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 clocksource.max_cswd_read_retries= [KNL]
585 Number of clocksource_watchdog() retries due to
586 external delays before the clock will be marked
587 unstable. Defaults to three retries, that is,
588 four attempts to read the clock under test.
590 clearcpuid=BITNUM[,BITNUM...] [X86]
591 Disable CPUID feature X for the kernel. See
592 arch/x86/include/asm/cpufeatures.h for the valid bit
593 numbers. Note the Linux specific bits are not necessarily
594 stable over kernel options, but the vendor specific
596 Also note that user programs calling CPUID directly
597 or using the feature without checking anything
598 will still see it. This just prevents it from
599 being used by the kernel or shown in /proc/cpuinfo.
600 Also note the kernel might malfunction if you disable
603 cma=nn[MG]@[start[MG][-end[MG]]]
605 Sets the size of kernel global memory area for
606 contiguous memory allocations and optionally the
607 placement constraint by the physical address range of
608 memory allocations. A value of 0 disables CMA
609 altogether. For more information, see
610 kernel/dma/contiguous.c
614 Sets the size of kernel per-numa memory area for
615 contiguous memory allocations. A value of 0 disables
616 per-numa CMA altogether. And If this option is not
617 specificed, the default value is 0.
618 With per-numa CMA enabled, DMA users on node nid will
619 first try to allocate buffer from the pernuma area
620 which is located in node nid, if the allocation fails,
621 they will fallback to the global default memory area.
623 cmo_free_hint= [PPC] Format: { yes | no }
624 Specify whether pages are marked as being inactive
625 when they are freed. This is used in CMO environments
626 to determine OS memory pressure for page stealing by
630 coherent_pool=nn[KMG] [ARM,KNL]
631 Sets the size of memory pool for coherent, atomic dma
632 allocations, by default set to 256K.
634 com20020= [HW,NET] ARCnet - COM20020 chipset
636 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
638 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
642 ARCnet - COM90xx chipset (memory-mapped buffers)
643 Format: <io>[,<irq>[,<memstart>]]
645 condev= [HW,S390] console device
648 console= [KNL] Output console device and options.
650 tty<n> Use the virtual console device <n>.
654 Use the specified serial port. The options are of
655 the form "bbbbpnf", where "bbbb" is the baud rate,
656 "p" is parity ("n", "o", or "e"), "n" is number of
657 bits, and "f" is flow control ("r" for RTS or
658 omit it). Default is "9600n8".
660 See Documentation/admin-guide/serial-console.rst for more
662 Documentation/networking/netconsole.rst for an
665 uart[8250],io,<addr>[,options]
666 uart[8250],mmio,<addr>[,options]
667 uart[8250],mmio16,<addr>[,options]
668 uart[8250],mmio32,<addr>[,options]
669 uart[8250],0x<addr>[,options]
670 Start an early, polled-mode console on the 8250/16550
671 UART at the specified I/O port or MMIO address,
672 switching to the matching ttyS device later.
673 MMIO inter-register address stride is either 8-bit
674 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
675 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
676 to be equivalent to 'mmio'. 'options' are specified in
677 the same format described for ttyS above; if unspecified,
678 the h/w is not re-initialized.
680 hvc<n> Use the hypervisor console device <n>. This is for
681 both Xen and PowerPC hypervisors.
683 If the device connected to the port is not a TTY but a braille
684 device, prepend "brl," before the device type, for instance
686 For now, only VisioBraille is supported.
689 [KNL] Change console messages format
691 By default we print messages on consoles in
692 "[time stamp] text\n" format (time stamp may not be
693 printed, depending on CONFIG_PRINTK_TIME or
694 `printk_time' param).
696 Switch to syslog format: "<%u>[time stamp] text\n"
697 IOW, each message will have a facility and loglevel
698 prefix. The format is similar to one used by syslog()
699 syscall, or to executing "dmesg -S --raw" or to reading
702 consoleblank= [KNL] The console blank (screen saver) timeout in
703 seconds. A value of 0 disables the blank timer.
707 [KNL] Change the default value for
708 /proc/<pid>/coredump_filter.
709 See also Documentation/filesystems/proc.rst.
711 coresight_cpu_debug.enable
714 Enable/disable the CPU sampling based debugging.
715 0: default value, disable debugging
716 1: enable debugging at boot time
718 cpuidle.off=1 [CPU_IDLE]
719 disable the cpuidle sub-system
722 [CPU_IDLE] Name of the cpuidle governor to use.
724 cpufreq.off=1 [CPU_FREQ]
725 disable the cpufreq sub-system
727 cpufreq.default_governor=
728 [CPU_FREQ] Name of the default cpufreq governor or
729 policy to use. This governor must be registered in the
730 kernel before the cpufreq driver probes.
733 [X86] Delay for N microsec between assert and de-assert
734 of APIC INIT to start processors. This delay occurs
735 on every CPU online, such as boot, and resume from suspend.
738 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
740 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
742 crashkernel=size[KMG][@offset[KMG]]
743 [KNL] Using kexec, Linux can switch to a 'crash kernel'
744 upon panic. This parameter reserves the physical
745 memory region [offset, offset + size] for that kernel
746 image. If '@offset' is omitted, then a suitable offset
747 is selected automatically.
748 [KNL, X86-64] Select a region under 4G first, and
749 fall back to reserve region above 4G when '@offset'
750 hasn't been specified.
751 See Documentation/admin-guide/kdump/kdump.rst for further details.
753 crashkernel=range1:size1[,range2:size2,...][@offset]
754 [KNL] Same as above, but depends on the memory
755 in the running system. The syntax of range is
756 start-[end] where start and end are both
757 a memory unit (amount[KMG]). See also
758 Documentation/admin-guide/kdump/kdump.rst for an example.
760 crashkernel=size[KMG],high
761 [KNL, X86-64] range could be above 4G. Allow kernel
762 to allocate physical memory region from top, so could
763 be above 4G if system have more than 4G ram installed.
764 Otherwise memory region will be allocated below 4G, if
766 It will be ignored if crashkernel=X is specified.
767 crashkernel=size[KMG],low
768 [KNL, X86-64] range under 4G. When crashkernel=X,high
769 is passed, kernel could allocate physical memory region
770 above 4G, that cause second kernel crash on system
771 that require some amount of low memory, e.g. swiotlb
772 requires at least 64M+32K low memory, also enough extra
773 low memory is needed to make sure DMA buffers for 32-bit
774 devices won't run out. Kernel would try to allocate at
775 at least 256M below 4G automatically.
776 This one let user to specify own low range under 4G
777 for second kernel instead.
778 0: to disable low allocation.
779 It will be ignored when crashkernel=X,high is not used
780 or memory reserved is below 4G.
783 [KNL] Disable crypto self-tests
788 cs89x0_media= [HW,NET]
789 Format: { rj45 | aui | bnc }
791 csdlock_debug= [KNL] Enable debug add-ons of cross-CPU function call
792 handling. When switched on, additional debug data is
793 printed to the console in case a hanging CPU is
794 detected, and that CPU is pinged again in order to try
795 to resolve the hang situation.
796 0: disable csdlock debugging (default)
797 1: enable basic csdlock debugging (minor impact)
798 ext: enable extended csdlock debugging (more impact,
802 See header of drivers/s390/block/dasd_devmap.c.
804 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
805 (one device per port)
806 Format: <port#>,<type>
807 See also Documentation/input/devices/joystick-parport.rst
809 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
811 Documentation/admin-guide/dynamic-debug-howto.rst for
812 details. Deprecated, see dyndbg.
814 debug [KNL] Enable kernel debugging (events log level).
817 [KNL] Enable printing [hashed] pointers early in the
818 boot sequence. If enabled, we use a weak hash instead
819 of siphash to hash pointers. Use this option if you are
820 seeing instances of '(___ptrval___)') and need to see a
821 value (hashed pointer) instead. Cryptographically
822 insecure, please do not use on production kernels.
825 [KNL] verbose locking self-tests
827 Print debugging info while doing the locking API
829 Bitmask for the various LOCKTYPE_ tests. Defaults to 0
830 (no extra messages), setting it to -1 (all bits set)
831 will print _a_lot_ more information - normally only
832 useful to lockdep developers.
834 debug_objects [KNL] Enable object debugging
837 [KNL] Disable object debugging
839 debug_guardpage_minorder=
840 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
841 parameter allows control of the order of pages that will
842 be intentionally kept free (and hence protected) by the
843 buddy allocator. Bigger value increase the probability
844 of catching random memory corruption, but reduce the
845 amount of memory for normal system use. The maximum
846 possible value is MAX_ORDER/2. Setting this parameter
847 to 1 or 2 should be enough to identify most random
848 memory corruption problems caused by bugs in kernel or
849 driver code when a CPU writes to (or reads from) a
850 random memory location. Note that there exists a class
851 of memory corruptions problems caused by buggy H/W or
852 F/W or by drivers badly programing DMA (basically when
853 memory is written at bus level and the CPU MMU is
854 bypassed) which are not detectable by
855 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
856 tracking down these problems.
859 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this parameter
860 enables the feature at boot time. By default, it is
861 disabled and the system will work mostly the same as a
862 kernel built without CONFIG_DEBUG_PAGEALLOC.
863 Note: to get most of debug_pagealloc error reports, it's
864 useful to also enable the page_owner functionality.
865 on: enable the feature
867 debugfs= [KNL] This parameter enables what is exposed to userspace
868 and debugfs internal clients.
869 Format: { on, no-mount, off }
870 on: All functions are enabled.
872 Filesystem is not registered but kernel clients can
873 access APIs and a crashkernel can be used to read
874 its content. There is nothing to mount.
