1 acpi= [HW,ACPI,X86,ARM64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
5 force -- enable ACPI if default was off
6 on -- enable ACPI but allow fallback to DT [arm64]
7 off -- disable ACPI if default was on
8 noirq -- do not use ACPI for IRQ routing
9 strict -- Be less tolerant of platforms that are not
10 strictly ACPI specification compliant.
11 rsdt -- prefer RSDT over (default) XSDT
12 copy_dsdt -- copy DSDT to memory
13 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force"
16 See also Documentation/power/runtime_pm.txt, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
27 If set to vendor, prefer vendor specific driver
28 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
29 of the ACPI video.ko driver.
31 acpi_force_32bit_fadt_addr
32 force FADT to use 32 bit addresses rather than the
33 64 bit X_* addresses. Some firmware have broken 64
34 bit addresses for force ACPI ignore these and use
35 the older legacy 32 bit addresses.
37 acpica_no_return_repair [HW, ACPI]
38 Disable AML predefined validation mechanism
39 This mechanism can repair the evaluation result to make
40 the return objects more ACPI specification compliant.
41 This option is useful for developers to identify the
42 root cause of an AML interpreter issue when the issue
43 has something to do with the repair mechanism.
45 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
46 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
48 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
49 debug output. Bits in debug_layer correspond to a
50 _COMPONENT in an ACPI source file, e.g.,
51 #define _COMPONENT ACPI_PCI_COMPONENT
52 Bits in debug_level correspond to a level in
53 ACPI_DEBUG_PRINT statements, e.g.,
54 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
55 The debug_level mask defaults to "info". See
56 Documentation/acpi/debug.txt for more information about
57 debug layers and levels.
59 Enable processor driver info messages:
60 acpi.debug_layer=0x20000000
61 Enable PCI/PCI interrupt routing info messages:
62 acpi.debug_layer=0x400000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_no_auto_serialize [HW,ACPI]
110 Disable auto-serialization of AML methods
111 AML control methods that contain the opcodes to create
112 named objects will be marked as "Serialized" by the
113 auto-serialization feature.
114 This feature is enabled by default.
115 This option allows to turn off the feature.
117 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
120 acpi_no_static_ssdt [HW,ACPI]
121 Disable installation of static SSDTs at early boot time
122 By default, SSDTs contained in the RSDT/XSDT will be
123 installed automatically and they will appear under
124 /sys/firmware/acpi/tables.
125 This option turns off this feature.
126 Note that specifying this option does not affect
127 dynamic table installation which will install SSDT
128 tables to /sys/firmware/acpi/tables/dynamic.
130 acpi_rsdp= [ACPI,EFI,KEXEC]
131 Pass the RSDP address to the kernel, mostly used
132 on machines running EFI runtime service to boot the
133 second kernel for kdump.
135 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
136 Format: To spoof as Windows 98: ="Microsoft Windows"
138 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
139 of 2 which is mandated by ACPI 6) as the supported ACPI
140 specification revision (when using this switch, it may
141 be necessary to carry out a cold reboot _twice_ in a
142 row to make it take effect on the platform firmware).
144 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
145 acpi_osi="string1" # add string1
146 acpi_osi="!string2" # remove string2
147 acpi_osi=!* # remove all strings
148 acpi_osi=! # disable all built-in OS vendor
150 acpi_osi=!! # enable all built-in OS vendor
152 acpi_osi= # disable all strings
154 'acpi_osi=!' can be used in combination with single or
155 multiple 'acpi_osi="string1"' to support specific OS
156 vendor string(s). Note that such command can only
157 affect the default state of the OS vendor strings, thus
158 it cannot affect the default state of the feature group
159 strings and the current state of the OS vendor strings,
160 specifying it multiple times through kernel command line
161 is meaningless. This command is useful when one do not
162 care about the state of the feature group strings which
163 should be controlled by the OSPM.
165 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
166 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
167 can make '_OSI("Windows 2000")' TRUE.
169 'acpi_osi=' cannot be used in combination with other
170 'acpi_osi=' command lines, the _OSI method will not
171 exist in the ACPI namespace. NOTE that such command can
172 only affect the _OSI support state, thus specifying it
173 multiple times through kernel command line is also
176 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
179 'acpi_osi=!*' can be used in combination with single or
180 multiple 'acpi_osi="string1"' to support specific
181 string(s). Note that such command can affect the
182 current state of both the OS vendor strings and the
183 feature group strings, thus specifying it multiple times
184 through kernel command line is meaningful. But it may
185 still not able to affect the final state of a string if
186 there are quirks related to this string. This command
187 is useful when one want to control the state of the
188 feature group strings to debug BIOS issues related to
191 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
192 '_OSI("Module Device")' FALSE.
193 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
194 '_OSI("Module Device")' TRUE.
195 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
197 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
199 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
200 they all will make '_OSI("Windows 2000")' TRUE.
203 Override the pmtimer bug detection: force the kernel
204 to assume that this machine's pmtimer latches its value
205 and always returns good values.
207 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
208 Format: { level | edge | high | low }
210 acpi_skip_timer_override [HW,ACPI]
211 Recognize and ignore IRQ0/pin2 Interrupt Override.
212 For broken nForce2 BIOS resulting in XT-PIC timer.
214 acpi_sleep= [HW,ACPI] Sleep options
215 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
216 old_ordering, nonvs, sci_force_enable }
217 See Documentation/power/video.txt for information on
219 s3_beep is for debugging; it makes the PC's speaker beep
220 as soon as the kernel's real-mode entry point is called.
221 s4_nohwsig prevents ACPI hardware signature from being
222 used during resume from hibernation.
223 old_ordering causes the ACPI 1.0 ordering of the _PTS
224 control method, with respect to putting devices into
225 low power states, to be enforced (the ACPI 2.0 ordering
226 of _PTS is used by default).
227 nonvs prevents the kernel from saving/restoring the
228 ACPI NVS memory during suspend/hibernation and resume.
229 sci_force_enable causes the kernel to set SCI_EN directly
230 on resume from S1/S3 (which is against the ACPI spec,
231 but some broken systems don't work without it).
233 acpi_use_timer_override [HW,ACPI]
234 Use timer override. For some broken Nvidia NF5 boards
235 that require a timer override, but don't have HPET
237 add_efi_memmap [EFI; X86] Include EFI memory map in
238 kernel's map of available physical RAM.
241 { off | try_unsupported }
242 off: disable AGP support
243 try_unsupported: try to drive unsupported chipsets
244 (may crash computer or cause data corruption)
247 See Documentation/sound/alsa/alsa-parameters.txt
250 Allow the default userspace alignment fault handler
251 behaviour to be specified. Bit 0 enables warnings,
252 bit 1 enables fixups, and bit 2 sends a segfault.
254 align_va_addr= [X86-64]
255 Align virtual addresses by clearing slice [14:12] when
256 allocating a VMA at process creation time. This option
257 gives you up to 3% performance improvement on AMD F15h
258 machines (where it is enabled by default) for a
259 CPU-intensive style benchmark, and it can vary highly in
260 a microbenchmark depending on workload and compiler.
262 32: only for 32-bit processes
263 64: only for 64-bit processes
264 on: enable for both 32- and 64-bit processes
265 off: disable for both 32- and 64-bit processes
267 alloc_snapshot [FTRACE]
268 Allocate the ftrace snapshot buffer on boot up when the
269 main buffer is allocated. This is handy if debugging
270 and you need to use tracing_snapshot() on boot up, and
271 do not want to use tracing_snapshot_alloc() as it needs
272 to be done where GFP_KERNEL allocations are allowed.
274 amd_iommu= [HW,X86-64]
275 Pass parameters to the AMD IOMMU driver in the system.
277 fullflush - enable flushing of IO/TLB entries when
278 they are unmapped. Otherwise they are
279 flushed before they will be reused, which
281 off - do not initialize any AMD IOMMU found in
283 force_isolation - Force device isolation for all
284 devices. The IOMMU driver is not
285 allowed anymore to lift isolation
286 requirements as needed. This option
287 does not override iommu=pt
289 amd_iommu_dump= [HW,X86-64]
290 Enable AMD IOMMU driver option to dump the ACPI table
291 for AMD IOMMU. With this option enabled, AMD IOMMU
292 driver will print ACPI tables for AMD IOMMU during
293 IOMMU initialization.
295 amd_iommu_intr= [HW,X86-64]
296 Specifies one of the following AMD IOMMU interrupt
298 legacy - Use legacy interrupt remapping mode.
299 vapic - Use virtual APIC mode, which allows IOMMU
300 to inject interrupts directly into guest.
301 This mode requires kvm-amd.avic=1.
302 (Default when IOMMU HW support is present.)
304 amijoy.map= [HW,JOY] Amiga joystick support
305 Map of devices attached to JOY0DAT and JOY1DAT
307 See also Documentation/input/joystick.txt
309 analog.map= [HW,JOY] Analog joystick and gamepad support
310 Specifies type or capabilities of an analog joystick
311 connected to one of 16 gameports
312 Format: <type1>,<type2>,..<type16>
315 Power management functions (SPARCstation-4/5 + deriv.)
317 Disable APC CPU standby support. SPARCstation-Fox does
318 not play well with APC CPU idle - disable it if you have
319 APC and your system crashes randomly.
321 apic= [APIC,X86-32] Advanced Programmable Interrupt Controller
322 Change the output verbosity whilst booting
323 Format: { quiet (default) | verbose | debug }
324 Change the amount of debugging information output
325 when initialising the APIC and IO-APIC components.
327 apic_extnmi= [APIC,X86] External NMI delivery setting
328 Format: { bsp (default) | all | none }
329 bsp: External NMI is delivered only to CPU 0
330 all: External NMIs are broadcast to all CPUs as a
332 none: External NMI is masked for all CPUs. This is
333 useful so that a dump capture kernel won't be
337 See Documentation/networking/ipv6.txt.
339 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
340 Limit apic dumping. The parameter defines the maximal
341 number of local apics being dumped. Also it is possible
342 to set it to "all" by meaning -- no limit here.
343 Format: { 1 (default) | 2 | ... | all }.
344 The parameter valid if only apic=debug or
345 apic=verbose is specified.
346 Example: apic=debug show_lapic=all
348 apm= [APM] Advanced Power Management
349 See header of arch/x86/kernel/apm_32.c.
351 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
352 Format: <io>,<irq>,<nodeID>
356 atarimouse= [HW,MOUSE] Atari Mouse
358 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
359 EzKey and similar keyboards
361 atkbd.reset= [HW] Reset keyboard during initialization
363 atkbd.set= [HW] Select keyboard code set
364 Format: <int> (2 = AT (default), 3 = PS/2)
366 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
369 atkbd.softraw= [HW] Choose between synthetic and real raw mode
370 Format: <bool> (0 = real, 1 = synthetic (default))
372 atkbd.softrepeat= [HW]
373 Use software keyboard repeat
375 audit= [KNL] Enable the audit sub-system
376 Format: { "0" | "1" } (0 = disabled, 1 = enabled)
377 0 - kernel audit is disabled and can not be enabled
378 until the next reboot
379 unset - kernel audit is initialized but disabled and
380 will be fully enabled by the userspace auditd.
381 1 - kernel audit is initialized and partially enabled,
382 storing at most audit_backlog_limit messages in
383 RAM until it is fully enabled by the userspace
387 audit_backlog_limit= [KNL] Set the audit queue size limit.
388 Format: <int> (must be >=0)
391 bau= [X86_UV] Enable the BAU on SGI UV. The default
392 behavior is to disable the BAU (i.e. bau=0).
393 Format: { "0" | "1" }
396 unset - Disable the BAU.
398 baycom_epp= [HW,AX25]
401 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
403 See header of drivers/net/hamradio/baycom_par.c.
405 baycom_ser_fdx= [HW,AX25]
406 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
407 Format: <io>,<irq>,<mode>[,<baud>]
408 See header of drivers/net/hamradio/baycom_ser_fdx.c.
410 baycom_ser_hdx= [HW,AX25]
411 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
412 Format: <io>,<irq>,<mode>
413 See header of drivers/net/hamradio/baycom_ser_hdx.c.
415 blkdevparts= Manual partition parsing of block device(s) for
416 embedded devices based on command line input.
417 See Documentation/block/cmdline-partition.txt
419 boot_delay= Milliseconds to delay each printk during boot.
420 Values larger than 10 seconds (10000) are changed to
424 bootmem_debug [KNL] Enable bootmem allocator debug messages.
427 Disable BERT OS support on buggy BIOSes.
429 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
430 bttv.radio= Most important insmod options are available as
432 bttv.pll= See Documentation/video4linux/bttv/Insmod-options
435 bulk_remove=off [PPC] This parameter disables the use of the pSeries
436 firmware feature for flushing multiple hpte entries
439 c101= [NET] Moxa C101 synchronous serial card
441 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
442 Sometimes CPU hardware bugs make them report the cache
443 size incorrectly. The kernel will attempt work arounds
444 to fix known problems, but for some CPUs it is not
445 possible to determine what the correct size should be.
446 This option provides an override for these situations.
448 ca_keys= [KEYS] This parameter identifies a specific key(s) on
449 the system trusted keyring to be used for certificate
451 format: { id:<keyid> | builtin }
453 cca= [MIPS] Override the kernel pages' cache coherency
454 algorithm. Accepted values range from 0 to 7
455 inclusive. See arch/mips/include/asm/pgtable-bits.h
456 for platform specific values (SB1, Loongson3 and
459 ccw_timeout_log [S390]
460 See Documentation/s390/CommonIO for details.
462 cgroup_disable= [KNL] Disable a particular controller
463 Format: {name of the controller(s) to disable}
464 The effects of cgroup_disable=foo are:
465 - foo isn't auto-mounted if you mount all cgroups in
467 - foo isn't visible as an individually mountable
469 {Currently only "memory" controller deal with this and
470 cut the overhead, others just disable the usage. So
471 only cgroup_disable=memory is actually worthy}
473 cgroup_no_v1= [KNL] Disable one, multiple, all cgroup controllers in v1
474 Format: { controller[,controller...] | "all" }
475 Like cgroup_disable, but only applies to cgroup v1;
476 the blacklisted controllers remain available in cgroup2.
478 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
480 nosocket -- Disable socket memory accounting.
481 nokmem -- Disable kernel memory accounting.
483 checkreqprot [SELINUX] Set initial checkreqprot flag value.
484 Format: { "0" | "1" }
485 See security/selinux/Kconfig help text.
486 0 -- check protection applied by kernel (includes
487 any implied execute protection).
488 1 -- check protection requested by application.
489 Default value is set via a kernel config option.
490 Value can be changed at runtime via
491 /selinux/checkreqprot.
494 See Documentation/s390/CommonIO for details.
497 Prevents the clock framework from automatically gating
498 clocks that have not been explicitly enabled by a Linux
499 device driver but are enabled in hardware at reset or
500 by the bootloader/firmware. Note that this does not
501 force such clocks to be always-on nor does it reserve
502 those clocks in any way. This parameter is useful for
503 debug and development, but should not be needed on a
504 platform with proper driver support. For more
505 information, see Documentation/clk.txt.
507 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
509 Forces specified clocksource (if available) to be used
510 when calculating gettimeofday(). If specified
511 clocksource is not available, it defaults to PIT.
512 Format: { pit | tsc | cyclone | pmtmr }
514 clocksource= Override the default clocksource
516 Override the default clocksource and use the clocksource
517 with the name specified.
518 Some clocksource names to choose from, depending on
520 [all] jiffies (this is the base, fallback clocksource)
522 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
523 pxa_timer,timer3,32k_counter,timer0_1
525 [X86-32] pit,hpet,tsc;
526 scx200_hrt on Geode; cyclone on IBM x440
534 clocksource.arm_arch_timer.evtstrm=
537 Enable/disable the eventstream feature of the ARM
538 architected timer so that code using WFE-based polling
539 loops can be debugged more effectively on production
542 clocksource.arm_arch_timer.fsl-a008585=
545 Enable/disable the workaround of Freescale/NXP
546 erratum A-008585. This can be useful for KVM
547 guests, if the guest device tree doesn't show the
548 erratum. If unspecified, the workaround is
549 enabled based on the device tree.
551 clearcpuid=BITNUM [X86]
552 Disable CPUID feature X for the kernel. See
553 arch/x86/include/asm/cpufeatures.h for the valid bit
554 numbers. Note the Linux specific bits are not necessarily
555 stable over kernel options, but the vendor specific
557 Also note that user programs calling CPUID directly
558 or using the feature without checking anything
559 will still see it. This just prevents it from
560 being used by the kernel or shown in /proc/cpuinfo.
561 Also note the kernel might malfunction if you disable
564 cma=nn[MG]@[start[MG][-end[MG]]]
566 Sets the size of kernel global memory area for
567 contiguous memory allocations and optionally the
568 placement constraint by the physical address range of
569 memory allocations. A value of 0 disables CMA
570 altogether. For more information, see
571 include/linux/dma-contiguous.h
573 cmo_free_hint= [PPC] Format: { yes | no }
574 Specify whether pages are marked as being inactive
575 when they are freed. This is used in CMO environments
576 to determine OS memory pressure for page stealing by
580 coherent_pool=nn[KMG] [ARM,KNL]
581 Sets the size of memory pool for coherent, atomic dma
582 allocations, by default set to 256K.
584 code_bytes [X86] How many bytes of object code to print
589 com20020= [HW,NET] ARCnet - COM20020 chipset
591 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
593 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
597 ARCnet - COM90xx chipset (memory-mapped buffers)
598 Format: <io>[,<irq>[,<memstart>]]
600 condev= [HW,S390] console device
603 console= [KNL] Output console device and options.
