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 Disable BERT OS support on buggy BIOSes.
443 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
444 bttv.radio= Most important insmod options are available as
446 bttv.pll= See Documentation/media/v4l-drivers/bttv.rst
449 bulk_remove=off [PPC] This parameter disables the use of the pSeries
450 firmware feature for flushing multiple hpte entries
453 c101= [NET] Moxa C101 synchronous serial card
455 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
456 Sometimes CPU hardware bugs make them report the cache
457 size incorrectly. The kernel will attempt work arounds
458 to fix known problems, but for some CPUs it is not
459 possible to determine what the correct size should be.
460 This option provides an override for these situations.
463 [NET] Specifies amount of time (in seconds) that
464 the kernel should wait for a network carrier. By default
465 it waits 120 seconds.
467 ca_keys= [KEYS] This parameter identifies a specific key(s) on
468 the system trusted keyring to be used for certificate
470 format: { id:<keyid> | builtin }
472 cca= [MIPS] Override the kernel pages' cache coherency
473 algorithm. Accepted values range from 0 to 7
474 inclusive. See arch/mips/include/asm/pgtable-bits.h
475 for platform specific values (SB1, Loongson3 and
478 ccw_timeout_log [S390]
479 See Documentation/s390/common_io.rst for details.
481 cgroup_disable= [KNL] Disable a particular controller
482 Format: {name of the controller(s) to disable}
483 The effects of cgroup_disable=foo are:
484 - foo isn't auto-mounted if you mount all cgroups in
486 - foo isn't visible as an individually mountable
488 {Currently only "memory" controller deal with this and
489 cut the overhead, others just disable the usage. So
490 only cgroup_disable=memory is actually worthy}
492 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
493 Format: { { controller | "all" | "named" }
494 [,{ controller | "all" | "named" }...] }
495 Like cgroup_disable, but only applies to cgroup v1;
496 the blacklisted controllers remain available in cgroup2.
497 "all" blacklists all controllers and "named" disables
498 named mounts. Specifying both "all" and "named" disables
501 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
503 nosocket -- Disable socket memory accounting.
504 nokmem -- Disable kernel memory accounting.
506 checkreqprot [SELINUX] Set initial checkreqprot flag value.
507 Format: { "0" | "1" }
508 See security/selinux/Kconfig help text.
509 0 -- check protection applied by kernel (includes
510 any implied execute protection).
511 1 -- check protection requested by application.
512 Default value is set via a kernel config option.
513 Value can be changed at runtime via
514 /sys/fs/selinux/checkreqprot.
517 See Documentation/s390/common_io.rst for details.
520 Prevents the clock framework from automatically gating
521 clocks that have not been explicitly enabled by a Linux
522 device driver but are enabled in hardware at reset or
523 by the bootloader/firmware. Note that this does not
524 force such clocks to be always-on nor does it reserve
525 those clocks in any way. This parameter is useful for
526 debug and development, but should not be needed on a
527 platform with proper driver support. For more
528 information, see Documentation/driver-api/clk.rst.
530 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
532 Forces specified clocksource (if available) to be used
533 when calculating gettimeofday(). If specified
534 clocksource is not available, it defaults to PIT.
535 Format: { pit | tsc | cyclone | pmtmr }
537 clocksource= Override the default clocksource
539 Override the default clocksource and use the clocksource
540 with the name specified.
541 Some clocksource names to choose from, depending on
543 [all] jiffies (this is the base, fallback clocksource)
545 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
546 pxa_timer,timer3,32k_counter,timer0_1
547 [X86-32] pit,hpet,tsc;
548 scx200_hrt on Geode; cyclone on IBM x440
556 clocksource.arm_arch_timer.evtstrm=
559 Enable/disable the eventstream feature of the ARM
560 architected timer so that code using WFE-based polling
561 loops can be debugged more effectively on production
564 clearcpuid=BITNUM [X86]
565 Disable CPUID feature X for the kernel. See
566 arch/x86/include/asm/cpufeatures.h for the valid bit
567 numbers. Note the Linux specific bits are not necessarily
568 stable over kernel options, but the vendor specific
570 Also note that user programs calling CPUID directly
571 or using the feature without checking anything
572 will still see it. This just prevents it from
573 being used by the kernel or shown in /proc/cpuinfo.
574 Also note the kernel might malfunction if you disable
577 cma=nn[MG]@[start[MG][-end[MG]]]
579 Sets the size of kernel global memory area for
580 contiguous memory allocations and optionally the
581 placement constraint by the physical address range of
582 memory allocations. A value of 0 disables CMA
583 altogether. For more information, see
584 include/linux/dma-contiguous.h
586 cmo_free_hint= [PPC] Format: { yes | no }
587 Specify whether pages are marked as being inactive
588 when they are freed. This is used in CMO environments
589 to determine OS memory pressure for page stealing by
593 coherent_pool=nn[KMG] [ARM,KNL]
594 Sets the size of memory pool for coherent, atomic dma
595 allocations, by default set to 256K.
597 com20020= [HW,NET] ARCnet - COM20020 chipset
599 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
601 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
605 ARCnet - COM90xx chipset (memory-mapped buffers)
606 Format: <io>[,<irq>[,<memstart>]]
608 condev= [HW,S390] console device
611 console= [KNL] Output console device and options.
613 tty<n> Use the virtual console device <n>.
617 Use the specified serial port. The options are of
618 the form "bbbbpnf", where "bbbb" is the baud rate,
619 "p" is parity ("n", "o", or "e"), "n" is number of
620 bits, and "f" is flow control ("r" for RTS or
621 omit it). Default is "9600n8".
623 See Documentation/admin-guide/serial-console.rst for more
625 Documentation/networking/netconsole.txt for an
628 uart[8250],io,<addr>[,options]
629 uart[8250],mmio,<addr>[,options]
630 uart[8250],mmio16,<addr>[,options]
631 uart[8250],mmio32,<addr>[,options]
632 uart[8250],0x<addr>[,options]
633 Start an early, polled-mode console on the 8250/16550
634 UART at the specified I/O port or MMIO address,
635 switching to the matching ttyS device later.
636 MMIO inter-register address stride is either 8-bit
637 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
638 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
639 to be equivalent to 'mmio'. 'options' are specified in
640 the same format described for ttyS above; if unspecified,
641 the h/w is not re-initialized.
643 hvc<n> Use the hypervisor console device <n>. This is for
644 both Xen and PowerPC hypervisors.
646 If the device connected to the port is not a TTY but a braille
647 device, prepend "brl," before the device type, for instance
649 For now, only VisioBraille is supported.
652 [KNL] Change console messages format
654 By default we print messages on consoles in
655 "[time stamp] text\n" format (time stamp may not be
656 printed, depending on CONFIG_PRINTK_TIME or
657 `printk_time' param).
659 Switch to syslog format: "<%u>[time stamp] text\n"
660 IOW, each message will have a facility and loglevel
661 prefix. The format is similar to one used by syslog()
662 syscall, or to executing "dmesg -S --raw" or to reading
665 consoleblank= [KNL] The console blank (screen saver) timeout in
666 seconds. A value of 0 disables the blank timer.
670 [KNL] Change the default value for
671 /proc/<pid>/coredump_filter.
672 See also Documentation/filesystems/proc.txt.
674 coresight_cpu_debug.enable
677 Enable/disable the CPU sampling based debugging.
678 0: default value, disable debugging
679 1: enable debugging at boot time
681 cpuidle.off=1 [CPU_IDLE]
682 disable the cpuidle sub-system
685 [CPU_IDLE] Name of the cpuidle governor to use.
687 cpufreq.off=1 [CPU_FREQ]
688 disable the cpufreq sub-system
691 [X86] Delay for N microsec between assert and de-assert
692 of APIC INIT to start processors. This delay occurs
693 on every CPU online, such as boot, and resume from suspend.
696 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
698 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
700 crashkernel=size[KMG][@offset[KMG]]
701 [KNL] Using kexec, Linux can switch to a 'crash kernel'
702 upon panic. This parameter reserves the physical
703 memory region [offset, offset + size] for that kernel
704 image. If '@offset' is omitted, then a suitable offset
705 is selected automatically.
706 [KNL, x86_64] select a region under 4G first, and
707 fall back to reserve region above 4G when '@offset'
708 hasn't been specified.
709 See Documentation/admin-guide/kdump/kdump.rst for further details.
711 crashkernel=range1:size1[,range2:size2,...][@offset]
712 [KNL] Same as above, but depends on the memory
713 in the running system. The syntax of range is
714 start-[end] where start and end are both
715 a memory unit (amount[KMG]). See also
716 Documentation/admin-guide/kdump/kdump.rst for an example.
718 crashkernel=size[KMG],high
719 [KNL, x86_64] range could be above 4G. Allow kernel
720 to allocate physical memory region from top, so could
721 be above 4G if system have more than 4G ram installed.
722 Otherwise memory region will be allocated below 4G, if
724 It will be ignored if crashkernel=X is specified.
725 crashkernel=size[KMG],low
726 [KNL, x86_64] range under 4G. When crashkernel=X,high
727 is passed, kernel could allocate physical memory region
728 above 4G, that cause second kernel crash on system
729 that require some amount of low memory, e.g. swiotlb
730 requires at least 64M+32K low memory, also enough extra
731 low memory is needed to make sure DMA buffers for 32-bit
732 devices won't run out. Kernel would try to allocate at
733 at least 256M below 4G automatically.
734 This one let user to specify own low range under 4G
735 for second kernel instead.
736 0: to disable low allocation.
737 It will be ignored when crashkernel=X,high is not used
738 or memory reserved is below 4G.
741 [KNL] Disable crypto self-tests
746 cs89x0_media= [HW,NET]
747 Format: { rj45 | aui | bnc }
750 See header of drivers/s390/block/dasd_devmap.c.
752 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
753 (one device per port)
754 Format: <port#>,<type>
755 See also Documentation/input/devices/joystick-parport.rst
757 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
759 Documentation/admin-guide/dynamic-debug-howto.rst for
760 details. Deprecated, see dyndbg.
762 debug [KNL] Enable kernel debugging (events log level).
765 [KNL] Enable printing [hashed] pointers early in the
766 boot sequence. If enabled, we use a weak hash instead
767 of siphash to hash pointers. Use this option if you are
768 seeing instances of '(___ptrval___)') and need to see a
769 value (hashed pointer) instead. Cryptographically
770 insecure, please do not use on production kernels.
773 [KNL] verbose self-tests
775 Print debugging info while doing the locking API
777 We default to 0 (no extra messages), setting it to
778 1 will print _a lot_ more information - normally
779 only useful to kernel developers.
781 debug_objects [KNL] Enable object debugging
784 [KNL] Disable object debugging
786 debug_guardpage_minorder=
787 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
788 parameter allows control of the order of pages that will
789 be intentionally kept free (and hence protected) by the
790 buddy allocator. Bigger value increase the probability
791 of catching random memory corruption, but reduce the
792 amount of memory for normal system use. The maximum
793 possible value is MAX_ORDER/2. Setting this parameter
794 to 1 or 2 should be enough to identify most random
795 memory corruption problems caused by bugs in kernel or
796 driver code when a CPU writes to (or reads from) a
797 random memory location. Note that there exists a class
798 of memory corruptions problems caused by buggy H/W or
799 F/W or by drivers badly programing DMA (basically when
800 memory is written at bus level and the CPU MMU is
801 bypassed) which are not detectable by
802 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
803 tracking down these problems.
806 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
807 enables the feature at boot time. By default, it is
808 disabled and the system will work mostly the same as a
809 kernel built without CONFIG_DEBUG_PAGEALLOC.
810 Note: to get most of debug_pagealloc error reports, it's
811 useful to also enable the page_owner functionality.
812 on: enable the feature
814 debugpat [X86] Enable PAT debugging
816 decnet.addr= [HW,NET]
817 Format: <area>[,<node>]
818 See also Documentation/networking/decnet.txt.
821 [same as hugepagesz=] The size of the default
822 HugeTLB page size. This is the size represented by
823 the legacy /proc/ hugepages APIs, used for SHM, and
824 default size when mounting hugetlbfs filesystems.
825 Defaults to the default architecture's huge page size
828 deferred_probe_timeout=
829 [KNL] Debugging option to set a timeout in seconds for
830 deferred probe to give up waiting on dependencies to
831 probe. Only specific dependencies (subsystems or
832 drivers) that have opted in will be ignored. A timeout of 0
833 will timeout at the end of initcalls. This option will also
834 dump out devices still on the deferred probe list after
838 Set number of hash buckets for dentry cache.
840 disable_1tb_segments [PPC]
841 Disables the use of 1TB hash page table segments. This
842 causes the kernel to fall back to 256MB segments which
843 can be useful when debugging issues that require an SLB
847 See Documentation/networking/ipv6.txt.
850 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
851 hardening is enabled for this boot. Hardened
852 usercopy checking is used to protect the kernel
853 from reading or writing beyond known memory
854 allocation boundaries as a proactive defense
855 against bounds-checking flaws in the kernel's
856 copy_to_user()/copy_from_user() interface.
857 on Perform hardened usercopy checks (default).
858 off Disable hardened usercopy checks.
861 Disable RADIX MMU mode on POWER9
864 Disable TLBIE instruction. Currently does not work
865 with KVM, with HASH MMU, or with coherent accelerators.
867 disable_cpu_apicid= [X86,APIC,SMP]
869 The number of initial APIC ID for the
870 corresponding CPU to be disabled at boot,
871 mostly used for the kdump 2nd kernel to
872 disable BSP to wake up multiple CPUs without
873 causing system reset or hang due to sending
876 perf_v4_pmi= [X86,INTEL]
878 Disable Intel PMU counter freezing feature.
879 The feature only exists starting from
880 Arch Perfmon v4 (Skylake and newer).
882 disable_ddw [PPC/PSERIES]
883 Disable Dynamic DMA Window support. Use this if
884 to workaround buggy firmware.
887 See Documentation/networking/ipv6.txt.
889 disable_mtrr_cleanup [X86]
890 The kernel tries to adjust MTRR layout from continuous
891 to discrete, to make X server driver able to add WB
892 entry later. This parameter disables that.
894 disable_mtrr_trim [X86, Intel and AMD only]
895 By default the kernel will trim any uncacheable
896 memory out of your available memory pool based on
897 MTRR settings. This parameter disables that behavior,
898 possibly causing your machine to run very slowly.
900 disable_timer_pin_1 [X86]
901 Disable PIN 1 of APIC timer
902 Can be useful to work around chipset bugs.
904 dis_ucode_ldr [X86] Disable the microcode loader.
906 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
907 this option disables the debugging code at boot.
909 dma_debug_entries=<number>
910 This option allows to tune the number of preallocated
911 entries for DMA-API debugging code. One entry is
912 required per DMA-API allocation. Use this if the
913 DMA-API debugging code disables itself because the
914 architectural default is too low.
916 dma_debug_driver=<driver_name>
917 With this option the DMA-API debugging driver
918 filter feature can be enabled at boot time. Just
919 pass the driver to filter for as the parameter.
920 The filter can be disabled or changed to another
921 driver later using sysfs.
923 driver_async_probe= [KNL]
924 List of driver names to be probed asynchronously.
925 Format: <driver_name1>,<driver_name2>...
927 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
928 Broken monitors, graphic adapters, KVMs and EDIDless
929 panels may send no or incorrect EDID data sets.
930 This parameter allows to specify an EDID data sets
931 in the /lib/firmware directory that are used instead.
932 Generic built-in EDID data sets are used, if one of
933 edid/1024x768.bin, edid/1280x1024.bin,
934 edid/1680x1050.bin, or edid/1920x1080.bin is given
935 and no file with the same name exists. Details and
936 instructions how to build your own EDID data are
937 available in Documentation/driver-api/edid.rst. An EDID
938 data set will only be used for a particular connector,
939 if its name and a colon are prepended to the EDID
940 name. Each connector may use a unique EDID data
941 set by separating the files with a comma. An EDID
942 data set with no connector name will be used for
943 any connectors not explicitly specified.
948 Format: {"off" | "known"}
949 Control how the dt_cpu_ftrs device-tree binding is
950 used for CPU feature discovery and setup (if it
952 off: Do not use it, fall back to legacy cpu table.
953 known: Do not pass through unknown features to guests
954 or userspace, only those that the kernel is aware of.
956 dump_apple_properties [X86]
957 Dump name and content of EFI device properties on
958 x86 Macs. Useful for driver authors to determine
959 what data is available or for reverse-engineering.
961 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
962 module.dyndbg[="val"]
963 Enable debug messages at boot time. See
964 Documentation/admin-guide/dynamic-debug-howto.rst
967 nompx [X86] Disables Intel Memory Protection Extensions.
968 See Documentation/x86/intel_mpx.rst for more
969 information about the feature.
971 nopku [X86] Disable Memory Protection Keys CPU feature found
974 module.async_probe [KNL]
975 Enable asynchronous probe on this module.
977 early_ioremap_debug [KNL]
978 Enable debug messages in early_ioremap support. This
979 is useful for tracking down temporary early mappings
980 which are not unmapped.
982 earlycon= [KNL] Output early console device and options.
984 When used with no options, the early console is
985 determined by stdout-path property in device tree's
986 chosen node or the ACPI SPCR table if supported by
989 cdns,<addr>[,options]
990 Start an early, polled-mode console on a Cadence
991 (xuartps) serial port at the specified address. Only
992 supported option is baud rate. If baud rate is not
993 specified, the serial port must already be setup and
996 uart[8250],io,<addr>[,options]
997 uart[8250],mmio,<addr>[,options]
998 uart[8250],mmio32,<addr>[,options]
999 uart[8250],mmio32be,<addr>[,options]
1000 uart[8250],0x<addr>[,options]
1001 Start an early, polled-mode console on the 8250/16550
1002 UART at the specified I/O port or MMIO address.
1003 MMIO inter-register address stride is either 8-bit
1004 (mmio) or 32-bit (mmio32 or mmio32be).
1005 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1006 to be equivalent to 'mmio'. 'options' are specified
1007 in the same format described for "console=ttyS<n>"; if
1008 unspecified, the h/w is not initialized.
1012 Start an early, polled-mode console on a pl011 serial
1013 port at the specified address. The pl011 serial port
1014 must already be setup and configured. Options are not
1015 yet supported. If 'mmio32' is specified, then only
1016 the driver will use only 32-bit accessors to read/write
1017 the device registers.
1020 Start an early, polled-mode console on a meson serial
1021 port at the specified address. The serial port must
1022 already be setup and configured. Options are not yet
1026 Start an early, polled-mode console on an msm serial
1027 port at the specified address. The serial port
1028 must already be setup and configured. Options are not
1031 msm_serial_dm,<addr>
1032 Start an early, polled-mode console on an msm serial
1033 dm port at the specified address. The serial port
1034 must already be setup and configured. Options are not
1038 Start an early, polled-mode console on a serial port
1039 of an Actions Semi SoC, such as S500 or S900, at the
1040 specified address. The serial port must already be
1041 setup and configured. Options are not yet supported.
1044 Start an early, polled-mode console on a serial port
1045 of an RDA Micro SoC, such as RDA8810PL, at the
1046 specified address. The serial port must already be
1047 setup and configured. Options are not yet supported.
1050 Use RISC-V SBI (Supervisor Binary Interface) for early
1053 smh Use ARM semihosting calls for early console.
1061 Use early console provided by serial driver available
1062 on Samsung SoCs, requires selecting proper type and
1063 a correct base address of the selected UART port. The
1064 serial port must already be setup and configured.
1065 Options are not yet supported.
1068 Start an early, polled-mode console on a lantiq serial
1069 (lqasc) port at the specified address. The serial port
1070 must already be setup and configured. Options are not
1075 Use early console provided by Freescale LP UART driver
1076 found on Freescale Vybrid and QorIQ LS1021A processors.
1077 A valid base address must be provided, and the serial
1078 port must already be setup and configured.
1081 Start an early, polled-mode console on the
1082 Armada 3700 serial port at the specified
1083 address. The serial port must already be setup
1084 and configured. Options are not yet supported.
1087 Start an early, polled-mode console on a Qualcomm
1088 Generic Interface (GENI) based serial port at the
1089 specified address. The serial port must already be
1090 setup and configured. Options are not yet supported.
1093 Start an early, unaccelerated console on the EFI
1094 memory mapped framebuffer (if available). On cache
1095 coherent non-x86 systems that use system memory for
1096 the framebuffer, pass the 'ram' option so that it is
1097 mapped with the correct attributes.
