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_hwsig,
229 s4_nohwsig, old_ordering, nonvs,
230 sci_force_enable, nobl }
231 See Documentation/power/video.rst for information on
233 s3_beep is for debugging; it makes the PC's speaker beep
234 as soon as the kernel's real-mode entry point is called.
235 s4_hwsig causes the kernel to check the ACPI hardware
236 signature during resume from hibernation, and gracefully
237 refuse to resume if it has changed. This complies with
238 the ACPI specification but not with reality, since
239 Windows does not do this and many laptops do change it
240 on docking. So the default behaviour is to allow resume
241 and simply warn when the signature changes, unless the
242 s4_hwsig option is enabled.
243 s4_nohwsig prevents ACPI hardware signature from being
244 used (or even warned about) during resume.
245 old_ordering causes the ACPI 1.0 ordering of the _PTS
246 control method, with respect to putting devices into
247 low power states, to be enforced (the ACPI 2.0 ordering
248 of _PTS is used by default).
249 nonvs prevents the kernel from saving/restoring the
250 ACPI NVS memory during suspend/hibernation and resume.
251 sci_force_enable causes the kernel to set SCI_EN directly
252 on resume from S1/S3 (which is against the ACPI spec,
253 but some broken systems don't work without it).
254 nobl causes the internal blacklist of systems known to
255 behave incorrectly in some ways with respect to system
256 suspend and resume to be ignored (use wisely).
258 acpi_use_timer_override [HW,ACPI]
259 Use timer override. For some broken Nvidia NF5 boards
260 that require a timer override, but don't have HPET
262 add_efi_memmap [EFI; X86] Include EFI memory map in
263 kernel's map of available physical RAM.
266 { off | try_unsupported }
267 off: disable AGP support
268 try_unsupported: try to drive unsupported chipsets
269 (may crash computer or cause data corruption)
272 See Documentation/sound/alsa-configuration.rst
275 Allow the default userspace alignment fault handler
276 behaviour to be specified. Bit 0 enables warnings,
277 bit 1 enables fixups, and bit 2 sends a segfault.
279 align_va_addr= [X86-64]
280 Align virtual addresses by clearing slice [14:12] when
281 allocating a VMA at process creation time. This option
282 gives you up to 3% performance improvement on AMD F15h
283 machines (where it is enabled by default) for a
284 CPU-intensive style benchmark, and it can vary highly in
285 a microbenchmark depending on workload and compiler.
287 32: only for 32-bit processes
288 64: only for 64-bit processes
289 on: enable for both 32- and 64-bit processes
290 off: disable for both 32- and 64-bit processes
292 alloc_snapshot [FTRACE]
293 Allocate the ftrace snapshot buffer on boot up when the
294 main buffer is allocated. This is handy if debugging
295 and you need to use tracing_snapshot() on boot up, and
296 do not want to use tracing_snapshot_alloc() as it needs
297 to be done where GFP_KERNEL allocations are allowed.
299 allow_mismatched_32bit_el0 [ARM64]
300 Allow execve() of 32-bit applications and setting of the
301 PER_LINUX32 personality on systems where only a strict
302 subset of the CPUs support 32-bit EL0. When this
303 parameter is present, the set of CPUs supporting 32-bit
304 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
305 and hot-unplug operations may be restricted.
307 See Documentation/arm64/asymmetric-32bit.rst for more
310 amd_iommu= [HW,X86-64]
311 Pass parameters to the AMD IOMMU driver in the system.
313 fullflush - Deprecated, equivalent to iommu.strict=1
314 off - do not initialize any AMD IOMMU found in
316 force_isolation - Force device isolation for all
317 devices. The IOMMU driver is not
318 allowed anymore to lift isolation
319 requirements as needed. This option
320 does not override iommu=pt
321 force_enable - Force enable the IOMMU on platforms known
322 to be buggy with IOMMU enabled. Use this
325 amd_iommu_dump= [HW,X86-64]
326 Enable AMD IOMMU driver option to dump the ACPI table
327 for AMD IOMMU. With this option enabled, AMD IOMMU
328 driver will print ACPI tables for AMD IOMMU during
329 IOMMU initialization.
331 amd_iommu_intr= [HW,X86-64]
332 Specifies one of the following AMD IOMMU interrupt
334 legacy - Use legacy interrupt remapping mode.
335 vapic - Use virtual APIC mode, which allows IOMMU
336 to inject interrupts directly into guest.
337 This mode requires kvm-amd.avic=1.
338 (Default when IOMMU HW support is present.)
340 amijoy.map= [HW,JOY] Amiga joystick support
341 Map of devices attached to JOY0DAT and JOY1DAT
343 See also Documentation/input/joydev/joystick.rst
345 analog.map= [HW,JOY] Analog joystick and gamepad support
346 Specifies type or capabilities of an analog joystick
347 connected to one of 16 gameports
348 Format: <type1>,<type2>,..<type16>
351 Power management functions (SPARCstation-4/5 + deriv.)
353 Disable APC CPU standby support. SPARCstation-Fox does
354 not play well with APC CPU idle - disable it if you have
355 APC and your system crashes randomly.
357 apic= [APIC,X86] Advanced Programmable Interrupt Controller
358 Change the output verbosity while booting
359 Format: { quiet (default) | verbose | debug }
360 Change the amount of debugging information output
361 when initialising the APIC and IO-APIC components.
362 For X86-32, this can also be used to specify an APIC
364 Format: apic=driver_name
365 Examples: apic=bigsmp
367 apic_extnmi= [APIC,X86] External NMI delivery setting
368 Format: { bsp (default) | all | none }
369 bsp: External NMI is delivered only to CPU 0
370 all: External NMIs are broadcast to all CPUs as a
372 none: External NMI is masked for all CPUs. This is
373 useful so that a dump capture kernel won't be
377 See Documentation/networking/ipv6.rst.
379 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
380 Limit apic dumping. The parameter defines the maximal
381 number of local apics being dumped. Also it is possible
382 to set it to "all" by meaning -- no limit here.
383 Format: { 1 (default) | 2 | ... | all }.
384 The parameter valid if only apic=debug or
385 apic=verbose is specified.
386 Example: apic=debug show_lapic=all
388 apm= [APM] Advanced Power Management
389 See header of arch/x86/kernel/apm_32.c.
391 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
392 Format: <io>,<irq>,<nodeID>
394 arm64.nobti [ARM64] Unconditionally disable Branch Target
395 Identification support
397 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
400 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
405 atarimouse= [HW,MOUSE] Atari Mouse
407 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
408 EzKey and similar keyboards
410 atkbd.reset= [HW] Reset keyboard during initialization
412 atkbd.set= [HW] Select keyboard code set
413 Format: <int> (2 = AT (default), 3 = PS/2)
415 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
418 atkbd.softraw= [HW] Choose between synthetic and real raw mode
419 Format: <bool> (0 = real, 1 = synthetic (default))
421 atkbd.softrepeat= [HW]
422 Use software keyboard repeat
424 audit= [KNL] Enable the audit sub-system
425 Format: { "0" | "1" | "off" | "on" }
426 0 | off - kernel audit is disabled and can not be
427 enabled until the next reboot
428 unset - kernel audit is initialized but disabled and
429 will be fully enabled by the userspace auditd.
430 1 | on - kernel audit is initialized and partially
431 enabled, storing at most audit_backlog_limit
432 messages in RAM until it is fully enabled by the
436 audit_backlog_limit= [KNL] Set the audit queue size limit.
437 Format: <int> (must be >=0)
440 bau= [X86_UV] Enable the BAU on SGI UV. The default
441 behavior is to disable the BAU (i.e. bau=0).
442 Format: { "0" | "1" }
445 unset - Disable the BAU.
447 baycom_epp= [HW,AX25]
450 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
452 See header of drivers/net/hamradio/baycom_par.c.
454 baycom_ser_fdx= [HW,AX25]
455 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
456 Format: <io>,<irq>,<mode>[,<baud>]
457 See header of drivers/net/hamradio/baycom_ser_fdx.c.
459 baycom_ser_hdx= [HW,AX25]
460 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
461 Format: <io>,<irq>,<mode>
462 See header of drivers/net/hamradio/baycom_ser_hdx.c.
464 blkdevparts= Manual partition parsing of block device(s) for
465 embedded devices based on command line input.
466 See Documentation/block/cmdline-partition.rst
468 boot_delay= Milliseconds to delay each printk during boot.
469 Values larger than 10 seconds (10000) are changed to
474 Extended command line options can be added to an initrd
475 and this will cause the kernel to look for it.
477 See Documentation/admin-guide/bootconfig.rst
480 Disable BERT OS support on buggy BIOSes.
482 bgrt_disable [ACPI][X86]
483 Disable BGRT to avoid flickering OEM logo.
485 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
486 bttv.radio= Most important insmod options are available as
488 bttv.pll= See Documentation/admin-guide/media/bttv.rst
491 bulk_remove=off [PPC] This parameter disables the use of the pSeries
492 firmware feature for flushing multiple hpte entries
495 c101= [NET] Moxa C101 synchronous serial card
497 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
498 Sometimes CPU hardware bugs make them report the cache
499 size incorrectly. The kernel will attempt work arounds
500 to fix known problems, but for some CPUs it is not
501 possible to determine what the correct size should be.
502 This option provides an override for these situations.
505 [NET] Specifies amount of time (in seconds) that
506 the kernel should wait for a network carrier. By default
507 it waits 120 seconds.
509 ca_keys= [KEYS] This parameter identifies a specific key(s) on
510 the system trusted keyring to be used for certificate
512 format: { id:<keyid> | builtin }
514 cca= [MIPS] Override the kernel pages' cache coherency
515 algorithm. Accepted values range from 0 to 7
516 inclusive. See arch/mips/include/asm/pgtable-bits.h
517 for platform specific values (SB1, Loongson3 and
520 ccw_timeout_log [S390]
521 See Documentation/s390/common_io.rst for details.
523 cgroup_disable= [KNL] Disable a particular controller or optional feature
524 Format: {name of the controller(s) or feature(s) to disable}
525 The effects of cgroup_disable=foo are:
526 - foo isn't auto-mounted if you mount all cgroups in
528 - foo isn't visible as an individually mountable
530 - if foo is an optional feature then the feature is
531 disabled and corresponding cgroup files are not
533 {Currently only "memory" controller deal with this and
534 cut the overhead, others just disable the usage. So
535 only cgroup_disable=memory is actually worthy}
536 Specifying "pressure" disables per-cgroup pressure
537 stall information accounting feature
539 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
540 Format: { { controller | "all" | "named" }
541 [,{ controller | "all" | "named" }...] }
542 Like cgroup_disable, but only applies to cgroup v1;
543 the blacklisted controllers remain available in cgroup2.
544 "all" blacklists all controllers and "named" disables
545 named mounts. Specifying both "all" and "named" disables
548 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
550 nosocket -- Disable socket memory accounting.
551 nokmem -- Disable kernel memory accounting.
553 checkreqprot [SELINUX] Set initial checkreqprot flag value.
554 Format: { "0" | "1" }
555 See security/selinux/Kconfig help text.
556 0 -- check protection applied by kernel (includes
557 any implied execute protection).
558 1 -- check protection requested by application.
559 Default value is set via a kernel config option.
560 Value can be changed at runtime via
561 /sys/fs/selinux/checkreqprot.
562 Setting checkreqprot to 1 is deprecated.
565 See Documentation/s390/common_io.rst for details.
568 Prevents the clock framework from automatically gating
569 clocks that have not been explicitly enabled by a Linux
570 device driver but are enabled in hardware at reset or
571 by the bootloader/firmware. Note that this does not
572 force such clocks to be always-on nor does it reserve
573 those clocks in any way. This parameter is useful for
574 debug and development, but should not be needed on a
575 platform with proper driver support. For more
576 information, see Documentation/driver-api/clk.rst.
578 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
580 Forces specified clocksource (if available) to be used
581 when calculating gettimeofday(). If specified
582 clocksource is not available, it defaults to PIT.
583 Format: { pit | tsc | cyclone | pmtmr }
585 clocksource= Override the default clocksource
587 Override the default clocksource and use the clocksource
588 with the name specified.
589 Some clocksource names to choose from, depending on
591 [all] jiffies (this is the base, fallback clocksource)
593 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
594 pxa_timer,timer3,32k_counter,timer0_1
595 [X86-32] pit,hpet,tsc;
596 scx200_hrt on Geode; cyclone on IBM x440
604 clocksource.arm_arch_timer.evtstrm=
607 Enable/disable the eventstream feature of the ARM
608 architected timer so that code using WFE-based polling
609 loops can be debugged more effectively on production
612 clocksource.max_cswd_read_retries= [KNL]
613 Number of clocksource_watchdog() retries due to
614 external delays before the clock will be marked
615 unstable. Defaults to two retries, that is,
616 three attempts to read the clock under test.
618 clocksource.verify_n_cpus= [KNL]
619 Limit the number of CPUs checked for clocksources
620 marked with CLOCK_SOURCE_VERIFY_PERCPU that
621 are marked unstable due to excessive skew.
622 A negative value says to check all CPUs, while
623 zero says not to check any. Values larger than
624 nr_cpu_ids are silently truncated to nr_cpu_ids.
625 The actual CPUs are chosen randomly, with
626 no replacement if the same CPU is chosen twice.
628 clocksource-wdtest.holdoff= [KNL]
629 Set the time in seconds that the clocksource
630 watchdog test waits before commencing its tests.
631 Defaults to zero when built as a module and to
632 10 seconds when built into the kernel.
634 clearcpuid=BITNUM[,BITNUM...] [X86]
635 Disable CPUID feature X for the kernel. See
636 arch/x86/include/asm/cpufeatures.h for the valid bit
637 numbers. Note the Linux specific bits are not necessarily
638 stable over kernel options, but the vendor specific
640 Also note that user programs calling CPUID directly
641 or using the feature without checking anything
642 will still see it. This just prevents it from
643 being used by the kernel or shown in /proc/cpuinfo.
644 Also note the kernel might malfunction if you disable
647 cma=nn[MG]@[start[MG][-end[MG]]]
649 Sets the size of kernel global memory area for
650 contiguous memory allocations and optionally the
651 placement constraint by the physical address range of
652 memory allocations. A value of 0 disables CMA
653 altogether. For more information, see
654 kernel/dma/contiguous.c
658 Sets the size of kernel per-numa memory area for
659 contiguous memory allocations. A value of 0 disables
660 per-numa CMA altogether. And If this option is not
661 specificed, the default value is 0.
662 With per-numa CMA enabled, DMA users on node nid will
663 first try to allocate buffer from the pernuma area
664 which is located in node nid, if the allocation fails,
665 they will fallback to the global default memory area.
667 cmo_free_hint= [PPC] Format: { yes | no }
668 Specify whether pages are marked as being inactive
669 when they are freed. This is used in CMO environments
670 to determine OS memory pressure for page stealing by
674 coherent_pool=nn[KMG] [ARM,KNL]
675 Sets the size of memory pool for coherent, atomic dma
676 allocations, by default set to 256K.
678 com20020= [HW,NET] ARCnet - COM20020 chipset
680 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
682 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
686 ARCnet - COM90xx chipset (memory-mapped buffers)
687 Format: <io>[,<irq>[,<memstart>]]
689 condev= [HW,S390] console device
692 console= [KNL] Output console device and options.
694 tty<n> Use the virtual console device <n>.
698 Use the specified serial port. The options are of
699 the form "bbbbpnf", where "bbbb" is the baud rate,
700 "p" is parity ("n", "o", or "e"), "n" is number of
701 bits, and "f" is flow control ("r" for RTS or
702 omit it). Default is "9600n8".
704 See Documentation/admin-guide/serial-console.rst for more
706 Documentation/networking/netconsole.rst for an
709 uart[8250],io,<addr>[,options]
710 uart[8250],mmio,<addr>[,options]
711 uart[8250],mmio16,<addr>[,options]
712 uart[8250],mmio32,<addr>[,options]
713 uart[8250],0x<addr>[,options]
714 Start an early, polled-mode console on the 8250/16550
715 UART at the specified I/O port or MMIO address,
716 switching to the matching ttyS device later.
717 MMIO inter-register address stride is either 8-bit
718 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
719 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
720 to be equivalent to 'mmio'. 'options' are specified in
721 the same format described for ttyS above; if unspecified,
722 the h/w is not re-initialized.
724 hvc<n> Use the hypervisor console device <n>. This is for
725 both Xen and PowerPC hypervisors.
728 Use to disable console output, i.e., to have kernel
729 console messages discarded.
730 This must be the only console= parameter used on the
733 If the device connected to the port is not a TTY but a braille
734 device, prepend "brl," before the device type, for instance
736 For now, only VisioBraille is supported.
739 [KNL] Change console messages format
741 By default we print messages on consoles in
742 "[time stamp] text\n" format (time stamp may not be
743 printed, depending on CONFIG_PRINTK_TIME or
744 `printk_time' param).
746 Switch to syslog format: "<%u>[time stamp] text\n"
747 IOW, each message will have a facility and loglevel
748 prefix. The format is similar to one used by syslog()
749 syscall, or to executing "dmesg -S --raw" or to reading
752 consoleblank= [KNL] The console blank (screen saver) timeout in
753 seconds. A value of 0 disables the blank timer.
757 [KNL] Change the default value for
758 /proc/<pid>/coredump_filter.
759 See also Documentation/filesystems/proc.rst.
761 coresight_cpu_debug.enable
764 Enable/disable the CPU sampling based debugging.
765 0: default value, disable debugging
766 1: enable debugging at boot time
768 cpuidle.off=1 [CPU_IDLE]
769 disable the cpuidle sub-system
772 [CPU_IDLE] Name of the cpuidle governor to use.
774 cpufreq.off=1 [CPU_FREQ]
775 disable the cpufreq sub-system
777 cpufreq.default_governor=
778 [CPU_FREQ] Name of the default cpufreq governor or
779 policy to use. This governor must be registered in the
780 kernel before the cpufreq driver probes.
783 [X86] Delay for N microsec between assert and de-assert
784 of APIC INIT to start processors. This delay occurs
785 on every CPU online, such as boot, and resume from suspend.
788 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
790 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
792 crashkernel=size[KMG][@offset[KMG]]
793 [KNL] Using kexec, Linux can switch to a 'crash kernel'
794 upon panic. This parameter reserves the physical
795 memory region [offset, offset + size] for that kernel
796 image. If '@offset' is omitted, then a suitable offset
797 is selected automatically.
798 [KNL, X86-64] Select a region under 4G first, and
799 fall back to reserve region above 4G when '@offset'
800 hasn't been specified.
801 See Documentation/admin-guide/kdump/kdump.rst for further details.
803 crashkernel=range1:size1[,range2:size2,...][@offset]
804 [KNL] Same as above, but depends on the memory
805 in the running system. The syntax of range is
806 start-[end] where start and end are both
807 a memory unit (amount[KMG]). See also
808 Documentation/admin-guide/kdump/kdump.rst for an example.
810 crashkernel=size[KMG],high
811 [KNL, X86-64] range could be above 4G. Allow kernel
812 to allocate physical memory region from top, so could
813 be above 4G if system have more than 4G ram installed.
814 Otherwise memory region will be allocated below 4G, if
816 It will be ignored if crashkernel=X is specified.
817 crashkernel=size[KMG],low
818 [KNL, X86-64] range under 4G. When crashkernel=X,high
819 is passed, kernel could allocate physical memory region
820 above 4G, that cause second kernel crash on system
821 that require some amount of low memory, e.g. swiotlb
822 requires at least 64M+32K low memory, also enough extra
823 low memory is needed to make sure DMA buffers for 32-bit
824 devices won't run out. Kernel would try to allocate at
825 at least 256M below 4G automatically.
826 This one let user to specify own low range under 4G
827 for second kernel instead.
828 0: to disable low allocation.
829 It will be ignored when crashkernel=X,high is not used
830 or memory reserved is below 4G.
833 [KNL] Disable crypto self-tests
838 cs89x0_media= [HW,NET]
839 Format: { rj45 | aui | bnc }
841 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
842 handling. When switched on, additional debug data is
843 printed to the console in case a hanging CPU is
844 detected, and that CPU is pinged again in order to try
845 to resolve the hang situation.
846 0: disable csdlock debugging (default)
847 1: enable basic csdlock debugging (minor impact)
848 ext: enable extended csdlock debugging (more impact,
852 See header of drivers/s390/block/dasd_devmap.c.
854 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
855 (one device per port)
856 Format: <port#>,<type>
857 See also Documentation/input/devices/joystick-parport.rst
859 debug [KNL] Enable kernel debugging (events log level).
862 [KNL] Enable printing [hashed] pointers early in the
863 boot sequence. If enabled, we use a weak hash instead
864 of siphash to hash pointers. Use this option if you are
865 seeing instances of '(___ptrval___)') and need to see a
866 value (hashed pointer) instead. Cryptographically
867 insecure, please do not use on production kernels.
870 [KNL] verbose locking self-tests
872 Print debugging info while doing the locking API
874 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
875 (no extra messages), setting it to -1 (all bits set)
876 will print _a_lot_ more information - normally only
877 useful to lockdep developers.
879 debug_objects [KNL] Enable object debugging
882 [KNL] Disable object debugging
884 debug_guardpage_minorder=
885 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
886 parameter allows control of the order of pages that will
887 be intentionally kept free (and hence protected) by the
888 buddy allocator. Bigger value increase the probability
889 of catching random memory corruption, but reduce the
890 amount of memory for normal system use. The maximum
891 possible value is MAX_ORDER/2. Setting this parameter
892 to 1 or 2 should be enough to identify most random
893 memory corruption problems caused by bugs in kernel or
894 driver code when a CPU writes to (or reads from) a
895 random memory location. Note that there exists a class
896 of memory corruptions problems caused by buggy H/W or
897 F/W or by drivers badly programing DMA (basically when
898 memory is written at bus level and the CPU MMU is
899 bypassed) which are not detectable by
900 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
901 tracking down these problems.
904 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
905 enables the feature at boot time. By default, it is
906 disabled and the system will work mostly the same as a
907 kernel built without CONFIG_DEBUG_PAGEALLOC.
908 Note: to get most of debug_pagealloc error reports, it's
909 useful to also enable the page_owner functionality.
910 on: enable the feature
912 debugfs= [KNL] This parameter enables what is exposed to userspace
913 and debugfs internal clients.
914 Format: { on, no-mount, off }
915 on: All functions are enabled.
917 Filesystem is not registered but kernel clients can
918 access APIs and a crashkernel can be used to read
919 its content. There is nothing to mount.
920 off: Filesystem is not registered and clients
921 get a -EPERM as result when trying to register files
922 or directories within debugfs.
923 This is equivalent of the runtime functionality if
924 debugfs was not enabled in the kernel at all.
925 Default value is set in build-time with a kernel configuration.
927 debugpat [X86] Enable PAT debugging
929 decnet.addr= [HW,NET]
930 Format: <area>[,<node>]
931 See also Documentation/networking/decnet.rst.
934 [HW] The size of the default HugeTLB page. This is
935 the size represented by the legacy /proc/ hugepages
936 APIs. In addition, this is the default hugetlb size
937 used for shmget(), mmap() and mounting hugetlbfs
938 filesystems. If not specified, defaults to the
939 architecture's default huge page size. Huge page
940 sizes are architecture dependent. See also
941 Documentation/admin-guide/mm/hugetlbpage.rst.
944 deferred_probe_timeout=
945 [KNL] Debugging option to set a timeout in seconds for
946 deferred probe to give up waiting on dependencies to
947 probe. Only specific dependencies (subsystems or
948 drivers) that have opted in will be ignored. A timeout of 0
949 will timeout at the end of initcalls. This option will also
950 dump out devices still on the deferred probe list after
953 dell_smm_hwmon.ignore_dmi=
954 [HW] Continue probing hardware even if DMI data
955 indicates that the driver is running on unsupported
958 dell_smm_hwmon.force=
959 [HW] Activate driver even if SMM BIOS signature does
960 not match list of supported models and enable otherwise
961 blacklisted features.
963 dell_smm_hwmon.power_status=
964 [HW] Report power status in /proc/i8k
965 (disabled by default).
967 dell_smm_hwmon.restricted=
968 [HW] Allow controlling fans only if SYS_ADMIN
971 dell_smm_hwmon.fan_mult=
972 [HW] Factor to multiply fan speed with.
974 dell_smm_hwmon.fan_max=
975 [HW] Maximum configurable fan speed.
978 Format: { on | off | def_only | inf_only | always }
979 on: s390 zlib hardware support for compression on
980 level 1 and decompression (default)
981 off: No s390 zlib hardware support
982 def_only: s390 zlib hardware support for deflate
983 only (compression on level 1)
984 inf_only: s390 zlib hardware support for inflate
986 always: Same as 'on' but ignores the selected compression
987 level always using hardware support (used for debugging)
990 Set number of hash buckets for dentry cache.
992 disable_1tb_segments [PPC]
993 Disables the use of 1TB hash page table segments. This
994 causes the kernel to fall back to 256MB segments which
995 can be useful when debugging issues that require an SLB
999 Limits the number of kernel SLB entries, and flushes
1000 them frequently to increase the rate of SLB faults
1001 on kernel addresses.
1004 See Documentation/networking/ipv6.rst.
1007 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1008 hardening is enabled for this boot. Hardened
1009 usercopy checking is used to protect the kernel
1010 from reading or writing beyond known memory
1011 allocation boundaries as a proactive defense
1012 against bounds-checking flaws in the kernel's
1013 copy_to_user()/copy_from_user() interface.
1014 on Perform hardened usercopy checks (default).
1015 off Disable hardened usercopy checks.
1018 Disable RADIX MMU mode on POWER9
1020 radix_hcall_invalidate=on [PPC/PSERIES]
1021 Disable RADIX GTSE feature and use hcall for TLB
1025 Disable TLBIE instruction. Currently does not work
1026 with KVM, with HASH MMU, or with coherent accelerators.
