1 acpi= [HW,ACPI,X86,ARM64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
5 force -- enable ACPI if default was off
6 on -- enable ACPI but allow fallback to DT [arm64]
7 off -- disable ACPI if default was on
8 noirq -- do not use ACPI for IRQ routing
9 strict -- Be less tolerant of platforms that are not
10 strictly ACPI specification compliant.
11 rsdt -- prefer RSDT over (default) XSDT
12 copy_dsdt -- copy DSDT to memory
13 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force"
16 See also Documentation/power/runtime_pm.rst, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
25 { vendor | video | native | none }
26 If set to vendor, prefer vendor-specific driver
27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
28 of the ACPI video.ko driver.
29 If set to video, use the ACPI video.ko driver.
30 If set to native, use the device's native backlight mode.
31 If set to none, disable the ACPI backlight interface.
33 acpi_force_32bit_fadt_addr
34 force FADT to use 32 bit addresses rather than the
35 64 bit X_* addresses. Some firmware have broken 64
36 bit addresses for force ACPI ignore these and use
37 the older legacy 32 bit addresses.
39 acpica_no_return_repair [HW, ACPI]
40 Disable AML predefined validation mechanism
41 This mechanism can repair the evaluation result to make
42 the return objects more ACPI specification compliant.
43 This option is useful for developers to identify the
44 root cause of an AML interpreter issue when the issue
45 has something to do with the repair mechanism.
47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
51 debug output. Bits in debug_layer correspond to a
52 _COMPONENT in an ACPI source file, e.g.,
53 #define _COMPONENT ACPI_EVENTS
54 Bits in debug_level correspond to a level in
55 ACPI_DEBUG_PRINT statements, e.g.,
56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
57 The debug_level mask defaults to "info". See
58 Documentation/firmware-guide/acpi/debug.rst for more information about
59 debug layers and levels.
61 Enable processor driver info messages:
62 acpi.debug_layer=0x20000000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
116 Format: <byte> or <bitmap-list>
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
163 acpi_osi=!! # enable all built-in OS vendor
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
229 old_ordering, nonvs, sci_force_enable, nobl }
230 See Documentation/power/video.rst for information on
232 s3_beep is for debugging; it makes the PC's speaker beep
233 as soon as the kernel's real-mode entry point is called.
234 s4_nohwsig prevents ACPI hardware signature from being
235 used during resume from hibernation.
236 old_ordering causes the ACPI 1.0 ordering of the _PTS
237 control method, with respect to putting devices into
238 low power states, to be enforced (the ACPI 2.0 ordering
239 of _PTS is used by default).
240 nonvs prevents the kernel from saving/restoring the
241 ACPI NVS memory during suspend/hibernation and resume.
242 sci_force_enable causes the kernel to set SCI_EN directly
243 on resume from S1/S3 (which is against the ACPI spec,
244 but some broken systems don't work without it).
245 nobl causes the internal blacklist of systems known to
246 behave incorrectly in some ways with respect to system
247 suspend and resume to be ignored (use wisely).
249 acpi_use_timer_override [HW,ACPI]
250 Use timer override. For some broken Nvidia NF5 boards
251 that require a timer override, but don't have HPET
253 add_efi_memmap [EFI; X86] Include EFI memory map in
254 kernel's map of available physical RAM.
257 { off | try_unsupported }
258 off: disable AGP support
259 try_unsupported: try to drive unsupported chipsets
260 (may crash computer or cause data corruption)
263 See Documentation/sound/alsa-configuration.rst
266 Allow the default userspace alignment fault handler
267 behaviour to be specified. Bit 0 enables warnings,
268 bit 1 enables fixups, and bit 2 sends a segfault.
270 align_va_addr= [X86-64]
271 Align virtual addresses by clearing slice [14:12] when
272 allocating a VMA at process creation time. This option
273 gives you up to 3% performance improvement on AMD F15h
274 machines (where it is enabled by default) for a
275 CPU-intensive style benchmark, and it can vary highly in
276 a microbenchmark depending on workload and compiler.
278 32: only for 32-bit processes
279 64: only for 64-bit processes
280 on: enable for both 32- and 64-bit processes
281 off: disable for both 32- and 64-bit processes
283 alloc_snapshot [FTRACE]
284 Allocate the ftrace snapshot buffer on boot up when the
285 main buffer is allocated. This is handy if debugging
286 and you need to use tracing_snapshot() on boot up, and
287 do not want to use tracing_snapshot_alloc() as it needs
288 to be done where GFP_KERNEL allocations are allowed.
290 allow_mismatched_32bit_el0 [ARM64]
291 Allow execve() of 32-bit applications and setting of the
292 PER_LINUX32 personality on systems where only a strict
293 subset of the CPUs support 32-bit EL0. When this
294 parameter is present, the set of CPUs supporting 32-bit
295 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
296 and hot-unplug operations may be restricted.
298 See Documentation/arm64/asymmetric-32bit.rst for more
301 amd_iommu= [HW,X86-64]
302 Pass parameters to the AMD IOMMU driver in the system.
304 fullflush - enable flushing of IO/TLB entries when
305 they are unmapped. Otherwise they are
306 flushed before they will be reused, which
308 off - do not initialize any AMD IOMMU found in
310 force_isolation - Force device isolation for all
311 devices. The IOMMU driver is not
312 allowed anymore to lift isolation
313 requirements as needed. This option
314 does not override iommu=pt
315 force_enable - Force enable the IOMMU on platforms known
316 to be buggy with IOMMU enabled. Use this
319 amd_iommu_dump= [HW,X86-64]
320 Enable AMD IOMMU driver option to dump the ACPI table
321 for AMD IOMMU. With this option enabled, AMD IOMMU
322 driver will print ACPI tables for AMD IOMMU during
323 IOMMU initialization.
325 amd_iommu_intr= [HW,X86-64]
326 Specifies one of the following AMD IOMMU interrupt
328 legacy - Use legacy interrupt remapping mode.
329 vapic - Use virtual APIC mode, which allows IOMMU
330 to inject interrupts directly into guest.
331 This mode requires kvm-amd.avic=1.
332 (Default when IOMMU HW support is present.)
334 amijoy.map= [HW,JOY] Amiga joystick support
335 Map of devices attached to JOY0DAT and JOY1DAT
337 See also Documentation/input/joydev/joystick.rst
339 analog.map= [HW,JOY] Analog joystick and gamepad support
340 Specifies type or capabilities of an analog joystick
341 connected to one of 16 gameports
342 Format: <type1>,<type2>,..<type16>
345 Power management functions (SPARCstation-4/5 + deriv.)
347 Disable APC CPU standby support. SPARCstation-Fox does
348 not play well with APC CPU idle - disable it if you have
349 APC and your system crashes randomly.
351 apic= [APIC,X86] Advanced Programmable Interrupt Controller
352 Change the output verbosity while booting
353 Format: { quiet (default) | verbose | debug }
354 Change the amount of debugging information output
355 when initialising the APIC and IO-APIC components.
356 For X86-32, this can also be used to specify an APIC
358 Format: apic=driver_name
359 Examples: apic=bigsmp
361 apic_extnmi= [APIC,X86] External NMI delivery setting
362 Format: { bsp (default) | all | none }
363 bsp: External NMI is delivered only to CPU 0
364 all: External NMIs are broadcast to all CPUs as a
366 none: External NMI is masked for all CPUs. This is
367 useful so that a dump capture kernel won't be
371 See Documentation/networking/ipv6.rst.
373 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
374 Limit apic dumping. The parameter defines the maximal
375 number of local apics being dumped. Also it is possible
376 to set it to "all" by meaning -- no limit here.
377 Format: { 1 (default) | 2 | ... | all }.
378 The parameter valid if only apic=debug or
379 apic=verbose is specified.
380 Example: apic=debug show_lapic=all
382 apm= [APM] Advanced Power Management
383 See header of arch/x86/kernel/apm_32.c.
385 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
386 Format: <io>,<irq>,<nodeID>
388 arm64.nobti [ARM64] Unconditionally disable Branch Target
389 Identification support
391 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
394 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
399 atarimouse= [HW,MOUSE] Atari Mouse
401 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
402 EzKey and similar keyboards
404 atkbd.reset= [HW] Reset keyboard during initialization
406 atkbd.set= [HW] Select keyboard code set
407 Format: <int> (2 = AT (default), 3 = PS/2)
409 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
412 atkbd.softraw= [HW] Choose between synthetic and real raw mode
413 Format: <bool> (0 = real, 1 = synthetic (default))
415 atkbd.softrepeat= [HW]
416 Use software keyboard repeat
418 audit= [KNL] Enable the audit sub-system
419 Format: { "0" | "1" | "off" | "on" }
420 0 | off - kernel audit is disabled and can not be
421 enabled until the next reboot
422 unset - kernel audit is initialized but disabled and
423 will be fully enabled by the userspace auditd.
424 1 | on - kernel audit is initialized and partially
425 enabled, storing at most audit_backlog_limit
426 messages in RAM until it is fully enabled by the
430 audit_backlog_limit= [KNL] Set the audit queue size limit.
431 Format: <int> (must be >=0)
434 bau= [X86_UV] Enable the BAU on SGI UV. The default
435 behavior is to disable the BAU (i.e. bau=0).
436 Format: { "0" | "1" }
439 unset - Disable the BAU.
441 baycom_epp= [HW,AX25]
444 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
446 See header of drivers/net/hamradio/baycom_par.c.
448 baycom_ser_fdx= [HW,AX25]
449 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
450 Format: <io>,<irq>,<mode>[,<baud>]
451 See header of drivers/net/hamradio/baycom_ser_fdx.c.
453 baycom_ser_hdx= [HW,AX25]
454 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
455 Format: <io>,<irq>,<mode>
456 See header of drivers/net/hamradio/baycom_ser_hdx.c.
458 blkdevparts= Manual partition parsing of block device(s) for
459 embedded devices based on command line input.
460 See Documentation/block/cmdline-partition.rst
462 boot_delay= Milliseconds to delay each printk during boot.
463 Values larger than 10 seconds (10000) are changed to
468 Extended command line options can be added to an initrd
469 and this will cause the kernel to look for it.
471 See Documentation/admin-guide/bootconfig.rst
474 Disable BERT OS support on buggy BIOSes.
476 bgrt_disable [ACPI][X86]
477 Disable BGRT to avoid flickering OEM logo.
479 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
480 bttv.radio= Most important insmod options are available as
482 bttv.pll= See Documentation/admin-guide/media/bttv.rst
485 bulk_remove=off [PPC] This parameter disables the use of the pSeries
486 firmware feature for flushing multiple hpte entries
489 c101= [NET] Moxa C101 synchronous serial card
491 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
492 Sometimes CPU hardware bugs make them report the cache
493 size incorrectly. The kernel will attempt work arounds
494 to fix known problems, but for some CPUs it is not
495 possible to determine what the correct size should be.
496 This option provides an override for these situations.
499 [NET] Specifies amount of time (in seconds) that
500 the kernel should wait for a network carrier. By default
501 it waits 120 seconds.
503 ca_keys= [KEYS] This parameter identifies a specific key(s) on
504 the system trusted keyring to be used for certificate
506 format: { id:<keyid> | builtin }
508 cca= [MIPS] Override the kernel pages' cache coherency
509 algorithm. Accepted values range from 0 to 7
510 inclusive. See arch/mips/include/asm/pgtable-bits.h
511 for platform specific values (SB1, Loongson3 and
514 ccw_timeout_log [S390]
515 See Documentation/s390/common_io.rst for details.
517 cgroup_disable= [KNL] Disable a particular controller or optional feature
518 Format: {name of the controller(s) or feature(s) to disable}
519 The effects of cgroup_disable=foo are:
520 - foo isn't auto-mounted if you mount all cgroups in
522 - foo isn't visible as an individually mountable
524 - if foo is an optional feature then the feature is
525 disabled and corresponding cgroup files are not
527 {Currently only "memory" controller deal with this and
528 cut the overhead, others just disable the usage. So
529 only cgroup_disable=memory is actually worthy}
530 Specifying "pressure" disables per-cgroup pressure
531 stall information accounting feature
533 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
534 Format: { { controller | "all" | "named" }
535 [,{ controller | "all" | "named" }...] }
536 Like cgroup_disable, but only applies to cgroup v1;
537 the blacklisted controllers remain available in cgroup2.
538 "all" blacklists all controllers and "named" disables
539 named mounts. Specifying both "all" and "named" disables
542 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
544 nosocket -- Disable socket memory accounting.
545 nokmem -- Disable kernel memory accounting.
547 checkreqprot [SELINUX] Set initial checkreqprot flag value.
548 Format: { "0" | "1" }
549 See security/selinux/Kconfig help text.
550 0 -- check protection applied by kernel (includes
551 any implied execute protection).
552 1 -- check protection requested by application.
553 Default value is set via a kernel config option.
554 Value can be changed at runtime via
555 /sys/fs/selinux/checkreqprot.
556 Setting checkreqprot to 1 is deprecated.
559 See Documentation/s390/common_io.rst for details.
562 Prevents the clock framework from automatically gating
563 clocks that have not been explicitly enabled by a Linux
564 device driver but are enabled in hardware at reset or
565 by the bootloader/firmware. Note that this does not
566 force such clocks to be always-on nor does it reserve
567 those clocks in any way. This parameter is useful for
568 debug and development, but should not be needed on a
569 platform with proper driver support. For more
570 information, see Documentation/driver-api/clk.rst.
572 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
574 Forces specified clocksource (if available) to be used
575 when calculating gettimeofday(). If specified
576 clocksource is not available, it defaults to PIT.
577 Format: { pit | tsc | cyclone | pmtmr }
579 clocksource= Override the default clocksource
581 Override the default clocksource and use the clocksource
582 with the name specified.
583 Some clocksource names to choose from, depending on
585 [all] jiffies (this is the base, fallback clocksource)
587 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
588 pxa_timer,timer3,32k_counter,timer0_1
589 [X86-32] pit,hpet,tsc;
590 scx200_hrt on Geode; cyclone on IBM x440
598 clocksource.arm_arch_timer.evtstrm=
601 Enable/disable the eventstream feature of the ARM
602 architected timer so that code using WFE-based polling
603 loops can be debugged more effectively on production
606 clocksource.max_cswd_read_retries= [KNL]
607 Number of clocksource_watchdog() retries due to
608 external delays before the clock will be marked
609 unstable. Defaults to three retries, that is,
610 four attempts to read the clock under test.
612 clocksource.verify_n_cpus= [KNL]
613 Limit the number of CPUs checked for clocksources
614 marked with CLOCK_SOURCE_VERIFY_PERCPU that
615 are marked unstable due to excessive skew.
616 A negative value says to check all CPUs, while
617 zero says not to check any. Values larger than
618 nr_cpu_ids are silently truncated to nr_cpu_ids.
619 The actual CPUs are chosen randomly, with
620 no replacement if the same CPU is chosen twice.
622 clocksource-wdtest.holdoff= [KNL]
623 Set the time in seconds that the clocksource
624 watchdog test waits before commencing its tests.
625 Defaults to zero when built as a module and to
626 10 seconds when built into the kernel.
628 clearcpuid=BITNUM[,BITNUM...] [X86]
629 Disable CPUID feature X for the kernel. See
630 arch/x86/include/asm/cpufeatures.h for the valid bit
631 numbers. Note the Linux specific bits are not necessarily
632 stable over kernel options, but the vendor specific
634 Also note that user programs calling CPUID directly
635 or using the feature without checking anything
636 will still see it. This just prevents it from
637 being used by the kernel or shown in /proc/cpuinfo.
638 Also note the kernel might malfunction if you disable
641 cma=nn[MG]@[start[MG][-end[MG]]]
643 Sets the size of kernel global memory area for
644 contiguous memory allocations and optionally the
645 placement constraint by the physical address range of
646 memory allocations. A value of 0 disables CMA
647 altogether. For more information, see
648 kernel/dma/contiguous.c
652 Sets the size of kernel per-numa memory area for
653 contiguous memory allocations. A value of 0 disables
654 per-numa CMA altogether. And If this option is not
655 specificed, the default value is 0.
656 With per-numa CMA enabled, DMA users on node nid will
657 first try to allocate buffer from the pernuma area
658 which is located in node nid, if the allocation fails,
659 they will fallback to the global default memory area.
661 cmo_free_hint= [PPC] Format: { yes | no }
662 Specify whether pages are marked as being inactive
663 when they are freed. This is used in CMO environments
664 to determine OS memory pressure for page stealing by
668 coherent_pool=nn[KMG] [ARM,KNL]
669 Sets the size of memory pool for coherent, atomic dma
670 allocations, by default set to 256K.
672 com20020= [HW,NET] ARCnet - COM20020 chipset
674 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
676 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
680 ARCnet - COM90xx chipset (memory-mapped buffers)
681 Format: <io>[,<irq>[,<memstart>]]
683 condev= [HW,S390] console device
686 console= [KNL] Output console device and options.
688 tty<n> Use the virtual console device <n>.
692 Use the specified serial port. The options are of
693 the form "bbbbpnf", where "bbbb" is the baud rate,
694 "p" is parity ("n", "o", or "e"), "n" is number of
695 bits, and "f" is flow control ("r" for RTS or
696 omit it). Default is "9600n8".
698 See Documentation/admin-guide/serial-console.rst for more
700 Documentation/networking/netconsole.rst for an
703 uart[8250],io,<addr>[,options]
704 uart[8250],mmio,<addr>[,options]
705 uart[8250],mmio16,<addr>[,options]
706 uart[8250],mmio32,<addr>[,options]
707 uart[8250],0x<addr>[,options]
708 Start an early, polled-mode console on the 8250/16550
709 UART at the specified I/O port or MMIO address,
710 switching to the matching ttyS device later.
711 MMIO inter-register address stride is either 8-bit
712 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
713 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
714 to be equivalent to 'mmio'. 'options' are specified in
715 the same format described for ttyS above; if unspecified,
716 the h/w is not re-initialized.
718 hvc<n> Use the hypervisor console device <n>. This is for
719 both Xen and PowerPC hypervisors.
721 If the device connected to the port is not a TTY but a braille
722 device, prepend "brl," before the device type, for instance
724 For now, only VisioBraille is supported.
727 [KNL] Change console messages format
729 By default we print messages on consoles in
730 "[time stamp] text\n" format (time stamp may not be
731 printed, depending on CONFIG_PRINTK_TIME or
732 `printk_time' param).
734 Switch to syslog format: "<%u>[time stamp] text\n"
735 IOW, each message will have a facility and loglevel
736 prefix. The format is similar to one used by syslog()
737 syscall, or to executing "dmesg -S --raw" or to reading
740 consoleblank= [KNL] The console blank (screen saver) timeout in
741 seconds. A value of 0 disables the blank timer.
745 [KNL] Change the default value for
746 /proc/<pid>/coredump_filter.
747 See also Documentation/filesystems/proc.rst.
749 coresight_cpu_debug.enable
752 Enable/disable the CPU sampling based debugging.
753 0: default value, disable debugging
754 1: enable debugging at boot time
756 cpuidle.off=1 [CPU_IDLE]
757 disable the cpuidle sub-system
760 [CPU_IDLE] Name of the cpuidle governor to use.
762 cpufreq.off=1 [CPU_FREQ]
763 disable the cpufreq sub-system
765 cpufreq.default_governor=
766 [CPU_FREQ] Name of the default cpufreq governor or
767 policy to use. This governor must be registered in the
768 kernel before the cpufreq driver probes.
771 [X86] Delay for N microsec between assert and de-assert
772 of APIC INIT to start processors. This delay occurs
773 on every CPU online, such as boot, and resume from suspend.
776 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
778 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
780 crashkernel=size[KMG][@offset[KMG]]
781 [KNL] Using kexec, Linux can switch to a 'crash kernel'
782 upon panic. This parameter reserves the physical
783 memory region [offset, offset + size] for that kernel
784 image. If '@offset' is omitted, then a suitable offset
785 is selected automatically.
786 [KNL, X86-64] Select a region under 4G first, and
787 fall back to reserve region above 4G when '@offset'
788 hasn't been specified.
789 See Documentation/admin-guide/kdump/kdump.rst for further details.
791 crashkernel=range1:size1[,range2:size2,...][@offset]
792 [KNL] Same as above, but depends on the memory
793 in the running system. The syntax of range is
794 start-[end] where start and end are both
795 a memory unit (amount[KMG]). See also
796 Documentation/admin-guide/kdump/kdump.rst for an example.
798 crashkernel=size[KMG],high
799 [KNL, X86-64] range could be above 4G. Allow kernel
800 to allocate physical memory region from top, so could
801 be above 4G if system have more than 4G ram installed.
802 Otherwise memory region will be allocated below 4G, if
804 It will be ignored if crashkernel=X is specified.
805 crashkernel=size[KMG],low
806 [KNL, X86-64] range under 4G. When crashkernel=X,high
807 is passed, kernel could allocate physical memory region
808 above 4G, that cause second kernel crash on system
809 that require some amount of low memory, e.g. swiotlb
810 requires at least 64M+32K low memory, also enough extra
811 low memory is needed to make sure DMA buffers for 32-bit
812 devices won't run out. Kernel would try to allocate at
813 at least 256M below 4G automatically.
814 This one let user to specify own low range under 4G
815 for second kernel instead.
816 0: to disable low allocation.
817 It will be ignored when crashkernel=X,high is not used
818 or memory reserved is below 4G.
821 [KNL] Disable crypto self-tests
826 cs89x0_media= [HW,NET]
827 Format: { rj45 | aui | bnc }
829 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
830 handling. When switched on, additional debug data is
831 printed to the console in case a hanging CPU is
832 detected, and that CPU is pinged again in order to try
833 to resolve the hang situation.
834 0: disable csdlock debugging (default)
835 1: enable basic csdlock debugging (minor impact)
836 ext: enable extended csdlock debugging (more impact,
840 See header of drivers/s390/block/dasd_devmap.c.
842 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
843 (one device per port)
844 Format: <port#>,<type>
845 See also Documentation/input/devices/joystick-parport.rst
847 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
849 Documentation/admin-guide/dynamic-debug-howto.rst for
850 details. Deprecated, see dyndbg.
852 debug [KNL] Enable kernel debugging (events log level).
855 [KNL] Enable printing [hashed] pointers early in the
856 boot sequence. If enabled, we use a weak hash instead
857 of siphash to hash pointers. Use this option if you are
858 seeing instances of '(___ptrval___)') and need to see a
859 value (hashed pointer) instead. Cryptographically
860 insecure, please do not use on production kernels.
863 [KNL] verbose locking self-tests
865 Print debugging info while doing the locking API
867 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
868 (no extra messages), setting it to -1 (all bits set)
869 will print _a_lot_ more information - normally only
870 useful to lockdep developers.
872 debug_objects [KNL] Enable object debugging
875 [KNL] Disable object debugging
877 debug_guardpage_minorder=
878 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
879 parameter allows control of the order of pages that will
880 be intentionally kept free (and hence protected) by the
881 buddy allocator. Bigger value increase the probability
882 of catching random memory corruption, but reduce the
883 amount of memory for normal system use. The maximum
884 possible value is MAX_ORDER/2. Setting this parameter
885 to 1 or 2 should be enough to identify most random
886 memory corruption problems caused by bugs in kernel or
887 driver code when a CPU writes to (or reads from) a
888 random memory location. Note that there exists a class
889 of memory corruptions problems caused by buggy H/W or
890 F/W or by drivers badly programing DMA (basically when
891 memory is written at bus level and the CPU MMU is
892 bypassed) which are not detectable by
893 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
894 tracking down these problems.
897 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
898 enables the feature at boot time. By default, it is
899 disabled and the system will work mostly the same as a
900 kernel built without CONFIG_DEBUG_PAGEALLOC.
901 Note: to get most of debug_pagealloc error reports, it's
902 useful to also enable the page_owner functionality.
903 on: enable the feature
905 debugfs= [KNL] This parameter enables what is exposed to userspace
906 and debugfs internal clients.
907 Format: { on, no-mount, off }
908 on: All functions are enabled.
910 Filesystem is not registered but kernel clients can
911 access APIs and a crashkernel can be used to read
912 its content. There is nothing to mount.
913 off: Filesystem is not registered and clients
914 get a -EPERM as result when trying to register files
915 or directories within debugfs.
916 This is equivalent of the runtime functionality if
917 debugfs was not enabled in the kernel at all.
