 and the maximum of the brightness value sets the backlight
1458 to maximum brightness. If this parameter is set to 0
1459 (default) and the machine requires it, or this parameter
1460 is set to 1, a brightness value of 0 sets the backlight
1461 to maximum brightness, and the maximum of the brightness
1462 value switches the backlight off.
1463 -1 -- never invert brightness
1464 0 -- machine default
1465 1 -- force brightness inversion
1468 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1470 ide-core.nodma= [HW] (E)IDE subsystem
1471 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1472 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1473 .cdrom .chs .ignore_cable are additional options
1474 See Documentation/ide/ide.txt.
1476 ide-generic.probe-mask= [HW] (E)IDE subsystem
1478 Probe mask for legacy ISA IDE ports. Depending on
1479 platform up to 6 ports are supported, enabled by
1480 setting corresponding bits in the mask to 1. The
1481 default value is 0x0, which has a special meaning.
1482 On systems that have PCI, it triggers scanning the
1483 PCI bus for the first and the second port, which
1484 are then probed. On systems without PCI the value
1485 of 0x0 enables probing the two first ports as if it
1488 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1489 Claim all unknown PCI IDE storage controllers.
1492 Format: idle=poll, idle=halt, idle=nomwait
1493 Poll forces a polling idle loop that can slightly
1494 improve the performance of waking up a idle CPU, but
1495 will use a lot of power and make the system run hot.
1497 idle=halt: Halt is forced to be used for CPU idle.
1498 In such case C2/C3 won't be used again.
1499 idle=nomwait: Disable mwait for CPU C-states
1501 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1502 Format: { strict | legacy | 2008 | relaxed }
1505 Choose which programs will be accepted for execution
1506 based on the IEEE 754 NaN encoding(s) supported by
1507 the FPU and the NaN encoding requested with the value
1508 of an ELF file header flag individually set by each
1509 binary. Hardware implementations are permitted to
1510 support either or both of the legacy and the 2008 NaN
1513 Available settings are as follows:
1514 strict accept binaries that request a NaN encoding
1515 supported by the FPU
1516 legacy only accept legacy-NaN binaries, if supported
1518 2008 only accept 2008-NaN binaries, if supported
1520 relaxed accept any binaries regardless of whether
1521 supported by the FPU
1523 The FPU emulator is always able to support both NaN
1524 encodings, so if no FPU hardware is present or it has
1525 been disabled with 'nofpu', then the settings of
1526 'legacy' and '2008' strap the emulator accordingly,
1527 'relaxed' straps the emulator for both legacy-NaN and
1528 2008-NaN, whereas 'strict' enables legacy-NaN only on
1529 legacy processors and both NaN encodings on MIPS32 or
1532 The setting for ABS.fmt/NEG.fmt instruction execution
1533 mode generally follows that for the NaN encoding,
1534 except where unsupported by hardware.
1536 ignore_loglevel [KNL]
1537 Ignore loglevel setting - this will print /all/
1538 kernel messages to the console. Useful for debugging.
1539 We also add it as printk module parameter, so users
1540 could change it dynamically, usually by
1541 /sys/module/printk/parameters/ignore_loglevel.
1544 Ignore RLIMIT_DATA setting for data mappings,
1545 print warning at first misuse. Can be changed via
1546 /sys/module/kernel/parameters/ignore_rlimit_data.
1548 ihash_entries= [KNL]
1549 Set number of hash buckets for inode cache.
1551 ima_appraise= [IMA] appraise integrity measurements
1552 Format: { "off" | "enforce" | "fix" | "log" }
1555 ima_appraise_tcb [IMA]
1556 The builtin appraise policy appraises all files
1560 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1564 The list of supported hash algorithms is defined
1565 in crypto/hash_info.h.
1568 The builtin measurement policy to load during IMA
1569 setup. Specyfing "tcb" as the value, measures all
1570 programs exec'd, files mmap'd for exec, and all files
1571 opened with the read mode bit set by either the
1572 effective uid (euid=0) or uid=0.
1575 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1576 Load a policy which meets the needs of the Trusted
1577 Computing Base. This means IMA will measure all
1578 programs exec'd, files mmap'd for exec, and all files
1579 opened for read by uid=0.
1582 Select one of defined IMA measurements template formats.
1583 Formats: { "ima" | "ima-ng" | "ima-sig" }
1587 [IMA] Define a custom template format.
1588 Format: { "field1|...|fieldN" }
1590 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1591 Format: <min_file_size>
1592 Set the minimal file size for using asynchronous hash.
1593 If left unspecified, ahash usage is disabled.
1595 ahash performance varies for different data sizes on
1596 different crypto accelerators. This option can be used
1597 to achieve the best performance for a particular HW.
1599 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1601 Set hashing buffer size. Default: 4k.
1603 ahash performance varies for different chunk sizes on
1604 different crypto accelerators. This option can be used
1605 to achieve best performance for particular HW.
1609 Run specified binary instead of /sbin/init as init
1612 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1613 for working out where the kernel is dying during
1616 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1617 initcall functions. Useful for debugging built-in
1618 modules and initcalls.
1620 initrd= [BOOT] Specify the location of the initial ramdisk
1622 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1625 int_pln_enable [x86] Enable power limit notification interrupt
1627 integrity_audit=[IMA]
1628 Format: { "0" | "1" }
1629 0 -- basic integrity auditing messages. (Default)
1630 1 -- additional integrity auditing messages.
1632 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1634 Enable intel iommu driver.
1636 Disable intel iommu driver.
1637 igfx_off [Default Off]
1638 By default, gfx is mapped as normal device. If a gfx
1639 device has a dedicated DMAR unit, the DMAR unit is
1640 bypassed by not enabling DMAR with this option. In
1641 this case, gfx device will use physical address for
1644 With this option iommu will not optimize to look
1645 for io virtual address below 32-bit forcing dual
1646 address cycle on pci bus for cards supporting greater
1647 than 32-bit addressing. The default is to look
1648 for translation below 32-bit and if not available
1649 then look in the higher range.
1650 strict [Default Off]
1651 With this option on every unmap_single operation will
1652 result in a hardware IOTLB flush operation as opposed
1653 to batching them for performance.
1654 sp_off [Default Off]
1655 By default, super page will be supported if Intel IOMMU
1656 has the capability. With this option, super page will
1658 ecs_off [Default Off]
1659 By default, extended context tables will be supported if
1660 the hardware advertises that it has support both for the
1661 extended tables themselves, and also PASID support. With
1662 this option set, extended tables will not be used even
1663 on hardware which claims to support them.
1665 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1666 0 disables intel_idle and fall back on acpi_idle.
1667 1 to 6 specify maximum depth of C-state.
1671 Do not enable intel_pstate as the default
1672 scaling driver for the supported processors
1674 Enable intel_pstate on systems that prohibit it by default
1675 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1676 instead of acpi-cpufreq may disable platform features, such
1677 as thermal controls and power capping, that rely on ACPI
1678 P-States information being indicated to OSPM and therefore
1679 should be used with caution. This option does not work with
1680 processors that aren't supported by the intel_pstate driver
1681 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1683 Do not enable hardware P state control (HWP)
1686 Only load intel_pstate on systems which support
1687 hardware P state control (HWP) if available.
1689 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1690 Description Table, specifies preferred power management
1691 profile as "Enterprise Server" or "Performance Server",
1692 then this feature is turned on by default.
1694 intremap= [X86-64, Intel-IOMMU]
1695 on enable Interrupt Remapping (default)
1696 off disable Interrupt Remapping
1697 nosid disable Source ID checking
1699 BIOS x2APIC opt-out request will be ignored
1700 nopost disable Interrupt Posting
1702 iomem= Disable strict checking of access to MMIO memory
1703 strict regions from userspace.
1718 nobypass [PPC/POWERNV]
1719 Disable IOMMU bypass, using IOMMU for PCI devices.
1722 io7= [HW] IO7 for Marvel based alpha systems
1723 See comment before marvel_specify_io7 in
1724 arch/alpha/kernel/core_marvel.c.
1726 io_delay= [X86] I/O delay method
1728 Standard port 0x80 based delay
1730 Alternate port 0xed based delay (needed on some systems)
1732 Simple two microseconds delay
1737 See Documentation/filesystems/nfs/nfsroot.txt.
1739 irqaffinity= [SMP] Set the default irq affinity mask
1741 <cpu number>,...,<cpu number>
1743 <cpu number>-<cpu number>
1744 (must be a positive range in ascending order)
1746 <cpu number>,...,<cpu number>-<cpu number>
1749 When an interrupt is not handled search all handlers
1750 for it. Intended to get systems with badly broken
1754 When an interrupt is not handled search all handlers
1755 for it. Also check all handlers each timer
1756 interrupt. Intended to get systems with badly broken
1760 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1762 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1764 <cpu number>,...,<cpu number>
1766 <cpu number>-<cpu number>
1767 (must be a positive range in ascending order)
1769 <cpu number>,...,<cpu number>-<cpu number>
1771 This option can be used to specify one or more CPUs
1772 to isolate from the general SMP balancing and scheduling
1773 algorithms. You can move a process onto or off an
1774 "isolated" CPU via the CPU affinity syscalls or cpuset.
1775 <cpu number> begins at 0 and the maximum value is
1776 "number of CPUs in system - 1".
1778 This option is the preferred way to isolate CPUs. The
1779 alternative -- manually setting the CPU mask of all
1780 tasks in the system -- can cause problems and
1781 suboptimal load balancer performance.
1785 ivrs_ioapic [HW,X86_64]
1786 Provide an override to the IOAPIC-ID<->DEVICE-ID
1787 mapping provided in the IVRS ACPI table. For
1788 example, to map IOAPIC-ID decimal 10 to
1789 PCI device 00:14.0 write the parameter as:
1790 ivrs_ioapic[10]=00:14.0
1792 ivrs_hpet [HW,X86_64]
1793 Provide an override to the HPET-ID<->DEVICE-ID
1794 mapping provided in the IVRS ACPI table. For
1795 example, to map HPET-ID decimal 0 to
1796 PCI device 00:14.0 write the parameter as:
1797 ivrs_hpet[0]=00:14.0
1799 ivrs_acpihid [HW,X86_64]
1800 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1801 mapping provided in the IVRS ACPI table. For
1802 example, to map UART-HID:UID AMD0020:0 to
1803 PCI device 00:14.5 write the parameter as:
1804 ivrs_acpihid[00:14.5]=AMD0020:0
1806 js= [HW,JOY] Analog joystick
1807 See Documentation/input/joystick.txt.
1810 Enable/disable kernel and module base offset ASLR
1811 (Address Space Layout Randomization) if built into
1812 the kernel. When CONFIG_HIBERNATION is selected,
1813 kASLR is disabled by default. When kASLR is enabled,
1814 hibernation will be disabled.
1818 kernelcore= [KNL,X86,IA-64,PPC]
1819 Format: nn[KMGTPE] | "mirror"
1821 specifies the amount of memory usable by the kernel
1822 for non-movable allocations. The requested amount is
1823 spread evenly throughout all nodes in the system. The
1824 remaining memory in each node is used for Movable
1825 pages. In the event, a node is too small to have both
1826 kernelcore and Movable pages, kernelcore pages will
1827 take priority and other nodes will have a larger number
1828 of Movable pages. The Movable zone is used for the
1829 allocation of pages that may be reclaimed or moved
1830 by the page migration subsystem. This means that
1831 HugeTLB pages may not be allocated from this zone.
1832 Note that allocations like PTEs-from-HighMem still
1833 use the HighMem zone if it exists, and the Normal
1834 zone if it does not.
1836 Instead of specifying the amount of memory (nn[KMGTPE]),
1837 you can specify "mirror" option. In case "mirror"
1838 option is specified, mirrored (reliable) memory is used
1839 for non-movable allocations and remaining memory is used
1840 for Movable pages. nn[KMGTPE] and "mirror" are exclusive,
1841 so you can NOT specify nn[KMGTPE] and "mirror" at the same
1844 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1845 Format: <Controller#>[,poll interval]
1846 The controller # is the number of the ehci usb debug
1847 port as it is probed via PCI. The poll interval is
1848 optional and is the number seconds in between
1849 each poll cycle to the debug port in case you need
1850 the functionality for interrupting the kernel with
1851 gdb or control-c on the dbgp connection. When
1852 not using this parameter you use sysrq-g to break into
1853 the kernel debugger.