1028 disable_cpu_apicid= [X86,APIC,SMP]
1030 The number of initial APIC ID for the
1031 corresponding CPU to be disabled at boot,
1032 mostly used for the kdump 2nd kernel to
1033 disable BSP to wake up multiple CPUs without
1034 causing system reset or hang due to sending
1035 INIT from AP to BSP.
1037 disable_ddw [PPC/PSERIES]
1038 Disable Dynamic DMA Window support. Use this
1039 to workaround buggy firmware.
1041 disable_ipv6= [IPV6]
1042 See Documentation/networking/ipv6.rst.
1044 disable_mtrr_cleanup [X86]
1045 The kernel tries to adjust MTRR layout from continuous
1046 to discrete, to make X server driver able to add WB
1047 entry later. This parameter disables that.
1049 disable_mtrr_trim [X86, Intel and AMD only]
1050 By default the kernel will trim any uncacheable
1051 memory out of your available memory pool based on
1052 MTRR settings. This parameter disables that behavior,
1053 possibly causing your machine to run very slowly.
1055 disable_timer_pin_1 [X86]
1056 Disable PIN 1 of APIC timer
1057 Can be useful to work around chipset bugs.
1059 dis_ucode_ldr [X86] Disable the microcode loader.
1061 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1062 this option disables the debugging code at boot.
1064 dma_debug_entries=<number>
1065 This option allows to tune the number of preallocated
1066 entries for DMA-API debugging code. One entry is
1067 required per DMA-API allocation. Use this if the
1068 DMA-API debugging code disables itself because the
1069 architectural default is too low.
1071 dma_debug_driver=<driver_name>
1072 With this option the DMA-API debugging driver
1073 filter feature can be enabled at boot time. Just
1074 pass the driver to filter for as the parameter.
1075 The filter can be disabled or changed to another
1076 driver later using sysfs.
1078 driver_async_probe= [KNL]
1079 List of driver names to be probed asynchronously.
1080 Format: <driver_name1>,<driver_name2>...
1082 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1083 Broken monitors, graphic adapters, KVMs and EDIDless
1084 panels may send no or incorrect EDID data sets.
1085 This parameter allows to specify an EDID data sets
1086 in the /lib/firmware directory that are used instead.
1087 Generic built-in EDID data sets are used, if one of
1088 edid/1024x768.bin, edid/1280x1024.bin,
1089 edid/1680x1050.bin, or edid/1920x1080.bin is given
1090 and no file with the same name exists. Details and
1091 instructions how to build your own EDID data are
1092 available in Documentation/admin-guide/edid.rst. An EDID
1093 data set will only be used for a particular connector,
1094 if its name and a colon are prepended to the EDID
1095 name. Each connector may use a unique EDID data
1096 set by separating the files with a comma. An EDID
1097 data set with no connector name will be used for
1098 any connectors not explicitly specified.
1103 Format: {"off" | "known"}
1104 Control how the dt_cpu_ftrs device-tree binding is
1105 used for CPU feature discovery and setup (if it
1107 off: Do not use it, fall back to legacy cpu table.
1108 known: Do not pass through unknown features to guests
1109 or userspace, only those that the kernel is aware of.
1111 dump_apple_properties [X86]
1112 Dump name and content of EFI device properties on
1113 x86 Macs. Useful for driver authors to determine
1114 what data is available or for reverse-engineering.
1116 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1117 <module>.dyndbg[="val"]
1118 Enable debug messages at boot time. See
1119 Documentation/admin-guide/dynamic-debug-howto.rst
1122 nopku [X86] Disable Memory Protection Keys CPU feature found
1125 <module>.async_probe [KNL]
1126 Enable asynchronous probe on this module.
1128 early_ioremap_debug [KNL]
1129 Enable debug messages in early_ioremap support. This
1130 is useful for tracking down temporary early mappings
1131 which are not unmapped.
1133 earlycon= [KNL] Output early console device and options.
1135 When used with no options, the early console is
1136 determined by stdout-path property in device tree's
1137 chosen node or the ACPI SPCR table if supported by
1140 cdns,<addr>[,options]
1141 Start an early, polled-mode console on a Cadence
1142 (xuartps) serial port at the specified address. Only
1143 supported option is baud rate. If baud rate is not
1144 specified, the serial port must already be setup and
1147 uart[8250],io,<addr>[,options]
1148 uart[8250],mmio,<addr>[,options]
1149 uart[8250],mmio32,<addr>[,options]
1150 uart[8250],mmio32be,<addr>[,options]
1151 uart[8250],0x<addr>[,options]
1152 Start an early, polled-mode console on the 8250/16550
1153 UART at the specified I/O port or MMIO address.
1154 MMIO inter-register address stride is either 8-bit
1155 (mmio) or 32-bit (mmio32 or mmio32be).
1156 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1157 to be equivalent to 'mmio'. 'options' are specified
1158 in the same format described for "console=ttyS<n>"; if
1159 unspecified, the h/w is not initialized.
1163 Start an early, polled-mode console on a pl011 serial
1164 port at the specified address. The pl011 serial port
1165 must already be setup and configured. Options are not
1166 yet supported. If 'mmio32' is specified, then only
1167 the driver will use only 32-bit accessors to read/write
1168 the device registers.
1171 Start an early console on a litex serial port at the
1172 specified address. The serial port must already be
1173 setup and configured. Options are not yet supported.
1176 Start an early, polled-mode console on a meson serial
1177 port at the specified address. The serial port must
1178 already be setup and configured. Options are not yet
1182 Start an early, polled-mode console on an msm serial
1183 port at the specified address. The serial port
1184 must already be setup and configured. Options are not
1187 msm_serial_dm,<addr>
1188 Start an early, polled-mode console on an msm serial
1189 dm port at the specified address. The serial port
1190 must already be setup and configured. Options are not
1194 Start an early, polled-mode console on a serial port
1195 of an Actions Semi SoC, such as S500 or S900, at the
1196 specified address. The serial port must already be
1197 setup and configured. Options are not yet supported.
1200 Start an early, polled-mode console on a serial port
1201 of an RDA Micro SoC, such as RDA8810PL, at the
1202 specified address. The serial port must already be
1203 setup and configured. Options are not yet supported.
1206 Use RISC-V SBI (Supervisor Binary Interface) for early
1209 smh Use ARM semihosting calls for early console.
1217 Use early console provided by serial driver available
1218 on Samsung SoCs, requires selecting proper type and
1219 a correct base address of the selected UART port. The
1220 serial port must already be setup and configured.
1221 Options are not yet supported.
1224 Start an early, polled-mode console on a lantiq serial
1225 (lqasc) port at the specified address. The serial port
1226 must already be setup and configured. Options are not
1231 Use early console provided by Freescale LP UART driver
1232 found on Freescale Vybrid and QorIQ LS1021A processors.
1233 A valid base address must be provided, and the serial
1234 port must already be setup and configured.
1238 Start an early, polled-mode, output-only console on the
1239 Freescale i.MX UART at the specified address. The UART
1240 must already be setup and configured.
1243 Start an early, polled-mode console on the
1244 Armada 3700 serial port at the specified
1245 address. The serial port must already be setup
1246 and configured. Options are not yet supported.
1249 Start an early, polled-mode console on a Qualcomm
1250 Generic Interface (GENI) based serial port at the
1251 specified address. The serial port must already be
1252 setup and configured. Options are not yet supported.
1255 Start an early, unaccelerated console on the EFI
1256 memory mapped framebuffer (if available). On cache
1257 coherent non-x86 systems that use system memory for
1258 the framebuffer, pass the 'ram' option so that it is
1259 mapped with the correct attributes.
1262 Use early console provided by Freescale LINFlexD UART
1263 serial driver for NXP S32V234 SoCs. A valid base
1264 address must be provided, and the serial port must
1265 already be setup and configured.
1267 earlyprintk= [X86,SH,ARM,M68k,S390]
1271 earlyprintk=serial[,ttySn[,baudrate]]
1272 earlyprintk=serial[,0x...[,baudrate]]
1273 earlyprintk=ttySn[,baudrate]
1274 earlyprintk=dbgp[debugController#]
1275 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1276 earlyprintk=xdbc[xhciController#]
1278 earlyprintk is useful when the kernel crashes before
1279 the normal console is initialized. It is not enabled by
1280 default because it has some cosmetic problems.
1282 Append ",keep" to not disable it when the real console
1285 Only one of vga, efi, serial, or usb debug port can
1288 Currently only ttyS0 and ttyS1 may be specified by
1289 name. Other I/O ports may be explicitly specified
1290 on some architectures (x86 and arm at least) by
1291 replacing ttySn with an I/O port address, like this:
1292 earlyprintk=serial,0x1008,115200
1293 You can find the port for a given device in
1294 /proc/tty/driver/serial:
1295 2: uart:ST16650V2 port:00001008 irq:18 ...
1297 Interaction with the standard serial driver is not
1300 The VGA and EFI output is eventually overwritten by
1303 The xen option can only be used in Xen domains.
1305 The sclp output can only be used on s390.
1307 The optional "force" to "pciserial" enables use of a
1308 PCI device even when its classcode is not of the
1311 edac_report= [HW,EDAC] Control how to report EDAC event
1312 Format: {"on" | "off" | "force"}
1313 on: enable EDAC to report H/W event. May be overridden
1314 by other higher priority error reporting module.
1315 off: disable H/W event reporting through EDAC.
1316 force: enforce the use of EDAC to report H/W event.
1319 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1322 This is designed to be used in conjunction with
1323 the boot argument: earlyprintk=vga
1325 This parameter works in place of the kgdboc parameter
1326 but can only be used if the backing tty is available
1327 very early in the boot process. For early debugging
1328 via a serial port see kgdboc_earlycon instead.
1331 Format: {"off" | "on" | "skip[mbr]"}
1334 Format: { "debug", "disable_early_pci_dma",
1335 "nochunk", "noruntime", "nosoftreserve",
1336 "novamap", "no_disable_early_pci_dma" }
1337 debug: enable misc debug output.
1338 disable_early_pci_dma: disable the busmaster bit on all
1339 PCI bridges while in the EFI boot stub.
1340 nochunk: disable reading files in "chunks" in the EFI
1341 boot stub, as chunking can cause problems with some
1342 firmware implementations.
1343 noruntime : disable EFI runtime services support
1344 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1345 attribute may cause the kernel to reserve the
1346 memory range for a memory mapping driver to
1347 claim. Specify efi=nosoftreserve to disable this
1348 reservation and treat the memory by its base type
1349 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1350 novamap: do not call SetVirtualAddressMap().
1351 no_disable_early_pci_dma: Leave the busmaster bit set
1352 on all PCI bridges while in the EFI boot stub
1354 efi_no_storage_paranoia [EFI; X86]
1355 Using this parameter you can use more than 50% of
1356 your efi variable storage. Use this parameter only if
1357 you are really sure that your UEFI does sane gc and
1358 fulfills the spec otherwise your board may brick.
1360 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1361 Add arbitrary attribute to specific memory range by
1362 updating original EFI memory map.
1363 Region of memory which aa attribute is added to is
1366 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1367 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1368 attribute is added to range 0x100000000-0x180000000 and
1369 0x10a0000000-0x1120000000.
1371 If efi_fake_mem=8G@9G:0x40000 is specified, the
1372 EFI_MEMORY_SP(0x40000) attribute is added to
1373 range 0x240000000-0x43fffffff.
1375 Using this parameter you can do debugging of EFI memmap
1376 related features. For example, you can do debugging of
1377 Address Range Mirroring feature even if your box
1378 doesn't support it, or mark specific memory as
1381 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1382 that is to be dynamically loaded by Linux. If there are
1383 multiple variables with the same name but with different
1384 vendor GUIDs, all of them will be loaded. See
1385 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1388 eisa_irq_edge= [PARISC,HW]
1389 See header of drivers/parisc/eisa.c.
1392 See comment before function elanfreq_setup() in
1393 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1395 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1396 Specifies physical address of start of kernel core
1397 image elf header and optionally the size. Generally
1398 kexec loader will pass this option to capture kernel.
1399 See Documentation/admin-guide/kdump/kdump.rst for details.
1401 enable_mtrr_cleanup [X86]
1402 The kernel tries to adjust MTRR layout from continuous
1403 to discrete, to make X server driver able to add WB
1404 entry later. This parameter enables that.
1406 enable_timer_pin_1 [X86]
1407 Enable PIN 1 of APIC timer
1408 Can be useful to work around chipset bugs
1409 (in particular on some ATI chipsets).
1410 The kernel tries to set a reasonable default.
1412 enforcing [SELINUX] Set initial enforcing status.
1414 See security/selinux/Kconfig help text.
1415 0 -- permissive (log only, no denials).
1416 1 -- enforcing (deny and log).
1418 Value can be changed at runtime via
1419 /sys/fs/selinux/enforce.
1422 Disable Error Record Serialization Table (ERST)
1425 ether= [HW,NET] Ethernet cards parameters
1426 This option is obsoleted by the "netdev=" option, which
1427 has equivalent usage. See its documentation for details.
1431 Permit 'security.evm' to be updated regardless of
1432 current integrity status.
1437 fail_make_request=[KNL]
1438 General fault injection mechanism.
1439 Format: <interval>,<probability>,<space>,<times>
1440 See also Documentation/fault-injection/.
1443 Format: { initns | none }
1444 See Documentation/admin-guide/sysctl/net.rst for
1445 fb_tunnels_only_for_init_ns
1448 See Documentation/admin-guide/blockdev/floppy.rst.
1450 force_pal_cache_flush
1451 [IA-64] Avoid check_sal_cache_flush which may hang on
1452 buggy SAL_CACHE_FLUSH implementations. Using this
1453 parameter will force ia64_sal_cache_flush to call
1454 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1457 Forcefully enable Physical Address Extension (PAE).
1458 Many Pentium M systems disable PAE but may have a
1459 functionally usable PAE implementation.
1460 Warning: use of this parameter will taint the kernel
1461 and may cause unknown problems.
1464 [FTRACE] will set and start the specified tracer
1465 as early as possible in order to facilitate early
1468 ftrace_boot_snapshot
1469 [FTRACE] On boot up, a snapshot will be taken of the
1470 ftrace ring buffer that can be read at:
1471 /sys/kernel/tracing/snapshot.
1472 This is useful if you need tracing information from kernel
1473 boot up that is likely to be overridden by user space
1474 start up functionality.
1476 ftrace_dump_on_oops[=orig_cpu]
1477 [FTRACE] will dump the trace buffers on oops.
1478 If no parameter is passed, ftrace will dump
1479 buffers of all CPUs, but if you pass orig_cpu, it will
1480 dump only the buffer of the CPU that triggered the
1483 ftrace_filter=[function-list]
1484 [FTRACE] Limit the functions traced by the function
1485 tracer at boot up. function-list is a comma-separated
1486 list of functions. This list can be changed at run
1487 time by the set_ftrace_filter file in the debugfs
1490 ftrace_notrace=[function-list]
1491 [FTRACE] Do not trace the functions specified in
1492 function-list. This list can be changed at run time
1493 by the set_ftrace_notrace file in the debugfs
1496 ftrace_graph_filter=[function-list]
1497 [FTRACE] Limit the top level callers functions traced
1498 by the function graph tracer at boot up.
1499 function-list is a comma-separated list of functions
1500 that can be changed at run time by the
1501 set_graph_function file in the debugfs tracing directory.
1503 ftrace_graph_notrace=[function-list]
1504 [FTRACE] Do not trace from the functions specified in
1505 function-list. This list is a comma-separated list of
1506 functions that can be changed at run time by the
1507 set_graph_notrace file in the debugfs tracing directory.
1509 ftrace_graph_max_depth=<uint>
1510 [FTRACE] Used with the function graph tracer. This is
1511 the max depth it will trace into a function. This value
1512 can be changed at run time by the max_graph_depth file
1513 in the tracefs tracing directory. default: 0 (no limit)
1515 fw_devlink= [KNL] Create device links between consumer and supplier
1516 devices by scanning the firmware to infer the
1517 consumer/supplier relationships. This feature is
1518 especially useful when drivers are loaded as modules as
1519 it ensures proper ordering of tasks like device probing
1520 (suppliers first, then consumers), supplier boot state
1521 clean up (only after all consumers have probed),
1522 suspend/resume & runtime PM (consumers first, then
1524 Format: { off | permissive | on | rpm }
1525 off -- Don't create device links from firmware info.
1526 permissive -- Create device links from firmware info
1527 but use it only for ordering boot state clean
1528 up (sync_state() calls).
1529 on -- Create device links from firmware info and use it
1530 to enforce probe and suspend/resume ordering.
1531 rpm -- Like "on", but also use to order runtime PM.
1533 fw_devlink.strict=<bool>
1534 [KNL] Treat all inferred dependencies as mandatory
1535 dependencies. This only applies for fw_devlink=on|rpm.
1539 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1540 support via parallel port (up to 5 devices per port)
1541 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1542 See also Documentation/input/devices/joystick-parport.rst
1546 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1550 gcov_persist= [GCOV] When non-zero (default), profiling data for
1551 kernel modules is saved and remains accessible via
1552 debugfs, even when the module is unloaded/reloaded.
1553 When zero, profiling data is discarded and associated
1554 debugfs files are removed at module unload time.
1556 goldfish [X86] Enable the goldfish android emulator platform.
1557 Don't use this when you are not running on the
1560 gpio-mockup.gpio_mockup_ranges
1561 [HW] Sets the ranges of gpiochip of for this device.
1562 Format: <start1>,<end1>,<start2>,<end2>...
1563 gpio-mockup.gpio_mockup_named_lines
1564 [HW] Let the driver know GPIO lines should be named.
1566 gpt [EFI] Forces disk with valid GPT signature but
1567 invalid Protective MBR to be treated as GPT. If the
1568 primary GPT is corrupted, it enables the backup/alternate
1569 GPT to be used instead.
1571 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1572 the "Enable 0" bit of the configuration register.
1575 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1576 the "Enable 0" bit of the configuration register.
1579 grcan.select= [HW] Select which physical interface to use.
1582 grcan.txsize= [HW] Sets the size of the tx buffer.
1583 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1585 grcan.rxsize= [HW] Sets the size of the rx buffer.
1586 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1589 hardlockup_all_cpu_backtrace=
1590 [KNL] Should the hard-lockup detector generate
1591 backtraces on all cpus.
1594 hashdist= [KNL,NUMA] Large hashes allocated during boot
1595 are distributed across NUMA nodes. Defaults on
1596 for 64-bit NUMA, off otherwise.
1597 Format: 0 | 1 (for off | on)
1599 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1601 hd= [EIDE] (E)IDE hard drive subsystem geometry
1602 Format: <cyl>,<head>,<sect>
1605 Disable Hardware Error Source Table (HEST) support;
1606 corresponding firmware-first mode error processing
1607 logic will be disabled.
1609 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1610 size of <nn>. This works even on boxes that have no
1611 highmem otherwise. This also works to reduce highmem
1612 size on bigger boxes.
1614 highres= [KNL] Enable/disable high resolution timer mode.
1615 Valid parameters: "on", "off"
1620 hpet= [X86-32,HPET] option to control HPET usage
1621 Format: { enable (default) | disable | force |
1623 disable: disable HPET and use PIT instead
1624 force: allow force enabled of undocumented chips (ICH4,
1626 verbose: show contents of HPET registers during setup
1628 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1629 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1631 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1632 of gigantic hugepages. Or using node format, the size
1633 of a CMA area per node can be specified.
1634 Format: nn[KMGTPE] or (node format)
1635 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1637 Reserve a CMA area of given size and allocate gigantic
1638 hugepages using the CMA allocator. If enabled, the
1639 boot-time allocation of gigantic hugepages is skipped.
1641 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1642 If this follows hugepagesz (below), it specifies
1643 the number of pages of hugepagesz to be allocated.
1644 If this is the first HugeTLB parameter on the command
1645 line, it specifies the number of pages to allocate for
1646 the default huge page size. If using node format, the
1647 number of pages to allocate per-node can be specified.
1648 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1649 Format: <integer> or (node format)
1650 <node>:<integer>[,<node>:<integer>]
1653 [HW] The size of the HugeTLB pages. This is used in
1654 conjunction with hugepages (above) to allocate huge
1655 pages of a specific size at boot. The pair
1656 hugepagesz=X hugepages=Y can be specified once for
1657 each supported huge page size. Huge page sizes are
1658 architecture dependent. See also
1659 Documentation/admin-guide/mm/hugetlbpage.rst.
1662 hugetlb_free_vmemmap=
1663 [KNL] Reguires CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
1665 Allows heavy hugetlb users to free up some more
1666 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1667 Format: { on | off (default) }
1669 on: enable the feature
1670 off: disable the feature
1672 Built with CONFIG_HUGETLB_PAGE_FREE_VMEMMAP_DEFAULT_ON=y,
1675 This is not compatible with memory_hotplug.memmap_on_memory.
1676 If both parameters are enabled, hugetlb_free_vmemmap takes
1677 precedence over memory_hotplug.memmap_on_memory.
1680 [KNL] Should the hung task detector generate panics.
1683 A value of 1 instructs the kernel to panic when a
1684 hung task is detected. The default value is controlled
1685 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1686 option. The value selected by this boot parameter can
1687 be changed later by the kernel.hung_task_panic sysctl.
1689 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1690 terminal devices. Valid values: 0..8
1691 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1692 If specified, z/VM IUCV HVC accepts connections
1693 from listed z/VM user IDs only.
1695 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1696 which allow the hypervisor to 'idle' the
1697 guest on lock contention.
1700 Do not unregister boot console at start. This is only
1701 useful for debugging when something happens in the window
1702 between unregistering the boot console and initializing
1705 i2c_bus= [HW] Override the default board specific I2C bus speed
1706 or register an additional I2C bus that is not
1707 registered from board initialization code.
1711 i8042.debug [HW] Toggle i8042 debug mode
1712 i8042.unmask_kbd_data
1713 [HW] Enable printing of interrupt data from the KBD port
1714 (disabled by default, and as a pre-condition
1715 requires that i8042.debug=1 be enabled)
1716 i8042.direct [HW] Put keyboard port into non-translated mode
1717 i8042.dumbkbd [HW] Pretend that controller can only read data from
1718 keyboard and cannot control its state
1719 (Don't attempt to blink the leds)
1720 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1721 i8042.nokbd [HW] Don't check/create keyboard port
1722 i8042.noloop [HW] Disable the AUX Loopback command while probing
1724 i8042.nomux [HW] Don't check presence of an active multiplexing
1726 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1728 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1729 i8042.reset [HW] Reset the controller during init, cleanup and
1730 suspend-to-ram transitions, only during s2r
1731 transitions, or never reset
1732 Format: { 1 | Y | y | 0 | N | n }
1733 1, Y, y: always reset controller
1734 0, N, n: don't ever reset controller
1735 Default: only on s2r transitions on x86; most other
1736 architectures force reset to be always executed
1737 i8042.unlock [HW] Unlock (ignore) the keylock
1738 i8042.kbdreset [HW] Reset device connected to KBD port
1740 [HW] Allow deferred probing upon i8042 probe errors
1744 i915.invert_brightness=
1745 [DRM] Invert the sense of the variable that is used to
1746 set the brightness of the panel backlight. Normally a
1747 brightness value of 0 indicates backlight switched off,
1748 and the maximum of the brightness value sets the backlight
1749 to maximum brightness. If this parameter is set to 0
1750 (default) and the machine requires it, or this parameter
1751 is set to 1, a brightness value of 0 sets the backlight
1752 to maximum brightness, and the maximum of the brightness
1753 value switches the backlight off.
1754 -1 -- never invert brightness
1755 0 -- machine default
1756 1 -- force brightness inversion
1759 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1761 ide-core.nodma= [HW] (E)IDE subsystem
1762 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1763 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1764 .cdrom .chs .ignore_cable are additional options
1765 See Documentation/ide/ide.rst.
1767 ide-generic.probe-mask= [HW] (E)IDE subsystem
1769 Probe mask for legacy ISA IDE ports. Depending on
1770 platform up to 6 ports are supported, enabled by
1771 setting corresponding bits in the mask to 1. The
1772 default value is 0x0, which has a special meaning.
1773 On systems that have PCI, it triggers scanning the
1774 PCI bus for the first and the second port, which
1775 are then probed. On systems without PCI the value
1776 of 0x0 enables probing the two first ports as if it
1779 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1780 Claim all unknown PCI IDE storage controllers.
1783 Format: idle=poll, idle=halt, idle=nomwait
1784 Poll forces a polling idle loop that can slightly
1785 improve the performance of waking up a idle CPU, but
1786 will use a lot of power and make the system run hot.
1788 idle=halt: Halt is forced to be used for CPU idle.
1789 In such case C2/C3 won't be used again.
1790 idle=nomwait: Disable mwait for CPU C-states
1794 Allow force disabling of Shared Virtual Memory (SVA)
1795 support for the idxd driver. By default it is set to
1798 idxd.tc_override= [HW]
1800 Allow override of default traffic class configuration
1801 for the device. By default it is set to false (0).
1803 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1804 Format: { strict | legacy | 2008 | relaxed }
1807 Choose which programs will be accepted for execution
1808 based on the IEEE 754 NaN encoding(s) supported by
1809 the FPU and the NaN encoding requested with the value
1810 of an ELF file header flag individually set by each
1811 binary. Hardware implementations are permitted to
1812 support either or both of the legacy and the 2008 NaN
1815 Available settings are as follows:
1816 strict accept binaries that request a NaN encoding
1817 supported by the FPU
1818 legacy only accept legacy-NaN binaries, if supported
1820 2008 only accept 2008-NaN binaries, if supported
1822 relaxed accept any binaries regardless of whether
1823 supported by the FPU
1825 The FPU emulator is always able to support both NaN
1826 encodings, so if no FPU hardware is present or it has
1827 been disabled with 'nofpu', then the settings of
1828 'legacy' and '2008' strap the emulator accordingly,
1829 'relaxed' straps the emulator for both legacy-NaN and
1830 2008-NaN, whereas 'strict' enables legacy-NaN only on
1831 legacy processors and both NaN encodings on MIPS32 or
1834 The setting for ABS.fmt/NEG.fmt instruction execution
1835 mode generally follows that for the NaN encoding,
1836 except where unsupported by hardware.
