1 acpi= [HW,ACPI,X86,ARM64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
5 force -- enable ACPI if default was off
6 on -- enable ACPI but allow fallback to DT [arm64]
7 off -- disable ACPI if default was on
8 noirq -- do not use ACPI for IRQ routing
9 strict -- Be less tolerant of platforms that are not
10 strictly ACPI specification compliant.
11 rsdt -- prefer RSDT over (default) XSDT
12 copy_dsdt -- copy DSDT to memory
13 For ARM64, ONLY "acpi=off", "acpi=on" or "acpi=force"
16 See also Documentation/power/runtime_pm.rst, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
27 If set to vendor, prefer vendor specific driver
28 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
29 of the ACPI video.ko driver.
31 acpi_force_32bit_fadt_addr
32 force FADT to use 32 bit addresses rather than the
33 64 bit X_* addresses. Some firmware have broken 64
34 bit addresses for force ACPI ignore these and use
35 the older legacy 32 bit addresses.
37 acpica_no_return_repair [HW, ACPI]
38 Disable AML predefined validation mechanism
39 This mechanism can repair the evaluation result to make
40 the return objects more ACPI specification compliant.
41 This option is useful for developers to identify the
42 root cause of an AML interpreter issue when the issue
43 has something to do with the repair mechanism.
45 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
46 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
48 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
49 debug output. Bits in debug_layer correspond to a
50 _COMPONENT in an ACPI source file, e.g.,
51 #define _COMPONENT ACPI_PCI_COMPONENT
52 Bits in debug_level correspond to a level in
53 ACPI_DEBUG_PRINT statements, e.g.,
54 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
55 The debug_level mask defaults to "info". See
56 Documentation/firmware-guide/acpi/debug.rst for more information about
57 debug layers and levels.
59 Enable processor driver info messages:
60 acpi.debug_layer=0x20000000
61 Enable PCI/PCI interrupt routing info messages:
62 acpi.debug_layer=0x400000
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
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
163 acpi_osi=!! # enable all built-in OS vendor
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
229 old_ordering, nonvs, sci_force_enable, nobl }
230 See Documentation/power/video.rst for information on
232 s3_beep is for debugging; it makes the PC's speaker beep
233 as soon as the kernel's real-mode entry point is called.
234 s4_nohwsig prevents ACPI hardware signature from being
235 used during resume from hibernation.
236 old_ordering causes the ACPI 1.0 ordering of the _PTS
237 control method, with respect to putting devices into
238 low power states, to be enforced (the ACPI 2.0 ordering
239 of _PTS is used by default).
240 nonvs prevents the kernel from saving/restoring the
241 ACPI NVS memory during suspend/hibernation and resume.
242 sci_force_enable causes the kernel to set SCI_EN directly
243 on resume from S1/S3 (which is against the ACPI spec,
244 but some broken systems don't work without it).
245 nobl causes the internal blacklist of systems known to
246 behave incorrectly in some ways with respect to system
247 suspend and resume to be ignored (use wisely).
249 acpi_use_timer_override [HW,ACPI]
250 Use timer override. For some broken Nvidia NF5 boards
251 that require a timer override, but don't have HPET
253 add_efi_memmap [EFI; X86] Include EFI memory map in
254 kernel's map of available physical RAM.
257 { off | try_unsupported }
258 off: disable AGP support
259 try_unsupported: try to drive unsupported chipsets
260 (may crash computer or cause data corruption)
263 See Documentation/sound/alsa-configuration.rst
266 Allow the default userspace alignment fault handler
267 behaviour to be specified. Bit 0 enables warnings,
268 bit 1 enables fixups, and bit 2 sends a segfault.
270 align_va_addr= [X86-64]
271 Align virtual addresses by clearing slice [14:12] when
272 allocating a VMA at process creation time. This option
273 gives you up to 3% performance improvement on AMD F15h
274 machines (where it is enabled by default) for a
275 CPU-intensive style benchmark, and it can vary highly in
276 a microbenchmark depending on workload and compiler.
278 32: only for 32-bit processes
279 64: only for 64-bit processes
280 on: enable for both 32- and 64-bit processes
281 off: disable for both 32- and 64-bit processes
283 alloc_snapshot [FTRACE]
284 Allocate the ftrace snapshot buffer on boot up when the
285 main buffer is allocated. This is handy if debugging
286 and you need to use tracing_snapshot() on boot up, and
287 do not want to use tracing_snapshot_alloc() as it needs
288 to be done where GFP_KERNEL allocations are allowed.
290 amd_iommu= [HW,X86-64]
291 Pass parameters to the AMD IOMMU driver in the system.
293 fullflush - enable flushing of IO/TLB entries when
294 they are unmapped. Otherwise they are
295 flushed before they will be reused, which
297 off - do not initialize any AMD IOMMU found in
299 force_isolation - Force device isolation for all
300 devices. The IOMMU driver is not
301 allowed anymore to lift isolation
302 requirements as needed. This option
303 does not override iommu=pt
305 amd_iommu_dump= [HW,X86-64]
306 Enable AMD IOMMU driver option to dump the ACPI table
307 for AMD IOMMU. With this option enabled, AMD IOMMU
308 driver will print ACPI tables for AMD IOMMU during
309 IOMMU initialization.
311 amd_iommu_intr= [HW,X86-64]
312 Specifies one of the following AMD IOMMU interrupt
314 legacy - Use legacy interrupt remapping mode.
315 vapic - Use virtual APIC mode, which allows IOMMU
316 to inject interrupts directly into guest.
317 This mode requires kvm-amd.avic=1.
318 (Default when IOMMU HW support is present.)
320 amijoy.map= [HW,JOY] Amiga joystick support
321 Map of devices attached to JOY0DAT and JOY1DAT
323 See also Documentation/input/joydev/joystick.rst
325 analog.map= [HW,JOY] Analog joystick and gamepad support
326 Specifies type or capabilities of an analog joystick
327 connected to one of 16 gameports
328 Format: <type1>,<type2>,..<type16>
331 Power management functions (SPARCstation-4/5 + deriv.)
333 Disable APC CPU standby support. SPARCstation-Fox does
334 not play well with APC CPU idle - disable it if you have
335 APC and your system crashes randomly.
337 apic= [APIC,X86] Advanced Programmable Interrupt Controller
338 Change the output verbosity while booting
339 Format: { quiet (default) | verbose | debug }
340 Change the amount of debugging information output
341 when initialising the APIC and IO-APIC components.
342 For X86-32, this can also be used to specify an APIC
344 Format: apic=driver_name
345 Examples: apic=bigsmp
347 apic_extnmi= [APIC,X86] External NMI delivery setting
348 Format: { bsp (default) | all | none }
349 bsp: External NMI is delivered only to CPU 0
350 all: External NMIs are broadcast to all CPUs as a
352 none: External NMI is masked for all CPUs. This is
353 useful so that a dump capture kernel won't be
357 See Documentation/networking/ipv6.txt.
359 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
360 Limit apic dumping. The parameter defines the maximal
361 number of local apics being dumped. Also it is possible
362 to set it to "all" by meaning -- no limit here.
363 Format: { 1 (default) | 2 | ... | all }.
364 The parameter valid if only apic=debug or
365 apic=verbose is specified.
366 Example: apic=debug show_lapic=all
368 apm= [APM] Advanced Power Management
369 See header of arch/x86/kernel/apm_32.c.
371 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
372 Format: <io>,<irq>,<nodeID>
376 atarimouse= [HW,MOUSE] Atari Mouse
378 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
379 EzKey and similar keyboards
381 atkbd.reset= [HW] Reset keyboard during initialization
383 atkbd.set= [HW] Select keyboard code set
384 Format: <int> (2 = AT (default), 3 = PS/2)
386 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
389 atkbd.softraw= [HW] Choose between synthetic and real raw mode
390 Format: <bool> (0 = real, 1 = synthetic (default))
392 atkbd.softrepeat= [HW]
393 Use software keyboard repeat
395 audit= [KNL] Enable the audit sub-system
396 Format: { "0" | "1" | "off" | "on" }
397 0 | off - kernel audit is disabled and can not be
398 enabled until the next reboot
399 unset - kernel audit is initialized but disabled and
400 will be fully enabled by the userspace auditd.
401 1 | on - kernel audit is initialized and partially
402 enabled, storing at most audit_backlog_limit
403 messages in RAM until it is fully enabled by the
407 audit_backlog_limit= [KNL] Set the audit queue size limit.
408 Format: <int> (must be >=0)
411 bau= [X86_UV] Enable the BAU on SGI UV. The default
412 behavior is to disable the BAU (i.e. bau=0).
413 Format: { "0" | "1" }
416 unset - Disable the BAU.
418 baycom_epp= [HW,AX25]
421 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
423 See header of drivers/net/hamradio/baycom_par.c.
425 baycom_ser_fdx= [HW,AX25]
426 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
427 Format: <io>,<irq>,<mode>[,<baud>]
428 See header of drivers/net/hamradio/baycom_ser_fdx.c.
430 baycom_ser_hdx= [HW,AX25]
431 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
432 Format: <io>,<irq>,<mode>
433 See header of drivers/net/hamradio/baycom_ser_hdx.c.
435 blkdevparts= Manual partition parsing of block device(s) for
436 embedded devices based on command line input.
437 See Documentation/block/cmdline-partition.rst
439 boot_delay= Milliseconds to delay each printk during boot.
440 Values larger than 10 seconds (10000) are changed to
445 Extended command line options can be added to an initrd
446 and this will cause the kernel to look for it.
448 See Documentation/admin-guide/bootconfig.rst
451 Disable BERT OS support on buggy BIOSes.
453 bgrt_disable [ACPI][X86]
454 Disable BGRT to avoid flickering OEM logo.
456 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
457 bttv.radio= Most important insmod options are available as
459 bttv.pll= See Documentation/media/v4l-drivers/bttv.rst
462 bulk_remove=off [PPC] This parameter disables the use of the pSeries
463 firmware feature for flushing multiple hpte entries
466 c101= [NET] Moxa C101 synchronous serial card
468 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
469 Sometimes CPU hardware bugs make them report the cache
470 size incorrectly. The kernel will attempt work arounds
471 to fix known problems, but for some CPUs it is not
472 possible to determine what the correct size should be.
473 This option provides an override for these situations.
476 [NET] Specifies amount of time (in seconds) that
477 the kernel should wait for a network carrier. By default
478 it waits 120 seconds.
480 ca_keys= [KEYS] This parameter identifies a specific key(s) on
481 the system trusted keyring to be used for certificate
483 format: { id:<keyid> | builtin }
485 cca= [MIPS] Override the kernel pages' cache coherency
486 algorithm. Accepted values range from 0 to 7
487 inclusive. See arch/mips/include/asm/pgtable-bits.h
488 for platform specific values (SB1, Loongson3 and
491 ccw_timeout_log [S390]
492 See Documentation/s390/common_io.rst for details.
494 cgroup_disable= [KNL] Disable a particular controller
495 Format: {name of the controller(s) to disable}
496 The effects of cgroup_disable=foo are:
497 - foo isn't auto-mounted if you mount all cgroups in
499 - foo isn't visible as an individually mountable
501 {Currently only "memory" controller deal with this and
502 cut the overhead, others just disable the usage. So
503 only cgroup_disable=memory is actually worthy}
505 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
506 Format: { { controller | "all" | "named" }
507 [,{ controller | "all" | "named" }...] }
508 Like cgroup_disable, but only applies to cgroup v1;
509 the blacklisted controllers remain available in cgroup2.
510 "all" blacklists all controllers and "named" disables
511 named mounts. Specifying both "all" and "named" disables
514 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
516 nosocket -- Disable socket memory accounting.
517 nokmem -- Disable kernel memory accounting.
519 checkreqprot [SELINUX] Set initial checkreqprot flag value.
520 Format: { "0" | "1" }
521 See security/selinux/Kconfig help text.
522 0 -- check protection applied by kernel (includes
523 any implied execute protection).
524 1 -- check protection requested by application.
525 Default value is set via a kernel config option.
526 Value can be changed at runtime via
527 /sys/fs/selinux/checkreqprot.
528 Setting checkreqprot to 1 is deprecated.
531 See Documentation/s390/common_io.rst for details.
534 Prevents the clock framework from automatically gating
535 clocks that have not been explicitly enabled by a Linux
536 device driver but are enabled in hardware at reset or
537 by the bootloader/firmware. Note that this does not
538 force such clocks to be always-on nor does it reserve
539 those clocks in any way. This parameter is useful for
540 debug and development, but should not be needed on a
541 platform with proper driver support. For more
542 information, see Documentation/driver-api/clk.rst.
544 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
546 Forces specified clocksource (if available) to be used
547 when calculating gettimeofday(). If specified
548 clocksource is not available, it defaults to PIT.
549 Format: { pit | tsc | cyclone | pmtmr }
551 clocksource= Override the default clocksource
553 Override the default clocksource and use the clocksource
554 with the name specified.
555 Some clocksource names to choose from, depending on
557 [all] jiffies (this is the base, fallback clocksource)
559 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
560 pxa_timer,timer3,32k_counter,timer0_1
561 [X86-32] pit,hpet,tsc;
562 scx200_hrt on Geode; cyclone on IBM x440
570 clocksource.arm_arch_timer.evtstrm=
573 Enable/disable the eventstream feature of the ARM
574 architected timer so that code using WFE-based polling
575 loops can be debugged more effectively on production
578 clearcpuid=BITNUM [X86]
579 Disable CPUID feature X for the kernel. See
580 arch/x86/include/asm/cpufeatures.h for the valid bit
581 numbers. Note the Linux specific bits are not necessarily
582 stable over kernel options, but the vendor specific
584 Also note that user programs calling CPUID directly
585 or using the feature without checking anything
586 will still see it. This just prevents it from
587 being used by the kernel or shown in /proc/cpuinfo.
588 Also note the kernel might malfunction if you disable
591 cma=nn[MG]@[start[MG][-end[MG]]]
593 Sets the size of kernel global memory area for
594 contiguous memory allocations and optionally the
595 placement constraint by the physical address range of
596 memory allocations. A value of 0 disables CMA
597 altogether. For more information, see
598 include/linux/dma-contiguous.h
600 cmo_free_hint= [PPC] Format: { yes | no }
601 Specify whether pages are marked as being inactive
602 when they are freed. This is used in CMO environments
603 to determine OS memory pressure for page stealing by
607 coherent_pool=nn[KMG] [ARM,KNL]
608 Sets the size of memory pool for coherent, atomic dma
609 allocations, by default set to 256K.
611 com20020= [HW,NET] ARCnet - COM20020 chipset
613 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
615 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
619 ARCnet - COM90xx chipset (memory-mapped buffers)
620 Format: <io>[,<irq>[,<memstart>]]
622 condev= [HW,S390] console device
625 console= [KNL] Output console device and options.
627 tty<n> Use the virtual console device <n>.
631 Use the specified serial port. The options are of
632 the form "bbbbpnf", where "bbbb" is the baud rate,
633 "p" is parity ("n", "o", or "e"), "n" is number of
634 bits, and "f" is flow control ("r" for RTS or
635 omit it). Default is "9600n8".
637 See Documentation/admin-guide/serial-console.rst for more
639 Documentation/networking/netconsole.txt for an
642 uart[8250],io,<addr>[,options]
643 uart[8250],mmio,<addr>[,options]
644 uart[8250],mmio16,<addr>[,options]
645 uart[8250],mmio32,<addr>[,options]
646 uart[8250],0x<addr>[,options]
647 Start an early, polled-mode console on the 8250/16550
648 UART at the specified I/O port or MMIO address,
649 switching to the matching ttyS device later.
650 MMIO inter-register address stride is either 8-bit
651 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
652 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
653 to be equivalent to 'mmio'. 'options' are specified in
654 the same format described for ttyS above; if unspecified,
655 the h/w is not re-initialized.
657 hvc<n> Use the hypervisor console device <n>. This is for
658 both Xen and PowerPC hypervisors.
660 If the device connected to the port is not a TTY but a braille
661 device, prepend "brl," before the device type, for instance
663 For now, only VisioBraille is supported.
666 [KNL] Change console messages format
668 By default we print messages on consoles in
669 "[time stamp] text\n" format (time stamp may not be
670 printed, depending on CONFIG_PRINTK_TIME or
671 `printk_time' param).
673 Switch to syslog format: "<%u>[time stamp] text\n"
674 IOW, each message will have a facility and loglevel
675 prefix. The format is similar to one used by syslog()
676 syscall, or to executing "dmesg -S --raw" or to reading
679 consoleblank= [KNL] The console blank (screen saver) timeout in
680 seconds. A value of 0 disables the blank timer.
684 [KNL] Change the default value for
685 /proc/<pid>/coredump_filter.
686 See also Documentation/filesystems/proc.txt.
688 coresight_cpu_debug.enable
691 Enable/disable the CPU sampling based debugging.
692 0: default value, disable debugging
693 1: enable debugging at boot time
695 cpuidle.off=1 [CPU_IDLE]
696 disable the cpuidle sub-system
699 [CPU_IDLE] Name of the cpuidle governor to use.
701 cpufreq.off=1 [CPU_FREQ]
702 disable the cpufreq sub-system
705 [X86] Delay for N microsec between assert and de-assert
706 of APIC INIT to start processors. This delay occurs
707 on every CPU online, such as boot, and resume from suspend.
710 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
712 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
714 crashkernel=size[KMG][@offset[KMG]]
715 [KNL] Using kexec, Linux can switch to a 'crash kernel'
716 upon panic. This parameter reserves the physical
717 memory region [offset, offset + size] for that kernel
718 image. If '@offset' is omitted, then a suitable offset
719 is selected automatically.
720 [KNL, x86_64] select a region under 4G first, and
721 fall back to reserve region above 4G when '@offset'
722 hasn't been specified.
723 See Documentation/admin-guide/kdump/kdump.rst for further details.
725 crashkernel=range1:size1[,range2:size2,...][@offset]
726 [KNL] Same as above, but depends on the memory
727 in the running system. The syntax of range is
728 start-[end] where start and end are both
729 a memory unit (amount[KMG]). See also
730 Documentation/admin-guide/kdump/kdump.rst for an example.
732 crashkernel=size[KMG],high
733 [KNL, x86_64] range could be above 4G. Allow kernel
734 to allocate physical memory region from top, so could
735 be above 4G if system have more than 4G ram installed.
736 Otherwise memory region will be allocated below 4G, if
738 It will be ignored if crashkernel=X is specified.
739 crashkernel=size[KMG],low
740 [KNL, x86_64] range under 4G. When crashkernel=X,high
741 is passed, kernel could allocate physical memory region
742 above 4G, that cause second kernel crash on system
743 that require some amount of low memory, e.g. swiotlb
744 requires at least 64M+32K low memory, also enough extra
745 low memory is needed to make sure DMA buffers for 32-bit
746 devices won't run out. Kernel would try to allocate at
747 at least 256M below 4G automatically.
748 This one let user to specify own low range under 4G
749 for second kernel instead.
750 0: to disable low allocation.
751 It will be ignored when crashkernel=X,high is not used
752 or memory reserved is below 4G.
755 [KNL] Disable crypto self-tests
760 cs89x0_media= [HW,NET]
761 Format: { rj45 | aui | bnc }
764 See header of drivers/s390/block/dasd_devmap.c.
766 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
767 (one device per port)
768 Format: <port#>,<type>
769 See also Documentation/input/devices/joystick-parport.rst
771 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
773 Documentation/admin-guide/dynamic-debug-howto.rst for
774 details. Deprecated, see dyndbg.
776 debug [KNL] Enable kernel debugging (events log level).
779 [KNL] Enable printing [hashed] pointers early in the
780 boot sequence. If enabled, we use a weak hash instead
781 of siphash to hash pointers. Use this option if you are
782 seeing instances of '(___ptrval___)') and need to see a
783 value (hashed pointer) instead. Cryptographically
784 insecure, please do not use on production kernels.
787 [KNL] verbose self-tests
789 Print debugging info while doing the locking API
791 We default to 0 (no extra messages), setting it to
792 1 will print _a lot_ more information - normally
793 only useful to kernel developers.
795 debug_objects [KNL] Enable object debugging
798 [KNL] Disable object debugging
800 debug_guardpage_minorder=
801 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
802 parameter allows control of the order of pages that will
803 be intentionally kept free (and hence protected) by the
804 buddy allocator. Bigger value increase the probability
805 of catching random memory corruption, but reduce the
806 amount of memory for normal system use. The maximum
807 possible value is MAX_ORDER/2. Setting this parameter
808 to 1 or 2 should be enough to identify most random
809 memory corruption problems caused by bugs in kernel or
810 driver code when a CPU writes to (or reads from) a
811 random memory location. Note that there exists a class
812 of memory corruptions problems caused by buggy H/W or
813 F/W or by drivers badly programing DMA (basically when
814 memory is written at bus level and the CPU MMU is
815 bypassed) which are not detectable by
816 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
817 tracking down these problems.
820 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
821 enables the feature at boot time. By default, it is
822 disabled and the system will work mostly the same as a
823 kernel built without CONFIG_DEBUG_PAGEALLOC.
824 Note: to get most of debug_pagealloc error reports, it's
825 useful to also enable the page_owner functionality.
826 on: enable the feature
828 debugpat [X86] Enable PAT debugging
830 decnet.addr= [HW,NET]
831 Format: <area>[,<node>]
832 See also Documentation/networking/decnet.txt.
835 [same as hugepagesz=] The size of the default
836 HugeTLB page size. This is the size represented by
837 the legacy /proc/ hugepages APIs, used for SHM, and
838 default size when mounting hugetlbfs filesystems.