1855 kgdboc= [KGDB,HW] kgdb over consoles.
1856 Requires a tty driver that supports console polling,
1857 or a supported polling keyboard driver (non-usb).
1858 Serial only format: <serial_device>[,baud]
1859 keyboard only format: kbd
1860 keyboard and serial format: kbd,<serial_device>[,baud]
1861 Optional Kernel mode setting:
1862 kms, kbd format: kms,kbd
1863 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1865 kgdbwait [KGDB] Stop kernel execution and enter the
1866 kernel debugger at the earliest opportunity.
1868 kmac= [MIPS] korina ethernet MAC address.
1869 Configure the RouterBoard 532 series on-chip
1870 Ethernet adapter MAC address.
1872 kmemleak= [KNL] Boot-time kmemleak enable/disable
1873 Valid arguments: on, off
1875 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1878 kmemcheck= [X86] Boot-time kmemcheck enable/disable/one-shot mode
1879 Valid arguments: 0, 1, 2
1880 kmemcheck=0 (disabled)
1881 kmemcheck=1 (enabled)
1882 kmemcheck=2 (one-shot mode)
1883 Default: 2 (one-shot mode)
1885 kstack=N [X86] Print N words from the kernel stack
1888 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1889 Default is 0 (don't ignore, but inject #GP)
1891 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1895 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1896 Default is 1 (enabled)
1898 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1900 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1902 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1903 (virtualized MMU) support on capable Intel chips.
1904 Default is 1 (enabled)
1906 kvm-intel.emulate_invalid_guest_state=
1907 [KVM,Intel] Enable emulation of invalid guest states
1908 Default is 0 (disabled)
1910 kvm-intel.flexpriority=
1911 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1912 Default is 1 (enabled)
1915 [KVM,Intel] Enable VMX nesting (nVMX).
1916 Default is 0 (disabled)
1918 kvm-intel.unrestricted_guest=
1919 [KVM,Intel] Disable unrestricted guest feature
1920 (virtualized real and unpaged mode) on capable
1921 Intel chips. Default is 1 (enabled)
1923 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1924 feature (tagged TLBs) on capable Intel chips.
1925 Default is 1 (enabled)
1931 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1934 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1935 value for LAPIC timer one-shot implementation. Default
1936 back to the programmable timer unit in the LAPIC.
1938 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1941 libata.dma= [LIBATA] DMA control
1942 libata.dma=0 Disable all PATA and SATA DMA
1943 libata.dma=1 PATA and SATA Disk DMA only
1944 libata.dma=2 ATAPI (CDROM) DMA only
1945 libata.dma=4 Compact Flash DMA only
1946 Combinations also work, so libata.dma=3 enables DMA
1947 for disks and CDROMs, but not CFs.
1949 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1950 libata.ignore_hpa=0 keep BIOS limits (default)
1951 libata.ignore_hpa=1 ignore limits, using full disk
1953 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
1957 libata.force= [LIBATA] Force configurations. The format is comma
1958 separated list of "[ID:]VAL" where ID is
1959 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
1960 matching port, link or device. Basically, it matches
1961 the ATA ID string printed on console by libata. If
1962 the whole ID part is omitted, the last PORT and DEVICE
1963 values are used. If ID hasn't been specified yet, the
1964 configuration applies to all ports, links and devices.
1966 If only DEVICE is omitted, the parameter applies to
1967 the port and all links and devices behind it. DEVICE
1968 number of 0 either selects the first device or the
1969 first fan-out link behind PMP device. It does not
1970 select the host link. DEVICE number of 15 selects the
1971 host link and device attached to it.
1973 The VAL specifies the configuration to force. As long
1974 as there's no ambiguity shortcut notation is allowed.
1975 For example, both 1.5 and 1.5G would work for 1.5Gbps.
1976 The following configurations can be forced.
1978 * Cable type: 40c, 80c, short40c, unk, ign or sata.
1979 Any ID with matching PORT is used.
1981 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
1983 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
1984 udma[/][16,25,33,44,66,100,133] notation is also
1987 * [no]ncq: Turn on or off NCQ.
1989 * [no]ncqtrim: Turn off queued DSM TRIM.
1991 * nohrst, nosrst, norst: suppress hard, soft
1994 * rstonce: only attempt one reset during
1995 hot-unplug link recovery
1997 * dump_id: dump IDENTIFY data.
1999 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2001 * disable: Disable this device.
2003 If there are multiple matching configurations changing
2004 the same attribute, the last one is used.
2006 memblock=debug [KNL] Enable memblock debug messages.
2008 load_ramdisk= [RAM] List of ramdisks to load from floppy
2009 See Documentation/blockdev/ramdisk.txt.
2011 lockd.nlm_grace_period=P [NFS] Assign grace period.
2014 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2017 lockd.nlm_timeout=T [NFS] Assign timeout value.
2020 lockd.nlm_udpport=M [NFS] Assign UDP port.
2023 locktorture.nreaders_stress= [KNL]
2024 Set the number of locking read-acquisition kthreads.
2025 Defaults to being automatically set based on the
2026 number of online CPUs.
2028 locktorture.nwriters_stress= [KNL]
2029 Set the number of locking write-acquisition kthreads.
2031 locktorture.onoff_holdoff= [KNL]
2032 Set time (s) after boot for CPU-hotplug testing.
2034 locktorture.onoff_interval= [KNL]
2035 Set time (s) between CPU-hotplug operations, or
2036 zero to disable CPU-hotplug testing.
2038 locktorture.shuffle_interval= [KNL]
2039 Set task-shuffle interval (jiffies). Shuffling
2040 tasks allows some CPUs to go into dyntick-idle
2041 mode during the locktorture test.
2043 locktorture.shutdown_secs= [KNL]
2044 Set time (s) after boot system shutdown. This
2045 is useful for hands-off automated testing.
2047 locktorture.stat_interval= [KNL]
2048 Time (s) between statistics printk()s.
2050 locktorture.stutter= [KNL]
2051 Time (s) to stutter testing, for example,
2052 specifying five seconds causes the test to run for
2053 five seconds, wait for five seconds, and so on.
2054 This tests the locking primitive's ability to
2055 transition abruptly to and from idle.
2057 locktorture.torture_runnable= [BOOT]
2058 Start locktorture running at boot time.
2060 locktorture.torture_type= [KNL]
2061 Specify the locking implementation to test.
2063 locktorture.verbose= [KNL]
2064 Enable additional printk() statements.
2066 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2069 loglevel= All Kernel Messages with a loglevel smaller than the
2070 console loglevel will be printed to the console. It can
2071 also be changed with klogd or other programs. The
2072 loglevels are defined as follows:
2074 0 (KERN_EMERG) system is unusable
2075 1 (KERN_ALERT) action must be taken immediately
2076 2 (KERN_CRIT) critical conditions
2077 3 (KERN_ERR) error conditions
2078 4 (KERN_WARNING) warning conditions
2079 5 (KERN_NOTICE) normal but significant condition
2080 6 (KERN_INFO) informational
2081 7 (KERN_DEBUG) debug-level messages
2083 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2084 in bytes. n must be a power of two and greater
2085 than the minimal size. The minimal size is defined
2086 by LOG_BUF_SHIFT kernel config parameter. There is
2087 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2088 that allows to increase the default size depending on
2089 the number of CPUs. See init/Kconfig for more details.
2091 logo.nologo [FB] Disables display of the built-in Linux logo.
2092 This may be used to provide more screen space for
2093 kernel log messages and is useful when debugging
2094 kernel boot problems.
2096 lp=0 [LP] Specify parallel ports to use, e.g,
2097 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2098 lp=reset first parallel port). 'lp=0' disables the
2099 lp=auto printer driver. 'lp=reset' (which can be
2100 specified in addition to the ports) causes
2101 attached printers to be reset. Using
2102 lp=port1,port2,... specifies the parallel ports
2103 to associate lp devices with, starting with
2104 lp0. A port specification may be 'none' to skip
2105 that lp device, or a parport name such as
2106 'parport0'. Specifying 'lp=auto' instead of a
2107 port specification list means that device IDs
2108 from each port should be examined, to see if
2109 an IEEE 1284-compliant printer is attached; if
2110 so, the driver will manage that printer.
2111 See also header of drivers/char/lp.c.
2114 Sets loops_per_jiffy to given constant, thus avoiding
2115 time-consuming boot-time autodetection (up to 250 ms per
2116 CPU). 0 enables autodetection (default). To determine
2117 the correct value for your kernel, boot with normal
2118 autodetection and see what value is printed. Note that
2119 on SMP systems the preset will be applied to all CPUs,
2120 which is likely to cause problems if your CPUs need
2121 significantly divergent settings. An incorrect value
2122 will cause delays in the kernel to be wrong, leading to
2123 unpredictable I/O errors and other breakage. Although
2124 unlikely, in the extreme case this might damage your
2128 Format: <io>,<irq>,<dma>
2130 machvec= [IA-64] Force the use of a particular machine-vector
2131 (machvec) in a generic kernel.
2132 Example: machvec=hpzx1_swiotlb
2134 machtype= [Loongson] Share the same kernel image file between different
2136 Example: machtype=lemote-yeeloong-2f-7inch
2138 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2139 than or equal to this physical address is ignored.
2141 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2142 should make use of. maxcpus=n : n >= 0 limits the
2143 kernel to using 'n' processors. n=0 is a special case,
2144 it is equivalent to "nosmp", which also disables
2147 max_loop= [LOOP] The number of loop block devices that get
2148 (loop.max_loop) unconditionally pre-created at init time. The default
2149 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2150 of statically allocating a predefined number, loop
2151 devices can be requested on-demand with the
2152 /dev/loop-control interface.
2154 mce [X86-32] Machine Check Exception
2156 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2158 md= [HW] RAID subsystems devices and level
2159 See Documentation/md.txt.
2162 Format: <first>,<last>
2163 Specifies range of consoles to be captured by the MDA.
2165 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2166 Amount of memory to be used when the kernel is not able
2167 to see the whole system memory or for test.
2168 [X86] Work as limiting max address. Use together
2169 with memmap= to avoid physical address space collisions.
2170 Without memmap= PCI devices could be placed at addresses
2171 belonging to unused RAM.
2173 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2177 [KNL,SH] Allow user to override the default size for
2178 per-device physically contiguous DMA buffers.
2180 memhp_default_state=online/offline
2181 [KNL] Set the initial state for the memory hotplug
2182 onlining policy. If not specified, the default value is
2183 set according to the
2184 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2186 See Documentation/memory-hotplug.txt.
2188 memmap=exactmap [KNL,X86] Enable setting of an exact
2189 E820 memory map, as specified by the user.
2190 Such memmap=exactmap lines can be constructed based on
2191 BIOS output or other requirements. See the memmap=nn@ss
2194 memmap=nn[KMG]@ss[KMG]
2195 [KNL] Force usage of a specific region of memory.
2196 Region of memory to be used is from ss to ss+nn.
2198 memmap=nn[KMG]#ss[KMG]
2199 [KNL,ACPI] Mark specific memory as ACPI data.
2200 Region of memory to be marked is from ss to ss+nn.
2202 memmap=nn[KMG]$ss[KMG]
2203 [KNL,ACPI] Mark specific memory as reserved.
2204 Region of memory to be reserved is from ss to ss+nn.
2205 Example: Exclude memory from 0x18690000-0x1869ffff
2206 memmap=64K$0x18690000
2208 memmap=0x10000$0x18690000
2210 memmap=nn[KMG]!ss[KMG]
2211 [KNL,X86] Mark specific memory as protected.
2212 Region of memory to be used, from ss to ss+nn.
2213 The memory region may be marked as e820 type 12 (0xc)
2214 and is NVDIMM or ADR memory.
2216 memory_corruption_check=0/1 [X86]
2217 Some BIOSes seem to corrupt the first 64k of
2218 memory when doing things like suspend/resume.
2219 Setting this option will scan the memory
2220 looking for corruption. Enabling this will
2221 both detect corruption and prevent the kernel
2222 from using the memory being corrupted.
2223 However, its intended as a diagnostic tool; if
2224 repeatable BIOS-originated corruption always
2225 affects the same memory, you can use memmap=
2226 to prevent the kernel from using that memory.