1838 ignore_loglevel [KNL]
1839 Ignore loglevel setting - this will print /all/
1840 kernel messages to the console. Useful for debugging.
1841 We also add it as printk module parameter, so users
1842 could change it dynamically, usually by
1843 /sys/module/printk/parameters/ignore_loglevel.
1846 Ignore RLIMIT_DATA setting for data mappings,
1847 print warning at first misuse. Can be changed via
1848 /sys/module/kernel/parameters/ignore_rlimit_data.
1850 ihash_entries= [KNL]
1851 Set number of hash buckets for inode cache.
1853 ima_appraise= [IMA] appraise integrity measurements
1854 Format: { "off" | "enforce" | "fix" | "log" }
1857 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1858 The builtin appraise policy appraises all files
1861 ima_canonical_fmt [IMA]
1862 Use the canonical format for the binary runtime
1863 measurements, instead of host native format.
1866 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1870 The list of supported hash algorithms is defined
1871 in crypto/hash_info.h.
1874 The builtin policies to load during IMA setup.
1875 Format: "tcb | appraise_tcb | secure_boot |
1876 fail_securely | critical_data"
1878 The "tcb" policy measures all programs exec'd, files
1879 mmap'd for exec, and all files opened with the read
1880 mode bit set by either the effective uid (euid=0) or
1883 The "appraise_tcb" policy appraises the integrity of
1884 all files owned by root.
1886 The "secure_boot" policy appraises the integrity
1887 of files (eg. kexec kernel image, kernel modules,
1888 firmware, policy, etc) based on file signatures.
1890 The "fail_securely" policy forces file signature
1891 verification failure also on privileged mounted
1892 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1895 The "critical_data" policy measures kernel integrity
1898 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1899 Load a policy which meets the needs of the Trusted
1900 Computing Base. This means IMA will measure all
1901 programs exec'd, files mmap'd for exec, and all files
1902 opened for read by uid=0.
1905 Select one of defined IMA measurements template formats.
1906 Formats: { "ima" | "ima-ng" | "ima-sig" }
1910 [IMA] Define a custom template format.
1911 Format: { "field1|...|fieldN" }
1913 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1914 Format: <min_file_size>
1915 Set the minimal file size for using asynchronous hash.
1916 If left unspecified, ahash usage is disabled.
1918 ahash performance varies for different data sizes on
1919 different crypto accelerators. This option can be used
1920 to achieve the best performance for a particular HW.
1922 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1924 Set hashing buffer size. Default: 4k.
1926 ahash performance varies for different chunk sizes on
1927 different crypto accelerators. This option can be used
1928 to achieve best performance for particular HW.
1932 Run specified binary instead of /sbin/init as init
1935 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1936 for working out where the kernel is dying during
1939 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1940 initcall functions. Useful for debugging built-in
1941 modules and initcalls.
1943 initramfs_async= [KNL]
1946 This parameter controls whether the initramfs
1947 image is unpacked asynchronously, concurrently
1948 with devices being probed and
1949 initialized. This should normally just work,
1950 but as a debugging aid, one can get the
1951 historical behaviour of the initramfs
1952 unpacking being completed before device_ and
1955 initrd= [BOOT] Specify the location of the initial ramdisk
1957 initrdmem= [KNL] Specify a physical address and size from which to
1958 load the initrd. If an initrd is compiled in or
1959 specified in the bootparams, it takes priority over this
1961 Format: ss[KMG],nn[KMG]
1964 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1967 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1969 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1971 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1973 init_pkru= [X86] Specify the default memory protection keys rights
1974 register contents for all processes. 0x55555554 by
1975 default (disallow access to all but pkey 0). Can
1976 override in debugfs after boot.
1978 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1981 int_pln_enable [X86] Enable power limit notification interrupt
1983 integrity_audit=[IMA]
1984 Format: { "0" | "1" }
1985 0 -- basic integrity auditing messages. (Default)
1986 1 -- additional integrity auditing messages.
1988 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1990 Enable intel iommu driver.
1992 Disable intel iommu driver.
1993 igfx_off [Default Off]
1994 By default, gfx is mapped as normal device. If a gfx
1995 device has a dedicated DMAR unit, the DMAR unit is
1996 bypassed by not enabling DMAR with this option. In
1997 this case, gfx device will use physical address for
1999 strict [Default Off]
2000 Deprecated, equivalent to iommu.strict=1.
2001 sp_off [Default Off]
2002 By default, super page will be supported if Intel IOMMU
2003 has the capability. With this option, super page will
2006 Enable the Intel IOMMU scalable mode if the hardware
2007 advertises that it has support for the scalable mode
2010 Disallow use of the Intel IOMMU scalable mode.
2011 tboot_noforce [Default Off]
2012 Do not force the Intel IOMMU enabled under tboot.
2013 By default, tboot will force Intel IOMMU on, which
2014 could harm performance of some high-throughput
2015 devices like 40GBit network cards, even if identity
2017 Note that using this option lowers the security
2018 provided by tboot because it makes the system
2019 vulnerable to DMA attacks.
2021 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2022 0 disables intel_idle and fall back on acpi_idle.
2023 1 to 9 specify maximum depth of C-state.
2027 Do not enable intel_pstate as the default
2028 scaling driver for the supported processors
2030 Use intel_pstate as a scaling driver, but configure it
2031 to work with generic cpufreq governors (instead of
2032 enabling its internal governor). This mode cannot be
2033 used along with the hardware-managed P-states (HWP)
2036 Enable intel_pstate on systems that prohibit it by default
2037 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2038 instead of acpi-cpufreq may disable platform features, such
2039 as thermal controls and power capping, that rely on ACPI
2040 P-States information being indicated to OSPM and therefore
2041 should be used with caution. This option does not work with
2042 processors that aren't supported by the intel_pstate driver
2043 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2045 Do not enable hardware P state control (HWP)
2048 Only load intel_pstate on systems which support
2049 hardware P state control (HWP) if available.
2051 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2052 Description Table, specifies preferred power management
2053 profile as "Enterprise Server" or "Performance Server",
2054 then this feature is turned on by default.
2056 Allow per-logical-CPU P-State performance control limits using
2057 cpufreq sysfs interface
2059 intremap= [X86-64, Intel-IOMMU]
2060 on enable Interrupt Remapping (default)
2061 off disable Interrupt Remapping
2062 nosid disable Source ID checking
2064 BIOS x2APIC opt-out request will be ignored
2065 nopost disable Interrupt Posting
2067 iomem= Disable strict checking of access to MMIO memory
2068 strict regions from userspace.
2083 nobypass [PPC/POWERNV]
2084 Disable IOMMU bypass, using IOMMU for PCI devices.
2086 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2087 Format: { "0" | "1" }
2088 0 - Try to allocate a 32-bit DMA address first, before
2089 falling back to the full range if needed.
2090 1 - Allocate directly from the full usable range,
2091 forcing Dual Address Cycle for PCI cards supporting
2092 greater than 32-bit addressing.
2094 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2095 Format: { "0" | "1" }
2097 Request that DMA unmap operations use deferred
2098 invalidation of hardware TLBs, for increased
2099 throughput at the cost of reduced device isolation.
2100 Will fall back to strict mode if not supported by
2101 the relevant IOMMU driver.
2103 DMA unmap operations invalidate IOMMU hardware TLBs
2105 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2106 Note: on x86, strict mode specified via one of the
2107 legacy driver-specific options takes precedence.
2110 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2111 Format: { "0" | "1" }
2112 0 - Use IOMMU translation for DMA.
2113 1 - Bypass the IOMMU for DMA.
2114 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2116 io7= [HW] IO7 for Marvel-based Alpha systems
2117 See comment before marvel_specify_io7 in
2118 arch/alpha/kernel/core_marvel.c.
2120 io_delay= [X86] I/O delay method
2122 Standard port 0x80 based delay
2124 Alternate port 0xed based delay (needed on some systems)
2126 Simple two microseconds delay
2131 See Documentation/admin-guide/nfs/nfsroot.rst.
2133 ipcmni_extend [KNL] Extend the maximum number of unique System V
2134 IPC identifiers from 32,768 to 16,777,216.
2136 irqaffinity= [SMP] Set the default irq affinity mask
2137 The argument is a cpu list, as described above.
2139 irqchip.gicv2_force_probe=
2142 Force the kernel to look for the second 4kB page
2143 of a GICv2 controller even if the memory range
2144 exposed by the device tree is too small.
2146 irqchip.gicv3_nolpi=
2148 Force the kernel to ignore the availability of
2149 LPIs (and by consequence ITSs). Intended for system
2150 that use the kernel as a bootloader, and thus want
2151 to let secondary kernels in charge of setting up
2154 irqchip.gicv3_pseudo_nmi= [ARM64]
2155 Enables support for pseudo-NMIs in the kernel. This
2156 requires the kernel to be built with
2157 CONFIG_ARM64_PSEUDO_NMI.
2160 When an interrupt is not handled search all handlers
2161 for it. Intended to get systems with badly broken
2165 When an interrupt is not handled search all handlers
2166 for it. Also check all handlers each timer
2167 interrupt. Intended to get systems with badly broken
2171 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2173 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2174 [Deprecated - use cpusets instead]
2175 Format: [flag-list,]<cpu-list>
2177 Specify one or more CPUs to isolate from disturbances
2178 specified in the flag list (default: domain):
2181 Disable the tick when a single task runs.
2183 A residual 1Hz tick is offloaded to workqueues, which you
2184 need to affine to housekeeping through the global
2185 workqueue's affinity configured via the
2186 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2187 by using the 'domain' flag described below.
2189 NOTE: by default the global workqueue runs on all CPUs,
2190 so to protect individual CPUs the 'cpumask' file has to
2191 be configured manually after bootup.
2194 Isolate from the general SMP balancing and scheduling
2195 algorithms. Note that performing domain isolation this way
2196 is irreversible: it's not possible to bring back a CPU to
2197 the domains once isolated through isolcpus. It's strongly
2198 advised to use cpusets instead to disable scheduler load
2199 balancing through the "cpuset.sched_load_balance" file.
2200 It offers a much more flexible interface where CPUs can
2201 move in and out of an isolated set anytime.
2203 You can move a process onto or off an "isolated" CPU via
2204 the CPU affinity syscalls or cpuset.
2205 <cpu number> begins at 0 and the maximum value is
2206 "number of CPUs in system - 1".
2210 Isolate from being targeted by managed interrupts
2211 which have an interrupt mask containing isolated
2212 CPUs. The affinity of managed interrupts is
2213 handled by the kernel and cannot be changed via
2214 the /proc/irq/* interfaces.
2216 This isolation is best effort and only effective
2217 if the automatically assigned interrupt mask of a
2218 device queue contains isolated and housekeeping
2219 CPUs. If housekeeping CPUs are online then such
2220 interrupts are directed to the housekeeping CPU
2221 so that IO submitted on the housekeeping CPU
2222 cannot disturb the isolated CPU.
2224 If a queue's affinity mask contains only isolated
2225 CPUs then this parameter has no effect on the
2226 interrupt routing decision, though interrupts are
2227 only delivered when tasks running on those
2228 isolated CPUs submit IO. IO submitted on
2229 housekeeping CPUs has no influence on those
2232 The format of <cpu-list> is described above.
2236 ivrs_ioapic [HW,X86-64]
2237 Provide an override to the IOAPIC-ID<->DEVICE-ID
2238 mapping provided in the IVRS ACPI table. For
2239 example, to map IOAPIC-ID decimal 10 to
2240 PCI device 00:14.0 write the parameter as:
2241 ivrs_ioapic[10]=00:14.0
2243 ivrs_hpet [HW,X86-64]
2244 Provide an override to the HPET-ID<->DEVICE-ID
2245 mapping provided in the IVRS ACPI table. For
2246 example, to map HPET-ID decimal 0 to
2247 PCI device 00:14.0 write the parameter as:
2248 ivrs_hpet[0]=00:14.0
2250 ivrs_acpihid [HW,X86-64]
2251 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2252 mapping provided in the IVRS ACPI table. For
2253 example, to map UART-HID:UID AMD0020:0 to
2254 PCI device 00:14.5 write the parameter as:
2255 ivrs_acpihid[00:14.5]=AMD0020:0
2257 js= [HW,JOY] Analog joystick
2258 See Documentation/input/joydev/joystick.rst.
2261 When CONFIG_RANDOMIZE_BASE is set, this disables
2262 kernel and module base offset ASLR (Address Space
2263 Layout Randomization).
2266 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2267 report on every invalid memory access. Without this
2268 parameter KASAN will print report only for the first
2273 kernelcore= [KNL,X86,IA-64,PPC]
2274 Format: nn[KMGTPE] | nn% | "mirror"
2275 This parameter specifies the amount of memory usable by
2276 the kernel for non-movable allocations. The requested
2277 amount is spread evenly throughout all nodes in the
2278 system as ZONE_NORMAL. The remaining memory is used for
2279 movable memory in its own zone, ZONE_MOVABLE. In the
2280 event, a node is too small to have both ZONE_NORMAL and
2281 ZONE_MOVABLE, kernelcore memory will take priority and
2282 other nodes will have a larger ZONE_MOVABLE.
2284 ZONE_MOVABLE is used for the allocation of pages that
2285 may be reclaimed or moved by the page migration
2286 subsystem. Note that allocations like PTEs-from-HighMem
2287 still use the HighMem zone if it exists, and the Normal
2288 zone if it does not.
2290 It is possible to specify the exact amount of memory in
2291 the form of "nn[KMGTPE]", a percentage of total system
2292 memory in the form of "nn%", or "mirror". If "mirror"
2293 option is specified, mirrored (reliable) memory is used
2294 for non-movable allocations and remaining memory is used
2295 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2296 are exclusive, so you cannot specify multiple forms.
2298 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2299 Format: <Controller#>[,poll interval]
2300 The controller # is the number of the ehci usb debug
2301 port as it is probed via PCI. The poll interval is
2302 optional and is the number seconds in between
2303 each poll cycle to the debug port in case you need
2304 the functionality for interrupting the kernel with
2305 gdb or control-c on the dbgp connection. When
2306 not using this parameter you use sysrq-g to break into
2307 the kernel debugger.
2309 kgdboc= [KGDB,HW] kgdb over consoles.
2310 Requires a tty driver that supports console polling,
2311 or a supported polling keyboard driver (non-usb).
2312 Serial only format: <serial_device>[,baud]
2313 keyboard only format: kbd
2314 keyboard and serial format: kbd,<serial_device>[,baud]
2315 Optional Kernel mode setting:
2316 kms, kbd format: kms,kbd
2317 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2319 kgdboc_earlycon= [KGDB,HW]
2320 If the boot console provides the ability to read
2321 characters and can work in polling mode, you can use
2322 this parameter to tell kgdb to use it as a backend
2323 until the normal console is registered. Intended to
2324 be used together with the kgdboc parameter which
2325 specifies the normal console to transition to.
2327 The name of the early console should be specified
2328 as the value of this parameter. Note that the name of
2329 the early console might be different than the tty
2330 name passed to kgdboc. It's OK to leave the value
2331 blank and the first boot console that implements
2332 read() will be picked.
2334 kgdbwait [KGDB] Stop kernel execution and enter the
2335 kernel debugger at the earliest opportunity.
2337 kmac= [MIPS] Korina ethernet MAC address.
2338 Configure the RouterBoard 532 series on-chip
2339 Ethernet adapter MAC address.
2341 kmemleak= [KNL] Boot-time kmemleak enable/disable
2342 Valid arguments: on, off
2344 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2347 kprobe_event=[probe-list]
2348 [FTRACE] Add kprobe events and enable at boot time.
2349 The probe-list is a semicolon delimited list of probe
2350 definitions. Each definition is same as kprobe_events
2351 interface, but the parameters are comma delimited.
2352 For example, to add a kprobe event on vfs_read with
2353 arg1 and arg2, add to the command line;
2355 kprobe_event=p,vfs_read,$arg1,$arg2
2357 See also Documentation/trace/kprobetrace.rst "Kernel
2358 Boot Parameter" section.
2360 kpti= [ARM64] Control page table isolation of user
2361 and kernel address spaces.
2362 Default: enabled on cores which need mitigation.
2366 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2367 Default is 0 (don't ignore, but inject #GP)
2369 kvm.eager_page_split=
2370 [KVM,X86] Controls whether or not KVM will try to
2371 proactively split all huge pages during dirty logging.
2372 Eager page splitting reduces interruptions to vCPU
2373 execution by eliminating the write-protection faults
2374 and MMU lock contention that would otherwise be
2375 required to split huge pages lazily.
2377 VM workloads that rarely perform writes or that write
2378 only to a small region of VM memory may benefit from
2379 disabling eager page splitting to allow huge pages to
2380 still be used for reads.
2382 The behavior of eager page splitting depends on whether
2383 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2384 disabled, all huge pages in a memslot will be eagerly
2385 split when dirty logging is enabled on that memslot. If
2386 enabled, eager page splitting will be performed during
2387 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2390 Eager page splitting currently only supports splitting
2391 huge pages mapped by the TDP MMU.
2395 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2396 Default is false (don't support).
2399 [KVM] Controls the software workaround for the
2400 X86_BUG_ITLB_MULTIHIT bug.
2401 force : Always deploy workaround.
2402 off : Never deploy workaround.
2403 auto : Deploy workaround based on the presence of
2404 X86_BUG_ITLB_MULTIHIT.
2408 If the software workaround is enabled for the host,
2409 guests do need not to enable it for nested guests.
2411 kvm.nx_huge_pages_recovery_ratio=
2412 [KVM] Controls how many 4KiB pages are periodically zapped
2413 back to huge pages. 0 disables the recovery, otherwise if
2414 the value is N KVM will zap 1/Nth of the 4KiB pages every
2415 period (see below). The default is 60.
2417 kvm.nx_huge_pages_recovery_period_ms=
2418 [KVM] Controls the time period at which KVM zaps 4KiB pages
2419 back to huge pages. If the value is a non-zero N, KVM will
2420 zap a portion (see ratio above) of the pages every N msecs.
2421 If the value is 0 (the default), KVM will pick a period based
2422 on the ratio, such that a page is zapped after 1 hour on average.
2424 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2425 Default is 1 (enabled)
2427 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2429 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2432 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2434 none: Forcefully disable KVM.
2436 nvhe: Standard nVHE-based mode, without support for
2439 protected: nVHE-based mode with support for guests whose
2440 state is kept private from the host.
2441 Not valid if the kernel is running in EL2.
2443 Defaults to VHE/nVHE based on hardware support. Setting
2444 mode to "protected" will disable kexec and hibernation
2447 kvm-arm.vgic_v3_group0_trap=
2448 [KVM,ARM] Trap guest accesses to GICv3 group-0
2451 kvm-arm.vgic_v3_group1_trap=
2452 [KVM,ARM] Trap guest accesses to GICv3 group-1
2455 kvm-arm.vgic_v3_common_trap=
2456 [KVM,ARM] Trap guest accesses to GICv3 common
2459 kvm-arm.vgic_v4_enable=
2460 [KVM,ARM] Allow use of GICv4 for direct injection of
2463 kvm_cma_resv_ratio=n [PPC]
2464 Reserves given percentage from system memory area for
2465 contiguous memory allocation for KVM hash pagetable
2467 By default it reserves 5% of total system memory.
2471 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2472 (virtualized MMU) support on capable Intel chips.
2473 Default is 1 (enabled)
2475 kvm-intel.emulate_invalid_guest_state=
2476 [KVM,Intel] Disable emulation of invalid guest state.
2477 Ignored if kvm-intel.enable_unrestricted_guest=1, as
2478 guest state is never invalid for unrestricted guests.
2479 This param doesn't apply to nested guests (L2), as KVM
2480 never emulates invalid L2 guest state.
2481 Default is 1 (enabled)
2483 kvm-intel.flexpriority=
2484 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2485 Default is 1 (enabled)
2488 [KVM,Intel] Enable VMX nesting (nVMX).
2489 Default is 0 (disabled)
2491 kvm-intel.unrestricted_guest=
2492 [KVM,Intel] Disable unrestricted guest feature
2493 (virtualized real and unpaged mode) on capable
2494 Intel chips. Default is 1 (enabled)
2496 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2499 Valid arguments: never, cond, always
2501 always: L1D cache flush on every VMENTER.
2502 cond: Flush L1D on VMENTER only when the code between
2503 VMEXIT and VMENTER can leak host memory.
2504 never: Disables the mitigation
2506 Default is cond (do L1 cache flush in specific instances)
2508 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2509 feature (tagged TLBs) on capable Intel chips.
2510 Default is 1 (enabled)
2512 l1d_flush= [X86,INTEL]
2513 Control mitigation for L1D based snooping vulnerability.
2515 Certain CPUs are vulnerable to an exploit against CPU
2516 internal buffers which can forward information to a
2517 disclosure gadget under certain conditions.
2519 In vulnerable processors, the speculatively
2520 forwarded data can be used in a cache side channel
2521 attack, to access data to which the attacker does
2522 not have direct access.
2524 This parameter controls the mitigation. The
2527 on - enable the interface for the mitigation
2529 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2532 The kernel PTE inversion protection is unconditionally
2533 enabled and cannot be disabled.
2536 Provides all available mitigations for the
2537 L1TF vulnerability. Disables SMT and
2538 enables all mitigations in the
2539 hypervisors, i.e. unconditional L1D flush.
2541 SMT control and L1D flush control via the
2542 sysfs interface is still possible after
2543 boot. Hypervisors will issue a warning
2544 when the first VM is started in a
2545 potentially insecure configuration,
2546 i.e. SMT enabled or L1D flush disabled.
2549 Same as 'full', but disables SMT and L1D
2550 flush runtime control. Implies the
2551 'nosmt=force' command line option.
2552 (i.e. sysfs control of SMT is disabled.)
2555 Leaves SMT enabled and enables the default
2556 hypervisor mitigation, i.e. conditional
2559 SMT control and L1D flush control via the
2560 sysfs interface is still possible after
2561 boot. Hypervisors will issue a warning
2562 when the first VM is started in a
2563 potentially insecure configuration,
2564 i.e. SMT enabled or L1D flush disabled.
2568 Disables SMT and enables the default
2569 hypervisor mitigation.
2571 SMT control and L1D flush control via the
2572 sysfs interface is still possible after
2573 boot. Hypervisors will issue a warning
2574 when the first VM is started in a
2575 potentially insecure configuration,
2576 i.e. SMT enabled or L1D flush disabled.
2579 Same as 'flush', but hypervisors will not
2580 warn when a VM is started in a potentially
2581 insecure configuration.
2584 Disables hypervisor mitigations and doesn't
2586 It also drops the swap size and available
2587 RAM limit restriction on both hypervisor and
2592 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2598 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2601 lapic= [X86,APIC] Do not use TSC deadline
2602 value for LAPIC timer one-shot implementation. Default
2603 back to the programmable timer unit in the LAPIC.
2604 Format: notscdeadline
2606 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2609 libata.dma= [LIBATA] DMA control
2610 libata.dma=0 Disable all PATA and SATA DMA
2611 libata.dma=1 PATA and SATA Disk DMA only
2612 libata.dma=2 ATAPI (CDROM) DMA only
2613 libata.dma=4 Compact Flash DMA only
2614 Combinations also work, so libata.dma=3 enables DMA
2615 for disks and CDROMs, but not CFs.
2617 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2618 libata.ignore_hpa=0 keep BIOS limits (default)
2619 libata.ignore_hpa=1 ignore limits, using full disk
2621 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2625 libata.force= [LIBATA] Force configurations. The format is comma-
2626 separated list of "[ID:]VAL" where ID is
2627 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2628 matching port, link or device. Basically, it matches
2629 the ATA ID string printed on console by libata. If
2630 the whole ID part is omitted, the last PORT and DEVICE
2631 values are used. If ID hasn't been specified yet, the
2632 configuration applies to all ports, links and devices.
2634 If only DEVICE is omitted, the parameter applies to
2635 the port and all links and devices behind it. DEVICE
2636 number of 0 either selects the first device or the
2637 first fan-out link behind PMP device. It does not
2638 select the host link. DEVICE number of 15 selects the
2639 host link and device attached to it.
2641 The VAL specifies the configuration to force. As long
2642 as there's no ambiguity shortcut notation is allowed.
2643 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2644 The following configurations can be forced.
2646 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2647 Any ID with matching PORT is used.
2649 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2651 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2652 udma[/][16,25,33,44,66,100,133] notation is also
2655 * [no]ncq: Turn on or off NCQ.
2657 * [no]ncqtrim: Turn off queued DSM TRIM.
2659 * nohrst, nosrst, norst: suppress hard, soft
2662 * rstonce: only attempt one reset during
2663 hot-unplug link recovery
2665 * dump_id: dump IDENTIFY data.
2667 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2669 * disable: Disable this device.
2671 If there are multiple matching configurations changing
2672 the same attribute, the last one is used.
2674 memblock=debug [KNL] Enable memblock debug messages.
2676 load_ramdisk= [RAM] [Deprecated]
2678 lockd.nlm_grace_period=P [NFS] Assign grace period.
2681 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2684 lockd.nlm_timeout=T [NFS] Assign timeout value.
2687 lockd.nlm_udpport=M [NFS] Assign UDP port.
2690 lockdown= [SECURITY]
2691 { integrity | confidentiality }
2692 Enable the kernel lockdown feature. If set to
2693 integrity, kernel features that allow userland to
2694 modify the running kernel are disabled. If set to
2695 confidentiality, kernel features that allow userland
2696 to extract confidential information from the kernel
2699 locktorture.nreaders_stress= [KNL]
2700 Set the number of locking read-acquisition kthreads.
2701 Defaults to being automatically set based on the
2702 number of online CPUs.
