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[,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
706 cpufreq.default_governor=
707 [CPU_FREQ] Name of the default cpufreq governor or
708 policy to use. This governor must be registered in the
709 kernel before the cpufreq driver probes.
712 [X86] Delay for N microsec between assert and de-assert
713 of APIC INIT to start processors. This delay occurs
714 on every CPU online, such as boot, and resume from suspend.
717 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
719 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
721 crashkernel=size[KMG][@offset[KMG]]
722 [KNL] Using kexec, Linux can switch to a 'crash kernel'
723 upon panic. This parameter reserves the physical
724 memory region [offset, offset + size] for that kernel
725 image. If '@offset' is omitted, then a suitable offset
726 is selected automatically.
727 [KNL, X86-64] Select a region under 4G first, and
728 fall back to reserve region above 4G when '@offset'
729 hasn't been specified.
730 See Documentation/admin-guide/kdump/kdump.rst for further details.
732 crashkernel=range1:size1[,range2:size2,...][@offset]
733 [KNL] Same as above, but depends on the memory
734 in the running system. The syntax of range is
735 start-[end] where start and end are both
736 a memory unit (amount[KMG]). See also
737 Documentation/admin-guide/kdump/kdump.rst for an example.
739 crashkernel=size[KMG],high
740 [KNL, X86-64] range could be above 4G. Allow kernel
741 to allocate physical memory region from top, so could
742 be above 4G if system have more than 4G ram installed.
743 Otherwise memory region will be allocated below 4G, if
745 It will be ignored if crashkernel=X is specified.
746 crashkernel=size[KMG],low
747 [KNL, X86-64] range under 4G. When crashkernel=X,high
748 is passed, kernel could allocate physical memory region
749 above 4G, that cause second kernel crash on system
750 that require some amount of low memory, e.g. swiotlb
751 requires at least 64M+32K low memory, also enough extra
752 low memory is needed to make sure DMA buffers for 32-bit
753 devices won't run out. Kernel would try to allocate at
754 at least 256M below 4G automatically.
755 This one let user to specify own low range under 4G
756 for second kernel instead.
757 0: to disable low allocation.
758 It will be ignored when crashkernel=X,high is not used
759 or memory reserved is below 4G.
762 [KNL] Disable crypto self-tests
767 cs89x0_media= [HW,NET]
768 Format: { rj45 | aui | bnc }
771 See header of drivers/s390/block/dasd_devmap.c.
773 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
774 (one device per port)
775 Format: <port#>,<type>
776 See also Documentation/input/devices/joystick-parport.rst
778 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
780 Documentation/admin-guide/dynamic-debug-howto.rst for
781 details. Deprecated, see dyndbg.
783 debug [KNL] Enable kernel debugging (events log level).
786 [KNL] Enable printing [hashed] pointers early in the
787 boot sequence. If enabled, we use a weak hash instead
788 of siphash to hash pointers. Use this option if you are
789 seeing instances of '(___ptrval___)') and need to see a
790 value (hashed pointer) instead. Cryptographically
791 insecure, please do not use on production kernels.
794 [KNL] verbose self-tests
796 Print debugging info while doing the locking API
798 We default to 0 (no extra messages), setting it to
799 1 will print _a lot_ more information - normally
800 only useful to kernel developers.
802 debug_objects [KNL] Enable object debugging
805 [KNL] Disable object debugging
807 debug_guardpage_minorder=
808 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
809 parameter allows control of the order of pages that will
810 be intentionally kept free (and hence protected) by the
811 buddy allocator. Bigger value increase the probability
812 of catching random memory corruption, but reduce the
813 amount of memory for normal system use. The maximum
814 possible value is MAX_ORDER/2. Setting this parameter
815 to 1 or 2 should be enough to identify most random
816 memory corruption problems caused by bugs in kernel or
817 driver code when a CPU writes to (or reads from) a
818 random memory location. Note that there exists a class
819 of memory corruptions problems caused by buggy H/W or
820 F/W or by drivers badly programing DMA (basically when
821 memory is written at bus level and the CPU MMU is
822 bypassed) which are not detectable by
823 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
824 tracking down these problems.
827 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
828 enables the feature at boot time. By default, it is
829 disabled and the system will work mostly the same as a
830 kernel built without CONFIG_DEBUG_PAGEALLOC.
831 Note: to get most of debug_pagealloc error reports, it's
832 useful to also enable the page_owner functionality.
833 on: enable the feature
835 debugfs= [KNL] This parameter enables what is exposed to userspace
836 and debugfs internal clients.
837 Format: { on, no-mount, off }
838 on: All functions are enabled.
840 Filesystem is not registered but kernel clients can
841 access APIs and a crashkernel can be used to read
842 its content. There is nothing to mount.
843 off: Filesystem is not registered and clients
844 get a -EPERM as result when trying to register files
845 or directories within debugfs.
846 This is equivalent of the runtime functionality if
847 debugfs was not enabled in the kernel at all.
848 Default value is set in build-time with a kernel configuration.
850 debugpat [X86] Enable PAT debugging
852 decnet.addr= [HW,NET]
853 Format: <area>[,<node>]
854 See also Documentation/networking/decnet.rst.
857 [HW] The size of the default HugeTLB page. This is
858 the size represented by the legacy /proc/ hugepages
859 APIs. In addition, this is the default hugetlb size
860 used for shmget(), mmap() and mounting hugetlbfs
861 filesystems. If not specified, defaults to the
862 architecture's default huge page size. Huge page
863 sizes are architecture dependent. See also
864 Documentation/admin-guide/mm/hugetlbpage.rst.
867 deferred_probe_timeout=
868 [KNL] Debugging option to set a timeout in seconds for
869 deferred probe to give up waiting on dependencies to
870 probe. Only specific dependencies (subsystems or
871 drivers) that have opted in will be ignored. A timeout of 0
872 will timeout at the end of initcalls. This option will also
873 dump out devices still on the deferred probe list after
877 Format: { on | off | def_only | inf_only | always }
878 on: s390 zlib hardware support for compression on
879 level 1 and decompression (default)
880 off: No s390 zlib hardware support
881 def_only: s390 zlib hardware support for deflate
882 only (compression on level 1)
883 inf_only: s390 zlib hardware support for inflate
885 always: Same as 'on' but ignores the selected compression
886 level always using hardware support (used for debugging)
889 Set number of hash buckets for dentry cache.
891 disable_1tb_segments [PPC]
892 Disables the use of 1TB hash page table segments. This
893 causes the kernel to fall back to 256MB segments which
894 can be useful when debugging issues that require an SLB
898 Limits the number of kernel SLB entries, and flushes
899 them frequently to increase the rate of SLB faults
903 See Documentation/networking/ipv6.rst.
906 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
907 hardening is enabled for this boot. Hardened
908 usercopy checking is used to protect the kernel
909 from reading or writing beyond known memory
910 allocation boundaries as a proactive defense
911 against bounds-checking flaws in the kernel's
912 copy_to_user()/copy_from_user() interface.
913 on Perform hardened usercopy checks (default).
914 off Disable hardened usercopy checks.
917 Disable RADIX MMU mode on POWER9
919 radix_hcall_invalidate=on [PPC/PSERIES]
920 Disable RADIX GTSE feature and use hcall for TLB
924 Disable TLBIE instruction. Currently does not work
925 with KVM, with HASH MMU, or with coherent accelerators.
927 disable_cpu_apicid= [X86,APIC,SMP]
929 The number of initial APIC ID for the
930 corresponding CPU to be disabled at boot,
931 mostly used for the kdump 2nd kernel to
932 disable BSP to wake up multiple CPUs without
933 causing system reset or hang due to sending
936 perf_v4_pmi= [X86,INTEL]
938 Disable Intel PMU counter freezing feature.
939 The feature only exists starting from
940 Arch Perfmon v4 (Skylake and newer).
942 disable_ddw [PPC/PSERIES]
943 Disable Dynamic DMA Window support. Use this
944 to workaround buggy firmware.
947 See Documentation/networking/ipv6.rst.
949 disable_mtrr_cleanup [X86]
950 The kernel tries to adjust MTRR layout from continuous
951 to discrete, to make X server driver able to add WB
952 entry later. This parameter disables that.
954 disable_mtrr_trim [X86, Intel and AMD only]
955 By default the kernel will trim any uncacheable
956 memory out of your available memory pool based on
957 MTRR settings. This parameter disables that behavior,
958 possibly causing your machine to run very slowly.
960 disable_timer_pin_1 [X86]
961 Disable PIN 1 of APIC timer
962 Can be useful to work around chipset bugs.
964 dis_ucode_ldr [X86] Disable the microcode loader.
966 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
967 this option disables the debugging code at boot.
969 dma_debug_entries=<number>
970 This option allows to tune the number of preallocated
971 entries for DMA-API debugging code. One entry is
972 required per DMA-API allocation. Use this if the
973 DMA-API debugging code disables itself because the
974 architectural default is too low.
976 dma_debug_driver=<driver_name>
977 With this option the DMA-API debugging driver
978 filter feature can be enabled at boot time. Just
979 pass the driver to filter for as the parameter.
980 The filter can be disabled or changed to another
981 driver later using sysfs.
983 driver_async_probe= [KNL]
984 List of driver names to be probed asynchronously.
985 Format: <driver_name1>,<driver_name2>...
987 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
988 Broken monitors, graphic adapters, KVMs and EDIDless
989 panels may send no or incorrect EDID data sets.
990 This parameter allows to specify an EDID data sets
991 in the /lib/firmware directory that are used instead.
992 Generic built-in EDID data sets are used, if one of
993 edid/1024x768.bin, edid/1280x1024.bin,
994 edid/1680x1050.bin, or edid/1920x1080.bin is given
995 and no file with the same name exists. Details and
996 instructions how to build your own EDID data are
997 available in Documentation/admin-guide/edid.rst. An EDID
998 data set will only be used for a particular connector,
999 if its name and a colon are prepended to the EDID
1000 name. Each connector may use a unique EDID data
1001 set by separating the files with a comma. An EDID
1002 data set with no connector name will be used for
1003 any connectors not explicitly specified.
1008 Format: {"off" | "known"}
1009 Control how the dt_cpu_ftrs device-tree binding is
1010 used for CPU feature discovery and setup (if it
1012 off: Do not use it, fall back to legacy cpu table.
1013 known: Do not pass through unknown features to guests
1014 or userspace, only those that the kernel is aware of.
1016 dump_apple_properties [X86]
1017 Dump name and content of EFI device properties on
1018 x86 Macs. Useful for driver authors to determine
1019 what data is available or for reverse-engineering.
1021 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1022 <module>.dyndbg[="val"]
1023 Enable debug messages at boot time. See
1024 Documentation/admin-guide/dynamic-debug-howto.rst
1027 nopku [X86] Disable Memory Protection Keys CPU feature found
1030 <module>.async_probe [KNL]
1031 Enable asynchronous probe on this module.
1033 early_ioremap_debug [KNL]
1034 Enable debug messages in early_ioremap support. This
1035 is useful for tracking down temporary early mappings
1036 which are not unmapped.
1038 earlycon= [KNL] Output early console device and options.
1040 When used with no options, the early console is
1041 determined by stdout-path property in device tree's
1042 chosen node or the ACPI SPCR table if supported by
1045 cdns,<addr>[,options]
1046 Start an early, polled-mode console on a Cadence
1047 (xuartps) serial port at the specified address. Only
1048 supported option is baud rate. If baud rate is not
1049 specified, the serial port must already be setup and
1052 uart[8250],io,<addr>[,options]
1053 uart[8250],mmio,<addr>[,options]
1054 uart[8250],mmio32,<addr>[,options]
1055 uart[8250],mmio32be,<addr>[,options]
1056 uart[8250],0x<addr>[,options]
1057 Start an early, polled-mode console on the 8250/16550
1058 UART at the specified I/O port or MMIO address.
1059 MMIO inter-register address stride is either 8-bit
1060 (mmio) or 32-bit (mmio32 or mmio32be).
1061 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1062 to be equivalent to 'mmio'. 'options' are specified
1063 in the same format described for "console=ttyS<n>"; if
1064 unspecified, the h/w is not initialized.
1068 Start an early, polled-mode console on a pl011 serial
1069 port at the specified address. The pl011 serial port
1070 must already be setup and configured. Options are not
1071 yet supported. If 'mmio32' is specified, then only
1072 the driver will use only 32-bit accessors to read/write
1073 the device registers.
1076 Start an early, polled-mode console on a meson serial
1077 port at the specified address. The serial port must
1078 already be setup and configured. Options are not yet
1082 Start an early, polled-mode console on an msm serial
1083 port at the specified address. The serial port
1084 must already be setup and configured. Options are not
1087 msm_serial_dm,<addr>
1088 Start an early, polled-mode console on an msm serial
1089 dm port at the specified address. The serial port
1090 must already be setup and configured. Options are not
1094 Start an early, polled-mode console on a serial port
1095 of an Actions Semi SoC, such as S500 or S900, at the
1096 specified address. The serial port must already be
1097 setup and configured. Options are not yet supported.
1100 Start an early, polled-mode console on a serial port
1101 of an RDA Micro SoC, such as RDA8810PL, at the
1102 specified address. The serial port must already be
1103 setup and configured. Options are not yet supported.
1106 Use RISC-V SBI (Supervisor Binary Interface) for early
1109 smh Use ARM semihosting calls for early console.
1117 Use early console provided by serial driver available
1118 on Samsung SoCs, requires selecting proper type and
1119 a correct base address of the selected UART port. The
1120 serial port must already be setup and configured.
1121 Options are not yet supported.
1124 Start an early, polled-mode console on a lantiq serial
1125 (lqasc) port at the specified address. The serial port
1126 must already be setup and configured. Options are not
1131 Use early console provided by Freescale LP UART driver
1132 found on Freescale Vybrid and QorIQ LS1021A processors.
1133 A valid base address must be provided, and the serial
1134 port must already be setup and configured.
1138 Start an early, polled-mode, output-only console on the
1139 Freescale i.MX UART at the specified address. The UART
1140 must already be setup and configured.
1143 Start an early, polled-mode console on the
1144 Armada 3700 serial port at the specified
1145 address. The serial port must already be setup
1146 and configured. Options are not yet supported.
1149 Start an early, polled-mode console on a Qualcomm
1150 Generic Interface (GENI) based serial port at the
1151 specified address. The serial port must already be
1152 setup and configured. Options are not yet supported.
1155 Start an early, unaccelerated console on the EFI
1156 memory mapped framebuffer (if available). On cache
1157 coherent non-x86 systems that use system memory for
1158 the framebuffer, pass the 'ram' option so that it is
1159 mapped with the correct attributes.
1162 Use early console provided by Freescale LINFlexD UART
1163 serial driver for NXP S32V234 SoCs. A valid base
1164 address must be provided, and the serial port must
1165 already be setup and configured.
1167 earlyprintk= [X86,SH,ARM,M68k,S390]
1171 earlyprintk=serial[,ttySn[,baudrate]]
1172 earlyprintk=serial[,0x...[,baudrate]]
1173 earlyprintk=ttySn[,baudrate]
1174 earlyprintk=dbgp[debugController#]
1175 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1176 earlyprintk=xdbc[xhciController#]
1178 earlyprintk is useful when the kernel crashes before
1179 the normal console is initialized. It is not enabled by
1180 default because it has some cosmetic problems.
1182 Append ",keep" to not disable it when the real console
1185 Only one of vga, efi, serial, or usb debug port can
1188 Currently only ttyS0 and ttyS1 may be specified by
1189 name. Other I/O ports may be explicitly specified
1190 on some architectures (x86 and arm at least) by
1191 replacing ttySn with an I/O port address, like this:
1192 earlyprintk=serial,0x1008,115200
1193 You can find the port for a given device in
1194 /proc/tty/driver/serial:
1195 2: uart:ST16650V2 port:00001008 irq:18 ...
1197 Interaction with the standard serial driver is not
1200 The VGA and EFI output is eventually overwritten by
1203 The xen output can only be used by Xen PV guests.
1205 The sclp output can only be used on s390.
1207 The optional "force" to "pciserial" enables use of a
1208 PCI device even when its classcode is not of the
1211 edac_report= [HW,EDAC] Control how to report EDAC event
1212 Format: {"on" | "off" | "force"}
1213 on: enable EDAC to report H/W event. May be overridden
1214 by other higher priority error reporting module.
1215 off: disable H/W event reporting through EDAC.
1216 force: enforce the use of EDAC to report H/W event.
1219 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1222 This is designed to be used in conjunction with
1223 the boot argument: earlyprintk=vga
1225 This parameter works in place of the kgdboc parameter
1226 but can only be used if the backing tty is available
1227 very early in the boot process. For early debugging
1228 via a serial port see kgdboc_earlycon instead.
1231 Format: {"off" | "on" | "skip[mbr]"}
1234 Format: { "debug", "disable_early_pci_dma",
1235 "nochunk", "noruntime", "nosoftreserve",
1236 "novamap", "no_disable_early_pci_dma" }
1237 debug: enable misc debug output.
1238 disable_early_pci_dma: disable the busmaster bit on all
1239 PCI bridges while in the EFI boot stub.
1240 nochunk: disable reading files in "chunks" in the EFI
1241 boot stub, as chunking can cause problems with some
1242 firmware implementations.
1243 noruntime : disable EFI runtime services support
1244 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1245 attribute may cause the kernel to reserve the
1246 memory range for a memory mapping driver to
1247 claim. Specify efi=nosoftreserve to disable this
1248 reservation and treat the memory by its base type
1249 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1250 novamap: do not call SetVirtualAddressMap().
1251 no_disable_early_pci_dma: Leave the busmaster bit set
1252 on all PCI bridges while in the EFI boot stub
1254 efi_no_storage_paranoia [EFI; X86]
1255 Using this parameter you can use more than 50% of
1256 your efi variable storage. Use this parameter only if
1257 you are really sure that your UEFI does sane gc and
1258 fulfills the spec otherwise your board may brick.
1260 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1261 Add arbitrary attribute to specific memory range by
1262 updating original EFI memory map.
1263 Region of memory which aa attribute is added to is
1266 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1267 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1268 attribute is added to range 0x100000000-0x180000000 and
1269 0x10a0000000-0x1120000000.
1271 If efi_fake_mem=8G@9G:0x40000 is specified, the
1272 EFI_MEMORY_SP(0x40000) attribute is added to
1273 range 0x240000000-0x43fffffff.
1275 Using this parameter you can do debugging of EFI memmap
1276 related features. For example, you can do debugging of
1277 Address Range Mirroring feature even if your box
1278 doesn't support it, or mark specific memory as
1281 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1282 that is to be dynamically loaded by Linux. If there are
1283 multiple variables with the same name but with different
1284 vendor GUIDs, all of them will be loaded. See
1285 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1288 eisa_irq_edge= [PARISC,HW]
1289 See header of drivers/parisc/eisa.c.
1292 See comment before function elanfreq_setup() in
1293 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1295 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1296 Specifies physical address of start of kernel core
1297 image elf header and optionally the size. Generally
1298 kexec loader will pass this option to capture kernel.
1299 See Documentation/admin-guide/kdump/kdump.rst for details.
1301 enable_mtrr_cleanup [X86]
1302 The kernel tries to adjust MTRR layout from continuous
1303 to discrete, to make X server driver able to add WB
1304 entry later. This parameter enables that.
1306 enable_timer_pin_1 [X86]
1307 Enable PIN 1 of APIC timer
1308 Can be useful to work around chipset bugs
1309 (in particular on some ATI chipsets).
1310 The kernel tries to set a reasonable default.
1312 enforcing [SELINUX] Set initial enforcing status.
1314 See security/selinux/Kconfig help text.
1315 0 -- permissive (log only, no denials).
1316 1 -- enforcing (deny and log).
1318 Value can be changed at runtime via
1319 /sys/fs/selinux/enforce.
1322 Disable Error Record Serialization Table (ERST)
1325 ether= [HW,NET] Ethernet cards parameters
1326 This option is obsoleted by the "netdev=" option, which
1327 has equivalent usage. See its documentation for details.
1331 Permit 'security.evm' to be updated regardless of
1332 current integrity status.
1336 fail_make_request=[KNL]
1337 General fault injection mechanism.
1338 Format: <interval>,<probability>,<space>,<times>
1339 See also Documentation/fault-injection/.
1342 See Documentation/admin-guide/blockdev/floppy.rst.
1344 force_pal_cache_flush
1345 [IA-64] Avoid check_sal_cache_flush which may hang on
1346 buggy SAL_CACHE_FLUSH implementations. Using this
1347 parameter will force ia64_sal_cache_flush to call
1348 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1351 Forcefully enable Physical Address Extension (PAE).
1352 Many Pentium M systems disable PAE but may have a
1353 functionally usable PAE implementation.
1354 Warning: use of this parameter will taint the kernel
1355 and may cause unknown problems.
1358 [FTRACE] will set and start the specified tracer
1359 as early as possible in order to facilitate early
1362 ftrace_dump_on_oops[=orig_cpu]
1363 [FTRACE] will dump the trace buffers on oops.
