1 acpi= [HW,ACPI,X86,ARM64,RISCV64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
5 force -- enable ACPI if default was off
6 on -- enable ACPI but allow fallback to DT [arm64,riscv64]
7 off -- disable ACPI if default was on
8 noirq -- do not use ACPI for IRQ routing
9 strict -- Be less tolerant of platforms that are not
10 strictly ACPI specification compliant.
11 rsdt -- prefer RSDT over (default) XSDT
12 copy_dsdt -- copy DSDT to memory
13 For ARM64 and RISCV64, ONLY "acpi=off", "acpi=on" or
14 "acpi=force" are available
16 See also Documentation/power/runtime_pm.rst, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
25 { vendor | video | native | none }
26 If set to vendor, prefer vendor-specific driver
27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
28 of the ACPI video.ko driver.
29 If set to video, use the ACPI video.ko driver.
30 If set to native, use the device's native backlight mode.
31 If set to none, disable the ACPI backlight interface.
33 acpi_force_32bit_fadt_addr
34 force FADT to use 32 bit addresses rather than the
35 64 bit X_* addresses. Some firmware have broken 64
36 bit addresses for force ACPI ignore these and use
37 the older legacy 32 bit addresses.
39 acpica_no_return_repair [HW, ACPI]
40 Disable AML predefined validation mechanism
41 This mechanism can repair the evaluation result to make
42 the return objects more ACPI specification compliant.
43 This option is useful for developers to identify the
44 root cause of an AML interpreter issue when the issue
45 has something to do with the repair mechanism.
47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
51 debug output. Bits in debug_layer correspond to a
52 _COMPONENT in an ACPI source file, e.g.,
53 #define _COMPONENT ACPI_EVENTS
54 Bits in debug_level correspond to a level in
55 ACPI_DEBUG_PRINT statements, e.g.,
56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
57 The debug_level mask defaults to "info". See
58 Documentation/firmware-guide/acpi/debug.rst for more information about
59 debug layers and levels.
61 Enable processor driver info messages:
62 acpi.debug_layer=0x20000000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
116 Format: <byte> or <bitmap-list>
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
163 acpi_osi=!! # enable all built-in OS vendor
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_hwsig,
229 s4_nohwsig, old_ordering, nonvs,
230 sci_force_enable, nobl }
231 See Documentation/power/video.rst for information on
233 s3_beep is for debugging; it makes the PC's speaker beep
234 as soon as the kernel's real-mode entry point is called.
235 s4_hwsig causes the kernel to check the ACPI hardware
236 signature during resume from hibernation, and gracefully
237 refuse to resume if it has changed. This complies with
238 the ACPI specification but not with reality, since
239 Windows does not do this and many laptops do change it
240 on docking. So the default behaviour is to allow resume
241 and simply warn when the signature changes, unless the
242 s4_hwsig option is enabled.
243 s4_nohwsig prevents ACPI hardware signature from being
244 used (or even warned about) during resume.
245 old_ordering causes the ACPI 1.0 ordering of the _PTS
246 control method, with respect to putting devices into
247 low power states, to be enforced (the ACPI 2.0 ordering
248 of _PTS is used by default).
249 nonvs prevents the kernel from saving/restoring the
250 ACPI NVS memory during suspend/hibernation and resume.
251 sci_force_enable causes the kernel to set SCI_EN directly
252 on resume from S1/S3 (which is against the ACPI spec,
253 but some broken systems don't work without it).
254 nobl causes the internal blacklist of systems known to
255 behave incorrectly in some ways with respect to system
256 suspend and resume to be ignored (use wisely).
258 acpi_use_timer_override [HW,ACPI]
259 Use timer override. For some broken Nvidia NF5 boards
260 that require a timer override, but don't have HPET
262 add_efi_memmap [EFI; X86] Include EFI memory map in
263 kernel's map of available physical RAM.
266 { off | try_unsupported }
267 off: disable AGP support
268 try_unsupported: try to drive unsupported chipsets
269 (may crash computer or cause data corruption)
272 See Documentation/sound/alsa-configuration.rst
275 Allow the default userspace alignment fault handler
276 behaviour to be specified. Bit 0 enables warnings,
277 bit 1 enables fixups, and bit 2 sends a segfault.
279 align_va_addr= [X86-64]
280 Align virtual addresses by clearing slice [14:12] when
281 allocating a VMA at process creation time. This option
282 gives you up to 3% performance improvement on AMD F15h
283 machines (where it is enabled by default) for a
284 CPU-intensive style benchmark, and it can vary highly in
285 a microbenchmark depending on workload and compiler.
287 32: only for 32-bit processes
288 64: only for 64-bit processes
289 on: enable for both 32- and 64-bit processes
290 off: disable for both 32- and 64-bit processes
292 alloc_snapshot [FTRACE]
293 Allocate the ftrace snapshot buffer on boot up when the
294 main buffer is allocated. This is handy if debugging
295 and you need to use tracing_snapshot() on boot up, and
296 do not want to use tracing_snapshot_alloc() as it needs
297 to be done where GFP_KERNEL allocations are allowed.
299 allow_mismatched_32bit_el0 [ARM64]
300 Allow execve() of 32-bit applications and setting of the
301 PER_LINUX32 personality on systems where only a strict
302 subset of the CPUs support 32-bit EL0. When this
303 parameter is present, the set of CPUs supporting 32-bit
304 EL0 is indicated by /sys/devices/system/cpu/aarch32_el0
305 and hot-unplug operations may be restricted.
307 See Documentation/arch/arm64/asymmetric-32bit.rst for more
310 amd_iommu= [HW,X86-64]
311 Pass parameters to the AMD IOMMU driver in the system.
313 fullflush - Deprecated, equivalent to iommu.strict=1
314 off - do not initialize any AMD IOMMU found in
316 force_isolation - Force device isolation for all
317 devices. The IOMMU driver is not
318 allowed anymore to lift isolation
319 requirements as needed. This option
320 does not override iommu=pt
321 force_enable - Force enable the IOMMU on platforms known
322 to be buggy with IOMMU enabled. Use this
324 pgtbl_v1 - Use v1 page table for DMA-API (Default).
325 pgtbl_v2 - Use v2 page table for DMA-API.
326 irtcachedis - Disable Interrupt Remapping Table (IRT) caching.
328 amd_iommu_dump= [HW,X86-64]
329 Enable AMD IOMMU driver option to dump the ACPI table
330 for AMD IOMMU. With this option enabled, AMD IOMMU
331 driver will print ACPI tables for AMD IOMMU during
332 IOMMU initialization.
334 amd_iommu_intr= [HW,X86-64]
335 Specifies one of the following AMD IOMMU interrupt
337 legacy - Use legacy interrupt remapping mode.
338 vapic - Use virtual APIC mode, which allows IOMMU
339 to inject interrupts directly into guest.
340 This mode requires kvm-amd.avic=1.
341 (Default when IOMMU HW support is present.)
345 Do not enable amd_pstate as the default
346 scaling driver for the supported processors
348 Use amd_pstate with passive mode as a scaling driver.
349 In this mode autonomous selection is disabled.
350 Driver requests a desired performance level and platform
351 tries to match the same performance level if it is
352 satisfied by guaranteed performance level.
354 Use amd_pstate_epp driver instance as the scaling driver,
355 driver provides a hint to the hardware if software wants
356 to bias toward performance (0x0) or energy efficiency (0xff)
357 to the CPPC firmware. then CPPC power algorithm will
358 calculate the runtime workload and adjust the realtime cores
361 Activate guided autonomous mode. Driver requests minimum and
362 maximum performance level and the platform autonomously
363 selects a performance level in this range and appropriate
364 to the current workload.
366 amijoy.map= [HW,JOY] Amiga joystick support
367 Map of devices attached to JOY0DAT and JOY1DAT
369 See also Documentation/input/joydev/joystick.rst
371 analog.map= [HW,JOY] Analog joystick and gamepad support
372 Specifies type or capabilities of an analog joystick
373 connected to one of 16 gameports
374 Format: <type1>,<type2>,..<type16>
377 Power management functions (SPARCstation-4/5 + deriv.)
379 Disable APC CPU standby support. SPARCstation-Fox does
380 not play well with APC CPU idle - disable it if you have
381 APC and your system crashes randomly.
383 apic= [APIC,X86] Advanced Programmable Interrupt Controller
384 Change the output verbosity while booting
385 Format: { quiet (default) | verbose | debug }
386 Change the amount of debugging information output
387 when initialising the APIC and IO-APIC components.
388 For X86-32, this can also be used to specify an APIC
390 Format: apic=driver_name
391 Examples: apic=bigsmp
393 apic_extnmi= [APIC,X86] External NMI delivery setting
394 Format: { bsp (default) | all | none }
395 bsp: External NMI is delivered only to CPU 0
396 all: External NMIs are broadcast to all CPUs as a
398 none: External NMI is masked for all CPUs. This is
399 useful so that a dump capture kernel won't be
403 See Documentation/networking/ipv6.rst.
405 apm= [APM] Advanced Power Management
406 See header of arch/x86/kernel/apm_32.c.
408 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
409 Format: { "0" | "1" }
410 See security/apparmor/Kconfig help text
413 Default value is set via kernel config option.
415 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
416 Format: <io>,<irq>,<nodeID>
418 arm64.nobti [ARM64] Unconditionally disable Branch Target
419 Identification support
421 arm64.nomops [ARM64] Unconditionally disable Memory Copy and Memory
422 Set instructions support
424 arm64.nomte [ARM64] Unconditionally disable Memory Tagging Extension
427 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
430 arm64.nosme [ARM64] Unconditionally disable Scalable Matrix
433 arm64.nosve [ARM64] Unconditionally disable Scalable Vector
438 atarimouse= [HW,MOUSE] Atari Mouse
440 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
441 EzKey and similar keyboards
443 atkbd.reset= [HW] Reset keyboard during initialization
445 atkbd.set= [HW] Select keyboard code set
446 Format: <int> (2 = AT (default), 3 = PS/2)
448 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
451 atkbd.softraw= [HW] Choose between synthetic and real raw mode
452 Format: <bool> (0 = real, 1 = synthetic (default))
454 atkbd.softrepeat= [HW]
455 Use software keyboard repeat
457 audit= [KNL] Enable the audit sub-system
458 Format: { "0" | "1" | "off" | "on" }
459 0 | off - kernel audit is disabled and can not be
460 enabled until the next reboot
461 unset - kernel audit is initialized but disabled and
462 will be fully enabled by the userspace auditd.
463 1 | on - kernel audit is initialized and partially
464 enabled, storing at most audit_backlog_limit
465 messages in RAM until it is fully enabled by the
469 audit_backlog_limit= [KNL] Set the audit queue size limit.
470 Format: <int> (must be >=0)
473 bau= [X86_UV] Enable the BAU on SGI UV. The default
474 behavior is to disable the BAU (i.e. bau=0).
475 Format: { "0" | "1" }
478 unset - Disable the BAU.
480 baycom_epp= [HW,AX25]
483 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
485 See header of drivers/net/hamradio/baycom_par.c.
487 baycom_ser_fdx= [HW,AX25]
488 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
489 Format: <io>,<irq>,<mode>[,<baud>]
490 See header of drivers/net/hamradio/baycom_ser_fdx.c.
492 baycom_ser_hdx= [HW,AX25]
493 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
494 Format: <io>,<irq>,<mode>
495 See header of drivers/net/hamradio/baycom_ser_hdx.c.
498 Disable BERT OS support on buggy BIOSes.
500 bgrt_disable [ACPI][X86]
501 Disable BGRT to avoid flickering OEM logo.
503 blkdevparts= Manual partition parsing of block device(s) for
504 embedded devices based on command line input.
505 See Documentation/block/cmdline-partition.rst
507 boot_delay= Milliseconds to delay each printk during boot.
508 Only works if CONFIG_BOOT_PRINTK_DELAY is enabled,
509 and you may also have to specify "lpj=". Boot_delay
510 values larger than 10 seconds (10000) are assumed
511 erroneous and ignored.
515 Extended command line options can be added to an initrd
516 and this will cause the kernel to look for it.
518 See Documentation/admin-guide/bootconfig.rst
520 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
521 bttv.radio= Most important insmod options are available as
523 bttv.pll= See Documentation/admin-guide/media/bttv.rst
526 bulk_remove=off [PPC] This parameter disables the use of the pSeries
527 firmware feature for flushing multiple hpte entries
530 c101= [NET] Moxa C101 synchronous serial card
532 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
533 Sometimes CPU hardware bugs make them report the cache
534 size incorrectly. The kernel will attempt work arounds
535 to fix known problems, but for some CPUs it is not
536 possible to determine what the correct size should be.
537 This option provides an override for these situations.
540 [NET] Specifies amount of time (in seconds) that
541 the kernel should wait for a network carrier. By default
542 it waits 120 seconds.
544 ca_keys= [KEYS] This parameter identifies a specific key(s) on
545 the system trusted keyring to be used for certificate
547 format: { id:<keyid> | builtin }
549 cca= [MIPS] Override the kernel pages' cache coherency
550 algorithm. Accepted values range from 0 to 7
551 inclusive. See arch/mips/include/asm/pgtable-bits.h
552 for platform specific values (SB1, Loongson3 and
555 ccw_timeout_log [S390]
556 See Documentation/arch/s390/common_io.rst for details.
558 cgroup_disable= [KNL] Disable a particular controller or optional feature
559 Format: {name of the controller(s) or feature(s) to disable}
560 The effects of cgroup_disable=foo are:
561 - foo isn't auto-mounted if you mount all cgroups in
563 - foo isn't visible as an individually mountable
565 - if foo is an optional feature then the feature is
566 disabled and corresponding cgroup files are not
568 {Currently only "memory" controller deal with this and
569 cut the overhead, others just disable the usage. So
570 only cgroup_disable=memory is actually worthy}
571 Specifying "pressure" disables per-cgroup pressure
572 stall information accounting feature
574 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
575 Format: { { controller | "all" | "named" }
576 [,{ controller | "all" | "named" }...] }
577 Like cgroup_disable, but only applies to cgroup v1;
578 the blacklisted controllers remain available in cgroup2.
579 "all" blacklists all controllers and "named" disables
580 named mounts. Specifying both "all" and "named" disables
583 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
585 nosocket -- Disable socket memory accounting.
586 nokmem -- Disable kernel memory accounting.
587 nobpf -- Disable BPF memory accounting.
589 checkreqprot= [SELINUX] Set initial checkreqprot flag value.
590 Format: { "0" | "1" }
591 See security/selinux/Kconfig help text.
592 0 -- check protection applied by kernel (includes
593 any implied execute protection).
594 1 -- check protection requested by application.
595 Default value is set via a kernel config option.
596 Value can be changed at runtime via
597 /sys/fs/selinux/checkreqprot.
598 Setting checkreqprot to 1 is deprecated.
601 See Documentation/arch/s390/common_io.rst for details.
603 clearcpuid=X[,X...] [X86]
604 Disable CPUID feature X for the kernel. See
605 arch/x86/include/asm/cpufeatures.h for the valid bit
606 numbers X. Note the Linux-specific bits are not necessarily
607 stable over kernel options, but the vendor-specific
609 X can also be a string as appearing in the flags: line
610 in /proc/cpuinfo which does not have the above
611 instability issue. However, not all features have names
613 Note that using this option will taint your kernel.
614 Also note that user programs calling CPUID directly
615 or using the feature without checking anything
616 will still see it. This just prevents it from
617 being used by the kernel or shown in /proc/cpuinfo.
618 Also note the kernel might malfunction if you disable
623 Prevents the clock framework from automatically gating
624 clocks that have not been explicitly enabled by a Linux
625 device driver but are enabled in hardware at reset or
626 by the bootloader/firmware. Note that this does not
627 force such clocks to be always-on nor does it reserve
628 those clocks in any way. This parameter is useful for
629 debug and development, but should not be needed on a
630 platform with proper driver support. For more
631 information, see Documentation/driver-api/clk.rst.
633 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
635 Forces specified clocksource (if available) to be used
636 when calculating gettimeofday(). If specified
637 clocksource is not available, it defaults to PIT.
638 Format: { pit | tsc | cyclone | pmtmr }
640 clocksource= Override the default clocksource
642 Override the default clocksource and use the clocksource
643 with the name specified.
644 Some clocksource names to choose from, depending on
646 [all] jiffies (this is the base, fallback clocksource)
648 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
649 pxa_timer,timer3,32k_counter,timer0_1
650 [X86-32] pit,hpet,tsc;
651 scx200_hrt on Geode; cyclone on IBM x440
659 clocksource.arm_arch_timer.evtstrm=
662 Enable/disable the eventstream feature of the ARM
663 architected timer so that code using WFE-based polling
664 loops can be debugged more effectively on production
667 clocksource.max_cswd_read_retries= [KNL]
668 Number of clocksource_watchdog() retries due to
669 external delays before the clock will be marked
670 unstable. Defaults to two retries, that is,
671 three attempts to read the clock under test.
673 clocksource.verify_n_cpus= [KNL]
674 Limit the number of CPUs checked for clocksources
675 marked with CLOCK_SOURCE_VERIFY_PERCPU that
676 are marked unstable due to excessive skew.
677 A negative value says to check all CPUs, while
678 zero says not to check any. Values larger than
679 nr_cpu_ids are silently truncated to nr_cpu_ids.
680 The actual CPUs are chosen randomly, with
681 no replacement if the same CPU is chosen twice.
683 clocksource-wdtest.holdoff= [KNL]
684 Set the time in seconds that the clocksource
685 watchdog test waits before commencing its tests.
686 Defaults to zero when built as a module and to
687 10 seconds when built into the kernel.
689 cma=nn[MG]@[start[MG][-end[MG]]]
691 Sets the size of kernel global memory area for
692 contiguous memory allocations and optionally the
693 placement constraint by the physical address range of
694 memory allocations. A value of 0 disables CMA
695 altogether. For more information, see
696 kernel/dma/contiguous.c
700 Sets the size of kernel per-numa memory area for
701 contiguous memory allocations. A value of 0 disables
702 per-numa CMA altogether. And If this option is not
703 specified, the default value is 0.
704 With per-numa CMA enabled, DMA users on node nid will
705 first try to allocate buffer from the pernuma area
706 which is located in node nid, if the allocation fails,
707 they will fallback to the global default memory area.
709 numa_cma=<node>:nn[MG][,<node>:nn[MG]]
711 Sets the size of kernel numa memory area for
712 contiguous memory allocations. It will reserve CMA
713 area for the specified node.
715 With numa CMA enabled, DMA users on node nid will
716 first try to allocate buffer from the numa area
717 which is located in node nid, if the allocation fails,
718 they will fallback to the global default memory area.
720 cmo_free_hint= [PPC] Format: { yes | no }
721 Specify whether pages are marked as being inactive
722 when they are freed. This is used in CMO environments
723 to determine OS memory pressure for page stealing by
727 coherent_pool=nn[KMG] [ARM,KNL]
728 Sets the size of memory pool for coherent, atomic dma
729 allocations, by default set to 256K.
731 com20020= [HW,NET] ARCnet - COM20020 chipset
733 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
735 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
739 ARCnet - COM90xx chipset (memory-mapped buffers)
740 Format: <io>[,<irq>[,<memstart>]]
742 condev= [HW,S390] console device
745 con3215_drop= [S390] 3215 console drop mode.
747 When set to true, drop data on the 3215 console when
748 the console buffer is full. In this case the
749 operator using a 3270 terminal emulator (for example
750 x3270) does not have to enter the clear key for the
751 console output to advance and the kernel to continue.
752 This leads to a much faster boot time when a 3270
753 terminal emulator is active. If no 3270 terminal
754 emulator is used, this parameter has no effect.
756 console= [KNL] Output console device and options.
758 tty<n> Use the virtual console device <n>.
762 Use the specified serial port. The options are of
763 the form "bbbbpnf", where "bbbb" is the baud rate,
764 "p" is parity ("n", "o", or "e"), "n" is number of
765 bits, and "f" is flow control ("r" for RTS or
766 omit it). Default is "9600n8".
768 See Documentation/admin-guide/serial-console.rst for more
770 Documentation/networking/netconsole.rst for an
773 uart[8250],io,<addr>[,options]
774 uart[8250],mmio,<addr>[,options]
775 uart[8250],mmio16,<addr>[,options]
776 uart[8250],mmio32,<addr>[,options]
777 uart[8250],0x<addr>[,options]
778 Start an early, polled-mode console on the 8250/16550
779 UART at the specified I/O port or MMIO address,
780 switching to the matching ttyS device later.
781 MMIO inter-register address stride is either 8-bit
782 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
783 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
784 to be equivalent to 'mmio'. 'options' are specified in
785 the same format described for ttyS above; if unspecified,
786 the h/w is not re-initialized.
788 hvc<n> Use the hypervisor console device <n>. This is for
789 both Xen and PowerPC hypervisors.
792 Use to disable console output, i.e., to have kernel
793 console messages discarded.
794 This must be the only console= parameter used on the
797 If the device connected to the port is not a TTY but a braille
798 device, prepend "brl," before the device type, for instance
800 For now, only VisioBraille is supported.
803 [KNL] Change console messages format
805 By default we print messages on consoles in
806 "[time stamp] text\n" format (time stamp may not be
807 printed, depending on CONFIG_PRINTK_TIME or
808 `printk_time' param).
810 Switch to syslog format: "<%u>[time stamp] text\n"
811 IOW, each message will have a facility and loglevel
812 prefix. The format is similar to one used by syslog()
813 syscall, or to executing "dmesg -S --raw" or to reading
816 consoleblank= [KNL] The console blank (screen saver) timeout in
817 seconds. A value of 0 disables the blank timer.
821 [KNL] Change the default value for
822 /proc/<pid>/coredump_filter.
823 See also Documentation/filesystems/proc.rst.
825 coresight_cpu_debug.enable
828 Enable/disable the CPU sampling based debugging.
829 0: default value, disable debugging
830 1: enable debugging at boot time
832 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
834 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
836 cpuidle.off=1 [CPU_IDLE]
837 disable the cpuidle sub-system
840 [CPU_IDLE] Name of the cpuidle governor to use.
842 cpufreq.off=1 [CPU_FREQ]
843 disable the cpufreq sub-system
845 cpufreq.default_governor=
846 [CPU_FREQ] Name of the default cpufreq governor or
847 policy to use. This governor must be registered in the
848 kernel before the cpufreq driver probes.
851 [X86] Delay for N microsec between assert and de-assert
852 of APIC INIT to start processors. This delay occurs
853 on every CPU online, such as boot, and resume from suspend.
857 [SMP] Enable/disable parallel bringup of secondary CPUs
859 Default is enabled if CONFIG_HOTPLUG_PARALLEL=y. Otherwise
860 the parameter has no effect.
862 crash_kexec_post_notifiers
863 Run kdump after running panic-notifiers and dumping
864 kmsg. This only for the users who doubt kdump always
865 succeeds in any situation.
866 Note that this also increases risks of kdump failure,
867 because some panic notifiers can make the crashed
868 kernel more unstable.
870 crashkernel=size[KMG][@offset[KMG]]
871 [KNL] Using kexec, Linux can switch to a 'crash kernel'
872 upon panic. This parameter reserves the physical
873 memory region [offset, offset + size] for that kernel
874 image. If '@offset' is omitted, then a suitable offset
875 is selected automatically.
876 [KNL, X86-64, ARM64, RISCV] Select a region under 4G first, and
877 fall back to reserve region above 4G when '@offset'
878 hasn't been specified.
879 See Documentation/admin-guide/kdump/kdump.rst for further details.
881 crashkernel=range1:size1[,range2:size2,...][@offset]
882 [KNL] Same as above, but depends on the memory
883 in the running system. The syntax of range is
884 start-[end] where start and end are both
885 a memory unit (amount[KMG]). See also
886 Documentation/admin-guide/kdump/kdump.rst for an example.
888 crashkernel=size[KMG],high
889 [KNL, X86-64, ARM64, RISCV] range could be above 4G.
890 Allow kernel to allocate physical memory region from top,
891 so could be above 4G if system have more than 4G ram
892 installed. Otherwise memory region will be allocated
893 below 4G, if available.
894 It will be ignored if crashkernel=X is specified.
895 crashkernel=size[KMG],low
896 [KNL, X86-64, ARM64, RISCV] range under 4G. When crashkernel=X,high
897 is passed, kernel could allocate physical memory region
898 above 4G, that cause second kernel crash on system
899 that require some amount of low memory, e.g. swiotlb
900 requires at least 64M+32K low memory, also enough extra
901 low memory is needed to make sure DMA buffers for 32-bit
902 devices won't run out. Kernel would try to allocate
903 default size of memory below 4G automatically. The default
904 size is platform dependent.
905 --> x86: max(swiotlb_size_or_default() + 8MiB, 256MiB)
908 This one lets the user specify own low range under 4G
909 for second kernel instead.
910 0: to disable low allocation.
911 It will be ignored when crashkernel=X,high is not used
912 or memory reserved is below 4G.
915 [KNL] Disable crypto self-tests
920 cs89x0_media= [HW,NET]
921 Format: { rj45 | aui | bnc }
923 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU
924 function call handling. When switched on,
925 additional debug data is printed to the console
926 in case a hanging CPU is detected, and that
927 CPU is pinged again in order to try to resolve
928 the hang situation. The default value of this
929 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT
933 See header of drivers/s390/block/dasd_devmap.c.
935 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
936 (one device per port)
937 Format: <port#>,<type>
938 See also Documentation/input/devices/joystick-parport.rst
940 debug [KNL] Enable kernel debugging (events log level).
943 [KNL] Enable printing [hashed] pointers early in the
944 boot sequence. If enabled, we use a weak hash instead
945 of siphash to hash pointers. Use this option if you are
946 seeing instances of '(___ptrval___)') and need to see a
947 value (hashed pointer) instead. Cryptographically
948 insecure, please do not use on production kernels.
951 [KNL] verbose locking self-tests
953 Print debugging info while doing the locking API
955 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
956 (no extra messages), setting it to -1 (all bits set)
957 will print _a_lot_ more information - normally only
958 useful to lockdep developers.
960 debug_objects [KNL] Enable object debugging
962 debug_guardpage_minorder=
963 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
964 parameter allows control of the order of pages that will
965 be intentionally kept free (and hence protected) by the
966 buddy allocator. Bigger value increase the probability
967 of catching random memory corruption, but reduce the
968 amount of memory for normal system use. The maximum
969 possible value is MAX_ORDER/2. Setting this parameter
970 to 1 or 2 should be enough to identify most random
971 memory corruption problems caused by bugs in kernel or
972 driver code when a CPU writes to (or reads from) a
973 random memory location. Note that there exists a class
974 of memory corruptions problems caused by buggy H/W or
975 F/W or by drivers badly programming DMA (basically when
976 memory is written at bus level and the CPU MMU is
977 bypassed) which are not detectable by
978 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
979 tracking down these problems.
982 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
983 enables the feature at boot time. By default, it is
984 disabled and the system will work mostly the same as a
985 kernel built without CONFIG_DEBUG_PAGEALLOC.
986 Note: to get most of debug_pagealloc error reports, it's
987 useful to also enable the page_owner functionality.
988 on: enable the feature
990 debugfs= [KNL] This parameter enables what is exposed to userspace
991 and debugfs internal clients.
992 Format: { on, no-mount, off }
993 on: All functions are enabled.
995 Filesystem is not registered but kernel clients can
996 access APIs and a crashkernel can be used to read
997 its content. There is nothing to mount.
998 off: Filesystem is not registered and clients
999 get a -EPERM as result when trying to register files
1000 or directories within debugfs.
1001 This is equivalent of the runtime functionality if
1002 debugfs was not enabled in the kernel at all.
1003 Default value is set in build-time with a kernel configuration.
1005 debugpat [X86] Enable PAT debugging
1008 [HW] The size of the default HugeTLB page. This is
1009 the size represented by the legacy /proc/ hugepages
1010 APIs. In addition, this is the default hugetlb size
1011 used for shmget(), mmap() and mounting hugetlbfs
1012 filesystems. If not specified, defaults to the
1013 architecture's default huge page size. Huge page
1014 sizes are architecture dependent. See also
1015 Documentation/admin-guide/mm/hugetlbpage.rst.
1018 deferred_probe_timeout=
1019 [KNL] Debugging option to set a timeout in seconds for
1020 deferred probe to give up waiting on dependencies to
1021 probe. Only specific dependencies (subsystems or
1022 drivers) that have opted in will be ignored. A timeout
1023 of 0 will timeout at the end of initcalls. If the time
1024 out hasn't expired, it'll be restarted by each
1025 successful driver registration. This option will also
1026 dump out devices still on the deferred probe list after
1029 delayacct [KNL] Enable per-task delay accounting
1031 dell_smm_hwmon.ignore_dmi=
1032 [HW] Continue probing hardware even if DMI data
1033 indicates that the driver is running on unsupported
1036 dell_smm_hwmon.force=
1037 [HW] Activate driver even if SMM BIOS signature does
1038 not match list of supported models and enable otherwise
1039 blacklisted features.
1041 dell_smm_hwmon.power_status=
1042 [HW] Report power status in /proc/i8k
1043 (disabled by default).
1045 dell_smm_hwmon.restricted=
1046 [HW] Allow controlling fans only if SYS_ADMIN
1049 dell_smm_hwmon.fan_mult=
1050 [HW] Factor to multiply fan speed with.
1052 dell_smm_hwmon.fan_max=
1053 [HW] Maximum configurable fan speed.
