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
324 pgtbl_v1 - Use v1 page table for DMA-API (Default).
325 pgtbl_v2 - Use v2 page table for DMA-API.
327 amd_iommu_dump= [HW,X86-64]
328 Enable AMD IOMMU driver option to dump the ACPI table
329 for AMD IOMMU. With this option enabled, AMD IOMMU
330 driver will print ACPI tables for AMD IOMMU during
331 IOMMU initialization.
333 amd_iommu_intr= [HW,X86-64]
334 Specifies one of the following AMD IOMMU interrupt
336 legacy - Use legacy interrupt remapping mode.
337 vapic - Use virtual APIC mode, which allows IOMMU
338 to inject interrupts directly into guest.
339 This mode requires kvm-amd.avic=1.
340 (Default when IOMMU HW support is present.)
342 amijoy.map= [HW,JOY] Amiga joystick support
343 Map of devices attached to JOY0DAT and JOY1DAT
345 See also Documentation/input/joydev/joystick.rst
347 analog.map= [HW,JOY] Analog joystick and gamepad support
348 Specifies type or capabilities of an analog joystick
349 connected to one of 16 gameports
350 Format: <type1>,<type2>,..<type16>
353 Power management functions (SPARCstation-4/5 + deriv.)
355 Disable APC CPU standby support. SPARCstation-Fox does
356 not play well with APC CPU idle - disable it if you have
357 APC and your system crashes randomly.
359 apic= [APIC,X86] Advanced Programmable Interrupt Controller
360 Change the output verbosity while booting
361 Format: { quiet (default) | verbose | debug }
362 Change the amount of debugging information output
363 when initialising the APIC and IO-APIC components.
364 For X86-32, this can also be used to specify an APIC
366 Format: apic=driver_name
367 Examples: apic=bigsmp
369 apic_extnmi= [APIC,X86] External NMI delivery setting
370 Format: { bsp (default) | all | none }
371 bsp: External NMI is delivered only to CPU 0
372 all: External NMIs are broadcast to all CPUs as a
374 none: External NMI is masked for all CPUs. This is
375 useful so that a dump capture kernel won't be
379 See Documentation/networking/ipv6.rst.
381 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
382 Limit apic dumping. The parameter defines the maximal
383 number of local apics being dumped. Also it is possible
384 to set it to "all" by meaning -- no limit here.
385 Format: { 1 (default) | 2 | ... | all }.
386 The parameter valid if only apic=debug or
387 apic=verbose is specified.
388 Example: apic=debug show_lapic=all
390 apm= [APM] Advanced Power Management
391 See header of arch/x86/kernel/apm_32.c.
393 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
394 Format: <io>,<irq>,<nodeID>
396 arm64.nobti [ARM64] Unconditionally disable Branch Target
397 Identification support
399 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
402 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
405 arm64.nosve [ARM64] Unconditionally disable Scalable Vector
408 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix
413 atarimouse= [HW,MOUSE] Atari Mouse
415 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
416 EzKey and similar keyboards
418 atkbd.reset= [HW] Reset keyboard during initialization
420 atkbd.set= [HW] Select keyboard code set
421 Format: <int> (2 = AT (default), 3 = PS/2)
423 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
426 atkbd.softraw= [HW] Choose between synthetic and real raw mode
427 Format: <bool> (0 = real, 1 = synthetic (default))
429 atkbd.softrepeat= [HW]
430 Use software keyboard repeat
432 audit= [KNL] Enable the audit sub-system
433 Format: { "0" | "1" | "off" | "on" }
434 0 | off - kernel audit is disabled and can not be
435 enabled until the next reboot
436 unset - kernel audit is initialized but disabled and
437 will be fully enabled by the userspace auditd.
438 1 | on - kernel audit is initialized and partially
439 enabled, storing at most audit_backlog_limit
440 messages in RAM until it is fully enabled by the
444 audit_backlog_limit= [KNL] Set the audit queue size limit.
445 Format: <int> (must be >=0)
448 bau= [X86_UV] Enable the BAU on SGI UV. The default
449 behavior is to disable the BAU (i.e. bau=0).
450 Format: { "0" | "1" }
453 unset - Disable the BAU.
455 baycom_epp= [HW,AX25]
458 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
460 See header of drivers/net/hamradio/baycom_par.c.
462 baycom_ser_fdx= [HW,AX25]
463 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
464 Format: <io>,<irq>,<mode>[,<baud>]
465 See header of drivers/net/hamradio/baycom_ser_fdx.c.
467 baycom_ser_hdx= [HW,AX25]
468 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
469 Format: <io>,<irq>,<mode>
470 See header of drivers/net/hamradio/baycom_ser_hdx.c.
473 Disable BERT OS support on buggy BIOSes.
475 bgrt_disable [ACPI][X86]
476 Disable BGRT to avoid flickering OEM logo.
478 blkdevparts= Manual partition parsing of block device(s) for
479 embedded devices based on command line input.
480 See Documentation/block/cmdline-partition.rst
482 boot_delay= Milliseconds to delay each printk during boot.
483 Values larger than 10 seconds (10000) are changed to
488 Extended command line options can be added to an initrd
489 and this will cause the kernel to look for it.
491 See Documentation/admin-guide/bootconfig.rst
493 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
494 bttv.radio= Most important insmod options are available as
496 bttv.pll= See Documentation/admin-guide/media/bttv.rst
499 bulk_remove=off [PPC] This parameter disables the use of the pSeries
500 firmware feature for flushing multiple hpte entries
503 c101= [NET] Moxa C101 synchronous serial card
505 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
506 Sometimes CPU hardware bugs make them report the cache
507 size incorrectly. The kernel will attempt work arounds
508 to fix known problems, but for some CPUs it is not
509 possible to determine what the correct size should be.
510 This option provides an override for these situations.
513 [NET] Specifies amount of time (in seconds) that
514 the kernel should wait for a network carrier. By default
515 it waits 120 seconds.
517 ca_keys= [KEYS] This parameter identifies a specific key(s) on
518 the system trusted keyring to be used for certificate
520 format: { id:<keyid> | builtin }
522 cca= [MIPS] Override the kernel pages' cache coherency
523 algorithm. Accepted values range from 0 to 7
524 inclusive. See arch/mips/include/asm/pgtable-bits.h
525 for platform specific values (SB1, Loongson3 and
528 ccw_timeout_log [S390]
529 See Documentation/s390/common_io.rst for details.
531 cgroup_disable= [KNL] Disable a particular controller or optional feature
532 Format: {name of the controller(s) or feature(s) to disable}
533 The effects of cgroup_disable=foo are:
534 - foo isn't auto-mounted if you mount all cgroups in
536 - foo isn't visible as an individually mountable
538 - if foo is an optional feature then the feature is
539 disabled and corresponding cgroup files are not
541 {Currently only "memory" controller deal with this and
542 cut the overhead, others just disable the usage. So
543 only cgroup_disable=memory is actually worthy}
544 Specifying "pressure" disables per-cgroup pressure
545 stall information accounting feature
547 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
548 Format: { { controller | "all" | "named" }
549 [,{ controller | "all" | "named" }...] }
550 Like cgroup_disable, but only applies to cgroup v1;
551 the blacklisted controllers remain available in cgroup2.
552 "all" blacklists all controllers and "named" disables
553 named mounts. Specifying both "all" and "named" disables
556 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
558 nosocket -- Disable socket memory accounting.
559 nokmem -- Disable kernel memory accounting.
561 checkreqprot= [SELINUX] Set initial checkreqprot flag value.
562 Format: { "0" | "1" }
563 See security/selinux/Kconfig help text.
564 0 -- check protection applied by kernel (includes
565 any implied execute protection).
566 1 -- check protection requested by application.
567 Default value is set via a kernel config option.
568 Value can be changed at runtime via
569 /sys/fs/selinux/checkreqprot.
570 Setting checkreqprot to 1 is deprecated.
573 See Documentation/s390/common_io.rst for details.
575 clearcpuid=X[,X...] [X86]
576 Disable CPUID feature X for the kernel. See
577 arch/x86/include/asm/cpufeatures.h for the valid bit
578 numbers X. Note the Linux-specific bits are not necessarily
579 stable over kernel options, but the vendor-specific
581 X can also be a string as appearing in the flags: line
582 in /proc/cpuinfo which does not have the above
583 instability issue. However, not all features have names
585 Note that using this option will taint your kernel.
586 Also note that user programs calling CPUID directly
587 or using the feature without checking anything
588 will still see it. This just prevents it from
589 being used by the kernel or shown in /proc/cpuinfo.
590 Also note the kernel might malfunction if you disable
595 Prevents the clock framework from automatically gating
596 clocks that have not been explicitly enabled by a Linux
597 device driver but are enabled in hardware at reset or
598 by the bootloader/firmware. Note that this does not
599 force such clocks to be always-on nor does it reserve
600 those clocks in any way. This parameter is useful for
601 debug and development, but should not be needed on a
602 platform with proper driver support. For more
603 information, see Documentation/driver-api/clk.rst.
605 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
607 Forces specified clocksource (if available) to be used
608 when calculating gettimeofday(). If specified
609 clocksource is not available, it defaults to PIT.
610 Format: { pit | tsc | cyclone | pmtmr }
612 clocksource= Override the default clocksource
614 Override the default clocksource and use the clocksource
615 with the name specified.
616 Some clocksource names to choose from, depending on
618 [all] jiffies (this is the base, fallback clocksource)
620 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
621 pxa_timer,timer3,32k_counter,timer0_1
622 [X86-32] pit,hpet,tsc;
623 scx200_hrt on Geode; cyclone on IBM x440
631 clocksource.arm_arch_timer.evtstrm=
634 Enable/disable the eventstream feature of the ARM
635 architected timer so that code using WFE-based polling
636 loops can be debugged more effectively on production
639 clocksource.max_cswd_read_retries= [KNL]
640 Number of clocksource_watchdog() retries due to
641 external delays before the clock will be marked
642 unstable. Defaults to two retries, that is,
643 three attempts to read the clock under test.
645 clocksource.verify_n_cpus= [KNL]
646 Limit the number of CPUs checked for clocksources
647 marked with CLOCK_SOURCE_VERIFY_PERCPU that
648 are marked unstable due to excessive skew.
649 A negative value says to check all CPUs, while
650 zero says not to check any. Values larger than
651 nr_cpu_ids are silently truncated to nr_cpu_ids.
652 The actual CPUs are chosen randomly, with
653 no replacement if the same CPU is chosen twice.
655 clocksource-wdtest.holdoff= [KNL]
656 Set the time in seconds that the clocksource
657 watchdog test waits before commencing its tests.
658 Defaults to zero when built as a module and to
659 10 seconds when built into the kernel.
661 cma=nn[MG]@[start[MG][-end[MG]]]
663 Sets the size of kernel global memory area for
664 contiguous memory allocations and optionally the
665 placement constraint by the physical address range of
666 memory allocations. A value of 0 disables CMA
667 altogether. For more information, see
668 kernel/dma/contiguous.c
672 Sets the size of kernel per-numa memory area for
673 contiguous memory allocations. A value of 0 disables
674 per-numa CMA altogether. And If this option is not
675 specificed, the default value is 0.
676 With per-numa CMA enabled, DMA users on node nid will
677 first try to allocate buffer from the pernuma area
678 which is located in node nid, if the allocation fails,
679 they will fallback to the global default memory area.
681 cmo_free_hint= [PPC] Format: { yes | no }
682 Specify whether pages are marked as being inactive
683 when they are freed. This is used in CMO environments
684 to determine OS memory pressure for page stealing by
688 coherent_pool=nn[KMG] [ARM,KNL]
689 Sets the size of memory pool for coherent, atomic dma
690 allocations, by default set to 256K.
692 com20020= [HW,NET] ARCnet - COM20020 chipset
694 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
696 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
700 ARCnet - COM90xx chipset (memory-mapped buffers)
701 Format: <io>[,<irq>[,<memstart>]]
703 condev= [HW,S390] console device
706 console= [KNL] Output console device and options.
708 tty<n> Use the virtual console device <n>.
712 Use the specified serial port. The options are of
713 the form "bbbbpnf", where "bbbb" is the baud rate,
714 "p" is parity ("n", "o", or "e"), "n" is number of
715 bits, and "f" is flow control ("r" for RTS or
716 omit it). Default is "9600n8".
718 See Documentation/admin-guide/serial-console.rst for more
720 Documentation/networking/netconsole.rst for an
723 uart[8250],io,<addr>[,options]
724 uart[8250],mmio,<addr>[,options]
725 uart[8250],mmio16,<addr>[,options]
726 uart[8250],mmio32,<addr>[,options]
727 uart[8250],0x<addr>[,options]
728 Start an early, polled-mode console on the 8250/16550
729 UART at the specified I/O port or MMIO address,
730 switching to the matching ttyS device later.
731 MMIO inter-register address stride is either 8-bit
732 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
733 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
734 to be equivalent to 'mmio'. 'options' are specified in
735 the same format described for ttyS above; if unspecified,
736 the h/w is not re-initialized.
738 hvc<n> Use the hypervisor console device <n>. This is for
739 both Xen and PowerPC hypervisors.
742 Use to disable console output, i.e., to have kernel
743 console messages discarded.
744 This must be the only console= parameter used on the
747 If the device connected to the port is not a TTY but a braille
748 device, prepend "brl," before the device type, for instance
750 For now, only VisioBraille is supported.
753 [KNL] Change console messages format
755 By default we print messages on consoles in
756 "[time stamp] text\n" format (time stamp may not be
757 printed, depending on CONFIG_PRINTK_TIME or
758 `printk_time' param).
760 Switch to syslog format: "<%u>[time stamp] text\n"
761 IOW, each message will have a facility and loglevel
762 prefix. The format is similar to one used by syslog()
763 syscall, or to executing "dmesg -S --raw" or to reading
766 consoleblank= [KNL] The console blank (screen saver) timeout in
767 seconds. A value of 0 disables the blank timer.
771 [KNL] Change the default value for
772 /proc/<pid>/coredump_filter.
773 See also Documentation/filesystems/proc.rst.
775 coresight_cpu_debug.enable
778 Enable/disable the CPU sampling based debugging.
779 0: default value, disable debugging
780 1: enable debugging at boot time
782 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
784 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
786 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
787 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
788 Some features depend on CPU0. Known dependencies are:
789 1. Resume from suspend/hibernate depends on CPU0.
790 Suspend/hibernate will fail if CPU0 is offline and you
791 need to online CPU0 before suspend/hibernate.
792 2. PIC interrupts also depend on CPU0. CPU0 can't be
793 removed if a PIC interrupt is detected.
794 It's said poweroff/reboot may depend on CPU0 on some
795 machines although I haven't seen such issues so far
796 after CPU0 is offline on a few tested machines.
797 If the dependencies are under your control, you can
798 turn on cpu0_hotplug.
800 cpuidle.off=1 [CPU_IDLE]
801 disable the cpuidle sub-system
804 [CPU_IDLE] Name of the cpuidle governor to use.
806 cpufreq.off=1 [CPU_FREQ]
807 disable the cpufreq sub-system
809 cpufreq.default_governor=
810 [CPU_FREQ] Name of the default cpufreq governor or
811 policy to use. This governor must be registered in the
812 kernel before the cpufreq driver probes.
815 [X86] Delay for N microsec between assert and de-assert
816 of APIC INIT to start processors. This delay occurs
817 on every CPU online, such as boot, and resume from suspend.
820 crash_kexec_post_notifiers
821 Run kdump after running panic-notifiers and dumping
822 kmsg. This only for the users who doubt kdump always
823 succeeds in any situation.
824 Note that this also increases risks of kdump failure,
825 because some panic notifiers can make the crashed
826 kernel more unstable.
828 crashkernel=size[KMG][@offset[KMG]]
829 [KNL] Using kexec, Linux can switch to a 'crash kernel'
830 upon panic. This parameter reserves the physical
831 memory region [offset, offset + size] for that kernel
832 image. If '@offset' is omitted, then a suitable offset
833 is selected automatically.
834 [KNL, X86-64] Select a region under 4G first, and
835 fall back to reserve region above 4G when '@offset'
836 hasn't been specified.
837 See Documentation/admin-guide/kdump/kdump.rst for further details.
839 crashkernel=range1:size1[,range2:size2,...][@offset]
840 [KNL] Same as above, but depends on the memory
841 in the running system. The syntax of range is
842 start-[end] where start and end are both
843 a memory unit (amount[KMG]). See also
844 Documentation/admin-guide/kdump/kdump.rst for an example.
846 crashkernel=size[KMG],high
847 [KNL, X86-64, ARM64] range could be above 4G. Allow kernel
848 to allocate physical memory region from top, so could
849 be above 4G if system have more than 4G ram installed.
850 Otherwise memory region will be allocated below 4G, if
852 It will be ignored if crashkernel=X is specified.
853 crashkernel=size[KMG],low
854 [KNL, X86-64] range under 4G. When crashkernel=X,high
855 is passed, kernel could allocate physical memory region
856 above 4G, that cause second kernel crash on system
857 that require some amount of low memory, e.g. swiotlb
858 requires at least 64M+32K low memory, also enough extra
859 low memory is needed to make sure DMA buffers for 32-bit
860 devices won't run out. Kernel would try to allocate
861 at least 256M below 4G automatically.
862 This one lets the user specify own low range under 4G
863 for second kernel instead.
864 0: to disable low allocation.
865 It will be ignored when crashkernel=X,high is not used
866 or memory reserved is below 4G.
868 [KNL, ARM64] range in low memory.
869 This one lets the user specify a low range in the
870 DMA zone for the crash dump kernel.
871 It will be ignored when crashkernel=X,high is not used
872 or memory reserved is located in the DMA zones.
875 [KNL] Disable crypto self-tests
880 cs89x0_media= [HW,NET]
881 Format: { rj45 | aui | bnc }
883 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
884 handling. When switched on, additional debug data is
885 printed to the console in case a hanging CPU is
886 detected, and that CPU is pinged again in order to try
887 to resolve the hang situation.
888 0: disable csdlock debugging (default)
889 1: enable basic csdlock debugging (minor impact)
890 ext: enable extended csdlock debugging (more impact,
894 See header of drivers/s390/block/dasd_devmap.c.
896 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
897 (one device per port)
898 Format: <port#>,<type>
899 See also Documentation/input/devices/joystick-parport.rst
901 debug [KNL] Enable kernel debugging (events log level).
904 [KNL] Enable printing [hashed] pointers early in the
905 boot sequence. If enabled, we use a weak hash instead
906 of siphash to hash pointers. Use this option if you are
907 seeing instances of '(___ptrval___)') and need to see a
908 value (hashed pointer) instead. Cryptographically
909 insecure, please do not use on production kernels.
912 [KNL] verbose locking self-tests
914 Print debugging info while doing the locking API
916 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
917 (no extra messages), setting it to -1 (all bits set)
918 will print _a_lot_ more information - normally only
919 useful to lockdep developers.
921 debug_objects [KNL] Enable object debugging
924 [KNL] Disable object debugging
926 debug_guardpage_minorder=
927 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
928 parameter allows control of the order of pages that will
929 be intentionally kept free (and hence protected) by the
930 buddy allocator. Bigger value increase the probability
931 of catching random memory corruption, but reduce the
932 amount of memory for normal system use. The maximum
933 possible value is MAX_ORDER/2. Setting this parameter
934 to 1 or 2 should be enough to identify most random
935 memory corruption problems caused by bugs in kernel or
936 driver code when a CPU writes to (or reads from) a
937 random memory location. Note that there exists a class
938 of memory corruptions problems caused by buggy H/W or
939 F/W or by drivers badly programing DMA (basically when
940 memory is written at bus level and the CPU MMU is
941 bypassed) which are not detectable by
942 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
943 tracking down these problems.
946 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
947 enables the feature at boot time. By default, it is
948 disabled and the system will work mostly the same as a
949 kernel built without CONFIG_DEBUG_PAGEALLOC.
950 Note: to get most of debug_pagealloc error reports, it's
951 useful to also enable the page_owner functionality.
952 on: enable the feature
954 debugfs= [KNL] This parameter enables what is exposed to userspace
955 and debugfs internal clients.
956 Format: { on, no-mount, off }
957 on: All functions are enabled.
959 Filesystem is not registered but kernel clients can
960 access APIs and a crashkernel can be used to read
961 its content. There is nothing to mount.
962 off: Filesystem is not registered and clients
963 get a -EPERM as result when trying to register files
964 or directories within debugfs.
965 This is equivalent of the runtime functionality if
966 debugfs was not enabled in the kernel at all.
967 Default value is set in build-time with a kernel configuration.
969 debugpat [X86] Enable PAT debugging
972 [HW] The size of the default HugeTLB page. This is
973 the size represented by the legacy /proc/ hugepages
974 APIs. In addition, this is the default hugetlb size
975 used for shmget(), mmap() and mounting hugetlbfs
976 filesystems. If not specified, defaults to the
977 architecture's default huge page size. Huge page
978 sizes are architecture dependent. See also
979 Documentation/admin-guide/mm/hugetlbpage.rst.
982 deferred_probe_timeout=
983 [KNL] Debugging option to set a timeout in seconds for
984 deferred probe to give up waiting on dependencies to
985 probe. Only specific dependencies (subsystems or
986 drivers) that have opted in will be ignored. A timeout
987 of 0 will timeout at the end of initcalls. If the time
988 out hasn't expired, it'll be restarted by each
989 successful driver registration. This option will also
990 dump out devices still on the deferred probe list after
993 delayacct [KNL] Enable per-task delay accounting
995 dell_smm_hwmon.ignore_dmi=
996 [HW] Continue probing hardware even if DMI data
997 indicates that the driver is running on unsupported
1000 dell_smm_hwmon.force=
1001 [HW] Activate driver even if SMM BIOS signature does
1002 not match list of supported models and enable otherwise
1003 blacklisted features.
1005 dell_smm_hwmon.power_status=
1006 [HW] Report power status in /proc/i8k
1007 (disabled by default).
1009 dell_smm_hwmon.restricted=
1010 [HW] Allow controlling fans only if SYS_ADMIN
1013 dell_smm_hwmon.fan_mult=
1014 [HW] Factor to multiply fan speed with.
1016 dell_smm_hwmon.fan_max=
1017 [HW] Maximum configurable fan speed.
1020 Format: { on | off | def_only | inf_only | always }
1021 on: s390 zlib hardware support for compression on
1022 level 1 and decompression (default)
1023 off: No s390 zlib hardware support
1024 def_only: s390 zlib hardware support for deflate
1025 only (compression on level 1)
1026 inf_only: s390 zlib hardware support for inflate
1027 only (decompression)
1028 always: Same as 'on' but ignores the selected compression
1029 level always using hardware support (used for debugging)
1031 dhash_entries= [KNL]
1032 Set number of hash buckets for dentry cache.
1034 disable_1tb_segments [PPC]
1035 Disables the use of 1TB hash page table segments. This
1036 causes the kernel to fall back to 256MB segments which
1037 can be useful when debugging issues that require an SLB
1041 Limits the number of kernel SLB entries, and flushes
1042 them frequently to increase the rate of SLB faults
1043 on kernel addresses.
1046 See Documentation/networking/ipv6.rst.
1049 Disable RADIX MMU mode on POWER9
1051 radix_hcall_invalidate=on [PPC/PSERIES]
1052 Disable RADIX GTSE feature and use hcall for TLB
1056 Disable TLBIE instruction. Currently does not work
1057 with KVM, with HASH MMU, or with coherent accelerators.
1059 disable_cpu_apicid= [X86,APIC,SMP]
1061 The number of initial APIC ID for the
1062 corresponding CPU to be disabled at boot,
1063 mostly used for the kdump 2nd kernel to
1064 disable BSP to wake up multiple CPUs without
1065 causing system reset or hang due to sending
1066 INIT from AP to BSP.
1068 disable_ddw [PPC/PSERIES]
1069 Disable Dynamic DMA Window support. Use this
1070 to workaround buggy firmware.
1072 disable_ipv6= [IPV6]
1073 See Documentation/networking/ipv6.rst.
1075 disable_mtrr_cleanup [X86]
1076 The kernel tries to adjust MTRR layout from continuous
1077 to discrete, to make X server driver able to add WB
1078 entry later. This parameter disables that.
1080 disable_mtrr_trim [X86, Intel and AMD only]
1081 By default the kernel will trim any uncacheable
1082 memory out of your available memory pool based on
1083 MTRR settings. This parameter disables that behavior,
1084 possibly causing your machine to run very slowly.
1086 disable_timer_pin_1 [X86]
1087 Disable PIN 1 of APIC timer
1088 Can be useful to work around chipset bugs.
1090 dis_ucode_ldr [X86] Disable the microcode loader.
1092 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1093 this option disables the debugging code at boot.
1095 dma_debug_entries=<number>
1096 This option allows to tune the number of preallocated
1097 entries for DMA-API debugging code. One entry is
1098 required per DMA-API allocation. Use this if the
1099 DMA-API debugging code disables itself because the
1100 architectural default is too low.
1102 dma_debug_driver=<driver_name>
1103 With this option the DMA-API debugging driver
1104 filter feature can be enabled at boot time. Just
1105 pass the driver to filter for as the parameter.
1106 The filter can be disabled or changed to another
1107 driver later using sysfs.
1109 driver_async_probe= [KNL]
1110 List of driver names to be probed asynchronously. *
1111 matches with all driver names. If * is specified, the
1112 rest of the listed driver names are those that will NOT
1114 Format: <driver_name1>,<driver_name2>...
1116 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1117 Broken monitors, graphic adapters, KVMs and EDIDless
1118 panels may send no or incorrect EDID data sets.
1119 This parameter allows to specify an EDID data sets
1120 in the /lib/firmware directory that are used instead.
1121 Generic built-in EDID data sets are used, if one of
1122 edid/1024x768.bin, edid/1280x1024.bin,
1123 edid/1680x1050.bin, or edid/1920x1080.bin is given
1124 and no file with the same name exists. Details and
1125 instructions how to build your own EDID data are
1126 available in Documentation/admin-guide/edid.rst. An EDID
1127 data set will only be used for a particular connector,
1128 if its name and a colon are prepended to the EDID
1129 name. Each connector may use a unique EDID data
1130 set by separating the files with a comma. An EDID
1131 data set with no connector name will be used for
1132 any connectors not explicitly specified.
1137 Format: {"off" | "known"}
1138 Control how the dt_cpu_ftrs device-tree binding is
1139 used for CPU feature discovery and setup (if it
1141 off: Do not use it, fall back to legacy cpu table.
1142 known: Do not pass through unknown features to guests
1143 or userspace, only those that the kernel is aware of.
1145 dump_apple_properties [X86]
1146 Dump name and content of EFI device properties on
1147 x86 Macs. Useful for driver authors to determine
1148 what data is available or for reverse-engineering.
1150 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1151 <module>.dyndbg[="val"]
1152 Enable debug messages at boot time. See
1153 Documentation/admin-guide/dynamic-debug-howto.rst
1156 nopku [X86] Disable Memory Protection Keys CPU feature found
1159 <module>.async_probe[=<bool>] [KNL]
1160 If no <bool> value is specified or if the value
1161 specified is not a valid <bool>, enable asynchronous
1162 probe on this module. Otherwise, enable/disable
1163 asynchronous probe on this module as indicated by the
1164 <bool> value. See also: module.async_probe
1166 early_ioremap_debug [KNL]
1167 Enable debug messages in early_ioremap support. This
1168 is useful for tracking down temporary early mappings
1169 which are not unmapped.