2228 memory_corruption_check_size=size [X86]
2229 By default it checks for corruption in the low
2230 64k, making this memory unavailable for normal
2231 use. Use this parameter to scan for
2232 corruption in more or less memory.
2234 memory_corruption_check_period=seconds [X86]
2235 By default it checks for corruption every 60
2236 seconds. Use this parameter to check at some
2237 other rate. 0 disables periodic checking.
2239 memtest= [KNL,X86,ARM] Enable memtest
2241 default : 0 <disable>
2242 Specifies the number of memtest passes to be
2243 performed. Each pass selects another test
2244 pattern from a given set of patterns. Memtest
2245 fills the memory with this pattern, validates
2246 memory contents and reserves bad memory
2247 regions that are detected.
2249 meye.*= [HW] Set MotionEye Camera parameters
2250 See Documentation/video4linux/meye.txt.
2252 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2253 Multi-Function General Purpose Timers on AMD Geode
2256 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2257 the BIOS has incorrectly applied a workaround. TinyBIOS
2258 version 0.98 is known to be affected, 0.99 fixes the
2259 problem by letting the user disable the workaround.
2263 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2264 physical address is ignored.
2266 mini2440= [ARM,HW,KNL]
2267 Format:[0..2][b][c][t]
2269 MINI2440 configuration specification:
2270 0 - The attached screen is the 3.5" TFT
2271 1 - The attached screen is the 7" TFT
2272 2 - The VGA Shield is attached (1024x768)
2273 Leaving out the screen size parameter will not load
2274 the TFT driver, and the framebuffer will be left
2276 b - Enable backlight. The TFT backlight pin will be
2277 linked to the kernel VESA blanking code and a GPIO
2278 LED. This parameter is not necessary when using the
2280 c - Enable the s3c camera interface.
2281 t - Reserved for enabling touchscreen support. The
2282 touchscreen support is not enabled in the mainstream
2283 kernel as of 2.6.30, a preliminary port can be found
2284 in the "bleeding edge" mini2440 support kernel at
2285 http://repo.or.cz/w/linux-2.6/mini2440.git
2288 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2289 parameter allows control of the logging verbosity for
2290 the additional memory initialisation checks. A value
2291 of 0 disables mminit logging and a level of 4 will
2292 log everything. Information is printed at KERN_DEBUG
2293 so loglevel=8 may also need to be specified.
2296 [KNL] When CONFIG_MODULE_SIG is set, this means that
2297 modules without (valid) signatures will fail to load.
2298 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2299 is always true, so this option does nothing.
2302 [MOUSE] Maximum time between finger touching and
2303 leaving touchpad surface for touch to be considered
2304 a tap and be reported as a left button click (for
2305 touchpads working in absolute mode only).
2307 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2308 reporting absolute coordinates, such as tablets
2309 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2310 reporting absolute coordinates, such as tablets
2312 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
2313 is similar to kernelcore except it specifies the
2314 amount of memory used for migratable allocations.
2315 If both kernelcore and movablecore is specified,
2316 then kernelcore will be at *least* the specified
2317 value but may be more. If movablecore on its own
2318 is specified, the administrator must be careful
2319 that the amount of memory usable for all allocations
2322 movable_node [KNL,X86] Boot-time switch to enable the effects
2323 of CONFIG_MOVABLE_NODE=y. See mm/Kconfig for details.
2325 MTD_Partition= [MTD]
2326 Format: <name>,<region-number>,<size>,<offset>
2328 MTD_Region= [MTD] Format:
2329 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2332 See drivers/mtd/cmdlinepart.c.
2334 multitce=off [PPC] This parameter disables the use of the pSeries
2335 firmware feature for updating multiple TCE entries
2338 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2340 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2342 boundary - index of last SLC block on Flex-OneNAND.
2343 The remaining blocks are configured as MLC blocks.
2344 lock - Configure if Flex-OneNAND boundary should be locked.
2345 Once locked, the boundary cannot be changed.
2346 1 indicates lock status, 0 indicates unlock status.
2349 ARM/S3C2412 JIVE boot control
2351 See arch/arm/mach-s3c2412/mach-jive.c
2353 mtouchusb.raw_coordinates=
2354 [HW] Make the MicroTouch USB driver use raw coordinates
2355 ('y', default) or cooked coordinates ('n')
2357 mtrr_chunk_size=nn[KMG] [X86]
2358 used for mtrr cleanup. It is largest continuous chunk
2359 that could hold holes aka. UC entries.
2361 mtrr_gran_size=nn[KMG] [X86]
2362 Used for mtrr cleanup. It is granularity of mtrr block.
2364 Large value could prevent small alignment from
2367 mtrr_spare_reg_nr=n [X86]
2369 Range: 0,7 : spare reg number
2371 Used for mtrr cleanup. It is spare mtrr entries number.
2372 Set to 2 or more if your graphical card needs more.
2374 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2376 netdev= [NET] Network devices parameters
2377 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2378 Note that mem_start is often overloaded to mean
2379 something different and driver-specific.
2380 This usage is only documented in each driver source
2384 [NETFILTER] Enable connection tracking flow accounting
2385 0 to disable accounting
2386 1 to enable accounting
2389 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2390 See Documentation/filesystems/nfs/nfsroot.txt.
2392 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2393 See Documentation/filesystems/nfs/nfsroot.txt.
2395 nfsrootdebug [NFS] enable nfsroot debugging messages.
2396 See Documentation/filesystems/nfs/nfsroot.txt.
2398 nfs.callback_tcpport=
2399 [NFS] set the TCP port on which the NFSv4 callback
2400 channel should listen.
2403 [NFS] sets the pathname to the program which is used
2404 to update the NFS client cache entries.
2406 nfs.cache_getent_timeout=
2407 [NFS] sets the timeout after which an attempt to
2408 update a cache entry is deemed to have failed.
2410 nfs.idmap_cache_timeout=
2411 [NFS] set the maximum lifetime for idmapper cache
2415 [NFS] enable 64-bit inode numbers.
2416 If zero, the NFS client will fake up a 32-bit inode
2417 number for the readdir() and stat() syscalls instead
2418 of returning the full 64-bit number.
2419 The default is to return 64-bit inode numbers.
2421 nfs.max_session_slots=
2422 [NFSv4.1] Sets the maximum number of session slots
2423 the client will attempt to negotiate with the server.
2424 This limits the number of simultaneous RPC requests
2425 that the client can send to the NFSv4.1 server.
2426 Note that there is little point in setting this
2427 value higher than the max_tcp_slot_table_limit.
2429 nfs.nfs4_disable_idmapping=
2430 [NFSv4] When set to the default of '1', this option
2431 ensures that both the RPC level authentication
2432 scheme and the NFS level operations agree to use
2433 numeric uids/gids if the mount is using the
2434 'sec=sys' security flavour. In effect it is
2435 disabling idmapping, which can make migration from
2436 legacy NFSv2/v3 systems to NFSv4 easier.
2437 Servers that do not support this mode of operation
2438 will be autodetected by the client, and it will fall
2439 back to using the idmapper.
2440 To turn off this behaviour, set the value to '0'.
2442 [NFS4] Specify an additional fixed unique ident-
2443 ification string that NFSv4 clients can insert into
2444 their nfs_client_id4 string. This is typically a
2445 UUID that is generated at system install time.
2447 nfs.send_implementation_id =
2448 [NFSv4.1] Send client implementation identification
2449 information in exchange_id requests.
2450 If zero, no implementation identification information
2452 The default is to send the implementation identification
2455 nfs.recover_lost_locks =
2456 [NFSv4] Attempt to recover locks that were lost due
2457 to a lease timeout on the server. Please note that
2458 doing this risks data corruption, since there are
2459 no guarantees that the file will remain unchanged
2460 after the locks are lost.
2461 If you want to enable the kernel legacy behaviour of
2462 attempting to recover these locks, then set this
2464 The default parameter value of '0' causes the kernel
2465 not to attempt recovery of lost locks.
2467 nfs4.layoutstats_timer =
2468 [NFSv4.2] Change the rate at which the kernel sends
2469 layoutstats to the pNFS metadata server.
2471 Setting this to value to 0 causes the kernel to use
2472 whatever value is the default set by the layout
2473 driver. A non-zero value sets the minimum interval
2474 in seconds between layoutstats transmissions.
2476 nfsd.nfs4_disable_idmapping=
2477 [NFSv4] When set to the default of '1', the NFSv4
2478 server will return only numeric uids and gids to
2479 clients using auth_sys, and will accept numeric uids
2480 and gids from such clients. This is intended to ease
2481 migration from NFSv2/v3.
2483 objlayoutdriver.osd_login_prog=
2484 [NFS] [OBJLAYOUT] sets the pathname to the program which
2485 is used to automatically discover and login into new
2486 osd-targets. Please see:
2487 Documentation/filesystems/pnfs.txt for more explanations
2489 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
2490 when a NMI is triggered.
2491 Format: [state][,regs][,debounce][,die]
2493 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2494 Format: [panic,][nopanic,][num]
2496 0 - turn hardlockup detector in nmi_watchdog off
2497 1 - turn hardlockup detector in nmi_watchdog on
2498 When panic is specified, panic when an NMI watchdog
2499 timeout occurs (or 'nopanic' to override the opposite
2500 default). To disable both hard and soft lockup detectors,
2501 please see 'nowatchdog'.
2502 This is useful when you use a panic=... timeout and
2503 need the box quickly up again.
2505 netpoll.carrier_timeout=
2506 [NET] Specifies amount of time (in seconds) that
2507 netpoll should wait for a carrier. By default netpoll
2510 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2511 emulation library even if a 387 maths coprocessor
2515 [HW] Never suspend the console
2516 Disable suspending of consoles during suspend and
2517 hibernate operations. Once disabled, debugging
2518 messages can reach various consoles while the rest
2519 of the system is being put to sleep (ie, while
2520 debugging driver suspend/resume hooks). This may
2521 not work reliably with all consoles, but is known
2522 to work with serial and VGA consoles.
2523 To facilitate more flexible debugging, we also add
2524 console_suspend, a printk module parameter to control
2525 it. Users could use console_suspend (usually
2526 /sys/module/printk/parameters/console_suspend) to
2527 turn on/off it dynamically.
2529 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2530 caches in the slab allocator. Saves per-node memory,
2531 but will impact performance.
2535 noapic [SMP,APIC] Tells the kernel to not make use of any
2536 IOAPICs that may be present in the system.
2538 noautogroup Disable scheduler automatic task group creation.
2540 nobats [PPC] Do not use BATs for mapping kernel lowmem
2541 on "Classic" PPC cores.
2545 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2547 nodelayacct [KNL] Disable per-task delay accounting
2549 nodisconnect [HW,SCSI,M68K] Disables SCSI disconnects.
2551 nodsp [SH] Disable hardware DSP at boot time.
2553 noefi Disable EFI runtime services support.
2558 On X86-32 available only on PAE configured kernels.
2559 noexec=on: enable non-executable mappings (default)
2560 noexec=off: disable non-executable mappings
2563 Disable SMAP (Supervisor Mode Access Prevention)
2564 even if it is supported by processor.
2567 Disable SMEP (Supervisor Mode Execution Prevention)
2568 even if it is supported by processor.
2571 This affects only 32-bit executables.
2572 noexec32=on: enable non-executable mappings (default)
2573 read doesn't imply executable mappings
2574 noexec32=off: disable non-executable mappings
2575 read implies executable mappings
2577 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2579 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2580 register save and restore. The kernel will only save
2581 legacy floating-point registers on task switch.
2583 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2585 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2586 Equivalent to smt=1.
2588 noxsave [BUGS=X86] Disables x86 extended register state save
2589 and restore using xsave. The kernel will fallback to
2590 enabling legacy floating-point and sse state.
2592 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2593 register states. The kernel will fall back to use
2594 xsave to save the states. By using this parameter,
2595 performance of saving the states is degraded because
2596 xsave doesn't support modified optimization while
2597 xsaveopt supports it on xsaveopt enabled systems.
2599 noxsaves [X86] Disables xsaves and xrstors used in saving and
2600 restoring x86 extended register state in compacted
2601 form of xsave area. The kernel will fall back to use
2602 xsaveopt and xrstor to save and restore the states
2603 in standard form of xsave area. By using this
2604 parameter, xsave area per process might occupy more
2605 memory on xsaves enabled systems.
