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.rst, 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/firmware-guide/acpi/debug.rst 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_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_rsdp= [ACPI,EFI,KEXEC]
140 Pass the RSDP address to the kernel, mostly used
141 on machines running EFI runtime service to boot the
142 second kernel for kdump.
144 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
145 Format: To spoof as Windows 98: ="Microsoft Windows"
147 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
148 of 2 which is mandated by ACPI 6) as the supported ACPI
149 specification revision (when using this switch, it may
150 be necessary to carry out a cold reboot _twice_ in a
151 row to make it take effect on the platform firmware).
153 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
154 acpi_osi="string1" # add string1
155 acpi_osi="!string2" # remove string2
156 acpi_osi=!* # remove all strings
157 acpi_osi=! # disable all built-in OS vendor
159 acpi_osi=!! # enable all built-in OS vendor
161 acpi_osi= # disable all strings
163 'acpi_osi=!' can be used in combination with single or
164 multiple 'acpi_osi="string1"' to support specific OS
165 vendor string(s). Note that such command can only
166 affect the default state of the OS vendor strings, thus
167 it cannot affect the default state of the feature group
168 strings and the current state of the OS vendor strings,
169 specifying it multiple times through kernel command line
170 is meaningless. This command is useful when one do not
171 care about the state of the feature group strings which
172 should be controlled by the OSPM.
174 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
175 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
176 can make '_OSI("Windows 2000")' TRUE.
178 'acpi_osi=' cannot be used in combination with other
179 'acpi_osi=' command lines, the _OSI method will not
180 exist in the ACPI namespace. NOTE that such command can
181 only affect the _OSI support state, thus specifying it
182 multiple times through kernel command line is also
185 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
188 'acpi_osi=!*' can be used in combination with single or
189 multiple 'acpi_osi="string1"' to support specific
190 string(s). Note that such command can affect the
191 current state of both the OS vendor strings and the
192 feature group strings, thus specifying it multiple times
193 through kernel command line is meaningful. But it may
194 still not able to affect the final state of a string if
195 there are quirks related to this string. This command
196 is useful when one want to control the state of the
197 feature group strings to debug BIOS issues related to
200 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
201 '_OSI("Module Device")' FALSE.
202 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
203 '_OSI("Module Device")' TRUE.
204 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
206 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
208 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
209 they all will make '_OSI("Windows 2000")' TRUE.
212 Override the pmtimer bug detection: force the kernel
213 to assume that this machine's pmtimer latches its value
214 and always returns good values.
216 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
217 Format: { level | edge | high | low }
219 acpi_skip_timer_override [HW,ACPI]
220 Recognize and ignore IRQ0/pin2 Interrupt Override.
221 For broken nForce2 BIOS resulting in XT-PIC timer.
223 acpi_sleep= [HW,ACPI] Sleep options
224 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
225 old_ordering, nonvs, sci_force_enable, nobl }
226 See Documentation/power/video.rst for information on
228 s3_beep is for debugging; it makes the PC's speaker beep
229 as soon as the kernel's real-mode entry point is called.
230 s4_nohwsig prevents ACPI hardware signature from being
231 used during resume from hibernation.
232 old_ordering causes the ACPI 1.0 ordering of the _PTS
233 control method, with respect to putting devices into
234 low power states, to be enforced (the ACPI 2.0 ordering
235 of _PTS is used by default).
236 nonvs prevents the kernel from saving/restoring the
237 ACPI NVS memory during suspend/hibernation and resume.
238 sci_force_enable causes the kernel to set SCI_EN directly
239 on resume from S1/S3 (which is against the ACPI spec,
240 but some broken systems don't work without it).
241 nobl causes the internal blacklist of systems known to
242 behave incorrectly in some ways with respect to system
243 suspend and resume to be ignored (use wisely).
245 acpi_use_timer_override [HW,ACPI]
246 Use timer override. For some broken Nvidia NF5 boards
247 that require a timer override, but don't have HPET
249 add_efi_memmap [EFI; X86] Include EFI memory map in
250 kernel's map of available physical RAM.
253 { off | try_unsupported }
254 off: disable AGP support
255 try_unsupported: try to drive unsupported chipsets
256 (may crash computer or cause data corruption)
259 See Documentation/sound/alsa-configuration.rst
262 Allow the default userspace alignment fault handler
263 behaviour to be specified. Bit 0 enables warnings,
264 bit 1 enables fixups, and bit 2 sends a segfault.
266 align_va_addr= [X86-64]
267 Align virtual addresses by clearing slice [14:12] when
268 allocating a VMA at process creation time. This option
269 gives you up to 3% performance improvement on AMD F15h
270 machines (where it is enabled by default) for a
271 CPU-intensive style benchmark, and it can vary highly in
272 a microbenchmark depending on workload and compiler.
274 32: only for 32-bit processes
275 64: only for 64-bit processes
276 on: enable for both 32- and 64-bit processes
277 off: disable for both 32- and 64-bit processes
279 alloc_snapshot [FTRACE]
280 Allocate the ftrace snapshot buffer on boot up when the
281 main buffer is allocated. This is handy if debugging
282 and you need to use tracing_snapshot() on boot up, and
283 do not want to use tracing_snapshot_alloc() as it needs
284 to be done where GFP_KERNEL allocations are allowed.
286 amd_iommu= [HW,X86-64]
287 Pass parameters to the AMD IOMMU driver in the system.
289 fullflush - enable flushing of IO/TLB entries when
290 they are unmapped. Otherwise they are
291 flushed before they will be reused, which
293 off - do not initialize any AMD IOMMU found in
295 force_isolation - Force device isolation for all
296 devices. The IOMMU driver is not
297 allowed anymore to lift isolation
298 requirements as needed. This option
299 does not override iommu=pt
301 amd_iommu_dump= [HW,X86-64]
302 Enable AMD IOMMU driver option to dump the ACPI table
303 for AMD IOMMU. With this option enabled, AMD IOMMU
304 driver will print ACPI tables for AMD IOMMU during
305 IOMMU initialization.
307 amd_iommu_intr= [HW,X86-64]
308 Specifies one of the following AMD IOMMU interrupt
310 legacy - Use legacy interrupt remapping mode.
311 vapic - Use virtual APIC mode, which allows IOMMU
312 to inject interrupts directly into guest.
313 This mode requires kvm-amd.avic=1.
314 (Default when IOMMU HW support is present.)
316 amijoy.map= [HW,JOY] Amiga joystick support
317 Map of devices attached to JOY0DAT and JOY1DAT
319 See also Documentation/input/joydev/joystick.rst
321 analog.map= [HW,JOY] Analog joystick and gamepad support
322 Specifies type or capabilities of an analog joystick
323 connected to one of 16 gameports
324 Format: <type1>,<type2>,..<type16>
327 Power management functions (SPARCstation-4/5 + deriv.)
329 Disable APC CPU standby support. SPARCstation-Fox does
330 not play well with APC CPU idle - disable it if you have
331 APC and your system crashes randomly.
333 apic= [APIC,X86] Advanced Programmable Interrupt Controller
334 Change the output verbosity while booting
335 Format: { quiet (default) | verbose | debug }
336 Change the amount of debugging information output
337 when initialising the APIC and IO-APIC components.
338 For X86-32, this can also be used to specify an APIC
340 Format: apic=driver_name
341 Examples: apic=bigsmp
343 apic_extnmi= [APIC,X86] External NMI delivery setting
344 Format: { bsp (default) | all | none }
345 bsp: External NMI is delivered only to CPU 0
346 all: External NMIs are broadcast to all CPUs as a
348 none: External NMI is masked for all CPUs. This is
349 useful so that a dump capture kernel won't be
353 See Documentation/networking/ipv6.txt.
355 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
356 Limit apic dumping. The parameter defines the maximal
357 number of local apics being dumped. Also it is possible
358 to set it to "all" by meaning -- no limit here.
359 Format: { 1 (default) | 2 | ... | all }.
360 The parameter valid if only apic=debug or
361 apic=verbose is specified.
362 Example: apic=debug show_lapic=all
364 apm= [APM] Advanced Power Management
365 See header of arch/x86/kernel/apm_32.c.
367 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
368 Format: <io>,<irq>,<nodeID>
372 atarimouse= [HW,MOUSE] Atari Mouse
374 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
375 EzKey and similar keyboards
377 atkbd.reset= [HW] Reset keyboard during initialization
379 atkbd.set= [HW] Select keyboard code set
380 Format: <int> (2 = AT (default), 3 = PS/2)
382 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
385 atkbd.softraw= [HW] Choose between synthetic and real raw mode
386 Format: <bool> (0 = real, 1 = synthetic (default))
388 atkbd.softrepeat= [HW]
389 Use software keyboard repeat
391 audit= [KNL] Enable the audit sub-system
392 Format: { "0" | "1" | "off" | "on" }
393 0 | off - kernel audit is disabled and can not be
394 enabled until the next reboot
395 unset - kernel audit is initialized but disabled and
396 will be fully enabled by the userspace auditd.
397 1 | on - kernel audit is initialized and partially
398 enabled, storing at most audit_backlog_limit
399 messages in RAM until it is fully enabled by the
403 audit_backlog_limit= [KNL] Set the audit queue size limit.
404 Format: <int> (must be >=0)
407 bau= [X86_UV] Enable the BAU on SGI UV. The default
408 behavior is to disable the BAU (i.e. bau=0).
409 Format: { "0" | "1" }
412 unset - Disable the BAU.
414 baycom_epp= [HW,AX25]
417 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
419 See header of drivers/net/hamradio/baycom_par.c.
421 baycom_ser_fdx= [HW,AX25]
422 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
423 Format: <io>,<irq>,<mode>[,<baud>]
424 See header of drivers/net/hamradio/baycom_ser_fdx.c.
426 baycom_ser_hdx= [HW,AX25]
427 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
428 Format: <io>,<irq>,<mode>
429 See header of drivers/net/hamradio/baycom_ser_hdx.c.
431 blkdevparts= Manual partition parsing of block device(s) for
432 embedded devices based on command line input.
433 See Documentation/block/cmdline-partition.rst
435 boot_delay= Milliseconds to delay each printk during boot.
436 Values larger than 10 seconds (10000) are changed to
441 Extended command line options can be added to an initrd
442 and this will cause the kernel to look for it.
444 See Documentation/admin-guide/bootconfig.rst
447 Disable BERT OS support on buggy BIOSes.
449 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
450 bttv.radio= Most important insmod options are available as
452 bttv.pll= See Documentation/media/v4l-drivers/bttv.rst
455 bulk_remove=off [PPC] This parameter disables the use of the pSeries
456 firmware feature for flushing multiple hpte entries
459 c101= [NET] Moxa C101 synchronous serial card
461 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
462 Sometimes CPU hardware bugs make them report the cache
463 size incorrectly. The kernel will attempt work arounds
464 to fix known problems, but for some CPUs it is not
465 possible to determine what the correct size should be.
466 This option provides an override for these situations.
469 [NET] Specifies amount of time (in seconds) that
470 the kernel should wait for a network carrier. By default
471 it waits 120 seconds.
473 ca_keys= [KEYS] This parameter identifies a specific key(s) on
474 the system trusted keyring to be used for certificate
476 format: { id:<keyid> | builtin }
478 cca= [MIPS] Override the kernel pages' cache coherency
479 algorithm. Accepted values range from 0 to 7
480 inclusive. See arch/mips/include/asm/pgtable-bits.h
481 for platform specific values (SB1, Loongson3 and
484 ccw_timeout_log [S390]
485 See Documentation/s390/common_io.rst for details.
487 cgroup_disable= [KNL] Disable a particular controller
488 Format: {name of the controller(s) to disable}
489 The effects of cgroup_disable=foo are:
490 - foo isn't auto-mounted if you mount all cgroups in
492 - foo isn't visible as an individually mountable
494 {Currently only "memory" controller deal with this and
495 cut the overhead, others just disable the usage. So
496 only cgroup_disable=memory is actually worthy}
498 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
499 Format: { { controller | "all" | "named" }
500 [,{ controller | "all" | "named" }...] }
501 Like cgroup_disable, but only applies to cgroup v1;
502 the blacklisted controllers remain available in cgroup2.
503 "all" blacklists all controllers and "named" disables
504 named mounts. Specifying both "all" and "named" disables
507 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
509 nosocket -- Disable socket memory accounting.
510 nokmem -- Disable kernel memory accounting.
512 checkreqprot [SELINUX] Set initial checkreqprot flag value.
513 Format: { "0" | "1" }
514 See security/selinux/Kconfig help text.
515 0 -- check protection applied by kernel (includes
516 any implied execute protection).
517 1 -- check protection requested by application.
518 Default value is set via a kernel config option.
519 Value can be changed at runtime via
520 /sys/fs/selinux/checkreqprot.
523 See Documentation/s390/common_io.rst for details.
526 Prevents the clock framework from automatically gating
527 clocks that have not been explicitly enabled by a Linux
528 device driver but are enabled in hardware at reset or
529 by the bootloader/firmware. Note that this does not
530 force such clocks to be always-on nor does it reserve
531 those clocks in any way. This parameter is useful for
532 debug and development, but should not be needed on a
533 platform with proper driver support. For more
534 information, see Documentation/driver-api/clk.rst.
536 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
538 Forces specified clocksource (if available) to be used
539 when calculating gettimeofday(). If specified
540 clocksource is not available, it defaults to PIT.
541 Format: { pit | tsc | cyclone | pmtmr }
543 clocksource= Override the default clocksource
545 Override the default clocksource and use the clocksource
546 with the name specified.
547 Some clocksource names to choose from, depending on
549 [all] jiffies (this is the base, fallback clocksource)
551 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
552 pxa_timer,timer3,32k_counter,timer0_1
553 [X86-32] pit,hpet,tsc;
554 scx200_hrt on Geode; cyclone on IBM x440
562 clocksource.arm_arch_timer.evtstrm=
565 Enable/disable the eventstream feature of the ARM
566 architected timer so that code using WFE-based polling
567 loops can be debugged more effectively on production
570 clearcpuid=BITNUM [X86]
571 Disable CPUID feature X for the kernel. See
572 arch/x86/include/asm/cpufeatures.h for the valid bit
573 numbers. Note the Linux specific bits are not necessarily
574 stable over kernel options, but the vendor specific
576 Also note that user programs calling CPUID directly
577 or using the feature without checking anything
578 will still see it. This just prevents it from
579 being used by the kernel or shown in /proc/cpuinfo.
580 Also note the kernel might malfunction if you disable
583 cma=nn[MG]@[start[MG][-end[MG]]]
585 Sets the size of kernel global memory area for
586 contiguous memory allocations and optionally the
587 placement constraint by the physical address range of
588 memory allocations. A value of 0 disables CMA
589 altogether. For more information, see
590 include/linux/dma-contiguous.h
592 cmo_free_hint= [PPC] Format: { yes | no }
593 Specify whether pages are marked as being inactive
594 when they are freed. This is used in CMO environments
595 to determine OS memory pressure for page stealing by
599 coherent_pool=nn[KMG] [ARM,KNL]
600 Sets the size of memory pool for coherent, atomic dma
601 allocations, by default set to 256K.
603 com20020= [HW,NET] ARCnet - COM20020 chipset
605 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
607 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
611 ARCnet - COM90xx chipset (memory-mapped buffers)
612 Format: <io>[,<irq>[,<memstart>]]
614 condev= [HW,S390] console device
617 console= [KNL] Output console device and options.
619 tty<n> Use the virtual console device <n>.
623 Use the specified serial port. The options are of
624 the form "bbbbpnf", where "bbbb" is the baud rate,
625 "p" is parity ("n", "o", or "e"), "n" is number of
626 bits, and "f" is flow control ("r" for RTS or
627 omit it). Default is "9600n8".
629 See Documentation/admin-guide/serial-console.rst for more
631 Documentation/networking/netconsole.txt for an
634 uart[8250],io,<addr>[,options]
635 uart[8250],mmio,<addr>[,options]
636 uart[8250],mmio16,<addr>[,options]
637 uart[8250],mmio32,<addr>[,options]
638 uart[8250],0x<addr>[,options]
639 Start an early, polled-mode console on the 8250/16550
640 UART at the specified I/O port or MMIO address,
641 switching to the matching ttyS device later.
642 MMIO inter-register address stride is either 8-bit
643 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
644 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
645 to be equivalent to 'mmio'. 'options' are specified in
646 the same format described for ttyS above; if unspecified,
647 the h/w is not re-initialized.
649 hvc<n> Use the hypervisor console device <n>. This is for
650 both Xen and PowerPC hypervisors.
652 If the device connected to the port is not a TTY but a braille
653 device, prepend "brl," before the device type, for instance
655 For now, only VisioBraille is supported.
658 [KNL] Change console messages format
660 By default we print messages on consoles in
661 "[time stamp] text\n" format (time stamp may not be
662 printed, depending on CONFIG_PRINTK_TIME or
663 `printk_time' param).
665 Switch to syslog format: "<%u>[time stamp] text\n"
666 IOW, each message will have a facility and loglevel
667 prefix. The format is similar to one used by syslog()
668 syscall, or to executing "dmesg -S --raw" or to reading
671 consoleblank= [KNL] The console blank (screen saver) timeout in
672 seconds. A value of 0 disables the blank timer.
676 [KNL] Change the default value for
677 /proc/<pid>/coredump_filter.
678 See also Documentation/filesystems/proc.txt.
680 coresight_cpu_debug.enable
683 Enable/disable the CPU sampling based debugging.
684 0: default value, disable debugging
685 1: enable debugging at boot time
687 cpuidle.off=1 [CPU_IDLE]
688 disable the cpuidle sub-system
691 [CPU_IDLE] Name of the cpuidle governor to use.
693 cpufreq.off=1 [CPU_FREQ]
694 disable the cpufreq sub-system
697 [X86] Delay for N microsec between assert and de-assert
698 of APIC INIT to start processors. This delay occurs
699 on every CPU online, such as boot, and resume from suspend.
702 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
704 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
706 crashkernel=size[KMG][@offset[KMG]]
707 [KNL] Using kexec, Linux can switch to a 'crash kernel'
708 upon panic. This parameter reserves the physical
709 memory region [offset, offset + size] for that kernel
710 image. If '@offset' is omitted, then a suitable offset
711 is selected automatically.
712 [KNL, x86_64] select a region under 4G first, and
713 fall back to reserve region above 4G when '@offset'
714 hasn't been specified.
715 See Documentation/admin-guide/kdump/kdump.rst for further details.
717 crashkernel=range1:size1[,range2:size2,...][@offset]
718 [KNL] Same as above, but depends on the memory
719 in the running system. The syntax of range is
720 start-[end] where start and end are both
721 a memory unit (amount[KMG]). See also
722 Documentation/admin-guide/kdump/kdump.rst for an example.
724 crashkernel=size[KMG],high
725 [KNL, x86_64] range could be above 4G. Allow kernel
726 to allocate physical memory region from top, so could
727 be above 4G if system have more than 4G ram installed.
728 Otherwise memory region will be allocated below 4G, if
730 It will be ignored if crashkernel=X is specified.
731 crashkernel=size[KMG],low
732 [KNL, x86_64] range under 4G. When crashkernel=X,high
733 is passed, kernel could allocate physical memory region
734 above 4G, that cause second kernel crash on system
735 that require some amount of low memory, e.g. swiotlb
736 requires at least 64M+32K low memory, also enough extra
737 low memory is needed to make sure DMA buffers for 32-bit
738 devices won't run out. Kernel would try to allocate at
739 at least 256M below 4G automatically.
740 This one let user to specify own low range under 4G
741 for second kernel instead.
742 0: to disable low allocation.
743 It will be ignored when crashkernel=X,high is not used
744 or memory reserved is below 4G.
747 [KNL] Disable crypto self-tests
752 cs89x0_media= [HW,NET]
753 Format: { rj45 | aui | bnc }
756 See header of drivers/s390/block/dasd_devmap.c.
758 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
759 (one device per port)
760 Format: <port#>,<type>
761 See also Documentation/input/devices/joystick-parport.rst
763 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
765 Documentation/admin-guide/dynamic-debug-howto.rst for
766 details. Deprecated, see dyndbg.
768 debug [KNL] Enable kernel debugging (events log level).
771 [KNL] Enable printing [hashed] pointers early in the
772 boot sequence. If enabled, we use a weak hash instead
773 of siphash to hash pointers. Use this option if you are
774 seeing instances of '(___ptrval___)') and need to see a
775 value (hashed pointer) instead. Cryptographically
776 insecure, please do not use on production kernels.
779 [KNL] verbose self-tests
781 Print debugging info while doing the locking API
783 We default to 0 (no extra messages), setting it to
784 1 will print _a lot_ more information - normally
785 only useful to kernel developers.
787 debug_objects [KNL] Enable object debugging
790 [KNL] Disable object debugging
792 debug_guardpage_minorder=
793 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
794 parameter allows control of the order of pages that will
795 be intentionally kept free (and hence protected) by the
796 buddy allocator. Bigger value increase the probability
797 of catching random memory corruption, but reduce the
798 amount of memory for normal system use. The maximum
799 possible value is MAX_ORDER/2. Setting this parameter
800 to 1 or 2 should be enough to identify most random
801 memory corruption problems caused by bugs in kernel or
802 driver code when a CPU writes to (or reads from) a
803 random memory location. Note that there exists a class
804 of memory corruptions problems caused by buggy H/W or
805 F/W or by drivers badly programing DMA (basically when
806 memory is written at bus level and the CPU MMU is
807 bypassed) which are not detectable by
808 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
809 tracking down these problems.
812 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
813 enables the feature at boot time. By default, it is
814 disabled and the system will work mostly the same as a
815 kernel built without CONFIG_DEBUG_PAGEALLOC.
816 Note: to get most of debug_pagealloc error reports, it's
817 useful to also enable the page_owner functionality.
