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_EVENTS
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 AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
116 Format: <byte> or <bitmap-list>
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
163 acpi_osi=!! # enable all built-in OS vendor
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
229 old_ordering, nonvs, sci_force_enable, nobl }
230 See Documentation/power/video.rst for information on
232 s3_beep is for debugging; it makes the PC's speaker beep
233 as soon as the kernel's real-mode entry point is called.
234 s4_nohwsig prevents ACPI hardware signature from being
235 used during resume from hibernation.
236 old_ordering causes the ACPI 1.0 ordering of the _PTS
237 control method, with respect to putting devices into
238 low power states, to be enforced (the ACPI 2.0 ordering
239 of _PTS is used by default).
240 nonvs prevents the kernel from saving/restoring the
241 ACPI NVS memory during suspend/hibernation and resume.
242 sci_force_enable causes the kernel to set SCI_EN directly
243 on resume from S1/S3 (which is against the ACPI spec,
244 but some broken systems don't work without it).
245 nobl causes the internal blacklist of systems known to
246 behave incorrectly in some ways with respect to system
247 suspend and resume to be ignored (use wisely).
249 acpi_use_timer_override [HW,ACPI]
250 Use timer override. For some broken Nvidia NF5 boards
251 that require a timer override, but don't have HPET
253 add_efi_memmap [EFI; X86] Include EFI memory map in
254 kernel's map of available physical RAM.
257 { off | try_unsupported }
258 off: disable AGP support
259 try_unsupported: try to drive unsupported chipsets
260 (may crash computer or cause data corruption)
263 See Documentation/sound/alsa-configuration.rst
266 Allow the default userspace alignment fault handler
267 behaviour to be specified. Bit 0 enables warnings,
268 bit 1 enables fixups, and bit 2 sends a segfault.
270 align_va_addr= [X86-64]
271 Align virtual addresses by clearing slice [14:12] when
272 allocating a VMA at process creation time. This option
273 gives you up to 3% performance improvement on AMD F15h
274 machines (where it is enabled by default) for a
275 CPU-intensive style benchmark, and it can vary highly in
276 a microbenchmark depending on workload and compiler.
278 32: only for 32-bit processes
279 64: only for 64-bit processes
280 on: enable for both 32- and 64-bit processes
281 off: disable for both 32- and 64-bit processes
283 alloc_snapshot [FTRACE]
284 Allocate the ftrace snapshot buffer on boot up when the
285 main buffer is allocated. This is handy if debugging
286 and you need to use tracing_snapshot() on boot up, and
287 do not want to use tracing_snapshot_alloc() as it needs
288 to be done where GFP_KERNEL allocations are allowed.
290 allow_mismatched_32bit_el0 [ARM64]
291 Allow execve() of 32-bit applications and setting of the
292 PER_LINUX32 personality on systems where only a strict
293 subset of the CPUs support 32-bit EL0. When this
294 parameter is present, the set of CPUs supporting 32-bit
295 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
296 and hot-unplug operations may be restricted.
298 See Documentation/arm64/asymmetric-32bit.rst for more
301 amd_iommu= [HW,X86-64]
302 Pass parameters to the AMD IOMMU driver in the system.
304 fullflush - Deprecated, equivalent to iommu.strict=1
305 off - do not initialize any AMD IOMMU found in
307 force_isolation - Force device isolation for all
308 devices. The IOMMU driver is not
309 allowed anymore to lift isolation
310 requirements as needed. This option
311 does not override iommu=pt
312 force_enable - Force enable the IOMMU on platforms known
313 to be buggy with IOMMU enabled. Use this
316 amd_iommu_dump= [HW,X86-64]
317 Enable AMD IOMMU driver option to dump the ACPI table
318 for AMD IOMMU. With this option enabled, AMD IOMMU
319 driver will print ACPI tables for AMD IOMMU during
320 IOMMU initialization.
322 amd_iommu_intr= [HW,X86-64]
323 Specifies one of the following AMD IOMMU interrupt
325 legacy - Use legacy interrupt remapping mode.
326 vapic - Use virtual APIC mode, which allows IOMMU
327 to inject interrupts directly into guest.
328 This mode requires kvm-amd.avic=1.
329 (Default when IOMMU HW support is present.)
331 amijoy.map= [HW,JOY] Amiga joystick support
332 Map of devices attached to JOY0DAT and JOY1DAT
334 See also Documentation/input/joydev/joystick.rst
336 analog.map= [HW,JOY] Analog joystick and gamepad support
337 Specifies type or capabilities of an analog joystick
338 connected to one of 16 gameports
339 Format: <type1>,<type2>,..<type16>
342 Power management functions (SPARCstation-4/5 + deriv.)
344 Disable APC CPU standby support. SPARCstation-Fox does
345 not play well with APC CPU idle - disable it if you have
346 APC and your system crashes randomly.
348 apic= [APIC,X86] Advanced Programmable Interrupt Controller
349 Change the output verbosity while booting
350 Format: { quiet (default) | verbose | debug }
351 Change the amount of debugging information output
352 when initialising the APIC and IO-APIC components.
353 For X86-32, this can also be used to specify an APIC
355 Format: apic=driver_name
356 Examples: apic=bigsmp
358 apic_extnmi= [APIC,X86] External NMI delivery setting
359 Format: { bsp (default) | all | none }
360 bsp: External NMI is delivered only to CPU 0
361 all: External NMIs are broadcast to all CPUs as a
363 none: External NMI is masked for all CPUs. This is
364 useful so that a dump capture kernel won't be
368 See Documentation/networking/ipv6.rst.
370 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
371 Limit apic dumping. The parameter defines the maximal
372 number of local apics being dumped. Also it is possible
373 to set it to "all" by meaning -- no limit here.
374 Format: { 1 (default) | 2 | ... | all }.
375 The parameter valid if only apic=debug or
376 apic=verbose is specified.
377 Example: apic=debug show_lapic=all
379 apm= [APM] Advanced Power Management
380 See header of arch/x86/kernel/apm_32.c.
382 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
383 Format: <io>,<irq>,<nodeID>
385 arm64.nobti [ARM64] Unconditionally disable Branch Target
386 Identification support
388 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
391 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
396 atarimouse= [HW,MOUSE] Atari Mouse
398 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
399 EzKey and similar keyboards
401 atkbd.reset= [HW] Reset keyboard during initialization
403 atkbd.set= [HW] Select keyboard code set
404 Format: <int> (2 = AT (default), 3 = PS/2)
406 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
409 atkbd.softraw= [HW] Choose between synthetic and real raw mode
410 Format: <bool> (0 = real, 1 = synthetic (default))
412 atkbd.softrepeat= [HW]
413 Use software keyboard repeat
415 audit= [KNL] Enable the audit sub-system
416 Format: { "0" | "1" | "off" | "on" }
417 0 | off - kernel audit is disabled and can not be
418 enabled until the next reboot
419 unset - kernel audit is initialized but disabled and
420 will be fully enabled by the userspace auditd.
421 1 | on - kernel audit is initialized and partially
422 enabled, storing at most audit_backlog_limit
423 messages in RAM until it is fully enabled by the
427 audit_backlog_limit= [KNL] Set the audit queue size limit.
428 Format: <int> (must be >=0)
431 bau= [X86_UV] Enable the BAU on SGI UV. The default
432 behavior is to disable the BAU (i.e. bau=0).
433 Format: { "0" | "1" }
436 unset - Disable the BAU.
438 baycom_epp= [HW,AX25]
441 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
443 See header of drivers/net/hamradio/baycom_par.c.
445 baycom_ser_fdx= [HW,AX25]
446 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
447 Format: <io>,<irq>,<mode>[,<baud>]
448 See header of drivers/net/hamradio/baycom_ser_fdx.c.
450 baycom_ser_hdx= [HW,AX25]
451 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
452 Format: <io>,<irq>,<mode>
453 See header of drivers/net/hamradio/baycom_ser_hdx.c.
455 blkdevparts= Manual partition parsing of block device(s) for
456 embedded devices based on command line input.
457 See Documentation/block/cmdline-partition.rst
459 boot_delay= Milliseconds to delay each printk during boot.
460 Values larger than 10 seconds (10000) are changed to
465 Extended command line options can be added to an initrd
466 and this will cause the kernel to look for it.
468 See Documentation/admin-guide/bootconfig.rst
471 Disable BERT OS support on buggy BIOSes.
473 bgrt_disable [ACPI][X86]
474 Disable BGRT to avoid flickering OEM logo.
476 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
477 bttv.radio= Most important insmod options are available as
479 bttv.pll= See Documentation/admin-guide/media/bttv.rst
482 bulk_remove=off [PPC] This parameter disables the use of the pSeries
483 firmware feature for flushing multiple hpte entries
486 c101= [NET] Moxa C101 synchronous serial card
488 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
489 Sometimes CPU hardware bugs make them report the cache
490 size incorrectly. The kernel will attempt work arounds
491 to fix known problems, but for some CPUs it is not
492 possible to determine what the correct size should be.
493 This option provides an override for these situations.
496 [NET] Specifies amount of time (in seconds) that
497 the kernel should wait for a network carrier. By default
498 it waits 120 seconds.
500 ca_keys= [KEYS] This parameter identifies a specific key(s) on
501 the system trusted keyring to be used for certificate
503 format: { id:<keyid> | builtin }
505 cca= [MIPS] Override the kernel pages' cache coherency
506 algorithm. Accepted values range from 0 to 7
507 inclusive. See arch/mips/include/asm/pgtable-bits.h
508 for platform specific values (SB1, Loongson3 and
511 ccw_timeout_log [S390]
512 See Documentation/s390/common_io.rst for details.
514 cgroup_disable= [KNL] Disable a particular controller or optional feature
515 Format: {name of the controller(s) or feature(s) to disable}
516 The effects of cgroup_disable=foo are:
517 - foo isn't auto-mounted if you mount all cgroups in
519 - foo isn't visible as an individually mountable
521 - if foo is an optional feature then the feature is
522 disabled and corresponding cgroup files are not
524 {Currently only "memory" controller deal with this and
525 cut the overhead, others just disable the usage. So
526 only cgroup_disable=memory is actually worthy}
527 Specifying "pressure" disables per-cgroup pressure
528 stall information accounting feature
530 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
531 Format: { { controller | "all" | "named" }
532 [,{ controller | "all" | "named" }...] }
533 Like cgroup_disable, but only applies to cgroup v1;
534 the blacklisted controllers remain available in cgroup2.
535 "all" blacklists all controllers and "named" disables
536 named mounts. Specifying both "all" and "named" disables
539 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
541 nosocket -- Disable socket memory accounting.
542 nokmem -- Disable kernel memory accounting.
544 checkreqprot [SELINUX] Set initial checkreqprot flag value.
545 Format: { "0" | "1" }
546 See security/selinux/Kconfig help text.
547 0 -- check protection applied by kernel (includes
548 any implied execute protection).
549 1 -- check protection requested by application.
550 Default value is set via a kernel config option.
551 Value can be changed at runtime via
552 /sys/fs/selinux/checkreqprot.
553 Setting checkreqprot to 1 is deprecated.
556 See Documentation/s390/common_io.rst for details.
559 Prevents the clock framework from automatically gating
560 clocks that have not been explicitly enabled by a Linux
561 device driver but are enabled in hardware at reset or
562 by the bootloader/firmware. Note that this does not
563 force such clocks to be always-on nor does it reserve
564 those clocks in any way. This parameter is useful for
565 debug and development, but should not be needed on a
566 platform with proper driver support. For more
567 information, see Documentation/driver-api/clk.rst.
569 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
571 Forces specified clocksource (if available) to be used
572 when calculating gettimeofday(). If specified
573 clocksource is not available, it defaults to PIT.
574 Format: { pit | tsc | cyclone | pmtmr }
576 clocksource= Override the default clocksource
578 Override the default clocksource and use the clocksource
579 with the name specified.
580 Some clocksource names to choose from, depending on
582 [all] jiffies (this is the base, fallback clocksource)
584 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
585 pxa_timer,timer3,32k_counter,timer0_1
586 [X86-32] pit,hpet,tsc;
587 scx200_hrt on Geode; cyclone on IBM x440
595 clocksource.arm_arch_timer.evtstrm=
598 Enable/disable the eventstream feature of the ARM
599 architected timer so that code using WFE-based polling
600 loops can be debugged more effectively on production
603 clocksource.max_cswd_read_retries= [KNL]
604 Number of clocksource_watchdog() retries due to
605 external delays before the clock will be marked
606 unstable. Defaults to three retries, that is,
607 four attempts to read the clock under test.
609 clocksource.verify_n_cpus= [KNL]
610 Limit the number of CPUs checked for clocksources
611 marked with CLOCK_SOURCE_VERIFY_PERCPU that
612 are marked unstable due to excessive skew.
613 A negative value says to check all CPUs, while
614 zero says not to check any. Values larger than
615 nr_cpu_ids are silently truncated to nr_cpu_ids.
616 The actual CPUs are chosen randomly, with
617 no replacement if the same CPU is chosen twice.
619 clocksource-wdtest.holdoff= [KNL]
620 Set the time in seconds that the clocksource
621 watchdog test waits before commencing its tests.
622 Defaults to zero when built as a module and to
623 10 seconds when built into the kernel.
625 clearcpuid=BITNUM[,BITNUM...] [X86]
626 Disable CPUID feature X for the kernel. See
627 arch/x86/include/asm/cpufeatures.h for the valid bit
628 numbers. Note the Linux specific bits are not necessarily
629 stable over kernel options, but the vendor specific
631 Also note that user programs calling CPUID directly
632 or using the feature without checking anything
633 will still see it. This just prevents it from
634 being used by the kernel or shown in /proc/cpuinfo.
635 Also note the kernel might malfunction if you disable
638 cma=nn[MG]@[start[MG][-end[MG]]]
640 Sets the size of kernel global memory area for
641 contiguous memory allocations and optionally the
642 placement constraint by the physical address range of
643 memory allocations. A value of 0 disables CMA
644 altogether. For more information, see
645 kernel/dma/contiguous.c
649 Sets the size of kernel per-numa memory area for
650 contiguous memory allocations. A value of 0 disables
651 per-numa CMA altogether. And If this option is not
652 specificed, the default value is 0.
653 With per-numa CMA enabled, DMA users on node nid will
654 first try to allocate buffer from the pernuma area
655 which is located in node nid, if the allocation fails,
656 they will fallback to the global default memory area.
658 cmo_free_hint= [PPC] Format: { yes | no }
659 Specify whether pages are marked as being inactive
660 when they are freed. This is used in CMO environments
661 to determine OS memory pressure for page stealing by
665 coherent_pool=nn[KMG] [ARM,KNL]
666 Sets the size of memory pool for coherent, atomic dma
667 allocations, by default set to 256K.
669 com20020= [HW,NET] ARCnet - COM20020 chipset
671 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
673 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
677 ARCnet - COM90xx chipset (memory-mapped buffers)
678 Format: <io>[,<irq>[,<memstart>]]
680 condev= [HW,S390] console device
683 console= [KNL] Output console device and options.
685 tty<n> Use the virtual console device <n>.
689 Use the specified serial port. The options are of
690 the form "bbbbpnf", where "bbbb" is the baud rate,
691 "p" is parity ("n", "o", or "e"), "n" is number of
692 bits, and "f" is flow control ("r" for RTS or
693 omit it). Default is "9600n8".
695 See Documentation/admin-guide/serial-console.rst for more
697 Documentation/networking/netconsole.rst for an
700 uart[8250],io,<addr>[,options]
701 uart[8250],mmio,<addr>[,options]
702 uart[8250],mmio16,<addr>[,options]
703 uart[8250],mmio32,<addr>[,options]
704 uart[8250],0x<addr>[,options]
705 Start an early, polled-mode console on the 8250/16550
706 UART at the specified I/O port or MMIO address,
707 switching to the matching ttyS device later.
708 MMIO inter-register address stride is either 8-bit
709 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
710 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
711 to be equivalent to 'mmio'. 'options' are specified in
712 the same format described for ttyS above; if unspecified,
713 the h/w is not re-initialized.
715 hvc<n> Use the hypervisor console device <n>. This is for
716 both Xen and PowerPC hypervisors.
718 If the device connected to the port is not a TTY but a braille
719 device, prepend "brl," before the device type, for instance
721 For now, only VisioBraille is supported.
724 [KNL] Change console messages format
726 By default we print messages on consoles in
727 "[time stamp] text\n" format (time stamp may not be
728 printed, depending on CONFIG_PRINTK_TIME or
729 `printk_time' param).
731 Switch to syslog format: "<%u>[time stamp] text\n"
732 IOW, each message will have a facility and loglevel
733 prefix. The format is similar to one used by syslog()
734 syscall, or to executing "dmesg -S --raw" or to reading
737 consoleblank= [KNL] The console blank (screen saver) timeout in
738 seconds. A value of 0 disables the blank timer.
742 [KNL] Change the default value for
743 /proc/<pid>/coredump_filter.
744 See also Documentation/filesystems/proc.rst.
746 coresight_cpu_debug.enable
749 Enable/disable the CPU sampling based debugging.
750 0: default value, disable debugging
751 1: enable debugging at boot time
753 cpuidle.off=1 [CPU_IDLE]
754 disable the cpuidle sub-system
757 [CPU_IDLE] Name of the cpuidle governor to use.
759 cpufreq.off=1 [CPU_FREQ]
760 disable the cpufreq sub-system
762 cpufreq.default_governor=
763 [CPU_FREQ] Name of the default cpufreq governor or
764 policy to use. This governor must be registered in the
765 kernel before the cpufreq driver probes.
768 [X86] Delay for N microsec between assert and de-assert
769 of APIC INIT to start processors. This delay occurs
770 on every CPU online, such as boot, and resume from suspend.
773 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
775 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
777 crashkernel=size[KMG][@offset[KMG]]
778 [KNL] Using kexec, Linux can switch to a 'crash kernel'
779 upon panic. This parameter reserves the physical
780 memory region [offset, offset + size] for that kernel
781 image. If '@offset' is omitted, then a suitable offset
782 is selected automatically.
783 [KNL, X86-64] Select a region under 4G first, and
784 fall back to reserve region above 4G when '@offset'
785 hasn't been specified.
786 See Documentation/admin-guide/kdump/kdump.rst for further details.
788 crashkernel=range1:size1[,range2:size2,...][@offset]
789 [KNL] Same as above, but depends on the memory
790 in the running system. The syntax of range is
791 start-[end] where start and end are both
792 a memory unit (amount[KMG]). See also
793 Documentation/admin-guide/kdump/kdump.rst for an example.
795 crashkernel=size[KMG],high
796 [KNL, X86-64] range could be above 4G. Allow kernel
797 to allocate physical memory region from top, so could
798 be above 4G if system have more than 4G ram installed.
799 Otherwise memory region will be allocated below 4G, if
801 It will be ignored if crashkernel=X is specified.
802 crashkernel=size[KMG],low
803 [KNL, X86-64] range under 4G. When crashkernel=X,high
804 is passed, kernel could allocate physical memory region
805 above 4G, that cause second kernel crash on system
806 that require some amount of low memory, e.g. swiotlb
807 requires at least 64M+32K low memory, also enough extra
808 low memory is needed to make sure DMA buffers for 32-bit
809 devices won't run out. Kernel would try to allocate at
810 at least 256M below 4G automatically.
811 This one let user to specify own low range under 4G
812 for second kernel instead.
813 0: to disable low allocation.
814 It will be ignored when crashkernel=X,high is not used
815 or memory reserved is below 4G.
818 [KNL] Disable crypto self-tests
823 cs89x0_media= [HW,NET]
824 Format: { rj45 | aui | bnc }
826 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
827 handling. When switched on, additional debug data is
828 printed to the console in case a hanging CPU is
829 detected, and that CPU is pinged again in order to try
830 to resolve the hang situation.
831 0: disable csdlock debugging (default)
832 1: enable basic csdlock debugging (minor impact)
833 ext: enable extended csdlock debugging (more impact,
837 See header of drivers/s390/block/dasd_devmap.c.
839 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
840 (one device per port)
841 Format: <port#>,<type>
842 See also Documentation/input/devices/joystick-parport.rst
844 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
846 Documentation/admin-guide/dynamic-debug-howto.rst for
847 details. Deprecated, see dyndbg.
849 debug [KNL] Enable kernel debugging (events log level).
852 [KNL] Enable printing [hashed] pointers early in the
853 boot sequence. If enabled, we use a weak hash instead
854 of siphash to hash pointers. Use this option if you are
855 seeing instances of '(___ptrval___)') and need to see a
856 value (hashed pointer) instead. Cryptographically
857 insecure, please do not use on production kernels.
860 [KNL] verbose locking self-tests
862 Print debugging info while doing the locking API
864 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
865 (no extra messages), setting it to -1 (all bits set)
866 will print _a_lot_ more information - normally only
867 useful to lockdep developers.
869 debug_objects [KNL] Enable object debugging
872 [KNL] Disable object debugging
874 debug_guardpage_minorder=
875 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
876 parameter allows control of the order of pages that will
877 be intentionally kept free (and hence protected) by the
878 buddy allocator. Bigger value increase the probability
879 of catching random memory corruption, but reduce the
880 amount of memory for normal system use. The maximum
881 possible value is MAX_ORDER/2. Setting this parameter
882 to 1 or 2 should be enough to identify most random
883 memory corruption problems caused by bugs in kernel or
884 driver code when a CPU writes to (or reads from) a
885 random memory location. Note that there exists a class
886 of memory corruptions problems caused by buggy H/W or
887 F/W or by drivers badly programing DMA (basically when
888 memory is written at bus level and the CPU MMU is
889 bypassed) which are not detectable by
890 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
891 tracking down these problems.
894 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
895 enables the feature at boot time. By default, it is
896 disabled and the system will work mostly the same as a
897 kernel built without CONFIG_DEBUG_PAGEALLOC.
898 Note: to get most of debug_pagealloc error reports, it's
899 useful to also enable the page_owner functionality.
900 on: enable the feature
902 debugfs= [KNL] This parameter enables what is exposed to userspace
903 and debugfs internal clients.
904 Format: { on, no-mount, off }
905 on: All functions are enabled.
907 Filesystem is not registered but kernel clients can
908 access APIs and a crashkernel can be used to read
909 its content. There is nothing to mount.
910 off: Filesystem is not registered and clients
911 get a -EPERM as result when trying to register files
912 or directories within debugfs.
913 This is equivalent of the runtime functionality if
914 debugfs was not enabled in the kernel at all.
915 Default value is set in build-time with a kernel configuration.
917 debugpat [X86] Enable PAT debugging
919 decnet.addr= [HW,NET]
920 Format: <area>[,<node>]
921 See also Documentation/networking/decnet.rst.
924 [HW] The size of the default HugeTLB page. This is
925 the size represented by the legacy /proc/ hugepages
926 APIs. In addition, this is the default hugetlb size
927 used for shmget(), mmap() and mounting hugetlbfs
928 filesystems. If not specified, defaults to the
929 architecture's default huge page size. Huge page
930 sizes are architecture dependent. See also
931 Documentation/admin-guide/mm/hugetlbpage.rst.
934 deferred_probe_timeout=
935 [KNL] Debugging option to set a timeout in seconds for
936 deferred probe to give up waiting on dependencies to
937 probe. Only specific dependencies (subsystems or
938 drivers) that have opted in will be ignored. A timeout of 0
939 will timeout at the end of initcalls. This option will also
940 dump out devices still on the deferred probe list after
944 Format: { on | off | def_only | inf_only | always }
945 on: s390 zlib hardware support for compression on
946 level 1 and decompression (default)
947 off: No s390 zlib hardware support
948 def_only: s390 zlib hardware support for deflate
949 only (compression on level 1)
950 inf_only: s390 zlib hardware support for inflate
952 always: Same as 'on' but ignores the selected compression
953 level always using hardware support (used for debugging)
956 Set number of hash buckets for dentry cache.
958 disable_1tb_segments [PPC]
959 Disables the use of 1TB hash page table segments. This
960 causes the kernel to fall back to 256MB segments which
961 can be useful when debugging issues that require an SLB
965 Limits the number of kernel SLB entries, and flushes
966 them frequently to increase the rate of SLB faults
970 See Documentation/networking/ipv6.rst.
973 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
974 hardening is enabled for this boot. Hardened
975 usercopy checking is used to protect the kernel
976 from reading or writing beyond known memory
977 allocation boundaries as a proactive defense
978 against bounds-checking flaws in the kernel's
979 copy_to_user()/copy_from_user() interface.
980 on Perform hardened usercopy checks (default).
981 off Disable hardened usercopy checks.
984 Disable RADIX MMU mode on POWER9
986 radix_hcall_invalidate=on [PPC/PSERIES]
987 Disable RADIX GTSE feature and use hcall for TLB
991 Disable TLBIE instruction. Currently does not work
992 with KVM, with HASH MMU, or with coherent accelerators.
994 disable_cpu_apicid= [X86,APIC,SMP]
996 The number of initial APIC ID for the
997 corresponding CPU to be disabled at boot,
998 mostly used for the kdump 2nd kernel to
999 disable BSP to wake up multiple CPUs without
1000 causing system reset or hang due to sending
1001 INIT from AP to BSP.
1003 disable_ddw [PPC/PSERIES]
1004 Disable Dynamic DMA Window support. Use this
1005 to workaround buggy firmware.
