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.
530 See Documentation/s390/common_io.rst for details.
533 Prevents the clock framework from automatically gating
534 clocks that have not been explicitly enabled by a Linux
535 device driver but are enabled in hardware at reset or
536 by the bootloader/firmware. Note that this does not
537 force such clocks to be always-on nor does it reserve
538 those clocks in any way. This parameter is useful for
539 debug and development, but should not be needed on a
540 platform with proper driver support. For more
541 information, see Documentation/driver-api/clk.rst.
543 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
545 Forces specified clocksource (if available) to be used
546 when calculating gettimeofday(). If specified
547 clocksource is not available, it defaults to PIT.
548 Format: { pit | tsc | cyclone | pmtmr }
550 clocksource= Override the default clocksource
552 Override the default clocksource and use the clocksource
553 with the name specified.
554 Some clocksource names to choose from, depending on
556 [all] jiffies (this is the base, fallback clocksource)
558 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
559 pxa_timer,timer3,32k_counter,timer0_1
560 [X86-32] pit,hpet,tsc;
561 scx200_hrt on Geode; cyclone on IBM x440
569 clocksource.arm_arch_timer.evtstrm=
572 Enable/disable the eventstream feature of the ARM
573 architected timer so that code using WFE-based polling
574 loops can be debugged more effectively on production
577 clearcpuid=BITNUM [X86]
578 Disable CPUID feature X for the kernel. See
579 arch/x86/include/asm/cpufeatures.h for the valid bit
580 numbers. Note the Linux specific bits are not necessarily
581 stable over kernel options, but the vendor specific
583 Also note that user programs calling CPUID directly
584 or using the feature without checking anything
585 will still see it. This just prevents it from
586 being used by the kernel or shown in /proc/cpuinfo.
587 Also note the kernel might malfunction if you disable
590 cma=nn[MG]@[start[MG][-end[MG]]]
592 Sets the size of kernel global memory area for
593 contiguous memory allocations and optionally the
594 placement constraint by the physical address range of
595 memory allocations. A value of 0 disables CMA
596 altogether. For more information, see
597 include/linux/dma-contiguous.h
599 cmo_free_hint= [PPC] Format: { yes | no }
600 Specify whether pages are marked as being inactive
601 when they are freed. This is used in CMO environments
602 to determine OS memory pressure for page stealing by
606 coherent_pool=nn[KMG] [ARM,KNL]
607 Sets the size of memory pool for coherent, atomic dma
608 allocations, by default set to 256K.
610 com20020= [HW,NET] ARCnet - COM20020 chipset
612 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
614 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
618 ARCnet - COM90xx chipset (memory-mapped buffers)
619 Format: <io>[,<irq>[,<memstart>]]
621 condev= [HW,S390] console device
624 console= [KNL] Output console device and options.
626 tty<n> Use the virtual console device <n>.
630 Use the specified serial port. The options are of
631 the form "bbbbpnf", where "bbbb" is the baud rate,
632 "p" is parity ("n", "o", or "e"), "n" is number of
633 bits, and "f" is flow control ("r" for RTS or
634 omit it). Default is "9600n8".
636 See Documentation/admin-guide/serial-console.rst for more
638 Documentation/networking/netconsole.txt for an
641 uart[8250],io,<addr>[,options]
642 uart[8250],mmio,<addr>[,options]
643 uart[8250],mmio16,<addr>[,options]
644 uart[8250],mmio32,<addr>[,options]
645 uart[8250],0x<addr>[,options]
646 Start an early, polled-mode console on the 8250/16550
647 UART at the specified I/O port or MMIO address,
648 switching to the matching ttyS device later.
649 MMIO inter-register address stride is either 8-bit
650 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
651 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
652 to be equivalent to 'mmio'. 'options' are specified in
653 the same format described for ttyS above; if unspecified,
654 the h/w is not re-initialized.
656 hvc<n> Use the hypervisor console device <n>. This is for
657 both Xen and PowerPC hypervisors.
659 If the device connected to the port is not a TTY but a braille
660 device, prepend "brl," before the device type, for instance
662 For now, only VisioBraille is supported.
665 [KNL] Change console messages format
667 By default we print messages on consoles in
668 "[time stamp] text\n" format (time stamp may not be
669 printed, depending on CONFIG_PRINTK_TIME or
670 `printk_time' param).
672 Switch to syslog format: "<%u>[time stamp] text\n"
673 IOW, each message will have a facility and loglevel
674 prefix. The format is similar to one used by syslog()
675 syscall, or to executing "dmesg -S --raw" or to reading
678 consoleblank= [KNL] The console blank (screen saver) timeout in
679 seconds. A value of 0 disables the blank timer.
683 [KNL] Change the default value for
684 /proc/<pid>/coredump_filter.
685 See also Documentation/filesystems/proc.txt.
687 coresight_cpu_debug.enable
690 Enable/disable the CPU sampling based debugging.
691 0: default value, disable debugging
692 1: enable debugging at boot time
694 cpuidle.off=1 [CPU_IDLE]
695 disable the cpuidle sub-system
698 [CPU_IDLE] Name of the cpuidle governor to use.
700 cpufreq.off=1 [CPU_FREQ]
701 disable the cpufreq sub-system
704 [X86] Delay for N microsec between assert and de-assert
705 of APIC INIT to start processors. This delay occurs
706 on every CPU online, such as boot, and resume from suspend.
709 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
711 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
713 crashkernel=size[KMG][@offset[KMG]]
714 [KNL] Using kexec, Linux can switch to a 'crash kernel'
715 upon panic. This parameter reserves the physical
716 memory region [offset, offset + size] for that kernel
717 image. If '@offset' is omitted, then a suitable offset
718 is selected automatically.
719 [KNL, x86_64] select a region under 4G first, and
720 fall back to reserve region above 4G when '@offset'
721 hasn't been specified.
722 See Documentation/admin-guide/kdump/kdump.rst for further details.
724 crashkernel=range1:size1[,range2:size2,...][@offset]
725 [KNL] Same as above, but depends on the memory
726 in the running system. The syntax of range is
727 start-[end] where start and end are both
728 a memory unit (amount[KMG]). See also
729 Documentation/admin-guide/kdump/kdump.rst for an example.
731 crashkernel=size[KMG],high
732 [KNL, x86_64] range could be above 4G. Allow kernel
733 to allocate physical memory region from top, so could
734 be above 4G if system have more than 4G ram installed.
735 Otherwise memory region will be allocated below 4G, if
737 It will be ignored if crashkernel=X is specified.
738 crashkernel=size[KMG],low
739 [KNL, x86_64] range under 4G. When crashkernel=X,high
740 is passed, kernel could allocate physical memory region
741 above 4G, that cause second kernel crash on system
742 that require some amount of low memory, e.g. swiotlb
743 requires at least 64M+32K low memory, also enough extra
744 low memory is needed to make sure DMA buffers for 32-bit
745 devices won't run out. Kernel would try to allocate at
746 at least 256M below 4G automatically.
747 This one let user to specify own low range under 4G
748 for second kernel instead.
749 0: to disable low allocation.
750 It will be ignored when crashkernel=X,high is not used
751 or memory reserved is below 4G.
754 [KNL] Disable crypto self-tests
759 cs89x0_media= [HW,NET]
760 Format: { rj45 | aui | bnc }
763 See header of drivers/s390/block/dasd_devmap.c.
765 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
766 (one device per port)
767 Format: <port#>,<type>
768 See also Documentation/input/devices/joystick-parport.rst
770 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
772 Documentation/admin-guide/dynamic-debug-howto.rst for
773 details. Deprecated, see dyndbg.
775 debug [KNL] Enable kernel debugging (events log level).
778 [KNL] Enable printing [hashed] pointers early in the
779 boot sequence. If enabled, we use a weak hash instead
780 of siphash to hash pointers. Use this option if you are
781 seeing instances of '(___ptrval___)') and need to see a
782 value (hashed pointer) instead. Cryptographically
783 insecure, please do not use on production kernels.
786 [KNL] verbose self-tests
788 Print debugging info while doing the locking API
790 We default to 0 (no extra messages), setting it to
791 1 will print _a lot_ more information - normally
792 only useful to kernel developers.
794 debug_objects [KNL] Enable object debugging
797 [KNL] Disable object debugging
799 debug_guardpage_minorder=
800 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
801 parameter allows control of the order of pages that will
802 be intentionally kept free (and hence protected) by the
803 buddy allocator. Bigger value increase the probability
804 of catching random memory corruption, but reduce the
805 amount of memory for normal system use. The maximum
806 possible value is MAX_ORDER/2. Setting this parameter
807 to 1 or 2 should be enough to identify most random
808 memory corruption problems caused by bugs in kernel or
809 driver code when a CPU writes to (or reads from) a
810 random memory location. Note that there exists a class
811 of memory corruptions problems caused by buggy H/W or
812 F/W or by drivers badly programing DMA (basically when
813 memory is written at bus level and the CPU MMU is
814 bypassed) which are not detectable by
815 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
816 tracking down these problems.
819 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
820 enables the feature at boot time. By default, it is
821 disabled and the system will work mostly the same as a
822 kernel built without CONFIG_DEBUG_PAGEALLOC.
823 Note: to get most of debug_pagealloc error reports, it's
824 useful to also enable the page_owner functionality.
825 on: enable the feature
827 debugpat [X86] Enable PAT debugging
829 decnet.addr= [HW,NET]
830 Format: <area>[,<node>]
831 See also Documentation/networking/decnet.txt.
834 [same as hugepagesz=] The size of the default
835 HugeTLB page size. This is the size represented by
836 the legacy /proc/ hugepages APIs, used for SHM, and
837 default size when mounting hugetlbfs filesystems.
838 Defaults to the default architecture's huge page size
841 deferred_probe_timeout=
842 [KNL] Debugging option to set a timeout in seconds for
843 deferred probe to give up waiting on dependencies to
844 probe. Only specific dependencies (subsystems or
845 drivers) that have opted in will be ignored. A timeout of 0
846 will timeout at the end of initcalls. This option will also
847 dump out devices still on the deferred probe list after
851 Format: { on | off | def_only | inf_only | always }
852 on: s390 zlib hardware support for compression on
853 level 1 and decompression (default)
854 off: No s390 zlib hardware support
855 def_only: s390 zlib hardware support for deflate
856 only (compression on level 1)
857 inf_only: s390 zlib hardware support for inflate
859 always: Same as 'on' but ignores the selected compression
860 level always using hardware support (used for debugging)
863 Set number of hash buckets for dentry cache.
865 disable_1tb_segments [PPC]
866 Disables the use of 1TB hash page table segments. This
867 causes the kernel to fall back to 256MB segments which
868 can be useful when debugging issues that require an SLB
872 See Documentation/networking/ipv6.txt.
875 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
876 hardening is enabled for this boot. Hardened
877 usercopy checking is used to protect the kernel
878 from reading or writing beyond known memory
879 allocation boundaries as a proactive defense
880 against bounds-checking flaws in the kernel's
881 copy_to_user()/copy_from_user() interface.
882 on Perform hardened usercopy checks (default).
883 off Disable hardened usercopy checks.
886 Disable RADIX MMU mode on POWER9
889 Disable TLBIE instruction. Currently does not work
890 with KVM, with HASH MMU, or with coherent accelerators.
892 disable_cpu_apicid= [X86,APIC,SMP]
894 The number of initial APIC ID for the
895 corresponding CPU to be disabled at boot,
896 mostly used for the kdump 2nd kernel to
897 disable BSP to wake up multiple CPUs without
898 causing system reset or hang due to sending
901 perf_v4_pmi= [X86,INTEL]
903 Disable Intel PMU counter freezing feature.
904 The feature only exists starting from
905 Arch Perfmon v4 (Skylake and newer).
907 disable_ddw [PPC/PSERIES]
908 Disable Dynamic DMA Window support. Use this if
909 to workaround buggy firmware.
912 See Documentation/networking/ipv6.txt.
914 disable_mtrr_cleanup [X86]
915 The kernel tries to adjust MTRR layout from continuous
916 to discrete, to make X server driver able to add WB
917 entry later. This parameter disables that.
919 disable_mtrr_trim [X86, Intel and AMD only]
920 By default the kernel will trim any uncacheable
921 memory out of your available memory pool based on
922 MTRR settings. This parameter disables that behavior,
923 possibly causing your machine to run very slowly.
925 disable_timer_pin_1 [X86]
926 Disable PIN 1 of APIC timer
927 Can be useful to work around chipset bugs.
929 dis_ucode_ldr [X86] Disable the microcode loader.
931 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
932 this option disables the debugging code at boot.
934 dma_debug_entries=<number>
935 This option allows to tune the number of preallocated
936 entries for DMA-API debugging code. One entry is
937 required per DMA-API allocation. Use this if the
938 DMA-API debugging code disables itself because the
939 architectural default is too low.
941 dma_debug_driver=<driver_name>
942 With this option the DMA-API debugging driver
943 filter feature can be enabled at boot time. Just
944 pass the driver to filter for as the parameter.
945 The filter can be disabled or changed to another
946 driver later using sysfs.
948 driver_async_probe= [KNL]
949 List of driver names to be probed asynchronously.
950 Format: <driver_name1>,<driver_name2>...
952 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
953 Broken monitors, graphic adapters, KVMs and EDIDless
954 panels may send no or incorrect EDID data sets.
955 This parameter allows to specify an EDID data sets
956 in the /lib/firmware directory that are used instead.
957 Generic built-in EDID data sets are used, if one of
958 edid/1024x768.bin, edid/1280x1024.bin,
959 edid/1680x1050.bin, or edid/1920x1080.bin is given
960 and no file with the same name exists. Details and
961 instructions how to build your own EDID data are
962 available in Documentation/driver-api/edid.rst. An EDID
963 data set will only be used for a particular connector,
964 if its name and a colon are prepended to the EDID
965 name. Each connector may use a unique EDID data
966 set by separating the files with a comma. An EDID
967 data set with no connector name will be used for
968 any connectors not explicitly specified.
973 Format: {"off" | "known"}
974 Control how the dt_cpu_ftrs device-tree binding is
975 used for CPU feature discovery and setup (if it
977 off: Do not use it, fall back to legacy cpu table.
978 known: Do not pass through unknown features to guests
979 or userspace, only those that the kernel is aware of.
981 dump_apple_properties [X86]
982 Dump name and content of EFI device properties on
983 x86 Macs. Useful for driver authors to determine
984 what data is available or for reverse-engineering.
986 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
987 module.dyndbg[="val"]
988 Enable debug messages at boot time. See
989 Documentation/admin-guide/dynamic-debug-howto.rst
992 nompx [X86] Disables Intel Memory Protection Extensions.
993 See Documentation/x86/intel_mpx.rst for more
994 information about the feature.
996 nopku [X86] Disable Memory Protection Keys CPU feature found
999 module.async_probe [KNL]
1000 Enable asynchronous probe on this module.
1002 early_ioremap_debug [KNL]
1003 Enable debug messages in early_ioremap support. This
1004 is useful for tracking down temporary early mappings
1005 which are not unmapped.
1007 earlycon= [KNL] Output early console device and options.
1009 When used with no options, the early console is
1010 determined by stdout-path property in device tree's
1011 chosen node or the ACPI SPCR table if supported by
1014 cdns,<addr>[,options]
1015 Start an early, polled-mode console on a Cadence
1016 (xuartps) serial port at the specified address. Only
1017 supported option is baud rate. If baud rate is not
1018 specified, the serial port must already be setup and
1021 uart[8250],io,<addr>[,options]
1022 uart[8250],mmio,<addr>[,options]
1023 uart[8250],mmio32,<addr>[,options]
1024 uart[8250],mmio32be,<addr>[,options]
1025 uart[8250],0x<addr>[,options]
1026 Start an early, polled-mode console on the 8250/16550
1027 UART at the specified I/O port or MMIO address.
1028 MMIO inter-register address stride is either 8-bit
1029 (mmio) or 32-bit (mmio32 or mmio32be).
1030 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1031 to be equivalent to 'mmio'. 'options' are specified
1032 in the same format described for "console=ttyS<n>"; if
1033 unspecified, the h/w is not initialized.
1037 Start an early, polled-mode console on a pl011 serial
1038 port at the specified address. The pl011 serial port
1039 must already be setup and configured. Options are not
1040 yet supported. If 'mmio32' is specified, then only
1041 the driver will use only 32-bit accessors to read/write
1042 the device registers.
1045 Start an early, polled-mode console on a meson serial
1046 port at the specified address. The serial port must
1047 already be setup and configured. Options are not yet
1051 Start an early, polled-mode console on an msm serial
1052 port at the specified address. The serial port
1053 must already be setup and configured. Options are not
1056 msm_serial_dm,<addr>
1057 Start an early, polled-mode console on an msm serial
1058 dm port at the specified address. The serial port
1059 must already be setup and configured. Options are not
1063 Start an early, polled-mode console on a serial port
1064 of an Actions Semi SoC, such as S500 or S900, at the
1065 specified address. The serial port must already be
1066 setup and configured. Options are not yet supported.
1069 Start an early, polled-mode console on a serial port
1070 of an RDA Micro SoC, such as RDA8810PL, at the
1071 specified address. The serial port must already be
1072 setup and configured. Options are not yet supported.
1075 Use RISC-V SBI (Supervisor Binary Interface) for early
1078 smh Use ARM semihosting calls for early console.
1086 Use early console provided by serial driver available
1087 on Samsung SoCs, requires selecting proper type and
1088 a correct base address of the selected UART port. The
1089 serial port must already be setup and configured.
1090 Options are not yet supported.
1093 Start an early, polled-mode console on a lantiq serial
1094 (lqasc) port at the specified address. The serial port
1095 must already be setup and configured. Options are not
1100 Use early console provided by Freescale LP UART driver
1101 found on Freescale Vybrid and QorIQ LS1021A processors.
1102 A valid base address must be provided, and the serial
1103 port must already be setup and configured.
1106 Start an early, polled-mode console on the
1107 Armada 3700 serial port at the specified
1108 address. The serial port must already be setup
1109 and configured. Options are not yet supported.
1112 Start an early, polled-mode console on a Qualcomm
1113 Generic Interface (GENI) based serial port at the
1114 specified address. The serial port must already be
1115 setup and configured. Options are not yet supported.
1118 Start an early, unaccelerated console on the EFI
1119 memory mapped framebuffer (if available). On cache
1120 coherent non-x86 systems that use system memory for
1121 the framebuffer, pass the 'ram' option so that it is
1122 mapped with the correct attributes.
1125 Use early console provided by Freescale LINFlexD UART
1126 serial driver for NXP S32V234 SoCs. A valid base
1127 address must be provided, and the serial port must
1128 already be setup and configured.
1130 earlyprintk= [X86,SH,ARM,M68k,S390]
1134 earlyprintk=serial[,ttySn[,baudrate]]
1135 earlyprintk=serial[,0x...[,baudrate]]
1136 earlyprintk=ttySn[,baudrate]
1137 earlyprintk=dbgp[debugController#]
1138 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1139 earlyprintk=xdbc[xhciController#]
1141 earlyprintk is useful when the kernel crashes before
1142 the normal console is initialized. It is not enabled by
1143 default because it has some cosmetic problems.
1145 Append ",keep" to not disable it when the real console
1148 Only one of vga, efi, serial, or usb debug port can
1151 Currently only ttyS0 and ttyS1 may be specified by
1152 name. Other I/O ports may be explicitly specified
1153 on some architectures (x86 and arm at least) by
1154 replacing ttySn with an I/O port address, like this:
1155 earlyprintk=serial,0x1008,115200
1156 You can find the port for a given device in
1157 /proc/tty/driver/serial:
1158 2: uart:ST16650V2 port:00001008 irq:18 ...
1160 Interaction with the standard serial driver is not
1163 The VGA and EFI output is eventually overwritten by
1166 The xen output can only be used by Xen PV guests.
1168 The sclp output can only be used on s390.
1170 The optional "force" to "pciserial" enables use of a
1171 PCI device even when its classcode is not of the
1174 edac_report= [HW,EDAC] Control how to report EDAC event
1175 Format: {"on" | "off" | "force"}
1176 on: enable EDAC to report H/W event. May be overridden
1177 by other higher priority error reporting module.
1178 off: disable H/W event reporting through EDAC.
