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] 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] range could be above 4G. Allow kernel
890 to allocate physical memory region from top, so could
891 be above 4G if system have more than 4G ram installed.
892 Otherwise memory region will be allocated below 4G, if
894 It will be ignored if crashkernel=X is specified.
895 crashkernel=size[KMG],low
896 [KNL, X86-64, ARM64] 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)
907 This one lets the user specify own low range under 4G
908 for second kernel instead.
909 0: to disable low allocation.
910 It will be ignored when crashkernel=X,high is not used
911 or memory reserved is below 4G.
914 [KNL] Disable crypto self-tests
919 cs89x0_media= [HW,NET]
920 Format: { rj45 | aui | bnc }
922 csdlock_debug= [KNL] Enable or disable debug add-ons of cross-CPU
923 function call handling. When switched on,
924 additional debug data is printed to the console
925 in case a hanging CPU is detected, and that
926 CPU is pinged again in order to try to resolve
927 the hang situation. The default value of this
928 option depends on the CSD_LOCK_WAIT_DEBUG_DEFAULT
932 See header of drivers/s390/block/dasd_devmap.c.
934 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
935 (one device per port)
936 Format: <port#>,<type>
937 See also Documentation/input/devices/joystick-parport.rst
939 debug [KNL] Enable kernel debugging (events log level).
942 [KNL] Enable printing [hashed] pointers early in the
943 boot sequence. If enabled, we use a weak hash instead
944 of siphash to hash pointers. Use this option if you are
945 seeing instances of '(___ptrval___)') and need to see a
946 value (hashed pointer) instead. Cryptographically
947 insecure, please do not use on production kernels.
950 [KNL] verbose locking self-tests
952 Print debugging info while doing the locking API
954 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
955 (no extra messages), setting it to -1 (all bits set)
956 will print _a_lot_ more information - normally only
957 useful to lockdep developers.
959 debug_objects [KNL] Enable object debugging
961 debug_guardpage_minorder=
962 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
963 parameter allows control of the order of pages that will
964 be intentionally kept free (and hence protected) by the
965 buddy allocator. Bigger value increase the probability
966 of catching random memory corruption, but reduce the
967 amount of memory for normal system use. The maximum
968 possible value is MAX_ORDER/2. Setting this parameter
969 to 1 or 2 should be enough to identify most random
970 memory corruption problems caused by bugs in kernel or
971 driver code when a CPU writes to (or reads from) a
972 random memory location. Note that there exists a class
973 of memory corruptions problems caused by buggy H/W or
974 F/W or by drivers badly programming DMA (basically when
975 memory is written at bus level and the CPU MMU is
976 bypassed) which are not detectable by
977 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
978 tracking down these problems.
981 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
982 enables the feature at boot time. By default, it is
983 disabled and the system will work mostly the same as a
984 kernel built without CONFIG_DEBUG_PAGEALLOC.
985 Note: to get most of debug_pagealloc error reports, it's
986 useful to also enable the page_owner functionality.
987 on: enable the feature
989 debugfs= [KNL] This parameter enables what is exposed to userspace
990 and debugfs internal clients.
991 Format: { on, no-mount, off }
992 on: All functions are enabled.
994 Filesystem is not registered but kernel clients can
995 access APIs and a crashkernel can be used to read
996 its content. There is nothing to mount.
997 off: Filesystem is not registered and clients
998 get a -EPERM as result when trying to register files
999 or directories within debugfs.
1000 This is equivalent of the runtime functionality if
1001 debugfs was not enabled in the kernel at all.
1002 Default value is set in build-time with a kernel configuration.
1004 debugpat [X86] Enable PAT debugging
1007 [HW] The size of the default HugeTLB page. This is
1008 the size represented by the legacy /proc/ hugepages
1009 APIs. In addition, this is the default hugetlb size
1010 used for shmget(), mmap() and mounting hugetlbfs
1011 filesystems. If not specified, defaults to the
1012 architecture's default huge page size. Huge page
1013 sizes are architecture dependent. See also
1014 Documentation/admin-guide/mm/hugetlbpage.rst.
1017 deferred_probe_timeout=
1018 [KNL] Debugging option to set a timeout in seconds for
1019 deferred probe to give up waiting on dependencies to
1020 probe. Only specific dependencies (subsystems or
1021 drivers) that have opted in will be ignored. A timeout
1022 of 0 will timeout at the end of initcalls. If the time
1023 out hasn't expired, it'll be restarted by each
1024 successful driver registration. This option will also
1025 dump out devices still on the deferred probe list after
1028 delayacct [KNL] Enable per-task delay accounting
1030 dell_smm_hwmon.ignore_dmi=
1031 [HW] Continue probing hardware even if DMI data
1032 indicates that the driver is running on unsupported
1035 dell_smm_hwmon.force=
1036 [HW] Activate driver even if SMM BIOS signature does
1037 not match list of supported models and enable otherwise
1038 blacklisted features.
1040 dell_smm_hwmon.power_status=
1041 [HW] Report power status in /proc/i8k
1042 (disabled by default).
1044 dell_smm_hwmon.restricted=
1045 [HW] Allow controlling fans only if SYS_ADMIN
1048 dell_smm_hwmon.fan_mult=
1049 [HW] Factor to multiply fan speed with.
1051 dell_smm_hwmon.fan_max=
1052 [HW] Maximum configurable fan speed.
1055 Format: { on | off | def_only | inf_only | always }
1056 on: s390 zlib hardware support for compression on
1057 level 1 and decompression (default)
1058 off: No s390 zlib hardware support
1059 def_only: s390 zlib hardware support for deflate
1060 only (compression on level 1)
1061 inf_only: s390 zlib hardware support for inflate
1062 only (decompression)
1063 always: Same as 'on' but ignores the selected compression
1064 level always using hardware support (used for debugging)
1066 dhash_entries= [KNL]
1067 Set number of hash buckets for dentry cache.
1069 disable_1tb_segments [PPC]
1070 Disables the use of 1TB hash page table segments. This
1071 causes the kernel to fall back to 256MB segments which
1072 can be useful when debugging issues that require an SLB
1076 See Documentation/networking/ipv6.rst.
1079 Disable RADIX MMU mode on POWER9
1082 Disable TLBIE instruction. Currently does not work
1083 with KVM, with HASH MMU, or with coherent accelerators.
1085 disable_cpu_apicid= [X86,APIC,SMP]
1087 The number of initial APIC ID for the
1088 corresponding CPU to be disabled at boot,
1089 mostly used for the kdump 2nd kernel to
1090 disable BSP to wake up multiple CPUs without
1091 causing system reset or hang due to sending
1092 INIT from AP to BSP.
1094 disable_ddw [PPC/PSERIES]
1095 Disable Dynamic DMA Window support. Use this
1096 to workaround buggy firmware.
1098 disable_ipv6= [IPV6]
1099 See Documentation/networking/ipv6.rst.
1101 disable_mtrr_cleanup [X86]
1102 The kernel tries to adjust MTRR layout from continuous
1103 to discrete, to make X server driver able to add WB
1104 entry later. This parameter disables that.
1106 disable_mtrr_trim [X86, Intel and AMD only]
1107 By default the kernel will trim any uncacheable
1108 memory out of your available memory pool based on
1109 MTRR settings. This parameter disables that behavior,
1110 possibly causing your machine to run very slowly.
1112 disable_timer_pin_1 [X86]
1113 Disable PIN 1 of APIC timer
1114 Can be useful to work around chipset bugs.
1116 dis_ucode_ldr [X86] Disable the microcode loader.
1118 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
1119 this option disables the debugging code at boot.
1121 dma_debug_entries=<number>
1122 This option allows to tune the number of preallocated
1123 entries for DMA-API debugging code. One entry is
1124 required per DMA-API allocation. Use this if the
1125 DMA-API debugging code disables itself because the
1126 architectural default is too low.
1128 dma_debug_driver=<driver_name>
1129 With this option the DMA-API debugging driver
1130 filter feature can be enabled at boot time. Just
1131 pass the driver to filter for as the parameter.
1132 The filter can be disabled or changed to another
1133 driver later using sysfs.
1135 driver_async_probe= [KNL]
1136 List of driver names to be probed asynchronously. *
1137 matches with all driver names. If * is specified, the
1138 rest of the listed driver names are those that will NOT
1140 Format: <driver_name1>,<driver_name2>...
1142 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1143 Broken monitors, graphic adapters, KVMs and EDIDless
1144 panels may send no or incorrect EDID data sets.
1145 This parameter allows to specify an EDID data sets
1146 in the /lib/firmware directory that are used instead.
1147 Generic built-in EDID data sets are used, if one of
1148 edid/1024x768.bin, edid/1280x1024.bin,
1149 edid/1680x1050.bin, or edid/1920x1080.bin is given
1150 and no file with the same name exists. Details and
1151 instructions how to build your own EDID data are
1152 available in Documentation/admin-guide/edid.rst. An EDID
1153 data set will only be used for a particular connector,
1154 if its name and a colon are prepended to the EDID
1155 name. Each connector may use a unique EDID data
1156 set by separating the files with a comma. An EDID
1157 data set with no connector name will be used for
1158 any connectors not explicitly specified.
1163 Format: {"off" | "known"}
1164 Control how the dt_cpu_ftrs device-tree binding is
1165 used for CPU feature discovery and setup (if it
1167 off: Do not use it, fall back to legacy cpu table.
1168 known: Do not pass through unknown features to guests
1169 or userspace, only those that the kernel is aware of.
1171 dump_apple_properties [X86]
1172 Dump name and content of EFI device properties on
1173 x86 Macs. Useful for driver authors to determine
1174 what data is available or for reverse-engineering.
1176 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1177 <module>.dyndbg[="val"]
1178 Enable debug messages at boot time. See
1179 Documentation/admin-guide/dynamic-debug-howto.rst
1182 early_ioremap_debug [KNL]
1183 Enable debug messages in early_ioremap support. This
1184 is useful for tracking down temporary early mappings
1185 which are not unmapped.
1187 earlycon= [KNL] Output early console device and options.
1189 When used with no options, the early console is
1190 determined by stdout-path property in device tree's
1191 chosen node or the ACPI SPCR table if supported by
1194 cdns,<addr>[,options]
1195 Start an early, polled-mode console on a Cadence
1196 (xuartps) serial port at the specified address. Only
1197 supported option is baud rate. If baud rate is not
1198 specified, the serial port must already be setup and
1201 uart[8250],io,<addr>[,options[,uartclk]]
1202 uart[8250],mmio,<addr>[,options[,uartclk]]
1203 uart[8250],mmio32,<addr>[,options[,uartclk]]
1204 uart[8250],mmio32be,<addr>[,options[,uartclk]]
1205 uart[8250],0x<addr>[,options]
1206 Start an early, polled-mode console on the 8250/16550
1207 UART at the specified I/O port or MMIO address.
1208 MMIO inter-register address stride is either 8-bit
1209 (mmio) or 32-bit (mmio32 or mmio32be).
1210 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1211 to be equivalent to 'mmio'. 'options' are specified
1212 in the same format described for "console=ttyS<n>"; if
1213 unspecified, the h/w is not initialized. 'uartclk' is
1214 the uart clock frequency; if unspecified, it is set
1215 to 'BASE_BAUD' * 16.
1219 Start an early, polled-mode console on a pl011 serial
1220 port at the specified address. The pl011 serial port
1221 must already be setup and configured. Options are not
1222 yet supported. If 'mmio32' is specified, then only
1223 the driver will use only 32-bit accessors to read/write
1224 the device registers.
1227 Start an early console on a litex serial port at the
1228 specified address. The serial port must already be
1229 setup and configured. Options are not yet supported.
1232 Start an early, polled-mode console on a meson serial
1233 port at the specified address. The serial port must
1234 already be setup and configured. Options are not yet
1238 Start an early, polled-mode console on an msm serial
1239 port at the specified address. The serial port
1240 must already be setup and configured. Options are not
1243 msm_serial_dm,<addr>
1244 Start an early, polled-mode console on an msm serial
1245 dm port at the specified address. The serial port
1246 must already be setup and configured. Options are not
1250 Start an early, polled-mode console on a serial port
1251 of an Actions Semi SoC, such as S500 or S900, at the
1252 specified address. The serial port must already be
1253 setup and configured. Options are not yet supported.
1256 Start an early, polled-mode console on a serial port
1257 of an RDA Micro SoC, such as RDA8810PL, at the
1258 specified address. The serial port must already be
1259 setup and configured. Options are not yet supported.
1262 Use RISC-V SBI (Supervisor Binary Interface) for early
1265 smh Use ARM semihosting calls for early console.
1273 Use early console provided by serial driver available
1274 on Samsung SoCs, requires selecting proper type and
1275 a correct base address of the selected UART port. The
1276 serial port must already be setup and configured.
1277 Options are not yet supported.
1280 Start an early, polled-mode console on a lantiq serial
1281 (lqasc) port at the specified address. The serial port
1282 must already be setup and configured. Options are not
1287 Use early console provided by Freescale LP UART driver
1288 found on Freescale Vybrid and QorIQ LS1021A processors.
1289 A valid base address must be provided, and the serial
1290 port must already be setup and configured.
1294 Start an early, polled-mode, output-only console on the
1295 Freescale i.MX UART at the specified address. The UART
1296 must already be setup and configured.
1299 Start an early, polled-mode console on the
1300 Armada 3700 serial port at the specified
1301 address. The serial port must already be setup
1302 and configured. Options are not yet supported.
1305 Start an early, polled-mode console on a Qualcomm
1306 Generic Interface (GENI) based serial port at the
1307 specified address. The serial port must already be
1308 setup and configured. Options are not yet supported.
1311 Start an early, unaccelerated console on the EFI
1312 memory mapped framebuffer (if available). On cache
1313 coherent non-x86 systems that use system memory for
1314 the framebuffer, pass the 'ram' option so that it is
1315 mapped with the correct attributes.
1318 Use early console provided by Freescale LINFlexD UART
1319 serial driver for NXP S32V234 SoCs. A valid base
1320 address must be provided, and the serial port must
1321 already be setup and configured.
1323 earlyprintk= [X86,SH,ARM,M68k,S390]
1327 earlyprintk=serial[,ttySn[,baudrate]]
1328 earlyprintk=serial[,0x...[,baudrate]]
1329 earlyprintk=ttySn[,baudrate]
1330 earlyprintk=dbgp[debugController#]
1331 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1332 earlyprintk=xdbc[xhciController#]
1334 earlyprintk is useful when the kernel crashes before
1335 the normal console is initialized. It is not enabled by
1336 default because it has some cosmetic problems.
1338 Append ",keep" to not disable it when the real console
1341 Only one of vga, serial, or usb debug port can
1344 Currently only ttyS0 and ttyS1 may be specified by
1345 name. Other I/O ports may be explicitly specified
1346 on some architectures (x86 and arm at least) by
1347 replacing ttySn with an I/O port address, like this:
1348 earlyprintk=serial,0x1008,115200
1349 You can find the port for a given device in
1350 /proc/tty/driver/serial:
1351 2: uart:ST16650V2 port:00001008 irq:18 ...
1353 Interaction with the standard serial driver is not
1356 The VGA output is eventually overwritten by
1359 The xen option can only be used in Xen domains.
1361 The sclp output can only be used on s390.
1363 The optional "force" to "pciserial" enables use of a
1364 PCI device even when its classcode is not of the
1367 edac_report= [HW,EDAC] Control how to report EDAC event
1368 Format: {"on" | "off" | "force"}
1369 on: enable EDAC to report H/W event. May be overridden
1370 by other higher priority error reporting module.
1371 off: disable H/W event reporting through EDAC.
1372 force: enforce the use of EDAC to report H/W event.
1376 Format: {"off" | "on" | "skip[mbr]"}
1379 Format: { "debug", "disable_early_pci_dma",
1380 "nochunk", "noruntime", "nosoftreserve",
1381 "novamap", "no_disable_early_pci_dma" }
1382 debug: enable misc debug output.
1383 disable_early_pci_dma: disable the busmaster bit on all
1384 PCI bridges while in the EFI boot stub.
1385 nochunk: disable reading files in "chunks" in the EFI
1386 boot stub, as chunking can cause problems with some
1387 firmware implementations.
1388 noruntime : disable EFI runtime services support
1389 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1390 attribute may cause the kernel to reserve the
1391 memory range for a memory mapping driver to
1392 claim. Specify efi=nosoftreserve to disable this
1393 reservation and treat the memory by its base type
1394 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1395 novamap: do not call SetVirtualAddressMap().
1396 no_disable_early_pci_dma: Leave the busmaster bit set
1397 on all PCI bridges while in the EFI boot stub
1399 efi_no_storage_paranoia [EFI; X86]
1400 Using this parameter you can use more than 50% of
1401 your efi variable storage. Use this parameter only if
1402 you are really sure that your UEFI does sane gc and
1403 fulfills the spec otherwise your board may brick.
1405 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1406 Add arbitrary attribute to specific memory range by
1407 updating original EFI memory map.
1408 Region of memory which aa attribute is added to is
1411 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1412 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1413 attribute is added to range 0x100000000-0x180000000 and
1414 0x10a0000000-0x1120000000.
1416 If efi_fake_mem=8G@9G:0x40000 is specified, the
1417 EFI_MEMORY_SP(0x40000) attribute is added to
1418 range 0x240000000-0x43fffffff.
1420 Using this parameter you can do debugging of EFI memmap
1421 related features. For example, you can do debugging of
1422 Address Range Mirroring feature even if your box
1423 doesn't support it, or mark specific memory as
1426 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1427 that is to be dynamically loaded by Linux. If there are
1428 multiple variables with the same name but with different
1429 vendor GUIDs, all of them will be loaded. See
1430 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1433 eisa_irq_edge= [PARISC,HW]
1434 See header of drivers/parisc/eisa.c.
1436 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1439 This is designed to be used in conjunction with
1440 the boot argument: earlyprintk=vga
1442 This parameter works in place of the kgdboc parameter
1443 but can only be used if the backing tty is available
1444 very early in the boot process. For early debugging
1445 via a serial port see kgdboc_earlycon instead.
1448 See comment before function elanfreq_setup() in
1449 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1451 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1452 Specifies physical address of start of kernel core
1453 image elf header and optionally the size. Generally
1454 kexec loader will pass this option to capture kernel.
1455 See Documentation/admin-guide/kdump/kdump.rst for details.
1457 enable_mtrr_cleanup [X86]
1458 The kernel tries to adjust MTRR layout from continuous
1459 to discrete, to make X server driver able to add WB
1460 entry later. This parameter enables that.
1462 enable_timer_pin_1 [X86]
1463 Enable PIN 1 of APIC timer
1464 Can be useful to work around chipset bugs
1465 (in particular on some ATI chipsets).
1466 The kernel tries to set a reasonable default.
1468 enforcing= [SELINUX] Set initial enforcing status.
1470 See security/selinux/Kconfig help text.
1471 0 -- permissive (log only, no denials).
1472 1 -- enforcing (deny and log).
1474 Value can be changed at runtime via
1475 /sys/fs/selinux/enforce.
1478 Disable Error Record Serialization Table (ERST)
1481 ether= [HW,NET] Ethernet cards parameters
1482 This option is obsoleted by the "netdev=" option, which
1483 has equivalent usage. See its documentation for details.
1487 Permit 'security.evm' to be updated regardless of
1488 current integrity status.
1490 early_page_ext [KNL] Enforces page_ext initialization to earlier
1491 stages so cover more early boot allocations.
1492 Please note that as side effect some optimizations
1493 might be disabled to achieve that (e.g. parallelized
1494 memory initialization is disabled) so the boot process
1495 might take longer, especially on systems with a lot of
1496 memory. Available with CONFIG_PAGE_EXTENSION=y.
1501 fail_make_request=[KNL]
1502 General fault injection mechanism.
1503 Format: <interval>,<probability>,<space>,<times>
1504 See also Documentation/fault-injection/.
1507 Format: { initns | none }
1508 See Documentation/admin-guide/sysctl/net.rst for
1509 fb_tunnels_only_for_init_ns
1512 See Documentation/admin-guide/blockdev/floppy.rst.
1514 force_pal_cache_flush
1515 [IA-64] Avoid check_sal_cache_flush which may hang on
1516 buggy SAL_CACHE_FLUSH implementations. Using this
1517 parameter will force ia64_sal_cache_flush to call
1518 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1521 Forcefully enable Physical Address Extension (PAE).
1522 Many Pentium M systems disable PAE but may have a
1523 functionally usable PAE implementation.
1524 Warning: use of this parameter will taint the kernel
1525 and may cause unknown problems.
1528 [FTRACE] will set and start the specified tracer
1529 as early as possible in order to facilitate early
1532 ftrace_boot_snapshot
1533 [FTRACE] On boot up, a snapshot will be taken of the
1534 ftrace ring buffer that can be read at:
1535 /sys/kernel/tracing/snapshot.
1536 This is useful if you need tracing information from kernel
1537 boot up that is likely to be overridden by user space
1538 start up functionality.
1540 Optionally, the snapshot can also be defined for a tracing
1541 instance that was created by the trace_instance= command
1544 trace_instance=foo,sched_switch ftrace_boot_snapshot=foo
1546 The above will cause the "foo" tracing instance to trigger
1547 a snapshot at the end of boot up.
1549 ftrace_dump_on_oops[=orig_cpu]
1550 [FTRACE] will dump the trace buffers on oops.
1551 If no parameter is passed, ftrace will dump
1552 buffers of all CPUs, but if you pass orig_cpu, it will
1553 dump only the buffer of the CPU that triggered the
1556 ftrace_filter=[function-list]
1557 [FTRACE] Limit the functions traced by the function
1558 tracer at boot up. function-list is a comma-separated
1559 list of functions. This list can be changed at run
1560 time by the set_ftrace_filter file in the debugfs
1563 ftrace_notrace=[function-list]
1564 [FTRACE] Do not trace the functions specified in
1565 function-list. This list can be changed at run time
1566 by the set_ftrace_notrace file in the debugfs
1569 ftrace_graph_filter=[function-list]
1570 [FTRACE] Limit the top level callers functions traced
1571 by the function graph tracer at boot up.
1572 function-list is a comma-separated list of functions
1573 that can be changed at run time by the
1574 set_graph_function file in the debugfs tracing directory.
1576 ftrace_graph_notrace=[function-list]
1577 [FTRACE] Do not trace from the functions specified in
1578 function-list. This list is a comma-separated list of
1579 functions that can be changed at run time by the
1580 set_graph_notrace file in the debugfs tracing directory.
1582 ftrace_graph_max_depth=<uint>
1583 [FTRACE] Used with the function graph tracer. This is
1584 the max depth it will trace into a function. This value
1585 can be changed at run time by the max_graph_depth file
1586 in the tracefs tracing directory. default: 0 (no limit)
1588 fw_devlink= [KNL] Create device links between consumer and supplier
1589 devices by scanning the firmware to infer the
1590 consumer/supplier relationships. This feature is
1591 especially useful when drivers are loaded as modules as
1592 it ensures proper ordering of tasks like device probing
1593 (suppliers first, then consumers), supplier boot state
1594 clean up (only after all consumers have probed),
1595 suspend/resume & runtime PM (consumers first, then
1597 Format: { off | permissive | on | rpm }
1598 off -- Don't create device links from firmware info.
1599 permissive -- Create device links from firmware info
1600 but use it only for ordering boot state clean
1601 up (sync_state() calls).
1602 on -- Create device links from firmware info and use it
1603 to enforce probe and suspend/resume ordering.
1604 rpm -- Like "on", but also use to order runtime PM.
1606 fw_devlink.strict=<bool>
1607 [KNL] Treat all inferred dependencies as mandatory
1608 dependencies. This only applies for fw_devlink=on|rpm.
1611 fw_devlink.sync_state =
1612 [KNL] When all devices that could probe have finished
1613 probing, this parameter controls what to do with
1614 devices that haven't yet received their sync_state()
1616 Format: { strict | timeout }
1617 strict -- Default. Continue waiting on consumers to
1619 timeout -- Give up waiting on consumers and call
1620 sync_state() on any devices that haven't yet
1621 received their sync_state() calls after
1622 deferred_probe_timeout has expired or by
1623 late_initcall() if !CONFIG_MODULES.
1626 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1627 support via parallel port (up to 5 devices per port)
1628 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1629 See also Documentation/input/devices/joystick-parport.rst
1633 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1637 gather_data_sampling=
1638 [X86,INTEL] Control the Gather Data Sampling (GDS)
1641 Gather Data Sampling is a hardware vulnerability which
1642 allows unprivileged speculative access to data which was
1643 previously stored in vector registers.
1645 This issue is mitigated by default in updated microcode.
1646 The mitigation may have a performance impact but can be
1647 disabled. On systems without the microcode mitigation
1648 disabling AVX serves as a mitigation.
1650 force: Disable AVX to mitigate systems without
1651 microcode mitigation. No effect if the microcode
1652 mitigation is present. Known to cause crashes in
1653 userspace with buggy AVX enumeration.
1655 off: Disable GDS mitigation.
1657 gcov_persist= [GCOV] When non-zero (default), profiling data for
1658 kernel modules is saved and remains accessible via
1659 debugfs, even when the module is unloaded/reloaded.
1660 When zero, profiling data is discarded and associated
1661 debugfs files are removed at module unload time.
1663 goldfish [X86] Enable the goldfish android emulator platform.
1664 Don't use this when you are not running on the
1667 gpio-mockup.gpio_mockup_ranges
1668 [HW] Sets the ranges of gpiochip of for this device.
1669 Format: <start1>,<end1>,<start2>,<end2>...
1670 gpio-mockup.gpio_mockup_named_lines
1671 [HW] Let the driver know GPIO lines should be named.
1673 gpt [EFI] Forces disk with valid GPT signature but
1674 invalid Protective MBR to be treated as GPT. If the
1675 primary GPT is corrupted, it enables the backup/alternate
1676 GPT to be used instead.
1678 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1679 the "Enable 0" bit of the configuration register.
1682 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1683 the "Enable 0" bit of the configuration register.
1686 grcan.select= [HW] Select which physical interface to use.
1689 grcan.txsize= [HW] Sets the size of the tx buffer.
1690 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1692 grcan.rxsize= [HW] Sets the size of the rx buffer.
1693 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1697 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
1698 hardening is enabled for this boot. Hardened
1699 usercopy checking is used to protect the kernel
1700 from reading or writing beyond known memory
1701 allocation boundaries as a proactive defense
1702 against bounds-checking flaws in the kernel's
1703 copy_to_user()/copy_from_user() interface.
1704 on Perform hardened usercopy checks (default).
1705 off Disable hardened usercopy checks.
1707 hardlockup_all_cpu_backtrace=
1708 [KNL] Should the hard-lockup detector generate
1709 backtraces on all cpus.
1712 hashdist= [KNL,NUMA] Large hashes allocated during boot
1713 are distributed across NUMA nodes. Defaults on
1714 for 64-bit NUMA, off otherwise.
