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]
25 { vendor | video | native | none }
26 If set to vendor, prefer vendor-specific driver
27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
28 of the ACPI video.ko driver.
29 If set to video, use the ACPI video.ko driver.
30 If set to native, use the device's native backlight mode.
31 If set to none, disable the ACPI backlight interface.
33 acpi_force_32bit_fadt_addr
34 force FADT to use 32 bit addresses rather than the
35 64 bit X_* addresses. Some firmware have broken 64
36 bit addresses for force ACPI ignore these and use
37 the older legacy 32 bit addresses.
39 acpica_no_return_repair [HW, ACPI]
40 Disable AML predefined validation mechanism
41 This mechanism can repair the evaluation result to make
42 the return objects more ACPI specification compliant.
43 This option is useful for developers to identify the
44 root cause of an AML interpreter issue when the issue
45 has something to do with the repair mechanism.
47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
51 debug output. Bits in debug_layer correspond to a
52 _COMPONENT in an ACPI source file, e.g.,
53 #define _COMPONENT ACPI_PCI_COMPONENT
54 Bits in debug_level correspond to a level in
55 ACPI_DEBUG_PRINT statements, e.g.,
56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
57 The debug_level mask defaults to "info". See
58 Documentation/firmware-guide/acpi/debug.rst for more information about
59 debug layers and levels.
61 Enable processor driver info messages:
62 acpi.debug_layer=0x20000000
63 Enable PCI/PCI interrupt routing info messages:
64 acpi.debug_layer=0x400000
65 Enable AML "Debug" output, i.e., stores to the Debug
66 object while interpreting AML:
67 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
68 Enable all messages related to ACPI hardware:
69 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
71 Some values produce so much output that the system is
72 unusable. The "log_buf_len" parameter may be useful
73 if you need to capture more output.
75 acpi_enforce_resources= [ACPI]
77 Check for resource conflicts between native drivers
78 and ACPI OperationRegions (SystemIO and SystemMemory
79 only). IO ports and memory declared in ACPI might be
80 used by the ACPI subsystem in arbitrary AML code and
81 can interfere with legacy drivers.
82 strict (default): access to resources claimed by ACPI
83 is denied; legacy drivers trying to access reserved
84 resources will fail to bind to device using them.
85 lax: access to resources claimed by ACPI is allowed;
86 legacy drivers trying to access reserved resources
87 will bind successfully but a warning message is logged.
88 no: ACPI OperationRegions are not marked as reserved,
89 no further checks are performed.
91 acpi_force_table_verification [HW,ACPI]
92 Enable table checksum verification during early stage.
93 By default, this is disabled due to x86 early mapping
96 acpi_irq_balance [HW,ACPI]
97 ACPI will balance active IRQs
100 acpi_irq_nobalance [HW,ACPI]
101 ACPI will not move active IRQs (default)
104 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
105 Format: <irq>,<irq>...
107 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
109 Format: <irq>,<irq>...
111 acpi_mask_gpe= [HW,ACPI]
112 Due to the existence of _Lxx/_Exx, some GPEs triggered
113 by unsupported hardware/firmware features can result in
114 GPE floodings that cannot be automatically disabled by
116 This facility can be used to prevent such uncontrolled
120 acpi_no_auto_serialize [HW,ACPI]
121 Disable auto-serialization of AML methods
122 AML control methods that contain the opcodes to create
123 named objects will be marked as "Serialized" by the
124 auto-serialization feature.
125 This feature is enabled by default.
126 This option allows to turn off the feature.
128 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
131 acpi_no_static_ssdt [HW,ACPI]
132 Disable installation of static SSDTs at early boot time
133 By default, SSDTs contained in the RSDT/XSDT will be
134 installed automatically and they will appear under
135 /sys/firmware/acpi/tables.
136 This option turns off this feature.
137 Note that specifying this option does not affect
138 dynamic table installation which will install SSDT
139 tables to /sys/firmware/acpi/tables/dynamic.
141 acpi_no_watchdog [HW,ACPI,WDT]
142 Ignore the ACPI-based watchdog interface (WDAT) and let
143 a native driver control the watchdog device instead.
145 acpi_rsdp= [ACPI,EFI,KEXEC]
146 Pass the RSDP address to the kernel, mostly used
147 on machines running EFI runtime service to boot the
148 second kernel for kdump.
150 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
151 Format: To spoof as Windows 98: ="Microsoft Windows"
153 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
154 of 2 which is mandated by ACPI 6) as the supported ACPI
155 specification revision (when using this switch, it may
156 be necessary to carry out a cold reboot _twice_ in a
157 row to make it take effect on the platform firmware).
159 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
160 acpi_osi="string1" # add string1
161 acpi_osi="!string2" # remove string2
162 acpi_osi=!* # remove all strings
163 acpi_osi=! # disable all built-in OS vendor
165 acpi_osi=!! # enable all built-in OS vendor
167 acpi_osi= # disable all strings
169 'acpi_osi=!' can be used in combination with single or
170 multiple 'acpi_osi="string1"' to support specific OS
171 vendor string(s). Note that such command can only
172 affect the default state of the OS vendor strings, thus
173 it cannot affect the default state of the feature group
174 strings and the current state of the OS vendor strings,
175 specifying it multiple times through kernel command line
176 is meaningless. This command is useful when one do not
177 care about the state of the feature group strings which
178 should be controlled by the OSPM.
180 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
181 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
182 can make '_OSI("Windows 2000")' TRUE.
184 'acpi_osi=' cannot be used in combination with other
185 'acpi_osi=' command lines, the _OSI method will not
186 exist in the ACPI namespace. NOTE that such command can
187 only affect the _OSI support state, thus specifying it
188 multiple times through kernel command line is also
191 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
194 'acpi_osi=!*' can be used in combination with single or
195 multiple 'acpi_osi="string1"' to support specific
196 string(s). Note that such command can affect the
197 current state of both the OS vendor strings and the
198 feature group strings, thus specifying it multiple times
199 through kernel command line is meaningful. But it may
200 still not able to affect the final state of a string if
201 there are quirks related to this string. This command
202 is useful when one want to control the state of the
203 feature group strings to debug BIOS issues related to
206 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
207 '_OSI("Module Device")' FALSE.
208 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
209 '_OSI("Module Device")' TRUE.
210 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
212 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
214 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
215 they all will make '_OSI("Windows 2000")' TRUE.
218 Override the pmtimer bug detection: force the kernel
219 to assume that this machine's pmtimer latches its value
220 and always returns good values.
222 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
223 Format: { level | edge | high | low }
225 acpi_skip_timer_override [HW,ACPI]
226 Recognize and ignore IRQ0/pin2 Interrupt Override.
227 For broken nForce2 BIOS resulting in XT-PIC timer.
229 acpi_sleep= [HW,ACPI] Sleep options
230 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
231 old_ordering, nonvs, sci_force_enable, nobl }
232 See Documentation/power/video.rst for information on
234 s3_beep is for debugging; it makes the PC's speaker beep
235 as soon as the kernel's real-mode entry point is called.
236 s4_nohwsig prevents ACPI hardware signature from being
237 used during resume from hibernation.
238 old_ordering causes the ACPI 1.0 ordering of the _PTS
239 control method, with respect to putting devices into
240 low power states, to be enforced (the ACPI 2.0 ordering
241 of _PTS is used by default).
242 nonvs prevents the kernel from saving/restoring the
243 ACPI NVS memory during suspend/hibernation and resume.
244 sci_force_enable causes the kernel to set SCI_EN directly
245 on resume from S1/S3 (which is against the ACPI spec,
246 but some broken systems don't work without it).
247 nobl causes the internal blacklist of systems known to
248 behave incorrectly in some ways with respect to system
249 suspend and resume to be ignored (use wisely).
251 acpi_use_timer_override [HW,ACPI]
252 Use timer override. For some broken Nvidia NF5 boards
253 that require a timer override, but don't have HPET
255 add_efi_memmap [EFI; X86] Include EFI memory map in
256 kernel's map of available physical RAM.
259 { off | try_unsupported }
260 off: disable AGP support
261 try_unsupported: try to drive unsupported chipsets
262 (may crash computer or cause data corruption)
265 See Documentation/sound/alsa-configuration.rst
268 Allow the default userspace alignment fault handler
269 behaviour to be specified. Bit 0 enables warnings,
270 bit 1 enables fixups, and bit 2 sends a segfault.
272 align_va_addr= [X86-64]
273 Align virtual addresses by clearing slice [14:12] when
274 allocating a VMA at process creation time. This option
275 gives you up to 3% performance improvement on AMD F15h
276 machines (where it is enabled by default) for a
277 CPU-intensive style benchmark, and it can vary highly in
278 a microbenchmark depending on workload and compiler.
280 32: only for 32-bit processes
281 64: only for 64-bit processes
282 on: enable for both 32- and 64-bit processes
283 off: disable for both 32- and 64-bit processes
285 alloc_snapshot [FTRACE]
286 Allocate the ftrace snapshot buffer on boot up when the
287 main buffer is allocated. This is handy if debugging
288 and you need to use tracing_snapshot() on boot up, and
289 do not want to use tracing_snapshot_alloc() as it needs
290 to be done where GFP_KERNEL allocations are allowed.
292 amd_iommu= [HW,X86-64]
293 Pass parameters to the AMD IOMMU driver in the system.
295 fullflush - enable flushing of IO/TLB entries when
296 they are unmapped. Otherwise they are
297 flushed before they will be reused, which
299 off - do not initialize any AMD IOMMU found in
301 force_isolation - Force device isolation for all
302 devices. The IOMMU driver is not
303 allowed anymore to lift isolation
304 requirements as needed. This option
305 does not override iommu=pt
307 amd_iommu_dump= [HW,X86-64]
308 Enable AMD IOMMU driver option to dump the ACPI table
309 for AMD IOMMU. With this option enabled, AMD IOMMU
310 driver will print ACPI tables for AMD IOMMU during
311 IOMMU initialization.
313 amd_iommu_intr= [HW,X86-64]
314 Specifies one of the following AMD IOMMU interrupt
316 legacy - Use legacy interrupt remapping mode.
317 vapic - Use virtual APIC mode, which allows IOMMU
318 to inject interrupts directly into guest.
319 This mode requires kvm-amd.avic=1.
320 (Default when IOMMU HW support is present.)
322 amijoy.map= [HW,JOY] Amiga joystick support
323 Map of devices attached to JOY0DAT and JOY1DAT
325 See also Documentation/input/joydev/joystick.rst
327 analog.map= [HW,JOY] Analog joystick and gamepad support
328 Specifies type or capabilities of an analog joystick
329 connected to one of 16 gameports
330 Format: <type1>,<type2>,..<type16>
333 Power management functions (SPARCstation-4/5 + deriv.)
335 Disable APC CPU standby support. SPARCstation-Fox does
336 not play well with APC CPU idle - disable it if you have
337 APC and your system crashes randomly.
339 apic= [APIC,X86] Advanced Programmable Interrupt Controller
340 Change the output verbosity while booting
341 Format: { quiet (default) | verbose | debug }
342 Change the amount of debugging information output
343 when initialising the APIC and IO-APIC components.
344 For X86-32, this can also be used to specify an APIC
346 Format: apic=driver_name
347 Examples: apic=bigsmp
349 apic_extnmi= [APIC,X86] External NMI delivery setting
350 Format: { bsp (default) | all | none }
351 bsp: External NMI is delivered only to CPU 0
352 all: External NMIs are broadcast to all CPUs as a
354 none: External NMI is masked for all CPUs. This is
355 useful so that a dump capture kernel won't be
359 See Documentation/networking/ipv6.rst.
361 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
362 Limit apic dumping. The parameter defines the maximal
363 number of local apics being dumped. Also it is possible
364 to set it to "all" by meaning -- no limit here.
365 Format: { 1 (default) | 2 | ... | all }.
366 The parameter valid if only apic=debug or
367 apic=verbose is specified.
368 Example: apic=debug show_lapic=all
370 apm= [APM] Advanced Power Management
371 See header of arch/x86/kernel/apm_32.c.
373 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
374 Format: <io>,<irq>,<nodeID>
378 atarimouse= [HW,MOUSE] Atari Mouse
380 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
381 EzKey and similar keyboards
383 atkbd.reset= [HW] Reset keyboard during initialization
385 atkbd.set= [HW] Select keyboard code set
386 Format: <int> (2 = AT (default), 3 = PS/2)
388 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
391 atkbd.softraw= [HW] Choose between synthetic and real raw mode
392 Format: <bool> (0 = real, 1 = synthetic (default))
394 atkbd.softrepeat= [HW]
395 Use software keyboard repeat
397 audit= [KNL] Enable the audit sub-system
398 Format: { "0" | "1" | "off" | "on" }
399 0 | off - kernel audit is disabled and can not be
400 enabled until the next reboot
401 unset - kernel audit is initialized but disabled and
402 will be fully enabled by the userspace auditd.
403 1 | on - kernel audit is initialized and partially
404 enabled, storing at most audit_backlog_limit
405 messages in RAM until it is fully enabled by the
409 audit_backlog_limit= [KNL] Set the audit queue size limit.
410 Format: <int> (must be >=0)
413 bau= [X86_UV] Enable the BAU on SGI UV. The default
414 behavior is to disable the BAU (i.e. bau=0).
415 Format: { "0" | "1" }
418 unset - Disable the BAU.
420 baycom_epp= [HW,AX25]
423 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
425 See header of drivers/net/hamradio/baycom_par.c.
427 baycom_ser_fdx= [HW,AX25]
428 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
429 Format: <io>,<irq>,<mode>[,<baud>]
430 See header of drivers/net/hamradio/baycom_ser_fdx.c.
432 baycom_ser_hdx= [HW,AX25]
433 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
434 Format: <io>,<irq>,<mode>
435 See header of drivers/net/hamradio/baycom_ser_hdx.c.
437 blkdevparts= Manual partition parsing of block device(s) for
438 embedded devices based on command line input.
439 See Documentation/block/cmdline-partition.rst
441 boot_delay= Milliseconds to delay each printk during boot.
442 Values larger than 10 seconds (10000) are changed to
447 Extended command line options can be added to an initrd
448 and this will cause the kernel to look for it.
450 See Documentation/admin-guide/bootconfig.rst
453 Disable BERT OS support on buggy BIOSes.
455 bgrt_disable [ACPI][X86]
456 Disable BGRT to avoid flickering OEM logo.
458 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
459 bttv.radio= Most important insmod options are available as
461 bttv.pll= See Documentation/admin-guide/media/bttv.rst
464 bulk_remove=off [PPC] This parameter disables the use of the pSeries
465 firmware feature for flushing multiple hpte entries
468 c101= [NET] Moxa C101 synchronous serial card
470 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
471 Sometimes CPU hardware bugs make them report the cache
472 size incorrectly. The kernel will attempt work arounds
473 to fix known problems, but for some CPUs it is not
474 possible to determine what the correct size should be.
475 This option provides an override for these situations.
478 [NET] Specifies amount of time (in seconds) that
479 the kernel should wait for a network carrier. By default
480 it waits 120 seconds.
482 ca_keys= [KEYS] This parameter identifies a specific key(s) on
483 the system trusted keyring to be used for certificate
485 format: { id:<keyid> | builtin }
487 cca= [MIPS] Override the kernel pages' cache coherency
488 algorithm. Accepted values range from 0 to 7
489 inclusive. See arch/mips/include/asm/pgtable-bits.h
490 for platform specific values (SB1, Loongson3 and
493 ccw_timeout_log [S390]
494 See Documentation/s390/common_io.rst for details.
496 cgroup_disable= [KNL] Disable a particular controller
497 Format: {name of the controller(s) to disable}
498 The effects of cgroup_disable=foo are:
499 - foo isn't auto-mounted if you mount all cgroups in
501 - foo isn't visible as an individually mountable
503 {Currently only "memory" controller deal with this and
504 cut the overhead, others just disable the usage. So
505 only cgroup_disable=memory is actually worthy}
507 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
508 Format: { { controller | "all" | "named" }
509 [,{ controller | "all" | "named" }...] }
510 Like cgroup_disable, but only applies to cgroup v1;
511 the blacklisted controllers remain available in cgroup2.
512 "all" blacklists all controllers and "named" disables
513 named mounts. Specifying both "all" and "named" disables
516 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
518 nosocket -- Disable socket memory accounting.
519 nokmem -- Disable kernel memory accounting.
521 checkreqprot [SELINUX] Set initial checkreqprot flag value.
522 Format: { "0" | "1" }
523 See security/selinux/Kconfig help text.
524 0 -- check protection applied by kernel (includes
525 any implied execute protection).
526 1 -- check protection requested by application.
527 Default value is set via a kernel config option.
528 Value can be changed at runtime via
529 /sys/fs/selinux/checkreqprot.
530 Setting checkreqprot to 1 is deprecated.
533 See Documentation/s390/common_io.rst for details.
536 Prevents the clock framework from automatically gating
537 clocks that have not been explicitly enabled by a Linux
538 device driver but are enabled in hardware at reset or
539 by the bootloader/firmware. Note that this does not
540 force such clocks to be always-on nor does it reserve
541 those clocks in any way. This parameter is useful for
542 debug and development, but should not be needed on a
543 platform with proper driver support. For more
544 information, see Documentation/driver-api/clk.rst.
546 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
548 Forces specified clocksource (if available) to be used
549 when calculating gettimeofday(). If specified
550 clocksource is not available, it defaults to PIT.
551 Format: { pit | tsc | cyclone | pmtmr }
553 clocksource= Override the default clocksource
555 Override the default clocksource and use the clocksource
556 with the name specified.
557 Some clocksource names to choose from, depending on
559 [all] jiffies (this is the base, fallback clocksource)
561 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
562 pxa_timer,timer3,32k_counter,timer0_1
563 [X86-32] pit,hpet,tsc;
564 scx200_hrt on Geode; cyclone on IBM x440
572 clocksource.arm_arch_timer.evtstrm=
575 Enable/disable the eventstream feature of the ARM
576 architected timer so that code using WFE-based polling
577 loops can be debugged more effectively on production
580 clearcpuid=BITNUM [X86]
581 Disable CPUID feature X for the kernel. See
582 arch/x86/include/asm/cpufeatures.h for the valid bit
583 numbers. Note the Linux specific bits are not necessarily
584 stable over kernel options, but the vendor specific
586 Also note that user programs calling CPUID directly
587 or using the feature without checking anything
588 will still see it. This just prevents it from
589 being used by the kernel or shown in /proc/cpuinfo.
590 Also note the kernel might malfunction if you disable
593 cma=nn[MG]@[start[MG][-end[MG]]]
595 Sets the size of kernel global memory area for
596 contiguous memory allocations and optionally the
597 placement constraint by the physical address range of
598 memory allocations. A value of 0 disables CMA
599 altogether. For more information, see
600 include/linux/dma-contiguous.h
602 cmo_free_hint= [PPC] Format: { yes | no }
603 Specify whether pages are marked as being inactive
604 when they are freed. This is used in CMO environments
605 to determine OS memory pressure for page stealing by
609 coherent_pool=nn[KMG] [ARM,KNL]
610 Sets the size of memory pool for coherent, atomic dma
611 allocations, by default set to 256K.
613 com20020= [HW,NET] ARCnet - COM20020 chipset
615 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
617 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
621 ARCnet - COM90xx chipset (memory-mapped buffers)
622 Format: <io>[,<irq>[,<memstart>]]
624 condev= [HW,S390] console device
627 console= [KNL] Output console device and options.
629 tty<n> Use the virtual console device <n>.
633 Use the specified serial port. The options are of
634 the form "bbbbpnf", where "bbbb" is the baud rate,
635 "p" is parity ("n", "o", or "e"), "n" is number of
636 bits, and "f" is flow control ("r" for RTS or
637 omit it). Default is "9600n8".
639 See Documentation/admin-guide/serial-console.rst for more
641 Documentation/networking/netconsole.rst for an
644 uart[8250],io,<addr>[,options]
645 uart[8250],mmio,<addr>[,options]
646 uart[8250],mmio16,<addr>[,options]
647 uart[8250],mmio32,<addr>[,options]
648 uart[8250],0x<addr>[,options]
649 Start an early, polled-mode console on the 8250/16550
650 UART at the specified I/O port or MMIO address,
651 switching to the matching ttyS device later.
652 MMIO inter-register address stride is either 8-bit
653 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
654 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
655 to be equivalent to 'mmio'. 'options' are specified in
656 the same format described for ttyS above; if unspecified,
657 the h/w is not re-initialized.
659 hvc<n> Use the hypervisor console device <n>. This is for
660 both Xen and PowerPC hypervisors.
662 If the device connected to the port is not a TTY but a braille
663 device, prepend "brl," before the device type, for instance
665 For now, only VisioBraille is supported.
668 [KNL] Change console messages format
670 By default we print messages on consoles in
671 "[time stamp] text\n" format (time stamp may not be
672 printed, depending on CONFIG_PRINTK_TIME or
673 `printk_time' param).
675 Switch to syslog format: "<%u>[time stamp] text\n"
676 IOW, each message will have a facility and loglevel
677 prefix. The format is similar to one used by syslog()
678 syscall, or to executing "dmesg -S --raw" or to reading
681 consoleblank= [KNL] The console blank (screen saver) timeout in
682 seconds. A value of 0 disables the blank timer.
686 [KNL] Change the default value for
687 /proc/<pid>/coredump_filter.
688 See also Documentation/filesystems/proc.rst.
690 coresight_cpu_debug.enable
693 Enable/disable the CPU sampling based debugging.
694 0: default value, disable debugging
695 1: enable debugging at boot time
697 cpuidle.off=1 [CPU_IDLE]
698 disable the cpuidle sub-system
701 [CPU_IDLE] Name of the cpuidle governor to use.
703 cpufreq.off=1 [CPU_FREQ]
704 disable the cpufreq sub-system
707 [X86] Delay for N microsec between assert and de-assert
708 of APIC INIT to start processors. This delay occurs
709 on every CPU online, such as boot, and resume from suspend.
712 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
714 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
716 crashkernel=size[KMG][@offset[KMG]]
717 [KNL] Using kexec, Linux can switch to a 'crash kernel'
718 upon panic. This parameter reserves the physical
719 memory region [offset, offset + size] for that kernel
720 image. If '@offset' is omitted, then a suitable offset
721 is selected automatically.
722 [KNL, x86_64] select a region under 4G first, and
723 fall back to reserve region above 4G when '@offset'
724 hasn't been specified.
725 See Documentation/admin-guide/kdump/kdump.rst for further details.
727 crashkernel=range1:size1[,range2:size2,...][@offset]
728 [KNL] Same as above, but depends on the memory
729 in the running system. The syntax of range is
730 start-[end] where start and end are both
731 a memory unit (amount[KMG]). See also
732 Documentation/admin-guide/kdump/kdump.rst for an example.
734 crashkernel=size[KMG],high
735 [KNL, x86_64] range could be above 4G. Allow kernel
736 to allocate physical memory region from top, so could
737 be above 4G if system have more than 4G ram installed.