1179 force: enforce the use of EDAC to report H/W event.
1182 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1185 This is designed to be used in conjunction with
1186 the boot argument: earlyprintk=vga
1189 Format: {"off" | "on" | "skip[mbr]"}
1192 Format: { "old_map", "nochunk", "noruntime", "debug",
1193 "nosoftreserve", "disable_early_pci_dma",
1194 "no_disable_early_pci_dma" }
1195 old_map [X86-64]: switch to the old ioremap-based EFI
1196 runtime services mapping. [Needs CONFIG_X86_UV=y]
1197 nochunk: disable reading files in "chunks" in the EFI
1198 boot stub, as chunking can cause problems with some
1199 firmware implementations.
1200 noruntime : disable EFI runtime services support
1201 debug: enable misc debug output
1202 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1203 attribute may cause the kernel to reserve the
1204 memory range for a memory mapping driver to
1205 claim. Specify efi=nosoftreserve to disable this
1206 reservation and treat the memory by its base type
1207 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1208 disable_early_pci_dma: Disable the busmaster bit on all
1209 PCI bridges while in the EFI boot stub
1210 no_disable_early_pci_dma: Leave the busmaster bit set
1211 on all PCI bridges while in the EFI boot stub
1213 efi_no_storage_paranoia [EFI; X86]
1214 Using this parameter you can use more than 50% of
1215 your efi variable storage. Use this parameter only if
1216 you are really sure that your UEFI does sane gc and
1217 fulfills the spec otherwise your board may brick.
1219 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1220 Add arbitrary attribute to specific memory range by
1221 updating original EFI memory map.
1222 Region of memory which aa attribute is added to is
1225 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1226 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1227 attribute is added to range 0x100000000-0x180000000 and
1228 0x10a0000000-0x1120000000.
1230 If efi_fake_mem=8G@9G:0x40000 is specified, the
1231 EFI_MEMORY_SP(0x40000) attribute is added to
1232 range 0x240000000-0x43fffffff.
1234 Using this parameter you can do debugging of EFI memmap
1235 related features. For example, you can do debugging of
1236 Address Range Mirroring feature even if your box
1237 doesn't support it, or mark specific memory as
1240 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1241 that is to be dynamically loaded by Linux. If there are
1242 multiple variables with the same name but with different
1243 vendor GUIDs, all of them will be loaded. See
1244 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1247 eisa_irq_edge= [PARISC,HW]
1248 See header of drivers/parisc/eisa.c.
1251 See comment before function elanfreq_setup() in
1252 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1254 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1255 Specifies physical address of start of kernel core
1256 image elf header and optionally the size. Generally
1257 kexec loader will pass this option to capture kernel.
1258 See Documentation/admin-guide/kdump/kdump.rst for details.
1260 enable_mtrr_cleanup [X86]
1261 The kernel tries to adjust MTRR layout from continuous
1262 to discrete, to make X server driver able to add WB
1263 entry later. This parameter enables that.
1265 enable_timer_pin_1 [X86]
1266 Enable PIN 1 of APIC timer
1267 Can be useful to work around chipset bugs
1268 (in particular on some ATI chipsets).
1269 The kernel tries to set a reasonable default.
1271 enforcing [SELINUX] Set initial enforcing status.
1273 See security/selinux/Kconfig help text.
1274 0 -- permissive (log only, no denials).
1275 1 -- enforcing (deny and log).
1277 Value can be changed at runtime via
1278 /sys/fs/selinux/enforce.
1281 Disable Error Record Serialization Table (ERST)
1284 ether= [HW,NET] Ethernet cards parameters
1285 This option is obsoleted by the "netdev=" option, which
1286 has equivalent usage. See its documentation for details.
1290 Permit 'security.evm' to be updated regardless of
1291 current integrity status.
1295 fail_make_request=[KNL]
1296 General fault injection mechanism.
1297 Format: <interval>,<probability>,<space>,<times>
1298 See also Documentation/fault-injection/.
1301 See Documentation/admin-guide/blockdev/floppy.rst.
1303 force_pal_cache_flush
1304 [IA-64] Avoid check_sal_cache_flush which may hang on
1305 buggy SAL_CACHE_FLUSH implementations. Using this
1306 parameter will force ia64_sal_cache_flush to call
1307 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1310 Forcefully enable Physical Address Extension (PAE).
1311 Many Pentium M systems disable PAE but may have a
1312 functionally usable PAE implementation.
1313 Warning: use of this parameter will taint the kernel
1314 and may cause unknown problems.
1317 [FTRACE] will set and start the specified tracer
1318 as early as possible in order to facilitate early
1321 ftrace_dump_on_oops[=orig_cpu]
1322 [FTRACE] will dump the trace buffers on oops.
1323 If no parameter is passed, ftrace will dump
1324 buffers of all CPUs, but if you pass orig_cpu, it will
1325 dump only the buffer of the CPU that triggered the
1328 ftrace_filter=[function-list]
1329 [FTRACE] Limit the functions traced by the function
1330 tracer at boot up. function-list is a comma separated
1331 list of functions. This list can be changed at run
1332 time by the set_ftrace_filter file in the debugfs
1335 ftrace_notrace=[function-list]
1336 [FTRACE] Do not trace the functions specified in
1337 function-list. This list can be changed at run time
1338 by the set_ftrace_notrace file in the debugfs
1341 ftrace_graph_filter=[function-list]
1342 [FTRACE] Limit the top level callers functions traced
1343 by the function graph tracer at boot up.
1344 function-list is a comma separated list of functions
1345 that can be changed at run time by the
1346 set_graph_function file in the debugfs tracing directory.
1348 ftrace_graph_notrace=[function-list]
1349 [FTRACE] Do not trace from the functions specified in
1350 function-list. This list is a comma separated list of
1351 functions that can be changed at run time by the
1352 set_graph_notrace file in the debugfs tracing directory.
1354 ftrace_graph_max_depth=<uint>
1355 [FTRACE] Used with the function graph tracer. This is
1356 the max depth it will trace into a function. This value
1357 can be changed at run time by the max_graph_depth file
1358 in the tracefs tracing directory. default: 0 (no limit)
1361 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1362 support via parallel port (up to 5 devices per port)
1363 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1364 See also Documentation/input/devices/joystick-parport.rst
1368 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1372 gcov_persist= [GCOV] When non-zero (default), profiling data for
1373 kernel modules is saved and remains accessible via
1374 debugfs, even when the module is unloaded/reloaded.
1375 When zero, profiling data is discarded and associated
1376 debugfs files are removed at module unload time.
1378 goldfish [X86] Enable the goldfish android emulator platform.
1379 Don't use this when you are not running on the
1382 gpt [EFI] Forces disk with valid GPT signature but
1383 invalid Protective MBR to be treated as GPT. If the
1384 primary GPT is corrupted, it enables the backup/alternate
1385 GPT to be used instead.
1387 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1388 the "Enable 0" bit of the configuration register.
1391 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1392 the "Enable 0" bit of the configuration register.
1395 grcan.select= [HW] Select which physical interface to use.
1398 grcan.txsize= [HW] Sets the size of the tx buffer.
1399 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1401 grcan.rxsize= [HW] Sets the size of the rx buffer.
1402 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1405 gpio-mockup.gpio_mockup_ranges
1406 [HW] Sets the ranges of gpiochip of for this device.
1407 Format: <start1>,<end1>,<start2>,<end2>...
1409 hardlockup_all_cpu_backtrace=
1410 [KNL] Should the hard-lockup detector generate
1411 backtraces on all cpus.
1414 hashdist= [KNL,NUMA] Large hashes allocated during boot
1415 are distributed across NUMA nodes. Defaults on
1416 for 64-bit NUMA, off otherwise.
1417 Format: 0 | 1 (for off | on)
1419 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1421 hd= [EIDE] (E)IDE hard drive subsystem geometry
1422 Format: <cyl>,<head>,<sect>
1425 Disable Hardware Error Source Table (HEST) support;
1426 corresponding firmware-first mode error processing
1427 logic will be disabled.
1429 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1430 size of <nn>. This works even on boxes that have no
1431 highmem otherwise. This also works to reduce highmem
1432 size on bigger boxes.
1434 highres= [KNL] Enable/disable high resolution timer mode.
1435 Valid parameters: "on", "off"
1440 hpet= [X86-32,HPET] option to control HPET usage
1441 Format: { enable (default) | disable | force |
1443 disable: disable HPET and use PIT instead
1444 force: allow force enabled of undocumented chips (ICH4,
1446 verbose: show contents of HPET registers during setup
1448 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1449 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1451 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1452 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1453 On x86-64 and powerpc, this option can be specified
1454 multiple times interleaved with hugepages= to reserve
1455 huge pages of different sizes. Valid pages sizes on
1456 x86-64 are 2M (when the CPU supports "pse") and 1G
1457 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1460 [KNL] Should the hung task detector generate panics.
1463 A nonzero value instructs the kernel to panic when a
1464 hung task is detected. The default value is controlled
1465 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1466 option. The value selected by this boot parameter can
1467 be changed later by the kernel.hung_task_panic sysctl.
1469 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1470 terminal devices. Valid values: 0..8
1471 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1472 If specified, z/VM IUCV HVC accepts connections
1473 from listed z/VM user IDs only.
1475 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1476 which allow the hypervisor to 'idle' the
1477 guest on lock contention.
1480 Do not unregister boot console at start. This is only
1481 useful for debugging when something happens in the window
1482 between unregistering the boot console and initializing
1485 i2c_bus= [HW] Override the default board specific I2C bus speed
1486 or register an additional I2C bus that is not
1487 registered from board initialization code.
1491 i8042.debug [HW] Toggle i8042 debug mode
1492 i8042.unmask_kbd_data
1493 [HW] Enable printing of interrupt data from the KBD port
1494 (disabled by default, and as a pre-condition
1495 requires that i8042.debug=1 be enabled)
1496 i8042.direct [HW] Put keyboard port into non-translated mode
1497 i8042.dumbkbd [HW] Pretend that controller can only read data from
1498 keyboard and cannot control its state
1499 (Don't attempt to blink the leds)
1500 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1501 i8042.nokbd [HW] Don't check/create keyboard port
1502 i8042.noloop [HW] Disable the AUX Loopback command while probing
1504 i8042.nomux [HW] Don't check presence of an active multiplexing
1506 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1508 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1509 i8042.reset [HW] Reset the controller during init, cleanup and
1510 suspend-to-ram transitions, only during s2r
1511 transitions, or never reset
1512 Format: { 1 | Y | y | 0 | N | n }
1513 1, Y, y: always reset controller
1514 0, N, n: don't ever reset controller
1515 Default: only on s2r transitions on x86; most other
1516 architectures force reset to be always executed
1517 i8042.unlock [HW] Unlock (ignore) the keylock
1518 i8042.kbdreset [HW] Reset device connected to KBD port
1522 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1523 indicates that the driver is running on unsupported
1525 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1526 does not match list of supported models.
1528 [HW] Report power status in /proc/i8k
1529 (disabled by default)
1530 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1533 i915.invert_brightness=
1534 [DRM] Invert the sense of the variable that is used to
1535 set the brightness of the panel backlight. Normally a
1536 brightness value of 0 indicates backlight switched off,
1537 and the maximum of the brightness value sets the backlight
1538 to maximum brightness. If this parameter is set to 0
1539 (default) and the machine requires it, or this parameter
1540 is set to 1, a brightness value of 0 sets the backlight
1541 to maximum brightness, and the maximum of the brightness
1542 value switches the backlight off.
1543 -1 -- never invert brightness
1544 0 -- machine default
1545 1 -- force brightness inversion
1548 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1550 ide-core.nodma= [HW] (E)IDE subsystem
1551 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1552 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1553 .cdrom .chs .ignore_cable are additional options
1554 See Documentation/ide/ide.rst.
1556 ide-generic.probe-mask= [HW] (E)IDE subsystem
1558 Probe mask for legacy ISA IDE ports. Depending on
1559 platform up to 6 ports are supported, enabled by
1560 setting corresponding bits in the mask to 1. The
1561 default value is 0x0, which has a special meaning.
1562 On systems that have PCI, it triggers scanning the
1563 PCI bus for the first and the second port, which
1564 are then probed. On systems without PCI the value
1565 of 0x0 enables probing the two first ports as if it
1568 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1569 Claim all unknown PCI IDE storage controllers.
1572 Format: idle=poll, idle=halt, idle=nomwait
1573 Poll forces a polling idle loop that can slightly
1574 improve the performance of waking up a idle CPU, but
1575 will use a lot of power and make the system run hot.
1577 idle=halt: Halt is forced to be used for CPU idle.
1578 In such case C2/C3 won't be used again.
1579 idle=nomwait: Disable mwait for CPU C-states
1581 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1582 Format: { strict | legacy | 2008 | relaxed }
1585 Choose which programs will be accepted for execution
1586 based on the IEEE 754 NaN encoding(s) supported by
1587 the FPU and the NaN encoding requested with the value
1588 of an ELF file header flag individually set by each
1589 binary. Hardware implementations are permitted to
1590 support either or both of the legacy and the 2008 NaN
1593 Available settings are as follows:
1594 strict accept binaries that request a NaN encoding
1595 supported by the FPU
1596 legacy only accept legacy-NaN binaries, if supported
1598 2008 only accept 2008-NaN binaries, if supported
1600 relaxed accept any binaries regardless of whether
1601 supported by the FPU
1603 The FPU emulator is always able to support both NaN
1604 encodings, so if no FPU hardware is present or it has
1605 been disabled with 'nofpu', then the settings of
1606 'legacy' and '2008' strap the emulator accordingly,
1607 'relaxed' straps the emulator for both legacy-NaN and
1608 2008-NaN, whereas 'strict' enables legacy-NaN only on
1609 legacy processors and both NaN encodings on MIPS32 or
1612 The setting for ABS.fmt/NEG.fmt instruction execution
1613 mode generally follows that for the NaN encoding,
1614 except where unsupported by hardware.
1616 ignore_loglevel [KNL]
1617 Ignore loglevel setting - this will print /all/
1618 kernel messages to the console. Useful for debugging.
1619 We also add it as printk module parameter, so users
1620 could change it dynamically, usually by
1621 /sys/module/printk/parameters/ignore_loglevel.
1624 Ignore RLIMIT_DATA setting for data mappings,
1625 print warning at first misuse. Can be changed via
1626 /sys/module/kernel/parameters/ignore_rlimit_data.
1628 ihash_entries= [KNL]
1629 Set number of hash buckets for inode cache.
1631 ima_appraise= [IMA] appraise integrity measurements
1632 Format: { "off" | "enforce" | "fix" | "log" }
1635 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1636 The builtin appraise policy appraises all files
1639 ima_canonical_fmt [IMA]
1640 Use the canonical format for the binary runtime
1641 measurements, instead of host native format.
1644 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1648 The list of supported hash algorithms is defined
1649 in crypto/hash_info.h.
1652 The builtin policies to load during IMA setup.
1653 Format: "tcb | appraise_tcb | secure_boot |
1656 The "tcb" policy measures all programs exec'd, files
1657 mmap'd for exec, and all files opened with the read
1658 mode bit set by either the effective uid (euid=0) or
1661 The "appraise_tcb" policy appraises the integrity of
1662 all files owned by root.
1664 The "secure_boot" policy appraises the integrity
1665 of files (eg. kexec kernel image, kernel modules,
1666 firmware, policy, etc) based on file signatures.
1668 The "fail_securely" policy forces file signature
1669 verification failure also on privileged mounted
1670 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1673 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1674 Load a policy which meets the needs of the Trusted
1675 Computing Base. This means IMA will measure all
1676 programs exec'd, files mmap'd for exec, and all files
1677 opened for read by uid=0.
1680 Select one of defined IMA measurements template formats.
1681 Formats: { "ima" | "ima-ng" | "ima-sig" }
1685 [IMA] Define a custom template format.
1686 Format: { "field1|...|fieldN" }
1688 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1689 Format: <min_file_size>
1690 Set the minimal file size for using asynchronous hash.
1691 If left unspecified, ahash usage is disabled.
1693 ahash performance varies for different data sizes on
1694 different crypto accelerators. This option can be used
1695 to achieve the best performance for a particular HW.
1697 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1699 Set hashing buffer size. Default: 4k.
1701 ahash performance varies for different chunk sizes on
1702 different crypto accelerators. This option can be used
1703 to achieve best performance for particular HW.
1707 Run specified binary instead of /sbin/init as init
1710 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1711 for working out where the kernel is dying during
1714 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1715 initcall functions. Useful for debugging built-in
1716 modules and initcalls.
1718 initrd= [BOOT] Specify the location of the initial ramdisk
1720 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1723 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1725 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1727 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1729 init_pkru= [x86] Specify the default memory protection keys rights
1730 register contents for all processes. 0x55555554 by
1731 default (disallow access to all but pkey 0). Can
1732 override in debugfs after boot.
1734 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1737 int_pln_enable [x86] Enable power limit notification interrupt
1739 integrity_audit=[IMA]
1740 Format: { "0" | "1" }
1741 0 -- basic integrity auditing messages. (Default)
1742 1 -- additional integrity auditing messages.
1744 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1746 Enable intel iommu driver.
1748 Disable intel iommu driver.
1749 igfx_off [Default Off]
1750 By default, gfx is mapped as normal device. If a gfx
1751 device has a dedicated DMAR unit, the DMAR unit is
1752 bypassed by not enabling DMAR with this option. In
1753 this case, gfx device will use physical address for
1756 With this option iommu will not optimize to look
1757 for io virtual address below 32-bit forcing dual
1758 address cycle on pci bus for cards supporting greater
1759 than 32-bit addressing. The default is to look
1760 for translation below 32-bit and if not available
1761 then look in the higher range.
1762 strict [Default Off]
1763 With this option on every unmap_single operation will
1764 result in a hardware IOTLB flush operation as opposed
1765 to batching them for performance.
1766 sp_off [Default Off]
1767 By default, super page will be supported if Intel IOMMU
1768 has the capability. With this option, super page will
1771 By default, scalable mode will be disabled even if the
1772 hardware advertises that it has support for the scalable
1773 mode translation. With this option set, scalable mode
1774 will be used on hardware which claims to support it.
1775 tboot_noforce [Default Off]
1776 Do not force the Intel IOMMU enabled under tboot.
1777 By default, tboot will force Intel IOMMU on, which
1778 could harm performance of some high-throughput
1779 devices like 40GBit network cards, even if identity
1781 Note that using this option lowers the security
1782 provided by tboot because it makes the system
1783 vulnerable to DMA attacks.
1784 nobounce [Default off]
1785 Disable bounce buffer for unstrusted devices such as
1786 the Thunderbolt devices. This will treat the untrusted
1787 devices as the trusted ones, hence might expose security
1788 risks of DMA attacks.
1790 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1791 0 disables intel_idle and fall back on acpi_idle.
1792 1 to 9 specify maximum depth of C-state.
1796 Do not enable intel_pstate as the default
1797 scaling driver for the supported processors
1799 Use intel_pstate as a scaling driver, but configure it
1800 to work with generic cpufreq governors (instead of
1801 enabling its internal governor). This mode cannot be
1802 used along with the hardware-managed P-states (HWP)
1805 Enable intel_pstate on systems that prohibit it by default
1806 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1807 instead of acpi-cpufreq may disable platform features, such
1808 as thermal controls and power capping, that rely on ACPI
1809 P-States information being indicated to OSPM and therefore
1810 should be used with caution. This option does not work with
1811 processors that aren't supported by the intel_pstate driver
1812 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1814 Do not enable hardware P state control (HWP)
1817 Only load intel_pstate on systems which support
1818 hardware P state control (HWP) if available.
1820 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1821 Description Table, specifies preferred power management
1822 profile as "Enterprise Server" or "Performance Server",
1823 then this feature is turned on by default.
1825 Allow per-logical-CPU P-State performance control limits using
1826 cpufreq sysfs interface
1828 intremap= [X86-64, Intel-IOMMU]
1829 on enable Interrupt Remapping (default)
1830 off disable Interrupt Remapping
1831 nosid disable Source ID checking
1833 BIOS x2APIC opt-out request will be ignored
1834 nopost disable Interrupt Posting
1836 iomem= Disable strict checking of access to MMIO memory
1837 strict regions from userspace.