1715 Format: 0 | 1 (for off | on)
1717 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1719 hd= [EIDE] (E)IDE hard drive subsystem geometry
1720 Format: <cyl>,<head>,<sect>
1723 Disable Hardware Error Source Table (HEST) support;
1724 corresponding firmware-first mode error processing
1725 logic will be disabled.
1727 hibernate= [HIBERNATION]
1728 noresume Don't check if there's a hibernation image
1729 present during boot.
1730 nocompress Don't compress/decompress hibernation images.
1731 no Disable hibernation and resume.
1732 protect_image Turn on image protection during restoration
1733 (that will set all pages holding image data
1734 during restoration read-only).
1736 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1737 size of <nn>. This works even on boxes that have no
1738 highmem otherwise. This also works to reduce highmem
1739 size on bigger boxes.
1741 highres= [KNL] Enable/disable high resolution timer mode.
1742 Valid parameters: "on", "off"
1747 hostname= [KNL] Set the hostname (aka UTS nodename).
1749 This allows setting the system's hostname during early
1750 startup. This sets the name returned by gethostname.
1751 Using this parameter to set the hostname makes it
1752 possible to ensure the hostname is correctly set before
1753 any userspace processes run, avoiding the possibility
1754 that a process may call gethostname before the hostname
1755 has been explicitly set, resulting in the calling
1756 process getting an incorrect result. The string must
1757 not exceed the maximum allowed hostname length (usually
1758 64 characters) and will be truncated otherwise.
1760 hpet= [X86-32,HPET] option to control HPET usage
1761 Format: { enable (default) | disable | force |
1763 disable: disable HPET and use PIT instead
1764 force: allow force enabled of undocumented chips (ICH4,
1766 verbose: show contents of HPET registers during setup
1768 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1769 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1771 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1772 If this follows hugepagesz (below), it specifies
1773 the number of pages of hugepagesz to be allocated.
1774 If this is the first HugeTLB parameter on the command
1775 line, it specifies the number of pages to allocate for
1776 the default huge page size. If using node format, the
1777 number of pages to allocate per-node can be specified.
1778 See also Documentation/admin-guide/mm/hugetlbpage.rst.
1779 Format: <integer> or (node format)
1780 <node>:<integer>[,<node>:<integer>]
1783 [HW] The size of the HugeTLB pages. This is used in
1784 conjunction with hugepages (above) to allocate huge
1785 pages of a specific size at boot. The pair
1786 hugepagesz=X hugepages=Y can be specified once for
1787 each supported huge page size. Huge page sizes are
1788 architecture dependent. See also
1789 Documentation/admin-guide/mm/hugetlbpage.rst.
1792 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1793 of gigantic hugepages. Or using node format, the size
1794 of a CMA area per node can be specified.
1795 Format: nn[KMGTPE] or (node format)
1796 <node>:nn[KMGTPE][,<node>:nn[KMGTPE]]
1798 Reserve a CMA area of given size and allocate gigantic
1799 hugepages using the CMA allocator. If enabled, the
1800 boot-time allocation of gigantic hugepages is skipped.
1802 hugetlb_free_vmemmap=
1803 [KNL] Requires CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP
1805 Control if HugeTLB Vmemmap Optimization (HVO) is enabled.
1806 Allows heavy hugetlb users to free up some more
1807 memory (7 * PAGE_SIZE for each 2MB hugetlb page).
1808 Format: { on | off (default) }
1813 Built with CONFIG_HUGETLB_PAGE_OPTIMIZE_VMEMMAP_DEFAULT_ON=y,
1816 Note that the vmemmap pages may be allocated from the added
1817 memory block itself when memory_hotplug.memmap_on_memory is
1818 enabled, those vmemmap pages cannot be optimized even if this
1819 feature is enabled. Other vmemmap pages not allocated from
1820 the added memory block itself do not be affected.
1823 [KNL] Should the hung task detector generate panics.
1826 A value of 1 instructs the kernel to panic when a
1827 hung task is detected. The default value is controlled
1828 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1829 option. The value selected by this boot parameter can
1830 be changed later by the kernel.hung_task_panic sysctl.
1832 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1833 terminal devices. Valid values: 0..8
1834 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1835 If specified, z/VM IUCV HVC accepts connections
1836 from listed z/VM user IDs only.
1838 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1839 which allow the hypervisor to 'idle' the
1840 guest on lock contention.
1842 i2c_bus= [HW] Override the default board specific I2C bus speed
1843 or register an additional I2C bus that is not
1844 registered from board initialization code.
1848 i8042.debug [HW] Toggle i8042 debug mode
1849 i8042.unmask_kbd_data
1850 [HW] Enable printing of interrupt data from the KBD port
1851 (disabled by default, and as a pre-condition
1852 requires that i8042.debug=1 be enabled)
1853 i8042.direct [HW] Put keyboard port into non-translated mode
1854 i8042.dumbkbd [HW] Pretend that controller can only read data from
1855 keyboard and cannot control its state
1856 (Don't attempt to blink the leds)
1857 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1858 i8042.nokbd [HW] Don't check/create keyboard port
1859 i8042.noloop [HW] Disable the AUX Loopback command while probing
1861 i8042.nomux [HW] Don't check presence of an active multiplexing
1863 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1865 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1866 i8042.reset [HW] Reset the controller during init, cleanup and
1867 suspend-to-ram transitions, only during s2r
1868 transitions, or never reset
1869 Format: { 1 | Y | y | 0 | N | n }
1870 1, Y, y: always reset controller
1871 0, N, n: don't ever reset controller
1872 Default: only on s2r transitions on x86; most other
1873 architectures force reset to be always executed
1874 i8042.unlock [HW] Unlock (ignore) the keylock
1875 i8042.kbdreset [HW] Reset device connected to KBD port
1877 [HW] Allow deferred probing upon i8042 probe errors
1881 i915.invert_brightness=
1882 [DRM] Invert the sense of the variable that is used to
1883 set the brightness of the panel backlight. Normally a
1884 brightness value of 0 indicates backlight switched off,
1885 and the maximum of the brightness value sets the backlight
1886 to maximum brightness. If this parameter is set to 0
1887 (default) and the machine requires it, or this parameter
1888 is set to 1, a brightness value of 0 sets the backlight
1889 to maximum brightness, and the maximum of the brightness
1890 value switches the backlight off.
1891 -1 -- never invert brightness
1892 0 -- machine default
1893 1 -- force brightness inversion
1896 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1900 Format: idle=poll, idle=halt, idle=nomwait
1901 Poll forces a polling idle loop that can slightly
1902 improve the performance of waking up a idle CPU, but
1903 will use a lot of power and make the system run hot.
1905 idle=halt: Halt is forced to be used for CPU idle.
1906 In such case C2/C3 won't be used again.
1907 idle=nomwait: Disable mwait for CPU C-states
1911 Allow force disabling of Shared Virtual Memory (SVA)
1912 support for the idxd driver. By default it is set to
1915 idxd.tc_override= [HW]
1917 Allow override of default traffic class configuration
1918 for the device. By default it is set to false (0).
1920 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1921 Format: { strict | legacy | 2008 | relaxed }
1924 Choose which programs will be accepted for execution
1925 based on the IEEE 754 NaN encoding(s) supported by
1926 the FPU and the NaN encoding requested with the value
1927 of an ELF file header flag individually set by each
1928 binary. Hardware implementations are permitted to
1929 support either or both of the legacy and the 2008 NaN
1932 Available settings are as follows:
1933 strict accept binaries that request a NaN encoding
1934 supported by the FPU
1935 legacy only accept legacy-NaN binaries, if supported
1937 2008 only accept 2008-NaN binaries, if supported
1939 relaxed accept any binaries regardless of whether
1940 supported by the FPU
1942 The FPU emulator is always able to support both NaN
1943 encodings, so if no FPU hardware is present or it has
1944 been disabled with 'nofpu', then the settings of
1945 'legacy' and '2008' strap the emulator accordingly,
1946 'relaxed' straps the emulator for both legacy-NaN and
1947 2008-NaN, whereas 'strict' enables legacy-NaN only on
1948 legacy processors and both NaN encodings on MIPS32 or
1951 The setting for ABS.fmt/NEG.fmt instruction execution
1952 mode generally follows that for the NaN encoding,
1953 except where unsupported by hardware.
1955 ignore_loglevel [KNL]
1956 Ignore loglevel setting - this will print /all/
1957 kernel messages to the console. Useful for debugging.
1958 We also add it as printk module parameter, so users
1959 could change it dynamically, usually by
1960 /sys/module/printk/parameters/ignore_loglevel.
1963 Ignore RLIMIT_DATA setting for data mappings,
1964 print warning at first misuse. Can be changed via
1965 /sys/module/kernel/parameters/ignore_rlimit_data.
1967 ihash_entries= [KNL]
1968 Set number of hash buckets for inode cache.
1970 ima_appraise= [IMA] appraise integrity measurements
1971 Format: { "off" | "enforce" | "fix" | "log" }
1974 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1975 The builtin appraise policy appraises all files
1978 ima_canonical_fmt [IMA]
1979 Use the canonical format for the binary runtime
1980 measurements, instead of host native format.
1983 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1987 The list of supported hash algorithms is defined
1988 in crypto/hash_info.h.
1991 The builtin policies to load during IMA setup.
1992 Format: "tcb | appraise_tcb | secure_boot |
1993 fail_securely | critical_data"
1995 The "tcb" policy measures all programs exec'd, files
1996 mmap'd for exec, and all files opened with the read
1997 mode bit set by either the effective uid (euid=0) or
2000 The "appraise_tcb" policy appraises the integrity of
2001 all files owned by root.
2003 The "secure_boot" policy appraises the integrity
2004 of files (eg. kexec kernel image, kernel modules,
2005 firmware, policy, etc) based on file signatures.
2007 The "fail_securely" policy forces file signature
2008 verification failure also on privileged mounted
2009 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
2012 The "critical_data" policy measures kernel integrity
2015 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
2016 Load a policy which meets the needs of the Trusted
2017 Computing Base. This means IMA will measure all
2018 programs exec'd, files mmap'd for exec, and all files
2019 opened for read by uid=0.
2022 Select one of defined IMA measurements template formats.
2023 Formats: { "ima" | "ima-ng" | "ima-ngv2" | "ima-sig" |
2028 [IMA] Define a custom template format.
2029 Format: { "field1|...|fieldN" }
2031 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
2032 Format: <min_file_size>
2033 Set the minimal file size for using asynchronous hash.
2034 If left unspecified, ahash usage is disabled.
2036 ahash performance varies for different data sizes on
2037 different crypto accelerators. This option can be used
2038 to achieve the best performance for a particular HW.
2040 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
2042 Set hashing buffer size. Default: 4k.
2044 ahash performance varies for different chunk sizes on
2045 different crypto accelerators. This option can be used
2046 to achieve best performance for particular HW.
2050 Run specified binary instead of /sbin/init as init
2053 initcall_debug [KNL] Trace initcalls as they are executed. Useful
2054 for working out where the kernel is dying during
2057 initcall_blacklist= [KNL] Do not execute a comma-separated list of
2058 initcall functions. Useful for debugging built-in
2059 modules and initcalls.
2061 initramfs_async= [KNL]
2064 This parameter controls whether the initramfs
2065 image is unpacked asynchronously, concurrently
2066 with devices being probed and
2067 initialized. This should normally just work,
2068 but as a debugging aid, one can get the
2069 historical behaviour of the initramfs
2070 unpacking being completed before device_ and
2073 initrd= [BOOT] Specify the location of the initial ramdisk
2075 initrdmem= [KNL] Specify a physical address and size from which to
2076 load the initrd. If an initrd is compiled in or
2077 specified in the bootparams, it takes priority over this
2079 Format: ss[KMG],nn[KMG]
2082 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
2085 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
2087 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
2089 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
2091 init_pkru= [X86] Specify the default memory protection keys rights
2092 register contents for all processes. 0x55555554 by
2093 default (disallow access to all but pkey 0). Can
2094 override in debugfs after boot.
2096 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
2099 int_pln_enable [X86] Enable power limit notification interrupt
2101 integrity_audit=[IMA]
2102 Format: { "0" | "1" }
2103 0 -- basic integrity auditing messages. (Default)
2104 1 -- additional integrity auditing messages.
2106 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
2108 Enable intel iommu driver.
2110 Disable intel iommu driver.
2111 igfx_off [Default Off]
2112 By default, gfx is mapped as normal device. If a gfx
2113 device has a dedicated DMAR unit, the DMAR unit is
2114 bypassed by not enabling DMAR with this option. In
2115 this case, gfx device will use physical address for
2117 strict [Default Off]
2118 Deprecated, equivalent to iommu.strict=1.
2119 sp_off [Default Off]
2120 By default, super page will be supported if Intel IOMMU
2121 has the capability. With this option, super page will
2124 Enable the Intel IOMMU scalable mode if the hardware
2125 advertises that it has support for the scalable mode
2128 Disallow use of the Intel IOMMU scalable mode.
2129 tboot_noforce [Default Off]
2130 Do not force the Intel IOMMU enabled under tboot.
2131 By default, tboot will force Intel IOMMU on, which
2132 could harm performance of some high-throughput
2133 devices like 40GBit network cards, even if identity
2135 Note that using this option lowers the security
2136 provided by tboot because it makes the system
2137 vulnerable to DMA attacks.
2139 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
2140 0 disables intel_idle and fall back on acpi_idle.
2141 1 to 9 specify maximum depth of C-state.
2145 Do not enable intel_pstate as the default
2146 scaling driver for the supported processors
2148 Use intel_pstate driver to bypass the scaling
2149 governors layer of cpufreq and provides it own
2150 algorithms for p-state selection. There are two
2151 P-state selection algorithms provided by
2152 intel_pstate in the active mode: powersave and
2153 performance. The way they both operate depends
2154 on whether or not the hardware managed P-states
2155 (HWP) feature has been enabled in the processor
2156 and possibly on the processor model.
2158 Use intel_pstate as a scaling driver, but configure it
2159 to work with generic cpufreq governors (instead of
2160 enabling its internal governor). This mode cannot be
2161 used along with the hardware-managed P-states (HWP)
2164 Enable intel_pstate on systems that prohibit it by default
2165 in favor of acpi-cpufreq. Forcing the intel_pstate driver
2166 instead of acpi-cpufreq may disable platform features, such
2167 as thermal controls and power capping, that rely on ACPI
2168 P-States information being indicated to OSPM and therefore
2169 should be used with caution. This option does not work with
2170 processors that aren't supported by the intel_pstate driver
2171 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
2173 Do not enable hardware P state control (HWP)
2176 Only load intel_pstate on systems which support
2177 hardware P state control (HWP) if available.
2179 Enforce ACPI _PPC performance limits. If the Fixed ACPI
2180 Description Table, specifies preferred power management
2181 profile as "Enterprise Server" or "Performance Server",
2182 then this feature is turned on by default.
2184 Allow per-logical-CPU P-State performance control limits using
2185 cpufreq sysfs interface
2187 intremap= [X86-64, Intel-IOMMU]
2188 on enable Interrupt Remapping (default)
2189 off disable Interrupt Remapping
2190 nosid disable Source ID checking
2192 BIOS x2APIC opt-out request will be ignored
2193 nopost disable Interrupt Posting
2195 iomem= Disable strict checking of access to MMIO memory
2196 strict regions from userspace.
2211 nobypass [PPC/POWERNV]
2212 Disable IOMMU bypass, using IOMMU for PCI devices.
2214 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
2215 Format: { "0" | "1" }
2216 0 - Try to allocate a 32-bit DMA address first, before
2217 falling back to the full range if needed.
2218 1 - Allocate directly from the full usable range,
2219 forcing Dual Address Cycle for PCI cards supporting
2220 greater than 32-bit addressing.
2222 iommu.strict= [ARM64, X86] Configure TLB invalidation behaviour
2223 Format: { "0" | "1" }
2225 Request that DMA unmap operations use deferred
2226 invalidation of hardware TLBs, for increased
2227 throughput at the cost of reduced device isolation.
2228 Will fall back to strict mode if not supported by
2229 the relevant IOMMU driver.
2231 DMA unmap operations invalidate IOMMU hardware TLBs
2233 unset - Use value of CONFIG_IOMMU_DEFAULT_DMA_{LAZY,STRICT}.
2234 Note: on x86, strict mode specified via one of the
2235 legacy driver-specific options takes precedence.
2238 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2239 Format: { "0" | "1" }
2240 0 - Use IOMMU translation for DMA.
2241 1 - Bypass the IOMMU for DMA.
2242 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2244 io7= [HW] IO7 for Marvel-based Alpha systems
2245 See comment before marvel_specify_io7 in
2246 arch/alpha/kernel/core_marvel.c.
2248 io_delay= [X86] I/O delay method
2250 Standard port 0x80 based delay
2252 Alternate port 0xed based delay (needed on some systems)
2254 Simple two microseconds delay
2259 See Documentation/admin-guide/nfs/nfsroot.rst.
2261 ipcmni_extend [KNL] Extend the maximum number of unique System V
2262 IPC identifiers from 32,768 to 16,777,216.
2264 irqaffinity= [SMP] Set the default irq affinity mask
2265 The argument is a cpu list, as described above.
2267 irqchip.gicv2_force_probe=
2270 Force the kernel to look for the second 4kB page
2271 of a GICv2 controller even if the memory range
2272 exposed by the device tree is too small.
2274 irqchip.gicv3_nolpi=
2276 Force the kernel to ignore the availability of
2277 LPIs (and by consequence ITSs). Intended for system
2278 that use the kernel as a bootloader, and thus want
2279 to let secondary kernels in charge of setting up
2282 irqchip.gicv3_pseudo_nmi= [ARM64]
2283 Enables support for pseudo-NMIs in the kernel. This
2284 requires the kernel to be built with
2285 CONFIG_ARM64_PSEUDO_NMI.
2288 When an interrupt is not handled search all handlers
2289 for it. Intended to get systems with badly broken
2293 When an interrupt is not handled search all handlers
2294 for it. Also check all handlers each timer
2295 interrupt. Intended to get systems with badly broken
2299 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2301 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2302 [Deprecated - use cpusets instead]
2303 Format: [flag-list,]<cpu-list>
2305 Specify one or more CPUs to isolate from disturbances
2306 specified in the flag list (default: domain):
2309 Disable the tick when a single task runs.
2311 A residual 1Hz tick is offloaded to workqueues, which you
2312 need to affine to housekeeping through the global
2313 workqueue's affinity configured via the
2314 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2315 by using the 'domain' flag described below.
2317 NOTE: by default the global workqueue runs on all CPUs,
2318 so to protect individual CPUs the 'cpumask' file has to
2319 be configured manually after bootup.
2322 Isolate from the general SMP balancing and scheduling
2323 algorithms. Note that performing domain isolation this way
2324 is irreversible: it's not possible to bring back a CPU to
2325 the domains once isolated through isolcpus. It's strongly
2326 advised to use cpusets instead to disable scheduler load
2327 balancing through the "cpuset.sched_load_balance" file.
2328 It offers a much more flexible interface where CPUs can
2329 move in and out of an isolated set anytime.
2331 You can move a process onto or off an "isolated" CPU via
2332 the CPU affinity syscalls or cpuset.
2333 <cpu number> begins at 0 and the maximum value is
2334 "number of CPUs in system - 1".
2338 Isolate from being targeted by managed interrupts
2339 which have an interrupt mask containing isolated
2340 CPUs. The affinity of managed interrupts is
2341 handled by the kernel and cannot be changed via
2342 the /proc/irq/* interfaces.
2344 This isolation is best effort and only effective
2345 if the automatically assigned interrupt mask of a
2346 device queue contains isolated and housekeeping
2347 CPUs. If housekeeping CPUs are online then such
2348 interrupts are directed to the housekeeping CPU
2349 so that IO submitted on the housekeeping CPU
2350 cannot disturb the isolated CPU.
2352 If a queue's affinity mask contains only isolated
2353 CPUs then this parameter has no effect on the
2354 interrupt routing decision, though interrupts are
2355 only delivered when tasks running on those
2356 isolated CPUs submit IO. IO submitted on
2357 housekeeping CPUs has no influence on those
2360 The format of <cpu-list> is described above.
2364 ivrs_ioapic [HW,X86-64]
2365 Provide an override to the IOAPIC-ID<->DEVICE-ID
2366 mapping provided in the IVRS ACPI table.
2367 By default, PCI segment is 0, and can be omitted.
2369 For example, to map IOAPIC-ID decimal 10 to
2370 PCI segment 0x1 and PCI device 00:14.0,
2371 write the parameter as:
2372 ivrs_ioapic=10@0001:00:14.0
2375 * To map IOAPIC-ID decimal 10 to PCI device 00:14.0
2376 write the parameter as:
2377 ivrs_ioapic[10]=00:14.0
2378 * To map IOAPIC-ID decimal 10 to PCI segment 0x1 and
2379 PCI device 00:14.0 write the parameter as:
2380 ivrs_ioapic[10]=0001:00:14.0
2382 ivrs_hpet [HW,X86-64]
2383 Provide an override to the HPET-ID<->DEVICE-ID
2384 mapping provided in the IVRS ACPI table.
2385 By default, PCI segment is 0, and can be omitted.
2387 For example, to map HPET-ID decimal 10 to
2388 PCI segment 0x1 and PCI device 00:14.0,
2389 write the parameter as:
2390 ivrs_hpet=10@0001:00:14.0
2393 * To map HPET-ID decimal 0 to PCI device 00:14.0
2394 write the parameter as:
2395 ivrs_hpet[0]=00:14.0
2396 * To map HPET-ID decimal 10 to PCI segment 0x1 and
2397 PCI device 00:14.0 write the parameter as:
2398 ivrs_ioapic[10]=0001:00:14.0
2400 ivrs_acpihid [HW,X86-64]
2401 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2402 mapping provided in the IVRS ACPI table.
2403 By default, PCI segment is 0, and can be omitted.
2405 For example, to map UART-HID:UID AMD0020:0 to
2406 PCI segment 0x1 and PCI device ID 00:14.5,
2407 write the parameter as:
2408 ivrs_acpihid=AMD0020:0@0001:00:14.5
2411 * To map UART-HID:UID AMD0020:0 to PCI segment is 0,
2412 PCI device ID 00:14.5, write the parameter as:
2413 ivrs_acpihid[00:14.5]=AMD0020:0
2414 * To map UART-HID:UID AMD0020:0 to PCI segment 0x1 and
2415 PCI device ID 00:14.5, write the parameter as:
2416 ivrs_acpihid[0001:00:14.5]=AMD0020:0
2418 js= [HW,JOY] Analog joystick
2419 See Documentation/input/joydev/joystick.rst.
2422 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2423 report on every invalid memory access. Without this
2424 parameter KASAN will print report only for the first
2428 Do not unregister boot console at start. This is only
2429 useful for debugging when something happens in the window
2430 between unregistering the boot console and initializing
2435 kernelcore= [KNL,X86,IA-64,PPC]
2436 Format: nn[KMGTPE] | nn% | "mirror"
2437 This parameter specifies the amount of memory usable by
2438 the kernel for non-movable allocations. The requested
2439 amount is spread evenly throughout all nodes in the
2440 system as ZONE_NORMAL. The remaining memory is used for
2441 movable memory in its own zone, ZONE_MOVABLE. In the
2442 event, a node is too small to have both ZONE_NORMAL and
2443 ZONE_MOVABLE, kernelcore memory will take priority and
2444 other nodes will have a larger ZONE_MOVABLE.
2446 ZONE_MOVABLE is used for the allocation of pages that
2447 may be reclaimed or moved by the page migration
2448 subsystem. Note that allocations like PTEs-from-HighMem
2449 still use the HighMem zone if it exists, and the Normal
2450 zone if it does not.
2452 It is possible to specify the exact amount of memory in
2453 the form of "nn[KMGTPE]", a percentage of total system
2454 memory in the form of "nn%", or "mirror". If "mirror"
2455 option is specified, mirrored (reliable) memory is used
2456 for non-movable allocations and remaining memory is used
2457 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2458 are exclusive, so you cannot specify multiple forms.
2460 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2461 Format: <Controller#>[,poll interval]
2462 The controller # is the number of the ehci usb debug
2463 port as it is probed via PCI. The poll interval is
2464 optional and is the number seconds in between
2465 each poll cycle to the debug port in case you need
2466 the functionality for interrupting the kernel with
2467 gdb or control-c on the dbgp connection. When
2468 not using this parameter you use sysrq-g to break into
2469 the kernel debugger.
2471 kgdboc= [KGDB,HW] kgdb over consoles.
2472 Requires a tty driver that supports console polling,
2473 or a supported polling keyboard driver (non-usb).
2474 Serial only format: <serial_device>[,baud]
2475 keyboard only format: kbd
2476 keyboard and serial format: kbd,<serial_device>[,baud]
2477 Optional Kernel mode setting:
2478 kms, kbd format: kms,kbd
2479 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2481 kgdboc_earlycon= [KGDB,HW]
2482 If the boot console provides the ability to read
2483 characters and can work in polling mode, you can use
2484 this parameter to tell kgdb to use it as a backend
2485 until the normal console is registered. Intended to
2486 be used together with the kgdboc parameter which
2487 specifies the normal console to transition to.
2489 The name of the early console should be specified
2490 as the value of this parameter. Note that the name of
2491 the early console might be different than the tty
2492 name passed to kgdboc. It's OK to leave the value
2493 blank and the first boot console that implements
2494 read() will be picked.
2496 kgdbwait [KGDB] Stop kernel execution and enter the
2497 kernel debugger at the earliest opportunity.
2499 kmac= [MIPS] Korina ethernet MAC address.
2500 Configure the RouterBoard 532 series on-chip
2501 Ethernet adapter MAC address.
2503 kmemleak= [KNL] Boot-time kmemleak enable/disable
2504 Valid arguments: on, off
2506 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2509 kprobe_event=[probe-list]
2510 [FTRACE] Add kprobe events and enable at boot time.
2511 The probe-list is a semicolon delimited list of probe
2512 definitions. Each definition is same as kprobe_events
2513 interface, but the parameters are comma delimited.
2514 For example, to add a kprobe event on vfs_read with
2515 arg1 and arg2, add to the command line;
2517 kprobe_event=p,vfs_read,$arg1,$arg2
2519 See also Documentation/trace/kprobetrace.rst "Kernel
2520 Boot Parameter" section.
2522 kpti= [ARM64] Control page table isolation of user
2523 and kernel address spaces.
2524 Default: enabled on cores which need mitigation.
2528 kunit.enable= [KUNIT] Enable executing KUnit tests. Requires
2529 CONFIG_KUNIT to be set to be fully enabled. The
2530 default value can be overridden via
2531 KUNIT_DEFAULT_ENABLED.
2532 Default is 1 (enabled)
2534 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2535 Default is 0 (don't ignore, but inject #GP)
2537 kvm.eager_page_split=
2538 [KVM,X86] Controls whether or not KVM will try to
2539 proactively split all huge pages during dirty logging.