1171 earlycon= [KNL] Output early console device and options.
1173 When used with no options, the early console is
1174 determined by stdout-path property in device tree's
1175 chosen node or the ACPI SPCR table if supported by
1178 cdns,<addr>[,options]
1179 Start an early, polled-mode console on a Cadence
1180 (xuartps) serial port at the specified address. Only
1181 supported option is baud rate. If baud rate is not
1182 specified, the serial port must already be setup and
1185 uart[8250],io,<addr>[,options]
1186 uart[8250],mmio,<addr>[,options]
1187 uart[8250],mmio32,<addr>[,options]
1188 uart[8250],mmio32be,<addr>[,options]
1189 uart[8250],0x<addr>[,options]
1190 Start an early, polled-mode console on the 8250/16550
1191 UART at the specified I/O port or MMIO address.
1192 MMIO inter-register address stride is either 8-bit
1193 (mmio) or 32-bit (mmio32 or mmio32be).
1194 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1195 to be equivalent to 'mmio'. 'options' are specified
1196 in the same format described for "console=ttyS<n>"; if
1197 unspecified, the h/w is not initialized.
1201 Start an early, polled-mode console on a pl011 serial
1202 port at the specified address. The pl011 serial port
1203 must already be setup and configured. Options are not
1204 yet supported. If 'mmio32' is specified, then only
1205 the driver will use only 32-bit accessors to read/write
1206 the device registers.
1209 Start an early console on a litex serial port at the
1210 specified address. The serial port must already be
1211 setup and configured. Options are not yet supported.
1214 Start an early, polled-mode console on a meson serial
1215 port at the specified address. The serial port must
1216 already be setup and configured. Options are not yet
1220 Start an early, polled-mode console on an msm serial
1221 port at the specified address. The serial port
1222 must already be setup and configured. Options are not
1225 msm_serial_dm,<addr>
1226 Start an early, polled-mode console on an msm serial
1227 dm port at the specified address. The serial port
1228 must already be setup and configured. Options are not
1232 Start an early, polled-mode console on a serial port
1233 of an Actions Semi SoC, such as S500 or S900, at the
1234 specified address. The serial port must already be
1235 setup and configured. Options are not yet supported.
1238 Start an early, polled-mode console on a serial port
1239 of an RDA Micro SoC, such as RDA8810PL, at the
1240 specified address. The serial port must already be
1241 setup and configured. Options are not yet supported.
1244 Use RISC-V SBI (Supervisor Binary Interface) for early
1247 smh Use ARM semihosting calls for early console.
1255 Use early console provided by serial driver available
1256 on Samsung SoCs, requires selecting proper type and
1257 a correct base address of the selected UART port. The
1258 serial port must already be setup and configured.
1259 Options are not yet supported.
1262 Start an early, polled-mode console on a lantiq serial
1263 (lqasc) port at the specified address. The serial port
1264 must already be setup and configured. Options are not
1269 Use early console provided by Freescale LP UART driver
1270 found on Freescale Vybrid and QorIQ LS1021A processors.
1271 A valid base address must be provided, and the serial
1272 port must already be setup and configured.
1276 Start an early, polled-mode, output-only console on the
1277 Freescale i.MX UART at the specified address. The UART
1278 must already be setup and configured.
1281 Start an early, polled-mode console on the
1282 Armada 3700 serial port at the specified
1283 address. The serial port must already be setup
1284 and configured. Options are not yet supported.
1287 Start an early, polled-mode console on a Qualcomm
1288 Generic Interface (GENI) based serial port at the
1289 specified address. The serial port must already be
1290 setup and configured. Options are not yet supported.
1293 Start an early, unaccelerated console on the EFI
1294 memory mapped framebuffer (if available). On cache
1295 coherent non-x86 systems that use system memory for
1296 the framebuffer, pass the 'ram' option so that it is
1297 mapped with the correct attributes.
1300 Use early console provided by Freescale LINFlexD UART
1301 serial driver for NXP S32V234 SoCs. A valid base
1302 address must be provided, and the serial port must
1303 already be setup and configured.
1305 earlyprintk= [X86,SH,ARM,M68k,S390]
1309 earlyprintk=serial[,ttySn[,baudrate]]
1310 earlyprintk=serial[,0x...[,baudrate]]
1311 earlyprintk=ttySn[,baudrate]
1312 earlyprintk=dbgp[debugController#]
1313 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1314 earlyprintk=xdbc[xhciController#]
1316 earlyprintk is useful when the kernel crashes before
1317 the normal console is initialized. It is not enabled by
1318 default because it has some cosmetic problems.
1320 Append ",keep" to not disable it when the real console
1323 Only one of vga, serial, or usb debug port can
1326 Currently only ttyS0 and ttyS1 may be specified by
1327 name. Other I/O ports may be explicitly specified
1328 on some architectures (x86 and arm at least) by
1329 replacing ttySn with an I/O port address, like this:
1330 earlyprintk=serial,0x1008,115200
1331 You can find the port for a given device in
1332 /proc/tty/driver/serial:
1333 2: uart:ST16650V2 port:00001008 irq:18 ...
1335 Interaction with the standard serial driver is not
1338 The VGA output is eventually overwritten by
1341 The xen option can only be used in Xen domains.
1343 The sclp output can only be used on s390.
1345 The optional "force" to "pciserial" enables use of a
1346 PCI device even when its classcode is not of the
1349 edac_report= [HW,EDAC] Control how to report EDAC event
1350 Format: {"on" | "off" | "force"}
1351 on: enable EDAC to report H/W event. May be overridden
1352 by other higher priority error reporting module.
1353 off: disable H/W event reporting through EDAC.
1354 force: enforce the use of EDAC to report H/W event.
1358 Format: {"off" | "on" | "skip[mbr]"}
1361 Format: { "debug", "disable_early_pci_dma",
1362 "nochunk", "noruntime", "nosoftreserve",
1363 "novamap", "no_disable_early_pci_dma" }
1364 debug: enable misc debug output.
1365 disable_early_pci_dma: disable the busmaster bit on all
1366 PCI bridges while in the EFI boot stub.
1367 nochunk: disable reading files in "chunks" in the EFI
1368 boot stub, as chunking can cause problems with some
1369 firmware implementations.
1370 noruntime : disable EFI runtime services support
1371 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1372 attribute may cause the kernel to reserve the
1373 memory range for a memory mapping driver to
1374 claim. Specify efi=nosoftreserve to disable this
1375 reservation and treat the memory by its base type
1376 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1377 novamap: do not call SetVirtualAddressMap().
1378 no_disable_early_pci_dma: Leave the busmaster bit set
1379 on all PCI bridges while in the EFI boot stub
1381 efi_no_storage_paranoia [EFI; X86]
1382 Using this parameter you can use more than 50% of
1383 your efi variable storage. Use this parameter only if
1384 you are really sure that your UEFI does sane gc and
1385 fulfills the spec otherwise your board may brick.
1387 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1388 Add arbitrary attribute to specific memory range by
1389 updating original EFI memory map.
1390 Region of memory which aa attribute is added to is
1393 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1394 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1395 attribute is added to range 0x100000000-0x180000000 and
1396 0x10a0000000-0x1120000000.
1398 If efi_fake_mem=8G@9G:0x40000 is specified, the
1399 EFI_MEMORY_SP(0x40000) attribute is added to
1400 range 0x240000000-0x43fffffff.
1402 Using this parameter you can do debugging of EFI memmap
1403 related features. For example, you can do debugging of
1404 Address Range Mirroring feature even if your box
1405 doesn't support it, or mark specific memory as
1408 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1409 that is to be dynamically loaded by Linux. If there are
1410 multiple variables with the same name but with different
1411 vendor GUIDs, all of them will be loaded. See
1412 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1415 eisa_irq_edge= [PARISC,HW]
1416 See header of drivers/parisc/eisa.c.
1418 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1421 This is designed to be used in conjunction with
1422 the boot argument: earlyprintk=vga
1424 This parameter works in place of the kgdboc parameter
1425 but can only be used if the backing tty is available
1426 very early in the boot process. For early debugging
1427 via a serial port see kgdboc_earlycon instead.
1430 See comment before function elanfreq_setup() in
1431 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1433 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1434 Specifies physical address of start of kernel core
1435 image elf header and optionally the size. Generally
1436 kexec loader will pass this option to capture kernel.
1437 See Documentation/admin-guide/kdump/kdump.rst for details.
1439 enable_mtrr_cleanup [X86]
1440 The kernel tries to adjust MTRR layout from continuous
1441 to discrete, to make X server driver able to add WB
1442 entry later. This parameter enables that.
1444 enable_timer_pin_1 [X86]
1445 Enable PIN 1 of APIC timer
1446 Can be useful to work around chipset bugs
1447 (in particular on some ATI chipsets).
1448 The kernel tries to set a reasonable default.
1450 enforcing= [SELINUX] Set initial enforcing status.
1452 See security/selinux/Kconfig help text.
1453 0 -- permissive (log only, no denials).
1454 1 -- enforcing (deny and log).
1456 Value can be changed at runtime via
1457 /sys/fs/selinux/enforce.
1460 Disable Error Record Serialization Table (ERST)
1463 ether= [HW,NET] Ethernet cards parameters
1464 This option is obsoleted by the "netdev=" option, which
1465 has equivalent usage. See its documentation for details.
1469 Permit 'security.evm' to be updated regardless of
1470 current integrity status.
1472 early_page_ext [KNL] Enforces page_ext initialization to earlier
1473 stages so cover more early boot allocations.
1474 Please note that as side effect some optimizations
1475 might be disabled to achieve that (e.g. parallelized
1476 memory initialization is disabled) so the boot process
1477 might take longer, especially on systems with a lot of
1478 memory. Available with CONFIG_PAGE_EXTENSION=y.
1483 fail_make_request=[KNL]
1484 General fault injection mechanism.
1485 Format: <interval>,<probability>,<space>,<times>
1486 See also Documentation/fault-injection/.
1489 Format: { initns | none }
1490 See Documentation/admin-guide/sysctl/net.rst for
1491 fb_tunnels_only_for_init_ns
1494 See Documentation/admin-guide/blockdev/floppy.rst.
1496 force_pal_cache_flush
1497 [IA-64] Avoid check_sal_cache_flush which may hang on
1498 buggy SAL_CACHE_FLUSH implementations. Using this
1499 parameter will force ia64_sal_cache_flush to call
1500 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1503 Forcefully enable Physical Address Extension (PAE).
1504 Many Pentium M systems disable PAE but may have a
1505 functionally usable PAE implementation.
1506 Warning: use of this parameter will taint the kernel
1507 and may cause unknown problems.
1510 [FTRACE] will set and start the specified tracer
1511 as early as possible in order to facilitate early
1514 ftrace_boot_snapshot
1515 [FTRACE] On boot up, a snapshot will be taken of the
1516 ftrace ring buffer that can be read at:
1517 /sys/kernel/tracing/snapshot.
1518 This is useful if you need tracing information from kernel
1519 boot up that is likely to be overridden by user space
1520 start up functionality.
1522 ftrace_dump_on_oops[=orig_cpu]
1523 [FTRACE] will dump the trace buffers on oops.
1524 If no parameter is passed, ftrace will dump
1525 buffers of all CPUs, but if you pass orig_cpu, it will
1526 dump only the buffer of the CPU that triggered the
1529 ftrace_filter=[function-list]
1530 [FTRACE] Limit the functions traced by the function
1531 tracer at boot up. function-list is a comma-separated
1532 list of functions. This list can be changed at run
1533 time by the set_ftrace_filter file in the debugfs
1536 ftrace_notrace=[function-list]
1537 [FTRACE] Do not trace the functions specified in
1538 function-list. This list can be changed at run time
1539 by the set_ftrace_notrace file in the debugfs
1542 ftrace_graph_filter=[function-list]
1543 [FTRACE] Limit the top level callers functions traced
1544 by the function graph tracer at boot up.
1545 function-list is a comma-separated list of functions
1546 that can be changed at run time by the
1547 set_graph_function file in the debugfs tracing directory.
1549 ftrace_graph_notrace=[function-list]
1550 [FTRACE] Do not trace from the functions specified in
1551 function-list. This list is a comma-separated list of
1552 functions that can be changed at run time by the
1553 set_graph_notrace file in the debugfs tracing directory.
1555 ftrace_graph_max_depth=<uint>
1556 [FTRACE] Used with the function graph tracer. This is
1557 the max depth it will trace into a function. This value
1558 can be changed at run time by the max_graph_depth file
1559 in the tracefs tracing directory. default: 0 (no limit)
1561 fw_devlink= [KNL] Create device links between consumer and supplier
1562 devices by scanning the firmware to infer the
1563 consumer/supplier relationships. This feature is
1564 especially useful when drivers are loaded as modules as
1565 it ensures proper ordering of tasks like device probing
1566 (suppliers first, then consumers), supplier boot state
1567 clean up (only after all consumers have probed),
1568 suspend/resume & runtime PM (consumers first, then
1570 Format: { off | permissive | on | rpm }
1571 off -- Don't create device links from firmware info.
1572 permissive -- Create device links from firmware info
1573 but use it only for ordering boot state clean
1574 up (sync_state() calls).
1575 on -- Create device links from firmware info and use it
1576 to enforce probe and suspend/resume ordering.
1577 rpm -- Like "on", but also use to order runtime PM.
1579 fw_devlink.strict=<bool>
1580 [KNL] Treat all inferred dependencies as mandatory
1581 dependencies. This only applies for fw_devlink=on|rpm.
1585 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1586 support via parallel port (up to 5 devices per port)
1587 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1588 See also Documentation/input/devices/joystick-parport.rst
1592 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1596 gcov_persist= [GCOV] When non-zero (default), profiling data for
1597 kernel modules is saved and remains accessible via
1598 debugfs, even when the module is unloaded/reloaded.
1599 When zero, profiling data is discarded and associated
1600 debugfs files are removed at module unload time.
1602 goldfish [X86] Enable the goldfish android emulator platform.
1603 Don't use this when you are not running on the
1606 gpio-mockup.gpio_mockup_ranges
1607 [HW] Sets the ranges of gpiochip of for this device.
1608 Format: <start1>,<end1>,<start2>,<end2>...
1609 gpio-mockup.gpio_mockup_named_lines
1610 [HW] Let the driver know GPIO lines should be named.
1612 gpt [EFI] Forces disk with valid GPT signature but
1613 invalid Protective MBR to be treated as GPT. If the
1614 primary GPT is corrupted, it enables the backup/alternate
1615 GPT to be used instead.
1617 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1618 the "Enable 0" bit of the configuration register.
1621 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1622 the "Enable 0" bit of the configuration register.
1625 grcan.select= [HW] Select which physical interface to use.
1628 grcan.txsize= [HW] Sets the size of the tx buffer.
1629 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1631 grcan.rxsize= [HW] Sets the size of the rx buffer.
1632 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1636 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1637 hardening is enabled for this boot. Hardened
1638 usercopy checking is used to protect the kernel
1639 from reading or writing beyond known memory
1640 allocation boundaries as a proactive defense
1641 against bounds-checking flaws in the kernel's
1642 copy_to_user()/copy_from_user() interface.
1643 on Perform hardened usercopy checks (default).
1644 off Disable hardened usercopy checks.
1646 hardlockup_all_cpu_backtrace=
1647 [KNL] Should the hard-lockup detector generate
1648 backtraces on all cpus.
1651 hashdist= [KNL,NUMA] Large hashes allocated during boot
1652 are distributed across NUMA nodes. Defaults on
1653 for 64-bit NUMA, off otherwise.
1654 Format: 0 | 1 (for off | on)
1656 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1658 hd= [EIDE] (E)IDE hard drive subsystem geometry
1659 Format: <cyl>,<head>,<sect>
1662 Disable Hardware Error Source Table (HEST) support;
1663 corresponding firmware-first mode error processing
1664 logic will be disabled.
1666 hibernate= [HIBERNATION]
1667 noresume Don't check if there's a hibernation image
1668 present during boot.
1669 nocompress Don't compress/decompress hibernation images.
1670 no Disable hibernation and resume.
1671 protect_image Turn on image protection during restoration
1672 (that will set all pages holding image data
1673 during restoration read-only).
1675 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1676 size of <nn>. This works even on boxes that have no
1677 highmem otherwise. This also works to reduce highmem
1678 size on bigger boxes.
1680 highres= [KNL] Enable/disable high resolution timer mode.
1681 Valid parameters: "on", "off"
1686 hostname= [KNL] Set the hostname (aka UTS nodename).
1688 This allows setting the system's hostname during early
1689 startup. This sets the name returned by gethostname.
1690 Using this parameter to set the hostname makes it
1691 possible to ensure the hostname is correctly set before
1692 any userspace processes run, avoiding the possibility
1693 that a process may call gethostname before the hostname
1694 has been explicitly set, resulting in the calling
1695 process getting an incorrect result. The string must
1696 not exceed the maximum allowed hostname length (usually
1697 64 characters) and will be truncated otherwise.
1699 hpet= [X86-32,HPET] option to control HPET usage
1700 Format: { enable (default) | disable | force |
1702 disable: disable HPET and use PIT instead
1703 force: allow force enabled of undocumented chips (ICH4,
1705 verbose: show contents of HPET registers during setup
1707 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1708 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1710 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1711 If this follows hugepagesz (below), it specifies
1712 the number of pages of hugepagesz to be allocated.
1713 If this is the first HugeTLB parameter on the command
1714 line, it specifies the number of pages to allocate for
1715 the default huge page size. If using node format, the
1716 number of pages to allocate per-node can be specified.
1717 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1718 Format: <integer> or (node format)
1719 <node>:<integer>[,<node>:<integer>]
1722 [HW] The size of the HugeTLB pages. This is used in
1723 conjunction with hugepages (above) to allocate huge
1724 pages of a specific size at boot. The pair
1725 hugepagesz=X hugepages=Y can be specified once for
1726 each supported huge page size. Huge page sizes are
1727 architecture dependent. See also
1728 Documentation/admin-guide/mm/hugetlbpage.rst.
1731 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1732 of gigantic hugepages. Or using node format, the size
1733 of a CMA area per node can be specified.
1734 Format: nn[KMGTPE] or (node format)
1735 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1737 Reserve a CMA area of given size and allocate gigantic
1738 hugepages using the CMA allocator. If enabled, the
1739 boot-time allocation of gigantic hugepages is skipped.
1741 hugetlb_free_vmemmap=
1742 [KNL] Reguires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1744 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1745 Allows heavy hugetlb users to free up some more
1746 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1747 Format: { on | off (default) }
1752 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1755 Note that the vmemmap pages may be allocated from the added
1756 memory block itself when memory_hotplug.memmap_on_memory is
1757 enabled, those vmemmap pages cannot be optimized even if this
1758 feature is enabled. Other vmemmap pages not allocated from
1759 the added memory block itself do not be affected.
1762 [KNL] Should the hung task detector generate panics.
1765 A value of 1 instructs the kernel to panic when a
1766 hung task is detected. The default value is controlled
1767 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1768 option. The value selected by this boot parameter can
1769 be changed later by the kernel.hung_task_panic sysctl.
1771 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1772 terminal devices. Valid values: 0..8
1773 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1774 If specified, z/VM IUCV HVC accepts connections
1775 from listed z/VM user IDs only.
1777 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1778 which allow the hypervisor to 'idle' the
1779 guest on lock contention.
1782 Do not unregister boot console at start. This is only
1783 useful for debugging when something happens in the window
1784 between unregistering the boot console and initializing
1787 i2c_bus= [HW] Override the default board specific I2C bus speed
1788 or register an additional I2C bus that is not
1789 registered from board initialization code.
1793 i8042.debug [HW] Toggle i8042 debug mode
1794 i8042.unmask_kbd_data
1795 [HW] Enable printing of interrupt data from the KBD port
1796 (disabled by default, and as a pre-condition
1797 requires that i8042.debug=1 be enabled)
1798 i8042.direct [HW] Put keyboard port into non-translated mode
1799 i8042.dumbkbd [HW] Pretend that controller can only read data from
1800 keyboard and cannot control its state
1801 (Don't attempt to blink the leds)
1802 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1803 i8042.nokbd [HW] Don't check/create keyboard port
1804 i8042.noloop [HW] Disable the AUX Loopback command while probing
1806 i8042.nomux [HW] Don't check presence of an active multiplexing
1808 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1810 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1811 i8042.reset [HW] Reset the controller during init, cleanup and
1812 suspend-to-ram transitions, only during s2r
1813 transitions, or never reset
1814 Format: { 1 | Y | y | 0 | N | n }
1815 1, Y, y: always reset controller
1816 0, N, n: don't ever reset controller
1817 Default: only on s2r transitions on x86; most other
1818 architectures force reset to be always executed
1819 i8042.unlock [HW] Unlock (ignore) the keylock
1820 i8042.kbdreset [HW] Reset device connected to KBD port
1822 [HW] Allow deferred probing upon i8042 probe errors
1826 i915.invert_brightness=
1827 [DRM] Invert the sense of the variable that is used to
1828 set the brightness of the panel backlight. Normally a
1829 brightness value of 0 indicates backlight switched off,
1830 and the maximum of the brightness value sets the backlight
1831 to maximum brightness. If this parameter is set to 0
1832 (default) and the machine requires it, or this parameter
1833 is set to 1, a brightness value of 0 sets the backlight
1834 to maximum brightness, and the maximum of the brightness
1835 value switches the backlight off.
1836 -1 -- never invert brightness
1837 0 -- machine default
1838 1 -- force brightness inversion
1841 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1845 Format: idle=poll, idle=halt, idle=nomwait
1846 Poll forces a polling idle loop that can slightly
1847 improve the performance of waking up a idle CPU, but
1848 will use a lot of power and make the system run hot.
1850 idle=halt: Halt is forced to be used for CPU idle.
1851 In such case C2/C3 won't be used again.
1852 idle=nomwait: Disable mwait for CPU C-states
1856 Allow force disabling of Shared Virtual Memory (SVA)
1857 support for the idxd driver. By default it is set to
1860 idxd.tc_override= [HW]
1862 Allow override of default traffic class configuration
1863 for the device. By default it is set to false (0).
1865 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1866 Format: { strict | legacy | 2008 | relaxed }
1869 Choose which programs will be accepted for execution
1870 based on the IEEE 754 NaN encoding(s) supported by
1871 the FPU and the NaN encoding requested with the value
1872 of an ELF file header flag individually set by each
1873 binary. Hardware implementations are permitted to
1874 support either or both of the legacy and the 2008 NaN
1877 Available settings are as follows:
1878 strict accept binaries that request a NaN encoding
1879 supported by the FPU
1880 legacy only accept legacy-NaN binaries, if supported
1882 2008 only accept 2008-NaN binaries, if supported
1884 relaxed accept any binaries regardless of whether
1885 supported by the FPU
1887 The FPU emulator is always able to support both NaN
1888 encodings, so if no FPU hardware is present or it has
1889 been disabled with 'nofpu', then the settings of
1890 'legacy' and '2008' strap the emulator accordingly,
1891 'relaxed' straps the emulator for both legacy-NaN and
1892 2008-NaN, whereas 'strict' enables legacy-NaN only on
1893 legacy processors and both NaN encodings on MIPS32 or
1896 The setting for ABS.fmt/NEG.fmt instruction execution
1897 mode generally follows that for the NaN encoding,
1898 except where unsupported by hardware.
1900 ignore_loglevel [KNL]
1901 Ignore loglevel setting - this will print /all/
1902 kernel messages to the console. Useful for debugging.
1903 We also add it as printk module parameter, so users
1904 could change it dynamically, usually by
1905 /sys/module/printk/parameters/ignore_loglevel.
1908 Ignore RLIMIT_DATA setting for data mappings,
1909 print warning at first misuse. Can be changed via
1910 /sys/module/kernel/parameters/ignore_rlimit_data.
1912 ihash_entries= [KNL]
1913 Set number of hash buckets for inode cache.
1915 ima_appraise= [IMA] appraise integrity measurements
1916 Format: { "off" | "enforce" | "fix" | "log" }
1919 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1920 The builtin appraise policy appraises all files
1923 ima_canonical_fmt [IMA]
1924 Use the canonical format for the binary runtime
1925 measurements, instead of host native format.
1928 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1932 The list of supported hash algorithms is defined
1933 in crypto/hash_info.h.
1936 The builtin policies to load during IMA setup.
1937 Format: "tcb | appraise_tcb | secure_boot |
1938 fail_securely | critical_data"
1940 The "tcb" policy measures all programs exec'd, files
1941 mmap'd for exec, and all files opened with the read
1942 mode bit set by either the effective uid (euid=0) or
1945 The "appraise_tcb" policy appraises the integrity of
1946 all files owned by root.
1948 The "secure_boot" policy appraises the integrity
1949 of files (eg. kexec kernel image, kernel modules,
1950 firmware, policy, etc) based on file signatures.
1952 The "fail_securely" policy forces file signature
1953 verification failure also on privileged mounted
1954 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1957 The "critical_data" policy measures kernel integrity
1960 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1961 Load a policy which meets the needs of the Trusted
1962 Computing Base. This means IMA will measure all
1963 programs exec'd, files mmap'd for exec, and all files
1964 opened for read by uid=0.
1967 Select one of defined IMA measurements template formats.
1968 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
1973 [IMA] Define a custom template format.
1974 Format: { "field1|...|fieldN" }
1976 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1977 Format: <min_file_size>
1978 Set the minimal file size for using asynchronous hash.
1979 If left unspecified, ahash usage is disabled.
1981 ahash performance varies for different data sizes on
1982 different crypto accelerators. This option can be used
1983 to achieve the best performance for a particular HW.
1985 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1987 Set hashing buffer size. Default: 4k.
1989 ahash performance varies for different chunk sizes on
1990 different crypto accelerators. This option can be used
1991 to achieve best performance for particular HW.
1995 Run specified binary instead of /sbin/init as init
1998 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1999 for working out where the kernel is dying during
2002 initcall_blacklist= [KNL] Do not execute a comma-separated list of
2003 initcall functions. Useful for debugging built-in
2004 modules and initcalls.
2006 initramfs_async= [KNL]
2009 This parameter controls whether the initramfs
2010 image is unpacked asynchronously, concurrently
2011 with devices being probed and
2012 initialized. This should normally just work,
2013 but as a debugging aid, one can get the
2014 historical behaviour of the initramfs
2015 unpacking being completed before device_ and
2018 initrd= [BOOT] Specify the location of the initial ramdisk
2020 initrdmem= [KNL] Specify a physical address and size from which to
2021 load the initrd. If an initrd is compiled in or
2022 specified in the bootparams, it takes priority over this
2024 Format: ss[KMG],nn[KMG]
2027 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
2030 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2032 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
2034 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2036 init_pkru= [X86] Specify the default memory protection keys rights
2037 register contents for all processes. 0x55555554 by
2038 default (disallow access to all but pkey 0). Can
2039 override in debugfs after boot.
