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]
25 { vendor | video | native | none }
26 If set to vendor, prefer vendor-specific driver
27 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
28 of the ACPI video.ko driver.
29 If set to video, use the ACPI video.ko driver.
30 If set to native, use the device's native backlight mode.
31 If set to none, disable the ACPI backlight interface.
33 acpi_force_32bit_fadt_addr
34 force FADT to use 32 bit addresses rather than the
35 64 bit X_* addresses. Some firmware have broken 64
36 bit addresses for force ACPI ignore these and use
37 the older legacy 32 bit addresses.
39 acpica_no_return_repair [HW, ACPI]
40 Disable AML predefined validation mechanism
41 This mechanism can repair the evaluation result to make
42 the return objects more ACPI specification compliant.
43 This option is useful for developers to identify the
44 root cause of an AML interpreter issue when the issue
45 has something to do with the repair mechanism.
47 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
48 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
50 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
51 debug output. Bits in debug_layer correspond to a
52 _COMPONENT in an ACPI source file, e.g.,
53 #define _COMPONENT ACPI_EVENTS
54 Bits in debug_level correspond to a level in
55 ACPI_DEBUG_PRINT statements, e.g.,
56 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
57 The debug_level mask defaults to "info". See
58 Documentation/firmware-guide/acpi/debug.rst for more information about
59 debug layers and levels.
61 Enable processor driver info messages:
62 acpi.debug_layer=0x20000000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
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.rst.
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>
374 arm64.nobti [ARM64] Unconditionally disable Branch Target
375 Identification support
377 arm64.nopauth [ARM64] Unconditionally disable Pointer Authentication
382 atarimouse= [HW,MOUSE] Atari Mouse
384 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
385 EzKey and similar keyboards
387 atkbd.reset= [HW] Reset keyboard during initialization
389 atkbd.set= [HW] Select keyboard code set
390 Format: <int> (2 = AT (default), 3 = PS/2)
392 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
395 atkbd.softraw= [HW] Choose between synthetic and real raw mode
396 Format: <bool> (0 = real, 1 = synthetic (default))
398 atkbd.softrepeat= [HW]
399 Use software keyboard repeat
401 audit= [KNL] Enable the audit sub-system
402 Format: { "0" | "1" | "off" | "on" }
403 0 | off - kernel audit is disabled and can not be
404 enabled until the next reboot
405 unset - kernel audit is initialized but disabled and
406 will be fully enabled by the userspace auditd.
407 1 | on - kernel audit is initialized and partially
408 enabled, storing at most audit_backlog_limit
409 messages in RAM until it is fully enabled by the
413 audit_backlog_limit= [KNL] Set the audit queue size limit.
414 Format: <int> (must be >=0)
417 bau= [X86_UV] Enable the BAU on SGI UV. The default
418 behavior is to disable the BAU (i.e. bau=0).
419 Format: { "0" | "1" }
422 unset - Disable the BAU.
424 baycom_epp= [HW,AX25]
427 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
429 See header of drivers/net/hamradio/baycom_par.c.
431 baycom_ser_fdx= [HW,AX25]
432 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
433 Format: <io>,<irq>,<mode>[,<baud>]
434 See header of drivers/net/hamradio/baycom_ser_fdx.c.
436 baycom_ser_hdx= [HW,AX25]
437 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
438 Format: <io>,<irq>,<mode>
439 See header of drivers/net/hamradio/baycom_ser_hdx.c.
441 blkdevparts= Manual partition parsing of block device(s) for
442 embedded devices based on command line input.
443 See Documentation/block/cmdline-partition.rst
445 boot_delay= Milliseconds to delay each printk during boot.
446 Values larger than 10 seconds (10000) are changed to
451 Extended command line options can be added to an initrd
452 and this will cause the kernel to look for it.
454 See Documentation/admin-guide/bootconfig.rst
457 Disable BERT OS support on buggy BIOSes.
459 bgrt_disable [ACPI][X86]
460 Disable BGRT to avoid flickering OEM logo.
462 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
463 bttv.radio= Most important insmod options are available as
465 bttv.pll= See Documentation/admin-guide/media/bttv.rst
468 bulk_remove=off [PPC] This parameter disables the use of the pSeries
469 firmware feature for flushing multiple hpte entries
472 c101= [NET] Moxa C101 synchronous serial card
474 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
475 Sometimes CPU hardware bugs make them report the cache
476 size incorrectly. The kernel will attempt work arounds
477 to fix known problems, but for some CPUs it is not
478 possible to determine what the correct size should be.
479 This option provides an override for these situations.
482 [NET] Specifies amount of time (in seconds) that
483 the kernel should wait for a network carrier. By default
484 it waits 120 seconds.
486 ca_keys= [KEYS] This parameter identifies a specific key(s) on
487 the system trusted keyring to be used for certificate
489 format: { id:<keyid> | builtin }
491 cca= [MIPS] Override the kernel pages' cache coherency
492 algorithm. Accepted values range from 0 to 7
493 inclusive. See arch/mips/include/asm/pgtable-bits.h
494 for platform specific values (SB1, Loongson3 and
497 ccw_timeout_log [S390]
498 See Documentation/s390/common_io.rst for details.
500 cgroup_disable= [KNL] Disable a particular controller
501 Format: {name of the controller(s) to disable}
502 The effects of cgroup_disable=foo are:
503 - foo isn't auto-mounted if you mount all cgroups in
505 - foo isn't visible as an individually mountable
507 {Currently only "memory" controller deal with this and
508 cut the overhead, others just disable the usage. So
509 only cgroup_disable=memory is actually worthy}
511 cgroup_no_v1= [KNL] Disable cgroup controllers and named hierarchies in v1
512 Format: { { controller | "all" | "named" }
513 [,{ controller | "all" | "named" }...] }
514 Like cgroup_disable, but only applies to cgroup v1;
515 the blacklisted controllers remain available in cgroup2.
516 "all" blacklists all controllers and "named" disables
517 named mounts. Specifying both "all" and "named" disables
520 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
522 nosocket -- Disable socket memory accounting.
523 nokmem -- Disable kernel memory accounting.
525 checkreqprot [SELINUX] Set initial checkreqprot flag value.
526 Format: { "0" | "1" }
527 See security/selinux/Kconfig help text.
528 0 -- check protection applied by kernel (includes
529 any implied execute protection).
530 1 -- check protection requested by application.
531 Default value is set via a kernel config option.
532 Value can be changed at runtime via
533 /sys/fs/selinux/checkreqprot.
534 Setting checkreqprot to 1 is deprecated.
537 See Documentation/s390/common_io.rst for details.
540 Prevents the clock framework from automatically gating
541 clocks that have not been explicitly enabled by a Linux
542 device driver but are enabled in hardware at reset or
543 by the bootloader/firmware. Note that this does not
544 force such clocks to be always-on nor does it reserve
545 those clocks in any way. This parameter is useful for
546 debug and development, but should not be needed on a
547 platform with proper driver support. For more
548 information, see Documentation/driver-api/clk.rst.
550 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
552 Forces specified clocksource (if available) to be used
553 when calculating gettimeofday(). If specified
554 clocksource is not available, it defaults to PIT.
555 Format: { pit | tsc | cyclone | pmtmr }
557 clocksource= Override the default clocksource
559 Override the default clocksource and use the clocksource
560 with the name specified.
561 Some clocksource names to choose from, depending on
563 [all] jiffies (this is the base, fallback clocksource)
565 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
566 pxa_timer,timer3,32k_counter,timer0_1
567 [X86-32] pit,hpet,tsc;
568 scx200_hrt on Geode; cyclone on IBM x440
576 clocksource.arm_arch_timer.evtstrm=
579 Enable/disable the eventstream feature of the ARM
580 architected timer so that code using WFE-based polling
581 loops can be debugged more effectively on production
584 clearcpuid=BITNUM[,BITNUM...] [X86]
585 Disable CPUID feature X for the kernel. See
586 arch/x86/include/asm/cpufeatures.h for the valid bit
587 numbers. Note the Linux specific bits are not necessarily
588 stable over kernel options, but the vendor specific
590 Also note that user programs calling CPUID directly
591 or using the feature without checking anything
592 will still see it. This just prevents it from
593 being used by the kernel or shown in /proc/cpuinfo.
594 Also note the kernel might malfunction if you disable
597 cma=nn[MG]@[start[MG][-end[MG]]]
599 Sets the size of kernel global memory area for
600 contiguous memory allocations and optionally the
601 placement constraint by the physical address range of
602 memory allocations. A value of 0 disables CMA
603 altogether. For more information, see
604 kernel/dma/contiguous.c
608 Sets the size of kernel per-numa memory area for
609 contiguous memory allocations. A value of 0 disables
610 per-numa CMA altogether. And If this option is not
611 specificed, the default value is 0.
612 With per-numa CMA enabled, DMA users on node nid will
613 first try to allocate buffer from the pernuma area
614 which is located in node nid, if the allocation fails,
615 they will fallback to the global default memory area.
617 cmo_free_hint= [PPC] Format: { yes | no }
618 Specify whether pages are marked as being inactive
619 when they are freed. This is used in CMO environments
620 to determine OS memory pressure for page stealing by
624 coherent_pool=nn[KMG] [ARM,KNL]
625 Sets the size of memory pool for coherent, atomic dma
626 allocations, by default set to 256K.
628 com20020= [HW,NET] ARCnet - COM20020 chipset
630 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
632 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
636 ARCnet - COM90xx chipset (memory-mapped buffers)
637 Format: <io>[,<irq>[,<memstart>]]
639 condev= [HW,S390] console device
642 console= [KNL] Output console device and options.
644 tty<n> Use the virtual console device <n>.
648 Use the specified serial port. The options are of
649 the form "bbbbpnf", where "bbbb" is the baud rate,
650 "p" is parity ("n", "o", or "e"), "n" is number of
651 bits, and "f" is flow control ("r" for RTS or
652 omit it). Default is "9600n8".
654 See Documentation/admin-guide/serial-console.rst for more
656 Documentation/networking/netconsole.rst for an
659 uart[8250],io,<addr>[,options]
660 uart[8250],mmio,<addr>[,options]
661 uart[8250],mmio16,<addr>[,options]
662 uart[8250],mmio32,<addr>[,options]
663 uart[8250],0x<addr>[,options]
664 Start an early, polled-mode console on the 8250/16550
665 UART at the specified I/O port or MMIO address,
666 switching to the matching ttyS device later.
667 MMIO inter-register address stride is either 8-bit
668 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
669 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
670 to be equivalent to 'mmio'. 'options' are specified in
671 the same format described for ttyS above; if unspecified,
672 the h/w is not re-initialized.
674 hvc<n> Use the hypervisor console device <n>. This is for
675 both Xen and PowerPC hypervisors.
677 If the device connected to the port is not a TTY but a braille
678 device, prepend "brl," before the device type, for instance
680 For now, only VisioBraille is supported.
683 [KNL] Change console messages format
685 By default we print messages on consoles in
686 "[time stamp] text\n" format (time stamp may not be
687 printed, depending on CONFIG_PRINTK_TIME or
688 `printk_time' param).
690 Switch to syslog format: "<%u>[time stamp] text\n"
691 IOW, each message will have a facility and loglevel
692 prefix. The format is similar to one used by syslog()
693 syscall, or to executing "dmesg -S --raw" or to reading
696 consoleblank= [KNL] The console blank (screen saver) timeout in
697 seconds. A value of 0 disables the blank timer.
701 [KNL] Change the default value for
702 /proc/<pid>/coredump_filter.
703 See also Documentation/filesystems/proc.rst.
705 coresight_cpu_debug.enable
708 Enable/disable the CPU sampling based debugging.
709 0: default value, disable debugging
710 1: enable debugging at boot time
712 cpuidle.off=1 [CPU_IDLE]
713 disable the cpuidle sub-system
716 [CPU_IDLE] Name of the cpuidle governor to use.
718 cpufreq.off=1 [CPU_FREQ]
719 disable the cpufreq sub-system
721 cpufreq.default_governor=
722 [CPU_FREQ] Name of the default cpufreq governor or
723 policy to use. This governor must be registered in the
724 kernel before the cpufreq driver probes.
727 [X86] Delay for N microsec between assert and de-assert
728 of APIC INIT to start processors. This delay occurs
729 on every CPU online, such as boot, and resume from suspend.
732 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
734 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
736 crashkernel=size[KMG][@offset[KMG]]
737 [KNL] Using kexec, Linux can switch to a 'crash kernel'
738 upon panic. This parameter reserves the physical
739 memory region [offset, offset + size] for that kernel
740 image. If '@offset' is omitted, then a suitable offset
741 is selected automatically.
742 [KNL, X86-64] Select a region under 4G first, and
743 fall back to reserve region above 4G when '@offset'
744 hasn't been specified.
745 See Documentation/admin-guide/kdump/kdump.rst for further details.
747 crashkernel=range1:size1[,range2:size2,...][@offset]
748 [KNL] Same as above, but depends on the memory
749 in the running system. The syntax of range is
750 start-[end] where start and end are both
751 a memory unit (amount[KMG]). See also
752 Documentation/admin-guide/kdump/kdump.rst for an example.
754 crashkernel=size[KMG],high
755 [KNL, X86-64] range could be above 4G. Allow kernel
756 to allocate physical memory region from top, so could
757 be above 4G if system have more than 4G ram installed.
758 Otherwise memory region will be allocated below 4G, if
760 It will be ignored if crashkernel=X is specified.
761 crashkernel=size[KMG],low
762 [KNL, X86-64] range under 4G. When crashkernel=X,high
763 is passed, kernel could allocate physical memory region
764 above 4G, that cause second kernel crash on system
765 that require some amount of low memory, e.g. swiotlb
766 requires at least 64M+32K low memory, also enough extra
767 low memory is needed to make sure DMA buffers for 32-bit
768 devices won't run out. Kernel would try to allocate at
769 at least 256M below 4G automatically.
770 This one let user to specify own low range under 4G
771 for second kernel instead.
772 0: to disable low allocation.
773 It will be ignored when crashkernel=X,high is not used
774 or memory reserved is below 4G.
777 [KNL] Disable crypto self-tests
782 cs89x0_media= [HW,NET]
783 Format: { rj45 | aui | bnc }
785 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
786 handling. When switched on, additional debug data is
787 printed to the console in case a hanging CPU is
788 detected, and that CPU is pinged again in order to try
789 to resolve the hang situation.
790 0: disable csdlock debugging (default)
791 1: enable basic csdlock debugging (minor impact)
792 ext: enable extended csdlock debugging (more impact,
796 See header of drivers/s390/block/dasd_devmap.c.
798 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
799 (one device per port)
800 Format: <port#>,<type>
801 See also Documentation/input/devices/joystick-parport.rst
803 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
805 Documentation/admin-guide/dynamic-debug-howto.rst for
806 details. Deprecated, see dyndbg.
808 debug [KNL] Enable kernel debugging (events log level).
811 [KNL] Enable printing [hashed] pointers early in the
812 boot sequence. If enabled, we use a weak hash instead
813 of siphash to hash pointers. Use this option if you are
814 seeing instances of '(___ptrval___)') and need to see a
815 value (hashed pointer) instead. Cryptographically
816 insecure, please do not use on production kernels.
819 [KNL] verbose locking self-tests
821 Print debugging info while doing the locking API
823 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
824 (no extra messages), setting it to -1 (all bits set)
825 will print _a_lot_ more information - normally only
826 useful to lockdep developers.
828 debug_objects [KNL] Enable object debugging
831 [KNL] Disable object debugging
833 debug_guardpage_minorder=
834 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
835 parameter allows control of the order of pages that will
836 be intentionally kept free (and hence protected) by the
837 buddy allocator. Bigger value increase the probability
838 of catching random memory corruption, but reduce the
839 amount of memory for normal system use. The maximum
840 possible value is MAX_ORDER/2. Setting this parameter
841 to 1 or 2 should be enough to identify most random
842 memory corruption problems caused by bugs in kernel or
843 driver code when a CPU writes to (or reads from) a
844 random memory location. Note that there exists a class
845 of memory corruptions problems caused by buggy H/W or
846 F/W or by drivers badly programing DMA (basically when
847 memory is written at bus level and the CPU MMU is
848 bypassed) which are not detectable by
849 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
850 tracking down these problems.
853 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
854 enables the feature at boot time. By default, it is
855 disabled and the system will work mostly the same as a
856 kernel built without CONFIG_DEBUG_PAGEALLOC.
857 Note: to get most of debug_pagealloc error reports, it's
858 useful to also enable the page_owner functionality.
859 on: enable the feature
861 debugfs= [KNL] This parameter enables what is exposed to userspace
862 and debugfs internal clients.
863 Format: { on, no-mount, off }
864 on: All functions are enabled.
866 Filesystem is not registered but kernel clients can
867 access APIs and a crashkernel can be used to read
868 its content. There is nothing to mount.
869 off: Filesystem is not registered and clients
870 get a -EPERM as result when trying to register files
871 or directories within debugfs.
872 This is equivalent of the runtime functionality if
873 debugfs was not enabled in the kernel at all.
874 Default value is set in build-time with a kernel configuration.
876 debugpat [X86] Enable PAT debugging
878 decnet.addr= [HW,NET]
879 Format: <area>[,<node>]
880 See also Documentation/networking/decnet.rst.
883 [HW] The size of the default HugeTLB page. This is
884 the size represented by the legacy /proc/ hugepages
885 APIs. In addition, this is the default hugetlb size
886 used for shmget(), mmap() and mounting hugetlbfs
887 filesystems. If not specified, defaults to the
888 architecture's default huge page size. Huge page
889 sizes are architecture dependent. See also
890 Documentation/admin-guide/mm/hugetlbpage.rst.
893 deferred_probe_timeout=
894 [KNL] Debugging option to set a timeout in seconds for
895 deferred probe to give up waiting on dependencies to
896 probe. Only specific dependencies (subsystems or
897 drivers) that have opted in will be ignored. A timeout of 0
898 will timeout at the end of initcalls. This option will also
899 dump out devices still on the deferred probe list after
903 Format: { on | off | def_only | inf_only | always }
904 on: s390 zlib hardware support for compression on
905 level 1 and decompression (default)
906 off: No s390 zlib hardware support
907 def_only: s390 zlib hardware support for deflate
908 only (compression on level 1)
909 inf_only: s390 zlib hardware support for inflate
911 always: Same as 'on' but ignores the selected compression
912 level always using hardware support (used for debugging)
915 Set number of hash buckets for dentry cache.
917 disable_1tb_segments [PPC]
918 Disables the use of 1TB hash page table segments. This
919 causes the kernel to fall back to 256MB segments which
920 can be useful when debugging issues that require an SLB
924 Limits the number of kernel SLB entries, and flushes
925 them frequently to increase the rate of SLB faults
929 See Documentation/networking/ipv6.rst.
932 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
933 hardening is enabled for this boot. Hardened
934 usercopy checking is used to protect the kernel
935 from reading or writing beyond known memory
936 allocation boundaries as a proactive defense
937 against bounds-checking flaws in the kernel's
938 copy_to_user()/copy_from_user() interface.
939 on Perform hardened usercopy checks (default).
940 off Disable hardened usercopy checks.
943 Disable RADIX MMU mode on POWER9
945 radix_hcall_invalidate=on [PPC/PSERIES]
946 Disable RADIX GTSE feature and use hcall for TLB
950 Disable TLBIE instruction. Currently does not work
951 with KVM, with HASH MMU, or with coherent accelerators.
953 disable_cpu_apicid= [X86,APIC,SMP]
955 The number of initial APIC ID for the
956 corresponding CPU to be disabled at boot,
957 mostly used for the kdump 2nd kernel to
958 disable BSP to wake up multiple CPUs without
959 causing system reset or hang due to sending
962 disable_ddw [PPC/PSERIES]
963 Disable Dynamic DMA Window support. Use this
964 to workaround buggy firmware.
967 See Documentation/networking/ipv6.rst.
969 disable_mtrr_cleanup [X86]
970 The kernel tries to adjust MTRR layout from continuous
971 to discrete, to make X server driver able to add WB
972 entry later. This parameter disables that.
974 disable_mtrr_trim [X86, Intel and AMD only]
975 By default the kernel will trim any uncacheable
976 memory out of your available memory pool based on
977 MTRR settings. This parameter disables that behavior,
978 possibly causing your machine to run very slowly.
980 disable_timer_pin_1 [X86]
981 Disable PIN 1 of APIC timer
982 Can be useful to work around chipset bugs.
984 dis_ucode_ldr [X86] Disable the microcode loader.
986 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
987 this option disables the debugging code at boot.
989 dma_debug_entries=<number>
990 This option allows to tune the number of preallocated
991 entries for DMA-API debugging code. One entry is
992 required per DMA-API allocation. Use this if the
993 DMA-API debugging code disables itself because the
994 architectural default is too low.
996 dma_debug_driver=<driver_name>
997 With this option the DMA-API debugging driver
998 filter feature can be enabled at boot time. Just
999 pass the driver to filter for as the parameter.
1000 The filter can be disabled or changed to another
1001 driver later using sysfs.
1003 driver_async_probe= [KNL]
1004 List of driver names to be probed asynchronously.
1005 Format: <driver_name1>,<driver_name2>...
1007 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1008 Broken monitors, graphic adapters, KVMs and EDIDless
1009 panels may send no or incorrect EDID data sets.
1010 This parameter allows to specify an EDID data sets
1011 in the /lib/firmware directory that are used instead.
1012 Generic built-in EDID data sets are used, if one of
1013 edid/1024x768.bin, edid/1280x1024.bin,
1014 edid/1680x1050.bin, or edid/1920x1080.bin is given
1015 and no file with the same name exists. Details and
1016 instructions how to build your own EDID data are
1017 available in Documentation/admin-guide/edid.rst. An EDID
1018 data set will only be used for a particular connector,
1019 if its name and a colon are prepended to the EDID
1020 name. Each connector may use a unique EDID data
1021 set by separating the files with a comma. An EDID
1022 data set with no connector name will be used for
1023 any connectors not explicitly specified.
1028 Format: {"off" | "known"}
1029 Control how the dt_cpu_ftrs device-tree binding is
1030 used for CPU feature discovery and setup (if it
1032 off: Do not use it, fall back to legacy cpu table.
1033 known: Do not pass through unknown features to guests
1034 or userspace, only those that the kernel is aware of.
1036 dump_apple_properties [X86]
1037 Dump name and content of EFI device properties on
1038 x86 Macs. Useful for driver authors to determine
1039 what data is available or for reverse-engineering.
1041 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1042 <module>.dyndbg[="val"]
1043 Enable debug messages at boot time. See
1044 Documentation/admin-guide/dynamic-debug-howto.rst
1047 nopku [X86] Disable Memory Protection Keys CPU feature found
1050 <module>.async_probe [KNL]
1051 Enable asynchronous probe on this module.
1053 early_ioremap_debug [KNL]
1054 Enable debug messages in early_ioremap support. This
1055 is useful for tracking down temporary early mappings
1056 which are not unmapped.
1058 earlycon= [KNL] Output early console device and options.
1060 When used with no options, the early console is
1061 determined by stdout-path property in device tree's
1062 chosen node or the ACPI SPCR table if supported by
1065 cdns,<addr>[,options]
1066 Start an early, polled-mode console on a Cadence
1067 (xuartps) serial port at the specified address. Only
1068 supported option is baud rate. If baud rate is not
1069 specified, the serial port must already be setup and
1072 uart[8250],io,<addr>[,options]
1073 uart[8250],mmio,<addr>[,options]
1074 uart[8250],mmio32,<addr>[,options]
1075 uart[8250],mmio32be,<addr>[,options]
1076 uart[8250],0x<addr>[,options]
1077 Start an early, polled-mode console on the 8250/16550
1078 UART at the specified I/O port or MMIO address.
1079 MMIO inter-register address stride is either 8-bit
1080 (mmio) or 32-bit (mmio32 or mmio32be).