2607 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2608 wfi(ARM) instruction doesn't work correctly and not to
2609 use it. This is also useful when using JTAG debugger.
2611 no_file_caps Tells the kernel not to honor file capabilities. The
2612 only way then for a file to be executed with privilege
2613 is to be setuid root or executed by root.
2615 nohalt [IA-64] Tells the kernel not to use the power saving
2616 function PAL_HALT_LIGHT when idle. This increases
2617 power-consumption. On the positive side, it reduces
2618 interrupt wake-up latency, which may improve performance
2619 in certain environments such as networked servers or
2622 nohibernate [HIBERNATION] Disable hibernation and resume.
2624 nohz= [KNL] Boottime enable/disable dynamic ticks
2625 Valid arguments: on, off
2628 nohz_full= [KNL,BOOT]
2629 In kernels built with CONFIG_NO_HZ_FULL=y, set
2630 the specified list of CPUs whose tick will be stopped
2631 whenever possible. The boot CPU will be forced outside
2632 the range to maintain the timekeeping.
2633 The CPUs in this range must also be included in the
2636 noiotrap [SH] Disables trapped I/O port accesses.
2638 noirqdebug [X86-32] Disables the code which attempts to detect and
2639 disable unhandled interrupt sources.
2641 no_timer_check [X86,APIC] Disables the code which tests for
2642 broken timer IRQ sources.
2644 noisapnp [ISAPNP] Disables ISA PnP code.
2646 noinitrd [RAM] Tells the kernel not to load any configured
2649 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2651 [Deprecated - use intremap=off]
2655 noinvpcid [X86] Disable the INVPCID cpu feature.
2657 nojitter [IA-64] Disables jitter checking for ITC timers.
2659 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2661 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2664 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2665 steal time is computed, but won't influence scheduler
2668 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2670 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2672 noltlbs [PPC] Do not use large page/tlb entries for kernel
2673 lowmem mapping on PPC40x and PPC8xx
2675 nomca [IA-64] Disable machine check abort handling
2677 nomce [X86-32] Disable Machine Check Exception
2679 nomfgpt [X86-32] Disable Multi-Function General Purpose
2680 Timer usage (for AMD Geode machines).
2682 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2683 shutdown the other cpus. Instead use the REBOOT_VECTOR
2686 nomodule Disable module load
2688 nopat [X86] Disable PAT (page attribute table extension of
2689 pagetables) support.
2691 norandmaps Don't use address space randomization. Equivalent to
2692 echo 0 > /proc/sys/kernel/randomize_va_space
2694 noreplace-paravirt [X86,IA-64,PV_OPS] Don't patch paravirt_ops
2696 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2697 with UP alternatives
2699 nordrand [X86] Disable kernel use of the RDRAND and
2700 RDSEED instructions even if they are supported
2701 by the processor. RDRAND and RDSEED are still
2702 available to user space applications.
2704 noresume [SWSUSP] Disables resume and restores original swap
2707 no-scroll [VGA] Disables scrollback.
2708 This is required for the Braillex ib80-piezo Braille
2709 reader made by F.H. Papenmeier (Germany).
2713 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2715 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2716 and disable the IO APIC. legacy for "maxcpus=0".
2718 nosoftlockup [KNL] Disable the soft-lockup detector.
2720 nosync [HW,M68K] Disables sync negotiation for all devices.
2722 notsc [BUGS=X86-32] Disable Time Stamp Counter
2724 nowatchdog [KNL] Disable both lockup detectors, i.e.
2725 soft-lockup and NMI watchdog (hard-lockup).
2729 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2731 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2732 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2733 Some features depend on CPU0. Known dependencies are:
2734 1. Resume from suspend/hibernate depends on CPU0.
2735 Suspend/hibernate will fail if CPU0 is offline and you
2736 need to online CPU0 before suspend/hibernate.
2737 2. PIC interrupts also depend on CPU0. CPU0 can't be
2738 removed if a PIC interrupt is detected.
2739 It's said poweroff/reboot may depend on CPU0 on some
2740 machines although I haven't seen such issues so far
2741 after CPU0 is offline on a few tested machines.
2742 If the dependencies are under your control, you can
2743 turn on cpu0_hotplug.
2745 nptcg= [IA-64] Override max number of concurrent global TLB
2746 purges which is reported from either PAL_VM_SUMMARY or
2749 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2750 could support. nr_cpus=n : n >= 1 limits the kernel to
2751 supporting 'n' processors. Later in runtime you can not
2752 use hotplug cpu feature to put more cpu back to online.
2753 just like you compile the kernel NR_CPUS=n
2755 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2757 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2758 Allowed values are enable and disable
2760 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2761 one of ['zone', 'node', 'default'] can be specified
2762 This can be set from sysctl after boot.
2763 See Documentation/sysctl/vm.txt for details.
2765 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2766 See Documentation/debugging-via-ohci1394.txt for more
2769 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2770 Rather than timing out after 20 ms if an EC
2771 command is not properly ACKed, override the length
2772 of the timeout. We have interrupts disabled while
2773 waiting for the ACK, so if this is set too high
2774 interrupts *may* be lost!
2776 omap_mux= [OMAP] Override bootloader pin multiplexing.
2777 Format: <mux_mode0.mode_name=value>...
2778 For example, to override I2C bus2:
2779 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2781 oprofile.timer= [HW]
2782 Use timer interrupt instead of performance counters
2784 oprofile.cpu_type= Force an oprofile cpu type
2785 This might be useful if you have an older oprofile
2786 userland or if you want common events.
2787 Format: { arch_perfmon }
2788 arch_perfmon: [X86] Force use of architectural
2789 perfmon on Intel CPUs instead of the
2790 CPU specific event set.
2791 timer: [X86] Force use of architectural NMI
2792 timer mode (see also oprofile.timer
2793 for generic hr timer mode)
2794 [s390] Force legacy basic mode sampling
2795 (report cpu_type "timer")
2797 oops=panic Always panic on oopses. Default is to just kill the
2798 process, but there is a small probability of
2799 deadlocking the machine.
2800 This will also cause panics on machine check exceptions.
2801 Useful together with panic=30 to trigger a reboot.
2804 See Documentation/sound/oss/oss-parameters.txt
2806 page_owner= [KNL] Boot-time page_owner enabling option.
2807 Storage of the information about who allocated
2808 each page is disabled in default. With this switch,
2810 on: enable the feature
2812 page_poison= [KNL] Boot-time parameter changing the state of
2813 poisoning on the buddy allocator.
2814 off: turn off poisoning
2815 on: turn on poisoning
2817 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2818 timeout > 0: seconds before rebooting
2819 timeout = 0: wait forever
2820 timeout < 0: reboot immediately
2823 panic_on_warn panic() instead of WARN(). Useful to cause kdump
2826 crash_kexec_post_notifiers
2827 Run kdump after running panic-notifiers and dumping
2828 kmsg. This only for the users who doubt kdump always
2829 succeeds in any situation.
2830 Note that this also increases risks of kdump failure,
2831 because some panic notifiers can make the crashed
2832 kernel more unstable.
2834 parkbd.port= [HW] Parallel port number the keyboard adapter is
2835 connected to, default is 0.
2837 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2838 0 for XT, 1 for AT (default is AT).
2841 parport= [HW,PPT] Specify parallel ports. 0 disables.
2842 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2843 Use 'auto' to force the driver to use any
2844 IRQ/DMA settings detected (the default is to
2845 ignore detected IRQ/DMA settings because of
2846 possible conflicts). You can specify the base
2847 address, IRQ, and DMA settings; IRQ and DMA
2848 should be numbers, or 'auto' (for using detected
2849 settings on that particular port), or 'nofifo'
2850 (to avoid using a FIFO even if it is detected).
2851 Parallel ports are assigned in the order they
2852 are specified on the command line, starting
2855 parport_init_mode= [HW,PPT]
2856 Configure VIA parallel port to operate in
2857 a specific mode. This is necessary on Pegasos
2858 computer where firmware has no options for setting
2859 up parallel port mode and sets it to spp.
2860 Currently this function knows 686a and 8231 chips.
2861 Format: [spp|ps2|epp|ecp|ecpepp]
2864 Halt all CPUs after the first oops has been printed for
2865 the specified number of seconds. This is to be used if
2866 your oopses keep scrolling off the screen.
2871 See header of drivers/block/paride/pcd.c.
2872 See also Documentation/blockdev/paride.txt.
2874 pci=option[,option...] [PCI] various PCI subsystem options:
2875 earlydump [X86] dump PCI config space before the kernel
2877 off [X86] don't probe for the PCI bus
2878 bios [X86-32] force use of PCI BIOS, don't access
2879 the hardware directly. Use this if your machine
2880 has a non-standard PCI host bridge.
2881 nobios [X86-32] disallow use of PCI BIOS, only direct
2882 hardware access methods are allowed. Use this
2883 if you experience crashes upon bootup and you
2884 suspect they are caused by the BIOS.
2885 conf1 [X86] Force use of PCI Configuration Access
2886 Mechanism 1 (config address in IO port 0xCF8,
2887 data in IO port 0xCFC, both 32-bit).
2888 conf2 [X86] Force use of PCI Configuration Access
2889 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
2890 the function, IO port 0xCFA, also 8-bit, sets
2891 bus number. The config space is then accessed
2892 through ports 0xC000-0xCFFF).
2893 See http://wiki.osdev.org/PCI for more info
2894 on the configuration access mechanisms.
2895 noaer [PCIE] If the PCIEAER kernel config parameter is
2896 enabled, this kernel boot option can be used to
2897 disable the use of PCIE advanced error reporting.
2898 nodomains [PCI] Disable support for multiple PCI
2899 root domains (aka PCI segments, in ACPI-speak).
2900 nommconf [X86] Disable use of MMCONFIG for PCI
2902 check_enable_amd_mmconf [X86] check for and enable
2903 properly configured MMIO access to PCI
2904 config space on AMD family 10h CPU
2905 nomsi [MSI] If the PCI_MSI kernel config parameter is
2906 enabled, this kernel boot option can be used to
2907 disable the use of MSI interrupts system-wide.
2908 noioapicquirk [APIC] Disable all boot interrupt quirks.
2909 Safety option to keep boot IRQs enabled. This
2910 should never be necessary.
2911 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
2912 primary IO-APIC for bridges that cannot disable
2913 boot IRQs. This fixes a source of spurious IRQs
2914 when the system masks IRQs.
2915 noioapicreroute [APIC] Disable workaround that uses the
2916 boot IRQ equivalent of an IRQ that connects to
2917 a chipset where boot IRQs cannot be disabled.
2918 The opposite of ioapicreroute.
2919 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
2920 routing table. These calls are known to be buggy
2921 on several machines and they hang the machine
2922 when used, but on other computers it's the only
2923 way to get the interrupt routing table. Try
2924 this option if the kernel is unable to allocate
2925 IRQs or discover secondary PCI buses on your
2927 rom [X86] Assign address space to expansion ROMs.
2928 Use with caution as certain devices share
2929 address decoders between ROMs and other
2931 norom [X86] Do not assign address space to
2932 expansion ROMs that do not already have
2933 BIOS assigned address ranges.
2934 nobar [X86] Do not assign address space to the
2935 BARs that weren't assigned by the BIOS.
2936 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
2937 assigned automatically to PCI devices. You can
2938 make the kernel exclude IRQs of your ISA cards
2940 pirqaddr=0xAAAAA [X86] Specify the physical address
2941 of the PIRQ table (normally generated
2942 by the BIOS) if it is outside the
2943 F0000h-100000h range.
2944 lastbus=N [X86] Scan all buses thru bus #N. Can be
2945 useful if the kernel is unable to find your
2946 secondary buses and you want to tell it
2947 explicitly which ones they are.
2948 assign-busses [X86] Always assign all PCI bus
2949 numbers ourselves, overriding
2950 whatever the firmware may have done.
2951 usepirqmask [X86] Honor the possible IRQ mask stored
2952 in the BIOS $PIR table. This is needed on
2953 some systems with broken BIOSes, notably
2954 some HP Pavilion N5400 and Omnibook XE3
2955 notebooks. This will have no effect if ACPI
2956 IRQ routing is enabled.
2957 noacpi [X86] Do not use ACPI for IRQ routing
2958 or for PCI scanning.