2041 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2044 int_pln_enable [X86] Enable power limit notification interrupt
2046 integrity_audit=[IMA]
2047 Format: { "0" | "1" }
2048 0 -- basic integrity auditing messages. (Default)
2049 1 -- additional integrity auditing messages.
2051 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2053 Enable intel iommu driver.
2055 Disable intel iommu driver.
2056 igfx_off [Default Off]
2057 By default, gfx is mapped as normal device. If a gfx
2058 device has a dedicated DMAR unit, the DMAR unit is
2059 bypassed by not enabling DMAR with this option. In
2060 this case, gfx device will use physical address for
2062 strict [Default Off]
2063 Deprecated, equivalent to iommu.strict=1.
2064 sp_off [Default Off]
2065 By default, super page will be supported if Intel IOMMU
2066 has the capability. With this option, super page will
2069 Enable the Intel IOMMU scalable mode if the hardware
2070 advertises that it has support for the scalable mode
2073 Disallow use of the Intel IOMMU scalable mode.
2074 tboot_noforce [Default Off]
2075 Do not force the Intel IOMMU enabled under tboot.
2076 By default, tboot will force Intel IOMMU on, which
2077 could harm performance of some high-throughput
2078 devices like 40GBit network cards, even if identity
2080 Note that using this option lowers the security
2081 provided by tboot because it makes the system
2082 vulnerable to DMA attacks.
2084 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2085 0 disables intel_idle and fall back on acpi_idle.
2086 1 to 9 specify maximum depth of C-state.
2090 Do not enable intel_pstate as the default
2091 scaling driver for the supported processors
2093 Use intel_pstate as a scaling driver, but configure it
2094 to work with generic cpufreq governors (instead of
2095 enabling its internal governor). This mode cannot be
2096 used along with the hardware-managed P-states (HWP)
2099 Enable intel_pstate on systems that prohibit it by default
2100 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2101 instead of acpi-cpufreq may disable platform features, such
2102 as thermal controls and power capping, that rely on ACPI
2103 P-States information being indicated to OSPM and therefore
2104 should be used with caution. This option does not work with
2105 processors that aren't supported by the intel_pstate driver
2106 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2108 Do not enable hardware P state control (HWP)
2111 Only load intel_pstate on systems which support
2112 hardware P state control (HWP) if available.
2114 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2115 Description Table, specifies preferred power management
2116 profile as "Enterprise Server" or "Performance Server",
2117 then this feature is turned on by default.
2119 Allow per-logical-CPU P-State performance control limits using
2120 cpufreq sysfs interface
2122 intremap= [X86-64, Intel-IOMMU]
2123 on enable Interrupt Remapping (default)
2124 off disable Interrupt Remapping
2125 nosid disable Source ID checking
2127 BIOS x2APIC opt-out request will be ignored
2128 nopost disable Interrupt Posting
2130 iomem= Disable strict checking of access to MMIO memory
2131 strict regions from userspace.
2146 nobypass [PPC/POWERNV]
2147 Disable IOMMU bypass, using IOMMU for PCI devices.
2149 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2150 Format: { "0" | "1" }
2151 0 - Try to allocate a 32-bit DMA address first, before
2152 falling back to the full range if needed.
2153 1 - Allocate directly from the full usable range,
2154 forcing Dual Address Cycle for PCI cards supporting
2155 greater than 32-bit addressing.
2157 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2158 Format: { "0" | "1" }
2160 Request that DMA unmap operations use deferred
2161 invalidation of hardware TLBs, for increased
2162 throughput at the cost of reduced device isolation.
2163 Will fall back to strict mode if not supported by
2164 the relevant IOMMU driver.
2166 DMA unmap operations invalidate IOMMU hardware TLBs
2168 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2169 Note: on x86, strict mode specified via one of the
2170 legacy driver-specific options takes precedence.
2173 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2174 Format: { "0" | "1" }
2175 0 - Use IOMMU translation for DMA.
2176 1 - Bypass the IOMMU for DMA.
2177 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2179 io7= [HW] IO7 for Marvel-based Alpha systems
2180 See comment before marvel_specify_io7 in
2181 arch/alpha/kernel/core_marvel.c.
2183 io_delay= [X86] I/O delay method
2185 Standard port 0x80 based delay
2187 Alternate port 0xed based delay (needed on some systems)
2189 Simple two microseconds delay
2194 See Documentation/admin-guide/nfs/nfsroot.rst.
2196 ipcmni_extend [KNL] Extend the maximum number of unique System V
2197 IPC identifiers from 32,768 to 16,777,216.
2199 irqaffinity= [SMP] Set the default irq affinity mask
2200 The argument is a cpu list, as described above.
2202 irqchip.gicv2_force_probe=
2205 Force the kernel to look for the second 4kB page
2206 of a GICv2 controller even if the memory range
2207 exposed by the device tree is too small.
2209 irqchip.gicv3_nolpi=
2211 Force the kernel to ignore the availability of
2212 LPIs (and by consequence ITSs). Intended for system
2213 that use the kernel as a bootloader, and thus want
2214 to let secondary kernels in charge of setting up
2217 irqchip.gicv3_pseudo_nmi= [ARM64]
2218 Enables support for pseudo-NMIs in the kernel. This
2219 requires the kernel to be built with
2220 CONFIG_ARM64_PSEUDO_NMI.
2223 When an interrupt is not handled search all handlers
2224 for it. Intended to get systems with badly broken
2228 When an interrupt is not handled search all handlers
2229 for it. Also check all handlers each timer
2230 interrupt. Intended to get systems with badly broken
2234 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2236 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2237 [Deprecated - use cpusets instead]
2238 Format: [flag-list,]<cpu-list>
2240 Specify one or more CPUs to isolate from disturbances
2241 specified in the flag list (default: domain):
2244 Disable the tick when a single task runs.
2246 A residual 1Hz tick is offloaded to workqueues, which you
2247 need to affine to housekeeping through the global
2248 workqueue's affinity configured via the
2249 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2250 by using the 'domain' flag described below.
2252 NOTE: by default the global workqueue runs on all CPUs,
2253 so to protect individual CPUs the 'cpumask' file has to
2254 be configured manually after bootup.
2257 Isolate from the general SMP balancing and scheduling
2258 algorithms. Note that performing domain isolation this way
2259 is irreversible: it's not possible to bring back a CPU to
2260 the domains once isolated through isolcpus. It's strongly
2261 advised to use cpusets instead to disable scheduler load
2262 balancing through the "cpuset.sched_load_balance" file.
2263 It offers a much more flexible interface where CPUs can
2264 move in and out of an isolated set anytime.
2266 You can move a process onto or off an "isolated" CPU via
2267 the CPU affinity syscalls or cpuset.
2268 <cpu number> begins at 0 and the maximum value is
2269 "number of CPUs in system - 1".
2273 Isolate from being targeted by managed interrupts
2274 which have an interrupt mask containing isolated
2275 CPUs. The affinity of managed interrupts is
2276 handled by the kernel and cannot be changed via
2277 the /proc/irq/* interfaces.
2279 This isolation is best effort and only effective
2280 if the automatically assigned interrupt mask of a
2281 device queue contains isolated and housekeeping
2282 CPUs. If housekeeping CPUs are online then such
2283 interrupts are directed to the housekeeping CPU
2284 so that IO submitted on the housekeeping CPU
2285 cannot disturb the isolated CPU.
2287 If a queue's affinity mask contains only isolated
2288 CPUs then this parameter has no effect on the
2289 interrupt routing decision, though interrupts are
2290 only delivered when tasks running on those
2291 isolated CPUs submit IO. IO submitted on
2292 housekeeping CPUs has no influence on those
2295 The format of <cpu-list> is described above.
2299 ivrs_ioapic [HW,X86-64]
2300 Provide an override to the IOAPIC-ID<->DEVICE-ID
2301 mapping provided in the IVRS ACPI table.
2302 By default, PCI segment is 0, and can be omitted.
2304 For example, to map IOAPIC-ID decimal 10 to
2305 PCI segment 0x1 and PCI device 00:14.0,
2306 write the parameter as:
2307 ivrs_ioapic=10@0001:00:14.0
2310 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2311 write the parameter as:
2312 ivrs_ioapic[10]=00:14.0
2313 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2314 PCI device 00:14.0 write the parameter as:
2315 ivrs_ioapic[10]=0001:00:14.0
2317 ivrs_hpet [HW,X86-64]
2318 Provide an override to the HPET-ID<->DEVICE-ID
2319 mapping provided in the IVRS ACPI table.
2320 By default, PCI segment is 0, and can be omitted.
2322 For example, to map HPET-ID decimal 10 to
2323 PCI segment 0x1 and PCI device 00:14.0,
2324 write the parameter as:
2325 ivrs_hpet=10@0001:00:14.0
2328 * To map HPET-ID decimal 0 to PCI device 00:14.0
2329 write the parameter as:
2330 ivrs_hpet[0]=00:14.0
2331 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2332 PCI device 00:14.0 write the parameter as:
2333 ivrs_ioapic[10]=0001:00:14.0
2335 ivrs_acpihid [HW,X86-64]
2336 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2337 mapping provided in the IVRS ACPI table.
2338 By default, PCI segment is 0, and can be omitted.
2340 For example, to map UART-HID:UID AMD0020:0 to
2341 PCI segment 0x1 and PCI device ID 00:14.5,
2342 write the parameter as:
2343 ivrs_acpihid=AMD0020:0@0001:00:14.5
2346 * To map UART-HID:UID AMD0020:0 to PCI segment is 0,
2347 PCI device ID 00:14.5, write the parameter as:
2348 ivrs_acpihid[00:14.5]=AMD0020:0
2349 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and
2350 PCI device ID 00:14.5, write the parameter as:
2351 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2353 js= [HW,JOY] Analog joystick
2354 See Documentation/input/joydev/joystick.rst.
2357 When CONFIG_RANDOMIZE_BASE is set, this disables
2358 kernel and module base offset ASLR (Address Space
2359 Layout Randomization).
2362 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2363 report on every invalid memory access. Without this
2364 parameter KASAN will print report only for the first
2369 kernelcore= [KNL,X86,IA-64,PPC]
2370 Format: nn[KMGTPE] | nn% | "mirror"
2371 This parameter specifies the amount of memory usable by
2372 the kernel for non-movable allocations. The requested
2373 amount is spread evenly throughout all nodes in the
2374 system as ZONE_NORMAL. The remaining memory is used for
2375 movable memory in its own zone, ZONE_MOVABLE. In the
2376 event, a node is too small to have both ZONE_NORMAL and
2377 ZONE_MOVABLE, kernelcore memory will take priority and
2378 other nodes will have a larger ZONE_MOVABLE.
2380 ZONE_MOVABLE is used for the allocation of pages that
2381 may be reclaimed or moved by the page migration
2382 subsystem. Note that allocations like PTEs-from-HighMem
2383 still use the HighMem zone if it exists, and the Normal
2384 zone if it does not.
2386 It is possible to specify the exact amount of memory in
2387 the form of "nn[KMGTPE]", a percentage of total system
2388 memory in the form of "nn%", or "mirror". If "mirror"
2389 option is specified, mirrored (reliable) memory is used
2390 for non-movable allocations and remaining memory is used
2391 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2392 are exclusive, so you cannot specify multiple forms.
2394 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2395 Format: <Controller#>[,poll interval]
2396 The controller # is the number of the ehci usb debug
2397 port as it is probed via PCI. The poll interval is
2398 optional and is the number seconds in between
2399 each poll cycle to the debug port in case you need
2400 the functionality for interrupting the kernel with
2401 gdb or control-c on the dbgp connection. When
2402 not using this parameter you use sysrq-g to break into
2403 the kernel debugger.
2405 kgdboc= [KGDB,HW] kgdb over consoles.
2406 Requires a tty driver that supports console polling,
2407 or a supported polling keyboard driver (non-usb).
2408 Serial only format: <serial_device>[,baud]
2409 keyboard only format: kbd
2410 keyboard and serial format: kbd,<serial_device>[,baud]
2411 Optional Kernel mode setting:
2412 kms, kbd format: kms,kbd
2413 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2415 kgdboc_earlycon= [KGDB,HW]
2416 If the boot console provides the ability to read
2417 characters and can work in polling mode, you can use
2418 this parameter to tell kgdb to use it as a backend
2419 until the normal console is registered. Intended to
2420 be used together with the kgdboc parameter which
2421 specifies the normal console to transition to.
2423 The name of the early console should be specified
2424 as the value of this parameter. Note that the name of
2425 the early console might be different than the tty
2426 name passed to kgdboc. It's OK to leave the value
2427 blank and the first boot console that implements
2428 read() will be picked.
2430 kgdbwait [KGDB] Stop kernel execution and enter the
2431 kernel debugger at the earliest opportunity.
2433 kmac= [MIPS] Korina ethernet MAC address.
2434 Configure the RouterBoard 532 series on-chip
2435 Ethernet adapter MAC address.
2437 kmemleak= [KNL] Boot-time kmemleak enable/disable
2438 Valid arguments: on, off
2440 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2443 kprobe_event=[probe-list]
2444 [FTRACE] Add kprobe events and enable at boot time.
2445 The probe-list is a semicolon delimited list of probe
2446 definitions. Each definition is same as kprobe_events
2447 interface, but the parameters are comma delimited.
2448 For example, to add a kprobe event on vfs_read with
2449 arg1 and arg2, add to the command line;
2451 kprobe_event=p,vfs_read,$arg1,$arg2
2453 See also Documentation/trace/kprobetrace.rst "Kernel
2454 Boot Parameter" section.
2456 kpti= [ARM64] Control page table isolation of user
2457 and kernel address spaces.
2458 Default: enabled on cores which need mitigation.
2462 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires
2463 CONFIG_KUNIT to be set to be fully enabled. The
2464 default value can be overridden via
2465 KUNIT_DEFAULT_ENABLED.
2466 Default is 1 (enabled)
2468 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2469 Default is 0 (don't ignore, but inject #GP)
2471 kvm.eager_page_split=
2472 [KVM,X86] Controls whether or not KVM will try to
2473 proactively split all huge pages during dirty logging.
2474 Eager page splitting reduces interruptions to vCPU
2475 execution by eliminating the write-protection faults
2476 and MMU lock contention that would otherwise be
2477 required to split huge pages lazily.
2479 VM workloads that rarely perform writes or that write
2480 only to a small region of VM memory may benefit from
2481 disabling eager page splitting to allow huge pages to
2482 still be used for reads.
2484 The behavior of eager page splitting depends on whether
2485 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2486 disabled, all huge pages in a memslot will be eagerly
2487 split when dirty logging is enabled on that memslot. If
2488 enabled, eager page splitting will be performed during
2489 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2492 Eager page splitting is only supported when kvm.tdp_mmu=Y.
2496 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2497 Default is false (don't support).
2500 [KVM] Controls the software workaround for the
2501 X86_BUG_ITLB_MULTIHIT bug.
2502 force : Always deploy workaround.
2503 off : Never deploy workaround.
2504 auto : Deploy workaround based on the presence of
2505 X86_BUG_ITLB_MULTIHIT.
2509 If the software workaround is enabled for the host,
2510 guests do need not to enable it for nested guests.
2512 kvm.nx_huge_pages_recovery_ratio=
2513 [KVM] Controls how many 4KiB pages are periodically zapped
2514 back to huge pages. 0 disables the recovery, otherwise if
2515 the value is N KVM will zap 1/Nth of the 4KiB pages every
2516 period (see below). The default is 60.
2518 kvm.nx_huge_pages_recovery_period_ms=
2519 [KVM] Controls the time period at which KVM zaps 4KiB pages
2520 back to huge pages. If the value is a non-zero N, KVM will
2521 zap a portion (see ratio above) of the pages every N msecs.
2522 If the value is 0 (the default), KVM will pick a period based
2523 on the ratio, such that a page is zapped after 1 hour on average.
2525 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2526 Default is 1 (enabled)
2528 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2530 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2533 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2535 none: Forcefully disable KVM.
2537 nvhe: Standard nVHE-based mode, without support for
2540 protected: nVHE-based mode with support for guests whose
2541 state is kept private from the host.
2543 Defaults to VHE/nVHE based on hardware support. Setting
2544 mode to "protected" will disable kexec and hibernation
2547 kvm-arm.vgic_v3_group0_trap=
2548 [KVM,ARM] Trap guest accesses to GICv3 group-0
2551 kvm-arm.vgic_v3_group1_trap=
2552 [KVM,ARM] Trap guest accesses to GICv3 group-1
2555 kvm-arm.vgic_v3_common_trap=
2556 [KVM,ARM] Trap guest accesses to GICv3 common
2559 kvm-arm.vgic_v4_enable=
2560 [KVM,ARM] Allow use of GICv4 for direct injection of
2563 kvm_cma_resv_ratio=n [PPC]
2564 Reserves given percentage from system memory area for
2565 contiguous memory allocation for KVM hash pagetable
2567 By default it reserves 5% of total system memory.
2571 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2572 (virtualized MMU) support on capable Intel chips.
2573 Default is 1 (enabled)
2575 kvm-intel.emulate_invalid_guest_state=
2576 [KVM,Intel] Disable emulation of invalid guest state.
2577 Ignored if kvm-intel.enable_unrestricted_guest=1, as
2578 guest state is never invalid for unrestricted guests.
2579 This param doesn't apply to nested guests (L2), as KVM
2580 never emulates invalid L2 guest state.
2581 Default is 1 (enabled)
2583 kvm-intel.flexpriority=
2584 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2585 Default is 1 (enabled)
2588 [KVM,Intel] Enable VMX nesting (nVMX).
2589 Default is 0 (disabled)
2591 kvm-intel.unrestricted_guest=
2592 [KVM,Intel] Disable unrestricted guest feature
2593 (virtualized real and unpaged mode) on capable
2594 Intel chips. Default is 1 (enabled)
2596 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2599 Valid arguments: never, cond, always
2601 always: L1D cache flush on every VMENTER.
2602 cond: Flush L1D on VMENTER only when the code between
2603 VMEXIT and VMENTER can leak host memory.
2604 never: Disables the mitigation
2606 Default is cond (do L1 cache flush in specific instances)
2608 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2609 feature (tagged TLBs) on capable Intel chips.
2610 Default is 1 (enabled)
2612 l1d_flush= [X86,INTEL]
2613 Control mitigation for L1D based snooping vulnerability.
2615 Certain CPUs are vulnerable to an exploit against CPU
2616 internal buffers which can forward information to a
2617 disclosure gadget under certain conditions.
2619 In vulnerable processors, the speculatively
2620 forwarded data can be used in a cache side channel
2621 attack, to access data to which the attacker does
2622 not have direct access.
2624 This parameter controls the mitigation. The
2627 on - enable the interface for the mitigation
2629 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2632 The kernel PTE inversion protection is unconditionally
2633 enabled and cannot be disabled.
2636 Provides all available mitigations for the
2637 L1TF vulnerability. Disables SMT and
2638 enables all mitigations in the
2639 hypervisors, i.e. unconditional L1D flush.
2641 SMT control and L1D flush control via the
2642 sysfs interface is still possible after
2643 boot. Hypervisors will issue a warning
2644 when the first VM is started in a
2645 potentially insecure configuration,
2646 i.e. SMT enabled or L1D flush disabled.
2649 Same as 'full', but disables SMT and L1D
2650 flush runtime control. Implies the
2651 'nosmt=force' command line option.
2652 (i.e. sysfs control of SMT is disabled.)
2655 Leaves SMT enabled and enables the default
2656 hypervisor mitigation, i.e. conditional
2659 SMT control and L1D flush control via the
2660 sysfs interface is still possible after
2661 boot. Hypervisors will issue a warning
2662 when the first VM is started in a
2663 potentially insecure configuration,
2664 i.e. SMT enabled or L1D flush disabled.
2668 Disables SMT and enables the default
2669 hypervisor mitigation.
2671 SMT control and L1D flush control via the
2672 sysfs interface is still possible after
2673 boot. Hypervisors will issue a warning
2674 when the first VM is started in a
2675 potentially insecure configuration,
2676 i.e. SMT enabled or L1D flush disabled.
2679 Same as 'flush', but hypervisors will not
2680 warn when a VM is started in a potentially
2681 insecure configuration.
2684 Disables hypervisor mitigations and doesn't
2686 It also drops the swap size and available
2687 RAM limit restriction on both hypervisor and
2692 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2698 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2701 lapic= [X86,APIC] Do not use TSC deadline
2702 value for LAPIC timer one-shot implementation. Default
2703 back to the programmable timer unit in the LAPIC.
2704 Format: notscdeadline
2706 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2709 libata.dma= [LIBATA] DMA control
2710 libata.dma=0 Disable all PATA and SATA DMA
2711 libata.dma=1 PATA and SATA Disk DMA only
2712 libata.dma=2 ATAPI (CDROM) DMA only
2713 libata.dma=4 Compact Flash DMA only
2714 Combinations also work, so libata.dma=3 enables DMA
2715 for disks and CDROMs, but not CFs.
2717 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2718 libata.ignore_hpa=0 keep BIOS limits (default)
2719 libata.ignore_hpa=1 ignore limits, using full disk
2721 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2725 libata.force= [LIBATA] Force configurations. The format is a comma-
2726 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2727 PORT and DEVICE are decimal numbers matching port, link
2728 or device. Basically, it matches the ATA ID string
2729 printed on console by libata. If the whole ID part is
2730 omitted, the last PORT and DEVICE values are used. If
2731 ID hasn't been specified yet, the configuration applies
2732 to all ports, links and devices.
2734 If only DEVICE is omitted, the parameter applies to
2735 the port and all links and devices behind it. DEVICE
2736 number of 0 either selects the first device or the
2737 first fan-out link behind PMP device. It does not
2738 select the host link. DEVICE number of 15 selects the
2739 host link and device attached to it.
2741 The VAL specifies the configuration to force. As long
2742 as there is no ambiguity, shortcut notation is allowed.
2743 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2744 The following configurations can be forced.
2746 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2747 Any ID with matching PORT is used.
2749 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2751 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2752 udma[/][16,25,33,44,66,100,133] notation is also
2755 * nohrst, nosrst, norst: suppress hard, soft and both
2758 * rstonce: only attempt one reset during hot-unplug
2761 * [no]dbdelay: Enable or disable the extra 200ms delay
2762 before debouncing a link PHY and device presence
2765 * [no]ncq: Turn on or off NCQ.
2767 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2769 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2771 * [no]trim: Enable or disable (unqueued) TRIM.
2773 * trim_zero: Indicate that TRIM command zeroes data.
2775 * max_trim_128m: Set 128M maximum trim size limit.
2777 * [no]dma: Turn on or off DMA transfers.
2779 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2781 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2782 commands that are not a multiple of 16 bytes.
2784 * [no]dmalog: Enable or disable the use of the
2785 READ LOG DMA EXT command to access logs.
2787 * [no]iddevlog: Enable or disable access to the
2788 identify device data log.
2790 * [no]logdir: Enable or disable access to the general
2791 purpose log directory.
2793 * max_sec_128: Set transfer size limit to 128 sectors.
2795 * max_sec_1024: Set or clear transfer size limit to
2798 * max_sec_lba48: Set or clear transfer size limit to
2801 * [no]lpm: Enable or disable link power management.
2803 * [no]setxfer: Indicate if transfer speed mode setting
2806 * dump_id: Dump IDENTIFY data.
2808 * disable: Disable this device.
2810 If there are multiple matching configurations changing
2811 the same attribute, the last one is used.
2813 load_ramdisk= [RAM] [Deprecated]
2815 lockd.nlm_grace_period=P [NFS] Assign grace period.
2818 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2821 lockd.nlm_timeout=T [NFS] Assign timeout value.
2824 lockd.nlm_udpport=M [NFS] Assign UDP port.
2827 lockdown= [SECURITY]
2828 { integrity | confidentiality }
2829 Enable the kernel lockdown feature. If set to
2830 integrity, kernel features that allow userland to
2831 modify the running kernel are disabled. If set to
2832 confidentiality, kernel features that allow userland
2833 to extract confidential information from the kernel
2836 locktorture.nreaders_stress= [KNL]
2837 Set the number of locking read-acquisition kthreads.
2838 Defaults to being automatically set based on the
2839 number of online CPUs.
2841 locktorture.nwriters_stress= [KNL]
2842 Set the number of locking write-acquisition kthreads.
2844 locktorture.onoff_holdoff= [KNL]
2845 Set time (s) after boot for CPU-hotplug testing.
2847 locktorture.onoff_interval= [KNL]
2848 Set time (s) between CPU-hotplug operations, or
2849 zero to disable CPU-hotplug testing.
2851 locktorture.shuffle_interval= [KNL]
2852 Set task-shuffle interval (jiffies). Shuffling
2853 tasks allows some CPUs to go into dyntick-idle
2854 mode during the locktorture test.
2856 locktorture.shutdown_secs= [KNL]
2857 Set time (s) after boot system shutdown. This
2858 is useful for hands-off automated testing.
2860 locktorture.stat_interval= [KNL]
2861 Time (s) between statistics printk()s.
2863 locktorture.stutter= [KNL]
2864 Time (s) to stutter testing, for example,
2865 specifying five seconds causes the test to run for
2866 five seconds, wait for five seconds, and so on.
2867 This tests the locking primitive's ability to
2868 transition abruptly to and from idle.
2870 locktorture.torture_type= [KNL]
2871 Specify the locking implementation to test.
2873 locktorture.verbose= [KNL]
2874 Enable additional printk() statements.
2876 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2879 loglevel= All Kernel Messages with a loglevel smaller than the
2880 console loglevel will be printed to the console. It can
2881 also be changed with klogd or other programs. The
2882 loglevels are defined as follows:
2884 0 (KERN_EMERG) system is unusable
2885 1 (KERN_ALERT) action must be taken immediately
2886 2 (KERN_CRIT) critical conditions
2887 3 (KERN_ERR) error conditions
2888 4 (KERN_WARNING) warning conditions
2889 5 (KERN_NOTICE) normal but significant condition
2890 6 (KERN_INFO) informational
2891 7 (KERN_DEBUG) debug-level messages
2893 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2894 in bytes. n must be a power of two and greater
2895 than the minimal size. The minimal size is defined
2896 by LOG_BUF_SHIFT kernel config parameter. There is
2897 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2898 that allows to increase the default size depending on
2899 the number of CPUs. See init/Kconfig for more details.
2901 logo.nologo [FB] Disables display of the built-in Linux logo.
2902 This may be used to provide more screen space for
2903 kernel log messages and is useful when debugging
2904 kernel boot problems.
2906 lp=0 [LP] Specify parallel ports to use, e.g,
2907 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2908 lp=reset first parallel port). 'lp=0' disables the
2909 lp=auto printer driver. 'lp=reset' (which can be
2910 specified in addition to the ports) causes
2911 attached printers to be reset. Using
2912 lp=port1,port2,... specifies the parallel ports
2913 to associate lp devices with, starting with
2914 lp0. A port specification may be 'none' to skip
2915 that lp device, or a parport name such as
2916 'parport0'. Specifying 'lp=auto' instead of a
2917 port specification list means that device IDs
2918 from each port should be examined, to see if
2919 an IEEE 1284-compliant printer is attached; if
2920 so, the driver will manage that printer.
