2 Format: { eager | lazy }
4 By default, unaccepted memory is accepted lazily to
5 avoid prolonged boot times. The lazy option will add
6 some runtime overhead until all memory is eventually
7 accepted. In most cases the overhead is negligible.
8 For some workloads or for debugging purposes
9 accept_memory=eager can be used to accept all memory
12 acpi= [HW,ACPI,X86,ARM64,RISCV64,EARLY]
13 Advanced Configuration and Power Interface
14 Format: { force | on | off | strict | noirq | rsdt |
16 force -- enable ACPI if default was off
17 on -- enable ACPI but allow fallback to DT [arm64,riscv64]
18 off -- disable ACPI if default was on
19 noirq -- do not use ACPI for IRQ routing
20 strict -- Be less tolerant of platforms that are not
21 strictly ACPI specification compliant.
22 rsdt -- prefer RSDT over (default) XSDT
23 copy_dsdt -- copy DSDT to memory
24 For ARM64 and RISCV64, ONLY "acpi=off", "acpi=on" or
25 "acpi=force" are available
27 See also Documentation/power/runtime_pm.rst, pci=noacpi
29 acpi_apic_instance= [ACPI,IOAPIC,EARLY]
31 2: use 2nd APIC table, if available
32 1,0: use 1st APIC table
35 acpi_backlight= [HW,ACPI]
36 { vendor | video | native | none }
37 If set to vendor, prefer vendor-specific driver
38 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
39 of the ACPI video.ko driver.
40 If set to video, use the ACPI video.ko driver.
41 If set to native, use the device's native backlight mode.
42 If set to none, disable the ACPI backlight interface.
44 acpi_force_32bit_fadt_addr [ACPI,EARLY]
45 force FADT to use 32 bit addresses rather than the
46 64 bit X_* addresses. Some firmware have broken 64
47 bit addresses for force ACPI ignore these and use
48 the older legacy 32 bit addresses.
50 acpica_no_return_repair [HW, ACPI]
51 Disable AML predefined validation mechanism
52 This mechanism can repair the evaluation result to make
53 the return objects more ACPI specification compliant.
54 This option is useful for developers to identify the
55 root cause of an AML interpreter issue when the issue
56 has something to do with the repair mechanism.
58 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
59 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
61 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
62 debug output. Bits in debug_layer correspond to a
63 _COMPONENT in an ACPI source file, e.g.,
64 #define _COMPONENT ACPI_EVENTS
65 Bits in debug_level correspond to a level in
66 ACPI_DEBUG_PRINT statements, e.g.,
67 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
68 The debug_level mask defaults to "info". See
69 Documentation/firmware-guide/acpi/debug.rst for more information about
70 debug layers and levels.
72 Enable processor driver info messages:
73 acpi.debug_layer=0x20000000
74 Enable AML "Debug" output, i.e., stores to the Debug
75 object while interpreting AML:
76 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
77 Enable all messages related to ACPI hardware:
78 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
80 Some values produce so much output that the system is
81 unusable. The "log_buf_len" parameter may be useful
82 if you need to capture more output.
84 acpi_enforce_resources= [ACPI]
86 Check for resource conflicts between native drivers
87 and ACPI OperationRegions (SystemIO and SystemMemory
88 only). IO ports and memory declared in ACPI might be
89 used by the ACPI subsystem in arbitrary AML code and
90 can interfere with legacy drivers.
91 strict (default): access to resources claimed by ACPI
92 is denied; legacy drivers trying to access reserved
93 resources will fail to bind to device using them.
94 lax: access to resources claimed by ACPI is allowed;
95 legacy drivers trying to access reserved resources
96 will bind successfully but a warning message is logged.
97 no: ACPI OperationRegions are not marked as reserved,
98 no further checks are performed.
100 acpi_force_table_verification [HW,ACPI,EARLY]
101 Enable table checksum verification during early stage.
102 By default, this is disabled due to x86 early mapping
105 acpi_irq_balance [HW,ACPI]
106 ACPI will balance active IRQs
109 acpi_irq_nobalance [HW,ACPI]
110 ACPI will not move active IRQs (default)
113 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
114 Format: <irq>,<irq>...
116 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
118 Format: <irq>,<irq>...
120 acpi_mask_gpe= [HW,ACPI]
121 Due to the existence of _Lxx/_Exx, some GPEs triggered
122 by unsupported hardware/firmware features can result in
123 GPE floodings that cannot be automatically disabled by
125 This facility can be used to prevent such uncontrolled
127 Format: <byte> or <bitmap-list>
129 acpi_no_auto_serialize [HW,ACPI]
130 Disable auto-serialization of AML methods
131 AML control methods that contain the opcodes to create
132 named objects will be marked as "Serialized" by the
133 auto-serialization feature.
134 This feature is enabled by default.
135 This option allows to turn off the feature.
137 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
140 acpi_no_static_ssdt [HW,ACPI,EARLY]
141 Disable installation of static SSDTs at early boot time
142 By default, SSDTs contained in the RSDT/XSDT will be
143 installed automatically and they will appear under
144 /sys/firmware/acpi/tables.
145 This option turns off this feature.
146 Note that specifying this option does not affect
147 dynamic table installation which will install SSDT
148 tables to /sys/firmware/acpi/tables/dynamic.
150 acpi_no_watchdog [HW,ACPI,WDT]
151 Ignore the ACPI-based watchdog interface (WDAT) and let
152 a native driver control the watchdog device instead.
154 acpi_rsdp= [ACPI,EFI,KEXEC,EARLY]
155 Pass the RSDP address to the kernel, mostly used
156 on machines running EFI runtime service to boot the
157 second kernel for kdump.
159 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
160 Format: To spoof as Windows 98: ="Microsoft Windows"
162 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
163 of 2 which is mandated by ACPI 6) as the supported ACPI
164 specification revision (when using this switch, it may
165 be necessary to carry out a cold reboot _twice_ in a
166 row to make it take effect on the platform firmware).
168 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
169 acpi_osi="string1" # add string1
170 acpi_osi="!string2" # remove string2
171 acpi_osi=!* # remove all strings
172 acpi_osi=! # disable all built-in OS vendor
174 acpi_osi=!! # enable all built-in OS vendor
176 acpi_osi= # disable all strings
178 'acpi_osi=!' can be used in combination with single or
179 multiple 'acpi_osi="string1"' to support specific OS
180 vendor string(s). Note that such command can only
181 affect the default state of the OS vendor strings, thus
182 it cannot affect the default state of the feature group
183 strings and the current state of the OS vendor strings,
184 specifying it multiple times through kernel command line
185 is meaningless. This command is useful when one do not
186 care about the state of the feature group strings which
187 should be controlled by the OSPM.
189 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
190 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
191 can make '_OSI("Windows 2000")' TRUE.
193 'acpi_osi=' cannot be used in combination with other
194 'acpi_osi=' command lines, the _OSI method will not
195 exist in the ACPI namespace. NOTE that such command can
196 only affect the _OSI support state, thus specifying it
197 multiple times through kernel command line is also
200 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
203 'acpi_osi=!*' can be used in combination with single or
204 multiple 'acpi_osi="string1"' to support specific
205 string(s). Note that such command can affect the
206 current state of both the OS vendor strings and the
207 feature group strings, thus specifying it multiple times
208 through kernel command line is meaningful. But it may
209 still not able to affect the final state of a string if
210 there are quirks related to this string. This command
211 is useful when one want to control the state of the
212 feature group strings to debug BIOS issues related to
215 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
216 '_OSI("Module Device")' FALSE.
217 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
218 '_OSI("Module Device")' TRUE.
219 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
221 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
223 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
224 they all will make '_OSI("Windows 2000")' TRUE.
227 Override the pmtimer bug detection: force the kernel
228 to assume that this machine's pmtimer latches its value
229 and always returns good values.
231 acpi_sci= [HW,ACPI,EARLY] ACPI System Control Interrupt trigger mode
232 Format: { level | edge | high | low }
234 acpi_skip_timer_override [HW,ACPI,EARLY]
235 Recognize and ignore IRQ0/pin2 Interrupt Override.
236 For broken nForce2 BIOS resulting in XT-PIC timer.
238 acpi_sleep= [HW,ACPI] Sleep options
239 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig,
240 s4_nohwsig, old_ordering, nonvs,
241 sci_force_enable, nobl }
242 See Documentation/power/video.rst for information on
244 s3_beep is for debugging; it makes the PC's speaker beep
245 as soon as the kernel's real-mode entry point is called.
246 s4_hwsig causes the kernel to check the ACPI hardware
247 signature during resume from hibernation, and gracefully
248 refuse to resume if it has changed. This complies with
249 the ACPI specification but not with reality, since
250 Windows does not do this and many laptops do change it
251 on docking. So the default behaviour is to allow resume
252 and simply warn when the signature changes, unless the
253 s4_hwsig option is enabled.
254 s4_nohwsig prevents ACPI hardware signature from being
255 used (or even warned about) during resume.
256 old_ordering causes the ACPI 1.0 ordering of the _PTS
257 control method, with respect to putting devices into
258 low power states, to be enforced (the ACPI 2.0 ordering
259 of _PTS is used by default).
260 nonvs prevents the kernel from saving/restoring the
261 ACPI NVS memory during suspend/hibernation and resume.
262 sci_force_enable causes the kernel to set SCI_EN directly
263 on resume from S1/S3 (which is against the ACPI spec,
264 but some broken systems don't work without it).
265 nobl causes the internal blacklist of systems known to
266 behave incorrectly in some ways with respect to system
267 suspend and resume to be ignored (use wisely).
269 acpi_use_timer_override [HW,ACPI,EARLY]
270 Use timer override. For some broken Nvidia NF5 boards
271 that require a timer override, but don't have HPET
273 add_efi_memmap [EFI,X86,EARLY] Include EFI memory map in
274 kernel's map of available physical RAM.
277 { off | try_unsupported }
278 off: disable AGP support
279 try_unsupported: try to drive unsupported chipsets
280 (may crash computer or cause data corruption)
283 See Documentation/sound/alsa-configuration.rst
286 Allow the default userspace alignment fault handler
287 behaviour to be specified. Bit 0 enables warnings,
288 bit 1 enables fixups, and bit 2 sends a segfault.
290 align_va_addr= [X86-64]
291 Align virtual addresses by clearing slice [14:12] when
292 allocating a VMA at process creation time. This option
293 gives you up to 3% performance improvement on AMD F15h
294 machines (where it is enabled by default) for a
295 CPU-intensive style benchmark, and it can vary highly in
296 a microbenchmark depending on workload and compiler.
298 32: only for 32-bit processes
299 64: only for 64-bit processes
300 on: enable for both 32- and 64-bit processes
301 off: disable for both 32- and 64-bit processes
303 alloc_snapshot [FTRACE]
304 Allocate the ftrace snapshot buffer on boot up when the
305 main buffer is allocated. This is handy if debugging
306 and you need to use tracing_snapshot() on boot up, and
307 do not want to use tracing_snapshot_alloc() as it needs
308 to be done where GFP_KERNEL allocations are allowed.
310 allow_mismatched_32bit_el0 [ARM64,EARLY]
311 Allow execve() of 32-bit applications and setting of the
312 PER_LINUX32 personality on systems where only a strict
313 subset of the CPUs support 32-bit EL0. When this
314 parameter is present, the set of CPUs supporting 32-bit
315 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
316 and hot-unplug operations may be restricted.
318 See Documentation/arch/arm64/asymmetric-32bit.rst for more
321 amd_iommu= [HW,X86-64]
322 Pass parameters to the AMD IOMMU driver in the system.
324 fullflush - Deprecated, equivalent to iommu.strict=1
325 off - do not initialize any AMD IOMMU found in
327 force_isolation - Force device isolation for all
328 devices. The IOMMU driver is not
329 allowed anymore to lift isolation
330 requirements as needed. This option
331 does not override iommu=pt
332 force_enable - Force enable the IOMMU on platforms known
333 to be buggy with IOMMU enabled. Use this
335 pgtbl_v1 - Use v1 page table for DMA-API (Default).
336 pgtbl_v2 - Use v2 page table for DMA-API.
337 irtcachedis - Disable Interrupt Remapping Table (IRT) caching.
339 amd_iommu_dump= [HW,X86-64]
340 Enable AMD IOMMU driver option to dump the ACPI table
341 for AMD IOMMU. With this option enabled, AMD IOMMU
342 driver will print ACPI tables for AMD IOMMU during
343 IOMMU initialization.
345 amd_iommu_intr= [HW,X86-64]
346 Specifies one of the following AMD IOMMU interrupt
348 legacy - Use legacy interrupt remapping mode.
349 vapic - Use virtual APIC mode, which allows IOMMU
350 to inject interrupts directly into guest.
351 This mode requires kvm-amd.avic=1.
352 (Default when IOMMU HW support is present.)
354 amd_pstate= [X86,EARLY]
356 Do not enable amd_pstate as the default
357 scaling driver for the supported processors
359 Use amd_pstate with passive mode as a scaling driver.
360 In this mode autonomous selection is disabled.
361 Driver requests a desired performance level and platform
362 tries to match the same performance level if it is
363 satisfied by guaranteed performance level.
365 Use amd_pstate_epp driver instance as the scaling driver,
366 driver provides a hint to the hardware if software wants
367 to bias toward performance (0x0) or energy efficiency (0xff)
368 to the CPPC firmware. then CPPC power algorithm will
369 calculate the runtime workload and adjust the realtime cores
372 Activate guided autonomous mode. Driver requests minimum and
373 maximum performance level and the platform autonomously
374 selects a performance level in this range and appropriate
375 to the current workload.
377 amijoy.map= [HW,JOY] Amiga joystick support
378 Map of devices attached to JOY0DAT and JOY1DAT
380 See also Documentation/input/joydev/joystick.rst
382 analog.map= [HW,JOY] Analog joystick and gamepad support
383 Specifies type or capabilities of an analog joystick
384 connected to one of 16 gameports
385 Format: <type1>,<type2>,..<type16>
388 Power management functions (SPARCstation-4/5 + deriv.)
390 Disable APC CPU standby support. SPARCstation-Fox does
391 not play well with APC CPU idle - disable it if you have
392 APC and your system crashes randomly.
394 apic= [APIC,X86,EARLY] Advanced Programmable Interrupt Controller
395 Change the output verbosity while booting
396 Format: { quiet (default) | verbose | debug }
397 Change the amount of debugging information output
398 when initialising the APIC and IO-APIC components.
399 For X86-32, this can also be used to specify an APIC
401 Format: apic=driver_name
402 Examples: apic=bigsmp
404 apic_extnmi= [APIC,X86,EARLY] External NMI delivery setting
405 Format: { bsp (default) | all | none }
406 bsp: External NMI is delivered only to CPU 0
407 all: External NMIs are broadcast to all CPUs as a
409 none: External NMI is masked for all CPUs. This is
410 useful so that a dump capture kernel won't be
414 See Documentation/networking/ipv6.rst.
416 apm= [APM] Advanced Power Management
417 See header of arch/x86/kernel/apm_32.c.
419 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
420 Format: { "0" | "1" }
421 See security/apparmor/Kconfig help text
424 Default value is set via kernel config option.
426 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
427 Format: <io>,<irq>,<nodeID>
429 arm64.nobti [ARM64] Unconditionally disable Branch Target
430 Identification support
432 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory
433 Set instructions support
435 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
438 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
441 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix
444 arm64.nosve [ARM64] Unconditionally disable Scalable Vector
449 atarimouse= [HW,MOUSE] Atari Mouse
451 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
452 EzKey and similar keyboards
454 atkbd.reset= [HW] Reset keyboard during initialization
456 atkbd.set= [HW] Select keyboard code set
457 Format: <int> (2 = AT (default), 3 = PS/2)
459 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
462 atkbd.softraw= [HW] Choose between synthetic and real raw mode
463 Format: <bool> (0 = real, 1 = synthetic (default))
465 atkbd.softrepeat= [HW]
466 Use software keyboard repeat
468 audit= [KNL] Enable the audit sub-system
469 Format: { "0" | "1" | "off" | "on" }
470 0 | off - kernel audit is disabled and can not be
471 enabled until the next reboot
472 unset - kernel audit is initialized but disabled and
473 will be fully enabled by the userspace auditd.
474 1 | on - kernel audit is initialized and partially
475 enabled, storing at most audit_backlog_limit
476 messages in RAM until it is fully enabled by the
480 audit_backlog_limit= [KNL] Set the audit queue size limit.
481 Format: <int> (must be >=0)
484 bau= [X86_UV] Enable the BAU on SGI UV. The default
485 behavior is to disable the BAU (i.e. bau=0).
486 Format: { "0" | "1" }
489 unset - Disable the BAU.
491 baycom_epp= [HW,AX25]
494 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
496 See header of drivers/net/hamradio/baycom_par.c.
498 baycom_ser_fdx= [HW,AX25]
499 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
500 Format: <io>,<irq>,<mode>[,<baud>]
501 See header of drivers/net/hamradio/baycom_ser_fdx.c.
503 baycom_ser_hdx= [HW,AX25]
504 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
505 Format: <io>,<irq>,<mode>
506 See header of drivers/net/hamradio/baycom_ser_hdx.c.
509 Disable BERT OS support on buggy BIOSes.
511 bgrt_disable [ACPI,X86,EARLY]
512 Disable BGRT to avoid flickering OEM logo.
514 blkdevparts= Manual partition parsing of block device(s) for
515 embedded devices based on command line input.
516 See Documentation/block/cmdline-partition.rst
518 boot_delay= [KNL,EARLY]
519 Milliseconds to delay each printk during boot.
520 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled,
521 and you may also have to specify "lpj=". Boot_delay
522 values larger than 10 seconds (10000) are assumed
523 erroneous and ignored.
526 bootconfig [KNL,EARLY]
527 Extended command line options can be added to an initrd
528 and this will cause the kernel to look for it.
530 See Documentation/admin-guide/bootconfig.rst
532 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
533 bttv.radio= Most important insmod options are available as
535 bttv.pll= See Documentation/admin-guide/media/bttv.rst
538 bulk_remove=off [PPC] This parameter disables the use of the pSeries
539 firmware feature for flushing multiple hpte entries
542 c101= [NET] Moxa C101 synchronous serial card
544 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
545 Sometimes CPU hardware bugs make them report the cache
546 size incorrectly. The kernel will attempt work arounds
547 to fix known problems, but for some CPUs it is not
548 possible to determine what the correct size should be.
549 This option provides an override for these situations.
552 [NET] Specifies amount of time (in seconds) that
553 the kernel should wait for a network carrier. By default
554 it waits 120 seconds.
556 ca_keys= [KEYS] This parameter identifies a specific key(s) on
557 the system trusted keyring to be used for certificate
559 format: { id:<keyid> | builtin }
561 cca= [MIPS,EARLY] Override the kernel pages' cache coherency
562 algorithm. Accepted values range from 0 to 7
563 inclusive. See arch/mips/include/asm/pgtable-bits.h
564 for platform specific values (SB1, Loongson3 and
567 ccw_timeout_log [S390]
568 See Documentation/arch/s390/common_io.rst for details.
570 cgroup_disable= [KNL] Disable a particular controller or optional feature
571 Format: {name of the controller(s) or feature(s) to disable}
572 The effects of cgroup_disable=foo are:
573 - foo isn't auto-mounted if you mount all cgroups in
575 - foo isn't visible as an individually mountable
577 - if foo is an optional feature then the feature is
578 disabled and corresponding cgroup files are not
580 {Currently only "memory" controller deal with this and
581 cut the overhead, others just disable the usage. So
582 only cgroup_disable=memory is actually worthy}
583 Specifying "pressure" disables per-cgroup pressure
584 stall information accounting feature
586 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
587 Format: { { controller | "all" | "named" }
588 [,{ controller | "all" | "named" }...] }
589 Like cgroup_disable, but only applies to cgroup v1;
590 the blacklisted controllers remain available in cgroup2.
591 "all" blacklists all controllers and "named" disables
592 named mounts. Specifying both "all" and "named" disables
595 cgroup_favordynmods= [KNL] Enable or Disable favordynmods.
596 Format: { "true" | "false" }
597 Defaults to the value of CONFIG_CGROUP_FAVOR_DYNMODS.
599 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
601 nosocket -- Disable socket memory accounting.
602 nokmem -- Disable kernel memory accounting.
603 nobpf -- Disable BPF memory accounting.
605 checkreqprot= [SELINUX] Set initial checkreqprot flag value.
606 Format: { "0" | "1" }
607 See security/selinux/Kconfig help text.
608 0 -- check protection applied by kernel (includes
609 any implied execute protection).
610 1 -- check protection requested by application.
611 Default value is set via a kernel config option.
612 Value can be changed at runtime via
613 /sys/fs/selinux/checkreqprot.
614 Setting checkreqprot to 1 is deprecated.
617 See Documentation/arch/s390/common_io.rst for details.
619 clearcpuid=X[,X...] [X86]
620 Disable CPUID feature X for the kernel. See
621 arch/x86/include/asm/cpufeatures.h for the valid bit
622 numbers X. Note the Linux-specific bits are not necessarily
623 stable over kernel options, but the vendor-specific
625 X can also be a string as appearing in the flags: line
626 in /proc/cpuinfo which does not have the above
627 instability issue. However, not all features have names
629 Note that using this option will taint your kernel.
630 Also note that user programs calling CPUID directly
631 or using the feature without checking anything
632 will still see it. This just prevents it from
633 being used by the kernel or shown in /proc/cpuinfo.
634 Also note the kernel might malfunction if you disable
639 Prevents the clock framework from automatically gating
640 clocks that have not been explicitly enabled by a Linux
641 device driver but are enabled in hardware at reset or
642 by the bootloader/firmware. Note that this does not
643 force such clocks to be always-on nor does it reserve
644 those clocks in any way. This parameter is useful for
645 debug and development, but should not be needed on a
646 platform with proper driver support. For more
647 information, see Documentation/driver-api/clk.rst.
649 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
651 Forces specified clocksource (if available) to be used
652 when calculating gettimeofday(). If specified
653 clocksource is not available, it defaults to PIT.
654 Format: { pit | tsc | cyclone | pmtmr }
656 clocksource= Override the default clocksource
658 Override the default clocksource and use the clocksource
659 with the name specified.
660 Some clocksource names to choose from, depending on
662 [all] jiffies (this is the base, fallback clocksource)
664 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
665 pxa_timer,timer3,32k_counter,timer0_1
666 [X86-32] pit,hpet,tsc;
667 scx200_hrt on Geode; cyclone on IBM x440
675 clocksource.arm_arch_timer.evtstrm=
678 Enable/disable the eventstream feature of the ARM
679 architected timer so that code using WFE-based polling
680 loops can be debugged more effectively on production
683 clocksource.verify_n_cpus= [KNL]
684 Limit the number of CPUs checked for clocksources
685 marked with CLOCK_SOURCE_VERIFY_PERCPU that
686 are marked unstable due to excessive skew.
687 A negative value says to check all CPUs, while
688 zero says not to check any. Values larger than
689 nr_cpu_ids are silently truncated to nr_cpu_ids.
690 The actual CPUs are chosen randomly, with
691 no replacement if the same CPU is chosen twice.
693 clocksource-wdtest.holdoff= [KNL]
694 Set the time in seconds that the clocksource
695 watchdog test waits before commencing its tests.
696 Defaults to zero when built as a module and to
697 10 seconds when built into the kernel.
699 cma=nn[MG]@[start[MG][-end[MG]]]
701 Sets the size of kernel global memory area for
702 contiguous memory allocations and optionally the
703 placement constraint by the physical address range of
704 memory allocations. A value of 0 disables CMA
705 altogether. For more information, see
706 kernel/dma/contiguous.c
710 Sets the size of kernel per-numa memory area for
711 contiguous memory allocations. A value of 0 disables
712 per-numa CMA altogether. And If this option is not
713 specified, the default value is 0.
714 With per-numa CMA enabled, DMA users on node nid will
715 first try to allocate buffer from the pernuma area
716 which is located in node nid, if the allocation fails,
717 they will fallback to the global default memory area.
719 numa_cma=<node>:nn[MG][,<node>:nn[MG]]
721 Sets the size of kernel numa memory area for
722 contiguous memory allocations. It will reserve CMA
723 area for the specified node.
725 With numa CMA enabled, DMA users on node nid will
726 first try to allocate buffer from the numa area
727 which is located in node nid, if the allocation fails,
728 they will fallback to the global default memory area.
730 cmo_free_hint= [PPC] Format: { yes | no }
731 Specify whether pages are marked as being inactive
732 when they are freed. This is used in CMO environments
733 to determine OS memory pressure for page stealing by
737 coherent_pool=nn[KMG] [ARM,KNL,EARLY]
738 Sets the size of memory pool for coherent, atomic dma
739 allocations, by default set to 256K.
741 com20020= [HW,NET] ARCnet - COM20020 chipset
743 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
745 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
749 ARCnet - COM90xx chipset (memory-mapped buffers)
750 Format: <io>[,<irq>[,<memstart>]]
752 condev= [HW,S390] console device
755 con3215_drop= [S390,EARLY] 3215 console drop mode.
757 When set to true, drop data on the 3215 console when
758 the console buffer is full. In this case the
759 operator using a 3270 terminal emulator (for example
760 x3270) does not have to enter the clear key for the
761 console output to advance and the kernel to continue.
762 This leads to a much faster boot time when a 3270
763 terminal emulator is active. If no 3270 terminal
764 emulator is used, this parameter has no effect.
766 console= [KNL] Output console device and options.
768 tty<n> Use the virtual console device <n>.
772 Use the specified serial port. The options are of
773 the form "bbbbpnf", where "bbbb" is the baud rate,
774 "p" is parity ("n", "o", or "e"), "n" is number of
775 bits, and "f" is flow control ("r" for RTS or
776 omit it). Default is "9600n8".
778 See Documentation/admin-guide/serial-console.rst for more
780 Documentation/networking/netconsole.rst for an
783 uart[8250],io,<addr>[,options]
784 uart[8250],mmio,<addr>[,options]
785 uart[8250],mmio16,<addr>[,options]
786 uart[8250],mmio32,<addr>[,options]
787 uart[8250],0x<addr>[,options]
788 Start an early, polled-mode console on the 8250/16550
789 UART at the specified I/O port or MMIO address,
790 switching to the matching ttyS device later.
791 MMIO inter-register address stride is either 8-bit
792 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
793 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
794 to be equivalent to 'mmio'. 'options' are specified in
795 the same format described for ttyS above; if unspecified,
796 the h/w is not re-initialized.
798 hvc<n> Use the hypervisor console device <n>. This is for
799 both Xen and PowerPC hypervisors.
802 Use to disable console output, i.e., to have kernel
803 console messages discarded.
804 This must be the only console= parameter used on the
807 If the device connected to the port is not a TTY but a braille
808 device, prepend "brl," before the device type, for instance
810 For now, only VisioBraille is supported.
813 [KNL] Change console messages format
815 By default we print messages on consoles in
816 "[time stamp] text\n" format (time stamp may not be
817 printed, depending on CONFIG_PRINTK_TIME or
818 `printk_time' param).
820 Switch to syslog format: "<%u>[time stamp] text\n"
821 IOW, each message will have a facility and loglevel
822 prefix. The format is similar to one used by syslog()
823 syscall, or to executing "dmesg -S --raw" or to reading
826 consoleblank= [KNL] The console blank (screen saver) timeout in
827 seconds. A value of 0 disables the blank timer.
831 [KNL] Change the default value for
832 /proc/<pid>/coredump_filter.
833 See also Documentation/filesystems/proc.rst.
835 coresight_cpu_debug.enable
838 Enable/disable the CPU sampling based debugging.
839 0: default value, disable debugging
840 1: enable debugging at boot time
842 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
844 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
846 cpuidle.off=1 [CPU_IDLE]
847 disable the cpuidle sub-system
850 [CPU_IDLE] Name of the cpuidle governor to use.
852 cpufreq.off=1 [CPU_FREQ]
853 disable the cpufreq sub-system
855 cpufreq.default_governor=
856 [CPU_FREQ] Name of the default cpufreq governor or
857 policy to use. This governor must be registered in the
858 kernel before the cpufreq driver probes.
861 [X86,EARLY] Delay for N microsec between assert and de-assert
862 of APIC INIT to start processors. This delay occurs
863 on every CPU online, such as boot, and resume from suspend.
867 [SMP] Enable/disable parallel bringup of secondary CPUs
869 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise
870 the parameter has no effect.
872 crash_kexec_post_notifiers
873 Run kdump after running panic-notifiers and dumping
874 kmsg. This only for the users who doubt kdump always
875 succeeds in any situation.
876 Note that this also increases risks of kdump failure,
877 because some panic notifiers can make the crashed
878 kernel more unstable.
880 crashkernel=size[KMG][@offset[KMG]]
881 [KNL,EARLY] Using kexec, Linux can switch to a 'crash kernel'
882 upon panic. This parameter reserves the physical
883 memory region [offset, offset + size] for that kernel
884 image. If '@offset' is omitted, then a suitable offset
885 is selected automatically.
886 [KNL, X86-64, ARM64, RISCV, LoongArch] Select a region
887 under 4G first, and fall back to reserve region above
888 4G when '@offset' hasn't been specified.
889 See Documentation/admin-guide/kdump/kdump.rst for further details.
891 crashkernel=range1:size1[,range2:size2,...][@offset]
892 [KNL] Same as above, but depends on the memory
893 in the running system. The syntax of range is
894 start-[end] where start and end are both
895 a memory unit (amount[KMG]). See also
896 Documentation/admin-guide/kdump/kdump.rst for an example.
898 crashkernel=size[KMG],high
899 [KNL, X86-64, ARM64, RISCV, LoongArch] range could be
901 Allow kernel to allocate physical memory region from top,
902 so could be above 4G if system have more than 4G ram
903 installed. Otherwise memory region will be allocated
904 below 4G, if available.
905 It will be ignored if crashkernel=X is specified.
906 crashkernel=size[KMG],low
907 [KNL, X86-64, ARM64, RISCV, LoongArch] range under 4G.
908 When crashkernel=X,high is passed, kernel could allocate
909 physical memory region above 4G, that cause second kernel
910 crash on system that require some amount of low memory,
911 e.g. swiotlb requires at least 64M+32K low memory, also
912 enough extra low memory is needed to make sure DMA buffers
913 for 32-bit devices won't run out. Kernel would try to allocate
914 default size of memory below 4G automatically. The default
915 size is platform dependent.
916 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB)
919 --> loongarch: 128MiB
920 This one lets the user specify own low range under 4G
921 for second kernel instead.
922 0: to disable low allocation.
923 It will be ignored when crashkernel=X,high is not used
924 or memory reserved is below 4G.
927 [KNL] Disable crypto self-tests
932 cs89x0_media= [HW,NET]
933 Format: { rj45 | aui | bnc }
935 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU
936 function call handling. When switched on,
937 additional debug data is printed to the console
938 in case a hanging CPU is detected, and that
939 CPU is pinged again in order to try to resolve
940 the hang situation. The default value of this
941 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT
945 See header of drivers/s390/block/dasd_devmap.c.
947 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
948 (one device per port)
949 Format: <port#>,<type>
950 See also Documentation/input/devices/joystick-parport.rst
952 debug [KNL,EARLY] Enable kernel debugging (events log level).
955 [KNL,EARLY] Enable printing [hashed] pointers early in the
956 boot sequence. If enabled, we use a weak hash instead
957 of siphash to hash pointers. Use this option if you are
958 seeing instances of '(___ptrval___)') and need to see a
959 value (hashed pointer) instead. Cryptographically
960 insecure, please do not use on production kernels.