839 Defaults to the default architecture's huge page size
842 deferred_probe_timeout=
843 [KNL] Debugging option to set a timeout in seconds for
844 deferred probe to give up waiting on dependencies to
845 probe. Only specific dependencies (subsystems or
846 drivers) that have opted in will be ignored. A timeout of 0
847 will timeout at the end of initcalls. This option will also
848 dump out devices still on the deferred probe list after
852 Format: { on | off | def_only | inf_only | always }
853 on: s390 zlib hardware support for compression on
854 level 1 and decompression (default)
855 off: No s390 zlib hardware support
856 def_only: s390 zlib hardware support for deflate
857 only (compression on level 1)
858 inf_only: s390 zlib hardware support for inflate
860 always: Same as 'on' but ignores the selected compression
861 level always using hardware support (used for debugging)
864 Set number of hash buckets for dentry cache.
866 disable_1tb_segments [PPC]
867 Disables the use of 1TB hash page table segments. This
868 causes the kernel to fall back to 256MB segments which
869 can be useful when debugging issues that require an SLB
873 See Documentation/networking/ipv6.txt.
876 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
877 hardening is enabled for this boot. Hardened
878 usercopy checking is used to protect the kernel
879 from reading or writing beyond known memory
880 allocation boundaries as a proactive defense
881 against bounds-checking flaws in the kernel's
882 copy_to_user()/copy_from_user() interface.
883 on Perform hardened usercopy checks (default).
884 off Disable hardened usercopy checks.
887 Disable RADIX MMU mode on POWER9
890 Disable TLBIE instruction. Currently does not work
891 with KVM, with HASH MMU, or with coherent accelerators.
893 disable_cpu_apicid= [X86,APIC,SMP]
895 The number of initial APIC ID for the
896 corresponding CPU to be disabled at boot,
897 mostly used for the kdump 2nd kernel to
898 disable BSP to wake up multiple CPUs without
899 causing system reset or hang due to sending
902 perf_v4_pmi= [X86,INTEL]
904 Disable Intel PMU counter freezing feature.
905 The feature only exists starting from
906 Arch Perfmon v4 (Skylake and newer).
908 disable_ddw [PPC/PSERIES]
909 Disable Dynamic DMA Window support. Use this if
910 to workaround buggy firmware.
913 See Documentation/networking/ipv6.txt.
915 disable_mtrr_cleanup [X86]
916 The kernel tries to adjust MTRR layout from continuous
917 to discrete, to make X server driver able to add WB
918 entry later. This parameter disables that.
920 disable_mtrr_trim [X86, Intel and AMD only]
921 By default the kernel will trim any uncacheable
922 memory out of your available memory pool based on
923 MTRR settings. This parameter disables that behavior,
924 possibly causing your machine to run very slowly.
926 disable_timer_pin_1 [X86]
927 Disable PIN 1 of APIC timer
928 Can be useful to work around chipset bugs.
930 dis_ucode_ldr [X86] Disable the microcode loader.
932 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
933 this option disables the debugging code at boot.
935 dma_debug_entries=<number>
936 This option allows to tune the number of preallocated
937 entries for DMA-API debugging code. One entry is
938 required per DMA-API allocation. Use this if the
939 DMA-API debugging code disables itself because the
940 architectural default is too low.
942 dma_debug_driver=<driver_name>
943 With this option the DMA-API debugging driver
944 filter feature can be enabled at boot time. Just
945 pass the driver to filter for as the parameter.
946 The filter can be disabled or changed to another
947 driver later using sysfs.
949 driver_async_probe= [KNL]
950 List of driver names to be probed asynchronously.
951 Format: <driver_name1>,<driver_name2>...
953 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
954 Broken monitors, graphic adapters, KVMs and EDIDless
955 panels may send no or incorrect EDID data sets.
956 This parameter allows to specify an EDID data sets
957 in the /lib/firmware directory that are used instead.
958 Generic built-in EDID data sets are used, if one of
959 edid/1024x768.bin, edid/1280x1024.bin,
960 edid/1680x1050.bin, or edid/1920x1080.bin is given
961 and no file with the same name exists. Details and
962 instructions how to build your own EDID data are
963 available in Documentation/driver-api/edid.rst. An EDID
964 data set will only be used for a particular connector,
965 if its name and a colon are prepended to the EDID
966 name. Each connector may use a unique EDID data
967 set by separating the files with a comma. An EDID
968 data set with no connector name will be used for
969 any connectors not explicitly specified.
974 Format: {"off" | "known"}
975 Control how the dt_cpu_ftrs device-tree binding is
976 used for CPU feature discovery and setup (if it
978 off: Do not use it, fall back to legacy cpu table.
979 known: Do not pass through unknown features to guests
980 or userspace, only those that the kernel is aware of.
982 dump_apple_properties [X86]
983 Dump name and content of EFI device properties on
984 x86 Macs. Useful for driver authors to determine
985 what data is available or for reverse-engineering.
987 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
988 module.dyndbg[="val"]
989 Enable debug messages at boot time. See
990 Documentation/admin-guide/dynamic-debug-howto.rst
993 nompx [X86] Disables Intel Memory Protection Extensions.
994 See Documentation/x86/intel_mpx.rst for more
995 information about the feature.
997 nopku [X86] Disable Memory Protection Keys CPU feature found
1000 module.async_probe [KNL]
1001 Enable asynchronous probe on this module.
1003 early_ioremap_debug [KNL]
1004 Enable debug messages in early_ioremap support. This
1005 is useful for tracking down temporary early mappings
1006 which are not unmapped.
1008 earlycon= [KNL] Output early console device and options.
1010 When used with no options, the early console is
1011 determined by stdout-path property in device tree's
1012 chosen node or the ACPI SPCR table if supported by
1015 cdns,<addr>[,options]
1016 Start an early, polled-mode console on a Cadence
1017 (xuartps) serial port at the specified address. Only
1018 supported option is baud rate. If baud rate is not
1019 specified, the serial port must already be setup and
1022 uart[8250],io,<addr>[,options]
1023 uart[8250],mmio,<addr>[,options]
1024 uart[8250],mmio32,<addr>[,options]
1025 uart[8250],mmio32be,<addr>[,options]
1026 uart[8250],0x<addr>[,options]
1027 Start an early, polled-mode console on the 8250/16550
1028 UART at the specified I/O port or MMIO address.
1029 MMIO inter-register address stride is either 8-bit
1030 (mmio) or 32-bit (mmio32 or mmio32be).
1031 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1032 to be equivalent to 'mmio'. 'options' are specified
1033 in the same format described for "console=ttyS<n>"; if
1034 unspecified, the h/w is not initialized.
1038 Start an early, polled-mode console on a pl011 serial
1039 port at the specified address. The pl011 serial port
1040 must already be setup and configured. Options are not
1041 yet supported. If 'mmio32' is specified, then only
1042 the driver will use only 32-bit accessors to read/write
1043 the device registers.
1046 Start an early, polled-mode console on a meson serial
1047 port at the specified address. The serial port must
1048 already be setup and configured. Options are not yet
1052 Start an early, polled-mode console on an msm serial
1053 port at the specified address. The serial port
1054 must already be setup and configured. Options are not
1057 msm_serial_dm,<addr>
1058 Start an early, polled-mode console on an msm serial
1059 dm port at the specified address. The serial port
1060 must already be setup and configured. Options are not
1064 Start an early, polled-mode console on a serial port
1065 of an Actions Semi SoC, such as S500 or S900, at the
1066 specified address. The serial port must already be
1067 setup and configured. Options are not yet supported.
1070 Start an early, polled-mode console on a serial port
1071 of an RDA Micro SoC, such as RDA8810PL, at the
1072 specified address. The serial port must already be
1073 setup and configured. Options are not yet supported.
1076 Use RISC-V SBI (Supervisor Binary Interface) for early
1079 smh Use ARM semihosting calls for early console.
1087 Use early console provided by serial driver available
1088 on Samsung SoCs, requires selecting proper type and
1089 a correct base address of the selected UART port. The
1090 serial port must already be setup and configured.
1091 Options are not yet supported.
1094 Start an early, polled-mode console on a lantiq serial
1095 (lqasc) port at the specified address. The serial port
1096 must already be setup and configured. Options are not
1101 Use early console provided by Freescale LP UART driver
1102 found on Freescale Vybrid and QorIQ LS1021A processors.
1103 A valid base address must be provided, and the serial
1104 port must already be setup and configured.
1108 Start an early, polled-mode, output-only console on the
1109 Freescale i.MX UART at the specified address. The UART
1110 must already be setup and configured.
1113 Start an early, polled-mode console on the
1114 Armada 3700 serial port at the specified
1115 address. The serial port must already be setup
1116 and configured. Options are not yet supported.
1119 Start an early, polled-mode console on a Qualcomm
1120 Generic Interface (GENI) based serial port at the
1121 specified address. The serial port must already be
1122 setup and configured. Options are not yet supported.
1125 Start an early, unaccelerated console on the EFI
1126 memory mapped framebuffer (if available). On cache
1127 coherent non-x86 systems that use system memory for
1128 the framebuffer, pass the 'ram' option so that it is
1129 mapped with the correct attributes.
1132 Use early console provided by Freescale LINFlexD UART
1133 serial driver for NXP S32V234 SoCs. A valid base
1134 address must be provided, and the serial port must
1135 already be setup and configured.
1137 earlyprintk= [X86,SH,ARM,M68k,S390]
1141 earlyprintk=serial[,ttySn[,baudrate]]
1142 earlyprintk=serial[,0x...[,baudrate]]
1143 earlyprintk=ttySn[,baudrate]
1144 earlyprintk=dbgp[debugController#]
1145 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1146 earlyprintk=xdbc[xhciController#]
1148 earlyprintk is useful when the kernel crashes before
1149 the normal console is initialized. It is not enabled by
1150 default because it has some cosmetic problems.
1152 Append ",keep" to not disable it when the real console
1155 Only one of vga, efi, serial, or usb debug port can
1158 Currently only ttyS0 and ttyS1 may be specified by
1159 name. Other I/O ports may be explicitly specified
1160 on some architectures (x86 and arm at least) by
1161 replacing ttySn with an I/O port address, like this:
1162 earlyprintk=serial,0x1008,115200
1163 You can find the port for a given device in
1164 /proc/tty/driver/serial:
1165 2: uart:ST16650V2 port:00001008 irq:18 ...
1167 Interaction with the standard serial driver is not
1170 The VGA and EFI output is eventually overwritten by
1173 The xen output can only be used by Xen PV guests.
1175 The sclp output can only be used on s390.
1177 The optional "force" to "pciserial" enables use of a
1178 PCI device even when its classcode is not of the
1181 edac_report= [HW,EDAC] Control how to report EDAC event
1182 Format: {"on" | "off" | "force"}
1183 on: enable EDAC to report H/W event. May be overridden
1184 by other higher priority error reporting module.
1185 off: disable H/W event reporting through EDAC.
1186 force: enforce the use of EDAC to report H/W event.
1189 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1192 This is designed to be used in conjunction with
1193 the boot argument: earlyprintk=vga
1196 Format: {"off" | "on" | "skip[mbr]"}
1199 Format: { "old_map", "nochunk", "noruntime", "debug",
1200 "nosoftreserve", "disable_early_pci_dma",
1201 "no_disable_early_pci_dma" }
1202 old_map [X86-64]: switch to the old ioremap-based EFI
1203 runtime services mapping. [Needs CONFIG_X86_UV=y]
1204 nochunk: disable reading files in "chunks" in the EFI
1205 boot stub, as chunking can cause problems with some
1206 firmware implementations.
1207 noruntime : disable EFI runtime services support
1208 debug: enable misc debug output
1209 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1210 attribute may cause the kernel to reserve the
1211 memory range for a memory mapping driver to
1212 claim. Specify efi=nosoftreserve to disable this
1213 reservation and treat the memory by its base type
1214 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1215 disable_early_pci_dma: Disable the busmaster bit on all
1216 PCI bridges while in the EFI boot stub
1217 no_disable_early_pci_dma: Leave the busmaster bit set
1218 on all PCI bridges while in the EFI boot stub
1220 efi_no_storage_paranoia [EFI; X86]
1221 Using this parameter you can use more than 50% of
1222 your efi variable storage. Use this parameter only if
1223 you are really sure that your UEFI does sane gc and
1224 fulfills the spec otherwise your board may brick.
1226 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1227 Add arbitrary attribute to specific memory range by
1228 updating original EFI memory map.
1229 Region of memory which aa attribute is added to is
1232 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1233 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1234 attribute is added to range 0x100000000-0x180000000 and
1235 0x10a0000000-0x1120000000.
1237 If efi_fake_mem=8G@9G:0x40000 is specified, the
1238 EFI_MEMORY_SP(0x40000) attribute is added to
1239 range 0x240000000-0x43fffffff.
1241 Using this parameter you can do debugging of EFI memmap
1242 related features. For example, you can do debugging of
1243 Address Range Mirroring feature even if your box
1244 doesn't support it, or mark specific memory as
1247 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1248 that is to be dynamically loaded by Linux. If there are
1249 multiple variables with the same name but with different
1250 vendor GUIDs, all of them will be loaded. See
1251 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1254 eisa_irq_edge= [PARISC,HW]
1255 See header of drivers/parisc/eisa.c.
1258 See comment before function elanfreq_setup() in
1259 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1261 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1262 Specifies physical address of start of kernel core
1263 image elf header and optionally the size. Generally
1264 kexec loader will pass this option to capture kernel.
1265 See Documentation/admin-guide/kdump/kdump.rst for details.
1267 enable_mtrr_cleanup [X86]
1268 The kernel tries to adjust MTRR layout from continuous
1269 to discrete, to make X server driver able to add WB
1270 entry later. This parameter enables that.
1272 enable_timer_pin_1 [X86]
1273 Enable PIN 1 of APIC timer
1274 Can be useful to work around chipset bugs
1275 (in particular on some ATI chipsets).
1276 The kernel tries to set a reasonable default.
1278 enforcing [SELINUX] Set initial enforcing status.
1280 See security/selinux/Kconfig help text.
1281 0 -- permissive (log only, no denials).
1282 1 -- enforcing (deny and log).
1284 Value can be changed at runtime via
1285 /sys/fs/selinux/enforce.
1288 Disable Error Record Serialization Table (ERST)
1291 ether= [HW,NET] Ethernet cards parameters
1292 This option is obsoleted by the "netdev=" option, which
1293 has equivalent usage. See its documentation for details.
1297 Permit 'security.evm' to be updated regardless of
1298 current integrity status.
1302 fail_make_request=[KNL]
1303 General fault injection mechanism.
1304 Format: <interval>,<probability>,<space>,<times>
1305 See also Documentation/fault-injection/.
1308 See Documentation/admin-guide/blockdev/floppy.rst.
1310 force_pal_cache_flush
1311 [IA-64] Avoid check_sal_cache_flush which may hang on
1312 buggy SAL_CACHE_FLUSH implementations. Using this
1313 parameter will force ia64_sal_cache_flush to call
1314 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1317 Forcefully enable Physical Address Extension (PAE).
1318 Many Pentium M systems disable PAE but may have a
1319 functionally usable PAE implementation.
1320 Warning: use of this parameter will taint the kernel
1321 and may cause unknown problems.
1324 [FTRACE] will set and start the specified tracer
1325 as early as possible in order to facilitate early
1328 ftrace_dump_on_oops[=orig_cpu]
1329 [FTRACE] will dump the trace buffers on oops.
1330 If no parameter is passed, ftrace will dump
1331 buffers of all CPUs, but if you pass orig_cpu, it will
1332 dump only the buffer of the CPU that triggered the
1335 ftrace_filter=[function-list]
1336 [FTRACE] Limit the functions traced by the function
1337 tracer at boot up. function-list is a comma separated
1338 list of functions. This list can be changed at run
1339 time by the set_ftrace_filter file in the debugfs
1342 ftrace_notrace=[function-list]
1343 [FTRACE] Do not trace the functions specified in
1344 function-list. This list can be changed at run time
1345 by the set_ftrace_notrace file in the debugfs
1348 ftrace_graph_filter=[function-list]
1349 [FTRACE] Limit the top level callers functions traced
1350 by the function graph tracer at boot up.
1351 function-list is a comma separated list of functions
1352 that can be changed at run time by the
1353 set_graph_function file in the debugfs tracing directory.
1355 ftrace_graph_notrace=[function-list]
1356 [FTRACE] Do not trace from the functions specified in
1357 function-list. This list is a comma separated list of
1358 functions that can be changed at run time by the
1359 set_graph_notrace file in the debugfs tracing directory.
1361 ftrace_graph_max_depth=<uint>
1362 [FTRACE] Used with the function graph tracer. This is
1363 the max depth it will trace into a function. This value
1364 can be changed at run time by the max_graph_depth file
1365 in the tracefs tracing directory. default: 0 (no limit)
1367 fw_devlink= [KNL] Create device links between consumer and supplier
1368 devices by scanning the firmware to infer the
1369 consumer/supplier relationships. This feature is
1370 especially useful when drivers are loaded as modules as
1371 it ensures proper ordering of tasks like device probing
1372 (suppliers first, then consumers), supplier boot state
1373 clean up (only after all consumers have probed),
1374 suspend/resume & runtime PM (consumers first, then
1376 Format: { off | permissive | on | rpm }
1377 off -- Don't create device links from firmware info.
1378 permissive -- Create device links from firmware info
1379 but use it only for ordering boot state clean
1380 up (sync_state() calls).
1381 on -- Create device links from firmware info and use it
1382 to enforce probe and suspend/resume ordering.
1383 rpm -- Like "on", but also use to order runtime PM.
1386 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1387 support via parallel port (up to 5 devices per port)
1388 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1389 See also Documentation/input/devices/joystick-parport.rst
1393 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1397 gcov_persist= [GCOV] When non-zero (default), profiling data for
1398 kernel modules is saved and remains accessible via
1399 debugfs, even when the module is unloaded/reloaded.
1400 When zero, profiling data is discarded and associated
1401 debugfs files are removed at module unload time.
1403 goldfish [X86] Enable the goldfish android emulator platform.
1404 Don't use this when you are not running on the
1407 gpt [EFI] Forces disk with valid GPT signature but
1408 invalid Protective MBR to be treated as GPT. If the
1409 primary GPT is corrupted, it enables the backup/alternate
1410 GPT to be used instead.
1412 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1413 the "Enable 0" bit of the configuration register.
1416 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1417 the "Enable 0" bit of the configuration register.
1420 grcan.select= [HW] Select which physical interface to use.
1423 grcan.txsize= [HW] Sets the size of the tx buffer.
1424 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1426 grcan.rxsize= [HW] Sets the size of the rx buffer.
1427 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1430 gpio-mockup.gpio_mockup_ranges
1431 [HW] Sets the ranges of gpiochip of for this device.
1432 Format: <start1>,<end1>,<start2>,<end2>...
1434 hardlockup_all_cpu_backtrace=
1435 [KNL] Should the hard-lockup detector generate
1436 backtraces on all cpus.
1439 hashdist= [KNL,NUMA] Large hashes allocated during boot
1440 are distributed across NUMA nodes. Defaults on
1441 for 64-bit NUMA, off otherwise.
1442 Format: 0 | 1 (for off | on)
1444 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1446 hd= [EIDE] (E)IDE hard drive subsystem geometry
1447 Format: <cyl>,<head>,<sect>
1450 Disable Hardware Error Source Table (HEST) support;
1451 corresponding firmware-first mode error processing
1452 logic will be disabled.
1454 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1455 size of <nn>. This works even on boxes that have no
1456 highmem otherwise. This also works to reduce highmem
1457 size on bigger boxes.
1459 highres= [KNL] Enable/disable high resolution timer mode.
1460 Valid parameters: "on", "off"
1465 hpet= [X86-32,HPET] option to control HPET usage
1466 Format: { enable (default) | disable | force |
1468 disable: disable HPET and use PIT instead
1469 force: allow force enabled of undocumented chips (ICH4,
1471 verbose: show contents of HPET registers during setup
1473 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1474 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1476 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1477 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1478 On x86-64 and powerpc, this option can be specified
1479 multiple times interleaved with hugepages= to reserve
1480 huge pages of different sizes. Valid pages sizes on
1481 x86-64 are 2M (when the CPU supports "pse") and 1G
1482 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1485 [KNL] Should the hung task detector generate panics.
1488 A nonzero value instructs the kernel to panic when a
1489 hung task is detected. The default value is controlled
1490 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1491 option. The value selected by this boot parameter can
1492 be changed later by the kernel.hung_task_panic sysctl.
1494 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1495 terminal devices. Valid values: 0..8
1496 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1497 If specified, z/VM IUCV HVC accepts connections
1498 from listed z/VM user IDs only.
1500 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1501 which allow the hypervisor to 'idle' the
1502 guest on lock contention.
1505 Do not unregister boot console at start. This is only
1506 useful for debugging when something happens in the window
1507 between unregistering the boot console and initializing
1510 i2c_bus= [HW] Override the default board specific I2C bus speed
1511 or register an additional I2C bus that is not
1512 registered from board initialization code.
1516 i8042.debug [HW] Toggle i8042 debug mode
1517 i8042.unmask_kbd_data
1518 [HW] Enable printing of interrupt data from the KBD port
1519 (disabled by default, and as a pre-condition
1520 requires that i8042.debug=1 be enabled)
1521 i8042.direct [HW] Put keyboard port into non-translated mode
1522 i8042.dumbkbd [HW] Pretend that controller can only read data from
1523 keyboard and cannot control its state
1524 (Don't attempt to blink the leds)
1525 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1526 i8042.nokbd [HW] Don't check/create keyboard port
1527 i8042.noloop [HW] Disable the AUX Loopback command while probing
1529 i8042.nomux [HW] Don't check presence of an active multiplexing
1531 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1533 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1534 i8042.reset [HW] Reset the controller during init, cleanup and
1535 suspend-to-ram transitions, only during s2r
1536 transitions, or never reset
1537 Format: { 1 | Y | y | 0 | N | n }
1538 1, Y, y: always reset controller
1539 0, N, n: don't ever reset controller
1540 Default: only on s2r transitions on x86; most other
1541 architectures force reset to be always executed
1542 i8042.unlock [HW] Unlock (ignore) the keylock
1543 i8042.kbdreset [HW] Reset device connected to KBD port
1547 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1548 indicates that the driver is running on unsupported
1550 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1551 does not match list of supported models.