2704 locktorture.nwriters_stress= [KNL]
2705 Set the number of locking write-acquisition kthreads.
2707 locktorture.onoff_holdoff= [KNL]
2708 Set time (s) after boot for CPU-hotplug testing.
2710 locktorture.onoff_interval= [KNL]
2711 Set time (s) between CPU-hotplug operations, or
2712 zero to disable CPU-hotplug testing.
2714 locktorture.shuffle_interval= [KNL]
2715 Set task-shuffle interval (jiffies). Shuffling
2716 tasks allows some CPUs to go into dyntick-idle
2717 mode during the locktorture test.
2719 locktorture.shutdown_secs= [KNL]
2720 Set time (s) after boot system shutdown. This
2721 is useful for hands-off automated testing.
2723 locktorture.stat_interval= [KNL]
2724 Time (s) between statistics printk()s.
2726 locktorture.stutter= [KNL]
2727 Time (s) to stutter testing, for example,
2728 specifying five seconds causes the test to run for
2729 five seconds, wait for five seconds, and so on.
2730 This tests the locking primitive's ability to
2731 transition abruptly to and from idle.
2733 locktorture.torture_type= [KNL]
2734 Specify the locking implementation to test.
2736 locktorture.verbose= [KNL]
2737 Enable additional printk() statements.
2739 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2742 loglevel= All Kernel Messages with a loglevel smaller than the
2743 console loglevel will be printed to the console. It can
2744 also be changed with klogd or other programs. The
2745 loglevels are defined as follows:
2747 0 (KERN_EMERG) system is unusable
2748 1 (KERN_ALERT) action must be taken immediately
2749 2 (KERN_CRIT) critical conditions
2750 3 (KERN_ERR) error conditions
2751 4 (KERN_WARNING) warning conditions
2752 5 (KERN_NOTICE) normal but significant condition
2753 6 (KERN_INFO) informational
2754 7 (KERN_DEBUG) debug-level messages
2756 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2757 in bytes. n must be a power of two and greater
2758 than the minimal size. The minimal size is defined
2759 by LOG_BUF_SHIFT kernel config parameter. There is
2760 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2761 that allows to increase the default size depending on
2762 the number of CPUs. See init/Kconfig for more details.
2764 logo.nologo [FB] Disables display of the built-in Linux logo.
2765 This may be used to provide more screen space for
2766 kernel log messages and is useful when debugging
2767 kernel boot problems.
2769 lp=0 [LP] Specify parallel ports to use, e.g,
2770 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2771 lp=reset first parallel port). 'lp=0' disables the
2772 lp=auto printer driver. 'lp=reset' (which can be
2773 specified in addition to the ports) causes
2774 attached printers to be reset. Using
2775 lp=port1,port2,... specifies the parallel ports
2776 to associate lp devices with, starting with
2777 lp0. A port specification may be 'none' to skip
2778 that lp device, or a parport name such as
2779 'parport0'. Specifying 'lp=auto' instead of a
2780 port specification list means that device IDs
2781 from each port should be examined, to see if
2782 an IEEE 1284-compliant printer is attached; if
2783 so, the driver will manage that printer.
2784 See also header of drivers/char/lp.c.
2787 Sets loops_per_jiffy to given constant, thus avoiding
2788 time-consuming boot-time autodetection (up to 250 ms per
2789 CPU). 0 enables autodetection (default). To determine
2790 the correct value for your kernel, boot with normal
2791 autodetection and see what value is printed. Note that
2792 on SMP systems the preset will be applied to all CPUs,
2793 which is likely to cause problems if your CPUs need
2794 significantly divergent settings. An incorrect value
2795 will cause delays in the kernel to be wrong, leading to
2796 unpredictable I/O errors and other breakage. Although
2797 unlikely, in the extreme case this might damage your
2801 Format: <io>,<irq>,<dma>
2803 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2806 [SECURITY] Choose order of LSM initialization. This
2807 overrides CONFIG_LSM, and the "security=" parameter.
2809 machvec= [IA-64] Force the use of a particular machine-vector
2810 (machvec) in a generic kernel.
2811 Example: machvec=hpzx1
2813 machtype= [Loongson] Share the same kernel image file between
2814 different yeeloong laptops.
2815 Example: machtype=lemote-yeeloong-2f-7inch
2817 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2818 than or equal to this physical address is ignored.
2820 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2821 will bring up during bootup. maxcpus=n : n >= 0 limits
2822 the kernel to bring up 'n' processors. Surely after
2823 bootup you can bring up the other plugged cpu by executing
2824 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2825 only takes effect during system bootup.
2826 While n=0 is a special case, it is equivalent to "nosmp",
2827 which also disables the IO APIC.
2829 max_loop= [LOOP] The number of loop block devices that get
2830 (loop.max_loop) unconditionally pre-created at init time. The default
2831 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2832 of statically allocating a predefined number, loop
2833 devices can be requested on-demand with the
2834 /dev/loop-control interface.
2836 mce [X86-32] Machine Check Exception
2838 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2840 md= [HW] RAID subsystems devices and level
2841 See Documentation/admin-guide/md.rst.
2844 Format: <first>,<last>
2845 Specifies range of consoles to be captured by the MDA.
2848 Control mitigation for the Micro-architectural Data
2849 Sampling (MDS) vulnerability.
2851 Certain CPUs are vulnerable to an exploit against CPU
2852 internal buffers which can forward information to a
2853 disclosure gadget under certain conditions.
2855 In vulnerable processors, the speculatively
2856 forwarded data can be used in a cache side channel
2857 attack, to access data to which the attacker does
2858 not have direct access.
2860 This parameter controls the MDS mitigation. The
2863 full - Enable MDS mitigation on vulnerable CPUs
2864 full,nosmt - Enable MDS mitigation and disable
2865 SMT on vulnerable CPUs
2866 off - Unconditionally disable MDS mitigation
2868 On TAA-affected machines, mds=off can be prevented by
2869 an active TAA mitigation as both vulnerabilities are
2870 mitigated with the same mechanism so in order to disable
2871 this mitigation, you need to specify tsx_async_abort=off
2874 Not specifying this option is equivalent to
2877 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2879 mem=nn[KMG] [HEXAGON] Set the memory size.
2880 Must be specified, otherwise memory size will be 0.
2882 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2883 Amount of memory to be used in cases as follows:
2886 2 when the kernel is not able to see the whole system memory;
2887 3 memory that lies after 'mem=' boundary is excluded from
2888 the hypervisor, then assigned to KVM guests.
2889 4 to limit the memory available for kdump kernel.
2891 [ARC,MICROBLAZE] - the limit applies only to low memory,
2892 high memory is not affected.
2894 [ARM64] - only limits memory covered by the linear
2895 mapping. The NOMAP regions are not affected.
2897 [X86] Work as limiting max address. Use together
2898 with memmap= to avoid physical address space collisions.
2899 Without memmap= PCI devices could be placed at addresses
2900 belonging to unused RAM.
2902 Note that this only takes effects during boot time since
2903 in above case 3, memory may need be hot added after boot
2904 if system memory of hypervisor is not sufficient.
2907 [ARM,MIPS] - override the memory layout reported by
2909 Define a memory region of size nn[KMG] starting at
2911 Multiple different regions can be specified with
2912 multiple mem= parameters on the command line.
2914 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2918 [KNL,SH] Allow user to override the default size for
2919 per-device physically contiguous DMA buffers.
2921 memhp_default_state=online/offline
2922 [KNL] Set the initial state for the memory hotplug
2923 onlining policy. If not specified, the default value is
2924 set according to the
2925 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2927 See Documentation/admin-guide/mm/memory-hotplug.rst.
2929 memmap=exactmap [KNL,X86] Enable setting of an exact
2930 E820 memory map, as specified by the user.
2931 Such memmap=exactmap lines can be constructed based on
2932 BIOS output or other requirements. See the memmap=nn@ss
2935 memmap=nn[KMG]@ss[KMG]
2936 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
2937 Region of memory to be used is from ss to ss+nn.
2938 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2939 which limits max address to nn[KMG].
2940 Multiple different regions can be specified,
2943 memmap=100M@2G,100M#3G,1G!1024G
2945 memmap=nn[KMG]#ss[KMG]
2946 [KNL,ACPI] Mark specific memory as ACPI data.
2947 Region of memory to be marked is from ss to ss+nn.
2949 memmap=nn[KMG]$ss[KMG]
2950 [KNL,ACPI] Mark specific memory as reserved.
2951 Region of memory to be reserved is from ss to ss+nn.
2952 Example: Exclude memory from 0x18690000-0x1869ffff
2953 memmap=64K$0x18690000
2955 memmap=0x10000$0x18690000
2956 Some bootloaders may need an escape character before '$',
2957 like Grub2, otherwise '$' and the following number
2960 memmap=nn[KMG]!ss[KMG]
2961 [KNL,X86] Mark specific memory as protected.
2962 Region of memory to be used, from ss to ss+nn.
2963 The memory region may be marked as e820 type 12 (0xc)
2964 and is NVDIMM or ADR memory.
2966 memmap=<size>%<offset>-<oldtype>+<newtype>
2967 [KNL,ACPI] Convert memory within the specified region
2968 from <oldtype> to <newtype>. If "-<oldtype>" is left
2969 out, the whole region will be marked as <newtype>,
2970 even if previously unavailable. If "+<newtype>" is left
2971 out, matching memory will be removed. Types are
2972 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2973 3 = ACPI, 12 = PRAM.
2975 memory_corruption_check=0/1 [X86]
2976 Some BIOSes seem to corrupt the first 64k of
2977 memory when doing things like suspend/resume.
2978 Setting this option will scan the memory
2979 looking for corruption. Enabling this will
2980 both detect corruption and prevent the kernel
2981 from using the memory being corrupted.
2982 However, its intended as a diagnostic tool; if
2983 repeatable BIOS-originated corruption always
2984 affects the same memory, you can use memmap=
2985 to prevent the kernel from using that memory.
2987 memory_corruption_check_size=size [X86]
2988 By default it checks for corruption in the low
2989 64k, making this memory unavailable for normal
2990 use. Use this parameter to scan for
2991 corruption in more or less memory.
2993 memory_corruption_check_period=seconds [X86]
2994 By default it checks for corruption every 60
2995 seconds. Use this parameter to check at some
2996 other rate. 0 disables periodic checking.
2998 memory_hotplug.memmap_on_memory
2999 [KNL,X86,ARM] Boolean flag to enable this feature.
3000 Format: {on | off (default)}
3001 When enabled, runtime hotplugged memory will
3002 allocate its internal metadata (struct pages)
3003 from the hotadded memory which will allow to
3004 hotadd a lot of memory without requiring
3005 additional memory to do so.
3006 This feature is disabled by default because it
3007 has some implication on large (e.g. GB)
3008 allocations in some configurations (e.g. small
3010 The state of the flag can be read in
3011 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3012 Note that even when enabled, there are a few cases where
3013 the feature is not effective.
3015 This is not compatible with hugetlb_free_vmemmap. If
3016 both parameters are enabled, hugetlb_free_vmemmap takes
3017 precedence over memory_hotplug.memmap_on_memory.
3019 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3021 default : 0 <disable>
3022 Specifies the number of memtest passes to be
3023 performed. Each pass selects another test
3024 pattern from a given set of patterns. Memtest
3025 fills the memory with this pattern, validates
3026 memory contents and reserves bad memory
3027 regions that are detected.
3029 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3030 Valid arguments: on, off
3031 Default (depends on kernel configuration option):
3032 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3033 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3034 mem_encrypt=on: Activate SME
3035 mem_encrypt=off: Do not activate SME
3037 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
3038 for details on when memory encryption can be activated.
3040 mem_sleep_default= [SUSPEND] Default system suspend mode:
3041 s2idle - Suspend-To-Idle
3042 shallow - Power-On Suspend or equivalent (if supported)
3043 deep - Suspend-To-RAM or equivalent (if supported)
3044 See Documentation/admin-guide/pm/sleep-states.rst.
3046 meye.*= [HW] Set MotionEye Camera parameters
3047 See Documentation/admin-guide/media/meye.rst.
3049 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3050 Multi-Function General Purpose Timers on AMD Geode
3053 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3054 the BIOS has incorrectly applied a workaround. TinyBIOS
3055 version 0.98 is known to be affected, 0.99 fixes the
3056 problem by letting the user disable the workaround.
3060 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
3061 physical address is ignored.
3063 mini2440= [ARM,HW,KNL]
3064 Format:[0..2][b][c][t]
3066 MINI2440 configuration specification:
3067 0 - The attached screen is the 3.5" TFT
3068 1 - The attached screen is the 7" TFT
3069 2 - The VGA Shield is attached (1024x768)
3070 Leaving out the screen size parameter will not load
3071 the TFT driver, and the framebuffer will be left
3073 b - Enable backlight. The TFT backlight pin will be
3074 linked to the kernel VESA blanking code and a GPIO
3075 LED. This parameter is not necessary when using the
3077 c - Enable the s3c camera interface.
3078 t - Reserved for enabling touchscreen support. The
3079 touchscreen support is not enabled in the mainstream
3080 kernel as of 2.6.30, a preliminary port can be found
3081 in the "bleeding edge" mini2440 support kernel at
3082 https://repo.or.cz/w/linux-2.6/mini2440.git
3085 [X86,PPC,S390,ARM64] Control optional mitigations for
3086 CPU vulnerabilities. This is a set of curated,
3087 arch-independent options, each of which is an
3088 aggregation of existing arch-specific options.
3091 Disable all optional CPU mitigations. This
3092 improves system performance, but it may also
3093 expose users to several CPU vulnerabilities.
3094 Equivalent to: nopti [X86,PPC]
3096 nospectre_v1 [X86,PPC]
3098 nospectre_v2 [X86,PPC,S390,ARM64]
3099 spectre_v2_user=off [X86]
3100 spec_store_bypass_disable=off [X86,PPC]
3101 ssbd=force-off [ARM64]
3104 tsx_async_abort=off [X86]
3105 kvm.nx_huge_pages=off [X86]
3106 no_entry_flush [PPC]
3107 no_uaccess_flush [PPC]
3110 This does not have any effect on
3111 kvm.nx_huge_pages when
3112 kvm.nx_huge_pages=force.
3115 Mitigate all CPU vulnerabilities, but leave SMT
3116 enabled, even if it's vulnerable. This is for
3117 users who don't want to be surprised by SMT
3118 getting disabled across kernel upgrades, or who
3119 have other ways of avoiding SMT-based attacks.
3120 Equivalent to: (default behavior)
3123 Mitigate all CPU vulnerabilities, disabling SMT
3124 if needed. This is for users who always want to
3125 be fully mitigated, even if it means losing SMT.
3126 Equivalent to: l1tf=flush,nosmt [X86]
3127 mds=full,nosmt [X86]
3128 tsx_async_abort=full,nosmt [X86]
3131 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3132 parameter allows control of the logging verbosity for
3133 the additional memory initialisation checks. A value
3134 of 0 disables mminit logging and a level of 4 will
3135 log everything. Information is printed at KERN_DEBUG
3136 so loglevel=8 may also need to be specified.
3139 [KNL] When CONFIG_MODULE_SIG is set, this means that
3140 modules without (valid) signatures will fail to load.
3141 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3142 is always true, so this option does nothing.
3144 module_blacklist= [KNL] Do not load a comma-separated list of
3145 modules. Useful for debugging problem modules.
3148 [MOUSE] Maximum time between finger touching and
3149 leaving touchpad surface for touch to be considered
3150 a tap and be reported as a left button click (for
3151 touchpads working in absolute mode only).
3153 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3154 reporting absolute coordinates, such as tablets
3155 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3156 reporting absolute coordinates, such as tablets
3158 movablecore= [KNL,X86,IA-64,PPC]
3159 Format: nn[KMGTPE] | nn%
3160 This parameter is the complement to kernelcore=, it
3161 specifies the amount of memory used for migratable
3162 allocations. If both kernelcore and movablecore is
3163 specified, then kernelcore will be at *least* the
3164 specified value but may be more. If movablecore on its
3165 own is specified, the administrator must be careful
3166 that the amount of memory usable for all allocations
3169 movable_node [KNL] Boot-time switch to make hotplugable memory
3170 NUMA nodes to be movable. This means that the memory
3171 of such nodes will be usable only for movable
3172 allocations which rules out almost all kernel
3173 allocations. Use with caution!
3175 MTD_Partition= [MTD]
3176 Format: <name>,<region-number>,<size>,<offset>
3178 MTD_Region= [MTD] Format:
3179 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3182 See drivers/mtd/parsers/cmdlinepart.c
3184 multitce=off [PPC] This parameter disables the use of the pSeries
3185 firmware feature for updating multiple TCE entries
3188 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3190 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3192 boundary - index of last SLC block on Flex-OneNAND.
3193 The remaining blocks are configured as MLC blocks.
3194 lock - Configure if Flex-OneNAND boundary should be locked.
3195 Once locked, the boundary cannot be changed.
3196 1 indicates lock status, 0 indicates unlock status.
3199 ARM/S3C2412 JIVE boot control
3201 See arch/arm/mach-s3c/mach-jive.c
3203 mtouchusb.raw_coordinates=
3204 [HW] Make the MicroTouch USB driver use raw coordinates
3205 ('y', default) or cooked coordinates ('n')
3207 mtrr_chunk_size=nn[KMG] [X86]
3208 used for mtrr cleanup. It is largest continuous chunk
3209 that could hold holes aka. UC entries.
3211 mtrr_gran_size=nn[KMG] [X86]
3212 Used for mtrr cleanup. It is granularity of mtrr block.
3214 Large value could prevent small alignment from
3217 mtrr_spare_reg_nr=n [X86]
3219 Range: 0,7 : spare reg number
3221 Used for mtrr cleanup. It is spare mtrr entries number.
3222 Set to 2 or more if your graphical card needs more.
3224 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3226 netdev= [NET] Network devices parameters
3227 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3228 Note that mem_start is often overloaded to mean
3229 something different and driver-specific.
3230 This usage is only documented in each driver source
3234 [NETFILTER] Enable connection tracking flow accounting
3235 0 to disable accounting
3236 1 to enable accounting
3239 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3240 See Documentation/admin-guide/nfs/nfsroot.rst.
3242 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3243 See Documentation/admin-guide/nfs/nfsroot.rst.
3245 nfsrootdebug [NFS] enable nfsroot debugging messages.
3246 See Documentation/admin-guide/nfs/nfsroot.rst.
3248 nfs.callback_nr_threads=
3249 [NFSv4] set the total number of threads that the
3250 NFS client will assign to service NFSv4 callback
3253 nfs.callback_tcpport=
3254 [NFS] set the TCP port on which the NFSv4 callback
3255 channel should listen.
3258 [NFS] sets the pathname to the program which is used
3259 to update the NFS client cache entries.
3261 nfs.cache_getent_timeout=
3262 [NFS] sets the timeout after which an attempt to
3263 update a cache entry is deemed to have failed.
3265 nfs.idmap_cache_timeout=
3266 [NFS] set the maximum lifetime for idmapper cache
3270 [NFS] enable 64-bit inode numbers.
3271 If zero, the NFS client will fake up a 32-bit inode
3272 number for the readdir() and stat() syscalls instead
3273 of returning the full 64-bit number.
3274 The default is to return 64-bit inode numbers.
3276 nfs.max_session_cb_slots=
3277 [NFSv4.1] Sets the maximum number of session
3278 slots the client will assign to the callback
3279 channel. This determines the maximum number of
3280 callbacks the client will process in parallel for
3281 a particular server.
3283 nfs.max_session_slots=
3284 [NFSv4.1] Sets the maximum number of session slots
3285 the client will attempt to negotiate with the server.
3286 This limits the number of simultaneous RPC requests
3287 that the client can send to the NFSv4.1 server.
3288 Note that there is little point in setting this
3289 value higher than the max_tcp_slot_table_limit.
3291 nfs.nfs4_disable_idmapping=
3292 [NFSv4] When set to the default of '1', this option
3293 ensures that both the RPC level authentication
3294 scheme and the NFS level operations agree to use
3295 numeric uids/gids if the mount is using the
3296 'sec=sys' security flavour. In effect it is
3297 disabling idmapping, which can make migration from
3298 legacy NFSv2/v3 systems to NFSv4 easier.
3299 Servers that do not support this mode of operation
3300 will be autodetected by the client, and it will fall
3301 back to using the idmapper.
3302 To turn off this behaviour, set the value to '0'.
3304 [NFS4] Specify an additional fixed unique ident-
3305 ification string that NFSv4 clients can insert into
3306 their nfs_client_id4 string. This is typically a
3307 UUID that is generated at system install time.
3309 nfs.send_implementation_id =
3310 [NFSv4.1] Send client implementation identification
3311 information in exchange_id requests.
3312 If zero, no implementation identification information
3314 The default is to send the implementation identification
3317 nfs.recover_lost_locks =
3318 [NFSv4] Attempt to recover locks that were lost due
3319 to a lease timeout on the server. Please note that
3320 doing this risks data corruption, since there are
3321 no guarantees that the file will remain unchanged
3322 after the locks are lost.
3323 If you want to enable the kernel legacy behaviour of
3324 attempting to recover these locks, then set this
3326 The default parameter value of '0' causes the kernel
3327 not to attempt recovery of lost locks.
3329 nfs4.layoutstats_timer =
3330 [NFSv4.2] Change the rate at which the kernel sends
3331 layoutstats to the pNFS metadata server.
3333 Setting this to value to 0 causes the kernel to use
3334 whatever value is the default set by the layout
3335 driver. A non-zero value sets the minimum interval
3336 in seconds between layoutstats transmissions.
3338 nfsd.inter_copy_offload_enable =
3339 [NFSv4.2] When set to 1, the server will support
3340 server-to-server copies for which this server is
3341 the destination of the copy.
3343 nfsd.nfsd4_ssc_umount_timeout =
3344 [NFSv4.2] When used as the destination of a
3345 server-to-server copy, knfsd temporarily mounts
3346 the source server. It caches the mount in case
3347 it will be needed again, and discards it if not
3348 used for the number of milliseconds specified by
3351 nfsd.nfs4_disable_idmapping=
3352 [NFSv4] When set to the default of '1', the NFSv4
3353 server will return only numeric uids and gids to
3354 clients using auth_sys, and will accept numeric uids
3355 and gids from such clients. This is intended to ease
3356 migration from NFSv2/v3.
3359 nmi_backtrace.backtrace_idle [KNL]
3360 Dump stacks even of idle CPUs in response to an
3361 NMI stack-backtrace request.
3363 nmi_debug= [KNL,SH] Specify one or more actions to take
3364 when a NMI is triggered.
3365 Format: [state][,regs][,debounce][,die]
3367 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3368 Format: [panic,][nopanic,][num]
3370 0 - turn hardlockup detector in nmi_watchdog off
3371 1 - turn hardlockup detector in nmi_watchdog on
3372 When panic is specified, panic when an NMI watchdog
3373 timeout occurs (or 'nopanic' to not panic on an NMI
3374 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3375 To disable both hard and soft lockup detectors,
3376 please see 'nowatchdog'.
3377 This is useful when you use a panic=... timeout and
3378 need the box quickly up again.
3380 These settings can be accessed at runtime via
3381 the nmi_watchdog and hardlockup_panic sysctls.
3383 netpoll.carrier_timeout=
3384 [NET] Specifies amount of time (in seconds) that
3385 netpoll should wait for a carrier. By default netpoll
3388 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3389 emulation library even if a 387 maths coprocessor
3392 no5lvl [X86-64] Disable 5-level paging mode. Forces
3393 kernel to use 4-level paging instead.
3395 nofsgsbase [X86] Disables FSGSBASE instructions.
3398 [HW] Never suspend the console
3399 Disable suspending of consoles during suspend and
3400 hibernate operations. Once disabled, debugging
3401 messages can reach various consoles while the rest
3402 of the system is being put to sleep (ie, while
3403 debugging driver suspend/resume hooks). This may
3404 not work reliably with all consoles, but is known
3405 to work with serial and VGA consoles.
3406 To facilitate more flexible debugging, we also add
3407 console_suspend, a printk module parameter to control
3408 it. Users could use console_suspend (usually
3409 /sys/module/printk/parameters/console_suspend) to
3410 turn on/off it dynamically.
3412 novmcoredd [KNL,KDUMP]
3413 Disable device dump. Device dump allows drivers to
3414 append dump data to vmcore so you can collect driver
3415 specified debug info. Drivers can append the data
3416 without any limit and this data is stored in memory,
3417 so this may cause significant memory stress. Disabling
3418 device dump can help save memory but the driver debug
3419 data will be no longer available. This parameter
3420 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3423 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3424 caches in the slab allocator. Saves per-node memory,
3425 but will impact performance.
3429 noaltinstr [S390] Disables alternative instructions patching
3430 (CPU alternatives feature).
3432 noapic [SMP,APIC] Tells the kernel to not make use of any
3433 IOAPICs that may be present in the system.
3435 noautogroup Disable scheduler automatic task group creation.
3437 nobats [PPC] Do not use BATs for mapping kernel lowmem
3438 on "Classic" PPC cores.
3442 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3444 delayacct [KNL] Enable per-task delay accounting
3446 nodsp [SH] Disable hardware DSP at boot time.
3448 noefi Disable EFI runtime services support.
3450 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3455 On X86-32 available only on PAE configured kernels.
3456 noexec=on: enable non-executable mappings (default)
3457 noexec=off: disable non-executable mappings
3460 Disable SMAP (Supervisor Mode Access Prevention)
3461 even if it is supported by processor.
3464 Disable SMEP (Supervisor Mode Execution Prevention)
3465 even if it is supported by processor.
3468 This affects only 32-bit executables.
3469 noexec32=on: enable non-executable mappings (default)
3470 read doesn't imply executable mappings
3471 noexec32=off: disable non-executable mappings
3472 read implies executable mappings
3474 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3476 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3477 register save and restore. The kernel will only save
3478 legacy floating-point registers on task switch.
