| blob | 61b73e42f488cd56c9c38abd269dc8c2a8b78137 |
1 Kernel Parameters
2 ~~~~~~~~~~~~~~~~~
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
26 loadable modules too.
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 cpu lists:
37 ----------
39 Some kernel parameters take a list of CPUs as a value, e.g. isolcpus,
40 nohz_full, irqaffinity, rcu_nocbs. The format of this list is:
42 <cpu number>,...,<cpu number>
44 or
46 <cpu number>-<cpu number>
47 (must be a positive range in ascending order)
49 or a mixture
51 <cpu number>,...,<cpu number>-<cpu number>
53 Note that for the special case of a range one can split the range into equal
54 sized groups and for each group use some amount from the beginning of that
55 group:
57 <cpu number>-cpu number>:<used size>/<group size>
59 For example one can add to the command line following parameter:
61 isolcpus=1,2,10-20,100-2000:2/25
63 where the final item represents CPUs 100,101,125,126,150,151,...
67 This document may not be entirely up to date and comprehensive. The command
68 "modinfo -p ${modulename}" shows a current list of all parameters of a loadable
69 module. Loadable modules, after being loaded into the running kernel, also
70 reveal their parameters in /sys/module/${modulename}/parameters/. Some of these
71 parameters may be changed at runtime by the command
72 "echo -n ${value} > /sys/module/${modulename}/parameters/${parm}".
74 The parameters listed below are only valid if certain kernel build options were
75 enabled and if respective hardware is present. The text in square brackets at
76 the beginning of each description states the restrictions within which a
77 parameter is applicable:
79 ACPI ACPI support is enabled.
80 AGP AGP (Accelerated Graphics Port) is enabled.
81 ALSA ALSA sound support is enabled.
82 APIC APIC support is enabled.
83 APM Advanced Power Management support is enabled.
84 ARM ARM architecture is enabled.
85 AVR32 AVR32 architecture is enabled.
86 AX25 Appropriate AX.25 support is enabled.
87 BLACKFIN Blackfin architecture is enabled.
88 CLK Common clock infrastructure is enabled.
89 CMA Contiguous Memory Area support is enabled.
90 DRM Direct Rendering Management support is enabled.
91 DYNAMIC_DEBUG Build in debug messages and enable them at runtime
92 EDD BIOS Enhanced Disk Drive Services (EDD) is enabled
93 EFI EFI Partitioning (GPT) is enabled
94 EIDE EIDE/ATAPI support is enabled.
95 EVM Extended Verification Module
96 FB The frame buffer device is enabled.
97 FTRACE Function tracing enabled.
98 GCOV GCOV profiling is enabled.
99 HW Appropriate hardware is enabled.
100 IA-64 IA-64 architecture is enabled.
101 IMA Integrity measurement architecture is enabled.
102 IOSCHED More than one I/O scheduler is enabled.
103 IP_PNP IP DHCP, BOOTP, or RARP is enabled.
104 IPV6 IPv6 support is enabled.
105 ISAPNP ISA PnP code is enabled.
106 ISDN Appropriate ISDN support is enabled.
107 JOY Appropriate joystick support is enabled.
108 KGDB Kernel debugger support is enabled.
109 KVM Kernel Virtual Machine support is enabled.
110 LIBATA Libata driver is enabled
111 LP Printer support is enabled.
112 LOOP Loopback device support is enabled.
113 M68k M68k architecture is enabled.
114 These options have more detailed description inside of
115 Documentation/m68k/kernel-options.txt.
116 MDA MDA console support is enabled.
117 MIPS MIPS architecture is enabled.
118 MOUSE Appropriate mouse support is enabled.
119 MSI Message Signaled Interrupts (PCI).
120 MTD MTD (Memory Technology Device) support is enabled.
121 NET Appropriate network support is enabled.
122 NUMA NUMA support is enabled.
123 NFS Appropriate NFS support is enabled.
124 OSS OSS sound support is enabled.
125 PV_OPS A paravirtualized kernel is enabled.
126 PARIDE The ParIDE (parallel port IDE) subsystem is enabled.
127 PARISC The PA-RISC architecture is enabled.
128 PCI PCI bus support is enabled.
129 PCIE PCI Express support is enabled.
130 PCMCIA The PCMCIA subsystem is enabled.
131 PNP Plug & Play support is enabled.
132 PPC PowerPC architecture is enabled.
133 PPT Parallel port support is enabled.
134 PS2 Appropriate PS/2 support is enabled.
135 RAM RAM disk support is enabled.
136 S390 S390 architecture is enabled.
137 SCSI Appropriate SCSI support is enabled.
138 A lot of drivers have their options described inside
139 the Documentation/scsi/ sub-directory.
140 SECURITY Different security models are enabled.
141 SELINUX SELinux support is enabled.
142 APPARMOR AppArmor support is enabled.
143 SERIAL Serial support is enabled.
144 SH SuperH architecture is enabled.
145 SMP The kernel is an SMP kernel.
146 SPARC Sparc architecture is enabled.
147 SWSUSP Software suspend (hibernation) is enabled.
148 SUSPEND System suspend states are enabled.
149 TPM TPM drivers are enabled.
150 TS Appropriate touchscreen support is enabled.
151 UMS USB Mass Storage support is enabled.
152 USB USB support is enabled.
153 USBHID USB Human Interface Device support is enabled.
154 V4L Video For Linux support is enabled.
155 VMMIO Driver for memory mapped virtio devices is enabled.
156 VGA The VGA console has been enabled.
157 VT Virtual terminal support is enabled.
158 WDT Watchdog support is enabled.
159 XT IBM PC/XT MFM hard disk support is enabled.
160 X86-32 X86-32, aka i386 architecture is enabled.
161 X86-64 X86-64 architecture is enabled.
162 More X86-64 boot options can be found in
163 Documentation/x86/x86_64/boot-options.txt .
164 X86 Either 32-bit or 64-bit x86 (same as X86-32+X86-64)
165 X86_UV SGI UV support is enabled.
166 XEN Xen support is enabled
168 In addition, the following text indicates that the option:
170 BUGS= Relates to possible processor bugs on the said processor.
171 KNL Is a kernel start-up parameter.
172 BOOT Is a boot loader parameter.
174 Parameters denoted with BOOT are actually interpreted by the boot
175 loader, and have no meaning to the kernel directly.
176 Do not modify the syntax of boot loader parameters without extreme
177 need or coordination with <Documentation/x86/boot.txt>.
179 There are also arch-specific kernel-parameters not documented here.
180 See for example <Documentation/x86/x86_64/boot-options.txt>.
182 Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
183 a trailing = on the name of any parameter states that that parameter will
184 be entered as an environment variable, whereas its absence indicates that
185 it will appear as a kernel argument readable via /proc/cmdline by programs
186 running once the system is up.
188 The number of kernel parameters is not limited, but the length of the
189 complete command line (parameters including spaces etc.) is limited to
190 a fixed number of characters. This limit depends on the architecture
191 and is between 256 and 4096 characters. It is defined in the file
192 ./include/asm/setup.h as COMMAND_LINE_SIZE.
194 Finally, the [KMG] suffix is commonly described after a number of kernel
195 parameter values. These 'K', 'M', and 'G' letters represent the _binary_
196 multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
197 bytes respectively. Such letter suffixes can also be entirely omitted.
200 acpi= [HW,ACPI,X86,ARM64]
201 Advanced Configuration and Power Interface
202 Format: { force | on | off | strict | noirq | rsdt |
203 copy_dsdt }
204 force -- enable ACPI if default was off
205 on -- enable ACPI but allow fallback to DT [arm64]
206 off -- disable ACPI if default was on
207 noirq -- do not use ACPI for IRQ routing
208 strict -- Be less tolerant of platforms that are not
209 strictly ACPI specification compliant.
210 rsdt -- prefer RSDT over (default) XSDT
211 copy_dsdt -- copy DSDT to memory
212 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force"
213 are available
215 See also Documentation/power/runtime_pm.txt, pci=noacpi
217 acpi_apic_instance= [ACPI, IOAPIC]
218 Format: <int>
219 2: use 2nd APIC table, if available
220 1,0: use 1st APIC table
221 default: 0
223 acpi_backlight= [HW,ACPI]
224 acpi_backlight=vendor
225 acpi_backlight=video
226 If set to vendor, prefer vendor specific driver
227 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
228 of the ACPI video.ko driver.
230 acpi_force_32bit_fadt_addr
231 force FADT to use 32 bit addresses rather than the
232 64 bit X_* addresses. Some firmware have broken 64
233 bit addresses for force ACPI ignore these and use
234 the older legacy 32 bit addresses.
236 acpica_no_return_repair [HW, ACPI]
237 Disable AML predefined validation mechanism
238 This mechanism can repair the evaluation result to make
239 the return objects more ACPI specification compliant.
240 This option is useful for developers to identify the
241 root cause of an AML interpreter issue when the issue
242 has something to do with the repair mechanism.
244 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
245 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
246 Format: <int>
247 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
248 debug output. Bits in debug_layer correspond to a
249 _COMPONENT in an ACPI source file, e.g.,
250 #define _COMPONENT ACPI_PCI_COMPONENT
251 Bits in debug_level correspond to a level in
252 ACPI_DEBUG_PRINT statements, e.g.,
253 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
254 The debug_level mask defaults to "info". See
255 Documentation/acpi/debug.txt for more information about
256 debug layers and levels.
258 Enable processor driver info messages:
259 acpi.debug_layer=0x20000000
260 Enable PCI/PCI interrupt routing info messages:
261 acpi.debug_layer=0x400000
262 Enable AML "Debug" output, i.e., stores to the Debug
263 object while interpreting AML:
264 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
265 Enable all messages related to ACPI hardware:
266 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
268 Some values produce so much output that the system is
269 unusable. The "log_buf_len" parameter may be useful
270 if you need to capture more output.
272 acpi_enforce_resources= [ACPI]
273 { strict | lax | no }
274 Check for resource conflicts between native drivers
275 and ACPI OperationRegions (SystemIO and SystemMemory
276 only). IO ports and memory declared in ACPI might be
277 used by the ACPI subsystem in arbitrary AML code and
278 can interfere with legacy drivers.
279 strict (default): access to resources claimed by ACPI
280 is denied; legacy drivers trying to access reserved
281 resources will fail to bind to device using them.
282 lax: access to resources claimed by ACPI is allowed;
283 legacy drivers trying to access reserved resources
284 will bind successfully but a warning message is logged.
285 no: ACPI OperationRegions are not marked as reserved,
286 no further checks are performed.
288 acpi_force_table_verification [HW,ACPI]
289 Enable table checksum verification during early stage.
290 By default, this is disabled due to x86 early mapping
291 size limitation.
293 acpi_irq_balance [HW,ACPI]
294 ACPI will balance active IRQs
295 default in APIC mode
297 acpi_irq_nobalance [HW,ACPI]
298 ACPI will not move active IRQs (default)
299 default in PIC mode
301 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
302 Format: <irq>,<irq>...
304 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
305 use by PCI
306 Format: <irq>,<irq>...
308 acpi_mask_gpe= [HW,ACPI]
309 Due to the existence of _Lxx/_Exx, some GPEs triggered
310 by unsupported hardware/firmware features can result in
311 GPE floodings that cannot be automatically disabled by
312 the GPE dispatcher.
313 This facility can be used to prevent such uncontrolled
314 GPE floodings.
315 Format: <int>
317 acpi_no_auto_serialize [HW,ACPI]
318 Disable auto-serialization of AML methods
319 AML control methods that contain the opcodes to create
320 named objects will be marked as "Serialized" by the
321 auto-serialization feature.
322 This feature is enabled by default.
323 This option allows to turn off the feature.
325 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
326 kernels.
328 acpi_no_static_ssdt [HW,ACPI]
329 Disable installation of static SSDTs at early boot time
330 By default, SSDTs contained in the RSDT/XSDT will be
331 installed automatically and they will appear under
332 /sys/firmware/acpi/tables.
333 This option turns off this feature.
334 Note that specifying this option does not affect
335 dynamic table installation which will install SSDT
336 tables to /sys/firmware/acpi/tables/dynamic.
338 acpi_rsdp= [ACPI,EFI,KEXEC]
339 Pass the RSDP address to the kernel, mostly used
340 on machines running EFI runtime service to boot the
341 second kernel for kdump.
343 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
344 Format: To spoof as Windows 98: ="Microsoft Windows"
346 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
347 of 2 which is mandated by ACPI 6) as the supported ACPI
348 specification revision (when using this switch, it may
349 be necessary to carry out a cold reboot _twice_ in a
350 row to make it take effect on the platform firmware).
352 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
353 acpi_osi="string1" # add string1
354 acpi_osi="!string2" # remove string2
355 acpi_osi=!* # remove all strings
356 acpi_osi=! # disable all built-in OS vendor
357 strings
358 acpi_osi=!! # enable all built-in OS vendor
359 strings
360 acpi_osi= # disable all strings
362 'acpi_osi=!' can be used in combination with single or
363 multiple 'acpi_osi="string1"' to support specific OS
364 vendor string(s). Note that such command can only
365 affect the default state of the OS vendor strings, thus
366 it cannot affect the default state of the feature group
367 strings and the current state of the OS vendor strings,
368 specifying it multiple times through kernel command line
369 is meaningless. This command is useful when one do not
370 care about the state of the feature group strings which
371 should be controlled by the OSPM.
372 Examples:
373 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
374 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
375 can make '_OSI("Windows 2000")' TRUE.
377 'acpi_osi=' cannot be used in combination with other
378 'acpi_osi=' command lines, the _OSI method will not
379 exist in the ACPI namespace. NOTE that such command can
380 only affect the _OSI support state, thus specifying it
381 multiple times through kernel command line is also
382 meaningless.
383 Examples:
384 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
385 FALSE.
387 'acpi_osi=!*' can be used in combination with single or
388 multiple 'acpi_osi="string1"' to support specific
389 string(s). Note that such command can affect the
390 current state of both the OS vendor strings and the
391 feature group strings, thus specifying it multiple times
392 through kernel command line is meaningful. But it may
393 still not able to affect the final state of a string if
394 there are quirks related to this string. This command
395 is useful when one want to control the state of the
396 feature group strings to debug BIOS issues related to
397 the OSPM features.
398 Examples:
399 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
400 '_OSI("Module Device")' FALSE.
401 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
402 '_OSI("Module Device")' TRUE.
403 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
404 equivalent to
405 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
406 and
407 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
408 they all will make '_OSI("Windows 2000")' TRUE.
410 acpi_pm_good [X86]
411 Override the pmtimer bug detection: force the kernel
412 to assume that this machine's pmtimer latches its value
413 and always returns good values.
415 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
416 Format: { level | edge | high | low }
418 acpi_skip_timer_override [HW,ACPI]
419 Recognize and ignore IRQ0/pin2 Interrupt Override.
420 For broken nForce2 BIOS resulting in XT-PIC timer.
422 acpi_sleep= [HW,ACPI] Sleep options
423 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
424 old_ordering, nonvs, sci_force_enable }
425 See Documentation/power/video.txt for information on
426 s3_bios and s3_mode.
427 s3_beep is for debugging; it makes the PC's speaker beep
428 as soon as the kernel's real-mode entry point is called.
429 s4_nohwsig prevents ACPI hardware signature from being
430 used during resume from hibernation.
431 old_ordering causes the ACPI 1.0 ordering of the _PTS
432 control method, with respect to putting devices into
433 low power states, to be enforced (the ACPI 2.0 ordering
434 of _PTS is used by default).
435 nonvs prevents the kernel from saving/restoring the
436 ACPI NVS memory during suspend/hibernation and resume.
437 sci_force_enable causes the kernel to set SCI_EN directly
438 on resume from S1/S3 (which is against the ACPI spec,
439 but some broken systems don't work without it).
441 acpi_use_timer_override [HW,ACPI]
442 Use timer override. For some broken Nvidia NF5 boards
443 that require a timer override, but don't have HPET
445 add_efi_memmap [EFI; X86] Include EFI memory map in
446 kernel's map of available physical RAM.
448 agp= [AGP]
449 { off | try_unsupported }
450 off: disable AGP support
451 try_unsupported: try to drive unsupported chipsets
452 (may crash computer or cause data corruption)
454 ALSA [HW,ALSA]
455 See Documentation/sound/alsa/alsa-parameters.txt
457 alignment= [KNL,ARM]
458 Allow the default userspace alignment fault handler
459 behaviour to be specified. Bit 0 enables warnings,
460 bit 1 enables fixups, and bit 2 sends a segfault.
462 align_va_addr= [X86-64]
463 Align virtual addresses by clearing slice [14:12] when
464 allocating a VMA at process creation time. This option
465 gives you up to 3% performance improvement on AMD F15h
466 machines (where it is enabled by default) for a
467 CPU-intensive style benchmark, and it can vary highly in
468 a microbenchmark depending on workload and compiler.
470 32: only for 32-bit processes
471 64: only for 64-bit processes
472 on: enable for both 32- and 64-bit processes
473 off: disable for both 32- and 64-bit processes
475 alloc_snapshot [FTRACE]
476 Allocate the ftrace snapshot buffer on boot up when the
477 main buffer is allocated. This is handy if debugging
478 and you need to use tracing_snapshot() on boot up, and
479 do not want to use tracing_snapshot_alloc() as it needs
480 to be done where GFP_KERNEL allocations are allowed.
482 amd_iommu= [HW,X86-64]
483 Pass parameters to the AMD IOMMU driver in the system.
484 Possible values are:
485 fullflush - enable flushing of IO/TLB entries when
486 they are unmapped. Otherwise they are
487 flushed before they will be reused, which
488 is a lot of faster
489 off - do not initialize any AMD IOMMU found in
490 the system
491 force_isolation - Force device isolation for all
492 devices. The IOMMU driver is not
493 allowed anymore to lift isolation
494 requirements as needed. This option
495 does not override iommu=pt
497 amd_iommu_dump= [HW,X86-64]
498 Enable AMD IOMMU driver option to dump the ACPI table
499 for AMD IOMMU. With this option enabled, AMD IOMMU
500 driver will print ACPI tables for AMD IOMMU during
501 IOMMU initialization.
503 amd_iommu_intr= [HW,X86-64]
504 Specifies one of the following AMD IOMMU interrupt
505 remapping modes:
506 legacy - Use legacy interrupt remapping mode.
507 vapic - Use virtual APIC mode, which allows IOMMU
508 to inject interrupts directly into guest.
509 This mode requires kvm-amd.avic=1.
510 (Default when IOMMU HW support is present.)
512 amijoy.map= [HW,JOY] Amiga joystick support
513 Map of devices attached to JOY0DAT and JOY1DAT
514 Format: <a>,<b>
515 See also Documentation/input/joystick.txt
517 analog.map= [HW,JOY] Analog joystick and gamepad support
518 Specifies type or capabilities of an analog joystick
519 connected to one of 16 gameports
520 Format: <type1>,<type2>,..<type16>
522 apc= [HW,SPARC]
523 Power management functions (SPARCstation-4/5 + deriv.)
524 Format: noidle
525 Disable APC CPU standby support. SPARCstation-Fox does
526 not play well with APC CPU idle - disable it if you have
527 APC and your system crashes randomly.
529 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
530 Change the output verbosity whilst booting
531 Format: { quiet (default) | verbose | debug }
532 Change the amount of debugging information output
533 when initialising the APIC and IO-APIC components.
535 apic_extnmi= [APIC,X86] External NMI delivery setting
536 Format: { bsp (default) | all | none }
537 bsp: External NMI is delivered only to CPU 0
538 all: External NMIs are broadcast to all CPUs as a
539 backup of CPU 0
540 none: External NMI is masked for all CPUs. This is
541 useful so that a dump capture kernel won't be
542 shot down by NMI
544 autoconf= [IPV6]
545 See Documentation/networking/ipv6.txt.
547 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
548 Limit apic dumping. The parameter defines the maximal
549 number of local apics being dumped. Also it is possible
550 to set it to "all" by meaning -- no limit here.
551 Format: { 1 (default) | 2 | ... | all }.
552 The parameter valid if only apic=debug or
553 apic=verbose is specified.
554 Example: apic=debug show_lapic=all
556 apm= [APM] Advanced Power Management
557 See header of arch/x86/kernel/apm_32.c.
559 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
560 Format: <io>,<irq>,<nodeID>
562 ataflop= [HW,M68k]
564 atarimouse= [HW,MOUSE] Atari Mouse
566 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
567 EzKey and similar keyboards
569 atkbd.reset= [HW] Reset keyboard during initialization
571 atkbd.set= [HW] Select keyboard code set
572 Format: <int> (2 = AT (default), 3 = PS/2)
574 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
575 keyboards
577 atkbd.softraw= [HW] Choose between synthetic and real raw mode
578 Format: <bool> (0 = real, 1 = synthetic (default))
580 atkbd.softrepeat= [HW]
581 Use software keyboard repeat
583 audit= [KNL] Enable the audit sub-system
584 Format: { "0" | "1" } (0 = disabled, 1 = enabled)
585 0 - kernel audit is disabled and can not be enabled
586 until the next reboot
587 unset - kernel audit is initialized but disabled and
588 will be fully enabled by the userspace auditd.
589 1 - kernel audit is initialized and partially enabled,
590 storing at most audit_backlog_limit messages in
591 RAM until it is fully enabled by the userspace
592 auditd.
593 Default: unset
595 audit_backlog_limit= [KNL] Set the audit queue size limit.
596 Format: <int> (must be >=0)
597 Default: 64
599 bau= [X86_UV] Enable the BAU on SGI UV. The default
600 behavior is to disable the BAU (i.e. bau=0).
601 Format: { "0" | "1" }
602 0 - Disable the BAU.
603 1 - Enable the BAU.
