4 The following is a consolidated list of the kernel parameters as
5 implemented by the __setup(), core_param() and module_param() macros
6 and sorted into English Dictionary order (defined as ignoring all
7 punctuation and sorting digits before letters in a case insensitive
8 manner), and with descriptions where known.
10 The kernel parses parameters from the kernel command line up to "--";
11 if it doesn't recognize a parameter and it doesn't contain a '.', the
12 parameter gets passed to init: parameters with '=' go into init's
13 environment, others are passed as command line arguments to init.
14 Everything after "--" is passed as an argument to init.
16 Module parameters can be specified in two ways: via the kernel command
17 line with a module name prefix, or via modprobe, e.g.:
19 (kernel command line) usbcore.blinkenlights=1
20 (modprobe command line) modprobe usbcore blinkenlights=1
22 Parameters for modules which are built into the kernel need to be
23 specified on the kernel command line. modprobe looks through the
24 kernel command line (/proc/cmdline) and collects module parameters
25 when it loads a module, so the kernel command line can be used for
28 Hyphens (dashes) and underscores are equivalent in parameter names, so
29 log_buf_len=1M print-fatal-signals=1
30 can also be entered as
31 log-buf-len=1M print_fatal_signals=1
33 Double-quotes can be used to protect spaces in values, e.g.:
34 param="spaces in here"
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>
46 <cpu number>-<cpu number>
47 (must be a positive range in ascending order)
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
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 |
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"
215 See also Documentation/power/runtime_pm.txt, pci=noacpi
217 acpi_apic_instance= [ACPI, IOAPIC]
219 2: use 2nd APIC table, if available
220 1,0: use 1st APIC table
223 acpi_backlight= [HW,ACPI]
224 acpi_backlight=vendor
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]
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
293 acpi_irq_balance [HW,ACPI]
294 ACPI will balance active IRQs
297 acpi_irq_nobalance [HW,ACPI]
298 ACPI will not move active IRQs (default)
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
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
313 This facility can be used to prevent such uncontrolled
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
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_no_watchdog [HW,ACPI,WDT]
339 Ignore the ACPI-based watchdog interface (WDAT) and let
340 a native driver control the watchdog device instead.
342 acpi_rsdp= [ACPI,EFI,KEXEC]
343 Pass the RSDP address to the kernel, mostly used
344 on machines running EFI runtime service to boot the
345 second kernel for kdump.
347 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
348 Format: To spoof as Windows 98: ="Microsoft Windows"
350 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
351 of 2 which is mandated by ACPI 6) as the supported ACPI
352 specification revision (when using this switch, it may
353 be necessary to carry out a cold reboot _twice_ in a
354 row to make it take effect on the platform firmware).
356 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
357 acpi_osi="string1" # add string1
358 acpi_osi="!string2" # remove string2
359 acpi_osi=!* # remove all strings
360 acpi_osi=! # disable all built-in OS vendor
362 acpi_osi=!! # enable all built-in OS vendor
364 acpi_osi= # disable all strings
366 'acpi_osi=!' can be used in combination with single or
367 multiple 'acpi_osi="string1"' to support specific OS
368 vendor string(s). Note that such command can only
369 affect the default state of the OS vendor strings, thus
370 it cannot affect the default state of the feature group
371 strings and the current state of the OS vendor strings,
372 specifying it multiple times through kernel command line
373 is meaningless. This command is useful when one do not
374 care about the state of the feature group strings which
375 should be controlled by the OSPM.
377 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
378 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
379 can make '_OSI("Windows 2000")' TRUE.
381 'acpi_osi=' cannot be used in combination with other
382 'acpi_osi=' command lines, the _OSI method will not
383 exist in the ACPI namespace. NOTE that such command can
384 only affect the _OSI support state, thus specifying it
385 multiple times through kernel command line is also
388 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
391 'acpi_osi=!*' can be used in combination with single or
392 multiple 'acpi_osi="string1"' to support specific
393 string(s). Note that such command can affect the
394 current state of both the OS vendor strings and the
395 feature group strings, thus specifying it multiple times
396 through kernel command line is meaningful. But it may
397 still not able to affect the final state of a string if
398 there are quirks related to this string. This command
399 is useful when one want to control the state of the
400 feature group strings to debug BIOS issues related to
403 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
404 '_OSI("Module Device")' FALSE.
405 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
406 '_OSI("Module Device")' TRUE.
407 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
409 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
411 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
412 they all will make '_OSI("Windows 2000")' TRUE.
415 Override the pmtimer bug detection: force the kernel
416 to assume that this machine's pmtimer latches its value
417 and always returns good values.
419 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
420 Format: { level | edge | high | low }
422 acpi_skip_timer_override [HW,ACPI]
423 Recognize and ignore IRQ0/pin2 Interrupt Override.
424 For broken nForce2 BIOS resulting in XT-PIC timer.
426 acpi_sleep= [HW,ACPI] Sleep options
427 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
428 old_ordering, nonvs, sci_force_enable }
429 See Documentation/power/video.txt for information on
431 s3_beep is for debugging; it makes the PC's speaker beep
432 as soon as the kernel's real-mode entry point is called.
433 s4_nohwsig prevents ACPI hardware signature from being
434 used during resume from hibernation.
435 old_ordering causes the ACPI 1.0 ordering of the _PTS
436 control method, with respect to putting devices into
437 low power states, to be enforced (the ACPI 2.0 ordering
438 of _PTS is used by default).
439 nonvs prevents the kernel from saving/restoring the
440 ACPI NVS memory during suspend/hibernation and resume.
441 sci_force_enable causes the kernel to set SCI_EN directly
442 on resume from S1/S3 (which is against the ACPI spec,
443 but some broken systems don't work without it).
445 acpi_use_timer_override [HW,ACPI]
446 Use timer override. For some broken Nvidia NF5 boards
447 that require a timer override, but don't have HPET
449 add_efi_memmap [EFI; X86] Include EFI memory map in
450 kernel's map of available physical RAM.
453 { off | try_unsupported }
454 off: disable AGP support
455 try_unsupported: try to drive unsupported chipsets
456 (may crash computer or cause data corruption)
459 See Documentation/sound/alsa/alsa-parameters.txt
462 Allow the default userspace alignment fault handler
463 behaviour to be specified. Bit 0 enables warnings,
464 bit 1 enables fixups, and bit 2 sends a segfault.
466 align_va_addr= [X86-64]
467 Align virtual addresses by clearing slice [14:12] when
468 allocating a VMA at process creation time. This option
469 gives you up to 3% performance improvement on AMD F15h
470 machines (where it is enabled by default) for a
471 CPU-intensive style benchmark, and it can vary highly in
472 a microbenchmark depending on workload and compiler.
474 32: only for 32-bit processes
475 64: only for 64-bit processes
476 on: enable for both 32- and 64-bit processes
477 off: disable for both 32- and 64-bit processes
479 alloc_snapshot [FTRACE]
480 Allocate the ftrace snapshot buffer on boot up when the
481 main buffer is allocated. This is handy if debugging
482 and you need to use tracing_snapshot() on boot up, and
483 do not want to use tracing_snapshot_alloc() as it needs
484 to be done where GFP_KERNEL allocations are allowed.
486 amd_iommu= [HW,X86-64]
487 Pass parameters to the AMD IOMMU driver in the system.
489 fullflush - enable flushing of IO/TLB entries when
490 they are unmapped. Otherwise they are
491 flushed before they will be reused, which
493 off - do not initialize any AMD IOMMU found in
495 force_isolation - Force device isolation for all
496 devices. The IOMMU driver is not
497 allowed anymore to lift isolation
498 requirements as needed. This option
499 does not override iommu=pt
501 amd_iommu_dump= [HW,X86-64]
502 Enable AMD IOMMU driver option to dump the ACPI table
503 for AMD IOMMU. With this option enabled, AMD IOMMU
504 driver will print ACPI tables for AMD IOMMU during
505 IOMMU initialization.
507 amd_iommu_intr= [HW,X86-64]
508 Specifies one of the following AMD IOMMU interrupt
510 legacy - Use legacy interrupt remapping mode.
511 vapic - Use virtual APIC mode, which allows IOMMU
512 to inject interrupts directly into guest.
513 This mode requires kvm-amd.avic=1.
514 (Default when IOMMU HW support is present.)
516 amijoy.map= [HW,JOY] Amiga joystick support
517 Map of devices attached to JOY0DAT and JOY1DAT
519 See also Documentation/input/joystick.txt
521 analog.map= [HW,JOY] Analog joystick and gamepad support
522 Specifies type or capabilities of an analog joystick
523 connected to one of 16 gameports
524 Format: <type1>,<type2>,..<type16>
527 Power management functions (SPARCstation-4/5 + deriv.)
529 Disable APC CPU standby support. SPARCstation-Fox does
530 not play well with APC CPU idle - disable it if you have
531 APC and your system crashes randomly.
533 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
534 Change the output verbosity whilst booting
535 Format: { quiet (default) | verbose | debug }
536 Change the amount of debugging information output
537 when initialising the APIC and IO-APIC components.
539 apic_extnmi= [APIC,X86] External NMI delivery setting
540 Format: { bsp (default) | all | none }
541 bsp: External NMI is delivered only to CPU 0
542 all: External NMIs are broadcast to all CPUs as a
544 none: External NMI is masked for all CPUs. This is
545 useful so that a dump capture kernel won't be
549 See Documentation/networking/ipv6.txt.
551 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
552 Limit apic dumping. The parameter defines the maximal
553 number of local apics being dumped. Also it is possible
554 to set it to "all" by meaning -- no limit here.
555 Format: { 1 (default) | 2 | ... | all }.
556 The parameter valid if only apic=debug or
557 apic=verbose is specified.
558 Example: apic=debug show_lapic=all
560 apm= [APM] Advanced Power Management
561 See header of arch/x86/kernel/apm_32.c.
563 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
564 Format: <io>,<irq>,<nodeID>
568 atarimouse= [HW,MOUSE] Atari Mouse
570 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
571 EzKey and similar keyboards
573 atkbd.reset= [HW] Reset keyboard during initialization
575 atkbd.set= [HW] Select keyboard code set
576 Format: <int> (2 = AT (default), 3 = PS/2)
578 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
581 atkbd.softraw= [HW] Choose between synthetic and real raw mode
582 Format: <bool> (0 = real, 1 = synthetic (default))
584 atkbd.softrepeat= [HW]
585 Use software keyboard repeat
587 audit= [KNL] Enable the audit sub-system
588 Format: { "0" | "1" } (0 = disabled, 1 = enabled)
589 0 - kernel audit is disabled and can not be enabled
590 until the next reboot
591 unset - kernel audit is initialized but disabled and
592 will be fully enabled by the userspace auditd.
593 1 - kernel audit is initialized and partially enabled,
594 storing at most audit_backlog_limit messages in
595 RAM until it is fully enabled by the userspace
599 audit_backlog_limit= [KNL] Set the audit queue size limit.
600 Format: <int> (must be >=0)
603 bau= [X86_UV] Enable the BAU on SGI UV. The default
604 behavior is to disable the BAU (i.e. bau=0).
605 Format: { "0" | "1" }
608 unset - Disable the BAU.
610 baycom_epp= [HW,AX25]
613 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
615 See header of drivers/net/hamradio/baycom_par.c.
617 baycom_ser_fdx= [HW,AX25]
618 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
619 Format: <io>,<irq>,<mode>[,<baud>]
620 See header of drivers/net/hamradio/baycom_ser_fdx.c.
622 baycom_ser_hdx= [HW,AX25]
623 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
624 Format: <io>,<irq>,<mode>
625 See header of drivers/net/hamradio/baycom_ser_hdx.c.
627 blkdevparts= Manual partition parsing of block device(s) for
628 embedded devices based on command line input.
629 See Documentation/block/cmdline-partition.txt
631 boot_delay= Milliseconds to delay each printk during boot.
632 Values larger than 10 seconds (10000) are changed to
636 bootmem_debug [KNL] Enable bootmem allocator debug messages.
639 Disable BERT OS support on buggy BIOSes.
641 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
642 bttv.radio= Most important insmod options are available as
644 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
647 bulk_remove=off [PPC] This parameter disables the use of the pSeries
648 firmware feature for flushing multiple hpte entries
651 c101= [NET] Moxa C101 synchronous serial card
653 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
654 Sometimes CPU hardware bugs make them report the cache
655 size incorrectly. The kernel will attempt work arounds
656 to fix known problems, but for some CPUs it is not
657 possible to determine what the correct size should be.
658 This option provides an override for these situations.
660 ca_keys= [KEYS] This parameter identifies a specific key(s) on
661 the system trusted keyring to be used for certificate
663 format: { id:<keyid> | builtin }
665 cca= [MIPS] Override the kernel pages' cache coherency
666 algorithm. Accepted values range from 0 to 7
667 inclusive. See arch/mips/include/asm/pgtable-bits.h
668 for platform specific values (SB1, Loongson3 and
671 ccw_timeout_log [S390]
672 See Documentation/s390/CommonIO for details.
674 cgroup_disable= [KNL] Disable a particular controller
675 Format: {name of the controller(s) to disable}
676 The effects of cgroup_disable=foo are:
677 - foo isn't auto-mounted if you mount all cgroups in
679 - foo isn't visible as an individually mountable
681 {Currently only "memory" controller deal with this and
682 cut the overhead, others just disable the usage. So
683 only cgroup_disable=memory is actually worthy}
685 cgroup_no_v1= [KNL] Disable one, multiple, all cgroup controllers in v1
686 Format: { controller[,controller...] | "all" }
687 Like cgroup_disable, but only applies to cgroup v1;
688 the blacklisted controllers remain available in cgroup2.
690 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
692 nosocket -- Disable socket memory accounting.
693 nokmem -- Disable kernel memory accounting.
695 checkreqprot [SELINUX] Set initial checkreqprot flag value.
696 Format: { "0" | "1" }
697 See security/selinux/Kconfig help text.
698 0 -- check protection applied by kernel (includes
699 any implied execute protection).
700 1 -- check protection requested by application.
701 Default value is set via a kernel config option.
702 Value can be changed at runtime via
703 /selinux/checkreqprot.
706 See Documentation/s390/CommonIO for details.
709 Prevents the clock framework from automatically gating
710 clocks that have not been explicitly enabled by a Linux
711 device driver but are enabled in hardware at reset or
712 by the bootloader/firmware. Note that this does not
713 force such clocks to be always-on nor does it reserve
714 those clocks in any way. This parameter is useful for
715 debug and development, but should not be needed on a
716 platform with proper driver support. For more
717 information, see Documentation/clk.txt.
719 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
721 Forces specified clocksource (if available) to be used
722 when calculating gettimeofday(). If specified
723 clocksource is not available, it defaults to PIT.
724 Format: { pit | tsc | cyclone | pmtmr }
726 clocksource= Override the default clocksource
728 Override the default clocksource and use the clocksource
729 with the name specified.
730 Some clocksource names to choose from, depending on
732 [all] jiffies (this is the base, fallback clocksource)
734 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
735 pxa_timer,timer3,32k_counter,timer0_1
737 [X86-32] pit,hpet,tsc;
738 scx200_hrt on Geode; cyclone on IBM x440
746 clocksource.arm_arch_timer.evtstrm=
749 Enable/disable the eventstream feature of the ARM
750 architected timer so that code using WFE-based polling
751 loops can be debugged more effectively on production
754 clearcpuid=BITNUM [X86]
755 Disable CPUID feature X for the kernel. See
756 arch/x86/include/asm/cpufeatures.h for the valid bit
757 numbers. Note the Linux specific bits are not necessarily
758 stable over kernel options, but the vendor specific
760 Also note that user programs calling CPUID directly
761 or using the feature without checking anything
762 will still see it. This just prevents it from
763 being used by the kernel or shown in /proc/cpuinfo.
764 Also note the kernel might malfunction if you disable
767 cma=nn[MG]@[start[MG][-end[MG]]]
769 Sets the size of kernel global memory area for
770 contiguous memory allocations and optionally the
771 placement constraint by the physical address range of
772 memory allocations. A value of 0 disables CMA
773 altogether. For more information, see
774 include/linux/dma-contiguous.h
776 cmo_free_hint= [PPC] Format: { yes | no }
777 Specify whether pages are marked as being inactive
778 when they are freed. This is used in CMO environments
779 to determine OS memory pressure for page stealing by
783 coherent_pool=nn[KMG] [ARM,KNL]
784 Sets the size of memory pool for coherent, atomic dma
785 allocations, by default set to 256K.
787 code_bytes [X86] How many bytes of object code to print
792 com20020= [HW,NET] ARCnet - COM20020 chipset
794 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
796 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
800 ARCnet - COM90xx chipset (memory-mapped buffers)
801 Format: <io>[,<irq>[,<memstart>]]
803 condev= [HW,S390] console device
806 console= [KNL] Output console device and options.
808 tty<n> Use the virtual console device <n>.
812 Use the specified serial port. The options are of
813 the form "bbbbpnf", where "bbbb" is the baud rate,
814 "p" is parity ("n", "o", or "e"), "n" is number of
815 bits, and "f" is flow control ("r" for RTS or
816 omit it). Default is "9600n8".
818 See Documentation/serial-console.txt for more
820 Documentation/networking/netconsole.txt for an
823 uart[8250],io,<addr>[,options]
824 uart[8250],mmio,<addr>[,options]
825 uart[8250],mmio16,<addr>[,options]
826 uart[8250],mmio32,<addr>[,options]
827 uart[8250],0x<addr>[,options]
828 Start an early, polled-mode console on the 8250/16550
829 UART at the specified I/O port or MMIO address,
830 switching to the matching ttyS device later.
831 MMIO inter-register address stride is either 8-bit
832 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
833 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
834 to be equivalent to 'mmio'. 'options' are specified in
835 the same format described for ttyS above; if unspecified,
836 the h/w is not re-initialized.
838 hvc<n> Use the hypervisor console device <n>. This is for
839 both Xen and PowerPC hypervisors.
841 If the device connected to the port is not a TTY but a braille
842 device, prepend "brl," before the device type, for instance
844 For now, only VisioBraille is supported.
846 consoleblank= [KNL] The console blank (screen saver) timeout in
847 seconds. Defaults to 10*60 = 10mins. A value of 0
848 disables the blank timer.
851 [KNL] Change the default value for
852 /proc/<pid>/coredump_filter.
853 See also Documentation/filesystems/proc.txt.
855 cpuidle.off=1 [CPU_IDLE]
856 disable the cpuidle sub-system
859 [X86] Delay for N microsec between assert and de-assert
860 of APIC INIT to start processors. This delay occurs
861 on every CPU online, such as boot, and resume from suspend.
864 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
866 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
868 crashkernel=size[KMG][@offset[KMG]]
869 [KNL] Using kexec, Linux can switch to a 'crash kernel'
870 upon panic. This parameter reserves the physical
871 memory region [offset, offset + size] for that kernel
872 image. If '@offset' is omitted, then a suitable offset
873 is selected automatically. Check
874 Documentation/kdump/kdump.txt for further details.
876 crashkernel=range1:size1[,range2:size2,...][@offset]
877 [KNL] Same as above, but depends on the memory
878 in the running system. The syntax of range is
879 start-[end] where start and end are both
880 a memory unit (amount[KMG]). See also
881 Documentation/kdump/kdump.txt for an example.
883 crashkernel=size[KMG],high
884 [KNL, x86_64] range could be above 4G. Allow kernel
885 to allocate physical memory region from top, so could
886 be above 4G if system have more than 4G ram installed.