1852 nobypass [PPC/POWERNV]
1853 Disable IOMMU bypass, using IOMMU for PCI devices.
1855 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1856 Format: { "0" | "1" }
1858 Request that DMA unmap operations use deferred
1859 invalidation of hardware TLBs, for increased
1860 throughput at the cost of reduced device isolation.
1861 Will fall back to strict mode if not supported by
1862 the relevant IOMMU driver.
1863 1 - Strict mode (default).
1864 DMA unmap operations invalidate IOMMU hardware TLBs
1868 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
1869 Format: { "0" | "1" }
1870 0 - Use IOMMU translation for DMA.
1871 1 - Bypass the IOMMU for DMA.
1872 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
1874 io7= [HW] IO7 for Marvel based alpha systems
1875 See comment before marvel_specify_io7 in
1876 arch/alpha/kernel/core_marvel.c.
1878 io_delay= [X86] I/O delay method
1880 Standard port 0x80 based delay
1882 Alternate port 0xed based delay (needed on some systems)
1884 Simple two microseconds delay
1889 See Documentation/filesystems/nfs/nfsroot.txt.
1891 ipcmni_extend [KNL] Extend the maximum number of unique System V
1892 IPC identifiers from 32,768 to 16,777,216.
1894 irqaffinity= [SMP] Set the default irq affinity mask
1895 The argument is a cpu list, as described above.
1897 irqchip.gicv2_force_probe=
1900 Force the kernel to look for the second 4kB page
1901 of a GICv2 controller even if the memory range
1902 exposed by the device tree is too small.
1904 irqchip.gicv3_nolpi=
1906 Force the kernel to ignore the availability of
1907 LPIs (and by consequence ITSs). Intended for system
1908 that use the kernel as a bootloader, and thus want
1909 to let secondary kernels in charge of setting up
1912 irqchip.gicv3_pseudo_nmi= [ARM64]
1913 Enables support for pseudo-NMIs in the kernel. This
1914 requires the kernel to be built with
1915 CONFIG_ARM64_PSEUDO_NMI.
1918 When an interrupt is not handled search all handlers
1919 for it. Intended to get systems with badly broken
1923 When an interrupt is not handled search all handlers
1924 for it. Also check all handlers each timer
1925 interrupt. Intended to get systems with badly broken
1929 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1931 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1932 [Deprecated - use cpusets instead]
1933 Format: [flag-list,]<cpu-list>
1935 Specify one or more CPUs to isolate from disturbances
1936 specified in the flag list (default: domain):
1939 Disable the tick when a single task runs.
1941 A residual 1Hz tick is offloaded to workqueues, which you
1942 need to affine to housekeeping through the global
1943 workqueue's affinity configured via the
1944 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1945 by using the 'domain' flag described below.
1947 NOTE: by default the global workqueue runs on all CPUs,
1948 so to protect individual CPUs the 'cpumask' file has to
1949 be configured manually after bootup.
1952 Isolate from the general SMP balancing and scheduling
1953 algorithms. Note that performing domain isolation this way
1954 is irreversible: it's not possible to bring back a CPU to
1955 the domains once isolated through isolcpus. It's strongly
1956 advised to use cpusets instead to disable scheduler load
1957 balancing through the "cpuset.sched_load_balance" file.
1958 It offers a much more flexible interface where CPUs can
1959 move in and out of an isolated set anytime.
1961 You can move a process onto or off an "isolated" CPU via
1962 the CPU affinity syscalls or cpuset.
1963 <cpu number> begins at 0 and the maximum value is
1964 "number of CPUs in system - 1".
1968 Isolate from being targeted by managed interrupts
1969 which have an interrupt mask containing isolated
1970 CPUs. The affinity of managed interrupts is
1971 handled by the kernel and cannot be changed via
1972 the /proc/irq/* interfaces.
1974 This isolation is best effort and only effective
1975 if the automatically assigned interrupt mask of a
1976 device queue contains isolated and housekeeping
1977 CPUs. If housekeeping CPUs are online then such
1978 interrupts are directed to the housekeeping CPU
1979 so that IO submitted on the housekeeping CPU
1980 cannot disturb the isolated CPU.
1982 If a queue's affinity mask contains only isolated
1983 CPUs then this parameter has no effect on the
1984 interrupt routing decision, though interrupts are
1985 only delivered when tasks running on those
1986 isolated CPUs submit IO. IO submitted on
1987 housekeeping CPUs has no influence on those
1990 The format of <cpu-list> is described above.
1994 ivrs_ioapic [HW,X86_64]
1995 Provide an override to the IOAPIC-ID<->DEVICE-ID
1996 mapping provided in the IVRS ACPI table. For
1997 example, to map IOAPIC-ID decimal 10 to
1998 PCI device 00:14.0 write the parameter as:
1999 ivrs_ioapic[10]=00:14.0
2001 ivrs_hpet [HW,X86_64]
2002 Provide an override to the HPET-ID<->DEVICE-ID
2003 mapping provided in the IVRS ACPI table. For
2004 example, to map HPET-ID decimal 0 to
2005 PCI device 00:14.0 write the parameter as:
2006 ivrs_hpet[0]=00:14.0
2008 ivrs_acpihid [HW,X86_64]
2009 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2010 mapping provided in the IVRS ACPI table. For
2011 example, to map UART-HID:UID AMD0020:0 to
2012 PCI device 00:14.5 write the parameter as:
2013 ivrs_acpihid[00:14.5]=AMD0020:0
2015 js= [HW,JOY] Analog joystick
2016 See Documentation/input/joydev/joystick.rst.
2019 When CONFIG_RANDOMIZE_BASE is set, this disables
2020 kernel and module base offset ASLR (Address Space
2021 Layout Randomization).
2024 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2025 report on every invalid memory access. Without this
2026 parameter KASAN will print report only for the first
2031 kernelcore= [KNL,X86,IA-64,PPC]
2032 Format: nn[KMGTPE] | nn% | "mirror"
2033 This parameter specifies the amount of memory usable by
2034 the kernel for non-movable allocations. The requested
2035 amount is spread evenly throughout all nodes in the
2036 system as ZONE_NORMAL. The remaining memory is used for
2037 movable memory in its own zone, ZONE_MOVABLE. In the
2038 event, a node is too small to have both ZONE_NORMAL and
2039 ZONE_MOVABLE, kernelcore memory will take priority and
2040 other nodes will have a larger ZONE_MOVABLE.
2042 ZONE_MOVABLE is used for the allocation of pages that
2043 may be reclaimed or moved by the page migration
2044 subsystem. Note that allocations like PTEs-from-HighMem
2045 still use the HighMem zone if it exists, and the Normal
2046 zone if it does not.
2048 It is possible to specify the exact amount of memory in
2049 the form of "nn[KMGTPE]", a percentage of total system
2050 memory in the form of "nn%", or "mirror". If "mirror"
2051 option is specified, mirrored (reliable) memory is used
2052 for non-movable allocations and remaining memory is used
2053 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2054 are exclusive, so you cannot specify multiple forms.
2056 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2057 Format: <Controller#>[,poll interval]
2058 The controller # is the number of the ehci usb debug
2059 port as it is probed via PCI. The poll interval is
2060 optional and is the number seconds in between
2061 each poll cycle to the debug port in case you need
2062 the functionality for interrupting the kernel with
2063 gdb or control-c on the dbgp connection. When
2064 not using this parameter you use sysrq-g to break into
2065 the kernel debugger.
2067 kgdboc= [KGDB,HW] kgdb over consoles.
2068 Requires a tty driver that supports console polling,
2069 or a supported polling keyboard driver (non-usb).
2070 Serial only format: <serial_device>[,baud]
2071 keyboard only format: kbd
2072 keyboard and serial format: kbd,<serial_device>[,baud]
2073 Optional Kernel mode setting:
2074 kms, kbd format: kms,kbd
2075 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2077 kgdbwait [KGDB] Stop kernel execution and enter the
2078 kernel debugger at the earliest opportunity.
2080 kmac= [MIPS] korina ethernet MAC address.
2081 Configure the RouterBoard 532 series on-chip
2082 Ethernet adapter MAC address.
2084 kmemleak= [KNL] Boot-time kmemleak enable/disable
2085 Valid arguments: on, off
2087 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2090 kprobe_event=[probe-list]
2091 [FTRACE] Add kprobe events and enable at boot time.
2092 The probe-list is a semicolon delimited list of probe
2093 definitions. Each definition is same as kprobe_events
2094 interface, but the parameters are comma delimited.
2095 For example, to add a kprobe event on vfs_read with
2096 arg1 and arg2, add to the command line;
2098 kprobe_event=p,vfs_read,$arg1,$arg2
2100 See also Documentation/trace/kprobetrace.rst "Kernel
2101 Boot Parameter" section.
2103 kpti= [ARM64] Control page table isolation of user
2104 and kernel address spaces.
2105 Default: enabled on cores which need mitigation.
2109 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2110 Default is 0 (don't ignore, but inject #GP)
2112 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2113 Default is false (don't support).
2115 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2120 [KVM] Controls the software workaround for the
2121 X86_BUG_ITLB_MULTIHIT bug.
2122 force : Always deploy workaround.
2123 off : Never deploy workaround.
2124 auto : Deploy workaround based on the presence of
2125 X86_BUG_ITLB_MULTIHIT.
2129 If the software workaround is enabled for the host,
2130 guests do need not to enable it for nested guests.
2132 kvm.nx_huge_pages_recovery_ratio=
2133 [KVM] Controls how many 4KiB pages are periodically zapped
2134 back to huge pages. 0 disables the recovery, otherwise if
2135 the value is N KVM will zap 1/Nth of the 4KiB pages every
2136 minute. The default is 60.
2138 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2139 Default is 1 (enabled)
2141 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2143 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2145 kvm-arm.vgic_v3_group0_trap=
2146 [KVM,ARM] Trap guest accesses to GICv3 group-0
2149 kvm-arm.vgic_v3_group1_trap=
2150 [KVM,ARM] Trap guest accesses to GICv3 group-1
2153 kvm-arm.vgic_v3_common_trap=
2154 [KVM,ARM] Trap guest accesses to GICv3 common
2157 kvm-arm.vgic_v4_enable=
2158 [KVM,ARM] Allow use of GICv4 for direct injection of
2161 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2162 (virtualized MMU) support on capable Intel chips.
2163 Default is 1 (enabled)
2165 kvm-intel.emulate_invalid_guest_state=
2166 [KVM,Intel] Enable emulation of invalid guest states
2167 Default is 0 (disabled)
2169 kvm-intel.flexpriority=
2170 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2171 Default is 1 (enabled)
2174 [KVM,Intel] Enable VMX nesting (nVMX).
2175 Default is 0 (disabled)
2177 kvm-intel.unrestricted_guest=
2178 [KVM,Intel] Disable unrestricted guest feature
2179 (virtualized real and unpaged mode) on capable
2180 Intel chips. Default is 1 (enabled)
2182 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2185 Valid arguments: never, cond, always
2187 always: L1D cache flush on every VMENTER.
2188 cond: Flush L1D on VMENTER only when the code between
2189 VMEXIT and VMENTER can leak host memory.
2190 never: Disables the mitigation
2192 Default is cond (do L1 cache flush in specific instances)
2194 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2195 feature (tagged TLBs) on capable Intel chips.
2196 Default is 1 (enabled)
2198 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2201 The kernel PTE inversion protection is unconditionally
2202 enabled and cannot be disabled.
2205 Provides all available mitigations for the
2206 L1TF vulnerability. Disables SMT and
2207 enables all mitigations in the
2208 hypervisors, i.e. unconditional L1D flush.
2210 SMT control and L1D flush control via the
2211 sysfs interface is still possible after
2212 boot. Hypervisors will issue a warning
2213 when the first VM is started in a
2214 potentially insecure configuration,
2215 i.e. SMT enabled or L1D flush disabled.
2218 Same as 'full', but disables SMT and L1D
2219 flush runtime control. Implies the
2220 'nosmt=force' command line option.
2221 (i.e. sysfs control of SMT is disabled.)
2224 Leaves SMT enabled and enables the default
2225 hypervisor mitigation, i.e. conditional
2228 SMT control and L1D flush control via the
2229 sysfs interface is still possible after
2230 boot. Hypervisors will issue a warning
2231 when the first VM is started in a
2232 potentially insecure configuration,
2233 i.e. SMT enabled or L1D flush disabled.
2237 Disables SMT and enables the default
2238 hypervisor mitigation.
2240 SMT control and L1D flush control via the
2241 sysfs interface is still possible after
2242 boot. Hypervisors will issue a warning
2243 when the first VM is started in a
2244 potentially insecure configuration,
2245 i.e. SMT enabled or L1D flush disabled.
2248 Same as 'flush', but hypervisors will not
2249 warn when a VM is started in a potentially
2250 insecure configuration.
2253 Disables hypervisor mitigations and doesn't
2255 It also drops the swap size and available
2256 RAM limit restriction on both hypervisor and
2261 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2267 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2270 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2271 value for LAPIC timer one-shot implementation. Default
2272 back to the programmable timer unit in the LAPIC.
2274 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2277 libata.dma= [LIBATA] DMA control
2278 libata.dma=0 Disable all PATA and SATA DMA
2279 libata.dma=1 PATA and SATA Disk DMA only
2280 libata.dma=2 ATAPI (CDROM) DMA only
2281 libata.dma=4 Compact Flash DMA only
2282 Combinations also work, so libata.dma=3 enables DMA
2283 for disks and CDROMs, but not CFs.
2285 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2286 libata.ignore_hpa=0 keep BIOS limits (default)
2287 libata.ignore_hpa=1 ignore limits, using full disk
2289 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2293 libata.force= [LIBATA] Force configurations. The format is comma
2294 separated list of "[ID:]VAL" where ID is
2295 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2296 matching port, link or device. Basically, it matches
2297 the ATA ID string printed on console by libata. If
2298 the whole ID part is omitted, the last PORT and DEVICE
2299 values are used. If ID hasn't been specified yet, the
2300 configuration applies to all ports, links and devices.
2302 If only DEVICE is omitted, the parameter applies to
2303 the port and all links and devices behind it. DEVICE
2304 number of 0 either selects the first device or the
2305 first fan-out link behind PMP device. It does not
2306 select the host link. DEVICE number of 15 selects the
2307 host link and device attached to it.
2309 The VAL specifies the configuration to force. As long
2310 as there's no ambiguity shortcut notation is allowed.
2311 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2312 The following configurations can be forced.
2314 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2315 Any ID with matching PORT is used.
2317 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2319 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2320 udma[/][16,25,33,44,66,100,133] notation is also
2323 * [no]ncq: Turn on or off NCQ.
2325 * [no]ncqtrim: Turn off queued DSM TRIM.
2327 * nohrst, nosrst, norst: suppress hard, soft
2330 * rstonce: only attempt one reset during
2331 hot-unplug link recovery
2333 * dump_id: dump IDENTIFY data.
2335 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2337 * disable: Disable this device.
2339 If there are multiple matching configurations changing
2340 the same attribute, the last one is used.
2342 memblock=debug [KNL] Enable memblock debug messages.
2344 load_ramdisk= [RAM] List of ramdisks to load from floppy
2345 See Documentation/admin-guide/blockdev/ramdisk.rst.
2347 lockd.nlm_grace_period=P [NFS] Assign grace period.
2350 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2353 lockd.nlm_timeout=T [NFS] Assign timeout value.
2356 lockd.nlm_udpport=M [NFS] Assign UDP port.
2359 lockdown= [SECURITY]
2360 { integrity | confidentiality }
2361 Enable the kernel lockdown feature. If set to
2362 integrity, kernel features that allow userland to
2363 modify the running kernel are disabled. If set to
2364 confidentiality, kernel features that allow userland
2365 to extract confidential information from the kernel
2368 locktorture.nreaders_stress= [KNL]
2369 Set the number of locking read-acquisition kthreads.
2370 Defaults to being automatically set based on the
2371 number of online CPUs.
2373 locktorture.nwriters_stress= [KNL]
2374 Set the number of locking write-acquisition kthreads.
2376 locktorture.onoff_holdoff= [KNL]
2377 Set time (s) after boot for CPU-hotplug testing.
2379 locktorture.onoff_interval= [KNL]
2380 Set time (s) between CPU-hotplug operations, or
2381 zero to disable CPU-hotplug testing.
2383 locktorture.shuffle_interval= [KNL]
2384 Set task-shuffle interval (jiffies). Shuffling
2385 tasks allows some CPUs to go into dyntick-idle
2386 mode during the locktorture test.
2388 locktorture.shutdown_secs= [KNL]
2389 Set time (s) after boot system shutdown. This
2390 is useful for hands-off automated testing.
2392 locktorture.stat_interval= [KNL]
2393 Time (s) between statistics printk()s.
2395 locktorture.stutter= [KNL]
2396 Time (s) to stutter testing, for example,
2397 specifying five seconds causes the test to run for
2398 five seconds, wait for five seconds, and so on.
2399 This tests the locking primitive's ability to
2400 transition abruptly to and from idle.
2402 locktorture.torture_type= [KNL]
2403 Specify the locking implementation to test.
2405 locktorture.verbose= [KNL]
2406 Enable additional printk() statements.
2408 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2411 loglevel= All Kernel Messages with a loglevel smaller than the
2412 console loglevel will be printed to the console. It can
2413 also be changed with klogd or other programs. The
2414 loglevels are defined as follows:
2416 0 (KERN_EMERG) system is unusable
2417 1 (KERN_ALERT) action must be taken immediately
2418 2 (KERN_CRIT) critical conditions
2419 3 (KERN_ERR) error conditions
2420 4 (KERN_WARNING) warning conditions
2421 5 (KERN_NOTICE) normal but significant condition
2422 6 (KERN_INFO) informational
2423 7 (KERN_DEBUG) debug-level messages
2425 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2426 in bytes. n must be a power of two and greater
2427 than the minimal size. The minimal size is defined
2428 by LOG_BUF_SHIFT kernel config parameter. There is
2429 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2430 that allows to increase the default size depending on
2431 the number of CPUs. See init/Kconfig for more details.
2433 logo.nologo [FB] Disables display of the built-in Linux logo.
2434 This may be used to provide more screen space for
2435 kernel log messages and is useful when debugging
2436 kernel boot problems.
2438 lp=0 [LP] Specify parallel ports to use, e.g,
2439 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2440 lp=reset first parallel port). 'lp=0' disables the
2441 lp=auto printer driver. 'lp=reset' (which can be
2442 specified in addition to the ports) causes
2443 attached printers to be reset. Using
2444 lp=port1,port2,... specifies the parallel ports
2445 to associate lp devices with, starting with
2446 lp0. A port specification may be 'none' to skip
2447 that lp device, or a parport name such as
2448 'parport0'. Specifying 'lp=auto' instead of a
2449 port specification list means that device IDs
2450 from each port should be examined, to see if
2451 an IEEE 1284-compliant printer is attached; if
2452 so, the driver will manage that printer.
2453 See also header of drivers/char/lp.c.
2456 Sets loops_per_jiffy to given constant, thus avoiding
2457 time-consuming boot-time autodetection (up to 250 ms per
2458 CPU). 0 enables autodetection (default). To determine
2459 the correct value for your kernel, boot with normal
2460 autodetection and see what value is printed. Note that
2461 on SMP systems the preset will be applied to all CPUs,
2462 which is likely to cause problems if your CPUs need
2463 significantly divergent settings. An incorrect value
2464 will cause delays in the kernel to be wrong, leading to
2465 unpredictable I/O errors and other breakage. Although
2466 unlikely, in the extreme case this might damage your
2470 Format: <io>,<irq>,<dma>
2472 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2475 [SECURITY] Choose order of LSM initialization. This
2476 overrides CONFIG_LSM, and the "security=" parameter.
2478 machvec= [IA-64] Force the use of a particular machine-vector
2479 (machvec) in a generic kernel.
2480 Example: machvec=hpzx1
2482 machtype= [Loongson] Share the same kernel image file between different
2484 Example: machtype=lemote-yeeloong-2f-7inch
2486 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2487 than or equal to this physical address is ignored.
2489 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2490 will bring up during bootup. maxcpus=n : n >= 0 limits
2491 the kernel to bring up 'n' processors. Surely after
2492 bootup you can bring up the other plugged cpu by executing
2493 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2494 only takes effect during system bootup.