918 Default value is set in build-time with a kernel configuration.
920 debugpat [X86] Enable PAT debugging
922 decnet.addr= [HW,NET]
923 Format: <area>[,<node>]
924 See also Documentation/networking/decnet.rst.
927 [HW] The size of the default HugeTLB page. This is
928 the size represented by the legacy /proc/ hugepages
929 APIs. In addition, this is the default hugetlb size
930 used for shmget(), mmap() and mounting hugetlbfs
931 filesystems. If not specified, defaults to the
932 architecture's default huge page size. Huge page
933 sizes are architecture dependent. See also
934 Documentation/admin-guide/mm/hugetlbpage.rst.
937 deferred_probe_timeout=
938 [KNL] Debugging option to set a timeout in seconds for
939 deferred probe to give up waiting on dependencies to
940 probe. Only specific dependencies (subsystems or
941 drivers) that have opted in will be ignored. A timeout of 0
942 will timeout at the end of initcalls. This option will also
943 dump out devices still on the deferred probe list after
947 Format: { on | off | def_only | inf_only | always }
948 on: s390 zlib hardware support for compression on
949 level 1 and decompression (default)
950 off: No s390 zlib hardware support
951 def_only: s390 zlib hardware support for deflate
952 only (compression on level 1)
953 inf_only: s390 zlib hardware support for inflate
955 always: Same as 'on' but ignores the selected compression
956 level always using hardware support (used for debugging)
959 Set number of hash buckets for dentry cache.
961 disable_1tb_segments [PPC]
962 Disables the use of 1TB hash page table segments. This
963 causes the kernel to fall back to 256MB segments which
964 can be useful when debugging issues that require an SLB
968 Limits the number of kernel SLB entries, and flushes
969 them frequently to increase the rate of SLB faults
973 See Documentation/networking/ipv6.rst.
976 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
977 hardening is enabled for this boot. Hardened
978 usercopy checking is used to protect the kernel
979 from reading or writing beyond known memory
980 allocation boundaries as a proactive defense
981 against bounds-checking flaws in the kernel's
982 copy_to_user()/copy_from_user() interface.
983 on Perform hardened usercopy checks (default).
984 off Disable hardened usercopy checks.
987 Disable RADIX MMU mode on POWER9
989 radix_hcall_invalidate=on [PPC/PSERIES]
990 Disable RADIX GTSE feature and use hcall for TLB
994 Disable TLBIE instruction. Currently does not work
995 with KVM, with HASH MMU, or with coherent accelerators.
997 disable_cpu_apicid= [X86,APIC,SMP]
999 The number of initial APIC ID for the
1000 corresponding CPU to be disabled at boot,
1001 mostly used for the kdump 2nd kernel to
1002 disable BSP to wake up multiple CPUs without
1003 causing system reset or hang due to sending
1004 INIT from AP to BSP.
1006 disable_ddw [PPC/PSERIES]
1007 Disable Dynamic DMA Window support. Use this
1008 to workaround buggy firmware.
1010 disable_ipv6= [IPV6]
1011 See Documentation/networking/ipv6.rst.
1013 disable_mtrr_cleanup [X86]
1014 The kernel tries to adjust MTRR layout from continuous
1015 to discrete, to make X server driver able to add WB
1016 entry later. This parameter disables that.
1018 disable_mtrr_trim [X86, Intel and AMD only]
1019 By default the kernel will trim any uncacheable
1020 memory out of your available memory pool based on
1021 MTRR settings. This parameter disables that behavior,
1022 possibly causing your machine to run very slowly.
1024 disable_timer_pin_1 [X86]
1025 Disable PIN 1 of APIC timer
1026 Can be useful to work around chipset bugs.
1028 dis_ucode_ldr [X86] Disable the microcode loader.
1030 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1031 this option disables the debugging code at boot.
1033 dma_debug_entries=<number>
1034 This option allows to tune the number of preallocated
1035 entries for DMA-API debugging code. One entry is
1036 required per DMA-API allocation. Use this if the
1037 DMA-API debugging code disables itself because the
1038 architectural default is too low.
1040 dma_debug_driver=<driver_name>
1041 With this option the DMA-API debugging driver
1042 filter feature can be enabled at boot time. Just
1043 pass the driver to filter for as the parameter.
1044 The filter can be disabled or changed to another
1045 driver later using sysfs.
1047 driver_async_probe= [KNL]
1048 List of driver names to be probed asynchronously.
1049 Format: <driver_name1>,<driver_name2>...
1051 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1052 Broken monitors, graphic adapters, KVMs and EDIDless
1053 panels may send no or incorrect EDID data sets.
1054 This parameter allows to specify an EDID data sets
1055 in the /lib/firmware directory that are used instead.
1056 Generic built-in EDID data sets are used, if one of
1057 edid/1024x768.bin, edid/1280x1024.bin,
1058 edid/1680x1050.bin, or edid/1920x1080.bin is given
1059 and no file with the same name exists. Details and
1060 instructions how to build your own EDID data are
1061 available in Documentation/admin-guide/edid.rst. An EDID
1062 data set will only be used for a particular connector,
1063 if its name and a colon are prepended to the EDID
1064 name. Each connector may use a unique EDID data
1065 set by separating the files with a comma. An EDID
1066 data set with no connector name will be used for
1067 any connectors not explicitly specified.
1072 Format: {"off" | "known"}
1073 Control how the dt_cpu_ftrs device-tree binding is
1074 used for CPU feature discovery and setup (if it
1076 off: Do not use it, fall back to legacy cpu table.
1077 known: Do not pass through unknown features to guests
1078 or userspace, only those that the kernel is aware of.
1080 dump_apple_properties [X86]
1081 Dump name and content of EFI device properties on
1082 x86 Macs. Useful for driver authors to determine
1083 what data is available or for reverse-engineering.
1085 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1086 <module>.dyndbg[="val"]
1087 Enable debug messages at boot time. See
1088 Documentation/admin-guide/dynamic-debug-howto.rst
1091 nopku [X86] Disable Memory Protection Keys CPU feature found
1094 <module>.async_probe [KNL]
1095 Enable asynchronous probe on this module.
1097 early_ioremap_debug [KNL]
1098 Enable debug messages in early_ioremap support. This
1099 is useful for tracking down temporary early mappings
1100 which are not unmapped.
1102 earlycon= [KNL] Output early console device and options.
1104 When used with no options, the early console is
1105 determined by stdout-path property in device tree's
1106 chosen node or the ACPI SPCR table if supported by
1109 cdns,<addr>[,options]
1110 Start an early, polled-mode console on a Cadence
1111 (xuartps) serial port at the specified address. Only
1112 supported option is baud rate. If baud rate is not
1113 specified, the serial port must already be setup and
1116 uart[8250],io,<addr>[,options]
1117 uart[8250],mmio,<addr>[,options]
1118 uart[8250],mmio32,<addr>[,options]
1119 uart[8250],mmio32be,<addr>[,options]
1120 uart[8250],0x<addr>[,options]
1121 Start an early, polled-mode console on the 8250/16550
1122 UART at the specified I/O port or MMIO address.
1123 MMIO inter-register address stride is either 8-bit
1124 (mmio) or 32-bit (mmio32 or mmio32be).
1125 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1126 to be equivalent to 'mmio'. 'options' are specified
1127 in the same format described for "console=ttyS<n>"; if
1128 unspecified, the h/w is not initialized.
1132 Start an early, polled-mode console on a pl011 serial
1133 port at the specified address. The pl011 serial port
1134 must already be setup and configured. Options are not
1135 yet supported. If 'mmio32' is specified, then only
1136 the driver will use only 32-bit accessors to read/write
1137 the device registers.
1140 Start an early console on a litex serial port at the
1141 specified address. The serial port must already be
1142 setup and configured. Options are not yet supported.
1145 Start an early, polled-mode console on a meson serial
1146 port at the specified address. The serial port must
1147 already be setup and configured. Options are not yet
1151 Start an early, polled-mode console on an msm serial
1152 port at the specified address. The serial port
1153 must already be setup and configured. Options are not
1156 msm_serial_dm,<addr>
1157 Start an early, polled-mode console on an msm serial
1158 dm port at the specified address. The serial port
1159 must already be setup and configured. Options are not
1163 Start an early, polled-mode console on a serial port
1164 of an Actions Semi SoC, such as S500 or S900, at the
1165 specified address. The serial port must already be
1166 setup and configured. Options are not yet supported.
1169 Start an early, polled-mode console on a serial port
1170 of an RDA Micro SoC, such as RDA8810PL, at the
1171 specified address. The serial port must already be
1172 setup and configured. Options are not yet supported.
1175 Use RISC-V SBI (Supervisor Binary Interface) for early
1178 smh Use ARM semihosting calls for early console.
1186 Use early console provided by serial driver available
1187 on Samsung SoCs, requires selecting proper type and
1188 a correct base address of the selected UART port. The
1189 serial port must already be setup and configured.
1190 Options are not yet supported.
1193 Start an early, polled-mode console on a lantiq serial
1194 (lqasc) port at the specified address. The serial port
1195 must already be setup and configured. Options are not
1200 Use early console provided by Freescale LP UART driver
1201 found on Freescale Vybrid and QorIQ LS1021A processors.
1202 A valid base address must be provided, and the serial
1203 port must already be setup and configured.
1207 Start an early, polled-mode, output-only console on the
1208 Freescale i.MX UART at the specified address. The UART
1209 must already be setup and configured.
1212 Start an early, polled-mode console on the
1213 Armada 3700 serial port at the specified
1214 address. The serial port must already be setup
1215 and configured. Options are not yet supported.
1218 Start an early, polled-mode console on a Qualcomm
1219 Generic Interface (GENI) based serial port at the
1220 specified address. The serial port must already be
1221 setup and configured. Options are not yet supported.
1224 Start an early, unaccelerated console on the EFI
1225 memory mapped framebuffer (if available). On cache
1226 coherent non-x86 systems that use system memory for
1227 the framebuffer, pass the 'ram' option so that it is
1228 mapped with the correct attributes.
1231 Use early console provided by Freescale LINFlexD UART
1232 serial driver for NXP S32V234 SoCs. A valid base
1233 address must be provided, and the serial port must
1234 already be setup and configured.
1236 earlyprintk= [X86,SH,ARM,M68k,S390]
1240 earlyprintk=serial[,ttySn[,baudrate]]
1241 earlyprintk=serial[,0x...[,baudrate]]
1242 earlyprintk=ttySn[,baudrate]
1243 earlyprintk=dbgp[debugController#]
1244 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1245 earlyprintk=xdbc[xhciController#]
1247 earlyprintk is useful when the kernel crashes before
1248 the normal console is initialized. It is not enabled by
1249 default because it has some cosmetic problems.
1251 Append ",keep" to not disable it when the real console
1254 Only one of vga, efi, serial, or usb debug port can
1257 Currently only ttyS0 and ttyS1 may be specified by
1258 name. Other I/O ports may be explicitly specified
1259 on some architectures (x86 and arm at least) by
1260 replacing ttySn with an I/O port address, like this:
1261 earlyprintk=serial,0x1008,115200
1262 You can find the port for a given device in
1263 /proc/tty/driver/serial:
1264 2: uart:ST16650V2 port:00001008 irq:18 ...
1266 Interaction with the standard serial driver is not
1269 The VGA and EFI output is eventually overwritten by
1272 The xen output can only be used by Xen PV guests.
1274 The sclp output can only be used on s390.
1276 The optional "force" to "pciserial" enables use of a
1277 PCI device even when its classcode is not of the
1280 edac_report= [HW,EDAC] Control how to report EDAC event
1281 Format: {"on" | "off" | "force"}
1282 on: enable EDAC to report H/W event. May be overridden
1283 by other higher priority error reporting module.
1284 off: disable H/W event reporting through EDAC.
1285 force: enforce the use of EDAC to report H/W event.
1288 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1291 This is designed to be used in conjunction with
1292 the boot argument: earlyprintk=vga
1294 This parameter works in place of the kgdboc parameter
1295 but can only be used if the backing tty is available
1296 very early in the boot process. For early debugging
1297 via a serial port see kgdboc_earlycon instead.
1300 Format: {"off" | "on" | "skip[mbr]"}
1303 Format: { "debug", "disable_early_pci_dma",
1304 "nochunk", "noruntime", "nosoftreserve",
1305 "novamap", "no_disable_early_pci_dma" }
1306 debug: enable misc debug output.
1307 disable_early_pci_dma: disable the busmaster bit on all
1308 PCI bridges while in the EFI boot stub.
1309 nochunk: disable reading files in "chunks" in the EFI
1310 boot stub, as chunking can cause problems with some
1311 firmware implementations.
1312 noruntime : disable EFI runtime services support
1313 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1314 attribute may cause the kernel to reserve the
1315 memory range for a memory mapping driver to
1316 claim. Specify efi=nosoftreserve to disable this
1317 reservation and treat the memory by its base type
1318 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1319 novamap: do not call SetVirtualAddressMap().
1320 no_disable_early_pci_dma: Leave the busmaster bit set
1321 on all PCI bridges while in the EFI boot stub
1323 efi_no_storage_paranoia [EFI; X86]
1324 Using this parameter you can use more than 50% of
1325 your efi variable storage. Use this parameter only if
1326 you are really sure that your UEFI does sane gc and
1327 fulfills the spec otherwise your board may brick.
1329 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1330 Add arbitrary attribute to specific memory range by
1331 updating original EFI memory map.
1332 Region of memory which aa attribute is added to is
1335 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1336 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1337 attribute is added to range 0x100000000-0x180000000 and
1338 0x10a0000000-0x1120000000.
1340 If efi_fake_mem=8G@9G:0x40000 is specified, the
1341 EFI_MEMORY_SP(0x40000) attribute is added to
1342 range 0x240000000-0x43fffffff.
1344 Using this parameter you can do debugging of EFI memmap
1345 related features. For example, you can do debugging of
1346 Address Range Mirroring feature even if your box
1347 doesn't support it, or mark specific memory as
1350 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1351 that is to be dynamically loaded by Linux. If there are
1352 multiple variables with the same name but with different
1353 vendor GUIDs, all of them will be loaded. See
1354 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1357 eisa_irq_edge= [PARISC,HW]
1358 See header of drivers/parisc/eisa.c.
1361 See comment before function elanfreq_setup() in
1362 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1364 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1365 Specifies physical address of start of kernel core
1366 image elf header and optionally the size. Generally
1367 kexec loader will pass this option to capture kernel.
1368 See Documentation/admin-guide/kdump/kdump.rst for details.
1370 enable_mtrr_cleanup [X86]
1371 The kernel tries to adjust MTRR layout from continuous
1372 to discrete, to make X server driver able to add WB
1373 entry later. This parameter enables that.
1375 enable_timer_pin_1 [X86]
1376 Enable PIN 1 of APIC timer
1377 Can be useful to work around chipset bugs
1378 (in particular on some ATI chipsets).
1379 The kernel tries to set a reasonable default.
1381 enforcing [SELINUX] Set initial enforcing status.
1383 See security/selinux/Kconfig help text.
1384 0 -- permissive (log only, no denials).
1385 1 -- enforcing (deny and log).
1387 Value can be changed at runtime via
1388 /sys/fs/selinux/enforce.
1391 Disable Error Record Serialization Table (ERST)
1394 ether= [HW,NET] Ethernet cards parameters
1395 This option is obsoleted by the "netdev=" option, which
1396 has equivalent usage. See its documentation for details.
1400 Permit 'security.evm' to be updated regardless of
1401 current integrity status.
1406 fail_make_request=[KNL]
1407 General fault injection mechanism.
1408 Format: <interval>,<probability>,<space>,<times>
1409 See also Documentation/fault-injection/.
1412 Format: { initns | none }
1413 See Documentation/admin-guide/sysctl/net.rst for
1414 fb_tunnels_only_for_init_ns
1417 See Documentation/admin-guide/blockdev/floppy.rst.
1419 force_pal_cache_flush
1420 [IA-64] Avoid check_sal_cache_flush which may hang on
1421 buggy SAL_CACHE_FLUSH implementations. Using this
1422 parameter will force ia64_sal_cache_flush to call
1423 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1426 Forcefully enable Physical Address Extension (PAE).
1427 Many Pentium M systems disable PAE but may have a
1428 functionally usable PAE implementation.
1429 Warning: use of this parameter will taint the kernel
1430 and may cause unknown problems.
1433 [FTRACE] will set and start the specified tracer
1434 as early as possible in order to facilitate early
1437 ftrace_dump_on_oops[=orig_cpu]
1438 [FTRACE] will dump the trace buffers on oops.
1439 If no parameter is passed, ftrace will dump
1440 buffers of all CPUs, but if you pass orig_cpu, it will
1441 dump only the buffer of the CPU that triggered the
1444 ftrace_filter=[function-list]
1445 [FTRACE] Limit the functions traced by the function
1446 tracer at boot up. function-list is a comma-separated
1447 list of functions. This list can be changed at run
1448 time by the set_ftrace_filter file in the debugfs
1451 ftrace_notrace=[function-list]
1452 [FTRACE] Do not trace the functions specified in
1453 function-list. This list can be changed at run time
1454 by the set_ftrace_notrace file in the debugfs
1457 ftrace_graph_filter=[function-list]
1458 [FTRACE] Limit the top level callers functions traced
1459 by the function graph tracer at boot up.
1460 function-list is a comma-separated list of functions
1461 that can be changed at run time by the
1462 set_graph_function file in the debugfs tracing directory.
1464 ftrace_graph_notrace=[function-list]
1465 [FTRACE] Do not trace from the functions specified in
1466 function-list. This list is a comma-separated list of
1467 functions that can be changed at run time by the
1468 set_graph_notrace file in the debugfs tracing directory.
1470 ftrace_graph_max_depth=<uint>
1471 [FTRACE] Used with the function graph tracer. This is
1472 the max depth it will trace into a function. This value
1473 can be changed at run time by the max_graph_depth file
1474 in the tracefs tracing directory. default: 0 (no limit)
1476 fw_devlink= [KNL] Create device links between consumer and supplier
1477 devices by scanning the firmware to infer the
1478 consumer/supplier relationships. This feature is
1479 especially useful when drivers are loaded as modules as
1480 it ensures proper ordering of tasks like device probing
1481 (suppliers first, then consumers), supplier boot state
1482 clean up (only after all consumers have probed),
1483 suspend/resume & runtime PM (consumers first, then
1485 Format: { off | permissive | on | rpm }
1486 off -- Don't create device links from firmware info.
1487 permissive -- Create device links from firmware info
1488 but use it only for ordering boot state clean
1489 up (sync_state() calls).
1490 on -- Create device links from firmware info and use it
1491 to enforce probe and suspend/resume ordering.
1492 rpm -- Like "on", but also use to order runtime PM.
1494 fw_devlink.strict=<bool>
1495 [KNL] Treat all inferred dependencies as mandatory
1496 dependencies. This only applies for fw_devlink=on|rpm.
1500 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1501 support via parallel port (up to 5 devices per port)
1502 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1503 See also Documentation/input/devices/joystick-parport.rst
1507 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1511 gcov_persist= [GCOV] When non-zero (default), profiling data for
1512 kernel modules is saved and remains accessible via
1513 debugfs, even when the module is unloaded/reloaded.
1514 When zero, profiling data is discarded and associated
1515 debugfs files are removed at module unload time.
1517 goldfish [X86] Enable the goldfish android emulator platform.
1518 Don't use this when you are not running on the
1521 gpio-mockup.gpio_mockup_ranges
1522 [HW] Sets the ranges of gpiochip of for this device.
1523 Format: <start1>,<end1>,<start2>,<end2>...
1524 gpio-mockup.gpio_mockup_named_lines
1525 [HW] Let the driver know GPIO lines should be named.
1527 gpt [EFI] Forces disk with valid GPT signature but
1528 invalid Protective MBR to be treated as GPT. If the
1529 primary GPT is corrupted, it enables the backup/alternate
1530 GPT to be used instead.
1532 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1533 the "Enable 0" bit of the configuration register.
1536 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1537 the "Enable 0" bit of the configuration register.
1540 grcan.select= [HW] Select which physical interface to use.
1543 grcan.txsize= [HW] Sets the size of the tx buffer.
1544 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1546 grcan.rxsize= [HW] Sets the size of the rx buffer.
1547 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1550 hardlockup_all_cpu_backtrace=
1551 [KNL] Should the hard-lockup detector generate
1552 backtraces on all cpus.
1555 hashdist= [KNL,NUMA] Large hashes allocated during boot
1556 are distributed across NUMA nodes. Defaults on
1557 for 64-bit NUMA, off otherwise.
1558 Format: 0 | 1 (for off | on)
1560 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1562 hd= [EIDE] (E)IDE hard drive subsystem geometry
1563 Format: <cyl>,<head>,<sect>
1566 Disable Hardware Error Source Table (HEST) support;
1567 corresponding firmware-first mode error processing
1568 logic will be disabled.
1570 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1571 size of <nn>. This works even on boxes that have no
1572 highmem otherwise. This also works to reduce highmem
1573 size on bigger boxes.
1575 highres= [KNL] Enable/disable high resolution timer mode.
1576 Valid parameters: "on", "off"
1581 hpet= [X86-32,HPET] option to control HPET usage
1582 Format: { enable (default) | disable | force |
1584 disable: disable HPET and use PIT instead
1585 force: allow force enabled of undocumented chips (ICH4,
1587 verbose: show contents of HPET registers during setup
1589 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1590 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1592 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1593 of gigantic hugepages.
1596 Reserve a CMA area of given size and allocate gigantic
1597 hugepages using the CMA allocator. If enabled, the
1598 boot-time allocation of gigantic hugepages is skipped.
1600 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1601 If this follows hugepagesz (below), it specifies
1602 the number of pages of hugepagesz to be allocated.
1603 If this is the first HugeTLB parameter on the command
1604 line, it specifies the number of pages to allocate for
1605 the default huge page size. See also
1606 Documentation/admin-guide/mm/hugetlbpage.rst.
1610 [HW] The size of the HugeTLB pages. This is used in
1611 conjunction with hugepages (above) to allocate huge
1612 pages of a specific size at boot. The pair
1613 hugepagesz=X hugepages=Y can be specified once for
1614 each supported huge page size. Huge page sizes are
1615 architecture dependent. See also
1616 Documentation/admin-guide/mm/hugetlbpage.rst.
1619 hugetlb_free_vmemmap=
1620 [KNL] Reguires CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
1622 Allows heavy hugetlb users to free up some more
1623 memory (6 * PAGE_SIZE for each 2MB hugetlb page).
1624 Format: { on | off (default) }
1626 on: enable the feature
1627 off: disable the feature
1629 Built with CONFIG_HUGETLB_PAGE_FREE_VMEMMAP_DEFAULT_ON=y,
1632 This is not compatible with memory_hotplug.memmap_on_memory.
1633 If both parameters are enabled, hugetlb_free_vmemmap takes
1634 precedence over memory_hotplug.memmap_on_memory.
1637 [KNL] Should the hung task detector generate panics.
1640 A value of 1 instructs the kernel to panic when a
1641 hung task is detected. The default value is controlled
1642 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1643 option. The value selected by this boot parameter can
1644 be changed later by the kernel.hung_task_panic sysctl.
1646 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1647 terminal devices. Valid values: 0..8
1648 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1649 If specified, z/VM IUCV HVC accepts connections
1650 from listed z/VM user IDs only.
1652 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1653 which allow the hypervisor to 'idle' the
1654 guest on lock contention.
1657 Do not unregister boot console at start. This is only
1658 useful for debugging when something happens in the window
1659 between unregistering the boot console and initializing
1662 i2c_bus= [HW] Override the default board specific I2C bus speed
1663 or register an additional I2C bus that is not
1664 registered from board initialization code.
1668 i8042.debug [HW] Toggle i8042 debug mode
1669 i8042.unmask_kbd_data
1670 [HW] Enable printing of interrupt data from the KBD port
1671 (disabled by default, and as a pre-condition
1672 requires that i8042.debug=1 be enabled)
1673 i8042.direct [HW] Put keyboard port into non-translated mode
1674 i8042.dumbkbd [HW] Pretend that controller can only read data from
1675 keyboard and cannot control its state
1676 (Don't attempt to blink the leds)
1677 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1678 i8042.nokbd [HW] Don't check/create keyboard port
1679 i8042.noloop [HW] Disable the AUX Loopback command while probing
1681 i8042.nomux [HW] Don't check presence of an active multiplexing
1683 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1685 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1686 i8042.reset [HW] Reset the controller during init, cleanup and
1687 suspend-to-ram transitions, only during s2r
1688 transitions, or never reset
1689 Format: { 1 | Y | y | 0 | N | n }
1690 1, Y, y: always reset controller
1691 0, N, n: don't ever reset controller
1692 Default: only on s2r transitions on x86; most other
1693 architectures force reset to be always executed
1694 i8042.unlock [HW] Unlock (ignore) the keylock
1695 i8042.kbdreset [HW] Reset device connected to KBD port
1699 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1700 indicates that the driver is running on unsupported
1702 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1703 does not match list of supported models.
1705 [HW] Report power status in /proc/i8k
1706 (disabled by default)
1707 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1710 i915.invert_brightness=
1711 [DRM] Invert the sense of the variable that is used to
1712 set the brightness of the panel backlight. Normally a
1713 brightness value of 0 indicates backlight switched off,
1714 and the maximum of the brightness value sets the backlight
1715 to maximum brightness. If this parameter is set to 0
1716 (default) and the machine requires it, or this parameter
1717 is set to 1, a brightness value of 0 sets the backlight
1718 to maximum brightness, and the maximum of the brightness
1719 value switches the backlight off.
1720 -1 -- never invert brightness
1721 0 -- machine default
1722 1 -- force brightness inversion
1725 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1727 ide-core.nodma= [HW] (E)IDE subsystem
1728 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1729 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1730 .cdrom .chs .ignore_cable are additional options
1731 See Documentation/ide/ide.rst.
1733 ide-generic.probe-mask= [HW] (E)IDE subsystem
1735 Probe mask for legacy ISA IDE ports. Depending on
1736 platform up to 6 ports are supported, enabled by
1737 setting corresponding bits in the mask to 1. The
1738 default value is 0x0, which has a special meaning.