1056 Format: { on | off | def_only | inf_only | always }
1057 on: s390 zlib hardware support for compression on
1058 level 1 and decompression (default)
1059 off: No s390 zlib hardware support
1060 def_only: s390 zlib hardware support for deflate
1061 only (compression on level 1)
1062 inf_only: s390 zlib hardware support for inflate
1063 only (decompression)
1064 always: Same as 'on' but ignores the selected compression
1065 level always using hardware support (used for debugging)
1067 dhash_entries= [KNL]
1068 Set number of hash buckets for dentry cache.
1070 disable_1tb_segments [PPC]
1071 Disables the use of 1TB hash page table segments. This
1072 causes the kernel to fall back to 256MB segments which
1073 can be useful when debugging issues that require an SLB
1077 See Documentation/networking/ipv6.rst.
1080 Disable RADIX MMU mode on POWER9
1083 Disable TLBIE instruction. Currently does not work
1084 with KVM, with HASH MMU, or with coherent accelerators.
1086 disable_cpu_apicid= [X86,APIC,SMP]
1088 The number of initial APIC ID for the
1089 corresponding CPU to be disabled at boot,
1090 mostly used for the kdump 2nd kernel to
1091 disable BSP to wake up multiple CPUs without
1092 causing system reset or hang due to sending
1093 INIT from AP to BSP.
1095 disable_ddw [PPC/PSERIES]
1096 Disable Dynamic DMA Window support. Use this
1097 to workaround buggy firmware.
1099 disable_ipv6= [IPV6]
1100 See Documentation/networking/ipv6.rst.
1102 disable_mtrr_cleanup [X86]
1103 The kernel tries to adjust MTRR layout from continuous
1104 to discrete, to make X server driver able to add WB
1105 entry later. This parameter disables that.
1107 disable_mtrr_trim [X86, Intel and AMD only]
1108 By default the kernel will trim any uncacheable
1109 memory out of your available memory pool based on
1110 MTRR settings. This parameter disables that behavior,
1111 possibly causing your machine to run very slowly.
1113 disable_timer_pin_1 [X86]
1114 Disable PIN 1 of APIC timer
1115 Can be useful to work around chipset bugs.
1117 dis_ucode_ldr [X86] Disable the microcode loader.
1119 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1120 this option disables the debugging code at boot.
1122 dma_debug_entries=<number>
1123 This option allows to tune the number of preallocated
1124 entries for DMA-API debugging code. One entry is
1125 required per DMA-API allocation. Use this if the
1126 DMA-API debugging code disables itself because the
1127 architectural default is too low.
1129 dma_debug_driver=<driver_name>
1130 With this option the DMA-API debugging driver
1131 filter feature can be enabled at boot time. Just
1132 pass the driver to filter for as the parameter.
1133 The filter can be disabled or changed to another
1134 driver later using sysfs.
1136 reg_file_data_sampling=
1137 [X86] Controls mitigation for Register File Data
1138 Sampling (RFDS) vulnerability. RFDS is a CPU
1139 vulnerability which may allow userspace to infer
1140 kernel data values previously stored in floating point
1141 registers, vector registers, or integer registers.
1142 RFDS only affects Intel Atom processors.
1144 on: Turns ON the mitigation.
1145 off: Turns OFF the mitigation.
1147 This parameter overrides the compile time default set
1148 by CONFIG_MITIGATION_RFDS. Mitigation cannot be
1149 disabled when other VERW based mitigations (like MDS)
1150 are enabled. In order to disable RFDS mitigation all
1151 VERW based mitigations need to be disabled.
1154 Documentation/admin-guide/hw-vuln/reg-file-data-sampling.rst
1156 driver_async_probe= [KNL]
1157 List of driver names to be probed asynchronously. *
1158 matches with all driver names. If * is specified, the
1159 rest of the listed driver names are those that will NOT
1161 Format: <driver_name1>,<driver_name2>...
1163 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1164 Broken monitors, graphic adapters, KVMs and EDIDless
1165 panels may send no or incorrect EDID data sets.
1166 This parameter allows to specify an EDID data sets
1167 in the /lib/firmware directory that are used instead.
1168 Generic built-in EDID data sets are used, if one of
1169 edid/1024x768.bin, edid/1280x1024.bin,
1170 edid/1680x1050.bin, or edid/1920x1080.bin is given
1171 and no file with the same name exists. Details and
1172 instructions how to build your own EDID data are
1173 available in Documentation/admin-guide/edid.rst. An EDID
1174 data set will only be used for a particular connector,
1175 if its name and a colon are prepended to the EDID
1176 name. Each connector may use a unique EDID data
1177 set by separating the files with a comma. An EDID
1178 data set with no connector name will be used for
1179 any connectors not explicitly specified.
1184 Format: {"off" | "known"}
1185 Control how the dt_cpu_ftrs device-tree binding is
1186 used for CPU feature discovery and setup (if it
1188 off: Do not use it, fall back to legacy cpu table.
1189 known: Do not pass through unknown features to guests
1190 or userspace, only those that the kernel is aware of.
1192 dump_apple_properties [X86]
1193 Dump name and content of EFI device properties on
1194 x86 Macs. Useful for driver authors to determine
1195 what data is available or for reverse-engineering.
1197 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1198 <module>.dyndbg[="val"]
1199 Enable debug messages at boot time. See
1200 Documentation/admin-guide/dynamic-debug-howto.rst
1203 early_ioremap_debug [KNL]
1204 Enable debug messages in early_ioremap support. This
1205 is useful for tracking down temporary early mappings
1206 which are not unmapped.
1208 earlycon= [KNL] Output early console device and options.
1210 When used with no options, the early console is
1211 determined by stdout-path property in device tree's
1212 chosen node or the ACPI SPCR table if supported by
1215 cdns,<addr>[,options]
1216 Start an early, polled-mode console on a Cadence
1217 (xuartps) serial port at the specified address. Only
1218 supported option is baud rate. If baud rate is not
1219 specified, the serial port must already be setup and
1222 uart[8250],io,<addr>[,options[,uartclk]]
1223 uart[8250],mmio,<addr>[,options[,uartclk]]
1224 uart[8250],mmio32,<addr>[,options[,uartclk]]
1225 uart[8250],mmio32be,<addr>[,options[,uartclk]]
1226 uart[8250],0x<addr>[,options]
1227 Start an early, polled-mode console on the 8250/16550
1228 UART at the specified I/O port or MMIO address.
1229 MMIO inter-register address stride is either 8-bit
1230 (mmio) or 32-bit (mmio32 or mmio32be).
1231 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1232 to be equivalent to 'mmio'. 'options' are specified
1233 in the same format described for "console=ttyS<n>"; if
1234 unspecified, the h/w is not initialized. 'uartclk' is
1235 the uart clock frequency; if unspecified, it is set
1236 to 'BASE_BAUD' * 16.
1240 Start an early, polled-mode console on a pl011 serial
1241 port at the specified address. The pl011 serial port
1242 must already be setup and configured. Options are not
1243 yet supported. If 'mmio32' is specified, then only
1244 the driver will use only 32-bit accessors to read/write
1245 the device registers.
1248 Start an early console on a litex serial port at the
1249 specified address. The serial port must already be
1250 setup and configured. Options are not yet supported.
1253 Start an early, polled-mode console on a meson serial
1254 port at the specified address. The serial port must
1255 already be setup and configured. Options are not yet
1259 Start an early, polled-mode console on an msm serial
1260 port at the specified address. The serial port
1261 must already be setup and configured. Options are not
1264 msm_serial_dm,<addr>
1265 Start an early, polled-mode console on an msm serial
1266 dm port at the specified address. The serial port
1267 must already be setup and configured. Options are not
1271 Start an early, polled-mode console on a serial port
1272 of an Actions Semi SoC, such as S500 or S900, at the
1273 specified address. The serial port must already be
1274 setup and configured. Options are not yet supported.
1277 Start an early, polled-mode console on a serial port
1278 of an RDA Micro SoC, such as RDA8810PL, at the
1279 specified address. The serial port must already be
1280 setup and configured. Options are not yet supported.
1283 Use RISC-V SBI (Supervisor Binary Interface) for early
1286 smh Use ARM semihosting calls for early console.
1294 Use early console provided by serial driver available
1295 on Samsung SoCs, requires selecting proper type and
1296 a correct base address of the selected UART port. The
1297 serial port must already be setup and configured.
1298 Options are not yet supported.
1301 Start an early, polled-mode console on a lantiq serial
1302 (lqasc) port at the specified address. The serial port
1303 must already be setup and configured. Options are not
1308 Use early console provided by Freescale LP UART driver
1309 found on Freescale Vybrid and QorIQ LS1021A processors.
1310 A valid base address must be provided, and the serial
1311 port must already be setup and configured.
1315 Start an early, polled-mode, output-only console on the
1316 Freescale i.MX UART at the specified address. The UART
1317 must already be setup and configured.
1320 Start an early, polled-mode console on the
1321 Armada 3700 serial port at the specified
1322 address. The serial port must already be setup
1323 and configured. Options are not yet supported.
1326 Start an early, polled-mode console on a Qualcomm
1327 Generic Interface (GENI) based serial port at the
1328 specified address. The serial port must already be
1329 setup and configured. Options are not yet supported.
1332 Start an early, unaccelerated console on the EFI
1333 memory mapped framebuffer (if available). On cache
1334 coherent non-x86 systems that use system memory for
1335 the framebuffer, pass the 'ram' option so that it is
1336 mapped with the correct attributes.
1339 Use early console provided by Freescale LINFlexD UART
1340 serial driver for NXP S32V234 SoCs. A valid base
1341 address must be provided, and the serial port must
1342 already be setup and configured.
1344 earlyprintk= [X86,SH,ARM,M68k,S390]
1348 earlyprintk=serial[,ttySn[,baudrate]]
1349 earlyprintk=serial[,0x...[,baudrate]]
1350 earlyprintk=ttySn[,baudrate]
1351 earlyprintk=dbgp[debugController#]
1352 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1353 earlyprintk=xdbc[xhciController#]
1355 earlyprintk is useful when the kernel crashes before
1356 the normal console is initialized. It is not enabled by
1357 default because it has some cosmetic problems.
1359 Append ",keep" to not disable it when the real console
1362 Only one of vga, serial, or usb debug port can
1365 Currently only ttyS0 and ttyS1 may be specified by
1366 name. Other I/O ports may be explicitly specified
1367 on some architectures (x86 and arm at least) by
1368 replacing ttySn with an I/O port address, like this:
1369 earlyprintk=serial,0x1008,115200
1370 You can find the port for a given device in
1371 /proc/tty/driver/serial:
1372 2: uart:ST16650V2 port:00001008 irq:18 ...
1374 Interaction with the standard serial driver is not
1377 The VGA output is eventually overwritten by
1380 The xen option can only be used in Xen domains.
1382 The sclp output can only be used on s390.
1384 The optional "force" to "pciserial" enables use of a
1385 PCI device even when its classcode is not of the
1388 edac_report= [HW,EDAC] Control how to report EDAC event
1389 Format: {"on" | "off" | "force"}
1390 on: enable EDAC to report H/W event. May be overridden
1391 by other higher priority error reporting module.
1392 off: disable H/W event reporting through EDAC.
1393 force: enforce the use of EDAC to report H/W event.
1397 Format: {"off" | "on" | "skip[mbr]"}
1400 Format: { "debug", "disable_early_pci_dma",
1401 "nochunk", "noruntime", "nosoftreserve",
1402 "novamap", "no_disable_early_pci_dma" }
1403 debug: enable misc debug output.
1404 disable_early_pci_dma: disable the busmaster bit on all
1405 PCI bridges while in the EFI boot stub.
1406 nochunk: disable reading files in "chunks" in the EFI
1407 boot stub, as chunking can cause problems with some
1408 firmware implementations.
1409 noruntime : disable EFI runtime services support
1410 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1411 attribute may cause the kernel to reserve the
1412 memory range for a memory mapping driver to
1413 claim. Specify efi=nosoftreserve to disable this
1414 reservation and treat the memory by its base type
1415 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1416 novamap: do not call SetVirtualAddressMap().
1417 no_disable_early_pci_dma: Leave the busmaster bit set
1418 on all PCI bridges while in the EFI boot stub
1420 efi_no_storage_paranoia [EFI; X86]
1421 Using this parameter you can use more than 50% of
1422 your efi variable storage. Use this parameter only if
1423 you are really sure that your UEFI does sane gc and
1424 fulfills the spec otherwise your board may brick.
1426 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1427 Add arbitrary attribute to specific memory range by
1428 updating original EFI memory map.
1429 Region of memory which aa attribute is added to is
1432 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1433 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1434 attribute is added to range 0x100000000-0x180000000 and
1435 0x10a0000000-0x1120000000.
1437 If efi_fake_mem=8G@9G:0x40000 is specified, the
1438 EFI_MEMORY_SP(0x40000) attribute is added to
1439 range 0x240000000-0x43fffffff.
1441 Using this parameter you can do debugging of EFI memmap
1442 related features. For example, you can do debugging of
1443 Address Range Mirroring feature even if your box
1444 doesn't support it, or mark specific memory as
1447 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1448 that is to be dynamically loaded by Linux. If there are
1449 multiple variables with the same name but with different
1450 vendor GUIDs, all of them will be loaded. See
1451 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1454 eisa_irq_edge= [PARISC,HW]
1455 See header of drivers/parisc/eisa.c.
1457 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1460 This is designed to be used in conjunction with
1461 the boot argument: earlyprintk=vga
1463 This parameter works in place of the kgdboc parameter
1464 but can only be used if the backing tty is available
1465 very early in the boot process. For early debugging
1466 via a serial port see kgdboc_earlycon instead.
1469 See comment before function elanfreq_setup() in
1470 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1472 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1473 Specifies physical address of start of kernel core
1474 image elf header and optionally the size. Generally
1475 kexec loader will pass this option to capture kernel.
1476 See Documentation/admin-guide/kdump/kdump.rst for details.
1478 enable_mtrr_cleanup [X86]
1479 The kernel tries to adjust MTRR layout from continuous
1480 to discrete, to make X server driver able to add WB
1481 entry later. This parameter enables that.
1483 enable_timer_pin_1 [X86]
1484 Enable PIN 1 of APIC timer
1485 Can be useful to work around chipset bugs
1486 (in particular on some ATI chipsets).
1487 The kernel tries to set a reasonable default.
1489 enforcing= [SELINUX] Set initial enforcing status.
1491 See security/selinux/Kconfig help text.
1492 0 -- permissive (log only, no denials).
1493 1 -- enforcing (deny and log).
1495 Value can be changed at runtime via
1496 /sys/fs/selinux/enforce.
1499 Disable Error Record Serialization Table (ERST)
1502 ether= [HW,NET] Ethernet cards parameters
1503 This option is obsoleted by the "netdev=" option, which
1504 has equivalent usage. See its documentation for details.
1508 Permit 'security.evm' to be updated regardless of
1509 current integrity status.
1511 early_page_ext [KNL] Enforces page_ext initialization to earlier
1512 stages so cover more early boot allocations.
1513 Please note that as side effect some optimizations
1514 might be disabled to achieve that (e.g. parallelized
1515 memory initialization is disabled) so the boot process
1516 might take longer, especially on systems with a lot of
1517 memory. Available with CONFIG_PAGE_EXTENSION=y.
1522 fail_make_request=[KNL]
1523 General fault injection mechanism.
1524 Format: <interval>,<probability>,<space>,<times>
1525 See also Documentation/fault-injection/.
1528 Format: { initns | none }
1529 See Documentation/admin-guide/sysctl/net.rst for
1530 fb_tunnels_only_for_init_ns
1533 See Documentation/admin-guide/blockdev/floppy.rst.
1535 force_pal_cache_flush
1536 [IA-64] Avoid check_sal_cache_flush which may hang on
1537 buggy SAL_CACHE_FLUSH implementations. Using this
1538 parameter will force ia64_sal_cache_flush to call
1539 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
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 [FTRACE] will set and start the specified tracer
1550 as early as possible in order to facilitate early
1553 ftrace_boot_snapshot
1554 [FTRACE] On boot up, a snapshot will be taken of the
1555 ftrace ring buffer that can be read at:
1556 /sys/kernel/tracing/snapshot.
1557 This is useful if you need tracing information from kernel
1558 boot up that is likely to be overridden by user space
1559 start up functionality.
1561 Optionally, the snapshot can also be defined for a tracing
1562 instance that was created by the trace_instance= command
1565 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo
1567 The above will cause the "foo" tracing instance to trigger
1568 a snapshot at the end of boot up.
1570 ftrace_dump_on_oops[=orig_cpu]
1571 [FTRACE] will dump the trace buffers on oops.
1572 If no parameter is passed, ftrace will dump
1573 buffers of all CPUs, but if you pass orig_cpu, it will
1574 dump only the buffer of the CPU that triggered the
1577 ftrace_filter=[function-list]
1578 [FTRACE] Limit the functions traced by the function
1579 tracer at boot up. function-list is a comma-separated
1580 list of functions. This list can be changed at run
1581 time by the set_ftrace_filter file in the debugfs
1584 ftrace_notrace=[function-list]
1585 [FTRACE] Do not trace the functions specified in
1586 function-list. This list can be changed at run time
1587 by the set_ftrace_notrace file in the debugfs
1590 ftrace_graph_filter=[function-list]
1591 [FTRACE] Limit the top level callers functions traced
1592 by the function graph tracer at boot up.
1593 function-list is a comma-separated list of functions
1594 that can be changed at run time by the
1595 set_graph_function file in the debugfs tracing directory.
1597 ftrace_graph_notrace=[function-list]
1598 [FTRACE] Do not trace from the functions specified in
1599 function-list. This list is a comma-separated list of
1600 functions that can be changed at run time by the
1601 set_graph_notrace file in the debugfs tracing directory.
1603 ftrace_graph_max_depth=<uint>
1604 [FTRACE] Used with the function graph tracer. This is
1605 the max depth it will trace into a function. This value
1606 can be changed at run time by the max_graph_depth file
1607 in the tracefs tracing directory. default: 0 (no limit)
1609 fw_devlink= [KNL] Create device links between consumer and supplier
1610 devices by scanning the firmware to infer the
1611 consumer/supplier relationships. This feature is
1612 especially useful when drivers are loaded as modules as
1613 it ensures proper ordering of tasks like device probing
1614 (suppliers first, then consumers), supplier boot state
1615 clean up (only after all consumers have probed),
1616 suspend/resume & runtime PM (consumers first, then
1618 Format: { off | permissive | on | rpm }
1619 off -- Don't create device links from firmware info.
1620 permissive -- Create device links from firmware info
1621 but use it only for ordering boot state clean
1622 up (sync_state() calls).
1623 on -- Create device links from firmware info and use it
1624 to enforce probe and suspend/resume ordering.
1625 rpm -- Like "on", but also use to order runtime PM.
1627 fw_devlink.strict=<bool>
1628 [KNL] Treat all inferred dependencies as mandatory
1629 dependencies. This only applies for fw_devlink=on|rpm.
1632 fw_devlink.sync_state =
1633 [KNL] When all devices that could probe have finished
1634 probing, this parameter controls what to do with
1635 devices that haven't yet received their sync_state()
1637 Format: { strict | timeout }
1638 strict -- Default. Continue waiting on consumers to
1640 timeout -- Give up waiting on consumers and call
1641 sync_state() on any devices that haven't yet
1642 received their sync_state() calls after
1643 deferred_probe_timeout has expired or by
1644 late_initcall() if !CONFIG_MODULES.
1647 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1648 support via parallel port (up to 5 devices per port)
1649 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1650 See also Documentation/input/devices/joystick-parport.rst
1654 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1658 gather_data_sampling=
1659 [X86,INTEL] Control the Gather Data Sampling (GDS)
1662 Gather Data Sampling is a hardware vulnerability which
1663 allows unprivileged speculative access to data which was
1664 previously stored in vector registers.
1666 This issue is mitigated by default in updated microcode.
1667 The mitigation may have a performance impact but can be
1668 disabled. On systems without the microcode mitigation
1669 disabling AVX serves as a mitigation.
1671 force: Disable AVX to mitigate systems without
1672 microcode mitigation. No effect if the microcode
1673 mitigation is present. Known to cause crashes in
1674 userspace with buggy AVX enumeration.
1676 off: Disable GDS mitigation.
1678 gcov_persist= [GCOV] When non-zero (default), profiling data for
1679 kernel modules is saved and remains accessible via
1680 debugfs, even when the module is unloaded/reloaded.
1681 When zero, profiling data is discarded and associated
1682 debugfs files are removed at module unload time.
1684 goldfish [X86] Enable the goldfish android emulator platform.
1685 Don't use this when you are not running on the
1688 gpio-mockup.gpio_mockup_ranges
1689 [HW] Sets the ranges of gpiochip of for this device.
1690 Format: <start1>,<end1>,<start2>,<end2>...
1691 gpio-mockup.gpio_mockup_named_lines
1692 [HW] Let the driver know GPIO lines should be named.
1694 gpt [EFI] Forces disk with valid GPT signature but
1695 invalid Protective MBR to be treated as GPT. If the
1696 primary GPT is corrupted, it enables the backup/alternate
1697 GPT to be used instead.
1699 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1700 the "Enable 0" bit of the configuration register.
1703 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1704 the "Enable 0" bit of the configuration register.
1707 grcan.select= [HW] Select which physical interface to use.
1710 grcan.txsize= [HW] Sets the size of the tx buffer.
1711 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1713 grcan.rxsize= [HW] Sets the size of the rx buffer.
1714 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1718 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1719 hardening is enabled for this boot. Hardened
1720 usercopy checking is used to protect the kernel
1721 from reading or writing beyond known memory
1722 allocation boundaries as a proactive defense
1723 against bounds-checking flaws in the kernel's
1724 copy_to_user()/copy_from_user() interface.
1725 on Perform hardened usercopy checks (default).
1726 off Disable hardened usercopy checks.
1728 hardlockup_all_cpu_backtrace=
1729 [KNL] Should the hard-lockup detector generate
1730 backtraces on all cpus.
1733 hashdist= [KNL,NUMA] Large hashes allocated during boot
1734 are distributed across NUMA nodes. Defaults on
1735 for 64-bit NUMA, off otherwise.
1736 Format: 0 | 1 (for off | on)
1738 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1740 hd= [EIDE] (E)IDE hard drive subsystem geometry
1741 Format: <cyl>,<head>,<sect>
1744 Disable Hardware Error Source Table (HEST) support;
1745 corresponding firmware-first mode error processing
1746 logic will be disabled.
1748 hibernate= [HIBERNATION]
1749 noresume Don't check if there's a hibernation image
1750 present during boot.
1751 nocompress Don't compress/decompress hibernation images.
1752 no Disable hibernation and resume.
1753 protect_image Turn on image protection during restoration
1754 (that will set all pages holding image data
1755 during restoration read-only).
1757 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1758 size of <nn>. This works even on boxes that have no
1759 highmem otherwise. This also works to reduce highmem
1760 size on bigger boxes.
1762 highres= [KNL] Enable/disable high resolution timer mode.
1763 Valid parameters: "on", "off"
1768 hostname= [KNL] Set the hostname (aka UTS nodename).
1770 This allows setting the system's hostname during early
1771 startup. This sets the name returned by gethostname.
1772 Using this parameter to set the hostname makes it
1773 possible to ensure the hostname is correctly set before
1774 any userspace processes run, avoiding the possibility
1775 that a process may call gethostname before the hostname
1776 has been explicitly set, resulting in the calling
1777 process getting an incorrect result. The string must
1778 not exceed the maximum allowed hostname length (usually
1779 64 characters) and will be truncated otherwise.
1781 hpet= [X86-32,HPET] option to control HPET usage
1782 Format: { enable (default) | disable | force |
1784 disable: disable HPET and use PIT instead
1785 force: allow force enabled of undocumented chips (ICH4,
1787 verbose: show contents of HPET registers during setup
1789 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1790 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1792 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1793 If this follows hugepagesz (below), it specifies
1794 the number of pages of hugepagesz to be allocated.
1795 If this is the first HugeTLB parameter on the command
1796 line, it specifies the number of pages to allocate for
1797 the default huge page size. If using node format, the
1798 number of pages to allocate per-node can be specified.
1799 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1800 Format: <integer> or (node format)
1801 <node>:<integer>[,<node>:<integer>]
1804 [HW] The size of the HugeTLB pages. This is used in
1805 conjunction with hugepages (above) to allocate huge
1806 pages of a specific size at boot. The pair
1807 hugepagesz=X hugepages=Y can be specified once for
1808 each supported huge page size. Huge page sizes are
1809 architecture dependent. See also
1810 Documentation/admin-guide/mm/hugetlbpage.rst.
1813 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1814 of gigantic hugepages. Or using node format, the size
1815 of a CMA area per node can be specified.
1816 Format: nn[KMGTPE] or (node format)
1817 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1819 Reserve a CMA area of given size and allocate gigantic
1820 hugepages using the CMA allocator. If enabled, the
1821 boot-time allocation of gigantic hugepages is skipped.
1823 hugetlb_free_vmemmap=
1824 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1826 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1827 Allows heavy hugetlb users to free up some more
1828 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1829 Format: { on | off (default) }
1834 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1837 Note that the vmemmap pages may be allocated from the added
1838 memory block itself when memory_hotplug.memmap_on_memory is
1839 enabled, those vmemmap pages cannot be optimized even if this
1840 feature is enabled. Other vmemmap pages not allocated from
1841 the added memory block itself do not be affected.
1844 [KNL] Should the hung task detector generate panics.
1847 A value of 1 instructs the kernel to panic when a
1848 hung task is detected. The default value is controlled
1849 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1850 option. The value selected by this boot parameter can
1851 be changed later by the kernel.hung_task_panic sysctl.
1853 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1854 terminal devices. Valid values: 0..8
1855 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1856 If specified, z/VM IUCV HVC accepts connections
1857 from listed z/VM user IDs only.
1859 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1860 which allow the hypervisor to 'idle' the
1861 guest on lock contention.
1863 i2c_bus= [HW] Override the default board specific I2C bus speed
1864 or register an additional I2C bus that is not
1865 registered from board initialization code.
1869 i8042.debug [HW] Toggle i8042 debug mode
1870 i8042.unmask_kbd_data
1871 [HW] Enable printing of interrupt data from the KBD port
1872 (disabled by default, and as a pre-condition
1873 requires that i8042.debug=1 be enabled)
1874 i8042.direct [HW] Put keyboard port into non-translated mode
1875 i8042.dumbkbd [HW] Pretend that controller can only read data from
1876 keyboard and cannot control its state
1877 (Don't attempt to blink the leds)
1878 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1879 i8042.nokbd [HW] Don't check/create keyboard port
1880 i8042.noloop [HW] Disable the AUX Loopback command while probing
1882 i8042.nomux [HW] Don't check presence of an active multiplexing
1884 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1886 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1887 i8042.reset [HW] Reset the controller during init, cleanup and
1888 suspend-to-ram transitions, only during s2r
1889 transitions, or never reset
1890 Format: { 1 | Y | y | 0 | N | n }
1891 1, Y, y: always reset controller
1892 0, N, n: don't ever reset controller
1893 Default: only on s2r transitions on x86; most other
1894 architectures force reset to be always executed
1895 i8042.unlock [HW] Unlock (ignore) the keylock
1896 i8042.kbdreset [HW] Reset device connected to KBD port
1898 [HW] Allow deferred probing upon i8042 probe errors
1902 i915.invert_brightness=
1903 [DRM] Invert the sense of the variable that is used to
1904 set the brightness of the panel backlight. Normally a
1905 brightness value of 0 indicates backlight switched off,
1906 and the maximum of the brightness value sets the backlight
1907 to maximum brightness. If this parameter is set to 0
1908 (default) and the machine requires it, or this parameter
1909 is set to 1, a brightness value of 0 sets the backlight
1910 to maximum brightness, and the maximum of the brightness
1911 value switches the backlight off.
1912 -1 -- never invert brightness
1913 0 -- machine default
1914 1 -- force brightness inversion
1917 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1921 Format: idle=poll, idle=halt, idle=nomwait
1922 Poll forces a polling idle loop that can slightly
1923 improve the performance of waking up a idle CPU, but
1924 will use a lot of power and make the system run hot.
1926 idle=halt: Halt is forced to be used for CPU idle.
1927 In such case C2/C3 won't be used again.
1928 idle=nomwait: Disable mwait for CPU C-states
1932 Allow force disabling of Shared Virtual Memory (SVA)
1933 support for the idxd driver. By default it is set to
1936 idxd.tc_override= [HW]
1938 Allow override of default traffic class configuration
1939 for the device. By default it is set to false (0).
1941 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1942 Format: { strict | legacy | 2008 | relaxed }
1945 Choose which programs will be accepted for execution
1946 based on the IEEE 754 NaN encoding(s) supported by
1947 the FPU and the NaN encoding requested with the value
1948 of an ELF file header flag individually set by each
1949 binary. Hardware implementations are permitted to
1950 support either or both of the legacy and the 2008 NaN
1953 Available settings are as follows:
1954 strict accept binaries that request a NaN encoding
1955 supported by the FPU
1956 legacy only accept legacy-NaN binaries, if supported
1958 2008 only accept 2008-NaN binaries, if supported
1960 relaxed accept any binaries regardless of whether
1961 supported by the FPU
1963 The FPU emulator is always able to support both NaN
1964 encodings, so if no FPU hardware is present or it has
1965 been disabled with 'nofpu', then the settings of
1966 'legacy' and '2008' strap the emulator accordingly,
1967 'relaxed' straps the emulator for both legacy-NaN and
1968 2008-NaN, whereas 'strict' enables legacy-NaN only on
1969 legacy processors and both NaN encodings on MIPS32 or
1972 The setting for ABS.fmt/NEG.fmt instruction execution
1973 mode generally follows that for the NaN encoding,
1974 except where unsupported by hardware.