1007 disable_ipv6= [IPV6]
1008 See Documentation/networking/ipv6.rst.
1010 disable_mtrr_cleanup [X86]
1011 The kernel tries to adjust MTRR layout from continuous
1012 to discrete, to make X server driver able to add WB
1013 entry later. This parameter disables that.
1015 disable_mtrr_trim [X86, Intel and AMD only]
1016 By default the kernel will trim any uncacheable
1017 memory out of your available memory pool based on
1018 MTRR settings. This parameter disables that behavior,
1019 possibly causing your machine to run very slowly.
1021 disable_timer_pin_1 [X86]
1022 Disable PIN 1 of APIC timer
1023 Can be useful to work around chipset bugs.
1025 dis_ucode_ldr [X86] Disable the microcode loader.
1027 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1028 this option disables the debugging code at boot.
1030 dma_debug_entries=<number>
1031 This option allows to tune the number of preallocated
1032 entries for DMA-API debugging code. One entry is
1033 required per DMA-API allocation. Use this if the
1034 DMA-API debugging code disables itself because the
1035 architectural default is too low.
1037 dma_debug_driver=<driver_name>
1038 With this option the DMA-API debugging driver
1039 filter feature can be enabled at boot time. Just
1040 pass the driver to filter for as the parameter.
1041 The filter can be disabled or changed to another
1042 driver later using sysfs.
1044 driver_async_probe= [KNL]
1045 List of driver names to be probed asynchronously.
1046 Format: <driver_name1>,<driver_name2>...
1048 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1049 Broken monitors, graphic adapters, KVMs and EDIDless
1050 panels may send no or incorrect EDID data sets.
1051 This parameter allows to specify an EDID data sets
1052 in the /lib/firmware directory that are used instead.
1053 Generic built-in EDID data sets are used, if one of
1054 edid/1024x768.bin, edid/1280x1024.bin,
1055 edid/1680x1050.bin, or edid/1920x1080.bin is given
1056 and no file with the same name exists. Details and
1057 instructions how to build your own EDID data are
1058 available in Documentation/admin-guide/edid.rst. An EDID
1059 data set will only be used for a particular connector,
1060 if its name and a colon are prepended to the EDID
1061 name. Each connector may use a unique EDID data
1062 set by separating the files with a comma. An EDID
1063 data set with no connector name will be used for
1064 any connectors not explicitly specified.
1069 Format: {"off" | "known"}
1070 Control how the dt_cpu_ftrs device-tree binding is
1071 used for CPU feature discovery and setup (if it
1073 off: Do not use it, fall back to legacy cpu table.
1074 known: Do not pass through unknown features to guests
1075 or userspace, only those that the kernel is aware of.
1077 dump_apple_properties [X86]
1078 Dump name and content of EFI device properties on
1079 x86 Macs. Useful for driver authors to determine
1080 what data is available or for reverse-engineering.
1082 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1083 <module>.dyndbg[="val"]
1084 Enable debug messages at boot time. See
1085 Documentation/admin-guide/dynamic-debug-howto.rst
1088 nopku [X86] Disable Memory Protection Keys CPU feature found
1091 <module>.async_probe [KNL]
1092 Enable asynchronous probe on this module.
1094 early_ioremap_debug [KNL]
1095 Enable debug messages in early_ioremap support. This
1096 is useful for tracking down temporary early mappings
1097 which are not unmapped.
1099 earlycon= [KNL] Output early console device and options.
1101 When used with no options, the early console is
1102 determined by stdout-path property in device tree's
1103 chosen node or the ACPI SPCR table if supported by
1106 cdns,<addr>[,options]
1107 Start an early, polled-mode console on a Cadence
1108 (xuartps) serial port at the specified address. Only
1109 supported option is baud rate. If baud rate is not
1110 specified, the serial port must already be setup and
1113 uart[8250],io,<addr>[,options]
1114 uart[8250],mmio,<addr>[,options]
1115 uart[8250],mmio32,<addr>[,options]
1116 uart[8250],mmio32be,<addr>[,options]
1117 uart[8250],0x<addr>[,options]
1118 Start an early, polled-mode console on the 8250/16550
1119 UART at the specified I/O port or MMIO address.
1120 MMIO inter-register address stride is either 8-bit
1121 (mmio) or 32-bit (mmio32 or mmio32be).
1122 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1123 to be equivalent to 'mmio'. 'options' are specified
1124 in the same format described for "console=ttyS<n>"; if
1125 unspecified, the h/w is not initialized.
1129 Start an early, polled-mode console on a pl011 serial
1130 port at the specified address. The pl011 serial port
1131 must already be setup and configured. Options are not
1132 yet supported. If 'mmio32' is specified, then only
1133 the driver will use only 32-bit accessors to read/write
1134 the device registers.
1137 Start an early console on a litex serial port at the
1138 specified address. The serial port must already be
1139 setup and configured. Options are not yet supported.
1142 Start an early, polled-mode console on a meson serial
1143 port at the specified address. The serial port must
1144 already be setup and configured. Options are not yet
1148 Start an early, polled-mode console on an msm serial
1149 port at the specified address. The serial port
1150 must already be setup and configured. Options are not
1153 msm_serial_dm,<addr>
1154 Start an early, polled-mode console on an msm serial
1155 dm port at the specified address. The serial port
1156 must already be setup and configured. Options are not
1160 Start an early, polled-mode console on a serial port
1161 of an Actions Semi SoC, such as S500 or S900, at the
1162 specified address. The serial port must already be
1163 setup and configured. Options are not yet supported.
1166 Start an early, polled-mode console on a serial port
1167 of an RDA Micro SoC, such as RDA8810PL, at the
1168 specified address. The serial port must already be
1169 setup and configured. Options are not yet supported.
1172 Use RISC-V SBI (Supervisor Binary Interface) for early
1175 smh Use ARM semihosting calls for early console.
1183 Use early console provided by serial driver available
1184 on Samsung SoCs, requires selecting proper type and
1185 a correct base address of the selected UART port. The
1186 serial port must already be setup and configured.
1187 Options are not yet supported.
1190 Start an early, polled-mode console on a lantiq serial
1191 (lqasc) port at the specified address. The serial port
1192 must already be setup and configured. Options are not
1197 Use early console provided by Freescale LP UART driver
1198 found on Freescale Vybrid and QorIQ LS1021A processors.
1199 A valid base address must be provided, and the serial
1200 port must already be setup and configured.
1204 Start an early, polled-mode, output-only console on the
1205 Freescale i.MX UART at the specified address. The UART
1206 must already be setup and configured.
1209 Start an early, polled-mode console on the
1210 Armada 3700 serial port at the specified
1211 address. The serial port must already be setup
1212 and configured. Options are not yet supported.
1215 Start an early, polled-mode console on a Qualcomm
1216 Generic Interface (GENI) based serial port at the
1217 specified address. The serial port must already be
1218 setup and configured. Options are not yet supported.
1221 Start an early, unaccelerated console on the EFI
1222 memory mapped framebuffer (if available). On cache
1223 coherent non-x86 systems that use system memory for
1224 the framebuffer, pass the 'ram' option so that it is
1225 mapped with the correct attributes.
1228 Use early console provided by Freescale LINFlexD UART
1229 serial driver for NXP S32V234 SoCs. A valid base
1230 address must be provided, and the serial port must
1231 already be setup and configured.
1233 earlyprintk= [X86,SH,ARM,M68k,S390]
1237 earlyprintk=serial[,ttySn[,baudrate]]
1238 earlyprintk=serial[,0x...[,baudrate]]
1239 earlyprintk=ttySn[,baudrate]
1240 earlyprintk=dbgp[debugController#]
1241 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1242 earlyprintk=xdbc[xhciController#]
1244 earlyprintk is useful when the kernel crashes before
1245 the normal console is initialized. It is not enabled by
1246 default because it has some cosmetic problems.
1248 Append ",keep" to not disable it when the real console
1251 Only one of vga, efi, serial, or usb debug port can
1254 Currently only ttyS0 and ttyS1 may be specified by
1255 name. Other I/O ports may be explicitly specified
1256 on some architectures (x86 and arm at least) by
1257 replacing ttySn with an I/O port address, like this:
1258 earlyprintk=serial,0x1008,115200
1259 You can find the port for a given device in
1260 /proc/tty/driver/serial:
1261 2: uart:ST16650V2 port:00001008 irq:18 ...
1263 Interaction with the standard serial driver is not
1266 The VGA and EFI output is eventually overwritten by
1269 The xen option can only be used in Xen domains.
1271 The sclp output can only be used on s390.
1273 The optional "force" to "pciserial" enables use of a
1274 PCI device even when its classcode is not of the
1277 edac_report= [HW,EDAC] Control how to report EDAC event
1278 Format: {"on" | "off" | "force"}
1279 on: enable EDAC to report H/W event. May be overridden
1280 by other higher priority error reporting module.
1281 off: disable H/W event reporting through EDAC.
1282 force: enforce the use of EDAC to report H/W event.
1285 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1288 This is designed to be used in conjunction with
1289 the boot argument: earlyprintk=vga
1291 This parameter works in place of the kgdboc parameter
1292 but can only be used if the backing tty is available
1293 very early in the boot process. For early debugging
1294 via a serial port see kgdboc_earlycon instead.
1297 Format: {"off" | "on" | "skip[mbr]"}
1300 Format: { "debug", "disable_early_pci_dma",
1301 "nochunk", "noruntime", "nosoftreserve",
1302 "novamap", "no_disable_early_pci_dma" }
1303 debug: enable misc debug output.
1304 disable_early_pci_dma: disable the busmaster bit on all
1305 PCI bridges while in the EFI boot stub.
1306 nochunk: disable reading files in "chunks" in the EFI
1307 boot stub, as chunking can cause problems with some
1308 firmware implementations.
1309 noruntime : disable EFI runtime services support
1310 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1311 attribute may cause the kernel to reserve the
1312 memory range for a memory mapping driver to
1313 claim. Specify efi=nosoftreserve to disable this
1314 reservation and treat the memory by its base type
1315 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1316 novamap: do not call SetVirtualAddressMap().
1317 no_disable_early_pci_dma: Leave the busmaster bit set
1318 on all PCI bridges while in the EFI boot stub
1320 efi_no_storage_paranoia [EFI; X86]
1321 Using this parameter you can use more than 50% of
1322 your efi variable storage. Use this parameter only if
1323 you are really sure that your UEFI does sane gc and
1324 fulfills the spec otherwise your board may brick.
1326 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1327 Add arbitrary attribute to specific memory range by
1328 updating original EFI memory map.
1329 Region of memory which aa attribute is added to is
1332 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1333 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1334 attribute is added to range 0x100000000-0x180000000 and
1335 0x10a0000000-0x1120000000.
1337 If efi_fake_mem=8G@9G:0x40000 is specified, the
1338 EFI_MEMORY_SP(0x40000) attribute is added to
1339 range 0x240000000-0x43fffffff.
1341 Using this parameter you can do debugging of EFI memmap
1342 related features. For example, you can do debugging of
1343 Address Range Mirroring feature even if your box
1344 doesn't support it, or mark specific memory as
1347 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1348 that is to be dynamically loaded by Linux. If there are
1349 multiple variables with the same name but with different
1350 vendor GUIDs, all of them will be loaded. See
1351 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1354 eisa_irq_edge= [PARISC,HW]
1355 See header of drivers/parisc/eisa.c.
1358 See comment before function elanfreq_setup() in
1359 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1361 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1362 Specifies physical address of start of kernel core
1363 image elf header and optionally the size. Generally
1364 kexec loader will pass this option to capture kernel.
1365 See Documentation/admin-guide/kdump/kdump.rst for details.
1367 enable_mtrr_cleanup [X86]
1368 The kernel tries to adjust MTRR layout from continuous
1369 to discrete, to make X server driver able to add WB
1370 entry later. This parameter enables that.
1372 enable_timer_pin_1 [X86]
1373 Enable PIN 1 of APIC timer
1374 Can be useful to work around chipset bugs
1375 (in particular on some ATI chipsets).
1376 The kernel tries to set a reasonable default.
1378 enforcing [SELINUX] Set initial enforcing status.
1380 See security/selinux/Kconfig help text.
1381 0 -- permissive (log only, no denials).
1382 1 -- enforcing (deny and log).
1384 Value can be changed at runtime via
1385 /sys/fs/selinux/enforce.
1388 Disable Error Record Serialization Table (ERST)
1391 ether= [HW,NET] Ethernet cards parameters
1392 This option is obsoleted by the "netdev=" option, which
1393 has equivalent usage. See its documentation for details.
1397 Permit 'security.evm' to be updated regardless of
1398 current integrity status.
1403 fail_make_request=[KNL]
1404 General fault injection mechanism.
1405 Format: <interval>,<probability>,<space>,<times>
1406 See also Documentation/fault-injection/.
1409 Format: { initns | none }
1410 See Documentation/admin-guide/sysctl/net.rst for
1411 fb_tunnels_only_for_init_ns
1414 See Documentation/admin-guide/blockdev/floppy.rst.
1416 force_pal_cache_flush
1417 [IA-64] Avoid check_sal_cache_flush which may hang on
1418 buggy SAL_CACHE_FLUSH implementations. Using this
1419 parameter will force ia64_sal_cache_flush to call
1420 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1423 Forcefully enable Physical Address Extension (PAE).
1424 Many Pentium M systems disable PAE but may have a
1425 functionally usable PAE implementation.
1426 Warning: use of this parameter will taint the kernel
1427 and may cause unknown problems.
1430 [FTRACE] will set and start the specified tracer
1431 as early as possible in order to facilitate early
1434 ftrace_dump_on_oops[=orig_cpu]
1435 [FTRACE] will dump the trace buffers on oops.
1436 If no parameter is passed, ftrace will dump
1437 buffers of all CPUs, but if you pass orig_cpu, it will
1438 dump only the buffer of the CPU that triggered the
1441 ftrace_filter=[function-list]
1442 [FTRACE] Limit the functions traced by the function
1443 tracer at boot up. function-list is a comma-separated
1444 list of functions. This list can be changed at run
1445 time by the set_ftrace_filter file in the debugfs
1448 ftrace_notrace=[function-list]
1449 [FTRACE] Do not trace the functions specified in
1450 function-list. This list can be changed at run time
1451 by the set_ftrace_notrace file in the debugfs
1454 ftrace_graph_filter=[function-list]
1455 [FTRACE] Limit the top level callers functions traced
1456 by the function graph tracer at boot up.
1457 function-list is a comma-separated list of functions
1458 that can be changed at run time by the
1459 set_graph_function file in the debugfs tracing directory.
1461 ftrace_graph_notrace=[function-list]
1462 [FTRACE] Do not trace from the functions specified in
1463 function-list. This list is a comma-separated list of
1464 functions that can be changed at run time by the
1465 set_graph_notrace file in the debugfs tracing directory.
1467 ftrace_graph_max_depth=<uint>
1468 [FTRACE] Used with the function graph tracer. This is
1469 the max depth it will trace into a function. This value
1470 can be changed at run time by the max_graph_depth file
1471 in the tracefs tracing directory. default: 0 (no limit)
1473 fw_devlink= [KNL] Create device links between consumer and supplier
1474 devices by scanning the firmware to infer the
1475 consumer/supplier relationships. This feature is
1476 especially useful when drivers are loaded as modules as
1477 it ensures proper ordering of tasks like device probing
1478 (suppliers first, then consumers), supplier boot state
1479 clean up (only after all consumers have probed),
1480 suspend/resume & runtime PM (consumers first, then
1482 Format: { off | permissive | on | rpm }
1483 off -- Don't create device links from firmware info.
1484 permissive -- Create device links from firmware info
1485 but use it only for ordering boot state clean
1486 up (sync_state() calls).
1487 on -- Create device links from firmware info and use it
1488 to enforce probe and suspend/resume ordering.
1489 rpm -- Like "on", but also use to order runtime PM.
1491 fw_devlink.strict=<bool>
1492 [KNL] Treat all inferred dependencies as mandatory
1493 dependencies. This only applies for fw_devlink=on|rpm.
1497 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1498 support via parallel port (up to 5 devices per port)
1499 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1500 See also Documentation/input/devices/joystick-parport.rst
1504 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1508 gcov_persist= [GCOV] When non-zero (default), profiling data for
1509 kernel modules is saved and remains accessible via
1510 debugfs, even when the module is unloaded/reloaded.
1511 When zero, profiling data is discarded and associated
1512 debugfs files are removed at module unload time.
1514 goldfish [X86] Enable the goldfish android emulator platform.
1515 Don't use this when you are not running on the
1518 gpio-mockup.gpio_mockup_ranges
1519 [HW] Sets the ranges of gpiochip of for this device.
1520 Format: <start1>,<end1>,<start2>,<end2>...
1521 gpio-mockup.gpio_mockup_named_lines
1522 [HW] Let the driver know GPIO lines should be named.
1524 gpt [EFI] Forces disk with valid GPT signature but
1525 invalid Protective MBR to be treated as GPT. If the
1526 primary GPT is corrupted, it enables the backup/alternate
1527 GPT to be used instead.
1529 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1530 the "Enable 0" bit of the configuration register.
1533 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1534 the "Enable 0" bit of the configuration register.
1537 grcan.select= [HW] Select which physical interface to use.
1540 grcan.txsize= [HW] Sets the size of the tx buffer.
1541 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1543 grcan.rxsize= [HW] Sets the size of the rx buffer.
1544 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1547 hardlockup_all_cpu_backtrace=
1548 [KNL] Should the hard-lockup detector generate
1549 backtraces on all cpus.
1552 hashdist= [KNL,NUMA] Large hashes allocated during boot
1553 are distributed across NUMA nodes. Defaults on
1554 for 64-bit NUMA, off otherwise.
1555 Format: 0 | 1 (for off | on)
1557 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1559 hd= [EIDE] (E)IDE hard drive subsystem geometry
1560 Format: <cyl>,<head>,<sect>
1563 Disable Hardware Error Source Table (HEST) support;
1564 corresponding firmware-first mode error processing
1565 logic will be disabled.
1567 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1568 size of <nn>. This works even on boxes that have no
1569 highmem otherwise. This also works to reduce highmem
1570 size on bigger boxes.
1572 highres= [KNL] Enable/disable high resolution timer mode.
1573 Valid parameters: "on", "off"
1578 hpet= [X86-32,HPET] option to control HPET usage
1579 Format: { enable (default) | disable | force |
1581 disable: disable HPET and use PIT instead
1582 force: allow force enabled of undocumented chips (ICH4,
1584 verbose: show contents of HPET registers during setup
1586 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1587 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1589 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1590 of gigantic hugepages.
1593 Reserve a CMA area of given size and allocate gigantic
1594 hugepages using the CMA allocator. If enabled, the
1595 boot-time allocation of gigantic hugepages is skipped.
1597 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1598 If this follows hugepagesz (below), it specifies
1599 the number of pages of hugepagesz to be allocated.
1600 If this is the first HugeTLB parameter on the command
1601 line, it specifies the number of pages to allocate for
1602 the default huge page size. See also
1603 Documentation/admin-guide/mm/hugetlbpage.rst.
1607 [HW] The size of the HugeTLB pages. This is used in
1608 conjunction with hugepages (above) to allocate huge
1609 pages of a specific size at boot. The pair
1610 hugepagesz=X hugepages=Y can be specified once for
1611 each supported huge page size. Huge page sizes are
1612 architecture dependent. See also
1613 Documentation/admin-guide/mm/hugetlbpage.rst.
1616 hugetlb_free_vmemmap=
1617 [KNL] Reguires CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
1619 Allows heavy hugetlb users to free up some more
1620 memory (6 * PAGE_SIZE for each 2MB hugetlb page).
1621 Format: { on | off (default) }
1623 on: enable the feature
1624 off: disable the feature
1626 Built with CONFIG_HUGETLB_PAGE_FREE_VMEMMAP_DEFAULT_ON=y,
1629 This is not compatible with memory_hotplug.memmap_on_memory.
1630 If both parameters are enabled, hugetlb_free_vmemmap takes
1631 precedence over memory_hotplug.memmap_on_memory.
1634 [KNL] Should the hung task detector generate panics.
1637 A value of 1 instructs the kernel to panic when a
1638 hung task is detected. The default value is controlled
1639 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1640 option. The value selected by this boot parameter can
1641 be changed later by the kernel.hung_task_panic sysctl.
1643 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1644 terminal devices. Valid values: 0..8
1645 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1646 If specified, z/VM IUCV HVC accepts connections
1647 from listed z/VM user IDs only.
1649 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1650 which allow the hypervisor to 'idle' the
1651 guest on lock contention.
1654 Do not unregister boot console at start. This is only
1655 useful for debugging when something happens in the window
1656 between unregistering the boot console and initializing
1659 i2c_bus= [HW] Override the default board specific I2C bus speed
1660 or register an additional I2C bus that is not
1661 registered from board initialization code.
1665 i8042.debug [HW] Toggle i8042 debug mode
1666 i8042.unmask_kbd_data
1667 [HW] Enable printing of interrupt data from the KBD port
1668 (disabled by default, and as a pre-condition
1669 requires that i8042.debug=1 be enabled)
1670 i8042.direct [HW] Put keyboard port into non-translated mode
1671 i8042.dumbkbd [HW] Pretend that controller can only read data from
1672 keyboard and cannot control its state
1673 (Don't attempt to blink the leds)
1674 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1675 i8042.nokbd [HW] Don't check/create keyboard port
1676 i8042.noloop [HW] Disable the AUX Loopback command while probing
1678 i8042.nomux [HW] Don't check presence of an active multiplexing
1680 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1682 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1683 i8042.reset [HW] Reset the controller during init, cleanup and
1684 suspend-to-ram transitions, only during s2r
1685 transitions, or never reset
1686 Format: { 1 | Y | y | 0 | N | n }
1687 1, Y, y: always reset controller
1688 0, N, n: don't ever reset controller
1689 Default: only on s2r transitions on x86; most other
1690 architectures force reset to be always executed
1691 i8042.unlock [HW] Unlock (ignore) the keylock
1692 i8042.kbdreset [HW] Reset device connected to KBD port
1694 [HW] Allow deferred probing upon i8042 probe errors
1698 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1699 indicates that the driver is running on unsupported
1701 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1702 does not match list of supported models.
1704 [HW] Report power status in /proc/i8k
1705 (disabled by default)
1706 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1709 i915.invert_brightness=
1710 [DRM] Invert the sense of the variable that is used to
1711 set the brightness of the panel backlight. Normally a
1712 brightness value of 0 indicates backlight switched off,
1713 and the maximum of the brightness value sets the backlight
1714 to maximum brightness. If this parameter is set to 0
1715 (default) and the machine requires it, or this parameter
1716 is set to 1, a brightness value of 0 sets the backlight
1717 to maximum brightness, and the maximum of the brightness
1718 value switches the backlight off.
1719 -1 -- never invert brightness
1720 0 -- machine default
1721 1 -- force brightness inversion
1724 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1726 ide-core.nodma= [HW] (E)IDE subsystem
1727 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1728 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1729 .cdrom .chs .ignore_cable are additional options
1730 See Documentation/ide/ide.rst.
1732 ide-generic.probe-mask= [HW] (E)IDE subsystem
1734 Probe mask for legacy ISA IDE ports. Depending on
1735 platform up to 6 ports are supported, enabled by
1736 setting corresponding bits in the mask to 1. The
1737 default value is 0x0, which has a special meaning.
1738 On systems that have PCI, it triggers scanning the
1739 PCI bus for the first and the second port, which
1740 are then probed. On systems without PCI the value
1741 of 0x0 enables probing the two first ports as if it
1744 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1745 Claim all unknown PCI IDE storage controllers.
1748 Format: idle=poll, idle=halt, idle=nomwait
1749 Poll forces a polling idle loop that can slightly
1750 improve the performance of waking up a idle CPU, but
1751 will use a lot of power and make the system run hot.
1753 idle=halt: Halt is forced to be used for CPU idle.
1754 In such case C2/C3 won't be used again.
1755 idle=nomwait: Disable mwait for CPU C-states
1759 Allow force disabling of Shared Virtual Memory (SVA)
1760 support for the idxd driver. By default it is set to
1763 idxd.tc_override= [HW]
1765 Allow override of default traffic class configuration
1766 for the device. By default it is set to false (0).
1768 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1769 Format: { strict | legacy | 2008 | relaxed }
1772 Choose which programs will be accepted for execution
1773 based on the IEEE 754 NaN encoding(s) supported by
1774 the FPU and the NaN encoding requested with the value
1775 of an ELF file header flag individually set by each
1776 binary. Hardware implementations are permitted to
1777 support either or both of the legacy and the 2008 NaN
1780 Available settings are as follows:
1781 strict accept binaries that request a NaN encoding
1782 supported by the FPU
1783 legacy only accept legacy-NaN binaries, if supported
1785 2008 only accept 2008-NaN binaries, if supported
1787 relaxed accept any binaries regardless of whether
1788 supported by the FPU
1790 The FPU emulator is always able to support both NaN
1791 encodings, so if no FPU hardware is present or it has
1792 been disabled with 'nofpu', then the settings of
1793 'legacy' and '2008' strap the emulator accordingly,
1794 'relaxed' straps the emulator for both legacy-NaN and
1795 2008-NaN, whereas 'strict' enables legacy-NaN only on
1796 legacy processors and both NaN encodings on MIPS32 or
1799 The setting for ABS.fmt/NEG.fmt instruction execution
1800 mode generally follows that for the NaN encoding,
1801 except where unsupported by hardware.