1553 [HW] Report power status in /proc/i8k
1554 (disabled by default)
1555 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1558 i915.invert_brightness=
1559 [DRM] Invert the sense of the variable that is used to
1560 set the brightness of the panel backlight. Normally a
1561 brightness value of 0 indicates backlight switched off,
1562 and the maximum of the brightness value sets the backlight
1563 to maximum brightness. If this parameter is set to 0
1564 (default) and the machine requires it, or this parameter
1565 is set to 1, a brightness value of 0 sets the backlight
1566 to maximum brightness, and the maximum of the brightness
1567 value switches the backlight off.
1568 -1 -- never invert brightness
1569 0 -- machine default
1570 1 -- force brightness inversion
1573 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1575 ide-core.nodma= [HW] (E)IDE subsystem
1576 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1577 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1578 .cdrom .chs .ignore_cable are additional options
1579 See Documentation/ide/ide.rst.
1581 ide-generic.probe-mask= [HW] (E)IDE subsystem
1583 Probe mask for legacy ISA IDE ports. Depending on
1584 platform up to 6 ports are supported, enabled by
1585 setting corresponding bits in the mask to 1. The
1586 default value is 0x0, which has a special meaning.
1587 On systems that have PCI, it triggers scanning the
1588 PCI bus for the first and the second port, which
1589 are then probed. On systems without PCI the value
1590 of 0x0 enables probing the two first ports as if it
1593 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1594 Claim all unknown PCI IDE storage controllers.
1597 Format: idle=poll, idle=halt, idle=nomwait
1598 Poll forces a polling idle loop that can slightly
1599 improve the performance of waking up a idle CPU, but
1600 will use a lot of power and make the system run hot.
1602 idle=halt: Halt is forced to be used for CPU idle.
1603 In such case C2/C3 won't be used again.
1604 idle=nomwait: Disable mwait for CPU C-states
1606 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1607 Format: { strict | legacy | 2008 | relaxed }
1610 Choose which programs will be accepted for execution
1611 based on the IEEE 754 NaN encoding(s) supported by
1612 the FPU and the NaN encoding requested with the value
1613 of an ELF file header flag individually set by each
1614 binary. Hardware implementations are permitted to
1615 support either or both of the legacy and the 2008 NaN
1618 Available settings are as follows:
1619 strict accept binaries that request a NaN encoding
1620 supported by the FPU
1621 legacy only accept legacy-NaN binaries, if supported
1623 2008 only accept 2008-NaN binaries, if supported
1625 relaxed accept any binaries regardless of whether
1626 supported by the FPU
1628 The FPU emulator is always able to support both NaN
1629 encodings, so if no FPU hardware is present or it has
1630 been disabled with 'nofpu', then the settings of
1631 'legacy' and '2008' strap the emulator accordingly,
1632 'relaxed' straps the emulator for both legacy-NaN and
1633 2008-NaN, whereas 'strict' enables legacy-NaN only on
1634 legacy processors and both NaN encodings on MIPS32 or
1637 The setting for ABS.fmt/NEG.fmt instruction execution
1638 mode generally follows that for the NaN encoding,
1639 except where unsupported by hardware.
1641 ignore_loglevel [KNL]
1642 Ignore loglevel setting - this will print /all/
1643 kernel messages to the console. Useful for debugging.
1644 We also add it as printk module parameter, so users
1645 could change it dynamically, usually by
1646 /sys/module/printk/parameters/ignore_loglevel.
1649 Ignore RLIMIT_DATA setting for data mappings,
1650 print warning at first misuse. Can be changed via
1651 /sys/module/kernel/parameters/ignore_rlimit_data.
1653 ihash_entries= [KNL]
1654 Set number of hash buckets for inode cache.
1656 ima_appraise= [IMA] appraise integrity measurements
1657 Format: { "off" | "enforce" | "fix" | "log" }
1660 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1661 The builtin appraise policy appraises all files
1664 ima_canonical_fmt [IMA]
1665 Use the canonical format for the binary runtime
1666 measurements, instead of host native format.
1669 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1673 The list of supported hash algorithms is defined
1674 in crypto/hash_info.h.
1677 The builtin policies to load during IMA setup.
1678 Format: "tcb | appraise_tcb | secure_boot |
1681 The "tcb" policy measures all programs exec'd, files
1682 mmap'd for exec, and all files opened with the read
1683 mode bit set by either the effective uid (euid=0) or
1686 The "appraise_tcb" policy appraises the integrity of
1687 all files owned by root.
1689 The "secure_boot" policy appraises the integrity
1690 of files (eg. kexec kernel image, kernel modules,
1691 firmware, policy, etc) based on file signatures.
1693 The "fail_securely" policy forces file signature
1694 verification failure also on privileged mounted
1695 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1698 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1699 Load a policy which meets the needs of the Trusted
1700 Computing Base. This means IMA will measure all
1701 programs exec'd, files mmap'd for exec, and all files
1702 opened for read by uid=0.
1705 Select one of defined IMA measurements template formats.
1706 Formats: { "ima" | "ima-ng" | "ima-sig" }
1710 [IMA] Define a custom template format.
1711 Format: { "field1|...|fieldN" }
1713 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1714 Format: <min_file_size>
1715 Set the minimal file size for using asynchronous hash.
1716 If left unspecified, ahash usage is disabled.
1718 ahash performance varies for different data sizes on
1719 different crypto accelerators. This option can be used
1720 to achieve the best performance for a particular HW.
1722 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1724 Set hashing buffer size. Default: 4k.
1726 ahash performance varies for different chunk sizes on
1727 different crypto accelerators. This option can be used
1728 to achieve best performance for particular HW.
1732 Run specified binary instead of /sbin/init as init
1735 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1736 for working out where the kernel is dying during
1739 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1740 initcall functions. Useful for debugging built-in
1741 modules and initcalls.
1743 initrd= [BOOT] Specify the location of the initial ramdisk
1745 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1748 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1750 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1752 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1754 init_pkru= [x86] Specify the default memory protection keys rights
1755 register contents for all processes. 0x55555554 by
1756 default (disallow access to all but pkey 0). Can
1757 override in debugfs after boot.
1759 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1762 int_pln_enable [x86] Enable power limit notification interrupt
1764 integrity_audit=[IMA]
1765 Format: { "0" | "1" }
1766 0 -- basic integrity auditing messages. (Default)
1767 1 -- additional integrity auditing messages.
1769 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1771 Enable intel iommu driver.
1773 Disable intel iommu driver.
1774 igfx_off [Default Off]
1775 By default, gfx is mapped as normal device. If a gfx
1776 device has a dedicated DMAR unit, the DMAR unit is
1777 bypassed by not enabling DMAR with this option. In
1778 this case, gfx device will use physical address for
1781 With this option iommu will not optimize to look
1782 for io virtual address below 32-bit forcing dual
1783 address cycle on pci bus for cards supporting greater
1784 than 32-bit addressing. The default is to look
1785 for translation below 32-bit and if not available
1786 then look in the higher range.
1787 strict [Default Off]
1788 With this option on every unmap_single operation will
1789 result in a hardware IOTLB flush operation as opposed
1790 to batching them for performance.
1791 sp_off [Default Off]
1792 By default, super page will be supported if Intel IOMMU
1793 has the capability. With this option, super page will
1796 By default, scalable mode will be disabled even if the
1797 hardware advertises that it has support for the scalable
1798 mode translation. With this option set, scalable mode
1799 will be used on hardware which claims to support it.
1800 tboot_noforce [Default Off]
1801 Do not force the Intel IOMMU enabled under tboot.
1802 By default, tboot will force Intel IOMMU on, which
1803 could harm performance of some high-throughput
1804 devices like 40GBit network cards, even if identity
1806 Note that using this option lowers the security
1807 provided by tboot because it makes the system
1808 vulnerable to DMA attacks.
1809 nobounce [Default off]
1810 Disable bounce buffer for untrusted devices such as
1811 the Thunderbolt devices. This will treat the untrusted
1812 devices as the trusted ones, hence might expose security
1813 risks of DMA attacks.
1815 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1816 0 disables intel_idle and fall back on acpi_idle.
1817 1 to 9 specify maximum depth of C-state.
1821 Do not enable intel_pstate as the default
1822 scaling driver for the supported processors
1824 Use intel_pstate as a scaling driver, but configure it
1825 to work with generic cpufreq governors (instead of
1826 enabling its internal governor). This mode cannot be
1827 used along with the hardware-managed P-states (HWP)
1830 Enable intel_pstate on systems that prohibit it by default
1831 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1832 instead of acpi-cpufreq may disable platform features, such
1833 as thermal controls and power capping, that rely on ACPI
1834 P-States information being indicated to OSPM and therefore
1835 should be used with caution. This option does not work with
1836 processors that aren't supported by the intel_pstate driver
1837 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1839 Do not enable hardware P state control (HWP)
1842 Only load intel_pstate on systems which support
1843 hardware P state control (HWP) if available.
1845 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1846 Description Table, specifies preferred power management
1847 profile as "Enterprise Server" or "Performance Server",
1848 then this feature is turned on by default.
1850 Allow per-logical-CPU P-State performance control limits using
1851 cpufreq sysfs interface
1853 intremap= [X86-64, Intel-IOMMU]
1854 on enable Interrupt Remapping (default)
1855 off disable Interrupt Remapping
1856 nosid disable Source ID checking
1858 BIOS x2APIC opt-out request will be ignored
1859 nopost disable Interrupt Posting
1861 iomem= Disable strict checking of access to MMIO memory
1862 strict regions from userspace.
1877 nobypass [PPC/POWERNV]
1878 Disable IOMMU bypass, using IOMMU for PCI devices.
1880 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1881 Format: { "0" | "1" }
1883 Request that DMA unmap operations use deferred
1884 invalidation of hardware TLBs, for increased
1885 throughput at the cost of reduced device isolation.
1886 Will fall back to strict mode if not supported by
1887 the relevant IOMMU driver.
1888 1 - Strict mode (default).
1889 DMA unmap operations invalidate IOMMU hardware TLBs
1893 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
1894 Format: { "0" | "1" }
1895 0 - Use IOMMU translation for DMA.
1896 1 - Bypass the IOMMU for DMA.
1897 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
1899 io7= [HW] IO7 for Marvel based alpha systems
1900 See comment before marvel_specify_io7 in
1901 arch/alpha/kernel/core_marvel.c.
1903 io_delay= [X86] I/O delay method
1905 Standard port 0x80 based delay
1907 Alternate port 0xed based delay (needed on some systems)
1909 Simple two microseconds delay
1914 See Documentation/admin-guide/nfs/nfsroot.rst.
1916 ipcmni_extend [KNL] Extend the maximum number of unique System V
1917 IPC identifiers from 32,768 to 16,777,216.
1919 irqaffinity= [SMP] Set the default irq affinity mask
1920 The argument is a cpu list, as described above.
1922 irqchip.gicv2_force_probe=
1925 Force the kernel to look for the second 4kB page
1926 of a GICv2 controller even if the memory range
1927 exposed by the device tree is too small.
1929 irqchip.gicv3_nolpi=
1931 Force the kernel to ignore the availability of
1932 LPIs (and by consequence ITSs). Intended for system
1933 that use the kernel as a bootloader, and thus want
1934 to let secondary kernels in charge of setting up
1937 irqchip.gicv3_pseudo_nmi= [ARM64]
1938 Enables support for pseudo-NMIs in the kernel. This
1939 requires the kernel to be built with
1940 CONFIG_ARM64_PSEUDO_NMI.
1943 When an interrupt is not handled search all handlers
1944 for it. Intended to get systems with badly broken
1948 When an interrupt is not handled search all handlers
1949 for it. Also check all handlers each timer
1950 interrupt. Intended to get systems with badly broken
1954 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1956 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1957 [Deprecated - use cpusets instead]
1958 Format: [flag-list,]<cpu-list>
1960 Specify one or more CPUs to isolate from disturbances
1961 specified in the flag list (default: domain):
1964 Disable the tick when a single task runs.
1966 A residual 1Hz tick is offloaded to workqueues, which you
1967 need to affine to housekeeping through the global
1968 workqueue's affinity configured via the
1969 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1970 by using the 'domain' flag described below.
1972 NOTE: by default the global workqueue runs on all CPUs,
1973 so to protect individual CPUs the 'cpumask' file has to
1974 be configured manually after bootup.
1977 Isolate from the general SMP balancing and scheduling
1978 algorithms. Note that performing domain isolation this way
1979 is irreversible: it's not possible to bring back a CPU to
1980 the domains once isolated through isolcpus. It's strongly
1981 advised to use cpusets instead to disable scheduler load
1982 balancing through the "cpuset.sched_load_balance" file.
1983 It offers a much more flexible interface where CPUs can
1984 move in and out of an isolated set anytime.
1986 You can move a process onto or off an "isolated" CPU via
1987 the CPU affinity syscalls or cpuset.
1988 <cpu number> begins at 0 and the maximum value is
1989 "number of CPUs in system - 1".
1993 Isolate from being targeted by managed interrupts
1994 which have an interrupt mask containing isolated
1995 CPUs. The affinity of managed interrupts is
1996 handled by the kernel and cannot be changed via
1997 the /proc/irq/* interfaces.
1999 This isolation is best effort and only effective
2000 if the automatically assigned interrupt mask of a
2001 device queue contains isolated and housekeeping
2002 CPUs. If housekeeping CPUs are online then such
2003 interrupts are directed to the housekeeping CPU
2004 so that IO submitted on the housekeeping CPU
2005 cannot disturb the isolated CPU.
2007 If a queue's affinity mask contains only isolated
2008 CPUs then this parameter has no effect on the
2009 interrupt routing decision, though interrupts are
2010 only delivered when tasks running on those
2011 isolated CPUs submit IO. IO submitted on
2012 housekeeping CPUs has no influence on those
2015 The format of <cpu-list> is described above.
2019 ivrs_ioapic [HW,X86_64]
2020 Provide an override to the IOAPIC-ID<->DEVICE-ID
2021 mapping provided in the IVRS ACPI table. For
2022 example, to map IOAPIC-ID decimal 10 to
2023 PCI device 00:14.0 write the parameter as:
2024 ivrs_ioapic[10]=00:14.0
2026 ivrs_hpet [HW,X86_64]
2027 Provide an override to the HPET-ID<->DEVICE-ID
2028 mapping provided in the IVRS ACPI table. For
2029 example, to map HPET-ID decimal 0 to
2030 PCI device 00:14.0 write the parameter as:
2031 ivrs_hpet[0]=00:14.0
2033 ivrs_acpihid [HW,X86_64]
2034 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2035 mapping provided in the IVRS ACPI table. For
2036 example, to map UART-HID:UID AMD0020:0 to
2037 PCI device 00:14.5 write the parameter as:
2038 ivrs_acpihid[00:14.5]=AMD0020:0
2040 js= [HW,JOY] Analog joystick
2041 See Documentation/input/joydev/joystick.rst.
2044 When CONFIG_RANDOMIZE_BASE is set, this disables
2045 kernel and module base offset ASLR (Address Space
2046 Layout Randomization).
2049 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2050 report on every invalid memory access. Without this
2051 parameter KASAN will print report only for the first
2056 kernelcore= [KNL,X86,IA-64,PPC]
2057 Format: nn[KMGTPE] | nn% | "mirror"
2058 This parameter specifies the amount of memory usable by
2059 the kernel for non-movable allocations. The requested
2060 amount is spread evenly throughout all nodes in the
2061 system as ZONE_NORMAL. The remaining memory is used for
2062 movable memory in its own zone, ZONE_MOVABLE. In the
2063 event, a node is too small to have both ZONE_NORMAL and
2064 ZONE_MOVABLE, kernelcore memory will take priority and
2065 other nodes will have a larger ZONE_MOVABLE.
2067 ZONE_MOVABLE is used for the allocation of pages that
2068 may be reclaimed or moved by the page migration
2069 subsystem. Note that allocations like PTEs-from-HighMem
2070 still use the HighMem zone if it exists, and the Normal
2071 zone if it does not.
2073 It is possible to specify the exact amount of memory in
2074 the form of "nn[KMGTPE]", a percentage of total system
2075 memory in the form of "nn%", or "mirror". If "mirror"
2076 option is specified, mirrored (reliable) memory is used
2077 for non-movable allocations and remaining memory is used
2078 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2079 are exclusive, so you cannot specify multiple forms.
2081 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2082 Format: <Controller#>[,poll interval]
2083 The controller # is the number of the ehci usb debug
2084 port as it is probed via PCI. The poll interval is
2085 optional and is the number seconds in between
2086 each poll cycle to the debug port in case you need
2087 the functionality for interrupting the kernel with
2088 gdb or control-c on the dbgp connection. When
2089 not using this parameter you use sysrq-g to break into
2090 the kernel debugger.
2092 kgdboc= [KGDB,HW] kgdb over consoles.
2093 Requires a tty driver that supports console polling,
2094 or a supported polling keyboard driver (non-usb).
2095 Serial only format: <serial_device>[,baud]
2096 keyboard only format: kbd
2097 keyboard and serial format: kbd,<serial_device>[,baud]
2098 Optional Kernel mode setting:
2099 kms, kbd format: kms,kbd
2100 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2102 kgdbwait [KGDB] Stop kernel execution and enter the
2103 kernel debugger at the earliest opportunity.
2105 kmac= [MIPS] korina ethernet MAC address.
2106 Configure the RouterBoard 532 series on-chip
2107 Ethernet adapter MAC address.
2109 kmemleak= [KNL] Boot-time kmemleak enable/disable
2110 Valid arguments: on, off
2112 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2115 kprobe_event=[probe-list]
2116 [FTRACE] Add kprobe events and enable at boot time.
2117 The probe-list is a semicolon delimited list of probe
2118 definitions. Each definition is same as kprobe_events
2119 interface, but the parameters are comma delimited.
2120 For example, to add a kprobe event on vfs_read with
2121 arg1 and arg2, add to the command line;
2123 kprobe_event=p,vfs_read,$arg1,$arg2
2125 See also Documentation/trace/kprobetrace.rst "Kernel
2126 Boot Parameter" section.
2128 kpti= [ARM64] Control page table isolation of user
2129 and kernel address spaces.
2130 Default: enabled on cores which need mitigation.
2134 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2135 Default is 0 (don't ignore, but inject #GP)
2137 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2138 Default is false (don't support).
2140 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2145 [KVM] Controls the software workaround for the
2146 X86_BUG_ITLB_MULTIHIT bug.
2147 force : Always deploy workaround.
2148 off : Never deploy workaround.
2149 auto : Deploy workaround based on the presence of
2150 X86_BUG_ITLB_MULTIHIT.
2154 If the software workaround is enabled for the host,
2155 guests do need not to enable it for nested guests.
2157 kvm.nx_huge_pages_recovery_ratio=
2158 [KVM] Controls how many 4KiB pages are periodically zapped
2159 back to huge pages. 0 disables the recovery, otherwise if
2160 the value is N KVM will zap 1/Nth of the 4KiB pages every
2161 minute. The default is 60.
2163 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2164 Default is 1 (enabled)
2166 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2168 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2170 kvm-arm.vgic_v3_group0_trap=
2171 [KVM,ARM] Trap guest accesses to GICv3 group-0
2174 kvm-arm.vgic_v3_group1_trap=
2175 [KVM,ARM] Trap guest accesses to GICv3 group-1
2178 kvm-arm.vgic_v3_common_trap=
2179 [KVM,ARM] Trap guest accesses to GICv3 common
2182 kvm-arm.vgic_v4_enable=
2183 [KVM,ARM] Allow use of GICv4 for direct injection of
2186 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2187 (virtualized MMU) support on capable Intel chips.
2188 Default is 1 (enabled)
2190 kvm-intel.emulate_invalid_guest_state=
2191 [KVM,Intel] Enable emulation of invalid guest states
2192 Default is 0 (disabled)
2194 kvm-intel.flexpriority=
2195 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2196 Default is 1 (enabled)
2199 [KVM,Intel] Enable VMX nesting (nVMX).
2200 Default is 0 (disabled)
2202 kvm-intel.unrestricted_guest=
2203 [KVM,Intel] Disable unrestricted guest feature
2204 (virtualized real and unpaged mode) on capable
2205 Intel chips. Default is 1 (enabled)
2207 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2210 Valid arguments: never, cond, always
2212 always: L1D cache flush on every VMENTER.
2213 cond: Flush L1D on VMENTER only when the code between
2214 VMEXIT and VMENTER can leak host memory.
2215 never: Disables the mitigation
2217 Default is cond (do L1 cache flush in specific instances)
2219 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2220 feature (tagged TLBs) on capable Intel chips.
2221 Default is 1 (enabled)
2223 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2226 The kernel PTE inversion protection is unconditionally
2227 enabled and cannot be disabled.
2230 Provides all available mitigations for the
2231 L1TF vulnerability. Disables SMT and
2232 enables all mitigations in the
2233 hypervisors, i.e. unconditional L1D flush.
2235 SMT control and L1D flush control via the
2236 sysfs interface is still possible after
2237 boot. Hypervisors will issue a warning
2238 when the first VM is started in a
2239 potentially insecure configuration,
2240 i.e. SMT enabled or L1D flush disabled.
2243 Same as 'full', but disables SMT and L1D
2244 flush runtime control. Implies the
2245 'nosmt=force' command line option.
2246 (i.e. sysfs control of SMT is disabled.)
2249 Leaves SMT enabled and enables the default
2250 hypervisor mitigation, i.e. conditional
2253 SMT control and L1D flush control via the
2254 sysfs interface is still possible after
2255 boot. Hypervisors will issue a warning
2256 when the first VM is started in a
2257 potentially insecure configuration,
2258 i.e. SMT enabled or L1D flush disabled.
2262 Disables SMT and enables the default
2263 hypervisor mitigation.
2265 SMT control and L1D flush control via the
2266 sysfs interface is still possible after
2267 boot. Hypervisors will issue a warning
2268 when the first VM is started in a
2269 potentially insecure configuration,
2270 i.e. SMT enabled or L1D flush disabled.
2273 Same as 'flush', but hypervisors will not
2274 warn when a VM is started in a potentially
2275 insecure configuration.