2540 Eager page splitting reduces interruptions to vCPU
2541 execution by eliminating the write-protection faults
2542 and MMU lock contention that would otherwise be
2543 required to split huge pages lazily.
2545 VM workloads that rarely perform writes or that write
2546 only to a small region of VM memory may benefit from
2547 disabling eager page splitting to allow huge pages to
2548 still be used for reads.
2550 The behavior of eager page splitting depends on whether
2551 KVM_DIRTY_LOG_INITIALLY_SET is enabled or disabled. If
2552 disabled, all huge pages in a memslot will be eagerly
2553 split when dirty logging is enabled on that memslot. If
2554 enabled, eager page splitting will be performed during
2555 the KVM_CLEAR_DIRTY ioctl, and only for the pages being
2558 Eager page splitting is only supported when kvm.tdp_mmu=Y.
2562 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2563 Default is false (don't support).
2566 [KVM] Controls the software workaround for the
2567 X86_BUG_ITLB_MULTIHIT bug.
2568 force : Always deploy workaround.
2569 off : Never deploy workaround.
2570 auto : Deploy workaround based on the presence of
2571 X86_BUG_ITLB_MULTIHIT.
2575 If the software workaround is enabled for the host,
2576 guests do need not to enable it for nested guests.
2578 kvm.nx_huge_pages_recovery_ratio=
2579 [KVM] Controls how many 4KiB pages are periodically zapped
2580 back to huge pages. 0 disables the recovery, otherwise if
2581 the value is N KVM will zap 1/Nth of the 4KiB pages every
2582 period (see below). The default is 60.
2584 kvm.nx_huge_pages_recovery_period_ms=
2585 [KVM] Controls the time period at which KVM zaps 4KiB pages
2586 back to huge pages. If the value is a non-zero N, KVM will
2587 zap a portion (see ratio above) of the pages every N msecs.
2588 If the value is 0 (the default), KVM will pick a period based
2589 on the ratio, such that a page is zapped after 1 hour on average.
2591 kvm-amd.nested= [KVM,AMD] Control nested virtualization feature in
2592 KVM/SVM. Default is 1 (enabled).
2594 kvm-amd.npt= [KVM,AMD] Control KVM's use of Nested Page Tables,
2595 a.k.a. Two-Dimensional Page Tables. Default is 1
2596 (enabled). Disable by KVM if hardware lacks support
2600 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2602 none: Forcefully disable KVM.
2604 nvhe: Standard nVHE-based mode, without support for
2607 protected: nVHE-based mode with support for guests whose
2608 state is kept private from the host.
2610 nested: VHE-based mode with support for nested
2611 virtualization. Requires at least ARMv8.3
2614 Defaults to VHE/nVHE based on hardware support. Setting
2615 mode to "protected" will disable kexec and hibernation
2616 for the host. "nested" is experimental and should be
2617 used with extreme caution.
2619 kvm-arm.vgic_v3_group0_trap=
2620 [KVM,ARM] Trap guest accesses to GICv3 group-0
2623 kvm-arm.vgic_v3_group1_trap=
2624 [KVM,ARM] Trap guest accesses to GICv3 group-1
2627 kvm-arm.vgic_v3_common_trap=
2628 [KVM,ARM] Trap guest accesses to GICv3 common
2631 kvm-arm.vgic_v4_enable=
2632 [KVM,ARM] Allow use of GICv4 for direct injection of
2635 kvm_cma_resv_ratio=n [PPC]
2636 Reserves given percentage from system memory area for
2637 contiguous memory allocation for KVM hash pagetable
2639 By default it reserves 5% of total system memory.
2643 kvm-intel.ept= [KVM,Intel] Control KVM's use of Extended Page Tables,
2644 a.k.a. Two-Dimensional Page Tables. Default is 1
2645 (enabled). Disable by KVM if hardware lacks support
2648 kvm-intel.emulate_invalid_guest_state=
2649 [KVM,Intel] Control whether to emulate invalid guest
2650 state. Ignored if kvm-intel.enable_unrestricted_guest=1,
2651 as guest state is never invalid for unrestricted
2652 guests. This param doesn't apply to nested guests (L2),
2653 as KVM never emulates invalid L2 guest state.
2654 Default is 1 (enabled).
2656 kvm-intel.flexpriority=
2657 [KVM,Intel] Control KVM's use of FlexPriority feature
2658 (TPR shadow). Default is 1 (enabled). Disable by KVM if
2659 hardware lacks support for it.
2662 [KVM,Intel] Control nested virtualization feature in
2663 KVM/VMX. Default is 1 (enabled).
2665 kvm-intel.unrestricted_guest=
2666 [KVM,Intel] Control KVM's use of unrestricted guest
2667 feature (virtualized real and unpaged mode). Default
2668 is 1 (enabled). Disable by KVM if EPT is disabled or
2669 hardware lacks support for it.
2671 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2674 Valid arguments: never, cond, always
2676 always: L1D cache flush on every VMENTER.
2677 cond: Flush L1D on VMENTER only when the code between
2678 VMEXIT and VMENTER can leak host memory.
2679 never: Disables the mitigation
2681 Default is cond (do L1 cache flush in specific instances)
2683 kvm-intel.vpid= [KVM,Intel] Control KVM's use of Virtual Processor
2684 Identification feature (tagged TLBs). Default is 1
2685 (enabled). Disable by KVM if hardware lacks support
2688 l1d_flush= [X86,INTEL]
2689 Control mitigation for L1D based snooping vulnerability.
2691 Certain CPUs are vulnerable to an exploit against CPU
2692 internal buffers which can forward information to a
2693 disclosure gadget under certain conditions.
2695 In vulnerable processors, the speculatively
2696 forwarded data can be used in a cache side channel
2697 attack, to access data to which the attacker does
2698 not have direct access.
2700 This parameter controls the mitigation. The
2703 on - enable the interface for the mitigation
2705 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2708 The kernel PTE inversion protection is unconditionally
2709 enabled and cannot be disabled.
2712 Provides all available mitigations for the
2713 L1TF vulnerability. Disables SMT and
2714 enables all mitigations in the
2715 hypervisors, i.e. unconditional L1D flush.
2717 SMT control and L1D flush control via the
2718 sysfs interface is still possible after
2719 boot. Hypervisors will issue a warning
2720 when the first VM is started in a
2721 potentially insecure configuration,
2722 i.e. SMT enabled or L1D flush disabled.
2725 Same as 'full', but disables SMT and L1D
2726 flush runtime control. Implies the
2727 'nosmt=force' command line option.
2728 (i.e. sysfs control of SMT is disabled.)
2731 Leaves SMT enabled and enables the default
2732 hypervisor mitigation, i.e. conditional
2735 SMT control and L1D flush control via the
2736 sysfs interface is still possible after
2737 boot. Hypervisors will issue a warning
2738 when the first VM is started in a
2739 potentially insecure configuration,
2740 i.e. SMT enabled or L1D flush disabled.
2744 Disables SMT and enables the default
2745 hypervisor mitigation.
2747 SMT control and L1D flush control via the
2748 sysfs interface is still possible after
2749 boot. Hypervisors will issue a warning
2750 when the first VM is started in a
2751 potentially insecure configuration,
2752 i.e. SMT enabled or L1D flush disabled.
2755 Same as 'flush', but hypervisors will not
2756 warn when a VM is started in a potentially
2757 insecure configuration.
2760 Disables hypervisor mitigations and doesn't
2762 It also drops the swap size and available
2763 RAM limit restriction on both hypervisor and
2768 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2774 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2777 lapic= [X86,APIC] Do not use TSC deadline
2778 value for LAPIC timer one-shot implementation. Default
2779 back to the programmable timer unit in the LAPIC.
2780 Format: notscdeadline
2782 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2785 libata.dma= [LIBATA] DMA control
2786 libata.dma=0 Disable all PATA and SATA DMA
2787 libata.dma=1 PATA and SATA Disk DMA only
2788 libata.dma=2 ATAPI (CDROM) DMA only
2789 libata.dma=4 Compact Flash DMA only
2790 Combinations also work, so libata.dma=3 enables DMA
2791 for disks and CDROMs, but not CFs.
2793 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2794 libata.ignore_hpa=0 keep BIOS limits (default)
2795 libata.ignore_hpa=1 ignore limits, using full disk
2797 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2801 libata.force= [LIBATA] Force configurations. The format is a comma-
2802 separated list of "[ID:]VAL" where ID is PORT[.DEVICE].
2803 PORT and DEVICE are decimal numbers matching port, link
2804 or device. Basically, it matches the ATA ID string
2805 printed on console by libata. If the whole ID part is
2806 omitted, the last PORT and DEVICE values are used. If
2807 ID hasn't been specified yet, the configuration applies
2808 to all ports, links and devices.
2810 If only DEVICE is omitted, the parameter applies to
2811 the port and all links and devices behind it. DEVICE
2812 number of 0 either selects the first device or the
2813 first fan-out link behind PMP device. It does not
2814 select the host link. DEVICE number of 15 selects the
2815 host link and device attached to it.
2817 The VAL specifies the configuration to force. As long
2818 as there is no ambiguity, shortcut notation is allowed.
2819 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2820 The following configurations can be forced.
2822 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2823 Any ID with matching PORT is used.
2825 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2827 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2828 udma[/][16,25,33,44,66,100,133] notation is also
2831 * nohrst, nosrst, norst: suppress hard, soft and both
2834 * rstonce: only attempt one reset during hot-unplug
2837 * [no]dbdelay: Enable or disable the extra 200ms delay
2838 before debouncing a link PHY and device presence
2841 * [no]ncq: Turn on or off NCQ.
2843 * [no]ncqtrim: Enable or disable queued DSM TRIM.
2845 * [no]ncqati: Enable or disable NCQ trim on ATI chipset.
2847 * [no]trim: Enable or disable (unqueued) TRIM.
2849 * trim_zero: Indicate that TRIM command zeroes data.
2851 * max_trim_128m: Set 128M maximum trim size limit.
2853 * [no]dma: Turn on or off DMA transfers.
2855 * atapi_dmadir: Enable ATAPI DMADIR bridge support.
2857 * atapi_mod16_dma: Enable the use of ATAPI DMA for
2858 commands that are not a multiple of 16 bytes.
2860 * [no]dmalog: Enable or disable the use of the
2861 READ LOG DMA EXT command to access logs.
2863 * [no]iddevlog: Enable or disable access to the
2864 identify device data log.
2866 * [no]logdir: Enable or disable access to the general
2867 purpose log directory.
2869 * max_sec_128: Set transfer size limit to 128 sectors.
2871 * max_sec_1024: Set or clear transfer size limit to
2874 * max_sec_lba48: Set or clear transfer size limit to
2877 * [no]lpm: Enable or disable link power management.
2879 * [no]setxfer: Indicate if transfer speed mode setting
2882 * [no]fua: Disable or enable FUA (Force Unit Access)
2883 support for devices supporting this feature.
2885 * dump_id: Dump IDENTIFY data.
2887 * disable: Disable this device.
2889 If there are multiple matching configurations changing
2890 the same attribute, the last one is used.
2892 load_ramdisk= [RAM] [Deprecated]
2894 lockd.nlm_grace_period=P [NFS] Assign grace period.
2897 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2900 lockd.nlm_timeout=T [NFS] Assign timeout value.
2903 lockd.nlm_udpport=M [NFS] Assign UDP port.
2906 lockdown= [SECURITY]
2907 { integrity | confidentiality }
2908 Enable the kernel lockdown feature. If set to
2909 integrity, kernel features that allow userland to
2910 modify the running kernel are disabled. If set to
2911 confidentiality, kernel features that allow userland
2912 to extract confidential information from the kernel
2915 locktorture.nreaders_stress= [KNL]
2916 Set the number of locking read-acquisition kthreads.
2917 Defaults to being automatically set based on the
2918 number of online CPUs.
2920 locktorture.nwriters_stress= [KNL]
2921 Set the number of locking write-acquisition kthreads.
2923 locktorture.onoff_holdoff= [KNL]
2924 Set time (s) after boot for CPU-hotplug testing.
2926 locktorture.onoff_interval= [KNL]
2927 Set time (s) between CPU-hotplug operations, or
2928 zero to disable CPU-hotplug testing.
2930 locktorture.shuffle_interval= [KNL]
2931 Set task-shuffle interval (jiffies). Shuffling
2932 tasks allows some CPUs to go into dyntick-idle
2933 mode during the locktorture test.
2935 locktorture.shutdown_secs= [KNL]
2936 Set time (s) after boot system shutdown. This
2937 is useful for hands-off automated testing.
2939 locktorture.stat_interval= [KNL]
2940 Time (s) between statistics printk()s.
2942 locktorture.stutter= [KNL]
2943 Time (s) to stutter testing, for example,
2944 specifying five seconds causes the test to run for
2945 five seconds, wait for five seconds, and so on.
2946 This tests the locking primitive's ability to
2947 transition abruptly to and from idle.
2949 locktorture.torture_type= [KNL]
2950 Specify the locking implementation to test.
2952 locktorture.writer_fifo= [KNL]
2953 Run the write-side locktorture kthreads at
2954 sched_set_fifo() real-time priority.
2956 locktorture.verbose= [KNL]
2957 Enable additional printk() statements.
2959 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2962 loglevel= All Kernel Messages with a loglevel smaller than the
2963 console loglevel will be printed to the console. It can
2964 also be changed with klogd or other programs. The
2965 loglevels are defined as follows:
2967 0 (KERN_EMERG) system is unusable
2968 1 (KERN_ALERT) action must be taken immediately
2969 2 (KERN_CRIT) critical conditions
2970 3 (KERN_ERR) error conditions
2971 4 (KERN_WARNING) warning conditions
2972 5 (KERN_NOTICE) normal but significant condition
2973 6 (KERN_INFO) informational
2974 7 (KERN_DEBUG) debug-level messages
2976 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2977 in bytes. n must be a power of two and greater
2978 than the minimal size. The minimal size is defined
2979 by LOG_BUF_SHIFT kernel config parameter. There is
2980 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2981 that allows to increase the default size depending on
2982 the number of CPUs. See init/Kconfig for more details.
2984 logo.nologo [FB] Disables display of the built-in Linux logo.
2985 This may be used to provide more screen space for
2986 kernel log messages and is useful when debugging
2987 kernel boot problems.
2989 lp=0 [LP] Specify parallel ports to use, e.g,
2990 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2991 lp=reset first parallel port). 'lp=0' disables the
2992 lp=auto printer driver. 'lp=reset' (which can be
2993 specified in addition to the ports) causes
2994 attached printers to be reset. Using
2995 lp=port1,port2,... specifies the parallel ports
2996 to associate lp devices with, starting with
2997 lp0. A port specification may be 'none' to skip
2998 that lp device, or a parport name such as
2999 'parport0'. Specifying 'lp=auto' instead of a
3000 port specification list means that device IDs
3001 from each port should be examined, to see if
3002 an IEEE 1284-compliant printer is attached; if
3003 so, the driver will manage that printer.
3004 See also header of drivers/char/lp.c.
3007 Sets loops_per_jiffy to given constant, thus avoiding
3008 time-consuming boot-time autodetection (up to 250 ms per
3009 CPU). 0 enables autodetection (default). To determine
3010 the correct value for your kernel, boot with normal
3011 autodetection and see what value is printed. Note that
3012 on SMP systems the preset will be applied to all CPUs,
3013 which is likely to cause problems if your CPUs need
3014 significantly divergent settings. An incorrect value
3015 will cause delays in the kernel to be wrong, leading to
3016 unpredictable I/O errors and other breakage. Although
3017 unlikely, in the extreme case this might damage your
3021 Format: <io>,<irq>,<dma>
3023 lsm.debug [SECURITY] Enable LSM initialization debugging output.
3026 [SECURITY] Choose order of LSM initialization. This
3027 overrides CONFIG_LSM, and the "security=" parameter.
3029 machvec= [IA-64] Force the use of a particular machine-vector
3030 (machvec) in a generic kernel.
3031 Example: machvec=hpzx1
3033 machtype= [Loongson] Share the same kernel image file between
3034 different yeeloong laptops.
3035 Example: machtype=lemote-yeeloong-2f-7inch
3037 max_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory greater
3038 than or equal to this physical address is ignored.
3040 maxcpus= [SMP] Maximum number of processors that an SMP kernel
3041 will bring up during bootup. maxcpus=n : n >= 0 limits
3042 the kernel to bring up 'n' processors. Surely after
3043 bootup you can bring up the other plugged cpu by executing
3044 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
3045 only takes effect during system bootup.
3046 While n=0 is a special case, it is equivalent to "nosmp",
3047 which also disables the IO APIC.
3049 max_loop= [LOOP] The number of loop block devices that get
3050 (loop.max_loop) unconditionally pre-created at init time. The default
3051 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
3052 of statically allocating a predefined number, loop
3053 devices can be requested on-demand with the
3054 /dev/loop-control interface.
3056 mce [X86-32] Machine Check Exception
3058 mce=option [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
3060 md= [HW] RAID subsystems devices and level
3061 See Documentation/admin-guide/md.rst.
3064 Format: <first>,<last>
3065 Specifies range of consoles to be captured by the MDA.
3068 Control mitigation for the Micro-architectural Data
3069 Sampling (MDS) vulnerability.
3071 Certain CPUs are vulnerable to an exploit against CPU
3072 internal buffers which can forward information to a
3073 disclosure gadget under certain conditions.
3075 In vulnerable processors, the speculatively
3076 forwarded data can be used in a cache side channel
3077 attack, to access data to which the attacker does
3078 not have direct access.
3080 This parameter controls the MDS mitigation. The
3083 full - Enable MDS mitigation on vulnerable CPUs
3084 full,nosmt - Enable MDS mitigation and disable
3085 SMT on vulnerable CPUs
3086 off - Unconditionally disable MDS mitigation
3088 On TAA-affected machines, mds=off can be prevented by
3089 an active TAA mitigation as both vulnerabilities are
3090 mitigated with the same mechanism so in order to disable
3091 this mitigation, you need to specify tsx_async_abort=off
3094 Not specifying this option is equivalent to
3097 For details see: Documentation/admin-guide/hw-vuln/mds.rst
3099 mem=nn[KMG] [HEXAGON] Set the memory size.
3100 Must be specified, otherwise memory size will be 0.
3102 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
3103 Amount of memory to be used in cases as follows:
3106 2 when the kernel is not able to see the whole system memory;
3107 3 memory that lies after 'mem=' boundary is excluded from
3108 the hypervisor, then assigned to KVM guests.
3109 4 to limit the memory available for kdump kernel.
3111 [ARC,MICROBLAZE] - the limit applies only to low memory,
3112 high memory is not affected.
3114 [ARM64] - only limits memory covered by the linear
3115 mapping. The NOMAP regions are not affected.
3117 [X86] Work as limiting max address. Use together
3118 with memmap= to avoid physical address space collisions.
3119 Without memmap= PCI devices could be placed at addresses
3120 belonging to unused RAM.
3122 Note that this only takes effects during boot time since
3123 in above case 3, memory may need be hot added after boot
3124 if system memory of hypervisor is not sufficient.
3127 [ARM,MIPS] - override the memory layout reported by
3129 Define a memory region of size nn[KMG] starting at
3131 Multiple different regions can be specified with
3132 multiple mem= parameters on the command line.
3134 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
3137 memblock=debug [KNL] Enable memblock debug messages.
3140 [KNL,SH] Allow user to override the default size for
3141 per-device physically contiguous DMA buffers.
3143 memhp_default_state=online/offline/online_kernel/online_movable
3144 [KNL] Set the initial state for the memory hotplug
3145 onlining policy. If not specified, the default value is
3146 set according to the
3147 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
3149 See Documentation/admin-guide/mm/memory-hotplug.rst.
3151 memmap=exactmap [KNL,X86] Enable setting of an exact
3152 E820 memory map, as specified by the user.
3153 Such memmap=exactmap lines can be constructed based on
3154 BIOS output or other requirements. See the memmap=nn@ss
3157 memmap=nn[KMG]@ss[KMG]
3158 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
3159 Region of memory to be used is from ss to ss+nn.
3160 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
3161 which limits max address to nn[KMG].
3162 Multiple different regions can be specified,
3165 memmap=100M@2G,100M#3G,1G!1024G
3167 memmap=nn[KMG]#ss[KMG]
3168 [KNL,ACPI] Mark specific memory as ACPI data.
3169 Region of memory to be marked is from ss to ss+nn.
3171 memmap=nn[KMG]$ss[KMG]
3172 [KNL,ACPI] Mark specific memory as reserved.
3173 Region of memory to be reserved is from ss to ss+nn.
3174 Example: Exclude memory from 0x18690000-0x1869ffff
3175 memmap=64K$0x18690000
3177 memmap=0x10000$0x18690000
3178 Some bootloaders may need an escape character before '$',
3179 like Grub2, otherwise '$' and the following number
3182 memmap=nn[KMG]!ss[KMG]
3183 [KNL,X86] Mark specific memory as protected.
3184 Region of memory to be used, from ss to ss+nn.
3185 The memory region may be marked as e820 type 12 (0xc)
3186 and is NVDIMM or ADR memory.
3188 memmap=<size>%<offset>-<oldtype>+<newtype>
3189 [KNL,ACPI] Convert memory within the specified region
3190 from <oldtype> to <newtype>. If "-<oldtype>" is left
3191 out, the whole region will be marked as <newtype>,
3192 even if previously unavailable. If "+<newtype>" is left
3193 out, matching memory will be removed. Types are
3194 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
3195 3 = ACPI, 12 = PRAM.
3197 memory_corruption_check=0/1 [X86]
3198 Some BIOSes seem to corrupt the first 64k of
3199 memory when doing things like suspend/resume.
3200 Setting this option will scan the memory
3201 looking for corruption. Enabling this will
3202 both detect corruption and prevent the kernel
3203 from using the memory being corrupted.
3204 However, its intended as a diagnostic tool; if
3205 repeatable BIOS-originated corruption always
3206 affects the same memory, you can use memmap=
3207 to prevent the kernel from using that memory.
3209 memory_corruption_check_size=size [X86]
3210 By default it checks for corruption in the low
3211 64k, making this memory unavailable for normal
3212 use. Use this parameter to scan for
3213 corruption in more or less memory.
3215 memory_corruption_check_period=seconds [X86]
3216 By default it checks for corruption every 60
3217 seconds. Use this parameter to check at some
3218 other rate. 0 disables periodic checking.
3220 memory_hotplug.memmap_on_memory
3221 [KNL,X86,ARM] Boolean flag to enable this feature.
3222 Format: {on | off (default)}
3223 When enabled, runtime hotplugged memory will
3224 allocate its internal metadata (struct pages,
3225 those vmemmap pages cannot be optimized even
3226 if hugetlb_free_vmemmap is enabled) from the
3227 hotadded memory which will allow to hotadd a
3228 lot of memory without requiring additional
3230 This feature is disabled by default because it
3231 has some implication on large (e.g. GB)
3232 allocations in some configurations (e.g. small
3234 The state of the flag can be read in
3235 /sys/module/memory_hotplug/parameters/memmap_on_memory.
3236 Note that even when enabled, there are a few cases where
3237 the feature is not effective.
3239 memtest= [KNL,X86,ARM,M68K,PPC,RISCV] Enable memtest
3241 default : 0 <disable>
3242 Specifies the number of memtest passes to be
3243 performed. Each pass selects another test
3244 pattern from a given set of patterns. Memtest
3245 fills the memory with this pattern, validates
3246 memory contents and reserves bad memory
3247 regions that are detected.
3249 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
3250 Valid arguments: on, off
3251 Default (depends on kernel configuration option):
3252 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
3253 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
3254 mem_encrypt=on: Activate SME
3255 mem_encrypt=off: Do not activate SME
3257 Refer to Documentation/virt/kvm/x86/amd-memory-encryption.rst
3258 for details on when memory encryption can be activated.
3260 mem_sleep_default= [SUSPEND] Default system suspend mode:
3261 s2idle - Suspend-To-Idle
3262 shallow - Power-On Suspend or equivalent (if supported)
3263 deep - Suspend-To-RAM or equivalent (if supported)
3264 See Documentation/admin-guide/pm/sleep-states.rst.
3266 mfgpt_irq= [IA-32] Specify the IRQ to use for the
3267 Multi-Function General Purpose Timers on AMD Geode
3270 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
3271 the BIOS has incorrectly applied a workaround. TinyBIOS
3272 version 0.98 is known to be affected, 0.99 fixes the
3273 problem by letting the user disable the workaround.
3277 min_addr=nn[KMG] [KNL,BOOT,IA-64] All physical memory below this
3278 physical address is ignored.
3280 mini2440= [ARM,HW,KNL]
3281 Format:[0..2][b][c][t]
3283 MINI2440 configuration specification:
3284 0 - The attached screen is the 3.5" TFT
3285 1 - The attached screen is the 7" TFT
3286 2 - The VGA Shield is attached (1024x768)
3287 Leaving out the screen size parameter will not load
3288 the TFT driver, and the framebuffer will be left
3290 b - Enable backlight. The TFT backlight pin will be
3291 linked to the kernel VESA blanking code and a GPIO
3292 LED. This parameter is not necessary when using the
3294 c - Enable the s3c camera interface.
3295 t - Reserved for enabling touchscreen support. The
3296 touchscreen support is not enabled in the mainstream
3297 kernel as of 2.6.30, a preliminary port can be found
3298 in the "bleeding edge" mini2440 support kernel at
3299 https://repo.or.cz/w/linux-2.6/mini2440.git
3302 [X86,PPC,S390,ARM64] Control optional mitigations for
3303 CPU vulnerabilities. This is a set of curated,
3304 arch-independent options, each of which is an
3305 aggregation of existing arch-specific options.
3308 Disable all optional CPU mitigations. This
3309 improves system performance, but it may also
3310 expose users to several CPU vulnerabilities.
3311 Equivalent to: if nokaslr then kpti=0 [ARM64]
3312 gather_data_sampling=off [X86]
3313 kvm.nx_huge_pages=off [X86]
3316 mmio_stale_data=off [X86]
3317 no_entry_flush [PPC]
3318 no_uaccess_flush [PPC]
3321 nospectre_bhb [ARM64]
3322 nospectre_v1 [X86,PPC]
3323 nospectre_v2 [X86,PPC,S390,ARM64]
3325 spec_store_bypass_disable=off [X86,PPC]
3326 spectre_v2_user=off [X86]
3327 srbds=off [X86,INTEL]
3328 ssbd=force-off [ARM64]
3329 tsx_async_abort=off [X86]
3332 This does not have any effect on
3333 kvm.nx_huge_pages when
3334 kvm.nx_huge_pages=force.
3337 Mitigate all CPU vulnerabilities, but leave SMT
3338 enabled, even if it's vulnerable. This is for
3339 users who don't want to be surprised by SMT
3340 getting disabled across kernel upgrades, or who
3341 have other ways of avoiding SMT-based attacks.