875 off: Filesystem is not registered and clients
876 get a -EPERM as result when trying to register files
877 or directories within debugfs.
878 This is equivalent of the runtime functionality if
879 debugfs was not enabled in the kernel at all.
880 Default value is set in build-time with a kernel configuration.
882 debugpat [X86] Enable PAT debugging
884 decnet.addr= [HW,NET]
885 Format: <area>[,<node>]
886 See also Documentation/networking/decnet.rst.
889 [HW] The size of the default HugeTLB page. This is
890 the size represented by the legacy /proc/ hugepages
891 APIs. In addition, this is the default hugetlb size
892 used for shmget(), mmap() and mounting hugetlbfs
893 filesystems. If not specified, defaults to the
894 architecture's default huge page size. Huge page
895 sizes are architecture dependent. See also
896 Documentation/admin-guide/mm/hugetlbpage.rst.
899 deferred_probe_timeout=
900 [KNL] Debugging option to set a timeout in seconds for
901 deferred probe to give up waiting on dependencies to
902 probe. Only specific dependencies (subsystems or
903 drivers) that have opted in will be ignored. A timeout of 0
904 will timeout at the end of initcalls. This option will also
905 dump out devices still on the deferred probe list after
909 Format: { on | off | def_only | inf_only | always }
910 on: s390 zlib hardware support for compression on
911 level 1 and decompression (default)
912 off: No s390 zlib hardware support
913 def_only: s390 zlib hardware support for deflate
914 only (compression on level 1)
915 inf_only: s390 zlib hardware support for inflate
917 always: Same as 'on' but ignores the selected compression
918 level always using hardware support (used for debugging)
921 Set number of hash buckets for dentry cache.
923 disable_1tb_segments [PPC]
924 Disables the use of 1TB hash page table segments. This
925 causes the kernel to fall back to 256MB segments which
926 can be useful when debugging issues that require an SLB
930 Limits the number of kernel SLB entries, and flushes
931 them frequently to increase the rate of SLB faults
935 See Documentation/networking/ipv6.rst.
938 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
939 hardening is enabled for this boot. Hardened
940 usercopy checking is used to protect the kernel
941 from reading or writing beyond known memory
942 allocation boundaries as a proactive defense
943 against bounds-checking flaws in the kernel's
944 copy_to_user()/copy_from_user() interface.
945 on Perform hardened usercopy checks (default).
946 off Disable hardened usercopy checks.
949 Disable RADIX MMU mode on POWER9
951 radix_hcall_invalidate=on [PPC/PSERIES]
952 Disable RADIX GTSE feature and use hcall for TLB
956 Disable TLBIE instruction. Currently does not work
957 with KVM, with HASH MMU, or with coherent accelerators.
959 disable_cpu_apicid= [X86,APIC,SMP]
961 The number of initial APIC ID for the
962 corresponding CPU to be disabled at boot,
963 mostly used for the kdump 2nd kernel to
964 disable BSP to wake up multiple CPUs without
965 causing system reset or hang due to sending
968 disable_ddw [PPC/PSERIES]
969 Disable Dynamic DMA Window support. Use this
970 to workaround buggy firmware.
973 See Documentation/networking/ipv6.rst.
975 disable_mtrr_cleanup [X86]
976 The kernel tries to adjust MTRR layout from continuous
977 to discrete, to make X server driver able to add WB
978 entry later. This parameter disables that.
980 disable_mtrr_trim [X86, Intel and AMD only]
981 By default the kernel will trim any uncacheable
982 memory out of your available memory pool based on
983 MTRR settings. This parameter disables that behavior,
984 possibly causing your machine to run very slowly.
986 disable_timer_pin_1 [X86]
987 Disable PIN 1 of APIC timer
988 Can be useful to work around chipset bugs.
990 dis_ucode_ldr [X86] Disable the microcode loader.
992 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
993 this option disables the debugging code at boot.
995 dma_debug_entries=<number>
996 This option allows to tune the number of preallocated
997 entries for DMA-API debugging code. One entry is
998 required per DMA-API allocation. Use this if the
999 DMA-API debugging code disables itself because the
1000 architectural default is too low.
1002 dma_debug_driver=<driver_name>
1003 With this option the DMA-API debugging driver
1004 filter feature can be enabled at boot time. Just
1005 pass the driver to filter for as the parameter.
1006 The filter can be disabled or changed to another
1007 driver later using sysfs.
1009 driver_async_probe= [KNL]
1010 List of driver names to be probed asynchronously.
1011 Format: <driver_name1>,<driver_name2>...
1013 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
1014 Broken monitors, graphic adapters, KVMs and EDIDless
1015 panels may send no or incorrect EDID data sets.
1016 This parameter allows to specify an EDID data sets
1017 in the /lib/firmware directory that are used instead.
1018 Generic built-in EDID data sets are used, if one of
1019 edid/1024x768.bin, edid/1280x1024.bin,
1020 edid/1680x1050.bin, or edid/1920x1080.bin is given
1021 and no file with the same name exists. Details and
1022 instructions how to build your own EDID data are
1023 available in Documentation/admin-guide/edid.rst. An EDID
1024 data set will only be used for a particular connector,
1025 if its name and a colon are prepended to the EDID
1026 name. Each connector may use a unique EDID data
1027 set by separating the files with a comma. An EDID
1028 data set with no connector name will be used for
1029 any connectors not explicitly specified.
1034 Format: {"off" | "known"}
1035 Control how the dt_cpu_ftrs device-tree binding is
1036 used for CPU feature discovery and setup (if it
1038 off: Do not use it, fall back to legacy cpu table.
1039 known: Do not pass through unknown features to guests
1040 or userspace, only those that the kernel is aware of.
1042 dump_apple_properties [X86]
1043 Dump name and content of EFI device properties on
1044 x86 Macs. Useful for driver authors to determine
1045 what data is available or for reverse-engineering.
1047 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
1048 <module>.dyndbg[="val"]
1049 Enable debug messages at boot time. See
1050 Documentation/admin-guide/dynamic-debug-howto.rst
1053 nopku [X86] Disable Memory Protection Keys CPU feature found
1056 <module>.async_probe [KNL]
1057 Enable asynchronous probe on this module.
1059 early_ioremap_debug [KNL]
1060 Enable debug messages in early_ioremap support. This
1061 is useful for tracking down temporary early mappings
1062 which are not unmapped.
1064 earlycon= [KNL] Output early console device and options.
1066 When used with no options, the early console is
1067 determined by stdout-path property in device tree's
1068 chosen node or the ACPI SPCR table if supported by
1071 cdns,<addr>[,options]
1072 Start an early, polled-mode console on a Cadence
1073 (xuartps) serial port at the specified address. Only
1074 supported option is baud rate. If baud rate is not
1075 specified, the serial port must already be setup and
1078 uart[8250],io,<addr>[,options]
1079 uart[8250],mmio,<addr>[,options]
1080 uart[8250],mmio32,<addr>[,options]
1081 uart[8250],mmio32be,<addr>[,options]
1082 uart[8250],0x<addr>[,options]
1083 Start an early, polled-mode console on the 8250/16550
1084 UART at the specified I/O port or MMIO address.
1085 MMIO inter-register address stride is either 8-bit
1086 (mmio) or 32-bit (mmio32 or mmio32be).
1087 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
1088 to be equivalent to 'mmio'. 'options' are specified
1089 in the same format described for "console=ttyS<n>"; if
1090 unspecified, the h/w is not initialized.
1094 Start an early, polled-mode console on a pl011 serial
1095 port at the specified address. The pl011 serial port
1096 must already be setup and configured. Options are not
1097 yet supported. If 'mmio32' is specified, then only
1098 the driver will use only 32-bit accessors to read/write
1099 the device registers.
1102 Start an early, polled-mode console on a meson serial
1103 port at the specified address. The serial port must
1104 already be setup and configured. Options are not yet
1108 Start an early, polled-mode console on an msm serial
1109 port at the specified address. The serial port
1110 must already be setup and configured. Options are not
1113 msm_serial_dm,<addr>
1114 Start an early, polled-mode console on an msm serial
1115 dm port at the specified address. The serial port
1116 must already be setup and configured. Options are not
1120 Start an early, polled-mode console on a serial port
1121 of an Actions Semi SoC, such as S500 or S900, at the
1122 specified address. The serial port must already be
1123 setup and configured. Options are not yet supported.
1126 Start an early, polled-mode console on a serial port
1127 of an RDA Micro SoC, such as RDA8810PL, at the
1128 specified address. The serial port must already be
1129 setup and configured. Options are not yet supported.
1132 Use RISC-V SBI (Supervisor Binary Interface) for early
1135 smh Use ARM semihosting calls for early console.
1143 Use early console provided by serial driver available
1144 on Samsung SoCs, requires selecting proper type and
1145 a correct base address of the selected UART port. The
1146 serial port must already be setup and configured.
1147 Options are not yet supported.
1150 Start an early, polled-mode console on a lantiq serial
1151 (lqasc) port at the specified address. The serial port
1152 must already be setup and configured. Options are not
1157 Use early console provided by Freescale LP UART driver
1158 found on Freescale Vybrid and QorIQ LS1021A processors.
1159 A valid base address must be provided, and the serial
1160 port must already be setup and configured.
1164 Start an early, polled-mode, output-only console on the
1165 Freescale i.MX UART at the specified address. The UART
1166 must already be setup and configured.
1169 Start an early, polled-mode console on the
1170 Armada 3700 serial port at the specified
1171 address. The serial port must already be setup
1172 and configured. Options are not yet supported.
1175 Start an early, polled-mode console on a Qualcomm
1176 Generic Interface (GENI) based serial port at the
1177 specified address. The serial port must already be
1178 setup and configured. Options are not yet supported.