1739 On systems that have PCI, it triggers scanning the
1740 PCI bus for the first and the second port, which
1741 are then probed. On systems without PCI the value
1742 of 0x0 enables probing the two first ports as if it
1745 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1746 Claim all unknown PCI IDE storage controllers.
1749 Format: idle=poll, idle=halt, idle=nomwait
1750 Poll forces a polling idle loop that can slightly
1751 improve the performance of waking up a idle CPU, but
1752 will use a lot of power and make the system run hot.
1754 idle=halt: Halt is forced to be used for CPU idle.
1755 In such case C2/C3 won't be used again.
1756 idle=nomwait: Disable mwait for CPU C-states
1760 Allow force disabling of Shared Virtual Memory (SVA)
1761 support for the idxd driver. By default it is set to
1764 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1765 Format: { strict | legacy | 2008 | relaxed }
1768 Choose which programs will be accepted for execution
1769 based on the IEEE 754 NaN encoding(s) supported by
1770 the FPU and the NaN encoding requested with the value
1771 of an ELF file header flag individually set by each
1772 binary. Hardware implementations are permitted to
1773 support either or both of the legacy and the 2008 NaN
1776 Available settings are as follows:
1777 strict accept binaries that request a NaN encoding
1778 supported by the FPU
1779 legacy only accept legacy-NaN binaries, if supported
1781 2008 only accept 2008-NaN binaries, if supported
1783 relaxed accept any binaries regardless of whether
1784 supported by the FPU
1786 The FPU emulator is always able to support both NaN
1787 encodings, so if no FPU hardware is present or it has
1788 been disabled with 'nofpu', then the settings of
1789 'legacy' and '2008' strap the emulator accordingly,
1790 'relaxed' straps the emulator for both legacy-NaN and
1791 2008-NaN, whereas 'strict' enables legacy-NaN only on
1792 legacy processors and both NaN encodings on MIPS32 or
1795 The setting for ABS.fmt/NEG.fmt instruction execution
1796 mode generally follows that for the NaN encoding,
1797 except where unsupported by hardware.
1799 ignore_loglevel [KNL]
1800 Ignore loglevel setting - this will print /all/
1801 kernel messages to the console. Useful for debugging.
1802 We also add it as printk module parameter, so users
1803 could change it dynamically, usually by
1804 /sys/module/printk/parameters/ignore_loglevel.
1807 Ignore RLIMIT_DATA setting for data mappings,
1808 print warning at first misuse. Can be changed via
1809 /sys/module/kernel/parameters/ignore_rlimit_data.
1811 ihash_entries= [KNL]
1812 Set number of hash buckets for inode cache.
1814 ima_appraise= [IMA] appraise integrity measurements
1815 Format: { "off" | "enforce" | "fix" | "log" }
1818 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1819 The builtin appraise policy appraises all files
1822 ima_canonical_fmt [IMA]
1823 Use the canonical format for the binary runtime
1824 measurements, instead of host native format.
1827 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1831 The list of supported hash algorithms is defined
1832 in crypto/hash_info.h.
1835 The builtin policies to load during IMA setup.
1836 Format: "tcb | appraise_tcb | secure_boot |
1837 fail_securely | critical_data"
1839 The "tcb" policy measures all programs exec'd, files
1840 mmap'd for exec, and all files opened with the read
1841 mode bit set by either the effective uid (euid=0) or
1844 The "appraise_tcb" policy appraises the integrity of
1845 all files owned by root.
1847 The "secure_boot" policy appraises the integrity
1848 of files (eg. kexec kernel image, kernel modules,
1849 firmware, policy, etc) based on file signatures.
1851 The "fail_securely" policy forces file signature
1852 verification failure also on privileged mounted
1853 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1856 The "critical_data" policy measures kernel integrity
1859 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1860 Load a policy which meets the needs of the Trusted
1861 Computing Base. This means IMA will measure all
1862 programs exec'd, files mmap'd for exec, and all files
1863 opened for read by uid=0.
1866 Select one of defined IMA measurements template formats.
1867 Formats: { "ima" | "ima-ng" | "ima-sig" }
1871 [IMA] Define a custom template format.
1872 Format: { "field1|...|fieldN" }
1874 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1875 Format: <min_file_size>
1876 Set the minimal file size for using asynchronous hash.
1877 If left unspecified, ahash usage is disabled.
1879 ahash performance varies for different data sizes on
1880 different crypto accelerators. This option can be used
1881 to achieve the best performance for a particular HW.
1883 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1885 Set hashing buffer size. Default: 4k.
1887 ahash performance varies for different chunk sizes on
1888 different crypto accelerators. This option can be used
1889 to achieve best performance for particular HW.
1893 Run specified binary instead of /sbin/init as init
1896 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1897 for working out where the kernel is dying during
1900 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1901 initcall functions. Useful for debugging built-in
1902 modules and initcalls.
1904 initramfs_async= [KNL]
1907 This parameter controls whether the initramfs
1908 image is unpacked asynchronously, concurrently
1909 with devices being probed and
1910 initialized. This should normally just work,
1911 but as a debugging aid, one can get the
1912 historical behaviour of the initramfs
1913 unpacking being completed before device_ and
1916 initrd= [BOOT] Specify the location of the initial ramdisk
1918 initrdmem= [KNL] Specify a physical address and size from which to
1919 load the initrd. If an initrd is compiled in or
1920 specified in the bootparams, it takes priority over this
1922 Format: ss[KMG],nn[KMG]
1925 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1928 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1930 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1932 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1934 init_pkru= [X86] Specify the default memory protection keys rights
1935 register contents for all processes. 0x55555554 by
1936 default (disallow access to all but pkey 0). Can
1937 override in debugfs after boot.
1939 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1942 int_pln_enable [X86] Enable power limit notification interrupt
1944 integrity_audit=[IMA]
1945 Format: { "0" | "1" }
1946 0 -- basic integrity auditing messages. (Default)
1947 1 -- additional integrity auditing messages.
1949 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1951 Enable intel iommu driver.
1953 Disable intel iommu driver.
1954 igfx_off [Default Off]
1955 By default, gfx is mapped as normal device. If a gfx
1956 device has a dedicated DMAR unit, the DMAR unit is
1957 bypassed by not enabling DMAR with this option. In
1958 this case, gfx device will use physical address for
1960 strict [Default Off]
1961 With this option on every unmap_single operation will
1962 result in a hardware IOTLB flush operation as opposed
1963 to batching them for performance.
1964 sp_off [Default Off]
1965 By default, super page will be supported if Intel IOMMU
1966 has the capability. With this option, super page will
1969 By default, scalable mode will be disabled even if the
1970 hardware advertises that it has support for the scalable
1971 mode translation. With this option set, scalable mode
1972 will be used on hardware which claims to support it.
1973 tboot_noforce [Default Off]
1974 Do not force the Intel IOMMU enabled under tboot.
1975 By default, tboot will force Intel IOMMU on, which
1976 could harm performance of some high-throughput
1977 devices like 40GBit network cards, even if identity
1979 Note that using this option lowers the security
1980 provided by tboot because it makes the system
1981 vulnerable to DMA attacks.
1983 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1984 0 disables intel_idle and fall back on acpi_idle.
1985 1 to 9 specify maximum depth of C-state.
1989 Do not enable intel_pstate as the default
1990 scaling driver for the supported processors
1992 Use intel_pstate as a scaling driver, but configure it
1993 to work with generic cpufreq governors (instead of
1994 enabling its internal governor). This mode cannot be
1995 used along with the hardware-managed P-states (HWP)
1998 Enable intel_pstate on systems that prohibit it by default
1999 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2000 instead of acpi-cpufreq may disable platform features, such
2001 as thermal controls and power capping, that rely on ACPI
2002 P-States information being indicated to OSPM and therefore
2003 should be used with caution. This option does not work with
2004 processors that aren't supported by the intel_pstate driver
2005 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2007 Do not enable hardware P state control (HWP)
2010 Only load intel_pstate on systems which support
2011 hardware P state control (HWP) if available.
2013 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2014 Description Table, specifies preferred power management
2015 profile as "Enterprise Server" or "Performance Server",
2016 then this feature is turned on by default.
2018 Allow per-logical-CPU P-State performance control limits using
2019 cpufreq sysfs interface
2021 intremap= [X86-64, Intel-IOMMU]
2022 on enable Interrupt Remapping (default)
2023 off disable Interrupt Remapping
2024 nosid disable Source ID checking
2026 BIOS x2APIC opt-out request will be ignored
2027 nopost disable Interrupt Posting
2029 iomem= Disable strict checking of access to MMIO memory
2030 strict regions from userspace.
2045 nobypass [PPC/POWERNV]
2046 Disable IOMMU bypass, using IOMMU for PCI devices.
2048 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2049 Format: { "0" | "1" }
2050 0 - Try to allocate a 32-bit DMA address first, before
2051 falling back to the full range if needed.
2052 1 - Allocate directly from the full usable range,
2053 forcing Dual Address Cycle for PCI cards supporting
2054 greater than 32-bit addressing.
2056 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2057 Format: { "0" | "1" }
2059 Request that DMA unmap operations use deferred
2060 invalidation of hardware TLBs, for increased
2061 throughput at the cost of reduced device isolation.
2062 Will fall back to strict mode if not supported by
2063 the relevant IOMMU driver.
2064 1 - Strict mode (default).
2065 DMA unmap operations invalidate IOMMU hardware TLBs
2067 Note: on x86, the default behaviour depends on the
2068 equivalent driver-specific parameters, but a strict
2069 mode explicitly specified by either method takes
2073 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2074 Format: { "0" | "1" }
2075 0 - Use IOMMU translation for DMA.
2076 1 - Bypass the IOMMU for DMA.
2077 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2079 io7= [HW] IO7 for Marvel-based Alpha systems
2080 See comment before marvel_specify_io7 in
2081 arch/alpha/kernel/core_marvel.c.
2083 io_delay= [X86] I/O delay method
2085 Standard port 0x80 based delay
2087 Alternate port 0xed based delay (needed on some systems)
2089 Simple two microseconds delay
2094 See Documentation/admin-guide/nfs/nfsroot.rst.
2096 ipcmni_extend [KNL] Extend the maximum number of unique System V
2097 IPC identifiers from 32,768 to 16,777,216.
2099 irqaffinity= [SMP] Set the default irq affinity mask
2100 The argument is a cpu list, as described above.
2102 irqchip.gicv2_force_probe=
2105 Force the kernel to look for the second 4kB page
2106 of a GICv2 controller even if the memory range
2107 exposed by the device tree is too small.
2109 irqchip.gicv3_nolpi=
2111 Force the kernel to ignore the availability of
2112 LPIs (and by consequence ITSs). Intended for system
2113 that use the kernel as a bootloader, and thus want
2114 to let secondary kernels in charge of setting up
2117 irqchip.gicv3_pseudo_nmi= [ARM64]
2118 Enables support for pseudo-NMIs in the kernel. This
2119 requires the kernel to be built with
2120 CONFIG_ARM64_PSEUDO_NMI.
2123 When an interrupt is not handled search all handlers
2124 for it. Intended to get systems with badly broken
2128 When an interrupt is not handled search all handlers
2129 for it. Also check all handlers each timer
2130 interrupt. Intended to get systems with badly broken
2134 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2136 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2137 [Deprecated - use cpusets instead]
2138 Format: [flag-list,]<cpu-list>
2140 Specify one or more CPUs to isolate from disturbances
2141 specified in the flag list (default: domain):
2144 Disable the tick when a single task runs.
2146 A residual 1Hz tick is offloaded to workqueues, which you
2147 need to affine to housekeeping through the global
2148 workqueue's affinity configured via the
2149 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2150 by using the 'domain' flag described below.
2152 NOTE: by default the global workqueue runs on all CPUs,
2153 so to protect individual CPUs the 'cpumask' file has to
2154 be configured manually after bootup.
2157 Isolate from the general SMP balancing and scheduling
2158 algorithms. Note that performing domain isolation this way
2159 is irreversible: it's not possible to bring back a CPU to
2160 the domains once isolated through isolcpus. It's strongly
2161 advised to use cpusets instead to disable scheduler load
2162 balancing through the "cpuset.sched_load_balance" file.
2163 It offers a much more flexible interface where CPUs can
2164 move in and out of an isolated set anytime.
2166 You can move a process onto or off an "isolated" CPU via
2167 the CPU affinity syscalls or cpuset.
2168 <cpu number> begins at 0 and the maximum value is
2169 "number of CPUs in system - 1".
2173 Isolate from being targeted by managed interrupts
2174 which have an interrupt mask containing isolated
2175 CPUs. The affinity of managed interrupts is
2176 handled by the kernel and cannot be changed via
2177 the /proc/irq/* interfaces.
2179 This isolation is best effort and only effective
2180 if the automatically assigned interrupt mask of a
2181 device queue contains isolated and housekeeping
2182 CPUs. If housekeeping CPUs are online then such
2183 interrupts are directed to the housekeeping CPU
2184 so that IO submitted on the housekeeping CPU
2185 cannot disturb the isolated CPU.
2187 If a queue's affinity mask contains only isolated
2188 CPUs then this parameter has no effect on the
2189 interrupt routing decision, though interrupts are
2190 only delivered when tasks running on those
2191 isolated CPUs submit IO. IO submitted on
2192 housekeeping CPUs has no influence on those
2195 The format of <cpu-list> is described above.
2199 ivrs_ioapic [HW,X86-64]
2200 Provide an override to the IOAPIC-ID<->DEVICE-ID
2201 mapping provided in the IVRS ACPI table. For
2202 example, to map IOAPIC-ID decimal 10 to
2203 PCI device 00:14.0 write the parameter as:
2204 ivrs_ioapic[10]=00:14.0
2206 ivrs_hpet [HW,X86-64]
2207 Provide an override to the HPET-ID<->DEVICE-ID
2208 mapping provided in the IVRS ACPI table. For
2209 example, to map HPET-ID decimal 0 to
2210 PCI device 00:14.0 write the parameter as:
2211 ivrs_hpet[0]=00:14.0
2213 ivrs_acpihid [HW,X86-64]
2214 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2215 mapping provided in the IVRS ACPI table. For
2216 example, to map UART-HID:UID AMD0020:0 to
2217 PCI device 00:14.5 write the parameter as:
2218 ivrs_acpihid[00:14.5]=AMD0020:0
2220 js= [HW,JOY] Analog joystick
2221 See Documentation/input/joydev/joystick.rst.
2224 When CONFIG_RANDOMIZE_BASE is set, this disables
2225 kernel and module base offset ASLR (Address Space
2226 Layout Randomization).
2229 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2230 report on every invalid memory access. Without this
2231 parameter KASAN will print report only for the first
2236 kernelcore= [KNL,X86,IA-64,PPC]
2237 Format: nn[KMGTPE] | nn% | "mirror"
2238 This parameter specifies the amount of memory usable by
2239 the kernel for non-movable allocations. The requested
2240 amount is spread evenly throughout all nodes in the
2241 system as ZONE_NORMAL. The remaining memory is used for
2242 movable memory in its own zone, ZONE_MOVABLE. In the
2243 event, a node is too small to have both ZONE_NORMAL and
2244 ZONE_MOVABLE, kernelcore memory will take priority and
2245 other nodes will have a larger ZONE_MOVABLE.
2247 ZONE_MOVABLE is used for the allocation of pages that
2248 may be reclaimed or moved by the page migration
2249 subsystem. Note that allocations like PTEs-from-HighMem
2250 still use the HighMem zone if it exists, and the Normal
2251 zone if it does not.
2253 It is possible to specify the exact amount of memory in
2254 the form of "nn[KMGTPE]", a percentage of total system
2255 memory in the form of "nn%", or "mirror". If "mirror"
2256 option is specified, mirrored (reliable) memory is used
2257 for non-movable allocations and remaining memory is used
2258 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2259 are exclusive, so you cannot specify multiple forms.
2261 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2262 Format: <Controller#>[,poll interval]
2263 The controller # is the number of the ehci usb debug
2264 port as it is probed via PCI. The poll interval is
2265 optional and is the number seconds in between
2266 each poll cycle to the debug port in case you need
2267 the functionality for interrupting the kernel with
2268 gdb or control-c on the dbgp connection. When
2269 not using this parameter you use sysrq-g to break into
2270 the kernel debugger.
2272 kgdboc= [KGDB,HW] kgdb over consoles.
2273 Requires a tty driver that supports console polling,
2274 or a supported polling keyboard driver (non-usb).
2275 Serial only format: <serial_device>[,baud]
2276 keyboard only format: kbd
2277 keyboard and serial format: kbd,<serial_device>[,baud]
2278 Optional Kernel mode setting:
2279 kms, kbd format: kms,kbd
2280 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2282 kgdboc_earlycon= [KGDB,HW]
2283 If the boot console provides the ability to read
2284 characters and can work in polling mode, you can use
2285 this parameter to tell kgdb to use it as a backend
2286 until the normal console is registered. Intended to
2287 be used together with the kgdboc parameter which
2288 specifies the normal console to transition to.
2290 The name of the early console should be specified
2291 as the value of this parameter. Note that the name of
2292 the early console might be different than the tty
2293 name passed to kgdboc. It's OK to leave the value
2294 blank and the first boot console that implements
2295 read() will be picked.
2297 kgdbwait [KGDB] Stop kernel execution and enter the
2298 kernel debugger at the earliest opportunity.
2300 kmac= [MIPS] Korina ethernet MAC address.
2301 Configure the RouterBoard 532 series on-chip
2302 Ethernet adapter MAC address.
2304 kmemleak= [KNL] Boot-time kmemleak enable/disable
2305 Valid arguments: on, off
2307 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2310 kprobe_event=[probe-list]
2311 [FTRACE] Add kprobe events and enable at boot time.
2312 The probe-list is a semicolon delimited list of probe
2313 definitions. Each definition is same as kprobe_events
2314 interface, but the parameters are comma delimited.
2315 For example, to add a kprobe event on vfs_read with
2316 arg1 and arg2, add to the command line;
2318 kprobe_event=p,vfs_read,$arg1,$arg2
2320 See also Documentation/trace/kprobetrace.rst "Kernel
2321 Boot Parameter" section.
2323 kpti= [ARM64] Control page table isolation of user
2324 and kernel address spaces.
2325 Default: enabled on cores which need mitigation.
2329 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2330 Default is 0 (don't ignore, but inject #GP)
2332 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2333 Default is false (don't support).
2335 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2340 [KVM] Controls the software workaround for the
2341 X86_BUG_ITLB_MULTIHIT bug.
2342 force : Always deploy workaround.
2343 off : Never deploy workaround.
2344 auto : Deploy workaround based on the presence of
2345 X86_BUG_ITLB_MULTIHIT.
2349 If the software workaround is enabled for the host,
2350 guests do need not to enable it for nested guests.
2352 kvm.nx_huge_pages_recovery_ratio=
2353 [KVM] Controls how many 4KiB pages are periodically zapped
2354 back to huge pages. 0 disables the recovery, otherwise if
2355 the value is N KVM will zap 1/Nth of the 4KiB pages every
2356 minute. The default is 60.
2358 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2359 Default is 1 (enabled)
2361 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2363 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2366 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2368 nvhe: Standard nVHE-based mode, without support for
2371 protected: nVHE-based mode with support for guests whose
2372 state is kept private from the host.
2373 Not valid if the kernel is running in EL2.
2375 Defaults to VHE/nVHE based on hardware support.
2377 kvm-arm.vgic_v3_group0_trap=
2378 [KVM,ARM] Trap guest accesses to GICv3 group-0
2381 kvm-arm.vgic_v3_group1_trap=
2382 [KVM,ARM] Trap guest accesses to GICv3 group-1
2385 kvm-arm.vgic_v3_common_trap=
2386 [KVM,ARM] Trap guest accesses to GICv3 common
2389 kvm-arm.vgic_v4_enable=
2390 [KVM,ARM] Allow use of GICv4 for direct injection of
2393 kvm_cma_resv_ratio=n [PPC]
2394 Reserves given percentage from system memory area for
2395 contiguous memory allocation for KVM hash pagetable
2397 By default it reserves 5% of total system memory.
2401 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2402 (virtualized MMU) support on capable Intel chips.
2403 Default is 1 (enabled)
2405 kvm-intel.emulate_invalid_guest_state=
2406 [KVM,Intel] Enable emulation of invalid guest states
2407 Default is 0 (disabled)
2409 kvm-intel.flexpriority=
2410 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2411 Default is 1 (enabled)
2414 [KVM,Intel] Enable VMX nesting (nVMX).
2415 Default is 0 (disabled)
2417 kvm-intel.unrestricted_guest=
2418 [KVM,Intel] Disable unrestricted guest feature
2419 (virtualized real and unpaged mode) on capable
2420 Intel chips. Default is 1 (enabled)
2422 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2425 Valid arguments: never, cond, always
2427 always: L1D cache flush on every VMENTER.
2428 cond: Flush L1D on VMENTER only when the code between
2429 VMEXIT and VMENTER can leak host memory.
2430 never: Disables the mitigation
2432 Default is cond (do L1 cache flush in specific instances)
2434 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2435 feature (tagged TLBs) on capable Intel chips.
2436 Default is 1 (enabled)
2438 l1d_flush= [X86,INTEL]
2439 Control mitigation for L1D based snooping vulnerability.
2441 Certain CPUs are vulnerable to an exploit against CPU
2442 internal buffers which can forward information to a
2443 disclosure gadget under certain conditions.
2445 In vulnerable processors, the speculatively
2446 forwarded data can be used in a cache side channel
2447 attack, to access data to which the attacker does
2448 not have direct access.
2450 This parameter controls the mitigation. The
2453 on - enable the interface for the mitigation
2455 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2458 The kernel PTE inversion protection is unconditionally
2459 enabled and cannot be disabled.
2462 Provides all available mitigations for the
2463 L1TF vulnerability. Disables SMT and
2464 enables all mitigations in the
2465 hypervisors, i.e. unconditional L1D flush.
2467 SMT control and L1D flush control via the
2468 sysfs interface is still possible after
2469 boot. Hypervisors will issue a warning
2470 when the first VM is started in a
2471 potentially insecure configuration,
2472 i.e. SMT enabled or L1D flush disabled.
2475 Same as 'full', but disables SMT and L1D
2476 flush runtime control. Implies the
2477 'nosmt=force' command line option.
2478 (i.e. sysfs control of SMT is disabled.)
2481 Leaves SMT enabled and enables the default
2482 hypervisor mitigation, i.e. conditional
2485 SMT control and L1D flush control via the
2486 sysfs interface is still possible after
2487 boot. Hypervisors will issue a warning
2488 when the first VM is started in a
2489 potentially insecure configuration,
2490 i.e. SMT enabled or L1D flush disabled.
2494 Disables SMT and enables the default
2495 hypervisor mitigation.
2497 SMT control and L1D flush control via the
2498 sysfs interface is still possible after
2499 boot. Hypervisors will issue a warning
2500 when the first VM is started in a
2501 potentially insecure configuration,
2502 i.e. SMT enabled or L1D flush disabled.
2505 Same as 'flush', but hypervisors will not
2506 warn when a VM is started in a potentially
2507 insecure configuration.
2510 Disables hypervisor mitigations and doesn't
2512 It also drops the swap size and available
2513 RAM limit restriction on both hypervisor and
2518 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2524 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2527 lapic= [X86,APIC] Do not use TSC deadline
2528 value for LAPIC timer one-shot implementation. Default
2529 back to the programmable timer unit in the LAPIC.
2530 Format: notscdeadline
2532 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2535 libata.dma= [LIBATA] DMA control
2536 libata.dma=0 Disable all PATA and SATA DMA
2537 libata.dma=1 PATA and SATA Disk DMA only
2538 libata.dma=2 ATAPI (CDROM) DMA only
2539 libata.dma=4 Compact Flash DMA only
2540 Combinations also work, so libata.dma=3 enables DMA
2541 for disks and CDROMs, but not CFs.
2543 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2544 libata.ignore_hpa=0 keep BIOS limits (default)
2545 libata.ignore_hpa=1 ignore limits, using full disk
2547 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2551 libata.force= [LIBATA] Force configurations. The format is comma-
2552 separated list of "[ID:]VAL" where ID is
2553 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2554 matching port, link or device. Basically, it matches
2555 the ATA ID string printed on console by libata. If
2556 the whole ID part is omitted, the last PORT and DEVICE
2557 values are used. If ID hasn't been specified yet, the
2558 configuration applies to all ports, links and devices.
2560 If only DEVICE is omitted, the parameter applies to
2561 the port and all links and devices behind it. DEVICE
2562 number of 0 either selects the first device or the
2563 first fan-out link behind PMP device. It does not
2564 select the host link. DEVICE number of 15 selects the
2565 host link and device attached to it.
2567 The VAL specifies the configuration to force. As long
2568 as there's no ambiguity shortcut notation is allowed.
2569 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2570 The following configurations can be forced.
2572 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2573 Any ID with matching PORT is used.