1081 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1082 to be equivalent to 'mmio'. 'options' are specified
1083 in the same format described for "console=ttyS<n>"; if
1084 unspecified, the h/w is not initialized.
1088 Start an early, polled-mode console on a pl011 serial
1089 port at the specified address. The pl011 serial port
1090 must already be setup and configured. Options are not
1091 yet supported. If 'mmio32' is specified, then only
1092 the driver will use only 32-bit accessors to read/write
1093 the device registers.
1096 Start an early, polled-mode console on a meson serial
1097 port at the specified address. The serial port must
1098 already be setup and configured. Options are not yet
1102 Start an early, polled-mode console on an msm serial
1103 port at the specified address. The serial port
1104 must already be setup and configured. Options are not
1107 msm_serial_dm,<addr>
1108 Start an early, polled-mode console on an msm serial
1109 dm port at the specified address. The serial port
1110 must already be setup and configured. Options are not
1114 Start an early, polled-mode console on a serial port
1115 of an Actions Semi SoC, such as S500 or S900, at the
1116 specified address. The serial port must already be
1117 setup and configured. Options are not yet supported.
1120 Start an early, polled-mode console on a serial port
1121 of an RDA Micro SoC, such as RDA8810PL, at the
1122 specified address. The serial port must already be
1123 setup and configured. Options are not yet supported.
1126 Use RISC-V SBI (Supervisor Binary Interface) for early
1129 smh Use ARM semihosting calls for early console.
1137 Use early console provided by serial driver available
1138 on Samsung SoCs, requires selecting proper type and
1139 a correct base address of the selected UART port. The
1140 serial port must already be setup and configured.
1141 Options are not yet supported.
1144 Start an early, polled-mode console on a lantiq serial
1145 (lqasc) port at the specified address. The serial port
1146 must already be setup and configured. Options are not
1151 Use early console provided by Freescale LP UART driver
1152 found on Freescale Vybrid and QorIQ LS1021A processors.
1153 A valid base address must be provided, and the serial
1154 port must already be setup and configured.
1158 Start an early, polled-mode, output-only console on the
1159 Freescale i.MX UART at the specified address. The UART
1160 must already be setup and configured.
1163 Start an early, polled-mode console on the
1164 Armada 3700 serial port at the specified
1165 address. The serial port must already be setup
1166 and configured. Options are not yet supported.
1169 Start an early, polled-mode console on a Qualcomm
1170 Generic Interface (GENI) based serial port at the
1171 specified address. The serial port must already be
1172 setup and configured. Options are not yet supported.
1175 Start an early, unaccelerated console on the EFI
1176 memory mapped framebuffer (if available). On cache
1177 coherent non-x86 systems that use system memory for
1178 the framebuffer, pass the 'ram' option so that it is
1179 mapped with the correct attributes.
1182 Use early console provided by Freescale LINFlexD UART
1183 serial driver for NXP S32V234 SoCs. A valid base
1184 address must be provided, and the serial port must
1185 already be setup and configured.
1187 earlyprintk= [X86,SH,ARM,M68k,S390]
1191 earlyprintk=serial[,ttySn[,baudrate]]
1192 earlyprintk=serial[,0x...[,baudrate]]
1193 earlyprintk=ttySn[,baudrate]
1194 earlyprintk=dbgp[debugController#]
1195 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1196 earlyprintk=xdbc[xhciController#]
1198 earlyprintk is useful when the kernel crashes before
1199 the normal console is initialized. It is not enabled by
1200 default because it has some cosmetic problems.
1202 Append ",keep" to not disable it when the real console
1205 Only one of vga, efi, serial, or usb debug port can
1208 Currently only ttyS0 and ttyS1 may be specified by
1209 name. Other I/O ports may be explicitly specified
1210 on some architectures (x86 and arm at least) by
1211 replacing ttySn with an I/O port address, like this:
1212 earlyprintk=serial,0x1008,115200
1213 You can find the port for a given device in
1214 /proc/tty/driver/serial:
1215 2: uart:ST16650V2 port:00001008 irq:18 ...
1217 Interaction with the standard serial driver is not
1220 The VGA and EFI output is eventually overwritten by
1223 The xen output can only be used by Xen PV guests.
1225 The sclp output can only be used on s390.
1227 The optional "force" to "pciserial" enables use of a
1228 PCI device even when its classcode is not of the
1231 edac_report= [HW,EDAC] Control how to report EDAC event
1232 Format: {"on" | "off" | "force"}
1233 on: enable EDAC to report H/W event. May be overridden
1234 by other higher priority error reporting module.
1235 off: disable H/W event reporting through EDAC.
1236 force: enforce the use of EDAC to report H/W event.
1239 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1242 This is designed to be used in conjunction with
1243 the boot argument: earlyprintk=vga
1245 This parameter works in place of the kgdboc parameter
1246 but can only be used if the backing tty is available
1247 very early in the boot process. For early debugging
1248 via a serial port see kgdboc_earlycon instead.
1251 Format: {"off" | "on" | "skip[mbr]"}
1254 Format: { "debug", "disable_early_pci_dma",
1255 "nochunk", "noruntime", "nosoftreserve",
1256 "novamap", "no_disable_early_pci_dma" }
1257 debug: enable misc debug output.
1258 disable_early_pci_dma: disable the busmaster bit on all
1259 PCI bridges while in the EFI boot stub.
1260 nochunk: disable reading files in "chunks" in the EFI
1261 boot stub, as chunking can cause problems with some
1262 firmware implementations.
1263 noruntime : disable EFI runtime services support
1264 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1265 attribute may cause the kernel to reserve the
1266 memory range for a memory mapping driver to
1267 claim. Specify efi=nosoftreserve to disable this
1268 reservation and treat the memory by its base type
1269 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1270 novamap: do not call SetVirtualAddressMap().
1271 no_disable_early_pci_dma: Leave the busmaster bit set
1272 on all PCI bridges while in the EFI boot stub
1274 efi_no_storage_paranoia [EFI; X86]
1275 Using this parameter you can use more than 50% of
1276 your efi variable storage. Use this parameter only if
1277 you are really sure that your UEFI does sane gc and
1278 fulfills the spec otherwise your board may brick.
1280 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1281 Add arbitrary attribute to specific memory range by
1282 updating original EFI memory map.
1283 Region of memory which aa attribute is added to is
1286 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1287 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1288 attribute is added to range 0x100000000-0x180000000 and
1289 0x10a0000000-0x1120000000.
1291 If efi_fake_mem=8G@9G:0x40000 is specified, the
1292 EFI_MEMORY_SP(0x40000) attribute is added to
1293 range 0x240000000-0x43fffffff.
1295 Using this parameter you can do debugging of EFI memmap
1296 related features. For example, you can do debugging of
1297 Address Range Mirroring feature even if your box
1298 doesn't support it, or mark specific memory as
1301 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1302 that is to be dynamically loaded by Linux. If there are
1303 multiple variables with the same name but with different
1304 vendor GUIDs, all of them will be loaded. See
1305 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1308 eisa_irq_edge= [PARISC,HW]
1309 See header of drivers/parisc/eisa.c.
1312 See comment before function elanfreq_setup() in
1313 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1315 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1316 Specifies physical address of start of kernel core
1317 image elf header and optionally the size. Generally
1318 kexec loader will pass this option to capture kernel.
1319 See Documentation/admin-guide/kdump/kdump.rst for details.
1321 enable_mtrr_cleanup [X86]
1322 The kernel tries to adjust MTRR layout from continuous
1323 to discrete, to make X server driver able to add WB
1324 entry later. This parameter enables that.
1326 enable_timer_pin_1 [X86]
1327 Enable PIN 1 of APIC timer
1328 Can be useful to work around chipset bugs
1329 (in particular on some ATI chipsets).
1330 The kernel tries to set a reasonable default.
1332 enforcing [SELINUX] Set initial enforcing status.
1334 See security/selinux/Kconfig help text.
1335 0 -- permissive (log only, no denials).
1336 1 -- enforcing (deny and log).
1338 Value can be changed at runtime via
1339 /sys/fs/selinux/enforce.
1342 Disable Error Record Serialization Table (ERST)
1345 ether= [HW,NET] Ethernet cards parameters
1346 This option is obsoleted by the "netdev=" option, which
1347 has equivalent usage. See its documentation for details.
1351 Permit 'security.evm' to be updated regardless of
1352 current integrity status.
1357 fail_make_request=[KNL]
1358 General fault injection mechanism.
1359 Format: <interval>,<probability>,<space>,<times>
1360 See also Documentation/fault-injection/.
1363 Format: { initns | none }
1364 See Documentation/admin-guide/sysctl/net.rst for
1365 fb_tunnels_only_for_init_ns
1368 See Documentation/admin-guide/blockdev/floppy.rst.
1370 force_pal_cache_flush
1371 [IA-64] Avoid check_sal_cache_flush which may hang on
1372 buggy SAL_CACHE_FLUSH implementations. Using this
1373 parameter will force ia64_sal_cache_flush to call
1374 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1377 Forcefully enable Physical Address Extension (PAE).
1378 Many Pentium M systems disable PAE but may have a
1379 functionally usable PAE implementation.
1380 Warning: use of this parameter will taint the kernel
1381 and may cause unknown problems.
1384 [FTRACE] will set and start the specified tracer
1385 as early as possible in order to facilitate early
1388 ftrace_dump_on_oops[=orig_cpu]
1389 [FTRACE] will dump the trace buffers on oops.
1390 If no parameter is passed, ftrace will dump
1391 buffers of all CPUs, but if you pass orig_cpu, it will
1392 dump only the buffer of the CPU that triggered the
1395 ftrace_filter=[function-list]
1396 [FTRACE] Limit the functions traced by the function
1397 tracer at boot up. function-list is a comma-separated
1398 list of functions. This list can be changed at run
1399 time by the set_ftrace_filter file in the debugfs
1402 ftrace_notrace=[function-list]
1403 [FTRACE] Do not trace the functions specified in
1404 function-list. This list can be changed at run time
1405 by the set_ftrace_notrace file in the debugfs
1408 ftrace_graph_filter=[function-list]
1409 [FTRACE] Limit the top level callers functions traced
1410 by the function graph tracer at boot up.
1411 function-list is a comma-separated list of functions
1412 that can be changed at run time by the
1413 set_graph_function file in the debugfs tracing directory.
1415 ftrace_graph_notrace=[function-list]
1416 [FTRACE] Do not trace from the functions specified in
1417 function-list. This list is a comma-separated list of
1418 functions that can be changed at run time by the
1419 set_graph_notrace file in the debugfs tracing directory.
1421 ftrace_graph_max_depth=<uint>
1422 [FTRACE] Used with the function graph tracer. This is
1423 the max depth it will trace into a function. This value
1424 can be changed at run time by the max_graph_depth file
1425 in the tracefs tracing directory. default: 0 (no limit)
1427 fw_devlink= [KNL] Create device links between consumer and supplier
1428 devices by scanning the firmware to infer the
1429 consumer/supplier relationships. This feature is
1430 especially useful when drivers are loaded as modules as
1431 it ensures proper ordering of tasks like device probing
1432 (suppliers first, then consumers), supplier boot state
1433 clean up (only after all consumers have probed),
1434 suspend/resume & runtime PM (consumers first, then
1436 Format: { off | permissive | on | rpm }
1437 off -- Don't create device links from firmware info.
1438 permissive -- Create device links from firmware info
1439 but use it only for ordering boot state clean
1440 up (sync_state() calls).
1441 on -- Create device links from firmware info and use it
1442 to enforce probe and suspend/resume ordering.
1443 rpm -- Like "on", but also use to order runtime PM.
1445 fw_devlink.strict=<bool>
1446 [KNL] Treat all inferred dependencies as mandatory
1447 dependencies. This only applies for fw_devlink=on|rpm.
1451 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1452 support via parallel port (up to 5 devices per port)
1453 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1454 See also Documentation/input/devices/joystick-parport.rst
1458 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1462 gcov_persist= [GCOV] When non-zero (default), profiling data for
1463 kernel modules is saved and remains accessible via
1464 debugfs, even when the module is unloaded/reloaded.
1465 When zero, profiling data is discarded and associated
1466 debugfs files are removed at module unload time.
1468 goldfish [X86] Enable the goldfish android emulator platform.
1469 Don't use this when you are not running on the
1472 gpio-mockup.gpio_mockup_ranges
1473 [HW] Sets the ranges of gpiochip of for this device.
1474 Format: <start1>,<end1>,<start2>,<end2>...
1475 gpio-mockup.gpio_mockup_named_lines
1476 [HW] Let the driver know GPIO lines should be named.
1478 gpt [EFI] Forces disk with valid GPT signature but
1479 invalid Protective MBR to be treated as GPT. If the
1480 primary GPT is corrupted, it enables the backup/alternate
1481 GPT to be used instead.
1483 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1484 the "Enable 0" bit of the configuration register.
1487 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1488 the "Enable 0" bit of the configuration register.
1491 grcan.select= [HW] Select which physical interface to use.
1494 grcan.txsize= [HW] Sets the size of the tx buffer.
1495 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1497 grcan.rxsize= [HW] Sets the size of the rx buffer.
1498 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1501 hardlockup_all_cpu_backtrace=
1502 [KNL] Should the hard-lockup detector generate
1503 backtraces on all cpus.
1506 hashdist= [KNL,NUMA] Large hashes allocated during boot
1507 are distributed across NUMA nodes. Defaults on
1508 for 64-bit NUMA, off otherwise.
1509 Format: 0 | 1 (for off | on)
1511 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1513 hd= [EIDE] (E)IDE hard drive subsystem geometry
1514 Format: <cyl>,<head>,<sect>
1517 Disable Hardware Error Source Table (HEST) support;
1518 corresponding firmware-first mode error processing
1519 logic will be disabled.
1521 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1522 size of <nn>. This works even on boxes that have no
1523 highmem otherwise. This also works to reduce highmem
1524 size on bigger boxes.
1526 highres= [KNL] Enable/disable high resolution timer mode.
1527 Valid parameters: "on", "off"
1532 hpet= [X86-32,HPET] option to control HPET usage
1533 Format: { enable (default) | disable | force |
1535 disable: disable HPET and use PIT instead
1536 force: allow force enabled of undocumented chips (ICH4,
1538 verbose: show contents of HPET registers during setup
1540 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1541 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1543 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1544 of gigantic hugepages.
1547 Reserve a CMA area of given size and allocate gigantic
1548 hugepages using the CMA allocator. If enabled, the
1549 boot-time allocation of gigantic hugepages is skipped.
1551 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1552 If this follows hugepagesz (below), it specifies
1553 the number of pages of hugepagesz to be allocated.
1554 If this is the first HugeTLB parameter on the command
1555 line, it specifies the number of pages to allocate for
1556 the default huge page size. See also
1557 Documentation/admin-guide/mm/hugetlbpage.rst.
1561 [HW] The size of the HugeTLB pages. This is used in
1562 conjunction with hugepages (above) to allocate huge
1563 pages of a specific size at boot. The pair
1564 hugepagesz=X hugepages=Y can be specified once for
1565 each supported huge page size. Huge page sizes are
1566 architecture dependent. See also
1567 Documentation/admin-guide/mm/hugetlbpage.rst.
1571 [KNL] Should the hung task detector generate panics.
1574 A value of 1 instructs the kernel to panic when a
1575 hung task is detected. The default value is controlled
1576 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1577 option. The value selected by this boot parameter can
1578 be changed later by the kernel.hung_task_panic sysctl.
1580 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1581 terminal devices. Valid values: 0..8
1582 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1583 If specified, z/VM IUCV HVC accepts connections
1584 from listed z/VM user IDs only.
1586 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1587 which allow the hypervisor to 'idle' the
1588 guest on lock contention.
1591 Do not unregister boot console at start. This is only
1592 useful for debugging when something happens in the window
1593 between unregistering the boot console and initializing
1596 i2c_bus= [HW] Override the default board specific I2C bus speed
1597 or register an additional I2C bus that is not
1598 registered from board initialization code.
1602 i8042.debug [HW] Toggle i8042 debug mode
1603 i8042.unmask_kbd_data
1604 [HW] Enable printing of interrupt data from the KBD port
1605 (disabled by default, and as a pre-condition
1606 requires that i8042.debug=1 be enabled)
1607 i8042.direct [HW] Put keyboard port into non-translated mode
1608 i8042.dumbkbd [HW] Pretend that controller can only read data from
1609 keyboard and cannot control its state
1610 (Don't attempt to blink the leds)
1611 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1612 i8042.nokbd [HW] Don't check/create keyboard port
1613 i8042.noloop [HW] Disable the AUX Loopback command while probing
1615 i8042.nomux [HW] Don't check presence of an active multiplexing
1617 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1619 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1620 i8042.reset [HW] Reset the controller during init, cleanup and
1621 suspend-to-ram transitions, only during s2r
1622 transitions, or never reset
1623 Format: { 1 | Y | y | 0 | N | n }
1624 1, Y, y: always reset controller
1625 0, N, n: don't ever reset controller
1626 Default: only on s2r transitions on x86; most other
1627 architectures force reset to be always executed
1628 i8042.unlock [HW] Unlock (ignore) the keylock
1629 i8042.kbdreset [HW] Reset device connected to KBD port
1633 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1634 indicates that the driver is running on unsupported
1636 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1637 does not match list of supported models.
1639 [HW] Report power status in /proc/i8k
1640 (disabled by default)
1641 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1644 i915.invert_brightness=
1645 [DRM] Invert the sense of the variable that is used to
1646 set the brightness of the panel backlight. Normally a
1647 brightness value of 0 indicates backlight switched off,
1648 and the maximum of the brightness value sets the backlight
1649 to maximum brightness. If this parameter is set to 0
1650 (default) and the machine requires it, or this parameter
1651 is set to 1, a brightness value of 0 sets the backlight
1652 to maximum brightness, and the maximum of the brightness
1653 value switches the backlight off.
1654 -1 -- never invert brightness
1655 0 -- machine default
1656 1 -- force brightness inversion
1659 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1661 ide-core.nodma= [HW] (E)IDE subsystem
1662 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1663 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1664 .cdrom .chs .ignore_cable are additional options
1665 See Documentation/ide/ide.rst.
1667 ide-generic.probe-mask= [HW] (E)IDE subsystem
1669 Probe mask for legacy ISA IDE ports. Depending on
1670 platform up to 6 ports are supported, enabled by
1671 setting corresponding bits in the mask to 1. The
1672 default value is 0x0, which has a special meaning.
1673 On systems that have PCI, it triggers scanning the
1674 PCI bus for the first and the second port, which
1675 are then probed. On systems without PCI the value
1676 of 0x0 enables probing the two first ports as if it
1679 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1680 Claim all unknown PCI IDE storage controllers.
1683 Format: idle=poll, idle=halt, idle=nomwait
1684 Poll forces a polling idle loop that can slightly
1685 improve the performance of waking up a idle CPU, but
1686 will use a lot of power and make the system run hot.
1688 idle=halt: Halt is forced to be used for CPU idle.
1689 In such case C2/C3 won't be used again.
1690 idle=nomwait: Disable mwait for CPU C-states
1694 Allow force disabling of Shared Virtual Memory (SVA)
1695 support for the idxd driver. By default it is set to
1698 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1699 Format: { strict | legacy | 2008 | relaxed }
1702 Choose which programs will be accepted for execution
1703 based on the IEEE 754 NaN encoding(s) supported by
1704 the FPU and the NaN encoding requested with the value
1705 of an ELF file header flag individually set by each
1706 binary. Hardware implementations are permitted to
1707 support either or both of the legacy and the 2008 NaN
1710 Available settings are as follows:
1711 strict accept binaries that request a NaN encoding
1712 supported by the FPU
1713 legacy only accept legacy-NaN binaries, if supported
1715 2008 only accept 2008-NaN binaries, if supported
1717 relaxed accept any binaries regardless of whether
1718 supported by the FPU
1720 The FPU emulator is always able to support both NaN
1721 encodings, so if no FPU hardware is present or it has
1722 been disabled with 'nofpu', then the settings of
1723 'legacy' and '2008' strap the emulator accordingly,
1724 'relaxed' straps the emulator for both legacy-NaN and
1725 2008-NaN, whereas 'strict' enables legacy-NaN only on
1726 legacy processors and both NaN encodings on MIPS32 or
1729 The setting for ABS.fmt/NEG.fmt instruction execution
1730 mode generally follows that for the NaN encoding,
1731 except where unsupported by hardware.
1733 ignore_loglevel [KNL]
1734 Ignore loglevel setting - this will print /all/
1735 kernel messages to the console. Useful for debugging.
1736 We also add it as printk module parameter, so users
1737 could change it dynamically, usually by
1738 /sys/module/printk/parameters/ignore_loglevel.
1741 Ignore RLIMIT_DATA setting for data mappings,
1742 print warning at first misuse. Can be changed via
1743 /sys/module/kernel/parameters/ignore_rlimit_data.
1745 ihash_entries= [KNL]
1746 Set number of hash buckets for inode cache.
1748 ima_appraise= [IMA] appraise integrity measurements
1749 Format: { "off" | "enforce" | "fix" | "log" }
1752 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1753 The builtin appraise policy appraises all files
1756 ima_canonical_fmt [IMA]
1757 Use the canonical format for the binary runtime
1758 measurements, instead of host native format.
1761 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1765 The list of supported hash algorithms is defined
1766 in crypto/hash_info.h.
1769 The builtin policies to load during IMA setup.
1770 Format: "tcb | appraise_tcb | secure_boot |
1771 fail_securely | critical_data"
1773 The "tcb" policy measures all programs exec'd, files
1774 mmap'd for exec, and all files opened with the read
1775 mode bit set by either the effective uid (euid=0) or
1778 The "appraise_tcb" policy appraises the integrity of
1779 all files owned by root.
1781 The "secure_boot" policy appraises the integrity
1782 of files (eg. kexec kernel image, kernel modules,
1783 firmware, policy, etc) based on file signatures.
1785 The "fail_securely" policy forces file signature
1786 verification failure also on privileged mounted
1787 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1790 The "critical_data" policy measures kernel integrity
1793 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1794 Load a policy which meets the needs of the Trusted
1795 Computing Base. This means IMA will measure all
1796 programs exec'd, files mmap'd for exec, and all files
1797 opened for read by uid=0.
1800 Select one of defined IMA measurements template formats.
1801 Formats: { "ima" | "ima-ng" | "ima-sig" }
1805 [IMA] Define a custom template format.
1806 Format: { "field1|...|fieldN" }
1808 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1809 Format: <min_file_size>
1810 Set the minimal file size for using asynchronous hash.
1811 If left unspecified, ahash usage is disabled.
1813 ahash performance varies for different data sizes on
1814 different crypto accelerators. This option can be used
1815 to achieve the best performance for a particular HW.
1817 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1819 Set hashing buffer size. Default: 4k.
1821 ahash performance varies for different chunk sizes on
1822 different crypto accelerators. This option can be used
1823 to achieve best performance for particular HW.
1827 Run specified binary instead of /sbin/init as init
1830 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1831 for working out where the kernel is dying during
1834 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1835 initcall functions. Useful for debugging built-in
1836 modules and initcalls.
1838 initramfs_async= [KNL]
1841 This parameter controls whether the initramfs
1842 image is unpacked asynchronously, concurrently
1843 with devices being probed and
1844 initialized. This should normally just work,
1845 but as a debugging aid, one can get the
1846 historical behaviour of the initramfs
1847 unpacking being completed before device_ and
1850 initrd= [BOOT] Specify the location of the initial ramdisk
1852 initrdmem= [KNL] Specify a physical address and size from which to
1853 load the initrd. If an initrd is compiled in or
1854 specified in the bootparams, it takes priority over this
1856 Format: ss[KMG],nn[KMG]
1859 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1862 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1864 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1866 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1868 init_pkru= [X86] Specify the default memory protection keys rights
1869 register contents for all processes. 0x55555554 by
1870 default (disallow access to all but pkey 0). Can
1871 override in debugfs after boot.