1181 Start an early, unaccelerated console on the EFI
1182 memory mapped framebuffer (if available). On cache
1183 coherent non-x86 systems that use system memory for
1184 the framebuffer, pass the 'ram' option so that it is
1185 mapped with the correct attributes.
1188 Use early console provided by Freescale LINFlexD UART
1189 serial driver for NXP S32V234 SoCs. A valid base
1190 address must be provided, and the serial port must
1191 already be setup and configured.
1193 earlyprintk= [X86,SH,ARM,M68k,S390]
1197 earlyprintk=serial[,ttySn[,baudrate]]
1198 earlyprintk=serial[,0x...[,baudrate]]
1199 earlyprintk=ttySn[,baudrate]
1200 earlyprintk=dbgp[debugController#]
1201 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1202 earlyprintk=xdbc[xhciController#]
1204 earlyprintk is useful when the kernel crashes before
1205 the normal console is initialized. It is not enabled by
1206 default because it has some cosmetic problems.
1208 Append ",keep" to not disable it when the real console
1211 Only one of vga, efi, serial, or usb debug port can
1214 Currently only ttyS0 and ttyS1 may be specified by
1215 name. Other I/O ports may be explicitly specified
1216 on some architectures (x86 and arm at least) by
1217 replacing ttySn with an I/O port address, like this:
1218 earlyprintk=serial,0x1008,115200
1219 You can find the port for a given device in
1220 /proc/tty/driver/serial:
1221 2: uart:ST16650V2 port:00001008 irq:18 ...
1223 Interaction with the standard serial driver is not
1226 The VGA and EFI output is eventually overwritten by
1229 The xen output can only be used by Xen PV guests.
1231 The sclp output can only be used on s390.
1233 The optional "force" to "pciserial" enables use of a
1234 PCI device even when its classcode is not of the
1237 edac_report= [HW,EDAC] Control how to report EDAC event
1238 Format: {"on" | "off" | "force"}
1239 on: enable EDAC to report H/W event. May be overridden
1240 by other higher priority error reporting module.
1241 off: disable H/W event reporting through EDAC.
1242 force: enforce the use of EDAC to report H/W event.
1245 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1248 This is designed to be used in conjunction with
1249 the boot argument: earlyprintk=vga
1251 This parameter works in place of the kgdboc parameter
1252 but can only be used if the backing tty is available
1253 very early in the boot process. For early debugging
1254 via a serial port see kgdboc_earlycon instead.
1257 Format: {"off" | "on" | "skip[mbr]"}
1260 Format: { "debug", "disable_early_pci_dma",
1261 "nochunk", "noruntime", "nosoftreserve",
1262 "novamap", "no_disable_early_pci_dma" }
1263 debug: enable misc debug output.
1264 disable_early_pci_dma: disable the busmaster bit on all
1265 PCI bridges while in the EFI boot stub.
1266 nochunk: disable reading files in "chunks" in the EFI
1267 boot stub, as chunking can cause problems with some
1268 firmware implementations.
1269 noruntime : disable EFI runtime services support
1270 nosoftreserve: The EFI_MEMORY_SP (Specific Purpose)
1271 attribute may cause the kernel to reserve the
1272 memory range for a memory mapping driver to
1273 claim. Specify efi=nosoftreserve to disable this
1274 reservation and treat the memory by its base type
1275 (i.e. EFI_CONVENTIONAL_MEMORY / "System RAM").
1276 novamap: do not call SetVirtualAddressMap().
1277 no_disable_early_pci_dma: Leave the busmaster bit set
1278 on all PCI bridges while in the EFI boot stub
1280 efi_no_storage_paranoia [EFI; X86]
1281 Using this parameter you can use more than 50% of
1282 your efi variable storage. Use this parameter only if
1283 you are really sure that your UEFI does sane gc and
1284 fulfills the spec otherwise your board may brick.
1286 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1287 Add arbitrary attribute to specific memory range by
1288 updating original EFI memory map.
1289 Region of memory which aa attribute is added to is
1292 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1293 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1294 attribute is added to range 0x100000000-0x180000000 and
1295 0x10a0000000-0x1120000000.
1297 If efi_fake_mem=8G@9G:0x40000 is specified, the
1298 EFI_MEMORY_SP(0x40000) attribute is added to
1299 range 0x240000000-0x43fffffff.
1301 Using this parameter you can do debugging of EFI memmap
1302 related features. For example, you can do debugging of
1303 Address Range Mirroring feature even if your box
1304 doesn't support it, or mark specific memory as
1307 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1308 that is to be dynamically loaded by Linux. If there are
1309 multiple variables with the same name but with different
1310 vendor GUIDs, all of them will be loaded. See
1311 Documentation/admin-guide/acpi/ssdt-overlays.rst for details.
1314 eisa_irq_edge= [PARISC,HW]
1315 See header of drivers/parisc/eisa.c.
1318 See comment before function elanfreq_setup() in
1319 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1321 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1322 Specifies physical address of start of kernel core
1323 image elf header and optionally the size. Generally
1324 kexec loader will pass this option to capture kernel.
1325 See Documentation/admin-guide/kdump/kdump.rst for details.
1327 enable_mtrr_cleanup [X86]
1328 The kernel tries to adjust MTRR layout from continuous
1329 to discrete, to make X server driver able to add WB
1330 entry later. This parameter enables that.
1332 enable_timer_pin_1 [X86]
1333 Enable PIN 1 of APIC timer
1334 Can be useful to work around chipset bugs
1335 (in particular on some ATI chipsets).
1336 The kernel tries to set a reasonable default.
1338 enforcing [SELINUX] Set initial enforcing status.
1340 See security/selinux/Kconfig help text.
1341 0 -- permissive (log only, no denials).
1342 1 -- enforcing (deny and log).
1344 Value can be changed at runtime via
1345 /sys/fs/selinux/enforce.
1348 Disable Error Record Serialization Table (ERST)
1351 ether= [HW,NET] Ethernet cards parameters
1352 This option is obsoleted by the "netdev=" option, which
1353 has equivalent usage. See its documentation for details.
1357 Permit 'security.evm' to be updated regardless of
1358 current integrity status.
1363 fail_make_request=[KNL]
1364 General fault injection mechanism.
1365 Format: <interval>,<probability>,<space>,<times>
1366 See also Documentation/fault-injection/.
1369 Format: { initns | none }
1370 See Documentation/admin-guide/sysctl/net.rst for
1371 fb_tunnels_only_for_init_ns
1374 See Documentation/admin-guide/blockdev/floppy.rst.
1376 force_pal_cache_flush
1377 [IA-64] Avoid check_sal_cache_flush which may hang on
1378 buggy SAL_CACHE_FLUSH implementations. Using this
1379 parameter will force ia64_sal_cache_flush to call
1380 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1383 Forcefully enable Physical Address Extension (PAE).
1384 Many Pentium M systems disable PAE but may have a
1385 functionally usable PAE implementation.
1386 Warning: use of this parameter will taint the kernel
1387 and may cause unknown problems.
1390 [FTRACE] will set and start the specified tracer
1391 as early as possible in order to facilitate early
1394 ftrace_dump_on_oops[=orig_cpu]
1395 [FTRACE] will dump the trace buffers on oops.
1396 If no parameter is passed, ftrace will dump
1397 buffers of all CPUs, but if you pass orig_cpu, it will
1398 dump only the buffer of the CPU that triggered the
1401 ftrace_filter=[function-list]
1402 [FTRACE] Limit the functions traced by the function
1403 tracer at boot up. function-list is a comma-separated
1404 list of functions. This list can be changed at run
1405 time by the set_ftrace_filter file in the debugfs
1408 ftrace_notrace=[function-list]
1409 [FTRACE] Do not trace the functions specified in
1410 function-list. This list can be changed at run time
1411 by the set_ftrace_notrace file in the debugfs
1414 ftrace_graph_filter=[function-list]
1415 [FTRACE] Limit the top level callers functions traced
1416 by the function graph tracer at boot up.
1417 function-list is a comma-separated list of functions
1418 that can be changed at run time by the
1419 set_graph_function file in the debugfs tracing directory.
1421 ftrace_graph_notrace=[function-list]
1422 [FTRACE] Do not trace from the functions specified in
1423 function-list. This list is a comma-separated list of
1424 functions that can be changed at run time by the
1425 set_graph_notrace file in the debugfs tracing directory.
1427 ftrace_graph_max_depth=<uint>
1428 [FTRACE] Used with the function graph tracer. This is
1429 the max depth it will trace into a function. This value
1430 can be changed at run time by the max_graph_depth file
1431 in the tracefs tracing directory. default: 0 (no limit)
1433 fw_devlink= [KNL] Create device links between consumer and supplier
1434 devices by scanning the firmware to infer the
1435 consumer/supplier relationships. This feature is
1436 especially useful when drivers are loaded as modules as
1437 it ensures proper ordering of tasks like device probing
1438 (suppliers first, then consumers), supplier boot state
1439 clean up (only after all consumers have probed),
1440 suspend/resume & runtime PM (consumers first, then
1442 Format: { off | permissive | on | rpm }
1443 off -- Don't create device links from firmware info.
1444 permissive -- Create device links from firmware info
1445 but use it only for ordering boot state clean
1446 up (sync_state() calls).
1447 on -- Create device links from firmware info and use it
1448 to enforce probe and suspend/resume ordering.
1449 rpm -- Like "on", but also use to order runtime PM.
1451 fw_devlink.strict=<bool>
1452 [KNL] Treat all inferred dependencies as mandatory
1453 dependencies. This only applies for fw_devlink=on|rpm.
