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 clocksource.max_cswd_read_retries= [KNL]
581 Number of clocksource_watchdog() retries due to
582 external delays before the clock will be marked
583 unstable. Defaults to three retries, that is,
584 four attempts to read the clock under test.
586 clearcpuid=BITNUM[,BITNUM...] [X86]
587 Disable CPUID feature X for the kernel. See
588 arch/x86/include/asm/cpufeatures.h for the valid bit
589 numbers. Note the Linux specific bits are not necessarily
590 stable over kernel options, but the vendor specific
592 Also note that user programs calling CPUID directly
593 or using the feature without checking anything
594 will still see it. This just prevents it from
595 being used by the kernel or shown in /proc/cpuinfo.
596 Also note the kernel might malfunction if you disable
599 cma=nn[MG]@[start[MG][-end[MG]]]
601 Sets the size of kernel global memory area for
602 contiguous memory allocations and optionally the
603 placement constraint by the physical address range of
604 memory allocations. A value of 0 disables CMA
605 altogether. For more information, see
606 kernel/dma/contiguous.c
610 Sets the size of kernel per-numa memory area for
611 contiguous memory allocations. A value of 0 disables
612 per-numa CMA altogether. And If this option is not
613 specificed, the default value is 0.
614 With per-numa CMA enabled, DMA users on node nid will
615 first try to allocate buffer from the pernuma area
616 which is located in node nid, if the allocation fails,
617 they will fallback to the global default memory area.
619 cmo_free_hint= [PPC] Format: { yes | no }
620 Specify whether pages are marked as being inactive
621 when they are freed. This is used in CMO environments
622 to determine OS memory pressure for page stealing by
626 coherent_pool=nn[KMG] [ARM,KNL]
627 Sets the size of memory pool for coherent, atomic dma
628 allocations, by default set to 256K.
630 com20020= [HW,NET] ARCnet - COM20020 chipset
632 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
634 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
638 ARCnet - COM90xx chipset (memory-mapped buffers)
639 Format: <io>[,<irq>[,<memstart>]]
641 condev= [HW,S390] console device
644 console= [KNL] Output console device and options.
646 tty<n> Use the virtual console device <n>.
650 Use the specified serial port. The options are of
651 the form "bbbbpnf", where "bbbb" is the baud rate,
652 "p" is parity ("n", "o", or "e"), "n" is number of
653 bits, and "f" is flow control ("r" for RTS or
654 omit it). Default is "9600n8".
656 See Documentation/admin-guide/serial-console.rst for more
658 Documentation/networking/netconsole.rst for an
661 uart[8250],io,<addr>[,options]
662 uart[8250],mmio,<addr>[,options]
663 uart[8250],mmio16,<addr>[,options]
664 uart[8250],mmio32,<addr>[,options]
665 uart[8250],0x<addr>[,options]
666 Start an early, polled-mode console on the 8250/16550
667 UART at the specified I/O port or MMIO address,
668 switching to the matching ttyS device later.
669 MMIO inter-register address stride is either 8-bit
670 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
671 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
672 to be equivalent to 'mmio'. 'options' are specified in
673 the same format described for ttyS above; if unspecified,
674 the h/w is not re-initialized.
676 hvc<n> Use the hypervisor console device <n>. This is for
677 both Xen and PowerPC hypervisors.
679 If the device connected to the port is not a TTY but a braille
680 device, prepend "brl," before the device type, for instance
682 For now, only VisioBraille is supported.
685 [KNL] Change console messages format
687 By default we print messages on consoles in
688 "[time stamp] text\n" format (time stamp may not be
689 printed, depending on CONFIG_PRINTK_TIME or
690 `printk_time' param).
692 Switch to syslog format: "<%u>[time stamp] text\n"
693 IOW, each message will have a facility and loglevel
694 prefix. The format is similar to one used by syslog()
695 syscall, or to executing "dmesg -S --raw" or to reading
698 consoleblank= [KNL] The console blank (screen saver) timeout in
699 seconds. A value of 0 disables the blank timer.
703 [KNL] Change the default value for
704 /proc/<pid>/coredump_filter.
705 See also Documentation/filesystems/proc.rst.
707 coresight_cpu_debug.enable
710 Enable/disable the CPU sampling based debugging.
711 0: default value, disable debugging
712 1: enable debugging at boot time
714 cpuidle.off=1 [CPU_IDLE]
715 disable the cpuidle sub-system
718 [CPU_IDLE] Name of the cpuidle governor to use.
720 cpufreq.off=1 [CPU_FREQ]
721 disable the cpufreq sub-system
723 cpufreq.default_governor=
724 [CPU_FREQ] Name of the default cpufreq governor or
725 policy to use. This governor must be registered in the
726 kernel before the cpufreq driver probes.
729 [X86] Delay for N microsec between assert and de-assert
730 of APIC INIT to start processors. This delay occurs
731 on every CPU online, such as boot, and resume from suspend.
734 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
736 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
738 crashkernel=size[KMG][@offset[KMG]]
739 [KNL] Using kexec, Linux can switch to a 'crash kernel'
740 upon panic. This parameter reserves the physical
741 memory region [offset, offset + size] for that kernel
742 image. If '@offset' is omitted, then a suitable offset
743 is selected automatically.
744 [KNL, X86-64] Select a region under 4G first, and
745 fall back to reserve region above 4G when '@offset'
746 hasn't been specified.
747 See Documentation/admin-guide/kdump/kdump.rst for further details.
749 crashkernel=range1:size1[,range2:size2,...][@offset]
750 [KNL] Same as above, but depends on the memory
751 in the running system. The syntax of range is
752 start-[end] where start and end are both
753 a memory unit (amount[KMG]). See also
754 Documentation/admin-guide/kdump/kdump.rst for an example.
756 crashkernel=size[KMG],high
757 [KNL, X86-64] range could be above 4G. Allow kernel
758 to allocate physical memory region from top, so could
759 be above 4G if system have more than 4G ram installed.
760 Otherwise memory region will be allocated below 4G, if
762 It will be ignored if crashkernel=X is specified.
763 crashkernel=size[KMG],low
764 [KNL, X86-64] range under 4G. When crashkernel=X,high
765 is passed, kernel could allocate physical memory region
766 above 4G, that cause second kernel crash on system
767 that require some amount of low memory, e.g. swiotlb
768 requires at least 64M+32K low memory, also enough extra
769 low memory is needed to make sure DMA buffers for 32-bit
770 devices won't run out. Kernel would try to allocate at
771 at least 256M below 4G automatically.
772 This one let user to specify own low range under 4G
773 for second kernel instead.
774 0: to disable low allocation.
775 It will be ignored when crashkernel=X,high is not used
776 or memory reserved is below 4G.
779 [KNL] Disable crypto self-tests
784 cs89x0_media= [HW,NET]
785 Format: { rj45 | aui | bnc }
788 See header of drivers/s390/block/dasd_devmap.c.
790 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
791 (one device per port)
792 Format: <port#>,<type>
793 See also Documentation/input/devices/joystick-parport.rst
795 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
797 Documentation/admin-guide/dynamic-debug-howto.rst for
798 details. Deprecated, see dyndbg.
800 debug [KNL] Enable kernel debugging (events log level).
803 [KNL] Enable printing [hashed] pointers early in the
804 boot sequence. If enabled, we use a weak hash instead
805 of siphash to hash pointers. Use this option if you are
806 seeing instances of '(___ptrval___)') and need to see a
807 value (hashed pointer) instead. Cryptographically
808 insecure, please do not use on production kernels.
811 [KNL] verbose self-tests
813 Print debugging info while doing the locking API
815 We default to 0 (no extra messages), setting it to
816 1 will print _a lot_ more information - normally
817 only useful to kernel developers.
819 debug_objects [KNL] Enable object debugging
822 [KNL] Disable object debugging
824 debug_guardpage_minorder=
825 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
826 parameter allows control of the order of pages that will
827 be intentionally kept free (and hence protected) by the
828 buddy allocator. Bigger value increase the probability
829 of catching random memory corruption, but reduce the
830 amount of memory for normal system use. The maximum
831 possible value is MAX_ORDER/2. Setting this parameter
832 to 1 or 2 should be enough to identify most random
833 memory corruption problems caused by bugs in kernel or
834 driver code when a CPU writes to (or reads from) a
835 random memory location. Note that there exists a class
836 of memory corruptions problems caused by buggy H/W or
837 F/W or by drivers badly programing DMA (basically when
838 memory is written at bus level and the CPU MMU is
839 bypassed) which are not detectable by
840 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
841 tracking down these problems.
844 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
845 enables the feature at boot time. By default, it is
846 disabled and the system will work mostly the same as a
847 kernel built without CONFIG_DEBUG_PAGEALLOC.
848 Note: to get most of debug_pagealloc error reports, it's
849 useful to also enable the page_owner functionality.
850 on: enable the feature
852 debugfs= [KNL] This parameter enables what is exposed to userspace
853 and debugfs internal clients.
854 Format: { on, no-mount, off }
855 on: All functions are enabled.
857 Filesystem is not registered but kernel clients can
858 access APIs and a crashkernel can be used to read
859 its content. There is nothing to mount.
860 off: Filesystem is not registered and clients
861 get a -EPERM as result when trying to register files
862 or directories within debugfs.
863 This is equivalent of the runtime functionality if
864 debugfs was not enabled in the kernel at all.
865 Default value is set in build-time with a kernel configuration.
867 debugpat [X86] Enable PAT debugging
870 [HW] The size of the default HugeTLB page. This is
871 the size represented by the legacy /proc/ hugepages
872 APIs. In addition, this is the default hugetlb size
873 used for shmget(), mmap() and mounting hugetlbfs
874 filesystems. If not specified, defaults to the
875 architecture's default huge page size. Huge page
876 sizes are architecture dependent. See also
877 Documentation/admin-guide/mm/hugetlbpage.rst.
880 deferred_probe_timeout=
881 [KNL] Debugging option to set a timeout in seconds for
882 deferred probe to give up waiting on dependencies to
883 probe. Only specific dependencies (subsystems or
884 drivers) that have opted in will be ignored. A timeout of 0
885 will timeout at the end of initcalls. This option will also
886 dump out devices still on the deferred probe list after
890 Format: { on | off | def_only | inf_only | always }
891 on: s390 zlib hardware support for compression on
892 level 1 and decompression (default)
893 off: No s390 zlib hardware support
894 def_only: s390 zlib hardware support for deflate
895 only (compression on level 1)
896 inf_only: s390 zlib hardware support for inflate
898 always: Same as 'on' but ignores the selected compression
899 level always using hardware support (used for debugging)
902 Set number of hash buckets for dentry cache.
904 disable_1tb_segments [PPC]
905 Disables the use of 1TB hash page table segments. This
906 causes the kernel to fall back to 256MB segments which
907 can be useful when debugging issues that require an SLB
911 Limits the number of kernel SLB entries, and flushes
912 them frequently to increase the rate of SLB faults
916 See Documentation/networking/ipv6.rst.
919 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
920 hardening is enabled for this boot. Hardened
921 usercopy checking is used to protect the kernel
922 from reading or writing beyond known memory
923 allocation boundaries as a proactive defense
924 against bounds-checking flaws in the kernel's
925 copy_to_user()/copy_from_user() interface.
926 on Perform hardened usercopy checks (default).
927 off Disable hardened usercopy checks.
930 Disable RADIX MMU mode on POWER9
932 radix_hcall_invalidate=on [PPC/PSERIES]
933 Disable RADIX GTSE feature and use hcall for TLB
937 Disable TLBIE instruction. Currently does not work
938 with KVM, with HASH MMU, or with coherent accelerators.
940 disable_cpu_apicid= [X86,APIC,SMP]
942 The number of initial APIC ID for the
943 corresponding CPU to be disabled at boot,
944 mostly used for the kdump 2nd kernel to
945 disable BSP to wake up multiple CPUs without
946 causing system reset or hang due to sending
949 perf_v4_pmi= [X86,INTEL]
951 Disable Intel PMU counter freezing feature.
952 The feature only exists starting from
953 Arch Perfmon v4 (Skylake and newer).
955 disable_ddw [PPC/PSERIES]
956 Disable Dynamic DMA Window support. Use this
957 to workaround buggy firmware.
960 See Documentation/networking/ipv6.rst.
962 disable_mtrr_cleanup [X86]
963 The kernel tries to adjust MTRR layout from continuous
964 to discrete, to make X server driver able to add WB
965 entry later. This parameter disables that.
967 disable_mtrr_trim [X86, Intel and AMD only]
968 By default the kernel will trim any uncacheable
969 memory out of your available memory pool based on
970 MTRR settings. This parameter disables that behavior,
971 possibly causing your machine to run very slowly.
973 disable_timer_pin_1 [X86]
974 Disable PIN 1 of APIC timer
975 Can be useful to work around chipset bugs.
977 dis_ucode_ldr [X86] Disable the microcode loader.
979 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
980 this option disables the debugging code at boot.
982 dma_debug_entries=<number>
983 This option allows to tune the number of preallocated
984 entries for DMA-API debugging code. One entry is
985 required per DMA-API allocation. Use this if the
986 DMA-API debugging code disables itself because the
987 architectural default is too low.
989 dma_debug_driver=<driver_name>
990 With this option the DMA-API debugging driver
991 filter feature can be enabled at boot time. Just
992 pass the driver to filter for as the parameter.
993 The filter can be disabled or changed to another
994 driver later using sysfs.
996 reg_file_data_sampling=
997 [X86] Controls mitigation for Register File Data
998 Sampling (RFDS) vulnerability. RFDS is a CPU
999 vulnerability which may allow userspace to infer
1000 kernel data values previously stored in floating point
1001 registers, vector registers, or integer registers.
1002 RFDS only affects Intel Atom processors.
1004 on: Turns ON the mitigation.
1005 off: Turns OFF the mitigation.
1007 This parameter overrides the compile time default set
1008 by CONFIG_MITIGATION_RFDS. Mitigation cannot be
1009 disabled when other VERW based mitigations (like MDS)
1010 are enabled. In order to disable RFDS mitigation all
1011 VERW based mitigations need to be disabled.
1014 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst
1016 driver_async_probe= [KNL]
1017 List of driver names to be probed asynchronously.
1018 Format: <driver_name1>,<driver_name2>...
1020 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1021 Broken monitors, graphic adapters, KVMs and EDIDless
1022 panels may send no or incorrect EDID data sets.
1023 This parameter allows to specify an EDID data sets
1024 in the /lib/firmware directory that are used instead.
1025 Generic built-in EDID data sets are used, if one of
1026 edid/1024x768.bin, edid/1280x1024.bin,
1027 edid/1680x1050.bin, or edid/1920x1080.bin is given
1028 and no file with the same name exists. Details and
1029 instructions how to build your own EDID data are
1030 available in Documentation/admin-guide/edid.rst. An EDID
1031 data set will only be used for a particular connector,
1032 if its name and a colon are prepended to the EDID
1033 name. Each connector may use a unique EDID data
1034 set by separating the files with a comma. An EDID
1035 data set with no connector name will be used for
1036 any connectors not explicitly specified.
1041 Format: {"off" | "known"}
1042 Control how the dt_cpu_ftrs device-tree binding is
1043 used for CPU feature discovery and setup (if it
1045 off: Do not use it, fall back to legacy cpu table.
1046 known: Do not pass through unknown features to guests
1047 or userspace, only those that the kernel is aware of.
1049 dump_apple_properties [X86]
1050 Dump name and content of EFI device properties on
1051 x86 Macs. Useful for driver authors to determine
1052 what data is available or for reverse-engineering.
1054 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1055 <module>.dyndbg[="val"]
1056 Enable debug messages at boot time. See
1057 Documentation/admin-guide/dynamic-debug-howto.rst
1060 nopku [X86] Disable Memory Protection Keys CPU feature found
1063 <module>.async_probe [KNL]
1064 Enable asynchronous probe on this module.
1066 early_ioremap_debug [KNL]
1067 Enable debug messages in early_ioremap support. This
1068 is useful for tracking down temporary early mappings
1069 which are not unmapped.
1071 earlycon= [KNL] Output early console device and options.
1073 When used with no options, the early console is
1074 determined by stdout-path property in device tree's
1075 chosen node or the ACPI SPCR table if supported by
1078 cdns,<addr>[,options]
1079 Start an early, polled-mode console on a Cadence
1080 (xuartps) serial port at the specified address. Only
1081 supported option is baud rate. If baud rate is not
1082 specified, the serial port must already be setup and
1085 uart[8250],io,<addr>[,options]
1086 uart[8250],mmio,<addr>[,options]
1087 uart[8250],mmio32,<addr>[,options]
1088 uart[8250],mmio32be,<addr>[,options]
1089 uart[8250],0x<addr>[,options]
1090 Start an early, polled-mode console on the 8250/16550
1091 UART at the specified I/O port or MMIO address.
1092 MMIO inter-register address stride is either 8-bit
1093 (mmio) or 32-bit (mmio32 or mmio32be).
1094 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1095 to be equivalent to 'mmio'. 'options' are specified
1096 in the same format described for "console=ttyS<n>"; if
1097 unspecified, the h/w is not initialized.
1101 Start an early, polled-mode console on a pl011 serial
1102 port at the specified address. The pl011 serial port
1103 must already be setup and configured. Options are not
1104 yet supported. If 'mmio32' is specified, then only
1105 the driver will use only 32-bit accessors to read/write
1106 the device registers.
1109 Start an early, polled-mode console on a meson serial
1110 port at the specified address. The serial port must
1111 already be setup and configured. Options are not yet
1115 Start an early, polled-mode console on an msm serial
1116 port at the specified address. The serial port
1117 must already be setup and configured. Options are not
1120 msm_serial_dm,<addr>
1121 Start an early, polled-mode console on an msm serial
1122 dm port at the specified address. The serial port
1123 must already be setup and configured. Options are not
1127 Start an early, polled-mode console on a serial port
1128 of an Actions Semi SoC, such as S500 or S900, at the
1129 specified address. The serial port must already be
1130 setup and configured. Options are not yet supported.
1133 Start an early, polled-mode console on a serial port
1134 of an RDA Micro SoC, such as RDA8810PL, at the
1135 specified address. The serial port must already be
1136 setup and configured. Options are not yet supported.
1139 Use RISC-V SBI (Supervisor Binary Interface) for early
1142 smh Use ARM semihosting calls for early console.
1150 Use early console provided by serial driver available
1151 on Samsung SoCs, requires selecting proper type and
1152 a correct base address of the selected UART port. The
1153 serial port must already be setup and configured.
1154 Options are not yet supported.
1157 Start an early, polled-mode console on a lantiq serial
1158 (lqasc) port at the specified address. The serial port
1159 must already be setup and configured. Options are not
1164 Use early console provided by Freescale LP UART driver
1165 found on Freescale Vybrid and QorIQ LS1021A processors.
1166 A valid base address must be provided, and the serial
1167 port must already be setup and configured.
1171 Start an early, polled-mode, output-only console on the
1172 Freescale i.MX UART at the specified address. The UART
1173 must already be setup and configured.
1176 Start an early, polled-mode console on the
1177 Armada 3700 serial port at the specified
1178 address. The serial port must already be setup
1179 and configured. Options are not yet supported.
1182 Start an early, polled-mode console on a Qualcomm
1183 Generic Interface (GENI) based serial port at the
1184 specified address. The serial port must already be
1185 setup and configured. Options are not yet supported.
1188 Start an early, unaccelerated console on the EFI
1189 memory mapped framebuffer (if available). On cache
1190 coherent non-x86 systems that use system memory for
1191 the framebuffer, pass the 'ram' option so that it is
1192 mapped with the correct attributes.
1195 Use early console provided by Freescale LINFlexD UART
1196 serial driver for NXP S32V234 SoCs. A valid base
1197 address must be provided, and the serial port must
1198 already be setup and configured.
1200 earlyprintk= [X86,SH,ARM,M68k,S390]
1204 earlyprintk=serial[,ttySn[,baudrate]]
1205 earlyprintk=serial[,0x...[,baudrate]]
1206 earlyprintk=ttySn[,baudrate]
1207 earlyprintk=dbgp[debugController#]
1208 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1209 earlyprintk=xdbc[xhciController#]
1211 earlyprintk is useful when the kernel crashes before
1212 the normal console is initialized. It is not enabled by
1213 default because it has some cosmetic problems.
1215 Append ",keep" to not disable it when the real console
1218 Only one of vga, efi, serial, or usb debug port can
1221 Currently only ttyS0 and ttyS1 may be specified by
1222 name. Other I/O ports may be explicitly specified
1223 on some architectures (x86 and arm at least) by
1224 replacing ttySn with an I/O port address, like this:
1225 earlyprintk=serial,0x1008,115200
1226 You can find the port for a given device in
1227 /proc/tty/driver/serial:
1228 2: uart:ST16650V2 port:00001008 irq:18 ...
1230 Interaction with the standard serial driver is not
1233 The VGA and EFI output is eventually overwritten by
1236 The xen output can only be used by Xen PV guests.
1238 The sclp output can only be used on s390.
1240 The optional "force" to "pciserial" enables use of a
1241 PCI device even when its classcode is not of the
1244 edac_report= [HW,EDAC] Control how to report EDAC event
1245 Format: {"on" | "off" | "force"}
1246 on: enable EDAC to report H/W event. May be overridden
1247 by other higher priority error reporting module.
1248 off: disable H/W event reporting through EDAC.
1249 force: enforce the use of EDAC to report H/W event.
1252 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1255 This is designed to be used in conjunction with
1256 the boot argument: earlyprintk=vga
1258 This parameter works in place of the kgdboc parameter
1259 but can only be used if the backing tty is available
1260 very early in the boot process. For early debugging
1261 via a serial port see kgdboc_earlycon instead.
1264 Format: {"off" | "on" | "skip[mbr]"}
1267 Format: { "debug", "disable_early_pci_dma",
1268 "nochunk", "noruntime", "nosoftreserve",
1269 "novamap", "no_disable_early_pci_dma" }
1270 debug: enable misc debug output.
1271 disable_early_pci_dma: disable the busmaster bit on all
1272 PCI bridges while in the EFI boot stub.
1273 nochunk: disable reading files in "chunks" in the EFI
1274 boot stub, as chunking can cause problems with some
1275 firmware implementations.
1276 noruntime : disable EFI runtime services support
1277 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1278 attribute may cause the kernel to reserve the
1279 memory range for a memory mapping driver to
1280 claim. Specify efi=nosoftreserve to disable this
1281 reservation and treat the memory by its base type
1282 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1283 novamap: do not call SetVirtualAddressMap().
1284 no_disable_early_pci_dma: Leave the busmaster bit set
1285 on all PCI bridges while in the EFI boot stub
1287 efi_no_storage_paranoia [EFI; X86]
1288 Using this parameter you can use more than 50% of
1289 your efi variable storage. Use this parameter only if
1290 you are really sure that your UEFI does sane gc and
1291 fulfills the spec otherwise your board may brick.