887 Otherwise memory region will be allocated below 4G, if
889 It will be ignored if crashkernel=X is specified.
890 crashkernel=size[KMG],low
891 [KNL, x86_64] range under 4G. When crashkernel=X,high
892 is passed, kernel could allocate physical memory region
893 above 4G, that cause second kernel crash on system
894 that require some amount of low memory, e.g. swiotlb
895 requires at least 64M+32K low memory, also enough extra
896 low memory is needed to make sure DMA buffers for 32-bit
897 devices won't run out. Kernel would try to allocate at
898 at least 256M below 4G automatically.
899 This one let user to specify own low range under 4G
900 for second kernel instead.
901 0: to disable low allocation.
902 It will be ignored when crashkernel=X,high is not used
903 or memory reserved is below 4G.
906 [KNL] Disable crypto self-tests
911 cs89x0_media= [HW,NET]
912 Format: { rj45 | aui | bnc }
915 See header of drivers/s390/block/dasd_devmap.c.
917 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
918 (one device per port)
919 Format: <port#>,<type>
920 See also Documentation/input/joystick-parport.txt
922 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
923 time. See Documentation/dynamic-debug-howto.txt for
924 details. Deprecated, see dyndbg.
926 debug [KNL] Enable kernel debugging (events log level).
929 [KNL] verbose self-tests
931 Print debugging info while doing the locking API
933 We default to 0 (no extra messages), setting it to
934 1 will print _a lot_ more information - normally
935 only useful to kernel developers.
937 debug_objects [KNL] Enable object debugging
940 [KNL] Disable object debugging
942 debug_guardpage_minorder=
943 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
944 parameter allows control of the order of pages that will
945 be intentionally kept free (and hence protected) by the
946 buddy allocator. Bigger value increase the probability
947 of catching random memory corruption, but reduce the
948 amount of memory for normal system use. The maximum
949 possible value is MAX_ORDER/2. Setting this parameter
950 to 1 or 2 should be enough to identify most random
951 memory corruption problems caused by bugs in kernel or
952 driver code when a CPU writes to (or reads from) a
953 random memory location. Note that there exists a class
954 of memory corruptions problems caused by buggy H/W or
955 F/W or by drivers badly programing DMA (basically when
956 memory is written at bus level and the CPU MMU is
957 bypassed) which are not detectable by
958 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
959 tracking down these problems.
962 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
963 parameter enables the feature at boot time. In
964 default, it is disabled. We can avoid allocating huge
965 chunk of memory for debug pagealloc if we don't enable
966 it at boot time and the system will work mostly same
967 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
968 on: enable the feature
970 debugpat [X86] Enable PAT debugging
972 decnet.addr= [HW,NET]
973 Format: <area>[,<node>]
974 See also Documentation/networking/decnet.txt.
977 [same as hugepagesz=] The size of the default
978 HugeTLB page size. This is the size represented by
979 the legacy /proc/ hugepages APIs, used for SHM, and
980 default size when mounting hugetlbfs filesystems.
981 Defaults to the default architecture's huge page size
985 Set number of hash buckets for dentry cache.
987 disable_1tb_segments [PPC]
988 Disables the use of 1TB hash page table segments. This
989 causes the kernel to fall back to 256MB segments which
990 can be useful when debugging issues that require an SLB
994 See Documentation/networking/ipv6.txt.
997 Disable RADIX MMU mode on POWER9
999 disable_cpu_apicid= [X86,APIC,SMP]
1001 The number of initial APIC ID for the
1002 corresponding CPU to be disabled at boot,
1003 mostly used for the kdump 2nd kernel to
1004 disable BSP to wake up multiple CPUs without
1005 causing system reset or hang due to sending
1006 INIT from AP to BSP.
1008 disable_ddw [PPC/PSERIES]
1009 Disable Dynamic DMA Window support. Use this if
1010 to workaround buggy firmware.
1012 disable_ipv6= [IPV6]
1013 See Documentation/networking/ipv6.txt.
1015 disable_mtrr_cleanup [X86]
1016 The kernel tries to adjust MTRR layout from continuous
1017 to discrete, to make X server driver able to add WB
1018 entry later. This parameter disables that.
1020 disable_mtrr_trim [X86, Intel and AMD only]
1021 By default the kernel will trim any uncacheable
1022 memory out of your available memory pool based on
1023 MTRR settings. This parameter disables that behavior,
1024 possibly causing your machine to run very slowly.
1026 disable_timer_pin_1 [X86]
1027 Disable PIN 1 of APIC timer
1028 Can be useful to work around chipset bugs.
1030 dis_ucode_ldr [X86] Disable the microcode loader.
1032 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1033 this option disables the debugging code at boot.
1035 dma_debug_entries=<number>
1036 This option allows to tune the number of preallocated
1037 entries for DMA-API debugging code. One entry is
1038 required per DMA-API allocation. Use this if the
1039 DMA-API debugging code disables itself because the
1040 architectural default is too low.
1042 dma_debug_driver=<driver_name>
1043 With this option the DMA-API debugging driver
1044 filter feature can be enabled at boot time. Just
1045 pass the driver to filter for as the parameter.
1046 The filter can be disabled or changed to another
1047 driver later using sysfs.
1049 drm_kms_helper.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1050 Broken monitors, graphic adapters, KVMs and EDIDless
1051 panels may send no or incorrect EDID data sets.
1052 This parameter allows to specify an EDID data sets
1053 in the /lib/firmware directory that are used instead.
1054 Generic built-in EDID data sets are used, if one of
1055 edid/1024x768.bin, edid/1280x1024.bin,
1056 edid/1680x1050.bin, or edid/1920x1080.bin is given
1057 and no file with the same name exists. Details and
1058 instructions how to build your own EDID data are
1059 available in Documentation/EDID/HOWTO.txt. An EDID
1060 data set will only be used for a particular connector,
1061 if its name and a colon are prepended to the EDID
1062 name. Each connector may use a unique EDID data
1063 set by separating the files with a comma. An EDID
1064 data set with no connector name will be used for
1065 any connectors not explicitly specified.
1069 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1070 module.dyndbg[="val"]
1071 Enable debug messages at boot time. See
1072 Documentation/dynamic-debug-howto.txt for details.
1074 nompx [X86] Disables Intel Memory Protection Extensions.
1075 See Documentation/x86/intel_mpx.txt for more
1076 information about the feature.
1078 nopku [X86] Disable Memory Protection Keys CPU feature found
1081 module.async_probe [KNL]
1082 Enable asynchronous probe on this module.
1084 early_ioremap_debug [KNL]
1085 Enable debug messages in early_ioremap support. This
1086 is useful for tracking down temporary early mappings
1087 which are not unmapped.
1089 earlycon= [KNL] Output early console device and options.
1091 When used with no options, the early console is
1092 determined by the stdout-path property in device
1095 cdns,<addr>[,options]
1096 Start an early, polled-mode console on a Cadence
1097 (xuartps) serial port at the specified address. Only
1098 supported option is baud rate. If baud rate is not
1099 specified, the serial port must already be setup and
1102 uart[8250],io,<addr>[,options]
1103 uart[8250],mmio,<addr>[,options]
1104 uart[8250],mmio32,<addr>[,options]
1105 uart[8250],mmio32be,<addr>[,options]
1106 uart[8250],0x<addr>[,options]
1107 Start an early, polled-mode console on the 8250/16550
1108 UART at the specified I/O port or MMIO address.
1109 MMIO inter-register address stride is either 8-bit
1110 (mmio) or 32-bit (mmio32 or mmio32be).
1111 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1112 to be equivalent to 'mmio'. 'options' are specified
1113 in the same format described for "console=ttyS<n>"; if
1114 unspecified, the h/w is not initialized.
1118 Start an early, polled-mode console on a pl011 serial
1119 port at the specified address. The pl011 serial port
1120 must already be setup and configured. Options are not
1121 yet supported. If 'mmio32' is specified, then only
1122 the driver will use only 32-bit accessors to read/write
1123 the device registers.
1126 Start an early, polled-mode console on a meson serial
1127 port at the specified address. The serial port must
1128 already be setup and configured. Options are not yet
1132 Start an early, polled-mode console on an msm serial
1133 port at the specified address. The serial port
1134 must already be setup and configured. Options are not
1137 msm_serial_dm,<addr>
1138 Start an early, polled-mode console on an msm serial
1139 dm port at the specified address. The serial port
1140 must already be setup and configured. Options are not
1143 smh Use ARM semihosting calls for early console.
1151 Use early console provided by serial driver available
1152 on Samsung SoCs, requires selecting proper type and
1153 a correct base address of the selected UART port. The
1154 serial port must already be setup and configured.
1155 Options are not yet supported.
1159 Use early console provided by Freescale LP UART driver
1160 found on Freescale Vybrid and QorIQ LS1021A processors.
1161 A valid base address must be provided, and the serial
1162 port must already be setup and configured.
1164 armada3700_uart,<addr>
1165 Start an early, polled-mode console on the
1166 Armada 3700 serial port at the specified
1167 address. The serial port must already be setup
1168 and configured. Options are not yet supported.
1170 earlyprintk= [X86,SH,BLACKFIN,ARM,M68k]
1174 earlyprintk=serial[,ttySn[,baudrate]]
1175 earlyprintk=serial[,0x...[,baudrate]]
1176 earlyprintk=ttySn[,baudrate]
1177 earlyprintk=dbgp[debugController#]
1178 earlyprintk=pciserial,bus:device.function[,baudrate]
1180 earlyprintk is useful when the kernel crashes before
1181 the normal console is initialized. It is not enabled by
1182 default because it has some cosmetic problems.
1184 Append ",keep" to not disable it when the real console
1187 Only one of vga, efi, serial, or usb debug port can
1190 Currently only ttyS0 and ttyS1 may be specified by
1191 name. Other I/O ports may be explicitly specified
1192 on some architectures (x86 and arm at least) by
1193 replacing ttySn with an I/O port address, like this:
1194 earlyprintk=serial,0x1008,115200
1195 You can find the port for a given device in
1196 /proc/tty/driver/serial:
1197 2: uart:ST16650V2 port:00001008 irq:18 ...
1199 Interaction with the standard serial driver is not
1202 The VGA and EFI output is eventually overwritten by
1205 The xen output can only be used by Xen PV guests.
1207 edac_report= [HW,EDAC] Control how to report EDAC event
1208 Format: {"on" | "off" | "force"}
1209 on: enable EDAC to report H/W event. May be overridden
1210 by other higher priority error reporting module.
1211 off: disable H/W event reporting through EDAC.
1212 force: enforce the use of EDAC to report H/W event.
1215 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1218 This is designed to be used in conjunction with
1219 the boot argument: earlyprintk=vga
1222 Format: {"off" | "on" | "skip[mbr]"}
1225 Format: { "old_map", "nochunk", "noruntime", "debug" }
1226 old_map [X86-64]: switch to the old ioremap-based EFI
1227 runtime services mapping. 32-bit still uses this one by
1229 nochunk: disable reading files in "chunks" in the EFI
1230 boot stub, as chunking can cause problems with some
1231 firmware implementations.
1232 noruntime : disable EFI runtime services support
1233 debug: enable misc debug output
1235 efi_no_storage_paranoia [EFI; X86]
1236 Using this parameter you can use more than 50% of
1237 your efi variable storage. Use this parameter only if
1238 you are really sure that your UEFI does sane gc and
1239 fulfills the spec otherwise your board may brick.
1241 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1242 Add arbitrary attribute to specific memory range by
1243 updating original EFI memory map.
1244 Region of memory which aa attribute is added to is
1246 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1247 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1248 attribute is added to range 0x100000000-0x180000000 and
1249 0x10a0000000-0x1120000000.
1251 Using this parameter you can do debugging of EFI memmap
1252 related feature. For example, you can do debugging of
1253 Address Range Mirroring feature even if your box
1256 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1257 that is to be dynamically loaded by Linux. If there are
1258 multiple variables with the same name but with different
1259 vendor GUIDs, all of them will be loaded. See
1260 Documentation/acpi/ssdt-overlays.txt for details.
1263 eisa_irq_edge= [PARISC,HW]
1264 See header of drivers/parisc/eisa.c.
1267 See comment before function elanfreq_setup() in
1268 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1271 Format: {"cfq" | "deadline" | "noop"}
1272 See Documentation/block/cfq-iosched.txt and
1273 Documentation/block/deadline-iosched.txt for details.
1275 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1276 Specifies physical address of start of kernel core
1277 image elf header and optionally the size. Generally
1278 kexec loader will pass this option to capture kernel.
1279 See Documentation/kdump/kdump.txt for details.
1281 enable_mtrr_cleanup [X86]
1282 The kernel tries to adjust MTRR layout from continuous
1283 to discrete, to make X server driver able to add WB
1284 entry later. This parameter enables that.
1286 enable_timer_pin_1 [X86]
1287 Enable PIN 1 of APIC timer
1288 Can be useful to work around chipset bugs
1289 (in particular on some ATI chipsets).
1290 The kernel tries to set a reasonable default.
1292 enforcing [SELINUX] Set initial enforcing status.
1294 See security/selinux/Kconfig help text.
1295 0 -- permissive (log only, no denials).
1296 1 -- enforcing (deny and log).
1298 Value can be changed at runtime via /selinux/enforce.
1301 Disable Error Record Serialization Table (ERST)
1304 ether= [HW,NET] Ethernet cards parameters
1305 This option is obsoleted by the "netdev=" option, which
1306 has equivalent usage. See its documentation for details.
1310 Permit 'security.evm' to be updated regardless of
1311 current integrity status.
1315 fail_make_request=[KNL]
1316 General fault injection mechanism.
1317 Format: <interval>,<probability>,<space>,<times>
1318 See also Documentation/fault-injection/.
1321 See Documentation/blockdev/floppy.txt.
1323 force_pal_cache_flush
1324 [IA-64] Avoid check_sal_cache_flush which may hang on
1325 buggy SAL_CACHE_FLUSH implementations. Using this
1326 parameter will force ia64_sal_cache_flush to call
1327 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1330 Forcefully enable Physical Address Extension (PAE).
1331 Many Pentium M systems disable PAE but may have a
1332 functionally usable PAE implementation.
1333 Warning: use of this parameter will taint the kernel
1334 and may cause unknown problems.
1337 [FTRACE] will set and start the specified tracer
1338 as early as possible in order to facilitate early
1341 ftrace_dump_on_oops[=orig_cpu]
1342 [FTRACE] will dump the trace buffers on oops.
1343 If no parameter is passed, ftrace will dump
1344 buffers of all CPUs, but if you pass orig_cpu, it will
1345 dump only the buffer of the CPU that triggered the
1348 ftrace_filter=[function-list]
1349 [FTRACE] Limit the functions traced by the function
1350 tracer at boot up. function-list is a comma separated
1351 list of functions. This list can be changed at run
1352 time by the set_ftrace_filter file in the debugfs
1355 ftrace_notrace=[function-list]
1356 [FTRACE] Do not trace the functions specified in
1357 function-list. This list can be changed at run time
1358 by the set_ftrace_notrace file in the debugfs
1361 ftrace_graph_filter=[function-list]
1362 [FTRACE] Limit the top level callers functions traced
1363 by the function graph tracer at boot up.
1364 function-list is a comma separated list of functions
1365 that can be changed at run time by the
1366 set_graph_function file in the debugfs tracing directory.
1368 ftrace_graph_notrace=[function-list]
1369 [FTRACE] Do not trace from the functions specified in
1370 function-list. This list is a comma separated list of
1371 functions that can be changed at run time by the
1372 set_graph_notrace file in the debugfs tracing directory.
1375 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1376 support via parallel port (up to 5 devices per port)
1377 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1378 See also Documentation/input/joystick-parport.txt
1382 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1386 gcov_persist= [GCOV] When non-zero (default), profiling data for
1387 kernel modules is saved and remains accessible via
1388 debugfs, even when the module is unloaded/reloaded.
1389 When zero, profiling data is discarded and associated
1390 debugfs files are removed at module unload time.
1392 goldfish [X86] Enable the goldfish android emulator platform.
1393 Don't use this when you are not running on the
1396 gpt [EFI] Forces disk with valid GPT signature but
1397 invalid Protective MBR to be treated as GPT. If the
1398 primary GPT is corrupted, it enables the backup/alternate
1399 GPT to be used instead.
1401 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1402 the "Enable 0" bit of the configuration register.
1405 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1406 the "Enable 0" bit of the configuration register.
1409 grcan.select= [HW] Select which physical interface to use.
1412 grcan.txsize= [HW] Sets the size of the tx buffer.
1413 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1415 grcan.rxsize= [HW] Sets the size of the rx buffer.
1416 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1419 gpio-mockup.gpio_mockup_ranges
1420 [HW] Sets the ranges of gpiochip of for this device.
1421 Format: <start1>,<end1>,<start2>,<end2>...
1423 hardlockup_all_cpu_backtrace=
1424 [KNL] Should the hard-lockup detector generate
1425 backtraces on all cpus.
1428 hashdist= [KNL,NUMA] Large hashes allocated during boot
1429 are distributed across NUMA nodes. Defaults on
1430 for 64-bit NUMA, off otherwise.
1431 Format: 0 | 1 (for off | on)
1433 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1435 hd= [EIDE] (E)IDE hard drive subsystem geometry
1436 Format: <cyl>,<head>,<sect>
1439 Disable Hardware Error Source Table (HEST) support;
1440 corresponding firmware-first mode error processing
1441 logic will be disabled.
1443 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1444 size of <nn>. This works even on boxes that have no
1445 highmem otherwise. This also works to reduce highmem
1446 size on bigger boxes.
1448 highres= [KNL] Enable/disable high resolution timer mode.
1449 Valid parameters: "on", "off"
1453 See Documentation/isdn/README.HiSax.
1457 hpet= [X86-32,HPET] option to control HPET usage
1458 Format: { enable (default) | disable | force |
1460 disable: disable HPET and use PIT instead
1461 force: allow force enabled of undocumented chips (ICH4,
1463 verbose: show contents of HPET registers during setup
1465 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1466 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1468 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1469 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1470 On x86-64 and powerpc, this option can be specified
1471 multiple times interleaved with hugepages= to reserve
1472 huge pages of different sizes. Valid pages sizes on
1473 x86-64 are 2M (when the CPU supports "pse") and 1G
1474 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1476 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1477 terminal devices. Valid values: 0..8
1478 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1479 If specified, z/VM IUCV HVC accepts connections
1480 from listed z/VM user IDs only.
1482 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1483 hardware thread id mappings.
1484 Format: <cpu>:<hwthread>
1487 Do not unregister boot console at start. This is only
1488 useful for debugging when something happens in the window
1489 between unregistering the boot console and initializing
1492 i2c_bus= [HW] Override the default board specific I2C bus speed
1493 or register an additional I2C bus that is not
1494 registered from board initialization code.