1873 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1876 int_pln_enable [X86] Enable power limit notification interrupt
1878 integrity_audit=[IMA]
1879 Format: { "0" | "1" }
1880 0 -- basic integrity auditing messages. (Default)
1881 1 -- additional integrity auditing messages.
1883 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1885 Enable intel iommu driver.
1887 Disable intel iommu driver.
1888 igfx_off [Default Off]
1889 By default, gfx is mapped as normal device. If a gfx
1890 device has a dedicated DMAR unit, the DMAR unit is
1891 bypassed by not enabling DMAR with this option. In
1892 this case, gfx device will use physical address for
1894 strict [Default Off]
1895 With this option on every unmap_single operation will
1896 result in a hardware IOTLB flush operation as opposed
1897 to batching them for performance.
1898 sp_off [Default Off]
1899 By default, super page will be supported if Intel IOMMU
1900 has the capability. With this option, super page will
1903 By default, scalable mode will be disabled even if the
1904 hardware advertises that it has support for the scalable
1905 mode translation. With this option set, scalable mode
1906 will be used on hardware which claims to support it.
1907 tboot_noforce [Default Off]
1908 Do not force the Intel IOMMU enabled under tboot.
1909 By default, tboot will force Intel IOMMU on, which
1910 could harm performance of some high-throughput
1911 devices like 40GBit network cards, even if identity
1913 Note that using this option lowers the security
1914 provided by tboot because it makes the system
1915 vulnerable to DMA attacks.
1917 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1918 0 disables intel_idle and fall back on acpi_idle.
1919 1 to 9 specify maximum depth of C-state.
1923 Do not enable intel_pstate as the default
1924 scaling driver for the supported processors
1926 Use intel_pstate as a scaling driver, but configure it
1927 to work with generic cpufreq governors (instead of
1928 enabling its internal governor). This mode cannot be
1929 used along with the hardware-managed P-states (HWP)
1932 Enable intel_pstate on systems that prohibit it by default
1933 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1934 instead of acpi-cpufreq may disable platform features, such
1935 as thermal controls and power capping, that rely on ACPI
1936 P-States information being indicated to OSPM and therefore
1937 should be used with caution. This option does not work with
1938 processors that aren't supported by the intel_pstate driver
1939 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1941 Do not enable hardware P state control (HWP)
1944 Only load intel_pstate on systems which support
1945 hardware P state control (HWP) if available.
1947 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1948 Description Table, specifies preferred power management
1949 profile as "Enterprise Server" or "Performance Server",
1950 then this feature is turned on by default.
1952 Allow per-logical-CPU P-State performance control limits using
1953 cpufreq sysfs interface
1955 intremap= [X86-64, Intel-IOMMU]
1956 on enable Interrupt Remapping (default)
1957 off disable Interrupt Remapping
1958 nosid disable Source ID checking
1960 BIOS x2APIC opt-out request will be ignored
1961 nopost disable Interrupt Posting
1963 iomem= Disable strict checking of access to MMIO memory
1964 strict regions from userspace.
1979 nobypass [PPC/POWERNV]
1980 Disable IOMMU bypass, using IOMMU for PCI devices.
1982 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
1983 Format: { "0" | "1" }
1984 0 - Try to allocate a 32-bit DMA address first, before
1985 falling back to the full range if needed.
1986 1 - Allocate directly from the full usable range,
1987 forcing Dual Address Cycle for PCI cards supporting
1988 greater than 32-bit addressing.
1990 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1991 Format: { "0" | "1" }
1993 Request that DMA unmap operations use deferred
1994 invalidation of hardware TLBs, for increased
1995 throughput at the cost of reduced device isolation.
1996 Will fall back to strict mode if not supported by
1997 the relevant IOMMU driver.
1998 1 - Strict mode (default).
1999 DMA unmap operations invalidate IOMMU hardware TLBs
2003 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2004 Format: { "0" | "1" }
2005 0 - Use IOMMU translation for DMA.
2006 1 - Bypass the IOMMU for DMA.
2007 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2009 io7= [HW] IO7 for Marvel-based Alpha systems
2010 See comment before marvel_specify_io7 in
2011 arch/alpha/kernel/core_marvel.c.
2013 io_delay= [X86] I/O delay method
2015 Standard port 0x80 based delay
2017 Alternate port 0xed based delay (needed on some systems)
2019 Simple two microseconds delay
2024 See Documentation/admin-guide/nfs/nfsroot.rst.
2026 ipcmni_extend [KNL] Extend the maximum number of unique System V
2027 IPC identifiers from 32,768 to 16,777,216.
2029 irqaffinity= [SMP] Set the default irq affinity mask
2030 The argument is a cpu list, as described above.
2032 irqchip.gicv2_force_probe=
2035 Force the kernel to look for the second 4kB page
2036 of a GICv2 controller even if the memory range
2037 exposed by the device tree is too small.
2039 irqchip.gicv3_nolpi=
2041 Force the kernel to ignore the availability of
2042 LPIs (and by consequence ITSs). Intended for system
2043 that use the kernel as a bootloader, and thus want
2044 to let secondary kernels in charge of setting up
2047 irqchip.gicv3_pseudo_nmi= [ARM64]
2048 Enables support for pseudo-NMIs in the kernel. This
2049 requires the kernel to be built with
2050 CONFIG_ARM64_PSEUDO_NMI.
2053 When an interrupt is not handled search all handlers
2054 for it. Intended to get systems with badly broken
2058 When an interrupt is not handled search all handlers
2059 for it. Also check all handlers each timer
2060 interrupt. Intended to get systems with badly broken
2064 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2066 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2067 [Deprecated - use cpusets instead]
2068 Format: [flag-list,]<cpu-list>
2070 Specify one or more CPUs to isolate from disturbances
2071 specified in the flag list (default: domain):
2074 Disable the tick when a single task runs.
2076 A residual 1Hz tick is offloaded to workqueues, which you
2077 need to affine to housekeeping through the global
2078 workqueue's affinity configured via the
2079 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2080 by using the 'domain' flag described below.
2082 NOTE: by default the global workqueue runs on all CPUs,
2083 so to protect individual CPUs the 'cpumask' file has to
2084 be configured manually after bootup.
2087 Isolate from the general SMP balancing and scheduling
2088 algorithms. Note that performing domain isolation this way
2089 is irreversible: it's not possible to bring back a CPU to
2090 the domains once isolated through isolcpus. It's strongly
2091 advised to use cpusets instead to disable scheduler load
2092 balancing through the "cpuset.sched_load_balance" file.
2093 It offers a much more flexible interface where CPUs can
2094 move in and out of an isolated set anytime.
2096 You can move a process onto or off an "isolated" CPU via
2097 the CPU affinity syscalls or cpuset.
2098 <cpu number> begins at 0 and the maximum value is
2099 "number of CPUs in system - 1".
2103 Isolate from being targeted by managed interrupts
2104 which have an interrupt mask containing isolated
2105 CPUs. The affinity of managed interrupts is
2106 handled by the kernel and cannot be changed via
2107 the /proc/irq/* interfaces.
2109 This isolation is best effort and only effective
2110 if the automatically assigned interrupt mask of a
2111 device queue contains isolated and housekeeping
2112 CPUs. If housekeeping CPUs are online then such
2113 interrupts are directed to the housekeeping CPU
2114 so that IO submitted on the housekeeping CPU
2115 cannot disturb the isolated CPU.
2117 If a queue's affinity mask contains only isolated
2118 CPUs then this parameter has no effect on the
2119 interrupt routing decision, though interrupts are
2120 only delivered when tasks running on those
2121 isolated CPUs submit IO. IO submitted on
2122 housekeeping CPUs has no influence on those
2125 The format of <cpu-list> is described above.
2129 ivrs_ioapic [HW,X86-64]
2130 Provide an override to the IOAPIC-ID<->DEVICE-ID
2131 mapping provided in the IVRS ACPI table. For
2132 example, to map IOAPIC-ID decimal 10 to
2133 PCI device 00:14.0 write the parameter as:
2134 ivrs_ioapic[10]=00:14.0
2136 ivrs_hpet [HW,X86-64]
2137 Provide an override to the HPET-ID<->DEVICE-ID
2138 mapping provided in the IVRS ACPI table. For
2139 example, to map HPET-ID decimal 0 to
2140 PCI device 00:14.0 write the parameter as:
2141 ivrs_hpet[0]=00:14.0
2143 ivrs_acpihid [HW,X86-64]
2144 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2145 mapping provided in the IVRS ACPI table. For
2146 example, to map UART-HID:UID AMD0020:0 to
2147 PCI device 00:14.5 write the parameter as:
2148 ivrs_acpihid[00:14.5]=AMD0020:0
2150 js= [HW,JOY] Analog joystick
2151 See Documentation/input/joydev/joystick.rst.
2154 When CONFIG_RANDOMIZE_BASE is set, this disables
2155 kernel and module base offset ASLR (Address Space
2156 Layout Randomization).
2159 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2160 report on every invalid memory access. Without this
2161 parameter KASAN will print report only for the first
2166 kernelcore= [KNL,X86,IA-64,PPC]
2167 Format: nn[KMGTPE] | nn% | "mirror"
2168 This parameter specifies the amount of memory usable by
2169 the kernel for non-movable allocations. The requested
2170 amount is spread evenly throughout all nodes in the
2171 system as ZONE_NORMAL. The remaining memory is used for
2172 movable memory in its own zone, ZONE_MOVABLE. In the
2173 event, a node is too small to have both ZONE_NORMAL and
2174 ZONE_MOVABLE, kernelcore memory will take priority and
2175 other nodes will have a larger ZONE_MOVABLE.
2177 ZONE_MOVABLE is used for the allocation of pages that
2178 may be reclaimed or moved by the page migration
2179 subsystem. Note that allocations like PTEs-from-HighMem
2180 still use the HighMem zone if it exists, and the Normal
2181 zone if it does not.
2183 It is possible to specify the exact amount of memory in
2184 the form of "nn[KMGTPE]", a percentage of total system
2185 memory in the form of "nn%", or "mirror". If "mirror"
2186 option is specified, mirrored (reliable) memory is used
2187 for non-movable allocations and remaining memory is used
2188 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2189 are exclusive, so you cannot specify multiple forms.
2191 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2192 Format: <Controller#>[,poll interval]
2193 The controller # is the number of the ehci usb debug
2194 port as it is probed via PCI. The poll interval is
2195 optional and is the number seconds in between
2196 each poll cycle to the debug port in case you need
2197 the functionality for interrupting the kernel with
2198 gdb or control-c on the dbgp connection. When
2199 not using this parameter you use sysrq-g to break into
2200 the kernel debugger.
2202 kgdboc= [KGDB,HW] kgdb over consoles.
2203 Requires a tty driver that supports console polling,
2204 or a supported polling keyboard driver (non-usb).
2205 Serial only format: <serial_device>[,baud]
2206 keyboard only format: kbd
2207 keyboard and serial format: kbd,<serial_device>[,baud]
2208 Optional Kernel mode setting:
2209 kms, kbd format: kms,kbd
2210 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2212 kgdboc_earlycon= [KGDB,HW]
2213 If the boot console provides the ability to read
2214 characters and can work in polling mode, you can use
2215 this parameter to tell kgdb to use it as a backend
2216 until the normal console is registered. Intended to
2217 be used together with the kgdboc parameter which
2218 specifies the normal console to transition to.
2220 The name of the early console should be specified
2221 as the value of this parameter. Note that the name of
2222 the early console might be different than the tty
2223 name passed to kgdboc. It's OK to leave the value
2224 blank and the first boot console that implements
2225 read() will be picked.
2227 kgdbwait [KGDB] Stop kernel execution and enter the
2228 kernel debugger at the earliest opportunity.
2230 kmac= [MIPS] Korina ethernet MAC address.
2231 Configure the RouterBoard 532 series on-chip
2232 Ethernet adapter MAC address.
2234 kmemleak= [KNL] Boot-time kmemleak enable/disable
2235 Valid arguments: on, off
2237 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2240 kprobe_event=[probe-list]
2241 [FTRACE] Add kprobe events and enable at boot time.
2242 The probe-list is a semicolon delimited list of probe
2243 definitions. Each definition is same as kprobe_events
2244 interface, but the parameters are comma delimited.
2245 For example, to add a kprobe event on vfs_read with
2246 arg1 and arg2, add to the command line;
2248 kprobe_event=p,vfs_read,$arg1,$arg2
2250 See also Documentation/trace/kprobetrace.rst "Kernel
2251 Boot Parameter" section.
2253 kpti= [ARM64] Control page table isolation of user
2254 and kernel address spaces.
2255 Default: enabled on cores which need mitigation.
2259 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2260 Default is 0 (don't ignore, but inject #GP)
2262 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2263 Default is false (don't support).
2265 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2270 [KVM] Controls the software workaround for the
2271 X86_BUG_ITLB_MULTIHIT bug.
2272 force : Always deploy workaround.
2273 off : Never deploy workaround.
2274 auto : Deploy workaround based on the presence of
2275 X86_BUG_ITLB_MULTIHIT.
2279 If the software workaround is enabled for the host,
2280 guests do need not to enable it for nested guests.
2282 kvm.nx_huge_pages_recovery_ratio=
2283 [KVM] Controls how many 4KiB pages are periodically zapped
2284 back to huge pages. 0 disables the recovery, otherwise if
2285 the value is N KVM will zap 1/Nth of the 4KiB pages every
2286 minute. The default is 60.
2288 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2289 Default is 1 (enabled)
2291 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2293 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2296 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2298 nvhe: Standard nVHE-based mode, without support for
2301 protected: nVHE-based mode with support for guests whose
2302 state is kept private from the host.
2303 Not valid if the kernel is running in EL2.
2305 Defaults to VHE/nVHE based on hardware support.
2307 kvm-arm.vgic_v3_group0_trap=
2308 [KVM,ARM] Trap guest accesses to GICv3 group-0
2311 kvm-arm.vgic_v3_group1_trap=
2312 [KVM,ARM] Trap guest accesses to GICv3 group-1
2315 kvm-arm.vgic_v3_common_trap=
2316 [KVM,ARM] Trap guest accesses to GICv3 common
2319 kvm-arm.vgic_v4_enable=
2320 [KVM,ARM] Allow use of GICv4 for direct injection of
2323 kvm_cma_resv_ratio=n [PPC]
2324 Reserves given percentage from system memory area for
2325 contiguous memory allocation for KVM hash pagetable
2327 By default it reserves 5% of total system memory.
2331 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2332 (virtualized MMU) support on capable Intel chips.
2333 Default is 1 (enabled)
2335 kvm-intel.emulate_invalid_guest_state=
2336 [KVM,Intel] Enable emulation of invalid guest states
2337 Default is 0 (disabled)
2339 kvm-intel.flexpriority=
2340 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2341 Default is 1 (enabled)
2344 [KVM,Intel] Enable VMX nesting (nVMX).
2345 Default is 0 (disabled)
2347 kvm-intel.unrestricted_guest=
2348 [KVM,Intel] Disable unrestricted guest feature
2349 (virtualized real and unpaged mode) on capable
2350 Intel chips. Default is 1 (enabled)
2352 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2355 Valid arguments: never, cond, always
2357 always: L1D cache flush on every VMENTER.
2358 cond: Flush L1D on VMENTER only when the code between
2359 VMEXIT and VMENTER can leak host memory.
2360 never: Disables the mitigation
2362 Default is cond (do L1 cache flush in specific instances)
2364 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2365 feature (tagged TLBs) on capable Intel chips.
2366 Default is 1 (enabled)
2368 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2371 The kernel PTE inversion protection is unconditionally
2372 enabled and cannot be disabled.
2375 Provides all available mitigations for the
2376 L1TF vulnerability. Disables SMT and
2377 enables all mitigations in the
2378 hypervisors, i.e. unconditional L1D flush.
2380 SMT control and L1D flush control via the
2381 sysfs interface is still possible after
2382 boot. Hypervisors will issue a warning
2383 when the first VM is started in a
2384 potentially insecure configuration,
2385 i.e. SMT enabled or L1D flush disabled.
2388 Same as 'full', but disables SMT and L1D
2389 flush runtime control. Implies the
2390 'nosmt=force' command line option.
2391 (i.e. sysfs control of SMT is disabled.)
2394 Leaves SMT enabled and enables the default
2395 hypervisor mitigation, i.e. conditional
2398 SMT control and L1D flush control via the
2399 sysfs interface is still possible after
2400 boot. Hypervisors will issue a warning
2401 when the first VM is started in a
2402 potentially insecure configuration,
2403 i.e. SMT enabled or L1D flush disabled.
2407 Disables SMT and enables the default
2408 hypervisor mitigation.
2410 SMT control and L1D flush control via the
2411 sysfs interface is still possible after
2412 boot. Hypervisors will issue a warning
2413 when the first VM is started in a
2414 potentially insecure configuration,
2415 i.e. SMT enabled or L1D flush disabled.
2418 Same as 'flush', but hypervisors will not
2419 warn when a VM is started in a potentially
2420 insecure configuration.
2423 Disables hypervisor mitigations and doesn't
2425 It also drops the swap size and available
2426 RAM limit restriction on both hypervisor and
2431 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2437 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2440 lapic= [X86,APIC] Do not use TSC deadline
2441 value for LAPIC timer one-shot implementation. Default
2442 back to the programmable timer unit in the LAPIC.
2443 Format: notscdeadline
2445 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2448 libata.dma= [LIBATA] DMA control
2449 libata.dma=0 Disable all PATA and SATA DMA
2450 libata.dma=1 PATA and SATA Disk DMA only
2451 libata.dma=2 ATAPI (CDROM) DMA only
2452 libata.dma=4 Compact Flash DMA only
2453 Combinations also work, so libata.dma=3 enables DMA
2454 for disks and CDROMs, but not CFs.
2456 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2457 libata.ignore_hpa=0 keep BIOS limits (default)
2458 libata.ignore_hpa=1 ignore limits, using full disk
2460 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2464 libata.force= [LIBATA] Force configurations. The format is comma-
2465 separated list of "[ID:]VAL" where ID is
2466 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2467 matching port, link or device. Basically, it matches
2468 the ATA ID string printed on console by libata. If
2469 the whole ID part is omitted, the last PORT and DEVICE
2470 values are used. If ID hasn't been specified yet, the
2471 configuration applies to all ports, links and devices.
2473 If only DEVICE is omitted, the parameter applies to
2474 the port and all links and devices behind it. DEVICE
2475 number of 0 either selects the first device or the
2476 first fan-out link behind PMP device. It does not
2477 select the host link. DEVICE number of 15 selects the
2478 host link and device attached to it.
2480 The VAL specifies the configuration to force. As long
2481 as there's no ambiguity shortcut notation is allowed.
2482 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2483 The following configurations can be forced.
2485 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2486 Any ID with matching PORT is used.
2488 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2490 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2491 udma[/][16,25,33,44,66,100,133] notation is also
2494 * [no]ncq: Turn on or off NCQ.
2496 * [no]ncqtrim: Turn off queued DSM TRIM.
2498 * nohrst, nosrst, norst: suppress hard, soft
2501 * rstonce: only attempt one reset during
2502 hot-unplug link recovery
2504 * dump_id: dump IDENTIFY data.
2506 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2508 * disable: Disable this device.
2510 If there are multiple matching configurations changing
2511 the same attribute, the last one is used.
2513 memblock=debug [KNL] Enable memblock debug messages.
2515 load_ramdisk= [RAM] [Deprecated]
2517 lockd.nlm_grace_period=P [NFS] Assign grace period.
2520 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2523 lockd.nlm_timeout=T [NFS] Assign timeout value.
2526 lockd.nlm_udpport=M [NFS] Assign UDP port.
2529 lockdown= [SECURITY]
2530 { integrity | confidentiality }
2531 Enable the kernel lockdown feature. If set to
2532 integrity, kernel features that allow userland to
2533 modify the running kernel are disabled. If set to
2534 confidentiality, kernel features that allow userland
2535 to extract confidential information from the kernel
2538 locktorture.nreaders_stress= [KNL]
2539 Set the number of locking read-acquisition kthreads.
2540 Defaults to being automatically set based on the
2541 number of online CPUs.
2543 locktorture.nwriters_stress= [KNL]
2544 Set the number of locking write-acquisition kthreads.
2546 locktorture.onoff_holdoff= [KNL]
2547 Set time (s) after boot for CPU-hotplug testing.
2549 locktorture.onoff_interval= [KNL]
2550 Set time (s) between CPU-hotplug operations, or
2551 zero to disable CPU-hotplug testing.
2553 locktorture.shuffle_interval= [KNL]
2554 Set task-shuffle interval (jiffies). Shuffling
2555 tasks allows some CPUs to go into dyntick-idle
2556 mode during the locktorture test.
2558 locktorture.shutdown_secs= [KNL]
2559 Set time (s) after boot system shutdown. This
2560 is useful for hands-off automated testing.
2562 locktorture.stat_interval= [KNL]
2563 Time (s) between statistics printk()s.
2565 locktorture.stutter= [KNL]
2566 Time (s) to stutter testing, for example,
2567 specifying five seconds causes the test to run for
2568 five seconds, wait for five seconds, and so on.
2569 This tests the locking primitive's ability to
2570 transition abruptly to and from idle.
2572 locktorture.torture_type= [KNL]
2573 Specify the locking implementation to test.
2575 locktorture.verbose= [KNL]
2576 Enable additional printk() statements.
2578 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2581 loglevel= All Kernel Messages with a loglevel smaller than the
2582 console loglevel will be printed to the console. It can
2583 also be changed with klogd or other programs. The
2584 loglevels are defined as follows:
2586 0 (KERN_EMERG) system is unusable
2587 1 (KERN_ALERT) action must be taken immediately
2588 2 (KERN_CRIT) critical conditions
2589 3 (KERN_ERR) error conditions
2590 4 (KERN_WARNING) warning conditions
2591 5 (KERN_NOTICE) normal but significant condition
2592 6 (KERN_INFO) informational
2593 7 (KERN_DEBUG) debug-level messages
2595 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2596 in bytes. n must be a power of two and greater
2597 than the minimal size. The minimal size is defined
2598 by LOG_BUF_SHIFT kernel config parameter. There is
2599 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2600 that allows to increase the default size depending on
2601 the number of CPUs. See init/Kconfig for more details.
2603 logo.nologo [FB] Disables display of the built-in Linux logo.
2604 This may be used to provide more screen space for
2605 kernel log messages and is useful when debugging
2606 kernel boot problems.
2608 lp=0 [LP] Specify parallel ports to use, e.g,
2609 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2610 lp=reset first parallel port). 'lp=0' disables the
2611 lp=auto printer driver. 'lp=reset' (which can be
2612 specified in addition to the ports) causes
2613 attached printers to be reset. Using
2614 lp=port1,port2,... specifies the parallel ports
2615 to associate lp devices with, starting with
2616 lp0. A port specification may be 'none' to skip
2617 that lp device, or a parport name such as
2618 'parport0'. Specifying 'lp=auto' instead of a
2619 port specification list means that device IDs
2620 from each port should be examined, to see if
2621 an IEEE 1284-compliant printer is attached; if
2622 so, the driver will manage that printer.
2623 See also header of drivers/char/lp.c.
2626 Sets loops_per_jiffy to given constant, thus avoiding
2627 time-consuming boot-time autodetection (up to 250 ms per
2628 CPU). 0 enables autodetection (default). To determine
2629 the correct value for your kernel, boot with normal
2630 autodetection and see what value is printed. Note that
2631 on SMP systems the preset will be applied to all CPUs,
2632 which is likely to cause problems if your CPUs need
2633 significantly divergent settings. An incorrect value
2634 will cause delays in the kernel to be wrong, leading to
2635 unpredictable I/O errors and other breakage. Although
2636 unlikely, in the extreme case this might damage your
2640 Format: <io>,<irq>,<dma>
2642 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2645 [SECURITY] Choose order of LSM initialization. This
2646 overrides CONFIG_LSM, and the "security=" parameter.
