1 acpi= [HW,ACPI,X86,ARM64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
5 force -- enable ACPI if default was off
6 on -- enable ACPI but allow fallback to DT [arm64]
7 off -- disable ACPI if default was on
8 noirq -- do not use ACPI for IRQ routing
9 strict -- Be less tolerant of platforms that are not
10 strictly ACPI specification compliant.
11 rsdt -- prefer RSDT over (default) XSDT
12 copy_dsdt -- copy DSDT to memory
13 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force"
16 See also Documentation/power/runtime_pm.rst, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
25 { vendor | video | native | none }
26 If set to vendor, prefer vendor-specific driver
27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
28 of the ACPI video.ko driver.
29 If set to video, use the ACPI video.ko driver.
30 If set to native, use the device's native backlight mode.
31 If set to none, disable the ACPI backlight interface.
33 acpi_force_32bit_fadt_addr
34 force FADT to use 32 bit addresses rather than the
35 64 bit X_* addresses. Some firmware have broken 64
36 bit addresses for force ACPI ignore these and use
37 the older legacy 32 bit addresses.
39 acpica_no_return_repair [HW, ACPI]
40 Disable AML predefined validation mechanism
41 This mechanism can repair the evaluation result to make
42 the return objects more ACPI specification compliant.
43 This option is useful for developers to identify the
44 root cause of an AML interpreter issue when the issue
45 has something to do with the repair mechanism.
47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
51 debug output. Bits in debug_layer correspond to a
52 _COMPONENT in an ACPI source file, e.g.,
53 #define _COMPONENT ACPI_EVENTS
54 Bits in debug_level correspond to a level in
55 ACPI_DEBUG_PRINT statements, e.g.,
56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
57 The debug_level mask defaults to "info". See
58 Documentation/firmware-guide/acpi/debug.rst for more information about
59 debug layers and levels.
61 Enable processor driver info messages:
62 acpi.debug_layer=0x20000000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
116 Format: <byte> or <bitmap-list>
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
163 acpi_osi=!! # enable all built-in OS vendor
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig,
229 s4_nohwsig, old_ordering, nonvs,
230 sci_force_enable, nobl }
231 See Documentation/power/video.rst for information on
233 s3_beep is for debugging; it makes the PC's speaker beep
234 as soon as the kernel's real-mode entry point is called.
235 s4_hwsig causes the kernel to check the ACPI hardware
236 signature during resume from hibernation, and gracefully
237 refuse to resume if it has changed. This complies with
238 the ACPI specification but not with reality, since
239 Windows does not do this and many laptops do change it
240 on docking. So the default behaviour is to allow resume
241 and simply warn when the signature changes, unless the
242 s4_hwsig option is enabled.
243 s4_nohwsig prevents ACPI hardware signature from being
244 used (or even warned about) during resume.
245 old_ordering causes the ACPI 1.0 ordering of the _PTS
246 control method, with respect to putting devices into
247 low power states, to be enforced (the ACPI 2.0 ordering
248 of _PTS is used by default).
249 nonvs prevents the kernel from saving/restoring the
250 ACPI NVS memory during suspend/hibernation and resume.
251 sci_force_enable causes the kernel to set SCI_EN directly
252 on resume from S1/S3 (which is against the ACPI spec,
253 but some broken systems don't work without it).
254 nobl causes the internal blacklist of systems known to
255 behave incorrectly in some ways with respect to system
256 suspend and resume to be ignored (use wisely).
258 acpi_use_timer_override [HW,ACPI]
259 Use timer override. For some broken Nvidia NF5 boards
260 that require a timer override, but don't have HPET
262 add_efi_memmap [EFI; X86] Include EFI memory map in
263 kernel's map of available physical RAM.
266 { off | try_unsupported }
267 off: disable AGP support
268 try_unsupported: try to drive unsupported chipsets
269 (may crash computer or cause data corruption)
272 See Documentation/sound/alsa-configuration.rst
275 Allow the default userspace alignment fault handler
276 behaviour to be specified. Bit 0 enables warnings,
277 bit 1 enables fixups, and bit 2 sends a segfault.
279 align_va_addr= [X86-64]
280 Align virtual addresses by clearing slice [14:12] when
281 allocating a VMA at process creation time. This option
282 gives you up to 3% performance improvement on AMD F15h
283 machines (where it is enabled by default) for a
284 CPU-intensive style benchmark, and it can vary highly in
285 a microbenchmark depending on workload and compiler.
287 32: only for 32-bit processes
288 64: only for 64-bit processes
289 on: enable for both 32- and 64-bit processes
290 off: disable for both 32- and 64-bit processes
292 alloc_snapshot [FTRACE]
293 Allocate the ftrace snapshot buffer on boot up when the
294 main buffer is allocated. This is handy if debugging
295 and you need to use tracing_snapshot() on boot up, and
296 do not want to use tracing_snapshot_alloc() as it needs
297 to be done where GFP_KERNEL allocations are allowed.
299 allow_mismatched_32bit_el0 [ARM64]
300 Allow execve() of 32-bit applications and setting of the
301 PER_LINUX32 personality on systems where only a strict
302 subset of the CPUs support 32-bit EL0. When this
303 parameter is present, the set of CPUs supporting 32-bit
304 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
305 and hot-unplug operations may be restricted.
307 See Documentation/arm64/asymmetric-32bit.rst for more
310 amd_iommu= [HW,X86-64]
311 Pass parameters to the AMD IOMMU driver in the system.
313 fullflush - Deprecated, equivalent to iommu.strict=1
314 off - do not initialize any AMD IOMMU found in
316 force_isolation - Force device isolation for all
317 devices. The IOMMU driver is not
318 allowed anymore to lift isolation
319 requirements as needed. This option
320 does not override iommu=pt
321 force_enable - Force enable the IOMMU on platforms known
322 to be buggy with IOMMU enabled. Use this
325 amd_iommu_dump= [HW,X86-64]
326 Enable AMD IOMMU driver option to dump the ACPI table
327 for AMD IOMMU. With this option enabled, AMD IOMMU
328 driver will print ACPI tables for AMD IOMMU during
329 IOMMU initialization.
331 amd_iommu_intr= [HW,X86-64]
332 Specifies one of the following AMD IOMMU interrupt
334 legacy - Use legacy interrupt remapping mode.
335 vapic - Use virtual APIC mode, which allows IOMMU
336 to inject interrupts directly into guest.
337 This mode requires kvm-amd.avic=1.
338 (Default when IOMMU HW support is present.)
340 amijoy.map= [HW,JOY] Amiga joystick support
341 Map of devices attached to JOY0DAT and JOY1DAT
343 See also Documentation/input/joydev/joystick.rst
345 analog.map= [HW,JOY] Analog joystick and gamepad support
346 Specifies type or capabilities of an analog joystick
347 connected to one of 16 gameports
348 Format: <type1>,<type2>,..<type16>
351 Power management functions (SPARCstation-4/5 + deriv.)
353 Disable APC CPU standby support. SPARCstation-Fox does
354 not play well with APC CPU idle - disable it if you have
355 APC and your system crashes randomly.
357 apic= [APIC,X86] Advanced Programmable Interrupt Controller
358 Change the output verbosity while booting
359 Format: { quiet (default) | verbose | debug }
360 Change the amount of debugging information output
361 when initialising the APIC and IO-APIC components.
362 For X86-32, this can also be used to specify an APIC
364 Format: apic=driver_name
365 Examples: apic=bigsmp
367 apic_extnmi= [APIC,X86] External NMI delivery setting
368 Format: { bsp (default) | all | none }
369 bsp: External NMI is delivered only to CPU 0
370 all: External NMIs are broadcast to all CPUs as a
372 none: External NMI is masked for all CPUs. This is
373 useful so that a dump capture kernel won't be
377 See Documentation/networking/ipv6.rst.
379 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
380 Limit apic dumping. The parameter defines the maximal
381 number of local apics being dumped. Also it is possible
382 to set it to "all" by meaning -- no limit here.
383 Format: { 1 (default) | 2 | ... | all }.
384 The parameter valid if only apic=debug or
385 apic=verbose is specified.
386 Example: apic=debug show_lapic=all
388 apm= [APM] Advanced Power Management
389 See header of arch/x86/kernel/apm_32.c.
391 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
392 Format: <io>,<irq>,<nodeID>
394 arm64.nobti [ARM64] Unconditionally disable Branch Target
395 Identification support
397 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
400 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
403 arm64.nosve [ARM64] Unconditionally disable Scalable Vector
406 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix
411 atarimouse= [HW,MOUSE] Atari Mouse
413 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
414 EzKey and similar keyboards
416 atkbd.reset= [HW] Reset keyboard during initialization
418 atkbd.set= [HW] Select keyboard code set
419 Format: <int> (2 = AT (default), 3 = PS/2)
421 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
424 atkbd.softraw= [HW] Choose between synthetic and real raw mode
425 Format: <bool> (0 = real, 1 = synthetic (default))
427 atkbd.softrepeat= [HW]
428 Use software keyboard repeat
430 audit= [KNL] Enable the audit sub-system
431 Format: { "0" | "1" | "off" | "on" }
432 0 | off - kernel audit is disabled and can not be
433 enabled until the next reboot
434 unset - kernel audit is initialized but disabled and
435 will be fully enabled by the userspace auditd.
436 1 | on - kernel audit is initialized and partially
437 enabled, storing at most audit_backlog_limit
438 messages in RAM until it is fully enabled by the
442 audit_backlog_limit= [KNL] Set the audit queue size limit.
443 Format: <int> (must be >=0)
446 bau= [X86_UV] Enable the BAU on SGI UV. The default
447 behavior is to disable the BAU (i.e. bau=0).
448 Format: { "0" | "1" }
451 unset - Disable the BAU.
453 baycom_epp= [HW,AX25]
456 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
458 See header of drivers/net/hamradio/baycom_par.c.
460 baycom_ser_fdx= [HW,AX25]
461 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
462 Format: <io>,<irq>,<mode>[,<baud>]
463 See header of drivers/net/hamradio/baycom_ser_fdx.c.
465 baycom_ser_hdx= [HW,AX25]
466 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
467 Format: <io>,<irq>,<mode>
468 See header of drivers/net/hamradio/baycom_ser_hdx.c.
471 Disable BERT OS support on buggy BIOSes.
473 bgrt_disable [ACPI][X86]
474 Disable BGRT to avoid flickering OEM logo.
476 blkdevparts= Manual partition parsing of block device(s) for
477 embedded devices based on command line input.
478 See Documentation/block/cmdline-partition.rst
480 boot_delay= Milliseconds to delay each printk during boot.
481 Values larger than 10 seconds (10000) are changed to
486 Extended command line options can be added to an initrd
487 and this will cause the kernel to look for it.
489 See Documentation/admin-guide/bootconfig.rst
491 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
492 bttv.radio= Most important insmod options are available as
494 bttv.pll= See Documentation/admin-guide/media/bttv.rst
497 bulk_remove=off [PPC] This parameter disables the use of the pSeries
498 firmware feature for flushing multiple hpte entries
501 c101= [NET] Moxa C101 synchronous serial card
503 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
504 Sometimes CPU hardware bugs make them report the cache
505 size incorrectly. The kernel will attempt work arounds
506 to fix known problems, but for some CPUs it is not
507 possible to determine what the correct size should be.
508 This option provides an override for these situations.
511 [NET] Specifies amount of time (in seconds) that
512 the kernel should wait for a network carrier. By default
513 it waits 120 seconds.
515 ca_keys= [KEYS] This parameter identifies a specific key(s) on
516 the system trusted keyring to be used for certificate
518 format: { id:<keyid> | builtin }
520 cca= [MIPS] Override the kernel pages' cache coherency
521 algorithm. Accepted values range from 0 to 7
522 inclusive. See arch/mips/include/asm/pgtable-bits.h
523 for platform specific values (SB1, Loongson3 and
526 ccw_timeout_log [S390]
527 See Documentation/s390/common_io.rst for details.
529 cgroup_disable= [KNL] Disable a particular controller or optional feature
530 Format: {name of the controller(s) or feature(s) to disable}
531 The effects of cgroup_disable=foo are:
532 - foo isn't auto-mounted if you mount all cgroups in
534 - foo isn't visible as an individually mountable
536 - if foo is an optional feature then the feature is
537 disabled and corresponding cgroup files are not
539 {Currently only "memory" controller deal with this and
540 cut the overhead, others just disable the usage. So
541 only cgroup_disable=memory is actually worthy}
542 Specifying "pressure" disables per-cgroup pressure
543 stall information accounting feature
545 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
546 Format: { { controller | "all" | "named" }
547 [,{ controller | "all" | "named" }...] }
548 Like cgroup_disable, but only applies to cgroup v1;
549 the blacklisted controllers remain available in cgroup2.
550 "all" blacklists all controllers and "named" disables
551 named mounts. Specifying both "all" and "named" disables
554 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
556 nosocket -- Disable socket memory accounting.
557 nokmem -- Disable kernel memory accounting.
559 checkreqprot= [SELINUX] Set initial checkreqprot flag value.
560 Format: { "0" | "1" }
561 See security/selinux/Kconfig help text.
562 0 -- check protection applied by kernel (includes
563 any implied execute protection).
564 1 -- check protection requested by application.
565 Default value is set via a kernel config option.
566 Value can be changed at runtime via
567 /sys/fs/selinux/checkreqprot.
568 Setting checkreqprot to 1 is deprecated.
571 See Documentation/s390/common_io.rst for details.
573 clearcpuid=X[,X...] [X86]
574 Disable CPUID feature X for the kernel. See
575 arch/x86/include/asm/cpufeatures.h for the valid bit
576 numbers X. Note the Linux-specific bits are not necessarily
577 stable over kernel options, but the vendor-specific
579 X can also be a string as appearing in the flags: line
580 in /proc/cpuinfo which does not have the above
581 instability issue. However, not all features have names
583 Note that using this option will taint your kernel.
584 Also note that user programs calling CPUID directly
585 or using the feature without checking anything
586 will still see it. This just prevents it from
587 being used by the kernel or shown in /proc/cpuinfo.
588 Also note the kernel might malfunction if you disable
593 Prevents the clock framework from automatically gating
594 clocks that have not been explicitly enabled by a Linux
595 device driver but are enabled in hardware at reset or
596 by the bootloader/firmware. Note that this does not
597 force such clocks to be always-on nor does it reserve
598 those clocks in any way. This parameter is useful for
599 debug and development, but should not be needed on a
600 platform with proper driver support. For more
601 information, see Documentation/driver-api/clk.rst.
603 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
605 Forces specified clocksource (if available) to be used
606 when calculating gettimeofday(). If specified
607 clocksource is not available, it defaults to PIT.
608 Format: { pit | tsc | cyclone | pmtmr }
610 clocksource= Override the default clocksource
612 Override the default clocksource and use the clocksource
613 with the name specified.
614 Some clocksource names to choose from, depending on
616 [all] jiffies (this is the base, fallback clocksource)
618 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
619 pxa_timer,timer3,32k_counter,timer0_1
620 [X86-32] pit,hpet,tsc;
621 scx200_hrt on Geode; cyclone on IBM x440
629 clocksource.arm_arch_timer.evtstrm=
632 Enable/disable the eventstream feature of the ARM
633 architected timer so that code using WFE-based polling
634 loops can be debugged more effectively on production
637 clocksource.max_cswd_read_retries= [KNL]
638 Number of clocksource_watchdog() retries due to
639 external delays before the clock will be marked
640 unstable. Defaults to two retries, that is,
641 three attempts to read the clock under test.
643 clocksource.verify_n_cpus= [KNL]
644 Limit the number of CPUs checked for clocksources
645 marked with CLOCK_SOURCE_VERIFY_PERCPU that
646 are marked unstable due to excessive skew.
647 A negative value says to check all CPUs, while
648 zero says not to check any. Values larger than
649 nr_cpu_ids are silently truncated to nr_cpu_ids.
650 The actual CPUs are chosen randomly, with
651 no replacement if the same CPU is chosen twice.
653 clocksource-wdtest.holdoff= [KNL]
654 Set the time in seconds that the clocksource
655 watchdog test waits before commencing its tests.
656 Defaults to zero when built as a module and to
657 10 seconds when built into the kernel.
659 cma=nn[MG]@[start[MG][-end[MG]]]
661 Sets the size of kernel global memory area for
662 contiguous memory allocations and optionally the
663 placement constraint by the physical address range of
664 memory allocations. A value of 0 disables CMA
665 altogether. For more information, see
666 kernel/dma/contiguous.c
670 Sets the size of kernel per-numa memory area for
671 contiguous memory allocations. A value of 0 disables
672 per-numa CMA altogether. And If this option is not
673 specificed, the default value is 0.
674 With per-numa CMA enabled, DMA users on node nid will
675 first try to allocate buffer from the pernuma area
676 which is located in node nid, if the allocation fails,
677 they will fallback to the global default memory area.
679 cmo_free_hint= [PPC] Format: { yes | no }
680 Specify whether pages are marked as being inactive
681 when they are freed. This is used in CMO environments
682 to determine OS memory pressure for page stealing by
686 coherent_pool=nn[KMG] [ARM,KNL]
687 Sets the size of memory pool for coherent, atomic dma
688 allocations, by default set to 256K.
690 com20020= [HW,NET] ARCnet - COM20020 chipset
692 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
694 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
698 ARCnet - COM90xx chipset (memory-mapped buffers)
699 Format: <io>[,<irq>[,<memstart>]]
701 condev= [HW,S390] console device
704 console= [KNL] Output console device and options.
706 tty<n> Use the virtual console device <n>.
710 Use the specified serial port. The options are of
711 the form "bbbbpnf", where "bbbb" is the baud rate,
712 "p" is parity ("n", "o", or "e"), "n" is number of
713 bits, and "f" is flow control ("r" for RTS or
714 omit it). Default is "9600n8".
716 See Documentation/admin-guide/serial-console.rst for more
718 Documentation/networking/netconsole.rst for an
721 uart[8250],io,<addr>[,options]
722 uart[8250],mmio,<addr>[,options]
723 uart[8250],mmio16,<addr>[,options]
724 uart[8250],mmio32,<addr>[,options]
725 uart[8250],0x<addr>[,options]
726 Start an early, polled-mode console on the 8250/16550
727 UART at the specified I/O port or MMIO address,
728 switching to the matching ttyS device later.
729 MMIO inter-register address stride is either 8-bit
730 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
731 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
732 to be equivalent to 'mmio'. 'options' are specified in
733 the same format described for ttyS above; if unspecified,
734 the h/w is not re-initialized.
736 hvc<n> Use the hypervisor console device <n>. This is for
737 both Xen and PowerPC hypervisors.
740 Use to disable console output, i.e., to have kernel
741 console messages discarded.
742 This must be the only console= parameter used on the
745 If the device connected to the port is not a TTY but a braille
746 device, prepend "brl," before the device type, for instance
748 For now, only VisioBraille is supported.
751 [KNL] Change console messages format
753 By default we print messages on consoles in
754 "[time stamp] text\n" format (time stamp may not be
755 printed, depending on CONFIG_PRINTK_TIME or
756 `printk_time' param).
758 Switch to syslog format: "<%u>[time stamp] text\n"
759 IOW, each message will have a facility and loglevel
760 prefix. The format is similar to one used by syslog()
761 syscall, or to executing "dmesg -S --raw" or to reading
764 consoleblank= [KNL] The console blank (screen saver) timeout in
765 seconds. A value of 0 disables the blank timer.
769 [KNL] Change the default value for
770 /proc/<pid>/coredump_filter.
771 See also Documentation/filesystems/proc.rst.
773 coresight_cpu_debug.enable
776 Enable/disable the CPU sampling based debugging.
777 0: default value, disable debugging
778 1: enable debugging at boot time
780 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
782 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
784 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
785 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
786 Some features depend on CPU0. Known dependencies are:
787 1. Resume from suspend/hibernate depends on CPU0.
788 Suspend/hibernate will fail if CPU0 is offline and you
789 need to online CPU0 before suspend/hibernate.
790 2. PIC interrupts also depend on CPU0. CPU0 can't be
791 removed if a PIC interrupt is detected.
792 It's said poweroff/reboot may depend on CPU0 on some
793 machines although I haven't seen such issues so far
794 after CPU0 is offline on a few tested machines.
795 If the dependencies are under your control, you can
796 turn on cpu0_hotplug.
798 cpuidle.off=1 [CPU_IDLE]
799 disable the cpuidle sub-system
802 [CPU_IDLE] Name of the cpuidle governor to use.
804 cpufreq.off=1 [CPU_FREQ]
805 disable the cpufreq sub-system
807 cpufreq.default_governor=
808 [CPU_FREQ] Name of the default cpufreq governor or
809 policy to use. This governor must be registered in the
810 kernel before the cpufreq driver probes.
813 [X86] Delay for N microsec between assert and de-assert
814 of APIC INIT to start processors. This delay occurs
815 on every CPU online, such as boot, and resume from suspend.
818 crash_kexec_post_notifiers
819 Run kdump after running panic-notifiers and dumping
820 kmsg. This only for the users who doubt kdump always
821 succeeds in any situation.
822 Note that this also increases risks of kdump failure,
823 because some panic notifiers can make the crashed
824 kernel more unstable.
826 crashkernel=size[KMG][@offset[KMG]]
827 [KNL] Using kexec, Linux can switch to a 'crash kernel'
828 upon panic. This parameter reserves the physical
829 memory region [offset, offset + size] for that kernel
830 image. If '@offset' is omitted, then a suitable offset
831 is selected automatically.
832 [KNL, X86-64] Select a region under 4G first, and
833 fall back to reserve region above 4G when '@offset'
834 hasn't been specified.
835 See Documentation/admin-guide/kdump/kdump.rst for further details.
837 crashkernel=range1:size1[,range2:size2,...][@offset]
838 [KNL] Same as above, but depends on the memory
839 in the running system. The syntax of range is
840 start-[end] where start and end are both
841 a memory unit (amount[KMG]). See also
842 Documentation/admin-guide/kdump/kdump.rst for an example.
844 crashkernel=size[KMG],high
845 [KNL, X86-64, ARM64] range could be above 4G. Allow kernel
846 to allocate physical memory region from top, so could
847 be above 4G if system have more than 4G ram installed.
848 Otherwise memory region will be allocated below 4G, if
850 It will be ignored if crashkernel=X is specified.
851 crashkernel=size[KMG],low
852 [KNL, X86-64] range under 4G. When crashkernel=X,high
853 is passed, kernel could allocate physical memory region
854 above 4G, that cause second kernel crash on system
855 that require some amount of low memory, e.g. swiotlb
856 requires at least 64M+32K low memory, also enough extra
857 low memory is needed to make sure DMA buffers for 32-bit
858 devices won't run out. Kernel would try to allocate
859 at least 256M below 4G automatically.
860 This one lets the user specify own low range under 4G
861 for second kernel instead.
862 0: to disable low allocation.
863 It will be ignored when crashkernel=X,high is not used
864 or memory reserved is below 4G.
866 [KNL, ARM64] range in low memory.
867 This one lets the user specify a low range in the
868 DMA zone for the crash dump kernel.
869 It will be ignored when crashkernel=X,high is not used
870 or memory reserved is located in the DMA zones.
873 [KNL] Disable crypto self-tests
878 cs89x0_media= [HW,NET]
879 Format: { rj45 | aui | bnc }
881 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
882 handling. When switched on, additional debug data is
883 printed to the console in case a hanging CPU is
884 detected, and that CPU is pinged again in order to try
885 to resolve the hang situation.
886 0: disable csdlock debugging (default)
887 1: enable basic csdlock debugging (minor impact)
888 ext: enable extended csdlock debugging (more impact,
892 See header of drivers/s390/block/dasd_devmap.c.
894 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
895 (one device per port)
896 Format: <port#>,<type>
897 See also Documentation/input/devices/joystick-parport.rst
899 debug [KNL] Enable kernel debugging (events log level).
902 [KNL] Enable printing [hashed] pointers early in the
903 boot sequence. If enabled, we use a weak hash instead
904 of siphash to hash pointers. Use this option if you are
905 seeing instances of '(___ptrval___)') and need to see a
906 value (hashed pointer) instead. Cryptographically
907 insecure, please do not use on production kernels.
910 [KNL] verbose locking self-tests
912 Print debugging info while doing the locking API
914 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
915 (no extra messages), setting it to -1 (all bits set)
916 will print _a_lot_ more information - normally only
917 useful to lockdep developers.
919 debug_objects [KNL] Enable object debugging
922 [KNL] Disable object debugging
924 debug_guardpage_minorder=
925 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
926 parameter allows control of the order of pages that will
927 be intentionally kept free (and hence protected) by the
928 buddy allocator. Bigger value increase the probability
929 of catching random memory corruption, but reduce the
930 amount of memory for normal system use. The maximum
931 possible value is MAX_ORDER/2. Setting this parameter
932 to 1 or 2 should be enough to identify most random
933 memory corruption problems caused by bugs in kernel or
934 driver code when a CPU writes to (or reads from) a
935 random memory location. Note that there exists a class
936 of memory corruptions problems caused by buggy H/W or
937 F/W or by drivers badly programing DMA (basically when
938 memory is written at bus level and the CPU MMU is
939 bypassed) which are not detectable by
940 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
941 tracking down these problems.
944 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
945 enables the feature at boot time. By default, it is
946 disabled and the system will work mostly the same as a
947 kernel built without CONFIG_DEBUG_PAGEALLOC.
948 Note: to get most of debug_pagealloc error reports, it's
949 useful to also enable the page_owner functionality.
950 on: enable the feature
952 debugfs= [KNL] This parameter enables what is exposed to userspace
953 and debugfs internal clients.
954 Format: { on, no-mount, off }
955 on: All functions are enabled.
957 Filesystem is not registered but kernel clients can
958 access APIs and a crashkernel can be used to read
959 its content. There is nothing to mount.
960 off: Filesystem is not registered and clients
961 get a -EPERM as result when trying to register files
962 or directories within debugfs.
963 This is equivalent of the runtime functionality if
964 debugfs was not enabled in the kernel at all.
965 Default value is set in build-time with a kernel configuration.
967 debugpat [X86] Enable PAT debugging
969 decnet.addr= [HW,NET]
970 Format: <area>[,<node>]
971 See also Documentation/networking/decnet.rst.
974 [HW] The size of the default HugeTLB page. This is
975 the size represented by the legacy /proc/ hugepages
976 APIs. In addition, this is the default hugetlb size
977 used for shmget(), mmap() and mounting hugetlbfs
978 filesystems. If not specified, defaults to the
979 architecture's default huge page size. Huge page
980 sizes are architecture dependent. See also
981 Documentation/admin-guide/mm/hugetlbpage.rst.
984 deferred_probe_timeout=
985 [KNL] Debugging option to set a timeout in seconds for
986 deferred probe to give up waiting on dependencies to
987 probe. Only specific dependencies (subsystems or
988 drivers) that have opted in will be ignored. A timeout
989 of 0 will timeout at the end of initcalls. If the time
990 out hasn't expired, it'll be restarted by each
991 successful driver registration. This option will also
992 dump out devices still on the deferred probe list after
995 delayacct [KNL] Enable per-task delay accounting
997 dell_smm_hwmon.ignore_dmi=
998 [HW] Continue probing hardware even if DMI data
999 indicates that the driver is running on unsupported
1002 dell_smm_hwmon.force=
1003 [HW] Activate driver even if SMM BIOS signature does
1004 not match list of supported models and enable otherwise
1005 blacklisted features.
1007 dell_smm_hwmon.power_status=
1008 [HW] Report power status in /proc/i8k
1009 (disabled by default).