1498 i8042.debug [HW] Toggle i8042 debug mode
1499 i8042.unmask_kbd_data
1500 [HW] Enable printing of interrupt data from the KBD port
1501 (disabled by default, and as a pre-condition
1502 requires that i8042.debug=1 be enabled)
1503 i8042.direct [HW] Put keyboard port into non-translated mode
1504 i8042.dumbkbd [HW] Pretend that controller can only read data from
1505 keyboard and cannot control its state
1506 (Don't attempt to blink the leds)
1507 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1508 i8042.nokbd [HW] Don't check/create keyboard port
1509 i8042.noloop [HW] Disable the AUX Loopback command while probing
1511 i8042.nomux [HW] Don't check presence of an active multiplexing
1513 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1515 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1516 i8042.reset [HW] Reset the controller during init, cleanup and
1517 suspend-to-ram transitions, only during s2r
1518 transitions, or never reset
1519 Format: { 1 | Y | y | 0 | N | n }
1520 1, Y, y: always reset controller
1521 0, N, n: don't ever reset controller
1522 Default: only on s2r transitions on x86; most other
1523 architectures force reset to be always executed
1524 i8042.unlock [HW] Unlock (ignore) the keylock
1525 i8042.kbdreset [HW] Reset device connected to KBD port
1529 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1530 indicates that the driver is running on unsupported
1532 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1533 does not match list of supported models.
1535 [HW] Report power status in /proc/i8k
1536 (disabled by default)
1537 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1540 i915.invert_brightness=
1541 [DRM] Invert the sense of the variable that is used to
1542 set the brightness of the panel backlight. Normally a
1543 brightness value of 0 indicates backlight switched off,
1544 and the maximum of the brightness value sets the backlight
1545 to maximum brightness. If this parameter is set to 0
1546 (default) and the machine requires it, or this parameter
1547 is set to 1, a brightness value of 0 sets the backlight
1548 to maximum brightness, and the maximum of the brightness
1549 value switches the backlight off.
1550 -1 -- never invert brightness
1551 0 -- machine default
1552 1 -- force brightness inversion
1555 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1557 ide-core.nodma= [HW] (E)IDE subsystem
1558 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1559 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1560 .cdrom .chs .ignore_cable are additional options
1561 See Documentation/ide/ide.txt.
1563 ide-generic.probe-mask= [HW] (E)IDE subsystem
1565 Probe mask for legacy ISA IDE ports. Depending on
1566 platform up to 6 ports are supported, enabled by
1567 setting corresponding bits in the mask to 1. The
1568 default value is 0x0, which has a special meaning.
1569 On systems that have PCI, it triggers scanning the
1570 PCI bus for the first and the second port, which
1571 are then probed. On systems without PCI the value
1572 of 0x0 enables probing the two first ports as if it
1575 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1576 Claim all unknown PCI IDE storage controllers.
1579 Format: idle=poll, idle=halt, idle=nomwait
1580 Poll forces a polling idle loop that can slightly
1581 improve the performance of waking up a idle CPU, but
1582 will use a lot of power and make the system run hot.
1584 idle=halt: Halt is forced to be used for CPU idle.
1585 In such case C2/C3 won't be used again.
1586 idle=nomwait: Disable mwait for CPU C-states
1588 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1589 Format: { strict | legacy | 2008 | relaxed }
1592 Choose which programs will be accepted for execution
1593 based on the IEEE 754 NaN encoding(s) supported by
1594 the FPU and the NaN encoding requested with the value
1595 of an ELF file header flag individually set by each
1596 binary. Hardware implementations are permitted to
1597 support either or both of the legacy and the 2008 NaN
1600 Available settings are as follows:
1601 strict accept binaries that request a NaN encoding
1602 supported by the FPU
1603 legacy only accept legacy-NaN binaries, if supported
1605 2008 only accept 2008-NaN binaries, if supported
1607 relaxed accept any binaries regardless of whether
1608 supported by the FPU
1610 The FPU emulator is always able to support both NaN
1611 encodings, so if no FPU hardware is present or it has
1612 been disabled with 'nofpu', then the settings of
1613 'legacy' and '2008' strap the emulator accordingly,
1614 'relaxed' straps the emulator for both legacy-NaN and
1615 2008-NaN, whereas 'strict' enables legacy-NaN only on
1616 legacy processors and both NaN encodings on MIPS32 or
1619 The setting for ABS.fmt/NEG.fmt instruction execution
1620 mode generally follows that for the NaN encoding,
1621 except where unsupported by hardware.
1623 ignore_loglevel [KNL]
1624 Ignore loglevel setting - this will print /all/
1625 kernel messages to the console. Useful for debugging.
1626 We also add it as printk module parameter, so users
1627 could change it dynamically, usually by
1628 /sys/module/printk/parameters/ignore_loglevel.
1631 Ignore RLIMIT_DATA setting for data mappings,
1632 print warning at first misuse. Can be changed via
1633 /sys/module/kernel/parameters/ignore_rlimit_data.
1635 ihash_entries= [KNL]
1636 Set number of hash buckets for inode cache.
1638 ima_appraise= [IMA] appraise integrity measurements
1639 Format: { "off" | "enforce" | "fix" | "log" }
1642 ima_appraise_tcb [IMA]
1643 The builtin appraise policy appraises all files
1647 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1651 The list of supported hash algorithms is defined
1652 in crypto/hash_info.h.
1655 The builtin measurement policy to load during IMA
1656 setup. Specyfing "tcb" as the value, measures all
1657 programs exec'd, files mmap'd for exec, and all files
1658 opened with the read mode bit set by either the
1659 effective uid (euid=0) or uid=0.
1662 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1663 Load a policy which meets the needs of the Trusted
1664 Computing Base. This means IMA will measure all
1665 programs exec'd, files mmap'd for exec, and all files
1666 opened for read by uid=0.
1669 Select one of defined IMA measurements template formats.
1670 Formats: { "ima" | "ima-ng" | "ima-sig" }
1674 [IMA] Define a custom template format.
1675 Format: { "field1|...|fieldN" }
1677 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1678 Format: <min_file_size>
1679 Set the minimal file size for using asynchronous hash.
1680 If left unspecified, ahash usage is disabled.
1682 ahash performance varies for different data sizes on
1683 different crypto accelerators. This option can be used
1684 to achieve the best performance for a particular HW.
1686 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1688 Set hashing buffer size. Default: 4k.
1690 ahash performance varies for different chunk sizes on
1691 different crypto accelerators. This option can be used
1692 to achieve best performance for particular HW.
1696 Run specified binary instead of /sbin/init as init
1699 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1700 for working out where the kernel is dying during
1703 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1704 initcall functions. Useful for debugging built-in
1705 modules and initcalls.
1707 initrd= [BOOT] Specify the location of the initial ramdisk
1709 init_pkru= [x86] Specify the default memory protection keys rights
1710 register contents for all processes. 0x55555554 by
1711 default (disallow access to all but pkey 0). Can
1712 override in debugfs after boot.
1714 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1717 int_pln_enable [x86] Enable power limit notification interrupt
1719 integrity_audit=[IMA]
1720 Format: { "0" | "1" }
1721 0 -- basic integrity auditing messages. (Default)
1722 1 -- additional integrity auditing messages.
1724 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1726 Enable intel iommu driver.
1728 Disable intel iommu driver.
1729 igfx_off [Default Off]
1730 By default, gfx is mapped as normal device. If a gfx
1731 device has a dedicated DMAR unit, the DMAR unit is
1732 bypassed by not enabling DMAR with this option. In
1733 this case, gfx device will use physical address for
1736 With this option iommu will not optimize to look
1737 for io virtual address below 32-bit forcing dual
1738 address cycle on pci bus for cards supporting greater
1739 than 32-bit addressing. The default is to look
1740 for translation below 32-bit and if not available
1741 then look in the higher range.
1742 strict [Default Off]
1743 With this option on every unmap_single operation will
1744 result in a hardware IOTLB flush operation as opposed
1745 to batching them for performance.
1746 sp_off [Default Off]
1747 By default, super page will be supported if Intel IOMMU
1748 has the capability. With this option, super page will
1750 ecs_off [Default Off]
1751 By default, extended context tables will be supported if
1752 the hardware advertises that it has support both for the
1753 extended tables themselves, and also PASID support. With
1754 this option set, extended tables will not be used even
1755 on hardware which claims to support them.
1757 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1758 0 disables intel_idle and fall back on acpi_idle.
1759 1 to 9 specify maximum depth of C-state.
1763 Do not enable intel_pstate as the default
1764 scaling driver for the supported processors
1766 Enable intel_pstate on systems that prohibit it by default
1767 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1768 instead of acpi-cpufreq may disable platform features, such
1769 as thermal controls and power capping, that rely on ACPI
1770 P-States information being indicated to OSPM and therefore
1771 should be used with caution. This option does not work with
1772 processors that aren't supported by the intel_pstate driver
1773 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1775 Do not enable hardware P state control (HWP)
1778 Only load intel_pstate on systems which support
1779 hardware P state control (HWP) if available.
1781 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1782 Description Table, specifies preferred power management
1783 profile as "Enterprise Server" or "Performance Server",
1784 then this feature is turned on by default.
1786 intremap= [X86-64, Intel-IOMMU]
1787 on enable Interrupt Remapping (default)
1788 off disable Interrupt Remapping
1789 nosid disable Source ID checking
1791 BIOS x2APIC opt-out request will be ignored
1792 nopost disable Interrupt Posting
1794 iomem= Disable strict checking of access to MMIO memory
1795 strict regions from userspace.
1810 nobypass [PPC/POWERNV]
1811 Disable IOMMU bypass, using IOMMU for PCI devices.
1814 io7= [HW] IO7 for Marvel based alpha systems
1815 See comment before marvel_specify_io7 in
1816 arch/alpha/kernel/core_marvel.c.
1818 io_delay= [X86] I/O delay method
1820 Standard port 0x80 based delay
1822 Alternate port 0xed based delay (needed on some systems)
1824 Simple two microseconds delay
1829 See Documentation/filesystems/nfs/nfsroot.txt.
1831 irqaffinity= [SMP] Set the default irq affinity mask
1832 The argument is a cpu list, as described above.
1835 When an interrupt is not handled search all handlers
1836 for it. Intended to get systems with badly broken
1840 When an interrupt is not handled search all handlers
1841 for it. Also check all handlers each timer
1842 interrupt. Intended to get systems with badly broken
1846 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1848 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1849 The argument is a cpu list, as described above.
1851 This option can be used to specify one or more CPUs
1852 to isolate from the general SMP balancing and scheduling
1853 algorithms. You can move a process onto or off an
1854 "isolated" CPU via the CPU affinity syscalls or cpuset.
1855 <cpu number> begins at 0 and the maximum value is
1856 "number of CPUs in system - 1".
1858 This option is the preferred way to isolate CPUs. The
1859 alternative -- manually setting the CPU mask of all
1860 tasks in the system -- can cause problems and
1861 suboptimal load balancer performance.
1865 ivrs_ioapic [HW,X86_64]
1866 Provide an override to the IOAPIC-ID<->DEVICE-ID
1867 mapping provided in the IVRS ACPI table. For
1868 example, to map IOAPIC-ID decimal 10 to
1869 PCI device 00:14.0 write the parameter as:
1870 ivrs_ioapic[10]=00:14.0
1872 ivrs_hpet [HW,X86_64]
1873 Provide an override to the HPET-ID<->DEVICE-ID
1874 mapping provided in the IVRS ACPI table. For
1875 example, to map HPET-ID decimal 0 to
1876 PCI device 00:14.0 write the parameter as:
1877 ivrs_hpet[0]=00:14.0
1879 ivrs_acpihid [HW,X86_64]
1880 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1881 mapping provided in the IVRS ACPI table. For
1882 example, to map UART-HID:UID AMD0020:0 to
1883 PCI device 00:14.5 write the parameter as:
1884 ivrs_acpihid[00:14.5]=AMD0020:0
1886 js= [HW,JOY] Analog joystick
1887 See Documentation/input/joystick.txt.
1890 When CONFIG_RANDOMIZE_BASE is set, this disables
1891 kernel and module base offset ASLR (Address Space
1892 Layout Randomization).
1896 kernelcore= [KNL,X86,IA-64,PPC]
1897 Format: nn[KMGTPE] | "mirror"
1899 specifies the amount of memory usable by the kernel
1900 for non-movable allocations. The requested amount is
1901 spread evenly throughout all nodes in the system. The
1902 remaining memory in each node is used for Movable
1903 pages. In the event, a node is too small to have both
1904 kernelcore and Movable pages, kernelcore pages will
1905 take priority and other nodes will have a larger number
1906 of Movable pages. The Movable zone is used for the
1907 allocation of pages that may be reclaimed or moved
1908 by the page migration subsystem. This means that
1909 HugeTLB pages may not be allocated from this zone.
1910 Note that allocations like PTEs-from-HighMem still
1911 use the HighMem zone if it exists, and the Normal
1912 zone if it does not.
1914 Instead of specifying the amount of memory (nn[KMGTPE]),
1915 you can specify "mirror" option. In case "mirror"
1916 option is specified, mirrored (reliable) memory is used
1917 for non-movable allocations and remaining memory is used
1918 for Movable pages. nn[KMGTPE] and "mirror" are exclusive,
1919 so you can NOT specify nn[KMGTPE] and "mirror" at the same
1922 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1923 Format: <Controller#>[,poll interval]
1924 The controller # is the number of the ehci usb debug
1925 port as it is probed via PCI. The poll interval is
1926 optional and is the number seconds in between
1927 each poll cycle to the debug port in case you need
1928 the functionality for interrupting the kernel with
1929 gdb or control-c on the dbgp connection. When
1930 not using this parameter you use sysrq-g to break into
1931 the kernel debugger.
1933 kgdboc= [KGDB,HW] kgdb over consoles.
1934 Requires a tty driver that supports console polling,
1935 or a supported polling keyboard driver (non-usb).
1936 Serial only format: <serial_device>[,baud]
1937 keyboard only format: kbd
1938 keyboard and serial format: kbd,<serial_device>[,baud]
1939 Optional Kernel mode setting:
1940 kms, kbd format: kms,kbd
1941 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1943 kgdbwait [KGDB] Stop kernel execution and enter the
1944 kernel debugger at the earliest opportunity.
1946 kmac= [MIPS] korina ethernet MAC address.
1947 Configure the RouterBoard 532 series on-chip
1948 Ethernet adapter MAC address.
1950 kmemleak= [KNL] Boot-time kmemleak enable/disable
1951 Valid arguments: on, off
1953 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1956 kmemcheck= [X86] Boot-time kmemcheck enable/disable/one-shot mode
1957 Valid arguments: 0, 1, 2
1958 kmemcheck=0 (disabled)
1959 kmemcheck=1 (enabled)
1960 kmemcheck=2 (one-shot mode)
1961 Default: 2 (one-shot mode)
1963 kpti= [ARM64] Control page table isolation of user
1964 and kernel address spaces.
1965 Default: enabled on cores which need mitigation.
1969 kstack=N [X86] Print N words from the kernel stack
1972 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1973 Default is 0 (don't ignore, but inject #GP)
1975 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1980 [KVM] Controls the software workaround for the
1981 X86_BUG_ITLB_MULTIHIT bug.
1982 force : Always deploy workaround.
1983 off : Never deploy workaround.
1984 auto : Deploy workaround based on the presence of
1985 X86_BUG_ITLB_MULTIHIT.
1989 If the software workaround is enabled for the host,
1990 guests do need not to enable it for nested guests.
1992 kvm.nx_huge_pages_recovery_ratio=
1993 [KVM] Controls how many 4KiB pages are periodically zapped
1994 back to huge pages. 0 disables the recovery, otherwise if
1995 the value is N KVM will zap 1/Nth of the 4KiB pages every
1996 minute. The default is 60.
1998 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1999 Default is 1 (enabled)
2001 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2003 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2005 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2006 (virtualized MMU) support on capable Intel chips.
2007 Default is 1 (enabled)
2009 kvm-intel.emulate_invalid_guest_state=
2010 [KVM,Intel] Enable emulation of invalid guest states
2011 Default is 0 (disabled)
2013 kvm-intel.flexpriority=
2014 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2015 Default is 1 (enabled)
2018 [KVM,Intel] Enable VMX nesting (nVMX).
2019 Default is 0 (disabled)
2021 kvm-intel.unrestricted_guest=
2022 [KVM,Intel] Disable unrestricted guest feature
2023 (virtualized real and unpaged mode) on capable
2024 Intel chips. Default is 1 (enabled)
2026 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2029 Valid arguments: never, cond, always
2031 always: L1D cache flush on every VMENTER.
2032 cond: Flush L1D on VMENTER only when the code between
2033 VMEXIT and VMENTER can leak host memory.
2034 never: Disables the mitigation
2036 Default is cond (do L1 cache flush in specific instances)
2038 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2039 feature (tagged TLBs) on capable Intel chips.
2040 Default is 1 (enabled)
2042 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2045 The kernel PTE inversion protection is unconditionally
2046 enabled and cannot be disabled.
2049 Provides all available mitigations for the
2050 L1TF vulnerability. Disables SMT and
2051 enables all mitigations in the
2052 hypervisors, i.e. unconditional L1D flush.
2054 SMT control and L1D flush control via the
2055 sysfs interface is still possible after
2056 boot. Hypervisors will issue a warning
2057 when the first VM is started in a
2058 potentially insecure configuration,
2059 i.e. SMT enabled or L1D flush disabled.
2062 Same as 'full', but disables SMT and L1D
2063 flush runtime control. Implies the
2064 'nosmt=force' command line option.
2065 (i.e. sysfs control of SMT is disabled.)
2068 Leaves SMT enabled and enables the default
2069 hypervisor mitigation, i.e. conditional
2072 SMT control and L1D flush control via the
2073 sysfs interface is still possible after
2074 boot. Hypervisors will issue a warning
2075 when the first VM is started in a
2076 potentially insecure configuration,
2077 i.e. SMT enabled or L1D flush disabled.
2081 Disables SMT and enables the default
2082 hypervisor mitigation.
2084 SMT control and L1D flush control via the
2085 sysfs interface is still possible after
2086 boot. Hypervisors will issue a warning
2087 when the first VM is started in a
2088 potentially insecure configuration,
2089 i.e. SMT enabled or L1D flush disabled.
2092 Same as 'flush', but hypervisors will not
2093 warn when a VM is started in a potentially
2094 insecure configuration.
2097 Disables hypervisor mitigations and doesn't
2099 It also drops the swap size and available
2100 RAM limit restriction on both hypervisor and
2105 For details see: Documentation/hw-vuln/l1tf.rst
2111 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2114 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2115 value for LAPIC timer one-shot implementation. Default
2116 back to the programmable timer unit in the LAPIC.
2118 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2121 libata.dma= [LIBATA] DMA control
2122 libata.dma=0 Disable all PATA and SATA DMA
2123 libata.dma=1 PATA and SATA Disk DMA only
2124 libata.dma=2 ATAPI (CDROM) DMA only
2125 libata.dma=4 Compact Flash DMA only
2126 Combinations also work, so libata.dma=3 enables DMA
2127 for disks and CDROMs, but not CFs.