2575 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2577 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2578 udma[/][16,25,33,44,66,100,133] notation is also
2581 * [no]ncq: Turn on or off NCQ.
2583 * [no]ncqtrim: Turn off queued DSM TRIM.
2585 * nohrst, nosrst, norst: suppress hard, soft
2588 * rstonce: only attempt one reset during
2589 hot-unplug link recovery
2591 * dump_id: dump IDENTIFY data.
2593 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2595 * disable: Disable this device.
2597 If there are multiple matching configurations changing
2598 the same attribute, the last one is used.
2600 memblock=debug [KNL] Enable memblock debug messages.
2602 load_ramdisk= [RAM] [Deprecated]
2604 lockd.nlm_grace_period=P [NFS] Assign grace period.
2607 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2610 lockd.nlm_timeout=T [NFS] Assign timeout value.
2613 lockd.nlm_udpport=M [NFS] Assign UDP port.
2616 lockdown= [SECURITY]
2617 { integrity | confidentiality }
2618 Enable the kernel lockdown feature. If set to
2619 integrity, kernel features that allow userland to
2620 modify the running kernel are disabled. If set to
2621 confidentiality, kernel features that allow userland
2622 to extract confidential information from the kernel
2625 locktorture.nreaders_stress= [KNL]
2626 Set the number of locking read-acquisition kthreads.
2627 Defaults to being automatically set based on the
2628 number of online CPUs.
2630 locktorture.nwriters_stress= [KNL]
2631 Set the number of locking write-acquisition kthreads.
2633 locktorture.onoff_holdoff= [KNL]
2634 Set time (s) after boot for CPU-hotplug testing.
2636 locktorture.onoff_interval= [KNL]
2637 Set time (s) between CPU-hotplug operations, or
2638 zero to disable CPU-hotplug testing.
2640 locktorture.shuffle_interval= [KNL]
2641 Set task-shuffle interval (jiffies). Shuffling
2642 tasks allows some CPUs to go into dyntick-idle
2643 mode during the locktorture test.
2645 locktorture.shutdown_secs= [KNL]
2646 Set time (s) after boot system shutdown. This
2647 is useful for hands-off automated testing.
2649 locktorture.stat_interval= [KNL]
2650 Time (s) between statistics printk()s.
2652 locktorture.stutter= [KNL]
2653 Time (s) to stutter testing, for example,
2654 specifying five seconds causes the test to run for
2655 five seconds, wait for five seconds, and so on.
2656 This tests the locking primitive's ability to
2657 transition abruptly to and from idle.
2659 locktorture.torture_type= [KNL]
2660 Specify the locking implementation to test.
2662 locktorture.verbose= [KNL]
2663 Enable additional printk() statements.
2665 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2668 loglevel= All Kernel Messages with a loglevel smaller than the
2669 console loglevel will be printed to the console. It can
2670 also be changed with klogd or other programs. The
2671 loglevels are defined as follows:
2673 0 (KERN_EMERG) system is unusable
2674 1 (KERN_ALERT) action must be taken immediately
2675 2 (KERN_CRIT) critical conditions
2676 3 (KERN_ERR) error conditions
2677 4 (KERN_WARNING) warning conditions
2678 5 (KERN_NOTICE) normal but significant condition
2679 6 (KERN_INFO) informational
2680 7 (KERN_DEBUG) debug-level messages
2682 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2683 in bytes. n must be a power of two and greater
2684 than the minimal size. The minimal size is defined
2685 by LOG_BUF_SHIFT kernel config parameter. There is
2686 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2687 that allows to increase the default size depending on
2688 the number of CPUs. See init/Kconfig for more details.
2690 logo.nologo [FB] Disables display of the built-in Linux logo.
2691 This may be used to provide more screen space for
2692 kernel log messages and is useful when debugging
2693 kernel boot problems.
2695 lp=0 [LP] Specify parallel ports to use, e.g,
2696 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2697 lp=reset first parallel port). 'lp=0' disables the
2698 lp=auto printer driver. 'lp=reset' (which can be
2699 specified in addition to the ports) causes
2700 attached printers to be reset. Using
2701 lp=port1,port2,... specifies the parallel ports
2702 to associate lp devices with, starting with
2703 lp0. A port specification may be 'none' to skip
2704 that lp device, or a parport name such as
2705 'parport0'. Specifying 'lp=auto' instead of a
2706 port specification list means that device IDs
2707 from each port should be examined, to see if
2708 an IEEE 1284-compliant printer is attached; if
2709 so, the driver will manage that printer.
2710 See also header of drivers/char/lp.c.
2713 Sets loops_per_jiffy to given constant, thus avoiding
2714 time-consuming boot-time autodetection (up to 250 ms per
2715 CPU). 0 enables autodetection (default). To determine
2716 the correct value for your kernel, boot with normal
2717 autodetection and see what value is printed. Note that
2718 on SMP systems the preset will be applied to all CPUs,
2719 which is likely to cause problems if your CPUs need
2720 significantly divergent settings. An incorrect value
2721 will cause delays in the kernel to be wrong, leading to
2722 unpredictable I/O errors and other breakage. Although
2723 unlikely, in the extreme case this might damage your
2727 Format: <io>,<irq>,<dma>
2729 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2732 [SECURITY] Choose order of LSM initialization. This
2733 overrides CONFIG_LSM, and the "security=" parameter.
2735 machvec= [IA-64] Force the use of a particular machine-vector
2736 (machvec) in a generic kernel.
2737 Example: machvec=hpzx1
2739 machtype= [Loongson] Share the same kernel image file between
2740 different yeeloong laptops.
2741 Example: machtype=lemote-yeeloong-2f-7inch
2743 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2744 than or equal to this physical address is ignored.
2746 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2747 will bring up during bootup. maxcpus=n : n >= 0 limits
2748 the kernel to bring up 'n' processors. Surely after
2749 bootup you can bring up the other plugged cpu by executing
2750 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2751 only takes effect during system bootup.
2752 While n=0 is a special case, it is equivalent to "nosmp",
2753 which also disables the IO APIC.
2755 max_loop= [LOOP] The number of loop block devices that get
2756 (loop.max_loop) unconditionally pre-created at init time. The default
2757 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2758 of statically allocating a predefined number, loop
2759 devices can be requested on-demand with the
2760 /dev/loop-control interface.
2762 mce [X86-32] Machine Check Exception
2764 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2766 md= [HW] RAID subsystems devices and level
2767 See Documentation/admin-guide/md.rst.
2770 Format: <first>,<last>
2771 Specifies range of consoles to be captured by the MDA.
2774 Control mitigation for the Micro-architectural Data
2775 Sampling (MDS) vulnerability.
2777 Certain CPUs are vulnerable to an exploit against CPU
2778 internal buffers which can forward information to a
2779 disclosure gadget under certain conditions.
2781 In vulnerable processors, the speculatively
2782 forwarded data can be used in a cache side channel
2783 attack, to access data to which the attacker does
2784 not have direct access.
2786 This parameter controls the MDS mitigation. The
2789 full - Enable MDS mitigation on vulnerable CPUs
2790 full,nosmt - Enable MDS mitigation and disable
2791 SMT on vulnerable CPUs
2792 off - Unconditionally disable MDS mitigation
2794 On TAA-affected machines, mds=off can be prevented by
2795 an active TAA mitigation as both vulnerabilities are
2796 mitigated with the same mechanism so in order to disable
2797 this mitigation, you need to specify tsx_async_abort=off
2800 Not specifying this option is equivalent to
2803 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2805 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2806 Amount of memory to be used in cases as follows:
2809 2 when the kernel is not able to see the whole system memory;
2810 3 memory that lies after 'mem=' boundary is excluded from
2811 the hypervisor, then assigned to KVM guests.
2813 [X86] Work as limiting max address. Use together
2814 with memmap= to avoid physical address space collisions.
2815 Without memmap= PCI devices could be placed at addresses
2816 belonging to unused RAM.
2818 Note that this only takes effects during boot time since
2819 in above case 3, memory may need be hot added after boot
2820 if system memory of hypervisor is not sufficient.
2822 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2826 [KNL,SH] Allow user to override the default size for
2827 per-device physically contiguous DMA buffers.
2829 memhp_default_state=online/offline
2830 [KNL] Set the initial state for the memory hotplug
2831 onlining policy. If not specified, the default value is
2832 set according to the
2833 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2835 See Documentation/admin-guide/mm/memory-hotplug.rst.
2837 memmap=exactmap [KNL,X86] Enable setting of an exact
2838 E820 memory map, as specified by the user.
2839 Such memmap=exactmap lines can be constructed based on
2840 BIOS output or other requirements. See the memmap=nn@ss
2843 memmap=nn[KMG]@ss[KMG]
2844 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
2845 Region of memory to be used is from ss to ss+nn.
2846 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2847 which limits max address to nn[KMG].
2848 Multiple different regions can be specified,
2851 memmap=100M@2G,100M#3G,1G!1024G
2853 memmap=nn[KMG]#ss[KMG]
2854 [KNL,ACPI] Mark specific memory as ACPI data.
2855 Region of memory to be marked is from ss to ss+nn.
2857 memmap=nn[KMG]$ss[KMG]
2858 [KNL,ACPI] Mark specific memory as reserved.
2859 Region of memory to be reserved is from ss to ss+nn.
2860 Example: Exclude memory from 0x18690000-0x1869ffff
2861 memmap=64K$0x18690000
2863 memmap=0x10000$0x18690000
2864 Some bootloaders may need an escape character before '$',
2865 like Grub2, otherwise '$' and the following number
2868 memmap=nn[KMG]!ss[KMG]
2869 [KNL,X86] Mark specific memory as protected.
2870 Region of memory to be used, from ss to ss+nn.
2871 The memory region may be marked as e820 type 12 (0xc)
2872 and is NVDIMM or ADR memory.
2874 memmap=<size>%<offset>-<oldtype>+<newtype>
2875 [KNL,ACPI] Convert memory within the specified region
2876 from <oldtype> to <newtype>. If "-<oldtype>" is left
2877 out, the whole region will be marked as <newtype>,
2878 even if previously unavailable. If "+<newtype>" is left
2879 out, matching memory will be removed. Types are
2880 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2881 3 = ACPI, 12 = PRAM.
2883 memory_corruption_check=0/1 [X86]
2884 Some BIOSes seem to corrupt the first 64k of
2885 memory when doing things like suspend/resume.
2886 Setting this option will scan the memory
2887 looking for corruption. Enabling this will
2888 both detect corruption and prevent the kernel
2889 from using the memory being corrupted.
2890 However, its intended as a diagnostic tool; if
2891 repeatable BIOS-originated corruption always
2892 affects the same memory, you can use memmap=
2893 to prevent the kernel from using that memory.
2895 memory_corruption_check_size=size [X86]
2896 By default it checks for corruption in the low
2897 64k, making this memory unavailable for normal
2898 use. Use this parameter to scan for
2899 corruption in more or less memory.
2901 memory_corruption_check_period=seconds [X86]
2902 By default it checks for corruption every 60
2903 seconds. Use this parameter to check at some
2904 other rate. 0 disables periodic checking.
2906 memory_hotplug.memmap_on_memory
2907 [KNL,X86,ARM] Boolean flag to enable this feature.
2908 Format: {on | off (default)}
2909 When enabled, runtime hotplugged memory will
2910 allocate its internal metadata (struct pages)
2911 from the hotadded memory which will allow to
2912 hotadd a lot of memory without requiring
2913 additional memory to do so.
2914 This feature is disabled by default because it
2915 has some implication on large (e.g. GB)
2916 allocations in some configurations (e.g. small
2918 The state of the flag can be read in
2919 /sys/module/memory_hotplug/parameters/memmap_on_memory.
2920 Note that even when enabled, there are a few cases where
2921 the feature is not effective.
2923 This is not compatible with hugetlb_free_vmemmap. If
2924 both parameters are enabled, hugetlb_free_vmemmap takes
2925 precedence over memory_hotplug.memmap_on_memory.
2927 memtest= [KNL,X86,ARM,PPC,RISCV] Enable memtest
2929 default : 0 <disable>
2930 Specifies the number of memtest passes to be
2931 performed. Each pass selects another test
2932 pattern from a given set of patterns. Memtest
2933 fills the memory with this pattern, validates
2934 memory contents and reserves bad memory
2935 regions that are detected.
2937 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2938 Valid arguments: on, off
2939 Default (depends on kernel configuration option):
2940 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2941 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2942 mem_encrypt=on: Activate SME
2943 mem_encrypt=off: Do not activate SME
2945 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2946 for details on when memory encryption can be activated.
2948 mem_sleep_default= [SUSPEND] Default system suspend mode:
2949 s2idle - Suspend-To-Idle
2950 shallow - Power-On Suspend or equivalent (if supported)
2951 deep - Suspend-To-RAM or equivalent (if supported)
2952 See Documentation/admin-guide/pm/sleep-states.rst.
2954 meye.*= [HW] Set MotionEye Camera parameters
2955 See Documentation/admin-guide/media/meye.rst.
2957 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2958 Multi-Function General Purpose Timers on AMD Geode
2961 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2962 the BIOS has incorrectly applied a workaround. TinyBIOS
2963 version 0.98 is known to be affected, 0.99 fixes the
2964 problem by letting the user disable the workaround.
2968 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2969 physical address is ignored.
2971 mini2440= [ARM,HW,KNL]
2972 Format:[0..2][b][c][t]
2974 MINI2440 configuration specification:
2975 0 - The attached screen is the 3.5" TFT
2976 1 - The attached screen is the 7" TFT
2977 2 - The VGA Shield is attached (1024x768)
2978 Leaving out the screen size parameter will not load
2979 the TFT driver, and the framebuffer will be left
2981 b - Enable backlight. The TFT backlight pin will be
2982 linked to the kernel VESA blanking code and a GPIO
2983 LED. This parameter is not necessary when using the
2985 c - Enable the s3c camera interface.
2986 t - Reserved for enabling touchscreen support. The
2987 touchscreen support is not enabled in the mainstream
2988 kernel as of 2.6.30, a preliminary port can be found
2989 in the "bleeding edge" mini2440 support kernel at
2990 https://repo.or.cz/w/linux-2.6/mini2440.git
2993 [X86,PPC,S390,ARM64] Control optional mitigations for
2994 CPU vulnerabilities. This is a set of curated,
2995 arch-independent options, each of which is an
2996 aggregation of existing arch-specific options.
2999 Disable all optional CPU mitigations. This
3000 improves system performance, but it may also
3001 expose users to several CPU vulnerabilities.
3002 Equivalent to: nopti [X86,PPC]
3004 nospectre_v1 [X86,PPC]
3006 nospectre_v2 [X86,PPC,S390,ARM64]
3007 spectre_v2_user=off [X86]
3008 spec_store_bypass_disable=off [X86,PPC]
3009 ssbd=force-off [ARM64]
3012 tsx_async_abort=off [X86]
3013 kvm.nx_huge_pages=off [X86]
3014 no_entry_flush [PPC]
3015 no_uaccess_flush [PPC]
3018 This does not have any effect on
3019 kvm.nx_huge_pages when
3020 kvm.nx_huge_pages=force.
3023 Mitigate all CPU vulnerabilities, but leave SMT
3024 enabled, even if it's vulnerable. This is for
3025 users who don't want to be surprised by SMT
3026 getting disabled across kernel upgrades, or who
3027 have other ways of avoiding SMT-based attacks.
3028 Equivalent to: (default behavior)
3031 Mitigate all CPU vulnerabilities, disabling SMT
3032 if needed. This is for users who always want to
3033 be fully mitigated, even if it means losing SMT.
3034 Equivalent to: l1tf=flush,nosmt [X86]
3035 mds=full,nosmt [X86]
3036 tsx_async_abort=full,nosmt [X86]
3039 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3040 parameter allows control of the logging verbosity for
3041 the additional memory initialisation checks. A value
3042 of 0 disables mminit logging and a level of 4 will
3043 log everything. Information is printed at KERN_DEBUG
3044 so loglevel=8 may also need to be specified.
3047 [KNL] When CONFIG_MODULE_SIG is set, this means that
3048 modules without (valid) signatures will fail to load.
3049 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3050 is always true, so this option does nothing.
3052 module_blacklist= [KNL] Do not load a comma-separated list of
3053 modules. Useful for debugging problem modules.
3056 [MOUSE] Maximum time between finger touching and
3057 leaving touchpad surface for touch to be considered
3058 a tap and be reported as a left button click (for
3059 touchpads working in absolute mode only).
3061 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3062 reporting absolute coordinates, such as tablets
3063 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3064 reporting absolute coordinates, such as tablets
3066 movablecore= [KNL,X86,IA-64,PPC]
3067 Format: nn[KMGTPE] | nn%
3068 This parameter is the complement to kernelcore=, it
3069 specifies the amount of memory used for migratable
3070 allocations. If both kernelcore and movablecore is
3071 specified, then kernelcore will be at *least* the
3072 specified value but may be more. If movablecore on its
3073 own is specified, the administrator must be careful
3074 that the amount of memory usable for all allocations
3077 movable_node [KNL] Boot-time switch to make hotplugable memory
3078 NUMA nodes to be movable. This means that the memory
3079 of such nodes will be usable only for movable
3080 allocations which rules out almost all kernel
3081 allocations. Use with caution!
3083 MTD_Partition= [MTD]
3084 Format: <name>,<region-number>,<size>,<offset>
3086 MTD_Region= [MTD] Format:
3087 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3090 See drivers/mtd/parsers/cmdlinepart.c
3092 multitce=off [PPC] This parameter disables the use of the pSeries
3093 firmware feature for updating multiple TCE entries
3096 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3098 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3100 boundary - index of last SLC block on Flex-OneNAND.
3101 The remaining blocks are configured as MLC blocks.
3102 lock - Configure if Flex-OneNAND boundary should be locked.
3103 Once locked, the boundary cannot be changed.
3104 1 indicates lock status, 0 indicates unlock status.
3107 ARM/S3C2412 JIVE boot control
3109 See arch/arm/mach-s3c/mach-jive.c
3111 mtouchusb.raw_coordinates=
3112 [HW] Make the MicroTouch USB driver use raw coordinates
3113 ('y', default) or cooked coordinates ('n')
3115 mtrr_chunk_size=nn[KMG] [X86]
3116 used for mtrr cleanup. It is largest continuous chunk
3117 that could hold holes aka. UC entries.
3119 mtrr_gran_size=nn[KMG] [X86]
3120 Used for mtrr cleanup. It is granularity of mtrr block.
3122 Large value could prevent small alignment from
3125 mtrr_spare_reg_nr=n [X86]
3127 Range: 0,7 : spare reg number
3129 Used for mtrr cleanup. It is spare mtrr entries number.
3130 Set to 2 or more if your graphical card needs more.
3132 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3134 netdev= [NET] Network devices parameters
3135 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3136 Note that mem_start is often overloaded to mean
3137 something different and driver-specific.
3138 This usage is only documented in each driver source
3142 [NETFILTER] Enable connection tracking flow accounting
3143 0 to disable accounting
3144 1 to enable accounting
3147 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3148 See Documentation/admin-guide/nfs/nfsroot.rst.
3150 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3151 See Documentation/admin-guide/nfs/nfsroot.rst.
3153 nfsrootdebug [NFS] enable nfsroot debugging messages.
3154 See Documentation/admin-guide/nfs/nfsroot.rst.
3156 nfs.callback_nr_threads=
3157 [NFSv4] set the total number of threads that the
3158 NFS client will assign to service NFSv4 callback
3161 nfs.callback_tcpport=
3162 [NFS] set the TCP port on which the NFSv4 callback
3163 channel should listen.
3166 [NFS] sets the pathname to the program which is used
3167 to update the NFS client cache entries.
3169 nfs.cache_getent_timeout=
3170 [NFS] sets the timeout after which an attempt to
3171 update a cache entry is deemed to have failed.
3173 nfs.idmap_cache_timeout=
3174 [NFS] set the maximum lifetime for idmapper cache
3178 [NFS] enable 64-bit inode numbers.
3179 If zero, the NFS client will fake up a 32-bit inode
3180 number for the readdir() and stat() syscalls instead
3181 of returning the full 64-bit number.
3182 The default is to return 64-bit inode numbers.
3184 nfs.max_session_cb_slots=
3185 [NFSv4.1] Sets the maximum number of session
3186 slots the client will assign to the callback
3187 channel. This determines the maximum number of
3188 callbacks the client will process in parallel for
3189 a particular server.
3191 nfs.max_session_slots=
3192 [NFSv4.1] Sets the maximum number of session slots
3193 the client will attempt to negotiate with the server.
3194 This limits the number of simultaneous RPC requests
3195 that the client can send to the NFSv4.1 server.
3196 Note that there is little point in setting this
3197 value higher than the max_tcp_slot_table_limit.
3199 nfs.nfs4_disable_idmapping=
3200 [NFSv4] When set to the default of '1', this option
3201 ensures that both the RPC level authentication
3202 scheme and the NFS level operations agree to use
3203 numeric uids/gids if the mount is using the
3204 'sec=sys' security flavour. In effect it is
3205 disabling idmapping, which can make migration from
3206 legacy NFSv2/v3 systems to NFSv4 easier.
3207 Servers that do not support this mode of operation
3208 will be autodetected by the client, and it will fall
3209 back to using the idmapper.
3210 To turn off this behaviour, set the value to '0'.
3212 [NFS4] Specify an additional fixed unique ident-
3213 ification string that NFSv4 clients can insert into
3214 their nfs_client_id4 string. This is typically a
3215 UUID that is generated at system install time.
3217 nfs.send_implementation_id =
3218 [NFSv4.1] Send client implementation identification
3219 information in exchange_id requests.
3220 If zero, no implementation identification information
3222 The default is to send the implementation identification
3225 nfs.recover_lost_locks =
3226 [NFSv4] Attempt to recover locks that were lost due
3227 to a lease timeout on the server. Please note that
3228 doing this risks data corruption, since there are
3229 no guarantees that the file will remain unchanged
3230 after the locks are lost.
3231 If you want to enable the kernel legacy behaviour of
3232 attempting to recover these locks, then set this
3234 The default parameter value of '0' causes the kernel
3235 not to attempt recovery of lost locks.
3237 nfs4.layoutstats_timer =
3238 [NFSv4.2] Change the rate at which the kernel sends
3239 layoutstats to the pNFS metadata server.
3241 Setting this to value to 0 causes the kernel to use
3242 whatever value is the default set by the layout
3243 driver. A non-zero value sets the minimum interval
3244 in seconds between layoutstats transmissions.
3246 nfsd.nfs4_disable_idmapping=
3247 [NFSv4] When set to the default of '1', the NFSv4
3248 server will return only numeric uids and gids to
3249 clients using auth_sys, and will accept numeric uids
3250 and gids from such clients. This is intended to ease
3251 migration from NFSv2/v3.
3253 nmi_backtrace.backtrace_idle [KNL]
3254 Dump stacks even of idle CPUs in response to an
3255 NMI stack-backtrace request.
3257 nmi_debug= [KNL,SH] Specify one or more actions to take
3258 when a NMI is triggered.
3259 Format: [state][,regs][,debounce][,die]
3261 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3262 Format: [panic,][nopanic,][num]
3264 0 - turn hardlockup detector in nmi_watchdog off
3265 1 - turn hardlockup detector in nmi_watchdog on
3266 When panic is specified, panic when an NMI watchdog
3267 timeout occurs (or 'nopanic' to not panic on an NMI
3268 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3269 To disable both hard and soft lockup detectors,
3270 please see 'nowatchdog'.
3271 This is useful when you use a panic=... timeout and
3272 need the box quickly up again.
3274 These settings can be accessed at runtime via
3275 the nmi_watchdog and hardlockup_panic sysctls.
3277 netpoll.carrier_timeout=
3278 [NET] Specifies amount of time (in seconds) that
3279 netpoll should wait for a carrier. By default netpoll
3282 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3283 emulation library even if a 387 maths coprocessor
3286 no5lvl [X86-64] Disable 5-level paging mode. Forces
3287 kernel to use 4-level paging instead.
3289 nofsgsbase [X86] Disables FSGSBASE instructions.
3292 [HW] Never suspend the console
3293 Disable suspending of consoles during suspend and
3294 hibernate operations. Once disabled, debugging
3295 messages can reach various consoles while the rest
3296 of the system is being put to sleep (ie, while
3297 debugging driver suspend/resume hooks). This may
3298 not work reliably with all consoles, but is known
3299 to work with serial and VGA consoles.
3300 To facilitate more flexible debugging, we also add
3301 console_suspend, a printk module parameter to control
3302 it. Users could use console_suspend (usually
3303 /sys/module/printk/parameters/console_suspend) to
3304 turn on/off it dynamically.
3306 novmcoredd [KNL,KDUMP]
3307 Disable device dump. Device dump allows drivers to
3308 append dump data to vmcore so you can collect driver
3309 specified debug info. Drivers can append the data
3310 without any limit and this data is stored in memory,
3311 so this may cause significant memory stress. Disabling
3312 device dump can help save memory but the driver debug
3313 data will be no longer available. This parameter
3314 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3317 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3318 caches in the slab allocator. Saves per-node memory,
3319 but will impact performance.
3323 noaltinstr [S390] Disables alternative instructions patching
3324 (CPU alternatives feature).