1011 dell_smm_hwmon.restricted=
1012 [HW] Allow controlling fans only if SYS_ADMIN
1015 dell_smm_hwmon.fan_mult=
1016 [HW] Factor to multiply fan speed with.
1018 dell_smm_hwmon.fan_max=
1019 [HW] Maximum configurable fan speed.
1022 Format: { on | off | def_only | inf_only | always }
1023 on: s390 zlib hardware support for compression on
1024 level 1 and decompression (default)
1025 off: No s390 zlib hardware support
1026 def_only: s390 zlib hardware support for deflate
1027 only (compression on level 1)
1028 inf_only: s390 zlib hardware support for inflate
1029 only (decompression)
1030 always: Same as 'on' but ignores the selected compression
1031 level always using hardware support (used for debugging)
1033 dhash_entries= [KNL]
1034 Set number of hash buckets for dentry cache.
1036 disable_1tb_segments [PPC]
1037 Disables the use of 1TB hash page table segments. This
1038 causes the kernel to fall back to 256MB segments which
1039 can be useful when debugging issues that require an SLB
1043 Limits the number of kernel SLB entries, and flushes
1044 them frequently to increase the rate of SLB faults
1045 on kernel addresses.
1048 See Documentation/networking/ipv6.rst.
1051 Disable RADIX MMU mode on POWER9
1053 radix_hcall_invalidate=on [PPC/PSERIES]
1054 Disable RADIX GTSE feature and use hcall for TLB
1058 Disable TLBIE instruction. Currently does not work
1059 with KVM, with HASH MMU, or with coherent accelerators.
1061 disable_cpu_apicid= [X86,APIC,SMP]
1063 The number of initial APIC ID for the
1064 corresponding CPU to be disabled at boot,
1065 mostly used for the kdump 2nd kernel to
1066 disable BSP to wake up multiple CPUs without
1067 causing system reset or hang due to sending
1068 INIT from AP to BSP.
1070 disable_ddw [PPC/PSERIES]
1071 Disable Dynamic DMA Window support. Use this
1072 to workaround buggy firmware.
1074 disable_ipv6= [IPV6]
1075 See Documentation/networking/ipv6.rst.
1077 disable_mtrr_cleanup [X86]
1078 The kernel tries to adjust MTRR layout from continuous
1079 to discrete, to make X server driver able to add WB
1080 entry later. This parameter disables that.
1082 disable_mtrr_trim [X86, Intel and AMD only]
1083 By default the kernel will trim any uncacheable
1084 memory out of your available memory pool based on
1085 MTRR settings. This parameter disables that behavior,
1086 possibly causing your machine to run very slowly.
1088 disable_timer_pin_1 [X86]
1089 Disable PIN 1 of APIC timer
1090 Can be useful to work around chipset bugs.
1092 dis_ucode_ldr [X86] Disable the microcode loader.
1094 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1095 this option disables the debugging code at boot.
1097 dma_debug_entries=<number>
1098 This option allows to tune the number of preallocated
1099 entries for DMA-API debugging code. One entry is
1100 required per DMA-API allocation. Use this if the
1101 DMA-API debugging code disables itself because the
1102 architectural default is too low.
1104 dma_debug_driver=<driver_name>
1105 With this option the DMA-API debugging driver
1106 filter feature can be enabled at boot time. Just
1107 pass the driver to filter for as the parameter.
1108 The filter can be disabled or changed to another
1109 driver later using sysfs.
1111 driver_async_probe= [KNL]
1112 List of driver names to be probed asynchronously. *
1113 matches with all driver names. If * is specified, the
1114 rest of the listed driver names are those that will NOT
1116 Format: <driver_name1>,<driver_name2>...
1118 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1119 Broken monitors, graphic adapters, KVMs and EDIDless
1120 panels may send no or incorrect EDID data sets.
1121 This parameter allows to specify an EDID data sets
1122 in the /lib/firmware directory that are used instead.
1123 Generic built-in EDID data sets are used, if one of
1124 edid/1024x768.bin, edid/1280x1024.bin,
1125 edid/1680x1050.bin, or edid/1920x1080.bin is given
1126 and no file with the same name exists. Details and
1127 instructions how to build your own EDID data are
1128 available in Documentation/admin-guide/edid.rst. An EDID
1129 data set will only be used for a particular connector,
1130 if its name and a colon are prepended to the EDID
1131 name. Each connector may use a unique EDID data
1132 set by separating the files with a comma. An EDID
1133 data set with no connector name will be used for
1134 any connectors not explicitly specified.
1139 Format: {"off" | "known"}
1140 Control how the dt_cpu_ftrs device-tree binding is
1141 used for CPU feature discovery and setup (if it
1143 off: Do not use it, fall back to legacy cpu table.
1144 known: Do not pass through unknown features to guests
1145 or userspace, only those that the kernel is aware of.
1147 dump_apple_properties [X86]
1148 Dump name and content of EFI device properties on
1149 x86 Macs. Useful for driver authors to determine
1150 what data is available or for reverse-engineering.
1152 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1153 <module>.dyndbg[="val"]
1154 Enable debug messages at boot time. See
1155 Documentation/admin-guide/dynamic-debug-howto.rst
1158 nopku [X86] Disable Memory Protection Keys CPU feature found
1161 <module>.async_probe[=<bool>] [KNL]
1162 If no <bool> value is specified or if the value
1163 specified is not a valid <bool>, enable asynchronous
1164 probe on this module. Otherwise, enable/disable
1165 asynchronous probe on this module as indicated by the
1166 <bool> value. See also: module.async_probe
1168 early_ioremap_debug [KNL]
1169 Enable debug messages in early_ioremap support. This
1170 is useful for tracking down temporary early mappings
1171 which are not unmapped.
1173 earlycon= [KNL] Output early console device and options.
1175 When used with no options, the early console is
1176 determined by stdout-path property in device tree's
1177 chosen node or the ACPI SPCR table if supported by
1180 cdns,<addr>[,options]
1181 Start an early, polled-mode console on a Cadence
1182 (xuartps) serial port at the specified address. Only
1183 supported option is baud rate. If baud rate is not
1184 specified, the serial port must already be setup and
1187 uart[8250],io,<addr>[,options]
1188 uart[8250],mmio,<addr>[,options]
1189 uart[8250],mmio32,<addr>[,options]
1190 uart[8250],mmio32be,<addr>[,options]
1191 uart[8250],0x<addr>[,options]
1192 Start an early, polled-mode console on the 8250/16550
1193 UART at the specified I/O port or MMIO address.
1194 MMIO inter-register address stride is either 8-bit
1195 (mmio) or 32-bit (mmio32 or mmio32be).
1196 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1197 to be equivalent to 'mmio'. 'options' are specified
1198 in the same format described for "console=ttyS<n>"; if
1199 unspecified, the h/w is not initialized.
1203 Start an early, polled-mode console on a pl011 serial
1204 port at the specified address. The pl011 serial port
1205 must already be setup and configured. Options are not
1206 yet supported. If 'mmio32' is specified, then only
1207 the driver will use only 32-bit accessors to read/write
1208 the device registers.
1211 Start an early console on a litex serial port at the
1212 specified address. The serial port must already be
1213 setup and configured. Options are not yet supported.
1216 Start an early, polled-mode console on a meson serial
1217 port at the specified address. The serial port must
1218 already be setup and configured. Options are not yet
1222 Start an early, polled-mode console on an msm serial
1223 port at the specified address. The serial port
1224 must already be setup and configured. Options are not
1227 msm_serial_dm,<addr>
1228 Start an early, polled-mode console on an msm serial
1229 dm port at the specified address. The serial port
1230 must already be setup and configured. Options are not
1234 Start an early, polled-mode console on a serial port
1235 of an Actions Semi SoC, such as S500 or S900, at the
1236 specified address. The serial port must already be
1237 setup and configured. Options are not yet supported.
1240 Start an early, polled-mode console on a serial port
1241 of an RDA Micro SoC, such as RDA8810PL, at the
1242 specified address. The serial port must already be
1243 setup and configured. Options are not yet supported.
1246 Use RISC-V SBI (Supervisor Binary Interface) for early
1249 smh Use ARM semihosting calls for early console.
1257 Use early console provided by serial driver available
1258 on Samsung SoCs, requires selecting proper type and
1259 a correct base address of the selected UART port. The
1260 serial port must already be setup and configured.
1261 Options are not yet supported.
1264 Start an early, polled-mode console on a lantiq serial
1265 (lqasc) port at the specified address. The serial port
1266 must already be setup and configured. Options are not
1271 Use early console provided by Freescale LP UART driver
1272 found on Freescale Vybrid and QorIQ LS1021A processors.
1273 A valid base address must be provided, and the serial
1274 port must already be setup and configured.
1278 Start an early, polled-mode, output-only console on the
1279 Freescale i.MX UART at the specified address. The UART
1280 must already be setup and configured.
1283 Start an early, polled-mode console on the
1284 Armada 3700 serial port at the specified
1285 address. The serial port must already be setup
1286 and configured. Options are not yet supported.
1289 Start an early, polled-mode console on a Qualcomm
1290 Generic Interface (GENI) based serial port at the
1291 specified address. The serial port must already be
1292 setup and configured. Options are not yet supported.
1295 Start an early, unaccelerated console on the EFI
1296 memory mapped framebuffer (if available). On cache
1297 coherent non-x86 systems that use system memory for
1298 the framebuffer, pass the 'ram' option so that it is
1299 mapped with the correct attributes.
1302 Use early console provided by Freescale LINFlexD UART
1303 serial driver for NXP S32V234 SoCs. A valid base
1304 address must be provided, and the serial port must
1305 already be setup and configured.
1307 earlyprintk= [X86,SH,ARM,M68k,S390]
1311 earlyprintk=serial[,ttySn[,baudrate]]
1312 earlyprintk=serial[,0x...[,baudrate]]
1313 earlyprintk=ttySn[,baudrate]
1314 earlyprintk=dbgp[debugController#]
1315 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1316 earlyprintk=xdbc[xhciController#]
1318 earlyprintk is useful when the kernel crashes before
1319 the normal console is initialized. It is not enabled by
1320 default because it has some cosmetic problems.
1322 Append ",keep" to not disable it when the real console
1325 Only one of vga, serial, or usb debug port can
1328 Currently only ttyS0 and ttyS1 may be specified by
1329 name. Other I/O ports may be explicitly specified
1330 on some architectures (x86 and arm at least) by
1331 replacing ttySn with an I/O port address, like this:
1332 earlyprintk=serial,0x1008,115200
1333 You can find the port for a given device in
1334 /proc/tty/driver/serial:
1335 2: uart:ST16650V2 port:00001008 irq:18 ...
1337 Interaction with the standard serial driver is not
1340 The VGA output is eventually overwritten by
1343 The xen option can only be used in Xen domains.
1345 The sclp output can only be used on s390.
1347 The optional "force" to "pciserial" enables use of a
1348 PCI device even when its classcode is not of the
1351 edac_report= [HW,EDAC] Control how to report EDAC event
1352 Format: {"on" | "off" | "force"}
1353 on: enable EDAC to report H/W event. May be overridden
1354 by other higher priority error reporting module.
1355 off: disable H/W event reporting through EDAC.
1356 force: enforce the use of EDAC to report H/W event.
1360 Format: {"off" | "on" | "skip[mbr]"}
1363 Format: { "debug", "disable_early_pci_dma",
1364 "nochunk", "noruntime", "nosoftreserve",
1365 "novamap", "no_disable_early_pci_dma" }
1366 debug: enable misc debug output.
1367 disable_early_pci_dma: disable the busmaster bit on all
1368 PCI bridges while in the EFI boot stub.
1369 nochunk: disable reading files in "chunks" in the EFI
1370 boot stub, as chunking can cause problems with some
1371 firmware implementations.
1372 noruntime : disable EFI runtime services support
1373 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1374 attribute may cause the kernel to reserve the
1375 memory range for a memory mapping driver to
1376 claim. Specify efi=nosoftreserve to disable this
1377 reservation and treat the memory by its base type
1378 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1379 novamap: do not call SetVirtualAddressMap().
1380 no_disable_early_pci_dma: Leave the busmaster bit set
1381 on all PCI bridges while in the EFI boot stub
1383 efi_no_storage_paranoia [EFI; X86]
1384 Using this parameter you can use more than 50% of
1385 your efi variable storage. Use this parameter only if
1386 you are really sure that your UEFI does sane gc and
1387 fulfills the spec otherwise your board may brick.
1389 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1390 Add arbitrary attribute to specific memory range by
1391 updating original EFI memory map.
1392 Region of memory which aa attribute is added to is
1395 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1396 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1397 attribute is added to range 0x100000000-0x180000000 and
1398 0x10a0000000-0x1120000000.
1400 If efi_fake_mem=8G@9G:0x40000 is specified, the
1401 EFI_MEMORY_SP(0x40000) attribute is added to
1402 range 0x240000000-0x43fffffff.
1404 Using this parameter you can do debugging of EFI memmap
1405 related features. For example, you can do debugging of
1406 Address Range Mirroring feature even if your box
1407 doesn't support it, or mark specific memory as
1410 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1411 that is to be dynamically loaded by Linux. If there are
1412 multiple variables with the same name but with different
1413 vendor GUIDs, all of them will be loaded. See
1414 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1417 eisa_irq_edge= [PARISC,HW]
1418 See header of drivers/parisc/eisa.c.
1420 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1423 This is designed to be used in conjunction with
1424 the boot argument: earlyprintk=vga
1426 This parameter works in place of the kgdboc parameter
1427 but can only be used if the backing tty is available
1428 very early in the boot process. For early debugging
1429 via a serial port see kgdboc_earlycon instead.
1432 See comment before function elanfreq_setup() in
1433 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1435 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1436 Specifies physical address of start of kernel core
1437 image elf header and optionally the size. Generally
1438 kexec loader will pass this option to capture kernel.
1439 See Documentation/admin-guide/kdump/kdump.rst for details.
1441 enable_mtrr_cleanup [X86]
1442 The kernel tries to adjust MTRR layout from continuous
1443 to discrete, to make X server driver able to add WB
1444 entry later. This parameter enables that.
1446 enable_timer_pin_1 [X86]
1447 Enable PIN 1 of APIC timer
1448 Can be useful to work around chipset bugs
1449 (in particular on some ATI chipsets).
1450 The kernel tries to set a reasonable default.
1452 enforcing= [SELINUX] Set initial enforcing status.
1454 See security/selinux/Kconfig help text.
1455 0 -- permissive (log only, no denials).
1456 1 -- enforcing (deny and log).
1458 Value can be changed at runtime via
1459 /sys/fs/selinux/enforce.
1462 Disable Error Record Serialization Table (ERST)
1465 ether= [HW,NET] Ethernet cards parameters
1466 This option is obsoleted by the "netdev=" option, which
1467 has equivalent usage. See its documentation for details.
1471 Permit 'security.evm' to be updated regardless of
1472 current integrity status.
1477 fail_make_request=[KNL]
1478 General fault injection mechanism.
1479 Format: <interval>,<probability>,<space>,<times>
1480 See also Documentation/fault-injection/.
1483 Format: { initns | none }
1484 See Documentation/admin-guide/sysctl/net.rst for
1485 fb_tunnels_only_for_init_ns
1488 See Documentation/admin-guide/blockdev/floppy.rst.
1490 force_pal_cache_flush
1491 [IA-64] Avoid check_sal_cache_flush which may hang on
1492 buggy SAL_CACHE_FLUSH implementations. Using this
1493 parameter will force ia64_sal_cache_flush to call
1494 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1497 Forcefully enable Physical Address Extension (PAE).
1498 Many Pentium M systems disable PAE but may have a
1499 functionally usable PAE implementation.
1500 Warning: use of this parameter will taint the kernel
1501 and may cause unknown problems.
1504 [FTRACE] will set and start the specified tracer
1505 as early as possible in order to facilitate early
1508 ftrace_boot_snapshot
1509 [FTRACE] On boot up, a snapshot will be taken of the
1510 ftrace ring buffer that can be read at:
1511 /sys/kernel/tracing/snapshot.
1512 This is useful if you need tracing information from kernel
1513 boot up that is likely to be overridden by user space
1514 start up functionality.
1516 ftrace_dump_on_oops[=orig_cpu]
1517 [FTRACE] will dump the trace buffers on oops.
1518 If no parameter is passed, ftrace will dump
1519 buffers of all CPUs, but if you pass orig_cpu, it will
1520 dump only the buffer of the CPU that triggered the
1523 ftrace_filter=[function-list]
1524 [FTRACE] Limit the functions traced by the function
1525 tracer at boot up. function-list is a comma-separated
1526 list of functions. This list can be changed at run
1527 time by the set_ftrace_filter file in the debugfs
1530 ftrace_notrace=[function-list]
1531 [FTRACE] Do not trace the functions specified in
1532 function-list. This list can be changed at run time
1533 by the set_ftrace_notrace file in the debugfs
1536 ftrace_graph_filter=[function-list]
1537 [FTRACE] Limit the top level callers functions traced
1538 by the function graph tracer at boot up.
1539 function-list is a comma-separated list of functions
1540 that can be changed at run time by the
1541 set_graph_function file in the debugfs tracing directory.
1543 ftrace_graph_notrace=[function-list]
1544 [FTRACE] Do not trace from the functions specified in
1545 function-list. This list is a comma-separated list of
1546 functions that can be changed at run time by the
1547 set_graph_notrace file in the debugfs tracing directory.
1549 ftrace_graph_max_depth=<uint>
1550 [FTRACE] Used with the function graph tracer. This is
1551 the max depth it will trace into a function. This value
1552 can be changed at run time by the max_graph_depth file
1553 in the tracefs tracing directory. default: 0 (no limit)
1555 fw_devlink= [KNL] Create device links between consumer and supplier
1556 devices by scanning the firmware to infer the
1557 consumer/supplier relationships. This feature is
1558 especially useful when drivers are loaded as modules as
1559 it ensures proper ordering of tasks like device probing
1560 (suppliers first, then consumers), supplier boot state
1561 clean up (only after all consumers have probed),
1562 suspend/resume & runtime PM (consumers first, then
1564 Format: { off | permissive | on | rpm }
1565 off -- Don't create device links from firmware info.
1566 permissive -- Create device links from firmware info
1567 but use it only for ordering boot state clean
1568 up (sync_state() calls).
1569 on -- Create device links from firmware info and use it
1570 to enforce probe and suspend/resume ordering.
1571 rpm -- Like "on", but also use to order runtime PM.
1573 fw_devlink.strict=<bool>
1574 [KNL] Treat all inferred dependencies as mandatory
1575 dependencies. This only applies for fw_devlink=on|rpm.
1579 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1580 support via parallel port (up to 5 devices per port)
1581 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1582 See also Documentation/input/devices/joystick-parport.rst
1586 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1590 gcov_persist= [GCOV] When non-zero (default), profiling data for
1591 kernel modules is saved and remains accessible via
1592 debugfs, even when the module is unloaded/reloaded.
1593 When zero, profiling data is discarded and associated
1594 debugfs files are removed at module unload time.
1596 goldfish [X86] Enable the goldfish android emulator platform.
1597 Don't use this when you are not running on the
1600 gpio-mockup.gpio_mockup_ranges
1601 [HW] Sets the ranges of gpiochip of for this device.
1602 Format: <start1>,<end1>,<start2>,<end2>...
1603 gpio-mockup.gpio_mockup_named_lines
1604 [HW] Let the driver know GPIO lines should be named.
1606 gpt [EFI] Forces disk with valid GPT signature but
1607 invalid Protective MBR to be treated as GPT. If the
1608 primary GPT is corrupted, it enables the backup/alternate
1609 GPT to be used instead.
1611 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1612 the "Enable 0" bit of the configuration register.
1615 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1616 the "Enable 0" bit of the configuration register.
1619 grcan.select= [HW] Select which physical interface to use.
1622 grcan.txsize= [HW] Sets the size of the tx buffer.
1623 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1625 grcan.rxsize= [HW] Sets the size of the rx buffer.
1626 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1630 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1631 hardening is enabled for this boot. Hardened
1632 usercopy checking is used to protect the kernel
1633 from reading or writing beyond known memory
1634 allocation boundaries as a proactive defense
1635 against bounds-checking flaws in the kernel's
1636 copy_to_user()/copy_from_user() interface.
1637 on Perform hardened usercopy checks (default).
1638 off Disable hardened usercopy checks.
1640 hardlockup_all_cpu_backtrace=
1641 [KNL] Should the hard-lockup detector generate
1642 backtraces on all cpus.
1645 hashdist= [KNL,NUMA] Large hashes allocated during boot
1646 are distributed across NUMA nodes. Defaults on
1647 for 64-bit NUMA, off otherwise.
1648 Format: 0 | 1 (for off | on)
1650 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1652 hd= [EIDE] (E)IDE hard drive subsystem geometry
1653 Format: <cyl>,<head>,<sect>
1656 Disable Hardware Error Source Table (HEST) support;
1657 corresponding firmware-first mode error processing
1658 logic will be disabled.
1660 hibernate= [HIBERNATION]
1661 noresume Don't check if there's a hibernation image
1662 present during boot.
1663 nocompress Don't compress/decompress hibernation images.
1664 no Disable hibernation and resume.
1665 protect_image Turn on image protection during restoration
1666 (that will set all pages holding image data
1667 during restoration read-only).
1669 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1670 size of <nn>. This works even on boxes that have no
1671 highmem otherwise. This also works to reduce highmem
1672 size on bigger boxes.
1674 highres= [KNL] Enable/disable high resolution timer mode.
1675 Valid parameters: "on", "off"
1680 hostname= [KNL] Set the hostname (aka UTS nodename).
1682 This allows setting the system's hostname during early
1683 startup. This sets the name returned by gethostname.
1684 Using this parameter to set the hostname makes it
1685 possible to ensure the hostname is correctly set before
1686 any userspace processes run, avoiding the possibility
1687 that a process may call gethostname before the hostname
1688 has been explicitly set, resulting in the calling
1689 process getting an incorrect result. The string must
1690 not exceed the maximum allowed hostname length (usually
1691 64 characters) and will be truncated otherwise.
1693 hpet= [X86-32,HPET] option to control HPET usage
1694 Format: { enable (default) | disable | force |
1696 disable: disable HPET and use PIT instead
1697 force: allow force enabled of undocumented chips (ICH4,
1699 verbose: show contents of HPET registers during setup
1701 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1702 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1704 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1705 If this follows hugepagesz (below), it specifies
1706 the number of pages of hugepagesz to be allocated.
1707 If this is the first HugeTLB parameter on the command
1708 line, it specifies the number of pages to allocate for
1709 the default huge page size. If using node format, the
1710 number of pages to allocate per-node can be specified.
1711 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1712 Format: <integer> or (node format)
1713 <node>:<integer>[,<node>:<integer>]
1716 [HW] The size of the HugeTLB pages. This is used in
1717 conjunction with hugepages (above) to allocate huge
1718 pages of a specific size at boot. The pair
1719 hugepagesz=X hugepages=Y can be specified once for
1720 each supported huge page size. Huge page sizes are
1721 architecture dependent. See also
1722 Documentation/admin-guide/mm/hugetlbpage.rst.
1725 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1726 of gigantic hugepages. Or using node format, the size
1727 of a CMA area per node can be specified.
1728 Format: nn[KMGTPE] or (node format)
1729 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1731 Reserve a CMA area of given size and allocate gigantic
1732 hugepages using the CMA allocator. If enabled, the
1733 boot-time allocation of gigantic hugepages is skipped.
1735 hugetlb_free_vmemmap=
1736 [KNL] Reguires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1738 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1739 Allows heavy hugetlb users to free up some more
1740 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1741 Format: { on | off (default) }
1746 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1749 Note that the vmemmap pages may be allocated from the added
1750 memory block itself when memory_hotplug.memmap_on_memory is
1751 enabled, those vmemmap pages cannot be optimized even if this
1752 feature is enabled. Other vmemmap pages not allocated from
1753 the added memory block itself do not be affected.
1756 [KNL] Should the hung task detector generate panics.
1759 A value of 1 instructs the kernel to panic when a
1760 hung task is detected. The default value is controlled
1761 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1762 option. The value selected by this boot parameter can
1763 be changed later by the kernel.hung_task_panic sysctl.
1765 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1766 terminal devices. Valid values: 0..8
1767 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1768 If specified, z/VM IUCV HVC accepts connections
1769 from listed z/VM user IDs only.
1771 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1772 which allow the hypervisor to 'idle' the
1773 guest on lock contention.
1776 Do not unregister boot console at start. This is only
1777 useful for debugging when something happens in the window
1778 between unregistering the boot console and initializing
1781 i2c_bus= [HW] Override the default board specific I2C bus speed
1782 or register an additional I2C bus that is not
1783 registered from board initialization code.
1787 i8042.debug [HW] Toggle i8042 debug mode
1788 i8042.unmask_kbd_data
1789 [HW] Enable printing of interrupt data from the KBD port
1790 (disabled by default, and as a pre-condition
1791 requires that i8042.debug=1 be enabled)
1792 i8042.direct [HW] Put keyboard port into non-translated mode
1793 i8042.dumbkbd [HW] Pretend that controller can only read data from
1794 keyboard and cannot control its state
1795 (Don't attempt to blink the leds)
1796 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1797 i8042.nokbd [HW] Don't check/create keyboard port
1798 i8042.noloop [HW] Disable the AUX Loopback command while probing
1800 i8042.nomux [HW] Don't check presence of an active multiplexing
1802 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1804 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1805 i8042.reset [HW] Reset the controller during init, cleanup and
1806 suspend-to-ram transitions, only during s2r
1807 transitions, or never reset
1808 Format: { 1 | Y | y | 0 | N | n }
1809 1, Y, y: always reset controller
1810 0, N, n: don't ever reset controller
1811 Default: only on s2r transitions on x86; most other
1812 architectures force reset to be always executed
1813 i8042.unlock [HW] Unlock (ignore) the keylock
1814 i8042.kbdreset [HW] Reset device connected to KBD port
1816 [HW] Allow deferred probing upon i8042 probe errors
1820 i915.invert_brightness=
1821 [DRM] Invert the sense of the variable that is used to
1822 set the brightness of the panel backlight. Normally a
1823 brightness value of 0 indicates backlight switched off,
1824 and the maximum of the brightness value sets the backlight
1825 to maximum brightness. If this parameter is set to 0
1826 (default) and the machine requires it, or this parameter
1827 is set to 1, a brightness value of 0 sets the backlight
1828 to maximum brightness, and the maximum of the brightness
1829 value switches the backlight off.
1830 -1 -- never invert brightness
1831 0 -- machine default
1832 1 -- force brightness inversion
1835 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1839 Format: idle=poll, idle=halt, idle=nomwait
1840 Poll forces a polling idle loop that can slightly
1841 improve the performance of waking up a idle CPU, but
1842 will use a lot of power and make the system run hot.
1844 idle=halt: Halt is forced to be used for CPU idle.
1845 In such case C2/C3 won't be used again.
1846 idle=nomwait: Disable mwait for CPU C-states
1850 Allow force disabling of Shared Virtual Memory (SVA)
1851 support for the idxd driver. By default it is set to
1854 idxd.tc_override= [HW]
1856 Allow override of default traffic class configuration
1857 for the device. By default it is set to false (0).