2129 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2130 libata.ignore_hpa=0 keep BIOS limits (default)
2131 libata.ignore_hpa=1 ignore limits, using full disk
2133 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2137 libata.force= [LIBATA] Force configurations. The format is comma
2138 separated list of "[ID:]VAL" where ID is
2139 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2140 matching port, link or device. Basically, it matches
2141 the ATA ID string printed on console by libata. If
2142 the whole ID part is omitted, the last PORT and DEVICE
2143 values are used. If ID hasn't been specified yet, the
2144 configuration applies to all ports, links and devices.
2146 If only DEVICE is omitted, the parameter applies to
2147 the port and all links and devices behind it. DEVICE
2148 number of 0 either selects the first device or the
2149 first fan-out link behind PMP device. It does not
2150 select the host link. DEVICE number of 15 selects the
2151 host link and device attached to it.
2153 The VAL specifies the configuration to force. As long
2154 as there's no ambiguity shortcut notation is allowed.
2155 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2156 The following configurations can be forced.
2158 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2159 Any ID with matching PORT is used.
2161 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2163 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2164 udma[/][16,25,33,44,66,100,133] notation is also
2167 * [no]ncq: Turn on or off NCQ.
2169 * [no]ncqtrim: Turn off queued DSM TRIM.
2171 * nohrst, nosrst, norst: suppress hard, soft
2174 * rstonce: only attempt one reset during
2175 hot-unplug link recovery
2177 * dump_id: dump IDENTIFY data.
2179 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2181 * disable: Disable this device.
2183 If there are multiple matching configurations changing
2184 the same attribute, the last one is used.
2186 memblock=debug [KNL] Enable memblock debug messages.
2188 load_ramdisk= [RAM] List of ramdisks to load from floppy
2189 See Documentation/blockdev/ramdisk.txt.
2191 lockd.nlm_grace_period=P [NFS] Assign grace period.
2194 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2197 lockd.nlm_timeout=T [NFS] Assign timeout value.
2200 lockd.nlm_udpport=M [NFS] Assign UDP port.
2203 locktorture.nreaders_stress= [KNL]
2204 Set the number of locking read-acquisition kthreads.
2205 Defaults to being automatically set based on the
2206 number of online CPUs.
2208 locktorture.nwriters_stress= [KNL]
2209 Set the number of locking write-acquisition kthreads.
2211 locktorture.onoff_holdoff= [KNL]
2212 Set time (s) after boot for CPU-hotplug testing.
2214 locktorture.onoff_interval= [KNL]
2215 Set time (s) between CPU-hotplug operations, or
2216 zero to disable CPU-hotplug testing.
2218 locktorture.shuffle_interval= [KNL]
2219 Set task-shuffle interval (jiffies). Shuffling
2220 tasks allows some CPUs to go into dyntick-idle
2221 mode during the locktorture test.
2223 locktorture.shutdown_secs= [KNL]
2224 Set time (s) after boot system shutdown. This
2225 is useful for hands-off automated testing.
2227 locktorture.stat_interval= [KNL]
2228 Time (s) between statistics printk()s.
2230 locktorture.stutter= [KNL]
2231 Time (s) to stutter testing, for example,
2232 specifying five seconds causes the test to run for
2233 five seconds, wait for five seconds, and so on.
2234 This tests the locking primitive's ability to
2235 transition abruptly to and from idle.
2237 locktorture.torture_runnable= [BOOT]
2238 Start locktorture running at boot time.
2240 locktorture.torture_type= [KNL]
2241 Specify the locking implementation to test.
2243 locktorture.verbose= [KNL]
2244 Enable additional printk() statements.
2246 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2249 loglevel= All Kernel Messages with a loglevel smaller than the
2250 console loglevel will be printed to the console. It can
2251 also be changed with klogd or other programs. The
2252 loglevels are defined as follows:
2254 0 (KERN_EMERG) system is unusable
2255 1 (KERN_ALERT) action must be taken immediately
2256 2 (KERN_CRIT) critical conditions
2257 3 (KERN_ERR) error conditions
2258 4 (KERN_WARNING) warning conditions
2259 5 (KERN_NOTICE) normal but significant condition
2260 6 (KERN_INFO) informational
2261 7 (KERN_DEBUG) debug-level messages
2263 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2264 in bytes. n must be a power of two and greater
2265 than the minimal size. The minimal size is defined
2266 by LOG_BUF_SHIFT kernel config parameter. There is
2267 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2268 that allows to increase the default size depending on
2269 the number of CPUs. See init/Kconfig for more details.
2271 logo.nologo [FB] Disables display of the built-in Linux logo.
2272 This may be used to provide more screen space for
2273 kernel log messages and is useful when debugging
2274 kernel boot problems.
2276 lp=0 [LP] Specify parallel ports to use, e.g,
2277 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2278 lp=reset first parallel port). 'lp=0' disables the
2279 lp=auto printer driver. 'lp=reset' (which can be
2280 specified in addition to the ports) causes
2281 attached printers to be reset. Using
2282 lp=port1,port2,... specifies the parallel ports
2283 to associate lp devices with, starting with
2284 lp0. A port specification may be 'none' to skip
2285 that lp device, or a parport name such as
2286 'parport0'. Specifying 'lp=auto' instead of a
2287 port specification list means that device IDs
2288 from each port should be examined, to see if
2289 an IEEE 1284-compliant printer is attached; if
2290 so, the driver will manage that printer.
2291 See also header of drivers/char/lp.c.
2294 Sets loops_per_jiffy to given constant, thus avoiding
2295 time-consuming boot-time autodetection (up to 250 ms per
2296 CPU). 0 enables autodetection (default). To determine
2297 the correct value for your kernel, boot with normal
2298 autodetection and see what value is printed. Note that
2299 on SMP systems the preset will be applied to all CPUs,
2300 which is likely to cause problems if your CPUs need
2301 significantly divergent settings. An incorrect value
2302 will cause delays in the kernel to be wrong, leading to
2303 unpredictable I/O errors and other breakage. Although
2304 unlikely, in the extreme case this might damage your
2308 Format: <io>,<irq>,<dma>
2310 machvec= [IA-64] Force the use of a particular machine-vector
2311 (machvec) in a generic kernel.
2312 Example: machvec=hpzx1_swiotlb
2314 machtype= [Loongson] Share the same kernel image file between different
2316 Example: machtype=lemote-yeeloong-2f-7inch
2318 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2319 than or equal to this physical address is ignored.
2321 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2322 will bring up during bootup. maxcpus=n : n >= 0 limits
2323 the kernel to bring up 'n' processors. Surely after
2324 bootup you can bring up the other plugged cpu by executing
2325 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2326 only takes effect during system bootup.
2327 While n=0 is a special case, it is equivalent to "nosmp",
2328 which also disables the IO APIC.
2330 max_loop= [LOOP] The number of loop block devices that get
2331 (loop.max_loop) unconditionally pre-created at init time. The default
2332 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2333 of statically allocating a predefined number, loop
2334 devices can be requested on-demand with the
2335 /dev/loop-control interface.
2337 mce [X86-32] Machine Check Exception
2339 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2341 md= [HW] RAID subsystems devices and level
2342 See Documentation/md.txt.
2345 Format: <first>,<last>
2346 Specifies range of consoles to be captured by the MDA.
2349 Control mitigation for the Micro-architectural Data
2350 Sampling (MDS) vulnerability.
2352 Certain CPUs are vulnerable to an exploit against CPU
2353 internal buffers which can forward information to a
2354 disclosure gadget under certain conditions.
2356 In vulnerable processors, the speculatively
2357 forwarded data can be used in a cache side channel
2358 attack, to access data to which the attacker does
2359 not have direct access.
2361 This parameter controls the MDS mitigation. The
2364 full - Enable MDS mitigation on vulnerable CPUs
2365 full,nosmt - Enable MDS mitigation and disable
2366 SMT on vulnerable CPUs
2367 off - Unconditionally disable MDS mitigation
2369 On TAA-affected machines, mds=off can be prevented by
2370 an active TAA mitigation as both vulnerabilities are
2371 mitigated with the same mechanism so in order to disable
2372 this mitigation, you need to specify tsx_async_abort=off
2375 Not specifying this option is equivalent to
2378 For details see: Documentation/hw-vuln/mds.rst
2380 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2381 Amount of memory to be used when the kernel is not able
2382 to see the whole system memory or for test.
2383 [X86] Work as limiting max address. Use together
2384 with memmap= to avoid physical address space collisions.
2385 Without memmap= PCI devices could be placed at addresses
2386 belonging to unused RAM.
2388 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2392 [KNL,SH] Allow user to override the default size for
2393 per-device physically contiguous DMA buffers.
2395 memhp_default_state=online/offline
2396 [KNL] Set the initial state for the memory hotplug
2397 onlining policy. If not specified, the default value is
2398 set according to the
2399 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2401 See Documentation/memory-hotplug.txt.
2403 memmap=exactmap [KNL,X86] Enable setting of an exact
2404 E820 memory map, as specified by the user.
2405 Such memmap=exactmap lines can be constructed based on
2406 BIOS output or other requirements. See the memmap=nn@ss
2409 memmap=nn[KMG]@ss[KMG]
2410 [KNL] Force usage of a specific region of memory.
2411 Region of memory to be used is from ss to ss+nn.
2413 memmap=nn[KMG]#ss[KMG]
2414 [KNL,ACPI] Mark specific memory as ACPI data.
2415 Region of memory to be marked is from ss to ss+nn.
2417 memmap=nn[KMG]$ss[KMG]
2418 [KNL,ACPI] Mark specific memory as reserved.
2419 Region of memory to be reserved is from ss to ss+nn.
2420 Example: Exclude memory from 0x18690000-0x1869ffff
2421 memmap=64K$0x18690000
2423 memmap=0x10000$0x18690000
2425 memmap=nn[KMG]!ss[KMG]
2426 [KNL,X86] Mark specific memory as protected.
2427 Region of memory to be used, from ss to ss+nn.
2428 The memory region may be marked as e820 type 12 (0xc)
2429 and is NVDIMM or ADR memory.
2431 memory_corruption_check=0/1 [X86]
2432 Some BIOSes seem to corrupt the first 64k of
2433 memory when doing things like suspend/resume.
2434 Setting this option will scan the memory
2435 looking for corruption. Enabling this will
2436 both detect corruption and prevent the kernel
2437 from using the memory being corrupted.
2438 However, its intended as a diagnostic tool; if
2439 repeatable BIOS-originated corruption always
2440 affects the same memory, you can use memmap=
2441 to prevent the kernel from using that memory.
2443 memory_corruption_check_size=size [X86]
2444 By default it checks for corruption in the low
2445 64k, making this memory unavailable for normal
2446 use. Use this parameter to scan for
2447 corruption in more or less memory.
2449 memory_corruption_check_period=seconds [X86]
2450 By default it checks for corruption every 60
2451 seconds. Use this parameter to check at some
2452 other rate. 0 disables periodic checking.
2454 memtest= [KNL,X86,ARM] Enable memtest
2456 default : 0 <disable>
2457 Specifies the number of memtest passes to be
2458 performed. Each pass selects another test
2459 pattern from a given set of patterns. Memtest
2460 fills the memory with this pattern, validates
2461 memory contents and reserves bad memory
2462 regions that are detected.
2464 meye.*= [HW] Set MotionEye Camera parameters
2465 See Documentation/video4linux/meye.txt.
2467 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2468 Multi-Function General Purpose Timers on AMD Geode
2471 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2472 the BIOS has incorrectly applied a workaround. TinyBIOS
2473 version 0.98 is known to be affected, 0.99 fixes the
2474 problem by letting the user disable the workaround.
2478 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2479 physical address is ignored.
2481 mini2440= [ARM,HW,KNL]
2482 Format:[0..2][b][c][t]
2484 MINI2440 configuration specification:
2485 0 - The attached screen is the 3.5" TFT
2486 1 - The attached screen is the 7" TFT
2487 2 - The VGA Shield is attached (1024x768)
2488 Leaving out the screen size parameter will not load
2489 the TFT driver, and the framebuffer will be left
2491 b - Enable backlight. The TFT backlight pin will be
2492 linked to the kernel VESA blanking code and a GPIO
2493 LED. This parameter is not necessary when using the
2495 c - Enable the s3c camera interface.
2496 t - Reserved for enabling touchscreen support. The
2497 touchscreen support is not enabled in the mainstream
2498 kernel as of 2.6.30, a preliminary port can be found
2499 in the "bleeding edge" mini2440 support kernel at
2500 http://repo.or.cz/w/linux-2.6/mini2440.git
2503 [X86] Control optional mitigations for CPU
2504 vulnerabilities. This is a set of curated,
2505 arch-independent options, each of which is an
2506 aggregation of existing arch-specific options.
2509 Disable all optional CPU mitigations. This
2510 improves system performance, but it may also
2511 expose users to several CPU vulnerabilities.
2512 Equivalent to: nopti [X86]
2515 spectre_v2_user=off [X86]
2516 spec_store_bypass_disable=off [X86]
2519 tsx_async_abort=off [X86]
2520 kvm.nx_huge_pages=off [X86]
2521 no_entry_flush [PPC]
2522 no_uaccess_flush [PPC]
2525 This does not have any effect on
2526 kvm.nx_huge_pages when
2527 kvm.nx_huge_pages=force.
2530 Mitigate all CPU vulnerabilities, but leave SMT
2531 enabled, even if it's vulnerable. This is for
2532 users who don't want to be surprised by SMT
2533 getting disabled across kernel upgrades, or who
2534 have other ways of avoiding SMT-based attacks.
2535 Equivalent to: (default behavior)
2538 Mitigate all CPU vulnerabilities, disabling SMT
2539 if needed. This is for users who always want to
2540 be fully mitigated, even if it means losing SMT.
2541 Equivalent to: l1tf=flush,nosmt [X86]
2542 mds=full,nosmt [X86]
2543 tsx_async_abort=full,nosmt [X86]
2546 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2547 parameter allows control of the logging verbosity for
2548 the additional memory initialisation checks. A value
2549 of 0 disables mminit logging and a level of 4 will
2550 log everything. Information is printed at KERN_DEBUG
2551 so loglevel=8 may also need to be specified.
2554 [KNL] When CONFIG_MODULE_SIG is set, this means that
2555 modules without (valid) signatures will fail to load.
2556 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2557 is always true, so this option does nothing.
2559 module_blacklist= [KNL] Do not load a comma-separated list of
2560 modules. Useful for debugging problem modules.
2563 [MOUSE] Maximum time between finger touching and
2564 leaving touchpad surface for touch to be considered
2565 a tap and be reported as a left button click (for
2566 touchpads working in absolute mode only).
2568 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2569 reporting absolute coordinates, such as tablets
2570 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2571 reporting absolute coordinates, such as tablets
2573 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
2574 is similar to kernelcore except it specifies the
2575 amount of memory used for migratable allocations.
2576 If both kernelcore and movablecore is specified,
2577 then kernelcore will be at *least* the specified
2578 value but may be more. If movablecore on its own
2579 is specified, the administrator must be careful
2580 that the amount of memory usable for all allocations
2583 movable_node [KNL,X86] Boot-time switch to enable the effects
2584 of CONFIG_MOVABLE_NODE=y. See mm/Kconfig for details.
2586 MTD_Partition= [MTD]
2587 Format: <name>,<region-number>,<size>,<offset>
2589 MTD_Region= [MTD] Format:
2590 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2593 See drivers/mtd/cmdlinepart.c.
2595 multitce=off [PPC] This parameter disables the use of the pSeries
2596 firmware feature for updating multiple TCE entries
2599 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2601 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2603 boundary - index of last SLC block on Flex-OneNAND.
2604 The remaining blocks are configured as MLC blocks.
2605 lock - Configure if Flex-OneNAND boundary should be locked.
2606 Once locked, the boundary cannot be changed.
2607 1 indicates lock status, 0 indicates unlock status.
2610 ARM/S3C2412 JIVE boot control
2612 See arch/arm/mach-s3c2412/mach-jive.c
2614 mtouchusb.raw_coordinates=
2615 [HW] Make the MicroTouch USB driver use raw coordinates
2616 ('y', default) or cooked coordinates ('n')
2618 mtrr_chunk_size=nn[KMG] [X86]
2619 used for mtrr cleanup. It is largest continuous chunk
2620 that could hold holes aka. UC entries.
2622 mtrr_gran_size=nn[KMG] [X86]
2623 Used for mtrr cleanup. It is granularity of mtrr block.
2625 Large value could prevent small alignment from
2628 mtrr_spare_reg_nr=n [X86]
2630 Range: 0,7 : spare reg number
2632 Used for mtrr cleanup. It is spare mtrr entries number.
2633 Set to 2 or more if your graphical card needs more.
2635 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2637 netdev= [NET] Network devices parameters
2638 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2639 Note that mem_start is often overloaded to mean
2640 something different and driver-specific.
2641 This usage is only documented in each driver source
2645 [NETFILTER] Enable connection tracking flow accounting
2646 0 to disable accounting
2647 1 to enable accounting
2650 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2651 See Documentation/filesystems/nfs/nfsroot.txt.
2653 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2654 See Documentation/filesystems/nfs/nfsroot.txt.
2656 nfsrootdebug [NFS] enable nfsroot debugging messages.
2657 See Documentation/filesystems/nfs/nfsroot.txt.
2659 nfs.callback_nr_threads=
2660 [NFSv4] set the total number of threads that the
2661 NFS client will assign to service NFSv4 callback
2664 nfs.callback_tcpport=
2665 [NFS] set the TCP port on which the NFSv4 callback
2666 channel should listen.
2669 [NFS] sets the pathname to the program which is used
2670 to update the NFS client cache entries.
2672 nfs.cache_getent_timeout=
2673 [NFS] sets the timeout after which an attempt to
2674 update a cache entry is deemed to have failed.
2676 nfs.idmap_cache_timeout=
2677 [NFS] set the maximum lifetime for idmapper cache
2681 [NFS] enable 64-bit inode numbers.
2682 If zero, the NFS client will fake up a 32-bit inode
2683 number for the readdir() and stat() syscalls instead
2684 of returning the full 64-bit number.
2685 The default is to return 64-bit inode numbers.
2687 nfs.max_session_cb_slots=
2688 [NFSv4.1] Sets the maximum number of session
2689 slots the client will assign to the callback
2690 channel. This determines the maximum number of
2691 callbacks the client will process in parallel for
2692 a particular server.
2694 nfs.max_session_slots=
2695 [NFSv4.1] Sets the maximum number of session slots
2696 the client will attempt to negotiate with the server.
2697 This limits the number of simultaneous RPC requests
2698 that the client can send to the NFSv4.1 server.
2699 Note that there is little point in setting this
2700 value higher than the max_tcp_slot_table_limit.
2702 nfs.nfs4_disable_idmapping=
2703 [NFSv4] When set to the default of '1', this option
2704 ensures that both the RPC level authentication
2705 scheme and the NFS level operations agree to use
2706 numeric uids/gids if the mount is using the
2707 'sec=sys' security flavour. In effect it is
2708 disabling idmapping, which can make migration from
2709 legacy NFSv2/v3 systems to NFSv4 easier.