2648 machvec= [IA-64] Force the use of a particular machine-vector
2649 (machvec) in a generic kernel.
2650 Example: machvec=hpzx1
2652 machtype= [Loongson] Share the same kernel image file between
2653 different yeeloong laptops.
2654 Example: machtype=lemote-yeeloong-2f-7inch
2656 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2657 than or equal to this physical address is ignored.
2659 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2660 will bring up during bootup. maxcpus=n : n >= 0 limits
2661 the kernel to bring up 'n' processors. Surely after
2662 bootup you can bring up the other plugged cpu by executing
2663 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2664 only takes effect during system bootup.
2665 While n=0 is a special case, it is equivalent to "nosmp",
2666 which also disables the IO APIC.
2668 max_loop= [LOOP] The number of loop block devices that get
2669 (loop.max_loop) unconditionally pre-created at init time. The default
2670 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2671 of statically allocating a predefined number, loop
2672 devices can be requested on-demand with the
2673 /dev/loop-control interface.
2675 mce [X86-32] Machine Check Exception
2677 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2679 md= [HW] RAID subsystems devices and level
2680 See Documentation/admin-guide/md.rst.
2683 Format: <first>,<last>
2684 Specifies range of consoles to be captured by the MDA.
2687 Control mitigation for the Micro-architectural Data
2688 Sampling (MDS) vulnerability.
2690 Certain CPUs are vulnerable to an exploit against CPU
2691 internal buffers which can forward information to a
2692 disclosure gadget under certain conditions.
2694 In vulnerable processors, the speculatively
2695 forwarded data can be used in a cache side channel
2696 attack, to access data to which the attacker does
2697 not have direct access.
2699 This parameter controls the MDS mitigation. The
2702 full - Enable MDS mitigation on vulnerable CPUs
2703 full,nosmt - Enable MDS mitigation and disable
2704 SMT on vulnerable CPUs
2705 off - Unconditionally disable MDS mitigation
2707 On TAA-affected machines, mds=off can be prevented by
2708 an active TAA mitigation as both vulnerabilities are
2709 mitigated with the same mechanism so in order to disable
2710 this mitigation, you need to specify tsx_async_abort=off
2713 Not specifying this option is equivalent to
2716 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2718 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2719 Amount of memory to be used in cases as follows:
2722 2 when the kernel is not able to see the whole system memory;
2723 3 memory that lies after 'mem=' boundary is excluded from
2724 the hypervisor, then assigned to KVM guests.
2726 [X86] Work as limiting max address. Use together
2727 with memmap= to avoid physical address space collisions.
2728 Without memmap= PCI devices could be placed at addresses
2729 belonging to unused RAM.
2731 Note that this only takes effects during boot time since
2732 in above case 3, memory may need be hot added after boot
2733 if system memory of hypervisor is not sufficient.
2735 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2739 [KNL,SH] Allow user to override the default size for
2740 per-device physically contiguous DMA buffers.
2742 memhp_default_state=online/offline
2743 [KNL] Set the initial state for the memory hotplug
2744 onlining policy. If not specified, the default value is
2745 set according to the
2746 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2748 See Documentation/admin-guide/mm/memory-hotplug.rst.
2750 memmap=exactmap [KNL,X86] Enable setting of an exact
2751 E820 memory map, as specified by the user.
2752 Such memmap=exactmap lines can be constructed based on
2753 BIOS output or other requirements. See the memmap=nn@ss
2756 memmap=nn[KMG]@ss[KMG]
2757 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
2758 Region of memory to be used is from ss to ss+nn.
2759 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2760 which limits max address to nn[KMG].
2761 Multiple different regions can be specified,
2764 memmap=100M@2G,100M#3G,1G!1024G
2766 memmap=nn[KMG]#ss[KMG]
2767 [KNL,ACPI] Mark specific memory as ACPI data.
2768 Region of memory to be marked is from ss to ss+nn.
2770 memmap=nn[KMG]$ss[KMG]
2771 [KNL,ACPI] Mark specific memory as reserved.
2772 Region of memory to be reserved is from ss to ss+nn.
2773 Example: Exclude memory from 0x18690000-0x1869ffff
2774 memmap=64K$0x18690000
2776 memmap=0x10000$0x18690000
2777 Some bootloaders may need an escape character before '$',
2778 like Grub2, otherwise '$' and the following number
2781 memmap=nn[KMG]!ss[KMG]
2782 [KNL,X86] Mark specific memory as protected.
2783 Region of memory to be used, from ss to ss+nn.
2784 The memory region may be marked as e820 type 12 (0xc)
2785 and is NVDIMM or ADR memory.
2787 memmap=<size>%<offset>-<oldtype>+<newtype>
2788 [KNL,ACPI] Convert memory within the specified region
2789 from <oldtype> to <newtype>. If "-<oldtype>" is left
2790 out, the whole region will be marked as <newtype>,
2791 even if previously unavailable. If "+<newtype>" is left
2792 out, matching memory will be removed. Types are
2793 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2794 3 = ACPI, 12 = PRAM.
2796 memory_corruption_check=0/1 [X86]
2797 Some BIOSes seem to corrupt the first 64k of
2798 memory when doing things like suspend/resume.
2799 Setting this option will scan the memory
2800 looking for corruption. Enabling this will
2801 both detect corruption and prevent the kernel
2802 from using the memory being corrupted.
2803 However, its intended as a diagnostic tool; if
2804 repeatable BIOS-originated corruption always
2805 affects the same memory, you can use memmap=
2806 to prevent the kernel from using that memory.
2808 memory_corruption_check_size=size [X86]
2809 By default it checks for corruption in the low
2810 64k, making this memory unavailable for normal
2811 use. Use this parameter to scan for
2812 corruption in more or less memory.
2814 memory_corruption_check_period=seconds [X86]
2815 By default it checks for corruption every 60
2816 seconds. Use this parameter to check at some
2817 other rate. 0 disables periodic checking.
2819 memory_hotplug.memmap_on_memory
2820 [KNL,X86,ARM] Boolean flag to enable this feature.
2821 Format: {on | off (default)}
2822 When enabled, runtime hotplugged memory will
2823 allocate its internal metadata (struct pages)
2824 from the hotadded memory which will allow to
2825 hotadd a lot of memory without requiring
2826 additional memory to do so.
2827 This feature is disabled by default because it
2828 has some implication on large (e.g. GB)
2829 allocations in some configurations (e.g. small
2831 The state of the flag can be read in
2832 /sys/module/memory_hotplug/parameters/memmap_on_memory.
2833 Note that even when enabled, there are a few cases where
2834 the feature is not effective.
2836 memtest= [KNL,X86,ARM,PPC,RISCV] Enable memtest
2838 default : 0 <disable>
2839 Specifies the number of memtest passes to be
2840 performed. Each pass selects another test
2841 pattern from a given set of patterns. Memtest
2842 fills the memory with this pattern, validates
2843 memory contents and reserves bad memory
2844 regions that are detected.
2846 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2847 Valid arguments: on, off
2848 Default (depends on kernel configuration option):
2849 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2850 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2851 mem_encrypt=on: Activate SME
2852 mem_encrypt=off: Do not activate SME
2854 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2855 for details on when memory encryption can be activated.
2857 mem_sleep_default= [SUSPEND] Default system suspend mode:
2858 s2idle - Suspend-To-Idle
2859 shallow - Power-On Suspend or equivalent (if supported)
2860 deep - Suspend-To-RAM or equivalent (if supported)
2861 See Documentation/admin-guide/pm/sleep-states.rst.
2863 meye.*= [HW] Set MotionEye Camera parameters
2864 See Documentation/admin-guide/media/meye.rst.
2866 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2867 Multi-Function General Purpose Timers on AMD Geode
2870 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2871 the BIOS has incorrectly applied a workaround. TinyBIOS
2872 version 0.98 is known to be affected, 0.99 fixes the
2873 problem by letting the user disable the workaround.
2877 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2878 physical address is ignored.
2880 mini2440= [ARM,HW,KNL]
2881 Format:[0..2][b][c][t]
2883 MINI2440 configuration specification:
2884 0 - The attached screen is the 3.5" TFT
2885 1 - The attached screen is the 7" TFT
2886 2 - The VGA Shield is attached (1024x768)
2887 Leaving out the screen size parameter will not load
2888 the TFT driver, and the framebuffer will be left
2890 b - Enable backlight. The TFT backlight pin will be
2891 linked to the kernel VESA blanking code and a GPIO
2892 LED. This parameter is not necessary when using the
2894 c - Enable the s3c camera interface.
2895 t - Reserved for enabling touchscreen support. The
2896 touchscreen support is not enabled in the mainstream
2897 kernel as of 2.6.30, a preliminary port can be found
2898 in the "bleeding edge" mini2440 support kernel at
2899 https://repo.or.cz/w/linux-2.6/mini2440.git
2902 [X86,PPC,S390,ARM64] Control optional mitigations for
2903 CPU vulnerabilities. This is a set of curated,
2904 arch-independent options, each of which is an
2905 aggregation of existing arch-specific options.
2908 Disable all optional CPU mitigations. This
2909 improves system performance, but it may also
2910 expose users to several CPU vulnerabilities.
2911 Equivalent to: nopti [X86,PPC]
2913 nospectre_v1 [X86,PPC]
2915 nospectre_v2 [X86,PPC,S390,ARM64]
2916 spectre_v2_user=off [X86]
2917 spec_store_bypass_disable=off [X86,PPC]
2918 ssbd=force-off [ARM64]
2921 tsx_async_abort=off [X86]
2922 kvm.nx_huge_pages=off [X86]
2923 no_entry_flush [PPC]
2924 no_uaccess_flush [PPC]
2927 This does not have any effect on
2928 kvm.nx_huge_pages when
2929 kvm.nx_huge_pages=force.
2932 Mitigate all CPU vulnerabilities, but leave SMT
2933 enabled, even if it's vulnerable. This is for
2934 users who don't want to be surprised by SMT
2935 getting disabled across kernel upgrades, or who
2936 have other ways of avoiding SMT-based attacks.
2937 Equivalent to: (default behavior)
2940 Mitigate all CPU vulnerabilities, disabling SMT
2941 if needed. This is for users who always want to
2942 be fully mitigated, even if it means losing SMT.
2943 Equivalent to: l1tf=flush,nosmt [X86]
2944 mds=full,nosmt [X86]
2945 tsx_async_abort=full,nosmt [X86]
2948 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2949 parameter allows control of the logging verbosity for
2950 the additional memory initialisation checks. A value
2951 of 0 disables mminit logging and a level of 4 will
2952 log everything. Information is printed at KERN_DEBUG
2953 so loglevel=8 may also need to be specified.
2956 [KNL] When CONFIG_MODULE_SIG is set, this means that
2957 modules without (valid) signatures will fail to load.
2958 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2959 is always true, so this option does nothing.
2961 module_blacklist= [KNL] Do not load a comma-separated list of
2962 modules. Useful for debugging problem modules.
2965 [MOUSE] Maximum time between finger touching and
2966 leaving touchpad surface for touch to be considered
2967 a tap and be reported as a left button click (for
2968 touchpads working in absolute mode only).
2970 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2971 reporting absolute coordinates, such as tablets
2972 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2973 reporting absolute coordinates, such as tablets
2975 movablecore= [KNL,X86,IA-64,PPC]
2976 Format: nn[KMGTPE] | nn%
2977 This parameter is the complement to kernelcore=, it
2978 specifies the amount of memory used for migratable
2979 allocations. If both kernelcore and movablecore is
2980 specified, then kernelcore will be at *least* the
2981 specified value but may be more. If movablecore on its
2982 own is specified, the administrator must be careful
2983 that the amount of memory usable for all allocations
2986 movable_node [KNL] Boot-time switch to make hotplugable memory
2987 NUMA nodes to be movable. This means that the memory
2988 of such nodes will be usable only for movable
2989 allocations which rules out almost all kernel
2990 allocations. Use with caution!
2992 MTD_Partition= [MTD]
2993 Format: <name>,<region-number>,<size>,<offset>
2995 MTD_Region= [MTD] Format:
2996 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2999 See drivers/mtd/parsers/cmdlinepart.c
3001 multitce=off [PPC] This parameter disables the use of the pSeries
3002 firmware feature for updating multiple TCE entries
3005 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3007 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3009 boundary - index of last SLC block on Flex-OneNAND.
3010 The remaining blocks are configured as MLC blocks.
3011 lock - Configure if Flex-OneNAND boundary should be locked.
3012 Once locked, the boundary cannot be changed.
3013 1 indicates lock status, 0 indicates unlock status.
3016 ARM/S3C2412 JIVE boot control
3018 See arch/arm/mach-s3c/mach-jive.c
3020 mtouchusb.raw_coordinates=
3021 [HW] Make the MicroTouch USB driver use raw coordinates
3022 ('y', default) or cooked coordinates ('n')
3024 mtrr_chunk_size=nn[KMG] [X86]
3025 used for mtrr cleanup. It is largest continuous chunk
3026 that could hold holes aka. UC entries.
3028 mtrr_gran_size=nn[KMG] [X86]
3029 Used for mtrr cleanup. It is granularity of mtrr block.
3031 Large value could prevent small alignment from
3034 mtrr_spare_reg_nr=n [X86]
3036 Range: 0,7 : spare reg number
3038 Used for mtrr cleanup. It is spare mtrr entries number.
3039 Set to 2 or more if your graphical card needs more.
3041 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3043 netdev= [NET] Network devices parameters
3044 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3045 Note that mem_start is often overloaded to mean
3046 something different and driver-specific.
3047 This usage is only documented in each driver source
3051 [NETFILTER] Enable connection tracking flow accounting
3052 0 to disable accounting
3053 1 to enable accounting
3056 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3057 See Documentation/admin-guide/nfs/nfsroot.rst.
3059 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3060 See Documentation/admin-guide/nfs/nfsroot.rst.
3062 nfsrootdebug [NFS] enable nfsroot debugging messages.
3063 See Documentation/admin-guide/nfs/nfsroot.rst.
3065 nfs.callback_nr_threads=
3066 [NFSv4] set the total number of threads that the
3067 NFS client will assign to service NFSv4 callback
3070 nfs.callback_tcpport=
3071 [NFS] set the TCP port on which the NFSv4 callback
3072 channel should listen.
3075 [NFS] sets the pathname to the program which is used
3076 to update the NFS client cache entries.
3078 nfs.cache_getent_timeout=
3079 [NFS] sets the timeout after which an attempt to
3080 update a cache entry is deemed to have failed.
3082 nfs.idmap_cache_timeout=
3083 [NFS] set the maximum lifetime for idmapper cache
3087 [NFS] enable 64-bit inode numbers.
3088 If zero, the NFS client will fake up a 32-bit inode
3089 number for the readdir() and stat() syscalls instead
3090 of returning the full 64-bit number.
3091 The default is to return 64-bit inode numbers.
3093 nfs.max_session_cb_slots=
3094 [NFSv4.1] Sets the maximum number of session
3095 slots the client will assign to the callback
3096 channel. This determines the maximum number of
3097 callbacks the client will process in parallel for
3098 a particular server.
3100 nfs.max_session_slots=
3101 [NFSv4.1] Sets the maximum number of session slots
3102 the client will attempt to negotiate with the server.
3103 This limits the number of simultaneous RPC requests
3104 that the client can send to the NFSv4.1 server.
3105 Note that there is little point in setting this
3106 value higher than the max_tcp_slot_table_limit.
3108 nfs.nfs4_disable_idmapping=
3109 [NFSv4] When set to the default of '1', this option
3110 ensures that both the RPC level authentication
3111 scheme and the NFS level operations agree to use
3112 numeric uids/gids if the mount is using the
3113 'sec=sys' security flavour. In effect it is
3114 disabling idmapping, which can make migration from
3115 legacy NFSv2/v3 systems to NFSv4 easier.
3116 Servers that do not support this mode of operation
3117 will be autodetected by the client, and it will fall
3118 back to using the idmapper.
3119 To turn off this behaviour, set the value to '0'.
3121 [NFS4] Specify an additional fixed unique ident-
3122 ification string that NFSv4 clients can insert into
3123 their nfs_client_id4 string. This is typically a
3124 UUID that is generated at system install time.
3126 nfs.send_implementation_id =
3127 [NFSv4.1] Send client implementation identification
3128 information in exchange_id requests.
3129 If zero, no implementation identification information
3131 The default is to send the implementation identification
3134 nfs.recover_lost_locks =
3135 [NFSv4] Attempt to recover locks that were lost due
3136 to a lease timeout on the server. Please note that
3137 doing this risks data corruption, since there are
3138 no guarantees that the file will remain unchanged
3139 after the locks are lost.
3140 If you want to enable the kernel legacy behaviour of
3141 attempting to recover these locks, then set this
3143 The default parameter value of '0' causes the kernel
3144 not to attempt recovery of lost locks.
3146 nfs4.layoutstats_timer =
3147 [NFSv4.2] Change the rate at which the kernel sends
3148 layoutstats to the pNFS metadata server.
3150 Setting this to value to 0 causes the kernel to use
3151 whatever value is the default set by the layout
3152 driver. A non-zero value sets the minimum interval
3153 in seconds between layoutstats transmissions.
3155 nfsd.nfs4_disable_idmapping=
3156 [NFSv4] When set to the default of '1', the NFSv4
3157 server will return only numeric uids and gids to
3158 clients using auth_sys, and will accept numeric uids
3159 and gids from such clients. This is intended to ease
3160 migration from NFSv2/v3.
3162 nmi_backtrace.backtrace_idle [KNL]
3163 Dump stacks even of idle CPUs in response to an
3164 NMI stack-backtrace request.
3166 nmi_debug= [KNL,SH] Specify one or more actions to take
3167 when a NMI is triggered.
3168 Format: [state][,regs][,debounce][,die]
3170 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3171 Format: [panic,][nopanic,][num]
3173 0 - turn hardlockup detector in nmi_watchdog off
3174 1 - turn hardlockup detector in nmi_watchdog on
3175 When panic is specified, panic when an NMI watchdog
3176 timeout occurs (or 'nopanic' to not panic on an NMI
3177 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3178 To disable both hard and soft lockup detectors,
3179 please see 'nowatchdog'.
3180 This is useful when you use a panic=... timeout and
3181 need the box quickly up again.
3183 These settings can be accessed at runtime via
3184 the nmi_watchdog and hardlockup_panic sysctls.
3186 netpoll.carrier_timeout=
3187 [NET] Specifies amount of time (in seconds) that
3188 netpoll should wait for a carrier. By default netpoll
3191 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3192 emulation library even if a 387 maths coprocessor
3195 no5lvl [X86-64] Disable 5-level paging mode. Forces
3196 kernel to use 4-level paging instead.
3198 nofsgsbase [X86] Disables FSGSBASE instructions.
3201 [HW] Never suspend the console
3202 Disable suspending of consoles during suspend and
3203 hibernate operations. Once disabled, debugging
3204 messages can reach various consoles while the rest
3205 of the system is being put to sleep (ie, while
3206 debugging driver suspend/resume hooks). This may
3207 not work reliably with all consoles, but is known
3208 to work with serial and VGA consoles.
3209 To facilitate more flexible debugging, we also add
3210 console_suspend, a printk module parameter to control
3211 it. Users could use console_suspend (usually
3212 /sys/module/printk/parameters/console_suspend) to
3213 turn on/off it dynamically.
3215 novmcoredd [KNL,KDUMP]
3216 Disable device dump. Device dump allows drivers to
3217 append dump data to vmcore so you can collect driver
3218 specified debug info. Drivers can append the data
3219 without any limit and this data is stored in memory,
3220 so this may cause significant memory stress. Disabling
3221 device dump can help save memory but the driver debug
3222 data will be no longer available. This parameter
3223 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3226 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3227 caches in the slab allocator. Saves per-node memory,
3228 but will impact performance.
3232 noaltinstr [S390] Disables alternative instructions patching
3233 (CPU alternatives feature).
3235 noapic [SMP,APIC] Tells the kernel to not make use of any
3236 IOAPICs that may be present in the system.
3238 noautogroup Disable scheduler automatic task group creation.
3240 nobats [PPC] Do not use BATs for mapping kernel lowmem
3241 on "Classic" PPC cores.
3245 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3247 nodelayacct [KNL] Disable per-task delay accounting
3249 nodsp [SH] Disable hardware DSP at boot time.
3251 noefi Disable EFI runtime services support.
3253 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3258 On X86-32 available only on PAE configured kernels.
3259 noexec=on: enable non-executable mappings (default)
3260 noexec=off: disable non-executable mappings
3263 Disable SMAP (Supervisor Mode Access Prevention)
3264 even if it is supported by processor.
3267 Disable SMEP (Supervisor Mode Execution Prevention)
3268 even if it is supported by processor.
3271 This affects only 32-bit executables.
3272 noexec32=on: enable non-executable mappings (default)
3273 read doesn't imply executable mappings
3274 noexec32=off: disable non-executable mappings
3275 read implies executable mappings
3277 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3279 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3280 register save and restore. The kernel will only save
3281 legacy floating-point registers on task switch.
3283 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3285 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3287 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3288 Equivalent to smt=1.
3290 [KNL,X86] Disable symmetric multithreading (SMT).
3291 nosmt=force: Force disable SMT, cannot be undone
3292 via the sysfs control file.
3294 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3295 (bounds check bypass). With this option data leaks are
3296 possible in the system.
3298 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3299 the Spectre variant 2 (indirect branch prediction)
3300 vulnerability. System may allow data leaks with this
3303 nospec_store_bypass_disable
3304 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3307 [PPC] Don't flush the L1-D cache after accessing user data.
3309 noxsave [BUGS=X86] Disables x86 extended register state save
3310 and restore using xsave. The kernel will fallback to
3311 enabling legacy floating-point and sse state.
3313 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3314 register states. The kernel will fall back to use
3315 xsave to save the states. By using this parameter,
3316 performance of saving the states is degraded because
3317 xsave doesn't support modified optimization while
3318 xsaveopt supports it on xsaveopt enabled systems.
3320 noxsaves [X86] Disables xsaves and xrstors used in saving and
3321 restoring x86 extended register state in compacted
3322 form of xsave area. The kernel will fall back to use
3323 xsaveopt and xrstor to save and restore the states
3324 in standard form of xsave area. By using this
3325 parameter, xsave area per process might occupy more
3326 memory on xsaves enabled systems.
3328 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3329 in do_idle() and not use the arch_cpu_idle()
3330 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3331 to be effective. This is useful on platforms where the
3332 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3333 correctly or when doing power measurements to evalute
3334 the impact of the sleep instructions. This is also
3335 useful when using JTAG debugger.
3337 no_file_caps Tells the kernel not to honor file capabilities. The
3338 only way then for a file to be executed with privilege
3339 is to be setuid root or executed by root.
3341 nohalt [IA-64] Tells the kernel not to use the power saving
3342 function PAL_HALT_LIGHT when idle. This increases
3343 power-consumption. On the positive side, it reduces
3344 interrupt wake-up latency, which may improve performance
3345 in certain environments such as networked servers or
3349 Force pointers printed to the console or buffers to be
3350 unhashed. By default, when a pointer is printed via %p
3351 format string, that pointer is "hashed", i.e. obscured
3352 by hashing the pointer value. This is a security feature
3353 that hides actual kernel addresses from unprivileged
3354 users, but it also makes debugging the kernel more
3355 difficult since unequal pointers can no longer be
3356 compared. However, if this command-line option is
3357 specified, then all normal pointers will have their true
3358 value printed. Pointers printed via %pK may still be
3359 hashed. This option should only be specified when
3360 debugging the kernel. Please do not use on production
3363 nohibernate [HIBERNATION] Disable hibernation and resume.
3365 nohz= [KNL] Boottime enable/disable dynamic ticks
3366 Valid arguments: on, off
3369 nohz_full= [KNL,BOOT,SMP,ISOL]
3370 The argument is a cpu list, as described above.
3371 In kernels built with CONFIG_NO_HZ_FULL=y, set
3372 the specified list of CPUs whose tick will be stopped
3373 whenever possible. The boot CPU will be forced outside
3374 the range to maintain the timekeeping. Any CPUs
3375 in this list will have their RCU callbacks offloaded,
3376 just as if they had also been called out in the
3377 rcu_nocbs= boot parameter.