1100 Use early console provided by Freescale LINFlexD UART
1101 serial driver for NXP S32V234 SoCs. A valid base
1102 address must be provided, and the serial port must
1103 already be setup and configured.
1105 earlyprintk= [X86,SH,ARM,M68k,S390]
1109 earlyprintk=serial[,ttySn[,baudrate]]
1110 earlyprintk=serial[,0x...[,baudrate]]
1111 earlyprintk=ttySn[,baudrate]
1112 earlyprintk=dbgp[debugController#]
1113 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1114 earlyprintk=xdbc[xhciController#]
1116 earlyprintk is useful when the kernel crashes before
1117 the normal console is initialized. It is not enabled by
1118 default because it has some cosmetic problems.
1120 Append ",keep" to not disable it when the real console
1123 Only one of vga, efi, serial, or usb debug port can
1126 Currently only ttyS0 and ttyS1 may be specified by
1127 name. Other I/O ports may be explicitly specified
1128 on some architectures (x86 and arm at least) by
1129 replacing ttySn with an I/O port address, like this:
1130 earlyprintk=serial,0x1008,115200
1131 You can find the port for a given device in
1132 /proc/tty/driver/serial:
1133 2: uart:ST16650V2 port:00001008 irq:18 ...
1135 Interaction with the standard serial driver is not
1138 The VGA and EFI output is eventually overwritten by
1141 The xen output can only be used by Xen PV guests.
1143 The sclp output can only be used on s390.
1145 The optional "force" to "pciserial" enables use of a
1146 PCI device even when its classcode is not of the
1149 edac_report= [HW,EDAC] Control how to report EDAC event
1150 Format: {"on" | "off" | "force"}
1151 on: enable EDAC to report H/W event. May be overridden
1152 by other higher priority error reporting module.
1153 off: disable H/W event reporting through EDAC.
1154 force: enforce the use of EDAC to report H/W event.
1157 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1160 This is designed to be used in conjunction with
1161 the boot argument: earlyprintk=vga
1164 Format: {"off" | "on" | "skip[mbr]"}
1167 Format: { "old_map", "nochunk", "noruntime", "debug",
1168 "nosoftreserve", "disable_early_pci_dma",
1169 "no_disable_early_pci_dma" }
1170 old_map [X86-64]: switch to the old ioremap-based EFI
1171 runtime services mapping. [Needs CONFIG_X86_UV=y]
1172 nochunk: disable reading files in "chunks" in the EFI
1173 boot stub, as chunking can cause problems with some
1174 firmware implementations.
1175 noruntime : disable EFI runtime services support
1176 debug: enable misc debug output
1177 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1178 attribute may cause the kernel to reserve the
1179 memory range for a memory mapping driver to
1180 claim. Specify efi=nosoftreserve to disable this
1181 reservation and treat the memory by its base type
1182 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1183 disable_early_pci_dma: Disable the busmaster bit on all
1184 PCI bridges while in the EFI boot stub
1185 no_disable_early_pci_dma: Leave the busmaster bit set
1186 on all PCI bridges while in the EFI boot stub
1188 efi_no_storage_paranoia [EFI; X86]
1189 Using this parameter you can use more than 50% of
1190 your efi variable storage. Use this parameter only if
1191 you are really sure that your UEFI does sane gc and
1192 fulfills the spec otherwise your board may brick.
1194 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1195 Add arbitrary attribute to specific memory range by
1196 updating original EFI memory map.
1197 Region of memory which aa attribute is added to is
1200 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1201 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1202 attribute is added to range 0x100000000-0x180000000 and
1203 0x10a0000000-0x1120000000.
1205 If efi_fake_mem=8G@9G:0x40000 is specified, the
1206 EFI_MEMORY_SP(0x40000) attribute is added to
1207 range 0x240000000-0x43fffffff.
1209 Using this parameter you can do debugging of EFI memmap
1210 related features. For example, you can do debugging of
1211 Address Range Mirroring feature even if your box
1212 doesn't support it, or mark specific memory as
1215 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1216 that is to be dynamically loaded by Linux. If there are
1217 multiple variables with the same name but with different
1218 vendor GUIDs, all of them will be loaded. See
1219 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1222 eisa_irq_edge= [PARISC,HW]
1223 See header of drivers/parisc/eisa.c.
1226 See comment before function elanfreq_setup() in
1227 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1229 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1230 Specifies physical address of start of kernel core
1231 image elf header and optionally the size. Generally
1232 kexec loader will pass this option to capture kernel.
1233 See Documentation/admin-guide/kdump/kdump.rst for details.
1235 enable_mtrr_cleanup [X86]
1236 The kernel tries to adjust MTRR layout from continuous
1237 to discrete, to make X server driver able to add WB
1238 entry later. This parameter enables that.
1240 enable_timer_pin_1 [X86]
1241 Enable PIN 1 of APIC timer
1242 Can be useful to work around chipset bugs
1243 (in particular on some ATI chipsets).
1244 The kernel tries to set a reasonable default.
1246 enforcing [SELINUX] Set initial enforcing status.
1248 See security/selinux/Kconfig help text.
1249 0 -- permissive (log only, no denials).
1250 1 -- enforcing (deny and log).
1252 Value can be changed at runtime via
1253 /sys/fs/selinux/enforce.
1256 Disable Error Record Serialization Table (ERST)
1259 ether= [HW,NET] Ethernet cards parameters
1260 This option is obsoleted by the "netdev=" option, which
1261 has equivalent usage. See its documentation for details.
1265 Permit 'security.evm' to be updated regardless of
1266 current integrity status.
1270 fail_make_request=[KNL]
1271 General fault injection mechanism.
1272 Format: <interval>,<probability>,<space>,<times>
1273 See also Documentation/fault-injection/.
1276 See Documentation/admin-guide/blockdev/floppy.rst.
1278 force_pal_cache_flush
1279 [IA-64] Avoid check_sal_cache_flush which may hang on
1280 buggy SAL_CACHE_FLUSH implementations. Using this
1281 parameter will force ia64_sal_cache_flush to call
1282 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1285 Forcefully enable Physical Address Extension (PAE).
1286 Many Pentium M systems disable PAE but may have a
1287 functionally usable PAE implementation.
1288 Warning: use of this parameter will taint the kernel
1289 and may cause unknown problems.
1292 [FTRACE] will set and start the specified tracer
1293 as early as possible in order to facilitate early
1296 ftrace_dump_on_oops[=orig_cpu]
1297 [FTRACE] will dump the trace buffers on oops.
1298 If no parameter is passed, ftrace will dump
1299 buffers of all CPUs, but if you pass orig_cpu, it will
1300 dump only the buffer of the CPU that triggered the
1303 ftrace_filter=[function-list]
1304 [FTRACE] Limit the functions traced by the function
1305 tracer at boot up. function-list is a comma separated
1306 list of functions. This list can be changed at run
1307 time by the set_ftrace_filter file in the debugfs
1310 ftrace_notrace=[function-list]
1311 [FTRACE] Do not trace the functions specified in
1312 function-list. This list can be changed at run time
1313 by the set_ftrace_notrace file in the debugfs
1316 ftrace_graph_filter=[function-list]
1317 [FTRACE] Limit the top level callers functions traced
1318 by the function graph tracer at boot up.
1319 function-list is a comma separated list of functions
1320 that can be changed at run time by the
1321 set_graph_function file in the debugfs tracing directory.
1323 ftrace_graph_notrace=[function-list]
1324 [FTRACE] Do not trace from the functions specified in
1325 function-list. This list is a comma separated list of
1326 functions that can be changed at run time by the
1327 set_graph_notrace file in the debugfs tracing directory.
1329 ftrace_graph_max_depth=<uint>
1330 [FTRACE] Used with the function graph tracer. This is
1331 the max depth it will trace into a function. This value
1332 can be changed at run time by the max_graph_depth file
1333 in the tracefs tracing directory. default: 0 (no limit)
1336 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1337 support via parallel port (up to 5 devices per port)
1338 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1339 See also Documentation/input/devices/joystick-parport.rst
1343 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1347 gcov_persist= [GCOV] When non-zero (default), profiling data for
1348 kernel modules is saved and remains accessible via
1349 debugfs, even when the module is unloaded/reloaded.
1350 When zero, profiling data is discarded and associated
1351 debugfs files are removed at module unload time.
1353 goldfish [X86] Enable the goldfish android emulator platform.
1354 Don't use this when you are not running on the
1357 gpt [EFI] Forces disk with valid GPT signature but
1358 invalid Protective MBR to be treated as GPT. If the
1359 primary GPT is corrupted, it enables the backup/alternate
1360 GPT to be used instead.
1362 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1363 the "Enable 0" bit of the configuration register.
1366 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1367 the "Enable 0" bit of the configuration register.
1370 grcan.select= [HW] Select which physical interface to use.
1373 grcan.txsize= [HW] Sets the size of the tx buffer.
1374 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1376 grcan.rxsize= [HW] Sets the size of the rx buffer.
1377 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1380 gpio-mockup.gpio_mockup_ranges
1381 [HW] Sets the ranges of gpiochip of for this device.
1382 Format: <start1>,<end1>,<start2>,<end2>...
1384 hardlockup_all_cpu_backtrace=
1385 [KNL] Should the hard-lockup detector generate
1386 backtraces on all cpus.
1389 hashdist= [KNL,NUMA] Large hashes allocated during boot
1390 are distributed across NUMA nodes. Defaults on
1391 for 64-bit NUMA, off otherwise.
1392 Format: 0 | 1 (for off | on)
1394 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1396 hd= [EIDE] (E)IDE hard drive subsystem geometry
1397 Format: <cyl>,<head>,<sect>
1400 Disable Hardware Error Source Table (HEST) support;
1401 corresponding firmware-first mode error processing
1402 logic will be disabled.
1404 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1405 size of <nn>. This works even on boxes that have no
1406 highmem otherwise. This also works to reduce highmem
1407 size on bigger boxes.
1409 highres= [KNL] Enable/disable high resolution timer mode.
1410 Valid parameters: "on", "off"
1415 hpet= [X86-32,HPET] option to control HPET usage
1416 Format: { enable (default) | disable | force |
1418 disable: disable HPET and use PIT instead
1419 force: allow force enabled of undocumented chips (ICH4,
1421 verbose: show contents of HPET registers during setup
1423 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1424 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1426 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1427 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1428 On x86-64 and powerpc, this option can be specified
1429 multiple times interleaved with hugepages= to reserve
1430 huge pages of different sizes. Valid pages sizes on
1431 x86-64 are 2M (when the CPU supports "pse") and 1G
1432 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1435 [KNL] Should the hung task detector generate panics.
1438 A nonzero value instructs the kernel to panic when a
1439 hung task is detected. The default value is controlled
1440 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1441 option. The value selected by this boot parameter can
1442 be changed later by the kernel.hung_task_panic sysctl.
1444 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1445 terminal devices. Valid values: 0..8
1446 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1447 If specified, z/VM IUCV HVC accepts connections
1448 from listed z/VM user IDs only.
1450 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1451 which allow the hypervisor to 'idle' the
1452 guest on lock contention.
1455 Do not unregister boot console at start. This is only
1456 useful for debugging when something happens in the window
1457 between unregistering the boot console and initializing
1460 i2c_bus= [HW] Override the default board specific I2C bus speed
1461 or register an additional I2C bus that is not
1462 registered from board initialization code.
1466 i8042.debug [HW] Toggle i8042 debug mode
1467 i8042.unmask_kbd_data
1468 [HW] Enable printing of interrupt data from the KBD port
1469 (disabled by default, and as a pre-condition
1470 requires that i8042.debug=1 be enabled)
1471 i8042.direct [HW] Put keyboard port into non-translated mode
1472 i8042.dumbkbd [HW] Pretend that controller can only read data from
1473 keyboard and cannot control its state
1474 (Don't attempt to blink the leds)
1475 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1476 i8042.nokbd [HW] Don't check/create keyboard port
1477 i8042.noloop [HW] Disable the AUX Loopback command while probing
1479 i8042.nomux [HW] Don't check presence of an active multiplexing
1481 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1483 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1484 i8042.reset [HW] Reset the controller during init, cleanup and
1485 suspend-to-ram transitions, only during s2r
1486 transitions, or never reset
1487 Format: { 1 | Y | y | 0 | N | n }
1488 1, Y, y: always reset controller
1489 0, N, n: don't ever reset controller
1490 Default: only on s2r transitions on x86; most other
1491 architectures force reset to be always executed
1492 i8042.unlock [HW] Unlock (ignore) the keylock
1493 i8042.kbdreset [HW] Reset device connected to KBD port
1497 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1498 indicates that the driver is running on unsupported
1500 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1501 does not match list of supported models.
1503 [HW] Report power status in /proc/i8k
1504 (disabled by default)
1505 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1508 i915.invert_brightness=
1509 [DRM] Invert the sense of the variable that is used to
1510 set the brightness of the panel backlight. Normally a
1511 brightness value of 0 indicates backlight switched off,
1512 and the maximum of the brightness value sets the backlight
1513 to maximum brightness. If this parameter is set to 0
1514 (default) and the machine requires it, or this parameter
1515 is set to 1, a brightness value of 0 sets the backlight
1516 to maximum brightness, and the maximum of the brightness
1517 value switches the backlight off.
1518 -1 -- never invert brightness
1519 0 -- machine default
1520 1 -- force brightness inversion
1523 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1525 ide-core.nodma= [HW] (E)IDE subsystem
1526 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1527 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1528 .cdrom .chs .ignore_cable are additional options
1529 See Documentation/ide/ide.rst.
1531 ide-generic.probe-mask= [HW] (E)IDE subsystem
1533 Probe mask for legacy ISA IDE ports. Depending on
1534 platform up to 6 ports are supported, enabled by
1535 setting corresponding bits in the mask to 1. The
1536 default value is 0x0, which has a special meaning.
1537 On systems that have PCI, it triggers scanning the
1538 PCI bus for the first and the second port, which
1539 are then probed. On systems without PCI the value
1540 of 0x0 enables probing the two first ports as if it
1543 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1544 Claim all unknown PCI IDE storage controllers.
1547 Format: idle=poll, idle=halt, idle=nomwait
1548 Poll forces a polling idle loop that can slightly
1549 improve the performance of waking up a idle CPU, but
1550 will use a lot of power and make the system run hot.
1552 idle=halt: Halt is forced to be used for CPU idle.
1553 In such case C2/C3 won't be used again.
1554 idle=nomwait: Disable mwait for CPU C-states
1556 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1557 Format: { strict | legacy | 2008 | relaxed }
1560 Choose which programs will be accepted for execution
1561 based on the IEEE 754 NaN encoding(s) supported by
1562 the FPU and the NaN encoding requested with the value
1563 of an ELF file header flag individually set by each
1564 binary. Hardware implementations are permitted to
1565 support either or both of the legacy and the 2008 NaN
1568 Available settings are as follows:
1569 strict accept binaries that request a NaN encoding
1570 supported by the FPU
1571 legacy only accept legacy-NaN binaries, if supported
1573 2008 only accept 2008-NaN binaries, if supported
1575 relaxed accept any binaries regardless of whether
1576 supported by the FPU
1578 The FPU emulator is always able to support both NaN
1579 encodings, so if no FPU hardware is present or it has
1580 been disabled with 'nofpu', then the settings of
1581 'legacy' and '2008' strap the emulator accordingly,
1582 'relaxed' straps the emulator for both legacy-NaN and
1583 2008-NaN, whereas 'strict' enables legacy-NaN only on
1584 legacy processors and both NaN encodings on MIPS32 or
1587 The setting for ABS.fmt/NEG.fmt instruction execution
1588 mode generally follows that for the NaN encoding,
1589 except where unsupported by hardware.
1591 ignore_loglevel [KNL]
1592 Ignore loglevel setting - this will print /all/
1593 kernel messages to the console. Useful for debugging.
1594 We also add it as printk module parameter, so users
1595 could change it dynamically, usually by
1596 /sys/module/printk/parameters/ignore_loglevel.
1599 Ignore RLIMIT_DATA setting for data mappings,
1600 print warning at first misuse. Can be changed via
1601 /sys/module/kernel/parameters/ignore_rlimit_data.
1603 ihash_entries= [KNL]
1604 Set number of hash buckets for inode cache.
1606 ima_appraise= [IMA] appraise integrity measurements
1607 Format: { "off" | "enforce" | "fix" | "log" }
1610 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1611 The builtin appraise policy appraises all files
1614 ima_canonical_fmt [IMA]
1615 Use the canonical format for the binary runtime
1616 measurements, instead of host native format.
1619 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1623 The list of supported hash algorithms is defined
1624 in crypto/hash_info.h.
1627 The builtin policies to load during IMA setup.
1628 Format: "tcb | appraise_tcb | secure_boot |
1631 The "tcb" policy measures all programs exec'd, files
1632 mmap'd for exec, and all files opened with the read
1633 mode bit set by either the effective uid (euid=0) or
1636 The "appraise_tcb" policy appraises the integrity of
1637 all files owned by root.
1639 The "secure_boot" policy appraises the integrity
1640 of files (eg. kexec kernel image, kernel modules,
1641 firmware, policy, etc) based on file signatures.
1643 The "fail_securely" policy forces file signature
1644 verification failure also on privileged mounted
1645 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1648 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1649 Load a policy which meets the needs of the Trusted
1650 Computing Base. This means IMA will measure all
1651 programs exec'd, files mmap'd for exec, and all files
1652 opened for read by uid=0.
1655 Select one of defined IMA measurements template formats.
1656 Formats: { "ima" | "ima-ng" | "ima-sig" }
1660 [IMA] Define a custom template format.
1661 Format: { "field1|...|fieldN" }
1663 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1664 Format: <min_file_size>
1665 Set the minimal file size for using asynchronous hash.
1666 If left unspecified, ahash usage is disabled.
1668 ahash performance varies for different data sizes on
1669 different crypto accelerators. This option can be used
1670 to achieve the best performance for a particular HW.
1672 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1674 Set hashing buffer size. Default: 4k.
1676 ahash performance varies for different chunk sizes on
1677 different crypto accelerators. This option can be used
1678 to achieve best performance for particular HW.
1682 Run specified binary instead of /sbin/init as init
1685 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1686 for working out where the kernel is dying during
1689 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1690 initcall functions. Useful for debugging built-in
1691 modules and initcalls.
1693 initrd= [BOOT] Specify the location of the initial ramdisk
1695 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1698 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1700 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1702 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1704 init_pkru= [x86] Specify the default memory protection keys rights
1705 register contents for all processes. 0x55555554 by
1706 default (disallow access to all but pkey 0). Can
1707 override in debugfs after boot.
1709 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1712 int_pln_enable [x86] Enable power limit notification interrupt
1714 integrity_audit=[IMA]
1715 Format: { "0" | "1" }
1716 0 -- basic integrity auditing messages. (Default)
1717 1 -- additional integrity auditing messages.
1719 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1721 Enable intel iommu driver.
1723 Disable intel iommu driver.
1724 igfx_off [Default Off]
1725 By default, gfx is mapped as normal device. If a gfx
1726 device has a dedicated DMAR unit, the DMAR unit is
1727 bypassed by not enabling DMAR with this option. In
1728 this case, gfx device will use physical address for
1731 With this option iommu will not optimize to look
1732 for io virtual address below 32-bit forcing dual
1733 address cycle on pci bus for cards supporting greater
1734 than 32-bit addressing. The default is to look
1735 for translation below 32-bit and if not available
1736 then look in the higher range.
1737 strict [Default Off]
1738 With this option on every unmap_single operation will
1739 result in a hardware IOTLB flush operation as opposed
1740 to batching them for performance.
1741 sp_off [Default Off]
1742 By default, super page will be supported if Intel IOMMU
1743 has the capability. With this option, super page will
1746 By default, scalable mode will be disabled even if the
1747 hardware advertises that it has support for the scalable
1748 mode translation. With this option set, scalable mode
1749 will be used on hardware which claims to support it.
1750 tboot_noforce [Default Off]
1751 Do not force the Intel IOMMU enabled under tboot.
1752 By default, tboot will force Intel IOMMU on, which
1753 could harm performance of some high-throughput
1754 devices like 40GBit network cards, even if identity
1756 Note that using this option lowers the security
1757 provided by tboot because it makes the system
1758 vulnerable to DMA attacks.
1759 nobounce [Default off]
1760 Disable bounce buffer for unstrusted devices such as
1761 the Thunderbolt devices. This will treat the untrusted
1762 devices as the trusted ones, hence might expose security
1763 risks of DMA attacks.