818 on: enable the feature
820 debugpat [X86] Enable PAT debugging
822 decnet.addr= [HW,NET]
823 Format: <area>[,<node>]
824 See also Documentation/networking/decnet.txt.
827 [same as hugepagesz=] The size of the default
828 HugeTLB page size. This is the size represented by
829 the legacy /proc/ hugepages APIs, used for SHM, and
830 default size when mounting hugetlbfs filesystems.
831 Defaults to the default architecture's huge page size
834 deferred_probe_timeout=
835 [KNL] Debugging option to set a timeout in seconds for
836 deferred probe to give up waiting on dependencies to
837 probe. Only specific dependencies (subsystems or
838 drivers) that have opted in will be ignored. A timeout of 0
839 will timeout at the end of initcalls. This option will also
840 dump out devices still on the deferred probe list after
844 Format: { on | off | def_only | inf_only | always }
845 on: s390 zlib hardware support for compression on
846 level 1 and decompression (default)
847 off: No s390 zlib hardware support
848 def_only: s390 zlib hardware support for deflate
849 only (compression on level 1)
850 inf_only: s390 zlib hardware support for inflate
852 always: Same as 'on' but ignores the selected compression
853 level always using hardware support (used for debugging)
856 Set number of hash buckets for dentry cache.
858 disable_1tb_segments [PPC]
859 Disables the use of 1TB hash page table segments. This
860 causes the kernel to fall back to 256MB segments which
861 can be useful when debugging issues that require an SLB
865 See Documentation/networking/ipv6.txt.
868 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
869 hardening is enabled for this boot. Hardened
870 usercopy checking is used to protect the kernel
871 from reading or writing beyond known memory
872 allocation boundaries as a proactive defense
873 against bounds-checking flaws in the kernel's
874 copy_to_user()/copy_from_user() interface.
875 on Perform hardened usercopy checks (default).
876 off Disable hardened usercopy checks.
879 Disable RADIX MMU mode on POWER9
882 Disable TLBIE instruction. Currently does not work
883 with KVM, with HASH MMU, or with coherent accelerators.
885 disable_cpu_apicid= [X86,APIC,SMP]
887 The number of initial APIC ID for the
888 corresponding CPU to be disabled at boot,
889 mostly used for the kdump 2nd kernel to
890 disable BSP to wake up multiple CPUs without
891 causing system reset or hang due to sending
894 perf_v4_pmi= [X86,INTEL]
896 Disable Intel PMU counter freezing feature.
897 The feature only exists starting from
898 Arch Perfmon v4 (Skylake and newer).
900 disable_ddw [PPC/PSERIES]
901 Disable Dynamic DMA Window support. Use this if
902 to workaround buggy firmware.
905 See Documentation/networking/ipv6.txt.
907 disable_mtrr_cleanup [X86]
908 The kernel tries to adjust MTRR layout from continuous
909 to discrete, to make X server driver able to add WB
910 entry later. This parameter disables that.
912 disable_mtrr_trim [X86, Intel and AMD only]
913 By default the kernel will trim any uncacheable
914 memory out of your available memory pool based on
915 MTRR settings. This parameter disables that behavior,
916 possibly causing your machine to run very slowly.
918 disable_timer_pin_1 [X86]
919 Disable PIN 1 of APIC timer
920 Can be useful to work around chipset bugs.
922 dis_ucode_ldr [X86] Disable the microcode loader.
924 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
925 this option disables the debugging code at boot.
927 dma_debug_entries=<number>
928 This option allows to tune the number of preallocated
929 entries for DMA-API debugging code. One entry is
930 required per DMA-API allocation. Use this if the
931 DMA-API debugging code disables itself because the
932 architectural default is too low.
934 dma_debug_driver=<driver_name>
935 With this option the DMA-API debugging driver
936 filter feature can be enabled at boot time. Just
937 pass the driver to filter for as the parameter.
938 The filter can be disabled or changed to another
939 driver later using sysfs.
941 driver_async_probe= [KNL]
942 List of driver names to be probed asynchronously.
943 Format: <driver_name1>,<driver_name2>...
945 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
946 Broken monitors, graphic adapters, KVMs and EDIDless
947 panels may send no or incorrect EDID data sets.
948 This parameter allows to specify an EDID data sets
949 in the /lib/firmware directory that are used instead.
950 Generic built-in EDID data sets are used, if one of
951 edid/1024x768.bin, edid/1280x1024.bin,
952 edid/1680x1050.bin, or edid/1920x1080.bin is given
953 and no file with the same name exists. Details and
954 instructions how to build your own EDID data are
955 available in Documentation/driver-api/edid.rst. An EDID
956 data set will only be used for a particular connector,
957 if its name and a colon are prepended to the EDID
958 name. Each connector may use a unique EDID data
959 set by separating the files with a comma. An EDID
960 data set with no connector name will be used for
961 any connectors not explicitly specified.
966 Format: {"off" | "known"}
967 Control how the dt_cpu_ftrs device-tree binding is
968 used for CPU feature discovery and setup (if it
970 off: Do not use it, fall back to legacy cpu table.
971 known: Do not pass through unknown features to guests
972 or userspace, only those that the kernel is aware of.
974 dump_apple_properties [X86]
975 Dump name and content of EFI device properties on
976 x86 Macs. Useful for driver authors to determine
977 what data is available or for reverse-engineering.
979 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
980 module.dyndbg[="val"]
981 Enable debug messages at boot time. See
982 Documentation/admin-guide/dynamic-debug-howto.rst
985 nompx [X86] Disables Intel Memory Protection Extensions.
986 See Documentation/x86/intel_mpx.rst for more
987 information about the feature.
989 nopku [X86] Disable Memory Protection Keys CPU feature found
992 module.async_probe [KNL]
993 Enable asynchronous probe on this module.
995 early_ioremap_debug [KNL]
996 Enable debug messages in early_ioremap support. This
997 is useful for tracking down temporary early mappings
998 which are not unmapped.
1000 earlycon= [KNL] Output early console device and options.
1002 When used with no options, the early console is
1003 determined by stdout-path property in device tree's
1004 chosen node or the ACPI SPCR table if supported by
1007 cdns,<addr>[,options]
1008 Start an early, polled-mode console on a Cadence
1009 (xuartps) serial port at the specified address. Only
1010 supported option is baud rate. If baud rate is not
1011 specified, the serial port must already be setup and
1014 uart[8250],io,<addr>[,options]
1015 uart[8250],mmio,<addr>[,options]
1016 uart[8250],mmio32,<addr>[,options]
1017 uart[8250],mmio32be,<addr>[,options]
1018 uart[8250],0x<addr>[,options]
1019 Start an early, polled-mode console on the 8250/16550
1020 UART at the specified I/O port or MMIO address.
1021 MMIO inter-register address stride is either 8-bit
1022 (mmio) or 32-bit (mmio32 or mmio32be).
1023 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1024 to be equivalent to 'mmio'. 'options' are specified
1025 in the same format described for "console=ttyS<n>"; if
1026 unspecified, the h/w is not initialized.
1030 Start an early, polled-mode console on a pl011 serial
1031 port at the specified address. The pl011 serial port
1032 must already be setup and configured. Options are not
1033 yet supported. If 'mmio32' is specified, then only
1034 the driver will use only 32-bit accessors to read/write
1035 the device registers.
1038 Start an early, polled-mode console on a meson serial
1039 port at the specified address. The serial port must
1040 already be setup and configured. Options are not yet
1044 Start an early, polled-mode console on an msm serial
1045 port at the specified address. The serial port
1046 must already be setup and configured. Options are not
1049 msm_serial_dm,<addr>
1050 Start an early, polled-mode console on an msm serial
1051 dm port at the specified address. The serial port
1052 must already be setup and configured. Options are not
1056 Start an early, polled-mode console on a serial port
1057 of an Actions Semi SoC, such as S500 or S900, at the
1058 specified address. The serial port must already be
1059 setup and configured. Options are not yet supported.
1062 Start an early, polled-mode console on a serial port
1063 of an RDA Micro SoC, such as RDA8810PL, at the
1064 specified address. The serial port must already be
1065 setup and configured. Options are not yet supported.
1068 Use RISC-V SBI (Supervisor Binary Interface) for early
1071 smh Use ARM semihosting calls for early console.
1079 Use early console provided by serial driver available
1080 on Samsung SoCs, requires selecting proper type and
1081 a correct base address of the selected UART port. The
1082 serial port must already be setup and configured.
1083 Options are not yet supported.
1086 Start an early, polled-mode console on a lantiq serial
1087 (lqasc) port at the specified address. The serial port
1088 must already be setup and configured. Options are not
1093 Use early console provided by Freescale LP UART driver
1094 found on Freescale Vybrid and QorIQ LS1021A processors.
1095 A valid base address must be provided, and the serial
1096 port must already be setup and configured.
1099 Start an early, polled-mode console on the
1100 Armada 3700 serial port at the specified
1101 address. The serial port must already be setup
1102 and configured. Options are not yet supported.
1105 Start an early, polled-mode console on a Qualcomm
1106 Generic Interface (GENI) based serial port at the
1107 specified address. The serial port must already be
1108 setup and configured. Options are not yet supported.
1111 Start an early, unaccelerated console on the EFI
1112 memory mapped framebuffer (if available). On cache
1113 coherent non-x86 systems that use system memory for
1114 the framebuffer, pass the 'ram' option so that it is
1115 mapped with the correct attributes.
1118 Use early console provided by Freescale LINFlexD UART
1119 serial driver for NXP S32V234 SoCs. A valid base
1120 address must be provided, and the serial port must
1121 already be setup and configured.
1123 earlyprintk= [X86,SH,ARM,M68k,S390]
1127 earlyprintk=serial[,ttySn[,baudrate]]
1128 earlyprintk=serial[,0x...[,baudrate]]
1129 earlyprintk=ttySn[,baudrate]
1130 earlyprintk=dbgp[debugController#]
1131 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1132 earlyprintk=xdbc[xhciController#]
1134 earlyprintk is useful when the kernel crashes before
1135 the normal console is initialized. It is not enabled by
1136 default because it has some cosmetic problems.
1138 Append ",keep" to not disable it when the real console
1141 Only one of vga, efi, serial, or usb debug port can
1144 Currently only ttyS0 and ttyS1 may be specified by
1145 name. Other I/O ports may be explicitly specified
1146 on some architectures (x86 and arm at least) by
1147 replacing ttySn with an I/O port address, like this:
1148 earlyprintk=serial,0x1008,115200
1149 You can find the port for a given device in
1150 /proc/tty/driver/serial:
1151 2: uart:ST16650V2 port:00001008 irq:18 ...
1153 Interaction with the standard serial driver is not
1156 The VGA and EFI output is eventually overwritten by
1159 The xen output can only be used by Xen PV guests.
1161 The sclp output can only be used on s390.
1163 The optional "force" to "pciserial" enables use of a
1164 PCI device even when its classcode is not of the
1167 edac_report= [HW,EDAC] Control how to report EDAC event
1168 Format: {"on" | "off" | "force"}
1169 on: enable EDAC to report H/W event. May be overridden
1170 by other higher priority error reporting module.
1171 off: disable H/W event reporting through EDAC.
1172 force: enforce the use of EDAC to report H/W event.
1175 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1178 This is designed to be used in conjunction with
1179 the boot argument: earlyprintk=vga
1182 Format: {"off" | "on" | "skip[mbr]"}
1185 Format: { "old_map", "nochunk", "noruntime", "debug",
1186 "nosoftreserve", "disable_early_pci_dma",
1187 "no_disable_early_pci_dma" }
1188 old_map [X86-64]: switch to the old ioremap-based EFI
1189 runtime services mapping. [Needs CONFIG_X86_UV=y]
1190 nochunk: disable reading files in "chunks" in the EFI
1191 boot stub, as chunking can cause problems with some
1192 firmware implementations.
1193 noruntime : disable EFI runtime services support
1194 debug: enable misc debug output
1195 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1196 attribute may cause the kernel to reserve the
1197 memory range for a memory mapping driver to
1198 claim. Specify efi=nosoftreserve to disable this
1199 reservation and treat the memory by its base type
1200 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1201 disable_early_pci_dma: Disable the busmaster bit on all
1202 PCI bridges while in the EFI boot stub
1203 no_disable_early_pci_dma: Leave the busmaster bit set
1204 on all PCI bridges while in the EFI boot stub
1206 efi_no_storage_paranoia [EFI; X86]
1207 Using this parameter you can use more than 50% of
1208 your efi variable storage. Use this parameter only if
1209 you are really sure that your UEFI does sane gc and
1210 fulfills the spec otherwise your board may brick.
1212 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1213 Add arbitrary attribute to specific memory range by
1214 updating original EFI memory map.
1215 Region of memory which aa attribute is added to is
1218 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1219 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1220 attribute is added to range 0x100000000-0x180000000 and
1221 0x10a0000000-0x1120000000.
1223 If efi_fake_mem=8G@9G:0x40000 is specified, the
1224 EFI_MEMORY_SP(0x40000) attribute is added to
1225 range 0x240000000-0x43fffffff.
1227 Using this parameter you can do debugging of EFI memmap
1228 related features. For example, you can do debugging of
1229 Address Range Mirroring feature even if your box
1230 doesn't support it, or mark specific memory as
1233 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1234 that is to be dynamically loaded by Linux. If there are
1235 multiple variables with the same name but with different
1236 vendor GUIDs, all of them will be loaded. See
1237 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1240 eisa_irq_edge= [PARISC,HW]
1241 See header of drivers/parisc/eisa.c.
1244 See comment before function elanfreq_setup() in
1245 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1247 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1248 Specifies physical address of start of kernel core
1249 image elf header and optionally the size. Generally
1250 kexec loader will pass this option to capture kernel.
1251 See Documentation/admin-guide/kdump/kdump.rst for details.
1253 enable_mtrr_cleanup [X86]
1254 The kernel tries to adjust MTRR layout from continuous
1255 to discrete, to make X server driver able to add WB
1256 entry later. This parameter enables that.
1258 enable_timer_pin_1 [X86]
1259 Enable PIN 1 of APIC timer
1260 Can be useful to work around chipset bugs
1261 (in particular on some ATI chipsets).
1262 The kernel tries to set a reasonable default.
1264 enforcing [SELINUX] Set initial enforcing status.
1266 See security/selinux/Kconfig help text.
1267 0 -- permissive (log only, no denials).
1268 1 -- enforcing (deny and log).
1270 Value can be changed at runtime via
1271 /sys/fs/selinux/enforce.
1274 Disable Error Record Serialization Table (ERST)
1277 ether= [HW,NET] Ethernet cards parameters
1278 This option is obsoleted by the "netdev=" option, which
1279 has equivalent usage. See its documentation for details.
1283 Permit 'security.evm' to be updated regardless of
1284 current integrity status.
1288 fail_make_request=[KNL]
1289 General fault injection mechanism.
1290 Format: <interval>,<probability>,<space>,<times>
1291 See also Documentation/fault-injection/.
1294 See Documentation/admin-guide/blockdev/floppy.rst.
1296 force_pal_cache_flush
1297 [IA-64] Avoid check_sal_cache_flush which may hang on
1298 buggy SAL_CACHE_FLUSH implementations. Using this
1299 parameter will force ia64_sal_cache_flush to call
1300 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1303 Forcefully enable Physical Address Extension (PAE).
1304 Many Pentium M systems disable PAE but may have a
1305 functionally usable PAE implementation.
1306 Warning: use of this parameter will taint the kernel
1307 and may cause unknown problems.
1310 [FTRACE] will set and start the specified tracer
1311 as early as possible in order to facilitate early
1314 ftrace_dump_on_oops[=orig_cpu]
1315 [FTRACE] will dump the trace buffers on oops.
1316 If no parameter is passed, ftrace will dump
1317 buffers of all CPUs, but if you pass orig_cpu, it will
1318 dump only the buffer of the CPU that triggered the
1321 ftrace_filter=[function-list]
1322 [FTRACE] Limit the functions traced by the function
1323 tracer at boot up. function-list is a comma separated
1324 list of functions. This list can be changed at run
1325 time by the set_ftrace_filter file in the debugfs
1328 ftrace_notrace=[function-list]
1329 [FTRACE] Do not trace the functions specified in
1330 function-list. This list can be changed at run time
1331 by the set_ftrace_notrace file in the debugfs
1334 ftrace_graph_filter=[function-list]
1335 [FTRACE] Limit the top level callers functions traced
1336 by the function graph tracer at boot up.
1337 function-list is a comma separated list of functions
1338 that can be changed at run time by the
1339 set_graph_function file in the debugfs tracing directory.
1341 ftrace_graph_notrace=[function-list]
1342 [FTRACE] Do not trace from the functions specified in
1343 function-list. This list is a comma separated list of
1344 functions that can be changed at run time by the
1345 set_graph_notrace file in the debugfs tracing directory.
1347 ftrace_graph_max_depth=<uint>
1348 [FTRACE] Used with the function graph tracer. This is
1349 the max depth it will trace into a function. This value
1350 can be changed at run time by the max_graph_depth file
1351 in the tracefs tracing directory. default: 0 (no limit)
1354 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1355 support via parallel port (up to 5 devices per port)
1356 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1357 See also Documentation/input/devices/joystick-parport.rst
1361 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1365 gcov_persist= [GCOV] When non-zero (default), profiling data for
1366 kernel modules is saved and remains accessible via
1367 debugfs, even when the module is unloaded/reloaded.
1368 When zero, profiling data is discarded and associated
1369 debugfs files are removed at module unload time.
1371 goldfish [X86] Enable the goldfish android emulator platform.
1372 Don't use this when you are not running on the
1375 gpt [EFI] Forces disk with valid GPT signature but
1376 invalid Protective MBR to be treated as GPT. If the
1377 primary GPT is corrupted, it enables the backup/alternate
1378 GPT to be used instead.
1380 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1381 the "Enable 0" bit of the configuration register.
1384 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1385 the "Enable 0" bit of the configuration register.
1388 grcan.select= [HW] Select which physical interface to use.
1391 grcan.txsize= [HW] Sets the size of the tx buffer.
1392 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1394 grcan.rxsize= [HW] Sets the size of the rx buffer.
1395 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1398 gpio-mockup.gpio_mockup_ranges
1399 [HW] Sets the ranges of gpiochip of for this device.
1400 Format: <start1>,<end1>,<start2>,<end2>...
1402 hardlockup_all_cpu_backtrace=
1403 [KNL] Should the hard-lockup detector generate
1404 backtraces on all cpus.
1407 hashdist= [KNL,NUMA] Large hashes allocated during boot
1408 are distributed across NUMA nodes. Defaults on
1409 for 64-bit NUMA, off otherwise.
1410 Format: 0 | 1 (for off | on)
1412 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1414 hd= [EIDE] (E)IDE hard drive subsystem geometry
1415 Format: <cyl>,<head>,<sect>
1418 Disable Hardware Error Source Table (HEST) support;
1419 corresponding firmware-first mode error processing
1420 logic will be disabled.
1422 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1423 size of <nn>. This works even on boxes that have no
1424 highmem otherwise. This also works to reduce highmem
1425 size on bigger boxes.
1427 highres= [KNL] Enable/disable high resolution timer mode.
1428 Valid parameters: "on", "off"
1433 hpet= [X86-32,HPET] option to control HPET usage
1434 Format: { enable (default) | disable | force |
1436 disable: disable HPET and use PIT instead
1437 force: allow force enabled of undocumented chips (ICH4,
1439 verbose: show contents of HPET registers during setup
1441 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1442 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1444 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1445 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1446 On x86-64 and powerpc, this option can be specified
1447 multiple times interleaved with hugepages= to reserve
1448 huge pages of different sizes. Valid pages sizes on
1449 x86-64 are 2M (when the CPU supports "pse") and 1G
1450 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1453 [KNL] Should the hung task detector generate panics.
1456 A nonzero value instructs the kernel to panic when a
1457 hung task is detected. The default value is controlled
1458 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1459 option. The value selected by this boot parameter can
1460 be changed later by the kernel.hung_task_panic sysctl.
1462 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1463 terminal devices. Valid values: 0..8
1464 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1465 If specified, z/VM IUCV HVC accepts connections
1466 from listed z/VM user IDs only.
1468 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1469 which allow the hypervisor to 'idle' the
1470 guest on lock contention.
1473 Do not unregister boot console at start. This is only
1474 useful for debugging when something happens in the window
1475 between unregistering the boot console and initializing
1478 i2c_bus= [HW] Override the default board specific I2C bus speed
1479 or register an additional I2C bus that is not
1480 registered from board initialization code.
1484 i8042.debug [HW] Toggle i8042 debug mode
1485 i8042.unmask_kbd_data
1486 [HW] Enable printing of interrupt data from the KBD port
1487 (disabled by default, and as a pre-condition
1488 requires that i8042.debug=1 be enabled)
1489 i8042.direct [HW] Put keyboard port into non-translated mode
1490 i8042.dumbkbd [HW] Pretend that controller can only read data from
1491 keyboard and cannot control its state
1492 (Don't attempt to blink the leds)
1493 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1494 i8042.nokbd [HW] Don't check/create keyboard port
1495 i8042.noloop [HW] Disable the AUX Loopback command while probing
1497 i8042.nomux [HW] Don't check presence of an active multiplexing
1499 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1501 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1502 i8042.reset [HW] Reset the controller during init, cleanup and
1503 suspend-to-ram transitions, only during s2r
1504 transitions, or never reset
1505 Format: { 1 | Y | y | 0 | N | n }
1506 1, Y, y: always reset controller
1507 0, N, n: don't ever reset controller
1508 Default: only on s2r transitions on x86; most other
1509 architectures force reset to be always executed
1510 i8042.unlock [HW] Unlock (ignore) the keylock
1511 i8042.kbdreset [HW] Reset device connected to KBD port
1515 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1516 indicates that the driver is running on unsupported
1518 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1519 does not match list of supported models.