604 unset - Disable the BAU.
606 baycom_epp= [HW,AX25]
607 Format: <io>,<mode>
609 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
610 Format: <io>,<mode>
611 See header of drivers/net/hamradio/baycom_par.c.
613 baycom_ser_fdx= [HW,AX25]
614 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
615 Format: <io>,<irq>,<mode>[,<baud>]
616 See header of drivers/net/hamradio/baycom_ser_fdx.c.
618 baycom_ser_hdx= [HW,AX25]
619 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
620 Format: <io>,<irq>,<mode>
621 See header of drivers/net/hamradio/baycom_ser_hdx.c.
623 blkdevparts= Manual partition parsing of block device(s) for
624 embedded devices based on command line input.
625 See Documentation/block/cmdline-partition.txt
627 boot_delay= Milliseconds to delay each printk during boot.
628 Values larger than 10 seconds (10000) are changed to
629 no delay (0).
630 Format: integer
632 bootmem_debug [KNL] Enable bootmem allocator debug messages.
634 bert_disable [ACPI]
635 Disable BERT OS support on buggy BIOSes.
637 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
638 bttv.radio= Most important insmod options are available as
639 kernel args too.
640 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
641 bttv.tuner=
643 bulk_remove=off [PPC] This parameter disables the use of the pSeries
644 firmware feature for flushing multiple hpte entries
645 at a time.
647 c101= [NET] Moxa C101 synchronous serial card
649 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
650 Sometimes CPU hardware bugs make them report the cache
651 size incorrectly. The kernel will attempt work arounds
652 to fix known problems, but for some CPUs it is not
653 possible to determine what the correct size should be.
654 This option provides an override for these situations.
656 ca_keys= [KEYS] This parameter identifies a specific key(s) on
657 the system trusted keyring to be used for certificate
658 trust validation.
659 format: { id:<keyid> | builtin }
661 cca= [MIPS] Override the kernel pages' cache coherency
662 algorithm. Accepted values range from 0 to 7
663 inclusive. See arch/mips/include/asm/pgtable-bits.h
664 for platform specific values (SB1, Loongson3 and
665 others).
667 ccw_timeout_log [S390]
668 See Documentation/s390/CommonIO for details.
670 cgroup_disable= [KNL] Disable a particular controller
671 Format: {name of the controller(s) to disable}
672 The effects of cgroup_disable=foo are:
673 - foo isn't auto-mounted if you mount all cgroups in
674 a single hierarchy
675 - foo isn't visible as an individually mountable
676 subsystem
677 {Currently only "memory" controller deal with this and
678 cut the overhead, others just disable the usage. So
679 only cgroup_disable=memory is actually worthy}
681 cgroup_no_v1= [KNL] Disable one, multiple, all cgroup controllers in v1
682 Format: { controller[,controller...] | "all" }
683 Like cgroup_disable, but only applies to cgroup v1;
684 the blacklisted controllers remain available in cgroup2.
686 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
687 Format: <string>
688 nosocket -- Disable socket memory accounting.
689 nokmem -- Disable kernel memory accounting.
691 checkreqprot [SELINUX] Set initial checkreqprot flag value.
692 Format: { "0" | "1" }
693 See security/selinux/Kconfig help text.
694 0 -- check protection applied by kernel (includes
695 any implied execute protection).
696 1 -- check protection requested by application.
697 Default value is set via a kernel config option.
698 Value can be changed at runtime via
699 /selinux/checkreqprot.
701 cio_ignore= [S390]
702 See Documentation/s390/CommonIO for details.
703 clk_ignore_unused
704 [CLK]
705 Prevents the clock framework from automatically gating
706 clocks that have not been explicitly enabled by a Linux
707 device driver but are enabled in hardware at reset or
708 by the bootloader/firmware. Note that this does not
709 force such clocks to be always-on nor does it reserve
710 those clocks in any way. This parameter is useful for
711 debug and development, but should not be needed on a
712 platform with proper driver support. For more
713 information, see Documentation/clk.txt.
715 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
716 [Deprecated]
717 Forces specified clocksource (if available) to be used
718 when calculating gettimeofday(). If specified
719 clocksource is not available, it defaults to PIT.
720 Format: { pit | tsc | cyclone | pmtmr }
722 clocksource= Override the default clocksource
723 Format: <string>
724 Override the default clocksource and use the clocksource
725 with the name specified.
726 Some clocksource names to choose from, depending on
727 the platform:
728 [all] jiffies (this is the base, fallback clocksource)
729 [ACPI] acpi_pm
730 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
731 pxa_timer,timer3,32k_counter,timer0_1
732 [AVR32] avr32
733 [X86-32] pit,hpet,tsc;
734 scx200_hrt on Geode; cyclone on IBM x440
735 [MIPS] MIPS
736 [PARISC] cr16
737 [S390] tod
738 [SH] SuperH
739 [SPARC64] tick
740 [X86-64] hpet,tsc
742 clocksource.arm_arch_timer.evtstrm=
743 [ARM,ARM64]
744 Format: <bool>
745 Enable/disable the eventstream feature of the ARM
746 architected timer so that code using WFE-based polling
747 loops can be debugged more effectively on production
748 systems.
750 clocksource.arm_arch_timer.fsl-a008585=
751 [ARM64]
752 Format: <bool>
753 Enable/disable the workaround of Freescale/NXP
754 erratum A-008585. This can be useful for KVM
755 guests, if the guest device tree doesn't show the
756 erratum. If unspecified, the workaround is
757 enabled based on the device tree.
759 clearcpuid=BITNUM [X86]
760 Disable CPUID feature X for the kernel. See
761 arch/x86/include/asm/cpufeatures.h for the valid bit
762 numbers. Note the Linux specific bits are not necessarily
763 stable over kernel options, but the vendor specific
764 ones should be.
765 Also note that user programs calling CPUID directly
766 or using the feature without checking anything
767 will still see it. This just prevents it from
768 being used by the kernel or shown in /proc/cpuinfo.
769 Also note the kernel might malfunction if you disable
770 some critical bits.
772 cma=nn[MG]@[start[MG][-end[MG]]]
773 [ARM,X86,KNL]
774 Sets the size of kernel global memory area for
775 contiguous memory allocations and optionally the
776 placement constraint by the physical address range of
777 memory allocations. A value of 0 disables CMA
778 altogether. For more information, see
779 include/linux/dma-contiguous.h
781 cmo_free_hint= [PPC] Format: { yes | no }
782 Specify whether pages are marked as being inactive
783 when they are freed. This is used in CMO environments
784 to determine OS memory pressure for page stealing by
785 a hypervisor.
786 Default: yes
788 coherent_pool=nn[KMG] [ARM,KNL]
789 Sets the size of memory pool for coherent, atomic dma
790 allocations, by default set to 256K.
792 code_bytes [X86] How many bytes of object code to print
793 in an oops report.
794 Range: 0 - 8192
795 Default: 64
797 com20020= [HW,NET] ARCnet - COM20020 chipset
798 Format:
799 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
801 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
802 Format: <io>[,<irq>]
804 com90xx= [HW,NET]
805 ARCnet - COM90xx chipset (memory-mapped buffers)
806 Format: <io>[,<irq>[,<memstart>]]
808 condev= [HW,S390] console device
809 conmode=
811 console= [KNL] Output console device and options.
813 tty<n> Use the virtual console device <n>.
815 ttyS<n>[,options]
816 ttyUSB0[,options]
817 Use the specified serial port. The options are of
818 the form "bbbbpnf", where "bbbb" is the baud rate,
819 "p" is parity ("n", "o", or "e"), "n" is number of
820 bits, and "f" is flow control ("r" for RTS or
821 omit it). Default is "9600n8".
823 See Documentation/serial-console.txt for more
824 information. See
825 Documentation/networking/netconsole.txt for an
826 alternative.
828 uart[8250],io,<addr>[,options]
829 uart[8250],mmio,<addr>[,options]
830 uart[8250],mmio16,<addr>[,options]
831 uart[8250],mmio32,<addr>[,options]
832 uart[8250],0x<addr>[,options]
833 Start an early, polled-mode console on the 8250/16550
834 UART at the specified I/O port or MMIO address,
835 switching to the matching ttyS device later.
836 MMIO inter-register address stride is either 8-bit
837 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
838 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
839 to be equivalent to 'mmio'. 'options' are specified in
840 the same format described for ttyS above; if unspecified,
841 the h/w is not re-initialized.
843 hvc<n> Use the hypervisor console device <n>. This is for
844 both Xen and PowerPC hypervisors.
846 If the device connected to the port is not a TTY but a braille
847 device, prepend "brl," before the device type, for instance
848 console=brl,ttyS0
849 For now, only VisioBraille is supported.
851 consoleblank= [KNL] The console blank (screen saver) timeout in
852 seconds. Defaults to 10*60 = 10mins. A value of 0
853 disables the blank timer.
855 coredump_filter=
856 [KNL] Change the default value for
857 /proc/<pid>/coredump_filter.
858 See also Documentation/filesystems/proc.txt.
860 cpuidle.off=1 [CPU_IDLE]
861 disable the cpuidle sub-system
863 cpu_init_udelay=N
864 [X86] Delay for N microsec between assert and de-assert
865 of APIC INIT to start processors. This delay occurs
866 on every CPU online, such as boot, and resume from suspend.
867 Default: 10000
869 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
870 Format:
871 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
873 crashkernel=size[KMG][@offset[KMG]]
874 [KNL] Using kexec, Linux can switch to a 'crash kernel'
875 upon panic. This parameter reserves the physical
876 memory region [offset, offset + size] for that kernel
877 image. If '@offset' is omitted, then a suitable offset
878 is selected automatically. Check
879 Documentation/kdump/kdump.txt for further details.
881 crashkernel=range1:size1[,range2:size2,...][@offset]
882 [KNL] Same as above, but depends on the memory
883 in the running system. The syntax of range is
884 start-[end] where start and end are both
885 a memory unit (amount[KMG]). See also
886 Documentation/kdump/kdump.txt for an example.
888 crashkernel=size[KMG],high
889 [KNL, x86_64] range could be above 4G. Allow kernel
890 to allocate physical memory region from top, so could
891 be above 4G if system have more than 4G ram installed.
892 Otherwise memory region will be allocated below 4G, if
893 available.
894 It will be ignored if crashkernel=X is specified.
895 crashkernel=size[KMG],low
896 [KNL, x86_64] range under 4G. When crashkernel=X,high
897 is passed, kernel could allocate physical memory region
898 above 4G, that cause second kernel crash on system
899 that require some amount of low memory, e.g. swiotlb
900 requires at least 64M+32K low memory, also enough extra
901 low memory is needed to make sure DMA buffers for 32-bit
902 devices won't run out. Kernel would try to allocate at
903 at least 256M below 4G automatically.
904 This one let user to specify own low range under 4G
905 for second kernel instead.
906 0: to disable low allocation.
907 It will be ignored when crashkernel=X,high is not used
908 or memory reserved is below 4G.
910 cryptomgr.notests
911 [KNL] Disable crypto self-tests
913 cs89x0_dma= [HW,NET]
914 Format: <dma>
916 cs89x0_media= [HW,NET]
917 Format: { rj45 | aui | bnc }
919 dasd= [HW,NET]
920 See header of drivers/s390/block/dasd_devmap.c.
922 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
923 (one device per port)
924 Format: <port#>,<type>
925 See also Documentation/input/joystick-parport.txt
927 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
928 time. See Documentation/dynamic-debug-howto.txt for
929 details. Deprecated, see dyndbg.
931 debug [KNL] Enable kernel debugging (events log level).
933 debug_locks_verbose=
934 [KNL] verbose self-tests
935 Format=<0|1>
936 Print debugging info while doing the locking API
937 self-tests.
938 We default to 0 (no extra messages), setting it to
939 1 will print _a lot_ more information - normally
940 only useful to kernel developers.
942 debug_objects [KNL] Enable object debugging
944 no_debug_objects
945 [KNL] Disable object debugging
947 debug_guardpage_minorder=
948 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
949 parameter allows control of the order of pages that will
950 be intentionally kept free (and hence protected) by the
951 buddy allocator. Bigger value increase the probability
952 of catching random memory corruption, but reduce the
953 amount of memory for normal system use. The maximum
954 possible value is MAX_ORDER/2. Setting this parameter
955 to 1 or 2 should be enough to identify most random
956 memory corruption problems caused by bugs in kernel or
957 driver code when a CPU writes to (or reads from) a
958 random memory location. Note that there exists a class
959 of memory corruptions problems caused by buggy H/W or
960 F/W or by drivers badly programing DMA (basically when
961 memory is written at bus level and the CPU MMU is
962 bypassed) which are not detectable by
963 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
964 tracking down these problems.
966 debug_pagealloc=
967 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
968 parameter enables the feature at boot time. In
969 default, it is disabled. We can avoid allocating huge
970 chunk of memory for debug pagealloc if we don't enable
971 it at boot time and the system will work mostly same
972 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
973 on: enable the feature
975 debugpat [X86] Enable PAT debugging
977 decnet.addr= [HW,NET]
978 Format: <area>[,<node>]
979 See also Documentation/networking/decnet.txt.
981 default_hugepagesz=
982 [same as hugepagesz=] The size of the default
983 HugeTLB page size. This is the size represented by
984 the legacy /proc/ hugepages APIs, used for SHM, and
985 default size when mounting hugetlbfs filesystems.
986 Defaults to the default architecture's huge page size
987 if not specified.
989 dhash_entries= [KNL]
990 Set number of hash buckets for dentry cache.
992 disable_1tb_segments [PPC]
993 Disables the use of 1TB hash page table segments. This
994 causes the kernel to fall back to 256MB segments which
995 can be useful when debugging issues that require an SLB
996 miss to occur.
998 disable= [IPV6]
999 See Documentation/networking/ipv6.txt.
1001 disable_radix [PPC]
1002 Disable RADIX MMU mode on POWER9
1004 disable_cpu_apicid= [X86,APIC,SMP]
1005 Format: <int>
1006 The number of initial APIC ID for the
1007 corresponding CPU to be disabled at boot,
1008 mostly used for the kdump 2nd kernel to
1009 disable BSP to wake up multiple CPUs without
1010 causing system reset or hang due to sending
1011 INIT from AP to BSP.
1013 disable_ddw [PPC/PSERIES]
1014 Disable Dynamic DMA Window support. Use this if
1015 to workaround buggy firmware.
1017 disable_ipv6= [IPV6]
1018 See Documentation/networking/ipv6.txt.
1020 disable_mtrr_cleanup [X86]
1021 The kernel tries to adjust MTRR layout from continuous
1022 to discrete, to make X server driver able to add WB
1023 entry later. This parameter disables that.
1025 disable_mtrr_trim [X86, Intel and AMD only]
1026 By default the kernel will trim any uncacheable
1027 memory out of your available memory pool based on
1028 MTRR settings. This parameter disables that behavior,
1029 possibly causing your machine to run very slowly.
1031 disable_timer_pin_1 [X86]
1032 Disable PIN 1 of APIC timer
1033 Can be useful to work around chipset bugs.
1035 dis_ucode_ldr [X86] Disable the microcode loader.
1037 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1038 this option disables the debugging code at boot.
1040 dma_debug_entries=<number>
1041 This option allows to tune the number of preallocated
1042 entries for DMA-API debugging code. One entry is
1043 required per DMA-API allocation. Use this if the
1044 DMA-API debugging code disables itself because the
1045 architectural default is too low.
1047 dma_debug_driver=<driver_name>
1048 With this option the DMA-API debugging driver
1049 filter feature can be enabled at boot time. Just
1050 pass the driver to filter for as the parameter.
1051 The filter can be disabled or changed to another
1052 driver later using sysfs.
1054 drm_kms_helper.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1055 Broken monitors, graphic adapters, KVMs and EDIDless
1056 panels may send no or incorrect EDID data sets.
1057 This parameter allows to specify an EDID data sets
1058 in the /lib/firmware directory that are used instead.
1059 Generic built-in EDID data sets are used, if one of
1060 edid/1024x768.bin, edid/1280x1024.bin,
1061 edid/1680x1050.bin, or edid/1920x1080.bin is given
1062 and no file with the same name exists. Details and
1063 instructions how to build your own EDID data are
1064 available in Documentation/EDID/HOWTO.txt. An EDID
1065 data set will only be used for a particular connector,
1066 if its name and a colon are prepended to the EDID
1067 name. Each connector may use a unique EDID data
1068 set by separating the files with a comma. An EDID
1069 data set with no connector name will be used for
1070 any connectors not explicitly specified.
1072 dscc4.setup= [NET]
1074 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1075 module.dyndbg[="val"]
1076 Enable debug messages at boot time. See
1077 Documentation/dynamic-debug-howto.txt for details.
1079 nompx [X86] Disables Intel Memory Protection Extensions.
1080 See Documentation/x86/intel_mpx.txt for more
1081 information about the feature.
1083 nopku [X86] Disable Memory Protection Keys CPU feature found
1084 in some Intel CPUs.
1086 module.async_probe [KNL]
1087 Enable asynchronous probe on this module.
1089 early_ioremap_debug [KNL]
1090 Enable debug messages in early_ioremap support. This
1091 is useful for tracking down temporary early mappings
1092 which are not unmapped.
1094 earlycon= [KNL] Output early console device and options.
1096 When used with no options, the early console is
1097 determined by the stdout-path property in device
1098 tree's chosen node.
1100 cdns,<addr>[,options]
1101 Start an early, polled-mode console on a Cadence
1102 (xuartps) serial port at the specified address. Only
1103 supported option is baud rate. If baud rate is not
1104 specified, the serial port must already be setup and
1105 configured.
1107 uart[8250],io,<addr>[,options]
1108 uart[8250],mmio,<addr>[,options]
1109 uart[8250],mmio32,<addr>[,options]
1110 uart[8250],mmio32be,<addr>[,options]
1111 uart[8250],0x<addr>[,options]
1112 Start an early, polled-mode console on the 8250/16550
1113 UART at the specified I/O port or MMIO address.
1114 MMIO inter-register address stride is either 8-bit
1115 (mmio) or 32-bit (mmio32 or mmio32be).
1116 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1117 to be equivalent to 'mmio'. 'options' are specified
1118 in the same format described for "console=ttyS<n>"; if
1119 unspecified, the h/w is not initialized.
1121 pl011,<addr>
1122 pl011,mmio32,<addr>
1123 Start an early, polled-mode console on a pl011 serial
1124 port at the specified address. The pl011 serial port
1125 must already be setup and configured. Options are not
1126 yet supported. If 'mmio32' is specified, then only
1127 the driver will use only 32-bit accessors to read/write
1128 the device registers.
1130 meson,<addr>
1131 Start an early, polled-mode console on a meson serial
1132 port at the specified address. The serial port must
1133 already be setup and configured. Options are not yet
1134 supported.
1136 msm_serial,<addr>
1137 Start an early, polled-mode console on an msm serial
1138 port at the specified address. The serial port
1139 must already be setup and configured. Options are not
1140 yet supported.
1142 msm_serial_dm,<addr>
1143 Start an early, polled-mode console on an msm serial
1144 dm port at the specified address. The serial port
1145 must already be setup and configured. Options are not
1146 yet supported.
1148 smh Use ARM semihosting calls for early console.
1150 s3c2410,<addr>
1151 s3c2412,<addr>
1152 s3c2440,<addr>
1153 s3c6400,<addr>
1154 s5pv210,<addr>
1155 exynos4210,<addr>
1156 Use early console provided by serial driver available
1157 on Samsung SoCs, requires selecting proper type and
1158 a correct base address of the selected UART port. The
1159 serial port must already be setup and configured.
1160 Options are not yet supported.
1162 lpuart,<addr>
1163 lpuart32,<addr>
1164 Use early console provided by Freescale LP UART driver
1165 found on Freescale Vybrid and QorIQ LS1021A processors.
1166 A valid base address must be provided, and the serial
1167 port must already be setup and configured.
1169 armada3700_uart,<addr>
1170 Start an early, polled-mode console on the
1171 Armada 3700 serial port at the specified
1172 address. The serial port must already be setup
1173 and configured. Options are not yet supported.
1175 earlyprintk= [X86,SH,BLACKFIN,ARM,M68k]
1176 earlyprintk=vga
1177 earlyprintk=efi
1178 earlyprintk=xen
1179 earlyprintk=serial[,ttySn[,baudrate]]
1180 earlyprintk=serial[,0x...[,baudrate]]
1181 earlyprintk=ttySn[,baudrate]
1182 earlyprintk=dbgp[debugController#]
1183 earlyprintk=pciserial,bus:device.function[,baudrate]
1185 earlyprintk is useful when the kernel crashes before
1186 the normal console is initialized. It is not enabled by
1187 default because it has some cosmetic problems.
1189 Append ",keep" to not disable it when the real console
1190 takes over.
1192 Only one of vga, efi, serial, or usb debug port can
1193 be used at a time.
1195 Currently only ttyS0 and ttyS1 may be specified by
1196 name. Other I/O ports may be explicitly specified
1197 on some architectures (x86 and arm at least) by
1198 replacing ttySn with an I/O port address, like this:
1199 earlyprintk=serial,0x1008,115200
1200 You can find the port for a given device in
1201 /proc/tty/driver/serial:
1202 2: uart:ST16650V2 port:00001008 irq:18 ...
1204 Interaction with the standard serial driver is not
1205 very good.
1207 The VGA and EFI output is eventually overwritten by
1208 the real console.
1210 The xen output can only be used by Xen PV guests.
1212 edac_report= [HW,EDAC] Control how to report EDAC event
1213 Format: {"on" | "off" | "force"}
1214 on: enable EDAC to report H/W event. May be overridden
1215 by other higher priority error reporting module.