738 Otherwise memory region will be allocated below 4G, if
740 It will be ignored if crashkernel=X is specified.
741 crashkernel=size[KMG],low
742 [KNL, x86_64] range under 4G. When crashkernel=X,high
743 is passed, kernel could allocate physical memory region
744 above 4G, that cause second kernel crash on system
745 that require some amount of low memory, e.g. swiotlb
746 requires at least 64M+32K low memory, also enough extra
747 low memory is needed to make sure DMA buffers for 32-bit
748 devices won't run out. Kernel would try to allocate at
749 at least 256M below 4G automatically.
750 This one let user to specify own low range under 4G
751 for second kernel instead.
752 0: to disable low allocation.
753 It will be ignored when crashkernel=X,high is not used
754 or memory reserved is below 4G.
757 [KNL] Disable crypto self-tests
762 cs89x0_media= [HW,NET]
763 Format: { rj45 | aui | bnc }
766 See header of drivers/s390/block/dasd_devmap.c.
768 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
769 (one device per port)
770 Format: <port#>,<type>
771 See also Documentation/input/devices/joystick-parport.rst
773 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
775 Documentation/admin-guide/dynamic-debug-howto.rst for
776 details. Deprecated, see dyndbg.
778 debug [KNL] Enable kernel debugging (events log level).
781 [KNL] Enable printing [hashed] pointers early in the
782 boot sequence. If enabled, we use a weak hash instead
783 of siphash to hash pointers. Use this option if you are
784 seeing instances of '(___ptrval___)') and need to see a
785 value (hashed pointer) instead. Cryptographically
786 insecure, please do not use on production kernels.
789 [KNL] verbose self-tests
791 Print debugging info while doing the locking API
793 We default to 0 (no extra messages), setting it to
794 1 will print _a lot_ more information - normally
795 only useful to kernel developers.
797 debug_objects [KNL] Enable object debugging
800 [KNL] Disable object debugging
802 debug_guardpage_minorder=
803 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
804 parameter allows control of the order of pages that will
805 be intentionally kept free (and hence protected) by the
806 buddy allocator. Bigger value increase the probability
807 of catching random memory corruption, but reduce the
808 amount of memory for normal system use. The maximum
809 possible value is MAX_ORDER/2. Setting this parameter
810 to 1 or 2 should be enough to identify most random
811 memory corruption problems caused by bugs in kernel or
812 driver code when a CPU writes to (or reads from) a
813 random memory location. Note that there exists a class
814 of memory corruptions problems caused by buggy H/W or
815 F/W or by drivers badly programing DMA (basically when
816 memory is written at bus level and the CPU MMU is
817 bypassed) which are not detectable by
818 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
819 tracking down these problems.
822 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
823 enables the feature at boot time. By default, it is
824 disabled and the system will work mostly the same as a
825 kernel built without CONFIG_DEBUG_PAGEALLOC.
826 Note: to get most of debug_pagealloc error reports, it's
827 useful to also enable the page_owner functionality.
828 on: enable the feature
830 debugpat [X86] Enable PAT debugging
832 decnet.addr= [HW,NET]
833 Format: <area>[,<node>]
834 See also Documentation/networking/decnet.rst.
837 [HW] The size of the default HugeTLB page. This is
838 the size represented by the legacy /proc/ hugepages
839 APIs. In addition, this is the default hugetlb size
840 used for shmget(), mmap() and mounting hugetlbfs
841 filesystems. If not specified, defaults to the
842 architecture's default huge page size. Huge page
843 sizes are architecture dependent. See also
844 Documentation/admin-guide/mm/hugetlbpage.rst.
847 deferred_probe_timeout=
848 [KNL] Debugging option to set a timeout in seconds for
849 deferred probe to give up waiting on dependencies to
850 probe. Only specific dependencies (subsystems or
851 drivers) that have opted in will be ignored. A timeout of 0
852 will timeout at the end of initcalls. This option will also
853 dump out devices still on the deferred probe list after
857 Format: { on | off | def_only | inf_only | always }
858 on: s390 zlib hardware support for compression on
859 level 1 and decompression (default)
860 off: No s390 zlib hardware support
861 def_only: s390 zlib hardware support for deflate
862 only (compression on level 1)
863 inf_only: s390 zlib hardware support for inflate
865 always: Same as 'on' but ignores the selected compression
866 level always using hardware support (used for debugging)
869 Set number of hash buckets for dentry cache.
871 disable_1tb_segments [PPC]
872 Disables the use of 1TB hash page table segments. This
873 causes the kernel to fall back to 256MB segments which
874 can be useful when debugging issues that require an SLB
878 Limits the number of kernel SLB entries, and flushes
879 them frequently to increase the rate of SLB faults
883 See Documentation/networking/ipv6.rst.
886 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
887 hardening is enabled for this boot. Hardened
888 usercopy checking is used to protect the kernel
889 from reading or writing beyond known memory
890 allocation boundaries as a proactive defense
891 against bounds-checking flaws in the kernel's
892 copy_to_user()/copy_from_user() interface.
893 on Perform hardened usercopy checks (default).
894 off Disable hardened usercopy checks.
897 Disable RADIX MMU mode on POWER9
900 Disable TLBIE instruction. Currently does not work
901 with KVM, with HASH MMU, or with coherent accelerators.
903 disable_cpu_apicid= [X86,APIC,SMP]
905 The number of initial APIC ID for the
906 corresponding CPU to be disabled at boot,
907 mostly used for the kdump 2nd kernel to
908 disable BSP to wake up multiple CPUs without
909 causing system reset or hang due to sending
912 perf_v4_pmi= [X86,INTEL]
914 Disable Intel PMU counter freezing feature.
915 The feature only exists starting from
916 Arch Perfmon v4 (Skylake and newer).
918 disable_ddw [PPC/PSERIES]
919 Disable Dynamic DMA Window support. Use this if
920 to workaround buggy firmware.
923 See Documentation/networking/ipv6.rst.
925 disable_mtrr_cleanup [X86]
926 The kernel tries to adjust MTRR layout from continuous
927 to discrete, to make X server driver able to add WB
928 entry later. This parameter disables that.
930 disable_mtrr_trim [X86, Intel and AMD only]
931 By default the kernel will trim any uncacheable
932 memory out of your available memory pool based on
933 MTRR settings. This parameter disables that behavior,
934 possibly causing your machine to run very slowly.
936 disable_timer_pin_1 [X86]
937 Disable PIN 1 of APIC timer
938 Can be useful to work around chipset bugs.
940 dis_ucode_ldr [X86] Disable the microcode loader.
942 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
943 this option disables the debugging code at boot.
945 dma_debug_entries=<number>
946 This option allows to tune the number of preallocated
947 entries for DMA-API debugging code. One entry is
948 required per DMA-API allocation. Use this if the
949 DMA-API debugging code disables itself because the
950 architectural default is too low.
952 dma_debug_driver=<driver_name>
953 With this option the DMA-API debugging driver
954 filter feature can be enabled at boot time. Just
955 pass the driver to filter for as the parameter.
956 The filter can be disabled or changed to another
957 driver later using sysfs.
959 driver_async_probe= [KNL]
960 List of driver names to be probed asynchronously.
961 Format: <driver_name1>,<driver_name2>...
963 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
964 Broken monitors, graphic adapters, KVMs and EDIDless
965 panels may send no or incorrect EDID data sets.
966 This parameter allows to specify an EDID data sets
967 in the /lib/firmware directory that are used instead.
968 Generic built-in EDID data sets are used, if one of
969 edid/1024x768.bin, edid/1280x1024.bin,
970 edid/1680x1050.bin, or edid/1920x1080.bin is given
971 and no file with the same name exists. Details and
972 instructions how to build your own EDID data are
973 available in Documentation/admin-guide/edid.rst. An EDID
974 data set will only be used for a particular connector,
975 if its name and a colon are prepended to the EDID
976 name. Each connector may use a unique EDID data
977 set by separating the files with a comma. An EDID
978 data set with no connector name will be used for
979 any connectors not explicitly specified.
984 Format: {"off" | "known"}
985 Control how the dt_cpu_ftrs device-tree binding is
986 used for CPU feature discovery and setup (if it
988 off: Do not use it, fall back to legacy cpu table.
989 known: Do not pass through unknown features to guests
990 or userspace, only those that the kernel is aware of.
992 dump_apple_properties [X86]
993 Dump name and content of EFI device properties on
994 x86 Macs. Useful for driver authors to determine
995 what data is available or for reverse-engineering.
997 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
998 module.dyndbg[="val"]
999 Enable debug messages at boot time. See
1000 Documentation/admin-guide/dynamic-debug-howto.rst
1003 nopku [X86] Disable Memory Protection Keys CPU feature found
1006 module.async_probe [KNL]
1007 Enable asynchronous probe on this module.
1009 early_ioremap_debug [KNL]
1010 Enable debug messages in early_ioremap support. This
1011 is useful for tracking down temporary early mappings
1012 which are not unmapped.
1014 earlycon= [KNL] Output early console device and options.
1016 When used with no options, the early console is
1017 determined by stdout-path property in device tree's
1018 chosen node or the ACPI SPCR table if supported by
1021 cdns,<addr>[,options]
1022 Start an early, polled-mode console on a Cadence
1023 (xuartps) serial port at the specified address. Only
1024 supported option is baud rate. If baud rate is not
1025 specified, the serial port must already be setup and
1028 uart[8250],io,<addr>[,options]
1029 uart[8250],mmio,<addr>[,options]
1030 uart[8250],mmio32,<addr>[,options]
1031 uart[8250],mmio32be,<addr>[,options]
1032 uart[8250],0x<addr>[,options]
1033 Start an early, polled-mode console on the 8250/16550
1034 UART at the specified I/O port or MMIO address.
1035 MMIO inter-register address stride is either 8-bit
1036 (mmio) or 32-bit (mmio32 or mmio32be).
1037 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1038 to be equivalent to 'mmio'. 'options' are specified
1039 in the same format described for "console=ttyS<n>"; if
1040 unspecified, the h/w is not initialized.
1044 Start an early, polled-mode console on a pl011 serial
1045 port at the specified address. The pl011 serial port
1046 must already be setup and configured. Options are not
1047 yet supported. If 'mmio32' is specified, then only
1048 the driver will use only 32-bit accessors to read/write
1049 the device registers.
1052 Start an early, polled-mode console on a meson serial
1053 port at the specified address. The serial port must
1054 already be setup and configured. Options are not yet
1058 Start an early, polled-mode console on an msm serial
1059 port at the specified address. The serial port
1060 must already be setup and configured. Options are not
1063 msm_serial_dm,<addr>
1064 Start an early, polled-mode console on an msm serial
1065 dm port at the specified address. The serial port
1066 must already be setup and configured. Options are not
1070 Start an early, polled-mode console on a serial port
1071 of an Actions Semi SoC, such as S500 or S900, at the
1072 specified address. The serial port must already be
1073 setup and configured. Options are not yet supported.
1076 Start an early, polled-mode console on a serial port
1077 of an RDA Micro SoC, such as RDA8810PL, at the
1078 specified address. The serial port must already be
1079 setup and configured. Options are not yet supported.
1082 Use RISC-V SBI (Supervisor Binary Interface) for early
1085 smh Use ARM semihosting calls for early console.
1093 Use early console provided by serial driver available
1094 on Samsung SoCs, requires selecting proper type and
1095 a correct base address of the selected UART port. The
1096 serial port must already be setup and configured.
1097 Options are not yet supported.
1100 Start an early, polled-mode console on a lantiq serial
1101 (lqasc) port at the specified address. The serial port
1102 must already be setup and configured. Options are not
1107 Use early console provided by Freescale LP UART driver
1108 found on Freescale Vybrid and QorIQ LS1021A processors.
1109 A valid base address must be provided, and the serial
1110 port must already be setup and configured.
1114 Start an early, polled-mode, output-only console on the
1115 Freescale i.MX UART at the specified address. The UART
1116 must already be setup and configured.
1119 Start an early, polled-mode console on the
1120 Armada 3700 serial port at the specified
1121 address. The serial port must already be setup
1122 and configured. Options are not yet supported.
1125 Start an early, polled-mode console on a Qualcomm
1126 Generic Interface (GENI) based serial port at the
1127 specified address. The serial port must already be
1128 setup and configured. Options are not yet supported.
1131 Start an early, unaccelerated console on the EFI
1132 memory mapped framebuffer (if available). On cache
1133 coherent non-x86 systems that use system memory for
1134 the framebuffer, pass the 'ram' option so that it is
1135 mapped with the correct attributes.
1138 Use early console provided by Freescale LINFlexD UART
1139 serial driver for NXP S32V234 SoCs. A valid base
1140 address must be provided, and the serial port must
1141 already be setup and configured.
1143 earlyprintk= [X86,SH,ARM,M68k,S390]
1147 earlyprintk=serial[,ttySn[,baudrate]]
1148 earlyprintk=serial[,0x...[,baudrate]]
1149 earlyprintk=ttySn[,baudrate]
1150 earlyprintk=dbgp[debugController#]
1151 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1152 earlyprintk=xdbc[xhciController#]
1154 earlyprintk is useful when the kernel crashes before
1155 the normal console is initialized. It is not enabled by
1156 default because it has some cosmetic problems.
1158 Append ",keep" to not disable it when the real console
1161 Only one of vga, efi, serial, or usb debug port can
1164 Currently only ttyS0 and ttyS1 may be specified by
1165 name. Other I/O ports may be explicitly specified
1166 on some architectures (x86 and arm at least) by
1167 replacing ttySn with an I/O port address, like this:
1168 earlyprintk=serial,0x1008,115200
1169 You can find the port for a given device in
1170 /proc/tty/driver/serial:
1171 2: uart:ST16650V2 port:00001008 irq:18 ...
1173 Interaction with the standard serial driver is not
1176 The VGA and EFI output is eventually overwritten by
1179 The xen output can only be used by Xen PV guests.
1181 The sclp output can only be used on s390.
1183 The optional "force" to "pciserial" enables use of a
1184 PCI device even when its classcode is not of the
1187 edac_report= [HW,EDAC] Control how to report EDAC event
1188 Format: {"on" | "off" | "force"}
1189 on: enable EDAC to report H/W event. May be overridden
1190 by other higher priority error reporting module.
1191 off: disable H/W event reporting through EDAC.
1192 force: enforce the use of EDAC to report H/W event.
1195 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1198 This is designed to be used in conjunction with
1199 the boot argument: earlyprintk=vga
1201 This parameter works in place of the kgdboc parameter
1202 but can only be used if the backing tty is available
1203 very early in the boot process. For early debugging
1204 via a serial port see kgdboc_earlycon instead.
1207 Format: {"off" | "on" | "skip[mbr]"}
1210 Format: { "old_map", "nochunk", "noruntime", "debug",
1211 "nosoftreserve", "disable_early_pci_dma",
1212 "no_disable_early_pci_dma" }
1213 old_map [X86-64]: switch to the old ioremap-based EFI
1214 runtime services mapping. [Needs CONFIG_X86_UV=y]
1215 nochunk: disable reading files in "chunks" in the EFI
1216 boot stub, as chunking can cause problems with some
1217 firmware implementations.
1218 noruntime : disable EFI runtime services support
1219 debug: enable misc debug output
1220 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1221 attribute may cause the kernel to reserve the
1222 memory range for a memory mapping driver to
1223 claim. Specify efi=nosoftreserve to disable this
1224 reservation and treat the memory by its base type
1225 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1226 disable_early_pci_dma: Disable the busmaster bit on all
1227 PCI bridges while in the EFI boot stub
1228 no_disable_early_pci_dma: Leave the busmaster bit set
1229 on all PCI bridges while in the EFI boot stub
1231 efi_no_storage_paranoia [EFI; X86]
1232 Using this parameter you can use more than 50% of
1233 your efi variable storage. Use this parameter only if
1234 you are really sure that your UEFI does sane gc and
1235 fulfills the spec otherwise your board may brick.
1237 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1238 Add arbitrary attribute to specific memory range by
1239 updating original EFI memory map.
1240 Region of memory which aa attribute is added to is
1243 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1244 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1245 attribute is added to range 0x100000000-0x180000000 and
1246 0x10a0000000-0x1120000000.
1248 If efi_fake_mem=8G@9G:0x40000 is specified, the
1249 EFI_MEMORY_SP(0x40000) attribute is added to
1250 range 0x240000000-0x43fffffff.
1252 Using this parameter you can do debugging of EFI memmap
1253 related features. For example, you can do debugging of
1254 Address Range Mirroring feature even if your box
1255 doesn't support it, or mark specific memory as
1258 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1259 that is to be dynamically loaded by Linux. If there are
1260 multiple variables with the same name but with different
1261 vendor GUIDs, all of them will be loaded. See
1262 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1265 eisa_irq_edge= [PARISC,HW]
1266 See header of drivers/parisc/eisa.c.
1269 See comment before function elanfreq_setup() in
1270 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1272 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1273 Specifies physical address of start of kernel core
1274 image elf header and optionally the size. Generally
1275 kexec loader will pass this option to capture kernel.
1276 See Documentation/admin-guide/kdump/kdump.rst for details.
1278 enable_mtrr_cleanup [X86]
1279 The kernel tries to adjust MTRR layout from continuous
1280 to discrete, to make X server driver able to add WB
1281 entry later. This parameter enables that.
1283 enable_timer_pin_1 [X86]
1284 Enable PIN 1 of APIC timer
1285 Can be useful to work around chipset bugs
1286 (in particular on some ATI chipsets).
1287 The kernel tries to set a reasonable default.
1289 enforcing [SELINUX] Set initial enforcing status.
1291 See security/selinux/Kconfig help text.
1292 0 -- permissive (log only, no denials).
1293 1 -- enforcing (deny and log).
1295 Value can be changed at runtime via
1296 /sys/fs/selinux/enforce.
1299 Disable Error Record Serialization Table (ERST)
1302 ether= [HW,NET] Ethernet cards parameters
1303 This option is obsoleted by the "netdev=" option, which
1304 has equivalent usage. See its documentation for details.
1308 Permit 'security.evm' to be updated regardless of
1309 current integrity status.
1313 fail_make_request=[KNL]
1314 General fault injection mechanism.
1315 Format: <interval>,<probability>,<space>,<times>
1316 See also Documentation/fault-injection/.
1319 See Documentation/admin-guide/blockdev/floppy.rst.
1321 force_pal_cache_flush
1322 [IA-64] Avoid check_sal_cache_flush which may hang on
1323 buggy SAL_CACHE_FLUSH implementations. Using this
1324 parameter will force ia64_sal_cache_flush to call
1325 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1328 Forcefully enable Physical Address Extension (PAE).
1329 Many Pentium M systems disable PAE but may have a
1330 functionally usable PAE implementation.
1331 Warning: use of this parameter will taint the kernel
1332 and may cause unknown problems.
1335 [FTRACE] will set and start the specified tracer
1336 as early as possible in order to facilitate early
1339 ftrace_dump_on_oops[=orig_cpu]
1340 [FTRACE] will dump the trace buffers on oops.
1341 If no parameter is passed, ftrace will dump
1342 buffers of all CPUs, but if you pass orig_cpu, it will
1343 dump only the buffer of the CPU that triggered the
1346 ftrace_filter=[function-list]
1347 [FTRACE] Limit the functions traced by the function
1348 tracer at boot up. function-list is a comma separated
1349 list of functions. This list can be changed at run
1350 time by the set_ftrace_filter file in the debugfs
1353 ftrace_notrace=[function-list]
1354 [FTRACE] Do not trace the functions specified in
1355 function-list. This list can be changed at run time
1356 by the set_ftrace_notrace file in the debugfs
1359 ftrace_graph_filter=[function-list]
1360 [FTRACE] Limit the top level callers functions traced
1361 by the function graph tracer at boot up.
1362 function-list is a comma separated list of functions
1363 that can be changed at run time by the
1364 set_graph_function file in the debugfs tracing directory.
1366 ftrace_graph_notrace=[function-list]
1367 [FTRACE] Do not trace from the functions specified in
1368 function-list. This list is a comma separated list of
1369 functions that can be changed at run time by the
1370 set_graph_notrace file in the debugfs tracing directory.
1372 ftrace_graph_max_depth=<uint>
1373 [FTRACE] Used with the function graph tracer. This is
1374 the max depth it will trace into a function. This value
1375 can be changed at run time by the max_graph_depth file
1376 in the tracefs tracing directory. default: 0 (no limit)
1378 fw_devlink= [KNL] Create device links between consumer and supplier
1379 devices by scanning the firmware to infer the
1380 consumer/supplier relationships. This feature is
1381 especially useful when drivers are loaded as modules as
1382 it ensures proper ordering of tasks like device probing
1383 (suppliers first, then consumers), supplier boot state
1384 clean up (only after all consumers have probed),
1385 suspend/resume & runtime PM (consumers first, then
1387 Format: { off | permissive | on | rpm }
1388 off -- Don't create device links from firmware info.
1389 permissive -- Create device links from firmware info
1390 but use it only for ordering boot state clean
1391 up (sync_state() calls).
1392 on -- Create device links from firmware info and use it
1393 to enforce probe and suspend/resume ordering.
1394 rpm -- Like "on", but also use to order runtime PM.
1397 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1398 support via parallel port (up to 5 devices per port)
1399 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1400 See also Documentation/input/devices/joystick-parport.rst
1404 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1408 gcov_persist= [GCOV] When non-zero (default), profiling data for
1409 kernel modules is saved and remains accessible via
1410 debugfs, even when the module is unloaded/reloaded.
1411 When zero, profiling data is discarded and associated
1412 debugfs files are removed at module unload time.
1414 goldfish [X86] Enable the goldfish android emulator platform.
1415 Don't use this when you are not running on the
1418 gpt [EFI] Forces disk with valid GPT signature but
1419 invalid Protective MBR to be treated as GPT. If the
1420 primary GPT is corrupted, it enables the backup/alternate
1421 GPT to be used instead.
1423 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1424 the "Enable 0" bit of the configuration register.
1427 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1428 the "Enable 0" bit of the configuration register.
1431 grcan.select= [HW] Select which physical interface to use.
1434 grcan.txsize= [HW] Sets the size of the tx buffer.
1435 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1437 grcan.rxsize= [HW] Sets the size of the rx buffer.
1438 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1441 gpio-mockup.gpio_mockup_ranges
1442 [HW] Sets the ranges of gpiochip of for this device.
1443 Format: <start1>,<end1>,<start2>,<end2>...
1445 hardlockup_all_cpu_backtrace=
1446 [KNL] Should the hard-lockup detector generate
1447 backtraces on all cpus.
1450 hashdist= [KNL,NUMA] Large hashes allocated during boot
1451 are distributed across NUMA nodes. Defaults on
1452 for 64-bit NUMA, off otherwise.
1453 Format: 0 | 1 (for off | on)
1455 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1457 hd= [EIDE] (E)IDE hard drive subsystem geometry
1458 Format: <cyl>,<head>,<sect>
1461 Disable Hardware Error Source Table (HEST) support;
1462 corresponding firmware-first mode error processing
1463 logic will be disabled.
1465 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1466 size of <nn>. This works even on boxes that have no
1467 highmem otherwise. This also works to reduce highmem
1468 size on bigger boxes.
1470 highres= [KNL] Enable/disable high resolution timer mode.