605 tty<n> Use the virtual console device <n>.
609 Use the specified serial port. The options are of
610 the form "bbbbpnf", where "bbbb" is the baud rate,
611 "p" is parity ("n", "o", or "e"), "n" is number of
612 bits, and "f" is flow control ("r" for RTS or
613 omit it). Default is "9600n8".
615 See Documentation/admin-guide/serial-console.rst for more
617 Documentation/networking/netconsole.txt for an
620 uart[8250],io,<addr>[,options]
621 uart[8250],mmio,<addr>[,options]
622 uart[8250],mmio16,<addr>[,options]
623 uart[8250],mmio32,<addr>[,options]
624 uart[8250],0x<addr>[,options]
625 Start an early, polled-mode console on the 8250/16550
626 UART at the specified I/O port or MMIO address,
627 switching to the matching ttyS device later.
628 MMIO inter-register address stride is either 8-bit
629 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
630 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
631 to be equivalent to 'mmio'. 'options' are specified in
632 the same format described for ttyS above; if unspecified,
633 the h/w is not re-initialized.
635 hvc<n> Use the hypervisor console device <n>. This is for
636 both Xen and PowerPC hypervisors.
638 If the device connected to the port is not a TTY but a braille
639 device, prepend "brl," before the device type, for instance
641 For now, only VisioBraille is supported.
643 consoleblank= [KNL] The console blank (screen saver) timeout in
644 seconds. Defaults to 10*60 = 10mins. A value of 0
645 disables the blank timer.
648 [KNL] Change the default value for
649 /proc/<pid>/coredump_filter.
650 See also Documentation/filesystems/proc.txt.
652 cpuidle.off=1 [CPU_IDLE]
653 disable the cpuidle sub-system
656 [X86] Delay for N microsec between assert and de-assert
657 of APIC INIT to start processors. This delay occurs
658 on every CPU online, such as boot, and resume from suspend.
661 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
663 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
665 crashkernel=size[KMG][@offset[KMG]]
666 [KNL] Using kexec, Linux can switch to a 'crash kernel'
667 upon panic. This parameter reserves the physical
668 memory region [offset, offset + size] for that kernel
669 image. If '@offset' is omitted, then a suitable offset
670 is selected automatically. Check
671 Documentation/kdump/kdump.txt for further details.
673 crashkernel=range1:size1[,range2:size2,...][@offset]
674 [KNL] Same as above, but depends on the memory
675 in the running system. The syntax of range is
676 start-[end] where start and end are both
677 a memory unit (amount[KMG]). See also
678 Documentation/kdump/kdump.txt for an example.
680 crashkernel=size[KMG],high
681 [KNL, x86_64] range could be above 4G. Allow kernel
682 to allocate physical memory region from top, so could
683 be above 4G if system have more than 4G ram installed.
684 Otherwise memory region will be allocated below 4G, if
686 It will be ignored if crashkernel=X is specified.
687 crashkernel=size[KMG],low
688 [KNL, x86_64] range under 4G. When crashkernel=X,high
689 is passed, kernel could allocate physical memory region
690 above 4G, that cause second kernel crash on system
691 that require some amount of low memory, e.g. swiotlb
692 requires at least 64M+32K low memory, also enough extra
693 low memory is needed to make sure DMA buffers for 32-bit
694 devices won't run out. Kernel would try to allocate at
695 at least 256M below 4G automatically.
696 This one let user to specify own low range under 4G
697 for second kernel instead.
698 0: to disable low allocation.
699 It will be ignored when crashkernel=X,high is not used
700 or memory reserved is below 4G.
703 [KNL] Disable crypto self-tests
708 cs89x0_media= [HW,NET]
709 Format: { rj45 | aui | bnc }
712 See header of drivers/s390/block/dasd_devmap.c.
714 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
715 (one device per port)
716 Format: <port#>,<type>
717 See also Documentation/input/joystick-parport.txt
719 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
720 time. See Documentation/dynamic-debug-howto.txt for
721 details. Deprecated, see dyndbg.
723 debug [KNL] Enable kernel debugging (events log level).
726 [KNL] verbose self-tests
728 Print debugging info while doing the locking API
730 We default to 0 (no extra messages), setting it to
731 1 will print _a lot_ more information - normally
732 only useful to kernel developers.
734 debug_objects [KNL] Enable object debugging
737 [KNL] Disable object debugging
739 debug_guardpage_minorder=
740 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
741 parameter allows control of the order of pages that will
742 be intentionally kept free (and hence protected) by the
743 buddy allocator. Bigger value increase the probability
744 of catching random memory corruption, but reduce the
745 amount of memory for normal system use. The maximum
746 possible value is MAX_ORDER/2. Setting this parameter
747 to 1 or 2 should be enough to identify most random
748 memory corruption problems caused by bugs in kernel or
749 driver code when a CPU writes to (or reads from) a
750 random memory location. Note that there exists a class
751 of memory corruptions problems caused by buggy H/W or
752 F/W or by drivers badly programing DMA (basically when
753 memory is written at bus level and the CPU MMU is
754 bypassed) which are not detectable by
755 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
756 tracking down these problems.
759 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
760 parameter enables the feature at boot time. In
761 default, it is disabled. We can avoid allocating huge
762 chunk of memory for debug pagealloc if we don't enable
763 it at boot time and the system will work mostly same
764 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
765 on: enable the feature
767 debugpat [X86] Enable PAT debugging
769 decnet.addr= [HW,NET]
770 Format: <area>[,<node>]
771 See also Documentation/networking/decnet.txt.
774 [same as hugepagesz=] The size of the default
775 HugeTLB page size. This is the size represented by
776 the legacy /proc/ hugepages APIs, used for SHM, and
777 default size when mounting hugetlbfs filesystems.
778 Defaults to the default architecture's huge page size
782 Set number of hash buckets for dentry cache.
784 disable_1tb_segments [PPC]
785 Disables the use of 1TB hash page table segments. This
786 causes the kernel to fall back to 256MB segments which
787 can be useful when debugging issues that require an SLB
791 See Documentation/networking/ipv6.txt.
794 Disable RADIX MMU mode on POWER9
796 disable_cpu_apicid= [X86,APIC,SMP]
798 The number of initial APIC ID for the
799 corresponding CPU to be disabled at boot,
800 mostly used for the kdump 2nd kernel to
801 disable BSP to wake up multiple CPUs without
802 causing system reset or hang due to sending
805 disable_ddw [PPC/PSERIES]
806 Disable Dynamic DMA Window support. Use this if
807 to workaround buggy firmware.
810 See Documentation/networking/ipv6.txt.
812 disable_mtrr_cleanup [X86]
813 The kernel tries to adjust MTRR layout from continuous
814 to discrete, to make X server driver able to add WB
815 entry later. This parameter disables that.
817 disable_mtrr_trim [X86, Intel and AMD only]
818 By default the kernel will trim any uncacheable
819 memory out of your available memory pool based on
820 MTRR settings. This parameter disables that behavior,
821 possibly causing your machine to run very slowly.
823 disable_timer_pin_1 [X86]
824 Disable PIN 1 of APIC timer
825 Can be useful to work around chipset bugs.
827 dis_ucode_ldr [X86] Disable the microcode loader.
829 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
830 this option disables the debugging code at boot.
832 dma_debug_entries=<number>
833 This option allows to tune the number of preallocated
834 entries for DMA-API debugging code. One entry is
835 required per DMA-API allocation. Use this if the
836 DMA-API debugging code disables itself because the
837 architectural default is too low.
839 dma_debug_driver=<driver_name>
840 With this option the DMA-API debugging driver
841 filter feature can be enabled at boot time. Just
842 pass the driver to filter for as the parameter.
843 The filter can be disabled or changed to another
844 driver later using sysfs.
846 drm_kms_helper.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
847 Broken monitors, graphic adapters, KVMs and EDIDless
848 panels may send no or incorrect EDID data sets.
849 This parameter allows to specify an EDID data sets
850 in the /lib/firmware directory that are used instead.
851 Generic built-in EDID data sets are used, if one of
852 edid/1024x768.bin, edid/1280x1024.bin,
853 edid/1680x1050.bin, or edid/1920x1080.bin is given
854 and no file with the same name exists. Details and
855 instructions how to build your own EDID data are
856 available in Documentation/EDID/HOWTO.txt. An EDID
857 data set will only be used for a particular connector,
858 if its name and a colon are prepended to the EDID
859 name. Each connector may use a unique EDID data
860 set by separating the files with a comma. An EDID
861 data set with no connector name will be used for
862 any connectors not explicitly specified.
866 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
867 module.dyndbg[="val"]
868 Enable debug messages at boot time. See
869 Documentation/dynamic-debug-howto.txt for details.
871 nompx [X86] Disables Intel Memory Protection Extensions.
872 See Documentation/x86/intel_mpx.txt for more
873 information about the feature.
875 nopku [X86] Disable Memory Protection Keys CPU feature found
879 on enable eager fpu restore
880 off disable eager fpu restore
881 auto selects the default scheme, which automatically
882 enables eagerfpu restore for xsaveopt.
884 module.async_probe [KNL]
885 Enable asynchronous probe on this module.
887 early_ioremap_debug [KNL]
888 Enable debug messages in early_ioremap support. This
889 is useful for tracking down temporary early mappings
890 which are not unmapped.
892 earlycon= [KNL] Output early console device and options.
894 When used with no options, the early console is
895 determined by the stdout-path property in device
898 cdns,<addr>[,options]
899 Start an early, polled-mode console on a Cadence
900 (xuartps) serial port at the specified address. Only
901 supported option is baud rate. If baud rate is not
902 specified, the serial port must already be setup and
905 uart[8250],io,<addr>[,options]
906 uart[8250],mmio,<addr>[,options]
907 uart[8250],mmio32,<addr>[,options]
908 uart[8250],mmio32be,<addr>[,options]
909 uart[8250],0x<addr>[,options]
910 Start an early, polled-mode console on the 8250/16550
911 UART at the specified I/O port or MMIO address.
912 MMIO inter-register address stride is either 8-bit
913 (mmio) or 32-bit (mmio32 or mmio32be).
914 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
915 to be equivalent to 'mmio'. 'options' are specified
916 in the same format described for "console=ttyS<n>"; if
917 unspecified, the h/w is not initialized.
921 Start an early, polled-mode console on a pl011 serial
922 port at the specified address. The pl011 serial port
923 must already be setup and configured. Options are not
924 yet supported. If 'mmio32' is specified, then only
925 the driver will use only 32-bit accessors to read/write
926 the device registers.
929 Start an early, polled-mode console on a meson serial
930 port at the specified address. The serial port must
931 already be setup and configured. Options are not yet
935 Start an early, polled-mode console on an msm serial
936 port at the specified address. The serial port
937 must already be setup and configured. Options are not
941 Start an early, polled-mode console on an msm serial
942 dm port at the specified address. The serial port
943 must already be setup and configured. Options are not
946 smh Use ARM semihosting calls for early console.
954 Use early console provided by serial driver available
955 on Samsung SoCs, requires selecting proper type and
956 a correct base address of the selected UART port. The
957 serial port must already be setup and configured.
958 Options are not yet supported.
962 Use early console provided by Freescale LP UART driver
963 found on Freescale Vybrid and QorIQ LS1021A processors.
964 A valid base address must be provided, and the serial
965 port must already be setup and configured.
967 armada3700_uart,<addr>
968 Start an early, polled-mode console on the
969 Armada 3700 serial port at the specified
970 address. The serial port must already be setup
971 and configured. Options are not yet supported.
973 earlyprintk= [X86,SH,BLACKFIN,ARM,M68k]
977 earlyprintk=serial[,ttySn[,baudrate]]
978 earlyprintk=serial[,0x...[,baudrate]]
979 earlyprintk=ttySn[,baudrate]
980 earlyprintk=dbgp[debugController#]
981 earlyprintk=pciserial,bus:device.function[,baudrate]
983 earlyprintk is useful when the kernel crashes before
984 the normal console is initialized. It is not enabled by
985 default because it has some cosmetic problems.
987 Append ",keep" to not disable it when the real console
990 Only one of vga, efi, serial, or usb debug port can
993 Currently only ttyS0 and ttyS1 may be specified by
994 name. Other I/O ports may be explicitly specified
995 on some architectures (x86 and arm at least) by
996 replacing ttySn with an I/O port address, like this:
997 earlyprintk=serial,0x1008,115200
998 You can find the port for a given device in
999 /proc/tty/driver/serial:
1000 2: uart:ST16650V2 port:00001008 irq:18 ...
1002 Interaction with the standard serial driver is not
1005 The VGA and EFI output is eventually overwritten by
1008 The xen output can only be used by Xen PV guests.
1010 edac_report= [HW,EDAC] Control how to report EDAC event
1011 Format: {"on" | "off" | "force"}
1012 on: enable EDAC to report H/W event. May be overridden
1013 by other higher priority error reporting module.
1014 off: disable H/W event reporting through EDAC.
1015 force: enforce the use of EDAC to report H/W event.
1018 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1021 This is designed to be used in conjunction with
1022 the boot argument: earlyprintk=vga
1025 Format: {"off" | "on" | "skip[mbr]"}
1028 Format: { "old_map", "nochunk", "noruntime", "debug" }
1029 old_map [X86-64]: switch to the old ioremap-based EFI
1030 runtime services mapping. 32-bit still uses this one by
1032 nochunk: disable reading files in "chunks" in the EFI
1033 boot stub, as chunking can cause problems with some
1034 firmware implementations.
1035 noruntime : disable EFI runtime services support
1036 debug: enable misc debug output
1038 efi_no_storage_paranoia [EFI; X86]
1039 Using this parameter you can use more than 50% of
1040 your efi variable storage. Use this parameter only if
1041 you are really sure that your UEFI does sane gc and
1042 fulfills the spec otherwise your board may brick.
1044 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1045 Add arbitrary attribute to specific memory range by
1046 updating original EFI memory map.
1047 Region of memory which aa attribute is added to is
1049 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1050 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1051 attribute is added to range 0x100000000-0x180000000 and
1052 0x10a0000000-0x1120000000.
1054 Using this parameter you can do debugging of EFI memmap
1055 related feature. For example, you can do debugging of
1056 Address Range Mirroring feature even if your box
1059 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1060 that is to be dynamically loaded by Linux. If there are
1061 multiple variables with the same name but with different
1062 vendor GUIDs, all of them will be loaded. See
1063 Documentation/acpi/ssdt-overlays.txt for details.
1066 eisa_irq_edge= [PARISC,HW]
1067 See header of drivers/parisc/eisa.c.
1070 See comment before function elanfreq_setup() in
1071 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1074 Format: {"cfq" | "deadline" | "noop"}
1075 See Documentation/block/cfq-iosched.txt and
1076 Documentation/block/deadline-iosched.txt for details.
1078 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1079 Specifies physical address of start of kernel core
1080 image elf header and optionally the size. Generally
1081 kexec loader will pass this option to capture kernel.
1082 See Documentation/kdump/kdump.txt for details.
1084 enable_mtrr_cleanup [X86]
1085 The kernel tries to adjust MTRR layout from continuous
1086 to discrete, to make X server driver able to add WB
1087 entry later. This parameter enables that.
1089 enable_timer_pin_1 [X86]
1090 Enable PIN 1 of APIC timer
1091 Can be useful to work around chipset bugs
1092 (in particular on some ATI chipsets).
1093 The kernel tries to set a reasonable default.
1095 enforcing [SELINUX] Set initial enforcing status.
1097 See security/selinux/Kconfig help text.
1098 0 -- permissive (log only, no denials).
1099 1 -- enforcing (deny and log).
1101 Value can be changed at runtime via /selinux/enforce.
1104 Disable Error Record Serialization Table (ERST)
1107 ether= [HW,NET] Ethernet cards parameters
1108 This option is obsoleted by the "netdev=" option, which
1109 has equivalent usage. See its documentation for details.
1113 Permit 'security.evm' to be updated regardless of
1114 current integrity status.
1118 fail_make_request=[KNL]
1119 General fault injection mechanism.
1120 Format: <interval>,<probability>,<space>,<times>
1121 See also Documentation/fault-injection/.
1124 See Documentation/blockdev/floppy.txt.
1126 force_pal_cache_flush
1127 [IA-64] Avoid check_sal_cache_flush which may hang on
1128 buggy SAL_CACHE_FLUSH implementations. Using this
1129 parameter will force ia64_sal_cache_flush to call
1130 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1133 Forcefully enable Physical Address Extension (PAE).
1134 Many Pentium M systems disable PAE but may have a
1135 functionally usable PAE implementation.
1136 Warning: use of this parameter will taint the kernel
1137 and may cause unknown problems.
1140 [FTRACE] will set and start the specified tracer
1141 as early as possible in order to facilitate early
1144 ftrace_dump_on_oops[=orig_cpu]
1145 [FTRACE] will dump the trace buffers on oops.
1146 If no parameter is passed, ftrace will dump
1147 buffers of all CPUs, but if you pass orig_cpu, it will
1148 dump only the buffer of the CPU that triggered the
1151 ftrace_filter=[function-list]
1152 [FTRACE] Limit the functions traced by the function
1153 tracer at boot up. function-list is a comma separated
1154 list of functions. This list can be changed at run
1155 time by the set_ftrace_filter file in the debugfs
1158 ftrace_notrace=[function-list]
1159 [FTRACE] Do not trace the functions specified in
1160 function-list. This list can be changed at run time
1161 by the set_ftrace_notrace file in the debugfs
1164 ftrace_graph_filter=[function-list]
1165 [FTRACE] Limit the top level callers functions traced
1166 by the function graph tracer at boot up.