2278 Disables hypervisor mitigations and doesn't
2280 It also drops the swap size and available
2281 RAM limit restriction on both hypervisor and
2286 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2292 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2295 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2296 value for LAPIC timer one-shot implementation. Default
2297 back to the programmable timer unit in the LAPIC.
2299 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2302 libata.dma= [LIBATA] DMA control
2303 libata.dma=0 Disable all PATA and SATA DMA
2304 libata.dma=1 PATA and SATA Disk DMA only
2305 libata.dma=2 ATAPI (CDROM) DMA only
2306 libata.dma=4 Compact Flash DMA only
2307 Combinations also work, so libata.dma=3 enables DMA
2308 for disks and CDROMs, but not CFs.
2310 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2311 libata.ignore_hpa=0 keep BIOS limits (default)
2312 libata.ignore_hpa=1 ignore limits, using full disk
2314 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2318 libata.force= [LIBATA] Force configurations. The format is comma
2319 separated list of "[ID:]VAL" where ID is
2320 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2321 matching port, link or device. Basically, it matches
2322 the ATA ID string printed on console by libata. If
2323 the whole ID part is omitted, the last PORT and DEVICE
2324 values are used. If ID hasn't been specified yet, the
2325 configuration applies to all ports, links and devices.
2327 If only DEVICE is omitted, the parameter applies to
2328 the port and all links and devices behind it. DEVICE
2329 number of 0 either selects the first device or the
2330 first fan-out link behind PMP device. It does not
2331 select the host link. DEVICE number of 15 selects the
2332 host link and device attached to it.
2334 The VAL specifies the configuration to force. As long
2335 as there's no ambiguity shortcut notation is allowed.
2336 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2337 The following configurations can be forced.
2339 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2340 Any ID with matching PORT is used.
2342 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2344 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2345 udma[/][16,25,33,44,66,100,133] notation is also
2348 * [no]ncq: Turn on or off NCQ.
2350 * [no]ncqtrim: Turn off queued DSM TRIM.
2352 * nohrst, nosrst, norst: suppress hard, soft
2355 * rstonce: only attempt one reset during
2356 hot-unplug link recovery
2358 * dump_id: dump IDENTIFY data.
2360 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2362 * disable: Disable this device.
2364 If there are multiple matching configurations changing
2365 the same attribute, the last one is used.
2367 memblock=debug [KNL] Enable memblock debug messages.
2369 load_ramdisk= [RAM] List of ramdisks to load from floppy
2370 See Documentation/admin-guide/blockdev/ramdisk.rst.
2372 lockd.nlm_grace_period=P [NFS] Assign grace period.
2375 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2378 lockd.nlm_timeout=T [NFS] Assign timeout value.
2381 lockd.nlm_udpport=M [NFS] Assign UDP port.
2384 lockdown= [SECURITY]
2385 { integrity | confidentiality }
2386 Enable the kernel lockdown feature. If set to
2387 integrity, kernel features that allow userland to
2388 modify the running kernel are disabled. If set to
2389 confidentiality, kernel features that allow userland
2390 to extract confidential information from the kernel
2393 locktorture.nreaders_stress= [KNL]
2394 Set the number of locking read-acquisition kthreads.
2395 Defaults to being automatically set based on the
2396 number of online CPUs.
2398 locktorture.nwriters_stress= [KNL]
2399 Set the number of locking write-acquisition kthreads.
2401 locktorture.onoff_holdoff= [KNL]
2402 Set time (s) after boot for CPU-hotplug testing.
2404 locktorture.onoff_interval= [KNL]
2405 Set time (s) between CPU-hotplug operations, or
2406 zero to disable CPU-hotplug testing.
2408 locktorture.shuffle_interval= [KNL]
2409 Set task-shuffle interval (jiffies). Shuffling
2410 tasks allows some CPUs to go into dyntick-idle
2411 mode during the locktorture test.
2413 locktorture.shutdown_secs= [KNL]
2414 Set time (s) after boot system shutdown. This
2415 is useful for hands-off automated testing.
2417 locktorture.stat_interval= [KNL]
2418 Time (s) between statistics printk()s.
2420 locktorture.stutter= [KNL]
2421 Time (s) to stutter testing, for example,
2422 specifying five seconds causes the test to run for
2423 five seconds, wait for five seconds, and so on.
2424 This tests the locking primitive's ability to
2425 transition abruptly to and from idle.
2427 locktorture.torture_type= [KNL]
2428 Specify the locking implementation to test.
2430 locktorture.verbose= [KNL]
2431 Enable additional printk() statements.
2433 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2436 loglevel= All Kernel Messages with a loglevel smaller than the
2437 console loglevel will be printed to the console. It can
2438 also be changed with klogd or other programs. The
2439 loglevels are defined as follows:
2441 0 (KERN_EMERG) system is unusable
2442 1 (KERN_ALERT) action must be taken immediately
2443 2 (KERN_CRIT) critical conditions
2444 3 (KERN_ERR) error conditions
2445 4 (KERN_WARNING) warning conditions
2446 5 (KERN_NOTICE) normal but significant condition
2447 6 (KERN_INFO) informational
2448 7 (KERN_DEBUG) debug-level messages
2450 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2451 in bytes. n must be a power of two and greater
2452 than the minimal size. The minimal size is defined
2453 by LOG_BUF_SHIFT kernel config parameter. There is
2454 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2455 that allows to increase the default size depending on
2456 the number of CPUs. See init/Kconfig for more details.
2458 logo.nologo [FB] Disables display of the built-in Linux logo.
2459 This may be used to provide more screen space for
2460 kernel log messages and is useful when debugging
2461 kernel boot problems.
2463 lp=0 [LP] Specify parallel ports to use, e.g,
2464 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2465 lp=reset first parallel port). 'lp=0' disables the
2466 lp=auto printer driver. 'lp=reset' (which can be
2467 specified in addition to the ports) causes
2468 attached printers to be reset. Using
2469 lp=port1,port2,... specifies the parallel ports
2470 to associate lp devices with, starting with
2471 lp0. A port specification may be 'none' to skip
2472 that lp device, or a parport name such as
2473 'parport0'. Specifying 'lp=auto' instead of a
2474 port specification list means that device IDs
2475 from each port should be examined, to see if
2476 an IEEE 1284-compliant printer is attached; if
2477 so, the driver will manage that printer.
2478 See also header of drivers/char/lp.c.
2481 Sets loops_per_jiffy to given constant, thus avoiding
2482 time-consuming boot-time autodetection (up to 250 ms per
2483 CPU). 0 enables autodetection (default). To determine
2484 the correct value for your kernel, boot with normal
2485 autodetection and see what value is printed. Note that
2486 on SMP systems the preset will be applied to all CPUs,
2487 which is likely to cause problems if your CPUs need
2488 significantly divergent settings. An incorrect value
2489 will cause delays in the kernel to be wrong, leading to
2490 unpredictable I/O errors and other breakage. Although
2491 unlikely, in the extreme case this might damage your
2495 Format: <io>,<irq>,<dma>
2497 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2500 [SECURITY] Choose order of LSM initialization. This
2501 overrides CONFIG_LSM, and the "security=" parameter.
2503 machvec= [IA-64] Force the use of a particular machine-vector
2504 (machvec) in a generic kernel.
2505 Example: machvec=hpzx1
2507 machtype= [Loongson] Share the same kernel image file between different
2509 Example: machtype=lemote-yeeloong-2f-7inch
2511 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2512 than or equal to this physical address is ignored.
2514 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2515 will bring up during bootup. maxcpus=n : n >= 0 limits
2516 the kernel to bring up 'n' processors. Surely after
2517 bootup you can bring up the other plugged cpu by executing
2518 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2519 only takes effect during system bootup.
2520 While n=0 is a special case, it is equivalent to "nosmp",
2521 which also disables the IO APIC.
2523 max_loop= [LOOP] The number of loop block devices that get
2524 (loop.max_loop) unconditionally pre-created at init time. The default
2525 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2526 of statically allocating a predefined number, loop
2527 devices can be requested on-demand with the
2528 /dev/loop-control interface.
2530 mce [X86-32] Machine Check Exception
2532 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2534 md= [HW] RAID subsystems devices and level
2535 See Documentation/admin-guide/md.rst.
2538 Format: <first>,<last>
2539 Specifies range of consoles to be captured by the MDA.
2542 Control mitigation for the Micro-architectural Data
2543 Sampling (MDS) vulnerability.
2545 Certain CPUs are vulnerable to an exploit against CPU
2546 internal buffers which can forward information to a
2547 disclosure gadget under certain conditions.
2549 In vulnerable processors, the speculatively
2550 forwarded data can be used in a cache side channel
2551 attack, to access data to which the attacker does
2552 not have direct access.
2554 This parameter controls the MDS mitigation. The
2557 full - Enable MDS mitigation on vulnerable CPUs
2558 full,nosmt - Enable MDS mitigation and disable
2559 SMT on vulnerable CPUs
2560 off - Unconditionally disable MDS mitigation
2562 On TAA-affected machines, mds=off can be prevented by
2563 an active TAA mitigation as both vulnerabilities are
2564 mitigated with the same mechanism so in order to disable
2565 this mitigation, you need to specify tsx_async_abort=off
2568 Not specifying this option is equivalent to
2571 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2573 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2574 Amount of memory to be used when the kernel is not able
2575 to see the whole system memory or for test.
2576 [X86] Work as limiting max address. Use together
2577 with memmap= to avoid physical address space collisions.
2578 Without memmap= PCI devices could be placed at addresses
2579 belonging to unused RAM.
2581 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2585 [KNL,SH] Allow user to override the default size for
2586 per-device physically contiguous DMA buffers.
2588 memhp_default_state=online/offline
2589 [KNL] Set the initial state for the memory hotplug
2590 onlining policy. If not specified, the default value is
2591 set according to the
2592 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2594 See Documentation/admin-guide/mm/memory-hotplug.rst.
2596 memmap=exactmap [KNL,X86] Enable setting of an exact
2597 E820 memory map, as specified by the user.
2598 Such memmap=exactmap lines can be constructed based on
2599 BIOS output or other requirements. See the memmap=nn@ss
2602 memmap=nn[KMG]@ss[KMG]
2603 [KNL] Force usage of a specific region of memory.
2604 Region of memory to be used is from ss to ss+nn.
2605 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2606 which limits max address to nn[KMG].
2607 Multiple different regions can be specified,
2610 memmap=100M@2G,100M#3G,1G!1024G
2612 memmap=nn[KMG]#ss[KMG]
2613 [KNL,ACPI] Mark specific memory as ACPI data.
2614 Region of memory to be marked is from ss to ss+nn.
2616 memmap=nn[KMG]$ss[KMG]
2617 [KNL,ACPI] Mark specific memory as reserved.
2618 Region of memory to be reserved is from ss to ss+nn.
2619 Example: Exclude memory from 0x18690000-0x1869ffff
2620 memmap=64K$0x18690000
2622 memmap=0x10000$0x18690000
2623 Some bootloaders may need an escape character before '$',
2624 like Grub2, otherwise '$' and the following number
2627 memmap=nn[KMG]!ss[KMG]
2628 [KNL,X86] Mark specific memory as protected.
2629 Region of memory to be used, from ss to ss+nn.
2630 The memory region may be marked as e820 type 12 (0xc)
2631 and is NVDIMM or ADR memory.
2633 memmap=<size>%<offset>-<oldtype>+<newtype>
2634 [KNL,ACPI] Convert memory within the specified region
2635 from <oldtype> to <newtype>. If "-<oldtype>" is left
2636 out, the whole region will be marked as <newtype>,
2637 even if previously unavailable. If "+<newtype>" is left
2638 out, matching memory will be removed. Types are
2639 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2640 3 = ACPI, 12 = PRAM.
2642 memory_corruption_check=0/1 [X86]
2643 Some BIOSes seem to corrupt the first 64k of
2644 memory when doing things like suspend/resume.
2645 Setting this option will scan the memory
2646 looking for corruption. Enabling this will
2647 both detect corruption and prevent the kernel
2648 from using the memory being corrupted.
2649 However, its intended as a diagnostic tool; if
2650 repeatable BIOS-originated corruption always
2651 affects the same memory, you can use memmap=
2652 to prevent the kernel from using that memory.
2654 memory_corruption_check_size=size [X86]
2655 By default it checks for corruption in the low
2656 64k, making this memory unavailable for normal
2657 use. Use this parameter to scan for
2658 corruption in more or less memory.
2660 memory_corruption_check_period=seconds [X86]
2661 By default it checks for corruption every 60
2662 seconds. Use this parameter to check at some
2663 other rate. 0 disables periodic checking.
2665 memtest= [KNL,X86,ARM,PPC] Enable memtest
2667 default : 0 <disable>
2668 Specifies the number of memtest passes to be
2669 performed. Each pass selects another test
2670 pattern from a given set of patterns. Memtest
2671 fills the memory with this pattern, validates
2672 memory contents and reserves bad memory
2673 regions that are detected.
2675 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2676 Valid arguments: on, off
2677 Default (depends on kernel configuration option):
2678 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2679 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2680 mem_encrypt=on: Activate SME
2681 mem_encrypt=off: Do not activate SME
2683 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2684 for details on when memory encryption can be activated.
2686 mem_sleep_default= [SUSPEND] Default system suspend mode:
2687 s2idle - Suspend-To-Idle
2688 shallow - Power-On Suspend or equivalent (if supported)
2689 deep - Suspend-To-RAM or equivalent (if supported)
2690 See Documentation/admin-guide/pm/sleep-states.rst.
2692 meye.*= [HW] Set MotionEye Camera parameters
2693 See Documentation/media/v4l-drivers/meye.rst.
2695 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2696 Multi-Function General Purpose Timers on AMD Geode
2699 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2700 the BIOS has incorrectly applied a workaround. TinyBIOS
2701 version 0.98 is known to be affected, 0.99 fixes the
2702 problem by letting the user disable the workaround.
2706 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2707 physical address is ignored.
2709 mini2440= [ARM,HW,KNL]
2710 Format:[0..2][b][c][t]
2712 MINI2440 configuration specification:
2713 0 - The attached screen is the 3.5" TFT
2714 1 - The attached screen is the 7" TFT
2715 2 - The VGA Shield is attached (1024x768)
2716 Leaving out the screen size parameter will not load
2717 the TFT driver, and the framebuffer will be left
2719 b - Enable backlight. The TFT backlight pin will be
2720 linked to the kernel VESA blanking code and a GPIO
2721 LED. This parameter is not necessary when using the
2723 c - Enable the s3c camera interface.
2724 t - Reserved for enabling touchscreen support. The
2725 touchscreen support is not enabled in the mainstream
2726 kernel as of 2.6.30, a preliminary port can be found
2727 in the "bleeding edge" mini2440 support kernel at
2728 http://repo.or.cz/w/linux-2.6/mini2440.git
2731 [X86,PPC,S390,ARM64] Control optional mitigations for
2732 CPU vulnerabilities. This is a set of curated,
2733 arch-independent options, each of which is an
2734 aggregation of existing arch-specific options.
2737 Disable all optional CPU mitigations. This
2738 improves system performance, but it may also
2739 expose users to several CPU vulnerabilities.
2740 Equivalent to: nopti [X86,PPC]
2742 nospectre_v1 [X86,PPC]
2744 nospectre_v2 [X86,PPC,S390,ARM64]
2745 spectre_v2_user=off [X86]
2746 spec_store_bypass_disable=off [X86,PPC]
2747 ssbd=force-off [ARM64]
2750 tsx_async_abort=off [X86]
2751 kvm.nx_huge_pages=off [X86]
2754 This does not have any effect on
2755 kvm.nx_huge_pages when
2756 kvm.nx_huge_pages=force.
2759 Mitigate all CPU vulnerabilities, but leave SMT
2760 enabled, even if it's vulnerable. This is for
2761 users who don't want to be surprised by SMT
2762 getting disabled across kernel upgrades, or who
2763 have other ways of avoiding SMT-based attacks.
2764 Equivalent to: (default behavior)
2767 Mitigate all CPU vulnerabilities, disabling SMT
2768 if needed. This is for users who always want to
2769 be fully mitigated, even if it means losing SMT.
2770 Equivalent to: l1tf=flush,nosmt [X86]
2771 mds=full,nosmt [X86]
2772 tsx_async_abort=full,nosmt [X86]
2775 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2776 parameter allows control of the logging verbosity for
2777 the additional memory initialisation checks. A value
2778 of 0 disables mminit logging and a level of 4 will
2779 log everything. Information is printed at KERN_DEBUG
2780 so loglevel=8 may also need to be specified.
2783 [KNL] When CONFIG_MODULE_SIG is set, this means that
2784 modules without (valid) signatures will fail to load.
2785 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2786 is always true, so this option does nothing.
2788 module_blacklist= [KNL] Do not load a comma-separated list of
2789 modules. Useful for debugging problem modules.
2792 [MOUSE] Maximum time between finger touching and
2793 leaving touchpad surface for touch to be considered
2794 a tap and be reported as a left button click (for
2795 touchpads working in absolute mode only).
2797 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2798 reporting absolute coordinates, such as tablets
2799 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2800 reporting absolute coordinates, such as tablets
2802 movablecore= [KNL,X86,IA-64,PPC]
2803 Format: nn[KMGTPE] | nn%
2804 This parameter is the complement to kernelcore=, it
2805 specifies the amount of memory used for migratable
2806 allocations. If both kernelcore and movablecore is
2807 specified, then kernelcore will be at *least* the
2808 specified value but may be more. If movablecore on its
2809 own is specified, the administrator must be careful
2810 that the amount of memory usable for all allocations
2813 movable_node [KNL] Boot-time switch to make hotplugable memory
2814 NUMA nodes to be movable. This means that the memory
2815 of such nodes will be usable only for movable
2816 allocations which rules out almost all kernel
2817 allocations. Use with caution!
2819 MTD_Partition= [MTD]
2820 Format: <name>,<region-number>,<size>,<offset>
2822 MTD_Region= [MTD] Format:
2823 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2826 See drivers/mtd/parsers/cmdlinepart.c
2828 multitce=off [PPC] This parameter disables the use of the pSeries
2829 firmware feature for updating multiple TCE entries
2832 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2834 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2836 boundary - index of last SLC block on Flex-OneNAND.
2837 The remaining blocks are configured as MLC blocks.
2838 lock - Configure if Flex-OneNAND boundary should be locked.
2839 Once locked, the boundary cannot be changed.
2840 1 indicates lock status, 0 indicates unlock status.
2843 ARM/S3C2412 JIVE boot control
2845 See arch/arm/mach-s3c2412/mach-jive.c
2847 mtouchusb.raw_coordinates=
2848 [HW] Make the MicroTouch USB driver use raw coordinates
2849 ('y', default) or cooked coordinates ('n')
2851 mtrr_chunk_size=nn[KMG] [X86]
2852 used for mtrr cleanup. It is largest continuous chunk
2853 that could hold holes aka. UC entries.
2855 mtrr_gran_size=nn[KMG] [X86]
2856 Used for mtrr cleanup. It is granularity of mtrr block.
2858 Large value could prevent small alignment from
2861 mtrr_spare_reg_nr=n [X86]
2863 Range: 0,7 : spare reg number
2865 Used for mtrr cleanup. It is spare mtrr entries number.
2866 Set to 2 or more if your graphical card needs more.
2868 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2870 netdev= [NET] Network devices parameters
2871 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2872 Note that mem_start is often overloaded to mean
2873 something different and driver-specific.
2874 This usage is only documented in each driver source
2878 [NETFILTER] Enable connection tracking flow accounting
2879 0 to disable accounting
2880 1 to enable accounting
2883 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2884 See Documentation/admin-guide/nfs/nfsroot.rst.
2886 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2887 See Documentation/admin-guide/nfs/nfsroot.rst.
2889 nfsrootdebug [NFS] enable nfsroot debugging messages.
2890 See Documentation/admin-guide/nfs/nfsroot.rst.
2892 nfs.callback_nr_threads=
2893 [NFSv4] set the total number of threads that the
2894 NFS client will assign to service NFSv4 callback
2897 nfs.callback_tcpport=
2898 [NFS] set the TCP port on which the NFSv4 callback
2899 channel should listen.
2902 [NFS] sets the pathname to the program which is used
2903 to update the NFS client cache entries.
2905 nfs.cache_getent_timeout=
2906 [NFS] sets the timeout after which an attempt to
2907 update a cache entry is deemed to have failed.
2909 nfs.idmap_cache_timeout=
2910 [NFS] set the maximum lifetime for idmapper cache
2914 [NFS] enable 64-bit inode numbers.
2915 If zero, the NFS client will fake up a 32-bit inode
2916 number for the readdir() and stat() syscalls instead
2917 of returning the full 64-bit number.
2918 The default is to return 64-bit inode numbers.
2920 nfs.max_session_cb_slots=
2921 [NFSv4.1] Sets the maximum number of session
2922 slots the client will assign to the callback
2923 channel. This determines the maximum number of
2924 callbacks the client will process in parallel for
2925 a particular server.
2927 nfs.max_session_slots=
2928 [NFSv4.1] Sets the maximum number of session slots
2929 the client will attempt to negotiate with the server.
2930 This limits the number of simultaneous RPC requests
2931 that the client can send to the NFSv4.1 server.
2932 Note that there is little point in setting this
2933 value higher than the max_tcp_slot_table_limit.
2935 nfs.nfs4_disable_idmapping=
2936 [NFSv4] When set to the default of '1', this option
2937 ensures that both the RPC level authentication
2938 scheme and the NFS level operations agree to use
2939 numeric uids/gids if the mount is using the
2940 'sec=sys' security flavour. In effect it is
2941 disabling idmapping, which can make migration from
2942 legacy NFSv2/v3 systems to NFSv4 easier.
2943 Servers that do not support this mode of operation
2944 will be autodetected by the client, and it will fall
2945 back to using the idmapper.
2946 To turn off this behaviour, set the value to '0'.
2948 [NFS4] Specify an additional fixed unique ident-
2949 ification string that NFSv4 clients can insert into
2950 their nfs_client_id4 string. This is typically a
2951 UUID that is generated at system install time.
2953 nfs.send_implementation_id =
2954 [NFSv4.1] Send client implementation identification
2955 information in exchange_id requests.