1364 If no parameter is passed, ftrace will dump
1365 buffers of all CPUs, but if you pass orig_cpu, it will
1366 dump only the buffer of the CPU that triggered the
1369 ftrace_filter=[function-list]
1370 [FTRACE] Limit the functions traced by the function
1371 tracer at boot up. function-list is a comma separated
1372 list of functions. This list can be changed at run
1373 time by the set_ftrace_filter file in the debugfs
1376 ftrace_notrace=[function-list]
1377 [FTRACE] Do not trace the functions specified in
1378 function-list. This list can be changed at run time
1379 by the set_ftrace_notrace file in the debugfs
1382 ftrace_graph_filter=[function-list]
1383 [FTRACE] Limit the top level callers functions traced
1384 by the function graph tracer at boot up.
1385 function-list is a comma separated list of functions
1386 that can be changed at run time by the
1387 set_graph_function file in the debugfs tracing directory.
1389 ftrace_graph_notrace=[function-list]
1390 [FTRACE] Do not trace from the functions specified in
1391 function-list. This list is a comma separated list of
1392 functions that can be changed at run time by the
1393 set_graph_notrace file in the debugfs tracing directory.
1395 ftrace_graph_max_depth=<uint>
1396 [FTRACE] Used with the function graph tracer. This is
1397 the max depth it will trace into a function. This value
1398 can be changed at run time by the max_graph_depth file
1399 in the tracefs tracing directory. default: 0 (no limit)
1401 fw_devlink= [KNL] Create device links between consumer and supplier
1402 devices by scanning the firmware to infer the
1403 consumer/supplier relationships. This feature is
1404 especially useful when drivers are loaded as modules as
1405 it ensures proper ordering of tasks like device probing
1406 (suppliers first, then consumers), supplier boot state
1407 clean up (only after all consumers have probed),
1408 suspend/resume & runtime PM (consumers first, then
1410 Format: { off | permissive | on | rpm }
1411 off -- Don't create device links from firmware info.
1412 permissive -- Create device links from firmware info
1413 but use it only for ordering boot state clean
1414 up (sync_state() calls).
1415 on -- Create device links from firmware info and use it
1416 to enforce probe and suspend/resume ordering.
1417 rpm -- Like "on", but also use to order runtime PM.
1420 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1421 support via parallel port (up to 5 devices per port)
1422 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1423 See also Documentation/input/devices/joystick-parport.rst
1427 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1431 gcov_persist= [GCOV] When non-zero (default), profiling data for
1432 kernel modules is saved and remains accessible via
1433 debugfs, even when the module is unloaded/reloaded.
1434 When zero, profiling data is discarded and associated
1435 debugfs files are removed at module unload time.
1437 goldfish [X86] Enable the goldfish android emulator platform.
1438 Don't use this when you are not running on the
1441 gpt [EFI] Forces disk with valid GPT signature but
1442 invalid Protective MBR to be treated as GPT. If the
1443 primary GPT is corrupted, it enables the backup/alternate
1444 GPT to be used instead.
1446 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1447 the "Enable 0" bit of the configuration register.
1450 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1451 the "Enable 0" bit of the configuration register.
1454 grcan.select= [HW] Select which physical interface to use.
1457 grcan.txsize= [HW] Sets the size of the tx buffer.
1458 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1460 grcan.rxsize= [HW] Sets the size of the rx buffer.
1461 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1464 gpio-mockup.gpio_mockup_ranges
1465 [HW] Sets the ranges of gpiochip of for this device.
1466 Format: <start1>,<end1>,<start2>,<end2>...
1468 hardlockup_all_cpu_backtrace=
1469 [KNL] Should the hard-lockup detector generate
1470 backtraces on all cpus.
1473 hashdist= [KNL,NUMA] Large hashes allocated during boot
1474 are distributed across NUMA nodes. Defaults on
1475 for 64-bit NUMA, off otherwise.
1476 Format: 0 | 1 (for off | on)
1478 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1480 hd= [EIDE] (E)IDE hard drive subsystem geometry
1481 Format: <cyl>,<head>,<sect>
1484 Disable Hardware Error Source Table (HEST) support;
1485 corresponding firmware-first mode error processing
1486 logic will be disabled.
1488 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1489 size of <nn>. This works even on boxes that have no
1490 highmem otherwise. This also works to reduce highmem
1491 size on bigger boxes.
1493 highres= [KNL] Enable/disable high resolution timer mode.
1494 Valid parameters: "on", "off"
1499 hpet= [X86-32,HPET] option to control HPET usage
1500 Format: { enable (default) | disable | force |
1502 disable: disable HPET and use PIT instead
1503 force: allow force enabled of undocumented chips (ICH4,
1505 verbose: show contents of HPET registers during setup
1507 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1508 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1510 hugetlb_cma= [HW] The size of a cma area used for allocation
1511 of gigantic hugepages.
1514 Reserve a cma area of given size and allocate gigantic
1515 hugepages using the cma allocator. If enabled, the
1516 boot-time allocation of gigantic hugepages is skipped.
1518 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1519 If this follows hugepagesz (below), it specifies
1520 the number of pages of hugepagesz to be allocated.
1521 If this is the first HugeTLB parameter on the command
1522 line, it specifies the number of pages to allocate for
1523 the default huge page size. See also
1524 Documentation/admin-guide/mm/hugetlbpage.rst.
1528 [HW] The size of the HugeTLB pages. This is used in
1529 conjunction with hugepages (above) to allocate huge
1530 pages of a specific size at boot. The pair
1531 hugepagesz=X hugepages=Y can be specified once for
1532 each supported huge page size. Huge page sizes are
1533 architecture dependent. See also
1534 Documentation/admin-guide/mm/hugetlbpage.rst.
1538 [KNL] Should the hung task detector generate panics.
1541 A value of 1 instructs the kernel to panic when a
1542 hung task is detected. The default value is controlled
1543 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1544 option. The value selected by this boot parameter can
1545 be changed later by the kernel.hung_task_panic sysctl.
1547 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1548 terminal devices. Valid values: 0..8
1549 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1550 If specified, z/VM IUCV HVC accepts connections
1551 from listed z/VM user IDs only.
1553 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1554 which allow the hypervisor to 'idle' the
1555 guest on lock contention.
1558 Do not unregister boot console at start. This is only
1559 useful for debugging when something happens in the window
1560 between unregistering the boot console and initializing
1563 i2c_bus= [HW] Override the default board specific I2C bus speed
1564 or register an additional I2C bus that is not
1565 registered from board initialization code.
1569 i8042.debug [HW] Toggle i8042 debug mode
1570 i8042.unmask_kbd_data
1571 [HW] Enable printing of interrupt data from the KBD port
1572 (disabled by default, and as a pre-condition
1573 requires that i8042.debug=1 be enabled)
1574 i8042.direct [HW] Put keyboard port into non-translated mode
1575 i8042.dumbkbd [HW] Pretend that controller can only read data from
1576 keyboard and cannot control its state
1577 (Don't attempt to blink the leds)
1578 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1579 i8042.nokbd [HW] Don't check/create keyboard port
1580 i8042.noloop [HW] Disable the AUX Loopback command while probing
1582 i8042.nomux [HW] Don't check presence of an active multiplexing
1584 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1586 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1587 i8042.reset [HW] Reset the controller during init, cleanup and
1588 suspend-to-ram transitions, only during s2r
1589 transitions, or never reset
1590 Format: { 1 | Y | y | 0 | N | n }
1591 1, Y, y: always reset controller
1592 0, N, n: don't ever reset controller
1593 Default: only on s2r transitions on x86; most other
1594 architectures force reset to be always executed
1595 i8042.unlock [HW] Unlock (ignore) the keylock
1596 i8042.kbdreset [HW] Reset device connected to KBD port
1600 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1601 indicates that the driver is running on unsupported
1603 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1604 does not match list of supported models.
1606 [HW] Report power status in /proc/i8k
1607 (disabled by default)
1608 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1611 i915.invert_brightness=
1612 [DRM] Invert the sense of the variable that is used to
1613 set the brightness of the panel backlight. Normally a
1614 brightness value of 0 indicates backlight switched off,
1615 and the maximum of the brightness value sets the backlight
1616 to maximum brightness. If this parameter is set to 0
1617 (default) and the machine requires it, or this parameter
1618 is set to 1, a brightness value of 0 sets the backlight
1619 to maximum brightness, and the maximum of the brightness
1620 value switches the backlight off.
1621 -1 -- never invert brightness
1622 0 -- machine default
1623 1 -- force brightness inversion
1626 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1628 ide-core.nodma= [HW] (E)IDE subsystem
1629 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1630 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1631 .cdrom .chs .ignore_cable are additional options
1632 See Documentation/ide/ide.rst.
1634 ide-generic.probe-mask= [HW] (E)IDE subsystem
1636 Probe mask for legacy ISA IDE ports. Depending on
1637 platform up to 6 ports are supported, enabled by
1638 setting corresponding bits in the mask to 1. The
1639 default value is 0x0, which has a special meaning.
1640 On systems that have PCI, it triggers scanning the
1641 PCI bus for the first and the second port, which
1642 are then probed. On systems without PCI the value
1643 of 0x0 enables probing the two first ports as if it
1646 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1647 Claim all unknown PCI IDE storage controllers.
1650 Format: idle=poll, idle=halt, idle=nomwait
1651 Poll forces a polling idle loop that can slightly
1652 improve the performance of waking up a idle CPU, but
1653 will use a lot of power and make the system run hot.
1655 idle=halt: Halt is forced to be used for CPU idle.
1656 In such case C2/C3 won't be used again.
1657 idle=nomwait: Disable mwait for CPU C-states
1659 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1660 Format: { strict | legacy | 2008 | relaxed }
1663 Choose which programs will be accepted for execution
1664 based on the IEEE 754 NaN encoding(s) supported by
1665 the FPU and the NaN encoding requested with the value
1666 of an ELF file header flag individually set by each
1667 binary. Hardware implementations are permitted to
1668 support either or both of the legacy and the 2008 NaN
1671 Available settings are as follows:
1672 strict accept binaries that request a NaN encoding
1673 supported by the FPU
1674 legacy only accept legacy-NaN binaries, if supported
1676 2008 only accept 2008-NaN binaries, if supported
1678 relaxed accept any binaries regardless of whether
1679 supported by the FPU
1681 The FPU emulator is always able to support both NaN
1682 encodings, so if no FPU hardware is present or it has
1683 been disabled with 'nofpu', then the settings of
1684 'legacy' and '2008' strap the emulator accordingly,
1685 'relaxed' straps the emulator for both legacy-NaN and
1686 2008-NaN, whereas 'strict' enables legacy-NaN only on
1687 legacy processors and both NaN encodings on MIPS32 or
1690 The setting for ABS.fmt/NEG.fmt instruction execution
1691 mode generally follows that for the NaN encoding,
1692 except where unsupported by hardware.
1694 ignore_loglevel [KNL]
1695 Ignore loglevel setting - this will print /all/
1696 kernel messages to the console. Useful for debugging.
1697 We also add it as printk module parameter, so users
1698 could change it dynamically, usually by
1699 /sys/module/printk/parameters/ignore_loglevel.
1702 Ignore RLIMIT_DATA setting for data mappings,
1703 print warning at first misuse. Can be changed via
1704 /sys/module/kernel/parameters/ignore_rlimit_data.
1706 ihash_entries= [KNL]
1707 Set number of hash buckets for inode cache.
1709 ima_appraise= [IMA] appraise integrity measurements
1710 Format: { "off" | "enforce" | "fix" | "log" }
1713 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1714 The builtin appraise policy appraises all files
1717 ima_canonical_fmt [IMA]
1718 Use the canonical format for the binary runtime
1719 measurements, instead of host native format.
1722 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1726 The list of supported hash algorithms is defined
1727 in crypto/hash_info.h.
1730 The builtin policies to load during IMA setup.
1731 Format: "tcb | appraise_tcb | secure_boot |
1734 The "tcb" policy measures all programs exec'd, files
1735 mmap'd for exec, and all files opened with the read
1736 mode bit set by either the effective uid (euid=0) or
1739 The "appraise_tcb" policy appraises the integrity of
1740 all files owned by root.
1742 The "secure_boot" policy appraises the integrity
1743 of files (eg. kexec kernel image, kernel modules,
1744 firmware, policy, etc) based on file signatures.
1746 The "fail_securely" policy forces file signature
1747 verification failure also on privileged mounted
1748 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1751 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1752 Load a policy which meets the needs of the Trusted
1753 Computing Base. This means IMA will measure all
1754 programs exec'd, files mmap'd for exec, and all files
1755 opened for read by uid=0.
1758 Select one of defined IMA measurements template formats.
1759 Formats: { "ima" | "ima-ng" | "ima-sig" }
1763 [IMA] Define a custom template format.
1764 Format: { "field1|...|fieldN" }
1766 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1767 Format: <min_file_size>
1768 Set the minimal file size for using asynchronous hash.
1769 If left unspecified, ahash usage is disabled.
1771 ahash performance varies for different data sizes on
1772 different crypto accelerators. This option can be used
1773 to achieve the best performance for a particular HW.
1775 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1777 Set hashing buffer size. Default: 4k.
1779 ahash performance varies for different chunk sizes on
1780 different crypto accelerators. This option can be used
1781 to achieve best performance for particular HW.
1785 Run specified binary instead of /sbin/init as init
1788 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1789 for working out where the kernel is dying during
1792 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1793 initcall functions. Useful for debugging built-in
1794 modules and initcalls.
1796 initrd= [BOOT] Specify the location of the initial ramdisk
1798 initrdmem= [KNL] Specify a physical address and size from which to
1799 load the initrd. If an initrd is compiled in or
1800 specified in the bootparams, it takes priority over this
1802 Format: ss[KMG],nn[KMG]
1805 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1808 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1810 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1812 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1814 init_pkru= [X86] Specify the default memory protection keys rights
1815 register contents for all processes. 0x55555554 by
1816 default (disallow access to all but pkey 0). Can
1817 override in debugfs after boot.
1819 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1822 int_pln_enable [X86] Enable power limit notification interrupt
1824 integrity_audit=[IMA]
1825 Format: { "0" | "1" }
1826 0 -- basic integrity auditing messages. (Default)
1827 1 -- additional integrity auditing messages.
1829 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1831 Enable intel iommu driver.
1833 Disable intel iommu driver.
1834 igfx_off [Default Off]
1835 By default, gfx is mapped as normal device. If a gfx
1836 device has a dedicated DMAR unit, the DMAR unit is
1837 bypassed by not enabling DMAR with this option. In
1838 this case, gfx device will use physical address for
1841 With this option iommu will not optimize to look
1842 for io virtual address below 32-bit forcing dual
1843 address cycle on pci bus for cards supporting greater
1844 than 32-bit addressing. The default is to look
1845 for translation below 32-bit and if not available
1846 then look in the higher range.
1847 strict [Default Off]
1848 With this option on every unmap_single operation will
1849 result in a hardware IOTLB flush operation as opposed
1850 to batching them for performance.
1851 sp_off [Default Off]
1852 By default, super page will be supported if Intel IOMMU
1853 has the capability. With this option, super page will
1856 By default, scalable mode will be disabled even if the
1857 hardware advertises that it has support for the scalable
1858 mode translation. With this option set, scalable mode
1859 will be used on hardware which claims to support it.
1860 tboot_noforce [Default Off]
1861 Do not force the Intel IOMMU enabled under tboot.
1862 By default, tboot will force Intel IOMMU on, which
1863 could harm performance of some high-throughput
1864 devices like 40GBit network cards, even if identity
1866 Note that using this option lowers the security
1867 provided by tboot because it makes the system
1868 vulnerable to DMA attacks.
1869 nobounce [Default off]
1870 Disable bounce buffer for untrusted devices such as
1871 the Thunderbolt devices. This will treat the untrusted
1872 devices as the trusted ones, hence might expose security
1873 risks of DMA attacks.
1875 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1876 0 disables intel_idle and fall back on acpi_idle.
1877 1 to 9 specify maximum depth of C-state.
1881 Do not enable intel_pstate as the default
1882 scaling driver for the supported processors
1884 Use intel_pstate as a scaling driver, but configure it
1885 to work with generic cpufreq governors (instead of
1886 enabling its internal governor). This mode cannot be
1887 used along with the hardware-managed P-states (HWP)
1890 Enable intel_pstate on systems that prohibit it by default
1891 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1892 instead of acpi-cpufreq may disable platform features, such
1893 as thermal controls and power capping, that rely on ACPI
1894 P-States information being indicated to OSPM and therefore
1895 should be used with caution. This option does not work with
1896 processors that aren't supported by the intel_pstate driver
1897 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1899 Do not enable hardware P state control (HWP)
1902 Only load intel_pstate on systems which support
1903 hardware P state control (HWP) if available.
1905 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1906 Description Table, specifies preferred power management
1907 profile as "Enterprise Server" or "Performance Server",
1908 then this feature is turned on by default.
1910 Allow per-logical-CPU P-State performance control limits using
1911 cpufreq sysfs interface
1913 intremap= [X86-64, Intel-IOMMU]
1914 on enable Interrupt Remapping (default)
1915 off disable Interrupt Remapping
1916 nosid disable Source ID checking
1918 BIOS x2APIC opt-out request will be ignored
1919 nopost disable Interrupt Posting
1921 iomem= Disable strict checking of access to MMIO memory
1922 strict regions from userspace.
1937 nobypass [PPC/POWERNV]
1938 Disable IOMMU bypass, using IOMMU for PCI devices.
1940 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1941 Format: { "0" | "1" }
1943 Request that DMA unmap operations use deferred
1944 invalidation of hardware TLBs, for increased
1945 throughput at the cost of reduced device isolation.
1946 Will fall back to strict mode if not supported by
1947 the relevant IOMMU driver.
1948 1 - Strict mode (default).
1949 DMA unmap operations invalidate IOMMU hardware TLBs
1953 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
1954 Format: { "0" | "1" }
1955 0 - Use IOMMU translation for DMA.
1956 1 - Bypass the IOMMU for DMA.
1957 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
1959 io7= [HW] IO7 for Marvel-based Alpha systems
1960 See comment before marvel_specify_io7 in
1961 arch/alpha/kernel/core_marvel.c.
1963 io_delay= [X86] I/O delay method
1965 Standard port 0x80 based delay
1967 Alternate port 0xed based delay (needed on some systems)
1969 Simple two microseconds delay
1974 See Documentation/admin-guide/nfs/nfsroot.rst.
1976 ipcmni_extend [KNL] Extend the maximum number of unique System V
1977 IPC identifiers from 32,768 to 16,777,216.
1979 irqaffinity= [SMP] Set the default irq affinity mask
1980 The argument is a cpu list, as described above.
1982 irqchip.gicv2_force_probe=
1985 Force the kernel to look for the second 4kB page
1986 of a GICv2 controller even if the memory range
1987 exposed by the device tree is too small.
1989 irqchip.gicv3_nolpi=
1991 Force the kernel to ignore the availability of
1992 LPIs (and by consequence ITSs). Intended for system
1993 that use the kernel as a bootloader, and thus want
1994 to let secondary kernels in charge of setting up
1997 irqchip.gicv3_pseudo_nmi= [ARM64]
1998 Enables support for pseudo-NMIs in the kernel. This
1999 requires the kernel to be built with
2000 CONFIG_ARM64_PSEUDO_NMI.
2003 When an interrupt is not handled search all handlers
2004 for it. Intended to get systems with badly broken
2008 When an interrupt is not handled search all handlers
2009 for it. Also check all handlers each timer
2010 interrupt. Intended to get systems with badly broken
2014 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2016 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2017 [Deprecated - use cpusets instead]
2018 Format: [flag-list,]<cpu-list>
2020 Specify one or more CPUs to isolate from disturbances
2021 specified in the flag list (default: domain):
2024 Disable the tick when a single task runs.
2026 A residual 1Hz tick is offloaded to workqueues, which you
2027 need to affine to housekeeping through the global
2028 workqueue's affinity configured via the
2029 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2030 by using the 'domain' flag described below.
2032 NOTE: by default the global workqueue runs on all CPUs,
2033 so to protect individual CPUs the 'cpumask' file has to
2034 be configured manually after bootup.
2037 Isolate from the general SMP balancing and scheduling
2038 algorithms. Note that performing domain isolation this way
2039 is irreversible: it's not possible to bring back a CPU to
2040 the domains once isolated through isolcpus. It's strongly
2041 advised to use cpusets instead to disable scheduler load
2042 balancing through the "cpuset.sched_load_balance" file.
2043 It offers a much more flexible interface where CPUs can
2044 move in and out of an isolated set anytime.
2046 You can move a process onto or off an "isolated" CPU via
2047 the CPU affinity syscalls or cpuset.
2048 <cpu number> begins at 0 and the maximum value is
2049 "number of CPUs in system - 1".