963 [KNL] verbose locking self-tests
965 Print debugging info while doing the locking API
967 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
968 (no extra messages), setting it to -1 (all bits set)
969 will print _a_lot_ more information - normally only
970 useful to lockdep developers.
972 debug_objects [KNL,EARLY] Enable object debugging
974 debug_guardpage_minorder=
975 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this
976 parameter allows control of the order of pages that will
977 be intentionally kept free (and hence protected) by the
978 buddy allocator. Bigger value increase the probability
979 of catching random memory corruption, but reduce the
980 amount of memory for normal system use. The maximum
981 possible value is MAX_PAGE_ORDER/2. Setting this
982 parameter to 1 or 2 should be enough to identify most
983 random memory corruption problems caused by bugs in
984 kernel or driver code when a CPU writes to (or reads
985 from) a random memory location. Note that there exists
986 a class of memory corruptions problems caused by buggy
987 H/W or F/W or by drivers badly programming DMA
988 (basically when memory is written at bus level and the
989 CPU MMU is bypassed) which are not detectable by
990 CONFIG_DEBUG_PAGEALLOC, hence this option will not
991 help tracking down these problems.
994 [KNL,EARLY] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
995 enables the feature at boot time. By default, it is
996 disabled and the system will work mostly the same as a
997 kernel built without CONFIG_DEBUG_PAGEALLOC.
998 Note: to get most of debug_pagealloc error reports, it's
999 useful to also enable the page_owner functionality.
1000 on: enable the feature
1002 debugfs= [KNL,EARLY] This parameter enables what is exposed to
1003 userspace and debugfs internal clients.
1004 Format: { on, no-mount, off }
1005 on: All functions are enabled.
1007 Filesystem is not registered but kernel clients can
1008 access APIs and a crashkernel can be used to read
1009 its content. There is nothing to mount.
1010 off: Filesystem is not registered and clients
1011 get a -EPERM as result when trying to register files
1012 or directories within debugfs.
1013 This is equivalent of the runtime functionality if
1014 debugfs was not enabled in the kernel at all.
1015 Default value is set in build-time with a kernel configuration.
1017 debugpat [X86] Enable PAT debugging
1020 [HW] The size of the default HugeTLB page. This is
1021 the size represented by the legacy /proc/ hugepages
1022 APIs. In addition, this is the default hugetlb size
1023 used for shmget(), mmap() and mounting hugetlbfs
1024 filesystems. If not specified, defaults to the
1025 architecture's default huge page size. Huge page
1026 sizes are architecture dependent. See also
1027 Documentation/admin-guide/mm/hugetlbpage.rst.
1030 deferred_probe_timeout=
1031 [KNL] Debugging option to set a timeout in seconds for
1032 deferred probe to give up waiting on dependencies to
1033 probe. Only specific dependencies (subsystems or
1034 drivers) that have opted in will be ignored. A timeout
1035 of 0 will timeout at the end of initcalls. If the time
1036 out hasn't expired, it'll be restarted by each
1037 successful driver registration. This option will also
1038 dump out devices still on the deferred probe list after
1041 delayacct [KNL] Enable per-task delay accounting
1043 dell_smm_hwmon.ignore_dmi=
1044 [HW] Continue probing hardware even if DMI data
1045 indicates that the driver is running on unsupported
1048 dell_smm_hwmon.force=
1049 [HW] Activate driver even if SMM BIOS signature does
1050 not match list of supported models and enable otherwise
1051 blacklisted features.
1053 dell_smm_hwmon.power_status=
1054 [HW] Report power status in /proc/i8k
1055 (disabled by default).
1057 dell_smm_hwmon.restricted=
1058 [HW] Allow controlling fans only if SYS_ADMIN
1061 dell_smm_hwmon.fan_mult=
1062 [HW] Factor to multiply fan speed with.
1064 dell_smm_hwmon.fan_max=
1065 [HW] Maximum configurable fan speed.
1068 Format: { on | off | def_only | inf_only | always }
1069 on: s390 zlib hardware support for compression on
1070 level 1 and decompression (default)
1071 off: No s390 zlib hardware support
1072 def_only: s390 zlib hardware support for deflate
1073 only (compression on level 1)
1074 inf_only: s390 zlib hardware support for inflate
1075 only (decompression)
1076 always: Same as 'on' but ignores the selected compression
1077 level always using hardware support (used for debugging)
1079 dhash_entries= [KNL]
1080 Set number of hash buckets for dentry cache.
1082 disable_1tb_segments [PPC,EARLY]
1083 Disables the use of 1TB hash page table segments. This
1084 causes the kernel to fall back to 256MB segments which
1085 can be useful when debugging issues that require an SLB
1089 See Documentation/networking/ipv6.rst.
1091 disable_radix [PPC,EARLY]
1092 Disable RADIX MMU mode on POWER9
1095 Disable TLBIE instruction. Currently does not work
1096 with KVM, with HASH MMU, or with coherent accelerators.
1098 disable_ddw [PPC/PSERIES,EARLY]
1099 Disable Dynamic DMA Window support. Use this
1100 to workaround buggy firmware.
1102 disable_ipv6= [IPV6]
1103 See Documentation/networking/ipv6.rst.
1105 disable_mtrr_cleanup [X86,EARLY]
1106 The kernel tries to adjust MTRR layout from continuous
1107 to discrete, to make X server driver able to add WB
1108 entry later. This parameter disables that.
1110 disable_mtrr_trim [X86, Intel and AMD only,EARLY]
1111 By default the kernel will trim any uncacheable
1112 memory out of your available memory pool based on
1113 MTRR settings. This parameter disables that behavior,
1114 possibly causing your machine to run very slowly.
1116 disable_timer_pin_1 [X86,EARLY]
1117 Disable PIN 1 of APIC timer
1118 Can be useful to work around chipset bugs.
1120 dis_ucode_ldr [X86] Disable the microcode loader.
1122 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1123 this option disables the debugging code at boot.
1125 dma_debug_entries=<number>
1126 This option allows to tune the number of preallocated
1127 entries for DMA-API debugging code. One entry is
1128 required per DMA-API allocation. Use this if the
1129 DMA-API debugging code disables itself because the
1130 architectural default is too low.
1132 dma_debug_driver=<driver_name>
1133 With this option the DMA-API debugging driver
1134 filter feature can be enabled at boot time. Just
1135 pass the driver to filter for as the parameter.
1136 The filter can be disabled or changed to another
1137 driver later using sysfs.
1139 reg_file_data_sampling=
1140 [X86] Controls mitigation for Register File Data
1141 Sampling (RFDS) vulnerability. RFDS is a CPU
1142 vulnerability which may allow userspace to infer
1143 kernel data values previously stored in floating point
1144 registers, vector registers, or integer registers.
1145 RFDS only affects Intel Atom processors.
1147 on: Turns ON the mitigation.
1148 off: Turns OFF the mitigation.
1150 This parameter overrides the compile time default set
1151 by CONFIG_MITIGATION_RFDS. Mitigation cannot be
1152 disabled when other VERW based mitigations (like MDS)
1153 are enabled. In order to disable RFDS mitigation all
1154 VERW based mitigations need to be disabled.
1157 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst
1159 driver_async_probe= [KNL]
1160 List of driver names to be probed asynchronously. *
1161 matches with all driver names. If * is specified, the
1162 rest of the listed driver names are those that will NOT
1164 Format: <driver_name1>,<driver_name2>...
1166 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1167 Broken monitors, graphic adapters, KVMs and EDIDless
1168 panels may send no or incorrect EDID data sets.
1169 This parameter allows to specify an EDID data sets
1170 in the /lib/firmware directory that are used instead.
1171 Generic built-in EDID data sets are used, if one of
1172 edid/1024x768.bin, edid/1280x1024.bin,
1173 edid/1680x1050.bin, or edid/1920x1080.bin is given
1174 and no file with the same name exists. Details and
1175 instructions how to build your own EDID data are
1176 available in Documentation/admin-guide/edid.rst. An EDID
1177 data set will only be used for a particular connector,
1178 if its name and a colon are prepended to the EDID
1179 name. Each connector may use a unique EDID data
1180 set by separating the files with a comma. An EDID
1181 data set with no connector name will be used for
1182 any connectors not explicitly specified.
1186 dt_cpu_ftrs= [PPC,EARLY]
1187 Format: {"off" | "known"}
1188 Control how the dt_cpu_ftrs device-tree binding is
1189 used for CPU feature discovery and setup (if it
1191 off: Do not use it, fall back to legacy cpu table.
1192 known: Do not pass through unknown features to guests
1193 or userspace, only those that the kernel is aware of.
1195 dump_apple_properties [X86]
1196 Dump name and content of EFI device properties on
1197 x86 Macs. Useful for driver authors to determine
1198 what data is available or for reverse-engineering.
1200 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1201 <module>.dyndbg[="val"]
1202 Enable debug messages at boot time. See
1203 Documentation/admin-guide/dynamic-debug-howto.rst
1206 early_ioremap_debug [KNL,EARLY]
1207 Enable debug messages in early_ioremap support. This
1208 is useful for tracking down temporary early mappings
1209 which are not unmapped.
1211 earlycon= [KNL,EARLY] Output early console device and options.
1213 When used with no options, the early console is
1214 determined by stdout-path property in device tree's
1215 chosen node or the ACPI SPCR table if supported by
1218 cdns,<addr>[,options]
1219 Start an early, polled-mode console on a Cadence
1220 (xuartps) serial port at the specified address. Only
1221 supported option is baud rate. If baud rate is not
1222 specified, the serial port must already be setup and
1225 uart[8250],io,<addr>[,options[,uartclk]]
1226 uart[8250],mmio,<addr>[,options[,uartclk]]
1227 uart[8250],mmio32,<addr>[,options[,uartclk]]
1228 uart[8250],mmio32be,<addr>[,options[,uartclk]]
1229 uart[8250],0x<addr>[,options]
1230 Start an early, polled-mode console on the 8250/16550
1231 UART at the specified I/O port or MMIO address.
1232 MMIO inter-register address stride is either 8-bit
1233 (mmio) or 32-bit (mmio32 or mmio32be).
1234 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1235 to be equivalent to 'mmio'. 'options' are specified
1236 in the same format described for "console=ttyS<n>"; if
1237 unspecified, the h/w is not initialized. 'uartclk' is
1238 the uart clock frequency; if unspecified, it is set
1239 to 'BASE_BAUD' * 16.
1243 Start an early, polled-mode console on a pl011 serial
1244 port at the specified address. The pl011 serial port
1245 must already be setup and configured. Options are not
1246 yet supported. If 'mmio32' is specified, then only
1247 the driver will use only 32-bit accessors to read/write
1248 the device registers.
1251 Start an early console on a litex serial port at the
1252 specified address. The serial port must already be
1253 setup and configured. Options are not yet supported.
1256 Start an early, polled-mode console on a meson serial
1257 port at the specified address. The serial port must
1258 already be setup and configured. Options are not yet
1262 Start an early, polled-mode console on an msm serial
1263 port at the specified address. The serial port
1264 must already be setup and configured. Options are not
1267 msm_serial_dm,<addr>
1268 Start an early, polled-mode console on an msm serial
1269 dm port at the specified address. The serial port
1270 must already be setup and configured. Options are not
1274 Start an early, polled-mode console on a serial port
1275 of an Actions Semi SoC, such as S500 or S900, at the
1276 specified address. The serial port must already be
1277 setup and configured. Options are not yet supported.
1280 Start an early, polled-mode console on a serial port
1281 of an RDA Micro SoC, such as RDA8810PL, at the
1282 specified address. The serial port must already be
1283 setup and configured. Options are not yet supported.
1286 Use RISC-V SBI (Supervisor Binary Interface) for early
1289 smh Use ARM semihosting calls for early console.
1297 Use early console provided by serial driver available
1298 on Samsung SoCs, requires selecting proper type and
1299 a correct base address of the selected UART port. The
1300 serial port must already be setup and configured.
1301 Options are not yet supported.
1304 Start an early, polled-mode console on a lantiq serial
1305 (lqasc) port at the specified address. The serial port
1306 must already be setup and configured. Options are not
1311 Use early console provided by Freescale LP UART driver
1312 found on Freescale Vybrid and QorIQ LS1021A processors.
1313 A valid base address must be provided, and the serial
1314 port must already be setup and configured.
1318 Start an early, polled-mode, output-only console on the
1319 Freescale i.MX UART at the specified address. The UART
1320 must already be setup and configured.
1323 Start an early, polled-mode console on the
1324 Armada 3700 serial port at the specified
1325 address. The serial port must already be setup
1326 and configured. Options are not yet supported.
1329 Start an early, polled-mode console on a Qualcomm
1330 Generic Interface (GENI) based serial port at the
1331 specified address. The serial port must already be
1332 setup and configured. Options are not yet supported.
1335 Start an early, unaccelerated console on the EFI
1336 memory mapped framebuffer (if available). On cache
1337 coherent non-x86 systems that use system memory for
1338 the framebuffer, pass the 'ram' option so that it is
1339 mapped with the correct attributes.
1342 Use early console provided by Freescale LINFlexD UART
1343 serial driver for NXP S32V234 SoCs. A valid base
1344 address must be provided, and the serial port must
1345 already be setup and configured.
1347 earlyprintk= [X86,SH,ARM,M68k,S390,UM,EARLY]
1351 earlyprintk=serial[,ttySn[,baudrate]]
1352 earlyprintk=serial[,0x...[,baudrate]]
1353 earlyprintk=ttySn[,baudrate]
1354 earlyprintk=dbgp[debugController#]
1355 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1356 earlyprintk=xdbc[xhciController#]
1359 earlyprintk is useful when the kernel crashes before
1360 the normal console is initialized. It is not enabled by
1361 default because it has some cosmetic problems.
1363 Append ",keep" to not disable it when the real console
1366 Only one of vga, serial, or usb debug port can
1369 Currently only ttyS0 and ttyS1 may be specified by
1370 name. Other I/O ports may be explicitly specified
1371 on some architectures (x86 and arm at least) by
1372 replacing ttySn with an I/O port address, like this:
1373 earlyprintk=serial,0x1008,115200
1374 You can find the port for a given device in
1375 /proc/tty/driver/serial:
1376 2: uart:ST16650V2 port:00001008 irq:18 ...
1378 Interaction with the standard serial driver is not
1381 The VGA output is eventually overwritten by
1384 The xen option can only be used in Xen domains.
1386 The sclp output can only be used on s390.
1388 The bios output can only be used on SuperH.
1390 The optional "force" to "pciserial" enables use of a
1391 PCI device even when its classcode is not of the
1394 edac_report= [HW,EDAC] Control how to report EDAC event
1395 Format: {"on" | "off" | "force"}
1396 on: enable EDAC to report H/W event. May be overridden
1397 by other higher priority error reporting module.
1398 off: disable H/W event reporting through EDAC.
1399 force: enforce the use of EDAC to report H/W event.
1403 Format: {"off" | "on" | "skip[mbr]"}
1406 Format: { "debug", "disable_early_pci_dma",
1407 "nochunk", "noruntime", "nosoftreserve",
1408 "novamap", "no_disable_early_pci_dma" }
1409 debug: enable misc debug output.
1410 disable_early_pci_dma: disable the busmaster bit on all
1411 PCI bridges while in the EFI boot stub.
1412 nochunk: disable reading files in "chunks" in the EFI
1413 boot stub, as chunking can cause problems with some
1414 firmware implementations.
1415 noruntime : disable EFI runtime services support
1416 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1417 attribute may cause the kernel to reserve the
1418 memory range for a memory mapping driver to
1419 claim. Specify efi=nosoftreserve to disable this
1420 reservation and treat the memory by its base type
1421 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1422 novamap: do not call SetVirtualAddressMap().
1423 no_disable_early_pci_dma: Leave the busmaster bit set
1424 on all PCI bridges while in the EFI boot stub
1426 efi_no_storage_paranoia [EFI,X86,EARLY]
1427 Using this parameter you can use more than 50% of
1428 your efi variable storage. Use this parameter only if
1429 you are really sure that your UEFI does sane gc and
1430 fulfills the spec otherwise your board may brick.
1432 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI,X86,EARLY]
1433 Add arbitrary attribute to specific memory range by
1434 updating original EFI memory map.
1435 Region of memory which aa attribute is added to is
1438 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1439 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1440 attribute is added to range 0x100000000-0x180000000 and
1441 0x10a0000000-0x1120000000.
1443 If efi_fake_mem=8G@9G:0x40000 is specified, the
1444 EFI_MEMORY_SP(0x40000) attribute is added to
1445 range 0x240000000-0x43fffffff.
1447 Using this parameter you can do debugging of EFI memmap
1448 related features. For example, you can do debugging of
1449 Address Range Mirroring feature even if your box
1450 doesn't support it, or mark specific memory as
1453 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1454 that is to be dynamically loaded by Linux. If there are
1455 multiple variables with the same name but with different
1456 vendor GUIDs, all of them will be loaded. See
1457 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1460 eisa_irq_edge= [PARISC,HW]
1461 See header of drivers/parisc/eisa.c.
1463 ekgdboc= [X86,KGDB,EARLY] Allow early kernel console debugging
1466 This is designed to be used in conjunction with
1467 the boot argument: earlyprintk=vga
1469 This parameter works in place of the kgdboc parameter
1470 but can only be used if the backing tty is available
1471 very early in the boot process. For early debugging
1472 via a serial port see kgdboc_earlycon instead.
1475 See comment before function elanfreq_setup() in
1476 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1478 elfcorehdr=[size[KMG]@]offset[KMG] [PPC,SH,X86,S390,EARLY]
1479 Specifies physical address of start of kernel core
1480 image elf header and optionally the size. Generally
1481 kexec loader will pass this option to capture kernel.
1482 See Documentation/admin-guide/kdump/kdump.rst for details.
1484 enable_mtrr_cleanup [X86,EARLY]
1485 The kernel tries to adjust MTRR layout from continuous
1486 to discrete, to make X server driver able to add WB
1487 entry later. This parameter enables that.
1489 enable_timer_pin_1 [X86]
1490 Enable PIN 1 of APIC timer
1491 Can be useful to work around chipset bugs
1492 (in particular on some ATI chipsets).
1493 The kernel tries to set a reasonable default.
1495 enforcing= [SELINUX] Set initial enforcing status.
1497 See security/selinux/Kconfig help text.
1498 0 -- permissive (log only, no denials).
1499 1 -- enforcing (deny and log).
1501 Value can be changed at runtime via
1502 /sys/fs/selinux/enforce.
1505 Disable Error Record Serialization Table (ERST)
1508 ether= [HW,NET] Ethernet cards parameters
1509 This option is obsoleted by the "netdev=" option, which
1510 has equivalent usage. See its documentation for details.
1514 Permit 'security.evm' to be updated regardless of
1515 current integrity status.
1517 early_page_ext [KNL,EARLY] Enforces page_ext initialization to earlier
1518 stages so cover more early boot allocations.
1519 Please note that as side effect some optimizations
1520 might be disabled to achieve that (e.g. parallelized
1521 memory initialization is disabled) so the boot process
1522 might take longer, especially on systems with a lot of
1523 memory. Available with CONFIG_PAGE_EXTENSION=y.
1528 fail_make_request=[KNL]
1529 General fault injection mechanism.
1530 Format: <interval>,<probability>,<space>,<times>
1531 See also Documentation/fault-injection/.
1534 Format: { initns | none }
1535 See Documentation/admin-guide/sysctl/net.rst for
1536 fb_tunnels_only_for_init_ns
1539 See Documentation/admin-guide/blockdev/floppy.rst.
1542 Forcefully enable Physical Address Extension (PAE).
1543 Many Pentium M systems disable PAE but may have a
1544 functionally usable PAE implementation.
1545 Warning: use of this parameter will taint the kernel
1546 and may cause unknown problems.
1549 Enable/disable Flexible Return and Event Delivery.
1550 Format: { on | off }
1551 on: enable FRED when it's present.
1552 off: disable FRED, the default setting.
1555 [FTRACE] will set and start the specified tracer
1556 as early as possible in order to facilitate early
1559 ftrace_boot_snapshot
1560 [FTRACE] On boot up, a snapshot will be taken of the
1561 ftrace ring buffer that can be read at:
1562 /sys/kernel/tracing/snapshot.
1563 This is useful if you need tracing information from kernel
1564 boot up that is likely to be overridden by user space
1565 start up functionality.
1567 Optionally, the snapshot can also be defined for a tracing
1568 instance that was created by the trace_instance= command
1571 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo
1573 The above will cause the "foo" tracing instance to trigger
1574 a snapshot at the end of boot up.
1576 ftrace_dump_on_oops[=orig_cpu]
1577 [FTRACE] will dump the trace buffers on oops.
1578 If no parameter is passed, ftrace will dump
1579 buffers of all CPUs, but if you pass orig_cpu, it will
1580 dump only the buffer of the CPU that triggered the
1583 ftrace_filter=[function-list]
1584 [FTRACE] Limit the functions traced by the function
1585 tracer at boot up. function-list is a comma-separated
1586 list of functions. This list can be changed at run
1587 time by the set_ftrace_filter file in the debugfs
1590 ftrace_notrace=[function-list]
1591 [FTRACE] Do not trace the functions specified in
1592 function-list. This list can be changed at run time
1593 by the set_ftrace_notrace file in the debugfs
1596 ftrace_graph_filter=[function-list]
1597 [FTRACE] Limit the top level callers functions traced
1598 by the function graph tracer at boot up.
1599 function-list is a comma-separated list of functions
1600 that can be changed at run time by the
1601 set_graph_function file in the debugfs tracing directory.
1603 ftrace_graph_notrace=[function-list]
1604 [FTRACE] Do not trace from the functions specified in
1605 function-list. This list is a comma-separated list of
1606 functions that can be changed at run time by the
1607 set_graph_notrace file in the debugfs tracing directory.
1609 ftrace_graph_max_depth=<uint>
1610 [FTRACE] Used with the function graph tracer. This is
1611 the max depth it will trace into a function. This value
1612 can be changed at run time by the max_graph_depth file
1613 in the tracefs tracing directory. default: 0 (no limit)
1615 fw_devlink= [KNL,EARLY] Create device links between consumer and supplier
1616 devices by scanning the firmware to infer the
1617 consumer/supplier relationships. This feature is
1618 especially useful when drivers are loaded as modules as
1619 it ensures proper ordering of tasks like device probing
1620 (suppliers first, then consumers), supplier boot state
1621 clean up (only after all consumers have probed),
1622 suspend/resume & runtime PM (consumers first, then
1624 Format: { off | permissive | on | rpm }
1625 off -- Don't create device links from firmware info.
1626 permissive -- Create device links from firmware info
1627 but use it only for ordering boot state clean
1628 up (sync_state() calls).
1629 on -- Create device links from firmware info and use it
1630 to enforce probe and suspend/resume ordering.
1631 rpm -- Like "on", but also use to order runtime PM.
1633 fw_devlink.strict=<bool>
1634 [KNL,EARLY] Treat all inferred dependencies as mandatory
1635 dependencies. This only applies for fw_devlink=on|rpm.
1638 fw_devlink.sync_state =
1639 [KNL,EARLY] When all devices that could probe have finished
1640 probing, this parameter controls what to do with
1641 devices that haven't yet received their sync_state()
1643 Format: { strict | timeout }
1644 strict -- Default. Continue waiting on consumers to
1646 timeout -- Give up waiting on consumers and call
1647 sync_state() on any devices that haven't yet
1648 received their sync_state() calls after
1649 deferred_probe_timeout has expired or by
1650 late_initcall() if !CONFIG_MODULES.
1653 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1654 support via parallel port (up to 5 devices per port)
1655 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1656 See also Documentation/input/devices/joystick-parport.rst
1660 gart_fix_e820= [X86-64,EARLY] disable the fix e820 for K8 GART
1664 gather_data_sampling=
1665 [X86,INTEL,EARLY] Control the Gather Data Sampling (GDS)
1668 Gather Data Sampling is a hardware vulnerability which
1669 allows unprivileged speculative access to data which was
1670 previously stored in vector registers.
1672 This issue is mitigated by default in updated microcode.
1673 The mitigation may have a performance impact but can be
1674 disabled. On systems without the microcode mitigation
1675 disabling AVX serves as a mitigation.
1677 force: Disable AVX to mitigate systems without
1678 microcode mitigation. No effect if the microcode
1679 mitigation is present. Known to cause crashes in
1680 userspace with buggy AVX enumeration.
1682 off: Disable GDS mitigation.
1684 gcov_persist= [GCOV] When non-zero (default), profiling data for
1685 kernel modules is saved and remains accessible via
1686 debugfs, even when the module is unloaded/reloaded.
1687 When zero, profiling data is discarded and associated
1688 debugfs files are removed at module unload time.
1690 goldfish [X86] Enable the goldfish android emulator platform.
1691 Don't use this when you are not running on the
1694 gpio-mockup.gpio_mockup_ranges
1695 [HW] Sets the ranges of gpiochip of for this device.
1696 Format: <start1>,<end1>,<start2>,<end2>...
1697 gpio-mockup.gpio_mockup_named_lines
1698 [HW] Let the driver know GPIO lines should be named.
1700 gpt [EFI] Forces disk with valid GPT signature but
1701 invalid Protective MBR to be treated as GPT. If the
1702 primary GPT is corrupted, it enables the backup/alternate
1703 GPT to be used instead.
1705 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1706 the "Enable 0" bit of the configuration register.
1709 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1710 the "Enable 0" bit of the configuration register.
1713 grcan.select= [HW] Select which physical interface to use.
1716 grcan.txsize= [HW] Sets the size of the tx buffer.
1717 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1719 grcan.rxsize= [HW] Sets the size of the rx buffer.
1720 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1724 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1725 hardening is enabled for this boot. Hardened
1726 usercopy checking is used to protect the kernel
1727 from reading or writing beyond known memory
1728 allocation boundaries as a proactive defense
1729 against bounds-checking flaws in the kernel's
1730 copy_to_user()/copy_from_user() interface.
1731 on Perform hardened usercopy checks (default).
1732 off Disable hardened usercopy checks.
1734 hardlockup_all_cpu_backtrace=
1735 [KNL] Should the hard-lockup detector generate
1736 backtraces on all cpus.
1739 hashdist= [KNL,NUMA] Large hashes allocated during boot
1740 are distributed across NUMA nodes. Defaults on
1741 for 64-bit NUMA, off otherwise.
1742 Format: 0 | 1 (for off | on)
1744 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1746 hd= [EIDE] (E)IDE hard drive subsystem geometry
1747 Format: <cyl>,<head>,<sect>
1750 Disable Hardware Error Source Table (HEST) support;
1751 corresponding firmware-first mode error processing
1752 logic will be disabled.
1754 hibernate= [HIBERNATION]
1755 noresume Don't check if there's a hibernation image
1756 present during boot.
1757 nocompress Don't compress/decompress hibernation images.
1758 no Disable hibernation and resume.
1759 protect_image Turn on image protection during restoration
1760 (that will set all pages holding image data
1761 during restoration read-only).
1763 highmem=nn[KMG] [KNL,BOOT,EARLY] forces the highmem zone to have an exact
1764 size of <nn>. This works even on boxes that have no
1765 highmem otherwise. This also works to reduce highmem
1766 size on bigger boxes.
1768 highres= [KNL] Enable/disable high resolution timer mode.
1769 Valid parameters: "on", "off"
1774 hostname= [KNL,EARLY] Set the hostname (aka UTS nodename).
1776 This allows setting the system's hostname during early
1777 startup. This sets the name returned by gethostname.
1778 Using this parameter to set the hostname makes it
1779 possible to ensure the hostname is correctly set before
1780 any userspace processes run, avoiding the possibility
1781 that a process may call gethostname before the hostname
1782 has been explicitly set, resulting in the calling
1783 process getting an incorrect result. The string must
1784 not exceed the maximum allowed hostname length (usually
1785 64 characters) and will be truncated otherwise.
1787 hpet= [X86-32,HPET] option to control HPET usage
1788 Format: { enable (default) | disable | force |
1790 disable: disable HPET and use PIT instead
1791 force: allow force enabled of undocumented chips (ICH4,
1793 verbose: show contents of HPET registers during setup
1795 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1796 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1798 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1799 If this follows hugepagesz (below), it specifies
1800 the number of pages of hugepagesz to be allocated.
1801 If this is the first HugeTLB parameter on the command
1802 line, it specifies the number of pages to allocate for
1803 the default huge page size. If using node format, the
1804 number of pages to allocate per-node can be specified.
1805 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1806 Format: <integer> or (node format)
1807 <node>:<integer>[,<node>:<integer>]
1810 [HW] The size of the HugeTLB pages. This is used in
1811 conjunction with hugepages (above) to allocate huge
1812 pages of a specific size at boot. The pair
1813 hugepagesz=X hugepages=Y can be specified once for
1814 each supported huge page size. Huge page sizes are
1815 architecture dependent. See also
1816 Documentation/admin-guide/mm/hugetlbpage.rst.
1819 hugetlb_cma= [HW,CMA,EARLY] The size of a CMA area used for allocation
1820 of gigantic hugepages. Or using node format, the size
1821 of a CMA area per node can be specified.
1822 Format: nn[KMGTPE] or (node format)
1823 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1825 Reserve a CMA area of given size and allocate gigantic
1826 hugepages using the CMA allocator. If enabled, the
1827 boot-time allocation of gigantic hugepages is skipped.
1829 hugetlb_free_vmemmap=
1830 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1832 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1833 Allows heavy hugetlb users to free up some more
1834 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1835 Format: { on | off (default) }
1840 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1843 Note that the vmemmap pages may be allocated from the added
1844 memory block itself when memory_hotplug.memmap_on_memory is
1845 enabled, those vmemmap pages cannot be optimized even if this
1846 feature is enabled. Other vmemmap pages not allocated from
1847 the added memory block itself do not be affected.
1850 [KNL] Should the hung task detector generate panics.
1853 A value of 1 instructs the kernel to panic when a
1854 hung task is detected. The default value is controlled
1855 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1856 option. The value selected by this boot parameter can
1857 be changed later by the kernel.hung_task_panic sysctl.
1859 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1860 terminal devices. Valid values: 0..8
1861 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1862 If specified, z/VM IUCV HVC accepts connections
1863 from listed z/VM user IDs only.
1865 hv_nopvspin [X86,HYPER_V,EARLY]
1866 Disables the paravirt spinlock optimizations
1867 which allow the hypervisor to 'idle' the guest
1870 i2c_bus= [HW] Override the default board specific I2C bus speed
1871 or register an additional I2C bus that is not
1872 registered from board initialization code.