1216 off: disable H/W event reporting through EDAC.
1217 force: enforce the use of EDAC to report H/W event.
1218 default: on.
1220 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1221 ekgdboc=kbd
1223 This is designed to be used in conjunction with
1224 the boot argument: earlyprintk=vga
1226 edd= [EDD]
1227 Format: {"off" | "on" | "skip[mbr]"}
1229 efi= [EFI]
1230 Format: { "old_map", "nochunk", "noruntime", "debug" }
1231 old_map [X86-64]: switch to the old ioremap-based EFI
1232 runtime services mapping. 32-bit still uses this one by
1233 default.
1234 nochunk: disable reading files in "chunks" in the EFI
1235 boot stub, as chunking can cause problems with some
1236 firmware implementations.
1237 noruntime : disable EFI runtime services support
1238 debug: enable misc debug output
1240 efi_no_storage_paranoia [EFI; X86]
1241 Using this parameter you can use more than 50% of
1242 your efi variable storage. Use this parameter only if
1243 you are really sure that your UEFI does sane gc and
1244 fulfills the spec otherwise your board may brick.
1246 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1247 Add arbitrary attribute to specific memory range by
1248 updating original EFI memory map.
1249 Region of memory which aa attribute is added to is
1250 from ss to ss+nn.
1251 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1252 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1253 attribute is added to range 0x100000000-0x180000000 and
1254 0x10a0000000-0x1120000000.
1256 Using this parameter you can do debugging of EFI memmap
1257 related feature. For example, you can do debugging of
1258 Address Range Mirroring feature even if your box
1259 doesn't support it.
1261 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1262 that is to be dynamically loaded by Linux. If there are
1263 multiple variables with the same name but with different
1264 vendor GUIDs, all of them will be loaded. See
1265 Documentation/acpi/ssdt-overlays.txt for details.
1268 eisa_irq_edge= [PARISC,HW]
1269 See header of drivers/parisc/eisa.c.
1271 elanfreq= [X86-32]
1272 See comment before function elanfreq_setup() in
1273 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1275 elevator= [IOSCHED]
1276 Format: {"cfq" | "deadline" | "noop"}
1277 See Documentation/block/cfq-iosched.txt and
1278 Documentation/block/deadline-iosched.txt for details.
1280 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1281 Specifies physical address of start of kernel core
1282 image elf header and optionally the size. Generally
1283 kexec loader will pass this option to capture kernel.
1284 See Documentation/kdump/kdump.txt for details.
1286 enable_mtrr_cleanup [X86]
1287 The kernel tries to adjust MTRR layout from continuous
1288 to discrete, to make X server driver able to add WB
1289 entry later. This parameter enables that.
1291 enable_timer_pin_1 [X86]
1292 Enable PIN 1 of APIC timer
1293 Can be useful to work around chipset bugs
1294 (in particular on some ATI chipsets).
1295 The kernel tries to set a reasonable default.
1297 enforcing [SELINUX] Set initial enforcing status.
1298 Format: {"0" | "1"}
1299 See security/selinux/Kconfig help text.
1300 0 -- permissive (log only, no denials).
1301 1 -- enforcing (deny and log).
1302 Default value is 0.
1303 Value can be changed at runtime via /selinux/enforce.
1305 erst_disable [ACPI]
1306 Disable Error Record Serialization Table (ERST)
1307 support.
1309 ether= [HW,NET] Ethernet cards parameters
1310 This option is obsoleted by the "netdev=" option, which
1311 has equivalent usage. See its documentation for details.
1313 evm= [EVM]
1314 Format: { "fix" }
1315 Permit 'security.evm' to be updated regardless of
1316 current integrity status.
1318 failslab=
1319 fail_page_alloc=
1320 fail_make_request=[KNL]
1321 General fault injection mechanism.
1322 Format: <interval>,<probability>,<space>,<times>
1323 See also Documentation/fault-injection/.
1325 floppy= [HW]
1326 See Documentation/blockdev/floppy.txt.
1328 force_pal_cache_flush
1329 [IA-64] Avoid check_sal_cache_flush which may hang on
1330 buggy SAL_CACHE_FLUSH implementations. Using this
1331 parameter will force ia64_sal_cache_flush to call
1332 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1334 forcepae [X86-32]
1335 Forcefully enable Physical Address Extension (PAE).
1336 Many Pentium M systems disable PAE but may have a
1337 functionally usable PAE implementation.
1338 Warning: use of this parameter will taint the kernel
1339 and may cause unknown problems.
1341 ftrace=[tracer]
1342 [FTRACE] will set and start the specified tracer
1343 as early as possible in order to facilitate early
1344 boot debugging.
1346 ftrace_dump_on_oops[=orig_cpu]
1347 [FTRACE] will dump the trace buffers on oops.
1348 If no parameter is passed, ftrace will dump
1349 buffers of all CPUs, but if you pass orig_cpu, it will
1350 dump only the buffer of the CPU that triggered the
1351 oops.
1353 ftrace_filter=[function-list]
1354 [FTRACE] Limit the functions traced by the function
1355 tracer at boot up. function-list is a comma separated
1356 list of functions. This list can be changed at run
1357 time by the set_ftrace_filter file in the debugfs
1358 tracing directory.
1360 ftrace_notrace=[function-list]
1361 [FTRACE] Do not trace the functions specified in
1362 function-list. This list can be changed at run time
1363 by the set_ftrace_notrace file in the debugfs
1364 tracing directory.
1366 ftrace_graph_filter=[function-list]
1367 [FTRACE] Limit the top level callers functions traced
1368 by the function graph tracer at boot up.
1369 function-list is a comma separated list of functions
1370 that can be changed at run time by the
1371 set_graph_function file in the debugfs tracing directory.
1373 ftrace_graph_notrace=[function-list]
1374 [FTRACE] Do not trace from the functions specified in
1375 function-list. This list is a comma separated list of
1376 functions that can be changed at run time by the
1377 set_graph_notrace file in the debugfs tracing directory.
1379 gamecon.map[2|3]=
1380 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1381 support via parallel port (up to 5 devices per port)
1382 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1383 See also Documentation/input/joystick-parport.txt
1385 gamma= [HW,DRM]
1387 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1388 Format: off | on
1389 default: on
1391 gcov_persist= [GCOV] When non-zero (default), profiling data for
1392 kernel modules is saved and remains accessible via
1393 debugfs, even when the module is unloaded/reloaded.
1394 When zero, profiling data is discarded and associated
1395 debugfs files are removed at module unload time.
1397 goldfish [X86] Enable the goldfish android emulator platform.
1398 Don't use this when you are not running on the
1399 android emulator
1401 gpt [EFI] Forces disk with valid GPT signature but
1402 invalid Protective MBR to be treated as GPT. If the
1403 primary GPT is corrupted, it enables the backup/alternate
1404 GPT to be used instead.
1406 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1407 the "Enable 0" bit of the configuration register.
1408 Format: 0 | 1
1409 Default: 0
1410 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1411 the "Enable 0" bit of the configuration register.
1412 Format: 0 | 1
1413 Default: 0
1414 grcan.select= [HW] Select which physical interface to use.
1415 Format: 0 | 1
1416 Default: 0
1417 grcan.txsize= [HW] Sets the size of the tx buffer.
1418 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1419 Default: 1024
1420 grcan.rxsize= [HW] Sets the size of the rx buffer.
1421 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1422 Default: 1024
1424 gpio-mockup.gpio_mockup_ranges
1425 [HW] Sets the ranges of gpiochip of for this device.
1426 Format: <start1>,<end1>,<start2>,<end2>...
1428 hardlockup_all_cpu_backtrace=
1429 [KNL] Should the hard-lockup detector generate
1430 backtraces on all cpus.
1431 Format: <integer>
1433 hashdist= [KNL,NUMA] Large hashes allocated during boot
1434 are distributed across NUMA nodes. Defaults on
1435 for 64-bit NUMA, off otherwise.
1436 Format: 0 | 1 (for off | on)
1438 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1440 hd= [EIDE] (E)IDE hard drive subsystem geometry
1441 Format: <cyl>,<head>,<sect>
1443 hest_disable [ACPI]
1444 Disable Hardware Error Source Table (HEST) support;
1445 corresponding firmware-first mode error processing
1446 logic will be disabled.
1448 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1449 size of <nn>. This works even on boxes that have no
1450 highmem otherwise. This also works to reduce highmem
1451 size on bigger boxes.
1453 highres= [KNL] Enable/disable high resolution timer mode.
1454 Valid parameters: "on", "off"
1455 Default: "on"
1457 hisax= [HW,ISDN]
1458 See Documentation/isdn/README.HiSax.
1460 hlt [BUGS=ARM,SH]
1462 hpet= [X86-32,HPET] option to control HPET usage
1463 Format: { enable (default) | disable | force |
1464 verbose }
1465 disable: disable HPET and use PIT instead
1466 force: allow force enabled of undocumented chips (ICH4,
1467 VIA, nVidia)
1468 verbose: show contents of HPET registers during setup
1470 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1471 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1473 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1474 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1475 On x86-64 and powerpc, this option can be specified
1476 multiple times interleaved with hugepages= to reserve
1477 huge pages of different sizes. Valid pages sizes on
1478 x86-64 are 2M (when the CPU supports "pse") and 1G
1479 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1481 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1482 terminal devices. Valid values: 0..8
1483 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1484 If specified, z/VM IUCV HVC accepts connections
1485 from listed z/VM user IDs only.
1487 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1488 hardware thread id mappings.
1489 Format: <cpu>:<hwthread>
1491 keep_bootcon [KNL]
1492 Do not unregister boot console at start. This is only
1493 useful for debugging when something happens in the window
1494 between unregistering the boot console and initializing
1495 the real console.
1497 i2c_bus= [HW] Override the default board specific I2C bus speed
1498 or register an additional I2C bus that is not
1499 registered from board initialization code.
1500 Format:
1501 <bus_id>,<clkrate>
1503 i8042.debug [HW] Toggle i8042 debug mode
1504 i8042.unmask_kbd_data
1505 [HW] Enable printing of interrupt data from the KBD port
1506 (disabled by default, and as a pre-condition
1507 requires that i8042.debug=1 be enabled)
1508 i8042.direct [HW] Put keyboard port into non-translated mode
1509 i8042.dumbkbd [HW] Pretend that controller can only read data from
1510 keyboard and cannot control its state
1511 (Don't attempt to blink the leds)
1512 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1513 i8042.nokbd [HW] Don't check/create keyboard port
1514 i8042.noloop [HW] Disable the AUX Loopback command while probing
1515 for the AUX port
1516 i8042.nomux [HW] Don't check presence of an active multiplexing
1517 controller
1518 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1519 controllers
1520 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1521 i8042.reset [HW] Reset the controller during init, cleanup and
1522 suspend-to-ram transitions, only during s2r
1523 transitions, or never reset
1524 Format: { 1 | Y | y | 0 | N | n }
1525 1, Y, y: always reset controller
1526 0, N, n: don't ever reset controller
1527 Default: only on s2r transitions on x86; most other
1528 architectures force reset to be always executed
1529 i8042.unlock [HW] Unlock (ignore) the keylock
1530 i8042.kbdreset [HW] Reset device connected to KBD port
1532 i810= [HW,DRM]
1534 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1535 indicates that the driver is running on unsupported
1536 hardware.
1537 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1538 does not match list of supported models.
1539 i8k.power_status
1540 [HW] Report power status in /proc/i8k
1541 (disabled by default)
1542 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1543 capability is set.
1545 i915.invert_brightness=
1546 [DRM] Invert the sense of the variable that is used to
1547 set the brightness of the panel backlight. Normally a
1548 brightness value of 0 indicates backlight switched off,
1549 and the maximum of the brightness value sets the backlight
1550 to maximum brightness. If this parameter is set to 0
1551 (default) and the machine requires it, or this parameter
1552 is set to 1, a brightness value of 0 sets the backlight
1553 to maximum brightness, and the maximum of the brightness
1554 value switches the backlight off.
1555 -1 -- never invert brightness
1556 0 -- machine default
1557 1 -- force brightness inversion
1559 icn= [HW,ISDN]
1560 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1562 ide-core.nodma= [HW] (E)IDE subsystem
1563 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1564 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1565 .cdrom .chs .ignore_cable are additional options
1566 See Documentation/ide/ide.txt.
1568 ide-generic.probe-mask= [HW] (E)IDE subsystem
1569 Format: <int>
1570 Probe mask for legacy ISA IDE ports. Depending on
1571 platform up to 6 ports are supported, enabled by
1572 setting corresponding bits in the mask to 1. The
1573 default value is 0x0, which has a special meaning.
1574 On systems that have PCI, it triggers scanning the
1575 PCI bus for the first and the second port, which
1576 are then probed. On systems without PCI the value
1577 of 0x0 enables probing the two first ports as if it
1578 was 0x3.
1580 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1581 Claim all unknown PCI IDE storage controllers.
1583 idle= [X86]
1584 Format: idle=poll, idle=halt, idle=nomwait
1585 Poll forces a polling idle loop that can slightly
1586 improve the performance of waking up a idle CPU, but
1587 will use a lot of power and make the system run hot.
1588 Not recommended.
1589 idle=halt: Halt is forced to be used for CPU idle.
1590 In such case C2/C3 won't be used again.
1591 idle=nomwait: Disable mwait for CPU C-states
1593 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1594 Format: { strict | legacy | 2008 | relaxed }
1595 Default: strict
1597 Choose which programs will be accepted for execution
1598 based on the IEEE 754 NaN encoding(s) supported by
1599 the FPU and the NaN encoding requested with the value
1600 of an ELF file header flag individually set by each
1601 binary. Hardware implementations are permitted to
1602 support either or both of the legacy and the 2008 NaN
1603 encoding mode.
1605 Available settings are as follows:
1606 strict accept binaries that request a NaN encoding
1607 supported by the FPU
1608 legacy only accept legacy-NaN binaries, if supported
1609 by the FPU
1610 2008 only accept 2008-NaN binaries, if supported
1611 by the FPU
1612 relaxed accept any binaries regardless of whether
1613 supported by the FPU
1615 The FPU emulator is always able to support both NaN
1616 encodings, so if no FPU hardware is present or it has
1617 been disabled with 'nofpu', then the settings of
1618 'legacy' and '2008' strap the emulator accordingly,
1619 'relaxed' straps the emulator for both legacy-NaN and
1620 2008-NaN, whereas 'strict' enables legacy-NaN only on
1621 legacy processors and both NaN encodings on MIPS32 or
1622 MIPS64 CPUs.
1624 The setting for ABS.fmt/NEG.fmt instruction execution
1625 mode generally follows that for the NaN encoding,
1626 except where unsupported by hardware.
1628 ignore_loglevel [KNL]
1629 Ignore loglevel setting - this will print /all/
1630 kernel messages to the console. Useful for debugging.
1631 We also add it as printk module parameter, so users
1632 could change it dynamically, usually by
1633 /sys/module/printk/parameters/ignore_loglevel.
1635 ignore_rlimit_data
1636 Ignore RLIMIT_DATA setting for data mappings,
1637 print warning at first misuse. Can be changed via
1638 /sys/module/kernel/parameters/ignore_rlimit_data.
1640 ihash_entries= [KNL]
1641 Set number of hash buckets for inode cache.
1643 ima_appraise= [IMA] appraise integrity measurements
1644 Format: { "off" | "enforce" | "fix" | "log" }
1645 default: "enforce"
1647 ima_appraise_tcb [IMA]
1648 The builtin appraise policy appraises all files
1649 owned by uid=0.
1651 ima_hash= [IMA]
1652 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1653 | sha512 | ... }
1654 default: "sha1"
1656 The list of supported hash algorithms is defined
1657 in crypto/hash_info.h.
1659 ima_policy= [IMA]
1660 The builtin measurement policy to load during IMA
1661 setup. Specyfing "tcb" as the value, measures all
1662 programs exec'd, files mmap'd for exec, and all files
1663 opened with the read mode bit set by either the
1664 effective uid (euid=0) or uid=0.
1665 Format: "tcb"
1667 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1668 Load a policy which meets the needs of the Trusted
1669 Computing Base. This means IMA will measure all
1670 programs exec'd, files mmap'd for exec, and all files
1671 opened for read by uid=0.
1673 ima_template= [IMA]
1674 Select one of defined IMA measurements template formats.
1675 Formats: { "ima" | "ima-ng" | "ima-sig" }
1676 Default: "ima-ng"
1678 ima_template_fmt=
1679 [IMA] Define a custom template format.
1680 Format: { "field1|...|fieldN" }
1682 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1683 Format: <min_file_size>
1684 Set the minimal file size for using asynchronous hash.
1685 If left unspecified, ahash usage is disabled.
1687 ahash performance varies for different data sizes on
1688 different crypto accelerators. This option can be used
1689 to achieve the best performance for a particular HW.
1691 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1692 Format: <bufsize>
1693 Set hashing buffer size. Default: 4k.
1695 ahash performance varies for different chunk sizes on
1696 different crypto accelerators. This option can be used
1697 to achieve best performance for particular HW.
1699 init= [KNL]
1700 Format: <full_path>
1701 Run specified binary instead of /sbin/init as init
1702 process.
1704 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1705 for working out where the kernel is dying during
1706 startup.
1708 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1709 initcall functions. Useful for debugging built-in
1710 modules and initcalls.
1712 initrd= [BOOT] Specify the location of the initial ramdisk
1714 init_pkru= [x86] Specify the default memory protection keys rights
1715 register contents for all processes. 0x55555554 by
1716 default (disallow access to all but pkey 0). Can
1717 override in debugfs after boot.
1719 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1720 Format: <irq>
1722 int_pln_enable [x86] Enable power limit notification interrupt
1724 integrity_audit=[IMA]
1725 Format: { "0" | "1" }
1726 0 -- basic integrity auditing messages. (Default)
1727 1 -- additional integrity auditing messages.
1729 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1730 on
1731 Enable intel iommu driver.
1732 off
1733 Disable intel iommu driver.
1734 igfx_off [Default Off]
1735 By default, gfx is mapped as normal device. If a gfx
1736 device has a dedicated DMAR unit, the DMAR unit is
1737 bypassed by not enabling DMAR with this option. In
1738 this case, gfx device will use physical address for
1739 DMA.
1740 forcedac [x86_64]
1741 With this option iommu will not optimize to look
1742 for io virtual address below 32-bit forcing dual
1743 address cycle on pci bus for cards supporting greater
1744 than 32-bit addressing. The default is to look
1745 for translation below 32-bit and if not available
1746 then look in the higher range.
1747 strict [Default Off]
1748 With this option on every unmap_single operation will
1749 result in a hardware IOTLB flush operation as opposed
1750 to batching them for performance.
1751 sp_off [Default Off]
1752 By default, super page will be supported if Intel IOMMU
1753 has the capability. With this option, super page will
1754 not be supported.
1755 ecs_off [Default Off]
1756 By default, extended context tables will be supported if
1757 the hardware advertises that it has support both for the
1758 extended tables themselves, and also PASID support. With
1759 this option set, extended tables will not be used even
1760 on hardware which claims to support them.
1762 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1763 0 disables intel_idle and fall back on acpi_idle.
1764 1 to 9 specify maximum depth of C-state.
1766 intel_pstate= [X86]
1767 disable
1768 Do not enable intel_pstate as the default
1769 scaling driver for the supported processors
1770 force
1771 Enable intel_pstate on systems that prohibit it by default
1772 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1773 instead of acpi-cpufreq may disable platform features, such
1774 as thermal controls and power capping, that rely on ACPI
1775 P-States information being indicated to OSPM and therefore
1776 should be used with caution. This option does not work with
1777 processors that aren't supported by the intel_pstate driver
1778 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1779 no_hwp
1780 Do not enable hardware P state control (HWP)
1781 if available.
1782 hwp_only
1783 Only load intel_pstate on systems which support
1784 hardware P state control (HWP) if available.
1785 support_acpi_ppc
1786 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1787 Description Table, specifies preferred power management
1788 profile as "Enterprise Server" or "Performance Server",
1789 then this feature is turned on by default.
1791 intremap= [X86-64, Intel-IOMMU]
1792 on enable Interrupt Remapping (default)
1793 off disable Interrupt Remapping
1794 nosid disable Source ID checking
1795 no_x2apic_optout
1796 BIOS x2APIC opt-out request will be ignored
1797 nopost disable Interrupt Posting
1799 iomem= Disable strict checking of access to MMIO memory
1800 strict regions from userspace.
1801 relaxed
1803 iommu= [x86]
1804 off
1805 force
1806 noforce
1807 biomerge
1808 panic
1809 nopanic
1810 merge
1811 nomerge
1812 forcesac
1813 soft
1814 pt [x86, IA-64]
1815 nobypass [PPC/POWERNV]
1816 Disable IOMMU bypass, using IOMMU for PCI devices.
1819 io7= [HW] IO7 for Marvel based alpha systems
1820 See comment before marvel_specify_io7 in
1821 arch/alpha/kernel/core_marvel.c.
1823 io_delay= [X86] I/O delay method
1824 0x80
1825 Standard port 0x80 based delay
1826 0xed
1827 Alternate port 0xed based delay (needed on some systems)
1828 udelay
1829 Simple two microseconds delay
1830 none
1831 No delay
1833 ip= [IP_PNP]
1834 See Documentation/filesystems/nfs/nfsroot.txt.
1836 irqaffinity= [SMP] Set the default irq affinity mask
1837 The argument is a cpu list, as described above.
1839 irqfixup [HW]
1840 When an interrupt is not handled search all handlers
1841 for it. Intended to get systems with badly broken
1842 firmware running.
1844 irqpoll [HW]
1845 When an interrupt is not handled search all handlers
1846 for it. Also check all handlers each timer
1847 interrupt. Intended to get systems with badly broken
1848 firmware running.