2921 See also header of drivers/char/lp.c.
2924 Sets loops_per_jiffy to given constant, thus avoiding
2925 time-consuming boot-time autodetection (up to 250 ms per
2926 CPU). 0 enables autodetection (default). To determine
2927 the correct value for your kernel, boot with normal
2928 autodetection and see what value is printed. Note that
2929 on SMP systems the preset will be applied to all CPUs,
2930 which is likely to cause problems if your CPUs need
2931 significantly divergent settings. An incorrect value
2932 will cause delays in the kernel to be wrong, leading to
2933 unpredictable I/O errors and other breakage. Although
2934 unlikely, in the extreme case this might damage your
2938 Format: <io>,<irq>,<dma>
2940 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2943 [SECURITY] Choose order of LSM initialization. This
2944 overrides CONFIG_LSM, and the "security=" parameter.
2946 machvec= [IA-64] Force the use of a particular machine-vector
2947 (machvec) in a generic kernel.
2948 Example: machvec=hpzx1
2950 machtype= [Loongson] Share the same kernel image file between
2951 different yeeloong laptops.
2952 Example: machtype=lemote-yeeloong-2f-7inch
2954 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
2955 than or equal to this physical address is ignored.
2957 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2958 will bring up during bootup. maxcpus=n : n >= 0 limits
2959 the kernel to bring up 'n' processors. Surely after
2960 bootup you can bring up the other plugged cpu by executing
2961 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2962 only takes effect during system bootup.
2963 While n=0 is a special case, it is equivalent to "nosmp",
2964 which also disables the IO APIC.
2966 max_loop= [LOOP] The number of loop block devices that get
2967 (loop.max_loop) unconditionally pre-created at init time. The default
2968 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2969 of statically allocating a predefined number, loop
2970 devices can be requested on-demand with the
2971 /dev/loop-control interface.
2973 mce [X86-32] Machine Check Exception
2975 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2977 md= [HW] RAID subsystems devices and level
2978 See Documentation/admin-guide/md.rst.
2981 Format: <first>,<last>
2982 Specifies range of consoles to be captured by the MDA.
2985 Control mitigation for the Micro-architectural Data
2986 Sampling (MDS) vulnerability.
2988 Certain CPUs are vulnerable to an exploit against CPU
2989 internal buffers which can forward information to a
2990 disclosure gadget under certain conditions.
2992 In vulnerable processors, the speculatively
2993 forwarded data can be used in a cache side channel
2994 attack, to access data to which the attacker does
2995 not have direct access.
2997 This parameter controls the MDS mitigation. The
3000 full - Enable MDS mitigation on vulnerable CPUs
3001 full,nosmt - Enable MDS mitigation and disable
3002 SMT on vulnerable CPUs
3003 off - Unconditionally disable MDS mitigation
3005 On TAA-affected machines, mds=off can be prevented by
3006 an active TAA mitigation as both vulnerabilities are
3007 mitigated with the same mechanism so in order to disable
3008 this mitigation, you need to specify tsx_async_abort=off
3011 Not specifying this option is equivalent to
3014 For details see: Documentation/admin-guide/hw-vuln/mds.rst
3016 mem=nn[KMG] [HEXAGON] Set the memory size.
3017 Must be specified, otherwise memory size will be 0.
3019 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
3020 Amount of memory to be used in cases as follows:
3023 2 when the kernel is not able to see the whole system memory;
3024 3 memory that lies after 'mem=' boundary is excluded from
3025 the hypervisor, then assigned to KVM guests.
3026 4 to limit the memory available for kdump kernel.
3028 [ARC,MICROBLAZE] - the limit applies only to low memory,
3029 high memory is not affected.
3031 [ARM64] - only limits memory covered by the linear
3032 mapping. The NOMAP regions are not affected.
3034 [X86] Work as limiting max address. Use together
3035 with memmap= to avoid physical address space collisions.
3036 Without memmap= PCI devices could be placed at addresses
3037 belonging to unused RAM.
3039 Note that this only takes effects during boot time since
3040 in above case 3, memory may need be hot added after boot
3041 if system memory of hypervisor is not sufficient.
3044 [ARM,MIPS] - override the memory layout reported by
3046 Define a memory region of size nn[KMG] starting at
3048 Multiple different regions can be specified with
3049 multiple mem= parameters on the command line.
3051 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3054 memblock=debug [KNL] Enable memblock debug messages.
3057 [KNL,SH] Allow user to override the default size for
3058 per-device physically contiguous DMA buffers.
3060 memhp_default_state=online/offline
3061 [KNL] Set the initial state for the memory hotplug
3062 onlining policy. If not specified, the default value is
3063 set according to the
3064 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3066 See Documentation/admin-guide/mm/memory-hotplug.rst.
3068 memmap=exactmap [KNL,X86] Enable setting of an exact
3069 E820 memory map, as specified by the user.
3070 Such memmap=exactmap lines can be constructed based on
3071 BIOS output or other requirements. See the memmap=nn@ss
3074 memmap=nn[KMG]@ss[KMG]
3075 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
3076 Region of memory to be used is from ss to ss+nn.
3077 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3078 which limits max address to nn[KMG].
3079 Multiple different regions can be specified,
3082 memmap=100M@2G,100M#3G,1G!1024G
3084 memmap=nn[KMG]#ss[KMG]
3085 [KNL,ACPI] Mark specific memory as ACPI data.
3086 Region of memory to be marked is from ss to ss+nn.
3088 memmap=nn[KMG]$ss[KMG]
3089 [KNL,ACPI] Mark specific memory as reserved.
3090 Region of memory to be reserved is from ss to ss+nn.
3091 Example: Exclude memory from 0x18690000-0x1869ffff
3092 memmap=64K$0x18690000
3094 memmap=0x10000$0x18690000
3095 Some bootloaders may need an escape character before '$',
3096 like Grub2, otherwise '$' and the following number
3099 memmap=nn[KMG]!ss[KMG]
3100 [KNL,X86] Mark specific memory as protected.
3101 Region of memory to be used, from ss to ss+nn.
3102 The memory region may be marked as e820 type 12 (0xc)
3103 and is NVDIMM or ADR memory.
3105 memmap=<size>%<offset>-<oldtype>+<newtype>
3106 [KNL,ACPI] Convert memory within the specified region
3107 from <oldtype> to <newtype>. If "-<oldtype>" is left
3108 out, the whole region will be marked as <newtype>,
3109 even if previously unavailable. If "+<newtype>" is left
3110 out, matching memory will be removed. Types are
3111 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3112 3 = ACPI, 12 = PRAM.
3114 memory_corruption_check=0/1 [X86]
3115 Some BIOSes seem to corrupt the first 64k of
3116 memory when doing things like suspend/resume.
3117 Setting this option will scan the memory
3118 looking for corruption. Enabling this will
3119 both detect corruption and prevent the kernel
3120 from using the memory being corrupted.
3121 However, its intended as a diagnostic tool; if
3122 repeatable BIOS-originated corruption always
3123 affects the same memory, you can use memmap=
3124 to prevent the kernel from using that memory.
3126 memory_corruption_check_size=size [X86]
3127 By default it checks for corruption in the low
3128 64k, making this memory unavailable for normal
3129 use. Use this parameter to scan for
3130 corruption in more or less memory.
3132 memory_corruption_check_period=seconds [X86]
3133 By default it checks for corruption every 60
3134 seconds. Use this parameter to check at some
3135 other rate. 0 disables periodic checking.
3137 memory_hotplug.memmap_on_memory
3138 [KNL,X86,ARM] Boolean flag to enable this feature.
3139 Format: {on | off (default)}
3140 When enabled, runtime hotplugged memory will
3141 allocate its internal metadata (struct pages,
3142 those vmemmap pages cannot be optimized even
3143 if hugetlb_free_vmemmap is enabled) from the
3144 hotadded memory which will allow to hotadd a
3145 lot of memory without requiring additional
3147 This feature is disabled by default because it
3148 has some implication on large (e.g. GB)
3149 allocations in some configurations (e.g. small
3151 The state of the flag can be read in
3152 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3153 Note that even when enabled, there are a few cases where
3154 the feature is not effective.
3156 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3158 default : 0 <disable>
3159 Specifies the number of memtest passes to be
3160 performed. Each pass selects another test
3161 pattern from a given set of patterns. Memtest
3162 fills the memory with this pattern, validates
3163 memory contents and reserves bad memory
3164 regions that are detected.
3166 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3167 Valid arguments: on, off
3168 Default (depends on kernel configuration option):
3169 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3170 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3171 mem_encrypt=on: Activate SME
3172 mem_encrypt=off: Do not activate SME
3174 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3175 for details on when memory encryption can be activated.
3177 mem_sleep_default= [SUSPEND] Default system suspend mode:
3178 s2idle - Suspend-To-Idle
3179 shallow - Power-On Suspend or equivalent (if supported)
3180 deep - Suspend-To-RAM or equivalent (if supported)
3181 See Documentation/admin-guide/pm/sleep-states.rst.
3183 meye.*= [HW] Set MotionEye Camera parameters
3184 See Documentation/admin-guide/media/meye.rst.
3186 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3187 Multi-Function General Purpose Timers on AMD Geode
3190 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3191 the BIOS has incorrectly applied a workaround. TinyBIOS
3192 version 0.98 is known to be affected, 0.99 fixes the
3193 problem by letting the user disable the workaround.
3197 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3198 physical address is ignored.
3200 mini2440= [ARM,HW,KNL]
3201 Format:[0..2][b][c][t]
3203 MINI2440 configuration specification:
3204 0 - The attached screen is the 3.5" TFT
3205 1 - The attached screen is the 7" TFT
3206 2 - The VGA Shield is attached (1024x768)
3207 Leaving out the screen size parameter will not load
3208 the TFT driver, and the framebuffer will be left
3210 b - Enable backlight. The TFT backlight pin will be
3211 linked to the kernel VESA blanking code and a GPIO
3212 LED. This parameter is not necessary when using the
3214 c - Enable the s3c camera interface.
3215 t - Reserved for enabling touchscreen support. The
3216 touchscreen support is not enabled in the mainstream
3217 kernel as of 2.6.30, a preliminary port can be found
3218 in the "bleeding edge" mini2440 support kernel at
3219 https://repo.or.cz/w/linux-2.6/mini2440.git
3222 [X86,PPC,S390,ARM64] Control optional mitigations for
3223 CPU vulnerabilities. This is a set of curated,
3224 arch-independent options, each of which is an
3225 aggregation of existing arch-specific options.
3228 Disable all optional CPU mitigations. This
3229 improves system performance, but it may also
3230 expose users to several CPU vulnerabilities.
3231 Equivalent to: nopti [X86,PPC]
3232 if nokaslr then kpti=0 [ARM64]
3233 nospectre_v1 [X86,PPC]
3235 nospectre_v2 [X86,PPC,S390,ARM64]
3236 spectre_v2_user=off [X86]
3237 spec_store_bypass_disable=off [X86,PPC]
3238 ssbd=force-off [ARM64]
3239 nospectre_bhb [ARM64]
3242 tsx_async_abort=off [X86]
3243 kvm.nx_huge_pages=off [X86]
3244 srbds=off [X86,INTEL]
3245 no_entry_flush [PPC]
3246 no_uaccess_flush [PPC]
3247 mmio_stale_data=off [X86]
3251 This does not have any effect on
3252 kvm.nx_huge_pages when
3253 kvm.nx_huge_pages=force.
3256 Mitigate all CPU vulnerabilities, but leave SMT
3257 enabled, even if it's vulnerable. This is for
3258 users who don't want to be surprised by SMT
3259 getting disabled across kernel upgrades, or who
3260 have other ways of avoiding SMT-based attacks.
3261 Equivalent to: (default behavior)
3264 Mitigate all CPU vulnerabilities, disabling SMT
3265 if needed. This is for users who always want to
3266 be fully mitigated, even if it means losing SMT.
3267 Equivalent to: l1tf=flush,nosmt [X86]
3268 mds=full,nosmt [X86]
3269 tsx_async_abort=full,nosmt [X86]
3270 mmio_stale_data=full,nosmt [X86]
3271 retbleed=auto,nosmt [X86]
3274 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3275 parameter allows control of the logging verbosity for
3276 the additional memory initialisation checks. A value
3277 of 0 disables mminit logging and a level of 4 will
3278 log everything. Information is printed at KERN_DEBUG
3279 so loglevel=8 may also need to be specified.
3282 [X86,INTEL] Control mitigation for the Processor
3283 MMIO Stale Data vulnerabilities.
3285 Processor MMIO Stale Data is a class of
3286 vulnerabilities that may expose data after an MMIO
3287 operation. Exposed data could originate or end in
3288 the same CPU buffers as affected by MDS and TAA.
3289 Therefore, similar to MDS and TAA, the mitigation
3290 is to clear the affected CPU buffers.
3292 This parameter controls the mitigation. The
3295 full - Enable mitigation on vulnerable CPUs
3297 full,nosmt - Enable mitigation and disable SMT on
3300 off - Unconditionally disable mitigation
3302 On MDS or TAA affected machines,
3303 mmio_stale_data=off can be prevented by an active
3304 MDS or TAA mitigation as these vulnerabilities are
3305 mitigated with the same mechanism so in order to
3306 disable this mitigation, you need to specify
3307 mds=off and tsx_async_abort=off too.
3309 Not specifying this option is equivalent to
3310 mmio_stale_data=full.
3313 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3315 module.async_probe=<bool>
3316 [KNL] When set to true, modules will use async probing
3317 by default. To enable/disable async probing for a
3318 specific module, use the module specific control that
3319 is documented under <module>.async_probe. When both
3320 module.async_probe and <module>.async_probe are
3321 specified, <module>.async_probe takes precedence for
3322 the specific module.
3325 [KNL] When CONFIG_MODULE_SIG is set, this means that
3326 modules without (valid) signatures will fail to load.
3327 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3328 is always true, so this option does nothing.
3330 module_blacklist= [KNL] Do not load a comma-separated list of
3331 modules. Useful for debugging problem modules.
3334 [MOUSE] Maximum time between finger touching and
3335 leaving touchpad surface for touch to be considered
3336 a tap and be reported as a left button click (for
3337 touchpads working in absolute mode only).
3339 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3340 reporting absolute coordinates, such as tablets
3341 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3342 reporting absolute coordinates, such as tablets
3344 movablecore= [KNL,X86,IA-64,PPC]
3345 Format: nn[KMGTPE] | nn%
3346 This parameter is the complement to kernelcore=, it
3347 specifies the amount of memory used for migratable
3348 allocations. If both kernelcore and movablecore is
3349 specified, then kernelcore will be at *least* the
3350 specified value but may be more. If movablecore on its
3351 own is specified, the administrator must be careful
3352 that the amount of memory usable for all allocations
3355 movable_node [KNL] Boot-time switch to make hotplugable memory
3356 NUMA nodes to be movable. This means that the memory
3357 of such nodes will be usable only for movable
3358 allocations which rules out almost all kernel
3359 allocations. Use with caution!
3361 MTD_Partition= [MTD]
3362 Format: <name>,<region-number>,<size>,<offset>
3364 MTD_Region= [MTD] Format:
3365 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3368 See drivers/mtd/parsers/cmdlinepart.c
3371 ARM/S3C2412 JIVE boot control
3373 See arch/arm/mach-s3c/mach-jive.c
3375 mtouchusb.raw_coordinates=
3376 [HW] Make the MicroTouch USB driver use raw coordinates
3377 ('y', default) or cooked coordinates ('n')
3379 mtrr_chunk_size=nn[KMG] [X86]
3380 used for mtrr cleanup. It is largest continuous chunk
3381 that could hold holes aka. UC entries.
3383 mtrr_gran_size=nn[KMG] [X86]
3384 Used for mtrr cleanup. It is granularity of mtrr block.
3386 Large value could prevent small alignment from
3389 mtrr_spare_reg_nr=n [X86]
3391 Range: 0,7 : spare reg number
3393 Used for mtrr cleanup. It is spare mtrr entries number.
3394 Set to 2 or more if your graphical card needs more.
3396 multitce=off [PPC] This parameter disables the use of the pSeries
3397 firmware feature for updating multiple TCE entries
3400 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3402 netdev= [NET] Network devices parameters
3403 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3404 Note that mem_start is often overloaded to mean
3405 something different and driver-specific.
3406 This usage is only documented in each driver source
3409 netpoll.carrier_timeout=
3410 [NET] Specifies amount of time (in seconds) that
3411 netpoll should wait for a carrier. By default netpoll
3415 [NETFILTER] Enable connection tracking flow accounting
3416 0 to disable accounting
3417 1 to enable accounting
3420 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3421 See Documentation/admin-guide/nfs/nfsroot.rst.
3423 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3424 See Documentation/admin-guide/nfs/nfsroot.rst.
3426 nfsrootdebug [NFS] enable nfsroot debugging messages.
3427 See Documentation/admin-guide/nfs/nfsroot.rst.
3429 nfs.callback_nr_threads=
3430 [NFSv4] set the total number of threads that the
3431 NFS client will assign to service NFSv4 callback
3434 nfs.callback_tcpport=
3435 [NFS] set the TCP port on which the NFSv4 callback
3436 channel should listen.
3439 [NFS] sets the pathname to the program which is used
3440 to update the NFS client cache entries.
3442 nfs.cache_getent_timeout=
3443 [NFS] sets the timeout after which an attempt to
3444 update a cache entry is deemed to have failed.
3446 nfs.idmap_cache_timeout=
3447 [NFS] set the maximum lifetime for idmapper cache
3451 [NFS] enable 64-bit inode numbers.
3452 If zero, the NFS client will fake up a 32-bit inode
3453 number for the readdir() and stat() syscalls instead
3454 of returning the full 64-bit number.
3455 The default is to return 64-bit inode numbers.
3457 nfs.max_session_cb_slots=
3458 [NFSv4.1] Sets the maximum number of session
3459 slots the client will assign to the callback
3460 channel. This determines the maximum number of
3461 callbacks the client will process in parallel for
3462 a particular server.
3464 nfs.max_session_slots=
3465 [NFSv4.1] Sets the maximum number of session slots
3466 the client will attempt to negotiate with the server.
3467 This limits the number of simultaneous RPC requests
3468 that the client can send to the NFSv4.1 server.
3469 Note that there is little point in setting this
3470 value higher than the max_tcp_slot_table_limit.
3472 nfs.nfs4_disable_idmapping=
3473 [NFSv4] When set to the default of '1', this option
3474 ensures that both the RPC level authentication
3475 scheme and the NFS level operations agree to use
3476 numeric uids/gids if the mount is using the
3477 'sec=sys' security flavour. In effect it is
3478 disabling idmapping, which can make migration from
3479 legacy NFSv2/v3 systems to NFSv4 easier.
3480 Servers that do not support this mode of operation
3481 will be autodetected by the client, and it will fall
3482 back to using the idmapper.
3483 To turn off this behaviour, set the value to '0'.
3485 [NFS4] Specify an additional fixed unique ident-
3486 ification string that NFSv4 clients can insert into
3487 their nfs_client_id4 string. This is typically a
3488 UUID that is generated at system install time.
3490 nfs.send_implementation_id =
3491 [NFSv4.1] Send client implementation identification
3492 information in exchange_id requests.
3493 If zero, no implementation identification information
3495 The default is to send the implementation identification
3498 nfs.recover_lost_locks =
3499 [NFSv4] Attempt to recover locks that were lost due
3500 to a lease timeout on the server. Please note that
3501 doing this risks data corruption, since there are
3502 no guarantees that the file will remain unchanged
3503 after the locks are lost.
3504 If you want to enable the kernel legacy behaviour of
3505 attempting to recover these locks, then set this
3507 The default parameter value of '0' causes the kernel
3508 not to attempt recovery of lost locks.
3510 nfs4.layoutstats_timer =
3511 [NFSv4.2] Change the rate at which the kernel sends
3512 layoutstats to the pNFS metadata server.
3514 Setting this to value to 0 causes the kernel to use
3515 whatever value is the default set by the layout
3516 driver. A non-zero value sets the minimum interval
3517 in seconds between layoutstats transmissions.
3519 nfsd.inter_copy_offload_enable =
3520 [NFSv4.2] When set to 1, the server will support
3521 server-to-server copies for which this server is
3522 the destination of the copy.
3524 nfsd.nfsd4_ssc_umount_timeout =
3525 [NFSv4.2] When used as the destination of a
3526 server-to-server copy, knfsd temporarily mounts
3527 the source server. It caches the mount in case
3528 it will be needed again, and discards it if not
3529 used for the number of milliseconds specified by
3532 nfsd.nfs4_disable_idmapping=
3533 [NFSv4] When set to the default of '1', the NFSv4
3534 server will return only numeric uids and gids to
3535 clients using auth_sys, and will accept numeric uids
3536 and gids from such clients. This is intended to ease
3537 migration from NFSv2/v3.
3540 nmi_backtrace.backtrace_idle [KNL]
3541 Dump stacks even of idle CPUs in response to an
3542 NMI stack-backtrace request.
3544 nmi_debug= [KNL,SH] Specify one or more actions to take
3545 when a NMI is triggered.
3546 Format: [state][,regs][,debounce][,die]
3548 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3549 Format: [panic,][nopanic,][num]
3551 0 - turn hardlockup detector in nmi_watchdog off
3552 1 - turn hardlockup detector in nmi_watchdog on
3553 When panic is specified, panic when an NMI watchdog
3554 timeout occurs (or 'nopanic' to not panic on an NMI
3555 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3556 To disable both hard and soft lockup detectors,
3557 please see 'nowatchdog'.
3558 This is useful when you use a panic=... timeout and
3559 need the box quickly up again.
3561 These settings can be accessed at runtime via
3562 the nmi_watchdog and hardlockup_panic sysctls.
3564 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3565 emulation library even if a 387 maths coprocessor
3568 no5lvl [X86-64] Disable 5-level paging mode. Forces
3569 kernel to use 4-level paging instead.
3571 nofsgsbase [X86] Disables FSGSBASE instructions.
3574 [HW] Never suspend the console
3575 Disable suspending of consoles during suspend and
3576 hibernate operations. Once disabled, debugging
3577 messages can reach various consoles while the rest
3578 of the system is being put to sleep (ie, while
3579 debugging driver suspend/resume hooks). This may
3580 not work reliably with all consoles, but is known
3581 to work with serial and VGA consoles.
3582 To facilitate more flexible debugging, we also add
3583 console_suspend, a printk module parameter to control
3584 it. Users could use console_suspend (usually
3585 /sys/module/printk/parameters/console_suspend) to
3586 turn on/off it dynamically.
3588 novmcoredd [KNL,KDUMP]
3589 Disable device dump. Device dump allows drivers to
3590 append dump data to vmcore so you can collect driver
3591 specified debug info. Drivers can append the data
3592 without any limit and this data is stored in memory,
3593 so this may cause significant memory stress. Disabling
3594 device dump can help save memory but the driver debug
3595 data will be no longer available. This parameter
3596 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3599 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3600 caches in the slab allocator. Saves per-node memory,
3601 but will impact performance.
3605 noaltinstr [S390] Disables alternative instructions patching
3606 (CPU alternatives feature).
3608 noapic [SMP,APIC] Tells the kernel to not make use of any
3609 IOAPICs that may be present in the system.
3611 noautogroup Disable scheduler automatic task group creation.
3615 nodsp [SH] Disable hardware DSP at boot time.
3617 noefi Disable EFI runtime services support.
3619 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3624 Disable SMAP (Supervisor Mode Access Prevention)
3625 even if it is supported by processor.
3628 Disable SMEP (Supervisor Mode Execution Prevention)
3629 even if it is supported by processor.
3632 This affects only 32-bit executables.
3633 noexec32=on: enable non-executable mappings (default)
3634 read doesn't imply executable mappings
3635 noexec32=off: disable non-executable mappings
3636 read implies executable mappings
3638 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3640 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3641 register save and restore. The kernel will only save
3642 legacy floating-point registers on task switch.
3644 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3646 nohugevmalloc [KNL,X86,PPC,ARM64] Disable kernel huge vmalloc mappings.
3648 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3649 Equivalent to smt=1.
3651 [KNL,X86] Disable symmetric multithreading (SMT).
3652 nosmt=force: Force disable SMT, cannot be undone
3653 via the sysfs control file.
3655 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3656 (bounds check bypass). With this option data leaks are
3657 possible in the system.
3659 nospectre_v2 [X86,PPC_E500,ARM64] Disable all mitigations for
3660 the Spectre variant 2 (indirect branch prediction)
3661 vulnerability. System may allow data leaks with this
3664 nospectre_bhb [ARM64] Disable all mitigations for Spectre-BHB (branch
3665 history injection) vulnerability. System may allow data leaks
3668 nospec_store_bypass_disable
3669 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3672 [PPC] Don't flush the L1-D cache after accessing user data.
3674 noxsave [BUGS=X86] Disables x86 extended register state save
3675 and restore using xsave. The kernel will fallback to
3676 enabling legacy floating-point and sse state.
3678 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3679 register states. The kernel will fall back to use
3680 xsave to save the states. By using this parameter,
3681 performance of saving the states is degraded because
3682 xsave doesn't support modified optimization while
3683 xsaveopt supports it on xsaveopt enabled systems.
3685 noxsaves [X86] Disables xsaves and xrstors used in saving and
3686 restoring x86 extended register state in compacted
3687 form of xsave area. The kernel will fall back to use
3688 xsaveopt and xrstor to save and restore the states
3689 in standard form of xsave area. By using this
3690 parameter, xsave area per process might occupy more
3691 memory on xsaves enabled systems.
3693 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3694 in do_idle() and not use the arch_cpu_idle()
3695 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3696 to be effective. This is useful on platforms where the
3697 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3698 correctly or when doing power measurements to evalute
3699 the impact of the sleep instructions. This is also
3700 useful when using JTAG debugger.
3702 no_file_caps Tells the kernel not to honor file capabilities. The
3703 only way then for a file to be executed with privilege
3704 is to be setuid root or executed by root.
3706 nohalt [IA-64] Tells the kernel not to use the power saving
3707 function PAL_HALT_LIGHT when idle. This increases
3708 power-consumption. On the positive side, it reduces
3709 interrupt wake-up latency, which may improve performance
3710 in certain environments such as networked servers or
3714 Force pointers printed to the console or buffers to be
3715 unhashed. By default, when a pointer is printed via %p
3716 format string, that pointer is "hashed", i.e. obscured
3717 by hashing the pointer value. This is a security feature
3718 that hides actual kernel addresses from unprivileged
3719 users, but it also makes debugging the kernel more
3720 difficult since unequal pointers can no longer be
3721 compared. However, if this command-line option is
3722 specified, then all normal pointers will have their true
3723 value printed. This option should only be specified when
3724 debugging the kernel. Please do not use on production
3727 nohibernate [HIBERNATION] Disable hibernation and resume.
3729 nohz= [KNL] Boottime enable/disable dynamic ticks
3730 Valid arguments: on, off
3733 nohz_full= [KNL,BOOT,SMP,ISOL]
3734 The argument is a cpu list, as described above.
3735 In kernels built with CONFIG_NO_HZ_FULL=y, set
3736 the specified list of CPUs whose tick will be stopped
3737 whenever possible. The boot CPU will be forced outside
3738 the range to maintain the timekeeping. Any CPUs
3739 in this list will have their RCU callbacks offloaded,
3740 just as if they had also been called out in the
3741 rcu_nocbs= boot parameter.