1167 function-list is a comma separated list of functions
1168 that can be changed at run time by the
1169 set_graph_function file in the debugfs tracing directory.
1171 ftrace_graph_notrace=[function-list]
1172 [FTRACE] Do not trace from the functions specified in
1173 function-list. This list is a comma separated list of
1174 functions that can be changed at run time by the
1175 set_graph_notrace file in the debugfs tracing directory.
1178 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1179 support via parallel port (up to 5 devices per port)
1180 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1181 See also Documentation/input/joystick-parport.txt
1185 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1189 gcov_persist= [GCOV] When non-zero (default), profiling data for
1190 kernel modules is saved and remains accessible via
1191 debugfs, even when the module is unloaded/reloaded.
1192 When zero, profiling data is discarded and associated
1193 debugfs files are removed at module unload time.
1195 gpt [EFI] Forces disk with valid GPT signature but
1196 invalid Protective MBR to be treated as GPT. If the
1197 primary GPT is corrupted, it enables the backup/alternate
1198 GPT to be used instead.
1200 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1201 the "Enable 0" bit of the configuration register.
1204 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1205 the "Enable 0" bit of the configuration register.
1208 grcan.select= [HW] Select which physical interface to use.
1211 grcan.txsize= [HW] Sets the size of the tx buffer.
1212 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1214 grcan.rxsize= [HW] Sets the size of the rx buffer.
1215 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1218 gpio-mockup.gpio_mockup_ranges
1219 [HW] Sets the ranges of gpiochip of for this device.
1220 Format: <start1>,<end1>,<start2>,<end2>...
1222 hardlockup_all_cpu_backtrace=
1223 [KNL] Should the hard-lockup detector generate
1224 backtraces on all cpus.
1227 hashdist= [KNL,NUMA] Large hashes allocated during boot
1228 are distributed across NUMA nodes. Defaults on
1229 for 64-bit NUMA, off otherwise.
1230 Format: 0 | 1 (for off | on)
1232 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1234 hd= [EIDE] (E)IDE hard drive subsystem geometry
1235 Format: <cyl>,<head>,<sect>
1238 Disable Hardware Error Source Table (HEST) support;
1239 corresponding firmware-first mode error processing
1240 logic will be disabled.
1242 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1243 size of <nn>. This works even on boxes that have no
1244 highmem otherwise. This also works to reduce highmem
1245 size on bigger boxes.
1247 highres= [KNL] Enable/disable high resolution timer mode.
1248 Valid parameters: "on", "off"
1252 See Documentation/isdn/README.HiSax.
1256 hpet= [X86-32,HPET] option to control HPET usage
1257 Format: { enable (default) | disable | force |
1259 disable: disable HPET and use PIT instead
1260 force: allow force enabled of undocumented chips (ICH4,
1262 verbose: show contents of HPET registers during setup
1264 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1265 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1267 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1268 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1269 On x86-64 and powerpc, this option can be specified
1270 multiple times interleaved with hugepages= to reserve
1271 huge pages of different sizes. Valid pages sizes on
1272 x86-64 are 2M (when the CPU supports "pse") and 1G
1273 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1275 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1276 terminal devices. Valid values: 0..8
1277 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1278 If specified, z/VM IUCV HVC accepts connections
1279 from listed z/VM user IDs only.
1281 hwthread_map= [METAG] Comma-separated list of Linux cpu id to
1282 hardware thread id mappings.
1283 Format: <cpu>:<hwthread>
1286 Do not unregister boot console at start. This is only
1287 useful for debugging when something happens in the window
1288 between unregistering the boot console and initializing
1291 i2c_bus= [HW] Override the default board specific I2C bus speed
1292 or register an additional I2C bus that is not
1293 registered from board initialization code.
1297 i8042.debug [HW] Toggle i8042 debug mode
1298 i8042.unmask_kbd_data
1299 [HW] Enable printing of interrupt data from the KBD port
1300 (disabled by default, and as a pre-condition
1301 requires that i8042.debug=1 be enabled)
1302 i8042.direct [HW] Put keyboard port into non-translated mode
1303 i8042.dumbkbd [HW] Pretend that controller can only read data from
1304 keyboard and cannot control its state
1305 (Don't attempt to blink the leds)
1306 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1307 i8042.nokbd [HW] Don't check/create keyboard port
1308 i8042.noloop [HW] Disable the AUX Loopback command while probing
1310 i8042.nomux [HW] Don't check presence of an active multiplexing
1312 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1314 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1315 i8042.reset [HW] Reset the controller during init, cleanup and
1316 suspend-to-ram transitions, only during s2r
1317 transitions, or never reset
1318 Format: { 1 | Y | y | 0 | N | n }
1319 1, Y, y: always reset controller
1320 0, N, n: don't ever reset controller
1321 Default: only on s2r transitions on x86; most other
1322 architectures force reset to be always executed
1323 i8042.unlock [HW] Unlock (ignore) the keylock
1324 i8042.kbdreset [HW] Reset device connected to KBD port
1328 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1329 indicates that the driver is running on unsupported
1331 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1332 does not match list of supported models.
1334 [HW] Report power status in /proc/i8k
1335 (disabled by default)
1336 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1339 i915.invert_brightness=
1340 [DRM] Invert the sense of the variable that is used to
1341 set the brightness of the panel backlight. Normally a
1342 brightness value of 0 indicates backlight switched off,
1343 and the maximum of the brightness value sets the backlight
1344 to maximum brightness. If this parameter is set to 0
1345 (default) and the machine requires it, or this parameter
1346 is set to 1, a brightness value of 0 sets the backlight
1347 to maximum brightness, and the maximum of the brightness
1348 value switches the backlight off.
1349 -1 -- never invert brightness
1350 0 -- machine default
1351 1 -- force brightness inversion
1354 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1356 ide-core.nodma= [HW] (E)IDE subsystem
1357 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1358 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1359 .cdrom .chs .ignore_cable are additional options
1360 See Documentation/ide/ide.txt.
1362 ide-generic.probe-mask= [HW] (E)IDE subsystem
1364 Probe mask for legacy ISA IDE ports. Depending on
1365 platform up to 6 ports are supported, enabled by
1366 setting corresponding bits in the mask to 1. The
1367 default value is 0x0, which has a special meaning.
1368 On systems that have PCI, it triggers scanning the
1369 PCI bus for the first and the second port, which
1370 are then probed. On systems without PCI the value
1371 of 0x0 enables probing the two first ports as if it
1374 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1375 Claim all unknown PCI IDE storage controllers.
1378 Format: idle=poll, idle=halt, idle=nomwait
1379 Poll forces a polling idle loop that can slightly
1380 improve the performance of waking up a idle CPU, but
1381 will use a lot of power and make the system run hot.
1383 idle=halt: Halt is forced to be used for CPU idle.
1384 In such case C2/C3 won't be used again.
1385 idle=nomwait: Disable mwait for CPU C-states
1387 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1388 Format: { strict | legacy | 2008 | relaxed }
1391 Choose which programs will be accepted for execution
1392 based on the IEEE 754 NaN encoding(s) supported by
1393 the FPU and the NaN encoding requested with the value
1394 of an ELF file header flag individually set by each
1395 binary. Hardware implementations are permitted to
1396 support either or both of the legacy and the 2008 NaN
1399 Available settings are as follows:
1400 strict accept binaries that request a NaN encoding
1401 supported by the FPU
1402 legacy only accept legacy-NaN binaries, if supported
1404 2008 only accept 2008-NaN binaries, if supported
1406 relaxed accept any binaries regardless of whether
1407 supported by the FPU
1409 The FPU emulator is always able to support both NaN
1410 encodings, so if no FPU hardware is present or it has
1411 been disabled with 'nofpu', then the settings of
1412 'legacy' and '2008' strap the emulator accordingly,
1413 'relaxed' straps the emulator for both legacy-NaN and
1414 2008-NaN, whereas 'strict' enables legacy-NaN only on
1415 legacy processors and both NaN encodings on MIPS32 or
1418 The setting for ABS.fmt/NEG.fmt instruction execution
1419 mode generally follows that for the NaN encoding,
1420 except where unsupported by hardware.
1422 ignore_loglevel [KNL]
1423 Ignore loglevel setting - this will print /all/
1424 kernel messages to the console. Useful for debugging.
1425 We also add it as printk module parameter, so users
1426 could change it dynamically, usually by
1427 /sys/module/printk/parameters/ignore_loglevel.
1430 Ignore RLIMIT_DATA setting for data mappings,
1431 print warning at first misuse. Can be changed via
1432 /sys/module/kernel/parameters/ignore_rlimit_data.
1434 ihash_entries= [KNL]
1435 Set number of hash buckets for inode cache.
1437 ima_appraise= [IMA] appraise integrity measurements
1438 Format: { "off" | "enforce" | "fix" | "log" }
1441 ima_appraise_tcb [IMA]
1442 The builtin appraise policy appraises all files
1446 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1450 The list of supported hash algorithms is defined
1451 in crypto/hash_info.h.
1454 The builtin measurement policy to load during IMA
1455 setup. Specyfing "tcb" as the value, measures all
1456 programs exec'd, files mmap'd for exec, and all files
1457 opened with the read mode bit set by either the
1458 effective uid (euid=0) or uid=0.
1461 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1462 Load a policy which meets the needs of the Trusted
1463 Computing Base. This means IMA will measure all
1464 programs exec'd, files mmap'd for exec, and all files
1465 opened for read by uid=0.
1468 Select one of defined IMA measurements template formats.
1469 Formats: { "ima" | "ima-ng" | "ima-sig" }
1473 [IMA] Define a custom template format.
1474 Format: { "field1|...|fieldN" }
1476 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1477 Format: <min_file_size>
1478 Set the minimal file size for using asynchronous hash.
1479 If left unspecified, ahash usage is disabled.
1481 ahash performance varies for different data sizes on
1482 different crypto accelerators. This option can be used
1483 to achieve the best performance for a particular HW.
1485 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1487 Set hashing buffer size. Default: 4k.
1489 ahash performance varies for different chunk sizes on
1490 different crypto accelerators. This option can be used
1491 to achieve best performance for particular HW.
1495 Run specified binary instead of /sbin/init as init
1498 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1499 for working out where the kernel is dying during
1502 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1503 initcall functions. Useful for debugging built-in
1504 modules and initcalls.
1506 initrd= [BOOT] Specify the location of the initial ramdisk
1508 init_pkru= [x86] Specify the default memory protection keys rights
1509 register contents for all processes. 0x55555554 by
1510 default (disallow access to all but pkey 0). Can
1511 override in debugfs after boot.
1513 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1516 int_pln_enable [x86] Enable power limit notification interrupt
1518 integrity_audit=[IMA]
1519 Format: { "0" | "1" }
1520 0 -- basic integrity auditing messages. (Default)
1521 1 -- additional integrity auditing messages.
1523 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1525 Enable intel iommu driver.
1527 Disable intel iommu driver.
1528 igfx_off [Default Off]
1529 By default, gfx is mapped as normal device. If a gfx
1530 device has a dedicated DMAR unit, the DMAR unit is
1531 bypassed by not enabling DMAR with this option. In
1532 this case, gfx device will use physical address for
1535 With this option iommu will not optimize to look
1536 for io virtual address below 32-bit forcing dual
1537 address cycle on pci bus for cards supporting greater
1538 than 32-bit addressing. The default is to look
1539 for translation below 32-bit and if not available
1540 then look in the higher range.
1541 strict [Default Off]
1542 With this option on every unmap_single operation will
1543 result in a hardware IOTLB flush operation as opposed
1544 to batching them for performance.
1545 sp_off [Default Off]
1546 By default, super page will be supported if Intel IOMMU
1547 has the capability. With this option, super page will
1549 ecs_off [Default Off]
1550 By default, extended context tables will be supported if
1551 the hardware advertises that it has support both for the
1552 extended tables themselves, and also PASID support. With
1553 this option set, extended tables will not be used even
1554 on hardware which claims to support them.
1556 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1557 0 disables intel_idle and fall back on acpi_idle.
1558 1 to 9 specify maximum depth of C-state.
1562 Do not enable intel_pstate as the default
1563 scaling driver for the supported processors
1565 Enable intel_pstate on systems that prohibit it by default
1566 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1567 instead of acpi-cpufreq may disable platform features, such
1568 as thermal controls and power capping, that rely on ACPI
1569 P-States information being indicated to OSPM and therefore
1570 should be used with caution. This option does not work with
1571 processors that aren't supported by the intel_pstate driver
1572 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1574 Do not enable hardware P state control (HWP)
1577 Only load intel_pstate on systems which support
1578 hardware P state control (HWP) if available.
1580 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1581 Description Table, specifies preferred power management
1582 profile as "Enterprise Server" or "Performance Server",
1583 then this feature is turned on by default.
1585 intremap= [X86-64, Intel-IOMMU]
1586 on enable Interrupt Remapping (default)
1587 off disable Interrupt Remapping
1588 nosid disable Source ID checking
1590 BIOS x2APIC opt-out request will be ignored
1591 nopost disable Interrupt Posting
1593 iomem= Disable strict checking of access to MMIO memory
1594 strict regions from userspace.
1609 nobypass [PPC/POWERNV]
1610 Disable IOMMU bypass, using IOMMU for PCI devices.
1613 io7= [HW] IO7 for Marvel based alpha systems
1614 See comment before marvel_specify_io7 in
1615 arch/alpha/kernel/core_marvel.c.
1617 io_delay= [X86] I/O delay method
1619 Standard port 0x80 based delay
1621 Alternate port 0xed based delay (needed on some systems)
1623 Simple two microseconds delay
1628 See Documentation/filesystems/nfs/nfsroot.txt.
1630 irqaffinity= [SMP] Set the default irq affinity mask
1631 The argument is a cpu list, as described above.
1634 When an interrupt is not handled search all handlers
1635 for it. Intended to get systems with badly broken
1639 When an interrupt is not handled search all handlers
1640 for it. Also check all handlers each timer
1641 interrupt. Intended to get systems with badly broken
1645 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1647 isolcpus= [KNL,SMP] Isolate CPUs from the general scheduler.
1648 The argument is a cpu list, as described above.
1650 This option can be used to specify one or more CPUs
1651 to isolate from the general SMP balancing and scheduling
1652 algorithms. You can move a process onto or off an
1653 "isolated" CPU via the CPU affinity syscalls or cpuset.
1654 <cpu number> begins at 0 and the maximum value is
1655 "number of CPUs in system - 1".
1657 This option is the preferred way to isolate CPUs. The
1658 alternative -- manually setting the CPU mask of all
1659 tasks in the system -- can cause problems and
1660 suboptimal load balancer performance.
1664 ivrs_ioapic [HW,X86_64]
1665 Provide an override to the IOAPIC-ID<->DEVICE-ID
1666 mapping provided in the IVRS ACPI table. For
1667 example, to map IOAPIC-ID decimal 10 to
1668 PCI device 00:14.0 write the parameter as:
1669 ivrs_ioapic[10]=00:14.0
1671 ivrs_hpet [HW,X86_64]
1672 Provide an override to the HPET-ID<->DEVICE-ID
1673 mapping provided in the IVRS ACPI table. For
1674 example, to map HPET-ID decimal 0 to
1675 PCI device 00:14.0 write the parameter as:
1676 ivrs_hpet[0]=00:14.0
1678 ivrs_acpihid [HW,X86_64]
1679 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1680 mapping provided in the IVRS ACPI table. For
1681 example, to map UART-HID:UID AMD0020:0 to
1682 PCI device 00:14.5 write the parameter as:
1683 ivrs_acpihid[00:14.5]=AMD0020:0
1685 js= [HW,JOY] Analog joystick
1686 See Documentation/input/joystick.txt.
1689 When CONFIG_RANDOMIZE_BASE is set, this disables
1690 kernel and module base offset ASLR (Address Space
1691 Layout Randomization).
1695 kernelcore= [KNL,X86,IA-64,PPC]
1696 Format: nn[KMGTPE] | "mirror"
1698 specifies the amount of memory usable by the kernel
1699 for non-movable allocations. The requested amount is
1700 spread evenly throughout all nodes in the system. The
1701 remaining memory in each node is used for Movable
1702 pages. In the event, a node is too small to have both
1703 kernelcore and Movable pages, kernelcore pages will
1704 take priority and other nodes will have a larger number
1705 of Movable pages. The Movable zone is used for the
1706 allocation of pages that may be reclaimed or moved
1707 by the page migration subsystem. This means that
1708 HugeTLB pages may not be allocated from this zone.
1709 Note that allocations like PTEs-from-HighMem still
1710 use the HighMem zone if it exists, and the Normal
1711 zone if it does not.
1713 Instead of specifying the amount of memory (nn[KMGTPE]),
1714 you can specify "mirror" option. In case "mirror"
1715 option is specified, mirrored (reliable) memory is used
1716 for non-movable allocations and remaining memory is used
1717 for Movable pages. nn[KMGTPE] and "mirror" are exclusive,
1718 so you can NOT specify nn[KMGTPE] and "mirror" at the same
1721 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1722 Format: <Controller#>[,poll interval]
1723 The controller # is the number of the ehci usb debug
1724 port as it is probed via PCI. The poll interval is
1725 optional and is the number seconds in between
1726 each poll cycle to the debug port in case you need
1727 the functionality for interrupting the kernel with
1728 gdb or control-c on the dbgp connection. When
1729 not using this parameter you use sysrq-g to break into
1730 the kernel debugger.
1732 kgdboc= [KGDB,HW] kgdb over consoles.