1293 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1294 Add arbitrary attribute to specific memory range by
1295 updating original EFI memory map.
1296 Region of memory which aa attribute is added to is
1299 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1300 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1301 attribute is added to range 0x100000000-0x180000000 and
1302 0x10a0000000-0x1120000000.
1304 If efi_fake_mem=8G@9G:0x40000 is specified, the
1305 EFI_MEMORY_SP(0x40000) attribute is added to
1306 range 0x240000000-0x43fffffff.
1308 Using this parameter you can do debugging of EFI memmap
1309 related features. For example, you can do debugging of
1310 Address Range Mirroring feature even if your box
1311 doesn't support it, or mark specific memory as
1314 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1315 that is to be dynamically loaded by Linux. If there are
1316 multiple variables with the same name but with different
1317 vendor GUIDs, all of them will be loaded. See
1318 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1321 eisa_irq_edge= [PARISC,HW]
1322 See header of drivers/parisc/eisa.c.
1325 See comment before function elanfreq_setup() in
1326 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1328 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1329 Specifies physical address of start of kernel core
1330 image elf header and optionally the size. Generally
1331 kexec loader will pass this option to capture kernel.
1332 See Documentation/admin-guide/kdump/kdump.rst for details.
1334 enable_mtrr_cleanup [X86]
1335 The kernel tries to adjust MTRR layout from continuous
1336 to discrete, to make X server driver able to add WB
1337 entry later. This parameter enables that.
1339 enable_timer_pin_1 [X86]
1340 Enable PIN 1 of APIC timer
1341 Can be useful to work around chipset bugs
1342 (in particular on some ATI chipsets).
1343 The kernel tries to set a reasonable default.
1345 enforcing [SELINUX] Set initial enforcing status.
1347 See security/selinux/Kconfig help text.
1348 0 -- permissive (log only, no denials).
1349 1 -- enforcing (deny and log).
1351 Value can be changed at runtime via
1352 /sys/fs/selinux/enforce.
1355 Disable Error Record Serialization Table (ERST)
1358 ether= [HW,NET] Ethernet cards parameters
1359 This option is obsoleted by the "netdev=" option, which
1360 has equivalent usage. See its documentation for details.
1364 Permit 'security.evm' to be updated regardless of
1365 current integrity status.
1370 fail_make_request=[KNL]
1371 General fault injection mechanism.
1372 Format: <interval>,<probability>,<space>,<times>
1373 See also Documentation/fault-injection/.
1376 Format: { initns | none }
1377 See Documentation/admin-guide/sysctl/net.rst for
1378 fb_tunnels_only_for_init_ns
1381 See Documentation/admin-guide/blockdev/floppy.rst.
1383 force_pal_cache_flush
1384 [IA-64] Avoid check_sal_cache_flush which may hang on
1385 buggy SAL_CACHE_FLUSH implementations. Using this
1386 parameter will force ia64_sal_cache_flush to call
1387 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1390 Forcefully enable Physical Address Extension (PAE).
1391 Many Pentium M systems disable PAE but may have a
1392 functionally usable PAE implementation.
1393 Warning: use of this parameter will taint the kernel
1394 and may cause unknown problems.
1397 [FTRACE] will set and start the specified tracer
1398 as early as possible in order to facilitate early
1401 ftrace_dump_on_oops[=orig_cpu]
1402 [FTRACE] will dump the trace buffers on oops.
1403 If no parameter is passed, ftrace will dump
1404 buffers of all CPUs, but if you pass orig_cpu, it will
1405 dump only the buffer of the CPU that triggered the
1408 ftrace_filter=[function-list]
1409 [FTRACE] Limit the functions traced by the function
1410 tracer at boot up. function-list is a comma separated
1411 list of functions. This list can be changed at run
1412 time by the set_ftrace_filter file in the debugfs
1415 ftrace_notrace=[function-list]
1416 [FTRACE] Do not trace the functions specified in
1417 function-list. This list can be changed at run time
1418 by the set_ftrace_notrace file in the debugfs
1421 ftrace_graph_filter=[function-list]
1422 [FTRACE] Limit the top level callers functions traced
1423 by the function graph tracer at boot up.
1424 function-list is a comma separated list of functions
1425 that can be changed at run time by the
1426 set_graph_function file in the debugfs tracing directory.
1428 ftrace_graph_notrace=[function-list]
1429 [FTRACE] Do not trace from the functions specified in
1430 function-list. This list is a comma separated list of
1431 functions that can be changed at run time by the
1432 set_graph_notrace file in the debugfs tracing directory.
1434 ftrace_graph_max_depth=<uint>
1435 [FTRACE] Used with the function graph tracer. This is
1436 the max depth it will trace into a function. This value
1437 can be changed at run time by the max_graph_depth file
1438 in the tracefs tracing directory. default: 0 (no limit)
1440 fw_devlink= [KNL] Create device links between consumer and supplier
1441 devices by scanning the firmware to infer the
1442 consumer/supplier relationships. This feature is
1443 especially useful when drivers are loaded as modules as
1444 it ensures proper ordering of tasks like device probing
1445 (suppliers first, then consumers), supplier boot state
1446 clean up (only after all consumers have probed),
1447 suspend/resume & runtime PM (consumers first, then
1449 Format: { off | permissive | on | rpm }
1450 off -- Don't create device links from firmware info.
1451 permissive -- Create device links from firmware info
1452 but use it only for ordering boot state clean
1453 up (sync_state() calls).
1454 on -- Create device links from firmware info and use it
1455 to enforce probe and suspend/resume ordering.
1456 rpm -- Like "on", but also use to order runtime PM.
1459 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1460 support via parallel port (up to 5 devices per port)
1461 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1462 See also Documentation/input/devices/joystick-parport.rst
1466 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1470 gather_data_sampling=
1471 [X86,INTEL] Control the Gather Data Sampling (GDS)
1474 Gather Data Sampling is a hardware vulnerability which
1475 allows unprivileged speculative access to data which was
1476 previously stored in vector registers.
1478 This issue is mitigated by default in updated microcode.
1479 The mitigation may have a performance impact but can be
1480 disabled. On systems without the microcode mitigation
1481 disabling AVX serves as a mitigation.
1483 force: Disable AVX to mitigate systems without
1484 microcode mitigation. No effect if the microcode
1485 mitigation is present. Known to cause crashes in
1486 userspace with buggy AVX enumeration.
1488 off: Disable GDS mitigation.
1490 gcov_persist= [GCOV] When non-zero (default), profiling data for
1491 kernel modules is saved and remains accessible via
1492 debugfs, even when the module is unloaded/reloaded.
1493 When zero, profiling data is discarded and associated
1494 debugfs files are removed at module unload time.
1496 goldfish [X86] Enable the goldfish android emulator platform.
1497 Don't use this when you are not running on the
1500 gpt [EFI] Forces disk with valid GPT signature but
1501 invalid Protective MBR to be treated as GPT. If the
1502 primary GPT is corrupted, it enables the backup/alternate
1503 GPT to be used instead.
1505 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1506 the "Enable 0" bit of the configuration register.
1509 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1510 the "Enable 0" bit of the configuration register.
1513 grcan.select= [HW] Select which physical interface to use.
1516 grcan.txsize= [HW] Sets the size of the tx buffer.
1517 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1519 grcan.rxsize= [HW] Sets the size of the rx buffer.
1520 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1523 gpio-mockup.gpio_mockup_ranges
1524 [HW] Sets the ranges of gpiochip of for this device.
1525 Format: <start1>,<end1>,<start2>,<end2>...
1527 hardlockup_all_cpu_backtrace=
1528 [KNL] Should the hard-lockup detector generate
1529 backtraces on all cpus.
1532 hashdist= [KNL,NUMA] Large hashes allocated during boot
1533 are distributed across NUMA nodes. Defaults on
1534 for 64-bit NUMA, off otherwise.
1535 Format: 0 | 1 (for off | on)
1537 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1539 hd= [EIDE] (E)IDE hard drive subsystem geometry
1540 Format: <cyl>,<head>,<sect>
1543 Disable Hardware Error Source Table (HEST) support;
1544 corresponding firmware-first mode error processing
1545 logic will be disabled.
1547 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1548 size of <nn>. This works even on boxes that have no
1549 highmem otherwise. This also works to reduce highmem
1550 size on bigger boxes.
1552 highres= [KNL] Enable/disable high resolution timer mode.
1553 Valid parameters: "on", "off"
1558 hpet= [X86-32,HPET] option to control HPET usage
1559 Format: { enable (default) | disable | force |
1561 disable: disable HPET and use PIT instead
1562 force: allow force enabled of undocumented chips (ICH4,
1564 verbose: show contents of HPET registers during setup
1566 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1567 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1569 hugetlb_cma= [HW] The size of a cma area used for allocation
1570 of gigantic hugepages.
1573 Reserve a cma area of given size and allocate gigantic
1574 hugepages using the cma allocator. If enabled, the
1575 boot-time allocation of gigantic hugepages is skipped.
1577 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1578 If this follows hugepagesz (below), it specifies
1579 the number of pages of hugepagesz to be allocated.
1580 If this is the first HugeTLB parameter on the command
1581 line, it specifies the number of pages to allocate for
1582 the default huge page size. See also
1583 Documentation/admin-guide/mm/hugetlbpage.rst.
1587 [HW] The size of the HugeTLB pages. This is used in
1588 conjunction with hugepages (above) to allocate huge
1589 pages of a specific size at boot. The pair
1590 hugepagesz=X hugepages=Y can be specified once for
1591 each supported huge page size. Huge page sizes are
1592 architecture dependent. See also
1593 Documentation/admin-guide/mm/hugetlbpage.rst.
1597 [KNL] Should the hung task detector generate panics.
1600 A value of 1 instructs the kernel to panic when a
1601 hung task is detected. The default value is controlled
1602 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1603 option. The value selected by this boot parameter can
1604 be changed later by the kernel.hung_task_panic sysctl.
1606 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1607 terminal devices. Valid values: 0..8
1608 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1609 If specified, z/VM IUCV HVC accepts connections
1610 from listed z/VM user IDs only.
1612 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1613 which allow the hypervisor to 'idle' the
1614 guest on lock contention.
1617 Do not unregister boot console at start. This is only
1618 useful for debugging when something happens in the window
1619 between unregistering the boot console and initializing
1622 i2c_bus= [HW] Override the default board specific I2C bus speed
1623 or register an additional I2C bus that is not
1624 registered from board initialization code.
1628 i8042.debug [HW] Toggle i8042 debug mode
1629 i8042.unmask_kbd_data
1630 [HW] Enable printing of interrupt data from the KBD port
1631 (disabled by default, and as a pre-condition
1632 requires that i8042.debug=1 be enabled)
1633 i8042.direct [HW] Put keyboard port into non-translated mode
1634 i8042.dumbkbd [HW] Pretend that controller can only read data from
1635 keyboard and cannot control its state
1636 (Don't attempt to blink the leds)
1637 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1638 i8042.nokbd [HW] Don't check/create keyboard port
1639 i8042.noloop [HW] Disable the AUX Loopback command while probing
1641 i8042.nomux [HW] Don't check presence of an active multiplexing
1643 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1645 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1646 i8042.reset [HW] Reset the controller during init, cleanup and
1647 suspend-to-ram transitions, only during s2r
1648 transitions, or never reset
1649 Format: { 1 | Y | y | 0 | N | n }
1650 1, Y, y: always reset controller
1651 0, N, n: don't ever reset controller
1652 Default: only on s2r transitions on x86; most other
1653 architectures force reset to be always executed
1654 i8042.unlock [HW] Unlock (ignore) the keylock
1655 i8042.kbdreset [HW] Reset device connected to KBD port
1657 [HW] Allow deferred probing upon i8042 probe errors
1661 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1662 indicates that the driver is running on unsupported
1664 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1665 does not match list of supported models.
1667 [HW] Report power status in /proc/i8k
1668 (disabled by default)
1669 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1672 i915.invert_brightness=
1673 [DRM] Invert the sense of the variable that is used to
1674 set the brightness of the panel backlight. Normally a
1675 brightness value of 0 indicates backlight switched off,
1676 and the maximum of the brightness value sets the backlight
1677 to maximum brightness. If this parameter is set to 0
1678 (default) and the machine requires it, or this parameter
1679 is set to 1, a brightness value of 0 sets the backlight
1680 to maximum brightness, and the maximum of the brightness
1681 value switches the backlight off.
1682 -1 -- never invert brightness
1683 0 -- machine default
1684 1 -- force brightness inversion
1687 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1689 ide-core.nodma= [HW] (E)IDE subsystem
1690 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1691 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1692 .cdrom .chs .ignore_cable are additional options
1693 See Documentation/ide/ide.rst.
1695 ide-generic.probe-mask= [HW] (E)IDE subsystem
1697 Probe mask for legacy ISA IDE ports. Depending on
1698 platform up to 6 ports are supported, enabled by
1699 setting corresponding bits in the mask to 1. The
1700 default value is 0x0, which has a special meaning.
1701 On systems that have PCI, it triggers scanning the
1702 PCI bus for the first and the second port, which
1703 are then probed. On systems without PCI the value
1704 of 0x0 enables probing the two first ports as if it
1707 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1708 Claim all unknown PCI IDE storage controllers.
1711 Format: idle=poll, idle=halt, idle=nomwait
1712 Poll forces a polling idle loop that can slightly
1713 improve the performance of waking up a idle CPU, but
1714 will use a lot of power and make the system run hot.
1716 idle=halt: Halt is forced to be used for CPU idle.
1717 In such case C2/C3 won't be used again.
1718 idle=nomwait: Disable mwait for CPU C-states
1720 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1721 Format: { strict | legacy | 2008 | relaxed }
1724 Choose which programs will be accepted for execution
1725 based on the IEEE 754 NaN encoding(s) supported by
1726 the FPU and the NaN encoding requested with the value
1727 of an ELF file header flag individually set by each
1728 binary. Hardware implementations are permitted to
1729 support either or both of the legacy and the 2008 NaN
1732 Available settings are as follows:
1733 strict accept binaries that request a NaN encoding
1734 supported by the FPU
1735 legacy only accept legacy-NaN binaries, if supported
1737 2008 only accept 2008-NaN binaries, if supported
1739 relaxed accept any binaries regardless of whether
1740 supported by the FPU
1742 The FPU emulator is always able to support both NaN
1743 encodings, so if no FPU hardware is present or it has
1744 been disabled with 'nofpu', then the settings of
1745 'legacy' and '2008' strap the emulator accordingly,
1746 'relaxed' straps the emulator for both legacy-NaN and
1747 2008-NaN, whereas 'strict' enables legacy-NaN only on
1748 legacy processors and both NaN encodings on MIPS32 or
1751 The setting for ABS.fmt/NEG.fmt instruction execution
1752 mode generally follows that for the NaN encoding,
1753 except where unsupported by hardware.
1755 ignore_loglevel [KNL]
1756 Ignore loglevel setting - this will print /all/
1757 kernel messages to the console. Useful for debugging.
1758 We also add it as printk module parameter, so users
1759 could change it dynamically, usually by
1760 /sys/module/printk/parameters/ignore_loglevel.
1763 Ignore RLIMIT_DATA setting for data mappings,
1764 print warning at first misuse. Can be changed via
1765 /sys/module/kernel/parameters/ignore_rlimit_data.
1767 ihash_entries= [KNL]
1768 Set number of hash buckets for inode cache.
1770 ima_appraise= [IMA] appraise integrity measurements
1771 Format: { "off" | "enforce" | "fix" | "log" }
1774 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1775 The builtin appraise policy appraises all files
1778 ima_canonical_fmt [IMA]
1779 Use the canonical format for the binary runtime
1780 measurements, instead of host native format.
1783 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1787 The list of supported hash algorithms is defined
1788 in crypto/hash_info.h.
1791 The builtin policies to load during IMA setup.
1792 Format: "tcb | appraise_tcb | secure_boot |
1795 The "tcb" policy measures all programs exec'd, files
1796 mmap'd for exec, and all files opened with the read
1797 mode bit set by either the effective uid (euid=0) or
1800 The "appraise_tcb" policy appraises the integrity of
1801 all files owned by root.
1803 The "secure_boot" policy appraises the integrity
1804 of files (eg. kexec kernel image, kernel modules,
1805 firmware, policy, etc) based on file signatures.
1807 The "fail_securely" policy forces file signature
1808 verification failure also on privileged mounted
1809 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1812 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1813 Load a policy which meets the needs of the Trusted
1814 Computing Base. This means IMA will measure all
1815 programs exec'd, files mmap'd for exec, and all files
1816 opened for read by uid=0.
1819 Select one of defined IMA measurements template formats.
1820 Formats: { "ima" | "ima-ng" | "ima-sig" }
1824 [IMA] Define a custom template format.
1825 Format: { "field1|...|fieldN" }
1827 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1828 Format: <min_file_size>
1829 Set the minimal file size for using asynchronous hash.
1830 If left unspecified, ahash usage is disabled.
1832 ahash performance varies for different data sizes on
1833 different crypto accelerators. This option can be used
1834 to achieve the best performance for a particular HW.
1836 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1838 Set hashing buffer size. Default: 4k.
1840 ahash performance varies for different chunk sizes on
1841 different crypto accelerators. This option can be used
1842 to achieve best performance for particular HW.
1846 Run specified binary instead of /sbin/init as init
1849 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1850 for working out where the kernel is dying during
1853 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1854 initcall functions. Useful for debugging built-in
1855 modules and initcalls.
1857 initrd= [BOOT] Specify the location of the initial ramdisk
1859 initrdmem= [KNL] Specify a physical address and size from which to
1860 load the initrd. If an initrd is compiled in or
1861 specified in the bootparams, it takes priority over this
1863 Format: ss[KMG],nn[KMG]
1866 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1869 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1871 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1873 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1875 init_pkru= [X86] Specify the default memory protection keys rights
1876 register contents for all processes. 0x55555554 by
1877 default (disallow access to all but pkey 0). Can
1878 override in debugfs after boot.
1880 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1883 int_pln_enable [X86] Enable power limit notification interrupt
1885 integrity_audit=[IMA]
1886 Format: { "0" | "1" }
1887 0 -- basic integrity auditing messages. (Default)
1888 1 -- additional integrity auditing messages.
1890 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1892 Enable intel iommu driver.
1894 Disable intel iommu driver.
1895 igfx_off [Default Off]
1896 By default, gfx is mapped as normal device. If a gfx
1897 device has a dedicated DMAR unit, the DMAR unit is
1898 bypassed by not enabling DMAR with this option. In
1899 this case, gfx device will use physical address for
1902 With this option iommu will not optimize to look
1903 for io virtual address below 32-bit forcing dual
1904 address cycle on pci bus for cards supporting greater
1905 than 32-bit addressing. The default is to look
1906 for translation below 32-bit and if not available
1907 then look in the higher range.
1908 strict [Default Off]
1909 With this option on every unmap_single operation will
1910 result in a hardware IOTLB flush operation as opposed
1911 to batching them for performance.
1912 sp_off [Default Off]
1913 By default, super page will be supported if Intel IOMMU
1914 has the capability. With this option, super page will
1917 By default, scalable mode will be disabled even if the
1918 hardware advertises that it has support for the scalable
1919 mode translation. With this option set, scalable mode
1920 will be used on hardware which claims to support it.
1921 tboot_noforce [Default Off]
1922 Do not force the Intel IOMMU enabled under tboot.
1923 By default, tboot will force Intel IOMMU on, which
1924 could harm performance of some high-throughput
1925 devices like 40GBit network cards, even if identity
1927 Note that using this option lowers the security
1928 provided by tboot because it makes the system
1929 vulnerable to DMA attacks.
1930 nobounce [Default off]
1931 Disable bounce buffer for untrusted devices such as
1932 the Thunderbolt devices. This will treat the untrusted
1933 devices as the trusted ones, hence might expose security
1934 risks of DMA attacks.
1936 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1937 0 disables intel_idle and fall back on acpi_idle.
1938 1 to 9 specify maximum depth of C-state.
1942 Do not enable intel_pstate as the default
1943 scaling driver for the supported processors
1945 Use intel_pstate as a scaling driver, but configure it
1946 to work with generic cpufreq governors (instead of
1947 enabling its internal governor). This mode cannot be
1948 used along with the hardware-managed P-states (HWP)
1951 Enable intel_pstate on systems that prohibit it by default
1952 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1953 instead of acpi-cpufreq may disable platform features, such
1954 as thermal controls and power capping, that rely on ACPI
1955 P-States information being indicated to OSPM and therefore
1956 should be used with caution. This option does not work with
1957 processors that aren't supported by the intel_pstate driver
1958 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1960 Do not enable hardware P state control (HWP)
1963 Only load intel_pstate on systems which support
1964 hardware P state control (HWP) if available.
1966 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1967 Description Table, specifies preferred power management
1968 profile as "Enterprise Server" or "Performance Server",
1969 then this feature is turned on by default.
1971 Allow per-logical-CPU P-State performance control limits using
1972 cpufreq sysfs interface
1974 intremap= [X86-64, Intel-IOMMU]
1975 on enable Interrupt Remapping (default)
1976 off disable Interrupt Remapping
1977 nosid disable Source ID checking
1979 BIOS x2APIC opt-out request will be ignored
1980 nopost disable Interrupt Posting
1982 iomem= Disable strict checking of access to MMIO memory
1983 strict regions from userspace.
1998 nobypass [PPC/POWERNV]
1999 Disable IOMMU bypass, using IOMMU for PCI devices.
2001 iommu.strict= [ARM64] Configure TLB invalidation behaviour
2002 Format: { "0" | "1" }
2004 Request that DMA unmap operations use deferred
2005 invalidation of hardware TLBs, for increased
2006 throughput at the cost of reduced device isolation.
2007 Will fall back to strict mode if not supported by
2008 the relevant IOMMU driver.
2009 1 - Strict mode (default).
2010 DMA unmap operations invalidate IOMMU hardware TLBs
2014 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2015 Format: { "0" | "1" }
2016 0 - Use IOMMU translation for DMA.
2017 1 - Bypass the IOMMU for DMA.
2018 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2020 io7= [HW] IO7 for Marvel-based Alpha systems
2021 See comment before marvel_specify_io7 in
2022 arch/alpha/kernel/core_marvel.c.
2024 io_delay= [X86] I/O delay method
2026 Standard port 0x80 based delay
2028 Alternate port 0xed based delay (needed on some systems)
2030 Simple two microseconds delay
2035 See Documentation/admin-guide/nfs/nfsroot.rst.
2037 ipcmni_extend [KNL] Extend the maximum number of unique System V
2038 IPC identifiers from 32,768 to 16,777,216.
2040 irqaffinity= [SMP] Set the default irq affinity mask
2041 The argument is a cpu list, as described above.
2043 irqchip.gicv2_force_probe=
2046 Force the kernel to look for the second 4kB page
2047 of a GICv2 controller even if the memory range
2048 exposed by the device tree is too small.