3379 noiotrap [SH] Disables trapped I/O port accesses.
3381 noirqdebug [X86-32] Disables the code which attempts to detect and
3382 disable unhandled interrupt sources.
3384 no_timer_check [X86,APIC] Disables the code which tests for
3385 broken timer IRQ sources.
3387 noisapnp [ISAPNP] Disables ISA PnP code.
3389 noinitrd [RAM] Tells the kernel not to load any configured
3392 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3394 [Deprecated - use intremap=off]
3398 noinvpcid [X86] Disable the INVPCID cpu feature.
3400 nojitter [IA-64] Disables jitter checking for ITC timers.
3402 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3404 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3408 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3409 clock and use the default one.
3411 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3412 accounting. steal time is computed, but won't
3413 influence scheduler behaviour
3415 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3417 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3419 noltlbs [PPC] Do not use large page/tlb entries for kernel
3420 lowmem mapping on PPC40x and PPC8xx
3422 nomca [IA-64] Disable machine check abort handling
3424 nomce [X86-32] Disable Machine Check Exception
3426 nomfgpt [X86-32] Disable Multi-Function General Purpose
3427 Timer usage (for AMD Geode machines).
3429 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3430 shutdown the other cpus. Instead use the REBOOT_VECTOR
3433 nomodule Disable module load
3435 nopat [X86] Disable PAT (page attribute table extension of
3436 pagetables) support.
3438 nopcid [X86-64] Disable the PCID cpu feature.
3440 norandmaps Don't use address space randomization. Equivalent to
3441 echo 0 > /proc/sys/kernel/randomize_va_space
3443 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3444 with UP alternatives
3446 nordrand [X86] Disable kernel use of the RDRAND and
3447 RDSEED instructions even if they are supported
3448 by the processor. RDRAND and RDSEED are still
3449 available to user space applications.
3451 noresume [SWSUSP] Disables resume and restores original swap
3454 no-scroll [VGA] Disables scrollback.
3455 This is required for the Braillex ib80-piezo Braille
3456 reader made by F.H. Papenmeier (Germany).
3460 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3462 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3464 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3465 and disable the IO APIC. legacy for "maxcpus=0".
3467 nosoftlockup [KNL] Disable the soft-lockup detector.
3469 nosync [HW,M68K] Disables sync negotiation for all devices.
3471 nowatchdog [KNL] Disable both lockup detectors, i.e.
3472 soft-lockup and NMI watchdog (hard-lockup).
3476 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3478 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3479 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3480 Some features depend on CPU0. Known dependencies are:
3481 1. Resume from suspend/hibernate depends on CPU0.
3482 Suspend/hibernate will fail if CPU0 is offline and you
3483 need to online CPU0 before suspend/hibernate.
3484 2. PIC interrupts also depend on CPU0. CPU0 can't be
3485 removed if a PIC interrupt is detected.
3486 It's said poweroff/reboot may depend on CPU0 on some
3487 machines although I haven't seen such issues so far
3488 after CPU0 is offline on a few tested machines.
3489 If the dependencies are under your control, you can
3490 turn on cpu0_hotplug.
3492 nps_mtm_hs_ctr= [KNL,ARC]
3493 This parameter sets the maximum duration, in
3494 cycles, each HW thread of the CTOP can run
3495 without interruptions, before HW switches it.
3496 The actual maximum duration is 16 times this
3498 Format: integer between 1 and 255
3501 nptcg= [IA-64] Override max number of concurrent global TLB
3502 purges which is reported from either PAL_VM_SUMMARY or
3505 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3506 could support. nr_cpus=n : n >= 1 limits the kernel to
3507 support 'n' processors. It could be larger than the
3508 number of already plugged CPU during bootup, later in
3509 runtime you can physically add extra cpu until it reaches
3510 n. So during boot up some boot time memory for per-cpu
3511 variables need be pre-allocated for later physical cpu
3514 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3516 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3518 Allowed values are enable and disable
3520 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3521 'node', 'default' can be specified
3522 This can be set from sysctl after boot.
3523 See Documentation/admin-guide/sysctl/vm.rst for details.
3525 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3526 See Documentation/core-api/debugging-via-ohci1394.rst for more
3529 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3530 Rather than timing out after 20 ms if an EC
3531 command is not properly ACKed, override the length
3532 of the timeout. We have interrupts disabled while
3533 waiting for the ACK, so if this is set too high
3534 interrupts *may* be lost!
3536 omap_mux= [OMAP] Override bootloader pin multiplexing.
3537 Format: <mux_mode0.mode_name=value>...
3538 For example, to override I2C bus2:
3539 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3541 oops=panic Always panic on oopses. Default is to just kill the
3542 process, but there is a small probability of
3543 deadlocking the machine.
3544 This will also cause panics on machine check exceptions.
3545 Useful together with panic=30 to trigger a reboot.
3548 [KNL] Boolean flag to control whether the page allocator
3549 should randomize its free lists. The randomization may
3550 be automatically enabled if the kernel detects it is
3551 running on a platform with a direct-mapped memory-side
3552 cache, and this parameter can be used to
3553 override/disable that behavior. The state of the flag
3554 can be read from sysfs at:
3555 /sys/module/page_alloc/parameters/shuffle.
3557 page_owner= [KNL] Boot-time page_owner enabling option.
3558 Storage of the information about who allocated
3559 each page is disabled in default. With this switch,
3561 on: enable the feature
3563 page_poison= [KNL] Boot-time parameter changing the state of
3564 poisoning on the buddy allocator, available with
3565 CONFIG_PAGE_POISONING=y.
3566 off: turn off poisoning (default)
3567 on: turn on poisoning
3569 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3570 timeout > 0: seconds before rebooting
3571 timeout = 0: wait forever
3572 timeout < 0: reboot immediately
3575 panic_print= Bitmask for printing system info when panic happens.
3576 User can chose combination of the following bits:
3577 bit 0: print all tasks info
3578 bit 1: print system memory info
3579 bit 2: print timer info
3580 bit 3: print locks info if CONFIG_LOCKDEP is on
3581 bit 4: print ftrace buffer
3582 bit 5: print all printk messages in buffer
3584 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3585 Format: <hex>[,nousertaint]
3586 Hexadecimal bitmask representing the set of TAINT flags
3587 that will cause the kernel to panic when add_taint() is
3588 called with any of the flags in this set.
3589 The optional switch "nousertaint" can be utilized to
3590 prevent userspace forced crashes by writing to sysctl
3591 /proc/sys/kernel/tainted any flagset matching with the
3592 bitmask set on panic_on_taint.
3593 See Documentation/admin-guide/tainted-kernels.rst for
3594 extra details on the taint flags that users can pick
3595 to compose the bitmask to assign to panic_on_taint.
3597 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3600 crash_kexec_post_notifiers
3601 Run kdump after running panic-notifiers and dumping
3602 kmsg. This only for the users who doubt kdump always
3603 succeeds in any situation.
3604 Note that this also increases risks of kdump failure,
3605 because some panic notifiers can make the crashed
3606 kernel more unstable.
3608 parkbd.port= [HW] Parallel port number the keyboard adapter is
3609 connected to, default is 0.
3611 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3612 0 for XT, 1 for AT (default is AT).
3615 parport= [HW,PPT] Specify parallel ports. 0 disables.
3616 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3617 Use 'auto' to force the driver to use any
3618 IRQ/DMA settings detected (the default is to
3619 ignore detected IRQ/DMA settings because of
3620 possible conflicts). You can specify the base
3621 address, IRQ, and DMA settings; IRQ and DMA
3622 should be numbers, or 'auto' (for using detected
3623 settings on that particular port), or 'nofifo'
3624 (to avoid using a FIFO even if it is detected).
3625 Parallel ports are assigned in the order they
3626 are specified on the command line, starting
3629 parport_init_mode= [HW,PPT]
3630 Configure VIA parallel port to operate in
3631 a specific mode. This is necessary on Pegasos
3632 computer where firmware has no options for setting
3633 up parallel port mode and sets it to spp.
3634 Currently this function knows 686a and 8231 chips.
3635 Format: [spp|ps2|epp|ecp|ecpepp]
3637 pata_legacy.all= [HW,LIBATA]
3639 Set to non-zero to probe primary and secondary ISA
3640 port ranges on PCI systems where no PCI PATA device
3641 has been found at either range. Disabled by default.
3643 pata_legacy.autospeed= [HW,LIBATA]
3645 Set to non-zero if a chip is present that snoops speed
3646 changes. Disabled by default.
3648 pata_legacy.ht6560a= [HW,LIBATA]
3650 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3651 the secondary channel, or both channels respectively.
3652 Disabled by default.
3654 pata_legacy.ht6560b= [HW,LIBATA]
3656 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3657 the secondary channel, or both channels respectively.
3658 Disabled by default.
3660 pata_legacy.iordy_mask= [HW,LIBATA]
3662 IORDY enable mask. Set individual bits to allow IORDY
3663 for the respective channel. Bit 0 is for the first
3664 legacy channel handled by this driver, bit 1 is for
3665 the second channel, and so on. The sequence will often
3666 correspond to the primary legacy channel, the secondary
3667 legacy channel, and so on, but the handling of a PCI
3668 bus and the use of other driver options may interfere
3669 with the sequence. By default IORDY is allowed across
3672 pata_legacy.opti82c46x= [HW,LIBATA]
3674 Set to 1, 2, or 3 for Opti 82c611A on the primary
3675 channel, the secondary channel, or both channels
3676 respectively. Disabled by default.
3678 pata_legacy.opti82c611a= [HW,LIBATA]
3680 Set to 1, 2, or 3 for Opti 82c465MV on the primary
3681 channel, the secondary channel, or both channels
3682 respectively. Disabled by default.
3684 pata_legacy.pio_mask= [HW,LIBATA]
3686 PIO mode mask for autospeed devices. Set individual
3687 bits to allow the use of the respective PIO modes.
3688 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
3689 All modes allowed by default.
3691 pata_legacy.probe_all= [HW,LIBATA]
3693 Set to non-zero to probe tertiary and further ISA
3694 port ranges on PCI systems. Disabled by default.
3696 pata_legacy.probe_mask= [HW,LIBATA]
3698 Probe mask for legacy ISA PATA ports. Depending on
3699 platform configuration and the use of other driver
3700 options up to 6 legacy ports are supported: 0x1f0,
3701 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
3702 of individual ports can be disabled by setting the
3703 corresponding bits in the mask to 1. Bit 0 is for
3704 the first port in the list above (0x1f0), and so on.
3705 By default all supported ports are probed.
3707 pata_legacy.qdi= [HW,LIBATA]
3709 Set to non-zero to probe QDI controllers. By default
3710 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
3712 pata_legacy.winbond= [HW,LIBATA]
3714 Set to non-zero to probe Winbond controllers. Use
3715 the standard I/O port (0x130) if 1, otherwise the
3716 value given is the I/O port to use (typically 0x1b0).
3717 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
3720 pata_platform.pio_mask= [HW,LIBATA]
3722 Supported PIO mode mask. Set individual bits to allow
3723 the use of the respective PIO modes. Bit 0 is for
3724 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
3728 Halt all CPUs after the first oops has been printed for
3729 the specified number of seconds. This is to be used if
3730 your oopses keep scrolling off the screen.
3735 See header of drivers/block/paride/pcd.c.
3736 See also Documentation/admin-guide/blockdev/paride.rst.
3738 pci=option[,option...] [PCI] various PCI subsystem options.
3740 Some options herein operate on a specific device
3741 or a set of devices (<pci_dev>). These are
3742 specified in one of the following formats:
3744 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3745 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3747 Note: the first format specifies a PCI
3748 bus/device/function address which may change
3749 if new hardware is inserted, if motherboard
3750 firmware changes, or due to changes caused
3751 by other kernel parameters. If the
3752 domain is left unspecified, it is
3753 taken to be zero. Optionally, a path
3754 to a device through multiple device/function
3755 addresses can be specified after the base
3756 address (this is more robust against
3757 renumbering issues). The second format
3758 selects devices using IDs from the
3759 configuration space which may match multiple
3760 devices in the system.
3762 earlydump dump PCI config space before the kernel
3764 off [X86] don't probe for the PCI bus
3765 bios [X86-32] force use of PCI BIOS, don't access
3766 the hardware directly. Use this if your machine
3767 has a non-standard PCI host bridge.
3768 nobios [X86-32] disallow use of PCI BIOS, only direct
3769 hardware access methods are allowed. Use this
3770 if you experience crashes upon bootup and you
3771 suspect they are caused by the BIOS.
3772 conf1 [X86] Force use of PCI Configuration Access
3773 Mechanism 1 (config address in IO port 0xCF8,
3774 data in IO port 0xCFC, both 32-bit).
3775 conf2 [X86] Force use of PCI Configuration Access
3776 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3777 the function, IO port 0xCFA, also 8-bit, sets
3778 bus number. The config space is then accessed
3779 through ports 0xC000-0xCFFF).
3780 See http://wiki.osdev.org/PCI for more info
3781 on the configuration access mechanisms.
3782 noaer [PCIE] If the PCIEAER kernel config parameter is
3783 enabled, this kernel boot option can be used to
3784 disable the use of PCIE advanced error reporting.
3785 nodomains [PCI] Disable support for multiple PCI
3786 root domains (aka PCI segments, in ACPI-speak).
3787 nommconf [X86] Disable use of MMCONFIG for PCI
3789 check_enable_amd_mmconf [X86] check for and enable
3790 properly configured MMIO access to PCI
3791 config space on AMD family 10h CPU
3792 nomsi [MSI] If the PCI_MSI kernel config parameter is
3793 enabled, this kernel boot option can be used to
3794 disable the use of MSI interrupts system-wide.
3795 noioapicquirk [APIC] Disable all boot interrupt quirks.
3796 Safety option to keep boot IRQs enabled. This
3797 should never be necessary.
3798 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3799 primary IO-APIC for bridges that cannot disable
3800 boot IRQs. This fixes a source of spurious IRQs
3801 when the system masks IRQs.
3802 noioapicreroute [APIC] Disable workaround that uses the
3803 boot IRQ equivalent of an IRQ that connects to
3804 a chipset where boot IRQs cannot be disabled.
3805 The opposite of ioapicreroute.
3806 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3807 routing table. These calls are known to be buggy
3808 on several machines and they hang the machine
3809 when used, but on other computers it's the only
3810 way to get the interrupt routing table. Try
3811 this option if the kernel is unable to allocate
3812 IRQs or discover secondary PCI buses on your
3814 rom [X86] Assign address space to expansion ROMs.
3815 Use with caution as certain devices share
3816 address decoders between ROMs and other
3818 norom [X86] Do not assign address space to
3819 expansion ROMs that do not already have
3820 BIOS assigned address ranges.
3821 nobar [X86] Do not assign address space to the
3822 BARs that weren't assigned by the BIOS.
3823 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3824 assigned automatically to PCI devices. You can
3825 make the kernel exclude IRQs of your ISA cards
3827 pirqaddr=0xAAAAA [X86] Specify the physical address
3828 of the PIRQ table (normally generated
3829 by the BIOS) if it is outside the
3830 F0000h-100000h range.
3831 lastbus=N [X86] Scan all buses thru bus #N. Can be
3832 useful if the kernel is unable to find your
3833 secondary buses and you want to tell it
3834 explicitly which ones they are.
3835 assign-busses [X86] Always assign all PCI bus
3836 numbers ourselves, overriding
3837 whatever the firmware may have done.
3838 usepirqmask [X86] Honor the possible IRQ mask stored
3839 in the BIOS $PIR table. This is needed on
3840 some systems with broken BIOSes, notably
3841 some HP Pavilion N5400 and Omnibook XE3
3842 notebooks. This will have no effect if ACPI
3843 IRQ routing is enabled.
3844 noacpi [X86] Do not use ACPI for IRQ routing
3845 or for PCI scanning.
3846 use_crs [X86] Use PCI host bridge window information
3847 from ACPI. On BIOSes from 2008 or later, this
3848 is enabled by default. If you need to use this,
3849 please report a bug.
3850 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3851 If you need to use this, please report a bug.
3852 routeirq Do IRQ routing for all PCI devices.
3853 This is normally done in pci_enable_device(),
3854 so this option is a temporary workaround
3855 for broken drivers that don't call it.
3856 skip_isa_align [X86] do not align io start addr, so can
3857 handle more pci cards
3858 noearly [X86] Don't do any early type 1 scanning.
3859 This might help on some broken boards which
3860 machine check when some devices' config space
3861 is read. But various workarounds are disabled
3862 and some IOMMU drivers will not work.
3863 bfsort Sort PCI devices into breadth-first order.
3864 This sorting is done to get a device
3865 order compatible with older (<= 2.4) kernels.
3866 nobfsort Don't sort PCI devices into breadth-first order.
3867 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3868 tuning and use the BIOS-configured MPS defaults.
3869 pcie_bus_safe Set every device's MPS to the largest value
3870 supported by all devices below the root complex.
3871 pcie_bus_perf Set device MPS to the largest allowable MPS
3872 based on its parent bus. Also set MRRS (Max
3873 Read Request Size) to the largest supported
3874 value (no larger than the MPS that the device
3875 or bus can support) for best performance.
3876 pcie_bus_peer2peer Set every device's MPS to 128B, which
3877 every device is guaranteed to support. This
3878 configuration allows peer-to-peer DMA between
3879 any pair of devices, possibly at the cost of
3880 reduced performance. This also guarantees
3881 that hot-added devices will work.
3882 cbiosize=nn[KMG] The fixed amount of bus space which is
3883 reserved for the CardBus bridge's IO window.
3884 The default value is 256 bytes.
3885 cbmemsize=nn[KMG] The fixed amount of bus space which is
3886 reserved for the CardBus bridge's memory
3887 window. The default value is 64 megabytes.
3890 [<order of align>@]<pci_dev>[; ...]
3891 Specifies alignment and device to reassign
3892 aligned memory resources. How to
3893 specify the device is described above.
3894 If <order of align> is not specified,
3895 PAGE_SIZE is used as alignment.
3896 A PCI-PCI bridge can be specified if resource
3897 windows need to be expanded.
3898 To specify the alignment for several
3899 instances of a device, the PCI vendor,
3900 device, subvendor, and subdevice may be
3901 specified, e.g., 12@pci:8086:9c22:103c:198f
3902 for 4096-byte alignment.
3903 ecrc= Enable/disable PCIe ECRC (transaction layer
3904 end-to-end CRC checking).
3905 bios: Use BIOS/firmware settings. This is the
3909 hpiosize=nn[KMG] The fixed amount of bus space which is
3910 reserved for hotplug bridge's IO window.
3911 Default size is 256 bytes.
3912 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3913 reserved for hotplug bridge's MMIO window.
3914 Default size is 2 megabytes.
3915 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3916 reserved for hotplug bridge's MMIO_PREF window.
3917 Default size is 2 megabytes.
3918 hpmemsize=nn[KMG] The fixed amount of bus space which is
3919 reserved for hotplug bridge's MMIO and
3921 Default size is 2 megabytes.
3922 hpbussize=nn The minimum amount of additional bus numbers
3923 reserved for buses below a hotplug bridge.
3925 realloc= Enable/disable reallocating PCI bridge resources
3926 if allocations done by BIOS are too small to
3927 accommodate resources required by all child
3929 off: Turn realloc off
3931 realloc same as realloc=on
3932 noari do not use PCIe ARI.
3933 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3934 do not use PCIe ATS (and IOMMU device IOTLB).
3935 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3936 only look for one device below a PCIe downstream
3938 big_root_window Try to add a big 64bit memory window to the PCIe
3939 root complex on AMD CPUs. Some GFX hardware
3940 can resize a BAR to allow access to all VRAM.
3941 Adding the window is slightly risky (it may
3942 conflict with unreported devices), so this
3944 disable_acs_redir=<pci_dev>[; ...]
3945 Specify one or more PCI devices (in the format
3946 specified above) separated by semicolons.
3947 Each device specified will have the PCI ACS
3948 redirect capabilities forced off which will
3949 allow P2P traffic between devices through
3950 bridges without forcing it upstream. Note:
3951 this removes isolation between devices and
3952 may put more devices in an IOMMU group.
3953 force_floating [S390] Force usage of floating interrupts.
3954 nomio [S390] Do not use MIO instructions.
3955 norid [S390] ignore the RID field and force use of
3956 one PCI domain per PCI function
3958 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3961 force Enable ASPM even on devices that claim not to support it.
3962 WARNING: Forcing ASPM on may cause system lockups.
3964 pcie_ports= [PCIE] PCIe port services handling:
3965 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3966 even if the platform doesn't give the OS permission to
3967 use them. This may cause conflicts if the platform
3968 also tries to use these services.
3969 dpc-native Use native PCIe service for DPC only. May
3970 cause conflicts if firmware uses AER or DPC.
3971 compat Disable native PCIe services (PME, AER, DPC, PCIe
3974 pcie_port_pm= [PCIE] PCIe port power management handling:
3975 off Disable power management of all PCIe ports
3976 force Forcibly enable power management of all PCIe ports
3978 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3979 nomsi Do not use MSI for native PCIe PME signaling (this makes
3980 all PCIe root ports use INTx for all services).
3982 pcmv= [HW,PCMCIA] BadgePAD 4
3986 Keep all power-domains already enabled by bootloader on,
3987 even if no driver has claimed them. This is useful
3988 for debug and development, but should not be
3989 needed on a platform with proper driver support.
3992 See Documentation/admin-guide/blockdev/paride.rst.
3994 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3997 See arch/parisc/kernel/pdc_chassis.c
3999 percpu_alloc= Select which percpu first chunk allocator to use.
4000 Currently supported values are "embed" and "page".
4001 Archs may support subset or none of the selections.
4002 See comments in mm/percpu.c for details on each
4003 allocator. This parameter is primarily for debugging
4004 and performance comparison.
4007 See Documentation/admin-guide/blockdev/paride.rst.
4010 See Documentation/admin-guide/blockdev/paride.rst.
4012 pirq= [SMP,APIC] Manual mp-table setup
4013 See Documentation/x86/i386/IO-APIC.rst.
4015 plip= [PPT,NET] Parallel port network link
4016 Format: { parport<nr> | timid | 0 }
4017 See also Documentation/admin-guide/parport.rst.
4019 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4020 Override pmtimer IOPort with a hex value.
4023 pm_debug_messages [SUSPEND,KNL]
4024 Enable suspend/resume debug messages during boot up.
4027 Enable PNP debug messages (depends on the
4028 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4029 via /sys/module/pnp/parameters/debug. We always show
4030 current resource usage; turning this on also shows
4031 possible settings and some assignment information.
4037 { on | off | curr | res | no-curr | no-res }
4040 [ISAPNP] Exclude IRQs for the autoconfiguration
4043 [ISAPNP] Exclude DMAs for the autoconfiguration
4045 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4046 Ranges are in pairs (I/O port base and size).
4049 [ISAPNP] Exclude memory regions for the
4051 Ranges are in pairs (memory base and size).
4053 ports= [IP_VS_FTP] IPVS ftp helper module
4055 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4057 Format: <port>,<port>....
4059 powersave=off [PPC] This option disables power saving features.
4060 It specifically disables cpuidle and sets the
4061 platform machine description specific power_save
4062 function to NULL. On Idle the CPU just reduces
4065 ppc_strict_facility_enable
4066 [PPC] This option catches any kernel floating point,
4067 Altivec, VSX and SPE outside of regions specifically
4068 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4069 There is some performance impact when enabling this.
4073 Disable Hardware Transactional Memory
4076 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4077 none - Limited to cond_resched() calls
4078 voluntary - Limited to cond_resched() and might_sleep() calls
4079 full - Any section that isn't explicitly preempt disabled
4080 can be preempted anytime.