2495 While n=0 is a special case, it is equivalent to "nosmp",
2496 which also disables the IO APIC.
2498 max_loop= [LOOP] The number of loop block devices that get
2499 (loop.max_loop) unconditionally pre-created at init time. The default
2500 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2501 of statically allocating a predefined number, loop
2502 devices can be requested on-demand with the
2503 /dev/loop-control interface.
2505 mce [X86-32] Machine Check Exception
2507 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2509 md= [HW] RAID subsystems devices and level
2510 See Documentation/admin-guide/md.rst.
2513 Format: <first>,<last>
2514 Specifies range of consoles to be captured by the MDA.
2517 Control mitigation for the Micro-architectural Data
2518 Sampling (MDS) vulnerability.
2520 Certain CPUs are vulnerable to an exploit against CPU
2521 internal buffers which can forward information to a
2522 disclosure gadget under certain conditions.
2524 In vulnerable processors, the speculatively
2525 forwarded data can be used in a cache side channel
2526 attack, to access data to which the attacker does
2527 not have direct access.
2529 This parameter controls the MDS mitigation. The
2532 full - Enable MDS mitigation on vulnerable CPUs
2533 full,nosmt - Enable MDS mitigation and disable
2534 SMT on vulnerable CPUs
2535 off - Unconditionally disable MDS mitigation
2537 On TAA-affected machines, mds=off can be prevented by
2538 an active TAA mitigation as both vulnerabilities are
2539 mitigated with the same mechanism so in order to disable
2540 this mitigation, you need to specify tsx_async_abort=off
2543 Not specifying this option is equivalent to
2546 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2548 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2549 Amount of memory to be used when the kernel is not able
2550 to see the whole system memory or for test.
2551 [X86] Work as limiting max address. Use together
2552 with memmap= to avoid physical address space collisions.
2553 Without memmap= PCI devices could be placed at addresses
2554 belonging to unused RAM.
2556 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2560 [KNL,SH] Allow user to override the default size for
2561 per-device physically contiguous DMA buffers.
2563 memhp_default_state=online/offline
2564 [KNL] Set the initial state for the memory hotplug
2565 onlining policy. If not specified, the default value is
2566 set according to the
2567 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2569 See Documentation/admin-guide/mm/memory-hotplug.rst.
2571 memmap=exactmap [KNL,X86] Enable setting of an exact
2572 E820 memory map, as specified by the user.
2573 Such memmap=exactmap lines can be constructed based on
2574 BIOS output or other requirements. See the memmap=nn@ss
2577 memmap=nn[KMG]@ss[KMG]
2578 [KNL] Force usage of a specific region of memory.
2579 Region of memory to be used is from ss to ss+nn.
2580 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2581 which limits max address to nn[KMG].
2582 Multiple different regions can be specified,
2585 memmap=100M@2G,100M#3G,1G!1024G
2587 memmap=nn[KMG]#ss[KMG]
2588 [KNL,ACPI] Mark specific memory as ACPI data.
2589 Region of memory to be marked is from ss to ss+nn.
2591 memmap=nn[KMG]$ss[KMG]
2592 [KNL,ACPI] Mark specific memory as reserved.
2593 Region of memory to be reserved is from ss to ss+nn.
2594 Example: Exclude memory from 0x18690000-0x1869ffff
2595 memmap=64K$0x18690000
2597 memmap=0x10000$0x18690000
2598 Some bootloaders may need an escape character before '$',
2599 like Grub2, otherwise '$' and the following number
2602 memmap=nn[KMG]!ss[KMG]
2603 [KNL,X86] Mark specific memory as protected.
2604 Region of memory to be used, from ss to ss+nn.
2605 The memory region may be marked as e820 type 12 (0xc)
2606 and is NVDIMM or ADR memory.
2608 memmap=<size>%<offset>-<oldtype>+<newtype>
2609 [KNL,ACPI] Convert memory within the specified region
2610 from <oldtype> to <newtype>. If "-<oldtype>" is left
2611 out, the whole region will be marked as <newtype>,
2612 even if previously unavailable. If "+<newtype>" is left
2613 out, matching memory will be removed. Types are
2614 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2615 3 = ACPI, 12 = PRAM.
2617 memory_corruption_check=0/1 [X86]
2618 Some BIOSes seem to corrupt the first 64k of
2619 memory when doing things like suspend/resume.
2620 Setting this option will scan the memory
2621 looking for corruption. Enabling this will
2622 both detect corruption and prevent the kernel
2623 from using the memory being corrupted.
2624 However, its intended as a diagnostic tool; if
2625 repeatable BIOS-originated corruption always
2626 affects the same memory, you can use memmap=
2627 to prevent the kernel from using that memory.
2629 memory_corruption_check_size=size [X86]
2630 By default it checks for corruption in the low
2631 64k, making this memory unavailable for normal
2632 use. Use this parameter to scan for
2633 corruption in more or less memory.
2635 memory_corruption_check_period=seconds [X86]
2636 By default it checks for corruption every 60
2637 seconds. Use this parameter to check at some
2638 other rate. 0 disables periodic checking.
2640 memtest= [KNL,X86,ARM,PPC] Enable memtest
2642 default : 0 <disable>
2643 Specifies the number of memtest passes to be
2644 performed. Each pass selects another test
2645 pattern from a given set of patterns. Memtest
2646 fills the memory with this pattern, validates
2647 memory contents and reserves bad memory
2648 regions that are detected.
2650 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2651 Valid arguments: on, off
2652 Default (depends on kernel configuration option):
2653 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2654 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2655 mem_encrypt=on: Activate SME
2656 mem_encrypt=off: Do not activate SME
2658 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2659 for details on when memory encryption can be activated.
2661 mem_sleep_default= [SUSPEND] Default system suspend mode:
2662 s2idle - Suspend-To-Idle
2663 shallow - Power-On Suspend or equivalent (if supported)
2664 deep - Suspend-To-RAM or equivalent (if supported)
2665 See Documentation/admin-guide/pm/sleep-states.rst.
2667 meye.*= [HW] Set MotionEye Camera parameters
2668 See Documentation/media/v4l-drivers/meye.rst.
2670 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2671 Multi-Function General Purpose Timers on AMD Geode
2674 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2675 the BIOS has incorrectly applied a workaround. TinyBIOS
2676 version 0.98 is known to be affected, 0.99 fixes the
2677 problem by letting the user disable the workaround.
2681 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2682 physical address is ignored.
2684 mini2440= [ARM,HW,KNL]
2685 Format:[0..2][b][c][t]
2687 MINI2440 configuration specification:
2688 0 - The attached screen is the 3.5" TFT
2689 1 - The attached screen is the 7" TFT
2690 2 - The VGA Shield is attached (1024x768)
2691 Leaving out the screen size parameter will not load
2692 the TFT driver, and the framebuffer will be left
2694 b - Enable backlight. The TFT backlight pin will be
2695 linked to the kernel VESA blanking code and a GPIO
2696 LED. This parameter is not necessary when using the
2698 c - Enable the s3c camera interface.
2699 t - Reserved for enabling touchscreen support. The
2700 touchscreen support is not enabled in the mainstream
2701 kernel as of 2.6.30, a preliminary port can be found
2702 in the "bleeding edge" mini2440 support kernel at
2703 http://repo.or.cz/w/linux-2.6/mini2440.git
2706 [X86,PPC,S390,ARM64] Control optional mitigations for
2707 CPU vulnerabilities. This is a set of curated,
2708 arch-independent options, each of which is an
2709 aggregation of existing arch-specific options.
2712 Disable all optional CPU mitigations. This
2713 improves system performance, but it may also
2714 expose users to several CPU vulnerabilities.
2715 Equivalent to: nopti [X86,PPC]
2717 nospectre_v1 [X86,PPC]
2719 nospectre_v2 [X86,PPC,S390,ARM64]
2720 spectre_v2_user=off [X86]
2721 spec_store_bypass_disable=off [X86,PPC]
2722 ssbd=force-off [ARM64]
2725 tsx_async_abort=off [X86]
2726 kvm.nx_huge_pages=off [X86]
2729 This does not have any effect on
2730 kvm.nx_huge_pages when
2731 kvm.nx_huge_pages=force.
2734 Mitigate all CPU vulnerabilities, but leave SMT
2735 enabled, even if it's vulnerable. This is for
2736 users who don't want to be surprised by SMT
2737 getting disabled across kernel upgrades, or who
2738 have other ways of avoiding SMT-based attacks.
2739 Equivalent to: (default behavior)
2742 Mitigate all CPU vulnerabilities, disabling SMT
2743 if needed. This is for users who always want to
2744 be fully mitigated, even if it means losing SMT.
2745 Equivalent to: l1tf=flush,nosmt [X86]
2746 mds=full,nosmt [X86]
2747 tsx_async_abort=full,nosmt [X86]
2750 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2751 parameter allows control of the logging verbosity for
2752 the additional memory initialisation checks. A value
2753 of 0 disables mminit logging and a level of 4 will
2754 log everything. Information is printed at KERN_DEBUG
2755 so loglevel=8 may also need to be specified.
2758 [KNL] When CONFIG_MODULE_SIG is set, this means that
2759 modules without (valid) signatures will fail to load.
2760 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2761 is always true, so this option does nothing.
2763 module_blacklist= [KNL] Do not load a comma-separated list of
2764 modules. Useful for debugging problem modules.
2767 [MOUSE] Maximum time between finger touching and
2768 leaving touchpad surface for touch to be considered
2769 a tap and be reported as a left button click (for
2770 touchpads working in absolute mode only).
2772 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2773 reporting absolute coordinates, such as tablets
2774 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2775 reporting absolute coordinates, such as tablets
2777 movablecore= [KNL,X86,IA-64,PPC]
2778 Format: nn[KMGTPE] | nn%
2779 This parameter is the complement to kernelcore=, it
2780 specifies the amount of memory used for migratable
2781 allocations. If both kernelcore and movablecore is
2782 specified, then kernelcore will be at *least* the
2783 specified value but may be more. If movablecore on its
2784 own is specified, the administrator must be careful
2785 that the amount of memory usable for all allocations
2788 movable_node [KNL] Boot-time switch to make hotplugable memory
2789 NUMA nodes to be movable. This means that the memory
2790 of such nodes will be usable only for movable
2791 allocations which rules out almost all kernel
2792 allocations. Use with caution!
2794 MTD_Partition= [MTD]
2795 Format: <name>,<region-number>,<size>,<offset>
2797 MTD_Region= [MTD] Format:
2798 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2801 See drivers/mtd/cmdlinepart.c.
2803 multitce=off [PPC] This parameter disables the use of the pSeries
2804 firmware feature for updating multiple TCE entries
2807 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2809 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2811 boundary - index of last SLC block on Flex-OneNAND.
2812 The remaining blocks are configured as MLC blocks.
2813 lock - Configure if Flex-OneNAND boundary should be locked.
2814 Once locked, the boundary cannot be changed.
2815 1 indicates lock status, 0 indicates unlock status.
2818 ARM/S3C2412 JIVE boot control
2820 See arch/arm/mach-s3c2412/mach-jive.c
2822 mtouchusb.raw_coordinates=
2823 [HW] Make the MicroTouch USB driver use raw coordinates
2824 ('y', default) or cooked coordinates ('n')
2826 mtrr_chunk_size=nn[KMG] [X86]
2827 used for mtrr cleanup. It is largest continuous chunk
2828 that could hold holes aka. UC entries.
2830 mtrr_gran_size=nn[KMG] [X86]
2831 Used for mtrr cleanup. It is granularity of mtrr block.
2833 Large value could prevent small alignment from
2836 mtrr_spare_reg_nr=n [X86]
2838 Range: 0,7 : spare reg number
2840 Used for mtrr cleanup. It is spare mtrr entries number.
2841 Set to 2 or more if your graphical card needs more.
2843 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2845 netdev= [NET] Network devices parameters
2846 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2847 Note that mem_start is often overloaded to mean
2848 something different and driver-specific.
2849 This usage is only documented in each driver source
2853 [NETFILTER] Enable connection tracking flow accounting
2854 0 to disable accounting
2855 1 to enable accounting
2858 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2859 See Documentation/filesystems/nfs/nfsroot.txt.
2861 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2862 See Documentation/filesystems/nfs/nfsroot.txt.
2864 nfsrootdebug [NFS] enable nfsroot debugging messages.
2865 See Documentation/filesystems/nfs/nfsroot.txt.
2867 nfs.callback_nr_threads=
2868 [NFSv4] set the total number of threads that the
2869 NFS client will assign to service NFSv4 callback
2872 nfs.callback_tcpport=
2873 [NFS] set the TCP port on which the NFSv4 callback
2874 channel should listen.
2877 [NFS] sets the pathname to the program which is used
2878 to update the NFS client cache entries.
2880 nfs.cache_getent_timeout=
2881 [NFS] sets the timeout after which an attempt to
2882 update a cache entry is deemed to have failed.
2884 nfs.idmap_cache_timeout=
2885 [NFS] set the maximum lifetime for idmapper cache
2889 [NFS] enable 64-bit inode numbers.
2890 If zero, the NFS client will fake up a 32-bit inode
2891 number for the readdir() and stat() syscalls instead
2892 of returning the full 64-bit number.
2893 The default is to return 64-bit inode numbers.
2895 nfs.max_session_cb_slots=
2896 [NFSv4.1] Sets the maximum number of session
2897 slots the client will assign to the callback
2898 channel. This determines the maximum number of
2899 callbacks the client will process in parallel for
2900 a particular server.
2902 nfs.max_session_slots=
2903 [NFSv4.1] Sets the maximum number of session slots
2904 the client will attempt to negotiate with the server.
2905 This limits the number of simultaneous RPC requests
2906 that the client can send to the NFSv4.1 server.
2907 Note that there is little point in setting this
2908 value higher than the max_tcp_slot_table_limit.
2910 nfs.nfs4_disable_idmapping=
2911 [NFSv4] When set to the default of '1', this option
2912 ensures that both the RPC level authentication
2913 scheme and the NFS level operations agree to use
2914 numeric uids/gids if the mount is using the
2915 'sec=sys' security flavour. In effect it is
2916 disabling idmapping, which can make migration from
2917 legacy NFSv2/v3 systems to NFSv4 easier.
2918 Servers that do not support this mode of operation
2919 will be autodetected by the client, and it will fall
2920 back to using the idmapper.
2921 To turn off this behaviour, set the value to '0'.
2923 [NFS4] Specify an additional fixed unique ident-
2924 ification string that NFSv4 clients can insert into
2925 their nfs_client_id4 string. This is typically a
2926 UUID that is generated at system install time.
2928 nfs.send_implementation_id =
2929 [NFSv4.1] Send client implementation identification
2930 information in exchange_id requests.
2931 If zero, no implementation identification information
2933 The default is to send the implementation identification
2936 nfs.recover_lost_locks =
2937 [NFSv4] Attempt to recover locks that were lost due
2938 to a lease timeout on the server. Please note that
2939 doing this risks data corruption, since there are
2940 no guarantees that the file will remain unchanged
2941 after the locks are lost.
2942 If you want to enable the kernel legacy behaviour of
2943 attempting to recover these locks, then set this
2945 The default parameter value of '0' causes the kernel
2946 not to attempt recovery of lost locks.
2948 nfs4.layoutstats_timer =
2949 [NFSv4.2] Change the rate at which the kernel sends
2950 layoutstats to the pNFS metadata server.
2952 Setting this to value to 0 causes the kernel to use
2953 whatever value is the default set by the layout
2954 driver. A non-zero value sets the minimum interval
2955 in seconds between layoutstats transmissions.
2957 nfsd.nfs4_disable_idmapping=
2958 [NFSv4] When set to the default of '1', the NFSv4
2959 server will return only numeric uids and gids to
2960 clients using auth_sys, and will accept numeric uids
2961 and gids from such clients. This is intended to ease
2962 migration from NFSv2/v3.
2964 nmi_debug= [KNL,SH] Specify one or more actions to take
2965 when a NMI is triggered.
2966 Format: [state][,regs][,debounce][,die]
2968 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2969 Format: [panic,][nopanic,][num]
2971 0 - turn hardlockup detector in nmi_watchdog off
2972 1 - turn hardlockup detector in nmi_watchdog on
2973 When panic is specified, panic when an NMI watchdog
2974 timeout occurs (or 'nopanic' to not panic on an NMI
2975 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
2976 To disable both hard and soft lockup detectors,
2977 please see 'nowatchdog'.
2978 This is useful when you use a panic=... timeout and
2979 need the box quickly up again.
2981 These settings can be accessed at runtime via
2982 the nmi_watchdog and hardlockup_panic sysctls.
2984 netpoll.carrier_timeout=
2985 [NET] Specifies amount of time (in seconds) that
2986 netpoll should wait for a carrier. By default netpoll
2989 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2990 emulation library even if a 387 maths coprocessor
2993 no5lvl [X86-64] Disable 5-level paging mode. Forces
2994 kernel to use 4-level paging instead.
2997 [HW] Never suspend the console
2998 Disable suspending of consoles during suspend and
2999 hibernate operations. Once disabled, debugging
3000 messages can reach various consoles while the rest
3001 of the system is being put to sleep (ie, while
3002 debugging driver suspend/resume hooks). This may
3003 not work reliably with all consoles, but is known
3004 to work with serial and VGA consoles.
3005 To facilitate more flexible debugging, we also add
3006 console_suspend, a printk module parameter to control
3007 it. Users could use console_suspend (usually
3008 /sys/module/printk/parameters/console_suspend) to
3009 turn on/off it dynamically.
3011 novmcoredd [KNL,KDUMP]
3012 Disable device dump. Device dump allows drivers to
3013 append dump data to vmcore so you can collect driver
3014 specified debug info. Drivers can append the data
3015 without any limit and this data is stored in memory,
3016 so this may cause significant memory stress. Disabling
3017 device dump can help save memory but the driver debug
3018 data will be no longer available. This parameter
3019 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3022 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3023 caches in the slab allocator. Saves per-node memory,
3024 but will impact performance.
3028 noaltinstr [S390] Disables alternative instructions patching
3029 (CPU alternatives feature).
3031 noapic [SMP,APIC] Tells the kernel to not make use of any
3032 IOAPICs that may be present in the system.
3034 noautogroup Disable scheduler automatic task group creation.
3036 nobats [PPC] Do not use BATs for mapping kernel lowmem
3037 on "Classic" PPC cores.
3041 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3043 nodelayacct [KNL] Disable per-task delay accounting
3045 nodsp [SH] Disable hardware DSP at boot time.
3047 noefi Disable EFI runtime services support.
3052 On X86-32 available only on PAE configured kernels.
3053 noexec=on: enable non-executable mappings (default)
3054 noexec=off: disable non-executable mappings
3057 Disable SMAP (Supervisor Mode Access Prevention)
3058 even if it is supported by processor.
3061 Disable SMEP (Supervisor Mode Execution Prevention)
3062 even if it is supported by processor.
3065 This affects only 32-bit executables.
3066 noexec32=on: enable non-executable mappings (default)
3067 read doesn't imply executable mappings
3068 noexec32=off: disable non-executable mappings
3069 read implies executable mappings
3071 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3073 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3074 register save and restore. The kernel will only save
3075 legacy floating-point registers on task switch.
3077 nohugeiomap [KNL,x86,PPC] Disable kernel huge I/O mappings.
3079 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3080 Equivalent to smt=1.
3082 [KNL,x86] Disable symmetric multithreading (SMT).
3083 nosmt=force: Force disable SMT, cannot be undone
3084 via the sysfs control file.
3086 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3087 (bounds check bypass). With this option data leaks are
3088 possible in the system.
3090 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3091 the Spectre variant 2 (indirect branch prediction)
3092 vulnerability. System may allow data leaks with this
3095 nospec_store_bypass_disable
3096 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3098 noxsave [BUGS=X86] Disables x86 extended register state save
3099 and restore using xsave. The kernel will fallback to
3100 enabling legacy floating-point and sse state.
3102 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3103 register states. The kernel will fall back to use
3104 xsave to save the states. By using this parameter,
3105 performance of saving the states is degraded because
3106 xsave doesn't support modified optimization while
3107 xsaveopt supports it on xsaveopt enabled systems.