2053 Isolate from being targeted by managed interrupts
2054 which have an interrupt mask containing isolated
2055 CPUs. The affinity of managed interrupts is
2056 handled by the kernel and cannot be changed via
2057 the /proc/irq/* interfaces.
2059 This isolation is best effort and only effective
2060 if the automatically assigned interrupt mask of a
2061 device queue contains isolated and housekeeping
2062 CPUs. If housekeeping CPUs are online then such
2063 interrupts are directed to the housekeeping CPU
2064 so that IO submitted on the housekeeping CPU
2065 cannot disturb the isolated CPU.
2067 If a queue's affinity mask contains only isolated
2068 CPUs then this parameter has no effect on the
2069 interrupt routing decision, though interrupts are
2070 only delivered when tasks running on those
2071 isolated CPUs submit IO. IO submitted on
2072 housekeeping CPUs has no influence on those
2075 The format of <cpu-list> is described above.
2079 ivrs_ioapic [HW,X86-64]
2080 Provide an override to the IOAPIC-ID<->DEVICE-ID
2081 mapping provided in the IVRS ACPI table. For
2082 example, to map IOAPIC-ID decimal 10 to
2083 PCI device 00:14.0 write the parameter as:
2084 ivrs_ioapic[10]=00:14.0
2086 ivrs_hpet [HW,X86-64]
2087 Provide an override to the HPET-ID<->DEVICE-ID
2088 mapping provided in the IVRS ACPI table. For
2089 example, to map HPET-ID decimal 0 to
2090 PCI device 00:14.0 write the parameter as:
2091 ivrs_hpet[0]=00:14.0
2093 ivrs_acpihid [HW,X86-64]
2094 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2095 mapping provided in the IVRS ACPI table. For
2096 example, to map UART-HID:UID AMD0020:0 to
2097 PCI device 00:14.5 write the parameter as:
2098 ivrs_acpihid[00:14.5]=AMD0020:0
2100 js= [HW,JOY] Analog joystick
2101 See Documentation/input/joydev/joystick.rst.
2104 When CONFIG_RANDOMIZE_BASE is set, this disables
2105 kernel and module base offset ASLR (Address Space
2106 Layout Randomization).
2109 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2110 report on every invalid memory access. Without this
2111 parameter KASAN will print report only for the first
2116 kernelcore= [KNL,X86,IA-64,PPC]
2117 Format: nn[KMGTPE] | nn% | "mirror"
2118 This parameter specifies the amount of memory usable by
2119 the kernel for non-movable allocations. The requested
2120 amount is spread evenly throughout all nodes in the
2121 system as ZONE_NORMAL. The remaining memory is used for
2122 movable memory in its own zone, ZONE_MOVABLE. In the
2123 event, a node is too small to have both ZONE_NORMAL and
2124 ZONE_MOVABLE, kernelcore memory will take priority and
2125 other nodes will have a larger ZONE_MOVABLE.
2127 ZONE_MOVABLE is used for the allocation of pages that
2128 may be reclaimed or moved by the page migration
2129 subsystem. Note that allocations like PTEs-from-HighMem
2130 still use the HighMem zone if it exists, and the Normal
2131 zone if it does not.
2133 It is possible to specify the exact amount of memory in
2134 the form of "nn[KMGTPE]", a percentage of total system
2135 memory in the form of "nn%", or "mirror". If "mirror"
2136 option is specified, mirrored (reliable) memory is used
2137 for non-movable allocations and remaining memory is used
2138 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2139 are exclusive, so you cannot specify multiple forms.
2141 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2142 Format: <Controller#>[,poll interval]
2143 The controller # is the number of the ehci usb debug
2144 port as it is probed via PCI. The poll interval is
2145 optional and is the number seconds in between
2146 each poll cycle to the debug port in case you need
2147 the functionality for interrupting the kernel with
2148 gdb or control-c on the dbgp connection. When
2149 not using this parameter you use sysrq-g to break into
2150 the kernel debugger.
2152 kgdboc= [KGDB,HW] kgdb over consoles.
2153 Requires a tty driver that supports console polling,
2154 or a supported polling keyboard driver (non-usb).
2155 Serial only format: <serial_device>[,baud]
2156 keyboard only format: kbd
2157 keyboard and serial format: kbd,<serial_device>[,baud]
2158 Optional Kernel mode setting:
2159 kms, kbd format: kms,kbd
2160 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2162 kgdboc_earlycon= [KGDB,HW]
2163 If the boot console provides the ability to read
2164 characters and can work in polling mode, you can use
2165 this parameter to tell kgdb to use it as a backend
2166 until the normal console is registered. Intended to
2167 be used together with the kgdboc parameter which
2168 specifies the normal console to transition to.
2170 The name of the early console should be specified
2171 as the value of this parameter. Note that the name of
2172 the early console might be different than the tty
2173 name passed to kgdboc. It's OK to leave the value
2174 blank and the first boot console that implements
2175 read() will be picked.
2177 kgdbwait [KGDB] Stop kernel execution and enter the
2178 kernel debugger at the earliest opportunity.
2180 kmac= [MIPS] Korina ethernet MAC address.
2181 Configure the RouterBoard 532 series on-chip
2182 Ethernet adapter MAC address.
2184 kmemleak= [KNL] Boot-time kmemleak enable/disable
2185 Valid arguments: on, off
2187 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2190 kprobe_event=[probe-list]
2191 [FTRACE] Add kprobe events and enable at boot time.
2192 The probe-list is a semicolon delimited list of probe
2193 definitions. Each definition is same as kprobe_events
2194 interface, but the parameters are comma delimited.
2195 For example, to add a kprobe event on vfs_read with
2196 arg1 and arg2, add to the command line;
2198 kprobe_event=p,vfs_read,$arg1,$arg2
2200 See also Documentation/trace/kprobetrace.rst "Kernel
2201 Boot Parameter" section.
2203 kpti= [ARM64] Control page table isolation of user
2204 and kernel address spaces.
2205 Default: enabled on cores which need mitigation.
2209 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2210 Default is 0 (don't ignore, but inject #GP)
2212 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2213 Default is false (don't support).
2215 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2220 [KVM] Controls the software workaround for the
2221 X86_BUG_ITLB_MULTIHIT bug.
2222 force : Always deploy workaround.
2223 off : Never deploy workaround.
2224 auto : Deploy workaround based on the presence of
2225 X86_BUG_ITLB_MULTIHIT.
2229 If the software workaround is enabled for the host,
2230 guests do need not to enable it for nested guests.
2232 kvm.nx_huge_pages_recovery_ratio=
2233 [KVM] Controls how many 4KiB pages are periodically zapped
2234 back to huge pages. 0 disables the recovery, otherwise if
2235 the value is N KVM will zap 1/Nth of the 4KiB pages every
2236 minute. The default is 60.
2238 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2239 Default is 1 (enabled)
2241 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2243 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2245 kvm-arm.vgic_v3_group0_trap=
2246 [KVM,ARM] Trap guest accesses to GICv3 group-0
2249 kvm-arm.vgic_v3_group1_trap=
2250 [KVM,ARM] Trap guest accesses to GICv3 group-1
2253 kvm-arm.vgic_v3_common_trap=
2254 [KVM,ARM] Trap guest accesses to GICv3 common
2257 kvm-arm.vgic_v4_enable=
2258 [KVM,ARM] Allow use of GICv4 for direct injection of
2261 kvm_cma_resv_ratio=n [PPC]
2262 Reserves given percentage from system memory area for
2263 contiguous memory allocation for KVM hash pagetable
2265 By default it reserves 5% of total system memory.
2269 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2270 (virtualized MMU) support on capable Intel chips.
2271 Default is 1 (enabled)
2273 kvm-intel.emulate_invalid_guest_state=
2274 [KVM,Intel] Enable emulation of invalid guest states
2275 Default is 0 (disabled)
2277 kvm-intel.flexpriority=
2278 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2279 Default is 1 (enabled)
2282 [KVM,Intel] Enable VMX nesting (nVMX).
2283 Default is 0 (disabled)
2285 kvm-intel.unrestricted_guest=
2286 [KVM,Intel] Disable unrestricted guest feature
2287 (virtualized real and unpaged mode) on capable
2288 Intel chips. Default is 1 (enabled)
2290 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2293 Valid arguments: never, cond, always
2295 always: L1D cache flush on every VMENTER.
2296 cond: Flush L1D on VMENTER only when the code between
2297 VMEXIT and VMENTER can leak host memory.
2298 never: Disables the mitigation
2300 Default is cond (do L1 cache flush in specific instances)
2302 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2303 feature (tagged TLBs) on capable Intel chips.
2304 Default is 1 (enabled)
2306 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2309 The kernel PTE inversion protection is unconditionally
2310 enabled and cannot be disabled.
2313 Provides all available mitigations for the
2314 L1TF vulnerability. Disables SMT and
2315 enables all mitigations in the
2316 hypervisors, i.e. unconditional L1D flush.
2318 SMT control and L1D flush control via the
2319 sysfs interface is still possible after
2320 boot. Hypervisors will issue a warning
2321 when the first VM is started in a
2322 potentially insecure configuration,
2323 i.e. SMT enabled or L1D flush disabled.
2326 Same as 'full', but disables SMT and L1D
2327 flush runtime control. Implies the
2328 'nosmt=force' command line option.
2329 (i.e. sysfs control of SMT is disabled.)
2332 Leaves SMT enabled and enables the default
2333 hypervisor mitigation, i.e. conditional
2336 SMT control and L1D flush control via the
2337 sysfs interface is still possible after
2338 boot. Hypervisors will issue a warning
2339 when the first VM is started in a
2340 potentially insecure configuration,
2341 i.e. SMT enabled or L1D flush disabled.
2345 Disables SMT and enables the default
2346 hypervisor mitigation.
2348 SMT control and L1D flush control via the
2349 sysfs interface is still possible after
2350 boot. Hypervisors will issue a warning
2351 when the first VM is started in a
2352 potentially insecure configuration,
2353 i.e. SMT enabled or L1D flush disabled.
2356 Same as 'flush', but hypervisors will not
2357 warn when a VM is started in a potentially
2358 insecure configuration.
2361 Disables hypervisor mitigations and doesn't
2363 It also drops the swap size and available
2364 RAM limit restriction on both hypervisor and
2369 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2375 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2378 lapic= [X86,APIC] Do not use TSC deadline
2379 value for LAPIC timer one-shot implementation. Default
2380 back to the programmable timer unit in the LAPIC.
2381 Format: notscdeadline
2383 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2386 libata.dma= [LIBATA] DMA control
2387 libata.dma=0 Disable all PATA and SATA DMA
2388 libata.dma=1 PATA and SATA Disk DMA only
2389 libata.dma=2 ATAPI (CDROM) DMA only
2390 libata.dma=4 Compact Flash DMA only
2391 Combinations also work, so libata.dma=3 enables DMA
2392 for disks and CDROMs, but not CFs.
2394 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2395 libata.ignore_hpa=0 keep BIOS limits (default)
2396 libata.ignore_hpa=1 ignore limits, using full disk
2398 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2402 libata.force= [LIBATA] Force configurations. The format is comma
2403 separated list of "[ID:]VAL" where ID is
2404 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2405 matching port, link or device. Basically, it matches
2406 the ATA ID string printed on console by libata. If
2407 the whole ID part is omitted, the last PORT and DEVICE
2408 values are used. If ID hasn't been specified yet, the
2409 configuration applies to all ports, links and devices.
2411 If only DEVICE is omitted, the parameter applies to
2412 the port and all links and devices behind it. DEVICE
2413 number of 0 either selects the first device or the
2414 first fan-out link behind PMP device. It does not
2415 select the host link. DEVICE number of 15 selects the
2416 host link and device attached to it.
2418 The VAL specifies the configuration to force. As long
2419 as there's no ambiguity shortcut notation is allowed.
2420 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2421 The following configurations can be forced.
2423 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2424 Any ID with matching PORT is used.
2426 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2428 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2429 udma[/][16,25,33,44,66,100,133] notation is also
2432 * [no]ncq: Turn on or off NCQ.
2434 * [no]ncqtrim: Turn off queued DSM TRIM.
2436 * nohrst, nosrst, norst: suppress hard, soft
2439 * rstonce: only attempt one reset during
2440 hot-unplug link recovery
2442 * dump_id: dump IDENTIFY data.
2444 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2446 * disable: Disable this device.
2448 If there are multiple matching configurations changing
2449 the same attribute, the last one is used.
2451 memblock=debug [KNL] Enable memblock debug messages.
2453 load_ramdisk= [RAM] [Deprecated]
2455 lockd.nlm_grace_period=P [NFS] Assign grace period.
2458 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2461 lockd.nlm_timeout=T [NFS] Assign timeout value.
2464 lockd.nlm_udpport=M [NFS] Assign UDP port.
2467 lockdown= [SECURITY]
2468 { integrity | confidentiality }
2469 Enable the kernel lockdown feature. If set to
2470 integrity, kernel features that allow userland to
2471 modify the running kernel are disabled. If set to
2472 confidentiality, kernel features that allow userland
2473 to extract confidential information from the kernel
2476 locktorture.nreaders_stress= [KNL]
2477 Set the number of locking read-acquisition kthreads.
2478 Defaults to being automatically set based on the
2479 number of online CPUs.
2481 locktorture.nwriters_stress= [KNL]
2482 Set the number of locking write-acquisition kthreads.
2484 locktorture.onoff_holdoff= [KNL]
2485 Set time (s) after boot for CPU-hotplug testing.
2487 locktorture.onoff_interval= [KNL]
2488 Set time (s) between CPU-hotplug operations, or
2489 zero to disable CPU-hotplug testing.
2491 locktorture.shuffle_interval= [KNL]
2492 Set task-shuffle interval (jiffies). Shuffling
2493 tasks allows some CPUs to go into dyntick-idle
2494 mode during the locktorture test.
2496 locktorture.shutdown_secs= [KNL]
2497 Set time (s) after boot system shutdown. This
2498 is useful for hands-off automated testing.
2500 locktorture.stat_interval= [KNL]
2501 Time (s) between statistics printk()s.
2503 locktorture.stutter= [KNL]
2504 Time (s) to stutter testing, for example,
2505 specifying five seconds causes the test to run for
2506 five seconds, wait for five seconds, and so on.
2507 This tests the locking primitive's ability to
2508 transition abruptly to and from idle.
2510 locktorture.torture_type= [KNL]
2511 Specify the locking implementation to test.
2513 locktorture.verbose= [KNL]
2514 Enable additional printk() statements.
2516 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2519 loglevel= All Kernel Messages with a loglevel smaller than the
2520 console loglevel will be printed to the console. It can
2521 also be changed with klogd or other programs. The
2522 loglevels are defined as follows:
2524 0 (KERN_EMERG) system is unusable
2525 1 (KERN_ALERT) action must be taken immediately
2526 2 (KERN_CRIT) critical conditions
2527 3 (KERN_ERR) error conditions
2528 4 (KERN_WARNING) warning conditions
2529 5 (KERN_NOTICE) normal but significant condition
2530 6 (KERN_INFO) informational
2531 7 (KERN_DEBUG) debug-level messages
2533 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2534 in bytes. n must be a power of two and greater
2535 than the minimal size. The minimal size is defined
2536 by LOG_BUF_SHIFT kernel config parameter. There is
2537 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2538 that allows to increase the default size depending on
2539 the number of CPUs. See init/Kconfig for more details.
2541 logo.nologo [FB] Disables display of the built-in Linux logo.
2542 This may be used to provide more screen space for
2543 kernel log messages and is useful when debugging
2544 kernel boot problems.
2546 lp=0 [LP] Specify parallel ports to use, e.g,
2547 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2548 lp=reset first parallel port). 'lp=0' disables the
2549 lp=auto printer driver. 'lp=reset' (which can be
2550 specified in addition to the ports) causes
2551 attached printers to be reset. Using
2552 lp=port1,port2,... specifies the parallel ports
2553 to associate lp devices with, starting with
2554 lp0. A port specification may be 'none' to skip
2555 that lp device, or a parport name such as
2556 'parport0'. Specifying 'lp=auto' instead of a
2557 port specification list means that device IDs
2558 from each port should be examined, to see if
2559 an IEEE 1284-compliant printer is attached; if
2560 so, the driver will manage that printer.
2561 See also header of drivers/char/lp.c.
2564 Sets loops_per_jiffy to given constant, thus avoiding
2565 time-consuming boot-time autodetection (up to 250 ms per
2566 CPU). 0 enables autodetection (default). To determine
2567 the correct value for your kernel, boot with normal
2568 autodetection and see what value is printed. Note that
2569 on SMP systems the preset will be applied to all CPUs,
2570 which is likely to cause problems if your CPUs need
2571 significantly divergent settings. An incorrect value
2572 will cause delays in the kernel to be wrong, leading to
2573 unpredictable I/O errors and other breakage. Although
2574 unlikely, in the extreme case this might damage your
2578 Format: <io>,<irq>,<dma>
2580 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2583 [SECURITY] Choose order of LSM initialization. This
2584 overrides CONFIG_LSM, and the "security=" parameter.
2586 machvec= [IA-64] Force the use of a particular machine-vector
2587 (machvec) in a generic kernel.
2588 Example: machvec=hpzx1
2590 machtype= [Loongson] Share the same kernel image file between
2591 different yeeloong laptops.
2592 Example: machtype=lemote-yeeloong-2f-7inch
2594 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2595 than or equal to this physical address is ignored.
2597 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2598 will bring up during bootup. maxcpus=n : n >= 0 limits
2599 the kernel to bring up 'n' processors. Surely after
2600 bootup you can bring up the other plugged cpu by executing
2601 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2602 only takes effect during system bootup.
2603 While n=0 is a special case, it is equivalent to "nosmp",
2604 which also disables the IO APIC.
2606 max_loop= [LOOP] The number of loop block devices that get
2607 (loop.max_loop) unconditionally pre-created at init time. The default
2608 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2609 of statically allocating a predefined number, loop
2610 devices can be requested on-demand with the
2611 /dev/loop-control interface.
2613 mce [X86-32] Machine Check Exception
2615 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2617 md= [HW] RAID subsystems devices and level
2618 See Documentation/admin-guide/md.rst.
2621 Format: <first>,<last>
2622 Specifies range of consoles to be captured by the MDA.
2625 Control mitigation for the Micro-architectural Data
2626 Sampling (MDS) vulnerability.
2628 Certain CPUs are vulnerable to an exploit against CPU
2629 internal buffers which can forward information to a
2630 disclosure gadget under certain conditions.
2632 In vulnerable processors, the speculatively
2633 forwarded data can be used in a cache side channel
2634 attack, to access data to which the attacker does
2635 not have direct access.
2637 This parameter controls the MDS mitigation. The
2640 full - Enable MDS mitigation on vulnerable CPUs
2641 full,nosmt - Enable MDS mitigation and disable
2642 SMT on vulnerable CPUs
2643 off - Unconditionally disable MDS mitigation
2645 On TAA-affected machines, mds=off can be prevented by
2646 an active TAA mitigation as both vulnerabilities are
2647 mitigated with the same mechanism so in order to disable
2648 this mitigation, you need to specify tsx_async_abort=off
2651 Not specifying this option is equivalent to
2654 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2656 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2657 Amount of memory to be used in cases as follows:
2660 2 when the kernel is not able to see the whole system memory;
2661 3 memory that lies after 'mem=' boundary is excluded from
2662 the hypervisor, then assigned to KVM guests.
2664 [X86] Work as limiting max address. Use together
2665 with memmap= to avoid physical address space collisions.
2666 Without memmap= PCI devices could be placed at addresses
2667 belonging to unused RAM.
2669 Note that this only takes effects during boot time since
2670 in above case 3, memory may need be hot added after boot
2671 if system memory of hypervisor is not sufficient.
2673 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2677 [KNL,SH] Allow user to override the default size for
2678 per-device physically contiguous DMA buffers.
2680 memhp_default_state=online/offline
2681 [KNL] Set the initial state for the memory hotplug
2682 onlining policy. If not specified, the default value is
2683 set according to the
2684 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2686 See Documentation/admin-guide/mm/memory-hotplug.rst.
2688 memmap=exactmap [KNL,X86] Enable setting of an exact
2689 E820 memory map, as specified by the user.
2690 Such memmap=exactmap lines can be constructed based on
2691 BIOS output or other requirements. See the memmap=nn@ss
2694 memmap=nn[KMG]@ss[KMG]
2695 [KNL] Force usage of a specific region of memory.
2696 Region of memory to be used is from ss to ss+nn.
2697 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2698 which limits max address to nn[KMG].
2699 Multiple different regions can be specified,
2702 memmap=100M@2G,100M#3G,1G!1024G
2704 memmap=nn[KMG]#ss[KMG]
2705 [KNL,ACPI] Mark specific memory as ACPI data.
2706 Region of memory to be marked is from ss to ss+nn.
2708 memmap=nn[KMG]$ss[KMG]
2709 [KNL,ACPI] Mark specific memory as reserved.