1803 ignore_loglevel [KNL]
1804 Ignore loglevel setting - this will print /all/
1805 kernel messages to the console. Useful for debugging.
1806 We also add it as printk module parameter, so users
1807 could change it dynamically, usually by
1808 /sys/module/printk/parameters/ignore_loglevel.
1811 Ignore RLIMIT_DATA setting for data mappings,
1812 print warning at first misuse. Can be changed via
1813 /sys/module/kernel/parameters/ignore_rlimit_data.
1815 ihash_entries= [KNL]
1816 Set number of hash buckets for inode cache.
1818 ima_appraise= [IMA] appraise integrity measurements
1819 Format: { "off" | "enforce" | "fix" | "log" }
1822 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1823 The builtin appraise policy appraises all files
1826 ima_canonical_fmt [IMA]
1827 Use the canonical format for the binary runtime
1828 measurements, instead of host native format.
1831 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1835 The list of supported hash algorithms is defined
1836 in crypto/hash_info.h.
1839 The builtin policies to load during IMA setup.
1840 Format: "tcb | appraise_tcb | secure_boot |
1841 fail_securely | critical_data"
1843 The "tcb" policy measures all programs exec'd, files
1844 mmap'd for exec, and all files opened with the read
1845 mode bit set by either the effective uid (euid=0) or
1848 The "appraise_tcb" policy appraises the integrity of
1849 all files owned by root.
1851 The "secure_boot" policy appraises the integrity
1852 of files (eg. kexec kernel image, kernel modules,
1853 firmware, policy, etc) based on file signatures.
1855 The "fail_securely" policy forces file signature
1856 verification failure also on privileged mounted
1857 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1860 The "critical_data" policy measures kernel integrity
1863 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1864 Load a policy which meets the needs of the Trusted
1865 Computing Base. This means IMA will measure all
1866 programs exec'd, files mmap'd for exec, and all files
1867 opened for read by uid=0.
1870 Select one of defined IMA measurements template formats.
1871 Formats: { "ima" | "ima-ng" | "ima-sig" }
1875 [IMA] Define a custom template format.
1876 Format: { "field1|...|fieldN" }
1878 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1879 Format: <min_file_size>
1880 Set the minimal file size for using asynchronous hash.
1881 If left unspecified, ahash usage is disabled.
1883 ahash performance varies for different data sizes on
1884 different crypto accelerators. This option can be used
1885 to achieve the best performance for a particular HW.
1887 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1889 Set hashing buffer size. Default: 4k.
1891 ahash performance varies for different chunk sizes on
1892 different crypto accelerators. This option can be used
1893 to achieve best performance for particular HW.
1897 Run specified binary instead of /sbin/init as init
1900 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1901 for working out where the kernel is dying during
1904 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1905 initcall functions. Useful for debugging built-in
1906 modules and initcalls.
1908 initramfs_async= [KNL]
1911 This parameter controls whether the initramfs
1912 image is unpacked asynchronously, concurrently
1913 with devices being probed and
1914 initialized. This should normally just work,
1915 but as a debugging aid, one can get the
1916 historical behaviour of the initramfs
1917 unpacking being completed before device_ and
1920 initrd= [BOOT] Specify the location of the initial ramdisk
1922 initrdmem= [KNL] Specify a physical address and size from which to
1923 load the initrd. If an initrd is compiled in or
1924 specified in the bootparams, it takes priority over this
1926 Format: ss[KMG],nn[KMG]
1929 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1932 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1934 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1936 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1938 init_pkru= [X86] Specify the default memory protection keys rights
1939 register contents for all processes. 0x55555554 by
1940 default (disallow access to all but pkey 0). Can
1941 override in debugfs after boot.
1943 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1946 int_pln_enable [X86] Enable power limit notification interrupt
1948 integrity_audit=[IMA]
1949 Format: { "0" | "1" }
1950 0 -- basic integrity auditing messages. (Default)
1951 1 -- additional integrity auditing messages.
1953 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1955 Enable intel iommu driver.
1957 Disable intel iommu driver.
1958 igfx_off [Default Off]
1959 By default, gfx is mapped as normal device. If a gfx
1960 device has a dedicated DMAR unit, the DMAR unit is
1961 bypassed by not enabling DMAR with this option. In
1962 this case, gfx device will use physical address for
1964 strict [Default Off]
1965 Deprecated, equivalent to iommu.strict=1.
1966 sp_off [Default Off]
1967 By default, super page will be supported if Intel IOMMU
1968 has the capability. With this option, super page will
1971 Enable the Intel IOMMU scalable mode if the hardware
1972 advertises that it has support for the scalable mode
1975 Disallow use of the Intel IOMMU scalable mode.
1976 tboot_noforce [Default Off]
1977 Do not force the Intel IOMMU enabled under tboot.
1978 By default, tboot will force Intel IOMMU on, which
1979 could harm performance of some high-throughput
1980 devices like 40GBit network cards, even if identity
1982 Note that using this option lowers the security
1983 provided by tboot because it makes the system
1984 vulnerable to DMA attacks.
1986 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1987 0 disables intel_idle and fall back on acpi_idle.
1988 1 to 9 specify maximum depth of C-state.
1992 Do not enable intel_pstate as the default
1993 scaling driver for the supported processors
1995 Use intel_pstate as a scaling driver, but configure it
1996 to work with generic cpufreq governors (instead of
1997 enabling its internal governor). This mode cannot be
1998 used along with the hardware-managed P-states (HWP)
2001 Enable intel_pstate on systems that prohibit it by default
2002 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2003 instead of acpi-cpufreq may disable platform features, such
2004 as thermal controls and power capping, that rely on ACPI
2005 P-States information being indicated to OSPM and therefore
2006 should be used with caution. This option does not work with
2007 processors that aren't supported by the intel_pstate driver
2008 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2010 Do not enable hardware P state control (HWP)
2013 Only load intel_pstate on systems which support
2014 hardware P state control (HWP) if available.
2016 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2017 Description Table, specifies preferred power management
2018 profile as "Enterprise Server" or "Performance Server",
2019 then this feature is turned on by default.
2021 Allow per-logical-CPU P-State performance control limits using
2022 cpufreq sysfs interface
2024 intremap= [X86-64, Intel-IOMMU]
2025 on enable Interrupt Remapping (default)
2026 off disable Interrupt Remapping
2027 nosid disable Source ID checking
2029 BIOS x2APIC opt-out request will be ignored
2030 nopost disable Interrupt Posting
2032 iomem= Disable strict checking of access to MMIO memory
2033 strict regions from userspace.
2048 nobypass [PPC/POWERNV]
2049 Disable IOMMU bypass, using IOMMU for PCI devices.
2051 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2052 Format: { "0" | "1" }
2053 0 - Try to allocate a 32-bit DMA address first, before
2054 falling back to the full range if needed.
2055 1 - Allocate directly from the full usable range,
2056 forcing Dual Address Cycle for PCI cards supporting
2057 greater than 32-bit addressing.
2059 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2060 Format: { "0" | "1" }
2062 Request that DMA unmap operations use deferred
2063 invalidation of hardware TLBs, for increased
2064 throughput at the cost of reduced device isolation.
2065 Will fall back to strict mode if not supported by
2066 the relevant IOMMU driver.
2068 DMA unmap operations invalidate IOMMU hardware TLBs
2070 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2071 Note: on x86, strict mode specified via one of the
2072 legacy driver-specific options takes precedence.
2075 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2076 Format: { "0" | "1" }
2077 0 - Use IOMMU translation for DMA.
2078 1 - Bypass the IOMMU for DMA.
2079 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2081 io7= [HW] IO7 for Marvel-based Alpha systems
2082 See comment before marvel_specify_io7 in
2083 arch/alpha/kernel/core_marvel.c.
2085 io_delay= [X86] I/O delay method
2087 Standard port 0x80 based delay
2089 Alternate port 0xed based delay (needed on some systems)
2091 Simple two microseconds delay
2096 See Documentation/admin-guide/nfs/nfsroot.rst.
2098 ipcmni_extend [KNL] Extend the maximum number of unique System V
2099 IPC identifiers from 32,768 to 16,777,216.
2101 irqaffinity= [SMP] Set the default irq affinity mask
2102 The argument is a cpu list, as described above.
2104 irqchip.gicv2_force_probe=
2107 Force the kernel to look for the second 4kB page
2108 of a GICv2 controller even if the memory range
2109 exposed by the device tree is too small.
2111 irqchip.gicv3_nolpi=
2113 Force the kernel to ignore the availability of
2114 LPIs (and by consequence ITSs). Intended for system
2115 that use the kernel as a bootloader, and thus want
2116 to let secondary kernels in charge of setting up
2119 irqchip.gicv3_pseudo_nmi= [ARM64]
2120 Enables support for pseudo-NMIs in the kernel. This
2121 requires the kernel to be built with
2122 CONFIG_ARM64_PSEUDO_NMI.
2125 When an interrupt is not handled search all handlers
2126 for it. Intended to get systems with badly broken
2130 When an interrupt is not handled search all handlers
2131 for it. Also check all handlers each timer
2132 interrupt. Intended to get systems with badly broken
2136 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2138 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2139 [Deprecated - use cpusets instead]
2140 Format: [flag-list,]<cpu-list>
2142 Specify one or more CPUs to isolate from disturbances
2143 specified in the flag list (default: domain):
2146 Disable the tick when a single task runs.
2148 A residual 1Hz tick is offloaded to workqueues, which you
2149 need to affine to housekeeping through the global
2150 workqueue's affinity configured via the
2151 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2152 by using the 'domain' flag described below.
2154 NOTE: by default the global workqueue runs on all CPUs,
2155 so to protect individual CPUs the 'cpumask' file has to
2156 be configured manually after bootup.
2159 Isolate from the general SMP balancing and scheduling
2160 algorithms. Note that performing domain isolation this way
2161 is irreversible: it's not possible to bring back a CPU to
2162 the domains once isolated through isolcpus. It's strongly
2163 advised to use cpusets instead to disable scheduler load
2164 balancing through the "cpuset.sched_load_balance" file.
2165 It offers a much more flexible interface where CPUs can
2166 move in and out of an isolated set anytime.
2168 You can move a process onto or off an "isolated" CPU via
2169 the CPU affinity syscalls or cpuset.
2170 <cpu number> begins at 0 and the maximum value is
2171 "number of CPUs in system - 1".
2175 Isolate from being targeted by managed interrupts
2176 which have an interrupt mask containing isolated
2177 CPUs. The affinity of managed interrupts is
2178 handled by the kernel and cannot be changed via
2179 the /proc/irq/* interfaces.
2181 This isolation is best effort and only effective
2182 if the automatically assigned interrupt mask of a
2183 device queue contains isolated and housekeeping
2184 CPUs. If housekeeping CPUs are online then such
2185 interrupts are directed to the housekeeping CPU
2186 so that IO submitted on the housekeeping CPU
2187 cannot disturb the isolated CPU.
2189 If a queue's affinity mask contains only isolated
2190 CPUs then this parameter has no effect on the
2191 interrupt routing decision, though interrupts are
2192 only delivered when tasks running on those
2193 isolated CPUs submit IO. IO submitted on
2194 housekeeping CPUs has no influence on those
2197 The format of <cpu-list> is described above.
2201 ivrs_ioapic [HW,X86-64]
2202 Provide an override to the IOAPIC-ID<->DEVICE-ID
2203 mapping provided in the IVRS ACPI table. For
2204 example, to map IOAPIC-ID decimal 10 to
2205 PCI device 00:14.0 write the parameter as:
2206 ivrs_ioapic[10]=00:14.0
2208 ivrs_hpet [HW,X86-64]
2209 Provide an override to the HPET-ID<->DEVICE-ID
2210 mapping provided in the IVRS ACPI table. For
2211 example, to map HPET-ID decimal 0 to
2212 PCI device 00:14.0 write the parameter as:
2213 ivrs_hpet[0]=00:14.0
2215 ivrs_acpihid [HW,X86-64]
2216 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2217 mapping provided in the IVRS ACPI table. For
2218 example, to map UART-HID:UID AMD0020:0 to
2219 PCI device 00:14.5 write the parameter as:
2220 ivrs_acpihid[00:14.5]=AMD0020:0
2222 js= [HW,JOY] Analog joystick
2223 See Documentation/input/joydev/joystick.rst.
2226 When CONFIG_RANDOMIZE_BASE is set, this disables
2227 kernel and module base offset ASLR (Address Space
2228 Layout Randomization).
2231 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2232 report on every invalid memory access. Without this
2233 parameter KASAN will print report only for the first
2238 kernelcore= [KNL,X86,IA-64,PPC]
2239 Format: nn[KMGTPE] | nn% | "mirror"
2240 This parameter specifies the amount of memory usable by
2241 the kernel for non-movable allocations. The requested
2242 amount is spread evenly throughout all nodes in the
2243 system as ZONE_NORMAL. The remaining memory is used for
2244 movable memory in its own zone, ZONE_MOVABLE. In the
2245 event, a node is too small to have both ZONE_NORMAL and
2246 ZONE_MOVABLE, kernelcore memory will take priority and
2247 other nodes will have a larger ZONE_MOVABLE.
2249 ZONE_MOVABLE is used for the allocation of pages that
2250 may be reclaimed or moved by the page migration
2251 subsystem. Note that allocations like PTEs-from-HighMem
2252 still use the HighMem zone if it exists, and the Normal
2253 zone if it does not.
2255 It is possible to specify the exact amount of memory in
2256 the form of "nn[KMGTPE]", a percentage of total system
2257 memory in the form of "nn%", or "mirror". If "mirror"
2258 option is specified, mirrored (reliable) memory is used
2259 for non-movable allocations and remaining memory is used
2260 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2261 are exclusive, so you cannot specify multiple forms.
2263 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2264 Format: <Controller#>[,poll interval]
2265 The controller # is the number of the ehci usb debug
2266 port as it is probed via PCI. The poll interval is
2267 optional and is the number seconds in between
2268 each poll cycle to the debug port in case you need
2269 the functionality for interrupting the kernel with
2270 gdb or control-c on the dbgp connection. When
2271 not using this parameter you use sysrq-g to break into
2272 the kernel debugger.
2274 kgdboc= [KGDB,HW] kgdb over consoles.
2275 Requires a tty driver that supports console polling,
2276 or a supported polling keyboard driver (non-usb).
2277 Serial only format: <serial_device>[,baud]
2278 keyboard only format: kbd
2279 keyboard and serial format: kbd,<serial_device>[,baud]
2280 Optional Kernel mode setting:
2281 kms, kbd format: kms,kbd
2282 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2284 kgdboc_earlycon= [KGDB,HW]
2285 If the boot console provides the ability to read
2286 characters and can work in polling mode, you can use
2287 this parameter to tell kgdb to use it as a backend
2288 until the normal console is registered. Intended to
2289 be used together with the kgdboc parameter which
2290 specifies the normal console to transition to.
2292 The name of the early console should be specified
2293 as the value of this parameter. Note that the name of
2294 the early console might be different than the tty
2295 name passed to kgdboc. It's OK to leave the value
2296 blank and the first boot console that implements
2297 read() will be picked.
2299 kgdbwait [KGDB] Stop kernel execution and enter the
2300 kernel debugger at the earliest opportunity.
2302 kmac= [MIPS] Korina ethernet MAC address.
2303 Configure the RouterBoard 532 series on-chip
2304 Ethernet adapter MAC address.
2306 kmemleak= [KNL] Boot-time kmemleak enable/disable
2307 Valid arguments: on, off
2309 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2312 kprobe_event=[probe-list]
2313 [FTRACE] Add kprobe events and enable at boot time.
2314 The probe-list is a semicolon delimited list of probe
2315 definitions. Each definition is same as kprobe_events
2316 interface, but the parameters are comma delimited.
2317 For example, to add a kprobe event on vfs_read with
2318 arg1 and arg2, add to the command line;
2320 kprobe_event=p,vfs_read,$arg1,$arg2
2322 See also Documentation/trace/kprobetrace.rst "Kernel
2323 Boot Parameter" section.
2325 kpti= [ARM64] Control page table isolation of user
2326 and kernel address spaces.
2327 Default: enabled on cores which need mitigation.
2331 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2332 Default is 0 (don't ignore, but inject #GP)
2334 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2335 Default is false (don't support).
2337 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2342 [KVM] Controls the software workaround for the
2343 X86_BUG_ITLB_MULTIHIT bug.
2344 force : Always deploy workaround.
2345 off : Never deploy workaround.
2346 auto : Deploy workaround based on the presence of
2347 X86_BUG_ITLB_MULTIHIT.
2351 If the software workaround is enabled for the host,
2352 guests do need not to enable it for nested guests.
2354 kvm.nx_huge_pages_recovery_ratio=
2355 [KVM] Controls how many 4KiB pages are periodically zapped
2356 back to huge pages. 0 disables the recovery, otherwise if
2357 the value is N KVM will zap 1/Nth of the 4KiB pages every
2358 minute. The default is 60.
2360 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2361 Default is 1 (enabled)
2363 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2365 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2368 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2370 nvhe: Standard nVHE-based mode, without support for
2373 protected: nVHE-based mode with support for guests whose
2374 state is kept private from the host.
2375 Not valid if the kernel is running in EL2.
2377 Defaults to VHE/nVHE based on hardware support.
2379 kvm-arm.vgic_v3_group0_trap=
2380 [KVM,ARM] Trap guest accesses to GICv3 group-0
2383 kvm-arm.vgic_v3_group1_trap=
2384 [KVM,ARM] Trap guest accesses to GICv3 group-1
2387 kvm-arm.vgic_v3_common_trap=
2388 [KVM,ARM] Trap guest accesses to GICv3 common
2391 kvm-arm.vgic_v4_enable=
2392 [KVM,ARM] Allow use of GICv4 for direct injection of
2395 kvm_cma_resv_ratio=n [PPC]
2396 Reserves given percentage from system memory area for
2397 contiguous memory allocation for KVM hash pagetable
2399 By default it reserves 5% of total system memory.
2403 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2404 (virtualized MMU) support on capable Intel chips.
2405 Default is 1 (enabled)
2407 kvm-intel.emulate_invalid_guest_state=
2408 [KVM,Intel] Disable emulation of invalid guest state.
2409 Ignored if kvm-intel.enable_unrestricted_guest=1, as
2410 guest state is never invalid for unrestricted guests.
2411 This param doesn't apply to nested guests (L2), as KVM
2412 never emulates invalid L2 guest state.
2413 Default is 1 (enabled)
2415 kvm-intel.flexpriority=
2416 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2417 Default is 1 (enabled)
2420 [KVM,Intel] Enable VMX nesting (nVMX).
2421 Default is 0 (disabled)
2423 kvm-intel.unrestricted_guest=
2424 [KVM,Intel] Disable unrestricted guest feature
2425 (virtualized real and unpaged mode) on capable
2426 Intel chips. Default is 1 (enabled)
2428 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2431 Valid arguments: never, cond, always
2433 always: L1D cache flush on every VMENTER.
2434 cond: Flush L1D on VMENTER only when the code between
2435 VMEXIT and VMENTER can leak host memory.
2436 never: Disables the mitigation
2438 Default is cond (do L1 cache flush in specific instances)
2440 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2441 feature (tagged TLBs) on capable Intel chips.
2442 Default is 1 (enabled)
2444 l1d_flush= [X86,INTEL]
2445 Control mitigation for L1D based snooping vulnerability.
2447 Certain CPUs are vulnerable to an exploit against CPU
2448 internal buffers which can forward information to a
2449 disclosure gadget under certain conditions.
2451 In vulnerable processors, the speculatively
2452 forwarded data can be used in a cache side channel
2453 attack, to access data to which the attacker does
2454 not have direct access.
2456 This parameter controls the mitigation. The
2459 on - enable the interface for the mitigation
2461 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2464 The kernel PTE inversion protection is unconditionally
2465 enabled and cannot be disabled.
2468 Provides all available mitigations for the
2469 L1TF vulnerability. Disables SMT and
2470 enables all mitigations in the
2471 hypervisors, i.e. unconditional L1D flush.
2473 SMT control and L1D flush control via the
2474 sysfs interface is still possible after
2475 boot. Hypervisors will issue a warning
2476 when the first VM is started in a
2477 potentially insecure configuration,
2478 i.e. SMT enabled or L1D flush disabled.
2481 Same as 'full', but disables SMT and L1D
2482 flush runtime control. Implies the
2483 'nosmt=force' command line option.
2484 (i.e. sysfs control of SMT is disabled.)
2487 Leaves SMT enabled and enables the default
2488 hypervisor mitigation, i.e. conditional
2491 SMT control and L1D flush control via the
2492 sysfs interface is still possible after
2493 boot. Hypervisors will issue a warning
2494 when the first VM is started in a
2495 potentially insecure configuration,
2496 i.e. SMT enabled or L1D flush disabled.
2500 Disables SMT and enables the default
2501 hypervisor mitigation.
2503 SMT control and L1D flush control via the
2504 sysfs interface is still possible after
2505 boot. Hypervisors will issue a warning
2506 when the first VM is started in a
2507 potentially insecure configuration,
2508 i.e. SMT enabled or L1D flush disabled.
2511 Same as 'flush', but hypervisors will not
2512 warn when a VM is started in a potentially
2513 insecure configuration.
2516 Disables hypervisor mitigations and doesn't
2518 It also drops the swap size and available
2519 RAM limit restriction on both hypervisor and
2524 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2530 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2533 lapic= [X86,APIC] Do not use TSC deadline
2534 value for LAPIC timer one-shot implementation. Default
2535 back to the programmable timer unit in the LAPIC.
2536 Format: notscdeadline
2538 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2541 libata.dma= [LIBATA] DMA control
2542 libata.dma=0 Disable all PATA and SATA DMA
2543 libata.dma=1 PATA and SATA Disk DMA only
2544 libata.dma=2 ATAPI (CDROM) DMA only
2545 libata.dma=4 Compact Flash DMA only
2546 Combinations also work, so libata.dma=3 enables DMA
2547 for disks and CDROMs, but not CFs.
2549 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2550 libata.ignore_hpa=0 keep BIOS limits (default)
2551 libata.ignore_hpa=1 ignore limits, using full disk
2553 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2557 libata.force= [LIBATA] Force configurations. The format is comma-
2558 separated list of "[ID:]VAL" where ID is
2559 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2560 matching port, link or device. Basically, it matches
2561 the ATA ID string printed on console by libata. If
2562 the whole ID part is omitted, the last PORT and DEVICE
2563 values are used. If ID hasn't been specified yet, the
2564 configuration applies to all ports, links and devices.
2566 If only DEVICE is omitted, the parameter applies to
2567 the port and all links and devices behind it. DEVICE
2568 number of 0 either selects the first device or the
2569 first fan-out link behind PMP device. It does not
2570 select the host link. DEVICE number of 15 selects the
2571 host link and device attached to it.
2573 The VAL specifies the configuration to force. As long
2574 as there's no ambiguity shortcut notation is allowed.
2575 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2576 The following configurations can be forced.
2578 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2579 Any ID with matching PORT is used.
2581 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2583 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2584 udma[/][16,25,33,44,66,100,133] notation is also
2587 * [no]ncq: Turn on or off NCQ.
2589 * [no]ncqtrim: Turn off queued DSM TRIM.
2591 * nohrst, nosrst, norst: suppress hard, soft
2594 * rstonce: only attempt one reset during
2595 hot-unplug link recovery
2597 * dump_id: dump IDENTIFY data.
2599 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2601 * disable: Disable this device.
2603 If there are multiple matching configurations changing
2604 the same attribute, the last one is used.
2606 memblock=debug [KNL] Enable memblock debug messages.
2608 load_ramdisk= [RAM] [Deprecated]
2610 lockd.nlm_grace_period=P [NFS] Assign grace period.
2613 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2616 lockd.nlm_timeout=T [NFS] Assign timeout value.
2619 lockd.nlm_udpport=M [NFS] Assign UDP port.
2622 lockdown= [SECURITY]
2623 { integrity | confidentiality }
2624 Enable the kernel lockdown feature. If set to
2625 integrity, kernel features that allow userland to
2626 modify the running kernel are disabled. If set to
2627 confidentiality, kernel features that allow userland
2628 to extract confidential information from the kernel
2631 locktorture.nreaders_stress= [KNL]
2632 Set the number of locking read-acquisition kthreads.
2633 Defaults to being automatically set based on the
2634 number of online CPUs.
2636 locktorture.nwriters_stress= [KNL]
2637 Set the number of locking write-acquisition kthreads.
2639 locktorture.onoff_holdoff= [KNL]
2640 Set time (s) after boot for CPU-hotplug testing.
2642 locktorture.onoff_interval= [KNL]
2643 Set time (s) between CPU-hotplug operations, or
2644 zero to disable CPU-hotplug testing.
2646 locktorture.shuffle_interval= [KNL]
2647 Set task-shuffle interval (jiffies). Shuffling
2648 tasks allows some CPUs to go into dyntick-idle
2649 mode during the locktorture test.
2651 locktorture.shutdown_secs= [KNL]
2652 Set time (s) after boot system shutdown. This
2653 is useful for hands-off automated testing.
2655 locktorture.stat_interval= [KNL]
2656 Time (s) between statistics printk()s.
2658 locktorture.stutter= [KNL]
2659 Time (s) to stutter testing, for example,
2660 specifying five seconds causes the test to run for
2661 five seconds, wait for five seconds, and so on.