4082 print-fatal-signals=
4083 [KNL] debug: print fatal signals
4085 If enabled, warn about various signal handling
4086 related application anomalies: too many signals,
4087 too many POSIX.1 timers, fatal signals causing a
4090 If you hit the warning due to signal overflow,
4091 you might want to try "ulimit -i unlimited".
4095 printk.always_kmsg_dump=
4096 Trigger kmsg_dump for cases other than kernel oops or
4098 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4101 printk.devkmsg={on,off,ratelimit}
4102 Control writing to /dev/kmsg.
4103 on - unlimited logging to /dev/kmsg from userspace
4104 off - logging to /dev/kmsg disabled
4105 ratelimit - ratelimit the logging
4108 printk.time= Show timing data prefixed to each printk message line
4109 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4111 processor.max_cstate= [HW,ACPI]
4112 Limit processor to maximum C-state
4113 max_cstate=9 overrides any DMI blacklist limit.
4115 processor.nocst [HW,ACPI]
4116 Ignore the _CST method to determine C-states,
4117 instead using the legacy FADT method
4119 profile= [KNL] Enable kernel profiling via /proc/profile
4120 Format: [<profiletype>,]<number>
4121 Param: <profiletype>: "schedule", "sleep", or "kvm"
4122 [defaults to kernel profiling]
4123 Param: "schedule" - profile schedule points.
4124 Param: "sleep" - profile D-state sleeping (millisecs).
4125 Requires CONFIG_SCHEDSTATS
4126 Param: "kvm" - profile VM exits.
4127 Param: <number> - step/bucket size as a power of 2 for
4128 statistical time based profiling.
4130 prompt_ramdisk= [RAM] [Deprecated]
4132 prot_virt= [S390] enable hosting protected virtual machines
4133 isolated from the hypervisor (if hardware supports
4137 psi= [KNL] Enable or disable pressure stall information
4141 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4142 probe for; one of (bare|imps|exps|lifebook|any).
4143 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4145 psmouse.resetafter= [HW,MOUSE]
4146 Try to reset the device after so many bad packets
4149 [HW,MOUSE] Set desired mouse resolution, in dpi.
4150 psmouse.smartscroll=
4151 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4152 0 = disabled, 1 = enabled (default).
4154 pstore.backend= Specify the name of the pstore backend to use
4157 See Documentation/admin-guide/blockdev/paride.rst.
4159 pti= [X86-64] Control Page Table Isolation of user and
4160 kernel address spaces. Disabling this feature
4161 removes hardening, but improves performance of
4162 system calls and interrupts.
4164 on - unconditionally enable
4165 off - unconditionally disable
4166 auto - kernel detects whether your CPU model is
4167 vulnerable to issues that PTI mitigates
4169 Not specifying this option is equivalent to pti=auto.
4172 Equivalent to pti=off
4175 [KNL] Number of legacy pty's. Overwrites compiled-in
4178 quiet [KNL] Disable most log messages
4183 See Documentation/admin-guide/md.rst.
4185 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4186 See Documentation/admin-guide/blockdev/ramdisk.rst.
4188 ramdisk_start= [RAM] RAM disk image start address
4190 random.trust_cpu={on,off}
4191 [KNL] Enable or disable trusting the use of the
4192 CPU's random number generator (if available) to
4193 fully seed the kernel's CRNG. Default is controlled
4194 by CONFIG_RANDOM_TRUST_CPU.
4196 randomize_kstack_offset=
4197 [KNL] Enable or disable kernel stack offset
4198 randomization, which provides roughly 5 bits of
4199 entropy, frustrating memory corruption attacks
4200 that depend on stack address determinism or
4201 cross-syscall address exposures. This is only
4202 available on architectures that have defined
4203 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4204 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4205 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4207 ras=option[,option,...] [KNL] RAS-specific options
4210 Disable the Correctable Errors Collector,
4211 see CONFIG_RAS_CEC help text.
4214 The argument is a cpu list, as described above.
4216 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
4217 the specified list of CPUs to be no-callback CPUs.
4218 Invocation of these CPUs' RCU callbacks will be
4219 offloaded to "rcuox/N" kthreads created for that
4220 purpose, where "x" is "p" for RCU-preempt, and
4221 "s" for RCU-sched, and "N" is the CPU number.
4222 This reduces OS jitter on the offloaded CPUs,
4223 which can be useful for HPC and real-time
4224 workloads. It can also improve energy efficiency
4225 for asymmetric multiprocessors.
4228 Rather than requiring that offloaded CPUs
4229 (specified by rcu_nocbs= above) explicitly
4230 awaken the corresponding "rcuoN" kthreads,
4231 make these kthreads poll for callbacks.
4232 This improves the real-time response for the
4233 offloaded CPUs by relieving them of the need to
4234 wake up the corresponding kthread, but degrades
4235 energy efficiency by requiring that the kthreads
4236 periodically wake up to do the polling.
4238 rcutree.blimit= [KNL]
4239 Set maximum number of finished RCU callbacks to
4240 process in one batch.
4242 rcutree.dump_tree= [KNL]
4243 Dump the structure of the rcu_node combining tree
4244 out at early boot. This is used for diagnostic
4245 purposes, to verify correct tree setup.
4247 rcutree.gp_cleanup_delay= [KNL]
4248 Set the number of jiffies to delay each step of
4249 RCU grace-period cleanup.
4251 rcutree.gp_init_delay= [KNL]
4252 Set the number of jiffies to delay each step of
4253 RCU grace-period initialization.
4255 rcutree.gp_preinit_delay= [KNL]
4256 Set the number of jiffies to delay each step of
4257 RCU grace-period pre-initialization, that is,
4258 the propagation of recent CPU-hotplug changes up
4259 the rcu_node combining tree.
4261 rcutree.use_softirq= [KNL]
4262 If set to zero, move all RCU_SOFTIRQ processing to
4263 per-CPU rcuc kthreads. Defaults to a non-zero
4264 value, meaning that RCU_SOFTIRQ is used by default.
4265 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4267 But note that CONFIG_PREEMPT_RT=y kernels disable
4268 this kernel boot parameter, forcibly setting it
4271 rcutree.rcu_fanout_exact= [KNL]
4272 Disable autobalancing of the rcu_node combining
4273 tree. This is used by rcutorture, and might
4274 possibly be useful for architectures having high
4275 cache-to-cache transfer latencies.
4277 rcutree.rcu_fanout_leaf= [KNL]
4278 Change the number of CPUs assigned to each
4279 leaf rcu_node structure. Useful for very
4280 large systems, which will choose the value 64,
4281 and for NUMA systems with large remote-access
4282 latencies, which will choose a value aligned
4283 with the appropriate hardware boundaries.
4285 rcutree.rcu_min_cached_objs= [KNL]
4286 Minimum number of objects which are cached and
4287 maintained per one CPU. Object size is equal
4288 to PAGE_SIZE. The cache allows to reduce the
4289 pressure to page allocator, also it makes the
4290 whole algorithm to behave better in low memory
4293 rcutree.jiffies_till_first_fqs= [KNL]
4294 Set delay from grace-period initialization to
4295 first attempt to force quiescent states.
4296 Units are jiffies, minimum value is zero,
4297 and maximum value is HZ.
4299 rcutree.jiffies_till_next_fqs= [KNL]
4300 Set delay between subsequent attempts to force
4301 quiescent states. Units are jiffies, minimum
4302 value is one, and maximum value is HZ.
4304 rcutree.jiffies_till_sched_qs= [KNL]
4305 Set required age in jiffies for a
4306 given grace period before RCU starts
4307 soliciting quiescent-state help from
4308 rcu_note_context_switch() and cond_resched().
4309 If not specified, the kernel will calculate
4310 a value based on the most recent settings
4311 of rcutree.jiffies_till_first_fqs
4312 and rcutree.jiffies_till_next_fqs.
4313 This calculated value may be viewed in
4314 rcutree.jiffies_to_sched_qs. Any attempt to set
4315 rcutree.jiffies_to_sched_qs will be cheerfully
4318 rcutree.kthread_prio= [KNL,BOOT]
4319 Set the SCHED_FIFO priority of the RCU per-CPU
4320 kthreads (rcuc/N). This value is also used for
4321 the priority of the RCU boost threads (rcub/N)
4322 and for the RCU grace-period kthreads (rcu_bh,
4323 rcu_preempt, and rcu_sched). If RCU_BOOST is
4324 set, valid values are 1-99 and the default is 1
4325 (the least-favored priority). Otherwise, when
4326 RCU_BOOST is not set, valid values are 0-99 and
4327 the default is zero (non-realtime operation).
4329 rcutree.rcu_nocb_gp_stride= [KNL]
4330 Set the number of NOCB callback kthreads in
4331 each group, which defaults to the square root
4332 of the number of CPUs. Larger numbers reduce
4333 the wakeup overhead on the global grace-period
4334 kthread, but increases that same overhead on
4335 each group's NOCB grace-period kthread.
4337 rcutree.qhimark= [KNL]
4338 Set threshold of queued RCU callbacks beyond which
4339 batch limiting is disabled.
4341 rcutree.qlowmark= [KNL]
4342 Set threshold of queued RCU callbacks below which
4343 batch limiting is re-enabled.
4345 rcutree.qovld= [KNL]
4346 Set threshold of queued RCU callbacks beyond which
4347 RCU's force-quiescent-state scan will aggressively
4348 enlist help from cond_resched() and sched IPIs to
4349 help CPUs more quickly reach quiescent states.
4350 Set to less than zero to make this be set based
4351 on rcutree.qhimark at boot time and to zero to
4352 disable more aggressive help enlistment.
4354 rcutree.rcu_idle_gp_delay= [KNL]
4355 Set wakeup interval for idle CPUs that have
4356 RCU callbacks (RCU_FAST_NO_HZ=y).
4358 rcutree.rcu_kick_kthreads= [KNL]
4359 Cause the grace-period kthread to get an extra
4360 wake_up() if it sleeps three times longer than
4361 it should at force-quiescent-state time.
4362 This wake_up() will be accompanied by a
4363 WARN_ONCE() splat and an ftrace_dump().
4365 rcutree.rcu_unlock_delay= [KNL]
4366 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4367 this specifies an rcu_read_unlock()-time delay
4368 in microseconds. This defaults to zero.
4369 Larger delays increase the probability of
4370 catching RCU pointer leaks, that is, buggy use
4371 of RCU-protected pointers after the relevant
4372 rcu_read_unlock() has completed.
4374 rcutree.sysrq_rcu= [KNL]
4375 Commandeer a sysrq key to dump out Tree RCU's
4376 rcu_node tree with an eye towards determining
4377 why a new grace period has not yet started.
4379 rcuscale.gp_async= [KNL]
4380 Measure performance of asynchronous
4381 grace-period primitives such as call_rcu().
4383 rcuscale.gp_async_max= [KNL]
4384 Specify the maximum number of outstanding
4385 callbacks per writer thread. When a writer
4386 thread exceeds this limit, it invokes the
4387 corresponding flavor of rcu_barrier() to allow
4388 previously posted callbacks to drain.
4390 rcuscale.gp_exp= [KNL]
4391 Measure performance of expedited synchronous
4392 grace-period primitives.
4394 rcuscale.holdoff= [KNL]
4395 Set test-start holdoff period. The purpose of
4396 this parameter is to delay the start of the
4397 test until boot completes in order to avoid
4400 rcuscale.kfree_rcu_test= [KNL]
4401 Set to measure performance of kfree_rcu() flooding.
4403 rcuscale.kfree_rcu_test_double= [KNL]
4404 Test the double-argument variant of kfree_rcu().
4405 If this parameter has the same value as
4406 rcuscale.kfree_rcu_test_single, both the single-
4407 and double-argument variants are tested.
4409 rcuscale.kfree_rcu_test_single= [KNL]
4410 Test the single-argument variant of kfree_rcu().
4411 If this parameter has the same value as
4412 rcuscale.kfree_rcu_test_double, both the single-
4413 and double-argument variants are tested.
4415 rcuscale.kfree_nthreads= [KNL]
4416 The number of threads running loops of kfree_rcu().
4418 rcuscale.kfree_alloc_num= [KNL]
4419 Number of allocations and frees done in an iteration.
4421 rcuscale.kfree_loops= [KNL]
4422 Number of loops doing rcuscale.kfree_alloc_num number
4423 of allocations and frees.
4425 rcuscale.nreaders= [KNL]
4426 Set number of RCU readers. The value -1 selects
4427 N, where N is the number of CPUs. A value
4428 "n" less than -1 selects N-n+1, where N is again
4429 the number of CPUs. For example, -2 selects N
4430 (the number of CPUs), -3 selects N+1, and so on.
4431 A value of "n" less than or equal to -N selects
4434 rcuscale.nwriters= [KNL]
4435 Set number of RCU writers. The values operate
4436 the same as for rcuscale.nreaders.
4437 N, where N is the number of CPUs
4439 rcuscale.perf_type= [KNL]
4440 Specify the RCU implementation to test.
4442 rcuscale.shutdown= [KNL]
4443 Shut the system down after performance tests
4444 complete. This is useful for hands-off automated
4447 rcuscale.verbose= [KNL]
4448 Enable additional printk() statements.
4450 rcuscale.writer_holdoff= [KNL]
4451 Write-side holdoff between grace periods,
4452 in microseconds. The default of zero says
4455 rcutorture.fqs_duration= [KNL]
4456 Set duration of force_quiescent_state bursts
4459 rcutorture.fqs_holdoff= [KNL]
4460 Set holdoff time within force_quiescent_state bursts
4463 rcutorture.fqs_stutter= [KNL]
4464 Set wait time between force_quiescent_state bursts
4467 rcutorture.fwd_progress= [KNL]
4468 Enable RCU grace-period forward-progress testing
4469 for the types of RCU supporting this notion.
4471 rcutorture.fwd_progress_div= [KNL]
4472 Specify the fraction of a CPU-stall-warning
4473 period to do tight-loop forward-progress testing.
4475 rcutorture.fwd_progress_holdoff= [KNL]
4476 Number of seconds to wait between successive
4477 forward-progress tests.
4479 rcutorture.fwd_progress_need_resched= [KNL]
4480 Enclose cond_resched() calls within checks for
4481 need_resched() during tight-loop forward-progress
4484 rcutorture.gp_cond= [KNL]
4485 Use conditional/asynchronous update-side
4486 primitives, if available.
4488 rcutorture.gp_exp= [KNL]
4489 Use expedited update-side primitives, if available.
4491 rcutorture.gp_normal= [KNL]
4492 Use normal (non-expedited) asynchronous
4493 update-side primitives, if available.
4495 rcutorture.gp_sync= [KNL]
4496 Use normal (non-expedited) synchronous
4497 update-side primitives, if available. If all
4498 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4499 rcutorture.gp_normal=, and rcutorture.gp_sync=
4500 are zero, rcutorture acts as if is interpreted
4501 they are all non-zero.
4503 rcutorture.irqreader= [KNL]
4504 Run RCU readers from irq handlers, or, more
4505 accurately, from a timer handler. Not all RCU
4506 flavors take kindly to this sort of thing.
4508 rcutorture.leakpointer= [KNL]
4509 Leak an RCU-protected pointer out of the reader.
4510 This can of course result in splats, and is
4511 intended to test the ability of things like
4512 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4515 rcutorture.n_barrier_cbs= [KNL]
4516 Set callbacks/threads for rcu_barrier() testing.
4518 rcutorture.nfakewriters= [KNL]
4519 Set number of concurrent RCU writers. These just
4520 stress RCU, they don't participate in the actual
4521 test, hence the "fake".
4523 rcutorture.nocbs_nthreads= [KNL]
4524 Set number of RCU callback-offload togglers.
4525 Zero (the default) disables toggling.
4527 rcutorture.nocbs_toggle= [KNL]
4528 Set the delay in milliseconds between successive
4529 callback-offload toggling attempts.
4531 rcutorture.nreaders= [KNL]
4532 Set number of RCU readers. The value -1 selects
4533 N-1, where N is the number of CPUs. A value
4534 "n" less than -1 selects N-n-2, where N is again
4535 the number of CPUs. For example, -2 selects N
4536 (the number of CPUs), -3 selects N+1, and so on.
4538 rcutorture.object_debug= [KNL]
4539 Enable debug-object double-call_rcu() testing.
4541 rcutorture.onoff_holdoff= [KNL]
4542 Set time (s) after boot for CPU-hotplug testing.
4544 rcutorture.onoff_interval= [KNL]
4545 Set time (jiffies) between CPU-hotplug operations,
4546 or zero to disable CPU-hotplug testing.
4548 rcutorture.read_exit= [KNL]
4549 Set the number of read-then-exit kthreads used
4550 to test the interaction of RCU updaters and
4551 task-exit processing.
4553 rcutorture.read_exit_burst= [KNL]
4554 The number of times in a given read-then-exit
4555 episode that a set of read-then-exit kthreads
4558 rcutorture.read_exit_delay= [KNL]
4559 The delay, in seconds, between successive
4560 read-then-exit testing episodes.
4562 rcutorture.shuffle_interval= [KNL]
4563 Set task-shuffle interval (s). Shuffling tasks
4564 allows some CPUs to go into dyntick-idle mode
4565 during the rcutorture test.
4567 rcutorture.shutdown_secs= [KNL]
4568 Set time (s) after boot system shutdown. This
4569 is useful for hands-off automated testing.
4571 rcutorture.stall_cpu= [KNL]
4572 Duration of CPU stall (s) to test RCU CPU stall
4573 warnings, zero to disable.
4575 rcutorture.stall_cpu_block= [KNL]
4576 Sleep while stalling if set. This will result
4577 in warnings from preemptible RCU in addition
4578 to any other stall-related activity.
4580 rcutorture.stall_cpu_holdoff= [KNL]
4581 Time to wait (s) after boot before inducing stall.
4583 rcutorture.stall_cpu_irqsoff= [KNL]
4584 Disable interrupts while stalling if set.
4586 rcutorture.stall_gp_kthread= [KNL]
4587 Duration (s) of forced sleep within RCU
4588 grace-period kthread to test RCU CPU stall
4589 warnings, zero to disable. If both stall_cpu
4590 and stall_gp_kthread are specified, the
4591 kthread is starved first, then the CPU.
4593 rcutorture.stat_interval= [KNL]
4594 Time (s) between statistics printk()s.
4596 rcutorture.stutter= [KNL]
4597 Time (s) to stutter testing, for example, specifying
4598 five seconds causes the test to run for five seconds,
4599 wait for five seconds, and so on. This tests RCU's
4600 ability to transition abruptly to and from idle.
4602 rcutorture.test_boost= [KNL]
4603 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4604 "Maybe" means test if the RCU implementation
4605 under test support RCU priority boosting.
4607 rcutorture.test_boost_duration= [KNL]
4608 Duration (s) of each individual boost test.
4610 rcutorture.test_boost_interval= [KNL]
4611 Interval (s) between each boost test.
4613 rcutorture.test_no_idle_hz= [KNL]
4614 Test RCU's dyntick-idle handling. See also the
4615 rcutorture.shuffle_interval parameter.
4617 rcutorture.torture_type= [KNL]
4618 Specify the RCU implementation to test.
4620 rcutorture.verbose= [KNL]
4621 Enable additional printk() statements.
4623 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4624 Dump ftrace buffer after reporting RCU CPU
4627 rcupdate.rcu_cpu_stall_suppress= [KNL]
4628 Suppress RCU CPU stall warning messages.
4630 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4631 Suppress RCU CPU stall warning messages and
4632 rcutorture writer stall warnings that occur
4633 during early boot, that is, during the time
4634 before the init task is spawned.
4636 rcupdate.rcu_cpu_stall_timeout= [KNL]
4637 Set timeout for RCU CPU stall warning messages.
4639 rcupdate.rcu_expedited= [KNL]
4640 Use expedited grace-period primitives, for
4641 example, synchronize_rcu_expedited() instead
4642 of synchronize_rcu(). This reduces latency,
4643 but can increase CPU utilization, degrade
4644 real-time latency, and degrade energy efficiency.
4645 No effect on CONFIG_TINY_RCU kernels.
4647 rcupdate.rcu_normal= [KNL]
4648 Use only normal grace-period primitives,
4649 for example, synchronize_rcu() instead of
4650 synchronize_rcu_expedited(). This improves
4651 real-time latency, CPU utilization, and
4652 energy efficiency, but can expose users to
4653 increased grace-period latency. This parameter
4654 overrides rcupdate.rcu_expedited. No effect on
4655 CONFIG_TINY_RCU kernels.
4657 rcupdate.rcu_normal_after_boot= [KNL]
4658 Once boot has completed (that is, after
4659 rcu_end_inkernel_boot() has been invoked), use
4660 only normal grace-period primitives. No effect
4661 on CONFIG_TINY_RCU kernels.
4663 But note that CONFIG_PREEMPT_RT=y kernels enables
4664 this kernel boot parameter, forcibly setting
4665 it to the value one, that is, converting any
4666 post-boot attempt at an expedited RCU grace
4667 period to instead use normal non-expedited
4668 grace-period processing.
4670 rcupdate.rcu_task_ipi_delay= [KNL]
4671 Set time in jiffies during which RCU tasks will
4672 avoid sending IPIs, starting with the beginning
4673 of a given grace period. Setting a large
4674 number avoids disturbing real-time workloads,
4675 but lengthens grace periods.
4677 rcupdate.rcu_task_stall_timeout= [KNL]
4678 Set timeout in jiffies for RCU task stall warning
4679 messages. Disable with a value less than or equal
4682 rcupdate.rcu_self_test= [KNL]
4683 Run the RCU early boot self tests
4687 Run specified binary instead of /init from the ramdisk,
4688 used for early userspace startup. See initrd.
4691 force - Override the decision by the kernel to hide the
4692 advertisement of RDRAND support (this affects
4693 certain AMD processors because of buggy BIOS
4694 support, specifically around the suspend/resume
4698 Turn on/off individual RDT features. List is:
4699 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4701 E.g. to turn on cmt and turn off mba use:
4705 Format (x86 or x86_64):
4706 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4708 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4710 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4711 (prefix with 'panic_' to set mode for panic
4713 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4714 reboot_force is either force or not specified,
4715 reboot_cpu is s[mp]#### with #### being the processor
4716 to be used for rebooting.
4718 refscale.holdoff= [KNL]
4719 Set test-start holdoff period. The purpose of
4720 this parameter is to delay the start of the
4721 test until boot completes in order to avoid
4724 refscale.loops= [KNL]
4725 Set the number of loops over the synchronization
4726 primitive under test. Increasing this number
4727 reduces noise due to loop start/end overhead,
4728 but the default has already reduced the per-pass
4729 noise to a handful of picoseconds on ca. 2020
4732 refscale.nreaders= [KNL]
4733 Set number of readers. The default value of -1
4734 selects N, where N is roughly 75% of the number
4735 of CPUs. A value of zero is an interesting choice.
4737 refscale.nruns= [KNL]
4738 Set number of runs, each of which is dumped onto
4741 refscale.readdelay= [KNL]
4742 Set the read-side critical-section duration,
4743 measured in microseconds.
4745 refscale.scale_type= [KNL]
4746 Specify the read-protection implementation to test.
4748 refscale.shutdown= [KNL]
4749 Shut down the system at the end of the performance
4750 test. This defaults to 1 (shut it down) when
4751 refscale is built into the kernel and to 0 (leave
4752 it running) when refscale is built as a module.
4754 refscale.verbose= [KNL]
4755 Enable additional printk() statements.
4757 refscale.verbose_batched= [KNL]
4758 Batch the additional printk() statements. If zero
4759 (the default) or negative, print everything. Otherwise,
4760 print every Nth verbose statement, where N is the value
4764 [KNL, SMP] Set scheduler's default relax_domain_level.
4765 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4767 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4768 Format: <base1>,<size1>[,<base2>,<size2>,...]
4769 Reserve I/O ports or memory so the kernel won't use
4770 them. If <base> is less than 0x10000, the region
4771 is assumed to be I/O ports; otherwise it is memory.