1471 Valid parameters: "on", "off"
1476 hpet= [X86-32,HPET] option to control HPET usage
1477 Format: { enable (default) | disable | force |
1479 disable: disable HPET and use PIT instead
1480 force: allow force enabled of undocumented chips (ICH4,
1482 verbose: show contents of HPET registers during setup
1484 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1485 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1487 hugetlb_cma= [HW] The size of a cma area used for allocation
1488 of gigantic hugepages.
1491 Reserve a cma area of given size and allocate gigantic
1492 hugepages using the cma allocator. If enabled, the
1493 boot-time allocation of gigantic hugepages is skipped.
1495 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1496 If this follows hugepagesz (below), it specifies
1497 the number of pages of hugepagesz to be allocated.
1498 If this is the first HugeTLB parameter on the command
1499 line, it specifies the number of pages to allocate for
1500 the default huge page size. See also
1501 Documentation/admin-guide/mm/hugetlbpage.rst.
1505 [HW] The size of the HugeTLB pages. This is used in
1506 conjunction with hugepages (above) to allocate huge
1507 pages of a specific size at boot. The pair
1508 hugepagesz=X hugepages=Y can be specified once for
1509 each supported huge page size. Huge page sizes are
1510 architecture dependent. See also
1511 Documentation/admin-guide/mm/hugetlbpage.rst.
1515 [KNL] Should the hung task detector generate panics.
1518 A value of 1 instructs the kernel to panic when a
1519 hung task is detected. The default value is controlled
1520 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1521 option. The value selected by this boot parameter can
1522 be changed later by the kernel.hung_task_panic sysctl.
1524 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1525 terminal devices. Valid values: 0..8
1526 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1527 If specified, z/VM IUCV HVC accepts connections
1528 from listed z/VM user IDs only.
1530 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1531 which allow the hypervisor to 'idle' the
1532 guest on lock contention.
1535 Do not unregister boot console at start. This is only
1536 useful for debugging when something happens in the window
1537 between unregistering the boot console and initializing
1540 i2c_bus= [HW] Override the default board specific I2C bus speed
1541 or register an additional I2C bus that is not
1542 registered from board initialization code.
1546 i8042.debug [HW] Toggle i8042 debug mode
1547 i8042.unmask_kbd_data
1548 [HW] Enable printing of interrupt data from the KBD port
1549 (disabled by default, and as a pre-condition
1550 requires that i8042.debug=1 be enabled)
1551 i8042.direct [HW] Put keyboard port into non-translated mode
1552 i8042.dumbkbd [HW] Pretend that controller can only read data from
1553 keyboard and cannot control its state
1554 (Don't attempt to blink the leds)
1555 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1556 i8042.nokbd [HW] Don't check/create keyboard port
1557 i8042.noloop [HW] Disable the AUX Loopback command while probing
1559 i8042.nomux [HW] Don't check presence of an active multiplexing
1561 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1563 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1564 i8042.reset [HW] Reset the controller during init, cleanup and
1565 suspend-to-ram transitions, only during s2r
1566 transitions, or never reset
1567 Format: { 1 | Y | y | 0 | N | n }
1568 1, Y, y: always reset controller
1569 0, N, n: don't ever reset controller
1570 Default: only on s2r transitions on x86; most other
1571 architectures force reset to be always executed
1572 i8042.unlock [HW] Unlock (ignore) the keylock
1573 i8042.kbdreset [HW] Reset device connected to KBD port
1577 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1578 indicates that the driver is running on unsupported
1580 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1581 does not match list of supported models.
1583 [HW] Report power status in /proc/i8k
1584 (disabled by default)
1585 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1588 i915.invert_brightness=
1589 [DRM] Invert the sense of the variable that is used to
1590 set the brightness of the panel backlight. Normally a
1591 brightness value of 0 indicates backlight switched off,
1592 and the maximum of the brightness value sets the backlight
1593 to maximum brightness. If this parameter is set to 0
1594 (default) and the machine requires it, or this parameter
1595 is set to 1, a brightness value of 0 sets the backlight
1596 to maximum brightness, and the maximum of the brightness
1597 value switches the backlight off.
1598 -1 -- never invert brightness
1599 0 -- machine default
1600 1 -- force brightness inversion
1603 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1605 ide-core.nodma= [HW] (E)IDE subsystem
1606 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1607 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1608 .cdrom .chs .ignore_cable are additional options
1609 See Documentation/ide/ide.rst.
1611 ide-generic.probe-mask= [HW] (E)IDE subsystem
1613 Probe mask for legacy ISA IDE ports. Depending on
1614 platform up to 6 ports are supported, enabled by
1615 setting corresponding bits in the mask to 1. The
1616 default value is 0x0, which has a special meaning.
1617 On systems that have PCI, it triggers scanning the
1618 PCI bus for the first and the second port, which
1619 are then probed. On systems without PCI the value
1620 of 0x0 enables probing the two first ports as if it
1623 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1624 Claim all unknown PCI IDE storage controllers.
1627 Format: idle=poll, idle=halt, idle=nomwait
1628 Poll forces a polling idle loop that can slightly
1629 improve the performance of waking up a idle CPU, but
1630 will use a lot of power and make the system run hot.
1632 idle=halt: Halt is forced to be used for CPU idle.
1633 In such case C2/C3 won't be used again.
1634 idle=nomwait: Disable mwait for CPU C-states
1636 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1637 Format: { strict | legacy | 2008 | relaxed }
1640 Choose which programs will be accepted for execution
1641 based on the IEEE 754 NaN encoding(s) supported by
1642 the FPU and the NaN encoding requested with the value
1643 of an ELF file header flag individually set by each
1644 binary. Hardware implementations are permitted to
1645 support either or both of the legacy and the 2008 NaN
1648 Available settings are as follows:
1649 strict accept binaries that request a NaN encoding
1650 supported by the FPU
1651 legacy only accept legacy-NaN binaries, if supported
1653 2008 only accept 2008-NaN binaries, if supported
1655 relaxed accept any binaries regardless of whether
1656 supported by the FPU
1658 The FPU emulator is always able to support both NaN
1659 encodings, so if no FPU hardware is present or it has
1660 been disabled with 'nofpu', then the settings of
1661 'legacy' and '2008' strap the emulator accordingly,
1662 'relaxed' straps the emulator for both legacy-NaN and
1663 2008-NaN, whereas 'strict' enables legacy-NaN only on
1664 legacy processors and both NaN encodings on MIPS32 or
1667 The setting for ABS.fmt/NEG.fmt instruction execution
1668 mode generally follows that for the NaN encoding,
1669 except where unsupported by hardware.
1671 ignore_loglevel [KNL]
1672 Ignore loglevel setting - this will print /all/
1673 kernel messages to the console. Useful for debugging.
1674 We also add it as printk module parameter, so users
1675 could change it dynamically, usually by
1676 /sys/module/printk/parameters/ignore_loglevel.
1679 Ignore RLIMIT_DATA setting for data mappings,
1680 print warning at first misuse. Can be changed via
1681 /sys/module/kernel/parameters/ignore_rlimit_data.
1683 ihash_entries= [KNL]
1684 Set number of hash buckets for inode cache.
1686 ima_appraise= [IMA] appraise integrity measurements
1687 Format: { "off" | "enforce" | "fix" | "log" }
1690 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1691 The builtin appraise policy appraises all files
1694 ima_canonical_fmt [IMA]
1695 Use the canonical format for the binary runtime
1696 measurements, instead of host native format.
1699 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1703 The list of supported hash algorithms is defined
1704 in crypto/hash_info.h.
1707 The builtin policies to load during IMA setup.
1708 Format: "tcb | appraise_tcb | secure_boot |
1711 The "tcb" policy measures all programs exec'd, files
1712 mmap'd for exec, and all files opened with the read
1713 mode bit set by either the effective uid (euid=0) or
1716 The "appraise_tcb" policy appraises the integrity of
1717 all files owned by root.
1719 The "secure_boot" policy appraises the integrity
1720 of files (eg. kexec kernel image, kernel modules,
1721 firmware, policy, etc) based on file signatures.
1723 The "fail_securely" policy forces file signature
1724 verification failure also on privileged mounted
1725 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1728 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1729 Load a policy which meets the needs of the Trusted
1730 Computing Base. This means IMA will measure all
1731 programs exec'd, files mmap'd for exec, and all files
1732 opened for read by uid=0.
1735 Select one of defined IMA measurements template formats.
1736 Formats: { "ima" | "ima-ng" | "ima-sig" }
1740 [IMA] Define a custom template format.
1741 Format: { "field1|...|fieldN" }
1743 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1744 Format: <min_file_size>
1745 Set the minimal file size for using asynchronous hash.
1746 If left unspecified, ahash usage is disabled.
1748 ahash performance varies for different data sizes on
1749 different crypto accelerators. This option can be used
1750 to achieve the best performance for a particular HW.
1752 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1754 Set hashing buffer size. Default: 4k.
1756 ahash performance varies for different chunk sizes on
1757 different crypto accelerators. This option can be used
1758 to achieve best performance for particular HW.
1762 Run specified binary instead of /sbin/init as init
1765 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1766 for working out where the kernel is dying during
1769 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1770 initcall functions. Useful for debugging built-in
1771 modules and initcalls.
1773 initrd= [BOOT] Specify the location of the initial ramdisk
1775 initrdmem= [KNL] Specify a physical address and size from which to
1776 load the initrd. If an initrd is compiled in or
1777 specified in the bootparams, it takes priority over this
1779 Format: ss[KMG],nn[KMG]
1782 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1785 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1787 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1789 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1791 init_pkru= [x86] Specify the default memory protection keys rights
1792 register contents for all processes. 0x55555554 by
1793 default (disallow access to all but pkey 0). Can
1794 override in debugfs after boot.
1796 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1799 int_pln_enable [x86] Enable power limit notification interrupt
1801 integrity_audit=[IMA]
1802 Format: { "0" | "1" }
1803 0 -- basic integrity auditing messages. (Default)
1804 1 -- additional integrity auditing messages.
1806 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1808 Enable intel iommu driver.
1810 Disable intel iommu driver.
1811 igfx_off [Default Off]
1812 By default, gfx is mapped as normal device. If a gfx
1813 device has a dedicated DMAR unit, the DMAR unit is
1814 bypassed by not enabling DMAR with this option. In
1815 this case, gfx device will use physical address for
1818 With this option iommu will not optimize to look
1819 for io virtual address below 32-bit forcing dual
1820 address cycle on pci bus for cards supporting greater
1821 than 32-bit addressing. The default is to look
1822 for translation below 32-bit and if not available
1823 then look in the higher range.
1824 strict [Default Off]
1825 With this option on every unmap_single operation will
1826 result in a hardware IOTLB flush operation as opposed
1827 to batching them for performance.
1828 sp_off [Default Off]
1829 By default, super page will be supported if Intel IOMMU
1830 has the capability. With this option, super page will
1833 By default, scalable mode will be disabled even if the
1834 hardware advertises that it has support for the scalable
1835 mode translation. With this option set, scalable mode
1836 will be used on hardware which claims to support it.
1837 tboot_noforce [Default Off]
1838 Do not force the Intel IOMMU enabled under tboot.
1839 By default, tboot will force Intel IOMMU on, which
1840 could harm performance of some high-throughput
1841 devices like 40GBit network cards, even if identity
1843 Note that using this option lowers the security
1844 provided by tboot because it makes the system
1845 vulnerable to DMA attacks.
1846 nobounce [Default off]
1847 Disable bounce buffer for untrusted devices such as
1848 the Thunderbolt devices. This will treat the untrusted
1849 devices as the trusted ones, hence might expose security
1850 risks of DMA attacks.
1852 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1853 0 disables intel_idle and fall back on acpi_idle.
1854 1 to 9 specify maximum depth of C-state.
1858 Do not enable intel_pstate as the default
1859 scaling driver for the supported processors
1861 Use intel_pstate as a scaling driver, but configure it
1862 to work with generic cpufreq governors (instead of
1863 enabling its internal governor). This mode cannot be
1864 used along with the hardware-managed P-states (HWP)
1867 Enable intel_pstate on systems that prohibit it by default
1868 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1869 instead of acpi-cpufreq may disable platform features, such
1870 as thermal controls and power capping, that rely on ACPI
1871 P-States information being indicated to OSPM and therefore
1872 should be used with caution. This option does not work with
1873 processors that aren't supported by the intel_pstate driver
1874 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1876 Do not enable hardware P state control (HWP)
1879 Only load intel_pstate on systems which support
1880 hardware P state control (HWP) if available.
1882 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1883 Description Table, specifies preferred power management
1884 profile as "Enterprise Server" or "Performance Server",
1885 then this feature is turned on by default.
1887 Allow per-logical-CPU P-State performance control limits using
1888 cpufreq sysfs interface
1890 intremap= [X86-64, Intel-IOMMU]
1891 on enable Interrupt Remapping (default)
1892 off disable Interrupt Remapping
1893 nosid disable Source ID checking
1895 BIOS x2APIC opt-out request will be ignored
1896 nopost disable Interrupt Posting
1898 iomem= Disable strict checking of access to MMIO memory
1899 strict regions from userspace.
1914 nobypass [PPC/POWERNV]
1915 Disable IOMMU bypass, using IOMMU for PCI devices.
1917 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1918 Format: { "0" | "1" }
1920 Request that DMA unmap operations use deferred
1921 invalidation of hardware TLBs, for increased
1922 throughput at the cost of reduced device isolation.
1923 Will fall back to strict mode if not supported by
1924 the relevant IOMMU driver.
1925 1 - Strict mode (default).
1926 DMA unmap operations invalidate IOMMU hardware TLBs
1930 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
1931 Format: { "0" | "1" }
1932 0 - Use IOMMU translation for DMA.
1933 1 - Bypass the IOMMU for DMA.
1934 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
1936 io7= [HW] IO7 for Marvel based alpha systems
1937 See comment before marvel_specify_io7 in
1938 arch/alpha/kernel/core_marvel.c.
1940 io_delay= [X86] I/O delay method
1942 Standard port 0x80 based delay
1944 Alternate port 0xed based delay (needed on some systems)
1946 Simple two microseconds delay
1951 See Documentation/admin-guide/nfs/nfsroot.rst.
1953 ipcmni_extend [KNL] Extend the maximum number of unique System V
1954 IPC identifiers from 32,768 to 16,777,216.
1956 irqaffinity= [SMP] Set the default irq affinity mask
1957 The argument is a cpu list, as described above.
1959 irqchip.gicv2_force_probe=
1962 Force the kernel to look for the second 4kB page
1963 of a GICv2 controller even if the memory range
1964 exposed by the device tree is too small.
1966 irqchip.gicv3_nolpi=
1968 Force the kernel to ignore the availability of
1969 LPIs (and by consequence ITSs). Intended for system
1970 that use the kernel as a bootloader, and thus want
1971 to let secondary kernels in charge of setting up
1974 irqchip.gicv3_pseudo_nmi= [ARM64]
1975 Enables support for pseudo-NMIs in the kernel. This
1976 requires the kernel to be built with
1977 CONFIG_ARM64_PSEUDO_NMI.
1980 When an interrupt is not handled search all handlers
1981 for it. Intended to get systems with badly broken
1985 When an interrupt is not handled search all handlers
1986 for it. Also check all handlers each timer
1987 interrupt. Intended to get systems with badly broken
1991 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1993 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1994 [Deprecated - use cpusets instead]
1995 Format: [flag-list,]<cpu-list>
1997 Specify one or more CPUs to isolate from disturbances
1998 specified in the flag list (default: domain):
2001 Disable the tick when a single task runs.
2003 A residual 1Hz tick is offloaded to workqueues, which you
2004 need to affine to housekeeping through the global
2005 workqueue's affinity configured via the
2006 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2007 by using the 'domain' flag described below.
2009 NOTE: by default the global workqueue runs on all CPUs,
2010 so to protect individual CPUs the 'cpumask' file has to
2011 be configured manually after bootup.
2014 Isolate from the general SMP balancing and scheduling
2015 algorithms. Note that performing domain isolation this way
2016 is irreversible: it's not possible to bring back a CPU to
2017 the domains once isolated through isolcpus. It's strongly
2018 advised to use cpusets instead to disable scheduler load
2019 balancing through the "cpuset.sched_load_balance" file.
2020 It offers a much more flexible interface where CPUs can
2021 move in and out of an isolated set anytime.
2023 You can move a process onto or off an "isolated" CPU via
2024 the CPU affinity syscalls or cpuset.
2025 <cpu number> begins at 0 and the maximum value is
2026 "number of CPUs in system - 1".
2030 Isolate from being targeted by managed interrupts
2031 which have an interrupt mask containing isolated
2032 CPUs. The affinity of managed interrupts is
2033 handled by the kernel and cannot be changed via
2034 the /proc/irq/* interfaces.
2036 This isolation is best effort and only effective
2037 if the automatically assigned interrupt mask of a
2038 device queue contains isolated and housekeeping
2039 CPUs. If housekeeping CPUs are online then such
2040 interrupts are directed to the housekeeping CPU
2041 so that IO submitted on the housekeeping CPU
2042 cannot disturb the isolated CPU.
2044 If a queue's affinity mask contains only isolated
2045 CPUs then this parameter has no effect on the
2046 interrupt routing decision, though interrupts are
2047 only delivered when tasks running on those
2048 isolated CPUs submit IO. IO submitted on
2049 housekeeping CPUs has no influence on those
2052 The format of <cpu-list> is described above.
2056 ivrs_ioapic [HW,X86_64]
2057 Provide an override to the IOAPIC-ID<->DEVICE-ID
2058 mapping provided in the IVRS ACPI table. For
2059 example, to map IOAPIC-ID decimal 10 to
2060 PCI device 00:14.0 write the parameter as:
2061 ivrs_ioapic[10]=00:14.0
2063 ivrs_hpet [HW,X86_64]
2064 Provide an override to the HPET-ID<->DEVICE-ID
2065 mapping provided in the IVRS ACPI table. For
2066 example, to map HPET-ID decimal 0 to
2067 PCI device 00:14.0 write the parameter as:
2068 ivrs_hpet[0]=00:14.0
2070 ivrs_acpihid [HW,X86_64]
2071 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2072 mapping provided in the IVRS ACPI table. For
2073 example, to map UART-HID:UID AMD0020:0 to
2074 PCI device 00:14.5 write the parameter as:
2075 ivrs_acpihid[00:14.5]=AMD0020:0
2077 js= [HW,JOY] Analog joystick
2078 See Documentation/input/joydev/joystick.rst.
2081 When CONFIG_RANDOMIZE_BASE is set, this disables
2082 kernel and module base offset ASLR (Address Space
2083 Layout Randomization).
2086 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2087 report on every invalid memory access. Without this
2088 parameter KASAN will print report only for the first
2093 kernelcore= [KNL,X86,IA-64,PPC]
2094 Format: nn[KMGTPE] | nn% | "mirror"
2095 This parameter specifies the amount of memory usable by
2096 the kernel for non-movable allocations. The requested
2097 amount is spread evenly throughout all nodes in the
2098 system as ZONE_NORMAL. The remaining memory is used for
2099 movable memory in its own zone, ZONE_MOVABLE. In the
2100 event, a node is too small to have both ZONE_NORMAL and
2101 ZONE_MOVABLE, kernelcore memory will take priority and
2102 other nodes will have a larger ZONE_MOVABLE.
2104 ZONE_MOVABLE is used for the allocation of pages that
2105 may be reclaimed or moved by the page migration
2106 subsystem. Note that allocations like PTEs-from-HighMem
2107 still use the HighMem zone if it exists, and the Normal
2108 zone if it does not.
2110 It is possible to specify the exact amount of memory in
2111 the form of "nn[KMGTPE]", a percentage of total system
2112 memory in the form of "nn%", or "mirror". If "mirror"
2113 option is specified, mirrored (reliable) memory is used
2114 for non-movable allocations and remaining memory is used
2115 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2116 are exclusive, so you cannot specify multiple forms.
2118 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2119 Format: <Controller#>[,poll interval]
2120 The controller # is the number of the ehci usb debug
2121 port as it is probed via PCI. The poll interval is
2122 optional and is the number seconds in between
2123 each poll cycle to the debug port in case you need
2124 the functionality for interrupting the kernel with
2125 gdb or control-c on the dbgp connection. When
2126 not using this parameter you use sysrq-g to break into
2127 the kernel debugger.
2129 kgdboc= [KGDB,HW] kgdb over consoles.
2130 Requires a tty driver that supports console polling,
2131 or a supported polling keyboard driver (non-usb).
2132 Serial only format: <serial_device>[,baud]
2133 keyboard only format: kbd
2134 keyboard and serial format: kbd,<serial_device>[,baud]
2135 Optional Kernel mode setting:
2136 kms, kbd format: kms,kbd
2137 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2139 kgdboc_earlycon= [KGDB,HW]
2140 If the boot console provides the ability to read
2141 characters and can work in polling mode, you can use
2142 this parameter to tell kgdb to use it as a backend
2143 until the normal console is registered. Intended to
2144 be used together with the kgdboc parameter which
2145 specifies the normal console to transition to.
2147 The name of the early console should be specified
2148 as the value of this parameter. Note that the name of
2149 the early console might be different than the tty
2150 name passed to kgdboc. It's OK to leave the value
2151 blank and the first boot console that implements
2152 read() will be picked.
2154 kgdbwait [KGDB] Stop kernel execution and enter the
2155 kernel debugger at the earliest opportunity.
2157 kmac= [MIPS] korina ethernet MAC address.
2158 Configure the RouterBoard 532 series on-chip
2159 Ethernet adapter MAC address.
2161 kmemleak= [KNL] Boot-time kmemleak enable/disable
2162 Valid arguments: on, off
2164 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2167 kprobe_event=[probe-list]
2168 [FTRACE] Add kprobe events and enable at boot time.
2169 The probe-list is a semicolon delimited list of probe
2170 definitions. Each definition is same as kprobe_events
2171 interface, but the parameters are comma delimited.
2172 For example, to add a kprobe event on vfs_read with
2173 arg1 and arg2, add to the command line;
2175 kprobe_event=p,vfs_read,$arg1,$arg2
2177 See also Documentation/trace/kprobetrace.rst "Kernel
2178 Boot Parameter" section.
2180 kpti= [ARM64] Control page table isolation of user
2181 and kernel address spaces.
2182 Default: enabled on cores which need mitigation.
2186 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2187 Default is 0 (don't ignore, but inject #GP)
2189 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2190 Default is false (don't support).
2192 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2197 [KVM] Controls the software workaround for the
2198 X86_BUG_ITLB_MULTIHIT bug.
2199 force : Always deploy workaround.
2200 off : Never deploy workaround.
2201 auto : Deploy workaround based on the presence of
2202 X86_BUG_ITLB_MULTIHIT.
2206 If the software workaround is enabled for the host,
2207 guests do need not to enable it for nested guests.
2209 kvm.nx_huge_pages_recovery_ratio=
2210 [KVM] Controls how many 4KiB pages are periodically zapped
2211 back to huge pages. 0 disables the recovery, otherwise if
2212 the value is N KVM will zap 1/Nth of the 4KiB pages every
2213 minute. The default is 60.