1457 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1458 support via parallel port (up to 5 devices per port)
1459 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1460 See also Documentation/input/devices/joystick-parport.rst
1464 gart_fix_e820= [X86-64] disable the fix e820 for K8 GART
1468 gcov_persist= [GCOV] When non-zero (default), profiling data for
1469 kernel modules is saved and remains accessible via
1470 debugfs, even when the module is unloaded/reloaded.
1471 When zero, profiling data is discarded and associated
1472 debugfs files are removed at module unload time.
1474 goldfish [X86] Enable the goldfish android emulator platform.
1475 Don't use this when you are not running on the
1478 gpio-mockup.gpio_mockup_ranges
1479 [HW] Sets the ranges of gpiochip of for this device.
1480 Format: <start1>,<end1>,<start2>,<end2>...
1481 gpio-mockup.gpio_mockup_named_lines
1482 [HW] Let the driver know GPIO lines should be named.
1484 gpt [EFI] Forces disk with valid GPT signature but
1485 invalid Protective MBR to be treated as GPT. If the
1486 primary GPT is corrupted, it enables the backup/alternate
1487 GPT to be used instead.
1489 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1490 the "Enable 0" bit of the configuration register.
1493 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1494 the "Enable 0" bit of the configuration register.
1497 grcan.select= [HW] Select which physical interface to use.
1500 grcan.txsize= [HW] Sets the size of the tx buffer.
1501 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1503 grcan.rxsize= [HW] Sets the size of the rx buffer.
1504 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1507 hardlockup_all_cpu_backtrace=
1508 [KNL] Should the hard-lockup detector generate
1509 backtraces on all cpus.
1512 hashdist= [KNL,NUMA] Large hashes allocated during boot
1513 are distributed across NUMA nodes. Defaults on
1514 for 64-bit NUMA, off otherwise.
1515 Format: 0 | 1 (for off | on)
1517 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1519 hd= [EIDE] (E)IDE hard drive subsystem geometry
1520 Format: <cyl>,<head>,<sect>
1523 Disable Hardware Error Source Table (HEST) support;
1524 corresponding firmware-first mode error processing
1525 logic will be disabled.
1527 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1528 size of <nn>. This works even on boxes that have no
1529 highmem otherwise. This also works to reduce highmem
1530 size on bigger boxes.
1532 highres= [KNL] Enable/disable high resolution timer mode.
1533 Valid parameters: "on", "off"
1538 hpet= [X86-32,HPET] option to control HPET usage
1539 Format: { enable (default) | disable | force |
1541 disable: disable HPET and use PIT instead
1542 force: allow force enabled of undocumented chips (ICH4,
1544 verbose: show contents of HPET registers during setup
1546 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1547 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1549 hugetlb_cma= [HW,CMA] The size of a CMA area used for allocation
1550 of gigantic hugepages.
1553 Reserve a CMA area of given size and allocate gigantic
1554 hugepages using the CMA allocator. If enabled, the
1555 boot-time allocation of gigantic hugepages is skipped.
1557 hugepages= [HW] Number of HugeTLB pages to allocate at boot.
1558 If this follows hugepagesz (below), it specifies
1559 the number of pages of hugepagesz to be allocated.
1560 If this is the first HugeTLB parameter on the command
1561 line, it specifies the number of pages to allocate for
1562 the default huge page size. See also
1563 Documentation/admin-guide/mm/hugetlbpage.rst.
1567 [HW] The size of the HugeTLB pages. This is used in
1568 conjunction with hugepages (above) to allocate huge
1569 pages of a specific size at boot. The pair
1570 hugepagesz=X hugepages=Y can be specified once for
1571 each supported huge page size. Huge page sizes are
1572 architecture dependent. See also
1573 Documentation/admin-guide/mm/hugetlbpage.rst.
1577 [KNL] Should the hung task detector generate panics.
1580 A value of 1 instructs the kernel to panic when a
1581 hung task is detected. The default value is controlled
1582 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1583 option. The value selected by this boot parameter can
1584 be changed later by the kernel.hung_task_panic sysctl.
1586 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1587 terminal devices. Valid values: 0..8
1588 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1589 If specified, z/VM IUCV HVC accepts connections
1590 from listed z/VM user IDs only.
1592 hv_nopvspin [X86,HYPER_V] Disables the paravirt spinlock optimizations
1593 which allow the hypervisor to 'idle' the
1594 guest on lock contention.
1597 Do not unregister boot console at start. This is only
1598 useful for debugging when something happens in the window
1599 between unregistering the boot console and initializing
1602 i2c_bus= [HW] Override the default board specific I2C bus speed
1603 or register an additional I2C bus that is not
1604 registered from board initialization code.
1608 i8042.debug [HW] Toggle i8042 debug mode
1609 i8042.unmask_kbd_data
1610 [HW] Enable printing of interrupt data from the KBD port
1611 (disabled by default, and as a pre-condition
1612 requires that i8042.debug=1 be enabled)
1613 i8042.direct [HW] Put keyboard port into non-translated mode
1614 i8042.dumbkbd [HW] Pretend that controller can only read data from
1615 keyboard and cannot control its state
1616 (Don't attempt to blink the leds)
1617 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1618 i8042.nokbd [HW] Don't check/create keyboard port
1619 i8042.noloop [HW] Disable the AUX Loopback command while probing
1621 i8042.nomux [HW] Don't check presence of an active multiplexing
1623 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1625 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1626 i8042.reset [HW] Reset the controller during init, cleanup and
1627 suspend-to-ram transitions, only during s2r
1628 transitions, or never reset
1629 Format: { 1 | Y | y | 0 | N | n }
1630 1, Y, y: always reset controller
1631 0, N, n: don't ever reset controller
1632 Default: only on s2r transitions on x86; most other
1633 architectures force reset to be always executed
1634 i8042.unlock [HW] Unlock (ignore) the keylock
1635 i8042.kbdreset [HW] Reset device connected to KBD port
1639 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1640 indicates that the driver is running on unsupported
1642 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1643 does not match list of supported models.
1645 [HW] Report power status in /proc/i8k
1646 (disabled by default)
1647 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1650 i915.invert_brightness=
1651 [DRM] Invert the sense of the variable that is used to
1652 set the brightness of the panel backlight. Normally a
1653 brightness value of 0 indicates backlight switched off,
1654 and the maximum of the brightness value sets the backlight
1655 to maximum brightness. If this parameter is set to 0
1656 (default) and the machine requires it, or this parameter
1657 is set to 1, a brightness value of 0 sets the backlight
1658 to maximum brightness, and the maximum of the brightness
1659 value switches the backlight off.
1660 -1 -- never invert brightness
1661 0 -- machine default
1662 1 -- force brightness inversion
1665 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1667 ide-core.nodma= [HW] (E)IDE subsystem
1668 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1669 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1670 .cdrom .chs .ignore_cable are additional options
1671 See Documentation/ide/ide.rst.
1673 ide-generic.probe-mask= [HW] (E)IDE subsystem
1675 Probe mask for legacy ISA IDE ports. Depending on
1676 platform up to 6 ports are supported, enabled by
1677 setting corresponding bits in the mask to 1. The
1678 default value is 0x0, which has a special meaning.
1679 On systems that have PCI, it triggers scanning the
1680 PCI bus for the first and the second port, which
1681 are then probed. On systems without PCI the value
1682 of 0x0 enables probing the two first ports as if it
1685 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1686 Claim all unknown PCI IDE storage controllers.
1689 Format: idle=poll, idle=halt, idle=nomwait
1690 Poll forces a polling idle loop that can slightly
1691 improve the performance of waking up a idle CPU, but
1692 will use a lot of power and make the system run hot.
1694 idle=halt: Halt is forced to be used for CPU idle.
1695 In such case C2/C3 won't be used again.
1696 idle=nomwait: Disable mwait for CPU C-states
1700 Allow force disabling of Shared Virtual Memory (SVA)
1701 support for the idxd driver. By default it is set to
1704 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1705 Format: { strict | legacy | 2008 | relaxed }
1708 Choose which programs will be accepted for execution
1709 based on the IEEE 754 NaN encoding(s) supported by
1710 the FPU and the NaN encoding requested with the value
1711 of an ELF file header flag individually set by each
1712 binary. Hardware implementations are permitted to
1713 support either or both of the legacy and the 2008 NaN
1716 Available settings are as follows:
1717 strict accept binaries that request a NaN encoding
1718 supported by the FPU
1719 legacy only accept legacy-NaN binaries, if supported
1721 2008 only accept 2008-NaN binaries, if supported
1723 relaxed accept any binaries regardless of whether
1724 supported by the FPU
1726 The FPU emulator is always able to support both NaN
1727 encodings, so if no FPU hardware is present or it has
1728 been disabled with 'nofpu', then the settings of
1729 'legacy' and '2008' strap the emulator accordingly,
1730 'relaxed' straps the emulator for both legacy-NaN and
1731 2008-NaN, whereas 'strict' enables legacy-NaN only on
1732 legacy processors and both NaN encodings on MIPS32 or
1735 The setting for ABS.fmt/NEG.fmt instruction execution
1736 mode generally follows that for the NaN encoding,
1737 except where unsupported by hardware.
1739 ignore_loglevel [KNL]
1740 Ignore loglevel setting - this will print /all/
1741 kernel messages to the console. Useful for debugging.
1742 We also add it as printk module parameter, so users
1743 could change it dynamically, usually by
1744 /sys/module/printk/parameters/ignore_loglevel.
1747 Ignore RLIMIT_DATA setting for data mappings,
1748 print warning at first misuse. Can be changed via
1749 /sys/module/kernel/parameters/ignore_rlimit_data.