1765 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1766 0 disables intel_idle and fall back on acpi_idle.
1767 1 to 9 specify maximum depth of C-state.
1771 Do not enable intel_pstate as the default
1772 scaling driver for the supported processors
1774 Use intel_pstate as a scaling driver, but configure it
1775 to work with generic cpufreq governors (instead of
1776 enabling its internal governor). This mode cannot be
1777 used along with the hardware-managed P-states (HWP)
1780 Enable intel_pstate on systems that prohibit it by default
1781 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1782 instead of acpi-cpufreq may disable platform features, such
1783 as thermal controls and power capping, that rely on ACPI
1784 P-States information being indicated to OSPM and therefore
1785 should be used with caution. This option does not work with
1786 processors that aren't supported by the intel_pstate driver
1787 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1789 Do not enable hardware P state control (HWP)
1792 Only load intel_pstate on systems which support
1793 hardware P state control (HWP) if available.
1795 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1796 Description Table, specifies preferred power management
1797 profile as "Enterprise Server" or "Performance Server",
1798 then this feature is turned on by default.
1800 Allow per-logical-CPU P-State performance control limits using
1801 cpufreq sysfs interface
1803 intremap= [X86-64, Intel-IOMMU]
1804 on enable Interrupt Remapping (default)
1805 off disable Interrupt Remapping
1806 nosid disable Source ID checking
1808 BIOS x2APIC opt-out request will be ignored
1809 nopost disable Interrupt Posting
1811 iomem= Disable strict checking of access to MMIO memory
1812 strict regions from userspace.
1827 nobypass [PPC/POWERNV]
1828 Disable IOMMU bypass, using IOMMU for PCI devices.
1830 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1831 Format: { "0" | "1" }
1833 Request that DMA unmap operations use deferred
1834 invalidation of hardware TLBs, for increased
1835 throughput at the cost of reduced device isolation.
1836 Will fall back to strict mode if not supported by
1837 the relevant IOMMU driver.
1838 1 - Strict mode (default).
1839 DMA unmap operations invalidate IOMMU hardware TLBs
1843 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
1844 Format: { "0" | "1" }
1845 0 - Use IOMMU translation for DMA.
1846 1 - Bypass the IOMMU for DMA.
1847 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
1849 io7= [HW] IO7 for Marvel based alpha systems
1850 See comment before marvel_specify_io7 in
1851 arch/alpha/kernel/core_marvel.c.
1853 io_delay= [X86] I/O delay method
1855 Standard port 0x80 based delay
1857 Alternate port 0xed based delay (needed on some systems)
1859 Simple two microseconds delay
1864 See Documentation/filesystems/nfs/nfsroot.txt.
1866 ipcmni_extend [KNL] Extend the maximum number of unique System V
1867 IPC identifiers from 32,768 to 16,777,216.
1869 irqaffinity= [SMP] Set the default irq affinity mask
1870 The argument is a cpu list, as described above.
1872 irqchip.gicv2_force_probe=
1875 Force the kernel to look for the second 4kB page
1876 of a GICv2 controller even if the memory range
1877 exposed by the device tree is too small.
1879 irqchip.gicv3_nolpi=
1881 Force the kernel to ignore the availability of
1882 LPIs (and by consequence ITSs). Intended for system
1883 that use the kernel as a bootloader, and thus want
1884 to let secondary kernels in charge of setting up
1887 irqchip.gicv3_pseudo_nmi= [ARM64]
1888 Enables support for pseudo-NMIs in the kernel. This
1889 requires the kernel to be built with
1890 CONFIG_ARM64_PSEUDO_NMI.
1893 When an interrupt is not handled search all handlers
1894 for it. Intended to get systems with badly broken
1898 When an interrupt is not handled search all handlers
1899 for it. Also check all handlers each timer
1900 interrupt. Intended to get systems with badly broken
1904 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1906 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1907 [Deprecated - use cpusets instead]
1908 Format: [flag-list,]<cpu-list>
1910 Specify one or more CPUs to isolate from disturbances
1911 specified in the flag list (default: domain):
1914 Disable the tick when a single task runs.
1916 A residual 1Hz tick is offloaded to workqueues, which you
1917 need to affine to housekeeping through the global
1918 workqueue's affinity configured via the
1919 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1920 by using the 'domain' flag described below.
1922 NOTE: by default the global workqueue runs on all CPUs,
1923 so to protect individual CPUs the 'cpumask' file has to
1924 be configured manually after bootup.
1927 Isolate from the general SMP balancing and scheduling
1928 algorithms. Note that performing domain isolation this way
1929 is irreversible: it's not possible to bring back a CPU to
1930 the domains once isolated through isolcpus. It's strongly
1931 advised to use cpusets instead to disable scheduler load
1932 balancing through the "cpuset.sched_load_balance" file.
1933 It offers a much more flexible interface where CPUs can
1934 move in and out of an isolated set anytime.
1936 You can move a process onto or off an "isolated" CPU via
1937 the CPU affinity syscalls or cpuset.
1938 <cpu number> begins at 0 and the maximum value is
1939 "number of CPUs in system - 1".
1943 Isolate from being targeted by managed interrupts
1944 which have an interrupt mask containing isolated
1945 CPUs. The affinity of managed interrupts is
1946 handled by the kernel and cannot be changed via
1947 the /proc/irq/* interfaces.
1949 This isolation is best effort and only effective
1950 if the automatically assigned interrupt mask of a
1951 device queue contains isolated and housekeeping
1952 CPUs. If housekeeping CPUs are online then such
1953 interrupts are directed to the housekeeping CPU
1954 so that IO submitted on the housekeeping CPU
1955 cannot disturb the isolated CPU.
1957 If a queue's affinity mask contains only isolated
1958 CPUs then this parameter has no effect on the
1959 interrupt routing decision, though interrupts are
1960 only delivered when tasks running on those
1961 isolated CPUs submit IO. IO submitted on
1962 housekeeping CPUs has no influence on those
1965 The format of <cpu-list> is described above.
1969 ivrs_ioapic [HW,X86_64]
1970 Provide an override to the IOAPIC-ID<->DEVICE-ID
1971 mapping provided in the IVRS ACPI table. For
1972 example, to map IOAPIC-ID decimal 10 to
1973 PCI device 00:14.0 write the parameter as:
1974 ivrs_ioapic[10]=00:14.0
1976 ivrs_hpet [HW,X86_64]
1977 Provide an override to the HPET-ID<->DEVICE-ID
1978 mapping provided in the IVRS ACPI table. For
1979 example, to map HPET-ID decimal 0 to
1980 PCI device 00:14.0 write the parameter as:
1981 ivrs_hpet[0]=00:14.0
1983 ivrs_acpihid [HW,X86_64]
1984 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1985 mapping provided in the IVRS ACPI table. For
1986 example, to map UART-HID:UID AMD0020:0 to
1987 PCI device 00:14.5 write the parameter as:
1988 ivrs_acpihid[00:14.5]=AMD0020:0
1990 js= [HW,JOY] Analog joystick
1991 See Documentation/input/joydev/joystick.rst.
1994 When CONFIG_RANDOMIZE_BASE is set, this disables
1995 kernel and module base offset ASLR (Address Space
1996 Layout Randomization).
1999 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2000 report on every invalid memory access. Without this
2001 parameter KASAN will print report only for the first
2006 kernelcore= [KNL,X86,IA-64,PPC]
2007 Format: nn[KMGTPE] | nn% | "mirror"
2008 This parameter specifies the amount of memory usable by
2009 the kernel for non-movable allocations. The requested
2010 amount is spread evenly throughout all nodes in the
2011 system as ZONE_NORMAL. The remaining memory is used for
2012 movable memory in its own zone, ZONE_MOVABLE. In the
2013 event, a node is too small to have both ZONE_NORMAL and
2014 ZONE_MOVABLE, kernelcore memory will take priority and
2015 other nodes will have a larger ZONE_MOVABLE.
2017 ZONE_MOVABLE is used for the allocation of pages that
2018 may be reclaimed or moved by the page migration
2019 subsystem. Note that allocations like PTEs-from-HighMem
2020 still use the HighMem zone if it exists, and the Normal
2021 zone if it does not.
2023 It is possible to specify the exact amount of memory in
2024 the form of "nn[KMGTPE]", a percentage of total system
2025 memory in the form of "nn%", or "mirror". If "mirror"
2026 option is specified, mirrored (reliable) memory is used
2027 for non-movable allocations and remaining memory is used
2028 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2029 are exclusive, so you cannot specify multiple forms.
2031 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2032 Format: <Controller#>[,poll interval]
2033 The controller # is the number of the ehci usb debug
2034 port as it is probed via PCI. The poll interval is
2035 optional and is the number seconds in between
2036 each poll cycle to the debug port in case you need
2037 the functionality for interrupting the kernel with
2038 gdb or control-c on the dbgp connection. When
2039 not using this parameter you use sysrq-g to break into
2040 the kernel debugger.
2042 kgdboc= [KGDB,HW] kgdb over consoles.
2043 Requires a tty driver that supports console polling,
2044 or a supported polling keyboard driver (non-usb).
2045 Serial only format: <serial_device>[,baud]
2046 keyboard only format: kbd
2047 keyboard and serial format: kbd,<serial_device>[,baud]
2048 Optional Kernel mode setting:
2049 kms, kbd format: kms,kbd
2050 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2052 kgdbwait [KGDB] Stop kernel execution and enter the
2053 kernel debugger at the earliest opportunity.
2055 kmac= [MIPS] korina ethernet MAC address.
2056 Configure the RouterBoard 532 series on-chip
2057 Ethernet adapter MAC address.
2059 kmemleak= [KNL] Boot-time kmemleak enable/disable
2060 Valid arguments: on, off
2062 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2065 kprobe_event=[probe-list]
2066 [FTRACE] Add kprobe events and enable at boot time.
2067 The probe-list is a semicolon delimited list of probe
2068 definitions. Each definition is same as kprobe_events
2069 interface, but the parameters are comma delimited.
2070 For example, to add a kprobe event on vfs_read with
2071 arg1 and arg2, add to the command line;
2073 kprobe_event=p,vfs_read,$arg1,$arg2
2075 See also Documentation/trace/kprobetrace.rst "Kernel
2076 Boot Parameter" section.
2078 kpti= [ARM64] Control page table isolation of user
2079 and kernel address spaces.
2080 Default: enabled on cores which need mitigation.
2084 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2085 Default is 0 (don't ignore, but inject #GP)
2087 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2088 Default is false (don't support).
2090 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2095 [KVM] Controls the software workaround for the
2096 X86_BUG_ITLB_MULTIHIT bug.
2097 force : Always deploy workaround.
2098 off : Never deploy workaround.
2099 auto : Deploy workaround based on the presence of
2100 X86_BUG_ITLB_MULTIHIT.
2104 If the software workaround is enabled for the host,
2105 guests do need not to enable it for nested guests.
2107 kvm.nx_huge_pages_recovery_ratio=
2108 [KVM] Controls how many 4KiB pages are periodically zapped
2109 back to huge pages. 0 disables the recovery, otherwise if
2110 the value is N KVM will zap 1/Nth of the 4KiB pages every
2111 minute. The default is 60.
2113 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2114 Default is 1 (enabled)
2116 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2118 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2120 kvm-arm.vgic_v3_group0_trap=
2121 [KVM,ARM] Trap guest accesses to GICv3 group-0
2124 kvm-arm.vgic_v3_group1_trap=
2125 [KVM,ARM] Trap guest accesses to GICv3 group-1
2128 kvm-arm.vgic_v3_common_trap=
2129 [KVM,ARM] Trap guest accesses to GICv3 common
2132 kvm-arm.vgic_v4_enable=
2133 [KVM,ARM] Allow use of GICv4 for direct injection of
2136 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2137 (virtualized MMU) support on capable Intel chips.
2138 Default is 1 (enabled)
2140 kvm-intel.emulate_invalid_guest_state=
2141 [KVM,Intel] Enable emulation of invalid guest states
2142 Default is 0 (disabled)
2144 kvm-intel.flexpriority=
2145 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2146 Default is 1 (enabled)
2149 [KVM,Intel] Enable VMX nesting (nVMX).
2150 Default is 0 (disabled)
2152 kvm-intel.unrestricted_guest=
2153 [KVM,Intel] Disable unrestricted guest feature
2154 (virtualized real and unpaged mode) on capable
2155 Intel chips. Default is 1 (enabled)
2157 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2160 Valid arguments: never, cond, always
2162 always: L1D cache flush on every VMENTER.
2163 cond: Flush L1D on VMENTER only when the code between
2164 VMEXIT and VMENTER can leak host memory.
2165 never: Disables the mitigation
2167 Default is cond (do L1 cache flush in specific instances)
2169 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2170 feature (tagged TLBs) on capable Intel chips.
2171 Default is 1 (enabled)
2173 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2176 The kernel PTE inversion protection is unconditionally
2177 enabled and cannot be disabled.
2180 Provides all available mitigations for the
2181 L1TF vulnerability. Disables SMT and
2182 enables all mitigations in the
2183 hypervisors, i.e. unconditional L1D flush.
2185 SMT control and L1D flush control via the
2186 sysfs interface is still possible after
2187 boot. Hypervisors will issue a warning
2188 when the first VM is started in a
2189 potentially insecure configuration,
2190 i.e. SMT enabled or L1D flush disabled.
2193 Same as 'full', but disables SMT and L1D
2194 flush runtime control. Implies the
2195 'nosmt=force' command line option.
2196 (i.e. sysfs control of SMT is disabled.)
2199 Leaves SMT enabled and enables the default
2200 hypervisor mitigation, i.e. conditional
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.
2212 Disables SMT and enables the default
2213 hypervisor mitigation.
2215 SMT control and L1D flush control via the
2216 sysfs interface is still possible after
2217 boot. Hypervisors will issue a warning
2218 when the first VM is started in a
2219 potentially insecure configuration,
2220 i.e. SMT enabled or L1D flush disabled.
2223 Same as 'flush', but hypervisors will not
2224 warn when a VM is started in a potentially
2225 insecure configuration.
2228 Disables hypervisor mitigations and doesn't
2230 It also drops the swap size and available
2231 RAM limit restriction on both hypervisor and
2236 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2242 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2245 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2246 value for LAPIC timer one-shot implementation. Default
2247 back to the programmable timer unit in the LAPIC.
2249 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2252 libata.dma= [LIBATA] DMA control
2253 libata.dma=0 Disable all PATA and SATA DMA
2254 libata.dma=1 PATA and SATA Disk DMA only
2255 libata.dma=2 ATAPI (CDROM) DMA only
2256 libata.dma=4 Compact Flash DMA only
2257 Combinations also work, so libata.dma=3 enables DMA
2258 for disks and CDROMs, but not CFs.
2260 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2261 libata.ignore_hpa=0 keep BIOS limits (default)
2262 libata.ignore_hpa=1 ignore limits, using full disk
2264 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2268 libata.force= [LIBATA] Force configurations. The format is comma
2269 separated list of "[ID:]VAL" where ID is
2270 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2271 matching port, link or device. Basically, it matches
2272 the ATA ID string printed on console by libata. If
2273 the whole ID part is omitted, the last PORT and DEVICE
2274 values are used. If ID hasn't been specified yet, the
2275 configuration applies to all ports, links and devices.
2277 If only DEVICE is omitted, the parameter applies to
2278 the port and all links and devices behind it. DEVICE
2279 number of 0 either selects the first device or the
2280 first fan-out link behind PMP device. It does not
2281 select the host link. DEVICE number of 15 selects the
2282 host link and device attached to it.
2284 The VAL specifies the configuration to force. As long
2285 as there's no ambiguity shortcut notation is allowed.
2286 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2287 The following configurations can be forced.
2289 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2290 Any ID with matching PORT is used.
2292 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2294 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2295 udma[/][16,25,33,44,66,100,133] notation is also
2298 * [no]ncq: Turn on or off NCQ.
2300 * [no]ncqtrim: Turn off queued DSM TRIM.
2302 * nohrst, nosrst, norst: suppress hard, soft
2305 * rstonce: only attempt one reset during
2306 hot-unplug link recovery
2308 * dump_id: dump IDENTIFY data.
2310 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2312 * disable: Disable this device.
2314 If there are multiple matching configurations changing
2315 the same attribute, the last one is used.
2317 memblock=debug [KNL] Enable memblock debug messages.
2319 load_ramdisk= [RAM] List of ramdisks to load from floppy
2320 See Documentation/admin-guide/blockdev/ramdisk.rst.
2322 lockd.nlm_grace_period=P [NFS] Assign grace period.
2325 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2328 lockd.nlm_timeout=T [NFS] Assign timeout value.
2331 lockd.nlm_udpport=M [NFS] Assign UDP port.
2334 lockdown= [SECURITY]
2335 { integrity | confidentiality }
2336 Enable the kernel lockdown feature. If set to
2337 integrity, kernel features that allow userland to
2338 modify the running kernel are disabled. If set to
2339 confidentiality, kernel features that allow userland
2340 to extract confidential information from the kernel
2343 locktorture.nreaders_stress= [KNL]
2344 Set the number of locking read-acquisition kthreads.
2345 Defaults to being automatically set based on the
2346 number of online CPUs.
2348 locktorture.nwriters_stress= [KNL]
2349 Set the number of locking write-acquisition kthreads.
2351 locktorture.onoff_holdoff= [KNL]
2352 Set time (s) after boot for CPU-hotplug testing.
2354 locktorture.onoff_interval= [KNL]
2355 Set time (s) between CPU-hotplug operations, or
2356 zero to disable CPU-hotplug testing.
2358 locktorture.shuffle_interval= [KNL]
2359 Set task-shuffle interval (jiffies). Shuffling
2360 tasks allows some CPUs to go into dyntick-idle
2361 mode during the locktorture test.
2363 locktorture.shutdown_secs= [KNL]
2364 Set time (s) after boot system shutdown. This
2365 is useful for hands-off automated testing.
2367 locktorture.stat_interval= [KNL]
2368 Time (s) between statistics printk()s.
2370 locktorture.stutter= [KNL]
2371 Time (s) to stutter testing, for example,
2372 specifying five seconds causes the test to run for
2373 five seconds, wait for five seconds, and so on.
2374 This tests the locking primitive's ability to
2375 transition abruptly to and from idle.
2377 locktorture.torture_type= [KNL]
2378 Specify the locking implementation to test.
2380 locktorture.verbose= [KNL]
2381 Enable additional printk() statements.
2383 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2386 loglevel= All Kernel Messages with a loglevel smaller than the
2387 console loglevel will be printed to the console. It can
2388 also be changed with klogd or other programs. The
2389 loglevels are defined as follows:
2391 0 (KERN_EMERG) system is unusable
2392 1 (KERN_ALERT) action must be taken immediately
2393 2 (KERN_CRIT) critical conditions
2394 3 (KERN_ERR) error conditions
2395 4 (KERN_WARNING) warning conditions
2396 5 (KERN_NOTICE) normal but significant condition
2397 6 (KERN_INFO) informational
2398 7 (KERN_DEBUG) debug-level messages
2400 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2401 in bytes. n must be a power of two and greater
2402 than the minimal size. The minimal size is defined
2403 by LOG_BUF_SHIFT kernel config parameter. There is
2404 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2405 that allows to increase the default size depending on
2406 the number of CPUs. See init/Kconfig for more details.
2408 logo.nologo [FB] Disables display of the built-in Linux logo.
2409 This may be used to provide more screen space for
2410 kernel log messages and is useful when debugging
2411 kernel boot problems.
2413 lp=0 [LP] Specify parallel ports to use, e.g,
2414 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2415 lp=reset first parallel port). 'lp=0' disables the
2416 lp=auto printer driver. 'lp=reset' (which can be
2417 specified in addition to the ports) causes
2418 attached printers to be reset. Using
2419 lp=port1,port2,... specifies the parallel ports
2420 to associate lp devices with, starting with
2421 lp0. A port specification may be 'none' to skip
2422 that lp device, or a parport name such as
2423 'parport0'. Specifying 'lp=auto' instead of a
2424 port specification list means that device IDs
2425 from each port should be examined, to see if
2426 an IEEE 1284-compliant printer is attached; if
2427 so, the driver will manage that printer.