2215 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2216 Default is 1 (enabled)
2218 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2220 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2222 kvm-arm.vgic_v3_group0_trap=
2223 [KVM,ARM] Trap guest accesses to GICv3 group-0
2226 kvm-arm.vgic_v3_group1_trap=
2227 [KVM,ARM] Trap guest accesses to GICv3 group-1
2230 kvm-arm.vgic_v3_common_trap=
2231 [KVM,ARM] Trap guest accesses to GICv3 common
2234 kvm-arm.vgic_v4_enable=
2235 [KVM,ARM] Allow use of GICv4 for direct injection of
2238 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2239 (virtualized MMU) support on capable Intel chips.
2240 Default is 1 (enabled)
2242 kvm-intel.emulate_invalid_guest_state=
2243 [KVM,Intel] Enable emulation of invalid guest states
2244 Default is 0 (disabled)
2246 kvm-intel.flexpriority=
2247 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2248 Default is 1 (enabled)
2251 [KVM,Intel] Enable VMX nesting (nVMX).
2252 Default is 0 (disabled)
2254 kvm-intel.unrestricted_guest=
2255 [KVM,Intel] Disable unrestricted guest feature
2256 (virtualized real and unpaged mode) on capable
2257 Intel chips. Default is 1 (enabled)
2259 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2262 Valid arguments: never, cond, always
2264 always: L1D cache flush on every VMENTER.
2265 cond: Flush L1D on VMENTER only when the code between
2266 VMEXIT and VMENTER can leak host memory.
2267 never: Disables the mitigation
2269 Default is cond (do L1 cache flush in specific instances)
2271 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2272 feature (tagged TLBs) on capable Intel chips.
2273 Default is 1 (enabled)
2275 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2278 The kernel PTE inversion protection is unconditionally
2279 enabled and cannot be disabled.
2282 Provides all available mitigations for the
2283 L1TF vulnerability. Disables SMT and
2284 enables all mitigations in the
2285 hypervisors, i.e. unconditional L1D flush.
2287 SMT control and L1D flush control via the
2288 sysfs interface is still possible after
2289 boot. Hypervisors will issue a warning
2290 when the first VM is started in a
2291 potentially insecure configuration,
2292 i.e. SMT enabled or L1D flush disabled.
2295 Same as 'full', but disables SMT and L1D
2296 flush runtime control. Implies the
2297 'nosmt=force' command line option.
2298 (i.e. sysfs control of SMT is disabled.)
2301 Leaves SMT enabled and enables the default
2302 hypervisor mitigation, i.e. conditional
2305 SMT control and L1D flush control via the
2306 sysfs interface is still possible after
2307 boot. Hypervisors will issue a warning
2308 when the first VM is started in a
2309 potentially insecure configuration,
2310 i.e. SMT enabled or L1D flush disabled.
2314 Disables SMT and enables the default
2315 hypervisor mitigation.
2317 SMT control and L1D flush control via the
2318 sysfs interface is still possible after
2319 boot. Hypervisors will issue a warning
2320 when the first VM is started in a
2321 potentially insecure configuration,
2322 i.e. SMT enabled or L1D flush disabled.
2325 Same as 'flush', but hypervisors will not
2326 warn when a VM is started in a potentially
2327 insecure configuration.
2330 Disables hypervisor mitigations and doesn't
2332 It also drops the swap size and available
2333 RAM limit restriction on both hypervisor and
2338 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2344 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2347 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2348 value for LAPIC timer one-shot implementation. Default
2349 back to the programmable timer unit in the LAPIC.
2351 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2354 libata.dma= [LIBATA] DMA control
2355 libata.dma=0 Disable all PATA and SATA DMA
2356 libata.dma=1 PATA and SATA Disk DMA only
2357 libata.dma=2 ATAPI (CDROM) DMA only
2358 libata.dma=4 Compact Flash DMA only
2359 Combinations also work, so libata.dma=3 enables DMA
2360 for disks and CDROMs, but not CFs.
2362 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2363 libata.ignore_hpa=0 keep BIOS limits (default)
2364 libata.ignore_hpa=1 ignore limits, using full disk
2366 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2370 libata.force= [LIBATA] Force configurations. The format is comma
2371 separated list of "[ID:]VAL" where ID is
2372 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2373 matching port, link or device. Basically, it matches
2374 the ATA ID string printed on console by libata. If
2375 the whole ID part is omitted, the last PORT and DEVICE
2376 values are used. If ID hasn't been specified yet, the
2377 configuration applies to all ports, links and devices.
2379 If only DEVICE is omitted, the parameter applies to
2380 the port and all links and devices behind it. DEVICE
2381 number of 0 either selects the first device or the
2382 first fan-out link behind PMP device. It does not
2383 select the host link. DEVICE number of 15 selects the
2384 host link and device attached to it.
2386 The VAL specifies the configuration to force. As long
2387 as there's no ambiguity shortcut notation is allowed.
2388 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2389 The following configurations can be forced.
2391 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2392 Any ID with matching PORT is used.
2394 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2396 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2397 udma[/][16,25,33,44,66,100,133] notation is also
2400 * [no]ncq: Turn on or off NCQ.
2402 * [no]ncqtrim: Turn off queued DSM TRIM.
2404 * nohrst, nosrst, norst: suppress hard, soft
2407 * rstonce: only attempt one reset during
2408 hot-unplug link recovery
2410 * dump_id: dump IDENTIFY data.
2412 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2414 * disable: Disable this device.
2416 If there are multiple matching configurations changing
2417 the same attribute, the last one is used.
2419 memblock=debug [KNL] Enable memblock debug messages.
2421 load_ramdisk= [RAM] List of ramdisks to load from floppy
2422 See Documentation/admin-guide/blockdev/ramdisk.rst.
2424 lockd.nlm_grace_period=P [NFS] Assign grace period.
2427 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2430 lockd.nlm_timeout=T [NFS] Assign timeout value.
2433 lockd.nlm_udpport=M [NFS] Assign UDP port.
2436 lockdown= [SECURITY]
2437 { integrity | confidentiality }
2438 Enable the kernel lockdown feature. If set to
2439 integrity, kernel features that allow userland to
2440 modify the running kernel are disabled. If set to
2441 confidentiality, kernel features that allow userland
2442 to extract confidential information from the kernel
2445 locktorture.nreaders_stress= [KNL]
2446 Set the number of locking read-acquisition kthreads.
2447 Defaults to being automatically set based on the
2448 number of online CPUs.
2450 locktorture.nwriters_stress= [KNL]
2451 Set the number of locking write-acquisition kthreads.
2453 locktorture.onoff_holdoff= [KNL]
2454 Set time (s) after boot for CPU-hotplug testing.
2456 locktorture.onoff_interval= [KNL]
2457 Set time (s) between CPU-hotplug operations, or
2458 zero to disable CPU-hotplug testing.
2460 locktorture.shuffle_interval= [KNL]
2461 Set task-shuffle interval (jiffies). Shuffling
2462 tasks allows some CPUs to go into dyntick-idle
2463 mode during the locktorture test.
2465 locktorture.shutdown_secs= [KNL]
2466 Set time (s) after boot system shutdown. This
2467 is useful for hands-off automated testing.
2469 locktorture.stat_interval= [KNL]
2470 Time (s) between statistics printk()s.
2472 locktorture.stutter= [KNL]
2473 Time (s) to stutter testing, for example,
2474 specifying five seconds causes the test to run for
2475 five seconds, wait for five seconds, and so on.
2476 This tests the locking primitive's ability to
2477 transition abruptly to and from idle.
2479 locktorture.torture_type= [KNL]
2480 Specify the locking implementation to test.
2482 locktorture.verbose= [KNL]
2483 Enable additional printk() statements.
2485 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2488 loglevel= All Kernel Messages with a loglevel smaller than the
2489 console loglevel will be printed to the console. It can
2490 also be changed with klogd or other programs. The
2491 loglevels are defined as follows:
2493 0 (KERN_EMERG) system is unusable
2494 1 (KERN_ALERT) action must be taken immediately
2495 2 (KERN_CRIT) critical conditions
2496 3 (KERN_ERR) error conditions
2497 4 (KERN_WARNING) warning conditions
2498 5 (KERN_NOTICE) normal but significant condition
2499 6 (KERN_INFO) informational
2500 7 (KERN_DEBUG) debug-level messages
2502 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2503 in bytes. n must be a power of two and greater
2504 than the minimal size. The minimal size is defined
2505 by LOG_BUF_SHIFT kernel config parameter. There is
2506 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2507 that allows to increase the default size depending on
2508 the number of CPUs. See init/Kconfig for more details.
2510 logo.nologo [FB] Disables display of the built-in Linux logo.
2511 This may be used to provide more screen space for
2512 kernel log messages and is useful when debugging
2513 kernel boot problems.
2515 lp=0 [LP] Specify parallel ports to use, e.g,
2516 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2517 lp=reset first parallel port). 'lp=0' disables the
2518 lp=auto printer driver. 'lp=reset' (which can be
2519 specified in addition to the ports) causes
2520 attached printers to be reset. Using
2521 lp=port1,port2,... specifies the parallel ports
2522 to associate lp devices with, starting with
2523 lp0. A port specification may be 'none' to skip
2524 that lp device, or a parport name such as
2525 'parport0'. Specifying 'lp=auto' instead of a
2526 port specification list means that device IDs
2527 from each port should be examined, to see if
2528 an IEEE 1284-compliant printer is attached; if
2529 so, the driver will manage that printer.
2530 See also header of drivers/char/lp.c.
2533 Sets loops_per_jiffy to given constant, thus avoiding
2534 time-consuming boot-time autodetection (up to 250 ms per
2535 CPU). 0 enables autodetection (default). To determine
2536 the correct value for your kernel, boot with normal
2537 autodetection and see what value is printed. Note that
2538 on SMP systems the preset will be applied to all CPUs,
2539 which is likely to cause problems if your CPUs need
2540 significantly divergent settings. An incorrect value
2541 will cause delays in the kernel to be wrong, leading to
2542 unpredictable I/O errors and other breakage. Although
2543 unlikely, in the extreme case this might damage your
2547 Format: <io>,<irq>,<dma>
2549 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2552 [SECURITY] Choose order of LSM initialization. This
2553 overrides CONFIG_LSM, and the "security=" parameter.
2555 machvec= [IA-64] Force the use of a particular machine-vector
2556 (machvec) in a generic kernel.
2557 Example: machvec=hpzx1
2559 machtype= [Loongson] Share the same kernel image file between different
2561 Example: machtype=lemote-yeeloong-2f-7inch
2563 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2564 than or equal to this physical address is ignored.
2566 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2567 will bring up during bootup. maxcpus=n : n >= 0 limits
2568 the kernel to bring up 'n' processors. Surely after
2569 bootup you can bring up the other plugged cpu by executing
2570 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2571 only takes effect during system bootup.
2572 While n=0 is a special case, it is equivalent to "nosmp",
2573 which also disables the IO APIC.
2575 max_loop= [LOOP] The number of loop block devices that get
2576 (loop.max_loop) unconditionally pre-created at init time. The default
2577 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2578 of statically allocating a predefined number, loop
2579 devices can be requested on-demand with the
2580 /dev/loop-control interface.
2582 mce [X86-32] Machine Check Exception
2584 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2586 md= [HW] RAID subsystems devices and level
2587 See Documentation/admin-guide/md.rst.
2590 Format: <first>,<last>
2591 Specifies range of consoles to be captured by the MDA.
2594 Control mitigation for the Micro-architectural Data
2595 Sampling (MDS) vulnerability.
2597 Certain CPUs are vulnerable to an exploit against CPU
2598 internal buffers which can forward information to a
2599 disclosure gadget under certain conditions.
2601 In vulnerable processors, the speculatively
2602 forwarded data can be used in a cache side channel
2603 attack, to access data to which the attacker does
2604 not have direct access.
2606 This parameter controls the MDS mitigation. The
2609 full - Enable MDS mitigation on vulnerable CPUs
2610 full,nosmt - Enable MDS mitigation and disable
2611 SMT on vulnerable CPUs
2612 off - Unconditionally disable MDS mitigation
2614 On TAA-affected machines, mds=off can be prevented by
2615 an active TAA mitigation as both vulnerabilities are
2616 mitigated with the same mechanism so in order to disable
2617 this mitigation, you need to specify tsx_async_abort=off
2620 Not specifying this option is equivalent to
2623 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2625 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2626 Amount of memory to be used in cases as follows:
2629 2 when the kernel is not able to see the whole system memory;
2630 3 memory that lies after 'mem=' boundary is excluded from
2631 the hypervisor, then assigned to KVM guests.
2633 [X86] Work as limiting max address. Use together
2634 with memmap= to avoid physical address space collisions.
2635 Without memmap= PCI devices could be placed at addresses
2636 belonging to unused RAM.
2638 Note that this only takes effects during boot time since
2639 in above case 3, memory may need be hot added after boot
2640 if system memory of hypervisor is not sufficient.
2642 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2646 [KNL,SH] Allow user to override the default size for
2647 per-device physically contiguous DMA buffers.
2649 memhp_default_state=online/offline
2650 [KNL] Set the initial state for the memory hotplug
2651 onlining policy. If not specified, the default value is
2652 set according to the
2653 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2655 See Documentation/admin-guide/mm/memory-hotplug.rst.
2657 memmap=exactmap [KNL,X86] Enable setting of an exact
2658 E820 memory map, as specified by the user.
2659 Such memmap=exactmap lines can be constructed based on
2660 BIOS output or other requirements. See the memmap=nn@ss
2663 memmap=nn[KMG]@ss[KMG]
2664 [KNL] Force usage of a specific region of memory.
2665 Region of memory to be used is from ss to ss+nn.
2666 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2667 which limits max address to nn[KMG].
2668 Multiple different regions can be specified,
2671 memmap=100M@2G,100M#3G,1G!1024G
2673 memmap=nn[KMG]#ss[KMG]
2674 [KNL,ACPI] Mark specific memory as ACPI data.
2675 Region of memory to be marked is from ss to ss+nn.
2677 memmap=nn[KMG]$ss[KMG]
2678 [KNL,ACPI] Mark specific memory as reserved.
2679 Region of memory to be reserved is from ss to ss+nn.
2680 Example: Exclude memory from 0x18690000-0x1869ffff
2681 memmap=64K$0x18690000
2683 memmap=0x10000$0x18690000
2684 Some bootloaders may need an escape character before '$',
2685 like Grub2, otherwise '$' and the following number
2688 memmap=nn[KMG]!ss[KMG]
2689 [KNL,X86] Mark specific memory as protected.
2690 Region of memory to be used, from ss to ss+nn.
2691 The memory region may be marked as e820 type 12 (0xc)
2692 and is NVDIMM or ADR memory.
2694 memmap=<size>%<offset>-<oldtype>+<newtype>
2695 [KNL,ACPI] Convert memory within the specified region
2696 from <oldtype> to <newtype>. If "-<oldtype>" is left
2697 out, the whole region will be marked as <newtype>,
2698 even if previously unavailable. If "+<newtype>" is left
2699 out, matching memory will be removed. Types are
2700 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2701 3 = ACPI, 12 = PRAM.
2703 memory_corruption_check=0/1 [X86]
2704 Some BIOSes seem to corrupt the first 64k of
2705 memory when doing things like suspend/resume.
2706 Setting this option will scan the memory
2707 looking for corruption. Enabling this will
2708 both detect corruption and prevent the kernel
2709 from using the memory being corrupted.
2710 However, its intended as a diagnostic tool; if
2711 repeatable BIOS-originated corruption always
2712 affects the same memory, you can use memmap=
2713 to prevent the kernel from using that memory.
2715 memory_corruption_check_size=size [X86]
2716 By default it checks for corruption in the low
2717 64k, making this memory unavailable for normal
2718 use. Use this parameter to scan for
2719 corruption in more or less memory.
2721 memory_corruption_check_period=seconds [X86]
2722 By default it checks for corruption every 60
2723 seconds. Use this parameter to check at some
2724 other rate. 0 disables periodic checking.
2726 memtest= [KNL,X86,ARM,PPC] Enable memtest
2728 default : 0 <disable>
2729 Specifies the number of memtest passes to be
2730 performed. Each pass selects another test
2731 pattern from a given set of patterns. Memtest
2732 fills the memory with this pattern, validates
2733 memory contents and reserves bad memory
2734 regions that are detected.
2736 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2737 Valid arguments: on, off
2738 Default (depends on kernel configuration option):
2739 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2740 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2741 mem_encrypt=on: Activate SME
2742 mem_encrypt=off: Do not activate SME
2744 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2745 for details on when memory encryption can be activated.
2747 mem_sleep_default= [SUSPEND] Default system suspend mode:
2748 s2idle - Suspend-To-Idle
2749 shallow - Power-On Suspend or equivalent (if supported)
2750 deep - Suspend-To-RAM or equivalent (if supported)
2751 See Documentation/admin-guide/pm/sleep-states.rst.
2753 meye.*= [HW] Set MotionEye Camera parameters
2754 See Documentation/admin-guide/media/meye.rst.
2756 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2757 Multi-Function General Purpose Timers on AMD Geode
2760 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2761 the BIOS has incorrectly applied a workaround. TinyBIOS
2762 version 0.98 is known to be affected, 0.99 fixes the
2763 problem by letting the user disable the workaround.
2767 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2768 physical address is ignored.
2770 mini2440= [ARM,HW,KNL]
2771 Format:[0..2][b][c][t]
2773 MINI2440 configuration specification:
2774 0 - The attached screen is the 3.5" TFT
2775 1 - The attached screen is the 7" TFT
2776 2 - The VGA Shield is attached (1024x768)
2777 Leaving out the screen size parameter will not load
2778 the TFT driver, and the framebuffer will be left
2780 b - Enable backlight. The TFT backlight pin will be
2781 linked to the kernel VESA blanking code and a GPIO
2782 LED. This parameter is not necessary when using the
2784 c - Enable the s3c camera interface.
2785 t - Reserved for enabling touchscreen support. The
2786 touchscreen support is not enabled in the mainstream
2787 kernel as of 2.6.30, a preliminary port can be found
2788 in the "bleeding edge" mini2440 support kernel at
2789 http://repo.or.cz/w/linux-2.6/mini2440.git
2792 [X86,PPC,S390,ARM64] Control optional mitigations for
2793 CPU vulnerabilities. This is a set of curated,
2794 arch-independent options, each of which is an
2795 aggregation of existing arch-specific options.
2798 Disable all optional CPU mitigations. This
2799 improves system performance, but it may also
2800 expose users to several CPU vulnerabilities.
2801 Equivalent to: nopti [X86,PPC]
2803 nospectre_v1 [X86,PPC]
2805 nospectre_v2 [X86,PPC,S390,ARM64]
2806 spectre_v2_user=off [X86]
2807 spec_store_bypass_disable=off [X86,PPC]
2808 ssbd=force-off [ARM64]
2811 tsx_async_abort=off [X86]
2812 kvm.nx_huge_pages=off [X86]
2815 This does not have any effect on
2816 kvm.nx_huge_pages when
2817 kvm.nx_huge_pages=force.
2820 Mitigate all CPU vulnerabilities, but leave SMT
2821 enabled, even if it's vulnerable. This is for
2822 users who don't want to be surprised by SMT
2823 getting disabled across kernel upgrades, or who
2824 have other ways of avoiding SMT-based attacks.
2825 Equivalent to: (default behavior)
2828 Mitigate all CPU vulnerabilities, disabling SMT
2829 if needed. This is for users who always want to
2830 be fully mitigated, even if it means losing SMT.
2831 Equivalent to: l1tf=flush,nosmt [X86]
2832 mds=full,nosmt [X86]
2833 tsx_async_abort=full,nosmt [X86]
2836 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2837 parameter allows control of the logging verbosity for
2838 the additional memory initialisation checks. A value
2839 of 0 disables mminit logging and a level of 4 will
2840 log everything. Information is printed at KERN_DEBUG
2841 so loglevel=8 may also need to be specified.
2844 [KNL] When CONFIG_MODULE_SIG is set, this means that
2845 modules without (valid) signatures will fail to load.
2846 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2847 is always true, so this option does nothing.
2849 module_blacklist= [KNL] Do not load a comma-separated list of
2850 modules. Useful for debugging problem modules.
2853 [MOUSE] Maximum time between finger touching and
2854 leaving touchpad surface for touch to be considered
2855 a tap and be reported as a left button click (for
2856 touchpads working in absolute mode only).
2858 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2859 reporting absolute coordinates, such as tablets
2860 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2861 reporting absolute coordinates, such as tablets
2863 movablecore= [KNL,X86,IA-64,PPC]
2864 Format: nn[KMGTPE] | nn%
2865 This parameter is the complement to kernelcore=, it
2866 specifies the amount of memory used for migratable
2867 allocations. If both kernelcore and movablecore is
2868 specified, then kernelcore will be at *least* the
2869 specified value but may be more. If movablecore on its
2870 own is specified, the administrator must be careful
2871 that the amount of memory usable for all allocations
2874 movable_node [KNL] Boot-time switch to make hotplugable memory
2875 NUMA nodes to be movable. This means that the memory
2876 of such nodes will be usable only for movable
2877 allocations which rules out almost all kernel
2878 allocations. Use with caution!
2880 MTD_Partition= [MTD]
2881 Format: <name>,<region-number>,<size>,<offset>
2883 MTD_Region= [MTD] Format:
2884 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2887 See drivers/mtd/parsers/cmdlinepart.c
2889 multitce=off [PPC] This parameter disables the use of the pSeries
2890 firmware feature for updating multiple TCE entries
2893 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2895 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2897 boundary - index of last SLC block on Flex-OneNAND.
2898 The remaining blocks are configured as MLC blocks.
2899 lock - Configure if Flex-OneNAND boundary should be locked.
2900 Once locked, the boundary cannot be changed.
2901 1 indicates lock status, 0 indicates unlock status.
2904 ARM/S3C2412 JIVE boot control
2906 See arch/arm/mach-s3c2412/mach-jive.c
2908 mtouchusb.raw_coordinates=
2909 [HW] Make the MicroTouch USB driver use raw coordinates
2910 ('y', default) or cooked coordinates ('n')
2912 mtrr_chunk_size=nn[KMG] [X86]
2913 used for mtrr cleanup. It is largest continuous chunk
2914 that could hold holes aka. UC entries.
2916 mtrr_gran_size=nn[KMG] [X86]
2917 Used for mtrr cleanup. It is granularity of mtrr block.
2919 Large value could prevent small alignment from
2922 mtrr_spare_reg_nr=n [X86]
2924 Range: 0,7 : spare reg number
2926 Used for mtrr cleanup. It is spare mtrr entries number.
2927 Set to 2 or more if your graphical card needs more.
2929 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2931 netdev= [NET] Network devices parameters
2932 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2933 Note that mem_start is often overloaded to mean
2934 something different and driver-specific.
2935 This usage is only documented in each driver source
2939 [NETFILTER] Enable connection tracking flow accounting
2940 0 to disable accounting
2941 1 to enable accounting
2944 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2945 See Documentation/admin-guide/nfs/nfsroot.rst.
2947 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2948 See Documentation/admin-guide/nfs/nfsroot.rst.
2950 nfsrootdebug [NFS] enable nfsroot debugging messages.
2951 See Documentation/admin-guide/nfs/nfsroot.rst.
2953 nfs.callback_nr_threads=
2954 [NFSv4] set the total number of threads that the
2955 NFS client will assign to service NFSv4 callback
2958 nfs.callback_tcpport=
2959 [NFS] set the TCP port on which the NFSv4 callback
2960 channel should listen.