1850 isapnp= [ISAPNP]
1851 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1853 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1854 The argument is a cpu list, as described above.
1856 This option can be used to specify one or more CPUs
1857 to isolate from the general SMP balancing and scheduling
1858 algorithms. You can move a process onto or off an
1859 "isolated" CPU via the CPU affinity syscalls or cpuset.
1860 <cpu number> begins at 0 and the maximum value is
1861 "number of CPUs in system - 1".
1863 This option is the preferred way to isolate CPUs. The
1864 alternative -- manually setting the CPU mask of all
1865 tasks in the system -- can cause problems and
1866 suboptimal load balancer performance.
1868 iucv= [HW,NET]
1870 ivrs_ioapic [HW,X86_64]
1871 Provide an override to the IOAPIC-ID<->DEVICE-ID
1872 mapping provided in the IVRS ACPI table. For
1873 example, to map IOAPIC-ID decimal 10 to
1874 PCI device 00:14.0 write the parameter as:
1875 ivrs_ioapic[10]=00:14.0
1877 ivrs_hpet [HW,X86_64]
1878 Provide an override to the HPET-ID<->DEVICE-ID
1879 mapping provided in the IVRS ACPI table. For
1880 example, to map HPET-ID decimal 0 to
1881 PCI device 00:14.0 write the parameter as:
1882 ivrs_hpet[0]=00:14.0
1884 ivrs_acpihid [HW,X86_64]
1885 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1886 mapping provided in the IVRS ACPI table. For
1887 example, to map UART-HID:UID AMD0020:0 to
1888 PCI device 00:14.5 write the parameter as:
1889 ivrs_acpihid[00:14.5]=AMD0020:0
1891 js= [HW,JOY] Analog joystick
1892 See Documentation/input/joystick.txt.
1894 nokaslr [KNL]
1895 When CONFIG_RANDOMIZE_BASE is set, this disables
1896 kernel and module base offset ASLR (Address Space
1897 Layout Randomization).
1899 keepinitrd [HW,ARM]
1901 kernelcore= [KNL,X86,IA-64,PPC]
1902 Format: nn[KMGTPE] | "mirror"
1903 This parameter
1904 specifies the amount of memory usable by the kernel
1905 for non-movable allocations. The requested amount is
1906 spread evenly throughout all nodes in the system. The
1907 remaining memory in each node is used for Movable
1908 pages. In the event, a node is too small to have both
1909 kernelcore and Movable pages, kernelcore pages will
1910 take priority and other nodes will have a larger number
1911 of Movable pages. The Movable zone is used for the
1912 allocation of pages that may be reclaimed or moved
1913 by the page migration subsystem. This means that
1914 HugeTLB pages may not be allocated from this zone.
1915 Note that allocations like PTEs-from-HighMem still
1916 use the HighMem zone if it exists, and the Normal
1917 zone if it does not.
1919 Instead of specifying the amount of memory (nn[KMGTPE]),
1920 you can specify "mirror" option. In case "mirror"
1921 option is specified, mirrored (reliable) memory is used
1922 for non-movable allocations and remaining memory is used
1923 for Movable pages. nn[KMGTPE] and "mirror" are exclusive,
1924 so you can NOT specify nn[KMGTPE] and "mirror" at the same
1925 time.
1927 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1928 Format: <Controller#>[,poll interval]
1929 The controller # is the number of the ehci usb debug
1930 port as it is probed via PCI. The poll interval is
1931 optional and is the number seconds in between
1932 each poll cycle to the debug port in case you need
1933 the functionality for interrupting the kernel with
1934 gdb or control-c on the dbgp connection. When
1935 not using this parameter you use sysrq-g to break into
1936 the kernel debugger.
1938 kgdboc= [KGDB,HW] kgdb over consoles.
1939 Requires a tty driver that supports console polling,
1940 or a supported polling keyboard driver (non-usb).
1941 Serial only format: <serial_device>[,baud]
1942 keyboard only format: kbd
1943 keyboard and serial format: kbd,<serial_device>[,baud]
1944 Optional Kernel mode setting:
1945 kms, kbd format: kms,kbd
1946 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1948 kgdbwait [KGDB] Stop kernel execution and enter the
1949 kernel debugger at the earliest opportunity.
1951 kmac= [MIPS] korina ethernet MAC address.
1952 Configure the RouterBoard 532 series on-chip
1953 Ethernet adapter MAC address.
1955 kmemleak= [KNL] Boot-time kmemleak enable/disable
1956 Valid arguments: on, off
1957 Default: on
1958 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1959 the default is off.
1961 kmemcheck= [X86] Boot-time kmemcheck enable/disable/one-shot mode
1962 Valid arguments: 0, 1, 2
1963 kmemcheck=0 (disabled)
1964 kmemcheck=1 (enabled)
1965 kmemcheck=2 (one-shot mode)
1966 Default: 2 (one-shot mode)
1968 kstack=N [X86] Print N words from the kernel stack
1969 in oops dumps.
1971 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1972 Default is 0 (don't ignore, but inject #GP)
1974 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1975 KVM MMU at runtime.
1976 Default is 0 (off)
1978 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1979 Default is 1 (enabled)
1981 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1982 for all guests.
1983 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1985 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1986 (virtualized MMU) support on capable Intel chips.
1987 Default is 1 (enabled)
1989 kvm-intel.emulate_invalid_guest_state=
1990 [KVM,Intel] Enable emulation of invalid guest states
1991 Default is 0 (disabled)
1993 kvm-intel.flexpriority=
1994 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1995 Default is 1 (enabled)
1997 kvm-intel.nested=
1998 [KVM,Intel] Enable VMX nesting (nVMX).
1999 Default is 0 (disabled)
2001 kvm-intel.unrestricted_guest=
2002 [KVM,Intel] Disable unrestricted guest feature
2003 (virtualized real and unpaged mode) on capable
2004 Intel chips. Default is 1 (enabled)
2006 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2007 CVE-2018-3620.
2009 Valid arguments: never, cond, always
2011 always: L1D cache flush on every VMENTER.
2012 cond: Flush L1D on VMENTER only when the code between
2013 VMEXIT and VMENTER can leak host memory.
2014 never: Disables the mitigation
2016 Default is cond (do L1 cache flush in specific instances)
2018 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2019 feature (tagged TLBs) on capable Intel chips.
2020 Default is 1 (enabled)
2022 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2023 affected CPUs
2025 The kernel PTE inversion protection is unconditionally
2026 enabled and cannot be disabled.
2028 full
2029 Provides all available mitigations for the
2030 L1TF vulnerability. Disables SMT and
2031 enables all mitigations in the
2032 hypervisors, i.e. unconditional L1D flush.
2034 SMT control and L1D flush control via the
2035 sysfs interface is still possible after
2036 boot. Hypervisors will issue a warning
2037 when the first VM is started in a
2038 potentially insecure configuration,
2039 i.e. SMT enabled or L1D flush disabled.
2041 full,force
2042 Same as 'full', but disables SMT and L1D
2043 flush runtime control. Implies the
2044 'nosmt=force' command line option.
2045 (i.e. sysfs control of SMT is disabled.)
2047 flush
2048 Leaves SMT enabled and enables the default
2049 hypervisor mitigation, i.e. conditional
2050 L1D flush.
2052 SMT control and L1D flush control via the
2053 sysfs interface is still possible after
2054 boot. Hypervisors will issue a warning
2055 when the first VM is started in a
2056 potentially insecure configuration,
2057 i.e. SMT enabled or L1D flush disabled.
2059 flush,nosmt
2061 Disables SMT and enables the default
2062 hypervisor mitigation.
2064 SMT control and L1D flush control via the
2065 sysfs interface is still possible after
2066 boot. Hypervisors will issue a warning
2067 when the first VM is started in a
2068 potentially insecure configuration,
2069 i.e. SMT enabled or L1D flush disabled.
2071 flush,nowarn
2072 Same as 'flush', but hypervisors will not
2073 warn when a VM is started in a potentially
2074 insecure configuration.
2076 off
2077 Disables hypervisor mitigations and doesn't
2078 emit any warnings.
2079 It also drops the swap size and available
2080 RAM limit restriction on both hypervisor and
2081 bare metal.
2083 Default is 'flush'.
2085 For details see: Documentation/hw-vuln/l1tf.rst
2087 l2cr= [PPC]
2089 l3cr= [PPC]
2091 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2092 disabled it.
2094 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2095 value for LAPIC timer one-shot implementation. Default
2096 back to the programmable timer unit in the LAPIC.
2098 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2099 in C2 power state.
2101 libata.dma= [LIBATA] DMA control
2102 libata.dma=0 Disable all PATA and SATA DMA
2103 libata.dma=1 PATA and SATA Disk DMA only
2104 libata.dma=2 ATAPI (CDROM) DMA only
2105 libata.dma=4 Compact Flash DMA only
2106 Combinations also work, so libata.dma=3 enables DMA
2107 for disks and CDROMs, but not CFs.
2109 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2110 libata.ignore_hpa=0 keep BIOS limits (default)
2111 libata.ignore_hpa=1 ignore limits, using full disk
2113 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2114 when set.
2115 Format: <int>
2117 libata.force= [LIBATA] Force configurations. The format is comma
2118 separated list of "[ID:]VAL" where ID is
2119 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2120 matching port, link or device. Basically, it matches
2121 the ATA ID string printed on console by libata. If
2122 the whole ID part is omitted, the last PORT and DEVICE
2123 values are used. If ID hasn't been specified yet, the
2124 configuration applies to all ports, links and devices.
2126 If only DEVICE is omitted, the parameter applies to
2127 the port and all links and devices behind it. DEVICE
2128 number of 0 either selects the first device or the
2129 first fan-out link behind PMP device. It does not
2130 select the host link. DEVICE number of 15 selects the
2131 host link and device attached to it.
2133 The VAL specifies the configuration to force. As long
2134 as there's no ambiguity shortcut notation is allowed.
2135 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2136 The following configurations can be forced.
2138 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2139 Any ID with matching PORT is used.
2141 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2143 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2144 udma[/][16,25,33,44,66,100,133] notation is also
2145 allowed.
2147 * [no]ncq: Turn on or off NCQ.
2149 * [no]ncqtrim: Turn off queued DSM TRIM.
2151 * nohrst, nosrst, norst: suppress hard, soft
2152 and both resets.
2154 * rstonce: only attempt one reset during
2155 hot-unplug link recovery
2157 * dump_id: dump IDENTIFY data.
2159 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2161 * disable: Disable this device.
2163 If there are multiple matching configurations changing
2164 the same attribute, the last one is used.
2166 memblock=debug [KNL] Enable memblock debug messages.
2168 load_ramdisk= [RAM] List of ramdisks to load from floppy
2169 See Documentation/blockdev/ramdisk.txt.
2171 lockd.nlm_grace_period=P [NFS] Assign grace period.
2172 Format: <integer>
2174 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2175 Format: <integer>
2177 lockd.nlm_timeout=T [NFS] Assign timeout value.
2178 Format: <integer>
2180 lockd.nlm_udpport=M [NFS] Assign UDP port.
2181 Format: <integer>
2183 locktorture.nreaders_stress= [KNL]
2184 Set the number of locking read-acquisition kthreads.
2185 Defaults to being automatically set based on the
2186 number of online CPUs.
2188 locktorture.nwriters_stress= [KNL]
2189 Set the number of locking write-acquisition kthreads.
2191 locktorture.onoff_holdoff= [KNL]
2192 Set time (s) after boot for CPU-hotplug testing.
2194 locktorture.onoff_interval= [KNL]
2195 Set time (s) between CPU-hotplug operations, or
2196 zero to disable CPU-hotplug testing.
2198 locktorture.shuffle_interval= [KNL]
2199 Set task-shuffle interval (jiffies). Shuffling
2200 tasks allows some CPUs to go into dyntick-idle
2201 mode during the locktorture test.
2203 locktorture.shutdown_secs= [KNL]
2204 Set time (s) after boot system shutdown. This
2205 is useful for hands-off automated testing.
2207 locktorture.stat_interval= [KNL]
2208 Time (s) between statistics printk()s.
2210 locktorture.stutter= [KNL]
2211 Time (s) to stutter testing, for example,
2212 specifying five seconds causes the test to run for
2213 five seconds, wait for five seconds, and so on.
2214 This tests the locking primitive's ability to
2215 transition abruptly to and from idle.
2217 locktorture.torture_runnable= [BOOT]
2218 Start locktorture running at boot time.
2220 locktorture.torture_type= [KNL]
2221 Specify the locking implementation to test.
2223 locktorture.verbose= [KNL]
2224 Enable additional printk() statements.
2226 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2227 Format: <irq>
2229 loglevel= All Kernel Messages with a loglevel smaller than the
2230 console loglevel will be printed to the console. It can
2231 also be changed with klogd or other programs. The
2232 loglevels are defined as follows:
2234 0 (KERN_EMERG) system is unusable
2235 1 (KERN_ALERT) action must be taken immediately
2236 2 (KERN_CRIT) critical conditions
2237 3 (KERN_ERR) error conditions
2238 4 (KERN_WARNING) warning conditions
2239 5 (KERN_NOTICE) normal but significant condition
2240 6 (KERN_INFO) informational
2241 7 (KERN_DEBUG) debug-level messages
2243 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2244 in bytes. n must be a power of two and greater
2245 than the minimal size. The minimal size is defined
2246 by LOG_BUF_SHIFT kernel config parameter. There is
2247 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2248 that allows to increase the default size depending on
2249 the number of CPUs. See init/Kconfig for more details.
2251 logo.nologo [FB] Disables display of the built-in Linux logo.
2252 This may be used to provide more screen space for
2253 kernel log messages and is useful when debugging
2254 kernel boot problems.
2256 lp=0 [LP] Specify parallel ports to use, e.g,
2257 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2258 lp=reset first parallel port). 'lp=0' disables the
2259 lp=auto printer driver. 'lp=reset' (which can be
2260 specified in addition to the ports) causes
2261 attached printers to be reset. Using
2262 lp=port1,port2,... specifies the parallel ports
2263 to associate lp devices with, starting with
2264 lp0. A port specification may be 'none' to skip
2265 that lp device, or a parport name such as
2266 'parport0'. Specifying 'lp=auto' instead of a
2267 port specification list means that device IDs
2268 from each port should be examined, to see if
2269 an IEEE 1284-compliant printer is attached; if
2270 so, the driver will manage that printer.
2271 See also header of drivers/char/lp.c.
2273 lpj=n [KNL]
2274 Sets loops_per_jiffy to given constant, thus avoiding
2275 time-consuming boot-time autodetection (up to 250 ms per
2276 CPU). 0 enables autodetection (default). To determine
2277 the correct value for your kernel, boot with normal
2278 autodetection and see what value is printed. Note that
2279 on SMP systems the preset will be applied to all CPUs,
2280 which is likely to cause problems if your CPUs need
2281 significantly divergent settings. An incorrect value
2282 will cause delays in the kernel to be wrong, leading to
2283 unpredictable I/O errors and other breakage. Although
2284 unlikely, in the extreme case this might damage your
2285 hardware.
2287 ltpc= [NET]
2288 Format: <io>,<irq>,<dma>
2290 machvec= [IA-64] Force the use of a particular machine-vector
2291 (machvec) in a generic kernel.
2292 Example: machvec=hpzx1_swiotlb
2294 machtype= [Loongson] Share the same kernel image file between different
2295 yeeloong laptop.
2296 Example: machtype=lemote-yeeloong-2f-7inch
2298 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2299 than or equal to this physical address is ignored.
2301 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2302 will bring up during bootup. maxcpus=n : n >= 0 limits
2303 the kernel to bring up 'n' processors. Surely after
2304 bootup you can bring up the other plugged cpu by executing
2305 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2306 only takes effect during system bootup.
2307 While n=0 is a special case, it is equivalent to "nosmp",
2308 which also disables the IO APIC.
2310 max_loop= [LOOP] The number of loop block devices that get
2311 (loop.max_loop) unconditionally pre-created at init time. The default
2312 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2313 of statically allocating a predefined number, loop
2314 devices can be requested on-demand with the
2315 /dev/loop-control interface.
2317 mce [X86-32] Machine Check Exception
2319 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2321 md= [HW] RAID subsystems devices and level
2322 See Documentation/md.txt.
2324 mdacon= [MDA]
2325 Format: <first>,<last>
2326 Specifies range of consoles to be captured by the MDA.
2328 mds= [X86,INTEL]
2329 Control mitigation for the Micro-architectural Data
2330 Sampling (MDS) vulnerability.
2332 Certain CPUs are vulnerable to an exploit against CPU
2333 internal buffers which can forward information to a
2334 disclosure gadget under certain conditions.
2336 In vulnerable processors, the speculatively
2337 forwarded data can be used in a cache side channel
2338 attack, to access data to which the attacker does
2339 not have direct access.
2341 This parameter controls the MDS mitigation. The
2342 options are:
2344 full - Enable MDS mitigation on vulnerable CPUs
2345 full,nosmt - Enable MDS mitigation and disable
2346 SMT on vulnerable CPUs
2347 off - Unconditionally disable MDS mitigation
2349 Not specifying this option is equivalent to
2350 mds=full.
2352 For details see: Documentation/hw-vuln/mds.rst
2354 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2355 Amount of memory to be used when the kernel is not able
2356 to see the whole system memory or for test.
2357 [X86] Work as limiting max address. Use together
2358 with memmap= to avoid physical address space collisions.
2359 Without memmap= PCI devices could be placed at addresses
2360 belonging to unused RAM.
2362 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2363 memory.
2365 memchunk=nn[KMG]
2366 [KNL,SH] Allow user to override the default size for
2367 per-device physically contiguous DMA buffers.
2369 memhp_default_state=online/offline
2370 [KNL] Set the initial state for the memory hotplug
2371 onlining policy. If not specified, the default value is
2372 set according to the
2373 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2374 option.
2375 See Documentation/memory-hotplug.txt.
2377 memmap=exactmap [KNL,X86] Enable setting of an exact
2378 E820 memory map, as specified by the user.
2379 Such memmap=exactmap lines can be constructed based on
2380 BIOS output or other requirements. See the memmap=nn@ss
2381 option description.
2383 memmap=nn[KMG]@ss[KMG]
2384 [KNL] Force usage of a specific region of memory.
2385 Region of memory to be used is from ss to ss+nn.
2387 memmap=nn[KMG]#ss[KMG]
2388 [KNL,ACPI] Mark specific memory as ACPI data.
2389 Region of memory to be marked is from ss to ss+nn.
2391 memmap=nn[KMG]$ss[KMG]
2392 [KNL,ACPI] Mark specific memory as reserved.
2393 Region of memory to be reserved is from ss to ss+nn.
2394 Example: Exclude memory from 0x18690000-0x1869ffff
2395 memmap=64K$0x18690000
2396 or
2397 memmap=0x10000$0x18690000
2399 memmap=nn[KMG]!ss[KMG]
2400 [KNL,X86] Mark specific memory as protected.
2401 Region of memory to be used, from ss to ss+nn.
2402 The memory region may be marked as e820 type 12 (0xc)
2403 and is NVDIMM or ADR memory.
2405 memory_corruption_check=0/1 [X86]
2406 Some BIOSes seem to corrupt the first 64k of
2407 memory when doing things like suspend/resume.
2408 Setting this option will scan the memory
2409 looking for corruption. Enabling this will
2410 both detect corruption and prevent the kernel
2411 from using the memory being corrupted.
2412 However, its intended as a diagnostic tool; if
2413 repeatable BIOS-originated corruption always
2414 affects the same memory, you can use memmap=
2415 to prevent the kernel from using that memory.
2417 memory_corruption_check_size=size [X86]
2418 By default it checks for corruption in the low
2419 64k, making this memory unavailable for normal
2420 use. Use this parameter to scan for
2421 corruption in more or less memory.
2423 memory_corruption_check_period=seconds [X86]
2424 By default it checks for corruption every 60
2425 seconds. Use this parameter to check at some
2426 other rate. 0 disables periodic checking.
2428 memtest= [KNL,X86,ARM] Enable memtest
2429 Format: <integer>
2430 default : 0 <disable>
2431 Specifies the number of memtest passes to be
2432 performed. Each pass selects another test
2433 pattern from a given set of patterns. Memtest
2434 fills the memory with this pattern, validates
2435 memory contents and reserves bad memory
2436 regions that are detected.
2438 meye.*= [HW] Set MotionEye Camera parameters
2439 See Documentation/video4linux/meye.txt.
2441 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2442 Multi-Function General Purpose Timers on AMD Geode
2443 platforms.
2445 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2446 the BIOS has incorrectly applied a workaround. TinyBIOS
2447 version 0.98 is known to be affected, 0.99 fixes the
2448 problem by letting the user disable the workaround.
2450 mga= [HW,DRM]
2452 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2453 physical address is ignored.
2455 mini2440= [ARM,HW,KNL]
2456 Format:[0..2][b][c][t]
2457 Default: "0tb"
2458 MINI2440 configuration specification:
2459 0 - The attached screen is the 3.5" TFT
2460 1 - The attached screen is the 7" TFT
2461 2 - The VGA Shield is attached (1024x768)
2462 Leaving out the screen size parameter will not load
2463 the TFT driver, and the framebuffer will be left
2464 unconfigured.
2465 b - Enable backlight. The TFT backlight pin will be
2466 linked to the kernel VESA blanking code and a GPIO
2467 LED. This parameter is not necessary when using the
2468 VGA shield.
2469 c - Enable the s3c camera interface.