1859 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1860 Format: { strict | legacy | 2008 | relaxed }
1863 Choose which programs will be accepted for execution
1864 based on the IEEE 754 NaN encoding(s) supported by
1865 the FPU and the NaN encoding requested with the value
1866 of an ELF file header flag individually set by each
1867 binary. Hardware implementations are permitted to
1868 support either or both of the legacy and the 2008 NaN
1871 Available settings are as follows:
1872 strict accept binaries that request a NaN encoding
1873 supported by the FPU
1874 legacy only accept legacy-NaN binaries, if supported
1876 2008 only accept 2008-NaN binaries, if supported
1878 relaxed accept any binaries regardless of whether
1879 supported by the FPU
1881 The FPU emulator is always able to support both NaN
1882 encodings, so if no FPU hardware is present or it has
1883 been disabled with 'nofpu', then the settings of
1884 'legacy' and '2008' strap the emulator accordingly,
1885 'relaxed' straps the emulator for both legacy-NaN and
1886 2008-NaN, whereas 'strict' enables legacy-NaN only on
1887 legacy processors and both NaN encodings on MIPS32 or
1890 The setting for ABS.fmt/NEG.fmt instruction execution
1891 mode generally follows that for the NaN encoding,
1892 except where unsupported by hardware.
1894 ignore_loglevel [KNL]
1895 Ignore loglevel setting - this will print /all/
1896 kernel messages to the console. Useful for debugging.
1897 We also add it as printk module parameter, so users
1898 could change it dynamically, usually by
1899 /sys/module/printk/parameters/ignore_loglevel.
1902 Ignore RLIMIT_DATA setting for data mappings,
1903 print warning at first misuse. Can be changed via
1904 /sys/module/kernel/parameters/ignore_rlimit_data.
1906 ihash_entries= [KNL]
1907 Set number of hash buckets for inode cache.
1909 ima_appraise= [IMA] appraise integrity measurements
1910 Format: { "off" | "enforce" | "fix" | "log" }
1913 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1914 The builtin appraise policy appraises all files
1917 ima_canonical_fmt [IMA]
1918 Use the canonical format for the binary runtime
1919 measurements, instead of host native format.
1922 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1926 The list of supported hash algorithms is defined
1927 in crypto/hash_info.h.
1930 The builtin policies to load during IMA setup.
1931 Format: "tcb | appraise_tcb | secure_boot |
1932 fail_securely | critical_data"
1934 The "tcb" policy measures all programs exec'd, files
1935 mmap'd for exec, and all files opened with the read
1936 mode bit set by either the effective uid (euid=0) or
1939 The "appraise_tcb" policy appraises the integrity of
1940 all files owned by root.
1942 The "secure_boot" policy appraises the integrity
1943 of files (eg. kexec kernel image, kernel modules,
1944 firmware, policy, etc) based on file signatures.
1946 The "fail_securely" policy forces file signature
1947 verification failure also on privileged mounted
1948 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1951 The "critical_data" policy measures kernel integrity
1954 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1955 Load a policy which meets the needs of the Trusted
1956 Computing Base. This means IMA will measure all
1957 programs exec'd, files mmap'd for exec, and all files
1958 opened for read by uid=0.
1961 Select one of defined IMA measurements template formats.
1962 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
1967 [IMA] Define a custom template format.
1968 Format: { "field1|...|fieldN" }
1970 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1971 Format: <min_file_size>
1972 Set the minimal file size for using asynchronous hash.
1973 If left unspecified, ahash usage is disabled.
1975 ahash performance varies for different data sizes on
1976 different crypto accelerators. This option can be used
1977 to achieve the best performance for a particular HW.
1979 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1981 Set hashing buffer size. Default: 4k.
1983 ahash performance varies for different chunk sizes on
1984 different crypto accelerators. This option can be used
1985 to achieve best performance for particular HW.
1989 Run specified binary instead of /sbin/init as init
1992 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1993 for working out where the kernel is dying during
1996 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1997 initcall functions. Useful for debugging built-in
1998 modules and initcalls.
2000 initramfs_async= [KNL]
2003 This parameter controls whether the initramfs
2004 image is unpacked asynchronously, concurrently
2005 with devices being probed and
2006 initialized. This should normally just work,
2007 but as a debugging aid, one can get the
2008 historical behaviour of the initramfs
2009 unpacking being completed before device_ and
2012 initrd= [BOOT] Specify the location of the initial ramdisk
2014 initrdmem= [KNL] Specify a physical address and size from which to
2015 load the initrd. If an initrd is compiled in or
2016 specified in the bootparams, it takes priority over this
2018 Format: ss[KMG],nn[KMG]
2021 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
2024 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2026 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
2028 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2030 init_pkru= [X86] Specify the default memory protection keys rights
2031 register contents for all processes. 0x55555554 by
2032 default (disallow access to all but pkey 0). Can
2033 override in debugfs after boot.
2035 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2038 int_pln_enable [X86] Enable power limit notification interrupt
2040 integrity_audit=[IMA]
2041 Format: { "0" | "1" }
2042 0 -- basic integrity auditing messages. (Default)
2043 1 -- additional integrity auditing messages.
2045 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2047 Enable intel iommu driver.
2049 Disable intel iommu driver.
2050 igfx_off [Default Off]
2051 By default, gfx is mapped as normal device. If a gfx
2052 device has a dedicated DMAR unit, the DMAR unit is
2053 bypassed by not enabling DMAR with this option. In
2054 this case, gfx device will use physical address for
2056 strict [Default Off]
2057 Deprecated, equivalent to iommu.strict=1.
2058 sp_off [Default Off]
2059 By default, super page will be supported if Intel IOMMU
2060 has the capability. With this option, super page will
2063 Enable the Intel IOMMU scalable mode if the hardware
2064 advertises that it has support for the scalable mode
2067 Disallow use of the Intel IOMMU scalable mode.
2068 tboot_noforce [Default Off]
2069 Do not force the Intel IOMMU enabled under tboot.
2070 By default, tboot will force Intel IOMMU on, which
2071 could harm performance of some high-throughput
2072 devices like 40GBit network cards, even if identity
2074 Note that using this option lowers the security
2075 provided by tboot because it makes the system
2076 vulnerable to DMA attacks.
2078 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2079 0 disables intel_idle and fall back on acpi_idle.
2080 1 to 9 specify maximum depth of C-state.
2084 Do not enable intel_pstate as the default
2085 scaling driver for the supported processors
2087 Use intel_pstate as a scaling driver, but configure it
2088 to work with generic cpufreq governors (instead of
2089 enabling its internal governor). This mode cannot be
2090 used along with the hardware-managed P-states (HWP)
2093 Enable intel_pstate on systems that prohibit it by default
2094 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2095 instead of acpi-cpufreq may disable platform features, such
2096 as thermal controls and power capping, that rely on ACPI
2097 P-States information being indicated to OSPM and therefore
2098 should be used with caution. This option does not work with
2099 processors that aren't supported by the intel_pstate driver
2100 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2102 Do not enable hardware P state control (HWP)
2105 Only load intel_pstate on systems which support
2106 hardware P state control (HWP) if available.
2108 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2109 Description Table, specifies preferred power management
2110 profile as "Enterprise Server" or "Performance Server",
2111 then this feature is turned on by default.
2113 Allow per-logical-CPU P-State performance control limits using
2114 cpufreq sysfs interface
2116 intremap= [X86-64, Intel-IOMMU]
2117 on enable Interrupt Remapping (default)
2118 off disable Interrupt Remapping
2119 nosid disable Source ID checking
2121 BIOS x2APIC opt-out request will be ignored
2122 nopost disable Interrupt Posting
2124 iomem= Disable strict checking of access to MMIO memory
2125 strict regions from userspace.
2140 nobypass [PPC/POWERNV]
2141 Disable IOMMU bypass, using IOMMU for PCI devices.
2143 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2144 Format: { "0" | "1" }
2145 0 - Try to allocate a 32-bit DMA address first, before
2146 falling back to the full range if needed.
2147 1 - Allocate directly from the full usable range,
2148 forcing Dual Address Cycle for PCI cards supporting
2149 greater than 32-bit addressing.
2151 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2152 Format: { "0" | "1" }
2154 Request that DMA unmap operations use deferred
2155 invalidation of hardware TLBs, for increased
2156 throughput at the cost of reduced device isolation.
2157 Will fall back to strict mode if not supported by
2158 the relevant IOMMU driver.
2160 DMA unmap operations invalidate IOMMU hardware TLBs
2162 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2163 Note: on x86, strict mode specified via one of the
2164 legacy driver-specific options takes precedence.
2167 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2168 Format: { "0" | "1" }
2169 0 - Use IOMMU translation for DMA.
2170 1 - Bypass the IOMMU for DMA.
2171 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2173 io7= [HW] IO7 for Marvel-based Alpha systems
2174 See comment before marvel_specify_io7 in
2175 arch/alpha/kernel/core_marvel.c.
2177 io_delay= [X86] I/O delay method
2179 Standard port 0x80 based delay
2181 Alternate port 0xed based delay (needed on some systems)
2183 Simple two microseconds delay
2188 See Documentation/admin-guide/nfs/nfsroot.rst.
2190 ipcmni_extend [KNL] Extend the maximum number of unique System V
2191 IPC identifiers from 32,768 to 16,777,216.
2193 irqaffinity= [SMP] Set the default irq affinity mask
2194 The argument is a cpu list, as described above.
2196 irqchip.gicv2_force_probe=
2199 Force the kernel to look for the second 4kB page
2200 of a GICv2 controller even if the memory range
2201 exposed by the device tree is too small.
2203 irqchip.gicv3_nolpi=
2205 Force the kernel to ignore the availability of
2206 LPIs (and by consequence ITSs). Intended for system
2207 that use the kernel as a bootloader, and thus want
2208 to let secondary kernels in charge of setting up
2211 irqchip.gicv3_pseudo_nmi= [ARM64]
2212 Enables support for pseudo-NMIs in the kernel. This
2213 requires the kernel to be built with
2214 CONFIG_ARM64_PSEUDO_NMI.
2217 When an interrupt is not handled search all handlers
2218 for it. Intended to get systems with badly broken
2222 When an interrupt is not handled search all handlers
2223 for it. Also check all handlers each timer
2224 interrupt. Intended to get systems with badly broken
2228 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2230 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2231 [Deprecated - use cpusets instead]
2232 Format: [flag-list,]<cpu-list>
2234 Specify one or more CPUs to isolate from disturbances
2235 specified in the flag list (default: domain):
2238 Disable the tick when a single task runs.
2240 A residual 1Hz tick is offloaded to workqueues, which you
2241 need to affine to housekeeping through the global
2242 workqueue's affinity configured via the
2243 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2244 by using the 'domain' flag described below.
2246 NOTE: by default the global workqueue runs on all CPUs,
2247 so to protect individual CPUs the 'cpumask' file has to
2248 be configured manually after bootup.
2251 Isolate from the general SMP balancing and scheduling
2252 algorithms. Note that performing domain isolation this way
2253 is irreversible: it's not possible to bring back a CPU to
2254 the domains once isolated through isolcpus. It's strongly
2255 advised to use cpusets instead to disable scheduler load
2256 balancing through the "cpuset.sched_load_balance" file.
2257 It offers a much more flexible interface where CPUs can
2258 move in and out of an isolated set anytime.
2260 You can move a process onto or off an "isolated" CPU via
2261 the CPU affinity syscalls or cpuset.
2262 <cpu number> begins at 0 and the maximum value is
2263 "number of CPUs in system - 1".
2267 Isolate from being targeted by managed interrupts
2268 which have an interrupt mask containing isolated
2269 CPUs. The affinity of managed interrupts is
2270 handled by the kernel and cannot be changed via
2271 the /proc/irq/* interfaces.
2273 This isolation is best effort and only effective
2274 if the automatically assigned interrupt mask of a
2275 device queue contains isolated and housekeeping
2276 CPUs. If housekeeping CPUs are online then such
2277 interrupts are directed to the housekeeping CPU
2278 so that IO submitted on the housekeeping CPU
2279 cannot disturb the isolated CPU.
2281 If a queue's affinity mask contains only isolated
2282 CPUs then this parameter has no effect on the
2283 interrupt routing decision, though interrupts are
2284 only delivered when tasks running on those
2285 isolated CPUs submit IO. IO submitted on
2286 housekeeping CPUs has no influence on those
2289 The format of <cpu-list> is described above.
2293 ivrs_ioapic [HW,X86-64]
2294 Provide an override to the IOAPIC-ID<->DEVICE-ID
2295 mapping provided in the IVRS ACPI table.
2296 By default, PCI segment is 0, and can be omitted.
2298 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2299 write the parameter as:
2300 ivrs_ioapic[10]=00:14.0
2301 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2302 PCI device 00:14.0 write the parameter as:
2303 ivrs_ioapic[10]=0001:00:14.0
2305 ivrs_hpet [HW,X86-64]
2306 Provide an override to the HPET-ID<->DEVICE-ID
2307 mapping provided in the IVRS ACPI table.
2308 By default, PCI segment is 0, and can be omitted.
2310 * To map HPET-ID decimal 0 to PCI device 00:14.0
2311 write the parameter as:
2312 ivrs_hpet[0]=00:14.0
2313 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2314 PCI device 00:14.0 write the parameter as:
2315 ivrs_ioapic[10]=0001:00:14.0
2317 ivrs_acpihid [HW,X86-64]
2318 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2319 mapping provided in the IVRS ACPI table.
2321 For example, to map UART-HID:UID AMD0020:0 to
2322 PCI segment 0x1 and PCI device ID 00:14.5,
2323 write the parameter as:
2324 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2326 By default, PCI segment is 0, and can be omitted.
2327 For example, PCI device 00:14.5 write the parameter as:
2328 ivrs_acpihid[00:14.5]=AMD0020:0
2330 js= [HW,JOY] Analog joystick
2331 See Documentation/input/joydev/joystick.rst.
2334 When CONFIG_RANDOMIZE_BASE is set, this disables
2335 kernel and module base offset ASLR (Address Space
2336 Layout Randomization).
2339 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2340 report on every invalid memory access. Without this
2341 parameter KASAN will print report only for the first
2346 kernelcore= [KNL,X86,IA-64,PPC]
2347 Format: nn[KMGTPE] | nn% | "mirror"
2348 This parameter specifies the amount of memory usable by
2349 the kernel for non-movable allocations. The requested
2350 amount is spread evenly throughout all nodes in the
2351 system as ZONE_NORMAL. The remaining memory is used for
2352 movable memory in its own zone, ZONE_MOVABLE. In the
2353 event, a node is too small to have both ZONE_NORMAL and
2354 ZONE_MOVABLE, kernelcore memory will take priority and
2355 other nodes will have a larger ZONE_MOVABLE.
2357 ZONE_MOVABLE is used for the allocation of pages that
2358 may be reclaimed or moved by the page migration
2359 subsystem. Note that allocations like PTEs-from-HighMem
2360 still use the HighMem zone if it exists, and the Normal
2361 zone if it does not.
2363 It is possible to specify the exact amount of memory in
2364 the form of "nn[KMGTPE]", a percentage of total system
2365 memory in the form of "nn%", or "mirror". If "mirror"
2366 option is specified, mirrored (reliable) memory is used
2367 for non-movable allocations and remaining memory is used
2368 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2369 are exclusive, so you cannot specify multiple forms.
2371 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2372 Format: <Controller#>[,poll interval]
2373 The controller # is the number of the ehci usb debug
2374 port as it is probed via PCI. The poll interval is
2375 optional and is the number seconds in between
2376 each poll cycle to the debug port in case you need
2377 the functionality for interrupting the kernel with
2378 gdb or control-c on the dbgp connection. When
2379 not using this parameter you use sysrq-g to break into
2380 the kernel debugger.
2382 kgdboc= [KGDB,HW] kgdb over consoles.
2383 Requires a tty driver that supports console polling,
2384 or a supported polling keyboard driver (non-usb).
2385 Serial only format: <serial_device>[,baud]
2386 keyboard only format: kbd
2387 keyboard and serial format: kbd,<serial_device>[,baud]
2388 Optional Kernel mode setting:
2389 kms, kbd format: kms,kbd
2390 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2392 kgdboc_earlycon= [KGDB,HW]
2393 If the boot console provides the ability to read
2394 characters and can work in polling mode, you can use
2395 this parameter to tell kgdb to use it as a backend
2396 until the normal console is registered. Intended to
2397 be used together with the kgdboc parameter which
2398 specifies the normal console to transition to.
2400 The name of the early console should be specified
2401 as the value of this parameter. Note that the name of
2402 the early console might be different than the tty
2403 name passed to kgdboc. It's OK to leave the value
2404 blank and the first boot console that implements
2405 read() will be picked.
2407 kgdbwait [KGDB] Stop kernel execution and enter the
2408 kernel debugger at the earliest opportunity.
2410 kmac= [MIPS] Korina ethernet MAC address.
2411 Configure the RouterBoard 532 series on-chip
2412 Ethernet adapter MAC address.
2414 kmemleak= [KNL] Boot-time kmemleak enable/disable
2415 Valid arguments: on, off
2417 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2420 kprobe_event=[probe-list]
2421 [FTRACE] Add kprobe events and enable at boot time.
2422 The probe-list is a semicolon delimited list of probe
2423 definitions. Each definition is same as kprobe_events
2424 interface, but the parameters are comma delimited.
2425 For example, to add a kprobe event on vfs_read with
2426 arg1 and arg2, add to the command line;
2428 kprobe_event=p,vfs_read,$arg1,$arg2
2430 See also Documentation/trace/kprobetrace.rst "Kernel
2431 Boot Parameter" section.
2433 kpti= [ARM64] Control page table isolation of user
2434 and kernel address spaces.
2435 Default: enabled on cores which need mitigation.
2439 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2440 Default is 0 (don't ignore, but inject #GP)
2442 kvm.eager_page_split=
2443 [KVM,X86] Controls whether or not KVM will try to
2444 proactively split all huge pages during dirty logging.
2445 Eager page splitting reduces interruptions to vCPU
2446 execution by eliminating the write-protection faults
2447 and MMU lock contention that would otherwise be
2448 required to split huge pages lazily.
2450 VM workloads that rarely perform writes or that write
2451 only to a small region of VM memory may benefit from
2452 disabling eager page splitting to allow huge pages to
2453 still be used for reads.
2455 The behavior of eager page splitting depends on whether
2456 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2457 disabled, all huge pages in a memslot will be eagerly
2458 split when dirty logging is enabled on that memslot. If
2459 enabled, eager page splitting will be performed during
2460 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2463 Eager page splitting is only supported when kvm.tdp_mmu=Y.
2467 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2468 Default is false (don't support).
2471 [KVM] Controls the software workaround for the
2472 X86_BUG_ITLB_MULTIHIT bug.
2473 force : Always deploy workaround.
2474 off : Never deploy workaround.
2475 auto : Deploy workaround based on the presence of
2476 X86_BUG_ITLB_MULTIHIT.
2480 If the software workaround is enabled for the host,
2481 guests do need not to enable it for nested guests.
2483 kvm.nx_huge_pages_recovery_ratio=
2484 [KVM] Controls how many 4KiB pages are periodically zapped
2485 back to huge pages. 0 disables the recovery, otherwise if
2486 the value is N KVM will zap 1/Nth of the 4KiB pages every
2487 period (see below). The default is 60.
2489 kvm.nx_huge_pages_recovery_period_ms=
2490 [KVM] Controls the time period at which KVM zaps 4KiB pages
2491 back to huge pages. If the value is a non-zero N, KVM will
2492 zap a portion (see ratio above) of the pages every N msecs.
2493 If the value is 0 (the default), KVM will pick a period based
2494 on the ratio, such that a page is zapped after 1 hour on average.
2496 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2497 Default is 1 (enabled)
2499 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2501 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2504 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2506 none: Forcefully disable KVM.
2508 nvhe: Standard nVHE-based mode, without support for
2511 protected: nVHE-based mode with support for guests whose
2512 state is kept private from the host.
2514 Defaults to VHE/nVHE based on hardware support. Setting
2515 mode to "protected" will disable kexec and hibernation
2518 kvm-arm.vgic_v3_group0_trap=
2519 [KVM,ARM] Trap guest accesses to GICv3 group-0
2522 kvm-arm.vgic_v3_group1_trap=
2523 [KVM,ARM] Trap guest accesses to GICv3 group-1
2526 kvm-arm.vgic_v3_common_trap=
2527 [KVM,ARM] Trap guest accesses to GICv3 common
2530 kvm-arm.vgic_v4_enable=
2531 [KVM,ARM] Allow use of GICv4 for direct injection of
2534 kvm_cma_resv_ratio=n [PPC]
2535 Reserves given percentage from system memory area for
2536 contiguous memory allocation for KVM hash pagetable
2538 By default it reserves 5% of total system memory.
2542 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2543 (virtualized MMU) support on capable Intel chips.
2544 Default is 1 (enabled)
2546 kvm-intel.emulate_invalid_guest_state=
2547 [KVM,Intel] Disable emulation of invalid guest state.
2548 Ignored if kvm-intel.enable_unrestricted_guest=1, as
2549 guest state is never invalid for unrestricted guests.
2550 This param doesn't apply to nested guests (L2), as KVM
2551 never emulates invalid L2 guest state.
2552 Default is 1 (enabled)
2554 kvm-intel.flexpriority=
2555 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2556 Default is 1 (enabled)
2559 [KVM,Intel] Enable VMX nesting (nVMX).
2560 Default is 0 (disabled)
2562 kvm-intel.unrestricted_guest=
2563 [KVM,Intel] Disable unrestricted guest feature
2564 (virtualized real and unpaged mode) on capable
2565 Intel chips. Default is 1 (enabled)
2567 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2570 Valid arguments: never, cond, always
2572 always: L1D cache flush on every VMENTER.
2573 cond: Flush L1D on VMENTER only when the code between
2574 VMEXIT and VMENTER can leak host memory.
2575 never: Disables the mitigation
2577 Default is cond (do L1 cache flush in specific instances)
2579 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2580 feature (tagged TLBs) on capable Intel chips.
2581 Default is 1 (enabled)
2583 l1d_flush= [X86,INTEL]
2584 Control mitigation for L1D based snooping vulnerability.
2586 Certain CPUs are vulnerable to an exploit against CPU
2587 internal buffers which can forward information to a
2588 disclosure gadget under certain conditions.
2590 In vulnerable processors, the speculatively
2591 forwarded data can be used in a cache side channel
2592 attack, to access data to which the attacker does
2593 not have direct access.
2595 This parameter controls the mitigation. The
2598 on - enable the interface for the mitigation
2600 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2603 The kernel PTE inversion protection is unconditionally
2604 enabled and cannot be disabled.
2607 Provides all available mitigations for the
2608 L1TF vulnerability. Disables SMT and
2609 enables all mitigations in the
2610 hypervisors, i.e. unconditional L1D flush.
2612 SMT control and L1D flush control via the
2613 sysfs interface is still possible after
2614 boot. Hypervisors will issue a warning
2615 when the first VM is started in a
2616 potentially insecure configuration,
2617 i.e. SMT enabled or L1D flush disabled.
2620 Same as 'full', but disables SMT and L1D
2621 flush runtime control. Implies the
2622 'nosmt=force' command line option.
2623 (i.e. sysfs control of SMT is disabled.)
2626 Leaves SMT enabled and enables the default
2627 hypervisor mitigation, i.e. conditional
2630 SMT control and L1D flush control via the
2631 sysfs interface is still possible after
2632 boot. Hypervisors will issue a warning
2633 when the first VM is started in a
2634 potentially insecure configuration,
2635 i.e. SMT enabled or L1D flush disabled.
2639 Disables SMT and enables the default
2640 hypervisor mitigation.
2642 SMT control and L1D flush control via the
2643 sysfs interface is still possible after
2644 boot. Hypervisors will issue a warning
2645 when the first VM is started in a
2646 potentially insecure configuration,
2647 i.e. SMT enabled or L1D flush disabled.
2650 Same as 'flush', but hypervisors will not
2651 warn when a VM is started in a potentially
2652 insecure configuration.
2655 Disables hypervisor mitigations and doesn't
2657 It also drops the swap size and available
2658 RAM limit restriction on both hypervisor and
2663 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2669 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2672 lapic= [X86,APIC] Do not use TSC deadline
2673 value for LAPIC timer one-shot implementation. Default
2674 back to the programmable timer unit in the LAPIC.
2675 Format: notscdeadline
2677 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2680 libata.dma= [LIBATA] DMA control
2681 libata.dma=0 Disable all PATA and SATA DMA
2682 libata.dma=1 PATA and SATA Disk DMA only
2683 libata.dma=2 ATAPI (CDROM) DMA only
2684 libata.dma=4 Compact Flash DMA only
2685 Combinations also work, so libata.dma=3 enables DMA
2686 for disks and CDROMs, but not CFs.
2688 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2689 libata.ignore_hpa=0 keep BIOS limits (default)
2690 libata.ignore_hpa=1 ignore limits, using full disk
2692 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2696 libata.force= [LIBATA] Force configurations. The format is a comma-
2697 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2698 PORT and DEVICE are decimal numbers matching port, link
2699 or device. Basically, it matches the ATA ID string
2700 printed on console by libata. If the whole ID part is
2701 omitted, the last PORT and DEVICE values are used. If
2702 ID hasn't been specified yet, the configuration applies
2703 to all ports, links and devices.
2705 If only DEVICE is omitted, the parameter applies to
2706 the port and all links and devices behind it. DEVICE
2707 number of 0 either selects the first device or the
2708 first fan-out link behind PMP device. It does not
2709 select the host link. DEVICE number of 15 selects the
2710 host link and device attached to it.
2712 The VAL specifies the configuration to force. As long
2713 as there is no ambiguity, shortcut notation is allowed.
2714 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2715 The following configurations can be forced.
2717 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2718 Any ID with matching PORT is used.
2720 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2722 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2723 udma[/][16,25,33,44,66,100,133] notation is also
2726 * nohrst, nosrst, norst: suppress hard, soft and both
2729 * rstonce: only attempt one reset during hot-unplug
2732 * [no]dbdelay: Enable or disable the extra 200ms delay
2733 before debouncing a link PHY and device presence
2736 * [no]ncq: Turn on or off NCQ.
2738 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2740 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2742 * [no]trim: Enable or disable (unqueued) TRIM.
2744 * trim_zero: Indicate that TRIM command zeroes data.
2746 * max_trim_128m: Set 128M maximum trim size limit.
2748 * [no]dma: Turn on or off DMA transfers.
2750 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2752 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2753 commands that are not a multiple of 16 bytes.
2755 * [no]dmalog: Enable or disable the use of the
2756 READ LOG DMA EXT command to access logs.
2758 * [no]iddevlog: Enable or disable access to the
2759 identify device data log.
2761 * [no]logdir: Enable or disable access to the general
2762 purpose log directory.
2764 * max_sec_128: Set transfer size limit to 128 sectors.
2766 * max_sec_1024: Set or clear transfer size limit to
2769 * max_sec_lba48: Set or clear transfer size limit to
2772 * [no]lpm: Enable or disable link power management.
2774 * [no]setxfer: Indicate if transfer speed mode setting
2777 * dump_id: Dump IDENTIFY data.
2779 * disable: Disable this device.
2781 If there are multiple matching configurations changing
2782 the same attribute, the last one is used.
2784 load_ramdisk= [RAM] [Deprecated]
2786 lockd.nlm_grace_period=P [NFS] Assign grace period.
2789 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2792 lockd.nlm_timeout=T [NFS] Assign timeout value.
2795 lockd.nlm_udpport=M [NFS] Assign UDP port.
2798 lockdown= [SECURITY]
2799 { integrity | confidentiality }
2800 Enable the kernel lockdown feature. If set to
2801 integrity, kernel features that allow userland to
2802 modify the running kernel are disabled. If set to
2803 confidentiality, kernel features that allow userland
2804 to extract confidential information from the kernel
2807 locktorture.nreaders_stress= [KNL]
2808 Set the number of locking read-acquisition kthreads.
2809 Defaults to being automatically set based on the
2810 number of online CPUs.