3109 noxsaves [X86] Disables xsaves and xrstors used in saving and
3110 restoring x86 extended register state in compacted
3111 form of xsave area. The kernel will fall back to use
3112 xsaveopt and xrstor to save and restore the states
3113 in standard form of xsave area. By using this
3114 parameter, xsave area per process might occupy more
3115 memory on xsaves enabled systems.
3117 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
3118 wfi(ARM) instruction doesn't work correctly and not to
3119 use it. This is also useful when using JTAG debugger.
3121 no_file_caps Tells the kernel not to honor file capabilities. The
3122 only way then for a file to be executed with privilege
3123 is to be setuid root or executed by root.
3125 nohalt [IA-64] Tells the kernel not to use the power saving
3126 function PAL_HALT_LIGHT when idle. This increases
3127 power-consumption. On the positive side, it reduces
3128 interrupt wake-up latency, which may improve performance
3129 in certain environments such as networked servers or
3132 nohibernate [HIBERNATION] Disable hibernation and resume.
3134 nohz= [KNL] Boottime enable/disable dynamic ticks
3135 Valid arguments: on, off
3138 nohz_full= [KNL,BOOT,SMP,ISOL]
3139 The argument is a cpu list, as described above.
3140 In kernels built with CONFIG_NO_HZ_FULL=y, set
3141 the specified list of CPUs whose tick will be stopped
3142 whenever possible. The boot CPU will be forced outside
3143 the range to maintain the timekeeping. Any CPUs
3144 in this list will have their RCU callbacks offloaded,
3145 just as if they had also been called out in the
3146 rcu_nocbs= boot parameter.
3148 noiotrap [SH] Disables trapped I/O port accesses.
3150 noirqdebug [X86-32] Disables the code which attempts to detect and
3151 disable unhandled interrupt sources.
3153 no_timer_check [X86,APIC] Disables the code which tests for
3154 broken timer IRQ sources.
3156 noisapnp [ISAPNP] Disables ISA PnP code.
3158 noinitrd [RAM] Tells the kernel not to load any configured
3161 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3163 [Deprecated - use intremap=off]
3167 noinvpcid [X86] Disable the INVPCID cpu feature.
3169 nojitter [IA-64] Disables jitter checking for ITC timers.
3171 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3173 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3177 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3178 clock and use the default one.
3180 no-steal-acc [X86,KVM,ARM64] Disable paravirtualized steal time
3181 accounting. steal time is computed, but won't
3182 influence scheduler behaviour
3184 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3186 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3188 noltlbs [PPC] Do not use large page/tlb entries for kernel
3189 lowmem mapping on PPC40x and PPC8xx
3191 nomca [IA-64] Disable machine check abort handling
3193 nomce [X86-32] Disable Machine Check Exception
3195 nomfgpt [X86-32] Disable Multi-Function General Purpose
3196 Timer usage (for AMD Geode machines).
3198 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3199 shutdown the other cpus. Instead use the REBOOT_VECTOR
3202 nomodule Disable module load
3204 nopat [X86] Disable PAT (page attribute table extension of
3205 pagetables) support.
3207 nopcid [X86-64] Disable the PCID cpu feature.
3209 norandmaps Don't use address space randomization. Equivalent to
3210 echo 0 > /proc/sys/kernel/randomize_va_space
3212 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3213 with UP alternatives
3215 nordrand [X86] Disable kernel use of the RDRAND and
3216 RDSEED instructions even if they are supported
3217 by the processor. RDRAND and RDSEED are still
3218 available to user space applications.
3220 noresume [SWSUSP] Disables resume and restores original swap
3223 no-scroll [VGA] Disables scrollback.
3224 This is required for the Braillex ib80-piezo Braille
3225 reader made by F.H. Papenmeier (Germany).
3229 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3231 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3232 and disable the IO APIC. legacy for "maxcpus=0".
3234 nosoftlockup [KNL] Disable the soft-lockup detector.
3236 nosync [HW,M68K] Disables sync negotiation for all devices.
3238 nowatchdog [KNL] Disable both lockup detectors, i.e.
3239 soft-lockup and NMI watchdog (hard-lockup).
3243 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3245 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3246 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3247 Some features depend on CPU0. Known dependencies are:
3248 1. Resume from suspend/hibernate depends on CPU0.
3249 Suspend/hibernate will fail if CPU0 is offline and you
3250 need to online CPU0 before suspend/hibernate.
3251 2. PIC interrupts also depend on CPU0. CPU0 can't be
3252 removed if a PIC interrupt is detected.
3253 It's said poweroff/reboot may depend on CPU0 on some
3254 machines although I haven't seen such issues so far
3255 after CPU0 is offline on a few tested machines.
3256 If the dependencies are under your control, you can
3257 turn on cpu0_hotplug.
3259 nps_mtm_hs_ctr= [KNL,ARC]
3260 This parameter sets the maximum duration, in
3261 cycles, each HW thread of the CTOP can run
3262 without interruptions, before HW switches it.
3263 The actual maximum duration is 16 times this
3265 Format: integer between 1 and 255
3268 nptcg= [IA-64] Override max number of concurrent global TLB
3269 purges which is reported from either PAL_VM_SUMMARY or
3272 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3273 could support. nr_cpus=n : n >= 1 limits the kernel to
3274 support 'n' processors. It could be larger than the
3275 number of already plugged CPU during bootup, later in
3276 runtime you can physically add extra cpu until it reaches
3277 n. So during boot up some boot time memory for per-cpu
3278 variables need be pre-allocated for later physical cpu
3281 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3283 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3284 Allowed values are enable and disable
3286 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3287 'node', 'default' can be specified
3288 This can be set from sysctl after boot.
3289 See Documentation/admin-guide/sysctl/vm.rst for details.
3291 of_devlink [OF, KNL] Create device links between consumer and
3292 supplier devices by scanning the devictree to infer the
3293 consumer/supplier relationships. A consumer device
3294 will not be probed until all the supplier devices have
3295 probed successfully.
3297 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3298 See Documentation/debugging-via-ohci1394.txt for more
3301 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3302 Rather than timing out after 20 ms if an EC
3303 command is not properly ACKed, override the length
3304 of the timeout. We have interrupts disabled while
3305 waiting for the ACK, so if this is set too high
3306 interrupts *may* be lost!
3308 omap_mux= [OMAP] Override bootloader pin multiplexing.
3309 Format: <mux_mode0.mode_name=value>...
3310 For example, to override I2C bus2:
3311 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3313 oprofile.timer= [HW]
3314 Use timer interrupt instead of performance counters
3316 oprofile.cpu_type= Force an oprofile cpu type
3317 This might be useful if you have an older oprofile
3318 userland or if you want common events.
3319 Format: { arch_perfmon }
3320 arch_perfmon: [X86] Force use of architectural
3321 perfmon on Intel CPUs instead of the
3322 CPU specific event set.
3323 timer: [X86] Force use of architectural NMI
3324 timer mode (see also oprofile.timer
3325 for generic hr timer mode)
3327 oops=panic Always panic on oopses. Default is to just kill the
3328 process, but there is a small probability of
3329 deadlocking the machine.
3330 This will also cause panics on machine check exceptions.
3331 Useful together with panic=30 to trigger a reboot.
3334 [KNL] Boolean flag to control whether the page allocator
3335 should randomize its free lists. The randomization may
3336 be automatically enabled if the kernel detects it is
3337 running on a platform with a direct-mapped memory-side
3338 cache, and this parameter can be used to
3339 override/disable that behavior. The state of the flag
3340 can be read from sysfs at:
3341 /sys/module/page_alloc/parameters/shuffle.
3343 page_owner= [KNL] Boot-time page_owner enabling option.
3344 Storage of the information about who allocated
3345 each page is disabled in default. With this switch,
3347 on: enable the feature
3349 page_poison= [KNL] Boot-time parameter changing the state of
3350 poisoning on the buddy allocator, available with
3351 CONFIG_PAGE_POISONING=y.
3352 off: turn off poisoning (default)
3353 on: turn on poisoning
3355 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3356 timeout > 0: seconds before rebooting
3357 timeout = 0: wait forever
3358 timeout < 0: reboot immediately
3361 panic_print= Bitmask for printing system info when panic happens.
3362 User can chose combination of the following bits:
3363 bit 0: print all tasks info
3364 bit 1: print system memory info
3365 bit 2: print timer info
3366 bit 3: print locks info if CONFIG_LOCKDEP is on
3367 bit 4: print ftrace buffer
3368 bit 5: print all printk messages in buffer
3370 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3373 crash_kexec_post_notifiers
3374 Run kdump after running panic-notifiers and dumping
3375 kmsg. This only for the users who doubt kdump always
3376 succeeds in any situation.
3377 Note that this also increases risks of kdump failure,
3378 because some panic notifiers can make the crashed
3379 kernel more unstable.
3381 parkbd.port= [HW] Parallel port number the keyboard adapter is
3382 connected to, default is 0.
3384 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3385 0 for XT, 1 for AT (default is AT).
3388 parport= [HW,PPT] Specify parallel ports. 0 disables.
3389 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3390 Use 'auto' to force the driver to use any
3391 IRQ/DMA settings detected (the default is to
3392 ignore detected IRQ/DMA settings because of
3393 possible conflicts). You can specify the base
3394 address, IRQ, and DMA settings; IRQ and DMA
3395 should be numbers, or 'auto' (for using detected
3396 settings on that particular port), or 'nofifo'
3397 (to avoid using a FIFO even if it is detected).
3398 Parallel ports are assigned in the order they
3399 are specified on the command line, starting
3402 parport_init_mode= [HW,PPT]
3403 Configure VIA parallel port to operate in
3404 a specific mode. This is necessary on Pegasos
3405 computer where firmware has no options for setting
3406 up parallel port mode and sets it to spp.
3407 Currently this function knows 686a and 8231 chips.
3408 Format: [spp|ps2|epp|ecp|ecpepp]
3411 Halt all CPUs after the first oops has been printed for
3412 the specified number of seconds. This is to be used if
3413 your oopses keep scrolling off the screen.
3418 See header of drivers/block/paride/pcd.c.
3419 See also Documentation/admin-guide/blockdev/paride.rst.
3421 pci=option[,option...] [PCI] various PCI subsystem options.
3423 Some options herein operate on a specific device
3424 or a set of devices (<pci_dev>). These are
3425 specified in one of the following formats:
3427 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3428 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3430 Note: the first format specifies a PCI
3431 bus/device/function address which may change
3432 if new hardware is inserted, if motherboard
3433 firmware changes, or due to changes caused
3434 by other kernel parameters. If the
3435 domain is left unspecified, it is
3436 taken to be zero. Optionally, a path
3437 to a device through multiple device/function
3438 addresses can be specified after the base
3439 address (this is more robust against
3440 renumbering issues). The second format
3441 selects devices using IDs from the
3442 configuration space which may match multiple
3443 devices in the system.
3445 earlydump dump PCI config space before the kernel
3447 off [X86] don't probe for the PCI bus
3448 bios [X86-32] force use of PCI BIOS, don't access
3449 the hardware directly. Use this if your machine
3450 has a non-standard PCI host bridge.
3451 nobios [X86-32] disallow use of PCI BIOS, only direct
3452 hardware access methods are allowed. Use this
3453 if you experience crashes upon bootup and you
3454 suspect they are caused by the BIOS.
3455 conf1 [X86] Force use of PCI Configuration Access
3456 Mechanism 1 (config address in IO port 0xCF8,
3457 data in IO port 0xCFC, both 32-bit).
3458 conf2 [X86] Force use of PCI Configuration Access
3459 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3460 the function, IO port 0xCFA, also 8-bit, sets
3461 bus number. The config space is then accessed
3462 through ports 0xC000-0xCFFF).
3463 See http://wiki.osdev.org/PCI for more info
3464 on the configuration access mechanisms.
3465 noaer [PCIE] If the PCIEAER kernel config parameter is
3466 enabled, this kernel boot option can be used to
3467 disable the use of PCIE advanced error reporting.
3468 nodomains [PCI] Disable support for multiple PCI
3469 root domains (aka PCI segments, in ACPI-speak).
3470 nommconf [X86] Disable use of MMCONFIG for PCI
3472 check_enable_amd_mmconf [X86] check for and enable
3473 properly configured MMIO access to PCI
3474 config space on AMD family 10h CPU
3475 nomsi [MSI] If the PCI_MSI kernel config parameter is
3476 enabled, this kernel boot option can be used to
3477 disable the use of MSI interrupts system-wide.
3478 noioapicquirk [APIC] Disable all boot interrupt quirks.
3479 Safety option to keep boot IRQs enabled. This
3480 should never be necessary.
3481 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3482 primary IO-APIC for bridges that cannot disable
3483 boot IRQs. This fixes a source of spurious IRQs
3484 when the system masks IRQs.
3485 noioapicreroute [APIC] Disable workaround that uses the
3486 boot IRQ equivalent of an IRQ that connects to
3487 a chipset where boot IRQs cannot be disabled.
3488 The opposite of ioapicreroute.
3489 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3490 routing table. These calls are known to be buggy
3491 on several machines and they hang the machine
3492 when used, but on other computers it's the only
3493 way to get the interrupt routing table. Try
3494 this option if the kernel is unable to allocate
3495 IRQs or discover secondary PCI buses on your
3497 rom [X86] Assign address space to expansion ROMs.
3498 Use with caution as certain devices share
3499 address decoders between ROMs and other
3501 norom [X86] Do not assign address space to
3502 expansion ROMs that do not already have
3503 BIOS assigned address ranges.
3504 nobar [X86] Do not assign address space to the
3505 BARs that weren't assigned by the BIOS.
3506 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3507 assigned automatically to PCI devices. You can
3508 make the kernel exclude IRQs of your ISA cards
3510 pirqaddr=0xAAAAA [X86] Specify the physical address
3511 of the PIRQ table (normally generated
3512 by the BIOS) if it is outside the
3513 F0000h-100000h range.
3514 lastbus=N [X86] Scan all buses thru bus #N. Can be
3515 useful if the kernel is unable to find your
3516 secondary buses and you want to tell it
3517 explicitly which ones they are.
3518 assign-busses [X86] Always assign all PCI bus
3519 numbers ourselves, overriding
3520 whatever the firmware may have done.
3521 usepirqmask [X86] Honor the possible IRQ mask stored
3522 in the BIOS $PIR table. This is needed on
3523 some systems with broken BIOSes, notably
3524 some HP Pavilion N5400 and Omnibook XE3
3525 notebooks. This will have no effect if ACPI
3526 IRQ routing is enabled.
3527 noacpi [X86] Do not use ACPI for IRQ routing
3528 or for PCI scanning.
3529 use_crs [X86] Use PCI host bridge window information
3530 from ACPI. On BIOSes from 2008 or later, this
3531 is enabled by default. If you need to use this,
3532 please report a bug.
3533 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3534 If you need to use this, please report a bug.
3535 routeirq Do IRQ routing for all PCI devices.
3536 This is normally done in pci_enable_device(),
3537 so this option is a temporary workaround
3538 for broken drivers that don't call it.
3539 skip_isa_align [X86] do not align io start addr, so can
3540 handle more pci cards
3541 noearly [X86] Don't do any early type 1 scanning.
3542 This might help on some broken boards which
3543 machine check when some devices' config space
3544 is read. But various workarounds are disabled
3545 and some IOMMU drivers will not work.
3546 bfsort Sort PCI devices into breadth-first order.
3547 This sorting is done to get a device
3548 order compatible with older (<= 2.4) kernels.
3549 nobfsort Don't sort PCI devices into breadth-first order.
3550 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3551 tuning and use the BIOS-configured MPS defaults.
3552 pcie_bus_safe Set every device's MPS to the largest value
3553 supported by all devices below the root complex.
3554 pcie_bus_perf Set device MPS to the largest allowable MPS
3555 based on its parent bus. Also set MRRS (Max
3556 Read Request Size) to the largest supported
3557 value (no larger than the MPS that the device
3558 or bus can support) for best performance.
3559 pcie_bus_peer2peer Set every device's MPS to 128B, which
3560 every device is guaranteed to support. This
3561 configuration allows peer-to-peer DMA between
3562 any pair of devices, possibly at the cost of
3563 reduced performance. This also guarantees
3564 that hot-added devices will work.
3565 cbiosize=nn[KMG] The fixed amount of bus space which is
3566 reserved for the CardBus bridge's IO window.
3567 The default value is 256 bytes.
3568 cbmemsize=nn[KMG] The fixed amount of bus space which is
3569 reserved for the CardBus bridge's memory
3570 window. The default value is 64 megabytes.
3573 [<order of align>@]<pci_dev>[; ...]
3574 Specifies alignment and device to reassign
3575 aligned memory resources. How to
3576 specify the device is described above.
3577 If <order of align> is not specified,
3578 PAGE_SIZE is used as alignment.
3579 A PCI-PCI bridge can be specified if resource
3580 windows need to be expanded.
3581 To specify the alignment for several
3582 instances of a device, the PCI vendor,
3583 device, subvendor, and subdevice may be
3584 specified, e.g., 12@pci:8086:9c22:103c:198f
3585 for 4096-byte alignment.
3586 ecrc= Enable/disable PCIe ECRC (transaction layer
3587 end-to-end CRC checking).
3588 bios: Use BIOS/firmware settings. This is the
3592 hpiosize=nn[KMG] The fixed amount of bus space which is
3593 reserved for hotplug bridge's IO window.
3594 Default size is 256 bytes.
3595 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3596 reserved for hotplug bridge's MMIO window.
3597 Default size is 2 megabytes.
3598 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3599 reserved for hotplug bridge's MMIO_PREF window.
3600 Default size is 2 megabytes.
3601 hpmemsize=nn[KMG] The fixed amount of bus space which is
3602 reserved for hotplug bridge's MMIO and
3604 Default size is 2 megabytes.
3605 hpbussize=nn The minimum amount of additional bus numbers
3606 reserved for buses below a hotplug bridge.
3608 realloc= Enable/disable reallocating PCI bridge resources
3609 if allocations done by BIOS are too small to
3610 accommodate resources required by all child
3612 off: Turn realloc off
3614 realloc same as realloc=on
3615 noari do not use PCIe ARI.
3616 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3617 do not use PCIe ATS (and IOMMU device IOTLB).
3618 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3619 only look for one device below a PCIe downstream
3621 big_root_window Try to add a big 64bit memory window to the PCIe
3622 root complex on AMD CPUs. Some GFX hardware
3623 can resize a BAR to allow access to all VRAM.
3624 Adding the window is slightly risky (it may
3625 conflict with unreported devices), so this
3627 disable_acs_redir=<pci_dev>[; ...]
3628 Specify one or more PCI devices (in the format
3629 specified above) separated by semicolons.
3630 Each device specified will have the PCI ACS
3631 redirect capabilities forced off which will
3632 allow P2P traffic between devices through
3633 bridges without forcing it upstream. Note:
3634 this removes isolation between devices and
3635 may put more devices in an IOMMU group.
3636 force_floating [S390] Force usage of floating interrupts.
3637 nomio [S390] Do not use MIO instructions.
3639 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3642 force Enable ASPM even on devices that claim not to support it.
3643 WARNING: Forcing ASPM on may cause system lockups.
3645 pcie_ports= [PCIE] PCIe port services handling:
3646 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3647 even if the platform doesn't give the OS permission to
3648 use them. This may cause conflicts if the platform
3649 also tries to use these services.
3650 dpc-native Use native PCIe service for DPC only. May
3651 cause conflicts if firmware uses AER or DPC.
3652 compat Disable native PCIe services (PME, AER, DPC, PCIe
3655 pcie_port_pm= [PCIE] PCIe port power management handling:
3656 off Disable power management of all PCIe ports
3657 force Forcibly enable power management of all PCIe ports
3659 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3660 nomsi Do not use MSI for native PCIe PME signaling (this makes
3661 all PCIe root ports use INTx for all services).
3663 pcmv= [HW,PCMCIA] BadgePAD 4
3667 Keep all power-domains already enabled by bootloader on,
3668 even if no driver has claimed them. This is useful
3669 for debug and development, but should not be
3670 needed on a platform with proper driver support.