2428 See also header of drivers/char/lp.c.
2431 Sets loops_per_jiffy to given constant, thus avoiding
2432 time-consuming boot-time autodetection (up to 250 ms per
2433 CPU). 0 enables autodetection (default). To determine
2434 the correct value for your kernel, boot with normal
2435 autodetection and see what value is printed. Note that
2436 on SMP systems the preset will be applied to all CPUs,
2437 which is likely to cause problems if your CPUs need
2438 significantly divergent settings. An incorrect value
2439 will cause delays in the kernel to be wrong, leading to
2440 unpredictable I/O errors and other breakage. Although
2441 unlikely, in the extreme case this might damage your
2445 Format: <io>,<irq>,<dma>
2447 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2450 [SECURITY] Choose order of LSM initialization. This
2451 overrides CONFIG_LSM, and the "security=" parameter.
2453 machvec= [IA-64] Force the use of a particular machine-vector
2454 (machvec) in a generic kernel.
2455 Example: machvec=hpzx1
2457 machtype= [Loongson] Share the same kernel image file between different
2459 Example: machtype=lemote-yeeloong-2f-7inch
2461 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2462 than or equal to this physical address is ignored.
2464 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2465 will bring up during bootup. maxcpus=n : n >= 0 limits
2466 the kernel to bring up 'n' processors. Surely after
2467 bootup you can bring up the other plugged cpu by executing
2468 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2469 only takes effect during system bootup.
2470 While n=0 is a special case, it is equivalent to "nosmp",
2471 which also disables the IO APIC.
2473 max_loop= [LOOP] The number of loop block devices that get
2474 (loop.max_loop) unconditionally pre-created at init time. The default
2475 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2476 of statically allocating a predefined number, loop
2477 devices can be requested on-demand with the
2478 /dev/loop-control interface.
2480 mce [X86-32] Machine Check Exception
2482 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2484 md= [HW] RAID subsystems devices and level
2485 See Documentation/admin-guide/md.rst.
2488 Format: <first>,<last>
2489 Specifies range of consoles to be captured by the MDA.
2492 Control mitigation for the Micro-architectural Data
2493 Sampling (MDS) vulnerability.
2495 Certain CPUs are vulnerable to an exploit against CPU
2496 internal buffers which can forward information to a
2497 disclosure gadget under certain conditions.
2499 In vulnerable processors, the speculatively
2500 forwarded data can be used in a cache side channel
2501 attack, to access data to which the attacker does
2502 not have direct access.
2504 This parameter controls the MDS mitigation. The
2507 full - Enable MDS mitigation on vulnerable CPUs
2508 full,nosmt - Enable MDS mitigation and disable
2509 SMT on vulnerable CPUs
2510 off - Unconditionally disable MDS mitigation
2512 On TAA-affected machines, mds=off can be prevented by
2513 an active TAA mitigation as both vulnerabilities are
2514 mitigated with the same mechanism so in order to disable
2515 this mitigation, you need to specify tsx_async_abort=off
2518 Not specifying this option is equivalent to
2521 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2523 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2524 Amount of memory to be used when the kernel is not able
2525 to see the whole system memory or for test.
2526 [X86] Work as limiting max address. Use together
2527 with memmap= to avoid physical address space collisions.
2528 Without memmap= PCI devices could be placed at addresses
2529 belonging to unused RAM.
2531 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2535 [KNL,SH] Allow user to override the default size for
2536 per-device physically contiguous DMA buffers.
2538 memhp_default_state=online/offline
2539 [KNL] Set the initial state for the memory hotplug
2540 onlining policy. If not specified, the default value is
2541 set according to the
2542 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2544 See Documentation/admin-guide/mm/memory-hotplug.rst.
2546 memmap=exactmap [KNL,X86] Enable setting of an exact
2547 E820 memory map, as specified by the user.
2548 Such memmap=exactmap lines can be constructed based on
2549 BIOS output or other requirements. See the memmap=nn@ss
2552 memmap=nn[KMG]@ss[KMG]
2553 [KNL] Force usage of a specific region of memory.
2554 Region of memory to be used is from ss to ss+nn.
2555 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2556 which limits max address to nn[KMG].
2557 Multiple different regions can be specified,
2560 memmap=100M@2G,100M#3G,1G!1024G
2562 memmap=nn[KMG]#ss[KMG]
2563 [KNL,ACPI] Mark specific memory as ACPI data.
2564 Region of memory to be marked is from ss to ss+nn.
2566 memmap=nn[KMG]$ss[KMG]
2567 [KNL,ACPI] Mark specific memory as reserved.
2568 Region of memory to be reserved is from ss to ss+nn.
2569 Example: Exclude memory from 0x18690000-0x1869ffff
2570 memmap=64K$0x18690000
2572 memmap=0x10000$0x18690000
2573 Some bootloaders may need an escape character before '$',
2574 like Grub2, otherwise '$' and the following number
2577 memmap=nn[KMG]!ss[KMG]
2578 [KNL,X86] Mark specific memory as protected.
2579 Region of memory to be used, from ss to ss+nn.
2580 The memory region may be marked as e820 type 12 (0xc)
2581 and is NVDIMM or ADR memory.
2583 memmap=<size>%<offset>-<oldtype>+<newtype>
2584 [KNL,ACPI] Convert memory within the specified region
2585 from <oldtype> to <newtype>. If "-<oldtype>" is left
2586 out, the whole region will be marked as <newtype>,
2587 even if previously unavailable. If "+<newtype>" is left
2588 out, matching memory will be removed. Types are
2589 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2590 3 = ACPI, 12 = PRAM.
2592 memory_corruption_check=0/1 [X86]
2593 Some BIOSes seem to corrupt the first 64k of
2594 memory when doing things like suspend/resume.
2595 Setting this option will scan the memory
2596 looking for corruption. Enabling this will
2597 both detect corruption and prevent the kernel
2598 from using the memory being corrupted.
2599 However, its intended as a diagnostic tool; if
2600 repeatable BIOS-originated corruption always
2601 affects the same memory, you can use memmap=
2602 to prevent the kernel from using that memory.
2604 memory_corruption_check_size=size [X86]
2605 By default it checks for corruption in the low
2606 64k, making this memory unavailable for normal
2607 use. Use this parameter to scan for
2608 corruption in more or less memory.
2610 memory_corruption_check_period=seconds [X86]
2611 By default it checks for corruption every 60
2612 seconds. Use this parameter to check at some
2613 other rate. 0 disables periodic checking.
2615 memtest= [KNL,X86,ARM,PPC] Enable memtest
2617 default : 0 <disable>
2618 Specifies the number of memtest passes to be
2619 performed. Each pass selects another test
2620 pattern from a given set of patterns. Memtest
2621 fills the memory with this pattern, validates
2622 memory contents and reserves bad memory
2623 regions that are detected.
2625 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2626 Valid arguments: on, off
2627 Default (depends on kernel configuration option):
2628 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2629 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2630 mem_encrypt=on: Activate SME
2631 mem_encrypt=off: Do not activate SME
2633 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2634 for details on when memory encryption can be activated.
2636 mem_sleep_default= [SUSPEND] Default system suspend mode:
2637 s2idle - Suspend-To-Idle
2638 shallow - Power-On Suspend or equivalent (if supported)
2639 deep - Suspend-To-RAM or equivalent (if supported)
2640 See Documentation/admin-guide/pm/sleep-states.rst.
2642 meye.*= [HW] Set MotionEye Camera parameters
2643 See Documentation/media/v4l-drivers/meye.rst.
2645 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2646 Multi-Function General Purpose Timers on AMD Geode
2649 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2650 the BIOS has incorrectly applied a workaround. TinyBIOS
2651 version 0.98 is known to be affected, 0.99 fixes the
2652 problem by letting the user disable the workaround.
2656 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2657 physical address is ignored.
2659 mini2440= [ARM,HW,KNL]
2660 Format:[0..2][b][c][t]
2662 MINI2440 configuration specification:
2663 0 - The attached screen is the 3.5" TFT
2664 1 - The attached screen is the 7" TFT
2665 2 - The VGA Shield is attached (1024x768)
2666 Leaving out the screen size parameter will not load
2667 the TFT driver, and the framebuffer will be left
2669 b - Enable backlight. The TFT backlight pin will be
2670 linked to the kernel VESA blanking code and a GPIO
2671 LED. This parameter is not necessary when using the
2673 c - Enable the s3c camera interface.
2674 t - Reserved for enabling touchscreen support. The
2675 touchscreen support is not enabled in the mainstream
2676 kernel as of 2.6.30, a preliminary port can be found
2677 in the "bleeding edge" mini2440 support kernel at
2678 http://repo.or.cz/w/linux-2.6/mini2440.git
2681 [X86,PPC,S390,ARM64] Control optional mitigations for
2682 CPU vulnerabilities. This is a set of curated,
2683 arch-independent options, each of which is an
2684 aggregation of existing arch-specific options.
2687 Disable all optional CPU mitigations. This
2688 improves system performance, but it may also
2689 expose users to several CPU vulnerabilities.
2690 Equivalent to: nopti [X86,PPC]
2692 nospectre_v1 [X86,PPC]
2694 nospectre_v2 [X86,PPC,S390,ARM64]
2695 spectre_v2_user=off [X86]
2696 spec_store_bypass_disable=off [X86,PPC]
2697 ssbd=force-off [ARM64]
2700 tsx_async_abort=off [X86]
2701 kvm.nx_huge_pages=off [X86]
2704 This does not have any effect on
2705 kvm.nx_huge_pages when
2706 kvm.nx_huge_pages=force.
2709 Mitigate all CPU vulnerabilities, but leave SMT
2710 enabled, even if it's vulnerable. This is for
2711 users who don't want to be surprised by SMT
2712 getting disabled across kernel upgrades, or who
2713 have other ways of avoiding SMT-based attacks.
2714 Equivalent to: (default behavior)
2717 Mitigate all CPU vulnerabilities, disabling SMT
2718 if needed. This is for users who always want to
2719 be fully mitigated, even if it means losing SMT.
2720 Equivalent to: l1tf=flush,nosmt [X86]
2721 mds=full,nosmt [X86]
2722 tsx_async_abort=full,nosmt [X86]
2725 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2726 parameter allows control of the logging verbosity for
2727 the additional memory initialisation checks. A value
2728 of 0 disables mminit logging and a level of 4 will
2729 log everything. Information is printed at KERN_DEBUG
2730 so loglevel=8 may also need to be specified.
2733 [KNL] When CONFIG_MODULE_SIG is set, this means that
2734 modules without (valid) signatures will fail to load.
2735 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2736 is always true, so this option does nothing.
2738 module_blacklist= [KNL] Do not load a comma-separated list of
2739 modules. Useful for debugging problem modules.
2742 [MOUSE] Maximum time between finger touching and
2743 leaving touchpad surface for touch to be considered
2744 a tap and be reported as a left button click (for
2745 touchpads working in absolute mode only).
2747 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2748 reporting absolute coordinates, such as tablets
2749 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2750 reporting absolute coordinates, such as tablets
2752 movablecore= [KNL,X86,IA-64,PPC]
2753 Format: nn[KMGTPE] | nn%
2754 This parameter is the complement to kernelcore=, it
2755 specifies the amount of memory used for migratable
2756 allocations. If both kernelcore and movablecore is
2757 specified, then kernelcore will be at *least* the
2758 specified value but may be more. If movablecore on its
2759 own is specified, the administrator must be careful
2760 that the amount of memory usable for all allocations
2763 movable_node [KNL] Boot-time switch to make hotplugable memory
2764 NUMA nodes to be movable. This means that the memory
2765 of such nodes will be usable only for movable
2766 allocations which rules out almost all kernel
2767 allocations. Use with caution!
2769 MTD_Partition= [MTD]
2770 Format: <name>,<region-number>,<size>,<offset>
2772 MTD_Region= [MTD] Format:
2773 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2776 See drivers/mtd/cmdlinepart.c.
2778 multitce=off [PPC] This parameter disables the use of the pSeries
2779 firmware feature for updating multiple TCE entries
2782 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2784 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2786 boundary - index of last SLC block on Flex-OneNAND.
2787 The remaining blocks are configured as MLC blocks.
2788 lock - Configure if Flex-OneNAND boundary should be locked.
2789 Once locked, the boundary cannot be changed.
2790 1 indicates lock status, 0 indicates unlock status.
2793 ARM/S3C2412 JIVE boot control
2795 See arch/arm/mach-s3c2412/mach-jive.c
2797 mtouchusb.raw_coordinates=
2798 [HW] Make the MicroTouch USB driver use raw coordinates
2799 ('y', default) or cooked coordinates ('n')
2801 mtrr_chunk_size=nn[KMG] [X86]
2802 used for mtrr cleanup. It is largest continuous chunk
2803 that could hold holes aka. UC entries.
2805 mtrr_gran_size=nn[KMG] [X86]
2806 Used for mtrr cleanup. It is granularity of mtrr block.
2808 Large value could prevent small alignment from
2811 mtrr_spare_reg_nr=n [X86]
2813 Range: 0,7 : spare reg number
2815 Used for mtrr cleanup. It is spare mtrr entries number.
2816 Set to 2 or more if your graphical card needs more.
2818 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2820 netdev= [NET] Network devices parameters
2821 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2822 Note that mem_start is often overloaded to mean
2823 something different and driver-specific.
2824 This usage is only documented in each driver source
2828 [NETFILTER] Enable connection tracking flow accounting
2829 0 to disable accounting
2830 1 to enable accounting
2833 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2834 See Documentation/filesystems/nfs/nfsroot.txt.
2836 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2837 See Documentation/filesystems/nfs/nfsroot.txt.
2839 nfsrootdebug [NFS] enable nfsroot debugging messages.
2840 See Documentation/filesystems/nfs/nfsroot.txt.
2842 nfs.callback_nr_threads=
2843 [NFSv4] set the total number of threads that the
2844 NFS client will assign to service NFSv4 callback
2847 nfs.callback_tcpport=
2848 [NFS] set the TCP port on which the NFSv4 callback
2849 channel should listen.
2852 [NFS] sets the pathname to the program which is used
2853 to update the NFS client cache entries.
2855 nfs.cache_getent_timeout=
2856 [NFS] sets the timeout after which an attempt to
2857 update a cache entry is deemed to have failed.
2859 nfs.idmap_cache_timeout=
2860 [NFS] set the maximum lifetime for idmapper cache
2864 [NFS] enable 64-bit inode numbers.
2865 If zero, the NFS client will fake up a 32-bit inode
2866 number for the readdir() and stat() syscalls instead
2867 of returning the full 64-bit number.
2868 The default is to return 64-bit inode numbers.
2870 nfs.max_session_cb_slots=
2871 [NFSv4.1] Sets the maximum number of session
2872 slots the client will assign to the callback
2873 channel. This determines the maximum number of
2874 callbacks the client will process in parallel for
2875 a particular server.
2877 nfs.max_session_slots=
2878 [NFSv4.1] Sets the maximum number of session slots
2879 the client will attempt to negotiate with the server.
2880 This limits the number of simultaneous RPC requests
2881 that the client can send to the NFSv4.1 server.
2882 Note that there is little point in setting this
2883 value higher than the max_tcp_slot_table_limit.
2885 nfs.nfs4_disable_idmapping=
2886 [NFSv4] When set to the default of '1', this option
2887 ensures that both the RPC level authentication
2888 scheme and the NFS level operations agree to use
2889 numeric uids/gids if the mount is using the
2890 'sec=sys' security flavour. In effect it is
2891 disabling idmapping, which can make migration from
2892 legacy NFSv2/v3 systems to NFSv4 easier.
2893 Servers that do not support this mode of operation
2894 will be autodetected by the client, and it will fall
2895 back to using the idmapper.
2896 To turn off this behaviour, set the value to '0'.
2898 [NFS4] Specify an additional fixed unique ident-
2899 ification string that NFSv4 clients can insert into
2900 their nfs_client_id4 string. This is typically a
2901 UUID that is generated at system install time.
2903 nfs.send_implementation_id =
2904 [NFSv4.1] Send client implementation identification
2905 information in exchange_id requests.
2906 If zero, no implementation identification information
2908 The default is to send the implementation identification
2911 nfs.recover_lost_locks =
2912 [NFSv4] Attempt to recover locks that were lost due
2913 to a lease timeout on the server. Please note that
2914 doing this risks data corruption, since there are
2915 no guarantees that the file will remain unchanged
2916 after the locks are lost.
2917 If you want to enable the kernel legacy behaviour of
2918 attempting to recover these locks, then set this
2920 The default parameter value of '0' causes the kernel
2921 not to attempt recovery of lost locks.
2923 nfs4.layoutstats_timer =
2924 [NFSv4.2] Change the rate at which the kernel sends
2925 layoutstats to the pNFS metadata server.
2927 Setting this to value to 0 causes the kernel to use
2928 whatever value is the default set by the layout
2929 driver. A non-zero value sets the minimum interval
2930 in seconds between layoutstats transmissions.
2932 nfsd.nfs4_disable_idmapping=
2933 [NFSv4] When set to the default of '1', the NFSv4
2934 server will return only numeric uids and gids to
2935 clients using auth_sys, and will accept numeric uids
2936 and gids from such clients. This is intended to ease
2937 migration from NFSv2/v3.
2939 nmi_debug= [KNL,SH] Specify one or more actions to take
2940 when a NMI is triggered.
2941 Format: [state][,regs][,debounce][,die]
2943 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2944 Format: [panic,][nopanic,][num]
2946 0 - turn hardlockup detector in nmi_watchdog off
2947 1 - turn hardlockup detector in nmi_watchdog on
2948 When panic is specified, panic when an NMI watchdog
2949 timeout occurs (or 'nopanic' to not panic on an NMI
2950 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
2951 To disable both hard and soft lockup detectors,
2952 please see 'nowatchdog'.
2953 This is useful when you use a panic=... timeout and
2954 need the box quickly up again.
2956 These settings can be accessed at runtime via
2957 the nmi_watchdog and hardlockup_panic sysctls.
2959 netpoll.carrier_timeout=
2960 [NET] Specifies amount of time (in seconds) that
2961 netpoll should wait for a carrier. By default netpoll
2964 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2965 emulation library even if a 387 maths coprocessor
2968 no5lvl [X86-64] Disable 5-level paging mode. Forces
2969 kernel to use 4-level paging instead.
2972 [HW] Never suspend the console
2973 Disable suspending of consoles during suspend and
2974 hibernate operations. Once disabled, debugging
2975 messages can reach various consoles while the rest
2976 of the system is being put to sleep (ie, while
2977 debugging driver suspend/resume hooks). This may
2978 not work reliably with all consoles, but is known
2979 to work with serial and VGA consoles.
2980 To facilitate more flexible debugging, we also add
2981 console_suspend, a printk module parameter to control
2982 it. Users could use console_suspend (usually
2983 /sys/module/printk/parameters/console_suspend) to
2984 turn on/off it dynamically.
2986 novmcoredd [KNL,KDUMP]
2987 Disable device dump. Device dump allows drivers to
2988 append dump data to vmcore so you can collect driver
2989 specified debug info. Drivers can append the data
2990 without any limit and this data is stored in memory,
2991 so this may cause significant memory stress. Disabling
2992 device dump can help save memory but the driver debug
2993 data will be no longer available. This parameter
2994 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
2997 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2998 caches in the slab allocator. Saves per-node memory,
2999 but will impact performance.
3003 noaltinstr [S390] Disables alternative instructions patching
3004 (CPU alternatives feature).
3006 noapic [SMP,APIC] Tells the kernel to not make use of any
3007 IOAPICs that may be present in the system.
3009 noautogroup Disable scheduler automatic task group creation.
3011 nobats [PPC] Do not use BATs for mapping kernel lowmem
3012 on "Classic" PPC cores.
3016 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3018 nodelayacct [KNL] Disable per-task delay accounting
3020 nodsp [SH] Disable hardware DSP at boot time.
3022 noefi Disable EFI runtime services support.
3027 On X86-32 available only on PAE configured kernels.
3028 noexec=on: enable non-executable mappings (default)
3029 noexec=off: disable non-executable mappings
3032 Disable SMAP (Supervisor Mode Access Prevention)
3033 even if it is supported by processor.
3036 Disable SMEP (Supervisor Mode Execution Prevention)
3037 even if it is supported by processor.
3040 This affects only 32-bit executables.
3041 noexec32=on: enable non-executable mappings (default)
3042 read doesn't imply executable mappings
3043 noexec32=off: disable non-executable mappings
3044 read implies executable mappings
3046 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3048 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3049 register save and restore. The kernel will only save
3050 legacy floating-point registers on task switch.