3326 noapic [SMP,APIC] Tells the kernel to not make use of any
3327 IOAPICs that may be present in the system.
3329 noautogroup Disable scheduler automatic task group creation.
3331 nobats [PPC] Do not use BATs for mapping kernel lowmem
3332 on "Classic" PPC cores.
3336 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3338 delayacct [KNL] Enable per-task delay accounting
3340 nodsp [SH] Disable hardware DSP at boot time.
3342 noefi Disable EFI runtime services support.
3344 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3349 On X86-32 available only on PAE configured kernels.
3350 noexec=on: enable non-executable mappings (default)
3351 noexec=off: disable non-executable mappings
3354 Disable SMAP (Supervisor Mode Access Prevention)
3355 even if it is supported by processor.
3358 Disable SMEP (Supervisor Mode Execution Prevention)
3359 even if it is supported by processor.
3362 This affects only 32-bit executables.
3363 noexec32=on: enable non-executable mappings (default)
3364 read doesn't imply executable mappings
3365 noexec32=off: disable non-executable mappings
3366 read implies executable mappings
3368 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3370 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3371 register save and restore. The kernel will only save
3372 legacy floating-point registers on task switch.
3374 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3376 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3378 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3379 Equivalent to smt=1.
3381 [KNL,X86] Disable symmetric multithreading (SMT).
3382 nosmt=force: Force disable SMT, cannot be undone
3383 via the sysfs control file.
3385 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3386 (bounds check bypass). With this option data leaks are
3387 possible in the system.
3389 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3390 the Spectre variant 2 (indirect branch prediction)
3391 vulnerability. System may allow data leaks with this
3394 nospec_store_bypass_disable
3395 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3398 [PPC] Don't flush the L1-D cache after accessing user data.
3400 noxsave [BUGS=X86] Disables x86 extended register state save
3401 and restore using xsave. The kernel will fallback to
3402 enabling legacy floating-point and sse state.
3404 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3405 register states. The kernel will fall back to use
3406 xsave to save the states. By using this parameter,
3407 performance of saving the states is degraded because
3408 xsave doesn't support modified optimization while
3409 xsaveopt supports it on xsaveopt enabled systems.
3411 noxsaves [X86] Disables xsaves and xrstors used in saving and
3412 restoring x86 extended register state in compacted
3413 form of xsave area. The kernel will fall back to use
3414 xsaveopt and xrstor to save and restore the states
3415 in standard form of xsave area. By using this
3416 parameter, xsave area per process might occupy more
3417 memory on xsaves enabled systems.
3419 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3420 in do_idle() and not use the arch_cpu_idle()
3421 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3422 to be effective. This is useful on platforms where the
3423 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3424 correctly or when doing power measurements to evalute
3425 the impact of the sleep instructions. This is also
3426 useful when using JTAG debugger.
3428 no_file_caps Tells the kernel not to honor file capabilities. The
3429 only way then for a file to be executed with privilege
3430 is to be setuid root or executed by root.
3432 nohalt [IA-64] Tells the kernel not to use the power saving
3433 function PAL_HALT_LIGHT when idle. This increases
3434 power-consumption. On the positive side, it reduces
3435 interrupt wake-up latency, which may improve performance
3436 in certain environments such as networked servers or
3440 Force pointers printed to the console or buffers to be
3441 unhashed. By default, when a pointer is printed via %p
3442 format string, that pointer is "hashed", i.e. obscured
3443 by hashing the pointer value. This is a security feature
3444 that hides actual kernel addresses from unprivileged
3445 users, but it also makes debugging the kernel more
3446 difficult since unequal pointers can no longer be
3447 compared. However, if this command-line option is
3448 specified, then all normal pointers will have their true
3449 value printed. Pointers printed via %pK may still be
3450 hashed. This option should only be specified when
3451 debugging the kernel. Please do not use on production
3454 nohibernate [HIBERNATION] Disable hibernation and resume.
3456 nohz= [KNL] Boottime enable/disable dynamic ticks
3457 Valid arguments: on, off
3460 nohz_full= [KNL,BOOT,SMP,ISOL]
3461 The argument is a cpu list, as described above.
3462 In kernels built with CONFIG_NO_HZ_FULL=y, set
3463 the specified list of CPUs whose tick will be stopped
3464 whenever possible. The boot CPU will be forced outside
3465 the range to maintain the timekeeping. Any CPUs
3466 in this list will have their RCU callbacks offloaded,
3467 just as if they had also been called out in the
3468 rcu_nocbs= boot parameter.
3470 noiotrap [SH] Disables trapped I/O port accesses.
3472 noirqdebug [X86-32] Disables the code which attempts to detect and
3473 disable unhandled interrupt sources.
3475 no_timer_check [X86,APIC] Disables the code which tests for
3476 broken timer IRQ sources.
3478 noisapnp [ISAPNP] Disables ISA PnP code.
3480 noinitrd [RAM] Tells the kernel not to load any configured
3483 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3485 [Deprecated - use intremap=off]
3489 noinvpcid [X86] Disable the INVPCID cpu feature.
3491 nojitter [IA-64] Disables jitter checking for ITC timers.
3493 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3495 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3499 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3500 clock and use the default one.
3502 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3503 accounting. steal time is computed, but won't
3504 influence scheduler behaviour
3506 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3508 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3510 noltlbs [PPC] Do not use large page/tlb entries for kernel
3511 lowmem mapping on PPC40x and PPC8xx
3513 nomca [IA-64] Disable machine check abort handling
3515 nomce [X86-32] Disable Machine Check Exception
3517 nomfgpt [X86-32] Disable Multi-Function General Purpose
3518 Timer usage (for AMD Geode machines).
3520 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3521 shutdown the other cpus. Instead use the REBOOT_VECTOR
3524 nomodule Disable module load
3526 nopat [X86] Disable PAT (page attribute table extension of
3527 pagetables) support.
3529 nopcid [X86-64] Disable the PCID cpu feature.
3531 norandmaps Don't use address space randomization. Equivalent to
3532 echo 0 > /proc/sys/kernel/randomize_va_space
3534 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3535 with UP alternatives
3537 nordrand [X86] Disable kernel use of the RDRAND and
3538 RDSEED instructions even if they are supported
3539 by the processor. RDRAND and RDSEED are still
3540 available to user space applications.
3542 noresume [SWSUSP] Disables resume and restores original swap
3545 no-scroll [VGA] Disables scrollback.
3546 This is required for the Braillex ib80-piezo Braille
3547 reader made by F.H. Papenmeier (Germany).
3551 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3553 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3555 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3556 and disable the IO APIC. legacy for "maxcpus=0".
3558 nosoftlockup [KNL] Disable the soft-lockup detector.
3560 nosync [HW,M68K] Disables sync negotiation for all devices.
3562 nowatchdog [KNL] Disable both lockup detectors, i.e.
3563 soft-lockup and NMI watchdog (hard-lockup).
3567 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3569 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3570 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3571 Some features depend on CPU0. Known dependencies are:
3572 1. Resume from suspend/hibernate depends on CPU0.
3573 Suspend/hibernate will fail if CPU0 is offline and you
3574 need to online CPU0 before suspend/hibernate.
3575 2. PIC interrupts also depend on CPU0. CPU0 can't be
3576 removed if a PIC interrupt is detected.
3577 It's said poweroff/reboot may depend on CPU0 on some
3578 machines although I haven't seen such issues so far
3579 after CPU0 is offline on a few tested machines.
3580 If the dependencies are under your control, you can
3581 turn on cpu0_hotplug.
3583 nps_mtm_hs_ctr= [KNL,ARC]
3584 This parameter sets the maximum duration, in
3585 cycles, each HW thread of the CTOP can run
3586 without interruptions, before HW switches it.
3587 The actual maximum duration is 16 times this
3589 Format: integer between 1 and 255
3592 nptcg= [IA-64] Override max number of concurrent global TLB
3593 purges which is reported from either PAL_VM_SUMMARY or
3596 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3597 could support. nr_cpus=n : n >= 1 limits the kernel to
3598 support 'n' processors. It could be larger than the
3599 number of already plugged CPU during bootup, later in
3600 runtime you can physically add extra cpu until it reaches
3601 n. So during boot up some boot time memory for per-cpu
3602 variables need be pre-allocated for later physical cpu
3605 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3607 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3608 set up a single NUMA node spanning all memory.
3610 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3612 Allowed values are enable and disable
3614 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3615 'node', 'default' can be specified
3616 This can be set from sysctl after boot.
3617 See Documentation/admin-guide/sysctl/vm.rst for details.
3619 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3620 See Documentation/core-api/debugging-via-ohci1394.rst for more
3623 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3624 Rather than timing out after 20 ms if an EC
3625 command is not properly ACKed, override the length
3626 of the timeout. We have interrupts disabled while
3627 waiting for the ACK, so if this is set too high
3628 interrupts *may* be lost!
3630 omap_mux= [OMAP] Override bootloader pin multiplexing.
3631 Format: <mux_mode0.mode_name=value>...
3632 For example, to override I2C bus2:
3633 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3635 oops=panic Always panic on oopses. Default is to just kill the
3636 process, but there is a small probability of
3637 deadlocking the machine.
3638 This will also cause panics on machine check exceptions.
3639 Useful together with panic=30 to trigger a reboot.
3642 [KNL] Boolean flag to control whether the page allocator
3643 should randomize its free lists. The randomization may
3644 be automatically enabled if the kernel detects it is
3645 running on a platform with a direct-mapped memory-side
3646 cache, and this parameter can be used to
3647 override/disable that behavior. The state of the flag
3648 can be read from sysfs at:
3649 /sys/module/page_alloc/parameters/shuffle.
3651 page_owner= [KNL] Boot-time page_owner enabling option.
3652 Storage of the information about who allocated
3653 each page is disabled in default. With this switch,
3655 on: enable the feature
3657 page_poison= [KNL] Boot-time parameter changing the state of
3658 poisoning on the buddy allocator, available with
3659 CONFIG_PAGE_POISONING=y.
3660 off: turn off poisoning (default)
3661 on: turn on poisoning
3663 page_reporting.page_reporting_order=
3664 [KNL] Minimal page reporting order
3666 Adjust the minimal page reporting order. The page
3667 reporting is disabled when it exceeds (MAX_ORDER-1).
3669 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3670 timeout > 0: seconds before rebooting
3671 timeout = 0: wait forever
3672 timeout < 0: reboot immediately
3675 panic_print= Bitmask for printing system info when panic happens.
3676 User can chose combination of the following bits:
3677 bit 0: print all tasks info
3678 bit 1: print system memory info
3679 bit 2: print timer info
3680 bit 3: print locks info if CONFIG_LOCKDEP is on
3681 bit 4: print ftrace buffer
3682 bit 5: print all printk messages in buffer
3684 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3685 Format: <hex>[,nousertaint]
3686 Hexadecimal bitmask representing the set of TAINT flags
3687 that will cause the kernel to panic when add_taint() is
3688 called with any of the flags in this set.
3689 The optional switch "nousertaint" can be utilized to
3690 prevent userspace forced crashes by writing to sysctl
3691 /proc/sys/kernel/tainted any flagset matching with the
3692 bitmask set on panic_on_taint.
3693 See Documentation/admin-guide/tainted-kernels.rst for
3694 extra details on the taint flags that users can pick
3695 to compose the bitmask to assign to panic_on_taint.
3697 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3700 crash_kexec_post_notifiers
3701 Run kdump after running panic-notifiers and dumping
3702 kmsg. This only for the users who doubt kdump always
3703 succeeds in any situation.
3704 Note that this also increases risks of kdump failure,
3705 because some panic notifiers can make the crashed
3706 kernel more unstable.
3708 parkbd.port= [HW] Parallel port number the keyboard adapter is
3709 connected to, default is 0.
3711 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3712 0 for XT, 1 for AT (default is AT).
3715 parport= [HW,PPT] Specify parallel ports. 0 disables.
3716 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3717 Use 'auto' to force the driver to use any
3718 IRQ/DMA settings detected (the default is to
3719 ignore detected IRQ/DMA settings because of
3720 possible conflicts). You can specify the base
3721 address, IRQ, and DMA settings; IRQ and DMA
3722 should be numbers, or 'auto' (for using detected
3723 settings on that particular port), or 'nofifo'
3724 (to avoid using a FIFO even if it is detected).
3725 Parallel ports are assigned in the order they
3726 are specified on the command line, starting
3729 parport_init_mode= [HW,PPT]
3730 Configure VIA parallel port to operate in
3731 a specific mode. This is necessary on Pegasos
3732 computer where firmware has no options for setting
3733 up parallel port mode and sets it to spp.
3734 Currently this function knows 686a and 8231 chips.
3735 Format: [spp|ps2|epp|ecp|ecpepp]
3737 pata_legacy.all= [HW,LIBATA]
3739 Set to non-zero to probe primary and secondary ISA
3740 port ranges on PCI systems where no PCI PATA device
3741 has been found at either range. Disabled by default.
3743 pata_legacy.autospeed= [HW,LIBATA]
3745 Set to non-zero if a chip is present that snoops speed
3746 changes. Disabled by default.
3748 pata_legacy.ht6560a= [HW,LIBATA]
3750 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3751 the secondary channel, or both channels respectively.
3752 Disabled by default.
3754 pata_legacy.ht6560b= [HW,LIBATA]
3756 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3757 the secondary channel, or both channels respectively.
3758 Disabled by default.
3760 pata_legacy.iordy_mask= [HW,LIBATA]
3762 IORDY enable mask. Set individual bits to allow IORDY
3763 for the respective channel. Bit 0 is for the first
3764 legacy channel handled by this driver, bit 1 is for
3765 the second channel, and so on. The sequence will often
3766 correspond to the primary legacy channel, the secondary
3767 legacy channel, and so on, but the handling of a PCI
3768 bus and the use of other driver options may interfere
3769 with the sequence. By default IORDY is allowed across
3772 pata_legacy.opti82c46x= [HW,LIBATA]
3774 Set to 1, 2, or 3 for Opti 82c611A on the primary
3775 channel, the secondary channel, or both channels
3776 respectively. Disabled by default.
3778 pata_legacy.opti82c611a= [HW,LIBATA]
3780 Set to 1, 2, or 3 for Opti 82c465MV on the primary
3781 channel, the secondary channel, or both channels
3782 respectively. Disabled by default.
3784 pata_legacy.pio_mask= [HW,LIBATA]
3786 PIO mode mask for autospeed devices. Set individual
3787 bits to allow the use of the respective PIO modes.
3788 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
3789 All modes allowed by default.
3791 pata_legacy.probe_all= [HW,LIBATA]
3793 Set to non-zero to probe tertiary and further ISA
3794 port ranges on PCI systems. Disabled by default.
3796 pata_legacy.probe_mask= [HW,LIBATA]
3798 Probe mask for legacy ISA PATA ports. Depending on
3799 platform configuration and the use of other driver
3800 options up to 6 legacy ports are supported: 0x1f0,
3801 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
3802 of individual ports can be disabled by setting the
3803 corresponding bits in the mask to 1. Bit 0 is for
3804 the first port in the list above (0x1f0), and so on.
3805 By default all supported ports are probed.
3807 pata_legacy.qdi= [HW,LIBATA]
3809 Set to non-zero to probe QDI controllers. By default
3810 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
3812 pata_legacy.winbond= [HW,LIBATA]
3814 Set to non-zero to probe Winbond controllers. Use
3815 the standard I/O port (0x130) if 1, otherwise the
3816 value given is the I/O port to use (typically 0x1b0).
3817 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
3820 pata_platform.pio_mask= [HW,LIBATA]
3822 Supported PIO mode mask. Set individual bits to allow
3823 the use of the respective PIO modes. Bit 0 is for
3824 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
3828 Halt all CPUs after the first oops has been printed for
3829 the specified number of seconds. This is to be used if
3830 your oopses keep scrolling off the screen.
3835 See header of drivers/block/paride/pcd.c.
3836 See also Documentation/admin-guide/blockdev/paride.rst.
3838 pci=option[,option...] [PCI] various PCI subsystem options.
3840 Some options herein operate on a specific device
3841 or a set of devices (<pci_dev>). These are
3842 specified in one of the following formats:
3844 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3845 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3847 Note: the first format specifies a PCI
3848 bus/device/function address which may change
3849 if new hardware is inserted, if motherboard
3850 firmware changes, or due to changes caused
3851 by other kernel parameters. If the
3852 domain is left unspecified, it is
3853 taken to be zero. Optionally, a path
3854 to a device through multiple device/function
3855 addresses can be specified after the base
3856 address (this is more robust against
3857 renumbering issues). The second format
3858 selects devices using IDs from the
3859 configuration space which may match multiple
3860 devices in the system.
3862 earlydump dump PCI config space before the kernel
3864 off [X86] don't probe for the PCI bus
3865 bios [X86-32] force use of PCI BIOS, don't access
3866 the hardware directly. Use this if your machine
3867 has a non-standard PCI host bridge.
3868 nobios [X86-32] disallow use of PCI BIOS, only direct
3869 hardware access methods are allowed. Use this
3870 if you experience crashes upon bootup and you
3871 suspect they are caused by the BIOS.
3872 conf1 [X86] Force use of PCI Configuration Access
3873 Mechanism 1 (config address in IO port 0xCF8,
3874 data in IO port 0xCFC, both 32-bit).
3875 conf2 [X86] Force use of PCI Configuration Access
3876 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3877 the function, IO port 0xCFA, also 8-bit, sets
3878 bus number. The config space is then accessed
3879 through ports 0xC000-0xCFFF).
3880 See http://wiki.osdev.org/PCI for more info
3881 on the configuration access mechanisms.
3882 noaer [PCIE] If the PCIEAER kernel config parameter is
3883 enabled, this kernel boot option can be used to
3884 disable the use of PCIE advanced error reporting.
3885 nodomains [PCI] Disable support for multiple PCI
3886 root domains (aka PCI segments, in ACPI-speak).
3887 nommconf [X86] Disable use of MMCONFIG for PCI
3889 check_enable_amd_mmconf [X86] check for and enable
3890 properly configured MMIO access to PCI
3891 config space on AMD family 10h CPU
3892 nomsi [MSI] If the PCI_MSI kernel config parameter is
3893 enabled, this kernel boot option can be used to
3894 disable the use of MSI interrupts system-wide.
3895 noioapicquirk [APIC] Disable all boot interrupt quirks.
3896 Safety option to keep boot IRQs enabled. This
3897 should never be necessary.
3898 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3899 primary IO-APIC for bridges that cannot disable
3900 boot IRQs. This fixes a source of spurious IRQs
3901 when the system masks IRQs.
3902 noioapicreroute [APIC] Disable workaround that uses the
3903 boot IRQ equivalent of an IRQ that connects to
3904 a chipset where boot IRQs cannot be disabled.
3905 The opposite of ioapicreroute.
3906 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3907 routing table. These calls are known to be buggy
3908 on several machines and they hang the machine
3909 when used, but on other computers it's the only
3910 way to get the interrupt routing table. Try
3911 this option if the kernel is unable to allocate
3912 IRQs or discover secondary PCI buses on your
3914 rom [X86] Assign address space to expansion ROMs.
3915 Use with caution as certain devices share
3916 address decoders between ROMs and other
3918 norom [X86] Do not assign address space to
3919 expansion ROMs that do not already have
3920 BIOS assigned address ranges.
3921 nobar [X86] Do not assign address space to the
3922 BARs that weren't assigned by the BIOS.
3923 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3924 assigned automatically to PCI devices. You can
3925 make the kernel exclude IRQs of your ISA cards
3927 pirqaddr=0xAAAAA [X86] Specify the physical address
3928 of the PIRQ table (normally generated
3929 by the BIOS) if it is outside the
3930 F0000h-100000h range.
3931 lastbus=N [X86] Scan all buses thru bus #N. Can be
3932 useful if the kernel is unable to find your
3933 secondary buses and you want to tell it
3934 explicitly which ones they are.
3935 assign-busses [X86] Always assign all PCI bus
3936 numbers ourselves, overriding
3937 whatever the firmware may have done.
3938 usepirqmask [X86] Honor the possible IRQ mask stored
3939 in the BIOS $PIR table. This is needed on
3940 some systems with broken BIOSes, notably
3941 some HP Pavilion N5400 and Omnibook XE3
3942 notebooks. This will have no effect if ACPI
3943 IRQ routing is enabled.
3944 noacpi [X86] Do not use ACPI for IRQ routing
3945 or for PCI scanning.
3946 use_crs [X86] Use PCI host bridge window information
3947 from ACPI. On BIOSes from 2008 or later, this
3948 is enabled by default. If you need to use this,
3949 please report a bug.
3950 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3951 If you need to use this, please report a bug.
3952 routeirq Do IRQ routing for all PCI devices.
3953 This is normally done in pci_enable_device(),
3954 so this option is a temporary workaround
3955 for broken drivers that don't call it.
3956 skip_isa_align [X86] do not align io start addr, so can
3957 handle more pci cards
3958 noearly [X86] Don't do any early type 1 scanning.
3959 This might help on some broken boards which
3960 machine check when some devices' config space
3961 is read. But various workarounds are disabled
3962 and some IOMMU drivers will not work.
3963 bfsort Sort PCI devices into breadth-first order.
3964 This sorting is done to get a device
3965 order compatible with older (<= 2.4) kernels.
3966 nobfsort Don't sort PCI devices into breadth-first order.
3967 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3968 tuning and use the BIOS-configured MPS defaults.
3969 pcie_bus_safe Set every device's MPS to the largest value
3970 supported by all devices below the root complex.
3971 pcie_bus_perf Set device MPS to the largest allowable MPS
3972 based on its parent bus. Also set MRRS (Max
3973 Read Request Size) to the largest supported
3974 value (no larger than the MPS that the device
3975 or bus can support) for best performance.
3976 pcie_bus_peer2peer Set every device's MPS to 128B, which
3977 every device is guaranteed to support. This
3978 configuration allows peer-to-peer DMA between
3979 any pair of devices, possibly at the cost of
3980 reduced performance. This also guarantees
3981 that hot-added devices will work.
3982 cbiosize=nn[KMG] The fixed amount of bus space which is
3983 reserved for the CardBus bridge's IO window.
3984 The default value is 256 bytes.
3985 cbmemsize=nn[KMG] The fixed amount of bus space which is
3986 reserved for the CardBus bridge's memory
3987 window. The default value is 64 megabytes.
3990 [<order of align>@]<pci_dev>[; ...]
3991 Specifies alignment and device to reassign
3992 aligned memory resources. How to
3993 specify the device is described above.
3994 If <order of align> is not specified,
3995 PAGE_SIZE is used as alignment.
3996 A PCI-PCI bridge can be specified if resource
3997 windows need to be expanded.
3998 To specify the alignment for several
3999 instances of a device, the PCI vendor,
4000 device, subvendor, and subdevice may be
4001 specified, e.g., 12@pci:8086:9c22:103c:198f
4002 for 4096-byte alignment.
4003 ecrc= Enable/disable PCIe ECRC (transaction layer
4004 end-to-end CRC checking).
4005 bios: Use BIOS/firmware settings. This is the
4009 hpiosize=nn[KMG] The fixed amount of bus space which is
4010 reserved for hotplug bridge's IO window.
4011 Default size is 256 bytes.
4012 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4013 reserved for hotplug bridge's MMIO window.
4014 Default size is 2 megabytes.
4015 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4016 reserved for hotplug bridge's MMIO_PREF window.
4017 Default size is 2 megabytes.
4018 hpmemsize=nn[KMG] The fixed amount of bus space which is
4019 reserved for hotplug bridge's MMIO and
4021 Default size is 2 megabytes.
4022 hpbussize=nn The minimum amount of additional bus numbers
4023 reserved for buses below a hotplug bridge.
4025 realloc= Enable/disable reallocating PCI bridge resources
4026 if allocations done by BIOS are too small to
4027 accommodate resources required by all child
4029 off: Turn realloc off
4031 realloc same as realloc=on
4032 noari do not use PCIe ARI.
4033 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4034 do not use PCIe ATS (and IOMMU device IOTLB).
4035 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4036 only look for one device below a PCIe downstream
4038 big_root_window Try to add a big 64bit memory window to the PCIe
4039 root complex on AMD CPUs. Some GFX hardware
4040 can resize a BAR to allow access to all VRAM.
4041 Adding the window is slightly risky (it may
4042 conflict with unreported devices), so this
4044 disable_acs_redir=<pci_dev>[; ...]
4045 Specify one or more PCI devices (in the format
4046 specified above) separated by semicolons.
4047 Each device specified will have the PCI ACS
4048 redirect capabilities forced off which will
4049 allow P2P traffic between devices through
4050 bridges without forcing it upstream. Note:
4051 this removes isolation between devices and
4052 may put more devices in an IOMMU group.
4053 force_floating [S390] Force usage of floating interrupts.
4054 nomio [S390] Do not use MIO instructions.
4055 norid [S390] ignore the RID field and force use of
4056 one PCI domain per PCI function
4058 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4061 force Enable ASPM even on devices that claim not to support it.
4062 WARNING: Forcing ASPM on may cause system lockups.
4064 pcie_ports= [PCIE] PCIe port services handling:
4065 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4066 even if the platform doesn't give the OS permission to
4067 use them. This may cause conflicts if the platform
4068 also tries to use these services.