2710 Servers that do not support this mode of operation
2711 will be autodetected by the client, and it will fall
2712 back to using the idmapper.
2713 To turn off this behaviour, set the value to '0'.
2715 [NFS4] Specify an additional fixed unique ident-
2716 ification string that NFSv4 clients can insert into
2717 their nfs_client_id4 string. This is typically a
2718 UUID that is generated at system install time.
2720 nfs.send_implementation_id =
2721 [NFSv4.1] Send client implementation identification
2722 information in exchange_id requests.
2723 If zero, no implementation identification information
2725 The default is to send the implementation identification
2728 nfs.recover_lost_locks =
2729 [NFSv4] Attempt to recover locks that were lost due
2730 to a lease timeout on the server. Please note that
2731 doing this risks data corruption, since there are
2732 no guarantees that the file will remain unchanged
2733 after the locks are lost.
2734 If you want to enable the kernel legacy behaviour of
2735 attempting to recover these locks, then set this
2737 The default parameter value of '0' causes the kernel
2738 not to attempt recovery of lost locks.
2740 nfs4.layoutstats_timer =
2741 [NFSv4.2] Change the rate at which the kernel sends
2742 layoutstats to the pNFS metadata server.
2744 Setting this to value to 0 causes the kernel to use
2745 whatever value is the default set by the layout
2746 driver. A non-zero value sets the minimum interval
2747 in seconds between layoutstats transmissions.
2749 nfsd.nfs4_disable_idmapping=
2750 [NFSv4] When set to the default of '1', the NFSv4
2751 server will return only numeric uids and gids to
2752 clients using auth_sys, and will accept numeric uids
2753 and gids from such clients. This is intended to ease
2754 migration from NFSv2/v3.
2756 objlayoutdriver.osd_login_prog=
2757 [NFS] [OBJLAYOUT] sets the pathname to the program which
2758 is used to automatically discover and login into new
2759 osd-targets. Please see:
2760 Documentation/filesystems/pnfs.txt for more explanations
2762 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
2763 when a NMI is triggered.
2764 Format: [state][,regs][,debounce][,die]
2766 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2767 Format: [panic,][nopanic,][num]
2769 0 - turn hardlockup detector in nmi_watchdog off
2770 1 - turn hardlockup detector in nmi_watchdog on
2771 When panic is specified, panic when an NMI watchdog
2772 timeout occurs (or 'nopanic' to override the opposite
2773 default). To disable both hard and soft lockup detectors,
2774 please see 'nowatchdog'.
2775 This is useful when you use a panic=... timeout and
2776 need the box quickly up again.
2778 netpoll.carrier_timeout=
2779 [NET] Specifies amount of time (in seconds) that
2780 netpoll should wait for a carrier. By default netpoll
2783 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2784 emulation library even if a 387 maths coprocessor
2788 [HW] Never suspend the console
2789 Disable suspending of consoles during suspend and
2790 hibernate operations. Once disabled, debugging
2791 messages can reach various consoles while the rest
2792 of the system is being put to sleep (ie, while
2793 debugging driver suspend/resume hooks). This may
2794 not work reliably with all consoles, but is known
2795 to work with serial and VGA consoles.
2796 To facilitate more flexible debugging, we also add
2797 console_suspend, a printk module parameter to control
2798 it. Users could use console_suspend (usually
2799 /sys/module/printk/parameters/console_suspend) to
2800 turn on/off it dynamically.
2802 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2803 caches in the slab allocator. Saves per-node memory,
2804 but will impact performance.
2808 noaltinstr [S390] Disables alternative instructions patching
2809 (CPU alternatives feature).
2811 noapic [SMP,APIC] Tells the kernel to not make use of any
2812 IOAPICs that may be present in the system.
2814 noautogroup Disable scheduler automatic task group creation.
2816 nobats [PPC] Do not use BATs for mapping kernel lowmem
2817 on "Classic" PPC cores.
2821 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2823 nodelayacct [KNL] Disable per-task delay accounting
2825 nodsp [SH] Disable hardware DSP at boot time.
2827 noefi Disable EFI runtime services support.
2829 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
2834 On X86-32 available only on PAE configured kernels.
2835 noexec=on: enable non-executable mappings (default)
2836 noexec=off: disable non-executable mappings
2839 Disable SMAP (Supervisor Mode Access Prevention)
2840 even if it is supported by processor.
2843 Disable SMEP (Supervisor Mode Execution Prevention)
2844 even if it is supported by processor.
2847 This affects only 32-bit executables.
2848 noexec32=on: enable non-executable mappings (default)
2849 read doesn't imply executable mappings
2850 noexec32=off: disable non-executable mappings
2851 read implies executable mappings
2853 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2855 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2856 register save and restore. The kernel will only save
2857 legacy floating-point registers on task switch.
2859 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2861 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2862 Equivalent to smt=1.
2864 [KNL,x86] Disable symmetric multithreading (SMT).
2865 nosmt=force: Force disable SMT, cannot be undone
2866 via the sysfs control file.
2868 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
2869 (bounds check bypass). With this option data leaks are
2870 possible in the system.
2872 nospectre_v2 [X86,PPC_FSL_BOOK3E] Disable all mitigations for the Spectre variant 2
2873 (indirect branch prediction) vulnerability. System may
2874 allow data leaks with this option, which is equivalent
2877 nospec_store_bypass_disable
2878 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
2881 [PPC] Don't flush the L1-D cache after accessing user data.
2883 noxsave [BUGS=X86] Disables x86 extended register state save
2884 and restore using xsave. The kernel will fallback to
2885 enabling legacy floating-point and sse state.
2887 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2888 register states. The kernel will fall back to use
2889 xsave to save the states. By using this parameter,
2890 performance of saving the states is degraded because
2891 xsave doesn't support modified optimization while
2892 xsaveopt supports it on xsaveopt enabled systems.
2894 noxsaves [X86] Disables xsaves and xrstors used in saving and
2895 restoring x86 extended register state in compacted
2896 form of xsave area. The kernel will fall back to use
2897 xsaveopt and xrstor to save and restore the states
2898 in standard form of xsave area. By using this
2899 parameter, xsave area per process might occupy more
2900 memory on xsaves enabled systems.
2902 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2903 wfi(ARM) instruction doesn't work correctly and not to
2904 use it. This is also useful when using JTAG debugger.
2906 no_file_caps Tells the kernel not to honor file capabilities. The
2907 only way then for a file to be executed with privilege
2908 is to be setuid root or executed by root.
2910 nohalt [IA-64] Tells the kernel not to use the power saving
2911 function PAL_HALT_LIGHT when idle. This increases
2912 power-consumption. On the positive side, it reduces
2913 interrupt wake-up latency, which may improve performance
2914 in certain environments such as networked servers or
2917 nohibernate [HIBERNATION] Disable hibernation and resume.
2919 nohz= [KNL] Boottime enable/disable dynamic ticks
2920 Valid arguments: on, off
2923 nohz_full= [KNL,BOOT]
2924 The argument is a cpu list, as described above.
2925 In kernels built with CONFIG_NO_HZ_FULL=y, set
2926 the specified list of CPUs whose tick will be stopped
2927 whenever possible. The boot CPU will be forced outside
2928 the range to maintain the timekeeping.
2929 The CPUs in this range must also be included in the
2932 noiotrap [SH] Disables trapped I/O port accesses.
2934 noirqdebug [X86-32] Disables the code which attempts to detect and
2935 disable unhandled interrupt sources.
2937 no_timer_check [X86,APIC] Disables the code which tests for
2938 broken timer IRQ sources.
2940 noisapnp [ISAPNP] Disables ISA PnP code.
2942 noinitrd [RAM] Tells the kernel not to load any configured
2945 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2947 [Deprecated - use intremap=off]
2951 noinvpcid [X86] Disable the INVPCID cpu feature.
2953 nojitter [IA-64] Disables jitter checking for ITC timers.
2955 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2957 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2960 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2961 steal time is computed, but won't influence scheduler
2964 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2966 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2968 noltlbs [PPC] Do not use large page/tlb entries for kernel
2969 lowmem mapping on PPC40x and PPC8xx
2971 nomca [IA-64] Disable machine check abort handling
2973 nomce [X86-32] Disable Machine Check Exception
2975 nomfgpt [X86-32] Disable Multi-Function General Purpose
2976 Timer usage (for AMD Geode machines).
2978 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2979 shutdown the other cpus. Instead use the REBOOT_VECTOR
2982 nomodule Disable module load
2984 nopat [X86] Disable PAT (page attribute table extension of
2985 pagetables) support.
2987 nopcid [X86-64] Disable the PCID cpu feature.
2989 norandmaps Don't use address space randomization. Equivalent to
2990 echo 0 > /proc/sys/kernel/randomize_va_space
2992 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2993 with UP alternatives
2995 nordrand [X86] Disable kernel use of the RDRAND and
2996 RDSEED instructions even if they are supported
2997 by the processor. RDRAND and RDSEED are still
2998 available to user space applications.
3000 noresume [SWSUSP] Disables resume and restores original swap
3003 no-scroll [VGA] Disables scrollback.
3004 This is required for the Braillex ib80-piezo Braille
3005 reader made by F.H. Papenmeier (Germany).
3009 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3011 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3012 and disable the IO APIC. legacy for "maxcpus=0".
3014 nosoftlockup [KNL] Disable the soft-lockup detector.
3016 nosync [HW,M68K] Disables sync negotiation for all devices.
3018 notsc [BUGS=X86-32] Disable Time Stamp Counter
3020 nowatchdog [KNL] Disable both lockup detectors, i.e.
3021 soft-lockup and NMI watchdog (hard-lockup).
3025 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3027 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3028 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3029 Some features depend on CPU0. Known dependencies are:
3030 1. Resume from suspend/hibernate depends on CPU0.
3031 Suspend/hibernate will fail if CPU0 is offline and you
3032 need to online CPU0 before suspend/hibernate.
3033 2. PIC interrupts also depend on CPU0. CPU0 can't be
3034 removed if a PIC interrupt is detected.
3035 It's said poweroff/reboot may depend on CPU0 on some
3036 machines although I haven't seen such issues so far
3037 after CPU0 is offline on a few tested machines.
3038 If the dependencies are under your control, you can
3039 turn on cpu0_hotplug.
3041 nptcg= [IA-64] Override max number of concurrent global TLB
3042 purges which is reported from either PAL_VM_SUMMARY or
3045 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3046 could support. nr_cpus=n : n >= 1 limits the kernel to
3047 support 'n' processors. It could be larger than the
3048 number of already plugged CPU during bootup, later in
3049 runtime you can physically add extra cpu until it reaches
3050 n. So during boot up some boot time memory for per-cpu
3051 variables need be pre-allocated for later physical cpu
3054 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3056 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3057 Allowed values are enable and disable
3059 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3060 one of ['zone', 'node', 'default'] can be specified
3061 This can be set from sysctl after boot.
3062 See Documentation/sysctl/vm.txt for details.
3064 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3065 See Documentation/debugging-via-ohci1394.txt for more
3068 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3069 Rather than timing out after 20 ms if an EC
3070 command is not properly ACKed, override the length
3071 of the timeout. We have interrupts disabled while
3072 waiting for the ACK, so if this is set too high
3073 interrupts *may* be lost!
3075 omap_mux= [OMAP] Override bootloader pin multiplexing.
3076 Format: <mux_mode0.mode_name=value>...
3077 For example, to override I2C bus2:
3078 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3080 oprofile.timer= [HW]
3081 Use timer interrupt instead of performance counters
3083 oprofile.cpu_type= Force an oprofile cpu type
3084 This might be useful if you have an older oprofile
3085 userland or if you want common events.
3086 Format: { arch_perfmon }
3087 arch_perfmon: [X86] Force use of architectural
3088 perfmon on Intel CPUs instead of the
3089 CPU specific event set.
3090 timer: [X86] Force use of architectural NMI
3091 timer mode (see also oprofile.timer
3092 for generic hr timer mode)
3094 oops=panic Always panic on oopses. Default is to just kill the
3095 process, but there is a small probability of
3096 deadlocking the machine.
3097 This will also cause panics on machine check exceptions.
3098 Useful together with panic=30 to trigger a reboot.
3101 See Documentation/sound/oss/oss-parameters.txt
3103 page_owner= [KNL] Boot-time page_owner enabling option.
3104 Storage of the information about who allocated
3105 each page is disabled in default. With this switch,
3107 on: enable the feature
3109 page_poison= [KNL] Boot-time parameter changing the state of
3110 poisoning on the buddy allocator.
3111 off: turn off poisoning
3112 on: turn on poisoning
3114 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3115 timeout > 0: seconds before rebooting
3116 timeout = 0: wait forever
3117 timeout < 0: reboot immediately
3120 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3123 crash_kexec_post_notifiers
3124 Run kdump after running panic-notifiers and dumping
3125 kmsg. This only for the users who doubt kdump always
3126 succeeds in any situation.
3127 Note that this also increases risks of kdump failure,
3128 because some panic notifiers can make the crashed
3129 kernel more unstable.
3131 parkbd.port= [HW] Parallel port number the keyboard adapter is
3132 connected to, default is 0.
3134 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3135 0 for XT, 1 for AT (default is AT).
3138 parport= [HW,PPT] Specify parallel ports. 0 disables.
3139 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3140 Use 'auto' to force the driver to use any
3141 IRQ/DMA settings detected (the default is to
3142 ignore detected IRQ/DMA settings because of
3143 possible conflicts). You can specify the base
3144 address, IRQ, and DMA settings; IRQ and DMA
3145 should be numbers, or 'auto' (for using detected
3146 settings on that particular port), or 'nofifo'
3147 (to avoid using a FIFO even if it is detected).
3148 Parallel ports are assigned in the order they
3149 are specified on the command line, starting
3152 parport_init_mode= [HW,PPT]
3153 Configure VIA parallel port to operate in
3154 a specific mode. This is necessary on Pegasos
3155 computer where firmware has no options for setting
3156 up parallel port mode and sets it to spp.
3157 Currently this function knows 686a and 8231 chips.
3158 Format: [spp|ps2|epp|ecp|ecpepp]
3161 Halt all CPUs after the first oops has been printed for
3162 the specified number of seconds. This is to be used if
3163 your oopses keep scrolling off the screen.
3168 See header of drivers/block/paride/pcd.c.
3169 See also Documentation/blockdev/paride.txt.
3171 pci=option[,option...] [PCI] various PCI subsystem options:
3172 earlydump [X86] dump PCI config space before the kernel
3174 off [X86] don't probe for the PCI bus
3175 bios [X86-32] force use of PCI BIOS, don't access
3176 the hardware directly. Use this if your machine
3177 has a non-standard PCI host bridge.
3178 nobios [X86-32] disallow use of PCI BIOS, only direct
3179 hardware access methods are allowed. Use this
3180 if you experience crashes upon bootup and you
3181 suspect they are caused by the BIOS.
3182 conf1 [X86] Force use of PCI Configuration Access
3183 Mechanism 1 (config address in IO port 0xCF8,
3184 data in IO port 0xCFC, both 32-bit).
3185 conf2 [X86] Force use of PCI Configuration Access
3186 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3187 the function, IO port 0xCFA, also 8-bit, sets
3188 bus number. The config space is then accessed
3189 through ports 0xC000-0xCFFF).
3190 See http://wiki.osdev.org/PCI for more info
3191 on the configuration access mechanisms.
3192 noaer [PCIE] If the PCIEAER kernel config parameter is
3193 enabled, this kernel boot option can be used to
3194 disable the use of PCIE advanced error reporting.
3195 nodomains [PCI] Disable support for multiple PCI
3196 root domains (aka PCI segments, in ACPI-speak).
3197 nommconf [X86] Disable use of MMCONFIG for PCI
3199 check_enable_amd_mmconf [X86] check for and enable
3200 properly configured MMIO access to PCI
3201 config space on AMD family 10h CPU
3202 nomsi [MSI] If the PCI_MSI kernel config parameter is
3203 enabled, this kernel boot option can be used to
3204 disable the use of MSI interrupts system-wide.
3205 noioapicquirk [APIC] Disable all boot interrupt quirks.
3206 Safety option to keep boot IRQs enabled. This
3207 should never be necessary.
3208 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3209 primary IO-APIC for bridges that cannot disable
3210 boot IRQs. This fixes a source of spurious IRQs
3211 when the system masks IRQs.
3212 noioapicreroute [APIC] Disable workaround that uses the
3213 boot IRQ equivalent of an IRQ that connects to
3214 a chipset where boot IRQs cannot be disabled.
3215 The opposite of ioapicreroute.
3216 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3217 routing table. These calls are known to be buggy
3218 on several machines and they hang the machine
3219 when used, but on other computers it's the only
3220 way to get the interrupt routing table. Try
3221 this option if the kernel is unable to allocate
3222 IRQs or discover secondary PCI buses on your
3224 rom [X86] Assign address space to expansion ROMs.
3225 Use with caution as certain devices share
3226 address decoders between ROMs and other
3228 norom [X86] Do not assign address space to
3229 expansion ROMs that do not already have
3230 BIOS assigned address ranges.
3231 nobar [X86] Do not assign address space to the
3232 BARs that weren't assigned by the BIOS.
3233 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3234 assigned automatically to PCI devices. You can
3235 make the kernel exclude IRQs of your ISA cards
3237 pirqaddr=0xAAAAA [X86] Specify the physical address
3238 of the PIRQ table (normally generated
3239 by the BIOS) if it is outside the
3240 F0000h-100000h range.
3241 lastbus=N [X86] Scan all buses thru bus #N. Can be
3242 useful if the kernel is unable to find your
3243 secondary buses and you want to tell it
3244 explicitly which ones they are.
3245 assign-busses [X86] Always assign all PCI bus
3246 numbers ourselves, overriding
3247 whatever the firmware may have done.
3248 usepirqmask [X86] Honor the possible IRQ mask stored
3249 in the BIOS $PIR table. This is needed on
3250 some systems with broken BIOSes, notably
3251 some HP Pavilion N5400 and Omnibook XE3
3252 notebooks. This will have no effect if ACPI
3253 IRQ routing is enabled.
3254 noacpi [X86] Do not use ACPI for IRQ routing
3255 or for PCI scanning.
3256 use_crs [X86] Use PCI host bridge window information
3257 from ACPI. On BIOSes from 2008 or later, this
3258 is enabled by default. If you need to use this,
3259 please report a bug.
3260 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3261 If you need to use this, please report a bug.
3262 routeirq Do IRQ routing for all PCI devices.
3263 This is normally done in pci_enable_device(),
3264 so this option is a temporary workaround
3265 for broken drivers that don't call it.
3266 skip_isa_align [X86] do not align io start addr, so can
3267 handle more pci cards
3268 noearly [X86] Don't do any early type 1 scanning.
3269 This might help on some broken boards which
3270 machine check when some devices' config space
3271 is read. But various workarounds are disabled
3272 and some IOMMU drivers will not work.
3273 bfsort Sort PCI devices into breadth-first order.
3274 This sorting is done to get a device
3275 order compatible with older (<= 2.4) kernels.