1521 [HW] Report power status in /proc/i8k
1522 (disabled by default)
1523 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1526 i915.invert_brightness=
1527 [DRM] Invert the sense of the variable that is used to
1528 set the brightness of the panel backlight. Normally a
1529 brightness value of 0 indicates backlight switched off,
1530 and the maximum of the brightness value sets the backlight
1531 to maximum brightness. If this parameter is set to 0
1532 (default) and the machine requires it, or this parameter
1533 is set to 1, a brightness value of 0 sets the backlight
1534 to maximum brightness, and the maximum of the brightness
1535 value switches the backlight off.
1536 -1 -- never invert brightness
1537 0 -- machine default
1538 1 -- force brightness inversion
1541 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1543 ide-core.nodma= [HW] (E)IDE subsystem
1544 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1545 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1546 .cdrom .chs .ignore_cable are additional options
1547 See Documentation/ide/ide.rst.
1549 ide-generic.probe-mask= [HW] (E)IDE subsystem
1551 Probe mask for legacy ISA IDE ports. Depending on
1552 platform up to 6 ports are supported, enabled by
1553 setting corresponding bits in the mask to 1. The
1554 default value is 0x0, which has a special meaning.
1555 On systems that have PCI, it triggers scanning the
1556 PCI bus for the first and the second port, which
1557 are then probed. On systems without PCI the value
1558 of 0x0 enables probing the two first ports as if it
1561 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1562 Claim all unknown PCI IDE storage controllers.
1565 Format: idle=poll, idle=halt, idle=nomwait
1566 Poll forces a polling idle loop that can slightly
1567 improve the performance of waking up a idle CPU, but
1568 will use a lot of power and make the system run hot.
1570 idle=halt: Halt is forced to be used for CPU idle.
1571 In such case C2/C3 won't be used again.
1572 idle=nomwait: Disable mwait for CPU C-states
1574 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1575 Format: { strict | legacy | 2008 | relaxed }
1578 Choose which programs will be accepted for execution
1579 based on the IEEE 754 NaN encoding(s) supported by
1580 the FPU and the NaN encoding requested with the value
1581 of an ELF file header flag individually set by each
1582 binary. Hardware implementations are permitted to
1583 support either or both of the legacy and the 2008 NaN
1586 Available settings are as follows:
1587 strict accept binaries that request a NaN encoding
1588 supported by the FPU
1589 legacy only accept legacy-NaN binaries, if supported
1591 2008 only accept 2008-NaN binaries, if supported
1593 relaxed accept any binaries regardless of whether
1594 supported by the FPU
1596 The FPU emulator is always able to support both NaN
1597 encodings, so if no FPU hardware is present or it has
1598 been disabled with 'nofpu', then the settings of
1599 'legacy' and '2008' strap the emulator accordingly,
1600 'relaxed' straps the emulator for both legacy-NaN and
1601 2008-NaN, whereas 'strict' enables legacy-NaN only on
1602 legacy processors and both NaN encodings on MIPS32 or
1605 The setting for ABS.fmt/NEG.fmt instruction execution
1606 mode generally follows that for the NaN encoding,
1607 except where unsupported by hardware.
1609 ignore_loglevel [KNL]
1610 Ignore loglevel setting - this will print /all/
1611 kernel messages to the console. Useful for debugging.
1612 We also add it as printk module parameter, so users
1613 could change it dynamically, usually by
1614 /sys/module/printk/parameters/ignore_loglevel.
1617 Ignore RLIMIT_DATA setting for data mappings,
1618 print warning at first misuse. Can be changed via
1619 /sys/module/kernel/parameters/ignore_rlimit_data.
1621 ihash_entries= [KNL]
1622 Set number of hash buckets for inode cache.
1624 ima_appraise= [IMA] appraise integrity measurements
1625 Format: { "off" | "enforce" | "fix" | "log" }
1628 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1629 The builtin appraise policy appraises all files
1632 ima_canonical_fmt [IMA]
1633 Use the canonical format for the binary runtime
1634 measurements, instead of host native format.
1637 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1641 The list of supported hash algorithms is defined
1642 in crypto/hash_info.h.
1645 The builtin policies to load during IMA setup.
1646 Format: "tcb | appraise_tcb | secure_boot |
1649 The "tcb" policy measures all programs exec'd, files
1650 mmap'd for exec, and all files opened with the read
1651 mode bit set by either the effective uid (euid=0) or
1654 The "appraise_tcb" policy appraises the integrity of
1655 all files owned by root.
1657 The "secure_boot" policy appraises the integrity
1658 of files (eg. kexec kernel image, kernel modules,
1659 firmware, policy, etc) based on file signatures.
1661 The "fail_securely" policy forces file signature
1662 verification failure also on privileged mounted
1663 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1666 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1667 Load a policy which meets the needs of the Trusted
1668 Computing Base. This means IMA will measure all
1669 programs exec'd, files mmap'd for exec, and all files
1670 opened for read by uid=0.
1673 Select one of defined IMA measurements template formats.
1674 Formats: { "ima" | "ima-ng" | "ima-sig" }
1678 [IMA] Define a custom template format.
1679 Format: { "field1|...|fieldN" }
1681 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1682 Format: <min_file_size>
1683 Set the minimal file size for using asynchronous hash.
1684 If left unspecified, ahash usage is disabled.
1686 ahash performance varies for different data sizes on
1687 different crypto accelerators. This option can be used
1688 to achieve the best performance for a particular HW.
1690 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1692 Set hashing buffer size. Default: 4k.
1694 ahash performance varies for different chunk sizes on
1695 different crypto accelerators. This option can be used
1696 to achieve best performance for particular HW.
1700 Run specified binary instead of /sbin/init as init
1703 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1704 for working out where the kernel is dying during
1707 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1708 initcall functions. Useful for debugging built-in
1709 modules and initcalls.
1711 initrd= [BOOT] Specify the location of the initial ramdisk
1713 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1716 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1718 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1720 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1722 init_pkru= [x86] Specify the default memory protection keys rights
1723 register contents for all processes. 0x55555554 by
1724 default (disallow access to all but pkey 0). Can
1725 override in debugfs after boot.
1727 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1730 int_pln_enable [x86] Enable power limit notification interrupt
1732 integrity_audit=[IMA]
1733 Format: { "0" | "1" }
1734 0 -- basic integrity auditing messages. (Default)
1735 1 -- additional integrity auditing messages.
1737 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1739 Enable intel iommu driver.
1741 Disable intel iommu driver.
1742 igfx_off [Default Off]
1743 By default, gfx is mapped as normal device. If a gfx
1744 device has a dedicated DMAR unit, the DMAR unit is
1745 bypassed by not enabling DMAR with this option. In
1746 this case, gfx device will use physical address for
1749 With this option iommu will not optimize to look
1750 for io virtual address below 32-bit forcing dual
1751 address cycle on pci bus for cards supporting greater
1752 than 32-bit addressing. The default is to look
1753 for translation below 32-bit and if not available
1754 then look in the higher range.
1755 strict [Default Off]
1756 With this option on every unmap_single operation will
1757 result in a hardware IOTLB flush operation as opposed
1758 to batching them for performance.
1759 sp_off [Default Off]
1760 By default, super page will be supported if Intel IOMMU
1761 has the capability. With this option, super page will
1764 By default, scalable mode will be disabled even if the
1765 hardware advertises that it has support for the scalable
1766 mode translation. With this option set, scalable mode
1767 will be used on hardware which claims to support it.
1768 tboot_noforce [Default Off]
1769 Do not force the Intel IOMMU enabled under tboot.
1770 By default, tboot will force Intel IOMMU on, which
1771 could harm performance of some high-throughput
1772 devices like 40GBit network cards, even if identity
1774 Note that using this option lowers the security
1775 provided by tboot because it makes the system
1776 vulnerable to DMA attacks.
1777 nobounce [Default off]
1778 Disable bounce buffer for unstrusted devices such as
1779 the Thunderbolt devices. This will treat the untrusted
1780 devices as the trusted ones, hence might expose security
1781 risks of DMA attacks.
1783 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1784 0 disables intel_idle and fall back on acpi_idle.
1785 1 to 9 specify maximum depth of C-state.
1789 Do not enable intel_pstate as the default
1790 scaling driver for the supported processors
1792 Use intel_pstate as a scaling driver, but configure it
1793 to work with generic cpufreq governors (instead of
1794 enabling its internal governor). This mode cannot be
1795 used along with the hardware-managed P-states (HWP)
1798 Enable intel_pstate on systems that prohibit it by default
1799 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1800 instead of acpi-cpufreq may disable platform features, such
1801 as thermal controls and power capping, that rely on ACPI
1802 P-States information being indicated to OSPM and therefore
1803 should be used with caution. This option does not work with
1804 processors that aren't supported by the intel_pstate driver
1805 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1807 Do not enable hardware P state control (HWP)
1810 Only load intel_pstate on systems which support
1811 hardware P state control (HWP) if available.
1813 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1814 Description Table, specifies preferred power management
1815 profile as "Enterprise Server" or "Performance Server",
1816 then this feature is turned on by default.
1818 Allow per-logical-CPU P-State performance control limits using
1819 cpufreq sysfs interface
1821 intremap= [X86-64, Intel-IOMMU]
1822 on enable Interrupt Remapping (default)
1823 off disable Interrupt Remapping
1824 nosid disable Source ID checking
1826 BIOS x2APIC opt-out request will be ignored
1827 nopost disable Interrupt Posting
1829 iomem= Disable strict checking of access to MMIO memory
1830 strict regions from userspace.
1845 nobypass [PPC/POWERNV]
1846 Disable IOMMU bypass, using IOMMU for PCI devices.
1848 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1849 Format: { "0" | "1" }
1851 Request that DMA unmap operations use deferred
1852 invalidation of hardware TLBs, for increased
1853 throughput at the cost of reduced device isolation.
1854 Will fall back to strict mode if not supported by
1855 the relevant IOMMU driver.
1856 1 - Strict mode (default).
1857 DMA unmap operations invalidate IOMMU hardware TLBs
1861 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
1862 Format: { "0" | "1" }
1863 0 - Use IOMMU translation for DMA.
1864 1 - Bypass the IOMMU for DMA.
1865 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
1867 io7= [HW] IO7 for Marvel based alpha systems
1868 See comment before marvel_specify_io7 in
1869 arch/alpha/kernel/core_marvel.c.
1871 io_delay= [X86] I/O delay method
1873 Standard port 0x80 based delay
1875 Alternate port 0xed based delay (needed on some systems)
1877 Simple two microseconds delay
1882 See Documentation/filesystems/nfs/nfsroot.txt.
1884 ipcmni_extend [KNL] Extend the maximum number of unique System V
1885 IPC identifiers from 32,768 to 16,777,216.
1887 irqaffinity= [SMP] Set the default irq affinity mask
1888 The argument is a cpu list, as described above.
1890 irqchip.gicv2_force_probe=
1893 Force the kernel to look for the second 4kB page
1894 of a GICv2 controller even if the memory range
1895 exposed by the device tree is too small.
1897 irqchip.gicv3_nolpi=
1899 Force the kernel to ignore the availability of
1900 LPIs (and by consequence ITSs). Intended for system
1901 that use the kernel as a bootloader, and thus want
1902 to let secondary kernels in charge of setting up
1905 irqchip.gicv3_pseudo_nmi= [ARM64]
1906 Enables support for pseudo-NMIs in the kernel. This
1907 requires the kernel to be built with
1908 CONFIG_ARM64_PSEUDO_NMI.
1911 When an interrupt is not handled search all handlers
1912 for it. Intended to get systems with badly broken
1916 When an interrupt is not handled search all handlers
1917 for it. Also check all handlers each timer
1918 interrupt. Intended to get systems with badly broken
1922 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1924 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1925 [Deprecated - use cpusets instead]
1926 Format: [flag-list,]<cpu-list>
1928 Specify one or more CPUs to isolate from disturbances
1929 specified in the flag list (default: domain):
1932 Disable the tick when a single task runs.
1934 A residual 1Hz tick is offloaded to workqueues, which you
1935 need to affine to housekeeping through the global
1936 workqueue's affinity configured via the
1937 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1938 by using the 'domain' flag described below.
1940 NOTE: by default the global workqueue runs on all CPUs,
1941 so to protect individual CPUs the 'cpumask' file has to
1942 be configured manually after bootup.
1945 Isolate from the general SMP balancing and scheduling
1946 algorithms. Note that performing domain isolation this way
1947 is irreversible: it's not possible to bring back a CPU to
1948 the domains once isolated through isolcpus. It's strongly
1949 advised to use cpusets instead to disable scheduler load
1950 balancing through the "cpuset.sched_load_balance" file.
1951 It offers a much more flexible interface where CPUs can
1952 move in and out of an isolated set anytime.
1954 You can move a process onto or off an "isolated" CPU via
1955 the CPU affinity syscalls or cpuset.
1956 <cpu number> begins at 0 and the maximum value is
1957 "number of CPUs in system - 1".
1961 Isolate from being targeted by managed interrupts
1962 which have an interrupt mask containing isolated
1963 CPUs. The affinity of managed interrupts is
1964 handled by the kernel and cannot be changed via
1965 the /proc/irq/* interfaces.
1967 This isolation is best effort and only effective
1968 if the automatically assigned interrupt mask of a
1969 device queue contains isolated and housekeeping
1970 CPUs. If housekeeping CPUs are online then such
1971 interrupts are directed to the housekeeping CPU
1972 so that IO submitted on the housekeeping CPU
1973 cannot disturb the isolated CPU.
1975 If a queue's affinity mask contains only isolated
1976 CPUs then this parameter has no effect on the
1977 interrupt routing decision, though interrupts are
1978 only delivered when tasks running on those
1979 isolated CPUs submit IO. IO submitted on
1980 housekeeping CPUs has no influence on those
1983 The format of <cpu-list> is described above.
1987 ivrs_ioapic [HW,X86_64]
1988 Provide an override to the IOAPIC-ID<->DEVICE-ID
1989 mapping provided in the IVRS ACPI table. For
1990 example, to map IOAPIC-ID decimal 10 to
1991 PCI device 00:14.0 write the parameter as:
1992 ivrs_ioapic[10]=00:14.0
1994 ivrs_hpet [HW,X86_64]
1995 Provide an override to the HPET-ID<->DEVICE-ID
1996 mapping provided in the IVRS ACPI table. For
1997 example, to map HPET-ID decimal 0 to
1998 PCI device 00:14.0 write the parameter as:
1999 ivrs_hpet[0]=00:14.0
2001 ivrs_acpihid [HW,X86_64]
2002 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2003 mapping provided in the IVRS ACPI table. For
2004 example, to map UART-HID:UID AMD0020:0 to
2005 PCI device 00:14.5 write the parameter as:
2006 ivrs_acpihid[00:14.5]=AMD0020:0
2008 js= [HW,JOY] Analog joystick
2009 See Documentation/input/joydev/joystick.rst.
2012 When CONFIG_RANDOMIZE_BASE is set, this disables
2013 kernel and module base offset ASLR (Address Space
2014 Layout Randomization).
2017 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2018 report on every invalid memory access. Without this
2019 parameter KASAN will print report only for the first
2024 kernelcore= [KNL,X86,IA-64,PPC]
2025 Format: nn[KMGTPE] | nn% | "mirror"
2026 This parameter specifies the amount of memory usable by
2027 the kernel for non-movable allocations. The requested
2028 amount is spread evenly throughout all nodes in the
2029 system as ZONE_NORMAL. The remaining memory is used for
2030 movable memory in its own zone, ZONE_MOVABLE. In the
2031 event, a node is too small to have both ZONE_NORMAL and
2032 ZONE_MOVABLE, kernelcore memory will take priority and
2033 other nodes will have a larger ZONE_MOVABLE.
2035 ZONE_MOVABLE is used for the allocation of pages that
2036 may be reclaimed or moved by the page migration
2037 subsystem. Note that allocations like PTEs-from-HighMem
2038 still use the HighMem zone if it exists, and the Normal
2039 zone if it does not.
2041 It is possible to specify the exact amount of memory in
2042 the form of "nn[KMGTPE]", a percentage of total system
2043 memory in the form of "nn%", or "mirror". If "mirror"
2044 option is specified, mirrored (reliable) memory is used
2045 for non-movable allocations and remaining memory is used
2046 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2047 are exclusive, so you cannot specify multiple forms.
2049 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2050 Format: <Controller#>[,poll interval]
2051 The controller # is the number of the ehci usb debug
2052 port as it is probed via PCI. The poll interval is
2053 optional and is the number seconds in between
2054 each poll cycle to the debug port in case you need
2055 the functionality for interrupting the kernel with
2056 gdb or control-c on the dbgp connection. When
2057 not using this parameter you use sysrq-g to break into
2058 the kernel debugger.
2060 kgdboc= [KGDB,HW] kgdb over consoles.
2061 Requires a tty driver that supports console polling,
2062 or a supported polling keyboard driver (non-usb).
2063 Serial only format: <serial_device>[,baud]
2064 keyboard only format: kbd
2065 keyboard and serial format: kbd,<serial_device>[,baud]
2066 Optional Kernel mode setting:
2067 kms, kbd format: kms,kbd
2068 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2070 kgdbwait [KGDB] Stop kernel execution and enter the
2071 kernel debugger at the earliest opportunity.
2073 kmac= [MIPS] korina ethernet MAC address.
2074 Configure the RouterBoard 532 series on-chip
2075 Ethernet adapter MAC address.
2077 kmemleak= [KNL] Boot-time kmemleak enable/disable
2078 Valid arguments: on, off
2080 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2083 kprobe_event=[probe-list]
2084 [FTRACE] Add kprobe events and enable at boot time.
2085 The probe-list is a semicolon delimited list of probe
2086 definitions. Each definition is same as kprobe_events
2087 interface, but the parameters are comma delimited.
2088 For example, to add a kprobe event on vfs_read with
2089 arg1 and arg2, add to the command line;
2091 kprobe_event=p,vfs_read,$arg1,$arg2
2093 See also Documentation/trace/kprobetrace.rst "Kernel
2094 Boot Parameter" section.
2096 kpti= [ARM64] Control page table isolation of user
2097 and kernel address spaces.
2098 Default: enabled on cores which need mitigation.
2102 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2103 Default is 0 (don't ignore, but inject #GP)
2105 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2106 Default is false (don't support).
2108 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2113 [KVM] Controls the software workaround for the
2114 X86_BUG_ITLB_MULTIHIT bug.
2115 force : Always deploy workaround.
2116 off : Never deploy workaround.
2117 auto : Deploy workaround based on the presence of
2118 X86_BUG_ITLB_MULTIHIT.
2122 If the software workaround is enabled for the host,
2123 guests do need not to enable it for nested guests.
2125 kvm.nx_huge_pages_recovery_ratio=
2126 [KVM] Controls how many 4KiB pages are periodically zapped
2127 back to huge pages. 0 disables the recovery, otherwise if
2128 the value is N KVM will zap 1/Nth of the 4KiB pages every
2129 minute. The default is 60.
2131 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2132 Default is 1 (enabled)
2134 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2136 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2138 kvm-arm.vgic_v3_group0_trap=
2139 [KVM,ARM] Trap guest accesses to GICv3 group-0
2142 kvm-arm.vgic_v3_group1_trap=
2143 [KVM,ARM] Trap guest accesses to GICv3 group-1
2146 kvm-arm.vgic_v3_common_trap=
2147 [KVM,ARM] Trap guest accesses to GICv3 common
2150 kvm-arm.vgic_v4_enable=
2151 [KVM,ARM] Allow use of GICv4 for direct injection of
2154 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2155 (virtualized MMU) support on capable Intel chips.
2156 Default is 1 (enabled)
2158 kvm-intel.emulate_invalid_guest_state=
2159 [KVM,Intel] Enable emulation of invalid guest states
2160 Default is 0 (disabled)
2162 kvm-intel.flexpriority=
2163 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2164 Default is 1 (enabled)
2167 [KVM,Intel] Enable VMX nesting (nVMX).
2168 Default is 0 (disabled)
2170 kvm-intel.unrestricted_guest=
2171 [KVM,Intel] Disable unrestricted guest feature
2172 (virtualized real and unpaged mode) on capable
2173 Intel chips. Default is 1 (enabled)
2175 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2178 Valid arguments: never, cond, always
2180 always: L1D cache flush on every VMENTER.
2181 cond: Flush L1D on VMENTER only when the code between
2182 VMEXIT and VMENTER can leak host memory.
2183 never: Disables the mitigation
2185 Default is cond (do L1 cache flush in specific instances)
2187 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2188 feature (tagged TLBs) on capable Intel chips.
2189 Default is 1 (enabled)
2191 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2194 The kernel PTE inversion protection is unconditionally
2195 enabled and cannot be disabled.
2198 Provides all available mitigations for the
2199 L1TF vulnerability. Disables SMT and
2200 enables all mitigations in the
2201 hypervisors, i.e. unconditional L1D flush.
2203 SMT control and L1D flush control via the
2204 sysfs interface is still possible after
2205 boot. Hypervisors will issue a warning
2206 when the first VM is started in a
2207 potentially insecure configuration,
2208 i.e. SMT enabled or L1D flush disabled.
2211 Same as 'full', but disables SMT and L1D
2212 flush runtime control. Implies the
2213 'nosmt=force' command line option.
2214 (i.e. sysfs control of SMT is disabled.)
2217 Leaves SMT enabled and enables the default
2218 hypervisor mitigation, i.e. conditional
2221 SMT control and L1D flush control via the
2222 sysfs interface is still possible after
2223 boot. Hypervisors will issue a warning
2224 when the first VM is started in a
2225 potentially insecure configuration,
2226 i.e. SMT enabled or L1D flush disabled.
2230 Disables SMT and enables the default
2231 hypervisor mitigation.