3480 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3482 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3484 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3485 Equivalent to smt=1.
3487 [KNL,X86] Disable symmetric multithreading (SMT).
3488 nosmt=force: Force disable SMT, cannot be undone
3489 via the sysfs control file.
3491 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3492 (bounds check bypass). With this option data leaks are
3493 possible in the system.
3495 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3496 the Spectre variant 2 (indirect branch prediction)
3497 vulnerability. System may allow data leaks with this
3500 nospec_store_bypass_disable
3501 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3504 [PPC] Don't flush the L1-D cache after accessing user data.
3506 noxsave [BUGS=X86] Disables x86 extended register state save
3507 and restore using xsave. The kernel will fallback to
3508 enabling legacy floating-point and sse state.
3510 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3511 register states. The kernel will fall back to use
3512 xsave to save the states. By using this parameter,
3513 performance of saving the states is degraded because
3514 xsave doesn't support modified optimization while
3515 xsaveopt supports it on xsaveopt enabled systems.
3517 noxsaves [X86] Disables xsaves and xrstors used in saving and
3518 restoring x86 extended register state in compacted
3519 form of xsave area. The kernel will fall back to use
3520 xsaveopt and xrstor to save and restore the states
3521 in standard form of xsave area. By using this
3522 parameter, xsave area per process might occupy more
3523 memory on xsaves enabled systems.
3525 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3526 in do_idle() and not use the arch_cpu_idle()
3527 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3528 to be effective. This is useful on platforms where the
3529 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3530 correctly or when doing power measurements to evalute
3531 the impact of the sleep instructions. This is also
3532 useful when using JTAG debugger.
3534 no_file_caps Tells the kernel not to honor file capabilities. The
3535 only way then for a file to be executed with privilege
3536 is to be setuid root or executed by root.
3538 nohalt [IA-64] Tells the kernel not to use the power saving
3539 function PAL_HALT_LIGHT when idle. This increases
3540 power-consumption. On the positive side, it reduces
3541 interrupt wake-up latency, which may improve performance
3542 in certain environments such as networked servers or
3546 Force pointers printed to the console or buffers to be
3547 unhashed. By default, when a pointer is printed via %p
3548 format string, that pointer is "hashed", i.e. obscured
3549 by hashing the pointer value. This is a security feature
3550 that hides actual kernel addresses from unprivileged
3551 users, but it also makes debugging the kernel more
3552 difficult since unequal pointers can no longer be
3553 compared. However, if this command-line option is
3554 specified, then all normal pointers will have their true
3555 value printed. This option should only be specified when
3556 debugging the kernel. Please do not use on production
3559 nohibernate [HIBERNATION] Disable hibernation and resume.
3561 nohz= [KNL] Boottime enable/disable dynamic ticks
3562 Valid arguments: on, off
3565 nohz_full= [KNL,BOOT,SMP,ISOL]
3566 The argument is a cpu list, as described above.
3567 In kernels built with CONFIG_NO_HZ_FULL=y, set
3568 the specified list of CPUs whose tick will be stopped
3569 whenever possible. The boot CPU will be forced outside
3570 the range to maintain the timekeeping. Any CPUs
3571 in this list will have their RCU callbacks offloaded,
3572 just as if they had also been called out in the
3573 rcu_nocbs= boot parameter.
3575 noiotrap [SH] Disables trapped I/O port accesses.
3577 noirqdebug [X86-32] Disables the code which attempts to detect and
3578 disable unhandled interrupt sources.
3580 no_timer_check [X86,APIC] Disables the code which tests for
3581 broken timer IRQ sources.
3583 noisapnp [ISAPNP] Disables ISA PnP code.
3585 noinitrd [RAM] Tells the kernel not to load any configured
3588 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3590 [Deprecated - use intremap=off]
3594 noinvpcid [X86] Disable the INVPCID cpu feature.
3596 nojitter [IA-64] Disables jitter checking for ITC timers.
3598 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3600 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3604 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3605 clock and use the default one.
3607 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3608 accounting. steal time is computed, but won't
3609 influence scheduler behaviour
3611 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3613 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3615 noltlbs [PPC] Do not use large page/tlb entries for kernel
3616 lowmem mapping on PPC40x and PPC8xx
3618 nomca [IA-64] Disable machine check abort handling
3620 nomce [X86-32] Disable Machine Check Exception
3622 nomfgpt [X86-32] Disable Multi-Function General Purpose
3623 Timer usage (for AMD Geode machines).
3625 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3626 shutdown the other cpus. Instead use the REBOOT_VECTOR
3629 nomodeset Disable kernel modesetting. DRM drivers will not perform
3630 display-mode changes or accelerated rendering. Only the
3631 system framebuffer will be available for use if this was
3632 set-up by the firmware or boot loader.
3634 Useful as fallback, or for testing and debugging.
3636 nomodule Disable module load
3638 nopat [X86] Disable PAT (page attribute table extension of
3639 pagetables) support.
3641 nopcid [X86-64] Disable the PCID cpu feature.
3643 norandmaps Don't use address space randomization. Equivalent to
3644 echo 0 > /proc/sys/kernel/randomize_va_space
3646 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3647 with UP alternatives
3649 nordrand [X86] Disable kernel use of the RDRAND and
3650 RDSEED instructions even if they are supported
3651 by the processor. RDRAND and RDSEED are still
3652 available to user space applications.
3654 noresume [SWSUSP] Disables resume and restores original swap
3657 no-scroll [VGA] Disables scrollback.
3658 This is required for the Braillex ib80-piezo Braille
3659 reader made by F.H. Papenmeier (Germany).
3663 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3665 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3667 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3668 and disable the IO APIC. legacy for "maxcpus=0".
3670 nosoftlockup [KNL] Disable the soft-lockup detector.
3672 nosync [HW,M68K] Disables sync negotiation for all devices.
3674 nowatchdog [KNL] Disable both lockup detectors, i.e.
3675 soft-lockup and NMI watchdog (hard-lockup).
3679 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3681 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3682 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3683 Some features depend on CPU0. Known dependencies are:
3684 1. Resume from suspend/hibernate depends on CPU0.
3685 Suspend/hibernate will fail if CPU0 is offline and you
3686 need to online CPU0 before suspend/hibernate.
3687 2. PIC interrupts also depend on CPU0. CPU0 can't be
3688 removed if a PIC interrupt is detected.
3689 It's said poweroff/reboot may depend on CPU0 on some
3690 machines although I haven't seen such issues so far
3691 after CPU0 is offline on a few tested machines.
3692 If the dependencies are under your control, you can
3693 turn on cpu0_hotplug.
3695 nps_mtm_hs_ctr= [KNL,ARC]
3696 This parameter sets the maximum duration, in
3697 cycles, each HW thread of the CTOP can run
3698 without interruptions, before HW switches it.
3699 The actual maximum duration is 16 times this
3701 Format: integer between 1 and 255
3704 nptcg= [IA-64] Override max number of concurrent global TLB
3705 purges which is reported from either PAL_VM_SUMMARY or
3708 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3709 could support. nr_cpus=n : n >= 1 limits the kernel to
3710 support 'n' processors. It could be larger than the
3711 number of already plugged CPU during bootup, later in
3712 runtime you can physically add extra cpu until it reaches
3713 n. So during boot up some boot time memory for per-cpu
3714 variables need be pre-allocated for later physical cpu
3717 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3719 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3720 set up a single NUMA node spanning all memory.
3722 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3724 Allowed values are enable and disable
3726 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3727 'node', 'default' can be specified
3728 This can be set from sysctl after boot.
3729 See Documentation/admin-guide/sysctl/vm.rst for details.
3731 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3732 See Documentation/core-api/debugging-via-ohci1394.rst for more
3735 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3736 Rather than timing out after 20 ms if an EC
3737 command is not properly ACKed, override the length
3738 of the timeout. We have interrupts disabled while
3739 waiting for the ACK, so if this is set too high
3740 interrupts *may* be lost!
3742 omap_mux= [OMAP] Override bootloader pin multiplexing.
3743 Format: <mux_mode0.mode_name=value>...
3744 For example, to override I2C bus2:
3745 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3747 oops=panic Always panic on oopses. Default is to just kill the
3748 process, but there is a small probability of
3749 deadlocking the machine.
3750 This will also cause panics on machine check exceptions.
3751 Useful together with panic=30 to trigger a reboot.
3754 [KNL] Boolean flag to control whether the page allocator
3755 should randomize its free lists. The randomization may
3756 be automatically enabled if the kernel detects it is
3757 running on a platform with a direct-mapped memory-side
3758 cache, and this parameter can be used to
3759 override/disable that behavior. The state of the flag
3760 can be read from sysfs at:
3761 /sys/module/page_alloc/parameters/shuffle.
3763 page_owner= [KNL] Boot-time page_owner enabling option.
3764 Storage of the information about who allocated
3765 each page is disabled in default. With this switch,
3767 on: enable the feature
3769 page_poison= [KNL] Boot-time parameter changing the state of
3770 poisoning on the buddy allocator, available with
3771 CONFIG_PAGE_POISONING=y.
3772 off: turn off poisoning (default)
3773 on: turn on poisoning
3775 page_reporting.page_reporting_order=
3776 [KNL] Minimal page reporting order
3778 Adjust the minimal page reporting order. The page
3779 reporting is disabled when it exceeds (MAX_ORDER-1).
3781 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3782 timeout > 0: seconds before rebooting
3783 timeout = 0: wait forever
3784 timeout < 0: reboot immediately
3787 panic_print= Bitmask for printing system info when panic happens.
3788 User can chose combination of the following bits:
3789 bit 0: print all tasks info
3790 bit 1: print system memory info
3791 bit 2: print timer info
3792 bit 3: print locks info if CONFIG_LOCKDEP is on
3793 bit 4: print ftrace buffer
3794 bit 5: print all printk messages in buffer
3795 bit 6: print all CPUs backtrace (if available in the arch)
3796 *Be aware* that this option may print a _lot_ of lines,
3797 so there are risks of losing older messages in the log.
3798 Use this option carefully, maybe worth to setup a
3799 bigger log buffer with "log_buf_len" along with this.
3801 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3802 Format: <hex>[,nousertaint]
3803 Hexadecimal bitmask representing the set of TAINT flags
3804 that will cause the kernel to panic when add_taint() is
3805 called with any of the flags in this set.
3806 The optional switch "nousertaint" can be utilized to
3807 prevent userspace forced crashes by writing to sysctl
3808 /proc/sys/kernel/tainted any flagset matching with the
3809 bitmask set on panic_on_taint.
3810 See Documentation/admin-guide/tainted-kernels.rst for
3811 extra details on the taint flags that users can pick
3812 to compose the bitmask to assign to panic_on_taint.
3814 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3817 crash_kexec_post_notifiers
3818 Run kdump after running panic-notifiers and dumping
3819 kmsg. This only for the users who doubt kdump always
3820 succeeds in any situation.
3821 Note that this also increases risks of kdump failure,
3822 because some panic notifiers can make the crashed
3823 kernel more unstable.
3825 parkbd.port= [HW] Parallel port number the keyboard adapter is
3826 connected to, default is 0.
3828 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3829 0 for XT, 1 for AT (default is AT).
3832 parport= [HW,PPT] Specify parallel ports. 0 disables.
3833 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3834 Use 'auto' to force the driver to use any
3835 IRQ/DMA settings detected (the default is to
3836 ignore detected IRQ/DMA settings because of
3837 possible conflicts). You can specify the base
3838 address, IRQ, and DMA settings; IRQ and DMA
3839 should be numbers, or 'auto' (for using detected
3840 settings on that particular port), or 'nofifo'
3841 (to avoid using a FIFO even if it is detected).
3842 Parallel ports are assigned in the order they
3843 are specified on the command line, starting
3846 parport_init_mode= [HW,PPT]
3847 Configure VIA parallel port to operate in
3848 a specific mode. This is necessary on Pegasos
3849 computer where firmware has no options for setting
3850 up parallel port mode and sets it to spp.
3851 Currently this function knows 686a and 8231 chips.
3852 Format: [spp|ps2|epp|ecp|ecpepp]
3854 pata_legacy.all= [HW,LIBATA]
3856 Set to non-zero to probe primary and secondary ISA
3857 port ranges on PCI systems where no PCI PATA device
3858 has been found at either range. Disabled by default.
3860 pata_legacy.autospeed= [HW,LIBATA]
3862 Set to non-zero if a chip is present that snoops speed
3863 changes. Disabled by default.
3865 pata_legacy.ht6560a= [HW,LIBATA]
3867 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3868 the secondary channel, or both channels respectively.
3869 Disabled by default.
3871 pata_legacy.ht6560b= [HW,LIBATA]
3873 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3874 the secondary channel, or both channels respectively.
3875 Disabled by default.
3877 pata_legacy.iordy_mask= [HW,LIBATA]
3879 IORDY enable mask. Set individual bits to allow IORDY
3880 for the respective channel. Bit 0 is for the first
3881 legacy channel handled by this driver, bit 1 is for
3882 the second channel, and so on. The sequence will often
3883 correspond to the primary legacy channel, the secondary
3884 legacy channel, and so on, but the handling of a PCI
3885 bus and the use of other driver options may interfere
3886 with the sequence. By default IORDY is allowed across
3889 pata_legacy.opti82c46x= [HW,LIBATA]
3891 Set to 1, 2, or 3 for Opti 82c611A on the primary
3892 channel, the secondary channel, or both channels
3893 respectively. Disabled by default.
3895 pata_legacy.opti82c611a= [HW,LIBATA]
3897 Set to 1, 2, or 3 for Opti 82c465MV on the primary
3898 channel, the secondary channel, or both channels
3899 respectively. Disabled by default.
3901 pata_legacy.pio_mask= [HW,LIBATA]
3903 PIO mode mask for autospeed devices. Set individual
3904 bits to allow the use of the respective PIO modes.
3905 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
3906 All modes allowed by default.
3908 pata_legacy.probe_all= [HW,LIBATA]
3910 Set to non-zero to probe tertiary and further ISA
3911 port ranges on PCI systems. Disabled by default.
3913 pata_legacy.probe_mask= [HW,LIBATA]
3915 Probe mask for legacy ISA PATA ports. Depending on
3916 platform configuration and the use of other driver
3917 options up to 6 legacy ports are supported: 0x1f0,
3918 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
3919 of individual ports can be disabled by setting the
3920 corresponding bits in the mask to 1. Bit 0 is for
3921 the first port in the list above (0x1f0), and so on.
3922 By default all supported ports are probed.
3924 pata_legacy.qdi= [HW,LIBATA]
3926 Set to non-zero to probe QDI controllers. By default
3927 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
3929 pata_legacy.winbond= [HW,LIBATA]
3931 Set to non-zero to probe Winbond controllers. Use
3932 the standard I/O port (0x130) if 1, otherwise the
3933 value given is the I/O port to use (typically 0x1b0).
3934 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
3937 pata_platform.pio_mask= [HW,LIBATA]
3939 Supported PIO mode mask. Set individual bits to allow
3940 the use of the respective PIO modes. Bit 0 is for
3941 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
3945 Halt all CPUs after the first oops has been printed for
3946 the specified number of seconds. This is to be used if
3947 your oopses keep scrolling off the screen.
3952 See header of drivers/block/paride/pcd.c.
3953 See also Documentation/admin-guide/blockdev/paride.rst.
3955 pci=option[,option...] [PCI] various PCI subsystem options.
3957 Some options herein operate on a specific device
3958 or a set of devices (<pci_dev>). These are
3959 specified in one of the following formats:
3961 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3962 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3964 Note: the first format specifies a PCI
3965 bus/device/function address which may change
3966 if new hardware is inserted, if motherboard
3967 firmware changes, or due to changes caused
3968 by other kernel parameters. If the
3969 domain is left unspecified, it is
3970 taken to be zero. Optionally, a path
3971 to a device through multiple device/function
3972 addresses can be specified after the base
3973 address (this is more robust against
3974 renumbering issues). The second format
3975 selects devices using IDs from the
3976 configuration space which may match multiple
3977 devices in the system.
3979 earlydump dump PCI config space before the kernel
3981 off [X86] don't probe for the PCI bus
3982 bios [X86-32] force use of PCI BIOS, don't access
3983 the hardware directly. Use this if your machine
3984 has a non-standard PCI host bridge.
3985 nobios [X86-32] disallow use of PCI BIOS, only direct
3986 hardware access methods are allowed. Use this
3987 if you experience crashes upon bootup and you
3988 suspect they are caused by the BIOS.
3989 conf1 [X86] Force use of PCI Configuration Access
3990 Mechanism 1 (config address in IO port 0xCF8,
3991 data in IO port 0xCFC, both 32-bit).
3992 conf2 [X86] Force use of PCI Configuration Access
3993 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3994 the function, IO port 0xCFA, also 8-bit, sets
3995 bus number. The config space is then accessed
3996 through ports 0xC000-0xCFFF).
3997 See http://wiki.osdev.org/PCI for more info
3998 on the configuration access mechanisms.
3999 noaer [PCIE] If the PCIEAER kernel config parameter is
4000 enabled, this kernel boot option can be used to
4001 disable the use of PCIE advanced error reporting.
4002 nodomains [PCI] Disable support for multiple PCI
4003 root domains (aka PCI segments, in ACPI-speak).
4004 nommconf [X86] Disable use of MMCONFIG for PCI
4006 check_enable_amd_mmconf [X86] check for and enable
4007 properly configured MMIO access to PCI
4008 config space on AMD family 10h CPU
4009 nomsi [MSI] If the PCI_MSI kernel config parameter is
4010 enabled, this kernel boot option can be used to
4011 disable the use of MSI interrupts system-wide.
4012 noioapicquirk [APIC] Disable all boot interrupt quirks.
4013 Safety option to keep boot IRQs enabled. This
4014 should never be necessary.
4015 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4016 primary IO-APIC for bridges that cannot disable
4017 boot IRQs. This fixes a source of spurious IRQs
4018 when the system masks IRQs.
4019 noioapicreroute [APIC] Disable workaround that uses the
4020 boot IRQ equivalent of an IRQ that connects to
4021 a chipset where boot IRQs cannot be disabled.
4022 The opposite of ioapicreroute.
4023 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4024 routing table. These calls are known to be buggy
4025 on several machines and they hang the machine
4026 when used, but on other computers it's the only
4027 way to get the interrupt routing table. Try
4028 this option if the kernel is unable to allocate
4029 IRQs or discover secondary PCI buses on your
4031 rom [X86] Assign address space to expansion ROMs.
4032 Use with caution as certain devices share
4033 address decoders between ROMs and other
4035 norom [X86] Do not assign address space to
4036 expansion ROMs that do not already have
4037 BIOS assigned address ranges.
4038 nobar [X86] Do not assign address space to the
4039 BARs that weren't assigned by the BIOS.
4040 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4041 assigned automatically to PCI devices. You can
4042 make the kernel exclude IRQs of your ISA cards
4044 pirqaddr=0xAAAAA [X86] Specify the physical address
4045 of the PIRQ table (normally generated
4046 by the BIOS) if it is outside the
4047 F0000h-100000h range.
4048 lastbus=N [X86] Scan all buses thru bus #N. Can be
4049 useful if the kernel is unable to find your
4050 secondary buses and you want to tell it
4051 explicitly which ones they are.
4052 assign-busses [X86] Always assign all PCI bus
4053 numbers ourselves, overriding
4054 whatever the firmware may have done.
4055 usepirqmask [X86] Honor the possible IRQ mask stored
4056 in the BIOS $PIR table. This is needed on
4057 some systems with broken BIOSes, notably
4058 some HP Pavilion N5400 and Omnibook XE3
4059 notebooks. This will have no effect if ACPI
4060 IRQ routing is enabled.
4061 noacpi [X86] Do not use ACPI for IRQ routing
4062 or for PCI scanning.
4063 use_crs [X86] Use PCI host bridge window information
4064 from ACPI. On BIOSes from 2008 or later, this
4065 is enabled by default. If you need to use this,
4066 please report a bug.
4067 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4068 If you need to use this, please report a bug.
4069 routeirq Do IRQ routing for all PCI devices.
4070 This is normally done in pci_enable_device(),
4071 so this option is a temporary workaround
4072 for broken drivers that don't call it.
4073 skip_isa_align [X86] do not align io start addr, so can
4074 handle more pci cards
4075 noearly [X86] Don't do any early type 1 scanning.
4076 This might help on some broken boards which
4077 machine check when some devices' config space
4078 is read. But various workarounds are disabled
4079 and some IOMMU drivers will not work.
4080 bfsort Sort PCI devices into breadth-first order.
4081 This sorting is done to get a device
4082 order compatible with older (<= 2.4) kernels.
4083 nobfsort Don't sort PCI devices into breadth-first order.
4084 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4085 tuning and use the BIOS-configured MPS defaults.
4086 pcie_bus_safe Set every device's MPS to the largest value
4087 supported by all devices below the root complex.
4088 pcie_bus_perf Set device MPS to the largest allowable MPS
4089 based on its parent bus. Also set MRRS (Max
4090 Read Request Size) to the largest supported
4091 value (no larger than the MPS that the device
4092 or bus can support) for best performance.
4093 pcie_bus_peer2peer Set every device's MPS to 128B, which
4094 every device is guaranteed to support. This
4095 configuration allows peer-to-peer DMA between
4096 any pair of devices, possibly at the cost of
4097 reduced performance. This also guarantees
4098 that hot-added devices will work.
4099 cbiosize=nn[KMG] The fixed amount of bus space which is
4100 reserved for the CardBus bridge's IO window.
4101 The default value is 256 bytes.
4102 cbmemsize=nn[KMG] The fixed amount of bus space which is
4103 reserved for the CardBus bridge's memory
4104 window. The default value is 64 megabytes.
4107 [<order of align>@]<pci_dev>[; ...]
4108 Specifies alignment and device to reassign
4109 aligned memory resources. How to
4110 specify the device is described above.
4111 If <order of align> is not specified,
4112 PAGE_SIZE is used as alignment.
4113 A PCI-PCI bridge can be specified if resource
4114 windows need to be expanded.
4115 To specify the alignment for several
4116 instances of a device, the PCI vendor,
4117 device, subvendor, and subdevice may be
4118 specified, e.g., 12@pci:8086:9c22:103c:198f
4119 for 4096-byte alignment.
4120 ecrc= Enable/disable PCIe ECRC (transaction layer
4121 end-to-end CRC checking).
4122 bios: Use BIOS/firmware settings. This is the
4126 hpiosize=nn[KMG] The fixed amount of bus space which is
4127 reserved for hotplug bridge's IO window.
4128 Default size is 256 bytes.
4129 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4130 reserved for hotplug bridge's MMIO window.
4131 Default size is 2 megabytes.
4132 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4133 reserved for hotplug bridge's MMIO_PREF window.
4134 Default size is 2 megabytes.
4135 hpmemsize=nn[KMG] The fixed amount of bus space which is
4136 reserved for hotplug bridge's MMIO and
4138 Default size is 2 megabytes.
4139 hpbussize=nn The minimum amount of additional bus numbers
4140 reserved for buses below a hotplug bridge.
4142 realloc= Enable/disable reallocating PCI bridge resources
4143 if allocations done by BIOS are too small to
4144 accommodate resources required by all child
4146 off: Turn realloc off
4148 realloc same as realloc=on
4149 noari do not use PCIe ARI.
4150 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4151 do not use PCIe ATS (and IOMMU device IOTLB).
4152 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4153 only look for one device below a PCIe downstream
4155 big_root_window Try to add a big 64bit memory window to the PCIe
4156 root complex on AMD CPUs. Some GFX hardware
4157 can resize a BAR to allow access to all VRAM.
4158 Adding the window is slightly risky (it may
4159 conflict with unreported devices), so this
4161 disable_acs_redir=<pci_dev>[; ...]
4162 Specify one or more PCI devices (in the format
4163 specified above) separated by semicolons.
4164 Each device specified will have the PCI ACS
4165 redirect capabilities forced off which will
4166 allow P2P traffic between devices through
4167 bridges without forcing it upstream. Note:
4168 this removes isolation between devices and
4169 may put more devices in an IOMMU group.
4170 force_floating [S390] Force usage of floating interrupts.
4171 nomio [S390] Do not use MIO instructions.
4172 norid [S390] ignore the RID field and force use of
4173 one PCI domain per PCI function
4175 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4178 force Enable ASPM even on devices that claim not to support it.
4179 WARNING: Forcing ASPM on may cause system lockups.
4181 pcie_ports= [PCIE] PCIe port services handling:
4182 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4183 even if the platform doesn't give the OS permission to
4184 use them. This may cause conflicts if the platform
4185 also tries to use these services.
4186 dpc-native Use native PCIe service for DPC only. May
4187 cause conflicts if firmware uses AER or DPC.
4188 compat Disable native PCIe services (PME, AER, DPC, PCIe
4191 pcie_port_pm= [PCIE] PCIe port power management handling:
4192 off Disable power management of all PCIe ports
4193 force Forcibly enable power management of all PCIe ports
4195 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4196 nomsi Do not use MSI for native PCIe PME signaling (this makes
4197 all PCIe root ports use INTx for all services).
4199 pcmv= [HW,PCMCIA] BadgePAD 4
4203 Keep all power-domains already enabled by bootloader on,
4204 even if no driver has claimed them. This is useful
4205 for debug and development, but should not be
4206 needed on a platform with proper driver support.
4209 See Documentation/admin-guide/blockdev/paride.rst.
4211 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4214 See arch/parisc/kernel/pdc_chassis.c
4216 percpu_alloc= Select which percpu first chunk allocator to use.
4217 Currently supported values are "embed" and "page".
4218 Archs may support subset or none of the selections.
4219 See comments in mm/percpu.c for details on each
4220 allocator. This parameter is primarily for debugging
4221 and performance comparison.