3276 nobfsort Don't sort PCI devices into breadth-first order.
3277 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3278 tuning and use the BIOS-configured MPS defaults.
3279 pcie_bus_safe Set every device's MPS to the largest value
3280 supported by all devices below the root complex.
3281 pcie_bus_perf Set device MPS to the largest allowable MPS
3282 based on its parent bus. Also set MRRS (Max
3283 Read Request Size) to the largest supported
3284 value (no larger than the MPS that the device
3285 or bus can support) for best performance.
3286 pcie_bus_peer2peer Set every device's MPS to 128B, which
3287 every device is guaranteed to support. This
3288 configuration allows peer-to-peer DMA between
3289 any pair of devices, possibly at the cost of
3290 reduced performance. This also guarantees
3291 that hot-added devices will work.
3292 cbiosize=nn[KMG] The fixed amount of bus space which is
3293 reserved for the CardBus bridge's IO window.
3294 The default value is 256 bytes.
3295 cbmemsize=nn[KMG] The fixed amount of bus space which is
3296 reserved for the CardBus bridge's memory
3297 window. The default value is 64 megabytes.
3300 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
3301 [<order of align>@]pci:<vendor>:<device>\
3302 [:<subvendor>:<subdevice>][; ...]
3303 Specifies alignment and device to reassign
3304 aligned memory resources.
3305 If <order of align> is not specified,
3306 PAGE_SIZE is used as alignment.
3307 PCI-PCI bridge can be specified, if resource
3308 windows need to be expanded.
3309 To specify the alignment for several
3310 instances of a device, the PCI vendor,
3311 device, subvendor, and subdevice may be
3312 specified, e.g., 4096@pci:8086:9c22:103c:198f
3313 ecrc= Enable/disable PCIe ECRC (transaction layer
3314 end-to-end CRC checking).
3315 bios: Use BIOS/firmware settings. This is the
3319 hpiosize=nn[KMG] The fixed amount of bus space which is
3320 reserved for hotplug bridge's IO window.
3321 Default size is 256 bytes.
3322 hpmemsize=nn[KMG] The fixed amount of bus space which is
3323 reserved for hotplug bridge's memory window.
3324 Default size is 2 megabytes.
3325 hpbussize=nn The minimum amount of additional bus numbers
3326 reserved for buses below a hotplug bridge.
3328 realloc= Enable/disable reallocating PCI bridge resources
3329 if allocations done by BIOS are too small to
3330 accommodate resources required by all child
3332 off: Turn realloc off
3334 realloc same as realloc=on
3335 noari do not use PCIe ARI.
3336 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3337 only look for one device below a PCIe downstream
3340 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3343 force Enable ASPM even on devices that claim not to support it.
3344 WARNING: Forcing ASPM on may cause system lockups.
3346 pcie_hp= [PCIE] PCI Express Hotplug driver options:
3347 nomsi Do not use MSI for PCI Express Native Hotplug (this
3348 makes all PCIe ports use INTx for hotplug services).
3350 pcie_ports= [PCIE] PCIe ports handling:
3351 auto Ask the BIOS whether or not to use native PCIe services
3352 associated with PCIe ports (PME, hot-plug, AER). Use
3353 them only if that is allowed by the BIOS.
3354 native Use native PCIe services associated with PCIe ports
3356 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
3359 pcie_port_pm= [PCIE] PCIe port power management handling:
3360 off Disable power management of all PCIe ports
3361 force Forcibly enable power management of all PCIe ports
3363 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3364 nomsi Do not use MSI for native PCIe PME signaling (this makes
3365 all PCIe root ports use INTx for all services).
3367 pcmv= [HW,PCMCIA] BadgePAD 4
3371 Keep all power-domains already enabled by bootloader on,
3372 even if no driver has claimed them. This is useful
3373 for debug and development, but should not be
3374 needed on a platform with proper driver support.
3377 See Documentation/blockdev/paride.txt.
3379 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3382 See arch/parisc/kernel/pdc_chassis.c
3384 percpu_alloc= Select which percpu first chunk allocator to use.
3385 Currently supported values are "embed" and "page".
3386 Archs may support subset or none of the selections.
3387 See comments in mm/percpu.c for details on each
3388 allocator. This parameter is primarily for debugging
3389 and performance comparison.
3392 See Documentation/blockdev/paride.txt.
3395 See Documentation/blockdev/paride.txt.
3397 pirq= [SMP,APIC] Manual mp-table setup
3398 See Documentation/x86/i386/IO-APIC.txt.
3400 plip= [PPT,NET] Parallel port network link
3401 Format: { parport<nr> | timid | 0 }
3402 See also Documentation/parport.txt.
3404 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3405 Override pmtimer IOPort with a hex value.
3409 Enable PNP debug messages (depends on the
3410 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3411 via /sys/module/pnp/parameters/debug. We always show
3412 current resource usage; turning this on also shows
3413 possible settings and some assignment information.
3419 { on | off | curr | res | no-curr | no-res }
3422 [ISAPNP] Exclude IRQs for the autoconfiguration
3425 [ISAPNP] Exclude DMAs for the autoconfiguration
3427 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3428 Ranges are in pairs (I/O port base and size).
3431 [ISAPNP] Exclude memory regions for the
3433 Ranges are in pairs (memory base and size).
3435 ports= [IP_VS_FTP] IPVS ftp helper module
3437 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3439 Format: <port>,<port>....
3441 ppc_strict_facility_enable
3442 [PPC] This option catches any kernel floating point,
3443 Altivec, VSX and SPE outside of regions specifically
3444 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3445 There is some performance impact when enabling this.
3447 print-fatal-signals=
3448 [KNL] debug: print fatal signals
3450 If enabled, warn about various signal handling
3451 related application anomalies: too many signals,
3452 too many POSIX.1 timers, fatal signals causing a
3455 If you hit the warning due to signal overflow,
3456 you might want to try "ulimit -i unlimited".
3460 printk.always_kmsg_dump=
3461 Trigger kmsg_dump for cases other than kernel oops or
3463 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3466 printk.devkmsg={on,off,ratelimit}
3467 Control writing to /dev/kmsg.
3468 on - unlimited logging to /dev/kmsg from userspace
3469 off - logging to /dev/kmsg disabled
3470 ratelimit - ratelimit the logging
3473 printk.time= Show timing data prefixed to each printk message line
3474 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3476 processor.max_cstate= [HW,ACPI]
3477 Limit processor to maximum C-state
3478 max_cstate=9 overrides any DMI blacklist limit.
3480 processor.nocst [HW,ACPI]
3481 Ignore the _CST method to determine C-states,
3482 instead using the legacy FADT method
3484 profile= [KNL] Enable kernel profiling via /proc/profile
3485 Format: [schedule,]<number>
3486 Param: "schedule" - profile schedule points.
3487 Param: <number> - step/bucket size as a power of 2 for
3488 statistical time based profiling.
3489 Param: "sleep" - profile D-state sleeping (millisecs).
3490 Requires CONFIG_SCHEDSTATS
3491 Param: "kvm" - profile VM exits.
3493 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3495 See Documentation/blockdev/ramdisk.txt.
3497 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3498 probe for; one of (bare|imps|exps|lifebook|any).
3499 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3501 psmouse.resetafter= [HW,MOUSE]
3502 Try to reset the device after so many bad packets
3505 [HW,MOUSE] Set desired mouse resolution, in dpi.
3506 psmouse.smartscroll=
3507 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3508 0 = disabled, 1 = enabled (default).
3510 pstore.backend= Specify the name of the pstore backend to use
3513 See Documentation/blockdev/paride.txt.
3515 pti= [X86_64] Control Page Table Isolation of user and
3516 kernel address spaces. Disabling this feature
3517 removes hardening, but improves performance of
3518 system calls and interrupts.
3520 on - unconditionally enable
3521 off - unconditionally disable
3522 auto - kernel detects whether your CPU model is
3523 vulnerable to issues that PTI mitigates
3525 Not specifying this option is equivalent to pti=auto.
3528 Equivalent to pti=off
3531 [KNL] Number of legacy pty's. Overwrites compiled-in
3534 quiet [KNL] Disable most log messages
3539 See Documentation/md.txt.
3541 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3542 See Documentation/blockdev/ramdisk.txt.
3545 The argument is a cpu list, as described above.
3547 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3548 the specified list of CPUs to be no-callback CPUs.
3549 Invocation of these CPUs' RCU callbacks will
3550 be offloaded to "rcuox/N" kthreads created for
3551 that purpose, where "x" is "b" for RCU-bh, "p"
3552 for RCU-preempt, and "s" for RCU-sched, and "N"
3553 is the CPU number. This reduces OS jitter on the
3554 offloaded CPUs, which can be useful for HPC and
3555 real-time workloads. It can also improve energy
3556 efficiency for asymmetric multiprocessors.
3559 Rather than requiring that offloaded CPUs
3560 (specified by rcu_nocbs= above) explicitly
3561 awaken the corresponding "rcuoN" kthreads,
3562 make these kthreads poll for callbacks.
3563 This improves the real-time response for the
3564 offloaded CPUs by relieving them of the need to
3565 wake up the corresponding kthread, but degrades
3566 energy efficiency by requiring that the kthreads
3567 periodically wake up to do the polling.
3569 rcutree.blimit= [KNL]
3570 Set maximum number of finished RCU callbacks to
3571 process in one batch.
3573 rcutree.dump_tree= [KNL]
3574 Dump the structure of the rcu_node combining tree
3575 out at early boot. This is used for diagnostic
3576 purposes, to verify correct tree setup.
3578 rcutree.gp_cleanup_delay= [KNL]
3579 Set the number of jiffies to delay each step of
3580 RCU grace-period cleanup. This only has effect
3581 when CONFIG_RCU_TORTURE_TEST_SLOW_CLEANUP is set.
3583 rcutree.gp_init_delay= [KNL]
3584 Set the number of jiffies to delay each step of
3585 RCU grace-period initialization. This only has
3586 effect when CONFIG_RCU_TORTURE_TEST_SLOW_INIT
3589 rcutree.gp_preinit_delay= [KNL]
3590 Set the number of jiffies to delay each step of
3591 RCU grace-period pre-initialization, that is,
3592 the propagation of recent CPU-hotplug changes up
3593 the rcu_node combining tree. This only has effect
3594 when CONFIG_RCU_TORTURE_TEST_SLOW_PREINIT is set.
3596 rcutree.rcu_fanout_exact= [KNL]
3597 Disable autobalancing of the rcu_node combining
3598 tree. This is used by rcutorture, and might
3599 possibly be useful for architectures having high
3600 cache-to-cache transfer latencies.
3602 rcutree.rcu_fanout_leaf= [KNL]
3603 Change the number of CPUs assigned to each
3604 leaf rcu_node structure. Useful for very
3605 large systems, which will choose the value 64,
3606 and for NUMA systems with large remote-access
3607 latencies, which will choose a value aligned
3608 with the appropriate hardware boundaries.
3610 rcutree.jiffies_till_sched_qs= [KNL]
3611 Set required age in jiffies for a
3612 given grace period before RCU starts
3613 soliciting quiescent-state help from
3614 rcu_note_context_switch().
3616 rcutree.jiffies_till_first_fqs= [KNL]
3617 Set delay from grace-period initialization to
3618 first attempt to force quiescent states.
3619 Units are jiffies, minimum value is zero,
3620 and maximum value is HZ.
3622 rcutree.jiffies_till_next_fqs= [KNL]
3623 Set delay between subsequent attempts to force
3624 quiescent states. Units are jiffies, minimum
3625 value is one, and maximum value is HZ.
3627 rcutree.kthread_prio= [KNL,BOOT]
3628 Set the SCHED_FIFO priority of the RCU per-CPU
3629 kthreads (rcuc/N). This value is also used for
3630 the priority of the RCU boost threads (rcub/N)
3631 and for the RCU grace-period kthreads (rcu_bh,
3632 rcu_preempt, and rcu_sched). If RCU_BOOST is
3633 set, valid values are 1-99 and the default is 1
3634 (the least-favored priority). Otherwise, when
3635 RCU_BOOST is not set, valid values are 0-99 and
3636 the default is zero (non-realtime operation).
3638 rcutree.rcu_nocb_leader_stride= [KNL]
3639 Set the number of NOCB kthread groups, which
3640 defaults to the square root of the number of
3641 CPUs. Larger numbers reduces the wakeup overhead
3642 on the per-CPU grace-period kthreads, but increases
3643 that same overhead on each group's leader.
3645 rcutree.qhimark= [KNL]
3646 Set threshold of queued RCU callbacks beyond which
3647 batch limiting is disabled.
3649 rcutree.qlowmark= [KNL]
3650 Set threshold of queued RCU callbacks below which
3651 batch limiting is re-enabled.
3653 rcutree.rcu_idle_gp_delay= [KNL]
3654 Set wakeup interval for idle CPUs that have
3655 RCU callbacks (RCU_FAST_NO_HZ=y).
3657 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3658 Set wakeup interval for idle CPUs that have
3659 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3660 Lazy RCU callbacks are those which RCU can
3661 prove do nothing more than free memory.
3663 rcuperf.gp_exp= [KNL]
3664 Measure performance of expedited synchronous
3665 grace-period primitives.
3667 rcuperf.holdoff= [KNL]
3668 Set test-start holdoff period. The purpose of
3669 this parameter is to delay the start of the
3670 test until boot completes in order to avoid
3673 rcuperf.nreaders= [KNL]
3674 Set number of RCU readers. The value -1 selects
3675 N, where N is the number of CPUs. A value
3676 "n" less than -1 selects N-n+1, where N is again
3677 the number of CPUs. For example, -2 selects N
3678 (the number of CPUs), -3 selects N+1, and so on.
3679 A value of "n" less than or equal to -N selects
3682 rcuperf.nwriters= [KNL]
3683 Set number of RCU writers. The values operate
3684 the same as for rcuperf.nreaders.
3685 N, where N is the number of CPUs
3687 rcuperf.perf_runnable= [BOOT]
3688 Start rcuperf running at boot time.
3690 rcuperf.shutdown= [KNL]
3691 Shut the system down after performance tests
3692 complete. This is useful for hands-off automated
3695 rcuperf.perf_type= [KNL]
3696 Specify the RCU implementation to test.
3698 rcuperf.verbose= [KNL]
3699 Enable additional printk() statements.
3701 rcutorture.cbflood_inter_holdoff= [KNL]
3702 Set holdoff time (jiffies) between successive
3703 callback-flood tests.
3705 rcutorture.cbflood_intra_holdoff= [KNL]
3706 Set holdoff time (jiffies) between successive
3707 bursts of callbacks within a given callback-flood
3710 rcutorture.cbflood_n_burst= [KNL]
3711 Set the number of bursts making up a given
3712 callback-flood test. Set this to zero to
3713 disable callback-flood testing.
3715 rcutorture.cbflood_n_per_burst= [KNL]
3716 Set the number of callbacks to be registered
3717 in a given burst of a callback-flood test.
3719 rcutorture.fqs_duration= [KNL]
3720 Set duration of force_quiescent_state bursts
3723 rcutorture.fqs_holdoff= [KNL]
3724 Set holdoff time within force_quiescent_state bursts
3727 rcutorture.fqs_stutter= [KNL]
3728 Set wait time between force_quiescent_state bursts
3731 rcutorture.gp_cond= [KNL]
3732 Use conditional/asynchronous update-side
3733 primitives, if available.
3735 rcutorture.gp_exp= [KNL]
3736 Use expedited update-side primitives, if available.
3738 rcutorture.gp_normal= [KNL]
3739 Use normal (non-expedited) asynchronous
3740 update-side primitives, if available.
3742 rcutorture.gp_sync= [KNL]
3743 Use normal (non-expedited) synchronous
3744 update-side primitives, if available. If all
3745 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3746 rcutorture.gp_normal=, and rcutorture.gp_sync=
3747 are zero, rcutorture acts as if is interpreted
3748 they are all non-zero.
3750 rcutorture.n_barrier_cbs= [KNL]
3751 Set callbacks/threads for rcu_barrier() testing.
3753 rcutorture.nfakewriters= [KNL]
3754 Set number of concurrent RCU writers. These just
3755 stress RCU, they don't participate in the actual
3756 test, hence the "fake".
3758 rcutorture.nreaders= [KNL]
3759 Set number of RCU readers. The value -1 selects
3760 N-1, where N is the number of CPUs. A value
3761 "n" less than -1 selects N-n-2, where N is again
3762 the number of CPUs. For example, -2 selects N
3763 (the number of CPUs), -3 selects N+1, and so on.
3765 rcutorture.object_debug= [KNL]
3766 Enable debug-object double-call_rcu() testing.
3768 rcutorture.onoff_holdoff= [KNL]
3769 Set time (s) after boot for CPU-hotplug testing.
3771 rcutorture.onoff_interval= [KNL]
3772 Set time (s) between CPU-hotplug operations, or
3773 zero to disable CPU-hotplug testing.
3775 rcutorture.shuffle_interval= [KNL]
3776 Set task-shuffle interval (s). Shuffling tasks
3777 allows some CPUs to go into dyntick-idle mode
3778 during the rcutorture test.
3780 rcutorture.shutdown_secs= [KNL]
3781 Set time (s) after boot system shutdown. This
3782 is useful for hands-off automated testing.
3784 rcutorture.stall_cpu= [KNL]
3785 Duration of CPU stall (s) to test RCU CPU stall
3786 warnings, zero to disable.
3788 rcutorture.stall_cpu_holdoff= [KNL]
3789 Time to wait (s) after boot before inducing stall.
3791 rcutorture.stat_interval= [KNL]
3792 Time (s) between statistics printk()s.
3794 rcutorture.stutter= [KNL]
3795 Time (s) to stutter testing, for example, specifying
3796 five seconds causes the test to run for five seconds,
3797 wait for five seconds, and so on. This tests RCU's
3798 ability to transition abruptly to and from idle.
3800 rcutorture.test_boost= [KNL]
3801 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3802 "Maybe" means test if the RCU implementation
3803 under test support RCU priority boosting.
3805 rcutorture.test_boost_duration= [KNL]
3806 Duration (s) of each individual boost test.
3808 rcutorture.test_boost_interval= [KNL]
3809 Interval (s) between each boost test.
3811 rcutorture.test_no_idle_hz= [KNL]
3812 Test RCU's dyntick-idle handling. See also the
3813 rcutorture.shuffle_interval parameter.
3815 rcutorture.torture_runnable= [BOOT]
3816 Start rcutorture running at boot time.
3818 rcutorture.torture_type= [KNL]
3819 Specify the RCU implementation to test.
3821 rcutorture.verbose= [KNL]
3822 Enable additional printk() statements.
3824 rcupdate.rcu_cpu_stall_suppress= [KNL]
3825 Suppress RCU CPU stall warning messages.
3827 rcupdate.rcu_cpu_stall_timeout= [KNL]
3828 Set timeout for RCU CPU stall warning messages.
3830 rcupdate.rcu_expedited= [KNL]
3831 Use expedited grace-period primitives, for
3832 example, synchronize_rcu_expedited() instead
3833 of synchronize_rcu(). This reduces latency,
3834 but can increase CPU utilization, degrade
3835 real-time latency, and degrade energy efficiency.