2956 If zero, no implementation identification information
2958 The default is to send the implementation identification
2961 nfs.recover_lost_locks =
2962 [NFSv4] Attempt to recover locks that were lost due
2963 to a lease timeout on the server. Please note that
2964 doing this risks data corruption, since there are
2965 no guarantees that the file will remain unchanged
2966 after the locks are lost.
2967 If you want to enable the kernel legacy behaviour of
2968 attempting to recover these locks, then set this
2970 The default parameter value of '0' causes the kernel
2971 not to attempt recovery of lost locks.
2973 nfs4.layoutstats_timer =
2974 [NFSv4.2] Change the rate at which the kernel sends
2975 layoutstats to the pNFS metadata server.
2977 Setting this to value to 0 causes the kernel to use
2978 whatever value is the default set by the layout
2979 driver. A non-zero value sets the minimum interval
2980 in seconds between layoutstats transmissions.
2982 nfsd.nfs4_disable_idmapping=
2983 [NFSv4] When set to the default of '1', the NFSv4
2984 server will return only numeric uids and gids to
2985 clients using auth_sys, and will accept numeric uids
2986 and gids from such clients. This is intended to ease
2987 migration from NFSv2/v3.
2989 nmi_debug= [KNL,SH] Specify one or more actions to take
2990 when a NMI is triggered.
2991 Format: [state][,regs][,debounce][,die]
2993 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2994 Format: [panic,][nopanic,][num]
2996 0 - turn hardlockup detector in nmi_watchdog off
2997 1 - turn hardlockup detector in nmi_watchdog on
2998 When panic is specified, panic when an NMI watchdog
2999 timeout occurs (or 'nopanic' to not panic on an NMI
3000 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3001 To disable both hard and soft lockup detectors,
3002 please see 'nowatchdog'.
3003 This is useful when you use a panic=... timeout and
3004 need the box quickly up again.
3006 These settings can be accessed at runtime via
3007 the nmi_watchdog and hardlockup_panic sysctls.
3009 netpoll.carrier_timeout=
3010 [NET] Specifies amount of time (in seconds) that
3011 netpoll should wait for a carrier. By default netpoll
3014 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3015 emulation library even if a 387 maths coprocessor
3018 no5lvl [X86-64] Disable 5-level paging mode. Forces
3019 kernel to use 4-level paging instead.
3022 [HW] Never suspend the console
3023 Disable suspending of consoles during suspend and
3024 hibernate operations. Once disabled, debugging
3025 messages can reach various consoles while the rest
3026 of the system is being put to sleep (ie, while
3027 debugging driver suspend/resume hooks). This may
3028 not work reliably with all consoles, but is known
3029 to work with serial and VGA consoles.
3030 To facilitate more flexible debugging, we also add
3031 console_suspend, a printk module parameter to control
3032 it. Users could use console_suspend (usually
3033 /sys/module/printk/parameters/console_suspend) to
3034 turn on/off it dynamically.
3036 novmcoredd [KNL,KDUMP]
3037 Disable device dump. Device dump allows drivers to
3038 append dump data to vmcore so you can collect driver
3039 specified debug info. Drivers can append the data
3040 without any limit and this data is stored in memory,
3041 so this may cause significant memory stress. Disabling
3042 device dump can help save memory but the driver debug
3043 data will be no longer available. This parameter
3044 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3047 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3048 caches in the slab allocator. Saves per-node memory,
3049 but will impact performance.
3053 noaltinstr [S390] Disables alternative instructions patching
3054 (CPU alternatives feature).
3056 noapic [SMP,APIC] Tells the kernel to not make use of any
3057 IOAPICs that may be present in the system.
3059 noautogroup Disable scheduler automatic task group creation.
3061 nobats [PPC] Do not use BATs for mapping kernel lowmem
3062 on "Classic" PPC cores.
3066 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3068 nodelayacct [KNL] Disable per-task delay accounting
3070 nodsp [SH] Disable hardware DSP at boot time.
3072 noefi Disable EFI runtime services support.
3077 On X86-32 available only on PAE configured kernels.
3078 noexec=on: enable non-executable mappings (default)
3079 noexec=off: disable non-executable mappings
3082 Disable SMAP (Supervisor Mode Access Prevention)
3083 even if it is supported by processor.
3086 Disable SMEP (Supervisor Mode Execution Prevention)
3087 even if it is supported by processor.
3090 This affects only 32-bit executables.
3091 noexec32=on: enable non-executable mappings (default)
3092 read doesn't imply executable mappings
3093 noexec32=off: disable non-executable mappings
3094 read implies executable mappings
3096 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3098 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3099 register save and restore. The kernel will only save
3100 legacy floating-point registers on task switch.
3102 nohugeiomap [KNL,x86,PPC] Disable kernel huge I/O mappings.
3104 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3105 Equivalent to smt=1.
3107 [KNL,x86] Disable symmetric multithreading (SMT).
3108 nosmt=force: Force disable SMT, cannot be undone
3109 via the sysfs control file.
3111 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3112 (bounds check bypass). With this option data leaks are
3113 possible in the system.
3115 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3116 the Spectre variant 2 (indirect branch prediction)
3117 vulnerability. System may allow data leaks with this
3120 nospec_store_bypass_disable
3121 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3123 noxsave [BUGS=X86] Disables x86 extended register state save
3124 and restore using xsave. The kernel will fallback to
3125 enabling legacy floating-point and sse state.
3127 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3128 register states. The kernel will fall back to use
3129 xsave to save the states. By using this parameter,
3130 performance of saving the states is degraded because
3131 xsave doesn't support modified optimization while
3132 xsaveopt supports it on xsaveopt enabled systems.
3134 noxsaves [X86] Disables xsaves and xrstors used in saving and
3135 restoring x86 extended register state in compacted
3136 form of xsave area. The kernel will fall back to use
3137 xsaveopt and xrstor to save and restore the states
3138 in standard form of xsave area. By using this
3139 parameter, xsave area per process might occupy more
3140 memory on xsaves enabled systems.
3142 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
3143 wfi(ARM) instruction doesn't work correctly and not to
3144 use it. This is also useful when using JTAG debugger.
3146 no_file_caps Tells the kernel not to honor file capabilities. The
3147 only way then for a file to be executed with privilege
3148 is to be setuid root or executed by root.
3150 nohalt [IA-64] Tells the kernel not to use the power saving
3151 function PAL_HALT_LIGHT when idle. This increases
3152 power-consumption. On the positive side, it reduces
3153 interrupt wake-up latency, which may improve performance
3154 in certain environments such as networked servers or
3157 nohibernate [HIBERNATION] Disable hibernation and resume.
3159 nohz= [KNL] Boottime enable/disable dynamic ticks
3160 Valid arguments: on, off
3163 nohz_full= [KNL,BOOT,SMP,ISOL]
3164 The argument is a cpu list, as described above.
3165 In kernels built with CONFIG_NO_HZ_FULL=y, set
3166 the specified list of CPUs whose tick will be stopped
3167 whenever possible. The boot CPU will be forced outside
3168 the range to maintain the timekeeping. Any CPUs
3169 in this list will have their RCU callbacks offloaded,
3170 just as if they had also been called out in the
3171 rcu_nocbs= boot parameter.
3173 noiotrap [SH] Disables trapped I/O port accesses.
3175 noirqdebug [X86-32] Disables the code which attempts to detect and
3176 disable unhandled interrupt sources.
3178 no_timer_check [X86,APIC] Disables the code which tests for
3179 broken timer IRQ sources.
3181 noisapnp [ISAPNP] Disables ISA PnP code.
3183 noinitrd [RAM] Tells the kernel not to load any configured
3186 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3188 [Deprecated - use intremap=off]
3192 noinvpcid [X86] Disable the INVPCID cpu feature.
3194 nojitter [IA-64] Disables jitter checking for ITC timers.
3196 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3198 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3202 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3203 clock and use the default one.
3205 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3206 accounting. steal time is computed, but won't
3207 influence scheduler behaviour
3209 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3211 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3213 noltlbs [PPC] Do not use large page/tlb entries for kernel
3214 lowmem mapping on PPC40x and PPC8xx
3216 nomca [IA-64] Disable machine check abort handling
3218 nomce [X86-32] Disable Machine Check Exception
3220 nomfgpt [X86-32] Disable Multi-Function General Purpose
3221 Timer usage (for AMD Geode machines).
3223 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3224 shutdown the other cpus. Instead use the REBOOT_VECTOR
3227 nomodule Disable module load
3229 nopat [X86] Disable PAT (page attribute table extension of
3230 pagetables) support.
3232 nopcid [X86-64] Disable the PCID cpu feature.
3234 norandmaps Don't use address space randomization. Equivalent to
3235 echo 0 > /proc/sys/kernel/randomize_va_space
3237 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3238 with UP alternatives
3240 nordrand [X86] Disable kernel use of the RDRAND and
3241 RDSEED instructions even if they are supported
3242 by the processor. RDRAND and RDSEED are still
3243 available to user space applications.
3245 noresume [SWSUSP] Disables resume and restores original swap
3248 no-scroll [VGA] Disables scrollback.
3249 This is required for the Braillex ib80-piezo Braille
3250 reader made by F.H. Papenmeier (Germany).
3254 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3256 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3257 and disable the IO APIC. legacy for "maxcpus=0".
3259 nosoftlockup [KNL] Disable the soft-lockup detector.
3261 nosync [HW,M68K] Disables sync negotiation for all devices.
3263 nowatchdog [KNL] Disable both lockup detectors, i.e.
3264 soft-lockup and NMI watchdog (hard-lockup).
3268 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3270 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3271 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3272 Some features depend on CPU0. Known dependencies are:
3273 1. Resume from suspend/hibernate depends on CPU0.
3274 Suspend/hibernate will fail if CPU0 is offline and you
3275 need to online CPU0 before suspend/hibernate.
3276 2. PIC interrupts also depend on CPU0. CPU0 can't be
3277 removed if a PIC interrupt is detected.
3278 It's said poweroff/reboot may depend on CPU0 on some
3279 machines although I haven't seen such issues so far
3280 after CPU0 is offline on a few tested machines.
3281 If the dependencies are under your control, you can
3282 turn on cpu0_hotplug.
3284 nps_mtm_hs_ctr= [KNL,ARC]
3285 This parameter sets the maximum duration, in
3286 cycles, each HW thread of the CTOP can run
3287 without interruptions, before HW switches it.
3288 The actual maximum duration is 16 times this
3290 Format: integer between 1 and 255
3293 nptcg= [IA-64] Override max number of concurrent global TLB
3294 purges which is reported from either PAL_VM_SUMMARY or
3297 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3298 could support. nr_cpus=n : n >= 1 limits the kernel to
3299 support 'n' processors. It could be larger than the
3300 number of already plugged CPU during bootup, later in
3301 runtime you can physically add extra cpu until it reaches
3302 n. So during boot up some boot time memory for per-cpu
3303 variables need be pre-allocated for later physical cpu
3306 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3308 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3309 Allowed values are enable and disable
3311 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3312 'node', 'default' can be specified
3313 This can be set from sysctl after boot.
3314 See Documentation/admin-guide/sysctl/vm.rst for details.
3316 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3317 See Documentation/debugging-via-ohci1394.txt for more
3320 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3321 Rather than timing out after 20 ms if an EC
3322 command is not properly ACKed, override the length
3323 of the timeout. We have interrupts disabled while
3324 waiting for the ACK, so if this is set too high
3325 interrupts *may* be lost!
3327 omap_mux= [OMAP] Override bootloader pin multiplexing.
3328 Format: <mux_mode0.mode_name=value>...
3329 For example, to override I2C bus2:
3330 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3332 oprofile.timer= [HW]
3333 Use timer interrupt instead of performance counters
3335 oprofile.cpu_type= Force an oprofile cpu type
3336 This might be useful if you have an older oprofile
3337 userland or if you want common events.
3338 Format: { arch_perfmon }
3339 arch_perfmon: [X86] Force use of architectural
3340 perfmon on Intel CPUs instead of the
3341 CPU specific event set.
3342 timer: [X86] Force use of architectural NMI
3343 timer mode (see also oprofile.timer
3344 for generic hr timer mode)
3346 oops=panic Always panic on oopses. Default is to just kill the
3347 process, but there is a small probability of
3348 deadlocking the machine.
3349 This will also cause panics on machine check exceptions.
3350 Useful together with panic=30 to trigger a reboot.
3353 [KNL] Boolean flag to control whether the page allocator
3354 should randomize its free lists. The randomization may
3355 be automatically enabled if the kernel detects it is
3356 running on a platform with a direct-mapped memory-side
3357 cache, and this parameter can be used to
3358 override/disable that behavior. The state of the flag
3359 can be read from sysfs at:
3360 /sys/module/page_alloc/parameters/shuffle.
3362 page_owner= [KNL] Boot-time page_owner enabling option.
3363 Storage of the information about who allocated
3364 each page is disabled in default. With this switch,
3366 on: enable the feature
3368 page_poison= [KNL] Boot-time parameter changing the state of
3369 poisoning on the buddy allocator, available with
3370 CONFIG_PAGE_POISONING=y.
3371 off: turn off poisoning (default)
3372 on: turn on poisoning
3374 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3375 timeout > 0: seconds before rebooting
3376 timeout = 0: wait forever
3377 timeout < 0: reboot immediately
3380 panic_print= Bitmask for printing system info when panic happens.
3381 User can chose combination of the following bits:
3382 bit 0: print all tasks info
3383 bit 1: print system memory info
3384 bit 2: print timer info
3385 bit 3: print locks info if CONFIG_LOCKDEP is on
3386 bit 4: print ftrace buffer
3387 bit 5: print all printk messages in buffer
3389 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3392 crash_kexec_post_notifiers
3393 Run kdump after running panic-notifiers and dumping
3394 kmsg. This only for the users who doubt kdump always
3395 succeeds in any situation.
3396 Note that this also increases risks of kdump failure,
3397 because some panic notifiers can make the crashed
3398 kernel more unstable.
3400 parkbd.port= [HW] Parallel port number the keyboard adapter is
3401 connected to, default is 0.
3403 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3404 0 for XT, 1 for AT (default is AT).
3407 parport= [HW,PPT] Specify parallel ports. 0 disables.
3408 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3409 Use 'auto' to force the driver to use any
3410 IRQ/DMA settings detected (the default is to
3411 ignore detected IRQ/DMA settings because of
3412 possible conflicts). You can specify the base
3413 address, IRQ, and DMA settings; IRQ and DMA
3414 should be numbers, or 'auto' (for using detected
3415 settings on that particular port), or 'nofifo'
3416 (to avoid using a FIFO even if it is detected).
3417 Parallel ports are assigned in the order they
3418 are specified on the command line, starting
3421 parport_init_mode= [HW,PPT]
3422 Configure VIA parallel port to operate in
3423 a specific mode. This is necessary on Pegasos
3424 computer where firmware has no options for setting
3425 up parallel port mode and sets it to spp.
3426 Currently this function knows 686a and 8231 chips.
3427 Format: [spp|ps2|epp|ecp|ecpepp]
3430 Halt all CPUs after the first oops has been printed for
3431 the specified number of seconds. This is to be used if
3432 your oopses keep scrolling off the screen.
3437 See header of drivers/block/paride/pcd.c.
3438 See also Documentation/admin-guide/blockdev/paride.rst.
3440 pci=option[,option...] [PCI] various PCI subsystem options.
3442 Some options herein operate on a specific device
3443 or a set of devices (<pci_dev>). These are
3444 specified in one of the following formats:
3446 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3447 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3449 Note: the first format specifies a PCI
3450 bus/device/function address which may change
3451 if new hardware is inserted, if motherboard
3452 firmware changes, or due to changes caused
3453 by other kernel parameters. If the
3454 domain is left unspecified, it is
3455 taken to be zero. Optionally, a path
3456 to a device through multiple device/function
3457 addresses can be specified after the base
3458 address (this is more robust against
3459 renumbering issues). The second format
3460 selects devices using IDs from the
3461 configuration space which may match multiple
3462 devices in the system.
3464 earlydump dump PCI config space before the kernel
3466 off [X86] don't probe for the PCI bus
3467 bios [X86-32] force use of PCI BIOS, don't access
3468 the hardware directly. Use this if your machine
3469 has a non-standard PCI host bridge.
3470 nobios [X86-32] disallow use of PCI BIOS, only direct
3471 hardware access methods are allowed. Use this
3472 if you experience crashes upon bootup and you
3473 suspect they are caused by the BIOS.
3474 conf1 [X86] Force use of PCI Configuration Access
3475 Mechanism 1 (config address in IO port 0xCF8,
3476 data in IO port 0xCFC, both 32-bit).
3477 conf2 [X86] Force use of PCI Configuration Access
3478 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3479 the function, IO port 0xCFA, also 8-bit, sets
3480 bus number. The config space is then accessed
3481 through ports 0xC000-0xCFFF).
3482 See http://wiki.osdev.org/PCI for more info
3483 on the configuration access mechanisms.
3484 noaer [PCIE] If the PCIEAER kernel config parameter is
3485 enabled, this kernel boot option can be used to
3486 disable the use of PCIE advanced error reporting.
3487 nodomains [PCI] Disable support for multiple PCI
3488 root domains (aka PCI segments, in ACPI-speak).
3489 nommconf [X86] Disable use of MMCONFIG for PCI
3491 check_enable_amd_mmconf [X86] check for and enable
3492 properly configured MMIO access to PCI
3493 config space on AMD family 10h CPU
3494 nomsi [MSI] If the PCI_MSI kernel config parameter is
3495 enabled, this kernel boot option can be used to
3496 disable the use of MSI interrupts system-wide.
3497 noioapicquirk [APIC] Disable all boot interrupt quirks.
3498 Safety option to keep boot IRQs enabled. This
3499 should never be necessary.
3500 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3501 primary IO-APIC for bridges that cannot disable
3502 boot IRQs. This fixes a source of spurious IRQs
3503 when the system masks IRQs.
3504 noioapicreroute [APIC] Disable workaround that uses the
3505 boot IRQ equivalent of an IRQ that connects to
3506 a chipset where boot IRQs cannot be disabled.
3507 The opposite of ioapicreroute.
3508 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3509 routing table. These calls are known to be buggy
3510 on several machines and they hang the machine
3511 when used, but on other computers it's the only
3512 way to get the interrupt routing table. Try
3513 this option if the kernel is unable to allocate
3514 IRQs or discover secondary PCI buses on your
3516 rom [X86] Assign address space to expansion ROMs.
3517 Use with caution as certain devices share
3518 address decoders between ROMs and other
3520 norom [X86] Do not assign address space to
3521 expansion ROMs that do not already have
3522 BIOS assigned address ranges.
3523 nobar [X86] Do not assign address space to the
3524 BARs that weren't assigned by the BIOS.
3525 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3526 assigned automatically to PCI devices. You can
3527 make the kernel exclude IRQs of your ISA cards
3529 pirqaddr=0xAAAAA [X86] Specify the physical address
3530 of the PIRQ table (normally generated
3531 by the BIOS) if it is outside the
3532 F0000h-100000h range.
3533 lastbus=N [X86] Scan all buses thru bus #N. Can be
3534 useful if the kernel is unable to find your
3535 secondary buses and you want to tell it
3536 explicitly which ones they are.
3537 assign-busses [X86] Always assign all PCI bus
3538 numbers ourselves, overriding
3539 whatever the firmware may have done.
3540 usepirqmask [X86] Honor the possible IRQ mask stored
3541 in the BIOS $PIR table. This is needed on
3542 some systems with broken BIOSes, notably
3543 some HP Pavilion N5400 and Omnibook XE3
3544 notebooks. This will have no effect if ACPI
3545 IRQ routing is enabled.
3546 noacpi [X86] Do not use ACPI for IRQ routing
3547 or for PCI scanning.
3548 use_crs [X86] Use PCI host bridge window information
3549 from ACPI. On BIOSes from 2008 or later, this
3550 is enabled by default. If you need to use this,
3551 please report a bug.
3552 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3553 If you need to use this, please report a bug.
3554 routeirq Do IRQ routing for all PCI devices.
3555 This is normally done in pci_enable_device(),
3556 so this option is a temporary workaround
3557 for broken drivers that don't call it.
3558 skip_isa_align [X86] do not align io start addr, so can
3559 handle more pci cards
3560 noearly [X86] Don't do any early type 1 scanning.
3561 This might help on some broken boards which
3562 machine check when some devices' config space
3563 is read. But various workarounds are disabled
3564 and some IOMMU drivers will not work.
3565 bfsort Sort PCI devices into breadth-first order.
3566 This sorting is done to get a device
3567 order compatible with older (<= 2.4) kernels.
3568 nobfsort Don't sort PCI devices into breadth-first order.
3569 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3570 tuning and use the BIOS-configured MPS defaults.
3571 pcie_bus_safe Set every device's MPS to the largest value
3572 supported by all devices below the root complex.
3573 pcie_bus_perf Set device MPS to the largest allowable MPS
3574 based on its parent bus. Also set MRRS (Max
3575 Read Request Size) to the largest supported
3576 value (no larger than the MPS that the device
3577 or bus can support) for best performance.
3578 pcie_bus_peer2peer Set every device's MPS to 128B, which
3579 every device is guaranteed to support. This
3580 configuration allows peer-to-peer DMA between
3581 any pair of devices, possibly at the cost of
3582 reduced performance. This also guarantees
3583 that hot-added devices will work.
3584 cbiosize=nn[KMG] The fixed amount of bus space which is
3585 reserved for the CardBus bridge's IO window.
3586 The default value is 256 bytes.
3587 cbmemsize=nn[KMG] The fixed amount of bus space which is
3588 reserved for the CardBus bridge's memory
3589 window. The default value is 64 megabytes.
3592 [<order of align>@]<pci_dev>[; ...]
3593 Specifies alignment and device to reassign
3594 aligned memory resources. How to
3595 specify the device is described above.