3673 See Documentation/admin-guide/blockdev/paride.rst.
3675 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3678 See arch/parisc/kernel/pdc_chassis.c
3680 percpu_alloc= Select which percpu first chunk allocator to use.
3681 Currently supported values are "embed" and "page".
3682 Archs may support subset or none of the selections.
3683 See comments in mm/percpu.c for details on each
3684 allocator. This parameter is primarily for debugging
3685 and performance comparison.
3688 See Documentation/admin-guide/blockdev/paride.rst.
3691 See Documentation/admin-guide/blockdev/paride.rst.
3693 pirq= [SMP,APIC] Manual mp-table setup
3694 See Documentation/x86/i386/IO-APIC.rst.
3696 plip= [PPT,NET] Parallel port network link
3697 Format: { parport<nr> | timid | 0 }
3698 See also Documentation/admin-guide/parport.rst.
3700 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3701 Override pmtimer IOPort with a hex value.
3705 Enable PNP debug messages (depends on the
3706 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3707 via /sys/module/pnp/parameters/debug. We always show
3708 current resource usage; turning this on also shows
3709 possible settings and some assignment information.
3715 { on | off | curr | res | no-curr | no-res }
3718 [ISAPNP] Exclude IRQs for the autoconfiguration
3721 [ISAPNP] Exclude DMAs for the autoconfiguration
3723 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3724 Ranges are in pairs (I/O port base and size).
3727 [ISAPNP] Exclude memory regions for the
3729 Ranges are in pairs (memory base and size).
3731 ports= [IP_VS_FTP] IPVS ftp helper module
3733 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3735 Format: <port>,<port>....
3737 powersave=off [PPC] This option disables power saving features.
3738 It specifically disables cpuidle and sets the
3739 platform machine description specific power_save
3740 function to NULL. On Idle the CPU just reduces
3743 ppc_strict_facility_enable
3744 [PPC] This option catches any kernel floating point,
3745 Altivec, VSX and SPE outside of regions specifically
3746 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3747 There is some performance impact when enabling this.
3751 Disable Hardware Transactional Memory
3753 print-fatal-signals=
3754 [KNL] debug: print fatal signals
3756 If enabled, warn about various signal handling
3757 related application anomalies: too many signals,
3758 too many POSIX.1 timers, fatal signals causing a
3761 If you hit the warning due to signal overflow,
3762 you might want to try "ulimit -i unlimited".
3766 printk.always_kmsg_dump=
3767 Trigger kmsg_dump for cases other than kernel oops or
3769 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3772 printk.devkmsg={on,off,ratelimit}
3773 Control writing to /dev/kmsg.
3774 on - unlimited logging to /dev/kmsg from userspace
3775 off - logging to /dev/kmsg disabled
3776 ratelimit - ratelimit the logging
3779 printk.time= Show timing data prefixed to each printk message line
3780 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3782 processor.max_cstate= [HW,ACPI]
3783 Limit processor to maximum C-state
3784 max_cstate=9 overrides any DMI blacklist limit.
3786 processor.nocst [HW,ACPI]
3787 Ignore the _CST method to determine C-states,
3788 instead using the legacy FADT method
3790 profile= [KNL] Enable kernel profiling via /proc/profile
3791 Format: [<profiletype>,]<number>
3792 Param: <profiletype>: "schedule", "sleep", or "kvm"
3793 [defaults to kernel profiling]
3794 Param: "schedule" - profile schedule points.
3795 Param: "sleep" - profile D-state sleeping (millisecs).
3796 Requires CONFIG_SCHEDSTATS
3797 Param: "kvm" - profile VM exits.
3798 Param: <number> - step/bucket size as a power of 2 for
3799 statistical time based profiling.
3801 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3803 See Documentation/admin-guide/blockdev/ramdisk.rst.
3805 psi= [KNL] Enable or disable pressure stall information
3809 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3810 probe for; one of (bare|imps|exps|lifebook|any).
3811 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3813 psmouse.resetafter= [HW,MOUSE]
3814 Try to reset the device after so many bad packets
3817 [HW,MOUSE] Set desired mouse resolution, in dpi.
3818 psmouse.smartscroll=
3819 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3820 0 = disabled, 1 = enabled (default).
3822 pstore.backend= Specify the name of the pstore backend to use
3825 See Documentation/admin-guide/blockdev/paride.rst.
3827 pti= [X86_64] Control Page Table Isolation of user and
3828 kernel address spaces. Disabling this feature
3829 removes hardening, but improves performance of
3830 system calls and interrupts.
3832 on - unconditionally enable
3833 off - unconditionally disable
3834 auto - kernel detects whether your CPU model is
3835 vulnerable to issues that PTI mitigates
3837 Not specifying this option is equivalent to pti=auto.
3840 Equivalent to pti=off
3843 [KNL] Number of legacy pty's. Overwrites compiled-in
3846 quiet [KNL] Disable most log messages
3851 See Documentation/admin-guide/md.rst.
3853 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3854 See Documentation/admin-guide/blockdev/ramdisk.rst.
3856 random.trust_cpu={on,off}
3857 [KNL] Enable or disable trusting the use of the
3858 CPU's random number generator (if available) to
3859 fully seed the kernel's CRNG. Default is controlled
3860 by CONFIG_RANDOM_TRUST_CPU.
3862 ras=option[,option,...] [KNL] RAS-specific options
3865 Disable the Correctable Errors Collector,
3866 see CONFIG_RAS_CEC help text.
3869 The argument is a cpu list, as described above,
3870 except that the string "all" can be used to
3871 specify every CPU on the system.
3873 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3874 the specified list of CPUs to be no-callback CPUs.
3875 Invocation of these CPUs' RCU callbacks will be
3876 offloaded to "rcuox/N" kthreads created for that
3877 purpose, where "x" is "p" for RCU-preempt, and
3878 "s" for RCU-sched, and "N" is the CPU number.
3879 This reduces OS jitter on the offloaded CPUs,
3880 which can be useful for HPC and real-time
3881 workloads. It can also improve energy efficiency
3882 for asymmetric multiprocessors.
3885 Rather than requiring that offloaded CPUs
3886 (specified by rcu_nocbs= above) explicitly
3887 awaken the corresponding "rcuoN" kthreads,
3888 make these kthreads poll for callbacks.
3889 This improves the real-time response for the
3890 offloaded CPUs by relieving them of the need to
3891 wake up the corresponding kthread, but degrades
3892 energy efficiency by requiring that the kthreads
3893 periodically wake up to do the polling.
3895 rcutree.blimit= [KNL]
3896 Set maximum number of finished RCU callbacks to
3897 process in one batch.
3899 rcutree.dump_tree= [KNL]
3900 Dump the structure of the rcu_node combining tree
3901 out at early boot. This is used for diagnostic
3902 purposes, to verify correct tree setup.
3904 rcutree.gp_cleanup_delay= [KNL]
3905 Set the number of jiffies to delay each step of
3906 RCU grace-period cleanup.
3908 rcutree.gp_init_delay= [KNL]
3909 Set the number of jiffies to delay each step of
3910 RCU grace-period initialization.
3912 rcutree.gp_preinit_delay= [KNL]
3913 Set the number of jiffies to delay each step of
3914 RCU grace-period pre-initialization, that is,
3915 the propagation of recent CPU-hotplug changes up
3916 the rcu_node combining tree.
3918 rcutree.use_softirq= [KNL]
3919 If set to zero, move all RCU_SOFTIRQ processing to
3920 per-CPU rcuc kthreads. Defaults to a non-zero
3921 value, meaning that RCU_SOFTIRQ is used by default.
3922 Specify rcutree.use_softirq=0 to use rcuc kthreads.
3924 rcutree.rcu_fanout_exact= [KNL]
3925 Disable autobalancing of the rcu_node combining
3926 tree. This is used by rcutorture, and might
3927 possibly be useful for architectures having high
3928 cache-to-cache transfer latencies.
3930 rcutree.rcu_fanout_leaf= [KNL]
3931 Change the number of CPUs assigned to each
3932 leaf rcu_node structure. Useful for very
3933 large systems, which will choose the value 64,
3934 and for NUMA systems with large remote-access
3935 latencies, which will choose a value aligned
3936 with the appropriate hardware boundaries.
3938 rcutree.jiffies_till_first_fqs= [KNL]
3939 Set delay from grace-period initialization to
3940 first attempt to force quiescent states.
3941 Units are jiffies, minimum value is zero,
3942 and maximum value is HZ.
3944 rcutree.jiffies_till_next_fqs= [KNL]
3945 Set delay between subsequent attempts to force
3946 quiescent states. Units are jiffies, minimum
3947 value is one, and maximum value is HZ.
3949 rcutree.jiffies_till_sched_qs= [KNL]
3950 Set required age in jiffies for a
3951 given grace period before RCU starts
3952 soliciting quiescent-state help from
3953 rcu_note_context_switch() and cond_resched().
3954 If not specified, the kernel will calculate
3955 a value based on the most recent settings
3956 of rcutree.jiffies_till_first_fqs
3957 and rcutree.jiffies_till_next_fqs.
3958 This calculated value may be viewed in
3959 rcutree.jiffies_to_sched_qs. Any attempt to set
3960 rcutree.jiffies_to_sched_qs will be cheerfully
3963 rcutree.kthread_prio= [KNL,BOOT]
3964 Set the SCHED_FIFO priority of the RCU per-CPU
3965 kthreads (rcuc/N). This value is also used for
3966 the priority of the RCU boost threads (rcub/N)
3967 and for the RCU grace-period kthreads (rcu_bh,
3968 rcu_preempt, and rcu_sched). If RCU_BOOST is
3969 set, valid values are 1-99 and the default is 1
3970 (the least-favored priority). Otherwise, when
3971 RCU_BOOST is not set, valid values are 0-99 and
3972 the default is zero (non-realtime operation).
3974 rcutree.rcu_nocb_gp_stride= [KNL]
3975 Set the number of NOCB callback kthreads in
3976 each group, which defaults to the square root
3977 of the number of CPUs. Larger numbers reduce
3978 the wakeup overhead on the global grace-period
3979 kthread, but increases that same overhead on
3980 each group's NOCB grace-period kthread.
3982 rcutree.qhimark= [KNL]
3983 Set threshold of queued RCU callbacks beyond which
3984 batch limiting is disabled.
3986 rcutree.qlowmark= [KNL]
3987 Set threshold of queued RCU callbacks below which
3988 batch limiting is re-enabled.
3990 rcutree.rcu_idle_gp_delay= [KNL]
3991 Set wakeup interval for idle CPUs that have
3992 RCU callbacks (RCU_FAST_NO_HZ=y).
3994 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3995 Set wakeup interval for idle CPUs that have
3996 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3997 Lazy RCU callbacks are those which RCU can
3998 prove do nothing more than free memory.
4000 rcutree.rcu_kick_kthreads= [KNL]
4001 Cause the grace-period kthread to get an extra
4002 wake_up() if it sleeps three times longer than
4003 it should at force-quiescent-state time.
4004 This wake_up() will be accompanied by a
4005 WARN_ONCE() splat and an ftrace_dump().
4007 rcutree.sysrq_rcu= [KNL]
4008 Commandeer a sysrq key to dump out Tree RCU's
4009 rcu_node tree with an eye towards determining
4010 why a new grace period has not yet started.
4012 rcuperf.gp_async= [KNL]
4013 Measure performance of asynchronous
4014 grace-period primitives such as call_rcu().
4016 rcuperf.gp_async_max= [KNL]
4017 Specify the maximum number of outstanding
4018 callbacks per writer thread. When a writer
4019 thread exceeds this limit, it invokes the
4020 corresponding flavor of rcu_barrier() to allow
4021 previously posted callbacks to drain.
4023 rcuperf.gp_exp= [KNL]
4024 Measure performance of expedited synchronous
4025 grace-period primitives.
4027 rcuperf.holdoff= [KNL]
4028 Set test-start holdoff period. The purpose of
4029 this parameter is to delay the start of the
4030 test until boot completes in order to avoid
4033 rcuperf.kfree_rcu_test= [KNL]
4034 Set to measure performance of kfree_rcu() flooding.
4036 rcuperf.kfree_nthreads= [KNL]
4037 The number of threads running loops of kfree_rcu().
4039 rcuperf.kfree_alloc_num= [KNL]
4040 Number of allocations and frees done in an iteration.
4042 rcuperf.kfree_loops= [KNL]
4043 Number of loops doing rcuperf.kfree_alloc_num number
4044 of allocations and frees.
4046 rcuperf.nreaders= [KNL]
4047 Set number of RCU readers. The value -1 selects
4048 N, where N is the number of CPUs. A value
4049 "n" less than -1 selects N-n+1, where N is again
4050 the number of CPUs. For example, -2 selects N
4051 (the number of CPUs), -3 selects N+1, and so on.
4052 A value of "n" less than or equal to -N selects
4055 rcuperf.nwriters= [KNL]
4056 Set number of RCU writers. The values operate
4057 the same as for rcuperf.nreaders.
4058 N, where N is the number of CPUs
4060 rcuperf.perf_type= [KNL]
4061 Specify the RCU implementation to test.
4063 rcuperf.shutdown= [KNL]
4064 Shut the system down after performance tests
4065 complete. This is useful for hands-off automated
4068 rcuperf.verbose= [KNL]
4069 Enable additional printk() statements.
4071 rcuperf.writer_holdoff= [KNL]
4072 Write-side holdoff between grace periods,
4073 in microseconds. The default of zero says
4076 rcutorture.fqs_duration= [KNL]
4077 Set duration of force_quiescent_state bursts
4080 rcutorture.fqs_holdoff= [KNL]
4081 Set holdoff time within force_quiescent_state bursts
4084 rcutorture.fqs_stutter= [KNL]
4085 Set wait time between force_quiescent_state bursts
4088 rcutorture.fwd_progress= [KNL]
4089 Enable RCU grace-period forward-progress testing
4090 for the types of RCU supporting this notion.
4092 rcutorture.fwd_progress_div= [KNL]
4093 Specify the fraction of a CPU-stall-warning
4094 period to do tight-loop forward-progress testing.
4096 rcutorture.fwd_progress_holdoff= [KNL]
4097 Number of seconds to wait between successive
4098 forward-progress tests.
4100 rcutorture.fwd_progress_need_resched= [KNL]
4101 Enclose cond_resched() calls within checks for
4102 need_resched() during tight-loop forward-progress
4105 rcutorture.gp_cond= [KNL]
4106 Use conditional/asynchronous update-side
4107 primitives, if available.
4109 rcutorture.gp_exp= [KNL]
4110 Use expedited update-side primitives, if available.
4112 rcutorture.gp_normal= [KNL]
4113 Use normal (non-expedited) asynchronous
4114 update-side primitives, if available.
4116 rcutorture.gp_sync= [KNL]
4117 Use normal (non-expedited) synchronous
4118 update-side primitives, if available. If all
4119 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4120 rcutorture.gp_normal=, and rcutorture.gp_sync=
4121 are zero, rcutorture acts as if is interpreted
4122 they are all non-zero.
4124 rcutorture.n_barrier_cbs= [KNL]
4125 Set callbacks/threads for rcu_barrier() testing.
4127 rcutorture.nfakewriters= [KNL]
4128 Set number of concurrent RCU writers. These just
4129 stress RCU, they don't participate in the actual
4130 test, hence the "fake".
4132 rcutorture.nreaders= [KNL]
4133 Set number of RCU readers. The value -1 selects
4134 N-1, where N is the number of CPUs. A value
4135 "n" less than -1 selects N-n-2, where N is again
4136 the number of CPUs. For example, -2 selects N
4137 (the number of CPUs), -3 selects N+1, and so on.
4139 rcutorture.object_debug= [KNL]
4140 Enable debug-object double-call_rcu() testing.
4142 rcutorture.onoff_holdoff= [KNL]
4143 Set time (s) after boot for CPU-hotplug testing.
4145 rcutorture.onoff_interval= [KNL]
4146 Set time (jiffies) between CPU-hotplug operations,
4147 or zero to disable CPU-hotplug testing.
4149 rcutorture.shuffle_interval= [KNL]
4150 Set task-shuffle interval (s). Shuffling tasks
4151 allows some CPUs to go into dyntick-idle mode
4152 during the rcutorture test.
4154 rcutorture.shutdown_secs= [KNL]
4155 Set time (s) after boot system shutdown. This
4156 is useful for hands-off automated testing.
4158 rcutorture.stall_cpu= [KNL]
4159 Duration of CPU stall (s) to test RCU CPU stall
4160 warnings, zero to disable.
4162 rcutorture.stall_cpu_holdoff= [KNL]
4163 Time to wait (s) after boot before inducing stall.
4165 rcutorture.stall_cpu_irqsoff= [KNL]
4166 Disable interrupts while stalling if set.
4168 rcutorture.stat_interval= [KNL]
4169 Time (s) between statistics printk()s.
4171 rcutorture.stutter= [KNL]
4172 Time (s) to stutter testing, for example, specifying
4173 five seconds causes the test to run for five seconds,
4174 wait for five seconds, and so on. This tests RCU's
4175 ability to transition abruptly to and from idle.
4177 rcutorture.test_boost= [KNL]
4178 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4179 "Maybe" means test if the RCU implementation
4180 under test support RCU priority boosting.
4182 rcutorture.test_boost_duration= [KNL]
4183 Duration (s) of each individual boost test.
4185 rcutorture.test_boost_interval= [KNL]
4186 Interval (s) between each boost test.
4188 rcutorture.test_no_idle_hz= [KNL]
4189 Test RCU's dyntick-idle handling. See also the
4190 rcutorture.shuffle_interval parameter.
4192 rcutorture.torture_type= [KNL]
4193 Specify the RCU implementation to test.
4195 rcutorture.verbose= [KNL]
4196 Enable additional printk() statements.
4198 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4199 Dump ftrace buffer after reporting RCU CPU
4202 rcupdate.rcu_cpu_stall_suppress= [KNL]
4203 Suppress RCU CPU stall warning messages.
4205 rcupdate.rcu_cpu_stall_timeout= [KNL]
4206 Set timeout for RCU CPU stall warning messages.
4208 rcupdate.rcu_expedited= [KNL]
4209 Use expedited grace-period primitives, for
4210 example, synchronize_rcu_expedited() instead
4211 of synchronize_rcu(). This reduces latency,
4212 but can increase CPU utilization, degrade
4213 real-time latency, and degrade energy efficiency.
4214 No effect on CONFIG_TINY_RCU kernels.
4216 rcupdate.rcu_normal= [KNL]
4217 Use only normal grace-period primitives,
4218 for example, synchronize_rcu() instead of
4219 synchronize_rcu_expedited(). This improves
4220 real-time latency, CPU utilization, and
4221 energy efficiency, but can expose users to
4222 increased grace-period latency. This parameter
4223 overrides rcupdate.rcu_expedited. No effect on
4224 CONFIG_TINY_RCU kernels.
4226 rcupdate.rcu_normal_after_boot= [KNL]
4227 Once boot has completed (that is, after
4228 rcu_end_inkernel_boot() has been invoked), use
4229 only normal grace-period primitives. No effect
4230 on CONFIG_TINY_RCU kernels.
4232 rcupdate.rcu_task_stall_timeout= [KNL]
4233 Set timeout in jiffies for RCU task stall warning
4234 messages. Disable with a value less than or equal
4237 rcupdate.rcu_self_test= [KNL]
4238 Run the RCU early boot self tests
4242 Run specified binary instead of /init from the ramdisk,
4243 used for early userspace startup. See initrd.
4246 force - Override the decision by the kernel to hide the
4247 advertisement of RDRAND support (this affects
4248 certain AMD processors because of buggy BIOS
4249 support, specifically around the suspend/resume
4253 Turn on/off individual RDT features. List is:
4254 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4256 E.g. to turn on cmt and turn off mba use:
4260 Format (x86 or x86_64):
4261 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4263 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4265 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4266 (prefix with 'panic_' to set mode for panic
4268 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4269 reboot_force is either force or not specified,
4270 reboot_cpu is s[mp]#### with #### being the processor
4271 to be used for rebooting.
4274 [KNL, SMP] Set scheduler's default relax_domain_level.
4275 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4277 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4278 Format: <base1>,<size1>[,<base2>,<size2>,...]
4279 Reserve I/O ports or memory so the kernel won't use
4280 them. If <base> is less than 0x10000, the region
4281 is assumed to be I/O ports; otherwise it is memory.