1976 ignore_loglevel [KNL]
1977 Ignore loglevel setting - this will print /all/
1978 kernel messages to the console. Useful for debugging.
1979 We also add it as printk module parameter, so users
1980 could change it dynamically, usually by
1981 /sys/module/printk/parameters/ignore_loglevel.
1984 Ignore RLIMIT_DATA setting for data mappings,
1985 print warning at first misuse. Can be changed via
1986 /sys/module/kernel/parameters/ignore_rlimit_data.
1988 ihash_entries= [KNL]
1989 Set number of hash buckets for inode cache.
1991 ima_appraise= [IMA] appraise integrity measurements
1992 Format: { "off" | "enforce" | "fix" | "log" }
1995 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1996 The builtin appraise policy appraises all files
1999 ima_canonical_fmt [IMA]
2000 Use the canonical format for the binary runtime
2001 measurements, instead of host native format.
2004 Format: { md5 | sha1 | rmd160 | sha256 | sha384
2008 The list of supported hash algorithms is defined
2009 in crypto/hash_info.h.
2012 The builtin policies to load during IMA setup.
2013 Format: "tcb | appraise_tcb | secure_boot |
2014 fail_securely | critical_data"
2016 The "tcb" policy measures all programs exec'd, files
2017 mmap'd for exec, and all files opened with the read
2018 mode bit set by either the effective uid (euid=0) or
2021 The "appraise_tcb" policy appraises the integrity of
2022 all files owned by root.
2024 The "secure_boot" policy appraises the integrity
2025 of files (eg. kexec kernel image, kernel modules,
2026 firmware, policy, etc) based on file signatures.
2028 The "fail_securely" policy forces file signature
2029 verification failure also on privileged mounted
2030 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
2033 The "critical_data" policy measures kernel integrity
2036 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
2037 Load a policy which meets the needs of the Trusted
2038 Computing Base. This means IMA will measure all
2039 programs exec'd, files mmap'd for exec, and all files
2040 opened for read by uid=0.
2043 Select one of defined IMA measurements template formats.
2044 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
2049 [IMA] Define a custom template format.
2050 Format: { "field1|...|fieldN" }
2052 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
2053 Format: <min_file_size>
2054 Set the minimal file size for using asynchronous hash.
2055 If left unspecified, ahash usage is disabled.
2057 ahash performance varies for different data sizes on
2058 different crypto accelerators. This option can be used
2059 to achieve the best performance for a particular HW.
2061 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
2063 Set hashing buffer size. Default: 4k.
2065 ahash performance varies for different chunk sizes on
2066 different crypto accelerators. This option can be used
2067 to achieve best performance for particular HW.
2071 Run specified binary instead of /sbin/init as init
2074 initcall_debug [KNL] Trace initcalls as they are executed. Useful
2075 for working out where the kernel is dying during
2078 initcall_blacklist= [KNL] Do not execute a comma-separated list of
2079 initcall functions. Useful for debugging built-in
2080 modules and initcalls.
2082 initramfs_async= [KNL]
2085 This parameter controls whether the initramfs
2086 image is unpacked asynchronously, concurrently
2087 with devices being probed and
2088 initialized. This should normally just work,
2089 but as a debugging aid, one can get the
2090 historical behaviour of the initramfs
2091 unpacking being completed before device_ and
2094 initrd= [BOOT] Specify the location of the initial ramdisk
2096 initrdmem= [KNL] Specify a physical address and size from which to
2097 load the initrd. If an initrd is compiled in or
2098 specified in the bootparams, it takes priority over this
2100 Format: ss[KMG],nn[KMG]
2103 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
2106 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2108 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
2110 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2112 init_pkru= [X86] Specify the default memory protection keys rights
2113 register contents for all processes. 0x55555554 by
2114 default (disallow access to all but pkey 0). Can
2115 override in debugfs after boot.
2117 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2120 int_pln_enable [X86] Enable power limit notification interrupt
2122 integrity_audit=[IMA]
2123 Format: { "0" | "1" }
2124 0 -- basic integrity auditing messages. (Default)
2125 1 -- additional integrity auditing messages.
2127 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2129 Enable intel iommu driver.
2131 Disable intel iommu driver.
2132 igfx_off [Default Off]
2133 By default, gfx is mapped as normal device. If a gfx
2134 device has a dedicated DMAR unit, the DMAR unit is
2135 bypassed by not enabling DMAR with this option. In
2136 this case, gfx device will use physical address for
2138 strict [Default Off]
2139 Deprecated, equivalent to iommu.strict=1.
2140 sp_off [Default Off]
2141 By default, super page will be supported if Intel IOMMU
2142 has the capability. With this option, super page will
2145 Enable the Intel IOMMU scalable mode if the hardware
2146 advertises that it has support for the scalable mode
2149 Disallow use of the Intel IOMMU scalable mode.
2150 tboot_noforce [Default Off]
2151 Do not force the Intel IOMMU enabled under tboot.
2152 By default, tboot will force Intel IOMMU on, which
2153 could harm performance of some high-throughput
2154 devices like 40GBit network cards, even if identity
2156 Note that using this option lowers the security
2157 provided by tboot because it makes the system
2158 vulnerable to DMA attacks.
2160 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2161 0 disables intel_idle and fall back on acpi_idle.
2162 1 to 9 specify maximum depth of C-state.
2166 Do not enable intel_pstate as the default
2167 scaling driver for the supported processors
2169 Use intel_pstate driver to bypass the scaling
2170 governors layer of cpufreq and provides it own
2171 algorithms for p-state selection. There are two
2172 P-state selection algorithms provided by
2173 intel_pstate in the active mode: powersave and
2174 performance. The way they both operate depends
2175 on whether or not the hardware managed P-states
2176 (HWP) feature has been enabled in the processor
2177 and possibly on the processor model.
2179 Use intel_pstate as a scaling driver, but configure it
2180 to work with generic cpufreq governors (instead of
2181 enabling its internal governor). This mode cannot be
2182 used along with the hardware-managed P-states (HWP)
2185 Enable intel_pstate on systems that prohibit it by default
2186 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2187 instead of acpi-cpufreq may disable platform features, such
2188 as thermal controls and power capping, that rely on ACPI
2189 P-States information being indicated to OSPM and therefore
2190 should be used with caution. This option does not work with
2191 processors that aren't supported by the intel_pstate driver
2192 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2194 Do not enable hardware P state control (HWP)
2197 Only load intel_pstate on systems which support
2198 hardware P state control (HWP) if available.
2200 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2201 Description Table, specifies preferred power management
2202 profile as "Enterprise Server" or "Performance Server",
2203 then this feature is turned on by default.
2205 Allow per-logical-CPU P-State performance control limits using
2206 cpufreq sysfs interface
2208 intremap= [X86-64, Intel-IOMMU]
2209 on enable Interrupt Remapping (default)
2210 off disable Interrupt Remapping
2211 nosid disable Source ID checking
2213 BIOS x2APIC opt-out request will be ignored
2214 nopost disable Interrupt Posting
2216 iomem= Disable strict checking of access to MMIO memory
2217 strict regions from userspace.
2232 nobypass [PPC/POWERNV]
2233 Disable IOMMU bypass, using IOMMU for PCI devices.
2235 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2236 Format: { "0" | "1" }
2237 0 - Try to allocate a 32-bit DMA address first, before
2238 falling back to the full range if needed.
2239 1 - Allocate directly from the full usable range,
2240 forcing Dual Address Cycle for PCI cards supporting
2241 greater than 32-bit addressing.
2243 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2244 Format: { "0" | "1" }
2246 Request that DMA unmap operations use deferred
2247 invalidation of hardware TLBs, for increased
2248 throughput at the cost of reduced device isolation.
2249 Will fall back to strict mode if not supported by
2250 the relevant IOMMU driver.
2252 DMA unmap operations invalidate IOMMU hardware TLBs
2254 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2255 Note: on x86, strict mode specified via one of the
2256 legacy driver-specific options takes precedence.
2259 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2260 Format: { "0" | "1" }
2261 0 - Use IOMMU translation for DMA.
2262 1 - Bypass the IOMMU for DMA.
2263 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2265 io7= [HW] IO7 for Marvel-based Alpha systems
2266 See comment before marvel_specify_io7 in
2267 arch/alpha/kernel/core_marvel.c.
2269 io_delay= [X86] I/O delay method
2271 Standard port 0x80 based delay
2273 Alternate port 0xed based delay (needed on some systems)
2275 Simple two microseconds delay
2280 See Documentation/admin-guide/nfs/nfsroot.rst.
2282 ipcmni_extend [KNL] Extend the maximum number of unique System V
2283 IPC identifiers from 32,768 to 16,777,216.
2285 irqaffinity= [SMP] Set the default irq affinity mask
2286 The argument is a cpu list, as described above.
2288 irqchip.gicv2_force_probe=
2291 Force the kernel to look for the second 4kB page
2292 of a GICv2 controller even if the memory range
2293 exposed by the device tree is too small.
2295 irqchip.gicv3_nolpi=
2297 Force the kernel to ignore the availability of
2298 LPIs (and by consequence ITSs). Intended for system
2299 that use the kernel as a bootloader, and thus want
2300 to let secondary kernels in charge of setting up
2303 irqchip.gicv3_pseudo_nmi= [ARM64]
2304 Enables support for pseudo-NMIs in the kernel. This
2305 requires the kernel to be built with
2306 CONFIG_ARM64_PSEUDO_NMI.
2309 When an interrupt is not handled search all handlers
2310 for it. Intended to get systems with badly broken
2314 When an interrupt is not handled search all handlers
2315 for it. Also check all handlers each timer
2316 interrupt. Intended to get systems with badly broken
2320 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2322 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2323 [Deprecated - use cpusets instead]
2324 Format: [flag-list,]<cpu-list>
2326 Specify one or more CPUs to isolate from disturbances
2327 specified in the flag list (default: domain):
2330 Disable the tick when a single task runs.
2332 A residual 1Hz tick is offloaded to workqueues, which you
2333 need to affine to housekeeping through the global
2334 workqueue's affinity configured via the
2335 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2336 by using the 'domain' flag described below.
2338 NOTE: by default the global workqueue runs on all CPUs,
2339 so to protect individual CPUs the 'cpumask' file has to
2340 be configured manually after bootup.
2343 Isolate from the general SMP balancing and scheduling
2344 algorithms. Note that performing domain isolation this way
2345 is irreversible: it's not possible to bring back a CPU to
2346 the domains once isolated through isolcpus. It's strongly
2347 advised to use cpusets instead to disable scheduler load
2348 balancing through the "cpuset.sched_load_balance" file.
2349 It offers a much more flexible interface where CPUs can
2350 move in and out of an isolated set anytime.
2352 You can move a process onto or off an "isolated" CPU via
2353 the CPU affinity syscalls or cpuset.
2354 <cpu number> begins at 0 and the maximum value is
2355 "number of CPUs in system - 1".
2359 Isolate from being targeted by managed interrupts
2360 which have an interrupt mask containing isolated
2361 CPUs. The affinity of managed interrupts is
2362 handled by the kernel and cannot be changed via
2363 the /proc/irq/* interfaces.
2365 This isolation is best effort and only effective
2366 if the automatically assigned interrupt mask of a
2367 device queue contains isolated and housekeeping
2368 CPUs. If housekeeping CPUs are online then such
2369 interrupts are directed to the housekeeping CPU
2370 so that IO submitted on the housekeeping CPU
2371 cannot disturb the isolated CPU.
2373 If a queue's affinity mask contains only isolated
2374 CPUs then this parameter has no effect on the
2375 interrupt routing decision, though interrupts are
2376 only delivered when tasks running on those
2377 isolated CPUs submit IO. IO submitted on
2378 housekeeping CPUs has no influence on those
2381 The format of <cpu-list> is described above.
2385 ivrs_ioapic [HW,X86-64]
2386 Provide an override to the IOAPIC-ID<->DEVICE-ID
2387 mapping provided in the IVRS ACPI table.
2388 By default, PCI segment is 0, and can be omitted.
2390 For example, to map IOAPIC-ID decimal 10 to
2391 PCI segment 0x1 and PCI device 00:14.0,
2392 write the parameter as:
2393 ivrs_ioapic=10@0001:00:14.0
2396 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2397 write the parameter as:
2398 ivrs_ioapic[10]=00:14.0
2399 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2400 PCI device 00:14.0 write the parameter as:
2401 ivrs_ioapic[10]=0001:00:14.0
2403 ivrs_hpet [HW,X86-64]
2404 Provide an override to the HPET-ID<->DEVICE-ID
2405 mapping provided in the IVRS ACPI table.
2406 By default, PCI segment is 0, and can be omitted.
2408 For example, to map HPET-ID decimal 10 to
2409 PCI segment 0x1 and PCI device 00:14.0,
2410 write the parameter as:
2411 ivrs_hpet=10@0001:00:14.0
2414 * To map HPET-ID decimal 0 to PCI device 00:14.0
2415 write the parameter as:
2416 ivrs_hpet[0]=00:14.0
2417 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2418 PCI device 00:14.0 write the parameter as:
2419 ivrs_ioapic[10]=0001:00:14.0
2421 ivrs_acpihid [HW,X86-64]
2422 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2423 mapping provided in the IVRS ACPI table.
2424 By default, PCI segment is 0, and can be omitted.
2426 For example, to map UART-HID:UID AMD0020:0 to
2427 PCI segment 0x1 and PCI device ID 00:14.5,
2428 write the parameter as:
2429 ivrs_acpihid=AMD0020:0@0001:00:14.5
2432 * To map UART-HID:UID AMD0020:0 to PCI segment is 0,
2433 PCI device ID 00:14.5, write the parameter as:
2434 ivrs_acpihid[00:14.5]=AMD0020:0
2435 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and
2436 PCI device ID 00:14.5, write the parameter as:
2437 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2439 js= [HW,JOY] Analog joystick
2440 See Documentation/input/joydev/joystick.rst.
2443 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2444 report on every invalid memory access. Without this
2445 parameter KASAN will print report only for the first
2449 Do not unregister boot console at start. This is only
2450 useful for debugging when something happens in the window
2451 between unregistering the boot console and initializing
2456 kernelcore= [KNL,X86,IA-64,PPC]
2457 Format: nn[KMGTPE] | nn% | "mirror"
2458 This parameter specifies the amount of memory usable by
2459 the kernel for non-movable allocations. The requested
2460 amount is spread evenly throughout all nodes in the
2461 system as ZONE_NORMAL. The remaining memory is used for
2462 movable memory in its own zone, ZONE_MOVABLE. In the
2463 event, a node is too small to have both ZONE_NORMAL and
2464 ZONE_MOVABLE, kernelcore memory will take priority and
2465 other nodes will have a larger ZONE_MOVABLE.
2467 ZONE_MOVABLE is used for the allocation of pages that
2468 may be reclaimed or moved by the page migration
2469 subsystem. Note that allocations like PTEs-from-HighMem
2470 still use the HighMem zone if it exists, and the Normal
2471 zone if it does not.
2473 It is possible to specify the exact amount of memory in
2474 the form of "nn[KMGTPE]", a percentage of total system
2475 memory in the form of "nn%", or "mirror". If "mirror"
2476 option is specified, mirrored (reliable) memory is used
2477 for non-movable allocations and remaining memory is used
2478 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2479 are exclusive, so you cannot specify multiple forms.
2481 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2482 Format: <Controller#>[,poll interval]
2483 The controller # is the number of the ehci usb debug
2484 port as it is probed via PCI. The poll interval is
2485 optional and is the number seconds in between
2486 each poll cycle to the debug port in case you need
2487 the functionality for interrupting the kernel with
2488 gdb or control-c on the dbgp connection. When
2489 not using this parameter you use sysrq-g to break into
2490 the kernel debugger.
2492 kgdboc= [KGDB,HW] kgdb over consoles.
2493 Requires a tty driver that supports console polling,
2494 or a supported polling keyboard driver (non-usb).
2495 Serial only format: <serial_device>[,baud]
2496 keyboard only format: kbd
2497 keyboard and serial format: kbd,<serial_device>[,baud]
2498 Optional Kernel mode setting:
2499 kms, kbd format: kms,kbd
2500 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2502 kgdboc_earlycon= [KGDB,HW]
2503 If the boot console provides the ability to read
2504 characters and can work in polling mode, you can use
2505 this parameter to tell kgdb to use it as a backend
2506 until the normal console is registered. Intended to
2507 be used together with the kgdboc parameter which
2508 specifies the normal console to transition to.
2510 The name of the early console should be specified
2511 as the value of this parameter. Note that the name of
2512 the early console might be different than the tty
2513 name passed to kgdboc. It's OK to leave the value
2514 blank and the first boot console that implements
2515 read() will be picked.
2517 kgdbwait [KGDB] Stop kernel execution and enter the
2518 kernel debugger at the earliest opportunity.
2520 kmac= [MIPS] Korina ethernet MAC address.
2521 Configure the RouterBoard 532 series on-chip
2522 Ethernet adapter MAC address.
2524 kmemleak= [KNL] Boot-time kmemleak enable/disable
2525 Valid arguments: on, off
2527 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2530 kprobe_event=[probe-list]
2531 [FTRACE] Add kprobe events and enable at boot time.
2532 The probe-list is a semicolon delimited list of probe
2533 definitions. Each definition is same as kprobe_events
2534 interface, but the parameters are comma delimited.
2535 For example, to add a kprobe event on vfs_read with
2536 arg1 and arg2, add to the command line;
2538 kprobe_event=p,vfs_read,$arg1,$arg2
2540 See also Documentation/trace/kprobetrace.rst "Kernel
2541 Boot Parameter" section.
2543 kpti= [ARM64] Control page table isolation of user
2544 and kernel address spaces.
2545 Default: enabled on cores which need mitigation.
2549 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires
2550 CONFIG_KUNIT to be set to be fully enabled. The
2551 default value can be overridden via
2552 KUNIT_DEFAULT_ENABLED.
2553 Default is 1 (enabled)
2555 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2556 Default is 0 (don't ignore, but inject #GP)
2558 kvm.eager_page_split=
2559 [KVM,X86] Controls whether or not KVM will try to
2560 proactively split all huge pages during dirty logging.
2561 Eager page splitting reduces interruptions to vCPU
2562 execution by eliminating the write-protection faults
2563 and MMU lock contention that would otherwise be
2564 required to split huge pages lazily.
2566 VM workloads that rarely perform writes or that write
2567 only to a small region of VM memory may benefit from
2568 disabling eager page splitting to allow huge pages to
2569 still be used for reads.
2571 The behavior of eager page splitting depends on whether
2572 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2573 disabled, all huge pages in a memslot will be eagerly
2574 split when dirty logging is enabled on that memslot. If
2575 enabled, eager page splitting will be performed during
2576 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2579 Eager page splitting is only supported when kvm.tdp_mmu=Y.
2583 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2584 Default is false (don't support).
2587 [KVM] Controls the software workaround for the
2588 X86_BUG_ITLB_MULTIHIT bug.
2589 force : Always deploy workaround.
2590 off : Never deploy workaround.
2591 auto : Deploy workaround based on the presence of
2592 X86_BUG_ITLB_MULTIHIT.
2596 If the software workaround is enabled for the host,
2597 guests do need not to enable it for nested guests.
2599 kvm.nx_huge_pages_recovery_ratio=
2600 [KVM] Controls how many 4KiB pages are periodically zapped
2601 back to huge pages. 0 disables the recovery, otherwise if
2602 the value is N KVM will zap 1/Nth of the 4KiB pages every
2603 period (see below). The default is 60.
2605 kvm.nx_huge_pages_recovery_period_ms=
2606 [KVM] Controls the time period at which KVM zaps 4KiB pages
2607 back to huge pages. If the value is a non-zero N, KVM will
2608 zap a portion (see ratio above) of the pages every N msecs.
2609 If the value is 0 (the default), KVM will pick a period based
2610 on the ratio, such that a page is zapped after 1 hour on average.
2612 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in
2613 KVM/SVM. Default is 1 (enabled).
2615 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables,
2616 a.k.a. Two-Dimensional Page Tables. Default is 1
2617 (enabled). Disable by KVM if hardware lacks support
2621 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2623 none: Forcefully disable KVM.
2625 nvhe: Standard nVHE-based mode, without support for
2628 protected: nVHE-based mode with support for guests whose
2629 state is kept private from the host.
2631 nested: VHE-based mode with support for nested
2632 virtualization. Requires at least ARMv8.3
2635 Defaults to VHE/nVHE based on hardware support. Setting
2636 mode to "protected" will disable kexec and hibernation
2637 for the host. "nested" is experimental and should be
2638 used with extreme caution.
2640 kvm-arm.vgic_v3_group0_trap=
2641 [KVM,ARM] Trap guest accesses to GICv3 group-0
2644 kvm-arm.vgic_v3_group1_trap=
2645 [KVM,ARM] Trap guest accesses to GICv3 group-1
2648 kvm-arm.vgic_v3_common_trap=
2649 [KVM,ARM] Trap guest accesses to GICv3 common
2652 kvm-arm.vgic_v4_enable=
2653 [KVM,ARM] Allow use of GICv4 for direct injection of
2656 kvm_cma_resv_ratio=n [PPC]
2657 Reserves given percentage from system memory area for
2658 contiguous memory allocation for KVM hash pagetable
2660 By default it reserves 5% of total system memory.
2664 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables,
2665 a.k.a. Two-Dimensional Page Tables. Default is 1
2666 (enabled). Disable by KVM if hardware lacks support
2669 kvm-intel.emulate_invalid_guest_state=
2670 [KVM,Intel] Control whether to emulate invalid guest
2671 state. Ignored if kvm-intel.enable_unrestricted_guest=1,
2672 as guest state is never invalid for unrestricted
2673 guests. This param doesn't apply to nested guests (L2),
2674 as KVM never emulates invalid L2 guest state.
2675 Default is 1 (enabled).
2677 kvm-intel.flexpriority=
2678 [KVM,Intel] Control KVM's use of FlexPriority feature
2679 (TPR shadow). Default is 1 (enabled). Disable by KVM if
2680 hardware lacks support for it.
2683 [KVM,Intel] Control nested virtualization feature in
2684 KVM/VMX. Default is 1 (enabled).
2686 kvm-intel.unrestricted_guest=
2687 [KVM,Intel] Control KVM's use of unrestricted guest
2688 feature (virtualized real and unpaged mode). Default
2689 is 1 (enabled). Disable by KVM if EPT is disabled or
2690 hardware lacks support for it.
2692 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2695 Valid arguments: never, cond, always
2697 always: L1D cache flush on every VMENTER.
2698 cond: Flush L1D on VMENTER only when the code between
2699 VMEXIT and VMENTER can leak host memory.
2700 never: Disables the mitigation
2702 Default is cond (do L1 cache flush in specific instances)
2704 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor
2705 Identification feature (tagged TLBs). Default is 1
2706 (enabled). Disable by KVM if hardware lacks support
2709 l1d_flush= [X86,INTEL]
2710 Control mitigation for L1D based snooping vulnerability.
2712 Certain CPUs are vulnerable to an exploit against CPU
2713 internal buffers which can forward information to a
2714 disclosure gadget under certain conditions.
2716 In vulnerable processors, the speculatively
2717 forwarded data can be used in a cache side channel
2718 attack, to access data to which the attacker does
2719 not have direct access.
2721 This parameter controls the mitigation. The
2724 on - enable the interface for the mitigation
2726 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2729 The kernel PTE inversion protection is unconditionally
2730 enabled and cannot be disabled.
2733 Provides all available mitigations for the
2734 L1TF vulnerability. Disables SMT and
2735 enables all mitigations in the
2736 hypervisors, i.e. unconditional L1D flush.
2738 SMT control and L1D flush control via the
2739 sysfs interface is still possible after
2740 boot. Hypervisors will issue a warning
2741 when the first VM is started in a
2742 potentially insecure configuration,
2743 i.e. SMT enabled or L1D flush disabled.
2746 Same as 'full', but disables SMT and L1D
2747 flush runtime control. Implies the
2748 'nosmt=force' command line option.
2749 (i.e. sysfs control of SMT is disabled.)
2752 Leaves SMT enabled and enables the default
2753 hypervisor mitigation, i.e. conditional
2756 SMT control and L1D flush control via the
2757 sysfs interface is still possible after
2758 boot. Hypervisors will issue a warning
2759 when the first VM is started in a
2760 potentially insecure configuration,
2761 i.e. SMT enabled or L1D flush disabled.
2765 Disables SMT and enables the default
2766 hypervisor mitigation.
2768 SMT control and L1D flush control via the
2769 sysfs interface is still possible after
2770 boot. Hypervisors will issue a warning
2771 when the first VM is started in a
2772 potentially insecure configuration,
2773 i.e. SMT enabled or L1D flush disabled.
2776 Same as 'flush', but hypervisors will not
2777 warn when a VM is started in a potentially
2778 insecure configuration.
2781 Disables hypervisor mitigations and doesn't
2783 It also drops the swap size and available
2784 RAM limit restriction on both hypervisor and
2789 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2795 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2798 lapic= [X86,APIC] Do not use TSC deadline
2799 value for LAPIC timer one-shot implementation. Default
2800 back to the programmable timer unit in the LAPIC.
2801 Format: notscdeadline
2803 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2806 libata.dma= [LIBATA] DMA control
2807 libata.dma=0 Disable all PATA and SATA DMA
2808 libata.dma=1 PATA and SATA Disk DMA only
2809 libata.dma=2 ATAPI (CDROM) DMA only
2810 libata.dma=4 Compact Flash DMA only
2811 Combinations also work, so libata.dma=3 enables DMA
2812 for disks and CDROMs, but not CFs.
2814 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2815 libata.ignore_hpa=0 keep BIOS limits (default)
2816 libata.ignore_hpa=1 ignore limits, using full disk
2818 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2822 libata.force= [LIBATA] Force configurations. The format is a comma-
2823 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2824 PORT and DEVICE are decimal numbers matching port, link
2825 or device. Basically, it matches the ATA ID string
2826 printed on console by libata. If the whole ID part is
2827 omitted, the last PORT and DEVICE values are used. If
2828 ID hasn't been specified yet, the configuration applies
2829 to all ports, links and devices.
2831 If only DEVICE is omitted, the parameter applies to
2832 the port and all links and devices behind it. DEVICE
2833 number of 0 either selects the first device or the
2834 first fan-out link behind PMP device. It does not
2835 select the host link. DEVICE number of 15 selects the
2836 host link and device attached to it.
2838 The VAL specifies the configuration to force. As long
2839 as there is no ambiguity, shortcut notation is allowed.
2840 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2841 The following configurations can be forced.
2843 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2844 Any ID with matching PORT is used.
2846 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2848 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2849 udma[/][16,25,33,44,66,100,133] notation is also
2852 * nohrst, nosrst, norst: suppress hard, soft and both
2855 * rstonce: only attempt one reset during hot-unplug
2858 * [no]dbdelay: Enable or disable the extra 200ms delay
2859 before debouncing a link PHY and device presence
2862 * [no]ncq: Turn on or off NCQ.
2864 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2866 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2868 * [no]trim: Enable or disable (unqueued) TRIM.
2870 * trim_zero: Indicate that TRIM command zeroes data.
2872 * max_trim_128m: Set 128M maximum trim size limit.
2874 * [no]dma: Turn on or off DMA transfers.
2876 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2878 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2879 commands that are not a multiple of 16 bytes.
2881 * [no]dmalog: Enable or disable the use of the
2882 READ LOG DMA EXT command to access logs.
2884 * [no]iddevlog: Enable or disable access to the
2885 identify device data log.
2887 * [no]logdir: Enable or disable access to the general
2888 purpose log directory.
2890 * max_sec_128: Set transfer size limit to 128 sectors.
2892 * max_sec_1024: Set or clear transfer size limit to
2895 * max_sec_lba48: Set or clear transfer size limit to
2898 * [no]lpm: Enable or disable link power management.
2900 * [no]setxfer: Indicate if transfer speed mode setting
2903 * [no]fua: Disable or enable FUA (Force Unit Access)
2904 support for devices supporting this feature.
2906 * dump_id: Dump IDENTIFY data.
2908 * disable: Disable this device.
2910 If there are multiple matching configurations changing
2911 the same attribute, the last one is used.
2913 load_ramdisk= [RAM] [Deprecated]
2915 lockd.nlm_grace_period=P [NFS] Assign grace period.
2918 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2921 lockd.nlm_timeout=T [NFS] Assign timeout value.
2924 lockd.nlm_udpport=M [NFS] Assign UDP port.
2927 lockdown= [SECURITY]
2928 { integrity | confidentiality }
2929 Enable the kernel lockdown feature. If set to
2930 integrity, kernel features that allow userland to
2931 modify the running kernel are disabled. If set to
2932 confidentiality, kernel features that allow userland
2933 to extract confidential information from the kernel
2936 locktorture.nreaders_stress= [KNL]
2937 Set the number of locking read-acquisition kthreads.
2938 Defaults to being automatically set based on the
2939 number of online CPUs.
2941 locktorture.nwriters_stress= [KNL]
2942 Set the number of locking write-acquisition kthreads.
2944 locktorture.onoff_holdoff= [KNL]
2945 Set time (s) after boot for CPU-hotplug testing.
2947 locktorture.onoff_interval= [KNL]
2948 Set time (s) between CPU-hotplug operations, or
2949 zero to disable CPU-hotplug testing.