2233 SMT control and L1D flush control via the
2234 sysfs interface is still possible after
2235 boot. Hypervisors will issue a warning
2236 when the first VM is started in a
2237 potentially insecure configuration,
2238 i.e. SMT enabled or L1D flush disabled.
2241 Same as 'flush', but hypervisors will not
2242 warn when a VM is started in a potentially
2243 insecure configuration.
2246 Disables hypervisor mitigations and doesn't
2248 It also drops the swap size and available
2249 RAM limit restriction on both hypervisor and
2254 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2260 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2263 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2264 value for LAPIC timer one-shot implementation. Default
2265 back to the programmable timer unit in the LAPIC.
2267 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2270 libata.dma= [LIBATA] DMA control
2271 libata.dma=0 Disable all PATA and SATA DMA
2272 libata.dma=1 PATA and SATA Disk DMA only
2273 libata.dma=2 ATAPI (CDROM) DMA only
2274 libata.dma=4 Compact Flash DMA only
2275 Combinations also work, so libata.dma=3 enables DMA
2276 for disks and CDROMs, but not CFs.
2278 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2279 libata.ignore_hpa=0 keep BIOS limits (default)
2280 libata.ignore_hpa=1 ignore limits, using full disk
2282 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2286 libata.force= [LIBATA] Force configurations. The format is comma
2287 separated list of "[ID:]VAL" where ID is
2288 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2289 matching port, link or device. Basically, it matches
2290 the ATA ID string printed on console by libata. If
2291 the whole ID part is omitted, the last PORT and DEVICE
2292 values are used. If ID hasn't been specified yet, the
2293 configuration applies to all ports, links and devices.
2295 If only DEVICE is omitted, the parameter applies to
2296 the port and all links and devices behind it. DEVICE
2297 number of 0 either selects the first device or the
2298 first fan-out link behind PMP device. It does not
2299 select the host link. DEVICE number of 15 selects the
2300 host link and device attached to it.
2302 The VAL specifies the configuration to force. As long
2303 as there's no ambiguity shortcut notation is allowed.
2304 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2305 The following configurations can be forced.
2307 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2308 Any ID with matching PORT is used.
2310 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2312 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2313 udma[/][16,25,33,44,66,100,133] notation is also
2316 * [no]ncq: Turn on or off NCQ.
2318 * [no]ncqtrim: Turn off queued DSM TRIM.
2320 * nohrst, nosrst, norst: suppress hard, soft
2323 * rstonce: only attempt one reset during
2324 hot-unplug link recovery
2326 * dump_id: dump IDENTIFY data.
2328 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2330 * disable: Disable this device.
2332 If there are multiple matching configurations changing
2333 the same attribute, the last one is used.
2335 memblock=debug [KNL] Enable memblock debug messages.
2337 load_ramdisk= [RAM] List of ramdisks to load from floppy
2338 See Documentation/admin-guide/blockdev/ramdisk.rst.
2340 lockd.nlm_grace_period=P [NFS] Assign grace period.
2343 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2346 lockd.nlm_timeout=T [NFS] Assign timeout value.
2349 lockd.nlm_udpport=M [NFS] Assign UDP port.
2352 lockdown= [SECURITY]
2353 { integrity | confidentiality }
2354 Enable the kernel lockdown feature. If set to
2355 integrity, kernel features that allow userland to
2356 modify the running kernel are disabled. If set to
2357 confidentiality, kernel features that allow userland
2358 to extract confidential information from the kernel
2361 locktorture.nreaders_stress= [KNL]
2362 Set the number of locking read-acquisition kthreads.
2363 Defaults to being automatically set based on the
2364 number of online CPUs.
2366 locktorture.nwriters_stress= [KNL]
2367 Set the number of locking write-acquisition kthreads.
2369 locktorture.onoff_holdoff= [KNL]
2370 Set time (s) after boot for CPU-hotplug testing.
2372 locktorture.onoff_interval= [KNL]
2373 Set time (s) between CPU-hotplug operations, or
2374 zero to disable CPU-hotplug testing.
2376 locktorture.shuffle_interval= [KNL]
2377 Set task-shuffle interval (jiffies). Shuffling
2378 tasks allows some CPUs to go into dyntick-idle
2379 mode during the locktorture test.
2381 locktorture.shutdown_secs= [KNL]
2382 Set time (s) after boot system shutdown. This
2383 is useful for hands-off automated testing.
2385 locktorture.stat_interval= [KNL]
2386 Time (s) between statistics printk()s.
2388 locktorture.stutter= [KNL]
2389 Time (s) to stutter testing, for example,
2390 specifying five seconds causes the test to run for
2391 five seconds, wait for five seconds, and so on.
2392 This tests the locking primitive's ability to
2393 transition abruptly to and from idle.
2395 locktorture.torture_type= [KNL]
2396 Specify the locking implementation to test.
2398 locktorture.verbose= [KNL]
2399 Enable additional printk() statements.
2401 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2404 loglevel= All Kernel Messages with a loglevel smaller than the
2405 console loglevel will be printed to the console. It can
2406 also be changed with klogd or other programs. The
2407 loglevels are defined as follows:
2409 0 (KERN_EMERG) system is unusable
2410 1 (KERN_ALERT) action must be taken immediately
2411 2 (KERN_CRIT) critical conditions
2412 3 (KERN_ERR) error conditions
2413 4 (KERN_WARNING) warning conditions
2414 5 (KERN_NOTICE) normal but significant condition
2415 6 (KERN_INFO) informational
2416 7 (KERN_DEBUG) debug-level messages
2418 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2419 in bytes. n must be a power of two and greater
2420 than the minimal size. The minimal size is defined
2421 by LOG_BUF_SHIFT kernel config parameter. There is
2422 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2423 that allows to increase the default size depending on
2424 the number of CPUs. See init/Kconfig for more details.
2426 logo.nologo [FB] Disables display of the built-in Linux logo.
2427 This may be used to provide more screen space for
2428 kernel log messages and is useful when debugging
2429 kernel boot problems.
2431 lp=0 [LP] Specify parallel ports to use, e.g,
2432 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2433 lp=reset first parallel port). 'lp=0' disables the
2434 lp=auto printer driver. 'lp=reset' (which can be
2435 specified in addition to the ports) causes
2436 attached printers to be reset. Using
2437 lp=port1,port2,... specifies the parallel ports
2438 to associate lp devices with, starting with
2439 lp0. A port specification may be 'none' to skip
2440 that lp device, or a parport name such as
2441 'parport0'. Specifying 'lp=auto' instead of a
2442 port specification list means that device IDs
2443 from each port should be examined, to see if
2444 an IEEE 1284-compliant printer is attached; if
2445 so, the driver will manage that printer.
2446 See also header of drivers/char/lp.c.
2449 Sets loops_per_jiffy to given constant, thus avoiding
2450 time-consuming boot-time autodetection (up to 250 ms per
2451 CPU). 0 enables autodetection (default). To determine
2452 the correct value for your kernel, boot with normal
2453 autodetection and see what value is printed. Note that
2454 on SMP systems the preset will be applied to all CPUs,
2455 which is likely to cause problems if your CPUs need
2456 significantly divergent settings. An incorrect value
2457 will cause delays in the kernel to be wrong, leading to
2458 unpredictable I/O errors and other breakage. Although
2459 unlikely, in the extreme case this might damage your
2463 Format: <io>,<irq>,<dma>
2465 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2468 [SECURITY] Choose order of LSM initialization. This
2469 overrides CONFIG_LSM, and the "security=" parameter.
2471 machvec= [IA-64] Force the use of a particular machine-vector
2472 (machvec) in a generic kernel.
2473 Example: machvec=hpzx1
2475 machtype= [Loongson] Share the same kernel image file between different
2477 Example: machtype=lemote-yeeloong-2f-7inch
2479 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2480 than or equal to this physical address is ignored.
2482 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2483 will bring up during bootup. maxcpus=n : n >= 0 limits
2484 the kernel to bring up 'n' processors. Surely after
2485 bootup you can bring up the other plugged cpu by executing
2486 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2487 only takes effect during system bootup.
2488 While n=0 is a special case, it is equivalent to "nosmp",
2489 which also disables the IO APIC.
2491 max_loop= [LOOP] The number of loop block devices that get
2492 (loop.max_loop) unconditionally pre-created at init time. The default
2493 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2494 of statically allocating a predefined number, loop
2495 devices can be requested on-demand with the
2496 /dev/loop-control interface.
2498 mce [X86-32] Machine Check Exception
2500 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2502 md= [HW] RAID subsystems devices and level
2503 See Documentation/admin-guide/md.rst.
2506 Format: <first>,<last>
2507 Specifies range of consoles to be captured by the MDA.
2510 Control mitigation for the Micro-architectural Data
2511 Sampling (MDS) vulnerability.
2513 Certain CPUs are vulnerable to an exploit against CPU
2514 internal buffers which can forward information to a
2515 disclosure gadget under certain conditions.
2517 In vulnerable processors, the speculatively
2518 forwarded data can be used in a cache side channel
2519 attack, to access data to which the attacker does
2520 not have direct access.
2522 This parameter controls the MDS mitigation. The
2525 full - Enable MDS mitigation on vulnerable CPUs
2526 full,nosmt - Enable MDS mitigation and disable
2527 SMT on vulnerable CPUs
2528 off - Unconditionally disable MDS mitigation
2530 On TAA-affected machines, mds=off can be prevented by
2531 an active TAA mitigation as both vulnerabilities are
2532 mitigated with the same mechanism so in order to disable
2533 this mitigation, you need to specify tsx_async_abort=off
2536 Not specifying this option is equivalent to
2539 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2541 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2542 Amount of memory to be used when the kernel is not able
2543 to see the whole system memory or for test.
2544 [X86] Work as limiting max address. Use together
2545 with memmap= to avoid physical address space collisions.
2546 Without memmap= PCI devices could be placed at addresses
2547 belonging to unused RAM.
2549 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2553 [KNL,SH] Allow user to override the default size for
2554 per-device physically contiguous DMA buffers.
2556 memhp_default_state=online/offline
2557 [KNL] Set the initial state for the memory hotplug
2558 onlining policy. If not specified, the default value is
2559 set according to the
2560 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2562 See Documentation/admin-guide/mm/memory-hotplug.rst.
2564 memmap=exactmap [KNL,X86] Enable setting of an exact
2565 E820 memory map, as specified by the user.
2566 Such memmap=exactmap lines can be constructed based on
2567 BIOS output or other requirements. See the memmap=nn@ss
2570 memmap=nn[KMG]@ss[KMG]
2571 [KNL] Force usage of a specific region of memory.
2572 Region of memory to be used is from ss to ss+nn.
2573 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2574 which limits max address to nn[KMG].
2575 Multiple different regions can be specified,
2578 memmap=100M@2G,100M#3G,1G!1024G
2580 memmap=nn[KMG]#ss[KMG]
2581 [KNL,ACPI] Mark specific memory as ACPI data.
2582 Region of memory to be marked is from ss to ss+nn.
2584 memmap=nn[KMG]$ss[KMG]
2585 [KNL,ACPI] Mark specific memory as reserved.
2586 Region of memory to be reserved is from ss to ss+nn.
2587 Example: Exclude memory from 0x18690000-0x1869ffff
2588 memmap=64K$0x18690000
2590 memmap=0x10000$0x18690000
2591 Some bootloaders may need an escape character before '$',
2592 like Grub2, otherwise '$' and the following number
2595 memmap=nn[KMG]!ss[KMG]
2596 [KNL,X86] Mark specific memory as protected.
2597 Region of memory to be used, from ss to ss+nn.
2598 The memory region may be marked as e820 type 12 (0xc)
2599 and is NVDIMM or ADR memory.
2601 memmap=<size>%<offset>-<oldtype>+<newtype>
2602 [KNL,ACPI] Convert memory within the specified region
2603 from <oldtype> to <newtype>. If "-<oldtype>" is left
2604 out, the whole region will be marked as <newtype>,
2605 even if previously unavailable. If "+<newtype>" is left
2606 out, matching memory will be removed. Types are
2607 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2608 3 = ACPI, 12 = PRAM.
2610 memory_corruption_check=0/1 [X86]
2611 Some BIOSes seem to corrupt the first 64k of
2612 memory when doing things like suspend/resume.
2613 Setting this option will scan the memory
2614 looking for corruption. Enabling this will
2615 both detect corruption and prevent the kernel
2616 from using the memory being corrupted.
2617 However, its intended as a diagnostic tool; if
2618 repeatable BIOS-originated corruption always
2619 affects the same memory, you can use memmap=
2620 to prevent the kernel from using that memory.
2622 memory_corruption_check_size=size [X86]
2623 By default it checks for corruption in the low
2624 64k, making this memory unavailable for normal
2625 use. Use this parameter to scan for
2626 corruption in more or less memory.
2628 memory_corruption_check_period=seconds [X86]
2629 By default it checks for corruption every 60
2630 seconds. Use this parameter to check at some
2631 other rate. 0 disables periodic checking.
2633 memtest= [KNL,X86,ARM,PPC] Enable memtest
2635 default : 0 <disable>
2636 Specifies the number of memtest passes to be
2637 performed. Each pass selects another test
2638 pattern from a given set of patterns. Memtest
2639 fills the memory with this pattern, validates
2640 memory contents and reserves bad memory
2641 regions that are detected.
2643 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2644 Valid arguments: on, off
2645 Default (depends on kernel configuration option):
2646 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2647 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2648 mem_encrypt=on: Activate SME
2649 mem_encrypt=off: Do not activate SME
2651 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2652 for details on when memory encryption can be activated.
2654 mem_sleep_default= [SUSPEND] Default system suspend mode:
2655 s2idle - Suspend-To-Idle
2656 shallow - Power-On Suspend or equivalent (if supported)
2657 deep - Suspend-To-RAM or equivalent (if supported)
2658 See Documentation/admin-guide/pm/sleep-states.rst.
2660 meye.*= [HW] Set MotionEye Camera parameters
2661 See Documentation/media/v4l-drivers/meye.rst.
2663 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2664 Multi-Function General Purpose Timers on AMD Geode
2667 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2668 the BIOS has incorrectly applied a workaround. TinyBIOS
2669 version 0.98 is known to be affected, 0.99 fixes the
2670 problem by letting the user disable the workaround.
2674 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2675 physical address is ignored.
2677 mini2440= [ARM,HW,KNL]
2678 Format:[0..2][b][c][t]
2680 MINI2440 configuration specification:
2681 0 - The attached screen is the 3.5" TFT
2682 1 - The attached screen is the 7" TFT
2683 2 - The VGA Shield is attached (1024x768)
2684 Leaving out the screen size parameter will not load
2685 the TFT driver, and the framebuffer will be left
2687 b - Enable backlight. The TFT backlight pin will be
2688 linked to the kernel VESA blanking code and a GPIO
2689 LED. This parameter is not necessary when using the
2691 c - Enable the s3c camera interface.
2692 t - Reserved for enabling touchscreen support. The
2693 touchscreen support is not enabled in the mainstream
2694 kernel as of 2.6.30, a preliminary port can be found
2695 in the "bleeding edge" mini2440 support kernel at
2696 http://repo.or.cz/w/linux-2.6/mini2440.git
2699 [X86,PPC,S390,ARM64] Control optional mitigations for
2700 CPU vulnerabilities. This is a set of curated,
2701 arch-independent options, each of which is an
2702 aggregation of existing arch-specific options.
2705 Disable all optional CPU mitigations. This
2706 improves system performance, but it may also
2707 expose users to several CPU vulnerabilities.
2708 Equivalent to: nopti [X86,PPC]
2710 nospectre_v1 [X86,PPC]
2712 nospectre_v2 [X86,PPC,S390,ARM64]
2713 spectre_v2_user=off [X86]
2714 spec_store_bypass_disable=off [X86,PPC]
2715 ssbd=force-off [ARM64]
2718 tsx_async_abort=off [X86]
2719 kvm.nx_huge_pages=off [X86]
2722 This does not have any effect on
2723 kvm.nx_huge_pages when
2724 kvm.nx_huge_pages=force.
2727 Mitigate all CPU vulnerabilities, but leave SMT
2728 enabled, even if it's vulnerable. This is for
2729 users who don't want to be surprised by SMT
2730 getting disabled across kernel upgrades, or who
2731 have other ways of avoiding SMT-based attacks.
2732 Equivalent to: (default behavior)
2735 Mitigate all CPU vulnerabilities, disabling SMT
2736 if needed. This is for users who always want to
2737 be fully mitigated, even if it means losing SMT.
2738 Equivalent to: l1tf=flush,nosmt [X86]
2739 mds=full,nosmt [X86]
2740 tsx_async_abort=full,nosmt [X86]
2743 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2744 parameter allows control of the logging verbosity for
2745 the additional memory initialisation checks. A value
2746 of 0 disables mminit logging and a level of 4 will
2747 log everything. Information is printed at KERN_DEBUG
2748 so loglevel=8 may also need to be specified.
2751 [KNL] When CONFIG_MODULE_SIG is set, this means that
2752 modules without (valid) signatures will fail to load.
2753 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2754 is always true, so this option does nothing.
2756 module_blacklist= [KNL] Do not load a comma-separated list of
2757 modules. Useful for debugging problem modules.
2760 [MOUSE] Maximum time between finger touching and
2761 leaving touchpad surface for touch to be considered
2762 a tap and be reported as a left button click (for
2763 touchpads working in absolute mode only).
2765 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2766 reporting absolute coordinates, such as tablets
2767 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2768 reporting absolute coordinates, such as tablets
2770 movablecore= [KNL,X86,IA-64,PPC]
2771 Format: nn[KMGTPE] | nn%
2772 This parameter is the complement to kernelcore=, it
2773 specifies the amount of memory used for migratable
2774 allocations. If both kernelcore and movablecore is
2775 specified, then kernelcore will be at *least* the
2776 specified value but may be more. If movablecore on its
2777 own is specified, the administrator must be careful
2778 that the amount of memory usable for all allocations
2781 movable_node [KNL] Boot-time switch to make hotplugable memory
2782 NUMA nodes to be movable. This means that the memory
2783 of such nodes will be usable only for movable
2784 allocations which rules out almost all kernel
2785 allocations. Use with caution!
2787 MTD_Partition= [MTD]
2788 Format: <name>,<region-number>,<size>,<offset>
2790 MTD_Region= [MTD] Format:
2791 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2794 See drivers/mtd/cmdlinepart.c.
2796 multitce=off [PPC] This parameter disables the use of the pSeries
2797 firmware feature for updating multiple TCE entries
2800 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2802 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2804 boundary - index of last SLC block on Flex-OneNAND.
2805 The remaining blocks are configured as MLC blocks.
2806 lock - Configure if Flex-OneNAND boundary should be locked.
2807 Once locked, the boundary cannot be changed.
2808 1 indicates lock status, 0 indicates unlock status.
2811 ARM/S3C2412 JIVE boot control
2813 See arch/arm/mach-s3c2412/mach-jive.c
2815 mtouchusb.raw_coordinates=
2816 [HW] Make the MicroTouch USB driver use raw coordinates
2817 ('y', default) or cooked coordinates ('n')
2819 mtrr_chunk_size=nn[KMG] [X86]
2820 used for mtrr cleanup. It is largest continuous chunk
2821 that could hold holes aka. UC entries.
2823 mtrr_gran_size=nn[KMG] [X86]
2824 Used for mtrr cleanup. It is granularity of mtrr block.
2826 Large value could prevent small alignment from
2829 mtrr_spare_reg_nr=n [X86]
2831 Range: 0,7 : spare reg number
2833 Used for mtrr cleanup. It is spare mtrr entries number.
2834 Set to 2 or more if your graphical card needs more.
2836 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2838 netdev= [NET] Network devices parameters
2839 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2840 Note that mem_start is often overloaded to mean
2841 something different and driver-specific.
2842 This usage is only documented in each driver source
2846 [NETFILTER] Enable connection tracking flow accounting
2847 0 to disable accounting
2848 1 to enable accounting
2851 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2852 See Documentation/filesystems/nfs/nfsroot.txt.
2854 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2855 See Documentation/filesystems/nfs/nfsroot.txt.
2857 nfsrootdebug [NFS] enable nfsroot debugging messages.
2858 See Documentation/filesystems/nfs/nfsroot.txt.
2860 nfs.callback_nr_threads=
2861 [NFSv4] set the total number of threads that the
2862 NFS client will assign to service NFSv4 callback
2865 nfs.callback_tcpport=
2866 [NFS] set the TCP port on which the NFSv4 callback
2867 channel should listen.
2870 [NFS] sets the pathname to the program which is used
2871 to update the NFS client cache entries.
2873 nfs.cache_getent_timeout=
2874 [NFS] sets the timeout after which an attempt to
2875 update a cache entry is deemed to have failed.
2877 nfs.idmap_cache_timeout=
2878 [NFS] set the maximum lifetime for idmapper cache
2882 [NFS] enable 64-bit inode numbers.
2883 If zero, the NFS client will fake up a 32-bit inode
2884 number for the readdir() and stat() syscalls instead
2885 of returning the full 64-bit number.
2886 The default is to return 64-bit inode numbers.
2888 nfs.max_session_cb_slots=
2889 [NFSv4.1] Sets the maximum number of session
2890 slots the client will assign to the callback
2891 channel. This determines the maximum number of
2892 callbacks the client will process in parallel for
2893 a particular server.
2895 nfs.max_session_slots=
2896 [NFSv4.1] Sets the maximum number of session slots
2897 the client will attempt to negotiate with the server.
2898 This limits the number of simultaneous RPC requests
2899 that the client can send to the NFSv4.1 server.
2900 Note that there is little point in setting this
2901 value higher than the max_tcp_slot_table_limit.