3836 No effect on CONFIG_TINY_RCU kernels.
3838 rcupdate.rcu_normal= [KNL]
3839 Use only normal grace-period primitives,
3840 for example, synchronize_rcu() instead of
3841 synchronize_rcu_expedited(). This improves
3842 real-time latency, CPU utilization, and
3843 energy efficiency, but can expose users to
3844 increased grace-period latency. This parameter
3845 overrides rcupdate.rcu_expedited. No effect on
3846 CONFIG_TINY_RCU kernels.
3848 rcupdate.rcu_normal_after_boot= [KNL]
3849 Once boot has completed (that is, after
3850 rcu_end_inkernel_boot() has been invoked), use
3851 only normal grace-period primitives. No effect
3852 on CONFIG_TINY_RCU kernels.
3854 rcupdate.rcu_task_stall_timeout= [KNL]
3855 Set timeout in jiffies for RCU task stall warning
3856 messages. Disable with a value less than or equal
3859 rcupdate.rcu_self_test= [KNL]
3860 Run the RCU early boot self tests
3862 rcupdate.rcu_self_test_bh= [KNL]
3863 Run the RCU bh early boot self tests
3865 rcupdate.rcu_self_test_sched= [KNL]
3866 Run the RCU sched early boot self tests
3870 Run specified binary instead of /init from the ramdisk,
3871 used for early userspace startup. See initrd.
3874 force - Override the decision by the kernel to hide the
3875 advertisement of RDRAND support (this affects
3876 certain AMD processors because of buggy BIOS
3877 support, specifically around the suspend/resume
3881 Format (x86 or x86_64):
3882 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3884 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3886 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3887 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3888 reboot_force is either force or not specified,
3889 reboot_cpu is s[mp]#### with #### being the processor
3890 to be used for rebooting.
3893 [KNL, SMP] Set scheduler's default relax_domain_level.
3894 See Documentation/cgroup-v1/cpusets.txt.
3896 relative_sleep_states=
3897 [SUSPEND] Use sleep state labeling where the deepest
3898 state available other than hibernation is always "mem".
3899 Format: { "0" | "1" }
3900 0 -- Traditional sleep state labels.
3901 1 -- Relative sleep state labels.
3903 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
3905 reservetop= [X86-32]
3907 Reserves a hole at the top of the kernel virtual
3912 Set the amount of memory to reserve for BIOS at
3913 the bottom of the address space.
3915 reset_devices [KNL] Force drivers to reset the underlying device
3916 during initialization.
3919 Specify the partition device for software suspend
3921 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3923 resume_offset= [SWSUSP]
3924 Specify the offset from the beginning of the partition
3925 given by "resume=" at which the swap header is located,
3926 in <PAGE_SIZE> units (needed only for swap files).
3927 See Documentation/power/swsusp-and-swap-files.txt
3929 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
3930 read the resume files
3932 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
3933 Useful for devices that are detected asynchronously
3934 (e.g. USB and MMC devices).
3936 hibernate= [HIBERNATION]
3937 noresume Don't check if there's a hibernation image
3938 present during boot.
3939 nocompress Don't compress/decompress hibernation images.
3940 no Disable hibernation and resume.
3941 protect_image Turn on image protection during restoration
3942 (that will set all pages holding image data
3943 during restoration read-only).
3945 retain_initrd [RAM] Keep initrd memory after extraction
3947 rfkill.default_state=
3948 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
3949 etc. communication is blocked by default.
3952 rfkill.master_switch_mode=
3953 0 The "airplane mode" button does nothing.
3954 1 The "airplane mode" button toggles between everything
3955 blocked and the previous configuration.
3956 2 The "airplane mode" button toggles between everything
3957 blocked and everything unblocked.
3959 rhash_entries= [KNL,NET]
3960 Set number of hash buckets for route cache
3962 ro [KNL] Mount root device read-only on boot
3965 on Mark read-only kernel memory as read-only (default).
3966 off Leave read-only kernel memory writable for debugging.
3969 Enable the uart passthrough on the designated usb port
3970 on Rockchip SoCs. When active, the signals of the
3971 debug-uart get routed to the D+ and D- pins of the usb
3972 port and the regular usb controller gets disabled.
3974 root= [KNL] Root filesystem
3975 See name_to_dev_t comment in init/do_mounts.c.
3977 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
3978 mount the root filesystem
3980 rootflags= [KNL] Set root filesystem mount option string
3982 rootfstype= [KNL] Set root filesystem type
3984 rootwait [KNL] Wait (indefinitely) for root device to show up.
3985 Useful for devices that are detected asynchronously
3986 (e.g. USB and MMC devices).
3988 rproc_mem=nn[KMG][@address]
3989 [KNL,ARM,CMA] Remoteproc physical memory block.
3990 Memory area to be used by remote processor image,
3993 rw [KNL] Mount root device read-write on boot
3995 S [KNL] Run init in single mode
3997 s390_iommu= [HW,S390]
3998 Set s390 IOTLB flushing mode
4000 With strict flushing every unmap operation will result in
4001 an IOTLB flush. Default is lazy flushing before reuse,
4005 See drivers/net/irda/sa1100_ir.c.
4007 sbni= [NET] Granch SBNI12 leased line adapter
4009 sched_debug [KNL] Enables verbose scheduler debug messages.
4011 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4012 Allowed values are enable and disable. This feature
4013 incurs a small amount of overhead in the scheduler
4014 but is useful for debugging and performance tuning.
4016 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4017 xtime_lock contention on larger systems, and/or RCU lock
4018 contention on all systems with CONFIG_MAXSMP set.
4019 Format: { "0" | "1" }
4020 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4022 Note: increases power consumption, thus should only be
4023 enabled if running jitter sensitive (HPC/RT) workloads.
4025 security= [SECURITY] Choose a security module to enable at boot.
4026 If this boot parameter is not specified, only the first
4027 security module asking for security registration will be
4028 loaded. An invalid security module name will be treated
4029 as if no module has been chosen.
4031 selinux= [SELINUX] Disable or enable SELinux at boot time.
4032 Format: { "0" | "1" }
4033 See security/selinux/Kconfig help text.
4036 Default value is set via kernel config option.
4037 If enabled at boot time, /selinux/disable can be used
4038 later to disable prior to initial policy load.
4040 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4041 Format: { "0" | "1" }
4042 See security/apparmor/Kconfig help text
4045 Default value is set via kernel config option.
4047 serialnumber [BUGS=X86-32]
4050 Maximal number of shapers.
4052 show_msr= [x86] show boot-time MSR settings
4053 Format: { <integer> }
4054 Show boot-time (BIOS-initialized) MSR settings.
4055 The parameter means the number of CPUs to show,
4056 for example 1 means boot CPU only.
4064 Disable merging of slabs with similar size. May be
4065 necessary if there is some reason to distinguish
4066 allocs to different slabs. Debug options disable
4067 merging on their own.
4068 For more information see Documentation/vm/slub.txt.
4070 slab_max_order= [MM, SLAB]
4071 Determines the maximum allowed order for slabs.
4072 A high setting may cause OOMs due to memory
4073 fragmentation. Defaults to 1 for systems with
4074 more than 32MB of RAM, 0 otherwise.
4076 slub_debug[=options[,slabs]] [MM, SLUB]
4077 Enabling slub_debug allows one to determine the
4078 culprit if slab objects become corrupted. Enabling
4079 slub_debug can create guard zones around objects and
4080 may poison objects when not in use. Also tracks the
4081 last alloc / free. For more information see
4082 Documentation/vm/slub.txt.
4084 slub_max_order= [MM, SLUB]
4085 Determines the maximum allowed order for slabs.
4086 A high setting may cause OOMs due to memory
4087 fragmentation. For more information see
4088 Documentation/vm/slub.txt.
4090 slub_min_objects= [MM, SLUB]
4091 The minimum number of objects per slab. SLUB will
4092 increase the slab order up to slub_max_order to
4093 generate a sufficiently large slab able to contain
4094 the number of objects indicated. The higher the number
4095 of objects the smaller the overhead of tracking slabs
4096 and the less frequently locks need to be acquired.
4097 For more information see Documentation/vm/slub.txt.
4099 slub_min_order= [MM, SLUB]
4100 Determines the minimum page order for slabs. Must be
4101 lower than slub_max_order.
4102 For more information see Documentation/vm/slub.txt.
4104 slub_nomerge [MM, SLUB]
4105 Same with slab_nomerge. This is supported for legacy.
4106 See slab_nomerge for more information.
4109 Format: <io1>[,<io2>[,...,<io8>]]
4111 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4112 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4113 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4114 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4115 smsc-ircc2.ircc_irq= [HW] IRQ line
4116 smsc-ircc2.ircc_dma= [HW] DMA channel
4117 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4118 0: Toshiba Satellite 1800 (GP data pin select)
4119 1: Fast pin select (default)
4122 smt [KNL,S390] Set the maximum number of threads (logical
4123 CPUs) to use per physical CPU on systems capable of
4124 symmetric multithreading (SMT). Will be capped to the
4125 actual hardware limit.
4127 Default: -1 (no limit)
4130 [KNL] Should the soft-lockup detector generate panics.
4133 softlockup_all_cpu_backtrace=
4134 [KNL] Should the soft-lockup detector generate
4135 backtraces on all cpus.
4138 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4139 See Documentation/laptops/sonypi.txt
4141 spectre_v2= [X86] Control mitigation of Spectre variant 2
4142 (indirect branch speculation) vulnerability.
4143 The default operation protects the kernel from
4146 on - unconditionally enable, implies
4148 off - unconditionally disable, implies
4150 auto - kernel detects whether your CPU model is
4153 Selecting 'on' will, and 'auto' may, choose a
4154 mitigation method at run time according to the
4155 CPU, the available microcode, the setting of the
4156 CONFIG_RETPOLINE configuration option, and the
4157 compiler with which the kernel was built.
4159 Selecting 'on' will also enable the mitigation
4160 against user space to user space task attacks.
4162 Selecting 'off' will disable both the kernel and
4163 the user space protections.
4165 Specific mitigations can also be selected manually:
4167 retpoline - replace indirect branches
4168 retpoline,generic - Retpolines
4169 retpoline,lfence - LFENCE; indirect branch
4170 retpoline,amd - alias for retpoline,lfence
4171 eibrs - enhanced IBRS
4172 eibrs,retpoline - enhanced IBRS + Retpolines
4173 eibrs,lfence - enhanced IBRS + LFENCE
4175 Not specifying this option is equivalent to
4179 [X86] Control mitigation of Spectre variant 2
4180 (indirect branch speculation) vulnerability between
4183 on - Unconditionally enable mitigations. Is
4184 enforced by spectre_v2=on
4186 off - Unconditionally disable mitigations. Is
4187 enforced by spectre_v2=off
4189 prctl - Indirect branch speculation is enabled,
4190 but mitigation can be enabled via prctl
4191 per thread. The mitigation control state
4192 is inherited on fork.
4195 - Like "prctl" above, but only STIBP is
4196 controlled per thread. IBPB is issued
4197 always when switching between different user
4201 - Same as "prctl" above, but all seccomp
4202 threads will enable the mitigation unless
4203 they explicitly opt out.
4206 - Like "seccomp" above, but only STIBP is
4207 controlled per thread. IBPB is issued
4208 always when switching between different
4209 user space processes.
4211 auto - Kernel selects the mitigation depending on
4212 the available CPU features and vulnerability.
4215 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4217 Not specifying this option is equivalent to
4218 spectre_v2_user=auto.
4220 spec_store_bypass_disable=
4221 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4222 (Speculative Store Bypass vulnerability)
4224 Certain CPUs are vulnerable to an exploit against a
4225 a common industry wide performance optimization known
4226 as "Speculative Store Bypass" in which recent stores
4227 to the same memory location may not be observed by
4228 later loads during speculative execution. The idea
4229 is that such stores are unlikely and that they can
4230 be detected prior to instruction retirement at the
4231 end of a particular speculation execution window.
4233 In vulnerable processors, the speculatively forwarded
4234 store can be used in a cache side channel attack, for
4235 example to read memory to which the attacker does not
4236 directly have access (e.g. inside sandboxed code).
4238 This parameter controls whether the Speculative Store
4239 Bypass optimization is used.
4241 on - Unconditionally disable Speculative Store Bypass
4242 off - Unconditionally enable Speculative Store Bypass
4243 auto - Kernel detects whether the CPU model contains an
4244 implementation of Speculative Store Bypass and
4245 picks the most appropriate mitigation. If the
4246 CPU is not vulnerable, "off" is selected. If the
4247 CPU is vulnerable the default mitigation is
4248 architecture and Kconfig dependent. See below.
4249 prctl - Control Speculative Store Bypass per thread
4250 via prctl. Speculative Store Bypass is enabled
4251 for a process by default. The state of the control
4252 is inherited on fork.
4253 seccomp - Same as "prctl" above, but all seccomp threads
4254 will disable SSB unless they explicitly opt out.
4256 Not specifying this option is equivalent to
4257 spec_store_bypass_disable=auto.
4259 Default mitigations:
4260 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4262 spia_io_base= [HW,MTD]
4268 Control the Special Register Buffer Data Sampling
4271 Certain CPUs are vulnerable to an MDS-like
4272 exploit which can leak bits from the random
4275 By default, this issue is mitigated by
4276 microcode. However, the microcode fix can cause
4277 the RDRAND and RDSEED instructions to become
4278 much slower. Among other effects, this will
4279 result in reduced throughput from /dev/urandom.
4281 The microcode mitigation can be disabled with
4282 the following option:
4284 off: Disable mitigation and remove
4285 performance impact to RDRAND and RDSEED
4288 Speculative Store Bypass Disable control
4290 On CPUs that are vulnerable to the Speculative
4291 Store Bypass vulnerability and offer a
4292 firmware based mitigation, this parameter
4293 indicates how the mitigation should be used:
4295 force-on: Unconditionally enable mitigation for
4296 for both kernel and userspace
4297 force-off: Unconditionally disable mitigation for
4298 for both kernel and userspace
4299 kernel: Always enable mitigation in the
4300 kernel, and offer a prctl interface
4301 to allow userspace to register its
4302 interest in being mitigated too.
4304 stack_guard_gap= [MM]
4305 override the default stack gap protection. The value
4306 is in page units and it defines how many pages prior
4307 to (for stacks growing down) resp. after (for stacks
4308 growing up) the main stack are reserved for no other
4309 mapping. Default value is 256 pages.
4312 Enabled the stack tracer on boot up.
4314 stacktrace_filter=[function-list]
4315 [FTRACE] Limit the functions that the stack tracer
4316 will trace at boot up. function-list is a comma separated
4317 list of functions. This list can be changed at run
4318 time by the stack_trace_filter file in the debugfs
4319 tracing directory. Note, this enables stack tracing
4320 and the stacktrace above is not needed.
4324 Set the STI (builtin display/keyboard on the HP-PARISC
4325 machines) console (graphic card) which should be used
4326 as the initial boot-console.
4327 See also comment in drivers/video/console/sticore.c.
4330 See comment in drivers/video/console/sticore.c.
4333 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4335 sunrpc.min_resvport=
4336 sunrpc.max_resvport=
4338 SunRPC servers often require that client requests
4339 originate from a privileged port (i.e. a port in the
4340 range 0 < portnr < 1024).
4341 An administrator who wishes to reserve some of these
4342 ports for other uses may adjust the range that the
4343 kernel's sunrpc client considers to be privileged
4344 using these two parameters to set the minimum and
4345 maximum port values.
4347 sunrpc.svc_rpc_per_connection_limit=
4349 Limit the number of requests that the server will
4350 process in parallel from a single connection.
4351 The default value is 0 (no limit).
4355 Control how the NFS server code allocates CPUs to
4356 service thread pools. Depending on how many NICs
4357 you have and where their interrupts are bound, this
4358 option will affect which CPUs will do NFS serving.
4359 Note: this parameter cannot be changed while the
4360 NFS server is running.
4362 auto the server chooses an appropriate mode
4363 automatically using heuristics
4364 global a single global pool contains all CPUs
4365 percpu one pool for each CPU
4366 pernode one pool for each NUMA node (equivalent
4367 to global on non-NUMA machines)
4369 sunrpc.tcp_slot_table_entries=
4370 sunrpc.udp_slot_table_entries=
4372 Sets the upper limit on the number of simultaneous
4373 RPC calls that can be sent from the client to a
4374 server. Increasing these values may allow you to
4375 improve throughput, but will also increase the
4376 amount of memory reserved for use by the client.
4378 suspend.pm_test_delay=
4380 Sets the number of seconds to remain in a suspend test
4381 mode before resuming the system (see
4382 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4383 is set. Default value is 5.
4386 [KNL] Enable accounting of swap in memory resource
4387 controller if no parameter or 1 is given or disable
4388 it if 0 is given (See Documentation/cgroup-v1/memory.txt)
4390 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4391 Format: { <int> | force | noforce }
4392 <int> -- Number of I/O TLB slabs
4393 force -- force using of bounce buffers even if they
4394 wouldn't be automatically used by the kernel
4395 noforce -- Never use bounce buffers (for debugging)
4399 sysfs.deprecated=0|1 [KNL]
4400 Enable/disable old style sysfs layout for old udev
4401 on older distributions. When this option is enabled
4402 very new udev will not work anymore. When this option
4403 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4404 in older udev will not work anymore.
4405 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4406 the kernel configuration.
4408 sysrq_always_enabled
4410 Ignore sysrq setting - this boot parameter will
4411 neutralize any effect of /proc/sys/kernel/sysrq.
4412 Useful for debugging.
4414 tcpmhash_entries= [KNL,NET]
4415 Set the number of tcp_metrics_hash slots.
4416 Default value is 8192 or 16384 depending on total
4417 ram pages. This is used to specify the TCP metrics
4418 cache size. See Documentation/networking/ip-sysctl.txt
4419 "tcp_no_metrics_save" section for more details.
4423 test_suspend= [SUSPEND][,N]
4424 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4425 standby suspend) or "freeze" (for suspend type freeze)
4426 as the system sleep state during system startup with
4427 the optional capability to repeat N number of times.
4428 The system is woken from this state using a
4429 wakeup-capable RTC alarm.
4431 thash_entries= [KNL,NET]
4432 Set number of hash buckets for TCP connection
4434 thermal.act= [HW,ACPI]
4435 -1: disable all active trip points in all thermal zones
4436 <degrees C>: override all lowest active trip points
4438 thermal.crt= [HW,ACPI]
4439 -1: disable all critical trip points in all thermal zones
4440 <degrees C>: override all critical trip points
4442 thermal.nocrt= [HW,ACPI]
4443 Set to disable actions on ACPI thermal zone
4444 critical and hot trip points.
4446 thermal.off= [HW,ACPI]
4447 1: disable ACPI thermal control
4449 thermal.psv= [HW,ACPI]
4450 -1: disable all passive trip points
4451 <degrees C>: override all passive trip points to this
4454 thermal.tzp= [HW,ACPI]
4455 Specify global default ACPI thermal zone polling rate
4456 <deci-seconds>: poll all this frequency
4457 0: no polling (default)
4460 Force threading of all interrupt handlers except those
4461 marked explicitly IRQF_NO_THREAD.