1876 i8042.debug [HW] Toggle i8042 debug mode
1877 i8042.unmask_kbd_data
1878 [HW] Enable printing of interrupt data from the KBD port
1879 (disabled by default, and as a pre-condition
1880 requires that i8042.debug=1 be enabled)
1881 i8042.direct [HW] Put keyboard port into non-translated mode
1882 i8042.dumbkbd [HW] Pretend that controller can only read data from
1883 keyboard and cannot control its state
1884 (Don't attempt to blink the leds)
1885 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1886 i8042.nokbd [HW] Don't check/create keyboard port
1887 i8042.noloop [HW] Disable the AUX Loopback command while probing
1889 i8042.nomux [HW] Don't check presence of an active multiplexing
1891 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1893 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1894 i8042.reset [HW] Reset the controller during init, cleanup and
1895 suspend-to-ram transitions, only during s2r
1896 transitions, or never reset
1897 Format: { 1 | Y | y | 0 | N | n }
1898 1, Y, y: always reset controller
1899 0, N, n: don't ever reset controller
1900 Default: only on s2r transitions on x86; most other
1901 architectures force reset to be always executed
1902 i8042.unlock [HW] Unlock (ignore) the keylock
1903 i8042.kbdreset [HW] Reset device connected to KBD port
1905 [HW] Allow deferred probing upon i8042 probe errors
1909 i915.invert_brightness=
1910 [DRM] Invert the sense of the variable that is used to
1911 set the brightness of the panel backlight. Normally a
1912 brightness value of 0 indicates backlight switched off,
1913 and the maximum of the brightness value sets the backlight
1914 to maximum brightness. If this parameter is set to 0
1915 (default) and the machine requires it, or this parameter
1916 is set to 1, a brightness value of 0 sets the backlight
1917 to maximum brightness, and the maximum of the brightness
1918 value switches the backlight off.
1919 -1 -- never invert brightness
1920 0 -- machine default
1921 1 -- force brightness inversion
1923 ia32_emulation= [X86-64]
1925 When true, allows loading 32-bit programs and executing 32-bit
1926 syscalls, essentially overriding IA32_EMULATION_DEFAULT_DISABLED at
1927 boot time. When false, unconditionally disables IA32 emulation.
1930 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1934 Format: idle=poll, idle=halt, idle=nomwait
1935 Poll forces a polling idle loop that can slightly
1936 improve the performance of waking up a idle CPU, but
1937 will use a lot of power and make the system run hot.
1939 idle=halt: Halt is forced to be used for CPU idle.
1940 In such case C2/C3 won't be used again.
1941 idle=nomwait: Disable mwait for CPU C-states
1945 Allow force disabling of Shared Virtual Memory (SVA)
1946 support for the idxd driver. By default it is set to
1949 idxd.tc_override= [HW]
1951 Allow override of default traffic class configuration
1952 for the device. By default it is set to false (0).
1954 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1955 Format: { strict | legacy | 2008 | relaxed }
1958 Choose which programs will be accepted for execution
1959 based on the IEEE 754 NaN encoding(s) supported by
1960 the FPU and the NaN encoding requested with the value
1961 of an ELF file header flag individually set by each
1962 binary. Hardware implementations are permitted to
1963 support either or both of the legacy and the 2008 NaN
1966 Available settings are as follows:
1967 strict accept binaries that request a NaN encoding
1968 supported by the FPU
1969 legacy only accept legacy-NaN binaries, if supported
1971 2008 only accept 2008-NaN binaries, if supported
1973 relaxed accept any binaries regardless of whether
1974 supported by the FPU
1976 The FPU emulator is always able to support both NaN
1977 encodings, so if no FPU hardware is present or it has
1978 been disabled with 'nofpu', then the settings of
1979 'legacy' and '2008' strap the emulator accordingly,
1980 'relaxed' straps the emulator for both legacy-NaN and
1981 2008-NaN, whereas 'strict' enables legacy-NaN only on
1982 legacy processors and both NaN encodings on MIPS32 or
1985 The setting for ABS.fmt/NEG.fmt instruction execution
1986 mode generally follows that for the NaN encoding,
1987 except where unsupported by hardware.
1989 ignore_loglevel [KNL,EARLY]
1990 Ignore loglevel setting - this will print /all/
1991 kernel messages to the console. Useful for debugging.
1992 We also add it as printk module parameter, so users
1993 could change it dynamically, usually by
1994 /sys/module/printk/parameters/ignore_loglevel.
1997 Ignore RLIMIT_DATA setting for data mappings,
1998 print warning at first misuse. Can be changed via
1999 /sys/module/kernel/parameters/ignore_rlimit_data.
2001 ihash_entries= [KNL]
2002 Set number of hash buckets for inode cache.
2004 ima_appraise= [IMA] appraise integrity measurements
2005 Format: { "off" | "enforce" | "fix" | "log" }
2008 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
2009 The builtin appraise policy appraises all files
2012 ima_canonical_fmt [IMA]
2013 Use the canonical format for the binary runtime
2014 measurements, instead of host native format.
2017 Format: { md5 | sha1 | rmd160 | sha256 | sha384
2021 The list of supported hash algorithms is defined
2022 in crypto/hash_info.h.
2025 The builtin policies to load during IMA setup.
2026 Format: "tcb | appraise_tcb | secure_boot |
2027 fail_securely | critical_data"
2029 The "tcb" policy measures all programs exec'd, files
2030 mmap'd for exec, and all files opened with the read
2031 mode bit set by either the effective uid (euid=0) or
2034 The "appraise_tcb" policy appraises the integrity of
2035 all files owned by root.
2037 The "secure_boot" policy appraises the integrity
2038 of files (eg. kexec kernel image, kernel modules,
2039 firmware, policy, etc) based on file signatures.
2041 The "fail_securely" policy forces file signature
2042 verification failure also on privileged mounted
2043 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
2046 The "critical_data" policy measures kernel integrity
2049 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
2050 Load a policy which meets the needs of the Trusted
2051 Computing Base. This means IMA will measure all
2052 programs exec'd, files mmap'd for exec, and all files
2053 opened for read by uid=0.
2056 Select one of defined IMA measurements template formats.
2057 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
2062 [IMA] Define a custom template format.
2063 Format: { "field1|...|fieldN" }
2065 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
2066 Format: <min_file_size>
2067 Set the minimal file size for using asynchronous hash.
2068 If left unspecified, ahash usage is disabled.
2070 ahash performance varies for different data sizes on
2071 different crypto accelerators. This option can be used
2072 to achieve the best performance for a particular HW.
2074 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
2076 Set hashing buffer size. Default: 4k.
2078 ahash performance varies for different chunk sizes on
2079 different crypto accelerators. This option can be used
2080 to achieve best performance for particular HW.
2084 Run specified binary instead of /sbin/init as init
2087 initcall_debug [KNL] Trace initcalls as they are executed. Useful
2088 for working out where the kernel is dying during
2091 initcall_blacklist= [KNL] Do not execute a comma-separated list of
2092 initcall functions. Useful for debugging built-in
2093 modules and initcalls.
2095 initramfs_async= [KNL]
2098 This parameter controls whether the initramfs
2099 image is unpacked asynchronously, concurrently
2100 with devices being probed and
2101 initialized. This should normally just work,
2102 but as a debugging aid, one can get the
2103 historical behaviour of the initramfs
2104 unpacking being completed before device_ and
2107 initrd= [BOOT,EARLY] Specify the location of the initial ramdisk
2109 initrdmem= [KNL,EARLY] Specify a physical address and size from which to
2110 load the initrd. If an initrd is compiled in or
2111 specified in the bootparams, it takes priority over this
2113 Format: ss[KMG],nn[KMG]
2116 init_on_alloc= [MM,EARLY] Fill newly allocated pages and heap objects with
2119 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2121 init_on_free= [MM,EARLY] Fill freed pages and heap objects with zeroes.
2123 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2125 init_pkru= [X86] Specify the default memory protection keys rights
2126 register contents for all processes. 0x55555554 by
2127 default (disallow access to all but pkey 0). Can
2128 override in debugfs after boot.
2130 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2133 int_pln_enable [X86] Enable power limit notification interrupt
2135 integrity_audit=[IMA]
2136 Format: { "0" | "1" }
2137 0 -- basic integrity auditing messages. (Default)
2138 1 -- additional integrity auditing messages.
2140 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2142 Enable intel iommu driver.
2144 Disable intel iommu driver.
2145 igfx_off [Default Off]
2146 By default, gfx is mapped as normal device. If a gfx
2147 device has a dedicated DMAR unit, the DMAR unit is
2148 bypassed by not enabling DMAR with this option. In
2149 this case, gfx device will use physical address for
2151 strict [Default Off]
2152 Deprecated, equivalent to iommu.strict=1.
2153 sp_off [Default Off]
2154 By default, super page will be supported if Intel IOMMU
2155 has the capability. With this option, super page will
2158 Enable the Intel IOMMU scalable mode if the hardware
2159 advertises that it has support for the scalable mode
2162 Disallow use of the Intel IOMMU scalable mode.
2163 tboot_noforce [Default Off]
2164 Do not force the Intel IOMMU enabled under tboot.
2165 By default, tboot will force Intel IOMMU on, which
2166 could harm performance of some high-throughput
2167 devices like 40GBit network cards, even if identity
2169 Note that using this option lowers the security
2170 provided by tboot because it makes the system
2171 vulnerable to DMA attacks.
2173 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2174 0 disables intel_idle and fall back on acpi_idle.
2175 1 to 9 specify maximum depth of C-state.
2177 intel_pstate= [X86,EARLY]
2179 Do not enable intel_pstate as the default
2180 scaling driver for the supported processors
2182 Use intel_pstate driver to bypass the scaling
2183 governors layer of cpufreq and provides it own
2184 algorithms for p-state selection. There are two
2185 P-state selection algorithms provided by
2186 intel_pstate in the active mode: powersave and
2187 performance. The way they both operate depends
2188 on whether or not the hardware managed P-states
2189 (HWP) feature has been enabled in the processor
2190 and possibly on the processor model.
2192 Use intel_pstate as a scaling driver, but configure it
2193 to work with generic cpufreq governors (instead of
2194 enabling its internal governor). This mode cannot be
2195 used along with the hardware-managed P-states (HWP)
2198 Enable intel_pstate on systems that prohibit it by default
2199 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2200 instead of acpi-cpufreq may disable platform features, such
2201 as thermal controls and power capping, that rely on ACPI
2202 P-States information being indicated to OSPM and therefore
2203 should be used with caution. This option does not work with
2204 processors that aren't supported by the intel_pstate driver
2205 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2207 Do not enable hardware P state control (HWP)
2210 Only load intel_pstate on systems which support
2211 hardware P state control (HWP) if available.
2213 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2214 Description Table, specifies preferred power management
2215 profile as "Enterprise Server" or "Performance Server",
2216 then this feature is turned on by default.
2218 Allow per-logical-CPU P-State performance control limits using
2219 cpufreq sysfs interface
2221 intremap= [X86-64,Intel-IOMMU,EARLY]
2222 on enable Interrupt Remapping (default)
2223 off disable Interrupt Remapping
2224 nosid disable Source ID checking
2226 BIOS x2APIC opt-out request will be ignored
2227 nopost disable Interrupt Posting
2229 iomem= Disable strict checking of access to MMIO memory
2230 strict regions from userspace.
2245 nobypass [PPC/POWERNV]
2246 Disable IOMMU bypass, using IOMMU for PCI devices.
2248 iommu.forcedac= [ARM64,X86,EARLY] Control IOVA allocation for PCI devices.
2249 Format: { "0" | "1" }
2250 0 - Try to allocate a 32-bit DMA address first, before
2251 falling back to the full range if needed.
2252 1 - Allocate directly from the full usable range,
2253 forcing Dual Address Cycle for PCI cards supporting
2254 greater than 32-bit addressing.
2256 iommu.strict= [ARM64,X86,S390,EARLY] Configure TLB invalidation behaviour
2257 Format: { "0" | "1" }
2259 Request that DMA unmap operations use deferred
2260 invalidation of hardware TLBs, for increased
2261 throughput at the cost of reduced device isolation.
2262 Will fall back to strict mode if not supported by
2263 the relevant IOMMU driver.
2265 DMA unmap operations invalidate IOMMU hardware TLBs
2267 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2268 Note: on x86, strict mode specified via one of the
2269 legacy driver-specific options takes precedence.
2272 [ARM64,X86,EARLY] Configure DMA to bypass the IOMMU by default.
2273 Format: { "0" | "1" }
2274 0 - Use IOMMU translation for DMA.
2275 1 - Bypass the IOMMU for DMA.
2276 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2278 io7= [HW] IO7 for Marvel-based Alpha systems
2279 See comment before marvel_specify_io7 in
2280 arch/alpha/kernel/core_marvel.c.
2282 io_delay= [X86,EARLY] I/O delay method
2284 Standard port 0x80 based delay
2286 Alternate port 0xed based delay (needed on some systems)
2288 Simple two microseconds delay
2293 See Documentation/admin-guide/nfs/nfsroot.rst.
2295 ipcmni_extend [KNL,EARLY] Extend the maximum number of unique System V
2296 IPC identifiers from 32,768 to 16,777,216.
2298 irqaffinity= [SMP] Set the default irq affinity mask
2299 The argument is a cpu list, as described above.
2301 irqchip.gicv2_force_probe=
2304 Force the kernel to look for the second 4kB page
2305 of a GICv2 controller even if the memory range
2306 exposed by the device tree is too small.
2308 irqchip.gicv3_nolpi=
2310 Force the kernel to ignore the availability of
2311 LPIs (and by consequence ITSs). Intended for system
2312 that use the kernel as a bootloader, and thus want
2313 to let secondary kernels in charge of setting up
2316 irqchip.gicv3_pseudo_nmi= [ARM64,EARLY]
2317 Enables support for pseudo-NMIs in the kernel. This
2318 requires the kernel to be built with
2319 CONFIG_ARM64_PSEUDO_NMI.
2322 When an interrupt is not handled search all handlers
2323 for it. Intended to get systems with badly broken
2327 When an interrupt is not handled search all handlers
2328 for it. Also check all handlers each timer
2329 interrupt. Intended to get systems with badly broken
2333 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2335 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2336 [Deprecated - use cpusets instead]
2337 Format: [flag-list,]<cpu-list>
2339 Specify one or more CPUs to isolate from disturbances
2340 specified in the flag list (default: domain):
2343 Disable the tick when a single task runs.
2345 A residual 1Hz tick is offloaded to workqueues, which you
2346 need to affine to housekeeping through the global
2347 workqueue's affinity configured via the
2348 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2349 by using the 'domain' flag described below.
2351 NOTE: by default the global workqueue runs on all CPUs,
2352 so to protect individual CPUs the 'cpumask' file has to
2353 be configured manually after bootup.
2356 Isolate from the general SMP balancing and scheduling
2357 algorithms. Note that performing domain isolation this way
2358 is irreversible: it's not possible to bring back a CPU to
2359 the domains once isolated through isolcpus. It's strongly
2360 advised to use cpusets instead to disable scheduler load
2361 balancing through the "cpuset.sched_load_balance" file.
2362 It offers a much more flexible interface where CPUs can
2363 move in and out of an isolated set anytime.
2365 You can move a process onto or off an "isolated" CPU via
2366 the CPU affinity syscalls or cpuset.
2367 <cpu number> begins at 0 and the maximum value is
2368 "number of CPUs in system - 1".
2372 Isolate from being targeted by managed interrupts
2373 which have an interrupt mask containing isolated
2374 CPUs. The affinity of managed interrupts is
2375 handled by the kernel and cannot be changed via
2376 the /proc/irq/* interfaces.
2378 This isolation is best effort and only effective
2379 if the automatically assigned interrupt mask of a
2380 device queue contains isolated and housekeeping
2381 CPUs. If housekeeping CPUs are online then such
2382 interrupts are directed to the housekeeping CPU
2383 so that IO submitted on the housekeeping CPU
2384 cannot disturb the isolated CPU.
2386 If a queue's affinity mask contains only isolated
2387 CPUs then this parameter has no effect on the
2388 interrupt routing decision, though interrupts are
2389 only delivered when tasks running on those
2390 isolated CPUs submit IO. IO submitted on
2391 housekeeping CPUs has no influence on those
2394 The format of <cpu-list> is described above.
2398 ivrs_ioapic [HW,X86-64]
2399 Provide an override to the IOAPIC-ID<->DEVICE-ID
2400 mapping provided in the IVRS ACPI table.
2401 By default, PCI segment is 0, and can be omitted.
2403 For example, to map IOAPIC-ID decimal 10 to
2404 PCI segment 0x1 and PCI device 00:14.0,
2405 write the parameter as:
2406 ivrs_ioapic=10@0001:00:14.0
2409 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2410 write the parameter as:
2411 ivrs_ioapic[10]=00:14.0
2412 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2413 PCI device 00:14.0 write the parameter as:
2414 ivrs_ioapic[10]=0001:00:14.0
2416 ivrs_hpet [HW,X86-64]
2417 Provide an override to the HPET-ID<->DEVICE-ID
2418 mapping provided in the IVRS ACPI table.
2419 By default, PCI segment is 0, and can be omitted.
2421 For example, to map HPET-ID decimal 10 to
2422 PCI segment 0x1 and PCI device 00:14.0,
2423 write the parameter as:
2424 ivrs_hpet=10@0001:00:14.0
2427 * To map HPET-ID decimal 0 to PCI device 00:14.0
2428 write the parameter as:
2429 ivrs_hpet[0]=00:14.0
2430 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2431 PCI device 00:14.0 write the parameter as:
2432 ivrs_ioapic[10]=0001:00:14.0
2434 ivrs_acpihid [HW,X86-64]
2435 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2436 mapping provided in the IVRS ACPI table.
2437 By default, PCI segment is 0, and can be omitted.
2439 For example, to map UART-HID:UID AMD0020:0 to
2440 PCI segment 0x1 and PCI device ID 00:14.5,
2441 write the parameter as:
2442 ivrs_acpihid=AMD0020:0@0001:00:14.5
2445 * To map UART-HID:UID AMD0020:0 to PCI segment is 0,
2446 PCI device ID 00:14.5, write the parameter as:
2447 ivrs_acpihid[00:14.5]=AMD0020:0
2448 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and
2449 PCI device ID 00:14.5, write the parameter as:
2450 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2452 js= [HW,JOY] Analog joystick
2453 See Documentation/input/joydev/joystick.rst.
2456 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2457 report on every invalid memory access. Without this
2458 parameter KASAN will print report only for the first
2461 keep_bootcon [KNL,EARLY]
2462 Do not unregister boot console at start. This is only
2463 useful for debugging when something happens in the window
2464 between unregistering the boot console and initializing
2467 keepinitrd [HW,ARM] See retain_initrd.
2469 kernelcore= [KNL,X86,IA-64,PPC,EARLY]
2470 Format: nn[KMGTPE] | nn% | "mirror"
2471 This parameter specifies the amount of memory usable by
2472 the kernel for non-movable allocations. The requested
2473 amount is spread evenly throughout all nodes in the
2474 system as ZONE_NORMAL. The remaining memory is used for
2475 movable memory in its own zone, ZONE_MOVABLE. In the
2476 event, a node is too small to have both ZONE_NORMAL and
2477 ZONE_MOVABLE, kernelcore memory will take priority and
2478 other nodes will have a larger ZONE_MOVABLE.
2480 ZONE_MOVABLE is used for the allocation of pages that
2481 may be reclaimed or moved by the page migration
2482 subsystem. Note that allocations like PTEs-from-HighMem
2483 still use the HighMem zone if it exists, and the Normal
2484 zone if it does not.
2486 It is possible to specify the exact amount of memory in
2487 the form of "nn[KMGTPE]", a percentage of total system
2488 memory in the form of "nn%", or "mirror". If "mirror"
2489 option is specified, mirrored (reliable) memory is used
2490 for non-movable allocations and remaining memory is used
2491 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2492 are exclusive, so you cannot specify multiple forms.
2494 kgdbdbgp= [KGDB,HW,EARLY] kgdb over EHCI usb debug port.
2495 Format: <Controller#>[,poll interval]
2496 The controller # is the number of the ehci usb debug
2497 port as it is probed via PCI. The poll interval is
2498 optional and is the number seconds in between
2499 each poll cycle to the debug port in case you need
2500 the functionality for interrupting the kernel with
2501 gdb or control-c on the dbgp connection. When
2502 not using this parameter you use sysrq-g to break into
2503 the kernel debugger.
2505 kgdboc= [KGDB,HW] kgdb over consoles.
2506 Requires a tty driver that supports console polling,
2507 or a supported polling keyboard driver (non-usb).
2508 Serial only format: <serial_device>[,baud]
2509 keyboard only format: kbd
2510 keyboard and serial format: kbd,<serial_device>[,baud]
2511 Optional Kernel mode setting:
2512 kms, kbd format: kms,kbd
2513 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2515 kgdboc_earlycon= [KGDB,HW,EARLY]
2516 If the boot console provides the ability to read
2517 characters and can work in polling mode, you can use
2518 this parameter to tell kgdb to use it as a backend
2519 until the normal console is registered. Intended to
2520 be used together with the kgdboc parameter which
2521 specifies the normal console to transition to.
2523 The name of the early console should be specified
2524 as the value of this parameter. Note that the name of
2525 the early console might be different than the tty
2526 name passed to kgdboc. It's OK to leave the value
2527 blank and the first boot console that implements
2528 read() will be picked.
2530 kgdbwait [KGDB,EARLY] Stop kernel execution and enter the
2531 kernel debugger at the earliest opportunity.
2533 kmac= [MIPS] Korina ethernet MAC address.
2534 Configure the RouterBoard 532 series on-chip
2535 Ethernet adapter MAC address.
2537 kmemleak= [KNL,EARLY] Boot-time kmemleak enable/disable
2538 Valid arguments: on, off
2540 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2543 kprobe_event=[probe-list]
2544 [FTRACE] Add kprobe events and enable at boot time.
2545 The probe-list is a semicolon delimited list of probe
2546 definitions. Each definition is same as kprobe_events
2547 interface, but the parameters are comma delimited.
2548 For example, to add a kprobe event on vfs_read with
2549 arg1 and arg2, add to the command line;
2551 kprobe_event=p,vfs_read,$arg1,$arg2
2553 See also Documentation/trace/kprobetrace.rst "Kernel
2554 Boot Parameter" section.
2556 kpti= [ARM64,EARLY] Control page table isolation of
2557 user and kernel address spaces.
2558 Default: enabled on cores which need mitigation.
2562 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires
2563 CONFIG_KUNIT to be set to be fully enabled. The
2564 default value can be overridden via
2565 KUNIT_DEFAULT_ENABLED.
2566 Default is 1 (enabled)
2568 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2569 Default is 0 (don't ignore, but inject #GP)
2571 kvm.eager_page_split=
2572 [KVM,X86] Controls whether or not KVM will try to
2573 proactively split all huge pages during dirty logging.
2574 Eager page splitting reduces interruptions to vCPU
2575 execution by eliminating the write-protection faults
2576 and MMU lock contention that would otherwise be
2577 required to split huge pages lazily.
2579 VM workloads that rarely perform writes or that write
2580 only to a small region of VM memory may benefit from
2581 disabling eager page splitting to allow huge pages to
2582 still be used for reads.
2584 The behavior of eager page splitting depends on whether
2585 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2586 disabled, all huge pages in a memslot will be eagerly
2587 split when dirty logging is enabled on that memslot. If
2588 enabled, eager page splitting will be performed during
2589 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2592 Eager page splitting is only supported when kvm.tdp_mmu=Y.
2596 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2597 Default is false (don't support).
2600 [KVM] Controls the software workaround for the
2601 X86_BUG_ITLB_MULTIHIT bug.
2602 force : Always deploy workaround.
2603 off : Never deploy workaround.
2604 auto : Deploy workaround based on the presence of
2605 X86_BUG_ITLB_MULTIHIT.
2609 If the software workaround is enabled for the host,
2610 guests do need not to enable it for nested guests.
2612 kvm.nx_huge_pages_recovery_ratio=
2613 [KVM] Controls how many 4KiB pages are periodically zapped
2614 back to huge pages. 0 disables the recovery, otherwise if
2615 the value is N KVM will zap 1/Nth of the 4KiB pages every
2616 period (see below). The default is 60.
2618 kvm.nx_huge_pages_recovery_period_ms=
2619 [KVM] Controls the time period at which KVM zaps 4KiB pages
2620 back to huge pages. If the value is a non-zero N, KVM will
2621 zap a portion (see ratio above) of the pages every N msecs.
2622 If the value is 0 (the default), KVM will pick a period based
2623 on the ratio, such that a page is zapped after 1 hour on average.
2625 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in
2626 KVM/SVM. Default is 1 (enabled).
2628 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables,
2629 a.k.a. Two-Dimensional Page Tables. Default is 1
2630 (enabled). Disable by KVM if hardware lacks support
2634 [KVM,ARM,EARLY] Select one of KVM/arm64's modes of
2637 none: Forcefully disable KVM.
2639 nvhe: Standard nVHE-based mode, without support for
2642 protected: nVHE-based mode with support for guests whose
2643 state is kept private from the host.
2645 nested: VHE-based mode with support for nested
2646 virtualization. Requires at least ARMv8.3
2649 Defaults to VHE/nVHE based on hardware support. Setting
2650 mode to "protected" will disable kexec and hibernation
2651 for the host. "nested" is experimental and should be
2652 used with extreme caution.
2654 kvm-arm.vgic_v3_group0_trap=
2655 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-0
2658 kvm-arm.vgic_v3_group1_trap=
2659 [KVM,ARM,EARLY] Trap guest accesses to GICv3 group-1
2662 kvm-arm.vgic_v3_common_trap=
2663 [KVM,ARM,EARLY] Trap guest accesses to GICv3 common
2666 kvm-arm.vgic_v4_enable=
2667 [KVM,ARM,EARLY] Allow use of GICv4 for direct
2670 kvm_cma_resv_ratio=n [PPC,EARLY]
2671 Reserves given percentage from system memory area for
2672 contiguous memory allocation for KVM hash pagetable
2674 By default it reserves 5% of total system memory.
2678 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables,
2679 a.k.a. Two-Dimensional Page Tables. Default is 1
2680 (enabled). Disable by KVM if hardware lacks support
2683 kvm-intel.emulate_invalid_guest_state=
2684 [KVM,Intel] Control whether to emulate invalid guest
2685 state. Ignored if kvm-intel.enable_unrestricted_guest=1,
2686 as guest state is never invalid for unrestricted
2687 guests. This param doesn't apply to nested guests (L2),
2688 as KVM never emulates invalid L2 guest state.
2689 Default is 1 (enabled).
2691 kvm-intel.flexpriority=
2692 [KVM,Intel] Control KVM's use of FlexPriority feature
2693 (TPR shadow). Default is 1 (enabled). Disable by KVM if
2694 hardware lacks support for it.
2697 [KVM,Intel] Control nested virtualization feature in
2698 KVM/VMX. Default is 1 (enabled).
2700 kvm-intel.unrestricted_guest=
2701 [KVM,Intel] Control KVM's use of unrestricted guest
2702 feature (virtualized real and unpaged mode). Default
2703 is 1 (enabled). Disable by KVM if EPT is disabled or
2704 hardware lacks support for it.
2706 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2709 Valid arguments: never, cond, always
2711 always: L1D cache flush on every VMENTER.
2712 cond: Flush L1D on VMENTER only when the code between
2713 VMEXIT and VMENTER can leak host memory.
2714 never: Disables the mitigation
2716 Default is cond (do L1 cache flush in specific instances)
2718 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor
2719 Identification feature (tagged TLBs). Default is 1
2720 (enabled). Disable by KVM if hardware lacks support
2723 l1d_flush= [X86,INTEL,EARLY]
2724 Control mitigation for L1D based snooping vulnerability.
2726 Certain CPUs are vulnerable to an exploit against CPU
2727 internal buffers which can forward information to a
2728 disclosure gadget under certain conditions.
2730 In vulnerable processors, the speculatively
2731 forwarded data can be used in a cache side channel
2732 attack, to access data to which the attacker does
2733 not have direct access.
2735 This parameter controls the mitigation. The
2738 on - enable the interface for the mitigation
2740 l1tf= [X86,EARLY] Control mitigation of the L1TF vulnerability on
2743 The kernel PTE inversion protection is unconditionally
2744 enabled and cannot be disabled.
2747 Provides all available mitigations for the
2748 L1TF vulnerability. Disables SMT and
2749 enables all mitigations in the
2750 hypervisors, i.e. unconditional L1D flush.
2752 SMT control and L1D flush control via the
2753 sysfs interface is still possible after
2754 boot. Hypervisors will issue a warning
2755 when the first VM is started in a
2756 potentially insecure configuration,
2757 i.e. SMT enabled or L1D flush disabled.
2760 Same as 'full', but disables SMT and L1D
2761 flush runtime control. Implies the
2762 'nosmt=force' command line option.
2763 (i.e. sysfs control of SMT is disabled.)
2766 Leaves SMT enabled and enables the default
2767 hypervisor mitigation, i.e. conditional
2770 SMT control and L1D flush control via the
2771 sysfs interface is still possible after
2772 boot. Hypervisors will issue a warning
2773 when the first VM is started in a
2774 potentially insecure configuration,
2775 i.e. SMT enabled or L1D flush disabled.
2779 Disables SMT and enables the default
2780 hypervisor mitigation.
2782 SMT control and L1D flush control via the
2783 sysfs interface is still possible after
2784 boot. Hypervisors will issue a warning
2785 when the first VM is started in a
2786 potentially insecure configuration,
2787 i.e. SMT enabled or L1D flush disabled.
2790 Same as 'flush', but hypervisors will not
2791 warn when a VM is started in a potentially
2792 insecure configuration.
2795 Disables hypervisor mitigations and doesn't
2797 It also drops the swap size and available
2798 RAM limit restriction on both hypervisor and
2803 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2809 lapic [X86-32,APIC,EARLY] Enable the local APIC even if BIOS
2812 lapic= [X86,APIC] Do not use TSC deadline
2813 value for LAPIC timer one-shot implementation. Default
2814 back to the programmable timer unit in the LAPIC.
2815 Format: notscdeadline
2817 lapic_timer_c2_ok [X86,APIC,EARLY] trust the local apic timer
2820 libata.dma= [LIBATA] DMA control
2821 libata.dma=0 Disable all PATA and SATA DMA
2822 libata.dma=1 PATA and SATA Disk DMA only
2823 libata.dma=2 ATAPI (CDROM) DMA only
2824 libata.dma=4 Compact Flash DMA only
2825 Combinations also work, so libata.dma=3 enables DMA
2826 for disks and CDROMs, but not CFs.
2828 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2829 libata.ignore_hpa=0 keep BIOS limits (default)
2830 libata.ignore_hpa=1 ignore limits, using full disk
2832 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2836 libata.force= [LIBATA] Force configurations. The format is a comma-
2837 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2838 PORT and DEVICE are decimal numbers matching port, link
2839 or device. Basically, it matches the ATA ID string
2840 printed on console by libata. If the whole ID part is
2841 omitted, the last PORT and DEVICE values are used. If
2842 ID hasn't been specified yet, the configuration applies
2843 to all ports, links and devices.
2845 If only DEVICE is omitted, the parameter applies to
2846 the port and all links and devices behind it. DEVICE
2847 number of 0 either selects the first device or the
2848 first fan-out link behind PMP device. It does not
2849 select the host link. DEVICE number of 15 selects the
2850 host link and device attached to it.
2852 The VAL specifies the configuration to force. As long
2853 as there is no ambiguity, shortcut notation is allowed.
2854 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2855 The following configurations can be forced.
2857 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2858 Any ID with matching PORT is used.
2860 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2862 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2863 udma[/][16,25,33,44,66,100,133] notation is also
2866 * nohrst, nosrst, norst: suppress hard, soft and both
2869 * rstonce: only attempt one reset during hot-unplug
2872 * [no]dbdelay: Enable or disable the extra 200ms delay
2873 before debouncing a link PHY and device presence
2876 * [no]ncq: Turn on or off NCQ.
2878 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2880 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2882 * [no]trim: Enable or disable (unqueued) TRIM.
2884 * trim_zero: Indicate that TRIM command zeroes data.
2886 * max_trim_128m: Set 128M maximum trim size limit.
2888 * [no]dma: Turn on or off DMA transfers.