1733 Requires a tty driver that supports console polling,
1734 or a supported polling keyboard driver (non-usb).
1735 Serial only format: <serial_device>[,baud]
1736 keyboard only format: kbd
1737 keyboard and serial format: kbd,<serial_device>[,baud]
1738 Optional Kernel mode setting:
1739 kms, kbd format: kms,kbd
1740 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1742 kgdbwait [KGDB] Stop kernel execution and enter the
1743 kernel debugger at the earliest opportunity.
1745 kmac= [MIPS] korina ethernet MAC address.
1746 Configure the RouterBoard 532 series on-chip
1747 Ethernet adapter MAC address.
1749 kmemleak= [KNL] Boot-time kmemleak enable/disable
1750 Valid arguments: on, off
1752 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1755 kmemcheck= [X86] Boot-time kmemcheck enable/disable/one-shot mode
1756 Valid arguments: 0, 1, 2
1757 kmemcheck=0 (disabled)
1758 kmemcheck=1 (enabled)
1759 kmemcheck=2 (one-shot mode)
1760 Default: 2 (one-shot mode)
1762 kstack=N [X86] Print N words from the kernel stack
1765 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1766 Default is 0 (don't ignore, but inject #GP)
1768 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1772 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1773 Default is 1 (enabled)
1775 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1777 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1779 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1780 (virtualized MMU) support on capable Intel chips.
1781 Default is 1 (enabled)
1783 kvm-intel.emulate_invalid_guest_state=
1784 [KVM,Intel] Enable emulation of invalid guest states
1785 Default is 0 (disabled)
1787 kvm-intel.flexpriority=
1788 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1789 Default is 1 (enabled)
1792 [KVM,Intel] Enable VMX nesting (nVMX).
1793 Default is 0 (disabled)
1795 kvm-intel.unrestricted_guest=
1796 [KVM,Intel] Disable unrestricted guest feature
1797 (virtualized real and unpaged mode) on capable
1798 Intel chips. Default is 1 (enabled)
1800 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1801 feature (tagged TLBs) on capable Intel chips.
1802 Default is 1 (enabled)
1808 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1811 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1812 value for LAPIC timer one-shot implementation. Default
1813 back to the programmable timer unit in the LAPIC.
1815 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1818 libata.dma= [LIBATA] DMA control
1819 libata.dma=0 Disable all PATA and SATA DMA
1820 libata.dma=1 PATA and SATA Disk DMA only
1821 libata.dma=2 ATAPI (CDROM) DMA only
1822 libata.dma=4 Compact Flash DMA only
1823 Combinations also work, so libata.dma=3 enables DMA
1824 for disks and CDROMs, but not CFs.
1826 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1827 libata.ignore_hpa=0 keep BIOS limits (default)
1828 libata.ignore_hpa=1 ignore limits, using full disk
1830 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
1834 libata.force= [LIBATA] Force configurations. The format is comma
1835 separated list of "[ID:]VAL" where ID is
1836 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
1837 matching port, link or device. Basically, it matches
1838 the ATA ID string printed on console by libata. If
1839 the whole ID part is omitted, the last PORT and DEVICE
1840 values are used. If ID hasn't been specified yet, the
1841 configuration applies to all ports, links and devices.
1843 If only DEVICE is omitted, the parameter applies to
1844 the port and all links and devices behind it. DEVICE
1845 number of 0 either selects the first device or the
1846 first fan-out link behind PMP device. It does not
1847 select the host link. DEVICE number of 15 selects the
1848 host link and device attached to it.
1850 The VAL specifies the configuration to force. As long
1851 as there's no ambiguity shortcut notation is allowed.
1852 For example, both 1.5 and 1.5G would work for 1.5Gbps.
1853 The following configurations can be forced.
1855 * Cable type: 40c, 80c, short40c, unk, ign or sata.
1856 Any ID with matching PORT is used.
1858 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
1860 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
1861 udma[/][16,25,33,44,66,100,133] notation is also
1864 * [no]ncq: Turn on or off NCQ.
1866 * [no]ncqtrim: Turn off queued DSM TRIM.
1868 * nohrst, nosrst, norst: suppress hard, soft
1871 * rstonce: only attempt one reset during
1872 hot-unplug link recovery
1874 * dump_id: dump IDENTIFY data.
1876 * atapi_dmadir: Enable ATAPI DMADIR bridge support
1878 * disable: Disable this device.
1880 If there are multiple matching configurations changing
1881 the same attribute, the last one is used.
1883 memblock=debug [KNL] Enable memblock debug messages.
1885 load_ramdisk= [RAM] List of ramdisks to load from floppy
1886 See Documentation/blockdev/ramdisk.txt.
1888 lockd.nlm_grace_period=P [NFS] Assign grace period.
1891 lockd.nlm_tcpport=N [NFS] Assign TCP port.
1894 lockd.nlm_timeout=T [NFS] Assign timeout value.
1897 lockd.nlm_udpport=M [NFS] Assign UDP port.
1900 locktorture.nreaders_stress= [KNL]
1901 Set the number of locking read-acquisition kthreads.
1902 Defaults to being automatically set based on the
1903 number of online CPUs.
1905 locktorture.nwriters_stress= [KNL]
1906 Set the number of locking write-acquisition kthreads.
1908 locktorture.onoff_holdoff= [KNL]
1909 Set time (s) after boot for CPU-hotplug testing.
1911 locktorture.onoff_interval= [KNL]
1912 Set time (s) between CPU-hotplug operations, or
1913 zero to disable CPU-hotplug testing.
1915 locktorture.shuffle_interval= [KNL]
1916 Set task-shuffle interval (jiffies). Shuffling
1917 tasks allows some CPUs to go into dyntick-idle
1918 mode during the locktorture test.
1920 locktorture.shutdown_secs= [KNL]
1921 Set time (s) after boot system shutdown. This
1922 is useful for hands-off automated testing.
1924 locktorture.stat_interval= [KNL]
1925 Time (s) between statistics printk()s.
1927 locktorture.stutter= [KNL]
1928 Time (s) to stutter testing, for example,
1929 specifying five seconds causes the test to run for
1930 five seconds, wait for five seconds, and so on.
1931 This tests the locking primitive's ability to
1932 transition abruptly to and from idle.
1934 locktorture.torture_runnable= [BOOT]
1935 Start locktorture running at boot time.
1937 locktorture.torture_type= [KNL]
1938 Specify the locking implementation to test.
1940 locktorture.verbose= [KNL]
1941 Enable additional printk() statements.
1943 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
1946 loglevel= All Kernel Messages with a loglevel smaller than the
1947 console loglevel will be printed to the console. It can
1948 also be changed with klogd or other programs. The
1949 loglevels are defined as follows:
1951 0 (KERN_EMERG) system is unusable
1952 1 (KERN_ALERT) action must be taken immediately
1953 2 (KERN_CRIT) critical conditions
1954 3 (KERN_ERR) error conditions
1955 4 (KERN_WARNING) warning conditions
1956 5 (KERN_NOTICE) normal but significant condition
1957 6 (KERN_INFO) informational
1958 7 (KERN_DEBUG) debug-level messages
1960 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
1961 in bytes. n must be a power of two and greater
1962 than the minimal size. The minimal size is defined
1963 by LOG_BUF_SHIFT kernel config parameter. There is
1964 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
1965 that allows to increase the default size depending on
1966 the number of CPUs. See init/Kconfig for more details.
1968 logo.nologo [FB] Disables display of the built-in Linux logo.
1969 This may be used to provide more screen space for
1970 kernel log messages and is useful when debugging
1971 kernel boot problems.
1973 lp=0 [LP] Specify parallel ports to use, e.g,
1974 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
1975 lp=reset first parallel port). 'lp=0' disables the
1976 lp=auto printer driver. 'lp=reset' (which can be
1977 specified in addition to the ports) causes
1978 attached printers to be reset. Using
1979 lp=port1,port2,... specifies the parallel ports
1980 to associate lp devices with, starting with
1981 lp0. A port specification may be 'none' to skip
1982 that lp device, or a parport name such as
1983 'parport0'. Specifying 'lp=auto' instead of a
1984 port specification list means that device IDs
1985 from each port should be examined, to see if
1986 an IEEE 1284-compliant printer is attached; if
1987 so, the driver will manage that printer.
1988 See also header of drivers/char/lp.c.
1991 Sets loops_per_jiffy to given constant, thus avoiding
1992 time-consuming boot-time autodetection (up to 250 ms per
1993 CPU). 0 enables autodetection (default). To determine
1994 the correct value for your kernel, boot with normal
1995 autodetection and see what value is printed. Note that
1996 on SMP systems the preset will be applied to all CPUs,
1997 which is likely to cause problems if your CPUs need
1998 significantly divergent settings. An incorrect value
1999 will cause delays in the kernel to be wrong, leading to
2000 unpredictable I/O errors and other breakage. Although
2001 unlikely, in the extreme case this might damage your
2005 Format: <io>,<irq>,<dma>
2007 machvec= [IA-64] Force the use of a particular machine-vector
2008 (machvec) in a generic kernel.
2009 Example: machvec=hpzx1_swiotlb
2011 machtype= [Loongson] Share the same kernel image file between different
2013 Example: machtype=lemote-yeeloong-2f-7inch
2015 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2016 than or equal to this physical address is ignored.
2018 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2019 will bring up during bootup. maxcpus=n : n >= 0 limits
2020 the kernel to bring up 'n' processors. Surely after
2021 bootup you can bring up the other plugged cpu by executing
2022 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2023 only takes effect during system bootup.
2024 While n=0 is a special case, it is equivalent to "nosmp",
2025 which also disables the IO APIC.
2027 max_loop= [LOOP] The number of loop block devices that get
2028 (loop.max_loop) unconditionally pre-created at init time. The default
2029 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2030 of statically allocating a predefined number, loop
2031 devices can be requested on-demand with the
2032 /dev/loop-control interface.
2034 mce [X86-32] Machine Check Exception
2036 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2038 md= [HW] RAID subsystems devices and level
2039 See Documentation/admin-guide/md.rst.
2042 Format: <first>,<last>
2043 Specifies range of consoles to be captured by the MDA.
2045 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2046 Amount of memory to be used when the kernel is not able
2047 to see the whole system memory or for test.
2048 [X86] Work as limiting max address. Use together
2049 with memmap= to avoid physical address space collisions.
2050 Without memmap= PCI devices could be placed at addresses
2051 belonging to unused RAM.
2053 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2057 [KNL,SH] Allow user to override the default size for
2058 per-device physically contiguous DMA buffers.
2060 memhp_default_state=online/offline
2061 [KNL] Set the initial state for the memory hotplug
2062 onlining policy. If not specified, the default value is
2063 set according to the
2064 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2066 See Documentation/memory-hotplug.txt.
2068 memmap=exactmap [KNL,X86] Enable setting of an exact
2069 E820 memory map, as specified by the user.
2070 Such memmap=exactmap lines can be constructed based on
2071 BIOS output or other requirements. See the memmap=nn@ss
2074 memmap=nn[KMG]@ss[KMG]
2075 [KNL] Force usage of a specific region of memory.
2076 Region of memory to be used is from ss to ss+nn.
2078 memmap=nn[KMG]#ss[KMG]
2079 [KNL,ACPI] Mark specific memory as ACPI data.
2080 Region of memory to be marked is from ss to ss+nn.
2082 memmap=nn[KMG]$ss[KMG]
2083 [KNL,ACPI] Mark specific memory as reserved.
2084 Region of memory to be reserved is from ss to ss+nn.
2085 Example: Exclude memory from 0x18690000-0x1869ffff
2086 memmap=64K$0x18690000
2088 memmap=0x10000$0x18690000
2090 memmap=nn[KMG]!ss[KMG]
2091 [KNL,X86] Mark specific memory as protected.
2092 Region of memory to be used, from ss to ss+nn.
2093 The memory region may be marked as e820 type 12 (0xc)
2094 and is NVDIMM or ADR memory.
2096 memory_corruption_check=0/1 [X86]
2097 Some BIOSes seem to corrupt the first 64k of
2098 memory when doing things like suspend/resume.
2099 Setting this option will scan the memory
2100 looking for corruption. Enabling this will
2101 both detect corruption and prevent the kernel
2102 from using the memory being corrupted.
2103 However, its intended as a diagnostic tool; if
2104 repeatable BIOS-originated corruption always
2105 affects the same memory, you can use memmap=
2106 to prevent the kernel from using that memory.
2108 memory_corruption_check_size=size [X86]
2109 By default it checks for corruption in the low
2110 64k, making this memory unavailable for normal
2111 use. Use this parameter to scan for
2112 corruption in more or less memory.
2114 memory_corruption_check_period=seconds [X86]
2115 By default it checks for corruption every 60
2116 seconds. Use this parameter to check at some
2117 other rate. 0 disables periodic checking.
2119 memtest= [KNL,X86,ARM] Enable memtest
2121 default : 0 <disable>
2122 Specifies the number of memtest passes to be
2123 performed. Each pass selects another test
2124 pattern from a given set of patterns. Memtest
2125 fills the memory with this pattern, validates
2126 memory contents and reserves bad memory
2127 regions that are detected.
2129 meye.*= [HW] Set MotionEye Camera parameters
2130 See Documentation/video4linux/meye.txt.
2132 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2133 Multi-Function General Purpose Timers on AMD Geode
2136 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2137 the BIOS has incorrectly applied a workaround. TinyBIOS
2138 version 0.98 is known to be affected, 0.99 fixes the
2139 problem by letting the user disable the workaround.
2143 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2144 physical address is ignored.
2146 mini2440= [ARM,HW,KNL]
2147 Format:[0..2][b][c][t]
2149 MINI2440 configuration specification:
2150 0 - The attached screen is the 3.5" TFT
2151 1 - The attached screen is the 7" TFT
2152 2 - The VGA Shield is attached (1024x768)
2153 Leaving out the screen size parameter will not load
2154 the TFT driver, and the framebuffer will be left
2156 b - Enable backlight. The TFT backlight pin will be
2157 linked to the kernel VESA blanking code and a GPIO
2158 LED. This parameter is not necessary when using the
2160 c - Enable the s3c camera interface.
2161 t - Reserved for enabling touchscreen support. The
2162 touchscreen support is not enabled in the mainstream
2163 kernel as of 2.6.30, a preliminary port can be found
2164 in the "bleeding edge" mini2440 support kernel at
2165 http://repo.or.cz/w/linux-2.6/mini2440.git
2168 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2169 parameter allows control of the logging verbosity for
2170 the additional memory initialisation checks. A value
2171 of 0 disables mminit logging and a level of 4 will
2172 log everything. Information is printed at KERN_DEBUG
2173 so loglevel=8 may also need to be specified.
2176 [KNL] When CONFIG_MODULE_SIG is set, this means that
2177 modules without (valid) signatures will fail to load.
2178 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2179 is always true, so this option does nothing.
2181 module_blacklist= [KNL] Do not load a comma-separated list of
2182 modules. Useful for debugging problem modules.
2185 [MOUSE] Maximum time between finger touching and
2186 leaving touchpad surface for touch to be considered
2187 a tap and be reported as a left button click (for
2188 touchpads working in absolute mode only).
2190 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2191 reporting absolute coordinates, such as tablets
2192 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2193 reporting absolute coordinates, such as tablets
2195 movablecore=nn[KMG] [KNL,X86,IA-64,PPC] This parameter
2196 is similar to kernelcore except it specifies the
2197 amount of memory used for migratable allocations.
2198 If both kernelcore and movablecore is specified,
2199 then kernelcore will be at *least* the specified
2200 value but may be more. If movablecore on its own
2201 is specified, the administrator must be careful
2202 that the amount of memory usable for all allocations
2205 movable_node [KNL,X86] Boot-time switch to enable the effects
2206 of CONFIG_MOVABLE_NODE=y. See mm/Kconfig for details.
2208 MTD_Partition= [MTD]
2209 Format: <name>,<region-number>,<size>,<offset>
2211 MTD_Region= [MTD] Format:
2212 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2215 See drivers/mtd/cmdlinepart.c.
2217 multitce=off [PPC] This parameter disables the use of the pSeries
2218 firmware feature for updating multiple TCE entries
2221 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2223 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2225 boundary - index of last SLC block on Flex-OneNAND.
2226 The remaining blocks are configured as MLC blocks.
2227 lock - Configure if Flex-OneNAND boundary should be locked.
2228 Once locked, the boundary cannot be changed.
2229 1 indicates lock status, 0 indicates unlock status.
2232 ARM/S3C2412 JIVE boot control
2234 See arch/arm/mach-s3c2412/mach-jive.c
2236 mtouchusb.raw_coordinates=
2237 [HW] Make the MicroTouch USB driver use raw coordinates
2238 ('y', default) or cooked coordinates ('n')
2240 mtrr_chunk_size=nn[KMG] [X86]
2241 used for mtrr cleanup. It is largest continuous chunk
2242 that could hold holes aka. UC entries.
2244 mtrr_gran_size=nn[KMG] [X86]
2245 Used for mtrr cleanup. It is granularity of mtrr block.
2247 Large value could prevent small alignment from
2250 mtrr_spare_reg_nr=n [X86]
2252 Range: 0,7 : spare reg number
2254 Used for mtrr cleanup. It is spare mtrr entries number.
2255 Set to 2 or more if your graphical card needs more.
2257 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2259 netdev= [NET] Network devices parameters
2260 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2261 Note that mem_start is often overloaded to mean
2262 something different and driver-specific.
2263 This usage is only documented in each driver source
2267 [NETFILTER] Enable connection tracking flow accounting
2268 0 to disable accounting
2269 1 to enable accounting
2272 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2273 See Documentation/filesystems/nfs/nfsroot.txt.