3596 If <order of align> is not specified,
3597 PAGE_SIZE is used as alignment.
3598 A PCI-PCI bridge can be specified if resource
3599 windows need to be expanded.
3600 To specify the alignment for several
3601 instances of a device, the PCI vendor,
3602 device, subvendor, and subdevice may be
3603 specified, e.g., 12@pci:8086:9c22:103c:198f
3604 for 4096-byte alignment.
3605 ecrc= Enable/disable PCIe ECRC (transaction layer
3606 end-to-end CRC checking).
3607 bios: Use BIOS/firmware settings. This is the
3611 hpiosize=nn[KMG] The fixed amount of bus space which is
3612 reserved for hotplug bridge's IO window.
3613 Default size is 256 bytes.
3614 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3615 reserved for hotplug bridge's MMIO window.
3616 Default size is 2 megabytes.
3617 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3618 reserved for hotplug bridge's MMIO_PREF window.
3619 Default size is 2 megabytes.
3620 hpmemsize=nn[KMG] The fixed amount of bus space which is
3621 reserved for hotplug bridge's MMIO and
3623 Default size is 2 megabytes.
3624 hpbussize=nn The minimum amount of additional bus numbers
3625 reserved for buses below a hotplug bridge.
3627 realloc= Enable/disable reallocating PCI bridge resources
3628 if allocations done by BIOS are too small to
3629 accommodate resources required by all child
3631 off: Turn realloc off
3633 realloc same as realloc=on
3634 noari do not use PCIe ARI.
3635 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3636 do not use PCIe ATS (and IOMMU device IOTLB).
3637 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3638 only look for one device below a PCIe downstream
3640 big_root_window Try to add a big 64bit memory window to the PCIe
3641 root complex on AMD CPUs. Some GFX hardware
3642 can resize a BAR to allow access to all VRAM.
3643 Adding the window is slightly risky (it may
3644 conflict with unreported devices), so this
3646 disable_acs_redir=<pci_dev>[; ...]
3647 Specify one or more PCI devices (in the format
3648 specified above) separated by semicolons.
3649 Each device specified will have the PCI ACS
3650 redirect capabilities forced off which will
3651 allow P2P traffic between devices through
3652 bridges without forcing it upstream. Note:
3653 this removes isolation between devices and
3654 may put more devices in an IOMMU group.
3655 force_floating [S390] Force usage of floating interrupts.
3656 nomio [S390] Do not use MIO instructions.
3658 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3661 force Enable ASPM even on devices that claim not to support it.
3662 WARNING: Forcing ASPM on may cause system lockups.
3664 pcie_ports= [PCIE] PCIe port services handling:
3665 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3666 even if the platform doesn't give the OS permission to
3667 use them. This may cause conflicts if the platform
3668 also tries to use these services.
3669 dpc-native Use native PCIe service for DPC only. May
3670 cause conflicts if firmware uses AER or DPC.
3671 compat Disable native PCIe services (PME, AER, DPC, PCIe
3674 pcie_port_pm= [PCIE] PCIe port power management handling:
3675 off Disable power management of all PCIe ports
3676 force Forcibly enable power management of all PCIe ports
3678 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3679 nomsi Do not use MSI for native PCIe PME signaling (this makes
3680 all PCIe root ports use INTx for all services).
3682 pcmv= [HW,PCMCIA] BadgePAD 4
3686 Keep all power-domains already enabled by bootloader on,
3687 even if no driver has claimed them. This is useful
3688 for debug and development, but should not be
3689 needed on a platform with proper driver support.
3692 See Documentation/admin-guide/blockdev/paride.rst.
3694 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3697 See arch/parisc/kernel/pdc_chassis.c
3699 percpu_alloc= Select which percpu first chunk allocator to use.
3700 Currently supported values are "embed" and "page".
3701 Archs may support subset or none of the selections.
3702 See comments in mm/percpu.c for details on each
3703 allocator. This parameter is primarily for debugging
3704 and performance comparison.
3707 See Documentation/admin-guide/blockdev/paride.rst.
3710 See Documentation/admin-guide/blockdev/paride.rst.
3712 pirq= [SMP,APIC] Manual mp-table setup
3713 See Documentation/x86/i386/IO-APIC.rst.
3715 plip= [PPT,NET] Parallel port network link
3716 Format: { parport<nr> | timid | 0 }
3717 See also Documentation/admin-guide/parport.rst.
3719 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3720 Override pmtimer IOPort with a hex value.
3724 Enable PNP debug messages (depends on the
3725 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3726 via /sys/module/pnp/parameters/debug. We always show
3727 current resource usage; turning this on also shows
3728 possible settings and some assignment information.
3734 { on | off | curr | res | no-curr | no-res }
3737 [ISAPNP] Exclude IRQs for the autoconfiguration
3740 [ISAPNP] Exclude DMAs for the autoconfiguration
3742 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3743 Ranges are in pairs (I/O port base and size).
3746 [ISAPNP] Exclude memory regions for the
3748 Ranges are in pairs (memory base and size).
3750 ports= [IP_VS_FTP] IPVS ftp helper module
3752 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3754 Format: <port>,<port>....
3756 powersave=off [PPC] This option disables power saving features.
3757 It specifically disables cpuidle and sets the
3758 platform machine description specific power_save
3759 function to NULL. On Idle the CPU just reduces
3762 ppc_strict_facility_enable
3763 [PPC] This option catches any kernel floating point,
3764 Altivec, VSX and SPE outside of regions specifically
3765 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3766 There is some performance impact when enabling this.
3770 Disable Hardware Transactional Memory
3772 print-fatal-signals=
3773 [KNL] debug: print fatal signals
3775 If enabled, warn about various signal handling
3776 related application anomalies: too many signals,
3777 too many POSIX.1 timers, fatal signals causing a
3780 If you hit the warning due to signal overflow,
3781 you might want to try "ulimit -i unlimited".
3785 printk.always_kmsg_dump=
3786 Trigger kmsg_dump for cases other than kernel oops or
3788 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3791 printk.devkmsg={on,off,ratelimit}
3792 Control writing to /dev/kmsg.
3793 on - unlimited logging to /dev/kmsg from userspace
3794 off - logging to /dev/kmsg disabled
3795 ratelimit - ratelimit the logging
3798 printk.time= Show timing data prefixed to each printk message line
3799 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3801 processor.max_cstate= [HW,ACPI]
3802 Limit processor to maximum C-state
3803 max_cstate=9 overrides any DMI blacklist limit.
3805 processor.nocst [HW,ACPI]
3806 Ignore the _CST method to determine C-states,
3807 instead using the legacy FADT method
3809 profile= [KNL] Enable kernel profiling via /proc/profile
3810 Format: [<profiletype>,]<number>
3811 Param: <profiletype>: "schedule", "sleep", or "kvm"
3812 [defaults to kernel profiling]
3813 Param: "schedule" - profile schedule points.
3814 Param: "sleep" - profile D-state sleeping (millisecs).
3815 Requires CONFIG_SCHEDSTATS
3816 Param: "kvm" - profile VM exits.
3817 Param: <number> - step/bucket size as a power of 2 for
3818 statistical time based profiling.
3820 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3822 See Documentation/admin-guide/blockdev/ramdisk.rst.
3824 psi= [KNL] Enable or disable pressure stall information
3828 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3829 probe for; one of (bare|imps|exps|lifebook|any).
3830 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3832 psmouse.resetafter= [HW,MOUSE]
3833 Try to reset the device after so many bad packets
3836 [HW,MOUSE] Set desired mouse resolution, in dpi.
3837 psmouse.smartscroll=
3838 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3839 0 = disabled, 1 = enabled (default).
3841 pstore.backend= Specify the name of the pstore backend to use
3844 See Documentation/admin-guide/blockdev/paride.rst.
3846 pti= [X86_64] Control Page Table Isolation of user and
3847 kernel address spaces. Disabling this feature
3848 removes hardening, but improves performance of
3849 system calls and interrupts.
3851 on - unconditionally enable
3852 off - unconditionally disable
3853 auto - kernel detects whether your CPU model is
3854 vulnerable to issues that PTI mitigates
3856 Not specifying this option is equivalent to pti=auto.
3859 Equivalent to pti=off
3862 [KNL] Number of legacy pty's. Overwrites compiled-in
3865 quiet [KNL] Disable most log messages
3870 See Documentation/admin-guide/md.rst.
3872 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3873 See Documentation/admin-guide/blockdev/ramdisk.rst.
3875 random.trust_cpu={on,off}
3876 [KNL] Enable or disable trusting the use of the
3877 CPU's random number generator (if available) to
3878 fully seed the kernel's CRNG. Default is controlled
3879 by CONFIG_RANDOM_TRUST_CPU.
3881 ras=option[,option,...] [KNL] RAS-specific options
3884 Disable the Correctable Errors Collector,
3885 see CONFIG_RAS_CEC help text.
3888 The argument is a cpu list, as described above,
3889 except that the string "all" can be used to
3890 specify every CPU on the system.
3892 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3893 the specified list of CPUs to be no-callback CPUs.
3894 Invocation of these CPUs' RCU callbacks will be
3895 offloaded to "rcuox/N" kthreads created for that
3896 purpose, where "x" is "p" for RCU-preempt, and
3897 "s" for RCU-sched, and "N" is the CPU number.
3898 This reduces OS jitter on the offloaded CPUs,
3899 which can be useful for HPC and real-time
3900 workloads. It can also improve energy efficiency
3901 for asymmetric multiprocessors.
3904 Rather than requiring that offloaded CPUs
3905 (specified by rcu_nocbs= above) explicitly
3906 awaken the corresponding "rcuoN" kthreads,
3907 make these kthreads poll for callbacks.
3908 This improves the real-time response for the
3909 offloaded CPUs by relieving them of the need to
3910 wake up the corresponding kthread, but degrades
3911 energy efficiency by requiring that the kthreads
3912 periodically wake up to do the polling.
3914 rcutree.blimit= [KNL]
3915 Set maximum number of finished RCU callbacks to
3916 process in one batch.
3918 rcutree.dump_tree= [KNL]
3919 Dump the structure of the rcu_node combining tree
3920 out at early boot. This is used for diagnostic
3921 purposes, to verify correct tree setup.
3923 rcutree.gp_cleanup_delay= [KNL]
3924 Set the number of jiffies to delay each step of
3925 RCU grace-period cleanup.
3927 rcutree.gp_init_delay= [KNL]
3928 Set the number of jiffies to delay each step of
3929 RCU grace-period initialization.
3931 rcutree.gp_preinit_delay= [KNL]
3932 Set the number of jiffies to delay each step of
3933 RCU grace-period pre-initialization, that is,
3934 the propagation of recent CPU-hotplug changes up
3935 the rcu_node combining tree.
3937 rcutree.use_softirq= [KNL]
3938 If set to zero, move all RCU_SOFTIRQ processing to
3939 per-CPU rcuc kthreads. Defaults to a non-zero
3940 value, meaning that RCU_SOFTIRQ is used by default.
3941 Specify rcutree.use_softirq=0 to use rcuc kthreads.
3943 rcutree.rcu_fanout_exact= [KNL]
3944 Disable autobalancing of the rcu_node combining
3945 tree. This is used by rcutorture, and might
3946 possibly be useful for architectures having high
3947 cache-to-cache transfer latencies.
3949 rcutree.rcu_fanout_leaf= [KNL]
3950 Change the number of CPUs assigned to each
3951 leaf rcu_node structure. Useful for very
3952 large systems, which will choose the value 64,
3953 and for NUMA systems with large remote-access
3954 latencies, which will choose a value aligned
3955 with the appropriate hardware boundaries.
3957 rcutree.jiffies_till_first_fqs= [KNL]
3958 Set delay from grace-period initialization to
3959 first attempt to force quiescent states.
3960 Units are jiffies, minimum value is zero,
3961 and maximum value is HZ.
3963 rcutree.jiffies_till_next_fqs= [KNL]
3964 Set delay between subsequent attempts to force
3965 quiescent states. Units are jiffies, minimum
3966 value is one, and maximum value is HZ.
3968 rcutree.jiffies_till_sched_qs= [KNL]
3969 Set required age in jiffies for a
3970 given grace period before RCU starts
3971 soliciting quiescent-state help from
3972 rcu_note_context_switch() and cond_resched().
3973 If not specified, the kernel will calculate
3974 a value based on the most recent settings
3975 of rcutree.jiffies_till_first_fqs
3976 and rcutree.jiffies_till_next_fqs.
3977 This calculated value may be viewed in
3978 rcutree.jiffies_to_sched_qs. Any attempt to set
3979 rcutree.jiffies_to_sched_qs will be cheerfully
3982 rcutree.kthread_prio= [KNL,BOOT]
3983 Set the SCHED_FIFO priority of the RCU per-CPU
3984 kthreads (rcuc/N). This value is also used for
3985 the priority of the RCU boost threads (rcub/N)
3986 and for the RCU grace-period kthreads (rcu_bh,
3987 rcu_preempt, and rcu_sched). If RCU_BOOST is
3988 set, valid values are 1-99 and the default is 1
3989 (the least-favored priority). Otherwise, when
3990 RCU_BOOST is not set, valid values are 0-99 and
3991 the default is zero (non-realtime operation).
3993 rcutree.rcu_nocb_gp_stride= [KNL]
3994 Set the number of NOCB callback kthreads in
3995 each group, which defaults to the square root
3996 of the number of CPUs. Larger numbers reduce
3997 the wakeup overhead on the global grace-period
3998 kthread, but increases that same overhead on
3999 each group's NOCB grace-period kthread.
4001 rcutree.qhimark= [KNL]
4002 Set threshold of queued RCU callbacks beyond which
4003 batch limiting is disabled.
4005 rcutree.qlowmark= [KNL]
4006 Set threshold of queued RCU callbacks below which
4007 batch limiting is re-enabled.
4009 rcutree.qovld= [KNL]
4010 Set threshold of queued RCU callbacks beyond which
4011 RCU's force-quiescent-state scan will aggressively
4012 enlist help from cond_resched() and sched IPIs to
4013 help CPUs more quickly reach quiescent states.
4014 Set to less than zero to make this be set based
4015 on rcutree.qhimark at boot time and to zero to
4016 disable more aggressive help enlistment.
4018 rcutree.rcu_idle_gp_delay= [KNL]
4019 Set wakeup interval for idle CPUs that have
4020 RCU callbacks (RCU_FAST_NO_HZ=y).
4022 rcutree.rcu_idle_lazy_gp_delay= [KNL]
4023 Set wakeup interval for idle CPUs that have
4024 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
4025 Lazy RCU callbacks are those which RCU can
4026 prove do nothing more than free memory.
4028 rcutree.rcu_kick_kthreads= [KNL]
4029 Cause the grace-period kthread to get an extra
4030 wake_up() if it sleeps three times longer than
4031 it should at force-quiescent-state time.
4032 This wake_up() will be accompanied by a
4033 WARN_ONCE() splat and an ftrace_dump().
4035 rcutree.sysrq_rcu= [KNL]
4036 Commandeer a sysrq key to dump out Tree RCU's
4037 rcu_node tree with an eye towards determining
4038 why a new grace period has not yet started.
4040 rcuperf.gp_async= [KNL]
4041 Measure performance of asynchronous
4042 grace-period primitives such as call_rcu().
4044 rcuperf.gp_async_max= [KNL]
4045 Specify the maximum number of outstanding
4046 callbacks per writer thread. When a writer
4047 thread exceeds this limit, it invokes the
4048 corresponding flavor of rcu_barrier() to allow
4049 previously posted callbacks to drain.
4051 rcuperf.gp_exp= [KNL]
4052 Measure performance of expedited synchronous
4053 grace-period primitives.
4055 rcuperf.holdoff= [KNL]
4056 Set test-start holdoff period. The purpose of
4057 this parameter is to delay the start of the
4058 test until boot completes in order to avoid
4061 rcuperf.kfree_rcu_test= [KNL]
4062 Set to measure performance of kfree_rcu() flooding.
4064 rcuperf.kfree_nthreads= [KNL]
4065 The number of threads running loops of kfree_rcu().
4067 rcuperf.kfree_alloc_num= [KNL]
4068 Number of allocations and frees done in an iteration.
4070 rcuperf.kfree_loops= [KNL]
4071 Number of loops doing rcuperf.kfree_alloc_num number
4072 of allocations and frees.
4074 rcuperf.nreaders= [KNL]
4075 Set number of RCU readers. The value -1 selects
4076 N, where N is the number of CPUs. A value
4077 "n" less than -1 selects N-n+1, where N is again
4078 the number of CPUs. For example, -2 selects N
4079 (the number of CPUs), -3 selects N+1, and so on.
4080 A value of "n" less than or equal to -N selects
4083 rcuperf.nwriters= [KNL]
4084 Set number of RCU writers. The values operate
4085 the same as for rcuperf.nreaders.
4086 N, where N is the number of CPUs
4088 rcuperf.perf_type= [KNL]
4089 Specify the RCU implementation to test.
4091 rcuperf.shutdown= [KNL]
4092 Shut the system down after performance tests
4093 complete. This is useful for hands-off automated
4096 rcuperf.verbose= [KNL]
4097 Enable additional printk() statements.
4099 rcuperf.writer_holdoff= [KNL]
4100 Write-side holdoff between grace periods,
4101 in microseconds. The default of zero says
4104 rcutorture.fqs_duration= [KNL]
4105 Set duration of force_quiescent_state bursts
4108 rcutorture.fqs_holdoff= [KNL]
4109 Set holdoff time within force_quiescent_state bursts
4112 rcutorture.fqs_stutter= [KNL]
4113 Set wait time between force_quiescent_state bursts
4116 rcutorture.fwd_progress= [KNL]
4117 Enable RCU grace-period forward-progress testing
4118 for the types of RCU supporting this notion.
4120 rcutorture.fwd_progress_div= [KNL]
4121 Specify the fraction of a CPU-stall-warning
4122 period to do tight-loop forward-progress testing.
4124 rcutorture.fwd_progress_holdoff= [KNL]
4125 Number of seconds to wait between successive
4126 forward-progress tests.
4128 rcutorture.fwd_progress_need_resched= [KNL]
4129 Enclose cond_resched() calls within checks for
4130 need_resched() during tight-loop forward-progress
4133 rcutorture.gp_cond= [KNL]
4134 Use conditional/asynchronous update-side
4135 primitives, if available.
4137 rcutorture.gp_exp= [KNL]
4138 Use expedited update-side primitives, if available.
4140 rcutorture.gp_normal= [KNL]
4141 Use normal (non-expedited) asynchronous
4142 update-side primitives, if available.
4144 rcutorture.gp_sync= [KNL]
4145 Use normal (non-expedited) synchronous
4146 update-side primitives, if available. If all
4147 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4148 rcutorture.gp_normal=, and rcutorture.gp_sync=
4149 are zero, rcutorture acts as if is interpreted
4150 they are all non-zero.
4152 rcutorture.n_barrier_cbs= [KNL]
4153 Set callbacks/threads for rcu_barrier() testing.
4155 rcutorture.nfakewriters= [KNL]
4156 Set number of concurrent RCU writers. These just
4157 stress RCU, they don't participate in the actual
4158 test, hence the "fake".
4160 rcutorture.nreaders= [KNL]
4161 Set number of RCU readers. The value -1 selects
4162 N-1, where N is the number of CPUs. A value
4163 "n" less than -1 selects N-n-2, where N is again
4164 the number of CPUs. For example, -2 selects N
4165 (the number of CPUs), -3 selects N+1, and so on.
4167 rcutorture.object_debug= [KNL]
4168 Enable debug-object double-call_rcu() testing.
4170 rcutorture.onoff_holdoff= [KNL]
4171 Set time (s) after boot for CPU-hotplug testing.
4173 rcutorture.onoff_interval= [KNL]
4174 Set time (jiffies) between CPU-hotplug operations,
4175 or zero to disable CPU-hotplug testing.
4177 rcutorture.shuffle_interval= [KNL]
4178 Set task-shuffle interval (s). Shuffling tasks
4179 allows some CPUs to go into dyntick-idle mode
4180 during the rcutorture test.
4182 rcutorture.shutdown_secs= [KNL]
4183 Set time (s) after boot system shutdown. This
4184 is useful for hands-off automated testing.
4186 rcutorture.stall_cpu= [KNL]
4187 Duration of CPU stall (s) to test RCU CPU stall
4188 warnings, zero to disable.
4190 rcutorture.stall_cpu_holdoff= [KNL]
4191 Time to wait (s) after boot before inducing stall.
4193 rcutorture.stall_cpu_irqsoff= [KNL]
4194 Disable interrupts while stalling if set.
4196 rcutorture.stat_interval= [KNL]
4197 Time (s) between statistics printk()s.
4199 rcutorture.stutter= [KNL]
4200 Time (s) to stutter testing, for example, specifying
4201 five seconds causes the test to run for five seconds,
4202 wait for five seconds, and so on. This tests RCU's
4203 ability to transition abruptly to and from idle.
4205 rcutorture.test_boost= [KNL]
4206 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4207 "Maybe" means test if the RCU implementation
4208 under test support RCU priority boosting.
4210 rcutorture.test_boost_duration= [KNL]
4211 Duration (s) of each individual boost test.
4213 rcutorture.test_boost_interval= [KNL]
4214 Interval (s) between each boost test.
4216 rcutorture.test_no_idle_hz= [KNL]
4217 Test RCU's dyntick-idle handling. See also the
4218 rcutorture.shuffle_interval parameter.
4220 rcutorture.torture_type= [KNL]
4221 Specify the RCU implementation to test.
4223 rcutorture.verbose= [KNL]
4224 Enable additional printk() statements.
4226 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4227 Dump ftrace buffer after reporting RCU CPU
4230 rcupdate.rcu_cpu_stall_suppress= [KNL]
4231 Suppress RCU CPU stall warning messages.
4233 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4234 Suppress RCU CPU stall warning messages and
4235 rcutorture writer stall warnings that occur
4236 during early boot, that is, during the time
4237 before the init task is spawned.