4069 dpc-native Use native PCIe service for DPC only. May
4070 cause conflicts if firmware uses AER or DPC.
4071 compat Disable native PCIe services (PME, AER, DPC, PCIe
4074 pcie_port_pm= [PCIE] PCIe port power management handling:
4075 off Disable power management of all PCIe ports
4076 force Forcibly enable power management of all PCIe ports
4078 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4079 nomsi Do not use MSI for native PCIe PME signaling (this makes
4080 all PCIe root ports use INTx for all services).
4082 pcmv= [HW,PCMCIA] BadgePAD 4
4086 Keep all power-domains already enabled by bootloader on,
4087 even if no driver has claimed them. This is useful
4088 for debug and development, but should not be
4089 needed on a platform with proper driver support.
4092 See Documentation/admin-guide/blockdev/paride.rst.
4094 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4097 See arch/parisc/kernel/pdc_chassis.c
4099 percpu_alloc= Select which percpu first chunk allocator to use.
4100 Currently supported values are "embed" and "page".
4101 Archs may support subset or none of the selections.
4102 See comments in mm/percpu.c for details on each
4103 allocator. This parameter is primarily for debugging
4104 and performance comparison.
4107 See Documentation/admin-guide/blockdev/paride.rst.
4110 See Documentation/admin-guide/blockdev/paride.rst.
4112 pirq= [SMP,APIC] Manual mp-table setup
4113 See Documentation/x86/i386/IO-APIC.rst.
4115 plip= [PPT,NET] Parallel port network link
4116 Format: { parport<nr> | timid | 0 }
4117 See also Documentation/admin-guide/parport.rst.
4119 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4120 Override pmtimer IOPort with a hex value.
4123 pm_debug_messages [SUSPEND,KNL]
4124 Enable suspend/resume debug messages during boot up.
4127 Enable PNP debug messages (depends on the
4128 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4129 via /sys/module/pnp/parameters/debug. We always show
4130 current resource usage; turning this on also shows
4131 possible settings and some assignment information.
4137 { on | off | curr | res | no-curr | no-res }
4140 [ISAPNP] Exclude IRQs for the autoconfiguration
4143 [ISAPNP] Exclude DMAs for the autoconfiguration
4145 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4146 Ranges are in pairs (I/O port base and size).
4149 [ISAPNP] Exclude memory regions for the
4151 Ranges are in pairs (memory base and size).
4153 ports= [IP_VS_FTP] IPVS ftp helper module
4155 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4157 Format: <port>,<port>....
4159 powersave=off [PPC] This option disables power saving features.
4160 It specifically disables cpuidle and sets the
4161 platform machine description specific power_save
4162 function to NULL. On Idle the CPU just reduces
4165 ppc_strict_facility_enable
4166 [PPC] This option catches any kernel floating point,
4167 Altivec, VSX and SPE outside of regions specifically
4168 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4169 There is some performance impact when enabling this.
4173 Disable Hardware Transactional Memory
4176 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4177 none - Limited to cond_resched() calls
4178 voluntary - Limited to cond_resched() and might_sleep() calls
4179 full - Any section that isn't explicitly preempt disabled
4180 can be preempted anytime.
4182 print-fatal-signals=
4183 [KNL] debug: print fatal signals
4185 If enabled, warn about various signal handling
4186 related application anomalies: too many signals,
4187 too many POSIX.1 timers, fatal signals causing a
4190 If you hit the warning due to signal overflow,
4191 you might want to try "ulimit -i unlimited".
4195 printk.always_kmsg_dump=
4196 Trigger kmsg_dump for cases other than kernel oops or
4198 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4201 printk.devkmsg={on,off,ratelimit}
4202 Control writing to /dev/kmsg.
4203 on - unlimited logging to /dev/kmsg from userspace
4204 off - logging to /dev/kmsg disabled
4205 ratelimit - ratelimit the logging
4208 printk.time= Show timing data prefixed to each printk message line
4209 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4211 processor.max_cstate= [HW,ACPI]
4212 Limit processor to maximum C-state
4213 max_cstate=9 overrides any DMI blacklist limit.
4215 processor.nocst [HW,ACPI]
4216 Ignore the _CST method to determine C-states,
4217 instead using the legacy FADT method
4219 profile= [KNL] Enable kernel profiling via /proc/profile
4220 Format: [<profiletype>,]<number>
4221 Param: <profiletype>: "schedule", "sleep", or "kvm"
4222 [defaults to kernel profiling]
4223 Param: "schedule" - profile schedule points.
4224 Param: "sleep" - profile D-state sleeping (millisecs).
4225 Requires CONFIG_SCHEDSTATS
4226 Param: "kvm" - profile VM exits.
4227 Param: <number> - step/bucket size as a power of 2 for
4228 statistical time based profiling.
4230 prompt_ramdisk= [RAM] [Deprecated]
4232 prot_virt= [S390] enable hosting protected virtual machines
4233 isolated from the hypervisor (if hardware supports
4237 psi= [KNL] Enable or disable pressure stall information
4241 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4242 probe for; one of (bare|imps|exps|lifebook|any).
4243 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4245 psmouse.resetafter= [HW,MOUSE]
4246 Try to reset the device after so many bad packets
4249 [HW,MOUSE] Set desired mouse resolution, in dpi.
4250 psmouse.smartscroll=
4251 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4252 0 = disabled, 1 = enabled (default).
4254 pstore.backend= Specify the name of the pstore backend to use
4257 See Documentation/admin-guide/blockdev/paride.rst.
4259 pti= [X86-64] Control Page Table Isolation of user and
4260 kernel address spaces. Disabling this feature
4261 removes hardening, but improves performance of
4262 system calls and interrupts.
4264 on - unconditionally enable
4265 off - unconditionally disable
4266 auto - kernel detects whether your CPU model is
4267 vulnerable to issues that PTI mitigates
4269 Not specifying this option is equivalent to pti=auto.
4272 Equivalent to pti=off
4275 [KNL] Number of legacy pty's. Overwrites compiled-in
4278 quiet [KNL] Disable most log messages
4283 See Documentation/admin-guide/md.rst.
4285 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4286 See Documentation/admin-guide/blockdev/ramdisk.rst.
4288 ramdisk_start= [RAM] RAM disk image start address
4290 random.trust_cpu={on,off}
4291 [KNL] Enable or disable trusting the use of the
4292 CPU's random number generator (if available) to
4293 fully seed the kernel's CRNG. Default is controlled
4294 by CONFIG_RANDOM_TRUST_CPU.
4296 randomize_kstack_offset=
4297 [KNL] Enable or disable kernel stack offset
4298 randomization, which provides roughly 5 bits of
4299 entropy, frustrating memory corruption attacks
4300 that depend on stack address determinism or
4301 cross-syscall address exposures. This is only
4302 available on architectures that have defined
4303 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4304 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4305 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4307 ras=option[,option,...] [KNL] RAS-specific options
4310 Disable the Correctable Errors Collector,
4311 see CONFIG_RAS_CEC help text.
4314 The argument is a cpu list, as described above.
4316 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
4317 the specified list of CPUs to be no-callback CPUs.
4318 Invocation of these CPUs' RCU callbacks will be
4319 offloaded to "rcuox/N" kthreads created for that
4320 purpose, where "x" is "p" for RCU-preempt, and
4321 "s" for RCU-sched, and "N" is the CPU number.
4322 This reduces OS jitter on the offloaded CPUs,
4323 which can be useful for HPC and real-time
4324 workloads. It can also improve energy efficiency
4325 for asymmetric multiprocessors.
4328 Rather than requiring that offloaded CPUs
4329 (specified by rcu_nocbs= above) explicitly
4330 awaken the corresponding "rcuoN" kthreads,
4331 make these kthreads poll for callbacks.
4332 This improves the real-time response for the
4333 offloaded CPUs by relieving them of the need to
4334 wake up the corresponding kthread, but degrades
4335 energy efficiency by requiring that the kthreads
4336 periodically wake up to do the polling.
4338 rcutree.blimit= [KNL]
4339 Set maximum number of finished RCU callbacks to
4340 process in one batch.
4342 rcutree.dump_tree= [KNL]
4343 Dump the structure of the rcu_node combining tree
4344 out at early boot. This is used for diagnostic
4345 purposes, to verify correct tree setup.
4347 rcutree.gp_cleanup_delay= [KNL]
4348 Set the number of jiffies to delay each step of
4349 RCU grace-period cleanup.
4351 rcutree.gp_init_delay= [KNL]
4352 Set the number of jiffies to delay each step of
4353 RCU grace-period initialization.
4355 rcutree.gp_preinit_delay= [KNL]
4356 Set the number of jiffies to delay each step of
4357 RCU grace-period pre-initialization, that is,
4358 the propagation of recent CPU-hotplug changes up
4359 the rcu_node combining tree.
4361 rcutree.use_softirq= [KNL]
4362 If set to zero, move all RCU_SOFTIRQ processing to
4363 per-CPU rcuc kthreads. Defaults to a non-zero
4364 value, meaning that RCU_SOFTIRQ is used by default.
4365 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4367 But note that CONFIG_PREEMPT_RT=y kernels disable
4368 this kernel boot parameter, forcibly setting it
4371 rcutree.rcu_fanout_exact= [KNL]
4372 Disable autobalancing of the rcu_node combining
4373 tree. This is used by rcutorture, and might
4374 possibly be useful for architectures having high
4375 cache-to-cache transfer latencies.
4377 rcutree.rcu_fanout_leaf= [KNL]
4378 Change the number of CPUs assigned to each
4379 leaf rcu_node structure. Useful for very
4380 large systems, which will choose the value 64,
4381 and for NUMA systems with large remote-access
4382 latencies, which will choose a value aligned
4383 with the appropriate hardware boundaries.
4385 rcutree.rcu_min_cached_objs= [KNL]
4386 Minimum number of objects which are cached and
4387 maintained per one CPU. Object size is equal
4388 to PAGE_SIZE. The cache allows to reduce the
4389 pressure to page allocator, also it makes the
4390 whole algorithm to behave better in low memory
4393 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4394 Set the page-cache refill delay (in milliseconds)
4395 in response to low-memory conditions. The range
4396 of permitted values is in the range 0:100000.
4398 rcutree.jiffies_till_first_fqs= [KNL]
4399 Set delay from grace-period initialization to
4400 first attempt to force quiescent states.
4401 Units are jiffies, minimum value is zero,
4402 and maximum value is HZ.
4404 rcutree.jiffies_till_next_fqs= [KNL]
4405 Set delay between subsequent attempts to force
4406 quiescent states. Units are jiffies, minimum
4407 value is one, and maximum value is HZ.
4409 rcutree.jiffies_till_sched_qs= [KNL]
4410 Set required age in jiffies for a
4411 given grace period before RCU starts
4412 soliciting quiescent-state help from
4413 rcu_note_context_switch() and cond_resched().
4414 If not specified, the kernel will calculate
4415 a value based on the most recent settings
4416 of rcutree.jiffies_till_first_fqs
4417 and rcutree.jiffies_till_next_fqs.
4418 This calculated value may be viewed in
4419 rcutree.jiffies_to_sched_qs. Any attempt to set
4420 rcutree.jiffies_to_sched_qs will be cheerfully
4423 rcutree.kthread_prio= [KNL,BOOT]
4424 Set the SCHED_FIFO priority of the RCU per-CPU
4425 kthreads (rcuc/N). This value is also used for
4426 the priority of the RCU boost threads (rcub/N)
4427 and for the RCU grace-period kthreads (rcu_bh,
4428 rcu_preempt, and rcu_sched). If RCU_BOOST is
4429 set, valid values are 1-99 and the default is 1
4430 (the least-favored priority). Otherwise, when
4431 RCU_BOOST is not set, valid values are 0-99 and
4432 the default is zero (non-realtime operation).
4434 rcutree.rcu_nocb_gp_stride= [KNL]
4435 Set the number of NOCB callback kthreads in
4436 each group, which defaults to the square root
4437 of the number of CPUs. Larger numbers reduce
4438 the wakeup overhead on the global grace-period
4439 kthread, but increases that same overhead on
4440 each group's NOCB grace-period kthread.
4442 rcutree.qhimark= [KNL]
4443 Set threshold of queued RCU callbacks beyond which
4444 batch limiting is disabled.
4446 rcutree.qlowmark= [KNL]
4447 Set threshold of queued RCU callbacks below which
4448 batch limiting is re-enabled.
4450 rcutree.qovld= [KNL]
4451 Set threshold of queued RCU callbacks beyond which
4452 RCU's force-quiescent-state scan will aggressively
4453 enlist help from cond_resched() and sched IPIs to
4454 help CPUs more quickly reach quiescent states.
4455 Set to less than zero to make this be set based
4456 on rcutree.qhimark at boot time and to zero to
4457 disable more aggressive help enlistment.
4459 rcutree.rcu_idle_gp_delay= [KNL]
4460 Set wakeup interval for idle CPUs that have
4461 RCU callbacks (RCU_FAST_NO_HZ=y).
4463 rcutree.rcu_kick_kthreads= [KNL]
4464 Cause the grace-period kthread to get an extra
4465 wake_up() if it sleeps three times longer than
4466 it should at force-quiescent-state time.
4467 This wake_up() will be accompanied by a
4468 WARN_ONCE() splat and an ftrace_dump().
4470 rcutree.rcu_unlock_delay= [KNL]
4471 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4472 this specifies an rcu_read_unlock()-time delay
4473 in microseconds. This defaults to zero.
4474 Larger delays increase the probability of
4475 catching RCU pointer leaks, that is, buggy use
4476 of RCU-protected pointers after the relevant
4477 rcu_read_unlock() has completed.
4479 rcutree.sysrq_rcu= [KNL]
4480 Commandeer a sysrq key to dump out Tree RCU's
4481 rcu_node tree with an eye towards determining
4482 why a new grace period has not yet started.
4484 rcuscale.gp_async= [KNL]
4485 Measure performance of asynchronous
4486 grace-period primitives such as call_rcu().
4488 rcuscale.gp_async_max= [KNL]
4489 Specify the maximum number of outstanding
4490 callbacks per writer thread. When a writer
4491 thread exceeds this limit, it invokes the
4492 corresponding flavor of rcu_barrier() to allow
4493 previously posted callbacks to drain.
4495 rcuscale.gp_exp= [KNL]
4496 Measure performance of expedited synchronous
4497 grace-period primitives.
4499 rcuscale.holdoff= [KNL]
4500 Set test-start holdoff period. The purpose of
4501 this parameter is to delay the start of the
4502 test until boot completes in order to avoid
4505 rcuscale.kfree_rcu_test= [KNL]
4506 Set to measure performance of kfree_rcu() flooding.
4508 rcuscale.kfree_rcu_test_double= [KNL]
4509 Test the double-argument variant of kfree_rcu().
4510 If this parameter has the same value as
4511 rcuscale.kfree_rcu_test_single, both the single-
4512 and double-argument variants are tested.
4514 rcuscale.kfree_rcu_test_single= [KNL]
4515 Test the single-argument variant of kfree_rcu().
4516 If this parameter has the same value as
4517 rcuscale.kfree_rcu_test_double, both the single-
4518 and double-argument variants are tested.
4520 rcuscale.kfree_nthreads= [KNL]
4521 The number of threads running loops of kfree_rcu().
4523 rcuscale.kfree_alloc_num= [KNL]
4524 Number of allocations and frees done in an iteration.
4526 rcuscale.kfree_loops= [KNL]
4527 Number of loops doing rcuscale.kfree_alloc_num number
4528 of allocations and frees.
4530 rcuscale.nreaders= [KNL]
4531 Set number of RCU readers. The value -1 selects
4532 N, where N is the number of CPUs. A value
4533 "n" less than -1 selects N-n+1, where N is again
4534 the number of CPUs. For example, -2 selects N
4535 (the number of CPUs), -3 selects N+1, and so on.
4536 A value of "n" less than or equal to -N selects
4539 rcuscale.nwriters= [KNL]
4540 Set number of RCU writers. The values operate
4541 the same as for rcuscale.nreaders.
4542 N, where N is the number of CPUs
4544 rcuscale.perf_type= [KNL]
4545 Specify the RCU implementation to test.
4547 rcuscale.shutdown= [KNL]
4548 Shut the system down after performance tests
4549 complete. This is useful for hands-off automated
4552 rcuscale.verbose= [KNL]
4553 Enable additional printk() statements.
4555 rcuscale.writer_holdoff= [KNL]
4556 Write-side holdoff between grace periods,
4557 in microseconds. The default of zero says
4560 rcutorture.fqs_duration= [KNL]
4561 Set duration of force_quiescent_state bursts
4564 rcutorture.fqs_holdoff= [KNL]
4565 Set holdoff time within force_quiescent_state bursts
4568 rcutorture.fqs_stutter= [KNL]
4569 Set wait time between force_quiescent_state bursts
4572 rcutorture.fwd_progress= [KNL]
4573 Enable RCU grace-period forward-progress testing
4574 for the types of RCU supporting this notion.
4576 rcutorture.fwd_progress_div= [KNL]
4577 Specify the fraction of a CPU-stall-warning
4578 period to do tight-loop forward-progress testing.
4580 rcutorture.fwd_progress_holdoff= [KNL]
4581 Number of seconds to wait between successive
4582 forward-progress tests.
4584 rcutorture.fwd_progress_need_resched= [KNL]
4585 Enclose cond_resched() calls within checks for
4586 need_resched() during tight-loop forward-progress
4589 rcutorture.gp_cond= [KNL]
4590 Use conditional/asynchronous update-side
4591 primitives, if available.
4593 rcutorture.gp_exp= [KNL]
4594 Use expedited update-side primitives, if available.
4596 rcutorture.gp_normal= [KNL]
4597 Use normal (non-expedited) asynchronous
4598 update-side primitives, if available.
4600 rcutorture.gp_sync= [KNL]
4601 Use normal (non-expedited) synchronous
4602 update-side primitives, if available. If all
4603 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4604 rcutorture.gp_normal=, and rcutorture.gp_sync=
4605 are zero, rcutorture acts as if is interpreted
4606 they are all non-zero.
4608 rcutorture.irqreader= [KNL]
4609 Run RCU readers from irq handlers, or, more
4610 accurately, from a timer handler. Not all RCU
4611 flavors take kindly to this sort of thing.
4613 rcutorture.leakpointer= [KNL]
4614 Leak an RCU-protected pointer out of the reader.
4615 This can of course result in splats, and is
4616 intended to test the ability of things like
4617 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4620 rcutorture.n_barrier_cbs= [KNL]
4621 Set callbacks/threads for rcu_barrier() testing.
4623 rcutorture.nfakewriters= [KNL]
4624 Set number of concurrent RCU writers. These just
4625 stress RCU, they don't participate in the actual
4626 test, hence the "fake".
4628 rcutorture.nocbs_nthreads= [KNL]
4629 Set number of RCU callback-offload togglers.
4630 Zero (the default) disables toggling.
4632 rcutorture.nocbs_toggle= [KNL]
4633 Set the delay in milliseconds between successive
4634 callback-offload toggling attempts.
4636 rcutorture.nreaders= [KNL]
4637 Set number of RCU readers. The value -1 selects
4638 N-1, where N is the number of CPUs. A value
4639 "n" less than -1 selects N-n-2, where N is again
4640 the number of CPUs. For example, -2 selects N
4641 (the number of CPUs), -3 selects N+1, and so on.
4643 rcutorture.object_debug= [KNL]
4644 Enable debug-object double-call_rcu() testing.
4646 rcutorture.onoff_holdoff= [KNL]
4647 Set time (s) after boot for CPU-hotplug testing.
4649 rcutorture.onoff_interval= [KNL]
4650 Set time (jiffies) between CPU-hotplug operations,
4651 or zero to disable CPU-hotplug testing.
4653 rcutorture.read_exit= [KNL]
4654 Set the number of read-then-exit kthreads used
4655 to test the interaction of RCU updaters and
4656 task-exit processing.
4658 rcutorture.read_exit_burst= [KNL]
4659 The number of times in a given read-then-exit
4660 episode that a set of read-then-exit kthreads
4663 rcutorture.read_exit_delay= [KNL]
4664 The delay, in seconds, between successive
4665 read-then-exit testing episodes.
4667 rcutorture.shuffle_interval= [KNL]
4668 Set task-shuffle interval (s). Shuffling tasks
4669 allows some CPUs to go into dyntick-idle mode
4670 during the rcutorture test.
4672 rcutorture.shutdown_secs= [KNL]
4673 Set time (s) after boot system shutdown. This
4674 is useful for hands-off automated testing.
4676 rcutorture.stall_cpu= [KNL]
4677 Duration of CPU stall (s) to test RCU CPU stall
4678 warnings, zero to disable.
4680 rcutorture.stall_cpu_block= [KNL]
4681 Sleep while stalling if set. This will result
4682 in warnings from preemptible RCU in addition
4683 to any other stall-related activity.
4685 rcutorture.stall_cpu_holdoff= [KNL]
4686 Time to wait (s) after boot before inducing stall.
4688 rcutorture.stall_cpu_irqsoff= [KNL]
4689 Disable interrupts while stalling if set.
4691 rcutorture.stall_gp_kthread= [KNL]
4692 Duration (s) of forced sleep within RCU
4693 grace-period kthread to test RCU CPU stall
4694 warnings, zero to disable. If both stall_cpu
4695 and stall_gp_kthread are specified, the
4696 kthread is starved first, then the CPU.
4698 rcutorture.stat_interval= [KNL]
4699 Time (s) between statistics printk()s.
4701 rcutorture.stutter= [KNL]
4702 Time (s) to stutter testing, for example, specifying
4703 five seconds causes the test to run for five seconds,
4704 wait for five seconds, and so on. This tests RCU's
4705 ability to transition abruptly to and from idle.
4707 rcutorture.test_boost= [KNL]
4708 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4709 "Maybe" means test if the RCU implementation
4710 under test support RCU priority boosting.
4712 rcutorture.test_boost_duration= [KNL]
4713 Duration (s) of each individual boost test.
4715 rcutorture.test_boost_interval= [KNL]
4716 Interval (s) between each boost test.
4718 rcutorture.test_no_idle_hz= [KNL]
4719 Test RCU's dyntick-idle handling. See also the
4720 rcutorture.shuffle_interval parameter.
4722 rcutorture.torture_type= [KNL]
4723 Specify the RCU implementation to test.
4725 rcutorture.verbose= [KNL]
4726 Enable additional printk() statements.
4728 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4729 Dump ftrace buffer after reporting RCU CPU
4732 rcupdate.rcu_cpu_stall_suppress= [KNL]
4733 Suppress RCU CPU stall warning messages.
4735 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4736 Suppress RCU CPU stall warning messages and
4737 rcutorture writer stall warnings that occur
4738 during early boot, that is, during the time
4739 before the init task is spawned.
4741 rcupdate.rcu_cpu_stall_timeout= [KNL]
4742 Set timeout for RCU CPU stall warning messages.
4744 rcupdate.rcu_expedited= [KNL]
4745 Use expedited grace-period primitives, for
4746 example, synchronize_rcu_expedited() instead
4747 of synchronize_rcu(). This reduces latency,
4748 but can increase CPU utilization, degrade
4749 real-time latency, and degrade energy efficiency.
4750 No effect on CONFIG_TINY_RCU kernels.
4752 rcupdate.rcu_normal= [KNL]
4753 Use only normal grace-period primitives,
4754 for example, synchronize_rcu() instead of
4755 synchronize_rcu_expedited(). This improves
4756 real-time latency, CPU utilization, and
4757 energy efficiency, but can expose users to
4758 increased grace-period latency. This parameter
4759 overrides rcupdate.rcu_expedited. No effect on
4760 CONFIG_TINY_RCU kernels.
4762 rcupdate.rcu_normal_after_boot= [KNL]
4763 Once boot has completed (that is, after
4764 rcu_end_inkernel_boot() has been invoked), use
4765 only normal grace-period primitives. No effect
4766 on CONFIG_TINY_RCU kernels.
4768 But note that CONFIG_PREEMPT_RT=y kernels enables
4769 this kernel boot parameter, forcibly setting
4770 it to the value one, that is, converting any
4771 post-boot attempt at an expedited RCU grace
4772 period to instead use normal non-expedited
4773 grace-period processing.
4775 rcupdate.rcu_task_ipi_delay= [KNL]
4776 Set time in jiffies during which RCU tasks will
4777 avoid sending IPIs, starting with the beginning
4778 of a given grace period. Setting a large
4779 number avoids disturbing real-time workloads,
4780 but lengthens grace periods.
4782 rcupdate.rcu_task_stall_timeout= [KNL]
4783 Set timeout in jiffies for RCU task stall warning
4784 messages. Disable with a value less than or equal
4787 rcupdate.rcu_self_test= [KNL]
4788 Run the RCU early boot self tests
4792 Run specified binary instead of /init from the ramdisk,
4793 used for early userspace startup. See initrd.