2050 irqchip.gicv3_nolpi=
2052 Force the kernel to ignore the availability of
2053 LPIs (and by consequence ITSs). Intended for system
2054 that use the kernel as a bootloader, and thus want
2055 to let secondary kernels in charge of setting up
2058 irqchip.gicv3_pseudo_nmi= [ARM64]
2059 Enables support for pseudo-NMIs in the kernel. This
2060 requires the kernel to be built with
2061 CONFIG_ARM64_PSEUDO_NMI.
2064 When an interrupt is not handled search all handlers
2065 for it. Intended to get systems with badly broken
2069 When an interrupt is not handled search all handlers
2070 for it. Also check all handlers each timer
2071 interrupt. Intended to get systems with badly broken
2075 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2077 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2078 [Deprecated - use cpusets instead]
2079 Format: [flag-list,]<cpu-list>
2081 Specify one or more CPUs to isolate from disturbances
2082 specified in the flag list (default: domain):
2085 Disable the tick when a single task runs.
2087 A residual 1Hz tick is offloaded to workqueues, which you
2088 need to affine to housekeeping through the global
2089 workqueue's affinity configured via the
2090 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2091 by using the 'domain' flag described below.
2093 NOTE: by default the global workqueue runs on all CPUs,
2094 so to protect individual CPUs the 'cpumask' file has to
2095 be configured manually after bootup.
2098 Isolate from the general SMP balancing and scheduling
2099 algorithms. Note that performing domain isolation this way
2100 is irreversible: it's not possible to bring back a CPU to
2101 the domains once isolated through isolcpus. It's strongly
2102 advised to use cpusets instead to disable scheduler load
2103 balancing through the "cpuset.sched_load_balance" file.
2104 It offers a much more flexible interface where CPUs can
2105 move in and out of an isolated set anytime.
2107 You can move a process onto or off an "isolated" CPU via
2108 the CPU affinity syscalls or cpuset.
2109 <cpu number> begins at 0 and the maximum value is
2110 "number of CPUs in system - 1".
2114 Isolate from being targeted by managed interrupts
2115 which have an interrupt mask containing isolated
2116 CPUs. The affinity of managed interrupts is
2117 handled by the kernel and cannot be changed via
2118 the /proc/irq/* interfaces.
2120 This isolation is best effort and only effective
2121 if the automatically assigned interrupt mask of a
2122 device queue contains isolated and housekeeping
2123 CPUs. If housekeeping CPUs are online then such
2124 interrupts are directed to the housekeeping CPU
2125 so that IO submitted on the housekeeping CPU
2126 cannot disturb the isolated CPU.
2128 If a queue's affinity mask contains only isolated
2129 CPUs then this parameter has no effect on the
2130 interrupt routing decision, though interrupts are
2131 only delivered when tasks running on those
2132 isolated CPUs submit IO. IO submitted on
2133 housekeeping CPUs has no influence on those
2136 The format of <cpu-list> is described above.
2140 ivrs_ioapic [HW,X86-64]
2141 Provide an override to the IOAPIC-ID<->DEVICE-ID
2142 mapping provided in the IVRS ACPI table.
2143 By default, PCI segment is 0, and can be omitted.
2145 For example, to map IOAPIC-ID decimal 10 to
2146 PCI segment 0x1 and PCI device 00:14.0,
2147 write the parameter as:
2148 ivrs_ioapic=10@0001:00:14.0
2151 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2152 write the parameter as:
2153 ivrs_ioapic[10]=00:14.0
2154 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2155 PCI device 00:14.0 write the parameter as:
2156 ivrs_ioapic[10]=0001:00:14.0
2158 ivrs_hpet [HW,X86-64]
2159 Provide an override to the HPET-ID<->DEVICE-ID
2160 mapping provided in the IVRS ACPI table.
2161 By default, PCI segment is 0, and can be omitted.
2163 For example, to map HPET-ID decimal 10 to
2164 PCI segment 0x1 and PCI device 00:14.0,
2165 write the parameter as:
2166 ivrs_hpet=10@0001:00:14.0
2169 * To map HPET-ID decimal 0 to PCI device 00:14.0
2170 write the parameter as:
2171 ivrs_hpet[0]=00:14.0
2172 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2173 PCI device 00:14.0 write the parameter as:
2174 ivrs_ioapic[10]=0001:00:14.0
2176 ivrs_acpihid [HW,X86-64]
2177 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2178 mapping provided in the IVRS ACPI table.
2179 By default, PCI segment is 0, and can be omitted.
2181 For example, to map UART-HID:UID AMD0020:0 to
2182 PCI segment 0x1 and PCI device ID 00:14.5,
2183 write the parameter as:
2184 ivrs_acpihid=AMD0020:0@0001:00:14.5
2187 * To map UART-HID:UID AMD0020:0 to PCI segment is 0,
2188 PCI device ID 00:14.5, write the parameter as:
2189 ivrs_acpihid[00:14.5]=AMD0020:0
2190 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and
2191 PCI device ID 00:14.5, write the parameter as:
2192 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2194 js= [HW,JOY] Analog joystick
2195 See Documentation/input/joydev/joystick.rst.
2198 When CONFIG_RANDOMIZE_BASE is set, this disables
2199 kernel and module base offset ASLR (Address Space
2200 Layout Randomization).
2203 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2204 report on every invalid memory access. Without this
2205 parameter KASAN will print report only for the first
2210 kernelcore= [KNL,X86,IA-64,PPC]
2211 Format: nn[KMGTPE] | nn% | "mirror"
2212 This parameter specifies the amount of memory usable by
2213 the kernel for non-movable allocations. The requested
2214 amount is spread evenly throughout all nodes in the
2215 system as ZONE_NORMAL. The remaining memory is used for
2216 movable memory in its own zone, ZONE_MOVABLE. In the
2217 event, a node is too small to have both ZONE_NORMAL and
2218 ZONE_MOVABLE, kernelcore memory will take priority and
2219 other nodes will have a larger ZONE_MOVABLE.
2221 ZONE_MOVABLE is used for the allocation of pages that
2222 may be reclaimed or moved by the page migration
2223 subsystem. Note that allocations like PTEs-from-HighMem
2224 still use the HighMem zone if it exists, and the Normal
2225 zone if it does not.
2227 It is possible to specify the exact amount of memory in
2228 the form of "nn[KMGTPE]", a percentage of total system
2229 memory in the form of "nn%", or "mirror". If "mirror"
2230 option is specified, mirrored (reliable) memory is used
2231 for non-movable allocations and remaining memory is used
2232 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2233 are exclusive, so you cannot specify multiple forms.
2235 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2236 Format: <Controller#>[,poll interval]
2237 The controller # is the number of the ehci usb debug
2238 port as it is probed via PCI. The poll interval is
2239 optional and is the number seconds in between
2240 each poll cycle to the debug port in case you need
2241 the functionality for interrupting the kernel with
2242 gdb or control-c on the dbgp connection. When
2243 not using this parameter you use sysrq-g to break into
2244 the kernel debugger.
2246 kgdboc= [KGDB,HW] kgdb over consoles.
2247 Requires a tty driver that supports console polling,
2248 or a supported polling keyboard driver (non-usb).
2249 Serial only format: <serial_device>[,baud]
2250 keyboard only format: kbd
2251 keyboard and serial format: kbd,<serial_device>[,baud]
2252 Optional Kernel mode setting:
2253 kms, kbd format: kms,kbd
2254 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2256 kgdboc_earlycon= [KGDB,HW]
2257 If the boot console provides the ability to read
2258 characters and can work in polling mode, you can use
2259 this parameter to tell kgdb to use it as a backend
2260 until the normal console is registered. Intended to
2261 be used together with the kgdboc parameter which
2262 specifies the normal console to transition to.
2264 The name of the early console should be specified
2265 as the value of this parameter. Note that the name of
2266 the early console might be different than the tty
2267 name passed to kgdboc. It's OK to leave the value
2268 blank and the first boot console that implements
2269 read() will be picked.
2271 kgdbwait [KGDB] Stop kernel execution and enter the
2272 kernel debugger at the earliest opportunity.
2274 kmac= [MIPS] Korina ethernet MAC address.
2275 Configure the RouterBoard 532 series on-chip
2276 Ethernet adapter MAC address.
2278 kmemleak= [KNL] Boot-time kmemleak enable/disable
2279 Valid arguments: on, off
2281 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2284 kprobe_event=[probe-list]
2285 [FTRACE] Add kprobe events and enable at boot time.
2286 The probe-list is a semicolon delimited list of probe
2287 definitions. Each definition is same as kprobe_events
2288 interface, but the parameters are comma delimited.
2289 For example, to add a kprobe event on vfs_read with
2290 arg1 and arg2, add to the command line;
2292 kprobe_event=p,vfs_read,$arg1,$arg2
2294 See also Documentation/trace/kprobetrace.rst "Kernel
2295 Boot Parameter" section.
2297 kpti= [ARM64] Control page table isolation of user
2298 and kernel address spaces.
2299 Default: enabled on cores which need mitigation.
2303 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2304 Default is 0 (don't ignore, but inject #GP)
2306 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2307 Default is false (don't support).
2309 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2314 [KVM] Controls the software workaround for the
2315 X86_BUG_ITLB_MULTIHIT bug.
2316 force : Always deploy workaround.
2317 off : Never deploy workaround.
2318 auto : Deploy workaround based on the presence of
2319 X86_BUG_ITLB_MULTIHIT.
2323 If the software workaround is enabled for the host,
2324 guests do need not to enable it for nested guests.
2326 kvm.nx_huge_pages_recovery_ratio=
2327 [KVM] Controls how many 4KiB pages are periodically zapped
2328 back to huge pages. 0 disables the recovery, otherwise if
2329 the value is N KVM will zap 1/Nth of the 4KiB pages every
2330 minute. The default is 60.
2332 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2333 Default is 1 (enabled)
2335 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2337 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2339 kvm-arm.vgic_v3_group0_trap=
2340 [KVM,ARM] Trap guest accesses to GICv3 group-0
2343 kvm-arm.vgic_v3_group1_trap=
2344 [KVM,ARM] Trap guest accesses to GICv3 group-1
2347 kvm-arm.vgic_v3_common_trap=
2348 [KVM,ARM] Trap guest accesses to GICv3 common
2351 kvm-arm.vgic_v4_enable=
2352 [KVM,ARM] Allow use of GICv4 for direct injection of
2355 kvm_cma_resv_ratio=n [PPC]
2356 Reserves given percentage from system memory area for
2357 contiguous memory allocation for KVM hash pagetable
2359 By default it reserves 5% of total system memory.
2363 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2364 (virtualized MMU) support on capable Intel chips.
2365 Default is 1 (enabled)
2367 kvm-intel.emulate_invalid_guest_state=
2368 [KVM,Intel] Disable emulation of invalid guest state.
2369 Ignored if kvm-intel.enable_unrestricted_guest=1, as
2370 guest state is never invalid for unrestricted guests.
2371 This param doesn't apply to nested guests (L2), as KVM
2372 never emulates invalid L2 guest state.
2373 Default is 1 (enabled)
2375 kvm-intel.flexpriority=
2376 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2377 Default is 1 (enabled)
2380 [KVM,Intel] Enable VMX nesting (nVMX).
2381 Default is 0 (disabled)
2383 kvm-intel.unrestricted_guest=
2384 [KVM,Intel] Disable unrestricted guest feature
2385 (virtualized real and unpaged mode) on capable
2386 Intel chips. Default is 1 (enabled)
2388 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2391 Valid arguments: never, cond, always
2393 always: L1D cache flush on every VMENTER.
2394 cond: Flush L1D on VMENTER only when the code between
2395 VMEXIT and VMENTER can leak host memory.
2396 never: Disables the mitigation
2398 Default is cond (do L1 cache flush in specific instances)
2400 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2401 feature (tagged TLBs) on capable Intel chips.
2402 Default is 1 (enabled)
2404 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2407 The kernel PTE inversion protection is unconditionally
2408 enabled and cannot be disabled.
2411 Provides all available mitigations for the
2412 L1TF vulnerability. Disables SMT and
2413 enables all mitigations in the
2414 hypervisors, i.e. unconditional L1D flush.
2416 SMT control and L1D flush control via the
2417 sysfs interface is still possible after
2418 boot. Hypervisors will issue a warning
2419 when the first VM is started in a
2420 potentially insecure configuration,
2421 i.e. SMT enabled or L1D flush disabled.
2424 Same as 'full', but disables SMT and L1D
2425 flush runtime control. Implies the
2426 'nosmt=force' command line option.
2427 (i.e. sysfs control of SMT is disabled.)
2430 Leaves SMT enabled and enables the default
2431 hypervisor mitigation, i.e. conditional
2434 SMT control and L1D flush control via the
2435 sysfs interface is still possible after
2436 boot. Hypervisors will issue a warning
2437 when the first VM is started in a
2438 potentially insecure configuration,
2439 i.e. SMT enabled or L1D flush disabled.
2443 Disables SMT and enables the default
2444 hypervisor mitigation.
2446 SMT control and L1D flush control via the
2447 sysfs interface is still possible after
2448 boot. Hypervisors will issue a warning
2449 when the first VM is started in a
2450 potentially insecure configuration,
2451 i.e. SMT enabled or L1D flush disabled.
2454 Same as 'flush', but hypervisors will not
2455 warn when a VM is started in a potentially
2456 insecure configuration.
2459 Disables hypervisor mitigations and doesn't
2461 It also drops the swap size and available
2462 RAM limit restriction on both hypervisor and
2467 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2473 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2476 lapic= [X86,APIC] Do not use TSC deadline
2477 value for LAPIC timer one-shot implementation. Default
2478 back to the programmable timer unit in the LAPIC.
2479 Format: notscdeadline
2481 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2484 libata.dma= [LIBATA] DMA control
2485 libata.dma=0 Disable all PATA and SATA DMA
2486 libata.dma=1 PATA and SATA Disk DMA only
2487 libata.dma=2 ATAPI (CDROM) DMA only
2488 libata.dma=4 Compact Flash DMA only
2489 Combinations also work, so libata.dma=3 enables DMA
2490 for disks and CDROMs, but not CFs.
2492 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2493 libata.ignore_hpa=0 keep BIOS limits (default)
2494 libata.ignore_hpa=1 ignore limits, using full disk
2496 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2500 libata.force= [LIBATA] Force configurations. The format is comma
2501 separated list of "[ID:]VAL" where ID is
2502 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2503 matching port, link or device. Basically, it matches
2504 the ATA ID string printed on console by libata. If
2505 the whole ID part is omitted, the last PORT and DEVICE
2506 values are used. If ID hasn't been specified yet, the
2507 configuration applies to all ports, links and devices.
2509 If only DEVICE is omitted, the parameter applies to
2510 the port and all links and devices behind it. DEVICE
2511 number of 0 either selects the first device or the
2512 first fan-out link behind PMP device. It does not
2513 select the host link. DEVICE number of 15 selects the
2514 host link and device attached to it.
2516 The VAL specifies the configuration to force. As long
2517 as there's no ambiguity shortcut notation is allowed.
2518 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2519 The following configurations can be forced.
2521 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2522 Any ID with matching PORT is used.
2524 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2526 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2527 udma[/][16,25,33,44,66,100,133] notation is also
2530 * [no]ncq: Turn on or off NCQ.
2532 * [no]ncqtrim: Turn off queued DSM TRIM.
2534 * nohrst, nosrst, norst: suppress hard, soft
2537 * rstonce: only attempt one reset during
2538 hot-unplug link recovery
2540 * dump_id: dump IDENTIFY data.
2542 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2544 * disable: Disable this device.
2546 If there are multiple matching configurations changing
2547 the same attribute, the last one is used.
2549 memblock=debug [KNL] Enable memblock debug messages.
2551 load_ramdisk= [RAM] [Deprecated]
2553 lockd.nlm_grace_period=P [NFS] Assign grace period.
2556 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2559 lockd.nlm_timeout=T [NFS] Assign timeout value.
2562 lockd.nlm_udpport=M [NFS] Assign UDP port.
2565 lockdown= [SECURITY]
2566 { integrity | confidentiality }
2567 Enable the kernel lockdown feature. If set to
2568 integrity, kernel features that allow userland to
2569 modify the running kernel are disabled. If set to
2570 confidentiality, kernel features that allow userland
2571 to extract confidential information from the kernel
2574 locktorture.nreaders_stress= [KNL]
2575 Set the number of locking read-acquisition kthreads.
2576 Defaults to being automatically set based on the
2577 number of online CPUs.
2579 locktorture.nwriters_stress= [KNL]
2580 Set the number of locking write-acquisition kthreads.
2582 locktorture.onoff_holdoff= [KNL]
2583 Set time (s) after boot for CPU-hotplug testing.
2585 locktorture.onoff_interval= [KNL]
2586 Set time (s) between CPU-hotplug operations, or
2587 zero to disable CPU-hotplug testing.
2589 locktorture.shuffle_interval= [KNL]
2590 Set task-shuffle interval (jiffies). Shuffling
2591 tasks allows some CPUs to go into dyntick-idle
2592 mode during the locktorture test.
2594 locktorture.shutdown_secs= [KNL]
2595 Set time (s) after boot system shutdown. This
2596 is useful for hands-off automated testing.
2598 locktorture.stat_interval= [KNL]
2599 Time (s) between statistics printk()s.
2601 locktorture.stutter= [KNL]
2602 Time (s) to stutter testing, for example,
2603 specifying five seconds causes the test to run for
2604 five seconds, wait for five seconds, and so on.
2605 This tests the locking primitive's ability to
2606 transition abruptly to and from idle.
2608 locktorture.torture_type= [KNL]
2609 Specify the locking implementation to test.
2611 locktorture.verbose= [KNL]
2612 Enable additional printk() statements.
2614 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2617 loglevel= All Kernel Messages with a loglevel smaller than the
2618 console loglevel will be printed to the console. It can
2619 also be changed with klogd or other programs. The
2620 loglevels are defined as follows:
2622 0 (KERN_EMERG) system is unusable
2623 1 (KERN_ALERT) action must be taken immediately
2624 2 (KERN_CRIT) critical conditions
2625 3 (KERN_ERR) error conditions
2626 4 (KERN_WARNING) warning conditions
2627 5 (KERN_NOTICE) normal but significant condition
2628 6 (KERN_INFO) informational
2629 7 (KERN_DEBUG) debug-level messages
2631 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2632 in bytes. n must be a power of two and greater
2633 than the minimal size. The minimal size is defined
2634 by LOG_BUF_SHIFT kernel config parameter. There is
2635 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2636 that allows to increase the default size depending on
2637 the number of CPUs. See init/Kconfig for more details.
2639 logo.nologo [FB] Disables display of the built-in Linux logo.
2640 This may be used to provide more screen space for
2641 kernel log messages and is useful when debugging
2642 kernel boot problems.
2644 lp=0 [LP] Specify parallel ports to use, e.g,
2645 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2646 lp=reset first parallel port). 'lp=0' disables the
2647 lp=auto printer driver. 'lp=reset' (which can be
2648 specified in addition to the ports) causes
2649 attached printers to be reset. Using
2650 lp=port1,port2,... specifies the parallel ports
2651 to associate lp devices with, starting with
2652 lp0. A port specification may be 'none' to skip
2653 that lp device, or a parport name such as
2654 'parport0'. Specifying 'lp=auto' instead of a
2655 port specification list means that device IDs
2656 from each port should be examined, to see if
2657 an IEEE 1284-compliant printer is attached; if
2658 so, the driver will manage that printer.
2659 See also header of drivers/char/lp.c.
2662 Sets loops_per_jiffy to given constant, thus avoiding
2663 time-consuming boot-time autodetection (up to 250 ms per
2664 CPU). 0 enables autodetection (default). To determine
2665 the correct value for your kernel, boot with normal
2666 autodetection and see what value is printed. Note that
2667 on SMP systems the preset will be applied to all CPUs,
2668 which is likely to cause problems if your CPUs need
2669 significantly divergent settings. An incorrect value
2670 will cause delays in the kernel to be wrong, leading to
2671 unpredictable I/O errors and other breakage. Although
2672 unlikely, in the extreme case this might damage your
2676 Format: <io>,<irq>,<dma>
2678 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2681 [SECURITY] Choose order of LSM initialization. This
2682 overrides CONFIG_LSM, and the "security=" parameter.
2684 machvec= [IA-64] Force the use of a particular machine-vector
2685 (machvec) in a generic kernel.
2686 Example: machvec=hpzx1
2688 machtype= [Loongson] Share the same kernel image file between
2689 different yeeloong laptops.
2690 Example: machtype=lemote-yeeloong-2f-7inch
2692 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2693 than or equal to this physical address is ignored.
2695 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2696 will bring up during bootup. maxcpus=n : n >= 0 limits
2697 the kernel to bring up 'n' processors. Surely after
2698 bootup you can bring up the other plugged cpu by executing
2699 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2700 only takes effect during system bootup.
2701 While n=0 is a special case, it is equivalent to "nosmp",
2702 which also disables the IO APIC.
2704 max_loop= [LOOP] The number of loop block devices that get
2705 (loop.max_loop) unconditionally pre-created at init time. The default
2706 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2707 of statically allocating a predefined number, loop
2708 devices can be requested on-demand with the
2709 /dev/loop-control interface.
2711 mce [X86-32] Machine Check Exception
2713 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2715 md= [HW] RAID subsystems devices and level
2716 See Documentation/admin-guide/md.rst.
2719 Format: <first>,<last>
2720 Specifies range of consoles to be captured by the MDA.
2723 Control mitigation for the Micro-architectural Data
2724 Sampling (MDS) vulnerability.
2726 Certain CPUs are vulnerable to an exploit against CPU
2727 internal buffers which can forward information to a
2728 disclosure gadget under certain conditions.
2730 In vulnerable processors, the speculatively
2731 forwarded data can be used in a cache side channel
2732 attack, to access data to which the attacker does
2733 not have direct access.
2735 This parameter controls the MDS mitigation. The
2738 full - Enable MDS mitigation on vulnerable CPUs
2739 full,nosmt - Enable MDS mitigation and disable
2740 SMT on vulnerable CPUs
2741 off - Unconditionally disable MDS mitigation
2743 On TAA-affected machines, mds=off can be prevented by
2744 an active TAA mitigation as both vulnerabilities are
2745 mitigated with the same mechanism so in order to disable
2746 this mitigation, you need to specify tsx_async_abort=off
2749 Not specifying this option is equivalent to
2752 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2754 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2755 Amount of memory to be used in cases as follows:
2758 2 when the kernel is not able to see the whole system memory;
2759 3 memory that lies after 'mem=' boundary is excluded from
2760 the hypervisor, then assigned to KVM guests.
2762 [X86] Work as limiting max address. Use together
2763 with memmap= to avoid physical address space collisions.
2764 Without memmap= PCI devices could be placed at addresses
2765 belonging to unused RAM.
2767 Note that this only takes effects during boot time since
2768 in above case 3, memory may need be hot added after boot
2769 if system memory of hypervisor is not sufficient.