2963 [NFS] sets the pathname to the program which is used
2964 to update the NFS client cache entries.
2966 nfs.cache_getent_timeout=
2967 [NFS] sets the timeout after which an attempt to
2968 update a cache entry is deemed to have failed.
2970 nfs.idmap_cache_timeout=
2971 [NFS] set the maximum lifetime for idmapper cache
2975 [NFS] enable 64-bit inode numbers.
2976 If zero, the NFS client will fake up a 32-bit inode
2977 number for the readdir() and stat() syscalls instead
2978 of returning the full 64-bit number.
2979 The default is to return 64-bit inode numbers.
2981 nfs.max_session_cb_slots=
2982 [NFSv4.1] Sets the maximum number of session
2983 slots the client will assign to the callback
2984 channel. This determines the maximum number of
2985 callbacks the client will process in parallel for
2986 a particular server.
2988 nfs.max_session_slots=
2989 [NFSv4.1] Sets the maximum number of session slots
2990 the client will attempt to negotiate with the server.
2991 This limits the number of simultaneous RPC requests
2992 that the client can send to the NFSv4.1 server.
2993 Note that there is little point in setting this
2994 value higher than the max_tcp_slot_table_limit.
2996 nfs.nfs4_disable_idmapping=
2997 [NFSv4] When set to the default of '1', this option
2998 ensures that both the RPC level authentication
2999 scheme and the NFS level operations agree to use
3000 numeric uids/gids if the mount is using the
3001 'sec=sys' security flavour. In effect it is
3002 disabling idmapping, which can make migration from
3003 legacy NFSv2/v3 systems to NFSv4 easier.
3004 Servers that do not support this mode of operation
3005 will be autodetected by the client, and it will fall
3006 back to using the idmapper.
3007 To turn off this behaviour, set the value to '0'.
3009 [NFS4] Specify an additional fixed unique ident-
3010 ification string that NFSv4 clients can insert into
3011 their nfs_client_id4 string. This is typically a
3012 UUID that is generated at system install time.
3014 nfs.send_implementation_id =
3015 [NFSv4.1] Send client implementation identification
3016 information in exchange_id requests.
3017 If zero, no implementation identification information
3019 The default is to send the implementation identification
3022 nfs.recover_lost_locks =
3023 [NFSv4] Attempt to recover locks that were lost due
3024 to a lease timeout on the server. Please note that
3025 doing this risks data corruption, since there are
3026 no guarantees that the file will remain unchanged
3027 after the locks are lost.
3028 If you want to enable the kernel legacy behaviour of
3029 attempting to recover these locks, then set this
3031 The default parameter value of '0' causes the kernel
3032 not to attempt recovery of lost locks.
3034 nfs4.layoutstats_timer =
3035 [NFSv4.2] Change the rate at which the kernel sends
3036 layoutstats to the pNFS metadata server.
3038 Setting this to value to 0 causes the kernel to use
3039 whatever value is the default set by the layout
3040 driver. A non-zero value sets the minimum interval
3041 in seconds between layoutstats transmissions.
3043 nfsd.nfs4_disable_idmapping=
3044 [NFSv4] When set to the default of '1', the NFSv4
3045 server will return only numeric uids and gids to
3046 clients using auth_sys, and will accept numeric uids
3047 and gids from such clients. This is intended to ease
3048 migration from NFSv2/v3.
3050 nmi_debug= [KNL,SH] Specify one or more actions to take
3051 when a NMI is triggered.
3052 Format: [state][,regs][,debounce][,die]
3054 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3055 Format: [panic,][nopanic,][num]
3057 0 - turn hardlockup detector in nmi_watchdog off
3058 1 - turn hardlockup detector in nmi_watchdog on
3059 When panic is specified, panic when an NMI watchdog
3060 timeout occurs (or 'nopanic' to not panic on an NMI
3061 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3062 To disable both hard and soft lockup detectors,
3063 please see 'nowatchdog'.
3064 This is useful when you use a panic=... timeout and
3065 need the box quickly up again.
3067 These settings can be accessed at runtime via
3068 the nmi_watchdog and hardlockup_panic sysctls.
3070 netpoll.carrier_timeout=
3071 [NET] Specifies amount of time (in seconds) that
3072 netpoll should wait for a carrier. By default netpoll
3075 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3076 emulation library even if a 387 maths coprocessor
3079 no5lvl [X86-64] Disable 5-level paging mode. Forces
3080 kernel to use 4-level paging instead.
3083 [HW] Never suspend the console
3084 Disable suspending of consoles during suspend and
3085 hibernate operations. Once disabled, debugging
3086 messages can reach various consoles while the rest
3087 of the system is being put to sleep (ie, while
3088 debugging driver suspend/resume hooks). This may
3089 not work reliably with all consoles, but is known
3090 to work with serial and VGA consoles.
3091 To facilitate more flexible debugging, we also add
3092 console_suspend, a printk module parameter to control
3093 it. Users could use console_suspend (usually
3094 /sys/module/printk/parameters/console_suspend) to
3095 turn on/off it dynamically.
3097 novmcoredd [KNL,KDUMP]
3098 Disable device dump. Device dump allows drivers to
3099 append dump data to vmcore so you can collect driver
3100 specified debug info. Drivers can append the data
3101 without any limit and this data is stored in memory,
3102 so this may cause significant memory stress. Disabling
3103 device dump can help save memory but the driver debug
3104 data will be no longer available. This parameter
3105 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3108 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3109 caches in the slab allocator. Saves per-node memory,
3110 but will impact performance.
3114 noaltinstr [S390] Disables alternative instructions patching
3115 (CPU alternatives feature).
3117 noapic [SMP,APIC] Tells the kernel to not make use of any
3118 IOAPICs that may be present in the system.
3120 noautogroup Disable scheduler automatic task group creation.
3122 nobats [PPC] Do not use BATs for mapping kernel lowmem
3123 on "Classic" PPC cores.
3127 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3129 nodelayacct [KNL] Disable per-task delay accounting
3131 nodsp [SH] Disable hardware DSP at boot time.
3133 noefi Disable EFI runtime services support.
3138 On X86-32 available only on PAE configured kernels.
3139 noexec=on: enable non-executable mappings (default)
3140 noexec=off: disable non-executable mappings
3143 Disable SMAP (Supervisor Mode Access Prevention)
3144 even if it is supported by processor.
3147 Disable SMEP (Supervisor Mode Execution Prevention)
3148 even if it is supported by processor.
3151 This affects only 32-bit executables.
3152 noexec32=on: enable non-executable mappings (default)
3153 read doesn't imply executable mappings
3154 noexec32=off: disable non-executable mappings
3155 read implies executable mappings
3157 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3159 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3160 register save and restore. The kernel will only save
3161 legacy floating-point registers on task switch.
3163 nohugeiomap [KNL,x86,PPC] Disable kernel huge I/O mappings.
3165 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3166 Equivalent to smt=1.
3168 [KNL,x86] Disable symmetric multithreading (SMT).
3169 nosmt=force: Force disable SMT, cannot be undone
3170 via the sysfs control file.
3172 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3173 (bounds check bypass). With this option data leaks are
3174 possible in the system.
3176 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3177 the Spectre variant 2 (indirect branch prediction)
3178 vulnerability. System may allow data leaks with this
3181 nospec_store_bypass_disable
3182 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3184 noxsave [BUGS=X86] Disables x86 extended register state save
3185 and restore using xsave. The kernel will fallback to
3186 enabling legacy floating-point and sse state.
3188 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3189 register states. The kernel will fall back to use
3190 xsave to save the states. By using this parameter,
3191 performance of saving the states is degraded because
3192 xsave doesn't support modified optimization while
3193 xsaveopt supports it on xsaveopt enabled systems.
3195 noxsaves [X86] Disables xsaves and xrstors used in saving and
3196 restoring x86 extended register state in compacted
3197 form of xsave area. The kernel will fall back to use
3198 xsaveopt and xrstor to save and restore the states
3199 in standard form of xsave area. By using this
3200 parameter, xsave area per process might occupy more
3201 memory on xsaves enabled systems.
3203 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
3204 wfi(ARM) instruction doesn't work correctly and not to
3205 use it. This is also useful when using JTAG debugger.
3207 no_file_caps Tells the kernel not to honor file capabilities. The
3208 only way then for a file to be executed with privilege
3209 is to be setuid root or executed by root.
3211 nohalt [IA-64] Tells the kernel not to use the power saving
3212 function PAL_HALT_LIGHT when idle. This increases
3213 power-consumption. On the positive side, it reduces
3214 interrupt wake-up latency, which may improve performance
3215 in certain environments such as networked servers or
3218 nohibernate [HIBERNATION] Disable hibernation and resume.
3220 nohz= [KNL] Boottime enable/disable dynamic ticks
3221 Valid arguments: on, off
3224 nohz_full= [KNL,BOOT,SMP,ISOL]
3225 The argument is a cpu list, as described above.
3226 In kernels built with CONFIG_NO_HZ_FULL=y, set
3227 the specified list of CPUs whose tick will be stopped
3228 whenever possible. The boot CPU will be forced outside
3229 the range to maintain the timekeeping. Any CPUs
3230 in this list will have their RCU callbacks offloaded,
3231 just as if they had also been called out in the
3232 rcu_nocbs= boot parameter.
3234 noiotrap [SH] Disables trapped I/O port accesses.
3236 noirqdebug [X86-32] Disables the code which attempts to detect and
3237 disable unhandled interrupt sources.
3239 no_timer_check [X86,APIC] Disables the code which tests for
3240 broken timer IRQ sources.
3242 noisapnp [ISAPNP] Disables ISA PnP code.
3244 noinitrd [RAM] Tells the kernel not to load any configured
3247 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3249 [Deprecated - use intremap=off]
3253 noinvpcid [X86] Disable the INVPCID cpu feature.
3255 nojitter [IA-64] Disables jitter checking for ITC timers.
3257 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3259 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3263 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3264 clock and use the default one.
3266 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3267 accounting. steal time is computed, but won't
3268 influence scheduler behaviour
3270 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3272 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3274 noltlbs [PPC] Do not use large page/tlb entries for kernel
3275 lowmem mapping on PPC40x and PPC8xx
3277 nomca [IA-64] Disable machine check abort handling
3279 nomce [X86-32] Disable Machine Check Exception
3281 nomfgpt [X86-32] Disable Multi-Function General Purpose
3282 Timer usage (for AMD Geode machines).
3284 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3285 shutdown the other cpus. Instead use the REBOOT_VECTOR
3288 nomodule Disable module load
3290 nopat [X86] Disable PAT (page attribute table extension of
3291 pagetables) support.
3293 nopcid [X86-64] Disable the PCID cpu feature.
3295 norandmaps Don't use address space randomization. Equivalent to
3296 echo 0 > /proc/sys/kernel/randomize_va_space
3298 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3299 with UP alternatives
3301 nordrand [X86] Disable kernel use of the RDRAND and
3302 RDSEED instructions even if they are supported
3303 by the processor. RDRAND and RDSEED are still
3304 available to user space applications.
3306 noresume [SWSUSP] Disables resume and restores original swap
3309 no-scroll [VGA] Disables scrollback.
3310 This is required for the Braillex ib80-piezo Braille
3311 reader made by F.H. Papenmeier (Germany).
3315 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3317 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3318 and disable the IO APIC. legacy for "maxcpus=0".
3320 nosoftlockup [KNL] Disable the soft-lockup detector.
3322 nosync [HW,M68K] Disables sync negotiation for all devices.
3324 nowatchdog [KNL] Disable both lockup detectors, i.e.
3325 soft-lockup and NMI watchdog (hard-lockup).
3329 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3331 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3332 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3333 Some features depend on CPU0. Known dependencies are:
3334 1. Resume from suspend/hibernate depends on CPU0.
3335 Suspend/hibernate will fail if CPU0 is offline and you
3336 need to online CPU0 before suspend/hibernate.
3337 2. PIC interrupts also depend on CPU0. CPU0 can't be
3338 removed if a PIC interrupt is detected.
3339 It's said poweroff/reboot may depend on CPU0 on some
3340 machines although I haven't seen such issues so far
3341 after CPU0 is offline on a few tested machines.
3342 If the dependencies are under your control, you can
3343 turn on cpu0_hotplug.
3345 nps_mtm_hs_ctr= [KNL,ARC]
3346 This parameter sets the maximum duration, in
3347 cycles, each HW thread of the CTOP can run
3348 without interruptions, before HW switches it.
3349 The actual maximum duration is 16 times this
3351 Format: integer between 1 and 255
3354 nptcg= [IA-64] Override max number of concurrent global TLB
3355 purges which is reported from either PAL_VM_SUMMARY or
3358 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3359 could support. nr_cpus=n : n >= 1 limits the kernel to
3360 support 'n' processors. It could be larger than the
3361 number of already plugged CPU during bootup, later in
3362 runtime you can physically add extra cpu until it reaches
3363 n. So during boot up some boot time memory for per-cpu
3364 variables need be pre-allocated for later physical cpu
3367 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3369 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3370 Allowed values are enable and disable
3372 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3373 'node', 'default' can be specified
3374 This can be set from sysctl after boot.
3375 See Documentation/admin-guide/sysctl/vm.rst for details.
3377 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3378 See Documentation/core-api/debugging-via-ohci1394.rst for more
3381 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3382 Rather than timing out after 20 ms if an EC
3383 command is not properly ACKed, override the length
3384 of the timeout. We have interrupts disabled while
3385 waiting for the ACK, so if this is set too high
3386 interrupts *may* be lost!
3388 omap_mux= [OMAP] Override bootloader pin multiplexing.
3389 Format: <mux_mode0.mode_name=value>...
3390 For example, to override I2C bus2:
3391 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3393 oprofile.timer= [HW]
3394 Use timer interrupt instead of performance counters
3396 oprofile.cpu_type= Force an oprofile cpu type
3397 This might be useful if you have an older oprofile
3398 userland or if you want common events.
3399 Format: { arch_perfmon }
3400 arch_perfmon: [X86] Force use of architectural
3401 perfmon on Intel CPUs instead of the
3402 CPU specific event set.
3403 timer: [X86] Force use of architectural NMI
3404 timer mode (see also oprofile.timer
3405 for generic hr timer mode)
3407 oops=panic Always panic on oopses. Default is to just kill the
3408 process, but there is a small probability of
3409 deadlocking the machine.
3410 This will also cause panics on machine check exceptions.
3411 Useful together with panic=30 to trigger a reboot.
3414 [KNL] Boolean flag to control whether the page allocator
3415 should randomize its free lists. The randomization may
3416 be automatically enabled if the kernel detects it is
3417 running on a platform with a direct-mapped memory-side
3418 cache, and this parameter can be used to
3419 override/disable that behavior. The state of the flag
3420 can be read from sysfs at:
3421 /sys/module/page_alloc/parameters/shuffle.
3423 page_owner= [KNL] Boot-time page_owner enabling option.
3424 Storage of the information about who allocated
3425 each page is disabled in default. With this switch,
3427 on: enable the feature
3429 page_poison= [KNL] Boot-time parameter changing the state of
3430 poisoning on the buddy allocator, available with
3431 CONFIG_PAGE_POISONING=y.
3432 off: turn off poisoning (default)
3433 on: turn on poisoning
3435 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3436 timeout > 0: seconds before rebooting
3437 timeout = 0: wait forever
3438 timeout < 0: reboot immediately
3441 panic_print= Bitmask for printing system info when panic happens.
3442 User can chose combination of the following bits:
3443 bit 0: print all tasks info
3444 bit 1: print system memory info
3445 bit 2: print timer info
3446 bit 3: print locks info if CONFIG_LOCKDEP is on
3447 bit 4: print ftrace buffer
3448 bit 5: print all printk messages in buffer
3450 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3451 Format: <hex>[,nousertaint]
3452 Hexadecimal bitmask representing the set of TAINT flags
3453 that will cause the kernel to panic when add_taint() is
3454 called with any of the flags in this set.
3455 The optional switch "nousertaint" can be utilized to
3456 prevent userspace forced crashes by writing to sysctl
3457 /proc/sys/kernel/tainted any flagset matching with the
3458 bitmask set on panic_on_taint.
3459 See Documentation/admin-guide/tainted-kernels.rst for
3460 extra details on the taint flags that users can pick
3461 to compose the bitmask to assign to panic_on_taint.
3463 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3466 crash_kexec_post_notifiers
3467 Run kdump after running panic-notifiers and dumping
3468 kmsg. This only for the users who doubt kdump always
3469 succeeds in any situation.
3470 Note that this also increases risks of kdump failure,
3471 because some panic notifiers can make the crashed
3472 kernel more unstable.
3474 parkbd.port= [HW] Parallel port number the keyboard adapter is
3475 connected to, default is 0.
3477 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3478 0 for XT, 1 for AT (default is AT).
3481 parport= [HW,PPT] Specify parallel ports. 0 disables.
3482 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3483 Use 'auto' to force the driver to use any
3484 IRQ/DMA settings detected (the default is to
3485 ignore detected IRQ/DMA settings because of
3486 possible conflicts). You can specify the base
3487 address, IRQ, and DMA settings; IRQ and DMA
3488 should be numbers, or 'auto' (for using detected
3489 settings on that particular port), or 'nofifo'
3490 (to avoid using a FIFO even if it is detected).
3491 Parallel ports are assigned in the order they
3492 are specified on the command line, starting
3495 parport_init_mode= [HW,PPT]
3496 Configure VIA parallel port to operate in
3497 a specific mode. This is necessary on Pegasos
3498 computer where firmware has no options for setting
3499 up parallel port mode and sets it to spp.
3500 Currently this function knows 686a and 8231 chips.
3501 Format: [spp|ps2|epp|ecp|ecpepp]
3504 Halt all CPUs after the first oops has been printed for
3505 the specified number of seconds. This is to be used if
3506 your oopses keep scrolling off the screen.
3511 See header of drivers/block/paride/pcd.c.
3512 See also Documentation/admin-guide/blockdev/paride.rst.
3514 pci=option[,option...] [PCI] various PCI subsystem options.
3516 Some options herein operate on a specific device
3517 or a set of devices (<pci_dev>). These are
3518 specified in one of the following formats:
3520 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3521 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3523 Note: the first format specifies a PCI
3524 bus/device/function address which may change
3525 if new hardware is inserted, if motherboard
3526 firmware changes, or due to changes caused
3527 by other kernel parameters. If the
3528 domain is left unspecified, it is
3529 taken to be zero. Optionally, a path
3530 to a device through multiple device/function
3531 addresses can be specified after the base
3532 address (this is more robust against
3533 renumbering issues). The second format
3534 selects devices using IDs from the
3535 configuration space which may match multiple
3536 devices in the system.
3538 earlydump dump PCI config space before the kernel
3540 off [X86] don't probe for the PCI bus
3541 bios [X86-32] force use of PCI BIOS, don't access
3542 the hardware directly. Use this if your machine
3543 has a non-standard PCI host bridge.
3544 nobios [X86-32] disallow use of PCI BIOS, only direct
3545 hardware access methods are allowed. Use this
3546 if you experience crashes upon bootup and you
3547 suspect they are caused by the BIOS.
3548 conf1 [X86] Force use of PCI Configuration Access
3549 Mechanism 1 (config address in IO port 0xCF8,
3550 data in IO port 0xCFC, both 32-bit).
3551 conf2 [X86] Force use of PCI Configuration Access
3552 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3553 the function, IO port 0xCFA, also 8-bit, sets
3554 bus number. The config space is then accessed
3555 through ports 0xC000-0xCFFF).
3556 See http://wiki.osdev.org/PCI for more info
3557 on the configuration access mechanisms.
3558 noaer [PCIE] If the PCIEAER kernel config parameter is
3559 enabled, this kernel boot option can be used to
3560 disable the use of PCIE advanced error reporting.
3561 nodomains [PCI] Disable support for multiple PCI
3562 root domains (aka PCI segments, in ACPI-speak).
3563 nommconf [X86] Disable use of MMCONFIG for PCI
3565 check_enable_amd_mmconf [X86] check for and enable
3566 properly configured MMIO access to PCI
3567 config space on AMD family 10h CPU
3568 nomsi [MSI] If the PCI_MSI kernel config parameter is
3569 enabled, this kernel boot option can be used to
3570 disable the use of MSI interrupts system-wide.
3571 noioapicquirk [APIC] Disable all boot interrupt quirks.
3572 Safety option to keep boot IRQs enabled. This
3573 should never be necessary.
3574 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3575 primary IO-APIC for bridges that cannot disable
3576 boot IRQs. This fixes a source of spurious IRQs
3577 when the system masks IRQs.
3578 noioapicreroute [APIC] Disable workaround that uses the
3579 boot IRQ equivalent of an IRQ that connects to
3580 a chipset where boot IRQs cannot be disabled.
3581 The opposite of ioapicreroute.
3582 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3583 routing table. These calls are known to be buggy
3584 on several machines and they hang the machine
3585 when used, but on other computers it's the only
3586 way to get the interrupt routing table. Try
3587 this option if the kernel is unable to allocate
3588 IRQs or discover secondary PCI buses on your
3590 rom [X86] Assign address space to expansion ROMs.
3591 Use with caution as certain devices share
3592 address decoders between ROMs and other
3594 norom [X86] Do not assign address space to
3595 expansion ROMs that do not already have
3596 BIOS assigned address ranges.
3597 nobar [X86] Do not assign address space to the
3598 BARs that weren't assigned by the BIOS.
3599 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3600 assigned automatically to PCI devices. You can
3601 make the kernel exclude IRQs of your ISA cards
3603 pirqaddr=0xAAAAA [X86] Specify the physical address
3604 of the PIRQ table (normally generated
3605 by the BIOS) if it is outside the
3606 F0000h-100000h range.
3607 lastbus=N [X86] Scan all buses thru bus #N. Can be
3608 useful if the kernel is unable to find your
3609 secondary buses and you want to tell it
3610 explicitly which ones they are.
3611 assign-busses [X86] Always assign all PCI bus
3612 numbers ourselves, overriding
3613 whatever the firmware may have done.
3614 usepirqmask [X86] Honor the possible IRQ mask stored
3615 in the BIOS $PIR table. This is needed on
3616 some systems with broken BIOSes, notably
3617 some HP Pavilion N5400 and Omnibook XE3
3618 notebooks. This will have no effect if ACPI
3619 IRQ routing is enabled.
3620 noacpi [X86] Do not use ACPI for IRQ routing
3621 or for PCI scanning.
3622 use_crs [X86] Use PCI host bridge window information
3623 from ACPI. On BIOSes from 2008 or later, this
3624 is enabled by default. If you need to use this,
3625 please report a bug.
3626 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3627 If you need to use this, please report a bug.
3628 routeirq Do IRQ routing for all PCI devices.
3629 This is normally done in pci_enable_device(),
3630 so this option is a temporary workaround
3631 for broken drivers that don't call it.
3632 skip_isa_align [X86] do not align io start addr, so can
3633 handle more pci cards
3634 noearly [X86] Don't do any early type 1 scanning.