2951 locktorture.shuffle_interval= [KNL]
2952 Set task-shuffle interval (jiffies). Shuffling
2953 tasks allows some CPUs to go into dyntick-idle
2954 mode during the locktorture test.
2956 locktorture.shutdown_secs= [KNL]
2957 Set time (s) after boot system shutdown. This
2958 is useful for hands-off automated testing.
2960 locktorture.stat_interval= [KNL]
2961 Time (s) between statistics printk()s.
2963 locktorture.stutter= [KNL]
2964 Time (s) to stutter testing, for example,
2965 specifying five seconds causes the test to run for
2966 five seconds, wait for five seconds, and so on.
2967 This tests the locking primitive's ability to
2968 transition abruptly to and from idle.
2970 locktorture.torture_type= [KNL]
2971 Specify the locking implementation to test.
2973 locktorture.writer_fifo= [KNL]
2974 Run the write-side locktorture kthreads at
2975 sched_set_fifo() real-time priority.
2977 locktorture.verbose= [KNL]
2978 Enable additional printk() statements.
2980 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2983 loglevel= All Kernel Messages with a loglevel smaller than the
2984 console loglevel will be printed to the console. It can
2985 also be changed with klogd or other programs. The
2986 loglevels are defined as follows:
2988 0 (KERN_EMERG) system is unusable
2989 1 (KERN_ALERT) action must be taken immediately
2990 2 (KERN_CRIT) critical conditions
2991 3 (KERN_ERR) error conditions
2992 4 (KERN_WARNING) warning conditions
2993 5 (KERN_NOTICE) normal but significant condition
2994 6 (KERN_INFO) informational
2995 7 (KERN_DEBUG) debug-level messages
2997 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2998 in bytes. n must be a power of two and greater
2999 than the minimal size. The minimal size is defined
3000 by LOG_BUF_SHIFT kernel config parameter. There is
3001 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
3002 that allows to increase the default size depending on
3003 the number of CPUs. See init/Kconfig for more details.
3005 logo.nologo [FB] Disables display of the built-in Linux logo.
3006 This may be used to provide more screen space for
3007 kernel log messages and is useful when debugging
3008 kernel boot problems.
3010 lp=0 [LP] Specify parallel ports to use, e.g,
3011 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
3012 lp=reset first parallel port). 'lp=0' disables the
3013 lp=auto printer driver. 'lp=reset' (which can be
3014 specified in addition to the ports) causes
3015 attached printers to be reset. Using
3016 lp=port1,port2,... specifies the parallel ports
3017 to associate lp devices with, starting with
3018 lp0. A port specification may be 'none' to skip
3019 that lp device, or a parport name such as
3020 'parport0'. Specifying 'lp=auto' instead of a
3021 port specification list means that device IDs
3022 from each port should be examined, to see if
3023 an IEEE 1284-compliant printer is attached; if
3024 so, the driver will manage that printer.
3025 See also header of drivers/char/lp.c.
3028 Sets loops_per_jiffy to given constant, thus avoiding
3029 time-consuming boot-time autodetection (up to 250 ms per
3030 CPU). 0 enables autodetection (default). To determine
3031 the correct value for your kernel, boot with normal
3032 autodetection and see what value is printed. Note that
3033 on SMP systems the preset will be applied to all CPUs,
3034 which is likely to cause problems if your CPUs need
3035 significantly divergent settings. An incorrect value
3036 will cause delays in the kernel to be wrong, leading to
3037 unpredictable I/O errors and other breakage. Although
3038 unlikely, in the extreme case this might damage your
3042 Format: <io>,<irq>,<dma>
3044 lsm.debug [SECURITY] Enable LSM initialization debugging output.
3047 [SECURITY] Choose order of LSM initialization. This
3048 overrides CONFIG_LSM, and the "security=" parameter.
3050 machvec= [IA-64] Force the use of a particular machine-vector
3051 (machvec) in a generic kernel.
3052 Example: machvec=hpzx1
3054 machtype= [Loongson] Share the same kernel image file between
3055 different yeeloong laptops.
3056 Example: machtype=lemote-yeeloong-2f-7inch
3058 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
3059 than or equal to this physical address is ignored.
3061 maxcpus= [SMP] Maximum number of processors that an SMP kernel
3062 will bring up during bootup. maxcpus=n : n >= 0 limits
3063 the kernel to bring up 'n' processors. Surely after
3064 bootup you can bring up the other plugged cpu by executing
3065 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
3066 only takes effect during system bootup.
3067 While n=0 is a special case, it is equivalent to "nosmp",
3068 which also disables the IO APIC.
3070 max_loop= [LOOP] The number of loop block devices that get
3071 (loop.max_loop) unconditionally pre-created at init time. The default
3072 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
3073 of statically allocating a predefined number, loop
3074 devices can be requested on-demand with the
3075 /dev/loop-control interface.
3077 mce [X86-32] Machine Check Exception
3079 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
3081 md= [HW] RAID subsystems devices and level
3082 See Documentation/admin-guide/md.rst.
3085 Format: <first>,<last>
3086 Specifies range of consoles to be captured by the MDA.
3089 Control mitigation for the Micro-architectural Data
3090 Sampling (MDS) vulnerability.
3092 Certain CPUs are vulnerable to an exploit against CPU
3093 internal buffers which can forward information to a
3094 disclosure gadget under certain conditions.
3096 In vulnerable processors, the speculatively
3097 forwarded data can be used in a cache side channel
3098 attack, to access data to which the attacker does
3099 not have direct access.
3101 This parameter controls the MDS mitigation. The
3104 full - Enable MDS mitigation on vulnerable CPUs
3105 full,nosmt - Enable MDS mitigation and disable
3106 SMT on vulnerable CPUs
3107 off - Unconditionally disable MDS mitigation
3109 On TAA-affected machines, mds=off can be prevented by
3110 an active TAA mitigation as both vulnerabilities are
3111 mitigated with the same mechanism so in order to disable
3112 this mitigation, you need to specify tsx_async_abort=off
3115 Not specifying this option is equivalent to
3118 For details see: Documentation/admin-guide/hw-vuln/mds.rst
3120 mem=nn[KMG] [HEXAGON] Set the memory size.
3121 Must be specified, otherwise memory size will be 0.
3123 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
3124 Amount of memory to be used in cases as follows:
3127 2 when the kernel is not able to see the whole system memory;
3128 3 memory that lies after 'mem=' boundary is excluded from
3129 the hypervisor, then assigned to KVM guests.
3130 4 to limit the memory available for kdump kernel.
3132 [ARC,MICROBLAZE] - the limit applies only to low memory,
3133 high memory is not affected.
3135 [ARM64] - only limits memory covered by the linear
3136 mapping. The NOMAP regions are not affected.
3138 [X86] Work as limiting max address. Use together
3139 with memmap= to avoid physical address space collisions.
3140 Without memmap= PCI devices could be placed at addresses
3141 belonging to unused RAM.
3143 Note that this only takes effects during boot time since
3144 in above case 3, memory may need be hot added after boot
3145 if system memory of hypervisor is not sufficient.
3148 [ARM,MIPS] - override the memory layout reported by
3150 Define a memory region of size nn[KMG] starting at
3152 Multiple different regions can be specified with
3153 multiple mem= parameters on the command line.
3155 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3158 memblock=debug [KNL] Enable memblock debug messages.
3161 [KNL,SH] Allow user to override the default size for
3162 per-device physically contiguous DMA buffers.
3164 memhp_default_state=online/offline/online_kernel/online_movable
3165 [KNL] Set the initial state for the memory hotplug
3166 onlining policy. If not specified, the default value is
3167 set according to the
3168 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3170 See Documentation/admin-guide/mm/memory-hotplug.rst.
3172 memmap=exactmap [KNL,X86] Enable setting of an exact
3173 E820 memory map, as specified by the user.
3174 Such memmap=exactmap lines can be constructed based on
3175 BIOS output or other requirements. See the memmap=nn@ss
3178 memmap=nn[KMG]@ss[KMG]
3179 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
3180 Region of memory to be used is from ss to ss+nn.
3181 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3182 which limits max address to nn[KMG].
3183 Multiple different regions can be specified,
3186 memmap=100M@2G,100M#3G,1G!1024G
3188 memmap=nn[KMG]#ss[KMG]
3189 [KNL,ACPI] Mark specific memory as ACPI data.
3190 Region of memory to be marked is from ss to ss+nn.
3192 memmap=nn[KMG]$ss[KMG]
3193 [KNL,ACPI] Mark specific memory as reserved.
3194 Region of memory to be reserved is from ss to ss+nn.
3195 Example: Exclude memory from 0x18690000-0x1869ffff
3196 memmap=64K$0x18690000
3198 memmap=0x10000$0x18690000
3199 Some bootloaders may need an escape character before '$',
3200 like Grub2, otherwise '$' and the following number
3203 memmap=nn[KMG]!ss[KMG]
3204 [KNL,X86] Mark specific memory as protected.
3205 Region of memory to be used, from ss to ss+nn.
3206 The memory region may be marked as e820 type 12 (0xc)
3207 and is NVDIMM or ADR memory.
3209 memmap=<size>%<offset>-<oldtype>+<newtype>
3210 [KNL,ACPI] Convert memory within the specified region
3211 from <oldtype> to <newtype>. If "-<oldtype>" is left
3212 out, the whole region will be marked as <newtype>,
3213 even if previously unavailable. If "+<newtype>" is left
3214 out, matching memory will be removed. Types are
3215 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3216 3 = ACPI, 12 = PRAM.
3218 memory_corruption_check=0/1 [X86]
3219 Some BIOSes seem to corrupt the first 64k of
3220 memory when doing things like suspend/resume.
3221 Setting this option will scan the memory
3222 looking for corruption. Enabling this will
3223 both detect corruption and prevent the kernel
3224 from using the memory being corrupted.
3225 However, its intended as a diagnostic tool; if
3226 repeatable BIOS-originated corruption always
3227 affects the same memory, you can use memmap=
3228 to prevent the kernel from using that memory.
3230 memory_corruption_check_size=size [X86]
3231 By default it checks for corruption in the low
3232 64k, making this memory unavailable for normal
3233 use. Use this parameter to scan for
3234 corruption in more or less memory.
3236 memory_corruption_check_period=seconds [X86]
3237 By default it checks for corruption every 60
3238 seconds. Use this parameter to check at some
3239 other rate. 0 disables periodic checking.
3241 memory_hotplug.memmap_on_memory
3242 [KNL,X86,ARM] Boolean flag to enable this feature.
3243 Format: {on | off (default)}
3244 When enabled, runtime hotplugged memory will
3245 allocate its internal metadata (struct pages,
3246 those vmemmap pages cannot be optimized even
3247 if hugetlb_free_vmemmap is enabled) from the
3248 hotadded memory which will allow to hotadd a
3249 lot of memory without requiring additional
3251 This feature is disabled by default because it
3252 has some implication on large (e.g. GB)
3253 allocations in some configurations (e.g. small
3255 The state of the flag can be read in
3256 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3257 Note that even when enabled, there are a few cases where
3258 the feature is not effective.
3260 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3262 default : 0 <disable>
3263 Specifies the number of memtest passes to be
3264 performed. Each pass selects another test
3265 pattern from a given set of patterns. Memtest
3266 fills the memory with this pattern, validates
3267 memory contents and reserves bad memory
3268 regions that are detected.
3270 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3271 Valid arguments: on, off
3273 mem_encrypt=on: Activate SME
3274 mem_encrypt=off: Do not activate SME
3276 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3277 for details on when memory encryption can be activated.
3279 mem_sleep_default= [SUSPEND] Default system suspend mode:
3280 s2idle - Suspend-To-Idle
3281 shallow - Power-On Suspend or equivalent (if supported)
3282 deep - Suspend-To-RAM or equivalent (if supported)
3283 See Documentation/admin-guide/pm/sleep-states.rst.
3285 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3286 Multi-Function General Purpose Timers on AMD Geode
3289 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3290 the BIOS has incorrectly applied a workaround. TinyBIOS
3291 version 0.98 is known to be affected, 0.99 fixes the
3292 problem by letting the user disable the workaround.
3296 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3297 physical address is ignored.
3299 mini2440= [ARM,HW,KNL]
3300 Format:[0..2][b][c][t]
3302 MINI2440 configuration specification:
3303 0 - The attached screen is the 3.5" TFT
3304 1 - The attached screen is the 7" TFT
3305 2 - The VGA Shield is attached (1024x768)
3306 Leaving out the screen size parameter will not load
3307 the TFT driver, and the framebuffer will be left
3309 b - Enable backlight. The TFT backlight pin will be
3310 linked to the kernel VESA blanking code and a GPIO
3311 LED. This parameter is not necessary when using the
3313 c - Enable the s3c camera interface.
3314 t - Reserved for enabling touchscreen support. The
3315 touchscreen support is not enabled in the mainstream
3316 kernel as of 2.6.30, a preliminary port can be found
3317 in the "bleeding edge" mini2440 support kernel at
3318 https://repo.or.cz/w/linux-2.6/mini2440.git
3321 [X86,PPC,S390,ARM64] Control optional mitigations for
3322 CPU vulnerabilities. This is a set of curated,
3323 arch-independent options, each of which is an
3324 aggregation of existing arch-specific options.
3326 Note, "mitigations" is supported if and only if the
3327 kernel was built with CPU_MITIGATIONS=y.
3330 Disable all optional CPU mitigations. This
3331 improves system performance, but it may also
3332 expose users to several CPU vulnerabilities.
3333 Equivalent to: if nokaslr then kpti=0 [ARM64]
3334 gather_data_sampling=off [X86]
3335 kvm.nx_huge_pages=off [X86]
3338 mmio_stale_data=off [X86]
3339 no_entry_flush [PPC]
3340 no_uaccess_flush [PPC]
3343 nospectre_bhb [ARM64]
3344 nospectre_v1 [X86,PPC]
3345 nospectre_v2 [X86,PPC,S390,ARM64]
3346 reg_file_data_sampling=off [X86]
3348 spec_store_bypass_disable=off [X86,PPC]
3349 spectre_bhi=off [X86]
3350 spectre_v2_user=off [X86]
3351 srbds=off [X86,INTEL]
3352 ssbd=force-off [ARM64]
3353 tsx_async_abort=off [X86]
3356 This does not have any effect on
3357 kvm.nx_huge_pages when
3358 kvm.nx_huge_pages=force.
3361 Mitigate all CPU vulnerabilities, but leave SMT
3362 enabled, even if it's vulnerable. This is for
3363 users who don't want to be surprised by SMT
3364 getting disabled across kernel upgrades, or who
3365 have other ways of avoiding SMT-based attacks.
3366 Equivalent to: (default behavior)
3369 Mitigate all CPU vulnerabilities, disabling SMT
3370 if needed. This is for users who always want to
3371 be fully mitigated, even if it means losing SMT.
3372 Equivalent to: l1tf=flush,nosmt [X86]
3373 mds=full,nosmt [X86]
3374 tsx_async_abort=full,nosmt [X86]
3375 mmio_stale_data=full,nosmt [X86]
3376 retbleed=auto,nosmt [X86]
3379 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3380 parameter allows control of the logging verbosity for
3381 the additional memory initialisation checks. A value
3382 of 0 disables mminit logging and a level of 4 will
3383 log everything. Information is printed at KERN_DEBUG
3384 so loglevel=8 may also need to be specified.
3387 [X86,INTEL] Control mitigation for the Processor
3388 MMIO Stale Data vulnerabilities.
3390 Processor MMIO Stale Data is a class of
3391 vulnerabilities that may expose data after an MMIO
3392 operation. Exposed data could originate or end in
3393 the same CPU buffers as affected by MDS and TAA.
3394 Therefore, similar to MDS and TAA, the mitigation
3395 is to clear the affected CPU buffers.
3397 This parameter controls the mitigation. The
3400 full - Enable mitigation on vulnerable CPUs
3402 full,nosmt - Enable mitigation and disable SMT on
3405 off - Unconditionally disable mitigation
3407 On MDS or TAA affected machines,
3408 mmio_stale_data=off can be prevented by an active
3409 MDS or TAA mitigation as these vulnerabilities are
3410 mitigated with the same mechanism so in order to
3411 disable this mitigation, you need to specify
3412 mds=off and tsx_async_abort=off too.
3414 Not specifying this option is equivalent to
3415 mmio_stale_data=full.
3418 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3420 <module>.async_probe[=<bool>] [KNL]
3421 If no <bool> value is specified or if the value
3422 specified is not a valid <bool>, enable asynchronous
3423 probe on this module. Otherwise, enable/disable
3424 asynchronous probe on this module as indicated by the
3425 <bool> value. See also: module.async_probe
3427 module.async_probe=<bool>
3428 [KNL] When set to true, modules will use async probing
3429 by default. To enable/disable async probing for a
3430 specific module, use the module specific control that
3431 is documented under <module>.async_probe. When both
3432 module.async_probe and <module>.async_probe are
3433 specified, <module>.async_probe takes precedence for
3434 the specific module.
3436 module.enable_dups_trace
3437 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set,
3438 this means that duplicate request_module() calls will
3439 trigger a WARN_ON() instead of a pr_warn(). Note that
3440 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s
3441 will always be issued and this option does nothing.
3443 [KNL] When CONFIG_MODULE_SIG is set, this means that
3444 modules without (valid) signatures will fail to load.
3445 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3446 is always true, so this option does nothing.
3448 module_blacklist= [KNL] Do not load a comma-separated list of
3449 modules. Useful for debugging problem modules.
3452 [MOUSE] Maximum time between finger touching and
3453 leaving touchpad surface for touch to be considered
3454 a tap and be reported as a left button click (for
3455 touchpads working in absolute mode only).
3457 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3458 reporting absolute coordinates, such as tablets
3459 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3460 reporting absolute coordinates, such as tablets
3462 movablecore= [KNL,X86,IA-64,PPC]
3463 Format: nn[KMGTPE] | nn%
3464 This parameter is the complement to kernelcore=, it
3465 specifies the amount of memory used for migratable
3466 allocations. If both kernelcore and movablecore is
3467 specified, then kernelcore will be at *least* the
3468 specified value but may be more. If movablecore on its
3469 own is specified, the administrator must be careful
3470 that the amount of memory usable for all allocations
3473 movable_node [KNL] Boot-time switch to make hotplugable memory
3474 NUMA nodes to be movable. This means that the memory
3475 of such nodes will be usable only for movable
3476 allocations which rules out almost all kernel
3477 allocations. Use with caution!
3479 MTD_Partition= [MTD]
3480 Format: <name>,<region-number>,<size>,<offset>
3482 MTD_Region= [MTD] Format:
3483 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3486 See drivers/mtd/parsers/cmdlinepart.c
3489 ARM/S3C2412 JIVE boot control
3491 See arch/arm/mach-s3c/mach-jive.c
3493 mtouchusb.raw_coordinates=
3494 [HW] Make the MicroTouch USB driver use raw coordinates
3495 ('y', default) or cooked coordinates ('n')
3498 Enable printing debug information related to MTRR
3499 registers at boot time.
3501 mtrr_chunk_size=nn[KMG] [X86]
3502 used for mtrr cleanup. It is largest continuous chunk
3503 that could hold holes aka. UC entries.
3505 mtrr_gran_size=nn[KMG] [X86]
3506 Used for mtrr cleanup. It is granularity of mtrr block.
3508 Large value could prevent small alignment from
3511 mtrr_spare_reg_nr=n [X86]
3513 Range: 0,7 : spare reg number
3515 Used for mtrr cleanup. It is spare mtrr entries number.
3516 Set to 2 or more if your graphical card needs more.
3518 multitce=off [PPC] This parameter disables the use of the pSeries
3519 firmware feature for updating multiple TCE entries
3522 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3524 netdev= [NET] Network devices parameters
3525 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3526 Note that mem_start is often overloaded to mean
3527 something different and driver-specific.
3528 This usage is only documented in each driver source
3531 netpoll.carrier_timeout=
3532 [NET] Specifies amount of time (in seconds) that
3533 netpoll should wait for a carrier. By default netpoll
3537 [NETFILTER] Enable connection tracking flow accounting
3538 0 to disable accounting
3539 1 to enable accounting
3543 [NFS] sets the pathname to the program which is used
3544 to update the NFS client cache entries.
3546 nfs.cache_getent_timeout=
3547 [NFS] sets the timeout after which an attempt to
3548 update a cache entry is deemed to have failed.
3550 nfs.callback_nr_threads=
3551 [NFSv4] set the total number of threads that the
3552 NFS client will assign to service NFSv4 callback
3555 nfs.callback_tcpport=
3556 [NFS] set the TCP port on which the NFSv4 callback
3557 channel should listen.
3560 [NFS] enable 64-bit inode numbers.
3561 If zero, the NFS client will fake up a 32-bit inode
3562 number for the readdir() and stat() syscalls instead
3563 of returning the full 64-bit number.
3564 The default is to return 64-bit inode numbers.
3566 nfs.idmap_cache_timeout=
3567 [NFS] set the maximum lifetime for idmapper cache
3570 nfs.max_session_cb_slots=
3571 [NFSv4.1] Sets the maximum number of session
3572 slots the client will assign to the callback
3573 channel. This determines the maximum number of
3574 callbacks the client will process in parallel for
3575 a particular server.
3577 nfs.max_session_slots=
3578 [NFSv4.1] Sets the maximum number of session slots
3579 the client will attempt to negotiate with the server.
3580 This limits the number of simultaneous RPC requests
3581 that the client can send to the NFSv4.1 server.
3582 Note that there is little point in setting this
3583 value higher than the max_tcp_slot_table_limit.
3585 nfs.nfs4_disable_idmapping=
3586 [NFSv4] When set to the default of '1', this option
3587 ensures that both the RPC level authentication
3588 scheme and the NFS level operations agree to use
3589 numeric uids/gids if the mount is using the
3590 'sec=sys' security flavour. In effect it is
3591 disabling idmapping, which can make migration from
3592 legacy NFSv2/v3 systems to NFSv4 easier.
3593 Servers that do not support this mode of operation
3594 will be autodetected by the client, and it will fall
3595 back to using the idmapper.
3596 To turn off this behaviour, set the value to '0'.
3599 [NFS4] Specify an additional fixed unique ident-
3600 ification string that NFSv4 clients can insert into
3601 their nfs_client_id4 string. This is typically a
3602 UUID that is generated at system install time.
3604 nfs.recover_lost_locks=
3605 [NFSv4] Attempt to recover locks that were lost due
3606 to a lease timeout on the server. Please note that
3607 doing this risks data corruption, since there are
3608 no guarantees that the file will remain unchanged
3609 after the locks are lost.
3610 If you want to enable the kernel legacy behaviour of
3611 attempting to recover these locks, then set this
3613 The default parameter value of '0' causes the kernel
3614 not to attempt recovery of lost locks.
3616 nfs.send_implementation_id=
3617 [NFSv4.1] Send client implementation identification
3618 information in exchange_id requests.
3619 If zero, no implementation identification information
3621 The default is to send the implementation identification
3624 nfs4.layoutstats_timer=
3625 [NFSv4.2] Change the rate at which the kernel sends
3626 layoutstats to the pNFS metadata server.
3628 Setting this to value to 0 causes the kernel to use
3629 whatever value is the default set by the layout
3630 driver. A non-zero value sets the minimum interval
3631 in seconds between layoutstats transmissions.
3633 nfsd.inter_copy_offload_enable=
3634 [NFSv4.2] When set to 1, the server will support
3635 server-to-server copies for which this server is
3636 the destination of the copy.
3638 nfsd.nfs4_disable_idmapping=
3639 [NFSv4] When set to the default of '1', the NFSv4
3640 server will return only numeric uids and gids to
3641 clients using auth_sys, and will accept numeric uids
3642 and gids from such clients. This is intended to ease
3643 migration from NFSv2/v3.
3645 nfsd.nfsd4_ssc_umount_timeout=
3646 [NFSv4.2] When used as the destination of a
3647 server-to-server copy, knfsd temporarily mounts
3648 the source server. It caches the mount in case
3649 it will be needed again, and discards it if not
3650 used for the number of milliseconds specified by
3653 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3654 See Documentation/admin-guide/nfs/nfsroot.rst.
3656 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3657 See Documentation/admin-guide/nfs/nfsroot.rst.
3659 nfsrootdebug [NFS] enable nfsroot debugging messages.
3660 See Documentation/admin-guide/nfs/nfsroot.rst.
3662 nmi_backtrace.backtrace_idle [KNL]
3663 Dump stacks even of idle CPUs in response to an
3664 NMI stack-backtrace request.
3666 nmi_debug= [KNL,SH] Specify one or more actions to take
3667 when a NMI is triggered.
3668 Format: [state][,regs][,debounce][,die]
3670 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3671 Format: [panic,][nopanic,][num]
3673 0 - turn hardlockup detector in nmi_watchdog off
3674 1 - turn hardlockup detector in nmi_watchdog on
3675 When panic is specified, panic when an NMI watchdog
3676 timeout occurs (or 'nopanic' to not panic on an NMI
3677 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3678 To disable both hard and soft lockup detectors,
3679 please see 'nowatchdog'.
3680 This is useful when you use a panic=... timeout and
3681 need the box quickly up again.
3683 These settings can be accessed at runtime via
3684 the nmi_watchdog and hardlockup_panic sysctls.
3686 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3687 emulation library even if a 387 maths coprocessor
3690 no4lvl [RISCV] Disable 4-level and 5-level paging modes. Forces
3691 kernel to use 3-level paging instead.
3693 no5lvl [X86-64,RISCV] Disable 5-level paging mode. Forces
3694 kernel to use 4-level paging instead.
3696 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3697 caches in the slab allocator. Saves per-node memory,
3698 but will impact performance.
3702 noaltinstr [S390] Disables alternative instructions patching
3703 (CPU alternatives feature).
3705 noapic [SMP,APIC] Tells the kernel to not make use of any
3706 IOAPICs that may be present in the system.
3708 noautogroup Disable scheduler automatic task group creation.
3713 [HW] Never suspend the console
3714 Disable suspending of consoles during suspend and
3715 hibernate operations. Once disabled, debugging
3716 messages can reach various consoles while the rest
3717 of the system is being put to sleep (ie, while
3718 debugging driver suspend/resume hooks). This may
3719 not work reliably with all consoles, but is known
3720 to work with serial and VGA consoles.
3721 To facilitate more flexible debugging, we also add
3722 console_suspend, a printk module parameter to control
3723 it. Users could use console_suspend (usually
3724 /sys/module/printk/parameters/console_suspend) to
3725 turn on/off it dynamically.
3728 [KNL] Disable object debugging
3730 nodsp [SH] Disable hardware DSP at boot time.
3732 noefi Disable EFI runtime services support.
3734 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3739 This affects only 32-bit executables.
3740 noexec32=on: enable non-executable mappings (default)
3741 read doesn't imply executable mappings
3742 noexec32=off: disable non-executable mappings
3743 read implies executable mappings
3745 no_file_caps Tells the kernel not to honor file capabilities. The
3746 only way then for a file to be executed with privilege
3747 is to be setuid root or executed by root.
3749 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3751 nofsgsbase [X86] Disables FSGSBASE instructions.
3753 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3754 register save and restore. The kernel will only save
3755 legacy floating-point registers on task switch.
3757 nohalt [IA-64] Tells the kernel not to use the power saving
3758 function PAL_HALT_LIGHT when idle. This increases
3759 power-consumption. On the positive side, it reduces
3760 interrupt wake-up latency, which may improve performance
3761 in certain environments such as networked servers or
3765 Force pointers printed to the console or buffers to be
3766 unhashed. By default, when a pointer is printed via %p
3767 format string, that pointer is "hashed", i.e. obscured
3768 by hashing the pointer value. This is a security feature
3769 that hides actual kernel addresses from unprivileged
3770 users, but it also makes debugging the kernel more
3771 difficult since unequal pointers can no longer be
3772 compared. However, if this command-line option is
3773 specified, then all normal pointers will have their true
3774 value printed. This option should only be specified when
3775 debugging the kernel. Please do not use on production
3778 nohibernate [HIBERNATION] Disable hibernation and resume.
3780 nohlt [ARM,ARM64,MICROBLAZE,MIPS,PPC,SH] Forces the kernel to
3781 busy wait in do_idle() and not use the arch_cpu_idle()
3782 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3783 to be effective. This is useful on platforms where the
3784 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3785 correctly or when doing power measurements to evaluate
3786 the impact of the sleep instructions. This is also
3787 useful when using JTAG debugger.
3789 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3791 nohugevmalloc [KNL,X86,PPC,ARM64] Disable kernel huge vmalloc mappings.
3793 nohz= [KNL] Boottime enable/disable dynamic ticks
3794 Valid arguments: on, off
3797 nohz_full= [KNL,BOOT,SMP,ISOL]
3798 The argument is a cpu list, as described above.
3799 In kernels built with CONFIG_NO_HZ_FULL=y, set
3800 the specified list of CPUs whose tick will be stopped
3801 whenever possible. The boot CPU will be forced outside
3802 the range to maintain the timekeeping. Any CPUs
3803 in this list will have their RCU callbacks offloaded,
3804 just as if they had also been called out in the
3805 rcu_nocbs= boot parameter.
3807 Note that this argument takes precedence over
3808 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3810 noinitrd [RAM] Tells the kernel not to load any configured
3813 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3815 [Deprecated - use intremap=off]
3819 noinvpcid [X86] Disable the INVPCID cpu feature.
3821 noiotrap [SH] Disables trapped I/O port accesses.
3823 noirqdebug [X86-32] Disables the code which attempts to detect and
3824 disable unhandled interrupt sources.
3826 noisapnp [ISAPNP] Disables ISA PnP code.