2771 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2775 [KNL,SH] Allow user to override the default size for
2776 per-device physically contiguous DMA buffers.
2778 memhp_default_state=online/offline
2779 [KNL] Set the initial state for the memory hotplug
2780 onlining policy. If not specified, the default value is
2781 set according to the
2782 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2784 See Documentation/admin-guide/mm/memory-hotplug.rst.
2786 memmap=exactmap [KNL,X86] Enable setting of an exact
2787 E820 memory map, as specified by the user.
2788 Such memmap=exactmap lines can be constructed based on
2789 BIOS output or other requirements. See the memmap=nn@ss
2792 memmap=nn[KMG]@ss[KMG]
2793 [KNL] Force usage of a specific region of memory.
2794 Region of memory to be used is from ss to ss+nn.
2795 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2796 which limits max address to nn[KMG].
2797 Multiple different regions can be specified,
2800 memmap=100M@2G,100M#3G,1G!1024G
2802 memmap=nn[KMG]#ss[KMG]
2803 [KNL,ACPI] Mark specific memory as ACPI data.
2804 Region of memory to be marked is from ss to ss+nn.
2806 memmap=nn[KMG]$ss[KMG]
2807 [KNL,ACPI] Mark specific memory as reserved.
2808 Region of memory to be reserved is from ss to ss+nn.
2809 Example: Exclude memory from 0x18690000-0x1869ffff
2810 memmap=64K$0x18690000
2812 memmap=0x10000$0x18690000
2813 Some bootloaders may need an escape character before '$',
2814 like Grub2, otherwise '$' and the following number
2817 memmap=nn[KMG]!ss[KMG]
2818 [KNL,X86] Mark specific memory as protected.
2819 Region of memory to be used, from ss to ss+nn.
2820 The memory region may be marked as e820 type 12 (0xc)
2821 and is NVDIMM or ADR memory.
2823 memmap=<size>%<offset>-<oldtype>+<newtype>
2824 [KNL,ACPI] Convert memory within the specified region
2825 from <oldtype> to <newtype>. If "-<oldtype>" is left
2826 out, the whole region will be marked as <newtype>,
2827 even if previously unavailable. If "+<newtype>" is left
2828 out, matching memory will be removed. Types are
2829 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2830 3 = ACPI, 12 = PRAM.
2832 memory_corruption_check=0/1 [X86]
2833 Some BIOSes seem to corrupt the first 64k of
2834 memory when doing things like suspend/resume.
2835 Setting this option will scan the memory
2836 looking for corruption. Enabling this will
2837 both detect corruption and prevent the kernel
2838 from using the memory being corrupted.
2839 However, its intended as a diagnostic tool; if
2840 repeatable BIOS-originated corruption always
2841 affects the same memory, you can use memmap=
2842 to prevent the kernel from using that memory.
2844 memory_corruption_check_size=size [X86]
2845 By default it checks for corruption in the low
2846 64k, making this memory unavailable for normal
2847 use. Use this parameter to scan for
2848 corruption in more or less memory.
2850 memory_corruption_check_period=seconds [X86]
2851 By default it checks for corruption every 60
2852 seconds. Use this parameter to check at some
2853 other rate. 0 disables periodic checking.
2855 memtest= [KNL,X86,ARM,PPC] Enable memtest
2857 default : 0 <disable>
2858 Specifies the number of memtest passes to be
2859 performed. Each pass selects another test
2860 pattern from a given set of patterns. Memtest
2861 fills the memory with this pattern, validates
2862 memory contents and reserves bad memory
2863 regions that are detected.
2865 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2866 Valid arguments: on, off
2867 Default (depends on kernel configuration option):
2868 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2869 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2870 mem_encrypt=on: Activate SME
2871 mem_encrypt=off: Do not activate SME
2873 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2874 for details on when memory encryption can be activated.
2876 mem_sleep_default= [SUSPEND] Default system suspend mode:
2877 s2idle - Suspend-To-Idle
2878 shallow - Power-On Suspend or equivalent (if supported)
2879 deep - Suspend-To-RAM or equivalent (if supported)
2880 See Documentation/admin-guide/pm/sleep-states.rst.
2882 meye.*= [HW] Set MotionEye Camera parameters
2883 See Documentation/admin-guide/media/meye.rst.
2885 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2886 Multi-Function General Purpose Timers on AMD Geode
2889 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2890 the BIOS has incorrectly applied a workaround. TinyBIOS
2891 version 0.98 is known to be affected, 0.99 fixes the
2892 problem by letting the user disable the workaround.
2896 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2897 physical address is ignored.
2899 mini2440= [ARM,HW,KNL]
2900 Format:[0..2][b][c][t]
2902 MINI2440 configuration specification:
2903 0 - The attached screen is the 3.5" TFT
2904 1 - The attached screen is the 7" TFT
2905 2 - The VGA Shield is attached (1024x768)
2906 Leaving out the screen size parameter will not load
2907 the TFT driver, and the framebuffer will be left
2909 b - Enable backlight. The TFT backlight pin will be
2910 linked to the kernel VESA blanking code and a GPIO
2911 LED. This parameter is not necessary when using the
2913 c - Enable the s3c camera interface.
2914 t - Reserved for enabling touchscreen support. The
2915 touchscreen support is not enabled in the mainstream
2916 kernel as of 2.6.30, a preliminary port can be found
2917 in the "bleeding edge" mini2440 support kernel at
2918 https://repo.or.cz/w/linux-2.6/mini2440.git
2921 [X86,PPC,S390,ARM64] Control optional mitigations for
2922 CPU vulnerabilities. This is a set of curated,
2923 arch-independent options, each of which is an
2924 aggregation of existing arch-specific options.
2927 Disable all optional CPU mitigations. This
2928 improves system performance, but it may also
2929 expose users to several CPU vulnerabilities.
2930 Equivalent to: gather_data_sampling=off [X86]
2932 kvm.nx_huge_pages=off [X86]
2935 mmio_stale_data=off [X86]
2936 no_entry_flush [PPC]
2937 no_uaccess_flush [PPC]
2940 nospectre_v1 [X86,PPC]
2941 nospectre_v2 [X86,PPC,S390,ARM64]
2942 reg_file_data_sampling=off [X86]
2944 spec_store_bypass_disable=off [X86,PPC]
2945 spectre_v2_user=off [X86]
2946 ssbd=force-off [ARM64]
2947 tsx_async_abort=off [X86]
2950 This does not have any effect on
2951 kvm.nx_huge_pages when
2952 kvm.nx_huge_pages=force.
2955 Mitigate all CPU vulnerabilities, but leave SMT
2956 enabled, even if it's vulnerable. This is for
2957 users who don't want to be surprised by SMT
2958 getting disabled across kernel upgrades, or who
2959 have other ways of avoiding SMT-based attacks.
2960 Equivalent to: (default behavior)
2963 Mitigate all CPU vulnerabilities, disabling SMT
2964 if needed. This is for users who always want to
2965 be fully mitigated, even if it means losing SMT.
2966 Equivalent to: l1tf=flush,nosmt [X86]
2967 mds=full,nosmt [X86]
2968 tsx_async_abort=full,nosmt [X86]
2969 mmio_stale_data=full,nosmt [X86]
2970 retbleed=auto,nosmt [X86]
2973 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2974 parameter allows control of the logging verbosity for
2975 the additional memory initialisation checks. A value
2976 of 0 disables mminit logging and a level of 4 will
2977 log everything. Information is printed at KERN_DEBUG
2978 so loglevel=8 may also need to be specified.
2981 [X86,INTEL] Control mitigation for the Processor
2982 MMIO Stale Data vulnerabilities.
2984 Processor MMIO Stale Data is a class of
2985 vulnerabilities that may expose data after an MMIO
2986 operation. Exposed data could originate or end in
2987 the same CPU buffers as affected by MDS and TAA.
2988 Therefore, similar to MDS and TAA, the mitigation
2989 is to clear the affected CPU buffers.
2991 This parameter controls the mitigation. The
2994 full - Enable mitigation on vulnerable CPUs
2996 full,nosmt - Enable mitigation and disable SMT on
2999 off - Unconditionally disable mitigation
3001 On MDS or TAA affected machines,
3002 mmio_stale_data=off can be prevented by an active
3003 MDS or TAA mitigation as these vulnerabilities are
3004 mitigated with the same mechanism so in order to
3005 disable this mitigation, you need to specify
3006 mds=off and tsx_async_abort=off too.
3008 Not specifying this option is equivalent to
3009 mmio_stale_data=full.
3012 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3015 [KNL] When CONFIG_MODULE_SIG is set, this means that
3016 modules without (valid) signatures will fail to load.
3017 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3018 is always true, so this option does nothing.
3020 module_blacklist= [KNL] Do not load a comma-separated list of
3021 modules. Useful for debugging problem modules.
3024 [MOUSE] Maximum time between finger touching and
3025 leaving touchpad surface for touch to be considered
3026 a tap and be reported as a left button click (for
3027 touchpads working in absolute mode only).
3029 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3030 reporting absolute coordinates, such as tablets
3031 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3032 reporting absolute coordinates, such as tablets
3034 movablecore= [KNL,X86,IA-64,PPC]
3035 Format: nn[KMGTPE] | nn%
3036 This parameter is the complement to kernelcore=, it
3037 specifies the amount of memory used for migratable
3038 allocations. If both kernelcore and movablecore is
3039 specified, then kernelcore will be at *least* the
3040 specified value but may be more. If movablecore on its
3041 own is specified, the administrator must be careful
3042 that the amount of memory usable for all allocations
3045 movable_node [KNL] Boot-time switch to make hotplugable memory
3046 NUMA nodes to be movable. This means that the memory
3047 of such nodes will be usable only for movable
3048 allocations which rules out almost all kernel
3049 allocations. Use with caution!
3051 MTD_Partition= [MTD]
3052 Format: <name>,<region-number>,<size>,<offset>
3054 MTD_Region= [MTD] Format:
3055 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3058 See drivers/mtd/parsers/cmdlinepart.c
3060 multitce=off [PPC] This parameter disables the use of the pSeries
3061 firmware feature for updating multiple TCE entries
3064 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3066 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3068 boundary - index of last SLC block on Flex-OneNAND.
3069 The remaining blocks are configured as MLC blocks.
3070 lock - Configure if Flex-OneNAND boundary should be locked.
3071 Once locked, the boundary cannot be changed.
3072 1 indicates lock status, 0 indicates unlock status.
3075 ARM/S3C2412 JIVE boot control
3077 See arch/arm/mach-s3c2412/mach-jive.c
3079 mtouchusb.raw_coordinates=
3080 [HW] Make the MicroTouch USB driver use raw coordinates
3081 ('y', default) or cooked coordinates ('n')
3083 mtrr_chunk_size=nn[KMG] [X86]
3084 used for mtrr cleanup. It is largest continuous chunk
3085 that could hold holes aka. UC entries.
3087 mtrr_gran_size=nn[KMG] [X86]
3088 Used for mtrr cleanup. It is granularity of mtrr block.
3090 Large value could prevent small alignment from
3093 mtrr_spare_reg_nr=n [X86]
3095 Range: 0,7 : spare reg number
3097 Used for mtrr cleanup. It is spare mtrr entries number.
3098 Set to 2 or more if your graphical card needs more.
3100 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3102 netdev= [NET] Network devices parameters
3103 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3104 Note that mem_start is often overloaded to mean
3105 something different and driver-specific.
3106 This usage is only documented in each driver source
3110 [NETFILTER] Enable connection tracking flow accounting
3111 0 to disable accounting
3112 1 to enable accounting
3115 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3116 See Documentation/admin-guide/nfs/nfsroot.rst.
3118 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3119 See Documentation/admin-guide/nfs/nfsroot.rst.
3121 nfsrootdebug [NFS] enable nfsroot debugging messages.
3122 See Documentation/admin-guide/nfs/nfsroot.rst.
3124 nfs.callback_nr_threads=
3125 [NFSv4] set the total number of threads that the
3126 NFS client will assign to service NFSv4 callback
3129 nfs.callback_tcpport=
3130 [NFS] set the TCP port on which the NFSv4 callback
3131 channel should listen.
3134 [NFS] sets the pathname to the program which is used
3135 to update the NFS client cache entries.
3137 nfs.cache_getent_timeout=
3138 [NFS] sets the timeout after which an attempt to
3139 update a cache entry is deemed to have failed.
3141 nfs.idmap_cache_timeout=
3142 [NFS] set the maximum lifetime for idmapper cache
3146 [NFS] enable 64-bit inode numbers.
3147 If zero, the NFS client will fake up a 32-bit inode
3148 number for the readdir() and stat() syscalls instead
3149 of returning the full 64-bit number.
3150 The default is to return 64-bit inode numbers.
3152 nfs.max_session_cb_slots=
3153 [NFSv4.1] Sets the maximum number of session
3154 slots the client will assign to the callback
3155 channel. This determines the maximum number of
3156 callbacks the client will process in parallel for
3157 a particular server.
3159 nfs.max_session_slots=
3160 [NFSv4.1] Sets the maximum number of session slots
3161 the client will attempt to negotiate with the server.
3162 This limits the number of simultaneous RPC requests
3163 that the client can send to the NFSv4.1 server.
3164 Note that there is little point in setting this
3165 value higher than the max_tcp_slot_table_limit.
3167 nfs.nfs4_disable_idmapping=
3168 [NFSv4] When set to the default of '1', this option
3169 ensures that both the RPC level authentication
3170 scheme and the NFS level operations agree to use
3171 numeric uids/gids if the mount is using the
3172 'sec=sys' security flavour. In effect it is
3173 disabling idmapping, which can make migration from
3174 legacy NFSv2/v3 systems to NFSv4 easier.
3175 Servers that do not support this mode of operation
3176 will be autodetected by the client, and it will fall
3177 back to using the idmapper.
3178 To turn off this behaviour, set the value to '0'.
3180 [NFS4] Specify an additional fixed unique ident-
3181 ification string that NFSv4 clients can insert into
3182 their nfs_client_id4 string. This is typically a
3183 UUID that is generated at system install time.
3185 nfs.send_implementation_id =
3186 [NFSv4.1] Send client implementation identification
3187 information in exchange_id requests.
3188 If zero, no implementation identification information
3190 The default is to send the implementation identification
3193 nfs.recover_lost_locks =
3194 [NFSv4] Attempt to recover locks that were lost due
3195 to a lease timeout on the server. Please note that
3196 doing this risks data corruption, since there are
3197 no guarantees that the file will remain unchanged
3198 after the locks are lost.
3199 If you want to enable the kernel legacy behaviour of
3200 attempting to recover these locks, then set this
3202 The default parameter value of '0' causes the kernel
3203 not to attempt recovery of lost locks.
3205 nfs4.layoutstats_timer =
3206 [NFSv4.2] Change the rate at which the kernel sends
3207 layoutstats to the pNFS metadata server.
3209 Setting this to value to 0 causes the kernel to use
3210 whatever value is the default set by the layout
3211 driver. A non-zero value sets the minimum interval
3212 in seconds between layoutstats transmissions.
3214 nfsd.nfs4_disable_idmapping=
3215 [NFSv4] When set to the default of '1', the NFSv4
3216 server will return only numeric uids and gids to
3217 clients using auth_sys, and will accept numeric uids
3218 and gids from such clients. This is intended to ease
3219 migration from NFSv2/v3.
3221 nmi_backtrace.backtrace_idle [KNL]
3222 Dump stacks even of idle CPUs in response to an
3223 NMI stack-backtrace request.
3225 nmi_debug= [KNL,SH] Specify one or more actions to take
3226 when a NMI is triggered.
3227 Format: [state][,regs][,debounce][,die]
3229 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3230 Format: [panic,][nopanic,][num]
3232 0 - turn hardlockup detector in nmi_watchdog off
3233 1 - turn hardlockup detector in nmi_watchdog on
3234 When panic is specified, panic when an NMI watchdog
3235 timeout occurs (or 'nopanic' to not panic on an NMI
3236 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3237 To disable both hard and soft lockup detectors,
3238 please see 'nowatchdog'.
3239 This is useful when you use a panic=... timeout and
3240 need the box quickly up again.
3242 These settings can be accessed at runtime via
3243 the nmi_watchdog and hardlockup_panic sysctls.
3245 netpoll.carrier_timeout=
3246 [NET] Specifies amount of time (in seconds) that
3247 netpoll should wait for a carrier. By default netpoll
3250 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3251 emulation library even if a 387 maths coprocessor
3254 no5lvl [X86-64] Disable 5-level paging mode. Forces
3255 kernel to use 4-level paging instead.
3257 nofsgsbase [X86] Disables FSGSBASE instructions.
3260 [HW] Never suspend the console
3261 Disable suspending of consoles during suspend and
3262 hibernate operations. Once disabled, debugging
3263 messages can reach various consoles while the rest
3264 of the system is being put to sleep (ie, while
3265 debugging driver suspend/resume hooks). This may
3266 not work reliably with all consoles, but is known
3267 to work with serial and VGA consoles.
3268 To facilitate more flexible debugging, we also add
3269 console_suspend, a printk module parameter to control
3270 it. Users could use console_suspend (usually
3271 /sys/module/printk/parameters/console_suspend) to
3272 turn on/off it dynamically.
3274 novmcoredd [KNL,KDUMP]
3275 Disable device dump. Device dump allows drivers to
3276 append dump data to vmcore so you can collect driver
3277 specified debug info. Drivers can append the data
3278 without any limit and this data is stored in memory,
3279 so this may cause significant memory stress. Disabling
3280 device dump can help save memory but the driver debug
3281 data will be no longer available. This parameter
3282 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3285 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3286 caches in the slab allocator. Saves per-node memory,
3287 but will impact performance.
3291 noaltinstr [S390] Disables alternative instructions patching
3292 (CPU alternatives feature).
3294 noapic [SMP,APIC] Tells the kernel to not make use of any
3295 IOAPICs that may be present in the system.
3297 noautogroup Disable scheduler automatic task group creation.
3299 nobats [PPC] Do not use BATs for mapping kernel lowmem
3300 on "Classic" PPC cores.
3304 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3306 nodelayacct [KNL] Disable per-task delay accounting
3308 nodsp [SH] Disable hardware DSP at boot time.
3310 noefi Disable EFI runtime services support.
3312 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3317 On X86-32 available only on PAE configured kernels.
3318 noexec=on: enable non-executable mappings (default)
3319 noexec=off: disable non-executable mappings
3322 Disable SMAP (Supervisor Mode Access Prevention)
3323 even if it is supported by processor.
3326 Disable SMEP (Supervisor Mode Execution Prevention)
3327 even if it is supported by processor.
3330 This affects only 32-bit executables.
3331 noexec32=on: enable non-executable mappings (default)
3332 read doesn't imply executable mappings
3333 noexec32=off: disable non-executable mappings
3334 read implies executable mappings
3336 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3338 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3339 register save and restore. The kernel will only save
3340 legacy floating-point registers on task switch.
3342 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3344 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3345 Equivalent to smt=1.
3347 [KNL,X86] Disable symmetric multithreading (SMT).
3348 nosmt=force: Force disable SMT, cannot be undone
3349 via the sysfs control file.
3351 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3352 (bounds check bypass). With this option data leaks are
3353 possible in the system.
3355 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3356 the Spectre variant 2 (indirect branch prediction)
3357 vulnerability. System may allow data leaks with this
3360 nospec_store_bypass_disable
3361 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3364 [PPC] Don't flush the L1-D cache after accessing user data.
3366 noxsave [BUGS=X86] Disables x86 extended register state save
3367 and restore using xsave. The kernel will fallback to
3368 enabling legacy floating-point and sse state.
3370 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3371 register states. The kernel will fall back to use
3372 xsave to save the states. By using this parameter,
3373 performance of saving the states is degraded because
3374 xsave doesn't support modified optimization while
3375 xsaveopt supports it on xsaveopt enabled systems.
3377 noxsaves [X86] Disables xsaves and xrstors used in saving and
3378 restoring x86 extended register state in compacted
3379 form of xsave area. The kernel will fall back to use
3380 xsaveopt and xrstor to save and restore the states
3381 in standard form of xsave area. By using this
3382 parameter, xsave area per process might occupy more
3383 memory on xsaves enabled systems.
3385 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
3386 wfi(ARM) instruction doesn't work correctly and not to
3387 use it. This is also useful when using JTAG debugger.
3389 no_file_caps Tells the kernel not to honor file capabilities. The
3390 only way then for a file to be executed with privilege
3391 is to be setuid root or executed by root.
3393 nohalt [IA-64] Tells the kernel not to use the power saving
3394 function PAL_HALT_LIGHT when idle. This increases
3395 power-consumption. On the positive side, it reduces
3396 interrupt wake-up latency, which may improve performance
3397 in certain environments such as networked servers or
3400 nohibernate [HIBERNATION] Disable hibernation and resume.
3402 nohz= [KNL] Boottime enable/disable dynamic ticks
3403 Valid arguments: on, off
3406 nohz_full= [KNL,BOOT,SMP,ISOL]
3407 The argument is a cpu list, as described above.
3408 In kernels built with CONFIG_NO_HZ_FULL=y, set
3409 the specified list of CPUs whose tick will be stopped
3410 whenever possible. The boot CPU will be forced outside
3411 the range to maintain the timekeeping. Any CPUs
3412 in this list will have their RCU callbacks offloaded,
3413 just as if they had also been called out in the
3414 rcu_nocbs= boot parameter.
3416 noiotrap [SH] Disables trapped I/O port accesses.
3418 noirqdebug [X86-32] Disables the code which attempts to detect and
3419 disable unhandled interrupt sources.
3421 no_timer_check [X86,APIC] Disables the code which tests for
3422 broken timer IRQ sources.
3424 noisapnp [ISAPNP] Disables ISA PnP code.
3426 noinitrd [RAM] Tells the kernel not to load any configured
3429 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3431 [Deprecated - use intremap=off]
3435 noinvpcid [X86] Disable the INVPCID cpu feature.
3437 nojitter [IA-64] Disables jitter checking for ITC timers.
3439 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3441 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3445 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3446 clock and use the default one.
3448 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3449 accounting. steal time is computed, but won't
3450 influence scheduler behaviour
3452 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3454 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3456 noltlbs [PPC] Do not use large page/tlb entries for kernel
3457 lowmem mapping on PPC40x and PPC8xx
3459 nomca [IA-64] Disable machine check abort handling
3461 nomce [X86-32] Disable Machine Check Exception
3463 nomfgpt [X86-32] Disable Multi-Function General Purpose
3464 Timer usage (for AMD Geode machines).
3466 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3467 shutdown the other cpus. Instead use the REBOOT_VECTOR
3470 nomodule Disable module load
3472 nopat [X86] Disable PAT (page attribute table extension of
3473 pagetables) support.
3475 nopcid [X86-64] Disable the PCID cpu feature.