2662 This tests the locking primitive's ability to
2663 transition abruptly to and from idle.
2665 locktorture.torture_type= [KNL]
2666 Specify the locking implementation to test.
2668 locktorture.verbose= [KNL]
2669 Enable additional printk() statements.
2671 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2674 loglevel= All Kernel Messages with a loglevel smaller than the
2675 console loglevel will be printed to the console. It can
2676 also be changed with klogd or other programs. The
2677 loglevels are defined as follows:
2679 0 (KERN_EMERG) system is unusable
2680 1 (KERN_ALERT) action must be taken immediately
2681 2 (KERN_CRIT) critical conditions
2682 3 (KERN_ERR) error conditions
2683 4 (KERN_WARNING) warning conditions
2684 5 (KERN_NOTICE) normal but significant condition
2685 6 (KERN_INFO) informational
2686 7 (KERN_DEBUG) debug-level messages
2688 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2689 in bytes. n must be a power of two and greater
2690 than the minimal size. The minimal size is defined
2691 by LOG_BUF_SHIFT kernel config parameter. There is
2692 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2693 that allows to increase the default size depending on
2694 the number of CPUs. See init/Kconfig for more details.
2696 logo.nologo [FB] Disables display of the built-in Linux logo.
2697 This may be used to provide more screen space for
2698 kernel log messages and is useful when debugging
2699 kernel boot problems.
2701 lp=0 [LP] Specify parallel ports to use, e.g,
2702 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2703 lp=reset first parallel port). 'lp=0' disables the
2704 lp=auto printer driver. 'lp=reset' (which can be
2705 specified in addition to the ports) causes
2706 attached printers to be reset. Using
2707 lp=port1,port2,... specifies the parallel ports
2708 to associate lp devices with, starting with
2709 lp0. A port specification may be 'none' to skip
2710 that lp device, or a parport name such as
2711 'parport0'. Specifying 'lp=auto' instead of a
2712 port specification list means that device IDs
2713 from each port should be examined, to see if
2714 an IEEE 1284-compliant printer is attached; if
2715 so, the driver will manage that printer.
2716 See also header of drivers/char/lp.c.
2719 Sets loops_per_jiffy to given constant, thus avoiding
2720 time-consuming boot-time autodetection (up to 250 ms per
2721 CPU). 0 enables autodetection (default). To determine
2722 the correct value for your kernel, boot with normal
2723 autodetection and see what value is printed. Note that
2724 on SMP systems the preset will be applied to all CPUs,
2725 which is likely to cause problems if your CPUs need
2726 significantly divergent settings. An incorrect value
2727 will cause delays in the kernel to be wrong, leading to
2728 unpredictable I/O errors and other breakage. Although
2729 unlikely, in the extreme case this might damage your
2733 Format: <io>,<irq>,<dma>
2735 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2738 [SECURITY] Choose order of LSM initialization. This
2739 overrides CONFIG_LSM, and the "security=" parameter.
2741 machvec= [IA-64] Force the use of a particular machine-vector
2742 (machvec) in a generic kernel.
2743 Example: machvec=hpzx1
2745 machtype= [Loongson] Share the same kernel image file between
2746 different yeeloong laptops.
2747 Example: machtype=lemote-yeeloong-2f-7inch
2749 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2750 than or equal to this physical address is ignored.
2752 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2753 will bring up during bootup. maxcpus=n : n >= 0 limits
2754 the kernel to bring up 'n' processors. Surely after
2755 bootup you can bring up the other plugged cpu by executing
2756 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2757 only takes effect during system bootup.
2758 While n=0 is a special case, it is equivalent to "nosmp",
2759 which also disables the IO APIC.
2761 max_loop= [LOOP] The number of loop block devices that get
2762 (loop.max_loop) unconditionally pre-created at init time. The default
2763 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2764 of statically allocating a predefined number, loop
2765 devices can be requested on-demand with the
2766 /dev/loop-control interface.
2768 mce [X86-32] Machine Check Exception
2770 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2772 md= [HW] RAID subsystems devices and level
2773 See Documentation/admin-guide/md.rst.
2776 Format: <first>,<last>
2777 Specifies range of consoles to be captured by the MDA.
2780 Control mitigation for the Micro-architectural Data
2781 Sampling (MDS) vulnerability.
2783 Certain CPUs are vulnerable to an exploit against CPU
2784 internal buffers which can forward information to a
2785 disclosure gadget under certain conditions.
2787 In vulnerable processors, the speculatively
2788 forwarded data can be used in a cache side channel
2789 attack, to access data to which the attacker does
2790 not have direct access.
2792 This parameter controls the MDS mitigation. The
2795 full - Enable MDS mitigation on vulnerable CPUs
2796 full,nosmt - Enable MDS mitigation and disable
2797 SMT on vulnerable CPUs
2798 off - Unconditionally disable MDS mitigation
2800 On TAA-affected machines, mds=off can be prevented by
2801 an active TAA mitigation as both vulnerabilities are
2802 mitigated with the same mechanism so in order to disable
2803 this mitigation, you need to specify tsx_async_abort=off
2806 Not specifying this option is equivalent to
2809 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2811 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2812 Amount of memory to be used in cases as follows:
2815 2 when the kernel is not able to see the whole system memory;
2816 3 memory that lies after 'mem=' boundary is excluded from
2817 the hypervisor, then assigned to KVM guests.
2819 [X86] Work as limiting max address. Use together
2820 with memmap= to avoid physical address space collisions.
2821 Without memmap= PCI devices could be placed at addresses
2822 belonging to unused RAM.
2824 Note that this only takes effects during boot time since
2825 in above case 3, memory may need be hot added after boot
2826 if system memory of hypervisor is not sufficient.
2828 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2832 [KNL,SH] Allow user to override the default size for
2833 per-device physically contiguous DMA buffers.
2835 memhp_default_state=online/offline
2836 [KNL] Set the initial state for the memory hotplug
2837 onlining policy. If not specified, the default value is
2838 set according to the
2839 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2841 See Documentation/admin-guide/mm/memory-hotplug.rst.
2843 memmap=exactmap [KNL,X86] Enable setting of an exact
2844 E820 memory map, as specified by the user.
2845 Such memmap=exactmap lines can be constructed based on
2846 BIOS output or other requirements. See the memmap=nn@ss
2849 memmap=nn[KMG]@ss[KMG]
2850 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
2851 Region of memory to be used is from ss to ss+nn.
2852 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2853 which limits max address to nn[KMG].
2854 Multiple different regions can be specified,
2857 memmap=100M@2G,100M#3G,1G!1024G
2859 memmap=nn[KMG]#ss[KMG]
2860 [KNL,ACPI] Mark specific memory as ACPI data.
2861 Region of memory to be marked is from ss to ss+nn.
2863 memmap=nn[KMG]$ss[KMG]
2864 [KNL,ACPI] Mark specific memory as reserved.
2865 Region of memory to be reserved is from ss to ss+nn.
2866 Example: Exclude memory from 0x18690000-0x1869ffff
2867 memmap=64K$0x18690000
2869 memmap=0x10000$0x18690000
2870 Some bootloaders may need an escape character before '$',
2871 like Grub2, otherwise '$' and the following number
2874 memmap=nn[KMG]!ss[KMG]
2875 [KNL,X86] Mark specific memory as protected.
2876 Region of memory to be used, from ss to ss+nn.
2877 The memory region may be marked as e820 type 12 (0xc)
2878 and is NVDIMM or ADR memory.
2880 memmap=<size>%<offset>-<oldtype>+<newtype>
2881 [KNL,ACPI] Convert memory within the specified region
2882 from <oldtype> to <newtype>. If "-<oldtype>" is left
2883 out, the whole region will be marked as <newtype>,
2884 even if previously unavailable. If "+<newtype>" is left
2885 out, matching memory will be removed. Types are
2886 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2887 3 = ACPI, 12 = PRAM.
2889 memory_corruption_check=0/1 [X86]
2890 Some BIOSes seem to corrupt the first 64k of
2891 memory when doing things like suspend/resume.
2892 Setting this option will scan the memory
2893 looking for corruption. Enabling this will
2894 both detect corruption and prevent the kernel
2895 from using the memory being corrupted.
2896 However, its intended as a diagnostic tool; if
2897 repeatable BIOS-originated corruption always
2898 affects the same memory, you can use memmap=
2899 to prevent the kernel from using that memory.
2901 memory_corruption_check_size=size [X86]
2902 By default it checks for corruption in the low
2903 64k, making this memory unavailable for normal
2904 use. Use this parameter to scan for
2905 corruption in more or less memory.
2907 memory_corruption_check_period=seconds [X86]
2908 By default it checks for corruption every 60
2909 seconds. Use this parameter to check at some
2910 other rate. 0 disables periodic checking.
2912 memory_hotplug.memmap_on_memory
2913 [KNL,X86,ARM] Boolean flag to enable this feature.
2914 Format: {on | off (default)}
2915 When enabled, runtime hotplugged memory will
2916 allocate its internal metadata (struct pages)
2917 from the hotadded memory which will allow to
2918 hotadd a lot of memory without requiring
2919 additional memory to do so.
2920 This feature is disabled by default because it
2921 has some implication on large (e.g. GB)
2922 allocations in some configurations (e.g. small
2924 The state of the flag can be read in
2925 /sys/module/memory_hotplug/parameters/memmap_on_memory.
2926 Note that even when enabled, there are a few cases where
2927 the feature is not effective.
2929 This is not compatible with hugetlb_free_vmemmap. If
2930 both parameters are enabled, hugetlb_free_vmemmap takes
2931 precedence over memory_hotplug.memmap_on_memory.
2933 memtest= [KNL,X86,ARM,PPC,RISCV] Enable memtest
2935 default : 0 <disable>
2936 Specifies the number of memtest passes to be
2937 performed. Each pass selects another test
2938 pattern from a given set of patterns. Memtest
2939 fills the memory with this pattern, validates
2940 memory contents and reserves bad memory
2941 regions that are detected.
2943 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2944 Valid arguments: on, off
2945 Default (depends on kernel configuration option):
2946 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2947 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2948 mem_encrypt=on: Activate SME
2949 mem_encrypt=off: Do not activate SME
2951 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2952 for details on when memory encryption can be activated.
2954 mem_sleep_default= [SUSPEND] Default system suspend mode:
2955 s2idle - Suspend-To-Idle
2956 shallow - Power-On Suspend or equivalent (if supported)
2957 deep - Suspend-To-RAM or equivalent (if supported)
2958 See Documentation/admin-guide/pm/sleep-states.rst.
2960 meye.*= [HW] Set MotionEye Camera parameters
2961 See Documentation/admin-guide/media/meye.rst.
2963 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2964 Multi-Function General Purpose Timers on AMD Geode
2967 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2968 the BIOS has incorrectly applied a workaround. TinyBIOS
2969 version 0.98 is known to be affected, 0.99 fixes the
2970 problem by letting the user disable the workaround.
2974 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2975 physical address is ignored.
2977 mini2440= [ARM,HW,KNL]
2978 Format:[0..2][b][c][t]
2980 MINI2440 configuration specification:
2981 0 - The attached screen is the 3.5" TFT
2982 1 - The attached screen is the 7" TFT
2983 2 - The VGA Shield is attached (1024x768)
2984 Leaving out the screen size parameter will not load
2985 the TFT driver, and the framebuffer will be left
2987 b - Enable backlight. The TFT backlight pin will be
2988 linked to the kernel VESA blanking code and a GPIO
2989 LED. This parameter is not necessary when using the
2991 c - Enable the s3c camera interface.
2992 t - Reserved for enabling touchscreen support. The
2993 touchscreen support is not enabled in the mainstream
2994 kernel as of 2.6.30, a preliminary port can be found
2995 in the "bleeding edge" mini2440 support kernel at
2996 https://repo.or.cz/w/linux-2.6/mini2440.git
2999 [X86,PPC,S390,ARM64] Control optional mitigations for
3000 CPU vulnerabilities. This is a set of curated,
3001 arch-independent options, each of which is an
3002 aggregation of existing arch-specific options.
3005 Disable all optional CPU mitigations. This
3006 improves system performance, but it may also
3007 expose users to several CPU vulnerabilities.
3008 Equivalent to: nopti [X86,PPC]
3010 nospectre_v1 [X86,PPC]
3012 nospectre_v2 [X86,PPC,S390,ARM64]
3013 spectre_v2_user=off [X86]
3014 spec_store_bypass_disable=off [X86,PPC]
3015 ssbd=force-off [ARM64]
3018 tsx_async_abort=off [X86]
3019 kvm.nx_huge_pages=off [X86]
3020 no_entry_flush [PPC]
3021 no_uaccess_flush [PPC]
3022 mmio_stale_data=off [X86]
3026 This does not have any effect on
3027 kvm.nx_huge_pages when
3028 kvm.nx_huge_pages=force.
3031 Mitigate all CPU vulnerabilities, but leave SMT
3032 enabled, even if it's vulnerable. This is for
3033 users who don't want to be surprised by SMT
3034 getting disabled across kernel upgrades, or who
3035 have other ways of avoiding SMT-based attacks.
3036 Equivalent to: (default behavior)
3039 Mitigate all CPU vulnerabilities, disabling SMT
3040 if needed. This is for users who always want to
3041 be fully mitigated, even if it means losing SMT.
3042 Equivalent to: l1tf=flush,nosmt [X86]
3043 mds=full,nosmt [X86]
3044 tsx_async_abort=full,nosmt [X86]
3045 mmio_stale_data=full,nosmt [X86]
3046 retbleed=auto,nosmt [X86]
3049 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3050 parameter allows control of the logging verbosity for
3051 the additional memory initialisation checks. A value
3052 of 0 disables mminit logging and a level of 4 will
3053 log everything. Information is printed at KERN_DEBUG
3054 so loglevel=8 may also need to be specified.
3057 [X86,INTEL] Control mitigation for the Processor
3058 MMIO Stale Data vulnerabilities.
3060 Processor MMIO Stale Data is a class of
3061 vulnerabilities that may expose data after an MMIO
3062 operation. Exposed data could originate or end in
3063 the same CPU buffers as affected by MDS and TAA.
3064 Therefore, similar to MDS and TAA, the mitigation
3065 is to clear the affected CPU buffers.
3067 This parameter controls the mitigation. The
3070 full - Enable mitigation on vulnerable CPUs
3072 full,nosmt - Enable mitigation and disable SMT on
3075 off - Unconditionally disable mitigation
3077 On MDS or TAA affected machines,
3078 mmio_stale_data=off can be prevented by an active
3079 MDS or TAA mitigation as these vulnerabilities are
3080 mitigated with the same mechanism so in order to
3081 disable this mitigation, you need to specify
3082 mds=off and tsx_async_abort=off too.
3084 Not specifying this option is equivalent to
3085 mmio_stale_data=full.
3088 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3091 [KNL] When CONFIG_MODULE_SIG is set, this means that
3092 modules without (valid) signatures will fail to load.
3093 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3094 is always true, so this option does nothing.
3096 module_blacklist= [KNL] Do not load a comma-separated list of
3097 modules. Useful for debugging problem modules.
3100 [MOUSE] Maximum time between finger touching and
3101 leaving touchpad surface for touch to be considered
3102 a tap and be reported as a left button click (for
3103 touchpads working in absolute mode only).
3105 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3106 reporting absolute coordinates, such as tablets
3107 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3108 reporting absolute coordinates, such as tablets
3110 movablecore= [KNL,X86,IA-64,PPC]
3111 Format: nn[KMGTPE] | nn%
3112 This parameter is the complement to kernelcore=, it
3113 specifies the amount of memory used for migratable
3114 allocations. If both kernelcore and movablecore is
3115 specified, then kernelcore will be at *least* the
3116 specified value but may be more. If movablecore on its
3117 own is specified, the administrator must be careful
3118 that the amount of memory usable for all allocations
3121 movable_node [KNL] Boot-time switch to make hotplugable memory
3122 NUMA nodes to be movable. This means that the memory
3123 of such nodes will be usable only for movable
3124 allocations which rules out almost all kernel
3125 allocations. Use with caution!
3127 MTD_Partition= [MTD]
3128 Format: <name>,<region-number>,<size>,<offset>
3130 MTD_Region= [MTD] Format:
3131 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3134 See drivers/mtd/parsers/cmdlinepart.c
3136 multitce=off [PPC] This parameter disables the use of the pSeries
3137 firmware feature for updating multiple TCE entries
3140 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3142 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3144 boundary - index of last SLC block on Flex-OneNAND.
3145 The remaining blocks are configured as MLC blocks.
3146 lock - Configure if Flex-OneNAND boundary should be locked.
3147 Once locked, the boundary cannot be changed.
3148 1 indicates lock status, 0 indicates unlock status.
3151 ARM/S3C2412 JIVE boot control
3153 See arch/arm/mach-s3c/mach-jive.c
3155 mtouchusb.raw_coordinates=
3156 [HW] Make the MicroTouch USB driver use raw coordinates
3157 ('y', default) or cooked coordinates ('n')
3159 mtrr_chunk_size=nn[KMG] [X86]
3160 used for mtrr cleanup. It is largest continuous chunk
3161 that could hold holes aka. UC entries.
3163 mtrr_gran_size=nn[KMG] [X86]
3164 Used for mtrr cleanup. It is granularity of mtrr block.
3166 Large value could prevent small alignment from
3169 mtrr_spare_reg_nr=n [X86]
3171 Range: 0,7 : spare reg number
3173 Used for mtrr cleanup. It is spare mtrr entries number.
3174 Set to 2 or more if your graphical card needs more.
3176 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3178 netdev= [NET] Network devices parameters
3179 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3180 Note that mem_start is often overloaded to mean
3181 something different and driver-specific.
3182 This usage is only documented in each driver source
3186 [NETFILTER] Enable connection tracking flow accounting
3187 0 to disable accounting
3188 1 to enable accounting
3191 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3192 See Documentation/admin-guide/nfs/nfsroot.rst.
3194 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3195 See Documentation/admin-guide/nfs/nfsroot.rst.
3197 nfsrootdebug [NFS] enable nfsroot debugging messages.
3198 See Documentation/admin-guide/nfs/nfsroot.rst.
3200 nfs.callback_nr_threads=
3201 [NFSv4] set the total number of threads that the
3202 NFS client will assign to service NFSv4 callback
3205 nfs.callback_tcpport=
3206 [NFS] set the TCP port on which the NFSv4 callback
3207 channel should listen.
3210 [NFS] sets the pathname to the program which is used
3211 to update the NFS client cache entries.
3213 nfs.cache_getent_timeout=
3214 [NFS] sets the timeout after which an attempt to
3215 update a cache entry is deemed to have failed.
3217 nfs.idmap_cache_timeout=
3218 [NFS] set the maximum lifetime for idmapper cache
3222 [NFS] enable 64-bit inode numbers.
3223 If zero, the NFS client will fake up a 32-bit inode
3224 number for the readdir() and stat() syscalls instead
3225 of returning the full 64-bit number.
3226 The default is to return 64-bit inode numbers.
3228 nfs.max_session_cb_slots=
3229 [NFSv4.1] Sets the maximum number of session
3230 slots the client will assign to the callback
3231 channel. This determines the maximum number of
3232 callbacks the client will process in parallel for
3233 a particular server.
3235 nfs.max_session_slots=
3236 [NFSv4.1] Sets the maximum number of session slots
3237 the client will attempt to negotiate with the server.
3238 This limits the number of simultaneous RPC requests
3239 that the client can send to the NFSv4.1 server.
3240 Note that there is little point in setting this
3241 value higher than the max_tcp_slot_table_limit.
3243 nfs.nfs4_disable_idmapping=
3244 [NFSv4] When set to the default of '1', this option
3245 ensures that both the RPC level authentication
3246 scheme and the NFS level operations agree to use
3247 numeric uids/gids if the mount is using the
3248 'sec=sys' security flavour. In effect it is
3249 disabling idmapping, which can make migration from
3250 legacy NFSv2/v3 systems to NFSv4 easier.
3251 Servers that do not support this mode of operation
3252 will be autodetected by the client, and it will fall
3253 back to using the idmapper.
3254 To turn off this behaviour, set the value to '0'.
3256 [NFS4] Specify an additional fixed unique ident-
3257 ification string that NFSv4 clients can insert into
3258 their nfs_client_id4 string. This is typically a
3259 UUID that is generated at system install time.
3261 nfs.send_implementation_id =
3262 [NFSv4.1] Send client implementation identification
3263 information in exchange_id requests.
3264 If zero, no implementation identification information
3266 The default is to send the implementation identification
3269 nfs.recover_lost_locks =
3270 [NFSv4] Attempt to recover locks that were lost due
3271 to a lease timeout on the server. Please note that
3272 doing this risks data corruption, since there are
3273 no guarantees that the file will remain unchanged
3274 after the locks are lost.
3275 If you want to enable the kernel legacy behaviour of
3276 attempting to recover these locks, then set this
3278 The default parameter value of '0' causes the kernel
3279 not to attempt recovery of lost locks.
3281 nfs4.layoutstats_timer =
3282 [NFSv4.2] Change the rate at which the kernel sends
3283 layoutstats to the pNFS metadata server.
3285 Setting this to value to 0 causes the kernel to use
3286 whatever value is the default set by the layout
3287 driver. A non-zero value sets the minimum interval
3288 in seconds between layoutstats transmissions.
3290 nfsd.nfs4_disable_idmapping=
3291 [NFSv4] When set to the default of '1', the NFSv4
3292 server will return only numeric uids and gids to
3293 clients using auth_sys, and will accept numeric uids
3294 and gids from such clients. This is intended to ease
3295 migration from NFSv2/v3.
3297 nmi_backtrace.backtrace_idle [KNL]
3298 Dump stacks even of idle CPUs in response to an
3299 NMI stack-backtrace request.
3301 nmi_debug= [KNL,SH] Specify one or more actions to take
3302 when a NMI is triggered.
3303 Format: [state][,regs][,debounce][,die]
3305 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3306 Format: [panic,][nopanic,][num]
3308 0 - turn hardlockup detector in nmi_watchdog off
3309 1 - turn hardlockup detector in nmi_watchdog on
3310 When panic is specified, panic when an NMI watchdog
3311 timeout occurs (or 'nopanic' to not panic on an NMI
3312 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3313 To disable both hard and soft lockup detectors,
3314 please see 'nowatchdog'.
3315 This is useful when you use a panic=... timeout and
3316 need the box quickly up again.
3318 These settings can be accessed at runtime via
3319 the nmi_watchdog and hardlockup_panic sysctls.
3321 netpoll.carrier_timeout=
3322 [NET] Specifies amount of time (in seconds) that
3323 netpoll should wait for a carrier. By default netpoll
3326 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3327 emulation library even if a 387 maths coprocessor
3330 no5lvl [X86-64] Disable 5-level paging mode. Forces
3331 kernel to use 4-level paging instead.
3333 nofsgsbase [X86] Disables FSGSBASE instructions.
3336 [HW] Never suspend the console
3337 Disable suspending of consoles during suspend and
3338 hibernate operations. Once disabled, debugging
3339 messages can reach various consoles while the rest
3340 of the system is being put to sleep (ie, while
3341 debugging driver suspend/resume hooks). This may
3342 not work reliably with all consoles, but is known
3343 to work with serial and VGA consoles.
3344 To facilitate more flexible debugging, we also add
3345 console_suspend, a printk module parameter to control
3346 it. Users could use console_suspend (usually
3347 /sys/module/printk/parameters/console_suspend) to
3348 turn on/off it dynamically.
3350 novmcoredd [KNL,KDUMP]
3351 Disable device dump. Device dump allows drivers to
3352 append dump data to vmcore so you can collect driver
3353 specified debug info. Drivers can append the data
3354 without any limit and this data is stored in memory,
3355 so this may cause significant memory stress. Disabling
3356 device dump can help save memory but the driver debug
3357 data will be no longer available. This parameter
3358 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3361 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3362 caches in the slab allocator. Saves per-node memory,
3363 but will impact performance.
3367 noaltinstr [S390] Disables alternative instructions patching
3368 (CPU alternatives feature).
3370 noapic [SMP,APIC] Tells the kernel to not make use of any
3371 IOAPICs that may be present in the system.
3373 noautogroup Disable scheduler automatic task group creation.
3375 nobats [PPC] Do not use BATs for mapping kernel lowmem
3376 on "Classic" PPC cores.
3380 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3382 delayacct [KNL] Enable per-task delay accounting
3384 nodsp [SH] Disable hardware DSP at boot time.
3386 noefi Disable EFI runtime services support.
3388 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3393 On X86-32 available only on PAE configured kernels.
3394 noexec=on: enable non-executable mappings (default)
3395 noexec=off: disable non-executable mappings
3398 Disable SMAP (Supervisor Mode Access Prevention)
3399 even if it is supported by processor.
3402 Disable SMEP (Supervisor Mode Execution Prevention)
3403 even if it is supported by processor.
3406 This affects only 32-bit executables.
3407 noexec32=on: enable non-executable mappings (default)
3408 read doesn't imply executable mappings
3409 noexec32=off: disable non-executable mappings
3410 read implies executable mappings
3412 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3414 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3415 register save and restore. The kernel will only save
3416 legacy floating-point registers on task switch.
3418 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3420 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3422 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3423 Equivalent to smt=1.
3425 [KNL,X86] Disable symmetric multithreading (SMT).
3426 nosmt=force: Force disable SMT, cannot be undone
3427 via the sysfs control file.