4224 See Documentation/admin-guide/blockdev/paride.rst.
4227 See Documentation/admin-guide/blockdev/paride.rst.
4229 pirq= [SMP,APIC] Manual mp-table setup
4230 See Documentation/x86/i386/IO-APIC.rst.
4232 plip= [PPT,NET] Parallel port network link
4233 Format: { parport<nr> | timid | 0 }
4234 See also Documentation/admin-guide/parport.rst.
4236 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4237 Override pmtimer IOPort with a hex value.
4240 pmu_override= [PPC] Override the PMU.
4241 This option takes over the PMU facility, so it is no
4242 longer usable by perf. Setting this option starts the
4243 PMU counters by setting MMCR0 to 0 (the FC bit is
4244 cleared). If a number is given, then MMCR1 is set to
4245 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4248 pm_debug_messages [SUSPEND,KNL]
4249 Enable suspend/resume debug messages during boot up.
4252 Enable PNP debug messages (depends on the
4253 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4254 via /sys/module/pnp/parameters/debug. We always show
4255 current resource usage; turning this on also shows
4256 possible settings and some assignment information.
4262 { on | off | curr | res | no-curr | no-res }
4265 [ISAPNP] Exclude IRQs for the autoconfiguration
4268 [ISAPNP] Exclude DMAs for the autoconfiguration
4270 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4271 Ranges are in pairs (I/O port base and size).
4274 [ISAPNP] Exclude memory regions for the
4276 Ranges are in pairs (memory base and size).
4278 ports= [IP_VS_FTP] IPVS ftp helper module
4280 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4282 Format: <port>,<port>....
4284 powersave=off [PPC] This option disables power saving features.
4285 It specifically disables cpuidle and sets the
4286 platform machine description specific power_save
4287 function to NULL. On Idle the CPU just reduces
4290 ppc_strict_facility_enable
4291 [PPC] This option catches any kernel floating point,
4292 Altivec, VSX and SPE outside of regions specifically
4293 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4294 There is some performance impact when enabling this.
4298 Disable Hardware Transactional Memory
4301 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4302 none - Limited to cond_resched() calls
4303 voluntary - Limited to cond_resched() and might_sleep() calls
4304 full - Any section that isn't explicitly preempt disabled
4305 can be preempted anytime.
4307 print-fatal-signals=
4308 [KNL] debug: print fatal signals
4310 If enabled, warn about various signal handling
4311 related application anomalies: too many signals,
4312 too many POSIX.1 timers, fatal signals causing a
4315 If you hit the warning due to signal overflow,
4316 you might want to try "ulimit -i unlimited".
4320 printk.always_kmsg_dump=
4321 Trigger kmsg_dump for cases other than kernel oops or
4323 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4326 printk.console_no_auto_verbose=
4327 Disable console loglevel raise on oops, panic
4328 or lockdep-detected issues (only if lock debug is on).
4329 With an exception to setups with low baudrate on
4330 serial console, keeping this 0 is a good choice
4331 in order to provide more debug information.
4333 default: 0 (auto_verbose is enabled)
4335 printk.devkmsg={on,off,ratelimit}
4336 Control writing to /dev/kmsg.
4337 on - unlimited logging to /dev/kmsg from userspace
4338 off - logging to /dev/kmsg disabled
4339 ratelimit - ratelimit the logging
4342 printk.time= Show timing data prefixed to each printk message line
4343 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4345 processor.max_cstate= [HW,ACPI]
4346 Limit processor to maximum C-state
4347 max_cstate=9 overrides any DMI blacklist limit.
4349 processor.nocst [HW,ACPI]
4350 Ignore the _CST method to determine C-states,
4351 instead using the legacy FADT method
4353 profile= [KNL] Enable kernel profiling via /proc/profile
4354 Format: [<profiletype>,]<number>
4355 Param: <profiletype>: "schedule", "sleep", or "kvm"
4356 [defaults to kernel profiling]
4357 Param: "schedule" - profile schedule points.
4358 Param: "sleep" - profile D-state sleeping (millisecs).
4359 Requires CONFIG_SCHEDSTATS
4360 Param: "kvm" - profile VM exits.
4361 Param: <number> - step/bucket size as a power of 2 for
4362 statistical time based profiling.
4364 prompt_ramdisk= [RAM] [Deprecated]
4366 prot_virt= [S390] enable hosting protected virtual machines
4367 isolated from the hypervisor (if hardware supports
4371 psi= [KNL] Enable or disable pressure stall information
4375 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4376 probe for; one of (bare|imps|exps|lifebook|any).
4377 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4379 psmouse.resetafter= [HW,MOUSE]
4380 Try to reset the device after so many bad packets
4383 [HW,MOUSE] Set desired mouse resolution, in dpi.
4384 psmouse.smartscroll=
4385 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4386 0 = disabled, 1 = enabled (default).
4388 pstore.backend= Specify the name of the pstore backend to use
4391 See Documentation/admin-guide/blockdev/paride.rst.
4393 pti= [X86-64] Control Page Table Isolation of user and
4394 kernel address spaces. Disabling this feature
4395 removes hardening, but improves performance of
4396 system calls and interrupts.
4398 on - unconditionally enable
4399 off - unconditionally disable
4400 auto - kernel detects whether your CPU model is
4401 vulnerable to issues that PTI mitigates
4403 Not specifying this option is equivalent to pti=auto.
4406 Equivalent to pti=off
4409 [KNL] Number of legacy pty's. Overwrites compiled-in
4412 quiet [KNL] Disable most log messages
4417 See Documentation/admin-guide/md.rst.
4419 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4420 See Documentation/admin-guide/blockdev/ramdisk.rst.
4422 ramdisk_start= [RAM] RAM disk image start address
4424 random.trust_cpu={on,off}
4425 [KNL] Enable or disable trusting the use of the
4426 CPU's random number generator (if available) to
4427 fully seed the kernel's CRNG. Default is controlled
4428 by CONFIG_RANDOM_TRUST_CPU.
4430 random.trust_bootloader={on,off}
4431 [KNL] Enable or disable trusting the use of a
4432 seed passed by the bootloader (if available) to
4433 fully seed the kernel's CRNG. Default is controlled
4434 by CONFIG_RANDOM_TRUST_BOOTLOADER.
4436 randomize_kstack_offset=
4437 [KNL] Enable or disable kernel stack offset
4438 randomization, which provides roughly 5 bits of
4439 entropy, frustrating memory corruption attacks
4440 that depend on stack address determinism or
4441 cross-syscall address exposures. This is only
4442 available on architectures that have defined
4443 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4444 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4445 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4447 ras=option[,option,...] [KNL] RAS-specific options
4450 Disable the Correctable Errors Collector,
4451 see CONFIG_RAS_CEC help text.
4453 rcu_nocbs[=cpu-list]
4454 [KNL] The optional argument is a cpu list,
4457 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4458 enable the no-callback CPU mode, which prevents
4459 such CPUs' callbacks from being invoked in
4460 softirq context. Invocation of such CPUs' RCU
4461 callbacks will instead be offloaded to "rcuox/N"
4462 kthreads created for that purpose, where "x" is
4463 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4464 for the kthreads that mediate grace periods; and
4465 "N" is the CPU number. This reduces OS jitter on
4466 the offloaded CPUs, which can be useful for HPC
4467 and real-time workloads. It can also improve
4468 energy efficiency for asymmetric multiprocessors.
4470 If a cpulist is passed as an argument, the specified
4471 list of CPUs is set to no-callback mode from boot.
4473 Otherwise, if the '=' sign and the cpulist
4474 arguments are omitted, no CPU will be set to
4475 no-callback mode from boot but the mode may be
4476 toggled at runtime via cpusets.
4479 Rather than requiring that offloaded CPUs
4480 (specified by rcu_nocbs= above) explicitly
4481 awaken the corresponding "rcuoN" kthreads,
4482 make these kthreads poll for callbacks.
4483 This improves the real-time response for the
4484 offloaded CPUs by relieving them of the need to
4485 wake up the corresponding kthread, but degrades
4486 energy efficiency by requiring that the kthreads
4487 periodically wake up to do the polling.
4489 rcutree.blimit= [KNL]
4490 Set maximum number of finished RCU callbacks to
4491 process in one batch.
4493 rcutree.dump_tree= [KNL]
4494 Dump the structure of the rcu_node combining tree
4495 out at early boot. This is used for diagnostic
4496 purposes, to verify correct tree setup.
4498 rcutree.gp_cleanup_delay= [KNL]
4499 Set the number of jiffies to delay each step of
4500 RCU grace-period cleanup.
4502 rcutree.gp_init_delay= [KNL]
4503 Set the number of jiffies to delay each step of
4504 RCU grace-period initialization.
4506 rcutree.gp_preinit_delay= [KNL]
4507 Set the number of jiffies to delay each step of
4508 RCU grace-period pre-initialization, that is,
4509 the propagation of recent CPU-hotplug changes up
4510 the rcu_node combining tree.
4512 rcutree.use_softirq= [KNL]
4513 If set to zero, move all RCU_SOFTIRQ processing to
4514 per-CPU rcuc kthreads. Defaults to a non-zero
4515 value, meaning that RCU_SOFTIRQ is used by default.
4516 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4518 But note that CONFIG_PREEMPT_RT=y kernels disable
4519 this kernel boot parameter, forcibly setting it
4522 rcutree.rcu_fanout_exact= [KNL]
4523 Disable autobalancing of the rcu_node combining
4524 tree. This is used by rcutorture, and might
4525 possibly be useful for architectures having high
4526 cache-to-cache transfer latencies.
4528 rcutree.rcu_fanout_leaf= [KNL]
4529 Change the number of CPUs assigned to each
4530 leaf rcu_node structure. Useful for very
4531 large systems, which will choose the value 64,
4532 and for NUMA systems with large remote-access
4533 latencies, which will choose a value aligned
4534 with the appropriate hardware boundaries.
4536 rcutree.rcu_min_cached_objs= [KNL]
4537 Minimum number of objects which are cached and
4538 maintained per one CPU. Object size is equal
4539 to PAGE_SIZE. The cache allows to reduce the
4540 pressure to page allocator, also it makes the
4541 whole algorithm to behave better in low memory
4544 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4545 Set the page-cache refill delay (in milliseconds)
4546 in response to low-memory conditions. The range
4547 of permitted values is in the range 0:100000.
4549 rcutree.jiffies_till_first_fqs= [KNL]
4550 Set delay from grace-period initialization to
4551 first attempt to force quiescent states.
4552 Units are jiffies, minimum value is zero,
4553 and maximum value is HZ.
4555 rcutree.jiffies_till_next_fqs= [KNL]
4556 Set delay between subsequent attempts to force
4557 quiescent states. Units are jiffies, minimum
4558 value is one, and maximum value is HZ.
4560 rcutree.jiffies_till_sched_qs= [KNL]
4561 Set required age in jiffies for a
4562 given grace period before RCU starts
4563 soliciting quiescent-state help from
4564 rcu_note_context_switch() and cond_resched().
4565 If not specified, the kernel will calculate
4566 a value based on the most recent settings
4567 of rcutree.jiffies_till_first_fqs
4568 and rcutree.jiffies_till_next_fqs.
4569 This calculated value may be viewed in
4570 rcutree.jiffies_to_sched_qs. Any attempt to set
4571 rcutree.jiffies_to_sched_qs will be cheerfully
4574 rcutree.kthread_prio= [KNL,BOOT]
4575 Set the SCHED_FIFO priority of the RCU per-CPU
4576 kthreads (rcuc/N). This value is also used for
4577 the priority of the RCU boost threads (rcub/N)
4578 and for the RCU grace-period kthreads (rcu_bh,
4579 rcu_preempt, and rcu_sched). If RCU_BOOST is
4580 set, valid values are 1-99 and the default is 1
4581 (the least-favored priority). Otherwise, when
4582 RCU_BOOST is not set, valid values are 0-99 and
4583 the default is zero (non-realtime operation).
4584 When RCU_NOCB_CPU is set, also adjust the
4585 priority of NOCB callback kthreads.
4587 rcutree.rcu_nocb_gp_stride= [KNL]
4588 Set the number of NOCB callback kthreads in
4589 each group, which defaults to the square root
4590 of the number of CPUs. Larger numbers reduce
4591 the wakeup overhead on the global grace-period
4592 kthread, but increases that same overhead on
4593 each group's NOCB grace-period kthread.
4595 rcutree.qhimark= [KNL]
4596 Set threshold of queued RCU callbacks beyond which
4597 batch limiting is disabled.
4599 rcutree.qlowmark= [KNL]
4600 Set threshold of queued RCU callbacks below which
4601 batch limiting is re-enabled.
4603 rcutree.qovld= [KNL]
4604 Set threshold of queued RCU callbacks beyond which
4605 RCU's force-quiescent-state scan will aggressively
4606 enlist help from cond_resched() and sched IPIs to
4607 help CPUs more quickly reach quiescent states.
4608 Set to less than zero to make this be set based
4609 on rcutree.qhimark at boot time and to zero to
4610 disable more aggressive help enlistment.
4612 rcutree.rcu_kick_kthreads= [KNL]
4613 Cause the grace-period kthread to get an extra
4614 wake_up() if it sleeps three times longer than
4615 it should at force-quiescent-state time.
4616 This wake_up() will be accompanied by a
4617 WARN_ONCE() splat and an ftrace_dump().
4619 rcutree.rcu_unlock_delay= [KNL]
4620 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4621 this specifies an rcu_read_unlock()-time delay
4622 in microseconds. This defaults to zero.
4623 Larger delays increase the probability of
4624 catching RCU pointer leaks, that is, buggy use
4625 of RCU-protected pointers after the relevant
4626 rcu_read_unlock() has completed.
4628 rcutree.sysrq_rcu= [KNL]
4629 Commandeer a sysrq key to dump out Tree RCU's
4630 rcu_node tree with an eye towards determining
4631 why a new grace period has not yet started.
4633 rcuscale.gp_async= [KNL]
4634 Measure performance of asynchronous
4635 grace-period primitives such as call_rcu().
4637 rcuscale.gp_async_max= [KNL]
4638 Specify the maximum number of outstanding
4639 callbacks per writer thread. When a writer
4640 thread exceeds this limit, it invokes the
4641 corresponding flavor of rcu_barrier() to allow
4642 previously posted callbacks to drain.
4644 rcuscale.gp_exp= [KNL]
4645 Measure performance of expedited synchronous
4646 grace-period primitives.
4648 rcuscale.holdoff= [KNL]
4649 Set test-start holdoff period. The purpose of
4650 this parameter is to delay the start of the
4651 test until boot completes in order to avoid
4654 rcuscale.kfree_rcu_test= [KNL]
4655 Set to measure performance of kfree_rcu() flooding.
4657 rcuscale.kfree_rcu_test_double= [KNL]
4658 Test the double-argument variant of kfree_rcu().
4659 If this parameter has the same value as
4660 rcuscale.kfree_rcu_test_single, both the single-
4661 and double-argument variants are tested.
4663 rcuscale.kfree_rcu_test_single= [KNL]
4664 Test the single-argument variant of kfree_rcu().
4665 If this parameter has the same value as
4666 rcuscale.kfree_rcu_test_double, both the single-
4667 and double-argument variants are tested.
4669 rcuscale.kfree_nthreads= [KNL]
4670 The number of threads running loops of kfree_rcu().
4672 rcuscale.kfree_alloc_num= [KNL]
4673 Number of allocations and frees done in an iteration.
4675 rcuscale.kfree_loops= [KNL]
4676 Number of loops doing rcuscale.kfree_alloc_num number
4677 of allocations and frees.
4679 rcuscale.nreaders= [KNL]
4680 Set number of RCU readers. The value -1 selects
4681 N, where N is the number of CPUs. A value
4682 "n" less than -1 selects N-n+1, where N is again
4683 the number of CPUs. For example, -2 selects N
4684 (the number of CPUs), -3 selects N+1, and so on.
4685 A value of "n" less than or equal to -N selects
4688 rcuscale.nwriters= [KNL]
4689 Set number of RCU writers. The values operate
4690 the same as for rcuscale.nreaders.
4691 N, where N is the number of CPUs
4693 rcuscale.perf_type= [KNL]
4694 Specify the RCU implementation to test.
4696 rcuscale.shutdown= [KNL]
4697 Shut the system down after performance tests
4698 complete. This is useful for hands-off automated
4701 rcuscale.verbose= [KNL]
4702 Enable additional printk() statements.
4704 rcuscale.writer_holdoff= [KNL]
4705 Write-side holdoff between grace periods,
4706 in microseconds. The default of zero says
4709 rcutorture.fqs_duration= [KNL]
4710 Set duration of force_quiescent_state bursts
4713 rcutorture.fqs_holdoff= [KNL]
4714 Set holdoff time within force_quiescent_state bursts
4717 rcutorture.fqs_stutter= [KNL]
4718 Set wait time between force_quiescent_state bursts
4721 rcutorture.fwd_progress= [KNL]
4722 Specifies the number of kthreads to be used
4723 for RCU grace-period forward-progress testing
4724 for the types of RCU supporting this notion.
4725 Defaults to 1 kthread, values less than zero or
4726 greater than the number of CPUs cause the number
4729 rcutorture.fwd_progress_div= [KNL]
4730 Specify the fraction of a CPU-stall-warning
4731 period to do tight-loop forward-progress testing.
4733 rcutorture.fwd_progress_holdoff= [KNL]
4734 Number of seconds to wait between successive
4735 forward-progress tests.
4737 rcutorture.fwd_progress_need_resched= [KNL]
4738 Enclose cond_resched() calls within checks for
4739 need_resched() during tight-loop forward-progress
4742 rcutorture.gp_cond= [KNL]
4743 Use conditional/asynchronous update-side
4744 primitives, if available.
4746 rcutorture.gp_exp= [KNL]
4747 Use expedited update-side primitives, if available.
4749 rcutorture.gp_normal= [KNL]
4750 Use normal (non-expedited) asynchronous
4751 update-side primitives, if available.
4753 rcutorture.gp_sync= [KNL]
4754 Use normal (non-expedited) synchronous
4755 update-side primitives, if available. If all
4756 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4757 rcutorture.gp_normal=, and rcutorture.gp_sync=
4758 are zero, rcutorture acts as if is interpreted
4759 they are all non-zero.
4761 rcutorture.irqreader= [KNL]
4762 Run RCU readers from irq handlers, or, more
4763 accurately, from a timer handler. Not all RCU
4764 flavors take kindly to this sort of thing.
4766 rcutorture.leakpointer= [KNL]
4767 Leak an RCU-protected pointer out of the reader.
4768 This can of course result in splats, and is
4769 intended to test the ability of things like
4770 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4773 rcutorture.n_barrier_cbs= [KNL]
4774 Set callbacks/threads for rcu_barrier() testing.
4776 rcutorture.nfakewriters= [KNL]
4777 Set number of concurrent RCU writers. These just
4778 stress RCU, they don't participate in the actual
4779 test, hence the "fake".
4781 rcutorture.nocbs_nthreads= [KNL]
4782 Set number of RCU callback-offload togglers.
4783 Zero (the default) disables toggling.
4785 rcutorture.nocbs_toggle= [KNL]
4786 Set the delay in milliseconds between successive
4787 callback-offload toggling attempts.
4789 rcutorture.nreaders= [KNL]
4790 Set number of RCU readers. The value -1 selects
4791 N-1, where N is the number of CPUs. A value
4792 "n" less than -1 selects N-n-2, where N is again
4793 the number of CPUs. For example, -2 selects N
4794 (the number of CPUs), -3 selects N+1, and so on.
4796 rcutorture.object_debug= [KNL]
4797 Enable debug-object double-call_rcu() testing.
4799 rcutorture.onoff_holdoff= [KNL]
4800 Set time (s) after boot for CPU-hotplug testing.
4802 rcutorture.onoff_interval= [KNL]
4803 Set time (jiffies) between CPU-hotplug operations,
4804 or zero to disable CPU-hotplug testing.
4806 rcutorture.read_exit= [KNL]
4807 Set the number of read-then-exit kthreads used
4808 to test the interaction of RCU updaters and
4809 task-exit processing.
4811 rcutorture.read_exit_burst= [KNL]
4812 The number of times in a given read-then-exit
4813 episode that a set of read-then-exit kthreads
4816 rcutorture.read_exit_delay= [KNL]
4817 The delay, in seconds, between successive
4818 read-then-exit testing episodes.
4820 rcutorture.shuffle_interval= [KNL]
4821 Set task-shuffle interval (s). Shuffling tasks
4822 allows some CPUs to go into dyntick-idle mode
4823 during the rcutorture test.
4825 rcutorture.shutdown_secs= [KNL]
4826 Set time (s) after boot system shutdown. This
4827 is useful for hands-off automated testing.
4829 rcutorture.stall_cpu= [KNL]
4830 Duration of CPU stall (s) to test RCU CPU stall
4831 warnings, zero to disable.
4833 rcutorture.stall_cpu_block= [KNL]
4834 Sleep while stalling if set. This will result
4835 in warnings from preemptible RCU in addition
4836 to any other stall-related activity.
4838 rcutorture.stall_cpu_holdoff= [KNL]
4839 Time to wait (s) after boot before inducing stall.
4841 rcutorture.stall_cpu_irqsoff= [KNL]
4842 Disable interrupts while stalling if set.
4844 rcutorture.stall_gp_kthread= [KNL]
4845 Duration (s) of forced sleep within RCU
4846 grace-period kthread to test RCU CPU stall
4847 warnings, zero to disable. If both stall_cpu
4848 and stall_gp_kthread are specified, the
4849 kthread is starved first, then the CPU.
4851 rcutorture.stat_interval= [KNL]
4852 Time (s) between statistics printk()s.
4854 rcutorture.stutter= [KNL]
4855 Time (s) to stutter testing, for example, specifying
4856 five seconds causes the test to run for five seconds,
4857 wait for five seconds, and so on. This tests RCU's
4858 ability to transition abruptly to and from idle.
4860 rcutorture.test_boost= [KNL]
4861 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4862 "Maybe" means test if the RCU implementation
4863 under test support RCU priority boosting.
4865 rcutorture.test_boost_duration= [KNL]
4866 Duration (s) of each individual boost test.
4868 rcutorture.test_boost_interval= [KNL]
4869 Interval (s) between each boost test.
4871 rcutorture.test_no_idle_hz= [KNL]
4872 Test RCU's dyntick-idle handling. See also the
4873 rcutorture.shuffle_interval parameter.
4875 rcutorture.torture_type= [KNL]
4876 Specify the RCU implementation to test.
4878 rcutorture.verbose= [KNL]
4879 Enable additional printk() statements.
4881 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4882 Dump ftrace buffer after reporting RCU CPU
4885 rcupdate.rcu_cpu_stall_suppress= [KNL]
4886 Suppress RCU CPU stall warning messages.
4888 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4889 Suppress RCU CPU stall warning messages and
4890 rcutorture writer stall warnings that occur
4891 during early boot, that is, during the time
4892 before the init task is spawned.
4894 rcupdate.rcu_cpu_stall_timeout= [KNL]
4895 Set timeout for RCU CPU stall warning messages.
4897 rcupdate.rcu_expedited= [KNL]
4898 Use expedited grace-period primitives, for
4899 example, synchronize_rcu_expedited() instead
4900 of synchronize_rcu(). This reduces latency,
4901 but can increase CPU utilization, degrade
4902 real-time latency, and degrade energy efficiency.
4903 No effect on CONFIG_TINY_RCU kernels.
4905 rcupdate.rcu_normal= [KNL]
4906 Use only normal grace-period primitives,
4907 for example, synchronize_rcu() instead of
4908 synchronize_rcu_expedited(). This improves
4909 real-time latency, CPU utilization, and
4910 energy efficiency, but can expose users to
4911 increased grace-period latency. This parameter
4912 overrides rcupdate.rcu_expedited. No effect on
4913 CONFIG_TINY_RCU kernels.
4915 rcupdate.rcu_normal_after_boot= [KNL]
4916 Once boot has completed (that is, after
4917 rcu_end_inkernel_boot() has been invoked), use
4918 only normal grace-period primitives. No effect
4919 on CONFIG_TINY_RCU kernels.
4921 But note that CONFIG_PREEMPT_RT=y kernels enables
4922 this kernel boot parameter, forcibly setting
4923 it to the value one, that is, converting any
4924 post-boot attempt at an expedited RCU grace
4925 period to instead use normal non-expedited
4926 grace-period processing.
4928 rcupdate.rcu_task_collapse_lim= [KNL]
4929 Set the maximum number of callbacks present
4930 at the beginning of a grace period that allows
4931 the RCU Tasks flavors to collapse back to using
4932 a single callback queue. This switching only
4933 occurs when rcupdate.rcu_task_enqueue_lim is
4934 set to the default value of -1.
4936 rcupdate.rcu_task_contend_lim= [KNL]
4937 Set the minimum number of callback-queuing-time
4938 lock-contention events per jiffy required to
4939 cause the RCU Tasks flavors to switch to per-CPU
4940 callback queuing. This switching only occurs
4941 when rcupdate.rcu_task_enqueue_lim is set to
4942 the default value of -1.
4944 rcupdate.rcu_task_enqueue_lim= [KNL]
4945 Set the number of callback queues to use for the
4946 RCU Tasks family of RCU flavors. The default
4947 of -1 allows this to be automatically (and
4948 dynamically) adjusted. This parameter is intended
4951 rcupdate.rcu_task_ipi_delay= [KNL]
4952 Set time in jiffies during which RCU tasks will
4953 avoid sending IPIs, starting with the beginning
4954 of a given grace period. Setting a large
4955 number avoids disturbing real-time workloads,
4956 but lengthens grace periods.
4958 rcupdate.rcu_task_stall_timeout= [KNL]
4959 Set timeout in jiffies for RCU task stall warning
4960 messages. Disable with a value less than or equal
4963 rcupdate.rcu_self_test= [KNL]
4964 Run the RCU early boot self tests
4968 Run specified binary instead of /init from the ramdisk,
4969 used for early userspace startup. See initrd.