2903 nfs.nfs4_disable_idmapping=
2904 [NFSv4] When set to the default of '1', this option
2905 ensures that both the RPC level authentication
2906 scheme and the NFS level operations agree to use
2907 numeric uids/gids if the mount is using the
2908 'sec=sys' security flavour. In effect it is
2909 disabling idmapping, which can make migration from
2910 legacy NFSv2/v3 systems to NFSv4 easier.
2911 Servers that do not support this mode of operation
2912 will be autodetected by the client, and it will fall
2913 back to using the idmapper.
2914 To turn off this behaviour, set the value to '0'.
2916 [NFS4] Specify an additional fixed unique ident-
2917 ification string that NFSv4 clients can insert into
2918 their nfs_client_id4 string. This is typically a
2919 UUID that is generated at system install time.
2921 nfs.send_implementation_id =
2922 [NFSv4.1] Send client implementation identification
2923 information in exchange_id requests.
2924 If zero, no implementation identification information
2926 The default is to send the implementation identification
2929 nfs.recover_lost_locks =
2930 [NFSv4] Attempt to recover locks that were lost due
2931 to a lease timeout on the server. Please note that
2932 doing this risks data corruption, since there are
2933 no guarantees that the file will remain unchanged
2934 after the locks are lost.
2935 If you want to enable the kernel legacy behaviour of
2936 attempting to recover these locks, then set this
2938 The default parameter value of '0' causes the kernel
2939 not to attempt recovery of lost locks.
2941 nfs4.layoutstats_timer =
2942 [NFSv4.2] Change the rate at which the kernel sends
2943 layoutstats to the pNFS metadata server.
2945 Setting this to value to 0 causes the kernel to use
2946 whatever value is the default set by the layout
2947 driver. A non-zero value sets the minimum interval
2948 in seconds between layoutstats transmissions.
2950 nfsd.nfs4_disable_idmapping=
2951 [NFSv4] When set to the default of '1', the NFSv4
2952 server will return only numeric uids and gids to
2953 clients using auth_sys, and will accept numeric uids
2954 and gids from such clients. This is intended to ease
2955 migration from NFSv2/v3.
2957 nmi_debug= [KNL,SH] Specify one or more actions to take
2958 when a NMI is triggered.
2959 Format: [state][,regs][,debounce][,die]
2961 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2962 Format: [panic,][nopanic,][num]
2964 0 - turn hardlockup detector in nmi_watchdog off
2965 1 - turn hardlockup detector in nmi_watchdog on
2966 When panic is specified, panic when an NMI watchdog
2967 timeout occurs (or 'nopanic' to not panic on an NMI
2968 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
2969 To disable both hard and soft lockup detectors,
2970 please see 'nowatchdog'.
2971 This is useful when you use a panic=... timeout and
2972 need the box quickly up again.
2974 These settings can be accessed at runtime via
2975 the nmi_watchdog and hardlockup_panic sysctls.
2977 netpoll.carrier_timeout=
2978 [NET] Specifies amount of time (in seconds) that
2979 netpoll should wait for a carrier. By default netpoll
2982 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2983 emulation library even if a 387 maths coprocessor
2986 no5lvl [X86-64] Disable 5-level paging mode. Forces
2987 kernel to use 4-level paging instead.
2990 [HW] Never suspend the console
2991 Disable suspending of consoles during suspend and
2992 hibernate operations. Once disabled, debugging
2993 messages can reach various consoles while the rest
2994 of the system is being put to sleep (ie, while
2995 debugging driver suspend/resume hooks). This may
2996 not work reliably with all consoles, but is known
2997 to work with serial and VGA consoles.
2998 To facilitate more flexible debugging, we also add
2999 console_suspend, a printk module parameter to control
3000 it. Users could use console_suspend (usually
3001 /sys/module/printk/parameters/console_suspend) to
3002 turn on/off it dynamically.
3004 novmcoredd [KNL,KDUMP]
3005 Disable device dump. Device dump allows drivers to
3006 append dump data to vmcore so you can collect driver
3007 specified debug info. Drivers can append the data
3008 without any limit and this data is stored in memory,
3009 so this may cause significant memory stress. Disabling
3010 device dump can help save memory but the driver debug
3011 data will be no longer available. This parameter
3012 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3015 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3016 caches in the slab allocator. Saves per-node memory,
3017 but will impact performance.
3021 noaltinstr [S390] Disables alternative instructions patching
3022 (CPU alternatives feature).
3024 noapic [SMP,APIC] Tells the kernel to not make use of any
3025 IOAPICs that may be present in the system.
3027 noautogroup Disable scheduler automatic task group creation.
3029 nobats [PPC] Do not use BATs for mapping kernel lowmem
3030 on "Classic" PPC cores.
3034 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3036 nodelayacct [KNL] Disable per-task delay accounting
3038 nodsp [SH] Disable hardware DSP at boot time.
3040 noefi Disable EFI runtime services support.
3045 On X86-32 available only on PAE configured kernels.
3046 noexec=on: enable non-executable mappings (default)
3047 noexec=off: disable non-executable mappings
3050 Disable SMAP (Supervisor Mode Access Prevention)
3051 even if it is supported by processor.
3054 Disable SMEP (Supervisor Mode Execution Prevention)
3055 even if it is supported by processor.
3058 This affects only 32-bit executables.
3059 noexec32=on: enable non-executable mappings (default)
3060 read doesn't imply executable mappings
3061 noexec32=off: disable non-executable mappings
3062 read implies executable mappings
3064 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3066 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3067 register save and restore. The kernel will only save
3068 legacy floating-point registers on task switch.
3070 nohugeiomap [KNL,x86,PPC] Disable kernel huge I/O mappings.
3072 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3073 Equivalent to smt=1.
3075 [KNL,x86] Disable symmetric multithreading (SMT).
3076 nosmt=force: Force disable SMT, cannot be undone
3077 via the sysfs control file.
3079 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3080 (bounds check bypass). With this option data leaks are
3081 possible in the system.
3083 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3084 the Spectre variant 2 (indirect branch prediction)
3085 vulnerability. System may allow data leaks with this
3088 nospec_store_bypass_disable
3089 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3091 noxsave [BUGS=X86] Disables x86 extended register state save
3092 and restore using xsave. The kernel will fallback to
3093 enabling legacy floating-point and sse state.
3095 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3096 register states. The kernel will fall back to use
3097 xsave to save the states. By using this parameter,
3098 performance of saving the states is degraded because
3099 xsave doesn't support modified optimization while
3100 xsaveopt supports it on xsaveopt enabled systems.
3102 noxsaves [X86] Disables xsaves and xrstors used in saving and
3103 restoring x86 extended register state in compacted
3104 form of xsave area. The kernel will fall back to use
3105 xsaveopt and xrstor to save and restore the states
3106 in standard form of xsave area. By using this
3107 parameter, xsave area per process might occupy more
3108 memory on xsaves enabled systems.
3110 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
3111 wfi(ARM) instruction doesn't work correctly and not to
3112 use it. This is also useful when using JTAG debugger.
3114 no_file_caps Tells the kernel not to honor file capabilities. The
3115 only way then for a file to be executed with privilege
3116 is to be setuid root or executed by root.
3118 nohalt [IA-64] Tells the kernel not to use the power saving
3119 function PAL_HALT_LIGHT when idle. This increases
3120 power-consumption. On the positive side, it reduces
3121 interrupt wake-up latency, which may improve performance
3122 in certain environments such as networked servers or
3125 nohibernate [HIBERNATION] Disable hibernation and resume.
3127 nohz= [KNL] Boottime enable/disable dynamic ticks
3128 Valid arguments: on, off
3131 nohz_full= [KNL,BOOT,SMP,ISOL]
3132 The argument is a cpu list, as described above.
3133 In kernels built with CONFIG_NO_HZ_FULL=y, set
3134 the specified list of CPUs whose tick will be stopped
3135 whenever possible. The boot CPU will be forced outside
3136 the range to maintain the timekeeping. Any CPUs
3137 in this list will have their RCU callbacks offloaded,
3138 just as if they had also been called out in the
3139 rcu_nocbs= boot parameter.
3141 noiotrap [SH] Disables trapped I/O port accesses.
3143 noirqdebug [X86-32] Disables the code which attempts to detect and
3144 disable unhandled interrupt sources.
3146 no_timer_check [X86,APIC] Disables the code which tests for
3147 broken timer IRQ sources.
3149 noisapnp [ISAPNP] Disables ISA PnP code.
3151 noinitrd [RAM] Tells the kernel not to load any configured
3154 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3156 [Deprecated - use intremap=off]
3160 noinvpcid [X86] Disable the INVPCID cpu feature.
3162 nojitter [IA-64] Disables jitter checking for ITC timers.
3164 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3166 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3170 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3171 clock and use the default one.
3173 no-steal-acc [X86,KVM,ARM64] Disable paravirtualized steal time
3174 accounting. steal time is computed, but won't
3175 influence scheduler behaviour
3177 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3179 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3181 noltlbs [PPC] Do not use large page/tlb entries for kernel
3182 lowmem mapping on PPC40x and PPC8xx
3184 nomca [IA-64] Disable machine check abort handling
3186 nomce [X86-32] Disable Machine Check Exception
3188 nomfgpt [X86-32] Disable Multi-Function General Purpose
3189 Timer usage (for AMD Geode machines).
3191 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3192 shutdown the other cpus. Instead use the REBOOT_VECTOR
3195 nomodule Disable module load
3197 nopat [X86] Disable PAT (page attribute table extension of
3198 pagetables) support.
3200 nopcid [X86-64] Disable the PCID cpu feature.
3202 norandmaps Don't use address space randomization. Equivalent to
3203 echo 0 > /proc/sys/kernel/randomize_va_space
3205 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3206 with UP alternatives
3208 nordrand [X86] Disable kernel use of the RDRAND and
3209 RDSEED instructions even if they are supported
3210 by the processor. RDRAND and RDSEED are still
3211 available to user space applications.
3213 noresume [SWSUSP] Disables resume and restores original swap
3216 no-scroll [VGA] Disables scrollback.
3217 This is required for the Braillex ib80-piezo Braille
3218 reader made by F.H. Papenmeier (Germany).
3222 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3224 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3225 and disable the IO APIC. legacy for "maxcpus=0".
3227 nosoftlockup [KNL] Disable the soft-lockup detector.
3229 nosync [HW,M68K] Disables sync negotiation for all devices.
3231 nowatchdog [KNL] Disable both lockup detectors, i.e.
3232 soft-lockup and NMI watchdog (hard-lockup).
3236 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3238 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3239 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3240 Some features depend on CPU0. Known dependencies are:
3241 1. Resume from suspend/hibernate depends on CPU0.
3242 Suspend/hibernate will fail if CPU0 is offline and you
3243 need to online CPU0 before suspend/hibernate.
3244 2. PIC interrupts also depend on CPU0. CPU0 can't be
3245 removed if a PIC interrupt is detected.
3246 It's said poweroff/reboot may depend on CPU0 on some
3247 machines although I haven't seen such issues so far
3248 after CPU0 is offline on a few tested machines.
3249 If the dependencies are under your control, you can
3250 turn on cpu0_hotplug.
3252 nps_mtm_hs_ctr= [KNL,ARC]
3253 This parameter sets the maximum duration, in
3254 cycles, each HW thread of the CTOP can run
3255 without interruptions, before HW switches it.
3256 The actual maximum duration is 16 times this
3258 Format: integer between 1 and 255
3261 nptcg= [IA-64] Override max number of concurrent global TLB
3262 purges which is reported from either PAL_VM_SUMMARY or
3265 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3266 could support. nr_cpus=n : n >= 1 limits the kernel to
3267 support 'n' processors. It could be larger than the
3268 number of already plugged CPU during bootup, later in
3269 runtime you can physically add extra cpu until it reaches
3270 n. So during boot up some boot time memory for per-cpu
3271 variables need be pre-allocated for later physical cpu
3274 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3276 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3277 Allowed values are enable and disable
3279 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3280 'node', 'default' can be specified
3281 This can be set from sysctl after boot.
3282 See Documentation/admin-guide/sysctl/vm.rst for details.
3284 of_devlink [OF, KNL] Create device links between consumer and
3285 supplier devices by scanning the devictree to infer the
3286 consumer/supplier relationships. A consumer device
3287 will not be probed until all the supplier devices have
3288 probed successfully.
3290 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3291 See Documentation/debugging-via-ohci1394.txt for more
3294 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3295 Rather than timing out after 20 ms if an EC
3296 command is not properly ACKed, override the length
3297 of the timeout. We have interrupts disabled while
3298 waiting for the ACK, so if this is set too high
3299 interrupts *may* be lost!
3301 omap_mux= [OMAP] Override bootloader pin multiplexing.
3302 Format: <mux_mode0.mode_name=value>...
3303 For example, to override I2C bus2:
3304 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3306 oprofile.timer= [HW]
3307 Use timer interrupt instead of performance counters
3309 oprofile.cpu_type= Force an oprofile cpu type
3310 This might be useful if you have an older oprofile
3311 userland or if you want common events.
3312 Format: { arch_perfmon }
3313 arch_perfmon: [X86] Force use of architectural
3314 perfmon on Intel CPUs instead of the
3315 CPU specific event set.
3316 timer: [X86] Force use of architectural NMI
3317 timer mode (see also oprofile.timer
3318 for generic hr timer mode)
3320 oops=panic Always panic on oopses. Default is to just kill the
3321 process, but there is a small probability of
3322 deadlocking the machine.
3323 This will also cause panics on machine check exceptions.
3324 Useful together with panic=30 to trigger a reboot.
3327 [KNL] Boolean flag to control whether the page allocator
3328 should randomize its free lists. The randomization may
3329 be automatically enabled if the kernel detects it is
3330 running on a platform with a direct-mapped memory-side
3331 cache, and this parameter can be used to
3332 override/disable that behavior. The state of the flag
3333 can be read from sysfs at:
3334 /sys/module/page_alloc/parameters/shuffle.
3336 page_owner= [KNL] Boot-time page_owner enabling option.
3337 Storage of the information about who allocated
3338 each page is disabled in default. With this switch,
3340 on: enable the feature
3342 page_poison= [KNL] Boot-time parameter changing the state of
3343 poisoning on the buddy allocator, available with
3344 CONFIG_PAGE_POISONING=y.
3345 off: turn off poisoning (default)
3346 on: turn on poisoning
3348 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3349 timeout > 0: seconds before rebooting
3350 timeout = 0: wait forever
3351 timeout < 0: reboot immediately
3354 panic_print= Bitmask for printing system info when panic happens.
3355 User can chose combination of the following bits:
3356 bit 0: print all tasks info
3357 bit 1: print system memory info
3358 bit 2: print timer info
3359 bit 3: print locks info if CONFIG_LOCKDEP is on
3360 bit 4: print ftrace buffer
3361 bit 5: print all printk messages in buffer
3363 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3366 crash_kexec_post_notifiers
3367 Run kdump after running panic-notifiers and dumping
3368 kmsg. This only for the users who doubt kdump always
3369 succeeds in any situation.
3370 Note that this also increases risks of kdump failure,
3371 because some panic notifiers can make the crashed
3372 kernel more unstable.
3374 parkbd.port= [HW] Parallel port number the keyboard adapter is
3375 connected to, default is 0.
3377 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3378 0 for XT, 1 for AT (default is AT).
3381 parport= [HW,PPT] Specify parallel ports. 0 disables.
3382 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3383 Use 'auto' to force the driver to use any
3384 IRQ/DMA settings detected (the default is to
3385 ignore detected IRQ/DMA settings because of
3386 possible conflicts). You can specify the base
3387 address, IRQ, and DMA settings; IRQ and DMA
3388 should be numbers, or 'auto' (for using detected
3389 settings on that particular port), or 'nofifo'
3390 (to avoid using a FIFO even if it is detected).
3391 Parallel ports are assigned in the order they
3392 are specified on the command line, starting
3395 parport_init_mode= [HW,PPT]
3396 Configure VIA parallel port to operate in
3397 a specific mode. This is necessary on Pegasos
3398 computer where firmware has no options for setting
3399 up parallel port mode and sets it to spp.
3400 Currently this function knows 686a and 8231 chips.
3401 Format: [spp|ps2|epp|ecp|ecpepp]
3404 Halt all CPUs after the first oops has been printed for
3405 the specified number of seconds. This is to be used if
3406 your oopses keep scrolling off the screen.
3411 See header of drivers/block/paride/pcd.c.
3412 See also Documentation/admin-guide/blockdev/paride.rst.
3414 pci=option[,option...] [PCI] various PCI subsystem options.
3416 Some options herein operate on a specific device
3417 or a set of devices (<pci_dev>). These are
3418 specified in one of the following formats:
3420 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3421 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3423 Note: the first format specifies a PCI
3424 bus/device/function address which may change
3425 if new hardware is inserted, if motherboard
3426 firmware changes, or due to changes caused
3427 by other kernel parameters. If the
3428 domain is left unspecified, it is
3429 taken to be zero. Optionally, a path
3430 to a device through multiple device/function
3431 addresses can be specified after the base
3432 address (this is more robust against
3433 renumbering issues). The second format
3434 selects devices using IDs from the
3435 configuration space which may match multiple
3436 devices in the system.
3438 earlydump dump PCI config space before the kernel
3440 off [X86] don't probe for the PCI bus
3441 bios [X86-32] force use of PCI BIOS, don't access
3442 the hardware directly. Use this if your machine
3443 has a non-standard PCI host bridge.
3444 nobios [X86-32] disallow use of PCI BIOS, only direct
3445 hardware access methods are allowed. Use this
3446 if you experience crashes upon bootup and you
3447 suspect they are caused by the BIOS.
3448 conf1 [X86] Force use of PCI Configuration Access
3449 Mechanism 1 (config address in IO port 0xCF8,
3450 data in IO port 0xCFC, both 32-bit).
3451 conf2 [X86] Force use of PCI Configuration Access
3452 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3453 the function, IO port 0xCFA, also 8-bit, sets
3454 bus number. The config space is then accessed
3455 through ports 0xC000-0xCFFF).
3456 See http://wiki.osdev.org/PCI for more info
3457 on the configuration access mechanisms.
3458 noaer [PCIE] If the PCIEAER kernel config parameter is
3459 enabled, this kernel boot option can be used to
3460 disable the use of PCIE advanced error reporting.
3461 nodomains [PCI] Disable support for multiple PCI
3462 root domains (aka PCI segments, in ACPI-speak).
3463 nommconf [X86] Disable use of MMCONFIG for PCI
3465 check_enable_amd_mmconf [X86] check for and enable
3466 properly configured MMIO access to PCI
3467 config space on AMD family 10h CPU
3468 nomsi [MSI] If the PCI_MSI kernel config parameter is
3469 enabled, this kernel boot option can be used to
3470 disable the use of MSI interrupts system-wide.
3471 noioapicquirk [APIC] Disable all boot interrupt quirks.
3472 Safety option to keep boot IRQs enabled. This
3473 should never be necessary.
3474 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3475 primary IO-APIC for bridges that cannot disable
3476 boot IRQs. This fixes a source of spurious IRQs
3477 when the system masks IRQs.
3478 noioapicreroute [APIC] Disable workaround that uses the
3479 boot IRQ equivalent of an IRQ that connects to
3480 a chipset where boot IRQs cannot be disabled.
3481 The opposite of ioapicreroute.
3482 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3483 routing table. These calls are known to be buggy
3484 on several machines and they hang the machine
3485 when used, but on other computers it's the only
3486 way to get the interrupt routing table. Try
3487 this option if the kernel is unable to allocate
3488 IRQs or discover secondary PCI buses on your
3490 rom [X86] Assign address space to expansion ROMs.
3491 Use with caution as certain devices share
3492 address decoders between ROMs and other
3494 norom [X86] Do not assign address space to
3495 expansion ROMs that do not already have
3496 BIOS assigned address ranges.
3497 nobar [X86] Do not assign address space to the
3498 BARs that weren't assigned by the BIOS.
3499 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3500 assigned automatically to PCI devices. You can
3501 make the kernel exclude IRQs of your ISA cards
3503 pirqaddr=0xAAAAA [X86] Specify the physical address
3504 of the PIRQ table (normally generated
3505 by the BIOS) if it is outside the
3506 F0000h-100000h range.
3507 lastbus=N [X86] Scan all buses thru bus #N. Can be
3508 useful if the kernel is unable to find your
3509 secondary buses and you want to tell it
3510 explicitly which ones they are.
3511 assign-busses [X86] Always assign all PCI bus
3512 numbers ourselves, overriding
3513 whatever the firmware may have done.
3514 usepirqmask [X86] Honor the possible IRQ mask stored
3515 in the BIOS $PIR table. This is needed on
3516 some systems with broken BIOSes, notably
3517 some HP Pavilion N5400 and Omnibook XE3
3518 notebooks. This will have no effect if ACPI
3519 IRQ routing is enabled.
3520 noacpi [X86] Do not use ACPI for IRQ routing
3521 or for PCI scanning.
3522 use_crs [X86] Use PCI host bridge window information
3523 from ACPI. On BIOSes from 2008 or later, this
3524 is enabled by default. If you need to use this,
3525 please report a bug.
3526 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3527 If you need to use this, please report a bug.
3528 routeirq Do IRQ routing for all PCI devices.
3529 This is normally done in pci_enable_device(),
3530 so this option is a temporary workaround
3531 for broken drivers that don't call it.
3532 skip_isa_align [X86] do not align io start addr, so can
3533 handle more pci cards
3534 noearly [X86] Don't do any early type 1 scanning.
3535 This might help on some broken boards which
3536 machine check when some devices' config space
3537 is read. But various workarounds are disabled
3538 and some IOMMU drivers will not work.
3539 bfsort Sort PCI devices into breadth-first order.
3540 This sorting is done to get a device
3541 order compatible with older (<= 2.4) kernels.
3542 nobfsort Don't sort PCI devices into breadth-first order.