3828 nojitter [IA-64] Disables jitter checking for ITC timers.
3831 When CONFIG_RANDOMIZE_BASE is set, this disables
3832 kernel and module base offset ASLR (Address Space
3833 Layout Randomization).
3835 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3838 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3840 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3842 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3844 nomca [IA-64] Disable machine check abort handling
3846 nomce [X86-32] Disable Machine Check Exception
3848 nomfgpt [X86-32] Disable Multi-Function General Purpose
3849 Timer usage (for AMD Geode machines).
3851 nomodeset Disable kernel modesetting. Most systems' firmware
3852 sets up a display mode and provides framebuffer memory
3853 for output. With nomodeset, DRM and fbdev drivers will
3854 not load if they could possibly displace the pre-
3855 initialized output. Only the system framebuffer will
3856 be available for use. The respective drivers will not
3857 perform display-mode changes or accelerated rendering.
3859 Useful as error fallback, or for testing and debugging.
3861 nomodule Disable module load
3863 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3864 shutdown the other cpus. Instead use the REBOOT_VECTOR
3867 nopat [X86] Disable PAT (page attribute table extension of
3868 pagetables) support.
3870 nopcid [X86-64] Disable the PCID cpu feature.
3872 nopku [X86] Disable Memory Protection Keys CPU feature found
3876 Equivalent to pti=off
3878 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
3879 Disables the PV optimizations forcing the guest to run
3880 as generic guest with no PV drivers. Currently support
3881 XEN HVM, KVM, HYPER_V and VMWARE guest.
3883 nopvspin [X86,XEN,KVM]
3884 Disables the qspinlock slow path using PV optimizations
3885 which allow the hypervisor to 'idle' the guest on lock
3888 norandmaps Don't use address space randomization. Equivalent to
3889 echo 0 > /proc/sys/kernel/randomize_va_space
3891 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3892 with UP alternatives
3894 noresume [SWSUSP] Disables resume and restores original swap
3899 no-scroll [VGA] Disables scrollback.
3900 This is required for the Braillex ib80-piezo Braille
3901 reader made by F.H. Papenmeier (Germany).
3903 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3906 Disable SMAP (Supervisor Mode Access Prevention)
3907 even if it is supported by processor.
3910 Disable SMEP (Supervisor Mode Execution Prevention)
3911 even if it is supported by processor.
3913 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3914 and disable the IO APIC. legacy for "maxcpus=0".
3916 nosmt [KNL,MIPS,PPC,S390] Disable symmetric multithreading (SMT).
3917 Equivalent to smt=1.
3919 [KNL,X86,PPC] Disable symmetric multithreading (SMT).
3920 nosmt=force: Force disable SMT, cannot be undone
3921 via the sysfs control file.
3923 nosoftlockup [KNL] Disable the soft-lockup detector.
3925 nospec_store_bypass_disable
3926 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3928 nospectre_bhb [ARM64] Disable all mitigations for Spectre-BHB (branch
3929 history injection) vulnerability. System may allow data leaks
3932 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3933 (bounds check bypass). With this option data leaks are
3934 possible in the system.
3936 nospectre_v2 [X86,PPC_E500,ARM64] Disable all mitigations for
3937 the Spectre variant 2 (indirect branch prediction)
3938 vulnerability. System may allow data leaks with this
3941 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES] Disable paravirtualized
3942 steal time accounting. steal time is computed, but
3943 won't influence scheduler behaviour
3945 nosync [HW,M68K] Disables sync negotiation for all devices.
3947 no_timer_check [X86,APIC] Disables the code which tests for
3948 broken timer IRQ sources.
3951 [PPC] Don't flush the L1-D cache after accessing user data.
3953 novmcoredd [KNL,KDUMP]
3954 Disable device dump. Device dump allows drivers to
3955 append dump data to vmcore so you can collect driver
3956 specified debug info. Drivers can append the data
3957 without any limit and this data is stored in memory,
3958 so this may cause significant memory stress. Disabling
3959 device dump can help save memory but the driver debug
3960 data will be no longer available. This parameter
3961 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3965 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3966 clock and use the default one.
3968 nowatchdog [KNL] Disable both lockup detectors, i.e.
3969 soft-lockup and NMI watchdog (hard-lockup).
3973 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3975 NOTE: this parameter will be ignored on systems with the
3976 LEGACY_XAPIC_DISABLED bit set in the
3977 IA32_XAPIC_DISABLE_STATUS MSR.
3979 noxsave [BUGS=X86] Disables x86 extended register state save
3980 and restore using xsave. The kernel will fallback to
3981 enabling legacy floating-point and sse state.
3983 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3984 register states. The kernel will fall back to use
3985 xsave to save the states. By using this parameter,
3986 performance of saving the states is degraded because
3987 xsave doesn't support modified optimization while
3988 xsaveopt supports it on xsaveopt enabled systems.
3990 noxsaves [X86] Disables xsaves and xrstors used in saving and
3991 restoring x86 extended register state in compacted
3992 form of xsave area. The kernel will fall back to use
3993 xsaveopt and xrstor to save and restore the states
3994 in standard form of xsave area. By using this
3995 parameter, xsave area per process might occupy more
3996 memory on xsaves enabled systems.
3998 nps_mtm_hs_ctr= [KNL,ARC]
3999 This parameter sets the maximum duration, in
4000 cycles, each HW thread of the CTOP can run
4001 without interruptions, before HW switches it.
4002 The actual maximum duration is 16 times this
4004 Format: integer between 1 and 255
4007 nptcg= [IA-64] Override max number of concurrent global TLB
4008 purges which is reported from either PAL_VM_SUMMARY or
4011 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
4012 could support. nr_cpus=n : n >= 1 limits the kernel to
4013 support 'n' processors. It could be larger than the
4014 number of already plugged CPU during bootup, later in
4015 runtime you can physically add extra cpu until it reaches
4016 n. So during boot up some boot time memory for per-cpu
4017 variables need be pre-allocated for later physical cpu
4020 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
4022 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
4023 set up a single NUMA node spanning all memory.
4025 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
4027 Allowed values are enable and disable
4029 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
4030 'node', 'default' can be specified
4031 This can be set from sysctl after boot.
4032 See Documentation/admin-guide/sysctl/vm.rst for details.
4034 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
4035 See Documentation/core-api/debugging-via-ohci1394.rst for more
4038 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
4039 Rather than timing out after 20 ms if an EC
4040 command is not properly ACKed, override the length
4041 of the timeout. We have interrupts disabled while
4042 waiting for the ACK, so if this is set too high
4043 interrupts *may* be lost!
4045 omap_mux= [OMAP] Override bootloader pin multiplexing.
4046 Format: <mux_mode0.mode_name=value>...
4047 For example, to override I2C bus2:
4048 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
4050 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
4052 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
4054 boundary - index of last SLC block on Flex-OneNAND.
4055 The remaining blocks are configured as MLC blocks.
4056 lock - Configure if Flex-OneNAND boundary should be locked.
4057 Once locked, the boundary cannot be changed.
4058 1 indicates lock status, 0 indicates unlock status.
4060 oops=panic Always panic on oopses. Default is to just kill the
4061 process, but there is a small probability of
4062 deadlocking the machine.
4063 This will also cause panics on machine check exceptions.
4064 Useful together with panic=30 to trigger a reboot.
4067 [KNL] Boolean flag to control whether the page allocator
4068 should randomize its free lists. The randomization may
4069 be automatically enabled if the kernel detects it is
4070 running on a platform with a direct-mapped memory-side
4071 cache, and this parameter can be used to
4072 override/disable that behavior. The state of the flag
4073 can be read from sysfs at:
4074 /sys/module/page_alloc/parameters/shuffle.
4076 page_owner= [KNL] Boot-time page_owner enabling option.
4077 Storage of the information about who allocated
4078 each page is disabled in default. With this switch,
4080 on: enable the feature
4082 page_poison= [KNL] Boot-time parameter changing the state of
4083 poisoning on the buddy allocator, available with
4084 CONFIG_PAGE_POISONING=y.
4085 off: turn off poisoning (default)
4086 on: turn on poisoning
4088 page_reporting.page_reporting_order=
4089 [KNL] Minimal page reporting order
4091 Adjust the minimal page reporting order. The page
4092 reporting is disabled when it exceeds MAX_ORDER.
4094 panic= [KNL] Kernel behaviour on panic: delay <timeout>
4095 timeout > 0: seconds before rebooting
4096 timeout = 0: wait forever
4097 timeout < 0: reboot immediately
4100 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
4101 Format: <hex>[,nousertaint]
4102 Hexadecimal bitmask representing the set of TAINT flags
4103 that will cause the kernel to panic when add_taint() is
4104 called with any of the flags in this set.
4105 The optional switch "nousertaint" can be utilized to
4106 prevent userspace forced crashes by writing to sysctl
4107 /proc/sys/kernel/tainted any flagset matching with the
4108 bitmask set on panic_on_taint.
4109 See Documentation/admin-guide/tainted-kernels.rst for
4110 extra details on the taint flags that users can pick
4111 to compose the bitmask to assign to panic_on_taint.
4113 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump
4116 panic_print= Bitmask for printing system info when panic happens.
4117 User can chose combination of the following bits:
4118 bit 0: print all tasks info
4119 bit 1: print system memory info
4120 bit 2: print timer info
4121 bit 3: print locks info if CONFIG_LOCKDEP is on
4122 bit 4: print ftrace buffer
4123 bit 5: print all printk messages in buffer
4124 bit 6: print all CPUs backtrace (if available in the arch)
4125 *Be aware* that this option may print a _lot_ of lines,
4126 so there are risks of losing older messages in the log.
4127 Use this option carefully, maybe worth to setup a
4128 bigger log buffer with "log_buf_len" along with this.
4130 parkbd.port= [HW] Parallel port number the keyboard adapter is
4131 connected to, default is 0.
4133 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
4134 0 for XT, 1 for AT (default is AT).
4137 parport= [HW,PPT] Specify parallel ports. 0 disables.
4138 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
4139 Use 'auto' to force the driver to use any
4140 IRQ/DMA settings detected (the default is to
4141 ignore detected IRQ/DMA settings because of
4142 possible conflicts). You can specify the base
4143 address, IRQ, and DMA settings; IRQ and DMA
4144 should be numbers, or 'auto' (for using detected
4145 settings on that particular port), or 'nofifo'
4146 (to avoid using a FIFO even if it is detected).
4147 Parallel ports are assigned in the order they
4148 are specified on the command line, starting
4151 parport_init_mode= [HW,PPT]
4152 Configure VIA parallel port to operate in
4153 a specific mode. This is necessary on Pegasos
4154 computer where firmware has no options for setting
4155 up parallel port mode and sets it to spp.
4156 Currently this function knows 686a and 8231 chips.
4157 Format: [spp|ps2|epp|ecp|ecpepp]
4159 pata_legacy.all= [HW,LIBATA]
4161 Set to non-zero to probe primary and secondary ISA
4162 port ranges on PCI systems where no PCI PATA device
4163 has been found at either range. Disabled by default.
4165 pata_legacy.autospeed= [HW,LIBATA]
4167 Set to non-zero if a chip is present that snoops speed
4168 changes. Disabled by default.
4170 pata_legacy.ht6560a= [HW,LIBATA]
4172 Set to 1, 2, or 3 for HT 6560A on the primary channel,
4173 the secondary channel, or both channels respectively.
4174 Disabled by default.
4176 pata_legacy.ht6560b= [HW,LIBATA]
4178 Set to 1, 2, or 3 for HT 6560B on the primary channel,
4179 the secondary channel, or both channels respectively.
4180 Disabled by default.
4182 pata_legacy.iordy_mask= [HW,LIBATA]
4184 IORDY enable mask. Set individual bits to allow IORDY
4185 for the respective channel. Bit 0 is for the first
4186 legacy channel handled by this driver, bit 1 is for
4187 the second channel, and so on. The sequence will often
4188 correspond to the primary legacy channel, the secondary
4189 legacy channel, and so on, but the handling of a PCI
4190 bus and the use of other driver options may interfere
4191 with the sequence. By default IORDY is allowed across
4194 pata_legacy.opti82c46x= [HW,LIBATA]
4196 Set to 1, 2, or 3 for Opti 82c611A on the primary
4197 channel, the secondary channel, or both channels
4198 respectively. Disabled by default.
4200 pata_legacy.opti82c611a= [HW,LIBATA]
4202 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4203 channel, the secondary channel, or both channels
4204 respectively. Disabled by default.
4206 pata_legacy.pio_mask= [HW,LIBATA]
4208 PIO mode mask for autospeed devices. Set individual
4209 bits to allow the use of the respective PIO modes.
4210 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4211 All modes allowed by default.
4213 pata_legacy.probe_all= [HW,LIBATA]
4215 Set to non-zero to probe tertiary and further ISA
4216 port ranges on PCI systems. Disabled by default.
4218 pata_legacy.probe_mask= [HW,LIBATA]
4220 Probe mask for legacy ISA PATA ports. Depending on
4221 platform configuration and the use of other driver
4222 options up to 6 legacy ports are supported: 0x1f0,
4223 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4224 of individual ports can be disabled by setting the
4225 corresponding bits in the mask to 1. Bit 0 is for
4226 the first port in the list above (0x1f0), and so on.
4227 By default all supported ports are probed.
4229 pata_legacy.qdi= [HW,LIBATA]
4231 Set to non-zero to probe QDI controllers. By default
4232 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4234 pata_legacy.winbond= [HW,LIBATA]
4236 Set to non-zero to probe Winbond controllers. Use
4237 the standard I/O port (0x130) if 1, otherwise the
4238 value given is the I/O port to use (typically 0x1b0).
4239 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4242 pata_platform.pio_mask= [HW,LIBATA]
4244 Supported PIO mode mask. Set individual bits to allow
4245 the use of the respective PIO modes. Bit 0 is for
4246 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4250 Halt all CPUs after the first oops has been printed for
4251 the specified number of seconds. This is to be used if
4252 your oopses keep scrolling off the screen.
4256 pci=option[,option...] [PCI] various PCI subsystem options.
4258 Some options herein operate on a specific device
4259 or a set of devices (<pci_dev>). These are
4260 specified in one of the following formats:
4262 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4263 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4265 Note: the first format specifies a PCI
4266 bus/device/function address which may change
4267 if new hardware is inserted, if motherboard
4268 firmware changes, or due to changes caused
4269 by other kernel parameters. If the
4270 domain is left unspecified, it is
4271 taken to be zero. Optionally, a path
4272 to a device through multiple device/function
4273 addresses can be specified after the base
4274 address (this is more robust against
4275 renumbering issues). The second format
4276 selects devices using IDs from the
4277 configuration space which may match multiple
4278 devices in the system.
4280 earlydump dump PCI config space before the kernel
4282 off [X86] don't probe for the PCI bus
4283 bios [X86-32] force use of PCI BIOS, don't access
4284 the hardware directly. Use this if your machine
4285 has a non-standard PCI host bridge.
4286 nobios [X86-32] disallow use of PCI BIOS, only direct
4287 hardware access methods are allowed. Use this
4288 if you experience crashes upon bootup and you
4289 suspect they are caused by the BIOS.
4290 conf1 [X86] Force use of PCI Configuration Access
4291 Mechanism 1 (config address in IO port 0xCF8,
4292 data in IO port 0xCFC, both 32-bit).
4293 conf2 [X86] Force use of PCI Configuration Access
4294 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4295 the function, IO port 0xCFA, also 8-bit, sets
4296 bus number. The config space is then accessed
4297 through ports 0xC000-0xCFFF).
4298 See http://wiki.osdev.org/PCI for more info
4299 on the configuration access mechanisms.
4300 noaer [PCIE] If the PCIEAER kernel config parameter is
4301 enabled, this kernel boot option can be used to
4302 disable the use of PCIE advanced error reporting.
4303 nodomains [PCI] Disable support for multiple PCI
4304 root domains (aka PCI segments, in ACPI-speak).
4305 nommconf [X86] Disable use of MMCONFIG for PCI
4307 check_enable_amd_mmconf [X86] check for and enable
4308 properly configured MMIO access to PCI
4309 config space on AMD family 10h CPU
4310 nomsi [MSI] If the PCI_MSI kernel config parameter is
4311 enabled, this kernel boot option can be used to
4312 disable the use of MSI interrupts system-wide.
4313 noioapicquirk [APIC] Disable all boot interrupt quirks.
4314 Safety option to keep boot IRQs enabled. This
4315 should never be necessary.
4316 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4317 primary IO-APIC for bridges that cannot disable
4318 boot IRQs. This fixes a source of spurious IRQs
4319 when the system masks IRQs.
4320 noioapicreroute [APIC] Disable workaround that uses the
4321 boot IRQ equivalent of an IRQ that connects to
4322 a chipset where boot IRQs cannot be disabled.
4323 The opposite of ioapicreroute.
4324 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4325 routing table. These calls are known to be buggy
4326 on several machines and they hang the machine
4327 when used, but on other computers it's the only
4328 way to get the interrupt routing table. Try
4329 this option if the kernel is unable to allocate
4330 IRQs or discover secondary PCI buses on your
4332 rom [X86] Assign address space to expansion ROMs.
4333 Use with caution as certain devices share
4334 address decoders between ROMs and other
4336 norom [X86] Do not assign address space to
4337 expansion ROMs that do not already have
4338 BIOS assigned address ranges.
4339 nobar [X86] Do not assign address space to the
4340 BARs that weren't assigned by the BIOS.
4341 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4342 assigned automatically to PCI devices. You can
4343 make the kernel exclude IRQs of your ISA cards
4345 pirqaddr=0xAAAAA [X86] Specify the physical address
4346 of the PIRQ table (normally generated
4347 by the BIOS) if it is outside the
4348 F0000h-100000h range.
4349 lastbus=N [X86] Scan all buses thru bus #N. Can be
4350 useful if the kernel is unable to find your
4351 secondary buses and you want to tell it
4352 explicitly which ones they are.
4353 assign-busses [X86] Always assign all PCI bus
4354 numbers ourselves, overriding
4355 whatever the firmware may have done.
4356 usepirqmask [X86] Honor the possible IRQ mask stored
4357 in the BIOS $PIR table. This is needed on
4358 some systems with broken BIOSes, notably
4359 some HP Pavilion N5400 and Omnibook XE3
4360 notebooks. This will have no effect if ACPI
4361 IRQ routing is enabled.
4362 noacpi [X86] Do not use ACPI for IRQ routing
4363 or for PCI scanning.
4364 use_crs [X86] Use PCI host bridge window information
4365 from ACPI. On BIOSes from 2008 or later, this
4366 is enabled by default. If you need to use this,
4367 please report a bug.
4368 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4369 If you need to use this, please report a bug.
4370 use_e820 [X86] Use E820 reservations to exclude parts of
4371 PCI host bridge windows. This is a workaround
4372 for BIOS defects in host bridge _CRS methods.
4373 If you need to use this, please report a bug to
4374 <linux-pci@vger.kernel.org>.
4375 no_e820 [X86] Ignore E820 reservations for PCI host
4376 bridge windows. This is the default on modern
4377 hardware. If you need to use this, please report
4378 a bug to <linux-pci@vger.kernel.org>.
4379 routeirq Do IRQ routing for all PCI devices.
4380 This is normally done in pci_enable_device(),
4381 so this option is a temporary workaround
4382 for broken drivers that don't call it.
4383 skip_isa_align [X86] do not align io start addr, so can
4384 handle more pci cards
4385 noearly [X86] Don't do any early type 1 scanning.
4386 This might help on some broken boards which
4387 machine check when some devices' config space
4388 is read. But various workarounds are disabled
4389 and some IOMMU drivers will not work.
4390 bfsort Sort PCI devices into breadth-first order.
4391 This sorting is done to get a device
4392 order compatible with older (<= 2.4) kernels.
4393 nobfsort Don't sort PCI devices into breadth-first order.
4394 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4395 tuning and use the BIOS-configured MPS defaults.
4396 pcie_bus_safe Set every device's MPS to the largest value
4397 supported by all devices below the root complex.
4398 pcie_bus_perf Set device MPS to the largest allowable MPS
4399 based on its parent bus. Also set MRRS (Max
4400 Read Request Size) to the largest supported
4401 value (no larger than the MPS that the device
4402 or bus can support) for best performance.
4403 pcie_bus_peer2peer Set every device's MPS to 128B, which
4404 every device is guaranteed to support. This
4405 configuration allows peer-to-peer DMA between
4406 any pair of devices, possibly at the cost of
4407 reduced performance. This also guarantees
4408 that hot-added devices will work.
4409 cbiosize=nn[KMG] The fixed amount of bus space which is
4410 reserved for the CardBus bridge's IO window.
4411 The default value is 256 bytes.
4412 cbmemsize=nn[KMG] The fixed amount of bus space which is
4413 reserved for the CardBus bridge's memory
4414 window. The default value is 64 megabytes.
4417 [<order of align>@]<pci_dev>[; ...]
4418 Specifies alignment and device to reassign
4419 aligned memory resources. How to
4420 specify the device is described above.
4421 If <order of align> is not specified,
4422 PAGE_SIZE is used as alignment.
4423 A PCI-PCI bridge can be specified if resource
4424 windows need to be expanded.
4425 To specify the alignment for several
4426 instances of a device, the PCI vendor,
4427 device, subvendor, and subdevice may be
4428 specified, e.g., 12@pci:8086:9c22:103c:198f
4429 for 4096-byte alignment.
4430 ecrc= Enable/disable PCIe ECRC (transaction layer
4431 end-to-end CRC checking). Only effective if
4432 OS has native AER control (either granted by
4433 ACPI _OSC or forced via "pcie_ports=native")
4434 bios: Use BIOS/firmware settings. This is the
4438 hpiosize=nn[KMG] The fixed amount of bus space which is
4439 reserved for hotplug bridge's IO window.
4440 Default size is 256 bytes.
4441 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4442 reserved for hotplug bridge's MMIO window.
4443 Default size is 2 megabytes.
4444 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4445 reserved for hotplug bridge's MMIO_PREF window.
4446 Default size is 2 megabytes.
4447 hpmemsize=nn[KMG] The fixed amount of bus space which is
4448 reserved for hotplug bridge's MMIO and
4450 Default size is 2 megabytes.
4451 hpbussize=nn The minimum amount of additional bus numbers
4452 reserved for buses below a hotplug bridge.
4454 realloc= Enable/disable reallocating PCI bridge resources
4455 if allocations done by BIOS are too small to
4456 accommodate resources required by all child
4458 off: Turn realloc off
4460 realloc same as realloc=on
4461 noari do not use PCIe ARI.
4462 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4463 do not use PCIe ATS (and IOMMU device IOTLB).
4464 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4465 only look for one device below a PCIe downstream
4467 big_root_window Try to add a big 64bit memory window to the PCIe
4468 root complex on AMD CPUs. Some GFX hardware
4469 can resize a BAR to allow access to all VRAM.
4470 Adding the window is slightly risky (it may
4471 conflict with unreported devices), so this
4473 disable_acs_redir=<pci_dev>[; ...]
4474 Specify one or more PCI devices (in the format
4475 specified above) separated by semicolons.
4476 Each device specified will have the PCI ACS
4477 redirect capabilities forced off which will
4478 allow P2P traffic between devices through
4479 bridges without forcing it upstream. Note:
4480 this removes isolation between devices and
4481 may put more devices in an IOMMU group.
4482 force_floating [S390] Force usage of floating interrupts.
4483 nomio [S390] Do not use MIO instructions.
4484 norid [S390] ignore the RID field and force use of
4485 one PCI domain per PCI function
4487 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4490 force Enable ASPM even on devices that claim not to support it.
4491 WARNING: Forcing ASPM on may cause system lockups.
4493 pcie_ports= [PCIE] PCIe port services handling:
4494 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4495 even if the platform doesn't give the OS permission to
4496 use them. This may cause conflicts if the platform
4497 also tries to use these services.
4498 dpc-native Use native PCIe service for DPC only. May
4499 cause conflicts if firmware uses AER or DPC.
4500 compat Disable native PCIe services (PME, AER, DPC, PCIe
4503 pcie_port_pm= [PCIE] PCIe port power management handling:
4504 off Disable power management of all PCIe ports
4505 force Forcibly enable power management of all PCIe ports
4507 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4508 nomsi Do not use MSI for native PCIe PME signaling (this makes
4509 all PCIe root ports use INTx for all services).
4511 pcmv= [HW,PCMCIA] BadgePAD 4
4515 Keep all power-domains already enabled by bootloader on,
4516 even if no driver has claimed them. This is useful
4517 for debug and development, but should not be
4518 needed on a platform with proper driver support.
4520 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4523 See arch/parisc/kernel/pdc_chassis.c
4525 percpu_alloc= Select which percpu first chunk allocator to use.
4526 Currently supported values are "embed" and "page".
4527 Archs may support subset or none of the selections.
4528 See comments in mm/percpu.c for details on each
4529 allocator. This parameter is primarily for debugging
4530 and performance comparison.
4532 pirq= [SMP,APIC] Manual mp-table setup
4533 See Documentation/arch/x86/i386/IO-APIC.rst.
4535 plip= [PPT,NET] Parallel port network link
4536 Format: { parport<nr> | timid | 0 }
4537 See also Documentation/admin-guide/parport.rst.
4539 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4540 Override pmtimer IOPort with a hex value.
4543 pmu_override= [PPC] Override the PMU.
4544 This option takes over the PMU facility, so it is no
4545 longer usable by perf. Setting this option starts the
4546 PMU counters by setting MMCR0 to 0 (the FC bit is
4547 cleared). If a number is given, then MMCR1 is set to
4548 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4551 pm_debug_messages [SUSPEND,KNL]
4552 Enable suspend/resume debug messages during boot up.
4555 Enable PNP debug messages (depends on the
4556 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4557 via /sys/module/pnp/parameters/debug. We always show
4558 current resource usage; turning this on also shows
4559 possible settings and some assignment information.
4565 { on | off | curr | res | no-curr | no-res }
4568 [ISAPNP] Exclude IRQs for the autoconfiguration
4571 [ISAPNP] Exclude DMAs for the autoconfiguration
4573 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4574 Ranges are in pairs (I/O port base and size).
4577 [ISAPNP] Exclude memory regions for the
4579 Ranges are in pairs (memory base and size).
4581 ports= [IP_VS_FTP] IPVS ftp helper module
4583 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4585 Format: <port>,<port>....
4587 powersave=off [PPC] This option disables power saving features.
4588 It specifically disables cpuidle and sets the
4589 platform machine description specific power_save
4590 function to NULL. On Idle the CPU just reduces
4593 ppc_strict_facility_enable
4594 [PPC] This option catches any kernel floating point,
4595 Altivec, VSX and SPE outside of regions specifically
4596 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4597 There is some performance impact when enabling this.
4601 Disable Hardware Transactional Memory
4604 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4605 none - Limited to cond_resched() calls
4606 voluntary - Limited to cond_resched() and might_sleep() calls
4607 full - Any section that isn't explicitly preempt disabled
4608 can be preempted anytime.
4610 print-fatal-signals=
4611 [KNL] debug: print fatal signals
4613 If enabled, warn about various signal handling
4614 related application anomalies: too many signals,
4615 too many POSIX.1 timers, fatal signals causing a
4618 If you hit the warning due to signal overflow,
4619 you might want to try "ulimit -i unlimited".
4623 printk.always_kmsg_dump=
4624 Trigger kmsg_dump for cases other than kernel oops or
4626 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4629 printk.console_no_auto_verbose=
4630 Disable console loglevel raise on oops, panic
4631 or lockdep-detected issues (only if lock debug is on).
4632 With an exception to setups with low baudrate on
4633 serial console, keeping this 0 is a good choice
4634 in order to provide more debug information.
4636 default: 0 (auto_verbose is enabled)
4638 printk.devkmsg={on,off,ratelimit}
4639 Control writing to /dev/kmsg.
4640 on - unlimited logging to /dev/kmsg from userspace
4641 off - logging to /dev/kmsg disabled
4642 ratelimit - ratelimit the logging
4645 printk.time= Show timing data prefixed to each printk message line
4646 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4648 processor.max_cstate= [HW,ACPI]
4649 Limit processor to maximum C-state
4650 max_cstate=9 overrides any DMI blacklist limit.
4652 processor.nocst [HW,ACPI]
4653 Ignore the _CST method to determine C-states,
4654 instead using the legacy FADT method
4656 profile= [KNL] Enable kernel profiling via /proc/profile
4657 Format: [<profiletype>,]<number>
4658 Param: <profiletype>: "schedule", "sleep", or "kvm"
4659 [defaults to kernel profiling]
4660 Param: "schedule" - profile schedule points.
4661 Param: "sleep" - profile D-state sleeping (millisecs).
4662 Requires CONFIG_SCHEDSTATS
4663 Param: "kvm" - profile VM exits.
4664 Param: <number> - step/bucket size as a power of 2 for
4665 statistical time based profiling.
4667 prompt_ramdisk= [RAM] [Deprecated]
4669 prot_virt= [S390] enable hosting protected virtual machines
4670 isolated from the hypervisor (if hardware supports
4674 psi= [KNL] Enable or disable pressure stall information
4678 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4679 probe for; one of (bare|imps|exps|lifebook|any).
4680 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4682 psmouse.resetafter= [HW,MOUSE]
4683 Try to reset the device after so many bad packets
4686 [HW,MOUSE] Set desired mouse resolution, in dpi.
4687 psmouse.smartscroll=
4688 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4689 0 = disabled, 1 = enabled (default).
4691 pstore.backend= Specify the name of the pstore backend to use
4693 pti= [X86-64] Control Page Table Isolation of user and
4694 kernel address spaces. Disabling this feature
4695 removes hardening, but improves performance of
4696 system calls and interrupts.