2959 use_crs [X86] Use PCI host bridge window information
2960 from ACPI. On BIOSes from 2008 or later, this
2961 is enabled by default. If you need to use this,
2962 please report a bug.
2963 nocrs [X86] Ignore PCI host bridge windows from ACPI.
2964 If you need to use this, please report a bug.
2965 routeirq Do IRQ routing for all PCI devices.
2966 This is normally done in pci_enable_device(),
2967 so this option is a temporary workaround
2968 for broken drivers that don't call it.
2969 skip_isa_align [X86] do not align io start addr, so can
2970 handle more pci cards
2971 noearly [X86] Don't do any early type 1 scanning.
2972 This might help on some broken boards which
2973 machine check when some devices' config space
2974 is read. But various workarounds are disabled
2975 and some IOMMU drivers will not work.
2976 bfsort Sort PCI devices into breadth-first order.
2977 This sorting is done to get a device
2978 order compatible with older (<= 2.4) kernels.
2979 nobfsort Don't sort PCI devices into breadth-first order.
2980 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
2981 tuning and use the BIOS-configured MPS defaults.
2982 pcie_bus_safe Set every device's MPS to the largest value
2983 supported by all devices below the root complex.
2984 pcie_bus_perf Set device MPS to the largest allowable MPS
2985 based on its parent bus. Also set MRRS (Max
2986 Read Request Size) to the largest supported
2987 value (no larger than the MPS that the device
2988 or bus can support) for best performance.
2989 pcie_bus_peer2peer Set every device's MPS to 128B, which
2990 every device is guaranteed to support. This
2991 configuration allows peer-to-peer DMA between
2992 any pair of devices, possibly at the cost of
2993 reduced performance. This also guarantees
2994 that hot-added devices will work.
2995 cbiosize=nn[KMG] The fixed amount of bus space which is
2996 reserved for the CardBus bridge's IO window.
2997 The default value is 256 bytes.
2998 cbmemsize=nn[KMG] The fixed amount of bus space which is
2999 reserved for the CardBus bridge's memory
3000 window. The default value is 64 megabytes.
3003 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
3004 Specifies alignment and device to reassign
3005 aligned memory resources.
3006 If <order of align> is not specified,
3007 PAGE_SIZE is used as alignment.
3008 PCI-PCI bridge can be specified, if resource
3009 windows need to be expanded.
3010 ecrc= Enable/disable PCIe ECRC (transaction layer
3011 end-to-end CRC checking).
3012 bios: Use BIOS/firmware settings. This is the
3016 hpiosize=nn[KMG] The fixed amount of bus space which is
3017 reserved for hotplug bridge's IO window.
3018 Default size is 256 bytes.
3019 hpmemsize=nn[KMG] The fixed amount of bus space which is
3020 reserved for hotplug bridge's memory window.
3021 Default size is 2 megabytes.
3022 realloc= Enable/disable reallocating PCI bridge resources
3023 if allocations done by BIOS are too small to
3024 accommodate resources required by all child
3026 off: Turn realloc off
3028 realloc same as realloc=on
3029 noari do not use PCIe ARI.
3030 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3031 only look for one device below a PCIe downstream
3034 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3037 force Enable ASPM even on devices that claim not to support it.
3038 WARNING: Forcing ASPM on may cause system lockups.
3040 pcie_hp= [PCIE] PCI Express Hotplug driver options:
3041 nomsi Do not use MSI for PCI Express Native Hotplug (this
3042 makes all PCIe ports use INTx for hotplug services).
3044 pcie_ports= [PCIE] PCIe ports handling:
3045 auto Ask the BIOS whether or not to use native PCIe services
3046 associated with PCIe ports (PME, hot-plug, AER). Use
3047 them only if that is allowed by the BIOS.
3048 native Use native PCIe services associated with PCIe ports
3050 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
3053 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3054 nomsi Do not use MSI for native PCIe PME signaling (this makes
3055 all PCIe root ports use INTx for all services).
3057 pcmv= [HW,PCMCIA] BadgePAD 4
3061 Keep all power-domains already enabled by bootloader on,
3062 even if no driver has claimed them. This is useful
3063 for debug and development, but should not be
3064 needed on a platform with proper driver support.
3067 See Documentation/blockdev/paride.txt.
3069 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3072 See arch/parisc/kernel/pdc_chassis.c
3074 percpu_alloc= Select which percpu first chunk allocator to use.
3075 Currently supported values are "embed" and "page".
3076 Archs may support subset or none of the selections.
3077 See comments in mm/percpu.c for details on each
3078 allocator. This parameter is primarily for debugging
3079 and performance comparison.
3082 See Documentation/blockdev/paride.txt.
3085 See Documentation/blockdev/paride.txt.
3087 pirq= [SMP,APIC] Manual mp-table setup
3088 See Documentation/x86/i386/IO-APIC.txt.
3090 plip= [PPT,NET] Parallel port network link
3091 Format: { parport<nr> | timid | 0 }
3092 See also Documentation/parport.txt.
3094 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3095 Override pmtimer IOPort with a hex value.
3099 Enable PNP debug messages (depends on the
3100 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3101 via /sys/module/pnp/parameters/debug. We always show
3102 current resource usage; turning this on also shows
3103 possible settings and some assignment information.
3109 { on | off | curr | res | no-curr | no-res }
3112 [ISAPNP] Exclude IRQs for the autoconfiguration
3115 [ISAPNP] Exclude DMAs for the autoconfiguration
3117 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3118 Ranges are in pairs (I/O port base and size).
3121 [ISAPNP] Exclude memory regions for the
3123 Ranges are in pairs (memory base and size).
3125 ports= [IP_VS_FTP] IPVS ftp helper module
3127 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3129 Format: <port>,<port>....
3131 ppc_strict_facility_enable
3132 [PPC] This option catches any kernel floating point,
3133 Altivec, VSX and SPE outside of regions specifically
3134 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3135 There is some performance impact when enabling this.
3137 print-fatal-signals=
3138 [KNL] debug: print fatal signals
3140 If enabled, warn about various signal handling
3141 related application anomalies: too many signals,
3142 too many POSIX.1 timers, fatal signals causing a
3145 If you hit the warning due to signal overflow,
3146 you might want to try "ulimit -i unlimited".
3150 printk.always_kmsg_dump=
3151 Trigger kmsg_dump for cases other than kernel oops or
3153 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3156 printk.time= Show timing data prefixed to each printk message line
3157 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3159 processor.max_cstate= [HW,ACPI]
3160 Limit processor to maximum C-state
3161 max_cstate=9 overrides any DMI blacklist limit.
3163 processor.nocst [HW,ACPI]
3164 Ignore the _CST method to determine C-states,
3165 instead using the legacy FADT method
3167 profile= [KNL] Enable kernel profiling via /proc/profile
3168 Format: [schedule,]<number>
3169 Param: "schedule" - profile schedule points.
3170 Param: <number> - step/bucket size as a power of 2 for
3171 statistical time based profiling.
3172 Param: "sleep" - profile D-state sleeping (millisecs).
3173 Requires CONFIG_SCHEDSTATS
3174 Param: "kvm" - profile VM exits.
3176 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3178 See Documentation/blockdev/ramdisk.txt.
3180 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3181 probe for; one of (bare|imps|exps|lifebook|any).
3182 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3184 psmouse.resetafter= [HW,MOUSE]
3185 Try to reset the device after so many bad packets
3188 [HW,MOUSE] Set desired mouse resolution, in dpi.
3189 psmouse.smartscroll=
3190 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3191 0 = disabled, 1 = enabled (default).
3193 pstore.backend= Specify the name of the pstore backend to use
3196 See Documentation/blockdev/paride.txt.
3199 [KNL] Number of legacy pty's. Overwrites compiled-in
3202 quiet [KNL] Disable most log messages
3207 See Documentation/md.txt.
3209 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3210 See Documentation/blockdev/ramdisk.txt.
3213 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3214 the specified list of CPUs to be no-callback CPUs.
3215 Invocation of these CPUs' RCU callbacks will
3216 be offloaded to "rcuox/N" kthreads created for
3217 that purpose, where "x" is "b" for RCU-bh, "p"
3218 for RCU-preempt, and "s" for RCU-sched, and "N"
3219 is the CPU number. This reduces OS jitter on the
3220 offloaded CPUs, which can be useful for HPC and
3221 real-time workloads. It can also improve energy
3222 efficiency for asymmetric multiprocessors.
3225 Rather than requiring that offloaded CPUs
3226 (specified by rcu_nocbs= above) explicitly
3227 awaken the corresponding "rcuoN" kthreads,
3228 make these kthreads poll for callbacks.
3229 This improves the real-time response for the
3230 offloaded CPUs by relieving them of the need to
3231 wake up the corresponding kthread, but degrades
3232 energy efficiency by requiring that the kthreads
3233 periodically wake up to do the polling.
3235 rcutree.blimit= [KNL]
3236 Set maximum number of finished RCU callbacks to
3237 process in one batch.
3239 rcutree.dump_tree= [KNL]
3240 Dump the structure of the rcu_node combining tree
3241 out at early boot. This is used for diagnostic
3242 purposes, to verify correct tree setup.
3244 rcutree.gp_cleanup_delay= [KNL]
3245 Set the number of jiffies to delay each step of
3246 RCU grace-period cleanup. This only has effect
3247 when CONFIG_RCU_TORTURE_TEST_SLOW_CLEANUP is set.
3249 rcutree.gp_init_delay= [KNL]
3250 Set the number of jiffies to delay each step of
3251 RCU grace-period initialization. This only has
3252 effect when CONFIG_RCU_TORTURE_TEST_SLOW_INIT
3255 rcutree.gp_preinit_delay= [KNL]
3256 Set the number of jiffies to delay each step of
3257 RCU grace-period pre-initialization, that is,
3258 the propagation of recent CPU-hotplug changes up
3259 the rcu_node combining tree. This only has effect
3260 when CONFIG_RCU_TORTURE_TEST_SLOW_PREINIT is set.
3262 rcutree.rcu_fanout_exact= [KNL]
3263 Disable autobalancing of the rcu_node combining
3264 tree. This is used by rcutorture, and might
3265 possibly be useful for architectures having high
3266 cache-to-cache transfer latencies.
3268 rcutree.rcu_fanout_leaf= [KNL]
3269 Change the number of CPUs assigned to each
3270 leaf rcu_node structure. Useful for very
3271 large systems, which will choose the value 64,
3272 and for NUMA systems with large remote-access
3273 latencies, which will choose a value aligned
3274 with the appropriate hardware boundaries.
3276 rcutree.jiffies_till_sched_qs= [KNL]
3277 Set required age in jiffies for a
3278 given grace period before RCU starts
3279 soliciting quiescent-state help from
3280 rcu_note_context_switch().
3282 rcutree.jiffies_till_first_fqs= [KNL]
3283 Set delay from grace-period initialization to
3284 first attempt to force quiescent states.
3285 Units are jiffies, minimum value is zero,
3286 and maximum value is HZ.
3288 rcutree.jiffies_till_next_fqs= [KNL]
3289 Set delay between subsequent attempts to force
3290 quiescent states. Units are jiffies, minimum
3291 value is one, and maximum value is HZ.
3293 rcutree.kthread_prio= [KNL,BOOT]
3294 Set the SCHED_FIFO priority of the RCU per-CPU
3295 kthreads (rcuc/N). This value is also used for
3296 the priority of the RCU boost threads (rcub/N)
3297 and for the RCU grace-period kthreads (rcu_bh,
3298 rcu_preempt, and rcu_sched). If RCU_BOOST is
3299 set, valid values are 1-99 and the default is 1
3300 (the least-favored priority). Otherwise, when
3301 RCU_BOOST is not set, valid values are 0-99 and
3302 the default is zero (non-realtime operation).
3304 rcutree.rcu_nocb_leader_stride= [KNL]
3305 Set the number of NOCB kthread groups, which
3306 defaults to the square root of the number of
3307 CPUs. Larger numbers reduces the wakeup overhead
3308 on the per-CPU grace-period kthreads, but increases
3309 that same overhead on each group's leader.
3311 rcutree.qhimark= [KNL]
3312 Set threshold of queued RCU callbacks beyond which
3313 batch limiting is disabled.
3315 rcutree.qlowmark= [KNL]
3316 Set threshold of queued RCU callbacks below which
3317 batch limiting is re-enabled.