3543 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3544 tuning and use the BIOS-configured MPS defaults.
3545 pcie_bus_safe Set every device's MPS to the largest value
3546 supported by all devices below the root complex.
3547 pcie_bus_perf Set device MPS to the largest allowable MPS
3548 based on its parent bus. Also set MRRS (Max
3549 Read Request Size) to the largest supported
3550 value (no larger than the MPS that the device
3551 or bus can support) for best performance.
3552 pcie_bus_peer2peer Set every device's MPS to 128B, which
3553 every device is guaranteed to support. This
3554 configuration allows peer-to-peer DMA between
3555 any pair of devices, possibly at the cost of
3556 reduced performance. This also guarantees
3557 that hot-added devices will work.
3558 cbiosize=nn[KMG] The fixed amount of bus space which is
3559 reserved for the CardBus bridge's IO window.
3560 The default value is 256 bytes.
3561 cbmemsize=nn[KMG] The fixed amount of bus space which is
3562 reserved for the CardBus bridge's memory
3563 window. The default value is 64 megabytes.
3566 [<order of align>@]<pci_dev>[; ...]
3567 Specifies alignment and device to reassign
3568 aligned memory resources. How to
3569 specify the device is described above.
3570 If <order of align> is not specified,
3571 PAGE_SIZE is used as alignment.
3572 A PCI-PCI bridge can be specified if resource
3573 windows need to be expanded.
3574 To specify the alignment for several
3575 instances of a device, the PCI vendor,
3576 device, subvendor, and subdevice may be
3577 specified, e.g., 12@pci:8086:9c22:103c:198f
3578 for 4096-byte alignment.
3579 ecrc= Enable/disable PCIe ECRC (transaction layer
3580 end-to-end CRC checking).
3581 bios: Use BIOS/firmware settings. This is the
3585 hpiosize=nn[KMG] The fixed amount of bus space which is
3586 reserved for hotplug bridge's IO window.
3587 Default size is 256 bytes.
3588 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3589 reserved for hotplug bridge's MMIO window.
3590 Default size is 2 megabytes.
3591 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3592 reserved for hotplug bridge's MMIO_PREF window.
3593 Default size is 2 megabytes.
3594 hpmemsize=nn[KMG] The fixed amount of bus space which is
3595 reserved for hotplug bridge's MMIO and
3597 Default size is 2 megabytes.
3598 hpbussize=nn The minimum amount of additional bus numbers
3599 reserved for buses below a hotplug bridge.
3601 realloc= Enable/disable reallocating PCI bridge resources
3602 if allocations done by BIOS are too small to
3603 accommodate resources required by all child
3605 off: Turn realloc off
3607 realloc same as realloc=on
3608 noari do not use PCIe ARI.
3609 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3610 do not use PCIe ATS (and IOMMU device IOTLB).
3611 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3612 only look for one device below a PCIe downstream
3614 big_root_window Try to add a big 64bit memory window to the PCIe
3615 root complex on AMD CPUs. Some GFX hardware
3616 can resize a BAR to allow access to all VRAM.
3617 Adding the window is slightly risky (it may
3618 conflict with unreported devices), so this
3620 disable_acs_redir=<pci_dev>[; ...]
3621 Specify one or more PCI devices (in the format
3622 specified above) separated by semicolons.
3623 Each device specified will have the PCI ACS
3624 redirect capabilities forced off which will
3625 allow P2P traffic between devices through
3626 bridges without forcing it upstream. Note:
3627 this removes isolation between devices and
3628 may put more devices in an IOMMU group.
3629 force_floating [S390] Force usage of floating interrupts.
3630 nomio [S390] Do not use MIO instructions.
3632 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3635 force Enable ASPM even on devices that claim not to support it.
3636 WARNING: Forcing ASPM on may cause system lockups.
3638 pcie_ports= [PCIE] PCIe port services handling:
3639 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3640 even if the platform doesn't give the OS permission to
3641 use them. This may cause conflicts if the platform
3642 also tries to use these services.
3643 dpc-native Use native PCIe service for DPC only. May
3644 cause conflicts if firmware uses AER or DPC.
3645 compat Disable native PCIe services (PME, AER, DPC, PCIe
3648 pcie_port_pm= [PCIE] PCIe port power management handling:
3649 off Disable power management of all PCIe ports
3650 force Forcibly enable power management of all PCIe ports
3652 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3653 nomsi Do not use MSI for native PCIe PME signaling (this makes
3654 all PCIe root ports use INTx for all services).
3656 pcmv= [HW,PCMCIA] BadgePAD 4
3660 Keep all power-domains already enabled by bootloader on,
3661 even if no driver has claimed them. This is useful
3662 for debug and development, but should not be
3663 needed on a platform with proper driver support.
3666 See Documentation/admin-guide/blockdev/paride.rst.
3668 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3671 See arch/parisc/kernel/pdc_chassis.c
3673 percpu_alloc= Select which percpu first chunk allocator to use.
3674 Currently supported values are "embed" and "page".
3675 Archs may support subset or none of the selections.
3676 See comments in mm/percpu.c for details on each
3677 allocator. This parameter is primarily for debugging
3678 and performance comparison.
3681 See Documentation/admin-guide/blockdev/paride.rst.
3684 See Documentation/admin-guide/blockdev/paride.rst.
3686 pirq= [SMP,APIC] Manual mp-table setup
3687 See Documentation/x86/i386/IO-APIC.rst.
3689 plip= [PPT,NET] Parallel port network link
3690 Format: { parport<nr> | timid | 0 }
3691 See also Documentation/admin-guide/parport.rst.
3693 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3694 Override pmtimer IOPort with a hex value.
3698 Enable PNP debug messages (depends on the
3699 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3700 via /sys/module/pnp/parameters/debug. We always show
3701 current resource usage; turning this on also shows
3702 possible settings and some assignment information.
3708 { on | off | curr | res | no-curr | no-res }
3711 [ISAPNP] Exclude IRQs for the autoconfiguration
3714 [ISAPNP] Exclude DMAs for the autoconfiguration
3716 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3717 Ranges are in pairs (I/O port base and size).
3720 [ISAPNP] Exclude memory regions for the
3722 Ranges are in pairs (memory base and size).
3724 ports= [IP_VS_FTP] IPVS ftp helper module
3726 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3728 Format: <port>,<port>....
3730 powersave=off [PPC] This option disables power saving features.
3731 It specifically disables cpuidle and sets the
3732 platform machine description specific power_save
3733 function to NULL. On Idle the CPU just reduces
3736 ppc_strict_facility_enable
3737 [PPC] This option catches any kernel floating point,
3738 Altivec, VSX and SPE outside of regions specifically
3739 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3740 There is some performance impact when enabling this.
3744 Disable Hardware Transactional Memory
3746 print-fatal-signals=
3747 [KNL] debug: print fatal signals
3749 If enabled, warn about various signal handling
3750 related application anomalies: too many signals,
3751 too many POSIX.1 timers, fatal signals causing a
3754 If you hit the warning due to signal overflow,
3755 you might want to try "ulimit -i unlimited".
3759 printk.always_kmsg_dump=
3760 Trigger kmsg_dump for cases other than kernel oops or
3762 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3765 printk.devkmsg={on,off,ratelimit}
3766 Control writing to /dev/kmsg.
3767 on - unlimited logging to /dev/kmsg from userspace
3768 off - logging to /dev/kmsg disabled
3769 ratelimit - ratelimit the logging
3772 printk.time= Show timing data prefixed to each printk message line
3773 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3775 processor.max_cstate= [HW,ACPI]
3776 Limit processor to maximum C-state
3777 max_cstate=9 overrides any DMI blacklist limit.
3779 processor.nocst [HW,ACPI]
3780 Ignore the _CST method to determine C-states,
3781 instead using the legacy FADT method
3783 profile= [KNL] Enable kernel profiling via /proc/profile
3784 Format: [<profiletype>,]<number>
3785 Param: <profiletype>: "schedule", "sleep", or "kvm"
3786 [defaults to kernel profiling]
3787 Param: "schedule" - profile schedule points.
3788 Param: "sleep" - profile D-state sleeping (millisecs).
3789 Requires CONFIG_SCHEDSTATS
3790 Param: "kvm" - profile VM exits.
3791 Param: <number> - step/bucket size as a power of 2 for
3792 statistical time based profiling.
3794 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3796 See Documentation/admin-guide/blockdev/ramdisk.rst.
3798 prot_virt= [S390] enable hosting protected virtual machines
3799 isolated from the hypervisor (if hardware supports
3803 psi= [KNL] Enable or disable pressure stall information
3807 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3808 probe for; one of (bare|imps|exps|lifebook|any).
3809 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3811 psmouse.resetafter= [HW,MOUSE]
3812 Try to reset the device after so many bad packets
3815 [HW,MOUSE] Set desired mouse resolution, in dpi.
3816 psmouse.smartscroll=
3817 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3818 0 = disabled, 1 = enabled (default).
3820 pstore.backend= Specify the name of the pstore backend to use
3823 See Documentation/admin-guide/blockdev/paride.rst.
3825 pti= [X86_64] Control Page Table Isolation of user and
3826 kernel address spaces. Disabling this feature
3827 removes hardening, but improves performance of
3828 system calls and interrupts.
3830 on - unconditionally enable
3831 off - unconditionally disable
3832 auto - kernel detects whether your CPU model is
3833 vulnerable to issues that PTI mitigates
3835 Not specifying this option is equivalent to pti=auto.
3838 Equivalent to pti=off
3841 [KNL] Number of legacy pty's. Overwrites compiled-in
3844 quiet [KNL] Disable most log messages
3849 See Documentation/admin-guide/md.rst.
3851 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3852 See Documentation/admin-guide/blockdev/ramdisk.rst.
3854 random.trust_cpu={on,off}
3855 [KNL] Enable or disable trusting the use of the
3856 CPU's random number generator (if available) to
3857 fully seed the kernel's CRNG. Default is controlled
3858 by CONFIG_RANDOM_TRUST_CPU.
3860 ras=option[,option,...] [KNL] RAS-specific options
3863 Disable the Correctable Errors Collector,
3864 see CONFIG_RAS_CEC help text.
3867 The argument is a cpu list, as described above,
3868 except that the string "all" can be used to
3869 specify every CPU on the system.
3871 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3872 the specified list of CPUs to be no-callback CPUs.
3873 Invocation of these CPUs' RCU callbacks will be
3874 offloaded to "rcuox/N" kthreads created for that
3875 purpose, where "x" is "p" for RCU-preempt, and
3876 "s" for RCU-sched, and "N" is the CPU number.
3877 This reduces OS jitter on the offloaded CPUs,
3878 which can be useful for HPC and real-time
3879 workloads. It can also improve energy efficiency
3880 for asymmetric multiprocessors.
3883 Rather than requiring that offloaded CPUs
3884 (specified by rcu_nocbs= above) explicitly
3885 awaken the corresponding "rcuoN" kthreads,
3886 make these kthreads poll for callbacks.
3887 This improves the real-time response for the
3888 offloaded CPUs by relieving them of the need to
3889 wake up the corresponding kthread, but degrades
3890 energy efficiency by requiring that the kthreads
3891 periodically wake up to do the polling.
3893 rcutree.blimit= [KNL]
3894 Set maximum number of finished RCU callbacks to
3895 process in one batch.
3897 rcutree.dump_tree= [KNL]
3898 Dump the structure of the rcu_node combining tree
3899 out at early boot. This is used for diagnostic
3900 purposes, to verify correct tree setup.
3902 rcutree.gp_cleanup_delay= [KNL]
3903 Set the number of jiffies to delay each step of
3904 RCU grace-period cleanup.
3906 rcutree.gp_init_delay= [KNL]
3907 Set the number of jiffies to delay each step of
3908 RCU grace-period initialization.
3910 rcutree.gp_preinit_delay= [KNL]
3911 Set the number of jiffies to delay each step of
3912 RCU grace-period pre-initialization, that is,
3913 the propagation of recent CPU-hotplug changes up
3914 the rcu_node combining tree.
3916 rcutree.use_softirq= [KNL]
3917 If set to zero, move all RCU_SOFTIRQ processing to
3918 per-CPU rcuc kthreads. Defaults to a non-zero
3919 value, meaning that RCU_SOFTIRQ is used by default.
3920 Specify rcutree.use_softirq=0 to use rcuc kthreads.
3922 rcutree.rcu_fanout_exact= [KNL]
3923 Disable autobalancing of the rcu_node combining
3924 tree. This is used by rcutorture, and might
3925 possibly be useful for architectures having high
3926 cache-to-cache transfer latencies.
3928 rcutree.rcu_fanout_leaf= [KNL]
3929 Change the number of CPUs assigned to each
3930 leaf rcu_node structure. Useful for very
3931 large systems, which will choose the value 64,
3932 and for NUMA systems with large remote-access
3933 latencies, which will choose a value aligned
3934 with the appropriate hardware boundaries.
3936 rcutree.jiffies_till_first_fqs= [KNL]
3937 Set delay from grace-period initialization to
3938 first attempt to force quiescent states.
3939 Units are jiffies, minimum value is zero,
3940 and maximum value is HZ.
3942 rcutree.jiffies_till_next_fqs= [KNL]
3943 Set delay between subsequent attempts to force
3944 quiescent states. Units are jiffies, minimum
3945 value is one, and maximum value is HZ.
3947 rcutree.jiffies_till_sched_qs= [KNL]
3948 Set required age in jiffies for a
3949 given grace period before RCU starts
3950 soliciting quiescent-state help from
3951 rcu_note_context_switch() and cond_resched().
3952 If not specified, the kernel will calculate
3953 a value based on the most recent settings
3954 of rcutree.jiffies_till_first_fqs
3955 and rcutree.jiffies_till_next_fqs.
3956 This calculated value may be viewed in
3957 rcutree.jiffies_to_sched_qs. Any attempt to set
3958 rcutree.jiffies_to_sched_qs will be cheerfully
3961 rcutree.kthread_prio= [KNL,BOOT]
3962 Set the SCHED_FIFO priority of the RCU per-CPU
3963 kthreads (rcuc/N). This value is also used for
3964 the priority of the RCU boost threads (rcub/N)
3965 and for the RCU grace-period kthreads (rcu_bh,
3966 rcu_preempt, and rcu_sched). If RCU_BOOST is
3967 set, valid values are 1-99 and the default is 1
3968 (the least-favored priority). Otherwise, when
3969 RCU_BOOST is not set, valid values are 0-99 and
3970 the default is zero (non-realtime operation).
3972 rcutree.rcu_nocb_gp_stride= [KNL]
3973 Set the number of NOCB callback kthreads in
3974 each group, which defaults to the square root
3975 of the number of CPUs. Larger numbers reduce
3976 the wakeup overhead on the global grace-period
3977 kthread, but increases that same overhead on
3978 each group's NOCB grace-period kthread.
3980 rcutree.qhimark= [KNL]
3981 Set threshold of queued RCU callbacks beyond which
3982 batch limiting is disabled.
3984 rcutree.qlowmark= [KNL]
3985 Set threshold of queued RCU callbacks below which
3986 batch limiting is re-enabled.
3988 rcutree.rcu_idle_gp_delay= [KNL]
3989 Set wakeup interval for idle CPUs that have
3990 RCU callbacks (RCU_FAST_NO_HZ=y).
3992 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3993 Set wakeup interval for idle CPUs that have
3994 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3995 Lazy RCU callbacks are those which RCU can
3996 prove do nothing more than free memory.
3998 rcutree.rcu_kick_kthreads= [KNL]
3999 Cause the grace-period kthread to get an extra
4000 wake_up() if it sleeps three times longer than
4001 it should at force-quiescent-state time.
4002 This wake_up() will be accompanied by a
4003 WARN_ONCE() splat and an ftrace_dump().
4005 rcutree.sysrq_rcu= [KNL]
4006 Commandeer a sysrq key to dump out Tree RCU's
4007 rcu_node tree with an eye towards determining
4008 why a new grace period has not yet started.
4010 rcuperf.gp_async= [KNL]
4011 Measure performance of asynchronous
4012 grace-period primitives such as call_rcu().
4014 rcuperf.gp_async_max= [KNL]
4015 Specify the maximum number of outstanding
4016 callbacks per writer thread. When a writer
4017 thread exceeds this limit, it invokes the
4018 corresponding flavor of rcu_barrier() to allow
4019 previously posted callbacks to drain.
4021 rcuperf.gp_exp= [KNL]
4022 Measure performance of expedited synchronous
4023 grace-period primitives.
4025 rcuperf.holdoff= [KNL]
4026 Set test-start holdoff period. The purpose of
4027 this parameter is to delay the start of the
4028 test until boot completes in order to avoid
4031 rcuperf.kfree_rcu_test= [KNL]
4032 Set to measure performance of kfree_rcu() flooding.
4034 rcuperf.kfree_nthreads= [KNL]
4035 The number of threads running loops of kfree_rcu().
4037 rcuperf.kfree_alloc_num= [KNL]
4038 Number of allocations and frees done in an iteration.
4040 rcuperf.kfree_loops= [KNL]
4041 Number of loops doing rcuperf.kfree_alloc_num number
4042 of allocations and frees.
4044 rcuperf.nreaders= [KNL]
4045 Set number of RCU readers. The value -1 selects
4046 N, where N is the number of CPUs. A value
4047 "n" less than -1 selects N-n+1, where N is again
4048 the number of CPUs. For example, -2 selects N
4049 (the number of CPUs), -3 selects N+1, and so on.
4050 A value of "n" less than or equal to -N selects
4053 rcuperf.nwriters= [KNL]
4054 Set number of RCU writers. The values operate
4055 the same as for rcuperf.nreaders.
4056 N, where N is the number of CPUs
4058 rcuperf.perf_type= [KNL]
4059 Specify the RCU implementation to test.
4061 rcuperf.shutdown= [KNL]
4062 Shut the system down after performance tests
4063 complete. This is useful for hands-off automated
4066 rcuperf.verbose= [KNL]
4067 Enable additional printk() statements.
4069 rcuperf.writer_holdoff= [KNL]
4070 Write-side holdoff between grace periods,
4071 in microseconds. The default of zero says
4074 rcutorture.fqs_duration= [KNL]
4075 Set duration of force_quiescent_state bursts
4078 rcutorture.fqs_holdoff= [KNL]
4079 Set holdoff time within force_quiescent_state bursts
4082 rcutorture.fqs_stutter= [KNL]
4083 Set wait time between force_quiescent_state bursts
4086 rcutorture.fwd_progress= [KNL]
4087 Enable RCU grace-period forward-progress testing
4088 for the types of RCU supporting this notion.
4090 rcutorture.fwd_progress_div= [KNL]
4091 Specify the fraction of a CPU-stall-warning
4092 period to do tight-loop forward-progress testing.
4094 rcutorture.fwd_progress_holdoff= [KNL]
4095 Number of seconds to wait between successive
4096 forward-progress tests.
4098 rcutorture.fwd_progress_need_resched= [KNL]
4099 Enclose cond_resched() calls within checks for
4100 need_resched() during tight-loop forward-progress
4103 rcutorture.gp_cond= [KNL]
4104 Use conditional/asynchronous update-side
4105 primitives, if available.
4107 rcutorture.gp_exp= [KNL]
4108 Use expedited update-side primitives, if available.
4110 rcutorture.gp_normal= [KNL]
4111 Use normal (non-expedited) asynchronous
4112 update-side primitives, if available.
4114 rcutorture.gp_sync= [KNL]
4115 Use normal (non-expedited) synchronous
4116 update-side primitives, if available. If all
4117 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4118 rcutorture.gp_normal=, and rcutorture.gp_sync=
4119 are zero, rcutorture acts as if is interpreted
4120 they are all non-zero.
4122 rcutorture.n_barrier_cbs= [KNL]
4123 Set callbacks/threads for rcu_barrier() testing.
4125 rcutorture.nfakewriters= [KNL]
4126 Set number of concurrent RCU writers. These just
4127 stress RCU, they don't participate in the actual
4128 test, hence the "fake".
4130 rcutorture.nreaders= [KNL]
4131 Set number of RCU readers. The value -1 selects
4132 N-1, where N is the number of CPUs. A value
4133 "n" less than -1 selects N-n-2, where N is again
4134 the number of CPUs. For example, -2 selects N
4135 (the number of CPUs), -3 selects N+1, and so on.
4137 rcutorture.object_debug= [KNL]
4138 Enable debug-object double-call_rcu() testing.
4140 rcutorture.onoff_holdoff= [KNL]
4141 Set time (s) after boot for CPU-hotplug testing.
4143 rcutorture.onoff_interval= [KNL]
4144 Set time (jiffies) between CPU-hotplug operations,
4145 or zero to disable CPU-hotplug testing.
4147 rcutorture.shuffle_interval= [KNL]
4148 Set task-shuffle interval (s). Shuffling tasks
4149 allows some CPUs to go into dyntick-idle mode
4150 during the rcutorture test.
4152 rcutorture.shutdown_secs= [KNL]
4153 Set time (s) after boot system shutdown. This
4154 is useful for hands-off automated testing.
4156 rcutorture.stall_cpu= [KNL]
4157 Duration of CPU stall (s) to test RCU CPU stall
4158 warnings, zero to disable.
4160 rcutorture.stall_cpu_holdoff= [KNL]
4161 Time to wait (s) after boot before inducing stall.
4163 rcutorture.stall_cpu_irqsoff= [KNL]
4164 Disable interrupts while stalling if set.
4166 rcutorture.stat_interval= [KNL]
4167 Time (s) between statistics printk()s.
4169 rcutorture.stutter= [KNL]
4170 Time (s) to stutter testing, for example, specifying
4171 five seconds causes the test to run for five seconds,
4172 wait for five seconds, and so on. This tests RCU's
4173 ability to transition abruptly to and from idle.
4175 rcutorture.test_boost= [KNL]
4176 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4177 "Maybe" means test if the RCU implementation
4178 under test support RCU priority boosting.