2812 locktorture.nwriters_stress= [KNL]
2813 Set the number of locking write-acquisition kthreads.
2815 locktorture.onoff_holdoff= [KNL]
2816 Set time (s) after boot for CPU-hotplug testing.
2818 locktorture.onoff_interval= [KNL]
2819 Set time (s) between CPU-hotplug operations, or
2820 zero to disable CPU-hotplug testing.
2822 locktorture.shuffle_interval= [KNL]
2823 Set task-shuffle interval (jiffies). Shuffling
2824 tasks allows some CPUs to go into dyntick-idle
2825 mode during the locktorture test.
2827 locktorture.shutdown_secs= [KNL]
2828 Set time (s) after boot system shutdown. This
2829 is useful for hands-off automated testing.
2831 locktorture.stat_interval= [KNL]
2832 Time (s) between statistics printk()s.
2834 locktorture.stutter= [KNL]
2835 Time (s) to stutter testing, for example,
2836 specifying five seconds causes the test to run for
2837 five seconds, wait for five seconds, and so on.
2838 This tests the locking primitive's ability to
2839 transition abruptly to and from idle.
2841 locktorture.torture_type= [KNL]
2842 Specify the locking implementation to test.
2844 locktorture.verbose= [KNL]
2845 Enable additional printk() statements.
2847 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2850 loglevel= All Kernel Messages with a loglevel smaller than the
2851 console loglevel will be printed to the console. It can
2852 also be changed with klogd or other programs. The
2853 loglevels are defined as follows:
2855 0 (KERN_EMERG) system is unusable
2856 1 (KERN_ALERT) action must be taken immediately
2857 2 (KERN_CRIT) critical conditions
2858 3 (KERN_ERR) error conditions
2859 4 (KERN_WARNING) warning conditions
2860 5 (KERN_NOTICE) normal but significant condition
2861 6 (KERN_INFO) informational
2862 7 (KERN_DEBUG) debug-level messages
2864 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2865 in bytes. n must be a power of two and greater
2866 than the minimal size. The minimal size is defined
2867 by LOG_BUF_SHIFT kernel config parameter. There is
2868 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2869 that allows to increase the default size depending on
2870 the number of CPUs. See init/Kconfig for more details.
2872 logo.nologo [FB] Disables display of the built-in Linux logo.
2873 This may be used to provide more screen space for
2874 kernel log messages and is useful when debugging
2875 kernel boot problems.
2877 lp=0 [LP] Specify parallel ports to use, e.g,
2878 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2879 lp=reset first parallel port). 'lp=0' disables the
2880 lp=auto printer driver. 'lp=reset' (which can be
2881 specified in addition to the ports) causes
2882 attached printers to be reset. Using
2883 lp=port1,port2,... specifies the parallel ports
2884 to associate lp devices with, starting with
2885 lp0. A port specification may be 'none' to skip
2886 that lp device, or a parport name such as
2887 'parport0'. Specifying 'lp=auto' instead of a
2888 port specification list means that device IDs
2889 from each port should be examined, to see if
2890 an IEEE 1284-compliant printer is attached; if
2891 so, the driver will manage that printer.
2892 See also header of drivers/char/lp.c.
2895 Sets loops_per_jiffy to given constant, thus avoiding
2896 time-consuming boot-time autodetection (up to 250 ms per
2897 CPU). 0 enables autodetection (default). To determine
2898 the correct value for your kernel, boot with normal
2899 autodetection and see what value is printed. Note that
2900 on SMP systems the preset will be applied to all CPUs,
2901 which is likely to cause problems if your CPUs need
2902 significantly divergent settings. An incorrect value
2903 will cause delays in the kernel to be wrong, leading to
2904 unpredictable I/O errors and other breakage. Although
2905 unlikely, in the extreme case this might damage your
2909 Format: <io>,<irq>,<dma>
2911 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2914 [SECURITY] Choose order of LSM initialization. This
2915 overrides CONFIG_LSM, and the "security=" parameter.
2917 machvec= [IA-64] Force the use of a particular machine-vector
2918 (machvec) in a generic kernel.
2919 Example: machvec=hpzx1
2921 machtype= [Loongson] Share the same kernel image file between
2922 different yeeloong laptops.
2923 Example: machtype=lemote-yeeloong-2f-7inch
2925 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
2926 than or equal to this physical address is ignored.
2928 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2929 will bring up during bootup. maxcpus=n : n >= 0 limits
2930 the kernel to bring up 'n' processors. Surely after
2931 bootup you can bring up the other plugged cpu by executing
2932 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2933 only takes effect during system bootup.
2934 While n=0 is a special case, it is equivalent to "nosmp",
2935 which also disables the IO APIC.
2937 max_loop= [LOOP] The number of loop block devices that get
2938 (loop.max_loop) unconditionally pre-created at init time. The default
2939 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2940 of statically allocating a predefined number, loop
2941 devices can be requested on-demand with the
2942 /dev/loop-control interface.
2944 mce [X86-32] Machine Check Exception
2946 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2948 md= [HW] RAID subsystems devices and level
2949 See Documentation/admin-guide/md.rst.
2952 Format: <first>,<last>
2953 Specifies range of consoles to be captured by the MDA.
2956 Control mitigation for the Micro-architectural Data
2957 Sampling (MDS) vulnerability.
2959 Certain CPUs are vulnerable to an exploit against CPU
2960 internal buffers which can forward information to a
2961 disclosure gadget under certain conditions.
2963 In vulnerable processors, the speculatively
2964 forwarded data can be used in a cache side channel
2965 attack, to access data to which the attacker does
2966 not have direct access.
2968 This parameter controls the MDS mitigation. The
2971 full - Enable MDS mitigation on vulnerable CPUs
2972 full,nosmt - Enable MDS mitigation and disable
2973 SMT on vulnerable CPUs
2974 off - Unconditionally disable MDS mitigation
2976 On TAA-affected machines, mds=off can be prevented by
2977 an active TAA mitigation as both vulnerabilities are
2978 mitigated with the same mechanism so in order to disable
2979 this mitigation, you need to specify tsx_async_abort=off
2982 Not specifying this option is equivalent to
2985 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2987 mem=nn[KMG] [HEXAGON] Set the memory size.
2988 Must be specified, otherwise memory size will be 0.
2990 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2991 Amount of memory to be used in cases as follows:
2994 2 when the kernel is not able to see the whole system memory;
2995 3 memory that lies after 'mem=' boundary is excluded from
2996 the hypervisor, then assigned to KVM guests.
2997 4 to limit the memory available for kdump kernel.
2999 [ARC,MICROBLAZE] - the limit applies only to low memory,
3000 high memory is not affected.
3002 [ARM64] - only limits memory covered by the linear
3003 mapping. The NOMAP regions are not affected.
3005 [X86] Work as limiting max address. Use together
3006 with memmap= to avoid physical address space collisions.
3007 Without memmap= PCI devices could be placed at addresses
3008 belonging to unused RAM.
3010 Note that this only takes effects during boot time since
3011 in above case 3, memory may need be hot added after boot
3012 if system memory of hypervisor is not sufficient.
3015 [ARM,MIPS] - override the memory layout reported by
3017 Define a memory region of size nn[KMG] starting at
3019 Multiple different regions can be specified with
3020 multiple mem= parameters on the command line.
3022 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3025 memblock=debug [KNL] Enable memblock debug messages.
3028 [KNL,SH] Allow user to override the default size for
3029 per-device physically contiguous DMA buffers.
3031 memhp_default_state=online/offline
3032 [KNL] Set the initial state for the memory hotplug
3033 onlining policy. If not specified, the default value is
3034 set according to the
3035 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3037 See Documentation/admin-guide/mm/memory-hotplug.rst.
3039 memmap=exactmap [KNL,X86] Enable setting of an exact
3040 E820 memory map, as specified by the user.
3041 Such memmap=exactmap lines can be constructed based on
3042 BIOS output or other requirements. See the memmap=nn@ss
3045 memmap=nn[KMG]@ss[KMG]
3046 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
3047 Region of memory to be used is from ss to ss+nn.
3048 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3049 which limits max address to nn[KMG].
3050 Multiple different regions can be specified,
3053 memmap=100M@2G,100M#3G,1G!1024G
3055 memmap=nn[KMG]#ss[KMG]
3056 [KNL,ACPI] Mark specific memory as ACPI data.
3057 Region of memory to be marked is from ss to ss+nn.
3059 memmap=nn[KMG]$ss[KMG]
3060 [KNL,ACPI] Mark specific memory as reserved.
3061 Region of memory to be reserved is from ss to ss+nn.
3062 Example: Exclude memory from 0x18690000-0x1869ffff
3063 memmap=64K$0x18690000
3065 memmap=0x10000$0x18690000
3066 Some bootloaders may need an escape character before '$',
3067 like Grub2, otherwise '$' and the following number
3070 memmap=nn[KMG]!ss[KMG]
3071 [KNL,X86] Mark specific memory as protected.
3072 Region of memory to be used, from ss to ss+nn.
3073 The memory region may be marked as e820 type 12 (0xc)
3074 and is NVDIMM or ADR memory.
3076 memmap=<size>%<offset>-<oldtype>+<newtype>
3077 [KNL,ACPI] Convert memory within the specified region
3078 from <oldtype> to <newtype>. If "-<oldtype>" is left
3079 out, the whole region will be marked as <newtype>,
3080 even if previously unavailable. If "+<newtype>" is left
3081 out, matching memory will be removed. Types are
3082 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3083 3 = ACPI, 12 = PRAM.
3085 memory_corruption_check=0/1 [X86]
3086 Some BIOSes seem to corrupt the first 64k of
3087 memory when doing things like suspend/resume.
3088 Setting this option will scan the memory
3089 looking for corruption. Enabling this will
3090 both detect corruption and prevent the kernel
3091 from using the memory being corrupted.
3092 However, its intended as a diagnostic tool; if
3093 repeatable BIOS-originated corruption always
3094 affects the same memory, you can use memmap=
3095 to prevent the kernel from using that memory.
3097 memory_corruption_check_size=size [X86]
3098 By default it checks for corruption in the low
3099 64k, making this memory unavailable for normal
3100 use. Use this parameter to scan for
3101 corruption in more or less memory.
3103 memory_corruption_check_period=seconds [X86]
3104 By default it checks for corruption every 60
3105 seconds. Use this parameter to check at some
3106 other rate. 0 disables periodic checking.
3108 memory_hotplug.memmap_on_memory
3109 [KNL,X86,ARM] Boolean flag to enable this feature.
3110 Format: {on | off (default)}
3111 When enabled, runtime hotplugged memory will
3112 allocate its internal metadata (struct pages,
3113 those vmemmap pages cannot be optimized even
3114 if hugetlb_free_vmemmap is enabled) from the
3115 hotadded memory which will allow to hotadd a
3116 lot of memory without requiring additional
3118 This feature is disabled by default because it
3119 has some implication on large (e.g. GB)
3120 allocations in some configurations (e.g. small
3122 The state of the flag can be read in
3123 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3124 Note that even when enabled, there are a few cases where
3125 the feature is not effective.
3127 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3129 default : 0 <disable>
3130 Specifies the number of memtest passes to be
3131 performed. Each pass selects another test
3132 pattern from a given set of patterns. Memtest
3133 fills the memory with this pattern, validates
3134 memory contents and reserves bad memory
3135 regions that are detected.
3137 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3138 Valid arguments: on, off
3139 Default (depends on kernel configuration option):
3140 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3141 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3142 mem_encrypt=on: Activate SME
3143 mem_encrypt=off: Do not activate SME
3145 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3146 for details on when memory encryption can be activated.
3148 mem_sleep_default= [SUSPEND] Default system suspend mode:
3149 s2idle - Suspend-To-Idle
3150 shallow - Power-On Suspend or equivalent (if supported)
3151 deep - Suspend-To-RAM or equivalent (if supported)
3152 See Documentation/admin-guide/pm/sleep-states.rst.
3154 meye.*= [HW] Set MotionEye Camera parameters
3155 See Documentation/admin-guide/media/meye.rst.
3157 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3158 Multi-Function General Purpose Timers on AMD Geode
3161 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3162 the BIOS has incorrectly applied a workaround. TinyBIOS
3163 version 0.98 is known to be affected, 0.99 fixes the
3164 problem by letting the user disable the workaround.
3168 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3169 physical address is ignored.
3171 mini2440= [ARM,HW,KNL]
3172 Format:[0..2][b][c][t]
3174 MINI2440 configuration specification:
3175 0 - The attached screen is the 3.5" TFT
3176 1 - The attached screen is the 7" TFT
3177 2 - The VGA Shield is attached (1024x768)
3178 Leaving out the screen size parameter will not load
3179 the TFT driver, and the framebuffer will be left
3181 b - Enable backlight. The TFT backlight pin will be
3182 linked to the kernel VESA blanking code and a GPIO
3183 LED. This parameter is not necessary when using the
3185 c - Enable the s3c camera interface.
3186 t - Reserved for enabling touchscreen support. The
3187 touchscreen support is not enabled in the mainstream
3188 kernel as of 2.6.30, a preliminary port can be found
3189 in the "bleeding edge" mini2440 support kernel at
3190 https://repo.or.cz/w/linux-2.6/mini2440.git
3193 [X86,PPC,S390,ARM64] Control optional mitigations for
3194 CPU vulnerabilities. This is a set of curated,
3195 arch-independent options, each of which is an
3196 aggregation of existing arch-specific options.
3199 Disable all optional CPU mitigations. This
3200 improves system performance, but it may also
3201 expose users to several CPU vulnerabilities.
3202 Equivalent to: nopti [X86,PPC]
3203 if nokaslr then kpti=0 [ARM64]
3204 nospectre_v1 [X86,PPC]
3206 nospectre_v2 [X86,PPC,S390,ARM64]
3207 spectre_v2_user=off [X86]
3208 spec_store_bypass_disable=off [X86,PPC]
3209 ssbd=force-off [ARM64]
3212 tsx_async_abort=off [X86]
3213 kvm.nx_huge_pages=off [X86]
3214 srbds=off [X86,INTEL]
3215 no_entry_flush [PPC]
3216 no_uaccess_flush [PPC]
3217 mmio_stale_data=off [X86]
3221 This does not have any effect on
3222 kvm.nx_huge_pages when
3223 kvm.nx_huge_pages=force.
3226 Mitigate all CPU vulnerabilities, but leave SMT
3227 enabled, even if it's vulnerable. This is for
3228 users who don't want to be surprised by SMT
3229 getting disabled across kernel upgrades, or who
3230 have other ways of avoiding SMT-based attacks.
3231 Equivalent to: (default behavior)
3234 Mitigate all CPU vulnerabilities, disabling SMT
3235 if needed. This is for users who always want to
3236 be fully mitigated, even if it means losing SMT.
3237 Equivalent to: l1tf=flush,nosmt [X86]
3238 mds=full,nosmt [X86]
3239 tsx_async_abort=full,nosmt [X86]
3240 mmio_stale_data=full,nosmt [X86]
3241 retbleed=auto,nosmt [X86]
3244 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3245 parameter allows control of the logging verbosity for
3246 the additional memory initialisation checks. A value
3247 of 0 disables mminit logging and a level of 4 will
3248 log everything. Information is printed at KERN_DEBUG
3249 so loglevel=8 may also need to be specified.
3252 [X86,INTEL] Control mitigation for the Processor
3253 MMIO Stale Data vulnerabilities.
3255 Processor MMIO Stale Data is a class of
3256 vulnerabilities that may expose data after an MMIO
3257 operation. Exposed data could originate or end in
3258 the same CPU buffers as affected by MDS and TAA.
3259 Therefore, similar to MDS and TAA, the mitigation
3260 is to clear the affected CPU buffers.
3262 This parameter controls the mitigation. The
3265 full - Enable mitigation on vulnerable CPUs
3267 full,nosmt - Enable mitigation and disable SMT on
3270 off - Unconditionally disable mitigation
3272 On MDS or TAA affected machines,
3273 mmio_stale_data=off can be prevented by an active
3274 MDS or TAA mitigation as these vulnerabilities are
3275 mitigated with the same mechanism so in order to
3276 disable this mitigation, you need to specify
3277 mds=off and tsx_async_abort=off too.
3279 Not specifying this option is equivalent to
3280 mmio_stale_data=full.
3283 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3285 module.async_probe=<bool>
3286 [KNL] When set to true, modules will use async probing
3287 by default. To enable/disable async probing for a
3288 specific module, use the module specific control that
3289 is documented under <module>.async_probe. When both
3290 module.async_probe and <module>.async_probe are
3291 specified, <module>.async_probe takes precedence for
3292 the specific module.
3295 [KNL] When CONFIG_MODULE_SIG is set, this means that
3296 modules without (valid) signatures will fail to load.
3297 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3298 is always true, so this option does nothing.
3300 module_blacklist= [KNL] Do not load a comma-separated list of
3301 modules. Useful for debugging problem modules.
3304 [MOUSE] Maximum time between finger touching and
3305 leaving touchpad surface for touch to be considered
3306 a tap and be reported as a left button click (for
3307 touchpads working in absolute mode only).
3309 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3310 reporting absolute coordinates, such as tablets
3311 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3312 reporting absolute coordinates, such as tablets
3314 movablecore= [KNL,X86,IA-64,PPC]
3315 Format: nn[KMGTPE] | nn%
3316 This parameter is the complement to kernelcore=, it
3317 specifies the amount of memory used for migratable
3318 allocations. If both kernelcore and movablecore is
3319 specified, then kernelcore will be at *least* the
3320 specified value but may be more. If movablecore on its
3321 own is specified, the administrator must be careful
3322 that the amount of memory usable for all allocations
3325 movable_node [KNL] Boot-time switch to make hotplugable memory
3326 NUMA nodes to be movable. This means that the memory
3327 of such nodes will be usable only for movable
3328 allocations which rules out almost all kernel
3329 allocations. Use with caution!
3331 MTD_Partition= [MTD]
3332 Format: <name>,<region-number>,<size>,<offset>
3334 MTD_Region= [MTD] Format:
3335 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3338 See drivers/mtd/parsers/cmdlinepart.c
3341 ARM/S3C2412 JIVE boot control
3343 See arch/arm/mach-s3c/mach-jive.c
3345 mtouchusb.raw_coordinates=
3346 [HW] Make the MicroTouch USB driver use raw coordinates
3347 ('y', default) or cooked coordinates ('n')
3349 mtrr_chunk_size=nn[KMG] [X86]
3350 used for mtrr cleanup. It is largest continuous chunk
3351 that could hold holes aka. UC entries.
3353 mtrr_gran_size=nn[KMG] [X86]
3354 Used for mtrr cleanup. It is granularity of mtrr block.
3356 Large value could prevent small alignment from
3359 mtrr_spare_reg_nr=n [X86]
3361 Range: 0,7 : spare reg number
3363 Used for mtrr cleanup. It is spare mtrr entries number.
3364 Set to 2 or more if your graphical card needs more.
3366 multitce=off [PPC] This parameter disables the use of the pSeries
3367 firmware feature for updating multiple TCE entries
3370 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3372 netdev= [NET] Network devices parameters
3373 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3374 Note that mem_start is often overloaded to mean
3375 something different and driver-specific.
3376 This usage is only documented in each driver source
3379 netpoll.carrier_timeout=
3380 [NET] Specifies amount of time (in seconds) that
3381 netpoll should wait for a carrier. By default netpoll
3385 [NETFILTER] Enable connection tracking flow accounting
3386 0 to disable accounting
3387 1 to enable accounting
3390 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3391 See Documentation/admin-guide/nfs/nfsroot.rst.
3393 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3394 See Documentation/admin-guide/nfs/nfsroot.rst.
3396 nfsrootdebug [NFS] enable nfsroot debugging messages.
3397 See Documentation/admin-guide/nfs/nfsroot.rst.
3399 nfs.callback_nr_threads=
3400 [NFSv4] set the total number of threads that the
3401 NFS client will assign to service NFSv4 callback
3404 nfs.callback_tcpport=
3405 [NFS] set the TCP port on which the NFSv4 callback
3406 channel should listen.
3409 [NFS] sets the pathname to the program which is used
3410 to update the NFS client cache entries.
3412 nfs.cache_getent_timeout=
3413 [NFS] sets the timeout after which an attempt to
3414 update a cache entry is deemed to have failed.
3416 nfs.idmap_cache_timeout=
3417 [NFS] set the maximum lifetime for idmapper cache
3421 [NFS] enable 64-bit inode numbers.
3422 If zero, the NFS client will fake up a 32-bit inode
3423 number for the readdir() and stat() syscalls instead
3424 of returning the full 64-bit number.
3425 The default is to return 64-bit inode numbers.
3427 nfs.max_session_cb_slots=
3428 [NFSv4.1] Sets the maximum number of session
3429 slots the client will assign to the callback
3430 channel. This determines the maximum number of
3431 callbacks the client will process in parallel for
3432 a particular server.
3434 nfs.max_session_slots=
3435 [NFSv4.1] Sets the maximum number of session slots
3436 the client will attempt to negotiate with the server.
3437 This limits the number of simultaneous RPC requests
3438 that the client can send to the NFSv4.1 server.
3439 Note that there is little point in setting this
3440 value higher than the max_tcp_slot_table_limit.
3442 nfs.nfs4_disable_idmapping=
3443 [NFSv4] When set to the default of '1', this option
3444 ensures that both the RPC level authentication
3445 scheme and the NFS level operations agree to use
3446 numeric uids/gids if the mount is using the
3447 'sec=sys' security flavour. In effect it is
3448 disabling idmapping, which can make migration from
3449 legacy NFSv2/v3 systems to NFSv4 easier.
3450 Servers that do not support this mode of operation
3451 will be autodetected by the client, and it will fall
3452 back to using the idmapper.
3453 To turn off this behaviour, set the value to '0'.
3455 [NFS4] Specify an additional fixed unique ident-
3456 ification string that NFSv4 clients can insert into
3457 their nfs_client_id4 string. This is typically a
3458 UUID that is generated at system install time.
3460 nfs.send_implementation_id =
3461 [NFSv4.1] Send client implementation identification
3462 information in exchange_id requests.
3463 If zero, no implementation identification information
3465 The default is to send the implementation identification
3468 nfs.recover_lost_locks =
3469 [NFSv4] Attempt to recover locks that were lost due
3470 to a lease timeout on the server. Please note that
3471 doing this risks data corruption, since there are
3472 no guarantees that the file will remain unchanged
3473 after the locks are lost.
3474 If you want to enable the kernel legacy behaviour of
3475 attempting to recover these locks, then set this
3477 The default parameter value of '0' causes the kernel
3478 not to attempt recovery of lost locks.
3480 nfs4.layoutstats_timer =
3481 [NFSv4.2] Change the rate at which the kernel sends
3482 layoutstats to the pNFS metadata server.
3484 Setting this to value to 0 causes the kernel to use
3485 whatever value is the default set by the layout
3486 driver. A non-zero value sets the minimum interval
3487 in seconds between layoutstats transmissions.
3489 nfsd.inter_copy_offload_enable =
3490 [NFSv4.2] When set to 1, the server will support
3491 server-to-server copies for which this server is
3492 the destination of the copy.
3494 nfsd.nfsd4_ssc_umount_timeout =
3495 [NFSv4.2] When used as the destination of a
3496 server-to-server copy, knfsd temporarily mounts
3497 the source server. It caches the mount in case
3498 it will be needed again, and discards it if not
3499 used for the number of milliseconds specified by
3502 nfsd.nfs4_disable_idmapping=
3503 [NFSv4] When set to the default of '1', the NFSv4
3504 server will return only numeric uids and gids to
3505 clients using auth_sys, and will accept numeric uids
3506 and gids from such clients. This is intended to ease
3507 migration from NFSv2/v3.
3510 nmi_backtrace.backtrace_idle [KNL]
3511 Dump stacks even of idle CPUs in response to an
3512 NMI stack-backtrace request.
3514 nmi_debug= [KNL,SH] Specify one or more actions to take
3515 when a NMI is triggered.
3516 Format: [state][,regs][,debounce][,die]
3518 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3519 Format: [panic,][nopanic,][num]
3521 0 - turn hardlockup detector in nmi_watchdog off
3522 1 - turn hardlockup detector in nmi_watchdog on
3523 When panic is specified, panic when an NMI watchdog
3524 timeout occurs (or 'nopanic' to not panic on an NMI
3525 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3526 To disable both hard and soft lockup detectors,
3527 please see 'nowatchdog'.
3528 This is useful when you use a panic=... timeout and
3529 need the box quickly up again.
3531 These settings can be accessed at runtime via
3532 the nmi_watchdog and hardlockup_panic sysctls.
3534 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3535 emulation library even if a 387 maths coprocessor
3538 no5lvl [X86-64] Disable 5-level paging mode. Forces
3539 kernel to use 4-level paging instead.
3541 nofsgsbase [X86] Disables FSGSBASE instructions.
3544 [HW] Never suspend the console
3545 Disable suspending of consoles during suspend and
3546 hibernate operations. Once disabled, debugging
3547 messages can reach various consoles while the rest
3548 of the system is being put to sleep (ie, while
3549 debugging driver suspend/resume hooks). This may
3550 not work reliably with all consoles, but is known
3551 to work with serial and VGA consoles.
3552 To facilitate more flexible debugging, we also add
3553 console_suspend, a printk module parameter to control
3554 it. Users could use console_suspend (usually
3555 /sys/module/printk/parameters/console_suspend) to
3556 turn on/off it dynamically.
3558 novmcoredd [KNL,KDUMP]
3559 Disable device dump. Device dump allows drivers to
3560 append dump data to vmcore so you can collect driver
3561 specified debug info. Drivers can append the data
3562 without any limit and this data is stored in memory,
3563 so this may cause significant memory stress. Disabling
3564 device dump can help save memory but the driver debug
3565 data will be no longer available. This parameter
3566 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3569 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3570 caches in the slab allocator. Saves per-node memory,
3571 but will impact performance.
3575 noaltinstr [S390] Disables alternative instructions patching
3576 (CPU alternatives feature).
3578 noapic [SMP,APIC] Tells the kernel to not make use of any
3579 IOAPICs that may be present in the system.
3581 noautogroup Disable scheduler automatic task group creation.
3585 nodsp [SH] Disable hardware DSP at boot time.
3587 noefi Disable EFI runtime services support.
3589 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3594 Disable SMAP (Supervisor Mode Access Prevention)
3595 even if it is supported by processor.
3598 Disable SMEP (Supervisor Mode Execution Prevention)
3599 even if it is supported by processor.
3602 This affects only 32-bit executables.
3603 noexec32=on: enable non-executable mappings (default)
3604 read doesn't imply executable mappings
3605 noexec32=off: disable non-executable mappings
3606 read implies executable mappings
3608 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3610 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3611 register save and restore. The kernel will only save
3612 legacy floating-point registers on task switch.
3614 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3616 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3618 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3619 Equivalent to smt=1.
3621 [KNL,X86] Disable symmetric multithreading (SMT).
3622 nosmt=force: Force disable SMT, cannot be undone
3623 via the sysfs control file.
3625 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3626 (bounds check bypass). With this option data leaks are
3627 possible in the system.
3629 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3630 the Spectre variant 2 (indirect branch prediction)
3631 vulnerability. System may allow data leaks with this
3634 nospec_store_bypass_disable
3635 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3638 [PPC] Don't flush the L1-D cache after accessing user data.