2275 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2276 See Documentation/filesystems/nfs/nfsroot.txt.
2278 nfsrootdebug [NFS] enable nfsroot debugging messages.
2279 See Documentation/filesystems/nfs/nfsroot.txt.
2281 nfs.callback_nr_threads=
2282 [NFSv4] set the total number of threads that the
2283 NFS client will assign to service NFSv4 callback
2286 nfs.callback_tcpport=
2287 [NFS] set the TCP port on which the NFSv4 callback
2288 channel should listen.
2291 [NFS] sets the pathname to the program which is used
2292 to update the NFS client cache entries.
2294 nfs.cache_getent_timeout=
2295 [NFS] sets the timeout after which an attempt to
2296 update a cache entry is deemed to have failed.
2298 nfs.idmap_cache_timeout=
2299 [NFS] set the maximum lifetime for idmapper cache
2303 [NFS] enable 64-bit inode numbers.
2304 If zero, the NFS client will fake up a 32-bit inode
2305 number for the readdir() and stat() syscalls instead
2306 of returning the full 64-bit number.
2307 The default is to return 64-bit inode numbers.
2309 nfs.max_session_cb_slots=
2310 [NFSv4.1] Sets the maximum number of session
2311 slots the client will assign to the callback
2312 channel. This determines the maximum number of
2313 callbacks the client will process in parallel for
2314 a particular server.
2316 nfs.max_session_slots=
2317 [NFSv4.1] Sets the maximum number of session slots
2318 the client will attempt to negotiate with the server.
2319 This limits the number of simultaneous RPC requests
2320 that the client can send to the NFSv4.1 server.
2321 Note that there is little point in setting this
2322 value higher than the max_tcp_slot_table_limit.
2324 nfs.nfs4_disable_idmapping=
2325 [NFSv4] When set to the default of '1', this option
2326 ensures that both the RPC level authentication
2327 scheme and the NFS level operations agree to use
2328 numeric uids/gids if the mount is using the
2329 'sec=sys' security flavour. In effect it is
2330 disabling idmapping, which can make migration from
2331 legacy NFSv2/v3 systems to NFSv4 easier.
2332 Servers that do not support this mode of operation
2333 will be autodetected by the client, and it will fall
2334 back to using the idmapper.
2335 To turn off this behaviour, set the value to '0'.
2337 [NFS4] Specify an additional fixed unique ident-
2338 ification string that NFSv4 clients can insert into
2339 their nfs_client_id4 string. This is typically a
2340 UUID that is generated at system install time.
2342 nfs.send_implementation_id =
2343 [NFSv4.1] Send client implementation identification
2344 information in exchange_id requests.
2345 If zero, no implementation identification information
2347 The default is to send the implementation identification
2350 nfs.recover_lost_locks =
2351 [NFSv4] Attempt to recover locks that were lost due
2352 to a lease timeout on the server. Please note that
2353 doing this risks data corruption, since there are
2354 no guarantees that the file will remain unchanged
2355 after the locks are lost.
2356 If you want to enable the kernel legacy behaviour of
2357 attempting to recover these locks, then set this
2359 The default parameter value of '0' causes the kernel
2360 not to attempt recovery of lost locks.
2362 nfs4.layoutstats_timer =
2363 [NFSv4.2] Change the rate at which the kernel sends
2364 layoutstats to the pNFS metadata server.
2366 Setting this to value to 0 causes the kernel to use
2367 whatever value is the default set by the layout
2368 driver. A non-zero value sets the minimum interval
2369 in seconds between layoutstats transmissions.
2371 nfsd.nfs4_disable_idmapping=
2372 [NFSv4] When set to the default of '1', the NFSv4
2373 server will return only numeric uids and gids to
2374 clients using auth_sys, and will accept numeric uids
2375 and gids from such clients. This is intended to ease
2376 migration from NFSv2/v3.
2378 objlayoutdriver.osd_login_prog=
2379 [NFS] [OBJLAYOUT] sets the pathname to the program which
2380 is used to automatically discover and login into new
2381 osd-targets. Please see:
2382 Documentation/filesystems/pnfs.txt for more explanations
2384 nmi_debug= [KNL,AVR32,SH] Specify one or more actions to take
2385 when a NMI is triggered.
2386 Format: [state][,regs][,debounce][,die]
2388 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2389 Format: [panic,][nopanic,][num]
2391 0 - turn hardlockup detector in nmi_watchdog off
2392 1 - turn hardlockup detector in nmi_watchdog on
2393 When panic is specified, panic when an NMI watchdog
2394 timeout occurs (or 'nopanic' to override the opposite
2395 default). To disable both hard and soft lockup detectors,
2396 please see 'nowatchdog'.
2397 This is useful when you use a panic=... timeout and
2398 need the box quickly up again.
2400 netpoll.carrier_timeout=
2401 [NET] Specifies amount of time (in seconds) that
2402 netpoll should wait for a carrier. By default netpoll
2405 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2406 emulation library even if a 387 maths coprocessor
2410 [HW] Never suspend the console
2411 Disable suspending of consoles during suspend and
2412 hibernate operations. Once disabled, debugging
2413 messages can reach various consoles while the rest
2414 of the system is being put to sleep (ie, while
2415 debugging driver suspend/resume hooks). This may
2416 not work reliably with all consoles, but is known
2417 to work with serial and VGA consoles.
2418 To facilitate more flexible debugging, we also add
2419 console_suspend, a printk module parameter to control
2420 it. Users could use console_suspend (usually
2421 /sys/module/printk/parameters/console_suspend) to
2422 turn on/off it dynamically.
2424 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2425 caches in the slab allocator. Saves per-node memory,
2426 but will impact performance.
2430 noapic [SMP,APIC] Tells the kernel to not make use of any
2431 IOAPICs that may be present in the system.
2433 noautogroup Disable scheduler automatic task group creation.
2435 nobats [PPC] Do not use BATs for mapping kernel lowmem
2436 on "Classic" PPC cores.
2440 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2442 nodelayacct [KNL] Disable per-task delay accounting
2444 nodsp [SH] Disable hardware DSP at boot time.
2446 noefi Disable EFI runtime services support.
2451 On X86-32 available only on PAE configured kernels.
2452 noexec=on: enable non-executable mappings (default)
2453 noexec=off: disable non-executable mappings
2456 Disable SMAP (Supervisor Mode Access Prevention)
2457 even if it is supported by processor.
2460 Disable SMEP (Supervisor Mode Execution Prevention)
2461 even if it is supported by processor.
2464 This affects only 32-bit executables.
2465 noexec32=on: enable non-executable mappings (default)
2466 read doesn't imply executable mappings
2467 noexec32=off: disable non-executable mappings
2468 read implies executable mappings
2470 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2472 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2473 register save and restore. The kernel will only save
2474 legacy floating-point registers on task switch.
2476 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2478 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2479 Equivalent to smt=1.
2481 noxsave [BUGS=X86] Disables x86 extended register state save
2482 and restore using xsave. The kernel will fallback to
2483 enabling legacy floating-point and sse state.
2485 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2486 register states. The kernel will fall back to use
2487 xsave to save the states. By using this parameter,
2488 performance of saving the states is degraded because
2489 xsave doesn't support modified optimization while
2490 xsaveopt supports it on xsaveopt enabled systems.
2492 noxsaves [X86] Disables xsaves and xrstors used in saving and
2493 restoring x86 extended register state in compacted
2494 form of xsave area. The kernel will fall back to use
2495 xsaveopt and xrstor to save and restore the states
2496 in standard form of xsave area. By using this
2497 parameter, xsave area per process might occupy more
2498 memory on xsaves enabled systems.
2500 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2501 wfi(ARM) instruction doesn't work correctly and not to
2502 use it. This is also useful when using JTAG debugger.
2504 no_file_caps Tells the kernel not to honor file capabilities. The
2505 only way then for a file to be executed with privilege
2506 is to be setuid root or executed by root.
2508 nohalt [IA-64] Tells the kernel not to use the power saving
2509 function PAL_HALT_LIGHT when idle. This increases
2510 power-consumption. On the positive side, it reduces
2511 interrupt wake-up latency, which may improve performance
2512 in certain environments such as networked servers or
2515 nohibernate [HIBERNATION] Disable hibernation and resume.
2517 nohz= [KNL] Boottime enable/disable dynamic ticks
2518 Valid arguments: on, off
2521 nohz_full= [KNL,BOOT]
2522 The argument is a cpu list, as described above.
2523 In kernels built with CONFIG_NO_HZ_FULL=y, set
2524 the specified list of CPUs whose tick will be stopped
2525 whenever possible. The boot CPU will be forced outside
2526 the range to maintain the timekeeping.
2527 The CPUs in this range must also be included in the
2530 noiotrap [SH] Disables trapped I/O port accesses.
2532 noirqdebug [X86-32] Disables the code which attempts to detect and
2533 disable unhandled interrupt sources.
2535 no_timer_check [X86,APIC] Disables the code which tests for
2536 broken timer IRQ sources.
2538 noisapnp [ISAPNP] Disables ISA PnP code.
2540 noinitrd [RAM] Tells the kernel not to load any configured
2543 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2545 [Deprecated - use intremap=off]
2549 noinvpcid [X86] Disable the INVPCID cpu feature.
2551 nojitter [IA-64] Disables jitter checking for ITC timers.
2553 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2555 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2558 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2559 steal time is computed, but won't influence scheduler
2562 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2564 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2566 noltlbs [PPC] Do not use large page/tlb entries for kernel
2567 lowmem mapping on PPC40x and PPC8xx
2569 nomca [IA-64] Disable machine check abort handling
2571 nomce [X86-32] Disable Machine Check Exception
2573 nomfgpt [X86-32] Disable Multi-Function General Purpose
2574 Timer usage (for AMD Geode machines).
2576 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2577 shutdown the other cpus. Instead use the REBOOT_VECTOR
2580 nomodule Disable module load
2582 nopat [X86] Disable PAT (page attribute table extension of
2583 pagetables) support.
2585 norandmaps Don't use address space randomization. Equivalent to
2586 echo 0 > /proc/sys/kernel/randomize_va_space
2588 noreplace-paravirt [X86,IA-64,PV_OPS] Don't patch paravirt_ops
2590 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2591 with UP alternatives
2593 nordrand [X86] Disable kernel use of the RDRAND and
2594 RDSEED instructions even if they are supported
2595 by the processor. RDRAND and RDSEED are still
2596 available to user space applications.
2598 noresume [SWSUSP] Disables resume and restores original swap
2601 no-scroll [VGA] Disables scrollback.
2602 This is required for the Braillex ib80-piezo Braille
2603 reader made by F.H. Papenmeier (Germany).
2607 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2609 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2610 and disable the IO APIC. legacy for "maxcpus=0".
2612 nosoftlockup [KNL] Disable the soft-lockup detector.
2614 nosync [HW,M68K] Disables sync negotiation for all devices.
2616 notsc [BUGS=X86-32] Disable Time Stamp Counter
2618 nowatchdog [KNL] Disable both lockup detectors, i.e.
2619 soft-lockup and NMI watchdog (hard-lockup).
2623 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2625 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2626 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2627 Some features depend on CPU0. Known dependencies are:
2628 1. Resume from suspend/hibernate depends on CPU0.
2629 Suspend/hibernate will fail if CPU0 is offline and you
2630 need to online CPU0 before suspend/hibernate.
2631 2. PIC interrupts also depend on CPU0. CPU0 can't be
2632 removed if a PIC interrupt is detected.
2633 It's said poweroff/reboot may depend on CPU0 on some
2634 machines although I haven't seen such issues so far
2635 after CPU0 is offline on a few tested machines.
2636 If the dependencies are under your control, you can
2637 turn on cpu0_hotplug.
2639 nptcg= [IA-64] Override max number of concurrent global TLB
2640 purges which is reported from either PAL_VM_SUMMARY or
2643 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2644 could support. nr_cpus=n : n >= 1 limits the kernel to
2645 support 'n' processors. It could be larger than the
2646 number of already plugged CPU during bootup, later in
2647 runtime you can physically add extra cpu until it reaches
2648 n. So during boot up some boot time memory for per-cpu
2649 variables need be pre-allocated for later physical cpu
2652 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2654 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2655 Allowed values are enable and disable
2657 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2658 one of ['zone', 'node', 'default'] can be specified
2659 This can be set from sysctl after boot.
2660 See Documentation/sysctl/vm.txt for details.
2662 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2663 See Documentation/debugging-via-ohci1394.txt for more
2666 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2667 Rather than timing out after 20 ms if an EC
2668 command is not properly ACKed, override the length
2669 of the timeout. We have interrupts disabled while
2670 waiting for the ACK, so if this is set too high
2671 interrupts *may* be lost!
2673 omap_mux= [OMAP] Override bootloader pin multiplexing.
2674 Format: <mux_mode0.mode_name=value>...
2675 For example, to override I2C bus2:
2676 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2678 oprofile.timer= [HW]
2679 Use timer interrupt instead of performance counters
2681 oprofile.cpu_type= Force an oprofile cpu type
2682 This might be useful if you have an older oprofile
2683 userland or if you want common events.
2684 Format: { arch_perfmon }
2685 arch_perfmon: [X86] Force use of architectural
2686 perfmon on Intel CPUs instead of the
2687 CPU specific event set.
2688 timer: [X86] Force use of architectural NMI
2689 timer mode (see also oprofile.timer
2690 for generic hr timer mode)
2692 oops=panic Always panic on oopses. Default is to just kill the
2693 process, but there is a small probability of
2694 deadlocking the machine.
2695 This will also cause panics on machine check exceptions.
2696 Useful together with panic=30 to trigger a reboot.
2699 See Documentation/sound/oss/oss-parameters.txt
2701 page_owner= [KNL] Boot-time page_owner enabling option.
2702 Storage of the information about who allocated
2703 each page is disabled in default. With this switch,
2705 on: enable the feature
2707 page_poison= [KNL] Boot-time parameter changing the state of
2708 poisoning on the buddy allocator.
2709 off: turn off poisoning
2710 on: turn on poisoning
2712 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2713 timeout > 0: seconds before rebooting
2714 timeout = 0: wait forever
2715 timeout < 0: reboot immediately
2718 panic_on_warn panic() instead of WARN(). Useful to cause kdump
2721 crash_kexec_post_notifiers
2722 Run kdump after running panic-notifiers and dumping
2723 kmsg. This only for the users who doubt kdump always
2724 succeeds in any situation.
2725 Note that this also increases risks of kdump failure,
2726 because some panic notifiers can make the crashed
2727 kernel more unstable.
2729 parkbd.port= [HW] Parallel port number the keyboard adapter is
2730 connected to, default is 0.
2732 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2733 0 for XT, 1 for AT (default is AT).
2736 parport= [HW,PPT] Specify parallel ports. 0 disables.
2737 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2738 Use 'auto' to force the driver to use any
2739 IRQ/DMA settings detected (the default is to
2740 ignore detected IRQ/DMA settings because of
2741 possible conflicts). You can specify the base
2742 address, IRQ, and DMA settings; IRQ and DMA
2743 should be numbers, or 'auto' (for using detected
2744 settings on that particular port), or 'nofifo'
2745 (to avoid using a FIFO even if it is detected).
2746 Parallel ports are assigned in the order they
2747 are specified on the command line, starting
2750 parport_init_mode= [HW,PPT]
2751 Configure VIA parallel port to operate in
2752 a specific mode. This is necessary on Pegasos
2753 computer where firmware has no options for setting
2754 up parallel port mode and sets it to spp.
2755 Currently this function knows 686a and 8231 chips.
2756 Format: [spp|ps2|epp|ecp|ecpepp]
2759 Halt all CPUs after the first oops has been printed for
2760 the specified number of seconds. This is to be used if
2761 your oopses keep scrolling off the screen.
2766 See header of drivers/block/paride/pcd.c.
2767 See also Documentation/blockdev/paride.txt.
2769 pci=option[,option...] [PCI] various PCI subsystem options:
2770 earlydump [X86] dump PCI config space before the kernel
2772 off [X86] don't probe for the PCI bus
2773 bios [X86-32] force use of PCI BIOS, don't access
2774 the hardware directly. Use this if your machine
2775 has a non-standard PCI host bridge.
2776 nobios [X86-32] disallow use of PCI BIOS, only direct
2777 hardware access methods are allowed. Use this
2778 if you experience crashes upon bootup and you
2779 suspect they are caused by the BIOS.
2780 conf1 [X86] Force use of PCI Configuration Access
2781 Mechanism 1 (config address in IO port 0xCF8,
2782 data in IO port 0xCFC, both 32-bit).
2783 conf2 [X86] Force use of PCI Configuration Access
2784 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
2785 the function, IO port 0xCFA, also 8-bit, sets
2786 bus number. The config space is then accessed
2787 through ports 0xC000-0xCFFF).
2788 See http://wiki.osdev.org/PCI for more info
2789 on the configuration access mechanisms.
2790 noaer [PCIE] If the PCIEAER kernel config parameter is
2791 enabled, this kernel boot option can be used to
2792 disable the use of PCIE advanced error reporting.
2793 nodomains [PCI] Disable support for multiple PCI
2794 root domains (aka PCI segments, in ACPI-speak).
2795 nommconf [X86] Disable use of MMCONFIG for PCI
2797 check_enable_amd_mmconf [X86] check for and enable
2798 properly configured MMIO access to PCI
2799 config space on AMD family 10h CPU
2800 nomsi [MSI] If the PCI_MSI kernel config parameter is
2801 enabled, this kernel boot option can be used to
2802 disable the use of MSI interrupts system-wide.