4180 rcutorture.test_boost_duration= [KNL]
4181 Duration (s) of each individual boost test.
4183 rcutorture.test_boost_interval= [KNL]
4184 Interval (s) between each boost test.
4186 rcutorture.test_no_idle_hz= [KNL]
4187 Test RCU's dyntick-idle handling. See also the
4188 rcutorture.shuffle_interval parameter.
4190 rcutorture.torture_type= [KNL]
4191 Specify the RCU implementation to test.
4193 rcutorture.verbose= [KNL]
4194 Enable additional printk() statements.
4196 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4197 Dump ftrace buffer after reporting RCU CPU
4200 rcupdate.rcu_cpu_stall_suppress= [KNL]
4201 Suppress RCU CPU stall warning messages.
4203 rcupdate.rcu_cpu_stall_timeout= [KNL]
4204 Set timeout for RCU CPU stall warning messages.
4206 rcupdate.rcu_expedited= [KNL]
4207 Use expedited grace-period primitives, for
4208 example, synchronize_rcu_expedited() instead
4209 of synchronize_rcu(). This reduces latency,
4210 but can increase CPU utilization, degrade
4211 real-time latency, and degrade energy efficiency.
4212 No effect on CONFIG_TINY_RCU kernels.
4214 rcupdate.rcu_normal= [KNL]
4215 Use only normal grace-period primitives,
4216 for example, synchronize_rcu() instead of
4217 synchronize_rcu_expedited(). This improves
4218 real-time latency, CPU utilization, and
4219 energy efficiency, but can expose users to
4220 increased grace-period latency. This parameter
4221 overrides rcupdate.rcu_expedited. No effect on
4222 CONFIG_TINY_RCU kernels.
4224 rcupdate.rcu_normal_after_boot= [KNL]
4225 Once boot has completed (that is, after
4226 rcu_end_inkernel_boot() has been invoked), use
4227 only normal grace-period primitives. No effect
4228 on CONFIG_TINY_RCU kernels.
4230 rcupdate.rcu_task_stall_timeout= [KNL]
4231 Set timeout in jiffies for RCU task stall warning
4232 messages. Disable with a value less than or equal
4235 rcupdate.rcu_self_test= [KNL]
4236 Run the RCU early boot self tests
4240 Run specified binary instead of /init from the ramdisk,
4241 used for early userspace startup. See initrd.
4244 force - Override the decision by the kernel to hide the
4245 advertisement of RDRAND support (this affects
4246 certain AMD processors because of buggy BIOS
4247 support, specifically around the suspend/resume
4251 Turn on/off individual RDT features. List is:
4252 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4254 E.g. to turn on cmt and turn off mba use:
4258 Format (x86 or x86_64):
4259 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4261 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4263 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4264 (prefix with 'panic_' to set mode for panic
4266 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4267 reboot_force is either force or not specified,
4268 reboot_cpu is s[mp]#### with #### being the processor
4269 to be used for rebooting.
4272 [KNL, SMP] Set scheduler's default relax_domain_level.
4273 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4275 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4276 Format: <base1>,<size1>[,<base2>,<size2>,...]
4277 Reserve I/O ports or memory so the kernel won't use
4278 them. If <base> is less than 0x10000, the region
4279 is assumed to be I/O ports; otherwise it is memory.
4281 reservetop= [X86-32]
4283 Reserves a hole at the top of the kernel virtual
4288 Set the amount of memory to reserve for BIOS at
4289 the bottom of the address space.
4291 reset_devices [KNL] Force drivers to reset the underlying device
4292 during initialization.
4295 Specify the partition device for software suspend
4297 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4299 resume_offset= [SWSUSP]
4300 Specify the offset from the beginning of the partition
4301 given by "resume=" at which the swap header is located,
4302 in <PAGE_SIZE> units (needed only for swap files).
4303 See Documentation/power/swsusp-and-swap-files.rst
4305 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4306 read the resume files
4308 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4309 Useful for devices that are detected asynchronously
4310 (e.g. USB and MMC devices).
4312 hibernate= [HIBERNATION]
4313 noresume Don't check if there's a hibernation image
4314 present during boot.
4315 nocompress Don't compress/decompress hibernation images.
4316 no Disable hibernation and resume.
4317 protect_image Turn on image protection during restoration
4318 (that will set all pages holding image data
4319 during restoration read-only).
4321 retain_initrd [RAM] Keep initrd memory after extraction
4323 rfkill.default_state=
4324 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4325 etc. communication is blocked by default.
4328 rfkill.master_switch_mode=
4329 0 The "airplane mode" button does nothing.
4330 1 The "airplane mode" button toggles between everything
4331 blocked and the previous configuration.
4332 2 The "airplane mode" button toggles between everything
4333 blocked and everything unblocked.
4335 rhash_entries= [KNL,NET]
4336 Set number of hash buckets for route cache
4339 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4342 ro [KNL] Mount root device read-only on boot
4345 on Mark read-only kernel memory as read-only (default).
4346 off Leave read-only kernel memory writable for debugging.
4349 Enable the uart passthrough on the designated usb port
4350 on Rockchip SoCs. When active, the signals of the
4351 debug-uart get routed to the D+ and D- pins of the usb
4352 port and the regular usb controller gets disabled.
4354 root= [KNL] Root filesystem
4355 See name_to_dev_t comment in init/do_mounts.c.
4357 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4358 mount the root filesystem
4360 rootflags= [KNL] Set root filesystem mount option string
4362 rootfstype= [KNL] Set root filesystem type
4364 rootwait [KNL] Wait (indefinitely) for root device to show up.
4365 Useful for devices that are detected asynchronously
4366 (e.g. USB and MMC devices).
4368 rproc_mem=nn[KMG][@address]
4369 [KNL,ARM,CMA] Remoteproc physical memory block.
4370 Memory area to be used by remote processor image,
4373 rw [KNL] Mount root device read-write on boot
4375 S [KNL] Run init in single mode
4377 s390_iommu= [HW,S390]
4378 Set s390 IOTLB flushing mode
4380 With strict flushing every unmap operation will result in
4381 an IOTLB flush. Default is lazy flushing before reuse,
4385 See drivers/net/irda/sa1100_ir.c.
4387 sbni= [NET] Granch SBNI12 leased line adapter
4389 sched_debug [KNL] Enables verbose scheduler debug messages.
4391 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4392 Allowed values are enable and disable. This feature
4393 incurs a small amount of overhead in the scheduler
4394 but is useful for debugging and performance tuning.
4396 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4397 xtime_lock contention on larger systems, and/or RCU lock
4398 contention on all systems with CONFIG_MAXSMP set.
4399 Format: { "0" | "1" }
4400 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4402 Note: increases power consumption, thus should only be
4403 enabled if running jitter sensitive (HPC/RT) workloads.
4405 security= [SECURITY] Choose a legacy "major" security module to
4406 enable at boot. This has been deprecated by the
4409 selinux= [SELINUX] Disable or enable SELinux at boot time.
4410 Format: { "0" | "1" }
4411 See security/selinux/Kconfig help text.
4416 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4417 Format: { "0" | "1" }
4418 See security/apparmor/Kconfig help text
4421 Default value is set via kernel config option.
4423 serialnumber [BUGS=X86-32]
4426 Maximal number of shapers.
4434 Disable merging of slabs with similar size. May be
4435 necessary if there is some reason to distinguish
4436 allocs to different slabs, especially in hardened
4437 environments where the risk of heap overflows and
4438 layout control by attackers can usually be
4439 frustrated by disabling merging. This will reduce
4440 most of the exposure of a heap attack to a single
4441 cache (risks via metadata attacks are mostly
4442 unchanged). Debug options disable merging on their
4444 For more information see Documentation/vm/slub.rst.
4446 slab_max_order= [MM, SLAB]
4447 Determines the maximum allowed order for slabs.
4448 A high setting may cause OOMs due to memory
4449 fragmentation. Defaults to 1 for systems with
4450 more than 32MB of RAM, 0 otherwise.
4452 slub_debug[=options[,slabs]] [MM, SLUB]
4453 Enabling slub_debug allows one to determine the
4454 culprit if slab objects become corrupted. Enabling
4455 slub_debug can create guard zones around objects and
4456 may poison objects when not in use. Also tracks the
4457 last alloc / free. For more information see
4458 Documentation/vm/slub.rst.
4460 slub_memcg_sysfs= [MM, SLUB]
4461 Determines whether to enable sysfs directories for
4462 memory cgroup sub-caches. 1 to enable, 0 to disable.
4463 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4464 Enabling this can lead to a very high number of debug
4465 directories and files being created under
4468 slub_max_order= [MM, SLUB]
4469 Determines the maximum allowed order for slabs.
4470 A high setting may cause OOMs due to memory
4471 fragmentation. For more information see
4472 Documentation/vm/slub.rst.
4474 slub_min_objects= [MM, SLUB]
4475 The minimum number of objects per slab. SLUB will
4476 increase the slab order up to slub_max_order to
4477 generate a sufficiently large slab able to contain
4478 the number of objects indicated. The higher the number
4479 of objects the smaller the overhead of tracking slabs
4480 and the less frequently locks need to be acquired.
4481 For more information see Documentation/vm/slub.rst.
4483 slub_min_order= [MM, SLUB]
4484 Determines the minimum page order for slabs. Must be
4485 lower than slub_max_order.
4486 For more information see Documentation/vm/slub.rst.
4488 slub_nomerge [MM, SLUB]
4489 Same with slab_nomerge. This is supported for legacy.
4490 See slab_nomerge for more information.
4493 Format: <io1>[,<io2>[,...,<io8>]]
4495 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4496 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4497 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4498 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4499 smsc-ircc2.ircc_irq= [HW] IRQ line
4500 smsc-ircc2.ircc_dma= [HW] DMA channel
4501 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4502 0: Toshiba Satellite 1800 (GP data pin select)
4503 1: Fast pin select (default)
4506 smt [KNL,S390] Set the maximum number of threads (logical
4507 CPUs) to use per physical CPU on systems capable of
4508 symmetric multithreading (SMT). Will be capped to the
4509 actual hardware limit.
4511 Default: -1 (no limit)
4514 [KNL] Should the soft-lockup detector generate panics.
4517 A nonzero value instructs the soft-lockup detector
4518 to panic the machine when a soft-lockup occurs. This
4519 is also controlled by CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC
4520 which is the respective build-time switch to that
4523 softlockup_all_cpu_backtrace=
4524 [KNL] Should the soft-lockup detector generate
4525 backtraces on all cpus.
4528 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4529 See Documentation/admin-guide/laptops/sonypi.rst
4531 spectre_v2= [X86] Control mitigation of Spectre variant 2
4532 (indirect branch speculation) vulnerability.
4533 The default operation protects the kernel from
4536 on - unconditionally enable, implies
4538 off - unconditionally disable, implies
4540 auto - kernel detects whether your CPU model is
4543 Selecting 'on' will, and 'auto' may, choose a
4544 mitigation method at run time according to the
4545 CPU, the available microcode, the setting of the
4546 CONFIG_RETPOLINE configuration option, and the
4547 compiler with which the kernel was built.
4549 Selecting 'on' will also enable the mitigation
4550 against user space to user space task attacks.
4552 Selecting 'off' will disable both the kernel and
4553 the user space protections.
4555 Specific mitigations can also be selected manually:
4557 retpoline - replace indirect branches
4558 retpoline,generic - google's original retpoline
4559 retpoline,amd - AMD-specific minimal thunk
4561 Not specifying this option is equivalent to
4565 [X86] Control mitigation of Spectre variant 2
4566 (indirect branch speculation) vulnerability between
4569 on - Unconditionally enable mitigations. Is
4570 enforced by spectre_v2=on
4572 off - Unconditionally disable mitigations. Is
4573 enforced by spectre_v2=off
4575 prctl - Indirect branch speculation is enabled,
4576 but mitigation can be enabled via prctl
4577 per thread. The mitigation control state
4578 is inherited on fork.
4581 - Like "prctl" above, but only STIBP is
4582 controlled per thread. IBPB is issued
4583 always when switching between different user
4587 - Same as "prctl" above, but all seccomp
4588 threads will enable the mitigation unless
4589 they explicitly opt out.
4592 - Like "seccomp" above, but only STIBP is
4593 controlled per thread. IBPB is issued
4594 always when switching between different
4595 user space processes.
4597 auto - Kernel selects the mitigation depending on
4598 the available CPU features and vulnerability.
4601 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4603 Not specifying this option is equivalent to
4604 spectre_v2_user=auto.
4606 spec_store_bypass_disable=
4607 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4608 (Speculative Store Bypass vulnerability)
4610 Certain CPUs are vulnerable to an exploit against a
4611 a common industry wide performance optimization known
4612 as "Speculative Store Bypass" in which recent stores
4613 to the same memory location may not be observed by
4614 later loads during speculative execution. The idea
4615 is that such stores are unlikely and that they can
4616 be detected prior to instruction retirement at the
4617 end of a particular speculation execution window.
4619 In vulnerable processors, the speculatively forwarded
4620 store can be used in a cache side channel attack, for
4621 example to read memory to which the attacker does not
4622 directly have access (e.g. inside sandboxed code).
4624 This parameter controls whether the Speculative Store
4625 Bypass optimization is used.
4627 On x86 the options are:
4629 on - Unconditionally disable Speculative Store Bypass
4630 off - Unconditionally enable Speculative Store Bypass
4631 auto - Kernel detects whether the CPU model contains an
4632 implementation of Speculative Store Bypass and
4633 picks the most appropriate mitigation. If the
4634 CPU is not vulnerable, "off" is selected. If the
4635 CPU is vulnerable the default mitigation is
4636 architecture and Kconfig dependent. See below.
4637 prctl - Control Speculative Store Bypass per thread
4638 via prctl. Speculative Store Bypass is enabled
4639 for a process by default. The state of the control
4640 is inherited on fork.
4641 seccomp - Same as "prctl" above, but all seccomp threads
4642 will disable SSB unless they explicitly opt out.
4644 Default mitigations:
4645 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4647 On powerpc the options are:
4649 on,auto - On Power8 and Power9 insert a store-forwarding
4650 barrier on kernel entry and exit. On Power7
4651 perform a software flush on kernel entry and
4655 Not specifying this option is equivalent to
4656 spec_store_bypass_disable=auto.
4658 spia_io_base= [HW,MTD]
4663 srcutree.counter_wrap_check [KNL]
4664 Specifies how frequently to check for
4665 grace-period sequence counter wrap for the
4666 srcu_data structure's ->srcu_gp_seq_needed field.
4667 The greater the number of bits set in this kernel
4668 parameter, the less frequently counter wrap will
4669 be checked for. Note that the bottom two bits
4672 srcutree.exp_holdoff [KNL]
4673 Specifies how many nanoseconds must elapse
4674 since the end of the last SRCU grace period for
4675 a given srcu_struct until the next normal SRCU
4676 grace period will be considered for automatic
4677 expediting. Set to zero to disable automatic
4681 Speculative Store Bypass Disable control
4683 On CPUs that are vulnerable to the Speculative
4684 Store Bypass vulnerability and offer a
4685 firmware based mitigation, this parameter
4686 indicates how the mitigation should be used:
4688 force-on: Unconditionally enable mitigation for
4689 for both kernel and userspace
4690 force-off: Unconditionally disable mitigation for
4691 for both kernel and userspace
4692 kernel: Always enable mitigation in the
4693 kernel, and offer a prctl interface
4694 to allow userspace to register its
4695 interest in being mitigated too.
4697 stack_guard_gap= [MM]
4698 override the default stack gap protection. The value
4699 is in page units and it defines how many pages prior
4700 to (for stacks growing down) resp. after (for stacks
4701 growing up) the main stack are reserved for no other
4702 mapping. Default value is 256 pages.
4705 Enabled the stack tracer on boot up.
4707 stacktrace_filter=[function-list]
4708 [FTRACE] Limit the functions that the stack tracer
4709 will trace at boot up. function-list is a comma separated
4710 list of functions. This list can be changed at run
4711 time by the stack_trace_filter file in the debugfs
4712 tracing directory. Note, this enables stack tracing
4713 and the stacktrace above is not needed.
4717 Set the STI (builtin display/keyboard on the HP-PARISC
4718 machines) console (graphic card) which should be used
4719 as the initial boot-console.
4720 See also comment in drivers/video/console/sticore.c.
4723 See comment in drivers/video/console/sticore.c.
4726 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4728 sunrpc.min_resvport=
4729 sunrpc.max_resvport=
4731 SunRPC servers often require that client requests
4732 originate from a privileged port (i.e. a port in the
4733 range 0 < portnr < 1024).
4734 An administrator who wishes to reserve some of these
4735 ports for other uses may adjust the range that the
4736 kernel's sunrpc client considers to be privileged
4737 using these two parameters to set the minimum and
4738 maximum port values.
4740 sunrpc.svc_rpc_per_connection_limit=
4742 Limit the number of requests that the server will
4743 process in parallel from a single connection.
4744 The default value is 0 (no limit).
4748 Control how the NFS server code allocates CPUs to
4749 service thread pools. Depending on how many NICs
4750 you have and where their interrupts are bound, this
4751 option will affect which CPUs will do NFS serving.
4752 Note: this parameter cannot be changed while the
4753 NFS server is running.
4755 auto the server chooses an appropriate mode
4756 automatically using heuristics
4757 global a single global pool contains all CPUs
4758 percpu one pool for each CPU
4759 pernode one pool for each NUMA node (equivalent
4760 to global on non-NUMA machines)
4762 sunrpc.tcp_slot_table_entries=
4763 sunrpc.udp_slot_table_entries=
4765 Sets the upper limit on the number of simultaneous
4766 RPC calls that can be sent from the client to a
4767 server. Increasing these values may allow you to
4768 improve throughput, but will also increase the
4769 amount of memory reserved for use by the client.
4771 suspend.pm_test_delay=
4773 Sets the number of seconds to remain in a suspend test
4774 mode before resuming the system (see
4775 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4776 is set. Default value is 5.
4779 Format: { on | off | y | n | 1 | 0 }
4780 This parameter controls use of the Protected
4781 Execution Facility on pSeries.
4784 [KNL] Enable accounting of swap in memory resource
4785 controller if no parameter or 1 is given or disable
4786 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
4788 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4789 Format: { <int> | force | noforce }
4790 <int> -- Number of I/O TLB slabs
4791 force -- force using of bounce buffers even if they
4792 wouldn't be automatically used by the kernel
4793 noforce -- Never use bounce buffers (for debugging)
4797 sysfs.deprecated=0|1 [KNL]
4798 Enable/disable old style sysfs layout for old udev
4799 on older distributions. When this option is enabled
4800 very new udev will not work anymore. When this option
4801 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4802 in older udev will not work anymore.
4803 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4804 the kernel configuration.
4806 sysrq_always_enabled
4808 Ignore sysrq setting - this boot parameter will
4809 neutralize any effect of /proc/sys/kernel/sysrq.
4810 Useful for debugging.
4812 tcpmhash_entries= [KNL,NET]
4813 Set the number of tcp_metrics_hash slots.
4814 Default value is 8192 or 16384 depending on total
4815 ram pages. This is used to specify the TCP metrics
4816 cache size. See Documentation/networking/ip-sysctl.txt
4817 "tcp_no_metrics_save" section for more details.
4821 test_suspend= [SUSPEND][,N]
4822 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4823 standby suspend) or "freeze" (for suspend type freeze)
4824 as the system sleep state during system startup with
4825 the optional capability to repeat N number of times.
4826 The system is woken from this state using a
4827 wakeup-capable RTC alarm.
4829 thash_entries= [KNL,NET]
4830 Set number of hash buckets for TCP connection
4832 thermal.act= [HW,ACPI]
4833 -1: disable all active trip points in all thermal zones
4834 <degrees C>: override all lowest active trip points
4836 thermal.crt= [HW,ACPI]
4837 -1: disable all critical trip points in all thermal zones
4838 <degrees C>: override all critical trip points
4840 thermal.nocrt= [HW,ACPI]
4841 Set to disable actions on ACPI thermal zone
4842 critical and hot trip points.
4844 thermal.off= [HW,ACPI]
4845 1: disable ACPI thermal control
4847 thermal.psv= [HW,ACPI]
4848 -1: disable all passive trip points
4849 <degrees C>: override all passive trip points to this
4852 thermal.tzp= [HW,ACPI]
4853 Specify global default ACPI thermal zone polling rate
4854 <deci-seconds>: poll all this frequency
4855 0: no polling (default)
4858 Force threading of all interrupt handlers except those
4859 marked explicitly IRQF_NO_THREAD.
4863 Specify if the kernel should make use of the cpu
4864 topology information if the hardware supports this.
4865 The scheduler will make use of this information and
4866 e.g. base its process migration decisions on it.
4869 topology_updates= [KNL, PPC, NUMA]
4871 Specify if the kernel should ignore (off)
4872 topology updates sent by the hypervisor to this
4877 tpm_suspend_pcr=[HW,TPM]
4878 Format: integer pcr id
4879 Specify that at suspend time, the tpm driver
4880 should extend the specified pcr with zeros,
4881 as a workaround for some chips which fail to
4882 flush the last written pcr on TPM_SaveState.
4883 This will guarantee that all the other pcrs
4886 trace_buf_size=nn[KMG]
4887 [FTRACE] will set tracing buffer size on each cpu.
4889 trace_event=[event-list]
4890 [FTRACE] Set and start specified trace events in order
4891 to facilitate early boot debugging. The event-list is a
4892 comma separated list of trace events to enable. See
4893 also Documentation/trace/events.rst
4895 trace_options=[option-list]
4896 [FTRACE] Enable or disable tracer options at boot.