4796 force - Override the decision by the kernel to hide the
4797 advertisement of RDRAND support (this affects
4798 certain AMD processors because of buggy BIOS
4799 support, specifically around the suspend/resume
4803 Turn on/off individual RDT features. List is:
4804 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4806 E.g. to turn on cmt and turn off mba use:
4810 Format (x86 or x86_64):
4811 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
4813 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4815 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4816 (prefix with 'panic_' to set mode for panic
4818 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4819 reboot_force is either force or not specified,
4820 reboot_cpu is s[mp]#### with #### being the processor
4821 to be used for rebooting.
4823 refscale.holdoff= [KNL]
4824 Set test-start holdoff period. The purpose of
4825 this parameter is to delay the start of the
4826 test until boot completes in order to avoid
4829 refscale.loops= [KNL]
4830 Set the number of loops over the synchronization
4831 primitive under test. Increasing this number
4832 reduces noise due to loop start/end overhead,
4833 but the default has already reduced the per-pass
4834 noise to a handful of picoseconds on ca. 2020
4837 refscale.nreaders= [KNL]
4838 Set number of readers. The default value of -1
4839 selects N, where N is roughly 75% of the number
4840 of CPUs. A value of zero is an interesting choice.
4842 refscale.nruns= [KNL]
4843 Set number of runs, each of which is dumped onto
4846 refscale.readdelay= [KNL]
4847 Set the read-side critical-section duration,
4848 measured in microseconds.
4850 refscale.scale_type= [KNL]
4851 Specify the read-protection implementation to test.
4853 refscale.shutdown= [KNL]
4854 Shut down the system at the end of the performance
4855 test. This defaults to 1 (shut it down) when
4856 refscale is built into the kernel and to 0 (leave
4857 it running) when refscale is built as a module.
4859 refscale.verbose= [KNL]
4860 Enable additional printk() statements.
4862 refscale.verbose_batched= [KNL]
4863 Batch the additional printk() statements. If zero
4864 (the default) or negative, print everything. Otherwise,
4865 print every Nth verbose statement, where N is the value
4869 [KNL, SMP] Set scheduler's default relax_domain_level.
4870 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4872 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4873 Format: <base1>,<size1>[,<base2>,<size2>,...]
4874 Reserve I/O ports or memory so the kernel won't use
4875 them. If <base> is less than 0x10000, the region
4876 is assumed to be I/O ports; otherwise it is memory.
4878 reservetop= [X86-32]
4880 Reserves a hole at the top of the kernel virtual
4883 reset_devices [KNL] Force drivers to reset the underlying device
4884 during initialization.
4887 Specify the partition device for software suspend
4889 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4891 resume_offset= [SWSUSP]
4892 Specify the offset from the beginning of the partition
4893 given by "resume=" at which the swap header is located,
4894 in <PAGE_SIZE> units (needed only for swap files).
4895 See Documentation/power/swsusp-and-swap-files.rst
4897 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4898 read the resume files
4900 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4901 Useful for devices that are detected asynchronously
4902 (e.g. USB and MMC devices).
4904 hibernate= [HIBERNATION]
4905 noresume Don't check if there's a hibernation image
4906 present during boot.
4907 nocompress Don't compress/decompress hibernation images.
4908 no Disable hibernation and resume.
4909 protect_image Turn on image protection during restoration
4910 (that will set all pages holding image data
4911 during restoration read-only).
4913 retain_initrd [RAM] Keep initrd memory after extraction
4915 rfkill.default_state=
4916 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4917 etc. communication is blocked by default.
4920 rfkill.master_switch_mode=
4921 0 The "airplane mode" button does nothing.
4922 1 The "airplane mode" button toggles between everything
4923 blocked and the previous configuration.
4924 2 The "airplane mode" button toggles between everything
4925 blocked and everything unblocked.
4927 rhash_entries= [KNL,NET]
4928 Set number of hash buckets for route cache
4931 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4934 ro [KNL] Mount root device read-only on boot
4937 on Mark read-only kernel memory as read-only (default).
4938 off Leave read-only kernel memory writable for debugging.
4941 Enable the uart passthrough on the designated usb port
4942 on Rockchip SoCs. When active, the signals of the
4943 debug-uart get routed to the D+ and D- pins of the usb
4944 port and the regular usb controller gets disabled.
4946 root= [KNL] Root filesystem
4947 See name_to_dev_t comment in init/do_mounts.c.
4949 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4950 mount the root filesystem
4952 rootflags= [KNL] Set root filesystem mount option string
4954 rootfstype= [KNL] Set root filesystem type
4956 rootwait [KNL] Wait (indefinitely) for root device to show up.
4957 Useful for devices that are detected asynchronously
4958 (e.g. USB and MMC devices).
4960 rproc_mem=nn[KMG][@address]
4961 [KNL,ARM,CMA] Remoteproc physical memory block.
4962 Memory area to be used by remote processor image,
4965 rw [KNL] Mount root device read-write on boot
4967 S [KNL] Run init in single mode
4969 s390_iommu= [HW,S390]
4970 Set s390 IOTLB flushing mode
4972 With strict flushing every unmap operation will result in
4973 an IOTLB flush. Default is lazy flushing before reuse,
4977 See drivers/net/irda/sa1100_ir.c.
4979 sched_verbose [KNL] Enables verbose scheduler debug messages.
4981 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4982 Allowed values are enable and disable. This feature
4983 incurs a small amount of overhead in the scheduler
4984 but is useful for debugging and performance tuning.
4986 sched_thermal_decay_shift=
4987 [KNL, SMP] Set a decay shift for scheduler thermal
4988 pressure signal. Thermal pressure signal follows the
4989 default decay period of other scheduler pelt
4990 signals(usually 32 ms but configurable). Setting
4991 sched_thermal_decay_shift will left shift the decay
4992 period for the thermal pressure signal by the shift
4994 i.e. with the default pelt decay period of 32 ms
4995 sched_thermal_decay_shift thermal pressure decay pr
4999 Format: integer between 0 and 10
5002 scftorture.holdoff= [KNL]
5003 Number of seconds to hold off before starting
5004 test. Defaults to zero for module insertion and
5005 to 10 seconds for built-in smp_call_function()
5008 scftorture.longwait= [KNL]
5009 Request ridiculously long waits randomly selected
5010 up to the chosen limit in seconds. Zero (the
5011 default) disables this feature. Please note
5012 that requesting even small non-zero numbers of
5013 seconds can result in RCU CPU stall warnings,
5014 softlockup complaints, and so on.
5016 scftorture.nthreads= [KNL]
5017 Number of kthreads to spawn to invoke the
5018 smp_call_function() family of functions.
5019 The default of -1 specifies a number of kthreads
5020 equal to the number of CPUs.
5022 scftorture.onoff_holdoff= [KNL]
5023 Number seconds to wait after the start of the
5024 test before initiating CPU-hotplug operations.
5026 scftorture.onoff_interval= [KNL]
5027 Number seconds to wait between successive
5028 CPU-hotplug operations. Specifying zero (which
5029 is the default) disables CPU-hotplug operations.
5031 scftorture.shutdown_secs= [KNL]
5032 The number of seconds following the start of the
5033 test after which to shut down the system. The
5034 default of zero avoids shutting down the system.
5035 Non-zero values are useful for automated tests.
5037 scftorture.stat_interval= [KNL]
5038 The number of seconds between outputting the
5039 current test statistics to the console. A value
5040 of zero disables statistics output.
5042 scftorture.stutter_cpus= [KNL]
5043 The number of jiffies to wait between each change
5044 to the set of CPUs under test.
5046 scftorture.use_cpus_read_lock= [KNL]
5047 Use use_cpus_read_lock() instead of the default
5048 preempt_disable() to disable CPU hotplug
5049 while invoking one of the smp_call_function*()
5052 scftorture.verbose= [KNL]
5053 Enable additional printk() statements.
5055 scftorture.weight_single= [KNL]
5056 The probability weighting to use for the
5057 smp_call_function_single() function with a zero
5058 "wait" parameter. A value of -1 selects the
5059 default if all other weights are -1. However,
5060 if at least one weight has some other value, a
5061 value of -1 will instead select a weight of zero.
5063 scftorture.weight_single_wait= [KNL]
5064 The probability weighting to use for the
5065 smp_call_function_single() function with a
5066 non-zero "wait" parameter. See weight_single.
5068 scftorture.weight_many= [KNL]
5069 The probability weighting to use for the
5070 smp_call_function_many() function with a zero
5071 "wait" parameter. See weight_single.
5072 Note well that setting a high probability for
5073 this weighting can place serious IPI load
5076 scftorture.weight_many_wait= [KNL]
5077 The probability weighting to use for the
5078 smp_call_function_many() function with a
5079 non-zero "wait" parameter. See weight_single
5082 scftorture.weight_all= [KNL]
5083 The probability weighting to use for the
5084 smp_call_function_all() function with a zero
5085 "wait" parameter. See weight_single and
5088 scftorture.weight_all_wait= [KNL]
5089 The probability weighting to use for the
5090 smp_call_function_all() function with a
5091 non-zero "wait" parameter. See weight_single
5094 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5095 xtime_lock contention on larger systems, and/or RCU lock
5096 contention on all systems with CONFIG_MAXSMP set.
5097 Format: { "0" | "1" }
5098 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5100 Note: increases power consumption, thus should only be
5101 enabled if running jitter sensitive (HPC/RT) workloads.
5103 security= [SECURITY] Choose a legacy "major" security module to
5104 enable at boot. This has been deprecated by the
5107 selinux= [SELINUX] Disable or enable SELinux at boot time.
5108 Format: { "0" | "1" }
5109 See security/selinux/Kconfig help text.
5114 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5115 Format: { "0" | "1" }
5116 See security/apparmor/Kconfig help text
5119 Default value is set via kernel config option.
5121 serialnumber [BUGS=X86-32]
5124 Maximal number of shapers.
5132 Enable merging of slabs with similar size when the
5133 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5136 Disable merging of slabs with similar size. May be
5137 necessary if there is some reason to distinguish
5138 allocs to different slabs, especially in hardened
5139 environments where the risk of heap overflows and
5140 layout control by attackers can usually be
5141 frustrated by disabling merging. This will reduce
5142 most of the exposure of a heap attack to a single
5143 cache (risks via metadata attacks are mostly
5144 unchanged). Debug options disable merging on their
5146 For more information see Documentation/vm/slub.rst.
5148 slab_max_order= [MM, SLAB]
5149 Determines the maximum allowed order for slabs.
5150 A high setting may cause OOMs due to memory
5151 fragmentation. Defaults to 1 for systems with
5152 more than 32MB of RAM, 0 otherwise.
5154 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5155 Enabling slub_debug allows one to determine the
5156 culprit if slab objects become corrupted. Enabling
5157 slub_debug can create guard zones around objects and
5158 may poison objects when not in use. Also tracks the
5159 last alloc / free. For more information see
5160 Documentation/vm/slub.rst.
5162 slub_max_order= [MM, SLUB]
5163 Determines the maximum allowed order for slabs.
5164 A high setting may cause OOMs due to memory
5165 fragmentation. For more information see
5166 Documentation/vm/slub.rst.
5168 slub_min_objects= [MM, SLUB]
5169 The minimum number of objects per slab. SLUB will
5170 increase the slab order up to slub_max_order to
5171 generate a sufficiently large slab able to contain
5172 the number of objects indicated. The higher the number
5173 of objects the smaller the overhead of tracking slabs
5174 and the less frequently locks need to be acquired.
5175 For more information see Documentation/vm/slub.rst.
5177 slub_min_order= [MM, SLUB]
5178 Determines the minimum page order for slabs. Must be
5179 lower than slub_max_order.
5180 For more information see Documentation/vm/slub.rst.
5182 slub_merge [MM, SLUB]
5183 Same with slab_merge.
5185 slub_nomerge [MM, SLUB]
5186 Same with slab_nomerge. This is supported for legacy.
5187 See slab_nomerge for more information.
5190 Format: <io1>[,<io2>[,...,<io8>]]
5192 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5193 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5194 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5195 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5196 smsc-ircc2.ircc_irq= [HW] IRQ line
5197 smsc-ircc2.ircc_dma= [HW] DMA channel
5198 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5199 0: Toshiba Satellite 1800 (GP data pin select)
5200 1: Fast pin select (default)
5203 smt [KNL,S390] Set the maximum number of threads (logical
5204 CPUs) to use per physical CPU on systems capable of
5205 symmetric multithreading (SMT). Will be capped to the
5206 actual hardware limit.
5208 Default: -1 (no limit)
5211 [KNL] Should the soft-lockup detector generate panics.
5214 A value of 1 instructs the soft-lockup detector
5215 to panic the machine when a soft-lockup occurs. It is
5216 also controlled by the kernel.softlockup_panic sysctl
5217 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5218 respective build-time switch to that functionality.
5220 softlockup_all_cpu_backtrace=
5221 [KNL] Should the soft-lockup detector generate
5222 backtraces on all cpus.
5225 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5226 See Documentation/admin-guide/laptops/sonypi.rst
5228 spectre_v2= [X86] Control mitigation of Spectre variant 2
5229 (indirect branch speculation) vulnerability.
5230 The default operation protects the kernel from
5233 on - unconditionally enable, implies
5235 off - unconditionally disable, implies
5237 auto - kernel detects whether your CPU model is
5240 Selecting 'on' will, and 'auto' may, choose a
5241 mitigation method at run time according to the
5242 CPU, the available microcode, the setting of the
5243 CONFIG_RETPOLINE configuration option, and the
5244 compiler with which the kernel was built.
5246 Selecting 'on' will also enable the mitigation
5247 against user space to user space task attacks.
5249 Selecting 'off' will disable both the kernel and
5250 the user space protections.
5252 Specific mitigations can also be selected manually:
5254 retpoline - replace indirect branches
5255 retpoline,generic - google's original retpoline
5256 retpoline,amd - AMD-specific minimal thunk
5258 Not specifying this option is equivalent to
5262 [X86] Control mitigation of Spectre variant 2
5263 (indirect branch speculation) vulnerability between
5266 on - Unconditionally enable mitigations. Is
5267 enforced by spectre_v2=on
5269 off - Unconditionally disable mitigations. Is
5270 enforced by spectre_v2=off
5272 prctl - Indirect branch speculation is enabled,
5273 but mitigation can be enabled via prctl
5274 per thread. The mitigation control state
5275 is inherited on fork.
5278 - Like "prctl" above, but only STIBP is
5279 controlled per thread. IBPB is issued
5280 always when switching between different user
5284 - Same as "prctl" above, but all seccomp
5285 threads will enable the mitigation unless
5286 they explicitly opt out.
5289 - Like "seccomp" above, but only STIBP is
5290 controlled per thread. IBPB is issued
5291 always when switching between different
5292 user space processes.
5294 auto - Kernel selects the mitigation depending on
5295 the available CPU features and vulnerability.
5298 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
5300 Not specifying this option is equivalent to
5301 spectre_v2_user=auto.
5303 spec_store_bypass_disable=
5304 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5305 (Speculative Store Bypass vulnerability)
5307 Certain CPUs are vulnerable to an exploit against a
5308 a common industry wide performance optimization known
5309 as "Speculative Store Bypass" in which recent stores
5310 to the same memory location may not be observed by
5311 later loads during speculative execution. The idea
5312 is that such stores are unlikely and that they can
5313 be detected prior to instruction retirement at the
5314 end of a particular speculation execution window.
5316 In vulnerable processors, the speculatively forwarded
5317 store can be used in a cache side channel attack, for
5318 example to read memory to which the attacker does not
5319 directly have access (e.g. inside sandboxed code).
5321 This parameter controls whether the Speculative Store
5322 Bypass optimization is used.
5324 On x86 the options are:
5326 on - Unconditionally disable Speculative Store Bypass
5327 off - Unconditionally enable Speculative Store Bypass
5328 auto - Kernel detects whether the CPU model contains an
5329 implementation of Speculative Store Bypass and
5330 picks the most appropriate mitigation. If the
5331 CPU is not vulnerable, "off" is selected. If the
5332 CPU is vulnerable the default mitigation is
5333 architecture and Kconfig dependent. See below.
5334 prctl - Control Speculative Store Bypass per thread
5335 via prctl. Speculative Store Bypass is enabled
5336 for a process by default. The state of the control
5337 is inherited on fork.
5338 seccomp - Same as "prctl" above, but all seccomp threads
5339 will disable SSB unless they explicitly opt out.
5341 Default mitigations:
5342 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
5344 On powerpc the options are:
5346 on,auto - On Power8 and Power9 insert a store-forwarding
5347 barrier on kernel entry and exit. On Power7
5348 perform a software flush on kernel entry and
5352 Not specifying this option is equivalent to
5353 spec_store_bypass_disable=auto.
5355 spia_io_base= [HW,MTD]
5361 [X86] Enable split lock detection or bus lock detection
5363 When enabled (and if hardware support is present), atomic
5364 instructions that access data across cache line
5365 boundaries will result in an alignment check exception
5366 for split lock detection or a debug exception for
5371 warn - the kernel will emit rate-limited warnings
5372 about applications triggering the #AC
5373 exception or the #DB exception. This mode is
5374 the default on CPUs that support split lock
5375 detection or bus lock detection. Default
5376 behavior is by #AC if both features are
5377 enabled in hardware.
5379 fatal - the kernel will send SIGBUS to applications
5380 that trigger the #AC exception or the #DB
5381 exception. Default behavior is by #AC if
5382 both features are enabled in hardware.
5385 Set system wide rate limit to N bus locks
5386 per second for bus lock detection.
5389 N/A for split lock detection.
5392 If an #AC exception is hit in the kernel or in
5393 firmware (i.e. not while executing in user mode)
5394 the kernel will oops in either "warn" or "fatal"
5397 #DB exception for bus lock is triggered only when
5401 Control the Special Register Buffer Data Sampling
5404 Certain CPUs are vulnerable to an MDS-like
5405 exploit which can leak bits from the random
5408 By default, this issue is mitigated by
5409 microcode. However, the microcode fix can cause
5410 the RDRAND and RDSEED instructions to become
5411 much slower. Among other effects, this will
5412 result in reduced throughput from /dev/urandom.
5414 The microcode mitigation can be disabled with
5415 the following option:
5417 off: Disable mitigation and remove
5418 performance impact to RDRAND and RDSEED
5420 srcutree.counter_wrap_check [KNL]
5421 Specifies how frequently to check for
5422 grace-period sequence counter wrap for the
5423 srcu_data structure's ->srcu_gp_seq_needed field.
5424 The greater the number of bits set in this kernel
5425 parameter, the less frequently counter wrap will
5426 be checked for. Note that the bottom two bits
5429 srcutree.exp_holdoff [KNL]
5430 Specifies how many nanoseconds must elapse
5431 since the end of the last SRCU grace period for
5432 a given srcu_struct until the next normal SRCU
5433 grace period will be considered for automatic
5434 expediting. Set to zero to disable automatic
5438 Speculative Store Bypass Disable control
5440 On CPUs that are vulnerable to the Speculative
5441 Store Bypass vulnerability and offer a
5442 firmware based mitigation, this parameter
5443 indicates how the mitigation should be used:
5445 force-on: Unconditionally enable mitigation for
5446 for both kernel and userspace
5447 force-off: Unconditionally disable mitigation for
5448 for both kernel and userspace
5449 kernel: Always enable mitigation in the
5450 kernel, and offer a prctl interface
5451 to allow userspace to register its
5452 interest in being mitigated too.
5454 stack_guard_gap= [MM]
5455 override the default stack gap protection. The value
5456 is in page units and it defines how many pages prior
5457 to (for stacks growing down) resp. after (for stacks
5458 growing up) the main stack are reserved for no other
5459 mapping. Default value is 256 pages.
5461 stack_depot_disable= [KNL]
5462 Setting this to true through kernel command line will
5463 disable the stack depot thereby saving the static memory
5464 consumed by the stack hash table. By default this is set
5468 Enabled the stack tracer on boot up.
5470 stacktrace_filter=[function-list]
5471 [FTRACE] Limit the functions that the stack tracer
5472 will trace at boot up. function-list is a comma-separated
5473 list of functions. This list can be changed at run
5474 time by the stack_trace_filter file in the debugfs
5475 tracing directory. Note, this enables stack tracing
5476 and the stacktrace above is not needed.
5480 Set the STI (builtin display/keyboard on the HP-PARISC
5481 machines) console (graphic card) which should be used
5482 as the initial boot-console.
5483 See also comment in drivers/video/console/sticore.c.
5486 See comment in drivers/video/console/sticore.c.
5489 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5491 sunrpc.min_resvport=
5492 sunrpc.max_resvport=
5494 SunRPC servers often require that client requests
5495 originate from a privileged port (i.e. a port in the
5496 range 0 < portnr < 1024).
5497 An administrator who wishes to reserve some of these
5498 ports for other uses may adjust the range that the
5499 kernel's sunrpc client considers to be privileged
5500 using these two parameters to set the minimum and
5501 maximum port values.
5503 sunrpc.svc_rpc_per_connection_limit=
5505 Limit the number of requests that the server will
5506 process in parallel from a single connection.
5507 The default value is 0 (no limit).
5511 Control how the NFS server code allocates CPUs to
5512 service thread pools. Depending on how many NICs
5513 you have and where their interrupts are bound, this
5514 option will affect which CPUs will do NFS serving.
5515 Note: this parameter cannot be changed while the
5516 NFS server is running.
5518 auto the server chooses an appropriate mode
5519 automatically using heuristics
5520 global a single global pool contains all CPUs
5521 percpu one pool for each CPU
5522 pernode one pool for each NUMA node (equivalent
5523 to global on non-NUMA machines)
5525 sunrpc.tcp_slot_table_entries=
5526 sunrpc.udp_slot_table_entries=
5528 Sets the upper limit on the number of simultaneous
5529 RPC calls that can be sent from the client to a
5530 server. Increasing these values may allow you to
5531 improve throughput, but will also increase the
5532 amount of memory reserved for use by the client.
5534 suspend.pm_test_delay=
5536 Sets the number of seconds to remain in a suspend test
5537 mode before resuming the system (see
5538 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5539 is set. Default value is 5.
5542 Format: { on | off | y | n | 1 | 0 }
5543 This parameter controls use of the Protected
5544 Execution Facility on pSeries.
5547 [KNL] Enable accounting of swap in memory resource
5548 controller if no parameter or 1 is given or disable
5549 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5551 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5552 Format: { <int> | force | noforce }
5553 <int> -- Number of I/O TLB slabs
5554 force -- force using of bounce buffers even if they
5555 wouldn't be automatically used by the kernel
5556 noforce -- Never use bounce buffers (for debugging)
5561 Set a sysctl parameter, right before loading the init
5562 process, as if the value was written to the respective
5563 /proc/sys/... file. Both '.' and '/' are recognized as
5564 separators. Unrecognized parameters and invalid values
5565 are reported in the kernel log. Sysctls registered
5566 later by a loaded module cannot be set this way.
5567 Example: sysctl.vm.swappiness=40
5569 sysfs.deprecated=0|1 [KNL]
5570 Enable/disable old style sysfs layout for old udev
5571 on older distributions. When this option is enabled
5572 very new udev will not work anymore. When this option
5573 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5574 in older udev will not work anymore.
5575 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5576 the kernel configuration.
5578 sysrq_always_enabled
5580 Ignore sysrq setting - this boot parameter will
5581 neutralize any effect of /proc/sys/kernel/sysrq.
5582 Useful for debugging.
5584 tcpmhash_entries= [KNL,NET]
5585 Set the number of tcp_metrics_hash slots.
5586 Default value is 8192 or 16384 depending on total
5587 ram pages. This is used to specify the TCP metrics
5588 cache size. See Documentation/networking/ip-sysctl.rst
5589 "tcp_no_metrics_save" section for more details.
5593 test_suspend= [SUSPEND][,N]
5594 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5595 standby suspend) or "freeze" (for suspend type freeze)
5596 as the system sleep state during system startup with
5597 the optional capability to repeat N number of times.
5598 The system is woken from this state using a
5599 wakeup-capable RTC alarm.
5601 thash_entries= [KNL,NET]
5602 Set number of hash buckets for TCP connection
5604 thermal.act= [HW,ACPI]
5605 -1: disable all active trip points in all thermal zones
5606 <degrees C>: override all lowest active trip points
5608 thermal.crt= [HW,ACPI]
5609 -1: disable all critical trip points in all thermal zones
5610 <degrees C>: override all critical trip points
5612 thermal.nocrt= [HW,ACPI]
5613 Set to disable actions on ACPI thermal zone
5614 critical and hot trip points.
5616 thermal.off= [HW,ACPI]
5617 1: disable ACPI thermal control
5619 thermal.psv= [HW,ACPI]
5620 -1: disable all passive trip points
5621 <degrees C>: override all passive trip points to this
5624 thermal.tzp= [HW,ACPI]
5625 Specify global default ACPI thermal zone polling rate
5626 <deci-seconds>: poll all this frequency
5627 0: no polling (default)
5630 Force threading of all interrupt handlers except those
5631 marked explicitly IRQF_NO_THREAD.
5635 Specify if the kernel should make use of the cpu
5636 topology information if the hardware supports this.
5637 The scheduler will make use of this information and
5638 e.g. base its process migration decisions on it.