2710 Region of memory to be reserved is from ss to ss+nn.
2711 Example: Exclude memory from 0x18690000-0x1869ffff
2712 memmap=64K$0x18690000
2714 memmap=0x10000$0x18690000
2715 Some bootloaders may need an escape character before '$',
2716 like Grub2, otherwise '$' and the following number
2719 memmap=nn[KMG]!ss[KMG]
2720 [KNL,X86] Mark specific memory as protected.
2721 Region of memory to be used, from ss to ss+nn.
2722 The memory region may be marked as e820 type 12 (0xc)
2723 and is NVDIMM or ADR memory.
2725 memmap=<size>%<offset>-<oldtype>+<newtype>
2726 [KNL,ACPI] Convert memory within the specified region
2727 from <oldtype> to <newtype>. If "-<oldtype>" is left
2728 out, the whole region will be marked as <newtype>,
2729 even if previously unavailable. If "+<newtype>" is left
2730 out, matching memory will be removed. Types are
2731 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2732 3 = ACPI, 12 = PRAM.
2734 memory_corruption_check=0/1 [X86]
2735 Some BIOSes seem to corrupt the first 64k of
2736 memory when doing things like suspend/resume.
2737 Setting this option will scan the memory
2738 looking for corruption. Enabling this will
2739 both detect corruption and prevent the kernel
2740 from using the memory being corrupted.
2741 However, its intended as a diagnostic tool; if
2742 repeatable BIOS-originated corruption always
2743 affects the same memory, you can use memmap=
2744 to prevent the kernel from using that memory.
2746 memory_corruption_check_size=size [X86]
2747 By default it checks for corruption in the low
2748 64k, making this memory unavailable for normal
2749 use. Use this parameter to scan for
2750 corruption in more or less memory.
2752 memory_corruption_check_period=seconds [X86]
2753 By default it checks for corruption every 60
2754 seconds. Use this parameter to check at some
2755 other rate. 0 disables periodic checking.
2757 memtest= [KNL,X86,ARM,PPC] Enable memtest
2759 default : 0 <disable>
2760 Specifies the number of memtest passes to be
2761 performed. Each pass selects another test
2762 pattern from a given set of patterns. Memtest
2763 fills the memory with this pattern, validates
2764 memory contents and reserves bad memory
2765 regions that are detected.
2767 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2768 Valid arguments: on, off
2769 Default (depends on kernel configuration option):
2770 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2771 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2772 mem_encrypt=on: Activate SME
2773 mem_encrypt=off: Do not activate SME
2775 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2776 for details on when memory encryption can be activated.
2778 mem_sleep_default= [SUSPEND] Default system suspend mode:
2779 s2idle - Suspend-To-Idle
2780 shallow - Power-On Suspend or equivalent (if supported)
2781 deep - Suspend-To-RAM or equivalent (if supported)
2782 See Documentation/admin-guide/pm/sleep-states.rst.
2784 meye.*= [HW] Set MotionEye Camera parameters
2785 See Documentation/admin-guide/media/meye.rst.
2787 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2788 Multi-Function General Purpose Timers on AMD Geode
2791 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2792 the BIOS has incorrectly applied a workaround. TinyBIOS
2793 version 0.98 is known to be affected, 0.99 fixes the
2794 problem by letting the user disable the workaround.
2798 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2799 physical address is ignored.
2801 mini2440= [ARM,HW,KNL]
2802 Format:[0..2][b][c][t]
2804 MINI2440 configuration specification:
2805 0 - The attached screen is the 3.5" TFT
2806 1 - The attached screen is the 7" TFT
2807 2 - The VGA Shield is attached (1024x768)
2808 Leaving out the screen size parameter will not load
2809 the TFT driver, and the framebuffer will be left
2811 b - Enable backlight. The TFT backlight pin will be
2812 linked to the kernel VESA blanking code and a GPIO
2813 LED. This parameter is not necessary when using the
2815 c - Enable the s3c camera interface.
2816 t - Reserved for enabling touchscreen support. The
2817 touchscreen support is not enabled in the mainstream
2818 kernel as of 2.6.30, a preliminary port can be found
2819 in the "bleeding edge" mini2440 support kernel at
2820 https://repo.or.cz/w/linux-2.6/mini2440.git
2823 [X86,PPC,S390,ARM64] Control optional mitigations for
2824 CPU vulnerabilities. This is a set of curated,
2825 arch-independent options, each of which is an
2826 aggregation of existing arch-specific options.
2829 Disable all optional CPU mitigations. This
2830 improves system performance, but it may also
2831 expose users to several CPU vulnerabilities.
2832 Equivalent to: nopti [X86,PPC]
2834 nospectre_v1 [X86,PPC]
2836 nospectre_v2 [X86,PPC,S390,ARM64]
2837 spectre_v2_user=off [X86]
2838 spec_store_bypass_disable=off [X86,PPC]
2839 ssbd=force-off [ARM64]
2842 tsx_async_abort=off [X86]
2843 kvm.nx_huge_pages=off [X86]
2846 This does not have any effect on
2847 kvm.nx_huge_pages when
2848 kvm.nx_huge_pages=force.
2851 Mitigate all CPU vulnerabilities, but leave SMT
2852 enabled, even if it's vulnerable. This is for
2853 users who don't want to be surprised by SMT
2854 getting disabled across kernel upgrades, or who
2855 have other ways of avoiding SMT-based attacks.
2856 Equivalent to: (default behavior)
2859 Mitigate all CPU vulnerabilities, disabling SMT
2860 if needed. This is for users who always want to
2861 be fully mitigated, even if it means losing SMT.
2862 Equivalent to: l1tf=flush,nosmt [X86]
2863 mds=full,nosmt [X86]
2864 tsx_async_abort=full,nosmt [X86]
2867 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2868 parameter allows control of the logging verbosity for
2869 the additional memory initialisation checks. A value
2870 of 0 disables mminit logging and a level of 4 will
2871 log everything. Information is printed at KERN_DEBUG
2872 so loglevel=8 may also need to be specified.
2875 [KNL] When CONFIG_MODULE_SIG is set, this means that
2876 modules without (valid) signatures will fail to load.
2877 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2878 is always true, so this option does nothing.
2880 module_blacklist= [KNL] Do not load a comma-separated list of
2881 modules. Useful for debugging problem modules.
2884 [MOUSE] Maximum time between finger touching and
2885 leaving touchpad surface for touch to be considered
2886 a tap and be reported as a left button click (for
2887 touchpads working in absolute mode only).
2889 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2890 reporting absolute coordinates, such as tablets
2891 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2892 reporting absolute coordinates, such as tablets
2894 movablecore= [KNL,X86,IA-64,PPC]
2895 Format: nn[KMGTPE] | nn%
2896 This parameter is the complement to kernelcore=, it
2897 specifies the amount of memory used for migratable
2898 allocations. If both kernelcore and movablecore is
2899 specified, then kernelcore will be at *least* the
2900 specified value but may be more. If movablecore on its
2901 own is specified, the administrator must be careful
2902 that the amount of memory usable for all allocations
2905 movable_node [KNL] Boot-time switch to make hotplugable memory
2906 NUMA nodes to be movable. This means that the memory
2907 of such nodes will be usable only for movable
2908 allocations which rules out almost all kernel
2909 allocations. Use with caution!
2911 MTD_Partition= [MTD]
2912 Format: <name>,<region-number>,<size>,<offset>
2914 MTD_Region= [MTD] Format:
2915 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2918 See drivers/mtd/parsers/cmdlinepart.c
2920 multitce=off [PPC] This parameter disables the use of the pSeries
2921 firmware feature for updating multiple TCE entries
2924 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2926 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2928 boundary - index of last SLC block on Flex-OneNAND.
2929 The remaining blocks are configured as MLC blocks.
2930 lock - Configure if Flex-OneNAND boundary should be locked.
2931 Once locked, the boundary cannot be changed.
2932 1 indicates lock status, 0 indicates unlock status.
2935 ARM/S3C2412 JIVE boot control
2937 See arch/arm/mach-s3c2412/mach-jive.c
2939 mtouchusb.raw_coordinates=
2940 [HW] Make the MicroTouch USB driver use raw coordinates
2941 ('y', default) or cooked coordinates ('n')
2943 mtrr_chunk_size=nn[KMG] [X86]
2944 used for mtrr cleanup. It is largest continuous chunk
2945 that could hold holes aka. UC entries.
2947 mtrr_gran_size=nn[KMG] [X86]
2948 Used for mtrr cleanup. It is granularity of mtrr block.
2950 Large value could prevent small alignment from
2953 mtrr_spare_reg_nr=n [X86]
2955 Range: 0,7 : spare reg number
2957 Used for mtrr cleanup. It is spare mtrr entries number.
2958 Set to 2 or more if your graphical card needs more.
2960 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2962 netdev= [NET] Network devices parameters
2963 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2964 Note that mem_start is often overloaded to mean
2965 something different and driver-specific.
2966 This usage is only documented in each driver source
2970 [NETFILTER] Enable connection tracking flow accounting
2971 0 to disable accounting
2972 1 to enable accounting
2975 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2976 See Documentation/admin-guide/nfs/nfsroot.rst.
2978 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2979 See Documentation/admin-guide/nfs/nfsroot.rst.
2981 nfsrootdebug [NFS] enable nfsroot debugging messages.
2982 See Documentation/admin-guide/nfs/nfsroot.rst.
2984 nfs.callback_nr_threads=
2985 [NFSv4] set the total number of threads that the
2986 NFS client will assign to service NFSv4 callback
2989 nfs.callback_tcpport=
2990 [NFS] set the TCP port on which the NFSv4 callback
2991 channel should listen.
2994 [NFS] sets the pathname to the program which is used
2995 to update the NFS client cache entries.
2997 nfs.cache_getent_timeout=
2998 [NFS] sets the timeout after which an attempt to
2999 update a cache entry is deemed to have failed.
3001 nfs.idmap_cache_timeout=
3002 [NFS] set the maximum lifetime for idmapper cache
3006 [NFS] enable 64-bit inode numbers.
3007 If zero, the NFS client will fake up a 32-bit inode
3008 number for the readdir() and stat() syscalls instead
3009 of returning the full 64-bit number.
3010 The default is to return 64-bit inode numbers.
3012 nfs.max_session_cb_slots=
3013 [NFSv4.1] Sets the maximum number of session
3014 slots the client will assign to the callback
3015 channel. This determines the maximum number of
3016 callbacks the client will process in parallel for
3017 a particular server.
3019 nfs.max_session_slots=
3020 [NFSv4.1] Sets the maximum number of session slots
3021 the client will attempt to negotiate with the server.
3022 This limits the number of simultaneous RPC requests
3023 that the client can send to the NFSv4.1 server.
3024 Note that there is little point in setting this
3025 value higher than the max_tcp_slot_table_limit.
3027 nfs.nfs4_disable_idmapping=
3028 [NFSv4] When set to the default of '1', this option
3029 ensures that both the RPC level authentication
3030 scheme and the NFS level operations agree to use
3031 numeric uids/gids if the mount is using the
3032 'sec=sys' security flavour. In effect it is
3033 disabling idmapping, which can make migration from
3034 legacy NFSv2/v3 systems to NFSv4 easier.
3035 Servers that do not support this mode of operation
3036 will be autodetected by the client, and it will fall
3037 back to using the idmapper.
3038 To turn off this behaviour, set the value to '0'.
3040 [NFS4] Specify an additional fixed unique ident-
3041 ification string that NFSv4 clients can insert into
3042 their nfs_client_id4 string. This is typically a
3043 UUID that is generated at system install time.
3045 nfs.send_implementation_id =
3046 [NFSv4.1] Send client implementation identification
3047 information in exchange_id requests.
3048 If zero, no implementation identification information
3050 The default is to send the implementation identification
3053 nfs.recover_lost_locks =
3054 [NFSv4] Attempt to recover locks that were lost due
3055 to a lease timeout on the server. Please note that
3056 doing this risks data corruption, since there are
3057 no guarantees that the file will remain unchanged
3058 after the locks are lost.
3059 If you want to enable the kernel legacy behaviour of
3060 attempting to recover these locks, then set this
3062 The default parameter value of '0' causes the kernel
3063 not to attempt recovery of lost locks.
3065 nfs4.layoutstats_timer =
3066 [NFSv4.2] Change the rate at which the kernel sends
3067 layoutstats to the pNFS metadata server.
3069 Setting this to value to 0 causes the kernel to use
3070 whatever value is the default set by the layout
3071 driver. A non-zero value sets the minimum interval
3072 in seconds between layoutstats transmissions.
3074 nfsd.nfs4_disable_idmapping=
3075 [NFSv4] When set to the default of '1', the NFSv4
3076 server will return only numeric uids and gids to
3077 clients using auth_sys, and will accept numeric uids
3078 and gids from such clients. This is intended to ease
3079 migration from NFSv2/v3.
3081 nmi_debug= [KNL,SH] Specify one or more actions to take
3082 when a NMI is triggered.
3083 Format: [state][,regs][,debounce][,die]
3085 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3086 Format: [panic,][nopanic,][num]
3088 0 - turn hardlockup detector in nmi_watchdog off
3089 1 - turn hardlockup detector in nmi_watchdog on
3090 When panic is specified, panic when an NMI watchdog
3091 timeout occurs (or 'nopanic' to not panic on an NMI
3092 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3093 To disable both hard and soft lockup detectors,
3094 please see 'nowatchdog'.
3095 This is useful when you use a panic=... timeout and
3096 need the box quickly up again.
3098 These settings can be accessed at runtime via
3099 the nmi_watchdog and hardlockup_panic sysctls.
3101 netpoll.carrier_timeout=
3102 [NET] Specifies amount of time (in seconds) that
3103 netpoll should wait for a carrier. By default netpoll
3106 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3107 emulation library even if a 387 maths coprocessor
3110 no5lvl [X86-64] Disable 5-level paging mode. Forces
3111 kernel to use 4-level paging instead.
3113 nofsgsbase [X86] Disables FSGSBASE instructions.
3116 [HW] Never suspend the console
3117 Disable suspending of consoles during suspend and
3118 hibernate operations. Once disabled, debugging
3119 messages can reach various consoles while the rest
3120 of the system is being put to sleep (ie, while
3121 debugging driver suspend/resume hooks). This may
3122 not work reliably with all consoles, but is known
3123 to work with serial and VGA consoles.
3124 To facilitate more flexible debugging, we also add
3125 console_suspend, a printk module parameter to control
3126 it. Users could use console_suspend (usually
3127 /sys/module/printk/parameters/console_suspend) to
3128 turn on/off it dynamically.
3130 novmcoredd [KNL,KDUMP]
3131 Disable device dump. Device dump allows drivers to
3132 append dump data to vmcore so you can collect driver
3133 specified debug info. Drivers can append the data
3134 without any limit and this data is stored in memory,
3135 so this may cause significant memory stress. Disabling
3136 device dump can help save memory but the driver debug
3137 data will be no longer available. This parameter
3138 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3141 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3142 caches in the slab allocator. Saves per-node memory,
3143 but will impact performance.
3147 noaltinstr [S390] Disables alternative instructions patching
3148 (CPU alternatives feature).
3150 noapic [SMP,APIC] Tells the kernel to not make use of any
3151 IOAPICs that may be present in the system.
3153 noautogroup Disable scheduler automatic task group creation.
3155 nobats [PPC] Do not use BATs for mapping kernel lowmem
3156 on "Classic" PPC cores.
3160 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3162 nodelayacct [KNL] Disable per-task delay accounting
3164 nodsp [SH] Disable hardware DSP at boot time.
3166 noefi Disable EFI runtime services support.
3171 On X86-32 available only on PAE configured kernels.
3172 noexec=on: enable non-executable mappings (default)
3173 noexec=off: disable non-executable mappings
3176 Disable SMAP (Supervisor Mode Access Prevention)
3177 even if it is supported by processor.
3180 Disable SMEP (Supervisor Mode Execution Prevention)
3181 even if it is supported by processor.
3184 This affects only 32-bit executables.
3185 noexec32=on: enable non-executable mappings (default)
3186 read doesn't imply executable mappings
3187 noexec32=off: disable non-executable mappings
3188 read implies executable mappings
3190 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3192 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3193 register save and restore. The kernel will only save
3194 legacy floating-point registers on task switch.
3196 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3198 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3199 Equivalent to smt=1.
3201 [KNL,X86] Disable symmetric multithreading (SMT).
3202 nosmt=force: Force disable SMT, cannot be undone
3203 via the sysfs control file.
3205 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3206 (bounds check bypass). With this option data leaks are
3207 possible in the system.
3209 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3210 the Spectre variant 2 (indirect branch prediction)
3211 vulnerability. System may allow data leaks with this
3214 nospec_store_bypass_disable
3215 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3217 noxsave [BUGS=X86] Disables x86 extended register state save
3218 and restore using xsave. The kernel will fallback to
3219 enabling legacy floating-point and sse state.
3221 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3222 register states. The kernel will fall back to use
3223 xsave to save the states. By using this parameter,
3224 performance of saving the states is degraded because
3225 xsave doesn't support modified optimization while
3226 xsaveopt supports it on xsaveopt enabled systems.
3228 noxsaves [X86] Disables xsaves and xrstors used in saving and
3229 restoring x86 extended register state in compacted
3230 form of xsave area. The kernel will fall back to use
3231 xsaveopt and xrstor to save and restore the states
3232 in standard form of xsave area. By using this
3233 parameter, xsave area per process might occupy more
3234 memory on xsaves enabled systems.
3236 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
3237 wfi(ARM) instruction doesn't work correctly and not to
3238 use it. This is also useful when using JTAG debugger.
3240 no_file_caps Tells the kernel not to honor file capabilities. The
3241 only way then for a file to be executed with privilege
3242 is to be setuid root or executed by root.
3244 nohalt [IA-64] Tells the kernel not to use the power saving
3245 function PAL_HALT_LIGHT when idle. This increases
3246 power-consumption. On the positive side, it reduces
3247 interrupt wake-up latency, which may improve performance
3248 in certain environments such as networked servers or
3251 nohibernate [HIBERNATION] Disable hibernation and resume.
3253 nohz= [KNL] Boottime enable/disable dynamic ticks
3254 Valid arguments: on, off
3257 nohz_full= [KNL,BOOT,SMP,ISOL]
3258 The argument is a cpu list, as described above.
3259 In kernels built with CONFIG_NO_HZ_FULL=y, set
3260 the specified list of CPUs whose tick will be stopped
3261 whenever possible. The boot CPU will be forced outside
3262 the range to maintain the timekeeping. Any CPUs
3263 in this list will have their RCU callbacks offloaded,
3264 just as if they had also been called out in the
3265 rcu_nocbs= boot parameter.
3267 noiotrap [SH] Disables trapped I/O port accesses.
3269 noirqdebug [X86-32] Disables the code which attempts to detect and
3270 disable unhandled interrupt sources.
3272 no_timer_check [X86,APIC] Disables the code which tests for
3273 broken timer IRQ sources.
3275 noisapnp [ISAPNP] Disables ISA PnP code.
3277 noinitrd [RAM] Tells the kernel not to load any configured
3280 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3282 [Deprecated - use intremap=off]
3286 noinvpcid [X86] Disable the INVPCID cpu feature.
3288 nojitter [IA-64] Disables jitter checking for ITC timers.
3290 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3292 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3296 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3297 clock and use the default one.
3299 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3300 accounting. steal time is computed, but won't
3301 influence scheduler behaviour
3303 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3305 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3307 noltlbs [PPC] Do not use large page/tlb entries for kernel
3308 lowmem mapping on PPC40x and PPC8xx
3310 nomca [IA-64] Disable machine check abort handling
3312 nomce [X86-32] Disable Machine Check Exception
3314 nomfgpt [X86-32] Disable Multi-Function General Purpose
3315 Timer usage (for AMD Geode machines).
3317 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3318 shutdown the other cpus. Instead use the REBOOT_VECTOR
3321 nomodule Disable module load
3323 nopat [X86] Disable PAT (page attribute table extension of
3324 pagetables) support.
3326 nopcid [X86-64] Disable the PCID cpu feature.
3328 norandmaps Don't use address space randomization. Equivalent to
3329 echo 0 > /proc/sys/kernel/randomize_va_space
3331 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3332 with UP alternatives
3334 nordrand [X86] Disable kernel use of the RDRAND and
3335 RDSEED instructions even if they are supported
3336 by the processor. RDRAND and RDSEED are still
3337 available to user space applications.
3339 noresume [SWSUSP] Disables resume and restores original swap
3342 no-scroll [VGA] Disables scrollback.
3343 This is required for the Braillex ib80-piezo Braille
3344 reader made by F.H. Papenmeier (Germany).
3348 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3350 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3351 and disable the IO APIC. legacy for "maxcpus=0".
3353 nosoftlockup [KNL] Disable the soft-lockup detector.
3355 nosync [HW,M68K] Disables sync negotiation for all devices.
3357 nowatchdog [KNL] Disable both lockup detectors, i.e.
3358 soft-lockup and NMI watchdog (hard-lockup).