4897 The option-list is a comma delimited list of options
4898 that can be enabled or disabled just as if you were
4899 to echo the option name into
4901 /sys/kernel/debug/tracing/trace_options
4903 For example, to enable stacktrace option (to dump the
4904 stack trace of each event), add to the command line:
4906 trace_options=stacktrace
4908 See also Documentation/trace/ftrace.rst "trace options"
4912 Have the tracepoints sent to printk as well as the
4913 tracing ring buffer. This is useful for early boot up
4914 where the system hangs or reboots and does not give the
4915 option for reading the tracing buffer or performing a
4916 ftrace_dump_on_oops.
4918 To turn off having tracepoints sent to printk,
4919 echo 0 > /proc/sys/kernel/tracepoint_printk
4920 Note, echoing 1 into this file without the
4921 tracepoint_printk kernel cmdline option has no effect.
4925 Having tracepoints sent to printk() and activating high
4926 frequency tracepoints such as irq or sched, can cause
4927 the system to live lock.
4930 [FTRACE] enable this option to disable tracing when a
4931 warning is hit. This turns off "tracing_on". Tracing can
4932 be enabled again by echoing '1' into the "tracing_on"
4933 file located in /sys/kernel/debug/tracing/
4935 This option is useful, as it disables the trace before
4936 the WARNING dump is called, which prevents the trace to
4937 be filled with content caused by the warning output.
4939 This option can also be set at run time via the sysctl
4940 option: kernel/traceoff_on_warning
4942 transparent_hugepage=
4944 Format: [always|madvise|never]
4945 Can be used to control the default behavior of the system
4946 with respect to transparent hugepages.
4947 See Documentation/admin-guide/mm/transhuge.rst
4950 tsc= Disable clocksource stability checks for TSC.
4952 [x86] reliable: mark tsc clocksource as reliable, this
4953 disables clocksource verification at runtime, as well
4954 as the stability checks done at bootup. Used to enable
4955 high-resolution timer mode on older hardware, and in
4956 virtualized environment.
4957 [x86] noirqtime: Do not use TSC to do irq accounting.
4958 Used to run time disable IRQ_TIME_ACCOUNTING on any
4959 platforms where RDTSC is slow and this accounting
4961 [x86] unstable: mark the TSC clocksource as unstable, this
4962 marks the TSC unconditionally unstable at bootup and
4963 avoids any further wobbles once the TSC watchdog notices.
4964 [x86] nowatchdog: disable clocksource watchdog. Used
4965 in situations with strict latency requirements (where
4966 interruptions from clocksource watchdog are not
4969 tsx= [X86] Control Transactional Synchronization
4970 Extensions (TSX) feature in Intel processors that
4971 support TSX control.
4973 This parameter controls the TSX feature. The options are:
4975 on - Enable TSX on the system. Although there are
4976 mitigations for all known security vulnerabilities,
4977 TSX has been known to be an accelerator for
4978 several previous speculation-related CVEs, and
4979 so there may be unknown security risks associated
4980 with leaving it enabled.
4982 off - Disable TSX on the system. (Note that this
4983 option takes effect only on newer CPUs which are
4984 not vulnerable to MDS, i.e., have
4985 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
4986 the new IA32_TSX_CTRL MSR through a microcode
4987 update. This new MSR allows for the reliable
4988 deactivation of the TSX functionality.)
4990 auto - Disable TSX if X86_BUG_TAA is present,
4991 otherwise enable TSX on the system.
4993 Not specifying this option is equivalent to tsx=off.
4995 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
4998 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
4999 Abort (TAA) vulnerability.
5001 Similar to Micro-architectural Data Sampling (MDS)
5002 certain CPUs that support Transactional
5003 Synchronization Extensions (TSX) are vulnerable to an
5004 exploit against CPU internal buffers which can forward
5005 information to a disclosure gadget under certain
5008 In vulnerable processors, the speculatively forwarded
5009 data can be used in a cache side channel attack, to
5010 access data to which the attacker does not have direct
5013 This parameter controls the TAA mitigation. The
5016 full - Enable TAA mitigation on vulnerable CPUs
5019 full,nosmt - Enable TAA mitigation and disable SMT on
5020 vulnerable CPUs. If TSX is disabled, SMT
5021 is not disabled because CPU is not
5022 vulnerable to cross-thread TAA attacks.
5023 off - Unconditionally disable TAA mitigation
5025 On MDS-affected machines, tsx_async_abort=off can be
5026 prevented by an active MDS mitigation as both vulnerabilities
5027 are mitigated with the same mechanism so in order to disable
5028 this mitigation, you need to specify mds=off too.
5030 Not specifying this option is equivalent to
5031 tsx_async_abort=full. On CPUs which are MDS affected
5032 and deploy MDS mitigation, TAA mitigation is not
5033 required and doesn't provide any additional
5037 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5039 turbografx.map[2|3]= [HW,JOY]
5040 TurboGraFX parallel port interface
5042 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5043 See also Documentation/input/devices/joystick-parport.rst
5045 udbg-immortal [PPC] When debugging early kernel crashes that
5046 happen after console_init() and before a proper
5047 console driver takes over, this boot options might
5048 help "seeing" what's going on.
5050 uhash_entries= [KNL,NET]
5051 Set number of hash buckets for UDP/UDP-Lite connections
5054 [USB] Ignore overcurrent events (default N).
5055 Some badly-designed motherboards generate lots of
5056 bogus events, for ports that aren't wired to
5057 anything. Set this parameter to avoid log spamming.
5058 Note that genuine overcurrent events won't be
5062 [X86] Cause panic on unknown NMI.
5064 usbcore.authorized_default=
5065 [USB] Default USB device authorization:
5066 (default -1 = authorized except for wireless USB,
5067 0 = not authorized, 1 = authorized, 2 = authorized
5068 if device connected to internal port)
5070 usbcore.autosuspend=
5071 [USB] The autosuspend time delay (in seconds) used
5072 for newly-detected USB devices (default 2). This
5073 is the time required before an idle device will be
5074 autosuspended. Devices for which the delay is set
5075 to a negative value won't be autosuspended at all.
5077 usbcore.usbfs_snoop=
5078 [USB] Set to log all usbfs traffic (default 0 = off).
5080 usbcore.usbfs_snoop_max=
5081 [USB] Maximum number of bytes to snoop in each URB
5084 usbcore.blinkenlights=
5085 [USB] Set to cycle leds on hubs (default 0 = off).
5087 usbcore.old_scheme_first=
5088 [USB] Start with the old device initialization
5089 scheme, applies only to low and full-speed devices
5092 usbcore.usbfs_memory_mb=
5093 [USB] Memory limit (in MB) for buffers allocated by
5094 usbfs (default = 16, 0 = max = 2047).
5096 usbcore.use_both_schemes=
5097 [USB] Try the other device initialization scheme
5098 if the first one fails (default 1 = enabled).
5100 usbcore.initial_descriptor_timeout=
5101 [USB] Specifies timeout for the initial 64-byte
5102 USB_REQ_GET_DESCRIPTOR request in milliseconds
5103 (default 5000 = 5.0 seconds).
5105 usbcore.nousb [USB] Disable the USB subsystem
5108 [USB] A list of quirk entries to augment the built-in
5109 usb core quirk list. List entries are separated by
5110 commas. Each entry has the form
5111 VendorID:ProductID:Flags. The IDs are 4-digit hex
5112 numbers and Flags is a set of letters. Each letter
5113 will change the built-in quirk; setting it if it is
5114 clear and clearing it if it is set. The letters have
5115 the following meanings:
5116 a = USB_QUIRK_STRING_FETCH_255 (string
5117 descriptors must not be fetched using
5119 b = USB_QUIRK_RESET_RESUME (device can't resume
5120 correctly so reset it instead);
5121 c = USB_QUIRK_NO_SET_INTF (device can't handle
5122 Set-Interface requests);
5123 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5124 handle its Configuration or Interface
5126 e = USB_QUIRK_RESET (device can't be reset
5127 (e.g morph devices), don't use reset);
5128 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5129 more interface descriptions than the
5130 bNumInterfaces count, and can't handle
5131 talking to these interfaces);
5132 g = USB_QUIRK_DELAY_INIT (device needs a pause
5133 during initialization, after we read
5134 the device descriptor);
5135 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5136 high speed and super speed interrupt
5137 endpoints, the USB 2.0 and USB 3.0 spec
5138 require the interval in microframes (1
5139 microframe = 125 microseconds) to be
5140 calculated as interval = 2 ^
5142 Devices with this quirk report their
5143 bInterval as the result of this
5144 calculation instead of the exponent
5145 variable used in the calculation);
5146 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5147 handle device_qualifier descriptor
5149 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5150 generates spurious wakeup, ignore
5151 remote wakeup capability);
5152 k = USB_QUIRK_NO_LPM (device can't handle Link
5154 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5155 (Device reports its bInterval as linear
5156 frames instead of the USB 2.0
5158 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5159 to be disconnected before suspend to
5160 prevent spurious wakeup);
5161 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5162 pause after every control message);
5163 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5164 delay after resetting its port);
5165 Example: quirks=0781:5580:bk,0a5c:5834:gij
5168 [USBHID] The interval which mice are to be polled at.
5171 [USBHID] The interval which joysticks are to be polled at.
5174 [USBHID] The interval which keyboards are to be polled at.
5176 usb-storage.delay_use=
5177 [UMS] The delay in seconds before a new device is
5178 scanned for Logical Units (default 1).
5181 [UMS] A list of quirks entries to supplement or
5182 override the built-in unusual_devs list. List
5183 entries are separated by commas. Each entry has
5184 the form VID:PID:Flags where VID and PID are Vendor
5185 and Product ID values (4-digit hex numbers) and
5186 Flags is a set of characters, each corresponding
5187 to a common usb-storage quirk flag as follows:
5188 a = SANE_SENSE (collect more than 18 bytes
5189 of sense data, not on uas);
5190 b = BAD_SENSE (don't collect more than 18
5191 bytes of sense data, not on uas);
5192 c = FIX_CAPACITY (decrease the reported
5193 device capacity by one sector);
5194 d = NO_READ_DISC_INFO (don't use
5195 READ_DISC_INFO command, not on uas);
5196 e = NO_READ_CAPACITY_16 (don't use
5197 READ_CAPACITY_16 command);
5198 f = NO_REPORT_OPCODES (don't use report opcodes
5200 g = MAX_SECTORS_240 (don't transfer more than
5201 240 sectors at a time, uas only);
5202 h = CAPACITY_HEURISTICS (decrease the
5203 reported device capacity by one
5204 sector if the number is odd);
5205 i = IGNORE_DEVICE (don't bind to this
5207 j = NO_REPORT_LUNS (don't use report luns
5209 l = NOT_LOCKABLE (don't try to lock and
5210 unlock ejectable media, not on uas);
5211 m = MAX_SECTORS_64 (don't transfer more
5212 than 64 sectors = 32 KB at a time,
5214 n = INITIAL_READ10 (force a retry of the
5215 initial READ(10) command, not on uas);
5216 o = CAPACITY_OK (accept the capacity
5217 reported by the device, not on uas);
5218 p = WRITE_CACHE (the device cache is ON
5219 by default, not on uas);
5220 r = IGNORE_RESIDUE (the device reports
5221 bogus residue values, not on uas);
5222 s = SINGLE_LUN (the device has only one
5224 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5225 commands, uas only);
5226 u = IGNORE_UAS (don't bind to the uas driver);
5227 w = NO_WP_DETECT (don't test whether the
5228 medium is write-protected).
5229 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5230 even if the device claims no cache,
5232 Example: quirks=0419:aaf5:rl,0421:0433:rc
5234 user_debug= [KNL,ARM]
5236 See arch/arm/Kconfig.debug help text.
5237 1 - undefined instruction events
5239 4 - invalid data aborts
5242 Example: user_debug=31
5245 [X86] Flags controlling user PTE allocations.
5247 nohigh = do not allocate PTE pages in
5248 HIGHMEM regardless of setting
5252 On X86_32, this is an alias for vdso32=. Otherwise:
5254 vdso=1: enable VDSO (the default)
5255 vdso=0: disable VDSO mapping
5257 vdso32= [X86] Control the 32-bit vDSO
5258 vdso32=1: enable 32-bit VDSO
5259 vdso32=0 or vdso32=2: disable 32-bit VDSO
5261 See the help text for CONFIG_COMPAT_VDSO for more
5262 details. If CONFIG_COMPAT_VDSO is set, the default is
5263 vdso32=0; otherwise, the default is vdso32=1.
5265 For compatibility with older kernels, vdso32=2 is an
5268 Try vdso32=0 if you encounter an error that says:
5269 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5272 vector=percpu: enable percpu vector domain
5274 video= [FB] Frame buffer configuration
5275 See Documentation/fb/modedb.rst.
5277 video.brightness_switch_enabled= [0,1]
5278 If set to 1, on receiving an ACPI notify event
5279 generated by hotkey, video driver will adjust brightness
5280 level and then send out the event to user space through
5281 the allocated input device; If set to 0, video driver
5282 will only send out the event without touching backlight
5287 [VMMIO] Memory mapped virtio (platform) device.
5289 <size>@<baseaddr>:<irq>[:<id>]
5291 <size> := size (can use standard suffixes
5293 <baseaddr> := physical base address
5294 <irq> := interrupt number (as passed to
5296 <id> := (optional) platform device id
5298 virtio_mmio.device=1K@0x100b0000:48:7
5300 Can be used multiple times for multiple devices.
5302 vga= [BOOT,X86-32] Select a particular video mode
5303 See Documentation/x86/boot.rst and
5304 Documentation/admin-guide/svga.rst.
5305 Use vga=ask for menu.
5306 This is actually a boot loader parameter; the value is
5307 passed to the kernel using a special protocol.
5309 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
5310 May slow down system boot speed, especially when
5311 enabled on systems with a large amount of memory.
5312 All options are enabled by default, and this
5313 interface is meant to allow for selectively
5314 enabling or disabling specific virtual memory
5317 Available options are:
5318 P Enable page structure init time poisoning
5319 - Disable all of the above options
5321 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5322 size of <nn>. This can be used to increase the
5323 minimum size (128MB on x86). It can also be used to
5324 decrease the size and leave more room for directly
5327 vmcp_cma=nn[MG] [KNL,S390]
5328 Sets the memory size reserved for contiguous memory
5329 allocations for the vmcp device driver.
5331 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5334 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5337 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5341 Controls the behavior of vsyscalls (i.e. calls to
5342 fixed addresses of 0xffffffffff600x00 from legacy
5343 code). Most statically-linked binaries and older
5344 versions of glibc use these calls. Because these
5345 functions are at fixed addresses, they make nice
5346 targets for exploits that can control RIP.
5348 emulate [default] Vsyscalls turn into traps and are
5349 emulated reasonably safely. The vsyscall
5352 xonly Vsyscalls turn into traps and are
5353 emulated reasonably safely. The vsyscall
5354 page is not readable.
5356 none Vsyscalls don't work at all. This makes
5357 them quite hard to use for exploits but
5358 might break your system.
5360 vt.color= [VT] Default text color.
5361 Format: 0xYX, X = foreground, Y = background.
5362 Default: 0x07 = light gray on black.
5364 vt.cur_default= [VT] Default cursor shape.
5365 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5366 the parameters of the <Esc>[?A;B;Cc escape sequence;
5367 see VGA-softcursor.txt. Default: 2 = underline.
5369 vt.default_blu= [VT]
5370 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5371 Change the default blue palette of the console.
5372 This is a 16-member array composed of values
5375 vt.default_grn= [VT]
5376 Format: <green0>,<green1>,<green2>,...,<green15>
5377 Change the default green palette of the console.
5378 This is a 16-member array composed of values
5381 vt.default_red= [VT]
5382 Format: <red0>,<red1>,<red2>,...,<red15>
5383 Change the default red palette of the console.
5384 This is a 16-member array composed of values
5390 Set system-wide default UTF-8 mode for all tty's.
5391 Default is 1, i.e. UTF-8 mode is enabled for all
5392 newly opened terminals.
5394 vt.global_cursor_default=
5397 Set system-wide default for whether a cursor
5398 is shown on new VTs. Default is -1,
5399 i.e. cursors will be created by default unless
5400 overridden by individual drivers. 0 will hide
5401 cursors, 1 will display them.
5403 vt.italic= [VT] Default color for italic text; 0-15.
5406 vt.underline= [VT] Default color for underlined text; 0-15.
5409 watchdog timers [HW,WDT] For information on watchdog timers,
5410 see Documentation/watchdog/watchdog-parameters.rst
5411 or other driver-specific files in the
5412 Documentation/watchdog/ directory.
5416 Set the hard lockup detector stall duration
5417 threshold in seconds. The soft lockup detector
5418 threshold is set to twice the value. A value of 0
5419 disables both lockup detectors. Default is 10
5422 workqueue.watchdog_thresh=
5423 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5424 warn stall conditions and dump internal state to
5425 help debugging. 0 disables workqueue stall
5426 detection; otherwise, it's the stall threshold
5427 duration in seconds. The default value is 30 and
5428 it can be updated at runtime by writing to the
5429 corresponding sysfs file.
5431 workqueue.disable_numa
5432 By default, all work items queued to unbound
5433 workqueues are affine to the NUMA nodes they're
5434 issued on, which results in better behavior in
5435 general. If NUMA affinity needs to be disabled for
5436 whatever reason, this option can be used. Note
5437 that this also can be controlled per-workqueue for
5438 workqueues visible under /sys/bus/workqueue/.
5440 workqueue.power_efficient
5441 Per-cpu workqueues are generally preferred because
5442 they show better performance thanks to cache
5443 locality; unfortunately, per-cpu workqueues tend to
5444 be more power hungry than unbound workqueues.
5446 Enabling this makes the per-cpu workqueues which
5447 were observed to contribute significantly to power
5448 consumption unbound, leading to measurably lower
5449 power usage at the cost of small performance
5452 The default value of this parameter is determined by
5453 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5455 workqueue.debug_force_rr_cpu
5456 Workqueue used to implicitly guarantee that work
5457 items queued without explicit CPU specified are put
5458 on the local CPU. This guarantee is no longer true
5459 and while local CPU is still preferred work items
5460 may be put on foreign CPUs. This debug option
5461 forces round-robin CPU selection to flush out
5462 usages which depend on the now broken guarantee.
5463 When enabled, memory and cache locality will be
5466 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5467 default x2apic cluster mode on platforms
5470 x86_intel_mid_timer= [X86-32,APBT]
5471 Choose timer option for x86 Intel MID platform.
5472 Two valid options are apbt timer only and lapic timer
5473 plus one apbt timer for broadcast timer.
5474 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5476 xen_512gb_limit [KNL,X86-64,XEN]
5477 Restricts the kernel running paravirtualized under Xen
5478 to use only up to 512 GB of RAM. The reason to do so is
5479 crash analysis tools and Xen tools for doing domain
5480 save/restore/migration must be enabled to handle larger
5483 xen_emul_unplug= [HW,X86,XEN]
5484 Unplug Xen emulated devices
5485 Format: [unplug0,][unplug1]
5486 ide-disks -- unplug primary master IDE devices
5487 aux-ide-disks -- unplug non-primary-master IDE devices
5488 nics -- unplug network devices
5489 all -- unplug all emulated devices (NICs and IDE disks)
5490 unnecessary -- unplugging emulated devices is
5491 unnecessary even if the host did not respond to
5493 never -- do not unplug even if version check succeeds
5495 xen_legacy_crash [X86,XEN]
5496 Crash from Xen panic notifier, without executing late
5497 panic() code such as dumping handler.
5499 xen_nopvspin [X86,XEN]
5500 Disables the ticketlock slowpath using Xen PV
5504 Disables the PV optimizations forcing the HVM guest to
5505 run as generic HVM guest with no PV drivers.
5506 This option is obsoleted by the "nopv" option, which
5507 has equivalent effect for XEN platform.
5509 xen_scrub_pages= [XEN]
5510 Boolean option to control scrubbing pages before giving them back
5511 to Xen, for use by other domains. Can be also changed at runtime
5512 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5513 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5515 xen_timer_slop= [X86-64,XEN]
5516 Set the timer slop (in nanoseconds) for the virtual Xen
5517 timers (default is 100000). This adjusts the minimum
5518 delta of virtualized Xen timers, where lower values
5519 improve timer resolution at the expense of processing
5520 more timer interrupts.
5522 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
5523 Disables the PV optimizations forcing the guest to run
5524 as generic guest with no PV drivers. Currently support
5525 XEN HVM, KVM, HYPER_V and VMWARE guest.
5527 xirc2ps_cs= [NET,PCMCIA]
5529 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5532 By default on POWER9 and above, the kernel will
5533 natively use the XIVE interrupt controller. This option
5534 allows the fallback firmware mode to be used:
5536 off Fallback to firmware control of XIVE interrupt
5537 controller on both pseries and powernv
5538 platforms. Only useful on POWER9 and above.
5540 xhci-hcd.quirks [USB,KNL]
5541 A hex value specifying bitmask with supplemental xhci
5542 host controller quirks. Meaning of each bit can be
5543 consulted in header drivers/usb/host/xhci.h.
5546 Format: { early | on | rw | ro | off }
5547 Controls if xmon debugger is enabled. Default is off.
5548 Passing only "xmon" is equivalent to "xmon=early".
5549 early Call xmon as early as possible on boot; xmon
5550 debugger is called from setup_arch().
5551 on xmon debugger hooks will be installed so xmon
5552 is only called on a kernel crash. Default mode,
5553 i.e. either "ro" or "rw" mode, is controlled
5554 with CONFIG_XMON_DEFAULT_RO_MODE.
5555 rw xmon debugger hooks will be installed so xmon
5556 is called only on a kernel crash, mode is write,
5557 meaning SPR registers, memory and, other data
5558 can be written using xmon commands.
5559 ro same as "rw" option above but SPR registers,
5560 memory, and other data can't be written using
5562 off xmon is disabled.