5641 topology_updates= [KNL, PPC, NUMA]
5643 Specify if the kernel should ignore (off)
5644 topology updates sent by the hypervisor to this
5647 torture.disable_onoff_at_boot= [KNL]
5648 Prevent the CPU-hotplug component of torturing
5649 until after init has spawned.
5651 torture.ftrace_dump_at_shutdown= [KNL]
5652 Dump the ftrace buffer at torture-test shutdown,
5653 even if there were no errors. This can be a
5654 very costly operation when many torture tests
5655 are running concurrently, especially on systems
5656 with rotating-rust storage.
5658 torture.verbose_sleep_frequency= [KNL]
5659 Specifies how many verbose printk()s should be
5660 emitted between each sleep. The default of zero
5661 disables verbose-printk() sleeping.
5663 torture.verbose_sleep_duration= [KNL]
5664 Duration of each verbose-printk() sleep in jiffies.
5668 tpm_suspend_pcr=[HW,TPM]
5669 Format: integer pcr id
5670 Specify that at suspend time, the tpm driver
5671 should extend the specified pcr with zeros,
5672 as a workaround for some chips which fail to
5673 flush the last written pcr on TPM_SaveState.
5674 This will guarantee that all the other pcrs
5677 trace_buf_size=nn[KMG]
5678 [FTRACE] will set tracing buffer size on each cpu.
5680 trace_event=[event-list]
5681 [FTRACE] Set and start specified trace events in order
5682 to facilitate early boot debugging. The event-list is a
5683 comma-separated list of trace events to enable. See
5684 also Documentation/trace/events.rst
5686 trace_options=[option-list]
5687 [FTRACE] Enable or disable tracer options at boot.
5688 The option-list is a comma delimited list of options
5689 that can be enabled or disabled just as if you were
5690 to echo the option name into
5692 /sys/kernel/debug/tracing/trace_options
5694 For example, to enable stacktrace option (to dump the
5695 stack trace of each event), add to the command line:
5697 trace_options=stacktrace
5699 See also Documentation/trace/ftrace.rst "trace options"
5703 Have the tracepoints sent to printk as well as the
5704 tracing ring buffer. This is useful for early boot up
5705 where the system hangs or reboots and does not give the
5706 option for reading the tracing buffer or performing a
5707 ftrace_dump_on_oops.
5709 To turn off having tracepoints sent to printk,
5710 echo 0 > /proc/sys/kernel/tracepoint_printk
5711 Note, echoing 1 into this file without the
5712 tracepoint_printk kernel cmdline option has no effect.
5714 The tp_printk_stop_on_boot (see below) can also be used
5715 to stop the printing of events to console at
5720 Having tracepoints sent to printk() and activating high
5721 frequency tracepoints such as irq or sched, can cause
5722 the system to live lock.
5724 tp_printk_stop_on_boot[FTRACE]
5725 When tp_printk (above) is set, it can cause a lot of noise
5726 on the console. It may be useful to only include the
5727 printing of events during boot up, as user space may
5728 make the system inoperable.
5730 This command line option will stop the printing of events
5731 to console at the late_initcall_sync() time frame.
5734 [FTRACE] enable this option to disable tracing when a
5735 warning is hit. This turns off "tracing_on". Tracing can
5736 be enabled again by echoing '1' into the "tracing_on"
5737 file located in /sys/kernel/debug/tracing/
5739 This option is useful, as it disables the trace before
5740 the WARNING dump is called, which prevents the trace to
5741 be filled with content caused by the warning output.
5743 This option can also be set at run time via the sysctl
5744 option: kernel/traceoff_on_warning
5746 transparent_hugepage=
5748 Format: [always|madvise|never]
5749 Can be used to control the default behavior of the system
5750 with respect to transparent hugepages.
5751 See Documentation/admin-guide/mm/transhuge.rst
5754 trusted.source= [KEYS]
5756 This parameter identifies the trust source as a backend
5757 for trusted keys implementation. Supported trust
5761 If not specified then it defaults to iterating through
5762 the trust source list starting with TPM and assigns the
5763 first trust source as a backend which is initialized
5764 successfully during iteration.
5766 tsc= Disable clocksource stability checks for TSC.
5768 [x86] reliable: mark tsc clocksource as reliable, this
5769 disables clocksource verification at runtime, as well
5770 as the stability checks done at bootup. Used to enable
5771 high-resolution timer mode on older hardware, and in
5772 virtualized environment.
5773 [x86] noirqtime: Do not use TSC to do irq accounting.
5774 Used to run time disable IRQ_TIME_ACCOUNTING on any
5775 platforms where RDTSC is slow and this accounting
5777 [x86] unstable: mark the TSC clocksource as unstable, this
5778 marks the TSC unconditionally unstable at bootup and
5779 avoids any further wobbles once the TSC watchdog notices.
5780 [x86] nowatchdog: disable clocksource watchdog. Used
5781 in situations with strict latency requirements (where
5782 interruptions from clocksource watchdog are not
5785 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5786 value instead. Useful when the early TSC frequency discovery
5787 procedure is not reliable, such as on overclocked systems
5788 with CPUID.16h support and partial CPUID.15h support.
5789 Format: <unsigned int>
5791 tsx= [X86] Control Transactional Synchronization
5792 Extensions (TSX) feature in Intel processors that
5793 support TSX control.
5795 This parameter controls the TSX feature. The options are:
5797 on - Enable TSX on the system. Although there are
5798 mitigations for all known security vulnerabilities,
5799 TSX has been known to be an accelerator for
5800 several previous speculation-related CVEs, and
5801 so there may be unknown security risks associated
5802 with leaving it enabled.
5804 off - Disable TSX on the system. (Note that this
5805 option takes effect only on newer CPUs which are
5806 not vulnerable to MDS, i.e., have
5807 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5808 the new IA32_TSX_CTRL MSR through a microcode
5809 update. This new MSR allows for the reliable
5810 deactivation of the TSX functionality.)
5812 auto - Disable TSX if X86_BUG_TAA is present,
5813 otherwise enable TSX on the system.
5815 Not specifying this option is equivalent to tsx=off.
5817 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5820 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5821 Abort (TAA) vulnerability.
5823 Similar to Micro-architectural Data Sampling (MDS)
5824 certain CPUs that support Transactional
5825 Synchronization Extensions (TSX) are vulnerable to an
5826 exploit against CPU internal buffers which can forward
5827 information to a disclosure gadget under certain
5830 In vulnerable processors, the speculatively forwarded
5831 data can be used in a cache side channel attack, to
5832 access data to which the attacker does not have direct
5835 This parameter controls the TAA mitigation. The
5838 full - Enable TAA mitigation on vulnerable CPUs
5841 full,nosmt - Enable TAA mitigation and disable SMT on
5842 vulnerable CPUs. If TSX is disabled, SMT
5843 is not disabled because CPU is not
5844 vulnerable to cross-thread TAA attacks.
5845 off - Unconditionally disable TAA mitigation
5847 On MDS-affected machines, tsx_async_abort=off can be
5848 prevented by an active MDS mitigation as both vulnerabilities
5849 are mitigated with the same mechanism so in order to disable
5850 this mitigation, you need to specify mds=off too.
5852 Not specifying this option is equivalent to
5853 tsx_async_abort=full. On CPUs which are MDS affected
5854 and deploy MDS mitigation, TAA mitigation is not
5855 required and doesn't provide any additional
5859 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5861 turbografx.map[2|3]= [HW,JOY]
5862 TurboGraFX parallel port interface
5864 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5865 See also Documentation/input/devices/joystick-parport.rst
5867 udbg-immortal [PPC] When debugging early kernel crashes that
5868 happen after console_init() and before a proper
5869 console driver takes over, this boot options might
5870 help "seeing" what's going on.
5872 uhash_entries= [KNL,NET]
5873 Set number of hash buckets for UDP/UDP-Lite connections
5876 [USB] Ignore overcurrent events (default N).
5877 Some badly-designed motherboards generate lots of
5878 bogus events, for ports that aren't wired to
5879 anything. Set this parameter to avoid log spamming.
5880 Note that genuine overcurrent events won't be
5884 [X86] Cause panic on unknown NMI.
5886 usbcore.authorized_default=
5887 [USB] Default USB device authorization:
5888 (default -1 = authorized except for wireless USB,
5889 0 = not authorized, 1 = authorized, 2 = authorized
5890 if device connected to internal port)
5892 usbcore.autosuspend=
5893 [USB] The autosuspend time delay (in seconds) used
5894 for newly-detected USB devices (default 2). This
5895 is the time required before an idle device will be
5896 autosuspended. Devices for which the delay is set
5897 to a negative value won't be autosuspended at all.
5899 usbcore.usbfs_snoop=
5900 [USB] Set to log all usbfs traffic (default 0 = off).
5902 usbcore.usbfs_snoop_max=
5903 [USB] Maximum number of bytes to snoop in each URB
5906 usbcore.blinkenlights=
5907 [USB] Set to cycle leds on hubs (default 0 = off).
5909 usbcore.old_scheme_first=
5910 [USB] Start with the old device initialization
5911 scheme (default 0 = off).
5913 usbcore.usbfs_memory_mb=
5914 [USB] Memory limit (in MB) for buffers allocated by
5915 usbfs (default = 16, 0 = max = 2047).
5917 usbcore.use_both_schemes=
5918 [USB] Try the other device initialization scheme
5919 if the first one fails (default 1 = enabled).
5921 usbcore.initial_descriptor_timeout=
5922 [USB] Specifies timeout for the initial 64-byte
5923 USB_REQ_GET_DESCRIPTOR request in milliseconds
5924 (default 5000 = 5.0 seconds).
5926 usbcore.nousb [USB] Disable the USB subsystem
5929 [USB] A list of quirk entries to augment the built-in
5930 usb core quirk list. List entries are separated by
5931 commas. Each entry has the form
5932 VendorID:ProductID:Flags. The IDs are 4-digit hex
5933 numbers and Flags is a set of letters. Each letter
5934 will change the built-in quirk; setting it if it is
5935 clear and clearing it if it is set. The letters have
5936 the following meanings:
5937 a = USB_QUIRK_STRING_FETCH_255 (string
5938 descriptors must not be fetched using
5940 b = USB_QUIRK_RESET_RESUME (device can't resume
5941 correctly so reset it instead);
5942 c = USB_QUIRK_NO_SET_INTF (device can't handle
5943 Set-Interface requests);
5944 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5945 handle its Configuration or Interface
5947 e = USB_QUIRK_RESET (device can't be reset
5948 (e.g morph devices), don't use reset);
5949 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5950 more interface descriptions than the
5951 bNumInterfaces count, and can't handle
5952 talking to these interfaces);
5953 g = USB_QUIRK_DELAY_INIT (device needs a pause
5954 during initialization, after we read
5955 the device descriptor);
5956 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5957 high speed and super speed interrupt
5958 endpoints, the USB 2.0 and USB 3.0 spec
5959 require the interval in microframes (1
5960 microframe = 125 microseconds) to be
5961 calculated as interval = 2 ^
5963 Devices with this quirk report their
5964 bInterval as the result of this
5965 calculation instead of the exponent
5966 variable used in the calculation);
5967 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5968 handle device_qualifier descriptor
5970 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5971 generates spurious wakeup, ignore
5972 remote wakeup capability);
5973 k = USB_QUIRK_NO_LPM (device can't handle Link
5975 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5976 (Device reports its bInterval as linear
5977 frames instead of the USB 2.0
5979 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5980 to be disconnected before suspend to
5981 prevent spurious wakeup);
5982 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5983 pause after every control message);
5984 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5985 delay after resetting its port);
5986 Example: quirks=0781:5580:bk,0a5c:5834:gij
5989 [USBHID] The interval which mice are to be polled at.
5992 [USBHID] The interval which joysticks are to be polled at.
5995 [USBHID] The interval which keyboards are to be polled at.
5997 usb-storage.delay_use=
5998 [UMS] The delay in seconds before a new device is
5999 scanned for Logical Units (default 1).
6002 [UMS] A list of quirks entries to supplement or
6003 override the built-in unusual_devs list. List
6004 entries are separated by commas. Each entry has
6005 the form VID:PID:Flags where VID and PID are Vendor
6006 and Product ID values (4-digit hex numbers) and
6007 Flags is a set of characters, each corresponding
6008 to a common usb-storage quirk flag as follows:
6009 a = SANE_SENSE (collect more than 18 bytes
6010 of sense data, not on uas);
6011 b = BAD_SENSE (don't collect more than 18
6012 bytes of sense data, not on uas);
6013 c = FIX_CAPACITY (decrease the reported
6014 device capacity by one sector);
6015 d = NO_READ_DISC_INFO (don't use
6016 READ_DISC_INFO command, not on uas);
6017 e = NO_READ_CAPACITY_16 (don't use
6018 READ_CAPACITY_16 command);
6019 f = NO_REPORT_OPCODES (don't use report opcodes
6021 g = MAX_SECTORS_240 (don't transfer more than
6022 240 sectors at a time, uas only);
6023 h = CAPACITY_HEURISTICS (decrease the
6024 reported device capacity by one
6025 sector if the number is odd);
6026 i = IGNORE_DEVICE (don't bind to this
6028 j = NO_REPORT_LUNS (don't use report luns
6030 k = NO_SAME (do not use WRITE_SAME, uas only)
6031 l = NOT_LOCKABLE (don't try to lock and
6032 unlock ejectable media, not on uas);
6033 m = MAX_SECTORS_64 (don't transfer more
6034 than 64 sectors = 32 KB at a time,
6036 n = INITIAL_READ10 (force a retry of the
6037 initial READ(10) command, not on uas);
6038 o = CAPACITY_OK (accept the capacity
6039 reported by the device, not on uas);
6040 p = WRITE_CACHE (the device cache is ON
6041 by default, not on uas);
6042 r = IGNORE_RESIDUE (the device reports
6043 bogus residue values, not on uas);
6044 s = SINGLE_LUN (the device has only one
6046 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6047 commands, uas only);
6048 u = IGNORE_UAS (don't bind to the uas driver);
6049 w = NO_WP_DETECT (don't test whether the
6050 medium is write-protected).
6051 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6052 even if the device claims no cache,
6054 Example: quirks=0419:aaf5:rl,0421:0433:rc
6056 user_debug= [KNL,ARM]
6058 See arch/arm/Kconfig.debug help text.
6059 1 - undefined instruction events
6061 4 - invalid data aborts
6064 Example: user_debug=31
6067 [X86] Flags controlling user PTE allocations.
6069 nohigh = do not allocate PTE pages in
6070 HIGHMEM regardless of setting
6074 On X86_32, this is an alias for vdso32=. Otherwise:
6076 vdso=1: enable VDSO (the default)
6077 vdso=0: disable VDSO mapping
6079 vdso32= [X86] Control the 32-bit vDSO
6080 vdso32=1: enable 32-bit VDSO
6081 vdso32=0 or vdso32=2: disable 32-bit VDSO
6083 See the help text for CONFIG_COMPAT_VDSO for more
6084 details. If CONFIG_COMPAT_VDSO is set, the default is
6085 vdso32=0; otherwise, the default is vdso32=1.
6087 For compatibility with older kernels, vdso32=2 is an
6090 Try vdso32=0 if you encounter an error that says:
6091 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6094 vector=percpu: enable percpu vector domain
6096 video= [FB] Frame buffer configuration
6097 See Documentation/fb/modedb.rst.
6099 video.brightness_switch_enabled= [0,1]
6100 If set to 1, on receiving an ACPI notify event
6101 generated by hotkey, video driver will adjust brightness
6102 level and then send out the event to user space through
6103 the allocated input device; If set to 0, video driver
6104 will only send out the event without touching backlight
6109 [VMMIO] Memory mapped virtio (platform) device.
6111 <size>@<baseaddr>:<irq>[:<id>]
6113 <size> := size (can use standard suffixes
6115 <baseaddr> := physical base address
6116 <irq> := interrupt number (as passed to
6118 <id> := (optional) platform device id
6120 virtio_mmio.device=1K@0x100b0000:48:7
6122 Can be used multiple times for multiple devices.
6124 vga= [BOOT,X86-32] Select a particular video mode
6125 See Documentation/x86/boot.rst and
6126 Documentation/admin-guide/svga.rst.
6127 Use vga=ask for menu.
6128 This is actually a boot loader parameter; the value is
6129 passed to the kernel using a special protocol.
6131 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6132 May slow down system boot speed, especially when
6133 enabled on systems with a large amount of memory.
6134 All options are enabled by default, and this
6135 interface is meant to allow for selectively
6136 enabling or disabling specific virtual memory
6139 Available options are:
6140 P Enable page structure init time poisoning
6141 - Disable all of the above options
6143 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6144 size of <nn>. This can be used to increase the
6145 minimum size (128MB on x86). It can also be used to
6146 decrease the size and leave more room for directly
6149 vmcp_cma=nn[MG] [KNL,S390]
6150 Sets the memory size reserved for contiguous memory
6151 allocations for the vmcp device driver.
6153 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6156 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6159 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6163 Controls the behavior of vsyscalls (i.e. calls to
6164 fixed addresses of 0xffffffffff600x00 from legacy
6165 code). Most statically-linked binaries and older
6166 versions of glibc use these calls. Because these
6167 functions are at fixed addresses, they make nice
6168 targets for exploits that can control RIP.
6170 emulate [default] Vsyscalls turn into traps and are
6171 emulated reasonably safely. The vsyscall
6174 xonly Vsyscalls turn into traps and are
6175 emulated reasonably safely. The vsyscall
6176 page is not readable.
6178 none Vsyscalls don't work at all. This makes
6179 them quite hard to use for exploits but
6180 might break your system.
6182 vt.color= [VT] Default text color.
6183 Format: 0xYX, X = foreground, Y = background.
6184 Default: 0x07 = light gray on black.
6186 vt.cur_default= [VT] Default cursor shape.
6187 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6188 the parameters of the <Esc>[?A;B;Cc escape sequence;
6189 see VGA-softcursor.txt. Default: 2 = underline.
6191 vt.default_blu= [VT]
6192 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6193 Change the default blue palette of the console.
6194 This is a 16-member array composed of values
6197 vt.default_grn= [VT]
6198 Format: <green0>,<green1>,<green2>,...,<green15>
6199 Change the default green palette of the console.
6200 This is a 16-member array composed of values
6203 vt.default_red= [VT]
6204 Format: <red0>,<red1>,<red2>,...,<red15>
6205 Change the default red palette of the console.
6206 This is a 16-member array composed of values
6212 Set system-wide default UTF-8 mode for all tty's.
6213 Default is 1, i.e. UTF-8 mode is enabled for all
6214 newly opened terminals.
6216 vt.global_cursor_default=
6219 Set system-wide default for whether a cursor
6220 is shown on new VTs. Default is -1,
6221 i.e. cursors will be created by default unless
6222 overridden by individual drivers. 0 will hide
6223 cursors, 1 will display them.
6225 vt.italic= [VT] Default color for italic text; 0-15.
6228 vt.underline= [VT] Default color for underlined text; 0-15.
6231 watchdog timers [HW,WDT] For information on watchdog timers,
6232 see Documentation/watchdog/watchdog-parameters.rst
6233 or other driver-specific files in the
6234 Documentation/watchdog/ directory.
6238 Set the hard lockup detector stall duration
6239 threshold in seconds. The soft lockup detector
6240 threshold is set to twice the value. A value of 0
6241 disables both lockup detectors. Default is 10
6244 workqueue.watchdog_thresh=
6245 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6246 warn stall conditions and dump internal state to
6247 help debugging. 0 disables workqueue stall
6248 detection; otherwise, it's the stall threshold
6249 duration in seconds. The default value is 30 and
6250 it can be updated at runtime by writing to the
6251 corresponding sysfs file.
6253 workqueue.disable_numa
6254 By default, all work items queued to unbound
6255 workqueues are affine to the NUMA nodes they're
6256 issued on, which results in better behavior in
6257 general. If NUMA affinity needs to be disabled for
6258 whatever reason, this option can be used. Note
6259 that this also can be controlled per-workqueue for
6260 workqueues visible under /sys/bus/workqueue/.
6262 workqueue.power_efficient
6263 Per-cpu workqueues are generally preferred because
6264 they show better performance thanks to cache
6265 locality; unfortunately, per-cpu workqueues tend to
6266 be more power hungry than unbound workqueues.
6268 Enabling this makes the per-cpu workqueues which
6269 were observed to contribute significantly to power
6270 consumption unbound, leading to measurably lower
6271 power usage at the cost of small performance
6274 The default value of this parameter is determined by
6275 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6277 workqueue.debug_force_rr_cpu
6278 Workqueue used to implicitly guarantee that work
6279 items queued without explicit CPU specified are put
6280 on the local CPU. This guarantee is no longer true
6281 and while local CPU is still preferred work items
6282 may be put on foreign CPUs. This debug option
6283 forces round-robin CPU selection to flush out
6284 usages which depend on the now broken guarantee.
6285 When enabled, memory and cache locality will be
6288 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6289 default x2apic cluster mode on platforms
6292 xen_512gb_limit [KNL,X86-64,XEN]
6293 Restricts the kernel running paravirtualized under Xen
6294 to use only up to 512 GB of RAM. The reason to do so is
6295 crash analysis tools and Xen tools for doing domain
6296 save/restore/migration must be enabled to handle larger
6299 xen_emul_unplug= [HW,X86,XEN]
6300 Unplug Xen emulated devices
6301 Format: [unplug0,][unplug1]
6302 ide-disks -- unplug primary master IDE devices
6303 aux-ide-disks -- unplug non-primary-master IDE devices
6304 nics -- unplug network devices
6305 all -- unplug all emulated devices (NICs and IDE disks)
6306 unnecessary -- unplugging emulated devices is
6307 unnecessary even if the host did not respond to
6309 never -- do not unplug even if version check succeeds
6311 xen_legacy_crash [X86,XEN]
6312 Crash from Xen panic notifier, without executing late
6313 panic() code such as dumping handler.
6315 xen_nopvspin [X86,XEN]
6316 Disables the qspinlock slowpath using Xen PV optimizations.
6317 This parameter is obsoleted by "nopvspin" parameter, which
6318 has equivalent effect for XEN platform.
6321 Disables the PV optimizations forcing the HVM guest to
6322 run as generic HVM guest with no PV drivers.
6323 This option is obsoleted by the "nopv" option, which
6324 has equivalent effect for XEN platform.
6326 xen_no_vector_callback
6327 [KNL,X86,XEN] Disable the vector callback for Xen
6328 event channel interrupts.
6330 xen_scrub_pages= [XEN]
6331 Boolean option to control scrubbing pages before giving them back
6332 to Xen, for use by other domains. Can be also changed at runtime
6333 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6334 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6336 xen_timer_slop= [X86-64,XEN]
6337 Set the timer slop (in nanoseconds) for the virtual Xen
6338 timers (default is 100000). This adjusts the minimum
6339 delta of virtualized Xen timers, where lower values
6340 improve timer resolution at the expense of processing
6341 more timer interrupts.
6343 xen.event_eoi_delay= [XEN]
6344 How long to delay EOI handling in case of event
6345 storms (jiffies). Default is 10.
6347 xen.event_loop_timeout= [XEN]
6348 After which time (jiffies) the event handling loop
6349 should start to delay EOI handling. Default is 2.
6351 xen.fifo_events= [XEN]
6352 Boolean parameter to disable using fifo event handling
6353 even if available. Normally fifo event handling is
6354 preferred over the 2-level event handling, as it is
6355 fairer and the number of possible event channels is
6356 much higher. Default is on (use fifo events).
6358 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6359 Disables the PV optimizations forcing the guest to run
6360 as generic guest with no PV drivers. Currently support
6361 XEN HVM, KVM, HYPER_V and VMWARE guest.
6363 nopvspin [X86,XEN,KVM]
6364 Disables the qspinlock slow path using PV optimizations
6365 which allow the hypervisor to 'idle' the guest on lock
6368 xirc2ps_cs= [NET,PCMCIA]
6370 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6373 By default on POWER9 and above, the kernel will
6374 natively use the XIVE interrupt controller. This option
6375 allows the fallback firmware mode to be used:
6377 off Fallback to firmware control of XIVE interrupt
6378 controller on both pseries and powernv
6379 platforms. Only useful on POWER9 and above.
6381 xhci-hcd.quirks [USB,KNL]
6382 A hex value specifying bitmask with supplemental xhci
6383 host controller quirks. Meaning of each bit can be
6384 consulted in header drivers/usb/host/xhci.h.
6387 Format: { early | on | rw | ro | off }
6388 Controls if xmon debugger is enabled. Default is off.
6389 Passing only "xmon" is equivalent to "xmon=early".
6390 early Call xmon as early as possible on boot; xmon
6391 debugger is called from setup_arch().
6392 on xmon debugger hooks will be installed so xmon
6393 is only called on a kernel crash. Default mode,
6394 i.e. either "ro" or "rw" mode, is controlled
6395 with CONFIG_XMON_DEFAULT_RO_MODE.
6396 rw xmon debugger hooks will be installed so xmon
6397 is called only on a kernel crash, mode is write,
6398 meaning SPR registers, memory and, other data
6399 can be written using xmon commands.
6400 ro same as "rw" option above but SPR registers,
6401 memory, and other data can't be written using
6403 off xmon is disabled.
projects / linux-2.6-microblaze.git / blob