4972 force - Override the decision by the kernel to hide the
4973 advertisement of RDRAND support (this affects
4974 certain AMD processors because of buggy BIOS
4975 support, specifically around the suspend/resume
4979 Turn on/off individual RDT features. List is:
4980 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4982 E.g. to turn on cmt and turn off mba use:
4986 Format (x86 or x86_64):
4987 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
4989 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4991 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4992 (prefix with 'panic_' to set mode for panic
4994 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4995 reboot_force is either force or not specified,
4996 reboot_cpu is s[mp]#### with #### being the processor
4997 to be used for rebooting.
4999 refscale.holdoff= [KNL]
5000 Set test-start holdoff period. The purpose of
5001 this parameter is to delay the start of the
5002 test until boot completes in order to avoid
5005 refscale.loops= [KNL]
5006 Set the number of loops over the synchronization
5007 primitive under test. Increasing this number
5008 reduces noise due to loop start/end overhead,
5009 but the default has already reduced the per-pass
5010 noise to a handful of picoseconds on ca. 2020
5013 refscale.nreaders= [KNL]
5014 Set number of readers. The default value of -1
5015 selects N, where N is roughly 75% of the number
5016 of CPUs. A value of zero is an interesting choice.
5018 refscale.nruns= [KNL]
5019 Set number of runs, each of which is dumped onto
5022 refscale.readdelay= [KNL]
5023 Set the read-side critical-section duration,
5024 measured in microseconds.
5026 refscale.scale_type= [KNL]
5027 Specify the read-protection implementation to test.
5029 refscale.shutdown= [KNL]
5030 Shut down the system at the end of the performance
5031 test. This defaults to 1 (shut it down) when
5032 refscale is built into the kernel and to 0 (leave
5033 it running) when refscale is built as a module.
5035 refscale.verbose= [KNL]
5036 Enable additional printk() statements.
5038 refscale.verbose_batched= [KNL]
5039 Batch the additional printk() statements. If zero
5040 (the default) or negative, print everything. Otherwise,
5041 print every Nth verbose statement, where N is the value
5045 [KNL, SMP] Set scheduler's default relax_domain_level.
5046 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5048 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5049 Format: <base1>,<size1>[,<base2>,<size2>,...]
5050 Reserve I/O ports or memory so the kernel won't use
5051 them. If <base> is less than 0x10000, the region
5052 is assumed to be I/O ports; otherwise it is memory.
5054 reservetop= [X86-32]
5056 Reserves a hole at the top of the kernel virtual
5059 reset_devices [KNL] Force drivers to reset the underlying device
5060 during initialization.
5063 Specify the partition device for software suspend
5065 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5067 resume_offset= [SWSUSP]
5068 Specify the offset from the beginning of the partition
5069 given by "resume=" at which the swap header is located,
5070 in <PAGE_SIZE> units (needed only for swap files).
5071 See Documentation/power/swsusp-and-swap-files.rst
5073 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5074 read the resume files
5076 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5077 Useful for devices that are detected asynchronously
5078 (e.g. USB and MMC devices).
5080 hibernate= [HIBERNATION]
5081 noresume Don't check if there's a hibernation image
5082 present during boot.
5083 nocompress Don't compress/decompress hibernation images.
5084 no Disable hibernation and resume.
5085 protect_image Turn on image protection during restoration
5086 (that will set all pages holding image data
5087 during restoration read-only).
5089 retain_initrd [RAM] Keep initrd memory after extraction
5091 rfkill.default_state=
5092 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5093 etc. communication is blocked by default.
5096 rfkill.master_switch_mode=
5097 0 The "airplane mode" button does nothing.
5098 1 The "airplane mode" button toggles between everything
5099 blocked and the previous configuration.
5100 2 The "airplane mode" button toggles between everything
5101 blocked and everything unblocked.
5103 rhash_entries= [KNL,NET]
5104 Set number of hash buckets for route cache
5107 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5110 ro [KNL] Mount root device read-only on boot
5113 on Mark read-only kernel memory as read-only (default).
5114 off Leave read-only kernel memory writable for debugging.
5117 Enable the uart passthrough on the designated usb port
5118 on Rockchip SoCs. When active, the signals of the
5119 debug-uart get routed to the D+ and D- pins of the usb
5120 port and the regular usb controller gets disabled.
5122 root= [KNL] Root filesystem
5123 See name_to_dev_t comment in init/do_mounts.c.
5125 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5126 mount the root filesystem
5128 rootflags= [KNL] Set root filesystem mount option string
5130 rootfstype= [KNL] Set root filesystem type
5132 rootwait [KNL] Wait (indefinitely) for root device to show up.
5133 Useful for devices that are detected asynchronously
5134 (e.g. USB and MMC devices).
5136 rproc_mem=nn[KMG][@address]
5137 [KNL,ARM,CMA] Remoteproc physical memory block.
5138 Memory area to be used by remote processor image,
5141 rw [KNL] Mount root device read-write on boot
5143 S [KNL] Run init in single mode
5145 s390_iommu= [HW,S390]
5146 Set s390 IOTLB flushing mode
5148 With strict flushing every unmap operation will result in
5149 an IOTLB flush. Default is lazy flushing before reuse,
5152 s390_iommu_aperture= [KNL,S390]
5153 Specifies the size of the per device DMA address space
5154 accessible through the DMA and IOMMU APIs as a decimal
5155 factor of the size of main memory.
5156 The default is 1 meaning that one can concurrently use
5157 as many DMA addresses as physical memory is installed,
5158 if supported by hardware, and thus map all of memory
5159 once. With a value of 2 one can map all of memory twice
5160 and so on. As a special case a factor of 0 imposes no
5161 restrictions other than those given by hardware at the
5162 cost of significant additional memory use for tables.
5165 See drivers/net/irda/sa1100_ir.c.
5167 sched_verbose [KNL] Enables verbose scheduler debug messages.
5169 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5170 Allowed values are enable and disable. This feature
5171 incurs a small amount of overhead in the scheduler
5172 but is useful for debugging and performance tuning.
5174 sched_thermal_decay_shift=
5175 [KNL, SMP] Set a decay shift for scheduler thermal
5176 pressure signal. Thermal pressure signal follows the
5177 default decay period of other scheduler pelt
5178 signals(usually 32 ms but configurable). Setting
5179 sched_thermal_decay_shift will left shift the decay
5180 period for the thermal pressure signal by the shift
5182 i.e. with the default pelt decay period of 32 ms
5183 sched_thermal_decay_shift thermal pressure decay pr
5187 Format: integer between 0 and 10
5190 scftorture.holdoff= [KNL]
5191 Number of seconds to hold off before starting
5192 test. Defaults to zero for module insertion and
5193 to 10 seconds for built-in smp_call_function()
5196 scftorture.longwait= [KNL]
5197 Request ridiculously long waits randomly selected
5198 up to the chosen limit in seconds. Zero (the
5199 default) disables this feature. Please note
5200 that requesting even small non-zero numbers of
5201 seconds can result in RCU CPU stall warnings,
5202 softlockup complaints, and so on.
5204 scftorture.nthreads= [KNL]
5205 Number of kthreads to spawn to invoke the
5206 smp_call_function() family of functions.
5207 The default of -1 specifies a number of kthreads
5208 equal to the number of CPUs.
5210 scftorture.onoff_holdoff= [KNL]
5211 Number seconds to wait after the start of the
5212 test before initiating CPU-hotplug operations.
5214 scftorture.onoff_interval= [KNL]
5215 Number seconds to wait between successive
5216 CPU-hotplug operations. Specifying zero (which
5217 is the default) disables CPU-hotplug operations.
5219 scftorture.shutdown_secs= [KNL]
5220 The number of seconds following the start of the
5221 test after which to shut down the system. The
5222 default of zero avoids shutting down the system.
5223 Non-zero values are useful for automated tests.
5225 scftorture.stat_interval= [KNL]
5226 The number of seconds between outputting the
5227 current test statistics to the console. A value
5228 of zero disables statistics output.
5230 scftorture.stutter_cpus= [KNL]
5231 The number of jiffies to wait between each change
5232 to the set of CPUs under test.
5234 scftorture.use_cpus_read_lock= [KNL]
5235 Use use_cpus_read_lock() instead of the default
5236 preempt_disable() to disable CPU hotplug
5237 while invoking one of the smp_call_function*()
5240 scftorture.verbose= [KNL]
5241 Enable additional printk() statements.
5243 scftorture.weight_single= [KNL]
5244 The probability weighting to use for the
5245 smp_call_function_single() function with a zero
5246 "wait" parameter. A value of -1 selects the
5247 default if all other weights are -1. However,
5248 if at least one weight has some other value, a
5249 value of -1 will instead select a weight of zero.
5251 scftorture.weight_single_wait= [KNL]
5252 The probability weighting to use for the
5253 smp_call_function_single() function with a
5254 non-zero "wait" parameter. See weight_single.
5256 scftorture.weight_many= [KNL]
5257 The probability weighting to use for the
5258 smp_call_function_many() function with a zero
5259 "wait" parameter. See weight_single.
5260 Note well that setting a high probability for
5261 this weighting can place serious IPI load
5264 scftorture.weight_many_wait= [KNL]
5265 The probability weighting to use for the
5266 smp_call_function_many() function with a
5267 non-zero "wait" parameter. See weight_single
5270 scftorture.weight_all= [KNL]
5271 The probability weighting to use for the
5272 smp_call_function_all() function with a zero
5273 "wait" parameter. See weight_single and
5276 scftorture.weight_all_wait= [KNL]
5277 The probability weighting to use for the
5278 smp_call_function_all() function with a
5279 non-zero "wait" parameter. See weight_single
5282 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5283 xtime_lock contention on larger systems, and/or RCU lock
5284 contention on all systems with CONFIG_MAXSMP set.
5285 Format: { "0" | "1" }
5286 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5288 Note: increases power consumption, thus should only be
5289 enabled if running jitter sensitive (HPC/RT) workloads.
5291 security= [SECURITY] Choose a legacy "major" security module to
5292 enable at boot. This has been deprecated by the
5295 selinux= [SELINUX] Disable or enable SELinux at boot time.
5296 Format: { "0" | "1" }
5297 See security/selinux/Kconfig help text.
5302 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5303 Format: { "0" | "1" }
5304 See security/apparmor/Kconfig help text
5307 Default value is set via kernel config option.
5309 serialnumber [BUGS=X86-32]
5312 Maximal number of shapers.
5320 Enable merging of slabs with similar size when the
5321 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5324 Disable merging of slabs with similar size. May be
5325 necessary if there is some reason to distinguish
5326 allocs to different slabs, especially in hardened
5327 environments where the risk of heap overflows and
5328 layout control by attackers can usually be
5329 frustrated by disabling merging. This will reduce
5330 most of the exposure of a heap attack to a single
5331 cache (risks via metadata attacks are mostly
5332 unchanged). Debug options disable merging on their
5334 For more information see Documentation/vm/slub.rst.
5336 slab_max_order= [MM, SLAB]
5337 Determines the maximum allowed order for slabs.
5338 A high setting may cause OOMs due to memory
5339 fragmentation. Defaults to 1 for systems with
5340 more than 32MB of RAM, 0 otherwise.
5342 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5343 Enabling slub_debug allows one to determine the
5344 culprit if slab objects become corrupted. Enabling
5345 slub_debug can create guard zones around objects and
5346 may poison objects when not in use. Also tracks the
5347 last alloc / free. For more information see
5348 Documentation/vm/slub.rst.
5350 slub_max_order= [MM, SLUB]
5351 Determines the maximum allowed order for slabs.
5352 A high setting may cause OOMs due to memory
5353 fragmentation. For more information see
5354 Documentation/vm/slub.rst.
5356 slub_min_objects= [MM, SLUB]
5357 The minimum number of objects per slab. SLUB will
5358 increase the slab order up to slub_max_order to
5359 generate a sufficiently large slab able to contain
5360 the number of objects indicated. The higher the number
5361 of objects the smaller the overhead of tracking slabs
5362 and the less frequently locks need to be acquired.
5363 For more information see Documentation/vm/slub.rst.
5365 slub_min_order= [MM, SLUB]
5366 Determines the minimum page order for slabs. Must be
5367 lower than slub_max_order.
5368 For more information see Documentation/vm/slub.rst.
5370 slub_merge [MM, SLUB]
5371 Same with slab_merge.
5373 slub_nomerge [MM, SLUB]
5374 Same with slab_nomerge. This is supported for legacy.
5375 See slab_nomerge for more information.
5378 Format: <io1>[,<io2>[,...,<io8>]]
5380 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5381 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5382 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5383 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5384 smsc-ircc2.ircc_irq= [HW] IRQ line
5385 smsc-ircc2.ircc_dma= [HW] DMA channel
5386 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5387 0: Toshiba Satellite 1800 (GP data pin select)
5388 1: Fast pin select (default)
5391 smt [KNL,S390] Set the maximum number of threads (logical
5392 CPUs) to use per physical CPU on systems capable of
5393 symmetric multithreading (SMT). Will be capped to the
5394 actual hardware limit.
5396 Default: -1 (no limit)
5399 [KNL] Should the soft-lockup detector generate panics.
5402 A value of 1 instructs the soft-lockup detector
5403 to panic the machine when a soft-lockup occurs. It is
5404 also controlled by the kernel.softlockup_panic sysctl
5405 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5406 respective build-time switch to that functionality.
5408 softlockup_all_cpu_backtrace=
5409 [KNL] Should the soft-lockup detector generate
5410 backtraces on all cpus.
5413 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5414 See Documentation/admin-guide/laptops/sonypi.rst
5416 spectre_v2= [X86] Control mitigation of Spectre variant 2
5417 (indirect branch speculation) vulnerability.
5418 The default operation protects the kernel from
5421 on - unconditionally enable, implies
5423 off - unconditionally disable, implies
5425 auto - kernel detects whether your CPU model is
5428 Selecting 'on' will, and 'auto' may, choose a
5429 mitigation method at run time according to the
5430 CPU, the available microcode, the setting of the
5431 CONFIG_RETPOLINE configuration option, and the
5432 compiler with which the kernel was built.
5434 Selecting 'on' will also enable the mitigation
5435 against user space to user space task attacks.
5437 Selecting 'off' will disable both the kernel and
5438 the user space protections.
5440 Specific mitigations can also be selected manually:
5442 retpoline - replace indirect branches
5443 retpoline,generic - Retpolines
5444 retpoline,lfence - LFENCE; indirect branch
5445 retpoline,amd - alias for retpoline,lfence
5446 eibrs - enhanced IBRS
5447 eibrs,retpoline - enhanced IBRS + Retpolines
5448 eibrs,lfence - enhanced IBRS + LFENCE
5450 Not specifying this option is equivalent to
5454 [X86] Control mitigation of Spectre variant 2
5455 (indirect branch speculation) vulnerability between
5458 on - Unconditionally enable mitigations. Is
5459 enforced by spectre_v2=on
5461 off - Unconditionally disable mitigations. Is
5462 enforced by spectre_v2=off
5464 prctl - Indirect branch speculation is enabled,
5465 but mitigation can be enabled via prctl
5466 per thread. The mitigation control state
5467 is inherited on fork.
5470 - Like "prctl" above, but only STIBP is
5471 controlled per thread. IBPB is issued
5472 always when switching between different user
5476 - Same as "prctl" above, but all seccomp
5477 threads will enable the mitigation unless
5478 they explicitly opt out.
5481 - Like "seccomp" above, but only STIBP is
5482 controlled per thread. IBPB is issued
5483 always when switching between different
5484 user space processes.
5486 auto - Kernel selects the mitigation depending on
5487 the available CPU features and vulnerability.
5489 Default mitigation: "prctl"
5491 Not specifying this option is equivalent to
5492 spectre_v2_user=auto.
5494 spec_store_bypass_disable=
5495 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5496 (Speculative Store Bypass vulnerability)
5498 Certain CPUs are vulnerable to an exploit against a
5499 a common industry wide performance optimization known
5500 as "Speculative Store Bypass" in which recent stores
5501 to the same memory location may not be observed by
5502 later loads during speculative execution. The idea
5503 is that such stores are unlikely and that they can
5504 be detected prior to instruction retirement at the
5505 end of a particular speculation execution window.
5507 In vulnerable processors, the speculatively forwarded
5508 store can be used in a cache side channel attack, for
5509 example to read memory to which the attacker does not
5510 directly have access (e.g. inside sandboxed code).
5512 This parameter controls whether the Speculative Store
5513 Bypass optimization is used.
5515 On x86 the options are:
5517 on - Unconditionally disable Speculative Store Bypass
5518 off - Unconditionally enable Speculative Store Bypass
5519 auto - Kernel detects whether the CPU model contains an
5520 implementation of Speculative Store Bypass and
5521 picks the most appropriate mitigation. If the
5522 CPU is not vulnerable, "off" is selected. If the
5523 CPU is vulnerable the default mitigation is
5524 architecture and Kconfig dependent. See below.
5525 prctl - Control Speculative Store Bypass per thread
5526 via prctl. Speculative Store Bypass is enabled
5527 for a process by default. The state of the control
5528 is inherited on fork.
5529 seccomp - Same as "prctl" above, but all seccomp threads
5530 will disable SSB unless they explicitly opt out.
5532 Default mitigations:
5535 On powerpc the options are:
5537 on,auto - On Power8 and Power9 insert a store-forwarding
5538 barrier on kernel entry and exit. On Power7
5539 perform a software flush on kernel entry and
5543 Not specifying this option is equivalent to
5544 spec_store_bypass_disable=auto.
5546 spia_io_base= [HW,MTD]
5552 [X86] Enable split lock detection or bus lock detection
5554 When enabled (and if hardware support is present), atomic
5555 instructions that access data across cache line
5556 boundaries will result in an alignment check exception
5557 for split lock detection or a debug exception for
5562 warn - the kernel will emit rate-limited warnings
5563 about applications triggering the #AC
5564 exception or the #DB exception. This mode is
5565 the default on CPUs that support split lock
5566 detection or bus lock detection. Default
5567 behavior is by #AC if both features are
5568 enabled in hardware.
5570 fatal - the kernel will send SIGBUS to applications
5571 that trigger the #AC exception or the #DB
5572 exception. Default behavior is by #AC if
5573 both features are enabled in hardware.
5576 Set system wide rate limit to N bus locks
5577 per second for bus lock detection.
5580 N/A for split lock detection.
5583 If an #AC exception is hit in the kernel or in
5584 firmware (i.e. not while executing in user mode)
5585 the kernel will oops in either "warn" or "fatal"
5588 #DB exception for bus lock is triggered only when
5592 Control the Special Register Buffer Data Sampling
5595 Certain CPUs are vulnerable to an MDS-like
5596 exploit which can leak bits from the random
5599 By default, this issue is mitigated by
5600 microcode. However, the microcode fix can cause
5601 the RDRAND and RDSEED instructions to become
5602 much slower. Among other effects, this will
5603 result in reduced throughput from /dev/urandom.
5605 The microcode mitigation can be disabled with
5606 the following option:
5608 off: Disable mitigation and remove
5609 performance impact to RDRAND and RDSEED
5611 srcutree.counter_wrap_check [KNL]
5612 Specifies how frequently to check for
5613 grace-period sequence counter wrap for the
5614 srcu_data structure's ->srcu_gp_seq_needed field.
5615 The greater the number of bits set in this kernel
5616 parameter, the less frequently counter wrap will
5617 be checked for. Note that the bottom two bits
5620 srcutree.exp_holdoff [KNL]
5621 Specifies how many nanoseconds must elapse
5622 since the end of the last SRCU grace period for
5623 a given srcu_struct until the next normal SRCU
5624 grace period will be considered for automatic
5625 expediting. Set to zero to disable automatic
5629 Speculative Store Bypass Disable control
5631 On CPUs that are vulnerable to the Speculative
5632 Store Bypass vulnerability and offer a
5633 firmware based mitigation, this parameter
5634 indicates how the mitigation should be used:
5636 force-on: Unconditionally enable mitigation for
5637 for both kernel and userspace
5638 force-off: Unconditionally disable mitigation for
5639 for both kernel and userspace
5640 kernel: Always enable mitigation in the
5641 kernel, and offer a prctl interface
5642 to allow userspace to register its
5643 interest in being mitigated too.
5645 stack_guard_gap= [MM]
5646 override the default stack gap protection. The value
5647 is in page units and it defines how many pages prior
5648 to (for stacks growing down) resp. after (for stacks
5649 growing up) the main stack are reserved for no other
5650 mapping. Default value is 256 pages.
5652 stack_depot_disable= [KNL]
5653 Setting this to true through kernel command line will
5654 disable the stack depot thereby saving the static memory
5655 consumed by the stack hash table. By default this is set
5659 Enabled the stack tracer on boot up.
5661 stacktrace_filter=[function-list]
5662 [FTRACE] Limit the functions that the stack tracer
5663 will trace at boot up. function-list is a comma-separated
5664 list of functions. This list can be changed at run
5665 time by the stack_trace_filter file in the debugfs
5666 tracing directory. Note, this enables stack tracing
5667 and the stacktrace above is not needed.
5671 Set the STI (builtin display/keyboard on the HP-PARISC
5672 machines) console (graphic card) which should be used
5673 as the initial boot-console.
5674 See also comment in drivers/video/console/sticore.c.
5677 See comment in drivers/video/console/sticore.c.
5680 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5685 Enable or disable strict sigaltstack size checks
5686 against the required signal frame size which
5687 depends on the supported FPU features. This can
5688 be used to filter out binaries which have
5689 not yet been made aware of AT_MINSIGSTKSZ.
5691 sunrpc.min_resvport=
5692 sunrpc.max_resvport=
5694 SunRPC servers often require that client requests
5695 originate from a privileged port (i.e. a port in the
5696 range 0 < portnr < 1024).
5697 An administrator who wishes to reserve some of these
5698 ports for other uses may adjust the range that the
5699 kernel's sunrpc client considers to be privileged
5700 using these two parameters to set the minimum and
5701 maximum port values.
5703 sunrpc.svc_rpc_per_connection_limit=
5705 Limit the number of requests that the server will
5706 process in parallel from a single connection.
5707 The default value is 0 (no limit).
5711 Control how the NFS server code allocates CPUs to
5712 service thread pools. Depending on how many NICs
5713 you have and where their interrupts are bound, this
5714 option will affect which CPUs will do NFS serving.
5715 Note: this parameter cannot be changed while the
5716 NFS server is running.
5718 auto the server chooses an appropriate mode
5719 automatically using heuristics
5720 global a single global pool contains all CPUs
5721 percpu one pool for each CPU
5722 pernode one pool for each NUMA node (equivalent
5723 to global on non-NUMA machines)
5725 sunrpc.tcp_slot_table_entries=
5726 sunrpc.udp_slot_table_entries=
5728 Sets the upper limit on the number of simultaneous
5729 RPC calls that can be sent from the client to a
5730 server. Increasing these values may allow you to
5731 improve throughput, but will also increase the
5732 amount of memory reserved for use by the client.
5734 suspend.pm_test_delay=
5736 Sets the number of seconds to remain in a suspend test
5737 mode before resuming the system (see
5738 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5739 is set. Default value is 5.
5742 Format: { on | off | y | n | 1 | 0 }
5743 This parameter controls use of the Protected
5744 Execution Facility on pSeries.
5747 [KNL] Enable accounting of swap in memory resource
5748 controller if no parameter or 1 is given or disable
5749 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5751 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5752 Format: { <int> | force | noforce }
5753 <int> -- Number of I/O TLB slabs
5754 force -- force using of bounce buffers even if they
5755 wouldn't be automatically used by the kernel
5756 noforce -- Never use bounce buffers (for debugging)
5761 Set a sysctl parameter, right before loading the init
5762 process, as if the value was written to the respective
5763 /proc/sys/... file. Both '.' and '/' are recognized as
5764 separators. Unrecognized parameters and invalid values
5765 are reported in the kernel log. Sysctls registered
5766 later by a loaded module cannot be set this way.
5767 Example: sysctl.vm.swappiness=40
5769 sysfs.deprecated=0|1 [KNL]
5770 Enable/disable old style sysfs layout for old udev
5771 on older distributions. When this option is enabled
5772 very new udev will not work anymore. When this option
5773 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5774 in older udev will not work anymore.
5775 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5776 the kernel configuration.
5778 sysrq_always_enabled
5780 Ignore sysrq setting - this boot parameter will
5781 neutralize any effect of /proc/sys/kernel/sysrq.
5782 Useful for debugging.
5784 tcpmhash_entries= [KNL,NET]
5785 Set the number of tcp_metrics_hash slots.
5786 Default value is 8192 or 16384 depending on total
5787 ram pages. This is used to specify the TCP metrics
5788 cache size. See Documentation/networking/ip-sysctl.rst
5789 "tcp_no_metrics_save" section for more details.
5793 test_suspend= [SUSPEND][,N]
5794 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5795 standby suspend) or "freeze" (for suspend type freeze)
5796 as the system sleep state during system startup with
5797 the optional capability to repeat N number of times.
5798 The system is woken from this state using a
5799 wakeup-capable RTC alarm.
5801 thash_entries= [KNL,NET]
5802 Set number of hash buckets for TCP connection
5804 thermal.act= [HW,ACPI]
5805 -1: disable all active trip points in all thermal zones
5806 <degrees C>: override all lowest active trip points
5808 thermal.crt= [HW,ACPI]
5809 -1: disable all critical trip points in all thermal zones
5810 <degrees C>: override all critical trip points
5812 thermal.nocrt= [HW,ACPI]
5813 Set to disable actions on ACPI thermal zone
5814 critical and hot trip points.
5816 thermal.off= [HW,ACPI]
5817 1: disable ACPI thermal control
5819 thermal.psv= [HW,ACPI]
5820 -1: disable all passive trip points
5821 <degrees C>: override all passive trip points to this
5824 thermal.tzp= [HW,ACPI]
5825 Specify global default ACPI thermal zone polling rate
5826 <deci-seconds>: poll all this frequency
5827 0: no polling (default)
5830 Force threading of all interrupt handlers except those
5831 marked explicitly IRQF_NO_THREAD.