3052 nohugeiomap [KNL,x86,PPC] Disable kernel huge I/O mappings.
3054 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3055 Equivalent to smt=1.
3057 [KNL,x86] Disable symmetric multithreading (SMT).
3058 nosmt=force: Force disable SMT, cannot be undone
3059 via the sysfs control file.
3061 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3062 (bounds check bypass). With this option data leaks are
3063 possible in the system.
3065 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3066 the Spectre variant 2 (indirect branch prediction)
3067 vulnerability. System may allow data leaks with this
3070 nospec_store_bypass_disable
3071 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3073 noxsave [BUGS=X86] Disables x86 extended register state save
3074 and restore using xsave. The kernel will fallback to
3075 enabling legacy floating-point and sse state.
3077 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3078 register states. The kernel will fall back to use
3079 xsave to save the states. By using this parameter,
3080 performance of saving the states is degraded because
3081 xsave doesn't support modified optimization while
3082 xsaveopt supports it on xsaveopt enabled systems.
3084 noxsaves [X86] Disables xsaves and xrstors used in saving and
3085 restoring x86 extended register state in compacted
3086 form of xsave area. The kernel will fall back to use
3087 xsaveopt and xrstor to save and restore the states
3088 in standard form of xsave area. By using this
3089 parameter, xsave area per process might occupy more
3090 memory on xsaves enabled systems.
3092 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
3093 wfi(ARM) instruction doesn't work correctly and not to
3094 use it. This is also useful when using JTAG debugger.
3096 no_file_caps Tells the kernel not to honor file capabilities. The
3097 only way then for a file to be executed with privilege
3098 is to be setuid root or executed by root.
3100 nohalt [IA-64] Tells the kernel not to use the power saving
3101 function PAL_HALT_LIGHT when idle. This increases
3102 power-consumption. On the positive side, it reduces
3103 interrupt wake-up latency, which may improve performance
3104 in certain environments such as networked servers or
3107 nohibernate [HIBERNATION] Disable hibernation and resume.
3109 nohz= [KNL] Boottime enable/disable dynamic ticks
3110 Valid arguments: on, off
3113 nohz_full= [KNL,BOOT,SMP,ISOL]
3114 The argument is a cpu list, as described above.
3115 In kernels built with CONFIG_NO_HZ_FULL=y, set
3116 the specified list of CPUs whose tick will be stopped
3117 whenever possible. The boot CPU will be forced outside
3118 the range to maintain the timekeeping. Any CPUs
3119 in this list will have their RCU callbacks offloaded,
3120 just as if they had also been called out in the
3121 rcu_nocbs= boot parameter.
3123 noiotrap [SH] Disables trapped I/O port accesses.
3125 noirqdebug [X86-32] Disables the code which attempts to detect and
3126 disable unhandled interrupt sources.
3128 no_timer_check [X86,APIC] Disables the code which tests for
3129 broken timer IRQ sources.
3131 noisapnp [ISAPNP] Disables ISA PnP code.
3133 noinitrd [RAM] Tells the kernel not to load any configured
3136 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3138 [Deprecated - use intremap=off]
3142 noinvpcid [X86] Disable the INVPCID cpu feature.
3144 nojitter [IA-64] Disables jitter checking for ITC timers.
3146 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3148 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3152 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3153 clock and use the default one.
3155 no-steal-acc [X86,KVM,ARM64] Disable paravirtualized steal time
3156 accounting. steal time is computed, but won't
3157 influence scheduler behaviour
3159 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3161 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3163 noltlbs [PPC] Do not use large page/tlb entries for kernel
3164 lowmem mapping on PPC40x and PPC8xx
3166 nomca [IA-64] Disable machine check abort handling
3168 nomce [X86-32] Disable Machine Check Exception
3170 nomfgpt [X86-32] Disable Multi-Function General Purpose
3171 Timer usage (for AMD Geode machines).
3173 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3174 shutdown the other cpus. Instead use the REBOOT_VECTOR
3177 nomodule Disable module load
3179 nopat [X86] Disable PAT (page attribute table extension of
3180 pagetables) support.
3182 nopcid [X86-64] Disable the PCID cpu feature.
3184 norandmaps Don't use address space randomization. Equivalent to
3185 echo 0 > /proc/sys/kernel/randomize_va_space
3187 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3188 with UP alternatives
3190 nordrand [X86] Disable kernel use of the RDRAND and
3191 RDSEED instructions even if they are supported
3192 by the processor. RDRAND and RDSEED are still
3193 available to user space applications.
3195 noresume [SWSUSP] Disables resume and restores original swap
3198 no-scroll [VGA] Disables scrollback.
3199 This is required for the Braillex ib80-piezo Braille
3200 reader made by F.H. Papenmeier (Germany).
3204 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3206 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3207 and disable the IO APIC. legacy for "maxcpus=0".
3209 nosoftlockup [KNL] Disable the soft-lockup detector.
3211 nosync [HW,M68K] Disables sync negotiation for all devices.
3213 nowatchdog [KNL] Disable both lockup detectors, i.e.
3214 soft-lockup and NMI watchdog (hard-lockup).
3218 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3220 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3221 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3222 Some features depend on CPU0. Known dependencies are:
3223 1. Resume from suspend/hibernate depends on CPU0.
3224 Suspend/hibernate will fail if CPU0 is offline and you
3225 need to online CPU0 before suspend/hibernate.
3226 2. PIC interrupts also depend on CPU0. CPU0 can't be
3227 removed if a PIC interrupt is detected.
3228 It's said poweroff/reboot may depend on CPU0 on some
3229 machines although I haven't seen such issues so far
3230 after CPU0 is offline on a few tested machines.
3231 If the dependencies are under your control, you can
3232 turn on cpu0_hotplug.
3234 nps_mtm_hs_ctr= [KNL,ARC]
3235 This parameter sets the maximum duration, in
3236 cycles, each HW thread of the CTOP can run
3237 without interruptions, before HW switches it.
3238 The actual maximum duration is 16 times this
3240 Format: integer between 1 and 255
3243 nptcg= [IA-64] Override max number of concurrent global TLB
3244 purges which is reported from either PAL_VM_SUMMARY or
3247 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3248 could support. nr_cpus=n : n >= 1 limits the kernel to
3249 support 'n' processors. It could be larger than the
3250 number of already plugged CPU during bootup, later in
3251 runtime you can physically add extra cpu until it reaches
3252 n. So during boot up some boot time memory for per-cpu
3253 variables need be pre-allocated for later physical cpu
3256 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3258 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3259 Allowed values are enable and disable
3261 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3262 'node', 'default' can be specified
3263 This can be set from sysctl after boot.
3264 See Documentation/admin-guide/sysctl/vm.rst for details.
3266 of_devlink [OF, KNL] Create device links between consumer and
3267 supplier devices by scanning the devictree to infer the
3268 consumer/supplier relationships. A consumer device
3269 will not be probed until all the supplier devices have
3270 probed successfully.
3272 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3273 See Documentation/debugging-via-ohci1394.txt for more
3276 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3277 Rather than timing out after 20 ms if an EC
3278 command is not properly ACKed, override the length
3279 of the timeout. We have interrupts disabled while
3280 waiting for the ACK, so if this is set too high
3281 interrupts *may* be lost!
3283 omap_mux= [OMAP] Override bootloader pin multiplexing.
3284 Format: <mux_mode0.mode_name=value>...
3285 For example, to override I2C bus2:
3286 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3288 oprofile.timer= [HW]
3289 Use timer interrupt instead of performance counters
3291 oprofile.cpu_type= Force an oprofile cpu type
3292 This might be useful if you have an older oprofile
3293 userland or if you want common events.
3294 Format: { arch_perfmon }
3295 arch_perfmon: [X86] Force use of architectural
3296 perfmon on Intel CPUs instead of the
3297 CPU specific event set.
3298 timer: [X86] Force use of architectural NMI
3299 timer mode (see also oprofile.timer
3300 for generic hr timer mode)
3302 oops=panic Always panic on oopses. Default is to just kill the
3303 process, but there is a small probability of
3304 deadlocking the machine.
3305 This will also cause panics on machine check exceptions.
3306 Useful together with panic=30 to trigger a reboot.
3309 [KNL] Boolean flag to control whether the page allocator
3310 should randomize its free lists. The randomization may
3311 be automatically enabled if the kernel detects it is
3312 running on a platform with a direct-mapped memory-side
3313 cache, and this parameter can be used to
3314 override/disable that behavior. The state of the flag
3315 can be read from sysfs at:
3316 /sys/module/page_alloc/parameters/shuffle.
3318 page_owner= [KNL] Boot-time page_owner enabling option.
3319 Storage of the information about who allocated
3320 each page is disabled in default. With this switch,
3322 on: enable the feature
3324 page_poison= [KNL] Boot-time parameter changing the state of
3325 poisoning on the buddy allocator, available with
3326 CONFIG_PAGE_POISONING=y.
3327 off: turn off poisoning (default)
3328 on: turn on poisoning
3330 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3331 timeout > 0: seconds before rebooting
3332 timeout = 0: wait forever
3333 timeout < 0: reboot immediately
3336 panic_print= Bitmask for printing system info when panic happens.
3337 User can chose combination of the following bits:
3338 bit 0: print all tasks info
3339 bit 1: print system memory info
3340 bit 2: print timer info
3341 bit 3: print locks info if CONFIG_LOCKDEP is on
3342 bit 4: print ftrace buffer
3343 bit 5: print all printk messages in buffer
3345 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3348 crash_kexec_post_notifiers
3349 Run kdump after running panic-notifiers and dumping
3350 kmsg. This only for the users who doubt kdump always
3351 succeeds in any situation.
3352 Note that this also increases risks of kdump failure,
3353 because some panic notifiers can make the crashed
3354 kernel more unstable.
3356 parkbd.port= [HW] Parallel port number the keyboard adapter is
3357 connected to, default is 0.
3359 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3360 0 for XT, 1 for AT (default is AT).
3363 parport= [HW,PPT] Specify parallel ports. 0 disables.
3364 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3365 Use 'auto' to force the driver to use any
3366 IRQ/DMA settings detected (the default is to
3367 ignore detected IRQ/DMA settings because of
3368 possible conflicts). You can specify the base
3369 address, IRQ, and DMA settings; IRQ and DMA
3370 should be numbers, or 'auto' (for using detected
3371 settings on that particular port), or 'nofifo'
3372 (to avoid using a FIFO even if it is detected).
3373 Parallel ports are assigned in the order they
3374 are specified on the command line, starting
3377 parport_init_mode= [HW,PPT]
3378 Configure VIA parallel port to operate in
3379 a specific mode. This is necessary on Pegasos
3380 computer where firmware has no options for setting
3381 up parallel port mode and sets it to spp.
3382 Currently this function knows 686a and 8231 chips.
3383 Format: [spp|ps2|epp|ecp|ecpepp]
3386 Halt all CPUs after the first oops has been printed for
3387 the specified number of seconds. This is to be used if
3388 your oopses keep scrolling off the screen.
3393 See header of drivers/block/paride/pcd.c.
3394 See also Documentation/admin-guide/blockdev/paride.rst.
3396 pci=option[,option...] [PCI] various PCI subsystem options.
3398 Some options herein operate on a specific device
3399 or a set of devices (<pci_dev>). These are
3400 specified in one of the following formats:
3402 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3403 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3405 Note: the first format specifies a PCI
3406 bus/device/function address which may change
3407 if new hardware is inserted, if motherboard
3408 firmware changes, or due to changes caused
3409 by other kernel parameters. If the
3410 domain is left unspecified, it is
3411 taken to be zero. Optionally, a path
3412 to a device through multiple device/function
3413 addresses can be specified after the base
3414 address (this is more robust against
3415 renumbering issues). The second format
3416 selects devices using IDs from the
3417 configuration space which may match multiple
3418 devices in the system.
3420 earlydump dump PCI config space before the kernel
3422 off [X86] don't probe for the PCI bus
3423 bios [X86-32] force use of PCI BIOS, don't access
3424 the hardware directly. Use this if your machine
3425 has a non-standard PCI host bridge.
3426 nobios [X86-32] disallow use of PCI BIOS, only direct
3427 hardware access methods are allowed. Use this
3428 if you experience crashes upon bootup and you
3429 suspect they are caused by the BIOS.
3430 conf1 [X86] Force use of PCI Configuration Access
3431 Mechanism 1 (config address in IO port 0xCF8,
3432 data in IO port 0xCFC, both 32-bit).
3433 conf2 [X86] Force use of PCI Configuration Access
3434 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3435 the function, IO port 0xCFA, also 8-bit, sets
3436 bus number. The config space is then accessed
3437 through ports 0xC000-0xCFFF).
3438 See http://wiki.osdev.org/PCI for more info
3439 on the configuration access mechanisms.
3440 noaer [PCIE] If the PCIEAER kernel config parameter is
3441 enabled, this kernel boot option can be used to
3442 disable the use of PCIE advanced error reporting.
3443 nodomains [PCI] Disable support for multiple PCI
3444 root domains (aka PCI segments, in ACPI-speak).
3445 nommconf [X86] Disable use of MMCONFIG for PCI
3447 check_enable_amd_mmconf [X86] check for and enable
3448 properly configured MMIO access to PCI
3449 config space on AMD family 10h CPU
3450 nomsi [MSI] If the PCI_MSI kernel config parameter is
3451 enabled, this kernel boot option can be used to
3452 disable the use of MSI interrupts system-wide.
3453 noioapicquirk [APIC] Disable all boot interrupt quirks.
3454 Safety option to keep boot IRQs enabled. This
3455 should never be necessary.
3456 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3457 primary IO-APIC for bridges that cannot disable
3458 boot IRQs. This fixes a source of spurious IRQs
3459 when the system masks IRQs.
3460 noioapicreroute [APIC] Disable workaround that uses the
3461 boot IRQ equivalent of an IRQ that connects to
3462 a chipset where boot IRQs cannot be disabled.
3463 The opposite of ioapicreroute.
3464 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3465 routing table. These calls are known to be buggy
3466 on several machines and they hang the machine
3467 when used, but on other computers it's the only
3468 way to get the interrupt routing table. Try
3469 this option if the kernel is unable to allocate
3470 IRQs or discover secondary PCI buses on your
3472 rom [X86] Assign address space to expansion ROMs.
3473 Use with caution as certain devices share
3474 address decoders between ROMs and other
3476 norom [X86] Do not assign address space to
3477 expansion ROMs that do not already have
3478 BIOS assigned address ranges.
3479 nobar [X86] Do not assign address space to the
3480 BARs that weren't assigned by the BIOS.
3481 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3482 assigned automatically to PCI devices. You can
3483 make the kernel exclude IRQs of your ISA cards
3485 pirqaddr=0xAAAAA [X86] Specify the physical address
3486 of the PIRQ table (normally generated
3487 by the BIOS) if it is outside the
3488 F0000h-100000h range.
3489 lastbus=N [X86] Scan all buses thru bus #N. Can be
3490 useful if the kernel is unable to find your
3491 secondary buses and you want to tell it
3492 explicitly which ones they are.
3493 assign-busses [X86] Always assign all PCI bus
3494 numbers ourselves, overriding
3495 whatever the firmware may have done.
3496 usepirqmask [X86] Honor the possible IRQ mask stored
3497 in the BIOS $PIR table. This is needed on
3498 some systems with broken BIOSes, notably
3499 some HP Pavilion N5400 and Omnibook XE3
3500 notebooks. This will have no effect if ACPI
3501 IRQ routing is enabled.
3502 noacpi [X86] Do not use ACPI for IRQ routing
3503 or for PCI scanning.
3504 use_crs [X86] Use PCI host bridge window information
3505 from ACPI. On BIOSes from 2008 or later, this
3506 is enabled by default. If you need to use this,
3507 please report a bug.
3508 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3509 If you need to use this, please report a bug.
3510 routeirq Do IRQ routing for all PCI devices.
3511 This is normally done in pci_enable_device(),
3512 so this option is a temporary workaround
3513 for broken drivers that don't call it.
3514 skip_isa_align [X86] do not align io start addr, so can
3515 handle more pci cards
3516 noearly [X86] Don't do any early type 1 scanning.
3517 This might help on some broken boards which
3518 machine check when some devices' config space
3519 is read. But various workarounds are disabled
3520 and some IOMMU drivers will not work.
3521 bfsort Sort PCI devices into breadth-first order.
3522 This sorting is done to get a device
3523 order compatible with older (<= 2.4) kernels.
3524 nobfsort Don't sort PCI devices into breadth-first order.
3525 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3526 tuning and use the BIOS-configured MPS defaults.
3527 pcie_bus_safe Set every device's MPS to the largest value
3528 supported by all devices below the root complex.
3529 pcie_bus_perf Set device MPS to the largest allowable MPS
3530 based on its parent bus. Also set MRRS (Max
3531 Read Request Size) to the largest supported
3532 value (no larger than the MPS that the device
3533 or bus can support) for best performance.
3534 pcie_bus_peer2peer Set every device's MPS to 128B, which
3535 every device is guaranteed to support. This
3536 configuration allows peer-to-peer DMA between
3537 any pair of devices, possibly at the cost of
3538 reduced performance. This also guarantees
3539 that hot-added devices will work.
3540 cbiosize=nn[KMG] The fixed amount of bus space which is
3541 reserved for the CardBus bridge's IO window.
3542 The default value is 256 bytes.
3543 cbmemsize=nn[KMG] The fixed amount of bus space which is
3544 reserved for the CardBus bridge's memory
3545 window. The default value is 64 megabytes.
3548 [<order of align>@]<pci_dev>[; ...]
3549 Specifies alignment and device to reassign
3550 aligned memory resources. How to
3551 specify the device is described above.
3552 If <order of align> is not specified,
3553 PAGE_SIZE is used as alignment.
3554 A PCI-PCI bridge can be specified if resource
3555 windows need to be expanded.
3556 To specify the alignment for several
3557 instances of a device, the PCI vendor,
3558 device, subvendor, and subdevice may be
3559 specified, e.g., 12@pci:8086:9c22:103c:198f
3560 for 4096-byte alignment.
3561 ecrc= Enable/disable PCIe ECRC (transaction layer
3562 end-to-end CRC checking).
3563 bios: Use BIOS/firmware settings. This is the
3567 hpiosize=nn[KMG] The fixed amount of bus space which is
3568 reserved for hotplug bridge's IO window.
3569 Default size is 256 bytes.
3570 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3571 reserved for hotplug bridge's MMIO window.
3572 Default size is 2 megabytes.
3573 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3574 reserved for hotplug bridge's MMIO_PREF window.
3575 Default size is 2 megabytes.
3576 hpmemsize=nn[KMG] The fixed amount of bus space which is
3577 reserved for hotplug bridge's MMIO and
3579 Default size is 2 megabytes.
3580 hpbussize=nn The minimum amount of additional bus numbers
3581 reserved for buses below a hotplug bridge.
3583 realloc= Enable/disable reallocating PCI bridge resources
3584 if allocations done by BIOS are too small to
3585 accommodate resources required by all child
3587 off: Turn realloc off
3589 realloc same as realloc=on
3590 noari do not use PCIe ARI.
3591 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3592 do not use PCIe ATS (and IOMMU device IOTLB).
3593 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3594 only look for one device below a PCIe downstream
3596 big_root_window Try to add a big 64bit memory window to the PCIe
3597 root complex on AMD CPUs. Some GFX hardware
3598 can resize a BAR to allow access to all VRAM.
3599 Adding the window is slightly risky (it may
3600 conflict with unreported devices), so this
3602 disable_acs_redir=<pci_dev>[; ...]
3603 Specify one or more PCI devices (in the format
3604 specified above) separated by semicolons.
3605 Each device specified will have the PCI ACS
3606 redirect capabilities forced off which will
3607 allow P2P traffic between devices through
3608 bridges without forcing it upstream. Note:
3609 this removes isolation between devices and
3610 may put more devices in an IOMMU group.
3611 force_floating [S390] Force usage of floating interrupts.
3612 nomio [S390] Do not use MIO instructions.
3614 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3617 force Enable ASPM even on devices that claim not to support it.
3618 WARNING: Forcing ASPM on may cause system lockups.
3620 pcie_ports= [PCIE] PCIe port services handling:
3621 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3622 even if the platform doesn't give the OS permission to
3623 use them. This may cause conflicts if the platform
3624 also tries to use these services.
3625 dpc-native Use native PCIe service for DPC only. May
3626 cause conflicts if firmware uses AER or DPC.