3342 Equivalent to: (default behavior)
3345 Mitigate all CPU vulnerabilities, disabling SMT
3346 if needed. This is for users who always want to
3347 be fully mitigated, even if it means losing SMT.
3348 Equivalent to: l1tf=flush,nosmt [X86]
3349 mds=full,nosmt [X86]
3350 tsx_async_abort=full,nosmt [X86]
3351 mmio_stale_data=full,nosmt [X86]
3352 retbleed=auto,nosmt [X86]
3355 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
3356 parameter allows control of the logging verbosity for
3357 the additional memory initialisation checks. A value
3358 of 0 disables mminit logging and a level of 4 will
3359 log everything. Information is printed at KERN_DEBUG
3360 so loglevel=8 may also need to be specified.
3363 [X86,INTEL] Control mitigation for the Processor
3364 MMIO Stale Data vulnerabilities.
3366 Processor MMIO Stale Data is a class of
3367 vulnerabilities that may expose data after an MMIO
3368 operation. Exposed data could originate or end in
3369 the same CPU buffers as affected by MDS and TAA.
3370 Therefore, similar to MDS and TAA, the mitigation
3371 is to clear the affected CPU buffers.
3373 This parameter controls the mitigation. The
3376 full - Enable mitigation on vulnerable CPUs
3378 full,nosmt - Enable mitigation and disable SMT on
3381 off - Unconditionally disable mitigation
3383 On MDS or TAA affected machines,
3384 mmio_stale_data=off can be prevented by an active
3385 MDS or TAA mitigation as these vulnerabilities are
3386 mitigated with the same mechanism so in order to
3387 disable this mitigation, you need to specify
3388 mds=off and tsx_async_abort=off too.
3390 Not specifying this option is equivalent to
3391 mmio_stale_data=full.
3394 Documentation/admin-guide/hw-vuln/processor_mmio_stale_data.rst
3396 <module>.async_probe[=<bool>] [KNL]
3397 If no <bool> value is specified or if the value
3398 specified is not a valid <bool>, enable asynchronous
3399 probe on this module. Otherwise, enable/disable
3400 asynchronous probe on this module as indicated by the
3401 <bool> value. See also: module.async_probe
3403 module.async_probe=<bool>
3404 [KNL] When set to true, modules will use async probing
3405 by default. To enable/disable async probing for a
3406 specific module, use the module specific control that
3407 is documented under <module>.async_probe. When both
3408 module.async_probe and <module>.async_probe are
3409 specified, <module>.async_probe takes precedence for
3410 the specific module.
3412 module.enable_dups_trace
3413 [KNL] When CONFIG_MODULE_DEBUG_AUTOLOAD_DUPS is set,
3414 this means that duplicate request_module() calls will
3415 trigger a WARN_ON() instead of a pr_warn(). Note that
3416 if MODULE_DEBUG_AUTOLOAD_DUPS_TRACE is set, WARN_ON()s
3417 will always be issued and this option does nothing.
3419 [KNL] When CONFIG_MODULE_SIG is set, this means that
3420 modules without (valid) signatures will fail to load.
3421 Note that if CONFIG_MODULE_SIG_FORCE is set, that
3422 is always true, so this option does nothing.
3424 module_blacklist= [KNL] Do not load a comma-separated list of
3425 modules. Useful for debugging problem modules.
3428 [MOUSE] Maximum time between finger touching and
3429 leaving touchpad surface for touch to be considered
3430 a tap and be reported as a left button click (for
3431 touchpads working in absolute mode only).
3433 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
3434 reporting absolute coordinates, such as tablets
3435 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
3436 reporting absolute coordinates, such as tablets
3438 movablecore= [KNL,X86,IA-64,PPC]
3439 Format: nn[KMGTPE] | nn%
3440 This parameter is the complement to kernelcore=, it
3441 specifies the amount of memory used for migratable
3442 allocations. If both kernelcore and movablecore is
3443 specified, then kernelcore will be at *least* the
3444 specified value but may be more. If movablecore on its
3445 own is specified, the administrator must be careful
3446 that the amount of memory usable for all allocations
3449 movable_node [KNL] Boot-time switch to make hotplugable memory
3450 NUMA nodes to be movable. This means that the memory
3451 of such nodes will be usable only for movable
3452 allocations which rules out almost all kernel
3453 allocations. Use with caution!
3455 MTD_Partition= [MTD]
3456 Format: <name>,<region-number>,<size>,<offset>
3458 MTD_Region= [MTD] Format:
3459 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3462 See drivers/mtd/parsers/cmdlinepart.c
3465 ARM/S3C2412 JIVE boot control
3467 See arch/arm/mach-s3c/mach-jive.c
3469 mtouchusb.raw_coordinates=
3470 [HW] Make the MicroTouch USB driver use raw coordinates
3471 ('y', default) or cooked coordinates ('n')
3474 Enable printing debug information related to MTRR
3475 registers at boot time.
3477 mtrr_chunk_size=nn[KMG] [X86]
3478 used for mtrr cleanup. It is largest continuous chunk
3479 that could hold holes aka. UC entries.
3481 mtrr_gran_size=nn[KMG] [X86]
3482 Used for mtrr cleanup. It is granularity of mtrr block.
3484 Large value could prevent small alignment from
3487 mtrr_spare_reg_nr=n [X86]
3489 Range: 0,7 : spare reg number
3491 Used for mtrr cleanup. It is spare mtrr entries number.
3492 Set to 2 or more if your graphical card needs more.
3494 multitce=off [PPC] This parameter disables the use of the pSeries
3495 firmware feature for updating multiple TCE entries
3498 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3500 netdev= [NET] Network devices parameters
3501 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3502 Note that mem_start is often overloaded to mean
3503 something different and driver-specific.
3504 This usage is only documented in each driver source
3507 netpoll.carrier_timeout=
3508 [NET] Specifies amount of time (in seconds) that
3509 netpoll should wait for a carrier. By default netpoll
3513 [NETFILTER] Enable connection tracking flow accounting
3514 0 to disable accounting
3515 1 to enable accounting
3519 [NFS] sets the pathname to the program which is used
3520 to update the NFS client cache entries.
3522 nfs.cache_getent_timeout=
3523 [NFS] sets the timeout after which an attempt to
3524 update a cache entry is deemed to have failed.
3526 nfs.callback_nr_threads=
3527 [NFSv4] set the total number of threads that the
3528 NFS client will assign to service NFSv4 callback
3531 nfs.callback_tcpport=
3532 [NFS] set the TCP port on which the NFSv4 callback
3533 channel should listen.
3536 [NFS] enable 64-bit inode numbers.
3537 If zero, the NFS client will fake up a 32-bit inode
3538 number for the readdir() and stat() syscalls instead
3539 of returning the full 64-bit number.
3540 The default is to return 64-bit inode numbers.
3542 nfs.idmap_cache_timeout=
3543 [NFS] set the maximum lifetime for idmapper cache
3546 nfs.max_session_cb_slots=
3547 [NFSv4.1] Sets the maximum number of session
3548 slots the client will assign to the callback
3549 channel. This determines the maximum number of
3550 callbacks the client will process in parallel for
3551 a particular server.
3553 nfs.max_session_slots=
3554 [NFSv4.1] Sets the maximum number of session slots
3555 the client will attempt to negotiate with the server.
3556 This limits the number of simultaneous RPC requests
3557 that the client can send to the NFSv4.1 server.
3558 Note that there is little point in setting this
3559 value higher than the max_tcp_slot_table_limit.
3561 nfs.nfs4_disable_idmapping=
3562 [NFSv4] When set to the default of '1', this option
3563 ensures that both the RPC level authentication
3564 scheme and the NFS level operations agree to use
3565 numeric uids/gids if the mount is using the
3566 'sec=sys' security flavour. In effect it is
3567 disabling idmapping, which can make migration from
3568 legacy NFSv2/v3 systems to NFSv4 easier.
3569 Servers that do not support this mode of operation
3570 will be autodetected by the client, and it will fall
3571 back to using the idmapper.
3572 To turn off this behaviour, set the value to '0'.
3575 [NFS4] Specify an additional fixed unique ident-
3576 ification string that NFSv4 clients can insert into
3577 their nfs_client_id4 string. This is typically a
3578 UUID that is generated at system install time.
3580 nfs.recover_lost_locks=
3581 [NFSv4] Attempt to recover locks that were lost due
3582 to a lease timeout on the server. Please note that
3583 doing this risks data corruption, since there are
3584 no guarantees that the file will remain unchanged
3585 after the locks are lost.
3586 If you want to enable the kernel legacy behaviour of
3587 attempting to recover these locks, then set this
3589 The default parameter value of '0' causes the kernel
3590 not to attempt recovery of lost locks.
3592 nfs.send_implementation_id=
3593 [NFSv4.1] Send client implementation identification
3594 information in exchange_id requests.
3595 If zero, no implementation identification information
3597 The default is to send the implementation identification
3600 nfs4.layoutstats_timer=
3601 [NFSv4.2] Change the rate at which the kernel sends
3602 layoutstats to the pNFS metadata server.
3604 Setting this to value to 0 causes the kernel to use
3605 whatever value is the default set by the layout
3606 driver. A non-zero value sets the minimum interval
3607 in seconds between layoutstats transmissions.
3609 nfsd.inter_copy_offload_enable=
3610 [NFSv4.2] When set to 1, the server will support
3611 server-to-server copies for which this server is
3612 the destination of the copy.
3614 nfsd.nfs4_disable_idmapping=
3615 [NFSv4] When set to the default of '1', the NFSv4
3616 server will return only numeric uids and gids to
3617 clients using auth_sys, and will accept numeric uids
3618 and gids from such clients. This is intended to ease
3619 migration from NFSv2/v3.
3621 nfsd.nfsd4_ssc_umount_timeout=
3622 [NFSv4.2] When used as the destination of a
3623 server-to-server copy, knfsd temporarily mounts
3624 the source server. It caches the mount in case
3625 it will be needed again, and discards it if not
3626 used for the number of milliseconds specified by
3629 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3630 See Documentation/admin-guide/nfs/nfsroot.rst.
3632 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3633 See Documentation/admin-guide/nfs/nfsroot.rst.
3635 nfsrootdebug [NFS] enable nfsroot debugging messages.
3636 See Documentation/admin-guide/nfs/nfsroot.rst.
3638 nmi_backtrace.backtrace_idle [KNL]
3639 Dump stacks even of idle CPUs in response to an
3640 NMI stack-backtrace request.
3642 nmi_debug= [KNL,SH] Specify one or more actions to take
3643 when a NMI is triggered.
3644 Format: [state][,regs][,debounce][,die]
3646 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3647 Format: [panic,][nopanic,][num]
3649 0 - turn hardlockup detector in nmi_watchdog off
3650 1 - turn hardlockup detector in nmi_watchdog on
3651 When panic is specified, panic when an NMI watchdog
3652 timeout occurs (or 'nopanic' to not panic on an NMI
3653 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3654 To disable both hard and soft lockup detectors,
3655 please see 'nowatchdog'.
3656 This is useful when you use a panic=... timeout and
3657 need the box quickly up again.
3659 These settings can be accessed at runtime via
3660 the nmi_watchdog and hardlockup_panic sysctls.
3662 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3663 emulation library even if a 387 maths coprocessor
3666 no4lvl [RISCV] Disable 4-level and 5-level paging modes. Forces
3667 kernel to use 3-level paging instead.
3669 no5lvl [X86-64,RISCV] Disable 5-level paging mode. Forces
3670 kernel to use 4-level paging instead.
3672 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3673 caches in the slab allocator. Saves per-node memory,
3674 but will impact performance.
3678 noaltinstr [S390] Disables alternative instructions patching
3679 (CPU alternatives feature).
3681 noapic [SMP,APIC] Tells the kernel to not make use of any
3682 IOAPICs that may be present in the system.
3684 noautogroup Disable scheduler automatic task group creation.
3689 [HW] Never suspend the console
3690 Disable suspending of consoles during suspend and
3691 hibernate operations. Once disabled, debugging
3692 messages can reach various consoles while the rest
3693 of the system is being put to sleep (ie, while
3694 debugging driver suspend/resume hooks). This may
3695 not work reliably with all consoles, but is known
3696 to work with serial and VGA consoles.
3697 To facilitate more flexible debugging, we also add
3698 console_suspend, a printk module parameter to control
3699 it. Users could use console_suspend (usually
3700 /sys/module/printk/parameters/console_suspend) to
3701 turn on/off it dynamically.
3704 [KNL] Disable object debugging
3706 nodsp [SH] Disable hardware DSP at boot time.
3708 noefi Disable EFI runtime services support.
3710 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3715 This affects only 32-bit executables.
3716 noexec32=on: enable non-executable mappings (default)
3717 read doesn't imply executable mappings
3718 noexec32=off: disable non-executable mappings
3719 read implies executable mappings
3721 no_file_caps Tells the kernel not to honor file capabilities. The
3722 only way then for a file to be executed with privilege
3723 is to be setuid root or executed by root.
3725 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3727 nofsgsbase [X86] Disables FSGSBASE instructions.
3729 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3730 register save and restore. The kernel will only save
3731 legacy floating-point registers on task switch.
3733 nohalt [IA-64] Tells the kernel not to use the power saving
3734 function PAL_HALT_LIGHT when idle. This increases
3735 power-consumption. On the positive side, it reduces
3736 interrupt wake-up latency, which may improve performance
3737 in certain environments such as networked servers or
3741 Force pointers printed to the console or buffers to be
3742 unhashed. By default, when a pointer is printed via %p
3743 format string, that pointer is "hashed", i.e. obscured
3744 by hashing the pointer value. This is a security feature
3745 that hides actual kernel addresses from unprivileged
3746 users, but it also makes debugging the kernel more
3747 difficult since unequal pointers can no longer be
3748 compared. However, if this command-line option is
3749 specified, then all normal pointers will have their true
3750 value printed. This option should only be specified when
3751 debugging the kernel. Please do not use on production
3754 nohibernate [HIBERNATION] Disable hibernation and resume.
3756 nohlt [ARM,ARM64,MICROBLAZE,MIPS,SH] Forces the kernel to
3757 busy wait in do_idle() and not use the arch_cpu_idle()
3758 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3759 to be effective. This is useful on platforms where the
3760 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3761 correctly or when doing power measurements to evaluate
3762 the impact of the sleep instructions. This is also
3763 useful when using JTAG debugger.
3765 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3767 nohugevmalloc [KNL,X86,PPC,ARM64] Disable kernel huge vmalloc mappings.
3769 nohz= [KNL] Boottime enable/disable dynamic ticks
3770 Valid arguments: on, off
3773 nohz_full= [KNL,BOOT,SMP,ISOL]
3774 The argument is a cpu list, as described above.
3775 In kernels built with CONFIG_NO_HZ_FULL=y, set
3776 the specified list of CPUs whose tick will be stopped
3777 whenever possible. The boot CPU will be forced outside
3778 the range to maintain the timekeeping. Any CPUs
3779 in this list will have their RCU callbacks offloaded,
3780 just as if they had also been called out in the
3781 rcu_nocbs= boot parameter.
3783 Note that this argument takes precedence over
3784 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
3786 noinitrd [RAM] Tells the kernel not to load any configured
3789 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3791 [Deprecated - use intremap=off]
3795 noinvpcid [X86] Disable the INVPCID cpu feature.
3797 noiotrap [SH] Disables trapped I/O port accesses.
3799 noirqdebug [X86-32] Disables the code which attempts to detect and
3800 disable unhandled interrupt sources.
3802 noisapnp [ISAPNP] Disables ISA PnP code.
3804 nojitter [IA-64] Disables jitter checking for ITC timers.
3807 When CONFIG_RANDOMIZE_BASE is set, this disables
3808 kernel and module base offset ASLR (Address Space
3809 Layout Randomization).
3811 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3814 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3816 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3818 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3820 nomca [IA-64] Disable machine check abort handling
3822 nomce [X86-32] Disable Machine Check Exception
3824 nomfgpt [X86-32] Disable Multi-Function General Purpose
3825 Timer usage (for AMD Geode machines).
3827 nomodeset Disable kernel modesetting. Most systems' firmware
3828 sets up a display mode and provides framebuffer memory
3829 for output. With nomodeset, DRM and fbdev drivers will
3830 not load if they could possibly displace the pre-
3831 initialized output. Only the system framebuffer will
3832 be available for use. The respective drivers will not
3833 perform display-mode changes or accelerated rendering.
3835 Useful as error fallback, or for testing and debugging.
3837 nomodule Disable module load
3839 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3840 shutdown the other cpus. Instead use the REBOOT_VECTOR
3843 nopat [X86] Disable PAT (page attribute table extension of
3844 pagetables) support.
3846 nopcid [X86-64] Disable the PCID cpu feature.
3848 nopku [X86] Disable Memory Protection Keys CPU feature found
3852 Equivalent to pti=off
3854 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
3855 Disables the PV optimizations forcing the guest to run
3856 as generic guest with no PV drivers. Currently support
3857 XEN HVM, KVM, HYPER_V and VMWARE guest.
3859 nopvspin [X86,XEN,KVM]
3860 Disables the qspinlock slow path using PV optimizations
3861 which allow the hypervisor to 'idle' the guest on lock
3864 norandmaps Don't use address space randomization. Equivalent to
3865 echo 0 > /proc/sys/kernel/randomize_va_space
3867 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3868 with UP alternatives
3870 noresume [SWSUSP] Disables resume and restores original swap
3875 no-scroll [VGA] Disables scrollback.
3876 This is required for the Braillex ib80-piezo Braille
3877 reader made by F.H. Papenmeier (Germany).
3879 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3882 Disable SMAP (Supervisor Mode Access Prevention)
3883 even if it is supported by processor.
3886 Disable SMEP (Supervisor Mode Execution Prevention)
3887 even if it is supported by processor.
3889 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3890 and disable the IO APIC. legacy for "maxcpus=0".
3892 nosmt [KNL,MIPS,S390] Disable symmetric multithreading (SMT).
3893 Equivalent to smt=1.
3895 [KNL,X86] Disable symmetric multithreading (SMT).
3896 nosmt=force: Force disable SMT, cannot be undone
3897 via the sysfs control file.
3899 nosoftlockup [KNL] Disable the soft-lockup detector.
3901 nospec_store_bypass_disable
3902 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3904 nospectre_bhb [ARM64] Disable all mitigations for Spectre-BHB (branch
3905 history injection) vulnerability. System may allow data leaks
3908 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3909 (bounds check bypass). With this option data leaks are
3910 possible in the system.
3912 nospectre_v2 [X86,PPC_E500,ARM64] Disable all mitigations for
3913 the Spectre variant 2 (indirect branch prediction)
3914 vulnerability. System may allow data leaks with this
3917 no-steal-acc [X86,PV_OPS,ARM64,PPC/PSERIES] Disable paravirtualized
3918 steal time accounting. steal time is computed, but
3919 won't influence scheduler behaviour
3921 nosync [HW,M68K] Disables sync negotiation for all devices.
3923 no_timer_check [X86,APIC] Disables the code which tests for
3924 broken timer IRQ sources.
3927 [PPC] Don't flush the L1-D cache after accessing user data.
3929 novmcoredd [KNL,KDUMP]
3930 Disable device dump. Device dump allows drivers to
3931 append dump data to vmcore so you can collect driver
3932 specified debug info. Drivers can append the data
3933 without any limit and this data is stored in memory,
3934 so this may cause significant memory stress. Disabling
3935 device dump can help save memory but the driver debug
3936 data will be no longer available. This parameter
3937 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3941 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3942 clock and use the default one.
3944 nowatchdog [KNL] Disable both lockup detectors, i.e.
3945 soft-lockup and NMI watchdog (hard-lockup).
3949 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3951 NOTE: this parameter will be ignored on systems with the
3952 LEGACY_XAPIC_DISABLED bit set in the
3953 IA32_XAPIC_DISABLE_STATUS MSR.
3955 noxsave [BUGS=X86] Disables x86 extended register state save
3956 and restore using xsave. The kernel will fallback to
3957 enabling legacy floating-point and sse state.
3959 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3960 register states. The kernel will fall back to use
3961 xsave to save the states. By using this parameter,
3962 performance of saving the states is degraded because
3963 xsave doesn't support modified optimization while
3964 xsaveopt supports it on xsaveopt enabled systems.
3966 noxsaves [X86] Disables xsaves and xrstors used in saving and
3967 restoring x86 extended register state in compacted
3968 form of xsave area. The kernel will fall back to use
3969 xsaveopt and xrstor to save and restore the states
3970 in standard form of xsave area. By using this
3971 parameter, xsave area per process might occupy more
3972 memory on xsaves enabled systems.
3974 nps_mtm_hs_ctr= [KNL,ARC]
3975 This parameter sets the maximum duration, in
3976 cycles, each HW thread of the CTOP can run
3977 without interruptions, before HW switches it.
3978 The actual maximum duration is 16 times this
3980 Format: integer between 1 and 255
3983 nptcg= [IA-64] Override max number of concurrent global TLB
3984 purges which is reported from either PAL_VM_SUMMARY or
3987 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3988 could support. nr_cpus=n : n >= 1 limits the kernel to
3989 support 'n' processors. It could be larger than the
3990 number of already plugged CPU during bootup, later in
3991 runtime you can physically add extra cpu until it reaches
3992 n. So during boot up some boot time memory for per-cpu
3993 variables need be pre-allocated for later physical cpu
3996 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3998 numa=off [KNL, ARM64, PPC, RISCV, SPARC, X86] Disable NUMA, Only
3999 set up a single NUMA node spanning all memory.
4001 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
4003 Allowed values are enable and disable
4005 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
4006 'node', 'default' can be specified
4007 This can be set from sysctl after boot.
4008 See Documentation/admin-guide/sysctl/vm.rst for details.
4010 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
4011 See Documentation/core-api/debugging-via-ohci1394.rst for more
4014 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
4015 Rather than timing out after 20 ms if an EC
4016 command is not properly ACKed, override the length
4017 of the timeout. We have interrupts disabled while
4018 waiting for the ACK, so if this is set too high
4019 interrupts *may* be lost!
4021 omap_mux= [OMAP] Override bootloader pin multiplexing.
4022 Format: <mux_mode0.mode_name=value>...
4023 For example, to override I2C bus2:
4024 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
4026 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
4028 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
4030 boundary - index of last SLC block on Flex-OneNAND.
4031 The remaining blocks are configured as MLC blocks.
4032 lock - Configure if Flex-OneNAND boundary should be locked.
4033 Once locked, the boundary cannot be changed.
4034 1 indicates lock status, 0 indicates unlock status.
4036 oops=panic Always panic on oopses. Default is to just kill the
4037 process, but there is a small probability of
4038 deadlocking the machine.
4039 This will also cause panics on machine check exceptions.
4040 Useful together with panic=30 to trigger a reboot.
4043 [KNL] Boolean flag to control whether the page allocator
4044 should randomize its free lists. The randomization may
4045 be automatically enabled if the kernel detects it is
4046 running on a platform with a direct-mapped memory-side
4047 cache, and this parameter can be used to
4048 override/disable that behavior. The state of the flag
4049 can be read from sysfs at:
4050 /sys/module/page_alloc/parameters/shuffle.
4052 page_owner= [KNL] Boot-time page_owner enabling option.
4053 Storage of the information about who allocated
4054 each page is disabled in default. With this switch,
4056 on: enable the feature
4058 page_poison= [KNL] Boot-time parameter changing the state of
4059 poisoning on the buddy allocator, available with
4060 CONFIG_PAGE_POISONING=y.
4061 off: turn off poisoning (default)
4062 on: turn on poisoning
4064 page_reporting.page_reporting_order=
4065 [KNL] Minimal page reporting order
4067 Adjust the minimal page reporting order. The page
4068 reporting is disabled when it exceeds MAX_ORDER.
4070 panic= [KNL] Kernel behaviour on panic: delay <timeout>
4071 timeout > 0: seconds before rebooting
4072 timeout = 0: wait forever
4073 timeout < 0: reboot immediately
4076 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
4077 Format: <hex>[,nousertaint]
4078 Hexadecimal bitmask representing the set of TAINT flags
4079 that will cause the kernel to panic when add_taint() is
4080 called with any of the flags in this set.
4081 The optional switch "nousertaint" can be utilized to
4082 prevent userspace forced crashes by writing to sysctl
4083 /proc/sys/kernel/tainted any flagset matching with the
4084 bitmask set on panic_on_taint.
4085 See Documentation/admin-guide/tainted-kernels.rst for
4086 extra details on the taint flags that users can pick
4087 to compose the bitmask to assign to panic_on_taint.
4089 panic_on_warn=1 panic() instead of WARN(). Useful to cause kdump
4092 panic_print= Bitmask for printing system info when panic happens.
4093 User can chose combination of the following bits:
4094 bit 0: print all tasks info
4095 bit 1: print system memory info
4096 bit 2: print timer info
4097 bit 3: print locks info if CONFIG_LOCKDEP is on
4098 bit 4: print ftrace buffer
4099 bit 5: print all printk messages in buffer
4100 bit 6: print all CPUs backtrace (if available in the arch)
4101 *Be aware* that this option may print a _lot_ of lines,
4102 so there are risks of losing older messages in the log.
4103 Use this option carefully, maybe worth to setup a
4104 bigger log buffer with "log_buf_len" along with this.
4106 parkbd.port= [HW] Parallel port number the keyboard adapter is
4107 connected to, default is 0.
4109 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
4110 0 for XT, 1 for AT (default is AT).
4113 parport= [HW,PPT] Specify parallel ports. 0 disables.
4114 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
4115 Use 'auto' to force the driver to use any
4116 IRQ/DMA settings detected (the default is to
4117 ignore detected IRQ/DMA settings because of
4118 possible conflicts). You can specify the base
4119 address, IRQ, and DMA settings; IRQ and DMA
4120 should be numbers, or 'auto' (for using detected
4121 settings on that particular port), or 'nofifo'
4122 (to avoid using a FIFO even if it is detected).
4123 Parallel ports are assigned in the order they
4124 are specified on the command line, starting
4127 parport_init_mode= [HW,PPT]
4128 Configure VIA parallel port to operate in
4129 a specific mode. This is necessary on Pegasos
4130 computer where firmware has no options for setting
4131 up parallel port mode and sets it to spp.
4132 Currently this function knows 686a and 8231 chips.
4133 Format: [spp|ps2|epp|ecp|ecpepp]
4135 pata_legacy.all= [HW,LIBATA]
4137 Set to non-zero to probe primary and secondary ISA
4138 port ranges on PCI systems where no PCI PATA device
4139 has been found at either range. Disabled by default.
4141 pata_legacy.autospeed= [HW,LIBATA]
4143 Set to non-zero if a chip is present that snoops speed
4144 changes. Disabled by default.