2470 t - Reserved for enabling touchscreen support. The
2471 touchscreen support is not enabled in the mainstream
2472 kernel as of 2.6.30, a preliminary port can be found
2473 in the "bleeding edge" mini2440 support kernel at
2474 http://repo.or.cz/w/linux-2.6/mini2440.git
2476 mitigations=
2477 [X86] Control optional mitigations for CPU
2478 vulnerabilities. This is a set of curated,
2479 arch-independent options, each of which is an
2480 aggregation of existing arch-specific options.
2482 off
2483 Disable all optional CPU mitigations. This
2484 improves system performance, but it may also
2485 expose users to several CPU vulnerabilities.
2486 Equivalent to: nopti [X86]
2487 nospectre_v1 [X86]
2488 nospectre_v2 [X86]
2489 spectre_v2_user=off [X86]
2490 spec_store_bypass_disable=off [X86]
2491 l1tf=off [X86]
2492 mds=off [X86]
2494 auto (default)
2495 Mitigate all CPU vulnerabilities, but leave SMT
2496 enabled, even if it's vulnerable. This is for
2497 users who don't want to be surprised by SMT
2498 getting disabled across kernel upgrades, or who
2499 have other ways of avoiding SMT-based attacks.
2500 Equivalent to: (default behavior)
2502 auto,nosmt
2503 Mitigate all CPU vulnerabilities, disabling SMT
2504 if needed. This is for users who always want to
2505 be fully mitigated, even if it means losing SMT.
2506 Equivalent to: l1tf=flush,nosmt [X86]
2507 mds=full,nosmt [X86]
2509 mminit_loglevel=
2510 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2511 parameter allows control of the logging verbosity for
2512 the additional memory initialisation checks. A value
2513 of 0 disables mminit logging and a level of 4 will
2514 log everything. Information is printed at KERN_DEBUG
2515 so loglevel=8 may also need to be specified.
2517 module.sig_enforce
2518 [KNL] When CONFIG_MODULE_SIG is set, this means that
2519 modules without (valid) signatures will fail to load.
2520 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2521 is always true, so this option does nothing.
2523 module_blacklist= [KNL] Do not load a comma-separated list of
2524 modules. Useful for debugging problem modules.
2526 mousedev.tap_time=
2527 [MOUSE] Maximum time between finger touching and
2528 leaving touchpad surface for touch to be considered
2529 a tap and be reported as a left button click (for
2530 touchpads working in absolute mode only).
2531 Format: <msecs>
2532 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2533 reporting absolute coordinates, such as tablets
2534 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2535 reporting absolute coordinates, such as tablets
2537 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
2538 is similar to kernelcore except it specifies the
2539 amount of memory used for migratable allocations.
2540 If both kernelcore and movablecore is specified,
2541 then kernelcore will be at *least* the specified
2542 value but may be more. If movablecore on its own
2543 is specified, the administrator must be careful
2544 that the amount of memory usable for all allocations
2545 is not too small.
2547 movable_node [KNL,X86] Boot-time switch to enable the effects
2548 of CONFIG_MOVABLE_NODE=y. See mm/Kconfig for details.
2550 MTD_Partition= [MTD]
2551 Format: <name>,<region-number>,<size>,<offset>
2553 MTD_Region= [MTD] Format:
2554 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2556 mtdparts= [MTD]
2557 See drivers/mtd/cmdlinepart.c.
2559 multitce=off [PPC] This parameter disables the use of the pSeries
2560 firmware feature for updating multiple TCE entries
2561 at a time.
2563 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2565 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2567 boundary - index of last SLC block on Flex-OneNAND.
2568 The remaining blocks are configured as MLC blocks.
2569 lock - Configure if Flex-OneNAND boundary should be locked.
2570 Once locked, the boundary cannot be changed.
2571 1 indicates lock status, 0 indicates unlock status.
2573 mtdset= [ARM]
2574 ARM/S3C2412 JIVE boot control
2576 See arch/arm/mach-s3c2412/mach-jive.c
2578 mtouchusb.raw_coordinates=
2579 [HW] Make the MicroTouch USB driver use raw coordinates
2580 ('y', default) or cooked coordinates ('n')
2582 mtrr_chunk_size=nn[KMG] [X86]
2583 used for mtrr cleanup. It is largest continuous chunk
2584 that could hold holes aka. UC entries.
2586 mtrr_gran_size=nn[KMG] [X86]
2587 Used for mtrr cleanup. It is granularity of mtrr block.
2588 Default is 1.
2589 Large value could prevent small alignment from
2590 using up MTRRs.
2592 mtrr_spare_reg_nr=n [X86]
2593 Format: <integer>
2594 Range: 0,7 : spare reg number
2595 Default : 1
2596 Used for mtrr cleanup. It is spare mtrr entries number.
2597 Set to 2 or more if your graphical card needs more.
2599 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2601 netdev= [NET] Network devices parameters
2602 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2603 Note that mem_start is often overloaded to mean
2604 something different and driver-specific.
2605 This usage is only documented in each driver source
2606 file if at all.
2608 nf_conntrack.acct=
2609 [NETFILTER] Enable connection tracking flow accounting
2610 0 to disable accounting
2611 1 to enable accounting
2612 Default value is 0.
2614 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2615 See Documentation/filesystems/nfs/nfsroot.txt.
2617 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2618 See Documentation/filesystems/nfs/nfsroot.txt.
2620 nfsrootdebug [NFS] enable nfsroot debugging messages.
2621 See Documentation/filesystems/nfs/nfsroot.txt.
2623 nfs.callback_nr_threads=
2624 [NFSv4] set the total number of threads that the
2625 NFS client will assign to service NFSv4 callback
2626 requests.
2628 nfs.callback_tcpport=
2629 [NFS] set the TCP port on which the NFSv4 callback
2630 channel should listen.
2632 nfs.cache_getent=
2633 [NFS] sets the pathname to the program which is used
2634 to update the NFS client cache entries.
2636 nfs.cache_getent_timeout=
2637 [NFS] sets the timeout after which an attempt to
2638 update a cache entry is deemed to have failed.
2640 nfs.idmap_cache_timeout=
2641 [NFS] set the maximum lifetime for idmapper cache
2642 entries.
2644 nfs.enable_ino64=
2645 [NFS] enable 64-bit inode numbers.
2646 If zero, the NFS client will fake up a 32-bit inode
2647 number for the readdir() and stat() syscalls instead
2648 of returning the full 64-bit number.
2649 The default is to return 64-bit inode numbers.
2651 nfs.max_session_cb_slots=
2652 [NFSv4.1] Sets the maximum number of session
2653 slots the client will assign to the callback
2654 channel. This determines the maximum number of
2655 callbacks the client will process in parallel for
2656 a particular server.
2658 nfs.max_session_slots=
2659 [NFSv4.1] Sets the maximum number of session slots
2660 the client will attempt to negotiate with the server.
2661 This limits the number of simultaneous RPC requests
2662 that the client can send to the NFSv4.1 server.
2663 Note that there is little point in setting this
2664 value higher than the max_tcp_slot_table_limit.
2666 nfs.nfs4_disable_idmapping=
2667 [NFSv4] When set to the default of '1', this option
2668 ensures that both the RPC level authentication
2669 scheme and the NFS level operations agree to use
2670 numeric uids/gids if the mount is using the
2671 'sec=sys' security flavour. In effect it is
2672 disabling idmapping, which can make migration from
2673 legacy NFSv2/v3 systems to NFSv4 easier.
2674 Servers that do not support this mode of operation
2675 will be autodetected by the client, and it will fall
2676 back to using the idmapper.
2677 To turn off this behaviour, set the value to '0'.
2678 nfs.nfs4_unique_id=
2679 [NFS4] Specify an additional fixed unique ident-
2680 ification string that NFSv4 clients can insert into
2681 their nfs_client_id4 string. This is typically a
2682 UUID that is generated at system install time.
2684 nfs.send_implementation_id =
2685 [NFSv4.1] Send client implementation identification
2686 information in exchange_id requests.
2687 If zero, no implementation identification information
2688 will be sent.
2689 The default is to send the implementation identification
2690 information.
2692 nfs.recover_lost_locks =
2693 [NFSv4] Attempt to recover locks that were lost due
2694 to a lease timeout on the server. Please note that
2695 doing this risks data corruption, since there are
2696 no guarantees that the file will remain unchanged
2697 after the locks are lost.
2698 If you want to enable the kernel legacy behaviour of
2699 attempting to recover these locks, then set this
2700 parameter to '1'.
2701 The default parameter value of '0' causes the kernel
2702 not to attempt recovery of lost locks.
2704 nfs4.layoutstats_timer =
2705 [NFSv4.2] Change the rate at which the kernel sends
2706 layoutstats to the pNFS metadata server.
2708 Setting this to value to 0 causes the kernel to use
2709 whatever value is the default set by the layout
2710 driver. A non-zero value sets the minimum interval
2711 in seconds between layoutstats transmissions.
2713 nfsd.nfs4_disable_idmapping=
2714 [NFSv4] When set to the default of '1', the NFSv4
2715 server will return only numeric uids and gids to
2716 clients using auth_sys, and will accept numeric uids
2717 and gids from such clients. This is intended to ease
2718 migration from NFSv2/v3.
2720 objlayoutdriver.osd_login_prog=
2721 [NFS] [OBJLAYOUT] sets the pathname to the program which
2722 is used to automatically discover and login into new
2723 osd-targets. Please see:
2724 Documentation/filesystems/pnfs.txt for more explanations
2726 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
2727 when a NMI is triggered.
2728 Format: [state][,regs][,debounce][,die]
2730 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2731 Format: [panic,][nopanic,][num]
2732 Valid num: 0 or 1
2733 0 - turn hardlockup detector in nmi_watchdog off
2734 1 - turn hardlockup detector in nmi_watchdog on
2735 When panic is specified, panic when an NMI watchdog
2736 timeout occurs (or 'nopanic' to override the opposite
2737 default). To disable both hard and soft lockup detectors,
2738 please see 'nowatchdog'.
2739 This is useful when you use a panic=... timeout and
2740 need the box quickly up again.
2742 netpoll.carrier_timeout=
2743 [NET] Specifies amount of time (in seconds) that
2744 netpoll should wait for a carrier. By default netpoll
2745 waits 4 seconds.
2747 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2748 emulation library even if a 387 maths coprocessor
2749 is present.
2751 no_console_suspend
2752 [HW] Never suspend the console
2753 Disable suspending of consoles during suspend and
2754 hibernate operations. Once disabled, debugging
2755 messages can reach various consoles while the rest
2756 of the system is being put to sleep (ie, while
2757 debugging driver suspend/resume hooks). This may
2758 not work reliably with all consoles, but is known
2759 to work with serial and VGA consoles.
2760 To facilitate more flexible debugging, we also add
2761 console_suspend, a printk module parameter to control
2762 it. Users could use console_suspend (usually
2763 /sys/module/printk/parameters/console_suspend) to
2764 turn on/off it dynamically.
2766 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2767 caches in the slab allocator. Saves per-node memory,
2768 but will impact performance.
2770 noalign [KNL,ARM]
2772 noaltinstr [S390] Disables alternative instructions patching
2773 (CPU alternatives feature).
2775 noapic [SMP,APIC] Tells the kernel to not make use of any
2776 IOAPICs that may be present in the system.
2778 noautogroup Disable scheduler automatic task group creation.
2780 nobats [PPC] Do not use BATs for mapping kernel lowmem
2781 on "Classic" PPC cores.
2783 nocache [ARM]
2785 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2787 nodelayacct [KNL] Disable per-task delay accounting
2789 nodsp [SH] Disable hardware DSP at boot time.
2791 noefi Disable EFI runtime services support.
2793 noexec [IA-64]
2795 noexec [X86]
2796 On X86-32 available only on PAE configured kernels.
2797 noexec=on: enable non-executable mappings (default)
2798 noexec=off: disable non-executable mappings
2800 nosmap [X86]
2801 Disable SMAP (Supervisor Mode Access Prevention)
2802 even if it is supported by processor.
2804 nosmep [X86]
2805 Disable SMEP (Supervisor Mode Execution Prevention)
2806 even if it is supported by processor.
2808 noexec32 [X86-64]
2809 This affects only 32-bit executables.
2810 noexec32=on: enable non-executable mappings (default)
2811 read doesn't imply executable mappings
2812 noexec32=off: disable non-executable mappings
2813 read implies executable mappings
2815 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2817 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2818 register save and restore. The kernel will only save
2819 legacy floating-point registers on task switch.
2821 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2823 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2824 Equivalent to smt=1.
2826 [KNL,x86] Disable symmetric multithreading (SMT).
2827 nosmt=force: Force disable SMT, cannot be undone
2828 via the sysfs control file.
2830 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
2831 (bounds check bypass). With this option data leaks are
2832 possible in the system.
2834 nospectre_v2 [X86,PPC_FSL_BOOK3E] Disable all mitigations for the Spectre variant 2
2835 (indirect branch prediction) vulnerability. System may
2836 allow data leaks with this option, which is equivalent
2837 to spectre_v2=off.
2839 nospec_store_bypass_disable
2840 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
2842 noxsave [BUGS=X86] Disables x86 extended register state save
2843 and restore using xsave. The kernel will fallback to
2844 enabling legacy floating-point and sse state.
2846 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2847 register states. The kernel will fall back to use
2848 xsave to save the states. By using this parameter,
2849 performance of saving the states is degraded because
2850 xsave doesn't support modified optimization while
2851 xsaveopt supports it on xsaveopt enabled systems.
2853 noxsaves [X86] Disables xsaves and xrstors used in saving and
2854 restoring x86 extended register state in compacted
2855 form of xsave area. The kernel will fall back to use
2856 xsaveopt and xrstor to save and restore the states
2857 in standard form of xsave area. By using this
2858 parameter, xsave area per process might occupy more
2859 memory on xsaves enabled systems.
2861 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2862 wfi(ARM) instruction doesn't work correctly and not to
2863 use it. This is also useful when using JTAG debugger.
2865 no_file_caps Tells the kernel not to honor file capabilities. The
2866 only way then for a file to be executed with privilege
2867 is to be setuid root or executed by root.
2869 nohalt [IA-64] Tells the kernel not to use the power saving
2870 function PAL_HALT_LIGHT when idle. This increases
2871 power-consumption. On the positive side, it reduces
2872 interrupt wake-up latency, which may improve performance
2873 in certain environments such as networked servers or
2874 real-time systems.
2876 nohibernate [HIBERNATION] Disable hibernation and resume.
2878 nohz= [KNL] Boottime enable/disable dynamic ticks
2879 Valid arguments: on, off
2880 Default: on
2882 nohz_full= [KNL,BOOT]
2883 The argument is a cpu list, as described above.
2884 In kernels built with CONFIG_NO_HZ_FULL=y, set
2885 the specified list of CPUs whose tick will be stopped
2886 whenever possible. The boot CPU will be forced outside
2887 the range to maintain the timekeeping.
2888 The CPUs in this range must also be included in the
2889 rcu_nocbs= set.
2891 noiotrap [SH] Disables trapped I/O port accesses.
2893 noirqdebug [X86-32] Disables the code which attempts to detect and
2894 disable unhandled interrupt sources.
2896 no_timer_check [X86,APIC] Disables the code which tests for
2897 broken timer IRQ sources.
2899 noisapnp [ISAPNP] Disables ISA PnP code.
2901 noinitrd [RAM] Tells the kernel not to load any configured
2902 initial RAM disk.
2904 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2905 remapping.
2906 [Deprecated - use intremap=off]
2908 nointroute [IA-64]
2910 noinvpcid [X86] Disable the INVPCID cpu feature.
2912 nojitter [IA-64] Disables jitter checking for ITC timers.
2914 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2916 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2917 fault handling.
2919 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2920 steal time is computed, but won't influence scheduler
2921 behaviour
2923 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2925 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2927 noltlbs [PPC] Do not use large page/tlb entries for kernel
2928 lowmem mapping on PPC40x and PPC8xx
2930 nomca [IA-64] Disable machine check abort handling
2932 nomce [X86-32] Disable Machine Check Exception
2934 nomfgpt [X86-32] Disable Multi-Function General Purpose
2935 Timer usage (for AMD Geode machines).
2937 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2938 shutdown the other cpus. Instead use the REBOOT_VECTOR
2939 irq.
2941 nomodule Disable module load
2943 nopat [X86] Disable PAT (page attribute table extension of
2944 pagetables) support.
2946 nopcid [X86-64] Disable the PCID cpu feature.
2948 norandmaps Don't use address space randomization. Equivalent to
2949 echo 0 > /proc/sys/kernel/randomize_va_space
2951 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2952 with UP alternatives
2954 nordrand [X86] Disable kernel use of the RDRAND and
2955 RDSEED instructions even if they are supported
2956 by the processor. RDRAND and RDSEED are still
2957 available to user space applications.
2959 noresume [SWSUSP] Disables resume and restores original swap
2960 space.
2962 no-scroll [VGA] Disables scrollback.
2963 This is required for the Braillex ib80-piezo Braille
2964 reader made by F.H. Papenmeier (Germany).
2966 nosbagart [IA-64]
2968 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2970 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2971 and disable the IO APIC. legacy for "maxcpus=0".
2973 nosoftlockup [KNL] Disable the soft-lockup detector.
2975 nosync [HW,M68K] Disables sync negotiation for all devices.
2977 notsc [BUGS=X86-32] Disable Time Stamp Counter
2979 nowatchdog [KNL] Disable both lockup detectors, i.e.
2980 soft-lockup and NMI watchdog (hard-lockup).
2982 nowb [ARM]
2984 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2986 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2987 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2988 Some features depend on CPU0. Known dependencies are:
2989 1. Resume from suspend/hibernate depends on CPU0.
2990 Suspend/hibernate will fail if CPU0 is offline and you
2991 need to online CPU0 before suspend/hibernate.
2992 2. PIC interrupts also depend on CPU0. CPU0 can't be
2993 removed if a PIC interrupt is detected.
2994 It's said poweroff/reboot may depend on CPU0 on some
2995 machines although I haven't seen such issues so far
2996 after CPU0 is offline on a few tested machines.
2997 If the dependencies are under your control, you can
2998 turn on cpu0_hotplug.
3000 nptcg= [IA-64] Override max number of concurrent global TLB
3001 purges which is reported from either PAL_VM_SUMMARY or
3002 SAL PALO.
3004 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3005 could support. nr_cpus=n : n >= 1 limits the kernel to
3006 support 'n' processors. It could be larger than the
3007 number of already plugged CPU during bootup, later in
3008 runtime you can physically add extra cpu until it reaches
3009 n. So during boot up some boot time memory for per-cpu
3010 variables need be pre-allocated for later physical cpu
3011 hot plugging.
3013 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3015 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3016 Allowed values are enable and disable
3018 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3019 one of ['zone', 'node', 'default'] can be specified
3020 This can be set from sysctl after boot.
3021 See Documentation/sysctl/vm.txt for details.
3023 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3024 See Documentation/debugging-via-ohci1394.txt for more
3025 info.
3027 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3028 Rather than timing out after 20 ms if an EC
3029 command is not properly ACKed, override the length
3030 of the timeout. We have interrupts disabled while
3031 waiting for the ACK, so if this is set too high
3032 interrupts *may* be lost!
3034 omap_mux= [OMAP] Override bootloader pin multiplexing.
3035 Format: <mux_mode0.mode_name=value>...
3036 For example, to override I2C bus2:
3037 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3039 oprofile.timer= [HW]
3040 Use timer interrupt instead of performance counters
3042 oprofile.cpu_type= Force an oprofile cpu type
3043 This might be useful if you have an older oprofile
3044 userland or if you want common events.
3045 Format: { arch_perfmon }
3046 arch_perfmon: [X86] Force use of architectural
3047 perfmon on Intel CPUs instead of the
3048 CPU specific event set.
3049 timer: [X86] Force use of architectural NMI
3050 timer mode (see also oprofile.timer
3051 for generic hr timer mode)
3053 oops=panic Always panic on oopses. Default is to just kill the
3054 process, but there is a small probability of
3055 deadlocking the machine.
3056 This will also cause panics on machine check exceptions.
3057 Useful together with panic=30 to trigger a reboot.
3059 OSS [HW,OSS]
3060 See Documentation/sound/oss/oss-parameters.txt
3062 page_owner= [KNL] Boot-time page_owner enabling option.
3063 Storage of the information about who allocated
3064 each page is disabled in default. With this switch,
3065 we can turn it on.
3066 on: enable the feature
3068 page_poison= [KNL] Boot-time parameter changing the state of
3069 poisoning on the buddy allocator.
3070 off: turn off poisoning
3071 on: turn on poisoning
3073 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3074 timeout > 0: seconds before rebooting
3075 timeout = 0: wait forever
3076 timeout < 0: reboot immediately
3077 Format: <timeout>
3079 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3080 on a WARN().
3082 crash_kexec_post_notifiers
3083 Run kdump after running panic-notifiers and dumping
3084 kmsg. This only for the users who doubt kdump always
3085 succeeds in any situation.
3086 Note that this also increases risks of kdump failure,
3087 because some panic notifiers can make the crashed
3088 kernel more unstable.
3090 parkbd.port= [HW] Parallel port number the keyboard adapter is
3091 connected to, default is 0.
3092 Format: <parport#>
3093 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3094 0 for XT, 1 for AT (default is AT).
3095 Format: <mode>
3097 parport= [HW,PPT] Specify parallel ports. 0 disables.
3098 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3099 Use 'auto' to force the driver to use any
3100 IRQ/DMA settings detected (the default is to
3101 ignore detected IRQ/DMA settings because of
3102 possible conflicts). You can specify the base
3103 address, IRQ, and DMA settings; IRQ and DMA
3104 should be numbers, or 'auto' (for using detected
3105 settings on that particular port), or 'nofifo'
3106 (to avoid using a FIFO even if it is detected).