3640 noxsave [BUGS=X86] Disables x86 extended register state save
3641 and restore using xsave. The kernel will fallback to
3642 enabling legacy floating-point and sse state.
3644 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3645 register states. The kernel will fall back to use
3646 xsave to save the states. By using this parameter,
3647 performance of saving the states is degraded because
3648 xsave doesn't support modified optimization while
3649 xsaveopt supports it on xsaveopt enabled systems.
3651 noxsaves [X86] Disables xsaves and xrstors used in saving and
3652 restoring x86 extended register state in compacted
3653 form of xsave area. The kernel will fall back to use
3654 xsaveopt and xrstor to save and restore the states
3655 in standard form of xsave area. By using this
3656 parameter, xsave area per process might occupy more
3657 memory on xsaves enabled systems.
3659 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3660 in do_idle() and not use the arch_cpu_idle()
3661 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3662 to be effective. This is useful on platforms where the
3663 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3664 correctly or when doing power measurements to evalute
3665 the impact of the sleep instructions. This is also
3666 useful when using JTAG debugger.
3668 no_file_caps Tells the kernel not to honor file capabilities. The
3669 only way then for a file to be executed with privilege
3670 is to be setuid root or executed by root.
3672 nohalt [IA-64] Tells the kernel not to use the power saving
3673 function PAL_HALT_LIGHT when idle. This increases
3674 power-consumption. On the positive side, it reduces
3675 interrupt wake-up latency, which may improve performance
3676 in certain environments such as networked servers or
3680 Force pointers printed to the console or buffers to be
3681 unhashed. By default, when a pointer is printed via %p
3682 format string, that pointer is "hashed", i.e. obscured
3683 by hashing the pointer value. This is a security feature
3684 that hides actual kernel addresses from unprivileged
3685 users, but it also makes debugging the kernel more
3686 difficult since unequal pointers can no longer be
3687 compared. However, if this command-line option is
3688 specified, then all normal pointers will have their true
3689 value printed. This option should only be specified when
3690 debugging the kernel. Please do not use on production
3693 nohibernate [HIBERNATION] Disable hibernation and resume.
3695 nohz= [KNL] Boottime enable/disable dynamic ticks
3696 Valid arguments: on, off
3699 nohz_full= [KNL,BOOT,SMP,ISOL]
3700 The argument is a cpu list, as described above.
3701 In kernels built with CONFIG_NO_HZ_FULL=y, set
3702 the specified list of CPUs whose tick will be stopped
3703 whenever possible. The boot CPU will be forced outside
3704 the range to maintain the timekeeping. Any CPUs
3705 in this list will have their RCU callbacks offloaded,
3706 just as if they had also been called out in the
3707 rcu_nocbs= boot parameter.
3709 Note that this argument takes precedence over
3710 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3712 noiotrap [SH] Disables trapped I/O port accesses.
3714 noirqdebug [X86-32] Disables the code which attempts to detect and
3715 disable unhandled interrupt sources.
3717 no_timer_check [X86,APIC] Disables the code which tests for
3718 broken timer IRQ sources.
3720 noisapnp [ISAPNP] Disables ISA PnP code.
3722 noinitrd [RAM] Tells the kernel not to load any configured
3725 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3727 [Deprecated - use intremap=off]
3731 noinvpcid [X86] Disable the INVPCID cpu feature.
3733 nojitter [IA-64] Disables jitter checking for ITC timers.
3735 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3737 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3741 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3742 clock and use the default one.
3744 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3745 accounting. steal time is computed, but won't
3746 influence scheduler behaviour
3748 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3750 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3752 nomca [IA-64] Disable machine check abort handling
3754 nomce [X86-32] Disable Machine Check Exception
3756 nomfgpt [X86-32] Disable Multi-Function General Purpose
3757 Timer usage (for AMD Geode machines).
3759 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3760 shutdown the other cpus. Instead use the REBOOT_VECTOR
3763 nomodeset Disable kernel modesetting. DRM drivers will not perform
3764 display-mode changes or accelerated rendering. Only the
3765 system framebuffer will be available for use if this was
3766 set-up by the firmware or boot loader.
3768 Useful as fallback, or for testing and debugging.
3770 nomodule Disable module load
3772 nopat [X86] Disable PAT (page attribute table extension of
3773 pagetables) support.
3775 nopcid [X86-64] Disable the PCID cpu feature.
3777 norandmaps Don't use address space randomization. Equivalent to
3778 echo 0 > /proc/sys/kernel/randomize_va_space
3780 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3781 with UP alternatives
3783 noresume [SWSUSP] Disables resume and restores original swap
3786 no-scroll [VGA] Disables scrollback.
3787 This is required for the Braillex ib80-piezo Braille
3788 reader made by F.H. Papenmeier (Germany).
3792 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3794 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3795 and disable the IO APIC. legacy for "maxcpus=0".
3797 nosoftlockup [KNL] Disable the soft-lockup detector.
3799 nosync [HW,M68K] Disables sync negotiation for all devices.
3801 nowatchdog [KNL] Disable both lockup detectors, i.e.
3802 soft-lockup and NMI watchdog (hard-lockup).
3806 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3808 nps_mtm_hs_ctr= [KNL,ARC]
3809 This parameter sets the maximum duration, in
3810 cycles, each HW thread of the CTOP can run
3811 without interruptions, before HW switches it.
3812 The actual maximum duration is 16 times this
3814 Format: integer between 1 and 255
3817 nptcg= [IA-64] Override max number of concurrent global TLB
3818 purges which is reported from either PAL_VM_SUMMARY or
3821 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3822 could support. nr_cpus=n : n >= 1 limits the kernel to
3823 support 'n' processors. It could be larger than the
3824 number of already plugged CPU during bootup, later in
3825 runtime you can physically add extra cpu until it reaches
3826 n. So during boot up some boot time memory for per-cpu
3827 variables need be pre-allocated for later physical cpu
3830 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3832 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3833 set up a single NUMA node spanning all memory.
3835 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3837 Allowed values are enable and disable
3839 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3840 'node', 'default' can be specified
3841 This can be set from sysctl after boot.
3842 See Documentation/admin-guide/sysctl/vm.rst for details.
3844 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3845 See Documentation/core-api/debugging-via-ohci1394.rst for more
3848 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3849 Rather than timing out after 20 ms if an EC
3850 command is not properly ACKed, override the length
3851 of the timeout. We have interrupts disabled while
3852 waiting for the ACK, so if this is set too high
3853 interrupts *may* be lost!
3855 omap_mux= [OMAP] Override bootloader pin multiplexing.
3856 Format: <mux_mode0.mode_name=value>...
3857 For example, to override I2C bus2:
3858 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3860 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3862 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3864 boundary - index of last SLC block on Flex-OneNAND.
3865 The remaining blocks are configured as MLC blocks.
3866 lock - Configure if Flex-OneNAND boundary should be locked.
3867 Once locked, the boundary cannot be changed.
3868 1 indicates lock status, 0 indicates unlock status.
3870 oops=panic Always panic on oopses. Default is to just kill the
3871 process, but there is a small probability of
3872 deadlocking the machine.
3873 This will also cause panics on machine check exceptions.
3874 Useful together with panic=30 to trigger a reboot.
3877 [KNL] Boolean flag to control whether the page allocator
3878 should randomize its free lists. The randomization may
3879 be automatically enabled if the kernel detects it is
3880 running on a platform with a direct-mapped memory-side
3881 cache, and this parameter can be used to
3882 override/disable that behavior. The state of the flag
3883 can be read from sysfs at:
3884 /sys/module/page_alloc/parameters/shuffle.
3886 page_owner= [KNL] Boot-time page_owner enabling option.
3887 Storage of the information about who allocated
3888 each page is disabled in default. With this switch,
3890 on: enable the feature
3892 page_poison= [KNL] Boot-time parameter changing the state of
3893 poisoning on the buddy allocator, available with
3894 CONFIG_PAGE_POISONING=y.
3895 off: turn off poisoning (default)
3896 on: turn on poisoning
3898 page_reporting.page_reporting_order=
3899 [KNL] Minimal page reporting order
3901 Adjust the minimal page reporting order. The page
3902 reporting is disabled when it exceeds (MAX_ORDER-1).
3904 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3905 timeout > 0: seconds before rebooting
3906 timeout = 0: wait forever
3907 timeout < 0: reboot immediately
3910 panic_print= Bitmask for printing system info when panic happens.
3911 User can chose combination of the following bits:
3912 bit 0: print all tasks info
3913 bit 1: print system memory info
3914 bit 2: print timer info
3915 bit 3: print locks info if CONFIG_LOCKDEP is on
3916 bit 4: print ftrace buffer
3917 bit 5: print all printk messages in buffer
3918 bit 6: print all CPUs backtrace (if available in the arch)
3919 *Be aware* that this option may print a _lot_ of lines,
3920 so there are risks of losing older messages in the log.
3921 Use this option carefully, maybe worth to setup a
3922 bigger log buffer with "log_buf_len" along with this.
3924 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3925 Format: <hex>[,nousertaint]
3926 Hexadecimal bitmask representing the set of TAINT flags
3927 that will cause the kernel to panic when add_taint() is
3928 called with any of the flags in this set.
3929 The optional switch "nousertaint" can be utilized to
3930 prevent userspace forced crashes by writing to sysctl
3931 /proc/sys/kernel/tainted any flagset matching with the
3932 bitmask set on panic_on_taint.
3933 See Documentation/admin-guide/tainted-kernels.rst for
3934 extra details on the taint flags that users can pick
3935 to compose the bitmask to assign to panic_on_taint.
3937 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3940 parkbd.port= [HW] Parallel port number the keyboard adapter is
3941 connected to, default is 0.
3943 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3944 0 for XT, 1 for AT (default is AT).
3947 parport= [HW,PPT] Specify parallel ports. 0 disables.
3948 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3949 Use 'auto' to force the driver to use any
3950 IRQ/DMA settings detected (the default is to
3951 ignore detected IRQ/DMA settings because of
3952 possible conflicts). You can specify the base
3953 address, IRQ, and DMA settings; IRQ and DMA
3954 should be numbers, or 'auto' (for using detected
3955 settings on that particular port), or 'nofifo'
3956 (to avoid using a FIFO even if it is detected).
3957 Parallel ports are assigned in the order they
3958 are specified on the command line, starting
3961 parport_init_mode= [HW,PPT]
3962 Configure VIA parallel port to operate in
3963 a specific mode. This is necessary on Pegasos
3964 computer where firmware has no options for setting
3965 up parallel port mode and sets it to spp.
3966 Currently this function knows 686a and 8231 chips.
3967 Format: [spp|ps2|epp|ecp|ecpepp]
3969 pata_legacy.all= [HW,LIBATA]
3971 Set to non-zero to probe primary and secondary ISA
3972 port ranges on PCI systems where no PCI PATA device
3973 has been found at either range. Disabled by default.
3975 pata_legacy.autospeed= [HW,LIBATA]
3977 Set to non-zero if a chip is present that snoops speed
3978 changes. Disabled by default.
3980 pata_legacy.ht6560a= [HW,LIBATA]
3982 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3983 the secondary channel, or both channels respectively.
3984 Disabled by default.
3986 pata_legacy.ht6560b= [HW,LIBATA]
3988 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3989 the secondary channel, or both channels respectively.
3990 Disabled by default.
3992 pata_legacy.iordy_mask= [HW,LIBATA]
3994 IORDY enable mask. Set individual bits to allow IORDY
3995 for the respective channel. Bit 0 is for the first
3996 legacy channel handled by this driver, bit 1 is for
3997 the second channel, and so on. The sequence will often
3998 correspond to the primary legacy channel, the secondary
3999 legacy channel, and so on, but the handling of a PCI
4000 bus and the use of other driver options may interfere
4001 with the sequence. By default IORDY is allowed across
4004 pata_legacy.opti82c46x= [HW,LIBATA]
4006 Set to 1, 2, or 3 for Opti 82c611A on the primary
4007 channel, the secondary channel, or both channels
4008 respectively. Disabled by default.
4010 pata_legacy.opti82c611a= [HW,LIBATA]
4012 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4013 channel, the secondary channel, or both channels
4014 respectively. Disabled by default.
4016 pata_legacy.pio_mask= [HW,LIBATA]
4018 PIO mode mask for autospeed devices. Set individual
4019 bits to allow the use of the respective PIO modes.
4020 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4021 All modes allowed by default.
4023 pata_legacy.probe_all= [HW,LIBATA]
4025 Set to non-zero to probe tertiary and further ISA
4026 port ranges on PCI systems. Disabled by default.
4028 pata_legacy.probe_mask= [HW,LIBATA]
4030 Probe mask for legacy ISA PATA ports. Depending on
4031 platform configuration and the use of other driver
4032 options up to 6 legacy ports are supported: 0x1f0,
4033 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4034 of individual ports can be disabled by setting the
4035 corresponding bits in the mask to 1. Bit 0 is for
4036 the first port in the list above (0x1f0), and so on.
4037 By default all supported ports are probed.
4039 pata_legacy.qdi= [HW,LIBATA]
4041 Set to non-zero to probe QDI controllers. By default
4042 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4044 pata_legacy.winbond= [HW,LIBATA]
4046 Set to non-zero to probe Winbond controllers. Use
4047 the standard I/O port (0x130) if 1, otherwise the
4048 value given is the I/O port to use (typically 0x1b0).
4049 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4052 pata_platform.pio_mask= [HW,LIBATA]
4054 Supported PIO mode mask. Set individual bits to allow
4055 the use of the respective PIO modes. Bit 0 is for
4056 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4060 Halt all CPUs after the first oops has been printed for
4061 the specified number of seconds. This is to be used if
4062 your oopses keep scrolling off the screen.
4067 See header of drivers/block/paride/pcd.c.
4068 See also Documentation/admin-guide/blockdev/paride.rst.
4070 pci=option[,option...] [PCI] various PCI subsystem options.
4072 Some options herein operate on a specific device
4073 or a set of devices (<pci_dev>). These are
4074 specified in one of the following formats:
4076 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4077 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4079 Note: the first format specifies a PCI
4080 bus/device/function address which may change
4081 if new hardware is inserted, if motherboard
4082 firmware changes, or due to changes caused
4083 by other kernel parameters. If the
4084 domain is left unspecified, it is
4085 taken to be zero. Optionally, a path
4086 to a device through multiple device/function
4087 addresses can be specified after the base
4088 address (this is more robust against
4089 renumbering issues). The second format
4090 selects devices using IDs from the
4091 configuration space which may match multiple
4092 devices in the system.
4094 earlydump dump PCI config space before the kernel
4096 off [X86] don't probe for the PCI bus
4097 bios [X86-32] force use of PCI BIOS, don't access
4098 the hardware directly. Use this if your machine
4099 has a non-standard PCI host bridge.
4100 nobios [X86-32] disallow use of PCI BIOS, only direct
4101 hardware access methods are allowed. Use this
4102 if you experience crashes upon bootup and you
4103 suspect they are caused by the BIOS.
4104 conf1 [X86] Force use of PCI Configuration Access
4105 Mechanism 1 (config address in IO port 0xCF8,
4106 data in IO port 0xCFC, both 32-bit).
4107 conf2 [X86] Force use of PCI Configuration Access
4108 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4109 the function, IO port 0xCFA, also 8-bit, sets
4110 bus number. The config space is then accessed
4111 through ports 0xC000-0xCFFF).
4112 See http://wiki.osdev.org/PCI for more info
4113 on the configuration access mechanisms.
4114 noaer [PCIE] If the PCIEAER kernel config parameter is
4115 enabled, this kernel boot option can be used to
4116 disable the use of PCIE advanced error reporting.
4117 nodomains [PCI] Disable support for multiple PCI
4118 root domains (aka PCI segments, in ACPI-speak).
4119 nommconf [X86] Disable use of MMCONFIG for PCI
4121 check_enable_amd_mmconf [X86] check for and enable
4122 properly configured MMIO access to PCI
4123 config space on AMD family 10h CPU
4124 nomsi [MSI] If the PCI_MSI kernel config parameter is
4125 enabled, this kernel boot option can be used to
4126 disable the use of MSI interrupts system-wide.
4127 noioapicquirk [APIC] Disable all boot interrupt quirks.
4128 Safety option to keep boot IRQs enabled. This
4129 should never be necessary.
4130 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4131 primary IO-APIC for bridges that cannot disable
4132 boot IRQs. This fixes a source of spurious IRQs
4133 when the system masks IRQs.
4134 noioapicreroute [APIC] Disable workaround that uses the
4135 boot IRQ equivalent of an IRQ that connects to
4136 a chipset where boot IRQs cannot be disabled.
4137 The opposite of ioapicreroute.
4138 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4139 routing table. These calls are known to be buggy
4140 on several machines and they hang the machine
4141 when used, but on other computers it's the only
4142 way to get the interrupt routing table. Try
4143 this option if the kernel is unable to allocate
4144 IRQs or discover secondary PCI buses on your
4146 rom [X86] Assign address space to expansion ROMs.
4147 Use with caution as certain devices share
4148 address decoders between ROMs and other
4150 norom [X86] Do not assign address space to
4151 expansion ROMs that do not already have
4152 BIOS assigned address ranges.
4153 nobar [X86] Do not assign address space to the
4154 BARs that weren't assigned by the BIOS.
4155 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4156 assigned automatically to PCI devices. You can
4157 make the kernel exclude IRQs of your ISA cards
4159 pirqaddr=0xAAAAA [X86] Specify the physical address
4160 of the PIRQ table (normally generated
4161 by the BIOS) if it is outside the
4162 F0000h-100000h range.
4163 lastbus=N [X86] Scan all buses thru bus #N. Can be
4164 useful if the kernel is unable to find your
4165 secondary buses and you want to tell it
4166 explicitly which ones they are.
4167 assign-busses [X86] Always assign all PCI bus
4168 numbers ourselves, overriding
4169 whatever the firmware may have done.
4170 usepirqmask [X86] Honor the possible IRQ mask stored
4171 in the BIOS $PIR table. This is needed on
4172 some systems with broken BIOSes, notably
4173 some HP Pavilion N5400 and Omnibook XE3
4174 notebooks. This will have no effect if ACPI
4175 IRQ routing is enabled.
4176 noacpi [X86] Do not use ACPI for IRQ routing
4177 or for PCI scanning.
4178 use_crs [X86] Use PCI host bridge window information
4179 from ACPI. On BIOSes from 2008 or later, this
4180 is enabled by default. If you need to use this,
4181 please report a bug.
4182 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4183 If you need to use this, please report a bug.
4184 use_e820 [X86] Use E820 reservations to exclude parts of
4185 PCI host bridge windows. This is a workaround
4186 for BIOS defects in host bridge _CRS methods.
4187 If you need to use this, please report a bug to
4188 <linux-pci@vger.kernel.org>.
4189 no_e820 [X86] Ignore E820 reservations for PCI host
4190 bridge windows. This is the default on modern
4191 hardware. If you need to use this, please report
4192 a bug to <linux-pci@vger.kernel.org>.
4193 routeirq Do IRQ routing for all PCI devices.
4194 This is normally done in pci_enable_device(),
4195 so this option is a temporary workaround
4196 for broken drivers that don't call it.
4197 skip_isa_align [X86] do not align io start addr, so can
4198 handle more pci cards
4199 noearly [X86] Don't do any early type 1 scanning.
4200 This might help on some broken boards which
4201 machine check when some devices' config space
4202 is read. But various workarounds are disabled
4203 and some IOMMU drivers will not work.
4204 bfsort Sort PCI devices into breadth-first order.
4205 This sorting is done to get a device
4206 order compatible with older (<= 2.4) kernels.
4207 nobfsort Don't sort PCI devices into breadth-first order.
4208 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4209 tuning and use the BIOS-configured MPS defaults.
4210 pcie_bus_safe Set every device's MPS to the largest value
4211 supported by all devices below the root complex.
4212 pcie_bus_perf Set device MPS to the largest allowable MPS
4213 based on its parent bus. Also set MRRS (Max
4214 Read Request Size) to the largest supported
4215 value (no larger than the MPS that the device
4216 or bus can support) for best performance.
4217 pcie_bus_peer2peer Set every device's MPS to 128B, which
4218 every device is guaranteed to support. This
4219 configuration allows peer-to-peer DMA between
4220 any pair of devices, possibly at the cost of
4221 reduced performance. This also guarantees
4222 that hot-added devices will work.
4223 cbiosize=nn[KMG] The fixed amount of bus space which is
4224 reserved for the CardBus bridge's IO window.
4225 The default value is 256 bytes.
4226 cbmemsize=nn[KMG] The fixed amount of bus space which is
4227 reserved for the CardBus bridge's memory
4228 window. The default value is 64 megabytes.
4231 [<order of align>@]<pci_dev>[; ...]
4232 Specifies alignment and device to reassign
4233 aligned memory resources. How to
4234 specify the device is described above.
4235 If <order of align> is not specified,
4236 PAGE_SIZE is used as alignment.
4237 A PCI-PCI bridge can be specified if resource
4238 windows need to be expanded.
4239 To specify the alignment for several
4240 instances of a device, the PCI vendor,
4241 device, subvendor, and subdevice may be
4242 specified, e.g., 12@pci:8086:9c22:103c:198f
4243 for 4096-byte alignment.
4244 ecrc= Enable/disable PCIe ECRC (transaction layer
4245 end-to-end CRC checking).
4246 bios: Use BIOS/firmware settings. This is the
4250 hpiosize=nn[KMG] The fixed amount of bus space which is
4251 reserved for hotplug bridge's IO window.
4252 Default size is 256 bytes.
4253 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4254 reserved for hotplug bridge's MMIO window.
4255 Default size is 2 megabytes.
4256 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4257 reserved for hotplug bridge's MMIO_PREF window.
4258 Default size is 2 megabytes.
4259 hpmemsize=nn[KMG] The fixed amount of bus space which is
4260 reserved for hotplug bridge's MMIO and
4262 Default size is 2 megabytes.
4263 hpbussize=nn The minimum amount of additional bus numbers
4264 reserved for buses below a hotplug bridge.
4266 realloc= Enable/disable reallocating PCI bridge resources
4267 if allocations done by BIOS are too small to
4268 accommodate resources required by all child
4270 off: Turn realloc off
4272 realloc same as realloc=on
4273 noari do not use PCIe ARI.
4274 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4275 do not use PCIe ATS (and IOMMU device IOTLB).
4276 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4277 only look for one device below a PCIe downstream
4279 big_root_window Try to add a big 64bit memory window to the PCIe
4280 root complex on AMD CPUs. Some GFX hardware
4281 can resize a BAR to allow access to all VRAM.
4282 Adding the window is slightly risky (it may
4283 conflict with unreported devices), so this
4285 disable_acs_redir=<pci_dev>[; ...]
4286 Specify one or more PCI devices (in the format
4287 specified above) separated by semicolons.
4288 Each device specified will have the PCI ACS
4289 redirect capabilities forced off which will
4290 allow P2P traffic between devices through
4291 bridges without forcing it upstream. Note:
4292 this removes isolation between devices and
4293 may put more devices in an IOMMU group.
4294 force_floating [S390] Force usage of floating interrupts.
4295 nomio [S390] Do not use MIO instructions.
4296 norid [S390] ignore the RID field and force use of
4297 one PCI domain per PCI function
4299 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4302 force Enable ASPM even on devices that claim not to support it.
4303 WARNING: Forcing ASPM on may cause system lockups.
4305 pcie_ports= [PCIE] PCIe port services handling:
4306 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4307 even if the platform doesn't give the OS permission to
4308 use them. This may cause conflicts if the platform
4309 also tries to use these services.
4310 dpc-native Use native PCIe service for DPC only. May
4311 cause conflicts if firmware uses AER or DPC.
4312 compat Disable native PCIe services (PME, AER, DPC, PCIe
4315 pcie_port_pm= [PCIE] PCIe port power management handling:
4316 off Disable power management of all PCIe ports
4317 force Forcibly enable power management of all PCIe ports
4319 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4320 nomsi Do not use MSI for native PCIe PME signaling (this makes
4321 all PCIe root ports use INTx for all services).
4323 pcmv= [HW,PCMCIA] BadgePAD 4
4327 Keep all power-domains already enabled by bootloader on,
4328 even if no driver has claimed them. This is useful
4329 for debug and development, but should not be
4330 needed on a platform with proper driver support.
4333 See Documentation/admin-guide/blockdev/paride.rst.
4335 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4338 See arch/parisc/kernel/pdc_chassis.c
4340 percpu_alloc= Select which percpu first chunk allocator to use.
4341 Currently supported values are "embed" and "page".
4342 Archs may support subset or none of the selections.
4343 See comments in mm/percpu.c for details on each
4344 allocator. This parameter is primarily for debugging
4345 and performance comparison.
4348 See Documentation/admin-guide/blockdev/paride.rst.
4351 See Documentation/admin-guide/blockdev/paride.rst.
4353 pirq= [SMP,APIC] Manual mp-table setup
4354 See Documentation/x86/i386/IO-APIC.rst.
4356 plip= [PPT,NET] Parallel port network link
4357 Format: { parport<nr> | timid | 0 }
4358 See also Documentation/admin-guide/parport.rst.
4360 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4361 Override pmtimer IOPort with a hex value.
4364 pmu_override= [PPC] Override the PMU.
4365 This option takes over the PMU facility, so it is no
4366 longer usable by perf. Setting this option starts the
4367 PMU counters by setting MMCR0 to 0 (the FC bit is
4368 cleared). If a number is given, then MMCR1 is set to
4369 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4372 pm_debug_messages [SUSPEND,KNL]
4373 Enable suspend/resume debug messages during boot up.
4376 Enable PNP debug messages (depends on the
4377 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4378 via /sys/module/pnp/parameters/debug. We always show
4379 current resource usage; turning this on also shows
4380 possible settings and some assignment information.
4386 { on | off | curr | res | no-curr | no-res }
4389 [ISAPNP] Exclude IRQs for the autoconfiguration
4392 [ISAPNP] Exclude DMAs for the autoconfiguration
4394 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4395 Ranges are in pairs (I/O port base and size).
4398 [ISAPNP] Exclude memory regions for the
4400 Ranges are in pairs (memory base and size).
4402 ports= [IP_VS_FTP] IPVS ftp helper module
4404 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4406 Format: <port>,<port>....
4408 powersave=off [PPC] This option disables power saving features.
4409 It specifically disables cpuidle and sets the
4410 platform machine description specific power_save
4411 function to NULL. On Idle the CPU just reduces
4414 ppc_strict_facility_enable
4415 [PPC] This option catches any kernel floating point,
4416 Altivec, VSX and SPE outside of regions specifically
4417 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4418 There is some performance impact when enabling this.
4422 Disable Hardware Transactional Memory
4425 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4426 none - Limited to cond_resched() calls
4427 voluntary - Limited to cond_resched() and might_sleep() calls
4428 full - Any section that isn't explicitly preempt disabled
4429 can be preempted anytime.
4431 print-fatal-signals=
4432 [KNL] debug: print fatal signals
4434 If enabled, warn about various signal handling
4435 related application anomalies: too many signals,
4436 too many POSIX.1 timers, fatal signals causing a
4439 If you hit the warning due to signal overflow,
4440 you might want to try "ulimit -i unlimited".
4444 printk.always_kmsg_dump=
4445 Trigger kmsg_dump for cases other than kernel oops or
4447 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4450 printk.console_no_auto_verbose=
4451 Disable console loglevel raise on oops, panic
4452 or lockdep-detected issues (only if lock debug is on).
4453 With an exception to setups with low baudrate on
4454 serial console, keeping this 0 is a good choice
4455 in order to provide more debug information.