3743 Note that this argument takes precedence over
3744 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3746 noiotrap [SH] Disables trapped I/O port accesses.
3748 noirqdebug [X86-32] Disables the code which attempts to detect and
3749 disable unhandled interrupt sources.
3751 no_timer_check [X86,APIC] Disables the code which tests for
3752 broken timer IRQ sources.
3754 noisapnp [ISAPNP] Disables ISA PnP code.
3756 noinitrd [RAM] Tells the kernel not to load any configured
3759 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3761 [Deprecated - use intremap=off]
3765 noinvpcid [X86] Disable the INVPCID cpu feature.
3767 nojitter [IA-64] Disables jitter checking for ITC timers.
3769 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3771 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3775 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3776 clock and use the default one.
3778 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES] Disable paravirtualized
3779 steal time accounting. steal time is computed, but
3780 won't influence scheduler behaviour
3782 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3784 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3786 nomca [IA-64] Disable machine check abort handling
3788 nomce [X86-32] Disable Machine Check Exception
3790 nomfgpt [X86-32] Disable Multi-Function General Purpose
3791 Timer usage (for AMD Geode machines).
3793 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3794 shutdown the other cpus. Instead use the REBOOT_VECTOR
3797 nomodeset Disable kernel modesetting. DRM drivers will not perform
3798 display-mode changes or accelerated rendering. Only the
3799 system framebuffer will be available for use if this was
3800 set-up by the firmware or boot loader.
3802 Useful as fallback, or for testing and debugging.
3804 nomodule Disable module load
3806 nopat [X86] Disable PAT (page attribute table extension of
3807 pagetables) support.
3809 nopcid [X86-64] Disable the PCID cpu feature.
3811 norandmaps Don't use address space randomization. Equivalent to
3812 echo 0 > /proc/sys/kernel/randomize_va_space
3814 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3815 with UP alternatives
3817 noresume [SWSUSP] Disables resume and restores original swap
3820 no-scroll [VGA] Disables scrollback.
3821 This is required for the Braillex ib80-piezo Braille
3822 reader made by F.H. Papenmeier (Germany).
3826 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3828 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3829 and disable the IO APIC. legacy for "maxcpus=0".
3831 nosoftlockup [KNL] Disable the soft-lockup detector.
3833 nosync [HW,M68K] Disables sync negotiation for all devices.
3835 nowatchdog [KNL] Disable both lockup detectors, i.e.
3836 soft-lockup and NMI watchdog (hard-lockup).
3840 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3842 NOTE: this parameter will be ignored on systems with the
3843 LEGACY_XAPIC_DISABLED bit set in the
3844 IA32_XAPIC_DISABLE_STATUS MSR.
3846 nps_mtm_hs_ctr= [KNL,ARC]
3847 This parameter sets the maximum duration, in
3848 cycles, each HW thread of the CTOP can run
3849 without interruptions, before HW switches it.
3850 The actual maximum duration is 16 times this
3852 Format: integer between 1 and 255
3855 nptcg= [IA-64] Override max number of concurrent global TLB
3856 purges which is reported from either PAL_VM_SUMMARY or
3859 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3860 could support. nr_cpus=n : n >= 1 limits the kernel to
3861 support 'n' processors. It could be larger than the
3862 number of already plugged CPU during bootup, later in
3863 runtime you can physically add extra cpu until it reaches
3864 n. So during boot up some boot time memory for per-cpu
3865 variables need be pre-allocated for later physical cpu
3868 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3870 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3871 set up a single NUMA node spanning all memory.
3873 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3875 Allowed values are enable and disable
3877 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3878 'node', 'default' can be specified
3879 This can be set from sysctl after boot.
3880 See Documentation/admin-guide/sysctl/vm.rst for details.
3882 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3883 See Documentation/core-api/debugging-via-ohci1394.rst for more
3886 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3887 Rather than timing out after 20 ms if an EC
3888 command is not properly ACKed, override the length
3889 of the timeout. We have interrupts disabled while
3890 waiting for the ACK, so if this is set too high
3891 interrupts *may* be lost!
3893 omap_mux= [OMAP] Override bootloader pin multiplexing.
3894 Format: <mux_mode0.mode_name=value>...
3895 For example, to override I2C bus2:
3896 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3898 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3900 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3902 boundary - index of last SLC block on Flex-OneNAND.
3903 The remaining blocks are configured as MLC blocks.
3904 lock - Configure if Flex-OneNAND boundary should be locked.
3905 Once locked, the boundary cannot be changed.
3906 1 indicates lock status, 0 indicates unlock status.
3908 oops=panic Always panic on oopses. Default is to just kill the
3909 process, but there is a small probability of
3910 deadlocking the machine.
3911 This will also cause panics on machine check exceptions.
3912 Useful together with panic=30 to trigger a reboot.
3915 [KNL] Boolean flag to control whether the page allocator
3916 should randomize its free lists. The randomization may
3917 be automatically enabled if the kernel detects it is
3918 running on a platform with a direct-mapped memory-side
3919 cache, and this parameter can be used to
3920 override/disable that behavior. The state of the flag
3921 can be read from sysfs at:
3922 /sys/module/page_alloc/parameters/shuffle.
3924 page_owner= [KNL] Boot-time page_owner enabling option.
3925 Storage of the information about who allocated
3926 each page is disabled in default. With this switch,
3928 on: enable the feature
3930 page_poison= [KNL] Boot-time parameter changing the state of
3931 poisoning on the buddy allocator, available with
3932 CONFIG_PAGE_POISONING=y.
3933 off: turn off poisoning (default)
3934 on: turn on poisoning
3936 page_reporting.page_reporting_order=
3937 [KNL] Minimal page reporting order
3939 Adjust the minimal page reporting order. The page
3940 reporting is disabled when it exceeds (MAX_ORDER-1).
3942 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3943 timeout > 0: seconds before rebooting
3944 timeout = 0: wait forever
3945 timeout < 0: reboot immediately
3948 panic_print= Bitmask for printing system info when panic happens.
3949 User can chose combination of the following bits:
3950 bit 0: print all tasks info
3951 bit 1: print system memory info
3952 bit 2: print timer info
3953 bit 3: print locks info if CONFIG_LOCKDEP is on
3954 bit 4: print ftrace buffer
3955 bit 5: print all printk messages in buffer
3956 bit 6: print all CPUs backtrace (if available in the arch)
3957 *Be aware* that this option may print a _lot_ of lines,
3958 so there are risks of losing older messages in the log.
3959 Use this option carefully, maybe worth to setup a
3960 bigger log buffer with "log_buf_len" along with this.
3962 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3963 Format: <hex>[,nousertaint]
3964 Hexadecimal bitmask representing the set of TAINT flags
3965 that will cause the kernel to panic when add_taint() is
3966 called with any of the flags in this set.
3967 The optional switch "nousertaint" can be utilized to
3968 prevent userspace forced crashes by writing to sysctl
3969 /proc/sys/kernel/tainted any flagset matching with the
3970 bitmask set on panic_on_taint.
3971 See Documentation/admin-guide/tainted-kernels.rst for
3972 extra details on the taint flags that users can pick
3973 to compose the bitmask to assign to panic_on_taint.
3975 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3978 parkbd.port= [HW] Parallel port number the keyboard adapter is
3979 connected to, default is 0.
3981 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3982 0 for XT, 1 for AT (default is AT).
3985 parport= [HW,PPT] Specify parallel ports. 0 disables.
3986 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3987 Use 'auto' to force the driver to use any
3988 IRQ/DMA settings detected (the default is to
3989 ignore detected IRQ/DMA settings because of
3990 possible conflicts). You can specify the base
3991 address, IRQ, and DMA settings; IRQ and DMA
3992 should be numbers, or 'auto' (for using detected
3993 settings on that particular port), or 'nofifo'
3994 (to avoid using a FIFO even if it is detected).
3995 Parallel ports are assigned in the order they
3996 are specified on the command line, starting
3999 parport_init_mode= [HW,PPT]
4000 Configure VIA parallel port to operate in
4001 a specific mode. This is necessary on Pegasos
4002 computer where firmware has no options for setting
4003 up parallel port mode and sets it to spp.
4004 Currently this function knows 686a and 8231 chips.
4005 Format: [spp|ps2|epp|ecp|ecpepp]
4007 pata_legacy.all= [HW,LIBATA]
4009 Set to non-zero to probe primary and secondary ISA
4010 port ranges on PCI systems where no PCI PATA device
4011 has been found at either range. Disabled by default.
4013 pata_legacy.autospeed= [HW,LIBATA]
4015 Set to non-zero if a chip is present that snoops speed
4016 changes. Disabled by default.
4018 pata_legacy.ht6560a= [HW,LIBATA]
4020 Set to 1, 2, or 3 for HT 6560A on the primary channel,
4021 the secondary channel, or both channels respectively.
4022 Disabled by default.
4024 pata_legacy.ht6560b= [HW,LIBATA]
4026 Set to 1, 2, or 3 for HT 6560B on the primary channel,
4027 the secondary channel, or both channels respectively.
4028 Disabled by default.
4030 pata_legacy.iordy_mask= [HW,LIBATA]
4032 IORDY enable mask. Set individual bits to allow IORDY
4033 for the respective channel. Bit 0 is for the first
4034 legacy channel handled by this driver, bit 1 is for
4035 the second channel, and so on. The sequence will often
4036 correspond to the primary legacy channel, the secondary
4037 legacy channel, and so on, but the handling of a PCI
4038 bus and the use of other driver options may interfere
4039 with the sequence. By default IORDY is allowed across
4042 pata_legacy.opti82c46x= [HW,LIBATA]
4044 Set to 1, 2, or 3 for Opti 82c611A on the primary
4045 channel, the secondary channel, or both channels
4046 respectively. Disabled by default.
4048 pata_legacy.opti82c611a= [HW,LIBATA]
4050 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4051 channel, the secondary channel, or both channels
4052 respectively. Disabled by default.
4054 pata_legacy.pio_mask= [HW,LIBATA]
4056 PIO mode mask for autospeed devices. Set individual
4057 bits to allow the use of the respective PIO modes.
4058 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4059 All modes allowed by default.
4061 pata_legacy.probe_all= [HW,LIBATA]
4063 Set to non-zero to probe tertiary and further ISA
4064 port ranges on PCI systems. Disabled by default.
4066 pata_legacy.probe_mask= [HW,LIBATA]
4068 Probe mask for legacy ISA PATA ports. Depending on
4069 platform configuration and the use of other driver
4070 options up to 6 legacy ports are supported: 0x1f0,
4071 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4072 of individual ports can be disabled by setting the
4073 corresponding bits in the mask to 1. Bit 0 is for
4074 the first port in the list above (0x1f0), and so on.
4075 By default all supported ports are probed.
4077 pata_legacy.qdi= [HW,LIBATA]
4079 Set to non-zero to probe QDI controllers. By default
4080 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4082 pata_legacy.winbond= [HW,LIBATA]
4084 Set to non-zero to probe Winbond controllers. Use
4085 the standard I/O port (0x130) if 1, otherwise the
4086 value given is the I/O port to use (typically 0x1b0).
4087 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4090 pata_platform.pio_mask= [HW,LIBATA]
4092 Supported PIO mode mask. Set individual bits to allow
4093 the use of the respective PIO modes. Bit 0 is for
4094 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4098 Halt all CPUs after the first oops has been printed for
4099 the specified number of seconds. This is to be used if
4100 your oopses keep scrolling off the screen.
4105 See header of drivers/block/paride/pcd.c.
4106 See also Documentation/admin-guide/blockdev/paride.rst.
4108 pci=option[,option...] [PCI] various PCI subsystem options.
4110 Some options herein operate on a specific device
4111 or a set of devices (<pci_dev>). These are
4112 specified in one of the following formats:
4114 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4115 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4117 Note: the first format specifies a PCI
4118 bus/device/function address which may change
4119 if new hardware is inserted, if motherboard
4120 firmware changes, or due to changes caused
4121 by other kernel parameters. If the
4122 domain is left unspecified, it is
4123 taken to be zero. Optionally, a path
4124 to a device through multiple device/function
4125 addresses can be specified after the base
4126 address (this is more robust against
4127 renumbering issues). The second format
4128 selects devices using IDs from the
4129 configuration space which may match multiple
4130 devices in the system.
4132 earlydump dump PCI config space before the kernel
4134 off [X86] don't probe for the PCI bus
4135 bios [X86-32] force use of PCI BIOS, don't access
4136 the hardware directly. Use this if your machine
4137 has a non-standard PCI host bridge.
4138 nobios [X86-32] disallow use of PCI BIOS, only direct
4139 hardware access methods are allowed. Use this
4140 if you experience crashes upon bootup and you
4141 suspect they are caused by the BIOS.
4142 conf1 [X86] Force use of PCI Configuration Access
4143 Mechanism 1 (config address in IO port 0xCF8,
4144 data in IO port 0xCFC, both 32-bit).
4145 conf2 [X86] Force use of PCI Configuration Access
4146 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4147 the function, IO port 0xCFA, also 8-bit, sets
4148 bus number. The config space is then accessed
4149 through ports 0xC000-0xCFFF).
4150 See http://wiki.osdev.org/PCI for more info
4151 on the configuration access mechanisms.
4152 noaer [PCIE] If the PCIEAER kernel config parameter is
4153 enabled, this kernel boot option can be used to
4154 disable the use of PCIE advanced error reporting.
4155 nodomains [PCI] Disable support for multiple PCI
4156 root domains (aka PCI segments, in ACPI-speak).
4157 nommconf [X86] Disable use of MMCONFIG for PCI
4159 check_enable_amd_mmconf [X86] check for and enable
4160 properly configured MMIO access to PCI
4161 config space on AMD family 10h CPU
4162 nomsi [MSI] If the PCI_MSI kernel config parameter is
4163 enabled, this kernel boot option can be used to
4164 disable the use of MSI interrupts system-wide.
4165 noioapicquirk [APIC] Disable all boot interrupt quirks.
4166 Safety option to keep boot IRQs enabled. This
4167 should never be necessary.
4168 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4169 primary IO-APIC for bridges that cannot disable
4170 boot IRQs. This fixes a source of spurious IRQs
4171 when the system masks IRQs.
4172 noioapicreroute [APIC] Disable workaround that uses the
4173 boot IRQ equivalent of an IRQ that connects to
4174 a chipset where boot IRQs cannot be disabled.
4175 The opposite of ioapicreroute.
4176 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4177 routing table. These calls are known to be buggy
4178 on several machines and they hang the machine
4179 when used, but on other computers it's the only
4180 way to get the interrupt routing table. Try
4181 this option if the kernel is unable to allocate
4182 IRQs or discover secondary PCI buses on your
4184 rom [X86] Assign address space to expansion ROMs.
4185 Use with caution as certain devices share
4186 address decoders between ROMs and other
4188 norom [X86] Do not assign address space to
4189 expansion ROMs that do not already have
4190 BIOS assigned address ranges.
4191 nobar [X86] Do not assign address space to the
4192 BARs that weren't assigned by the BIOS.
4193 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4194 assigned automatically to PCI devices. You can
4195 make the kernel exclude IRQs of your ISA cards
4197 pirqaddr=0xAAAAA [X86] Specify the physical address
4198 of the PIRQ table (normally generated
4199 by the BIOS) if it is outside the
4200 F0000h-100000h range.
4201 lastbus=N [X86] Scan all buses thru bus #N. Can be
4202 useful if the kernel is unable to find your
4203 secondary buses and you want to tell it
4204 explicitly which ones they are.
4205 assign-busses [X86] Always assign all PCI bus
4206 numbers ourselves, overriding
4207 whatever the firmware may have done.
4208 usepirqmask [X86] Honor the possible IRQ mask stored
4209 in the BIOS $PIR table. This is needed on
4210 some systems with broken BIOSes, notably
4211 some HP Pavilion N5400 and Omnibook XE3
4212 notebooks. This will have no effect if ACPI
4213 IRQ routing is enabled.
4214 noacpi [X86] Do not use ACPI for IRQ routing
4215 or for PCI scanning.
4216 use_crs [X86] Use PCI host bridge window information
4217 from ACPI. On BIOSes from 2008 or later, this
4218 is enabled by default. If you need to use this,
4219 please report a bug.
4220 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4221 If you need to use this, please report a bug.
4222 use_e820 [X86] Use E820 reservations to exclude parts of
4223 PCI host bridge windows. This is a workaround
4224 for BIOS defects in host bridge _CRS methods.
4225 If you need to use this, please report a bug to
4226 <linux-pci@vger.kernel.org>.
4227 no_e820 [X86] Ignore E820 reservations for PCI host
4228 bridge windows. This is the default on modern
4229 hardware. If you need to use this, please report
4230 a bug to <linux-pci@vger.kernel.org>.
4231 routeirq Do IRQ routing for all PCI devices.
4232 This is normally done in pci_enable_device(),
4233 so this option is a temporary workaround
4234 for broken drivers that don't call it.
4235 skip_isa_align [X86] do not align io start addr, so can
4236 handle more pci cards
4237 noearly [X86] Don't do any early type 1 scanning.
4238 This might help on some broken boards which
4239 machine check when some devices' config space
4240 is read. But various workarounds are disabled
4241 and some IOMMU drivers will not work.
4242 bfsort Sort PCI devices into breadth-first order.
4243 This sorting is done to get a device
4244 order compatible with older (<= 2.4) kernels.
4245 nobfsort Don't sort PCI devices into breadth-first order.
4246 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4247 tuning and use the BIOS-configured MPS defaults.
4248 pcie_bus_safe Set every device's MPS to the largest value
4249 supported by all devices below the root complex.
4250 pcie_bus_perf Set device MPS to the largest allowable MPS
4251 based on its parent bus. Also set MRRS (Max
4252 Read Request Size) to the largest supported
4253 value (no larger than the MPS that the device
4254 or bus can support) for best performance.
4255 pcie_bus_peer2peer Set every device's MPS to 128B, which
4256 every device is guaranteed to support. This
4257 configuration allows peer-to-peer DMA between
4258 any pair of devices, possibly at the cost of
4259 reduced performance. This also guarantees
4260 that hot-added devices will work.
4261 cbiosize=nn[KMG] The fixed amount of bus space which is
4262 reserved for the CardBus bridge's IO window.
4263 The default value is 256 bytes.
4264 cbmemsize=nn[KMG] The fixed amount of bus space which is
4265 reserved for the CardBus bridge's memory
4266 window. The default value is 64 megabytes.
4269 [<order of align>@]<pci_dev>[; ...]
4270 Specifies alignment and device to reassign
4271 aligned memory resources. How to
4272 specify the device is described above.
4273 If <order of align> is not specified,
4274 PAGE_SIZE is used as alignment.
4275 A PCI-PCI bridge can be specified if resource
4276 windows need to be expanded.
4277 To specify the alignment for several
4278 instances of a device, the PCI vendor,
4279 device, subvendor, and subdevice may be
4280 specified, e.g., 12@pci:8086:9c22:103c:198f
4281 for 4096-byte alignment.
4282 ecrc= Enable/disable PCIe ECRC (transaction layer
4283 end-to-end CRC checking).
4284 bios: Use BIOS/firmware settings. This is the
4288 hpiosize=nn[KMG] The fixed amount of bus space which is
4289 reserved for hotplug bridge's IO window.
4290 Default size is 256 bytes.
4291 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4292 reserved for hotplug bridge's MMIO window.
4293 Default size is 2 megabytes.
4294 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4295 reserved for hotplug bridge's MMIO_PREF window.
4296 Default size is 2 megabytes.
4297 hpmemsize=nn[KMG] The fixed amount of bus space which is
4298 reserved for hotplug bridge's MMIO and
4300 Default size is 2 megabytes.
4301 hpbussize=nn The minimum amount of additional bus numbers
4302 reserved for buses below a hotplug bridge.
4304 realloc= Enable/disable reallocating PCI bridge resources
4305 if allocations done by BIOS are too small to
4306 accommodate resources required by all child
4308 off: Turn realloc off
4310 realloc same as realloc=on
4311 noari do not use PCIe ARI.
4312 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4313 do not use PCIe ATS (and IOMMU device IOTLB).
4314 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4315 only look for one device below a PCIe downstream
4317 big_root_window Try to add a big 64bit memory window to the PCIe
4318 root complex on AMD CPUs. Some GFX hardware
4319 can resize a BAR to allow access to all VRAM.
4320 Adding the window is slightly risky (it may
4321 conflict with unreported devices), so this
4323 disable_acs_redir=<pci_dev>[; ...]
4324 Specify one or more PCI devices (in the format
4325 specified above) separated by semicolons.
4326 Each device specified will have the PCI ACS
4327 redirect capabilities forced off which will
4328 allow P2P traffic between devices through
4329 bridges without forcing it upstream. Note:
4330 this removes isolation between devices and
4331 may put more devices in an IOMMU group.
4332 force_floating [S390] Force usage of floating interrupts.
4333 nomio [S390] Do not use MIO instructions.
4334 norid [S390] ignore the RID field and force use of
4335 one PCI domain per PCI function
4337 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4340 force Enable ASPM even on devices that claim not to support it.
4341 WARNING: Forcing ASPM on may cause system lockups.
4343 pcie_ports= [PCIE] PCIe port services handling:
4344 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4345 even if the platform doesn't give the OS permission to
4346 use them. This may cause conflicts if the platform
4347 also tries to use these services.
4348 dpc-native Use native PCIe service for DPC only. May
4349 cause conflicts if firmware uses AER or DPC.
4350 compat Disable native PCIe services (PME, AER, DPC, PCIe
4353 pcie_port_pm= [PCIE] PCIe port power management handling:
4354 off Disable power management of all PCIe ports
4355 force Forcibly enable power management of all PCIe ports
4357 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4358 nomsi Do not use MSI for native PCIe PME signaling (this makes
4359 all PCIe root ports use INTx for all services).
4361 pcmv= [HW,PCMCIA] BadgePAD 4
4365 Keep all power-domains already enabled by bootloader on,
4366 even if no driver has claimed them. This is useful
4367 for debug and development, but should not be
4368 needed on a platform with proper driver support.
4371 See Documentation/admin-guide/blockdev/paride.rst.
4373 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4376 See arch/parisc/kernel/pdc_chassis.c
4378 percpu_alloc= Select which percpu first chunk allocator to use.
4379 Currently supported values are "embed" and "page".
4380 Archs may support subset or none of the selections.
4381 See comments in mm/percpu.c for details on each
4382 allocator. This parameter is primarily for debugging
4383 and performance comparison.
4386 See Documentation/admin-guide/blockdev/paride.rst.
4389 See Documentation/admin-guide/blockdev/paride.rst.
4391 pirq= [SMP,APIC] Manual mp-table setup
4392 See Documentation/x86/i386/IO-APIC.rst.
4394 plip= [PPT,NET] Parallel port network link
4395 Format: { parport<nr> | timid | 0 }
4396 See also Documentation/admin-guide/parport.rst.
4398 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4399 Override pmtimer IOPort with a hex value.
4402 pmu_override= [PPC] Override the PMU.
4403 This option takes over the PMU facility, so it is no
4404 longer usable by perf. Setting this option starts the
4405 PMU counters by setting MMCR0 to 0 (the FC bit is
4406 cleared). If a number is given, then MMCR1 is set to
4407 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4410 pm_debug_messages [SUSPEND,KNL]
4411 Enable suspend/resume debug messages during boot up.
4414 Enable PNP debug messages (depends on the
4415 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4416 via /sys/module/pnp/parameters/debug. We always show
4417 current resource usage; turning this on also shows
4418 possible settings and some assignment information.
4424 { on | off | curr | res | no-curr | no-res }
4427 [ISAPNP] Exclude IRQs for the autoconfiguration
4430 [ISAPNP] Exclude DMAs for the autoconfiguration
4432 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4433 Ranges are in pairs (I/O port base and size).
4436 [ISAPNP] Exclude memory regions for the
4438 Ranges are in pairs (memory base and size).
4440 ports= [IP_VS_FTP] IPVS ftp helper module
4442 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4444 Format: <port>,<port>....
4446 powersave=off [PPC] This option disables power saving features.
4447 It specifically disables cpuidle and sets the
4448 platform machine description specific power_save
4449 function to NULL. On Idle the CPU just reduces
4452 ppc_strict_facility_enable
4453 [PPC] This option catches any kernel floating point,
4454 Altivec, VSX and SPE outside of regions specifically
4455 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4456 There is some performance impact when enabling this.
4460 Disable Hardware Transactional Memory
4463 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4464 none - Limited to cond_resched() calls
4465 voluntary - Limited to cond_resched() and might_sleep() calls
4466 full - Any section that isn't explicitly preempt disabled
4467 can be preempted anytime.
4469 print-fatal-signals=
4470 [KNL] debug: print fatal signals
4472 If enabled, warn about various signal handling
4473 related application anomalies: too many signals,
4474 too many POSIX.1 timers, fatal signals causing a
4477 If you hit the warning due to signal overflow,
4478 you might want to try "ulimit -i unlimited".
4482 printk.always_kmsg_dump=
4483 Trigger kmsg_dump for cases other than kernel oops or
4485 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4488 printk.console_no_auto_verbose=
4489 Disable console loglevel raise on oops, panic
4490 or lockdep-detected issues (only if lock debug is on).
4491 With an exception to setups with low baudrate on
4492 serial console, keeping this 0 is a good choice
4493 in order to provide more debug information.
4495 default: 0 (auto_verbose is enabled)
4497 printk.devkmsg={on,off,ratelimit}
4498 Control writing to /dev/kmsg.
4499 on - unlimited logging to /dev/kmsg from userspace
4500 off - logging to /dev/kmsg disabled
4501 ratelimit - ratelimit the logging
4504 printk.time= Show timing data prefixed to each printk message line
4505 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4507 processor.max_cstate= [HW,ACPI]
4508 Limit processor to maximum C-state
4509 max_cstate=9 overrides any DMI blacklist limit.
4511 processor.nocst [HW,ACPI]
4512 Ignore the _CST method to determine C-states,
4513 instead using the legacy FADT method
4515 profile= [KNL] Enable kernel profiling via /proc/profile
4516 Format: [<profiletype>,]<number>
4517 Param: <profiletype>: "schedule", "sleep", or "kvm"
4518 [defaults to kernel profiling]
4519 Param: "schedule" - profile schedule points.
4520 Param: "sleep" - profile D-state sleeping (millisecs).
4521 Requires CONFIG_SCHEDSTATS
4522 Param: "kvm" - profile VM exits.
4523 Param: <number> - step/bucket size as a power of 2 for
4524 statistical time based profiling.
4526 prompt_ramdisk= [RAM] [Deprecated]
4528 prot_virt= [S390] enable hosting protected virtual machines
4529 isolated from the hypervisor (if hardware supports
4533 psi= [KNL] Enable or disable pressure stall information
4537 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4538 probe for; one of (bare|imps|exps|lifebook|any).
4539 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4541 psmouse.resetafter= [HW,MOUSE]
4542 Try to reset the device after so many bad packets
4545 [HW,MOUSE] Set desired mouse resolution, in dpi.