3477 norandmaps Don't use address space randomization. Equivalent to
3478 echo 0 > /proc/sys/kernel/randomize_va_space
3480 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3481 with UP alternatives
3483 nordrand [X86] Disable kernel use of the RDRAND and
3484 RDSEED instructions even if they are supported
3485 by the processor. RDRAND and RDSEED are still
3486 available to user space applications.
3488 noresume [SWSUSP] Disables resume and restores original swap
3491 no-scroll [VGA] Disables scrollback.
3492 This is required for the Braillex ib80-piezo Braille
3493 reader made by F.H. Papenmeier (Germany).
3497 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3499 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3500 and disable the IO APIC. legacy for "maxcpus=0".
3502 nosoftlockup [KNL] Disable the soft-lockup detector.
3504 nosync [HW,M68K] Disables sync negotiation for all devices.
3506 nowatchdog [KNL] Disable both lockup detectors, i.e.
3507 soft-lockup and NMI watchdog (hard-lockup).
3511 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3513 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3514 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3515 Some features depend on CPU0. Known dependencies are:
3516 1. Resume from suspend/hibernate depends on CPU0.
3517 Suspend/hibernate will fail if CPU0 is offline and you
3518 need to online CPU0 before suspend/hibernate.
3519 2. PIC interrupts also depend on CPU0. CPU0 can't be
3520 removed if a PIC interrupt is detected.
3521 It's said poweroff/reboot may depend on CPU0 on some
3522 machines although I haven't seen such issues so far
3523 after CPU0 is offline on a few tested machines.
3524 If the dependencies are under your control, you can
3525 turn on cpu0_hotplug.
3527 nps_mtm_hs_ctr= [KNL,ARC]
3528 This parameter sets the maximum duration, in
3529 cycles, each HW thread of the CTOP can run
3530 without interruptions, before HW switches it.
3531 The actual maximum duration is 16 times this
3533 Format: integer between 1 and 255
3536 nptcg= [IA-64] Override max number of concurrent global TLB
3537 purges which is reported from either PAL_VM_SUMMARY or
3540 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3541 could support. nr_cpus=n : n >= 1 limits the kernel to
3542 support 'n' processors. It could be larger than the
3543 number of already plugged CPU during bootup, later in
3544 runtime you can physically add extra cpu until it reaches
3545 n. So during boot up some boot time memory for per-cpu
3546 variables need be pre-allocated for later physical cpu
3549 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3551 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3552 Allowed values are enable and disable
3554 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3555 'node', 'default' can be specified
3556 This can be set from sysctl after boot.
3557 See Documentation/admin-guide/sysctl/vm.rst for details.
3559 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3560 See Documentation/core-api/debugging-via-ohci1394.rst for more
3563 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3564 Rather than timing out after 20 ms if an EC
3565 command is not properly ACKed, override the length
3566 of the timeout. We have interrupts disabled while
3567 waiting for the ACK, so if this is set too high
3568 interrupts *may* be lost!
3570 omap_mux= [OMAP] Override bootloader pin multiplexing.
3571 Format: <mux_mode0.mode_name=value>...
3572 For example, to override I2C bus2:
3573 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3575 oprofile.timer= [HW]
3576 Use timer interrupt instead of performance counters
3578 oprofile.cpu_type= Force an oprofile cpu type
3579 This might be useful if you have an older oprofile
3580 userland or if you want common events.
3581 Format: { arch_perfmon }
3582 arch_perfmon: [X86] Force use of architectural
3583 perfmon on Intel CPUs instead of the
3584 CPU specific event set.
3585 timer: [X86] Force use of architectural NMI
3586 timer mode (see also oprofile.timer
3587 for generic hr timer mode)
3589 oops=panic Always panic on oopses. Default is to just kill the
3590 process, but there is a small probability of
3591 deadlocking the machine.
3592 This will also cause panics on machine check exceptions.
3593 Useful together with panic=30 to trigger a reboot.
3596 [KNL] Boolean flag to control whether the page allocator
3597 should randomize its free lists. The randomization may
3598 be automatically enabled if the kernel detects it is
3599 running on a platform with a direct-mapped memory-side
3600 cache, and this parameter can be used to
3601 override/disable that behavior. The state of the flag
3602 can be read from sysfs at:
3603 /sys/module/page_alloc/parameters/shuffle.
3605 page_owner= [KNL] Boot-time page_owner enabling option.
3606 Storage of the information about who allocated
3607 each page is disabled in default. With this switch,
3609 on: enable the feature
3611 page_poison= [KNL] Boot-time parameter changing the state of
3612 poisoning on the buddy allocator, available with
3613 CONFIG_PAGE_POISONING=y.
3614 off: turn off poisoning (default)
3615 on: turn on poisoning
3617 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3618 timeout > 0: seconds before rebooting
3619 timeout = 0: wait forever
3620 timeout < 0: reboot immediately
3623 panic_print= Bitmask for printing system info when panic happens.
3624 User can chose combination of the following bits:
3625 bit 0: print all tasks info
3626 bit 1: print system memory info
3627 bit 2: print timer info
3628 bit 3: print locks info if CONFIG_LOCKDEP is on
3629 bit 4: print ftrace buffer
3630 bit 5: print all printk messages in buffer
3632 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3633 Format: <hex>[,nousertaint]
3634 Hexadecimal bitmask representing the set of TAINT flags
3635 that will cause the kernel to panic when add_taint() is
3636 called with any of the flags in this set.
3637 The optional switch "nousertaint" can be utilized to
3638 prevent userspace forced crashes by writing to sysctl
3639 /proc/sys/kernel/tainted any flagset matching with the
3640 bitmask set on panic_on_taint.
3641 See Documentation/admin-guide/tainted-kernels.rst for
3642 extra details on the taint flags that users can pick
3643 to compose the bitmask to assign to panic_on_taint.
3645 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3648 crash_kexec_post_notifiers
3649 Run kdump after running panic-notifiers and dumping
3650 kmsg. This only for the users who doubt kdump always
3651 succeeds in any situation.
3652 Note that this also increases risks of kdump failure,
3653 because some panic notifiers can make the crashed
3654 kernel more unstable.
3656 parkbd.port= [HW] Parallel port number the keyboard adapter is
3657 connected to, default is 0.
3659 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3660 0 for XT, 1 for AT (default is AT).
3663 parport= [HW,PPT] Specify parallel ports. 0 disables.
3664 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3665 Use 'auto' to force the driver to use any
3666 IRQ/DMA settings detected (the default is to
3667 ignore detected IRQ/DMA settings because of
3668 possible conflicts). You can specify the base
3669 address, IRQ, and DMA settings; IRQ and DMA
3670 should be numbers, or 'auto' (for using detected
3671 settings on that particular port), or 'nofifo'
3672 (to avoid using a FIFO even if it is detected).
3673 Parallel ports are assigned in the order they
3674 are specified on the command line, starting
3677 parport_init_mode= [HW,PPT]
3678 Configure VIA parallel port to operate in
3679 a specific mode. This is necessary on Pegasos
3680 computer where firmware has no options for setting
3681 up parallel port mode and sets it to spp.
3682 Currently this function knows 686a and 8231 chips.
3683 Format: [spp|ps2|epp|ecp|ecpepp]
3686 Halt all CPUs after the first oops has been printed for
3687 the specified number of seconds. This is to be used if
3688 your oopses keep scrolling off the screen.
3693 See header of drivers/block/paride/pcd.c.
3694 See also Documentation/admin-guide/blockdev/paride.rst.
3696 pci=option[,option...] [PCI] various PCI subsystem options.
3698 Some options herein operate on a specific device
3699 or a set of devices (<pci_dev>). These are
3700 specified in one of the following formats:
3702 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3703 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3705 Note: the first format specifies a PCI
3706 bus/device/function address which may change
3707 if new hardware is inserted, if motherboard
3708 firmware changes, or due to changes caused
3709 by other kernel parameters. If the
3710 domain is left unspecified, it is
3711 taken to be zero. Optionally, a path
3712 to a device through multiple device/function
3713 addresses can be specified after the base
3714 address (this is more robust against
3715 renumbering issues). The second format
3716 selects devices using IDs from the
3717 configuration space which may match multiple
3718 devices in the system.
3720 earlydump dump PCI config space before the kernel
3722 off [X86] don't probe for the PCI bus
3723 bios [X86-32] force use of PCI BIOS, don't access
3724 the hardware directly. Use this if your machine
3725 has a non-standard PCI host bridge.
3726 nobios [X86-32] disallow use of PCI BIOS, only direct
3727 hardware access methods are allowed. Use this
3728 if you experience crashes upon bootup and you
3729 suspect they are caused by the BIOS.
3730 conf1 [X86] Force use of PCI Configuration Access
3731 Mechanism 1 (config address in IO port 0xCF8,
3732 data in IO port 0xCFC, both 32-bit).
3733 conf2 [X86] Force use of PCI Configuration Access
3734 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3735 the function, IO port 0xCFA, also 8-bit, sets
3736 bus number. The config space is then accessed
3737 through ports 0xC000-0xCFFF).
3738 See http://wiki.osdev.org/PCI for more info
3739 on the configuration access mechanisms.
3740 noaer [PCIE] If the PCIEAER kernel config parameter is
3741 enabled, this kernel boot option can be used to
3742 disable the use of PCIE advanced error reporting.
3743 nodomains [PCI] Disable support for multiple PCI
3744 root domains (aka PCI segments, in ACPI-speak).
3745 nommconf [X86] Disable use of MMCONFIG for PCI
3747 check_enable_amd_mmconf [X86] check for and enable
3748 properly configured MMIO access to PCI
3749 config space on AMD family 10h CPU
3750 nomsi [MSI] If the PCI_MSI kernel config parameter is
3751 enabled, this kernel boot option can be used to
3752 disable the use of MSI interrupts system-wide.
3753 noioapicquirk [APIC] Disable all boot interrupt quirks.
3754 Safety option to keep boot IRQs enabled. This
3755 should never be necessary.
3756 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3757 primary IO-APIC for bridges that cannot disable
3758 boot IRQs. This fixes a source of spurious IRQs
3759 when the system masks IRQs.
3760 noioapicreroute [APIC] Disable workaround that uses the
3761 boot IRQ equivalent of an IRQ that connects to
3762 a chipset where boot IRQs cannot be disabled.
3763 The opposite of ioapicreroute.
3764 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3765 routing table. These calls are known to be buggy
3766 on several machines and they hang the machine
3767 when used, but on other computers it's the only
3768 way to get the interrupt routing table. Try
3769 this option if the kernel is unable to allocate
3770 IRQs or discover secondary PCI buses on your
3772 rom [X86] Assign address space to expansion ROMs.
3773 Use with caution as certain devices share
3774 address decoders between ROMs and other
3776 norom [X86] Do not assign address space to
3777 expansion ROMs that do not already have
3778 BIOS assigned address ranges.
3779 nobar [X86] Do not assign address space to the
3780 BARs that weren't assigned by the BIOS.
3781 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3782 assigned automatically to PCI devices. You can
3783 make the kernel exclude IRQs of your ISA cards
3785 pirqaddr=0xAAAAA [X86] Specify the physical address
3786 of the PIRQ table (normally generated
3787 by the BIOS) if it is outside the
3788 F0000h-100000h range.
3789 lastbus=N [X86] Scan all buses thru bus #N. Can be
3790 useful if the kernel is unable to find your
3791 secondary buses and you want to tell it
3792 explicitly which ones they are.
3793 assign-busses [X86] Always assign all PCI bus
3794 numbers ourselves, overriding
3795 whatever the firmware may have done.
3796 usepirqmask [X86] Honor the possible IRQ mask stored
3797 in the BIOS $PIR table. This is needed on
3798 some systems with broken BIOSes, notably
3799 some HP Pavilion N5400 and Omnibook XE3
3800 notebooks. This will have no effect if ACPI
3801 IRQ routing is enabled.
3802 noacpi [X86] Do not use ACPI for IRQ routing
3803 or for PCI scanning.
3804 use_crs [X86] Use PCI host bridge window information
3805 from ACPI. On BIOSes from 2008 or later, this
3806 is enabled by default. If you need to use this,
3807 please report a bug.
3808 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3809 If you need to use this, please report a bug.
3810 routeirq Do IRQ routing for all PCI devices.
3811 This is normally done in pci_enable_device(),
3812 so this option is a temporary workaround
3813 for broken drivers that don't call it.
3814 skip_isa_align [X86] do not align io start addr, so can
3815 handle more pci cards
3816 noearly [X86] Don't do any early type 1 scanning.
3817 This might help on some broken boards which
3818 machine check when some devices' config space
3819 is read. But various workarounds are disabled
3820 and some IOMMU drivers will not work.
3821 bfsort Sort PCI devices into breadth-first order.
3822 This sorting is done to get a device
3823 order compatible with older (<= 2.4) kernels.
3824 nobfsort Don't sort PCI devices into breadth-first order.
3825 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3826 tuning and use the BIOS-configured MPS defaults.
3827 pcie_bus_safe Set every device's MPS to the largest value
3828 supported by all devices below the root complex.
3829 pcie_bus_perf Set device MPS to the largest allowable MPS
3830 based on its parent bus. Also set MRRS (Max
3831 Read Request Size) to the largest supported
3832 value (no larger than the MPS that the device
3833 or bus can support) for best performance.
3834 pcie_bus_peer2peer Set every device's MPS to 128B, which
3835 every device is guaranteed to support. This
3836 configuration allows peer-to-peer DMA between
3837 any pair of devices, possibly at the cost of
3838 reduced performance. This also guarantees
3839 that hot-added devices will work.
3840 cbiosize=nn[KMG] The fixed amount of bus space which is
3841 reserved for the CardBus bridge's IO window.
3842 The default value is 256 bytes.
3843 cbmemsize=nn[KMG] The fixed amount of bus space which is
3844 reserved for the CardBus bridge's memory
3845 window. The default value is 64 megabytes.
3848 [<order of align>@]<pci_dev>[; ...]
3849 Specifies alignment and device to reassign
3850 aligned memory resources. How to
3851 specify the device is described above.
3852 If <order of align> is not specified,
3853 PAGE_SIZE is used as alignment.
3854 A PCI-PCI bridge can be specified if resource
3855 windows need to be expanded.
3856 To specify the alignment for several
3857 instances of a device, the PCI vendor,
3858 device, subvendor, and subdevice may be
3859 specified, e.g., 12@pci:8086:9c22:103c:198f
3860 for 4096-byte alignment.
3861 ecrc= Enable/disable PCIe ECRC (transaction layer
3862 end-to-end CRC checking).
3863 bios: Use BIOS/firmware settings. This is the
3867 hpiosize=nn[KMG] The fixed amount of bus space which is
3868 reserved for hotplug bridge's IO window.
3869 Default size is 256 bytes.
3870 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3871 reserved for hotplug bridge's MMIO window.
3872 Default size is 2 megabytes.
3873 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3874 reserved for hotplug bridge's MMIO_PREF window.
3875 Default size is 2 megabytes.
3876 hpmemsize=nn[KMG] The fixed amount of bus space which is
3877 reserved for hotplug bridge's MMIO and
3879 Default size is 2 megabytes.
3880 hpbussize=nn The minimum amount of additional bus numbers
3881 reserved for buses below a hotplug bridge.
3883 realloc= Enable/disable reallocating PCI bridge resources
3884 if allocations done by BIOS are too small to
3885 accommodate resources required by all child
3887 off: Turn realloc off
3889 realloc same as realloc=on
3890 noari do not use PCIe ARI.
3891 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3892 do not use PCIe ATS (and IOMMU device IOTLB).
3893 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3894 only look for one device below a PCIe downstream
3896 big_root_window Try to add a big 64bit memory window to the PCIe
3897 root complex on AMD CPUs. Some GFX hardware
3898 can resize a BAR to allow access to all VRAM.
3899 Adding the window is slightly risky (it may
3900 conflict with unreported devices), so this
3902 disable_acs_redir=<pci_dev>[; ...]
3903 Specify one or more PCI devices (in the format
3904 specified above) separated by semicolons.
3905 Each device specified will have the PCI ACS
3906 redirect capabilities forced off which will
3907 allow P2P traffic between devices through
3908 bridges without forcing it upstream. Note:
3909 this removes isolation between devices and
3910 may put more devices in an IOMMU group.
3911 force_floating [S390] Force usage of floating interrupts.
3912 nomio [S390] Do not use MIO instructions.
3913 norid [S390] ignore the RID field and force use of
3914 one PCI domain per PCI function
3916 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3919 force Enable ASPM even on devices that claim not to support it.
3920 WARNING: Forcing ASPM on may cause system lockups.
3922 pcie_ports= [PCIE] PCIe port services handling:
3923 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3924 even if the platform doesn't give the OS permission to
3925 use them. This may cause conflicts if the platform
3926 also tries to use these services.
3927 dpc-native Use native PCIe service for DPC only. May
3928 cause conflicts if firmware uses AER or DPC.
3929 compat Disable native PCIe services (PME, AER, DPC, PCIe
3932 pcie_port_pm= [PCIE] PCIe port power management handling:
3933 off Disable power management of all PCIe ports
3934 force Forcibly enable power management of all PCIe ports
3936 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3937 nomsi Do not use MSI for native PCIe PME signaling (this makes
3938 all PCIe root ports use INTx for all services).
3940 pcmv= [HW,PCMCIA] BadgePAD 4
3944 Keep all power-domains already enabled by bootloader on,
3945 even if no driver has claimed them. This is useful
3946 for debug and development, but should not be
3947 needed on a platform with proper driver support.
3950 See Documentation/admin-guide/blockdev/paride.rst.
3952 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3955 See arch/parisc/kernel/pdc_chassis.c
3957 percpu_alloc= Select which percpu first chunk allocator to use.
3958 Currently supported values are "embed" and "page".
3959 Archs may support subset or none of the selections.
3960 See comments in mm/percpu.c for details on each
3961 allocator. This parameter is primarily for debugging
3962 and performance comparison.
3965 See Documentation/admin-guide/blockdev/paride.rst.
3968 See Documentation/admin-guide/blockdev/paride.rst.
3970 pirq= [SMP,APIC] Manual mp-table setup
3971 See Documentation/x86/i386/IO-APIC.rst.
3973 plip= [PPT,NET] Parallel port network link
3974 Format: { parport<nr> | timid | 0 }
3975 See also Documentation/admin-guide/parport.rst.
3977 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3978 Override pmtimer IOPort with a hex value.
3981 pm_debug_messages [SUSPEND,KNL]
3982 Enable suspend/resume debug messages during boot up.
3985 Enable PNP debug messages (depends on the
3986 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3987 via /sys/module/pnp/parameters/debug. We always show
3988 current resource usage; turning this on also shows
3989 possible settings and some assignment information.
3995 { on | off | curr | res | no-curr | no-res }
3998 [ISAPNP] Exclude IRQs for the autoconfiguration
4001 [ISAPNP] Exclude DMAs for the autoconfiguration
4003 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4004 Ranges are in pairs (I/O port base and size).
4007 [ISAPNP] Exclude memory regions for the
4009 Ranges are in pairs (memory base and size).
4011 ports= [IP_VS_FTP] IPVS ftp helper module
4013 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4015 Format: <port>,<port>....
4017 powersave=off [PPC] This option disables power saving features.
4018 It specifically disables cpuidle and sets the
4019 platform machine description specific power_save
4020 function to NULL. On Idle the CPU just reduces
4023 ppc_strict_facility_enable
4024 [PPC] This option catches any kernel floating point,
4025 Altivec, VSX and SPE outside of regions specifically
4026 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4027 There is some performance impact when enabling this.
4031 Disable Hardware Transactional Memory
4033 print-fatal-signals=
4034 [KNL] debug: print fatal signals
4036 If enabled, warn about various signal handling
4037 related application anomalies: too many signals,
4038 too many POSIX.1 timers, fatal signals causing a
4041 If you hit the warning due to signal overflow,
4042 you might want to try "ulimit -i unlimited".
4046 printk.always_kmsg_dump=
4047 Trigger kmsg_dump for cases other than kernel oops or
4049 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4052 printk.devkmsg={on,off,ratelimit}
4053 Control writing to /dev/kmsg.
4054 on - unlimited logging to /dev/kmsg from userspace
4055 off - logging to /dev/kmsg disabled
4056 ratelimit - ratelimit the logging
4059 printk.time= Show timing data prefixed to each printk message line
4060 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4062 processor.max_cstate= [HW,ACPI]
4063 Limit processor to maximum C-state
4064 max_cstate=9 overrides any DMI blacklist limit.
4066 processor.nocst [HW,ACPI]
4067 Ignore the _CST method to determine C-states,
4068 instead using the legacy FADT method
4070 profile= [KNL] Enable kernel profiling via /proc/profile
4071 Format: [<profiletype>,]<number>
4072 Param: <profiletype>: "schedule", "sleep", or "kvm"
4073 [defaults to kernel profiling]
4074 Param: "schedule" - profile schedule points.
4075 Param: "sleep" - profile D-state sleeping (millisecs).
4076 Requires CONFIG_SCHEDSTATS
4077 Param: "kvm" - profile VM exits.
4078 Param: <number> - step/bucket size as a power of 2 for
4079 statistical time based profiling.
4081 prompt_ramdisk= [RAM] [Deprecated]
4083 prot_virt= [S390] enable hosting protected virtual machines
4084 isolated from the hypervisor (if hardware supports
4088 psi= [KNL] Enable or disable pressure stall information
4092 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4093 probe for; one of (bare|imps|exps|lifebook|any).
4094 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4096 psmouse.resetafter= [HW,MOUSE]
4097 Try to reset the device after so many bad packets
4100 [HW,MOUSE] Set desired mouse resolution, in dpi.
4101 psmouse.smartscroll=
4102 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4103 0 = disabled, 1 = enabled (default).
4105 pstore.backend= Specify the name of the pstore backend to use
4108 See Documentation/admin-guide/blockdev/paride.rst.
4110 pti= [X86-64] Control Page Table Isolation of user and
4111 kernel address spaces. Disabling this feature
4112 removes hardening, but improves performance of
4113 system calls and interrupts.
4115 on - unconditionally enable
4116 off - unconditionally disable
4117 auto - kernel detects whether your CPU model is
4118 vulnerable to issues that PTI mitigates
4120 Not specifying this option is equivalent to pti=auto.
4123 Equivalent to pti=off
4126 [KNL] Number of legacy pty's. Overwrites compiled-in
4129 quiet [KNL] Disable most log messages
4134 See Documentation/admin-guide/md.rst.
4136 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4137 See Documentation/admin-guide/blockdev/ramdisk.rst.
4139 ramdisk_start= [RAM] RAM disk image start address
4141 random.trust_cpu={on,off}
4142 [KNL] Enable or disable trusting the use of the
4143 CPU's random number generator (if available) to
4144 fully seed the kernel's CRNG. Default is controlled
4145 by CONFIG_RANDOM_TRUST_CPU.