4457 default: 0 (auto_verbose is enabled)
4459 printk.devkmsg={on,off,ratelimit}
4460 Control writing to /dev/kmsg.
4461 on - unlimited logging to /dev/kmsg from userspace
4462 off - logging to /dev/kmsg disabled
4463 ratelimit - ratelimit the logging
4466 printk.time= Show timing data prefixed to each printk message line
4467 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4469 processor.max_cstate= [HW,ACPI]
4470 Limit processor to maximum C-state
4471 max_cstate=9 overrides any DMI blacklist limit.
4473 processor.nocst [HW,ACPI]
4474 Ignore the _CST method to determine C-states,
4475 instead using the legacy FADT method
4477 profile= [KNL] Enable kernel profiling via /proc/profile
4478 Format: [<profiletype>,]<number>
4479 Param: <profiletype>: "schedule", "sleep", or "kvm"
4480 [defaults to kernel profiling]
4481 Param: "schedule" - profile schedule points.
4482 Param: "sleep" - profile D-state sleeping (millisecs).
4483 Requires CONFIG_SCHEDSTATS
4484 Param: "kvm" - profile VM exits.
4485 Param: <number> - step/bucket size as a power of 2 for
4486 statistical time based profiling.
4488 prompt_ramdisk= [RAM] [Deprecated]
4490 prot_virt= [S390] enable hosting protected virtual machines
4491 isolated from the hypervisor (if hardware supports
4495 psi= [KNL] Enable or disable pressure stall information
4499 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4500 probe for; one of (bare|imps|exps|lifebook|any).
4501 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4503 psmouse.resetafter= [HW,MOUSE]
4504 Try to reset the device after so many bad packets
4507 [HW,MOUSE] Set desired mouse resolution, in dpi.
4508 psmouse.smartscroll=
4509 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4510 0 = disabled, 1 = enabled (default).
4512 pstore.backend= Specify the name of the pstore backend to use
4515 See Documentation/admin-guide/blockdev/paride.rst.
4517 pti= [X86-64] Control Page Table Isolation of user and
4518 kernel address spaces. Disabling this feature
4519 removes hardening, but improves performance of
4520 system calls and interrupts.
4522 on - unconditionally enable
4523 off - unconditionally disable
4524 auto - kernel detects whether your CPU model is
4525 vulnerable to issues that PTI mitigates
4527 Not specifying this option is equivalent to pti=auto.
4530 Equivalent to pti=off
4533 [KNL] Number of legacy pty's. Overwrites compiled-in
4536 quiet [KNL] Disable most log messages
4541 See Documentation/admin-guide/md.rst.
4543 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4544 See Documentation/admin-guide/blockdev/ramdisk.rst.
4546 ramdisk_start= [RAM] RAM disk image start address
4548 random.trust_cpu={on,off}
4549 [KNL] Enable or disable trusting the use of the
4550 CPU's random number generator (if available) to
4551 fully seed the kernel's CRNG. Default is controlled
4552 by CONFIG_RANDOM_TRUST_CPU.
4554 random.trust_bootloader={on,off}
4555 [KNL] Enable or disable trusting the use of a
4556 seed passed by the bootloader (if available) to
4557 fully seed the kernel's CRNG. Default is controlled
4558 by CONFIG_RANDOM_TRUST_BOOTLOADER.
4560 randomize_kstack_offset=
4561 [KNL] Enable or disable kernel stack offset
4562 randomization, which provides roughly 5 bits of
4563 entropy, frustrating memory corruption attacks
4564 that depend on stack address determinism or
4565 cross-syscall address exposures. This is only
4566 available on architectures that have defined
4567 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4568 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4569 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4571 ras=option[,option,...] [KNL] RAS-specific options
4574 Disable the Correctable Errors Collector,
4575 see CONFIG_RAS_CEC help text.
4577 rcu_nocbs[=cpu-list]
4578 [KNL] The optional argument is a cpu list,
4581 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4582 enable the no-callback CPU mode, which prevents
4583 such CPUs' callbacks from being invoked in
4584 softirq context. Invocation of such CPUs' RCU
4585 callbacks will instead be offloaded to "rcuox/N"
4586 kthreads created for that purpose, where "x" is
4587 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4588 for the kthreads that mediate grace periods; and
4589 "N" is the CPU number. This reduces OS jitter on
4590 the offloaded CPUs, which can be useful for HPC
4591 and real-time workloads. It can also improve
4592 energy efficiency for asymmetric multiprocessors.
4594 If a cpulist is passed as an argument, the specified
4595 list of CPUs is set to no-callback mode from boot.
4597 Otherwise, if the '=' sign and the cpulist
4598 arguments are omitted, no CPU will be set to
4599 no-callback mode from boot but the mode may be
4600 toggled at runtime via cpusets.
4602 Note that this argument takes precedence over
4603 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4606 Rather than requiring that offloaded CPUs
4607 (specified by rcu_nocbs= above) explicitly
4608 awaken the corresponding "rcuoN" kthreads,
4609 make these kthreads poll for callbacks.
4610 This improves the real-time response for the
4611 offloaded CPUs by relieving them of the need to
4612 wake up the corresponding kthread, but degrades
4613 energy efficiency by requiring that the kthreads
4614 periodically wake up to do the polling.
4616 rcutree.blimit= [KNL]
4617 Set maximum number of finished RCU callbacks to
4618 process in one batch.
4620 rcutree.dump_tree= [KNL]
4621 Dump the structure of the rcu_node combining tree
4622 out at early boot. This is used for diagnostic
4623 purposes, to verify correct tree setup.
4625 rcutree.gp_cleanup_delay= [KNL]
4626 Set the number of jiffies to delay each step of
4627 RCU grace-period cleanup.
4629 rcutree.gp_init_delay= [KNL]
4630 Set the number of jiffies to delay each step of
4631 RCU grace-period initialization.
4633 rcutree.gp_preinit_delay= [KNL]
4634 Set the number of jiffies to delay each step of
4635 RCU grace-period pre-initialization, that is,
4636 the propagation of recent CPU-hotplug changes up
4637 the rcu_node combining tree.
4639 rcutree.use_softirq= [KNL]
4640 If set to zero, move all RCU_SOFTIRQ processing to
4641 per-CPU rcuc kthreads. Defaults to a non-zero
4642 value, meaning that RCU_SOFTIRQ is used by default.
4643 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4645 But note that CONFIG_PREEMPT_RT=y kernels disable
4646 this kernel boot parameter, forcibly setting it
4649 rcutree.rcu_fanout_exact= [KNL]
4650 Disable autobalancing of the rcu_node combining
4651 tree. This is used by rcutorture, and might
4652 possibly be useful for architectures having high
4653 cache-to-cache transfer latencies.
4655 rcutree.rcu_fanout_leaf= [KNL]
4656 Change the number of CPUs assigned to each
4657 leaf rcu_node structure. Useful for very
4658 large systems, which will choose the value 64,
4659 and for NUMA systems with large remote-access
4660 latencies, which will choose a value aligned
4661 with the appropriate hardware boundaries.
4663 rcutree.rcu_min_cached_objs= [KNL]
4664 Minimum number of objects which are cached and
4665 maintained per one CPU. Object size is equal
4666 to PAGE_SIZE. The cache allows to reduce the
4667 pressure to page allocator, also it makes the
4668 whole algorithm to behave better in low memory
4671 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4672 Set the page-cache refill delay (in milliseconds)
4673 in response to low-memory conditions. The range
4674 of permitted values is in the range 0:100000.
4676 rcutree.jiffies_till_first_fqs= [KNL]
4677 Set delay from grace-period initialization to
4678 first attempt to force quiescent states.
4679 Units are jiffies, minimum value is zero,
4680 and maximum value is HZ.
4682 rcutree.jiffies_till_next_fqs= [KNL]
4683 Set delay between subsequent attempts to force
4684 quiescent states. Units are jiffies, minimum
4685 value is one, and maximum value is HZ.
4687 rcutree.jiffies_till_sched_qs= [KNL]
4688 Set required age in jiffies for a
4689 given grace period before RCU starts
4690 soliciting quiescent-state help from
4691 rcu_note_context_switch() and cond_resched().
4692 If not specified, the kernel will calculate
4693 a value based on the most recent settings
4694 of rcutree.jiffies_till_first_fqs
4695 and rcutree.jiffies_till_next_fqs.
4696 This calculated value may be viewed in
4697 rcutree.jiffies_to_sched_qs. Any attempt to set
4698 rcutree.jiffies_to_sched_qs will be cheerfully
4701 rcutree.kthread_prio= [KNL,BOOT]
4702 Set the SCHED_FIFO priority of the RCU per-CPU
4703 kthreads (rcuc/N). This value is also used for
4704 the priority of the RCU boost threads (rcub/N)
4705 and for the RCU grace-period kthreads (rcu_bh,
4706 rcu_preempt, and rcu_sched). If RCU_BOOST is
4707 set, valid values are 1-99 and the default is 1
4708 (the least-favored priority). Otherwise, when
4709 RCU_BOOST is not set, valid values are 0-99 and
4710 the default is zero (non-realtime operation).
4711 When RCU_NOCB_CPU is set, also adjust the
4712 priority of NOCB callback kthreads.
4714 rcutree.rcu_divisor= [KNL]
4715 Set the shift-right count to use to compute
4716 the callback-invocation batch limit bl from
4717 the number of callbacks queued on this CPU.
4718 The result will be bounded below by the value of
4719 the rcutree.blimit kernel parameter. Every bl
4720 callbacks, the softirq handler will exit in
4721 order to allow the CPU to do other work.
4723 Please note that this callback-invocation batch
4724 limit applies only to non-offloaded callback
4725 invocation. Offloaded callbacks are instead
4726 invoked in the context of an rcuoc kthread, which
4727 scheduler will preempt as it does any other task.
4729 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4730 On callback-offloaded (rcu_nocbs) CPUs,
4731 RCU reduces the lock contention that would
4732 otherwise be caused by callback floods through
4733 use of the ->nocb_bypass list. However, in the
4734 common non-flooded case, RCU queues directly to
4735 the main ->cblist in order to avoid the extra
4736 overhead of the ->nocb_bypass list and its lock.
4737 But if there are too many callbacks queued during
4738 a single jiffy, RCU pre-queues the callbacks into
4739 the ->nocb_bypass queue. The definition of "too
4740 many" is supplied by this kernel boot parameter.
4742 rcutree.rcu_nocb_gp_stride= [KNL]
4743 Set the number of NOCB callback kthreads in
4744 each group, which defaults to the square root
4745 of the number of CPUs. Larger numbers reduce
4746 the wakeup overhead on the global grace-period
4747 kthread, but increases that same overhead on
4748 each group's NOCB grace-period kthread.
4750 rcutree.qhimark= [KNL]
4751 Set threshold of queued RCU callbacks beyond which
4752 batch limiting is disabled.
4754 rcutree.qlowmark= [KNL]
4755 Set threshold of queued RCU callbacks below which
4756 batch limiting is re-enabled.
4758 rcutree.qovld= [KNL]
4759 Set threshold of queued RCU callbacks beyond which
4760 RCU's force-quiescent-state scan will aggressively
4761 enlist help from cond_resched() and sched IPIs to
4762 help CPUs more quickly reach quiescent states.
4763 Set to less than zero to make this be set based
4764 on rcutree.qhimark at boot time and to zero to
4765 disable more aggressive help enlistment.
4767 rcutree.rcu_kick_kthreads= [KNL]
4768 Cause the grace-period kthread to get an extra
4769 wake_up() if it sleeps three times longer than
4770 it should at force-quiescent-state time.
4771 This wake_up() will be accompanied by a
4772 WARN_ONCE() splat and an ftrace_dump().
4774 rcutree.rcu_unlock_delay= [KNL]
4775 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4776 this specifies an rcu_read_unlock()-time delay
4777 in microseconds. This defaults to zero.
4778 Larger delays increase the probability of
4779 catching RCU pointer leaks, that is, buggy use
4780 of RCU-protected pointers after the relevant
4781 rcu_read_unlock() has completed.
4783 rcutree.sysrq_rcu= [KNL]
4784 Commandeer a sysrq key to dump out Tree RCU's
4785 rcu_node tree with an eye towards determining
4786 why a new grace period has not yet started.
4788 rcuscale.gp_async= [KNL]
4789 Measure performance of asynchronous
4790 grace-period primitives such as call_rcu().
4792 rcuscale.gp_async_max= [KNL]
4793 Specify the maximum number of outstanding
4794 callbacks per writer thread. When a writer
4795 thread exceeds this limit, it invokes the
4796 corresponding flavor of rcu_barrier() to allow
4797 previously posted callbacks to drain.
4799 rcuscale.gp_exp= [KNL]
4800 Measure performance of expedited synchronous
4801 grace-period primitives.
4803 rcuscale.holdoff= [KNL]
4804 Set test-start holdoff period. The purpose of
4805 this parameter is to delay the start of the
4806 test until boot completes in order to avoid
4809 rcuscale.kfree_rcu_test= [KNL]
4810 Set to measure performance of kfree_rcu() flooding.
4812 rcuscale.kfree_rcu_test_double= [KNL]
4813 Test the double-argument variant of kfree_rcu().
4814 If this parameter has the same value as
4815 rcuscale.kfree_rcu_test_single, both the single-
4816 and double-argument variants are tested.
4818 rcuscale.kfree_rcu_test_single= [KNL]
4819 Test the single-argument variant of kfree_rcu().
4820 If this parameter has the same value as
4821 rcuscale.kfree_rcu_test_double, both the single-
4822 and double-argument variants are tested.
4824 rcuscale.kfree_nthreads= [KNL]
4825 The number of threads running loops of kfree_rcu().
4827 rcuscale.kfree_alloc_num= [KNL]
4828 Number of allocations and frees done in an iteration.
4830 rcuscale.kfree_loops= [KNL]
4831 Number of loops doing rcuscale.kfree_alloc_num number
4832 of allocations and frees.
4834 rcuscale.nreaders= [KNL]
4835 Set number of RCU readers. The value -1 selects
4836 N, where N is the number of CPUs. A value
4837 "n" less than -1 selects N-n+1, where N is again
4838 the number of CPUs. For example, -2 selects N
4839 (the number of CPUs), -3 selects N+1, and so on.
4840 A value of "n" less than or equal to -N selects
4843 rcuscale.nwriters= [KNL]
4844 Set number of RCU writers. The values operate
4845 the same as for rcuscale.nreaders.
4846 N, where N is the number of CPUs
4848 rcuscale.perf_type= [KNL]
4849 Specify the RCU implementation to test.
4851 rcuscale.shutdown= [KNL]
4852 Shut the system down after performance tests
4853 complete. This is useful for hands-off automated
4856 rcuscale.verbose= [KNL]
4857 Enable additional printk() statements.
4859 rcuscale.writer_holdoff= [KNL]
4860 Write-side holdoff between grace periods,
4861 in microseconds. The default of zero says
4864 rcutorture.fqs_duration= [KNL]
4865 Set duration of force_quiescent_state bursts
4868 rcutorture.fqs_holdoff= [KNL]
4869 Set holdoff time within force_quiescent_state bursts
4872 rcutorture.fqs_stutter= [KNL]
4873 Set wait time between force_quiescent_state bursts
4876 rcutorture.fwd_progress= [KNL]
4877 Specifies the number of kthreads to be used
4878 for RCU grace-period forward-progress testing
4879 for the types of RCU supporting this notion.
4880 Defaults to 1 kthread, values less than zero or
4881 greater than the number of CPUs cause the number
4884 rcutorture.fwd_progress_div= [KNL]
4885 Specify the fraction of a CPU-stall-warning
4886 period to do tight-loop forward-progress testing.
4888 rcutorture.fwd_progress_holdoff= [KNL]
4889 Number of seconds to wait between successive
4890 forward-progress tests.
4892 rcutorture.fwd_progress_need_resched= [KNL]
4893 Enclose cond_resched() calls within checks for
4894 need_resched() during tight-loop forward-progress
4897 rcutorture.gp_cond= [KNL]
4898 Use conditional/asynchronous update-side
4899 primitives, if available.
4901 rcutorture.gp_exp= [KNL]
4902 Use expedited update-side primitives, if available.
4904 rcutorture.gp_normal= [KNL]
4905 Use normal (non-expedited) asynchronous
4906 update-side primitives, if available.
4908 rcutorture.gp_sync= [KNL]
4909 Use normal (non-expedited) synchronous
4910 update-side primitives, if available. If all
4911 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4912 rcutorture.gp_normal=, and rcutorture.gp_sync=
4913 are zero, rcutorture acts as if is interpreted
4914 they are all non-zero.
4916 rcutorture.irqreader= [KNL]
4917 Run RCU readers from irq handlers, or, more
4918 accurately, from a timer handler. Not all RCU
4919 flavors take kindly to this sort of thing.
4921 rcutorture.leakpointer= [KNL]
4922 Leak an RCU-protected pointer out of the reader.
4923 This can of course result in splats, and is
4924 intended to test the ability of things like
4925 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4928 rcutorture.n_barrier_cbs= [KNL]
4929 Set callbacks/threads for rcu_barrier() testing.
4931 rcutorture.nfakewriters= [KNL]
4932 Set number of concurrent RCU writers. These just
4933 stress RCU, they don't participate in the actual
4934 test, hence the "fake".
4936 rcutorture.nocbs_nthreads= [KNL]
4937 Set number of RCU callback-offload togglers.
4938 Zero (the default) disables toggling.
4940 rcutorture.nocbs_toggle= [KNL]
4941 Set the delay in milliseconds between successive
4942 callback-offload toggling attempts.
4944 rcutorture.nreaders= [KNL]
4945 Set number of RCU readers. The value -1 selects
4946 N-1, where N is the number of CPUs. A value
4947 "n" less than -1 selects N-n-2, where N is again
4948 the number of CPUs. For example, -2 selects N
4949 (the number of CPUs), -3 selects N+1, and so on.
4951 rcutorture.object_debug= [KNL]
4952 Enable debug-object double-call_rcu() testing.
4954 rcutorture.onoff_holdoff= [KNL]
4955 Set time (s) after boot for CPU-hotplug testing.
4957 rcutorture.onoff_interval= [KNL]
4958 Set time (jiffies) between CPU-hotplug operations,
4959 or zero to disable CPU-hotplug testing.
4961 rcutorture.read_exit= [KNL]
4962 Set the number of read-then-exit kthreads used
4963 to test the interaction of RCU updaters and
4964 task-exit processing.
4966 rcutorture.read_exit_burst= [KNL]
4967 The number of times in a given read-then-exit
4968 episode that a set of read-then-exit kthreads
4971 rcutorture.read_exit_delay= [KNL]
4972 The delay, in seconds, between successive
4973 read-then-exit testing episodes.
4975 rcutorture.shuffle_interval= [KNL]
4976 Set task-shuffle interval (s). Shuffling tasks
4977 allows some CPUs to go into dyntick-idle mode
4978 during the rcutorture test.
4980 rcutorture.shutdown_secs= [KNL]
4981 Set time (s) after boot system shutdown. This
4982 is useful for hands-off automated testing.
4984 rcutorture.stall_cpu= [KNL]
4985 Duration of CPU stall (s) to test RCU CPU stall
4986 warnings, zero to disable.
4988 rcutorture.stall_cpu_block= [KNL]
4989 Sleep while stalling if set. This will result
4990 in warnings from preemptible RCU in addition
4991 to any other stall-related activity.
4993 rcutorture.stall_cpu_holdoff= [KNL]
4994 Time to wait (s) after boot before inducing stall.
4996 rcutorture.stall_cpu_irqsoff= [KNL]
4997 Disable interrupts while stalling if set.
4999 rcutorture.stall_gp_kthread= [KNL]
5000 Duration (s) of forced sleep within RCU
5001 grace-period kthread to test RCU CPU stall
5002 warnings, zero to disable. If both stall_cpu
5003 and stall_gp_kthread are specified, the
5004 kthread is starved first, then the CPU.
5006 rcutorture.stat_interval= [KNL]
5007 Time (s) between statistics printk()s.
5009 rcutorture.stutter= [KNL]
5010 Time (s) to stutter testing, for example, specifying
5011 five seconds causes the test to run for five seconds,
5012 wait for five seconds, and so on. This tests RCU's
5013 ability to transition abruptly to and from idle.
5015 rcutorture.test_boost= [KNL]
5016 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5017 "Maybe" means test if the RCU implementation
5018 under test support RCU priority boosting.
5020 rcutorture.test_boost_duration= [KNL]
5021 Duration (s) of each individual boost test.
5023 rcutorture.test_boost_interval= [KNL]
5024 Interval (s) between each boost test.
5026 rcutorture.test_no_idle_hz= [KNL]
5027 Test RCU's dyntick-idle handling. See also the
5028 rcutorture.shuffle_interval parameter.
5030 rcutorture.torture_type= [KNL]
5031 Specify the RCU implementation to test.
5033 rcutorture.verbose= [KNL]
5034 Enable additional printk() statements.
5036 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5037 Dump ftrace buffer after reporting RCU CPU
5040 rcupdate.rcu_cpu_stall_suppress= [KNL]
5041 Suppress RCU CPU stall warning messages.
5043 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5044 Suppress RCU CPU stall warning messages and
5045 rcutorture writer stall warnings that occur
5046 during early boot, that is, during the time
5047 before the init task is spawned.
5049 rcupdate.rcu_cpu_stall_timeout= [KNL]
5050 Set timeout for RCU CPU stall warning messages.
5051 The value is in seconds and the maximum allowed
5052 value is 300 seconds.
5054 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5055 Set timeout for expedited RCU CPU stall warning
5056 messages. The value is in milliseconds
5057 and the maximum allowed value is 21000
5058 milliseconds. Please note that this value is
5059 adjusted to an arch timer tick resolution.
5060 Setting this to zero causes the value from
5061 rcupdate.rcu_cpu_stall_timeout to be used (after
5062 conversion from seconds to milliseconds).
5064 rcupdate.rcu_expedited= [KNL]
5065 Use expedited grace-period primitives, for
5066 example, synchronize_rcu_expedited() instead
5067 of synchronize_rcu(). This reduces latency,
5068 but can increase CPU utilization, degrade
5069 real-time latency, and degrade energy efficiency.
5070 No effect on CONFIG_TINY_RCU kernels.
5072 rcupdate.rcu_normal= [KNL]
5073 Use only normal grace-period primitives,
5074 for example, synchronize_rcu() instead of
5075 synchronize_rcu_expedited(). This improves
5076 real-time latency, CPU utilization, and
5077 energy efficiency, but can expose users to
5078 increased grace-period latency. This parameter
5079 overrides rcupdate.rcu_expedited. No effect on
5080 CONFIG_TINY_RCU kernels.
5082 rcupdate.rcu_normal_after_boot= [KNL]
5083 Once boot has completed (that is, after
5084 rcu_end_inkernel_boot() has been invoked), use
5085 only normal grace-period primitives. No effect
5086 on CONFIG_TINY_RCU kernels.
5088 But note that CONFIG_PREEMPT_RT=y kernels enables
5089 this kernel boot parameter, forcibly setting
5090 it to the value one, that is, converting any
5091 post-boot attempt at an expedited RCU grace
5092 period to instead use normal non-expedited
5093 grace-period processing.
5095 rcupdate.rcu_task_collapse_lim= [KNL]
5096 Set the maximum number of callbacks present
5097 at the beginning of a grace period that allows
5098 the RCU Tasks flavors to collapse back to using
5099 a single callback queue. This switching only
5100 occurs when rcupdate.rcu_task_enqueue_lim is
5101 set to the default value of -1.
5103 rcupdate.rcu_task_contend_lim= [KNL]
5104 Set the minimum number of callback-queuing-time
5105 lock-contention events per jiffy required to
5106 cause the RCU Tasks flavors to switch to per-CPU
5107 callback queuing. This switching only occurs
5108 when rcupdate.rcu_task_enqueue_lim is set to
5109 the default value of -1.
5111 rcupdate.rcu_task_enqueue_lim= [KNL]
5112 Set the number of callback queues to use for the
5113 RCU Tasks family of RCU flavors. The default
5114 of -1 allows this to be automatically (and
5115 dynamically) adjusted. This parameter is intended
5118 rcupdate.rcu_task_ipi_delay= [KNL]
5119 Set time in jiffies during which RCU tasks will
5120 avoid sending IPIs, starting with the beginning
5121 of a given grace period. Setting a large
5122 number avoids disturbing real-time workloads,
5123 but lengthens grace periods.
5125 rcupdate.rcu_task_stall_info= [KNL]
5126 Set initial timeout in jiffies for RCU task stall
5127 informational messages, which give some indication
5128 of the problem for those not patient enough to
5129 wait for ten minutes. Informational messages are
5130 only printed prior to the stall-warning message
5131 for a given grace period. Disable with a value
5132 less than or equal to zero. Defaults to ten
5133 seconds. A change in value does not take effect
5134 until the beginning of the next grace period.
5136 rcupdate.rcu_task_stall_info_mult= [KNL]
5137 Multiplier for time interval between successive
5138 RCU task stall informational messages for a given
5139 RCU tasks grace period. This value is clamped
5140 to one through ten, inclusive. It defaults to
5141 the value three, so that the first informational
5142 message is printed 10 seconds into the grace
5143 period, the second at 40 seconds, the third at
5144 160 seconds, and then the stall warning at 600
5145 seconds would prevent a fourth at 640 seconds.
5147 rcupdate.rcu_task_stall_timeout= [KNL]
5148 Set timeout in jiffies for RCU task stall
5149 warning messages. Disable with a value less
5150 than or equal to zero. Defaults to ten minutes.
5151 A change in value does not take effect until
5152 the beginning of the next grace period.
5154 rcupdate.rcu_self_test= [KNL]
5155 Run the RCU early boot self tests
5159 Run specified binary instead of /init from the ramdisk,
5160 used for early userspace startup. See initrd.
5163 force - Override the decision by the kernel to hide the
5164 advertisement of RDRAND support (this affects
5165 certain AMD processors because of buggy BIOS
5166 support, specifically around the suspend/resume
5170 Turn on/off individual RDT features. List is:
5171 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5173 E.g. to turn on cmt and turn off mba use:
5177 Format (x86 or x86_64):
5178 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5180 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5182 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5183 (prefix with 'panic_' to set mode for panic
5185 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5186 reboot_force is either force or not specified,
5187 reboot_cpu is s[mp]#### with #### being the processor
5188 to be used for rebooting.
5190 refscale.holdoff= [KNL]
5191 Set test-start holdoff period. The purpose of
5192 this parameter is to delay the start of the
5193 test until boot completes in order to avoid
5196 refscale.loops= [KNL]
5197 Set the number of loops over the synchronization
5198 primitive under test. Increasing this number
5199 reduces noise due to loop start/end overhead,
5200 but the default has already reduced the per-pass
5201 noise to a handful of picoseconds on ca. 2020
5204 refscale.nreaders= [KNL]
5205 Set number of readers. The default value of -1
5206 selects N, where N is roughly 75% of the number
5207 of CPUs. A value of zero is an interesting choice.
5209 refscale.nruns= [KNL]
5210 Set number of runs, each of which is dumped onto
5213 refscale.readdelay= [KNL]
5214 Set the read-side critical-section duration,
5215 measured in microseconds.
5217 refscale.scale_type= [KNL]
5218 Specify the read-protection implementation to test.
5220 refscale.shutdown= [KNL]
5221 Shut down the system at the end of the performance
5222 test. This defaults to 1 (shut it down) when
5223 refscale is built into the kernel and to 0 (leave
5224 it running) when refscale is built as a module.