3319 rcutree.rcu_idle_gp_delay= [KNL]
3320 Set wakeup interval for idle CPUs that have
3321 RCU callbacks (RCU_FAST_NO_HZ=y).
3323 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3324 Set wakeup interval for idle CPUs that have
3325 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3326 Lazy RCU callbacks are those which RCU can
3327 prove do nothing more than free memory.
3329 rcuperf.gp_exp= [KNL]
3330 Measure performance of expedited synchronous
3331 grace-period primitives.
3333 rcuperf.holdoff= [KNL]
3334 Set test-start holdoff period. The purpose of
3335 this parameter is to delay the start of the
3336 test until boot completes in order to avoid
3339 rcuperf.nreaders= [KNL]
3340 Set number of RCU readers. The value -1 selects
3341 N, where N is the number of CPUs. A value
3342 "n" less than -1 selects N-n+1, where N is again
3343 the number of CPUs. For example, -2 selects N
3344 (the number of CPUs), -3 selects N+1, and so on.
3345 A value of "n" less than or equal to -N selects
3348 rcuperf.nwriters= [KNL]
3349 Set number of RCU writers. The values operate
3350 the same as for rcuperf.nreaders.
3351 N, where N is the number of CPUs
3353 rcuperf.perf_runnable= [BOOT]
3354 Start rcuperf running at boot time.
3356 rcuperf.shutdown= [KNL]
3357 Shut the system down after performance tests
3358 complete. This is useful for hands-off automated
3361 rcuperf.perf_type= [KNL]
3362 Specify the RCU implementation to test.
3364 rcuperf.verbose= [KNL]
3365 Enable additional printk() statements.
3367 rcutorture.cbflood_inter_holdoff= [KNL]
3368 Set holdoff time (jiffies) between successive
3369 callback-flood tests.
3371 rcutorture.cbflood_intra_holdoff= [KNL]
3372 Set holdoff time (jiffies) between successive
3373 bursts of callbacks within a given callback-flood
3376 rcutorture.cbflood_n_burst= [KNL]
3377 Set the number of bursts making up a given
3378 callback-flood test. Set this to zero to
3379 disable callback-flood testing.
3381 rcutorture.cbflood_n_per_burst= [KNL]
3382 Set the number of callbacks to be registered
3383 in a given burst of a callback-flood test.
3385 rcutorture.fqs_duration= [KNL]
3386 Set duration of force_quiescent_state bursts
3389 rcutorture.fqs_holdoff= [KNL]
3390 Set holdoff time within force_quiescent_state bursts
3393 rcutorture.fqs_stutter= [KNL]
3394 Set wait time between force_quiescent_state bursts
3397 rcutorture.gp_cond= [KNL]
3398 Use conditional/asynchronous update-side
3399 primitives, if available.
3401 rcutorture.gp_exp= [KNL]
3402 Use expedited update-side primitives, if available.
3404 rcutorture.gp_normal= [KNL]
3405 Use normal (non-expedited) asynchronous
3406 update-side primitives, if available.
3408 rcutorture.gp_sync= [KNL]
3409 Use normal (non-expedited) synchronous
3410 update-side primitives, if available. If all
3411 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3412 rcutorture.gp_normal=, and rcutorture.gp_sync=
3413 are zero, rcutorture acts as if is interpreted
3414 they are all non-zero.
3416 rcutorture.n_barrier_cbs= [KNL]
3417 Set callbacks/threads for rcu_barrier() testing.
3419 rcutorture.nfakewriters= [KNL]
3420 Set number of concurrent RCU writers. These just
3421 stress RCU, they don't participate in the actual
3422 test, hence the "fake".
3424 rcutorture.nreaders= [KNL]
3425 Set number of RCU readers. The value -1 selects
3426 N-1, where N is the number of CPUs. A value
3427 "n" less than -1 selects N-n-2, where N is again
3428 the number of CPUs. For example, -2 selects N
3429 (the number of CPUs), -3 selects N+1, and so on.
3431 rcutorture.object_debug= [KNL]
3432 Enable debug-object double-call_rcu() testing.
3434 rcutorture.onoff_holdoff= [KNL]
3435 Set time (s) after boot for CPU-hotplug testing.
3437 rcutorture.onoff_interval= [KNL]
3438 Set time (s) between CPU-hotplug operations, or
3439 zero to disable CPU-hotplug testing.
3441 rcutorture.shuffle_interval= [KNL]
3442 Set task-shuffle interval (s). Shuffling tasks
3443 allows some CPUs to go into dyntick-idle mode
3444 during the rcutorture test.
3446 rcutorture.shutdown_secs= [KNL]
3447 Set time (s) after boot system shutdown. This
3448 is useful for hands-off automated testing.
3450 rcutorture.stall_cpu= [KNL]
3451 Duration of CPU stall (s) to test RCU CPU stall
3452 warnings, zero to disable.
3454 rcutorture.stall_cpu_holdoff= [KNL]
3455 Time to wait (s) after boot before inducing stall.
3457 rcutorture.stat_interval= [KNL]
3458 Time (s) between statistics printk()s.
3460 rcutorture.stutter= [KNL]
3461 Time (s) to stutter testing, for example, specifying
3462 five seconds causes the test to run for five seconds,
3463 wait for five seconds, and so on. This tests RCU's
3464 ability to transition abruptly to and from idle.
3466 rcutorture.test_boost= [KNL]
3467 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3468 "Maybe" means test if the RCU implementation
3469 under test support RCU priority boosting.
3471 rcutorture.test_boost_duration= [KNL]
3472 Duration (s) of each individual boost test.
3474 rcutorture.test_boost_interval= [KNL]
3475 Interval (s) between each boost test.
3477 rcutorture.test_no_idle_hz= [KNL]
3478 Test RCU's dyntick-idle handling. See also the
3479 rcutorture.shuffle_interval parameter.
3481 rcutorture.torture_runnable= [BOOT]
3482 Start rcutorture running at boot time.
3484 rcutorture.torture_type= [KNL]
3485 Specify the RCU implementation to test.
3487 rcutorture.verbose= [KNL]
3488 Enable additional printk() statements.
3490 rcupdate.rcu_cpu_stall_suppress= [KNL]
3491 Suppress RCU CPU stall warning messages.
3493 rcupdate.rcu_cpu_stall_timeout= [KNL]
3494 Set timeout for RCU CPU stall warning messages.
3496 rcupdate.rcu_expedited= [KNL]
3497 Use expedited grace-period primitives, for
3498 example, synchronize_rcu_expedited() instead
3499 of synchronize_rcu(). This reduces latency,
3500 but can increase CPU utilization, degrade
3501 real-time latency, and degrade energy efficiency.
3502 No effect on CONFIG_TINY_RCU kernels.
3504 rcupdate.rcu_normal= [KNL]
3505 Use only normal grace-period primitives,
3506 for example, synchronize_rcu() instead of
3507 synchronize_rcu_expedited(). This improves
3508 real-time latency, CPU utilization, and
3509 energy efficiency, but can expose users to
3510 increased grace-period latency. This parameter
3511 overrides rcupdate.rcu_expedited. No effect on
3512 CONFIG_TINY_RCU kernels.
3514 rcupdate.rcu_normal_after_boot= [KNL]
3515 Once boot has completed (that is, after
3516 rcu_end_inkernel_boot() has been invoked), use
3517 only normal grace-period primitives. No effect
3518 on CONFIG_TINY_RCU kernels.
3520 rcupdate.rcu_task_stall_timeout= [KNL]
3521 Set timeout in jiffies for RCU task stall warning
3522 messages. Disable with a value less than or equal
3525 rcupdate.rcu_self_test= [KNL]
3526 Run the RCU early boot self tests
3528 rcupdate.rcu_self_test_bh= [KNL]
3529 Run the RCU bh early boot self tests
3531 rcupdate.rcu_self_test_sched= [KNL]
3532 Run the RCU sched early boot self tests
3536 Run specified binary instead of /init from the ramdisk,
3537 used for early userspace startup. See initrd.
3540 Format (x86 or x86_64):
3541 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3543 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3545 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3546 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3547 reboot_force is either force or not specified,
3548 reboot_cpu is s[mp]#### with #### being the processor
3549 to be used for rebooting.
3552 [KNL, SMP] Set scheduler's default relax_domain_level.
3553 See Documentation/cgroups/cpusets.txt.
3555 relative_sleep_states=
3556 [SUSPEND] Use sleep state labeling where the deepest
3557 state available other than hibernation is always "mem".
3558 Format: { "0" | "1" }
3559 0 -- Traditional sleep state labels.
3560 1 -- Relative sleep state labels.
3562 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
3564 reservetop= [X86-32]
3566 Reserves a hole at the top of the kernel virtual
3571 Set the amount of memory to reserve for BIOS at
3572 the bottom of the address space.
3574 reset_devices [KNL] Force drivers to reset the underlying device
3575 during initialization.
3578 Specify the partition device for software suspend
3580 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3582 resume_offset= [SWSUSP]
3583 Specify the offset from the beginning of the partition
3584 given by "resume=" at which the swap header is located,
3585 in <PAGE_SIZE> units (needed only for swap files).
3586 See Documentation/power/swsusp-and-swap-files.txt
3588 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
3589 read the resume files
3591 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
3592 Useful for devices that are detected asynchronously
3593 (e.g. USB and MMC devices).
3595 hibernate= [HIBERNATION]
3596 noresume Don't check if there's a hibernation image
3597 present during boot.
3598 nocompress Don't compress/decompress hibernation images.
3599 no Disable hibernation and resume.
3601 retain_initrd [RAM] Keep initrd memory after extraction
3603 rfkill.default_state=
3604 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
3605 etc. communication is blocked by default.
3608 rfkill.master_switch_mode=
3609 0 The "airplane mode" button does nothing.
3610 1 The "airplane mode" button toggles between everything
3611 blocked and the previous configuration.
3612 2 The "airplane mode" button toggles between everything
3613 blocked and everything unblocked.
3615 rhash_entries= [KNL,NET]
3616 Set number of hash buckets for route cache
3618 ro [KNL] Mount root device read-only on boot
3621 on Mark read-only kernel memory as read-only (default).
3622 off Leave read-only kernel memory writable for debugging.
3625 Enable the uart passthrough on the designated usb port
3626 on Rockchip SoCs. When active, the signals of the
3627 debug-uart get routed to the D+ and D- pins of the usb
3628 port and the regular usb controller gets disabled.
3630 root= [KNL] Root filesystem
3631 See name_to_dev_t comment in init/do_mounts.c.
3633 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
3634 mount the root filesystem
3636 rootflags= [KNL] Set root filesystem mount option string
3638 rootfstype= [KNL] Set root filesystem type
3640 rootwait [KNL] Wait (indefinitely) for root device to show up.
3641 Useful for devices that are detected asynchronously
3642 (e.g. USB and MMC devices).
3644 rproc_mem=nn[KMG][@address]
3645 [KNL,ARM,CMA] Remoteproc physical memory block.
3646 Memory area to be used by remote processor image,
3649 rw [KNL] Mount root device read-write on boot
3651 S [KNL] Run init in single mode
3653 s390_iommu= [HW,S390]
3654 Set s390 IOTLB flushing mode
3656 With strict flushing every unmap operation will result in
3657 an IOTLB flush. Default is lazy flushing before reuse,
3661 See drivers/net/irda/sa1100_ir.c.
3663 sbni= [NET] Granch SBNI12 leased line adapter
3665 sched_debug [KNL] Enables verbose scheduler debug messages.
3667 schedstats= [KNL,X86] Enable or disable scheduled statistics.
3668 Allowed values are enable and disable. This feature
3669 incurs a small amount of overhead in the scheduler
3670 but is useful for debugging and performance tuning.
3672 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
3673 xtime_lock contention on larger systems, and/or RCU lock
3674 contention on all systems with CONFIG_MAXSMP set.
3675 Format: { "0" | "1" }
3676 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
3678 Note: increases power consumption, thus should only be
3679 enabled if running jitter sensitive (HPC/RT) workloads.
3681 security= [SECURITY] Choose a security module to enable at boot.
3682 If this boot parameter is not specified, only the first
3683 security module asking for security registration will be
3684 loaded. An invalid security module name will be treated
3685 as if no module has been chosen.