4464 Enable the Transcendent memory driver if built-in.
4466 tmem.cleancache=0|1 [KNL, XEN]
4467 Default is on (1). Disable the usage of the cleancache
4468 API to send anonymous pages to the hypervisor.
4470 tmem.frontswap=0|1 [KNL, XEN]
4471 Default is on (1). Disable the usage of the frontswap
4472 API to send swap pages to the hypervisor. If disabled
4473 the selfballooning and selfshrinking are force disabled.
4475 tmem.selfballooning=0|1 [KNL, XEN]
4476 Default is on (1). Disable the driving of swap pages
4479 tmem.selfshrinking=0|1 [KNL, XEN]
4480 Default is on (1). Partial swapoff that immediately
4481 transfers pages from Xen hypervisor back to the
4482 kernel based on different criteria.
4486 Specify if the kernel should make use of the cpu
4487 topology information if the hardware supports this.
4488 The scheduler will make use of this information and
4489 e.g. base its process migration decisions on it.
4492 topology_updates= [KNL, PPC, NUMA]
4494 Specify if the kernel should ignore (off)
4495 topology updates sent by the hypervisor to this
4500 tpm_suspend_pcr=[HW,TPM]
4501 Format: integer pcr id
4502 Specify that at suspend time, the tpm driver
4503 should extend the specified pcr with zeros,
4504 as a workaround for some chips which fail to
4505 flush the last written pcr on TPM_SaveState.
4506 This will guarantee that all the other pcrs
4509 trace_buf_size=nn[KMG]
4510 [FTRACE] will set tracing buffer size on each cpu.
4512 trace_event=[event-list]
4513 [FTRACE] Set and start specified trace events in order
4514 to facilitate early boot debugging. The event-list is a
4515 comma separated list of trace events to enable. See
4516 also Documentation/trace/events.txt
4518 trace_options=[option-list]
4519 [FTRACE] Enable or disable tracer options at boot.
4520 The option-list is a comma delimited list of options
4521 that can be enabled or disabled just as if you were
4522 to echo the option name into
4524 /sys/kernel/debug/tracing/trace_options
4526 For example, to enable stacktrace option (to dump the
4527 stack trace of each event), add to the command line:
4529 trace_options=stacktrace
4531 See also Documentation/trace/ftrace.txt "trace options"
4535 Have the tracepoints sent to printk as well as the
4536 tracing ring buffer. This is useful for early boot up
4537 where the system hangs or reboots and does not give the
4538 option for reading the tracing buffer or performing a
4539 ftrace_dump_on_oops.
4541 To turn off having tracepoints sent to printk,
4542 echo 0 > /proc/sys/kernel/tracepoint_printk
4543 Note, echoing 1 into this file without the
4544 tracepoint_printk kernel cmdline option has no effect.
4548 Having tracepoints sent to printk() and activating high
4549 frequency tracepoints such as irq or sched, can cause
4550 the system to live lock.
4553 [FTRACE] enable this option to disable tracing when a
4554 warning is hit. This turns off "tracing_on". Tracing can
4555 be enabled again by echoing '1' into the "tracing_on"
4556 file located in /sys/kernel/debug/tracing/
4558 This option is useful, as it disables the trace before
4559 the WARNING dump is called, which prevents the trace to
4560 be filled with content caused by the warning output.
4562 This option can also be set at run time via the sysctl
4563 option: kernel/traceoff_on_warning
4565 transparent_hugepage=
4567 Format: [always|madvise|never]
4568 Can be used to control the default behavior of the system
4569 with respect to transparent hugepages.
4570 See Documentation/vm/transhuge.txt for more details.
4572 tsc= Disable clocksource stability checks for TSC.
4574 [x86] reliable: mark tsc clocksource as reliable, this
4575 disables clocksource verification at runtime, as well
4576 as the stability checks done at bootup. Used to enable
4577 high-resolution timer mode on older hardware, and in
4578 virtualized environment.
4579 [x86] noirqtime: Do not use TSC to do irq accounting.
4580 Used to run time disable IRQ_TIME_ACCOUNTING on any
4581 platforms where RDTSC is slow and this accounting
4584 tsx= [X86] Control Transactional Synchronization
4585 Extensions (TSX) feature in Intel processors that
4586 support TSX control.
4588 This parameter controls the TSX feature. The options are:
4590 on - Enable TSX on the system. Although there are
4591 mitigations for all known security vulnerabilities,
4592 TSX has been known to be an accelerator for
4593 several previous speculation-related CVEs, and
4594 so there may be unknown security risks associated
4595 with leaving it enabled.
4597 off - Disable TSX on the system. (Note that this
4598 option takes effect only on newer CPUs which are
4599 not vulnerable to MDS, i.e., have
4600 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
4601 the new IA32_TSX_CTRL MSR through a microcode
4602 update. This new MSR allows for the reliable
4603 deactivation of the TSX functionality.)
4605 auto - Disable TSX if X86_BUG_TAA is present,
4606 otherwise enable TSX on the system.
4608 Not specifying this option is equivalent to tsx=off.
4610 See Documentation/hw-vuln/tsx_async_abort.rst
4613 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
4614 Abort (TAA) vulnerability.
4616 Similar to Micro-architectural Data Sampling (MDS)
4617 certain CPUs that support Transactional
4618 Synchronization Extensions (TSX) are vulnerable to an
4619 exploit against CPU internal buffers which can forward
4620 information to a disclosure gadget under certain
4623 In vulnerable processors, the speculatively forwarded
4624 data can be used in a cache side channel attack, to
4625 access data to which the attacker does not have direct
4628 This parameter controls the TAA mitigation. The
4631 full - Enable TAA mitigation on vulnerable CPUs
4634 full,nosmt - Enable TAA mitigation and disable SMT on
4635 vulnerable CPUs. If TSX is disabled, SMT
4636 is not disabled because CPU is not
4637 vulnerable to cross-thread TAA attacks.
4638 off - Unconditionally disable TAA mitigation
4640 On MDS-affected machines, tsx_async_abort=off can be
4641 prevented by an active MDS mitigation as both vulnerabilities
4642 are mitigated with the same mechanism so in order to disable
4643 this mitigation, you need to specify mds=off too.
4645 Not specifying this option is equivalent to
4646 tsx_async_abort=full. On CPUs which are MDS affected
4647 and deploy MDS mitigation, TAA mitigation is not
4648 required and doesn't provide any additional
4652 Documentation/hw-vuln/tsx_async_abort.rst
4654 turbografx.map[2|3]= [HW,JOY]
4655 TurboGraFX parallel port interface
4657 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4658 See also Documentation/input/joystick-parport.txt
4660 udbg-immortal [PPC] When debugging early kernel crashes that
4661 happen after console_init() and before a proper
4662 console driver takes over, this boot options might
4663 help "seeing" what's going on.
4665 uhash_entries= [KNL,NET]
4666 Set number of hash buckets for UDP/UDP-Lite connections
4669 [USB] Ignore overcurrent events (default N).
4670 Some badly-designed motherboards generate lots of
4671 bogus events, for ports that aren't wired to
4672 anything. Set this parameter to avoid log spamming.
4673 Note that genuine overcurrent events won't be
4677 [X86] Cause panic on unknown NMI.
4679 usbcore.authorized_default=
4680 [USB] Default USB device authorization:
4681 (default -1 = authorized except for wireless USB,
4682 0 = not authorized, 1 = authorized)
4684 usbcore.autosuspend=
4685 [USB] The autosuspend time delay (in seconds) used
4686 for newly-detected USB devices (default 2). This
4687 is the time required before an idle device will be
4688 autosuspended. Devices for which the delay is set
4689 to a negative value won't be autosuspended at all.
4691 usbcore.usbfs_snoop=
4692 [USB] Set to log all usbfs traffic (default 0 = off).
4694 usbcore.usbfs_snoop_max=
4695 [USB] Maximum number of bytes to snoop in each URB
4698 usbcore.blinkenlights=
4699 [USB] Set to cycle leds on hubs (default 0 = off).
4701 usbcore.old_scheme_first=
4702 [USB] Start with the old device initialization
4703 scheme (default 0 = off).
4705 usbcore.usbfs_memory_mb=
4706 [USB] Memory limit (in MB) for buffers allocated by
4707 usbfs (default = 16, 0 = max = 2047).
4709 usbcore.use_both_schemes=
4710 [USB] Try the other device initialization scheme
4711 if the first one fails (default 1 = enabled).
4713 usbcore.initial_descriptor_timeout=
4714 [USB] Specifies timeout for the initial 64-byte
4715 USB_REQ_GET_DESCRIPTOR request in milliseconds
4716 (default 5000 = 5.0 seconds).
4718 usbcore.nousb [USB] Disable the USB subsystem
4721 [USBHID] The interval which mice are to be polled at.
4723 usb-storage.delay_use=
4724 [UMS] The delay in seconds before a new device is
4725 scanned for Logical Units (default 1).
4728 [UMS] A list of quirks entries to supplement or
4729 override the built-in unusual_devs list. List
4730 entries are separated by commas. Each entry has
4731 the form VID:PID:Flags where VID and PID are Vendor
4732 and Product ID values (4-digit hex numbers) and
4733 Flags is a set of characters, each corresponding
4734 to a common usb-storage quirk flag as follows:
4735 a = SANE_SENSE (collect more than 18 bytes
4737 b = BAD_SENSE (don't collect more than 18
4738 bytes of sense data);
4739 c = FIX_CAPACITY (decrease the reported
4740 device capacity by one sector);
4741 d = NO_READ_DISC_INFO (don't use
4742 READ_DISC_INFO command);
4743 e = NO_READ_CAPACITY_16 (don't use
4744 READ_CAPACITY_16 command);
4745 f = NO_REPORT_OPCODES (don't use report opcodes
4747 g = MAX_SECTORS_240 (don't transfer more than
4748 240 sectors at a time, uas only);
4749 h = CAPACITY_HEURISTICS (decrease the
4750 reported device capacity by one
4751 sector if the number is odd);
4752 i = IGNORE_DEVICE (don't bind to this
4754 j = NO_REPORT_LUNS (don't use report luns
4756 l = NOT_LOCKABLE (don't try to lock and
4757 unlock ejectable media);
4758 m = MAX_SECTORS_64 (don't transfer more
4759 than 64 sectors = 32 KB at a time);
4760 n = INITIAL_READ10 (force a retry of the
4761 initial READ(10) command);
4762 o = CAPACITY_OK (accept the capacity
4763 reported by the device);
4764 p = WRITE_CACHE (the device cache is ON
4766 r = IGNORE_RESIDUE (the device reports
4767 bogus residue values);
4768 s = SINGLE_LUN (the device has only one
4770 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
4771 commands, uas only);
4772 u = IGNORE_UAS (don't bind to the uas driver);
4773 w = NO_WP_DETECT (don't test whether the
4774 medium is write-protected).
4775 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
4776 even if the device claims no cache)
4777 Example: quirks=0419:aaf5:rl,0421:0433:rc
4779 user_debug= [KNL,ARM]
4781 See arch/arm/Kconfig.debug help text.
4782 1 - undefined instruction events
4784 4 - invalid data aborts
4787 Example: user_debug=31
4790 [X86] Flags controlling user PTE allocations.
4792 nohigh = do not allocate PTE pages in
4793 HIGHMEM regardless of setting
4797 On X86_32, this is an alias for vdso32=. Otherwise:
4799 vdso=1: enable VDSO (the default)
4800 vdso=0: disable VDSO mapping
4802 vdso32= [X86] Control the 32-bit vDSO
4803 vdso32=1: enable 32-bit VDSO
4804 vdso32=0 or vdso32=2: disable 32-bit VDSO
4806 See the help text for CONFIG_COMPAT_VDSO for more
4807 details. If CONFIG_COMPAT_VDSO is set, the default is
4808 vdso32=0; otherwise, the default is vdso32=1.
4810 For compatibility with older kernels, vdso32=2 is an
4813 Try vdso32=0 if you encounter an error that says:
4814 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
4817 vector=percpu: enable percpu vector domain
4819 video= [FB] Frame buffer configuration
4820 See Documentation/fb/modedb.txt.
4822 video.brightness_switch_enabled= [0,1]
4823 If set to 1, on receiving an ACPI notify event
4824 generated by hotkey, video driver will adjust brightness
4825 level and then send out the event to user space through
4826 the allocated input device; If set to 0, video driver
4827 will only send out the event without touching backlight
4832 [VMMIO] Memory mapped virtio (platform) device.
4834 <size>@<baseaddr>:<irq>[:<id>]
4836 <size> := size (can use standard suffixes
4838 <baseaddr> := physical base address
4839 <irq> := interrupt number (as passed to
4841 <id> := (optional) platform device id
4843 virtio_mmio.device=1K@0x100b0000:48:7
4845 Can be used multiple times for multiple devices.
4847 vga= [BOOT,X86-32] Select a particular video mode
4848 See Documentation/x86/boot.txt and
4849 Documentation/svga.txt.
4850 Use vga=ask for menu.
4851 This is actually a boot loader parameter; the value is
4852 passed to the kernel using a special protocol.
4854 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
4855 size of <nn>. This can be used to increase the
4856 minimum size (128MB on x86). It can also be used to
4857 decrease the size and leave more room for directly
4860 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
4863 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
4866 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
4870 Controls the behavior of vsyscalls (i.e. calls to
4871 fixed addresses of 0xffffffffff600x00 from legacy
4872 code). Most statically-linked binaries and older
4873 versions of glibc use these calls. Because these
4874 functions are at fixed addresses, they make nice
4875 targets for exploits that can control RIP.
4877 emulate [default] Vsyscalls turn into traps and are
4878 emulated reasonably safely.
4880 native Vsyscalls are native syscall instructions.
4881 This is a little bit faster than trapping
4882 and makes a few dynamic recompilers work
4883 better than they would in emulation mode.
4884 It also makes exploits much easier to write.
4886 none Vsyscalls don't work at all. This makes
4887 them quite hard to use for exploits but
4888 might break your system.
4890 vt.color= [VT] Default text color.
4891 Format: 0xYX, X = foreground, Y = background.
4892 Default: 0x07 = light gray on black.
4894 vt.cur_default= [VT] Default cursor shape.
4895 Format: 0xCCBBAA, where AA, BB, and CC are the same as
4896 the parameters of the <Esc>[?A;B;Cc escape sequence;
4897 see VGA-softcursor.txt. Default: 2 = underline.
4899 vt.default_blu= [VT]
4900 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
4901 Change the default blue palette of the console.
4902 This is a 16-member array composed of values
4905 vt.default_grn= [VT]
4906 Format: <green0>,<green1>,<green2>,...,<green15>
4907 Change the default green palette of the console.
4908 This is a 16-member array composed of values
4911 vt.default_red= [VT]
4912 Format: <red0>,<red1>,<red2>,...,<red15>
4913 Change the default red palette of the console.
4914 This is a 16-member array composed of values
4920 Set system-wide default UTF-8 mode for all tty's.
4921 Default is 1, i.e. UTF-8 mode is enabled for all
4922 newly opened terminals.
4924 vt.global_cursor_default=
4927 Set system-wide default for whether a cursor
4928 is shown on new VTs. Default is -1,
4929 i.e. cursors will be created by default unless
4930 overridden by individual drivers. 0 will hide
4931 cursors, 1 will display them.
4933 vt.italic= [VT] Default color for italic text; 0-15.
4936 vt.underline= [VT] Default color for underlined text; 0-15.
4939 watchdog timers [HW,WDT] For information on watchdog timers,
4940 see Documentation/watchdog/watchdog-parameters.txt
4941 or other driver-specific files in the
4942 Documentation/watchdog/ directory.
4944 workqueue.watchdog_thresh=
4945 If CONFIG_WQ_WATCHDOG is configured, workqueue can
4946 warn stall conditions and dump internal state to
4947 help debugging. 0 disables workqueue stall
4948 detection; otherwise, it's the stall threshold
4949 duration in seconds. The default value is 30 and
4950 it can be updated at runtime by writing to the
4951 corresponding sysfs file.
4953 workqueue.disable_numa
4954 By default, all work items queued to unbound
4955 workqueues are affine to the NUMA nodes they're
4956 issued on, which results in better behavior in
4957 general. If NUMA affinity needs to be disabled for
4958 whatever reason, this option can be used. Note
4959 that this also can be controlled per-workqueue for
4960 workqueues visible under /sys/bus/workqueue/.
4962 workqueue.power_efficient
4963 Per-cpu workqueues are generally preferred because
4964 they show better performance thanks to cache
4965 locality; unfortunately, per-cpu workqueues tend to
4966 be more power hungry than unbound workqueues.
4968 Enabling this makes the per-cpu workqueues which
4969 were observed to contribute significantly to power
4970 consumption unbound, leading to measurably lower
4971 power usage at the cost of small performance
4974 The default value of this parameter is determined by
4975 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
4977 workqueue.debug_force_rr_cpu
4978 Workqueue used to implicitly guarantee that work
4979 items queued without explicit CPU specified are put
4980 on the local CPU. This guarantee is no longer true
4981 and while local CPU is still preferred work items
4982 may be put on foreign CPUs. This debug option
4983 forces round-robin CPU selection to flush out
4984 usages which depend on the now broken guarantee.
4985 When enabled, memory and cache locality will be
4988 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
4989 default x2apic cluster mode on platforms
4992 x86_intel_mid_timer= [X86-32,APBT]
4993 Choose timer option for x86 Intel MID platform.
4994 Two valid options are apbt timer only and lapic timer
4995 plus one apbt timer for broadcast timer.
4996 x86_intel_mid_timer=apbt_only | lapic_and_apbt
4998 xen_512gb_limit [KNL,X86-64,XEN]
4999 Restricts the kernel running paravirtualized under Xen
5000 to use only up to 512 GB of RAM. The reason to do so is
5001 crash analysis tools and Xen tools for doing domain
5002 save/restore/migration must be enabled to handle larger
5005 xen_emul_unplug= [HW,X86,XEN]
5006 Unplug Xen emulated devices
5007 Format: [unplug0,][unplug1]
5008 ide-disks -- unplug primary master IDE devices
5009 aux-ide-disks -- unplug non-primary-master IDE devices
5010 nics -- unplug network devices
5011 all -- unplug all emulated devices (NICs and IDE disks)
5012 unnecessary -- unplugging emulated devices is
5013 unnecessary even if the host did not respond to
5015 never -- do not unplug even if version check succeeds
5017 xen_nopvspin [X86,XEN]
5018 Disables the ticketlock slowpath using Xen PV
5022 Disables the PV optimizations forcing the HVM guest to
5023 run as generic HVM guest with no PV drivers.
5025 xen.event_eoi_delay= [XEN]
5026 How long to delay EOI handling in case of event
5027 storms (jiffies). Default is 10.
5029 xen.event_loop_timeout= [XEN]
5030 After which time (jiffies) the event handling loop
5031 should start to delay EOI handling. Default is 2.
5033 xirc2ps_cs= [NET,PCMCIA]
5035 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5037 ______________________________________________________________________
5041 Add more DRM drivers.