3107 Parallel ports are assigned in the order they
3108 are specified on the command line, starting
3109 with parport0.
3111 parport_init_mode= [HW,PPT]
3112 Configure VIA parallel port to operate in
3113 a specific mode. This is necessary on Pegasos
3114 computer where firmware has no options for setting
3115 up parallel port mode and sets it to spp.
3116 Currently this function knows 686a and 8231 chips.
3117 Format: [spp|ps2|epp|ecp|ecpepp]
3119 pause_on_oops=
3120 Halt all CPUs after the first oops has been printed for
3121 the specified number of seconds. This is to be used if
3122 your oopses keep scrolling off the screen.
3124 pcbit= [HW,ISDN]
3126 pcd. [PARIDE]
3127 See header of drivers/block/paride/pcd.c.
3128 See also Documentation/blockdev/paride.txt.
3130 pci=option[,option...] [PCI] various PCI subsystem options:
3131 earlydump [X86] dump PCI config space before the kernel
3132 changes anything
3133 off [X86] don't probe for the PCI bus
3134 bios [X86-32] force use of PCI BIOS, don't access
3135 the hardware directly. Use this if your machine
3136 has a non-standard PCI host bridge.
3137 nobios [X86-32] disallow use of PCI BIOS, only direct
3138 hardware access methods are allowed. Use this
3139 if you experience crashes upon bootup and you
3140 suspect they are caused by the BIOS.
3141 conf1 [X86] Force use of PCI Configuration Access
3142 Mechanism 1 (config address in IO port 0xCF8,
3143 data in IO port 0xCFC, both 32-bit).
3144 conf2 [X86] Force use of PCI Configuration Access
3145 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3146 the function, IO port 0xCFA, also 8-bit, sets
3147 bus number. The config space is then accessed
3148 through ports 0xC000-0xCFFF).
3149 See http://wiki.osdev.org/PCI for more info
3150 on the configuration access mechanisms.
3151 noaer [PCIE] If the PCIEAER kernel config parameter is
3152 enabled, this kernel boot option can be used to
3153 disable the use of PCIE advanced error reporting.
3154 nodomains [PCI] Disable support for multiple PCI
3155 root domains (aka PCI segments, in ACPI-speak).
3156 nommconf [X86] Disable use of MMCONFIG for PCI
3157 Configuration
3158 check_enable_amd_mmconf [X86] check for and enable
3159 properly configured MMIO access to PCI
3160 config space on AMD family 10h CPU
3161 nomsi [MSI] If the PCI_MSI kernel config parameter is
3162 enabled, this kernel boot option can be used to
3163 disable the use of MSI interrupts system-wide.
3164 noioapicquirk [APIC] Disable all boot interrupt quirks.
3165 Safety option to keep boot IRQs enabled. This
3166 should never be necessary.
3167 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3168 primary IO-APIC for bridges that cannot disable
3169 boot IRQs. This fixes a source of spurious IRQs
3170 when the system masks IRQs.
3171 noioapicreroute [APIC] Disable workaround that uses the
3172 boot IRQ equivalent of an IRQ that connects to
3173 a chipset where boot IRQs cannot be disabled.
3174 The opposite of ioapicreroute.
3175 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3176 routing table. These calls are known to be buggy
3177 on several machines and they hang the machine
3178 when used, but on other computers it's the only
3179 way to get the interrupt routing table. Try
3180 this option if the kernel is unable to allocate
3181 IRQs or discover secondary PCI buses on your
3182 motherboard.
3183 rom [X86] Assign address space to expansion ROMs.
3184 Use with caution as certain devices share
3185 address decoders between ROMs and other
3186 resources.
3187 norom [X86] Do not assign address space to
3188 expansion ROMs that do not already have
3189 BIOS assigned address ranges.
3190 nobar [X86] Do not assign address space to the
3191 BARs that weren't assigned by the BIOS.
3192 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3193 assigned automatically to PCI devices. You can
3194 make the kernel exclude IRQs of your ISA cards
3195 this way.
3196 pirqaddr=0xAAAAA [X86] Specify the physical address
3197 of the PIRQ table (normally generated
3198 by the BIOS) if it is outside the
3199 F0000h-100000h range.
3200 lastbus=N [X86] Scan all buses thru bus #N. Can be
3201 useful if the kernel is unable to find your
3202 secondary buses and you want to tell it
3203 explicitly which ones they are.
3204 assign-busses [X86] Always assign all PCI bus
3205 numbers ourselves, overriding
3206 whatever the firmware may have done.
3207 usepirqmask [X86] Honor the possible IRQ mask stored
3208 in the BIOS $PIR table. This is needed on
3209 some systems with broken BIOSes, notably
3210 some HP Pavilion N5400 and Omnibook XE3
3211 notebooks. This will have no effect if ACPI
3212 IRQ routing is enabled.
3213 noacpi [X86] Do not use ACPI for IRQ routing
3214 or for PCI scanning.
3215 use_crs [X86] Use PCI host bridge window information
3216 from ACPI. On BIOSes from 2008 or later, this
3217 is enabled by default. If you need to use this,
3218 please report a bug.
3219 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3220 If you need to use this, please report a bug.
3221 routeirq Do IRQ routing for all PCI devices.
3222 This is normally done in pci_enable_device(),
3223 so this option is a temporary workaround
3224 for broken drivers that don't call it.
3225 skip_isa_align [X86] do not align io start addr, so can
3226 handle more pci cards
3227 noearly [X86] Don't do any early type 1 scanning.
3228 This might help on some broken boards which
3229 machine check when some devices' config space
3230 is read. But various workarounds are disabled
3231 and some IOMMU drivers will not work.
3232 bfsort Sort PCI devices into breadth-first order.
3233 This sorting is done to get a device
3234 order compatible with older (<= 2.4) kernels.
3235 nobfsort Don't sort PCI devices into breadth-first order.
3236 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3237 tuning and use the BIOS-configured MPS defaults.
3238 pcie_bus_safe Set every device's MPS to the largest value
3239 supported by all devices below the root complex.
3240 pcie_bus_perf Set device MPS to the largest allowable MPS
3241 based on its parent bus. Also set MRRS (Max
3242 Read Request Size) to the largest supported
3243 value (no larger than the MPS that the device
3244 or bus can support) for best performance.
3245 pcie_bus_peer2peer Set every device's MPS to 128B, which
3246 every device is guaranteed to support. This
3247 configuration allows peer-to-peer DMA between
3248 any pair of devices, possibly at the cost of
3249 reduced performance. This also guarantees
3250 that hot-added devices will work.
3251 cbiosize=nn[KMG] The fixed amount of bus space which is
3252 reserved for the CardBus bridge's IO window.
3253 The default value is 256 bytes.
3254 cbmemsize=nn[KMG] The fixed amount of bus space which is
3255 reserved for the CardBus bridge's memory
3256 window. The default value is 64 megabytes.
3257 resource_alignment=
3258 Format:
3259 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
3260 [<order of align>@]pci:<vendor>:<device>\
3261 [:<subvendor>:<subdevice>][; ...]
3262 Specifies alignment and device to reassign
3263 aligned memory resources.
3264 If <order of align> is not specified,
3265 PAGE_SIZE is used as alignment.
3266 PCI-PCI bridge can be specified, if resource
3267 windows need to be expanded.
3268 To specify the alignment for several
3269 instances of a device, the PCI vendor,
3270 device, subvendor, and subdevice may be
3271 specified, e.g., 4096@pci:8086:9c22:103c:198f
3272 ecrc= Enable/disable PCIe ECRC (transaction layer
3273 end-to-end CRC checking).
3274 bios: Use BIOS/firmware settings. This is the
3275 the default.
3276 off: Turn ECRC off
3277 on: Turn ECRC on.
3278 hpiosize=nn[KMG] The fixed amount of bus space which is
3279 reserved for hotplug bridge's IO window.
3280 Default size is 256 bytes.
3281 hpmemsize=nn[KMG] The fixed amount of bus space which is
3282 reserved for hotplug bridge's memory window.
3283 Default size is 2 megabytes.
3284 hpbussize=nn The minimum amount of additional bus numbers
3285 reserved for buses below a hotplug bridge.
3286 Default is 1.
3287 realloc= Enable/disable reallocating PCI bridge resources
3288 if allocations done by BIOS are too small to
3289 accommodate resources required by all child
3290 devices.
3291 off: Turn realloc off
3292 on: Turn realloc on
3293 realloc same as realloc=on
3294 noari do not use PCIe ARI.
3295 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3296 only look for one device below a PCIe downstream
3297 port.
3299 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3300 Management.
3301 off Disable ASPM.
3302 force Enable ASPM even on devices that claim not to support it.
3303 WARNING: Forcing ASPM on may cause system lockups.
3305 pcie_hp= [PCIE] PCI Express Hotplug driver options:
3306 nomsi Do not use MSI for PCI Express Native Hotplug (this
3307 makes all PCIe ports use INTx for hotplug services).
3309 pcie_ports= [PCIE] PCIe ports handling:
3310 auto Ask the BIOS whether or not to use native PCIe services
3311 associated with PCIe ports (PME, hot-plug, AER). Use
3312 them only if that is allowed by the BIOS.
3313 native Use native PCIe services associated with PCIe ports
3314 unconditionally.
3315 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
3316 ports driver.
3318 pcie_port_pm= [PCIE] PCIe port power management handling:
3319 off Disable power management of all PCIe ports
3320 force Forcibly enable power management of all PCIe ports
3322 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3323 nomsi Do not use MSI for native PCIe PME signaling (this makes
3324 all PCIe root ports use INTx for all services).
3326 pcmv= [HW,PCMCIA] BadgePAD 4
3328 pd_ignore_unused
3329 [PM]
3330 Keep all power-domains already enabled by bootloader on,
3331 even if no driver has claimed them. This is useful
3332 for debug and development, but should not be
3333 needed on a platform with proper driver support.
3335 pd. [PARIDE]
3336 See Documentation/blockdev/paride.txt.
3338 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3339 boot time.
3340 Format: { 0 | 1 }
3341 See arch/parisc/kernel/pdc_chassis.c
3343 percpu_alloc= Select which percpu first chunk allocator to use.
3344 Currently supported values are "embed" and "page".
3345 Archs may support subset or none of the selections.
3346 See comments in mm/percpu.c for details on each
3347 allocator. This parameter is primarily for debugging
3348 and performance comparison.
3350 pf. [PARIDE]
3351 See Documentation/blockdev/paride.txt.
3353 pg. [PARIDE]
3354 See Documentation/blockdev/paride.txt.
3356 pirq= [SMP,APIC] Manual mp-table setup
3357 See Documentation/x86/i386/IO-APIC.txt.
3359 plip= [PPT,NET] Parallel port network link
3360 Format: { parport<nr> | timid | 0 }
3361 See also Documentation/parport.txt.
3363 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3364 Override pmtimer IOPort with a hex value.
3365 e.g. pmtmr=0x508
3367 pnp.debug=1 [PNP]
3368 Enable PNP debug messages (depends on the
3369 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3370 via /sys/module/pnp/parameters/debug. We always show
3371 current resource usage; turning this on also shows
3372 possible settings and some assignment information.
3374 pnpacpi= [ACPI]
3375 { off }
3377 pnpbios= [ISAPNP]
3378 { on | off | curr | res | no-curr | no-res }
3380 pnp_reserve_irq=
3381 [ISAPNP] Exclude IRQs for the autoconfiguration
3383 pnp_reserve_dma=
3384 [ISAPNP] Exclude DMAs for the autoconfiguration
3386 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3387 Ranges are in pairs (I/O port base and size).
3389 pnp_reserve_mem=
3390 [ISAPNP] Exclude memory regions for the
3391 autoconfiguration.
3392 Ranges are in pairs (memory base and size).
3394 ports= [IP_VS_FTP] IPVS ftp helper module
3395 Default is 21.
3396 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3397 may be specified.
3398 Format: <port>,<port>....
3400 ppc_strict_facility_enable
3401 [PPC] This option catches any kernel floating point,
3402 Altivec, VSX and SPE outside of regions specifically
3403 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3404 There is some performance impact when enabling this.
3406 print-fatal-signals=
3407 [KNL] debug: print fatal signals
3409 If enabled, warn about various signal handling
3410 related application anomalies: too many signals,
3411 too many POSIX.1 timers, fatal signals causing a
3412 coredump - etc.
3414 If you hit the warning due to signal overflow,
3415 you might want to try "ulimit -i unlimited".
3417 default: off.
3419 printk.always_kmsg_dump=
3420 Trigger kmsg_dump for cases other than kernel oops or
3421 panics
3422 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3423 default: disabled
3425 printk.devkmsg={on,off,ratelimit}
3426 Control writing to /dev/kmsg.
3427 on - unlimited logging to /dev/kmsg from userspace
3428 off - logging to /dev/kmsg disabled
3429 ratelimit - ratelimit the logging
3430 Default: ratelimit
3432 printk.time= Show timing data prefixed to each printk message line
3433 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3435 processor.max_cstate= [HW,ACPI]
3436 Limit processor to maximum C-state
3437 max_cstate=9 overrides any DMI blacklist limit.
3439 processor.nocst [HW,ACPI]
3440 Ignore the _CST method to determine C-states,
3441 instead using the legacy FADT method
3443 profile= [KNL] Enable kernel profiling via /proc/profile
3444 Format: [schedule,]<number>
3445 Param: "schedule" - profile schedule points.
3446 Param: <number> - step/bucket size as a power of 2 for
3447 statistical time based profiling.
3448 Param: "sleep" - profile D-state sleeping (millisecs).
3449 Requires CONFIG_SCHEDSTATS
3450 Param: "kvm" - profile VM exits.
3452 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3453 before loading.
3454 See Documentation/blockdev/ramdisk.txt.
3456 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3457 probe for; one of (bare|imps|exps|lifebook|any).
3458 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3459 per second.
3460 psmouse.resetafter= [HW,MOUSE]
3461 Try to reset the device after so many bad packets
3462 (0 = never).
3463 psmouse.resolution=
3464 [HW,MOUSE] Set desired mouse resolution, in dpi.
3465 psmouse.smartscroll=
3466 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3467 0 = disabled, 1 = enabled (default).
3469 pstore.backend= Specify the name of the pstore backend to use
3471 pt. [PARIDE]
3472 See Documentation/blockdev/paride.txt.
3474 pti= [X86_64] Control Page Table Isolation of user and
3475 kernel address spaces. Disabling this feature
3476 removes hardening, but improves performance of
3477 system calls and interrupts.
3479 on - unconditionally enable
3480 off - unconditionally disable
3481 auto - kernel detects whether your CPU model is
3482 vulnerable to issues that PTI mitigates
3484 Not specifying this option is equivalent to pti=auto.
3486 nopti [X86_64]
3487 Equivalent to pti=off
3489 pty.legacy_count=
3490 [KNL] Number of legacy pty's. Overwrites compiled-in
3491 default number.
3493 quiet [KNL] Disable most log messages
3495 r128= [HW,DRM]
3497 raid= [HW,RAID]
3498 See Documentation/md.txt.
3500 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3501 See Documentation/blockdev/ramdisk.txt.
3503 rcu_nocbs= [KNL]
3504 The argument is a cpu list, as described above.
3506 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3507 the specified list of CPUs to be no-callback CPUs.
3508 Invocation of these CPUs' RCU callbacks will
3509 be offloaded to "rcuox/N" kthreads created for
3510 that purpose, where "x" is "b" for RCU-bh, "p"
3511 for RCU-preempt, and "s" for RCU-sched, and "N"
3512 is the CPU number. This reduces OS jitter on the
3513 offloaded CPUs, which can be useful for HPC and
3514 real-time workloads. It can also improve energy
3515 efficiency for asymmetric multiprocessors.
3517 rcu_nocb_poll [KNL]
3518 Rather than requiring that offloaded CPUs
3519 (specified by rcu_nocbs= above) explicitly
3520 awaken the corresponding "rcuoN" kthreads,
3521 make these kthreads poll for callbacks.
3522 This improves the real-time response for the
3523 offloaded CPUs by relieving them of the need to
3524 wake up the corresponding kthread, but degrades
3525 energy efficiency by requiring that the kthreads
3526 periodically wake up to do the polling.
3528 rcutree.blimit= [KNL]
3529 Set maximum number of finished RCU callbacks to
3530 process in one batch.
3532 rcutree.dump_tree= [KNL]
3533 Dump the structure of the rcu_node combining tree
3534 out at early boot. This is used for diagnostic
3535 purposes, to verify correct tree setup.
3537 rcutree.gp_cleanup_delay= [KNL]
3538 Set the number of jiffies to delay each step of
3539 RCU grace-period cleanup. This only has effect
3540 when CONFIG_RCU_TORTURE_TEST_SLOW_CLEANUP is set.
3542 rcutree.gp_init_delay= [KNL]
3543 Set the number of jiffies to delay each step of
3544 RCU grace-period initialization. This only has
3545 effect when CONFIG_RCU_TORTURE_TEST_SLOW_INIT
3546 is set.
3548 rcutree.gp_preinit_delay= [KNL]
3549 Set the number of jiffies to delay each step of
3550 RCU grace-period pre-initialization, that is,
3551 the propagation of recent CPU-hotplug changes up
3552 the rcu_node combining tree. This only has effect
3553 when CONFIG_RCU_TORTURE_TEST_SLOW_PREINIT is set.
3555 rcutree.rcu_fanout_exact= [KNL]
3556 Disable autobalancing of the rcu_node combining
3557 tree. This is used by rcutorture, and might
3558 possibly be useful for architectures having high
3559 cache-to-cache transfer latencies.
3561 rcutree.rcu_fanout_leaf= [KNL]
3562 Change the number of CPUs assigned to each
3563 leaf rcu_node structure. Useful for very
3564 large systems, which will choose the value 64,
3565 and for NUMA systems with large remote-access
3566 latencies, which will choose a value aligned
3567 with the appropriate hardware boundaries.
3569 rcutree.jiffies_till_sched_qs= [KNL]
3570 Set required age in jiffies for a
3571 given grace period before RCU starts
3572 soliciting quiescent-state help from
3573 rcu_note_context_switch().
3575 rcutree.jiffies_till_first_fqs= [KNL]
3576 Set delay from grace-period initialization to
3577 first attempt to force quiescent states.
3578 Units are jiffies, minimum value is zero,
3579 and maximum value is HZ.
3581 rcutree.jiffies_till_next_fqs= [KNL]
3582 Set delay between subsequent attempts to force
3583 quiescent states. Units are jiffies, minimum
3584 value is one, and maximum value is HZ.
3586 rcutree.kthread_prio= [KNL,BOOT]
3587 Set the SCHED_FIFO priority of the RCU per-CPU
3588 kthreads (rcuc/N). This value is also used for
3589 the priority of the RCU boost threads (rcub/N)
3590 and for the RCU grace-period kthreads (rcu_bh,
3591 rcu_preempt, and rcu_sched). If RCU_BOOST is
3592 set, valid values are 1-99 and the default is 1
3593 (the least-favored priority). Otherwise, when
3594 RCU_BOOST is not set, valid values are 0-99 and
3595 the default is zero (non-realtime operation).
3597 rcutree.rcu_nocb_leader_stride= [KNL]
3598 Set the number of NOCB kthread groups, which
3599 defaults to the square root of the number of
3600 CPUs. Larger numbers reduces the wakeup overhead
3601 on the per-CPU grace-period kthreads, but increases
3602 that same overhead on each group's leader.
3604 rcutree.qhimark= [KNL]
3605 Set threshold of queued RCU callbacks beyond which
3606 batch limiting is disabled.
3608 rcutree.qlowmark= [KNL]
3609 Set threshold of queued RCU callbacks below which
3610 batch limiting is re-enabled.
3612 rcutree.rcu_idle_gp_delay= [KNL]
3613 Set wakeup interval for idle CPUs that have
3614 RCU callbacks (RCU_FAST_NO_HZ=y).
3616 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3617 Set wakeup interval for idle CPUs that have
3618 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3619 Lazy RCU callbacks are those which RCU can
3620 prove do nothing more than free memory.
3622 rcuperf.gp_exp= [KNL]
3623 Measure performance of expedited synchronous
3624 grace-period primitives.
3626 rcuperf.holdoff= [KNL]
3627 Set test-start holdoff period. The purpose of
3628 this parameter is to delay the start of the
3629 test until boot completes in order to avoid
3630 interference.
3632 rcuperf.nreaders= [KNL]
3633 Set number of RCU readers. The value -1 selects
3634 N, where N is the number of CPUs. A value
3635 "n" less than -1 selects N-n+1, where N is again
3636 the number of CPUs. For example, -2 selects N
3637 (the number of CPUs), -3 selects N+1, and so on.
3638 A value of "n" less than or equal to -N selects
3639 a single reader.
3641 rcuperf.nwriters= [KNL]
3642 Set number of RCU writers. The values operate
3643 the same as for rcuperf.nreaders.
3644 N, where N is the number of CPUs
3646 rcuperf.perf_runnable= [BOOT]
3647 Start rcuperf running at boot time.
3649 rcuperf.shutdown= [KNL]
3650 Shut the system down after performance tests
3651 complete. This is useful for hands-off automated
3652 testing.
3654 rcuperf.perf_type= [KNL]
3655 Specify the RCU implementation to test.
3657 rcuperf.verbose= [KNL]
3658 Enable additional printk() statements.
3660 rcutorture.cbflood_inter_holdoff= [KNL]
3661 Set holdoff time (jiffies) between successive
3662 callback-flood tests.
3664 rcutorture.cbflood_intra_holdoff= [KNL]
3665 Set holdoff time (jiffies) between successive
3666 bursts of callbacks within a given callback-flood
3667 test.
3669 rcutorture.cbflood_n_burst= [KNL]
3670 Set the number of bursts making up a given
3671 callback-flood test. Set this to zero to
3672 disable callback-flood testing.