3635 This might help on some broken boards which
3636 machine check when some devices' config space
3637 is read. But various workarounds are disabled
3638 and some IOMMU drivers will not work.
3639 bfsort Sort PCI devices into breadth-first order.
3640 This sorting is done to get a device
3641 order compatible with older (<= 2.4) kernels.
3642 nobfsort Don't sort PCI devices into breadth-first order.
3643 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3644 tuning and use the BIOS-configured MPS defaults.
3645 pcie_bus_safe Set every device's MPS to the largest value
3646 supported by all devices below the root complex.
3647 pcie_bus_perf Set device MPS to the largest allowable MPS
3648 based on its parent bus. Also set MRRS (Max
3649 Read Request Size) to the largest supported
3650 value (no larger than the MPS that the device
3651 or bus can support) for best performance.
3652 pcie_bus_peer2peer Set every device's MPS to 128B, which
3653 every device is guaranteed to support. This
3654 configuration allows peer-to-peer DMA between
3655 any pair of devices, possibly at the cost of
3656 reduced performance. This also guarantees
3657 that hot-added devices will work.
3658 cbiosize=nn[KMG] The fixed amount of bus space which is
3659 reserved for the CardBus bridge's IO window.
3660 The default value is 256 bytes.
3661 cbmemsize=nn[KMG] The fixed amount of bus space which is
3662 reserved for the CardBus bridge's memory
3663 window. The default value is 64 megabytes.
3666 [<order of align>@]<pci_dev>[; ...]
3667 Specifies alignment and device to reassign
3668 aligned memory resources. How to
3669 specify the device is described above.
3670 If <order of align> is not specified,
3671 PAGE_SIZE is used as alignment.
3672 A PCI-PCI bridge can be specified if resource
3673 windows need to be expanded.
3674 To specify the alignment for several
3675 instances of a device, the PCI vendor,
3676 device, subvendor, and subdevice may be
3677 specified, e.g., 12@pci:8086:9c22:103c:198f
3678 for 4096-byte alignment.
3679 ecrc= Enable/disable PCIe ECRC (transaction layer
3680 end-to-end CRC checking).
3681 bios: Use BIOS/firmware settings. This is the
3685 hpiosize=nn[KMG] The fixed amount of bus space which is
3686 reserved for hotplug bridge's IO window.
3687 Default size is 256 bytes.
3688 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3689 reserved for hotplug bridge's MMIO window.
3690 Default size is 2 megabytes.
3691 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3692 reserved for hotplug bridge's MMIO_PREF window.
3693 Default size is 2 megabytes.
3694 hpmemsize=nn[KMG] The fixed amount of bus space which is
3695 reserved for hotplug bridge's MMIO and
3697 Default size is 2 megabytes.
3698 hpbussize=nn The minimum amount of additional bus numbers
3699 reserved for buses below a hotplug bridge.
3701 realloc= Enable/disable reallocating PCI bridge resources
3702 if allocations done by BIOS are too small to
3703 accommodate resources required by all child
3705 off: Turn realloc off
3707 realloc same as realloc=on
3708 noari do not use PCIe ARI.
3709 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3710 do not use PCIe ATS (and IOMMU device IOTLB).
3711 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3712 only look for one device below a PCIe downstream
3714 big_root_window Try to add a big 64bit memory window to the PCIe
3715 root complex on AMD CPUs. Some GFX hardware
3716 can resize a BAR to allow access to all VRAM.
3717 Adding the window is slightly risky (it may
3718 conflict with unreported devices), so this
3720 disable_acs_redir=<pci_dev>[; ...]
3721 Specify one or more PCI devices (in the format
3722 specified above) separated by semicolons.
3723 Each device specified will have the PCI ACS
3724 redirect capabilities forced off which will
3725 allow P2P traffic between devices through
3726 bridges without forcing it upstream. Note:
3727 this removes isolation between devices and
3728 may put more devices in an IOMMU group.
3729 force_floating [S390] Force usage of floating interrupts.
3730 nomio [S390] Do not use MIO instructions.
3731 norid [S390] ignore the RID field and force use of
3732 one PCI domain per PCI function
3734 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3737 force Enable ASPM even on devices that claim not to support it.
3738 WARNING: Forcing ASPM on may cause system lockups.
3740 pcie_ports= [PCIE] PCIe port services handling:
3741 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3742 even if the platform doesn't give the OS permission to
3743 use them. This may cause conflicts if the platform
3744 also tries to use these services.
3745 dpc-native Use native PCIe service for DPC only. May
3746 cause conflicts if firmware uses AER or DPC.
3747 compat Disable native PCIe services (PME, AER, DPC, PCIe
3750 pcie_port_pm= [PCIE] PCIe port power management handling:
3751 off Disable power management of all PCIe ports
3752 force Forcibly enable power management of all PCIe ports
3754 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3755 nomsi Do not use MSI for native PCIe PME signaling (this makes
3756 all PCIe root ports use INTx for all services).
3758 pcmv= [HW,PCMCIA] BadgePAD 4
3762 Keep all power-domains already enabled by bootloader on,
3763 even if no driver has claimed them. This is useful
3764 for debug and development, but should not be
3765 needed on a platform with proper driver support.
3768 See Documentation/admin-guide/blockdev/paride.rst.
3770 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3773 See arch/parisc/kernel/pdc_chassis.c
3775 percpu_alloc= Select which percpu first chunk allocator to use.
3776 Currently supported values are "embed" and "page".
3777 Archs may support subset or none of the selections.
3778 See comments in mm/percpu.c for details on each
3779 allocator. This parameter is primarily for debugging
3780 and performance comparison.
3783 See Documentation/admin-guide/blockdev/paride.rst.
3786 See Documentation/admin-guide/blockdev/paride.rst.
3788 pirq= [SMP,APIC] Manual mp-table setup
3789 See Documentation/x86/i386/IO-APIC.rst.
3791 plip= [PPT,NET] Parallel port network link
3792 Format: { parport<nr> | timid | 0 }
3793 See also Documentation/admin-guide/parport.rst.
3795 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3796 Override pmtimer IOPort with a hex value.
3799 pm_debug_messages [SUSPEND,KNL]
3800 Enable suspend/resume debug messages during boot up.
3803 Enable PNP debug messages (depends on the
3804 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3805 via /sys/module/pnp/parameters/debug. We always show
3806 current resource usage; turning this on also shows
3807 possible settings and some assignment information.
3813 { on | off | curr | res | no-curr | no-res }
3816 [ISAPNP] Exclude IRQs for the autoconfiguration
3819 [ISAPNP] Exclude DMAs for the autoconfiguration
3821 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3822 Ranges are in pairs (I/O port base and size).
3825 [ISAPNP] Exclude memory regions for the
3827 Ranges are in pairs (memory base and size).
3829 ports= [IP_VS_FTP] IPVS ftp helper module
3831 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3833 Format: <port>,<port>....
3835 powersave=off [PPC] This option disables power saving features.
3836 It specifically disables cpuidle and sets the
3837 platform machine description specific power_save
3838 function to NULL. On Idle the CPU just reduces
3841 ppc_strict_facility_enable
3842 [PPC] This option catches any kernel floating point,
3843 Altivec, VSX and SPE outside of regions specifically
3844 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3845 There is some performance impact when enabling this.
3849 Disable Hardware Transactional Memory
3851 print-fatal-signals=
3852 [KNL] debug: print fatal signals
3854 If enabled, warn about various signal handling
3855 related application anomalies: too many signals,
3856 too many POSIX.1 timers, fatal signals causing a
3859 If you hit the warning due to signal overflow,
3860 you might want to try "ulimit -i unlimited".
3864 printk.always_kmsg_dump=
3865 Trigger kmsg_dump for cases other than kernel oops or
3867 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3870 printk.devkmsg={on,off,ratelimit}
3871 Control writing to /dev/kmsg.
3872 on - unlimited logging to /dev/kmsg from userspace
3873 off - logging to /dev/kmsg disabled
3874 ratelimit - ratelimit the logging
3877 printk.time= Show timing data prefixed to each printk message line
3878 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3880 processor.max_cstate= [HW,ACPI]
3881 Limit processor to maximum C-state
3882 max_cstate=9 overrides any DMI blacklist limit.
3884 processor.nocst [HW,ACPI]
3885 Ignore the _CST method to determine C-states,
3886 instead using the legacy FADT method
3888 profile= [KNL] Enable kernel profiling via /proc/profile
3889 Format: [<profiletype>,]<number>
3890 Param: <profiletype>: "schedule", "sleep", or "kvm"
3891 [defaults to kernel profiling]
3892 Param: "schedule" - profile schedule points.
3893 Param: "sleep" - profile D-state sleeping (millisecs).
3894 Requires CONFIG_SCHEDSTATS
3895 Param: "kvm" - profile VM exits.
3896 Param: <number> - step/bucket size as a power of 2 for
3897 statistical time based profiling.
3899 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3901 See Documentation/admin-guide/blockdev/ramdisk.rst.
3903 prot_virt= [S390] enable hosting protected virtual machines
3904 isolated from the hypervisor (if hardware supports
3908 psi= [KNL] Enable or disable pressure stall information
3912 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3913 probe for; one of (bare|imps|exps|lifebook|any).
3914 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3916 psmouse.resetafter= [HW,MOUSE]
3917 Try to reset the device after so many bad packets
3920 [HW,MOUSE] Set desired mouse resolution, in dpi.
3921 psmouse.smartscroll=
3922 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3923 0 = disabled, 1 = enabled (default).
3925 pstore.backend= Specify the name of the pstore backend to use
3928 See Documentation/admin-guide/blockdev/paride.rst.
3930 pti= [X86_64] Control Page Table Isolation of user and
3931 kernel address spaces. Disabling this feature
3932 removes hardening, but improves performance of
3933 system calls and interrupts.
3935 on - unconditionally enable
3936 off - unconditionally disable
3937 auto - kernel detects whether your CPU model is
3938 vulnerable to issues that PTI mitigates
3940 Not specifying this option is equivalent to pti=auto.
3943 Equivalent to pti=off
3946 [KNL] Number of legacy pty's. Overwrites compiled-in
3949 quiet [KNL] Disable most log messages
3954 See Documentation/admin-guide/md.rst.
3956 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3957 See Documentation/admin-guide/blockdev/ramdisk.rst.
3959 random.trust_cpu={on,off}
3960 [KNL] Enable or disable trusting the use of the
3961 CPU's random number generator (if available) to
3962 fully seed the kernel's CRNG. Default is controlled
3963 by CONFIG_RANDOM_TRUST_CPU.
3965 ras=option[,option,...] [KNL] RAS-specific options
3968 Disable the Correctable Errors Collector,
3969 see CONFIG_RAS_CEC help text.
3972 The argument is a cpu list, as described above,
3973 except that the string "all" can be used to
3974 specify every CPU on the system.
3976 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3977 the specified list of CPUs to be no-callback CPUs.
3978 Invocation of these CPUs' RCU callbacks will be
3979 offloaded to "rcuox/N" kthreads created for that
3980 purpose, where "x" is "p" for RCU-preempt, and
3981 "s" for RCU-sched, and "N" is the CPU number.
3982 This reduces OS jitter on the offloaded CPUs,
3983 which can be useful for HPC and real-time
3984 workloads. It can also improve energy efficiency
3985 for asymmetric multiprocessors.
3988 Rather than requiring that offloaded CPUs
3989 (specified by rcu_nocbs= above) explicitly
3990 awaken the corresponding "rcuoN" kthreads,
3991 make these kthreads poll for callbacks.
3992 This improves the real-time response for the
3993 offloaded CPUs by relieving them of the need to
3994 wake up the corresponding kthread, but degrades
3995 energy efficiency by requiring that the kthreads
3996 periodically wake up to do the polling.
3998 rcutree.blimit= [KNL]
3999 Set maximum number of finished RCU callbacks to
4000 process in one batch.
4002 rcutree.dump_tree= [KNL]
4003 Dump the structure of the rcu_node combining tree
4004 out at early boot. This is used for diagnostic
4005 purposes, to verify correct tree setup.
4007 rcutree.gp_cleanup_delay= [KNL]
4008 Set the number of jiffies to delay each step of
4009 RCU grace-period cleanup.
4011 rcutree.gp_init_delay= [KNL]
4012 Set the number of jiffies to delay each step of
4013 RCU grace-period initialization.
4015 rcutree.gp_preinit_delay= [KNL]
4016 Set the number of jiffies to delay each step of
4017 RCU grace-period pre-initialization, that is,
4018 the propagation of recent CPU-hotplug changes up
4019 the rcu_node combining tree.
4021 rcutree.use_softirq= [KNL]
4022 If set to zero, move all RCU_SOFTIRQ processing to
4023 per-CPU rcuc kthreads. Defaults to a non-zero
4024 value, meaning that RCU_SOFTIRQ is used by default.
4025 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4027 rcutree.rcu_fanout_exact= [KNL]
4028 Disable autobalancing of the rcu_node combining
4029 tree. This is used by rcutorture, and might
4030 possibly be useful for architectures having high
4031 cache-to-cache transfer latencies.
4033 rcutree.rcu_fanout_leaf= [KNL]
4034 Change the number of CPUs assigned to each
4035 leaf rcu_node structure. Useful for very
4036 large systems, which will choose the value 64,
4037 and for NUMA systems with large remote-access
4038 latencies, which will choose a value aligned
4039 with the appropriate hardware boundaries.
4041 rcutree.jiffies_till_first_fqs= [KNL]
4042 Set delay from grace-period initialization to
4043 first attempt to force quiescent states.
4044 Units are jiffies, minimum value is zero,
4045 and maximum value is HZ.
4047 rcutree.jiffies_till_next_fqs= [KNL]
4048 Set delay between subsequent attempts to force
4049 quiescent states. Units are jiffies, minimum
4050 value is one, and maximum value is HZ.
4052 rcutree.jiffies_till_sched_qs= [KNL]
4053 Set required age in jiffies for a
4054 given grace period before RCU starts
4055 soliciting quiescent-state help from
4056 rcu_note_context_switch() and cond_resched().
4057 If not specified, the kernel will calculate
4058 a value based on the most recent settings
4059 of rcutree.jiffies_till_first_fqs
4060 and rcutree.jiffies_till_next_fqs.
4061 This calculated value may be viewed in
4062 rcutree.jiffies_to_sched_qs. Any attempt to set
4063 rcutree.jiffies_to_sched_qs will be cheerfully
4066 rcutree.kthread_prio= [KNL,BOOT]
4067 Set the SCHED_FIFO priority of the RCU per-CPU
4068 kthreads (rcuc/N). This value is also used for
4069 the priority of the RCU boost threads (rcub/N)
4070 and for the RCU grace-period kthreads (rcu_bh,
4071 rcu_preempt, and rcu_sched). If RCU_BOOST is
4072 set, valid values are 1-99 and the default is 1
4073 (the least-favored priority). Otherwise, when
4074 RCU_BOOST is not set, valid values are 0-99 and
4075 the default is zero (non-realtime operation).
4077 rcutree.rcu_nocb_gp_stride= [KNL]
4078 Set the number of NOCB callback kthreads in
4079 each group, which defaults to the square root
4080 of the number of CPUs. Larger numbers reduce
4081 the wakeup overhead on the global grace-period
4082 kthread, but increases that same overhead on
4083 each group's NOCB grace-period kthread.
4085 rcutree.qhimark= [KNL]
4086 Set threshold of queued RCU callbacks beyond which
4087 batch limiting is disabled.
4089 rcutree.qlowmark= [KNL]
4090 Set threshold of queued RCU callbacks below which
4091 batch limiting is re-enabled.
4093 rcutree.qovld= [KNL]
4094 Set threshold of queued RCU callbacks beyond which
4095 RCU's force-quiescent-state scan will aggressively
4096 enlist help from cond_resched() and sched IPIs to
4097 help CPUs more quickly reach quiescent states.
4098 Set to less than zero to make this be set based
4099 on rcutree.qhimark at boot time and to zero to
4100 disable more aggressive help enlistment.
4102 rcutree.rcu_idle_gp_delay= [KNL]
4103 Set wakeup interval for idle CPUs that have
4104 RCU callbacks (RCU_FAST_NO_HZ=y).
4106 rcutree.rcu_idle_lazy_gp_delay= [KNL]
4107 Set wakeup interval for idle CPUs that have
4108 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
4109 Lazy RCU callbacks are those which RCU can
4110 prove do nothing more than free memory.
4112 rcutree.rcu_kick_kthreads= [KNL]
4113 Cause the grace-period kthread to get an extra
4114 wake_up() if it sleeps three times longer than
4115 it should at force-quiescent-state time.
4116 This wake_up() will be accompanied by a
4117 WARN_ONCE() splat and an ftrace_dump().
4119 rcutree.sysrq_rcu= [KNL]
4120 Commandeer a sysrq key to dump out Tree RCU's
4121 rcu_node tree with an eye towards determining
4122 why a new grace period has not yet started.
4124 rcuperf.gp_async= [KNL]
4125 Measure performance of asynchronous
4126 grace-period primitives such as call_rcu().
4128 rcuperf.gp_async_max= [KNL]
4129 Specify the maximum number of outstanding
4130 callbacks per writer thread. When a writer
4131 thread exceeds this limit, it invokes the
4132 corresponding flavor of rcu_barrier() to allow
4133 previously posted callbacks to drain.
4135 rcuperf.gp_exp= [KNL]
4136 Measure performance of expedited synchronous
4137 grace-period primitives.
4139 rcuperf.holdoff= [KNL]
4140 Set test-start holdoff period. The purpose of
4141 this parameter is to delay the start of the
4142 test until boot completes in order to avoid
4145 rcuperf.kfree_rcu_test= [KNL]
4146 Set to measure performance of kfree_rcu() flooding.
4148 rcuperf.kfree_nthreads= [KNL]
4149 The number of threads running loops of kfree_rcu().
4151 rcuperf.kfree_alloc_num= [KNL]
4152 Number of allocations and frees done in an iteration.
4154 rcuperf.kfree_loops= [KNL]
4155 Number of loops doing rcuperf.kfree_alloc_num number
4156 of allocations and frees.
4158 rcuperf.nreaders= [KNL]
4159 Set number of RCU readers. The value -1 selects
4160 N, where N is the number of CPUs. A value
4161 "n" less than -1 selects N-n+1, where N is again
4162 the number of CPUs. For example, -2 selects N
4163 (the number of CPUs), -3 selects N+1, and so on.
4164 A value of "n" less than or equal to -N selects
4167 rcuperf.nwriters= [KNL]
4168 Set number of RCU writers. The values operate
4169 the same as for rcuperf.nreaders.
4170 N, where N is the number of CPUs
4172 rcuperf.perf_type= [KNL]
4173 Specify the RCU implementation to test.
4175 rcuperf.shutdown= [KNL]
4176 Shut the system down after performance tests
4177 complete. This is useful for hands-off automated
4180 rcuperf.verbose= [KNL]
4181 Enable additional printk() statements.
4183 rcuperf.writer_holdoff= [KNL]
4184 Write-side holdoff between grace periods,
4185 in microseconds. The default of zero says
4188 rcutorture.fqs_duration= [KNL]
4189 Set duration of force_quiescent_state bursts
4192 rcutorture.fqs_holdoff= [KNL]
4193 Set holdoff time within force_quiescent_state bursts
4196 rcutorture.fqs_stutter= [KNL]
4197 Set wait time between force_quiescent_state bursts
4200 rcutorture.fwd_progress= [KNL]
4201 Enable RCU grace-period forward-progress testing
4202 for the types of RCU supporting this notion.
4204 rcutorture.fwd_progress_div= [KNL]
4205 Specify the fraction of a CPU-stall-warning
4206 period to do tight-loop forward-progress testing.
4208 rcutorture.fwd_progress_holdoff= [KNL]
4209 Number of seconds to wait between successive
4210 forward-progress tests.
4212 rcutorture.fwd_progress_need_resched= [KNL]
4213 Enclose cond_resched() calls within checks for
4214 need_resched() during tight-loop forward-progress
4217 rcutorture.gp_cond= [KNL]
4218 Use conditional/asynchronous update-side
4219 primitives, if available.
4221 rcutorture.gp_exp= [KNL]
4222 Use expedited update-side primitives, if available.
4224 rcutorture.gp_normal= [KNL]
4225 Use normal (non-expedited) asynchronous
4226 update-side primitives, if available.
4228 rcutorture.gp_sync= [KNL]
4229 Use normal (non-expedited) synchronous
4230 update-side primitives, if available. If all
4231 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4232 rcutorture.gp_normal=, and rcutorture.gp_sync=
4233 are zero, rcutorture acts as if is interpreted
4234 they are all non-zero.
4236 rcutorture.n_barrier_cbs= [KNL]
4237 Set callbacks/threads for rcu_barrier() testing.
4239 rcutorture.nfakewriters= [KNL]
4240 Set number of concurrent RCU writers. These just
4241 stress RCU, they don't participate in the actual
4242 test, hence the "fake".
4244 rcutorture.nreaders= [KNL]
4245 Set number of RCU readers. The value -1 selects
4246 N-1, where N is the number of CPUs. A value
4247 "n" less than -1 selects N-n-2, where N is again
4248 the number of CPUs. For example, -2 selects N
4249 (the number of CPUs), -3 selects N+1, and so on.
4251 rcutorture.object_debug= [KNL]
4252 Enable debug-object double-call_rcu() testing.
4254 rcutorture.onoff_holdoff= [KNL]
4255 Set time (s) after boot for CPU-hotplug testing.
4257 rcutorture.onoff_interval= [KNL]
4258 Set time (jiffies) between CPU-hotplug operations,
4259 or zero to disable CPU-hotplug testing.
4261 rcutorture.shuffle_interval= [KNL]
4262 Set task-shuffle interval (s). Shuffling tasks
4263 allows some CPUs to go into dyntick-idle mode
4264 during the rcutorture test.
4266 rcutorture.shutdown_secs= [KNL]
4267 Set time (s) after boot system shutdown. This
4268 is useful for hands-off automated testing.
4270 rcutorture.stall_cpu= [KNL]
4271 Duration of CPU stall (s) to test RCU CPU stall
4272 warnings, zero to disable.
4274 rcutorture.stall_cpu_block= [KNL]
4275 Sleep while stalling if set. This will result
4276 in warnings from preemptible RCU in addition
4277 to any other stall-related activity.
4279 rcutorture.stall_cpu_holdoff= [KNL]
4280 Time to wait (s) after boot before inducing stall.
4282 rcutorture.stall_cpu_irqsoff= [KNL]
4283 Disable interrupts while stalling if set.
4285 rcutorture.stall_gp_kthread= [KNL]
4286 Duration (s) of forced sleep within RCU
4287 grace-period kthread to test RCU CPU stall
4288 warnings, zero to disable. If both stall_cpu
4289 and stall_gp_kthread are specified, the
4290 kthread is starved first, then the CPU.
4292 rcutorture.stat_interval= [KNL]
4293 Time (s) between statistics printk()s.
4295 rcutorture.stutter= [KNL]
4296 Time (s) to stutter testing, for example, specifying
4297 five seconds causes the test to run for five seconds,
4298 wait for five seconds, and so on. This tests RCU's
4299 ability to transition abruptly to and from idle.
4301 rcutorture.test_boost= [KNL]
4302 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4303 "Maybe" means test if the RCU implementation
4304 under test support RCU priority boosting.