4283 reservetop= [X86-32]
4285 Reserves a hole at the top of the kernel virtual
4290 Set the amount of memory to reserve for BIOS at
4291 the bottom of the address space.
4293 reset_devices [KNL] Force drivers to reset the underlying device
4294 during initialization.
4297 Specify the partition device for software suspend
4299 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4301 resume_offset= [SWSUSP]
4302 Specify the offset from the beginning of the partition
4303 given by "resume=" at which the swap header is located,
4304 in <PAGE_SIZE> units (needed only for swap files).
4305 See Documentation/power/swsusp-and-swap-files.rst
4307 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4308 read the resume files
4310 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4311 Useful for devices that are detected asynchronously
4312 (e.g. USB and MMC devices).
4314 hibernate= [HIBERNATION]
4315 noresume Don't check if there's a hibernation image
4316 present during boot.
4317 nocompress Don't compress/decompress hibernation images.
4318 no Disable hibernation and resume.
4319 protect_image Turn on image protection during restoration
4320 (that will set all pages holding image data
4321 during restoration read-only).
4323 retain_initrd [RAM] Keep initrd memory after extraction
4325 rfkill.default_state=
4326 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4327 etc. communication is blocked by default.
4330 rfkill.master_switch_mode=
4331 0 The "airplane mode" button does nothing.
4332 1 The "airplane mode" button toggles between everything
4333 blocked and the previous configuration.
4334 2 The "airplane mode" button toggles between everything
4335 blocked and everything unblocked.
4337 rhash_entries= [KNL,NET]
4338 Set number of hash buckets for route cache
4341 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4344 ro [KNL] Mount root device read-only on boot
4347 on Mark read-only kernel memory as read-only (default).
4348 off Leave read-only kernel memory writable for debugging.
4351 Enable the uart passthrough on the designated usb port
4352 on Rockchip SoCs. When active, the signals of the
4353 debug-uart get routed to the D+ and D- pins of the usb
4354 port and the regular usb controller gets disabled.
4356 root= [KNL] Root filesystem
4357 See name_to_dev_t comment in init/do_mounts.c.
4359 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4360 mount the root filesystem
4362 rootflags= [KNL] Set root filesystem mount option string
4364 rootfstype= [KNL] Set root filesystem type
4366 rootwait [KNL] Wait (indefinitely) for root device to show up.
4367 Useful for devices that are detected asynchronously
4368 (e.g. USB and MMC devices).
4370 rproc_mem=nn[KMG][@address]
4371 [KNL,ARM,CMA] Remoteproc physical memory block.
4372 Memory area to be used by remote processor image,
4375 rw [KNL] Mount root device read-write on boot
4377 S [KNL] Run init in single mode
4379 s390_iommu= [HW,S390]
4380 Set s390 IOTLB flushing mode
4382 With strict flushing every unmap operation will result in
4383 an IOTLB flush. Default is lazy flushing before reuse,
4387 See drivers/net/irda/sa1100_ir.c.
4389 sbni= [NET] Granch SBNI12 leased line adapter
4391 sched_debug [KNL] Enables verbose scheduler debug messages.
4393 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4394 Allowed values are enable and disable. This feature
4395 incurs a small amount of overhead in the scheduler
4396 but is useful for debugging and performance tuning.
4398 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4399 xtime_lock contention on larger systems, and/or RCU lock
4400 contention on all systems with CONFIG_MAXSMP set.
4401 Format: { "0" | "1" }
4402 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4404 Note: increases power consumption, thus should only be
4405 enabled if running jitter sensitive (HPC/RT) workloads.
4407 security= [SECURITY] Choose a legacy "major" security module to
4408 enable at boot. This has been deprecated by the
4411 selinux= [SELINUX] Disable or enable SELinux at boot time.
4412 Format: { "0" | "1" }
4413 See security/selinux/Kconfig help text.
4418 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4419 Format: { "0" | "1" }
4420 See security/apparmor/Kconfig help text
4423 Default value is set via kernel config option.
4425 serialnumber [BUGS=X86-32]
4428 Maximal number of shapers.
4436 Disable merging of slabs with similar size. May be
4437 necessary if there is some reason to distinguish
4438 allocs to different slabs, especially in hardened
4439 environments where the risk of heap overflows and
4440 layout control by attackers can usually be
4441 frustrated by disabling merging. This will reduce
4442 most of the exposure of a heap attack to a single
4443 cache (risks via metadata attacks are mostly
4444 unchanged). Debug options disable merging on their
4446 For more information see Documentation/vm/slub.rst.
4448 slab_max_order= [MM, SLAB]
4449 Determines the maximum allowed order for slabs.
4450 A high setting may cause OOMs due to memory
4451 fragmentation. Defaults to 1 for systems with
4452 more than 32MB of RAM, 0 otherwise.
4454 slub_debug[=options[,slabs]] [MM, SLUB]
4455 Enabling slub_debug allows one to determine the
4456 culprit if slab objects become corrupted. Enabling
4457 slub_debug can create guard zones around objects and
4458 may poison objects when not in use. Also tracks the
4459 last alloc / free. For more information see
4460 Documentation/vm/slub.rst.
4462 slub_memcg_sysfs= [MM, SLUB]
4463 Determines whether to enable sysfs directories for
4464 memory cgroup sub-caches. 1 to enable, 0 to disable.
4465 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4466 Enabling this can lead to a very high number of debug
4467 directories and files being created under
4470 slub_max_order= [MM, SLUB]
4471 Determines the maximum allowed order for slabs.
4472 A high setting may cause OOMs due to memory
4473 fragmentation. For more information see
4474 Documentation/vm/slub.rst.
4476 slub_min_objects= [MM, SLUB]
4477 The minimum number of objects per slab. SLUB will
4478 increase the slab order up to slub_max_order to
4479 generate a sufficiently large slab able to contain
4480 the number of objects indicated. The higher the number
4481 of objects the smaller the overhead of tracking slabs
4482 and the less frequently locks need to be acquired.
4483 For more information see Documentation/vm/slub.rst.
4485 slub_min_order= [MM, SLUB]
4486 Determines the minimum page order for slabs. Must be
4487 lower than slub_max_order.
4488 For more information see Documentation/vm/slub.rst.
4490 slub_nomerge [MM, SLUB]
4491 Same with slab_nomerge. This is supported for legacy.
4492 See slab_nomerge for more information.
4495 Format: <io1>[,<io2>[,...,<io8>]]
4497 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4498 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4499 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4500 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4501 smsc-ircc2.ircc_irq= [HW] IRQ line
4502 smsc-ircc2.ircc_dma= [HW] DMA channel
4503 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4504 0: Toshiba Satellite 1800 (GP data pin select)
4505 1: Fast pin select (default)
4508 smt [KNL,S390] Set the maximum number of threads (logical
4509 CPUs) to use per physical CPU on systems capable of
4510 symmetric multithreading (SMT). Will be capped to the
4511 actual hardware limit.
4513 Default: -1 (no limit)
4516 [KNL] Should the soft-lockup detector generate panics.
4519 A nonzero value instructs the soft-lockup detector
4520 to panic the machine when a soft-lockup occurs. This
4521 is also controlled by CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC
4522 which is the respective build-time switch to that
4525 softlockup_all_cpu_backtrace=
4526 [KNL] Should the soft-lockup detector generate
4527 backtraces on all cpus.
4530 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4531 See Documentation/admin-guide/laptops/sonypi.rst
4533 spectre_v2= [X86] Control mitigation of Spectre variant 2
4534 (indirect branch speculation) vulnerability.
4535 The default operation protects the kernel from
4538 on - unconditionally enable, implies
4540 off - unconditionally disable, implies
4542 auto - kernel detects whether your CPU model is
4545 Selecting 'on' will, and 'auto' may, choose a
4546 mitigation method at run time according to the
4547 CPU, the available microcode, the setting of the
4548 CONFIG_RETPOLINE configuration option, and the
4549 compiler with which the kernel was built.
4551 Selecting 'on' will also enable the mitigation
4552 against user space to user space task attacks.
4554 Selecting 'off' will disable both the kernel and
4555 the user space protections.
4557 Specific mitigations can also be selected manually:
4559 retpoline - replace indirect branches
4560 retpoline,generic - google's original retpoline
4561 retpoline,amd - AMD-specific minimal thunk
4563 Not specifying this option is equivalent to
4567 [X86] Control mitigation of Spectre variant 2
4568 (indirect branch speculation) vulnerability between
4571 on - Unconditionally enable mitigations. Is
4572 enforced by spectre_v2=on
4574 off - Unconditionally disable mitigations. Is
4575 enforced by spectre_v2=off
4577 prctl - Indirect branch speculation is enabled,
4578 but mitigation can be enabled via prctl
4579 per thread. The mitigation control state
4580 is inherited on fork.
4583 - Like "prctl" above, but only STIBP is
4584 controlled per thread. IBPB is issued
4585 always when switching between different user
4589 - Same as "prctl" above, but all seccomp
4590 threads will enable the mitigation unless
4591 they explicitly opt out.
4594 - Like "seccomp" above, but only STIBP is
4595 controlled per thread. IBPB is issued
4596 always when switching between different
4597 user space processes.
4599 auto - Kernel selects the mitigation depending on
4600 the available CPU features and vulnerability.
4603 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4605 Not specifying this option is equivalent to
4606 spectre_v2_user=auto.
4608 spec_store_bypass_disable=
4609 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4610 (Speculative Store Bypass vulnerability)
4612 Certain CPUs are vulnerable to an exploit against a
4613 a common industry wide performance optimization known
4614 as "Speculative Store Bypass" in which recent stores
4615 to the same memory location may not be observed by
4616 later loads during speculative execution. The idea
4617 is that such stores are unlikely and that they can
4618 be detected prior to instruction retirement at the
4619 end of a particular speculation execution window.
4621 In vulnerable processors, the speculatively forwarded
4622 store can be used in a cache side channel attack, for
4623 example to read memory to which the attacker does not
4624 directly have access (e.g. inside sandboxed code).
4626 This parameter controls whether the Speculative Store
4627 Bypass optimization is used.
4629 On x86 the options are:
4631 on - Unconditionally disable Speculative Store Bypass
4632 off - Unconditionally enable Speculative Store Bypass
4633 auto - Kernel detects whether the CPU model contains an
4634 implementation of Speculative Store Bypass and
4635 picks the most appropriate mitigation. If the
4636 CPU is not vulnerable, "off" is selected. If the
4637 CPU is vulnerable the default mitigation is
4638 architecture and Kconfig dependent. See below.
4639 prctl - Control Speculative Store Bypass per thread
4640 via prctl. Speculative Store Bypass is enabled
4641 for a process by default. The state of the control
4642 is inherited on fork.
4643 seccomp - Same as "prctl" above, but all seccomp threads
4644 will disable SSB unless they explicitly opt out.
4646 Default mitigations:
4647 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4649 On powerpc the options are:
4651 on,auto - On Power8 and Power9 insert a store-forwarding
4652 barrier on kernel entry and exit. On Power7
4653 perform a software flush on kernel entry and
4657 Not specifying this option is equivalent to
4658 spec_store_bypass_disable=auto.
4660 spia_io_base= [HW,MTD]
4665 srcutree.counter_wrap_check [KNL]
4666 Specifies how frequently to check for
4667 grace-period sequence counter wrap for the
4668 srcu_data structure's ->srcu_gp_seq_needed field.
4669 The greater the number of bits set in this kernel
4670 parameter, the less frequently counter wrap will
4671 be checked for. Note that the bottom two bits
4674 srcutree.exp_holdoff [KNL]
4675 Specifies how many nanoseconds must elapse
4676 since the end of the last SRCU grace period for
4677 a given srcu_struct until the next normal SRCU
4678 grace period will be considered for automatic
4679 expediting. Set to zero to disable automatic
4683 Speculative Store Bypass Disable control
4685 On CPUs that are vulnerable to the Speculative
4686 Store Bypass vulnerability and offer a
4687 firmware based mitigation, this parameter
4688 indicates how the mitigation should be used:
4690 force-on: Unconditionally enable mitigation for
4691 for both kernel and userspace
4692 force-off: Unconditionally disable mitigation for
4693 for both kernel and userspace
4694 kernel: Always enable mitigation in the
4695 kernel, and offer a prctl interface
4696 to allow userspace to register its
4697 interest in being mitigated too.
4699 stack_guard_gap= [MM]
4700 override the default stack gap protection. The value
4701 is in page units and it defines how many pages prior
4702 to (for stacks growing down) resp. after (for stacks
4703 growing up) the main stack are reserved for no other
4704 mapping. Default value is 256 pages.
4707 Enabled the stack tracer on boot up.
4709 stacktrace_filter=[function-list]
4710 [FTRACE] Limit the functions that the stack tracer
4711 will trace at boot up. function-list is a comma separated
4712 list of functions. This list can be changed at run
4713 time by the stack_trace_filter file in the debugfs
4714 tracing directory. Note, this enables stack tracing
4715 and the stacktrace above is not needed.
4719 Set the STI (builtin display/keyboard on the HP-PARISC
4720 machines) console (graphic card) which should be used
4721 as the initial boot-console.
4722 See also comment in drivers/video/console/sticore.c.
4725 See comment in drivers/video/console/sticore.c.
4728 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4730 sunrpc.min_resvport=
4731 sunrpc.max_resvport=
4733 SunRPC servers often require that client requests
4734 originate from a privileged port (i.e. a port in the
4735 range 0 < portnr < 1024).
4736 An administrator who wishes to reserve some of these
4737 ports for other uses may adjust the range that the
4738 kernel's sunrpc client considers to be privileged
4739 using these two parameters to set the minimum and
4740 maximum port values.
4742 sunrpc.svc_rpc_per_connection_limit=
4744 Limit the number of requests that the server will
4745 process in parallel from a single connection.
4746 The default value is 0 (no limit).
4750 Control how the NFS server code allocates CPUs to
4751 service thread pools. Depending on how many NICs
4752 you have and where their interrupts are bound, this
4753 option will affect which CPUs will do NFS serving.
4754 Note: this parameter cannot be changed while the
4755 NFS server is running.
4757 auto the server chooses an appropriate mode
4758 automatically using heuristics
4759 global a single global pool contains all CPUs
4760 percpu one pool for each CPU
4761 pernode one pool for each NUMA node (equivalent
4762 to global on non-NUMA machines)
4764 sunrpc.tcp_slot_table_entries=
4765 sunrpc.udp_slot_table_entries=
4767 Sets the upper limit on the number of simultaneous
4768 RPC calls that can be sent from the client to a
4769 server. Increasing these values may allow you to
4770 improve throughput, but will also increase the
4771 amount of memory reserved for use by the client.
4773 suspend.pm_test_delay=
4775 Sets the number of seconds to remain in a suspend test
4776 mode before resuming the system (see
4777 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4778 is set. Default value is 5.
4781 Format: { on | off | y | n | 1 | 0 }
4782 This parameter controls use of the Protected
4783 Execution Facility on pSeries.
4786 [KNL] Enable accounting of swap in memory resource
4787 controller if no parameter or 1 is given or disable
4788 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
4790 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4791 Format: { <int> | force | noforce }
4792 <int> -- Number of I/O TLB slabs
4793 force -- force using of bounce buffers even if they
4794 wouldn't be automatically used by the kernel
4795 noforce -- Never use bounce buffers (for debugging)
4799 sysfs.deprecated=0|1 [KNL]
4800 Enable/disable old style sysfs layout for old udev
4801 on older distributions. When this option is enabled
4802 very new udev will not work anymore. When this option
4803 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4804 in older udev will not work anymore.
4805 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4806 the kernel configuration.
4808 sysrq_always_enabled
4810 Ignore sysrq setting - this boot parameter will
4811 neutralize any effect of /proc/sys/kernel/sysrq.
4812 Useful for debugging.
4814 tcpmhash_entries= [KNL,NET]
4815 Set the number of tcp_metrics_hash slots.
4816 Default value is 8192 or 16384 depending on total
4817 ram pages. This is used to specify the TCP metrics
4818 cache size. See Documentation/networking/ip-sysctl.txt
4819 "tcp_no_metrics_save" section for more details.
4823 test_suspend= [SUSPEND][,N]
4824 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4825 standby suspend) or "freeze" (for suspend type freeze)
4826 as the system sleep state during system startup with
4827 the optional capability to repeat N number of times.
4828 The system is woken from this state using a
4829 wakeup-capable RTC alarm.
4831 thash_entries= [KNL,NET]
4832 Set number of hash buckets for TCP connection
4834 thermal.act= [HW,ACPI]
4835 -1: disable all active trip points in all thermal zones
4836 <degrees C>: override all lowest active trip points
4838 thermal.crt= [HW,ACPI]
4839 -1: disable all critical trip points in all thermal zones
4840 <degrees C>: override all critical trip points
4842 thermal.nocrt= [HW,ACPI]
4843 Set to disable actions on ACPI thermal zone
4844 critical and hot trip points.
4846 thermal.off= [HW,ACPI]
4847 1: disable ACPI thermal control
4849 thermal.psv= [HW,ACPI]
4850 -1: disable all passive trip points
4851 <degrees C>: override all passive trip points to this
4854 thermal.tzp= [HW,ACPI]
4855 Specify global default ACPI thermal zone polling rate
4856 <deci-seconds>: poll all this frequency
4857 0: no polling (default)
4860 Force threading of all interrupt handlers except those
4861 marked explicitly IRQF_NO_THREAD.
4865 Specify if the kernel should make use of the cpu
4866 topology information if the hardware supports this.
4867 The scheduler will make use of this information and
4868 e.g. base its process migration decisions on it.
4871 topology_updates= [KNL, PPC, NUMA]
4873 Specify if the kernel should ignore (off)
4874 topology updates sent by the hypervisor to this
4879 tpm_suspend_pcr=[HW,TPM]
4880 Format: integer pcr id
4881 Specify that at suspend time, the tpm driver
4882 should extend the specified pcr with zeros,
4883 as a workaround for some chips which fail to
4884 flush the last written pcr on TPM_SaveState.
4885 This will guarantee that all the other pcrs
4888 trace_buf_size=nn[KMG]
4889 [FTRACE] will set tracing buffer size on each cpu.
4891 trace_event=[event-list]
4892 [FTRACE] Set and start specified trace events in order
4893 to facilitate early boot debugging. The event-list is a
4894 comma separated list of trace events to enable. See
4895 also Documentation/trace/events.rst
4897 trace_options=[option-list]
4898 [FTRACE] Enable or disable tracer options at boot.
4899 The option-list is a comma delimited list of options
4900 that can be enabled or disabled just as if you were
4901 to echo the option name into
4903 /sys/kernel/debug/tracing/trace_options
4905 For example, to enable stacktrace option (to dump the
4906 stack trace of each event), add to the command line:
4908 trace_options=stacktrace
4910 See also Documentation/trace/ftrace.rst "trace options"
4914 Have the tracepoints sent to printk as well as the
4915 tracing ring buffer. This is useful for early boot up
4916 where the system hangs or reboots and does not give the
4917 option for reading the tracing buffer or performing a
4918 ftrace_dump_on_oops.
4920 To turn off having tracepoints sent to printk,
4921 echo 0 > /proc/sys/kernel/tracepoint_printk
4922 Note, echoing 1 into this file without the
4923 tracepoint_printk kernel cmdline option has no effect.
4927 Having tracepoints sent to printk() and activating high
4928 frequency tracepoints such as irq or sched, can cause
4929 the system to live lock.
4932 [FTRACE] enable this option to disable tracing when a
4933 warning is hit. This turns off "tracing_on". Tracing can
4934 be enabled again by echoing '1' into the "tracing_on"
4935 file located in /sys/kernel/debug/tracing/
4937 This option is useful, as it disables the trace before
4938 the WARNING dump is called, which prevents the trace to
4939 be filled with content caused by the warning output.
4941 This option can also be set at run time via the sysctl
4942 option: kernel/traceoff_on_warning
4944 transparent_hugepage=
4946 Format: [always|madvise|never]
4947 Can be used to control the default behavior of the system
4948 with respect to transparent hugepages.
4949 See Documentation/admin-guide/mm/transhuge.rst
4952 tsc= Disable clocksource stability checks for TSC.