4546 psmouse.smartscroll=
4547 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4548 0 = disabled, 1 = enabled (default).
4550 pstore.backend= Specify the name of the pstore backend to use
4553 See Documentation/admin-guide/blockdev/paride.rst.
4555 pti= [X86-64] Control Page Table Isolation of user and
4556 kernel address spaces. Disabling this feature
4557 removes hardening, but improves performance of
4558 system calls and interrupts.
4560 on - unconditionally enable
4561 off - unconditionally disable
4562 auto - kernel detects whether your CPU model is
4563 vulnerable to issues that PTI mitigates
4565 Not specifying this option is equivalent to pti=auto.
4568 Equivalent to pti=off
4571 [KNL] Number of legacy pty's. Overwrites compiled-in
4574 quiet [KNL] Disable most log messages
4579 See Documentation/admin-guide/md.rst.
4581 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4582 See Documentation/admin-guide/blockdev/ramdisk.rst.
4584 ramdisk_start= [RAM] RAM disk image start address
4586 random.trust_cpu={on,off}
4587 [KNL] Enable or disable trusting the use of the
4588 CPU's random number generator (if available) to
4589 fully seed the kernel's CRNG. Default is controlled
4590 by CONFIG_RANDOM_TRUST_CPU.
4592 random.trust_bootloader={on,off}
4593 [KNL] Enable or disable trusting the use of a
4594 seed passed by the bootloader (if available) to
4595 fully seed the kernel's CRNG. Default is controlled
4596 by CONFIG_RANDOM_TRUST_BOOTLOADER.
4598 randomize_kstack_offset=
4599 [KNL] Enable or disable kernel stack offset
4600 randomization, which provides roughly 5 bits of
4601 entropy, frustrating memory corruption attacks
4602 that depend on stack address determinism or
4603 cross-syscall address exposures. This is only
4604 available on architectures that have defined
4605 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4606 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4607 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4609 ras=option[,option,...] [KNL] RAS-specific options
4612 Disable the Correctable Errors Collector,
4613 see CONFIG_RAS_CEC help text.
4615 rcu_nocbs[=cpu-list]
4616 [KNL] The optional argument is a cpu list,
4619 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4620 enable the no-callback CPU mode, which prevents
4621 such CPUs' callbacks from being invoked in
4622 softirq context. Invocation of such CPUs' RCU
4623 callbacks will instead be offloaded to "rcuox/N"
4624 kthreads created for that purpose, where "x" is
4625 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4626 for the kthreads that mediate grace periods; and
4627 "N" is the CPU number. This reduces OS jitter on
4628 the offloaded CPUs, which can be useful for HPC
4629 and real-time workloads. It can also improve
4630 energy efficiency for asymmetric multiprocessors.
4632 If a cpulist is passed as an argument, the specified
4633 list of CPUs is set to no-callback mode from boot.
4635 Otherwise, if the '=' sign and the cpulist
4636 arguments are omitted, no CPU will be set to
4637 no-callback mode from boot but the mode may be
4638 toggled at runtime via cpusets.
4640 Note that this argument takes precedence over
4641 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4644 Rather than requiring that offloaded CPUs
4645 (specified by rcu_nocbs= above) explicitly
4646 awaken the corresponding "rcuoN" kthreads,
4647 make these kthreads poll for callbacks.
4648 This improves the real-time response for the
4649 offloaded CPUs by relieving them of the need to
4650 wake up the corresponding kthread, but degrades
4651 energy efficiency by requiring that the kthreads
4652 periodically wake up to do the polling.
4654 rcutree.blimit= [KNL]
4655 Set maximum number of finished RCU callbacks to
4656 process in one batch.
4658 rcutree.dump_tree= [KNL]
4659 Dump the structure of the rcu_node combining tree
4660 out at early boot. This is used for diagnostic
4661 purposes, to verify correct tree setup.
4663 rcutree.gp_cleanup_delay= [KNL]
4664 Set the number of jiffies to delay each step of
4665 RCU grace-period cleanup.
4667 rcutree.gp_init_delay= [KNL]
4668 Set the number of jiffies to delay each step of
4669 RCU grace-period initialization.
4671 rcutree.gp_preinit_delay= [KNL]
4672 Set the number of jiffies to delay each step of
4673 RCU grace-period pre-initialization, that is,
4674 the propagation of recent CPU-hotplug changes up
4675 the rcu_node combining tree.
4677 rcutree.use_softirq= [KNL]
4678 If set to zero, move all RCU_SOFTIRQ processing to
4679 per-CPU rcuc kthreads. Defaults to a non-zero
4680 value, meaning that RCU_SOFTIRQ is used by default.
4681 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4683 But note that CONFIG_PREEMPT_RT=y kernels disable
4684 this kernel boot parameter, forcibly setting it
4687 rcutree.rcu_fanout_exact= [KNL]
4688 Disable autobalancing of the rcu_node combining
4689 tree. This is used by rcutorture, and might
4690 possibly be useful for architectures having high
4691 cache-to-cache transfer latencies.
4693 rcutree.rcu_fanout_leaf= [KNL]
4694 Change the number of CPUs assigned to each
4695 leaf rcu_node structure. Useful for very
4696 large systems, which will choose the value 64,
4697 and for NUMA systems with large remote-access
4698 latencies, which will choose a value aligned
4699 with the appropriate hardware boundaries.
4701 rcutree.rcu_min_cached_objs= [KNL]
4702 Minimum number of objects which are cached and
4703 maintained per one CPU. Object size is equal
4704 to PAGE_SIZE. The cache allows to reduce the
4705 pressure to page allocator, also it makes the
4706 whole algorithm to behave better in low memory
4709 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4710 Set the page-cache refill delay (in milliseconds)
4711 in response to low-memory conditions. The range
4712 of permitted values is in the range 0:100000.
4714 rcutree.jiffies_till_first_fqs= [KNL]
4715 Set delay from grace-period initialization to
4716 first attempt to force quiescent states.
4717 Units are jiffies, minimum value is zero,
4718 and maximum value is HZ.
4720 rcutree.jiffies_till_next_fqs= [KNL]
4721 Set delay between subsequent attempts to force
4722 quiescent states. Units are jiffies, minimum
4723 value is one, and maximum value is HZ.
4725 rcutree.jiffies_till_sched_qs= [KNL]
4726 Set required age in jiffies for a
4727 given grace period before RCU starts
4728 soliciting quiescent-state help from
4729 rcu_note_context_switch() and cond_resched().
4730 If not specified, the kernel will calculate
4731 a value based on the most recent settings
4732 of rcutree.jiffies_till_first_fqs
4733 and rcutree.jiffies_till_next_fqs.
4734 This calculated value may be viewed in
4735 rcutree.jiffies_to_sched_qs. Any attempt to set
4736 rcutree.jiffies_to_sched_qs will be cheerfully
4739 rcutree.kthread_prio= [KNL,BOOT]
4740 Set the SCHED_FIFO priority of the RCU per-CPU
4741 kthreads (rcuc/N). This value is also used for
4742 the priority of the RCU boost threads (rcub/N)
4743 and for the RCU grace-period kthreads (rcu_bh,
4744 rcu_preempt, and rcu_sched). If RCU_BOOST is
4745 set, valid values are 1-99 and the default is 1
4746 (the least-favored priority). Otherwise, when
4747 RCU_BOOST is not set, valid values are 0-99 and
4748 the default is zero (non-realtime operation).
4749 When RCU_NOCB_CPU is set, also adjust the
4750 priority of NOCB callback kthreads.
4752 rcutree.rcu_divisor= [KNL]
4753 Set the shift-right count to use to compute
4754 the callback-invocation batch limit bl from
4755 the number of callbacks queued on this CPU.
4756 The result will be bounded below by the value of
4757 the rcutree.blimit kernel parameter. Every bl
4758 callbacks, the softirq handler will exit in
4759 order to allow the CPU to do other work.
4761 Please note that this callback-invocation batch
4762 limit applies only to non-offloaded callback
4763 invocation. Offloaded callbacks are instead
4764 invoked in the context of an rcuoc kthread, which
4765 scheduler will preempt as it does any other task.
4767 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4768 On callback-offloaded (rcu_nocbs) CPUs,
4769 RCU reduces the lock contention that would
4770 otherwise be caused by callback floods through
4771 use of the ->nocb_bypass list. However, in the
4772 common non-flooded case, RCU queues directly to
4773 the main ->cblist in order to avoid the extra
4774 overhead of the ->nocb_bypass list and its lock.
4775 But if there are too many callbacks queued during
4776 a single jiffy, RCU pre-queues the callbacks into
4777 the ->nocb_bypass queue. The definition of "too
4778 many" is supplied by this kernel boot parameter.
4780 rcutree.rcu_nocb_gp_stride= [KNL]
4781 Set the number of NOCB callback kthreads in
4782 each group, which defaults to the square root
4783 of the number of CPUs. Larger numbers reduce
4784 the wakeup overhead on the global grace-period
4785 kthread, but increases that same overhead on
4786 each group's NOCB grace-period kthread.
4788 rcutree.qhimark= [KNL]
4789 Set threshold of queued RCU callbacks beyond which
4790 batch limiting is disabled.
4792 rcutree.qlowmark= [KNL]
4793 Set threshold of queued RCU callbacks below which
4794 batch limiting is re-enabled.
4796 rcutree.qovld= [KNL]
4797 Set threshold of queued RCU callbacks beyond which
4798 RCU's force-quiescent-state scan will aggressively
4799 enlist help from cond_resched() and sched IPIs to
4800 help CPUs more quickly reach quiescent states.
4801 Set to less than zero to make this be set based
4802 on rcutree.qhimark at boot time and to zero to
4803 disable more aggressive help enlistment.
4805 rcutree.rcu_kick_kthreads= [KNL]
4806 Cause the grace-period kthread to get an extra
4807 wake_up() if it sleeps three times longer than
4808 it should at force-quiescent-state time.
4809 This wake_up() will be accompanied by a
4810 WARN_ONCE() splat and an ftrace_dump().
4812 rcutree.rcu_unlock_delay= [KNL]
4813 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4814 this specifies an rcu_read_unlock()-time delay
4815 in microseconds. This defaults to zero.
4816 Larger delays increase the probability of
4817 catching RCU pointer leaks, that is, buggy use
4818 of RCU-protected pointers after the relevant
4819 rcu_read_unlock() has completed.
4821 rcutree.sysrq_rcu= [KNL]
4822 Commandeer a sysrq key to dump out Tree RCU's
4823 rcu_node tree with an eye towards determining
4824 why a new grace period has not yet started.
4826 rcuscale.gp_async= [KNL]
4827 Measure performance of asynchronous
4828 grace-period primitives such as call_rcu().
4830 rcuscale.gp_async_max= [KNL]
4831 Specify the maximum number of outstanding
4832 callbacks per writer thread. When a writer
4833 thread exceeds this limit, it invokes the
4834 corresponding flavor of rcu_barrier() to allow
4835 previously posted callbacks to drain.
4837 rcuscale.gp_exp= [KNL]
4838 Measure performance of expedited synchronous
4839 grace-period primitives.
4841 rcuscale.holdoff= [KNL]
4842 Set test-start holdoff period. The purpose of
4843 this parameter is to delay the start of the
4844 test until boot completes in order to avoid
4847 rcuscale.kfree_rcu_test= [KNL]
4848 Set to measure performance of kfree_rcu() flooding.
4850 rcuscale.kfree_rcu_test_double= [KNL]
4851 Test the double-argument variant of kfree_rcu().
4852 If this parameter has the same value as
4853 rcuscale.kfree_rcu_test_single, both the single-
4854 and double-argument variants are tested.
4856 rcuscale.kfree_rcu_test_single= [KNL]
4857 Test the single-argument variant of kfree_rcu().
4858 If this parameter has the same value as
4859 rcuscale.kfree_rcu_test_double, both the single-
4860 and double-argument variants are tested.
4862 rcuscale.kfree_nthreads= [KNL]
4863 The number of threads running loops of kfree_rcu().
4865 rcuscale.kfree_alloc_num= [KNL]
4866 Number of allocations and frees done in an iteration.
4868 rcuscale.kfree_loops= [KNL]
4869 Number of loops doing rcuscale.kfree_alloc_num number
4870 of allocations and frees.
4872 rcuscale.nreaders= [KNL]
4873 Set number of RCU readers. The value -1 selects
4874 N, where N is the number of CPUs. A value
4875 "n" less than -1 selects N-n+1, where N is again
4876 the number of CPUs. For example, -2 selects N
4877 (the number of CPUs), -3 selects N+1, and so on.
4878 A value of "n" less than or equal to -N selects
4881 rcuscale.nwriters= [KNL]
4882 Set number of RCU writers. The values operate
4883 the same as for rcuscale.nreaders.
4884 N, where N is the number of CPUs
4886 rcuscale.perf_type= [KNL]
4887 Specify the RCU implementation to test.
4889 rcuscale.shutdown= [KNL]
4890 Shut the system down after performance tests
4891 complete. This is useful for hands-off automated
4894 rcuscale.verbose= [KNL]
4895 Enable additional printk() statements.
4897 rcuscale.writer_holdoff= [KNL]
4898 Write-side holdoff between grace periods,
4899 in microseconds. The default of zero says
4902 rcutorture.fqs_duration= [KNL]
4903 Set duration of force_quiescent_state bursts
4906 rcutorture.fqs_holdoff= [KNL]
4907 Set holdoff time within force_quiescent_state bursts
4910 rcutorture.fqs_stutter= [KNL]
4911 Set wait time between force_quiescent_state bursts
4914 rcutorture.fwd_progress= [KNL]
4915 Specifies the number of kthreads to be used
4916 for RCU grace-period forward-progress testing
4917 for the types of RCU supporting this notion.
4918 Defaults to 1 kthread, values less than zero or
4919 greater than the number of CPUs cause the number
4922 rcutorture.fwd_progress_div= [KNL]
4923 Specify the fraction of a CPU-stall-warning
4924 period to do tight-loop forward-progress testing.
4926 rcutorture.fwd_progress_holdoff= [KNL]
4927 Number of seconds to wait between successive
4928 forward-progress tests.
4930 rcutorture.fwd_progress_need_resched= [KNL]
4931 Enclose cond_resched() calls within checks for
4932 need_resched() during tight-loop forward-progress
4935 rcutorture.gp_cond= [KNL]
4936 Use conditional/asynchronous update-side
4937 primitives, if available.
4939 rcutorture.gp_exp= [KNL]
4940 Use expedited update-side primitives, if available.
4942 rcutorture.gp_normal= [KNL]
4943 Use normal (non-expedited) asynchronous
4944 update-side primitives, if available.
4946 rcutorture.gp_sync= [KNL]
4947 Use normal (non-expedited) synchronous
4948 update-side primitives, if available. If all
4949 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4950 rcutorture.gp_normal=, and rcutorture.gp_sync=
4951 are zero, rcutorture acts as if is interpreted
4952 they are all non-zero.
4954 rcutorture.irqreader= [KNL]
4955 Run RCU readers from irq handlers, or, more
4956 accurately, from a timer handler. Not all RCU
4957 flavors take kindly to this sort of thing.
4959 rcutorture.leakpointer= [KNL]
4960 Leak an RCU-protected pointer out of the reader.
4961 This can of course result in splats, and is
4962 intended to test the ability of things like
4963 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4966 rcutorture.n_barrier_cbs= [KNL]
4967 Set callbacks/threads for rcu_barrier() testing.
4969 rcutorture.nfakewriters= [KNL]
4970 Set number of concurrent RCU writers. These just
4971 stress RCU, they don't participate in the actual
4972 test, hence the "fake".
4974 rcutorture.nocbs_nthreads= [KNL]
4975 Set number of RCU callback-offload togglers.
4976 Zero (the default) disables toggling.
4978 rcutorture.nocbs_toggle= [KNL]
4979 Set the delay in milliseconds between successive
4980 callback-offload toggling attempts.
4982 rcutorture.nreaders= [KNL]
4983 Set number of RCU readers. The value -1 selects
4984 N-1, where N is the number of CPUs. A value
4985 "n" less than -1 selects N-n-2, where N is again
4986 the number of CPUs. For example, -2 selects N
4987 (the number of CPUs), -3 selects N+1, and so on.
4989 rcutorture.object_debug= [KNL]
4990 Enable debug-object double-call_rcu() testing.
4992 rcutorture.onoff_holdoff= [KNL]
4993 Set time (s) after boot for CPU-hotplug testing.
4995 rcutorture.onoff_interval= [KNL]
4996 Set time (jiffies) between CPU-hotplug operations,
4997 or zero to disable CPU-hotplug testing.
4999 rcutorture.read_exit= [KNL]
5000 Set the number of read-then-exit kthreads used
5001 to test the interaction of RCU updaters and
5002 task-exit processing.
5004 rcutorture.read_exit_burst= [KNL]
5005 The number of times in a given read-then-exit
5006 episode that a set of read-then-exit kthreads
5009 rcutorture.read_exit_delay= [KNL]
5010 The delay, in seconds, between successive
5011 read-then-exit testing episodes.
5013 rcutorture.shuffle_interval= [KNL]
5014 Set task-shuffle interval (s). Shuffling tasks
5015 allows some CPUs to go into dyntick-idle mode
5016 during the rcutorture test.
5018 rcutorture.shutdown_secs= [KNL]
5019 Set time (s) after boot system shutdown. This
5020 is useful for hands-off automated testing.
5022 rcutorture.stall_cpu= [KNL]
5023 Duration of CPU stall (s) to test RCU CPU stall
5024 warnings, zero to disable.
5026 rcutorture.stall_cpu_block= [KNL]
5027 Sleep while stalling if set. This will result
5028 in warnings from preemptible RCU in addition
5029 to any other stall-related activity.
5031 rcutorture.stall_cpu_holdoff= [KNL]
5032 Time to wait (s) after boot before inducing stall.
5034 rcutorture.stall_cpu_irqsoff= [KNL]
5035 Disable interrupts while stalling if set.
5037 rcutorture.stall_gp_kthread= [KNL]
5038 Duration (s) of forced sleep within RCU
5039 grace-period kthread to test RCU CPU stall
5040 warnings, zero to disable. If both stall_cpu
5041 and stall_gp_kthread are specified, the
5042 kthread is starved first, then the CPU.
5044 rcutorture.stat_interval= [KNL]
5045 Time (s) between statistics printk()s.
5047 rcutorture.stutter= [KNL]
5048 Time (s) to stutter testing, for example, specifying
5049 five seconds causes the test to run for five seconds,
5050 wait for five seconds, and so on. This tests RCU's
5051 ability to transition abruptly to and from idle.
5053 rcutorture.test_boost= [KNL]
5054 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5055 "Maybe" means test if the RCU implementation
5056 under test support RCU priority boosting.
5058 rcutorture.test_boost_duration= [KNL]
5059 Duration (s) of each individual boost test.
5061 rcutorture.test_boost_interval= [KNL]
5062 Interval (s) between each boost test.
5064 rcutorture.test_no_idle_hz= [KNL]
5065 Test RCU's dyntick-idle handling. See also the
5066 rcutorture.shuffle_interval parameter.
5068 rcutorture.torture_type= [KNL]
5069 Specify the RCU implementation to test.
5071 rcutorture.verbose= [KNL]
5072 Enable additional printk() statements.
5074 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5075 Dump ftrace buffer after reporting RCU CPU
5078 rcupdate.rcu_cpu_stall_suppress= [KNL]
5079 Suppress RCU CPU stall warning messages.
5081 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5082 Suppress RCU CPU stall warning messages and
5083 rcutorture writer stall warnings that occur
5084 during early boot, that is, during the time
5085 before the init task is spawned.
5087 rcupdate.rcu_cpu_stall_timeout= [KNL]
5088 Set timeout for RCU CPU stall warning messages.
5089 The value is in seconds and the maximum allowed
5090 value is 300 seconds.
5092 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5093 Set timeout for expedited RCU CPU stall warning
5094 messages. The value is in milliseconds
5095 and the maximum allowed value is 21000
5096 milliseconds. Please note that this value is
5097 adjusted to an arch timer tick resolution.
5098 Setting this to zero causes the value from
5099 rcupdate.rcu_cpu_stall_timeout to be used (after
5100 conversion from seconds to milliseconds).
5102 rcupdate.rcu_expedited= [KNL]
5103 Use expedited grace-period primitives, for
5104 example, synchronize_rcu_expedited() instead
5105 of synchronize_rcu(). This reduces latency,
5106 but can increase CPU utilization, degrade
5107 real-time latency, and degrade energy efficiency.
5108 No effect on CONFIG_TINY_RCU kernels.
5110 rcupdate.rcu_normal= [KNL]
5111 Use only normal grace-period primitives,
5112 for example, synchronize_rcu() instead of
5113 synchronize_rcu_expedited(). This improves
5114 real-time latency, CPU utilization, and
5115 energy efficiency, but can expose users to
5116 increased grace-period latency. This parameter
5117 overrides rcupdate.rcu_expedited. No effect on
5118 CONFIG_TINY_RCU kernels.
5120 rcupdate.rcu_normal_after_boot= [KNL]
5121 Once boot has completed (that is, after
5122 rcu_end_inkernel_boot() has been invoked), use
5123 only normal grace-period primitives. No effect
5124 on CONFIG_TINY_RCU kernels.
5126 But note that CONFIG_PREEMPT_RT=y kernels enables
5127 this kernel boot parameter, forcibly setting
5128 it to the value one, that is, converting any
5129 post-boot attempt at an expedited RCU grace
5130 period to instead use normal non-expedited
5131 grace-period processing.
5133 rcupdate.rcu_task_collapse_lim= [KNL]
5134 Set the maximum number of callbacks present
5135 at the beginning of a grace period that allows
5136 the RCU Tasks flavors to collapse back to using
5137 a single callback queue. This switching only
5138 occurs when rcupdate.rcu_task_enqueue_lim is
5139 set to the default value of -1.
5141 rcupdate.rcu_task_contend_lim= [KNL]
5142 Set the minimum number of callback-queuing-time
5143 lock-contention events per jiffy required to
5144 cause the RCU Tasks flavors to switch to per-CPU
5145 callback queuing. This switching only occurs
5146 when rcupdate.rcu_task_enqueue_lim is set to
5147 the default value of -1.
5149 rcupdate.rcu_task_enqueue_lim= [KNL]
5150 Set the number of callback queues to use for the
5151 RCU Tasks family of RCU flavors. The default
5152 of -1 allows this to be automatically (and
5153 dynamically) adjusted. This parameter is intended
5156 rcupdate.rcu_task_ipi_delay= [KNL]
5157 Set time in jiffies during which RCU tasks will
5158 avoid sending IPIs, starting with the beginning
5159 of a given grace period. Setting a large
5160 number avoids disturbing real-time workloads,
5161 but lengthens grace periods.
5163 rcupdate.rcu_task_stall_info= [KNL]
5164 Set initial timeout in jiffies for RCU task stall
5165 informational messages, which give some indication
5166 of the problem for those not patient enough to
5167 wait for ten minutes. Informational messages are
5168 only printed prior to the stall-warning message
5169 for a given grace period. Disable with a value
5170 less than or equal to zero. Defaults to ten
5171 seconds. A change in value does not take effect
5172 until the beginning of the next grace period.
5174 rcupdate.rcu_task_stall_info_mult= [KNL]
5175 Multiplier for time interval between successive
5176 RCU task stall informational messages for a given
5177 RCU tasks grace period. This value is clamped
5178 to one through ten, inclusive. It defaults to
5179 the value three, so that the first informational
5180 message is printed 10 seconds into the grace
5181 period, the second at 40 seconds, the third at
5182 160 seconds, and then the stall warning at 600
5183 seconds would prevent a fourth at 640 seconds.
5185 rcupdate.rcu_task_stall_timeout= [KNL]
5186 Set timeout in jiffies for RCU task stall
5187 warning messages. Disable with a value less
5188 than or equal to zero. Defaults to ten minutes.
5189 A change in value does not take effect until
5190 the beginning of the next grace period.
5192 rcupdate.rcu_self_test= [KNL]
5193 Run the RCU early boot self tests
5197 Run specified binary instead of /init from the ramdisk,
5198 used for early userspace startup. See initrd.
5201 force - Override the decision by the kernel to hide the
5202 advertisement of RDRAND support (this affects
5203 certain AMD processors because of buggy BIOS
5204 support, specifically around the suspend/resume
5208 Turn on/off individual RDT features. List is:
5209 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5211 E.g. to turn on cmt and turn off mba use:
5215 Format (x86 or x86_64):
5216 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5218 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5220 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5221 (prefix with 'panic_' to set mode for panic
5223 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5224 reboot_force is either force or not specified,
5225 reboot_cpu is s[mp]#### with #### being the processor
5226 to be used for rebooting.
5228 refscale.holdoff= [KNL]
5229 Set test-start holdoff period. The purpose of
5230 this parameter is to delay the start of the
5231 test until boot completes in order to avoid
5234 refscale.loops= [KNL]
5235 Set the number of loops over the synchronization
5236 primitive under test. Increasing this number
5237 reduces noise due to loop start/end overhead,
5238 but the default has already reduced the per-pass
5239 noise to a handful of picoseconds on ca. 2020
5242 refscale.nreaders= [KNL]
5243 Set number of readers. The default value of -1
5244 selects N, where N is roughly 75% of the number
5245 of CPUs. A value of zero is an interesting choice.
5247 refscale.nruns= [KNL]
5248 Set number of runs, each of which is dumped onto
5251 refscale.readdelay= [KNL]
5252 Set the read-side critical-section duration,
5253 measured in microseconds.
5255 refscale.scale_type= [KNL]
5256 Specify the read-protection implementation to test.
5258 refscale.shutdown= [KNL]
5259 Shut down the system at the end of the performance
5260 test. This defaults to 1 (shut it down) when
5261 refscale is built into the kernel and to 0 (leave
5262 it running) when refscale is built as a module.
5264 refscale.verbose= [KNL]
5265 Enable additional printk() statements.
5267 refscale.verbose_batched= [KNL]
5268 Batch the additional printk() statements. If zero
5269 (the default) or negative, print everything. Otherwise,
5270 print every Nth verbose statement, where N is the value
5274 [KNL, SMP] Set scheduler's default relax_domain_level.
5275 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5277 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5278 Format: <base1>,<size1>[,<base2>,<size2>,...]
5279 Reserve I/O ports or memory so the kernel won't use
5280 them. If <base> is less than 0x10000, the region
5281 is assumed to be I/O ports; otherwise it is memory.
5283 reservetop= [X86-32]
5285 Reserves a hole at the top of the kernel virtual
5288 reset_devices [KNL] Force drivers to reset the underlying device
5289 during initialization.
5292 Specify the partition device for software suspend
5294 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5296 resume_offset= [SWSUSP]
5297 Specify the offset from the beginning of the partition
5298 given by "resume=" at which the swap header is located,
5299 in <PAGE_SIZE> units (needed only for swap files).
5300 See Documentation/power/swsusp-and-swap-files.rst
5302 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5303 read the resume files
5305 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5306 Useful for devices that are detected asynchronously
5307 (e.g. USB and MMC devices).