3627 compat Disable native PCIe services (PME, AER, DPC, PCIe
3630 pcie_port_pm= [PCIE] PCIe port power management handling:
3631 off Disable power management of all PCIe ports
3632 force Forcibly enable power management of all PCIe ports
3634 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3635 nomsi Do not use MSI for native PCIe PME signaling (this makes
3636 all PCIe root ports use INTx for all services).
3638 pcmv= [HW,PCMCIA] BadgePAD 4
3642 Keep all power-domains already enabled by bootloader on,
3643 even if no driver has claimed them. This is useful
3644 for debug and development, but should not be
3645 needed on a platform with proper driver support.
3648 See Documentation/admin-guide/blockdev/paride.rst.
3650 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3653 See arch/parisc/kernel/pdc_chassis.c
3655 percpu_alloc= Select which percpu first chunk allocator to use.
3656 Currently supported values are "embed" and "page".
3657 Archs may support subset or none of the selections.
3658 See comments in mm/percpu.c for details on each
3659 allocator. This parameter is primarily for debugging
3660 and performance comparison.
3663 See Documentation/admin-guide/blockdev/paride.rst.
3666 See Documentation/admin-guide/blockdev/paride.rst.
3668 pirq= [SMP,APIC] Manual mp-table setup
3669 See Documentation/x86/i386/IO-APIC.rst.
3671 plip= [PPT,NET] Parallel port network link
3672 Format: { parport<nr> | timid | 0 }
3673 See also Documentation/admin-guide/parport.rst.
3675 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3676 Override pmtimer IOPort with a hex value.
3680 Enable PNP debug messages (depends on the
3681 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3682 via /sys/module/pnp/parameters/debug. We always show
3683 current resource usage; turning this on also shows
3684 possible settings and some assignment information.
3690 { on | off | curr | res | no-curr | no-res }
3693 [ISAPNP] Exclude IRQs for the autoconfiguration
3696 [ISAPNP] Exclude DMAs for the autoconfiguration
3698 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3699 Ranges are in pairs (I/O port base and size).
3702 [ISAPNP] Exclude memory regions for the
3704 Ranges are in pairs (memory base and size).
3706 ports= [IP_VS_FTP] IPVS ftp helper module
3708 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3710 Format: <port>,<port>....
3712 powersave=off [PPC] This option disables power saving features.
3713 It specifically disables cpuidle and sets the
3714 platform machine description specific power_save
3715 function to NULL. On Idle the CPU just reduces
3718 ppc_strict_facility_enable
3719 [PPC] This option catches any kernel floating point,
3720 Altivec, VSX and SPE outside of regions specifically
3721 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3722 There is some performance impact when enabling this.
3726 Disable Hardware Transactional Memory
3728 print-fatal-signals=
3729 [KNL] debug: print fatal signals
3731 If enabled, warn about various signal handling
3732 related application anomalies: too many signals,
3733 too many POSIX.1 timers, fatal signals causing a
3736 If you hit the warning due to signal overflow,
3737 you might want to try "ulimit -i unlimited".
3741 printk.always_kmsg_dump=
3742 Trigger kmsg_dump for cases other than kernel oops or
3744 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3747 printk.devkmsg={on,off,ratelimit}
3748 Control writing to /dev/kmsg.
3749 on - unlimited logging to /dev/kmsg from userspace
3750 off - logging to /dev/kmsg disabled
3751 ratelimit - ratelimit the logging
3754 printk.time= Show timing data prefixed to each printk message line
3755 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3757 processor.max_cstate= [HW,ACPI]
3758 Limit processor to maximum C-state
3759 max_cstate=9 overrides any DMI blacklist limit.
3761 processor.nocst [HW,ACPI]
3762 Ignore the _CST method to determine C-states,
3763 instead using the legacy FADT method
3765 profile= [KNL] Enable kernel profiling via /proc/profile
3766 Format: [<profiletype>,]<number>
3767 Param: <profiletype>: "schedule", "sleep", or "kvm"
3768 [defaults to kernel profiling]
3769 Param: "schedule" - profile schedule points.
3770 Param: "sleep" - profile D-state sleeping (millisecs).
3771 Requires CONFIG_SCHEDSTATS
3772 Param: "kvm" - profile VM exits.
3773 Param: <number> - step/bucket size as a power of 2 for
3774 statistical time based profiling.
3776 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3778 See Documentation/admin-guide/blockdev/ramdisk.rst.
3780 psi= [KNL] Enable or disable pressure stall information
3784 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3785 probe for; one of (bare|imps|exps|lifebook|any).
3786 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3788 psmouse.resetafter= [HW,MOUSE]
3789 Try to reset the device after so many bad packets
3792 [HW,MOUSE] Set desired mouse resolution, in dpi.
3793 psmouse.smartscroll=
3794 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3795 0 = disabled, 1 = enabled (default).
3797 pstore.backend= Specify the name of the pstore backend to use
3800 See Documentation/admin-guide/blockdev/paride.rst.
3802 pti= [X86_64] Control Page Table Isolation of user and
3803 kernel address spaces. Disabling this feature
3804 removes hardening, but improves performance of
3805 system calls and interrupts.
3807 on - unconditionally enable
3808 off - unconditionally disable
3809 auto - kernel detects whether your CPU model is
3810 vulnerable to issues that PTI mitigates
3812 Not specifying this option is equivalent to pti=auto.
3815 Equivalent to pti=off
3818 [KNL] Number of legacy pty's. Overwrites compiled-in
3821 quiet [KNL] Disable most log messages
3826 See Documentation/admin-guide/md.rst.
3828 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3829 See Documentation/admin-guide/blockdev/ramdisk.rst.
3831 random.trust_cpu={on,off}
3832 [KNL] Enable or disable trusting the use of the
3833 CPU's random number generator (if available) to
3834 fully seed the kernel's CRNG. Default is controlled
3835 by CONFIG_RANDOM_TRUST_CPU.
3837 ras=option[,option,...] [KNL] RAS-specific options
3840 Disable the Correctable Errors Collector,
3841 see CONFIG_RAS_CEC help text.
3844 The argument is a cpu list, as described above,
3845 except that the string "all" can be used to
3846 specify every CPU on the system.
3848 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3849 the specified list of CPUs to be no-callback CPUs.
3850 Invocation of these CPUs' RCU callbacks will be
3851 offloaded to "rcuox/N" kthreads created for that
3852 purpose, where "x" is "p" for RCU-preempt, and
3853 "s" for RCU-sched, and "N" is the CPU number.
3854 This reduces OS jitter on the offloaded CPUs,
3855 which can be useful for HPC and real-time
3856 workloads. It can also improve energy efficiency
3857 for asymmetric multiprocessors.
3860 Rather than requiring that offloaded CPUs
3861 (specified by rcu_nocbs= above) explicitly
3862 awaken the corresponding "rcuoN" kthreads,
3863 make these kthreads poll for callbacks.
3864 This improves the real-time response for the
3865 offloaded CPUs by relieving them of the need to
3866 wake up the corresponding kthread, but degrades
3867 energy efficiency by requiring that the kthreads
3868 periodically wake up to do the polling.
3870 rcutree.blimit= [KNL]
3871 Set maximum number of finished RCU callbacks to
3872 process in one batch.
3874 rcutree.dump_tree= [KNL]
3875 Dump the structure of the rcu_node combining tree
3876 out at early boot. This is used for diagnostic
3877 purposes, to verify correct tree setup.
3879 rcutree.gp_cleanup_delay= [KNL]
3880 Set the number of jiffies to delay each step of
3881 RCU grace-period cleanup.
3883 rcutree.gp_init_delay= [KNL]
3884 Set the number of jiffies to delay each step of
3885 RCU grace-period initialization.
3887 rcutree.gp_preinit_delay= [KNL]
3888 Set the number of jiffies to delay each step of
3889 RCU grace-period pre-initialization, that is,
3890 the propagation of recent CPU-hotplug changes up
3891 the rcu_node combining tree.
3893 rcutree.use_softirq= [KNL]
3894 If set to zero, move all RCU_SOFTIRQ processing to
3895 per-CPU rcuc kthreads. Defaults to a non-zero
3896 value, meaning that RCU_SOFTIRQ is used by default.
3897 Specify rcutree.use_softirq=0 to use rcuc kthreads.
3899 rcutree.rcu_fanout_exact= [KNL]
3900 Disable autobalancing of the rcu_node combining
3901 tree. This is used by rcutorture, and might
3902 possibly be useful for architectures having high
3903 cache-to-cache transfer latencies.
3905 rcutree.rcu_fanout_leaf= [KNL]
3906 Change the number of CPUs assigned to each
3907 leaf rcu_node structure. Useful for very
3908 large systems, which will choose the value 64,
3909 and for NUMA systems with large remote-access
3910 latencies, which will choose a value aligned
3911 with the appropriate hardware boundaries.
3913 rcutree.jiffies_till_first_fqs= [KNL]
3914 Set delay from grace-period initialization to
3915 first attempt to force quiescent states.
3916 Units are jiffies, minimum value is zero,
3917 and maximum value is HZ.
3919 rcutree.jiffies_till_next_fqs= [KNL]
3920 Set delay between subsequent attempts to force
3921 quiescent states. Units are jiffies, minimum
3922 value is one, and maximum value is HZ.
3924 rcutree.jiffies_till_sched_qs= [KNL]
3925 Set required age in jiffies for a
3926 given grace period before RCU starts
3927 soliciting quiescent-state help from
3928 rcu_note_context_switch() and cond_resched().
3929 If not specified, the kernel will calculate
3930 a value based on the most recent settings
3931 of rcutree.jiffies_till_first_fqs
3932 and rcutree.jiffies_till_next_fqs.
3933 This calculated value may be viewed in
3934 rcutree.jiffies_to_sched_qs. Any attempt to set
3935 rcutree.jiffies_to_sched_qs will be cheerfully
3938 rcutree.kthread_prio= [KNL,BOOT]
3939 Set the SCHED_FIFO priority of the RCU per-CPU
3940 kthreads (rcuc/N). This value is also used for
3941 the priority of the RCU boost threads (rcub/N)
3942 and for the RCU grace-period kthreads (rcu_bh,
3943 rcu_preempt, and rcu_sched). If RCU_BOOST is
3944 set, valid values are 1-99 and the default is 1
3945 (the least-favored priority). Otherwise, when
3946 RCU_BOOST is not set, valid values are 0-99 and
3947 the default is zero (non-realtime operation).
3949 rcutree.rcu_nocb_gp_stride= [KNL]
3950 Set the number of NOCB callback kthreads in
3951 each group, which defaults to the square root
3952 of the number of CPUs. Larger numbers reduce
3953 the wakeup overhead on the global grace-period
3954 kthread, but increases that same overhead on
3955 each group's NOCB grace-period kthread.
3957 rcutree.qhimark= [KNL]
3958 Set threshold of queued RCU callbacks beyond which
3959 batch limiting is disabled.
3961 rcutree.qlowmark= [KNL]
3962 Set threshold of queued RCU callbacks below which
3963 batch limiting is re-enabled.
3965 rcutree.rcu_idle_gp_delay= [KNL]
3966 Set wakeup interval for idle CPUs that have
3967 RCU callbacks (RCU_FAST_NO_HZ=y).
3969 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3970 Set wakeup interval for idle CPUs that have
3971 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3972 Lazy RCU callbacks are those which RCU can
3973 prove do nothing more than free memory.
3975 rcutree.rcu_kick_kthreads= [KNL]
3976 Cause the grace-period kthread to get an extra
3977 wake_up() if it sleeps three times longer than
3978 it should at force-quiescent-state time.
3979 This wake_up() will be accompanied by a
3980 WARN_ONCE() splat and an ftrace_dump().
3982 rcutree.sysrq_rcu= [KNL]
3983 Commandeer a sysrq key to dump out Tree RCU's
3984 rcu_node tree with an eye towards determining
3985 why a new grace period has not yet started.
3987 rcuperf.gp_async= [KNL]
3988 Measure performance of asynchronous
3989 grace-period primitives such as call_rcu().
3991 rcuperf.gp_async_max= [KNL]
3992 Specify the maximum number of outstanding
3993 callbacks per writer thread. When a writer
3994 thread exceeds this limit, it invokes the
3995 corresponding flavor of rcu_barrier() to allow
3996 previously posted callbacks to drain.
3998 rcuperf.gp_exp= [KNL]
3999 Measure performance of expedited synchronous
4000 grace-period primitives.
4002 rcuperf.holdoff= [KNL]
4003 Set test-start holdoff period. The purpose of
4004 this parameter is to delay the start of the
4005 test until boot completes in order to avoid
4008 rcuperf.kfree_rcu_test= [KNL]
4009 Set to measure performance of kfree_rcu() flooding.
4011 rcuperf.kfree_nthreads= [KNL]
4012 The number of threads running loops of kfree_rcu().
4014 rcuperf.kfree_alloc_num= [KNL]
4015 Number of allocations and frees done in an iteration.
4017 rcuperf.kfree_loops= [KNL]
4018 Number of loops doing rcuperf.kfree_alloc_num number
4019 of allocations and frees.
4021 rcuperf.nreaders= [KNL]
4022 Set number of RCU readers. The value -1 selects
4023 N, where N is the number of CPUs. A value
4024 "n" less than -1 selects N-n+1, where N is again
4025 the number of CPUs. For example, -2 selects N
4026 (the number of CPUs), -3 selects N+1, and so on.
4027 A value of "n" less than or equal to -N selects
4030 rcuperf.nwriters= [KNL]
4031 Set number of RCU writers. The values operate
4032 the same as for rcuperf.nreaders.
4033 N, where N is the number of CPUs
4035 rcuperf.perf_type= [KNL]
4036 Specify the RCU implementation to test.
4038 rcuperf.shutdown= [KNL]
4039 Shut the system down after performance tests
4040 complete. This is useful for hands-off automated
4043 rcuperf.verbose= [KNL]
4044 Enable additional printk() statements.
4046 rcuperf.writer_holdoff= [KNL]
4047 Write-side holdoff between grace periods,
4048 in microseconds. The default of zero says
4051 rcutorture.fqs_duration= [KNL]
4052 Set duration of force_quiescent_state bursts
4055 rcutorture.fqs_holdoff= [KNL]
4056 Set holdoff time within force_quiescent_state bursts
4059 rcutorture.fqs_stutter= [KNL]
4060 Set wait time between force_quiescent_state bursts
4063 rcutorture.fwd_progress= [KNL]
4064 Enable RCU grace-period forward-progress testing
4065 for the types of RCU supporting this notion.
4067 rcutorture.fwd_progress_div= [KNL]
4068 Specify the fraction of a CPU-stall-warning
4069 period to do tight-loop forward-progress testing.
4071 rcutorture.fwd_progress_holdoff= [KNL]
4072 Number of seconds to wait between successive
4073 forward-progress tests.
4075 rcutorture.fwd_progress_need_resched= [KNL]
4076 Enclose cond_resched() calls within checks for
4077 need_resched() during tight-loop forward-progress
4080 rcutorture.gp_cond= [KNL]
4081 Use conditional/asynchronous update-side
4082 primitives, if available.
4084 rcutorture.gp_exp= [KNL]
4085 Use expedited update-side primitives, if available.
4087 rcutorture.gp_normal= [KNL]
4088 Use normal (non-expedited) asynchronous
4089 update-side primitives, if available.
4091 rcutorture.gp_sync= [KNL]
4092 Use normal (non-expedited) synchronous
4093 update-side primitives, if available. If all
4094 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4095 rcutorture.gp_normal=, and rcutorture.gp_sync=
4096 are zero, rcutorture acts as if is interpreted
4097 they are all non-zero.
4099 rcutorture.n_barrier_cbs= [KNL]
4100 Set callbacks/threads for rcu_barrier() testing.
4102 rcutorture.nfakewriters= [KNL]
4103 Set number of concurrent RCU writers. These just
4104 stress RCU, they don't participate in the actual
4105 test, hence the "fake".
4107 rcutorture.nreaders= [KNL]
4108 Set number of RCU readers. The value -1 selects
4109 N-1, where N is the number of CPUs. A value
4110 "n" less than -1 selects N-n-2, where N is again
4111 the number of CPUs. For example, -2 selects N
4112 (the number of CPUs), -3 selects N+1, and so on.
4114 rcutorture.object_debug= [KNL]
4115 Enable debug-object double-call_rcu() testing.
4117 rcutorture.onoff_holdoff= [KNL]
4118 Set time (s) after boot for CPU-hotplug testing.
4120 rcutorture.onoff_interval= [KNL]
4121 Set time (jiffies) between CPU-hotplug operations,
4122 or zero to disable CPU-hotplug testing.
4124 rcutorture.shuffle_interval= [KNL]
4125 Set task-shuffle interval (s). Shuffling tasks
4126 allows some CPUs to go into dyntick-idle mode
4127 during the rcutorture test.
4129 rcutorture.shutdown_secs= [KNL]
4130 Set time (s) after boot system shutdown. This
4131 is useful for hands-off automated testing.
4133 rcutorture.stall_cpu= [KNL]
4134 Duration of CPU stall (s) to test RCU CPU stall
4135 warnings, zero to disable.
4137 rcutorture.stall_cpu_holdoff= [KNL]
4138 Time to wait (s) after boot before inducing stall.
4140 rcutorture.stall_cpu_irqsoff= [KNL]
4141 Disable interrupts while stalling if set.
4143 rcutorture.stat_interval= [KNL]
4144 Time (s) between statistics printk()s.
4146 rcutorture.stutter= [KNL]
4147 Time (s) to stutter testing, for example, specifying
4148 five seconds causes the test to run for five seconds,
4149 wait for five seconds, and so on. This tests RCU's
4150 ability to transition abruptly to and from idle.
4152 rcutorture.test_boost= [KNL]
4153 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4154 "Maybe" means test if the RCU implementation
4155 under test support RCU priority boosting.
4157 rcutorture.test_boost_duration= [KNL]
4158 Duration (s) of each individual boost test.
4160 rcutorture.test_boost_interval= [KNL]
4161 Interval (s) between each boost test.
4163 rcutorture.test_no_idle_hz= [KNL]
4164 Test RCU's dyntick-idle handling. See also the
4165 rcutorture.shuffle_interval parameter.
4167 rcutorture.torture_type= [KNL]
4168 Specify the RCU implementation to test.
4170 rcutorture.verbose= [KNL]
4171 Enable additional printk() statements.
4173 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4174 Dump ftrace buffer after reporting RCU CPU
4177 rcupdate.rcu_cpu_stall_suppress= [KNL]
4178 Suppress RCU CPU stall warning messages.
4180 rcupdate.rcu_cpu_stall_timeout= [KNL]
4181 Set timeout for RCU CPU stall warning messages.
4183 rcupdate.rcu_expedited= [KNL]
4184 Use expedited grace-period primitives, for
4185 example, synchronize_rcu_expedited() instead
4186 of synchronize_rcu(). This reduces latency,
4187 but can increase CPU utilization, degrade
4188 real-time latency, and degrade energy efficiency.
4189 No effect on CONFIG_TINY_RCU kernels.
4191 rcupdate.rcu_normal= [KNL]
4192 Use only normal grace-period primitives,
4193 for example, synchronize_rcu() instead of
4194 synchronize_rcu_expedited(). This improves
4195 real-time latency, CPU utilization, and
4196 energy efficiency, but can expose users to
4197 increased grace-period latency. This parameter
4198 overrides rcupdate.rcu_expedited. No effect on
4199 CONFIG_TINY_RCU kernels.
4201 rcupdate.rcu_normal_after_boot= [KNL]
4202 Once boot has completed (that is, after
4203 rcu_end_inkernel_boot() has been invoked), use
4204 only normal grace-period primitives. No effect
4205 on CONFIG_TINY_RCU kernels.
4207 rcupdate.rcu_task_stall_timeout= [KNL]
4208 Set timeout in jiffies for RCU task stall warning
4209 messages. Disable with a value less than or equal
4212 rcupdate.rcu_self_test= [KNL]
4213 Run the RCU early boot self tests
4217 Run specified binary instead of /init from the ramdisk,
4218 used for early userspace startup. See initrd.
4221 force - Override the decision by the kernel to hide the
4222 advertisement of RDRAND support (this affects
4223 certain AMD processors because of buggy BIOS
4224 support, specifically around the suspend/resume
4228 Turn on/off individual RDT features. List is:
4229 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4231 E.g. to turn on cmt and turn off mba use:
4235 Format (x86 or x86_64):
4236 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4238 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4240 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4241 (prefix with 'panic_' to set mode for panic
4243 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4244 reboot_force is either force or not specified,
4245 reboot_cpu is s[mp]#### with #### being the processor
4246 to be used for rebooting.