2803 noioapicquirk [APIC] Disable all boot interrupt quirks.
2804 Safety option to keep boot IRQs enabled. This
2805 should never be necessary.
2806 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
2807 primary IO-APIC for bridges that cannot disable
2808 boot IRQs. This fixes a source of spurious IRQs
2809 when the system masks IRQs.
2810 noioapicreroute [APIC] Disable workaround that uses the
2811 boot IRQ equivalent of an IRQ that connects to
2812 a chipset where boot IRQs cannot be disabled.
2813 The opposite of ioapicreroute.
2814 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
2815 routing table. These calls are known to be buggy
2816 on several machines and they hang the machine
2817 when used, but on other computers it's the only
2818 way to get the interrupt routing table. Try
2819 this option if the kernel is unable to allocate
2820 IRQs or discover secondary PCI buses on your
2822 rom [X86] Assign address space to expansion ROMs.
2823 Use with caution as certain devices share
2824 address decoders between ROMs and other
2826 norom [X86] Do not assign address space to
2827 expansion ROMs that do not already have
2828 BIOS assigned address ranges.
2829 nobar [X86] Do not assign address space to the
2830 BARs that weren't assigned by the BIOS.
2831 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
2832 assigned automatically to PCI devices. You can
2833 make the kernel exclude IRQs of your ISA cards
2835 pirqaddr=0xAAAAA [X86] Specify the physical address
2836 of the PIRQ table (normally generated
2837 by the BIOS) if it is outside the
2838 F0000h-100000h range.
2839 lastbus=N [X86] Scan all buses thru bus #N. Can be
2840 useful if the kernel is unable to find your
2841 secondary buses and you want to tell it
2842 explicitly which ones they are.
2843 assign-busses [X86] Always assign all PCI bus
2844 numbers ourselves, overriding
2845 whatever the firmware may have done.
2846 usepirqmask [X86] Honor the possible IRQ mask stored
2847 in the BIOS $PIR table. This is needed on
2848 some systems with broken BIOSes, notably
2849 some HP Pavilion N5400 and Omnibook XE3
2850 notebooks. This will have no effect if ACPI
2851 IRQ routing is enabled.
2852 noacpi [X86] Do not use ACPI for IRQ routing
2853 or for PCI scanning.
2854 use_crs [X86] Use PCI host bridge window information
2855 from ACPI. On BIOSes from 2008 or later, this
2856 is enabled by default. If you need to use this,
2857 please report a bug.
2858 nocrs [X86] Ignore PCI host bridge windows from ACPI.
2859 If you need to use this, please report a bug.
2860 routeirq Do IRQ routing for all PCI devices.
2861 This is normally done in pci_enable_device(),
2862 so this option is a temporary workaround
2863 for broken drivers that don't call it.
2864 skip_isa_align [X86] do not align io start addr, so can
2865 handle more pci cards
2866 noearly [X86] Don't do any early type 1 scanning.
2867 This might help on some broken boards which
2868 machine check when some devices' config space
2869 is read. But various workarounds are disabled
2870 and some IOMMU drivers will not work.
2871 bfsort Sort PCI devices into breadth-first order.
2872 This sorting is done to get a device
2873 order compatible with older (<= 2.4) kernels.
2874 nobfsort Don't sort PCI devices into breadth-first order.
2875 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
2876 tuning and use the BIOS-configured MPS defaults.
2877 pcie_bus_safe Set every device's MPS to the largest value
2878 supported by all devices below the root complex.
2879 pcie_bus_perf Set device MPS to the largest allowable MPS
2880 based on its parent bus. Also set MRRS (Max
2881 Read Request Size) to the largest supported
2882 value (no larger than the MPS that the device
2883 or bus can support) for best performance.
2884 pcie_bus_peer2peer Set every device's MPS to 128B, which
2885 every device is guaranteed to support. This
2886 configuration allows peer-to-peer DMA between
2887 any pair of devices, possibly at the cost of
2888 reduced performance. This also guarantees
2889 that hot-added devices will work.
2890 cbiosize=nn[KMG] The fixed amount of bus space which is
2891 reserved for the CardBus bridge's IO window.
2892 The default value is 256 bytes.
2893 cbmemsize=nn[KMG] The fixed amount of bus space which is
2894 reserved for the CardBus bridge's memory
2895 window. The default value is 64 megabytes.
2898 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
2899 [<order of align>@]pci:<vendor>:<device>\
2900 [:<subvendor>:<subdevice>][; ...]
2901 Specifies alignment and device to reassign
2902 aligned memory resources.
2903 If <order of align> is not specified,
2904 PAGE_SIZE is used as alignment.
2905 PCI-PCI bridge can be specified, if resource
2906 windows need to be expanded.
2907 To specify the alignment for several
2908 instances of a device, the PCI vendor,
2909 device, subvendor, and subdevice may be
2910 specified, e.g., 4096@pci:8086:9c22:103c:198f
2911 ecrc= Enable/disable PCIe ECRC (transaction layer
2912 end-to-end CRC checking).
2913 bios: Use BIOS/firmware settings. This is the
2917 hpiosize=nn[KMG] The fixed amount of bus space which is
2918 reserved for hotplug bridge's IO window.
2919 Default size is 256 bytes.
2920 hpmemsize=nn[KMG] The fixed amount of bus space which is
2921 reserved for hotplug bridge's memory window.
2922 Default size is 2 megabytes.
2923 hpbussize=nn The minimum amount of additional bus numbers
2924 reserved for buses below a hotplug bridge.
2926 realloc= Enable/disable reallocating PCI bridge resources
2927 if allocations done by BIOS are too small to
2928 accommodate resources required by all child
2930 off: Turn realloc off
2932 realloc same as realloc=on
2933 noari do not use PCIe ARI.
2934 pcie_scan_all Scan all possible PCIe devices. Otherwise we
2935 only look for one device below a PCIe downstream
2938 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
2941 force Enable ASPM even on devices that claim not to support it.
2942 WARNING: Forcing ASPM on may cause system lockups.
2944 pcie_hp= [PCIE] PCI Express Hotplug driver options:
2945 nomsi Do not use MSI for PCI Express Native Hotplug (this
2946 makes all PCIe ports use INTx for hotplug services).
2948 pcie_ports= [PCIE] PCIe ports handling:
2949 auto Ask the BIOS whether or not to use native PCIe services
2950 associated with PCIe ports (PME, hot-plug, AER). Use
2951 them only if that is allowed by the BIOS.
2952 native Use native PCIe services associated with PCIe ports
2954 compat Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
2957 pcie_port_pm= [PCIE] PCIe port power management handling:
2958 off Disable power management of all PCIe ports
2959 force Forcibly enable power management of all PCIe ports
2961 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
2962 nomsi Do not use MSI for native PCIe PME signaling (this makes
2963 all PCIe root ports use INTx for all services).
2965 pcmv= [HW,PCMCIA] BadgePAD 4
2969 Keep all power-domains already enabled by bootloader on,
2970 even if no driver has claimed them. This is useful
2971 for debug and development, but should not be
2972 needed on a platform with proper driver support.
2975 See Documentation/blockdev/paride.txt.
2977 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
2980 See arch/parisc/kernel/pdc_chassis.c
2982 percpu_alloc= Select which percpu first chunk allocator to use.
2983 Currently supported values are "embed" and "page".
2984 Archs may support subset or none of the selections.
2985 See comments in mm/percpu.c for details on each
2986 allocator. This parameter is primarily for debugging
2987 and performance comparison.
2990 See Documentation/blockdev/paride.txt.
2993 See Documentation/blockdev/paride.txt.
2995 pirq= [SMP,APIC] Manual mp-table setup
2996 See Documentation/x86/i386/IO-APIC.txt.
2998 plip= [PPT,NET] Parallel port network link
2999 Format: { parport<nr> | timid | 0 }
3000 See also Documentation/parport.txt.
3002 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3003 Override pmtimer IOPort with a hex value.
3007 Enable PNP debug messages (depends on the
3008 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3009 via /sys/module/pnp/parameters/debug. We always show
3010 current resource usage; turning this on also shows
3011 possible settings and some assignment information.
3017 { on | off | curr | res | no-curr | no-res }
3020 [ISAPNP] Exclude IRQs for the autoconfiguration
3023 [ISAPNP] Exclude DMAs for the autoconfiguration
3025 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3026 Ranges are in pairs (I/O port base and size).
3029 [ISAPNP] Exclude memory regions for the
3031 Ranges are in pairs (memory base and size).
3033 ports= [IP_VS_FTP] IPVS ftp helper module
3035 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3037 Format: <port>,<port>....
3039 powersave=off [PPC] This option disables power saving features.
3040 It specifically disables cpuidle and sets the
3041 platform machine description specific power_save
3042 function to NULL. On Idle the CPU just reduces
3045 ppc_strict_facility_enable
3046 [PPC] This option catches any kernel floating point,
3047 Altivec, VSX and SPE outside of regions specifically
3048 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3049 There is some performance impact when enabling this.
3051 print-fatal-signals=
3052 [KNL] debug: print fatal signals
3054 If enabled, warn about various signal handling
3055 related application anomalies: too many signals,
3056 too many POSIX.1 timers, fatal signals causing a
3059 If you hit the warning due to signal overflow,
3060 you might want to try "ulimit -i unlimited".
3064 printk.always_kmsg_dump=
3065 Trigger kmsg_dump for cases other than kernel oops or
3067 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3070 printk.devkmsg={on,off,ratelimit}
3071 Control writing to /dev/kmsg.
3072 on - unlimited logging to /dev/kmsg from userspace
3073 off - logging to /dev/kmsg disabled
3074 ratelimit - ratelimit the logging
3077 printk.time= Show timing data prefixed to each printk message line
3078 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3080 processor.max_cstate= [HW,ACPI]
3081 Limit processor to maximum C-state
3082 max_cstate=9 overrides any DMI blacklist limit.
3084 processor.nocst [HW,ACPI]
3085 Ignore the _CST method to determine C-states,
3086 instead using the legacy FADT method
3088 profile= [KNL] Enable kernel profiling via /proc/profile
3089 Format: [schedule,]<number>
3090 Param: "schedule" - profile schedule points.
3091 Param: <number> - step/bucket size as a power of 2 for
3092 statistical time based profiling.
3093 Param: "sleep" - profile D-state sleeping (millisecs).
3094 Requires CONFIG_SCHEDSTATS
3095 Param: "kvm" - profile VM exits.
3097 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3099 See Documentation/blockdev/ramdisk.txt.
3101 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3102 probe for; one of (bare|imps|exps|lifebook|any).
3103 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3105 psmouse.resetafter= [HW,MOUSE]
3106 Try to reset the device after so many bad packets
3109 [HW,MOUSE] Set desired mouse resolution, in dpi.
3110 psmouse.smartscroll=
3111 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3112 0 = disabled, 1 = enabled (default).
3114 pstore.backend= Specify the name of the pstore backend to use
3117 See Documentation/blockdev/paride.txt.
3120 [KNL] Number of legacy pty's. Overwrites compiled-in
3123 quiet [KNL] Disable most log messages
3128 See Documentation/admin-guide/md.rst.
3130 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3131 See Documentation/blockdev/ramdisk.txt.
3134 The argument is a cpu list, as described above.
3136 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3137 the specified list of CPUs to be no-callback CPUs.
3138 Invocation of these CPUs' RCU callbacks will
3139 be offloaded to "rcuox/N" kthreads created for
3140 that purpose, where "x" is "b" for RCU-bh, "p"
3141 for RCU-preempt, and "s" for RCU-sched, and "N"
3142 is the CPU number. This reduces OS jitter on the
3143 offloaded CPUs, which can be useful for HPC and
3144 real-time workloads. It can also improve energy
3145 efficiency for asymmetric multiprocessors.
3148 Rather than requiring that offloaded CPUs
3149 (specified by rcu_nocbs= above) explicitly
3150 awaken the corresponding "rcuoN" kthreads,
3151 make these kthreads poll for callbacks.
3152 This improves the real-time response for the
3153 offloaded CPUs by relieving them of the need to
3154 wake up the corresponding kthread, but degrades
3155 energy efficiency by requiring that the kthreads
3156 periodically wake up to do the polling.
3158 rcutree.blimit= [KNL]
3159 Set maximum number of finished RCU callbacks to
3160 process in one batch.
3162 rcutree.dump_tree= [KNL]
3163 Dump the structure of the rcu_node combining tree
3164 out at early boot. This is used for diagnostic
3165 purposes, to verify correct tree setup.
3167 rcutree.gp_cleanup_delay= [KNL]
3168 Set the number of jiffies to delay each step of
3169 RCU grace-period cleanup. This only has effect
3170 when CONFIG_RCU_TORTURE_TEST_SLOW_CLEANUP is set.
3172 rcutree.gp_init_delay= [KNL]
3173 Set the number of jiffies to delay each step of
3174 RCU grace-period initialization. This only has
3175 effect when CONFIG_RCU_TORTURE_TEST_SLOW_INIT
3178 rcutree.gp_preinit_delay= [KNL]
3179 Set the number of jiffies to delay each step of
3180 RCU grace-period pre-initialization, that is,
3181 the propagation of recent CPU-hotplug changes up
3182 the rcu_node combining tree. This only has effect
3183 when CONFIG_RCU_TORTURE_TEST_SLOW_PREINIT is set.
3185 rcutree.rcu_fanout_exact= [KNL]
3186 Disable autobalancing of the rcu_node combining
3187 tree. This is used by rcutorture, and might
3188 possibly be useful for architectures having high
3189 cache-to-cache transfer latencies.
3191 rcutree.rcu_fanout_leaf= [KNL]
3192 Change the number of CPUs assigned to each
3193 leaf rcu_node structure. Useful for very
3194 large systems, which will choose the value 64,
3195 and for NUMA systems with large remote-access
3196 latencies, which will choose a value aligned
3197 with the appropriate hardware boundaries.
3199 rcutree.jiffies_till_sched_qs= [KNL]
3200 Set required age in jiffies for a
3201 given grace period before RCU starts
3202 soliciting quiescent-state help from
3203 rcu_note_context_switch().
3205 rcutree.jiffies_till_first_fqs= [KNL]
3206 Set delay from grace-period initialization to
3207 first attempt to force quiescent states.
3208 Units are jiffies, minimum value is zero,
3209 and maximum value is HZ.
3211 rcutree.jiffies_till_next_fqs= [KNL]
3212 Set delay between subsequent attempts to force
3213 quiescent states. Units are jiffies, minimum
3214 value is one, and maximum value is HZ.
3216 rcutree.kthread_prio= [KNL,BOOT]
3217 Set the SCHED_FIFO priority of the RCU per-CPU
3218 kthreads (rcuc/N). This value is also used for
3219 the priority of the RCU boost threads (rcub/N)
3220 and for the RCU grace-period kthreads (rcu_bh,
3221 rcu_preempt, and rcu_sched). If RCU_BOOST is
3222 set, valid values are 1-99 and the default is 1
3223 (the least-favored priority). Otherwise, when
3224 RCU_BOOST is not set, valid values are 0-99 and
3225 the default is zero (non-realtime operation).
3227 rcutree.rcu_nocb_leader_stride= [KNL]
3228 Set the number of NOCB kthread groups, which
3229 defaults to the square root of the number of
3230 CPUs. Larger numbers reduces the wakeup overhead
3231 on the per-CPU grace-period kthreads, but increases
3232 that same overhead on each group's leader.
3234 rcutree.qhimark= [KNL]
3235 Set threshold of queued RCU callbacks beyond which
3236 batch limiting is disabled.
3238 rcutree.qlowmark= [KNL]
3239 Set threshold of queued RCU callbacks below which
3240 batch limiting is re-enabled.
3242 rcutree.rcu_idle_gp_delay= [KNL]
3243 Set wakeup interval for idle CPUs that have
3244 RCU callbacks (RCU_FAST_NO_HZ=y).
3246 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3247 Set wakeup interval for idle CPUs that have
3248 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3249 Lazy RCU callbacks are those which RCU can
3250 prove do nothing more than free memory.
3252 rcuperf.gp_exp= [KNL]
3253 Measure performance of expedited synchronous
3254 grace-period primitives.
3256 rcuperf.holdoff= [KNL]
3257 Set test-start holdoff period. The purpose of
3258 this parameter is to delay the start of the
3259 test until boot completes in order to avoid
3262 rcuperf.nreaders= [KNL]
3263 Set number of RCU readers. The value -1 selects
3264 N, where N is the number of CPUs. A value
3265 "n" less than -1 selects N-n+1, where N is again
3266 the number of CPUs. For example, -2 selects N
3267 (the number of CPUs), -3 selects N+1, and so on.
3268 A value of "n" less than or equal to -N selects
3271 rcuperf.nwriters= [KNL]
3272 Set number of RCU writers. The values operate
3273 the same as for rcuperf.nreaders.
3274 N, where N is the number of CPUs
3276 rcuperf.perf_runnable= [BOOT]
3277 Start rcuperf running at boot time.
3279 rcuperf.shutdown= [KNL]
3280 Shut the system down after performance tests
3281 complete. This is useful for hands-off automated
3284 rcuperf.perf_type= [KNL]
3285 Specify the RCU implementation to test.
3287 rcuperf.verbose= [KNL]
3288 Enable additional printk() statements.
3290 rcutorture.cbflood_inter_holdoff= [KNL]
3291 Set holdoff time (jiffies) between successive
3292 callback-flood tests.