2890 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2892 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2893 commands that are not a multiple of 16 bytes.
2895 * [no]dmalog: Enable or disable the use of the
2896 READ LOG DMA EXT command to access logs.
2898 * [no]iddevlog: Enable or disable access to the
2899 identify device data log.
2901 * [no]logdir: Enable or disable access to the general
2902 purpose log directory.
2904 * max_sec_128: Set transfer size limit to 128 sectors.
2906 * max_sec_1024: Set or clear transfer size limit to
2909 * max_sec_lba48: Set or clear transfer size limit to
2912 * [no]lpm: Enable or disable link power management.
2914 * [no]setxfer: Indicate if transfer speed mode setting
2917 * [no]fua: Disable or enable FUA (Force Unit Access)
2918 support for devices supporting this feature.
2920 * dump_id: Dump IDENTIFY data.
2922 * disable: Disable this device.
2924 If there are multiple matching configurations changing
2925 the same attribute, the last one is used.
2927 load_ramdisk= [RAM] [Deprecated]
2929 lockd.nlm_grace_period=P [NFS] Assign grace period.
2932 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2935 lockd.nlm_timeout=T [NFS] Assign timeout value.
2938 lockd.nlm_udpport=M [NFS] Assign UDP port.
2941 lockdown= [SECURITY,EARLY]
2942 { integrity | confidentiality }
2943 Enable the kernel lockdown feature. If set to
2944 integrity, kernel features that allow userland to
2945 modify the running kernel are disabled. If set to
2946 confidentiality, kernel features that allow userland
2947 to extract confidential information from the kernel
2950 locktorture.acq_writer_lim= [KNL]
2951 Set the time limit in jiffies for a lock
2952 acquisition. Acquisitions exceeding this limit
2953 will result in a splat once they do complete.
2955 locktorture.bind_readers= [KNL]
2956 Specify the list of CPUs to which the readers are
2959 locktorture.bind_writers= [KNL]
2960 Specify the list of CPUs to which the writers are
2963 locktorture.call_rcu_chains= [KNL]
2964 Specify the number of self-propagating call_rcu()
2965 chains to set up. These are used to ensure that
2966 there is a high probability of an RCU grace period
2967 in progress at any given time. Defaults to 0,
2968 which disables these call_rcu() chains.
2970 locktorture.long_hold= [KNL]
2971 Specify the duration in milliseconds for the
2972 occasional long-duration lock hold time. Defaults
2973 to 100 milliseconds. Select 0 to disable.
2975 locktorture.nested_locks= [KNL]
2976 Specify the maximum lock nesting depth that
2977 locktorture is to exercise, up to a limit of 8
2978 (MAX_NESTED_LOCKS). Specify zero to disable.
2979 Note that this parameter is ineffective on types
2980 of locks that do not support nested acquisition.
2982 locktorture.nreaders_stress= [KNL]
2983 Set the number of locking read-acquisition kthreads.
2984 Defaults to being automatically set based on the
2985 number of online CPUs.
2987 locktorture.nwriters_stress= [KNL]
2988 Set the number of locking write-acquisition kthreads.
2990 locktorture.onoff_holdoff= [KNL]
2991 Set time (s) after boot for CPU-hotplug testing.
2993 locktorture.onoff_interval= [KNL]
2994 Set time (s) between CPU-hotplug operations, or
2995 zero to disable CPU-hotplug testing.
2997 locktorture.rt_boost= [KNL]
2998 Do periodic testing of real-time lock priority
2999 boosting. Select 0 to disable, 1 to boost
3000 only rt_mutex, and 2 to boost unconditionally.
3001 Defaults to 2, which might seem to be an
3002 odd choice, but which should be harmless for
3003 non-real-time spinlocks, due to their disabling
3004 of preemption. Note that non-realtime mutexes
3007 locktorture.rt_boost_factor= [KNL]
3008 Number that determines how often and for how
3009 long priority boosting is exercised. This is
3010 scaled down by the number of writers, so that the
3011 number of boosts per unit time remains roughly
3012 constant as the number of writers increases.
3013 On the other hand, the duration of each boost
3014 increases with the number of writers.
3016 locktorture.shuffle_interval= [KNL]
3017 Set task-shuffle interval (jiffies). Shuffling
3018 tasks allows some CPUs to go into dyntick-idle
3019 mode during the locktorture test.
3021 locktorture.shutdown_secs= [KNL]
3022 Set time (s) after boot system shutdown. This
3023 is useful for hands-off automated testing.
3025 locktorture.stat_interval= [KNL]
3026 Time (s) between statistics printk()s.
3028 locktorture.stutter= [KNL]
3029 Time (s) to stutter testing, for example,
3030 specifying five seconds causes the test to run for
3031 five seconds, wait for five seconds, and so on.
3032 This tests the locking primitive's ability to
3033 transition abruptly to and from idle.
3035 locktorture.torture_type= [KNL]
3036 Specify the locking implementation to test.
3038 locktorture.verbose= [KNL]
3039 Enable additional printk() statements.
3041 locktorture.writer_fifo= [KNL]
3042 Run the write-side locktorture kthreads at
3043 sched_set_fifo() real-time priority.
3045 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
3048 loglevel= [KNL,EARLY]
3049 All Kernel Messages with a loglevel smaller than the
3050 console loglevel will be printed to the console. It can
3051 also be changed with klogd or other programs. The
3052 loglevels are defined as follows:
3054 0 (KERN_EMERG) system is unusable
3055 1 (KERN_ALERT) action must be taken immediately
3056 2 (KERN_CRIT) critical conditions
3057 3 (KERN_ERR) error conditions
3058 4 (KERN_WARNING) warning conditions
3059 5 (KERN_NOTICE) normal but significant condition
3060 6 (KERN_INFO) informational
3061 7 (KERN_DEBUG) debug-level messages
3063 log_buf_len=n[KMG] [KNL,EARLY]
3064 Sets the size of the printk ring buffer, in bytes.
3065 n must be a power of two and greater than the
3066 minimal size. The minimal size is defined by
3067 LOG_BUF_SHIFT kernel config parameter. There
3068 is also CONFIG_LOG_CPU_MAX_BUF_SHIFT config
3069 parameter that allows to increase the default size
3070 depending on the number of CPUs. See init/Kconfig
3073 logo.nologo [FB] Disables display of the built-in Linux logo.
3074 This may be used to provide more screen space for
3075 kernel log messages and is useful when debugging
3076 kernel boot problems.
3078 lp=0 [LP] Specify parallel ports to use, e.g,
3079 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
3080 lp=reset first parallel port). 'lp=0' disables the
3081 lp=auto printer driver. 'lp=reset' (which can be
3082 specified in addition to the ports) causes
3083 attached printers to be reset. Using
3084 lp=port1,port2,... specifies the parallel ports
3085 to associate lp devices with, starting with
3086 lp0. A port specification may be 'none' to skip
3087 that lp device, or a parport name such as
3088 'parport0'. Specifying 'lp=auto' instead of a
3089 port specification list means that device IDs
3090 from each port should be examined, to see if
3091 an IEEE 1284-compliant printer is attached; if
3092 so, the driver will manage that printer.
3093 See also header of drivers/char/lp.c.
3096 Sets loops_per_jiffy to given constant, thus avoiding
3097 time-consuming boot-time autodetection (up to 250 ms per
3098 CPU). 0 enables autodetection (default). To determine
3099 the correct value for your kernel, boot with normal
3100 autodetection and see what value is printed. Note that
3101 on SMP systems the preset will be applied to all CPUs,
3102 which is likely to cause problems if your CPUs need
3103 significantly divergent settings. An incorrect value
3104 will cause delays in the kernel to be wrong, leading to
3105 unpredictable I/O errors and other breakage. Although
3106 unlikely, in the extreme case this might damage your
3110 Format: <io>,<irq>,<dma>
3112 lsm.debug [SECURITY] Enable LSM initialization debugging output.
3115 [SECURITY] Choose order of LSM initialization. This
3116 overrides CONFIG_LSM, and the "security=" parameter.
3118 machvec= [IA-64] Force the use of a particular machine-vector
3119 (machvec) in a generic kernel.
3120 Example: machvec=hpzx1
3122 machtype= [Loongson] Share the same kernel image file between
3123 different yeeloong laptops.
3124 Example: machtype=lemote-yeeloong-2f-7inch
3126 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
3127 than or equal to this physical address is ignored.
3129 maxcpus= [SMP,EARLY] Maximum number of processors that an SMP kernel
3130 will bring up during bootup. maxcpus=n : n >= 0 limits
3131 the kernel to bring up 'n' processors. Surely after
3132 bootup you can bring up the other plugged cpu by executing
3133 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
3134 only takes effect during system bootup.
3135 While n=0 is a special case, it is equivalent to "nosmp",
3136 which also disables the IO APIC.
3138 max_loop= [LOOP] The number of loop block devices that get
3139 (loop.max_loop) unconditionally pre-created at init time. The default
3140 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
3141 of statically allocating a predefined number, loop
3142 devices can be requested on-demand with the
3143 /dev/loop-control interface.
3145 mce [X86-32] Machine Check Exception
3147 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
3149 md= [HW] RAID subsystems devices and level
3150 See Documentation/admin-guide/md.rst.
3153 Format: <first>,<last>
3154 Specifies range of consoles to be captured by the MDA.
3156 mds= [X86,INTEL,EARLY]
3157 Control mitigation for the Micro-architectural Data
3158 Sampling (MDS) vulnerability.
3160 Certain CPUs are vulnerable to an exploit against CPU
3161 internal buffers which can forward information to a
3162 disclosure gadget under certain conditions.
3164 In vulnerable processors, the speculatively
3165 forwarded data can be used in a cache side channel
3166 attack, to access data to which the attacker does
3167 not have direct access.
3169 This parameter controls the MDS mitigation. The
3172 full - Enable MDS mitigation on vulnerable CPUs
3173 full,nosmt - Enable MDS mitigation and disable
3174 SMT on vulnerable CPUs
3175 off - Unconditionally disable MDS mitigation
3177 On TAA-affected machines, mds=off can be prevented by
3178 an active TAA mitigation as both vulnerabilities are
3179 mitigated with the same mechanism so in order to disable
3180 this mitigation, you need to specify tsx_async_abort=off
3183 Not specifying this option is equivalent to
3186 For details see: Documentation/admin-guide/hw-vuln/mds.rst
3188 mem=nn[KMG] [HEXAGON,EARLY] Set the memory size.
3189 Must be specified, otherwise memory size will be 0.
3191 mem=nn[KMG] [KNL,BOOT,EARLY] Force usage of a specific amount
3192 of memory Amount of memory to be used in cases
3196 2 when the kernel is not able to see the whole system memory;
3197 3 memory that lies after 'mem=' boundary is excluded from
3198 the hypervisor, then assigned to KVM guests.
3199 4 to limit the memory available for kdump kernel.
3201 [ARC,MICROBLAZE] - the limit applies only to low memory,
3202 high memory is not affected.
3204 [ARM64] - only limits memory covered by the linear
3205 mapping. The NOMAP regions are not affected.
3207 [X86] Work as limiting max address. Use together
3208 with memmap= to avoid physical address space collisions.
3209 Without memmap= PCI devices could be placed at addresses
3210 belonging to unused RAM.
3212 Note that this only takes effects during boot time since
3213 in above case 3, memory may need be hot added after boot
3214 if system memory of hypervisor is not sufficient.
3217 [ARM,MIPS,EARLY] - override the memory layout
3218 reported by firmware.
3219 Define a memory region of size nn[KMG] starting at
3221 Multiple different regions can be specified with
3222 multiple mem= parameters on the command line.
3224 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3227 memblock=debug [KNL,EARLY] Enable memblock debug messages.
3230 [KNL,SH] Allow user to override the default size for
3231 per-device physically contiguous DMA buffers.
3233 memhp_default_state=online/offline/online_kernel/online_movable
3234 [KNL] Set the initial state for the memory hotplug
3235 onlining policy. If not specified, the default value is
3236 set according to the
3237 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3239 See Documentation/admin-guide/mm/memory-hotplug.rst.
3241 memmap=exactmap [KNL,X86,EARLY] Enable setting of an exact
3242 E820 memory map, as specified by the user.
3243 Such memmap=exactmap lines can be constructed based on
3244 BIOS output or other requirements. See the memmap=nn@ss
3247 memmap=nn[KMG]@ss[KMG]
3248 [KNL, X86,MIPS,XTENSA,EARLY] Force usage of a specific region of memory.
3249 Region of memory to be used is from ss to ss+nn.
3250 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3251 which limits max address to nn[KMG].
3252 Multiple different regions can be specified,
3255 memmap=100M@2G,100M#3G,1G!1024G
3257 memmap=nn[KMG]#ss[KMG]
3258 [KNL,ACPI,EARLY] Mark specific memory as ACPI data.
3259 Region of memory to be marked is from ss to ss+nn.
3261 memmap=nn[KMG]$ss[KMG]
3262 [KNL,ACPI,EARLY] Mark specific memory as reserved.
3263 Region of memory to be reserved is from ss to ss+nn.
3264 Example: Exclude memory from 0x18690000-0x1869ffff
3265 memmap=64K$0x18690000
3267 memmap=0x10000$0x18690000
3268 Some bootloaders may need an escape character before '$',
3269 like Grub2, otherwise '$' and the following number
3272 memmap=nn[KMG]!ss[KMG,EARLY]
3273 [KNL,X86] Mark specific memory as protected.
3274 Region of memory to be used, from ss to ss+nn.
3275 The memory region may be marked as e820 type 12 (0xc)
3276 and is NVDIMM or ADR memory.
3278 memmap=<size>%<offset>-<oldtype>+<newtype>
3279 [KNL,ACPI,EARLY] Convert memory within the specified region
3280 from <oldtype> to <newtype>. If "-<oldtype>" is left
3281 out, the whole region will be marked as <newtype>,
3282 even if previously unavailable. If "+<newtype>" is left
3283 out, matching memory will be removed. Types are
3284 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3285 3 = ACPI, 12 = PRAM.
3287 memory_corruption_check=0/1 [X86,EARLY]
3288 Some BIOSes seem to corrupt the first 64k of
3289 memory when doing things like suspend/resume.
3290 Setting this option will scan the memory
3291 looking for corruption. Enabling this will
3292 both detect corruption and prevent the kernel
3293 from using the memory being corrupted.
3294 However, its intended as a diagnostic tool; if
3295 repeatable BIOS-originated corruption always
3296 affects the same memory, you can use memmap=
3297 to prevent the kernel from using that memory.
3299 memory_corruption_check_size=size [X86,EARLY]
3300 By default it checks for corruption in the low
3301 64k, making this memory unavailable for normal
3302 use. Use this parameter to scan for
3303 corruption in more or less memory.
3305 memory_corruption_check_period=seconds [X86,EARLY]
3306 By default it checks for corruption every 60
3307 seconds. Use this parameter to check at some
3308 other rate. 0 disables periodic checking.
3310 memory_hotplug.memmap_on_memory
3311 [KNL,X86,ARM] Boolean flag to enable this feature.
3312 Format: {on | off (default)}
3313 When enabled, runtime hotplugged memory will
3314 allocate its internal metadata (struct pages,
3315 those vmemmap pages cannot be optimized even
3316 if hugetlb_free_vmemmap is enabled) from the
3317 hotadded memory which will allow to hotadd a
3318 lot of memory without requiring additional
3320 This feature is disabled by default because it
3321 has some implication on large (e.g. GB)
3322 allocations in some configurations (e.g. small
3324 The state of the flag can be read in
3325 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3326 Note that even when enabled, there are a few cases where
3327 the feature is not effective.
3329 memtest= [KNL,X86,ARM,M68K,PPC,RISCV,EARLY] Enable memtest
3331 default : 0 <disable>
3332 Specifies the number of memtest passes to be
3333 performed. Each pass selects another test
3334 pattern from a given set of patterns. Memtest
3335 fills the memory with this pattern, validates
3336 memory contents and reserves bad memory
3337 regions that are detected.
3339 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3340 Valid arguments: on, off
3342 mem_encrypt=on: Activate SME
3343 mem_encrypt=off: Do not activate SME
3345 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3346 for details on when memory encryption can be activated.
3348 mem_sleep_default= [SUSPEND] Default system suspend mode:
3349 s2idle - Suspend-To-Idle
3350 shallow - Power-On Suspend or equivalent (if supported)
3351 deep - Suspend-To-RAM or equivalent (if supported)
3352 See Documentation/admin-guide/pm/sleep-states.rst.
3354 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3355 Multi-Function General Purpose Timers on AMD Geode
3358 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3359 the BIOS has incorrectly applied a workaround. TinyBIOS
3360 version 0.98 is known to be affected, 0.99 fixes the
3361 problem by letting the user disable the workaround.
3365 microcode.force_minrev= [X86]
3367 Enable or disable the microcode minimal revision
3368 enforcement for the runtime microcode loader.
3370 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3371 physical address is ignored.
3373 mini2440= [ARM,HW,KNL]
3374 Format:[0..2][b][c][t]
3376 MINI2440 configuration specification:
3377 0 - The attached screen is the 3.5" TFT
3378 1 - The attached screen is the 7" TFT
3379 2 - The VGA Shield is attached (1024x768)
3380 Leaving out the screen size parameter will not load
3381 the TFT driver, and the framebuffer will be left
3383 b - Enable backlight. The TFT backlight pin will be
3384 linked to the kernel VESA blanking code and a GPIO
3385 LED. This parameter is not necessary when using the
3387 c - Enable the s3c camera interface.
3388 t - Reserved for enabling touchscreen support. The
3389 touchscreen support is not enabled in the mainstream
3390 kernel as of 2.6.30, a preliminary port can be found
3391 in the "bleeding edge" mini2440 support kernel at
3392 https://repo.or.cz/w/linux-2.6/mini2440.git
3395 [X86,PPC,S390,ARM64,EARLY] Control optional mitigations for
3396 CPU vulnerabilities. This is a set of curated,
3397 arch-independent options, each of which is an
3398 aggregation of existing arch-specific options.
3401 Disable all optional CPU mitigations. This
3402 improves system performance, but it may also
3403 expose users to several CPU vulnerabilities.
3404 Equivalent to: if nokaslr then kpti=0 [ARM64]
3405 gather_data_sampling=off [X86]
3406 kvm.nx_huge_pages=off [X86]
3409 mmio_stale_data=off [X86]
3410 no_entry_flush [PPC]
3411 no_uaccess_flush [PPC]
3414 nospectre_bhb [ARM64]
3415 nospectre_v1 [X86,PPC]
3416 nospectre_v2 [X86,PPC,S390,ARM64]
3417 reg_file_data_sampling=off [X86]
3419 spec_rstack_overflow=off [X86]
3420 spec_store_bypass_disable=off [X86,PPC]
3421 spectre_v2_user=off [X86]
3422 srbds=off [X86,INTEL]
3423 ssbd=force-off [ARM64]
3424 tsx_async_abort=off [X86]
3427 This does not have any effect on
3428 kvm.nx_huge_pages when
3429 kvm.nx_huge_pages=force.
3432 Mitigate all CPU vulnerabilities, but leave SMT
3433 enabled, even if it's vulnerable. This is for
3434 users who don't want to be surprised by SMT
3435 getting disabled across kernel upgrades, or who
3436 have other ways of avoiding SMT-based attacks.
3437 Equivalent to: (default behavior)
3440 Mitigate all CPU vulnerabilities, disabling SMT
3441 if needed. This is for users who always want to
3442 be fully mitigated, even if it means losing SMT.
3443 Equivalent to: l1tf=flush,nosmt [X86]
3444 mds=full,nosmt [X86]
3445 tsx_async_abort=full,nosmt [X86]
3446 mmio_stale_data=full,nosmt [X86]
3447 retbleed=auto,nosmt [X86]
3450 [KNL,EARLY] When CONFIG_DEBUG_MEMORY_INIT is set, this
3451 parameter allows control of the logging verbosity for
3452 the additional memory initialisation checks. A value
3453 of 0 disables mminit logging and a level of 4 will
3454 log everything. Information is printed at KERN_DEBUG
3455 so loglevel=8 may also need to be specified.
3458 [X86,INTEL,EARLY] Control mitigation for the Processor
3459 MMIO Stale Data vulnerabilities.
3461 Processor MMIO Stale Data is a class of
3462 vulnerabilities that may expose data after an MMIO
3463 operation. Exposed data could originate or end in
3464 the same CPU buffers as affected by MDS and TAA.
3465 Therefore, similar to MDS and TAA, the mitigation
3466 is to clear the affected CPU buffers.
3468 This parameter controls the mitigation. The
3471 full - Enable mitigation on vulnerable CPUs
3473 full,nosmt - Enable mitigation and disable SMT on
3476 off - Unconditionally disable mitigation
3478 On MDS or TAA affected machines,
3479 mmio_stale_data=off can be prevented by an active
3480 MDS or TAA mitigation as these vulnerabilities are
3481 mitigated with the same mechanism so in order to
3482 disable this mitigation, you need to specify
3483 mds=off and tsx_async_abort=off too.
3485 Not specifying this option is equivalent to
3486 mmio_stale_data=full.
3489 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3491 <module>.async_probe[=<bool>] [KNL]
3492 If no <bool> value is specified or if the value
3493 specified is not a valid <bool>, enable asynchronous
3494 probe on this module. Otherwise, enable/disable
3495 asynchronous probe on this module as indicated by the
3496 <bool> value. See also: module.async_probe
3498 module.async_probe=<bool>
3499 [KNL] When set to true, modules will use async probing
3500 by default. To enable/disable async probing for a
3501 specific module, use the module specific control that
3502 is documented under <module>.async_probe. When both
3503 module.async_probe and <module>.async_probe are
3504 specified, <module>.async_probe takes precedence for
3505 the specific module.
3507 module.enable_dups_trace
3508 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set,
3509 this means that duplicate request_module() calls will
3510 trigger a WARN_ON() instead of a pr_warn(). Note that
3511 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s
3512 will always be issued and this option does nothing.
3514 [KNL] When CONFIG_MODULE_SIG is set, this means that
3515 modules without (valid) signatures will fail to load.
3516 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3517 is always true, so this option does nothing.
3519 module_blacklist= [KNL] Do not load a comma-separated list of
3520 modules. Useful for debugging problem modules.
3523 [MOUSE] Maximum time between finger touching and
3524 leaving touchpad surface for touch to be considered
3525 a tap and be reported as a left button click (for
3526 touchpads working in absolute mode only).
3528 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3529 reporting absolute coordinates, such as tablets
3530 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3531 reporting absolute coordinates, such as tablets
3533 movablecore= [KNL,X86,IA-64,PPC,EARLY]
3534 Format: nn[KMGTPE] | nn%
3535 This parameter is the complement to kernelcore=, it
3536 specifies the amount of memory used for migratable
3537 allocations. If both kernelcore and movablecore is
3538 specified, then kernelcore will be at *least* the
3539 specified value but may be more. If movablecore on its
3540 own is specified, the administrator must be careful
3541 that the amount of memory usable for all allocations
3544 movable_node [KNL,EARLY] Boot-time switch to make hotplugable memory
3545 NUMA nodes to be movable. This means that the memory
3546 of such nodes will be usable only for movable
3547 allocations which rules out almost all kernel
3548 allocations. Use with caution!
3550 MTD_Partition= [MTD]
3551 Format: <name>,<region-number>,<size>,<offset>
3553 MTD_Region= [MTD] Format:
3554 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3557 See drivers/mtd/parsers/cmdlinepart.c
3560 ARM/S3C2412 JIVE boot control
3562 See arch/arm/mach-s3c/mach-jive.c
3564 mtouchusb.raw_coordinates=
3565 [HW] Make the MicroTouch USB driver use raw coordinates
3566 ('y', default) or cooked coordinates ('n')
3568 mtrr=debug [X86,EARLY]
3569 Enable printing debug information related to MTRR
3570 registers at boot time.
3572 mtrr_chunk_size=nn[KMG,X86,EARLY]
3573 used for mtrr cleanup. It is largest continuous chunk
3574 that could hold holes aka. UC entries.
3576 mtrr_gran_size=nn[KMG,X86,EARLY]
3577 Used for mtrr cleanup. It is granularity of mtrr block.
3579 Large value could prevent small alignment from
3582 mtrr_spare_reg_nr=n [X86,EARLY]
3584 Range: 0,7 : spare reg number
3586 Used for mtrr cleanup. It is spare mtrr entries number.
3587 Set to 2 or more if your graphical card needs more.
3589 multitce=off [PPC] This parameter disables the use of the pSeries
3590 firmware feature for updating multiple TCE entries
3593 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3595 netdev= [NET] Network devices parameters
3596 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3597 Note that mem_start is often overloaded to mean
3598 something different and driver-specific.
3599 This usage is only documented in each driver source
3602 netpoll.carrier_timeout=
3603 [NET] Specifies amount of time (in seconds) that
3604 netpoll should wait for a carrier. By default netpoll
3608 [NETFILTER] Enable connection tracking flow accounting
3609 0 to disable accounting
3610 1 to enable accounting
3614 [NFS] sets the pathname to the program which is used
3615 to update the NFS client cache entries.
3617 nfs.cache_getent_timeout=
3618 [NFS] sets the timeout after which an attempt to
3619 update a cache entry is deemed to have failed.
3621 nfs.callback_nr_threads=
3622 [NFSv4] set the total number of threads that the
3623 NFS client will assign to service NFSv4 callback
3626 nfs.callback_tcpport=
3627 [NFS] set the TCP port on which the NFSv4 callback
3628 channel should listen.
3631 [NFS] specifies the number of times the NFSv4 client
3632 retries the request before returning an EAGAIN error,
3633 after a reply of NFS4ERR_DELAY from the server.
3634 Only applies if the softerr mount option is enabled,
3635 and the specified value is >= 0.
3638 [NFS] enable 64-bit inode numbers.
3639 If zero, the NFS client will fake up a 32-bit inode
3640 number for the readdir() and stat() syscalls instead
3641 of returning the full 64-bit number.
3642 The default is to return 64-bit inode numbers.
3644 nfs.idmap_cache_timeout=
3645 [NFS] set the maximum lifetime for idmapper cache
3648 nfs.max_session_cb_slots=
3649 [NFSv4.1] Sets the maximum number of session
3650 slots the client will assign to the callback
3651 channel. This determines the maximum number of
3652 callbacks the client will process in parallel for
3653 a particular server.
3655 nfs.max_session_slots=
3656 [NFSv4.1] Sets the maximum number of session slots
3657 the client will attempt to negotiate with the server.
3658 This limits the number of simultaneous RPC requests
3659 that the client can send to the NFSv4.1 server.
3660 Note that there is little point in setting this
3661 value higher than the max_tcp_slot_table_limit.
3663 nfs.nfs4_disable_idmapping=
3664 [NFSv4] When set to the default of '1', this option
3665 ensures that both the RPC level authentication
3666 scheme and the NFS level operations agree to use
3667 numeric uids/gids if the mount is using the
3668 'sec=sys' security flavour. In effect it is
3669 disabling idmapping, which can make migration from
3670 legacy NFSv2/v3 systems to NFSv4 easier.
3671 Servers that do not support this mode of operation
3672 will be autodetected by the client, and it will fall
3673 back to using the idmapper.
3674 To turn off this behaviour, set the value to '0'.
3677 [NFS4] Specify an additional fixed unique ident-
3678 ification string that NFSv4 clients can insert into
3679 their nfs_client_id4 string. This is typically a
3680 UUID that is generated at system install time.
3682 nfs.recover_lost_locks=
3683 [NFSv4] Attempt to recover locks that were lost due
3684 to a lease timeout on the server. Please note that
3685 doing this risks data corruption, since there are
3686 no guarantees that the file will remain unchanged
3687 after the locks are lost.
3688 If you want to enable the kernel legacy behaviour of
3689 attempting to recover these locks, then set this
3691 The default parameter value of '0' causes the kernel
3692 not to attempt recovery of lost locks.
3694 nfs.send_implementation_id=
3695 [NFSv4.1] Send client implementation identification
3696 information in exchange_id requests.
3697 If zero, no implementation identification information
3699 The default is to send the implementation identification
3702 nfs4.layoutstats_timer=
3703 [NFSv4.2] Change the rate at which the kernel sends
3704 layoutstats to the pNFS metadata server.
3706 Setting this to value to 0 causes the kernel to use
3707 whatever value is the default set by the layout
3708 driver. A non-zero value sets the minimum interval
3709 in seconds between layoutstats transmissions.
3711 nfsd.inter_copy_offload_enable=
3712 [NFSv4.2] When set to 1, the server will support
3713 server-to-server copies for which this server is
3714 the destination of the copy.
3716 nfsd.nfs4_disable_idmapping=
3717 [NFSv4] When set to the default of '1', the NFSv4
3718 server will return only numeric uids and gids to
3719 clients using auth_sys, and will accept numeric uids
3720 and gids from such clients. This is intended to ease
3721 migration from NFSv2/v3.
3723 nfsd.nfsd4_ssc_umount_timeout=
3724 [NFSv4.2] When used as the destination of a
3725 server-to-server copy, knfsd temporarily mounts
3726 the source server. It caches the mount in case
3727 it will be needed again, and discards it if not
3728 used for the number of milliseconds specified by
3731 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3732 See Documentation/admin-guide/nfs/nfsroot.rst.
3734 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3735 See Documentation/admin-guide/nfs/nfsroot.rst.
3737 nfsrootdebug [NFS] enable nfsroot debugging messages.
3738 See Documentation/admin-guide/nfs/nfsroot.rst.
3740 nmi_backtrace.backtrace_idle [KNL]
3741 Dump stacks even of idle CPUs in response to an
3742 NMI stack-backtrace request.
3744 nmi_debug= [KNL,SH] Specify one or more actions to take
3745 when a NMI is triggered.
3746 Format: [state][,regs][,debounce][,die]
3748 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3749 Format: [panic,][nopanic,][num]
3751 0 - turn hardlockup detector in nmi_watchdog off
3752 1 - turn hardlockup detector in nmi_watchdog on
3753 When panic is specified, panic when an NMI watchdog
3754 timeout occurs (or 'nopanic' to not panic on an NMI
3755 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3756 To disable both hard and soft lockup detectors,
3757 please see 'nowatchdog'.
3758 This is useful when you use a panic=... timeout and
3759 need the box quickly up again.
3761 These settings can be accessed at runtime via
3762 the nmi_watchdog and hardlockup_panic sysctls.
3764 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3765 emulation library even if a 387 maths coprocessor
3768 no4lvl [RISCV,EARLY] Disable 4-level and 5-level paging modes.
3769 Forces kernel to use 3-level paging instead.
3771 no5lvl [X86-64,RISCV,EARLY] Disable 5-level paging mode. Forces
3772 kernel to use 4-level paging instead.
3776 noaltinstr [S390,EARLY] Disables alternative instructions
3777 patching (CPU alternatives feature).
3779 noapic [SMP,APIC,EARLY] Tells the kernel to not make use of any
3780 IOAPICs that may be present in the system.
3782 noautogroup Disable scheduler automatic task group creation.
3787 [HW] Never suspend the console
3788 Disable suspending of consoles during suspend and
3789 hibernate operations. Once disabled, debugging
3790 messages can reach various consoles while the rest
3791 of the system is being put to sleep (ie, while
3792 debugging driver suspend/resume hooks). This may
3793 not work reliably with all consoles, but is known
3794 to work with serial and VGA consoles.
3795 To facilitate more flexible debugging, we also add
3796 console_suspend, a printk module parameter to control
3797 it. Users could use console_suspend (usually
3798 /sys/module/printk/parameters/console_suspend) to
3799 turn on/off it dynamically.