3362 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3364 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3365 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3366 Some features depend on CPU0. Known dependencies are:
3367 1. Resume from suspend/hibernate depends on CPU0.
3368 Suspend/hibernate will fail if CPU0 is offline and you
3369 need to online CPU0 before suspend/hibernate.
3370 2. PIC interrupts also depend on CPU0. CPU0 can't be
3371 removed if a PIC interrupt is detected.
3372 It's said poweroff/reboot may depend on CPU0 on some
3373 machines although I haven't seen such issues so far
3374 after CPU0 is offline on a few tested machines.
3375 If the dependencies are under your control, you can
3376 turn on cpu0_hotplug.
3378 nps_mtm_hs_ctr= [KNL,ARC]
3379 This parameter sets the maximum duration, in
3380 cycles, each HW thread of the CTOP can run
3381 without interruptions, before HW switches it.
3382 The actual maximum duration is 16 times this
3384 Format: integer between 1 and 255
3387 nptcg= [IA-64] Override max number of concurrent global TLB
3388 purges which is reported from either PAL_VM_SUMMARY or
3391 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3392 could support. nr_cpus=n : n >= 1 limits the kernel to
3393 support 'n' processors. It could be larger than the
3394 number of already plugged CPU during bootup, later in
3395 runtime you can physically add extra cpu until it reaches
3396 n. So during boot up some boot time memory for per-cpu
3397 variables need be pre-allocated for later physical cpu
3400 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3402 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3403 Allowed values are enable and disable
3405 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3406 'node', 'default' can be specified
3407 This can be set from sysctl after boot.
3408 See Documentation/admin-guide/sysctl/vm.rst for details.
3410 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3411 See Documentation/core-api/debugging-via-ohci1394.rst for more
3414 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3415 Rather than timing out after 20 ms if an EC
3416 command is not properly ACKed, override the length
3417 of the timeout. We have interrupts disabled while
3418 waiting for the ACK, so if this is set too high
3419 interrupts *may* be lost!
3421 omap_mux= [OMAP] Override bootloader pin multiplexing.
3422 Format: <mux_mode0.mode_name=value>...
3423 For example, to override I2C bus2:
3424 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3426 oprofile.timer= [HW]
3427 Use timer interrupt instead of performance counters
3429 oprofile.cpu_type= Force an oprofile cpu type
3430 This might be useful if you have an older oprofile
3431 userland or if you want common events.
3432 Format: { arch_perfmon }
3433 arch_perfmon: [X86] Force use of architectural
3434 perfmon on Intel CPUs instead of the
3435 CPU specific event set.
3436 timer: [X86] Force use of architectural NMI
3437 timer mode (see also oprofile.timer
3438 for generic hr timer mode)
3440 oops=panic Always panic on oopses. Default is to just kill the
3441 process, but there is a small probability of
3442 deadlocking the machine.
3443 This will also cause panics on machine check exceptions.
3444 Useful together with panic=30 to trigger a reboot.
3447 [KNL] Boolean flag to control whether the page allocator
3448 should randomize its free lists. The randomization may
3449 be automatically enabled if the kernel detects it is
3450 running on a platform with a direct-mapped memory-side
3451 cache, and this parameter can be used to
3452 override/disable that behavior. The state of the flag
3453 can be read from sysfs at:
3454 /sys/module/page_alloc/parameters/shuffle.
3456 page_owner= [KNL] Boot-time page_owner enabling option.
3457 Storage of the information about who allocated
3458 each page is disabled in default. With this switch,
3460 on: enable the feature
3462 page_poison= [KNL] Boot-time parameter changing the state of
3463 poisoning on the buddy allocator, available with
3464 CONFIG_PAGE_POISONING=y.
3465 off: turn off poisoning (default)
3466 on: turn on poisoning
3468 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3469 timeout > 0: seconds before rebooting
3470 timeout = 0: wait forever
3471 timeout < 0: reboot immediately
3474 panic_print= Bitmask for printing system info when panic happens.
3475 User can chose combination of the following bits:
3476 bit 0: print all tasks info
3477 bit 1: print system memory info
3478 bit 2: print timer info
3479 bit 3: print locks info if CONFIG_LOCKDEP is on
3480 bit 4: print ftrace buffer
3481 bit 5: print all printk messages in buffer
3483 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3484 Format: <hex>[,nousertaint]
3485 Hexadecimal bitmask representing the set of TAINT flags
3486 that will cause the kernel to panic when add_taint() is
3487 called with any of the flags in this set.
3488 The optional switch "nousertaint" can be utilized to
3489 prevent userspace forced crashes by writing to sysctl
3490 /proc/sys/kernel/tainted any flagset matching with the
3491 bitmask set on panic_on_taint.
3492 See Documentation/admin-guide/tainted-kernels.rst for
3493 extra details on the taint flags that users can pick
3494 to compose the bitmask to assign to panic_on_taint.
3496 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3499 crash_kexec_post_notifiers
3500 Run kdump after running panic-notifiers and dumping
3501 kmsg. This only for the users who doubt kdump always
3502 succeeds in any situation.
3503 Note that this also increases risks of kdump failure,
3504 because some panic notifiers can make the crashed
3505 kernel more unstable.
3507 parkbd.port= [HW] Parallel port number the keyboard adapter is
3508 connected to, default is 0.
3510 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3511 0 for XT, 1 for AT (default is AT).
3514 parport= [HW,PPT] Specify parallel ports. 0 disables.
3515 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3516 Use 'auto' to force the driver to use any
3517 IRQ/DMA settings detected (the default is to
3518 ignore detected IRQ/DMA settings because of
3519 possible conflicts). You can specify the base
3520 address, IRQ, and DMA settings; IRQ and DMA
3521 should be numbers, or 'auto' (for using detected
3522 settings on that particular port), or 'nofifo'
3523 (to avoid using a FIFO even if it is detected).
3524 Parallel ports are assigned in the order they
3525 are specified on the command line, starting
3528 parport_init_mode= [HW,PPT]
3529 Configure VIA parallel port to operate in
3530 a specific mode. This is necessary on Pegasos
3531 computer where firmware has no options for setting
3532 up parallel port mode and sets it to spp.
3533 Currently this function knows 686a and 8231 chips.
3534 Format: [spp|ps2|epp|ecp|ecpepp]
3537 Halt all CPUs after the first oops has been printed for
3538 the specified number of seconds. This is to be used if
3539 your oopses keep scrolling off the screen.
3544 See header of drivers/block/paride/pcd.c.
3545 See also Documentation/admin-guide/blockdev/paride.rst.
3547 pci=option[,option...] [PCI] various PCI subsystem options.
3549 Some options herein operate on a specific device
3550 or a set of devices (<pci_dev>). These are
3551 specified in one of the following formats:
3553 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3554 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3556 Note: the first format specifies a PCI
3557 bus/device/function address which may change
3558 if new hardware is inserted, if motherboard
3559 firmware changes, or due to changes caused
3560 by other kernel parameters. If the
3561 domain is left unspecified, it is
3562 taken to be zero. Optionally, a path
3563 to a device through multiple device/function
3564 addresses can be specified after the base
3565 address (this is more robust against
3566 renumbering issues). The second format
3567 selects devices using IDs from the
3568 configuration space which may match multiple
3569 devices in the system.
3571 earlydump dump PCI config space before the kernel
3573 off [X86] don't probe for the PCI bus
3574 bios [X86-32] force use of PCI BIOS, don't access
3575 the hardware directly. Use this if your machine
3576 has a non-standard PCI host bridge.
3577 nobios [X86-32] disallow use of PCI BIOS, only direct
3578 hardware access methods are allowed. Use this
3579 if you experience crashes upon bootup and you
3580 suspect they are caused by the BIOS.
3581 conf1 [X86] Force use of PCI Configuration Access
3582 Mechanism 1 (config address in IO port 0xCF8,
3583 data in IO port 0xCFC, both 32-bit).
3584 conf2 [X86] Force use of PCI Configuration Access
3585 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3586 the function, IO port 0xCFA, also 8-bit, sets
3587 bus number. The config space is then accessed
3588 through ports 0xC000-0xCFFF).
3589 See http://wiki.osdev.org/PCI for more info
3590 on the configuration access mechanisms.
3591 noaer [PCIE] If the PCIEAER kernel config parameter is
3592 enabled, this kernel boot option can be used to
3593 disable the use of PCIE advanced error reporting.
3594 nodomains [PCI] Disable support for multiple PCI
3595 root domains (aka PCI segments, in ACPI-speak).
3596 nommconf [X86] Disable use of MMCONFIG for PCI
3598 check_enable_amd_mmconf [X86] check for and enable
3599 properly configured MMIO access to PCI
3600 config space on AMD family 10h CPU
3601 nomsi [MSI] If the PCI_MSI kernel config parameter is
3602 enabled, this kernel boot option can be used to
3603 disable the use of MSI interrupts system-wide.
3604 noioapicquirk [APIC] Disable all boot interrupt quirks.
3605 Safety option to keep boot IRQs enabled. This
3606 should never be necessary.
3607 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3608 primary IO-APIC for bridges that cannot disable
3609 boot IRQs. This fixes a source of spurious IRQs
3610 when the system masks IRQs.
3611 noioapicreroute [APIC] Disable workaround that uses the
3612 boot IRQ equivalent of an IRQ that connects to
3613 a chipset where boot IRQs cannot be disabled.
3614 The opposite of ioapicreroute.
3615 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3616 routing table. These calls are known to be buggy
3617 on several machines and they hang the machine
3618 when used, but on other computers it's the only
3619 way to get the interrupt routing table. Try
3620 this option if the kernel is unable to allocate
3621 IRQs or discover secondary PCI buses on your
3623 rom [X86] Assign address space to expansion ROMs.
3624 Use with caution as certain devices share
3625 address decoders between ROMs and other
3627 norom [X86] Do not assign address space to
3628 expansion ROMs that do not already have
3629 BIOS assigned address ranges.
3630 nobar [X86] Do not assign address space to the
3631 BARs that weren't assigned by the BIOS.
3632 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3633 assigned automatically to PCI devices. You can
3634 make the kernel exclude IRQs of your ISA cards
3636 pirqaddr=0xAAAAA [X86] Specify the physical address
3637 of the PIRQ table (normally generated
3638 by the BIOS) if it is outside the
3639 F0000h-100000h range.
3640 lastbus=N [X86] Scan all buses thru bus #N. Can be
3641 useful if the kernel is unable to find your
3642 secondary buses and you want to tell it
3643 explicitly which ones they are.
3644 assign-busses [X86] Always assign all PCI bus
3645 numbers ourselves, overriding
3646 whatever the firmware may have done.
3647 usepirqmask [X86] Honor the possible IRQ mask stored
3648 in the BIOS $PIR table. This is needed on
3649 some systems with broken BIOSes, notably
3650 some HP Pavilion N5400 and Omnibook XE3
3651 notebooks. This will have no effect if ACPI
3652 IRQ routing is enabled.
3653 noacpi [X86] Do not use ACPI for IRQ routing
3654 or for PCI scanning.
3655 use_crs [X86] Use PCI host bridge window information
3656 from ACPI. On BIOSes from 2008 or later, this
3657 is enabled by default. If you need to use this,
3658 please report a bug.
3659 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3660 If you need to use this, please report a bug.
3661 routeirq Do IRQ routing for all PCI devices.
3662 This is normally done in pci_enable_device(),
3663 so this option is a temporary workaround
3664 for broken drivers that don't call it.
3665 skip_isa_align [X86] do not align io start addr, so can
3666 handle more pci cards
3667 noearly [X86] Don't do any early type 1 scanning.
3668 This might help on some broken boards which
3669 machine check when some devices' config space
3670 is read. But various workarounds are disabled
3671 and some IOMMU drivers will not work.
3672 bfsort Sort PCI devices into breadth-first order.
3673 This sorting is done to get a device
3674 order compatible with older (<= 2.4) kernels.
3675 nobfsort Don't sort PCI devices into breadth-first order.
3676 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3677 tuning and use the BIOS-configured MPS defaults.
3678 pcie_bus_safe Set every device's MPS to the largest value
3679 supported by all devices below the root complex.
3680 pcie_bus_perf Set device MPS to the largest allowable MPS
3681 based on its parent bus. Also set MRRS (Max
3682 Read Request Size) to the largest supported
3683 value (no larger than the MPS that the device
3684 or bus can support) for best performance.
3685 pcie_bus_peer2peer Set every device's MPS to 128B, which
3686 every device is guaranteed to support. This
3687 configuration allows peer-to-peer DMA between
3688 any pair of devices, possibly at the cost of
3689 reduced performance. This also guarantees
3690 that hot-added devices will work.
3691 cbiosize=nn[KMG] The fixed amount of bus space which is
3692 reserved for the CardBus bridge's IO window.
3693 The default value is 256 bytes.
3694 cbmemsize=nn[KMG] The fixed amount of bus space which is
3695 reserved for the CardBus bridge's memory
3696 window. The default value is 64 megabytes.
3699 [<order of align>@]<pci_dev>[; ...]
3700 Specifies alignment and device to reassign
3701 aligned memory resources. How to
3702 specify the device is described above.
3703 If <order of align> is not specified,
3704 PAGE_SIZE is used as alignment.
3705 A PCI-PCI bridge can be specified if resource
3706 windows need to be expanded.
3707 To specify the alignment for several
3708 instances of a device, the PCI vendor,
3709 device, subvendor, and subdevice may be
3710 specified, e.g., 12@pci:8086:9c22:103c:198f
3711 for 4096-byte alignment.
3712 ecrc= Enable/disable PCIe ECRC (transaction layer
3713 end-to-end CRC checking).
3714 bios: Use BIOS/firmware settings. This is the
3718 hpiosize=nn[KMG] The fixed amount of bus space which is
3719 reserved for hotplug bridge's IO window.
3720 Default size is 256 bytes.
3721 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3722 reserved for hotplug bridge's MMIO window.
3723 Default size is 2 megabytes.
3724 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3725 reserved for hotplug bridge's MMIO_PREF window.
3726 Default size is 2 megabytes.
3727 hpmemsize=nn[KMG] The fixed amount of bus space which is
3728 reserved for hotplug bridge's MMIO and
3730 Default size is 2 megabytes.
3731 hpbussize=nn The minimum amount of additional bus numbers
3732 reserved for buses below a hotplug bridge.
3734 realloc= Enable/disable reallocating PCI bridge resources
3735 if allocations done by BIOS are too small to
3736 accommodate resources required by all child
3738 off: Turn realloc off
3740 realloc same as realloc=on
3741 noari do not use PCIe ARI.
3742 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3743 do not use PCIe ATS (and IOMMU device IOTLB).
3744 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3745 only look for one device below a PCIe downstream
3747 big_root_window Try to add a big 64bit memory window to the PCIe
3748 root complex on AMD CPUs. Some GFX hardware
3749 can resize a BAR to allow access to all VRAM.
3750 Adding the window is slightly risky (it may
3751 conflict with unreported devices), so this
3753 disable_acs_redir=<pci_dev>[; ...]
3754 Specify one or more PCI devices (in the format
3755 specified above) separated by semicolons.
3756 Each device specified will have the PCI ACS
3757 redirect capabilities forced off which will
3758 allow P2P traffic between devices through
3759 bridges without forcing it upstream. Note:
3760 this removes isolation between devices and
3761 may put more devices in an IOMMU group.
3762 force_floating [S390] Force usage of floating interrupts.
3763 nomio [S390] Do not use MIO instructions.
3764 norid [S390] ignore the RID field and force use of
3765 one PCI domain per PCI function
3767 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3770 force Enable ASPM even on devices that claim not to support it.
3771 WARNING: Forcing ASPM on may cause system lockups.
3773 pcie_ports= [PCIE] PCIe port services handling:
3774 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3775 even if the platform doesn't give the OS permission to
3776 use them. This may cause conflicts if the platform
3777 also tries to use these services.
3778 dpc-native Use native PCIe service for DPC only. May
3779 cause conflicts if firmware uses AER or DPC.
3780 compat Disable native PCIe services (PME, AER, DPC, PCIe
3783 pcie_port_pm= [PCIE] PCIe port power management handling:
3784 off Disable power management of all PCIe ports
3785 force Forcibly enable power management of all PCIe ports
3787 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3788 nomsi Do not use MSI for native PCIe PME signaling (this makes
3789 all PCIe root ports use INTx for all services).
3791 pcmv= [HW,PCMCIA] BadgePAD 4
3795 Keep all power-domains already enabled by bootloader on,
3796 even if no driver has claimed them. This is useful
3797 for debug and development, but should not be
3798 needed on a platform with proper driver support.
3801 See Documentation/admin-guide/blockdev/paride.rst.
3803 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3806 See arch/parisc/kernel/pdc_chassis.c
3808 percpu_alloc= Select which percpu first chunk allocator to use.
3809 Currently supported values are "embed" and "page".
3810 Archs may support subset or none of the selections.
3811 See comments in mm/percpu.c for details on each
3812 allocator. This parameter is primarily for debugging
3813 and performance comparison.
3816 See Documentation/admin-guide/blockdev/paride.rst.
3819 See Documentation/admin-guide/blockdev/paride.rst.
3821 pirq= [SMP,APIC] Manual mp-table setup
3822 See Documentation/x86/i386/IO-APIC.rst.
3824 plip= [PPT,NET] Parallel port network link
3825 Format: { parport<nr> | timid | 0 }
3826 See also Documentation/admin-guide/parport.rst.
3828 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3829 Override pmtimer IOPort with a hex value.
3832 pm_debug_messages [SUSPEND,KNL]
3833 Enable suspend/resume debug messages during boot up.
3836 Enable PNP debug messages (depends on the
3837 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3838 via /sys/module/pnp/parameters/debug. We always show
3839 current resource usage; turning this on also shows
3840 possible settings and some assignment information.
3846 { on | off | curr | res | no-curr | no-res }
3849 [ISAPNP] Exclude IRQs for the autoconfiguration
3852 [ISAPNP] Exclude DMAs for the autoconfiguration
3854 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3855 Ranges are in pairs (I/O port base and size).
3858 [ISAPNP] Exclude memory regions for the
3860 Ranges are in pairs (memory base and size).
3862 ports= [IP_VS_FTP] IPVS ftp helper module
3864 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3866 Format: <port>,<port>....
3868 powersave=off [PPC] This option disables power saving features.
3869 It specifically disables cpuidle and sets the
3870 platform machine description specific power_save
3871 function to NULL. On Idle the CPU just reduces
3874 ppc_strict_facility_enable
3875 [PPC] This option catches any kernel floating point,
3876 Altivec, VSX and SPE outside of regions specifically
3877 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3878 There is some performance impact when enabling this.
3882 Disable Hardware Transactional Memory
3884 print-fatal-signals=
3885 [KNL] debug: print fatal signals
3887 If enabled, warn about various signal handling
3888 related application anomalies: too many signals,
3889 too many POSIX.1 timers, fatal signals causing a
3892 If you hit the warning due to signal overflow,
3893 you might want to try "ulimit -i unlimited".
3897 printk.always_kmsg_dump=
3898 Trigger kmsg_dump for cases other than kernel oops or
3900 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3903 printk.devkmsg={on,off,ratelimit}
3904 Control writing to /dev/kmsg.
3905 on - unlimited logging to /dev/kmsg from userspace
3906 off - logging to /dev/kmsg disabled
3907 ratelimit - ratelimit the logging
3910 printk.time= Show timing data prefixed to each printk message line
3911 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3913 processor.max_cstate= [HW,ACPI]
3914 Limit processor to maximum C-state
3915 max_cstate=9 overrides any DMI blacklist limit.
3917 processor.nocst [HW,ACPI]
3918 Ignore the _CST method to determine C-states,
3919 instead using the legacy FADT method
3921 profile= [KNL] Enable kernel profiling via /proc/profile
3922 Format: [<profiletype>,]<number>
3923 Param: <profiletype>: "schedule", "sleep", or "kvm"
3924 [defaults to kernel profiling]
3925 Param: "schedule" - profile schedule points.
3926 Param: "sleep" - profile D-state sleeping (millisecs).
3927 Requires CONFIG_SCHEDSTATS
3928 Param: "kvm" - profile VM exits.
3929 Param: <number> - step/bucket size as a power of 2 for
3930 statistical time based profiling.
3932 prompt_ramdisk= [RAM] [Deprecated]
3934 prot_virt= [S390] enable hosting protected virtual machines
3935 isolated from the hypervisor (if hardware supports
3939 psi= [KNL] Enable or disable pressure stall information
3943 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3944 probe for; one of (bare|imps|exps|lifebook|any).
3945 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3947 psmouse.resetafter= [HW,MOUSE]
3948 Try to reset the device after so many bad packets
3951 [HW,MOUSE] Set desired mouse resolution, in dpi.
3952 psmouse.smartscroll=
3953 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3954 0 = disabled, 1 = enabled (default).