4146 pata_legacy.ht6560a= [HW,LIBATA]
4148 Set to 1, 2, or 3 for HT 6560A on the primary channel,
4149 the secondary channel, or both channels respectively.
4150 Disabled by default.
4152 pata_legacy.ht6560b= [HW,LIBATA]
4154 Set to 1, 2, or 3 for HT 6560B on the primary channel,
4155 the secondary channel, or both channels respectively.
4156 Disabled by default.
4158 pata_legacy.iordy_mask= [HW,LIBATA]
4160 IORDY enable mask. Set individual bits to allow IORDY
4161 for the respective channel. Bit 0 is for the first
4162 legacy channel handled by this driver, bit 1 is for
4163 the second channel, and so on. The sequence will often
4164 correspond to the primary legacy channel, the secondary
4165 legacy channel, and so on, but the handling of a PCI
4166 bus and the use of other driver options may interfere
4167 with the sequence. By default IORDY is allowed across
4170 pata_legacy.opti82c46x= [HW,LIBATA]
4172 Set to 1, 2, or 3 for Opti 82c611A on the primary
4173 channel, the secondary channel, or both channels
4174 respectively. Disabled by default.
4176 pata_legacy.opti82c611a= [HW,LIBATA]
4178 Set to 1, 2, or 3 for Opti 82c465MV on the primary
4179 channel, the secondary channel, or both channels
4180 respectively. Disabled by default.
4182 pata_legacy.pio_mask= [HW,LIBATA]
4184 PIO mode mask for autospeed devices. Set individual
4185 bits to allow the use of the respective PIO modes.
4186 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
4187 All modes allowed by default.
4189 pata_legacy.probe_all= [HW,LIBATA]
4191 Set to non-zero to probe tertiary and further ISA
4192 port ranges on PCI systems. Disabled by default.
4194 pata_legacy.probe_mask= [HW,LIBATA]
4196 Probe mask for legacy ISA PATA ports. Depending on
4197 platform configuration and the use of other driver
4198 options up to 6 legacy ports are supported: 0x1f0,
4199 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
4200 of individual ports can be disabled by setting the
4201 corresponding bits in the mask to 1. Bit 0 is for
4202 the first port in the list above (0x1f0), and so on.
4203 By default all supported ports are probed.
4205 pata_legacy.qdi= [HW,LIBATA]
4207 Set to non-zero to probe QDI controllers. By default
4208 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
4210 pata_legacy.winbond= [HW,LIBATA]
4212 Set to non-zero to probe Winbond controllers. Use
4213 the standard I/O port (0x130) if 1, otherwise the
4214 value given is the I/O port to use (typically 0x1b0).
4215 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
4218 pata_platform.pio_mask= [HW,LIBATA]
4220 Supported PIO mode mask. Set individual bits to allow
4221 the use of the respective PIO modes. Bit 0 is for
4222 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
4226 Halt all CPUs after the first oops has been printed for
4227 the specified number of seconds. This is to be used if
4228 your oopses keep scrolling off the screen.
4232 pci=option[,option...] [PCI] various PCI subsystem options.
4234 Some options herein operate on a specific device
4235 or a set of devices (<pci_dev>). These are
4236 specified in one of the following formats:
4238 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
4239 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
4241 Note: the first format specifies a PCI
4242 bus/device/function address which may change
4243 if new hardware is inserted, if motherboard
4244 firmware changes, or due to changes caused
4245 by other kernel parameters. If the
4246 domain is left unspecified, it is
4247 taken to be zero. Optionally, a path
4248 to a device through multiple device/function
4249 addresses can be specified after the base
4250 address (this is more robust against
4251 renumbering issues). The second format
4252 selects devices using IDs from the
4253 configuration space which may match multiple
4254 devices in the system.
4256 earlydump dump PCI config space before the kernel
4258 off [X86] don't probe for the PCI bus
4259 bios [X86-32] force use of PCI BIOS, don't access
4260 the hardware directly. Use this if your machine
4261 has a non-standard PCI host bridge.
4262 nobios [X86-32] disallow use of PCI BIOS, only direct
4263 hardware access methods are allowed. Use this
4264 if you experience crashes upon bootup and you
4265 suspect they are caused by the BIOS.
4266 conf1 [X86] Force use of PCI Configuration Access
4267 Mechanism 1 (config address in IO port 0xCF8,
4268 data in IO port 0xCFC, both 32-bit).
4269 conf2 [X86] Force use of PCI Configuration Access
4270 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
4271 the function, IO port 0xCFA, also 8-bit, sets
4272 bus number. The config space is then accessed
4273 through ports 0xC000-0xCFFF).
4274 See http://wiki.osdev.org/PCI for more info
4275 on the configuration access mechanisms.
4276 noaer [PCIE] If the PCIEAER kernel config parameter is
4277 enabled, this kernel boot option can be used to
4278 disable the use of PCIE advanced error reporting.
4279 nodomains [PCI] Disable support for multiple PCI
4280 root domains (aka PCI segments, in ACPI-speak).
4281 nommconf [X86] Disable use of MMCONFIG for PCI
4283 check_enable_amd_mmconf [X86] check for and enable
4284 properly configured MMIO access to PCI
4285 config space on AMD family 10h CPU
4286 nomsi [MSI] If the PCI_MSI kernel config parameter is
4287 enabled, this kernel boot option can be used to
4288 disable the use of MSI interrupts system-wide.
4289 noioapicquirk [APIC] Disable all boot interrupt quirks.
4290 Safety option to keep boot IRQs enabled. This
4291 should never be necessary.
4292 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
4293 primary IO-APIC for bridges that cannot disable
4294 boot IRQs. This fixes a source of spurious IRQs
4295 when the system masks IRQs.
4296 noioapicreroute [APIC] Disable workaround that uses the
4297 boot IRQ equivalent of an IRQ that connects to
4298 a chipset where boot IRQs cannot be disabled.
4299 The opposite of ioapicreroute.
4300 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
4301 routing table. These calls are known to be buggy
4302 on several machines and they hang the machine
4303 when used, but on other computers it's the only
4304 way to get the interrupt routing table. Try
4305 this option if the kernel is unable to allocate
4306 IRQs or discover secondary PCI buses on your
4308 rom [X86] Assign address space to expansion ROMs.
4309 Use with caution as certain devices share
4310 address decoders between ROMs and other
4312 norom [X86] Do not assign address space to
4313 expansion ROMs that do not already have
4314 BIOS assigned address ranges.
4315 nobar [X86] Do not assign address space to the
4316 BARs that weren't assigned by the BIOS.
4317 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
4318 assigned automatically to PCI devices. You can
4319 make the kernel exclude IRQs of your ISA cards
4321 pirqaddr=0xAAAAA [X86] Specify the physical address
4322 of the PIRQ table (normally generated
4323 by the BIOS) if it is outside the
4324 F0000h-100000h range.
4325 lastbus=N [X86] Scan all buses thru bus #N. Can be
4326 useful if the kernel is unable to find your
4327 secondary buses and you want to tell it
4328 explicitly which ones they are.
4329 assign-busses [X86] Always assign all PCI bus
4330 numbers ourselves, overriding
4331 whatever the firmware may have done.
4332 usepirqmask [X86] Honor the possible IRQ mask stored
4333 in the BIOS $PIR table. This is needed on
4334 some systems with broken BIOSes, notably
4335 some HP Pavilion N5400 and Omnibook XE3
4336 notebooks. This will have no effect if ACPI
4337 IRQ routing is enabled.
4338 noacpi [X86] Do not use ACPI for IRQ routing
4339 or for PCI scanning.
4340 use_crs [X86] Use PCI host bridge window information
4341 from ACPI. On BIOSes from 2008 or later, this
4342 is enabled by default. If you need to use this,
4343 please report a bug.
4344 nocrs [X86] Ignore PCI host bridge windows from ACPI.
4345 If you need to use this, please report a bug.
4346 use_e820 [X86] Use E820 reservations to exclude parts of
4347 PCI host bridge windows. This is a workaround
4348 for BIOS defects in host bridge _CRS methods.
4349 If you need to use this, please report a bug to
4350 <linux-pci@vger.kernel.org>.
4351 no_e820 [X86] Ignore E820 reservations for PCI host
4352 bridge windows. This is the default on modern
4353 hardware. If you need to use this, please report
4354 a bug to <linux-pci@vger.kernel.org>.
4355 routeirq Do IRQ routing for all PCI devices.
4356 This is normally done in pci_enable_device(),
4357 so this option is a temporary workaround
4358 for broken drivers that don't call it.
4359 skip_isa_align [X86] do not align io start addr, so can
4360 handle more pci cards
4361 noearly [X86] Don't do any early type 1 scanning.
4362 This might help on some broken boards which
4363 machine check when some devices' config space
4364 is read. But various workarounds are disabled
4365 and some IOMMU drivers will not work.
4366 bfsort Sort PCI devices into breadth-first order.
4367 This sorting is done to get a device
4368 order compatible with older (<= 2.4) kernels.
4369 nobfsort Don't sort PCI devices into breadth-first order.
4370 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
4371 tuning and use the BIOS-configured MPS defaults.
4372 pcie_bus_safe Set every device's MPS to the largest value
4373 supported by all devices below the root complex.
4374 pcie_bus_perf Set device MPS to the largest allowable MPS
4375 based on its parent bus. Also set MRRS (Max
4376 Read Request Size) to the largest supported
4377 value (no larger than the MPS that the device
4378 or bus can support) for best performance.
4379 pcie_bus_peer2peer Set every device's MPS to 128B, which
4380 every device is guaranteed to support. This
4381 configuration allows peer-to-peer DMA between
4382 any pair of devices, possibly at the cost of
4383 reduced performance. This also guarantees
4384 that hot-added devices will work.
4385 cbiosize=nn[KMG] The fixed amount of bus space which is
4386 reserved for the CardBus bridge's IO window.
4387 The default value is 256 bytes.
4388 cbmemsize=nn[KMG] The fixed amount of bus space which is
4389 reserved for the CardBus bridge's memory
4390 window. The default value is 64 megabytes.
4393 [<order of align>@]<pci_dev>[; ...]
4394 Specifies alignment and device to reassign
4395 aligned memory resources. How to
4396 specify the device is described above.
4397 If <order of align> is not specified,
4398 PAGE_SIZE is used as alignment.
4399 A PCI-PCI bridge can be specified if resource
4400 windows need to be expanded.
4401 To specify the alignment for several
4402 instances of a device, the PCI vendor,
4403 device, subvendor, and subdevice may be
4404 specified, e.g., 12@pci:8086:9c22:103c:198f
4405 for 4096-byte alignment.
4406 ecrc= Enable/disable PCIe ECRC (transaction layer
4407 end-to-end CRC checking). Only effective if
4408 OS has native AER control (either granted by
4409 ACPI _OSC or forced via "pcie_ports=native")
4410 bios: Use BIOS/firmware settings. This is the
4414 hpiosize=nn[KMG] The fixed amount of bus space which is
4415 reserved for hotplug bridge's IO window.
4416 Default size is 256 bytes.
4417 hpmmiosize=nn[KMG] The fixed amount of bus space which is
4418 reserved for hotplug bridge's MMIO window.
4419 Default size is 2 megabytes.
4420 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
4421 reserved for hotplug bridge's MMIO_PREF window.
4422 Default size is 2 megabytes.
4423 hpmemsize=nn[KMG] The fixed amount of bus space which is
4424 reserved for hotplug bridge's MMIO and
4426 Default size is 2 megabytes.
4427 hpbussize=nn The minimum amount of additional bus numbers
4428 reserved for buses below a hotplug bridge.
4430 realloc= Enable/disable reallocating PCI bridge resources
4431 if allocations done by BIOS are too small to
4432 accommodate resources required by all child
4434 off: Turn realloc off
4436 realloc same as realloc=on
4437 noari do not use PCIe ARI.
4438 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
4439 do not use PCIe ATS (and IOMMU device IOTLB).
4440 pcie_scan_all Scan all possible PCIe devices. Otherwise we
4441 only look for one device below a PCIe downstream
4443 big_root_window Try to add a big 64bit memory window to the PCIe
4444 root complex on AMD CPUs. Some GFX hardware
4445 can resize a BAR to allow access to all VRAM.
4446 Adding the window is slightly risky (it may
4447 conflict with unreported devices), so this
4449 disable_acs_redir=<pci_dev>[; ...]
4450 Specify one or more PCI devices (in the format
4451 specified above) separated by semicolons.
4452 Each device specified will have the PCI ACS
4453 redirect capabilities forced off which will
4454 allow P2P traffic between devices through
4455 bridges without forcing it upstream. Note:
4456 this removes isolation between devices and
4457 may put more devices in an IOMMU group.
4458 force_floating [S390] Force usage of floating interrupts.
4459 nomio [S390] Do not use MIO instructions.
4460 norid [S390] ignore the RID field and force use of
4461 one PCI domain per PCI function
4463 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
4466 force Enable ASPM even on devices that claim not to support it.
4467 WARNING: Forcing ASPM on may cause system lockups.
4469 pcie_ports= [PCIE] PCIe port services handling:
4470 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
4471 even if the platform doesn't give the OS permission to
4472 use them. This may cause conflicts if the platform
4473 also tries to use these services.
4474 dpc-native Use native PCIe service for DPC only. May
4475 cause conflicts if firmware uses AER or DPC.
4476 compat Disable native PCIe services (PME, AER, DPC, PCIe
4479 pcie_port_pm= [PCIE] PCIe port power management handling:
4480 off Disable power management of all PCIe ports
4481 force Forcibly enable power management of all PCIe ports
4483 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
4484 nomsi Do not use MSI for native PCIe PME signaling (this makes
4485 all PCIe root ports use INTx for all services).
4487 pcmv= [HW,PCMCIA] BadgePAD 4
4491 Keep all power-domains already enabled by bootloader on,
4492 even if no driver has claimed them. This is useful
4493 for debug and development, but should not be
4494 needed on a platform with proper driver support.
4496 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4499 See arch/parisc/kernel/pdc_chassis.c
4501 percpu_alloc= Select which percpu first chunk allocator to use.
4502 Currently supported values are "embed" and "page".
4503 Archs may support subset or none of the selections.
4504 See comments in mm/percpu.c for details on each
4505 allocator. This parameter is primarily for debugging
4506 and performance comparison.
4508 pirq= [SMP,APIC] Manual mp-table setup
4509 See Documentation/arch/x86/i386/IO-APIC.rst.
4511 plip= [PPT,NET] Parallel port network link
4512 Format: { parport<nr> | timid | 0 }
4513 See also Documentation/admin-guide/parport.rst.
4515 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4516 Override pmtimer IOPort with a hex value.
4519 pmu_override= [PPC] Override the PMU.
4520 This option takes over the PMU facility, so it is no
4521 longer usable by perf. Setting this option starts the
4522 PMU counters by setting MMCR0 to 0 (the FC bit is
4523 cleared). If a number is given, then MMCR1 is set to
4524 that number, otherwise (e.g., 'pmu_override=on'), MMCR1
4527 pm_debug_messages [SUSPEND,KNL]
4528 Enable suspend/resume debug messages during boot up.
4531 Enable PNP debug messages (depends on the
4532 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4533 via /sys/module/pnp/parameters/debug. We always show
4534 current resource usage; turning this on also shows
4535 possible settings and some assignment information.
4541 { on | off | curr | res | no-curr | no-res }
4544 [ISAPNP] Exclude IRQs for the autoconfiguration
4547 [ISAPNP] Exclude DMAs for the autoconfiguration
4549 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4550 Ranges are in pairs (I/O port base and size).
4553 [ISAPNP] Exclude memory regions for the
4555 Ranges are in pairs (memory base and size).
4557 ports= [IP_VS_FTP] IPVS ftp helper module
4559 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4561 Format: <port>,<port>....
4563 powersave=off [PPC] This option disables power saving features.
4564 It specifically disables cpuidle and sets the
4565 platform machine description specific power_save
4566 function to NULL. On Idle the CPU just reduces
4569 ppc_strict_facility_enable
4570 [PPC] This option catches any kernel floating point,
4571 Altivec, VSX and SPE outside of regions specifically
4572 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4573 There is some performance impact when enabling this.
4577 Disable Hardware Transactional Memory
4580 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4581 none - Limited to cond_resched() calls
4582 voluntary - Limited to cond_resched() and might_sleep() calls
4583 full - Any section that isn't explicitly preempt disabled
4584 can be preempted anytime.
4586 print-fatal-signals=
4587 [KNL] debug: print fatal signals
4589 If enabled, warn about various signal handling
4590 related application anomalies: too many signals,
4591 too many POSIX.1 timers, fatal signals causing a
4594 If you hit the warning due to signal overflow,
4595 you might want to try "ulimit -i unlimited".
4599 printk.always_kmsg_dump=
4600 Trigger kmsg_dump for cases other than kernel oops or
4602 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4605 printk.console_no_auto_verbose=
4606 Disable console loglevel raise on oops, panic
4607 or lockdep-detected issues (only if lock debug is on).
4608 With an exception to setups with low baudrate on
4609 serial console, keeping this 0 is a good choice
4610 in order to provide more debug information.
4612 default: 0 (auto_verbose is enabled)
4614 printk.devkmsg={on,off,ratelimit}
4615 Control writing to /dev/kmsg.
4616 on - unlimited logging to /dev/kmsg from userspace
4617 off - logging to /dev/kmsg disabled
4618 ratelimit - ratelimit the logging
4621 printk.time= Show timing data prefixed to each printk message line
4622 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4624 processor.max_cstate= [HW,ACPI]
4625 Limit processor to maximum C-state
4626 max_cstate=9 overrides any DMI blacklist limit.
4628 processor.nocst [HW,ACPI]
4629 Ignore the _CST method to determine C-states,
4630 instead using the legacy FADT method
4632 profile= [KNL] Enable kernel profiling via /proc/profile
4633 Format: [<profiletype>,]<number>
4634 Param: <profiletype>: "schedule", "sleep", or "kvm"
4635 [defaults to kernel profiling]
4636 Param: "schedule" - profile schedule points.
4637 Param: "sleep" - profile D-state sleeping (millisecs).
4638 Requires CONFIG_SCHEDSTATS
4639 Param: "kvm" - profile VM exits.
4640 Param: <number> - step/bucket size as a power of 2 for
4641 statistical time based profiling.
4643 prompt_ramdisk= [RAM] [Deprecated]
4645 prot_virt= [S390] enable hosting protected virtual machines
4646 isolated from the hypervisor (if hardware supports
4650 psi= [KNL] Enable or disable pressure stall information
4654 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4655 probe for; one of (bare|imps|exps|lifebook|any).
4656 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4658 psmouse.resetafter= [HW,MOUSE]
4659 Try to reset the device after so many bad packets
4662 [HW,MOUSE] Set desired mouse resolution, in dpi.
4663 psmouse.smartscroll=
4664 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4665 0 = disabled, 1 = enabled (default).
4667 pstore.backend= Specify the name of the pstore backend to use
4669 pti= [X86-64] Control Page Table Isolation of user and
4670 kernel address spaces. Disabling this feature
4671 removes hardening, but improves performance of
4672 system calls and interrupts.
4674 on - unconditionally enable
4675 off - unconditionally disable
4676 auto - kernel detects whether your CPU model is
4677 vulnerable to issues that PTI mitigates
4679 Not specifying this option is equivalent to pti=auto.
4682 [KNL] Number of legacy pty's. Overwrites compiled-in
4685 quiet [KNL] Disable most log messages
4689 radix_hcall_invalidate=on [PPC/PSERIES]
4690 Disable RADIX GTSE feature and use hcall for TLB
4694 See Documentation/admin-guide/md.rst.
4696 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4697 See Documentation/admin-guide/blockdev/ramdisk.rst.
4699 ramdisk_start= [RAM] RAM disk image start address
4701 random.trust_cpu=off
4702 [KNL] Disable trusting the use of the CPU's
4703 random number generator (if available) to
4704 initialize the kernel's RNG.
4706 random.trust_bootloader=off
4707 [KNL] Disable trusting the use of the a seed
4708 passed by the bootloader (if available) to
4709 initialize the kernel's RNG.
4711 randomize_kstack_offset=
4712 [KNL] Enable or disable kernel stack offset
4713 randomization, which provides roughly 5 bits of
4714 entropy, frustrating memory corruption attacks
4715 that depend on stack address determinism or
4716 cross-syscall address exposures. This is only
4717 available on architectures that have defined
4718 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4719 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4720 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4722 ras=option[,option,...] [KNL] RAS-specific options
4725 Disable the Correctable Errors Collector,
4726 see CONFIG_RAS_CEC help text.
4728 rcu_nocbs[=cpu-list]
4729 [KNL] The optional argument is a cpu list,
4732 In kernels built with CONFIG_RCU_NOCB_CPU=y,
4733 enable the no-callback CPU mode, which prevents
4734 such CPUs' callbacks from being invoked in
4735 softirq context. Invocation of such CPUs' RCU
4736 callbacks will instead be offloaded to "rcuox/N"
4737 kthreads created for that purpose, where "x" is
4738 "p" for RCU-preempt, "s" for RCU-sched, and "g"
4739 for the kthreads that mediate grace periods; and
4740 "N" is the CPU number. This reduces OS jitter on
4741 the offloaded CPUs, which can be useful for HPC
4742 and real-time workloads. It can also improve
4743 energy efficiency for asymmetric multiprocessors.
4745 If a cpulist is passed as an argument, the specified
4746 list of CPUs is set to no-callback mode from boot.
4748 Otherwise, if the '=' sign and the cpulist
4749 arguments are omitted, no CPU will be set to
4750 no-callback mode from boot but the mode may be
4751 toggled at runtime via cpusets.
4753 Note that this argument takes precedence over
4754 the CONFIG_RCU_NOCB_CPU_DEFAULT_ALL option.
4757 Rather than requiring that offloaded CPUs
4758 (specified by rcu_nocbs= above) explicitly
4759 awaken the corresponding "rcuoN" kthreads,
4760 make these kthreads poll for callbacks.
4761 This improves the real-time response for the
4762 offloaded CPUs by relieving them of the need to
4763 wake up the corresponding kthread, but degrades
4764 energy efficiency by requiring that the kthreads
4765 periodically wake up to do the polling.
4767 rcutree.blimit= [KNL]
4768 Set maximum number of finished RCU callbacks to
4769 process in one batch.
4771 rcutree.dump_tree= [KNL]
4772 Dump the structure of the rcu_node combining tree
4773 out at early boot. This is used for diagnostic
4774 purposes, to verify correct tree setup.
4776 rcutree.gp_cleanup_delay= [KNL]
4777 Set the number of jiffies to delay each step of
4778 RCU grace-period cleanup.
4780 rcutree.gp_init_delay= [KNL]
4781 Set the number of jiffies to delay each step of
4782 RCU grace-period initialization.
4784 rcutree.gp_preinit_delay= [KNL]
4785 Set the number of jiffies to delay each step of
4786 RCU grace-period pre-initialization, that is,
4787 the propagation of recent CPU-hotplug changes up
4788 the rcu_node combining tree.
4790 rcutree.jiffies_till_first_fqs= [KNL]
4791 Set delay from grace-period initialization to
4792 first attempt to force quiescent states.
4793 Units are jiffies, minimum value is zero,
4794 and maximum value is HZ.
4796 rcutree.jiffies_till_next_fqs= [KNL]
4797 Set delay between subsequent attempts to force
4798 quiescent states. Units are jiffies, minimum
4799 value is one, and maximum value is HZ.
4801 rcutree.jiffies_till_sched_qs= [KNL]
4802 Set required age in jiffies for a
4803 given grace period before RCU starts
4804 soliciting quiescent-state help from
4805 rcu_note_context_switch() and cond_resched().
4806 If not specified, the kernel will calculate
4807 a value based on the most recent settings
4808 of rcutree.jiffies_till_first_fqs
4809 and rcutree.jiffies_till_next_fqs.
4810 This calculated value may be viewed in
4811 rcutree.jiffies_to_sched_qs. Any attempt to set
4812 rcutree.jiffies_to_sched_qs will be cheerfully
4815 rcutree.kthread_prio= [KNL,BOOT]
4816 Set the SCHED_FIFO priority of the RCU per-CPU
4817 kthreads (rcuc/N). This value is also used for
4818 the priority of the RCU boost threads (rcub/N)
4819 and for the RCU grace-period kthreads (rcu_bh,
4820 rcu_preempt, and rcu_sched). If RCU_BOOST is
4821 set, valid values are 1-99 and the default is 1
4822 (the least-favored priority). Otherwise, when
4823 RCU_BOOST is not set, valid values are 0-99 and
4824 the default is zero (non-realtime operation).
4825 When RCU_NOCB_CPU is set, also adjust the
4826 priority of NOCB callback kthreads.
4828 rcutree.nocb_nobypass_lim_per_jiffy= [KNL]
4829 On callback-offloaded (rcu_nocbs) CPUs,
4830 RCU reduces the lock contention that would
4831 otherwise be caused by callback floods through
4832 use of the ->nocb_bypass list. However, in the
4833 common non-flooded case, RCU queues directly to
4834 the main ->cblist in order to avoid the extra
4835 overhead of the ->nocb_bypass list and its lock.
4836 But if there are too many callbacks queued during
4837 a single jiffy, RCU pre-queues the callbacks into
4838 the ->nocb_bypass queue. The definition of "too
4839 many" is supplied by this kernel boot parameter.
4841 rcutree.qhimark= [KNL]
4842 Set threshold of queued RCU callbacks beyond which
4843 batch limiting is disabled.
4845 rcutree.qlowmark= [KNL]
4846 Set threshold of queued RCU callbacks below which
4847 batch limiting is re-enabled.
4849 rcutree.qovld= [KNL]
4850 Set threshold of queued RCU callbacks beyond which
4851 RCU's force-quiescent-state scan will aggressively
4852 enlist help from cond_resched() and sched IPIs to
4853 help CPUs more quickly reach quiescent states.
4854 Set to less than zero to make this be set based
4855 on rcutree.qhimark at boot time and to zero to
4856 disable more aggressive help enlistment.
4858 rcutree.rcu_delay_page_cache_fill_msec= [KNL]
4859 Set the page-cache refill delay (in milliseconds)
4860 in response to low-memory conditions. The range
4861 of permitted values is in the range 0:100000.
4863 rcutree.rcu_divisor= [KNL]
4864 Set the shift-right count to use to compute
4865 the callback-invocation batch limit bl from
4866 the number of callbacks queued on this CPU.
4867 The result will be bounded below by the value of
4868 the rcutree.blimit kernel parameter. Every bl
4869 callbacks, the softirq handler will exit in
4870 order to allow the CPU to do other work.
4872 Please note that this callback-invocation batch
4873 limit applies only to non-offloaded callback
4874 invocation. Offloaded callbacks are instead
4875 invoked in the context of an rcuoc kthread, which
4876 scheduler will preempt as it does any other task.