3294 rcutorture.cbflood_intra_holdoff= [KNL]
3295 Set holdoff time (jiffies) between successive
3296 bursts of callbacks within a given callback-flood
3299 rcutorture.cbflood_n_burst= [KNL]
3300 Set the number of bursts making up a given
3301 callback-flood test. Set this to zero to
3302 disable callback-flood testing.
3304 rcutorture.cbflood_n_per_burst= [KNL]
3305 Set the number of callbacks to be registered
3306 in a given burst of a callback-flood test.
3308 rcutorture.fqs_duration= [KNL]
3309 Set duration of force_quiescent_state bursts
3312 rcutorture.fqs_holdoff= [KNL]
3313 Set holdoff time within force_quiescent_state bursts
3316 rcutorture.fqs_stutter= [KNL]
3317 Set wait time between force_quiescent_state bursts
3320 rcutorture.gp_cond= [KNL]
3321 Use conditional/asynchronous update-side
3322 primitives, if available.
3324 rcutorture.gp_exp= [KNL]
3325 Use expedited update-side primitives, if available.
3327 rcutorture.gp_normal= [KNL]
3328 Use normal (non-expedited) asynchronous
3329 update-side primitives, if available.
3331 rcutorture.gp_sync= [KNL]
3332 Use normal (non-expedited) synchronous
3333 update-side primitives, if available. If all
3334 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3335 rcutorture.gp_normal=, and rcutorture.gp_sync=
3336 are zero, rcutorture acts as if is interpreted
3337 they are all non-zero.
3339 rcutorture.n_barrier_cbs= [KNL]
3340 Set callbacks/threads for rcu_barrier() testing.
3342 rcutorture.nfakewriters= [KNL]
3343 Set number of concurrent RCU writers. These just
3344 stress RCU, they don't participate in the actual
3345 test, hence the "fake".
3347 rcutorture.nreaders= [KNL]
3348 Set number of RCU readers. The value -1 selects
3349 N-1, where N is the number of CPUs. A value
3350 "n" less than -1 selects N-n-2, where N is again
3351 the number of CPUs. For example, -2 selects N
3352 (the number of CPUs), -3 selects N+1, and so on.
3354 rcutorture.object_debug= [KNL]
3355 Enable debug-object double-call_rcu() testing.
3357 rcutorture.onoff_holdoff= [KNL]
3358 Set time (s) after boot for CPU-hotplug testing.
3360 rcutorture.onoff_interval= [KNL]
3361 Set time (s) between CPU-hotplug operations, or
3362 zero to disable CPU-hotplug testing.
3364 rcutorture.shuffle_interval= [KNL]
3365 Set task-shuffle interval (s). Shuffling tasks
3366 allows some CPUs to go into dyntick-idle mode
3367 during the rcutorture test.
3369 rcutorture.shutdown_secs= [KNL]
3370 Set time (s) after boot system shutdown. This
3371 is useful for hands-off automated testing.
3373 rcutorture.stall_cpu= [KNL]
3374 Duration of CPU stall (s) to test RCU CPU stall
3375 warnings, zero to disable.
3377 rcutorture.stall_cpu_holdoff= [KNL]
3378 Time to wait (s) after boot before inducing stall.
3380 rcutorture.stat_interval= [KNL]
3381 Time (s) between statistics printk()s.
3383 rcutorture.stutter= [KNL]
3384 Time (s) to stutter testing, for example, specifying
3385 five seconds causes the test to run for five seconds,
3386 wait for five seconds, and so on. This tests RCU's
3387 ability to transition abruptly to and from idle.
3389 rcutorture.test_boost= [KNL]
3390 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3391 "Maybe" means test if the RCU implementation
3392 under test support RCU priority boosting.
3394 rcutorture.test_boost_duration= [KNL]
3395 Duration (s) of each individual boost test.
3397 rcutorture.test_boost_interval= [KNL]
3398 Interval (s) between each boost test.
3400 rcutorture.test_no_idle_hz= [KNL]
3401 Test RCU's dyntick-idle handling. See also the
3402 rcutorture.shuffle_interval parameter.
3404 rcutorture.torture_runnable= [BOOT]
3405 Start rcutorture running at boot time.
3407 rcutorture.torture_type= [KNL]
3408 Specify the RCU implementation to test.
3410 rcutorture.verbose= [KNL]
3411 Enable additional printk() statements.
3413 rcupdate.rcu_cpu_stall_suppress= [KNL]
3414 Suppress RCU CPU stall warning messages.
3416 rcupdate.rcu_cpu_stall_timeout= [KNL]
3417 Set timeout for RCU CPU stall warning messages.
3419 rcupdate.rcu_expedited= [KNL]
3420 Use expedited grace-period primitives, for
3421 example, synchronize_rcu_expedited() instead
3422 of synchronize_rcu(). This reduces latency,
3423 but can increase CPU utilization, degrade
3424 real-time latency, and degrade energy efficiency.
3425 No effect on CONFIG_TINY_RCU kernels.
3427 rcupdate.rcu_normal= [KNL]
3428 Use only normal grace-period primitives,
3429 for example, synchronize_rcu() instead of
3430 synchronize_rcu_expedited(). This improves
3431 real-time latency, CPU utilization, and
3432 energy efficiency, but can expose users to
3433 increased grace-period latency. This parameter
3434 overrides rcupdate.rcu_expedited. No effect on
3435 CONFIG_TINY_RCU kernels.
3437 rcupdate.rcu_normal_after_boot= [KNL]
3438 Once boot has completed (that is, after
3439 rcu_end_inkernel_boot() has been invoked), use
3440 only normal grace-period primitives. No effect
3441 on CONFIG_TINY_RCU kernels.
3443 rcupdate.rcu_task_stall_timeout= [KNL]
3444 Set timeout in jiffies for RCU task stall warning
3445 messages. Disable with a value less than or equal
3448 rcupdate.rcu_self_test= [KNL]
3449 Run the RCU early boot self tests
3451 rcupdate.rcu_self_test_bh= [KNL]
3452 Run the RCU bh early boot self tests
3454 rcupdate.rcu_self_test_sched= [KNL]
3455 Run the RCU sched early boot self tests
3459 Run specified binary instead of /init from the ramdisk,
3460 used for early userspace startup. See initrd.
3463 Format (x86 or x86_64):
3464 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3466 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3468 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3469 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3470 reboot_force is either force or not specified,
3471 reboot_cpu is s[mp]#### with #### being the processor
3472 to be used for rebooting.
3475 [KNL, SMP] Set scheduler's default relax_domain_level.
3476 See Documentation/cgroup-v1/cpusets.txt.
3478 relative_sleep_states=
3479 [SUSPEND] Use sleep state labeling where the deepest
3480 state available other than hibernation is always "mem".
3481 Format: { "0" | "1" }
3482 0 -- Traditional sleep state labels.
3483 1 -- Relative sleep state labels.
3485 reserve= [KNL,BUGS] Force the kernel to ignore some iomem area
3487 reservetop= [X86-32]
3489 Reserves a hole at the top of the kernel virtual
3494 Set the amount of memory to reserve for BIOS at
3495 the bottom of the address space.
3497 reset_devices [KNL] Force drivers to reset the underlying device
3498 during initialization.
3501 Specify the partition device for software suspend
3503 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3505 resume_offset= [SWSUSP]
3506 Specify the offset from the beginning of the partition
3507 given by "resume=" at which the swap header is located,
3508 in <PAGE_SIZE> units (needed only for swap files).
3509 See Documentation/power/swsusp-and-swap-files.txt
3511 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
3512 read the resume files
3514 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
3515 Useful for devices that are detected asynchronously
3516 (e.g. USB and MMC devices).
3518 hibernate= [HIBERNATION]
3519 noresume Don't check if there's a hibernation image
3520 present during boot.
3521 nocompress Don't compress/decompress hibernation images.
3522 no Disable hibernation and resume.
3523 protect_image Turn on image protection during restoration
3524 (that will set all pages holding image data
3525 during restoration read-only).
3527 retain_initrd [RAM] Keep initrd memory after extraction
3529 rfkill.default_state=
3530 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
3531 etc. communication is blocked by default.
3534 rfkill.master_switch_mode=
3535 0 The "airplane mode" button does nothing.
3536 1 The "airplane mode" button toggles between everything
3537 blocked and the previous configuration.
3538 2 The "airplane mode" button toggles between everything
3539 blocked and everything unblocked.
3541 rhash_entries= [KNL,NET]
3542 Set number of hash buckets for route cache
3544 ro [KNL] Mount root device read-only on boot
3547 on Mark read-only kernel memory as read-only (default).
3548 off Leave read-only kernel memory writable for debugging.
3551 Enable the uart passthrough on the designated usb port
3552 on Rockchip SoCs. When active, the signals of the
3553 debug-uart get routed to the D+ and D- pins of the usb
3554 port and the regular usb controller gets disabled.
3556 root= [KNL] Root filesystem
3557 See name_to_dev_t comment in init/do_mounts.c.
3559 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
3560 mount the root filesystem
3562 rootflags= [KNL] Set root filesystem mount option string
3564 rootfstype= [KNL] Set root filesystem type
3566 rootwait [KNL] Wait (indefinitely) for root device to show up.
3567 Useful for devices that are detected asynchronously
3568 (e.g. USB and MMC devices).
3570 rproc_mem=nn[KMG][@address]
3571 [KNL,ARM,CMA] Remoteproc physical memory block.
3572 Memory area to be used by remote processor image,
3575 rw [KNL] Mount root device read-write on boot
3577 S [KNL] Run init in single mode
3579 s390_iommu= [HW,S390]
3580 Set s390 IOTLB flushing mode
3582 With strict flushing every unmap operation will result in
3583 an IOTLB flush. Default is lazy flushing before reuse,
3587 See drivers/net/irda/sa1100_ir.c.
3589 sbni= [NET] Granch SBNI12 leased line adapter
3591 sched_debug [KNL] Enables verbose scheduler debug messages.
3593 schedstats= [KNL,X86] Enable or disable scheduled statistics.
3594 Allowed values are enable and disable. This feature
3595 incurs a small amount of overhead in the scheduler
3596 but is useful for debugging and performance tuning.
3598 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
3599 xtime_lock contention on larger systems, and/or RCU lock
3600 contention on all systems with CONFIG_MAXSMP set.
3601 Format: { "0" | "1" }
3602 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
3604 Note: increases power consumption, thus should only be
3605 enabled if running jitter sensitive (HPC/RT) workloads.
3607 security= [SECURITY] Choose a security module to enable at boot.
3608 If this boot parameter is not specified, only the first
3609 security module asking for security registration will be
3610 loaded. An invalid security module name will be treated
3611 as if no module has been chosen.
3613 selinux= [SELINUX] Disable or enable SELinux at boot time.
3614 Format: { "0" | "1" }
3615 See security/selinux/Kconfig help text.
3618 Default value is set via kernel config option.
3619 If enabled at boot time, /selinux/disable can be used
3620 later to disable prior to initial policy load.
3622 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
3623 Format: { "0" | "1" }
3624 See security/apparmor/Kconfig help text
3627 Default value is set via kernel config option.
3629 serialnumber [BUGS=X86-32]
3632 Maximal number of shapers.
3634 show_msr= [x86] show boot-time MSR settings
3635 Format: { <integer> }
3636 Show boot-time (BIOS-initialized) MSR settings.
3637 The parameter means the number of CPUs to show,
3638 for example 1 means boot CPU only.
3646 Disable merging of slabs with similar size. May be
3647 necessary if there is some reason to distinguish
3648 allocs to different slabs. Debug options disable
3649 merging on their own.
3650 For more information see Documentation/vm/slub.txt.
3652 slab_max_order= [MM, SLAB]
3653 Determines the maximum allowed order for slabs.
3654 A high setting may cause OOMs due to memory
3655 fragmentation. Defaults to 1 for systems with
3656 more than 32MB of RAM, 0 otherwise.
3658 slub_debug[=options[,slabs]] [MM, SLUB]
3659 Enabling slub_debug allows one to determine the
3660 culprit if slab objects become corrupted. Enabling
3661 slub_debug can create guard zones around objects and
3662 may poison objects when not in use. Also tracks the
3663 last alloc / free. For more information see
3664 Documentation/vm/slub.txt.
3666 slub_max_order= [MM, SLUB]
3667 Determines the maximum allowed order for slabs.
3668 A high setting may cause OOMs due to memory
3669 fragmentation. For more information see
3670 Documentation/vm/slub.txt.
3672 slub_min_objects= [MM, SLUB]
3673 The minimum number of objects per slab. SLUB will
3674 increase the slab order up to slub_max_order to
3675 generate a sufficiently large slab able to contain
3676 the number of objects indicated. The higher the number
3677 of objects the smaller the overhead of tracking slabs
3678 and the less frequently locks need to be acquired.
3679 For more information see Documentation/vm/slub.txt.
3681 slub_min_order= [MM, SLUB]
3682 Determines the minimum page order for slabs. Must be
3683 lower than slub_max_order.
3684 For more information see Documentation/vm/slub.txt.
3686 slub_nomerge [MM, SLUB]
3687 Same with slab_nomerge. This is supported for legacy.
3688 See slab_nomerge for more information.
3691 Format: <io1>[,<io2>[,...,<io8>]]
3693 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
3694 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
3695 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
3696 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
3697 smsc-ircc2.ircc_irq= [HW] IRQ line
3698 smsc-ircc2.ircc_dma= [HW] DMA channel
3699 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
3700 0: Toshiba Satellite 1800 (GP data pin select)
3701 1: Fast pin select (default)
3704 smt [KNL,S390] Set the maximum number of threads (logical
3705 CPUs) to use per physical CPU on systems capable of
3706 symmetric multithreading (SMT). Will be capped to the
3707 actual hardware limit.
3709 Default: -1 (no limit)
3712 [KNL] Should the soft-lockup detector generate panics.
3715 softlockup_all_cpu_backtrace=
3716 [KNL] Should the soft-lockup detector generate
3717 backtraces on all cpus.
3720 sonypi.*= [HW] Sony Programmable I/O Control Device driver
3721 See Documentation/laptops/sonypi.txt
3723 spia_io_base= [HW,MTD]
3729 Enabled the stack tracer on boot up.
3731 stacktrace_filter=[function-list]
3732 [FTRACE] Limit the functions that the stack tracer
3733 will trace at boot up. function-list is a comma separated
3734 list of functions. This list can be changed at run
3735 time by the stack_trace_filter file in the debugfs
3736 tracing directory. Note, this enables stack tracing
3737 and the stacktrace above is not needed.
3741 Set the STI (builtin display/keyboard on the HP-PARISC
3742 machines) console (graphic card) which should be used
3743 as the initial boot-console.
3744 See also comment in drivers/video/console/sticore.c.
3747 See comment in drivers/video/console/sticore.c.
3750 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
3752 sunrpc.min_resvport=
3753 sunrpc.max_resvport=
3755 SunRPC servers often require that client requests
3756 originate from a privileged port (i.e. a port in the
3757 range 0 < portnr < 1024).
3758 An administrator who wishes to reserve some of these
3759 ports for other uses may adjust the range that the
3760 kernel's sunrpc client considers to be privileged
3761 using these two parameters to set the minimum and
3762 maximum port values.
3764 sunrpc.svc_rpc_per_connection_limit=
3766 Limit the number of requests that the server will
3767 process in parallel from a single connection.
3768 The default value is 0 (no limit).
3772 Control how the NFS server code allocates CPUs to
3773 service thread pools. Depending on how many NICs
3774 you have and where their interrupts are bound, this
3775 option will affect which CPUs will do NFS serving.
3776 Note: this parameter cannot be changed while the
3777 NFS server is running.
3779 auto the server chooses an appropriate mode
3780 automatically using heuristics
3781 global a single global pool contains all CPUs
3782 percpu one pool for each CPU
3783 pernode one pool for each NUMA node (equivalent
3784 to global on non-NUMA machines)
3786 sunrpc.tcp_slot_table_entries=
3787 sunrpc.udp_slot_table_entries=
3789 Sets the upper limit on the number of simultaneous
3790 RPC calls that can be sent from the client to a
3791 server. Increasing these values may allow you to
3792 improve throughput, but will also increase the
3793 amount of memory reserved for use by the client.
3795 suspend.pm_test_delay=
3797 Sets the number of seconds to remain in a suspend test
3798 mode before resuming the system (see
3799 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
3800 is set. Default value is 5.
3803 [KNL] Enable accounting of swap in memory resource
3804 controller if no parameter or 1 is given or disable
3805 it if 0 is given (See Documentation/cgroup-v1/memory.txt)
3807 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
3808 Format: { <int> | force }
3809 <int> -- Number of I/O TLB slabs
3810 force -- force using of bounce buffers even if they
3811 wouldn't be automatically used by the kernel
3815 sysfs.deprecated=0|1 [KNL]
3816 Enable/disable old style sysfs layout for old udev
3817 on older distributions. When this option is enabled
3818 very new udev will not work anymore. When this option
3819 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
3820 in older udev will not work anymore.
3821 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
3822 the kernel configuration.
3824 sysrq_always_enabled
3826 Ignore sysrq setting - this boot parameter will
3827 neutralize any effect of /proc/sys/kernel/sysrq.
3828 Useful for debugging.
3830 tcpmhash_entries= [KNL,NET]
3831 Set the number of tcp_metrics_hash slots.
3832 Default value is 8192 or 16384 depending on total
3833 ram pages. This is used to specify the TCP metrics
3834 cache size. See Documentation/networking/ip-sysctl.txt
3835 "tcp_no_metrics_save" section for more details.
3839 test_suspend= [SUSPEND][,N]
3840 Specify "mem" (for Suspend-to-RAM) or "standby" (for
3841 standby suspend) or "freeze" (for suspend type freeze)
3842 as the system sleep state during system startup with
3843 the optional capability to repeat N number of times.