3802 [KNL,EARLY] Disable object debugging
3804 nodsp [SH] Disable hardware DSP at boot time.
3806 noefi [EFI,EARLY] Disable EFI runtime services support.
3808 no_entry_flush [PPC,EARLY] Don't flush the L1-D cache when entering the kernel.
3813 This affects only 32-bit executables.
3814 noexec32=on: enable non-executable mappings (default)
3815 read doesn't imply executable mappings
3816 noexec32=off: disable non-executable mappings
3817 read implies executable mappings
3819 no_file_caps Tells the kernel not to honor file capabilities. The
3820 only way then for a file to be executed with privilege
3821 is to be setuid root or executed by root.
3823 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3825 nofsgsbase [X86] Disables FSGSBASE instructions.
3827 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3828 register save and restore. The kernel will only save
3829 legacy floating-point registers on task switch.
3831 nohalt [IA-64] Tells the kernel not to use the power saving
3832 function PAL_HALT_LIGHT when idle. This increases
3833 power-consumption. On the positive side, it reduces
3834 interrupt wake-up latency, which may improve performance
3835 in certain environments such as networked servers or
3840 Force pointers printed to the console or buffers to be
3841 unhashed. By default, when a pointer is printed via %p
3842 format string, that pointer is "hashed", i.e. obscured
3843 by hashing the pointer value. This is a security feature
3844 that hides actual kernel addresses from unprivileged
3845 users, but it also makes debugging the kernel more
3846 difficult since unequal pointers can no longer be
3847 compared. However, if this command-line option is
3848 specified, then all normal pointers will have their true
3849 value printed. This option should only be specified when
3850 debugging the kernel. Please do not use on production
3853 nohibernate [HIBERNATION] Disable hibernation and resume.
3855 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,SH] Forces the kernel to
3856 busy wait in do_idle() and not use the arch_cpu_idle()
3857 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3858 to be effective. This is useful on platforms where the
3859 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3860 correctly or when doing power measurements to evaluate
3861 the impact of the sleep instructions. This is also
3862 useful when using JTAG debugger.
3864 nohugeiomap [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge I/O mappings.
3866 nohugevmalloc [KNL,X86,PPC,ARM64,EARLY] Disable kernel huge vmalloc mappings.
3868 nohz= [KNL] Boottime enable/disable dynamic ticks
3869 Valid arguments: on, off
3872 nohz_full= [KNL,BOOT,SMP,ISOL]
3873 The argument is a cpu list, as described above.
3874 In kernels built with CONFIG_NO_HZ_FULL=y, set
3875 the specified list of CPUs whose tick will be stopped
3876 whenever possible. The boot CPU will be forced outside
3877 the range to maintain the timekeeping. Any CPUs
3878 in this list will have their RCU callbacks offloaded,
3879 just as if they had also been called out in the
3880 rcu_nocbs= boot parameter.
3882 Note that this argument takes precedence over
3883 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3885 noinitrd [RAM] Tells the kernel not to load any configured
3888 nointremap [X86-64,Intel-IOMMU,EARLY] Do not enable interrupt
3890 [Deprecated - use intremap=off]
3894 noinvpcid [X86,EARLY] Disable the INVPCID cpu feature.
3896 noiotrap [SH] Disables trapped I/O port accesses.
3898 noirqdebug [X86-32] Disables the code which attempts to detect and
3899 disable unhandled interrupt sources.
3901 noisapnp [ISAPNP] Disables ISA PnP code.
3903 nojitter [IA-64] Disables jitter checking for ITC timers.
3906 When CONFIG_RANDOMIZE_BASE is set, this disables
3907 kernel and module base offset ASLR (Address Space
3908 Layout Randomization).
3910 no-kvmapf [X86,KVM,EARLY] Disable paravirtualized asynchronous page
3913 no-kvmclock [X86,KVM,EARLY] Disable paravirtualized KVM clock driver
3915 nolapic [X86-32,APIC,EARLY] Do not enable or use the local APIC.
3917 nolapic_timer [X86-32,APIC,EARLY] Do not use the local APIC timer.
3919 nomca [IA-64] Disable machine check abort handling
3921 nomce [X86-32] Disable Machine Check Exception
3923 nomfgpt [X86-32] Disable Multi-Function General Purpose
3924 Timer usage (for AMD Geode machines).
3926 nomodeset Disable kernel modesetting. Most systems' firmware
3927 sets up a display mode and provides framebuffer memory
3928 for output. With nomodeset, DRM and fbdev drivers will
3929 not load if they could possibly displace the pre-
3930 initialized output. Only the system framebuffer will
3931 be available for use. The respective drivers will not
3932 perform display-mode changes or accelerated rendering.
3934 Useful as error fallback, or for testing and debugging.
3936 nomodule Disable module load
3938 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3939 shutdown the other cpus. Instead use the REBOOT_VECTOR
3942 nopat [X86,EARLY] Disable PAT (page attribute table extension of
3943 pagetables) support.
3945 nopcid [X86-64,EARLY] Disable the PCID cpu feature.
3947 nopku [X86] Disable Memory Protection Keys CPU feature found
3950 nopti [X86-64,EARLY]
3951 Equivalent to pti=off
3953 nopv= [X86,XEN,KVM,HYPER_V,VMWARE,EARLY]
3954 Disables the PV optimizations forcing the guest to run
3955 as generic guest with no PV drivers. Currently support
3956 XEN HVM, KVM, HYPER_V and VMWARE guest.
3958 nopvspin [X86,XEN,KVM,EARLY]
3959 Disables the qspinlock slow path using PV optimizations
3960 which allow the hypervisor to 'idle' the guest on lock
3963 norandmaps Don't use address space randomization. Equivalent to
3964 echo 0 > /proc/sys/kernel/randomize_va_space
3966 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3967 with UP alternatives
3969 noresume [SWSUSP] Disables resume and restores original swap
3974 no-scroll [VGA] Disables scrollback.
3975 This is required for the Braillex ib80-piezo Braille
3976 reader made by F.H. Papenmeier (Germany).
3978 nosgx [X86-64,SGX,EARLY] Disables Intel SGX kernel support.
3981 Disable SMAP (Supervisor Mode Access Prevention)
3982 even if it is supported by processor.
3984 nosmep [PPC64s,EARLY]
3985 Disable SMEP (Supervisor Mode Execution Prevention)
3986 even if it is supported by processor.
3988 nosmp [SMP,EARLY] Tells an SMP kernel to act as a UP kernel,
3989 and disable the IO APIC. legacy for "maxcpus=0".
3991 nosmt [KNL,MIPS,PPC,S390,EARLY] Disable symmetric multithreading (SMT).
3992 Equivalent to smt=1.
3994 [KNL,X86,PPC] Disable symmetric multithreading (SMT).
3995 nosmt=force: Force disable SMT, cannot be undone
3996 via the sysfs control file.
3998 nosoftlockup [KNL] Disable the soft-lockup detector.
4000 nospec_store_bypass_disable
4001 [HW,EARLY] Disable all mitigations for the Speculative
4002 Store Bypass vulnerability
4004 nospectre_bhb [ARM64,EARLY] Disable all mitigations for Spectre-BHB (branch
4005 history injection) vulnerability. System may allow data leaks
4008 nospectre_v1 [X86,PPC,EARLY] Disable mitigations for Spectre Variant 1
4009 (bounds check bypass). With this option data leaks are
4010 possible in the system.
4012 nospectre_v2 [X86,PPC_E500,ARM64,EARLY] Disable all mitigations
4013 for the Spectre variant 2 (indirect branch
4014 prediction) vulnerability. System may allow data
4015 leaks with this option.
4017 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES,RISCV,EARLY] Disable
4018 paravirtualized steal time accounting. steal time is
4019 computed, but won't influence scheduler behaviour
4021 nosync [HW,M68K] Disables sync negotiation for all devices.
4023 no_timer_check [X86,APIC] Disables the code which tests for
4024 broken timer IRQ sources.
4027 [PPC,EARLY] Don't flush the L1-D cache after accessing user data.
4029 novmcoredd [KNL,KDUMP]
4030 Disable device dump. Device dump allows drivers to
4031 append dump data to vmcore so you can collect driver
4032 specified debug info. Drivers can append the data
4033 without any limit and this data is stored in memory,
4034 so this may cause significant memory stress. Disabling
4035 device dump can help save memory but the driver debug
4036 data will be no longer available. This parameter
4037 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
4041 [X86,PV_OPS,EARLY] Disable paravirtualized VMware
4042 scheduler clock and use the default one.
4044 nowatchdog [KNL] Disable both lockup detectors, i.e.
4045 soft-lockup and NMI watchdog (hard-lockup).
4049 nox2apic [X86-64,APIC,EARLY] Do not enable x2APIC mode.
4051 NOTE: this parameter will be ignored on systems with the
4052 LEGACY_XAPIC_DISABLED bit set in the
4053 IA32_XAPIC_DISABLE_STATUS MSR.
4055 noxsave [BUGS=X86] Disables x86 extended register state save
4056 and restore using xsave. The kernel will fallback to
4057 enabling legacy floating-point and sse state.
4059 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
4060 register states. The kernel will fall back to use
4061 xsave to save the states. By using this parameter,
4062 performance of saving the states is degraded because
4063 xsave doesn't support modified optimization while
4064 xsaveopt supports it on xsaveopt enabled systems.
4066 noxsaves [X86] Disables xsaves and xrstors used in saving and
4067 restoring x86 extended register state in compacted
4068 form of xsave area. The kernel will fall back to use
4069 xsaveopt and xrstor to save and restore the states
4070 in standard form of xsave area. By using this
4071 parameter, xsave area per process might occupy more
4072 memory on xsaves enabled systems.
4074 nps_mtm_hs_ctr= [KNL,ARC]
4075 This parameter sets the maximum duration, in
4076 cycles, each HW thread of the CTOP can run
4077 without interruptions, before HW switches it.
4078 The actual maximum duration is 16 times this
4080 Format: integer between 1 and 255
4083 nptcg= [IA-64] Override max number of concurrent global TLB
4084 purges which is reported from either PAL_VM_SUMMARY or
4087 nr_cpus= [SMP,EARLY] Maximum number of processors that an SMP kernel
4088 could support. nr_cpus=n : n >= 1 limits the kernel to
4089 support 'n' processors. It could be larger than the
4090 number of already plugged CPU during bootup, later in
4091 runtime you can physically add extra cpu until it reaches
4092 n. So during boot up some boot time memory for per-cpu
4093 variables need be pre-allocated for later physical cpu
4096 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
4098 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86, EARLY]
4099 Disable NUMA, Only set up a single NUMA node
4100 spanning all memory.
4102 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
4104 Allowed values are enable and disable
4106 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
4107 'node', 'default' can be specified
4108 This can be set from sysctl after boot.
4109 See Documentation/admin-guide/sysctl/vm.rst for details.
4111 ohci1394_dma=early [HW,EARLY] enable debugging via the ohci1394 driver.
4112 See Documentation/core-api/debugging-via-ohci1394.rst for more
4115 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
4116 Rather than timing out after 20 ms if an EC
4117 command is not properly ACKed, override the length
4118 of the timeout. We have interrupts disabled while
4119 waiting for the ACK, so if this is set too high
4120 interrupts *may* be lost!
4122 omap_mux= [OMAP] Override bootloader pin multiplexing.
4123 Format: <mux_mode0.mode_name=value>...
4124 For example, to override I2C bus2:
4125 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
4127 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
4129 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
4131 boundary - index of last SLC block on Flex-OneNAND.
4132 The remaining blocks are configured as MLC blocks.
4133 lock - Configure if Flex-OneNAND boundary should be locked.
4134 Once locked, the boundary cannot be changed.
4135 1 indicates lock status, 0 indicates unlock status.
4137 oops=panic [KNL,EARLY]
4138 Always panic on oopses. Default is to just kill the
4139 process, but there is a small probability of
4140 deadlocking the machine.
4141 This will also cause panics on machine check exceptions.
4142 Useful together with panic=30 to trigger a reboot.
4145 [KNL] Boolean flag to control whether the page allocator
4146 should randomize its free lists. The randomization may
4147 be automatically enabled if the kernel detects it is
4148 running on a platform with a direct-mapped memory-side
4149 cache, and this parameter can be used to
4150 override/disable that behavior. The state of the flag
4151 can be read from sysfs at:
4152 /sys/module/page_alloc/parameters/shuffle.
4154 page_owner= [KNL,EARLY] Boot-time page_owner enabling option.
4155 Storage of the information about who allocated
4156 each page is disabled in default. With this switch,
4158 on: enable the feature
4160 page_poison= [KNL,EARLY] Boot-time parameter changing the state of
4161 poisoning on the buddy allocator, available with
4162 CONFIG_PAGE_POISONING=y.
4163 off: turn off poisoning (default)
4164 on: turn on poisoning
4166 page_reporting.page_reporting_order=
4167 [KNL] Minimal page reporting order
4169 Adjust the minimal page reporting order. The page
4170 reporting is disabled when it exceeds MAX_PAGE_ORDER.
4172 panic= [KNL] Kernel behaviour on panic: delay <timeout>
4173 timeout > 0: seconds before rebooting
4174 timeout = 0: wait forever
4175 timeout < 0: reboot immediately
4178 panic_on_taint= [KNL,EARLY]
4179 Bitmask for conditionally calling panic() in add_taint()
4180 Format: <hex>[,nousertaint]
4181 Hexadecimal bitmask representing the set of TAINT flags
4182 that will cause the kernel to panic when add_taint() is
4183 called with any of the flags in this set.
4184 The optional switch "nousertaint" can be utilized to
4185 prevent userspace forced crashes by writing to sysctl
4186 /proc/sys/kernel/tainted any flagset matching with the
4187 bitmask set on panic_on_taint.
4188 See Documentation/admin-guide/tainted-kernels.rst for
4189 extra details on the taint flags that users can pick
4190 to compose the bitmask to assign to panic_on_taint.
4192 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump
4195 panic_print= Bitmask for printing system info when panic happens.
4196 User can chose combination of the following bits:
4197 bit 0: print all tasks info
4198 bit 1: print system memory info
4199 bit 2: print timer info
4200 bit 3: print locks info if CONFIG_LOCKDEP is on
4201 bit 4: print ftrace buffer
4202 bit 5: print all printk messages in buffer
4203 bit 6: print all CPUs backtrace (if available in the arch)
4204 *Be aware* that this option may print a _lot_ of lines,
4205 so there are risks of losing older messages in the log.
4206 Use this option carefully, maybe worth to setup a
4207 bigger log buffer with "log_buf_len" along with this.
4209 parkbd.port= [HW] Parallel port number the keyboard adapter is
4210 connected to, default is 0.
4212 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
4213 0 for XT, 1 for AT (default is AT).
4216 parport= [HW,PPT] Specify parallel ports. 0 disables.
4217 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
4218 Use 'auto' to force the driver to use any
4219 IRQ/DMA settings detected (the default is to
4220 ignore detected IRQ/DMA settings because of
4221 possible conflicts). You can specify the base
4222 address, IRQ, and DMA settings; IRQ and DMA
4223 should be numbers, or 'auto' (for using detected
4224 settings on that particular port), or 'nofifo'
4225 (to avoid using a FIFO even if it is detected).
4226 Parallel ports are assigned in the order they
4227 are specified on the command line, starting
4230 parport_init_mode= [HW,PPT]
4231 Configure VIA parallel port to operate in
4232 a specific mode. This is necessary on Pegasos
4233 computer where firmware has no options for setting
4234 up parallel port mode and sets it to spp.
4235 Currently this function knows 686a and 8231 chips.
4236 Format: [spp|ps2|epp|ecp|ecpepp]
4238 pata_legacy.all= [HW,LIBATA]
4240 Set to non-zero to probe primary and secondary ISA
4241 port ranges on PCI systems where no PCI PATA device
4242 has been found at either range. Disabled by default.
4244 pata_legacy.autospeed= [HW,LIBATA]
4246 Set to non-zero if a chip is present that snoops speed
4247 changes. Disabled by default.
4249 pata_legacy.ht6560a= [HW,LIBATA]
4251 Set to 1, 2, or 3 for HT 6560A on the primary channel,
4252 the secondary channel, or both channels respectively.
4253 Disabled by default.
4255 pata_legacy.ht6560b= [HW,LIBATA]
4257 Set to 1, 2, or 3 for HT 6560B on the primary channel,
4258 the secondary channel, or both channels respectively.
4259 Disabled by default.
4261 pata_legacy.iordy_mask= [HW,LIBATA]
4263 IORDY enable mask. Set individual bits to allow IORDY
4264 for the respective channel. Bit 0 is for the first
4265 legacy channel handled by this driver, bit 1 is for
4266 the second channel, and so on. The sequence will often
4267 correspond to the primary legacy channel, the secondary
4268 legacy channel, and so on, but the handling of a PCI
4269 bus and the use of other driver options may interfere
4270 with the sequence. By default IORDY is allowed across
4273 pata_legacy.opti82c46x= [HW,LIBATA]
4275 Set to 1, 2, or 3 for Opti 82c611A on the primary
4276 channel, the secondary channel, or both channels
4277 respectively. Disabled by default.
4279 pata_legacy.opti82c611a= [HW,LIBATA]
4281 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4282 channel, the secondary channel, or both channels
4283 respectively. Disabled by default.
4285 pata_legacy.pio_mask= [HW,LIBATA]
4287 PIO mode mask for autospeed devices. Set individual
4288 bits to allow the use of the respective PIO modes.
4289 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4290 All modes allowed by default.
4292 pata_legacy.probe_all= [HW,LIBATA]
4294 Set to non-zero to probe tertiary and further ISA
4295 port ranges on PCI systems. Disabled by default.
4297 pata_legacy.probe_mask= [HW,LIBATA]
4299 Probe mask for legacy ISA PATA ports. Depending on
4300 platform configuration and the use of other driver
4301 options up to 6 legacy ports are supported: 0x1f0,
4302 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4303 of individual ports can be disabled by setting the
4304 corresponding bits in the mask to 1. Bit 0 is for
4305 the first port in the list above (0x1f0), and so on.
4306 By default all supported ports are probed.
4308 pata_legacy.qdi= [HW,LIBATA]
4310 Set to non-zero to probe QDI controllers. By default
4311 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4313 pata_legacy.winbond= [HW,LIBATA]
4315 Set to non-zero to probe Winbond controllers. Use
4316 the standard I/O port (0x130) if 1, otherwise the
4317 value given is the I/O port to use (typically 0x1b0).
4318 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4321 pata_platform.pio_mask= [HW,LIBATA]
4323 Supported PIO mode mask. Set individual bits to allow
4324 the use of the respective PIO modes. Bit 0 is for
4325 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4329 Halt all CPUs after the first oops has been printed for
4330 the specified number of seconds. This is to be used if
4331 your oopses keep scrolling off the screen.
4335 pci=option[,option...] [PCI,EARLY] various PCI subsystem options.
4337 Some options herein operate on a specific device
4338 or a set of devices (<pci_dev>). These are
4339 specified in one of the following formats:
4341 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4342 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4344 Note: the first format specifies a PCI
4345 bus/device/function address which may change
4346 if new hardware is inserted, if motherboard
4347 firmware changes, or due to changes caused
4348 by other kernel parameters. If the
4349 domain is left unspecified, it is
4350 taken to be zero. Optionally, a path
4351 to a device through multiple device/function
4352 addresses can be specified after the base
4353 address (this is more robust against
4354 renumbering issues). The second format
4355 selects devices using IDs from the
4356 configuration space which may match multiple
4357 devices in the system.
4359 earlydump dump PCI config space before the kernel
4361 off [X86] don't probe for the PCI bus
4362 bios [X86-32] force use of PCI BIOS, don't access
4363 the hardware directly. Use this if your machine
4364 has a non-standard PCI host bridge.
4365 nobios [X86-32] disallow use of PCI BIOS, only direct
4366 hardware access methods are allowed. Use this
4367 if you experience crashes upon bootup and you
4368 suspect they are caused by the BIOS.
4369 conf1 [X86] Force use of PCI Configuration Access
4370 Mechanism 1 (config address in IO port 0xCF8,
4371 data in IO port 0xCFC, both 32-bit).
4372 conf2 [X86] Force use of PCI Configuration Access
4373 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4374 the function, IO port 0xCFA, also 8-bit, sets
4375 bus number. The config space is then accessed
4376 through ports 0xC000-0xCFFF).
4377 See http://wiki.osdev.org/PCI for more info
4378 on the configuration access mechanisms.
4379 noaer [PCIE] If the PCIEAER kernel config parameter is
4380 enabled, this kernel boot option can be used to
4381 disable the use of PCIE advanced error reporting.
4382 nodomains [PCI] Disable support for multiple PCI
4383 root domains (aka PCI segments, in ACPI-speak).
4384 nommconf [X86] Disable use of MMCONFIG for PCI
4386 check_enable_amd_mmconf [X86] check for and enable
4387 properly configured MMIO access to PCI
4388 config space on AMD family 10h CPU
4389 nomsi [MSI] If the PCI_MSI kernel config parameter is
4390 enabled, this kernel boot option can be used to
4391 disable the use of MSI interrupts system-wide.
4392 noioapicquirk [APIC] Disable all boot interrupt quirks.
4393 Safety option to keep boot IRQs enabled. This
4394 should never be necessary.
4395 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4396 primary IO-APIC for bridges that cannot disable
4397 boot IRQs. This fixes a source of spurious IRQs
4398 when the system masks IRQs.
4399 noioapicreroute [APIC] Disable workaround that uses the
4400 boot IRQ equivalent of an IRQ that connects to
4401 a chipset where boot IRQs cannot be disabled.
4402 The opposite of ioapicreroute.
4403 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4404 routing table. These calls are known to be buggy
4405 on several machines and they hang the machine
4406 when used, but on other computers it's the only
4407 way to get the interrupt routing table. Try
4408 this option if the kernel is unable to allocate
4409 IRQs or discover secondary PCI buses on your
4411 rom [X86] Assign address space to expansion ROMs.
4412 Use with caution as certain devices share
4413 address decoders between ROMs and other
4415 norom [X86] Do not assign address space to
4416 expansion ROMs that do not already have
4417 BIOS assigned address ranges.
4418 nobar [X86] Do not assign address space to the
4419 BARs that weren't assigned by the BIOS.
4420 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4421 assigned automatically to PCI devices. You can
4422 make the kernel exclude IRQs of your ISA cards
4424 pirqaddr=0xAAAAA [X86] Specify the physical address
4425 of the PIRQ table (normally generated
4426 by the BIOS) if it is outside the
4427 F0000h-100000h range.
4428 lastbus=N [X86] Scan all buses thru bus #N. Can be
4429 useful if the kernel is unable to find your
4430 secondary buses and you want to tell it
4431 explicitly which ones they are.
4432 assign-busses [X86] Always assign all PCI bus
4433 numbers ourselves, overriding
4434 whatever the firmware may have done.
4435 usepirqmask [X86] Honor the possible IRQ mask stored
4436 in the BIOS $PIR table. This is needed on
4437 some systems with broken BIOSes, notably
4438 some HP Pavilion N5400 and Omnibook XE3
4439 notebooks. This will have no effect if ACPI
4440 IRQ routing is enabled.
4441 noacpi [X86] Do not use ACPI for IRQ routing
4442 or for PCI scanning.
4443 use_crs [X86] Use PCI host bridge window information
4444 from ACPI. On BIOSes from 2008 or later, this
4445 is enabled by default. If you need to use this,
4446 please report a bug.
4447 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4448 If you need to use this, please report a bug.
4449 use_e820 [X86] Use E820 reservations to exclude parts of
4450 PCI host bridge windows. This is a workaround
4451 for BIOS defects in host bridge _CRS methods.
4452 If you need to use this, please report a bug to
4453 <linux-pci@vger.kernel.org>.
4454 no_e820 [X86] Ignore E820 reservations for PCI host
4455 bridge windows. This is the default on modern
4456 hardware. If you need to use this, please report
4457 a bug to <linux-pci@vger.kernel.org>.
4458 routeirq Do IRQ routing for all PCI devices.
4459 This is normally done in pci_enable_device(),
4460 so this option is a temporary workaround
4461 for broken drivers that don't call it.
4462 skip_isa_align [X86] do not align io start addr, so can
4463 handle more pci cards
4464 noearly [X86] Don't do any early type 1 scanning.
4465 This might help on some broken boards which
4466 machine check when some devices' config space
4467 is read. But various workarounds are disabled
4468 and some IOMMU drivers will not work.
4469 bfsort Sort PCI devices into breadth-first order.
4470 This sorting is done to get a device
4471 order compatible with older (<= 2.4) kernels.
4472 nobfsort Don't sort PCI devices into breadth-first order.
4473 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4474 tuning and use the BIOS-configured MPS defaults.
4475 pcie_bus_safe Set every device's MPS to the largest value
4476 supported by all devices below the root complex.
4477 pcie_bus_perf Set device MPS to the largest allowable MPS
4478 based on its parent bus. Also set MRRS (Max
4479 Read Request Size) to the largest supported
4480 value (no larger than the MPS that the device
4481 or bus can support) for best performance.
4482 pcie_bus_peer2peer Set every device's MPS to 128B, which
4483 every device is guaranteed to support. This
4484 configuration allows peer-to-peer DMA between
4485 any pair of devices, possibly at the cost of
4486 reduced performance. This also guarantees
4487 that hot-added devices will work.
4488 cbiosize=nn[KMG] The fixed amount of bus space which is
4489 reserved for the CardBus bridge's IO window.
4490 The default value is 256 bytes.
4491 cbmemsize=nn[KMG] The fixed amount of bus space which is
4492 reserved for the CardBus bridge's memory
4493 window. The default value is 64 megabytes.
4496 [<order of align>@]<pci_dev>[; ...]
4497 Specifies alignment and device to reassign
4498 aligned memory resources. How to
4499 specify the device is described above.
4500 If <order of align> is not specified,
4501 PAGE_SIZE is used as alignment.
4502 A PCI-PCI bridge can be specified if resource
4503 windows need to be expanded.
4504 To specify the alignment for several
4505 instances of a device, the PCI vendor,
4506 device, subvendor, and subdevice may be
4507 specified, e.g., 12@pci:8086:9c22:103c:198f
4508 for 4096-byte alignment.
4509 ecrc= Enable/disable PCIe ECRC (transaction layer
4510 end-to-end CRC checking). Only effective if
4511 OS has native AER control (either granted by
4512 ACPI _OSC or forced via "pcie_ports=native")
4513 bios: Use BIOS/firmware settings. This is the
4517 hpiosize=nn[KMG] The fixed amount of bus space which is
4518 reserved for hotplug bridge's IO window.
4519 Default size is 256 bytes.
4520 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4521 reserved for hotplug bridge's MMIO window.
4522 Default size is 2 megabytes.
4523 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4524 reserved for hotplug bridge's MMIO_PREF window.
4525 Default size is 2 megabytes.
4526 hpmemsize=nn[KMG] The fixed amount of bus space which is
4527 reserved for hotplug bridge's MMIO and
4529 Default size is 2 megabytes.
4530 hpbussize=nn The minimum amount of additional bus numbers
4531 reserved for buses below a hotplug bridge.
4533 realloc= Enable/disable reallocating PCI bridge resources
4534 if allocations done by BIOS are too small to
4535 accommodate resources required by all child
4537 off: Turn realloc off
4539 realloc same as realloc=on
4540 noari do not use PCIe ARI.
4541 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4542 do not use PCIe ATS (and IOMMU device IOTLB).
4543 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4544 only look for one device below a PCIe downstream
4546 big_root_window Try to add a big 64bit memory window to the PCIe
4547 root complex on AMD CPUs. Some GFX hardware
4548 can resize a BAR to allow access to all VRAM.
4549 Adding the window is slightly risky (it may
4550 conflict with unreported devices), so this
4552 disable_acs_redir=<pci_dev>[; ...]
4553 Specify one or more PCI devices (in the format
4554 specified above) separated by semicolons.
4555 Each device specified will have the PCI ACS
4556 redirect capabilities forced off which will
4557 allow P2P traffic between devices through
4558 bridges without forcing it upstream. Note:
4559 this removes isolation between devices and
4560 may put more devices in an IOMMU group.
4561 force_floating [S390] Force usage of floating interrupts.
4562 nomio [S390] Do not use MIO instructions.
4563 norid [S390] ignore the RID field and force use of
4564 one PCI domain per PCI function
4566 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4569 force Enable ASPM even on devices that claim not to support it.
4570 WARNING: Forcing ASPM on may cause system lockups.
4572 pcie_ports= [PCIE] PCIe port services handling:
4573 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4574 even if the platform doesn't give the OS permission to
4575 use them. This may cause conflicts if the platform
4576 also tries to use these services.
4577 dpc-native Use native PCIe service for DPC only. May
4578 cause conflicts if firmware uses AER or DPC.
4579 compat Disable native PCIe services (PME, AER, DPC, PCIe
4582 pcie_port_pm= [PCIE] PCIe port power management handling:
4583 off Disable power management of all PCIe ports
4584 force Forcibly enable power management of all PCIe ports
4586 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4587 nomsi Do not use MSI for native PCIe PME signaling (this makes
4588 all PCIe root ports use INTx for all services).
4590 pcmv= [HW,PCMCIA] BadgePAD 4
4594 Keep all power-domains already enabled by bootloader on,
4595 even if no driver has claimed them. This is useful
4596 for debug and development, but should not be
4597 needed on a platform with proper driver support.
4599 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4602 See arch/parisc/kernel/pdc_chassis.c
4604 percpu_alloc= [MM,EARLY]
4605 Select which percpu first chunk allocator to use.
4606 Currently supported values are "embed" and "page".
4607 Archs may support subset or none of the selections.
4608 See comments in mm/percpu.c for details on each
4609 allocator. This parameter is primarily for debugging
4610 and performance comparison.
4612 pirq= [SMP,APIC] Manual mp-table setup
4613 See Documentation/arch/x86/i386/IO-APIC.rst.
4615 plip= [PPT,NET] Parallel port network link
4616 Format: { parport<nr> | timid | 0 }
4617 See also Documentation/admin-guide/parport.rst.
4619 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4620 Override pmtimer IOPort with a hex value.
4623 pmu_override= [PPC] Override the PMU.
4624 This option takes over the PMU facility, so it is no
4625 longer usable by perf. Setting this option starts the
4626 PMU counters by setting MMCR0 to 0 (the FC bit is
4627 cleared). If a number is given, then MMCR1 is set to
4628 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4631 pm_debug_messages [SUSPEND,KNL]
4632 Enable suspend/resume debug messages during boot up.
4635 Enable PNP debug messages (depends on the
4636 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4637 via /sys/module/pnp/parameters/debug. We always show
4638 current resource usage; turning this on also shows
4639 possible settings and some assignment information.
4645 { on | off | curr | res | no-curr | no-res }
4648 [ISAPNP] Exclude IRQs for the autoconfiguration
4651 [ISAPNP] Exclude DMAs for the autoconfiguration
4653 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4654 Ranges are in pairs (I/O port base and size).
4657 [ISAPNP] Exclude memory regions for the
4659 Ranges are in pairs (memory base and size).
4661 ports= [IP_VS_FTP] IPVS ftp helper module
4663 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4665 Format: <port>,<port>....
4667 possible_cpus= [SMP,S390,X86]
4668 Format: <unsigned int>
4669 Set the number of possible CPUs, overriding the
4670 regular discovery mechanisms (such as ACPI/FW, etc).