3687 selinux= [SELINUX] Disable or enable SELinux at boot time.
3688 Format: { "0" | "1" }
3689 See security/selinux/Kconfig help text.
3692 Default value is set via kernel config option.
3693 If enabled at boot time, /selinux/disable can be used
3694 later to disable prior to initial policy load.
3696 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
3697 Format: { "0" | "1" }
3698 See security/apparmor/Kconfig help text
3701 Default value is set via kernel config option.
3703 serialnumber [BUGS=X86-32]
3706 Maximal number of shapers.
3708 show_msr= [x86] show boot-time MSR settings
3709 Format: { <integer> }
3710 Show boot-time (BIOS-initialized) MSR settings.
3711 The parameter means the number of CPUs to show,
3712 for example 1 means boot CPU only.
3720 Disable merging of slabs with similar size. May be
3721 necessary if there is some reason to distinguish
3722 allocs to different slabs. Debug options disable
3723 merging on their own.
3724 For more information see Documentation/vm/slub.txt.
3726 slab_max_order= [MM, SLAB]
3727 Determines the maximum allowed order for slabs.
3728 A high setting may cause OOMs due to memory
3729 fragmentation. Defaults to 1 for systems with
3730 more than 32MB of RAM, 0 otherwise.
3732 slub_debug[=options[,slabs]] [MM, SLUB]
3733 Enabling slub_debug allows one to determine the
3734 culprit if slab objects become corrupted. Enabling
3735 slub_debug can create guard zones around objects and
3736 may poison objects when not in use. Also tracks the
3737 last alloc / free. For more information see
3738 Documentation/vm/slub.txt.
3740 slub_max_order= [MM, SLUB]
3741 Determines the maximum allowed order for slabs.
3742 A high setting may cause OOMs due to memory
3743 fragmentation. For more information see
3744 Documentation/vm/slub.txt.
3746 slub_min_objects= [MM, SLUB]
3747 The minimum number of objects per slab. SLUB will
3748 increase the slab order up to slub_max_order to
3749 generate a sufficiently large slab able to contain
3750 the number of objects indicated. The higher the number
3751 of objects the smaller the overhead of tracking slabs
3752 and the less frequently locks need to be acquired.
3753 For more information see Documentation/vm/slub.txt.
3755 slub_min_order= [MM, SLUB]
3756 Determines the minimum page order for slabs. Must be
3757 lower than slub_max_order.
3758 For more information see Documentation/vm/slub.txt.
3760 slub_nomerge [MM, SLUB]
3761 Same with slab_nomerge. This is supported for legacy.
3762 See slab_nomerge for more information.
3765 Format: <io1>[,<io2>[,...,<io8>]]
3767 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
3768 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
3769 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
3770 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
3771 smsc-ircc2.ircc_irq= [HW] IRQ line
3772 smsc-ircc2.ircc_dma= [HW] DMA channel
3773 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
3774 0: Toshiba Satellite 1800 (GP data pin select)
3775 1: Fast pin select (default)
3778 smt [KNL,S390] Set the maximum number of threads (logical
3779 CPUs) to use per physical CPU on systems capable of
3780 symmetric multithreading (SMT). Will be capped to the
3781 actual hardware limit.
3783 Default: -1 (no limit)
3786 [KNL] Should the soft-lockup detector generate panics.
3789 softlockup_all_cpu_backtrace=
3790 [KNL] Should the soft-lockup detector generate
3791 backtraces on all cpus.
3794 sonypi.*= [HW] Sony Programmable I/O Control Device driver
3795 See Documentation/laptops/sonypi.txt
3797 spia_io_base= [HW,MTD]
3803 Enabled the stack tracer on boot up.
3805 stacktrace_filter=[function-list]
3806 [FTRACE] Limit the functions that the stack tracer
3807 will trace at boot up. function-list is a comma separated
3808 list of functions. This list can be changed at run
3809 time by the stack_trace_filter file in the debugfs
3810 tracing directory. Note, this enables stack tracing
3811 and the stacktrace above is not needed.
3815 Set the STI (builtin display/keyboard on the HP-PARISC
3816 machines) console (graphic card) which should be used
3817 as the initial boot-console.
3818 See also comment in drivers/video/console/sticore.c.
3821 See comment in drivers/video/console/sticore.c.
3824 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
3826 sunrpc.min_resvport=
3827 sunrpc.max_resvport=
3829 SunRPC servers often require that client requests
3830 originate from a privileged port (i.e. a port in the
3831 range 0 < portnr < 1024).
3832 An administrator who wishes to reserve some of these
3833 ports for other uses may adjust the range that the
3834 kernel's sunrpc client considers to be privileged
3835 using these two parameters to set the minimum and
3836 maximum port values.
3840 Control how the NFS server code allocates CPUs to
3841 service thread pools. Depending on how many NICs
3842 you have and where their interrupts are bound, this
3843 option will affect which CPUs will do NFS serving.
3844 Note: this parameter cannot be changed while the
3845 NFS server is running.
3847 auto the server chooses an appropriate mode
3848 automatically using heuristics
3849 global a single global pool contains all CPUs
3850 percpu one pool for each CPU
3851 pernode one pool for each NUMA node (equivalent
3852 to global on non-NUMA machines)
3854 sunrpc.tcp_slot_table_entries=
3855 sunrpc.udp_slot_table_entries=
3857 Sets the upper limit on the number of simultaneous
3858 RPC calls that can be sent from the client to a
3859 server. Increasing these values may allow you to
3860 improve throughput, but will also increase the
3861 amount of memory reserved for use by the client.
3863 suspend.pm_test_delay=
3865 Sets the number of seconds to remain in a suspend test
3866 mode before resuming the system (see
3867 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
3868 is set. Default value is 5.
3871 [KNL] Enable accounting of swap in memory resource
3872 controller if no parameter or 1 is given or disable
3873 it if 0 is given (See Documentation/cgroups/memory.txt)
3875 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
3876 Format: { <int> | force }
3877 <int> -- Number of I/O TLB slabs
3878 force -- force using of bounce buffers even if they
3879 wouldn't be automatically used by the kernel
3883 sysfs.deprecated=0|1 [KNL]
3884 Enable/disable old style sysfs layout for old udev
3885 on older distributions. When this option is enabled
3886 very new udev will not work anymore. When this option
3887 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
3888 in older udev will not work anymore.
3889 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
3890 the kernel configuration.
3892 sysrq_always_enabled
3894 Ignore sysrq setting - this boot parameter will
3895 neutralize any effect of /proc/sys/kernel/sysrq.
3896 Useful for debugging.
3898 tcpmhash_entries= [KNL,NET]
3899 Set the number of tcp_metrics_hash slots.
3900 Default value is 8192 or 16384 depending on total
3901 ram pages. This is used to specify the TCP metrics
3902 cache size. See Documentation/networking/ip-sysctl.txt
3903 "tcp_no_metrics_save" section for more details.
3907 test_suspend= [SUSPEND][,N]
3908 Specify "mem" (for Suspend-to-RAM) or "standby" (for
3909 standby suspend) or "freeze" (for suspend type freeze)
3910 as the system sleep state during system startup with
3911 the optional capability to repeat N number of times.
3912 The system is woken from this state using a
3913 wakeup-capable RTC alarm.
3915 thash_entries= [KNL,NET]
3916 Set number of hash buckets for TCP connection
3918 thermal.act= [HW,ACPI]
3919 -1: disable all active trip points in all thermal zones
3920 <degrees C>: override all lowest active trip points
3922 thermal.crt= [HW,ACPI]
3923 -1: disable all critical trip points in all thermal zones
3924 <degrees C>: override all critical trip points
3926 thermal.nocrt= [HW,ACPI]
3927 Set to disable actions on ACPI thermal zone
3928 critical and hot trip points.
3930 thermal.off= [HW,ACPI]
3931 1: disable ACPI thermal control
3933 thermal.psv= [HW,ACPI]
3934 -1: disable all passive trip points
3935 <degrees C>: override all passive trip points to this
3938 thermal.tzp= [HW,ACPI]
3939 Specify global default ACPI thermal zone polling rate
3940 <deci-seconds>: poll all this frequency
3941 0: no polling (default)
3944 Force threading of all interrupt handlers except those
3945 marked explicitly IRQF_NO_THREAD.
3948 Enable the Transcendent memory driver if built-in.
3950 tmem.cleancache=0|1 [KNL, XEN]
3951 Default is on (1). Disable the usage of the cleancache
3952 API to send anonymous pages to the hypervisor.
3954 tmem.frontswap=0|1 [KNL, XEN]
3955 Default is on (1). Disable the usage of the frontswap
3956 API to send swap pages to the hypervisor. If disabled
3957 the selfballooning and selfshrinking are force disabled.
3959 tmem.selfballooning=0|1 [KNL, XEN]
3960 Default is on (1). Disable the driving of swap pages
3963 tmem.selfshrinking=0|1 [KNL, XEN]
3964 Default is on (1). Partial swapoff that immediately
3965 transfers pages from Xen hypervisor back to the
3966 kernel based on different criteria.
3970 Specify if the kernel should make use of the cpu
3971 topology information if the hardware supports this.
3972 The scheduler will make use of this information and
3973 e.g. base its process migration decisions on it.
3976 topology_updates= [KNL, PPC, NUMA]
3978 Specify if the kernel should ignore (off)
3979 topology updates sent by the hypervisor to this
3984 tpm_suspend_pcr=[HW,TPM]
3985 Format: integer pcr id
3986 Specify that at suspend time, the tpm driver
3987 should extend the specified pcr with zeros,
3988 as a workaround for some chips which fail to
3989 flush the last written pcr on TPM_SaveState.
3990 This will guarantee that all the other pcrs
3993 trace_buf_size=nn[KMG]
3994 [FTRACE] will set tracing buffer size on each cpu.
3996 trace_event=[event-list]
3997 [FTRACE] Set and start specified trace events in order
3998 to facilitate early boot debugging.
3999 See also Documentation/trace/events.txt
4001 trace_options=[option-list]
4002 [FTRACE] Enable or disable tracer options at boot.
4003 The option-list is a comma delimited list of options
4004 that can be enabled or disabled just as if you were
4005 to echo the option name into
4007 /sys/kernel/debug/tracing/trace_options
4009 For example, to enable stacktrace option (to dump the
4010 stack trace of each event), add to the command line:
4012 trace_options=stacktrace
4014 See also Documentation/trace/ftrace.txt "trace options"
4018 Have the tracepoints sent to printk as well as the
4019 tracing ring buffer. This is useful for early boot up
4020 where the system hangs or reboots and does not give the
4021 option for reading the tracing buffer or performing a
4022 ftrace_dump_on_oops.
4024 To turn off having tracepoints sent to printk,
4025 echo 0 > /proc/sys/kernel/tracepoint_printk
4026 Note, echoing 1 into this file without the
4027 tracepoint_printk kernel cmdline option has no effect.
4031 Having tracepoints sent to printk() and activating high
4032 frequency tracepoints such as irq or sched, can cause
4033 the system to live lock.
4036 [FTRACE] enable this option to disable tracing when a
4037 warning is hit. This turns off "tracing_on". Tracing can
4038 be enabled again by echoing '1' into the "tracing_on"
4039 file located in /sys/kernel/debug/tracing/
4041 This option is useful, as it disables the trace before
4042 the WARNING dump is called, which prevents the trace to
4043 be filled with content caused by the warning output.
4045 This option can also be set at run time via the sysctl
4046 option: kernel/traceoff_on_warning
4048 transparent_hugepage=
4050 Format: [always|madvise|never]
4051 Can be used to control the default behavior of the system
4052 with respect to transparent hugepages.
4053 See Documentation/vm/transhuge.txt for more details.
4055 tsc= Disable clocksource stability checks for TSC.
4057 [x86] reliable: mark tsc clocksource as reliable, this
4058 disables clocksource verification at runtime, as well
4059 as the stability checks done at bootup. Used to enable
4060 high-resolution timer mode on older hardware, and in
4061 virtualized environment.
4062 [x86] noirqtime: Do not use TSC to do irq accounting.
4063 Used to run time disable IRQ_TIME_ACCOUNTING on any
4064 platforms where RDTSC is slow and this accounting
4067 turbografx.map[2|3]= [HW,JOY]
4068 TurboGraFX parallel port interface
4070 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4071 See also Documentation/input/joystick-parport.txt
4073 udbg-immortal [PPC] When debugging early kernel crashes that
4074 happen after console_init() and before a proper
4075 console driver takes over, this boot options might
4076 help "seeing" what's going on.