4147 random.trust_bootloader={on,off}
4148 [KNL] Enable or disable trusting the use of a
4149 seed passed by the bootloader (if available) to
4150 fully seed the kernel's CRNG. Default is controlled
4151 by CONFIG_RANDOM_TRUST_BOOTLOADER.
4153 ras=option[,option,...] [KNL] RAS-specific options
4156 Disable the Correctable Errors Collector,
4157 see CONFIG_RAS_CEC help text.
4160 The argument is a cpu list, as described above,
4161 except that the string "all" can be used to
4162 specify every CPU on the system.
4164 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
4165 the specified list of CPUs to be no-callback CPUs.
4166 Invocation of these CPUs' RCU callbacks will be
4167 offloaded to "rcuox/N" kthreads created for that
4168 purpose, where "x" is "p" for RCU-preempt, and
4169 "s" for RCU-sched, and "N" is the CPU number.
4170 This reduces OS jitter on the offloaded CPUs,
4171 which can be useful for HPC and real-time
4172 workloads. It can also improve energy efficiency
4173 for asymmetric multiprocessors.
4176 Rather than requiring that offloaded CPUs
4177 (specified by rcu_nocbs= above) explicitly
4178 awaken the corresponding "rcuoN" kthreads,
4179 make these kthreads poll for callbacks.
4180 This improves the real-time response for the
4181 offloaded CPUs by relieving them of the need to
4182 wake up the corresponding kthread, but degrades
4183 energy efficiency by requiring that the kthreads
4184 periodically wake up to do the polling.
4186 rcutree.blimit= [KNL]
4187 Set maximum number of finished RCU callbacks to
4188 process in one batch.
4190 rcutree.dump_tree= [KNL]
4191 Dump the structure of the rcu_node combining tree
4192 out at early boot. This is used for diagnostic
4193 purposes, to verify correct tree setup.
4195 rcutree.gp_cleanup_delay= [KNL]
4196 Set the number of jiffies to delay each step of
4197 RCU grace-period cleanup.
4199 rcutree.gp_init_delay= [KNL]
4200 Set the number of jiffies to delay each step of
4201 RCU grace-period initialization.
4203 rcutree.gp_preinit_delay= [KNL]
4204 Set the number of jiffies to delay each step of
4205 RCU grace-period pre-initialization, that is,
4206 the propagation of recent CPU-hotplug changes up
4207 the rcu_node combining tree.
4209 rcutree.use_softirq= [KNL]
4210 If set to zero, move all RCU_SOFTIRQ processing to
4211 per-CPU rcuc kthreads. Defaults to a non-zero
4212 value, meaning that RCU_SOFTIRQ is used by default.
4213 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4215 rcutree.rcu_fanout_exact= [KNL]
4216 Disable autobalancing of the rcu_node combining
4217 tree. This is used by rcutorture, and might
4218 possibly be useful for architectures having high
4219 cache-to-cache transfer latencies.
4221 rcutree.rcu_fanout_leaf= [KNL]
4222 Change the number of CPUs assigned to each
4223 leaf rcu_node structure. Useful for very
4224 large systems, which will choose the value 64,
4225 and for NUMA systems with large remote-access
4226 latencies, which will choose a value aligned
4227 with the appropriate hardware boundaries.
4229 rcutree.rcu_min_cached_objs= [KNL]
4230 Minimum number of objects which are cached and
4231 maintained per one CPU. Object size is equal
4232 to PAGE_SIZE. The cache allows to reduce the
4233 pressure to page allocator, also it makes the
4234 whole algorithm to behave better in low memory
4237 rcutree.jiffies_till_first_fqs= [KNL]
4238 Set delay from grace-period initialization to
4239 first attempt to force quiescent states.
4240 Units are jiffies, minimum value is zero,
4241 and maximum value is HZ.
4243 rcutree.jiffies_till_next_fqs= [KNL]
4244 Set delay between subsequent attempts to force
4245 quiescent states. Units are jiffies, minimum
4246 value is one, and maximum value is HZ.
4248 rcutree.jiffies_till_sched_qs= [KNL]
4249 Set required age in jiffies for a
4250 given grace period before RCU starts
4251 soliciting quiescent-state help from
4252 rcu_note_context_switch() and cond_resched().
4253 If not specified, the kernel will calculate
4254 a value based on the most recent settings
4255 of rcutree.jiffies_till_first_fqs
4256 and rcutree.jiffies_till_next_fqs.
4257 This calculated value may be viewed in
4258 rcutree.jiffies_to_sched_qs. Any attempt to set
4259 rcutree.jiffies_to_sched_qs will be cheerfully
4262 rcutree.kthread_prio= [KNL,BOOT]
4263 Set the SCHED_FIFO priority of the RCU per-CPU
4264 kthreads (rcuc/N). This value is also used for
4265 the priority of the RCU boost threads (rcub/N)
4266 and for the RCU grace-period kthreads (rcu_bh,
4267 rcu_preempt, and rcu_sched). If RCU_BOOST is
4268 set, valid values are 1-99 and the default is 1
4269 (the least-favored priority). Otherwise, when
4270 RCU_BOOST is not set, valid values are 0-99 and
4271 the default is zero (non-realtime operation).
4273 rcutree.rcu_nocb_gp_stride= [KNL]
4274 Set the number of NOCB callback kthreads in
4275 each group, which defaults to the square root
4276 of the number of CPUs. Larger numbers reduce
4277 the wakeup overhead on the global grace-period
4278 kthread, but increases that same overhead on
4279 each group's NOCB grace-period kthread.
4281 rcutree.qhimark= [KNL]
4282 Set threshold of queued RCU callbacks beyond which
4283 batch limiting is disabled.
4285 rcutree.qlowmark= [KNL]
4286 Set threshold of queued RCU callbacks below which
4287 batch limiting is re-enabled.
4289 rcutree.qovld= [KNL]
4290 Set threshold of queued RCU callbacks beyond which
4291 RCU's force-quiescent-state scan will aggressively
4292 enlist help from cond_resched() and sched IPIs to
4293 help CPUs more quickly reach quiescent states.
4294 Set to less than zero to make this be set based
4295 on rcutree.qhimark at boot time and to zero to
4296 disable more aggressive help enlistment.
4298 rcutree.rcu_idle_gp_delay= [KNL]
4299 Set wakeup interval for idle CPUs that have
4300 RCU callbacks (RCU_FAST_NO_HZ=y).
4302 rcutree.rcu_idle_lazy_gp_delay= [KNL]
4303 Set wakeup interval for idle CPUs that have
4304 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
4305 Lazy RCU callbacks are those which RCU can
4306 prove do nothing more than free memory.
4308 rcutree.rcu_kick_kthreads= [KNL]
4309 Cause the grace-period kthread to get an extra
4310 wake_up() if it sleeps three times longer than
4311 it should at force-quiescent-state time.
4312 This wake_up() will be accompanied by a
4313 WARN_ONCE() splat and an ftrace_dump().
4315 rcutree.rcu_unlock_delay= [KNL]
4316 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4317 this specifies an rcu_read_unlock()-time delay
4318 in microseconds. This defaults to zero.
4319 Larger delays increase the probability of
4320 catching RCU pointer leaks, that is, buggy use
4321 of RCU-protected pointers after the relevant
4322 rcu_read_unlock() has completed.
4324 rcutree.sysrq_rcu= [KNL]
4325 Commandeer a sysrq key to dump out Tree RCU's
4326 rcu_node tree with an eye towards determining
4327 why a new grace period has not yet started.
4329 rcuscale.gp_async= [KNL]
4330 Measure performance of asynchronous
4331 grace-period primitives such as call_rcu().
4333 rcuscale.gp_async_max= [KNL]
4334 Specify the maximum number of outstanding
4335 callbacks per writer thread. When a writer
4336 thread exceeds this limit, it invokes the
4337 corresponding flavor of rcu_barrier() to allow
4338 previously posted callbacks to drain.
4340 rcuscale.gp_exp= [KNL]
4341 Measure performance of expedited synchronous
4342 grace-period primitives.
4344 rcuscale.holdoff= [KNL]
4345 Set test-start holdoff period. The purpose of
4346 this parameter is to delay the start of the
4347 test until boot completes in order to avoid
4350 rcuscale.kfree_rcu_test= [KNL]
4351 Set to measure performance of kfree_rcu() flooding.
4353 rcuscale.kfree_nthreads= [KNL]
4354 The number of threads running loops of kfree_rcu().
4356 rcuscale.kfree_alloc_num= [KNL]
4357 Number of allocations and frees done in an iteration.
4359 rcuscale.kfree_loops= [KNL]
4360 Number of loops doing rcuscale.kfree_alloc_num number
4361 of allocations and frees.
4363 rcuscale.nreaders= [KNL]
4364 Set number of RCU readers. The value -1 selects
4365 N, where N is the number of CPUs. A value
4366 "n" less than -1 selects N-n+1, where N is again
4367 the number of CPUs. For example, -2 selects N
4368 (the number of CPUs), -3 selects N+1, and so on.
4369 A value of "n" less than or equal to -N selects
4372 rcuscale.nwriters= [KNL]
4373 Set number of RCU writers. The values operate
4374 the same as for rcuscale.nreaders.
4375 N, where N is the number of CPUs
4377 rcuscale.perf_type= [KNL]
4378 Specify the RCU implementation to test.
4380 rcuscale.shutdown= [KNL]
4381 Shut the system down after performance tests
4382 complete. This is useful for hands-off automated
4385 rcuscale.verbose= [KNL]
4386 Enable additional printk() statements.
4388 rcuscale.writer_holdoff= [KNL]
4389 Write-side holdoff between grace periods,
4390 in microseconds. The default of zero says
4393 rcutorture.fqs_duration= [KNL]
4394 Set duration of force_quiescent_state bursts
4397 rcutorture.fqs_holdoff= [KNL]
4398 Set holdoff time within force_quiescent_state bursts
4401 rcutorture.fqs_stutter= [KNL]
4402 Set wait time between force_quiescent_state bursts
4405 rcutorture.fwd_progress= [KNL]
4406 Enable RCU grace-period forward-progress testing
4407 for the types of RCU supporting this notion.
4409 rcutorture.fwd_progress_div= [KNL]
4410 Specify the fraction of a CPU-stall-warning
4411 period to do tight-loop forward-progress testing.
4413 rcutorture.fwd_progress_holdoff= [KNL]
4414 Number of seconds to wait between successive
4415 forward-progress tests.
4417 rcutorture.fwd_progress_need_resched= [KNL]
4418 Enclose cond_resched() calls within checks for
4419 need_resched() during tight-loop forward-progress
4422 rcutorture.gp_cond= [KNL]
4423 Use conditional/asynchronous update-side
4424 primitives, if available.
4426 rcutorture.gp_exp= [KNL]
4427 Use expedited update-side primitives, if available.
4429 rcutorture.gp_normal= [KNL]
4430 Use normal (non-expedited) asynchronous
4431 update-side primitives, if available.
4433 rcutorture.gp_sync= [KNL]
4434 Use normal (non-expedited) synchronous
4435 update-side primitives, if available. If all
4436 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4437 rcutorture.gp_normal=, and rcutorture.gp_sync=
4438 are zero, rcutorture acts as if is interpreted
4439 they are all non-zero.
4441 rcutorture.irqreader= [KNL]
4442 Run RCU readers from irq handlers, or, more
4443 accurately, from a timer handler. Not all RCU
4444 flavors take kindly to this sort of thing.
4446 rcutorture.leakpointer= [KNL]
4447 Leak an RCU-protected pointer out of the reader.
4448 This can of course result in splats, and is
4449 intended to test the ability of things like
4450 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4453 rcutorture.n_barrier_cbs= [KNL]
4454 Set callbacks/threads for rcu_barrier() testing.
4456 rcutorture.nfakewriters= [KNL]
4457 Set number of concurrent RCU writers. These just
4458 stress RCU, they don't participate in the actual
4459 test, hence the "fake".
4461 rcutorture.nreaders= [KNL]
4462 Set number of RCU readers. The value -1 selects
4463 N-1, where N is the number of CPUs. A value
4464 "n" less than -1 selects N-n-2, where N is again
4465 the number of CPUs. For example, -2 selects N
4466 (the number of CPUs), -3 selects N+1, and so on.
4468 rcutorture.object_debug= [KNL]
4469 Enable debug-object double-call_rcu() testing.
4471 rcutorture.onoff_holdoff= [KNL]
4472 Set time (s) after boot for CPU-hotplug testing.
4474 rcutorture.onoff_interval= [KNL]
4475 Set time (jiffies) between CPU-hotplug operations,
4476 or zero to disable CPU-hotplug testing.
4478 rcutorture.read_exit= [KNL]
4479 Set the number of read-then-exit kthreads used
4480 to test the interaction of RCU updaters and
4481 task-exit processing.
4483 rcutorture.read_exit_burst= [KNL]
4484 The number of times in a given read-then-exit
4485 episode that a set of read-then-exit kthreads
4488 rcutorture.read_exit_delay= [KNL]
4489 The delay, in seconds, between successive
4490 read-then-exit testing episodes.
4492 rcutorture.shuffle_interval= [KNL]
4493 Set task-shuffle interval (s). Shuffling tasks
4494 allows some CPUs to go into dyntick-idle mode
4495 during the rcutorture test.
4497 rcutorture.shutdown_secs= [KNL]
4498 Set time (s) after boot system shutdown. This
4499 is useful for hands-off automated testing.
4501 rcutorture.stall_cpu= [KNL]
4502 Duration of CPU stall (s) to test RCU CPU stall
4503 warnings, zero to disable.
4505 rcutorture.stall_cpu_block= [KNL]
4506 Sleep while stalling if set. This will result
4507 in warnings from preemptible RCU in addition
4508 to any other stall-related activity.
4510 rcutorture.stall_cpu_holdoff= [KNL]
4511 Time to wait (s) after boot before inducing stall.
4513 rcutorture.stall_cpu_irqsoff= [KNL]
4514 Disable interrupts while stalling if set.
4516 rcutorture.stall_gp_kthread= [KNL]
4517 Duration (s) of forced sleep within RCU
4518 grace-period kthread to test RCU CPU stall
4519 warnings, zero to disable. If both stall_cpu
4520 and stall_gp_kthread are specified, the
4521 kthread is starved first, then the CPU.
4523 rcutorture.stat_interval= [KNL]
4524 Time (s) between statistics printk()s.
4526 rcutorture.stutter= [KNL]
4527 Time (s) to stutter testing, for example, specifying
4528 five seconds causes the test to run for five seconds,
4529 wait for five seconds, and so on. This tests RCU's
4530 ability to transition abruptly to and from idle.
4532 rcutorture.test_boost= [KNL]
4533 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4534 "Maybe" means test if the RCU implementation
4535 under test support RCU priority boosting.
4537 rcutorture.test_boost_duration= [KNL]
4538 Duration (s) of each individual boost test.
4540 rcutorture.test_boost_interval= [KNL]
4541 Interval (s) between each boost test.
4543 rcutorture.test_no_idle_hz= [KNL]
4544 Test RCU's dyntick-idle handling. See also the
4545 rcutorture.shuffle_interval parameter.
4547 rcutorture.torture_type= [KNL]
4548 Specify the RCU implementation to test.
4550 rcutorture.verbose= [KNL]
4551 Enable additional printk() statements.
4553 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4554 Dump ftrace buffer after reporting RCU CPU
4557 rcupdate.rcu_cpu_stall_suppress= [KNL]
4558 Suppress RCU CPU stall warning messages.
4560 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4561 Suppress RCU CPU stall warning messages and
4562 rcutorture writer stall warnings that occur
4563 during early boot, that is, during the time
4564 before the init task is spawned.
4566 rcupdate.rcu_cpu_stall_timeout= [KNL]
4567 Set timeout for RCU CPU stall warning messages.
4569 rcupdate.rcu_expedited= [KNL]
4570 Use expedited grace-period primitives, for
4571 example, synchronize_rcu_expedited() instead
4572 of synchronize_rcu(). This reduces latency,
4573 but can increase CPU utilization, degrade
4574 real-time latency, and degrade energy efficiency.
4575 No effect on CONFIG_TINY_RCU kernels.
4577 rcupdate.rcu_normal= [KNL]
4578 Use only normal grace-period primitives,
4579 for example, synchronize_rcu() instead of
4580 synchronize_rcu_expedited(). This improves
4581 real-time latency, CPU utilization, and
4582 energy efficiency, but can expose users to
4583 increased grace-period latency. This parameter
4584 overrides rcupdate.rcu_expedited. No effect on
4585 CONFIG_TINY_RCU kernels.
4587 rcupdate.rcu_normal_after_boot= [KNL]
4588 Once boot has completed (that is, after
4589 rcu_end_inkernel_boot() has been invoked), use
4590 only normal grace-period primitives. No effect
4591 on CONFIG_TINY_RCU kernels.
4593 rcupdate.rcu_task_ipi_delay= [KNL]
4594 Set time in jiffies during which RCU tasks will
4595 avoid sending IPIs, starting with the beginning
4596 of a given grace period. Setting a large
4597 number avoids disturbing real-time workloads,
4598 but lengthens grace periods.
4600 rcupdate.rcu_task_stall_timeout= [KNL]
4601 Set timeout in jiffies for RCU task stall warning
4602 messages. Disable with a value less than or equal
4605 rcupdate.rcu_self_test= [KNL]
4606 Run the RCU early boot self tests
4610 Run specified binary instead of /init from the ramdisk,
4611 used for early userspace startup. See initrd.
4614 force - Override the decision by the kernel to hide the
4615 advertisement of RDRAND support (this affects
4616 certain AMD processors because of buggy BIOS
4617 support, specifically around the suspend/resume
4621 Turn on/off individual RDT features. List is:
4622 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4624 E.g. to turn on cmt and turn off mba use:
4628 Format (x86 or x86_64):
4629 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4631 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4633 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4634 (prefix with 'panic_' to set mode for panic
4636 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4637 reboot_force is either force or not specified,
4638 reboot_cpu is s[mp]#### with #### being the processor
4639 to be used for rebooting.
4641 refscale.holdoff= [KNL]
4642 Set test-start holdoff period. The purpose of
4643 this parameter is to delay the start of the
4644 test until boot completes in order to avoid
4647 refscale.loops= [KNL]
4648 Set the number of loops over the synchronization
4649 primitive under test. Increasing this number
4650 reduces noise due to loop start/end overhead,
4651 but the default has already reduced the per-pass
4652 noise to a handful of picoseconds on ca. 2020
4655 refscale.nreaders= [KNL]
4656 Set number of readers. The default value of -1
4657 selects N, where N is roughly 75% of the number
4658 of CPUs. A value of zero is an interesting choice.
4660 refscale.nruns= [KNL]
4661 Set number of runs, each of which is dumped onto
4664 refscale.readdelay= [KNL]
4665 Set the read-side critical-section duration,
4666 measured in microseconds.
4668 refscale.scale_type= [KNL]
4669 Specify the read-protection implementation to test.
4671 refscale.shutdown= [KNL]
4672 Shut down the system at the end of the performance
4673 test. This defaults to 1 (shut it down) when
4674 refscale is built into the kernel and to 0 (leave
4675 it running) when refscale is built as a module.
4677 refscale.verbose= [KNL]
4678 Enable additional printk() statements.
4681 [KNL, SMP] Set scheduler's default relax_domain_level.
4682 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4684 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4685 Format: <base1>,<size1>[,<base2>,<size2>,...]
4686 Reserve I/O ports or memory so the kernel won't use
4687 them. If <base> is less than 0x10000, the region
4688 is assumed to be I/O ports; otherwise it is memory.
4690 reservetop= [X86-32]
4692 Reserves a hole at the top of the kernel virtual
4697 Set the amount of memory to reserve for BIOS at
4698 the bottom of the address space.
4700 reset_devices [KNL] Force drivers to reset the underlying device
4701 during initialization.
4704 Specify the partition device for software suspend
4706 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4708 resume_offset= [SWSUSP]
4709 Specify the offset from the beginning of the partition
4710 given by "resume=" at which the swap header is located,
4711 in <PAGE_SIZE> units (needed only for swap files).
4712 See Documentation/power/swsusp-and-swap-files.rst
4714 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4715 read the resume files
4717 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4718 Useful for devices that are detected asynchronously
4719 (e.g. USB and MMC devices).
4721 hibernate= [HIBERNATION]
4722 noresume Don't check if there's a hibernation image
4723 present during boot.
4724 nocompress Don't compress/decompress hibernation images.
4725 no Disable hibernation and resume.
4726 protect_image Turn on image protection during restoration
4727 (that will set all pages holding image data
4728 during restoration read-only).
4730 retain_initrd [RAM] Keep initrd memory after extraction
4732 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
4733 Speculative Code Execution with Return Instructions)
4736 AMD-based UNRET and IBPB mitigations alone do not stop
4737 sibling threads from influencing the predictions of other
4738 sibling threads. For that reason, STIBP is used on pro-
4739 cessors that support it, and mitigate SMT on processors
4743 auto - automatically select a migitation
4744 auto,nosmt - automatically select a mitigation,
4745 disabling SMT if necessary for
4746 the full mitigation (only on Zen1
4747 and older without STIBP).
4748 ibpb - On AMD, mitigate short speculation
4749 windows on basic block boundaries too.
4750 Safe, highest perf impact. It also
4751 enables STIBP if present. Not suitable
4753 ibpb,nosmt - Like "ibpb" above but will disable SMT
4754 when STIBP is not available. This is
4755 the alternative for systems which do not
4757 unret - Force enable untrained return thunks,
4758 only effective on AMD f15h-f17h based
4760 unret,nosmt - Like unret, but will disable SMT when STIBP
4761 is not available. This is the alternative for
4762 systems which do not have STIBP.
4764 Selecting 'auto' will choose a mitigation method at run
4765 time according to the CPU.
4767 Not specifying this option is equivalent to retbleed=auto.
4769 rfkill.default_state=
4770 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4771 etc. communication is blocked by default.
4774 rfkill.master_switch_mode=
4775 0 The "airplane mode" button does nothing.
4776 1 The "airplane mode" button toggles between everything
4777 blocked and the previous configuration.
4778 2 The "airplane mode" button toggles between everything
4779 blocked and everything unblocked.
4781 rhash_entries= [KNL,NET]
4782 Set number of hash buckets for route cache
4785 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4788 ro [KNL] Mount root device read-only on boot
4791 on Mark read-only kernel memory as read-only (default).
4792 off Leave read-only kernel memory writable for debugging.
4795 Enable the uart passthrough on the designated usb port
4796 on Rockchip SoCs. When active, the signals of the
4797 debug-uart get routed to the D+ and D- pins of the usb
4798 port and the regular usb controller gets disabled.
4800 root= [KNL] Root filesystem
4801 See name_to_dev_t comment in init/do_mounts.c.
4803 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4804 mount the root filesystem
4806 rootflags= [KNL] Set root filesystem mount option string
4808 rootfstype= [KNL] Set root filesystem type
4810 rootwait [KNL] Wait (indefinitely) for root device to show up.