4672 powersave=off [PPC] This option disables power saving features.
4673 It specifically disables cpuidle and sets the
4674 platform machine description specific power_save
4675 function to NULL. On Idle the CPU just reduces
4678 ppc_strict_facility_enable
4679 [PPC,ENABLE] This option catches any kernel floating point,
4680 Altivec, VSX and SPE outside of regions specifically
4681 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4682 There is some performance impact when enabling this.
4686 Disable Hardware Transactional Memory
4689 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4690 none - Limited to cond_resched() calls
4691 voluntary - Limited to cond_resched() and might_sleep() calls
4692 full - Any section that isn't explicitly preempt disabled
4693 can be preempted anytime.
4695 print-fatal-signals=
4696 [KNL] debug: print fatal signals
4698 If enabled, warn about various signal handling
4699 related application anomalies: too many signals,
4700 too many POSIX.1 timers, fatal signals causing a
4703 If you hit the warning due to signal overflow,
4704 you might want to try "ulimit -i unlimited".
4708 printk.always_kmsg_dump=
4709 Trigger kmsg_dump for cases other than kernel oops or
4711 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4714 printk.console_no_auto_verbose=
4715 Disable console loglevel raise on oops, panic
4716 or lockdep-detected issues (only if lock debug is on).
4717 With an exception to setups with low baudrate on
4718 serial console, keeping this 0 is a good choice
4719 in order to provide more debug information.
4721 default: 0 (auto_verbose is enabled)
4723 printk.devkmsg={on,off,ratelimit}
4724 Control writing to /dev/kmsg.
4725 on - unlimited logging to /dev/kmsg from userspace
4726 off - logging to /dev/kmsg disabled
4727 ratelimit - ratelimit the logging
4730 printk.time= Show timing data prefixed to each printk message line
4731 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4733 processor.max_cstate= [HW,ACPI]
4734 Limit processor to maximum C-state
4735 max_cstate=9 overrides any DMI blacklist limit.
4737 processor.nocst [HW,ACPI]
4738 Ignore the _CST method to determine C-states,
4739 instead using the legacy FADT method
4741 profile= [KNL] Enable kernel profiling via /proc/profile
4742 Format: [<profiletype>,]<number>
4743 Param: <profiletype>: "schedule", "sleep", or "kvm"
4744 [defaults to kernel profiling]
4745 Param: "schedule" - profile schedule points.
4746 Param: "sleep" - profile D-state sleeping (millisecs).
4747 Requires CONFIG_SCHEDSTATS
4748 Param: "kvm" - profile VM exits.
4749 Param: <number> - step/bucket size as a power of 2 for
4750 statistical time based profiling.
4752 prompt_ramdisk= [RAM] [Deprecated]
4754 prot_virt= [S390] enable hosting protected virtual machines
4755 isolated from the hypervisor (if hardware supports
4759 psi= [KNL] Enable or disable pressure stall information
4763 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4764 probe for; one of (bare|imps|exps|lifebook|any).
4765 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4767 psmouse.resetafter= [HW,MOUSE]
4768 Try to reset the device after so many bad packets
4771 [HW,MOUSE] Set desired mouse resolution, in dpi.
4772 psmouse.smartscroll=
4773 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4774 0 = disabled, 1 = enabled (default).
4776 pstore.backend= Specify the name of the pstore backend to use
4778 pti= [X86-64] Control Page Table Isolation of user and
4779 kernel address spaces. Disabling this feature
4780 removes hardening, but improves performance of
4781 system calls and interrupts.
4783 on - unconditionally enable
4784 off - unconditionally disable
4785 auto - kernel detects whether your CPU model is
4786 vulnerable to issues that PTI mitigates
4788 Not specifying this option is equivalent to pti=auto.
4791 [KNL] Number of legacy pty's. Overwrites compiled-in
4794 quiet [KNL,EARLY] Disable most log messages
4798 radix_hcall_invalidate=on [PPC/PSERIES]
4799 Disable RADIX GTSE feature and use hcall for TLB
4803 See Documentation/admin-guide/md.rst.
4805 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4806 See Documentation/admin-guide/blockdev/ramdisk.rst.
4808 ramdisk_start= [RAM] RAM disk image start address
4810 random.trust_cpu=off
4811 [KNL,EARLY] Disable trusting the use of the CPU's
4812 random number generator (if available) to
4813 initialize the kernel's RNG.
4815 random.trust_bootloader=off
4816 [KNL,EARLY] Disable trusting the use of the a seed
4817 passed by the bootloader (if available) to
4818 initialize the kernel's RNG.
4820 randomize_kstack_offset=
4821 [KNL,EARLY] Enable or disable kernel stack offset
4822 randomization, which provides roughly 5 bits of
4823 entropy, frustrating memory corruption attacks
4824 that depend on stack address determinism or
4825 cross-syscall address exposures. This is only
4826 available on architectures that have defined
4827 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4828 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4829 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4831 ras=option[,option,...] [KNL] RAS-specific options
4834 Disable the Correctable Errors Collector,
4835 see CONFIG_RAS_CEC help text.
4837 rcu_nocbs[=cpu-list]
4838 [KNL] The optional argument is a cpu list,
4841 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4842 enable the no-callback CPU mode, which prevents
4843 such CPUs' callbacks from being invoked in
4844 softirq context. Invocation of such CPUs' RCU
4845 callbacks will instead be offloaded to "rcuox/N"
4846 kthreads created for that purpose, where "x" is
4847 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4848 for the kthreads that mediate grace periods; and
4849 "N" is the CPU number. This reduces OS jitter on
4850 the offloaded CPUs, which can be useful for HPC
4851 and real-time workloads. It can also improve
4852 energy efficiency for asymmetric multiprocessors.
4854 If a cpulist is passed as an argument, the specified
4855 list of CPUs is set to no-callback mode from boot.
4857 Otherwise, if the '=' sign and the cpulist
4858 arguments are omitted, no CPU will be set to
4859 no-callback mode from boot but the mode may be
4860 toggled at runtime via cpusets.
4862 Note that this argument takes precedence over
4863 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4866 Rather than requiring that offloaded CPUs
4867 (specified by rcu_nocbs= above) explicitly
4868 awaken the corresponding "rcuoN" kthreads,
4869 make these kthreads poll for callbacks.
4870 This improves the real-time response for the
4871 offloaded CPUs by relieving them of the need to
4872 wake up the corresponding kthread, but degrades
4873 energy efficiency by requiring that the kthreads
4874 periodically wake up to do the polling.
4876 rcutree.blimit= [KNL]
4877 Set maximum number of finished RCU callbacks to
4878 process in one batch.
4880 rcutree.do_rcu_barrier= [KNL]
4881 Request a call to rcu_barrier(). This is
4882 throttled so that userspace tests can safely
4883 hammer on the sysfs variable if they so choose.
4884 If triggered before the RCU grace-period machinery
4885 is fully active, this will error out with EAGAIN.
4887 rcutree.dump_tree= [KNL]
4888 Dump the structure of the rcu_node combining tree
4889 out at early boot. This is used for diagnostic
4890 purposes, to verify correct tree setup.
4892 rcutree.gp_cleanup_delay= [KNL]
4893 Set the number of jiffies to delay each step of
4894 RCU grace-period cleanup.
4896 rcutree.gp_init_delay= [KNL]
4897 Set the number of jiffies to delay each step of
4898 RCU grace-period initialization.
4900 rcutree.gp_preinit_delay= [KNL]
4901 Set the number of jiffies to delay each step of
4902 RCU grace-period pre-initialization, that is,
4903 the propagation of recent CPU-hotplug changes up
4904 the rcu_node combining tree.
4906 rcutree.jiffies_till_first_fqs= [KNL]
4907 Set delay from grace-period initialization to
4908 first attempt to force quiescent states.
4909 Units are jiffies, minimum value is zero,
4910 and maximum value is HZ.
4912 rcutree.jiffies_till_next_fqs= [KNL]
4913 Set delay between subsequent attempts to force
4914 quiescent states. Units are jiffies, minimum
4915 value is one, and maximum value is HZ.
4917 rcutree.jiffies_till_sched_qs= [KNL]
4918 Set required age in jiffies for a
4919 given grace period before RCU starts
4920 soliciting quiescent-state help from
4921 rcu_note_context_switch() and cond_resched().
4922 If not specified, the kernel will calculate
4923 a value based on the most recent settings
4924 of rcutree.jiffies_till_first_fqs
4925 and rcutree.jiffies_till_next_fqs.
4926 This calculated value may be viewed in
4927 rcutree.jiffies_to_sched_qs. Any attempt to set
4928 rcutree.jiffies_to_sched_qs will be cheerfully
4931 rcutree.kthread_prio= [KNL,BOOT]
4932 Set the SCHED_FIFO priority of the RCU per-CPU
4933 kthreads (rcuc/N). This value is also used for
4934 the priority of the RCU boost threads (rcub/N)
4935 and for the RCU grace-period kthreads (rcu_bh,
4936 rcu_preempt, and rcu_sched). If RCU_BOOST is
4937 set, valid values are 1-99 and the default is 1
4938 (the least-favored priority). Otherwise, when
4939 RCU_BOOST is not set, valid values are 0-99 and
4940 the default is zero (non-realtime operation).
4941 When RCU_NOCB_CPU is set, also adjust the
4942 priority of NOCB callback kthreads.
4944 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4945 On callback-offloaded (rcu_nocbs) CPUs,
4946 RCU reduces the lock contention that would
4947 otherwise be caused by callback floods through
4948 use of the ->nocb_bypass list. However, in the
4949 common non-flooded case, RCU queues directly to
4950 the main ->cblist in order to avoid the extra
4951 overhead of the ->nocb_bypass list and its lock.
4952 But if there are too many callbacks queued during
4953 a single jiffy, RCU pre-queues the callbacks into
4954 the ->nocb_bypass queue. The definition of "too
4955 many" is supplied by this kernel boot parameter.
4957 rcutree.qhimark= [KNL]
4958 Set threshold of queued RCU callbacks beyond which
4959 batch limiting is disabled.
4961 rcutree.qlowmark= [KNL]
4962 Set threshold of queued RCU callbacks below which
4963 batch limiting is re-enabled.
4965 rcutree.qovld= [KNL]
4966 Set threshold of queued RCU callbacks beyond which
4967 RCU's force-quiescent-state scan will aggressively
4968 enlist help from cond_resched() and sched IPIs to
4969 help CPUs more quickly reach quiescent states.
4970 Set to less than zero to make this be set based
4971 on rcutree.qhimark at boot time and to zero to
4972 disable more aggressive help enlistment.
4974 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4975 Set the page-cache refill delay (in milliseconds)
4976 in response to low-memory conditions. The range
4977 of permitted values is in the range 0:100000.
4979 rcutree.rcu_divisor= [KNL]
4980 Set the shift-right count to use to compute
4981 the callback-invocation batch limit bl from
4982 the number of callbacks queued on this CPU.
4983 The result will be bounded below by the value of
4984 the rcutree.blimit kernel parameter. Every bl
4985 callbacks, the softirq handler will exit in
4986 order to allow the CPU to do other work.
4988 Please note that this callback-invocation batch
4989 limit applies only to non-offloaded callback
4990 invocation. Offloaded callbacks are instead
4991 invoked in the context of an rcuoc kthread, which
4992 scheduler will preempt as it does any other task.
4994 rcutree.rcu_fanout_exact= [KNL]
4995 Disable autobalancing of the rcu_node combining
4996 tree. This is used by rcutorture, and might
4997 possibly be useful for architectures having high
4998 cache-to-cache transfer latencies.
5000 rcutree.rcu_fanout_leaf= [KNL]
5001 Change the number of CPUs assigned to each
5002 leaf rcu_node structure. Useful for very
5003 large systems, which will choose the value 64,
5004 and for NUMA systems with large remote-access
5005 latencies, which will choose a value aligned
5006 with the appropriate hardware boundaries.
5008 rcutree.rcu_min_cached_objs= [KNL]
5009 Minimum number of objects which are cached and
5010 maintained per one CPU. Object size is equal
5011 to PAGE_SIZE. The cache allows to reduce the
5012 pressure to page allocator, also it makes the
5013 whole algorithm to behave better in low memory
5016 rcutree.rcu_nocb_gp_stride= [KNL]
5017 Set the number of NOCB callback kthreads in
5018 each group, which defaults to the square root
5019 of the number of CPUs. Larger numbers reduce
5020 the wakeup overhead on the global grace-period
5021 kthread, but increases that same overhead on
5022 each group's NOCB grace-period kthread.
5024 rcutree.rcu_kick_kthreads= [KNL]
5025 Cause the grace-period kthread to get an extra
5026 wake_up() if it sleeps three times longer than
5027 it should at force-quiescent-state time.
5028 This wake_up() will be accompanied by a
5029 WARN_ONCE() splat and an ftrace_dump().
5031 rcutree.rcu_resched_ns= [KNL]
5032 Limit the time spend invoking a batch of RCU
5033 callbacks to the specified number of nanoseconds.
5034 By default, this limit is checked only once
5035 every 32 callbacks in order to limit the pain
5036 inflicted by local_clock() overhead.
5038 rcutree.rcu_unlock_delay= [KNL]
5039 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
5040 this specifies an rcu_read_unlock()-time delay
5041 in microseconds. This defaults to zero.
5042 Larger delays increase the probability of
5043 catching RCU pointer leaks, that is, buggy use
5044 of RCU-protected pointers after the relevant
5045 rcu_read_unlock() has completed.
5047 rcutree.sysrq_rcu= [KNL]
5048 Commandeer a sysrq key to dump out Tree RCU's
5049 rcu_node tree with an eye towards determining
5050 why a new grace period has not yet started.
5052 rcutree.use_softirq= [KNL]
5053 If set to zero, move all RCU_SOFTIRQ processing to
5054 per-CPU rcuc kthreads. Defaults to a non-zero
5055 value, meaning that RCU_SOFTIRQ is used by default.
5056 Specify rcutree.use_softirq=0 to use rcuc kthreads.
5058 But note that CONFIG_PREEMPT_RT=y kernels disable
5059 this kernel boot parameter, forcibly setting it
5062 rcutree.enable_rcu_lazy= [KNL]
5063 To save power, batch RCU callbacks and flush after
5064 delay, memory pressure or callback list growing too
5067 rcuscale.gp_async= [KNL]
5068 Measure performance of asynchronous
5069 grace-period primitives such as call_rcu().
5071 rcuscale.gp_async_max= [KNL]
5072 Specify the maximum number of outstanding
5073 callbacks per writer thread. When a writer
5074 thread exceeds this limit, it invokes the
5075 corresponding flavor of rcu_barrier() to allow
5076 previously posted callbacks to drain.
5078 rcuscale.gp_exp= [KNL]
5079 Measure performance of expedited synchronous
5080 grace-period primitives.
5082 rcuscale.holdoff= [KNL]
5083 Set test-start holdoff period. The purpose of
5084 this parameter is to delay the start of the
5085 test until boot completes in order to avoid
5088 rcuscale.kfree_by_call_rcu= [KNL]
5089 In kernels built with CONFIG_RCU_LAZY=y, test
5090 call_rcu() instead of kfree_rcu().
5092 rcuscale.kfree_mult= [KNL]
5093 Instead of allocating an object of size kfree_obj,
5094 allocate one of kfree_mult * sizeof(kfree_obj).
5097 rcuscale.kfree_rcu_test= [KNL]
5098 Set to measure performance of kfree_rcu() flooding.
5100 rcuscale.kfree_rcu_test_double= [KNL]
5101 Test the double-argument variant of kfree_rcu().
5102 If this parameter has the same value as
5103 rcuscale.kfree_rcu_test_single, both the single-
5104 and double-argument variants are tested.
5106 rcuscale.kfree_rcu_test_single= [KNL]
5107 Test the single-argument variant of kfree_rcu().
5108 If this parameter has the same value as
5109 rcuscale.kfree_rcu_test_double, both the single-
5110 and double-argument variants are tested.
5112 rcuscale.kfree_nthreads= [KNL]
5113 The number of threads running loops of kfree_rcu().
5115 rcuscale.kfree_alloc_num= [KNL]
5116 Number of allocations and frees done in an iteration.
5118 rcuscale.kfree_loops= [KNL]
5119 Number of loops doing rcuscale.kfree_alloc_num number
5120 of allocations and frees.
5122 rcuscale.minruntime= [KNL]
5123 Set the minimum test run time in seconds. This
5124 does not affect the data-collection interval,
5125 but instead allows better measurement of things
5126 like CPU consumption.
5128 rcuscale.nreaders= [KNL]
5129 Set number of RCU readers. The value -1 selects
5130 N, where N is the number of CPUs. A value
5131 "n" less than -1 selects N-n+1, where N is again
5132 the number of CPUs. For example, -2 selects N
5133 (the number of CPUs), -3 selects N+1, and so on.
5134 A value of "n" less than or equal to -N selects
5137 rcuscale.nwriters= [KNL]
5138 Set number of RCU writers. The values operate
5139 the same as for rcuscale.nreaders.
5140 N, where N is the number of CPUs
5142 rcuscale.scale_type= [KNL]
5143 Specify the RCU implementation to test.
5145 rcuscale.shutdown= [KNL]
5146 Shut the system down after performance tests
5147 complete. This is useful for hands-off automated
5150 rcuscale.verbose= [KNL]
5151 Enable additional printk() statements.
5153 rcuscale.writer_holdoff= [KNL]
5154 Write-side holdoff between grace periods,
5155 in microseconds. The default of zero says
5158 rcuscale.writer_holdoff_jiffies= [KNL]
5159 Additional write-side holdoff between grace
5160 periods, but in jiffies. The default of zero
5163 rcutorture.fqs_duration= [KNL]
5164 Set duration of force_quiescent_state bursts
5167 rcutorture.fqs_holdoff= [KNL]
5168 Set holdoff time within force_quiescent_state bursts
5171 rcutorture.fqs_stutter= [KNL]
5172 Set wait time between force_quiescent_state bursts
5175 rcutorture.fwd_progress= [KNL]
5176 Specifies the number of kthreads to be used
5177 for RCU grace-period forward-progress testing
5178 for the types of RCU supporting this notion.
5179 Defaults to 1 kthread, values less than zero or
5180 greater than the number of CPUs cause the number
5183 rcutorture.fwd_progress_div= [KNL]
5184 Specify the fraction of a CPU-stall-warning
5185 period to do tight-loop forward-progress testing.
5187 rcutorture.fwd_progress_holdoff= [KNL]
5188 Number of seconds to wait between successive
5189 forward-progress tests.
5191 rcutorture.fwd_progress_need_resched= [KNL]
5192 Enclose cond_resched() calls within checks for
5193 need_resched() during tight-loop forward-progress
5196 rcutorture.gp_cond= [KNL]
5197 Use conditional/asynchronous update-side
5198 primitives, if available.
5200 rcutorture.gp_exp= [KNL]
5201 Use expedited update-side primitives, if available.
5203 rcutorture.gp_normal= [KNL]
5204 Use normal (non-expedited) asynchronous
5205 update-side primitives, if available.
5207 rcutorture.gp_sync= [KNL]
5208 Use normal (non-expedited) synchronous
5209 update-side primitives, if available. If all
5210 of rcutorture.gp_cond=, rcutorture.gp_exp=,
5211 rcutorture.gp_normal=, and rcutorture.gp_sync=
5212 are zero, rcutorture acts as if is interpreted
5213 they are all non-zero.
5215 rcutorture.irqreader= [KNL]
5216 Run RCU readers from irq handlers, or, more
5217 accurately, from a timer handler. Not all RCU
5218 flavors take kindly to this sort of thing.
5220 rcutorture.leakpointer= [KNL]
5221 Leak an RCU-protected pointer out of the reader.
5222 This can of course result in splats, and is
5223 intended to test the ability of things like
5224 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
5227 rcutorture.n_barrier_cbs= [KNL]
5228 Set callbacks/threads for rcu_barrier() testing.
5230 rcutorture.nfakewriters= [KNL]
5231 Set number of concurrent RCU writers. These just
5232 stress RCU, they don't participate in the actual
5233 test, hence the "fake".
5235 rcutorture.nocbs_nthreads= [KNL]
5236 Set number of RCU callback-offload togglers.
5237 Zero (the default) disables toggling.
5239 rcutorture.nocbs_toggle= [KNL]
5240 Set the delay in milliseconds between successive
5241 callback-offload toggling attempts.
5243 rcutorture.nreaders= [KNL]
5244 Set number of RCU readers. The value -1 selects
5245 N-1, where N is the number of CPUs. A value
5246 "n" less than -1 selects N-n-2, where N is again
5247 the number of CPUs. For example, -2 selects N
5248 (the number of CPUs), -3 selects N+1, and so on.
5250 rcutorture.object_debug= [KNL]
5251 Enable debug-object double-call_rcu() testing.
5253 rcutorture.onoff_holdoff= [KNL]
5254 Set time (s) after boot for CPU-hotplug testing.
5256 rcutorture.onoff_interval= [KNL]
5257 Set time (jiffies) between CPU-hotplug operations,
5258 or zero to disable CPU-hotplug testing.
5260 rcutorture.read_exit= [KNL]
5261 Set the number of read-then-exit kthreads used
5262 to test the interaction of RCU updaters and
5263 task-exit processing.
5265 rcutorture.read_exit_burst= [KNL]
5266 The number of times in a given read-then-exit
5267 episode that a set of read-then-exit kthreads
5270 rcutorture.read_exit_delay= [KNL]
5271 The delay, in seconds, between successive
5272 read-then-exit testing episodes.
5274 rcutorture.shuffle_interval= [KNL]
5275 Set task-shuffle interval (s). Shuffling tasks
5276 allows some CPUs to go into dyntick-idle mode
5277 during the rcutorture test.
5279 rcutorture.shutdown_secs= [KNL]
5280 Set time (s) after boot system shutdown. This
5281 is useful for hands-off automated testing.
5283 rcutorture.stall_cpu= [KNL]
5284 Duration of CPU stall (s) to test RCU CPU stall
5285 warnings, zero to disable.
5287 rcutorture.stall_cpu_block= [KNL]
5288 Sleep while stalling if set. This will result
5289 in warnings from preemptible RCU in addition to
5290 any other stall-related activity. Note that
5291 in kernels built with CONFIG_PREEMPTION=n and
5292 CONFIG_PREEMPT_COUNT=y, this parameter will
5293 cause the CPU to pass through a quiescent state.
5294 Given CONFIG_PREEMPTION=n, this will suppress
5295 RCU CPU stall warnings, but will instead result
5296 in scheduling-while-atomic splats.
5298 Use of this module parameter results in splats.
5301 rcutorture.stall_cpu_holdoff= [KNL]
5302 Time to wait (s) after boot before inducing stall.
5304 rcutorture.stall_cpu_irqsoff= [KNL]
5305 Disable interrupts while stalling if set.
5307 rcutorture.stall_gp_kthread= [KNL]
5308 Duration (s) of forced sleep within RCU
5309 grace-period kthread to test RCU CPU stall
5310 warnings, zero to disable. If both stall_cpu
5311 and stall_gp_kthread are specified, the
5312 kthread is starved first, then the CPU.
5314 rcutorture.stat_interval= [KNL]
5315 Time (s) between statistics printk()s.
5317 rcutorture.stutter= [KNL]
5318 Time (s) to stutter testing, for example, specifying
5319 five seconds causes the test to run for five seconds,
5320 wait for five seconds, and so on. This tests RCU's
5321 ability to transition abruptly to and from idle.
5323 rcutorture.test_boost= [KNL]
5324 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5325 "Maybe" means test if the RCU implementation
5326 under test support RCU priority boosting.
5328 rcutorture.test_boost_duration= [KNL]
5329 Duration (s) of each individual boost test.
5331 rcutorture.test_boost_interval= [KNL]
5332 Interval (s) between each boost test.
5334 rcutorture.test_no_idle_hz= [KNL]
5335 Test RCU's dyntick-idle handling. See also the
5336 rcutorture.shuffle_interval parameter.
5338 rcutorture.torture_type= [KNL]
5339 Specify the RCU implementation to test.
5341 rcutorture.verbose= [KNL]
5342 Enable additional printk() statements.
5344 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5345 Dump ftrace buffer after reporting RCU CPU
5348 rcupdate.rcu_cpu_stall_notifiers= [KNL]
5349 Provide RCU CPU stall notifiers, but see the
5350 warnings in the RCU_CPU_STALL_NOTIFIER Kconfig
5351 option's help text. TL;DR: You almost certainly
5352 do not want rcupdate.rcu_cpu_stall_notifiers.
5354 rcupdate.rcu_cpu_stall_suppress= [KNL]
5355 Suppress RCU CPU stall warning messages.
5357 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5358 Suppress RCU CPU stall warning messages and
5359 rcutorture writer stall warnings that occur
5360 during early boot, that is, during the time
5361 before the init task is spawned.
5363 rcupdate.rcu_cpu_stall_timeout= [KNL]
5364 Set timeout for RCU CPU stall warning messages.
5365 The value is in seconds and the maximum allowed
5366 value is 300 seconds.
5368 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5369 Set timeout for expedited RCU CPU stall warning
5370 messages. The value is in milliseconds
5371 and the maximum allowed value is 21000
5372 milliseconds. Please note that this value is
5373 adjusted to an arch timer tick resolution.
5374 Setting this to zero causes the value from
5375 rcupdate.rcu_cpu_stall_timeout to be used (after
5376 conversion from seconds to milliseconds).
5378 rcupdate.rcu_cpu_stall_cputime= [KNL]
5379 Provide statistics on the cputime and count of
5380 interrupts and tasks during the sampling period. For
5381 multiple continuous RCU stalls, all sampling periods
5382 begin at half of the first RCU stall timeout.
5384 rcupdate.rcu_exp_stall_task_details= [KNL]
5385 Print stack dumps of any tasks blocking the
5386 current expedited RCU grace period during an
5387 expedited RCU CPU stall warning.
5389 rcupdate.rcu_expedited= [KNL]
5390 Use expedited grace-period primitives, for
5391 example, synchronize_rcu_expedited() instead
5392 of synchronize_rcu(). This reduces latency,
5393 but can increase CPU utilization, degrade
5394 real-time latency, and degrade energy efficiency.
5395 No effect on CONFIG_TINY_RCU kernels.
5397 rcupdate.rcu_normal= [KNL]
5398 Use only normal grace-period primitives,
5399 for example, synchronize_rcu() instead of
5400 synchronize_rcu_expedited(). This improves
5401 real-time latency, CPU utilization, and
5402 energy efficiency, but can expose users to
5403 increased grace-period latency. This parameter
5404 overrides rcupdate.rcu_expedited. No effect on
5405 CONFIG_TINY_RCU kernels.
5407 rcupdate.rcu_normal_after_boot= [KNL]
5408 Once boot has completed (that is, after
5409 rcu_end_inkernel_boot() has been invoked), use
5410 only normal grace-period primitives. No effect
5411 on CONFIG_TINY_RCU kernels.
5413 But note that CONFIG_PREEMPT_RT=y kernels enables
5414 this kernel boot parameter, forcibly setting
5415 it to the value one, that is, converting any
5416 post-boot attempt at an expedited RCU grace
5417 period to instead use normal non-expedited
5418 grace-period processing.
5420 rcupdate.rcu_task_collapse_lim= [KNL]
5421 Set the maximum number of callbacks present
5422 at the beginning of a grace period that allows
5423 the RCU Tasks flavors to collapse back to using
5424 a single callback queue. This switching only
5425 occurs when rcupdate.rcu_task_enqueue_lim is
5426 set to the default value of -1.
5428 rcupdate.rcu_task_contend_lim= [KNL]
5429 Set the minimum number of callback-queuing-time
5430 lock-contention events per jiffy required to
5431 cause the RCU Tasks flavors to switch to per-CPU
5432 callback queuing. This switching only occurs
5433 when rcupdate.rcu_task_enqueue_lim is set to
5434 the default value of -1.
5436 rcupdate.rcu_task_enqueue_lim= [KNL]
5437 Set the number of callback queues to use for the
5438 RCU Tasks family of RCU flavors. The default
5439 of -1 allows this to be automatically (and
5440 dynamically) adjusted. This parameter is intended
5443 rcupdate.rcu_task_ipi_delay= [KNL]
5444 Set time in jiffies during which RCU tasks will
5445 avoid sending IPIs, starting with the beginning
5446 of a given grace period. Setting a large
5447 number avoids disturbing real-time workloads,
5448 but lengthens grace periods.
5450 rcupdate.rcu_task_lazy_lim= [KNL]
5451 Number of callbacks on a given CPU that will
5452 cancel laziness on that CPU. Use -1 to disable
5453 cancellation of laziness, but be advised that
5454 doing so increases the danger of OOM due to
5457 rcupdate.rcu_task_stall_info= [KNL]
5458 Set initial timeout in jiffies for RCU task stall
5459 informational messages, which give some indication
5460 of the problem for those not patient enough to
5461 wait for ten minutes. Informational messages are
5462 only printed prior to the stall-warning message
5463 for a given grace period. Disable with a value
5464 less than or equal to zero. Defaults to ten
5465 seconds. A change in value does not take effect
5466 until the beginning of the next grace period.
5468 rcupdate.rcu_task_stall_info_mult= [KNL]
5469 Multiplier for time interval between successive
5470 RCU task stall informational messages for a given
5471 RCU tasks grace period. This value is clamped
5472 to one through ten, inclusive. It defaults to
5473 the value three, so that the first informational
5474 message is printed 10 seconds into the grace
5475 period, the second at 40 seconds, the third at
5476 160 seconds, and then the stall warning at 600
5477 seconds would prevent a fourth at 640 seconds.
5479 rcupdate.rcu_task_stall_timeout= [KNL]
5480 Set timeout in jiffies for RCU task stall
5481 warning messages. Disable with a value less
5482 than or equal to zero. Defaults to ten minutes.
5483 A change in value does not take effect until
5484 the beginning of the next grace period.
5486 rcupdate.rcu_tasks_lazy_ms= [KNL]
5487 Set timeout in milliseconds RCU Tasks asynchronous
5488 callback batching for call_rcu_tasks().
5489 A negative value will take the default. A value
5490 of zero will disable batching. Batching is
5491 always disabled for synchronize_rcu_tasks().
5493 rcupdate.rcu_tasks_rude_lazy_ms= [KNL]
5494 Set timeout in milliseconds RCU Tasks
5495 Rude asynchronous callback batching for
5496 call_rcu_tasks_rude(). A negative value
5497 will take the default. A value of zero will
5498 disable batching. Batching is always disabled
5499 for synchronize_rcu_tasks_rude().
5501 rcupdate.rcu_tasks_trace_lazy_ms= [KNL]
5502 Set timeout in milliseconds RCU Tasks
5503 Trace asynchronous callback batching for
5504 call_rcu_tasks_trace(). A negative value
5505 will take the default. A value of zero will
5506 disable batching. Batching is always disabled
5507 for synchronize_rcu_tasks_trace().
5509 rcupdate.rcu_self_test= [KNL]
5510 Run the RCU early boot self tests
5514 Run specified binary instead of /init from the ramdisk,
5515 used for early userspace startup. See initrd.
5518 force - Override the decision by the kernel to hide the
5519 advertisement of RDRAND support (this affects
5520 certain AMD processors because of buggy BIOS
5521 support, specifically around the suspend/resume
5525 Turn on/off individual RDT features. List is:
5526 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5528 E.g. to turn on cmt and turn off mba use:
5532 Format (x86 or x86_64):
5533 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5535 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5537 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5538 (prefix with 'panic_' to set mode for panic
5540 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5541 reboot_force is either force or not specified,
5542 reboot_cpu is s[mp]#### with #### being the processor
5543 to be used for rebooting.