5309 retain_initrd [RAM] Keep initrd memory after extraction
5311 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5312 Speculative Code Execution with Return Instructions)
5315 AMD-based UNRET and IBPB mitigations alone do not stop
5316 sibling threads from influencing the predictions of other
5317 sibling threads. For that reason, STIBP is used on pro-
5318 cessors that support it, and mitigate SMT on processors
5322 auto - automatically select a migitation
5323 auto,nosmt - automatically select a mitigation,
5324 disabling SMT if necessary for
5325 the full mitigation (only on Zen1
5326 and older without STIBP).
5327 ibpb - On AMD, mitigate short speculation
5328 windows on basic block boundaries too.
5329 Safe, highest perf impact. It also
5330 enables STIBP if present. Not suitable
5332 ibpb,nosmt - Like "ibpb" above but will disable SMT
5333 when STIBP is not available. This is
5334 the alternative for systems which do not
5336 unret - Force enable untrained return thunks,
5337 only effective on AMD f15h-f17h based
5339 unret,nosmt - Like unret, but will disable SMT when STIBP
5340 is not available. This is the alternative for
5341 systems which do not have STIBP.
5343 Selecting 'auto' will choose a mitigation method at run
5344 time according to the CPU.
5346 Not specifying this option is equivalent to retbleed=auto.
5348 rfkill.default_state=
5349 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5350 etc. communication is blocked by default.
5353 rfkill.master_switch_mode=
5354 0 The "airplane mode" button does nothing.
5355 1 The "airplane mode" button toggles between everything
5356 blocked and the previous configuration.
5357 2 The "airplane mode" button toggles between everything
5358 blocked and everything unblocked.
5360 rhash_entries= [KNL,NET]
5361 Set number of hash buckets for route cache
5364 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5367 ro [KNL] Mount root device read-only on boot
5370 on Mark read-only kernel memory as read-only (default).
5371 off Leave read-only kernel memory writable for debugging.
5372 full Mark read-only kernel memory and aliases as read-only
5376 Enable the uart passthrough on the designated usb port
5377 on Rockchip SoCs. When active, the signals of the
5378 debug-uart get routed to the D+ and D- pins of the usb
5379 port and the regular usb controller gets disabled.
5381 root= [KNL] Root filesystem
5382 See name_to_dev_t comment in init/do_mounts.c.
5384 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5385 mount the root filesystem
5387 rootflags= [KNL] Set root filesystem mount option string
5389 rootfstype= [KNL] Set root filesystem type
5391 rootwait [KNL] Wait (indefinitely) for root device to show up.
5392 Useful for devices that are detected asynchronously
5393 (e.g. USB and MMC devices).
5395 rproc_mem=nn[KMG][@address]
5396 [KNL,ARM,CMA] Remoteproc physical memory block.
5397 Memory area to be used by remote processor image,
5400 rw [KNL] Mount root device read-write on boot
5402 S [KNL] Run init in single mode
5404 s390_iommu= [HW,S390]
5405 Set s390 IOTLB flushing mode
5407 With strict flushing every unmap operation will result in
5408 an IOTLB flush. Default is lazy flushing before reuse,
5411 s390_iommu_aperture= [KNL,S390]
5412 Specifies the size of the per device DMA address space
5413 accessible through the DMA and IOMMU APIs as a decimal
5414 factor of the size of main memory.
5415 The default is 1 meaning that one can concurrently use
5416 as many DMA addresses as physical memory is installed,
5417 if supported by hardware, and thus map all of memory
5418 once. With a value of 2 one can map all of memory twice
5419 and so on. As a special case a factor of 0 imposes no
5420 restrictions other than those given by hardware at the
5421 cost of significant additional memory use for tables.
5424 See drivers/net/irda/sa1100_ir.c.
5426 sched_verbose [KNL] Enables verbose scheduler debug messages.
5428 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5429 Allowed values are enable and disable. This feature
5430 incurs a small amount of overhead in the scheduler
5431 but is useful for debugging and performance tuning.
5433 sched_thermal_decay_shift=
5434 [KNL, SMP] Set a decay shift for scheduler thermal
5435 pressure signal. Thermal pressure signal follows the
5436 default decay period of other scheduler pelt
5437 signals(usually 32 ms but configurable). Setting
5438 sched_thermal_decay_shift will left shift the decay
5439 period for the thermal pressure signal by the shift
5441 i.e. with the default pelt decay period of 32 ms
5442 sched_thermal_decay_shift thermal pressure decay pr
5446 Format: integer between 0 and 10
5449 scftorture.holdoff= [KNL]
5450 Number of seconds to hold off before starting
5451 test. Defaults to zero for module insertion and
5452 to 10 seconds for built-in smp_call_function()
5455 scftorture.longwait= [KNL]
5456 Request ridiculously long waits randomly selected
5457 up to the chosen limit in seconds. Zero (the
5458 default) disables this feature. Please note
5459 that requesting even small non-zero numbers of
5460 seconds can result in RCU CPU stall warnings,
5461 softlockup complaints, and so on.
5463 scftorture.nthreads= [KNL]
5464 Number of kthreads to spawn to invoke the
5465 smp_call_function() family of functions.
5466 The default of -1 specifies a number of kthreads
5467 equal to the number of CPUs.
5469 scftorture.onoff_holdoff= [KNL]
5470 Number seconds to wait after the start of the
5471 test before initiating CPU-hotplug operations.
5473 scftorture.onoff_interval= [KNL]
5474 Number seconds to wait between successive
5475 CPU-hotplug operations. Specifying zero (which
5476 is the default) disables CPU-hotplug operations.
5478 scftorture.shutdown_secs= [KNL]
5479 The number of seconds following the start of the
5480 test after which to shut down the system. The
5481 default of zero avoids shutting down the system.
5482 Non-zero values are useful for automated tests.
5484 scftorture.stat_interval= [KNL]
5485 The number of seconds between outputting the
5486 current test statistics to the console. A value
5487 of zero disables statistics output.
5489 scftorture.stutter_cpus= [KNL]
5490 The number of jiffies to wait between each change
5491 to the set of CPUs under test.
5493 scftorture.use_cpus_read_lock= [KNL]
5494 Use use_cpus_read_lock() instead of the default
5495 preempt_disable() to disable CPU hotplug
5496 while invoking one of the smp_call_function*()
5499 scftorture.verbose= [KNL]
5500 Enable additional printk() statements.
5502 scftorture.weight_single= [KNL]
5503 The probability weighting to use for the
5504 smp_call_function_single() function with a zero
5505 "wait" parameter. A value of -1 selects the
5506 default if all other weights are -1. However,
5507 if at least one weight has some other value, a
5508 value of -1 will instead select a weight of zero.
5510 scftorture.weight_single_wait= [KNL]
5511 The probability weighting to use for the
5512 smp_call_function_single() function with a
5513 non-zero "wait" parameter. See weight_single.
5515 scftorture.weight_many= [KNL]
5516 The probability weighting to use for the
5517 smp_call_function_many() function with a zero
5518 "wait" parameter. See weight_single.
5519 Note well that setting a high probability for
5520 this weighting can place serious IPI load
5523 scftorture.weight_many_wait= [KNL]
5524 The probability weighting to use for the
5525 smp_call_function_many() function with a
5526 non-zero "wait" parameter. See weight_single
5529 scftorture.weight_all= [KNL]
5530 The probability weighting to use for the
5531 smp_call_function_all() function with a zero
5532 "wait" parameter. See weight_single and
5535 scftorture.weight_all_wait= [KNL]
5536 The probability weighting to use for the
5537 smp_call_function_all() function with a
5538 non-zero "wait" parameter. See weight_single
5541 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5542 xtime_lock contention on larger systems, and/or RCU lock
5543 contention on all systems with CONFIG_MAXSMP set.
5544 Format: { "0" | "1" }
5545 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5547 Note: increases power consumption, thus should only be
5548 enabled if running jitter sensitive (HPC/RT) workloads.
5550 security= [SECURITY] Choose a legacy "major" security module to
5551 enable at boot. This has been deprecated by the
5554 selinux= [SELINUX] Disable or enable SELinux at boot time.
5555 Format: { "0" | "1" }
5556 See security/selinux/Kconfig help text.
5561 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5562 Format: { "0" | "1" }
5563 See security/apparmor/Kconfig help text
5566 Default value is set via kernel config option.
5568 serialnumber [BUGS=X86-32]
5570 sev=option[,option...] [X86-64] See Documentation/x86/x86_64/boot-options.rst
5573 Maximal number of shapers.
5581 Enable merging of slabs with similar size when the
5582 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5585 Disable merging of slabs with similar size. May be
5586 necessary if there is some reason to distinguish
5587 allocs to different slabs, especially in hardened
5588 environments where the risk of heap overflows and
5589 layout control by attackers can usually be
5590 frustrated by disabling merging. This will reduce
5591 most of the exposure of a heap attack to a single
5592 cache (risks via metadata attacks are mostly
5593 unchanged). Debug options disable merging on their
5595 For more information see Documentation/mm/slub.rst.
5597 slab_max_order= [MM, SLAB]
5598 Determines the maximum allowed order for slabs.
5599 A high setting may cause OOMs due to memory
5600 fragmentation. Defaults to 1 for systems with
5601 more than 32MB of RAM, 0 otherwise.
5603 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5604 Enabling slub_debug allows one to determine the
5605 culprit if slab objects become corrupted. Enabling
5606 slub_debug can create guard zones around objects and
5607 may poison objects when not in use. Also tracks the
5608 last alloc / free. For more information see
5609 Documentation/mm/slub.rst.
5611 slub_max_order= [MM, SLUB]
5612 Determines the maximum allowed order for slabs.
5613 A high setting may cause OOMs due to memory
5614 fragmentation. For more information see
5615 Documentation/mm/slub.rst.
5617 slub_min_objects= [MM, SLUB]
5618 The minimum number of objects per slab. SLUB will
5619 increase the slab order up to slub_max_order to
5620 generate a sufficiently large slab able to contain
5621 the number of objects indicated. The higher the number
5622 of objects the smaller the overhead of tracking slabs
5623 and the less frequently locks need to be acquired.
5624 For more information see Documentation/mm/slub.rst.
5626 slub_min_order= [MM, SLUB]
5627 Determines the minimum page order for slabs. Must be
5628 lower than slub_max_order.
5629 For more information see Documentation/mm/slub.rst.
5631 slub_merge [MM, SLUB]
5632 Same with slab_merge.
5634 slub_nomerge [MM, SLUB]
5635 Same with slab_nomerge. This is supported for legacy.
5636 See slab_nomerge for more information.
5639 Format: <io1>[,<io2>[,...,<io8>]]
5641 smp.csd_lock_timeout= [KNL]
5642 Specify the period of time in milliseconds
5643 that smp_call_function() and friends will wait
5644 for a CPU to release the CSD lock. This is
5645 useful when diagnosing bugs involving CPUs
5646 disabling interrupts for extended periods
5647 of time. Defaults to 5,000 milliseconds, and
5648 setting a value of zero disables this feature.
5649 This feature may be more efficiently disabled
5650 using the csdlock_debug- kernel parameter.
5652 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5653 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5654 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5655 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5656 smsc-ircc2.ircc_irq= [HW] IRQ line
5657 smsc-ircc2.ircc_dma= [HW] DMA channel
5658 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5659 0: Toshiba Satellite 1800 (GP data pin select)
5660 1: Fast pin select (default)
5663 smt= [KNL,S390] Set the maximum number of threads (logical
5664 CPUs) to use per physical CPU on systems capable of
5665 symmetric multithreading (SMT). Will be capped to the
5666 actual hardware limit.
5668 Default: -1 (no limit)
5671 [KNL] Should the soft-lockup detector generate panics.
5674 A value of 1 instructs the soft-lockup detector
5675 to panic the machine when a soft-lockup occurs. It is
5676 also controlled by the kernel.softlockup_panic sysctl
5677 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5678 respective build-time switch to that functionality.
5680 softlockup_all_cpu_backtrace=
5681 [KNL] Should the soft-lockup detector generate
5682 backtraces on all cpus.
5685 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5686 See Documentation/admin-guide/laptops/sonypi.rst
5688 spectre_v2= [X86] Control mitigation of Spectre variant 2
5689 (indirect branch speculation) vulnerability.
5690 The default operation protects the kernel from
5693 on - unconditionally enable, implies
5695 off - unconditionally disable, implies
5697 auto - kernel detects whether your CPU model is
5700 Selecting 'on' will, and 'auto' may, choose a
5701 mitigation method at run time according to the
5702 CPU, the available microcode, the setting of the
5703 CONFIG_RETPOLINE configuration option, and the
5704 compiler with which the kernel was built.
5706 Selecting 'on' will also enable the mitigation
5707 against user space to user space task attacks.
5709 Selecting 'off' will disable both the kernel and
5710 the user space protections.
5712 Specific mitigations can also be selected manually:
5714 retpoline - replace indirect branches
5715 retpoline,generic - Retpolines
5716 retpoline,lfence - LFENCE; indirect branch
5717 retpoline,amd - alias for retpoline,lfence
5718 eibrs - enhanced IBRS
5719 eibrs,retpoline - enhanced IBRS + Retpolines
5720 eibrs,lfence - enhanced IBRS + LFENCE
5721 ibrs - use IBRS to protect kernel
5723 Not specifying this option is equivalent to
5727 [X86] Control mitigation of Spectre variant 2
5728 (indirect branch speculation) vulnerability between
5731 on - Unconditionally enable mitigations. Is
5732 enforced by spectre_v2=on
5734 off - Unconditionally disable mitigations. Is
5735 enforced by spectre_v2=off
5737 prctl - Indirect branch speculation is enabled,
5738 but mitigation can be enabled via prctl
5739 per thread. The mitigation control state
5740 is inherited on fork.
5743 - Like "prctl" above, but only STIBP is
5744 controlled per thread. IBPB is issued
5745 always when switching between different user
5749 - Same as "prctl" above, but all seccomp
5750 threads will enable the mitigation unless
5751 they explicitly opt out.
5754 - Like "seccomp" above, but only STIBP is
5755 controlled per thread. IBPB is issued
5756 always when switching between different
5757 user space processes.
5759 auto - Kernel selects the mitigation depending on
5760 the available CPU features and vulnerability.
5762 Default mitigation: "prctl"
5764 Not specifying this option is equivalent to
5765 spectre_v2_user=auto.
5767 spec_store_bypass_disable=
5768 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5769 (Speculative Store Bypass vulnerability)
5771 Certain CPUs are vulnerable to an exploit against a
5772 a common industry wide performance optimization known
5773 as "Speculative Store Bypass" in which recent stores
5774 to the same memory location may not be observed by
5775 later loads during speculative execution. The idea
5776 is that such stores are unlikely and that they can
5777 be detected prior to instruction retirement at the
5778 end of a particular speculation execution window.
5780 In vulnerable processors, the speculatively forwarded
5781 store can be used in a cache side channel attack, for
5782 example to read memory to which the attacker does not
5783 directly have access (e.g. inside sandboxed code).
5785 This parameter controls whether the Speculative Store
5786 Bypass optimization is used.
5788 On x86 the options are:
5790 on - Unconditionally disable Speculative Store Bypass
5791 off - Unconditionally enable Speculative Store Bypass
5792 auto - Kernel detects whether the CPU model contains an
5793 implementation of Speculative Store Bypass and
5794 picks the most appropriate mitigation. If the
5795 CPU is not vulnerable, "off" is selected. If the
5796 CPU is vulnerable the default mitigation is
5797 architecture and Kconfig dependent. See below.
5798 prctl - Control Speculative Store Bypass per thread
5799 via prctl. Speculative Store Bypass is enabled
5800 for a process by default. The state of the control
5801 is inherited on fork.
5802 seccomp - Same as "prctl" above, but all seccomp threads
5803 will disable SSB unless they explicitly opt out.
5805 Default mitigations:
5808 On powerpc the options are:
5810 on,auto - On Power8 and Power9 insert a store-forwarding
5811 barrier on kernel entry and exit. On Power7
5812 perform a software flush on kernel entry and
5816 Not specifying this option is equivalent to
5817 spec_store_bypass_disable=auto.
5819 spia_io_base= [HW,MTD]
5825 [X86] Enable split lock detection or bus lock detection
5827 When enabled (and if hardware support is present), atomic
5828 instructions that access data across cache line
5829 boundaries will result in an alignment check exception
5830 for split lock detection or a debug exception for
5835 warn - the kernel will emit rate-limited warnings
5836 about applications triggering the #AC
5837 exception or the #DB exception. This mode is
5838 the default on CPUs that support split lock
5839 detection or bus lock detection. Default
5840 behavior is by #AC if both features are
5841 enabled in hardware.
5843 fatal - the kernel will send SIGBUS to applications
5844 that trigger the #AC exception or the #DB
5845 exception. Default behavior is by #AC if
5846 both features are enabled in hardware.
5849 Set system wide rate limit to N bus locks
5850 per second for bus lock detection.
5853 N/A for split lock detection.
5856 If an #AC exception is hit in the kernel or in
5857 firmware (i.e. not while executing in user mode)
5858 the kernel will oops in either "warn" or "fatal"
5861 #DB exception for bus lock is triggered only when
5865 Control the Special Register Buffer Data Sampling
5868 Certain CPUs are vulnerable to an MDS-like
5869 exploit which can leak bits from the random
5872 By default, this issue is mitigated by
5873 microcode. However, the microcode fix can cause
5874 the RDRAND and RDSEED instructions to become
5875 much slower. Among other effects, this will
5876 result in reduced throughput from /dev/urandom.
5878 The microcode mitigation can be disabled with
5879 the following option:
5881 off: Disable mitigation and remove
5882 performance impact to RDRAND and RDSEED
5884 srcutree.big_cpu_lim [KNL]
5885 Specifies the number of CPUs constituting a
5886 large system, such that srcu_struct structures
5887 should immediately allocate an srcu_node array.
5888 This kernel-boot parameter defaults to 128,
5889 but takes effect only when the low-order four
5890 bits of srcutree.convert_to_big is equal to 3
5893 srcutree.convert_to_big [KNL]
5894 Specifies under what conditions an SRCU tree
5895 srcu_struct structure will be converted to big
5896 form, that is, with an rcu_node tree:
5899 1: At init_srcu_struct() time.
5900 2: When rcutorture decides to.
5901 3: Decide at boot time (default).
5902 0x1X: Above plus if high contention.
5904 Either way, the srcu_node tree will be sized based
5905 on the actual runtime number of CPUs (nr_cpu_ids)
5906 instead of the compile-time CONFIG_NR_CPUS.
5908 srcutree.counter_wrap_check [KNL]
5909 Specifies how frequently to check for
5910 grace-period sequence counter wrap for the
5911 srcu_data structure's ->srcu_gp_seq_needed field.
5912 The greater the number of bits set in this kernel
5913 parameter, the less frequently counter wrap will
5914 be checked for. Note that the bottom two bits
5917 srcutree.exp_holdoff [KNL]
5918 Specifies how many nanoseconds must elapse
5919 since the end of the last SRCU grace period for
5920 a given srcu_struct until the next normal SRCU
5921 grace period will be considered for automatic
5922 expediting. Set to zero to disable automatic
5925 srcutree.srcu_max_nodelay [KNL]
5926 Specifies the number of no-delay instances
5927 per jiffy for which the SRCU grace period
5928 worker thread will be rescheduled with zero
5929 delay. Beyond this limit, worker thread will
5930 be rescheduled with a sleep delay of one jiffy.
5932 srcutree.srcu_max_nodelay_phase [KNL]
5933 Specifies the per-grace-period phase, number of
5934 non-sleeping polls of readers. Beyond this limit,
5935 grace period worker thread will be rescheduled
5936 with a sleep delay of one jiffy, between each
5937 rescan of the readers, for a grace period phase.
5939 srcutree.srcu_retry_check_delay [KNL]
5940 Specifies number of microseconds of non-sleeping
5941 delay between each non-sleeping poll of readers.
5943 srcutree.small_contention_lim [KNL]
5944 Specifies the number of update-side contention
5945 events per jiffy will be tolerated before
5946 initiating a conversion of an srcu_struct
5947 structure to big form. Note that the value of
5948 srcutree.convert_to_big must have the 0x10 bit
5949 set for contention-based conversions to occur.
5952 Speculative Store Bypass Disable control
5954 On CPUs that are vulnerable to the Speculative
5955 Store Bypass vulnerability and offer a
5956 firmware based mitigation, this parameter
5957 indicates how the mitigation should be used:
5959 force-on: Unconditionally enable mitigation for
5960 for both kernel and userspace
5961 force-off: Unconditionally disable mitigation for
5962 for both kernel and userspace
5963 kernel: Always enable mitigation in the
5964 kernel, and offer a prctl interface
5965 to allow userspace to register its
5966 interest in being mitigated too.
5968 stack_guard_gap= [MM]
5969 override the default stack gap protection. The value
5970 is in page units and it defines how many pages prior
5971 to (for stacks growing down) resp. after (for stacks
5972 growing up) the main stack are reserved for no other
5973 mapping. Default value is 256 pages.
5975 stack_depot_disable= [KNL]
5976 Setting this to true through kernel command line will
5977 disable the stack depot thereby saving the static memory
5978 consumed by the stack hash table. By default this is set
5982 Enabled the stack tracer on boot up.
5984 stacktrace_filter=[function-list]
5985 [FTRACE] Limit the functions that the stack tracer
5986 will trace at boot up. function-list is a comma-separated
5987 list of functions. This list can be changed at run
5988 time by the stack_trace_filter file in the debugfs
5989 tracing directory. Note, this enables stack tracing
5990 and the stacktrace above is not needed.
5994 Set the STI (builtin display/keyboard on the HP-PARISC
5995 machines) console (graphic card) which should be used
5996 as the initial boot-console.
5997 See also comment in drivers/video/console/sticore.c.
6000 See comment in drivers/video/console/sticore.c.
6003 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
6008 Enable or disable strict sigaltstack size checks
6009 against the required signal frame size which
6010 depends on the supported FPU features. This can
6011 be used to filter out binaries which have
6012 not yet been made aware of AT_MINSIGSTKSZ.
6014 sunrpc.min_resvport=
6015 sunrpc.max_resvport=
6017 SunRPC servers often require that client requests
6018 originate from a privileged port (i.e. a port in the
6019 range 0 < portnr < 1024).
6020 An administrator who wishes to reserve some of these
6021 ports for other uses may adjust the range that the
6022 kernel's sunrpc client considers to be privileged
6023 using these two parameters to set the minimum and
6024 maximum port values.
6026 sunrpc.svc_rpc_per_connection_limit=
6028 Limit the number of requests that the server will
6029 process in parallel from a single connection.
6030 The default value is 0 (no limit).
6034 Control how the NFS server code allocates CPUs to
6035 service thread pools. Depending on how many NICs
6036 you have and where their interrupts are bound, this
6037 option will affect which CPUs will do NFS serving.
6038 Note: this parameter cannot be changed while the
6039 NFS server is running.
6041 auto the server chooses an appropriate mode
6042 automatically using heuristics
6043 global a single global pool contains all CPUs
6044 percpu one pool for each CPU
6045 pernode one pool for each NUMA node (equivalent
6046 to global on non-NUMA machines)
6048 sunrpc.tcp_slot_table_entries=
6049 sunrpc.udp_slot_table_entries=
6051 Sets the upper limit on the number of simultaneous
6052 RPC calls that can be sent from the client to a
6053 server. Increasing these values may allow you to
6054 improve throughput, but will also increase the
6055 amount of memory reserved for use by the client.
6057 suspend.pm_test_delay=
6059 Sets the number of seconds to remain in a suspend test
6060 mode before resuming the system (see
6061 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6062 is set. Default value is 5.
6065 Format: { on | off | y | n | 1 | 0 }
6066 This parameter controls use of the Protected
6067 Execution Facility on pSeries.
6069 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
6070 Format: { <int> [,<int>] | force | noforce }
6071 <int> -- Number of I/O TLB slabs
6072 <int> -- Second integer after comma. Number of swiotlb
6073 areas with their own lock. Will be rounded up
6075 force -- force using of bounce buffers even if they
6076 wouldn't be automatically used by the kernel
6077 noforce -- Never use bounce buffers (for debugging)
6082 Set a sysctl parameter, right before loading the init
6083 process, as if the value was written to the respective
6084 /proc/sys/... file. Both '.' and '/' are recognized as
6085 separators. Unrecognized parameters and invalid values
6086 are reported in the kernel log. Sysctls registered
6087 later by a loaded module cannot be set this way.
6088 Example: sysctl.vm.swappiness=40
6090 sysfs.deprecated=0|1 [KNL]
6091 Enable/disable old style sysfs layout for old udev
6092 on older distributions. When this option is enabled
6093 very new udev will not work anymore. When this option
6094 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
6095 in older udev will not work anymore.
6096 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
6097 the kernel configuration.
6099 sysrq_always_enabled
6101 Ignore sysrq setting - this boot parameter will
6102 neutralize any effect of /proc/sys/kernel/sysrq.
6103 Useful for debugging.
6105 tcpmhash_entries= [KNL,NET]
6106 Set the number of tcp_metrics_hash slots.
6107 Default value is 8192 or 16384 depending on total
6108 ram pages. This is used to specify the TCP metrics
6109 cache size. See Documentation/networking/ip-sysctl.rst
6110 "tcp_no_metrics_save" section for more details.
6114 test_suspend= [SUSPEND]
6115 Format: { "mem" | "standby" | "freeze" }[,N]
6116 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6117 standby suspend) or "freeze" (for suspend type freeze)
6118 as the system sleep state during system startup with
6119 the optional capability to repeat N number of times.
6120 The system is woken from this state using a
6121 wakeup-capable RTC alarm.
6123 thash_entries= [KNL,NET]
6124 Set number of hash buckets for TCP connection
6126 thermal.act= [HW,ACPI]
6127 -1: disable all active trip points in all thermal zones
6128 <degrees C>: override all lowest active trip points
6130 thermal.crt= [HW,ACPI]
6131 -1: disable all critical trip points in all thermal zones
6132 <degrees C>: override all critical trip points
6134 thermal.nocrt= [HW,ACPI]
6135 Set to disable actions on ACPI thermal zone
6136 critical and hot trip points.
6138 thermal.off= [HW,ACPI]
6139 1: disable ACPI thermal control
6141 thermal.psv= [HW,ACPI]
6142 -1: disable all passive trip points
6143 <degrees C>: override all passive trip points to this
6146 thermal.tzp= [HW,ACPI]
6147 Specify global default ACPI thermal zone polling rate
6148 <deci-seconds>: poll all this frequency
6149 0: no polling (default)
6152 Force threading of all interrupt handlers except those
6153 marked explicitly IRQF_NO_THREAD.
6157 Specify if the kernel should make use of the cpu
6158 topology information if the hardware supports this.
6159 The scheduler will make use of this information and
6160 e.g. base its process migration decisions on it.
6163 topology_updates= [KNL, PPC, NUMA]
6165 Specify if the kernel should ignore (off)
6166 topology updates sent by the hypervisor to this
6169 torture.disable_onoff_at_boot= [KNL]
6170 Prevent the CPU-hotplug component of torturing
6171 until after init has spawned.