4773 reservetop= [X86-32]
4775 Reserves a hole at the top of the kernel virtual
4780 Set the amount of memory to reserve for BIOS at
4781 the bottom of the address space.
4783 reset_devices [KNL] Force drivers to reset the underlying device
4784 during initialization.
4787 Specify the partition device for software suspend
4789 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4791 resume_offset= [SWSUSP]
4792 Specify the offset from the beginning of the partition
4793 given by "resume=" at which the swap header is located,
4794 in <PAGE_SIZE> units (needed only for swap files).
4795 See Documentation/power/swsusp-and-swap-files.rst
4797 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4798 read the resume files
4800 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4801 Useful for devices that are detected asynchronously
4802 (e.g. USB and MMC devices).
4804 hibernate= [HIBERNATION]
4805 noresume Don't check if there's a hibernation image
4806 present during boot.
4807 nocompress Don't compress/decompress hibernation images.
4808 no Disable hibernation and resume.
4809 protect_image Turn on image protection during restoration
4810 (that will set all pages holding image data
4811 during restoration read-only).
4813 retain_initrd [RAM] Keep initrd memory after extraction
4815 rfkill.default_state=
4816 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4817 etc. communication is blocked by default.
4820 rfkill.master_switch_mode=
4821 0 The "airplane mode" button does nothing.
4822 1 The "airplane mode" button toggles between everything
4823 blocked and the previous configuration.
4824 2 The "airplane mode" button toggles between everything
4825 blocked and everything unblocked.
4827 rhash_entries= [KNL,NET]
4828 Set number of hash buckets for route cache
4831 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4834 ro [KNL] Mount root device read-only on boot
4837 on Mark read-only kernel memory as read-only (default).
4838 off Leave read-only kernel memory writable for debugging.
4841 Enable the uart passthrough on the designated usb port
4842 on Rockchip SoCs. When active, the signals of the
4843 debug-uart get routed to the D+ and D- pins of the usb
4844 port and the regular usb controller gets disabled.
4846 root= [KNL] Root filesystem
4847 See name_to_dev_t comment in init/do_mounts.c.
4849 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4850 mount the root filesystem
4852 rootflags= [KNL] Set root filesystem mount option string
4854 rootfstype= [KNL] Set root filesystem type
4856 rootwait [KNL] Wait (indefinitely) for root device to show up.
4857 Useful for devices that are detected asynchronously
4858 (e.g. USB and MMC devices).
4860 rproc_mem=nn[KMG][@address]
4861 [KNL,ARM,CMA] Remoteproc physical memory block.
4862 Memory area to be used by remote processor image,
4865 rw [KNL] Mount root device read-write on boot
4867 S [KNL] Run init in single mode
4869 s390_iommu= [HW,S390]
4870 Set s390 IOTLB flushing mode
4872 With strict flushing every unmap operation will result in
4873 an IOTLB flush. Default is lazy flushing before reuse,
4877 See drivers/net/irda/sa1100_ir.c.
4879 sbni= [NET] Granch SBNI12 leased line adapter
4881 sched_verbose [KNL] Enables verbose scheduler debug messages.
4883 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4884 Allowed values are enable and disable. This feature
4885 incurs a small amount of overhead in the scheduler
4886 but is useful for debugging and performance tuning.
4888 sched_thermal_decay_shift=
4889 [KNL, SMP] Set a decay shift for scheduler thermal
4890 pressure signal. Thermal pressure signal follows the
4891 default decay period of other scheduler pelt
4892 signals(usually 32 ms but configurable). Setting
4893 sched_thermal_decay_shift will left shift the decay
4894 period for the thermal pressure signal by the shift
4896 i.e. with the default pelt decay period of 32 ms
4897 sched_thermal_decay_shift thermal pressure decay pr
4901 Format: integer between 0 and 10
4904 scftorture.holdoff= [KNL]
4905 Number of seconds to hold off before starting
4906 test. Defaults to zero for module insertion and
4907 to 10 seconds for built-in smp_call_function()
4910 scftorture.longwait= [KNL]
4911 Request ridiculously long waits randomly selected
4912 up to the chosen limit in seconds. Zero (the
4913 default) disables this feature. Please note
4914 that requesting even small non-zero numbers of
4915 seconds can result in RCU CPU stall warnings,
4916 softlockup complaints, and so on.
4918 scftorture.nthreads= [KNL]
4919 Number of kthreads to spawn to invoke the
4920 smp_call_function() family of functions.
4921 The default of -1 specifies a number of kthreads
4922 equal to the number of CPUs.
4924 scftorture.onoff_holdoff= [KNL]
4925 Number seconds to wait after the start of the
4926 test before initiating CPU-hotplug operations.
4928 scftorture.onoff_interval= [KNL]
4929 Number seconds to wait between successive
4930 CPU-hotplug operations. Specifying zero (which
4931 is the default) disables CPU-hotplug operations.
4933 scftorture.shutdown_secs= [KNL]
4934 The number of seconds following the start of the
4935 test after which to shut down the system. The
4936 default of zero avoids shutting down the system.
4937 Non-zero values are useful for automated tests.
4939 scftorture.stat_interval= [KNL]
4940 The number of seconds between outputting the
4941 current test statistics to the console. A value
4942 of zero disables statistics output.
4944 scftorture.stutter_cpus= [KNL]
4945 The number of jiffies to wait between each change
4946 to the set of CPUs under test.
4948 scftorture.use_cpus_read_lock= [KNL]
4949 Use use_cpus_read_lock() instead of the default
4950 preempt_disable() to disable CPU hotplug
4951 while invoking one of the smp_call_function*()
4954 scftorture.verbose= [KNL]
4955 Enable additional printk() statements.
4957 scftorture.weight_single= [KNL]
4958 The probability weighting to use for the
4959 smp_call_function_single() function with a zero
4960 "wait" parameter. A value of -1 selects the
4961 default if all other weights are -1. However,
4962 if at least one weight has some other value, a
4963 value of -1 will instead select a weight of zero.
4965 scftorture.weight_single_wait= [KNL]
4966 The probability weighting to use for the
4967 smp_call_function_single() function with a
4968 non-zero "wait" parameter. See weight_single.
4970 scftorture.weight_many= [KNL]
4971 The probability weighting to use for the
4972 smp_call_function_many() function with a zero
4973 "wait" parameter. See weight_single.
4974 Note well that setting a high probability for
4975 this weighting can place serious IPI load
4978 scftorture.weight_many_wait= [KNL]
4979 The probability weighting to use for the
4980 smp_call_function_many() function with a
4981 non-zero "wait" parameter. See weight_single
4984 scftorture.weight_all= [KNL]
4985 The probability weighting to use for the
4986 smp_call_function_all() function with a zero
4987 "wait" parameter. See weight_single and
4990 scftorture.weight_all_wait= [KNL]
4991 The probability weighting to use for the
4992 smp_call_function_all() function with a
4993 non-zero "wait" parameter. See weight_single
4996 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4997 xtime_lock contention on larger systems, and/or RCU lock
4998 contention on all systems with CONFIG_MAXSMP set.
4999 Format: { "0" | "1" }
5000 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5002 Note: increases power consumption, thus should only be
5003 enabled if running jitter sensitive (HPC/RT) workloads.
5005 security= [SECURITY] Choose a legacy "major" security module to
5006 enable at boot. This has been deprecated by the
5009 selinux= [SELINUX] Disable or enable SELinux at boot time.
5010 Format: { "0" | "1" }
5011 See security/selinux/Kconfig help text.
5016 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5017 Format: { "0" | "1" }
5018 See security/apparmor/Kconfig help text
5021 Default value is set via kernel config option.
5023 serialnumber [BUGS=X86-32]
5026 Maximal number of shapers.
5034 Enable merging of slabs with similar size when the
5035 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5038 Disable merging of slabs with similar size. May be
5039 necessary if there is some reason to distinguish
5040 allocs to different slabs, especially in hardened
5041 environments where the risk of heap overflows and
5042 layout control by attackers can usually be
5043 frustrated by disabling merging. This will reduce
5044 most of the exposure of a heap attack to a single
5045 cache (risks via metadata attacks are mostly
5046 unchanged). Debug options disable merging on their
5048 For more information see Documentation/vm/slub.rst.
5050 slab_max_order= [MM, SLAB]
5051 Determines the maximum allowed order for slabs.
5052 A high setting may cause OOMs due to memory
5053 fragmentation. Defaults to 1 for systems with
5054 more than 32MB of RAM, 0 otherwise.
5056 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5057 Enabling slub_debug allows one to determine the
5058 culprit if slab objects become corrupted. Enabling
5059 slub_debug can create guard zones around objects and
5060 may poison objects when not in use. Also tracks the
5061 last alloc / free. For more information see
5062 Documentation/vm/slub.rst.
5064 slub_max_order= [MM, SLUB]
5065 Determines the maximum allowed order for slabs.
5066 A high setting may cause OOMs due to memory
5067 fragmentation. For more information see
5068 Documentation/vm/slub.rst.
5070 slub_min_objects= [MM, SLUB]
5071 The minimum number of objects per slab. SLUB will
5072 increase the slab order up to slub_max_order to
5073 generate a sufficiently large slab able to contain
5074 the number of objects indicated. The higher the number
5075 of objects the smaller the overhead of tracking slabs
5076 and the less frequently locks need to be acquired.
5077 For more information see Documentation/vm/slub.rst.
5079 slub_min_order= [MM, SLUB]
5080 Determines the minimum page order for slabs. Must be
5081 lower than slub_max_order.
5082 For more information see Documentation/vm/slub.rst.
5084 slub_merge [MM, SLUB]
5085 Same with slab_merge.
5087 slub_nomerge [MM, SLUB]
5088 Same with slab_nomerge. This is supported for legacy.
5089 See slab_nomerge for more information.
5092 Format: <io1>[,<io2>[,...,<io8>]]
5094 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5095 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5096 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5097 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5098 smsc-ircc2.ircc_irq= [HW] IRQ line
5099 smsc-ircc2.ircc_dma= [HW] DMA channel
5100 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5101 0: Toshiba Satellite 1800 (GP data pin select)
5102 1: Fast pin select (default)
5105 smt [KNL,S390] Set the maximum number of threads (logical
5106 CPUs) to use per physical CPU on systems capable of
5107 symmetric multithreading (SMT). Will be capped to the
5108 actual hardware limit.
5110 Default: -1 (no limit)
5113 [KNL] Should the soft-lockup detector generate panics.
5116 A value of 1 instructs the soft-lockup detector
5117 to panic the machine when a soft-lockup occurs. It is
5118 also controlled by the kernel.softlockup_panic sysctl
5119 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5120 respective build-time switch to that functionality.
5122 softlockup_all_cpu_backtrace=
5123 [KNL] Should the soft-lockup detector generate
5124 backtraces on all cpus.
5127 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5128 See Documentation/admin-guide/laptops/sonypi.rst
5130 spectre_v2= [X86] Control mitigation of Spectre variant 2
5131 (indirect branch speculation) vulnerability.
5132 The default operation protects the kernel from
5135 on - unconditionally enable, implies
5137 off - unconditionally disable, implies
5139 auto - kernel detects whether your CPU model is
5142 Selecting 'on' will, and 'auto' may, choose a
5143 mitigation method at run time according to the
5144 CPU, the available microcode, the setting of the
5145 CONFIG_RETPOLINE configuration option, and the
5146 compiler with which the kernel was built.
5148 Selecting 'on' will also enable the mitigation
5149 against user space to user space task attacks.
5151 Selecting 'off' will disable both the kernel and
5152 the user space protections.
5154 Specific mitigations can also be selected manually:
5156 retpoline - replace indirect branches
5157 retpoline,generic - google's original retpoline
5158 retpoline,amd - AMD-specific minimal thunk
5160 Not specifying this option is equivalent to
5164 [X86] Control mitigation of Spectre variant 2
5165 (indirect branch speculation) vulnerability between
5168 on - Unconditionally enable mitigations. Is
5169 enforced by spectre_v2=on
5171 off - Unconditionally disable mitigations. Is
5172 enforced by spectre_v2=off
5174 prctl - Indirect branch speculation is enabled,
5175 but mitigation can be enabled via prctl
5176 per thread. The mitigation control state
5177 is inherited on fork.
5180 - Like "prctl" above, but only STIBP is
5181 controlled per thread. IBPB is issued
5182 always when switching between different user
5186 - Same as "prctl" above, but all seccomp
5187 threads will enable the mitigation unless
5188 they explicitly opt out.
5191 - Like "seccomp" above, but only STIBP is
5192 controlled per thread. IBPB is issued
5193 always when switching between different
5194 user space processes.
5196 auto - Kernel selects the mitigation depending on
5197 the available CPU features and vulnerability.
5200 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
5202 Not specifying this option is equivalent to
5203 spectre_v2_user=auto.
5205 spec_store_bypass_disable=
5206 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5207 (Speculative Store Bypass vulnerability)
5209 Certain CPUs are vulnerable to an exploit against a
5210 a common industry wide performance optimization known
5211 as "Speculative Store Bypass" in which recent stores
5212 to the same memory location may not be observed by
5213 later loads during speculative execution. The idea
5214 is that such stores are unlikely and that they can
5215 be detected prior to instruction retirement at the
5216 end of a particular speculation execution window.
5218 In vulnerable processors, the speculatively forwarded
5219 store can be used in a cache side channel attack, for
5220 example to read memory to which the attacker does not
5221 directly have access (e.g. inside sandboxed code).
5223 This parameter controls whether the Speculative Store
5224 Bypass optimization is used.
5226 On x86 the options are:
5228 on - Unconditionally disable Speculative Store Bypass
5229 off - Unconditionally enable Speculative Store Bypass
5230 auto - Kernel detects whether the CPU model contains an
5231 implementation of Speculative Store Bypass and
5232 picks the most appropriate mitigation. If the
5233 CPU is not vulnerable, "off" is selected. If the
5234 CPU is vulnerable the default mitigation is
5235 architecture and Kconfig dependent. See below.
5236 prctl - Control Speculative Store Bypass per thread
5237 via prctl. Speculative Store Bypass is enabled
5238 for a process by default. The state of the control
5239 is inherited on fork.
5240 seccomp - Same as "prctl" above, but all seccomp threads
5241 will disable SSB unless they explicitly opt out.
5243 Default mitigations:
5244 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
5246 On powerpc the options are:
5248 on,auto - On Power8 and Power9 insert a store-forwarding
5249 barrier on kernel entry and exit. On Power7
5250 perform a software flush on kernel entry and
5254 Not specifying this option is equivalent to
5255 spec_store_bypass_disable=auto.
5257 spia_io_base= [HW,MTD]
5263 [X86] Enable split lock detection or bus lock detection
5265 When enabled (and if hardware support is present), atomic
5266 instructions that access data across cache line
5267 boundaries will result in an alignment check exception
5268 for split lock detection or a debug exception for
5273 warn - the kernel will emit rate-limited warnings
5274 about applications triggering the #AC
5275 exception or the #DB exception. This mode is
5276 the default on CPUs that support split lock
5277 detection or bus lock detection. Default
5278 behavior is by #AC if both features are
5279 enabled in hardware.
5281 fatal - the kernel will send SIGBUS to applications
5282 that trigger the #AC exception or the #DB
5283 exception. Default behavior is by #AC if
5284 both features are enabled in hardware.
5286 If an #AC exception is hit in the kernel or in
5287 firmware (i.e. not while executing in user mode)
5288 the kernel will oops in either "warn" or "fatal"
5291 #DB exception for bus lock is triggered only when
5295 Control the Special Register Buffer Data Sampling
5298 Certain CPUs are vulnerable to an MDS-like
5299 exploit which can leak bits from the random
5302 By default, this issue is mitigated by
5303 microcode. However, the microcode fix can cause
5304 the RDRAND and RDSEED instructions to become
5305 much slower. Among other effects, this will
5306 result in reduced throughput from /dev/urandom.
5308 The microcode mitigation can be disabled with
5309 the following option:
5311 off: Disable mitigation and remove
5312 performance impact to RDRAND and RDSEED
5314 srcutree.counter_wrap_check [KNL]
5315 Specifies how frequently to check for
5316 grace-period sequence counter wrap for the
5317 srcu_data structure's ->srcu_gp_seq_needed field.
5318 The greater the number of bits set in this kernel
5319 parameter, the less frequently counter wrap will
5320 be checked for. Note that the bottom two bits
5323 srcutree.exp_holdoff [KNL]
5324 Specifies how many nanoseconds must elapse
5325 since the end of the last SRCU grace period for
5326 a given srcu_struct until the next normal SRCU
5327 grace period will be considered for automatic
5328 expediting. Set to zero to disable automatic
5332 Speculative Store Bypass Disable control
5334 On CPUs that are vulnerable to the Speculative
5335 Store Bypass vulnerability and offer a
5336 firmware based mitigation, this parameter
5337 indicates how the mitigation should be used:
5339 force-on: Unconditionally enable mitigation for
5340 for both kernel and userspace
5341 force-off: Unconditionally disable mitigation for
5342 for both kernel and userspace
5343 kernel: Always enable mitigation in the
5344 kernel, and offer a prctl interface
5345 to allow userspace to register its
5346 interest in being mitigated too.
5348 stack_guard_gap= [MM]
5349 override the default stack gap protection. The value
5350 is in page units and it defines how many pages prior
5351 to (for stacks growing down) resp. after (for stacks
5352 growing up) the main stack are reserved for no other
5353 mapping. Default value is 256 pages.
5355 stack_depot_disable= [KNL]
5356 Setting this to true through kernel command line will
5357 disable the stack depot thereby saving the static memory
5358 consumed by the stack hash table. By default this is set
5362 Enabled the stack tracer on boot up.
5364 stacktrace_filter=[function-list]
5365 [FTRACE] Limit the functions that the stack tracer
5366 will trace at boot up. function-list is a comma-separated
5367 list of functions. This list can be changed at run
5368 time by the stack_trace_filter file in the debugfs
5369 tracing directory. Note, this enables stack tracing
5370 and the stacktrace above is not needed.
5374 Set the STI (builtin display/keyboard on the HP-PARISC
5375 machines) console (graphic card) which should be used
5376 as the initial boot-console.
5377 See also comment in drivers/video/console/sticore.c.
5380 See comment in drivers/video/console/sticore.c.
5383 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5385 sunrpc.min_resvport=
5386 sunrpc.max_resvport=
5388 SunRPC servers often require that client requests
5389 originate from a privileged port (i.e. a port in the
5390 range 0 < portnr < 1024).
5391 An administrator who wishes to reserve some of these
5392 ports for other uses may adjust the range that the
5393 kernel's sunrpc client considers to be privileged
5394 using these two parameters to set the minimum and
5395 maximum port values.
5397 sunrpc.svc_rpc_per_connection_limit=
5399 Limit the number of requests that the server will
5400 process in parallel from a single connection.
5401 The default value is 0 (no limit).
5405 Control how the NFS server code allocates CPUs to
5406 service thread pools. Depending on how many NICs
5407 you have and where their interrupts are bound, this
5408 option will affect which CPUs will do NFS serving.
5409 Note: this parameter cannot be changed while the
5410 NFS server is running.
5412 auto the server chooses an appropriate mode
5413 automatically using heuristics
5414 global a single global pool contains all CPUs
5415 percpu one pool for each CPU
5416 pernode one pool for each NUMA node (equivalent
5417 to global on non-NUMA machines)
5419 sunrpc.tcp_slot_table_entries=
5420 sunrpc.udp_slot_table_entries=
5422 Sets the upper limit on the number of simultaneous
5423 RPC calls that can be sent from the client to a
5424 server. Increasing these values may allow you to
5425 improve throughput, but will also increase the
5426 amount of memory reserved for use by the client.
5428 suspend.pm_test_delay=
5430 Sets the number of seconds to remain in a suspend test
5431 mode before resuming the system (see
5432 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5433 is set. Default value is 5.
5436 Format: { on | off | y | n | 1 | 0 }
5437 This parameter controls use of the Protected
5438 Execution Facility on pSeries.
5441 [KNL] Enable accounting of swap in memory resource
5442 controller if no parameter or 1 is given or disable
5443 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5445 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5446 Format: { <int> | force | noforce }
5447 <int> -- Number of I/O TLB slabs
5448 force -- force using of bounce buffers even if they
5449 wouldn't be automatically used by the kernel
5450 noforce -- Never use bounce buffers (for debugging)
5455 Set a sysctl parameter, right before loading the init
5456 process, as if the value was written to the respective
5457 /proc/sys/... file. Both '.' and '/' are recognized as
5458 separators. Unrecognized parameters and invalid values
5459 are reported in the kernel log. Sysctls registered
5460 later by a loaded module cannot be set this way.
5461 Example: sysctl.vm.swappiness=40
5463 sysfs.deprecated=0|1 [KNL]
5464 Enable/disable old style sysfs layout for old udev
5465 on older distributions. When this option is enabled
5466 very new udev will not work anymore. When this option
5467 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5468 in older udev will not work anymore.
5469 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5470 the kernel configuration.
5472 sysrq_always_enabled
5474 Ignore sysrq setting - this boot parameter will
5475 neutralize any effect of /proc/sys/kernel/sysrq.
5476 Useful for debugging.
5478 tcpmhash_entries= [KNL,NET]
5479 Set the number of tcp_metrics_hash slots.
5480 Default value is 8192 or 16384 depending on total
5481 ram pages. This is used to specify the TCP metrics
5482 cache size. See Documentation/networking/ip-sysctl.rst
5483 "tcp_no_metrics_save" section for more details.
5487 test_suspend= [SUSPEND][,N]
5488 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5489 standby suspend) or "freeze" (for suspend type freeze)
5490 as the system sleep state during system startup with
5491 the optional capability to repeat N number of times.
5492 The system is woken from this state using a
5493 wakeup-capable RTC alarm.
5495 thash_entries= [KNL,NET]
5496 Set number of hash buckets for TCP connection
5498 thermal.act= [HW,ACPI]
5499 -1: disable all active trip points in all thermal zones
5500 <degrees C>: override all lowest active trip points
5502 thermal.crt= [HW,ACPI]
5503 -1: disable all critical trip points in all thermal zones
5504 <degrees C>: override all critical trip points
5506 thermal.nocrt= [HW,ACPI]
5507 Set to disable actions on ACPI thermal zone
5508 critical and hot trip points.
5510 thermal.off= [HW,ACPI]
5511 1: disable ACPI thermal control
5513 thermal.psv= [HW,ACPI]
5514 -1: disable all passive trip points
5515 <degrees C>: override all passive trip points to this
5518 thermal.tzp= [HW,ACPI]
5519 Specify global default ACPI thermal zone polling rate
5520 <deci-seconds>: poll all this frequency
5521 0: no polling (default)
5524 Force threading of all interrupt handlers except those
5525 marked explicitly IRQF_NO_THREAD.
5529 Specify if the kernel should make use of the cpu
5530 topology information if the hardware supports this.
5531 The scheduler will make use of this information and
5532 e.g. base its process migration decisions on it.
5535 topology_updates= [KNL, PPC, NUMA]
5537 Specify if the kernel should ignore (off)
5538 topology updates sent by the hypervisor to this
5541 torture.disable_onoff_at_boot= [KNL]
5542 Prevent the CPU-hotplug component of torturing
5543 until after init has spawned.
5545 torture.ftrace_dump_at_shutdown= [KNL]
5546 Dump the ftrace buffer at torture-test shutdown,
5547 even if there were no errors. This can be a
5548 very costly operation when many torture tests
5549 are running concurrently, especially on systems
5550 with rotating-rust storage.
5552 torture.verbose_sleep_frequency= [KNL]
5553 Specifies how many verbose printk()s should be
5554 emitted between each sleep. The default of zero
5555 disables verbose-printk() sleeping.