1751 ihash_entries= [KNL]
1752 Set number of hash buckets for inode cache.
1754 ima_appraise= [IMA] appraise integrity measurements
1755 Format: { "off" | "enforce" | "fix" | "log" }
1758 ima_appraise_tcb [IMA] Deprecated. Use ima_policy= instead.
1759 The builtin appraise policy appraises all files
1762 ima_canonical_fmt [IMA]
1763 Use the canonical format for the binary runtime
1764 measurements, instead of host native format.
1767 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1771 The list of supported hash algorithms is defined
1772 in crypto/hash_info.h.
1775 The builtin policies to load during IMA setup.
1776 Format: "tcb | appraise_tcb | secure_boot |
1777 fail_securely | critical_data"
1779 The "tcb" policy measures all programs exec'd, files
1780 mmap'd for exec, and all files opened with the read
1781 mode bit set by either the effective uid (euid=0) or
1784 The "appraise_tcb" policy appraises the integrity of
1785 all files owned by root.
1787 The "secure_boot" policy appraises the integrity
1788 of files (eg. kexec kernel image, kernel modules,
1789 firmware, policy, etc) based on file signatures.
1791 The "fail_securely" policy forces file signature
1792 verification failure also on privileged mounted
1793 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1796 The "critical_data" policy measures kernel integrity
1799 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1800 Load a policy which meets the needs of the Trusted
1801 Computing Base. This means IMA will measure all
1802 programs exec'd, files mmap'd for exec, and all files
1803 opened for read by uid=0.
1806 Select one of defined IMA measurements template formats.
1807 Formats: { "ima" | "ima-ng" | "ima-sig" }
1811 [IMA] Define a custom template format.
1812 Format: { "field1|...|fieldN" }
1814 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1815 Format: <min_file_size>
1816 Set the minimal file size for using asynchronous hash.
1817 If left unspecified, ahash usage is disabled.
1819 ahash performance varies for different data sizes on
1820 different crypto accelerators. This option can be used
1821 to achieve the best performance for a particular HW.
1823 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1825 Set hashing buffer size. Default: 4k.
1827 ahash performance varies for different chunk sizes on
1828 different crypto accelerators. This option can be used
1829 to achieve best performance for particular HW.
1833 Run specified binary instead of /sbin/init as init
1836 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1837 for working out where the kernel is dying during
1840 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1841 initcall functions. Useful for debugging built-in
1842 modules and initcalls.
1844 initramfs_async= [KNL]
1847 This parameter controls whether the initramfs
1848 image is unpacked asynchronously, concurrently
1849 with devices being probed and
1850 initialized. This should normally just work,
1851 but as a debugging aid, one can get the
1852 historical behaviour of the initramfs
1853 unpacking being completed before device_ and
1856 initrd= [BOOT] Specify the location of the initial ramdisk
1858 initrdmem= [KNL] Specify a physical address and size from which to
1859 load the initrd. If an initrd is compiled in or
1860 specified in the bootparams, it takes priority over this
1862 Format: ss[KMG],nn[KMG]
1865 init_on_alloc= [MM] Fill newly allocated pages and heap objects with
1868 Default set by CONFIG_INIT_ON_ALLOC_DEFAULT_ON.
1870 init_on_free= [MM] Fill freed pages and heap objects with zeroes.
1872 Default set by CONFIG_INIT_ON_FREE_DEFAULT_ON.
1874 init_pkru= [X86] Specify the default memory protection keys rights
1875 register contents for all processes. 0x55555554 by
1876 default (disallow access to all but pkey 0). Can
1877 override in debugfs after boot.
1879 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1882 int_pln_enable [X86] Enable power limit notification interrupt
1884 integrity_audit=[IMA]
1885 Format: { "0" | "1" }
1886 0 -- basic integrity auditing messages. (Default)
1887 1 -- additional integrity auditing messages.
1889 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1891 Enable intel iommu driver.
1893 Disable intel iommu driver.
1894 igfx_off [Default Off]
1895 By default, gfx is mapped as normal device. If a gfx
1896 device has a dedicated DMAR unit, the DMAR unit is
1897 bypassed by not enabling DMAR with this option. In
1898 this case, gfx device will use physical address for
1900 strict [Default Off]
1901 With this option on every unmap_single operation will
1902 result in a hardware IOTLB flush operation as opposed
1903 to batching them for performance.
1904 sp_off [Default Off]
1905 By default, super page will be supported if Intel IOMMU
1906 has the capability. With this option, super page will
1909 By default, scalable mode will be disabled even if the
1910 hardware advertises that it has support for the scalable
1911 mode translation. With this option set, scalable mode
1912 will be used on hardware which claims to support it.
1913 tboot_noforce [Default Off]
1914 Do not force the Intel IOMMU enabled under tboot.
1915 By default, tboot will force Intel IOMMU on, which
1916 could harm performance of some high-throughput
1917 devices like 40GBit network cards, even if identity
1919 Note that using this option lowers the security
1920 provided by tboot because it makes the system
1921 vulnerable to DMA attacks.
1923 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1924 0 disables intel_idle and fall back on acpi_idle.
1925 1 to 9 specify maximum depth of C-state.
1929 Do not enable intel_pstate as the default
1930 scaling driver for the supported processors
1932 Use intel_pstate as a scaling driver, but configure it
1933 to work with generic cpufreq governors (instead of
1934 enabling its internal governor). This mode cannot be
1935 used along with the hardware-managed P-states (HWP)
1938 Enable intel_pstate on systems that prohibit it by default
1939 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1940 instead of acpi-cpufreq may disable platform features, such
1941 as thermal controls and power capping, that rely on ACPI
1942 P-States information being indicated to OSPM and therefore
1943 should be used with caution. This option does not work with
1944 processors that aren't supported by the intel_pstate driver
1945 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1947 Do not enable hardware P state control (HWP)
1950 Only load intel_pstate on systems which support
1951 hardware P state control (HWP) if available.
1953 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1954 Description Table, specifies preferred power management
1955 profile as "Enterprise Server" or "Performance Server",
1956 then this feature is turned on by default.
1958 Allow per-logical-CPU P-State performance control limits using
1959 cpufreq sysfs interface
1961 intremap= [X86-64, Intel-IOMMU]
1962 on enable Interrupt Remapping (default)
1963 off disable Interrupt Remapping
1964 nosid disable Source ID checking
1966 BIOS x2APIC opt-out request will be ignored
1967 nopost disable Interrupt Posting
1969 iomem= Disable strict checking of access to MMIO memory
1970 strict regions from userspace.
1985 nobypass [PPC/POWERNV]
1986 Disable IOMMU bypass, using IOMMU for PCI devices.
1988 iommu.forcedac= [ARM64, X86] Control IOVA allocation for PCI devices.
1989 Format: { "0" | "1" }
1990 0 - Try to allocate a 32-bit DMA address first, before
1991 falling back to the full range if needed.
1992 1 - Allocate directly from the full usable range,
1993 forcing Dual Address Cycle for PCI cards supporting
1994 greater than 32-bit addressing.
1996 iommu.strict= [ARM64] Configure TLB invalidation behaviour
1997 Format: { "0" | "1" }
1999 Request that DMA unmap operations use deferred
2000 invalidation of hardware TLBs, for increased
2001 throughput at the cost of reduced device isolation.
2002 Will fall back to strict mode if not supported by
2003 the relevant IOMMU driver.
2004 1 - Strict mode (default).
2005 DMA unmap operations invalidate IOMMU hardware TLBs
2009 [ARM64, X86] Configure DMA to bypass the IOMMU by default.
2010 Format: { "0" | "1" }
2011 0 - Use IOMMU translation for DMA.
2012 1 - Bypass the IOMMU for DMA.
2013 unset - Use value of CONFIG_IOMMU_DEFAULT_PASSTHROUGH.
2015 io7= [HW] IO7 for Marvel-based Alpha systems
2016 See comment before marvel_specify_io7 in
2017 arch/alpha/kernel/core_marvel.c.
2019 io_delay= [X86] I/O delay method
2021 Standard port 0x80 based delay
2023 Alternate port 0xed based delay (needed on some systems)
2025 Simple two microseconds delay
2030 See Documentation/admin-guide/nfs/nfsroot.rst.
2032 ipcmni_extend [KNL] Extend the maximum number of unique System V
2033 IPC identifiers from 32,768 to 16,777,216.
2035 irqaffinity= [SMP] Set the default irq affinity mask
2036 The argument is a cpu list, as described above.
2038 irqchip.gicv2_force_probe=
2041 Force the kernel to look for the second 4kB page
2042 of a GICv2 controller even if the memory range
2043 exposed by the device tree is too small.
2045 irqchip.gicv3_nolpi=
2047 Force the kernel to ignore the availability of
2048 LPIs (and by consequence ITSs). Intended for system
2049 that use the kernel as a bootloader, and thus want
2050 to let secondary kernels in charge of setting up
2053 irqchip.gicv3_pseudo_nmi= [ARM64]
2054 Enables support for pseudo-NMIs in the kernel. This
2055 requires the kernel to be built with
2056 CONFIG_ARM64_PSEUDO_NMI.
2059 When an interrupt is not handled search all handlers
2060 for it. Intended to get systems with badly broken
2064 When an interrupt is not handled search all handlers
2065 for it. Also check all handlers each timer
2066 interrupt. Intended to get systems with badly broken
2070 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
2072 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
2073 [Deprecated - use cpusets instead]
2074 Format: [flag-list,]<cpu-list>
2076 Specify one or more CPUs to isolate from disturbances
2077 specified in the flag list (default: domain):
2080 Disable the tick when a single task runs.
2082 A residual 1Hz tick is offloaded to workqueues, which you
2083 need to affine to housekeeping through the global
2084 workqueue's affinity configured via the
2085 /sys/devices/virtual/workqueue/cpumask sysfs file, or
2086 by using the 'domain' flag described below.