4954 [x86] reliable: mark tsc clocksource as reliable, this
4955 disables clocksource verification at runtime, as well
4956 as the stability checks done at bootup. Used to enable
4957 high-resolution timer mode on older hardware, and in
4958 virtualized environment.
4959 [x86] noirqtime: Do not use TSC to do irq accounting.
4960 Used to run time disable IRQ_TIME_ACCOUNTING on any
4961 platforms where RDTSC is slow and this accounting
4963 [x86] unstable: mark the TSC clocksource as unstable, this
4964 marks the TSC unconditionally unstable at bootup and
4965 avoids any further wobbles once the TSC watchdog notices.
4966 [x86] nowatchdog: disable clocksource watchdog. Used
4967 in situations with strict latency requirements (where
4968 interruptions from clocksource watchdog are not
4971 tsx= [X86] Control Transactional Synchronization
4972 Extensions (TSX) feature in Intel processors that
4973 support TSX control.
4975 This parameter controls the TSX feature. The options are:
4977 on - Enable TSX on the system. Although there are
4978 mitigations for all known security vulnerabilities,
4979 TSX has been known to be an accelerator for
4980 several previous speculation-related CVEs, and
4981 so there may be unknown security risks associated
4982 with leaving it enabled.
4984 off - Disable TSX on the system. (Note that this
4985 option takes effect only on newer CPUs which are
4986 not vulnerable to MDS, i.e., have
4987 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
4988 the new IA32_TSX_CTRL MSR through a microcode
4989 update. This new MSR allows for the reliable
4990 deactivation of the TSX functionality.)
4992 auto - Disable TSX if X86_BUG_TAA is present,
4993 otherwise enable TSX on the system.
4995 Not specifying this option is equivalent to tsx=off.
4997 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5000 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5001 Abort (TAA) vulnerability.
5003 Similar to Micro-architectural Data Sampling (MDS)
5004 certain CPUs that support Transactional
5005 Synchronization Extensions (TSX) are vulnerable to an
5006 exploit against CPU internal buffers which can forward
5007 information to a disclosure gadget under certain
5010 In vulnerable processors, the speculatively forwarded
5011 data can be used in a cache side channel attack, to
5012 access data to which the attacker does not have direct
5015 This parameter controls the TAA mitigation. The
5018 full - Enable TAA mitigation on vulnerable CPUs
5021 full,nosmt - Enable TAA mitigation and disable SMT on
5022 vulnerable CPUs. If TSX is disabled, SMT
5023 is not disabled because CPU is not
5024 vulnerable to cross-thread TAA attacks.
5025 off - Unconditionally disable TAA mitigation
5027 On MDS-affected machines, tsx_async_abort=off can be
5028 prevented by an active MDS mitigation as both vulnerabilities
5029 are mitigated with the same mechanism so in order to disable
5030 this mitigation, you need to specify mds=off too.
5032 Not specifying this option is equivalent to
5033 tsx_async_abort=full. On CPUs which are MDS affected
5034 and deploy MDS mitigation, TAA mitigation is not
5035 required and doesn't provide any additional
5039 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5041 turbografx.map[2|3]= [HW,JOY]
5042 TurboGraFX parallel port interface
5044 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5045 See also Documentation/input/devices/joystick-parport.rst
5047 udbg-immortal [PPC] When debugging early kernel crashes that
5048 happen after console_init() and before a proper
5049 console driver takes over, this boot options might
5050 help "seeing" what's going on.
5052 uhash_entries= [KNL,NET]
5053 Set number of hash buckets for UDP/UDP-Lite connections
5056 [USB] Ignore overcurrent events (default N).
5057 Some badly-designed motherboards generate lots of
5058 bogus events, for ports that aren't wired to
5059 anything. Set this parameter to avoid log spamming.
5060 Note that genuine overcurrent events won't be
5064 [X86] Cause panic on unknown NMI.
5066 usbcore.authorized_default=
5067 [USB] Default USB device authorization:
5068 (default -1 = authorized except for wireless USB,
5069 0 = not authorized, 1 = authorized, 2 = authorized
5070 if device connected to internal port)
5072 usbcore.autosuspend=
5073 [USB] The autosuspend time delay (in seconds) used
5074 for newly-detected USB devices (default 2). This
5075 is the time required before an idle device will be
5076 autosuspended. Devices for which the delay is set
5077 to a negative value won't be autosuspended at all.
5079 usbcore.usbfs_snoop=
5080 [USB] Set to log all usbfs traffic (default 0 = off).
5082 usbcore.usbfs_snoop_max=
5083 [USB] Maximum number of bytes to snoop in each URB
5086 usbcore.blinkenlights=
5087 [USB] Set to cycle leds on hubs (default 0 = off).
5089 usbcore.old_scheme_first=
5090 [USB] Start with the old device initialization
5091 scheme, applies only to low and full-speed devices
5094 usbcore.usbfs_memory_mb=
5095 [USB] Memory limit (in MB) for buffers allocated by
5096 usbfs (default = 16, 0 = max = 2047).
5098 usbcore.use_both_schemes=
5099 [USB] Try the other device initialization scheme
5100 if the first one fails (default 1 = enabled).
5102 usbcore.initial_descriptor_timeout=
5103 [USB] Specifies timeout for the initial 64-byte
5104 USB_REQ_GET_DESCRIPTOR request in milliseconds
5105 (default 5000 = 5.0 seconds).
5107 usbcore.nousb [USB] Disable the USB subsystem
5110 [USB] A list of quirk entries to augment the built-in
5111 usb core quirk list. List entries are separated by
5112 commas. Each entry has the form
5113 VendorID:ProductID:Flags. The IDs are 4-digit hex
5114 numbers and Flags is a set of letters. Each letter
5115 will change the built-in quirk; setting it if it is
5116 clear and clearing it if it is set. The letters have
5117 the following meanings:
5118 a = USB_QUIRK_STRING_FETCH_255 (string
5119 descriptors must not be fetched using
5121 b = USB_QUIRK_RESET_RESUME (device can't resume
5122 correctly so reset it instead);
5123 c = USB_QUIRK_NO_SET_INTF (device can't handle
5124 Set-Interface requests);
5125 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5126 handle its Configuration or Interface
5128 e = USB_QUIRK_RESET (device can't be reset
5129 (e.g morph devices), don't use reset);
5130 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5131 more interface descriptions than the
5132 bNumInterfaces count, and can't handle
5133 talking to these interfaces);
5134 g = USB_QUIRK_DELAY_INIT (device needs a pause
5135 during initialization, after we read
5136 the device descriptor);
5137 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5138 high speed and super speed interrupt
5139 endpoints, the USB 2.0 and USB 3.0 spec
5140 require the interval in microframes (1
5141 microframe = 125 microseconds) to be
5142 calculated as interval = 2 ^
5144 Devices with this quirk report their
5145 bInterval as the result of this
5146 calculation instead of the exponent
5147 variable used in the calculation);
5148 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5149 handle device_qualifier descriptor
5151 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5152 generates spurious wakeup, ignore
5153 remote wakeup capability);
5154 k = USB_QUIRK_NO_LPM (device can't handle Link
5156 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5157 (Device reports its bInterval as linear
5158 frames instead of the USB 2.0
5160 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5161 to be disconnected before suspend to
5162 prevent spurious wakeup);
5163 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5164 pause after every control message);
5165 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5166 delay after resetting its port);
5167 Example: quirks=0781:5580:bk,0a5c:5834:gij
5170 [USBHID] The interval which mice are to be polled at.
5173 [USBHID] The interval which joysticks are to be polled at.
5176 [USBHID] The interval which keyboards are to be polled at.
5178 usb-storage.delay_use=
5179 [UMS] The delay in seconds before a new device is
5180 scanned for Logical Units (default 1).
5183 [UMS] A list of quirks entries to supplement or
5184 override the built-in unusual_devs list. List
5185 entries are separated by commas. Each entry has
5186 the form VID:PID:Flags where VID and PID are Vendor
5187 and Product ID values (4-digit hex numbers) and
5188 Flags is a set of characters, each corresponding
5189 to a common usb-storage quirk flag as follows:
5190 a = SANE_SENSE (collect more than 18 bytes
5191 of sense data, not on uas);
5192 b = BAD_SENSE (don't collect more than 18
5193 bytes of sense data, not on uas);
5194 c = FIX_CAPACITY (decrease the reported
5195 device capacity by one sector);
5196 d = NO_READ_DISC_INFO (don't use
5197 READ_DISC_INFO command, not on uas);
5198 e = NO_READ_CAPACITY_16 (don't use
5199 READ_CAPACITY_16 command);
5200 f = NO_REPORT_OPCODES (don't use report opcodes
5202 g = MAX_SECTORS_240 (don't transfer more than
5203 240 sectors at a time, uas only);
5204 h = CAPACITY_HEURISTICS (decrease the
5205 reported device capacity by one
5206 sector if the number is odd);
5207 i = IGNORE_DEVICE (don't bind to this
5209 j = NO_REPORT_LUNS (don't use report luns
5211 l = NOT_LOCKABLE (don't try to lock and
5212 unlock ejectable media, not on uas);
5213 m = MAX_SECTORS_64 (don't transfer more
5214 than 64 sectors = 32 KB at a time,
5216 n = INITIAL_READ10 (force a retry of the
5217 initial READ(10) command, not on uas);
5218 o = CAPACITY_OK (accept the capacity
5219 reported by the device, not on uas);
5220 p = WRITE_CACHE (the device cache is ON
5221 by default, not on uas);
5222 r = IGNORE_RESIDUE (the device reports
5223 bogus residue values, not on uas);
5224 s = SINGLE_LUN (the device has only one
5226 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5227 commands, uas only);
5228 u = IGNORE_UAS (don't bind to the uas driver);
5229 w = NO_WP_DETECT (don't test whether the
5230 medium is write-protected).
5231 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5232 even if the device claims no cache,
5234 Example: quirks=0419:aaf5:rl,0421:0433:rc
5236 user_debug= [KNL,ARM]
5238 See arch/arm/Kconfig.debug help text.
5239 1 - undefined instruction events
5241 4 - invalid data aborts
5244 Example: user_debug=31
5247 [X86] Flags controlling user PTE allocations.
5249 nohigh = do not allocate PTE pages in
5250 HIGHMEM regardless of setting
5254 On X86_32, this is an alias for vdso32=. Otherwise:
5256 vdso=1: enable VDSO (the default)
5257 vdso=0: disable VDSO mapping
5259 vdso32= [X86] Control the 32-bit vDSO
5260 vdso32=1: enable 32-bit VDSO
5261 vdso32=0 or vdso32=2: disable 32-bit VDSO
5263 See the help text for CONFIG_COMPAT_VDSO for more
5264 details. If CONFIG_COMPAT_VDSO is set, the default is
5265 vdso32=0; otherwise, the default is vdso32=1.
5267 For compatibility with older kernels, vdso32=2 is an
5270 Try vdso32=0 if you encounter an error that says:
5271 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5274 vector=percpu: enable percpu vector domain
5276 video= [FB] Frame buffer configuration
5277 See Documentation/fb/modedb.rst.
5279 video.brightness_switch_enabled= [0,1]
5280 If set to 1, on receiving an ACPI notify event
5281 generated by hotkey, video driver will adjust brightness
5282 level and then send out the event to user space through
5283 the allocated input device; If set to 0, video driver
5284 will only send out the event without touching backlight
5289 [VMMIO] Memory mapped virtio (platform) device.
5291 <size>@<baseaddr>:<irq>[:<id>]
5293 <size> := size (can use standard suffixes
5295 <baseaddr> := physical base address
5296 <irq> := interrupt number (as passed to
5298 <id> := (optional) platform device id
5300 virtio_mmio.device=1K@0x100b0000:48:7
5302 Can be used multiple times for multiple devices.
5304 vga= [BOOT,X86-32] Select a particular video mode
5305 See Documentation/x86/boot.rst and
5306 Documentation/admin-guide/svga.rst.
5307 Use vga=ask for menu.
5308 This is actually a boot loader parameter; the value is
5309 passed to the kernel using a special protocol.
5311 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
5312 May slow down system boot speed, especially when
5313 enabled on systems with a large amount of memory.
5314 All options are enabled by default, and this
5315 interface is meant to allow for selectively
5316 enabling or disabling specific virtual memory
5319 Available options are:
5320 P Enable page structure init time poisoning
5321 - Disable all of the above options
5323 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5324 size of <nn>. This can be used to increase the
5325 minimum size (128MB on x86). It can also be used to
5326 decrease the size and leave more room for directly
5329 vmcp_cma=nn[MG] [KNL,S390]
5330 Sets the memory size reserved for contiguous memory
5331 allocations for the vmcp device driver.
5333 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5336 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5339 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5343 Controls the behavior of vsyscalls (i.e. calls to
5344 fixed addresses of 0xffffffffff600x00 from legacy
5345 code). Most statically-linked binaries and older
5346 versions of glibc use these calls. Because these
5347 functions are at fixed addresses, they make nice
5348 targets for exploits that can control RIP.
5350 emulate [default] Vsyscalls turn into traps and are
5351 emulated reasonably safely. The vsyscall
5354 xonly Vsyscalls turn into traps and are
5355 emulated reasonably safely. The vsyscall
5356 page is not readable.
5358 none Vsyscalls don't work at all. This makes
5359 them quite hard to use for exploits but
5360 might break your system.
5362 vt.color= [VT] Default text color.
5363 Format: 0xYX, X = foreground, Y = background.
5364 Default: 0x07 = light gray on black.
5366 vt.cur_default= [VT] Default cursor shape.
5367 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5368 the parameters of the <Esc>[?A;B;Cc escape sequence;
5369 see VGA-softcursor.txt. Default: 2 = underline.
5371 vt.default_blu= [VT]
5372 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5373 Change the default blue palette of the console.
5374 This is a 16-member array composed of values
5377 vt.default_grn= [VT]
5378 Format: <green0>,<green1>,<green2>,...,<green15>
5379 Change the default green palette of the console.
5380 This is a 16-member array composed of values
5383 vt.default_red= [VT]
5384 Format: <red0>,<red1>,<red2>,...,<red15>
5385 Change the default red palette of the console.
5386 This is a 16-member array composed of values
5392 Set system-wide default UTF-8 mode for all tty's.
5393 Default is 1, i.e. UTF-8 mode is enabled for all
5394 newly opened terminals.
5396 vt.global_cursor_default=
5399 Set system-wide default for whether a cursor
5400 is shown on new VTs. Default is -1,
5401 i.e. cursors will be created by default unless
5402 overridden by individual drivers. 0 will hide
5403 cursors, 1 will display them.
5405 vt.italic= [VT] Default color for italic text; 0-15.
5408 vt.underline= [VT] Default color for underlined text; 0-15.
5411 watchdog timers [HW,WDT] For information on watchdog timers,
5412 see Documentation/watchdog/watchdog-parameters.rst
5413 or other driver-specific files in the
5414 Documentation/watchdog/ directory.
5418 Set the hard lockup detector stall duration
5419 threshold in seconds. The soft lockup detector
5420 threshold is set to twice the value. A value of 0
5421 disables both lockup detectors. Default is 10
5424 workqueue.watchdog_thresh=
5425 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5426 warn stall conditions and dump internal state to
5427 help debugging. 0 disables workqueue stall
5428 detection; otherwise, it's the stall threshold
5429 duration in seconds. The default value is 30 and
5430 it can be updated at runtime by writing to the
5431 corresponding sysfs file.
5433 workqueue.disable_numa
5434 By default, all work items queued to unbound
5435 workqueues are affine to the NUMA nodes they're
5436 issued on, which results in better behavior in
5437 general. If NUMA affinity needs to be disabled for
5438 whatever reason, this option can be used. Note
5439 that this also can be controlled per-workqueue for
5440 workqueues visible under /sys/bus/workqueue/.
5442 workqueue.power_efficient
5443 Per-cpu workqueues are generally preferred because
5444 they show better performance thanks to cache
5445 locality; unfortunately, per-cpu workqueues tend to
5446 be more power hungry than unbound workqueues.
5448 Enabling this makes the per-cpu workqueues which
5449 were observed to contribute significantly to power
5450 consumption unbound, leading to measurably lower
5451 power usage at the cost of small performance
5454 The default value of this parameter is determined by
5455 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5457 workqueue.debug_force_rr_cpu
5458 Workqueue used to implicitly guarantee that work
5459 items queued without explicit CPU specified are put
5460 on the local CPU. This guarantee is no longer true
5461 and while local CPU is still preferred work items
5462 may be put on foreign CPUs. This debug option
5463 forces round-robin CPU selection to flush out
5464 usages which depend on the now broken guarantee.
5465 When enabled, memory and cache locality will be
5468 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5469 default x2apic cluster mode on platforms
5472 x86_intel_mid_timer= [X86-32,APBT]
5473 Choose timer option for x86 Intel MID platform.
5474 Two valid options are apbt timer only and lapic timer
5475 plus one apbt timer for broadcast timer.
5476 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5478 xen_512gb_limit [KNL,X86-64,XEN]
5479 Restricts the kernel running paravirtualized under Xen
5480 to use only up to 512 GB of RAM. The reason to do so is
5481 crash analysis tools and Xen tools for doing domain
5482 save/restore/migration must be enabled to handle larger
5485 xen_emul_unplug= [HW,X86,XEN]
5486 Unplug Xen emulated devices
5487 Format: [unplug0,][unplug1]
5488 ide-disks -- unplug primary master IDE devices
5489 aux-ide-disks -- unplug non-primary-master IDE devices
5490 nics -- unplug network devices
5491 all -- unplug all emulated devices (NICs and IDE disks)
5492 unnecessary -- unplugging emulated devices is
5493 unnecessary even if the host did not respond to
5495 never -- do not unplug even if version check succeeds
5497 xen_legacy_crash [X86,XEN]
5498 Crash from Xen panic notifier, without executing late
5499 panic() code such as dumping handler.
5501 xen_nopvspin [X86,XEN]
5502 Disables the ticketlock slowpath using Xen PV
5506 Disables the PV optimizations forcing the HVM guest to
5507 run as generic HVM guest with no PV drivers.
5508 This option is obsoleted by the "nopv" option, which
5509 has equivalent effect for XEN platform.
5511 xen_scrub_pages= [XEN]
5512 Boolean option to control scrubbing pages before giving them back
5513 to Xen, for use by other domains. Can be also changed at runtime
5514 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5515 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5517 xen_timer_slop= [X86-64,XEN]
5518 Set the timer slop (in nanoseconds) for the virtual Xen
5519 timers (default is 100000). This adjusts the minimum
5520 delta of virtualized Xen timers, where lower values
5521 improve timer resolution at the expense of processing
5522 more timer interrupts.
5524 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
5525 Disables the PV optimizations forcing the guest to run
5526 as generic guest with no PV drivers. Currently support
5527 XEN HVM, KVM, HYPER_V and VMWARE guest.
5529 xirc2ps_cs= [NET,PCMCIA]
5531 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5534 By default on POWER9 and above, the kernel will
5535 natively use the XIVE interrupt controller. This option
5536 allows the fallback firmware mode to be used:
5538 off Fallback to firmware control of XIVE interrupt
5539 controller on both pseries and powernv
5540 platforms. Only useful on POWER9 and above.
5542 xhci-hcd.quirks [USB,KNL]
5543 A hex value specifying bitmask with supplemental xhci
5544 host controller quirks. Meaning of each bit can be
5545 consulted in header drivers/usb/host/xhci.h.
5548 Format: { early | on | rw | ro | off }
5549 Controls if xmon debugger is enabled. Default is off.
5550 Passing only "xmon" is equivalent to "xmon=early".
5551 early Call xmon as early as possible on boot; xmon
5552 debugger is called from setup_arch().
5553 on xmon debugger hooks will be installed so xmon
5554 is only called on a kernel crash. Default mode,
5555 i.e. either "ro" or "rw" mode, is controlled
5556 with CONFIG_XMON_DEFAULT_RO_MODE.
5557 rw xmon debugger hooks will be installed so xmon
5558 is called only on a kernel crash, mode is write,
5559 meaning SPR registers, memory and, other data
5560 can be written using xmon commands.
5561 ro same as "rw" option above but SPR registers,
5562 memory, and other data can't be written using
5564 off xmon is disabled.