5557 torture.verbose_sleep_duration= [KNL]
5558 Duration of each verbose-printk() sleep in jiffies.
5562 tpm_suspend_pcr=[HW,TPM]
5563 Format: integer pcr id
5564 Specify that at suspend time, the tpm driver
5565 should extend the specified pcr with zeros,
5566 as a workaround for some chips which fail to
5567 flush the last written pcr on TPM_SaveState.
5568 This will guarantee that all the other pcrs
5571 trace_buf_size=nn[KMG]
5572 [FTRACE] will set tracing buffer size on each cpu.
5574 trace_event=[event-list]
5575 [FTRACE] Set and start specified trace events in order
5576 to facilitate early boot debugging. The event-list is a
5577 comma-separated list of trace events to enable. See
5578 also Documentation/trace/events.rst
5580 trace_options=[option-list]
5581 [FTRACE] Enable or disable tracer options at boot.
5582 The option-list is a comma delimited list of options
5583 that can be enabled or disabled just as if you were
5584 to echo the option name into
5586 /sys/kernel/debug/tracing/trace_options
5588 For example, to enable stacktrace option (to dump the
5589 stack trace of each event), add to the command line:
5591 trace_options=stacktrace
5593 See also Documentation/trace/ftrace.rst "trace options"
5597 Have the tracepoints sent to printk as well as the
5598 tracing ring buffer. This is useful for early boot up
5599 where the system hangs or reboots and does not give the
5600 option for reading the tracing buffer or performing a
5601 ftrace_dump_on_oops.
5603 To turn off having tracepoints sent to printk,
5604 echo 0 > /proc/sys/kernel/tracepoint_printk
5605 Note, echoing 1 into this file without the
5606 tracepoint_printk kernel cmdline option has no effect.
5610 Having tracepoints sent to printk() and activating high
5611 frequency tracepoints such as irq or sched, can cause
5612 the system to live lock.
5615 [FTRACE] enable this option to disable tracing when a
5616 warning is hit. This turns off "tracing_on". Tracing can
5617 be enabled again by echoing '1' into the "tracing_on"
5618 file located in /sys/kernel/debug/tracing/
5620 This option is useful, as it disables the trace before
5621 the WARNING dump is called, which prevents the trace to
5622 be filled with content caused by the warning output.
5624 This option can also be set at run time via the sysctl
5625 option: kernel/traceoff_on_warning
5627 transparent_hugepage=
5629 Format: [always|madvise|never]
5630 Can be used to control the default behavior of the system
5631 with respect to transparent hugepages.
5632 See Documentation/admin-guide/mm/transhuge.rst
5635 trusted.source= [KEYS]
5637 This parameter identifies the trust source as a backend
5638 for trusted keys implementation. Supported trust
5642 If not specified then it defaults to iterating through
5643 the trust source list starting with TPM and assigns the
5644 first trust source as a backend which is initialized
5645 successfully during iteration.
5647 tsc= Disable clocksource stability checks for TSC.
5649 [x86] reliable: mark tsc clocksource as reliable, this
5650 disables clocksource verification at runtime, as well
5651 as the stability checks done at bootup. Used to enable
5652 high-resolution timer mode on older hardware, and in
5653 virtualized environment.
5654 [x86] noirqtime: Do not use TSC to do irq accounting.
5655 Used to run time disable IRQ_TIME_ACCOUNTING on any
5656 platforms where RDTSC is slow and this accounting
5658 [x86] unstable: mark the TSC clocksource as unstable, this
5659 marks the TSC unconditionally unstable at bootup and
5660 avoids any further wobbles once the TSC watchdog notices.
5661 [x86] nowatchdog: disable clocksource watchdog. Used
5662 in situations with strict latency requirements (where
5663 interruptions from clocksource watchdog are not
5666 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5667 value instead. Useful when the early TSC frequency discovery
5668 procedure is not reliable, such as on overclocked systems
5669 with CPUID.16h support and partial CPUID.15h support.
5670 Format: <unsigned int>
5672 tsx= [X86] Control Transactional Synchronization
5673 Extensions (TSX) feature in Intel processors that
5674 support TSX control.
5676 This parameter controls the TSX feature. The options are:
5678 on - Enable TSX on the system. Although there are
5679 mitigations for all known security vulnerabilities,
5680 TSX has been known to be an accelerator for
5681 several previous speculation-related CVEs, and
5682 so there may be unknown security risks associated
5683 with leaving it enabled.
5685 off - Disable TSX on the system. (Note that this
5686 option takes effect only on newer CPUs which are
5687 not vulnerable to MDS, i.e., have
5688 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5689 the new IA32_TSX_CTRL MSR through a microcode
5690 update. This new MSR allows for the reliable
5691 deactivation of the TSX functionality.)
5693 auto - Disable TSX if X86_BUG_TAA is present,
5694 otherwise enable TSX on the system.
5696 Not specifying this option is equivalent to tsx=off.
5698 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5701 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5702 Abort (TAA) vulnerability.
5704 Similar to Micro-architectural Data Sampling (MDS)
5705 certain CPUs that support Transactional
5706 Synchronization Extensions (TSX) are vulnerable to an
5707 exploit against CPU internal buffers which can forward
5708 information to a disclosure gadget under certain
5711 In vulnerable processors, the speculatively forwarded
5712 data can be used in a cache side channel attack, to
5713 access data to which the attacker does not have direct
5716 This parameter controls the TAA mitigation. The
5719 full - Enable TAA mitigation on vulnerable CPUs
5722 full,nosmt - Enable TAA mitigation and disable SMT on
5723 vulnerable CPUs. If TSX is disabled, SMT
5724 is not disabled because CPU is not
5725 vulnerable to cross-thread TAA attacks.
5726 off - Unconditionally disable TAA mitigation
5728 On MDS-affected machines, tsx_async_abort=off can be
5729 prevented by an active MDS mitigation as both vulnerabilities
5730 are mitigated with the same mechanism so in order to disable
5731 this mitigation, you need to specify mds=off too.
5733 Not specifying this option is equivalent to
5734 tsx_async_abort=full. On CPUs which are MDS affected
5735 and deploy MDS mitigation, TAA mitigation is not
5736 required and doesn't provide any additional
5740 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5742 turbografx.map[2|3]= [HW,JOY]
5743 TurboGraFX parallel port interface
5745 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5746 See also Documentation/input/devices/joystick-parport.rst
5748 udbg-immortal [PPC] When debugging early kernel crashes that
5749 happen after console_init() and before a proper
5750 console driver takes over, this boot options might
5751 help "seeing" what's going on.
5753 uhash_entries= [KNL,NET]
5754 Set number of hash buckets for UDP/UDP-Lite connections
5757 [USB] Ignore overcurrent events (default N).
5758 Some badly-designed motherboards generate lots of
5759 bogus events, for ports that aren't wired to
5760 anything. Set this parameter to avoid log spamming.
5761 Note that genuine overcurrent events won't be
5765 [X86] Cause panic on unknown NMI.
5767 usbcore.authorized_default=
5768 [USB] Default USB device authorization:
5769 (default -1 = authorized except for wireless USB,
5770 0 = not authorized, 1 = authorized, 2 = authorized
5771 if device connected to internal port)
5773 usbcore.autosuspend=
5774 [USB] The autosuspend time delay (in seconds) used
5775 for newly-detected USB devices (default 2). This
5776 is the time required before an idle device will be
5777 autosuspended. Devices for which the delay is set
5778 to a negative value won't be autosuspended at all.
5780 usbcore.usbfs_snoop=
5781 [USB] Set to log all usbfs traffic (default 0 = off).
5783 usbcore.usbfs_snoop_max=
5784 [USB] Maximum number of bytes to snoop in each URB
5787 usbcore.blinkenlights=
5788 [USB] Set to cycle leds on hubs (default 0 = off).
5790 usbcore.old_scheme_first=
5791 [USB] Start with the old device initialization
5792 scheme (default 0 = off).
5794 usbcore.usbfs_memory_mb=
5795 [USB] Memory limit (in MB) for buffers allocated by
5796 usbfs (default = 16, 0 = max = 2047).
5798 usbcore.use_both_schemes=
5799 [USB] Try the other device initialization scheme
5800 if the first one fails (default 1 = enabled).
5802 usbcore.initial_descriptor_timeout=
5803 [USB] Specifies timeout for the initial 64-byte
5804 USB_REQ_GET_DESCRIPTOR request in milliseconds
5805 (default 5000 = 5.0 seconds).
5807 usbcore.nousb [USB] Disable the USB subsystem
5810 [USB] A list of quirk entries to augment the built-in
5811 usb core quirk list. List entries are separated by
5812 commas. Each entry has the form
5813 VendorID:ProductID:Flags. The IDs are 4-digit hex
5814 numbers and Flags is a set of letters. Each letter
5815 will change the built-in quirk; setting it if it is
5816 clear and clearing it if it is set. The letters have
5817 the following meanings:
5818 a = USB_QUIRK_STRING_FETCH_255 (string
5819 descriptors must not be fetched using
5821 b = USB_QUIRK_RESET_RESUME (device can't resume
5822 correctly so reset it instead);
5823 c = USB_QUIRK_NO_SET_INTF (device can't handle
5824 Set-Interface requests);
5825 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5826 handle its Configuration or Interface
5828 e = USB_QUIRK_RESET (device can't be reset
5829 (e.g morph devices), don't use reset);
5830 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5831 more interface descriptions than the
5832 bNumInterfaces count, and can't handle
5833 talking to these interfaces);
5834 g = USB_QUIRK_DELAY_INIT (device needs a pause
5835 during initialization, after we read
5836 the device descriptor);
5837 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5838 high speed and super speed interrupt
5839 endpoints, the USB 2.0 and USB 3.0 spec
5840 require the interval in microframes (1
5841 microframe = 125 microseconds) to be
5842 calculated as interval = 2 ^
5844 Devices with this quirk report their
5845 bInterval as the result of this
5846 calculation instead of the exponent
5847 variable used in the calculation);
5848 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5849 handle device_qualifier descriptor
5851 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5852 generates spurious wakeup, ignore
5853 remote wakeup capability);
5854 k = USB_QUIRK_NO_LPM (device can't handle Link
5856 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5857 (Device reports its bInterval as linear
5858 frames instead of the USB 2.0
5860 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5861 to be disconnected before suspend to
5862 prevent spurious wakeup);
5863 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5864 pause after every control message);
5865 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5866 delay after resetting its port);
5867 Example: quirks=0781:5580:bk,0a5c:5834:gij
5870 [USBHID] The interval which mice are to be polled at.
5873 [USBHID] The interval which joysticks are to be polled at.
5876 [USBHID] The interval which keyboards are to be polled at.
5878 usb-storage.delay_use=
5879 [UMS] The delay in seconds before a new device is
5880 scanned for Logical Units (default 1).
5883 [UMS] A list of quirks entries to supplement or
5884 override the built-in unusual_devs list. List
5885 entries are separated by commas. Each entry has
5886 the form VID:PID:Flags where VID and PID are Vendor
5887 and Product ID values (4-digit hex numbers) and
5888 Flags is a set of characters, each corresponding
5889 to a common usb-storage quirk flag as follows:
5890 a = SANE_SENSE (collect more than 18 bytes
5891 of sense data, not on uas);
5892 b = BAD_SENSE (don't collect more than 18
5893 bytes of sense data, not on uas);
5894 c = FIX_CAPACITY (decrease the reported
5895 device capacity by one sector);
5896 d = NO_READ_DISC_INFO (don't use
5897 READ_DISC_INFO command, not on uas);
5898 e = NO_READ_CAPACITY_16 (don't use
5899 READ_CAPACITY_16 command);
5900 f = NO_REPORT_OPCODES (don't use report opcodes
5902 g = MAX_SECTORS_240 (don't transfer more than
5903 240 sectors at a time, uas only);
5904 h = CAPACITY_HEURISTICS (decrease the
5905 reported device capacity by one
5906 sector if the number is odd);
5907 i = IGNORE_DEVICE (don't bind to this
5909 j = NO_REPORT_LUNS (don't use report luns
5911 k = NO_SAME (do not use WRITE_SAME, uas only)
5912 l = NOT_LOCKABLE (don't try to lock and
5913 unlock ejectable media, not on uas);
5914 m = MAX_SECTORS_64 (don't transfer more
5915 than 64 sectors = 32 KB at a time,
5917 n = INITIAL_READ10 (force a retry of the
5918 initial READ(10) command, not on uas);
5919 o = CAPACITY_OK (accept the capacity
5920 reported by the device, not on uas);
5921 p = WRITE_CACHE (the device cache is ON
5922 by default, not on uas);
5923 r = IGNORE_RESIDUE (the device reports
5924 bogus residue values, not on uas);
5925 s = SINGLE_LUN (the device has only one
5927 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5928 commands, uas only);
5929 u = IGNORE_UAS (don't bind to the uas driver);
5930 w = NO_WP_DETECT (don't test whether the
5931 medium is write-protected).
5932 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5933 even if the device claims no cache,
5935 Example: quirks=0419:aaf5:rl,0421:0433:rc
5937 user_debug= [KNL,ARM]
5939 See arch/arm/Kconfig.debug help text.
5940 1 - undefined instruction events
5942 4 - invalid data aborts
5945 Example: user_debug=31
5948 [X86] Flags controlling user PTE allocations.
5950 nohigh = do not allocate PTE pages in
5951 HIGHMEM regardless of setting
5955 On X86_32, this is an alias for vdso32=. Otherwise:
5957 vdso=1: enable VDSO (the default)
5958 vdso=0: disable VDSO mapping
5960 vdso32= [X86] Control the 32-bit vDSO
5961 vdso32=1: enable 32-bit VDSO
5962 vdso32=0 or vdso32=2: disable 32-bit VDSO
5964 See the help text for CONFIG_COMPAT_VDSO for more
5965 details. If CONFIG_COMPAT_VDSO is set, the default is
5966 vdso32=0; otherwise, the default is vdso32=1.
5968 For compatibility with older kernels, vdso32=2 is an
5971 Try vdso32=0 if you encounter an error that says:
5972 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5975 vector=percpu: enable percpu vector domain
5977 video= [FB] Frame buffer configuration
5978 See Documentation/fb/modedb.rst.
5980 video.brightness_switch_enabled= [0,1]
5981 If set to 1, on receiving an ACPI notify event
5982 generated by hotkey, video driver will adjust brightness
5983 level and then send out the event to user space through
5984 the allocated input device; If set to 0, video driver
5985 will only send out the event without touching backlight
5990 [VMMIO] Memory mapped virtio (platform) device.
5992 <size>@<baseaddr>:<irq>[:<id>]
5994 <size> := size (can use standard suffixes
5996 <baseaddr> := physical base address
5997 <irq> := interrupt number (as passed to
5999 <id> := (optional) platform device id
6001 virtio_mmio.device=1K@0x100b0000:48:7
6003 Can be used multiple times for multiple devices.
6005 vga= [BOOT,X86-32] Select a particular video mode
6006 See Documentation/x86/boot.rst and
6007 Documentation/admin-guide/svga.rst.
6008 Use vga=ask for menu.
6009 This is actually a boot loader parameter; the value is
6010 passed to the kernel using a special protocol.
6012 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6013 May slow down system boot speed, especially when
6014 enabled on systems with a large amount of memory.
6015 All options are enabled by default, and this
6016 interface is meant to allow for selectively
6017 enabling or disabling specific virtual memory
6020 Available options are:
6021 P Enable page structure init time poisoning
6022 - Disable all of the above options
6024 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6025 size of <nn>. This can be used to increase the
6026 minimum size (128MB on x86). It can also be used to
6027 decrease the size and leave more room for directly
6030 vmcp_cma=nn[MG] [KNL,S390]
6031 Sets the memory size reserved for contiguous memory
6032 allocations for the vmcp device driver.
6034 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6037 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6040 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6044 Controls the behavior of vsyscalls (i.e. calls to
6045 fixed addresses of 0xffffffffff600x00 from legacy
6046 code). Most statically-linked binaries and older
6047 versions of glibc use these calls. Because these
6048 functions are at fixed addresses, they make nice
6049 targets for exploits that can control RIP.
6051 emulate [default] Vsyscalls turn into traps and are
6052 emulated reasonably safely. The vsyscall
6055 xonly Vsyscalls turn into traps and are
6056 emulated reasonably safely. The vsyscall
6057 page is not readable.
6059 none Vsyscalls don't work at all. This makes
6060 them quite hard to use for exploits but
6061 might break your system.
6063 vt.color= [VT] Default text color.
6064 Format: 0xYX, X = foreground, Y = background.
6065 Default: 0x07 = light gray on black.
6067 vt.cur_default= [VT] Default cursor shape.
6068 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6069 the parameters of the <Esc>[?A;B;Cc escape sequence;
6070 see VGA-softcursor.txt. Default: 2 = underline.
6072 vt.default_blu= [VT]
6073 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6074 Change the default blue palette of the console.
6075 This is a 16-member array composed of values
6078 vt.default_grn= [VT]
6079 Format: <green0>,<green1>,<green2>,...,<green15>
6080 Change the default green palette of the console.
6081 This is a 16-member array composed of values
6084 vt.default_red= [VT]
6085 Format: <red0>,<red1>,<red2>,...,<red15>
6086 Change the default red palette of the console.
6087 This is a 16-member array composed of values
6093 Set system-wide default UTF-8 mode for all tty's.
6094 Default is 1, i.e. UTF-8 mode is enabled for all
6095 newly opened terminals.
6097 vt.global_cursor_default=
6100 Set system-wide default for whether a cursor
6101 is shown on new VTs. Default is -1,
6102 i.e. cursors will be created by default unless
6103 overridden by individual drivers. 0 will hide
6104 cursors, 1 will display them.
6106 vt.italic= [VT] Default color for italic text; 0-15.
6109 vt.underline= [VT] Default color for underlined text; 0-15.
6112 watchdog timers [HW,WDT] For information on watchdog timers,
6113 see Documentation/watchdog/watchdog-parameters.rst
6114 or other driver-specific files in the
6115 Documentation/watchdog/ directory.
6119 Set the hard lockup detector stall duration
6120 threshold in seconds. The soft lockup detector
6121 threshold is set to twice the value. A value of 0
6122 disables both lockup detectors. Default is 10
6125 workqueue.watchdog_thresh=
6126 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6127 warn stall conditions and dump internal state to
6128 help debugging. 0 disables workqueue stall
6129 detection; otherwise, it's the stall threshold
6130 duration in seconds. The default value is 30 and
6131 it can be updated at runtime by writing to the
6132 corresponding sysfs file.
6134 workqueue.disable_numa
6135 By default, all work items queued to unbound
6136 workqueues are affine to the NUMA nodes they're
6137 issued on, which results in better behavior in
6138 general. If NUMA affinity needs to be disabled for
6139 whatever reason, this option can be used. Note
6140 that this also can be controlled per-workqueue for
6141 workqueues visible under /sys/bus/workqueue/.
6143 workqueue.power_efficient
6144 Per-cpu workqueues are generally preferred because
6145 they show better performance thanks to cache
6146 locality; unfortunately, per-cpu workqueues tend to
6147 be more power hungry than unbound workqueues.
6149 Enabling this makes the per-cpu workqueues which
6150 were observed to contribute significantly to power
6151 consumption unbound, leading to measurably lower
6152 power usage at the cost of small performance
6155 The default value of this parameter is determined by
6156 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6158 workqueue.debug_force_rr_cpu
6159 Workqueue used to implicitly guarantee that work
6160 items queued without explicit CPU specified are put
6161 on the local CPU. This guarantee is no longer true
6162 and while local CPU is still preferred work items
6163 may be put on foreign CPUs. This debug option
6164 forces round-robin CPU selection to flush out
6165 usages which depend on the now broken guarantee.
6166 When enabled, memory and cache locality will be
6169 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6170 default x2apic cluster mode on platforms
6173 xen_512gb_limit [KNL,X86-64,XEN]
6174 Restricts the kernel running paravirtualized under Xen
6175 to use only up to 512 GB of RAM. The reason to do so is
6176 crash analysis tools and Xen tools for doing domain
6177 save/restore/migration must be enabled to handle larger
6180 xen_emul_unplug= [HW,X86,XEN]
6181 Unplug Xen emulated devices
6182 Format: [unplug0,][unplug1]
6183 ide-disks -- unplug primary master IDE devices
6184 aux-ide-disks -- unplug non-primary-master IDE devices
6185 nics -- unplug network devices
6186 all -- unplug all emulated devices (NICs and IDE disks)
6187 unnecessary -- unplugging emulated devices is
6188 unnecessary even if the host did not respond to
6190 never -- do not unplug even if version check succeeds
6192 xen_legacy_crash [X86,XEN]
6193 Crash from Xen panic notifier, without executing late
6194 panic() code such as dumping handler.
6196 xen_nopvspin [X86,XEN]
6197 Disables the qspinlock slowpath using Xen PV optimizations.
6198 This parameter is obsoleted by "nopvspin" parameter, which
6199 has equivalent effect for XEN platform.
6202 Disables the PV optimizations forcing the HVM guest to
6203 run as generic HVM guest with no PV drivers.
6204 This option is obsoleted by the "nopv" option, which
6205 has equivalent effect for XEN platform.
6207 xen_no_vector_callback
6208 [KNL,X86,XEN] Disable the vector callback for Xen
6209 event channel interrupts.
6211 xen_scrub_pages= [XEN]
6212 Boolean option to control scrubbing pages before giving them back
6213 to Xen, for use by other domains. Can be also changed at runtime
6214 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6215 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6217 xen_timer_slop= [X86-64,XEN]
6218 Set the timer slop (in nanoseconds) for the virtual Xen
6219 timers (default is 100000). This adjusts the minimum
6220 delta of virtualized Xen timers, where lower values
6221 improve timer resolution at the expense of processing
6222 more timer interrupts.
6224 xen.event_eoi_delay= [XEN]
6225 How long to delay EOI handling in case of event
6226 storms (jiffies). Default is 10.
6228 xen.event_loop_timeout= [XEN]
6229 After which time (jiffies) the event handling loop
6230 should start to delay EOI handling. Default is 2.
6232 xen.fifo_events= [XEN]
6233 Boolean parameter to disable using fifo event handling
6234 even if available. Normally fifo event handling is
6235 preferred over the 2-level event handling, as it is
6236 fairer and the number of possible event channels is
6237 much higher. Default is on (use fifo events).
6239 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6240 Disables the PV optimizations forcing the guest to run
6241 as generic guest with no PV drivers. Currently support
6242 XEN HVM, KVM, HYPER_V and VMWARE guest.
6244 nopvspin [X86,XEN,KVM]
6245 Disables the qspinlock slow path using PV optimizations
6246 which allow the hypervisor to 'idle' the guest on lock
6249 xirc2ps_cs= [NET,PCMCIA]
6251 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6254 By default on POWER9 and above, the kernel will
6255 natively use the XIVE interrupt controller. This option
6256 allows the fallback firmware mode to be used:
6258 off Fallback to firmware control of XIVE interrupt
6259 controller on both pseries and powernv
6260 platforms. Only useful on POWER9 and above.
6262 xhci-hcd.quirks [USB,KNL]
6263 A hex value specifying bitmask with supplemental xhci
6264 host controller quirks. Meaning of each bit can be
6265 consulted in header drivers/usb/host/xhci.h.
6268 Format: { early | on | rw | ro | off }
6269 Controls if xmon debugger is enabled. Default is off.
6270 Passing only "xmon" is equivalent to "xmon=early".
6271 early Call xmon as early as possible on boot; xmon
6272 debugger is called from setup_arch().
6273 on xmon debugger hooks will be installed so xmon
6274 is only called on a kernel crash. Default mode,
6275 i.e. either "ro" or "rw" mode, is controlled
6276 with CONFIG_XMON_DEFAULT_RO_MODE.
6277 rw xmon debugger hooks will be installed so xmon
6278 is called only on a kernel crash, mode is write,
6279 meaning SPR registers, memory and, other data
6280 can be written using xmon commands.
6281 ro same as "rw" option above but SPR registers,
6282 memory, and other data can't be written using
6284 off xmon is disabled.