3956 pstore.backend= Specify the name of the pstore backend to use
3959 See Documentation/admin-guide/blockdev/paride.rst.
3961 pti= [X86-64] Control Page Table Isolation of user and
3962 kernel address spaces. Disabling this feature
3963 removes hardening, but improves performance of
3964 system calls and interrupts.
3966 on - unconditionally enable
3967 off - unconditionally disable
3968 auto - kernel detects whether your CPU model is
3969 vulnerable to issues that PTI mitigates
3971 Not specifying this option is equivalent to pti=auto.
3974 Equivalent to pti=off
3977 [KNL] Number of legacy pty's. Overwrites compiled-in
3980 quiet [KNL] Disable most log messages
3985 See Documentation/admin-guide/md.rst.
3987 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3988 See Documentation/admin-guide/blockdev/ramdisk.rst.
3990 ramdisk_start= [RAM] RAM disk image start address
3992 random.trust_cpu={on,off}
3993 [KNL] Enable or disable trusting the use of the
3994 CPU's random number generator (if available) to
3995 fully seed the kernel's CRNG. Default is controlled
3996 by CONFIG_RANDOM_TRUST_CPU.
3998 ras=option[,option,...] [KNL] RAS-specific options
4001 Disable the Correctable Errors Collector,
4002 see CONFIG_RAS_CEC help text.
4005 The argument is a cpu list, as described above,
4006 except that the string "all" can be used to
4007 specify every CPU on the system.
4009 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
4010 the specified list of CPUs to be no-callback CPUs.
4011 Invocation of these CPUs' RCU callbacks will be
4012 offloaded to "rcuox/N" kthreads created for that
4013 purpose, where "x" is "p" for RCU-preempt, and
4014 "s" for RCU-sched, and "N" is the CPU number.
4015 This reduces OS jitter on the offloaded CPUs,
4016 which can be useful for HPC and real-time
4017 workloads. It can also improve energy efficiency
4018 for asymmetric multiprocessors.
4021 Rather than requiring that offloaded CPUs
4022 (specified by rcu_nocbs= above) explicitly
4023 awaken the corresponding "rcuoN" kthreads,
4024 make these kthreads poll for callbacks.
4025 This improves the real-time response for the
4026 offloaded CPUs by relieving them of the need to
4027 wake up the corresponding kthread, but degrades
4028 energy efficiency by requiring that the kthreads
4029 periodically wake up to do the polling.
4031 rcutree.blimit= [KNL]
4032 Set maximum number of finished RCU callbacks to
4033 process in one batch.
4035 rcutree.dump_tree= [KNL]
4036 Dump the structure of the rcu_node combining tree
4037 out at early boot. This is used for diagnostic
4038 purposes, to verify correct tree setup.
4040 rcutree.gp_cleanup_delay= [KNL]
4041 Set the number of jiffies to delay each step of
4042 RCU grace-period cleanup.
4044 rcutree.gp_init_delay= [KNL]
4045 Set the number of jiffies to delay each step of
4046 RCU grace-period initialization.
4048 rcutree.gp_preinit_delay= [KNL]
4049 Set the number of jiffies to delay each step of
4050 RCU grace-period pre-initialization, that is,
4051 the propagation of recent CPU-hotplug changes up
4052 the rcu_node combining tree.
4054 rcutree.use_softirq= [KNL]
4055 If set to zero, move all RCU_SOFTIRQ processing to
4056 per-CPU rcuc kthreads. Defaults to a non-zero
4057 value, meaning that RCU_SOFTIRQ is used by default.
4058 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4060 rcutree.rcu_fanout_exact= [KNL]
4061 Disable autobalancing of the rcu_node combining
4062 tree. This is used by rcutorture, and might
4063 possibly be useful for architectures having high
4064 cache-to-cache transfer latencies.
4066 rcutree.rcu_fanout_leaf= [KNL]
4067 Change the number of CPUs assigned to each
4068 leaf rcu_node structure. Useful for very
4069 large systems, which will choose the value 64,
4070 and for NUMA systems with large remote-access
4071 latencies, which will choose a value aligned
4072 with the appropriate hardware boundaries.
4074 rcutree.rcu_min_cached_objs= [KNL]
4075 Minimum number of objects which are cached and
4076 maintained per one CPU. Object size is equal
4077 to PAGE_SIZE. The cache allows to reduce the
4078 pressure to page allocator, also it makes the
4079 whole algorithm to behave better in low memory
4082 rcutree.jiffies_till_first_fqs= [KNL]
4083 Set delay from grace-period initialization to
4084 first attempt to force quiescent states.
4085 Units are jiffies, minimum value is zero,
4086 and maximum value is HZ.
4088 rcutree.jiffies_till_next_fqs= [KNL]
4089 Set delay between subsequent attempts to force
4090 quiescent states. Units are jiffies, minimum
4091 value is one, and maximum value is HZ.
4093 rcutree.jiffies_till_sched_qs= [KNL]
4094 Set required age in jiffies for a
4095 given grace period before RCU starts
4096 soliciting quiescent-state help from
4097 rcu_note_context_switch() and cond_resched().
4098 If not specified, the kernel will calculate
4099 a value based on the most recent settings
4100 of rcutree.jiffies_till_first_fqs
4101 and rcutree.jiffies_till_next_fqs.
4102 This calculated value may be viewed in
4103 rcutree.jiffies_to_sched_qs. Any attempt to set
4104 rcutree.jiffies_to_sched_qs will be cheerfully
4107 rcutree.kthread_prio= [KNL,BOOT]
4108 Set the SCHED_FIFO priority of the RCU per-CPU
4109 kthreads (rcuc/N). This value is also used for
4110 the priority of the RCU boost threads (rcub/N)
4111 and for the RCU grace-period kthreads (rcu_bh,
4112 rcu_preempt, and rcu_sched). If RCU_BOOST is
4113 set, valid values are 1-99 and the default is 1
4114 (the least-favored priority). Otherwise, when
4115 RCU_BOOST is not set, valid values are 0-99 and
4116 the default is zero (non-realtime operation).
4118 rcutree.rcu_nocb_gp_stride= [KNL]
4119 Set the number of NOCB callback kthreads in
4120 each group, which defaults to the square root
4121 of the number of CPUs. Larger numbers reduce
4122 the wakeup overhead on the global grace-period
4123 kthread, but increases that same overhead on
4124 each group's NOCB grace-period kthread.
4126 rcutree.qhimark= [KNL]
4127 Set threshold of queued RCU callbacks beyond which
4128 batch limiting is disabled.
4130 rcutree.qlowmark= [KNL]
4131 Set threshold of queued RCU callbacks below which
4132 batch limiting is re-enabled.
4134 rcutree.qovld= [KNL]
4135 Set threshold of queued RCU callbacks beyond which
4136 RCU's force-quiescent-state scan will aggressively
4137 enlist help from cond_resched() and sched IPIs to
4138 help CPUs more quickly reach quiescent states.
4139 Set to less than zero to make this be set based
4140 on rcutree.qhimark at boot time and to zero to
4141 disable more aggressive help enlistment.
4143 rcutree.rcu_idle_gp_delay= [KNL]
4144 Set wakeup interval for idle CPUs that have
4145 RCU callbacks (RCU_FAST_NO_HZ=y).
4147 rcutree.rcu_idle_lazy_gp_delay= [KNL]
4148 Set wakeup interval for idle CPUs that have
4149 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
4150 Lazy RCU callbacks are those which RCU can
4151 prove do nothing more than free memory.
4153 rcutree.rcu_kick_kthreads= [KNL]
4154 Cause the grace-period kthread to get an extra
4155 wake_up() if it sleeps three times longer than
4156 it should at force-quiescent-state time.
4157 This wake_up() will be accompanied by a
4158 WARN_ONCE() splat and an ftrace_dump().
4160 rcutree.sysrq_rcu= [KNL]
4161 Commandeer a sysrq key to dump out Tree RCU's
4162 rcu_node tree with an eye towards determining
4163 why a new grace period has not yet started.
4165 rcuperf.gp_async= [KNL]
4166 Measure performance of asynchronous
4167 grace-period primitives such as call_rcu().
4169 rcuperf.gp_async_max= [KNL]
4170 Specify the maximum number of outstanding
4171 callbacks per writer thread. When a writer
4172 thread exceeds this limit, it invokes the
4173 corresponding flavor of rcu_barrier() to allow
4174 previously posted callbacks to drain.
4176 rcuperf.gp_exp= [KNL]
4177 Measure performance of expedited synchronous
4178 grace-period primitives.
4180 rcuperf.holdoff= [KNL]
4181 Set test-start holdoff period. The purpose of
4182 this parameter is to delay the start of the
4183 test until boot completes in order to avoid
4186 rcuperf.kfree_rcu_test= [KNL]
4187 Set to measure performance of kfree_rcu() flooding.
4189 rcuperf.kfree_nthreads= [KNL]
4190 The number of threads running loops of kfree_rcu().
4192 rcuperf.kfree_alloc_num= [KNL]
4193 Number of allocations and frees done in an iteration.
4195 rcuperf.kfree_loops= [KNL]
4196 Number of loops doing rcuperf.kfree_alloc_num number
4197 of allocations and frees.
4199 rcuperf.nreaders= [KNL]
4200 Set number of RCU readers. The value -1 selects
4201 N, where N is the number of CPUs. A value
4202 "n" less than -1 selects N-n+1, where N is again
4203 the number of CPUs. For example, -2 selects N
4204 (the number of CPUs), -3 selects N+1, and so on.
4205 A value of "n" less than or equal to -N selects
4208 rcuperf.nwriters= [KNL]
4209 Set number of RCU writers. The values operate
4210 the same as for rcuperf.nreaders.
4211 N, where N is the number of CPUs
4213 rcuperf.perf_type= [KNL]
4214 Specify the RCU implementation to test.
4216 rcuperf.shutdown= [KNL]
4217 Shut the system down after performance tests
4218 complete. This is useful for hands-off automated
4221 rcuperf.verbose= [KNL]
4222 Enable additional printk() statements.
4224 rcuperf.writer_holdoff= [KNL]
4225 Write-side holdoff between grace periods,
4226 in microseconds. The default of zero says
4229 rcutorture.fqs_duration= [KNL]
4230 Set duration of force_quiescent_state bursts
4233 rcutorture.fqs_holdoff= [KNL]
4234 Set holdoff time within force_quiescent_state bursts
4237 rcutorture.fqs_stutter= [KNL]
4238 Set wait time between force_quiescent_state bursts
4241 rcutorture.fwd_progress= [KNL]
4242 Enable RCU grace-period forward-progress testing
4243 for the types of RCU supporting this notion.
4245 rcutorture.fwd_progress_div= [KNL]
4246 Specify the fraction of a CPU-stall-warning
4247 period to do tight-loop forward-progress testing.
4249 rcutorture.fwd_progress_holdoff= [KNL]
4250 Number of seconds to wait between successive
4251 forward-progress tests.
4253 rcutorture.fwd_progress_need_resched= [KNL]
4254 Enclose cond_resched() calls within checks for
4255 need_resched() during tight-loop forward-progress
4258 rcutorture.gp_cond= [KNL]
4259 Use conditional/asynchronous update-side
4260 primitives, if available.
4262 rcutorture.gp_exp= [KNL]
4263 Use expedited update-side primitives, if available.
4265 rcutorture.gp_normal= [KNL]
4266 Use normal (non-expedited) asynchronous
4267 update-side primitives, if available.
4269 rcutorture.gp_sync= [KNL]
4270 Use normal (non-expedited) synchronous
4271 update-side primitives, if available. If all
4272 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4273 rcutorture.gp_normal=, and rcutorture.gp_sync=
4274 are zero, rcutorture acts as if is interpreted
4275 they are all non-zero.
4277 rcutorture.n_barrier_cbs= [KNL]
4278 Set callbacks/threads for rcu_barrier() testing.
4280 rcutorture.nfakewriters= [KNL]
4281 Set number of concurrent RCU writers. These just
4282 stress RCU, they don't participate in the actual
4283 test, hence the "fake".
4285 rcutorture.nreaders= [KNL]
4286 Set number of RCU readers. The value -1 selects
4287 N-1, where N is the number of CPUs. A value
4288 "n" less than -1 selects N-n-2, where N is again
4289 the number of CPUs. For example, -2 selects N
4290 (the number of CPUs), -3 selects N+1, and so on.
4292 rcutorture.object_debug= [KNL]
4293 Enable debug-object double-call_rcu() testing.
4295 rcutorture.onoff_holdoff= [KNL]
4296 Set time (s) after boot for CPU-hotplug testing.
4298 rcutorture.onoff_interval= [KNL]
4299 Set time (jiffies) between CPU-hotplug operations,
4300 or zero to disable CPU-hotplug testing.
4302 rcutorture.read_exit= [KNL]
4303 Set the number of read-then-exit kthreads used
4304 to test the interaction of RCU updaters and
4305 task-exit processing.
4307 rcutorture.read_exit_burst= [KNL]
4308 The number of times in a given read-then-exit
4309 episode that a set of read-then-exit kthreads
4312 rcutorture.read_exit_delay= [KNL]
4313 The delay, in seconds, between successive
4314 read-then-exit testing episodes.
4316 rcutorture.shuffle_interval= [KNL]
4317 Set task-shuffle interval (s). Shuffling tasks
4318 allows some CPUs to go into dyntick-idle mode
4319 during the rcutorture test.
4321 rcutorture.shutdown_secs= [KNL]
4322 Set time (s) after boot system shutdown. This
4323 is useful for hands-off automated testing.
4325 rcutorture.stall_cpu= [KNL]
4326 Duration of CPU stall (s) to test RCU CPU stall
4327 warnings, zero to disable.
4329 rcutorture.stall_cpu_block= [KNL]
4330 Sleep while stalling if set. This will result
4331 in warnings from preemptible RCU in addition
4332 to any other stall-related activity.
4334 rcutorture.stall_cpu_holdoff= [KNL]
4335 Time to wait (s) after boot before inducing stall.
4337 rcutorture.stall_cpu_irqsoff= [KNL]
4338 Disable interrupts while stalling if set.
4340 rcutorture.stall_gp_kthread= [KNL]
4341 Duration (s) of forced sleep within RCU
4342 grace-period kthread to test RCU CPU stall
4343 warnings, zero to disable. If both stall_cpu
4344 and stall_gp_kthread are specified, the
4345 kthread is starved first, then the CPU.
4347 rcutorture.stat_interval= [KNL]
4348 Time (s) between statistics printk()s.
4350 rcutorture.stutter= [KNL]
4351 Time (s) to stutter testing, for example, specifying
4352 five seconds causes the test to run for five seconds,
4353 wait for five seconds, and so on. This tests RCU's
4354 ability to transition abruptly to and from idle.
4356 rcutorture.test_boost= [KNL]
4357 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4358 "Maybe" means test if the RCU implementation
4359 under test support RCU priority boosting.
4361 rcutorture.test_boost_duration= [KNL]
4362 Duration (s) of each individual boost test.
4364 rcutorture.test_boost_interval= [KNL]
4365 Interval (s) between each boost test.
4367 rcutorture.test_no_idle_hz= [KNL]
4368 Test RCU's dyntick-idle handling. See also the
4369 rcutorture.shuffle_interval parameter.
4371 rcutorture.torture_type= [KNL]
4372 Specify the RCU implementation to test.
4374 rcutorture.verbose= [KNL]
4375 Enable additional printk() statements.
4377 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4378 Dump ftrace buffer after reporting RCU CPU
4381 rcupdate.rcu_cpu_stall_suppress= [KNL]
4382 Suppress RCU CPU stall warning messages.
4384 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4385 Suppress RCU CPU stall warning messages and
4386 rcutorture writer stall warnings that occur
4387 during early boot, that is, during the time
4388 before the init task is spawned.
4390 rcupdate.rcu_cpu_stall_timeout= [KNL]
4391 Set timeout for RCU CPU stall warning messages.
4393 rcupdate.rcu_expedited= [KNL]
4394 Use expedited grace-period primitives, for
4395 example, synchronize_rcu_expedited() instead
4396 of synchronize_rcu(). This reduces latency,
4397 but can increase CPU utilization, degrade
4398 real-time latency, and degrade energy efficiency.
4399 No effect on CONFIG_TINY_RCU kernels.
4401 rcupdate.rcu_normal= [KNL]
4402 Use only normal grace-period primitives,
4403 for example, synchronize_rcu() instead of
4404 synchronize_rcu_expedited(). This improves
4405 real-time latency, CPU utilization, and
4406 energy efficiency, but can expose users to
4407 increased grace-period latency. This parameter
4408 overrides rcupdate.rcu_expedited. No effect on
4409 CONFIG_TINY_RCU kernels.
4411 rcupdate.rcu_normal_after_boot= [KNL]
4412 Once boot has completed (that is, after
4413 rcu_end_inkernel_boot() has been invoked), use
4414 only normal grace-period primitives. No effect
4415 on CONFIG_TINY_RCU kernels.
4417 rcupdate.rcu_task_ipi_delay= [KNL]
4418 Set time in jiffies during which RCU tasks will
4419 avoid sending IPIs, starting with the beginning
4420 of a given grace period. Setting a large
4421 number avoids disturbing real-time workloads,
4422 but lengthens grace periods.
4424 rcupdate.rcu_task_stall_timeout= [KNL]
4425 Set timeout in jiffies for RCU task stall warning
4426 messages. Disable with a value less than or equal
4429 rcupdate.rcu_self_test= [KNL]
4430 Run the RCU early boot self tests
4434 Run specified binary instead of /init from the ramdisk,
4435 used for early userspace startup. See initrd.
4438 force - Override the decision by the kernel to hide the
4439 advertisement of RDRAND support (this affects
4440 certain AMD processors because of buggy BIOS
4441 support, specifically around the suspend/resume
4445 Turn on/off individual RDT features. List is:
4446 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4448 E.g. to turn on cmt and turn off mba use:
4452 Format (x86 or x86_64):
4453 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4455 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4457 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4458 (prefix with 'panic_' to set mode for panic
4460 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4461 reboot_force is either force or not specified,
4462 reboot_cpu is s[mp]#### with #### being the processor
4463 to be used for rebooting.
4465 refscale.holdoff= [KNL]
4466 Set test-start holdoff period. The purpose of
4467 this parameter is to delay the start of the
4468 test until boot completes in order to avoid
4471 refscale.loops= [KNL]
4472 Set the number of loops over the synchronization
4473 primitive under test. Increasing this number
4474 reduces noise due to loop start/end overhead,
4475 but the default has already reduced the per-pass
4476 noise to a handful of picoseconds on ca. 2020
4479 refscale.nreaders= [KNL]
4480 Set number of readers. The default value of -1
4481 selects N, where N is roughly 75% of the number
4482 of CPUs. A value of zero is an interesting choice.
4484 refscale.nruns= [KNL]
4485 Set number of runs, each of which is dumped onto
4488 refscale.readdelay= [KNL]
4489 Set the read-side critical-section duration,
4490 measured in microseconds.
4492 refscale.scale_type= [KNL]
4493 Specify the read-protection implementation to test.
4495 refscale.shutdown= [KNL]
4496 Shut down the system at the end of the performance
4497 test. This defaults to 1 (shut it down) when
4498 rcuperf is built into the kernel and to 0 (leave
4499 it running) when rcuperf is built as a module.
4501 refscale.verbose= [KNL]
4502 Enable additional printk() statements.
4505 [KNL, SMP] Set scheduler's default relax_domain_level.
4506 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4508 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4509 Format: <base1>,<size1>[,<base2>,<size2>,...]
4510 Reserve I/O ports or memory so the kernel won't use
4511 them. If <base> is less than 0x10000, the region
4512 is assumed to be I/O ports; otherwise it is memory.
4514 reservetop= [X86-32]
4516 Reserves a hole at the top of the kernel virtual
4521 Set the amount of memory to reserve for BIOS at
4522 the bottom of the address space.
4524 reset_devices [KNL] Force drivers to reset the underlying device
4525 during initialization.
4528 Specify the partition device for software suspend
4530 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4532 resume_offset= [SWSUSP]
4533 Specify the offset from the beginning of the partition
4534 given by "resume=" at which the swap header is located,
4535 in <PAGE_SIZE> units (needed only for swap files).
4536 See Documentation/power/swsusp-and-swap-files.rst
4538 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4539 read the resume files
4541 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4542 Useful for devices that are detected asynchronously
4543 (e.g. USB and MMC devices).
4545 hibernate= [HIBERNATION]
4546 noresume Don't check if there's a hibernation image
4547 present during boot.
4548 nocompress Don't compress/decompress hibernation images.
4549 no Disable hibernation and resume.
4550 protect_image Turn on image protection during restoration
4551 (that will set all pages holding image data
4552 during restoration read-only).
4554 retain_initrd [RAM] Keep initrd memory after extraction
4556 rfkill.default_state=
4557 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4558 etc. communication is blocked by default.
4561 rfkill.master_switch_mode=
4562 0 The "airplane mode" button does nothing.
4563 1 The "airplane mode" button toggles between everything
4564 blocked and the previous configuration.