3429 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3430 (bounds check bypass). With this option data leaks are
3431 possible in the system.
3433 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3434 the Spectre variant 2 (indirect branch prediction)
3435 vulnerability. System may allow data leaks with this
3438 nospec_store_bypass_disable
3439 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3442 [PPC] Don't flush the L1-D cache after accessing user data.
3444 noxsave [BUGS=X86] Disables x86 extended register state save
3445 and restore using xsave. The kernel will fallback to
3446 enabling legacy floating-point and sse state.
3448 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3449 register states. The kernel will fall back to use
3450 xsave to save the states. By using this parameter,
3451 performance of saving the states is degraded because
3452 xsave doesn't support modified optimization while
3453 xsaveopt supports it on xsaveopt enabled systems.
3455 noxsaves [X86] Disables xsaves and xrstors used in saving and
3456 restoring x86 extended register state in compacted
3457 form of xsave area. The kernel will fall back to use
3458 xsaveopt and xrstor to save and restore the states
3459 in standard form of xsave area. By using this
3460 parameter, xsave area per process might occupy more
3461 memory on xsaves enabled systems.
3463 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3464 in do_idle() and not use the arch_cpu_idle()
3465 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3466 to be effective. This is useful on platforms where the
3467 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3468 correctly or when doing power measurements to evalute
3469 the impact of the sleep instructions. This is also
3470 useful when using JTAG debugger.
3472 no_file_caps Tells the kernel not to honor file capabilities. The
3473 only way then for a file to be executed with privilege
3474 is to be setuid root or executed by root.
3476 nohalt [IA-64] Tells the kernel not to use the power saving
3477 function PAL_HALT_LIGHT when idle. This increases
3478 power-consumption. On the positive side, it reduces
3479 interrupt wake-up latency, which may improve performance
3480 in certain environments such as networked servers or
3484 Force pointers printed to the console or buffers to be
3485 unhashed. By default, when a pointer is printed via %p
3486 format string, that pointer is "hashed", i.e. obscured
3487 by hashing the pointer value. This is a security feature
3488 that hides actual kernel addresses from unprivileged
3489 users, but it also makes debugging the kernel more
3490 difficult since unequal pointers can no longer be
3491 compared. However, if this command-line option is
3492 specified, then all normal pointers will have their true
3493 value printed. This option should only be specified when
3494 debugging the kernel. Please do not use on production
3497 nohibernate [HIBERNATION] Disable hibernation and resume.
3499 nohz= [KNL] Boottime enable/disable dynamic ticks
3500 Valid arguments: on, off
3503 nohz_full= [KNL,BOOT,SMP,ISOL]
3504 The argument is a cpu list, as described above.
3505 In kernels built with CONFIG_NO_HZ_FULL=y, set
3506 the specified list of CPUs whose tick will be stopped
3507 whenever possible. The boot CPU will be forced outside
3508 the range to maintain the timekeeping. Any CPUs
3509 in this list will have their RCU callbacks offloaded,
3510 just as if they had also been called out in the
3511 rcu_nocbs= boot parameter.
3513 noiotrap [SH] Disables trapped I/O port accesses.
3515 noirqdebug [X86-32] Disables the code which attempts to detect and
3516 disable unhandled interrupt sources.
3518 no_timer_check [X86,APIC] Disables the code which tests for
3519 broken timer IRQ sources.
3521 noisapnp [ISAPNP] Disables ISA PnP code.
3523 noinitrd [RAM] Tells the kernel not to load any configured
3526 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3528 [Deprecated - use intremap=off]
3532 noinvpcid [X86] Disable the INVPCID cpu feature.
3534 nojitter [IA-64] Disables jitter checking for ITC timers.
3536 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3538 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3542 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3543 clock and use the default one.
3545 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3546 accounting. steal time is computed, but won't
3547 influence scheduler behaviour
3549 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3551 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3553 noltlbs [PPC] Do not use large page/tlb entries for kernel
3554 lowmem mapping on PPC40x and PPC8xx
3556 nomca [IA-64] Disable machine check abort handling
3558 nomce [X86-32] Disable Machine Check Exception
3560 nomfgpt [X86-32] Disable Multi-Function General Purpose
3561 Timer usage (for AMD Geode machines).
3563 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3564 shutdown the other cpus. Instead use the REBOOT_VECTOR
3567 nomodule Disable module load
3569 nopat [X86] Disable PAT (page attribute table extension of
3570 pagetables) support.
3572 nopcid [X86-64] Disable the PCID cpu feature.
3574 norandmaps Don't use address space randomization. Equivalent to
3575 echo 0 > /proc/sys/kernel/randomize_va_space
3577 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3578 with UP alternatives
3580 nordrand [X86] Disable kernel use of the RDRAND and
3581 RDSEED instructions even if they are supported
3582 by the processor. RDRAND and RDSEED are still
3583 available to user space applications.
3585 noresume [SWSUSP] Disables resume and restores original swap
3588 no-scroll [VGA] Disables scrollback.
3589 This is required for the Braillex ib80-piezo Braille
3590 reader made by F.H. Papenmeier (Germany).
3594 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3596 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3598 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3599 and disable the IO APIC. legacy for "maxcpus=0".
3601 nosoftlockup [KNL] Disable the soft-lockup detector.
3603 nosync [HW,M68K] Disables sync negotiation for all devices.
3605 nowatchdog [KNL] Disable both lockup detectors, i.e.
3606 soft-lockup and NMI watchdog (hard-lockup).
3610 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3612 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3613 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3614 Some features depend on CPU0. Known dependencies are:
3615 1. Resume from suspend/hibernate depends on CPU0.
3616 Suspend/hibernate will fail if CPU0 is offline and you
3617 need to online CPU0 before suspend/hibernate.
3618 2. PIC interrupts also depend on CPU0. CPU0 can't be
3619 removed if a PIC interrupt is detected.
3620 It's said poweroff/reboot may depend on CPU0 on some
3621 machines although I haven't seen such issues so far
3622 after CPU0 is offline on a few tested machines.
3623 If the dependencies are under your control, you can
3624 turn on cpu0_hotplug.
3626 nps_mtm_hs_ctr= [KNL,ARC]
3627 This parameter sets the maximum duration, in
3628 cycles, each HW thread of the CTOP can run
3629 without interruptions, before HW switches it.
3630 The actual maximum duration is 16 times this
3632 Format: integer between 1 and 255
3635 nptcg= [IA-64] Override max number of concurrent global TLB
3636 purges which is reported from either PAL_VM_SUMMARY or
3639 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3640 could support. nr_cpus=n : n >= 1 limits the kernel to
3641 support 'n' processors. It could be larger than the
3642 number of already plugged CPU during bootup, later in
3643 runtime you can physically add extra cpu until it reaches
3644 n. So during boot up some boot time memory for per-cpu
3645 variables need be pre-allocated for later physical cpu
3648 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3650 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3651 set up a single NUMA node spanning all memory.
3653 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3655 Allowed values are enable and disable
3657 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3658 'node', 'default' can be specified
3659 This can be set from sysctl after boot.
3660 See Documentation/admin-guide/sysctl/vm.rst for details.
3662 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3663 See Documentation/core-api/debugging-via-ohci1394.rst for more
3666 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3667 Rather than timing out after 20 ms if an EC
3668 command is not properly ACKed, override the length
3669 of the timeout. We have interrupts disabled while
3670 waiting for the ACK, so if this is set too high
3671 interrupts *may* be lost!
3673 omap_mux= [OMAP] Override bootloader pin multiplexing.
3674 Format: <mux_mode0.mode_name=value>...
3675 For example, to override I2C bus2:
3676 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3678 oops=panic Always panic on oopses. Default is to just kill the
3679 process, but there is a small probability of
3680 deadlocking the machine.
3681 This will also cause panics on machine check exceptions.
3682 Useful together with panic=30 to trigger a reboot.
3685 [KNL] Boolean flag to control whether the page allocator
3686 should randomize its free lists. The randomization may
3687 be automatically enabled if the kernel detects it is
3688 running on a platform with a direct-mapped memory-side
3689 cache, and this parameter can be used to
3690 override/disable that behavior. The state of the flag
3691 can be read from sysfs at:
3692 /sys/module/page_alloc/parameters/shuffle.
3694 page_owner= [KNL] Boot-time page_owner enabling option.
3695 Storage of the information about who allocated
3696 each page is disabled in default. With this switch,
3698 on: enable the feature
3700 page_poison= [KNL] Boot-time parameter changing the state of
3701 poisoning on the buddy allocator, available with
3702 CONFIG_PAGE_POISONING=y.
3703 off: turn off poisoning (default)
3704 on: turn on poisoning
3706 page_reporting.page_reporting_order=
3707 [KNL] Minimal page reporting order
3709 Adjust the minimal page reporting order. The page
3710 reporting is disabled when it exceeds (MAX_ORDER-1).
3712 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3713 timeout > 0: seconds before rebooting
3714 timeout = 0: wait forever
3715 timeout < 0: reboot immediately
3718 panic_print= Bitmask for printing system info when panic happens.
3719 User can chose combination of the following bits:
3720 bit 0: print all tasks info
3721 bit 1: print system memory info
3722 bit 2: print timer info
3723 bit 3: print locks info if CONFIG_LOCKDEP is on
3724 bit 4: print ftrace buffer
3725 bit 5: print all printk messages in buffer
3727 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3728 Format: <hex>[,nousertaint]
3729 Hexadecimal bitmask representing the set of TAINT flags
3730 that will cause the kernel to panic when add_taint() is
3731 called with any of the flags in this set.
3732 The optional switch "nousertaint" can be utilized to
3733 prevent userspace forced crashes by writing to sysctl
3734 /proc/sys/kernel/tainted any flagset matching with the
3735 bitmask set on panic_on_taint.
3736 See Documentation/admin-guide/tainted-kernels.rst for
3737 extra details on the taint flags that users can pick
3738 to compose the bitmask to assign to panic_on_taint.
3740 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3743 crash_kexec_post_notifiers
3744 Run kdump after running panic-notifiers and dumping
3745 kmsg. This only for the users who doubt kdump always
3746 succeeds in any situation.
3747 Note that this also increases risks of kdump failure,
3748 because some panic notifiers can make the crashed
3749 kernel more unstable.
3751 parkbd.port= [HW] Parallel port number the keyboard adapter is
3752 connected to, default is 0.
3754 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3755 0 for XT, 1 for AT (default is AT).
3758 parport= [HW,PPT] Specify parallel ports. 0 disables.
3759 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3760 Use 'auto' to force the driver to use any
3761 IRQ/DMA settings detected (the default is to
3762 ignore detected IRQ/DMA settings because of
3763 possible conflicts). You can specify the base
3764 address, IRQ, and DMA settings; IRQ and DMA
3765 should be numbers, or 'auto' (for using detected
3766 settings on that particular port), or 'nofifo'
3767 (to avoid using a FIFO even if it is detected).
3768 Parallel ports are assigned in the order they
3769 are specified on the command line, starting
3772 parport_init_mode= [HW,PPT]
3773 Configure VIA parallel port to operate in
3774 a specific mode. This is necessary on Pegasos
3775 computer where firmware has no options for setting
3776 up parallel port mode and sets it to spp.
3777 Currently this function knows 686a and 8231 chips.
3778 Format: [spp|ps2|epp|ecp|ecpepp]
3780 pata_legacy.all= [HW,LIBATA]
3782 Set to non-zero to probe primary and secondary ISA
3783 port ranges on PCI systems where no PCI PATA device
3784 has been found at either range. Disabled by default.
3786 pata_legacy.autospeed= [HW,LIBATA]
3788 Set to non-zero if a chip is present that snoops speed
3789 changes. Disabled by default.
3791 pata_legacy.ht6560a= [HW,LIBATA]
3793 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3794 the secondary channel, or both channels respectively.
3795 Disabled by default.
3797 pata_legacy.ht6560b= [HW,LIBATA]
3799 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3800 the secondary channel, or both channels respectively.
3801 Disabled by default.
3803 pata_legacy.iordy_mask= [HW,LIBATA]
3805 IORDY enable mask. Set individual bits to allow IORDY
3806 for the respective channel. Bit 0 is for the first
3807 legacy channel handled by this driver, bit 1 is for
3808 the second channel, and so on. The sequence will often
3809 correspond to the primary legacy channel, the secondary
3810 legacy channel, and so on, but the handling of a PCI
3811 bus and the use of other driver options may interfere
3812 with the sequence. By default IORDY is allowed across
3815 pata_legacy.opti82c46x= [HW,LIBATA]
3817 Set to 1, 2, or 3 for Opti 82c611A on the primary
3818 channel, the secondary channel, or both channels
3819 respectively. Disabled by default.
3821 pata_legacy.opti82c611a= [HW,LIBATA]
3823 Set to 1, 2, or 3 for Opti 82c465MV on the primary
3824 channel, the secondary channel, or both channels
3825 respectively. Disabled by default.
3827 pata_legacy.pio_mask= [HW,LIBATA]
3829 PIO mode mask for autospeed devices. Set individual
3830 bits to allow the use of the respective PIO modes.
3831 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
3832 All modes allowed by default.
3834 pata_legacy.probe_all= [HW,LIBATA]
3836 Set to non-zero to probe tertiary and further ISA
3837 port ranges on PCI systems. Disabled by default.
3839 pata_legacy.probe_mask= [HW,LIBATA]
3841 Probe mask for legacy ISA PATA ports. Depending on
3842 platform configuration and the use of other driver
3843 options up to 6 legacy ports are supported: 0x1f0,
3844 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
3845 of individual ports can be disabled by setting the
3846 corresponding bits in the mask to 1. Bit 0 is for
3847 the first port in the list above (0x1f0), and so on.
3848 By default all supported ports are probed.
3850 pata_legacy.qdi= [HW,LIBATA]
3852 Set to non-zero to probe QDI controllers. By default
3853 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
3855 pata_legacy.winbond= [HW,LIBATA]
3857 Set to non-zero to probe Winbond controllers. Use
3858 the standard I/O port (0x130) if 1, otherwise the
3859 value given is the I/O port to use (typically 0x1b0).
3860 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
3863 pata_platform.pio_mask= [HW,LIBATA]
3865 Supported PIO mode mask. Set individual bits to allow
3866 the use of the respective PIO modes. Bit 0 is for
3867 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
3871 Halt all CPUs after the first oops has been printed for
3872 the specified number of seconds. This is to be used if
3873 your oopses keep scrolling off the screen.
3878 See header of drivers/block/paride/pcd.c.
3879 See also Documentation/admin-guide/blockdev/paride.rst.
3881 pci=option[,option...] [PCI] various PCI subsystem options.
3883 Some options herein operate on a specific device
3884 or a set of devices (<pci_dev>). These are
3885 specified in one of the following formats:
3887 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3888 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3890 Note: the first format specifies a PCI
3891 bus/device/function address which may change
3892 if new hardware is inserted, if motherboard
3893 firmware changes, or due to changes caused
3894 by other kernel parameters. If the
3895 domain is left unspecified, it is
3896 taken to be zero. Optionally, a path
3897 to a device through multiple device/function
3898 addresses can be specified after the base
3899 address (this is more robust against
3900 renumbering issues). The second format
3901 selects devices using IDs from the
3902 configuration space which may match multiple
3903 devices in the system.
3905 earlydump dump PCI config space before the kernel
3907 off [X86] don't probe for the PCI bus
3908 bios [X86-32] force use of PCI BIOS, don't access
3909 the hardware directly. Use this if your machine
3910 has a non-standard PCI host bridge.
3911 nobios [X86-32] disallow use of PCI BIOS, only direct
3912 hardware access methods are allowed. Use this
3913 if you experience crashes upon bootup and you
3914 suspect they are caused by the BIOS.
3915 conf1 [X86] Force use of PCI Configuration Access
3916 Mechanism 1 (config address in IO port 0xCF8,
3917 data in IO port 0xCFC, both 32-bit).
3918 conf2 [X86] Force use of PCI Configuration Access
3919 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3920 the function, IO port 0xCFA, also 8-bit, sets
3921 bus number. The config space is then accessed
3922 through ports 0xC000-0xCFFF).
3923 See http://wiki.osdev.org/PCI for more info
3924 on the configuration access mechanisms.
3925 noaer [PCIE] If the PCIEAER kernel config parameter is
3926 enabled, this kernel boot option can be used to
3927 disable the use of PCIE advanced error reporting.
3928 nodomains [PCI] Disable support for multiple PCI
3929 root domains (aka PCI segments, in ACPI-speak).
3930 nommconf [X86] Disable use of MMCONFIG for PCI
3932 check_enable_amd_mmconf [X86] check for and enable
3933 properly configured MMIO access to PCI
3934 config space on AMD family 10h CPU
3935 nomsi [MSI] If the PCI_MSI kernel config parameter is
3936 enabled, this kernel boot option can be used to
3937 disable the use of MSI interrupts system-wide.
3938 noioapicquirk [APIC] Disable all boot interrupt quirks.
3939 Safety option to keep boot IRQs enabled. This
3940 should never be necessary.
3941 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3942 primary IO-APIC for bridges that cannot disable
3943 boot IRQs. This fixes a source of spurious IRQs
3944 when the system masks IRQs.
3945 noioapicreroute [APIC] Disable workaround that uses the
3946 boot IRQ equivalent of an IRQ that connects to
3947 a chipset where boot IRQs cannot be disabled.
3948 The opposite of ioapicreroute.
3949 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3950 routing table. These calls are known to be buggy
3951 on several machines and they hang the machine
3952 when used, but on other computers it's the only
3953 way to get the interrupt routing table. Try
3954 this option if the kernel is unable to allocate
3955 IRQs or discover secondary PCI buses on your
3957 rom [X86] Assign address space to expansion ROMs.
3958 Use with caution as certain devices share
3959 address decoders between ROMs and other
3961 norom [X86] Do not assign address space to
3962 expansion ROMs that do not already have
3963 BIOS assigned address ranges.
3964 nobar [X86] Do not assign address space to the
3965 BARs that weren't assigned by the BIOS.
3966 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3967 assigned automatically to PCI devices. You can
3968 make the kernel exclude IRQs of your ISA cards
3970 pirqaddr=0xAAAAA [X86] Specify the physical address
3971 of the PIRQ table (normally generated
3972 by the BIOS) if it is outside the
3973 F0000h-100000h range.
3974 lastbus=N [X86] Scan all buses thru bus #N. Can be
3975 useful if the kernel is unable to find your
3976 secondary buses and you want to tell it
3977 explicitly which ones they are.
3978 assign-busses [X86] Always assign all PCI bus
3979 numbers ourselves, overriding
3980 whatever the firmware may have done.
3981 usepirqmask [X86] Honor the possible IRQ mask stored
3982 in the BIOS $PIR table. This is needed on
3983 some systems with broken BIOSes, notably
3984 some HP Pavilion N5400 and Omnibook XE3
3985 notebooks. This will have no effect if ACPI
3986 IRQ routing is enabled.
3987 noacpi [X86] Do not use ACPI for IRQ routing
3988 or for PCI scanning.
3989 use_crs [X86] Use PCI host bridge window information
3990 from ACPI. On BIOSes from 2008 or later, this
3991 is enabled by default. If you need to use this,
3992 please report a bug.
3993 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3994 If you need to use this, please report a bug.
3995 routeirq Do IRQ routing for all PCI devices.
3996 This is normally done in pci_enable_device(),
3997 so this option is a temporary workaround
3998 for broken drivers that don't call it.
3999 skip_isa_align [X86] do not align io start addr, so can
4000 handle more pci cards
4001 noearly [X86] Don't do any early type 1 scanning.
4002 This might help on some broken boards which
4003 machine check when some devices' config space
4004 is read. But various workarounds are disabled
4005 and some IOMMU drivers will not work.
4006 bfsort Sort PCI devices into breadth-first order.
4007 This sorting is done to get a device
4008 order compatible with older (<= 2.4) kernels.
4009 nobfsort Don't sort PCI devices into breadth-first order.
4010 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4011 tuning and use the BIOS-configured MPS defaults.
4012 pcie_bus_safe Set every device's MPS to the largest value
4013 supported by all devices below the root complex.
4014 pcie_bus_perf Set device MPS to the largest allowable MPS
4015 based on its parent bus. Also set MRRS (Max
4016 Read Request Size) to the largest supported
4017 value (no larger than the MPS that the device
4018 or bus can support) for best performance.
4019 pcie_bus_peer2peer Set every device's MPS to 128B, which
4020 every device is guaranteed to support. This
4021 configuration allows peer-to-peer DMA between
4022 any pair of devices, possibly at the cost of
4023 reduced performance. This also guarantees
4024 that hot-added devices will work.
4025 cbiosize=nn[KMG] The fixed amount of bus space which is
4026 reserved for the CardBus bridge's IO window.
4027 The default value is 256 bytes.
4028 cbmemsize=nn[KMG] The fixed amount of bus space which is
4029 reserved for the CardBus bridge's memory
4030 window. The default value is 64 megabytes.
4033 [<order of align>@]<pci_dev>[; ...]
4034 Specifies alignment and device to reassign
4035 aligned memory resources. How to
4036 specify the device is described above.
4037 If <order of align> is not specified,
4038 PAGE_SIZE is used as alignment.
4039 A PCI-PCI bridge can be specified if resource
4040 windows need to be expanded.
4041 To specify the alignment for several
4042 instances of a device, the PCI vendor,
4043 device, subvendor, and subdevice may be
4044 specified, e.g., 12@pci:8086:9c22:103c:198f
4045 for 4096-byte alignment.
4046 ecrc= Enable/disable PCIe ECRC (transaction layer
4047 end-to-end CRC checking).
4048 bios: Use BIOS/firmware settings. This is the
4052 hpiosize=nn[KMG] The fixed amount of bus space which is
4053 reserved for hotplug bridge's IO window.
4054 Default size is 256 bytes.
4055 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4056 reserved for hotplug bridge's MMIO window.
4057 Default size is 2 megabytes.
4058 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4059 reserved for hotplug bridge's MMIO_PREF window.
4060 Default size is 2 megabytes.
4061 hpmemsize=nn[KMG] The fixed amount of bus space which is
4062 reserved for hotplug bridge's MMIO and
4064 Default size is 2 megabytes.
4065 hpbussize=nn The minimum amount of additional bus numbers
4066 reserved for buses below a hotplug bridge.
4068 realloc= Enable/disable reallocating PCI bridge resources
4069 if allocations done by BIOS are too small to
4070 accommodate resources required by all child
4072 off: Turn realloc off
4074 realloc same as realloc=on
4075 noari do not use PCIe ARI.
4076 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4077 do not use PCIe ATS (and IOMMU device IOTLB).
4078 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4079 only look for one device below a PCIe downstream
4081 big_root_window Try to add a big 64bit memory window to the PCIe
4082 root complex on AMD CPUs. Some GFX hardware
4083 can resize a BAR to allow access to all VRAM.
4084 Adding the window is slightly risky (it may
4085 conflict with unreported devices), so this
4087 disable_acs_redir=<pci_dev>[; ...]
4088 Specify one or more PCI devices (in the format
4089 specified above) separated by semicolons.
4090 Each device specified will have the PCI ACS
4091 redirect capabilities forced off which will
4092 allow P2P traffic between devices through
4093 bridges without forcing it upstream. Note:
4094 this removes isolation between devices and
4095 may put more devices in an IOMMU group.
4096 force_floating [S390] Force usage of floating interrupts.
4097 nomio [S390] Do not use MIO instructions.
4098 norid [S390] ignore the RID field and force use of
4099 one PCI domain per PCI function
4101 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4104 force Enable ASPM even on devices that claim not to support it.
4105 WARNING: Forcing ASPM on may cause system lockups.
4107 pcie_ports= [PCIE] PCIe port services handling:
4108 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4109 even if the platform doesn't give the OS permission to
4110 use them. This may cause conflicts if the platform
4111 also tries to use these services.
4112 dpc-native Use native PCIe service for DPC only. May
4113 cause conflicts if firmware uses AER or DPC.
4114 compat Disable native PCIe services (PME, AER, DPC, PCIe
4117 pcie_port_pm= [PCIE] PCIe port power management handling:
4118 off Disable power management of all PCIe ports
4119 force Forcibly enable power management of all PCIe ports
4121 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4122 nomsi Do not use MSI for native PCIe PME signaling (this makes
4123 all PCIe root ports use INTx for all services).
4125 pcmv= [HW,PCMCIA] BadgePAD 4
4129 Keep all power-domains already enabled by bootloader on,
4130 even if no driver has claimed them. This is useful
4131 for debug and development, but should not be
4132 needed on a platform with proper driver support.
4135 See Documentation/admin-guide/blockdev/paride.rst.
4137 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4140 See arch/parisc/kernel/pdc_chassis.c
4142 percpu_alloc= Select which percpu first chunk allocator to use.
4143 Currently supported values are "embed" and "page".
4144 Archs may support subset or none of the selections.
4145 See comments in mm/percpu.c for details on each
4146 allocator. This parameter is primarily for debugging
4147 and performance comparison.
4150 See Documentation/admin-guide/blockdev/paride.rst.
4153 See Documentation/admin-guide/blockdev/paride.rst.
4155 pirq= [SMP,APIC] Manual mp-table setup
4156 See Documentation/x86/i386/IO-APIC.rst.
4158 plip= [PPT,NET] Parallel port network link
4159 Format: { parport<nr> | timid | 0 }
4160 See also Documentation/admin-guide/parport.rst.
4162 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4163 Override pmtimer IOPort with a hex value.