4078 uhash_entries= [KNL,NET]
4079 Set number of hash buckets for UDP/UDP-Lite connections
4082 [USB] Ignore overcurrent events (default N).
4083 Some badly-designed motherboards generate lots of
4084 bogus events, for ports that aren't wired to
4085 anything. Set this parameter to avoid log spamming.
4086 Note that genuine overcurrent events won't be
4090 [X86] Cause panic on unknown NMI.
4092 usbcore.authorized_default=
4093 [USB] Default USB device authorization:
4094 (default -1 = authorized except for wireless USB,
4095 0 = not authorized, 1 = authorized)
4097 usbcore.autosuspend=
4098 [USB] The autosuspend time delay (in seconds) used
4099 for newly-detected USB devices (default 2). This
4100 is the time required before an idle device will be
4101 autosuspended. Devices for which the delay is set
4102 to a negative value won't be autosuspended at all.
4104 usbcore.usbfs_snoop=
4105 [USB] Set to log all usbfs traffic (default 0 = off).
4107 usbcore.usbfs_snoop_max=
4108 [USB] Maximum number of bytes to snoop in each URB
4111 usbcore.blinkenlights=
4112 [USB] Set to cycle leds on hubs (default 0 = off).
4114 usbcore.old_scheme_first=
4115 [USB] Start with the old device initialization
4116 scheme (default 0 = off).
4118 usbcore.usbfs_memory_mb=
4119 [USB] Memory limit (in MB) for buffers allocated by
4120 usbfs (default = 16, 0 = max = 2047).
4122 usbcore.use_both_schemes=
4123 [USB] Try the other device initialization scheme
4124 if the first one fails (default 1 = enabled).
4126 usbcore.initial_descriptor_timeout=
4127 [USB] Specifies timeout for the initial 64-byte
4128 USB_REQ_GET_DESCRIPTOR request in milliseconds
4129 (default 5000 = 5.0 seconds).
4131 usbcore.nousb [USB] Disable the USB subsystem
4134 [USBHID] The interval which mice are to be polled at.
4136 usb-storage.delay_use=
4137 [UMS] The delay in seconds before a new device is
4138 scanned for Logical Units (default 1).
4141 [UMS] A list of quirks entries to supplement or
4142 override the built-in unusual_devs list. List
4143 entries are separated by commas. Each entry has
4144 the form VID:PID:Flags where VID and PID are Vendor
4145 and Product ID values (4-digit hex numbers) and
4146 Flags is a set of characters, each corresponding
4147 to a common usb-storage quirk flag as follows:
4148 a = SANE_SENSE (collect more than 18 bytes
4150 b = BAD_SENSE (don't collect more than 18
4151 bytes of sense data);
4152 c = FIX_CAPACITY (decrease the reported
4153 device capacity by one sector);
4154 d = NO_READ_DISC_INFO (don't use
4155 READ_DISC_INFO command);
4156 e = NO_READ_CAPACITY_16 (don't use
4157 READ_CAPACITY_16 command);
4158 f = NO_REPORT_OPCODES (don't use report opcodes
4160 g = MAX_SECTORS_240 (don't transfer more than
4161 240 sectors at a time, uas only);
4162 h = CAPACITY_HEURISTICS (decrease the
4163 reported device capacity by one
4164 sector if the number is odd);
4165 i = IGNORE_DEVICE (don't bind to this
4167 j = NO_REPORT_LUNS (don't use report luns
4169 l = NOT_LOCKABLE (don't try to lock and
4170 unlock ejectable media);
4171 m = MAX_SECTORS_64 (don't transfer more
4172 than 64 sectors = 32 KB at a time);
4173 n = INITIAL_READ10 (force a retry of the
4174 initial READ(10) command);
4175 o = CAPACITY_OK (accept the capacity
4176 reported by the device);
4177 p = WRITE_CACHE (the device cache is ON
4179 r = IGNORE_RESIDUE (the device reports
4180 bogus residue values);
4181 s = SINGLE_LUN (the device has only one
4183 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
4184 commands, uas only);
4185 u = IGNORE_UAS (don't bind to the uas driver);
4186 w = NO_WP_DETECT (don't test whether the
4187 medium is write-protected).
4188 Example: quirks=0419:aaf5:rl,0421:0433:rc
4190 user_debug= [KNL,ARM]
4192 See arch/arm/Kconfig.debug help text.
4193 1 - undefined instruction events
4195 4 - invalid data aborts
4198 Example: user_debug=31
4201 [X86] Flags controlling user PTE allocations.
4203 nohigh = do not allocate PTE pages in
4204 HIGHMEM regardless of setting
4208 On X86_32, this is an alias for vdso32=. Otherwise:
4210 vdso=1: enable VDSO (the default)
4211 vdso=0: disable VDSO mapping
4213 vdso32= [X86] Control the 32-bit vDSO
4214 vdso32=1: enable 32-bit VDSO
4215 vdso32=0 or vdso32=2: disable 32-bit VDSO
4217 See the help text for CONFIG_COMPAT_VDSO for more
4218 details. If CONFIG_COMPAT_VDSO is set, the default is
4219 vdso32=0; otherwise, the default is vdso32=1.
4221 For compatibility with older kernels, vdso32=2 is an
4224 Try vdso32=0 if you encounter an error that says:
4225 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
4228 vector=percpu: enable percpu vector domain
4230 video= [FB] Frame buffer configuration
4231 See Documentation/fb/modedb.txt.
4233 video.brightness_switch_enabled= [0,1]
4234 If set to 1, on receiving an ACPI notify event
4235 generated by hotkey, video driver will adjust brightness
4236 level and then send out the event to user space through
4237 the allocated input device; If set to 0, video driver
4238 will only send out the event without touching backlight
4243 [VMMIO] Memory mapped virtio (platform) device.
4245 <size>@<baseaddr>:<irq>[:<id>]
4247 <size> := size (can use standard suffixes
4249 <baseaddr> := physical base address
4250 <irq> := interrupt number (as passed to
4252 <id> := (optional) platform device id
4254 virtio_mmio.device=1K@0x100b0000:48:7
4256 Can be used multiple times for multiple devices.
4258 vga= [BOOT,X86-32] Select a particular video mode
4259 See Documentation/x86/boot.txt and
4260 Documentation/svga.txt.
4261 Use vga=ask for menu.
4262 This is actually a boot loader parameter; the value is
4263 passed to the kernel using a special protocol.
4265 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
4266 size of <nn>. This can be used to increase the
4267 minimum size (128MB on x86). It can also be used to
4268 decrease the size and leave more room for directly
4271 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
4274 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
4277 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
4281 Controls the behavior of vsyscalls (i.e. calls to
4282 fixed addresses of 0xffffffffff600x00 from legacy
4283 code). Most statically-linked binaries and older
4284 versions of glibc use these calls. Because these
4285 functions are at fixed addresses, they make nice
4286 targets for exploits that can control RIP.
4288 emulate [default] Vsyscalls turn into traps and are
4289 emulated reasonably safely.
4291 native Vsyscalls are native syscall instructions.
4292 This is a little bit faster than trapping
4293 and makes a few dynamic recompilers work
4294 better than they would in emulation mode.
4295 It also makes exploits much easier to write.
4297 none Vsyscalls don't work at all. This makes
4298 them quite hard to use for exploits but
4299 might break your system.
4301 vt.color= [VT] Default text color.
4302 Format: 0xYX, X = foreground, Y = background.
4303 Default: 0x07 = light gray on black.
4305 vt.cur_default= [VT] Default cursor shape.
4306 Format: 0xCCBBAA, where AA, BB, and CC are the same as
4307 the parameters of the <Esc>[?A;B;Cc escape sequence;
4308 see VGA-softcursor.txt. Default: 2 = underline.
4310 vt.default_blu= [VT]
4311 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
4312 Change the default blue palette of the console.
4313 This is a 16-member array composed of values
4316 vt.default_grn= [VT]
4317 Format: <green0>,<green1>,<green2>,...,<green15>
4318 Change the default green palette of the console.
4319 This is a 16-member array composed of values
4322 vt.default_red= [VT]
4323 Format: <red0>,<red1>,<red2>,...,<red15>
4324 Change the default red palette of the console.
4325 This is a 16-member array composed of values
4331 Set system-wide default UTF-8 mode for all tty's.
4332 Default is 1, i.e. UTF-8 mode is enabled for all
4333 newly opened terminals.
4335 vt.global_cursor_default=
4338 Set system-wide default for whether a cursor
4339 is shown on new VTs. Default is -1,
4340 i.e. cursors will be created by default unless
4341 overridden by individual drivers. 0 will hide
4342 cursors, 1 will display them.
4344 vt.italic= [VT] Default color for italic text; 0-15.
4347 vt.underline= [VT] Default color for underlined text; 0-15.
4350 watchdog timers [HW,WDT] For information on watchdog timers,
4351 see Documentation/watchdog/watchdog-parameters.txt
4352 or other driver-specific files in the
4353 Documentation/watchdog/ directory.
4355 workqueue.watchdog_thresh=
4356 If CONFIG_WQ_WATCHDOG is configured, workqueue can
4357 warn stall conditions and dump internal state to
4358 help debugging. 0 disables workqueue stall
4359 detection; otherwise, it's the stall threshold
4360 duration in seconds. The default value is 30 and
4361 it can be updated at runtime by writing to the
4362 corresponding sysfs file.
4364 workqueue.disable_numa
4365 By default, all work items queued to unbound
4366 workqueues are affine to the NUMA nodes they're
4367 issued on, which results in better behavior in
4368 general. If NUMA affinity needs to be disabled for
4369 whatever reason, this option can be used. Note
4370 that this also can be controlled per-workqueue for
4371 workqueues visible under /sys/bus/workqueue/.
4373 workqueue.power_efficient
4374 Per-cpu workqueues are generally preferred because
4375 they show better performance thanks to cache
4376 locality; unfortunately, per-cpu workqueues tend to
4377 be more power hungry than unbound workqueues.
4379 Enabling this makes the per-cpu workqueues which
4380 were observed to contribute significantly to power
4381 consumption unbound, leading to measurably lower
4382 power usage at the cost of small performance
4385 The default value of this parameter is determined by
4386 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
4388 workqueue.debug_force_rr_cpu
4389 Workqueue used to implicitly guarantee that work
4390 items queued without explicit CPU specified are put
4391 on the local CPU. This guarantee is no longer true
4392 and while local CPU is still preferred work items
4393 may be put on foreign CPUs. This debug option
4394 forces round-robin CPU selection to flush out
4395 usages which depend on the now broken guarantee.
4396 When enabled, memory and cache locality will be
4399 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
4400 default x2apic cluster mode on platforms
4403 x86_intel_mid_timer= [X86-32,APBT]
4404 Choose timer option for x86 Intel MID platform.
4405 Two valid options are apbt timer only and lapic timer
4406 plus one apbt timer for broadcast timer.
4407 x86_intel_mid_timer=apbt_only | lapic_and_apbt
4409 xen_512gb_limit [KNL,X86-64,XEN]
4410 Restricts the kernel running paravirtualized under Xen
4411 to use only up to 512 GB of RAM. The reason to do so is
4412 crash analysis tools and Xen tools for doing domain
4413 save/restore/migration must be enabled to handle larger
4416 xen_emul_unplug= [HW,X86,XEN]
4417 Unplug Xen emulated devices
4418 Format: [unplug0,][unplug1]
4419 ide-disks -- unplug primary master IDE devices
4420 aux-ide-disks -- unplug non-primary-master IDE devices
4421 nics -- unplug network devices
4422 all -- unplug all emulated devices (NICs and IDE disks)
4423 unnecessary -- unplugging emulated devices is
4424 unnecessary even if the host did not respond to
4426 never -- do not unplug even if version check succeeds
4428 xen_nopvspin [X86,XEN]
4429 Disables the ticketlock slowpath using Xen PV
4433 Disables the PV optimizations forcing the HVM guest to
4434 run as generic HVM guest with no PV drivers.
4436 xirc2ps_cs= [NET,PCMCIA]
4438 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
4440 ______________________________________________________________________
4444 Add more DRM drivers.