2088 NOTE: by default the global workqueue runs on all CPUs,
2089 so to protect individual CPUs the 'cpumask' file has to
2090 be configured manually after bootup.
2093 Isolate from the general SMP balancing and scheduling
2094 algorithms. Note that performing domain isolation this way
2095 is irreversible: it's not possible to bring back a CPU to
2096 the domains once isolated through isolcpus. It's strongly
2097 advised to use cpusets instead to disable scheduler load
2098 balancing through the "cpuset.sched_load_balance" file.
2099 It offers a much more flexible interface where CPUs can
2100 move in and out of an isolated set anytime.
2102 You can move a process onto or off an "isolated" CPU via
2103 the CPU affinity syscalls or cpuset.
2104 <cpu number> begins at 0 and the maximum value is
2105 "number of CPUs in system - 1".
2109 Isolate from being targeted by managed interrupts
2110 which have an interrupt mask containing isolated
2111 CPUs. The affinity of managed interrupts is
2112 handled by the kernel and cannot be changed via
2113 the /proc/irq/* interfaces.
2115 This isolation is best effort and only effective
2116 if the automatically assigned interrupt mask of a
2117 device queue contains isolated and housekeeping
2118 CPUs. If housekeeping CPUs are online then such
2119 interrupts are directed to the housekeeping CPU
2120 so that IO submitted on the housekeeping CPU
2121 cannot disturb the isolated CPU.
2123 If a queue's affinity mask contains only isolated
2124 CPUs then this parameter has no effect on the
2125 interrupt routing decision, though interrupts are
2126 only delivered when tasks running on those
2127 isolated CPUs submit IO. IO submitted on
2128 housekeeping CPUs has no influence on those
2131 The format of <cpu-list> is described above.
2135 ivrs_ioapic [HW,X86-64]
2136 Provide an override to the IOAPIC-ID<->DEVICE-ID
2137 mapping provided in the IVRS ACPI table. For
2138 example, to map IOAPIC-ID decimal 10 to
2139 PCI device 00:14.0 write the parameter as:
2140 ivrs_ioapic[10]=00:14.0
2142 ivrs_hpet [HW,X86-64]
2143 Provide an override to the HPET-ID<->DEVICE-ID
2144 mapping provided in the IVRS ACPI table. For
2145 example, to map HPET-ID decimal 0 to
2146 PCI device 00:14.0 write the parameter as:
2147 ivrs_hpet[0]=00:14.0
2149 ivrs_acpihid [HW,X86-64]
2150 Provide an override to the ACPI-HID:UID<->DEVICE-ID
2151 mapping provided in the IVRS ACPI table. For
2152 example, to map UART-HID:UID AMD0020:0 to
2153 PCI device 00:14.5 write the parameter as:
2154 ivrs_acpihid[00:14.5]=AMD0020:0
2156 js= [HW,JOY] Analog joystick
2157 See Documentation/input/joydev/joystick.rst.
2160 When CONFIG_RANDOMIZE_BASE is set, this disables
2161 kernel and module base offset ASLR (Address Space
2162 Layout Randomization).
2165 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
2166 report on every invalid memory access. Without this
2167 parameter KASAN will print report only for the first
2172 kernelcore= [KNL,X86,IA-64,PPC]
2173 Format: nn[KMGTPE] | nn% | "mirror"
2174 This parameter specifies the amount of memory usable by
2175 the kernel for non-movable allocations. The requested
2176 amount is spread evenly throughout all nodes in the
2177 system as ZONE_NORMAL. The remaining memory is used for
2178 movable memory in its own zone, ZONE_MOVABLE. In the
2179 event, a node is too small to have both ZONE_NORMAL and
2180 ZONE_MOVABLE, kernelcore memory will take priority and
2181 other nodes will have a larger ZONE_MOVABLE.
2183 ZONE_MOVABLE is used for the allocation of pages that
2184 may be reclaimed or moved by the page migration
2185 subsystem. Note that allocations like PTEs-from-HighMem
2186 still use the HighMem zone if it exists, and the Normal
2187 zone if it does not.
2189 It is possible to specify the exact amount of memory in
2190 the form of "nn[KMGTPE]", a percentage of total system
2191 memory in the form of "nn%", or "mirror". If "mirror"
2192 option is specified, mirrored (reliable) memory is used
2193 for non-movable allocations and remaining memory is used
2194 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
2195 are exclusive, so you cannot specify multiple forms.
2197 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
2198 Format: <Controller#>[,poll interval]
2199 The controller # is the number of the ehci usb debug
2200 port as it is probed via PCI. The poll interval is
2201 optional and is the number seconds in between
2202 each poll cycle to the debug port in case you need
2203 the functionality for interrupting the kernel with
2204 gdb or control-c on the dbgp connection. When
2205 not using this parameter you use sysrq-g to break into
2206 the kernel debugger.
2208 kgdboc= [KGDB,HW] kgdb over consoles.
2209 Requires a tty driver that supports console polling,
2210 or a supported polling keyboard driver (non-usb).
2211 Serial only format: <serial_device>[,baud]
2212 keyboard only format: kbd
2213 keyboard and serial format: kbd,<serial_device>[,baud]
2214 Optional Kernel mode setting:
2215 kms, kbd format: kms,kbd
2216 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
2218 kgdboc_earlycon= [KGDB,HW]
2219 If the boot console provides the ability to read
2220 characters and can work in polling mode, you can use
2221 this parameter to tell kgdb to use it as a backend
2222 until the normal console is registered. Intended to
2223 be used together with the kgdboc parameter which
2224 specifies the normal console to transition to.
2226 The name of the early console should be specified
2227 as the value of this parameter. Note that the name of
2228 the early console might be different than the tty
2229 name passed to kgdboc. It's OK to leave the value
2230 blank and the first boot console that implements
2231 read() will be picked.
2233 kgdbwait [KGDB] Stop kernel execution and enter the
2234 kernel debugger at the earliest opportunity.
2236 kmac= [MIPS] Korina ethernet MAC address.
2237 Configure the RouterBoard 532 series on-chip
2238 Ethernet adapter MAC address.
2240 kmemleak= [KNL] Boot-time kmemleak enable/disable
2241 Valid arguments: on, off
2243 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
2246 kprobe_event=[probe-list]
2247 [FTRACE] Add kprobe events and enable at boot time.
2248 The probe-list is a semicolon delimited list of probe
2249 definitions. Each definition is same as kprobe_events
2250 interface, but the parameters are comma delimited.
2251 For example, to add a kprobe event on vfs_read with
2252 arg1 and arg2, add to the command line;
2254 kprobe_event=p,vfs_read,$arg1,$arg2
2256 See also Documentation/trace/kprobetrace.rst "Kernel
2257 Boot Parameter" section.
2259 kpti= [ARM64] Control page table isolation of user
2260 and kernel address spaces.
2261 Default: enabled on cores which need mitigation.
2265 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
2266 Default is 0 (don't ignore, but inject #GP)
2268 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
2269 Default is false (don't support).
2271 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
2276 [KVM] Controls the software workaround for the
2277 X86_BUG_ITLB_MULTIHIT bug.
2278 force : Always deploy workaround.
2279 off : Never deploy workaround.
2280 auto : Deploy workaround based on the presence of
2281 X86_BUG_ITLB_MULTIHIT.
2285 If the software workaround is enabled for the host,
2286 guests do need not to enable it for nested guests.
2288 kvm.nx_huge_pages_recovery_ratio=
2289 [KVM] Controls how many 4KiB pages are periodically zapped
2290 back to huge pages. 0 disables the recovery, otherwise if
2291 the value is N KVM will zap 1/Nth of the 4KiB pages every
2292 minute. The default is 60.
2294 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
2295 Default is 1 (enabled)
2297 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
2299 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
2302 [KVM,ARM] Select one of KVM/arm64's modes of operation.
2304 nvhe: Standard nVHE-based mode, without support for
2307 protected: nVHE-based mode with support for guests whose
2308 state is kept private from the host.
2309 Not valid if the kernel is running in EL2.
2311 Defaults to VHE/nVHE based on hardware support.
2313 kvm-arm.vgic_v3_group0_trap=
2314 [KVM,ARM] Trap guest accesses to GICv3 group-0
2317 kvm-arm.vgic_v3_group1_trap=
2318 [KVM,ARM] Trap guest accesses to GICv3 group-1
2321 kvm-arm.vgic_v3_common_trap=
2322 [KVM,ARM] Trap guest accesses to GICv3 common
2325 kvm-arm.vgic_v4_enable=
2326 [KVM,ARM] Allow use of GICv4 for direct injection of
2329 kvm_cma_resv_ratio=n [PPC]
2330 Reserves given percentage from system memory area for
2331 contiguous memory allocation for KVM hash pagetable
2333 By default it reserves 5% of total system memory.
2337 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2338 (virtualized MMU) support on capable Intel chips.
2339 Default is 1 (enabled)
2341 kvm-intel.emulate_invalid_guest_state=
2342 [KVM,Intel] Enable emulation of invalid guest states
2343 Default is 0 (disabled)
2345 kvm-intel.flexpriority=
2346 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2347 Default is 1 (enabled)
2350 [KVM,Intel] Enable VMX nesting (nVMX).
2351 Default is 0 (disabled)
2353 kvm-intel.unrestricted_guest=
2354 [KVM,Intel] Disable unrestricted guest feature
2355 (virtualized real and unpaged mode) on capable
2356 Intel chips. Default is 1 (enabled)
2358 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2361 Valid arguments: never, cond, always
2363 always: L1D cache flush on every VMENTER.