4166 pm_debug_messages [SUSPEND,KNL]
4167 Enable suspend/resume debug messages during boot up.
4170 Enable PNP debug messages (depends on the
4171 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4172 via /sys/module/pnp/parameters/debug. We always show
4173 current resource usage; turning this on also shows
4174 possible settings and some assignment information.
4180 { on | off | curr | res | no-curr | no-res }
4183 [ISAPNP] Exclude IRQs for the autoconfiguration
4186 [ISAPNP] Exclude DMAs for the autoconfiguration
4188 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4189 Ranges are in pairs (I/O port base and size).
4192 [ISAPNP] Exclude memory regions for the
4194 Ranges are in pairs (memory base and size).
4196 ports= [IP_VS_FTP] IPVS ftp helper module
4198 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4200 Format: <port>,<port>....
4202 powersave=off [PPC] This option disables power saving features.
4203 It specifically disables cpuidle and sets the
4204 platform machine description specific power_save
4205 function to NULL. On Idle the CPU just reduces
4208 ppc_strict_facility_enable
4209 [PPC] This option catches any kernel floating point,
4210 Altivec, VSX and SPE outside of regions specifically
4211 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4212 There is some performance impact when enabling this.
4216 Disable Hardware Transactional Memory
4219 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4220 none - Limited to cond_resched() calls
4221 voluntary - Limited to cond_resched() and might_sleep() calls
4222 full - Any section that isn't explicitly preempt disabled
4223 can be preempted anytime.
4225 print-fatal-signals=
4226 [KNL] debug: print fatal signals
4228 If enabled, warn about various signal handling
4229 related application anomalies: too many signals,
4230 too many POSIX.1 timers, fatal signals causing a
4233 If you hit the warning due to signal overflow,
4234 you might want to try "ulimit -i unlimited".
4238 printk.always_kmsg_dump=
4239 Trigger kmsg_dump for cases other than kernel oops or
4241 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4244 printk.console_no_auto_verbose=
4245 Disable console loglevel raise on oops, panic
4246 or lockdep-detected issues (only if lock debug is on).
4247 With an exception to setups with low baudrate on
4248 serial console, keeping this 0 is a good choice
4249 in order to provide more debug information.
4251 default: 0 (auto_verbose is enabled)
4253 printk.devkmsg={on,off,ratelimit}
4254 Control writing to /dev/kmsg.
4255 on - unlimited logging to /dev/kmsg from userspace
4256 off - logging to /dev/kmsg disabled
4257 ratelimit - ratelimit the logging
4260 printk.time= Show timing data prefixed to each printk message line
4261 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4263 processor.max_cstate= [HW,ACPI]
4264 Limit processor to maximum C-state
4265 max_cstate=9 overrides any DMI blacklist limit.
4267 processor.nocst [HW,ACPI]
4268 Ignore the _CST method to determine C-states,
4269 instead using the legacy FADT method
4271 profile= [KNL] Enable kernel profiling via /proc/profile
4272 Format: [<profiletype>,]<number>
4273 Param: <profiletype>: "schedule", "sleep", or "kvm"
4274 [defaults to kernel profiling]
4275 Param: "schedule" - profile schedule points.
4276 Param: "sleep" - profile D-state sleeping (millisecs).
4277 Requires CONFIG_SCHEDSTATS
4278 Param: "kvm" - profile VM exits.
4279 Param: <number> - step/bucket size as a power of 2 for
4280 statistical time based profiling.
4282 prompt_ramdisk= [RAM] [Deprecated]
4284 prot_virt= [S390] enable hosting protected virtual machines
4285 isolated from the hypervisor (if hardware supports
4289 psi= [KNL] Enable or disable pressure stall information
4293 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4294 probe for; one of (bare|imps|exps|lifebook|any).
4295 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4297 psmouse.resetafter= [HW,MOUSE]
4298 Try to reset the device after so many bad packets
4301 [HW,MOUSE] Set desired mouse resolution, in dpi.
4302 psmouse.smartscroll=
4303 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4304 0 = disabled, 1 = enabled (default).
4306 pstore.backend= Specify the name of the pstore backend to use
4309 See Documentation/admin-guide/blockdev/paride.rst.
4311 pti= [X86-64] Control Page Table Isolation of user and
4312 kernel address spaces. Disabling this feature
4313 removes hardening, but improves performance of
4314 system calls and interrupts.
4316 on - unconditionally enable
4317 off - unconditionally disable
4318 auto - kernel detects whether your CPU model is
4319 vulnerable to issues that PTI mitigates
4321 Not specifying this option is equivalent to pti=auto.
4324 Equivalent to pti=off
4327 [KNL] Number of legacy pty's. Overwrites compiled-in
4330 quiet [KNL] Disable most log messages
4335 See Documentation/admin-guide/md.rst.
4337 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4338 See Documentation/admin-guide/blockdev/ramdisk.rst.
4340 ramdisk_start= [RAM] RAM disk image start address
4342 random.trust_cpu={on,off}
4343 [KNL] Enable or disable trusting the use of the
4344 CPU's random number generator (if available) to
4345 fully seed the kernel's CRNG. Default is controlled
4346 by CONFIG_RANDOM_TRUST_CPU.
4348 random.trust_bootloader={on,off}
4349 [KNL] Enable or disable trusting the use of a
4350 seed passed by the bootloader (if available) to
4351 fully seed the kernel's CRNG. Default is controlled
4352 by CONFIG_RANDOM_TRUST_BOOTLOADER.
4354 randomize_kstack_offset=
4355 [KNL] Enable or disable kernel stack offset
4356 randomization, which provides roughly 5 bits of
4357 entropy, frustrating memory corruption attacks
4358 that depend on stack address determinism or
4359 cross-syscall address exposures. This is only
4360 available on architectures that have defined
4361 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4362 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4363 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4365 ras=option[,option,...] [KNL] RAS-specific options
4368 Disable the Correctable Errors Collector,
4369 see CONFIG_RAS_CEC help text.
4372 The argument is a cpu list, as described above.
4374 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
4375 the specified list of CPUs to be no-callback CPUs.
4376 Invocation of these CPUs' RCU callbacks will be
4377 offloaded to "rcuox/N" kthreads created for that
4378 purpose, where "x" is "p" for RCU-preempt, and
4379 "s" for RCU-sched, and "N" is the CPU number.
4380 This reduces OS jitter on the offloaded CPUs,
4381 which can be useful for HPC and real-time
4382 workloads. It can also improve energy efficiency
4383 for asymmetric multiprocessors.
4386 Rather than requiring that offloaded CPUs
4387 (specified by rcu_nocbs= above) explicitly
4388 awaken the corresponding "rcuoN" kthreads,
4389 make these kthreads poll for callbacks.
4390 This improves the real-time response for the
4391 offloaded CPUs by relieving them of the need to
4392 wake up the corresponding kthread, but degrades
4393 energy efficiency by requiring that the kthreads
4394 periodically wake up to do the polling.
4396 rcutree.blimit= [KNL]
4397 Set maximum number of finished RCU callbacks to
4398 process in one batch.
4400 rcutree.dump_tree= [KNL]
4401 Dump the structure of the rcu_node combining tree
4402 out at early boot. This is used for diagnostic
4403 purposes, to verify correct tree setup.
4405 rcutree.gp_cleanup_delay= [KNL]
4406 Set the number of jiffies to delay each step of
4407 RCU grace-period cleanup.
4409 rcutree.gp_init_delay= [KNL]
4410 Set the number of jiffies to delay each step of
4411 RCU grace-period initialization.
4413 rcutree.gp_preinit_delay= [KNL]
4414 Set the number of jiffies to delay each step of
4415 RCU grace-period pre-initialization, that is,
4416 the propagation of recent CPU-hotplug changes up
4417 the rcu_node combining tree.
4419 rcutree.use_softirq= [KNL]
4420 If set to zero, move all RCU_SOFTIRQ processing to
4421 per-CPU rcuc kthreads. Defaults to a non-zero
4422 value, meaning that RCU_SOFTIRQ is used by default.
4423 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4425 But note that CONFIG_PREEMPT_RT=y kernels disable
4426 this kernel boot parameter, forcibly setting it
4429 rcutree.rcu_fanout_exact= [KNL]
4430 Disable autobalancing of the rcu_node combining
4431 tree. This is used by rcutorture, and might
4432 possibly be useful for architectures having high
4433 cache-to-cache transfer latencies.
4435 rcutree.rcu_fanout_leaf= [KNL]
4436 Change the number of CPUs assigned to each
4437 leaf rcu_node structure. Useful for very
4438 large systems, which will choose the value 64,
4439 and for NUMA systems with large remote-access
4440 latencies, which will choose a value aligned
4441 with the appropriate hardware boundaries.
4443 rcutree.rcu_min_cached_objs= [KNL]
4444 Minimum number of objects which are cached and
4445 maintained per one CPU. Object size is equal
4446 to PAGE_SIZE. The cache allows to reduce the
4447 pressure to page allocator, also it makes the
4448 whole algorithm to behave better in low memory
4451 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4452 Set the page-cache refill delay (in milliseconds)
4453 in response to low-memory conditions. The range
4454 of permitted values is in the range 0:100000.
4456 rcutree.jiffies_till_first_fqs= [KNL]
4457 Set delay from grace-period initialization to
4458 first attempt to force quiescent states.
4459 Units are jiffies, minimum value is zero,
4460 and maximum value is HZ.
4462 rcutree.jiffies_till_next_fqs= [KNL]
4463 Set delay between subsequent attempts to force
4464 quiescent states. Units are jiffies, minimum
4465 value is one, and maximum value is HZ.
4467 rcutree.jiffies_till_sched_qs= [KNL]
4468 Set required age in jiffies for a
4469 given grace period before RCU starts
4470 soliciting quiescent-state help from
4471 rcu_note_context_switch() and cond_resched().
4472 If not specified, the kernel will calculate
4473 a value based on the most recent settings
4474 of rcutree.jiffies_till_first_fqs
4475 and rcutree.jiffies_till_next_fqs.
4476 This calculated value may be viewed in
4477 rcutree.jiffies_to_sched_qs. Any attempt to set
4478 rcutree.jiffies_to_sched_qs will be cheerfully
4481 rcutree.kthread_prio= [KNL,BOOT]
4482 Set the SCHED_FIFO priority of the RCU per-CPU
4483 kthreads (rcuc/N). This value is also used for
4484 the priority of the RCU boost threads (rcub/N)
4485 and for the RCU grace-period kthreads (rcu_bh,
4486 rcu_preempt, and rcu_sched). If RCU_BOOST is
4487 set, valid values are 1-99 and the default is 1
4488 (the least-favored priority). Otherwise, when
4489 RCU_BOOST is not set, valid values are 0-99 and
4490 the default is zero (non-realtime operation).
4492 rcutree.rcu_nocb_gp_stride= [KNL]
4493 Set the number of NOCB callback kthreads in
4494 each group, which defaults to the square root
4495 of the number of CPUs. Larger numbers reduce
4496 the wakeup overhead on the global grace-period
4497 kthread, but increases that same overhead on
4498 each group's NOCB grace-period kthread.
4500 rcutree.qhimark= [KNL]
4501 Set threshold of queued RCU callbacks beyond which
4502 batch limiting is disabled.
4504 rcutree.qlowmark= [KNL]
4505 Set threshold of queued RCU callbacks below which
4506 batch limiting is re-enabled.
4508 rcutree.qovld= [KNL]
4509 Set threshold of queued RCU callbacks beyond which
4510 RCU's force-quiescent-state scan will aggressively
4511 enlist help from cond_resched() and sched IPIs to
4512 help CPUs more quickly reach quiescent states.
4513 Set to less than zero to make this be set based
4514 on rcutree.qhimark at boot time and to zero to
4515 disable more aggressive help enlistment.
4517 rcutree.rcu_idle_gp_delay= [KNL]
4518 Set wakeup interval for idle CPUs that have
4519 RCU callbacks (RCU_FAST_NO_HZ=y).
4521 rcutree.rcu_kick_kthreads= [KNL]
4522 Cause the grace-period kthread to get an extra
4523 wake_up() if it sleeps three times longer than
4524 it should at force-quiescent-state time.
4525 This wake_up() will be accompanied by a
4526 WARN_ONCE() splat and an ftrace_dump().
4528 rcutree.rcu_unlock_delay= [KNL]
4529 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4530 this specifies an rcu_read_unlock()-time delay
4531 in microseconds. This defaults to zero.
4532 Larger delays increase the probability of
4533 catching RCU pointer leaks, that is, buggy use
4534 of RCU-protected pointers after the relevant
4535 rcu_read_unlock() has completed.
4537 rcutree.sysrq_rcu= [KNL]
4538 Commandeer a sysrq key to dump out Tree RCU's
4539 rcu_node tree with an eye towards determining
4540 why a new grace period has not yet started.
4542 rcuscale.gp_async= [KNL]
4543 Measure performance of asynchronous
4544 grace-period primitives such as call_rcu().
4546 rcuscale.gp_async_max= [KNL]
4547 Specify the maximum number of outstanding
4548 callbacks per writer thread. When a writer
4549 thread exceeds this limit, it invokes the
4550 corresponding flavor of rcu_barrier() to allow
4551 previously posted callbacks to drain.
4553 rcuscale.gp_exp= [KNL]
4554 Measure performance of expedited synchronous
4555 grace-period primitives.
4557 rcuscale.holdoff= [KNL]
4558 Set test-start holdoff period. The purpose of
4559 this parameter is to delay the start of the
4560 test until boot completes in order to avoid
4563 rcuscale.kfree_rcu_test= [KNL]
4564 Set to measure performance of kfree_rcu() flooding.
4566 rcuscale.kfree_rcu_test_double= [KNL]
4567 Test the double-argument variant of kfree_rcu().
4568 If this parameter has the same value as
4569 rcuscale.kfree_rcu_test_single, both the single-
4570 and double-argument variants are tested.
4572 rcuscale.kfree_rcu_test_single= [KNL]
4573 Test the single-argument variant of kfree_rcu().
4574 If this parameter has the same value as
4575 rcuscale.kfree_rcu_test_double, both the single-
4576 and double-argument variants are tested.
4578 rcuscale.kfree_nthreads= [KNL]
4579 The number of threads running loops of kfree_rcu().
4581 rcuscale.kfree_alloc_num= [KNL]
4582 Number of allocations and frees done in an iteration.
4584 rcuscale.kfree_loops= [KNL]
4585 Number of loops doing rcuscale.kfree_alloc_num number
4586 of allocations and frees.
4588 rcuscale.nreaders= [KNL]
4589 Set number of RCU readers. The value -1 selects
4590 N, where N is the number of CPUs. A value
4591 "n" less than -1 selects N-n+1, where N is again
4592 the number of CPUs. For example, -2 selects N
4593 (the number of CPUs), -3 selects N+1, and so on.
4594 A value of "n" less than or equal to -N selects
4597 rcuscale.nwriters= [KNL]
4598 Set number of RCU writers. The values operate
4599 the same as for rcuscale.nreaders.
4600 N, where N is the number of CPUs
4602 rcuscale.perf_type= [KNL]
4603 Specify the RCU implementation to test.
4605 rcuscale.shutdown= [KNL]
4606 Shut the system down after performance tests
4607 complete. This is useful for hands-off automated
4610 rcuscale.verbose= [KNL]
4611 Enable additional printk() statements.
4613 rcuscale.writer_holdoff= [KNL]
4614 Write-side holdoff between grace periods,
4615 in microseconds. The default of zero says
4618 rcutorture.fqs_duration= [KNL]
4619 Set duration of force_quiescent_state bursts
4622 rcutorture.fqs_holdoff= [KNL]
4623 Set holdoff time within force_quiescent_state bursts
4626 rcutorture.fqs_stutter= [KNL]
4627 Set wait time between force_quiescent_state bursts
4630 rcutorture.fwd_progress= [KNL]
4631 Enable RCU grace-period forward-progress testing
4632 for the types of RCU supporting this notion.
4634 rcutorture.fwd_progress_div= [KNL]
4635 Specify the fraction of a CPU-stall-warning
4636 period to do tight-loop forward-progress testing.
4638 rcutorture.fwd_progress_holdoff= [KNL]
4639 Number of seconds to wait between successive
4640 forward-progress tests.
4642 rcutorture.fwd_progress_need_resched= [KNL]
4643 Enclose cond_resched() calls within checks for
4644 need_resched() during tight-loop forward-progress
4647 rcutorture.gp_cond= [KNL]
4648 Use conditional/asynchronous update-side
4649 primitives, if available.
4651 rcutorture.gp_exp= [KNL]
4652 Use expedited update-side primitives, if available.
4654 rcutorture.gp_normal= [KNL]
4655 Use normal (non-expedited) asynchronous
4656 update-side primitives, if available.
4658 rcutorture.gp_sync= [KNL]
4659 Use normal (non-expedited) synchronous
4660 update-side primitives, if available. If all
4661 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4662 rcutorture.gp_normal=, and rcutorture.gp_sync=
4663 are zero, rcutorture acts as if is interpreted
4664 they are all non-zero.
4666 rcutorture.irqreader= [KNL]
4667 Run RCU readers from irq handlers, or, more
4668 accurately, from a timer handler. Not all RCU
4669 flavors take kindly to this sort of thing.
4671 rcutorture.leakpointer= [KNL]
4672 Leak an RCU-protected pointer out of the reader.
4673 This can of course result in splats, and is
4674 intended to test the ability of things like
4675 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4678 rcutorture.n_barrier_cbs= [KNL]
4679 Set callbacks/threads for rcu_barrier() testing.
4681 rcutorture.nfakewriters= [KNL]
4682 Set number of concurrent RCU writers. These just
4683 stress RCU, they don't participate in the actual
4684 test, hence the "fake".
4686 rcutorture.nocbs_nthreads= [KNL]
4687 Set number of RCU callback-offload togglers.
4688 Zero (the default) disables toggling.
4690 rcutorture.nocbs_toggle= [KNL]
4691 Set the delay in milliseconds between successive
4692 callback-offload toggling attempts.
4694 rcutorture.nreaders= [KNL]
4695 Set number of RCU readers. The value -1 selects
4696 N-1, where N is the number of CPUs. A value
4697 "n" less than -1 selects N-n-2, where N is again
4698 the number of CPUs. For example, -2 selects N
4699 (the number of CPUs), -3 selects N+1, and so on.
4701 rcutorture.object_debug= [KNL]
4702 Enable debug-object double-call_rcu() testing.
4704 rcutorture.onoff_holdoff= [KNL]
4705 Set time (s) after boot for CPU-hotplug testing.
4707 rcutorture.onoff_interval= [KNL]
4708 Set time (jiffies) between CPU-hotplug operations,
4709 or zero to disable CPU-hotplug testing.
4711 rcutorture.read_exit= [KNL]
4712 Set the number of read-then-exit kthreads used
4713 to test the interaction of RCU updaters and
4714 task-exit processing.
4716 rcutorture.read_exit_burst= [KNL]
4717 The number of times in a given read-then-exit
4718 episode that a set of read-then-exit kthreads
4721 rcutorture.read_exit_delay= [KNL]
4722 The delay, in seconds, between successive
4723 read-then-exit testing episodes.
4725 rcutorture.shuffle_interval= [KNL]
4726 Set task-shuffle interval (s). Shuffling tasks
4727 allows some CPUs to go into dyntick-idle mode
4728 during the rcutorture test.
4730 rcutorture.shutdown_secs= [KNL]
4731 Set time (s) after boot system shutdown. This
4732 is useful for hands-off automated testing.
4734 rcutorture.stall_cpu= [KNL]
4735 Duration of CPU stall (s) to test RCU CPU stall
4736 warnings, zero to disable.
4738 rcutorture.stall_cpu_block= [KNL]
4739 Sleep while stalling if set. This will result
4740 in warnings from preemptible RCU in addition
4741 to any other stall-related activity.
4743 rcutorture.stall_cpu_holdoff= [KNL]
4744 Time to wait (s) after boot before inducing stall.
4746 rcutorture.stall_cpu_irqsoff= [KNL]
4747 Disable interrupts while stalling if set.
4749 rcutorture.stall_gp_kthread= [KNL]
4750 Duration (s) of forced sleep within RCU
4751 grace-period kthread to test RCU CPU stall
4752 warnings, zero to disable. If both stall_cpu
4753 and stall_gp_kthread are specified, the
4754 kthread is starved first, then the CPU.
4756 rcutorture.stat_interval= [KNL]
4757 Time (s) between statistics printk()s.
4759 rcutorture.stutter= [KNL]
4760 Time (s) to stutter testing, for example, specifying
4761 five seconds causes the test to run for five seconds,
4762 wait for five seconds, and so on. This tests RCU's
4763 ability to transition abruptly to and from idle.
4765 rcutorture.test_boost= [KNL]
4766 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4767 "Maybe" means test if the RCU implementation
4768 under test support RCU priority boosting.
4770 rcutorture.test_boost_duration= [KNL]
4771 Duration (s) of each individual boost test.
4773 rcutorture.test_boost_interval= [KNL]
4774 Interval (s) between each boost test.
4776 rcutorture.test_no_idle_hz= [KNL]
4777 Test RCU's dyntick-idle handling. See also the
4778 rcutorture.shuffle_interval parameter.
4780 rcutorture.torture_type= [KNL]
4781 Specify the RCU implementation to test.
4783 rcutorture.verbose= [KNL]
4784 Enable additional printk() statements.
4786 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4787 Dump ftrace buffer after reporting RCU CPU
4790 rcupdate.rcu_cpu_stall_suppress= [KNL]
4791 Suppress RCU CPU stall warning messages.
4793 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4794 Suppress RCU CPU stall warning messages and
4795 rcutorture writer stall warnings that occur
4796 during early boot, that is, during the time
4797 before the init task is spawned.
4799 rcupdate.rcu_cpu_stall_timeout= [KNL]
4800 Set timeout for RCU CPU stall warning messages.
4802 rcupdate.rcu_expedited= [KNL]
4803 Use expedited grace-period primitives, for
4804 example, synchronize_rcu_expedited() instead
4805 of synchronize_rcu(). This reduces latency,
4806 but can increase CPU utilization, degrade
4807 real-time latency, and degrade energy efficiency.
4808 No effect on CONFIG_TINY_RCU kernels.
4810 rcupdate.rcu_normal= [KNL]
4811 Use only normal grace-period primitives,
4812 for example, synchronize_rcu() instead of
4813 synchronize_rcu_expedited(). This improves
4814 real-time latency, CPU utilization, and
4815 energy efficiency, but can expose users to
4816 increased grace-period latency. This parameter
4817 overrides rcupdate.rcu_expedited. No effect on
4818 CONFIG_TINY_RCU kernels.
4820 rcupdate.rcu_normal_after_boot= [KNL]
4821 Once boot has completed (that is, after
4822 rcu_end_inkernel_boot() has been invoked), use
4823 only normal grace-period primitives. No effect
4824 on CONFIG_TINY_RCU kernels.
4826 But note that CONFIG_PREEMPT_RT=y kernels enables
4827 this kernel boot parameter, forcibly setting
4828 it to the value one, that is, converting any
4829 post-boot attempt at an expedited RCU grace
4830 period to instead use normal non-expedited
4831 grace-period processing.
4833 rcupdate.rcu_task_ipi_delay= [KNL]
4834 Set time in jiffies during which RCU tasks will
4835 avoid sending IPIs, starting with the beginning
4836 of a given grace period. Setting a large
4837 number avoids disturbing real-time workloads,
4838 but lengthens grace periods.
4840 rcupdate.rcu_task_stall_timeout= [KNL]
4841 Set timeout in jiffies for RCU task stall warning
4842 messages. Disable with a value less than or equal
4845 rcupdate.rcu_self_test= [KNL]
4846 Run the RCU early boot self tests
4850 Run specified binary instead of /init from the ramdisk,
4851 used for early userspace startup. See initrd.
4854 force - Override the decision by the kernel to hide the
4855 advertisement of RDRAND support (this affects
4856 certain AMD processors because of buggy BIOS
4857 support, specifically around the suspend/resume
4861 Turn on/off individual RDT features. List is:
4862 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4864 E.g. to turn on cmt and turn off mba use:
4868 Format (x86 or x86_64):
4869 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
4871 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4873 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4874 (prefix with 'panic_' to set mode for panic
4876 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4877 reboot_force is either force or not specified,
4878 reboot_cpu is s[mp]#### with #### being the processor
4879 to be used for rebooting.
4881 refscale.holdoff= [KNL]
4882 Set test-start holdoff period. The purpose of
4883 this parameter is to delay the start of the
4884 test until boot completes in order to avoid
4887 refscale.loops= [KNL]
4888 Set the number of loops over the synchronization
4889 primitive under test. Increasing this number
4890 reduces noise due to loop start/end overhead,
4891 but the default has already reduced the per-pass
4892 noise to a handful of picoseconds on ca. 2020
4895 refscale.nreaders= [KNL]
4896 Set number of readers. The default value of -1
4897 selects N, where N is roughly 75% of the number
4898 of CPUs. A value of zero is an interesting choice.
4900 refscale.nruns= [KNL]
4901 Set number of runs, each of which is dumped onto
4904 refscale.readdelay= [KNL]
4905 Set the read-side critical-section duration,
4906 measured in microseconds.