4698 on - unconditionally enable
4699 off - unconditionally disable
4700 auto - kernel detects whether your CPU model is
4701 vulnerable to issues that PTI mitigates
4703 Not specifying this option is equivalent to pti=auto.
4706 [KNL] Number of legacy pty's. Overwrites compiled-in
4709 quiet [KNL] Disable most log messages
4713 radix_hcall_invalidate=on [PPC/PSERIES]
4714 Disable RADIX GTSE feature and use hcall for TLB
4718 See Documentation/admin-guide/md.rst.
4720 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4721 See Documentation/admin-guide/blockdev/ramdisk.rst.
4723 ramdisk_start= [RAM] RAM disk image start address
4725 random.trust_cpu=off
4726 [KNL] Disable trusting the use of the CPU's
4727 random number generator (if available) to
4728 initialize the kernel's RNG.
4730 random.trust_bootloader=off
4731 [KNL] Disable trusting the use of the a seed
4732 passed by the bootloader (if available) to
4733 initialize the kernel's RNG.
4735 randomize_kstack_offset=
4736 [KNL] Enable or disable kernel stack offset
4737 randomization, which provides roughly 5 bits of
4738 entropy, frustrating memory corruption attacks
4739 that depend on stack address determinism or
4740 cross-syscall address exposures. This is only
4741 available on architectures that have defined
4742 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4743 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4744 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4746 ras=option[,option,...] [KNL] RAS-specific options
4749 Disable the Correctable Errors Collector,
4750 see CONFIG_RAS_CEC help text.
4752 rcu_nocbs[=cpu-list]
4753 [KNL] The optional argument is a cpu list,
4756 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4757 enable the no-callback CPU mode, which prevents
4758 such CPUs' callbacks from being invoked in
4759 softirq context. Invocation of such CPUs' RCU
4760 callbacks will instead be offloaded to "rcuox/N"
4761 kthreads created for that purpose, where "x" is
4762 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4763 for the kthreads that mediate grace periods; and
4764 "N" is the CPU number. This reduces OS jitter on
4765 the offloaded CPUs, which can be useful for HPC
4766 and real-time workloads. It can also improve
4767 energy efficiency for asymmetric multiprocessors.
4769 If a cpulist is passed as an argument, the specified
4770 list of CPUs is set to no-callback mode from boot.
4772 Otherwise, if the '=' sign and the cpulist
4773 arguments are omitted, no CPU will be set to
4774 no-callback mode from boot but the mode may be
4775 toggled at runtime via cpusets.
4777 Note that this argument takes precedence over
4778 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4781 Rather than requiring that offloaded CPUs
4782 (specified by rcu_nocbs= above) explicitly
4783 awaken the corresponding "rcuoN" kthreads,
4784 make these kthreads poll for callbacks.
4785 This improves the real-time response for the
4786 offloaded CPUs by relieving them of the need to
4787 wake up the corresponding kthread, but degrades
4788 energy efficiency by requiring that the kthreads
4789 periodically wake up to do the polling.
4791 rcutree.blimit= [KNL]
4792 Set maximum number of finished RCU callbacks to
4793 process in one batch.
4795 rcutree.dump_tree= [KNL]
4796 Dump the structure of the rcu_node combining tree
4797 out at early boot. This is used for diagnostic
4798 purposes, to verify correct tree setup.
4800 rcutree.gp_cleanup_delay= [KNL]
4801 Set the number of jiffies to delay each step of
4802 RCU grace-period cleanup.
4804 rcutree.gp_init_delay= [KNL]
4805 Set the number of jiffies to delay each step of
4806 RCU grace-period initialization.
4808 rcutree.gp_preinit_delay= [KNL]
4809 Set the number of jiffies to delay each step of
4810 RCU grace-period pre-initialization, that is,
4811 the propagation of recent CPU-hotplug changes up
4812 the rcu_node combining tree.
4814 rcutree.jiffies_till_first_fqs= [KNL]
4815 Set delay from grace-period initialization to
4816 first attempt to force quiescent states.
4817 Units are jiffies, minimum value is zero,
4818 and maximum value is HZ.
4820 rcutree.jiffies_till_next_fqs= [KNL]
4821 Set delay between subsequent attempts to force
4822 quiescent states. Units are jiffies, minimum
4823 value is one, and maximum value is HZ.
4825 rcutree.jiffies_till_sched_qs= [KNL]
4826 Set required age in jiffies for a
4827 given grace period before RCU starts
4828 soliciting quiescent-state help from
4829 rcu_note_context_switch() and cond_resched().
4830 If not specified, the kernel will calculate
4831 a value based on the most recent settings
4832 of rcutree.jiffies_till_first_fqs
4833 and rcutree.jiffies_till_next_fqs.
4834 This calculated value may be viewed in
4835 rcutree.jiffies_to_sched_qs. Any attempt to set
4836 rcutree.jiffies_to_sched_qs will be cheerfully
4839 rcutree.kthread_prio= [KNL,BOOT]
4840 Set the SCHED_FIFO priority of the RCU per-CPU
4841 kthreads (rcuc/N). This value is also used for
4842 the priority of the RCU boost threads (rcub/N)
4843 and for the RCU grace-period kthreads (rcu_bh,
4844 rcu_preempt, and rcu_sched). If RCU_BOOST is
4845 set, valid values are 1-99 and the default is 1
4846 (the least-favored priority). Otherwise, when
4847 RCU_BOOST is not set, valid values are 0-99 and
4848 the default is zero (non-realtime operation).
4849 When RCU_NOCB_CPU is set, also adjust the
4850 priority of NOCB callback kthreads.
4852 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4853 On callback-offloaded (rcu_nocbs) CPUs,
4854 RCU reduces the lock contention that would
4855 otherwise be caused by callback floods through
4856 use of the ->nocb_bypass list. However, in the
4857 common non-flooded case, RCU queues directly to
4858 the main ->cblist in order to avoid the extra
4859 overhead of the ->nocb_bypass list and its lock.
4860 But if there are too many callbacks queued during
4861 a single jiffy, RCU pre-queues the callbacks into
4862 the ->nocb_bypass queue. The definition of "too
4863 many" is supplied by this kernel boot parameter.
4865 rcutree.qhimark= [KNL]
4866 Set threshold of queued RCU callbacks beyond which
4867 batch limiting is disabled.
4869 rcutree.qlowmark= [KNL]
4870 Set threshold of queued RCU callbacks below which
4871 batch limiting is re-enabled.
4873 rcutree.qovld= [KNL]
4874 Set threshold of queued RCU callbacks beyond which
4875 RCU's force-quiescent-state scan will aggressively
4876 enlist help from cond_resched() and sched IPIs to
4877 help CPUs more quickly reach quiescent states.
4878 Set to less than zero to make this be set based
4879 on rcutree.qhimark at boot time and to zero to
4880 disable more aggressive help enlistment.
4882 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4883 Set the page-cache refill delay (in milliseconds)
4884 in response to low-memory conditions. The range
4885 of permitted values is in the range 0:100000.
4887 rcutree.rcu_divisor= [KNL]
4888 Set the shift-right count to use to compute
4889 the callback-invocation batch limit bl from
4890 the number of callbacks queued on this CPU.
4891 The result will be bounded below by the value of
4892 the rcutree.blimit kernel parameter. Every bl
4893 callbacks, the softirq handler will exit in
4894 order to allow the CPU to do other work.
4896 Please note that this callback-invocation batch
4897 limit applies only to non-offloaded callback
4898 invocation. Offloaded callbacks are instead
4899 invoked in the context of an rcuoc kthread, which
4900 scheduler will preempt as it does any other task.
4902 rcutree.rcu_fanout_exact= [KNL]
4903 Disable autobalancing of the rcu_node combining
4904 tree. This is used by rcutorture, and might
4905 possibly be useful for architectures having high
4906 cache-to-cache transfer latencies.
4908 rcutree.rcu_fanout_leaf= [KNL]
4909 Change the number of CPUs assigned to each
4910 leaf rcu_node structure. Useful for very
4911 large systems, which will choose the value 64,
4912 and for NUMA systems with large remote-access
4913 latencies, which will choose a value aligned
4914 with the appropriate hardware boundaries.
4916 rcutree.rcu_min_cached_objs= [KNL]
4917 Minimum number of objects which are cached and
4918 maintained per one CPU. Object size is equal
4919 to PAGE_SIZE. The cache allows to reduce the
4920 pressure to page allocator, also it makes the
4921 whole algorithm to behave better in low memory
4924 rcutree.rcu_nocb_gp_stride= [KNL]
4925 Set the number of NOCB callback kthreads in
4926 each group, which defaults to the square root
4927 of the number of CPUs. Larger numbers reduce
4928 the wakeup overhead on the global grace-period
4929 kthread, but increases that same overhead on
4930 each group's NOCB grace-period kthread.
4932 rcutree.rcu_kick_kthreads= [KNL]
4933 Cause the grace-period kthread to get an extra
4934 wake_up() if it sleeps three times longer than
4935 it should at force-quiescent-state time.
4936 This wake_up() will be accompanied by a
4937 WARN_ONCE() splat and an ftrace_dump().
4939 rcutree.rcu_resched_ns= [KNL]
4940 Limit the time spend invoking a batch of RCU
4941 callbacks to the specified number of nanoseconds.
4942 By default, this limit is checked only once
4943 every 32 callbacks in order to limit the pain
4944 inflicted by local_clock() overhead.
4946 rcutree.rcu_unlock_delay= [KNL]
4947 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4948 this specifies an rcu_read_unlock()-time delay
4949 in microseconds. This defaults to zero.
4950 Larger delays increase the probability of
4951 catching RCU pointer leaks, that is, buggy use
4952 of RCU-protected pointers after the relevant
4953 rcu_read_unlock() has completed.
4955 rcutree.sysrq_rcu= [KNL]
4956 Commandeer a sysrq key to dump out Tree RCU's
4957 rcu_node tree with an eye towards determining
4958 why a new grace period has not yet started.
4960 rcutree.use_softirq= [KNL]
4961 If set to zero, move all RCU_SOFTIRQ processing to
4962 per-CPU rcuc kthreads. Defaults to a non-zero
4963 value, meaning that RCU_SOFTIRQ is used by default.
4964 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4966 But note that CONFIG_PREEMPT_RT=y kernels disable
4967 this kernel boot parameter, forcibly setting it
4970 rcuscale.gp_async= [KNL]
4971 Measure performance of asynchronous
4972 grace-period primitives such as call_rcu().
4974 rcuscale.gp_async_max= [KNL]
4975 Specify the maximum number of outstanding
4976 callbacks per writer thread. When a writer
4977 thread exceeds this limit, it invokes the
4978 corresponding flavor of rcu_barrier() to allow
4979 previously posted callbacks to drain.
4981 rcuscale.gp_exp= [KNL]
4982 Measure performance of expedited synchronous
4983 grace-period primitives.
4985 rcuscale.holdoff= [KNL]
4986 Set test-start holdoff period. The purpose of
4987 this parameter is to delay the start of the
4988 test until boot completes in order to avoid
4991 rcuscale.kfree_by_call_rcu= [KNL]
4992 In kernels built with CONFIG_RCU_LAZY=y, test
4993 call_rcu() instead of kfree_rcu().
4995 rcuscale.kfree_mult= [KNL]
4996 Instead of allocating an object of size kfree_obj,
4997 allocate one of kfree_mult * sizeof(kfree_obj).
5000 rcuscale.kfree_rcu_test= [KNL]
5001 Set to measure performance of kfree_rcu() flooding.
5003 rcuscale.kfree_rcu_test_double= [KNL]
5004 Test the double-argument variant of kfree_rcu().
5005 If this parameter has the same value as
5006 rcuscale.kfree_rcu_test_single, both the single-
5007 and double-argument variants are tested.
5009 rcuscale.kfree_rcu_test_single= [KNL]
5010 Test the single-argument variant of kfree_rcu().
5011 If this parameter has the same value as
5012 rcuscale.kfree_rcu_test_double, both the single-
5013 and double-argument variants are tested.
5015 rcuscale.kfree_nthreads= [KNL]
5016 The number of threads running loops of kfree_rcu().
5018 rcuscale.kfree_alloc_num= [KNL]
5019 Number of allocations and frees done in an iteration.
5021 rcuscale.kfree_loops= [KNL]
5022 Number of loops doing rcuscale.kfree_alloc_num number
5023 of allocations and frees.
5025 rcuscale.minruntime= [KNL]
5026 Set the minimum test run time in seconds. This
5027 does not affect the data-collection interval,
5028 but instead allows better measurement of things
5029 like CPU consumption.
5031 rcuscale.nreaders= [KNL]
5032 Set number of RCU readers. The value -1 selects
5033 N, where N is the number of CPUs. A value
5034 "n" less than -1 selects N-n+1, where N is again
5035 the number of CPUs. For example, -2 selects N
5036 (the number of CPUs), -3 selects N+1, and so on.
5037 A value of "n" less than or equal to -N selects
5040 rcuscale.nwriters= [KNL]
5041 Set number of RCU writers. The values operate
5042 the same as for rcuscale.nreaders.
5043 N, where N is the number of CPUs
5045 rcuscale.scale_type= [KNL]
5046 Specify the RCU implementation to test.
5048 rcuscale.shutdown= [KNL]
5049 Shut the system down after performance tests
5050 complete. This is useful for hands-off automated
5053 rcuscale.verbose= [KNL]
5054 Enable additional printk() statements.
5056 rcuscale.writer_holdoff= [KNL]
5057 Write-side holdoff between grace periods,
5058 in microseconds. The default of zero says
5061 rcuscale.writer_holdoff_jiffies= [KNL]
5062 Additional write-side holdoff between grace
5063 periods, but in jiffies. The default of zero
5066 rcutorture.fqs_duration= [KNL]
5067 Set duration of force_quiescent_state bursts
5070 rcutorture.fqs_holdoff= [KNL]
5071 Set holdoff time within force_quiescent_state bursts
5074 rcutorture.fqs_stutter= [KNL]
5075 Set wait time between force_quiescent_state bursts
5078 rcutorture.fwd_progress= [KNL]
5079 Specifies the number of kthreads to be used
5080 for RCU grace-period forward-progress testing
5081 for the types of RCU supporting this notion.
5082 Defaults to 1 kthread, values less than zero or
5083 greater than the number of CPUs cause the number
5086 rcutorture.fwd_progress_div= [KNL]
5087 Specify the fraction of a CPU-stall-warning
5088 period to do tight-loop forward-progress testing.
5090 rcutorture.fwd_progress_holdoff= [KNL]
5091 Number of seconds to wait between successive
5092 forward-progress tests.
5094 rcutorture.fwd_progress_need_resched= [KNL]
5095 Enclose cond_resched() calls within checks for
5096 need_resched() during tight-loop forward-progress
5099 rcutorture.gp_cond= [KNL]
5100 Use conditional/asynchronous update-side
5101 primitives, if available.
5103 rcutorture.gp_exp= [KNL]
5104 Use expedited update-side primitives, if available.
5106 rcutorture.gp_normal= [KNL]
5107 Use normal (non-expedited) asynchronous
5108 update-side primitives, if available.
5110 rcutorture.gp_sync= [KNL]
5111 Use normal (non-expedited) synchronous
5112 update-side primitives, if available. If all
5113 of rcutorture.gp_cond=, rcutorture.gp_exp=,
5114 rcutorture.gp_normal=, and rcutorture.gp_sync=
5115 are zero, rcutorture acts as if is interpreted
5116 they are all non-zero.
5118 rcutorture.irqreader= [KNL]
5119 Run RCU readers from irq handlers, or, more
5120 accurately, from a timer handler. Not all RCU
5121 flavors take kindly to this sort of thing.
5123 rcutorture.leakpointer= [KNL]
5124 Leak an RCU-protected pointer out of the reader.
5125 This can of course result in splats, and is
5126 intended to test the ability of things like
5127 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
5130 rcutorture.n_barrier_cbs= [KNL]
5131 Set callbacks/threads for rcu_barrier() testing.
5133 rcutorture.nfakewriters= [KNL]
5134 Set number of concurrent RCU writers. These just
5135 stress RCU, they don't participate in the actual
5136 test, hence the "fake".
5138 rcutorture.nocbs_nthreads= [KNL]
5139 Set number of RCU callback-offload togglers.
5140 Zero (the default) disables toggling.
5142 rcutorture.nocbs_toggle= [KNL]
5143 Set the delay in milliseconds between successive
5144 callback-offload toggling attempts.
5146 rcutorture.nreaders= [KNL]
5147 Set number of RCU readers. The value -1 selects
5148 N-1, where N is the number of CPUs. A value
5149 "n" less than -1 selects N-n-2, where N is again
5150 the number of CPUs. For example, -2 selects N
5151 (the number of CPUs), -3 selects N+1, and so on.
5153 rcutorture.object_debug= [KNL]
5154 Enable debug-object double-call_rcu() testing.
5156 rcutorture.onoff_holdoff= [KNL]
5157 Set time (s) after boot for CPU-hotplug testing.
5159 rcutorture.onoff_interval= [KNL]
5160 Set time (jiffies) between CPU-hotplug operations,
5161 or zero to disable CPU-hotplug testing.
5163 rcutorture.read_exit= [KNL]
5164 Set the number of read-then-exit kthreads used
5165 to test the interaction of RCU updaters and
5166 task-exit processing.
5168 rcutorture.read_exit_burst= [KNL]
5169 The number of times in a given read-then-exit
5170 episode that a set of read-then-exit kthreads
5173 rcutorture.read_exit_delay= [KNL]
5174 The delay, in seconds, between successive
5175 read-then-exit testing episodes.
5177 rcutorture.shuffle_interval= [KNL]
5178 Set task-shuffle interval (s). Shuffling tasks
5179 allows some CPUs to go into dyntick-idle mode
5180 during the rcutorture test.
5182 rcutorture.shutdown_secs= [KNL]
5183 Set time (s) after boot system shutdown. This
5184 is useful for hands-off automated testing.
5186 rcutorture.stall_cpu= [KNL]
5187 Duration of CPU stall (s) to test RCU CPU stall
5188 warnings, zero to disable.
5190 rcutorture.stall_cpu_block= [KNL]
5191 Sleep while stalling if set. This will result
5192 in warnings from preemptible RCU in addition to
5193 any other stall-related activity. Note that
5194 in kernels built with CONFIG_PREEMPTION=n and
5195 CONFIG_PREEMPT_COUNT=y, this parameter will
5196 cause the CPU to pass through a quiescent state.
5197 Given CONFIG_PREEMPTION=n, this will suppress
5198 RCU CPU stall warnings, but will instead result
5199 in scheduling-while-atomic splats.
5201 Use of this module parameter results in splats.
5204 rcutorture.stall_cpu_holdoff= [KNL]
5205 Time to wait (s) after boot before inducing stall.
5207 rcutorture.stall_cpu_irqsoff= [KNL]
5208 Disable interrupts while stalling if set.
5210 rcutorture.stall_gp_kthread= [KNL]
5211 Duration (s) of forced sleep within RCU
5212 grace-period kthread to test RCU CPU stall
5213 warnings, zero to disable. If both stall_cpu
5214 and stall_gp_kthread are specified, the
5215 kthread is starved first, then the CPU.
5217 rcutorture.stat_interval= [KNL]
5218 Time (s) between statistics printk()s.
5220 rcutorture.stutter= [KNL]
5221 Time (s) to stutter testing, for example, specifying
5222 five seconds causes the test to run for five seconds,
5223 wait for five seconds, and so on. This tests RCU's
5224 ability to transition abruptly to and from idle.
5226 rcutorture.test_boost= [KNL]
5227 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5228 "Maybe" means test if the RCU implementation
5229 under test support RCU priority boosting.
5231 rcutorture.test_boost_duration= [KNL]
5232 Duration (s) of each individual boost test.
5234 rcutorture.test_boost_interval= [KNL]
5235 Interval (s) between each boost test.
5237 rcutorture.test_no_idle_hz= [KNL]
5238 Test RCU's dyntick-idle handling. See also the
5239 rcutorture.shuffle_interval parameter.
5241 rcutorture.torture_type= [KNL]
5242 Specify the RCU implementation to test.
5244 rcutorture.verbose= [KNL]
5245 Enable additional printk() statements.
5247 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5248 Dump ftrace buffer after reporting RCU CPU
5251 rcupdate.rcu_cpu_stall_suppress= [KNL]
5252 Suppress RCU CPU stall warning messages.
5254 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5255 Suppress RCU CPU stall warning messages and
5256 rcutorture writer stall warnings that occur
5257 during early boot, that is, during the time
5258 before the init task is spawned.
5260 rcupdate.rcu_cpu_stall_timeout= [KNL]
5261 Set timeout for RCU CPU stall warning messages.
5262 The value is in seconds and the maximum allowed
5263 value is 300 seconds.
5265 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5266 Set timeout for expedited RCU CPU stall warning
5267 messages. The value is in milliseconds
5268 and the maximum allowed value is 21000
5269 milliseconds. Please note that this value is
5270 adjusted to an arch timer tick resolution.
5271 Setting this to zero causes the value from
5272 rcupdate.rcu_cpu_stall_timeout to be used (after
5273 conversion from seconds to milliseconds).
5275 rcupdate.rcu_cpu_stall_cputime= [KNL]
5276 Provide statistics on the cputime and count of
5277 interrupts and tasks during the sampling period. For
5278 multiple continuous RCU stalls, all sampling periods
5279 begin at half of the first RCU stall timeout.
5281 rcupdate.rcu_exp_stall_task_details= [KNL]
5282 Print stack dumps of any tasks blocking the
5283 current expedited RCU grace period during an
5284 expedited RCU CPU stall warning.
5286 rcupdate.rcu_expedited= [KNL]
5287 Use expedited grace-period primitives, for
5288 example, synchronize_rcu_expedited() instead
5289 of synchronize_rcu(). This reduces latency,
5290 but can increase CPU utilization, degrade
5291 real-time latency, and degrade energy efficiency.
5292 No effect on CONFIG_TINY_RCU kernels.
5294 rcupdate.rcu_normal= [KNL]
5295 Use only normal grace-period primitives,
5296 for example, synchronize_rcu() instead of
5297 synchronize_rcu_expedited(). This improves
5298 real-time latency, CPU utilization, and
5299 energy efficiency, but can expose users to
5300 increased grace-period latency. This parameter
5301 overrides rcupdate.rcu_expedited. No effect on
5302 CONFIG_TINY_RCU kernels.
5304 rcupdate.rcu_normal_after_boot= [KNL]
5305 Once boot has completed (that is, after
5306 rcu_end_inkernel_boot() has been invoked), use
5307 only normal grace-period primitives. No effect
5308 on CONFIG_TINY_RCU kernels.
5310 But note that CONFIG_PREEMPT_RT=y kernels enables
5311 this kernel boot parameter, forcibly setting
5312 it to the value one, that is, converting any
5313 post-boot attempt at an expedited RCU grace
5314 period to instead use normal non-expedited
5315 grace-period processing.
5317 rcupdate.rcu_task_collapse_lim= [KNL]
5318 Set the maximum number of callbacks present
5319 at the beginning of a grace period that allows
5320 the RCU Tasks flavors to collapse back to using
5321 a single callback queue. This switching only
5322 occurs when rcupdate.rcu_task_enqueue_lim is
5323 set to the default value of -1.
5325 rcupdate.rcu_task_contend_lim= [KNL]
5326 Set the minimum number of callback-queuing-time
5327 lock-contention events per jiffy required to
5328 cause the RCU Tasks flavors to switch to per-CPU
5329 callback queuing. This switching only occurs
5330 when rcupdate.rcu_task_enqueue_lim is set to
5331 the default value of -1.
5333 rcupdate.rcu_task_enqueue_lim= [KNL]
5334 Set the number of callback queues to use for the
5335 RCU Tasks family of RCU flavors. The default
5336 of -1 allows this to be automatically (and
5337 dynamically) adjusted. This parameter is intended
5340 rcupdate.rcu_task_ipi_delay= [KNL]
5341 Set time in jiffies during which RCU tasks will
5342 avoid sending IPIs, starting with the beginning
5343 of a given grace period. Setting a large
5344 number avoids disturbing real-time workloads,
5345 but lengthens grace periods.
5347 rcupdate.rcu_task_lazy_lim= [KNL]
5348 Number of callbacks on a given CPU that will
5349 cancel laziness on that CPU. Use -1 to disable
5350 cancellation of laziness, but be advised that
5351 doing so increases the danger of OOM due to
5354 rcupdate.rcu_task_stall_info= [KNL]
5355 Set initial timeout in jiffies for RCU task stall
5356 informational messages, which give some indication
5357 of the problem for those not patient enough to
5358 wait for ten minutes. Informational messages are
5359 only printed prior to the stall-warning message
5360 for a given grace period. Disable with a value
5361 less than or equal to zero. Defaults to ten
5362 seconds. A change in value does not take effect
5363 until the beginning of the next grace period.
5365 rcupdate.rcu_task_stall_info_mult= [KNL]
5366 Multiplier for time interval between successive
5367 RCU task stall informational messages for a given
5368 RCU tasks grace period. This value is clamped
5369 to one through ten, inclusive. It defaults to
5370 the value three, so that the first informational
5371 message is printed 10 seconds into the grace
5372 period, the second at 40 seconds, the third at
5373 160 seconds, and then the stall warning at 600
5374 seconds would prevent a fourth at 640 seconds.
5376 rcupdate.rcu_task_stall_timeout= [KNL]
5377 Set timeout in jiffies for RCU task stall
5378 warning messages. Disable with a value less
5379 than or equal to zero. Defaults to ten minutes.
5380 A change in value does not take effect until
5381 the beginning of the next grace period.
5383 rcupdate.rcu_tasks_lazy_ms= [KNL]
5384 Set timeout in milliseconds RCU Tasks asynchronous
5385 callback batching for call_rcu_tasks().
5386 A negative value will take the default. A value
5387 of zero will disable batching. Batching is
5388 always disabled for synchronize_rcu_tasks().
5390 rcupdate.rcu_tasks_rude_lazy_ms= [KNL]
5391 Set timeout in milliseconds RCU Tasks
5392 Rude asynchronous callback batching for
5393 call_rcu_tasks_rude(). A negative value
5394 will take the default. A value of zero will
5395 disable batching. Batching is always disabled
5396 for synchronize_rcu_tasks_rude().
5398 rcupdate.rcu_tasks_trace_lazy_ms= [KNL]
5399 Set timeout in milliseconds RCU Tasks
5400 Trace asynchronous callback batching for
5401 call_rcu_tasks_trace(). A negative value
5402 will take the default. A value of zero will
5403 disable batching. Batching is always disabled
5404 for synchronize_rcu_tasks_trace().
5406 rcupdate.rcu_self_test= [KNL]
5407 Run the RCU early boot self tests
5411 Run specified binary instead of /init from the ramdisk,
5412 used for early userspace startup. See initrd.
5415 force - Override the decision by the kernel to hide the
5416 advertisement of RDRAND support (this affects
5417 certain AMD processors because of buggy BIOS
5418 support, specifically around the suspend/resume
5422 Turn on/off individual RDT features. List is:
5423 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5425 E.g. to turn on cmt and turn off mba use:
5429 Format (x86 or x86_64):
5430 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5432 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5434 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5435 (prefix with 'panic_' to set mode for panic
5437 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5438 reboot_force is either force or not specified,
5439 reboot_cpu is s[mp]#### with #### being the processor
5440 to be used for rebooting.
5442 refscale.holdoff= [KNL]
5443 Set test-start holdoff period. The purpose of
5444 this parameter is to delay the start of the
5445 test until boot completes in order to avoid
5448 refscale.loops= [KNL]
5449 Set the number of loops over the synchronization
5450 primitive under test. Increasing this number
5451 reduces noise due to loop start/end overhead,
5452 but the default has already reduced the per-pass
5453 noise to a handful of picoseconds on ca. 2020
5456 refscale.nreaders= [KNL]
5457 Set number of readers. The default value of -1
5458 selects N, where N is roughly 75% of the number
5459 of CPUs. A value of zero is an interesting choice.
5461 refscale.nruns= [KNL]
5462 Set number of runs, each of which is dumped onto
5465 refscale.readdelay= [KNL]
5466 Set the read-side critical-section duration,
5467 measured in microseconds.
5469 refscale.scale_type= [KNL]
5470 Specify the read-protection implementation to test.
5472 refscale.shutdown= [KNL]
5473 Shut down the system at the end of the performance
5474 test. This defaults to 1 (shut it down) when
5475 refscale is built into the kernel and to 0 (leave
5476 it running) when refscale is built as a module.
5478 refscale.verbose= [KNL]
5479 Enable additional printk() statements.
5481 refscale.verbose_batched= [KNL]
5482 Batch the additional printk() statements. If zero
5483 (the default) or negative, print everything. Otherwise,
5484 print every Nth verbose statement, where N is the value
5488 [KNL, SMP] Set scheduler's default relax_domain_level.
5489 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5491 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5492 Format: <base1>,<size1>[,<base2>,<size2>,...]
5493 Reserve I/O ports or memory so the kernel won't use
5494 them. If <base> is less than 0x10000, the region
5495 is assumed to be I/O ports; otherwise it is memory.
5497 reservetop= [X86-32]
5499 Reserves a hole at the top of the kernel virtual
5502 reset_devices [KNL] Force drivers to reset the underlying device
5503 during initialization.
5506 Specify the partition device for software suspend
5508 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5510 resume_offset= [SWSUSP]
5511 Specify the offset from the beginning of the partition
5512 given by "resume=" at which the swap header is located,
5513 in <PAGE_SIZE> units (needed only for swap files).