2364 cond: Flush L1D on VMENTER only when the code between
2365 VMEXIT and VMENTER can leak host memory.
2366 never: Disables the mitigation
2368 Default is cond (do L1 cache flush in specific instances)
2370 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2371 feature (tagged TLBs) on capable Intel chips.
2372 Default is 1 (enabled)
2374 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2377 The kernel PTE inversion protection is unconditionally
2378 enabled and cannot be disabled.
2381 Provides all available mitigations for the
2382 L1TF vulnerability. Disables SMT and
2383 enables all mitigations in the
2384 hypervisors, i.e. unconditional L1D flush.
2386 SMT control and L1D flush control via the
2387 sysfs interface is still possible after
2388 boot. Hypervisors will issue a warning
2389 when the first VM is started in a
2390 potentially insecure configuration,
2391 i.e. SMT enabled or L1D flush disabled.
2394 Same as 'full', but disables SMT and L1D
2395 flush runtime control. Implies the
2396 'nosmt=force' command line option.
2397 (i.e. sysfs control of SMT is disabled.)
2400 Leaves SMT enabled and enables the default
2401 hypervisor mitigation, i.e. conditional
2404 SMT control and L1D flush control via the
2405 sysfs interface is still possible after
2406 boot. Hypervisors will issue a warning
2407 when the first VM is started in a
2408 potentially insecure configuration,
2409 i.e. SMT enabled or L1D flush disabled.
2413 Disables SMT and enables the default
2414 hypervisor mitigation.
2416 SMT control and L1D flush control via the
2417 sysfs interface is still possible after
2418 boot. Hypervisors will issue a warning
2419 when the first VM is started in a
2420 potentially insecure configuration,
2421 i.e. SMT enabled or L1D flush disabled.
2424 Same as 'flush', but hypervisors will not
2425 warn when a VM is started in a potentially
2426 insecure configuration.
2429 Disables hypervisor mitigations and doesn't
2431 It also drops the swap size and available
2432 RAM limit restriction on both hypervisor and
2437 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2443 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2446 lapic= [X86,APIC] Do not use TSC deadline
2447 value for LAPIC timer one-shot implementation. Default
2448 back to the programmable timer unit in the LAPIC.
2449 Format: notscdeadline
2451 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2454 libata.dma= [LIBATA] DMA control
2455 libata.dma=0 Disable all PATA and SATA DMA
2456 libata.dma=1 PATA and SATA Disk DMA only
2457 libata.dma=2 ATAPI (CDROM) DMA only
2458 libata.dma=4 Compact Flash DMA only
2459 Combinations also work, so libata.dma=3 enables DMA
2460 for disks and CDROMs, but not CFs.
2462 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2463 libata.ignore_hpa=0 keep BIOS limits (default)
2464 libata.ignore_hpa=1 ignore limits, using full disk
2466 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2470 libata.force= [LIBATA] Force configurations. The format is comma-
2471 separated list of "[ID:]VAL" where ID is
2472 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2473 matching port, link or device. Basically, it matches
2474 the ATA ID string printed on console by libata. If
2475 the whole ID part is omitted, the last PORT and DEVICE
2476 values are used. If ID hasn't been specified yet, the
2477 configuration applies to all ports, links and devices.
2479 If only DEVICE is omitted, the parameter applies to
2480 the port and all links and devices behind it. DEVICE
2481 number of 0 either selects the first device or the
2482 first fan-out link behind PMP device. It does not
2483 select the host link. DEVICE number of 15 selects the
2484 host link and device attached to it.
2486 The VAL specifies the configuration to force. As long
2487 as there's no ambiguity shortcut notation is allowed.
2488 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2489 The following configurations can be forced.
2491 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2492 Any ID with matching PORT is used.
2494 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2496 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2497 udma[/][16,25,33,44,66,100,133] notation is also
2500 * [no]ncq: Turn on or off NCQ.
2502 * [no]ncqtrim: Turn off queued DSM TRIM.
2504 * nohrst, nosrst, norst: suppress hard, soft
2507 * rstonce: only attempt one reset during
2508 hot-unplug link recovery
2510 * dump_id: dump IDENTIFY data.
2512 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2514 * disable: Disable this device.
2516 If there are multiple matching configurations changing
2517 the same attribute, the last one is used.
2519 memblock=debug [KNL] Enable memblock debug messages.
2521 load_ramdisk= [RAM] [Deprecated]
2523 lockd.nlm_grace_period=P [NFS] Assign grace period.
2526 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2529 lockd.nlm_timeout=T [NFS] Assign timeout value.
2532 lockd.nlm_udpport=M [NFS] Assign UDP port.
2535 lockdown= [SECURITY]
2536 { integrity | confidentiality }
2537 Enable the kernel lockdown feature. If set to
2538 integrity, kernel features that allow userland to
2539 modify the running kernel are disabled. If set to
2540 confidentiality, kernel features that allow userland
2541 to extract confidential information from the kernel
2544 locktorture.nreaders_stress= [KNL]
2545 Set the number of locking read-acquisition kthreads.
2546 Defaults to being automatically set based on the
2547 number of online CPUs.
2549 locktorture.nwriters_stress= [KNL]
2550 Set the number of locking write-acquisition kthreads.
2552 locktorture.onoff_holdoff= [KNL]
2553 Set time (s) after boot for CPU-hotplug testing.
2555 locktorture.onoff_interval= [KNL]
2556 Set time (s) between CPU-hotplug operations, or
2557 zero to disable CPU-hotplug testing.
2559 locktorture.shuffle_interval= [KNL]
2560 Set task-shuffle interval (jiffies). Shuffling
2561 tasks allows some CPUs to go into dyntick-idle
2562 mode during the locktorture test.
2564 locktorture.shutdown_secs= [KNL]
2565 Set time (s) after boot system shutdown. This
2566 is useful for hands-off automated testing.
2568 locktorture.stat_interval= [KNL]
2569 Time (s) between statistics printk()s.
2571 locktorture.stutter= [KNL]
2572 Time (s) to stutter testing, for example,
2573 specifying five seconds causes the test to run for
2574 five seconds, wait for five seconds, and so on.
2575 This tests the locking primitive's ability to
2576 transition abruptly to and from idle.
2578 locktorture.torture_type= [KNL]
2579 Specify the locking implementation to test.
2581 locktorture.verbose= [KNL]
2582 Enable additional printk() statements.
2584 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2587 loglevel= All Kernel Messages with a loglevel smaller than the
2588 console loglevel will be printed to the console. It can
2589 also be changed with klogd or other programs. The
2590 loglevels are defined as follows:
2592 0 (KERN_EMERG) system is unusable
2593 1 (KERN_ALERT) action must be taken immediately
2594 2 (KERN_CRIT) critical conditions
2595 3 (KERN_ERR) error conditions
2596 4 (KERN_WARNING) warning conditions
2597 5 (KERN_NOTICE) normal but significant condition
2598 6 (KERN_INFO) informational
2599 7 (KERN_DEBUG) debug-level messages
2601 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2602 in bytes. n must be a power of two and greater
2603 than the minimal size. The minimal size is defined
2604 by LOG_BUF_SHIFT kernel config parameter. There is
2605 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2606 that allows to increase the default size depending on
2607 the number of CPUs. See init/Kconfig for more details.
2609 logo.nologo [FB] Disables display of the built-in Linux logo.
2610 This may be used to provide more screen space for
2611 kernel log messages and is useful when debugging
2612 kernel boot problems.
2614 lp=0 [LP] Specify parallel ports to use, e.g,
2615 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2616 lp=reset first parallel port). 'lp=0' disables the
2617 lp=auto printer driver. 'lp=reset' (which can be
2618 specified in addition to the ports) causes
2619 attached printers to be reset. Using
2620 lp=port1,port2,... specifies the parallel ports
2621 to associate lp devices with, starting with
2622 lp0. A port specification may be 'none' to skip
2623 that lp device, or a parport name such as
2624 'parport0'. Specifying 'lp=auto' instead of a
2625 port specification list means that device IDs
2626 from each port should be examined, to see if
2627 an IEEE 1284-compliant printer is attached; if
2628 so, the driver will manage that printer.
2629 See also header of drivers/char/lp.c.
2632 Sets loops_per_jiffy to given constant, thus avoiding
2633 time-consuming boot-time autodetection (up to 250 ms per
2634 CPU). 0 enables autodetection (default). To determine
2635 the correct value for your kernel, boot with normal
2636 autodetection and see what value is printed. Note that
2637 on SMP systems the preset will be applied to all CPUs,
2638 which is likely to cause problems if your CPUs need
2639 significantly divergent settings. An incorrect value
2640 will cause delays in the kernel to be wrong, leading to
2641 unpredictable I/O errors and other breakage. Although
2642 unlikely, in the extreme case this might damage your
2646 Format: <io>,<irq>,<dma>
2648 lsm.debug [SECURITY] Enable LSM initialization debugging output.
2651 [SECURITY] Choose order of LSM initialization. This
2652 overrides CONFIG_LSM, and the "security=" parameter.
2654 machvec= [IA-64] Force the use of a particular machine-vector
2655 (machvec) in a generic kernel.
2656 Example: machvec=hpzx1
2658 machtype= [Loongson] Share the same kernel image file between
2659 different yeeloong laptops.
2660 Example: machtype=lemote-yeeloong-2f-7inch
2662 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2663 than or equal to this physical address is ignored.
2665 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2666 will bring up during bootup. maxcpus=n : n >= 0 limits
2667 the kernel to bring up 'n' processors. Surely after
2668 bootup you can bring up the other plugged cpu by executing
2669 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2670 only takes effect during system bootup.
2671 While n=0 is a special case, it is equivalent to "nosmp",
2672 which also disables the IO APIC.