4565 2 The "airplane mode" button toggles between everything
4566 blocked and everything unblocked.
4568 rhash_entries= [KNL,NET]
4569 Set number of hash buckets for route cache
4572 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4575 ro [KNL] Mount root device read-only on boot
4578 on Mark read-only kernel memory as read-only (default).
4579 off Leave read-only kernel memory writable for debugging.
4582 Enable the uart passthrough on the designated usb port
4583 on Rockchip SoCs. When active, the signals of the
4584 debug-uart get routed to the D+ and D- pins of the usb
4585 port and the regular usb controller gets disabled.
4587 root= [KNL] Root filesystem
4588 See name_to_dev_t comment in init/do_mounts.c.
4590 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4591 mount the root filesystem
4593 rootflags= [KNL] Set root filesystem mount option string
4595 rootfstype= [KNL] Set root filesystem type
4597 rootwait [KNL] Wait (indefinitely) for root device to show up.
4598 Useful for devices that are detected asynchronously
4599 (e.g. USB and MMC devices).
4601 rproc_mem=nn[KMG][@address]
4602 [KNL,ARM,CMA] Remoteproc physical memory block.
4603 Memory area to be used by remote processor image,
4606 rw [KNL] Mount root device read-write on boot
4608 S [KNL] Run init in single mode
4610 s390_iommu= [HW,S390]
4611 Set s390 IOTLB flushing mode
4613 With strict flushing every unmap operation will result in
4614 an IOTLB flush. Default is lazy flushing before reuse,
4618 See drivers/net/irda/sa1100_ir.c.
4620 sbni= [NET] Granch SBNI12 leased line adapter
4622 sched_debug [KNL] Enables verbose scheduler debug messages.
4624 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4625 Allowed values are enable and disable. This feature
4626 incurs a small amount of overhead in the scheduler
4627 but is useful for debugging and performance tuning.
4629 sched_thermal_decay_shift=
4630 [KNL, SMP] Set a decay shift for scheduler thermal
4631 pressure signal. Thermal pressure signal follows the
4632 default decay period of other scheduler pelt
4633 signals(usually 32 ms but configurable). Setting
4634 sched_thermal_decay_shift will left shift the decay
4635 period for the thermal pressure signal by the shift
4637 i.e. with the default pelt decay period of 32 ms
4638 sched_thermal_decay_shift thermal pressure decay pr
4642 Format: integer between 0 and 10
4645 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4646 xtime_lock contention on larger systems, and/or RCU lock
4647 contention on all systems with CONFIG_MAXSMP set.
4648 Format: { "0" | "1" }
4649 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4651 Note: increases power consumption, thus should only be
4652 enabled if running jitter sensitive (HPC/RT) workloads.
4654 security= [SECURITY] Choose a legacy "major" security module to
4655 enable at boot. This has been deprecated by the
4658 selinux= [SELINUX] Disable or enable SELinux at boot time.
4659 Format: { "0" | "1" }
4660 See security/selinux/Kconfig help text.
4665 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4666 Format: { "0" | "1" }
4667 See security/apparmor/Kconfig help text
4670 Default value is set via kernel config option.
4672 serialnumber [BUGS=X86-32]
4675 Maximal number of shapers.
4683 Disable merging of slabs with similar size. May be
4684 necessary if there is some reason to distinguish
4685 allocs to different slabs, especially in hardened
4686 environments where the risk of heap overflows and
4687 layout control by attackers can usually be
4688 frustrated by disabling merging. This will reduce
4689 most of the exposure of a heap attack to a single
4690 cache (risks via metadata attacks are mostly
4691 unchanged). Debug options disable merging on their
4693 For more information see Documentation/vm/slub.rst.
4695 slab_max_order= [MM, SLAB]
4696 Determines the maximum allowed order for slabs.
4697 A high setting may cause OOMs due to memory
4698 fragmentation. Defaults to 1 for systems with
4699 more than 32MB of RAM, 0 otherwise.
4701 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
4702 Enabling slub_debug allows one to determine the
4703 culprit if slab objects become corrupted. Enabling
4704 slub_debug can create guard zones around objects and
4705 may poison objects when not in use. Also tracks the
4706 last alloc / free. For more information see
4707 Documentation/vm/slub.rst.
4709 slub_memcg_sysfs= [MM, SLUB]
4710 Determines whether to enable sysfs directories for
4711 memory cgroup sub-caches. 1 to enable, 0 to disable.
4712 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4713 Enabling this can lead to a very high number of debug
4714 directories and files being created under
4717 slub_max_order= [MM, SLUB]
4718 Determines the maximum allowed order for slabs.
4719 A high setting may cause OOMs due to memory
4720 fragmentation. For more information see
4721 Documentation/vm/slub.rst.
4723 slub_min_objects= [MM, SLUB]
4724 The minimum number of objects per slab. SLUB will
4725 increase the slab order up to slub_max_order to
4726 generate a sufficiently large slab able to contain
4727 the number of objects indicated. The higher the number
4728 of objects the smaller the overhead of tracking slabs
4729 and the less frequently locks need to be acquired.
4730 For more information see Documentation/vm/slub.rst.
4732 slub_min_order= [MM, SLUB]
4733 Determines the minimum page order for slabs. Must be
4734 lower than slub_max_order.
4735 For more information see Documentation/vm/slub.rst.
4737 slub_nomerge [MM, SLUB]
4738 Same with slab_nomerge. This is supported for legacy.
4739 See slab_nomerge for more information.
4742 Format: <io1>[,<io2>[,...,<io8>]]
4744 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4745 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4746 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4747 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4748 smsc-ircc2.ircc_irq= [HW] IRQ line
4749 smsc-ircc2.ircc_dma= [HW] DMA channel
4750 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4751 0: Toshiba Satellite 1800 (GP data pin select)
4752 1: Fast pin select (default)
4755 smt [KNL,S390] Set the maximum number of threads (logical
4756 CPUs) to use per physical CPU on systems capable of
4757 symmetric multithreading (SMT). Will be capped to the
4758 actual hardware limit.
4760 Default: -1 (no limit)
4763 [KNL] Should the soft-lockup detector generate panics.
4766 A value of 1 instructs the soft-lockup detector
4767 to panic the machine when a soft-lockup occurs. It is
4768 also controlled by the kernel.softlockup_panic sysctl
4769 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
4770 respective build-time switch to that functionality.
4772 softlockup_all_cpu_backtrace=
4773 [KNL] Should the soft-lockup detector generate
4774 backtraces on all cpus.
4777 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4778 See Documentation/admin-guide/laptops/sonypi.rst
4780 spectre_v2= [X86] Control mitigation of Spectre variant 2
4781 (indirect branch speculation) vulnerability.
4782 The default operation protects the kernel from
4785 on - unconditionally enable, implies
4787 off - unconditionally disable, implies
4789 auto - kernel detects whether your CPU model is
4792 Selecting 'on' will, and 'auto' may, choose a
4793 mitigation method at run time according to the
4794 CPU, the available microcode, the setting of the
4795 CONFIG_RETPOLINE configuration option, and the
4796 compiler with which the kernel was built.
4798 Selecting 'on' will also enable the mitigation
4799 against user space to user space task attacks.
4801 Selecting 'off' will disable both the kernel and
4802 the user space protections.
4804 Specific mitigations can also be selected manually:
4806 retpoline - replace indirect branches
4807 retpoline,generic - google's original retpoline
4808 retpoline,amd - AMD-specific minimal thunk
4810 Not specifying this option is equivalent to
4814 [X86] Control mitigation of Spectre variant 2
4815 (indirect branch speculation) vulnerability between
4818 on - Unconditionally enable mitigations. Is
4819 enforced by spectre_v2=on
4821 off - Unconditionally disable mitigations. Is
4822 enforced by spectre_v2=off
4824 prctl - Indirect branch speculation is enabled,
4825 but mitigation can be enabled via prctl
4826 per thread. The mitigation control state
4827 is inherited on fork.
4830 - Like "prctl" above, but only STIBP is
4831 controlled per thread. IBPB is issued
4832 always when switching between different user
4836 - Same as "prctl" above, but all seccomp
4837 threads will enable the mitigation unless
4838 they explicitly opt out.
4841 - Like "seccomp" above, but only STIBP is
4842 controlled per thread. IBPB is issued
4843 always when switching between different
4844 user space processes.
4846 auto - Kernel selects the mitigation depending on
4847 the available CPU features and vulnerability.
4850 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4852 Not specifying this option is equivalent to
4853 spectre_v2_user=auto.
4855 spec_store_bypass_disable=
4856 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4857 (Speculative Store Bypass vulnerability)
4859 Certain CPUs are vulnerable to an exploit against a
4860 a common industry wide performance optimization known
4861 as "Speculative Store Bypass" in which recent stores
4862 to the same memory location may not be observed by
4863 later loads during speculative execution. The idea
4864 is that such stores are unlikely and that they can
4865 be detected prior to instruction retirement at the
4866 end of a particular speculation execution window.
4868 In vulnerable processors, the speculatively forwarded
4869 store can be used in a cache side channel attack, for
4870 example to read memory to which the attacker does not
4871 directly have access (e.g. inside sandboxed code).
4873 This parameter controls whether the Speculative Store
4874 Bypass optimization is used.
4876 On x86 the options are:
4878 on - Unconditionally disable Speculative Store Bypass
4879 off - Unconditionally enable Speculative Store Bypass
4880 auto - Kernel detects whether the CPU model contains an
4881 implementation of Speculative Store Bypass and
4882 picks the most appropriate mitigation. If the
4883 CPU is not vulnerable, "off" is selected. If the
4884 CPU is vulnerable the default mitigation is
4885 architecture and Kconfig dependent. See below.
4886 prctl - Control Speculative Store Bypass per thread
4887 via prctl. Speculative Store Bypass is enabled
4888 for a process by default. The state of the control
4889 is inherited on fork.
4890 seccomp - Same as "prctl" above, but all seccomp threads
4891 will disable SSB unless they explicitly opt out.
4893 Default mitigations:
4894 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4896 On powerpc the options are:
4898 on,auto - On Power8 and Power9 insert a store-forwarding
4899 barrier on kernel entry and exit. On Power7
4900 perform a software flush on kernel entry and
4904 Not specifying this option is equivalent to
4905 spec_store_bypass_disable=auto.
4907 spia_io_base= [HW,MTD]
4913 [X86] Enable split lock detection
4915 When enabled (and if hardware support is present), atomic
4916 instructions that access data across cache line
4917 boundaries will result in an alignment check exception.
4921 warn - the kernel will emit rate limited warnings
4922 about applications triggering the #AC
4923 exception. This mode is the default on CPUs
4924 that supports split lock detection.
4926 fatal - the kernel will send SIGBUS to applications
4927 that trigger the #AC exception.
4929 If an #AC exception is hit in the kernel or in
4930 firmware (i.e. not while executing in user mode)
4931 the kernel will oops in either "warn" or "fatal"
4935 Control the Special Register Buffer Data Sampling
4938 Certain CPUs are vulnerable to an MDS-like
4939 exploit which can leak bits from the random
4942 By default, this issue is mitigated by
4943 microcode. However, the microcode fix can cause
4944 the RDRAND and RDSEED instructions to become
4945 much slower. Among other effects, this will
4946 result in reduced throughput from /dev/urandom.
4948 The microcode mitigation can be disabled with
4949 the following option:
4951 off: Disable mitigation and remove
4952 performance impact to RDRAND and RDSEED
4954 srcutree.counter_wrap_check [KNL]
4955 Specifies how frequently to check for
4956 grace-period sequence counter wrap for the
4957 srcu_data structure's ->srcu_gp_seq_needed field.
4958 The greater the number of bits set in this kernel
4959 parameter, the less frequently counter wrap will
4960 be checked for. Note that the bottom two bits
4963 srcutree.exp_holdoff [KNL]
4964 Specifies how many nanoseconds must elapse
4965 since the end of the last SRCU grace period for
4966 a given srcu_struct until the next normal SRCU
4967 grace period will be considered for automatic
4968 expediting. Set to zero to disable automatic
4972 Speculative Store Bypass Disable control
4974 On CPUs that are vulnerable to the Speculative
4975 Store Bypass vulnerability and offer a
4976 firmware based mitigation, this parameter
4977 indicates how the mitigation should be used:
4979 force-on: Unconditionally enable mitigation for
4980 for both kernel and userspace
4981 force-off: Unconditionally disable mitigation for
4982 for both kernel and userspace
4983 kernel: Always enable mitigation in the
4984 kernel, and offer a prctl interface
4985 to allow userspace to register its
4986 interest in being mitigated too.
4988 stack_guard_gap= [MM]
4989 override the default stack gap protection. The value
4990 is in page units and it defines how many pages prior
4991 to (for stacks growing down) resp. after (for stacks
4992 growing up) the main stack are reserved for no other
4993 mapping. Default value is 256 pages.
4996 Enabled the stack tracer on boot up.
4998 stacktrace_filter=[function-list]
4999 [FTRACE] Limit the functions that the stack tracer
5000 will trace at boot up. function-list is a comma separated
5001 list of functions. This list can be changed at run
5002 time by the stack_trace_filter file in the debugfs
5003 tracing directory. Note, this enables stack tracing
5004 and the stacktrace above is not needed.
5008 Set the STI (builtin display/keyboard on the HP-PARISC
5009 machines) console (graphic card) which should be used
5010 as the initial boot-console.
5011 See also comment in drivers/video/console/sticore.c.
5014 See comment in drivers/video/console/sticore.c.
5017 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5019 sunrpc.min_resvport=
5020 sunrpc.max_resvport=
5022 SunRPC servers often require that client requests
5023 originate from a privileged port (i.e. a port in the
5024 range 0 < portnr < 1024).
5025 An administrator who wishes to reserve some of these
5026 ports for other uses may adjust the range that the
5027 kernel's sunrpc client considers to be privileged
5028 using these two parameters to set the minimum and
5029 maximum port values.
5031 sunrpc.svc_rpc_per_connection_limit=
5033 Limit the number of requests that the server will
5034 process in parallel from a single connection.
5035 The default value is 0 (no limit).
5039 Control how the NFS server code allocates CPUs to
5040 service thread pools. Depending on how many NICs
5041 you have and where their interrupts are bound, this
5042 option will affect which CPUs will do NFS serving.
5043 Note: this parameter cannot be changed while the
5044 NFS server is running.
5046 auto the server chooses an appropriate mode
5047 automatically using heuristics
5048 global a single global pool contains all CPUs
5049 percpu one pool for each CPU
5050 pernode one pool for each NUMA node (equivalent
5051 to global on non-NUMA machines)
5053 sunrpc.tcp_slot_table_entries=
5054 sunrpc.udp_slot_table_entries=
5056 Sets the upper limit on the number of simultaneous
5057 RPC calls that can be sent from the client to a
5058 server. Increasing these values may allow you to
5059 improve throughput, but will also increase the
5060 amount of memory reserved for use by the client.
5062 suspend.pm_test_delay=
5064 Sets the number of seconds to remain in a suspend test
5065 mode before resuming the system (see
5066 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5067 is set. Default value is 5.
5070 Format: { on | off | y | n | 1 | 0 }
5071 This parameter controls use of the Protected
5072 Execution Facility on pSeries.
5075 [KNL] Enable accounting of swap in memory resource
5076 controller if no parameter or 1 is given or disable
5077 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5079 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5080 Format: { <int> | force | noforce }
5081 <int> -- Number of I/O TLB slabs
5082 force -- force using of bounce buffers even if they
5083 wouldn't be automatically used by the kernel
5084 noforce -- Never use bounce buffers (for debugging)
5089 Set a sysctl parameter, right before loading the init
5090 process, as if the value was written to the respective
5091 /proc/sys/... file. Both '.' and '/' are recognized as
5092 separators. Unrecognized parameters and invalid values
5093 are reported in the kernel log. Sysctls registered
5094 later by a loaded module cannot be set this way.
5095 Example: sysctl.vm.swappiness=40
5097 sysfs.deprecated=0|1 [KNL]
5098 Enable/disable old style sysfs layout for old udev
5099 on older distributions. When this option is enabled
5100 very new udev will not work anymore. When this option
5101 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5102 in older udev will not work anymore.
5103 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5104 the kernel configuration.
5106 sysrq_always_enabled
5108 Ignore sysrq setting - this boot parameter will
5109 neutralize any effect of /proc/sys/kernel/sysrq.
5110 Useful for debugging.
5112 tcpmhash_entries= [KNL,NET]
5113 Set the number of tcp_metrics_hash slots.
5114 Default value is 8192 or 16384 depending on total
5115 ram pages. This is used to specify the TCP metrics
5116 cache size. See Documentation/networking/ip-sysctl.rst
5117 "tcp_no_metrics_save" section for more details.
5121 test_suspend= [SUSPEND][,N]
5122 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5123 standby suspend) or "freeze" (for suspend type freeze)
5124 as the system sleep state during system startup with
5125 the optional capability to repeat N number of times.
5126 The system is woken from this state using a
5127 wakeup-capable RTC alarm.
5129 thash_entries= [KNL,NET]
5130 Set number of hash buckets for TCP connection
5132 thermal.act= [HW,ACPI]
5133 -1: disable all active trip points in all thermal zones
5134 <degrees C>: override all lowest active trip points
5136 thermal.crt= [HW,ACPI]
5137 -1: disable all critical trip points in all thermal zones
5138 <degrees C>: override all critical trip points
5140 thermal.nocrt= [HW,ACPI]
5141 Set to disable actions on ACPI thermal zone
5142 critical and hot trip points.
5144 thermal.off= [HW,ACPI]
5145 1: disable ACPI thermal control
5147 thermal.psv= [HW,ACPI]
5148 -1: disable all passive trip points
5149 <degrees C>: override all passive trip points to this
5152 thermal.tzp= [HW,ACPI]
5153 Specify global default ACPI thermal zone polling rate
5154 <deci-seconds>: poll all this frequency
5155 0: no polling (default)
5158 Force threading of all interrupt handlers except those
5159 marked explicitly IRQF_NO_THREAD.
5163 Specify if the kernel should make use of the cpu
5164 topology information if the hardware supports this.
5165 The scheduler will make use of this information and
5166 e.g. base its process migration decisions on it.
5169 topology_updates= [KNL, PPC, NUMA]
5171 Specify if the kernel should ignore (off)
5172 topology updates sent by the hypervisor to this
5175 torture.disable_onoff_at_boot= [KNL]
5176 Prevent the CPU-hotplug component of torturing
5177 until after init has spawned.
5179 torture.ftrace_dump_at_shutdown= [KNL]
5180 Dump the ftrace buffer at torture-test shutdown,
5181 even if there were no errors. This can be a
5182 very costly operation when many torture tests
5183 are running concurrently, especially on systems
5184 with rotating-rust storage.
5188 tpm_suspend_pcr=[HW,TPM]
5189 Format: integer pcr id
5190 Specify that at suspend time, the tpm driver
5191 should extend the specified pcr with zeros,
5192 as a workaround for some chips which fail to
5193 flush the last written pcr on TPM_SaveState.
5194 This will guarantee that all the other pcrs
5197 trace_buf_size=nn[KMG]
5198 [FTRACE] will set tracing buffer size on each cpu.
5200 trace_event=[event-list]
5201 [FTRACE] Set and start specified trace events in order
5202 to facilitate early boot debugging. The event-list is a
5203 comma separated list of trace events to enable. See
5204 also Documentation/trace/events.rst
5206 trace_options=[option-list]
5207 [FTRACE] Enable or disable tracer options at boot.
5208 The option-list is a comma delimited list of options
5209 that can be enabled or disabled just as if you were
5210 to echo the option name into
5212 /sys/kernel/debug/tracing/trace_options
5214 For example, to enable stacktrace option (to dump the
5215 stack trace of each event), add to the command line:
5217 trace_options=stacktrace
5219 See also Documentation/trace/ftrace.rst "trace options"
5223 Have the tracepoints sent to printk as well as the
5224 tracing ring buffer. This is useful for early boot up
5225 where the system hangs or reboots and does not give the
5226 option for reading the tracing buffer or performing a
5227 ftrace_dump_on_oops.
5229 To turn off having tracepoints sent to printk,
5230 echo 0 > /proc/sys/kernel/tracepoint_printk
5231 Note, echoing 1 into this file without the
5232 tracepoint_printk kernel cmdline option has no effect.
5236 Having tracepoints sent to printk() and activating high
5237 frequency tracepoints such as irq or sched, can cause
5238 the system to live lock.
5241 [FTRACE] enable this option to disable tracing when a
5242 warning is hit. This turns off "tracing_on". Tracing can
5243 be enabled again by echoing '1' into the "tracing_on"
5244 file located in /sys/kernel/debug/tracing/
5246 This option is useful, as it disables the trace before
5247 the WARNING dump is called, which prevents the trace to
5248 be filled with content caused by the warning output.
5250 This option can also be set at run time via the sysctl
5251 option: kernel/traceoff_on_warning
5253 transparent_hugepage=
5255 Format: [always|madvise|never]
5256 Can be used to control the default behavior of the system
5257 with respect to transparent hugepages.