3674 rcutorture.cbflood_n_per_burst= [KNL]
3675 Set the number of callbacks to be registered
3676 in a given burst of a callback-flood test.
3678 rcutorture.fqs_duration= [KNL]
3679 Set duration of force_quiescent_state bursts
3680 in microseconds.
3682 rcutorture.fqs_holdoff= [KNL]
3683 Set holdoff time within force_quiescent_state bursts
3684 in microseconds.
3686 rcutorture.fqs_stutter= [KNL]
3687 Set wait time between force_quiescent_state bursts
3688 in seconds.
3690 rcutorture.gp_cond= [KNL]
3691 Use conditional/asynchronous update-side
3692 primitives, if available.
3694 rcutorture.gp_exp= [KNL]
3695 Use expedited update-side primitives, if available.
3697 rcutorture.gp_normal= [KNL]
3698 Use normal (non-expedited) asynchronous
3699 update-side primitives, if available.
3701 rcutorture.gp_sync= [KNL]
3702 Use normal (non-expedited) synchronous
3703 update-side primitives, if available. If all
3704 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3705 rcutorture.gp_normal=, and rcutorture.gp_sync=
3706 are zero, rcutorture acts as if is interpreted
3707 they are all non-zero.
3709 rcutorture.n_barrier_cbs= [KNL]
3710 Set callbacks/threads for rcu_barrier() testing.
3712 rcutorture.nfakewriters= [KNL]
3713 Set number of concurrent RCU writers. These just
3714 stress RCU, they don't participate in the actual
3715 test, hence the "fake".
3717 rcutorture.nreaders= [KNL]
3718 Set number of RCU readers. The value -1 selects
3719 N-1, where N is the number of CPUs. A value
3720 "n" less than -1 selects N-n-2, where N is again
3721 the number of CPUs. For example, -2 selects N
3722 (the number of CPUs), -3 selects N+1, and so on.
3724 rcutorture.object_debug= [KNL]
3725 Enable debug-object double-call_rcu() testing.
3727 rcutorture.onoff_holdoff= [KNL]
3728 Set time (s) after boot for CPU-hotplug testing.
3730 rcutorture.onoff_interval= [KNL]
3731 Set time (s) between CPU-hotplug operations, or
3732 zero to disable CPU-hotplug testing.
3734 rcutorture.shuffle_interval= [KNL]
3735 Set task-shuffle interval (s). Shuffling tasks
3736 allows some CPUs to go into dyntick-idle mode
3737 during the rcutorture test.
3739 rcutorture.shutdown_secs= [KNL]
3740 Set time (s) after boot system shutdown. This
3741 is useful for hands-off automated testing.
3743 rcutorture.stall_cpu= [KNL]
3744 Duration of CPU stall (s) to test RCU CPU stall
3745 warnings, zero to disable.
3747 rcutorture.stall_cpu_holdoff= [KNL]
3748 Time to wait (s) after boot before inducing stall.
3750 rcutorture.stat_interval= [KNL]
3751 Time (s) between statistics printk()s.
3753 rcutorture.stutter= [KNL]
3754 Time (s) to stutter testing, for example, specifying
3755 five seconds causes the test to run for five seconds,
3756 wait for five seconds, and so on. This tests RCU's
3757 ability to transition abruptly to and from idle.
3759 rcutorture.test_boost= [KNL]
3760 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3761 "Maybe" means test if the RCU implementation
3762 under test support RCU priority boosting.
3764 rcutorture.test_boost_duration= [KNL]
3765 Duration (s) of each individual boost test.
3767 rcutorture.test_boost_interval= [KNL]
3768 Interval (s) between each boost test.
3770 rcutorture.test_no_idle_hz= [KNL]
3771 Test RCU's dyntick-idle handling. See also the
3772 rcutorture.shuffle_interval parameter.
3774 rcutorture.torture_runnable= [BOOT]
3775 Start rcutorture running at boot time.
3777 rcutorture.torture_type= [KNL]
3778 Specify the RCU implementation to test.
3780 rcutorture.verbose= [KNL]
3781 Enable additional printk() statements.
3783 rcupdate.rcu_cpu_stall_suppress= [KNL]
3784 Suppress RCU CPU stall warning messages.
3786 rcupdate.rcu_cpu_stall_timeout= [KNL]
3787 Set timeout for RCU CPU stall warning messages.
3789 rcupdate.rcu_expedited= [KNL]
3790 Use expedited grace-period primitives, for
3791 example, synchronize_rcu_expedited() instead
3792 of synchronize_rcu(). This reduces latency,
3793 but can increase CPU utilization, degrade
3794 real-time latency, and degrade energy efficiency.
3795 No effect on CONFIG_TINY_RCU kernels.
3797 rcupdate.rcu_normal= [KNL]
3798 Use only normal grace-period primitives,
3799 for example, synchronize_rcu() instead of
3800 synchronize_rcu_expedited(). This improves
3801 real-time latency, CPU utilization, and
3802 energy efficiency, but can expose users to
3803 increased grace-period latency. This parameter
3804 overrides rcupdate.rcu_expedited. No effect on
3805 CONFIG_TINY_RCU kernels.
3807 rcupdate.rcu_normal_after_boot= [KNL]
3808 Once boot has completed (that is, after
3809 rcu_end_inkernel_boot() has been invoked), use
3810 only normal grace-period primitives. No effect
3811 on CONFIG_TINY_RCU kernels.
3813 rcupdate.rcu_task_stall_timeout= [KNL]
3814 Set timeout in jiffies for RCU task stall warning
3815 messages. Disable with a value less than or equal
3816 to zero.
3818 rcupdate.rcu_self_test= [KNL]
3819 Run the RCU early boot self tests
3821 rcupdate.rcu_self_test_bh= [KNL]
3822 Run the RCU bh early boot self tests
3824 rcupdate.rcu_self_test_sched= [KNL]
3825 Run the RCU sched early boot self tests
3827 rdinit= [KNL]
3828 Format: <full_path>
3829 Run specified binary instead of /init from the ramdisk,
3830 used for early userspace startup. See initrd.
3832 rdrand= [X86]
3833 force - Override the decision by the kernel to hide the
3834 advertisement of RDRAND support (this affects
3835 certain AMD processors because of buggy BIOS
3836 support, specifically around the suspend/resume
3837 path).
3839 reboot= [KNL]
3840 Format (x86 or x86_64):
3841 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3842 [[,]s[mp]#### \
3843 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3844 [[,]f[orce]
3845 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3846 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3847 reboot_force is either force or not specified,
3848 reboot_cpu is s[mp]#### with #### being the processor
3849 to be used for rebooting.
3851 relax_domain_level=
3852 [KNL, SMP] Set scheduler's default relax_domain_level.
3853 See Documentation/cgroup-v1/cpusets.txt.
3855 relative_sleep_states=
3856 [SUSPEND] Use sleep state labeling where the deepest
3857 state available other than hibernation is always "mem".
3858 Format: { "0" | "1" }
3859 0 -- Traditional sleep state labels.
3860 1 -- Relative sleep state labels.
3862 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
3864 reservetop= [X86-32]
3865 Format: nn[KMG]
3866 Reserves a hole at the top of the kernel virtual
3867 address space.
3869 reservelow= [X86]
3870 Format: nn[K]
3871 Set the amount of memory to reserve for BIOS at
3872 the bottom of the address space.
3874 reset_devices [KNL] Force drivers to reset the underlying device
3875 during initialization.
3877 resume= [SWSUSP]
3878 Specify the partition device for software suspend
3879 Format:
3880 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3882 resume_offset= [SWSUSP]
3883 Specify the offset from the beginning of the partition
3884 given by "resume=" at which the swap header is located,
3885 in <PAGE_SIZE> units (needed only for swap files).
3886 See Documentation/power/swsusp-and-swap-files.txt
3888 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
3889 read the resume files
3891 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
3892 Useful for devices that are detected asynchronously
3893 (e.g. USB and MMC devices).
3895 hibernate= [HIBERNATION]
3896 noresume Don't check if there's a hibernation image
3897 present during boot.
3898 nocompress Don't compress/decompress hibernation images.
3899 no Disable hibernation and resume.
3900 protect_image Turn on image protection during restoration
3901 (that will set all pages holding image data
3902 during restoration read-only).
3904 retain_initrd [RAM] Keep initrd memory after extraction
3906 rfkill.default_state=
3907 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
3908 etc. communication is blocked by default.
3909 1 Unblocked.
3911 rfkill.master_switch_mode=
3912 0 The "airplane mode" button does nothing.
3913 1 The "airplane mode" button toggles between everything
3914 blocked and the previous configuration.
3915 2 The "airplane mode" button toggles between everything
3916 blocked and everything unblocked.
3918 rhash_entries= [KNL,NET]
3919 Set number of hash buckets for route cache
3921 ro [KNL] Mount root device read-only on boot
3923 rodata= [KNL]
3924 on Mark read-only kernel memory as read-only (default).
3925 off Leave read-only kernel memory writable for debugging.
3927 rockchip.usb_uart
3928 Enable the uart passthrough on the designated usb port
3929 on Rockchip SoCs. When active, the signals of the
3930 debug-uart get routed to the D+ and D- pins of the usb
3931 port and the regular usb controller gets disabled.
3933 root= [KNL] Root filesystem
3934 See name_to_dev_t comment in init/do_mounts.c.
3936 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
3937 mount the root filesystem
3939 rootflags= [KNL] Set root filesystem mount option string
3941 rootfstype= [KNL] Set root filesystem type
3943 rootwait [KNL] Wait (indefinitely) for root device to show up.
3944 Useful for devices that are detected asynchronously
3945 (e.g. USB and MMC devices).
3947 rproc_mem=nn[KMG][@address]
3948 [KNL,ARM,CMA] Remoteproc physical memory block.
3949 Memory area to be used by remote processor image,
3950 managed by CMA.
3952 rw [KNL] Mount root device read-write on boot
3954 S [KNL] Run init in single mode
3956 s390_iommu= [HW,S390]
3957 Set s390 IOTLB flushing mode
3958 strict
3959 With strict flushing every unmap operation will result in
3960 an IOTLB flush. Default is lazy flushing before reuse,
3961 which is faster.
3963 sa1100ir [NET]
3964 See drivers/net/irda/sa1100_ir.c.
3966 sbni= [NET] Granch SBNI12 leased line adapter
3968 sched_debug [KNL] Enables verbose scheduler debug messages.
3970 schedstats= [KNL,X86] Enable or disable scheduled statistics.
3971 Allowed values are enable and disable. This feature
3972 incurs a small amount of overhead in the scheduler
3973 but is useful for debugging and performance tuning.
3975 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
3976 xtime_lock contention on larger systems, and/or RCU lock
3977 contention on all systems with CONFIG_MAXSMP set.
3978 Format: { "0" | "1" }
3979 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
3980 1 -- enable.
3981 Note: increases power consumption, thus should only be
3982 enabled if running jitter sensitive (HPC/RT) workloads.
3984 security= [SECURITY] Choose a security module to enable at boot.
3985 If this boot parameter is not specified, only the first
3986 security module asking for security registration will be
3987 loaded. An invalid security module name will be treated
3988 as if no module has been chosen.
3990 selinux= [SELINUX] Disable or enable SELinux at boot time.
3991 Format: { "0" | "1" }
3992 See security/selinux/Kconfig help text.
3993 0 -- disable.
3994 1 -- enable.
3995 Default value is set via kernel config option.
3996 If enabled at boot time, /selinux/disable can be used
3997 later to disable prior to initial policy load.
3999 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4000 Format: { "0" | "1" }
4001 See security/apparmor/Kconfig help text
4002 0 -- disable.
4003 1 -- enable.
4004 Default value is set via kernel config option.
4006 serialnumber [BUGS=X86-32]
4008 shapers= [NET]
4009 Maximal number of shapers.
4011 show_msr= [x86] show boot-time MSR settings
4012 Format: { <integer> }
4013 Show boot-time (BIOS-initialized) MSR settings.
4014 The parameter means the number of CPUs to show,
4015 for example 1 means boot CPU only.
4017 simeth= [IA-64]
4018 simscsi=
4020 slram= [HW,MTD]
4022 slab_nomerge [MM]
4023 Disable merging of slabs with similar size. May be
4024 necessary if there is some reason to distinguish
4025 allocs to different slabs. Debug options disable
4026 merging on their own.
4027 For more information see Documentation/vm/slub.txt.
4029 slab_max_order= [MM, SLAB]
4030 Determines the maximum allowed order for slabs.
4031 A high setting may cause OOMs due to memory
4032 fragmentation. Defaults to 1 for systems with
4033 more than 32MB of RAM, 0 otherwise.
4035 slub_debug[=options[,slabs]] [MM, SLUB]
4036 Enabling slub_debug allows one to determine the
4037 culprit if slab objects become corrupted. Enabling
4038 slub_debug can create guard zones around objects and
4039 may poison objects when not in use. Also tracks the
4040 last alloc / free. For more information see
4041 Documentation/vm/slub.txt.
4043 slub_max_order= [MM, SLUB]
4044 Determines the maximum allowed order for slabs.
4045 A high setting may cause OOMs due to memory
4046 fragmentation. For more information see
4047 Documentation/vm/slub.txt.
4049 slub_min_objects= [MM, SLUB]
4050 The minimum number of objects per slab. SLUB will
4051 increase the slab order up to slub_max_order to
4052 generate a sufficiently large slab able to contain
4053 the number of objects indicated. The higher the number
4054 of objects the smaller the overhead of tracking slabs
4055 and the less frequently locks need to be acquired.
4056 For more information see Documentation/vm/slub.txt.
4058 slub_min_order= [MM, SLUB]
4059 Determines the minimum page order for slabs. Must be
4060 lower than slub_max_order.
4061 For more information see Documentation/vm/slub.txt.
4063 slub_nomerge [MM, SLUB]
4064 Same with slab_nomerge. This is supported for legacy.
4065 See slab_nomerge for more information.
4067 smart2= [HW]
4068 Format: <io1>[,<io2>[,...,<io8>]]
4070 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4071 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4072 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4073 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4074 smsc-ircc2.ircc_irq= [HW] IRQ line
4075 smsc-ircc2.ircc_dma= [HW] DMA channel
4076 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4077 0: Toshiba Satellite 1800 (GP data pin select)
4078 1: Fast pin select (default)
4079 2: ATC IRMode
4081 smt [KNL,S390] Set the maximum number of threads (logical
4082 CPUs) to use per physical CPU on systems capable of
4083 symmetric multithreading (SMT). Will be capped to the
4084 actual hardware limit.
4085 Format: <integer>
4086 Default: -1 (no limit)
4088 softlockup_panic=
4089 [KNL] Should the soft-lockup detector generate panics.
4090 Format: <integer>
4092 softlockup_all_cpu_backtrace=
4093 [KNL] Should the soft-lockup detector generate
4094 backtraces on all cpus.
4095 Format: <integer>
4097 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4098 See Documentation/laptops/sonypi.txt
4100 spectre_v2= [X86] Control mitigation of Spectre variant 2
4101 (indirect branch speculation) vulnerability.
4102 The default operation protects the kernel from
4103 user space attacks.
4105 on - unconditionally enable, implies
4106 spectre_v2_user=on
4107 off - unconditionally disable, implies
4108 spectre_v2_user=off
4109 auto - kernel detects whether your CPU model is
4110 vulnerable
4112 Selecting 'on' will, and 'auto' may, choose a
4113 mitigation method at run time according to the
4114 CPU, the available microcode, the setting of the
4115 CONFIG_RETPOLINE configuration option, and the
4116 compiler with which the kernel was built.
4118 Selecting 'on' will also enable the mitigation
4119 against user space to user space task attacks.
4121 Selecting 'off' will disable both the kernel and
4122 the user space protections.
4124 Specific mitigations can also be selected manually:
4126 retpoline - replace indirect branches
4127 retpoline,generic - google's original retpoline
4128 retpoline,amd - AMD-specific minimal thunk
4130 Not specifying this option is equivalent to
4131 spectre_v2=auto.
4133 spectre_v2_user=
4134 [X86] Control mitigation of Spectre variant 2
4135 (indirect branch speculation) vulnerability between
4136 user space tasks
4138 on - Unconditionally enable mitigations. Is
4139 enforced by spectre_v2=on
4141 off - Unconditionally disable mitigations. Is
4142 enforced by spectre_v2=off
4144 prctl - Indirect branch speculation is enabled,
4145 but mitigation can be enabled via prctl
4146 per thread. The mitigation control state
4147 is inherited on fork.
4149 prctl,ibpb
4150 - Like "prctl" above, but only STIBP is
4151 controlled per thread. IBPB is issued
4152 always when switching between different user
4153 space processes.
4155 seccomp
4156 - Same as "prctl" above, but all seccomp
4157 threads will enable the mitigation unless
4158 they explicitly opt out.
4160 seccomp,ibpb
4161 - Like "seccomp" above, but only STIBP is
4162 controlled per thread. IBPB is issued
4163 always when switching between different
4164 user space processes.
4166 auto - Kernel selects the mitigation depending on
4167 the available CPU features and vulnerability.
4169 Default mitigation:
4170 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4172 Not specifying this option is equivalent to
4173 spectre_v2_user=auto.
4175 spec_store_bypass_disable=
4176 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4177 (Speculative Store Bypass vulnerability)
4179 Certain CPUs are vulnerable to an exploit against a
4180 a common industry wide performance optimization known
4181 as "Speculative Store Bypass" in which recent stores
4182 to the same memory location may not be observed by
4183 later loads during speculative execution. The idea
4184 is that such stores are unlikely and that they can
4185 be detected prior to instruction retirement at the
4186 end of a particular speculation execution window.
4188 In vulnerable processors, the speculatively forwarded
4189 store can be used in a cache side channel attack, for
4190 example to read memory to which the attacker does not
4191 directly have access (e.g. inside sandboxed code).
4193 This parameter controls whether the Speculative Store
4194 Bypass optimization is used.
4196 on - Unconditionally disable Speculative Store Bypass
4197 off - Unconditionally enable Speculative Store Bypass
4198 auto - Kernel detects whether the CPU model contains an
4199 implementation of Speculative Store Bypass and
4200 picks the most appropriate mitigation. If the
4201 CPU is not vulnerable, "off" is selected. If the
4202 CPU is vulnerable the default mitigation is
4203 architecture and Kconfig dependent. See below.
4204 prctl - Control Speculative Store Bypass per thread
4205 via prctl. Speculative Store Bypass is enabled
4206 for a process by default. The state of the control
4207 is inherited on fork.
4208 seccomp - Same as "prctl" above, but all seccomp threads
4209 will disable SSB unless they explicitly opt out.
4211 Not specifying this option is equivalent to
4212 spec_store_bypass_disable=auto.
4214 Default mitigations:
4215 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4217 spia_io_base= [HW,MTD]
4218 spia_fio_base=
4219 spia_pedr=
4220 spia_peddr=
4222 ssbd= [ARM64,HW]
4223 Speculative Store Bypass Disable control
4225 On CPUs that are vulnerable to the Speculative
4226 Store Bypass vulnerability and offer a
4227 firmware based mitigation, this parameter
4228 indicates how the mitigation should be used:
4230 force-on: Unconditionally enable mitigation for
4231 for both kernel and userspace
4232 force-off: Unconditionally disable mitigation for
4233 for both kernel and userspace
4234 kernel: Always enable mitigation in the
4235 kernel, and offer a prctl interface
4236 to allow userspace to register its
4237 interest in being mitigated too.
4239 stack_guard_gap= [MM]
4240 override the default stack gap protection. The value
4241 is in page units and it defines how many pages prior
4242 to (for stacks growing down) resp. after (for stacks
4243 growing up) the main stack are reserved for no other
4244 mapping. Default value is 256 pages.
4246 stacktrace [FTRACE]
4247 Enabled the stack tracer on boot up.
4249 stacktrace_filter=[function-list]
4250 [FTRACE] Limit the functions that the stack tracer
4251 will trace at boot up. function-list is a comma separated
4252 list of functions. This list can be changed at run
4253 time by the stack_trace_filter file in the debugfs
4254 tracing directory. Note, this enables stack tracing
4255 and the stacktrace above is not needed.
4257 sti= [PARISC,HW]
4258 Format: <num>
4259 Set the STI (builtin display/keyboard on the HP-PARISC
4260 machines) console (graphic card) which should be used
4261 as the initial boot-console.
4262 See also comment in drivers/video/console/sticore.c.
4264 sti_font= [HW]
4265 See comment in drivers/video/console/sticore.c.
4267 stifb= [HW]
4268 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4270 sunrpc.min_resvport=
4271 sunrpc.max_resvport=
4272 [NFS,SUNRPC]
4273 SunRPC servers often require that client requests
4274 originate from a privileged port (i.e. a port in the
4275 range 0 < portnr < 1024).
4276 An administrator who wishes to reserve some of these
4277 ports for other uses may adjust the range that the
4278 kernel's sunrpc client considers to be privileged
4279 using these two parameters to set the minimum and
4280 maximum port values.
4282 sunrpc.svc_rpc_per_connection_limit=
4283 [NFS,SUNRPC]
4284 Limit the number of requests that the server will
4285 process in parallel from a single connection.
4286 The default value is 0 (no limit).
4288 sunrpc.pool_mode=
4289 [NFS]
4290 Control how the NFS server code allocates CPUs to
4291 service thread pools. Depending on how many NICs
4292 you have and where their interrupts are bound, this
4293 option will affect which CPUs will do NFS serving.