6173 torture.ftrace_dump_at_shutdown= [KNL]
6174 Dump the ftrace buffer at torture-test shutdown,
6175 even if there were no errors. This can be a
6176 very costly operation when many torture tests
6177 are running concurrently, especially on systems
6178 with rotating-rust storage.
6180 torture.verbose_sleep_frequency= [KNL]
6181 Specifies how many verbose printk()s should be
6182 emitted between each sleep. The default of zero
6183 disables verbose-printk() sleeping.
6185 torture.verbose_sleep_duration= [KNL]
6186 Duration of each verbose-printk() sleep in jiffies.
6190 tpm_suspend_pcr=[HW,TPM]
6191 Format: integer pcr id
6192 Specify that at suspend time, the tpm driver
6193 should extend the specified pcr with zeros,
6194 as a workaround for some chips which fail to
6195 flush the last written pcr on TPM_SaveState.
6196 This will guarantee that all the other pcrs
6200 Have the tracepoints sent to printk as well as the
6201 tracing ring buffer. This is useful for early boot up
6202 where the system hangs or reboots and does not give the
6203 option for reading the tracing buffer or performing a
6204 ftrace_dump_on_oops.
6206 To turn off having tracepoints sent to printk,
6207 echo 0 > /proc/sys/kernel/tracepoint_printk
6208 Note, echoing 1 into this file without the
6209 tracepoint_printk kernel cmdline option has no effect.
6211 The tp_printk_stop_on_boot (see below) can also be used
6212 to stop the printing of events to console at
6217 Having tracepoints sent to printk() and activating high
6218 frequency tracepoints such as irq or sched, can cause
6219 the system to live lock.
6221 tp_printk_stop_on_boot [FTRACE]
6222 When tp_printk (above) is set, it can cause a lot of noise
6223 on the console. It may be useful to only include the
6224 printing of events during boot up, as user space may
6225 make the system inoperable.
6227 This command line option will stop the printing of events
6228 to console at the late_initcall_sync() time frame.
6230 trace_buf_size=nn[KMG]
6231 [FTRACE] will set tracing buffer size on each cpu.
6233 trace_clock= [FTRACE] Set the clock used for tracing events
6235 local - Use the per CPU time stamp counter
6236 (converted into nanoseconds). Fast, but
6237 depending on the architecture, may not be
6238 in sync between CPUs.
6239 global - Event time stamps are synchronize across
6240 CPUs. May be slower than the local clock,
6241 but better for some race conditions.
6242 counter - Simple counting of events (1, 2, ..)
6243 note, some counts may be skipped due to the
6244 infrastructure grabbing the clock more than
6246 uptime - Use jiffies as the time stamp.
6247 perf - Use the same clock that perf uses.
6248 mono - Use ktime_get_mono_fast_ns() for time stamps.
6249 mono_raw - Use ktime_get_raw_fast_ns() for time
6251 boot - Use ktime_get_boot_fast_ns() for time stamps.
6252 Architectures may add more clocks. See
6253 Documentation/trace/ftrace.rst for more details.
6255 trace_event=[event-list]
6256 [FTRACE] Set and start specified trace events in order
6257 to facilitate early boot debugging. The event-list is a
6258 comma-separated list of trace events to enable. See
6259 also Documentation/trace/events.rst
6261 trace_options=[option-list]
6262 [FTRACE] Enable or disable tracer options at boot.
6263 The option-list is a comma delimited list of options
6264 that can be enabled or disabled just as if you were
6265 to echo the option name into
6267 /sys/kernel/debug/tracing/trace_options
6269 For example, to enable stacktrace option (to dump the
6270 stack trace of each event), add to the command line:
6272 trace_options=stacktrace
6274 See also Documentation/trace/ftrace.rst "trace options"
6278 [FTRACE] enable this option to disable tracing when a
6279 warning is hit. This turns off "tracing_on". Tracing can
6280 be enabled again by echoing '1' into the "tracing_on"
6281 file located in /sys/kernel/debug/tracing/
6283 This option is useful, as it disables the trace before
6284 the WARNING dump is called, which prevents the trace to
6285 be filled with content caused by the warning output.
6287 This option can also be set at run time via the sysctl
6288 option: kernel/traceoff_on_warning
6290 transparent_hugepage=
6292 Format: [always|madvise|never]
6293 Can be used to control the default behavior of the system
6294 with respect to transparent hugepages.
6295 See Documentation/admin-guide/mm/transhuge.rst
6298 trusted.source= [KEYS]
6300 This parameter identifies the trust source as a backend
6301 for trusted keys implementation. Supported trust
6306 If not specified then it defaults to iterating through
6307 the trust source list starting with TPM and assigns the
6308 first trust source as a backend which is initialized
6309 successfully during iteration.
6313 The RNG used to generate key material for trusted keys.
6316 - the same value as trusted.source: "tpm" or "tee"
6318 If not specified, "default" is used. In this case,
6319 the RNG's choice is left to each individual trust source.
6321 tsc= Disable clocksource stability checks for TSC.
6323 [x86] reliable: mark tsc clocksource as reliable, this
6324 disables clocksource verification at runtime, as well
6325 as the stability checks done at bootup. Used to enable
6326 high-resolution timer mode on older hardware, and in
6327 virtualized environment.
6328 [x86] noirqtime: Do not use TSC to do irq accounting.
6329 Used to run time disable IRQ_TIME_ACCOUNTING on any
6330 platforms where RDTSC is slow and this accounting
6332 [x86] unstable: mark the TSC clocksource as unstable, this
6333 marks the TSC unconditionally unstable at bootup and
6334 avoids any further wobbles once the TSC watchdog notices.
6335 [x86] nowatchdog: disable clocksource watchdog. Used
6336 in situations with strict latency requirements (where
6337 interruptions from clocksource watchdog are not
6340 tsc_early_khz= [X86] Skip early TSC calibration and use the given
6341 value instead. Useful when the early TSC frequency discovery
6342 procedure is not reliable, such as on overclocked systems
6343 with CPUID.16h support and partial CPUID.15h support.
6344 Format: <unsigned int>
6346 tsx= [X86] Control Transactional Synchronization
6347 Extensions (TSX) feature in Intel processors that
6348 support TSX control.
6350 This parameter controls the TSX feature. The options are:
6352 on - Enable TSX on the system. Although there are
6353 mitigations for all known security vulnerabilities,
6354 TSX has been known to be an accelerator for
6355 several previous speculation-related CVEs, and
6356 so there may be unknown security risks associated
6357 with leaving it enabled.
6359 off - Disable TSX on the system. (Note that this
6360 option takes effect only on newer CPUs which are
6361 not vulnerable to MDS, i.e., have
6362 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6363 the new IA32_TSX_CTRL MSR through a microcode
6364 update. This new MSR allows for the reliable
6365 deactivation of the TSX functionality.)
6367 auto - Disable TSX if X86_BUG_TAA is present,
6368 otherwise enable TSX on the system.
6370 Not specifying this option is equivalent to tsx=off.
6372 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6375 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6376 Abort (TAA) vulnerability.
6378 Similar to Micro-architectural Data Sampling (MDS)
6379 certain CPUs that support Transactional
6380 Synchronization Extensions (TSX) are vulnerable to an
6381 exploit against CPU internal buffers which can forward
6382 information to a disclosure gadget under certain
6385 In vulnerable processors, the speculatively forwarded
6386 data can be used in a cache side channel attack, to
6387 access data to which the attacker does not have direct
6390 This parameter controls the TAA mitigation. The
6393 full - Enable TAA mitigation on vulnerable CPUs
6396 full,nosmt - Enable TAA mitigation and disable SMT on
6397 vulnerable CPUs. If TSX is disabled, SMT
6398 is not disabled because CPU is not
6399 vulnerable to cross-thread TAA attacks.
6400 off - Unconditionally disable TAA mitigation
6402 On MDS-affected machines, tsx_async_abort=off can be
6403 prevented by an active MDS mitigation as both vulnerabilities
6404 are mitigated with the same mechanism so in order to disable
6405 this mitigation, you need to specify mds=off too.
6407 Not specifying this option is equivalent to
6408 tsx_async_abort=full. On CPUs which are MDS affected
6409 and deploy MDS mitigation, TAA mitigation is not
6410 required and doesn't provide any additional
6414 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6416 turbografx.map[2|3]= [HW,JOY]
6417 TurboGraFX parallel port interface
6419 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6420 See also Documentation/input/devices/joystick-parport.rst
6422 udbg-immortal [PPC] When debugging early kernel crashes that
6423 happen after console_init() and before a proper
6424 console driver takes over, this boot options might
6425 help "seeing" what's going on.
6427 uhash_entries= [KNL,NET]
6428 Set number of hash buckets for UDP/UDP-Lite connections
6431 [USB] Ignore overcurrent events (default N).
6432 Some badly-designed motherboards generate lots of
6433 bogus events, for ports that aren't wired to
6434 anything. Set this parameter to avoid log spamming.
6435 Note that genuine overcurrent events won't be
6439 [X86] Cause panic on unknown NMI.
6441 usbcore.authorized_default=
6442 [USB] Default USB device authorization:
6443 (default -1 = authorized except for wireless USB,
6444 0 = not authorized, 1 = authorized, 2 = authorized
6445 if device connected to internal port)
6447 usbcore.autosuspend=
6448 [USB] The autosuspend time delay (in seconds) used
6449 for newly-detected USB devices (default 2). This
6450 is the time required before an idle device will be
6451 autosuspended. Devices for which the delay is set
6452 to a negative value won't be autosuspended at all.
6454 usbcore.usbfs_snoop=
6455 [USB] Set to log all usbfs traffic (default 0 = off).
6457 usbcore.usbfs_snoop_max=
6458 [USB] Maximum number of bytes to snoop in each URB
6461 usbcore.blinkenlights=
6462 [USB] Set to cycle leds on hubs (default 0 = off).
6464 usbcore.old_scheme_first=
6465 [USB] Start with the old device initialization
6466 scheme (default 0 = off).
6468 usbcore.usbfs_memory_mb=
6469 [USB] Memory limit (in MB) for buffers allocated by
6470 usbfs (default = 16, 0 = max = 2047).
6472 usbcore.use_both_schemes=
6473 [USB] Try the other device initialization scheme
6474 if the first one fails (default 1 = enabled).
6476 usbcore.initial_descriptor_timeout=
6477 [USB] Specifies timeout for the initial 64-byte
6478 USB_REQ_GET_DESCRIPTOR request in milliseconds
6479 (default 5000 = 5.0 seconds).
6481 usbcore.nousb [USB] Disable the USB subsystem
6484 [USB] A list of quirk entries to augment the built-in
6485 usb core quirk list. List entries are separated by
6486 commas. Each entry has the form
6487 VendorID:ProductID:Flags. The IDs are 4-digit hex
6488 numbers and Flags is a set of letters. Each letter
6489 will change the built-in quirk; setting it if it is
6490 clear and clearing it if it is set. The letters have
6491 the following meanings:
6492 a = USB_QUIRK_STRING_FETCH_255 (string
6493 descriptors must not be fetched using
6495 b = USB_QUIRK_RESET_RESUME (device can't resume
6496 correctly so reset it instead);
6497 c = USB_QUIRK_NO_SET_INTF (device can't handle
6498 Set-Interface requests);
6499 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6500 handle its Configuration or Interface
6502 e = USB_QUIRK_RESET (device can't be reset
6503 (e.g morph devices), don't use reset);
6504 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6505 more interface descriptions than the
6506 bNumInterfaces count, and can't handle
6507 talking to these interfaces);
6508 g = USB_QUIRK_DELAY_INIT (device needs a pause
6509 during initialization, after we read
6510 the device descriptor);
6511 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6512 high speed and super speed interrupt
6513 endpoints, the USB 2.0 and USB 3.0 spec
6514 require the interval in microframes (1
6515 microframe = 125 microseconds) to be
6516 calculated as interval = 2 ^
6518 Devices with this quirk report their
6519 bInterval as the result of this
6520 calculation instead of the exponent
6521 variable used in the calculation);
6522 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6523 handle device_qualifier descriptor
6525 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6526 generates spurious wakeup, ignore
6527 remote wakeup capability);
6528 k = USB_QUIRK_NO_LPM (device can't handle Link
6530 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6531 (Device reports its bInterval as linear
6532 frames instead of the USB 2.0
6534 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6535 to be disconnected before suspend to
6536 prevent spurious wakeup);
6537 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6538 pause after every control message);
6539 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6540 delay after resetting its port);
6541 Example: quirks=0781:5580:bk,0a5c:5834:gij
6544 [USBHID] The interval which mice are to be polled at.
6547 [USBHID] The interval which joysticks are to be polled at.
6550 [USBHID] The interval which keyboards are to be polled at.
6552 usb-storage.delay_use=
6553 [UMS] The delay in seconds before a new device is
6554 scanned for Logical Units (default 1).
6557 [UMS] A list of quirks entries to supplement or
6558 override the built-in unusual_devs list. List
6559 entries are separated by commas. Each entry has
6560 the form VID:PID:Flags where VID and PID are Vendor
6561 and Product ID values (4-digit hex numbers) and
6562 Flags is a set of characters, each corresponding
6563 to a common usb-storage quirk flag as follows:
6564 a = SANE_SENSE (collect more than 18 bytes
6565 of sense data, not on uas);
6566 b = BAD_SENSE (don't collect more than 18
6567 bytes of sense data, not on uas);
6568 c = FIX_CAPACITY (decrease the reported
6569 device capacity by one sector);
6570 d = NO_READ_DISC_INFO (don't use
6571 READ_DISC_INFO command, not on uas);
6572 e = NO_READ_CAPACITY_16 (don't use
6573 READ_CAPACITY_16 command);
6574 f = NO_REPORT_OPCODES (don't use report opcodes
6576 g = MAX_SECTORS_240 (don't transfer more than
6577 240 sectors at a time, uas only);
6578 h = CAPACITY_HEURISTICS (decrease the
6579 reported device capacity by one
6580 sector if the number is odd);
6581 i = IGNORE_DEVICE (don't bind to this
6583 j = NO_REPORT_LUNS (don't use report luns
6585 k = NO_SAME (do not use WRITE_SAME, uas only)
6586 l = NOT_LOCKABLE (don't try to lock and
6587 unlock ejectable media, not on uas);
6588 m = MAX_SECTORS_64 (don't transfer more
6589 than 64 sectors = 32 KB at a time,
6591 n = INITIAL_READ10 (force a retry of the
6592 initial READ(10) command, not on uas);
6593 o = CAPACITY_OK (accept the capacity
6594 reported by the device, not on uas);
6595 p = WRITE_CACHE (the device cache is ON
6596 by default, not on uas);
6597 r = IGNORE_RESIDUE (the device reports
6598 bogus residue values, not on uas);
6599 s = SINGLE_LUN (the device has only one
6601 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6602 commands, uas only);
6603 u = IGNORE_UAS (don't bind to the uas driver);
6604 w = NO_WP_DETECT (don't test whether the
6605 medium is write-protected).
6606 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6607 even if the device claims no cache,
6609 Example: quirks=0419:aaf5:rl,0421:0433:rc
6611 user_debug= [KNL,ARM]
6613 See arch/arm/Kconfig.debug help text.
6614 1 - undefined instruction events
6616 4 - invalid data aborts
6619 Example: user_debug=31
6622 [X86] Flags controlling user PTE allocations.
6624 nohigh = do not allocate PTE pages in
6625 HIGHMEM regardless of setting
6628 vdso= [X86,SH,SPARC]
6629 On X86_32, this is an alias for vdso32=. Otherwise:
6631 vdso=1: enable VDSO (the default)
6632 vdso=0: disable VDSO mapping
6634 vdso32= [X86] Control the 32-bit vDSO
6635 vdso32=1: enable 32-bit VDSO
6636 vdso32=0 or vdso32=2: disable 32-bit VDSO
6638 See the help text for CONFIG_COMPAT_VDSO for more
6639 details. If CONFIG_COMPAT_VDSO is set, the default is
6640 vdso32=0; otherwise, the default is vdso32=1.
6642 For compatibility with older kernels, vdso32=2 is an
6645 Try vdso32=0 if you encounter an error that says:
6646 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6649 vector=percpu: enable percpu vector domain
6651 video= [FB] Frame buffer configuration
6652 See Documentation/fb/modedb.rst.
6654 video.brightness_switch_enabled= [ACPI]
6656 If set to 1, on receiving an ACPI notify event
6657 generated by hotkey, video driver will adjust brightness
6658 level and then send out the event to user space through
6659 the allocated input device. If set to 0, video driver
6660 will only send out the event without touching backlight
6665 [VMMIO] Memory mapped virtio (platform) device.
6667 <size>@<baseaddr>:<irq>[:<id>]
6669 <size> := size (can use standard suffixes
6671 <baseaddr> := physical base address
6672 <irq> := interrupt number (as passed to
6674 <id> := (optional) platform device id
6676 virtio_mmio.device=1K@0x100b0000:48:7
6678 Can be used multiple times for multiple devices.
6680 vga= [BOOT,X86-32] Select a particular video mode
6681 See Documentation/x86/boot.rst and
6682 Documentation/admin-guide/svga.rst.
6683 Use vga=ask for menu.
6684 This is actually a boot loader parameter; the value is
6685 passed to the kernel using a special protocol.
6687 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6688 May slow down system boot speed, especially when
6689 enabled on systems with a large amount of memory.
6690 All options are enabled by default, and this
6691 interface is meant to allow for selectively
6692 enabling or disabling specific virtual memory
6695 Available options are:
6696 P Enable page structure init time poisoning
6697 - Disable all of the above options
6699 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6700 size of <nn>. This can be used to increase the
6701 minimum size (128MB on x86). It can also be used to
6702 decrease the size and leave more room for directly
6705 vmcp_cma=nn[MG] [KNL,S390]
6706 Sets the memory size reserved for contiguous memory
6707 allocations for the vmcp device driver.
6709 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6712 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6715 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6719 Controls the behavior of vsyscalls (i.e. calls to
6720 fixed addresses of 0xffffffffff600x00 from legacy
6721 code). Most statically-linked binaries and older
6722 versions of glibc use these calls. Because these
6723 functions are at fixed addresses, they make nice
6724 targets for exploits that can control RIP.
6726 emulate [default] Vsyscalls turn into traps and are
6727 emulated reasonably safely. The vsyscall
6730 xonly Vsyscalls turn into traps and are
6731 emulated reasonably safely. The vsyscall
6732 page is not readable.
6734 none Vsyscalls don't work at all. This makes
6735 them quite hard to use for exploits but
6736 might break your system.
6738 vt.color= [VT] Default text color.
6739 Format: 0xYX, X = foreground, Y = background.
6740 Default: 0x07 = light gray on black.
6742 vt.cur_default= [VT] Default cursor shape.
6743 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6744 the parameters of the <Esc>[?A;B;Cc escape sequence;
6745 see VGA-softcursor.txt. Default: 2 = underline.
6747 vt.default_blu= [VT]
6748 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6749 Change the default blue palette of the console.
6750 This is a 16-member array composed of values
6753 vt.default_grn= [VT]
6754 Format: <green0>,<green1>,<green2>,...,<green15>
6755 Change the default green palette of the console.
6756 This is a 16-member array composed of values
6759 vt.default_red= [VT]
6760 Format: <red0>,<red1>,<red2>,...,<red15>
6761 Change the default red palette of the console.
6762 This is a 16-member array composed of values
6768 Set system-wide default UTF-8 mode for all tty's.
6769 Default is 1, i.e. UTF-8 mode is enabled for all
6770 newly opened terminals.
6772 vt.global_cursor_default=
6775 Set system-wide default for whether a cursor
6776 is shown on new VTs. Default is -1,
6777 i.e. cursors will be created by default unless
6778 overridden by individual drivers. 0 will hide
6779 cursors, 1 will display them.
6781 vt.italic= [VT] Default color for italic text; 0-15.
6784 vt.underline= [VT] Default color for underlined text; 0-15.
6787 watchdog timers [HW,WDT] For information on watchdog timers,
6788 see Documentation/watchdog/watchdog-parameters.rst
6789 or other driver-specific files in the
6790 Documentation/watchdog/ directory.
6794 Set the hard lockup detector stall duration
6795 threshold in seconds. The soft lockup detector
6796 threshold is set to twice the value. A value of 0
6797 disables both lockup detectors. Default is 10
6800 workqueue.watchdog_thresh=
6801 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6802 warn stall conditions and dump internal state to
6803 help debugging. 0 disables workqueue stall
6804 detection; otherwise, it's the stall threshold
6805 duration in seconds. The default value is 30 and
6806 it can be updated at runtime by writing to the
6807 corresponding sysfs file.
6809 workqueue.disable_numa
6810 By default, all work items queued to unbound
6811 workqueues are affine to the NUMA nodes they're
6812 issued on, which results in better behavior in
6813 general. If NUMA affinity needs to be disabled for
6814 whatever reason, this option can be used. Note
6815 that this also can be controlled per-workqueue for
6816 workqueues visible under /sys/bus/workqueue/.
6818 workqueue.power_efficient
6819 Per-cpu workqueues are generally preferred because
6820 they show better performance thanks to cache
6821 locality; unfortunately, per-cpu workqueues tend to
6822 be more power hungry than unbound workqueues.
6824 Enabling this makes the per-cpu workqueues which
6825 were observed to contribute significantly to power
6826 consumption unbound, leading to measurably lower
6827 power usage at the cost of small performance
6830 The default value of this parameter is determined by
6831 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6833 workqueue.debug_force_rr_cpu
6834 Workqueue used to implicitly guarantee that work
6835 items queued without explicit CPU specified are put
6836 on the local CPU. This guarantee is no longer true
6837 and while local CPU is still preferred work items
6838 may be put on foreign CPUs. This debug option
6839 forces round-robin CPU selection to flush out
6840 usages which depend on the now broken guarantee.
6841 When enabled, memory and cache locality will be
6844 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6845 default x2apic cluster mode on platforms
6848 xen_512gb_limit [KNL,X86-64,XEN]
6849 Restricts the kernel running paravirtualized under Xen
6850 to use only up to 512 GB of RAM. The reason to do so is
6851 crash analysis tools and Xen tools for doing domain
6852 save/restore/migration must be enabled to handle larger
6855 xen_emul_unplug= [HW,X86,XEN]
6856 Unplug Xen emulated devices
6857 Format: [unplug0,][unplug1]
6858 ide-disks -- unplug primary master IDE devices
6859 aux-ide-disks -- unplug non-primary-master IDE devices
6860 nics -- unplug network devices
6861 all -- unplug all emulated devices (NICs and IDE disks)
6862 unnecessary -- unplugging emulated devices is
6863 unnecessary even if the host did not respond to
6865 never -- do not unplug even if version check succeeds
6867 xen_legacy_crash [X86,XEN]
6868 Crash from Xen panic notifier, without executing late
6869 panic() code such as dumping handler.
6871 xen_msr_safe= [X86,XEN]
6873 Select whether to always use non-faulting (safe) MSR
6874 access functions when running as Xen PV guest. The
6875 default value is controlled by CONFIG_XEN_PV_MSR_SAFE.
6877 xen_nopvspin [X86,XEN]
6878 Disables the qspinlock slowpath using Xen PV optimizations.
6879 This parameter is obsoleted by "nopvspin" parameter, which
6880 has equivalent effect for XEN platform.
6883 Disables the PV optimizations forcing the HVM guest to
6884 run as generic HVM guest with no PV drivers.
6885 This option is obsoleted by the "nopv" option, which
6886 has equivalent effect for XEN platform.
6888 xen_no_vector_callback
6889 [KNL,X86,XEN] Disable the vector callback for Xen
6890 event channel interrupts.
6892 xen_scrub_pages= [XEN]
6893 Boolean option to control scrubbing pages before giving them back
6894 to Xen, for use by other domains. Can be also changed at runtime
6895 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6896 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6898 xen_timer_slop= [X86-64,XEN]
6899 Set the timer slop (in nanoseconds) for the virtual Xen
6900 timers (default is 100000). This adjusts the minimum
6901 delta of virtualized Xen timers, where lower values
6902 improve timer resolution at the expense of processing
6903 more timer interrupts.
6905 xen.balloon_boot_timeout= [XEN]
6906 The time (in seconds) to wait before giving up to boot
6907 in case initial ballooning fails to free enough memory.
6908 Applies only when running as HVM or PVH guest and
6909 started with less memory configured than allowed at
6910 max. Default is 180.
6912 xen.event_eoi_delay= [XEN]
6913 How long to delay EOI handling in case of event
6914 storms (jiffies). Default is 10.
6916 xen.event_loop_timeout= [XEN]
6917 After which time (jiffies) the event handling loop
6918 should start to delay EOI handling. Default is 2.
6920 xen.fifo_events= [XEN]
6921 Boolean parameter to disable using fifo event handling
6922 even if available. Normally fifo event handling is
6923 preferred over the 2-level event handling, as it is
6924 fairer and the number of possible event channels is
6925 much higher. Default is on (use fifo events).
6927 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6928 Disables the PV optimizations forcing the guest to run
6929 as generic guest with no PV drivers. Currently support
6930 XEN HVM, KVM, HYPER_V and VMWARE guest.
6932 nopvspin [X86,XEN,KVM]
6933 Disables the qspinlock slow path using PV optimizations
6934 which allow the hypervisor to 'idle' the guest on lock
6937 xirc2ps_cs= [NET,PCMCIA]
6939 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6942 By default on POWER9 and above, the kernel will
6943 natively use the XIVE interrupt controller. This option
6944 allows the fallback firmware mode to be used:
6946 off Fallback to firmware control of XIVE interrupt
6947 controller on both pseries and powernv
6948 platforms. Only useful on POWER9 and above.
6950 xive.store-eoi=off [PPC]
6951 By default on POWER10 and above, the kernel will use
6952 stores for EOI handling when the XIVE interrupt mode
6953 is active. This option allows the XIVE driver to use
6954 loads instead, as on POWER9.
6956 xhci-hcd.quirks [USB,KNL]
6957 A hex value specifying bitmask with supplemental xhci
6958 host controller quirks. Meaning of each bit can be
6959 consulted in header drivers/usb/host/xhci.h.
6962 Format: { early | on | rw | ro | off }
6963 Controls if xmon debugger is enabled. Default is off.
6964 Passing only "xmon" is equivalent to "xmon=early".
6965 early Call xmon as early as possible on boot; xmon
6966 debugger is called from setup_arch().
6967 on xmon debugger hooks will be installed so xmon
6968 is only called on a kernel crash. Default mode,
6969 i.e. either "ro" or "rw" mode, is controlled
6970 with CONFIG_XMON_DEFAULT_RO_MODE.
6971 rw xmon debugger hooks will be installed so xmon
6972 is called only on a kernel crash, mode is write,
6973 meaning SPR registers, memory and, other data
6974 can be written using xmon commands.
6975 ro same as "rw" option above but SPR registers,
6976 memory, and other data can't be written using
6978 off xmon is disabled.
6982 Do not enable amd_pstate as the default
6983 scaling driver for the supported processors
6985 Use amd_pstate as a scaling driver, driver requests a
6986 desired performance on this abstract scale and the power
6987 management firmware translates the requests into actual
6988 hardware states (core frequency, data fabric and memory