4306 rcutorture.test_boost_duration= [KNL]
4307 Duration (s) of each individual boost test.
4309 rcutorture.test_boost_interval= [KNL]
4310 Interval (s) between each boost test.
4312 rcutorture.test_no_idle_hz= [KNL]
4313 Test RCU's dyntick-idle handling. See also the
4314 rcutorture.shuffle_interval parameter.
4316 rcutorture.torture_type= [KNL]
4317 Specify the RCU implementation to test.
4319 rcutorture.verbose= [KNL]
4320 Enable additional printk() statements.
4322 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4323 Dump ftrace buffer after reporting RCU CPU
4326 rcupdate.rcu_cpu_stall_suppress= [KNL]
4327 Suppress RCU CPU stall warning messages.
4329 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4330 Suppress RCU CPU stall warning messages and
4331 rcutorture writer stall warnings that occur
4332 during early boot, that is, during the time
4333 before the init task is spawned.
4335 rcupdate.rcu_cpu_stall_timeout= [KNL]
4336 Set timeout for RCU CPU stall warning messages.
4338 rcupdate.rcu_expedited= [KNL]
4339 Use expedited grace-period primitives, for
4340 example, synchronize_rcu_expedited() instead
4341 of synchronize_rcu(). This reduces latency,
4342 but can increase CPU utilization, degrade
4343 real-time latency, and degrade energy efficiency.
4344 No effect on CONFIG_TINY_RCU kernels.
4346 rcupdate.rcu_normal= [KNL]
4347 Use only normal grace-period primitives,
4348 for example, synchronize_rcu() instead of
4349 synchronize_rcu_expedited(). This improves
4350 real-time latency, CPU utilization, and
4351 energy efficiency, but can expose users to
4352 increased grace-period latency. This parameter
4353 overrides rcupdate.rcu_expedited. No effect on
4354 CONFIG_TINY_RCU kernels.
4356 rcupdate.rcu_normal_after_boot= [KNL]
4357 Once boot has completed (that is, after
4358 rcu_end_inkernel_boot() has been invoked), use
4359 only normal grace-period primitives. No effect
4360 on CONFIG_TINY_RCU kernels.
4362 rcupdate.rcu_task_ipi_delay= [KNL]
4363 Set time in jiffies during which RCU tasks will
4364 avoid sending IPIs, starting with the beginning
4365 of a given grace period. Setting a large
4366 number avoids disturbing real-time workloads,
4367 but lengthens grace periods.
4369 rcupdate.rcu_task_stall_timeout= [KNL]
4370 Set timeout in jiffies for RCU task stall warning
4371 messages. Disable with a value less than or equal
4374 rcupdate.rcu_self_test= [KNL]
4375 Run the RCU early boot self tests
4379 Run specified binary instead of /init from the ramdisk,
4380 used for early userspace startup. See initrd.
4383 force - Override the decision by the kernel to hide the
4384 advertisement of RDRAND support (this affects
4385 certain AMD processors because of buggy BIOS
4386 support, specifically around the suspend/resume
4390 Turn on/off individual RDT features. List is:
4391 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4393 E.g. to turn on cmt and turn off mba use:
4397 Format (x86 or x86_64):
4398 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4400 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4402 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4403 (prefix with 'panic_' to set mode for panic
4405 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4406 reboot_force is either force or not specified,
4407 reboot_cpu is s[mp]#### with #### being the processor
4408 to be used for rebooting.
4411 [KNL, SMP] Set scheduler's default relax_domain_level.
4412 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4414 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4415 Format: <base1>,<size1>[,<base2>,<size2>,...]
4416 Reserve I/O ports or memory so the kernel won't use
4417 them. If <base> is less than 0x10000, the region
4418 is assumed to be I/O ports; otherwise it is memory.
4420 reservetop= [X86-32]
4422 Reserves a hole at the top of the kernel virtual
4427 Set the amount of memory to reserve for BIOS at
4428 the bottom of the address space.
4430 reset_devices [KNL] Force drivers to reset the underlying device
4431 during initialization.
4434 Specify the partition device for software suspend
4436 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4438 resume_offset= [SWSUSP]
4439 Specify the offset from the beginning of the partition
4440 given by "resume=" at which the swap header is located,
4441 in <PAGE_SIZE> units (needed only for swap files).
4442 See Documentation/power/swsusp-and-swap-files.rst
4444 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4445 read the resume files
4447 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4448 Useful for devices that are detected asynchronously
4449 (e.g. USB and MMC devices).
4451 hibernate= [HIBERNATION]
4452 noresume Don't check if there's a hibernation image
4453 present during boot.
4454 nocompress Don't compress/decompress hibernation images.
4455 no Disable hibernation and resume.
4456 protect_image Turn on image protection during restoration
4457 (that will set all pages holding image data
4458 during restoration read-only).
4460 retain_initrd [RAM] Keep initrd memory after extraction
4462 rfkill.default_state=
4463 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4464 etc. communication is blocked by default.
4467 rfkill.master_switch_mode=
4468 0 The "airplane mode" button does nothing.
4469 1 The "airplane mode" button toggles between everything
4470 blocked and the previous configuration.
4471 2 The "airplane mode" button toggles between everything
4472 blocked and everything unblocked.
4474 rhash_entries= [KNL,NET]
4475 Set number of hash buckets for route cache
4478 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4481 ro [KNL] Mount root device read-only on boot
4484 on Mark read-only kernel memory as read-only (default).
4485 off Leave read-only kernel memory writable for debugging.
4488 Enable the uart passthrough on the designated usb port
4489 on Rockchip SoCs. When active, the signals of the
4490 debug-uart get routed to the D+ and D- pins of the usb
4491 port and the regular usb controller gets disabled.
4493 root= [KNL] Root filesystem
4494 See name_to_dev_t comment in init/do_mounts.c.
4496 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4497 mount the root filesystem
4499 rootflags= [KNL] Set root filesystem mount option string
4501 rootfstype= [KNL] Set root filesystem type
4503 rootwait [KNL] Wait (indefinitely) for root device to show up.
4504 Useful for devices that are detected asynchronously
4505 (e.g. USB and MMC devices).
4507 rproc_mem=nn[KMG][@address]
4508 [KNL,ARM,CMA] Remoteproc physical memory block.
4509 Memory area to be used by remote processor image,
4512 rw [KNL] Mount root device read-write on boot
4514 S [KNL] Run init in single mode
4516 s390_iommu= [HW,S390]
4517 Set s390 IOTLB flushing mode
4519 With strict flushing every unmap operation will result in
4520 an IOTLB flush. Default is lazy flushing before reuse,
4524 See drivers/net/irda/sa1100_ir.c.
4526 sbni= [NET] Granch SBNI12 leased line adapter
4528 sched_debug [KNL] Enables verbose scheduler debug messages.
4530 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4531 Allowed values are enable and disable. This feature
4532 incurs a small amount of overhead in the scheduler
4533 but is useful for debugging and performance tuning.
4535 sched_thermal_decay_shift=
4536 [KNL, SMP] Set a decay shift for scheduler thermal
4537 pressure signal. Thermal pressure signal follows the
4538 default decay period of other scheduler pelt
4539 signals(usually 32 ms but configurable). Setting
4540 sched_thermal_decay_shift will left shift the decay
4541 period for the thermal pressure signal by the shift
4543 i.e. with the default pelt decay period of 32 ms
4544 sched_thermal_decay_shift thermal pressure decay pr
4548 Format: integer between 0 and 10
4551 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4552 xtime_lock contention on larger systems, and/or RCU lock
4553 contention on all systems with CONFIG_MAXSMP set.
4554 Format: { "0" | "1" }
4555 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4557 Note: increases power consumption, thus should only be
4558 enabled if running jitter sensitive (HPC/RT) workloads.
4560 security= [SECURITY] Choose a legacy "major" security module to
4561 enable at boot. This has been deprecated by the
4564 selinux= [SELINUX] Disable or enable SELinux at boot time.
4565 Format: { "0" | "1" }
4566 See security/selinux/Kconfig help text.
4571 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4572 Format: { "0" | "1" }
4573 See security/apparmor/Kconfig help text
4576 Default value is set via kernel config option.
4578 serialnumber [BUGS=X86-32]
4581 Maximal number of shapers.
4589 Disable merging of slabs with similar size. May be
4590 necessary if there is some reason to distinguish
4591 allocs to different slabs, especially in hardened
4592 environments where the risk of heap overflows and
4593 layout control by attackers can usually be
4594 frustrated by disabling merging. This will reduce
4595 most of the exposure of a heap attack to a single
4596 cache (risks via metadata attacks are mostly
4597 unchanged). Debug options disable merging on their
4599 For more information see Documentation/vm/slub.rst.
4601 slab_max_order= [MM, SLAB]
4602 Determines the maximum allowed order for slabs.
4603 A high setting may cause OOMs due to memory
4604 fragmentation. Defaults to 1 for systems with
4605 more than 32MB of RAM, 0 otherwise.
4607 slub_debug[=options[,slabs]] [MM, SLUB]
4608 Enabling slub_debug allows one to determine the
4609 culprit if slab objects become corrupted. Enabling
4610 slub_debug can create guard zones around objects and
4611 may poison objects when not in use. Also tracks the
4612 last alloc / free. For more information see
4613 Documentation/vm/slub.rst.
4615 slub_memcg_sysfs= [MM, SLUB]
4616 Determines whether to enable sysfs directories for
4617 memory cgroup sub-caches. 1 to enable, 0 to disable.
4618 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4619 Enabling this can lead to a very high number of debug
4620 directories and files being created under
4623 slub_max_order= [MM, SLUB]
4624 Determines the maximum allowed order for slabs.
4625 A high setting may cause OOMs due to memory
4626 fragmentation. For more information see
4627 Documentation/vm/slub.rst.
4629 slub_min_objects= [MM, SLUB]
4630 The minimum number of objects per slab. SLUB will
4631 increase the slab order up to slub_max_order to
4632 generate a sufficiently large slab able to contain
4633 the number of objects indicated. The higher the number
4634 of objects the smaller the overhead of tracking slabs
4635 and the less frequently locks need to be acquired.
4636 For more information see Documentation/vm/slub.rst.
4638 slub_min_order= [MM, SLUB]
4639 Determines the minimum page order for slabs. Must be
4640 lower than slub_max_order.
4641 For more information see Documentation/vm/slub.rst.
4643 slub_nomerge [MM, SLUB]
4644 Same with slab_nomerge. This is supported for legacy.
4645 See slab_nomerge for more information.
4648 Format: <io1>[,<io2>[,...,<io8>]]
4650 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4651 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4652 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4653 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4654 smsc-ircc2.ircc_irq= [HW] IRQ line
4655 smsc-ircc2.ircc_dma= [HW] DMA channel
4656 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4657 0: Toshiba Satellite 1800 (GP data pin select)
4658 1: Fast pin select (default)
4661 smt [KNL,S390] Set the maximum number of threads (logical
4662 CPUs) to use per physical CPU on systems capable of
4663 symmetric multithreading (SMT). Will be capped to the
4664 actual hardware limit.
4666 Default: -1 (no limit)
4669 [KNL] Should the soft-lockup detector generate panics.
4672 A value of 1 instructs the soft-lockup detector
4673 to panic the machine when a soft-lockup occurs. It is
4674 also controlled by the kernel.softlockup_panic sysctl
4675 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
4676 respective build-time switch to that functionality.
4678 softlockup_all_cpu_backtrace=
4679 [KNL] Should the soft-lockup detector generate
4680 backtraces on all cpus.
4683 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4684 See Documentation/admin-guide/laptops/sonypi.rst
4686 spectre_v2= [X86] Control mitigation of Spectre variant 2
4687 (indirect branch speculation) vulnerability.
4688 The default operation protects the kernel from
4691 on - unconditionally enable, implies
4693 off - unconditionally disable, implies
4695 auto - kernel detects whether your CPU model is
4698 Selecting 'on' will, and 'auto' may, choose a
4699 mitigation method at run time according to the
4700 CPU, the available microcode, the setting of the
4701 CONFIG_RETPOLINE configuration option, and the
4702 compiler with which the kernel was built.
4704 Selecting 'on' will also enable the mitigation
4705 against user space to user space task attacks.
4707 Selecting 'off' will disable both the kernel and
4708 the user space protections.
4710 Specific mitigations can also be selected manually:
4712 retpoline - replace indirect branches
4713 retpoline,generic - google's original retpoline
4714 retpoline,amd - AMD-specific minimal thunk
4716 Not specifying this option is equivalent to
4720 [X86] Control mitigation of Spectre variant 2
4721 (indirect branch speculation) vulnerability between
4724 on - Unconditionally enable mitigations. Is
4725 enforced by spectre_v2=on
4727 off - Unconditionally disable mitigations. Is
4728 enforced by spectre_v2=off
4730 prctl - Indirect branch speculation is enabled,
4731 but mitigation can be enabled via prctl
4732 per thread. The mitigation control state
4733 is inherited on fork.
4736 - Like "prctl" above, but only STIBP is
4737 controlled per thread. IBPB is issued
4738 always when switching between different user
4742 - Same as "prctl" above, but all seccomp
4743 threads will enable the mitigation unless
4744 they explicitly opt out.
4747 - Like "seccomp" above, but only STIBP is
4748 controlled per thread. IBPB is issued
4749 always when switching between different
4750 user space processes.
4752 auto - Kernel selects the mitigation depending on
4753 the available CPU features and vulnerability.
4756 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4758 Not specifying this option is equivalent to
4759 spectre_v2_user=auto.
4761 spec_store_bypass_disable=
4762 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4763 (Speculative Store Bypass vulnerability)
4765 Certain CPUs are vulnerable to an exploit against a
4766 a common industry wide performance optimization known
4767 as "Speculative Store Bypass" in which recent stores
4768 to the same memory location may not be observed by
4769 later loads during speculative execution. The idea
4770 is that such stores are unlikely and that they can
4771 be detected prior to instruction retirement at the
4772 end of a particular speculation execution window.
4774 In vulnerable processors, the speculatively forwarded
4775 store can be used in a cache side channel attack, for
4776 example to read memory to which the attacker does not
4777 directly have access (e.g. inside sandboxed code).
4779 This parameter controls whether the Speculative Store
4780 Bypass optimization is used.
4782 On x86 the options are:
4784 on - Unconditionally disable Speculative Store Bypass
4785 off - Unconditionally enable Speculative Store Bypass
4786 auto - Kernel detects whether the CPU model contains an
4787 implementation of Speculative Store Bypass and
4788 picks the most appropriate mitigation. If the
4789 CPU is not vulnerable, "off" is selected. If the
4790 CPU is vulnerable the default mitigation is
4791 architecture and Kconfig dependent. See below.
4792 prctl - Control Speculative Store Bypass per thread
4793 via prctl. Speculative Store Bypass is enabled
4794 for a process by default. The state of the control
4795 is inherited on fork.
4796 seccomp - Same as "prctl" above, but all seccomp threads
4797 will disable SSB unless they explicitly opt out.
4799 Default mitigations:
4800 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4802 On powerpc the options are:
4804 on,auto - On Power8 and Power9 insert a store-forwarding
4805 barrier on kernel entry and exit. On Power7
4806 perform a software flush on kernel entry and
4810 Not specifying this option is equivalent to
4811 spec_store_bypass_disable=auto.
4813 spia_io_base= [HW,MTD]
4819 [X86] Enable split lock detection
4821 When enabled (and if hardware support is present), atomic
4822 instructions that access data across cache line
4823 boundaries will result in an alignment check exception.
4827 warn - the kernel will emit rate limited warnings
4828 about applications triggering the #AC
4829 exception. This mode is the default on CPUs
4830 that supports split lock detection.
4832 fatal - the kernel will send SIGBUS to applications
4833 that trigger the #AC exception.
4835 If an #AC exception is hit in the kernel or in
4836 firmware (i.e. not while executing in user mode)
4837 the kernel will oops in either "warn" or "fatal"
4841 Control the Special Register Buffer Data Sampling
4844 Certain CPUs are vulnerable to an MDS-like
4845 exploit which can leak bits from the random
4848 By default, this issue is mitigated by
4849 microcode. However, the microcode fix can cause
4850 the RDRAND and RDSEED instructions to become
4851 much slower. Among other effects, this will
4852 result in reduced throughput from /dev/urandom.
4854 The microcode mitigation can be disabled with
4855 the following option:
4857 off: Disable mitigation and remove
4858 performance impact to RDRAND and RDSEED
4860 srcutree.counter_wrap_check [KNL]
4861 Specifies how frequently to check for
4862 grace-period sequence counter wrap for the
4863 srcu_data structure's ->srcu_gp_seq_needed field.
4864 The greater the number of bits set in this kernel
4865 parameter, the less frequently counter wrap will
4866 be checked for. Note that the bottom two bits
4869 srcutree.exp_holdoff [KNL]
4870 Specifies how many nanoseconds must elapse
4871 since the end of the last SRCU grace period for
4872 a given srcu_struct until the next normal SRCU
4873 grace period will be considered for automatic
4874 expediting. Set to zero to disable automatic
4878 Speculative Store Bypass Disable control
4880 On CPUs that are vulnerable to the Speculative
4881 Store Bypass vulnerability and offer a
4882 firmware based mitigation, this parameter
4883 indicates how the mitigation should be used:
4885 force-on: Unconditionally enable mitigation for
4886 for both kernel and userspace
4887 force-off: Unconditionally disable mitigation for
4888 for both kernel and userspace
4889 kernel: Always enable mitigation in the
4890 kernel, and offer a prctl interface
4891 to allow userspace to register its
4892 interest in being mitigated too.
4894 stack_guard_gap= [MM]
4895 override the default stack gap protection. The value
4896 is in page units and it defines how many pages prior
4897 to (for stacks growing down) resp. after (for stacks
4898 growing up) the main stack are reserved for no other
4899 mapping. Default value is 256 pages.
4902 Enabled the stack tracer on boot up.
4904 stacktrace_filter=[function-list]
4905 [FTRACE] Limit the functions that the stack tracer
4906 will trace at boot up. function-list is a comma separated
4907 list of functions. This list can be changed at run
4908 time by the stack_trace_filter file in the debugfs
4909 tracing directory. Note, this enables stack tracing
4910 and the stacktrace above is not needed.
4914 Set the STI (builtin display/keyboard on the HP-PARISC
4915 machines) console (graphic card) which should be used
4916 as the initial boot-console.
4917 See also comment in drivers/video/console/sticore.c.
4920 See comment in drivers/video/console/sticore.c.
4923 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4925 sunrpc.min_resvport=
4926 sunrpc.max_resvport=
4928 SunRPC servers often require that client requests
4929 originate from a privileged port (i.e. a port in the
4930 range 0 < portnr < 1024).
4931 An administrator who wishes to reserve some of these
4932 ports for other uses may adjust the range that the
4933 kernel's sunrpc client considers to be privileged
4934 using these two parameters to set the minimum and
4935 maximum port values.
4937 sunrpc.svc_rpc_per_connection_limit=
4939 Limit the number of requests that the server will
4940 process in parallel from a single connection.
4941 The default value is 0 (no limit).
4945 Control how the NFS server code allocates CPUs to
4946 service thread pools. Depending on how many NICs
4947 you have and where their interrupts are bound, this
4948 option will affect which CPUs will do NFS serving.
4949 Note: this parameter cannot be changed while the
4950 NFS server is running.
4952 auto the server chooses an appropriate mode
4953 automatically using heuristics
4954 global a single global pool contains all CPUs
4955 percpu one pool for each CPU
4956 pernode one pool for each NUMA node (equivalent
4957 to global on non-NUMA machines)
4959 sunrpc.tcp_slot_table_entries=
4960 sunrpc.udp_slot_table_entries=
4962 Sets the upper limit on the number of simultaneous
4963 RPC calls that can be sent from the client to a
4964 server. Increasing these values may allow you to
4965 improve throughput, but will also increase the
4966 amount of memory reserved for use by the client.
4968 suspend.pm_test_delay=
4970 Sets the number of seconds to remain in a suspend test
4971 mode before resuming the system (see
4972 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4973 is set. Default value is 5.
4976 Format: { on | off | y | n | 1 | 0 }
4977 This parameter controls use of the Protected
4978 Execution Facility on pSeries.
4981 [KNL] Enable accounting of swap in memory resource
4982 controller if no parameter or 1 is given or disable
4983 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
4985 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4986 Format: { <int> | force | noforce }
4987 <int> -- Number of I/O TLB slabs
4988 force -- force using of bounce buffers even if they
4989 wouldn't be automatically used by the kernel
4990 noforce -- Never use bounce buffers (for debugging)
4995 Set a sysctl parameter, right before loading the init
4996 process, as if the value was written to the respective
4997 /proc/sys/... file. Both '.' and '/' are recognized as
4998 separators. Unrecognized parameters and invalid values
4999 are reported in the kernel log. Sysctls registered
5000 later by a loaded module cannot be set this way.
5001 Example: sysctl.vm.swappiness=40
5003 sysfs.deprecated=0|1 [KNL]
5004 Enable/disable old style sysfs layout for old udev
5005 on older distributions. When this option is enabled
5006 very new udev will not work anymore. When this option
5007 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5008 in older udev will not work anymore.
5009 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5010 the kernel configuration.
5012 sysrq_always_enabled
5014 Ignore sysrq setting - this boot parameter will
5015 neutralize any effect of /proc/sys/kernel/sysrq.
5016 Useful for debugging.
5018 tcpmhash_entries= [KNL,NET]
5019 Set the number of tcp_metrics_hash slots.
5020 Default value is 8192 or 16384 depending on total
5021 ram pages. This is used to specify the TCP metrics
5022 cache size. See Documentation/networking/ip-sysctl.rst
5023 "tcp_no_metrics_save" section for more details.
5027 test_suspend= [SUSPEND][,N]
5028 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5029 standby suspend) or "freeze" (for suspend type freeze)
5030 as the system sleep state during system startup with
5031 the optional capability to repeat N number of times.
5032 The system is woken from this state using a
5033 wakeup-capable RTC alarm.
5035 thash_entries= [KNL,NET]
5036 Set number of hash buckets for TCP connection
5038 thermal.act= [HW,ACPI]
5039 -1: disable all active trip points in all thermal zones
5040 <degrees C>: override all lowest active trip points
5042 thermal.crt= [HW,ACPI]
5043 -1: disable all critical trip points in all thermal zones
5044 <degrees C>: override all critical trip points
5046 thermal.nocrt= [HW,ACPI]
5047 Set to disable actions on ACPI thermal zone
5048 critical and hot trip points.
5050 thermal.off= [HW,ACPI]
5051 1: disable ACPI thermal control
5053 thermal.psv= [HW,ACPI]
5054 -1: disable all passive trip points
5055 <degrees C>: override all passive trip points to this
5058 thermal.tzp= [HW,ACPI]
5059 Specify global default ACPI thermal zone polling rate
5060 <deci-seconds>: poll all this frequency
5061 0: no polling (default)
5064 Force threading of all interrupt handlers except those
5065 marked explicitly IRQF_NO_THREAD.
5069 Specify if the kernel should make use of the cpu
5070 topology information if the hardware supports this.