5258 See Documentation/admin-guide/mm/transhuge.rst
5261 tsc= Disable clocksource stability checks for TSC.
5263 [x86] reliable: mark tsc clocksource as reliable, this
5264 disables clocksource verification at runtime, as well
5265 as the stability checks done at bootup. Used to enable
5266 high-resolution timer mode on older hardware, and in
5267 virtualized environment.
5268 [x86] noirqtime: Do not use TSC to do irq accounting.
5269 Used to run time disable IRQ_TIME_ACCOUNTING on any
5270 platforms where RDTSC is slow and this accounting
5272 [x86] unstable: mark the TSC clocksource as unstable, this
5273 marks the TSC unconditionally unstable at bootup and
5274 avoids any further wobbles once the TSC watchdog notices.
5275 [x86] nowatchdog: disable clocksource watchdog. Used
5276 in situations with strict latency requirements (where
5277 interruptions from clocksource watchdog are not
5280 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5281 value instead. Useful when the early TSC frequency discovery
5282 procedure is not reliable, such as on overclocked systems
5283 with CPUID.16h support and partial CPUID.15h support.
5284 Format: <unsigned int>
5286 tsx= [X86] Control Transactional Synchronization
5287 Extensions (TSX) feature in Intel processors that
5288 support TSX control.
5290 This parameter controls the TSX feature. The options are:
5292 on - Enable TSX on the system. Although there are
5293 mitigations for all known security vulnerabilities,
5294 TSX has been known to be an accelerator for
5295 several previous speculation-related CVEs, and
5296 so there may be unknown security risks associated
5297 with leaving it enabled.
5299 off - Disable TSX on the system. (Note that this
5300 option takes effect only on newer CPUs which are
5301 not vulnerable to MDS, i.e., have
5302 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5303 the new IA32_TSX_CTRL MSR through a microcode
5304 update. This new MSR allows for the reliable
5305 deactivation of the TSX functionality.)
5307 auto - Disable TSX if X86_BUG_TAA is present,
5308 otherwise enable TSX on the system.
5310 Not specifying this option is equivalent to tsx=off.
5312 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5315 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5316 Abort (TAA) vulnerability.
5318 Similar to Micro-architectural Data Sampling (MDS)
5319 certain CPUs that support Transactional
5320 Synchronization Extensions (TSX) are vulnerable to an
5321 exploit against CPU internal buffers which can forward
5322 information to a disclosure gadget under certain
5325 In vulnerable processors, the speculatively forwarded
5326 data can be used in a cache side channel attack, to
5327 access data to which the attacker does not have direct
5330 This parameter controls the TAA mitigation. The
5333 full - Enable TAA mitigation on vulnerable CPUs
5336 full,nosmt - Enable TAA mitigation and disable SMT on
5337 vulnerable CPUs. If TSX is disabled, SMT
5338 is not disabled because CPU is not
5339 vulnerable to cross-thread TAA attacks.
5340 off - Unconditionally disable TAA mitigation
5342 On MDS-affected machines, tsx_async_abort=off can be
5343 prevented by an active MDS mitigation as both vulnerabilities
5344 are mitigated with the same mechanism so in order to disable
5345 this mitigation, you need to specify mds=off too.
5347 Not specifying this option is equivalent to
5348 tsx_async_abort=full. On CPUs which are MDS affected
5349 and deploy MDS mitigation, TAA mitigation is not
5350 required and doesn't provide any additional
5354 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5356 turbografx.map[2|3]= [HW,JOY]
5357 TurboGraFX parallel port interface
5359 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5360 See also Documentation/input/devices/joystick-parport.rst
5362 udbg-immortal [PPC] When debugging early kernel crashes that
5363 happen after console_init() and before a proper
5364 console driver takes over, this boot options might
5365 help "seeing" what's going on.
5367 uhash_entries= [KNL,NET]
5368 Set number of hash buckets for UDP/UDP-Lite connections
5371 [USB] Ignore overcurrent events (default N).
5372 Some badly-designed motherboards generate lots of
5373 bogus events, for ports that aren't wired to
5374 anything. Set this parameter to avoid log spamming.
5375 Note that genuine overcurrent events won't be
5379 [X86] Cause panic on unknown NMI.
5381 usbcore.authorized_default=
5382 [USB] Default USB device authorization:
5383 (default -1 = authorized except for wireless USB,
5384 0 = not authorized, 1 = authorized, 2 = authorized
5385 if device connected to internal port)
5387 usbcore.autosuspend=
5388 [USB] The autosuspend time delay (in seconds) used
5389 for newly-detected USB devices (default 2). This
5390 is the time required before an idle device will be
5391 autosuspended. Devices for which the delay is set
5392 to a negative value won't be autosuspended at all.
5394 usbcore.usbfs_snoop=
5395 [USB] Set to log all usbfs traffic (default 0 = off).
5397 usbcore.usbfs_snoop_max=
5398 [USB] Maximum number of bytes to snoop in each URB
5401 usbcore.blinkenlights=
5402 [USB] Set to cycle leds on hubs (default 0 = off).
5404 usbcore.old_scheme_first=
5405 [USB] Start with the old device initialization
5406 scheme (default 0 = off).
5408 usbcore.usbfs_memory_mb=
5409 [USB] Memory limit (in MB) for buffers allocated by
5410 usbfs (default = 16, 0 = max = 2047).
5412 usbcore.use_both_schemes=
5413 [USB] Try the other device initialization scheme
5414 if the first one fails (default 1 = enabled).
5416 usbcore.initial_descriptor_timeout=
5417 [USB] Specifies timeout for the initial 64-byte
5418 USB_REQ_GET_DESCRIPTOR request in milliseconds
5419 (default 5000 = 5.0 seconds).
5421 usbcore.nousb [USB] Disable the USB subsystem
5424 [USB] A list of quirk entries to augment the built-in
5425 usb core quirk list. List entries are separated by
5426 commas. Each entry has the form
5427 VendorID:ProductID:Flags. The IDs are 4-digit hex
5428 numbers and Flags is a set of letters. Each letter
5429 will change the built-in quirk; setting it if it is
5430 clear and clearing it if it is set. The letters have
5431 the following meanings:
5432 a = USB_QUIRK_STRING_FETCH_255 (string
5433 descriptors must not be fetched using
5435 b = USB_QUIRK_RESET_RESUME (device can't resume
5436 correctly so reset it instead);
5437 c = USB_QUIRK_NO_SET_INTF (device can't handle
5438 Set-Interface requests);
5439 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5440 handle its Configuration or Interface
5442 e = USB_QUIRK_RESET (device can't be reset
5443 (e.g morph devices), don't use reset);
5444 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5445 more interface descriptions than the
5446 bNumInterfaces count, and can't handle
5447 talking to these interfaces);
5448 g = USB_QUIRK_DELAY_INIT (device needs a pause
5449 during initialization, after we read
5450 the device descriptor);
5451 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5452 high speed and super speed interrupt
5453 endpoints, the USB 2.0 and USB 3.0 spec
5454 require the interval in microframes (1
5455 microframe = 125 microseconds) to be
5456 calculated as interval = 2 ^
5458 Devices with this quirk report their
5459 bInterval as the result of this
5460 calculation instead of the exponent
5461 variable used in the calculation);
5462 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5463 handle device_qualifier descriptor
5465 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5466 generates spurious wakeup, ignore
5467 remote wakeup capability);
5468 k = USB_QUIRK_NO_LPM (device can't handle Link
5470 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5471 (Device reports its bInterval as linear
5472 frames instead of the USB 2.0
5474 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5475 to be disconnected before suspend to
5476 prevent spurious wakeup);
5477 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5478 pause after every control message);
5479 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5480 delay after resetting its port);
5481 Example: quirks=0781:5580:bk,0a5c:5834:gij
5484 [USBHID] The interval which mice are to be polled at.
5487 [USBHID] The interval which joysticks are to be polled at.
5490 [USBHID] The interval which keyboards are to be polled at.
5492 usb-storage.delay_use=
5493 [UMS] The delay in seconds before a new device is
5494 scanned for Logical Units (default 1).
5497 [UMS] A list of quirks entries to supplement or
5498 override the built-in unusual_devs list. List
5499 entries are separated by commas. Each entry has
5500 the form VID:PID:Flags where VID and PID are Vendor
5501 and Product ID values (4-digit hex numbers) and
5502 Flags is a set of characters, each corresponding
5503 to a common usb-storage quirk flag as follows:
5504 a = SANE_SENSE (collect more than 18 bytes
5505 of sense data, not on uas);
5506 b = BAD_SENSE (don't collect more than 18
5507 bytes of sense data, not on uas);
5508 c = FIX_CAPACITY (decrease the reported
5509 device capacity by one sector);
5510 d = NO_READ_DISC_INFO (don't use
5511 READ_DISC_INFO command, not on uas);
5512 e = NO_READ_CAPACITY_16 (don't use
5513 READ_CAPACITY_16 command);
5514 f = NO_REPORT_OPCODES (don't use report opcodes
5516 g = MAX_SECTORS_240 (don't transfer more than
5517 240 sectors at a time, uas only);
5518 h = CAPACITY_HEURISTICS (decrease the
5519 reported device capacity by one
5520 sector if the number is odd);
5521 i = IGNORE_DEVICE (don't bind to this
5523 j = NO_REPORT_LUNS (don't use report luns
5525 l = NOT_LOCKABLE (don't try to lock and
5526 unlock ejectable media, not on uas);
5527 m = MAX_SECTORS_64 (don't transfer more
5528 than 64 sectors = 32 KB at a time,
5530 n = INITIAL_READ10 (force a retry of the
5531 initial READ(10) command, not on uas);
5532 o = CAPACITY_OK (accept the capacity
5533 reported by the device, not on uas);
5534 p = WRITE_CACHE (the device cache is ON
5535 by default, not on uas);
5536 r = IGNORE_RESIDUE (the device reports
5537 bogus residue values, not on uas);
5538 s = SINGLE_LUN (the device has only one
5540 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5541 commands, uas only);
5542 u = IGNORE_UAS (don't bind to the uas driver);
5543 w = NO_WP_DETECT (don't test whether the
5544 medium is write-protected).
5545 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5546 even if the device claims no cache,
5548 Example: quirks=0419:aaf5:rl,0421:0433:rc
5550 user_debug= [KNL,ARM]
5552 See arch/arm/Kconfig.debug help text.
5553 1 - undefined instruction events
5555 4 - invalid data aborts
5558 Example: user_debug=31
5561 [X86] Flags controlling user PTE allocations.
5563 nohigh = do not allocate PTE pages in
5564 HIGHMEM regardless of setting
5568 On X86_32, this is an alias for vdso32=. Otherwise:
5570 vdso=1: enable VDSO (the default)
5571 vdso=0: disable VDSO mapping
5573 vdso32= [X86] Control the 32-bit vDSO
5574 vdso32=1: enable 32-bit VDSO
5575 vdso32=0 or vdso32=2: disable 32-bit VDSO
5577 See the help text for CONFIG_COMPAT_VDSO for more
5578 details. If CONFIG_COMPAT_VDSO is set, the default is
5579 vdso32=0; otherwise, the default is vdso32=1.
5581 For compatibility with older kernels, vdso32=2 is an
5584 Try vdso32=0 if you encounter an error that says:
5585 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5588 vector=percpu: enable percpu vector domain
5590 video= [FB] Frame buffer configuration
5591 See Documentation/fb/modedb.rst.
5593 video.brightness_switch_enabled= [0,1]
5594 If set to 1, on receiving an ACPI notify event
5595 generated by hotkey, video driver will adjust brightness
5596 level and then send out the event to user space through
5597 the allocated input device; If set to 0, video driver
5598 will only send out the event without touching backlight
5603 [VMMIO] Memory mapped virtio (platform) device.
5605 <size>@<baseaddr>:<irq>[:<id>]
5607 <size> := size (can use standard suffixes
5609 <baseaddr> := physical base address
5610 <irq> := interrupt number (as passed to
5612 <id> := (optional) platform device id
5614 virtio_mmio.device=1K@0x100b0000:48:7
5616 Can be used multiple times for multiple devices.
5618 vga= [BOOT,X86-32] Select a particular video mode
5619 See Documentation/x86/boot.rst and
5620 Documentation/admin-guide/svga.rst.
5621 Use vga=ask for menu.
5622 This is actually a boot loader parameter; the value is
5623 passed to the kernel using a special protocol.
5625 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
5626 May slow down system boot speed, especially when
5627 enabled on systems with a large amount of memory.
5628 All options are enabled by default, and this
5629 interface is meant to allow for selectively
5630 enabling or disabling specific virtual memory
5633 Available options are:
5634 P Enable page structure init time poisoning
5635 - Disable all of the above options
5637 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5638 size of <nn>. This can be used to increase the
5639 minimum size (128MB on x86). It can also be used to
5640 decrease the size and leave more room for directly
5643 vmcp_cma=nn[MG] [KNL,S390]
5644 Sets the memory size reserved for contiguous memory
5645 allocations for the vmcp device driver.
5647 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5650 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5653 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5657 Controls the behavior of vsyscalls (i.e. calls to
5658 fixed addresses of 0xffffffffff600x00 from legacy
5659 code). Most statically-linked binaries and older
5660 versions of glibc use these calls. Because these
5661 functions are at fixed addresses, they make nice
5662 targets for exploits that can control RIP.
5664 emulate [default] Vsyscalls turn into traps and are
5665 emulated reasonably safely. The vsyscall
5668 xonly Vsyscalls turn into traps and are
5669 emulated reasonably safely. The vsyscall
5670 page is not readable.
5672 none Vsyscalls don't work at all. This makes
5673 them quite hard to use for exploits but
5674 might break your system.
5676 vt.color= [VT] Default text color.
5677 Format: 0xYX, X = foreground, Y = background.
5678 Default: 0x07 = light gray on black.
5680 vt.cur_default= [VT] Default cursor shape.
5681 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5682 the parameters of the <Esc>[?A;B;Cc escape sequence;
5683 see VGA-softcursor.txt. Default: 2 = underline.
5685 vt.default_blu= [VT]
5686 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5687 Change the default blue palette of the console.
5688 This is a 16-member array composed of values
5691 vt.default_grn= [VT]
5692 Format: <green0>,<green1>,<green2>,...,<green15>
5693 Change the default green palette of the console.
5694 This is a 16-member array composed of values
5697 vt.default_red= [VT]
5698 Format: <red0>,<red1>,<red2>,...,<red15>
5699 Change the default red palette of the console.
5700 This is a 16-member array composed of values
5706 Set system-wide default UTF-8 mode for all tty's.
5707 Default is 1, i.e. UTF-8 mode is enabled for all
5708 newly opened terminals.
5710 vt.global_cursor_default=
5713 Set system-wide default for whether a cursor
5714 is shown on new VTs. Default is -1,
5715 i.e. cursors will be created by default unless
5716 overridden by individual drivers. 0 will hide
5717 cursors, 1 will display them.
5719 vt.italic= [VT] Default color for italic text; 0-15.
5722 vt.underline= [VT] Default color for underlined text; 0-15.
5725 watchdog timers [HW,WDT] For information on watchdog timers,
5726 see Documentation/watchdog/watchdog-parameters.rst
5727 or other driver-specific files in the
5728 Documentation/watchdog/ directory.
5732 Set the hard lockup detector stall duration
5733 threshold in seconds. The soft lockup detector
5734 threshold is set to twice the value. A value of 0
5735 disables both lockup detectors. Default is 10
5738 workqueue.watchdog_thresh=
5739 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5740 warn stall conditions and dump internal state to
5741 help debugging. 0 disables workqueue stall
5742 detection; otherwise, it's the stall threshold
5743 duration in seconds. The default value is 30 and
5744 it can be updated at runtime by writing to the
5745 corresponding sysfs file.
5747 workqueue.disable_numa
5748 By default, all work items queued to unbound
5749 workqueues are affine to the NUMA nodes they're
5750 issued on, which results in better behavior in
5751 general. If NUMA affinity needs to be disabled for
5752 whatever reason, this option can be used. Note
5753 that this also can be controlled per-workqueue for
5754 workqueues visible under /sys/bus/workqueue/.
5756 workqueue.power_efficient
5757 Per-cpu workqueues are generally preferred because
5758 they show better performance thanks to cache
5759 locality; unfortunately, per-cpu workqueues tend to
5760 be more power hungry than unbound workqueues.
5762 Enabling this makes the per-cpu workqueues which
5763 were observed to contribute significantly to power
5764 consumption unbound, leading to measurably lower
5765 power usage at the cost of small performance
5768 The default value of this parameter is determined by
5769 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5771 workqueue.debug_force_rr_cpu
5772 Workqueue used to implicitly guarantee that work
5773 items queued without explicit CPU specified are put
5774 on the local CPU. This guarantee is no longer true
5775 and while local CPU is still preferred work items
5776 may be put on foreign CPUs. This debug option
5777 forces round-robin CPU selection to flush out
5778 usages which depend on the now broken guarantee.
5779 When enabled, memory and cache locality will be
5782 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5783 default x2apic cluster mode on platforms
5786 x86_intel_mid_timer= [X86-32,APBT]
5787 Choose timer option for x86 Intel MID platform.
5788 Two valid options are apbt timer only and lapic timer
5789 plus one apbt timer for broadcast timer.
5790 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5792 xen_512gb_limit [KNL,X86-64,XEN]
5793 Restricts the kernel running paravirtualized under Xen
5794 to use only up to 512 GB of RAM. The reason to do so is
5795 crash analysis tools and Xen tools for doing domain
5796 save/restore/migration must be enabled to handle larger
5799 xen_emul_unplug= [HW,X86,XEN]
5800 Unplug Xen emulated devices
5801 Format: [unplug0,][unplug1]
5802 ide-disks -- unplug primary master IDE devices
5803 aux-ide-disks -- unplug non-primary-master IDE devices
5804 nics -- unplug network devices
5805 all -- unplug all emulated devices (NICs and IDE disks)
5806 unnecessary -- unplugging emulated devices is
5807 unnecessary even if the host did not respond to
5809 never -- do not unplug even if version check succeeds
5811 xen_legacy_crash [X86,XEN]
5812 Crash from Xen panic notifier, without executing late
5813 panic() code such as dumping handler.
5815 xen_nopvspin [X86,XEN]
5816 Disables the qspinlock slowpath using Xen PV optimizations.
5817 This parameter is obsoleted by "nopvspin" parameter, which
5818 has equivalent effect for XEN platform.
5821 Disables the PV optimizations forcing the HVM guest to
5822 run as generic HVM guest with no PV drivers.
5823 This option is obsoleted by the "nopv" option, which
5824 has equivalent effect for XEN platform.
5826 xen_scrub_pages= [XEN]
5827 Boolean option to control scrubbing pages before giving them back
5828 to Xen, for use by other domains. Can be also changed at runtime
5829 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5830 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5832 xen_timer_slop= [X86-64,XEN]
5833 Set the timer slop (in nanoseconds) for the virtual Xen
5834 timers (default is 100000). This adjusts the minimum
5835 delta of virtualized Xen timers, where lower values
5836 improve timer resolution at the expense of processing
5837 more timer interrupts.
5839 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
5840 Disables the PV optimizations forcing the guest to run
5841 as generic guest with no PV drivers. Currently support
5842 XEN HVM, KVM, HYPER_V and VMWARE guest.
5844 nopvspin [X86,XEN,KVM]
5845 Disables the qspinlock slow path using PV optimizations
5846 which allow the hypervisor to 'idle' the guest on lock
5849 xirc2ps_cs= [NET,PCMCIA]
5851 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5854 By default on POWER9 and above, the kernel will
5855 natively use the XIVE interrupt controller. This option
5856 allows the fallback firmware mode to be used:
5858 off Fallback to firmware control of XIVE interrupt
5859 controller on both pseries and powernv
5860 platforms. Only useful on POWER9 and above.
5862 xhci-hcd.quirks [USB,KNL]
5863 A hex value specifying bitmask with supplemental xhci
5864 host controller quirks. Meaning of each bit can be
5865 consulted in header drivers/usb/host/xhci.h.
5868 Format: { early | on | rw | ro | off }
5869 Controls if xmon debugger is enabled. Default is off.
5870 Passing only "xmon" is equivalent to "xmon=early".
5871 early Call xmon as early as possible on boot; xmon
5872 debugger is called from setup_arch().
5873 on xmon debugger hooks will be installed so xmon
5874 is only called on a kernel crash. Default mode,
5875 i.e. either "ro" or "rw" mode, is controlled
5876 with CONFIG_XMON_DEFAULT_RO_MODE.
5877 rw xmon debugger hooks will be installed so xmon
5878 is called only on a kernel crash, mode is write,
5879 meaning SPR registers, memory and, other data
5880 can be written using xmon commands.
5881 ro same as "rw" option above but SPR registers,
5882 memory, and other data can't be written using
5884 off xmon is disabled.