4908 refscale.scale_type= [KNL]
4909 Specify the read-protection implementation to test.
4911 refscale.shutdown= [KNL]
4912 Shut down the system at the end of the performance
4913 test. This defaults to 1 (shut it down) when
4914 refscale is built into the kernel and to 0 (leave
4915 it running) when refscale is built as a module.
4917 refscale.verbose= [KNL]
4918 Enable additional printk() statements.
4920 refscale.verbose_batched= [KNL]
4921 Batch the additional printk() statements. If zero
4922 (the default) or negative, print everything. Otherwise,
4923 print every Nth verbose statement, where N is the value
4927 [KNL, SMP] Set scheduler's default relax_domain_level.
4928 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4930 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4931 Format: <base1>,<size1>[,<base2>,<size2>,...]
4932 Reserve I/O ports or memory so the kernel won't use
4933 them. If <base> is less than 0x10000, the region
4934 is assumed to be I/O ports; otherwise it is memory.
4936 reservetop= [X86-32]
4938 Reserves a hole at the top of the kernel virtual
4941 reset_devices [KNL] Force drivers to reset the underlying device
4942 during initialization.
4945 Specify the partition device for software suspend
4947 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4949 resume_offset= [SWSUSP]
4950 Specify the offset from the beginning of the partition
4951 given by "resume=" at which the swap header is located,
4952 in <PAGE_SIZE> units (needed only for swap files).
4953 See Documentation/power/swsusp-and-swap-files.rst
4955 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4956 read the resume files
4958 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4959 Useful for devices that are detected asynchronously
4960 (e.g. USB and MMC devices).
4962 hibernate= [HIBERNATION]
4963 noresume Don't check if there's a hibernation image
4964 present during boot.
4965 nocompress Don't compress/decompress hibernation images.
4966 no Disable hibernation and resume.
4967 protect_image Turn on image protection during restoration
4968 (that will set all pages holding image data
4969 during restoration read-only).
4971 retain_initrd [RAM] Keep initrd memory after extraction
4973 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
4974 Speculative Code Execution with Return Instructions)
4977 AMD-based UNRET and IBPB mitigations alone do not stop
4978 sibling threads from influencing the predictions of other
4979 sibling threads. For that reason, STIBP is used on pro-
4980 cessors that support it, and mitigate SMT on processors
4984 auto - automatically select a migitation
4985 auto,nosmt - automatically select a mitigation,
4986 disabling SMT if necessary for
4987 the full mitigation (only on Zen1
4988 and older without STIBP).
4989 ibpb - On AMD, mitigate short speculation
4990 windows on basic block boundaries too.
4991 Safe, highest perf impact. It also
4992 enables STIBP if present. Not suitable
4994 ibpb,nosmt - Like "ibpb" above but will disable SMT
4995 when STIBP is not available. This is
4996 the alternative for systems which do not
4998 unret - Force enable untrained return thunks,
4999 only effective on AMD f15h-f17h based
5001 unret,nosmt - Like unret, but will disable SMT when STIBP
5002 is not available. This is the alternative for
5003 systems which do not have STIBP.
5005 Selecting 'auto' will choose a mitigation method at run
5006 time according to the CPU.
5008 Not specifying this option is equivalent to retbleed=auto.
5010 rfkill.default_state=
5011 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5012 etc. communication is blocked by default.
5015 rfkill.master_switch_mode=
5016 0 The "airplane mode" button does nothing.
5017 1 The "airplane mode" button toggles between everything
5018 blocked and the previous configuration.
5019 2 The "airplane mode" button toggles between everything
5020 blocked and everything unblocked.
5022 rhash_entries= [KNL,NET]
5023 Set number of hash buckets for route cache
5026 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5029 ro [KNL] Mount root device read-only on boot
5032 on Mark read-only kernel memory as read-only (default).
5033 off Leave read-only kernel memory writable for debugging.
5036 Enable the uart passthrough on the designated usb port
5037 on Rockchip SoCs. When active, the signals of the
5038 debug-uart get routed to the D+ and D- pins of the usb
5039 port and the regular usb controller gets disabled.
5041 root= [KNL] Root filesystem
5042 See name_to_dev_t comment in init/do_mounts.c.
5044 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5045 mount the root filesystem
5047 rootflags= [KNL] Set root filesystem mount option string
5049 rootfstype= [KNL] Set root filesystem type
5051 rootwait [KNL] Wait (indefinitely) for root device to show up.
5052 Useful for devices that are detected asynchronously
5053 (e.g. USB and MMC devices).
5055 rproc_mem=nn[KMG][@address]
5056 [KNL,ARM,CMA] Remoteproc physical memory block.
5057 Memory area to be used by remote processor image,
5060 rw [KNL] Mount root device read-write on boot
5062 S [KNL] Run init in single mode
5064 s390_iommu= [HW,S390]
5065 Set s390 IOTLB flushing mode
5067 With strict flushing every unmap operation will result in
5068 an IOTLB flush. Default is lazy flushing before reuse,
5072 See drivers/net/irda/sa1100_ir.c.
5074 sched_verbose [KNL] Enables verbose scheduler debug messages.
5076 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5077 Allowed values are enable and disable. This feature
5078 incurs a small amount of overhead in the scheduler
5079 but is useful for debugging and performance tuning.
5081 sched_thermal_decay_shift=
5082 [KNL, SMP] Set a decay shift for scheduler thermal
5083 pressure signal. Thermal pressure signal follows the
5084 default decay period of other scheduler pelt
5085 signals(usually 32 ms but configurable). Setting
5086 sched_thermal_decay_shift will left shift the decay
5087 period for the thermal pressure signal by the shift
5089 i.e. with the default pelt decay period of 32 ms
5090 sched_thermal_decay_shift thermal pressure decay pr
5094 Format: integer between 0 and 10
5097 scftorture.holdoff= [KNL]
5098 Number of seconds to hold off before starting
5099 test. Defaults to zero for module insertion and
5100 to 10 seconds for built-in smp_call_function()
5103 scftorture.longwait= [KNL]
5104 Request ridiculously long waits randomly selected
5105 up to the chosen limit in seconds. Zero (the
5106 default) disables this feature. Please note
5107 that requesting even small non-zero numbers of
5108 seconds can result in RCU CPU stall warnings,
5109 softlockup complaints, and so on.
5111 scftorture.nthreads= [KNL]
5112 Number of kthreads to spawn to invoke the
5113 smp_call_function() family of functions.
5114 The default of -1 specifies a number of kthreads
5115 equal to the number of CPUs.
5117 scftorture.onoff_holdoff= [KNL]
5118 Number seconds to wait after the start of the
5119 test before initiating CPU-hotplug operations.
5121 scftorture.onoff_interval= [KNL]
5122 Number seconds to wait between successive
5123 CPU-hotplug operations. Specifying zero (which
5124 is the default) disables CPU-hotplug operations.
5126 scftorture.shutdown_secs= [KNL]
5127 The number of seconds following the start of the
5128 test after which to shut down the system. The
5129 default of zero avoids shutting down the system.
5130 Non-zero values are useful for automated tests.
5132 scftorture.stat_interval= [KNL]
5133 The number of seconds between outputting the
5134 current test statistics to the console. A value
5135 of zero disables statistics output.
5137 scftorture.stutter_cpus= [KNL]
5138 The number of jiffies to wait between each change
5139 to the set of CPUs under test.
5141 scftorture.use_cpus_read_lock= [KNL]
5142 Use use_cpus_read_lock() instead of the default
5143 preempt_disable() to disable CPU hotplug
5144 while invoking one of the smp_call_function*()
5147 scftorture.verbose= [KNL]
5148 Enable additional printk() statements.
5150 scftorture.weight_single= [KNL]
5151 The probability weighting to use for the
5152 smp_call_function_single() function with a zero
5153 "wait" parameter. A value of -1 selects the
5154 default if all other weights are -1. However,
5155 if at least one weight has some other value, a
5156 value of -1 will instead select a weight of zero.
5158 scftorture.weight_single_wait= [KNL]
5159 The probability weighting to use for the
5160 smp_call_function_single() function with a
5161 non-zero "wait" parameter. See weight_single.
5163 scftorture.weight_many= [KNL]
5164 The probability weighting to use for the
5165 smp_call_function_many() function with a zero
5166 "wait" parameter. See weight_single.
5167 Note well that setting a high probability for
5168 this weighting can place serious IPI load
5171 scftorture.weight_many_wait= [KNL]
5172 The probability weighting to use for the
5173 smp_call_function_many() function with a
5174 non-zero "wait" parameter. See weight_single
5177 scftorture.weight_all= [KNL]
5178 The probability weighting to use for the
5179 smp_call_function_all() function with a zero
5180 "wait" parameter. See weight_single and
5183 scftorture.weight_all_wait= [KNL]
5184 The probability weighting to use for the
5185 smp_call_function_all() function with a
5186 non-zero "wait" parameter. See weight_single
5189 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5190 xtime_lock contention on larger systems, and/or RCU lock
5191 contention on all systems with CONFIG_MAXSMP set.
5192 Format: { "0" | "1" }
5193 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5195 Note: increases power consumption, thus should only be
5196 enabled if running jitter sensitive (HPC/RT) workloads.
5198 security= [SECURITY] Choose a legacy "major" security module to
5199 enable at boot. This has been deprecated by the
5202 selinux= [SELINUX] Disable or enable SELinux at boot time.
5203 Format: { "0" | "1" }
5204 See security/selinux/Kconfig help text.
5209 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5210 Format: { "0" | "1" }
5211 See security/apparmor/Kconfig help text
5214 Default value is set via kernel config option.
5216 serialnumber [BUGS=X86-32]
5219 Maximal number of shapers.
5227 Enable merging of slabs with similar size when the
5228 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5231 Disable merging of slabs with similar size. May be
5232 necessary if there is some reason to distinguish
5233 allocs to different slabs, especially in hardened
5234 environments where the risk of heap overflows and
5235 layout control by attackers can usually be
5236 frustrated by disabling merging. This will reduce
5237 most of the exposure of a heap attack to a single
5238 cache (risks via metadata attacks are mostly
5239 unchanged). Debug options disable merging on their
5241 For more information see Documentation/vm/slub.rst.
5243 slab_max_order= [MM, SLAB]
5244 Determines the maximum allowed order for slabs.
5245 A high setting may cause OOMs due to memory
5246 fragmentation. Defaults to 1 for systems with
5247 more than 32MB of RAM, 0 otherwise.
5249 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5250 Enabling slub_debug allows one to determine the
5251 culprit if slab objects become corrupted. Enabling
5252 slub_debug can create guard zones around objects and
5253 may poison objects when not in use. Also tracks the
5254 last alloc / free. For more information see
5255 Documentation/vm/slub.rst.
5257 slub_max_order= [MM, SLUB]
5258 Determines the maximum allowed order for slabs.
5259 A high setting may cause OOMs due to memory
5260 fragmentation. For more information see
5261 Documentation/vm/slub.rst.
5263 slub_min_objects= [MM, SLUB]
5264 The minimum number of objects per slab. SLUB will
5265 increase the slab order up to slub_max_order to
5266 generate a sufficiently large slab able to contain
5267 the number of objects indicated. The higher the number
5268 of objects the smaller the overhead of tracking slabs
5269 and the less frequently locks need to be acquired.
5270 For more information see Documentation/vm/slub.rst.
5272 slub_min_order= [MM, SLUB]
5273 Determines the minimum page order for slabs. Must be
5274 lower than slub_max_order.
5275 For more information see Documentation/vm/slub.rst.
5277 slub_merge [MM, SLUB]
5278 Same with slab_merge.
5280 slub_nomerge [MM, SLUB]
5281 Same with slab_nomerge. This is supported for legacy.
5282 See slab_nomerge for more information.
5285 Format: <io1>[,<io2>[,...,<io8>]]
5287 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5288 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5289 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5290 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5291 smsc-ircc2.ircc_irq= [HW] IRQ line
5292 smsc-ircc2.ircc_dma= [HW] DMA channel
5293 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5294 0: Toshiba Satellite 1800 (GP data pin select)
5295 1: Fast pin select (default)
5298 smt [KNL,S390] Set the maximum number of threads (logical
5299 CPUs) to use per physical CPU on systems capable of
5300 symmetric multithreading (SMT). Will be capped to the
5301 actual hardware limit.
5303 Default: -1 (no limit)
5306 [KNL] Should the soft-lockup detector generate panics.
5309 A value of 1 instructs the soft-lockup detector
5310 to panic the machine when a soft-lockup occurs. It is
5311 also controlled by the kernel.softlockup_panic sysctl
5312 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5313 respective build-time switch to that functionality.
5315 softlockup_all_cpu_backtrace=
5316 [KNL] Should the soft-lockup detector generate
5317 backtraces on all cpus.
5320 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5321 See Documentation/admin-guide/laptops/sonypi.rst
5323 spectre_v2= [X86] Control mitigation of Spectre variant 2
5324 (indirect branch speculation) vulnerability.
5325 The default operation protects the kernel from
5328 on - unconditionally enable, implies
5330 off - unconditionally disable, implies
5332 auto - kernel detects whether your CPU model is
5335 Selecting 'on' will, and 'auto' may, choose a
5336 mitigation method at run time according to the
5337 CPU, the available microcode, the setting of the
5338 CONFIG_RETPOLINE configuration option, and the
5339 compiler with which the kernel was built.
5341 Selecting 'on' will also enable the mitigation
5342 against user space to user space task attacks.
5344 Selecting 'off' will disable both the kernel and
5345 the user space protections.
5347 Specific mitigations can also be selected manually:
5349 retpoline - replace indirect branches
5350 retpoline,generic - Retpolines
5351 retpoline,lfence - LFENCE; indirect branch
5352 retpoline,amd - alias for retpoline,lfence
5353 eibrs - enhanced IBRS
5354 eibrs,retpoline - enhanced IBRS + Retpolines
5355 eibrs,lfence - enhanced IBRS + LFENCE
5356 ibrs - use IBRS to protect kernel
5358 Not specifying this option is equivalent to
5362 [X86] Control mitigation of Spectre variant 2
5363 (indirect branch speculation) vulnerability between
5366 on - Unconditionally enable mitigations. Is
5367 enforced by spectre_v2=on
5369 off - Unconditionally disable mitigations. Is
5370 enforced by spectre_v2=off
5372 prctl - Indirect branch speculation is enabled,
5373 but mitigation can be enabled via prctl
5374 per thread. The mitigation control state
5375 is inherited on fork.
5378 - Like "prctl" above, but only STIBP is
5379 controlled per thread. IBPB is issued
5380 always when switching between different user
5384 - Same as "prctl" above, but all seccomp
5385 threads will enable the mitigation unless
5386 they explicitly opt out.
5389 - Like "seccomp" above, but only STIBP is
5390 controlled per thread. IBPB is issued
5391 always when switching between different
5392 user space processes.
5394 auto - Kernel selects the mitigation depending on
5395 the available CPU features and vulnerability.
5398 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
5400 Not specifying this option is equivalent to
5401 spectre_v2_user=auto.
5403 spec_store_bypass_disable=
5404 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5405 (Speculative Store Bypass vulnerability)
5407 Certain CPUs are vulnerable to an exploit against a
5408 a common industry wide performance optimization known
5409 as "Speculative Store Bypass" in which recent stores
5410 to the same memory location may not be observed by
5411 later loads during speculative execution. The idea
5412 is that such stores are unlikely and that they can
5413 be detected prior to instruction retirement at the
5414 end of a particular speculation execution window.
5416 In vulnerable processors, the speculatively forwarded
5417 store can be used in a cache side channel attack, for
5418 example to read memory to which the attacker does not
5419 directly have access (e.g. inside sandboxed code).
5421 This parameter controls whether the Speculative Store
5422 Bypass optimization is used.
5424 On x86 the options are:
5426 on - Unconditionally disable Speculative Store Bypass
5427 off - Unconditionally enable Speculative Store Bypass
5428 auto - Kernel detects whether the CPU model contains an
5429 implementation of Speculative Store Bypass and
5430 picks the most appropriate mitigation. If the
5431 CPU is not vulnerable, "off" is selected. If the
5432 CPU is vulnerable the default mitigation is
5433 architecture and Kconfig dependent. See below.
5434 prctl - Control Speculative Store Bypass per thread
5435 via prctl. Speculative Store Bypass is enabled
5436 for a process by default. The state of the control
5437 is inherited on fork.
5438 seccomp - Same as "prctl" above, but all seccomp threads
5439 will disable SSB unless they explicitly opt out.
5441 Default mitigations:
5442 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
5444 On powerpc the options are:
5446 on,auto - On Power8 and Power9 insert a store-forwarding
5447 barrier on kernel entry and exit. On Power7
5448 perform a software flush on kernel entry and
5452 Not specifying this option is equivalent to
5453 spec_store_bypass_disable=auto.
5455 spia_io_base= [HW,MTD]
5461 [X86] Enable split lock detection or bus lock detection
5463 When enabled (and if hardware support is present), atomic
5464 instructions that access data across cache line
5465 boundaries will result in an alignment check exception
5466 for split lock detection or a debug exception for
5471 warn - the kernel will emit rate-limited warnings
5472 about applications triggering the #AC
5473 exception or the #DB exception. This mode is
5474 the default on CPUs that support split lock
5475 detection or bus lock detection. Default
5476 behavior is by #AC if both features are
5477 enabled in hardware.
5479 fatal - the kernel will send SIGBUS to applications
5480 that trigger the #AC exception or the #DB
5481 exception. Default behavior is by #AC if
5482 both features are enabled in hardware.
5485 Set system wide rate limit to N bus locks
5486 per second for bus lock detection.
5489 N/A for split lock detection.
5492 If an #AC exception is hit in the kernel or in
5493 firmware (i.e. not while executing in user mode)
5494 the kernel will oops in either "warn" or "fatal"
5497 #DB exception for bus lock is triggered only when
5501 Control the Special Register Buffer Data Sampling
5504 Certain CPUs are vulnerable to an MDS-like
5505 exploit which can leak bits from the random
5508 By default, this issue is mitigated by
5509 microcode. However, the microcode fix can cause
5510 the RDRAND and RDSEED instructions to become
5511 much slower. Among other effects, this will
5512 result in reduced throughput from /dev/urandom.
5514 The microcode mitigation can be disabled with
5515 the following option:
5517 off: Disable mitigation and remove
5518 performance impact to RDRAND and RDSEED
5520 srcutree.counter_wrap_check [KNL]
5521 Specifies how frequently to check for
5522 grace-period sequence counter wrap for the
5523 srcu_data structure's ->srcu_gp_seq_needed field.
5524 The greater the number of bits set in this kernel
5525 parameter, the less frequently counter wrap will
5526 be checked for. Note that the bottom two bits
5529 srcutree.exp_holdoff [KNL]
5530 Specifies how many nanoseconds must elapse
5531 since the end of the last SRCU grace period for
5532 a given srcu_struct until the next normal SRCU
5533 grace period will be considered for automatic
5534 expediting. Set to zero to disable automatic
5538 Speculative Store Bypass Disable control
5540 On CPUs that are vulnerable to the Speculative
5541 Store Bypass vulnerability and offer a
5542 firmware based mitigation, this parameter
5543 indicates how the mitigation should be used:
5545 force-on: Unconditionally enable mitigation for
5546 for both kernel and userspace
5547 force-off: Unconditionally disable mitigation for
5548 for both kernel and userspace
5549 kernel: Always enable mitigation in the
5550 kernel, and offer a prctl interface
5551 to allow userspace to register its
5552 interest in being mitigated too.
5554 stack_guard_gap= [MM]
5555 override the default stack gap protection. The value
5556 is in page units and it defines how many pages prior
5557 to (for stacks growing down) resp. after (for stacks
5558 growing up) the main stack are reserved for no other
5559 mapping. Default value is 256 pages.
5561 stack_depot_disable= [KNL]
5562 Setting this to true through kernel command line will
5563 disable the stack depot thereby saving the static memory
5564 consumed by the stack hash table. By default this is set
5568 Enabled the stack tracer on boot up.
5570 stacktrace_filter=[function-list]
5571 [FTRACE] Limit the functions that the stack tracer
5572 will trace at boot up. function-list is a comma-separated
5573 list of functions. This list can be changed at run
5574 time by the stack_trace_filter file in the debugfs
5575 tracing directory. Note, this enables stack tracing
5576 and the stacktrace above is not needed.
5580 Set the STI (builtin display/keyboard on the HP-PARISC
5581 machines) console (graphic card) which should be used
5582 as the initial boot-console.
5583 See also comment in drivers/video/console/sticore.c.
5586 See comment in drivers/video/console/sticore.c.
5589 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5591 sunrpc.min_resvport=
5592 sunrpc.max_resvport=
5594 SunRPC servers often require that client requests
5595 originate from a privileged port (i.e. a port in the
5596 range 0 < portnr < 1024).
5597 An administrator who wishes to reserve some of these
5598 ports for other uses may adjust the range that the
5599 kernel's sunrpc client considers to be privileged
5600 using these two parameters to set the minimum and
5601 maximum port values.
5603 sunrpc.svc_rpc_per_connection_limit=
5605 Limit the number of requests that the server will
5606 process in parallel from a single connection.
5607 The default value is 0 (no limit).
5611 Control how the NFS server code allocates CPUs to
5612 service thread pools. Depending on how many NICs
5613 you have and where their interrupts are bound, this
5614 option will affect which CPUs will do NFS serving.
5615 Note: this parameter cannot be changed while the
5616 NFS server is running.
5618 auto the server chooses an appropriate mode
5619 automatically using heuristics
5620 global a single global pool contains all CPUs
5621 percpu one pool for each CPU
5622 pernode one pool for each NUMA node (equivalent
5623 to global on non-NUMA machines)
5625 sunrpc.tcp_slot_table_entries=
5626 sunrpc.udp_slot_table_entries=
5628 Sets the upper limit on the number of simultaneous
5629 RPC calls that can be sent from the client to a
5630 server. Increasing these values may allow you to
5631 improve throughput, but will also increase the
5632 amount of memory reserved for use by the client.
5634 suspend.pm_test_delay=
5636 Sets the number of seconds to remain in a suspend test
5637 mode before resuming the system (see
5638 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5639 is set. Default value is 5.
5642 Format: { on | off | y | n | 1 | 0 }
5643 This parameter controls use of the Protected
5644 Execution Facility on pSeries.
5647 [KNL] Enable accounting of swap in memory resource
5648 controller if no parameter or 1 is given or disable
5649 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5651 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5652 Format: { <int> | force | noforce }
5653 <int> -- Number of I/O TLB slabs
5654 force -- force using of bounce buffers even if they
5655 wouldn't be automatically used by the kernel
5656 noforce -- Never use bounce buffers (for debugging)
5661 Set a sysctl parameter, right before loading the init
5662 process, as if the value was written to the respective
5663 /proc/sys/... file. Both '.' and '/' are recognized as
5664 separators. Unrecognized parameters and invalid values
5665 are reported in the kernel log. Sysctls registered
5666 later by a loaded module cannot be set this way.
5667 Example: sysctl.vm.swappiness=40
5669 sysfs.deprecated=0|1 [KNL]
5670 Enable/disable old style sysfs layout for old udev
5671 on older distributions. When this option is enabled
5672 very new udev will not work anymore. When this option
5673 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5674 in older udev will not work anymore.
5675 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5676 the kernel configuration.
5678 sysrq_always_enabled
5680 Ignore sysrq setting - this boot parameter will
5681 neutralize any effect of /proc/sys/kernel/sysrq.
5682 Useful for debugging.
5684 tcpmhash_entries= [KNL,NET]
5685 Set the number of tcp_metrics_hash slots.
5686 Default value is 8192 or 16384 depending on total
5687 ram pages. This is used to specify the TCP metrics
5688 cache size. See Documentation/networking/ip-sysctl.rst
5689 "tcp_no_metrics_save" section for more details.
5693 test_suspend= [SUSPEND][,N]
5694 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5695 standby suspend) or "freeze" (for suspend type freeze)
5696 as the system sleep state during system startup with
5697 the optional capability to repeat N number of times.
5698 The system is woken from this state using a
5699 wakeup-capable RTC alarm.
5701 thash_entries= [KNL,NET]
5702 Set number of hash buckets for TCP connection
5704 thermal.act= [HW,ACPI]
5705 -1: disable all active trip points in all thermal zones
5706 <degrees C>: override all lowest active trip points
5708 thermal.crt= [HW,ACPI]
5709 -1: disable all critical trip points in all thermal zones
5710 <degrees C>: override all critical trip points
5712 thermal.nocrt= [HW,ACPI]
5713 Set to disable actions on ACPI thermal zone
5714 critical and hot trip points.
5716 thermal.off= [HW,ACPI]
5717 1: disable ACPI thermal control
5719 thermal.psv= [HW,ACPI]
5720 -1: disable all passive trip points
5721 <degrees C>: override all passive trip points to this
5724 thermal.tzp= [HW,ACPI]
5725 Specify global default ACPI thermal zone polling rate
5726 <deci-seconds>: poll all this frequency
5727 0: no polling (default)
5730 Force threading of all interrupt handlers except those
5731 marked explicitly IRQF_NO_THREAD.
5735 Specify if the kernel should make use of the cpu
5736 topology information if the hardware supports this.
5737 The scheduler will make use of this information and
5738 e.g. base its process migration decisions on it.