5071 The scheduler will make use of this information and
5072 e.g. base its process migration decisions on it.
5075 topology_updates= [KNL, PPC, NUMA]
5077 Specify if the kernel should ignore (off)
5078 topology updates sent by the hypervisor to this
5081 torture.disable_onoff_at_boot= [KNL]
5082 Prevent the CPU-hotplug component of torturing
5083 until after init has spawned.
5087 tpm_suspend_pcr=[HW,TPM]
5088 Format: integer pcr id
5089 Specify that at suspend time, the tpm driver
5090 should extend the specified pcr with zeros,
5091 as a workaround for some chips which fail to
5092 flush the last written pcr on TPM_SaveState.
5093 This will guarantee that all the other pcrs
5096 trace_buf_size=nn[KMG]
5097 [FTRACE] will set tracing buffer size on each cpu.
5099 trace_event=[event-list]
5100 [FTRACE] Set and start specified trace events in order
5101 to facilitate early boot debugging. The event-list is a
5102 comma separated list of trace events to enable. See
5103 also Documentation/trace/events.rst
5105 trace_options=[option-list]
5106 [FTRACE] Enable or disable tracer options at boot.
5107 The option-list is a comma delimited list of options
5108 that can be enabled or disabled just as if you were
5109 to echo the option name into
5111 /sys/kernel/debug/tracing/trace_options
5113 For example, to enable stacktrace option (to dump the
5114 stack trace of each event), add to the command line:
5116 trace_options=stacktrace
5118 See also Documentation/trace/ftrace.rst "trace options"
5122 Have the tracepoints sent to printk as well as the
5123 tracing ring buffer. This is useful for early boot up
5124 where the system hangs or reboots and does not give the
5125 option for reading the tracing buffer or performing a
5126 ftrace_dump_on_oops.
5128 To turn off having tracepoints sent to printk,
5129 echo 0 > /proc/sys/kernel/tracepoint_printk
5130 Note, echoing 1 into this file without the
5131 tracepoint_printk kernel cmdline option has no effect.
5135 Having tracepoints sent to printk() and activating high
5136 frequency tracepoints such as irq or sched, can cause
5137 the system to live lock.
5140 [FTRACE] enable this option to disable tracing when a
5141 warning is hit. This turns off "tracing_on". Tracing can
5142 be enabled again by echoing '1' into the "tracing_on"
5143 file located in /sys/kernel/debug/tracing/
5145 This option is useful, as it disables the trace before
5146 the WARNING dump is called, which prevents the trace to
5147 be filled with content caused by the warning output.
5149 This option can also be set at run time via the sysctl
5150 option: kernel/traceoff_on_warning
5152 transparent_hugepage=
5154 Format: [always|madvise|never]
5155 Can be used to control the default behavior of the system
5156 with respect to transparent hugepages.
5157 See Documentation/admin-guide/mm/transhuge.rst
5160 tsc= Disable clocksource stability checks for TSC.
5162 [x86] reliable: mark tsc clocksource as reliable, this
5163 disables clocksource verification at runtime, as well
5164 as the stability checks done at bootup. Used to enable
5165 high-resolution timer mode on older hardware, and in
5166 virtualized environment.
5167 [x86] noirqtime: Do not use TSC to do irq accounting.
5168 Used to run time disable IRQ_TIME_ACCOUNTING on any
5169 platforms where RDTSC is slow and this accounting
5171 [x86] unstable: mark the TSC clocksource as unstable, this
5172 marks the TSC unconditionally unstable at bootup and
5173 avoids any further wobbles once the TSC watchdog notices.
5174 [x86] nowatchdog: disable clocksource watchdog. Used
5175 in situations with strict latency requirements (where
5176 interruptions from clocksource watchdog are not
5179 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5180 value instead. Useful when the early TSC frequency discovery
5181 procedure is not reliable, such as on overclocked systems
5182 with CPUID.16h support and partial CPUID.15h support.
5183 Format: <unsigned int>
5185 tsx= [X86] Control Transactional Synchronization
5186 Extensions (TSX) feature in Intel processors that
5187 support TSX control.
5189 This parameter controls the TSX feature. The options are:
5191 on - Enable TSX on the system. Although there are
5192 mitigations for all known security vulnerabilities,
5193 TSX has been known to be an accelerator for
5194 several previous speculation-related CVEs, and
5195 so there may be unknown security risks associated
5196 with leaving it enabled.
5198 off - Disable TSX on the system. (Note that this
5199 option takes effect only on newer CPUs which are
5200 not vulnerable to MDS, i.e., have
5201 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5202 the new IA32_TSX_CTRL MSR through a microcode
5203 update. This new MSR allows for the reliable
5204 deactivation of the TSX functionality.)
5206 auto - Disable TSX if X86_BUG_TAA is present,
5207 otherwise enable TSX on the system.
5209 Not specifying this option is equivalent to tsx=off.
5211 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5214 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5215 Abort (TAA) vulnerability.
5217 Similar to Micro-architectural Data Sampling (MDS)
5218 certain CPUs that support Transactional
5219 Synchronization Extensions (TSX) are vulnerable to an
5220 exploit against CPU internal buffers which can forward
5221 information to a disclosure gadget under certain
5224 In vulnerable processors, the speculatively forwarded
5225 data can be used in a cache side channel attack, to
5226 access data to which the attacker does not have direct
5229 This parameter controls the TAA mitigation. The
5232 full - Enable TAA mitigation on vulnerable CPUs
5235 full,nosmt - Enable TAA mitigation and disable SMT on
5236 vulnerable CPUs. If TSX is disabled, SMT
5237 is not disabled because CPU is not
5238 vulnerable to cross-thread TAA attacks.
5239 off - Unconditionally disable TAA mitigation
5241 On MDS-affected machines, tsx_async_abort=off can be
5242 prevented by an active MDS mitigation as both vulnerabilities
5243 are mitigated with the same mechanism so in order to disable
5244 this mitigation, you need to specify mds=off too.
5246 Not specifying this option is equivalent to
5247 tsx_async_abort=full. On CPUs which are MDS affected
5248 and deploy MDS mitigation, TAA mitigation is not
5249 required and doesn't provide any additional
5253 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5255 turbografx.map[2|3]= [HW,JOY]
5256 TurboGraFX parallel port interface
5258 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5259 See also Documentation/input/devices/joystick-parport.rst
5261 udbg-immortal [PPC] When debugging early kernel crashes that
5262 happen after console_init() and before a proper
5263 console driver takes over, this boot options might
5264 help "seeing" what's going on.
5266 uhash_entries= [KNL,NET]
5267 Set number of hash buckets for UDP/UDP-Lite connections
5270 [USB] Ignore overcurrent events (default N).
5271 Some badly-designed motherboards generate lots of
5272 bogus events, for ports that aren't wired to
5273 anything. Set this parameter to avoid log spamming.
5274 Note that genuine overcurrent events won't be
5278 [X86] Cause panic on unknown NMI.
5280 usbcore.authorized_default=
5281 [USB] Default USB device authorization:
5282 (default -1 = authorized except for wireless USB,
5283 0 = not authorized, 1 = authorized, 2 = authorized
5284 if device connected to internal port)
5286 usbcore.autosuspend=
5287 [USB] The autosuspend time delay (in seconds) used
5288 for newly-detected USB devices (default 2). This
5289 is the time required before an idle device will be
5290 autosuspended. Devices for which the delay is set
5291 to a negative value won't be autosuspended at all.
5293 usbcore.usbfs_snoop=
5294 [USB] Set to log all usbfs traffic (default 0 = off).
5296 usbcore.usbfs_snoop_max=
5297 [USB] Maximum number of bytes to snoop in each URB
5300 usbcore.blinkenlights=
5301 [USB] Set to cycle leds on hubs (default 0 = off).
5303 usbcore.old_scheme_first=
5304 [USB] Start with the old device initialization
5305 scheme (default 0 = off).
5307 usbcore.usbfs_memory_mb=
5308 [USB] Memory limit (in MB) for buffers allocated by
5309 usbfs (default = 16, 0 = max = 2047).
5311 usbcore.use_both_schemes=
5312 [USB] Try the other device initialization scheme
5313 if the first one fails (default 1 = enabled).
5315 usbcore.initial_descriptor_timeout=
5316 [USB] Specifies timeout for the initial 64-byte
5317 USB_REQ_GET_DESCRIPTOR request in milliseconds
5318 (default 5000 = 5.0 seconds).
5320 usbcore.nousb [USB] Disable the USB subsystem
5323 [USB] A list of quirk entries to augment the built-in
5324 usb core quirk list. List entries are separated by
5325 commas. Each entry has the form
5326 VendorID:ProductID:Flags. The IDs are 4-digit hex
5327 numbers and Flags is a set of letters. Each letter
5328 will change the built-in quirk; setting it if it is
5329 clear and clearing it if it is set. The letters have
5330 the following meanings:
5331 a = USB_QUIRK_STRING_FETCH_255 (string
5332 descriptors must not be fetched using
5334 b = USB_QUIRK_RESET_RESUME (device can't resume
5335 correctly so reset it instead);
5336 c = USB_QUIRK_NO_SET_INTF (device can't handle
5337 Set-Interface requests);
5338 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5339 handle its Configuration or Interface
5341 e = USB_QUIRK_RESET (device can't be reset
5342 (e.g morph devices), don't use reset);
5343 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5344 more interface descriptions than the
5345 bNumInterfaces count, and can't handle
5346 talking to these interfaces);
5347 g = USB_QUIRK_DELAY_INIT (device needs a pause
5348 during initialization, after we read
5349 the device descriptor);
5350 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5351 high speed and super speed interrupt
5352 endpoints, the USB 2.0 and USB 3.0 spec
5353 require the interval in microframes (1
5354 microframe = 125 microseconds) to be
5355 calculated as interval = 2 ^
5357 Devices with this quirk report their
5358 bInterval as the result of this
5359 calculation instead of the exponent
5360 variable used in the calculation);
5361 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5362 handle device_qualifier descriptor
5364 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5365 generates spurious wakeup, ignore
5366 remote wakeup capability);
5367 k = USB_QUIRK_NO_LPM (device can't handle Link
5369 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5370 (Device reports its bInterval as linear
5371 frames instead of the USB 2.0
5373 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5374 to be disconnected before suspend to
5375 prevent spurious wakeup);
5376 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5377 pause after every control message);
5378 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5379 delay after resetting its port);
5380 Example: quirks=0781:5580:bk,0a5c:5834:gij
5383 [USBHID] The interval which mice are to be polled at.
5386 [USBHID] The interval which joysticks are to be polled at.
5389 [USBHID] The interval which keyboards are to be polled at.
5391 usb-storage.delay_use=
5392 [UMS] The delay in seconds before a new device is
5393 scanned for Logical Units (default 1).
5396 [UMS] A list of quirks entries to supplement or
5397 override the built-in unusual_devs list. List
5398 entries are separated by commas. Each entry has
5399 the form VID:PID:Flags where VID and PID are Vendor
5400 and Product ID values (4-digit hex numbers) and
5401 Flags is a set of characters, each corresponding
5402 to a common usb-storage quirk flag as follows:
5403 a = SANE_SENSE (collect more than 18 bytes
5404 of sense data, not on uas);
5405 b = BAD_SENSE (don't collect more than 18
5406 bytes of sense data, not on uas);
5407 c = FIX_CAPACITY (decrease the reported
5408 device capacity by one sector);
5409 d = NO_READ_DISC_INFO (don't use
5410 READ_DISC_INFO command, not on uas);
5411 e = NO_READ_CAPACITY_16 (don't use
5412 READ_CAPACITY_16 command);
5413 f = NO_REPORT_OPCODES (don't use report opcodes
5415 g = MAX_SECTORS_240 (don't transfer more than
5416 240 sectors at a time, uas only);
5417 h = CAPACITY_HEURISTICS (decrease the
5418 reported device capacity by one
5419 sector if the number is odd);
5420 i = IGNORE_DEVICE (don't bind to this
5422 j = NO_REPORT_LUNS (don't use report luns
5424 l = NOT_LOCKABLE (don't try to lock and
5425 unlock ejectable media, not on uas);
5426 m = MAX_SECTORS_64 (don't transfer more
5427 than 64 sectors = 32 KB at a time,
5429 n = INITIAL_READ10 (force a retry of the
5430 initial READ(10) command, not on uas);
5431 o = CAPACITY_OK (accept the capacity
5432 reported by the device, not on uas);
5433 p = WRITE_CACHE (the device cache is ON
5434 by default, not on uas);
5435 r = IGNORE_RESIDUE (the device reports
5436 bogus residue values, not on uas);
5437 s = SINGLE_LUN (the device has only one
5439 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5440 commands, uas only);
5441 u = IGNORE_UAS (don't bind to the uas driver);
5442 w = NO_WP_DETECT (don't test whether the
5443 medium is write-protected).
5444 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5445 even if the device claims no cache,
5447 Example: quirks=0419:aaf5:rl,0421:0433:rc
5449 user_debug= [KNL,ARM]
5451 See arch/arm/Kconfig.debug help text.
5452 1 - undefined instruction events
5454 4 - invalid data aborts
5457 Example: user_debug=31
5460 [X86] Flags controlling user PTE allocations.
5462 nohigh = do not allocate PTE pages in
5463 HIGHMEM regardless of setting
5467 On X86_32, this is an alias for vdso32=. Otherwise:
5469 vdso=1: enable VDSO (the default)
5470 vdso=0: disable VDSO mapping
5472 vdso32= [X86] Control the 32-bit vDSO
5473 vdso32=1: enable 32-bit VDSO
5474 vdso32=0 or vdso32=2: disable 32-bit VDSO
5476 See the help text for CONFIG_COMPAT_VDSO for more
5477 details. If CONFIG_COMPAT_VDSO is set, the default is
5478 vdso32=0; otherwise, the default is vdso32=1.
5480 For compatibility with older kernels, vdso32=2 is an
5483 Try vdso32=0 if you encounter an error that says:
5484 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5487 vector=percpu: enable percpu vector domain
5489 video= [FB] Frame buffer configuration
5490 See Documentation/fb/modedb.rst.
5492 video.brightness_switch_enabled= [0,1]
5493 If set to 1, on receiving an ACPI notify event
5494 generated by hotkey, video driver will adjust brightness
5495 level and then send out the event to user space through
5496 the allocated input device; If set to 0, video driver
5497 will only send out the event without touching backlight
5502 [VMMIO] Memory mapped virtio (platform) device.
5504 <size>@<baseaddr>:<irq>[:<id>]
5506 <size> := size (can use standard suffixes
5508 <baseaddr> := physical base address
5509 <irq> := interrupt number (as passed to
5511 <id> := (optional) platform device id
5513 virtio_mmio.device=1K@0x100b0000:48:7
5515 Can be used multiple times for multiple devices.
5517 vga= [BOOT,X86-32] Select a particular video mode
5518 See Documentation/x86/boot.rst and
5519 Documentation/admin-guide/svga.rst.
5520 Use vga=ask for menu.
5521 This is actually a boot loader parameter; the value is
5522 passed to the kernel using a special protocol.
5524 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
5525 May slow down system boot speed, especially when
5526 enabled on systems with a large amount of memory.
5527 All options are enabled by default, and this
5528 interface is meant to allow for selectively
5529 enabling or disabling specific virtual memory
5532 Available options are:
5533 P Enable page structure init time poisoning
5534 - Disable all of the above options
5536 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5537 size of <nn>. This can be used to increase the
5538 minimum size (128MB on x86). It can also be used to
5539 decrease the size and leave more room for directly
5542 vmcp_cma=nn[MG] [KNL,S390]
5543 Sets the memory size reserved for contiguous memory
5544 allocations for the vmcp device driver.
5546 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5549 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5552 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5556 Controls the behavior of vsyscalls (i.e. calls to
5557 fixed addresses of 0xffffffffff600x00 from legacy
5558 code). Most statically-linked binaries and older
5559 versions of glibc use these calls. Because these
5560 functions are at fixed addresses, they make nice
5561 targets for exploits that can control RIP.
5563 emulate [default] Vsyscalls turn into traps and are
5564 emulated reasonably safely. The vsyscall
5567 xonly Vsyscalls turn into traps and are
5568 emulated reasonably safely. The vsyscall
5569 page is not readable.
5571 none Vsyscalls don't work at all. This makes
5572 them quite hard to use for exploits but
5573 might break your system.
5575 vt.color= [VT] Default text color.
5576 Format: 0xYX, X = foreground, Y = background.
5577 Default: 0x07 = light gray on black.
5579 vt.cur_default= [VT] Default cursor shape.
5580 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5581 the parameters of the <Esc>[?A;B;Cc escape sequence;
5582 see VGA-softcursor.txt. Default: 2 = underline.
5584 vt.default_blu= [VT]
5585 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5586 Change the default blue palette of the console.
5587 This is a 16-member array composed of values
5590 vt.default_grn= [VT]
5591 Format: <green0>,<green1>,<green2>,...,<green15>
5592 Change the default green palette of the console.
5593 This is a 16-member array composed of values
5596 vt.default_red= [VT]
5597 Format: <red0>,<red1>,<red2>,...,<red15>
5598 Change the default red palette of the console.
5599 This is a 16-member array composed of values
5605 Set system-wide default UTF-8 mode for all tty's.
5606 Default is 1, i.e. UTF-8 mode is enabled for all
5607 newly opened terminals.
5609 vt.global_cursor_default=
5612 Set system-wide default for whether a cursor
5613 is shown on new VTs. Default is -1,
5614 i.e. cursors will be created by default unless
5615 overridden by individual drivers. 0 will hide
5616 cursors, 1 will display them.
5618 vt.italic= [VT] Default color for italic text; 0-15.
5621 vt.underline= [VT] Default color for underlined text; 0-15.
5624 watchdog timers [HW,WDT] For information on watchdog timers,
5625 see Documentation/watchdog/watchdog-parameters.rst
5626 or other driver-specific files in the
5627 Documentation/watchdog/ directory.
5631 Set the hard lockup detector stall duration
5632 threshold in seconds. The soft lockup detector
5633 threshold is set to twice the value. A value of 0
5634 disables both lockup detectors. Default is 10
5637 workqueue.watchdog_thresh=
5638 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5639 warn stall conditions and dump internal state to
5640 help debugging. 0 disables workqueue stall
5641 detection; otherwise, it's the stall threshold
5642 duration in seconds. The default value is 30 and
5643 it can be updated at runtime by writing to the
5644 corresponding sysfs file.
5646 workqueue.disable_numa
5647 By default, all work items queued to unbound
5648 workqueues are affine to the NUMA nodes they're
5649 issued on, which results in better behavior in
5650 general. If NUMA affinity needs to be disabled for
5651 whatever reason, this option can be used. Note
5652 that this also can be controlled per-workqueue for
5653 workqueues visible under /sys/bus/workqueue/.
5655 workqueue.power_efficient
5656 Per-cpu workqueues are generally preferred because
5657 they show better performance thanks to cache
5658 locality; unfortunately, per-cpu workqueues tend to
5659 be more power hungry than unbound workqueues.
5661 Enabling this makes the per-cpu workqueues which
5662 were observed to contribute significantly to power
5663 consumption unbound, leading to measurably lower
5664 power usage at the cost of small performance
5667 The default value of this parameter is determined by
5668 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5670 workqueue.debug_force_rr_cpu
5671 Workqueue used to implicitly guarantee that work
5672 items queued without explicit CPU specified are put
5673 on the local CPU. This guarantee is no longer true
5674 and while local CPU is still preferred work items
5675 may be put on foreign CPUs. This debug option
5676 forces round-robin CPU selection to flush out
5677 usages which depend on the now broken guarantee.
5678 When enabled, memory and cache locality will be
5681 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5682 default x2apic cluster mode on platforms
5685 x86_intel_mid_timer= [X86-32,APBT]
5686 Choose timer option for x86 Intel MID platform.
5687 Two valid options are apbt timer only and lapic timer
5688 plus one apbt timer for broadcast timer.
5689 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5691 xen_512gb_limit [KNL,X86-64,XEN]
5692 Restricts the kernel running paravirtualized under Xen
5693 to use only up to 512 GB of RAM. The reason to do so is
5694 crash analysis tools and Xen tools for doing domain
5695 save/restore/migration must be enabled to handle larger
5698 xen_emul_unplug= [HW,X86,XEN]
5699 Unplug Xen emulated devices
5700 Format: [unplug0,][unplug1]
5701 ide-disks -- unplug primary master IDE devices
5702 aux-ide-disks -- unplug non-primary-master IDE devices
5703 nics -- unplug network devices
5704 all -- unplug all emulated devices (NICs and IDE disks)
5705 unnecessary -- unplugging emulated devices is
5706 unnecessary even if the host did not respond to
5708 never -- do not unplug even if version check succeeds
5710 xen_legacy_crash [X86,XEN]
5711 Crash from Xen panic notifier, without executing late
5712 panic() code such as dumping handler.
5714 xen_nopvspin [X86,XEN]
5715 Disables the ticketlock slowpath using Xen PV
5719 Disables the PV optimizations forcing the HVM guest to
5720 run as generic HVM guest with no PV drivers.
5721 This option is obsoleted by the "nopv" option, which
5722 has equivalent effect for XEN platform.
5724 xen_scrub_pages= [XEN]
5725 Boolean option to control scrubbing pages before giving them back
5726 to Xen, for use by other domains. Can be also changed at runtime
5727 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5728 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5730 xen_timer_slop= [X86-64,XEN]
5731 Set the timer slop (in nanoseconds) for the virtual Xen
5732 timers (default is 100000). This adjusts the minimum
5733 delta of virtualized Xen timers, where lower values
5734 improve timer resolution at the expense of processing
5735 more timer interrupts.
5737 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
5738 Disables the PV optimizations forcing the guest to run
5739 as generic guest with no PV drivers. Currently support
5740 XEN HVM, KVM, HYPER_V and VMWARE guest.
5742 xirc2ps_cs= [NET,PCMCIA]
5744 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5747 By default on POWER9 and above, the kernel will
5748 natively use the XIVE interrupt controller. This option
5749 allows the fallback firmware mode to be used:
5751 off Fallback to firmware control of XIVE interrupt
5752 controller on both pseries and powernv
5753 platforms. Only useful on POWER9 and above.
5755 xhci-hcd.quirks [USB,KNL]
5756 A hex value specifying bitmask with supplemental xhci
5757 host controller quirks. Meaning of each bit can be
5758 consulted in header drivers/usb/host/xhci.h.
5761 Format: { early | on | rw | ro | off }
5762 Controls if xmon debugger is enabled. Default is off.
5763 Passing only "xmon" is equivalent to "xmon=early".
5764 early Call xmon as early as possible on boot; xmon
5765 debugger is called from setup_arch().
5766 on xmon debugger hooks will be installed so xmon
5767 is only called on a kernel crash. Default mode,
5768 i.e. either "ro" or "rw" mode, is controlled
5769 with CONFIG_XMON_DEFAULT_RO_MODE.
5770 rw xmon debugger hooks will be installed so xmon
5771 is called only on a kernel crash, mode is write,
5772 meaning SPR registers, memory and, other data
5773 can be written using xmon commands.
5774 ro same as "rw" option above but SPR registers,
5775 memory, and other data can't be written using
5777 off xmon is disabled.