5514 See Documentation/power/swsusp-and-swap-files.rst
5516 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5517 read the resume files
5519 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5520 Useful for devices that are detected asynchronously
5521 (e.g. USB and MMC devices).
5523 retain_initrd [RAM] Keep initrd memory after extraction
5525 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5526 Speculative Code Execution with Return Instructions)
5529 AMD-based UNRET and IBPB mitigations alone do not stop
5530 sibling threads from influencing the predictions of other
5531 sibling threads. For that reason, STIBP is used on pro-
5532 cessors that support it, and mitigate SMT on processors
5536 auto - automatically select a migitation
5537 auto,nosmt - automatically select a mitigation,
5538 disabling SMT if necessary for
5539 the full mitigation (only on Zen1
5540 and older without STIBP).
5541 ibpb - On AMD, mitigate short speculation
5542 windows on basic block boundaries too.
5543 Safe, highest perf impact. It also
5544 enables STIBP if present. Not suitable
5546 ibpb,nosmt - Like "ibpb" above but will disable SMT
5547 when STIBP is not available. This is
5548 the alternative for systems which do not
5550 unret - Force enable untrained return thunks,
5551 only effective on AMD f15h-f17h based
5553 unret,nosmt - Like unret, but will disable SMT when STIBP
5554 is not available. This is the alternative for
5555 systems which do not have STIBP.
5557 Selecting 'auto' will choose a mitigation method at run
5558 time according to the CPU.
5560 Not specifying this option is equivalent to retbleed=auto.
5562 rfkill.default_state=
5563 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5564 etc. communication is blocked by default.
5567 rfkill.master_switch_mode=
5568 0 The "airplane mode" button does nothing.
5569 1 The "airplane mode" button toggles between everything
5570 blocked and the previous configuration.
5571 2 The "airplane mode" button toggles between everything
5572 blocked and everything unblocked.
5574 rhash_entries= [KNL,NET]
5575 Set number of hash buckets for route cache
5578 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5581 riscv_isa_fallback [RISCV]
5582 When CONFIG_RISCV_ISA_FALLBACK is not enabled, permit
5583 falling back to detecting extension support by parsing
5584 "riscv,isa" property on devicetree systems when the
5585 replacement properties are not found. See the Kconfig
5586 entry for RISCV_ISA_FALLBACK.
5588 ro [KNL] Mount root device read-only on boot
5591 on Mark read-only kernel memory as read-only (default).
5592 off Leave read-only kernel memory writable for debugging.
5593 full Mark read-only kernel memory and aliases as read-only
5597 Enable the uart passthrough on the designated usb port
5598 on Rockchip SoCs. When active, the signals of the
5599 debug-uart get routed to the D+ and D- pins of the usb
5600 port and the regular usb controller gets disabled.
5602 root= [KNL] Root filesystem
5603 Usually this a a block device specifier of some kind,
5604 see the early_lookup_bdev comment in
5605 block/early-lookup.c for details.
5606 Alternatively this can be "ram" for the legacy initial
5607 ramdisk, "nfs" and "cifs" for root on a network file
5608 system, or "mtd" and "ubi" for mounting from raw flash.
5610 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5611 mount the root filesystem
5613 rootflags= [KNL] Set root filesystem mount option string
5615 rootfstype= [KNL] Set root filesystem type
5617 rootwait [KNL] Wait (indefinitely) for root device to show up.
5618 Useful for devices that are detected asynchronously
5619 (e.g. USB and MMC devices).
5621 rootwait= [KNL] Maximum time (in seconds) to wait for root device
5622 to show up before attempting to mount the root
5625 rproc_mem=nn[KMG][@address]
5626 [KNL,ARM,CMA] Remoteproc physical memory block.
5627 Memory area to be used by remote processor image,
5630 rw [KNL] Mount root device read-write on boot
5632 S [KNL] Run init in single mode
5634 s390_iommu= [HW,S390]
5635 Set s390 IOTLB flushing mode
5637 With strict flushing every unmap operation will result in
5638 an IOTLB flush. Default is lazy flushing before reuse,
5641 s390_iommu_aperture= [KNL,S390]
5642 Specifies the size of the per device DMA address space
5643 accessible through the DMA and IOMMU APIs as a decimal
5644 factor of the size of main memory.
5645 The default is 1 meaning that one can concurrently use
5646 as many DMA addresses as physical memory is installed,
5647 if supported by hardware, and thus map all of memory
5648 once. With a value of 2 one can map all of memory twice
5649 and so on. As a special case a factor of 0 imposes no
5650 restrictions other than those given by hardware at the
5651 cost of significant additional memory use for tables.
5654 See drivers/net/irda/sa1100_ir.c.
5656 sched_verbose [KNL] Enables verbose scheduler debug messages.
5658 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5659 Allowed values are enable and disable. This feature
5660 incurs a small amount of overhead in the scheduler
5661 but is useful for debugging and performance tuning.
5663 sched_thermal_decay_shift=
5664 [KNL, SMP] Set a decay shift for scheduler thermal
5665 pressure signal. Thermal pressure signal follows the
5666 default decay period of other scheduler pelt
5667 signals(usually 32 ms but configurable). Setting
5668 sched_thermal_decay_shift will left shift the decay
5669 period for the thermal pressure signal by the shift
5671 i.e. with the default pelt decay period of 32 ms
5672 sched_thermal_decay_shift thermal pressure decay pr
5676 Format: integer between 0 and 10
5679 scftorture.holdoff= [KNL]
5680 Number of seconds to hold off before starting
5681 test. Defaults to zero for module insertion and
5682 to 10 seconds for built-in smp_call_function()
5685 scftorture.longwait= [KNL]
5686 Request ridiculously long waits randomly selected
5687 up to the chosen limit in seconds. Zero (the
5688 default) disables this feature. Please note
5689 that requesting even small non-zero numbers of
5690 seconds can result in RCU CPU stall warnings,
5691 softlockup complaints, and so on.
5693 scftorture.nthreads= [KNL]
5694 Number of kthreads to spawn to invoke the
5695 smp_call_function() family of functions.
5696 The default of -1 specifies a number of kthreads
5697 equal to the number of CPUs.
5699 scftorture.onoff_holdoff= [KNL]
5700 Number seconds to wait after the start of the
5701 test before initiating CPU-hotplug operations.
5703 scftorture.onoff_interval= [KNL]
5704 Number seconds to wait between successive
5705 CPU-hotplug operations. Specifying zero (which
5706 is the default) disables CPU-hotplug operations.
5708 scftorture.shutdown_secs= [KNL]
5709 The number of seconds following the start of the
5710 test after which to shut down the system. The
5711 default of zero avoids shutting down the system.
5712 Non-zero values are useful for automated tests.
5714 scftorture.stat_interval= [KNL]
5715 The number of seconds between outputting the
5716 current test statistics to the console. A value
5717 of zero disables statistics output.
5719 scftorture.stutter_cpus= [KNL]
5720 The number of jiffies to wait between each change
5721 to the set of CPUs under test.
5723 scftorture.use_cpus_read_lock= [KNL]
5724 Use use_cpus_read_lock() instead of the default
5725 preempt_disable() to disable CPU hotplug
5726 while invoking one of the smp_call_function*()
5729 scftorture.verbose= [KNL]
5730 Enable additional printk() statements.
5732 scftorture.weight_single= [KNL]
5733 The probability weighting to use for the
5734 smp_call_function_single() function with a zero
5735 "wait" parameter. A value of -1 selects the
5736 default if all other weights are -1. However,
5737 if at least one weight has some other value, a
5738 value of -1 will instead select a weight of zero.
5740 scftorture.weight_single_wait= [KNL]
5741 The probability weighting to use for the
5742 smp_call_function_single() function with a
5743 non-zero "wait" parameter. See weight_single.
5745 scftorture.weight_many= [KNL]
5746 The probability weighting to use for the
5747 smp_call_function_many() function with a zero
5748 "wait" parameter. See weight_single.
5749 Note well that setting a high probability for
5750 this weighting can place serious IPI load
5753 scftorture.weight_many_wait= [KNL]
5754 The probability weighting to use for the
5755 smp_call_function_many() function with a
5756 non-zero "wait" parameter. See weight_single
5759 scftorture.weight_all= [KNL]
5760 The probability weighting to use for the
5761 smp_call_function_all() function with a zero
5762 "wait" parameter. See weight_single and
5765 scftorture.weight_all_wait= [KNL]
5766 The probability weighting to use for the
5767 smp_call_function_all() function with a
5768 non-zero "wait" parameter. See weight_single
5771 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5772 xtime_lock contention on larger systems, and/or RCU lock
5773 contention on all systems with CONFIG_MAXSMP set.
5774 Format: { "0" | "1" }
5775 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5777 Note: increases power consumption, thus should only be
5778 enabled if running jitter sensitive (HPC/RT) workloads.
5780 security= [SECURITY] Choose a legacy "major" security module to
5781 enable at boot. This has been deprecated by the
5784 selinux= [SELINUX] Disable or enable SELinux at boot time.
5785 Format: { "0" | "1" }
5786 See security/selinux/Kconfig help text.
5791 serialnumber [BUGS=X86-32]
5793 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
5796 Maximal number of shapers.
5798 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
5799 Limit apic dumping. The parameter defines the maximal
5800 number of local apics being dumped. Also it is possible
5801 to set it to "all" by meaning -- no limit here.
5802 Format: { 1 (default) | 2 | ... | all }.
5803 The parameter valid if only apic=debug or
5804 apic=verbose is specified.
5805 Example: apic=debug show_lapic=all
5813 Enable merging of slabs with similar size when the
5814 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5817 Disable merging of slabs with similar size. May be
5818 necessary if there is some reason to distinguish
5819 allocs to different slabs, especially in hardened
5820 environments where the risk of heap overflows and
5821 layout control by attackers can usually be
5822 frustrated by disabling merging. This will reduce
5823 most of the exposure of a heap attack to a single
5824 cache (risks via metadata attacks are mostly
5825 unchanged). Debug options disable merging on their
5827 For more information see Documentation/mm/slub.rst.
5829 slab_max_order= [MM, SLAB]
5830 Determines the maximum allowed order for slabs.
5831 A high setting may cause OOMs due to memory
5832 fragmentation. Defaults to 1 for systems with
5833 more than 32MB of RAM, 0 otherwise.
5835 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5836 Enabling slub_debug allows one to determine the
5837 culprit if slab objects become corrupted. Enabling
5838 slub_debug can create guard zones around objects and
5839 may poison objects when not in use. Also tracks the
5840 last alloc / free. For more information see
5841 Documentation/mm/slub.rst.
5843 slub_max_order= [MM, SLUB]
5844 Determines the maximum allowed order for slabs.
5845 A high setting may cause OOMs due to memory
5846 fragmentation. For more information see
5847 Documentation/mm/slub.rst.
5849 slub_min_objects= [MM, SLUB]
5850 The minimum number of objects per slab. SLUB will
5851 increase the slab order up to slub_max_order to
5852 generate a sufficiently large slab able to contain
5853 the number of objects indicated. The higher the number
5854 of objects the smaller the overhead of tracking slabs
5855 and the less frequently locks need to be acquired.
5856 For more information see Documentation/mm/slub.rst.
5858 slub_min_order= [MM, SLUB]
5859 Determines the minimum page order for slabs. Must be
5860 lower than slub_max_order.
5861 For more information see Documentation/mm/slub.rst.
5863 slub_merge [MM, SLUB]
5864 Same with slab_merge.
5866 slub_nomerge [MM, SLUB]
5867 Same with slab_nomerge. This is supported for legacy.
5868 See slab_nomerge for more information.
5871 Format: <io1>[,<io2>[,...,<io8>]]
5873 smp.csd_lock_timeout= [KNL]
5874 Specify the period of time in milliseconds
5875 that smp_call_function() and friends will wait
5876 for a CPU to release the CSD lock. This is
5877 useful when diagnosing bugs involving CPUs
5878 disabling interrupts for extended periods
5879 of time. Defaults to 5,000 milliseconds, and
5880 setting a value of zero disables this feature.
5881 This feature may be more efficiently disabled
5882 using the csdlock_debug- kernel parameter.
5884 smp.panic_on_ipistall= [KNL]
5885 If a csd_lock_timeout extends for more than
5886 the specified number of milliseconds, panic the
5887 system. By default, let CSD-lock acquisition
5888 take as long as they take. Specifying 300,000
5889 for this value provides a 5-minute timeout.
5891 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5892 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5893 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5894 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5895 smsc-ircc2.ircc_irq= [HW] IRQ line
5896 smsc-ircc2.ircc_dma= [HW] DMA channel
5897 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5898 0: Toshiba Satellite 1800 (GP data pin select)
5899 1: Fast pin select (default)
5902 smt= [KNL,MIPS,S390] Set the maximum number of threads (logical
5903 CPUs) to use per physical CPU on systems capable of
5904 symmetric multithreading (SMT). Will be capped to the
5905 actual hardware limit.
5907 Default: -1 (no limit)
5910 [KNL] Should the soft-lockup detector generate panics.
5913 A value of 1 instructs the soft-lockup detector
5914 to panic the machine when a soft-lockup occurs. It is
5915 also controlled by the kernel.softlockup_panic sysctl
5916 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5917 respective build-time switch to that functionality.
5919 softlockup_all_cpu_backtrace=
5920 [KNL] Should the soft-lockup detector generate
5921 backtraces on all cpus.
5924 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5925 See Documentation/admin-guide/laptops/sonypi.rst
5927 spectre_bhi= [X86] Control mitigation of Branch History Injection
5928 (BHI) vulnerability. This setting affects the
5929 deployment of the HW BHI control and the SW BHB
5932 on - (default) Enable the HW or SW mitigation
5934 off - Disable the mitigation.
5936 spectre_v2= [X86] Control mitigation of Spectre variant 2
5937 (indirect branch speculation) vulnerability.
5938 The default operation protects the kernel from
5941 on - unconditionally enable, implies
5943 off - unconditionally disable, implies
5945 auto - kernel detects whether your CPU model is
5948 Selecting 'on' will, and 'auto' may, choose a
5949 mitigation method at run time according to the
5950 CPU, the available microcode, the setting of the
5951 CONFIG_RETPOLINE configuration option, and the
5952 compiler with which the kernel was built.
5954 Selecting 'on' will also enable the mitigation
5955 against user space to user space task attacks.
5957 Selecting 'off' will disable both the kernel and
5958 the user space protections.
5960 Specific mitigations can also be selected manually:
5962 retpoline - replace indirect branches
5963 retpoline,generic - Retpolines
5964 retpoline,lfence - LFENCE; indirect branch
5965 retpoline,amd - alias for retpoline,lfence
5966 eibrs - Enhanced/Auto IBRS
5967 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines
5968 eibrs,lfence - Enhanced/Auto IBRS + LFENCE
5969 ibrs - use IBRS to protect kernel
5971 Not specifying this option is equivalent to
5975 [X86] Control mitigation of Spectre variant 2
5976 (indirect branch speculation) vulnerability between
5979 on - Unconditionally enable mitigations. Is
5980 enforced by spectre_v2=on
5982 off - Unconditionally disable mitigations. Is
5983 enforced by spectre_v2=off
5985 prctl - Indirect branch speculation is enabled,
5986 but mitigation can be enabled via prctl
5987 per thread. The mitigation control state
5988 is inherited on fork.
5991 - Like "prctl" above, but only STIBP is
5992 controlled per thread. IBPB is issued
5993 always when switching between different user
5997 - Same as "prctl" above, but all seccomp
5998 threads will enable the mitigation unless
5999 they explicitly opt out.
6002 - Like "seccomp" above, but only STIBP is
6003 controlled per thread. IBPB is issued
6004 always when switching between different
6005 user space processes.
6007 auto - Kernel selects the mitigation depending on
6008 the available CPU features and vulnerability.
6010 Default mitigation: "prctl"
6012 Not specifying this option is equivalent to
6013 spectre_v2_user=auto.
6015 spec_rstack_overflow=
6016 [X86] Control RAS overflow mitigation on AMD Zen CPUs
6018 off - Disable mitigation
6019 microcode - Enable microcode mitigation only
6020 safe-ret - Enable sw-only safe RET mitigation (default)
6021 ibpb - Enable mitigation by issuing IBPB on
6023 ibpb-vmexit - Issue IBPB only on VMEXIT
6024 (cloud-specific mitigation)
6026 spec_store_bypass_disable=
6027 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
6028 (Speculative Store Bypass vulnerability)
6030 Certain CPUs are vulnerable to an exploit against a
6031 a common industry wide performance optimization known
6032 as "Speculative Store Bypass" in which recent stores
6033 to the same memory location may not be observed by
6034 later loads during speculative execution. The idea
6035 is that such stores are unlikely and that they can
6036 be detected prior to instruction retirement at the
6037 end of a particular speculation execution window.
6039 In vulnerable processors, the speculatively forwarded
6040 store can be used in a cache side channel attack, for
6041 example to read memory to which the attacker does not
6042 directly have access (e.g. inside sandboxed code).
6044 This parameter controls whether the Speculative Store
6045 Bypass optimization is used.
6047 On x86 the options are:
6049 on - Unconditionally disable Speculative Store Bypass
6050 off - Unconditionally enable Speculative Store Bypass
6051 auto - Kernel detects whether the CPU model contains an
6052 implementation of Speculative Store Bypass and
6053 picks the most appropriate mitigation. If the
6054 CPU is not vulnerable, "off" is selected. If the
6055 CPU is vulnerable the default mitigation is
6056 architecture and Kconfig dependent. See below.
6057 prctl - Control Speculative Store Bypass per thread
6058 via prctl. Speculative Store Bypass is enabled
6059 for a process by default. The state of the control
6060 is inherited on fork.
6061 seccomp - Same as "prctl" above, but all seccomp threads
6062 will disable SSB unless they explicitly opt out.
6064 Default mitigations:
6067 On powerpc the options are:
6069 on,auto - On Power8 and Power9 insert a store-forwarding
6070 barrier on kernel entry and exit. On Power7
6071 perform a software flush on kernel entry and
6075 Not specifying this option is equivalent to
6076 spec_store_bypass_disable=auto.
6078 spia_io_base= [HW,MTD]
6084 [X86] Enable split lock detection or bus lock detection
6086 When enabled (and if hardware support is present), atomic
6087 instructions that access data across cache line
6088 boundaries will result in an alignment check exception
6089 for split lock detection or a debug exception for
6094 warn - the kernel will emit rate-limited warnings
6095 about applications triggering the #AC
6096 exception or the #DB exception. This mode is
6097 the default on CPUs that support split lock
6098 detection or bus lock detection. Default
6099 behavior is by #AC if both features are
6100 enabled in hardware.
6102 fatal - the kernel will send SIGBUS to applications
6103 that trigger the #AC exception or the #DB
6104 exception. Default behavior is by #AC if
6105 both features are enabled in hardware.
6108 Set system wide rate limit to N bus locks
6109 per second for bus lock detection.
6112 N/A for split lock detection.
6115 If an #AC exception is hit in the kernel or in
6116 firmware (i.e. not while executing in user mode)
6117 the kernel will oops in either "warn" or "fatal"
6120 #DB exception for bus lock is triggered only when
6124 Control the Special Register Buffer Data Sampling
6127 Certain CPUs are vulnerable to an MDS-like
6128 exploit which can leak bits from the random
6131 By default, this issue is mitigated by
6132 microcode. However, the microcode fix can cause
6133 the RDRAND and RDSEED instructions to become
6134 much slower. Among other effects, this will
6135 result in reduced throughput from /dev/urandom.
6137 The microcode mitigation can be disabled with
6138 the following option:
6140 off: Disable mitigation and remove
6141 performance impact to RDRAND and RDSEED
6143 srcutree.big_cpu_lim [KNL]
6144 Specifies the number of CPUs constituting a
6145 large system, such that srcu_struct structures
6146 should immediately allocate an srcu_node array.
6147 This kernel-boot parameter defaults to 128,
6148 but takes effect only when the low-order four
6149 bits of srcutree.convert_to_big is equal to 3
6152 srcutree.convert_to_big [KNL]
6153 Specifies under what conditions an SRCU tree
6154 srcu_struct structure will be converted to big
6155 form, that is, with an rcu_node tree:
6158 1: At init_srcu_struct() time.
6159 2: When rcutorture decides to.
6160 3: Decide at boot time (default).
6161 0x1X: Above plus if high contention.
6163 Either way, the srcu_node tree will be sized based
6164 on the actual runtime number of CPUs (nr_cpu_ids)
6165 instead of the compile-time CONFIG_NR_CPUS.
6167 srcutree.counter_wrap_check [KNL]
6168 Specifies how frequently to check for
6169 grace-period sequence counter wrap for the
6170 srcu_data structure's ->srcu_gp_seq_needed field.
6171 The greater the number of bits set in this kernel
6172 parameter, the less frequently counter wrap will
6173 be checked for. Note that the bottom two bits
6176 srcutree.exp_holdoff [KNL]
6177 Specifies how many nanoseconds must elapse
6178 since the end of the last SRCU grace period for
6179 a given srcu_struct until the next normal SRCU
6180 grace period will be considered for automatic
6181 expediting. Set to zero to disable automatic
6184 srcutree.srcu_max_nodelay [KNL]
6185 Specifies the number of no-delay instances
6186 per jiffy for which the SRCU grace period
6187 worker thread will be rescheduled with zero
6188 delay. Beyond this limit, worker thread will
6189 be rescheduled with a sleep delay of one jiffy.
6191 srcutree.srcu_max_nodelay_phase [KNL]
6192 Specifies the per-grace-period phase, number of
6193 non-sleeping polls of readers. Beyond this limit,
6194 grace period worker thread will be rescheduled
6195 with a sleep delay of one jiffy, between each
6196 rescan of the readers, for a grace period phase.
6198 srcutree.srcu_retry_check_delay [KNL]
6199 Specifies number of microseconds of non-sleeping
6200 delay between each non-sleeping poll of readers.
6202 srcutree.small_contention_lim [KNL]
6203 Specifies the number of update-side contention
6204 events per jiffy will be tolerated before
6205 initiating a conversion of an srcu_struct
6206 structure to big form. Note that the value of
6207 srcutree.convert_to_big must have the 0x10 bit
6208 set for contention-based conversions to occur.
6211 Speculative Store Bypass Disable control
6213 On CPUs that are vulnerable to the Speculative
6214 Store Bypass vulnerability and offer a
6215 firmware based mitigation, this parameter
6216 indicates how the mitigation should be used:
6218 force-on: Unconditionally enable mitigation for
6219 for both kernel and userspace
6220 force-off: Unconditionally disable mitigation for
6221 for both kernel and userspace
6222 kernel: Always enable mitigation in the
6223 kernel, and offer a prctl interface
6224 to allow userspace to register its
6225 interest in being mitigated too.
6227 stack_guard_gap= [MM]
6228 override the default stack gap protection. The value
6229 is in page units and it defines how many pages prior
6230 to (for stacks growing down) resp. after (for stacks
6231 growing up) the main stack are reserved for no other
6232 mapping. Default value is 256 pages.
6234 stack_depot_disable= [KNL]
6235 Setting this to true through kernel command line will
6236 disable the stack depot thereby saving the static memory
6237 consumed by the stack hash table. By default this is set
6241 Enabled the stack tracer on boot up.
6243 stacktrace_filter=[function-list]
6244 [FTRACE] Limit the functions that the stack tracer
6245 will trace at boot up. function-list is a comma-separated
6246 list of functions. This list can be changed at run
6247 time by the stack_trace_filter file in the debugfs
6248 tracing directory. Note, this enables stack tracing
6249 and the stacktrace above is not needed.
6253 Set the STI (builtin display/keyboard on the HP-PARISC
6254 machines) console (graphic card) which should be used
6255 as the initial boot-console.
6256 See also comment in drivers/video/console/sticore.c.
6259 See comment in drivers/video/console/sticore.c.
6262 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
6267 Enable or disable strict sigaltstack size checks
6268 against the required signal frame size which
6269 depends on the supported FPU features. This can
6270 be used to filter out binaries which have
6271 not yet been made aware of AT_MINSIGSTKSZ.
6274 Limits the number of kernel HPT entries in the hash
6275 page table to increase the rate of hash page table
6276 faults on kernel addresses.
6279 Limits the number of kernel SLB entries, and flushes
6280 them frequently to increase the rate of SLB faults
6281 on kernel addresses.
6283 sunrpc.min_resvport=
6284 sunrpc.max_resvport=
6286 SunRPC servers often require that client requests
6287 originate from a privileged port (i.e. a port in the
6288 range 0 < portnr < 1024).
6289 An administrator who wishes to reserve some of these
6290 ports for other uses may adjust the range that the
6291 kernel's sunrpc client considers to be privileged
6292 using these two parameters to set the minimum and
6293 maximum port values.
6295 sunrpc.svc_rpc_per_connection_limit=
6297 Limit the number of requests that the server will
6298 process in parallel from a single connection.
6299 The default value is 0 (no limit).
6303 Control how the NFS server code allocates CPUs to
6304 service thread pools. Depending on how many NICs
6305 you have and where their interrupts are bound, this
6306 option will affect which CPUs will do NFS serving.
6307 Note: this parameter cannot be changed while the
6308 NFS server is running.
6310 auto the server chooses an appropriate mode
6311 automatically using heuristics
6312 global a single global pool contains all CPUs
6313 percpu one pool for each CPU
6314 pernode one pool for each NUMA node (equivalent
6315 to global on non-NUMA machines)
6317 sunrpc.tcp_slot_table_entries=
6318 sunrpc.udp_slot_table_entries=
6320 Sets the upper limit on the number of simultaneous
6321 RPC calls that can be sent from the client to a
6322 server. Increasing these values may allow you to
6323 improve throughput, but will also increase the
6324 amount of memory reserved for use by the client.
6326 suspend.pm_test_delay=
6328 Sets the number of seconds to remain in a suspend test
6329 mode before resuming the system (see
6330 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6331 is set. Default value is 5.
6334 Format: { on | off | y | n | 1 | 0 }
6335 This parameter controls use of the Protected
6336 Execution Facility on pSeries.
6338 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
6339 Format: { <int> [,<int>] | force | noforce }
6340 <int> -- Number of I/O TLB slabs
6341 <int> -- Second integer after comma. Number of swiotlb
6342 areas with their own lock. Will be rounded up
6344 force -- force using of bounce buffers even if they
6345 wouldn't be automatically used by the kernel
6346 noforce -- Never use bounce buffers (for debugging)
6351 Set a sysctl parameter, right before loading the init
6352 process, as if the value was written to the respective
6353 /proc/sys/... file. Both '.' and '/' are recognized as
6354 separators. Unrecognized parameters and invalid values
6355 are reported in the kernel log. Sysctls registered
6356 later by a loaded module cannot be set this way.
6357 Example: sysctl.vm.swappiness=40
6359 sysrq_always_enabled
6361 Ignore sysrq setting - this boot parameter will
6362 neutralize any effect of /proc/sys/kernel/sysrq.
6363 Useful for debugging.
6365 tcpmhash_entries= [KNL,NET]
6366 Set the number of tcp_metrics_hash slots.
6367 Default value is 8192 or 16384 depending on total
6368 ram pages. This is used to specify the TCP metrics
6369 cache size. See Documentation/networking/ip-sysctl.rst
6370 "tcp_no_metrics_save" section for more details.
6374 test_suspend= [SUSPEND]
6375 Format: { "mem" | "standby" | "freeze" }[,N]
6376 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6377 standby suspend) or "freeze" (for suspend type freeze)
6378 as the system sleep state during system startup with
6379 the optional capability to repeat N number of times.
6380 The system is woken from this state using a
6381 wakeup-capable RTC alarm.
6383 thash_entries= [KNL,NET]
6384 Set number of hash buckets for TCP connection
6386 thermal.act= [HW,ACPI]
6387 -1: disable all active trip points in all thermal zones
6388 <degrees C>: override all lowest active trip points
6390 thermal.crt= [HW,ACPI]
6391 -1: disable all critical trip points in all thermal zones
6392 <degrees C>: override all critical trip points
6394 thermal.off= [HW,ACPI]
6395 1: disable ACPI thermal control
6397 thermal.psv= [HW,ACPI]
6398 -1: disable all passive trip points
6399 <degrees C>: override all passive trip points to this
6402 thermal.tzp= [HW,ACPI]
6403 Specify global default ACPI thermal zone polling rate
6404 <deci-seconds>: poll all this frequency
6405 0: no polling (default)
6408 Force threading of all interrupt handlers except those
6409 marked explicitly IRQF_NO_THREAD.
6413 Specify if the kernel should make use of the cpu
6414 topology information if the hardware supports this.
6415 The scheduler will make use of this information and
6416 e.g. base its process migration decisions on it.
6419 topology_updates= [KNL, PPC, NUMA]
6421 Specify if the kernel should ignore (off)
6422 topology updates sent by the hypervisor to this
6425 torture.disable_onoff_at_boot= [KNL]
6426 Prevent the CPU-hotplug component of torturing
6427 until after init has spawned.
6429 torture.ftrace_dump_at_shutdown= [KNL]
6430 Dump the ftrace buffer at torture-test shutdown,
6431 even if there were no errors. This can be a
6432 very costly operation when many torture tests
6433 are running concurrently, especially on systems
6434 with rotating-rust storage.
6436 torture.verbose_sleep_frequency= [KNL]
6437 Specifies how many verbose printk()s should be
6438 emitted between each sleep. The default of zero
6439 disables verbose-printk() sleeping.