4811 Useful for devices that are detected asynchronously
4812 (e.g. USB and MMC devices).
4814 rproc_mem=nn[KMG][@address]
4815 [KNL,ARM,CMA] Remoteproc physical memory block.
4816 Memory area to be used by remote processor image,
4819 rw [KNL] Mount root device read-write on boot
4821 S [KNL] Run init in single mode
4823 s390_iommu= [HW,S390]
4824 Set s390 IOTLB flushing mode
4826 With strict flushing every unmap operation will result in
4827 an IOTLB flush. Default is lazy flushing before reuse,
4831 See drivers/net/irda/sa1100_ir.c.
4833 sbni= [NET] Granch SBNI12 leased line adapter
4835 sched_debug [KNL] Enables verbose scheduler debug messages.
4837 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4838 Allowed values are enable and disable. This feature
4839 incurs a small amount of overhead in the scheduler
4840 but is useful for debugging and performance tuning.
4842 sched_thermal_decay_shift=
4843 [KNL, SMP] Set a decay shift for scheduler thermal
4844 pressure signal. Thermal pressure signal follows the
4845 default decay period of other scheduler pelt
4846 signals(usually 32 ms but configurable). Setting
4847 sched_thermal_decay_shift will left shift the decay
4848 period for the thermal pressure signal by the shift
4850 i.e. with the default pelt decay period of 32 ms
4851 sched_thermal_decay_shift thermal pressure decay pr
4855 Format: integer between 0 and 10
4858 scftorture.holdoff= [KNL]
4859 Number of seconds to hold off before starting
4860 test. Defaults to zero for module insertion and
4861 to 10 seconds for built-in smp_call_function()
4864 scftorture.longwait= [KNL]
4865 Request ridiculously long waits randomly selected
4866 up to the chosen limit in seconds. Zero (the
4867 default) disables this feature. Please note
4868 that requesting even small non-zero numbers of
4869 seconds can result in RCU CPU stall warnings,
4870 softlockup complaints, and so on.
4872 scftorture.nthreads= [KNL]
4873 Number of kthreads to spawn to invoke the
4874 smp_call_function() family of functions.
4875 The default of -1 specifies a number of kthreads
4876 equal to the number of CPUs.
4878 scftorture.onoff_holdoff= [KNL]
4879 Number seconds to wait after the start of the
4880 test before initiating CPU-hotplug operations.
4882 scftorture.onoff_interval= [KNL]
4883 Number seconds to wait between successive
4884 CPU-hotplug operations. Specifying zero (which
4885 is the default) disables CPU-hotplug operations.
4887 scftorture.shutdown_secs= [KNL]
4888 The number of seconds following the start of the
4889 test after which to shut down the system. The
4890 default of zero avoids shutting down the system.
4891 Non-zero values are useful for automated tests.
4893 scftorture.stat_interval= [KNL]
4894 The number of seconds between outputting the
4895 current test statistics to the console. A value
4896 of zero disables statistics output.
4898 scftorture.stutter_cpus= [KNL]
4899 The number of jiffies to wait between each change
4900 to the set of CPUs under test.
4902 scftorture.use_cpus_read_lock= [KNL]
4903 Use use_cpus_read_lock() instead of the default
4904 preempt_disable() to disable CPU hotplug
4905 while invoking one of the smp_call_function*()
4908 scftorture.verbose= [KNL]
4909 Enable additional printk() statements.
4911 scftorture.weight_single= [KNL]
4912 The probability weighting to use for the
4913 smp_call_function_single() function with a zero
4914 "wait" parameter. A value of -1 selects the
4915 default if all other weights are -1. However,
4916 if at least one weight has some other value, a
4917 value of -1 will instead select a weight of zero.
4919 scftorture.weight_single_wait= [KNL]
4920 The probability weighting to use for the
4921 smp_call_function_single() function with a
4922 non-zero "wait" parameter. See weight_single.
4924 scftorture.weight_many= [KNL]
4925 The probability weighting to use for the
4926 smp_call_function_many() function with a zero
4927 "wait" parameter. See weight_single.
4928 Note well that setting a high probability for
4929 this weighting can place serious IPI load
4932 scftorture.weight_many_wait= [KNL]
4933 The probability weighting to use for the
4934 smp_call_function_many() function with a
4935 non-zero "wait" parameter. See weight_single
4938 scftorture.weight_all= [KNL]
4939 The probability weighting to use for the
4940 smp_call_function_all() function with a zero
4941 "wait" parameter. See weight_single and
4944 scftorture.weight_all_wait= [KNL]
4945 The probability weighting to use for the
4946 smp_call_function_all() function with a
4947 non-zero "wait" parameter. See weight_single
4950 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4951 xtime_lock contention on larger systems, and/or RCU lock
4952 contention on all systems with CONFIG_MAXSMP set.
4953 Format: { "0" | "1" }
4954 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4956 Note: increases power consumption, thus should only be
4957 enabled if running jitter sensitive (HPC/RT) workloads.
4959 security= [SECURITY] Choose a legacy "major" security module to
4960 enable at boot. This has been deprecated by the
4963 selinux= [SELINUX] Disable or enable SELinux at boot time.
4964 Format: { "0" | "1" }
4965 See security/selinux/Kconfig help text.
4970 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4971 Format: { "0" | "1" }
4972 See security/apparmor/Kconfig help text
4975 Default value is set via kernel config option.
4977 serialnumber [BUGS=X86-32]
4980 Maximal number of shapers.
4988 Disable merging of slabs with similar size. May be
4989 necessary if there is some reason to distinguish
4990 allocs to different slabs, especially in hardened
4991 environments where the risk of heap overflows and
4992 layout control by attackers can usually be
4993 frustrated by disabling merging. This will reduce
4994 most of the exposure of a heap attack to a single
4995 cache (risks via metadata attacks are mostly
4996 unchanged). Debug options disable merging on their
4998 For more information see Documentation/vm/slub.rst.
5000 slab_max_order= [MM, SLAB]
5001 Determines the maximum allowed order for slabs.
5002 A high setting may cause OOMs due to memory
5003 fragmentation. Defaults to 1 for systems with
5004 more than 32MB of RAM, 0 otherwise.
5006 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5007 Enabling slub_debug allows one to determine the
5008 culprit if slab objects become corrupted. Enabling
5009 slub_debug can create guard zones around objects and
5010 may poison objects when not in use. Also tracks the
5011 last alloc / free. For more information see
5012 Documentation/vm/slub.rst.
5014 slub_memcg_sysfs= [MM, SLUB]
5015 Determines whether to enable sysfs directories for
5016 memory cgroup sub-caches. 1 to enable, 0 to disable.
5017 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
5018 Enabling this can lead to a very high number of debug
5019 directories and files being created under
5022 slub_max_order= [MM, SLUB]
5023 Determines the maximum allowed order for slabs.
5024 A high setting may cause OOMs due to memory
5025 fragmentation. For more information see
5026 Documentation/vm/slub.rst.
5028 slub_min_objects= [MM, SLUB]
5029 The minimum number of objects per slab. SLUB will
5030 increase the slab order up to slub_max_order to
5031 generate a sufficiently large slab able to contain
5032 the number of objects indicated. The higher the number
5033 of objects the smaller the overhead of tracking slabs
5034 and the less frequently locks need to be acquired.
5035 For more information see Documentation/vm/slub.rst.
5037 slub_min_order= [MM, SLUB]
5038 Determines the minimum page order for slabs. Must be
5039 lower than slub_max_order.
5040 For more information see Documentation/vm/slub.rst.
5042 slub_nomerge [MM, SLUB]
5043 Same with slab_nomerge. This is supported for legacy.
5044 See slab_nomerge for more information.
5047 Format: <io1>[,<io2>[,...,<io8>]]
5049 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5050 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5051 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5052 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5053 smsc-ircc2.ircc_irq= [HW] IRQ line
5054 smsc-ircc2.ircc_dma= [HW] DMA channel
5055 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5056 0: Toshiba Satellite 1800 (GP data pin select)
5057 1: Fast pin select (default)
5060 smt [KNL,S390] Set the maximum number of threads (logical
5061 CPUs) to use per physical CPU on systems capable of
5062 symmetric multithreading (SMT). Will be capped to the
5063 actual hardware limit.
5065 Default: -1 (no limit)
5068 [KNL] Should the soft-lockup detector generate panics.
5071 A value of 1 instructs the soft-lockup detector
5072 to panic the machine when a soft-lockup occurs. It is
5073 also controlled by the kernel.softlockup_panic sysctl
5074 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5075 respective build-time switch to that functionality.
5077 softlockup_all_cpu_backtrace=
5078 [KNL] Should the soft-lockup detector generate
5079 backtraces on all cpus.
5082 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5083 See Documentation/admin-guide/laptops/sonypi.rst
5085 spectre_v2= [X86] Control mitigation of Spectre variant 2
5086 (indirect branch speculation) vulnerability.
5087 The default operation protects the kernel from
5090 on - unconditionally enable, implies
5092 off - unconditionally disable, implies
5094 auto - kernel detects whether your CPU model is
5097 Selecting 'on' will, and 'auto' may, choose a
5098 mitigation method at run time according to the
5099 CPU, the available microcode, the setting of the
5100 CONFIG_RETPOLINE configuration option, and the
5101 compiler with which the kernel was built.
5103 Selecting 'on' will also enable the mitigation
5104 against user space to user space task attacks.
5106 Selecting 'off' will disable both the kernel and
5107 the user space protections.
5109 Specific mitigations can also be selected manually:
5111 retpoline - replace indirect branches
5112 retpoline,generic - Retpolines
5113 retpoline,lfence - LFENCE; indirect branch
5114 retpoline,amd - alias for retpoline,lfence
5115 eibrs - Enhanced/Auto IBRS
5116 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines
5117 eibrs,lfence - Enhanced/Auto IBRS + LFENCE
5118 ibrs - use IBRS to protect kernel
5120 Not specifying this option is equivalent to
5124 [X86] Control mitigation of Spectre variant 2
5125 (indirect branch speculation) vulnerability between
5128 on - Unconditionally enable mitigations. Is
5129 enforced by spectre_v2=on
5131 off - Unconditionally disable mitigations. Is
5132 enforced by spectre_v2=off
5134 prctl - Indirect branch speculation is enabled,
5135 but mitigation can be enabled via prctl
5136 per thread. The mitigation control state
5137 is inherited on fork.
5140 - Like "prctl" above, but only STIBP is
5141 controlled per thread. IBPB is issued
5142 always when switching between different user
5146 - Same as "prctl" above, but all seccomp
5147 threads will enable the mitigation unless
5148 they explicitly opt out.
5151 - Like "seccomp" above, but only STIBP is
5152 controlled per thread. IBPB is issued
5153 always when switching between different
5154 user space processes.
5156 auto - Kernel selects the mitigation depending on
5157 the available CPU features and vulnerability.
5160 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
5162 Not specifying this option is equivalent to
5163 spectre_v2_user=auto.
5165 spec_rstack_overflow=
5166 [X86] Control RAS overflow mitigation on AMD Zen CPUs
5168 off - Disable mitigation
5169 microcode - Enable microcode mitigation only
5170 safe-ret - Enable sw-only safe RET mitigation (default)
5171 ibpb - Enable mitigation by issuing IBPB on
5173 ibpb-vmexit - Issue IBPB only on VMEXIT
5174 (cloud-specific mitigation)
5176 spec_store_bypass_disable=
5177 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5178 (Speculative Store Bypass vulnerability)
5180 Certain CPUs are vulnerable to an exploit against a
5181 a common industry wide performance optimization known
5182 as "Speculative Store Bypass" in which recent stores
5183 to the same memory location may not be observed by
5184 later loads during speculative execution. The idea
5185 is that such stores are unlikely and that they can
5186 be detected prior to instruction retirement at the
5187 end of a particular speculation execution window.
5189 In vulnerable processors, the speculatively forwarded
5190 store can be used in a cache side channel attack, for
5191 example to read memory to which the attacker does not
5192 directly have access (e.g. inside sandboxed code).
5194 This parameter controls whether the Speculative Store
5195 Bypass optimization is used.
5197 On x86 the options are:
5199 on - Unconditionally disable Speculative Store Bypass
5200 off - Unconditionally enable Speculative Store Bypass
5201 auto - Kernel detects whether the CPU model contains an
5202 implementation of Speculative Store Bypass and
5203 picks the most appropriate mitigation. If the
5204 CPU is not vulnerable, "off" is selected. If the
5205 CPU is vulnerable the default mitigation is
5206 architecture and Kconfig dependent. See below.
5207 prctl - Control Speculative Store Bypass per thread
5208 via prctl. Speculative Store Bypass is enabled
5209 for a process by default. The state of the control
5210 is inherited on fork.
5211 seccomp - Same as "prctl" above, but all seccomp threads
5212 will disable SSB unless they explicitly opt out.
5214 Default mitigations:
5215 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
5217 On powerpc the options are:
5219 on,auto - On Power8 and Power9 insert a store-forwarding
5220 barrier on kernel entry and exit. On Power7
5221 perform a software flush on kernel entry and
5225 Not specifying this option is equivalent to
5226 spec_store_bypass_disable=auto.
5228 spia_io_base= [HW,MTD]
5234 [X86] Enable split lock detection
5236 When enabled (and if hardware support is present), atomic
5237 instructions that access data across cache line
5238 boundaries will result in an alignment check exception.
5242 warn - the kernel will emit rate limited warnings
5243 about applications triggering the #AC
5244 exception. This mode is the default on CPUs
5245 that supports split lock detection.
5247 fatal - the kernel will send SIGBUS to applications
5248 that trigger the #AC exception.
5250 If an #AC exception is hit in the kernel or in
5251 firmware (i.e. not while executing in user mode)
5252 the kernel will oops in either "warn" or "fatal"
5256 Control the Special Register Buffer Data Sampling
5259 Certain CPUs are vulnerable to an MDS-like
5260 exploit which can leak bits from the random
5263 By default, this issue is mitigated by
5264 microcode. However, the microcode fix can cause
5265 the RDRAND and RDSEED instructions to become
5266 much slower. Among other effects, this will
5267 result in reduced throughput from /dev/urandom.
5269 The microcode mitigation can be disabled with
5270 the following option:
5272 off: Disable mitigation and remove
5273 performance impact to RDRAND and RDSEED
5275 srcutree.counter_wrap_check [KNL]
5276 Specifies how frequently to check for
5277 grace-period sequence counter wrap for the
5278 srcu_data structure's ->srcu_gp_seq_needed field.
5279 The greater the number of bits set in this kernel
5280 parameter, the less frequently counter wrap will
5281 be checked for. Note that the bottom two bits
5284 srcutree.exp_holdoff [KNL]
5285 Specifies how many nanoseconds must elapse
5286 since the end of the last SRCU grace period for
5287 a given srcu_struct until the next normal SRCU
5288 grace period will be considered for automatic
5289 expediting. Set to zero to disable automatic
5293 Speculative Store Bypass Disable control
5295 On CPUs that are vulnerable to the Speculative
5296 Store Bypass vulnerability and offer a
5297 firmware based mitigation, this parameter
5298 indicates how the mitigation should be used:
5300 force-on: Unconditionally enable mitigation for
5301 for both kernel and userspace
5302 force-off: Unconditionally disable mitigation for
5303 for both kernel and userspace
5304 kernel: Always enable mitigation in the
5305 kernel, and offer a prctl interface
5306 to allow userspace to register its
5307 interest in being mitigated too.
5309 stack_guard_gap= [MM]
5310 override the default stack gap protection. The value
5311 is in page units and it defines how many pages prior
5312 to (for stacks growing down) resp. after (for stacks
5313 growing up) the main stack are reserved for no other
5314 mapping. Default value is 256 pages.
5317 Enabled the stack tracer on boot up.
5319 stacktrace_filter=[function-list]
5320 [FTRACE] Limit the functions that the stack tracer
5321 will trace at boot up. function-list is a comma separated
5322 list of functions. This list can be changed at run
5323 time by the stack_trace_filter file in the debugfs
5324 tracing directory. Note, this enables stack tracing
5325 and the stacktrace above is not needed.
5329 Set the STI (builtin display/keyboard on the HP-PARISC
5330 machines) console (graphic card) which should be used
5331 as the initial boot-console.
5332 See also comment in drivers/video/console/sticore.c.
5335 See comment in drivers/video/console/sticore.c.
5338 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5340 sunrpc.min_resvport=
5341 sunrpc.max_resvport=
5343 SunRPC servers often require that client requests
5344 originate from a privileged port (i.e. a port in the
5345 range 0 < portnr < 1024).
5346 An administrator who wishes to reserve some of these
5347 ports for other uses may adjust the range that the
5348 kernel's sunrpc client considers to be privileged
5349 using these two parameters to set the minimum and
5350 maximum port values.
5352 sunrpc.svc_rpc_per_connection_limit=
5354 Limit the number of requests that the server will
5355 process in parallel from a single connection.
5356 The default value is 0 (no limit).
5360 Control how the NFS server code allocates CPUs to
5361 service thread pools. Depending on how many NICs
5362 you have and where their interrupts are bound, this
5363 option will affect which CPUs will do NFS serving.
5364 Note: this parameter cannot be changed while the
5365 NFS server is running.
5367 auto the server chooses an appropriate mode
5368 automatically using heuristics
5369 global a single global pool contains all CPUs
5370 percpu one pool for each CPU
5371 pernode one pool for each NUMA node (equivalent
5372 to global on non-NUMA machines)
5374 sunrpc.tcp_slot_table_entries=
5375 sunrpc.udp_slot_table_entries=
5377 Sets the upper limit on the number of simultaneous
5378 RPC calls that can be sent from the client to a
5379 server. Increasing these values may allow you to
5380 improve throughput, but will also increase the
5381 amount of memory reserved for use by the client.
5383 suspend.pm_test_delay=
5385 Sets the number of seconds to remain in a suspend test
5386 mode before resuming the system (see
5387 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5388 is set. Default value is 5.
5391 Format: { on | off | y | n | 1 | 0 }
5392 This parameter controls use of the Protected
5393 Execution Facility on pSeries.
5396 [KNL] Enable accounting of swap in memory resource
5397 controller if no parameter or 1 is given or disable
5398 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5400 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5401 Format: { <int> | force | noforce }
5402 <int> -- Number of I/O TLB slabs
5403 force -- force using of bounce buffers even if they
5404 wouldn't be automatically used by the kernel
5405 noforce -- Never use bounce buffers (for debugging)
5410 Set a sysctl parameter, right before loading the init
5411 process, as if the value was written to the respective
5412 /proc/sys/... file. Both '.' and '/' are recognized as
5413 separators. Unrecognized parameters and invalid values
5414 are reported in the kernel log. Sysctls registered
5415 later by a loaded module cannot be set this way.
5416 Example: sysctl.vm.swappiness=40
5418 sysfs.deprecated=0|1 [KNL]
5419 Enable/disable old style sysfs layout for old udev
5420 on older distributions. When this option is enabled
5421 very new udev will not work anymore. When this option
5422 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5423 in older udev will not work anymore.
5424 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5425 the kernel configuration.
5427 sysrq_always_enabled
5429 Ignore sysrq setting - this boot parameter will
5430 neutralize any effect of /proc/sys/kernel/sysrq.
5431 Useful for debugging.
5433 tcpmhash_entries= [KNL,NET]
5434 Set the number of tcp_metrics_hash slots.
5435 Default value is 8192 or 16384 depending on total
5436 ram pages. This is used to specify the TCP metrics
5437 cache size. See Documentation/networking/ip-sysctl.rst
5438 "tcp_no_metrics_save" section for more details.
5442 test_suspend= [SUSPEND][,N]
5443 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5444 standby suspend) or "freeze" (for suspend type freeze)
5445 as the system sleep state during system startup with
5446 the optional capability to repeat N number of times.
5447 The system is woken from this state using a
5448 wakeup-capable RTC alarm.
5450 thash_entries= [KNL,NET]
5451 Set number of hash buckets for TCP connection
5453 thermal.act= [HW,ACPI]
5454 -1: disable all active trip points in all thermal zones
5455 <degrees C>: override all lowest active trip points
5457 thermal.crt= [HW,ACPI]
5458 -1: disable all critical trip points in all thermal zones
5459 <degrees C>: override all critical trip points
5461 thermal.nocrt= [HW,ACPI]
5462 Set to disable actions on ACPI thermal zone
5463 critical and hot trip points.
5465 thermal.off= [HW,ACPI]
5466 1: disable ACPI thermal control
5468 thermal.psv= [HW,ACPI]
5469 -1: disable all passive trip points
5470 <degrees C>: override all passive trip points to this
5473 thermal.tzp= [HW,ACPI]
5474 Specify global default ACPI thermal zone polling rate
5475 <deci-seconds>: poll all this frequency
5476 0: no polling (default)
5479 Force threading of all interrupt handlers except those
5480 marked explicitly IRQF_NO_THREAD.
5484 Specify if the kernel should make use of the cpu
5485 topology information if the hardware supports this.
5486 The scheduler will make use of this information and
5487 e.g. base its process migration decisions on it.
5490 topology_updates= [KNL, PPC, NUMA]
5492 Specify if the kernel should ignore (off)
5493 topology updates sent by the hypervisor to this
5496 torture.disable_onoff_at_boot= [KNL]
5497 Prevent the CPU-hotplug component of torturing
5498 until after init has spawned.
5500 torture.ftrace_dump_at_shutdown= [KNL]
5501 Dump the ftrace buffer at torture-test shutdown,
5502 even if there were no errors. This can be a
5503 very costly operation when many torture tests
5504 are running concurrently, especially on systems
5505 with rotating-rust storage.
5509 tpm_suspend_pcr=[HW,TPM]
5510 Format: integer pcr id
5511 Specify that at suspend time, the tpm driver
5512 should extend the specified pcr with zeros,
5513 as a workaround for some chips which fail to
5514 flush the last written pcr on TPM_SaveState.
5515 This will guarantee that all the other pcrs
5518 trace_buf_size=nn[KMG]
5519 [FTRACE] will set tracing buffer size on each cpu.
5521 trace_event=[event-list]
5522 [FTRACE] Set and start specified trace events in order
5523 to facilitate early boot debugging. The event-list is a
5524 comma separated list of trace events to enable. See
5525 also Documentation/trace/events.rst
5527 trace_options=[option-list]
5528 [FTRACE] Enable or disable tracer options at boot.
5529 The option-list is a comma delimited list of options
5530 that can be enabled or disabled just as if you were
5531 to echo the option name into
5533 /sys/kernel/debug/tracing/trace_options
5535 For example, to enable stacktrace option (to dump the
5536 stack trace of each event), add to the command line:
5538 trace_options=stacktrace
5540 See also Documentation/trace/ftrace.rst "trace options"
5544 Have the tracepoints sent to printk as well as the
5545 tracing ring buffer. This is useful for early boot up
5546 where the system hangs or reboots and does not give the
5547 option for reading the tracing buffer or performing a
5548 ftrace_dump_on_oops.