5226 refscale.verbose= [KNL]
5227 Enable additional printk() statements.
5229 refscale.verbose_batched= [KNL]
5230 Batch the additional printk() statements. If zero
5231 (the default) or negative, print everything. Otherwise,
5232 print every Nth verbose statement, where N is the value
5236 [KNL, SMP] Set scheduler's default relax_domain_level.
5237 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5239 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5240 Format: <base1>,<size1>[,<base2>,<size2>,...]
5241 Reserve I/O ports or memory so the kernel won't use
5242 them. If <base> is less than 0x10000, the region
5243 is assumed to be I/O ports; otherwise it is memory.
5245 reservetop= [X86-32]
5247 Reserves a hole at the top of the kernel virtual
5250 reset_devices [KNL] Force drivers to reset the underlying device
5251 during initialization.
5254 Specify the partition device for software suspend
5256 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5258 resume_offset= [SWSUSP]
5259 Specify the offset from the beginning of the partition
5260 given by "resume=" at which the swap header is located,
5261 in <PAGE_SIZE> units (needed only for swap files).
5262 See Documentation/power/swsusp-and-swap-files.rst
5264 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5265 read the resume files
5267 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5268 Useful for devices that are detected asynchronously
5269 (e.g. USB and MMC devices).
5271 retain_initrd [RAM] Keep initrd memory after extraction
5273 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5274 Speculative Code Execution with Return Instructions)
5277 AMD-based UNRET and IBPB mitigations alone do not stop
5278 sibling threads from influencing the predictions of other
5279 sibling threads. For that reason, STIBP is used on pro-
5280 cessors that support it, and mitigate SMT on processors
5284 auto - automatically select a migitation
5285 auto,nosmt - automatically select a mitigation,
5286 disabling SMT if necessary for
5287 the full mitigation (only on Zen1
5288 and older without STIBP).
5289 ibpb - On AMD, mitigate short speculation
5290 windows on basic block boundaries too.
5291 Safe, highest perf impact. It also
5292 enables STIBP if present. Not suitable
5294 ibpb,nosmt - Like "ibpb" above but will disable SMT
5295 when STIBP is not available. This is
5296 the alternative for systems which do not
5298 unret - Force enable untrained return thunks,
5299 only effective on AMD f15h-f17h based
5301 unret,nosmt - Like unret, but will disable SMT when STIBP
5302 is not available. This is the alternative for
5303 systems which do not have STIBP.
5305 Selecting 'auto' will choose a mitigation method at run
5306 time according to the CPU.
5308 Not specifying this option is equivalent to retbleed=auto.
5310 rfkill.default_state=
5311 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5312 etc. communication is blocked by default.
5315 rfkill.master_switch_mode=
5316 0 The "airplane mode" button does nothing.
5317 1 The "airplane mode" button toggles between everything
5318 blocked and the previous configuration.
5319 2 The "airplane mode" button toggles between everything
5320 blocked and everything unblocked.
5322 rhash_entries= [KNL,NET]
5323 Set number of hash buckets for route cache
5326 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5329 ro [KNL] Mount root device read-only on boot
5332 on Mark read-only kernel memory as read-only (default).
5333 off Leave read-only kernel memory writable for debugging.
5336 Enable the uart passthrough on the designated usb port
5337 on Rockchip SoCs. When active, the signals of the
5338 debug-uart get routed to the D+ and D- pins of the usb
5339 port and the regular usb controller gets disabled.
5341 root= [KNL] Root filesystem
5342 See name_to_dev_t comment in init/do_mounts.c.
5344 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5345 mount the root filesystem
5347 rootflags= [KNL] Set root filesystem mount option string
5349 rootfstype= [KNL] Set root filesystem type
5351 rootwait [KNL] Wait (indefinitely) for root device to show up.
5352 Useful for devices that are detected asynchronously
5353 (e.g. USB and MMC devices).
5355 rproc_mem=nn[KMG][@address]
5356 [KNL,ARM,CMA] Remoteproc physical memory block.
5357 Memory area to be used by remote processor image,
5360 rw [KNL] Mount root device read-write on boot
5362 S [KNL] Run init in single mode
5364 s390_iommu= [HW,S390]
5365 Set s390 IOTLB flushing mode
5367 With strict flushing every unmap operation will result in
5368 an IOTLB flush. Default is lazy flushing before reuse,
5371 s390_iommu_aperture= [KNL,S390]
5372 Specifies the size of the per device DMA address space
5373 accessible through the DMA and IOMMU APIs as a decimal
5374 factor of the size of main memory.
5375 The default is 1 meaning that one can concurrently use
5376 as many DMA addresses as physical memory is installed,
5377 if supported by hardware, and thus map all of memory
5378 once. With a value of 2 one can map all of memory twice
5379 and so on. As a special case a factor of 0 imposes no
5380 restrictions other than those given by hardware at the
5381 cost of significant additional memory use for tables.
5384 See drivers/net/irda/sa1100_ir.c.
5386 sched_verbose [KNL] Enables verbose scheduler debug messages.
5388 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5389 Allowed values are enable and disable. This feature
5390 incurs a small amount of overhead in the scheduler
5391 but is useful for debugging and performance tuning.
5393 sched_thermal_decay_shift=
5394 [KNL, SMP] Set a decay shift for scheduler thermal
5395 pressure signal. Thermal pressure signal follows the
5396 default decay period of other scheduler pelt
5397 signals(usually 32 ms but configurable). Setting
5398 sched_thermal_decay_shift will left shift the decay
5399 period for the thermal pressure signal by the shift
5401 i.e. with the default pelt decay period of 32 ms
5402 sched_thermal_decay_shift thermal pressure decay pr
5406 Format: integer between 0 and 10
5409 scftorture.holdoff= [KNL]
5410 Number of seconds to hold off before starting
5411 test. Defaults to zero for module insertion and
5412 to 10 seconds for built-in smp_call_function()
5415 scftorture.longwait= [KNL]
5416 Request ridiculously long waits randomly selected
5417 up to the chosen limit in seconds. Zero (the
5418 default) disables this feature. Please note
5419 that requesting even small non-zero numbers of
5420 seconds can result in RCU CPU stall warnings,
5421 softlockup complaints, and so on.
5423 scftorture.nthreads= [KNL]
5424 Number of kthreads to spawn to invoke the
5425 smp_call_function() family of functions.
5426 The default of -1 specifies a number of kthreads
5427 equal to the number of CPUs.
5429 scftorture.onoff_holdoff= [KNL]
5430 Number seconds to wait after the start of the
5431 test before initiating CPU-hotplug operations.
5433 scftorture.onoff_interval= [KNL]
5434 Number seconds to wait between successive
5435 CPU-hotplug operations. Specifying zero (which
5436 is the default) disables CPU-hotplug operations.
5438 scftorture.shutdown_secs= [KNL]
5439 The number of seconds following the start of the
5440 test after which to shut down the system. The
5441 default of zero avoids shutting down the system.
5442 Non-zero values are useful for automated tests.
5444 scftorture.stat_interval= [KNL]
5445 The number of seconds between outputting the
5446 current test statistics to the console. A value
5447 of zero disables statistics output.
5449 scftorture.stutter_cpus= [KNL]
5450 The number of jiffies to wait between each change
5451 to the set of CPUs under test.
5453 scftorture.use_cpus_read_lock= [KNL]
5454 Use use_cpus_read_lock() instead of the default
5455 preempt_disable() to disable CPU hotplug
5456 while invoking one of the smp_call_function*()
5459 scftorture.verbose= [KNL]
5460 Enable additional printk() statements.
5462 scftorture.weight_single= [KNL]
5463 The probability weighting to use for the
5464 smp_call_function_single() function with a zero
5465 "wait" parameter. A value of -1 selects the
5466 default if all other weights are -1. However,
5467 if at least one weight has some other value, a
5468 value of -1 will instead select a weight of zero.
5470 scftorture.weight_single_wait= [KNL]
5471 The probability weighting to use for the
5472 smp_call_function_single() function with a
5473 non-zero "wait" parameter. See weight_single.
5475 scftorture.weight_many= [KNL]
5476 The probability weighting to use for the
5477 smp_call_function_many() function with a zero
5478 "wait" parameter. See weight_single.
5479 Note well that setting a high probability for
5480 this weighting can place serious IPI load
5483 scftorture.weight_many_wait= [KNL]
5484 The probability weighting to use for the
5485 smp_call_function_many() function with a
5486 non-zero "wait" parameter. See weight_single
5489 scftorture.weight_all= [KNL]
5490 The probability weighting to use for the
5491 smp_call_function_all() function with a zero
5492 "wait" parameter. See weight_single and
5495 scftorture.weight_all_wait= [KNL]
5496 The probability weighting to use for the
5497 smp_call_function_all() function with a
5498 non-zero "wait" parameter. See weight_single
5501 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5502 xtime_lock contention on larger systems, and/or RCU lock
5503 contention on all systems with CONFIG_MAXSMP set.
5504 Format: { "0" | "1" }
5505 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5507 Note: increases power consumption, thus should only be
5508 enabled if running jitter sensitive (HPC/RT) workloads.
5510 security= [SECURITY] Choose a legacy "major" security module to
5511 enable at boot. This has been deprecated by the
5514 selinux= [SELINUX] Disable or enable SELinux at boot time.
5515 Format: { "0" | "1" }
5516 See security/selinux/Kconfig help text.
5521 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5522 Format: { "0" | "1" }
5523 See security/apparmor/Kconfig help text
5526 Default value is set via kernel config option.
5528 serialnumber [BUGS=X86-32]
5530 sev=option[,option...] [X86-64] See Documentation/x86/x86_64/boot-options.rst
5533 Maximal number of shapers.
5541 Enable merging of slabs with similar size when the
5542 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5545 Disable merging of slabs with similar size. May be
5546 necessary if there is some reason to distinguish
5547 allocs to different slabs, especially in hardened
5548 environments where the risk of heap overflows and
5549 layout control by attackers can usually be
5550 frustrated by disabling merging. This will reduce
5551 most of the exposure of a heap attack to a single
5552 cache (risks via metadata attacks are mostly
5553 unchanged). Debug options disable merging on their
5555 For more information see Documentation/mm/slub.rst.
5557 slab_max_order= [MM, SLAB]
5558 Determines the maximum allowed order for slabs.
5559 A high setting may cause OOMs due to memory
5560 fragmentation. Defaults to 1 for systems with
5561 more than 32MB of RAM, 0 otherwise.
5563 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5564 Enabling slub_debug allows one to determine the
5565 culprit if slab objects become corrupted. Enabling
5566 slub_debug can create guard zones around objects and
5567 may poison objects when not in use. Also tracks the
5568 last alloc / free. For more information see
5569 Documentation/mm/slub.rst.
5571 slub_max_order= [MM, SLUB]
5572 Determines the maximum allowed order for slabs.
5573 A high setting may cause OOMs due to memory
5574 fragmentation. For more information see
5575 Documentation/mm/slub.rst.
5577 slub_min_objects= [MM, SLUB]
5578 The minimum number of objects per slab. SLUB will
5579 increase the slab order up to slub_max_order to
5580 generate a sufficiently large slab able to contain
5581 the number of objects indicated. The higher the number
5582 of objects the smaller the overhead of tracking slabs
5583 and the less frequently locks need to be acquired.
5584 For more information see Documentation/mm/slub.rst.
5586 slub_min_order= [MM, SLUB]
5587 Determines the minimum page order for slabs. Must be
5588 lower than slub_max_order.
5589 For more information see Documentation/mm/slub.rst.
5591 slub_merge [MM, SLUB]
5592 Same with slab_merge.
5594 slub_nomerge [MM, SLUB]
5595 Same with slab_nomerge. This is supported for legacy.
5596 See slab_nomerge for more information.
5599 Format: <io1>[,<io2>[,...,<io8>]]
5601 smp.csd_lock_timeout= [KNL]
5602 Specify the period of time in milliseconds
5603 that smp_call_function() and friends will wait
5604 for a CPU to release the CSD lock. This is
5605 useful when diagnosing bugs involving CPUs
5606 disabling interrupts for extended periods
5607 of time. Defaults to 5,000 milliseconds, and
5608 setting a value of zero disables this feature.
5609 This feature may be more efficiently disabled
5610 using the csdlock_debug- kernel parameter.
5612 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5613 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5614 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5615 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5616 smsc-ircc2.ircc_irq= [HW] IRQ line
5617 smsc-ircc2.ircc_dma= [HW] DMA channel
5618 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5619 0: Toshiba Satellite 1800 (GP data pin select)
5620 1: Fast pin select (default)
5623 smt= [KNL,S390] Set the maximum number of threads (logical
5624 CPUs) to use per physical CPU on systems capable of
5625 symmetric multithreading (SMT). Will be capped to the
5626 actual hardware limit.
5628 Default: -1 (no limit)
5631 [KNL] Should the soft-lockup detector generate panics.
5634 A value of 1 instructs the soft-lockup detector
5635 to panic the machine when a soft-lockup occurs. It is
5636 also controlled by the kernel.softlockup_panic sysctl
5637 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5638 respective build-time switch to that functionality.
5640 softlockup_all_cpu_backtrace=
5641 [KNL] Should the soft-lockup detector generate
5642 backtraces on all cpus.
5645 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5646 See Documentation/admin-guide/laptops/sonypi.rst
5648 spectre_v2= [X86] Control mitigation of Spectre variant 2
5649 (indirect branch speculation) vulnerability.
5650 The default operation protects the kernel from
5653 on - unconditionally enable, implies
5655 off - unconditionally disable, implies
5657 auto - kernel detects whether your CPU model is
5660 Selecting 'on' will, and 'auto' may, choose a
5661 mitigation method at run time according to the
5662 CPU, the available microcode, the setting of the
5663 CONFIG_RETPOLINE configuration option, and the
5664 compiler with which the kernel was built.
5666 Selecting 'on' will also enable the mitigation
5667 against user space to user space task attacks.
5669 Selecting 'off' will disable both the kernel and
5670 the user space protections.
5672 Specific mitigations can also be selected manually:
5674 retpoline - replace indirect branches
5675 retpoline,generic - Retpolines
5676 retpoline,lfence - LFENCE; indirect branch
5677 retpoline,amd - alias for retpoline,lfence
5678 eibrs - enhanced IBRS
5679 eibrs,retpoline - enhanced IBRS + Retpolines
5680 eibrs,lfence - enhanced IBRS + LFENCE
5681 ibrs - use IBRS to protect kernel
5683 Not specifying this option is equivalent to
5687 [X86] Control mitigation of Spectre variant 2
5688 (indirect branch speculation) vulnerability between
5691 on - Unconditionally enable mitigations. Is
5692 enforced by spectre_v2=on
5694 off - Unconditionally disable mitigations. Is
5695 enforced by spectre_v2=off
5697 prctl - Indirect branch speculation is enabled,
5698 but mitigation can be enabled via prctl
5699 per thread. The mitigation control state
5700 is inherited on fork.
5703 - Like "prctl" above, but only STIBP is
5704 controlled per thread. IBPB is issued
5705 always when switching between different user
5709 - Same as "prctl" above, but all seccomp
5710 threads will enable the mitigation unless
5711 they explicitly opt out.
5714 - Like "seccomp" above, but only STIBP is
5715 controlled per thread. IBPB is issued
5716 always when switching between different
5717 user space processes.
5719 auto - Kernel selects the mitigation depending on
5720 the available CPU features and vulnerability.
5722 Default mitigation: "prctl"
5724 Not specifying this option is equivalent to
5725 spectre_v2_user=auto.
5727 spec_store_bypass_disable=
5728 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5729 (Speculative Store Bypass vulnerability)
5731 Certain CPUs are vulnerable to an exploit against a
5732 a common industry wide performance optimization known
5733 as "Speculative Store Bypass" in which recent stores
5734 to the same memory location may not be observed by
5735 later loads during speculative execution. The idea
5736 is that such stores are unlikely and that they can
5737 be detected prior to instruction retirement at the
5738 end of a particular speculation execution window.
5740 In vulnerable processors, the speculatively forwarded
5741 store can be used in a cache side channel attack, for
5742 example to read memory to which the attacker does not
5743 directly have access (e.g. inside sandboxed code).
5745 This parameter controls whether the Speculative Store
5746 Bypass optimization is used.
5748 On x86 the options are:
5750 on - Unconditionally disable Speculative Store Bypass
5751 off - Unconditionally enable Speculative Store Bypass
5752 auto - Kernel detects whether the CPU model contains an
5753 implementation of Speculative Store Bypass and
5754 picks the most appropriate mitigation. If the
5755 CPU is not vulnerable, "off" is selected. If the
5756 CPU is vulnerable the default mitigation is
5757 architecture and Kconfig dependent. See below.
5758 prctl - Control Speculative Store Bypass per thread
5759 via prctl. Speculative Store Bypass is enabled
5760 for a process by default. The state of the control
5761 is inherited on fork.
5762 seccomp - Same as "prctl" above, but all seccomp threads
5763 will disable SSB unless they explicitly opt out.
5765 Default mitigations:
5768 On powerpc the options are:
5770 on,auto - On Power8 and Power9 insert a store-forwarding
5771 barrier on kernel entry and exit. On Power7
5772 perform a software flush on kernel entry and
5776 Not specifying this option is equivalent to
5777 spec_store_bypass_disable=auto.
5779 spia_io_base= [HW,MTD]
5785 [X86] Enable split lock detection or bus lock detection
5787 When enabled (and if hardware support is present), atomic
5788 instructions that access data across cache line
5789 boundaries will result in an alignment check exception
5790 for split lock detection or a debug exception for
5795 warn - the kernel will emit rate-limited warnings
5796 about applications triggering the #AC
5797 exception or the #DB exception. This mode is
5798 the default on CPUs that support split lock
5799 detection or bus lock detection. Default
5800 behavior is by #AC if both features are
5801 enabled in hardware.
5803 fatal - the kernel will send SIGBUS to applications
5804 that trigger the #AC exception or the #DB
5805 exception. Default behavior is by #AC if
5806 both features are enabled in hardware.
5809 Set system wide rate limit to N bus locks
5810 per second for bus lock detection.
5813 N/A for split lock detection.
5816 If an #AC exception is hit in the kernel or in
5817 firmware (i.e. not while executing in user mode)
5818 the kernel will oops in either "warn" or "fatal"
5821 #DB exception for bus lock is triggered only when
5825 Control the Special Register Buffer Data Sampling
5828 Certain CPUs are vulnerable to an MDS-like
5829 exploit which can leak bits from the random
5832 By default, this issue is mitigated by
5833 microcode. However, the microcode fix can cause
5834 the RDRAND and RDSEED instructions to become
5835 much slower. Among other effects, this will
5836 result in reduced throughput from /dev/urandom.
5838 The microcode mitigation can be disabled with
5839 the following option:
5841 off: Disable mitigation and remove
5842 performance impact to RDRAND and RDSEED
5844 srcutree.big_cpu_lim [KNL]
5845 Specifies the number of CPUs constituting a
5846 large system, such that srcu_struct structures
5847 should immediately allocate an srcu_node array.
5848 This kernel-boot parameter defaults to 128,
5849 but takes effect only when the low-order four
5850 bits of srcutree.convert_to_big is equal to 3
5853 srcutree.convert_to_big [KNL]
5854 Specifies under what conditions an SRCU tree
5855 srcu_struct structure will be converted to big
5856 form, that is, with an rcu_node tree:
5859 1: At init_srcu_struct() time.
5860 2: When rcutorture decides to.
5861 3: Decide at boot time (default).
5862 0x1X: Above plus if high contention.
5864 Either way, the srcu_node tree will be sized based
5865 on the actual runtime number of CPUs (nr_cpu_ids)
5866 instead of the compile-time CONFIG_NR_CPUS.
5868 srcutree.counter_wrap_check [KNL]
5869 Specifies how frequently to check for
5870 grace-period sequence counter wrap for the
5871 srcu_data structure's ->srcu_gp_seq_needed field.
5872 The greater the number of bits set in this kernel
5873 parameter, the less frequently counter wrap will
5874 be checked for. Note that the bottom two bits
5877 srcutree.exp_holdoff [KNL]
5878 Specifies how many nanoseconds must elapse
5879 since the end of the last SRCU grace period for
5880 a given srcu_struct until the next normal SRCU
5881 grace period will be considered for automatic
5882 expediting. Set to zero to disable automatic
5885 srcutree.srcu_max_nodelay [KNL]
5886 Specifies the number of no-delay instances
5887 per jiffy for which the SRCU grace period
5888 worker thread will be rescheduled with zero
5889 delay. Beyond this limit, worker thread will
5890 be rescheduled with a sleep delay of one jiffy.
5892 srcutree.srcu_max_nodelay_phase [KNL]
5893 Specifies the per-grace-period phase, number of
5894 non-sleeping polls of readers. Beyond this limit,
5895 grace period worker thread will be rescheduled
5896 with a sleep delay of one jiffy, between each
5897 rescan of the readers, for a grace period phase.
5899 srcutree.srcu_retry_check_delay [KNL]
5900 Specifies number of microseconds of non-sleeping
5901 delay between each non-sleeping poll of readers.
5903 srcutree.small_contention_lim [KNL]
5904 Specifies the number of update-side contention
5905 events per jiffy will be tolerated before
5906 initiating a conversion of an srcu_struct
5907 structure to big form. Note that the value of
5908 srcutree.convert_to_big must have the 0x10 bit
5909 set for contention-based conversions to occur.
5912 Speculative Store Bypass Disable control
5914 On CPUs that are vulnerable to the Speculative
5915 Store Bypass vulnerability and offer a
5916 firmware based mitigation, this parameter
5917 indicates how the mitigation should be used:
5919 force-on: Unconditionally enable mitigation for
5920 for both kernel and userspace
5921 force-off: Unconditionally disable mitigation for
5922 for both kernel and userspace
5923 kernel: Always enable mitigation in the
5924 kernel, and offer a prctl interface
5925 to allow userspace to register its
5926 interest in being mitigated too.
5928 stack_guard_gap= [MM]
5929 override the default stack gap protection. The value
5930 is in page units and it defines how many pages prior
5931 to (for stacks growing down) resp. after (for stacks
5932 growing up) the main stack are reserved for no other
5933 mapping. Default value is 256 pages.
5935 stack_depot_disable= [KNL]
5936 Setting this to true through kernel command line will
5937 disable the stack depot thereby saving the static memory
5938 consumed by the stack hash table. By default this is set
5942 Enabled the stack tracer on boot up.
5944 stacktrace_filter=[function-list]
5945 [FTRACE] Limit the functions that the stack tracer
5946 will trace at boot up. function-list is a comma-separated
5947 list of functions. This list can be changed at run
5948 time by the stack_trace_filter file in the debugfs
5949 tracing directory. Note, this enables stack tracing
5950 and the stacktrace above is not needed.
5954 Set the STI (builtin display/keyboard on the HP-PARISC
5955 machines) console (graphic card) which should be used
5956 as the initial boot-console.
5957 See also comment in drivers/video/console/sticore.c.
5960 See comment in drivers/video/console/sticore.c.
5963 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5968 Enable or disable strict sigaltstack size checks
5969 against the required signal frame size which
5970 depends on the supported FPU features. This can
5971 be used to filter out binaries which have
5972 not yet been made aware of AT_MINSIGSTKSZ.
5974 sunrpc.min_resvport=
5975 sunrpc.max_resvport=
5977 SunRPC servers often require that client requests
5978 originate from a privileged port (i.e. a port in the
5979 range 0 < portnr < 1024).
5980 An administrator who wishes to reserve some of these
5981 ports for other uses may adjust the range that the
5982 kernel's sunrpc client considers to be privileged
5983 using these two parameters to set the minimum and
5984 maximum port values.
5986 sunrpc.svc_rpc_per_connection_limit=
5988 Limit the number of requests that the server will
5989 process in parallel from a single connection.
5990 The default value is 0 (no limit).
5994 Control how the NFS server code allocates CPUs to
5995 service thread pools. Depending on how many NICs
5996 you have and where their interrupts are bound, this
5997 option will affect which CPUs will do NFS serving.
5998 Note: this parameter cannot be changed while the
5999 NFS server is running.
6001 auto the server chooses an appropriate mode
6002 automatically using heuristics
6003 global a single global pool contains all CPUs
6004 percpu one pool for each CPU
6005 pernode one pool for each NUMA node (equivalent
6006 to global on non-NUMA machines)
6008 sunrpc.tcp_slot_table_entries=
6009 sunrpc.udp_slot_table_entries=
6011 Sets the upper limit on the number of simultaneous
6012 RPC calls that can be sent from the client to a
6013 server. Increasing these values may allow you to
6014 improve throughput, but will also increase the
6015 amount of memory reserved for use by the client.
6017 suspend.pm_test_delay=
6019 Sets the number of seconds to remain in a suspend test
6020 mode before resuming the system (see
6021 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6022 is set. Default value is 5.
6025 Format: { on | off | y | n | 1 | 0 }
6026 This parameter controls use of the Protected
6027 Execution Facility on pSeries.
6031 Enable accounting of swap in memory resource
6032 controller if no parameter or 1 is given or disable
6033 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
6035 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
6036 Format: { <int> [,<int>] | force | noforce }
6037 <int> -- Number of I/O TLB slabs
6038 <int> -- Second integer after comma. Number of swiotlb
6039 areas with their own lock. Will be rounded up
6041 force -- force using of bounce buffers even if they
6042 wouldn't be automatically used by the kernel
6043 noforce -- Never use bounce buffers (for debugging)
6048 Set a sysctl parameter, right before loading the init
6049 process, as if the value was written to the respective
6050 /proc/sys/... file. Both '.' and '/' are recognized as
6051 separators. Unrecognized parameters and invalid values
6052 are reported in the kernel log. Sysctls registered
6053 later by a loaded module cannot be set this way.
6054 Example: sysctl.vm.swappiness=40
6056 sysfs.deprecated=0|1 [KNL]
6057 Enable/disable old style sysfs layout for old udev
6058 on older distributions. When this option is enabled
6059 very new udev will not work anymore. When this option
6060 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
6061 in older udev will not work anymore.
6062 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
6063 the kernel configuration.
6065 sysrq_always_enabled
6067 Ignore sysrq setting - this boot parameter will
6068 neutralize any effect of /proc/sys/kernel/sysrq.
6069 Useful for debugging.
6071 tcpmhash_entries= [KNL,NET]
6072 Set the number of tcp_metrics_hash slots.
6073 Default value is 8192 or 16384 depending on total
6074 ram pages. This is used to specify the TCP metrics
6075 cache size. See Documentation/networking/ip-sysctl.rst
6076 "tcp_no_metrics_save" section for more details.
6080 test_suspend= [SUSPEND]
6081 Format: { "mem" | "standby" | "freeze" }[,N]
6082 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6083 standby suspend) or "freeze" (for suspend type freeze)
6084 as the system sleep state during system startup with
6085 the optional capability to repeat N number of times.
6086 The system is woken from this state using a
6087 wakeup-capable RTC alarm.
6089 thash_entries= [KNL,NET]
6090 Set number of hash buckets for TCP connection
6092 thermal.act= [HW,ACPI]
6093 -1: disable all active trip points in all thermal zones
6094 <degrees C>: override all lowest active trip points
6096 thermal.crt= [HW,ACPI]
6097 -1: disable all critical trip points in all thermal zones
6098 <degrees C>: override all critical trip points
6100 thermal.nocrt= [HW,ACPI]
6101 Set to disable actions on ACPI thermal zone
6102 critical and hot trip points.