6441 torture.verbose_sleep_duration= [KNL]
6442 Duration of each verbose-printk() sleep in jiffies.
6446 tpm_suspend_pcr=[HW,TPM]
6447 Format: integer pcr id
6448 Specify that at suspend time, the tpm driver
6449 should extend the specified pcr with zeros,
6450 as a workaround for some chips which fail to
6451 flush the last written pcr on TPM_SaveState.
6452 This will guarantee that all the other pcrs
6455 tpm_tis.interrupts= [HW,TPM]
6456 Enable interrupts for the MMIO based physical layer
6457 for the FIFO interface. By default it is set to false
6458 (0). For more information about TPM hardware interfaces
6459 defined by Trusted Computing Group (TCG) see
6460 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/
6463 Have the tracepoints sent to printk as well as the
6464 tracing ring buffer. This is useful for early boot up
6465 where the system hangs or reboots and does not give the
6466 option for reading the tracing buffer or performing a
6467 ftrace_dump_on_oops.
6469 To turn off having tracepoints sent to printk,
6470 echo 0 > /proc/sys/kernel/tracepoint_printk
6471 Note, echoing 1 into this file without the
6472 tracepoint_printk kernel cmdline option has no effect.
6474 The tp_printk_stop_on_boot (see below) can also be used
6475 to stop the printing of events to console at
6480 Having tracepoints sent to printk() and activating high
6481 frequency tracepoints such as irq or sched, can cause
6482 the system to live lock.
6484 tp_printk_stop_on_boot [FTRACE]
6485 When tp_printk (above) is set, it can cause a lot of noise
6486 on the console. It may be useful to only include the
6487 printing of events during boot up, as user space may
6488 make the system inoperable.
6490 This command line option will stop the printing of events
6491 to console at the late_initcall_sync() time frame.
6493 trace_buf_size=nn[KMG]
6494 [FTRACE] will set tracing buffer size on each cpu.
6496 trace_clock= [FTRACE] Set the clock used for tracing events
6498 local - Use the per CPU time stamp counter
6499 (converted into nanoseconds). Fast, but
6500 depending on the architecture, may not be
6501 in sync between CPUs.
6502 global - Event time stamps are synchronize across
6503 CPUs. May be slower than the local clock,
6504 but better for some race conditions.
6505 counter - Simple counting of events (1, 2, ..)
6506 note, some counts may be skipped due to the
6507 infrastructure grabbing the clock more than
6509 uptime - Use jiffies as the time stamp.
6510 perf - Use the same clock that perf uses.
6511 mono - Use ktime_get_mono_fast_ns() for time stamps.
6512 mono_raw - Use ktime_get_raw_fast_ns() for time
6514 boot - Use ktime_get_boot_fast_ns() for time stamps.
6515 Architectures may add more clocks. See
6516 Documentation/trace/ftrace.rst for more details.
6518 trace_event=[event-list]
6519 [FTRACE] Set and start specified trace events in order
6520 to facilitate early boot debugging. The event-list is a
6521 comma-separated list of trace events to enable. See
6522 also Documentation/trace/events.rst
6524 trace_instance=[instance-info]
6525 [FTRACE] Create a ring buffer instance early in boot up.
6526 This will be listed in:
6528 /sys/kernel/tracing/instances
6530 Events can be enabled at the time the instance is created
6533 trace_instance=<name>,<system1>:<event1>,<system2>:<event2>
6535 Note, the "<system*>:" portion is optional if the event is
6538 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall
6540 will enable the "sched_switch" event (note, the "sched:" is optional, and
6541 the same thing would happen if it was left off). The irq_handler_entry
6542 event, and all events under the "initcall" system.
6544 trace_options=[option-list]
6545 [FTRACE] Enable or disable tracer options at boot.
6546 The option-list is a comma delimited list of options
6547 that can be enabled or disabled just as if you were
6548 to echo the option name into
6550 /sys/kernel/tracing/trace_options
6552 For example, to enable stacktrace option (to dump the
6553 stack trace of each event), add to the command line:
6555 trace_options=stacktrace
6557 See also Documentation/trace/ftrace.rst "trace options"
6560 trace_trigger=[trigger-list]
6561 [FTRACE] Add a event trigger on specific events.
6562 Set a trigger on top of a specific event, with an optional
6565 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..."
6566 Where more than one trigger may be specified that are comma deliminated.
6570 trace_trigger="sched_switch.stacktrace if prev_state == 2"
6572 The above will enable the "stacktrace" trigger on the "sched_switch"
6573 event but only trigger it if the "prev_state" of the "sched_switch"
6574 event is "2" (TASK_UNINTERUPTIBLE).
6576 See also "Event triggers" in Documentation/trace/events.rst
6580 [FTRACE] enable this option to disable tracing when a
6581 warning is hit. This turns off "tracing_on". Tracing can
6582 be enabled again by echoing '1' into the "tracing_on"
6583 file located in /sys/kernel/tracing/
6585 This option is useful, as it disables the trace before
6586 the WARNING dump is called, which prevents the trace to
6587 be filled with content caused by the warning output.
6589 This option can also be set at run time via the sysctl
6590 option: kernel/traceoff_on_warning
6592 transparent_hugepage=
6594 Format: [always|madvise|never]
6595 Can be used to control the default behavior of the system
6596 with respect to transparent hugepages.
6597 See Documentation/admin-guide/mm/transhuge.rst
6600 trusted.source= [KEYS]
6602 This parameter identifies the trust source as a backend
6603 for trusted keys implementation. Supported trust
6608 If not specified then it defaults to iterating through
6609 the trust source list starting with TPM and assigns the
6610 first trust source as a backend which is initialized
6611 successfully during iteration.
6615 The RNG used to generate key material for trusted keys.
6618 - the same value as trusted.source: "tpm" or "tee"
6620 If not specified, "default" is used. In this case,
6621 the RNG's choice is left to each individual trust source.
6623 tsc= Disable clocksource stability checks for TSC.
6625 [x86] reliable: mark tsc clocksource as reliable, this
6626 disables clocksource verification at runtime, as well
6627 as the stability checks done at bootup. Used to enable
6628 high-resolution timer mode on older hardware, and in
6629 virtualized environment.
6630 [x86] noirqtime: Do not use TSC to do irq accounting.
6631 Used to run time disable IRQ_TIME_ACCOUNTING on any
6632 platforms where RDTSC is slow and this accounting
6634 [x86] unstable: mark the TSC clocksource as unstable, this
6635 marks the TSC unconditionally unstable at bootup and
6636 avoids any further wobbles once the TSC watchdog notices.
6637 [x86] nowatchdog: disable clocksource watchdog. Used
6638 in situations with strict latency requirements (where
6639 interruptions from clocksource watchdog are not
6641 [x86] recalibrate: force recalibration against a HW timer
6642 (HPET or PM timer) on systems whose TSC frequency was
6643 obtained from HW or FW using either an MSR or CPUID(0x15).
6644 Warn if the difference is more than 500 ppm.
6645 [x86] watchdog: Use TSC as the watchdog clocksource with
6646 which to check other HW timers (HPET or PM timer), but
6647 only on systems where TSC has been deemed trustworthy.
6648 This will be suppressed by an earlier tsc=nowatchdog and
6649 can be overridden by a later tsc=nowatchdog. A console
6650 message will flag any such suppression or overriding.
6652 tsc_early_khz= [X86] Skip early TSC calibration and use the given
6653 value instead. Useful when the early TSC frequency discovery
6654 procedure is not reliable, such as on overclocked systems
6655 with CPUID.16h support and partial CPUID.15h support.
6656 Format: <unsigned int>
6658 tsx= [X86] Control Transactional Synchronization
6659 Extensions (TSX) feature in Intel processors that
6660 support TSX control.
6662 This parameter controls the TSX feature. The options are:
6664 on - Enable TSX on the system. Although there are
6665 mitigations for all known security vulnerabilities,
6666 TSX has been known to be an accelerator for
6667 several previous speculation-related CVEs, and
6668 so there may be unknown security risks associated
6669 with leaving it enabled.
6671 off - Disable TSX on the system. (Note that this
6672 option takes effect only on newer CPUs which are
6673 not vulnerable to MDS, i.e., have
6674 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6675 the new IA32_TSX_CTRL MSR through a microcode
6676 update. This new MSR allows for the reliable
6677 deactivation of the TSX functionality.)
6679 auto - Disable TSX if X86_BUG_TAA is present,
6680 otherwise enable TSX on the system.
6682 Not specifying this option is equivalent to tsx=off.
6684 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6687 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6688 Abort (TAA) vulnerability.
6690 Similar to Micro-architectural Data Sampling (MDS)
6691 certain CPUs that support Transactional
6692 Synchronization Extensions (TSX) are vulnerable to an
6693 exploit against CPU internal buffers which can forward
6694 information to a disclosure gadget under certain
6697 In vulnerable processors, the speculatively forwarded
6698 data can be used in a cache side channel attack, to
6699 access data to which the attacker does not have direct
6702 This parameter controls the TAA mitigation. The
6705 full - Enable TAA mitigation on vulnerable CPUs
6708 full,nosmt - Enable TAA mitigation and disable SMT on
6709 vulnerable CPUs. If TSX is disabled, SMT
6710 is not disabled because CPU is not
6711 vulnerable to cross-thread TAA attacks.
6712 off - Unconditionally disable TAA mitigation
6714 On MDS-affected machines, tsx_async_abort=off can be
6715 prevented by an active MDS mitigation as both vulnerabilities
6716 are mitigated with the same mechanism so in order to disable
6717 this mitigation, you need to specify mds=off too.
6719 Not specifying this option is equivalent to
6720 tsx_async_abort=full. On CPUs which are MDS affected
6721 and deploy MDS mitigation, TAA mitigation is not
6722 required and doesn't provide any additional
6726 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6728 turbografx.map[2|3]= [HW,JOY]
6729 TurboGraFX parallel port interface
6731 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6732 See also Documentation/input/devices/joystick-parport.rst
6734 udbg-immortal [PPC] When debugging early kernel crashes that
6735 happen after console_init() and before a proper
6736 console driver takes over, this boot options might
6737 help "seeing" what's going on.
6739 uhash_entries= [KNL,NET]
6740 Set number of hash buckets for UDP/UDP-Lite connections
6743 [USB] Ignore overcurrent events (default N).
6744 Some badly-designed motherboards generate lots of
6745 bogus events, for ports that aren't wired to
6746 anything. Set this parameter to avoid log spamming.
6747 Note that genuine overcurrent events won't be
6751 [X86] Cause panic on unknown NMI.
6753 unwind_debug [X86-64]
6754 Enable unwinder debug output. This can be
6755 useful for debugging certain unwinder error
6756 conditions, including corrupt stacks and
6757 bad/missing unwinder metadata.
6759 usbcore.authorized_default=
6760 [USB] Default USB device authorization:
6761 (default -1 = authorized (same as 1),
6762 0 = not authorized, 1 = authorized, 2 = authorized
6763 if device connected to internal port)
6765 usbcore.autosuspend=
6766 [USB] The autosuspend time delay (in seconds) used
6767 for newly-detected USB devices (default 2). This
6768 is the time required before an idle device will be
6769 autosuspended. Devices for which the delay is set
6770 to a negative value won't be autosuspended at all.
6772 usbcore.usbfs_snoop=
6773 [USB] Set to log all usbfs traffic (default 0 = off).
6775 usbcore.usbfs_snoop_max=
6776 [USB] Maximum number of bytes to snoop in each URB
6779 usbcore.blinkenlights=
6780 [USB] Set to cycle leds on hubs (default 0 = off).
6782 usbcore.old_scheme_first=
6783 [USB] Start with the old device initialization
6784 scheme (default 0 = off).
6786 usbcore.usbfs_memory_mb=
6787 [USB] Memory limit (in MB) for buffers allocated by
6788 usbfs (default = 16, 0 = max = 2047).
6790 usbcore.use_both_schemes=
6791 [USB] Try the other device initialization scheme
6792 if the first one fails (default 1 = enabled).
6794 usbcore.initial_descriptor_timeout=
6795 [USB] Specifies timeout for the initial 64-byte
6796 USB_REQ_GET_DESCRIPTOR request in milliseconds
6797 (default 5000 = 5.0 seconds).
6799 usbcore.nousb [USB] Disable the USB subsystem
6802 [USB] A list of quirk entries to augment the built-in
6803 usb core quirk list. List entries are separated by
6804 commas. Each entry has the form
6805 VendorID:ProductID:Flags. The IDs are 4-digit hex
6806 numbers and Flags is a set of letters. Each letter
6807 will change the built-in quirk; setting it if it is
6808 clear and clearing it if it is set. The letters have
6809 the following meanings:
6810 a = USB_QUIRK_STRING_FETCH_255 (string
6811 descriptors must not be fetched using
6813 b = USB_QUIRK_RESET_RESUME (device can't resume
6814 correctly so reset it instead);
6815 c = USB_QUIRK_NO_SET_INTF (device can't handle
6816 Set-Interface requests);
6817 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6818 handle its Configuration or Interface
6820 e = USB_QUIRK_RESET (device can't be reset
6821 (e.g morph devices), don't use reset);
6822 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6823 more interface descriptions than the
6824 bNumInterfaces count, and can't handle
6825 talking to these interfaces);
6826 g = USB_QUIRK_DELAY_INIT (device needs a pause
6827 during initialization, after we read
6828 the device descriptor);
6829 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6830 high speed and super speed interrupt
6831 endpoints, the USB 2.0 and USB 3.0 spec
6832 require the interval in microframes (1
6833 microframe = 125 microseconds) to be
6834 calculated as interval = 2 ^
6836 Devices with this quirk report their
6837 bInterval as the result of this
6838 calculation instead of the exponent
6839 variable used in the calculation);
6840 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6841 handle device_qualifier descriptor
6843 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6844 generates spurious wakeup, ignore
6845 remote wakeup capability);
6846 k = USB_QUIRK_NO_LPM (device can't handle Link
6848 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6849 (Device reports its bInterval as linear
6850 frames instead of the USB 2.0
6852 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6853 to be disconnected before suspend to
6854 prevent spurious wakeup);
6855 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6856 pause after every control message);
6857 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6858 delay after resetting its port);
6859 p = USB_QUIRK_SHORT_SET_ADDRESS_REQ_TIMEOUT
6860 (Reduce timeout of the SET_ADDRESS
6861 request from 5000 ms to 500 ms);
6862 Example: quirks=0781:5580:bk,0a5c:5834:gij
6865 [USBHID] The interval which mice are to be polled at.
6868 [USBHID] The interval which joysticks are to be polled at.
6871 [USBHID] The interval which keyboards are to be polled at.
6873 usb-storage.delay_use=
6874 [UMS] The delay in seconds before a new device is
6875 scanned for Logical Units (default 1).
6878 [UMS] A list of quirks entries to supplement or
6879 override the built-in unusual_devs list. List
6880 entries are separated by commas. Each entry has
6881 the form VID:PID:Flags where VID and PID are Vendor
6882 and Product ID values (4-digit hex numbers) and
6883 Flags is a set of characters, each corresponding
6884 to a common usb-storage quirk flag as follows:
6885 a = SANE_SENSE (collect more than 18 bytes
6886 of sense data, not on uas);
6887 b = BAD_SENSE (don't collect more than 18
6888 bytes of sense data, not on uas);
6889 c = FIX_CAPACITY (decrease the reported
6890 device capacity by one sector);
6891 d = NO_READ_DISC_INFO (don't use
6892 READ_DISC_INFO command, not on uas);
6893 e = NO_READ_CAPACITY_16 (don't use
6894 READ_CAPACITY_16 command);
6895 f = NO_REPORT_OPCODES (don't use report opcodes
6897 g = MAX_SECTORS_240 (don't transfer more than
6898 240 sectors at a time, uas only);
6899 h = CAPACITY_HEURISTICS (decrease the
6900 reported device capacity by one
6901 sector if the number is odd);
6902 i = IGNORE_DEVICE (don't bind to this
6904 j = NO_REPORT_LUNS (don't use report luns
6906 k = NO_SAME (do not use WRITE_SAME, uas only)
6907 l = NOT_LOCKABLE (don't try to lock and
6908 unlock ejectable media, not on uas);
6909 m = MAX_SECTORS_64 (don't transfer more
6910 than 64 sectors = 32 KB at a time,
6912 n = INITIAL_READ10 (force a retry of the
6913 initial READ(10) command, not on uas);
6914 o = CAPACITY_OK (accept the capacity
6915 reported by the device, not on uas);
6916 p = WRITE_CACHE (the device cache is ON
6917 by default, not on uas);
6918 r = IGNORE_RESIDUE (the device reports
6919 bogus residue values, not on uas);
6920 s = SINGLE_LUN (the device has only one
6922 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6923 commands, uas only);
6924 u = IGNORE_UAS (don't bind to the uas driver);
6925 w = NO_WP_DETECT (don't test whether the
6926 medium is write-protected).
6927 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6928 even if the device claims no cache,
6930 Example: quirks=0419:aaf5:rl,0421:0433:rc
6932 user_debug= [KNL,ARM]
6934 See arch/arm/Kconfig.debug help text.
6935 1 - undefined instruction events
6937 4 - invalid data aborts
6940 Example: user_debug=31
6943 [X86] Flags controlling user PTE allocations.
6945 nohigh = do not allocate PTE pages in
6946 HIGHMEM regardless of setting
6949 vdso= [X86,SH,SPARC]
6950 On X86_32, this is an alias for vdso32=. Otherwise:
6952 vdso=1: enable VDSO (the default)
6953 vdso=0: disable VDSO mapping
6955 vdso32= [X86] Control the 32-bit vDSO
6956 vdso32=1: enable 32-bit VDSO
6957 vdso32=0 or vdso32=2: disable 32-bit VDSO
6959 See the help text for CONFIG_COMPAT_VDSO for more
6960 details. If CONFIG_COMPAT_VDSO is set, the default is
6961 vdso32=0; otherwise, the default is vdso32=1.
6963 For compatibility with older kernels, vdso32=2 is an
6966 Try vdso32=0 if you encounter an error that says:
6967 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6970 vector=percpu: enable percpu vector domain
6972 video= [FB] Frame buffer configuration
6973 See Documentation/fb/modedb.rst.
6975 video.brightness_switch_enabled= [ACPI]
6977 If set to 1, on receiving an ACPI notify event
6978 generated by hotkey, video driver will adjust brightness
6979 level and then send out the event to user space through
6980 the allocated input device. If set to 0, video driver
6981 will only send out the event without touching backlight
6986 [VMMIO] Memory mapped virtio (platform) device.
6988 <size>@<baseaddr>:<irq>[:<id>]
6990 <size> := size (can use standard suffixes
6992 <baseaddr> := physical base address
6993 <irq> := interrupt number (as passed to
6995 <id> := (optional) platform device id
6997 virtio_mmio.device=1K@0x100b0000:48:7
6999 Can be used multiple times for multiple devices.
7001 vga= [BOOT,X86-32] Select a particular video mode
7002 See Documentation/arch/x86/boot.rst and
7003 Documentation/admin-guide/svga.rst.
7004 Use vga=ask for menu.
7005 This is actually a boot loader parameter; the value is
7006 passed to the kernel using a special protocol.
7008 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
7009 May slow down system boot speed, especially when
7010 enabled on systems with a large amount of memory.
7011 All options are enabled by default, and this
7012 interface is meant to allow for selectively
7013 enabling or disabling specific virtual memory
7016 Available options are:
7017 P Enable page structure init time poisoning
7018 - Disable all of the above options
7020 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
7021 size of <nn>. This can be used to increase the
7022 minimum size (128MB on x86). It can also be used to
7023 decrease the size and leave more room for directly
7026 vmcp_cma=nn[MG] [KNL,S390]
7027 Sets the memory size reserved for contiguous memory
7028 allocations for the vmcp device driver.
7030 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
7033 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
7036 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
7040 Controls the behavior of vsyscalls (i.e. calls to
7041 fixed addresses of 0xffffffffff600x00 from legacy
7042 code). Most statically-linked binaries and older
7043 versions of glibc use these calls. Because these
7044 functions are at fixed addresses, they make nice
7045 targets for exploits that can control RIP.
7047 emulate Vsyscalls turn into traps and are emulated
7048 reasonably safely. The vsyscall page is
7051 xonly [default] Vsyscalls turn into traps and are
7052 emulated reasonably safely. The vsyscall
7053 page is not readable.
7055 none Vsyscalls don't work at all. This makes
7056 them quite hard to use for exploits but
7057 might break your system.
7059 vt.color= [VT] Default text color.
7060 Format: 0xYX, X = foreground, Y = background.
7061 Default: 0x07 = light gray on black.
7063 vt.cur_default= [VT] Default cursor shape.
7064 Format: 0xCCBBAA, where AA, BB, and CC are the same as
7065 the parameters of the <Esc>[?A;B;Cc escape sequence;
7066 see VGA-softcursor.txt. Default: 2 = underline.
7068 vt.default_blu= [VT]
7069 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
7070 Change the default blue palette of the console.
7071 This is a 16-member array composed of values
7074 vt.default_grn= [VT]
7075 Format: <green0>,<green1>,<green2>,...,<green15>
7076 Change the default green palette of the console.
7077 This is a 16-member array composed of values
7080 vt.default_red= [VT]
7081 Format: <red0>,<red1>,<red2>,...,<red15>
7082 Change the default red palette of the console.
7083 This is a 16-member array composed of values
7089 Set system-wide default UTF-8 mode for all tty's.
7090 Default is 1, i.e. UTF-8 mode is enabled for all
7091 newly opened terminals.
7093 vt.global_cursor_default=
7096 Set system-wide default for whether a cursor
7097 is shown on new VTs. Default is -1,
7098 i.e. cursors will be created by default unless
7099 overridden by individual drivers. 0 will hide
7100 cursors, 1 will display them.
7102 vt.italic= [VT] Default color for italic text; 0-15.
7105 vt.underline= [VT] Default color for underlined text; 0-15.
7108 watchdog timers [HW,WDT] For information on watchdog timers,
7109 see Documentation/watchdog/watchdog-parameters.rst
7110 or other driver-specific files in the
7111 Documentation/watchdog/ directory.
7115 Set the hard lockup detector stall duration
7116 threshold in seconds. The soft lockup detector
7117 threshold is set to twice the value. A value of 0
7118 disables both lockup detectors. Default is 10
7121 workqueue.unbound_cpus=
7122 [KNL,SMP] Specify to constrain one or some CPUs
7123 to use in unbound workqueues.
7125 By default, all online CPUs are available for
7128 workqueue.watchdog_thresh=
7129 If CONFIG_WQ_WATCHDOG is configured, workqueue can
7130 warn stall conditions and dump internal state to
7131 help debugging. 0 disables workqueue stall
7132 detection; otherwise, it's the stall threshold
7133 duration in seconds. The default value is 30 and
7134 it can be updated at runtime by writing to the
7135 corresponding sysfs file.
7137 workqueue.cpu_intensive_thresh_us=
7138 Per-cpu work items which run for longer than this
7139 threshold are automatically considered CPU intensive
7140 and excluded from concurrency management to prevent
7141 them from noticeably delaying other per-cpu work
7142 items. Default is 10000 (10ms).
7144 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel
7145 will report the work functions which violate this
7146 threshold repeatedly. They are likely good
7147 candidates for using WQ_UNBOUND workqueues instead.
7149 workqueue.power_efficient
7150 Per-cpu workqueues are generally preferred because
7151 they show better performance thanks to cache
7152 locality; unfortunately, per-cpu workqueues tend to
7153 be more power hungry than unbound workqueues.
7155 Enabling this makes the per-cpu workqueues which
7156 were observed to contribute significantly to power
7157 consumption unbound, leading to measurably lower
7158 power usage at the cost of small performance
7161 The default value of this parameter is determined by
7162 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
7164 workqueue.default_affinity_scope=
7165 Select the default affinity scope to use for unbound
7166 workqueues. Can be one of "cpu", "smt", "cache",
7167 "numa" and "system". Default is "cache". For more
7168 information, see the Affinity Scopes section in
7169 Documentation/core-api/workqueue.rst.
7171 This can be changed after boot by writing to the
7172 matching /sys/module/workqueue/parameters file. All
7173 workqueues with the "default" affinity scope will be
7174 updated accordignly.
7176 workqueue.debug_force_rr_cpu
7177 Workqueue used to implicitly guarantee that work
7178 items queued without explicit CPU specified are put
7179 on the local CPU. This guarantee is no longer true
7180 and while local CPU is still preferred work items
7181 may be put on foreign CPUs. This debug option
7182 forces round-robin CPU selection to flush out
7183 usages which depend on the now broken guarantee.
7184 When enabled, memory and cache locality will be
7187 writecombine= [LOONGARCH] Control the MAT (Memory Access Type) of
7190 on - Enable writecombine, use WUC for ioremap_wc()
7191 off - Disable writecombine, use SUC for ioremap_wc()
7193 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
7194 default x2apic cluster mode on platforms
7197 xen_512gb_limit [KNL,X86-64,XEN]
7198 Restricts the kernel running paravirtualized under Xen
7199 to use only up to 512 GB of RAM. The reason to do so is
7200 crash analysis tools and Xen tools for doing domain
7201 save/restore/migration must be enabled to handle larger
7204 xen_emul_unplug= [HW,X86,XEN]
7205 Unplug Xen emulated devices
7206 Format: [unplug0,][unplug1]
7207 ide-disks -- unplug primary master IDE devices
7208 aux-ide-disks -- unplug non-primary-master IDE devices
7209 nics -- unplug network devices
7210 all -- unplug all emulated devices (NICs and IDE disks)
7211 unnecessary -- unplugging emulated devices is
7212 unnecessary even if the host did not respond to
7214 never -- do not unplug even if version check succeeds
7216 xen_legacy_crash [X86,XEN]
7217 Crash from Xen panic notifier, without executing late
7218 panic() code such as dumping handler.
7220 xen_msr_safe= [X86,XEN]
7222 Select whether to always use non-faulting (safe) MSR
7223 access functions when running as Xen PV guest. The
7224 default value is controlled by CONFIG_XEN_PV_MSR_SAFE.
7226 xen_nopvspin [X86,XEN]
7227 Disables the qspinlock slowpath using Xen PV optimizations.
7228 This parameter is obsoleted by "nopvspin" parameter, which
7229 has equivalent effect for XEN platform.
7232 Disables the PV optimizations forcing the HVM guest to
7233 run as generic HVM guest with no PV drivers.
7234 This option is obsoleted by the "nopv" option, which
7235 has equivalent effect for XEN platform.
7237 xen_no_vector_callback
7238 [KNL,X86,XEN] Disable the vector callback for Xen
7239 event channel interrupts.
7241 xen_scrub_pages= [XEN]
7242 Boolean option to control scrubbing pages before giving them back
7243 to Xen, for use by other domains. Can be also changed at runtime
7244 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
7245 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
7247 xen_timer_slop= [X86-64,XEN]
7248 Set the timer slop (in nanoseconds) for the virtual Xen
7249 timers (default is 100000). This adjusts the minimum
7250 delta of virtualized Xen timers, where lower values
7251 improve timer resolution at the expense of processing
7252 more timer interrupts.
7254 xen.balloon_boot_timeout= [XEN]
7255 The time (in seconds) to wait before giving up to boot
7256 in case initial ballooning fails to free enough memory.
7257 Applies only when running as HVM or PVH guest and
7258 started with less memory configured than allowed at
7259 max. Default is 180.
7261 xen.event_eoi_delay= [XEN]
7262 How long to delay EOI handling in case of event
7263 storms (jiffies). Default is 10.
7265 xen.event_loop_timeout= [XEN]
7266 After which time (jiffies) the event handling loop
7267 should start to delay EOI handling. Default is 2.
7269 xen.fifo_events= [XEN]
7270 Boolean parameter to disable using fifo event handling
7271 even if available. Normally fifo event handling is
7272 preferred over the 2-level event handling, as it is
7273 fairer and the number of possible event channels is
7274 much higher. Default is on (use fifo events).
7276 xirc2ps_cs= [NET,PCMCIA]
7278 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
7281 By default on POWER9 and above, the kernel will
7282 natively use the XIVE interrupt controller. This option
7283 allows the fallback firmware mode to be used:
7285 off Fallback to firmware control of XIVE interrupt
7286 controller on both pseries and powernv
7287 platforms. Only useful on POWER9 and above.
7289 xive.store-eoi=off [PPC]
7290 By default on POWER10 and above, the kernel will use
7291 stores for EOI handling when the XIVE interrupt mode
7292 is active. This option allows the XIVE driver to use
7293 loads instead, as on POWER9.
7295 xhci-hcd.quirks [USB,KNL]
7296 A hex value specifying bitmask with supplemental xhci
7297 host controller quirks. Meaning of each bit can be
7298 consulted in header drivers/usb/host/xhci.h.
7301 Format: { early | on | rw | ro | off }
7302 Controls if xmon debugger is enabled. Default is off.
7303 Passing only "xmon" is equivalent to "xmon=early".
7304 early Call xmon as early as possible on boot; xmon
7305 debugger is called from setup_arch().
7306 on xmon debugger hooks will be installed so xmon
7307 is only called on a kernel crash. Default mode,
7308 i.e. either "ro" or "rw" mode, is controlled
7309 with CONFIG_XMON_DEFAULT_RO_MODE.
7310 rw xmon debugger hooks will be installed so xmon
7311 is called only on a kernel crash, mode is write,
7312 meaning SPR registers, memory and, other data
7313 can be written using xmon commands.
7314 ro same as "rw" option above but SPR registers,
7315 memory, and other data can't be written using
7317 off xmon is disabled.