5545 refscale.holdoff= [KNL]
5546 Set test-start holdoff period. The purpose of
5547 this parameter is to delay the start of the
5548 test until boot completes in order to avoid
5551 refscale.lookup_instances= [KNL]
5552 Number of data elements to use for the forms of
5553 SLAB_TYPESAFE_BY_RCU testing. A negative number
5554 is negated and multiplied by nr_cpu_ids, while
5555 zero specifies nr_cpu_ids.
5557 refscale.loops= [KNL]
5558 Set the number of loops over the synchronization
5559 primitive under test. Increasing this number
5560 reduces noise due to loop start/end overhead,
5561 but the default has already reduced the per-pass
5562 noise to a handful of picoseconds on ca. 2020
5565 refscale.nreaders= [KNL]
5566 Set number of readers. The default value of -1
5567 selects N, where N is roughly 75% of the number
5568 of CPUs. A value of zero is an interesting choice.
5570 refscale.nruns= [KNL]
5571 Set number of runs, each of which is dumped onto
5574 refscale.readdelay= [KNL]
5575 Set the read-side critical-section duration,
5576 measured in microseconds.
5578 refscale.scale_type= [KNL]
5579 Specify the read-protection implementation to test.
5581 refscale.shutdown= [KNL]
5582 Shut down the system at the end of the performance
5583 test. This defaults to 1 (shut it down) when
5584 refscale is built into the kernel and to 0 (leave
5585 it running) when refscale is built as a module.
5587 refscale.verbose= [KNL]
5588 Enable additional printk() statements.
5590 refscale.verbose_batched= [KNL]
5591 Batch the additional printk() statements. If zero
5592 (the default) or negative, print everything. Otherwise,
5593 print every Nth verbose statement, where N is the value
5596 regulator_ignore_unused
5598 Prevents regulator framework from disabling regulators
5599 that are unused, due no driver claiming them. This may
5600 be useful for debug and development, but should not be
5601 needed on a platform with proper driver support.
5604 [KNL, SMP] Set scheduler's default relax_domain_level.
5605 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5607 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5608 Format: <base1>,<size1>[,<base2>,<size2>,...]
5609 Reserve I/O ports or memory so the kernel won't use
5610 them. If <base> is less than 0x10000, the region
5611 is assumed to be I/O ports; otherwise it is memory.
5613 reservetop= [X86-32,EARLY]
5615 Reserves a hole at the top of the kernel virtual
5618 reset_devices [KNL] Force drivers to reset the underlying device
5619 during initialization.
5622 Specify the partition device for software suspend
5624 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5626 resume_offset= [SWSUSP]
5627 Specify the offset from the beginning of the partition
5628 given by "resume=" at which the swap header is located,
5629 in <PAGE_SIZE> units (needed only for swap files).
5630 See Documentation/power/swsusp-and-swap-files.rst
5632 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5633 read the resume files
5635 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5636 Useful for devices that are detected asynchronously
5637 (e.g. USB and MMC devices).
5639 retain_initrd [RAM] Keep initrd memory after extraction. After boot, it will
5640 be accessible via /sys/firmware/initrd.
5642 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5643 Speculative Code Execution with Return Instructions)
5646 AMD-based UNRET and IBPB mitigations alone do not stop
5647 sibling threads from influencing the predictions of other
5648 sibling threads. For that reason, STIBP is used on pro-
5649 cessors that support it, and mitigate SMT on processors
5653 auto - automatically select a migitation
5654 auto,nosmt - automatically select a mitigation,
5655 disabling SMT if necessary for
5656 the full mitigation (only on Zen1
5657 and older without STIBP).
5658 ibpb - On AMD, mitigate short speculation
5659 windows on basic block boundaries too.
5660 Safe, highest perf impact. It also
5661 enables STIBP if present. Not suitable
5663 ibpb,nosmt - Like "ibpb" above but will disable SMT
5664 when STIBP is not available. This is
5665 the alternative for systems which do not
5667 unret - Force enable untrained return thunks,
5668 only effective on AMD f15h-f17h based
5670 unret,nosmt - Like unret, but will disable SMT when STIBP
5671 is not available. This is the alternative for
5672 systems which do not have STIBP.
5674 Selecting 'auto' will choose a mitigation method at run
5675 time according to the CPU.
5677 Not specifying this option is equivalent to retbleed=auto.
5679 rfkill.default_state=
5680 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5681 etc. communication is blocked by default.
5684 rfkill.master_switch_mode=
5685 0 The "airplane mode" button does nothing.
5686 1 The "airplane mode" button toggles between everything
5687 blocked and the previous configuration.
5688 2 The "airplane mode" button toggles between everything
5689 blocked and everything unblocked.
5691 rhash_entries= [KNL,NET]
5692 Set number of hash buckets for route cache
5695 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5698 riscv_isa_fallback [RISCV,EARLY]
5699 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit
5700 falling back to detecting extension support by parsing
5701 "riscv,isa" property on devicetree systems when the
5702 replacement properties are not found. See the Kconfig
5703 entry for RISCV_ISA_FALLBACK.
5705 ro [KNL] Mount root device read-only on boot
5708 on Mark read-only kernel memory as read-only (default).
5709 off Leave read-only kernel memory writable for debugging.
5710 full Mark read-only kernel memory and aliases as read-only
5715 Enable the uart passthrough on the designated usb port
5716 on Rockchip SoCs. When active, the signals of the
5717 debug-uart get routed to the D+ and D- pins of the usb
5718 port and the regular usb controller gets disabled.
5720 root= [KNL] Root filesystem
5721 Usually this a a block device specifier of some kind,
5722 see the early_lookup_bdev comment in
5723 block/early-lookup.c for details.
5724 Alternatively this can be "ram" for the legacy initial
5725 ramdisk, "nfs" and "cifs" for root on a network file
5726 system, or "mtd" and "ubi" for mounting from raw flash.
5728 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5729 mount the root filesystem
5731 rootflags= [KNL] Set root filesystem mount option string
5733 rootfstype= [KNL] Set root filesystem type
5735 rootwait [KNL] Wait (indefinitely) for root device to show up.
5736 Useful for devices that are detected asynchronously
5737 (e.g. USB and MMC devices).
5739 rootwait= [KNL] Maximum time (in seconds) to wait for root device
5740 to show up before attempting to mount the root
5743 rproc_mem=nn[KMG][@address]
5744 [KNL,ARM,CMA] Remoteproc physical memory block.
5745 Memory area to be used by remote processor image,
5748 rw [KNL] Mount root device read-write on boot
5750 S [KNL] Run init in single mode
5752 s390_iommu= [HW,S390]
5753 Set s390 IOTLB flushing mode
5755 With strict flushing every unmap operation will result
5756 in an IOTLB flush. Default is lazy flushing before
5757 reuse, which is faster. Deprecated, equivalent to
5760 s390_iommu_aperture= [KNL,S390]
5761 Specifies the size of the per device DMA address space
5762 accessible through the DMA and IOMMU APIs as a decimal
5763 factor of the size of main memory.
5764 The default is 1 meaning that one can concurrently use
5765 as many DMA addresses as physical memory is installed,
5766 if supported by hardware, and thus map all of memory
5767 once. With a value of 2 one can map all of memory twice
5768 and so on. As a special case a factor of 0 imposes no
5769 restrictions other than those given by hardware at the
5770 cost of significant additional memory use for tables.
5773 See drivers/net/irda/sa1100_ir.c.
5775 sched_verbose [KNL,EARLY] Enables verbose scheduler debug messages.
5777 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5778 Allowed values are enable and disable. This feature
5779 incurs a small amount of overhead in the scheduler
5780 but is useful for debugging and performance tuning.
5782 sched_thermal_decay_shift=
5783 [KNL, SMP] Set a decay shift for scheduler thermal
5784 pressure signal. Thermal pressure signal follows the
5785 default decay period of other scheduler pelt
5786 signals(usually 32 ms but configurable). Setting
5787 sched_thermal_decay_shift will left shift the decay
5788 period for the thermal pressure signal by the shift
5790 i.e. with the default pelt decay period of 32 ms
5791 sched_thermal_decay_shift thermal pressure decay pr
5795 Format: integer between 0 and 10
5798 scftorture.holdoff= [KNL]
5799 Number of seconds to hold off before starting
5800 test. Defaults to zero for module insertion and
5801 to 10 seconds for built-in smp_call_function()
5804 scftorture.longwait= [KNL]
5805 Request ridiculously long waits randomly selected
5806 up to the chosen limit in seconds. Zero (the
5807 default) disables this feature. Please note
5808 that requesting even small non-zero numbers of
5809 seconds can result in RCU CPU stall warnings,
5810 softlockup complaints, and so on.
5812 scftorture.nthreads= [KNL]
5813 Number of kthreads to spawn to invoke the
5814 smp_call_function() family of functions.
5815 The default of -1 specifies a number of kthreads
5816 equal to the number of CPUs.
5818 scftorture.onoff_holdoff= [KNL]
5819 Number seconds to wait after the start of the
5820 test before initiating CPU-hotplug operations.
5822 scftorture.onoff_interval= [KNL]
5823 Number seconds to wait between successive
5824 CPU-hotplug operations. Specifying zero (which
5825 is the default) disables CPU-hotplug operations.
5827 scftorture.shutdown_secs= [KNL]
5828 The number of seconds following the start of the
5829 test after which to shut down the system. The
5830 default of zero avoids shutting down the system.
5831 Non-zero values are useful for automated tests.
5833 scftorture.stat_interval= [KNL]
5834 The number of seconds between outputting the
5835 current test statistics to the console. A value
5836 of zero disables statistics output.
5838 scftorture.stutter_cpus= [KNL]
5839 The number of jiffies to wait between each change
5840 to the set of CPUs under test.
5842 scftorture.use_cpus_read_lock= [KNL]
5843 Use use_cpus_read_lock() instead of the default
5844 preempt_disable() to disable CPU hotplug
5845 while invoking one of the smp_call_function*()
5848 scftorture.verbose= [KNL]
5849 Enable additional printk() statements.
5851 scftorture.weight_single= [KNL]
5852 The probability weighting to use for the
5853 smp_call_function_single() function with a zero
5854 "wait" parameter. A value of -1 selects the
5855 default if all other weights are -1. However,
5856 if at least one weight has some other value, a
5857 value of -1 will instead select a weight of zero.
5859 scftorture.weight_single_wait= [KNL]
5860 The probability weighting to use for the
5861 smp_call_function_single() function with a
5862 non-zero "wait" parameter. See weight_single.
5864 scftorture.weight_many= [KNL]
5865 The probability weighting to use for the
5866 smp_call_function_many() function with a zero
5867 "wait" parameter. See weight_single.
5868 Note well that setting a high probability for
5869 this weighting can place serious IPI load
5872 scftorture.weight_many_wait= [KNL]
5873 The probability weighting to use for the
5874 smp_call_function_many() function with a
5875 non-zero "wait" parameter. See weight_single
5878 scftorture.weight_all= [KNL]
5879 The probability weighting to use for the
5880 smp_call_function_all() function with a zero
5881 "wait" parameter. See weight_single and
5884 scftorture.weight_all_wait= [KNL]
5885 The probability weighting to use for the
5886 smp_call_function_all() function with a
5887 non-zero "wait" parameter. See weight_single
5890 skew_tick= [KNL,EARLY] Offset the periodic timer tick per cpu to mitigate
5891 xtime_lock contention on larger systems, and/or RCU lock
5892 contention on all systems with CONFIG_MAXSMP set.
5893 Format: { "0" | "1" }
5894 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5896 Note: increases power consumption, thus should only be
5897 enabled if running jitter sensitive (HPC/RT) workloads.
5899 security= [SECURITY] Choose a legacy "major" security module to
5900 enable at boot. This has been deprecated by the
5903 selinux= [SELINUX] Disable or enable SELinux at boot time.
5904 Format: { "0" | "1" }
5905 See security/selinux/Kconfig help text.
5910 serialnumber [BUGS=X86-32]
5912 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
5915 Maximal number of shapers.
5917 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
5918 Limit apic dumping. The parameter defines the maximal
5919 number of local apics being dumped. Also it is possible
5920 to set it to "all" by meaning -- no limit here.
5921 Format: { 1 (default) | 2 | ... | all }.
5922 The parameter valid if only apic=debug or
5923 apic=verbose is specified.
5924 Example: apic=debug show_lapic=all
5929 slab_debug[=options[,slabs][;[options[,slabs]]...] [MM]
5930 Enabling slab_debug allows one to determine the
5931 culprit if slab objects become corrupted. Enabling
5932 slab_debug can create guard zones around objects and
5933 may poison objects when not in use. Also tracks the
5934 last alloc / free. For more information see
5935 Documentation/mm/slub.rst.
5936 (slub_debug legacy name also accepted for now)
5938 slab_max_order= [MM]
5939 Determines the maximum allowed order for slabs.
5940 A high setting may cause OOMs due to memory
5941 fragmentation. For more information see
5942 Documentation/mm/slub.rst.
5943 (slub_max_order legacy name also accepted for now)
5946 Enable merging of slabs with similar size when the
5947 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5948 (slub_merge legacy name also accepted for now)
5950 slab_min_objects= [MM]
5951 The minimum number of objects per slab. SLUB will
5952 increase the slab order up to slab_max_order to
5953 generate a sufficiently large slab able to contain
5954 the number of objects indicated. The higher the number
5955 of objects the smaller the overhead of tracking slabs
5956 and the less frequently locks need to be acquired.
5957 For more information see Documentation/mm/slub.rst.
5958 (slub_min_objects legacy name also accepted for now)
5960 slab_min_order= [MM]
5961 Determines the minimum page order for slabs. Must be
5962 lower or equal to slab_max_order. For more information see
5963 Documentation/mm/slub.rst.
5964 (slub_min_order legacy name also accepted for now)
5967 Disable merging of slabs with similar size. May be
5968 necessary if there is some reason to distinguish
5969 allocs to different slabs, especially in hardened
5970 environments where the risk of heap overflows and
5971 layout control by attackers can usually be
5972 frustrated by disabling merging. This will reduce
5973 most of the exposure of a heap attack to a single
5974 cache (risks via metadata attacks are mostly
5975 unchanged). Debug options disable merging on their
5977 For more information see Documentation/mm/slub.rst.
5978 (slub_nomerge legacy name also accepted for now)
5983 Format: <io1>[,<io2>[,...,<io8>]]
5985 smp.csd_lock_timeout= [KNL]
5986 Specify the period of time in milliseconds
5987 that smp_call_function() and friends will wait
5988 for a CPU to release the CSD lock. This is
5989 useful when diagnosing bugs involving CPUs
5990 disabling interrupts for extended periods
5991 of time. Defaults to 5,000 milliseconds, and
5992 setting a value of zero disables this feature.
5993 This feature may be more efficiently disabled
5994 using the csdlock_debug- kernel parameter.
5996 smp.panic_on_ipistall= [KNL]
5997 If a csd_lock_timeout extends for more than
5998 the specified number of milliseconds, panic the
5999 system. By default, let CSD-lock acquisition
6000 take as long as they take. Specifying 300,000
6001 for this value provides a 5-minute timeout.
6003 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
6004 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
6005 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
6006 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
6007 smsc-ircc2.ircc_irq= [HW] IRQ line
6008 smsc-ircc2.ircc_dma= [HW] DMA channel
6009 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
6010 0: Toshiba Satellite 1800 (GP data pin select)
6011 1: Fast pin select (default)
6014 smt= [KNL,MIPS,S390,EARLY] Set the maximum number of threads
6015 (logical CPUs) to use per physical CPU on systems
6016 capable of symmetric multithreading (SMT). Will
6017 be capped to the actual hardware limit.
6019 Default: -1 (no limit)
6022 [KNL] Should the soft-lockup detector generate panics.
6025 A value of 1 instructs the soft-lockup detector
6026 to panic the machine when a soft-lockup occurs. It is
6027 also controlled by the kernel.softlockup_panic sysctl
6028 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
6029 respective build-time switch to that functionality.
6031 softlockup_all_cpu_backtrace=
6032 [KNL] Should the soft-lockup detector generate
6033 backtraces on all cpus.
6036 sonypi.*= [HW] Sony Programmable I/O Control Device driver
6037 See Documentation/admin-guide/laptops/sonypi.rst
6039 spectre_v2= [X86,EARLY] Control mitigation of Spectre variant 2
6040 (indirect branch speculation) vulnerability.
6041 The default operation protects the kernel from
6044 on - unconditionally enable, implies
6046 off - unconditionally disable, implies
6048 auto - kernel detects whether your CPU model is
6051 Selecting 'on' will, and 'auto' may, choose a
6052 mitigation method at run time according to the
6053 CPU, the available microcode, the setting of the
6054 CONFIG_MITIGATION_RETPOLINE configuration option,
6055 and the compiler with which the kernel was built.
6057 Selecting 'on' will also enable the mitigation
6058 against user space to user space task attacks.
6060 Selecting 'off' will disable both the kernel and
6061 the user space protections.
6063 Specific mitigations can also be selected manually:
6065 retpoline - replace indirect branches
6066 retpoline,generic - Retpolines
6067 retpoline,lfence - LFENCE; indirect branch
6068 retpoline,amd - alias for retpoline,lfence
6069 eibrs - Enhanced/Auto IBRS
6070 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines
6071 eibrs,lfence - Enhanced/Auto IBRS + LFENCE
6072 ibrs - use IBRS to protect kernel
6074 Not specifying this option is equivalent to
6078 [X86] Control mitigation of Spectre variant 2
6079 (indirect branch speculation) vulnerability between
6082 on - Unconditionally enable mitigations. Is
6083 enforced by spectre_v2=on
6085 off - Unconditionally disable mitigations. Is
6086 enforced by spectre_v2=off
6088 prctl - Indirect branch speculation is enabled,
6089 but mitigation can be enabled via prctl
6090 per thread. The mitigation control state
6091 is inherited on fork.
6094 - Like "prctl" above, but only STIBP is
6095 controlled per thread. IBPB is issued
6096 always when switching between different user
6100 - Same as "prctl" above, but all seccomp
6101 threads will enable the mitigation unless
6102 they explicitly opt out.
6105 - Like "seccomp" above, but only STIBP is
6106 controlled per thread. IBPB is issued
6107 always when switching between different
6108 user space processes.
6110 auto - Kernel selects the mitigation depending on
6111 the available CPU features and vulnerability.
6113 Default mitigation: "prctl"
6115 Not specifying this option is equivalent to
6116 spectre_v2_user=auto.
6118 spec_rstack_overflow=
6119 [X86,EARLY] Control RAS overflow mitigation on AMD Zen CPUs
6121 off - Disable mitigation
6122 microcode - Enable microcode mitigation only
6123 safe-ret - Enable sw-only safe RET mitigation (default)
6124 ibpb - Enable mitigation by issuing IBPB on
6126 ibpb-vmexit - Issue IBPB only on VMEXIT
6127 (cloud-specific mitigation)
6129 spec_store_bypass_disable=
6130 [HW,EARLY] Control Speculative Store Bypass (SSB) Disable mitigation
6131 (Speculative Store Bypass vulnerability)
6133 Certain CPUs are vulnerable to an exploit against a
6134 a common industry wide performance optimization known
6135 as "Speculative Store Bypass" in which recent stores
6136 to the same memory location may not be observed by
6137 later loads during speculative execution. The idea
6138 is that such stores are unlikely and that they can
6139 be detected prior to instruction retirement at the
6140 end of a particular speculation execution window.
6142 In vulnerable processors, the speculatively forwarded
6143 store can be used in a cache side channel attack, for
6144 example to read memory to which the attacker does not
6145 directly have access (e.g. inside sandboxed code).
6147 This parameter controls whether the Speculative Store
6148 Bypass optimization is used.
6150 On x86 the options are:
6152 on - Unconditionally disable Speculative Store Bypass
6153 off - Unconditionally enable Speculative Store Bypass
6154 auto - Kernel detects whether the CPU model contains an
6155 implementation of Speculative Store Bypass and
6156 picks the most appropriate mitigation. If the
6157 CPU is not vulnerable, "off" is selected. If the
6158 CPU is vulnerable the default mitigation is
6159 architecture and Kconfig dependent. See below.
6160 prctl - Control Speculative Store Bypass per thread
6161 via prctl. Speculative Store Bypass is enabled
6162 for a process by default. The state of the control
6163 is inherited on fork.
6164 seccomp - Same as "prctl" above, but all seccomp threads
6165 will disable SSB unless they explicitly opt out.
6167 Default mitigations:
6170 On powerpc the options are:
6172 on,auto - On Power8 and Power9 insert a store-forwarding
6173 barrier on kernel entry and exit. On Power7
6174 perform a software flush on kernel entry and
6178 Not specifying this option is equivalent to
6179 spec_store_bypass_disable=auto.
6181 spia_io_base= [HW,MTD]
6187 [X86] Enable split lock detection or bus lock detection
6189 When enabled (and if hardware support is present), atomic
6190 instructions that access data across cache line
6191 boundaries will result in an alignment check exception
6192 for split lock detection or a debug exception for
6197 warn - the kernel will emit rate-limited warnings
6198 about applications triggering the #AC
6199 exception or the #DB exception. This mode is
6200 the default on CPUs that support split lock
6201 detection or bus lock detection. Default
6202 behavior is by #AC if both features are
6203 enabled in hardware.
6205 fatal - the kernel will send SIGBUS to applications
6206 that trigger the #AC exception or the #DB
6207 exception. Default behavior is by #AC if
6208 both features are enabled in hardware.
6211 Set system wide rate limit to N bus locks
6212 per second for bus lock detection.
6215 N/A for split lock detection.
6218 If an #AC exception is hit in the kernel or in
6219 firmware (i.e. not while executing in user mode)
6220 the kernel will oops in either "warn" or "fatal"
6223 #DB exception for bus lock is triggered only when
6226 srbds= [X86,INTEL,EARLY]
6227 Control the Special Register Buffer Data Sampling
6230 Certain CPUs are vulnerable to an MDS-like
6231 exploit which can leak bits from the random
6234 By default, this issue is mitigated by
6235 microcode. However, the microcode fix can cause
6236 the RDRAND and RDSEED instructions to become
6237 much slower. Among other effects, this will
6238 result in reduced throughput from /dev/urandom.
6240 The microcode mitigation can be disabled with
6241 the following option:
6243 off: Disable mitigation and remove
6244 performance impact to RDRAND and RDSEED
6246 srcutree.big_cpu_lim [KNL]
6247 Specifies the number of CPUs constituting a
6248 large system, such that srcu_struct structures
6249 should immediately allocate an srcu_node array.
6250 This kernel-boot parameter defaults to 128,
6251 but takes effect only when the low-order four
6252 bits of srcutree.convert_to_big is equal to 3
6255 srcutree.convert_to_big [KNL]
6256 Specifies under what conditions an SRCU tree
6257 srcu_struct structure will be converted to big
6258 form, that is, with an rcu_node tree:
6261 1: At init_srcu_struct() time.
6262 2: When rcutorture decides to.
6263 3: Decide at boot time (default).
6264 0x1X: Above plus if high contention.
6266 Either way, the srcu_node tree will be sized based
6267 on the actual runtime number of CPUs (nr_cpu_ids)
6268 instead of the compile-time CONFIG_NR_CPUS.
6270 srcutree.counter_wrap_check [KNL]
6271 Specifies how frequently to check for
6272 grace-period sequence counter wrap for the
6273 srcu_data structure's ->srcu_gp_seq_needed field.
6274 The greater the number of bits set in this kernel
6275 parameter, the less frequently counter wrap will
6276 be checked for. Note that the bottom two bits
6279 srcutree.exp_holdoff [KNL]
6280 Specifies how many nanoseconds must elapse
6281 since the end of the last SRCU grace period for
6282 a given srcu_struct until the next normal SRCU
6283 grace period will be considered for automatic
6284 expediting. Set to zero to disable automatic
6287 srcutree.srcu_max_nodelay [KNL]
6288 Specifies the number of no-delay instances
6289 per jiffy for which the SRCU grace period
6290 worker thread will be rescheduled with zero
6291 delay. Beyond this limit, worker thread will
6292 be rescheduled with a sleep delay of one jiffy.
6294 srcutree.srcu_max_nodelay_phase [KNL]
6295 Specifies the per-grace-period phase, number of
6296 non-sleeping polls of readers. Beyond this limit,
6297 grace period worker thread will be rescheduled
6298 with a sleep delay of one jiffy, between each
6299 rescan of the readers, for a grace period phase.
6301 srcutree.srcu_retry_check_delay [KNL]
6302 Specifies number of microseconds of non-sleeping
6303 delay between each non-sleeping poll of readers.
6305 srcutree.small_contention_lim [KNL]
6306 Specifies the number of update-side contention
6307 events per jiffy will be tolerated before
6308 initiating a conversion of an srcu_struct
6309 structure to big form. Note that the value of
6310 srcutree.convert_to_big must have the 0x10 bit
6311 set for contention-based conversions to occur.
6313 ssbd= [ARM64,HW,EARLY]
6314 Speculative Store Bypass Disable control
6316 On CPUs that are vulnerable to the Speculative
6317 Store Bypass vulnerability and offer a
6318 firmware based mitigation, this parameter
6319 indicates how the mitigation should be used:
6321 force-on: Unconditionally enable mitigation for
6322 for both kernel and userspace
6323 force-off: Unconditionally disable mitigation for
6324 for both kernel and userspace
6325 kernel: Always enable mitigation in the
6326 kernel, and offer a prctl interface
6327 to allow userspace to register its
6328 interest in being mitigated too.
6330 stack_guard_gap= [MM]
6331 override the default stack gap protection. The value
6332 is in page units and it defines how many pages prior
6333 to (for stacks growing down) resp. after (for stacks
6334 growing up) the main stack are reserved for no other
6335 mapping. Default value is 256 pages.
6337 stack_depot_disable= [KNL,EARLY]
6338 Setting this to true through kernel command line will
6339 disable the stack depot thereby saving the static memory
6340 consumed by the stack hash table. By default this is set
6344 Enabled the stack tracer on boot up.
6346 stacktrace_filter=[function-list]
6347 [FTRACE] Limit the functions that the stack tracer
6348 will trace at boot up. function-list is a comma-separated
6349 list of functions. This list can be changed at run
6350 time by the stack_trace_filter file in the debugfs
6351 tracing directory. Note, this enables stack tracing
6352 and the stacktrace above is not needed.
6356 Set the STI (builtin display/keyboard on the HP-PARISC
6357 machines) console (graphic card) which should be used
6358 as the initial boot-console.
6359 See also comment in drivers/video/console/sticore.c.
6362 See comment in drivers/video/console/sticore.c.
6365 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
6370 Enable or disable strict sigaltstack size checks
6371 against the required signal frame size which
6372 depends on the supported FPU features. This can
6373 be used to filter out binaries which have
6374 not yet been made aware of AT_MINSIGSTKSZ.
6376 stress_hpt [PPC,EARLY]
6377 Limits the number of kernel HPT entries in the hash
6378 page table to increase the rate of hash page table
6379 faults on kernel addresses.
6381 stress_slb [PPC,EARLY]
6382 Limits the number of kernel SLB entries, and flushes
6383 them frequently to increase the rate of SLB faults
6384 on kernel addresses.
6386 sunrpc.min_resvport=
6387 sunrpc.max_resvport=
6389 SunRPC servers often require that client requests
6390 originate from a privileged port (i.e. a port in the
6391 range 0 < portnr < 1024).
6392 An administrator who wishes to reserve some of these
6393 ports for other uses may adjust the range that the
6394 kernel's sunrpc client considers to be privileged
6395 using these two parameters to set the minimum and
6396 maximum port values.
6398 sunrpc.svc_rpc_per_connection_limit=
6400 Limit the number of requests that the server will
6401 process in parallel from a single connection.
6402 The default value is 0 (no limit).
6406 Control how the NFS server code allocates CPUs to
6407 service thread pools. Depending on how many NICs
6408 you have and where their interrupts are bound, this
6409 option will affect which CPUs will do NFS serving.
6410 Note: this parameter cannot be changed while the
6411 NFS server is running.
6413 auto the server chooses an appropriate mode
6414 automatically using heuristics
6415 global a single global pool contains all CPUs
6416 percpu one pool for each CPU
6417 pernode one pool for each NUMA node (equivalent
6418 to global on non-NUMA machines)
6420 sunrpc.tcp_slot_table_entries=
6421 sunrpc.udp_slot_table_entries=
6423 Sets the upper limit on the number of simultaneous
6424 RPC calls that can be sent from the client to a
6425 server. Increasing these values may allow you to
6426 improve throughput, but will also increase the
6427 amount of memory reserved for use by the client.
6429 suspend.pm_test_delay=
6431 Sets the number of seconds to remain in a suspend test
6432 mode before resuming the system (see
6433 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6434 is set. Default value is 5.
6437 Format: { on | off | y | n | 1 | 0 }
6438 This parameter controls use of the Protected
6439 Execution Facility on pSeries.
6441 swiotlb= [ARM,IA-64,PPC,MIPS,X86,EARLY]
6442 Format: { <int> [,<int>] | force | noforce }
6443 <int> -- Number of I/O TLB slabs
6444 <int> -- Second integer after comma. Number of swiotlb
6445 areas with their own lock. Will be rounded up
6447 force -- force using of bounce buffers even if they
6448 wouldn't be automatically used by the kernel
6449 noforce -- Never use bounce buffers (for debugging)
6451 switches= [HW,M68k,EARLY]
6454 Set a sysctl parameter, right before loading the init
6455 process, as if the value was written to the respective
6456 /proc/sys/... file. Both '.' and '/' are recognized as
6457 separators. Unrecognized parameters and invalid values
6458 are reported in the kernel log. Sysctls registered
6459 later by a loaded module cannot be set this way.
6460 Example: sysctl.vm.swappiness=40
6462 sysrq_always_enabled
6464 Ignore sysrq setting - this boot parameter will
6465 neutralize any effect of /proc/sys/kernel/sysrq.
6466 Useful for debugging.
6468 tcpmhash_entries= [KNL,NET]
6469 Set the number of tcp_metrics_hash slots.
6470 Default value is 8192 or 16384 depending on total
6471 ram pages. This is used to specify the TCP metrics
6472 cache size. See Documentation/networking/ip-sysctl.rst
6473 "tcp_no_metrics_save" section for more details.
6477 test_suspend= [SUSPEND]
6478 Format: { "mem" | "standby" | "freeze" }[,N]
6479 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6480 standby suspend) or "freeze" (for suspend type freeze)
6481 as the system sleep state during system startup with
6482 the optional capability to repeat N number of times.
6483 The system is woken from this state using a
6484 wakeup-capable RTC alarm.
6486 thash_entries= [KNL,NET]
6487 Set number of hash buckets for TCP connection
6489 thermal.act= [HW,ACPI]
6490 -1: disable all active trip points in all thermal zones
6491 <degrees C>: override all lowest active trip points
6493 thermal.crt= [HW,ACPI]
6494 -1: disable all critical trip points in all thermal zones
6495 <degrees C>: override all critical trip points
6497 thermal.off= [HW,ACPI]
6498 1: disable ACPI thermal control
6500 thermal.psv= [HW,ACPI]
6501 -1: disable all passive trip points
6502 <degrees C>: override all passive trip points to this
6505 thermal.tzp= [HW,ACPI]
6506 Specify global default ACPI thermal zone polling rate
6507 <deci-seconds>: poll all this frequency
6508 0: no polling (default)
6510 threadirqs [KNL,EARLY]
6511 Force threading of all interrupt handlers except those
6512 marked explicitly IRQF_NO_THREAD.