4249 [KNL, SMP] Set scheduler's default relax_domain_level.
4250 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4252 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4253 Format: <base1>,<size1>[,<base2>,<size2>,...]
4254 Reserve I/O ports or memory so the kernel won't use
4255 them. If <base> is less than 0x10000, the region
4256 is assumed to be I/O ports; otherwise it is memory.
4258 reservetop= [X86-32]
4260 Reserves a hole at the top of the kernel virtual
4265 Set the amount of memory to reserve for BIOS at
4266 the bottom of the address space.
4268 reset_devices [KNL] Force drivers to reset the underlying device
4269 during initialization.
4272 Specify the partition device for software suspend
4274 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4276 resume_offset= [SWSUSP]
4277 Specify the offset from the beginning of the partition
4278 given by "resume=" at which the swap header is located,
4279 in <PAGE_SIZE> units (needed only for swap files).
4280 See Documentation/power/swsusp-and-swap-files.rst
4282 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4283 read the resume files
4285 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4286 Useful for devices that are detected asynchronously
4287 (e.g. USB and MMC devices).
4289 hibernate= [HIBERNATION]
4290 noresume Don't check if there's a hibernation image
4291 present during boot.
4292 nocompress Don't compress/decompress hibernation images.
4293 no Disable hibernation and resume.
4294 protect_image Turn on image protection during restoration
4295 (that will set all pages holding image data
4296 during restoration read-only).
4298 retain_initrd [RAM] Keep initrd memory after extraction
4300 rfkill.default_state=
4301 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4302 etc. communication is blocked by default.
4305 rfkill.master_switch_mode=
4306 0 The "airplane mode" button does nothing.
4307 1 The "airplane mode" button toggles between everything
4308 blocked and the previous configuration.
4309 2 The "airplane mode" button toggles between everything
4310 blocked and everything unblocked.
4312 rhash_entries= [KNL,NET]
4313 Set number of hash buckets for route cache
4316 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4319 ro [KNL] Mount root device read-only on boot
4322 on Mark read-only kernel memory as read-only (default).
4323 off Leave read-only kernel memory writable for debugging.
4326 Enable the uart passthrough on the designated usb port
4327 on Rockchip SoCs. When active, the signals of the
4328 debug-uart get routed to the D+ and D- pins of the usb
4329 port and the regular usb controller gets disabled.
4331 root= [KNL] Root filesystem
4332 See name_to_dev_t comment in init/do_mounts.c.
4334 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4335 mount the root filesystem
4337 rootflags= [KNL] Set root filesystem mount option string
4339 rootfstype= [KNL] Set root filesystem type
4341 rootwait [KNL] Wait (indefinitely) for root device to show up.
4342 Useful for devices that are detected asynchronously
4343 (e.g. USB and MMC devices).
4345 rproc_mem=nn[KMG][@address]
4346 [KNL,ARM,CMA] Remoteproc physical memory block.
4347 Memory area to be used by remote processor image,
4350 rw [KNL] Mount root device read-write on boot
4352 S [KNL] Run init in single mode
4354 s390_iommu= [HW,S390]
4355 Set s390 IOTLB flushing mode
4357 With strict flushing every unmap operation will result in
4358 an IOTLB flush. Default is lazy flushing before reuse,
4362 See drivers/net/irda/sa1100_ir.c.
4364 sbni= [NET] Granch SBNI12 leased line adapter
4366 sched_debug [KNL] Enables verbose scheduler debug messages.
4368 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4369 Allowed values are enable and disable. This feature
4370 incurs a small amount of overhead in the scheduler
4371 but is useful for debugging and performance tuning.
4373 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4374 xtime_lock contention on larger systems, and/or RCU lock
4375 contention on all systems with CONFIG_MAXSMP set.
4376 Format: { "0" | "1" }
4377 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4379 Note: increases power consumption, thus should only be
4380 enabled if running jitter sensitive (HPC/RT) workloads.
4382 security= [SECURITY] Choose a legacy "major" security module to
4383 enable at boot. This has been deprecated by the
4386 selinux= [SELINUX] Disable or enable SELinux at boot time.
4387 Format: { "0" | "1" }
4388 See security/selinux/Kconfig help text.
4393 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4394 Format: { "0" | "1" }
4395 See security/apparmor/Kconfig help text
4398 Default value is set via kernel config option.
4400 serialnumber [BUGS=X86-32]
4403 Maximal number of shapers.
4411 Disable merging of slabs with similar size. May be
4412 necessary if there is some reason to distinguish
4413 allocs to different slabs, especially in hardened
4414 environments where the risk of heap overflows and
4415 layout control by attackers can usually be
4416 frustrated by disabling merging. This will reduce
4417 most of the exposure of a heap attack to a single
4418 cache (risks via metadata attacks are mostly
4419 unchanged). Debug options disable merging on their
4421 For more information see Documentation/vm/slub.rst.
4423 slab_max_order= [MM, SLAB]
4424 Determines the maximum allowed order for slabs.
4425 A high setting may cause OOMs due to memory
4426 fragmentation. Defaults to 1 for systems with
4427 more than 32MB of RAM, 0 otherwise.
4429 slub_debug[=options[,slabs]] [MM, SLUB]
4430 Enabling slub_debug allows one to determine the
4431 culprit if slab objects become corrupted. Enabling
4432 slub_debug can create guard zones around objects and
4433 may poison objects when not in use. Also tracks the
4434 last alloc / free. For more information see
4435 Documentation/vm/slub.rst.
4437 slub_memcg_sysfs= [MM, SLUB]
4438 Determines whether to enable sysfs directories for
4439 memory cgroup sub-caches. 1 to enable, 0 to disable.
4440 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4441 Enabling this can lead to a very high number of debug
4442 directories and files being created under
4445 slub_max_order= [MM, SLUB]
4446 Determines the maximum allowed order for slabs.
4447 A high setting may cause OOMs due to memory
4448 fragmentation. For more information see
4449 Documentation/vm/slub.rst.
4451 slub_min_objects= [MM, SLUB]
4452 The minimum number of objects per slab. SLUB will
4453 increase the slab order up to slub_max_order to
4454 generate a sufficiently large slab able to contain
4455 the number of objects indicated. The higher the number
4456 of objects the smaller the overhead of tracking slabs
4457 and the less frequently locks need to be acquired.
4458 For more information see Documentation/vm/slub.rst.
4460 slub_min_order= [MM, SLUB]
4461 Determines the minimum page order for slabs. Must be
4462 lower than slub_max_order.
4463 For more information see Documentation/vm/slub.rst.
4465 slub_nomerge [MM, SLUB]
4466 Same with slab_nomerge. This is supported for legacy.
4467 See slab_nomerge for more information.
4470 Format: <io1>[,<io2>[,...,<io8>]]
4472 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4473 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4474 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4475 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4476 smsc-ircc2.ircc_irq= [HW] IRQ line
4477 smsc-ircc2.ircc_dma= [HW] DMA channel
4478 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4479 0: Toshiba Satellite 1800 (GP data pin select)
4480 1: Fast pin select (default)
4483 smt [KNL,S390] Set the maximum number of threads (logical
4484 CPUs) to use per physical CPU on systems capable of
4485 symmetric multithreading (SMT). Will be capped to the
4486 actual hardware limit.
4488 Default: -1 (no limit)
4491 [KNL] Should the soft-lockup detector generate panics.
4494 A nonzero value instructs the soft-lockup detector
4495 to panic the machine when a soft-lockup occurs. This
4496 is also controlled by CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC
4497 which is the respective build-time switch to that
4500 softlockup_all_cpu_backtrace=
4501 [KNL] Should the soft-lockup detector generate
4502 backtraces on all cpus.
4505 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4506 See Documentation/admin-guide/laptops/sonypi.rst
4508 spectre_v2= [X86] Control mitigation of Spectre variant 2
4509 (indirect branch speculation) vulnerability.
4510 The default operation protects the kernel from
4513 on - unconditionally enable, implies
4515 off - unconditionally disable, implies
4517 auto - kernel detects whether your CPU model is
4520 Selecting 'on' will, and 'auto' may, choose a
4521 mitigation method at run time according to the
4522 CPU, the available microcode, the setting of the
4523 CONFIG_RETPOLINE configuration option, and the
4524 compiler with which the kernel was built.
4526 Selecting 'on' will also enable the mitigation
4527 against user space to user space task attacks.
4529 Selecting 'off' will disable both the kernel and
4530 the user space protections.
4532 Specific mitigations can also be selected manually:
4534 retpoline - replace indirect branches
4535 retpoline,generic - google's original retpoline
4536 retpoline,amd - AMD-specific minimal thunk
4538 Not specifying this option is equivalent to
4542 [X86] Control mitigation of Spectre variant 2
4543 (indirect branch speculation) vulnerability between
4546 on - Unconditionally enable mitigations. Is
4547 enforced by spectre_v2=on
4549 off - Unconditionally disable mitigations. Is
4550 enforced by spectre_v2=off
4552 prctl - Indirect branch speculation is enabled,
4553 but mitigation can be enabled via prctl
4554 per thread. The mitigation control state
4555 is inherited on fork.
4558 - Like "prctl" above, but only STIBP is
4559 controlled per thread. IBPB is issued
4560 always when switching between different user
4564 - Same as "prctl" above, but all seccomp
4565 threads will enable the mitigation unless
4566 they explicitly opt out.
4569 - Like "seccomp" above, but only STIBP is
4570 controlled per thread. IBPB is issued
4571 always when switching between different
4572 user space processes.
4574 auto - Kernel selects the mitigation depending on
4575 the available CPU features and vulnerability.
4578 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4580 Not specifying this option is equivalent to
4581 spectre_v2_user=auto.
4583 spec_store_bypass_disable=
4584 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4585 (Speculative Store Bypass vulnerability)
4587 Certain CPUs are vulnerable to an exploit against a
4588 a common industry wide performance optimization known
4589 as "Speculative Store Bypass" in which recent stores
4590 to the same memory location may not be observed by
4591 later loads during speculative execution. The idea
4592 is that such stores are unlikely and that they can
4593 be detected prior to instruction retirement at the
4594 end of a particular speculation execution window.
4596 In vulnerable processors, the speculatively forwarded
4597 store can be used in a cache side channel attack, for
4598 example to read memory to which the attacker does not
4599 directly have access (e.g. inside sandboxed code).
4601 This parameter controls whether the Speculative Store
4602 Bypass optimization is used.
4604 On x86 the options are:
4606 on - Unconditionally disable Speculative Store Bypass
4607 off - Unconditionally enable Speculative Store Bypass
4608 auto - Kernel detects whether the CPU model contains an
4609 implementation of Speculative Store Bypass and
4610 picks the most appropriate mitigation. If the
4611 CPU is not vulnerable, "off" is selected. If the
4612 CPU is vulnerable the default mitigation is
4613 architecture and Kconfig dependent. See below.
4614 prctl - Control Speculative Store Bypass per thread
4615 via prctl. Speculative Store Bypass is enabled
4616 for a process by default. The state of the control
4617 is inherited on fork.
4618 seccomp - Same as "prctl" above, but all seccomp threads
4619 will disable SSB unless they explicitly opt out.
4621 Default mitigations:
4622 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4624 On powerpc the options are:
4626 on,auto - On Power8 and Power9 insert a store-forwarding
4627 barrier on kernel entry and exit. On Power7
4628 perform a software flush on kernel entry and
4632 Not specifying this option is equivalent to
4633 spec_store_bypass_disable=auto.
4635 spia_io_base= [HW,MTD]
4640 srcutree.counter_wrap_check [KNL]
4641 Specifies how frequently to check for
4642 grace-period sequence counter wrap for the
4643 srcu_data structure's ->srcu_gp_seq_needed field.
4644 The greater the number of bits set in this kernel
4645 parameter, the less frequently counter wrap will
4646 be checked for. Note that the bottom two bits
4649 srcutree.exp_holdoff [KNL]
4650 Specifies how many nanoseconds must elapse
4651 since the end of the last SRCU grace period for
4652 a given srcu_struct until the next normal SRCU
4653 grace period will be considered for automatic
4654 expediting. Set to zero to disable automatic
4658 Speculative Store Bypass Disable control
4660 On CPUs that are vulnerable to the Speculative
4661 Store Bypass vulnerability and offer a
4662 firmware based mitigation, this parameter
4663 indicates how the mitigation should be used:
4665 force-on: Unconditionally enable mitigation for
4666 for both kernel and userspace
4667 force-off: Unconditionally disable mitigation for
4668 for both kernel and userspace
4669 kernel: Always enable mitigation in the
4670 kernel, and offer a prctl interface
4671 to allow userspace to register its
4672 interest in being mitigated too.
4674 stack_guard_gap= [MM]
4675 override the default stack gap protection. The value
4676 is in page units and it defines how many pages prior
4677 to (for stacks growing down) resp. after (for stacks
4678 growing up) the main stack are reserved for no other
4679 mapping. Default value is 256 pages.
4682 Enabled the stack tracer on boot up.
4684 stacktrace_filter=[function-list]
4685 [FTRACE] Limit the functions that the stack tracer
4686 will trace at boot up. function-list is a comma separated
4687 list of functions. This list can be changed at run
4688 time by the stack_trace_filter file in the debugfs
4689 tracing directory. Note, this enables stack tracing
4690 and the stacktrace above is not needed.
4694 Set the STI (builtin display/keyboard on the HP-PARISC
4695 machines) console (graphic card) which should be used
4696 as the initial boot-console.
4697 See also comment in drivers/video/console/sticore.c.
4700 See comment in drivers/video/console/sticore.c.
4703 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4705 sunrpc.min_resvport=
4706 sunrpc.max_resvport=
4708 SunRPC servers often require that client requests
4709 originate from a privileged port (i.e. a port in the
4710 range 0 < portnr < 1024).
4711 An administrator who wishes to reserve some of these
4712 ports for other uses may adjust the range that the
4713 kernel's sunrpc client considers to be privileged
4714 using these two parameters to set the minimum and
4715 maximum port values.
4717 sunrpc.svc_rpc_per_connection_limit=
4719 Limit the number of requests that the server will
4720 process in parallel from a single connection.
4721 The default value is 0 (no limit).
4725 Control how the NFS server code allocates CPUs to
4726 service thread pools. Depending on how many NICs
4727 you have and where their interrupts are bound, this
4728 option will affect which CPUs will do NFS serving.
4729 Note: this parameter cannot be changed while the
4730 NFS server is running.
4732 auto the server chooses an appropriate mode
4733 automatically using heuristics
4734 global a single global pool contains all CPUs
4735 percpu one pool for each CPU
4736 pernode one pool for each NUMA node (equivalent
4737 to global on non-NUMA machines)
4739 sunrpc.tcp_slot_table_entries=
4740 sunrpc.udp_slot_table_entries=
4742 Sets the upper limit on the number of simultaneous
4743 RPC calls that can be sent from the client to a
4744 server. Increasing these values may allow you to
4745 improve throughput, but will also increase the
4746 amount of memory reserved for use by the client.
4748 suspend.pm_test_delay=
4750 Sets the number of seconds to remain in a suspend test
4751 mode before resuming the system (see
4752 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4753 is set. Default value is 5.
4756 Format: { on | off | y | n | 1 | 0 }
4757 This parameter controls use of the Protected
4758 Execution Facility on pSeries.
4761 [KNL] Enable accounting of swap in memory resource
4762 controller if no parameter or 1 is given or disable
4763 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
4765 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4766 Format: { <int> | force | noforce }
4767 <int> -- Number of I/O TLB slabs
4768 force -- force using of bounce buffers even if they
4769 wouldn't be automatically used by the kernel
4770 noforce -- Never use bounce buffers (for debugging)
4774 sysfs.deprecated=0|1 [KNL]
4775 Enable/disable old style sysfs layout for old udev
4776 on older distributions. When this option is enabled
4777 very new udev will not work anymore. When this option
4778 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4779 in older udev will not work anymore.
4780 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4781 the kernel configuration.
4783 sysrq_always_enabled
4785 Ignore sysrq setting - this boot parameter will
4786 neutralize any effect of /proc/sys/kernel/sysrq.
4787 Useful for debugging.
4789 tcpmhash_entries= [KNL,NET]
4790 Set the number of tcp_metrics_hash slots.
4791 Default value is 8192 or 16384 depending on total
4792 ram pages. This is used to specify the TCP metrics
4793 cache size. See Documentation/networking/ip-sysctl.txt
4794 "tcp_no_metrics_save" section for more details.
4798 test_suspend= [SUSPEND][,N]
4799 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4800 standby suspend) or "freeze" (for suspend type freeze)
4801 as the system sleep state during system startup with
4802 the optional capability to repeat N number of times.
4803 The system is woken from this state using a
4804 wakeup-capable RTC alarm.
4806 thash_entries= [KNL,NET]
4807 Set number of hash buckets for TCP connection
4809 thermal.act= [HW,ACPI]
4810 -1: disable all active trip points in all thermal zones
4811 <degrees C>: override all lowest active trip points
4813 thermal.crt= [HW,ACPI]
4814 -1: disable all critical trip points in all thermal zones
4815 <degrees C>: override all critical trip points
4817 thermal.nocrt= [HW,ACPI]
4818 Set to disable actions on ACPI thermal zone
4819 critical and hot trip points.
4821 thermal.off= [HW,ACPI]
4822 1: disable ACPI thermal control
4824 thermal.psv= [HW,ACPI]
4825 -1: disable all passive trip points
4826 <degrees C>: override all passive trip points to this
4829 thermal.tzp= [HW,ACPI]
4830 Specify global default ACPI thermal zone polling rate
4831 <deci-seconds>: poll all this frequency
4832 0: no polling (default)
4835 Force threading of all interrupt handlers except those
4836 marked explicitly IRQF_NO_THREAD.
4840 Specify if the kernel should make use of the cpu
4841 topology information if the hardware supports this.
4842 The scheduler will make use of this information and
4843 e.g. base its process migration decisions on it.
4846 topology_updates= [KNL, PPC, NUMA]
4848 Specify if the kernel should ignore (off)
4849 topology updates sent by the hypervisor to this
4854 tpm_suspend_pcr=[HW,TPM]
4855 Format: integer pcr id
4856 Specify that at suspend time, the tpm driver
4857 should extend the specified pcr with zeros,
4858 as a workaround for some chips which fail to
4859 flush the last written pcr on TPM_SaveState.
4860 This will guarantee that all the other pcrs
4863 trace_buf_size=nn[KMG]
4864 [FTRACE] will set tracing buffer size on each cpu.
4866 trace_event=[event-list]
4867 [FTRACE] Set and start specified trace events in order
4868 to facilitate early boot debugging. The event-list is a
4869 comma separated list of trace events to enable. See
4870 also Documentation/trace/events.rst
4872 trace_options=[option-list]
4873 [FTRACE] Enable or disable tracer options at boot.
4874 The option-list is a comma delimited list of options
4875 that can be enabled or disabled just as if you were
4876 to echo the option name into
4878 /sys/kernel/debug/tracing/trace_options
4880 For example, to enable stacktrace option (to dump the
4881 stack trace of each event), add to the command line:
4883 trace_options=stacktrace
4885 See also Documentation/trace/ftrace.rst "trace options"
4889 Have the tracepoints sent to printk as well as the
4890 tracing ring buffer. This is useful for early boot up
4891 where the system hangs or reboots and does not give the
4892 option for reading the tracing buffer or performing a
4893 ftrace_dump_on_oops.
4895 To turn off having tracepoints sent to printk,
4896 echo 0 > /proc/sys/kernel/tracepoint_printk
4897 Note, echoing 1 into this file without the
4898 tracepoint_printk kernel cmdline option has no effect.
4902 Having tracepoints sent to printk() and activating high
4903 frequency tracepoints such as irq or sched, can cause
4904 the system to live lock.
4907 [FTRACE] enable this option to disable tracing when a
4908 warning is hit. This turns off "tracing_on". Tracing can
4909 be enabled again by echoing '1' into the "tracing_on"
4910 file located in /sys/kernel/debug/tracing/
4912 This option is useful, as it disables the trace before
4913 the WARNING dump is called, which prevents the trace to
4914 be filled with content caused by the warning output.
4916 This option can also be set at run time via the sysctl
4917 option: kernel/traceoff_on_warning
4919 transparent_hugepage=
4921 Format: [always|madvise|never]
4922 Can be used to control the default behavior of the system
4923 with respect to transparent hugepages.