4239 rcupdate.rcu_cpu_stall_timeout= [KNL]
4240 Set timeout for RCU CPU stall warning messages.
4242 rcupdate.rcu_expedited= [KNL]
4243 Use expedited grace-period primitives, for
4244 example, synchronize_rcu_expedited() instead
4245 of synchronize_rcu(). This reduces latency,
4246 but can increase CPU utilization, degrade
4247 real-time latency, and degrade energy efficiency.
4248 No effect on CONFIG_TINY_RCU kernels.
4250 rcupdate.rcu_normal= [KNL]
4251 Use only normal grace-period primitives,
4252 for example, synchronize_rcu() instead of
4253 synchronize_rcu_expedited(). This improves
4254 real-time latency, CPU utilization, and
4255 energy efficiency, but can expose users to
4256 increased grace-period latency. This parameter
4257 overrides rcupdate.rcu_expedited. No effect on
4258 CONFIG_TINY_RCU kernels.
4260 rcupdate.rcu_normal_after_boot= [KNL]
4261 Once boot has completed (that is, after
4262 rcu_end_inkernel_boot() has been invoked), use
4263 only normal grace-period primitives. No effect
4264 on CONFIG_TINY_RCU kernels.
4266 rcupdate.rcu_task_stall_timeout= [KNL]
4267 Set timeout in jiffies for RCU task stall warning
4268 messages. Disable with a value less than or equal
4271 rcupdate.rcu_self_test= [KNL]
4272 Run the RCU early boot self tests
4276 Run specified binary instead of /init from the ramdisk,
4277 used for early userspace startup. See initrd.
4280 force - Override the decision by the kernel to hide the
4281 advertisement of RDRAND support (this affects
4282 certain AMD processors because of buggy BIOS
4283 support, specifically around the suspend/resume
4287 Turn on/off individual RDT features. List is:
4288 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4290 E.g. to turn on cmt and turn off mba use:
4294 Format (x86 or x86_64):
4295 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4297 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4299 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4300 (prefix with 'panic_' to set mode for panic
4302 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4303 reboot_force is either force or not specified,
4304 reboot_cpu is s[mp]#### with #### being the processor
4305 to be used for rebooting.
4308 [KNL, SMP] Set scheduler's default relax_domain_level.
4309 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4311 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4312 Format: <base1>,<size1>[,<base2>,<size2>,...]
4313 Reserve I/O ports or memory so the kernel won't use
4314 them. If <base> is less than 0x10000, the region
4315 is assumed to be I/O ports; otherwise it is memory.
4317 reservetop= [X86-32]
4319 Reserves a hole at the top of the kernel virtual
4324 Set the amount of memory to reserve for BIOS at
4325 the bottom of the address space.
4327 reset_devices [KNL] Force drivers to reset the underlying device
4328 during initialization.
4331 Specify the partition device for software suspend
4333 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4335 resume_offset= [SWSUSP]
4336 Specify the offset from the beginning of the partition
4337 given by "resume=" at which the swap header is located,
4338 in <PAGE_SIZE> units (needed only for swap files).
4339 See Documentation/power/swsusp-and-swap-files.rst
4341 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4342 read the resume files
4344 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4345 Useful for devices that are detected asynchronously
4346 (e.g. USB and MMC devices).
4348 hibernate= [HIBERNATION]
4349 noresume Don't check if there's a hibernation image
4350 present during boot.
4351 nocompress Don't compress/decompress hibernation images.
4352 no Disable hibernation and resume.
4353 protect_image Turn on image protection during restoration
4354 (that will set all pages holding image data
4355 during restoration read-only).
4357 retain_initrd [RAM] Keep initrd memory after extraction
4359 rfkill.default_state=
4360 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4361 etc. communication is blocked by default.
4364 rfkill.master_switch_mode=
4365 0 The "airplane mode" button does nothing.
4366 1 The "airplane mode" button toggles between everything
4367 blocked and the previous configuration.
4368 2 The "airplane mode" button toggles between everything
4369 blocked and everything unblocked.
4371 rhash_entries= [KNL,NET]
4372 Set number of hash buckets for route cache
4375 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4378 ro [KNL] Mount root device read-only on boot
4381 on Mark read-only kernel memory as read-only (default).
4382 off Leave read-only kernel memory writable for debugging.
4385 Enable the uart passthrough on the designated usb port
4386 on Rockchip SoCs. When active, the signals of the
4387 debug-uart get routed to the D+ and D- pins of the usb
4388 port and the regular usb controller gets disabled.
4390 root= [KNL] Root filesystem
4391 See name_to_dev_t comment in init/do_mounts.c.
4393 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4394 mount the root filesystem
4396 rootflags= [KNL] Set root filesystem mount option string
4398 rootfstype= [KNL] Set root filesystem type
4400 rootwait [KNL] Wait (indefinitely) for root device to show up.
4401 Useful for devices that are detected asynchronously
4402 (e.g. USB and MMC devices).
4404 rproc_mem=nn[KMG][@address]
4405 [KNL,ARM,CMA] Remoteproc physical memory block.
4406 Memory area to be used by remote processor image,
4409 rw [KNL] Mount root device read-write on boot
4411 S [KNL] Run init in single mode
4413 s390_iommu= [HW,S390]
4414 Set s390 IOTLB flushing mode
4416 With strict flushing every unmap operation will result in
4417 an IOTLB flush. Default is lazy flushing before reuse,
4421 See drivers/net/irda/sa1100_ir.c.
4423 sbni= [NET] Granch SBNI12 leased line adapter
4425 sched_debug [KNL] Enables verbose scheduler debug messages.
4427 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4428 Allowed values are enable and disable. This feature
4429 incurs a small amount of overhead in the scheduler
4430 but is useful for debugging and performance tuning.
4432 sched_thermal_decay_shift=
4433 [KNL, SMP] Set a decay shift for scheduler thermal
4434 pressure signal. Thermal pressure signal follows the
4435 default decay period of other scheduler pelt
4436 signals(usually 32 ms but configurable). Setting
4437 sched_thermal_decay_shift will left shift the decay
4438 period for the thermal pressure signal by the shift
4440 i.e. with the default pelt decay period of 32 ms
4441 sched_thermal_decay_shift thermal pressure decay pr
4445 Format: integer between 0 and 10
4448 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4449 xtime_lock contention on larger systems, and/or RCU lock
4450 contention on all systems with CONFIG_MAXSMP set.
4451 Format: { "0" | "1" }
4452 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4454 Note: increases power consumption, thus should only be
4455 enabled if running jitter sensitive (HPC/RT) workloads.
4457 security= [SECURITY] Choose a legacy "major" security module to
4458 enable at boot. This has been deprecated by the
4461 selinux= [SELINUX] Disable or enable SELinux at boot time.
4462 Format: { "0" | "1" }
4463 See security/selinux/Kconfig help text.
4468 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4469 Format: { "0" | "1" }
4470 See security/apparmor/Kconfig help text
4473 Default value is set via kernel config option.
4475 serialnumber [BUGS=X86-32]
4478 Maximal number of shapers.
4486 Disable merging of slabs with similar size. May be
4487 necessary if there is some reason to distinguish
4488 allocs to different slabs, especially in hardened
4489 environments where the risk of heap overflows and
4490 layout control by attackers can usually be
4491 frustrated by disabling merging. This will reduce
4492 most of the exposure of a heap attack to a single
4493 cache (risks via metadata attacks are mostly
4494 unchanged). Debug options disable merging on their
4496 For more information see Documentation/vm/slub.rst.
4498 slab_max_order= [MM, SLAB]
4499 Determines the maximum allowed order for slabs.
4500 A high setting may cause OOMs due to memory
4501 fragmentation. Defaults to 1 for systems with
4502 more than 32MB of RAM, 0 otherwise.
4504 slub_debug[=options[,slabs]] [MM, SLUB]
4505 Enabling slub_debug allows one to determine the
4506 culprit if slab objects become corrupted. Enabling
4507 slub_debug can create guard zones around objects and
4508 may poison objects when not in use. Also tracks the
4509 last alloc / free. For more information see
4510 Documentation/vm/slub.rst.
4512 slub_memcg_sysfs= [MM, SLUB]
4513 Determines whether to enable sysfs directories for
4514 memory cgroup sub-caches. 1 to enable, 0 to disable.
4515 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4516 Enabling this can lead to a very high number of debug
4517 directories and files being created under
4520 slub_max_order= [MM, SLUB]
4521 Determines the maximum allowed order for slabs.
4522 A high setting may cause OOMs due to memory
4523 fragmentation. For more information see
4524 Documentation/vm/slub.rst.
4526 slub_min_objects= [MM, SLUB]
4527 The minimum number of objects per slab. SLUB will
4528 increase the slab order up to slub_max_order to
4529 generate a sufficiently large slab able to contain
4530 the number of objects indicated. The higher the number
4531 of objects the smaller the overhead of tracking slabs
4532 and the less frequently locks need to be acquired.
4533 For more information see Documentation/vm/slub.rst.
4535 slub_min_order= [MM, SLUB]
4536 Determines the minimum page order for slabs. Must be
4537 lower than slub_max_order.
4538 For more information see Documentation/vm/slub.rst.
4540 slub_nomerge [MM, SLUB]
4541 Same with slab_nomerge. This is supported for legacy.
4542 See slab_nomerge for more information.
4545 Format: <io1>[,<io2>[,...,<io8>]]
4547 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4548 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4549 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4550 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4551 smsc-ircc2.ircc_irq= [HW] IRQ line
4552 smsc-ircc2.ircc_dma= [HW] DMA channel
4553 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4554 0: Toshiba Satellite 1800 (GP data pin select)
4555 1: Fast pin select (default)
4558 smt [KNL,S390] Set the maximum number of threads (logical
4559 CPUs) to use per physical CPU on systems capable of
4560 symmetric multithreading (SMT). Will be capped to the
4561 actual hardware limit.
4563 Default: -1 (no limit)
4566 [KNL] Should the soft-lockup detector generate panics.
4569 A nonzero value instructs the soft-lockup detector
4570 to panic the machine when a soft-lockup occurs. It is
4571 also controlled by the kernel.softlockup_panic sysctl
4572 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
4573 respective build-time switch to that functionality.
4575 softlockup_all_cpu_backtrace=
4576 [KNL] Should the soft-lockup detector generate
4577 backtraces on all cpus.
4580 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4581 See Documentation/admin-guide/laptops/sonypi.rst
4583 spectre_v2= [X86] Control mitigation of Spectre variant 2
4584 (indirect branch speculation) vulnerability.
4585 The default operation protects the kernel from
4588 on - unconditionally enable, implies
4590 off - unconditionally disable, implies
4592 auto - kernel detects whether your CPU model is
4595 Selecting 'on' will, and 'auto' may, choose a
4596 mitigation method at run time according to the
4597 CPU, the available microcode, the setting of the
4598 CONFIG_RETPOLINE configuration option, and the
4599 compiler with which the kernel was built.
4601 Selecting 'on' will also enable the mitigation
4602 against user space to user space task attacks.
4604 Selecting 'off' will disable both the kernel and
4605 the user space protections.
4607 Specific mitigations can also be selected manually:
4609 retpoline - replace indirect branches
4610 retpoline,generic - google's original retpoline
4611 retpoline,amd - AMD-specific minimal thunk
4613 Not specifying this option is equivalent to
4617 [X86] Control mitigation of Spectre variant 2
4618 (indirect branch speculation) vulnerability between
4621 on - Unconditionally enable mitigations. Is
4622 enforced by spectre_v2=on
4624 off - Unconditionally disable mitigations. Is
4625 enforced by spectre_v2=off
4627 prctl - Indirect branch speculation is enabled,
4628 but mitigation can be enabled via prctl
4629 per thread. The mitigation control state
4630 is inherited on fork.
4633 - Like "prctl" above, but only STIBP is
4634 controlled per thread. IBPB is issued
4635 always when switching between different user
4639 - Same as "prctl" above, but all seccomp
4640 threads will enable the mitigation unless
4641 they explicitly opt out.
4644 - Like "seccomp" above, but only STIBP is
4645 controlled per thread. IBPB is issued
4646 always when switching between different
4647 user space processes.
4649 auto - Kernel selects the mitigation depending on
4650 the available CPU features and vulnerability.
4653 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4655 Not specifying this option is equivalent to
4656 spectre_v2_user=auto.
4658 spec_store_bypass_disable=
4659 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4660 (Speculative Store Bypass vulnerability)
4662 Certain CPUs are vulnerable to an exploit against a
4663 a common industry wide performance optimization known
4664 as "Speculative Store Bypass" in which recent stores
4665 to the same memory location may not be observed by
4666 later loads during speculative execution. The idea
4667 is that such stores are unlikely and that they can
4668 be detected prior to instruction retirement at the
4669 end of a particular speculation execution window.
4671 In vulnerable processors, the speculatively forwarded
4672 store can be used in a cache side channel attack, for
4673 example to read memory to which the attacker does not
4674 directly have access (e.g. inside sandboxed code).
4676 This parameter controls whether the Speculative Store
4677 Bypass optimization is used.
4679 On x86 the options are:
4681 on - Unconditionally disable Speculative Store Bypass
4682 off - Unconditionally enable Speculative Store Bypass
4683 auto - Kernel detects whether the CPU model contains an
4684 implementation of Speculative Store Bypass and
4685 picks the most appropriate mitigation. If the
4686 CPU is not vulnerable, "off" is selected. If the
4687 CPU is vulnerable the default mitigation is
4688 architecture and Kconfig dependent. See below.
4689 prctl - Control Speculative Store Bypass per thread
4690 via prctl. Speculative Store Bypass is enabled
4691 for a process by default. The state of the control
4692 is inherited on fork.
4693 seccomp - Same as "prctl" above, but all seccomp threads
4694 will disable SSB unless they explicitly opt out.
4696 Default mitigations:
4697 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4699 On powerpc the options are:
4701 on,auto - On Power8 and Power9 insert a store-forwarding
4702 barrier on kernel entry and exit. On Power7
4703 perform a software flush on kernel entry and
4707 Not specifying this option is equivalent to
4708 spec_store_bypass_disable=auto.
4710 spia_io_base= [HW,MTD]
4716 [X86] Enable split lock detection
4718 When enabled (and if hardware support is present), atomic
4719 instructions that access data across cache line
4720 boundaries will result in an alignment check exception.
4724 warn - the kernel will emit rate limited warnings
4725 about applications triggering the #AC
4726 exception. This mode is the default on CPUs
4727 that supports split lock detection.
4729 fatal - the kernel will send SIGBUS to applications
4730 that trigger the #AC exception.
4732 If an #AC exception is hit in the kernel or in
4733 firmware (i.e. not while executing in user mode)
4734 the kernel will oops in either "warn" or "fatal"
4737 srcutree.counter_wrap_check [KNL]
4738 Specifies how frequently to check for
4739 grace-period sequence counter wrap for the
4740 srcu_data structure's ->srcu_gp_seq_needed field.
4741 The greater the number of bits set in this kernel
4742 parameter, the less frequently counter wrap will
4743 be checked for. Note that the bottom two bits
4746 srcutree.exp_holdoff [KNL]
4747 Specifies how many nanoseconds must elapse
4748 since the end of the last SRCU grace period for
4749 a given srcu_struct until the next normal SRCU
4750 grace period will be considered for automatic
4751 expediting. Set to zero to disable automatic
4755 Speculative Store Bypass Disable control
4757 On CPUs that are vulnerable to the Speculative
4758 Store Bypass vulnerability and offer a
4759 firmware based mitigation, this parameter
4760 indicates how the mitigation should be used:
4762 force-on: Unconditionally enable mitigation for
4763 for both kernel and userspace
4764 force-off: Unconditionally disable mitigation for
4765 for both kernel and userspace
4766 kernel: Always enable mitigation in the
4767 kernel, and offer a prctl interface
4768 to allow userspace to register its
4769 interest in being mitigated too.
4771 stack_guard_gap= [MM]
4772 override the default stack gap protection. The value
4773 is in page units and it defines how many pages prior
4774 to (for stacks growing down) resp. after (for stacks
4775 growing up) the main stack are reserved for no other
4776 mapping. Default value is 256 pages.
4779 Enabled the stack tracer on boot up.
4781 stacktrace_filter=[function-list]
4782 [FTRACE] Limit the functions that the stack tracer
4783 will trace at boot up. function-list is a comma separated
4784 list of functions. This list can be changed at run
4785 time by the stack_trace_filter file in the debugfs
4786 tracing directory. Note, this enables stack tracing
4787 and the stacktrace above is not needed.
4791 Set the STI (builtin display/keyboard on the HP-PARISC
4792 machines) console (graphic card) which should be used
4793 as the initial boot-console.
4794 See also comment in drivers/video/console/sticore.c.
4797 See comment in drivers/video/console/sticore.c.
4800 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4802 sunrpc.min_resvport=
4803 sunrpc.max_resvport=
4805 SunRPC servers often require that client requests
4806 originate from a privileged port (i.e. a port in the
4807 range 0 < portnr < 1024).
4808 An administrator who wishes to reserve some of these
4809 ports for other uses may adjust the range that the
4810 kernel's sunrpc client considers to be privileged
4811 using these two parameters to set the minimum and
4812 maximum port values.
4814 sunrpc.svc_rpc_per_connection_limit=
4816 Limit the number of requests that the server will
4817 process in parallel from a single connection.
4818 The default value is 0 (no limit).
4822 Control how the NFS server code allocates CPUs to
4823 service thread pools. Depending on how many NICs
4824 you have and where their interrupts are bound, this
4825 option will affect which CPUs will do NFS serving.
4826 Note: this parameter cannot be changed while the
4827 NFS server is running.
4829 auto the server chooses an appropriate mode
4830 automatically using heuristics
4831 global a single global pool contains all CPUs
4832 percpu one pool for each CPU
4833 pernode one pool for each NUMA node (equivalent
4834 to global on non-NUMA machines)
4836 sunrpc.tcp_slot_table_entries=
4837 sunrpc.udp_slot_table_entries=
4839 Sets the upper limit on the number of simultaneous
4840 RPC calls that can be sent from the client to a
4841 server. Increasing these values may allow you to
4842 improve throughput, but will also increase the
4843 amount of memory reserved for use by the client.
4845 suspend.pm_test_delay=
4847 Sets the number of seconds to remain in a suspend test
4848 mode before resuming the system (see
4849 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4850 is set. Default value is 5.
4853 Format: { on | off | y | n | 1 | 0 }
4854 This parameter controls use of the Protected
4855 Execution Facility on pSeries.
4858 [KNL] Enable accounting of swap in memory resource
4859 controller if no parameter or 1 is given or disable
4860 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
4862 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4863 Format: { <int> | force | noforce }
4864 <int> -- Number of I/O TLB slabs
4865 force -- force using of bounce buffers even if they
4866 wouldn't be automatically used by the kernel
4867 noforce -- Never use bounce buffers (for debugging)
4871 sysfs.deprecated=0|1 [KNL]
4872 Enable/disable old style sysfs layout for old udev
4873 on older distributions. When this option is enabled
4874 very new udev will not work anymore. When this option
4875 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4876 in older udev will not work anymore.
4877 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4878 the kernel configuration.
4880 sysrq_always_enabled
4882 Ignore sysrq setting - this boot parameter will
4883 neutralize any effect of /proc/sys/kernel/sysrq.
4884 Useful for debugging.
4886 tcpmhash_entries= [KNL,NET]
4887 Set the number of tcp_metrics_hash slots.
4888 Default value is 8192 or 16384 depending on total
4889 ram pages. This is used to specify the TCP metrics
4890 cache size. See Documentation/networking/ip-sysctl.txt
4891 "tcp_no_metrics_save" section for more details.
4895 test_suspend= [SUSPEND][,N]
4896 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4897 standby suspend) or "freeze" (for suspend type freeze)
4898 as the system sleep state during system startup with
4899 the optional capability to repeat N number of times.
4900 The system is woken from this state using a
4901 wakeup-capable RTC alarm.
4903 thash_entries= [KNL,NET]
4904 Set number of hash buckets for TCP connection
4906 thermal.act= [HW,ACPI]
4907 -1: disable all active trip points in all thermal zones
4908 <degrees C>: override all lowest active trip points
4910 thermal.crt= [HW,ACPI]
4911 -1: disable all critical trip points in all thermal zones
4912 <degrees C>: override all critical trip points
4914 thermal.nocrt= [HW,ACPI]
4915 Set to disable actions on ACPI thermal zone
4916 critical and hot trip points.
4918 thermal.off= [HW,ACPI]
4919 1: disable ACPI thermal control
4921 thermal.psv= [HW,ACPI]
4922 -1: disable all passive trip points
4923 <degrees C>: override all passive trip points to this
4926 thermal.tzp= [HW,ACPI]
4927 Specify global default ACPI thermal zone polling rate
4928 <deci-seconds>: poll all this frequency
4929 0: no polling (default)
4932 Force threading of all interrupt handlers except those
4933 marked explicitly IRQF_NO_THREAD.
4937 Specify if the kernel should make use of the cpu
4938 topology information if the hardware supports this.
4939 The scheduler will make use of this information and
4940 e.g. base its process migration decisions on it.
4943 topology_updates= [KNL, PPC, NUMA]
4945 Specify if the kernel should ignore (off)
4946 topology updates sent by the hypervisor to this
4949 torture.disable_onoff_at_boot= [KNL]
4950 Prevent the CPU-hotplug component of torturing
4951 until after init has spawned.
4955 tpm_suspend_pcr=[HW,TPM]
4956 Format: integer pcr id
4957 Specify that at suspend time, the tpm driver
4958 should extend the specified pcr with zeros,
4959 as a workaround for some chips which fail to
4960 flush the last written pcr on TPM_SaveState.
4961 This will guarantee that all the other pcrs
4964 trace_buf_size=nn[KMG]
4965 [FTRACE] will set tracing buffer size on each cpu.
4967 trace_event=[event-list]
4968 [FTRACE] Set and start specified trace events in order
4969 to facilitate early boot debugging. The event-list is a
4970 comma separated list of trace events to enable. See
4971 also Documentation/trace/events.rst
4973 trace_options=[option-list]
4974 [FTRACE] Enable or disable tracer options at boot.