5741 topology_updates= [KNL, PPC, NUMA]
5743 Specify if the kernel should ignore (off)
5744 topology updates sent by the hypervisor to this
5747 torture.disable_onoff_at_boot= [KNL]
5748 Prevent the CPU-hotplug component of torturing
5749 until after init has spawned.
5751 torture.ftrace_dump_at_shutdown= [KNL]
5752 Dump the ftrace buffer at torture-test shutdown,
5753 even if there were no errors. This can be a
5754 very costly operation when many torture tests
5755 are running concurrently, especially on systems
5756 with rotating-rust storage.
5758 torture.verbose_sleep_frequency= [KNL]
5759 Specifies how many verbose printk()s should be
5760 emitted between each sleep. The default of zero
5761 disables verbose-printk() sleeping.
5763 torture.verbose_sleep_duration= [KNL]
5764 Duration of each verbose-printk() sleep in jiffies.
5768 tpm_suspend_pcr=[HW,TPM]
5769 Format: integer pcr id
5770 Specify that at suspend time, the tpm driver
5771 should extend the specified pcr with zeros,
5772 as a workaround for some chips which fail to
5773 flush the last written pcr on TPM_SaveState.
5774 This will guarantee that all the other pcrs
5777 trace_buf_size=nn[KMG]
5778 [FTRACE] will set tracing buffer size on each cpu.
5780 trace_event=[event-list]
5781 [FTRACE] Set and start specified trace events in order
5782 to facilitate early boot debugging. The event-list is a
5783 comma-separated list of trace events to enable. See
5784 also Documentation/trace/events.rst
5786 trace_options=[option-list]
5787 [FTRACE] Enable or disable tracer options at boot.
5788 The option-list is a comma delimited list of options
5789 that can be enabled or disabled just as if you were
5790 to echo the option name into
5792 /sys/kernel/debug/tracing/trace_options
5794 For example, to enable stacktrace option (to dump the
5795 stack trace of each event), add to the command line:
5797 trace_options=stacktrace
5799 See also Documentation/trace/ftrace.rst "trace options"
5803 Have the tracepoints sent to printk as well as the
5804 tracing ring buffer. This is useful for early boot up
5805 where the system hangs or reboots and does not give the
5806 option for reading the tracing buffer or performing a
5807 ftrace_dump_on_oops.
5809 To turn off having tracepoints sent to printk,
5810 echo 0 > /proc/sys/kernel/tracepoint_printk
5811 Note, echoing 1 into this file without the
5812 tracepoint_printk kernel cmdline option has no effect.
5814 The tp_printk_stop_on_boot (see below) can also be used
5815 to stop the printing of events to console at
5820 Having tracepoints sent to printk() and activating high
5821 frequency tracepoints such as irq or sched, can cause
5822 the system to live lock.
5824 tp_printk_stop_on_boot[FTRACE]
5825 When tp_printk (above) is set, it can cause a lot of noise
5826 on the console. It may be useful to only include the
5827 printing of events during boot up, as user space may
5828 make the system inoperable.
5830 This command line option will stop the printing of events
5831 to console at the late_initcall_sync() time frame.
5834 [FTRACE] enable this option to disable tracing when a
5835 warning is hit. This turns off "tracing_on". Tracing can
5836 be enabled again by echoing '1' into the "tracing_on"
5837 file located in /sys/kernel/debug/tracing/
5839 This option is useful, as it disables the trace before
5840 the WARNING dump is called, which prevents the trace to
5841 be filled with content caused by the warning output.
5843 This option can also be set at run time via the sysctl
5844 option: kernel/traceoff_on_warning
5846 transparent_hugepage=
5848 Format: [always|madvise|never]
5849 Can be used to control the default behavior of the system
5850 with respect to transparent hugepages.
5851 See Documentation/admin-guide/mm/transhuge.rst
5854 trusted.source= [KEYS]
5856 This parameter identifies the trust source as a backend
5857 for trusted keys implementation. Supported trust
5861 If not specified then it defaults to iterating through
5862 the trust source list starting with TPM and assigns the
5863 first trust source as a backend which is initialized
5864 successfully during iteration.
5866 tsc= Disable clocksource stability checks for TSC.
5868 [x86] reliable: mark tsc clocksource as reliable, this
5869 disables clocksource verification at runtime, as well
5870 as the stability checks done at bootup. Used to enable
5871 high-resolution timer mode on older hardware, and in
5872 virtualized environment.
5873 [x86] noirqtime: Do not use TSC to do irq accounting.
5874 Used to run time disable IRQ_TIME_ACCOUNTING on any
5875 platforms where RDTSC is slow and this accounting
5877 [x86] unstable: mark the TSC clocksource as unstable, this
5878 marks the TSC unconditionally unstable at bootup and
5879 avoids any further wobbles once the TSC watchdog notices.
5880 [x86] nowatchdog: disable clocksource watchdog. Used
5881 in situations with strict latency requirements (where
5882 interruptions from clocksource watchdog are not
5885 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5886 value instead. Useful when the early TSC frequency discovery
5887 procedure is not reliable, such as on overclocked systems
5888 with CPUID.16h support and partial CPUID.15h support.
5889 Format: <unsigned int>
5891 tsx= [X86] Control Transactional Synchronization
5892 Extensions (TSX) feature in Intel processors that
5893 support TSX control.
5895 This parameter controls the TSX feature. The options are:
5897 on - Enable TSX on the system. Although there are
5898 mitigations for all known security vulnerabilities,
5899 TSX has been known to be an accelerator for
5900 several previous speculation-related CVEs, and
5901 so there may be unknown security risks associated
5902 with leaving it enabled.
5904 off - Disable TSX on the system. (Note that this
5905 option takes effect only on newer CPUs which are
5906 not vulnerable to MDS, i.e., have
5907 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5908 the new IA32_TSX_CTRL MSR through a microcode
5909 update. This new MSR allows for the reliable
5910 deactivation of the TSX functionality.)
5912 auto - Disable TSX if X86_BUG_TAA is present,
5913 otherwise enable TSX on the system.
5915 Not specifying this option is equivalent to tsx=off.
5917 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5920 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5921 Abort (TAA) vulnerability.
5923 Similar to Micro-architectural Data Sampling (MDS)
5924 certain CPUs that support Transactional
5925 Synchronization Extensions (TSX) are vulnerable to an
5926 exploit against CPU internal buffers which can forward
5927 information to a disclosure gadget under certain
5930 In vulnerable processors, the speculatively forwarded
5931 data can be used in a cache side channel attack, to
5932 access data to which the attacker does not have direct
5935 This parameter controls the TAA mitigation. The
5938 full - Enable TAA mitigation on vulnerable CPUs
5941 full,nosmt - Enable TAA mitigation and disable SMT on
5942 vulnerable CPUs. If TSX is disabled, SMT
5943 is not disabled because CPU is not
5944 vulnerable to cross-thread TAA attacks.
5945 off - Unconditionally disable TAA mitigation
5947 On MDS-affected machines, tsx_async_abort=off can be
5948 prevented by an active MDS mitigation as both vulnerabilities
5949 are mitigated with the same mechanism so in order to disable
5950 this mitigation, you need to specify mds=off too.
5952 Not specifying this option is equivalent to
5953 tsx_async_abort=full. On CPUs which are MDS affected
5954 and deploy MDS mitigation, TAA mitigation is not
5955 required and doesn't provide any additional
5959 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5961 turbografx.map[2|3]= [HW,JOY]
5962 TurboGraFX parallel port interface
5964 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5965 See also Documentation/input/devices/joystick-parport.rst
5967 udbg-immortal [PPC] When debugging early kernel crashes that
5968 happen after console_init() and before a proper
5969 console driver takes over, this boot options might
5970 help "seeing" what's going on.
5972 uhash_entries= [KNL,NET]
5973 Set number of hash buckets for UDP/UDP-Lite connections
5976 [USB] Ignore overcurrent events (default N).
5977 Some badly-designed motherboards generate lots of
5978 bogus events, for ports that aren't wired to
5979 anything. Set this parameter to avoid log spamming.
5980 Note that genuine overcurrent events won't be
5984 [X86] Cause panic on unknown NMI.
5986 usbcore.authorized_default=
5987 [USB] Default USB device authorization:
5988 (default -1 = authorized except for wireless USB,
5989 0 = not authorized, 1 = authorized, 2 = authorized
5990 if device connected to internal port)
5992 usbcore.autosuspend=
5993 [USB] The autosuspend time delay (in seconds) used
5994 for newly-detected USB devices (default 2). This
5995 is the time required before an idle device will be
5996 autosuspended. Devices for which the delay is set
5997 to a negative value won't be autosuspended at all.
5999 usbcore.usbfs_snoop=
6000 [USB] Set to log all usbfs traffic (default 0 = off).
6002 usbcore.usbfs_snoop_max=
6003 [USB] Maximum number of bytes to snoop in each URB
6006 usbcore.blinkenlights=
6007 [USB] Set to cycle leds on hubs (default 0 = off).
6009 usbcore.old_scheme_first=
6010 [USB] Start with the old device initialization
6011 scheme (default 0 = off).
6013 usbcore.usbfs_memory_mb=
6014 [USB] Memory limit (in MB) for buffers allocated by
6015 usbfs (default = 16, 0 = max = 2047).
6017 usbcore.use_both_schemes=
6018 [USB] Try the other device initialization scheme
6019 if the first one fails (default 1 = enabled).
6021 usbcore.initial_descriptor_timeout=
6022 [USB] Specifies timeout for the initial 64-byte
6023 USB_REQ_GET_DESCRIPTOR request in milliseconds
6024 (default 5000 = 5.0 seconds).
6026 usbcore.nousb [USB] Disable the USB subsystem
6029 [USB] A list of quirk entries to augment the built-in
6030 usb core quirk list. List entries are separated by
6031 commas. Each entry has the form
6032 VendorID:ProductID:Flags. The IDs are 4-digit hex
6033 numbers and Flags is a set of letters. Each letter
6034 will change the built-in quirk; setting it if it is
6035 clear and clearing it if it is set. The letters have
6036 the following meanings:
6037 a = USB_QUIRK_STRING_FETCH_255 (string
6038 descriptors must not be fetched using
6040 b = USB_QUIRK_RESET_RESUME (device can't resume
6041 correctly so reset it instead);
6042 c = USB_QUIRK_NO_SET_INTF (device can't handle
6043 Set-Interface requests);
6044 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6045 handle its Configuration or Interface
6047 e = USB_QUIRK_RESET (device can't be reset
6048 (e.g morph devices), don't use reset);
6049 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6050 more interface descriptions than the
6051 bNumInterfaces count, and can't handle
6052 talking to these interfaces);
6053 g = USB_QUIRK_DELAY_INIT (device needs a pause
6054 during initialization, after we read
6055 the device descriptor);
6056 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6057 high speed and super speed interrupt
6058 endpoints, the USB 2.0 and USB 3.0 spec
6059 require the interval in microframes (1
6060 microframe = 125 microseconds) to be
6061 calculated as interval = 2 ^
6063 Devices with this quirk report their
6064 bInterval as the result of this
6065 calculation instead of the exponent
6066 variable used in the calculation);
6067 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6068 handle device_qualifier descriptor
6070 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6071 generates spurious wakeup, ignore
6072 remote wakeup capability);
6073 k = USB_QUIRK_NO_LPM (device can't handle Link
6075 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6076 (Device reports its bInterval as linear
6077 frames instead of the USB 2.0
6079 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6080 to be disconnected before suspend to
6081 prevent spurious wakeup);
6082 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6083 pause after every control message);
6084 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6085 delay after resetting its port);
6086 Example: quirks=0781:5580:bk,0a5c:5834:gij
6089 [USBHID] The interval which mice are to be polled at.
6092 [USBHID] The interval which joysticks are to be polled at.
6095 [USBHID] The interval which keyboards are to be polled at.
6097 usb-storage.delay_use=
6098 [UMS] The delay in seconds before a new device is
6099 scanned for Logical Units (default 1).
6102 [UMS] A list of quirks entries to supplement or
6103 override the built-in unusual_devs list. List
6104 entries are separated by commas. Each entry has
6105 the form VID:PID:Flags where VID and PID are Vendor
6106 and Product ID values (4-digit hex numbers) and
6107 Flags is a set of characters, each corresponding
6108 to a common usb-storage quirk flag as follows:
6109 a = SANE_SENSE (collect more than 18 bytes
6110 of sense data, not on uas);
6111 b = BAD_SENSE (don't collect more than 18
6112 bytes of sense data, not on uas);
6113 c = FIX_CAPACITY (decrease the reported
6114 device capacity by one sector);
6115 d = NO_READ_DISC_INFO (don't use
6116 READ_DISC_INFO command, not on uas);
6117 e = NO_READ_CAPACITY_16 (don't use
6118 READ_CAPACITY_16 command);
6119 f = NO_REPORT_OPCODES (don't use report opcodes
6121 g = MAX_SECTORS_240 (don't transfer more than
6122 240 sectors at a time, uas only);
6123 h = CAPACITY_HEURISTICS (decrease the
6124 reported device capacity by one
6125 sector if the number is odd);
6126 i = IGNORE_DEVICE (don't bind to this
6128 j = NO_REPORT_LUNS (don't use report luns
6130 k = NO_SAME (do not use WRITE_SAME, uas only)
6131 l = NOT_LOCKABLE (don't try to lock and
6132 unlock ejectable media, not on uas);
6133 m = MAX_SECTORS_64 (don't transfer more
6134 than 64 sectors = 32 KB at a time,
6136 n = INITIAL_READ10 (force a retry of the
6137 initial READ(10) command, not on uas);
6138 o = CAPACITY_OK (accept the capacity
6139 reported by the device, not on uas);
6140 p = WRITE_CACHE (the device cache is ON
6141 by default, not on uas);
6142 r = IGNORE_RESIDUE (the device reports
6143 bogus residue values, not on uas);
6144 s = SINGLE_LUN (the device has only one
6146 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6147 commands, uas only);
6148 u = IGNORE_UAS (don't bind to the uas driver);
6149 w = NO_WP_DETECT (don't test whether the
6150 medium is write-protected).
6151 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6152 even if the device claims no cache,
6154 Example: quirks=0419:aaf5:rl,0421:0433:rc
6156 user_debug= [KNL,ARM]
6158 See arch/arm/Kconfig.debug help text.
6159 1 - undefined instruction events
6161 4 - invalid data aborts
6164 Example: user_debug=31
6167 [X86] Flags controlling user PTE allocations.
6169 nohigh = do not allocate PTE pages in
6170 HIGHMEM regardless of setting
6174 On X86_32, this is an alias for vdso32=. Otherwise:
6176 vdso=1: enable VDSO (the default)
6177 vdso=0: disable VDSO mapping
6179 vdso32= [X86] Control the 32-bit vDSO
6180 vdso32=1: enable 32-bit VDSO
6181 vdso32=0 or vdso32=2: disable 32-bit VDSO
6183 See the help text for CONFIG_COMPAT_VDSO for more
6184 details. If CONFIG_COMPAT_VDSO is set, the default is
6185 vdso32=0; otherwise, the default is vdso32=1.
6187 For compatibility with older kernels, vdso32=2 is an
6190 Try vdso32=0 if you encounter an error that says:
6191 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6194 vector=percpu: enable percpu vector domain
6196 video= [FB] Frame buffer configuration
6197 See Documentation/fb/modedb.rst.
6199 video.brightness_switch_enabled= [0,1]
6200 If set to 1, on receiving an ACPI notify event
6201 generated by hotkey, video driver will adjust brightness
6202 level and then send out the event to user space through
6203 the allocated input device; If set to 0, video driver
6204 will only send out the event without touching backlight
6209 [VMMIO] Memory mapped virtio (platform) device.
6211 <size>@<baseaddr>:<irq>[:<id>]
6213 <size> := size (can use standard suffixes
6215 <baseaddr> := physical base address
6216 <irq> := interrupt number (as passed to
6218 <id> := (optional) platform device id
6220 virtio_mmio.device=1K@0x100b0000:48:7
6222 Can be used multiple times for multiple devices.
6224 vga= [BOOT,X86-32] Select a particular video mode
6225 See Documentation/x86/boot.rst and
6226 Documentation/admin-guide/svga.rst.
6227 Use vga=ask for menu.
6228 This is actually a boot loader parameter; the value is
6229 passed to the kernel using a special protocol.
6231 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6232 May slow down system boot speed, especially when
6233 enabled on systems with a large amount of memory.
6234 All options are enabled by default, and this
6235 interface is meant to allow for selectively
6236 enabling or disabling specific virtual memory
6239 Available options are:
6240 P Enable page structure init time poisoning
6241 - Disable all of the above options
6243 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6244 size of <nn>. This can be used to increase the
6245 minimum size (128MB on x86). It can also be used to
6246 decrease the size and leave more room for directly
6249 vmcp_cma=nn[MG] [KNL,S390]
6250 Sets the memory size reserved for contiguous memory
6251 allocations for the vmcp device driver.
6253 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6256 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6259 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6263 Controls the behavior of vsyscalls (i.e. calls to
6264 fixed addresses of 0xffffffffff600x00 from legacy
6265 code). Most statically-linked binaries and older
6266 versions of glibc use these calls. Because these
6267 functions are at fixed addresses, they make nice
6268 targets for exploits that can control RIP.
6270 emulate [default] Vsyscalls turn into traps and are
6271 emulated reasonably safely. The vsyscall
6274 xonly Vsyscalls turn into traps and are
6275 emulated reasonably safely. The vsyscall
6276 page is not readable.
6278 none Vsyscalls don't work at all. This makes
6279 them quite hard to use for exploits but
6280 might break your system.
6282 vt.color= [VT] Default text color.
6283 Format: 0xYX, X = foreground, Y = background.
6284 Default: 0x07 = light gray on black.
6286 vt.cur_default= [VT] Default cursor shape.
6287 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6288 the parameters of the <Esc>[?A;B;Cc escape sequence;
6289 see VGA-softcursor.txt. Default: 2 = underline.
6291 vt.default_blu= [VT]
6292 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6293 Change the default blue palette of the console.
6294 This is a 16-member array composed of values
6297 vt.default_grn= [VT]
6298 Format: <green0>,<green1>,<green2>,...,<green15>
6299 Change the default green palette of the console.
6300 This is a 16-member array composed of values
6303 vt.default_red= [VT]
6304 Format: <red0>,<red1>,<red2>,...,<red15>
6305 Change the default red palette of the console.
6306 This is a 16-member array composed of values
6312 Set system-wide default UTF-8 mode for all tty's.
6313 Default is 1, i.e. UTF-8 mode is enabled for all
6314 newly opened terminals.
6316 vt.global_cursor_default=
6319 Set system-wide default for whether a cursor
6320 is shown on new VTs. Default is -1,
6321 i.e. cursors will be created by default unless
6322 overridden by individual drivers. 0 will hide
6323 cursors, 1 will display them.
6325 vt.italic= [VT] Default color for italic text; 0-15.
6328 vt.underline= [VT] Default color for underlined text; 0-15.
6331 watchdog timers [HW,WDT] For information on watchdog timers,
6332 see Documentation/watchdog/watchdog-parameters.rst
6333 or other driver-specific files in the
6334 Documentation/watchdog/ directory.
6338 Set the hard lockup detector stall duration
6339 threshold in seconds. The soft lockup detector
6340 threshold is set to twice the value. A value of 0
6341 disables both lockup detectors. Default is 10
6344 workqueue.watchdog_thresh=
6345 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6346 warn stall conditions and dump internal state to
6347 help debugging. 0 disables workqueue stall
6348 detection; otherwise, it's the stall threshold
6349 duration in seconds. The default value is 30 and
6350 it can be updated at runtime by writing to the
6351 corresponding sysfs file.
6353 workqueue.disable_numa
6354 By default, all work items queued to unbound
6355 workqueues are affine to the NUMA nodes they're
6356 issued on, which results in better behavior in
6357 general. If NUMA affinity needs to be disabled for
6358 whatever reason, this option can be used. Note
6359 that this also can be controlled per-workqueue for
6360 workqueues visible under /sys/bus/workqueue/.
6362 workqueue.power_efficient
6363 Per-cpu workqueues are generally preferred because
6364 they show better performance thanks to cache
6365 locality; unfortunately, per-cpu workqueues tend to
6366 be more power hungry than unbound workqueues.
6368 Enabling this makes the per-cpu workqueues which
6369 were observed to contribute significantly to power
6370 consumption unbound, leading to measurably lower
6371 power usage at the cost of small performance
6374 The default value of this parameter is determined by
6375 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6377 workqueue.debug_force_rr_cpu
6378 Workqueue used to implicitly guarantee that work
6379 items queued without explicit CPU specified are put
6380 on the local CPU. This guarantee is no longer true
6381 and while local CPU is still preferred work items
6382 may be put on foreign CPUs. This debug option
6383 forces round-robin CPU selection to flush out
6384 usages which depend on the now broken guarantee.
6385 When enabled, memory and cache locality will be
6388 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6389 default x2apic cluster mode on platforms
6392 xen_512gb_limit [KNL,X86-64,XEN]
6393 Restricts the kernel running paravirtualized under Xen
6394 to use only up to 512 GB of RAM. The reason to do so is
6395 crash analysis tools and Xen tools for doing domain
6396 save/restore/migration must be enabled to handle larger
6399 xen_emul_unplug= [HW,X86,XEN]
6400 Unplug Xen emulated devices
6401 Format: [unplug0,][unplug1]
6402 ide-disks -- unplug primary master IDE devices
6403 aux-ide-disks -- unplug non-primary-master IDE devices
6404 nics -- unplug network devices
6405 all -- unplug all emulated devices (NICs and IDE disks)
6406 unnecessary -- unplugging emulated devices is
6407 unnecessary even if the host did not respond to
6409 never -- do not unplug even if version check succeeds
6411 xen_legacy_crash [X86,XEN]
6412 Crash from Xen panic notifier, without executing late
6413 panic() code such as dumping handler.
6415 xen_nopvspin [X86,XEN]
6416 Disables the qspinlock slowpath using Xen PV optimizations.
6417 This parameter is obsoleted by "nopvspin" parameter, which
6418 has equivalent effect for XEN platform.
6421 Disables the PV optimizations forcing the HVM guest to
6422 run as generic HVM guest with no PV drivers.
6423 This option is obsoleted by the "nopv" option, which
6424 has equivalent effect for XEN platform.
6426 xen_no_vector_callback
6427 [KNL,X86,XEN] Disable the vector callback for Xen
6428 event channel interrupts.
6430 xen_scrub_pages= [XEN]
6431 Boolean option to control scrubbing pages before giving them back
6432 to Xen, for use by other domains. Can be also changed at runtime
6433 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6434 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6436 xen_timer_slop= [X86-64,XEN]
6437 Set the timer slop (in nanoseconds) for the virtual Xen
6438 timers (default is 100000). This adjusts the minimum
6439 delta of virtualized Xen timers, where lower values
6440 improve timer resolution at the expense of processing
6441 more timer interrupts.
6443 xen.balloon_boot_timeout= [XEN]
6444 The time (in seconds) to wait before giving up to boot
6445 in case initial ballooning fails to free enough memory.
6446 Applies only when running as HVM or PVH guest and
6447 started with less memory configured than allowed at
6448 max. Default is 180.
6450 xen.event_eoi_delay= [XEN]
6451 How long to delay EOI handling in case of event
6452 storms (jiffies). Default is 10.
6454 xen.event_loop_timeout= [XEN]
6455 After which time (jiffies) the event handling loop
6456 should start to delay EOI handling. Default is 2.
6458 xen.fifo_events= [XEN]
6459 Boolean parameter to disable using fifo event handling
6460 even if available. Normally fifo event handling is
6461 preferred over the 2-level event handling, as it is
6462 fairer and the number of possible event channels is
6463 much higher. Default is on (use fifo events).
6465 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6466 Disables the PV optimizations forcing the guest to run
6467 as generic guest with no PV drivers. Currently support
6468 XEN HVM, KVM, HYPER_V and VMWARE guest.
6470 nopvspin [X86,XEN,KVM]
6471 Disables the qspinlock slow path using PV optimizations
6472 which allow the hypervisor to 'idle' the guest on lock
6475 xirc2ps_cs= [NET,PCMCIA]
6477 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6480 By default on POWER9 and above, the kernel will
6481 natively use the XIVE interrupt controller. This option
6482 allows the fallback firmware mode to be used:
6484 off Fallback to firmware control of XIVE interrupt
6485 controller on both pseries and powernv
6486 platforms. Only useful on POWER9 and above.
6488 xhci-hcd.quirks [USB,KNL]
6489 A hex value specifying bitmask with supplemental xhci
6490 host controller quirks. Meaning of each bit can be
6491 consulted in header drivers/usb/host/xhci.h.
6494 Format: { early | on | rw | ro | off }
6495 Controls if xmon debugger is enabled. Default is off.
6496 Passing only "xmon" is equivalent to "xmon=early".
6497 early Call xmon as early as possible on boot; xmon
6498 debugger is called from setup_arch().
6499 on xmon debugger hooks will be installed so xmon
6500 is only called on a kernel crash. Default mode,
6501 i.e. either "ro" or "rw" mode, is controlled
6502 with CONFIG_XMON_DEFAULT_RO_MODE.
6503 rw xmon debugger hooks will be installed so xmon
6504 is called only on a kernel crash, mode is write,
6505 meaning SPR registers, memory and, other data
6506 can be written using xmon commands.
6507 ro same as "rw" option above but SPR registers,
6508 memory, and other data can't be written using
6510 off xmon is disabled.