4924 See Documentation/admin-guide/mm/transhuge.rst
4927 tsc= Disable clocksource stability checks for TSC.
4929 [x86] reliable: mark tsc clocksource as reliable, this
4930 disables clocksource verification at runtime, as well
4931 as the stability checks done at bootup. Used to enable
4932 high-resolution timer mode on older hardware, and in
4933 virtualized environment.
4934 [x86] noirqtime: Do not use TSC to do irq accounting.
4935 Used to run time disable IRQ_TIME_ACCOUNTING on any
4936 platforms where RDTSC is slow and this accounting
4938 [x86] unstable: mark the TSC clocksource as unstable, this
4939 marks the TSC unconditionally unstable at bootup and
4940 avoids any further wobbles once the TSC watchdog notices.
4941 [x86] nowatchdog: disable clocksource watchdog. Used
4942 in situations with strict latency requirements (where
4943 interruptions from clocksource watchdog are not
4946 tsx= [X86] Control Transactional Synchronization
4947 Extensions (TSX) feature in Intel processors that
4948 support TSX control.
4950 This parameter controls the TSX feature. The options are:
4952 on - Enable TSX on the system. Although there are
4953 mitigations for all known security vulnerabilities,
4954 TSX has been known to be an accelerator for
4955 several previous speculation-related CVEs, and
4956 so there may be unknown security risks associated
4957 with leaving it enabled.
4959 off - Disable TSX on the system. (Note that this
4960 option takes effect only on newer CPUs which are
4961 not vulnerable to MDS, i.e., have
4962 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
4963 the new IA32_TSX_CTRL MSR through a microcode
4964 update. This new MSR allows for the reliable
4965 deactivation of the TSX functionality.)
4967 auto - Disable TSX if X86_BUG_TAA is present,
4968 otherwise enable TSX on the system.
4970 Not specifying this option is equivalent to tsx=off.
4972 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
4975 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
4976 Abort (TAA) vulnerability.
4978 Similar to Micro-architectural Data Sampling (MDS)
4979 certain CPUs that support Transactional
4980 Synchronization Extensions (TSX) are vulnerable to an
4981 exploit against CPU internal buffers which can forward
4982 information to a disclosure gadget under certain
4985 In vulnerable processors, the speculatively forwarded
4986 data can be used in a cache side channel attack, to
4987 access data to which the attacker does not have direct
4990 This parameter controls the TAA mitigation. The
4993 full - Enable TAA mitigation on vulnerable CPUs
4996 full,nosmt - Enable TAA mitigation and disable SMT on
4997 vulnerable CPUs. If TSX is disabled, SMT
4998 is not disabled because CPU is not
4999 vulnerable to cross-thread TAA attacks.
5000 off - Unconditionally disable TAA mitigation
5002 On MDS-affected machines, tsx_async_abort=off can be
5003 prevented by an active MDS mitigation as both vulnerabilities
5004 are mitigated with the same mechanism so in order to disable
5005 this mitigation, you need to specify mds=off too.
5007 Not specifying this option is equivalent to
5008 tsx_async_abort=full. On CPUs which are MDS affected
5009 and deploy MDS mitigation, TAA mitigation is not
5010 required and doesn't provide any additional
5014 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5016 turbografx.map[2|3]= [HW,JOY]
5017 TurboGraFX parallel port interface
5019 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5020 See also Documentation/input/devices/joystick-parport.rst
5022 udbg-immortal [PPC] When debugging early kernel crashes that
5023 happen after console_init() and before a proper
5024 console driver takes over, this boot options might
5025 help "seeing" what's going on.
5027 uhash_entries= [KNL,NET]
5028 Set number of hash buckets for UDP/UDP-Lite connections
5031 [USB] Ignore overcurrent events (default N).
5032 Some badly-designed motherboards generate lots of
5033 bogus events, for ports that aren't wired to
5034 anything. Set this parameter to avoid log spamming.
5035 Note that genuine overcurrent events won't be
5039 [X86] Cause panic on unknown NMI.
5041 usbcore.authorized_default=
5042 [USB] Default USB device authorization:
5043 (default -1 = authorized except for wireless USB,
5044 0 = not authorized, 1 = authorized, 2 = authorized
5045 if device connected to internal port)
5047 usbcore.autosuspend=
5048 [USB] The autosuspend time delay (in seconds) used
5049 for newly-detected USB devices (default 2). This
5050 is the time required before an idle device will be
5051 autosuspended. Devices for which the delay is set
5052 to a negative value won't be autosuspended at all.
5054 usbcore.usbfs_snoop=
5055 [USB] Set to log all usbfs traffic (default 0 = off).
5057 usbcore.usbfs_snoop_max=
5058 [USB] Maximum number of bytes to snoop in each URB
5061 usbcore.blinkenlights=
5062 [USB] Set to cycle leds on hubs (default 0 = off).
5064 usbcore.old_scheme_first=
5065 [USB] Start with the old device initialization
5066 scheme, applies only to low and full-speed devices
5069 usbcore.usbfs_memory_mb=
5070 [USB] Memory limit (in MB) for buffers allocated by
5071 usbfs (default = 16, 0 = max = 2047).
5073 usbcore.use_both_schemes=
5074 [USB] Try the other device initialization scheme
5075 if the first one fails (default 1 = enabled).
5077 usbcore.initial_descriptor_timeout=
5078 [USB] Specifies timeout for the initial 64-byte
5079 USB_REQ_GET_DESCRIPTOR request in milliseconds
5080 (default 5000 = 5.0 seconds).
5082 usbcore.nousb [USB] Disable the USB subsystem
5085 [USB] A list of quirk entries to augment the built-in
5086 usb core quirk list. List entries are separated by
5087 commas. Each entry has the form
5088 VendorID:ProductID:Flags. The IDs are 4-digit hex
5089 numbers and Flags is a set of letters. Each letter
5090 will change the built-in quirk; setting it if it is
5091 clear and clearing it if it is set. The letters have
5092 the following meanings:
5093 a = USB_QUIRK_STRING_FETCH_255 (string
5094 descriptors must not be fetched using
5096 b = USB_QUIRK_RESET_RESUME (device can't resume
5097 correctly so reset it instead);
5098 c = USB_QUIRK_NO_SET_INTF (device can't handle
5099 Set-Interface requests);
5100 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5101 handle its Configuration or Interface
5103 e = USB_QUIRK_RESET (device can't be reset
5104 (e.g morph devices), don't use reset);
5105 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5106 more interface descriptions than the
5107 bNumInterfaces count, and can't handle
5108 talking to these interfaces);
5109 g = USB_QUIRK_DELAY_INIT (device needs a pause
5110 during initialization, after we read
5111 the device descriptor);
5112 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5113 high speed and super speed interrupt
5114 endpoints, the USB 2.0 and USB 3.0 spec
5115 require the interval in microframes (1
5116 microframe = 125 microseconds) to be
5117 calculated as interval = 2 ^
5119 Devices with this quirk report their
5120 bInterval as the result of this
5121 calculation instead of the exponent
5122 variable used in the calculation);
5123 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5124 handle device_qualifier descriptor
5126 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5127 generates spurious wakeup, ignore
5128 remote wakeup capability);
5129 k = USB_QUIRK_NO_LPM (device can't handle Link
5131 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5132 (Device reports its bInterval as linear
5133 frames instead of the USB 2.0
5135 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5136 to be disconnected before suspend to
5137 prevent spurious wakeup);
5138 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5139 pause after every control message);
5140 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5141 delay after resetting its port);
5142 Example: quirks=0781:5580:bk,0a5c:5834:gij
5145 [USBHID] The interval which mice are to be polled at.
5148 [USBHID] The interval which joysticks are to be polled at.
5151 [USBHID] The interval which keyboards are to be polled at.
5153 usb-storage.delay_use=
5154 [UMS] The delay in seconds before a new device is
5155 scanned for Logical Units (default 1).
5158 [UMS] A list of quirks entries to supplement or
5159 override the built-in unusual_devs list. List
5160 entries are separated by commas. Each entry has
5161 the form VID:PID:Flags where VID and PID are Vendor
5162 and Product ID values (4-digit hex numbers) and
5163 Flags is a set of characters, each corresponding
5164 to a common usb-storage quirk flag as follows:
5165 a = SANE_SENSE (collect more than 18 bytes
5166 of sense data, not on uas);
5167 b = BAD_SENSE (don't collect more than 18
5168 bytes of sense data, not on uas);
5169 c = FIX_CAPACITY (decrease the reported
5170 device capacity by one sector);
5171 d = NO_READ_DISC_INFO (don't use
5172 READ_DISC_INFO command, not on uas);
5173 e = NO_READ_CAPACITY_16 (don't use
5174 READ_CAPACITY_16 command);
5175 f = NO_REPORT_OPCODES (don't use report opcodes
5177 g = MAX_SECTORS_240 (don't transfer more than
5178 240 sectors at a time, uas only);
5179 h = CAPACITY_HEURISTICS (decrease the
5180 reported device capacity by one
5181 sector if the number is odd);
5182 i = IGNORE_DEVICE (don't bind to this
5184 j = NO_REPORT_LUNS (don't use report luns
5186 l = NOT_LOCKABLE (don't try to lock and
5187 unlock ejectable media, not on uas);
5188 m = MAX_SECTORS_64 (don't transfer more
5189 than 64 sectors = 32 KB at a time,
5191 n = INITIAL_READ10 (force a retry of the
5192 initial READ(10) command, not on uas);
5193 o = CAPACITY_OK (accept the capacity
5194 reported by the device, not on uas);
5195 p = WRITE_CACHE (the device cache is ON
5196 by default, not on uas);
5197 r = IGNORE_RESIDUE (the device reports
5198 bogus residue values, not on uas);
5199 s = SINGLE_LUN (the device has only one
5201 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5202 commands, uas only);
5203 u = IGNORE_UAS (don't bind to the uas driver);
5204 w = NO_WP_DETECT (don't test whether the
5205 medium is write-protected).
5206 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5207 even if the device claims no cache,
5209 Example: quirks=0419:aaf5:rl,0421:0433:rc
5211 user_debug= [KNL,ARM]
5213 See arch/arm/Kconfig.debug help text.
5214 1 - undefined instruction events
5216 4 - invalid data aborts
5219 Example: user_debug=31
5222 [X86] Flags controlling user PTE allocations.
5224 nohigh = do not allocate PTE pages in
5225 HIGHMEM regardless of setting
5229 On X86_32, this is an alias for vdso32=. Otherwise:
5231 vdso=1: enable VDSO (the default)
5232 vdso=0: disable VDSO mapping
5234 vdso32= [X86] Control the 32-bit vDSO
5235 vdso32=1: enable 32-bit VDSO
5236 vdso32=0 or vdso32=2: disable 32-bit VDSO
5238 See the help text for CONFIG_COMPAT_VDSO for more
5239 details. If CONFIG_COMPAT_VDSO is set, the default is
5240 vdso32=0; otherwise, the default is vdso32=1.
5242 For compatibility with older kernels, vdso32=2 is an
5245 Try vdso32=0 if you encounter an error that says:
5246 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5249 vector=percpu: enable percpu vector domain
5251 video= [FB] Frame buffer configuration
5252 See Documentation/fb/modedb.rst.
5254 video.brightness_switch_enabled= [0,1]
5255 If set to 1, on receiving an ACPI notify event
5256 generated by hotkey, video driver will adjust brightness
5257 level and then send out the event to user space through
5258 the allocated input device; If set to 0, video driver
5259 will only send out the event without touching backlight
5264 [VMMIO] Memory mapped virtio (platform) device.
5266 <size>@<baseaddr>:<irq>[:<id>]
5268 <size> := size (can use standard suffixes
5270 <baseaddr> := physical base address
5271 <irq> := interrupt number (as passed to
5273 <id> := (optional) platform device id
5275 virtio_mmio.device=1K@0x100b0000:48:7
5277 Can be used multiple times for multiple devices.
5279 vga= [BOOT,X86-32] Select a particular video mode
5280 See Documentation/x86/boot.rst and
5281 Documentation/admin-guide/svga.rst.
5282 Use vga=ask for menu.
5283 This is actually a boot loader parameter; the value is
5284 passed to the kernel using a special protocol.
5286 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
5287 May slow down system boot speed, especially when
5288 enabled on systems with a large amount of memory.
5289 All options are enabled by default, and this
5290 interface is meant to allow for selectively
5291 enabling or disabling specific virtual memory
5294 Available options are:
5295 P Enable page structure init time poisoning
5296 - Disable all of the above options
5298 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5299 size of <nn>. This can be used to increase the
5300 minimum size (128MB on x86). It can also be used to
5301 decrease the size and leave more room for directly
5304 vmcp_cma=nn[MG] [KNL,S390]
5305 Sets the memory size reserved for contiguous memory
5306 allocations for the vmcp device driver.
5308 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5311 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5314 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5318 Controls the behavior of vsyscalls (i.e. calls to
5319 fixed addresses of 0xffffffffff600x00 from legacy
5320 code). Most statically-linked binaries and older
5321 versions of glibc use these calls. Because these
5322 functions are at fixed addresses, they make nice
5323 targets for exploits that can control RIP.
5325 emulate [default] Vsyscalls turn into traps and are
5326 emulated reasonably safely. The vsyscall
5329 xonly Vsyscalls turn into traps and are
5330 emulated reasonably safely. The vsyscall
5331 page is not readable.
5333 none Vsyscalls don't work at all. This makes
5334 them quite hard to use for exploits but
5335 might break your system.
5337 vt.color= [VT] Default text color.
5338 Format: 0xYX, X = foreground, Y = background.
5339 Default: 0x07 = light gray on black.
5341 vt.cur_default= [VT] Default cursor shape.
5342 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5343 the parameters of the <Esc>[?A;B;Cc escape sequence;
5344 see VGA-softcursor.txt. Default: 2 = underline.
5346 vt.default_blu= [VT]
5347 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5348 Change the default blue palette of the console.
5349 This is a 16-member array composed of values
5352 vt.default_grn= [VT]
5353 Format: <green0>,<green1>,<green2>,...,<green15>
5354 Change the default green palette of the console.
5355 This is a 16-member array composed of values
5358 vt.default_red= [VT]
5359 Format: <red0>,<red1>,<red2>,...,<red15>
5360 Change the default red palette of the console.
5361 This is a 16-member array composed of values
5367 Set system-wide default UTF-8 mode for all tty's.
5368 Default is 1, i.e. UTF-8 mode is enabled for all
5369 newly opened terminals.
5371 vt.global_cursor_default=
5374 Set system-wide default for whether a cursor
5375 is shown on new VTs. Default is -1,
5376 i.e. cursors will be created by default unless
5377 overridden by individual drivers. 0 will hide
5378 cursors, 1 will display them.
5380 vt.italic= [VT] Default color for italic text; 0-15.
5383 vt.underline= [VT] Default color for underlined text; 0-15.
5386 watchdog timers [HW,WDT] For information on watchdog timers,
5387 see Documentation/watchdog/watchdog-parameters.rst
5388 or other driver-specific files in the
5389 Documentation/watchdog/ directory.
5393 Set the hard lockup detector stall duration
5394 threshold in seconds. The soft lockup detector
5395 threshold is set to twice the value. A value of 0
5396 disables both lockup detectors. Default is 10
5399 workqueue.watchdog_thresh=
5400 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5401 warn stall conditions and dump internal state to
5402 help debugging. 0 disables workqueue stall
5403 detection; otherwise, it's the stall threshold
5404 duration in seconds. The default value is 30 and
5405 it can be updated at runtime by writing to the
5406 corresponding sysfs file.
5408 workqueue.disable_numa
5409 By default, all work items queued to unbound
5410 workqueues are affine to the NUMA nodes they're
5411 issued on, which results in better behavior in
5412 general. If NUMA affinity needs to be disabled for
5413 whatever reason, this option can be used. Note
5414 that this also can be controlled per-workqueue for
5415 workqueues visible under /sys/bus/workqueue/.
5417 workqueue.power_efficient
5418 Per-cpu workqueues are generally preferred because
5419 they show better performance thanks to cache
5420 locality; unfortunately, per-cpu workqueues tend to
5421 be more power hungry than unbound workqueues.
5423 Enabling this makes the per-cpu workqueues which
5424 were observed to contribute significantly to power
5425 consumption unbound, leading to measurably lower
5426 power usage at the cost of small performance
5429 The default value of this parameter is determined by
5430 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5432 workqueue.debug_force_rr_cpu
5433 Workqueue used to implicitly guarantee that work
5434 items queued without explicit CPU specified are put
5435 on the local CPU. This guarantee is no longer true
5436 and while local CPU is still preferred work items
5437 may be put on foreign CPUs. This debug option
5438 forces round-robin CPU selection to flush out
5439 usages which depend on the now broken guarantee.
5440 When enabled, memory and cache locality will be
5443 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5444 default x2apic cluster mode on platforms
5447 x86_intel_mid_timer= [X86-32,APBT]
5448 Choose timer option for x86 Intel MID platform.
5449 Two valid options are apbt timer only and lapic timer
5450 plus one apbt timer for broadcast timer.
5451 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5453 xen_512gb_limit [KNL,X86-64,XEN]
5454 Restricts the kernel running paravirtualized under Xen
5455 to use only up to 512 GB of RAM. The reason to do so is
5456 crash analysis tools and Xen tools for doing domain
5457 save/restore/migration must be enabled to handle larger
5460 xen_emul_unplug= [HW,X86,XEN]
5461 Unplug Xen emulated devices
5462 Format: [unplug0,][unplug1]
5463 ide-disks -- unplug primary master IDE devices
5464 aux-ide-disks -- unplug non-primary-master IDE devices
5465 nics -- unplug network devices
5466 all -- unplug all emulated devices (NICs and IDE disks)
5467 unnecessary -- unplugging emulated devices is
5468 unnecessary even if the host did not respond to
5470 never -- do not unplug even if version check succeeds
5472 xen_legacy_crash [X86,XEN]
5473 Crash from Xen panic notifier, without executing late
5474 panic() code such as dumping handler.
5476 xen_nopvspin [X86,XEN]
5477 Disables the ticketlock slowpath using Xen PV
5481 Disables the PV optimizations forcing the HVM guest to
5482 run as generic HVM guest with no PV drivers.
5483 This option is obsoleted by the "nopv" option, which
5484 has equivalent effect for XEN platform.
5486 xen_scrub_pages= [XEN]
5487 Boolean option to control scrubbing pages before giving them back
5488 to Xen, for use by other domains. Can be also changed at runtime
5489 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5490 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5492 xen_timer_slop= [X86-64,XEN]
5493 Set the timer slop (in nanoseconds) for the virtual Xen
5494 timers (default is 100000). This adjusts the minimum
5495 delta of virtualized Xen timers, where lower values
5496 improve timer resolution at the expense of processing
5497 more timer interrupts.
5499 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
5500 Disables the PV optimizations forcing the guest to run
5501 as generic guest with no PV drivers. Currently support
5502 XEN HVM, KVM, HYPER_V and VMWARE guest.
5504 xirc2ps_cs= [NET,PCMCIA]
5506 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5509 By default on POWER9 and above, the kernel will
5510 natively use the XIVE interrupt controller. This option
5511 allows the fallback firmware mode to be used:
5513 off Fallback to firmware control of XIVE interrupt
5514 controller on both pseries and powernv
5515 platforms. Only useful on POWER9 and above.
5517 xhci-hcd.quirks [USB,KNL]
5518 A hex value specifying bitmask with supplemental xhci
5519 host controller quirks. Meaning of each bit can be
5520 consulted in header drivers/usb/host/xhci.h.
5523 Format: { early | on | rw | ro | off }
5524 Controls if xmon debugger is enabled. Default is off.
5525 Passing only "xmon" is equivalent to "xmon=early".
5526 early Call xmon as early as possible on boot; xmon
5527 debugger is called from setup_arch().
5528 on xmon debugger hooks will be installed so xmon
5529 is only called on a kernel crash. Default mode,
5530 i.e. either "ro" or "rw" mode, is controlled
5531 with CONFIG_XMON_DEFAULT_RO_MODE.
5532 rw xmon debugger hooks will be installed so xmon
5533 is called only on a kernel crash, mode is write,
5534 meaning SPR registers, memory and, other data
5535 can be written using xmon commands.
5536 ro same as "rw" option above but SPR registers,
5537 memory, and other data can't be written using
5539 off xmon is disabled.