6514 topology= [S390,EARLY]
6516 Specify if the kernel should make use of the cpu
6517 topology information if the hardware supports this.
6518 The scheduler will make use of this information and
6519 e.g. base its process migration decisions on it.
6522 topology_updates= [KNL, PPC, NUMA]
6524 Specify if the kernel should ignore (off)
6525 topology updates sent by the hypervisor to this
6528 torture.disable_onoff_at_boot= [KNL]
6529 Prevent the CPU-hotplug component of torturing
6530 until after init has spawned.
6532 torture.ftrace_dump_at_shutdown= [KNL]
6533 Dump the ftrace buffer at torture-test shutdown,
6534 even if there were no errors. This can be a
6535 very costly operation when many torture tests
6536 are running concurrently, especially on systems
6537 with rotating-rust storage.
6539 torture.verbose_sleep_frequency= [KNL]
6540 Specifies how many verbose printk()s should be
6541 emitted between each sleep. The default of zero
6542 disables verbose-printk() sleeping.
6544 torture.verbose_sleep_duration= [KNL]
6545 Duration of each verbose-printk() sleep in jiffies.
6549 tpm_suspend_pcr=[HW,TPM]
6550 Format: integer pcr id
6551 Specify that at suspend time, the tpm driver
6552 should extend the specified pcr with zeros,
6553 as a workaround for some chips which fail to
6554 flush the last written pcr on TPM_SaveState.
6555 This will guarantee that all the other pcrs
6558 tpm_tis.interrupts= [HW,TPM]
6559 Enable interrupts for the MMIO based physical layer
6560 for the FIFO interface. By default it is set to false
6561 (0). For more information about TPM hardware interfaces
6562 defined by Trusted Computing Group (TCG) see
6563 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/
6566 Have the tracepoints sent to printk as well as the
6567 tracing ring buffer. This is useful for early boot up
6568 where the system hangs or reboots and does not give the
6569 option for reading the tracing buffer or performing a
6570 ftrace_dump_on_oops.
6572 To turn off having tracepoints sent to printk,
6573 echo 0 > /proc/sys/kernel/tracepoint_printk
6574 Note, echoing 1 into this file without the
6575 tracepoint_printk kernel cmdline option has no effect.
6577 The tp_printk_stop_on_boot (see below) can also be used
6578 to stop the printing of events to console at
6583 Having tracepoints sent to printk() and activating high
6584 frequency tracepoints such as irq or sched, can cause
6585 the system to live lock.
6587 tp_printk_stop_on_boot [FTRACE]
6588 When tp_printk (above) is set, it can cause a lot of noise
6589 on the console. It may be useful to only include the
6590 printing of events during boot up, as user space may
6591 make the system inoperable.
6593 This command line option will stop the printing of events
6594 to console at the late_initcall_sync() time frame.
6596 trace_buf_size=nn[KMG]
6597 [FTRACE] will set tracing buffer size on each cpu.
6599 trace_clock= [FTRACE] Set the clock used for tracing events
6601 local - Use the per CPU time stamp counter
6602 (converted into nanoseconds). Fast, but
6603 depending on the architecture, may not be
6604 in sync between CPUs.
6605 global - Event time stamps are synchronize across
6606 CPUs. May be slower than the local clock,
6607 but better for some race conditions.
6608 counter - Simple counting of events (1, 2, ..)
6609 note, some counts may be skipped due to the
6610 infrastructure grabbing the clock more than
6612 uptime - Use jiffies as the time stamp.
6613 perf - Use the same clock that perf uses.
6614 mono - Use ktime_get_mono_fast_ns() for time stamps.
6615 mono_raw - Use ktime_get_raw_fast_ns() for time
6617 boot - Use ktime_get_boot_fast_ns() for time stamps.
6618 Architectures may add more clocks. See
6619 Documentation/trace/ftrace.rst for more details.
6621 trace_event=[event-list]
6622 [FTRACE] Set and start specified trace events in order
6623 to facilitate early boot debugging. The event-list is a
6624 comma-separated list of trace events to enable. See
6625 also Documentation/trace/events.rst
6627 trace_instance=[instance-info]
6628 [FTRACE] Create a ring buffer instance early in boot up.
6629 This will be listed in:
6631 /sys/kernel/tracing/instances
6633 Events can be enabled at the time the instance is created
6636 trace_instance=<name>,<system1>:<event1>,<system2>:<event2>
6638 Note, the "<system*>:" portion is optional if the event is
6641 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall
6643 will enable the "sched_switch" event (note, the "sched:" is optional, and
6644 the same thing would happen if it was left off). The irq_handler_entry
6645 event, and all events under the "initcall" system.
6647 trace_options=[option-list]
6648 [FTRACE] Enable or disable tracer options at boot.
6649 The option-list is a comma delimited list of options
6650 that can be enabled or disabled just as if you were
6651 to echo the option name into
6653 /sys/kernel/tracing/trace_options
6655 For example, to enable stacktrace option (to dump the
6656 stack trace of each event), add to the command line:
6658 trace_options=stacktrace
6660 See also Documentation/trace/ftrace.rst "trace options"
6663 trace_trigger=[trigger-list]
6664 [FTRACE] Add a event trigger on specific events.
6665 Set a trigger on top of a specific event, with an optional
6668 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..."
6669 Where more than one trigger may be specified that are comma deliminated.
6673 trace_trigger="sched_switch.stacktrace if prev_state == 2"
6675 The above will enable the "stacktrace" trigger on the "sched_switch"
6676 event but only trigger it if the "prev_state" of the "sched_switch"
6677 event is "2" (TASK_UNINTERUPTIBLE).
6679 See also "Event triggers" in Documentation/trace/events.rst
6683 [FTRACE] enable this option to disable tracing when a
6684 warning is hit. This turns off "tracing_on". Tracing can
6685 be enabled again by echoing '1' into the "tracing_on"
6686 file located in /sys/kernel/tracing/
6688 This option is useful, as it disables the trace before
6689 the WARNING dump is called, which prevents the trace to
6690 be filled with content caused by the warning output.
6692 This option can also be set at run time via the sysctl
6693 option: kernel/traceoff_on_warning
6695 transparent_hugepage=
6697 Format: [always|madvise|never]
6698 Can be used to control the default behavior of the system
6699 with respect to transparent hugepages.
6700 See Documentation/admin-guide/mm/transhuge.rst
6703 trusted.source= [KEYS]
6705 This parameter identifies the trust source as a backend
6706 for trusted keys implementation. Supported trust
6711 If not specified then it defaults to iterating through
6712 the trust source list starting with TPM and assigns the
6713 first trust source as a backend which is initialized
6714 successfully during iteration.
6718 The RNG used to generate key material for trusted keys.
6721 - the same value as trusted.source: "tpm" or "tee"
6723 If not specified, "default" is used. In this case,
6724 the RNG's choice is left to each individual trust source.
6726 tsc= Disable clocksource stability checks for TSC.
6728 [x86] reliable: mark tsc clocksource as reliable, this
6729 disables clocksource verification at runtime, as well
6730 as the stability checks done at bootup. Used to enable
6731 high-resolution timer mode on older hardware, and in
6732 virtualized environment.
6733 [x86] noirqtime: Do not use TSC to do irq accounting.
6734 Used to run time disable IRQ_TIME_ACCOUNTING on any
6735 platforms where RDTSC is slow and this accounting
6737 [x86] unstable: mark the TSC clocksource as unstable, this
6738 marks the TSC unconditionally unstable at bootup and
6739 avoids any further wobbles once the TSC watchdog notices.
6740 [x86] nowatchdog: disable clocksource watchdog. Used
6741 in situations with strict latency requirements (where
6742 interruptions from clocksource watchdog are not
6744 [x86] recalibrate: force recalibration against a HW timer
6745 (HPET or PM timer) on systems whose TSC frequency was
6746 obtained from HW or FW using either an MSR or CPUID(0x15).
6747 Warn if the difference is more than 500 ppm.
6748 [x86] watchdog: Use TSC as the watchdog clocksource with
6749 which to check other HW timers (HPET or PM timer), but
6750 only on systems where TSC has been deemed trustworthy.
6751 This will be suppressed by an earlier tsc=nowatchdog and
6752 can be overridden by a later tsc=nowatchdog. A console
6753 message will flag any such suppression or overriding.
6755 tsc_early_khz= [X86,EARLY] Skip early TSC calibration and use the given
6756 value instead. Useful when the early TSC frequency discovery
6757 procedure is not reliable, such as on overclocked systems
6758 with CPUID.16h support and partial CPUID.15h support.
6759 Format: <unsigned int>
6761 tsx= [X86] Control Transactional Synchronization
6762 Extensions (TSX) feature in Intel processors that
6763 support TSX control.
6765 This parameter controls the TSX feature. The options are:
6767 on - Enable TSX on the system. Although there are
6768 mitigations for all known security vulnerabilities,
6769 TSX has been known to be an accelerator for
6770 several previous speculation-related CVEs, and
6771 so there may be unknown security risks associated
6772 with leaving it enabled.
6774 off - Disable TSX on the system. (Note that this
6775 option takes effect only on newer CPUs which are
6776 not vulnerable to MDS, i.e., have
6777 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6778 the new IA32_TSX_CTRL MSR through a microcode
6779 update. This new MSR allows for the reliable
6780 deactivation of the TSX functionality.)
6782 auto - Disable TSX if X86_BUG_TAA is present,
6783 otherwise enable TSX on the system.
6785 Not specifying this option is equivalent to tsx=off.
6787 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6790 tsx_async_abort= [X86,INTEL,EARLY] Control mitigation for the TSX Async
6791 Abort (TAA) vulnerability.
6793 Similar to Micro-architectural Data Sampling (MDS)
6794 certain CPUs that support Transactional
6795 Synchronization Extensions (TSX) are vulnerable to an
6796 exploit against CPU internal buffers which can forward
6797 information to a disclosure gadget under certain
6800 In vulnerable processors, the speculatively forwarded
6801 data can be used in a cache side channel attack, to
6802 access data to which the attacker does not have direct
6805 This parameter controls the TAA mitigation. The
6808 full - Enable TAA mitigation on vulnerable CPUs
6811 full,nosmt - Enable TAA mitigation and disable SMT on
6812 vulnerable CPUs. If TSX is disabled, SMT
6813 is not disabled because CPU is not
6814 vulnerable to cross-thread TAA attacks.
6815 off - Unconditionally disable TAA mitigation
6817 On MDS-affected machines, tsx_async_abort=off can be
6818 prevented by an active MDS mitigation as both vulnerabilities
6819 are mitigated with the same mechanism so in order to disable
6820 this mitigation, you need to specify mds=off too.
6822 Not specifying this option is equivalent to
6823 tsx_async_abort=full. On CPUs which are MDS affected
6824 and deploy MDS mitigation, TAA mitigation is not
6825 required and doesn't provide any additional
6829 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6831 turbografx.map[2|3]= [HW,JOY]
6832 TurboGraFX parallel port interface
6834 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6835 See also Documentation/input/devices/joystick-parport.rst
6837 udbg-immortal [PPC] When debugging early kernel crashes that
6838 happen after console_init() and before a proper
6839 console driver takes over, this boot options might
6840 help "seeing" what's going on.
6842 uhash_entries= [KNL,NET]
6843 Set number of hash buckets for UDP/UDP-Lite connections
6846 [USB] Ignore overcurrent events (default N).
6847 Some badly-designed motherboards generate lots of
6848 bogus events, for ports that aren't wired to
6849 anything. Set this parameter to avoid log spamming.
6850 Note that genuine overcurrent events won't be
6854 [X86] Cause panic on unknown NMI.
6856 unwind_debug [X86-64,EARLY]
6857 Enable unwinder debug output. This can be
6858 useful for debugging certain unwinder error
6859 conditions, including corrupt stacks and
6860 bad/missing unwinder metadata.
6862 usbcore.authorized_default=
6863 [USB] Default USB device authorization:
6864 (default -1 = authorized (same as 1),
6865 0 = not authorized, 1 = authorized, 2 = authorized
6866 if device connected to internal port)
6868 usbcore.autosuspend=
6869 [USB] The autosuspend time delay (in seconds) used
6870 for newly-detected USB devices (default 2). This
6871 is the time required before an idle device will be
6872 autosuspended. Devices for which the delay is set
6873 to a negative value won't be autosuspended at all.
6875 usbcore.usbfs_snoop=
6876 [USB] Set to log all usbfs traffic (default 0 = off).
6878 usbcore.usbfs_snoop_max=
6879 [USB] Maximum number of bytes to snoop in each URB
6882 usbcore.blinkenlights=
6883 [USB] Set to cycle leds on hubs (default 0 = off).
6885 usbcore.old_scheme_first=
6886 [USB] Start with the old device initialization
6887 scheme (default 0 = off).
6889 usbcore.usbfs_memory_mb=
6890 [USB] Memory limit (in MB) for buffers allocated by
6891 usbfs (default = 16, 0 = max = 2047).
6893 usbcore.use_both_schemes=
6894 [USB] Try the other device initialization scheme
6895 if the first one fails (default 1 = enabled).
6897 usbcore.initial_descriptor_timeout=
6898 [USB] Specifies timeout for the initial 64-byte
6899 USB_REQ_GET_DESCRIPTOR request in milliseconds
6900 (default 5000 = 5.0 seconds).
6902 usbcore.nousb [USB] Disable the USB subsystem
6905 [USB] A list of quirk entries to augment the built-in
6906 usb core quirk list. List entries are separated by
6907 commas. Each entry has the form
6908 VendorID:ProductID:Flags. The IDs are 4-digit hex
6909 numbers and Flags is a set of letters. Each letter
6910 will change the built-in quirk; setting it if it is
6911 clear and clearing it if it is set. The letters have
6912 the following meanings:
6913 a = USB_QUIRK_STRING_FETCH_255 (string
6914 descriptors must not be fetched using
6916 b = USB_QUIRK_RESET_RESUME (device can't resume
6917 correctly so reset it instead);
6918 c = USB_QUIRK_NO_SET_INTF (device can't handle
6919 Set-Interface requests);
6920 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6921 handle its Configuration or Interface
6923 e = USB_QUIRK_RESET (device can't be reset
6924 (e.g morph devices), don't use reset);
6925 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6926 more interface descriptions than the
6927 bNumInterfaces count, and can't handle
6928 talking to these interfaces);
6929 g = USB_QUIRK_DELAY_INIT (device needs a pause
6930 during initialization, after we read
6931 the device descriptor);
6932 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6933 high speed and super speed interrupt
6934 endpoints, the USB 2.0 and USB 3.0 spec
6935 require the interval in microframes (1
6936 microframe = 125 microseconds) to be
6937 calculated as interval = 2 ^
6939 Devices with this quirk report their
6940 bInterval as the result of this
6941 calculation instead of the exponent
6942 variable used in the calculation);
6943 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6944 handle device_qualifier descriptor
6946 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6947 generates spurious wakeup, ignore
6948 remote wakeup capability);
6949 k = USB_QUIRK_NO_LPM (device can't handle Link
6951 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6952 (Device reports its bInterval as linear
6953 frames instead of the USB 2.0
6955 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6956 to be disconnected before suspend to
6957 prevent spurious wakeup);
6958 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6959 pause after every control message);
6960 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6961 delay after resetting its port);
6962 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT
6963 (Reduce timeout of the SET_ADDRESS
6964 request from 5000 ms to 500 ms);
6965 Example: quirks=0781:5580:bk,0a5c:5834:gij
6968 [USBHID] The interval which mice are to be polled at.
6971 [USBHID] The interval which joysticks are to be polled at.
6974 [USBHID] The interval which keyboards are to be polled at.
6976 usb-storage.delay_use=
6977 [UMS] The delay in seconds before a new device is
6978 scanned for Logical Units (default 1).
6981 [UMS] A list of quirks entries to supplement or
6982 override the built-in unusual_devs list. List
6983 entries are separated by commas. Each entry has
6984 the form VID:PID:Flags where VID and PID are Vendor
6985 and Product ID values (4-digit hex numbers) and
6986 Flags is a set of characters, each corresponding
6987 to a common usb-storage quirk flag as follows:
6988 a = SANE_SENSE (collect more than 18 bytes
6989 of sense data, not on uas);
6990 b = BAD_SENSE (don't collect more than 18
6991 bytes of sense data, not on uas);
6992 c = FIX_CAPACITY (decrease the reported
6993 device capacity by one sector);
6994 d = NO_READ_DISC_INFO (don't use
6995 READ_DISC_INFO command, not on uas);
6996 e = NO_READ_CAPACITY_16 (don't use
6997 READ_CAPACITY_16 command);
6998 f = NO_REPORT_OPCODES (don't use report opcodes
7000 g = MAX_SECTORS_240 (don't transfer more than
7001 240 sectors at a time, uas only);
7002 h = CAPACITY_HEURISTICS (decrease the
7003 reported device capacity by one
7004 sector if the number is odd);
7005 i = IGNORE_DEVICE (don't bind to this
7007 j = NO_REPORT_LUNS (don't use report luns
7009 k = NO_SAME (do not use WRITE_SAME, uas only)
7010 l = NOT_LOCKABLE (don't try to lock and
7011 unlock ejectable media, not on uas);
7012 m = MAX_SECTORS_64 (don't transfer more
7013 than 64 sectors = 32 KB at a time,
7015 n = INITIAL_READ10 (force a retry of the
7016 initial READ(10) command, not on uas);
7017 o = CAPACITY_OK (accept the capacity
7018 reported by the device, not on uas);
7019 p = WRITE_CACHE (the device cache is ON
7020 by default, not on uas);
7021 r = IGNORE_RESIDUE (the device reports
7022 bogus residue values, not on uas);
7023 s = SINGLE_LUN (the device has only one
7025 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
7026 commands, uas only);
7027 u = IGNORE_UAS (don't bind to the uas driver);
7028 w = NO_WP_DETECT (don't test whether the
7029 medium is write-protected).
7030 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
7031 even if the device claims no cache,
7033 Example: quirks=0419:aaf5:rl,0421:0433:rc
7035 user_debug= [KNL,ARM]
7037 See arch/arm/Kconfig.debug help text.
7038 1 - undefined instruction events
7040 4 - invalid data aborts
7043 Example: user_debug=31
7046 [X86,EARLY] Flags controlling user PTE allocations.
7048 nohigh = do not allocate PTE pages in
7049 HIGHMEM regardless of setting
7052 vdso= [X86,SH,SPARC]
7053 On X86_32, this is an alias for vdso32=. Otherwise:
7055 vdso=1: enable VDSO (the default)
7056 vdso=0: disable VDSO mapping
7058 vdso32= [X86] Control the 32-bit vDSO
7059 vdso32=1: enable 32-bit VDSO
7060 vdso32=0 or vdso32=2: disable 32-bit VDSO
7062 See the help text for CONFIG_COMPAT_VDSO for more
7063 details. If CONFIG_COMPAT_VDSO is set, the default is
7064 vdso32=0; otherwise, the default is vdso32=1.
7066 For compatibility with older kernels, vdso32=2 is an
7069 Try vdso32=0 if you encounter an error that says:
7070 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
7073 vector=percpu: enable percpu vector domain
7075 video= [FB,EARLY] Frame buffer configuration
7076 See Documentation/fb/modedb.rst.
7078 video.brightness_switch_enabled= [ACPI]
7080 If set to 1, on receiving an ACPI notify event
7081 generated by hotkey, video driver will adjust brightness
7082 level and then send out the event to user space through
7083 the allocated input device. If set to 0, video driver
7084 will only send out the event without touching backlight
7089 [VMMIO] Memory mapped virtio (platform) device.
7091 <size>@<baseaddr>:<irq>[:<id>]
7093 <size> := size (can use standard suffixes
7095 <baseaddr> := physical base address
7096 <irq> := interrupt number (as passed to
7098 <id> := (optional) platform device id
7100 virtio_mmio.device=1K@0x100b0000:48:7
7102 Can be used multiple times for multiple devices.
7104 vga= [BOOT,X86-32] Select a particular video mode
7105 See Documentation/arch/x86/boot.rst and
7106 Documentation/admin-guide/svga.rst.
7107 Use vga=ask for menu.
7108 This is actually a boot loader parameter; the value is
7109 passed to the kernel using a special protocol.
7111 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
7112 May slow down system boot speed, especially when
7113 enabled on systems with a large amount of memory.
7114 All options are enabled by default, and this
7115 interface is meant to allow for selectively
7116 enabling or disabling specific virtual memory
7119 Available options are:
7120 P Enable page structure init time poisoning
7121 - Disable all of the above options
7123 vmalloc=nn[KMG] [KNL,BOOT,EARLY] Forces the vmalloc area to have an
7124 exact size of <nn>. This can be used to increase
7125 the minimum size (128MB on x86). It can also be
7126 used to decrease the size and leave more room
7127 for directly mapped kernel RAM.
7129 vmcp_cma=nn[MG] [KNL,S390,EARLY]
7130 Sets the memory size reserved for contiguous memory
7131 allocations for the vmcp device driver.
7133 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
7136 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
7139 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
7142 vsyscall= [X86-64,EARLY]
7143 Controls the behavior of vsyscalls (i.e. calls to
7144 fixed addresses of 0xffffffffff600x00 from legacy
7145 code). Most statically-linked binaries and older
7146 versions of glibc use these calls. Because these
7147 functions are at fixed addresses, they make nice
7148 targets for exploits that can control RIP.
7150 emulate Vsyscalls turn into traps and are emulated
7151 reasonably safely. The vsyscall page is
7154 xonly [default] Vsyscalls turn into traps and are
7155 emulated reasonably safely. The vsyscall
7156 page is not readable.
7158 none Vsyscalls don't work at all. This makes
7159 them quite hard to use for exploits but
7160 might break your system.
7162 vt.color= [VT] Default text color.
7163 Format: 0xYX, X = foreground, Y = background.
7164 Default: 0x07 = light gray on black.
7166 vt.cur_default= [VT] Default cursor shape.
7167 Format: 0xCCBBAA, where AA, BB, and CC are the same as
7168 the parameters of the <Esc>[?A;B;Cc escape sequence;
7169 see VGA-softcursor.txt. Default: 2 = underline.
7171 vt.default_blu= [VT]
7172 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
7173 Change the default blue palette of the console.
7174 This is a 16-member array composed of values
7177 vt.default_grn= [VT]
7178 Format: <green0>,<green1>,<green2>,...,<green15>
7179 Change the default green palette of the console.
7180 This is a 16-member array composed of values
7183 vt.default_red= [VT]
7184 Format: <red0>,<red1>,<red2>,...,<red15>
7185 Change the default red palette of the console.
7186 This is a 16-member array composed of values
7192 Set system-wide default UTF-8 mode for all tty's.
7193 Default is 1, i.e. UTF-8 mode is enabled for all
7194 newly opened terminals.
7196 vt.global_cursor_default=
7199 Set system-wide default for whether a cursor
7200 is shown on new VTs. Default is -1,
7201 i.e. cursors will be created by default unless
7202 overridden by individual drivers. 0 will hide
7203 cursors, 1 will display them.
7205 vt.italic= [VT] Default color for italic text; 0-15.
7208 vt.underline= [VT] Default color for underlined text; 0-15.
7211 watchdog timers [HW,WDT] For information on watchdog timers,
7212 see Documentation/watchdog/watchdog-parameters.rst
7213 or other driver-specific files in the
7214 Documentation/watchdog/ directory.
7218 Set the hard lockup detector stall duration
7219 threshold in seconds. The soft lockup detector
7220 threshold is set to twice the value. A value of 0
7221 disables both lockup detectors. Default is 10
7224 workqueue.unbound_cpus=
7225 [KNL,SMP] Specify to constrain one or some CPUs
7226 to use in unbound workqueues.
7228 By default, all online CPUs are available for
7231 workqueue.watchdog_thresh=
7232 If CONFIG_WQ_WATCHDOG is configured, workqueue can
7233 warn stall conditions and dump internal state to
7234 help debugging. 0 disables workqueue stall
7235 detection; otherwise, it's the stall threshold
7236 duration in seconds. The default value is 30 and
7237 it can be updated at runtime by writing to the
7238 corresponding sysfs file.
7240 workqueue.cpu_intensive_thresh_us=
7241 Per-cpu work items which run for longer than this
7242 threshold are automatically considered CPU intensive
7243 and excluded from concurrency management to prevent
7244 them from noticeably delaying other per-cpu work
7245 items. Default is 10000 (10ms).
7247 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel
7248 will report the work functions which violate this
7249 threshold repeatedly. They are likely good
7250 candidates for using WQ_UNBOUND workqueues instead.
7252 workqueue.cpu_intensive_warning_thresh=<uint>
7253 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel
7254 will report the work functions which violate the
7255 intensive_threshold_us repeatedly. In order to prevent
7256 spurious warnings, start printing only after a work
7257 function has violated this threshold number of times.
7259 The default is 4 times. 0 disables the warning.
7261 workqueue.power_efficient
7262 Per-cpu workqueues are generally preferred because
7263 they show better performance thanks to cache
7264 locality; unfortunately, per-cpu workqueues tend to
7265 be more power hungry than unbound workqueues.
7267 Enabling this makes the per-cpu workqueues which
7268 were observed to contribute significantly to power
7269 consumption unbound, leading to measurably lower
7270 power usage at the cost of small performance
7273 The default value of this parameter is determined by
7274 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
7276 workqueue.default_affinity_scope=
7277 Select the default affinity scope to use for unbound
7278 workqueues. Can be one of "cpu", "smt", "cache",
7279 "numa" and "system". Default is "cache". For more
7280 information, see the Affinity Scopes section in
7281 Documentation/core-api/workqueue.rst.
7283 This can be changed after boot by writing to the
7284 matching /sys/module/workqueue/parameters file. All
7285 workqueues with the "default" affinity scope will be
7286 updated accordignly.
7288 workqueue.debug_force_rr_cpu
7289 Workqueue used to implicitly guarantee that work
7290 items queued without explicit CPU specified are put
7291 on the local CPU. This guarantee is no longer true
7292 and while local CPU is still preferred work items
7293 may be put on foreign CPUs. This debug option
7294 forces round-robin CPU selection to flush out
7295 usages which depend on the now broken guarantee.
7296 When enabled, memory and cache locality will be
7299 writecombine= [LOONGARCH,EARLY] Control the MAT (Memory Access
7300 Type) of ioremap_wc().
7302 on - Enable writecombine, use WUC for ioremap_wc()
7303 off - Disable writecombine, use SUC for ioremap_wc()
7305 x2apic_phys [X86-64,APIC,EARLY] Use x2apic physical mode instead of
7306 default x2apic cluster mode on platforms
7309 xen_512gb_limit [KNL,X86-64,XEN]
7310 Restricts the kernel running paravirtualized under Xen
7311 to use only up to 512 GB of RAM. The reason to do so is
7312 crash analysis tools and Xen tools for doing domain
7313 save/restore/migration must be enabled to handle larger
7316 xen_emul_unplug= [HW,X86,XEN,EARLY]
7317 Unplug Xen emulated devices
7318 Format: [unplug0,][unplug1]
7319 ide-disks -- unplug primary master IDE devices
7320 aux-ide-disks -- unplug non-primary-master IDE devices
7321 nics -- unplug network devices
7322 all -- unplug all emulated devices (NICs and IDE disks)
7323 unnecessary -- unplugging emulated devices is
7324 unnecessary even if the host did not respond to
7326 never -- do not unplug even if version check succeeds
7328 xen_legacy_crash [X86,XEN,EARLY]
7329 Crash from Xen panic notifier, without executing late
7330 panic() code such as dumping handler.
7332 xen_msr_safe= [X86,XEN,EARLY]
7334 Select whether to always use non-faulting (safe) MSR
7335 access functions when running as Xen PV guest. The
7336 default value is controlled by CONFIG_XEN_PV_MSR_SAFE.
7338 xen_nopvspin [X86,XEN,EARLY]
7339 Disables the qspinlock slowpath using Xen PV optimizations.
7340 This parameter is obsoleted by "nopvspin" parameter, which
7341 has equivalent effect for XEN platform.
7344 Disables the PV optimizations forcing the HVM guest to
7345 run as generic HVM guest with no PV drivers.
7346 This option is obsoleted by the "nopv" option, which
7347 has equivalent effect for XEN platform.
7349 xen_no_vector_callback
7350 [KNL,X86,XEN,EARLY] Disable the vector callback for Xen
7351 event channel interrupts.
7353 xen_scrub_pages= [XEN]
7354 Boolean option to control scrubbing pages before giving them back
7355 to Xen, for use by other domains. Can be also changed at runtime
7356 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
7357 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
7359 xen_timer_slop= [X86-64,XEN,EARLY]
7360 Set the timer slop (in nanoseconds) for the virtual Xen
7361 timers (default is 100000). This adjusts the minimum
7362 delta of virtualized Xen timers, where lower values
7363 improve timer resolution at the expense of processing
7364 more timer interrupts.
7366 xen.balloon_boot_timeout= [XEN]
7367 The time (in seconds) to wait before giving up to boot
7368 in case initial ballooning fails to free enough memory.
7369 Applies only when running as HVM or PVH guest and
7370 started with less memory configured than allowed at
7371 max. Default is 180.
7373 xen.event_eoi_delay= [XEN]
7374 How long to delay EOI handling in case of event
7375 storms (jiffies). Default is 10.
7377 xen.event_loop_timeout= [XEN]
7378 After which time (jiffies) the event handling loop
7379 should start to delay EOI handling. Default is 2.
7381 xen.fifo_events= [XEN]
7382 Boolean parameter to disable using fifo event handling
7383 even if available. Normally fifo event handling is
7384 preferred over the 2-level event handling, as it is
7385 fairer and the number of possible event channels is
7386 much higher. Default is on (use fifo events).
7388 xirc2ps_cs= [NET,PCMCIA]
7390 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
7393 By default on POWER9 and above, the kernel will
7394 natively use the XIVE interrupt controller. This option
7395 allows the fallback firmware mode to be used:
7397 off Fallback to firmware control of XIVE interrupt
7398 controller on both pseries and powernv
7399 platforms. Only useful on POWER9 and above.
7401 xive.store-eoi=off [PPC]
7402 By default on POWER10 and above, the kernel will use
7403 stores for EOI handling when the XIVE interrupt mode
7404 is active. This option allows the XIVE driver to use
7405 loads instead, as on POWER9.
7407 xhci-hcd.quirks [USB,KNL]
7408 A hex value specifying bitmask with supplemental xhci
7409 host controller quirks. Meaning of each bit can be
7410 consulted in header drivers/usb/host/xhci.h.
7413 Format: { early | on | rw | ro | off }
7414 Controls if xmon debugger is enabled. Default is off.
7415 Passing only "xmon" is equivalent to "xmon=early".
7416 early Call xmon as early as possible on boot; xmon
7417 debugger is called from setup_arch().
7418 on xmon debugger hooks will be installed so xmon
7419 is only called on a kernel crash. Default mode,
7420 i.e. either "ro" or "rw" mode, is controlled
7421 with CONFIG_XMON_DEFAULT_RO_MODE.
7422 rw xmon debugger hooks will be installed so xmon
7423 is called only on a kernel crash, mode is write,
7424 meaning SPR registers, memory and, other data
7425 can be written using xmon commands.
7426 ro same as "rw" option above but SPR registers,
7427 memory, and other data can't be written using
7429 off xmon is disabled.