4878 rcutree.rcu_fanout_exact= [KNL]
4879 Disable autobalancing of the rcu_node combining
4880 tree. This is used by rcutorture, and might
4881 possibly be useful for architectures having high
4882 cache-to-cache transfer latencies.
4884 rcutree.rcu_fanout_leaf= [KNL]
4885 Change the number of CPUs assigned to each
4886 leaf rcu_node structure. Useful for very
4887 large systems, which will choose the value 64,
4888 and for NUMA systems with large remote-access
4889 latencies, which will choose a value aligned
4890 with the appropriate hardware boundaries.
4892 rcutree.rcu_min_cached_objs= [KNL]
4893 Minimum number of objects which are cached and
4894 maintained per one CPU. Object size is equal
4895 to PAGE_SIZE. The cache allows to reduce the
4896 pressure to page allocator, also it makes the
4897 whole algorithm to behave better in low memory
4900 rcutree.rcu_nocb_gp_stride= [KNL]
4901 Set the number of NOCB callback kthreads in
4902 each group, which defaults to the square root
4903 of the number of CPUs. Larger numbers reduce
4904 the wakeup overhead on the global grace-period
4905 kthread, but increases that same overhead on
4906 each group's NOCB grace-period kthread.
4908 rcutree.rcu_kick_kthreads= [KNL]
4909 Cause the grace-period kthread to get an extra
4910 wake_up() if it sleeps three times longer than
4911 it should at force-quiescent-state time.
4912 This wake_up() will be accompanied by a
4913 WARN_ONCE() splat and an ftrace_dump().
4915 rcutree.rcu_resched_ns= [KNL]
4916 Limit the time spend invoking a batch of RCU
4917 callbacks to the specified number of nanoseconds.
4918 By default, this limit is checked only once
4919 every 32 callbacks in order to limit the pain
4920 inflicted by local_clock() overhead.
4922 rcutree.rcu_unlock_delay= [KNL]
4923 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4924 this specifies an rcu_read_unlock()-time delay
4925 in microseconds. This defaults to zero.
4926 Larger delays increase the probability of
4927 catching RCU pointer leaks, that is, buggy use
4928 of RCU-protected pointers after the relevant
4929 rcu_read_unlock() has completed.
4931 rcutree.sysrq_rcu= [KNL]
4932 Commandeer a sysrq key to dump out Tree RCU's
4933 rcu_node tree with an eye towards determining
4934 why a new grace period has not yet started.
4936 rcutree.use_softirq= [KNL]
4937 If set to zero, move all RCU_SOFTIRQ processing to
4938 per-CPU rcuc kthreads. Defaults to a non-zero
4939 value, meaning that RCU_SOFTIRQ is used by default.
4940 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4942 But note that CONFIG_PREEMPT_RT=y kernels disable
4943 this kernel boot parameter, forcibly setting it
4946 rcuscale.gp_async= [KNL]
4947 Measure performance of asynchronous
4948 grace-period primitives such as call_rcu().
4950 rcuscale.gp_async_max= [KNL]
4951 Specify the maximum number of outstanding
4952 callbacks per writer thread. When a writer
4953 thread exceeds this limit, it invokes the
4954 corresponding flavor of rcu_barrier() to allow
4955 previously posted callbacks to drain.
4957 rcuscale.gp_exp= [KNL]
4958 Measure performance of expedited synchronous
4959 grace-period primitives.
4961 rcuscale.holdoff= [KNL]
4962 Set test-start holdoff period. The purpose of
4963 this parameter is to delay the start of the
4964 test until boot completes in order to avoid
4967 rcuscale.kfree_by_call_rcu= [KNL]
4968 In kernels built with CONFIG_RCU_LAZY=y, test
4969 call_rcu() instead of kfree_rcu().
4971 rcuscale.kfree_mult= [KNL]
4972 Instead of allocating an object of size kfree_obj,
4973 allocate one of kfree_mult * sizeof(kfree_obj).
4976 rcuscale.kfree_rcu_test= [KNL]
4977 Set to measure performance of kfree_rcu() flooding.
4979 rcuscale.kfree_rcu_test_double= [KNL]
4980 Test the double-argument variant of kfree_rcu().
4981 If this parameter has the same value as
4982 rcuscale.kfree_rcu_test_single, both the single-
4983 and double-argument variants are tested.
4985 rcuscale.kfree_rcu_test_single= [KNL]
4986 Test the single-argument variant of kfree_rcu().
4987 If this parameter has the same value as
4988 rcuscale.kfree_rcu_test_double, both the single-
4989 and double-argument variants are tested.
4991 rcuscale.kfree_nthreads= [KNL]
4992 The number of threads running loops of kfree_rcu().
4994 rcuscale.kfree_alloc_num= [KNL]
4995 Number of allocations and frees done in an iteration.
4997 rcuscale.kfree_loops= [KNL]
4998 Number of loops doing rcuscale.kfree_alloc_num number
4999 of allocations and frees.
5001 rcuscale.minruntime= [KNL]
5002 Set the minimum test run time in seconds. This
5003 does not affect the data-collection interval,
5004 but instead allows better measurement of things
5005 like CPU consumption.
5007 rcuscale.nreaders= [KNL]
5008 Set number of RCU readers. The value -1 selects
5009 N, where N is the number of CPUs. A value
5010 "n" less than -1 selects N-n+1, where N is again
5011 the number of CPUs. For example, -2 selects N
5012 (the number of CPUs), -3 selects N+1, and so on.
5013 A value of "n" less than or equal to -N selects
5016 rcuscale.nwriters= [KNL]
5017 Set number of RCU writers. The values operate
5018 the same as for rcuscale.nreaders.
5019 N, where N is the number of CPUs
5021 rcuscale.scale_type= [KNL]
5022 Specify the RCU implementation to test.
5024 rcuscale.shutdown= [KNL]
5025 Shut the system down after performance tests
5026 complete. This is useful for hands-off automated
5029 rcuscale.verbose= [KNL]
5030 Enable additional printk() statements.
5032 rcuscale.writer_holdoff= [KNL]
5033 Write-side holdoff between grace periods,
5034 in microseconds. The default of zero says
5037 rcuscale.writer_holdoff_jiffies= [KNL]
5038 Additional write-side holdoff between grace
5039 periods, but in jiffies. The default of zero
5042 rcutorture.fqs_duration= [KNL]
5043 Set duration of force_quiescent_state bursts
5046 rcutorture.fqs_holdoff= [KNL]
5047 Set holdoff time within force_quiescent_state bursts
5050 rcutorture.fqs_stutter= [KNL]
5051 Set wait time between force_quiescent_state bursts
5054 rcutorture.fwd_progress= [KNL]
5055 Specifies the number of kthreads to be used
5056 for RCU grace-period forward-progress testing
5057 for the types of RCU supporting this notion.
5058 Defaults to 1 kthread, values less than zero or
5059 greater than the number of CPUs cause the number
5062 rcutorture.fwd_progress_div= [KNL]
5063 Specify the fraction of a CPU-stall-warning
5064 period to do tight-loop forward-progress testing.
5066 rcutorture.fwd_progress_holdoff= [KNL]
5067 Number of seconds to wait between successive
5068 forward-progress tests.
5070 rcutorture.fwd_progress_need_resched= [KNL]
5071 Enclose cond_resched() calls within checks for
5072 need_resched() during tight-loop forward-progress
5075 rcutorture.gp_cond= [KNL]
5076 Use conditional/asynchronous update-side
5077 primitives, if available.
5079 rcutorture.gp_exp= [KNL]
5080 Use expedited update-side primitives, if available.
5082 rcutorture.gp_normal= [KNL]
5083 Use normal (non-expedited) asynchronous
5084 update-side primitives, if available.
5086 rcutorture.gp_sync= [KNL]
5087 Use normal (non-expedited) synchronous
5088 update-side primitives, if available. If all
5089 of rcutorture.gp_cond=, rcutorture.gp_exp=,
5090 rcutorture.gp_normal=, and rcutorture.gp_sync=
5091 are zero, rcutorture acts as if is interpreted
5092 they are all non-zero.
5094 rcutorture.irqreader= [KNL]
5095 Run RCU readers from irq handlers, or, more
5096 accurately, from a timer handler. Not all RCU
5097 flavors take kindly to this sort of thing.
5099 rcutorture.leakpointer= [KNL]
5100 Leak an RCU-protected pointer out of the reader.
5101 This can of course result in splats, and is
5102 intended to test the ability of things like
5103 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
5106 rcutorture.n_barrier_cbs= [KNL]
5107 Set callbacks/threads for rcu_barrier() testing.
5109 rcutorture.nfakewriters= [KNL]
5110 Set number of concurrent RCU writers. These just
5111 stress RCU, they don't participate in the actual
5112 test, hence the "fake".
5114 rcutorture.nocbs_nthreads= [KNL]
5115 Set number of RCU callback-offload togglers.
5116 Zero (the default) disables toggling.
5118 rcutorture.nocbs_toggle= [KNL]
5119 Set the delay in milliseconds between successive
5120 callback-offload toggling attempts.
5122 rcutorture.nreaders= [KNL]
5123 Set number of RCU readers. The value -1 selects
5124 N-1, where N is the number of CPUs. A value
5125 "n" less than -1 selects N-n-2, where N is again
5126 the number of CPUs. For example, -2 selects N
5127 (the number of CPUs), -3 selects N+1, and so on.
5129 rcutorture.object_debug= [KNL]
5130 Enable debug-object double-call_rcu() testing.
5132 rcutorture.onoff_holdoff= [KNL]
5133 Set time (s) after boot for CPU-hotplug testing.
5135 rcutorture.onoff_interval= [KNL]
5136 Set time (jiffies) between CPU-hotplug operations,
5137 or zero to disable CPU-hotplug testing.
5139 rcutorture.read_exit= [KNL]
5140 Set the number of read-then-exit kthreads used
5141 to test the interaction of RCU updaters and
5142 task-exit processing.
5144 rcutorture.read_exit_burst= [KNL]
5145 The number of times in a given read-then-exit
5146 episode that a set of read-then-exit kthreads
5149 rcutorture.read_exit_delay= [KNL]
5150 The delay, in seconds, between successive
5151 read-then-exit testing episodes.
5153 rcutorture.shuffle_interval= [KNL]
5154 Set task-shuffle interval (s). Shuffling tasks
5155 allows some CPUs to go into dyntick-idle mode
5156 during the rcutorture test.
5158 rcutorture.shutdown_secs= [KNL]
5159 Set time (s) after boot system shutdown. This
5160 is useful for hands-off automated testing.
5162 rcutorture.stall_cpu= [KNL]
5163 Duration of CPU stall (s) to test RCU CPU stall
5164 warnings, zero to disable.
5166 rcutorture.stall_cpu_block= [KNL]
5167 Sleep while stalling if set. This will result
5168 in warnings from preemptible RCU in addition to
5169 any other stall-related activity. Note that
5170 in kernels built with CONFIG_PREEMPTION=n and
5171 CONFIG_PREEMPT_COUNT=y, this parameter will
5172 cause the CPU to pass through a quiescent state.
5173 Given CONFIG_PREEMPTION=n, this will suppress
5174 RCU CPU stall warnings, but will instead result
5175 in scheduling-while-atomic splats.
5177 Use of this module parameter results in splats.
5180 rcutorture.stall_cpu_holdoff= [KNL]
5181 Time to wait (s) after boot before inducing stall.
5183 rcutorture.stall_cpu_irqsoff= [KNL]
5184 Disable interrupts while stalling if set.
5186 rcutorture.stall_gp_kthread= [KNL]
5187 Duration (s) of forced sleep within RCU
5188 grace-period kthread to test RCU CPU stall
5189 warnings, zero to disable. If both stall_cpu
5190 and stall_gp_kthread are specified, the
5191 kthread is starved first, then the CPU.
5193 rcutorture.stat_interval= [KNL]
5194 Time (s) between statistics printk()s.
5196 rcutorture.stutter= [KNL]
5197 Time (s) to stutter testing, for example, specifying
5198 five seconds causes the test to run for five seconds,
5199 wait for five seconds, and so on. This tests RCU's
5200 ability to transition abruptly to and from idle.
5202 rcutorture.test_boost= [KNL]
5203 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
5204 "Maybe" means test if the RCU implementation
5205 under test support RCU priority boosting.
5207 rcutorture.test_boost_duration= [KNL]
5208 Duration (s) of each individual boost test.
5210 rcutorture.test_boost_interval= [KNL]
5211 Interval (s) between each boost test.
5213 rcutorture.test_no_idle_hz= [KNL]
5214 Test RCU's dyntick-idle handling. See also the
5215 rcutorture.shuffle_interval parameter.
5217 rcutorture.torture_type= [KNL]
5218 Specify the RCU implementation to test.
5220 rcutorture.verbose= [KNL]
5221 Enable additional printk() statements.
5223 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
5224 Dump ftrace buffer after reporting RCU CPU
5227 rcupdate.rcu_cpu_stall_suppress= [KNL]
5228 Suppress RCU CPU stall warning messages.
5230 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
5231 Suppress RCU CPU stall warning messages and
5232 rcutorture writer stall warnings that occur
5233 during early boot, that is, during the time
5234 before the init task is spawned.
5236 rcupdate.rcu_cpu_stall_timeout= [KNL]
5237 Set timeout for RCU CPU stall warning messages.
5238 The value is in seconds and the maximum allowed
5239 value is 300 seconds.
5241 rcupdate.rcu_exp_cpu_stall_timeout= [KNL]
5242 Set timeout for expedited RCU CPU stall warning
5243 messages. The value is in milliseconds
5244 and the maximum allowed value is 21000
5245 milliseconds. Please note that this value is
5246 adjusted to an arch timer tick resolution.
5247 Setting this to zero causes the value from
5248 rcupdate.rcu_cpu_stall_timeout to be used (after
5249 conversion from seconds to milliseconds).
5251 rcupdate.rcu_cpu_stall_cputime= [KNL]
5252 Provide statistics on the cputime and count of
5253 interrupts and tasks during the sampling period. For
5254 multiple continuous RCU stalls, all sampling periods
5255 begin at half of the first RCU stall timeout.
5257 rcupdate.rcu_exp_stall_task_details= [KNL]
5258 Print stack dumps of any tasks blocking the
5259 current expedited RCU grace period during an
5260 expedited RCU CPU stall warning.
5262 rcupdate.rcu_expedited= [KNL]
5263 Use expedited grace-period primitives, for
5264 example, synchronize_rcu_expedited() instead
5265 of synchronize_rcu(). This reduces latency,
5266 but can increase CPU utilization, degrade
5267 real-time latency, and degrade energy efficiency.
5268 No effect on CONFIG_TINY_RCU kernels.
5270 rcupdate.rcu_normal= [KNL]
5271 Use only normal grace-period primitives,
5272 for example, synchronize_rcu() instead of
5273 synchronize_rcu_expedited(). This improves
5274 real-time latency, CPU utilization, and
5275 energy efficiency, but can expose users to
5276 increased grace-period latency. This parameter
5277 overrides rcupdate.rcu_expedited. No effect on
5278 CONFIG_TINY_RCU kernels.
5280 rcupdate.rcu_normal_after_boot= [KNL]
5281 Once boot has completed (that is, after
5282 rcu_end_inkernel_boot() has been invoked), use
5283 only normal grace-period primitives. No effect
5284 on CONFIG_TINY_RCU kernels.
5286 But note that CONFIG_PREEMPT_RT=y kernels enables
5287 this kernel boot parameter, forcibly setting
5288 it to the value one, that is, converting any
5289 post-boot attempt at an expedited RCU grace
5290 period to instead use normal non-expedited
5291 grace-period processing.
5293 rcupdate.rcu_task_collapse_lim= [KNL]
5294 Set the maximum number of callbacks present
5295 at the beginning of a grace period that allows
5296 the RCU Tasks flavors to collapse back to using
5297 a single callback queue. This switching only
5298 occurs when rcupdate.rcu_task_enqueue_lim is
5299 set to the default value of -1.
5301 rcupdate.rcu_task_contend_lim= [KNL]
5302 Set the minimum number of callback-queuing-time
5303 lock-contention events per jiffy required to
5304 cause the RCU Tasks flavors to switch to per-CPU
5305 callback queuing. This switching only occurs
5306 when rcupdate.rcu_task_enqueue_lim is set to
5307 the default value of -1.
5309 rcupdate.rcu_task_enqueue_lim= [KNL]
5310 Set the number of callback queues to use for the
5311 RCU Tasks family of RCU flavors. The default
5312 of -1 allows this to be automatically (and
5313 dynamically) adjusted. This parameter is intended
5316 rcupdate.rcu_task_ipi_delay= [KNL]
5317 Set time in jiffies during which RCU tasks will
5318 avoid sending IPIs, starting with the beginning
5319 of a given grace period. Setting a large
5320 number avoids disturbing real-time workloads,
5321 but lengthens grace periods.
5323 rcupdate.rcu_task_lazy_lim= [KNL]
5324 Number of callbacks on a given CPU that will
5325 cancel laziness on that CPU. Use -1 to disable
5326 cancellation of laziness, but be advised that
5327 doing so increases the danger of OOM due to
5330 rcupdate.rcu_task_stall_info= [KNL]
5331 Set initial timeout in jiffies for RCU task stall
5332 informational messages, which give some indication
5333 of the problem for those not patient enough to
5334 wait for ten minutes. Informational messages are
5335 only printed prior to the stall-warning message
5336 for a given grace period. Disable with a value
5337 less than or equal to zero. Defaults to ten
5338 seconds. A change in value does not take effect
5339 until the beginning of the next grace period.
5341 rcupdate.rcu_task_stall_info_mult= [KNL]
5342 Multiplier for time interval between successive
5343 RCU task stall informational messages for a given
5344 RCU tasks grace period. This value is clamped
5345 to one through ten, inclusive. It defaults to
5346 the value three, so that the first informational
5347 message is printed 10 seconds into the grace
5348 period, the second at 40 seconds, the third at
5349 160 seconds, and then the stall warning at 600
5350 seconds would prevent a fourth at 640 seconds.
5352 rcupdate.rcu_task_stall_timeout= [KNL]
5353 Set timeout in jiffies for RCU task stall
5354 warning messages. Disable with a value less
5355 than or equal to zero. Defaults to ten minutes.
5356 A change in value does not take effect until
5357 the beginning of the next grace period.
5359 rcupdate.rcu_tasks_lazy_ms= [KNL]
5360 Set timeout in milliseconds RCU Tasks asynchronous
5361 callback batching for call_rcu_tasks().
5362 A negative value will take the default. A value
5363 of zero will disable batching. Batching is
5364 always disabled for synchronize_rcu_tasks().
5366 rcupdate.rcu_tasks_rude_lazy_ms= [KNL]
5367 Set timeout in milliseconds RCU Tasks
5368 Rude asynchronous callback batching for
5369 call_rcu_tasks_rude(). A negative value
5370 will take the default. A value of zero will
5371 disable batching. Batching is always disabled
5372 for synchronize_rcu_tasks_rude().
5374 rcupdate.rcu_tasks_trace_lazy_ms= [KNL]
5375 Set timeout in milliseconds RCU Tasks
5376 Trace asynchronous callback batching for
5377 call_rcu_tasks_trace(). A negative value
5378 will take the default. A value of zero will
5379 disable batching. Batching is always disabled
5380 for synchronize_rcu_tasks_trace().
5382 rcupdate.rcu_self_test= [KNL]
5383 Run the RCU early boot self tests
5387 Run specified binary instead of /init from the ramdisk,
5388 used for early userspace startup. See initrd.
5391 force - Override the decision by the kernel to hide the
5392 advertisement of RDRAND support (this affects
5393 certain AMD processors because of buggy BIOS
5394 support, specifically around the suspend/resume
5398 Turn on/off individual RDT features. List is:
5399 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
5401 E.g. to turn on cmt and turn off mba use:
5405 Format (x86 or x86_64):
5406 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] | d[efault] \
5408 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
5410 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
5411 (prefix with 'panic_' to set mode for panic
5413 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
5414 reboot_force is either force or not specified,
5415 reboot_cpu is s[mp]#### with #### being the processor
5416 to be used for rebooting.
5418 refscale.holdoff= [KNL]
5419 Set test-start holdoff period. The purpose of
5420 this parameter is to delay the start of the
5421 test until boot completes in order to avoid
5424 refscale.loops= [KNL]
5425 Set the number of loops over the synchronization
5426 primitive under test. Increasing this number
5427 reduces noise due to loop start/end overhead,
5428 but the default has already reduced the per-pass
5429 noise to a handful of picoseconds on ca. 2020
5432 refscale.nreaders= [KNL]
5433 Set number of readers. The default value of -1
5434 selects N, where N is roughly 75% of the number
5435 of CPUs. A value of zero is an interesting choice.
5437 refscale.nruns= [KNL]
5438 Set number of runs, each of which is dumped onto
5441 refscale.readdelay= [KNL]
5442 Set the read-side critical-section duration,
5443 measured in microseconds.
5445 refscale.scale_type= [KNL]
5446 Specify the read-protection implementation to test.
5448 refscale.shutdown= [KNL]
5449 Shut down the system at the end of the performance
5450 test. This defaults to 1 (shut it down) when
5451 refscale is built into the kernel and to 0 (leave
5452 it running) when refscale is built as a module.
5454 refscale.verbose= [KNL]
5455 Enable additional printk() statements.
5457 refscale.verbose_batched= [KNL]
5458 Batch the additional printk() statements. If zero
5459 (the default) or negative, print everything. Otherwise,
5460 print every Nth verbose statement, where N is the value
5464 [KNL, SMP] Set scheduler's default relax_domain_level.
5465 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
5467 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
5468 Format: <base1>,<size1>[,<base2>,<size2>,...]
5469 Reserve I/O ports or memory so the kernel won't use
5470 them. If <base> is less than 0x10000, the region
5471 is assumed to be I/O ports; otherwise it is memory.
5473 reservetop= [X86-32]
5475 Reserves a hole at the top of the kernel virtual
5478 reset_devices [KNL] Force drivers to reset the underlying device
5479 during initialization.
5482 Specify the partition device for software suspend
5484 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
5486 resume_offset= [SWSUSP]
5487 Specify the offset from the beginning of the partition
5488 given by "resume=" at which the swap header is located,
5489 in <PAGE_SIZE> units (needed only for swap files).
5490 See Documentation/power/swsusp-and-swap-files.rst
5492 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
5493 read the resume files
5495 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
5496 Useful for devices that are detected asynchronously
5497 (e.g. USB and MMC devices).
5499 retain_initrd [RAM] Keep initrd memory after extraction
5501 retbleed= [X86] Control mitigation of RETBleed (Arbitrary
5502 Speculative Code Execution with Return Instructions)
5505 AMD-based UNRET and IBPB mitigations alone do not stop
5506 sibling threads from influencing the predictions of other
5507 sibling threads. For that reason, STIBP is used on pro-
5508 cessors that support it, and mitigate SMT on processors
5512 auto - automatically select a migitation
5513 auto,nosmt - automatically select a mitigation,
5514 disabling SMT if necessary for
5515 the full mitigation (only on Zen1
5516 and older without STIBP).
5517 ibpb - On AMD, mitigate short speculation
5518 windows on basic block boundaries too.
5519 Safe, highest perf impact. It also
5520 enables STIBP if present. Not suitable
5522 ibpb,nosmt - Like "ibpb" above but will disable SMT
5523 when STIBP is not available. This is
5524 the alternative for systems which do not
5526 unret - Force enable untrained return thunks,
5527 only effective on AMD f15h-f17h based
5529 unret,nosmt - Like unret, but will disable SMT when STIBP
5530 is not available. This is the alternative for
5531 systems which do not have STIBP.
5533 Selecting 'auto' will choose a mitigation method at run
5534 time according to the CPU.
5536 Not specifying this option is equivalent to retbleed=auto.
5538 rfkill.default_state=
5539 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
5540 etc. communication is blocked by default.
5543 rfkill.master_switch_mode=
5544 0 The "airplane mode" button does nothing.
5545 1 The "airplane mode" button toggles between everything
5546 blocked and the previous configuration.
5547 2 The "airplane mode" button toggles between everything
5548 blocked and everything unblocked.
5550 rhash_entries= [KNL,NET]
5551 Set number of hash buckets for route cache
5554 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
5557 ro [KNL] Mount root device read-only on boot
5560 on Mark read-only kernel memory as read-only (default).
5561 off Leave read-only kernel memory writable for debugging.
5562 full Mark read-only kernel memory and aliases as read-only
5566 Enable the uart passthrough on the designated usb port
5567 on Rockchip SoCs. When active, the signals of the
5568 debug-uart get routed to the D+ and D- pins of the usb
5569 port and the regular usb controller gets disabled.
5571 root= [KNL] Root filesystem
5572 Usually this a a block device specifier of some kind,
5573 see the early_lookup_bdev comment in
5574 block/early-lookup.c for details.
5575 Alternatively this can be "ram" for the legacy initial
5576 ramdisk, "nfs" and "cifs" for root on a network file
5577 system, or "mtd" and "ubi" for mounting from raw flash.
5579 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
5580 mount the root filesystem
5582 rootflags= [KNL] Set root filesystem mount option string
5584 rootfstype= [KNL] Set root filesystem type
5586 rootwait [KNL] Wait (indefinitely) for root device to show up.
5587 Useful for devices that are detected asynchronously
5588 (e.g. USB and MMC devices).
5590 rootwait= [KNL] Maximum time (in seconds) to wait for root device
5591 to show up before attempting to mount the root
5594 rproc_mem=nn[KMG][@address]
5595 [KNL,ARM,CMA] Remoteproc physical memory block.
5596 Memory area to be used by remote processor image,
5599 rw [KNL] Mount root device read-write on boot
5601 S [KNL] Run init in single mode
5603 s390_iommu= [HW,S390]
5604 Set s390 IOTLB flushing mode
5606 With strict flushing every unmap operation will result in
5607 an IOTLB flush. Default is lazy flushing before reuse,
5610 s390_iommu_aperture= [KNL,S390]
5611 Specifies the size of the per device DMA address space
5612 accessible through the DMA and IOMMU APIs as a decimal
5613 factor of the size of main memory.
5614 The default is 1 meaning that one can concurrently use
5615 as many DMA addresses as physical memory is installed,
5616 if supported by hardware, and thus map all of memory
5617 once. With a value of 2 one can map all of memory twice
5618 and so on. As a special case a factor of 0 imposes no
5619 restrictions other than those given by hardware at the
5620 cost of significant additional memory use for tables.
5623 See drivers/net/irda/sa1100_ir.c.
5625 sched_verbose [KNL] Enables verbose scheduler debug messages.
5627 schedstats= [KNL,X86] Enable or disable scheduled statistics.
5628 Allowed values are enable and disable. This feature
5629 incurs a small amount of overhead in the scheduler
5630 but is useful for debugging and performance tuning.