4975 The option-list is a comma delimited list of options
4976 that can be enabled or disabled just as if you were
4977 to echo the option name into
4979 /sys/kernel/debug/tracing/trace_options
4981 For example, to enable stacktrace option (to dump the
4982 stack trace of each event), add to the command line:
4984 trace_options=stacktrace
4986 See also Documentation/trace/ftrace.rst "trace options"
4990 Have the tracepoints sent to printk as well as the
4991 tracing ring buffer. This is useful for early boot up
4992 where the system hangs or reboots and does not give the
4993 option for reading the tracing buffer or performing a
4994 ftrace_dump_on_oops.
4996 To turn off having tracepoints sent to printk,
4997 echo 0 > /proc/sys/kernel/tracepoint_printk
4998 Note, echoing 1 into this file without the
4999 tracepoint_printk kernel cmdline option has no effect.
5003 Having tracepoints sent to printk() and activating high
5004 frequency tracepoints such as irq or sched, can cause
5005 the system to live lock.
5008 [FTRACE] enable this option to disable tracing when a
5009 warning is hit. This turns off "tracing_on". Tracing can
5010 be enabled again by echoing '1' into the "tracing_on"
5011 file located in /sys/kernel/debug/tracing/
5013 This option is useful, as it disables the trace before
5014 the WARNING dump is called, which prevents the trace to
5015 be filled with content caused by the warning output.
5017 This option can also be set at run time via the sysctl
5018 option: kernel/traceoff_on_warning
5020 transparent_hugepage=
5022 Format: [always|madvise|never]
5023 Can be used to control the default behavior of the system
5024 with respect to transparent hugepages.
5025 See Documentation/admin-guide/mm/transhuge.rst
5028 tsc= Disable clocksource stability checks for TSC.
5030 [x86] reliable: mark tsc clocksource as reliable, this
5031 disables clocksource verification at runtime, as well
5032 as the stability checks done at bootup. Used to enable
5033 high-resolution timer mode on older hardware, and in
5034 virtualized environment.
5035 [x86] noirqtime: Do not use TSC to do irq accounting.
5036 Used to run time disable IRQ_TIME_ACCOUNTING on any
5037 platforms where RDTSC is slow and this accounting
5039 [x86] unstable: mark the TSC clocksource as unstable, this
5040 marks the TSC unconditionally unstable at bootup and
5041 avoids any further wobbles once the TSC watchdog notices.
5042 [x86] nowatchdog: disable clocksource watchdog. Used
5043 in situations with strict latency requirements (where
5044 interruptions from clocksource watchdog are not
5047 tsx= [X86] Control Transactional Synchronization
5048 Extensions (TSX) feature in Intel processors that
5049 support TSX control.
5051 This parameter controls the TSX feature. The options are:
5053 on - Enable TSX on the system. Although there are
5054 mitigations for all known security vulnerabilities,
5055 TSX has been known to be an accelerator for
5056 several previous speculation-related CVEs, and
5057 so there may be unknown security risks associated
5058 with leaving it enabled.
5060 off - Disable TSX on the system. (Note that this
5061 option takes effect only on newer CPUs which are
5062 not vulnerable to MDS, i.e., have
5063 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5064 the new IA32_TSX_CTRL MSR through a microcode
5065 update. This new MSR allows for the reliable
5066 deactivation of the TSX functionality.)
5068 auto - Disable TSX if X86_BUG_TAA is present,
5069 otherwise enable TSX on the system.
5071 Not specifying this option is equivalent to tsx=off.
5073 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5076 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5077 Abort (TAA) vulnerability.
5079 Similar to Micro-architectural Data Sampling (MDS)
5080 certain CPUs that support Transactional
5081 Synchronization Extensions (TSX) are vulnerable to an
5082 exploit against CPU internal buffers which can forward
5083 information to a disclosure gadget under certain
5086 In vulnerable processors, the speculatively forwarded
5087 data can be used in a cache side channel attack, to
5088 access data to which the attacker does not have direct
5091 This parameter controls the TAA mitigation. The
5094 full - Enable TAA mitigation on vulnerable CPUs
5097 full,nosmt - Enable TAA mitigation and disable SMT on
5098 vulnerable CPUs. If TSX is disabled, SMT
5099 is not disabled because CPU is not
5100 vulnerable to cross-thread TAA attacks.
5101 off - Unconditionally disable TAA mitigation
5103 On MDS-affected machines, tsx_async_abort=off can be
5104 prevented by an active MDS mitigation as both vulnerabilities
5105 are mitigated with the same mechanism so in order to disable
5106 this mitigation, you need to specify mds=off too.
5108 Not specifying this option is equivalent to
5109 tsx_async_abort=full. On CPUs which are MDS affected
5110 and deploy MDS mitigation, TAA mitigation is not
5111 required and doesn't provide any additional
5115 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5117 turbografx.map[2|3]= [HW,JOY]
5118 TurboGraFX parallel port interface
5120 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5121 See also Documentation/input/devices/joystick-parport.rst
5123 udbg-immortal [PPC] When debugging early kernel crashes that
5124 happen after console_init() and before a proper
5125 console driver takes over, this boot options might
5126 help "seeing" what's going on.
5128 uhash_entries= [KNL,NET]
5129 Set number of hash buckets for UDP/UDP-Lite connections
5132 [USB] Ignore overcurrent events (default N).
5133 Some badly-designed motherboards generate lots of
5134 bogus events, for ports that aren't wired to
5135 anything. Set this parameter to avoid log spamming.
5136 Note that genuine overcurrent events won't be
5140 [X86] Cause panic on unknown NMI.
5142 usbcore.authorized_default=
5143 [USB] Default USB device authorization:
5144 (default -1 = authorized except for wireless USB,
5145 0 = not authorized, 1 = authorized, 2 = authorized
5146 if device connected to internal port)
5148 usbcore.autosuspend=
5149 [USB] The autosuspend time delay (in seconds) used
5150 for newly-detected USB devices (default 2). This
5151 is the time required before an idle device will be
5152 autosuspended. Devices for which the delay is set
5153 to a negative value won't be autosuspended at all.
5155 usbcore.usbfs_snoop=
5156 [USB] Set to log all usbfs traffic (default 0 = off).
5158 usbcore.usbfs_snoop_max=
5159 [USB] Maximum number of bytes to snoop in each URB
5162 usbcore.blinkenlights=
5163 [USB] Set to cycle leds on hubs (default 0 = off).
5165 usbcore.old_scheme_first=
5166 [USB] Start with the old device initialization
5167 scheme, applies only to low and full-speed devices
5170 usbcore.usbfs_memory_mb=
5171 [USB] Memory limit (in MB) for buffers allocated by
5172 usbfs (default = 16, 0 = max = 2047).
5174 usbcore.use_both_schemes=
5175 [USB] Try the other device initialization scheme
5176 if the first one fails (default 1 = enabled).
5178 usbcore.initial_descriptor_timeout=
5179 [USB] Specifies timeout for the initial 64-byte
5180 USB_REQ_GET_DESCRIPTOR request in milliseconds
5181 (default 5000 = 5.0 seconds).
5183 usbcore.nousb [USB] Disable the USB subsystem
5186 [USB] A list of quirk entries to augment the built-in
5187 usb core quirk list. List entries are separated by
5188 commas. Each entry has the form
5189 VendorID:ProductID:Flags. The IDs are 4-digit hex
5190 numbers and Flags is a set of letters. Each letter
5191 will change the built-in quirk; setting it if it is
5192 clear and clearing it if it is set. The letters have
5193 the following meanings:
5194 a = USB_QUIRK_STRING_FETCH_255 (string
5195 descriptors must not be fetched using
5197 b = USB_QUIRK_RESET_RESUME (device can't resume
5198 correctly so reset it instead);
5199 c = USB_QUIRK_NO_SET_INTF (device can't handle
5200 Set-Interface requests);
5201 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5202 handle its Configuration or Interface
5204 e = USB_QUIRK_RESET (device can't be reset
5205 (e.g morph devices), don't use reset);
5206 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5207 more interface descriptions than the
5208 bNumInterfaces count, and can't handle
5209 talking to these interfaces);
5210 g = USB_QUIRK_DELAY_INIT (device needs a pause
5211 during initialization, after we read
5212 the device descriptor);
5213 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5214 high speed and super speed interrupt
5215 endpoints, the USB 2.0 and USB 3.0 spec
5216 require the interval in microframes (1
5217 microframe = 125 microseconds) to be
5218 calculated as interval = 2 ^
5220 Devices with this quirk report their
5221 bInterval as the result of this
5222 calculation instead of the exponent
5223 variable used in the calculation);
5224 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5225 handle device_qualifier descriptor
5227 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5228 generates spurious wakeup, ignore
5229 remote wakeup capability);
5230 k = USB_QUIRK_NO_LPM (device can't handle Link
5232 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5233 (Device reports its bInterval as linear
5234 frames instead of the USB 2.0
5236 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5237 to be disconnected before suspend to
5238 prevent spurious wakeup);
5239 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5240 pause after every control message);
5241 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5242 delay after resetting its port);
5243 Example: quirks=0781:5580:bk,0a5c:5834:gij
5246 [USBHID] The interval which mice are to be polled at.
5249 [USBHID] The interval which joysticks are to be polled at.
5252 [USBHID] The interval which keyboards are to be polled at.
5254 usb-storage.delay_use=
5255 [UMS] The delay in seconds before a new device is
5256 scanned for Logical Units (default 1).
5259 [UMS] A list of quirks entries to supplement or
5260 override the built-in unusual_devs list. List
5261 entries are separated by commas. Each entry has
5262 the form VID:PID:Flags where VID and PID are Vendor
5263 and Product ID values (4-digit hex numbers) and
5264 Flags is a set of characters, each corresponding
5265 to a common usb-storage quirk flag as follows:
5266 a = SANE_SENSE (collect more than 18 bytes
5267 of sense data, not on uas);
5268 b = BAD_SENSE (don't collect more than 18
5269 bytes of sense data, not on uas);
5270 c = FIX_CAPACITY (decrease the reported
5271 device capacity by one sector);
5272 d = NO_READ_DISC_INFO (don't use
5273 READ_DISC_INFO command, not on uas);
5274 e = NO_READ_CAPACITY_16 (don't use
5275 READ_CAPACITY_16 command);
5276 f = NO_REPORT_OPCODES (don't use report opcodes
5278 g = MAX_SECTORS_240 (don't transfer more than
5279 240 sectors at a time, uas only);
5280 h = CAPACITY_HEURISTICS (decrease the
5281 reported device capacity by one
5282 sector if the number is odd);
5283 i = IGNORE_DEVICE (don't bind to this
5285 j = NO_REPORT_LUNS (don't use report luns
5287 l = NOT_LOCKABLE (don't try to lock and
5288 unlock ejectable media, not on uas);
5289 m = MAX_SECTORS_64 (don't transfer more
5290 than 64 sectors = 32 KB at a time,
5292 n = INITIAL_READ10 (force a retry of the
5293 initial READ(10) command, not on uas);
5294 o = CAPACITY_OK (accept the capacity
5295 reported by the device, not on uas);
5296 p = WRITE_CACHE (the device cache is ON
5297 by default, not on uas);
5298 r = IGNORE_RESIDUE (the device reports
5299 bogus residue values, not on uas);
5300 s = SINGLE_LUN (the device has only one
5302 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5303 commands, uas only);
5304 u = IGNORE_UAS (don't bind to the uas driver);
5305 w = NO_WP_DETECT (don't test whether the
5306 medium is write-protected).
5307 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5308 even if the device claims no cache,
5310 Example: quirks=0419:aaf5:rl,0421:0433:rc
5312 user_debug= [KNL,ARM]
5314 See arch/arm/Kconfig.debug help text.
5315 1 - undefined instruction events
5317 4 - invalid data aborts
5320 Example: user_debug=31
5323 [X86] Flags controlling user PTE allocations.
5325 nohigh = do not allocate PTE pages in
5326 HIGHMEM regardless of setting
5330 On X86_32, this is an alias for vdso32=. Otherwise:
5332 vdso=1: enable VDSO (the default)
5333 vdso=0: disable VDSO mapping
5335 vdso32= [X86] Control the 32-bit vDSO
5336 vdso32=1: enable 32-bit VDSO
5337 vdso32=0 or vdso32=2: disable 32-bit VDSO
5339 See the help text for CONFIG_COMPAT_VDSO for more
5340 details. If CONFIG_COMPAT_VDSO is set, the default is
5341 vdso32=0; otherwise, the default is vdso32=1.
5343 For compatibility with older kernels, vdso32=2 is an
5346 Try vdso32=0 if you encounter an error that says:
5347 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5350 vector=percpu: enable percpu vector domain
5352 video= [FB] Frame buffer configuration
5353 See Documentation/fb/modedb.rst.
5355 video.brightness_switch_enabled= [0,1]
5356 If set to 1, on receiving an ACPI notify event
5357 generated by hotkey, video driver will adjust brightness
5358 level and then send out the event to user space through
5359 the allocated input device; If set to 0, video driver
5360 will only send out the event without touching backlight
5365 [VMMIO] Memory mapped virtio (platform) device.
5367 <size>@<baseaddr>:<irq>[:<id>]
5369 <size> := size (can use standard suffixes
5371 <baseaddr> := physical base address
5372 <irq> := interrupt number (as passed to
5374 <id> := (optional) platform device id
5376 virtio_mmio.device=1K@0x100b0000:48:7
5378 Can be used multiple times for multiple devices.
5380 vga= [BOOT,X86-32] Select a particular video mode
5381 See Documentation/x86/boot.rst and
5382 Documentation/admin-guide/svga.rst.
5383 Use vga=ask for menu.
5384 This is actually a boot loader parameter; the value is
5385 passed to the kernel using a special protocol.
5387 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
5388 May slow down system boot speed, especially when
5389 enabled on systems with a large amount of memory.
5390 All options are enabled by default, and this
5391 interface is meant to allow for selectively
5392 enabling or disabling specific virtual memory
5395 Available options are:
5396 P Enable page structure init time poisoning
5397 - Disable all of the above options
5399 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5400 size of <nn>. This can be used to increase the
5401 minimum size (128MB on x86). It can also be used to
5402 decrease the size and leave more room for directly
5405 vmcp_cma=nn[MG] [KNL,S390]
5406 Sets the memory size reserved for contiguous memory
5407 allocations for the vmcp device driver.
5409 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5412 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5415 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5419 Controls the behavior of vsyscalls (i.e. calls to
5420 fixed addresses of 0xffffffffff600x00 from legacy
5421 code). Most statically-linked binaries and older
5422 versions of glibc use these calls. Because these
5423 functions are at fixed addresses, they make nice
5424 targets for exploits that can control RIP.
5426 emulate [default] Vsyscalls turn into traps and are
5427 emulated reasonably safely. The vsyscall
5430 xonly Vsyscalls turn into traps and are
5431 emulated reasonably safely. The vsyscall
5432 page is not readable.
5434 none Vsyscalls don't work at all. This makes
5435 them quite hard to use for exploits but
5436 might break your system.
5438 vt.color= [VT] Default text color.
5439 Format: 0xYX, X = foreground, Y = background.
5440 Default: 0x07 = light gray on black.
5442 vt.cur_default= [VT] Default cursor shape.
5443 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5444 the parameters of the <Esc>[?A;B;Cc escape sequence;
5445 see VGA-softcursor.txt. Default: 2 = underline.
5447 vt.default_blu= [VT]
5448 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5449 Change the default blue palette of the console.
5450 This is a 16-member array composed of values
5453 vt.default_grn= [VT]
5454 Format: <green0>,<green1>,<green2>,...,<green15>
5455 Change the default green palette of the console.
5456 This is a 16-member array composed of values
5459 vt.default_red= [VT]
5460 Format: <red0>,<red1>,<red2>,...,<red15>
5461 Change the default red palette of the console.
5462 This is a 16-member array composed of values
5468 Set system-wide default UTF-8 mode for all tty's.
5469 Default is 1, i.e. UTF-8 mode is enabled for all
5470 newly opened terminals.
5472 vt.global_cursor_default=
5475 Set system-wide default for whether a cursor
5476 is shown on new VTs. Default is -1,
5477 i.e. cursors will be created by default unless
5478 overridden by individual drivers. 0 will hide
5479 cursors, 1 will display them.
5481 vt.italic= [VT] Default color for italic text; 0-15.
5484 vt.underline= [VT] Default color for underlined text; 0-15.
5487 watchdog timers [HW,WDT] For information on watchdog timers,
5488 see Documentation/watchdog/watchdog-parameters.rst
5489 or other driver-specific files in the
5490 Documentation/watchdog/ directory.
5494 Set the hard lockup detector stall duration
5495 threshold in seconds. The soft lockup detector
5496 threshold is set to twice the value. A value of 0
5497 disables both lockup detectors. Default is 10
5500 workqueue.watchdog_thresh=
5501 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5502 warn stall conditions and dump internal state to
5503 help debugging. 0 disables workqueue stall
5504 detection; otherwise, it's the stall threshold
5505 duration in seconds. The default value is 30 and
5506 it can be updated at runtime by writing to the
5507 corresponding sysfs file.
5509 workqueue.disable_numa
5510 By default, all work items queued to unbound
5511 workqueues are affine to the NUMA nodes they're
5512 issued on, which results in better behavior in
5513 general. If NUMA affinity needs to be disabled for
5514 whatever reason, this option can be used. Note
5515 that this also can be controlled per-workqueue for
5516 workqueues visible under /sys/bus/workqueue/.
5518 workqueue.power_efficient
5519 Per-cpu workqueues are generally preferred because
5520 they show better performance thanks to cache
5521 locality; unfortunately, per-cpu workqueues tend to
5522 be more power hungry than unbound workqueues.
5524 Enabling this makes the per-cpu workqueues which
5525 were observed to contribute significantly to power
5526 consumption unbound, leading to measurably lower
5527 power usage at the cost of small performance
5530 The default value of this parameter is determined by
5531 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5533 workqueue.debug_force_rr_cpu
5534 Workqueue used to implicitly guarantee that work
5535 items queued without explicit CPU specified are put
5536 on the local CPU. This guarantee is no longer true
5537 and while local CPU is still preferred work items
5538 may be put on foreign CPUs. This debug option
5539 forces round-robin CPU selection to flush out
5540 usages which depend on the now broken guarantee.
5541 When enabled, memory and cache locality will be
5544 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5545 default x2apic cluster mode on platforms
5548 x86_intel_mid_timer= [X86-32,APBT]
5549 Choose timer option for x86 Intel MID platform.
5550 Two valid options are apbt timer only and lapic timer
5551 plus one apbt timer for broadcast timer.
5552 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5554 xen_512gb_limit [KNL,X86-64,XEN]
5555 Restricts the kernel running paravirtualized under Xen
5556 to use only up to 512 GB of RAM. The reason to do so is
5557 crash analysis tools and Xen tools for doing domain
5558 save/restore/migration must be enabled to handle larger
5561 xen_emul_unplug= [HW,X86,XEN]
5562 Unplug Xen emulated devices
5563 Format: [unplug0,][unplug1]
5564 ide-disks -- unplug primary master IDE devices
5565 aux-ide-disks -- unplug non-primary-master IDE devices
5566 nics -- unplug network devices
5567 all -- unplug all emulated devices (NICs and IDE disks)
5568 unnecessary -- unplugging emulated devices is
5569 unnecessary even if the host did not respond to
5571 never -- do not unplug even if version check succeeds
5573 xen_legacy_crash [X86,XEN]
5574 Crash from Xen panic notifier, without executing late
5575 panic() code such as dumping handler.
5577 xen_nopvspin [X86,XEN]
5578 Disables the ticketlock slowpath using Xen PV
5582 Disables the PV optimizations forcing the HVM guest to
5583 run as generic HVM guest with no PV drivers.
5584 This option is obsoleted by the "nopv" option, which
5585 has equivalent effect for XEN platform.
5587 xen_scrub_pages= [XEN]
5588 Boolean option to control scrubbing pages before giving them back
5589 to Xen, for use by other domains. Can be also changed at runtime
5590 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5591 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5593 xen_timer_slop= [X86-64,XEN]
5594 Set the timer slop (in nanoseconds) for the virtual Xen
5595 timers (default is 100000). This adjusts the minimum
5596 delta of virtualized Xen timers, where lower values
5597 improve timer resolution at the expense of processing
5598 more timer interrupts.
5600 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
5601 Disables the PV optimizations forcing the guest to run
5602 as generic guest with no PV drivers. Currently support
5603 XEN HVM, KVM, HYPER_V and VMWARE guest.
5605 xirc2ps_cs= [NET,PCMCIA]
5607 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5610 By default on POWER9 and above, the kernel will
5611 natively use the XIVE interrupt controller. This option
5612 allows the fallback firmware mode to be used:
5614 off Fallback to firmware control of XIVE interrupt
5615 controller on both pseries and powernv
5616 platforms. Only useful on POWER9 and above.
5618 xhci-hcd.quirks [USB,KNL]
5619 A hex value specifying bitmask with supplemental xhci
5620 host controller quirks. Meaning of each bit can be
5621 consulted in header drivers/usb/host/xhci.h.
5624 Format: { early | on | rw | ro | off }
5625 Controls if xmon debugger is enabled. Default is off.
5626 Passing only "xmon" is equivalent to "xmon=early".
5627 early Call xmon as early as possible on boot; xmon
5628 debugger is called from setup_arch().
5629 on xmon debugger hooks will be installed so xmon
5630 is only called on a kernel crash. Default mode,
5631 i.e. either "ro" or "rw" mode, is controlled
5632 with CONFIG_XMON_DEFAULT_RO_MODE.
5633 rw xmon debugger hooks will be installed so xmon
5634 is called only on a kernel crash, mode is write,
5635 meaning SPR registers, memory and, other data
5636 can be written using xmon commands.
5637 ro same as "rw" option above but SPR registers,
5638 memory, and other data can't be written using
5640 off xmon is disabled.
projects / platform / kernel / linux-rpi.git / blob