5835 Specify if the kernel should make use of the cpu
5836 topology information if the hardware supports this.
5837 The scheduler will make use of this information and
5838 e.g. base its process migration decisions on it.
5841 topology_updates= [KNL, PPC, NUMA]
5843 Specify if the kernel should ignore (off)
5844 topology updates sent by the hypervisor to this
5847 torture.disable_onoff_at_boot= [KNL]
5848 Prevent the CPU-hotplug component of torturing
5849 until after init has spawned.
5851 torture.ftrace_dump_at_shutdown= [KNL]
5852 Dump the ftrace buffer at torture-test shutdown,
5853 even if there were no errors. This can be a
5854 very costly operation when many torture tests
5855 are running concurrently, especially on systems
5856 with rotating-rust storage.
5858 torture.verbose_sleep_frequency= [KNL]
5859 Specifies how many verbose printk()s should be
5860 emitted between each sleep. The default of zero
5861 disables verbose-printk() sleeping.
5863 torture.verbose_sleep_duration= [KNL]
5864 Duration of each verbose-printk() sleep in jiffies.
5868 tpm_suspend_pcr=[HW,TPM]
5869 Format: integer pcr id
5870 Specify that at suspend time, the tpm driver
5871 should extend the specified pcr with zeros,
5872 as a workaround for some chips which fail to
5873 flush the last written pcr on TPM_SaveState.
5874 This will guarantee that all the other pcrs
5877 trace_buf_size=nn[KMG]
5878 [FTRACE] will set tracing buffer size on each cpu.
5880 trace_event=[event-list]
5881 [FTRACE] Set and start specified trace events in order
5882 to facilitate early boot debugging. The event-list is a
5883 comma-separated list of trace events to enable. See
5884 also Documentation/trace/events.rst
5886 trace_options=[option-list]
5887 [FTRACE] Enable or disable tracer options at boot.
5888 The option-list is a comma delimited list of options
5889 that can be enabled or disabled just as if you were
5890 to echo the option name into
5892 /sys/kernel/debug/tracing/trace_options
5894 For example, to enable stacktrace option (to dump the
5895 stack trace of each event), add to the command line:
5897 trace_options=stacktrace
5899 See also Documentation/trace/ftrace.rst "trace options"
5903 Have the tracepoints sent to printk as well as the
5904 tracing ring buffer. This is useful for early boot up
5905 where the system hangs or reboots and does not give the
5906 option for reading the tracing buffer or performing a
5907 ftrace_dump_on_oops.
5909 To turn off having tracepoints sent to printk,
5910 echo 0 > /proc/sys/kernel/tracepoint_printk
5911 Note, echoing 1 into this file without the
5912 tracepoint_printk kernel cmdline option has no effect.
5914 The tp_printk_stop_on_boot (see below) can also be used
5915 to stop the printing of events to console at
5920 Having tracepoints sent to printk() and activating high
5921 frequency tracepoints such as irq or sched, can cause
5922 the system to live lock.
5924 tp_printk_stop_on_boot[FTRACE]
5925 When tp_printk (above) is set, it can cause a lot of noise
5926 on the console. It may be useful to only include the
5927 printing of events during boot up, as user space may
5928 make the system inoperable.
5930 This command line option will stop the printing of events
5931 to console at the late_initcall_sync() time frame.
5934 [FTRACE] enable this option to disable tracing when a
5935 warning is hit. This turns off "tracing_on". Tracing can
5936 be enabled again by echoing '1' into the "tracing_on"
5937 file located in /sys/kernel/debug/tracing/
5939 This option is useful, as it disables the trace before
5940 the WARNING dump is called, which prevents the trace to
5941 be filled with content caused by the warning output.
5943 This option can also be set at run time via the sysctl
5944 option: kernel/traceoff_on_warning
5946 transparent_hugepage=
5948 Format: [always|madvise|never]
5949 Can be used to control the default behavior of the system
5950 with respect to transparent hugepages.
5951 See Documentation/admin-guide/mm/transhuge.rst
5954 trusted.source= [KEYS]
5956 This parameter identifies the trust source as a backend
5957 for trusted keys implementation. Supported trust
5961 If not specified then it defaults to iterating through
5962 the trust source list starting with TPM and assigns the
5963 first trust source as a backend which is initialized
5964 successfully during iteration.
5966 tsc= Disable clocksource stability checks for TSC.
5968 [x86] reliable: mark tsc clocksource as reliable, this
5969 disables clocksource verification at runtime, as well
5970 as the stability checks done at bootup. Used to enable
5971 high-resolution timer mode on older hardware, and in
5972 virtualized environment.
5973 [x86] noirqtime: Do not use TSC to do irq accounting.
5974 Used to run time disable IRQ_TIME_ACCOUNTING on any
5975 platforms where RDTSC is slow and this accounting
5977 [x86] unstable: mark the TSC clocksource as unstable, this
5978 marks the TSC unconditionally unstable at bootup and
5979 avoids any further wobbles once the TSC watchdog notices.
5980 [x86] nowatchdog: disable clocksource watchdog. Used
5981 in situations with strict latency requirements (where
5982 interruptions from clocksource watchdog are not
5985 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5986 value instead. Useful when the early TSC frequency discovery
5987 procedure is not reliable, such as on overclocked systems
5988 with CPUID.16h support and partial CPUID.15h support.
5989 Format: <unsigned int>
5991 tsx= [X86] Control Transactional Synchronization
5992 Extensions (TSX) feature in Intel processors that
5993 support TSX control.
5995 This parameter controls the TSX feature. The options are:
5997 on - Enable TSX on the system. Although there are
5998 mitigations for all known security vulnerabilities,
5999 TSX has been known to be an accelerator for
6000 several previous speculation-related CVEs, and
6001 so there may be unknown security risks associated
6002 with leaving it enabled.
6004 off - Disable TSX on the system. (Note that this
6005 option takes effect only on newer CPUs which are
6006 not vulnerable to MDS, i.e., have
6007 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6008 the new IA32_TSX_CTRL MSR through a microcode
6009 update. This new MSR allows for the reliable
6010 deactivation of the TSX functionality.)
6012 auto - Disable TSX if X86_BUG_TAA is present,
6013 otherwise enable TSX on the system.
6015 Not specifying this option is equivalent to tsx=off.
6017 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6020 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6021 Abort (TAA) vulnerability.
6023 Similar to Micro-architectural Data Sampling (MDS)
6024 certain CPUs that support Transactional
6025 Synchronization Extensions (TSX) are vulnerable to an
6026 exploit against CPU internal buffers which can forward
6027 information to a disclosure gadget under certain
6030 In vulnerable processors, the speculatively forwarded
6031 data can be used in a cache side channel attack, to
6032 access data to which the attacker does not have direct
6035 This parameter controls the TAA mitigation. The
6038 full - Enable TAA mitigation on vulnerable CPUs
6041 full,nosmt - Enable TAA mitigation and disable SMT on
6042 vulnerable CPUs. If TSX is disabled, SMT
6043 is not disabled because CPU is not
6044 vulnerable to cross-thread TAA attacks.
6045 off - Unconditionally disable TAA mitigation
6047 On MDS-affected machines, tsx_async_abort=off can be
6048 prevented by an active MDS mitigation as both vulnerabilities
6049 are mitigated with the same mechanism so in order to disable
6050 this mitigation, you need to specify mds=off too.
6052 Not specifying this option is equivalent to
6053 tsx_async_abort=full. On CPUs which are MDS affected
6054 and deploy MDS mitigation, TAA mitigation is not
6055 required and doesn't provide any additional
6059 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6061 turbografx.map[2|3]= [HW,JOY]
6062 TurboGraFX parallel port interface
6064 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6065 See also Documentation/input/devices/joystick-parport.rst
6067 udbg-immortal [PPC] When debugging early kernel crashes that
6068 happen after console_init() and before a proper
6069 console driver takes over, this boot options might
6070 help "seeing" what's going on.
6072 uhash_entries= [KNL,NET]
6073 Set number of hash buckets for UDP/UDP-Lite connections
6076 [USB] Ignore overcurrent events (default N).
6077 Some badly-designed motherboards generate lots of
6078 bogus events, for ports that aren't wired to
6079 anything. Set this parameter to avoid log spamming.
6080 Note that genuine overcurrent events won't be
6084 [X86] Cause panic on unknown NMI.
6086 usbcore.authorized_default=
6087 [USB] Default USB device authorization:
6088 (default -1 = authorized except for wireless USB,
6089 0 = not authorized, 1 = authorized, 2 = authorized
6090 if device connected to internal port)
6092 usbcore.autosuspend=
6093 [USB] The autosuspend time delay (in seconds) used
6094 for newly-detected USB devices (default 2). This
6095 is the time required before an idle device will be
6096 autosuspended. Devices for which the delay is set
6097 to a negative value won't be autosuspended at all.
6099 usbcore.usbfs_snoop=
6100 [USB] Set to log all usbfs traffic (default 0 = off).
6102 usbcore.usbfs_snoop_max=
6103 [USB] Maximum number of bytes to snoop in each URB
6106 usbcore.blinkenlights=
6107 [USB] Set to cycle leds on hubs (default 0 = off).
6109 usbcore.old_scheme_first=
6110 [USB] Start with the old device initialization
6111 scheme (default 0 = off).
6113 usbcore.usbfs_memory_mb=
6114 [USB] Memory limit (in MB) for buffers allocated by
6115 usbfs (default = 16, 0 = max = 2047).
6117 usbcore.use_both_schemes=
6118 [USB] Try the other device initialization scheme
6119 if the first one fails (default 1 = enabled).
6121 usbcore.initial_descriptor_timeout=
6122 [USB] Specifies timeout for the initial 64-byte
6123 USB_REQ_GET_DESCRIPTOR request in milliseconds
6124 (default 5000 = 5.0 seconds).
6126 usbcore.nousb [USB] Disable the USB subsystem
6129 [USB] A list of quirk entries to augment the built-in
6130 usb core quirk list. List entries are separated by
6131 commas. Each entry has the form
6132 VendorID:ProductID:Flags. The IDs are 4-digit hex
6133 numbers and Flags is a set of letters. Each letter
6134 will change the built-in quirk; setting it if it is
6135 clear and clearing it if it is set. The letters have
6136 the following meanings:
6137 a = USB_QUIRK_STRING_FETCH_255 (string
6138 descriptors must not be fetched using
6140 b = USB_QUIRK_RESET_RESUME (device can't resume
6141 correctly so reset it instead);
6142 c = USB_QUIRK_NO_SET_INTF (device can't handle
6143 Set-Interface requests);
6144 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6145 handle its Configuration or Interface
6147 e = USB_QUIRK_RESET (device can't be reset
6148 (e.g morph devices), don't use reset);
6149 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6150 more interface descriptions than the
6151 bNumInterfaces count, and can't handle
6152 talking to these interfaces);
6153 g = USB_QUIRK_DELAY_INIT (device needs a pause
6154 during initialization, after we read
6155 the device descriptor);
6156 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6157 high speed and super speed interrupt
6158 endpoints, the USB 2.0 and USB 3.0 spec
6159 require the interval in microframes (1
6160 microframe = 125 microseconds) to be
6161 calculated as interval = 2 ^
6163 Devices with this quirk report their
6164 bInterval as the result of this
6165 calculation instead of the exponent
6166 variable used in the calculation);
6167 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6168 handle device_qualifier descriptor
6170 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6171 generates spurious wakeup, ignore
6172 remote wakeup capability);
6173 k = USB_QUIRK_NO_LPM (device can't handle Link
6175 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6176 (Device reports its bInterval as linear
6177 frames instead of the USB 2.0
6179 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6180 to be disconnected before suspend to
6181 prevent spurious wakeup);
6182 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6183 pause after every control message);
6184 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6185 delay after resetting its port);
6186 Example: quirks=0781:5580:bk,0a5c:5834:gij
6189 [USBHID] The interval which mice are to be polled at.
6192 [USBHID] The interval which joysticks are to be polled at.
6195 [USBHID] The interval which keyboards are to be polled at.
6197 usb-storage.delay_use=
6198 [UMS] The delay in seconds before a new device is
6199 scanned for Logical Units (default 1).
6202 [UMS] A list of quirks entries to supplement or
6203 override the built-in unusual_devs list. List
6204 entries are separated by commas. Each entry has
6205 the form VID:PID:Flags where VID and PID are Vendor
6206 and Product ID values (4-digit hex numbers) and
6207 Flags is a set of characters, each corresponding
6208 to a common usb-storage quirk flag as follows:
6209 a = SANE_SENSE (collect more than 18 bytes
6210 of sense data, not on uas);
6211 b = BAD_SENSE (don't collect more than 18
6212 bytes of sense data, not on uas);
6213 c = FIX_CAPACITY (decrease the reported
6214 device capacity by one sector);
6215 d = NO_READ_DISC_INFO (don't use
6216 READ_DISC_INFO command, not on uas);
6217 e = NO_READ_CAPACITY_16 (don't use
6218 READ_CAPACITY_16 command);
6219 f = NO_REPORT_OPCODES (don't use report opcodes
6221 g = MAX_SECTORS_240 (don't transfer more than
6222 240 sectors at a time, uas only);
6223 h = CAPACITY_HEURISTICS (decrease the
6224 reported device capacity by one
6225 sector if the number is odd);
6226 i = IGNORE_DEVICE (don't bind to this
6228 j = NO_REPORT_LUNS (don't use report luns
6230 k = NO_SAME (do not use WRITE_SAME, uas only)
6231 l = NOT_LOCKABLE (don't try to lock and
6232 unlock ejectable media, not on uas);
6233 m = MAX_SECTORS_64 (don't transfer more
6234 than 64 sectors = 32 KB at a time,
6236 n = INITIAL_READ10 (force a retry of the
6237 initial READ(10) command, not on uas);
6238 o = CAPACITY_OK (accept the capacity
6239 reported by the device, not on uas);
6240 p = WRITE_CACHE (the device cache is ON
6241 by default, not on uas);
6242 r = IGNORE_RESIDUE (the device reports
6243 bogus residue values, not on uas);
6244 s = SINGLE_LUN (the device has only one
6246 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6247 commands, uas only);
6248 u = IGNORE_UAS (don't bind to the uas driver);
6249 w = NO_WP_DETECT (don't test whether the
6250 medium is write-protected).
6251 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6252 even if the device claims no cache,
6254 Example: quirks=0419:aaf5:rl,0421:0433:rc
6256 user_debug= [KNL,ARM]
6258 See arch/arm/Kconfig.debug help text.
6259 1 - undefined instruction events
6261 4 - invalid data aborts
6264 Example: user_debug=31
6267 [X86] Flags controlling user PTE allocations.
6269 nohigh = do not allocate PTE pages in
6270 HIGHMEM regardless of setting
6274 On X86_32, this is an alias for vdso32=. Otherwise:
6276 vdso=1: enable VDSO (the default)
6277 vdso=0: disable VDSO mapping
6279 vdso32= [X86] Control the 32-bit vDSO
6280 vdso32=1: enable 32-bit VDSO
6281 vdso32=0 or vdso32=2: disable 32-bit VDSO
6283 See the help text for CONFIG_COMPAT_VDSO for more
6284 details. If CONFIG_COMPAT_VDSO is set, the default is
6285 vdso32=0; otherwise, the default is vdso32=1.
6287 For compatibility with older kernels, vdso32=2 is an
6290 Try vdso32=0 if you encounter an error that says:
6291 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6294 vector=percpu: enable percpu vector domain
6296 video= [FB] Frame buffer configuration
6297 See Documentation/fb/modedb.rst.
6299 video.brightness_switch_enabled= [0,1]
6300 If set to 1, on receiving an ACPI notify event
6301 generated by hotkey, video driver will adjust brightness
6302 level and then send out the event to user space through
6303 the allocated input device; If set to 0, video driver
6304 will only send out the event without touching backlight
6309 [VMMIO] Memory mapped virtio (platform) device.
6311 <size>@<baseaddr>:<irq>[:<id>]
6313 <size> := size (can use standard suffixes
6315 <baseaddr> := physical base address
6316 <irq> := interrupt number (as passed to
6318 <id> := (optional) platform device id
6320 virtio_mmio.device=1K@0x100b0000:48:7
6322 Can be used multiple times for multiple devices.
6324 vga= [BOOT,X86-32] Select a particular video mode
6325 See Documentation/x86/boot.rst and
6326 Documentation/admin-guide/svga.rst.
6327 Use vga=ask for menu.
6328 This is actually a boot loader parameter; the value is
6329 passed to the kernel using a special protocol.
6331 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6332 May slow down system boot speed, especially when
6333 enabled on systems with a large amount of memory.
6334 All options are enabled by default, and this
6335 interface is meant to allow for selectively
6336 enabling or disabling specific virtual memory
6339 Available options are:
6340 P Enable page structure init time poisoning
6341 - Disable all of the above options
6343 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6344 size of <nn>. This can be used to increase the
6345 minimum size (128MB on x86). It can also be used to
6346 decrease the size and leave more room for directly
6349 vmcp_cma=nn[MG] [KNL,S390]
6350 Sets the memory size reserved for contiguous memory
6351 allocations for the vmcp device driver.
6353 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6356 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6359 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6363 Controls the behavior of vsyscalls (i.e. calls to
6364 fixed addresses of 0xffffffffff600x00 from legacy
6365 code). Most statically-linked binaries and older
6366 versions of glibc use these calls. Because these
6367 functions are at fixed addresses, they make nice
6368 targets for exploits that can control RIP.
6370 emulate [default] Vsyscalls turn into traps and are
6371 emulated reasonably safely. The vsyscall
6374 xonly Vsyscalls turn into traps and are
6375 emulated reasonably safely. The vsyscall
6376 page is not readable.
6378 none Vsyscalls don't work at all. This makes
6379 them quite hard to use for exploits but
6380 might break your system.
6382 vt.color= [VT] Default text color.
6383 Format: 0xYX, X = foreground, Y = background.
6384 Default: 0x07 = light gray on black.
6386 vt.cur_default= [VT] Default cursor shape.
6387 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6388 the parameters of the <Esc>[?A;B;Cc escape sequence;
6389 see VGA-softcursor.txt. Default: 2 = underline.
6391 vt.default_blu= [VT]
6392 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6393 Change the default blue palette of the console.
6394 This is a 16-member array composed of values
6397 vt.default_grn= [VT]
6398 Format: <green0>,<green1>,<green2>,...,<green15>
6399 Change the default green palette of the console.
6400 This is a 16-member array composed of values
6403 vt.default_red= [VT]
6404 Format: <red0>,<red1>,<red2>,...,<red15>
6405 Change the default red palette of the console.
6406 This is a 16-member array composed of values
6412 Set system-wide default UTF-8 mode for all tty's.
6413 Default is 1, i.e. UTF-8 mode is enabled for all
6414 newly opened terminals.
6416 vt.global_cursor_default=
6419 Set system-wide default for whether a cursor
6420 is shown on new VTs. Default is -1,
6421 i.e. cursors will be created by default unless
6422 overridden by individual drivers. 0 will hide
6423 cursors, 1 will display them.
6425 vt.italic= [VT] Default color for italic text; 0-15.
6428 vt.underline= [VT] Default color for underlined text; 0-15.
6431 watchdog timers [HW,WDT] For information on watchdog timers,
6432 see Documentation/watchdog/watchdog-parameters.rst
6433 or other driver-specific files in the
6434 Documentation/watchdog/ directory.
6438 Set the hard lockup detector stall duration
6439 threshold in seconds. The soft lockup detector
6440 threshold is set to twice the value. A value of 0
6441 disables both lockup detectors. Default is 10
6444 workqueue.watchdog_thresh=
6445 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6446 warn stall conditions and dump internal state to
6447 help debugging. 0 disables workqueue stall
6448 detection; otherwise, it's the stall threshold
6449 duration in seconds. The default value is 30 and
6450 it can be updated at runtime by writing to the
6451 corresponding sysfs file.
6453 workqueue.disable_numa
6454 By default, all work items queued to unbound
6455 workqueues are affine to the NUMA nodes they're
6456 issued on, which results in better behavior in
6457 general. If NUMA affinity needs to be disabled for
6458 whatever reason, this option can be used. Note
6459 that this also can be controlled per-workqueue for
6460 workqueues visible under /sys/bus/workqueue/.
6462 workqueue.power_efficient
6463 Per-cpu workqueues are generally preferred because
6464 they show better performance thanks to cache
6465 locality; unfortunately, per-cpu workqueues tend to
6466 be more power hungry than unbound workqueues.
6468 Enabling this makes the per-cpu workqueues which
6469 were observed to contribute significantly to power
6470 consumption unbound, leading to measurably lower
6471 power usage at the cost of small performance
6474 The default value of this parameter is determined by
6475 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6477 workqueue.debug_force_rr_cpu
6478 Workqueue used to implicitly guarantee that work
6479 items queued without explicit CPU specified are put
6480 on the local CPU. This guarantee is no longer true
6481 and while local CPU is still preferred work items
6482 may be put on foreign CPUs. This debug option
6483 forces round-robin CPU selection to flush out
6484 usages which depend on the now broken guarantee.
6485 When enabled, memory and cache locality will be
6488 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6489 default x2apic cluster mode on platforms
6492 xen_512gb_limit [KNL,X86-64,XEN]
6493 Restricts the kernel running paravirtualized under Xen
6494 to use only up to 512 GB of RAM. The reason to do so is
6495 crash analysis tools and Xen tools for doing domain
6496 save/restore/migration must be enabled to handle larger
6499 xen_emul_unplug= [HW,X86,XEN]
6500 Unplug Xen emulated devices
6501 Format: [unplug0,][unplug1]
6502 ide-disks -- unplug primary master IDE devices
6503 aux-ide-disks -- unplug non-primary-master IDE devices
6504 nics -- unplug network devices
6505 all -- unplug all emulated devices (NICs and IDE disks)
6506 unnecessary -- unplugging emulated devices is
6507 unnecessary even if the host did not respond to
6509 never -- do not unplug even if version check succeeds
6511 xen_legacy_crash [X86,XEN]
6512 Crash from Xen panic notifier, without executing late
6513 panic() code such as dumping handler.
6515 xen_nopvspin [X86,XEN]
6516 Disables the qspinlock slowpath using Xen PV optimizations.
6517 This parameter is obsoleted by "nopvspin" parameter, which
6518 has equivalent effect for XEN platform.
6521 Disables the PV optimizations forcing the HVM guest to
6522 run as generic HVM guest with no PV drivers.
6523 This option is obsoleted by the "nopv" option, which
6524 has equivalent effect for XEN platform.
6526 xen_no_vector_callback
6527 [KNL,X86,XEN] Disable the vector callback for Xen
6528 event channel interrupts.
6530 xen_scrub_pages= [XEN]
6531 Boolean option to control scrubbing pages before giving them back
6532 to Xen, for use by other domains. Can be also changed at runtime
6533 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6534 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6536 xen_timer_slop= [X86-64,XEN]
6537 Set the timer slop (in nanoseconds) for the virtual Xen
6538 timers (default is 100000). This adjusts the minimum
6539 delta of virtualized Xen timers, where lower values
6540 improve timer resolution at the expense of processing
6541 more timer interrupts.
6543 xen.balloon_boot_timeout= [XEN]
6544 The time (in seconds) to wait before giving up to boot
6545 in case initial ballooning fails to free enough memory.
6546 Applies only when running as HVM or PVH guest and
6547 started with less memory configured than allowed at
6548 max. Default is 180.
6550 xen.event_eoi_delay= [XEN]
6551 How long to delay EOI handling in case of event
6552 storms (jiffies). Default is 10.
6554 xen.event_loop_timeout= [XEN]
6555 After which time (jiffies) the event handling loop
6556 should start to delay EOI handling. Default is 2.
6558 xen.fifo_events= [XEN]
6559 Boolean parameter to disable using fifo event handling
6560 even if available. Normally fifo event handling is
6561 preferred over the 2-level event handling, as it is
6562 fairer and the number of possible event channels is
6563 much higher. Default is on (use fifo events).
6565 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6566 Disables the PV optimizations forcing the guest to run
6567 as generic guest with no PV drivers. Currently support
6568 XEN HVM, KVM, HYPER_V and VMWARE guest.
6570 nopvspin [X86,XEN,KVM]
6571 Disables the qspinlock slow path using PV optimizations
6572 which allow the hypervisor to 'idle' the guest on lock
6575 xirc2ps_cs= [NET,PCMCIA]
6577 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6580 By default on POWER9 and above, the kernel will
6581 natively use the XIVE interrupt controller. This option
6582 allows the fallback firmware mode to be used:
6584 off Fallback to firmware control of XIVE interrupt
6585 controller on both pseries and powernv
6586 platforms. Only useful on POWER9 and above.
6588 xive.store-eoi=off [PPC]
6589 By default on POWER10 and above, the kernel will use
6590 stores for EOI handling when the XIVE interrupt mode
6591 is active. This option allows the XIVE driver to use
6592 loads instead, as on POWER9.
6594 xhci-hcd.quirks [USB,KNL]
6595 A hex value specifying bitmask with supplemental xhci
6596 host controller quirks. Meaning of each bit can be
6597 consulted in header drivers/usb/host/xhci.h.
6600 Format: { early | on | rw | ro | off }
6601 Controls if xmon debugger is enabled. Default is off.
6602 Passing only "xmon" is equivalent to "xmon=early".
6603 early Call xmon as early as possible on boot; xmon
6604 debugger is called from setup_arch().
6605 on xmon debugger hooks will be installed so xmon
6606 is only called on a kernel crash. Default mode,
6607 i.e. either "ro" or "rw" mode, is controlled
6608 with CONFIG_XMON_DEFAULT_RO_MODE.
6609 rw xmon debugger hooks will be installed so xmon
6610 is called only on a kernel crash, mode is write,
6611 meaning SPR registers, memory and, other data
6612 can be written using xmon commands.
6613 ro same as "rw" option above but SPR registers,
6614 memory, and other data can't be written using
6616 off xmon is disabled.