3844 The system is woken from this state using a
3845 wakeup-capable RTC alarm.
3847 thash_entries= [KNL,NET]
3848 Set number of hash buckets for TCP connection
3850 thermal.act= [HW,ACPI]
3851 -1: disable all active trip points in all thermal zones
3852 <degrees C>: override all lowest active trip points
3854 thermal.crt= [HW,ACPI]
3855 -1: disable all critical trip points in all thermal zones
3856 <degrees C>: override all critical trip points
3858 thermal.nocrt= [HW,ACPI]
3859 Set to disable actions on ACPI thermal zone
3860 critical and hot trip points.
3862 thermal.off= [HW,ACPI]
3863 1: disable ACPI thermal control
3865 thermal.psv= [HW,ACPI]
3866 -1: disable all passive trip points
3867 <degrees C>: override all passive trip points to this
3870 thermal.tzp= [HW,ACPI]
3871 Specify global default ACPI thermal zone polling rate
3872 <deci-seconds>: poll all this frequency
3873 0: no polling (default)
3876 Force threading of all interrupt handlers except those
3877 marked explicitly IRQF_NO_THREAD.
3880 Enable the Transcendent memory driver if built-in.
3882 tmem.cleancache=0|1 [KNL, XEN]
3883 Default is on (1). Disable the usage of the cleancache
3884 API to send anonymous pages to the hypervisor.
3886 tmem.frontswap=0|1 [KNL, XEN]
3887 Default is on (1). Disable the usage of the frontswap
3888 API to send swap pages to the hypervisor. If disabled
3889 the selfballooning and selfshrinking are force disabled.
3891 tmem.selfballooning=0|1 [KNL, XEN]
3892 Default is on (1). Disable the driving of swap pages
3895 tmem.selfshrinking=0|1 [KNL, XEN]
3896 Default is on (1). Partial swapoff that immediately
3897 transfers pages from Xen hypervisor back to the
3898 kernel based on different criteria.
3902 Specify if the kernel should make use of the cpu
3903 topology information if the hardware supports this.
3904 The scheduler will make use of this information and
3905 e.g. base its process migration decisions on it.
3908 topology_updates= [KNL, PPC, NUMA]
3910 Specify if the kernel should ignore (off)
3911 topology updates sent by the hypervisor to this
3916 tpm_suspend_pcr=[HW,TPM]
3917 Format: integer pcr id
3918 Specify that at suspend time, the tpm driver
3919 should extend the specified pcr with zeros,
3920 as a workaround for some chips which fail to
3921 flush the last written pcr on TPM_SaveState.
3922 This will guarantee that all the other pcrs
3925 trace_buf_size=nn[KMG]
3926 [FTRACE] will set tracing buffer size on each cpu.
3928 trace_event=[event-list]
3929 [FTRACE] Set and start specified trace events in order
3930 to facilitate early boot debugging. The event-list is a
3931 comma separated list of trace events to enable. See
3932 also Documentation/trace/events.txt
3934 trace_options=[option-list]
3935 [FTRACE] Enable or disable tracer options at boot.
3936 The option-list is a comma delimited list of options
3937 that can be enabled or disabled just as if you were
3938 to echo the option name into
3940 /sys/kernel/debug/tracing/trace_options
3942 For example, to enable stacktrace option (to dump the
3943 stack trace of each event), add to the command line:
3945 trace_options=stacktrace
3947 See also Documentation/trace/ftrace.txt "trace options"
3951 Have the tracepoints sent to printk as well as the
3952 tracing ring buffer. This is useful for early boot up
3953 where the system hangs or reboots and does not give the
3954 option for reading the tracing buffer or performing a
3955 ftrace_dump_on_oops.
3957 To turn off having tracepoints sent to printk,
3958 echo 0 > /proc/sys/kernel/tracepoint_printk
3959 Note, echoing 1 into this file without the
3960 tracepoint_printk kernel cmdline option has no effect.
3964 Having tracepoints sent to printk() and activating high
3965 frequency tracepoints such as irq or sched, can cause
3966 the system to live lock.
3969 [FTRACE] enable this option to disable tracing when a
3970 warning is hit. This turns off "tracing_on". Tracing can
3971 be enabled again by echoing '1' into the "tracing_on"
3972 file located in /sys/kernel/debug/tracing/
3974 This option is useful, as it disables the trace before
3975 the WARNING dump is called, which prevents the trace to
3976 be filled with content caused by the warning output.
3978 This option can also be set at run time via the sysctl
3979 option: kernel/traceoff_on_warning
3981 transparent_hugepage=
3983 Format: [always|madvise|never]
3984 Can be used to control the default behavior of the system
3985 with respect to transparent hugepages.
3986 See Documentation/vm/transhuge.txt for more details.
3988 tsc= Disable clocksource stability checks for TSC.
3990 [x86] reliable: mark tsc clocksource as reliable, this
3991 disables clocksource verification at runtime, as well
3992 as the stability checks done at bootup. Used to enable
3993 high-resolution timer mode on older hardware, and in
3994 virtualized environment.
3995 [x86] noirqtime: Do not use TSC to do irq accounting.
3996 Used to run time disable IRQ_TIME_ACCOUNTING on any
3997 platforms where RDTSC is slow and this accounting
4000 turbografx.map[2|3]= [HW,JOY]
4001 TurboGraFX parallel port interface
4003 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4004 See also Documentation/input/joystick-parport.txt
4006 udbg-immortal [PPC] When debugging early kernel crashes that
4007 happen after console_init() and before a proper
4008 console driver takes over, this boot options might
4009 help "seeing" what's going on.
4011 uhash_entries= [KNL,NET]
4012 Set number of hash buckets for UDP/UDP-Lite connections
4015 [USB] Ignore overcurrent events (default N).
4016 Some badly-designed motherboards generate lots of
4017 bogus events, for ports that aren't wired to
4018 anything. Set this parameter to avoid log spamming.
4019 Note that genuine overcurrent events won't be
4023 [X86] Cause panic on unknown NMI.
4025 usbcore.authorized_default=
4026 [USB] Default USB device authorization:
4027 (default -1 = authorized except for wireless USB,
4028 0 = not authorized, 1 = authorized)
4030 usbcore.autosuspend=
4031 [USB] The autosuspend time delay (in seconds) used
4032 for newly-detected USB devices (default 2). This
4033 is the time required before an idle device will be
4034 autosuspended. Devices for which the delay is set
4035 to a negative value won't be autosuspended at all.
4037 usbcore.usbfs_snoop=
4038 [USB] Set to log all usbfs traffic (default 0 = off).
4040 usbcore.usbfs_snoop_max=
4041 [USB] Maximum number of bytes to snoop in each URB
4044 usbcore.blinkenlights=
4045 [USB] Set to cycle leds on hubs (default 0 = off).
4047 usbcore.old_scheme_first=
4048 [USB] Start with the old device initialization
4049 scheme (default 0 = off).
4051 usbcore.usbfs_memory_mb=
4052 [USB] Memory limit (in MB) for buffers allocated by
4053 usbfs (default = 16, 0 = max = 2047).
4055 usbcore.use_both_schemes=
4056 [USB] Try the other device initialization scheme
4057 if the first one fails (default 1 = enabled).
4059 usbcore.initial_descriptor_timeout=
4060 [USB] Specifies timeout for the initial 64-byte
4061 USB_REQ_GET_DESCRIPTOR request in milliseconds
4062 (default 5000 = 5.0 seconds).
4064 usbcore.nousb [USB] Disable the USB subsystem
4067 [USBHID] The interval which mice are to be polled at.
4069 usb-storage.delay_use=
4070 [UMS] The delay in seconds before a new device is
4071 scanned for Logical Units (default 1).
4074 [UMS] A list of quirks entries to supplement or
4075 override the built-in unusual_devs list. List
4076 entries are separated by commas. Each entry has
4077 the form VID:PID:Flags where VID and PID are Vendor
4078 and Product ID values (4-digit hex numbers) and
4079 Flags is a set of characters, each corresponding
4080 to a common usb-storage quirk flag as follows:
4081 a = SANE_SENSE (collect more than 18 bytes
4083 b = BAD_SENSE (don't collect more than 18
4084 bytes of sense data);
4085 c = FIX_CAPACITY (decrease the reported
4086 device capacity by one sector);
4087 d = NO_READ_DISC_INFO (don't use
4088 READ_DISC_INFO command);
4089 e = NO_READ_CAPACITY_16 (don't use
4090 READ_CAPACITY_16 command);
4091 f = NO_REPORT_OPCODES (don't use report opcodes
4093 g = MAX_SECTORS_240 (don't transfer more than
4094 240 sectors at a time, uas only);
4095 h = CAPACITY_HEURISTICS (decrease the
4096 reported device capacity by one
4097 sector if the number is odd);
4098 i = IGNORE_DEVICE (don't bind to this
4100 j = NO_REPORT_LUNS (don't use report luns
4102 l = NOT_LOCKABLE (don't try to lock and
4103 unlock ejectable media);
4104 m = MAX_SECTORS_64 (don't transfer more
4105 than 64 sectors = 32 KB at a time);
4106 n = INITIAL_READ10 (force a retry of the
4107 initial READ(10) command);
4108 o = CAPACITY_OK (accept the capacity
4109 reported by the device);
4110 p = WRITE_CACHE (the device cache is ON
4112 r = IGNORE_RESIDUE (the device reports
4113 bogus residue values);
4114 s = SINGLE_LUN (the device has only one
4116 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
4117 commands, uas only);
4118 u = IGNORE_UAS (don't bind to the uas driver);
4119 w = NO_WP_DETECT (don't test whether the
4120 medium is write-protected).
4121 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
4122 even if the device claims no cache)
4123 Example: quirks=0419:aaf5:rl,0421:0433:rc
4125 user_debug= [KNL,ARM]
4127 See arch/arm/Kconfig.debug help text.
4128 1 - undefined instruction events
4130 4 - invalid data aborts
4133 Example: user_debug=31
4136 [X86] Flags controlling user PTE allocations.
4138 nohigh = do not allocate PTE pages in
4139 HIGHMEM regardless of setting
4143 On X86_32, this is an alias for vdso32=. Otherwise:
4145 vdso=1: enable VDSO (the default)
4146 vdso=0: disable VDSO mapping
4148 vdso32= [X86] Control the 32-bit vDSO
4149 vdso32=1: enable 32-bit VDSO
4150 vdso32=0 or vdso32=2: disable 32-bit VDSO
4152 See the help text for CONFIG_COMPAT_VDSO for more
4153 details. If CONFIG_COMPAT_VDSO is set, the default is
4154 vdso32=0; otherwise, the default is vdso32=1.
4156 For compatibility with older kernels, vdso32=2 is an
4159 Try vdso32=0 if you encounter an error that says:
4160 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
4163 vector=percpu: enable percpu vector domain
4165 video= [FB] Frame buffer configuration
4166 See Documentation/fb/modedb.txt.
4168 video.brightness_switch_enabled= [0,1]
4169 If set to 1, on receiving an ACPI notify event
4170 generated by hotkey, video driver will adjust brightness
4171 level and then send out the event to user space through
4172 the allocated input device; If set to 0, video driver
4173 will only send out the event without touching backlight
4178 [VMMIO] Memory mapped virtio (platform) device.
4180 <size>@<baseaddr>:<irq>[:<id>]
4182 <size> := size (can use standard suffixes
4184 <baseaddr> := physical base address
4185 <irq> := interrupt number (as passed to
4187 <id> := (optional) platform device id
4189 virtio_mmio.device=1K@0x100b0000:48:7
4191 Can be used multiple times for multiple devices.
4193 vga= [BOOT,X86-32] Select a particular video mode
4194 See Documentation/x86/boot.txt and
4195 Documentation/svga.txt.
4196 Use vga=ask for menu.
4197 This is actually a boot loader parameter; the value is
4198 passed to the kernel using a special protocol.
4200 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
4201 size of <nn>. This can be used to increase the
4202 minimum size (128MB on x86). It can also be used to
4203 decrease the size and leave more room for directly
4206 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
4209 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
4212 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
4216 Controls the behavior of vsyscalls (i.e. calls to
4217 fixed addresses of 0xffffffffff600x00 from legacy
4218 code). Most statically-linked binaries and older
4219 versions of glibc use these calls. Because these
4220 functions are at fixed addresses, they make nice
4221 targets for exploits that can control RIP.
4223 emulate [default] Vsyscalls turn into traps and are
4224 emulated reasonably safely.
4226 native Vsyscalls are native syscall instructions.
4227 This is a little bit faster than trapping
4228 and makes a few dynamic recompilers work
4229 better than they would in emulation mode.
4230 It also makes exploits much easier to write.
4232 none Vsyscalls don't work at all. This makes
4233 them quite hard to use for exploits but
4234 might break your system.
4236 vt.color= [VT] Default text color.
4237 Format: 0xYX, X = foreground, Y = background.
4238 Default: 0x07 = light gray on black.
4240 vt.cur_default= [VT] Default cursor shape.
4241 Format: 0xCCBBAA, where AA, BB, and CC are the same as
4242 the parameters of the <Esc>[?A;B;Cc escape sequence;
4243 see VGA-softcursor.txt. Default: 2 = underline.
4245 vt.default_blu= [VT]
4246 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
4247 Change the default blue palette of the console.
4248 This is a 16-member array composed of values
4251 vt.default_grn= [VT]
4252 Format: <green0>,<green1>,<green2>,...,<green15>
4253 Change the default green palette of the console.
4254 This is a 16-member array composed of values
4257 vt.default_red= [VT]
4258 Format: <red0>,<red1>,<red2>,...,<red15>
4259 Change the default red palette of the console.
4260 This is a 16-member array composed of values
4266 Set system-wide default UTF-8 mode for all tty's.
4267 Default is 1, i.e. UTF-8 mode is enabled for all
4268 newly opened terminals.
4270 vt.global_cursor_default=
4273 Set system-wide default for whether a cursor
4274 is shown on new VTs. Default is -1,
4275 i.e. cursors will be created by default unless
4276 overridden by individual drivers. 0 will hide
4277 cursors, 1 will display them.
4279 vt.italic= [VT] Default color for italic text; 0-15.
4282 vt.underline= [VT] Default color for underlined text; 0-15.
4285 watchdog timers [HW,WDT] For information on watchdog timers,
4286 see Documentation/watchdog/watchdog-parameters.txt
4287 or other driver-specific files in the
4288 Documentation/watchdog/ directory.
4290 workqueue.watchdog_thresh=
4291 If CONFIG_WQ_WATCHDOG is configured, workqueue can
4292 warn stall conditions and dump internal state to
4293 help debugging. 0 disables workqueue stall
4294 detection; otherwise, it's the stall threshold
4295 duration in seconds. The default value is 30 and
4296 it can be updated at runtime by writing to the
4297 corresponding sysfs file.
4299 workqueue.disable_numa
4300 By default, all work items queued to unbound
4301 workqueues are affine to the NUMA nodes they're
4302 issued on, which results in better behavior in
4303 general. If NUMA affinity needs to be disabled for
4304 whatever reason, this option can be used. Note
4305 that this also can be controlled per-workqueue for
4306 workqueues visible under /sys/bus/workqueue/.
4308 workqueue.power_efficient
4309 Per-cpu workqueues are generally preferred because
4310 they show better performance thanks to cache
4311 locality; unfortunately, per-cpu workqueues tend to
4312 be more power hungry than unbound workqueues.
4314 Enabling this makes the per-cpu workqueues which
4315 were observed to contribute significantly to power
4316 consumption unbound, leading to measurably lower
4317 power usage at the cost of small performance
4320 The default value of this parameter is determined by
4321 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
4323 workqueue.debug_force_rr_cpu
4324 Workqueue used to implicitly guarantee that work
4325 items queued without explicit CPU specified are put
4326 on the local CPU. This guarantee is no longer true
4327 and while local CPU is still preferred work items
4328 may be put on foreign CPUs. This debug option
4329 forces round-robin CPU selection to flush out
4330 usages which depend on the now broken guarantee.
4331 When enabled, memory and cache locality will be
4334 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
4335 default x2apic cluster mode on platforms
4338 x86_intel_mid_timer= [X86-32,APBT]
4339 Choose timer option for x86 Intel MID platform.
4340 Two valid options are apbt timer only and lapic timer
4341 plus one apbt timer for broadcast timer.
4342 x86_intel_mid_timer=apbt_only | lapic_and_apbt
4344 xen_512gb_limit [KNL,X86-64,XEN]
4345 Restricts the kernel running paravirtualized under Xen
4346 to use only up to 512 GB of RAM. The reason to do so is
4347 crash analysis tools and Xen tools for doing domain
4348 save/restore/migration must be enabled to handle larger
4351 xen_emul_unplug= [HW,X86,XEN]
4352 Unplug Xen emulated devices
4353 Format: [unplug0,][unplug1]
4354 ide-disks -- unplug primary master IDE devices
4355 aux-ide-disks -- unplug non-primary-master IDE devices
4356 nics -- unplug network devices
4357 all -- unplug all emulated devices (NICs and IDE disks)
4358 unnecessary -- unplugging emulated devices is
4359 unnecessary even if the host did not respond to
4361 never -- do not unplug even if version check succeeds
4363 xen_nopvspin [X86,XEN]
4364 Disables the ticketlock slowpath using Xen PV
4368 Disables the PV optimizations forcing the HVM guest to
4369 run as generic HVM guest with no PV drivers.
4371 xirc2ps_cs= [NET,PCMCIA]
4373 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]