5550 To turn off having tracepoints sent to printk,
5551 echo 0 > /proc/sys/kernel/tracepoint_printk
5552 Note, echoing 1 into this file without the
5553 tracepoint_printk kernel cmdline option has no effect.
5557 Having tracepoints sent to printk() and activating high
5558 frequency tracepoints such as irq or sched, can cause
5559 the system to live lock.
5562 [FTRACE] enable this option to disable tracing when a
5563 warning is hit. This turns off "tracing_on". Tracing can
5564 be enabled again by echoing '1' into the "tracing_on"
5565 file located in /sys/kernel/debug/tracing/
5567 This option is useful, as it disables the trace before
5568 the WARNING dump is called, which prevents the trace to
5569 be filled with content caused by the warning output.
5571 This option can also be set at run time via the sysctl
5572 option: kernel/traceoff_on_warning
5574 transparent_hugepage=
5576 Format: [always|madvise|never]
5577 Can be used to control the default behavior of the system
5578 with respect to transparent hugepages.
5579 See Documentation/admin-guide/mm/transhuge.rst
5582 tsc= Disable clocksource stability checks for TSC.
5584 [x86] reliable: mark tsc clocksource as reliable, this
5585 disables clocksource verification at runtime, as well
5586 as the stability checks done at bootup. Used to enable
5587 high-resolution timer mode on older hardware, and in
5588 virtualized environment.
5589 [x86] noirqtime: Do not use TSC to do irq accounting.
5590 Used to run time disable IRQ_TIME_ACCOUNTING on any
5591 platforms where RDTSC is slow and this accounting
5593 [x86] unstable: mark the TSC clocksource as unstable, this
5594 marks the TSC unconditionally unstable at bootup and
5595 avoids any further wobbles once the TSC watchdog notices.
5596 [x86] nowatchdog: disable clocksource watchdog. Used
5597 in situations with strict latency requirements (where
5598 interruptions from clocksource watchdog are not
5601 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5602 value instead. Useful when the early TSC frequency discovery
5603 procedure is not reliable, such as on overclocked systems
5604 with CPUID.16h support and partial CPUID.15h support.
5605 Format: <unsigned int>
5607 tsx= [X86] Control Transactional Synchronization
5608 Extensions (TSX) feature in Intel processors that
5609 support TSX control.
5611 This parameter controls the TSX feature. The options are:
5613 on - Enable TSX on the system. Although there are
5614 mitigations for all known security vulnerabilities,
5615 TSX has been known to be an accelerator for
5616 several previous speculation-related CVEs, and
5617 so there may be unknown security risks associated
5618 with leaving it enabled.
5620 off - Disable TSX on the system. (Note that this
5621 option takes effect only on newer CPUs which are
5622 not vulnerable to MDS, i.e., have
5623 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5624 the new IA32_TSX_CTRL MSR through a microcode
5625 update. This new MSR allows for the reliable
5626 deactivation of the TSX functionality.)
5628 auto - Disable TSX if X86_BUG_TAA is present,
5629 otherwise enable TSX on the system.
5631 Not specifying this option is equivalent to tsx=off.
5633 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5636 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5637 Abort (TAA) vulnerability.
5639 Similar to Micro-architectural Data Sampling (MDS)
5640 certain CPUs that support Transactional
5641 Synchronization Extensions (TSX) are vulnerable to an
5642 exploit against CPU internal buffers which can forward
5643 information to a disclosure gadget under certain
5646 In vulnerable processors, the speculatively forwarded
5647 data can be used in a cache side channel attack, to
5648 access data to which the attacker does not have direct
5651 This parameter controls the TAA mitigation. The
5654 full - Enable TAA mitigation on vulnerable CPUs
5657 full,nosmt - Enable TAA mitigation and disable SMT on
5658 vulnerable CPUs. If TSX is disabled, SMT
5659 is not disabled because CPU is not
5660 vulnerable to cross-thread TAA attacks.
5661 off - Unconditionally disable TAA mitigation
5663 On MDS-affected machines, tsx_async_abort=off can be
5664 prevented by an active MDS mitigation as both vulnerabilities
5665 are mitigated with the same mechanism so in order to disable
5666 this mitigation, you need to specify mds=off too.
5668 Not specifying this option is equivalent to
5669 tsx_async_abort=full. On CPUs which are MDS affected
5670 and deploy MDS mitigation, TAA mitigation is not
5671 required and doesn't provide any additional
5675 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5677 turbografx.map[2|3]= [HW,JOY]
5678 TurboGraFX parallel port interface
5680 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5681 See also Documentation/input/devices/joystick-parport.rst
5683 udbg-immortal [PPC] When debugging early kernel crashes that
5684 happen after console_init() and before a proper
5685 console driver takes over, this boot options might
5686 help "seeing" what's going on.
5688 uhash_entries= [KNL,NET]
5689 Set number of hash buckets for UDP/UDP-Lite connections
5692 [USB] Ignore overcurrent events (default N).
5693 Some badly-designed motherboards generate lots of
5694 bogus events, for ports that aren't wired to
5695 anything. Set this parameter to avoid log spamming.
5696 Note that genuine overcurrent events won't be
5700 [X86] Cause panic on unknown NMI.
5702 usbcore.authorized_default=
5703 [USB] Default USB device authorization:
5704 (default -1 = authorized except for wireless USB,
5705 0 = not authorized, 1 = authorized, 2 = authorized
5706 if device connected to internal port)
5708 usbcore.autosuspend=
5709 [USB] The autosuspend time delay (in seconds) used
5710 for newly-detected USB devices (default 2). This
5711 is the time required before an idle device will be
5712 autosuspended. Devices for which the delay is set
5713 to a negative value won't be autosuspended at all.
5715 usbcore.usbfs_snoop=
5716 [USB] Set to log all usbfs traffic (default 0 = off).
5718 usbcore.usbfs_snoop_max=
5719 [USB] Maximum number of bytes to snoop in each URB
5722 usbcore.blinkenlights=
5723 [USB] Set to cycle leds on hubs (default 0 = off).
5725 usbcore.old_scheme_first=
5726 [USB] Start with the old device initialization
5727 scheme (default 0 = off).
5729 usbcore.usbfs_memory_mb=
5730 [USB] Memory limit (in MB) for buffers allocated by
5731 usbfs (default = 16, 0 = max = 2047).
5733 usbcore.use_both_schemes=
5734 [USB] Try the other device initialization scheme
5735 if the first one fails (default 1 = enabled).
5737 usbcore.initial_descriptor_timeout=
5738 [USB] Specifies timeout for the initial 64-byte
5739 USB_REQ_GET_DESCRIPTOR request in milliseconds
5740 (default 5000 = 5.0 seconds).
5742 usbcore.nousb [USB] Disable the USB subsystem
5745 [USB] A list of quirk entries to augment the built-in
5746 usb core quirk list. List entries are separated by
5747 commas. Each entry has the form
5748 VendorID:ProductID:Flags. The IDs are 4-digit hex
5749 numbers and Flags is a set of letters. Each letter
5750 will change the built-in quirk; setting it if it is
5751 clear and clearing it if it is set. The letters have
5752 the following meanings:
5753 a = USB_QUIRK_STRING_FETCH_255 (string
5754 descriptors must not be fetched using
5756 b = USB_QUIRK_RESET_RESUME (device can't resume
5757 correctly so reset it instead);
5758 c = USB_QUIRK_NO_SET_INTF (device can't handle
5759 Set-Interface requests);
5760 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5761 handle its Configuration or Interface
5763 e = USB_QUIRK_RESET (device can't be reset
5764 (e.g morph devices), don't use reset);
5765 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5766 more interface descriptions than the
5767 bNumInterfaces count, and can't handle
5768 talking to these interfaces);
5769 g = USB_QUIRK_DELAY_INIT (device needs a pause
5770 during initialization, after we read
5771 the device descriptor);
5772 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5773 high speed and super speed interrupt
5774 endpoints, the USB 2.0 and USB 3.0 spec
5775 require the interval in microframes (1
5776 microframe = 125 microseconds) to be
5777 calculated as interval = 2 ^
5779 Devices with this quirk report their
5780 bInterval as the result of this
5781 calculation instead of the exponent
5782 variable used in the calculation);
5783 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5784 handle device_qualifier descriptor
5786 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5787 generates spurious wakeup, ignore
5788 remote wakeup capability);
5789 k = USB_QUIRK_NO_LPM (device can't handle Link
5791 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5792 (Device reports its bInterval as linear
5793 frames instead of the USB 2.0
5795 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5796 to be disconnected before suspend to
5797 prevent spurious wakeup);
5798 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5799 pause after every control message);
5800 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5801 delay after resetting its port);
5802 Example: quirks=0781:5580:bk,0a5c:5834:gij
5805 [USBHID] The interval which mice are to be polled at.
5808 [USBHID] The interval which joysticks are to be polled at.
5811 [USBHID] The interval which keyboards are to be polled at.
5813 usb-storage.delay_use=
5814 [UMS] The delay in seconds before a new device is
5815 scanned for Logical Units (default 1).
5818 [UMS] A list of quirks entries to supplement or
5819 override the built-in unusual_devs list. List
5820 entries are separated by commas. Each entry has
5821 the form VID:PID:Flags where VID and PID are Vendor
5822 and Product ID values (4-digit hex numbers) and
5823 Flags is a set of characters, each corresponding
5824 to a common usb-storage quirk flag as follows:
5825 a = SANE_SENSE (collect more than 18 bytes
5826 of sense data, not on uas);
5827 b = BAD_SENSE (don't collect more than 18
5828 bytes of sense data, not on uas);
5829 c = FIX_CAPACITY (decrease the reported
5830 device capacity by one sector);
5831 d = NO_READ_DISC_INFO (don't use
5832 READ_DISC_INFO command, not on uas);
5833 e = NO_READ_CAPACITY_16 (don't use
5834 READ_CAPACITY_16 command);
5835 f = NO_REPORT_OPCODES (don't use report opcodes
5837 g = MAX_SECTORS_240 (don't transfer more than
5838 240 sectors at a time, uas only);
5839 h = CAPACITY_HEURISTICS (decrease the
5840 reported device capacity by one
5841 sector if the number is odd);
5842 i = IGNORE_DEVICE (don't bind to this
5844 j = NO_REPORT_LUNS (don't use report luns
5846 k = NO_SAME (do not use WRITE_SAME, uas only)
5847 l = NOT_LOCKABLE (don't try to lock and
5848 unlock ejectable media, not on uas);
5849 m = MAX_SECTORS_64 (don't transfer more
5850 than 64 sectors = 32 KB at a time,
5852 n = INITIAL_READ10 (force a retry of the
5853 initial READ(10) command, not on uas);
5854 o = CAPACITY_OK (accept the capacity
5855 reported by the device, not on uas);
5856 p = WRITE_CACHE (the device cache is ON
5857 by default, not on uas);
5858 r = IGNORE_RESIDUE (the device reports
5859 bogus residue values, not on uas);
5860 s = SINGLE_LUN (the device has only one
5862 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5863 commands, uas only);
5864 u = IGNORE_UAS (don't bind to the uas driver);
5865 w = NO_WP_DETECT (don't test whether the
5866 medium is write-protected).
5867 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5868 even if the device claims no cache,
5870 Example: quirks=0419:aaf5:rl,0421:0433:rc
5872 user_debug= [KNL,ARM]
5874 See arch/arm/Kconfig.debug help text.
5875 1 - undefined instruction events
5877 4 - invalid data aborts
5880 Example: user_debug=31
5883 [X86] Flags controlling user PTE allocations.
5885 nohigh = do not allocate PTE pages in
5886 HIGHMEM regardless of setting
5890 On X86_32, this is an alias for vdso32=. Otherwise:
5892 vdso=1: enable VDSO (the default)
5893 vdso=0: disable VDSO mapping
5895 vdso32= [X86] Control the 32-bit vDSO
5896 vdso32=1: enable 32-bit VDSO
5897 vdso32=0 or vdso32=2: disable 32-bit VDSO
5899 See the help text for CONFIG_COMPAT_VDSO for more
5900 details. If CONFIG_COMPAT_VDSO is set, the default is
5901 vdso32=0; otherwise, the default is vdso32=1.
5903 For compatibility with older kernels, vdso32=2 is an
5906 Try vdso32=0 if you encounter an error that says:
5907 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5910 vector=percpu: enable percpu vector domain
5912 video= [FB] Frame buffer configuration
5913 See Documentation/fb/modedb.rst.
5915 video.brightness_switch_enabled= [0,1]
5916 If set to 1, on receiving an ACPI notify event
5917 generated by hotkey, video driver will adjust brightness
5918 level and then send out the event to user space through
5919 the allocated input device; If set to 0, video driver
5920 will only send out the event without touching backlight
5925 [VMMIO] Memory mapped virtio (platform) device.
5927 <size>@<baseaddr>:<irq>[:<id>]
5929 <size> := size (can use standard suffixes
5931 <baseaddr> := physical base address
5932 <irq> := interrupt number (as passed to
5934 <id> := (optional) platform device id
5936 virtio_mmio.device=1K@0x100b0000:48:7
5938 Can be used multiple times for multiple devices.
5940 vga= [BOOT,X86-32] Select a particular video mode
5941 See Documentation/x86/boot.rst and
5942 Documentation/admin-guide/svga.rst.
5943 Use vga=ask for menu.
5944 This is actually a boot loader parameter; the value is
5945 passed to the kernel using a special protocol.
5947 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
5948 May slow down system boot speed, especially when
5949 enabled on systems with a large amount of memory.
5950 All options are enabled by default, and this
5951 interface is meant to allow for selectively
5952 enabling or disabling specific virtual memory
5955 Available options are:
5956 P Enable page structure init time poisoning
5957 - Disable all of the above options
5959 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5960 size of <nn>. This can be used to increase the
5961 minimum size (128MB on x86). It can also be used to
5962 decrease the size and leave more room for directly
5965 vmcp_cma=nn[MG] [KNL,S390]
5966 Sets the memory size reserved for contiguous memory
5967 allocations for the vmcp device driver.
5969 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5972 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5975 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5979 Controls the behavior of vsyscalls (i.e. calls to
5980 fixed addresses of 0xffffffffff600x00 from legacy
5981 code). Most statically-linked binaries and older
5982 versions of glibc use these calls. Because these
5983 functions are at fixed addresses, they make nice
5984 targets for exploits that can control RIP.
5986 emulate [default] Vsyscalls turn into traps and are
5987 emulated reasonably safely. The vsyscall
5990 xonly Vsyscalls turn into traps and are
5991 emulated reasonably safely. The vsyscall
5992 page is not readable.
5994 none Vsyscalls don't work at all. This makes
5995 them quite hard to use for exploits but
5996 might break your system.
5998 vt.color= [VT] Default text color.
5999 Format: 0xYX, X = foreground, Y = background.
6000 Default: 0x07 = light gray on black.
6002 vt.cur_default= [VT] Default cursor shape.
6003 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6004 the parameters of the <Esc>[?A;B;Cc escape sequence;
6005 see VGA-softcursor.txt. Default: 2 = underline.
6007 vt.default_blu= [VT]
6008 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6009 Change the default blue palette of the console.
6010 This is a 16-member array composed of values
6013 vt.default_grn= [VT]
6014 Format: <green0>,<green1>,<green2>,...,<green15>
6015 Change the default green palette of the console.
6016 This is a 16-member array composed of values
6019 vt.default_red= [VT]
6020 Format: <red0>,<red1>,<red2>,...,<red15>
6021 Change the default red palette of the console.
6022 This is a 16-member array composed of values
6028 Set system-wide default UTF-8 mode for all tty's.
6029 Default is 1, i.e. UTF-8 mode is enabled for all
6030 newly opened terminals.
6032 vt.global_cursor_default=
6035 Set system-wide default for whether a cursor
6036 is shown on new VTs. Default is -1,
6037 i.e. cursors will be created by default unless
6038 overridden by individual drivers. 0 will hide
6039 cursors, 1 will display them.
6041 vt.italic= [VT] Default color for italic text; 0-15.
6044 vt.underline= [VT] Default color for underlined text; 0-15.
6047 watchdog timers [HW,WDT] For information on watchdog timers,
6048 see Documentation/watchdog/watchdog-parameters.rst
6049 or other driver-specific files in the
6050 Documentation/watchdog/ directory.
6054 Set the hard lockup detector stall duration
6055 threshold in seconds. The soft lockup detector
6056 threshold is set to twice the value. A value of 0
6057 disables both lockup detectors. Default is 10
6060 workqueue.watchdog_thresh=
6061 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6062 warn stall conditions and dump internal state to
6063 help debugging. 0 disables workqueue stall
6064 detection; otherwise, it's the stall threshold
6065 duration in seconds. The default value is 30 and
6066 it can be updated at runtime by writing to the
6067 corresponding sysfs file.
6069 workqueue.disable_numa
6070 By default, all work items queued to unbound
6071 workqueues are affine to the NUMA nodes they're
6072 issued on, which results in better behavior in
6073 general. If NUMA affinity needs to be disabled for
6074 whatever reason, this option can be used. Note
6075 that this also can be controlled per-workqueue for
6076 workqueues visible under /sys/bus/workqueue/.
6078 workqueue.power_efficient
6079 Per-cpu workqueues are generally preferred because
6080 they show better performance thanks to cache
6081 locality; unfortunately, per-cpu workqueues tend to
6082 be more power hungry than unbound workqueues.
6084 Enabling this makes the per-cpu workqueues which
6085 were observed to contribute significantly to power
6086 consumption unbound, leading to measurably lower
6087 power usage at the cost of small performance
6090 The default value of this parameter is determined by
6091 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6093 workqueue.debug_force_rr_cpu
6094 Workqueue used to implicitly guarantee that work
6095 items queued without explicit CPU specified are put
6096 on the local CPU. This guarantee is no longer true
6097 and while local CPU is still preferred work items
6098 may be put on foreign CPUs. This debug option
6099 forces round-robin CPU selection to flush out
6100 usages which depend on the now broken guarantee.
6101 When enabled, memory and cache locality will be
6104 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6105 default x2apic cluster mode on platforms
6108 x86_intel_mid_timer= [X86-32,APBT]
6109 Choose timer option for x86 Intel MID platform.
6110 Two valid options are apbt timer only and lapic timer
6111 plus one apbt timer for broadcast timer.
6112 x86_intel_mid_timer=apbt_only | lapic_and_apbt
6114 xen_512gb_limit [KNL,X86-64,XEN]
6115 Restricts the kernel running paravirtualized under Xen
6116 to use only up to 512 GB of RAM. The reason to do so is
6117 crash analysis tools and Xen tools for doing domain
6118 save/restore/migration must be enabled to handle larger
6121 xen_emul_unplug= [HW,X86,XEN]
6122 Unplug Xen emulated devices
6123 Format: [unplug0,][unplug1]
6124 ide-disks -- unplug primary master IDE devices
6125 aux-ide-disks -- unplug non-primary-master IDE devices
6126 nics -- unplug network devices
6127 all -- unplug all emulated devices (NICs and IDE disks)
6128 unnecessary -- unplugging emulated devices is
6129 unnecessary even if the host did not respond to
6131 never -- do not unplug even if version check succeeds
6133 xen_legacy_crash [X86,XEN]
6134 Crash from Xen panic notifier, without executing late
6135 panic() code such as dumping handler.
6137 xen_nopvspin [X86,XEN]
6138 Disables the qspinlock slowpath using Xen PV optimizations.
6139 This parameter is obsoleted by "nopvspin" parameter, which
6140 has equivalent effect for XEN platform.
6143 Disables the PV optimizations forcing the HVM guest to
6144 run as generic HVM guest with no PV drivers.
6145 This option is obsoleted by the "nopv" option, which
6146 has equivalent effect for XEN platform.
6148 xen_no_vector_callback
6149 [KNL,X86,XEN] Disable the vector callback for Xen
6150 event channel interrupts.
6152 xen_scrub_pages= [XEN]
6153 Boolean option to control scrubbing pages before giving them back
6154 to Xen, for use by other domains. Can be also changed at runtime
6155 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6156 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6158 xen_timer_slop= [X86-64,XEN]
6159 Set the timer slop (in nanoseconds) for the virtual Xen
6160 timers (default is 100000). This adjusts the minimum
6161 delta of virtualized Xen timers, where lower values
6162 improve timer resolution at the expense of processing
6163 more timer interrupts.
6165 xen.balloon_boot_timeout= [XEN]
6166 The time (in seconds) to wait before giving up to boot
6167 in case initial ballooning fails to free enough memory.
6168 Applies only when running as HVM or PVH guest and
6169 started with less memory configured than allowed at
6170 max. Default is 180.
6172 xen.event_eoi_delay= [XEN]
6173 How long to delay EOI handling in case of event
6174 storms (jiffies). Default is 10.
6176 xen.event_loop_timeout= [XEN]
6177 After which time (jiffies) the event handling loop
6178 should start to delay EOI handling. Default is 2.
6180 xen.fifo_events= [XEN]
6181 Boolean parameter to disable using fifo event handling
6182 even if available. Normally fifo event handling is
6183 preferred over the 2-level event handling, as it is
6184 fairer and the number of possible event channels is
6185 much higher. Default is on (use fifo events).
6187 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6188 Disables the PV optimizations forcing the guest to run
6189 as generic guest with no PV drivers. Currently support
6190 XEN HVM, KVM, HYPER_V and VMWARE guest.
6192 nopvspin [X86,XEN,KVM]
6193 Disables the qspinlock slow path using PV optimizations
6194 which allow the hypervisor to 'idle' the guest on lock
6197 xirc2ps_cs= [NET,PCMCIA]
6199 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6202 By default on POWER9 and above, the kernel will
6203 natively use the XIVE interrupt controller. This option
6204 allows the fallback firmware mode to be used:
6206 off Fallback to firmware control of XIVE interrupt
6207 controller on both pseries and powernv
6208 platforms. Only useful on POWER9 and above.
6210 xhci-hcd.quirks [USB,KNL]
6211 A hex value specifying bitmask with supplemental xhci
6212 host controller quirks. Meaning of each bit can be
6213 consulted in header drivers/usb/host/xhci.h.
6216 Format: { early | on | rw | ro | off }
6217 Controls if xmon debugger is enabled. Default is off.
6218 Passing only "xmon" is equivalent to "xmon=early".
6219 early Call xmon as early as possible on boot; xmon
6220 debugger is called from setup_arch().
6221 on xmon debugger hooks will be installed so xmon
6222 is only called on a kernel crash. Default mode,
6223 i.e. either "ro" or "rw" mode, is controlled
6224 with CONFIG_XMON_DEFAULT_RO_MODE.
6225 rw xmon debugger hooks will be installed so xmon
6226 is called only on a kernel crash, mode is write,
6227 meaning SPR registers, memory and, other data
6228 can be written using xmon commands.
6229 ro same as "rw" option above but SPR registers,
6230 memory, and other data can't be written using
6232 off xmon is disabled.