5632 sched_thermal_decay_shift=
5633 [KNL, SMP] Set a decay shift for scheduler thermal
5634 pressure signal. Thermal pressure signal follows the
5635 default decay period of other scheduler pelt
5636 signals(usually 32 ms but configurable). Setting
5637 sched_thermal_decay_shift will left shift the decay
5638 period for the thermal pressure signal by the shift
5640 i.e. with the default pelt decay period of 32 ms
5641 sched_thermal_decay_shift thermal pressure decay pr
5645 Format: integer between 0 and 10
5648 scftorture.holdoff= [KNL]
5649 Number of seconds to hold off before starting
5650 test. Defaults to zero for module insertion and
5651 to 10 seconds for built-in smp_call_function()
5654 scftorture.longwait= [KNL]
5655 Request ridiculously long waits randomly selected
5656 up to the chosen limit in seconds. Zero (the
5657 default) disables this feature. Please note
5658 that requesting even small non-zero numbers of
5659 seconds can result in RCU CPU stall warnings,
5660 softlockup complaints, and so on.
5662 scftorture.nthreads= [KNL]
5663 Number of kthreads to spawn to invoke the
5664 smp_call_function() family of functions.
5665 The default of -1 specifies a number of kthreads
5666 equal to the number of CPUs.
5668 scftorture.onoff_holdoff= [KNL]
5669 Number seconds to wait after the start of the
5670 test before initiating CPU-hotplug operations.
5672 scftorture.onoff_interval= [KNL]
5673 Number seconds to wait between successive
5674 CPU-hotplug operations. Specifying zero (which
5675 is the default) disables CPU-hotplug operations.
5677 scftorture.shutdown_secs= [KNL]
5678 The number of seconds following the start of the
5679 test after which to shut down the system. The
5680 default of zero avoids shutting down the system.
5681 Non-zero values are useful for automated tests.
5683 scftorture.stat_interval= [KNL]
5684 The number of seconds between outputting the
5685 current test statistics to the console. A value
5686 of zero disables statistics output.
5688 scftorture.stutter_cpus= [KNL]
5689 The number of jiffies to wait between each change
5690 to the set of CPUs under test.
5692 scftorture.use_cpus_read_lock= [KNL]
5693 Use use_cpus_read_lock() instead of the default
5694 preempt_disable() to disable CPU hotplug
5695 while invoking one of the smp_call_function*()
5698 scftorture.verbose= [KNL]
5699 Enable additional printk() statements.
5701 scftorture.weight_single= [KNL]
5702 The probability weighting to use for the
5703 smp_call_function_single() function with a zero
5704 "wait" parameter. A value of -1 selects the
5705 default if all other weights are -1. However,
5706 if at least one weight has some other value, a
5707 value of -1 will instead select a weight of zero.
5709 scftorture.weight_single_wait= [KNL]
5710 The probability weighting to use for the
5711 smp_call_function_single() function with a
5712 non-zero "wait" parameter. See weight_single.
5714 scftorture.weight_many= [KNL]
5715 The probability weighting to use for the
5716 smp_call_function_many() function with a zero
5717 "wait" parameter. See weight_single.
5718 Note well that setting a high probability for
5719 this weighting can place serious IPI load
5722 scftorture.weight_many_wait= [KNL]
5723 The probability weighting to use for the
5724 smp_call_function_many() function with a
5725 non-zero "wait" parameter. See weight_single
5728 scftorture.weight_all= [KNL]
5729 The probability weighting to use for the
5730 smp_call_function_all() function with a zero
5731 "wait" parameter. See weight_single and
5734 scftorture.weight_all_wait= [KNL]
5735 The probability weighting to use for the
5736 smp_call_function_all() function with a
5737 non-zero "wait" parameter. See weight_single
5740 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5741 xtime_lock contention on larger systems, and/or RCU lock
5742 contention on all systems with CONFIG_MAXSMP set.
5743 Format: { "0" | "1" }
5744 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5746 Note: increases power consumption, thus should only be
5747 enabled if running jitter sensitive (HPC/RT) workloads.
5749 security= [SECURITY] Choose a legacy "major" security module to
5750 enable at boot. This has been deprecated by the
5753 selinux= [SELINUX] Disable or enable SELinux at boot time.
5754 Format: { "0" | "1" }
5755 See security/selinux/Kconfig help text.
5760 serialnumber [BUGS=X86-32]
5762 sev=option[,option...] [X86-64] See Documentation/arch/x86/x86_64/boot-options.rst
5765 Maximal number of shapers.
5767 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
5768 Limit apic dumping. The parameter defines the maximal
5769 number of local apics being dumped. Also it is possible
5770 to set it to "all" by meaning -- no limit here.
5771 Format: { 1 (default) | 2 | ... | all }.
5772 The parameter valid if only apic=debug or
5773 apic=verbose is specified.
5774 Example: apic=debug show_lapic=all
5782 Enable merging of slabs with similar size when the
5783 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5786 Disable merging of slabs with similar size. May be
5787 necessary if there is some reason to distinguish
5788 allocs to different slabs, especially in hardened
5789 environments where the risk of heap overflows and
5790 layout control by attackers can usually be
5791 frustrated by disabling merging. This will reduce
5792 most of the exposure of a heap attack to a single
5793 cache (risks via metadata attacks are mostly
5794 unchanged). Debug options disable merging on their
5796 For more information see Documentation/mm/slub.rst.
5798 slab_max_order= [MM, SLAB]
5799 Determines the maximum allowed order for slabs.
5800 A high setting may cause OOMs due to memory
5801 fragmentation. Defaults to 1 for systems with
5802 more than 32MB of RAM, 0 otherwise.
5804 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5805 Enabling slub_debug allows one to determine the
5806 culprit if slab objects become corrupted. Enabling
5807 slub_debug can create guard zones around objects and
5808 may poison objects when not in use. Also tracks the
5809 last alloc / free. For more information see
5810 Documentation/mm/slub.rst.
5812 slub_max_order= [MM, SLUB]
5813 Determines the maximum allowed order for slabs.
5814 A high setting may cause OOMs due to memory
5815 fragmentation. For more information see
5816 Documentation/mm/slub.rst.
5818 slub_min_objects= [MM, SLUB]
5819 The minimum number of objects per slab. SLUB will
5820 increase the slab order up to slub_max_order to
5821 generate a sufficiently large slab able to contain
5822 the number of objects indicated. The higher the number
5823 of objects the smaller the overhead of tracking slabs
5824 and the less frequently locks need to be acquired.
5825 For more information see Documentation/mm/slub.rst.
5827 slub_min_order= [MM, SLUB]
5828 Determines the minimum page order for slabs. Must be
5829 lower than slub_max_order.
5830 For more information see Documentation/mm/slub.rst.
5832 slub_merge [MM, SLUB]
5833 Same with slab_merge.
5835 slub_nomerge [MM, SLUB]
5836 Same with slab_nomerge. This is supported for legacy.
5837 See slab_nomerge for more information.
5840 Format: <io1>[,<io2>[,...,<io8>]]
5842 smp.csd_lock_timeout= [KNL]
5843 Specify the period of time in milliseconds
5844 that smp_call_function() and friends will wait
5845 for a CPU to release the CSD lock. This is
5846 useful when diagnosing bugs involving CPUs
5847 disabling interrupts for extended periods
5848 of time. Defaults to 5,000 milliseconds, and
5849 setting a value of zero disables this feature.
5850 This feature may be more efficiently disabled
5851 using the csdlock_debug- kernel parameter.
5853 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5854 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5855 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5856 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5857 smsc-ircc2.ircc_irq= [HW] IRQ line
5858 smsc-ircc2.ircc_dma= [HW] DMA channel
5859 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5860 0: Toshiba Satellite 1800 (GP data pin select)
5861 1: Fast pin select (default)
5864 smt= [KNL,MIPS,S390] Set the maximum number of threads (logical
5865 CPUs) to use per physical CPU on systems capable of
5866 symmetric multithreading (SMT). Will be capped to the
5867 actual hardware limit.
5869 Default: -1 (no limit)
5872 [KNL] Should the soft-lockup detector generate panics.
5875 A value of 1 instructs the soft-lockup detector
5876 to panic the machine when a soft-lockup occurs. It is
5877 also controlled by the kernel.softlockup_panic sysctl
5878 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5879 respective build-time switch to that functionality.
5881 softlockup_all_cpu_backtrace=
5882 [KNL] Should the soft-lockup detector generate
5883 backtraces on all cpus.
5886 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5887 See Documentation/admin-guide/laptops/sonypi.rst
5889 spectre_v2= [X86] Control mitigation of Spectre variant 2
5890 (indirect branch speculation) vulnerability.
5891 The default operation protects the kernel from
5894 on - unconditionally enable, implies
5896 off - unconditionally disable, implies
5898 auto - kernel detects whether your CPU model is
5901 Selecting 'on' will, and 'auto' may, choose a
5902 mitigation method at run time according to the
5903 CPU, the available microcode, the setting of the
5904 CONFIG_RETPOLINE configuration option, and the
5905 compiler with which the kernel was built.
5907 Selecting 'on' will also enable the mitigation
5908 against user space to user space task attacks.
5910 Selecting 'off' will disable both the kernel and
5911 the user space protections.
5913 Specific mitigations can also be selected manually:
5915 retpoline - replace indirect branches
5916 retpoline,generic - Retpolines
5917 retpoline,lfence - LFENCE; indirect branch
5918 retpoline,amd - alias for retpoline,lfence
5919 eibrs - Enhanced/Auto IBRS
5920 eibrs,retpoline - Enhanced/Auto IBRS + Retpolines
5921 eibrs,lfence - Enhanced/Auto IBRS + LFENCE
5922 ibrs - use IBRS to protect kernel
5924 Not specifying this option is equivalent to
5928 [X86] Control mitigation of Spectre variant 2
5929 (indirect branch speculation) vulnerability between
5932 on - Unconditionally enable mitigations. Is
5933 enforced by spectre_v2=on
5935 off - Unconditionally disable mitigations. Is
5936 enforced by spectre_v2=off
5938 prctl - Indirect branch speculation is enabled,
5939 but mitigation can be enabled via prctl
5940 per thread. The mitigation control state
5941 is inherited on fork.
5944 - Like "prctl" above, but only STIBP is
5945 controlled per thread. IBPB is issued
5946 always when switching between different user
5950 - Same as "prctl" above, but all seccomp
5951 threads will enable the mitigation unless
5952 they explicitly opt out.
5955 - Like "seccomp" above, but only STIBP is
5956 controlled per thread. IBPB is issued
5957 always when switching between different
5958 user space processes.
5960 auto - Kernel selects the mitigation depending on
5961 the available CPU features and vulnerability.
5963 Default mitigation: "prctl"
5965 Not specifying this option is equivalent to
5966 spectre_v2_user=auto.
5968 spec_rstack_overflow=
5969 [X86] Control RAS overflow mitigation on AMD Zen CPUs
5971 off - Disable mitigation
5972 microcode - Enable microcode mitigation only
5973 safe-ret - Enable sw-only safe RET mitigation (default)
5974 ibpb - Enable mitigation by issuing IBPB on
5976 ibpb-vmexit - Issue IBPB only on VMEXIT
5977 (cloud-specific mitigation)
5979 spec_store_bypass_disable=
5980 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5981 (Speculative Store Bypass vulnerability)
5983 Certain CPUs are vulnerable to an exploit against a
5984 a common industry wide performance optimization known
5985 as "Speculative Store Bypass" in which recent stores
5986 to the same memory location may not be observed by
5987 later loads during speculative execution. The idea
5988 is that such stores are unlikely and that they can
5989 be detected prior to instruction retirement at the
5990 end of a particular speculation execution window.
5992 In vulnerable processors, the speculatively forwarded
5993 store can be used in a cache side channel attack, for
5994 example to read memory to which the attacker does not
5995 directly have access (e.g. inside sandboxed code).
5997 This parameter controls whether the Speculative Store
5998 Bypass optimization is used.
6000 On x86 the options are:
6002 on - Unconditionally disable Speculative Store Bypass
6003 off - Unconditionally enable Speculative Store Bypass
6004 auto - Kernel detects whether the CPU model contains an
6005 implementation of Speculative Store Bypass and
6006 picks the most appropriate mitigation. If the
6007 CPU is not vulnerable, "off" is selected. If the
6008 CPU is vulnerable the default mitigation is
6009 architecture and Kconfig dependent. See below.
6010 prctl - Control Speculative Store Bypass per thread
6011 via prctl. Speculative Store Bypass is enabled
6012 for a process by default. The state of the control
6013 is inherited on fork.
6014 seccomp - Same as "prctl" above, but all seccomp threads
6015 will disable SSB unless they explicitly opt out.
6017 Default mitigations:
6020 On powerpc the options are:
6022 on,auto - On Power8 and Power9 insert a store-forwarding
6023 barrier on kernel entry and exit. On Power7
6024 perform a software flush on kernel entry and
6028 Not specifying this option is equivalent to
6029 spec_store_bypass_disable=auto.
6031 spia_io_base= [HW,MTD]
6037 [X86] Enable split lock detection or bus lock detection
6039 When enabled (and if hardware support is present), atomic
6040 instructions that access data across cache line
6041 boundaries will result in an alignment check exception
6042 for split lock detection or a debug exception for
6047 warn - the kernel will emit rate-limited warnings
6048 about applications triggering the #AC
6049 exception or the #DB exception. This mode is
6050 the default on CPUs that support split lock
6051 detection or bus lock detection. Default
6052 behavior is by #AC if both features are
6053 enabled in hardware.
6055 fatal - the kernel will send SIGBUS to applications
6056 that trigger the #AC exception or the #DB
6057 exception. Default behavior is by #AC if
6058 both features are enabled in hardware.
6061 Set system wide rate limit to N bus locks
6062 per second for bus lock detection.
6065 N/A for split lock detection.
6068 If an #AC exception is hit in the kernel or in
6069 firmware (i.e. not while executing in user mode)
6070 the kernel will oops in either "warn" or "fatal"
6073 #DB exception for bus lock is triggered only when
6077 Control the Special Register Buffer Data Sampling
6080 Certain CPUs are vulnerable to an MDS-like
6081 exploit which can leak bits from the random
6084 By default, this issue is mitigated by
6085 microcode. However, the microcode fix can cause
6086 the RDRAND and RDSEED instructions to become
6087 much slower. Among other effects, this will
6088 result in reduced throughput from /dev/urandom.
6090 The microcode mitigation can be disabled with
6091 the following option:
6093 off: Disable mitigation and remove
6094 performance impact to RDRAND and RDSEED
6096 srcutree.big_cpu_lim [KNL]
6097 Specifies the number of CPUs constituting a
6098 large system, such that srcu_struct structures
6099 should immediately allocate an srcu_node array.
6100 This kernel-boot parameter defaults to 128,
6101 but takes effect only when the low-order four
6102 bits of srcutree.convert_to_big is equal to 3
6105 srcutree.convert_to_big [KNL]
6106 Specifies under what conditions an SRCU tree
6107 srcu_struct structure will be converted to big
6108 form, that is, with an rcu_node tree:
6111 1: At init_srcu_struct() time.
6112 2: When rcutorture decides to.
6113 3: Decide at boot time (default).
6114 0x1X: Above plus if high contention.
6116 Either way, the srcu_node tree will be sized based
6117 on the actual runtime number of CPUs (nr_cpu_ids)
6118 instead of the compile-time CONFIG_NR_CPUS.
6120 srcutree.counter_wrap_check [KNL]
6121 Specifies how frequently to check for
6122 grace-period sequence counter wrap for the
6123 srcu_data structure's ->srcu_gp_seq_needed field.
6124 The greater the number of bits set in this kernel
6125 parameter, the less frequently counter wrap will
6126 be checked for. Note that the bottom two bits
6129 srcutree.exp_holdoff [KNL]
6130 Specifies how many nanoseconds must elapse
6131 since the end of the last SRCU grace period for
6132 a given srcu_struct until the next normal SRCU
6133 grace period will be considered for automatic
6134 expediting. Set to zero to disable automatic
6137 srcutree.srcu_max_nodelay [KNL]
6138 Specifies the number of no-delay instances
6139 per jiffy for which the SRCU grace period
6140 worker thread will be rescheduled with zero
6141 delay. Beyond this limit, worker thread will
6142 be rescheduled with a sleep delay of one jiffy.
6144 srcutree.srcu_max_nodelay_phase [KNL]
6145 Specifies the per-grace-period phase, number of
6146 non-sleeping polls of readers. Beyond this limit,
6147 grace period worker thread will be rescheduled
6148 with a sleep delay of one jiffy, between each
6149 rescan of the readers, for a grace period phase.
6151 srcutree.srcu_retry_check_delay [KNL]
6152 Specifies number of microseconds of non-sleeping
6153 delay between each non-sleeping poll of readers.
6155 srcutree.small_contention_lim [KNL]
6156 Specifies the number of update-side contention
6157 events per jiffy will be tolerated before
6158 initiating a conversion of an srcu_struct
6159 structure to big form. Note that the value of
6160 srcutree.convert_to_big must have the 0x10 bit
6161 set for contention-based conversions to occur.
6164 Speculative Store Bypass Disable control
6166 On CPUs that are vulnerable to the Speculative
6167 Store Bypass vulnerability and offer a
6168 firmware based mitigation, this parameter
6169 indicates how the mitigation should be used:
6171 force-on: Unconditionally enable mitigation for
6172 for both kernel and userspace
6173 force-off: Unconditionally disable mitigation for
6174 for both kernel and userspace
6175 kernel: Always enable mitigation in the
6176 kernel, and offer a prctl interface
6177 to allow userspace to register its
6178 interest in being mitigated too.
6180 stack_guard_gap= [MM]
6181 override the default stack gap protection. The value
6182 is in page units and it defines how many pages prior
6183 to (for stacks growing down) resp. after (for stacks
6184 growing up) the main stack are reserved for no other
6185 mapping. Default value is 256 pages.
6187 stack_depot_disable= [KNL]
6188 Setting this to true through kernel command line will
6189 disable the stack depot thereby saving the static memory
6190 consumed by the stack hash table. By default this is set
6194 Enabled the stack tracer on boot up.
6196 stacktrace_filter=[function-list]
6197 [FTRACE] Limit the functions that the stack tracer
6198 will trace at boot up. function-list is a comma-separated
6199 list of functions. This list can be changed at run
6200 time by the stack_trace_filter file in the debugfs
6201 tracing directory. Note, this enables stack tracing
6202 and the stacktrace above is not needed.
6206 Set the STI (builtin display/keyboard on the HP-PARISC
6207 machines) console (graphic card) which should be used
6208 as the initial boot-console.
6209 See also comment in drivers/video/console/sticore.c.
6212 See comment in drivers/video/console/sticore.c.
6215 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
6220 Enable or disable strict sigaltstack size checks
6221 against the required signal frame size which
6222 depends on the supported FPU features. This can
6223 be used to filter out binaries which have
6224 not yet been made aware of AT_MINSIGSTKSZ.
6227 Limits the number of kernel HPT entries in the hash
6228 page table to increase the rate of hash page table
6229 faults on kernel addresses.
6232 Limits the number of kernel SLB entries, and flushes
6233 them frequently to increase the rate of SLB faults
6234 on kernel addresses.
6236 sunrpc.min_resvport=
6237 sunrpc.max_resvport=
6239 SunRPC servers often require that client requests
6240 originate from a privileged port (i.e. a port in the
6241 range 0 < portnr < 1024).
6242 An administrator who wishes to reserve some of these
6243 ports for other uses may adjust the range that the
6244 kernel's sunrpc client considers to be privileged
6245 using these two parameters to set the minimum and
6246 maximum port values.
6248 sunrpc.svc_rpc_per_connection_limit=
6250 Limit the number of requests that the server will
6251 process in parallel from a single connection.
6252 The default value is 0 (no limit).
6256 Control how the NFS server code allocates CPUs to
6257 service thread pools. Depending on how many NICs
6258 you have and where their interrupts are bound, this
6259 option will affect which CPUs will do NFS serving.
6260 Note: this parameter cannot be changed while the
6261 NFS server is running.
6263 auto the server chooses an appropriate mode
6264 automatically using heuristics
6265 global a single global pool contains all CPUs
6266 percpu one pool for each CPU
6267 pernode one pool for each NUMA node (equivalent
6268 to global on non-NUMA machines)
6270 sunrpc.tcp_slot_table_entries=
6271 sunrpc.udp_slot_table_entries=
6273 Sets the upper limit on the number of simultaneous
6274 RPC calls that can be sent from the client to a
6275 server. Increasing these values may allow you to
6276 improve throughput, but will also increase the
6277 amount of memory reserved for use by the client.
6279 suspend.pm_test_delay=
6281 Sets the number of seconds to remain in a suspend test
6282 mode before resuming the system (see
6283 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
6284 is set. Default value is 5.
6287 Format: { on | off | y | n | 1 | 0 }
6288 This parameter controls use of the Protected
6289 Execution Facility on pSeries.
6291 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
6292 Format: { <int> [,<int>] | force | noforce }
6293 <int> -- Number of I/O TLB slabs
6294 <int> -- Second integer after comma. Number of swiotlb
6295 areas with their own lock. Will be rounded up
6297 force -- force using of bounce buffers even if they
6298 wouldn't be automatically used by the kernel
6299 noforce -- Never use bounce buffers (for debugging)
6304 Set a sysctl parameter, right before loading the init
6305 process, as if the value was written to the respective
6306 /proc/sys/... file. Both '.' and '/' are recognized as
6307 separators. Unrecognized parameters and invalid values
6308 are reported in the kernel log. Sysctls registered
6309 later by a loaded module cannot be set this way.
6310 Example: sysctl.vm.swappiness=40
6312 sysrq_always_enabled
6314 Ignore sysrq setting - this boot parameter will
6315 neutralize any effect of /proc/sys/kernel/sysrq.
6316 Useful for debugging.
6318 tcpmhash_entries= [KNL,NET]
6319 Set the number of tcp_metrics_hash slots.
6320 Default value is 8192 or 16384 depending on total
6321 ram pages. This is used to specify the TCP metrics
6322 cache size. See Documentation/networking/ip-sysctl.rst
6323 "tcp_no_metrics_save" section for more details.
6327 test_suspend= [SUSPEND]
6328 Format: { "mem" | "standby" | "freeze" }[,N]
6329 Specify "mem" (for Suspend-to-RAM) or "standby" (for
6330 standby suspend) or "freeze" (for suspend type freeze)
6331 as the system sleep state during system startup with
6332 the optional capability to repeat N number of times.
6333 The system is woken from this state using a
6334 wakeup-capable RTC alarm.
6336 thash_entries= [KNL,NET]
6337 Set number of hash buckets for TCP connection
6339 thermal.act= [HW,ACPI]
6340 -1: disable all active trip points in all thermal zones
6341 <degrees C>: override all lowest active trip points
6343 thermal.crt= [HW,ACPI]
6344 -1: disable all critical trip points in all thermal zones
6345 <degrees C>: override all critical trip points
6347 thermal.off= [HW,ACPI]
6348 1: disable ACPI thermal control
6350 thermal.psv= [HW,ACPI]
6351 -1: disable all passive trip points
6352 <degrees C>: override all passive trip points to this
6355 thermal.tzp= [HW,ACPI]
6356 Specify global default ACPI thermal zone polling rate
6357 <deci-seconds>: poll all this frequency
6358 0: no polling (default)
6361 Force threading of all interrupt handlers except those
6362 marked explicitly IRQF_NO_THREAD.
6366 Specify if the kernel should make use of the cpu
6367 topology information if the hardware supports this.
6368 The scheduler will make use of this information and
6369 e.g. base its process migration decisions on it.
6372 topology_updates= [KNL, PPC, NUMA]
6374 Specify if the kernel should ignore (off)
6375 topology updates sent by the hypervisor to this
6378 torture.disable_onoff_at_boot= [KNL]
6379 Prevent the CPU-hotplug component of torturing
6380 until after init has spawned.
6382 torture.ftrace_dump_at_shutdown= [KNL]
6383 Dump the ftrace buffer at torture-test shutdown,
6384 even if there were no errors. This can be a
6385 very costly operation when many torture tests
6386 are running concurrently, especially on systems
6387 with rotating-rust storage.
6389 torture.verbose_sleep_frequency= [KNL]
6390 Specifies how many verbose printk()s should be
6391 emitted between each sleep. The default of zero
6392 disables verbose-printk() sleeping.
6394 torture.verbose_sleep_duration= [KNL]
6395 Duration of each verbose-printk() sleep in jiffies.
6399 tpm_suspend_pcr=[HW,TPM]
6400 Format: integer pcr id
6401 Specify that at suspend time, the tpm driver
6402 should extend the specified pcr with zeros,
6403 as a workaround for some chips which fail to
6404 flush the last written pcr on TPM_SaveState.
6405 This will guarantee that all the other pcrs
6408 tpm_tis.interrupts= [HW,TPM]
6409 Enable interrupts for the MMIO based physical layer
6410 for the FIFO interface. By default it is set to false
6411 (0). For more information about TPM hardware interfaces
6412 defined by Trusted Computing Group (TCG) see
6413 https://trustedcomputinggroup.org/resource/pc-client-platform-tpm-profile-ptp-specification/
6416 Have the tracepoints sent to printk as well as the
6417 tracing ring buffer. This is useful for early boot up
6418 where the system hangs or reboots and does not give the
6419 option for reading the tracing buffer or performing a
6420 ftrace_dump_on_oops.
6422 To turn off having tracepoints sent to printk,
6423 echo 0 > /proc/sys/kernel/tracepoint_printk
6424 Note, echoing 1 into this file without the
6425 tracepoint_printk kernel cmdline option has no effect.
6427 The tp_printk_stop_on_boot (see below) can also be used
6428 to stop the printing of events to console at
6433 Having tracepoints sent to printk() and activating high
6434 frequency tracepoints such as irq or sched, can cause
6435 the system to live lock.
6437 tp_printk_stop_on_boot [FTRACE]
6438 When tp_printk (above) is set, it can cause a lot of noise
6439 on the console. It may be useful to only include the
6440 printing of events during boot up, as user space may
6441 make the system inoperable.
6443 This command line option will stop the printing of events
6444 to console at the late_initcall_sync() time frame.
6446 trace_buf_size=nn[KMG]
6447 [FTRACE] will set tracing buffer size on each cpu.
6449 trace_clock= [FTRACE] Set the clock used for tracing events
6451 local - Use the per CPU time stamp counter
6452 (converted into nanoseconds). Fast, but
6453 depending on the architecture, may not be
6454 in sync between CPUs.
6455 global - Event time stamps are synchronize across
6456 CPUs. May be slower than the local clock,
6457 but better for some race conditions.
6458 counter - Simple counting of events (1, 2, ..)
6459 note, some counts may be skipped due to the
6460 infrastructure grabbing the clock more than
6462 uptime - Use jiffies as the time stamp.
6463 perf - Use the same clock that perf uses.
6464 mono - Use ktime_get_mono_fast_ns() for time stamps.