4294 Note: this parameter cannot be changed while the
4295 NFS server is running.
4297 auto the server chooses an appropriate mode
4298 automatically using heuristics
4299 global a single global pool contains all CPUs
4300 percpu one pool for each CPU
4301 pernode one pool for each NUMA node (equivalent
4302 to global on non-NUMA machines)
4304 sunrpc.tcp_slot_table_entries=
4305 sunrpc.udp_slot_table_entries=
4306 [NFS,SUNRPC]
4307 Sets the upper limit on the number of simultaneous
4308 RPC calls that can be sent from the client to a
4309 server. Increasing these values may allow you to
4310 improve throughput, but will also increase the
4311 amount of memory reserved for use by the client.
4313 suspend.pm_test_delay=
4314 [SUSPEND]
4315 Sets the number of seconds to remain in a suspend test
4316 mode before resuming the system (see
4317 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4318 is set. Default value is 5.
4320 swapaccount=[0|1]
4321 [KNL] Enable accounting of swap in memory resource
4322 controller if no parameter or 1 is given or disable
4323 it if 0 is given (See Documentation/cgroup-v1/memory.txt)
4325 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4326 Format: { <int> | force | noforce }
4327 <int> -- Number of I/O TLB slabs
4328 force -- force using of bounce buffers even if they
4329 wouldn't be automatically used by the kernel
4330 noforce -- Never use bounce buffers (for debugging)
4332 switches= [HW,M68k]
4334 sysfs.deprecated=0|1 [KNL]
4335 Enable/disable old style sysfs layout for old udev
4336 on older distributions. When this option is enabled
4337 very new udev will not work anymore. When this option
4338 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4339 in older udev will not work anymore.
4340 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4341 the kernel configuration.
4343 sysrq_always_enabled
4344 [KNL]
4345 Ignore sysrq setting - this boot parameter will
4346 neutralize any effect of /proc/sys/kernel/sysrq.
4347 Useful for debugging.
4349 tcpmhash_entries= [KNL,NET]
4350 Set the number of tcp_metrics_hash slots.
4351 Default value is 8192 or 16384 depending on total
4352 ram pages. This is used to specify the TCP metrics
4353 cache size. See Documentation/networking/ip-sysctl.txt
4354 "tcp_no_metrics_save" section for more details.
4356 tdfx= [HW,DRM]
4358 test_suspend= [SUSPEND][,N]
4359 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4360 standby suspend) or "freeze" (for suspend type freeze)
4361 as the system sleep state during system startup with
4362 the optional capability to repeat N number of times.
4363 The system is woken from this state using a
4364 wakeup-capable RTC alarm.
4366 thash_entries= [KNL,NET]
4367 Set number of hash buckets for TCP connection
4369 thermal.act= [HW,ACPI]
4370 -1: disable all active trip points in all thermal zones
4371 <degrees C>: override all lowest active trip points
4373 thermal.crt= [HW,ACPI]
4374 -1: disable all critical trip points in all thermal zones
4375 <degrees C>: override all critical trip points
4377 thermal.nocrt= [HW,ACPI]
4378 Set to disable actions on ACPI thermal zone
4379 critical and hot trip points.
4381 thermal.off= [HW,ACPI]
4382 1: disable ACPI thermal control
4384 thermal.psv= [HW,ACPI]
4385 -1: disable all passive trip points
4386 <degrees C>: override all passive trip points to this
4387 value
4389 thermal.tzp= [HW,ACPI]
4390 Specify global default ACPI thermal zone polling rate
4391 <deci-seconds>: poll all this frequency
4392 0: no polling (default)
4394 threadirqs [KNL]
4395 Force threading of all interrupt handlers except those
4396 marked explicitly IRQF_NO_THREAD.
4398 tmem [KNL,XEN]
4399 Enable the Transcendent memory driver if built-in.
4401 tmem.cleancache=0|1 [KNL, XEN]
4402 Default is on (1). Disable the usage of the cleancache
4403 API to send anonymous pages to the hypervisor.
4405 tmem.frontswap=0|1 [KNL, XEN]
4406 Default is on (1). Disable the usage of the frontswap
4407 API to send swap pages to the hypervisor. If disabled
4408 the selfballooning and selfshrinking are force disabled.
4410 tmem.selfballooning=0|1 [KNL, XEN]
4411 Default is on (1). Disable the driving of swap pages
4412 to the hypervisor.
4414 tmem.selfshrinking=0|1 [KNL, XEN]
4415 Default is on (1). Partial swapoff that immediately
4416 transfers pages from Xen hypervisor back to the
4417 kernel based on different criteria.
4419 topology= [S390]
4420 Format: {off | on}
4421 Specify if the kernel should make use of the cpu
4422 topology information if the hardware supports this.
4423 The scheduler will make use of this information and
4424 e.g. base its process migration decisions on it.
4425 Default is on.
4427 topology_updates= [KNL, PPC, NUMA]
4428 Format: {off}
4429 Specify if the kernel should ignore (off)
4430 topology updates sent by the hypervisor to this
4431 LPAR.
4433 tp720= [HW,PS2]
4435 tpm_suspend_pcr=[HW,TPM]
4436 Format: integer pcr id
4437 Specify that at suspend time, the tpm driver
4438 should extend the specified pcr with zeros,
4439 as a workaround for some chips which fail to
4440 flush the last written pcr on TPM_SaveState.
4441 This will guarantee that all the other pcrs
4442 are saved.
4444 trace_buf_size=nn[KMG]
4445 [FTRACE] will set tracing buffer size on each cpu.
4447 trace_event=[event-list]
4448 [FTRACE] Set and start specified trace events in order
4449 to facilitate early boot debugging. The event-list is a
4450 comma separated list of trace events to enable. See
4451 also Documentation/trace/events.txt
4453 trace_options=[option-list]
4454 [FTRACE] Enable or disable tracer options at boot.
4455 The option-list is a comma delimited list of options
4456 that can be enabled or disabled just as if you were
4457 to echo the option name into
4459 /sys/kernel/debug/tracing/trace_options
4461 For example, to enable stacktrace option (to dump the
4462 stack trace of each event), add to the command line:
4464 trace_options=stacktrace
4466 See also Documentation/trace/ftrace.txt "trace options"
4467 section.
4469 tp_printk[FTRACE]
4470 Have the tracepoints sent to printk as well as the
4471 tracing ring buffer. This is useful for early boot up
4472 where the system hangs or reboots and does not give the
4473 option for reading the tracing buffer or performing a
4474 ftrace_dump_on_oops.
4476 To turn off having tracepoints sent to printk,
4477 echo 0 > /proc/sys/kernel/tracepoint_printk
4478 Note, echoing 1 into this file without the
4479 tracepoint_printk kernel cmdline option has no effect.
4481 ** CAUTION **
4483 Having tracepoints sent to printk() and activating high
4484 frequency tracepoints such as irq or sched, can cause
4485 the system to live lock.
4487 traceoff_on_warning
4488 [FTRACE] enable this option to disable tracing when a
4489 warning is hit. This turns off "tracing_on". Tracing can
4490 be enabled again by echoing '1' into the "tracing_on"
4491 file located in /sys/kernel/debug/tracing/
4493 This option is useful, as it disables the trace before
4494 the WARNING dump is called, which prevents the trace to
4495 be filled with content caused by the warning output.
4497 This option can also be set at run time via the sysctl
4498 option: kernel/traceoff_on_warning
4500 transparent_hugepage=
4501 [KNL]
4502 Format: [always|madvise|never]
4503 Can be used to control the default behavior of the system
4504 with respect to transparent hugepages.
4505 See Documentation/vm/transhuge.txt for more details.
4507 tsc= Disable clocksource stability checks for TSC.
4508 Format: <string>
4509 [x86] reliable: mark tsc clocksource as reliable, this
4510 disables clocksource verification at runtime, as well
4511 as the stability checks done at bootup. Used to enable
4512 high-resolution timer mode on older hardware, and in
4513 virtualized environment.
4514 [x86] noirqtime: Do not use TSC to do irq accounting.
4515 Used to run time disable IRQ_TIME_ACCOUNTING on any
4516 platforms where RDTSC is slow and this accounting
4517 can add overhead.
4519 turbografx.map[2|3]= [HW,JOY]
4520 TurboGraFX parallel port interface
4521 Format:
4522 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4523 See also Documentation/input/joystick-parport.txt
4525 udbg-immortal [PPC] When debugging early kernel crashes that
4526 happen after console_init() and before a proper
4527 console driver takes over, this boot options might
4528 help "seeing" what's going on.
4530 uhash_entries= [KNL,NET]
4531 Set number of hash buckets for UDP/UDP-Lite connections
4533 uhci-hcd.ignore_oc=
4534 [USB] Ignore overcurrent events (default N).
4535 Some badly-designed motherboards generate lots of
4536 bogus events, for ports that aren't wired to
4537 anything. Set this parameter to avoid log spamming.
4538 Note that genuine overcurrent events won't be
4539 reported either.
4541 unknown_nmi_panic
4542 [X86] Cause panic on unknown NMI.
4544 usbcore.authorized_default=
4545 [USB] Default USB device authorization:
4546 (default -1 = authorized except for wireless USB,
4547 0 = not authorized, 1 = authorized)
4549 usbcore.autosuspend=
4550 [USB] The autosuspend time delay (in seconds) used
4551 for newly-detected USB devices (default 2). This
4552 is the time required before an idle device will be
4553 autosuspended. Devices for which the delay is set
4554 to a negative value won't be autosuspended at all.
4556 usbcore.usbfs_snoop=
4557 [USB] Set to log all usbfs traffic (default 0 = off).
4559 usbcore.usbfs_snoop_max=
4560 [USB] Maximum number of bytes to snoop in each URB
4561 (default = 65536).
4563 usbcore.blinkenlights=
4564 [USB] Set to cycle leds on hubs (default 0 = off).
4566 usbcore.old_scheme_first=
4567 [USB] Start with the old device initialization
4568 scheme (default 0 = off).
4570 usbcore.usbfs_memory_mb=
4571 [USB] Memory limit (in MB) for buffers allocated by
4572 usbfs (default = 16, 0 = max = 2047).
4574 usbcore.use_both_schemes=
4575 [USB] Try the other device initialization scheme
4576 if the first one fails (default 1 = enabled).
4578 usbcore.initial_descriptor_timeout=
4579 [USB] Specifies timeout for the initial 64-byte
4580 USB_REQ_GET_DESCRIPTOR request in milliseconds
4581 (default 5000 = 5.0 seconds).
4583 usbcore.nousb [USB] Disable the USB subsystem
4585 usbhid.mousepoll=
4586 [USBHID] The interval which mice are to be polled at.
4588 usb-storage.delay_use=
4589 [UMS] The delay in seconds before a new device is
4590 scanned for Logical Units (default 1).
4592 usb-storage.quirks=
4593 [UMS] A list of quirks entries to supplement or
4594 override the built-in unusual_devs list. List
4595 entries are separated by commas. Each entry has
4596 the form VID:PID:Flags where VID and PID are Vendor
4597 and Product ID values (4-digit hex numbers) and
4598 Flags is a set of characters, each corresponding
4599 to a common usb-storage quirk flag as follows:
4600 a = SANE_SENSE (collect more than 18 bytes
4601 of sense data);
4602 b = BAD_SENSE (don't collect more than 18
4603 bytes of sense data);
4604 c = FIX_CAPACITY (decrease the reported
4605 device capacity by one sector);
4606 d = NO_READ_DISC_INFO (don't use
4607 READ_DISC_INFO command);
4608 e = NO_READ_CAPACITY_16 (don't use
4609 READ_CAPACITY_16 command);
4610 f = NO_REPORT_OPCODES (don't use report opcodes
4611 command, uas only);
4612 g = MAX_SECTORS_240 (don't transfer more than
4613 240 sectors at a time, uas only);
4614 h = CAPACITY_HEURISTICS (decrease the
4615 reported device capacity by one
4616 sector if the number is odd);
4617 i = IGNORE_DEVICE (don't bind to this
4618 device);
4619 j = NO_REPORT_LUNS (don't use report luns
4620 command, uas only);
4621 l = NOT_LOCKABLE (don't try to lock and
4622 unlock ejectable media);
4623 m = MAX_SECTORS_64 (don't transfer more
4624 than 64 sectors = 32 KB at a time);
4625 n = INITIAL_READ10 (force a retry of the
4626 initial READ(10) command);
4627 o = CAPACITY_OK (accept the capacity
4628 reported by the device);
4629 p = WRITE_CACHE (the device cache is ON
4630 by default);
4631 r = IGNORE_RESIDUE (the device reports
4632 bogus residue values);
4633 s = SINGLE_LUN (the device has only one
4634 Logical Unit);
4635 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
4636 commands, uas only);
4637 u = IGNORE_UAS (don't bind to the uas driver);
4638 w = NO_WP_DETECT (don't test whether the
4639 medium is write-protected).
4640 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
4641 even if the device claims no cache)
4642 Example: quirks=0419:aaf5:rl,0421:0433:rc
4644 user_debug= [KNL,ARM]
4645 Format: <int>
4646 See arch/arm/Kconfig.debug help text.
4647 1 - undefined instruction events
4648 2 - system calls
4649 4 - invalid data aborts
4650 8 - SIGSEGV faults
4651 16 - SIGBUS faults
4652 Example: user_debug=31
4654 userpte=
4655 [X86] Flags controlling user PTE allocations.
4657 nohigh = do not allocate PTE pages in
4658 HIGHMEM regardless of setting
4659 of CONFIG_HIGHPTE.
4661 vdso= [X86,SH]
4662 On X86_32, this is an alias for vdso32=. Otherwise:
4664 vdso=1: enable VDSO (the default)
4665 vdso=0: disable VDSO mapping
4667 vdso32= [X86] Control the 32-bit vDSO
4668 vdso32=1: enable 32-bit VDSO
4669 vdso32=0 or vdso32=2: disable 32-bit VDSO
4671 See the help text for CONFIG_COMPAT_VDSO for more
4672 details. If CONFIG_COMPAT_VDSO is set, the default is
4673 vdso32=0; otherwise, the default is vdso32=1.
4675 For compatibility with older kernels, vdso32=2 is an
4676 alias for vdso32=0.
4678 Try vdso32=0 if you encounter an error that says:
4679 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
4681 vector= [IA-64,SMP]
4682 vector=percpu: enable percpu vector domain
4684 video= [FB] Frame buffer configuration
4685 See Documentation/fb/modedb.txt.
4687 video.brightness_switch_enabled= [0,1]
4688 If set to 1, on receiving an ACPI notify event
4689 generated by hotkey, video driver will adjust brightness
4690 level and then send out the event to user space through
4691 the allocated input device; If set to 0, video driver
4692 will only send out the event without touching backlight
4693 brightness level.
4694 default: 1
4696 virtio_mmio.device=
4697 [VMMIO] Memory mapped virtio (platform) device.
4699 <size>@<baseaddr>:<irq>[:<id>]
4700 where:
4701 <size> := size (can use standard suffixes
4702 like K, M and G)
4703 <baseaddr> := physical base address
4704 <irq> := interrupt number (as passed to
4705 request_irq())
4706 <id> := (optional) platform device id
4707 example:
4708 virtio_mmio.device=1K@0x100b0000:48:7
4710 Can be used multiple times for multiple devices.
4712 vga= [BOOT,X86-32] Select a particular video mode
4713 See Documentation/x86/boot.txt and
4714 Documentation/svga.txt.
4715 Use vga=ask for menu.
4716 This is actually a boot loader parameter; the value is
4717 passed to the kernel using a special protocol.
4719 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
4720 size of <nn>. This can be used to increase the
4721 minimum size (128MB on x86). It can also be used to
4722 decrease the size and leave more room for directly
4723 mapped kernel RAM.
4725 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
4726 Format: <command>
4728 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
4729 Format: <command>
4731 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
4732 Format: <command>
4734 vsyscall= [X86-64]
4735 Controls the behavior of vsyscalls (i.e. calls to
4736 fixed addresses of 0xffffffffff600x00 from legacy
4737 code). Most statically-linked binaries and older
4738 versions of glibc use these calls. Because these
4739 functions are at fixed addresses, they make nice
4740 targets for exploits that can control RIP.
4742 emulate [default] Vsyscalls turn into traps and are
4743 emulated reasonably safely.
4745 native Vsyscalls are native syscall instructions.
4746 This is a little bit faster than trapping
4747 and makes a few dynamic recompilers work
4748 better than they would in emulation mode.
4749 It also makes exploits much easier to write.
4751 none Vsyscalls don't work at all. This makes
4752 them quite hard to use for exploits but
4753 might break your system.
4755 vt.color= [VT] Default text color.
4756 Format: 0xYX, X = foreground, Y = background.
4757 Default: 0x07 = light gray on black.
4759 vt.cur_default= [VT] Default cursor shape.
4760 Format: 0xCCBBAA, where AA, BB, and CC are the same as
4761 the parameters of the <Esc>[?A;B;Cc escape sequence;
4762 see VGA-softcursor.txt. Default: 2 = underline.
4764 vt.default_blu= [VT]
4765 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
4766 Change the default blue palette of the console.
4767 This is a 16-member array composed of values
4768 ranging from 0-255.
4770 vt.default_grn= [VT]
4771 Format: <green0>,<green1>,<green2>,...,<green15>
4772 Change the default green palette of the console.
4773 This is a 16-member array composed of values
4774 ranging from 0-255.
4776 vt.default_red= [VT]
4777 Format: <red0>,<red1>,<red2>,...,<red15>
4778 Change the default red palette of the console.
4779 This is a 16-member array composed of values
4780 ranging from 0-255.
4782 vt.default_utf8=
4783 [VT]
4784 Format=<0|1>
4785 Set system-wide default UTF-8 mode for all tty's.
4786 Default is 1, i.e. UTF-8 mode is enabled for all
4787 newly opened terminals.
4789 vt.global_cursor_default=
4790 [VT]
4791 Format=<-1|0|1>
4792 Set system-wide default for whether a cursor
4793 is shown on new VTs. Default is -1,
4794 i.e. cursors will be created by default unless
4795 overridden by individual drivers. 0 will hide
4796 cursors, 1 will display them.
4798 vt.italic= [VT] Default color for italic text; 0-15.
4799 Default: 2 = green.
4801 vt.underline= [VT] Default color for underlined text; 0-15.
4802 Default: 3 = cyan.
4804 watchdog timers [HW,WDT] For information on watchdog timers,
4805 see Documentation/watchdog/watchdog-parameters.txt
4806 or other driver-specific files in the
4807 Documentation/watchdog/ directory.
4809 workqueue.watchdog_thresh=
4810 If CONFIG_WQ_WATCHDOG is configured, workqueue can
4811 warn stall conditions and dump internal state to
4812 help debugging. 0 disables workqueue stall
4813 detection; otherwise, it's the stall threshold
4814 duration in seconds. The default value is 30 and
4815 it can be updated at runtime by writing to the
4816 corresponding sysfs file.
4818 workqueue.disable_numa
4819 By default, all work items queued to unbound
4820 workqueues are affine to the NUMA nodes they're
4821 issued on, which results in better behavior in
4822 general. If NUMA affinity needs to be disabled for
4823 whatever reason, this option can be used. Note
4824 that this also can be controlled per-workqueue for
4825 workqueues visible under /sys/bus/workqueue/.
4827 workqueue.power_efficient
4828 Per-cpu workqueues are generally preferred because
4829 they show better performance thanks to cache
4830 locality; unfortunately, per-cpu workqueues tend to
4831 be more power hungry than unbound workqueues.
4833 Enabling this makes the per-cpu workqueues which
4834 were observed to contribute significantly to power
4835 consumption unbound, leading to measurably lower
4836 power usage at the cost of small performance
4837 overhead.
4839 The default value of this parameter is determined by
4840 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
4842 workqueue.debug_force_rr_cpu
4843 Workqueue used to implicitly guarantee that work
4844 items queued without explicit CPU specified are put
4845 on the local CPU. This guarantee is no longer true
4846 and while local CPU is still preferred work items
4847 may be put on foreign CPUs. This debug option
4848 forces round-robin CPU selection to flush out
4849 usages which depend on the now broken guarantee.
4850 When enabled, memory and cache locality will be
4851 impacted.
4853 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
4854 default x2apic cluster mode on platforms
4855 supporting x2apic.
4857 x86_intel_mid_timer= [X86-32,APBT]
4858 Choose timer option for x86 Intel MID platform.
4859 Two valid options are apbt timer only and lapic timer
4860 plus one apbt timer for broadcast timer.
4861 x86_intel_mid_timer=apbt_only | lapic_and_apbt
4863 xen_512gb_limit [KNL,X86-64,XEN]
4864 Restricts the kernel running paravirtualized under Xen
4865 to use only up to 512 GB of RAM. The reason to do so is
4866 crash analysis tools and Xen tools for doing domain
4867 save/restore/migration must be enabled to handle larger
4868 domains.
4870 xen_emul_unplug= [HW,X86,XEN]
4871 Unplug Xen emulated devices
4872 Format: [unplug0,][unplug1]
4873 ide-disks -- unplug primary master IDE devices
4874 aux-ide-disks -- unplug non-primary-master IDE devices
4875 nics -- unplug network devices
4876 all -- unplug all emulated devices (NICs and IDE disks)
4877 unnecessary -- unplugging emulated devices is
4878 unnecessary even if the host did not respond to
4879 the unplug protocol
4880 never -- do not unplug even if version check succeeds
4882 xen_nopvspin [X86,XEN]
4883 Disables the ticketlock slowpath using Xen PV
4884 optimizations.
4886 xen_nopv [X86]
4887 Disables the PV optimizations forcing the HVM guest to
4888 run as generic HVM guest with no PV drivers.
4890 xirc2ps_cs= [NET,PCMCIA]
4891 Format:
4892 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
4894 ______________________________________________________________________
4896 TODO:
4898 Add more DRM drivers.