6104 thermal.off= [HW,ACPI]
6105 1: disable ACPI thermal control
6107 thermal.psv= [HW,ACPI]
6108 -1: disable all passive trip points
6109 <degrees C>: override all passive trip points to this
6112 thermal.tzp= [HW,ACPI]
6113 Specify global default ACPI thermal zone polling rate
6114 <deci-seconds>: poll all this frequency
6115 0: no polling (default)
6118 Force threading of all interrupt handlers except those
6119 marked explicitly IRQF_NO_THREAD.
6123 Specify if the kernel should make use of the cpu
6124 topology information if the hardware supports this.
6125 The scheduler will make use of this information and
6126 e.g. base its process migration decisions on it.
6129 topology_updates= [KNL, PPC, NUMA]
6131 Specify if the kernel should ignore (off)
6132 topology updates sent by the hypervisor to this
6135 torture.disable_onoff_at_boot= [KNL]
6136 Prevent the CPU-hotplug component of torturing
6137 until after init has spawned.
6139 torture.ftrace_dump_at_shutdown= [KNL]
6140 Dump the ftrace buffer at torture-test shutdown,
6141 even if there were no errors. This can be a
6142 very costly operation when many torture tests
6143 are running concurrently, especially on systems
6144 with rotating-rust storage.
6146 torture.verbose_sleep_frequency= [KNL]
6147 Specifies how many verbose printk()s should be
6148 emitted between each sleep. The default of zero
6149 disables verbose-printk() sleeping.
6151 torture.verbose_sleep_duration= [KNL]
6152 Duration of each verbose-printk() sleep in jiffies.
6156 tpm_suspend_pcr=[HW,TPM]
6157 Format: integer pcr id
6158 Specify that at suspend time, the tpm driver
6159 should extend the specified pcr with zeros,
6160 as a workaround for some chips which fail to
6161 flush the last written pcr on TPM_SaveState.
6162 This will guarantee that all the other pcrs
6166 Have the tracepoints sent to printk as well as the
6167 tracing ring buffer. This is useful for early boot up
6168 where the system hangs or reboots and does not give the
6169 option for reading the tracing buffer or performing a
6170 ftrace_dump_on_oops.
6172 To turn off having tracepoints sent to printk,
6173 echo 0 > /proc/sys/kernel/tracepoint_printk
6174 Note, echoing 1 into this file without the
6175 tracepoint_printk kernel cmdline option has no effect.
6177 The tp_printk_stop_on_boot (see below) can also be used
6178 to stop the printing of events to console at
6183 Having tracepoints sent to printk() and activating high
6184 frequency tracepoints such as irq or sched, can cause
6185 the system to live lock.
6187 tp_printk_stop_on_boot [FTRACE]
6188 When tp_printk (above) is set, it can cause a lot of noise
6189 on the console. It may be useful to only include the
6190 printing of events during boot up, as user space may
6191 make the system inoperable.
6193 This command line option will stop the printing of events
6194 to console at the late_initcall_sync() time frame.
6196 trace_buf_size=nn[KMG]
6197 [FTRACE] will set tracing buffer size on each cpu.
6199 trace_clock= [FTRACE] Set the clock used for tracing events
6201 local - Use the per CPU time stamp counter
6202 (converted into nanoseconds). Fast, but
6203 depending on the architecture, may not be
6204 in sync between CPUs.
6205 global - Event time stamps are synchronize across
6206 CPUs. May be slower than the local clock,
6207 but better for some race conditions.
6208 counter - Simple counting of events (1, 2, ..)
6209 note, some counts may be skipped due to the
6210 infrastructure grabbing the clock more than
6212 uptime - Use jiffies as the time stamp.
6213 perf - Use the same clock that perf uses.
6214 mono - Use ktime_get_mono_fast_ns() for time stamps.
6215 mono_raw - Use ktime_get_raw_fast_ns() for time
6217 boot - Use ktime_get_boot_fast_ns() for time stamps.
6218 Architectures may add more clocks. See
6219 Documentation/trace/ftrace.rst for more details.
6221 trace_event=[event-list]
6222 [FTRACE] Set and start specified trace events in order
6223 to facilitate early boot debugging. The event-list is a
6224 comma-separated list of trace events to enable. See
6225 also Documentation/trace/events.rst
6227 trace_options=[option-list]
6228 [FTRACE] Enable or disable tracer options at boot.
6229 The option-list is a comma delimited list of options
6230 that can be enabled or disabled just as if you were
6231 to echo the option name into
6233 /sys/kernel/debug/tracing/trace_options
6235 For example, to enable stacktrace option (to dump the
6236 stack trace of each event), add to the command line:
6238 trace_options=stacktrace
6240 See also Documentation/trace/ftrace.rst "trace options"
6244 [FTRACE] enable this option to disable tracing when a
6245 warning is hit. This turns off "tracing_on". Tracing can
6246 be enabled again by echoing '1' into the "tracing_on"
6247 file located in /sys/kernel/debug/tracing/
6249 This option is useful, as it disables the trace before
6250 the WARNING dump is called, which prevents the trace to
6251 be filled with content caused by the warning output.
6253 This option can also be set at run time via the sysctl
6254 option: kernel/traceoff_on_warning
6256 transparent_hugepage=
6258 Format: [always|madvise|never]
6259 Can be used to control the default behavior of the system
6260 with respect to transparent hugepages.
6261 See Documentation/admin-guide/mm/transhuge.rst
6264 trusted.source= [KEYS]
6266 This parameter identifies the trust source as a backend
6267 for trusted keys implementation. Supported trust
6272 If not specified then it defaults to iterating through
6273 the trust source list starting with TPM and assigns the
6274 first trust source as a backend which is initialized
6275 successfully during iteration.
6279 The RNG used to generate key material for trusted keys.
6282 - the same value as trusted.source: "tpm" or "tee"
6284 If not specified, "default" is used. In this case,
6285 the RNG's choice is left to each individual trust source.
6287 tsc= Disable clocksource stability checks for TSC.
6289 [x86] reliable: mark tsc clocksource as reliable, this
6290 disables clocksource verification at runtime, as well
6291 as the stability checks done at bootup. Used to enable
6292 high-resolution timer mode on older hardware, and in
6293 virtualized environment.
6294 [x86] noirqtime: Do not use TSC to do irq accounting.
6295 Used to run time disable IRQ_TIME_ACCOUNTING on any
6296 platforms where RDTSC is slow and this accounting
6298 [x86] unstable: mark the TSC clocksource as unstable, this
6299 marks the TSC unconditionally unstable at bootup and
6300 avoids any further wobbles once the TSC watchdog notices.
6301 [x86] nowatchdog: disable clocksource watchdog. Used
6302 in situations with strict latency requirements (where
6303 interruptions from clocksource watchdog are not
6306 tsc_early_khz= [X86] Skip early TSC calibration and use the given
6307 value instead. Useful when the early TSC frequency discovery
6308 procedure is not reliable, such as on overclocked systems
6309 with CPUID.16h support and partial CPUID.15h support.
6310 Format: <unsigned int>
6312 tsx= [X86] Control Transactional Synchronization
6313 Extensions (TSX) feature in Intel processors that
6314 support TSX control.
6316 This parameter controls the TSX feature. The options are:
6318 on - Enable TSX on the system. Although there are
6319 mitigations for all known security vulnerabilities,
6320 TSX has been known to be an accelerator for
6321 several previous speculation-related CVEs, and
6322 so there may be unknown security risks associated
6323 with leaving it enabled.
6325 off - Disable TSX on the system. (Note that this
6326 option takes effect only on newer CPUs which are
6327 not vulnerable to MDS, i.e., have
6328 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6329 the new IA32_TSX_CTRL MSR through a microcode
6330 update. This new MSR allows for the reliable
6331 deactivation of the TSX functionality.)
6333 auto - Disable TSX if X86_BUG_TAA is present,
6334 otherwise enable TSX on the system.
6336 Not specifying this option is equivalent to tsx=off.
6338 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6341 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6342 Abort (TAA) vulnerability.
6344 Similar to Micro-architectural Data Sampling (MDS)
6345 certain CPUs that support Transactional
6346 Synchronization Extensions (TSX) are vulnerable to an
6347 exploit against CPU internal buffers which can forward
6348 information to a disclosure gadget under certain
6351 In vulnerable processors, the speculatively forwarded
6352 data can be used in a cache side channel attack, to
6353 access data to which the attacker does not have direct
6356 This parameter controls the TAA mitigation. The
6359 full - Enable TAA mitigation on vulnerable CPUs
6362 full,nosmt - Enable TAA mitigation and disable SMT on
6363 vulnerable CPUs. If TSX is disabled, SMT
6364 is not disabled because CPU is not
6365 vulnerable to cross-thread TAA attacks.
6366 off - Unconditionally disable TAA mitigation
6368 On MDS-affected machines, tsx_async_abort=off can be
6369 prevented by an active MDS mitigation as both vulnerabilities
6370 are mitigated with the same mechanism so in order to disable
6371 this mitigation, you need to specify mds=off too.
6373 Not specifying this option is equivalent to
6374 tsx_async_abort=full. On CPUs which are MDS affected
6375 and deploy MDS mitigation, TAA mitigation is not
6376 required and doesn't provide any additional
6380 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6382 turbografx.map[2|3]= [HW,JOY]
6383 TurboGraFX parallel port interface
6385 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6386 See also Documentation/input/devices/joystick-parport.rst
6388 udbg-immortal [PPC] When debugging early kernel crashes that
6389 happen after console_init() and before a proper
6390 console driver takes over, this boot options might
6391 help "seeing" what's going on.
6393 uhash_entries= [KNL,NET]
6394 Set number of hash buckets for UDP/UDP-Lite connections
6397 [USB] Ignore overcurrent events (default N).
6398 Some badly-designed motherboards generate lots of
6399 bogus events, for ports that aren't wired to
6400 anything. Set this parameter to avoid log spamming.
6401 Note that genuine overcurrent events won't be
6405 [X86] Cause panic on unknown NMI.
6407 usbcore.authorized_default=
6408 [USB] Default USB device authorization:
6409 (default -1 = authorized except for wireless USB,
6410 0 = not authorized, 1 = authorized, 2 = authorized
6411 if device connected to internal port)
6413 usbcore.autosuspend=
6414 [USB] The autosuspend time delay (in seconds) used
6415 for newly-detected USB devices (default 2). This
6416 is the time required before an idle device will be
6417 autosuspended. Devices for which the delay is set
6418 to a negative value won't be autosuspended at all.
6420 usbcore.usbfs_snoop=
6421 [USB] Set to log all usbfs traffic (default 0 = off).
6423 usbcore.usbfs_snoop_max=
6424 [USB] Maximum number of bytes to snoop in each URB
6427 usbcore.blinkenlights=
6428 [USB] Set to cycle leds on hubs (default 0 = off).
6430 usbcore.old_scheme_first=
6431 [USB] Start with the old device initialization
6432 scheme (default 0 = off).
6434 usbcore.usbfs_memory_mb=
6435 [USB] Memory limit (in MB) for buffers allocated by
6436 usbfs (default = 16, 0 = max = 2047).
6438 usbcore.use_both_schemes=
6439 [USB] Try the other device initialization scheme
6440 if the first one fails (default 1 = enabled).
6442 usbcore.initial_descriptor_timeout=
6443 [USB] Specifies timeout for the initial 64-byte
6444 USB_REQ_GET_DESCRIPTOR request in milliseconds
6445 (default 5000 = 5.0 seconds).
6447 usbcore.nousb [USB] Disable the USB subsystem
6450 [USB] A list of quirk entries to augment the built-in
6451 usb core quirk list. List entries are separated by
6452 commas. Each entry has the form
6453 VendorID:ProductID:Flags. The IDs are 4-digit hex
6454 numbers and Flags is a set of letters. Each letter
6455 will change the built-in quirk; setting it if it is
6456 clear and clearing it if it is set. The letters have
6457 the following meanings:
6458 a = USB_QUIRK_STRING_FETCH_255 (string
6459 descriptors must not be fetched using
6461 b = USB_QUIRK_RESET_RESUME (device can't resume
6462 correctly so reset it instead);
6463 c = USB_QUIRK_NO_SET_INTF (device can't handle
6464 Set-Interface requests);
6465 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6466 handle its Configuration or Interface
6468 e = USB_QUIRK_RESET (device can't be reset
6469 (e.g morph devices), don't use reset);
6470 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6471 more interface descriptions than the
6472 bNumInterfaces count, and can't handle
6473 talking to these interfaces);
6474 g = USB_QUIRK_DELAY_INIT (device needs a pause
6475 during initialization, after we read
6476 the device descriptor);
6477 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6478 high speed and super speed interrupt
6479 endpoints, the USB 2.0 and USB 3.0 spec
6480 require the interval in microframes (1
6481 microframe = 125 microseconds) to be
6482 calculated as interval = 2 ^
6484 Devices with this quirk report their
6485 bInterval as the result of this
6486 calculation instead of the exponent
6487 variable used in the calculation);
6488 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6489 handle device_qualifier descriptor
6491 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6492 generates spurious wakeup, ignore
6493 remote wakeup capability);
6494 k = USB_QUIRK_NO_LPM (device can't handle Link
6496 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6497 (Device reports its bInterval as linear
6498 frames instead of the USB 2.0
6500 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6501 to be disconnected before suspend to
6502 prevent spurious wakeup);
6503 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6504 pause after every control message);
6505 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6506 delay after resetting its port);
6507 Example: quirks=0781:5580:bk,0a5c:5834:gij
6510 [USBHID] The interval which mice are to be polled at.
6513 [USBHID] The interval which joysticks are to be polled at.
6516 [USBHID] The interval which keyboards are to be polled at.
6518 usb-storage.delay_use=
6519 [UMS] The delay in seconds before a new device is
6520 scanned for Logical Units (default 1).
6523 [UMS] A list of quirks entries to supplement or
6524 override the built-in unusual_devs list. List
6525 entries are separated by commas. Each entry has
6526 the form VID:PID:Flags where VID and PID are Vendor
6527 and Product ID values (4-digit hex numbers) and
6528 Flags is a set of characters, each corresponding
6529 to a common usb-storage quirk flag as follows:
6530 a = SANE_SENSE (collect more than 18 bytes
6531 of sense data, not on uas);
6532 b = BAD_SENSE (don't collect more than 18
6533 bytes of sense data, not on uas);
6534 c = FIX_CAPACITY (decrease the reported
6535 device capacity by one sector);
6536 d = NO_READ_DISC_INFO (don't use
6537 READ_DISC_INFO command, not on uas);
6538 e = NO_READ_CAPACITY_16 (don't use
6539 READ_CAPACITY_16 command);
6540 f = NO_REPORT_OPCODES (don't use report opcodes
6542 g = MAX_SECTORS_240 (don't transfer more than
6543 240 sectors at a time, uas only);
6544 h = CAPACITY_HEURISTICS (decrease the
6545 reported device capacity by one
6546 sector if the number is odd);
6547 i = IGNORE_DEVICE (don't bind to this
6549 j = NO_REPORT_LUNS (don't use report luns
6551 k = NO_SAME (do not use WRITE_SAME, uas only)
6552 l = NOT_LOCKABLE (don't try to lock and
6553 unlock ejectable media, not on uas);
6554 m = MAX_SECTORS_64 (don't transfer more
6555 than 64 sectors = 32 KB at a time,
6557 n = INITIAL_READ10 (force a retry of the
6558 initial READ(10) command, not on uas);
6559 o = CAPACITY_OK (accept the capacity
6560 reported by the device, not on uas);
6561 p = WRITE_CACHE (the device cache is ON
6562 by default, not on uas);
6563 r = IGNORE_RESIDUE (the device reports
6564 bogus residue values, not on uas);
6565 s = SINGLE_LUN (the device has only one
6567 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6568 commands, uas only);
6569 u = IGNORE_UAS (don't bind to the uas driver);
6570 w = NO_WP_DETECT (don't test whether the
6571 medium is write-protected).
6572 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6573 even if the device claims no cache,
6575 Example: quirks=0419:aaf5:rl,0421:0433:rc
6577 user_debug= [KNL,ARM]
6579 See arch/arm/Kconfig.debug help text.
6580 1 - undefined instruction events
6582 4 - invalid data aborts
6585 Example: user_debug=31
6588 [X86] Flags controlling user PTE allocations.
6590 nohigh = do not allocate PTE pages in
6591 HIGHMEM regardless of setting
6594 vdso= [X86,SH,SPARC]
6595 On X86_32, this is an alias for vdso32=. Otherwise:
6597 vdso=1: enable VDSO (the default)
6598 vdso=0: disable VDSO mapping
6600 vdso32= [X86] Control the 32-bit vDSO
6601 vdso32=1: enable 32-bit VDSO
6602 vdso32=0 or vdso32=2: disable 32-bit VDSO
6604 See the help text for CONFIG_COMPAT_VDSO for more
6605 details. If CONFIG_COMPAT_VDSO is set, the default is
6606 vdso32=0; otherwise, the default is vdso32=1.
6608 For compatibility with older kernels, vdso32=2 is an
6611 Try vdso32=0 if you encounter an error that says:
6612 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6615 vector=percpu: enable percpu vector domain
6617 video= [FB] Frame buffer configuration
6618 See Documentation/fb/modedb.rst.
6620 video.brightness_switch_enabled= [ACPI]
6622 If set to 1, on receiving an ACPI notify event
6623 generated by hotkey, video driver will adjust brightness
6624 level and then send out the event to user space through
6625 the allocated input device. If set to 0, video driver
6626 will only send out the event without touching backlight
6631 [VMMIO] Memory mapped virtio (platform) device.
6633 <size>@<baseaddr>:<irq>[:<id>]
6635 <size> := size (can use standard suffixes
6637 <baseaddr> := physical base address
6638 <irq> := interrupt number (as passed to
6640 <id> := (optional) platform device id
6642 virtio_mmio.device=1K@0x100b0000:48:7
6644 Can be used multiple times for multiple devices.
6646 vga= [BOOT,X86-32] Select a particular video mode
6647 See Documentation/x86/boot.rst and
6648 Documentation/admin-guide/svga.rst.
6649 Use vga=ask for menu.
6650 This is actually a boot loader parameter; the value is
6651 passed to the kernel using a special protocol.
6653 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6654 May slow down system boot speed, especially when
6655 enabled on systems with a large amount of memory.
6656 All options are enabled by default, and this
6657 interface is meant to allow for selectively
6658 enabling or disabling specific virtual memory
6661 Available options are:
6662 P Enable page structure init time poisoning
6663 - Disable all of the above options
6665 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6666 size of <nn>. This can be used to increase the
6667 minimum size (128MB on x86). It can also be used to
6668 decrease the size and leave more room for directly
6671 vmcp_cma=nn[MG] [KNL,S390]
6672 Sets the memory size reserved for contiguous memory
6673 allocations for the vmcp device driver.
6675 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6678 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6681 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6685 Controls the behavior of vsyscalls (i.e. calls to
6686 fixed addresses of 0xffffffffff600x00 from legacy
6687 code). Most statically-linked binaries and older
6688 versions of glibc use these calls. Because these
6689 functions are at fixed addresses, they make nice
6690 targets for exploits that can control RIP.
6692 emulate [default] Vsyscalls turn into traps and are
6693 emulated reasonably safely. The vsyscall
6696 xonly Vsyscalls turn into traps and are
6697 emulated reasonably safely. The vsyscall
6698 page is not readable.
6700 none Vsyscalls don't work at all. This makes
6701 them quite hard to use for exploits but
6702 might break your system.
6704 vt.color= [VT] Default text color.
6705 Format: 0xYX, X = foreground, Y = background.
6706 Default: 0x07 = light gray on black.
6708 vt.cur_default= [VT] Default cursor shape.
6709 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6710 the parameters of the <Esc>[?A;B;Cc escape sequence;
6711 see VGA-softcursor.txt. Default: 2 = underline.
6713 vt.default_blu= [VT]
6714 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6715 Change the default blue palette of the console.
6716 This is a 16-member array composed of values
6719 vt.default_grn= [VT]
6720 Format: <green0>,<green1>,<green2>,...,<green15>
6721 Change the default green palette of the console.
6722 This is a 16-member array composed of values
6725 vt.default_red= [VT]
6726 Format: <red0>,<red1>,<red2>,...,<red15>
6727 Change the default red palette of the console.
6728 This is a 16-member array composed of values
6734 Set system-wide default UTF-8 mode for all tty's.
6735 Default is 1, i.e. UTF-8 mode is enabled for all
6736 newly opened terminals.
6738 vt.global_cursor_default=
6741 Set system-wide default for whether a cursor
6742 is shown on new VTs. Default is -1,
6743 i.e. cursors will be created by default unless
6744 overridden by individual drivers. 0 will hide
6745 cursors, 1 will display them.
6747 vt.italic= [VT] Default color for italic text; 0-15.
6750 vt.underline= [VT] Default color for underlined text; 0-15.
6753 watchdog timers [HW,WDT] For information on watchdog timers,
6754 see Documentation/watchdog/watchdog-parameters.rst
6755 or other driver-specific files in the
6756 Documentation/watchdog/ directory.
6760 Set the hard lockup detector stall duration
6761 threshold in seconds. The soft lockup detector
6762 threshold is set to twice the value. A value of 0
6763 disables both lockup detectors. Default is 10
6766 workqueue.watchdog_thresh=
6767 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6768 warn stall conditions and dump internal state to
6769 help debugging. 0 disables workqueue stall
6770 detection; otherwise, it's the stall threshold
6771 duration in seconds. The default value is 30 and
6772 it can be updated at runtime by writing to the
6773 corresponding sysfs file.
6775 workqueue.disable_numa
6776 By default, all work items queued to unbound
6777 workqueues are affine to the NUMA nodes they're
6778 issued on, which results in better behavior in
6779 general. If NUMA affinity needs to be disabled for
6780 whatever reason, this option can be used. Note
6781 that this also can be controlled per-workqueue for
6782 workqueues visible under /sys/bus/workqueue/.
6784 workqueue.power_efficient
6785 Per-cpu workqueues are generally preferred because
6786 they show better performance thanks to cache
6787 locality; unfortunately, per-cpu workqueues tend to
6788 be more power hungry than unbound workqueues.
6790 Enabling this makes the per-cpu workqueues which
6791 were observed to contribute significantly to power
6792 consumption unbound, leading to measurably lower
6793 power usage at the cost of small performance
6796 The default value of this parameter is determined by
6797 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6799 workqueue.debug_force_rr_cpu
6800 Workqueue used to implicitly guarantee that work
6801 items queued without explicit CPU specified are put
6802 on the local CPU. This guarantee is no longer true
6803 and while local CPU is still preferred work items
6804 may be put on foreign CPUs. This debug option
6805 forces round-robin CPU selection to flush out
6806 usages which depend on the now broken guarantee.
6807 When enabled, memory and cache locality will be
6810 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6811 default x2apic cluster mode on platforms
6814 xen_512gb_limit [KNL,X86-64,XEN]
6815 Restricts the kernel running paravirtualized under Xen
6816 to use only up to 512 GB of RAM. The reason to do so is
6817 crash analysis tools and Xen tools for doing domain
6818 save/restore/migration must be enabled to handle larger
6821 xen_emul_unplug= [HW,X86,XEN]
6822 Unplug Xen emulated devices
6823 Format: [unplug0,][unplug1]
6824 ide-disks -- unplug primary master IDE devices
6825 aux-ide-disks -- unplug non-primary-master IDE devices
6826 nics -- unplug network devices
6827 all -- unplug all emulated devices (NICs and IDE disks)
6828 unnecessary -- unplugging emulated devices is
6829 unnecessary even if the host did not respond to
6831 never -- do not unplug even if version check succeeds
6833 xen_legacy_crash [X86,XEN]
6834 Crash from Xen panic notifier, without executing late
6835 panic() code such as dumping handler.
6837 xen_nopvspin [X86,XEN]
6838 Disables the qspinlock slowpath using Xen PV optimizations.
6839 This parameter is obsoleted by "nopvspin" parameter, which
6840 has equivalent effect for XEN platform.
6843 Disables the PV optimizations forcing the HVM guest to
6844 run as generic HVM guest with no PV drivers.
6845 This option is obsoleted by the "nopv" option, which
6846 has equivalent effect for XEN platform.
6848 xen_no_vector_callback
6849 [KNL,X86,XEN] Disable the vector callback for Xen
6850 event channel interrupts.
6852 xen_scrub_pages= [XEN]
6853 Boolean option to control scrubbing pages before giving them back
6854 to Xen, for use by other domains. Can be also changed at runtime
6855 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6856 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6858 xen_timer_slop= [X86-64,XEN]
6859 Set the timer slop (in nanoseconds) for the virtual Xen
6860 timers (default is 100000). This adjusts the minimum
6861 delta of virtualized Xen timers, where lower values
6862 improve timer resolution at the expense of processing
6863 more timer interrupts.
6865 xen.balloon_boot_timeout= [XEN]
6866 The time (in seconds) to wait before giving up to boot
6867 in case initial ballooning fails to free enough memory.
6868 Applies only when running as HVM or PVH guest and
6869 started with less memory configured than allowed at
6870 max. Default is 180.
6872 xen.event_eoi_delay= [XEN]
6873 How long to delay EOI handling in case of event
6874 storms (jiffies). Default is 10.
6876 xen.event_loop_timeout= [XEN]
6877 After which time (jiffies) the event handling loop
6878 should start to delay EOI handling. Default is 2.
6880 xen.fifo_events= [XEN]
6881 Boolean parameter to disable using fifo event handling
6882 even if available. Normally fifo event handling is
6883 preferred over the 2-level event handling, as it is
6884 fairer and the number of possible event channels is
6885 much higher. Default is on (use fifo events).
6887 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6888 Disables the PV optimizations forcing the guest to run
6889 as generic guest with no PV drivers. Currently support
6890 XEN HVM, KVM, HYPER_V and VMWARE guest.
6892 nopvspin [X86,XEN,KVM]
6893 Disables the qspinlock slow path using PV optimizations
6894 which allow the hypervisor to 'idle' the guest on lock
6897 xirc2ps_cs= [NET,PCMCIA]
6899 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6902 By default on POWER9 and above, the kernel will
6903 natively use the XIVE interrupt controller. This option
6904 allows the fallback firmware mode to be used:
6906 off Fallback to firmware control of XIVE interrupt
6907 controller on both pseries and powernv
6908 platforms. Only useful on POWER9 and above.
6910 xive.store-eoi=off [PPC]
6911 By default on POWER10 and above, the kernel will use
6912 stores for EOI handling when the XIVE interrupt mode
6913 is active. This option allows the XIVE driver to use
6914 loads instead, as on POWER9.
6916 xhci-hcd.quirks [USB,KNL]
6917 A hex value specifying bitmask with supplemental xhci
6918 host controller quirks. Meaning of each bit can be
6919 consulted in header drivers/usb/host/xhci.h.
6922 Format: { early | on | rw | ro | off }
6923 Controls if xmon debugger is enabled. Default is off.
6924 Passing only "xmon" is equivalent to "xmon=early".
6925 early Call xmon as early as possible on boot; xmon
6926 debugger is called from setup_arch().
6927 on xmon debugger hooks will be installed so xmon
6928 is only called on a kernel crash. Default mode,
6929 i.e. either "ro" or "rw" mode, is controlled
6930 with CONFIG_XMON_DEFAULT_RO_MODE.
6931 rw xmon debugger hooks will be installed so xmon
6932 is called only on a kernel crash, mode is write,
6933 meaning SPR registers, memory and, other data
6934 can be written using xmon commands.
6935 ro same as "rw" option above but SPR registers,
6936 memory, and other data can't be written using
6938 off xmon is disabled.