6465 mono_raw - Use ktime_get_raw_fast_ns() for time
6467 boot - Use ktime_get_boot_fast_ns() for time stamps.
6468 Architectures may add more clocks. See
6469 Documentation/trace/ftrace.rst for more details.
6471 trace_event=[event-list]
6472 [FTRACE] Set and start specified trace events in order
6473 to facilitate early boot debugging. The event-list is a
6474 comma-separated list of trace events to enable. See
6475 also Documentation/trace/events.rst
6477 trace_instance=[instance-info]
6478 [FTRACE] Create a ring buffer instance early in boot up.
6479 This will be listed in:
6481 /sys/kernel/tracing/instances
6483 Events can be enabled at the time the instance is created
6486 trace_instance=<name>,<system1>:<event1>,<system2>:<event2>
6488 Note, the "<system*>:" portion is optional if the event is
6491 trace_instance=foo,sched:sched_switch,irq_handler_entry,initcall
6493 will enable the "sched_switch" event (note, the "sched:" is optional, and
6494 the same thing would happen if it was left off). The irq_handler_entry
6495 event, and all events under the "initcall" system.
6497 trace_options=[option-list]
6498 [FTRACE] Enable or disable tracer options at boot.
6499 The option-list is a comma delimited list of options
6500 that can be enabled or disabled just as if you were
6501 to echo the option name into
6503 /sys/kernel/tracing/trace_options
6505 For example, to enable stacktrace option (to dump the
6506 stack trace of each event), add to the command line:
6508 trace_options=stacktrace
6510 See also Documentation/trace/ftrace.rst "trace options"
6513 trace_trigger=[trigger-list]
6514 [FTRACE] Add a event trigger on specific events.
6515 Set a trigger on top of a specific event, with an optional
6518 The format is is "trace_trigger=<event>.<trigger>[ if <filter>],..."
6519 Where more than one trigger may be specified that are comma deliminated.
6523 trace_trigger="sched_switch.stacktrace if prev_state == 2"
6525 The above will enable the "stacktrace" trigger on the "sched_switch"
6526 event but only trigger it if the "prev_state" of the "sched_switch"
6527 event is "2" (TASK_UNINTERUPTIBLE).
6529 See also "Event triggers" in Documentation/trace/events.rst
6533 [FTRACE] enable this option to disable tracing when a
6534 warning is hit. This turns off "tracing_on". Tracing can
6535 be enabled again by echoing '1' into the "tracing_on"
6536 file located in /sys/kernel/tracing/
6538 This option is useful, as it disables the trace before
6539 the WARNING dump is called, which prevents the trace to
6540 be filled with content caused by the warning output.
6542 This option can also be set at run time via the sysctl
6543 option: kernel/traceoff_on_warning
6545 transparent_hugepage=
6547 Format: [always|madvise|never]
6548 Can be used to control the default behavior of the system
6549 with respect to transparent hugepages.
6550 See Documentation/admin-guide/mm/transhuge.rst
6553 trusted.source= [KEYS]
6555 This parameter identifies the trust source as a backend
6556 for trusted keys implementation. Supported trust
6561 If not specified then it defaults to iterating through
6562 the trust source list starting with TPM and assigns the
6563 first trust source as a backend which is initialized
6564 successfully during iteration.
6568 The RNG used to generate key material for trusted keys.
6571 - the same value as trusted.source: "tpm" or "tee"
6573 If not specified, "default" is used. In this case,
6574 the RNG's choice is left to each individual trust source.
6576 tsc= Disable clocksource stability checks for TSC.
6578 [x86] reliable: mark tsc clocksource as reliable, this
6579 disables clocksource verification at runtime, as well
6580 as the stability checks done at bootup. Used to enable
6581 high-resolution timer mode on older hardware, and in
6582 virtualized environment.
6583 [x86] noirqtime: Do not use TSC to do irq accounting.
6584 Used to run time disable IRQ_TIME_ACCOUNTING on any
6585 platforms where RDTSC is slow and this accounting
6587 [x86] unstable: mark the TSC clocksource as unstable, this
6588 marks the TSC unconditionally unstable at bootup and
6589 avoids any further wobbles once the TSC watchdog notices.
6590 [x86] nowatchdog: disable clocksource watchdog. Used
6591 in situations with strict latency requirements (where
6592 interruptions from clocksource watchdog are not
6594 [x86] recalibrate: force recalibration against a HW timer
6595 (HPET or PM timer) on systems whose TSC frequency was
6596 obtained from HW or FW using either an MSR or CPUID(0x15).
6597 Warn if the difference is more than 500 ppm.
6598 [x86] watchdog: Use TSC as the watchdog clocksource with
6599 which to check other HW timers (HPET or PM timer), but
6600 only on systems where TSC has been deemed trustworthy.
6601 This will be suppressed by an earlier tsc=nowatchdog and
6602 can be overridden by a later tsc=nowatchdog. A console
6603 message will flag any such suppression or overriding.
6605 tsc_early_khz= [X86] Skip early TSC calibration and use the given
6606 value instead. Useful when the early TSC frequency discovery
6607 procedure is not reliable, such as on overclocked systems
6608 with CPUID.16h support and partial CPUID.15h support.
6609 Format: <unsigned int>
6611 tsx= [X86] Control Transactional Synchronization
6612 Extensions (TSX) feature in Intel processors that
6613 support TSX control.
6615 This parameter controls the TSX feature. The options are:
6617 on - Enable TSX on the system. Although there are
6618 mitigations for all known security vulnerabilities,
6619 TSX has been known to be an accelerator for
6620 several previous speculation-related CVEs, and
6621 so there may be unknown security risks associated
6622 with leaving it enabled.
6624 off - Disable TSX on the system. (Note that this
6625 option takes effect only on newer CPUs which are
6626 not vulnerable to MDS, i.e., have
6627 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
6628 the new IA32_TSX_CTRL MSR through a microcode
6629 update. This new MSR allows for the reliable
6630 deactivation of the TSX functionality.)
6632 auto - Disable TSX if X86_BUG_TAA is present,
6633 otherwise enable TSX on the system.
6635 Not specifying this option is equivalent to tsx=off.
6637 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6640 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
6641 Abort (TAA) vulnerability.
6643 Similar to Micro-architectural Data Sampling (MDS)
6644 certain CPUs that support Transactional
6645 Synchronization Extensions (TSX) are vulnerable to an
6646 exploit against CPU internal buffers which can forward
6647 information to a disclosure gadget under certain
6650 In vulnerable processors, the speculatively forwarded
6651 data can be used in a cache side channel attack, to
6652 access data to which the attacker does not have direct
6655 This parameter controls the TAA mitigation. The
6658 full - Enable TAA mitigation on vulnerable CPUs
6661 full,nosmt - Enable TAA mitigation and disable SMT on
6662 vulnerable CPUs. If TSX is disabled, SMT
6663 is not disabled because CPU is not
6664 vulnerable to cross-thread TAA attacks.
6665 off - Unconditionally disable TAA mitigation
6667 On MDS-affected machines, tsx_async_abort=off can be
6668 prevented by an active MDS mitigation as both vulnerabilities
6669 are mitigated with the same mechanism so in order to disable
6670 this mitigation, you need to specify mds=off too.
6672 Not specifying this option is equivalent to
6673 tsx_async_abort=full. On CPUs which are MDS affected
6674 and deploy MDS mitigation, TAA mitigation is not
6675 required and doesn't provide any additional
6679 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
6681 turbografx.map[2|3]= [HW,JOY]
6682 TurboGraFX parallel port interface
6684 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
6685 See also Documentation/input/devices/joystick-parport.rst
6687 udbg-immortal [PPC] When debugging early kernel crashes that
6688 happen after console_init() and before a proper
6689 console driver takes over, this boot options might
6690 help "seeing" what's going on.
6692 uhash_entries= [KNL,NET]
6693 Set number of hash buckets for UDP/UDP-Lite connections
6696 [USB] Ignore overcurrent events (default N).
6697 Some badly-designed motherboards generate lots of
6698 bogus events, for ports that aren't wired to
6699 anything. Set this parameter to avoid log spamming.
6700 Note that genuine overcurrent events won't be
6704 [X86] Cause panic on unknown NMI.
6706 unwind_debug [X86-64]
6707 Enable unwinder debug output. This can be
6708 useful for debugging certain unwinder error
6709 conditions, including corrupt stacks and
6710 bad/missing unwinder metadata.
6712 usbcore.authorized_default=
6713 [USB] Default USB device authorization:
6714 (default -1 = authorized except for wireless USB,
6715 0 = not authorized, 1 = authorized, 2 = authorized
6716 if device connected to internal port)
6718 usbcore.autosuspend=
6719 [USB] The autosuspend time delay (in seconds) used
6720 for newly-detected USB devices (default 2). This
6721 is the time required before an idle device will be
6722 autosuspended. Devices for which the delay is set
6723 to a negative value won't be autosuspended at all.
6725 usbcore.usbfs_snoop=
6726 [USB] Set to log all usbfs traffic (default 0 = off).
6728 usbcore.usbfs_snoop_max=
6729 [USB] Maximum number of bytes to snoop in each URB
6732 usbcore.blinkenlights=
6733 [USB] Set to cycle leds on hubs (default 0 = off).
6735 usbcore.old_scheme_first=
6736 [USB] Start with the old device initialization
6737 scheme (default 0 = off).
6739 usbcore.usbfs_memory_mb=
6740 [USB] Memory limit (in MB) for buffers allocated by
6741 usbfs (default = 16, 0 = max = 2047).
6743 usbcore.use_both_schemes=
6744 [USB] Try the other device initialization scheme
6745 if the first one fails (default 1 = enabled).
6747 usbcore.initial_descriptor_timeout=
6748 [USB] Specifies timeout for the initial 64-byte
6749 USB_REQ_GET_DESCRIPTOR request in milliseconds
6750 (default 5000 = 5.0 seconds).
6752 usbcore.nousb [USB] Disable the USB subsystem
6755 [USB] A list of quirk entries to augment the built-in
6756 usb core quirk list. List entries are separated by
6757 commas. Each entry has the form
6758 VendorID:ProductID:Flags. The IDs are 4-digit hex
6759 numbers and Flags is a set of letters. Each letter
6760 will change the built-in quirk; setting it if it is
6761 clear and clearing it if it is set. The letters have
6762 the following meanings:
6763 a = USB_QUIRK_STRING_FETCH_255 (string
6764 descriptors must not be fetched using
6766 b = USB_QUIRK_RESET_RESUME (device can't resume
6767 correctly so reset it instead);
6768 c = USB_QUIRK_NO_SET_INTF (device can't handle
6769 Set-Interface requests);
6770 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
6771 handle its Configuration or Interface
6773 e = USB_QUIRK_RESET (device can't be reset
6774 (e.g morph devices), don't use reset);
6775 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
6776 more interface descriptions than the
6777 bNumInterfaces count, and can't handle
6778 talking to these interfaces);
6779 g = USB_QUIRK_DELAY_INIT (device needs a pause
6780 during initialization, after we read
6781 the device descriptor);
6782 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
6783 high speed and super speed interrupt
6784 endpoints, the USB 2.0 and USB 3.0 spec
6785 require the interval in microframes (1
6786 microframe = 125 microseconds) to be
6787 calculated as interval = 2 ^
6789 Devices with this quirk report their
6790 bInterval as the result of this
6791 calculation instead of the exponent
6792 variable used in the calculation);
6793 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
6794 handle device_qualifier descriptor
6796 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
6797 generates spurious wakeup, ignore
6798 remote wakeup capability);
6799 k = USB_QUIRK_NO_LPM (device can't handle Link
6801 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
6802 (Device reports its bInterval as linear
6803 frames instead of the USB 2.0
6805 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
6806 to be disconnected before suspend to
6807 prevent spurious wakeup);
6808 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
6809 pause after every control message);
6810 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
6811 delay after resetting its port);
6812 Example: quirks=0781:5580:bk,0a5c:5834:gij
6815 [USBHID] The interval which mice are to be polled at.
6818 [USBHID] The interval which joysticks are to be polled at.
6821 [USBHID] The interval which keyboards are to be polled at.
6823 usb-storage.delay_use=
6824 [UMS] The delay in seconds before a new device is
6825 scanned for Logical Units (default 1).
6828 [UMS] A list of quirks entries to supplement or
6829 override the built-in unusual_devs list. List
6830 entries are separated by commas. Each entry has
6831 the form VID:PID:Flags where VID and PID are Vendor
6832 and Product ID values (4-digit hex numbers) and
6833 Flags is a set of characters, each corresponding
6834 to a common usb-storage quirk flag as follows:
6835 a = SANE_SENSE (collect more than 18 bytes
6836 of sense data, not on uas);
6837 b = BAD_SENSE (don't collect more than 18
6838 bytes of sense data, not on uas);
6839 c = FIX_CAPACITY (decrease the reported
6840 device capacity by one sector);
6841 d = NO_READ_DISC_INFO (don't use
6842 READ_DISC_INFO command, not on uas);
6843 e = NO_READ_CAPACITY_16 (don't use
6844 READ_CAPACITY_16 command);
6845 f = NO_REPORT_OPCODES (don't use report opcodes
6847 g = MAX_SECTORS_240 (don't transfer more than
6848 240 sectors at a time, uas only);
6849 h = CAPACITY_HEURISTICS (decrease the
6850 reported device capacity by one
6851 sector if the number is odd);
6852 i = IGNORE_DEVICE (don't bind to this
6854 j = NO_REPORT_LUNS (don't use report luns
6856 k = NO_SAME (do not use WRITE_SAME, uas only)
6857 l = NOT_LOCKABLE (don't try to lock and
6858 unlock ejectable media, not on uas);
6859 m = MAX_SECTORS_64 (don't transfer more
6860 than 64 sectors = 32 KB at a time,
6862 n = INITIAL_READ10 (force a retry of the
6863 initial READ(10) command, not on uas);
6864 o = CAPACITY_OK (accept the capacity
6865 reported by the device, not on uas);
6866 p = WRITE_CACHE (the device cache is ON
6867 by default, not on uas);
6868 r = IGNORE_RESIDUE (the device reports
6869 bogus residue values, not on uas);
6870 s = SINGLE_LUN (the device has only one
6872 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
6873 commands, uas only);
6874 u = IGNORE_UAS (don't bind to the uas driver);
6875 w = NO_WP_DETECT (don't test whether the
6876 medium is write-protected).
6877 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
6878 even if the device claims no cache,
6880 Example: quirks=0419:aaf5:rl,0421:0433:rc
6882 user_debug= [KNL,ARM]
6884 See arch/arm/Kconfig.debug help text.
6885 1 - undefined instruction events
6887 4 - invalid data aborts
6890 Example: user_debug=31
6893 [X86] Flags controlling user PTE allocations.
6895 nohigh = do not allocate PTE pages in
6896 HIGHMEM regardless of setting
6899 vdso= [X86,SH,SPARC]
6900 On X86_32, this is an alias for vdso32=. Otherwise:
6902 vdso=1: enable VDSO (the default)
6903 vdso=0: disable VDSO mapping
6905 vdso32= [X86] Control the 32-bit vDSO
6906 vdso32=1: enable 32-bit VDSO
6907 vdso32=0 or vdso32=2: disable 32-bit VDSO
6909 See the help text for CONFIG_COMPAT_VDSO for more
6910 details. If CONFIG_COMPAT_VDSO is set, the default is
6911 vdso32=0; otherwise, the default is vdso32=1.
6913 For compatibility with older kernels, vdso32=2 is an
6916 Try vdso32=0 if you encounter an error that says:
6917 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
6920 vector=percpu: enable percpu vector domain
6922 video= [FB] Frame buffer configuration
6923 See Documentation/fb/modedb.rst.
6925 video.brightness_switch_enabled= [ACPI]
6927 If set to 1, on receiving an ACPI notify event
6928 generated by hotkey, video driver will adjust brightness
6929 level and then send out the event to user space through
6930 the allocated input device. If set to 0, video driver
6931 will only send out the event without touching backlight
6936 [VMMIO] Memory mapped virtio (platform) device.
6938 <size>@<baseaddr>:<irq>[:<id>]
6940 <size> := size (can use standard suffixes
6942 <baseaddr> := physical base address
6943 <irq> := interrupt number (as passed to
6945 <id> := (optional) platform device id
6947 virtio_mmio.device=1K@0x100b0000:48:7
6949 Can be used multiple times for multiple devices.
6951 vga= [BOOT,X86-32] Select a particular video mode
6952 See Documentation/arch/x86/boot.rst and
6953 Documentation/admin-guide/svga.rst.
6954 Use vga=ask for menu.
6955 This is actually a boot loader parameter; the value is
6956 passed to the kernel using a special protocol.
6958 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6959 May slow down system boot speed, especially when
6960 enabled on systems with a large amount of memory.
6961 All options are enabled by default, and this
6962 interface is meant to allow for selectively
6963 enabling or disabling specific virtual memory
6966 Available options are:
6967 P Enable page structure init time poisoning
6968 - Disable all of the above options
6970 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6971 size of <nn>. This can be used to increase the
6972 minimum size (128MB on x86). It can also be used to
6973 decrease the size and leave more room for directly
6976 vmcp_cma=nn[MG] [KNL,S390]
6977 Sets the memory size reserved for contiguous memory
6978 allocations for the vmcp device driver.
6980 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6983 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6986 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6990 Controls the behavior of vsyscalls (i.e. calls to
6991 fixed addresses of 0xffffffffff600x00 from legacy
6992 code). Most statically-linked binaries and older
6993 versions of glibc use these calls. Because these
6994 functions are at fixed addresses, they make nice
6995 targets for exploits that can control RIP.
6997 emulate Vsyscalls turn into traps and are emulated
6998 reasonably safely. The vsyscall page is
7001 xonly [default] Vsyscalls turn into traps and are
7002 emulated reasonably safely. The vsyscall
7003 page is not readable.
7005 none Vsyscalls don't work at all. This makes
7006 them quite hard to use for exploits but
7007 might break your system.
7009 vt.color= [VT] Default text color.
7010 Format: 0xYX, X = foreground, Y = background.
7011 Default: 0x07 = light gray on black.
7013 vt.cur_default= [VT] Default cursor shape.
7014 Format: 0xCCBBAA, where AA, BB, and CC are the same as
7015 the parameters of the <Esc>[?A;B;Cc escape sequence;
7016 see VGA-softcursor.txt. Default: 2 = underline.
7018 vt.default_blu= [VT]
7019 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
7020 Change the default blue palette of the console.
7021 This is a 16-member array composed of values
7024 vt.default_grn= [VT]
7025 Format: <green0>,<green1>,<green2>,...,<green15>
7026 Change the default green palette of the console.
7027 This is a 16-member array composed of values
7030 vt.default_red= [VT]
7031 Format: <red0>,<red1>,<red2>,...,<red15>
7032 Change the default red palette of the console.
7033 This is a 16-member array composed of values
7039 Set system-wide default UTF-8 mode for all tty's.
7040 Default is 1, i.e. UTF-8 mode is enabled for all
7041 newly opened terminals.
7043 vt.global_cursor_default=
7046 Set system-wide default for whether a cursor
7047 is shown on new VTs. Default is -1,
7048 i.e. cursors will be created by default unless
7049 overridden by individual drivers. 0 will hide
7050 cursors, 1 will display them.
7052 vt.italic= [VT] Default color for italic text; 0-15.
7055 vt.underline= [VT] Default color for underlined text; 0-15.
7058 watchdog timers [HW,WDT] For information on watchdog timers,
7059 see Documentation/watchdog/watchdog-parameters.rst
7060 or other driver-specific files in the
7061 Documentation/watchdog/ directory.
7065 Set the hard lockup detector stall duration
7066 threshold in seconds. The soft lockup detector
7067 threshold is set to twice the value. A value of 0
7068 disables both lockup detectors. Default is 10
7071 workqueue.watchdog_thresh=
7072 If CONFIG_WQ_WATCHDOG is configured, workqueue can
7073 warn stall conditions and dump internal state to
7074 help debugging. 0 disables workqueue stall
7075 detection; otherwise, it's the stall threshold
7076 duration in seconds. The default value is 30 and
7077 it can be updated at runtime by writing to the
7078 corresponding sysfs file.
7080 workqueue.cpu_intensive_thresh_us=
7081 Per-cpu work items which run for longer than this
7082 threshold are automatically considered CPU intensive
7083 and excluded from concurrency management to prevent
7084 them from noticeably delaying other per-cpu work
7085 items. Default is 10000 (10ms).
7087 If CONFIG_WQ_CPU_INTENSIVE_REPORT is set, the kernel
7088 will report the work functions which violate this
7089 threshold repeatedly. They are likely good
7090 candidates for using WQ_UNBOUND workqueues instead.
7092 workqueue.disable_numa
7093 By default, all work items queued to unbound
7094 workqueues are affine to the NUMA nodes they're
7095 issued on, which results in better behavior in
7096 general. If NUMA affinity needs to be disabled for
7097 whatever reason, this option can be used. Note
7098 that this also can be controlled per-workqueue for
7099 workqueues visible under /sys/bus/workqueue/.
7101 workqueue.power_efficient
7102 Per-cpu workqueues are generally preferred because
7103 they show better performance thanks to cache
7104 locality; unfortunately, per-cpu workqueues tend to
7105 be more power hungry than unbound workqueues.
7107 Enabling this makes the per-cpu workqueues which
7108 were observed to contribute significantly to power
7109 consumption unbound, leading to measurably lower
7110 power usage at the cost of small performance
7113 The default value of this parameter is determined by
7114 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
7116 workqueue.debug_force_rr_cpu
7117 Workqueue used to implicitly guarantee that work
7118 items queued without explicit CPU specified are put
7119 on the local CPU. This guarantee is no longer true
7120 and while local CPU is still preferred work items
7121 may be put on foreign CPUs. This debug option
7122 forces round-robin CPU selection to flush out
7123 usages which depend on the now broken guarantee.
7124 When enabled, memory and cache locality will be
7127 writecombine= [LOONGARCH] Control the MAT (Memory Access Type) of
7130 on - Enable writecombine, use WUC for ioremap_wc()
7131 off - Disable writecombine, use SUC for ioremap_wc()
7133 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
7134 default x2apic cluster mode on platforms
7137 xen_512gb_limit [KNL,X86-64,XEN]
7138 Restricts the kernel running paravirtualized under Xen
7139 to use only up to 512 GB of RAM. The reason to do so is
7140 crash analysis tools and Xen tools for doing domain
7141 save/restore/migration must be enabled to handle larger
7144 xen_emul_unplug= [HW,X86,XEN]
7145 Unplug Xen emulated devices
7146 Format: [unplug0,][unplug1]
7147 ide-disks -- unplug primary master IDE devices
7148 aux-ide-disks -- unplug non-primary-master IDE devices
7149 nics -- unplug network devices
7150 all -- unplug all emulated devices (NICs and IDE disks)
7151 unnecessary -- unplugging emulated devices is
7152 unnecessary even if the host did not respond to
7154 never -- do not unplug even if version check succeeds
7156 xen_legacy_crash [X86,XEN]
7157 Crash from Xen panic notifier, without executing late
7158 panic() code such as dumping handler.
7160 xen_msr_safe= [X86,XEN]
7162 Select whether to always use non-faulting (safe) MSR
7163 access functions when running as Xen PV guest. The
7164 default value is controlled by CONFIG_XEN_PV_MSR_SAFE.
7166 xen_nopvspin [X86,XEN]
7167 Disables the qspinlock slowpath using Xen PV optimizations.
7168 This parameter is obsoleted by "nopvspin" parameter, which
7169 has equivalent effect for XEN platform.
7172 Disables the PV optimizations forcing the HVM guest to
7173 run as generic HVM guest with no PV drivers.
7174 This option is obsoleted by the "nopv" option, which
7175 has equivalent effect for XEN platform.
7177 xen_no_vector_callback
7178 [KNL,X86,XEN] Disable the vector callback for Xen
7179 event channel interrupts.
7181 xen_scrub_pages= [XEN]
7182 Boolean option to control scrubbing pages before giving them back
7183 to Xen, for use by other domains. Can be also changed at runtime
7184 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
7185 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
7187 xen_timer_slop= [X86-64,XEN]
7188 Set the timer slop (in nanoseconds) for the virtual Xen
7189 timers (default is 100000). This adjusts the minimum
7190 delta of virtualized Xen timers, where lower values
7191 improve timer resolution at the expense of processing
7192 more timer interrupts.
7194 xen.balloon_boot_timeout= [XEN]
7195 The time (in seconds) to wait before giving up to boot
7196 in case initial ballooning fails to free enough memory.
7197 Applies only when running as HVM or PVH guest and
7198 started with less memory configured than allowed at
7199 max. Default is 180.
7201 xen.event_eoi_delay= [XEN]
7202 How long to delay EOI handling in case of event
7203 storms (jiffies). Default is 10.
7205 xen.event_loop_timeout= [XEN]
7206 After which time (jiffies) the event handling loop
7207 should start to delay EOI handling. Default is 2.
7209 xen.fifo_events= [XEN]
7210 Boolean parameter to disable using fifo event handling
7211 even if available. Normally fifo event handling is
7212 preferred over the 2-level event handling, as it is
7213 fairer and the number of possible event channels is
7214 much higher. Default is on (use fifo events).
7216 xirc2ps_cs= [NET,PCMCIA]
7218 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
7221 By default on POWER9 and above, the kernel will
7222 natively use the XIVE interrupt controller. This option
7223 allows the fallback firmware mode to be used:
7225 off Fallback to firmware control of XIVE interrupt
7226 controller on both pseries and powernv
7227 platforms. Only useful on POWER9 and above.
7229 xive.store-eoi=off [PPC]
7230 By default on POWER10 and above, the kernel will use
7231 stores for EOI handling when the XIVE interrupt mode
7232 is active. This option allows the XIVE driver to use
7233 loads instead, as on POWER9.
7235 xhci-hcd.quirks [USB,KNL]
7236 A hex value specifying bitmask with supplemental xhci
7237 host controller quirks. Meaning of each bit can be
7238 consulted in header drivers/usb/host/xhci.h.
7241 Format: { early | on | rw | ro | off }
7242 Controls if xmon debugger is enabled. Default is off.
7243 Passing only "xmon" is equivalent to "xmon=early".
7244 early Call xmon as early as possible on boot; xmon
7245 debugger is called from setup_arch().
7246 on xmon debugger hooks will be installed so xmon
7247 is only called on a kernel crash. Default mode,
7248 i.e. either "ro" or "rw" mode, is controlled
7249 with CONFIG_XMON_DEFAULT_RO_MODE.
7250 rw xmon debugger hooks will be installed so xmon
7251 is called only on a kernel crash, mode is write,
7252 meaning SPR registers, memory and, other data
7253 can be written using xmon commands.
7254 ro same as "rw" option above but SPR registers,
7255 memory, and other data can't be written using
7257 off xmon is disabled.