2674 max_loop= [LOOP] The number of loop block devices that get
2675 (loop.max_loop) unconditionally pre-created at init time. The default
2676 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2677 of statically allocating a predefined number, loop
2678 devices can be requested on-demand with the
2679 /dev/loop-control interface.
2681 mce [X86-32] Machine Check Exception
2683 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.rst
2685 md= [HW] RAID subsystems devices and level
2686 See Documentation/admin-guide/md.rst.
2689 Format: <first>,<last>
2690 Specifies range of consoles to be captured by the MDA.
2693 Control mitigation for the Micro-architectural Data
2694 Sampling (MDS) vulnerability.
2696 Certain CPUs are vulnerable to an exploit against CPU
2697 internal buffers which can forward information to a
2698 disclosure gadget under certain conditions.
2700 In vulnerable processors, the speculatively
2701 forwarded data can be used in a cache side channel
2702 attack, to access data to which the attacker does
2703 not have direct access.
2705 This parameter controls the MDS mitigation. The
2708 full - Enable MDS mitigation on vulnerable CPUs
2709 full,nosmt - Enable MDS mitigation and disable
2710 SMT on vulnerable CPUs
2711 off - Unconditionally disable MDS mitigation
2713 On TAA-affected machines, mds=off can be prevented by
2714 an active TAA mitigation as both vulnerabilities are
2715 mitigated with the same mechanism so in order to disable
2716 this mitigation, you need to specify tsx_async_abort=off
2719 Not specifying this option is equivalent to
2722 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2724 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2725 Amount of memory to be used in cases as follows:
2728 2 when the kernel is not able to see the whole system memory;
2729 3 memory that lies after 'mem=' boundary is excluded from
2730 the hypervisor, then assigned to KVM guests.
2732 [X86] Work as limiting max address. Use together
2733 with memmap= to avoid physical address space collisions.
2734 Without memmap= PCI devices could be placed at addresses
2735 belonging to unused RAM.
2737 Note that this only takes effects during boot time since
2738 in above case 3, memory may need be hot added after boot
2739 if system memory of hypervisor is not sufficient.
2741 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2745 [KNL,SH] Allow user to override the default size for
2746 per-device physically contiguous DMA buffers.
2748 memhp_default_state=online/offline
2749 [KNL] Set the initial state for the memory hotplug
2750 onlining policy. If not specified, the default value is
2751 set according to the
2752 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2754 See Documentation/admin-guide/mm/memory-hotplug.rst.
2756 memmap=exactmap [KNL,X86] Enable setting of an exact
2757 E820 memory map, as specified by the user.
2758 Such memmap=exactmap lines can be constructed based on
2759 BIOS output or other requirements. See the memmap=nn@ss
2762 memmap=nn[KMG]@ss[KMG]
2763 [KNL, X86, MIPS, XTENSA] Force usage of a specific region of memory.
2764 Region of memory to be used is from ss to ss+nn.
2765 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2766 which limits max address to nn[KMG].
2767 Multiple different regions can be specified,
2770 memmap=100M@2G,100M#3G,1G!1024G
2772 memmap=nn[KMG]#ss[KMG]
2773 [KNL,ACPI] Mark specific memory as ACPI data.
2774 Region of memory to be marked is from ss to ss+nn.
2776 memmap=nn[KMG]$ss[KMG]
2777 [KNL,ACPI] Mark specific memory as reserved.
2778 Region of memory to be reserved is from ss to ss+nn.
2779 Example: Exclude memory from 0x18690000-0x1869ffff
2780 memmap=64K$0x18690000
2782 memmap=0x10000$0x18690000
2783 Some bootloaders may need an escape character before '$',
2784 like Grub2, otherwise '$' and the following number
2787 memmap=nn[KMG]!ss[KMG]
2788 [KNL,X86] Mark specific memory as protected.
2789 Region of memory to be used, from ss to ss+nn.
2790 The memory region may be marked as e820 type 12 (0xc)
2791 and is NVDIMM or ADR memory.
2793 memmap=<size>%<offset>-<oldtype>+<newtype>
2794 [KNL,ACPI] Convert memory within the specified region
2795 from <oldtype> to <newtype>. If "-<oldtype>" is left
2796 out, the whole region will be marked as <newtype>,
2797 even if previously unavailable. If "+<newtype>" is left
2798 out, matching memory will be removed. Types are
2799 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2800 3 = ACPI, 12 = PRAM.
2802 memory_corruption_check=0/1 [X86]
2803 Some BIOSes seem to corrupt the first 64k of
2804 memory when doing things like suspend/resume.
2805 Setting this option will scan the memory
2806 looking for corruption. Enabling this will
2807 both detect corruption and prevent the kernel
2808 from using the memory being corrupted.
2809 However, its intended as a diagnostic tool; if
2810 repeatable BIOS-originated corruption always
2811 affects the same memory, you can use memmap=
2812 to prevent the kernel from using that memory.
2814 memory_corruption_check_size=size [X86]
2815 By default it checks for corruption in the low
2816 64k, making this memory unavailable for normal
2817 use. Use this parameter to scan for
2818 corruption in more or less memory.
2820 memory_corruption_check_period=seconds [X86]
2821 By default it checks for corruption every 60
2822 seconds. Use this parameter to check at some
2823 other rate. 0 disables periodic checking.
2825 memory_hotplug.memmap_on_memory
2826 [KNL,X86,ARM] Boolean flag to enable this feature.
2827 Format: {on | off (default)}
2828 When enabled, runtime hotplugged memory will
2829 allocate its internal metadata (struct pages)
2830 from the hotadded memory which will allow to
2831 hotadd a lot of memory without requiring
2832 additional memory to do so.
2833 This feature is disabled by default because it
2834 has some implication on large (e.g. GB)
2835 allocations in some configurations (e.g. small
2837 The state of the flag can be read in
2838 /sys/module/memory_hotplug/parameters/memmap_on_memory.
2839 Note that even when enabled, there are a few cases where
2840 the feature is not effective.
2842 memtest= [KNL,X86,ARM,PPC,RISCV] Enable memtest
2844 default : 0 <disable>
2845 Specifies the number of memtest passes to be
2846 performed. Each pass selects another test
2847 pattern from a given set of patterns. Memtest
2848 fills the memory with this pattern, validates
2849 memory contents and reserves bad memory
2850 regions that are detected.
2852 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2853 Valid arguments: on, off
2854 Default (depends on kernel configuration option):
2855 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2856 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2857 mem_encrypt=on: Activate SME
2858 mem_encrypt=off: Do not activate SME
2860 Refer to Documentation/virt/kvm/amd-memory-encryption.rst
2861 for details on when memory encryption can be activated.
2863 mem_sleep_default= [SUSPEND] Default system suspend mode:
2864 s2idle - Suspend-To-Idle
2865 shallow - Power-On Suspend or equivalent (if supported)
2866 deep - Suspend-To-RAM or equivalent (if supported)
2867 See Documentation/admin-guide/pm/sleep-states.rst.
2869 meye.*= [HW] Set MotionEye Camera parameters
2870 See Documentation/admin-guide/media/meye.rst.
2872 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2873 Multi-Function General Purpose Timers on AMD Geode
2876 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2877 the BIOS has incorrectly applied a workaround. TinyBIOS
2878 version 0.98 is known to be affected, 0.99 fixes the
2879 problem by letting the user disable the workaround.
2883 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2884 physical address is ignored.
2886 mini2440= [ARM,HW,KNL]
2887 Format:[0..2][b][c][t]
2889 MINI2440 configuration specification:
2890 0 - The attached screen is the 3.5" TFT
2891 1 - The attached screen is the 7" TFT
2892 2 - The VGA Shield is attached (1024x768)
2893 Leaving out the screen size parameter will not load
2894 the TFT driver, and the framebuffer will be left
2896 b - Enable backlight. The TFT backlight pin will be
2897 linked to the kernel VESA blanking code and a GPIO
2898 LED. This parameter is not necessary when using the
2900 c - Enable the s3c camera interface.
2901 t - Reserved for enabling touchscreen support. The
2902 touchscreen support is not enabled in the mainstream
2903 kernel as of 2.6.30, a preliminary port can be found
2904 in the "bleeding edge" mini2440 support kernel at
2905 https://repo.or.cz/w/linux-2.6/mini2440.git
2908 [X86,PPC,S390,ARM64] Control optional mitigations for
2909 CPU vulnerabilities. This is a set of curated,
2910 arch-independent options, each of which is an
2911 aggregation of existing arch-specific options.
2914 Disable all optional CPU mitigations. This
2915 improves system performance, but it may also
2916 expose users to several CPU vulnerabilities.
2917 Equivalent to: nopti [X86,PPC]
2919 nospectre_v1 [X86,PPC]
2921 nospectre_v2 [X86,PPC,S390,ARM64]
2922 spectre_v2_user=off [X86]
2923 spec_store_bypass_disable=off [X86,PPC]
2924 ssbd=force-off [ARM64]
2927 tsx_async_abort=off [X86]
2928 kvm.nx_huge_pages=off [X86]
2929 no_entry_flush [PPC]
2930 no_uaccess_flush [PPC]
2933 This does not have any effect on
2934 kvm.nx_huge_pages when
2935 kvm.nx_huge_pages=force.
2938 Mitigate all CPU vulnerabilities, but leave SMT
2939 enabled, even if it's vulnerable. This is for
2940 users who don't want to be surprised by SMT
2941 getting disabled across kernel upgrades, or who
2942 have other ways of avoiding SMT-based attacks.
2943 Equivalent to: (default behavior)
2946 Mitigate all CPU vulnerabilities, disabling SMT
2947 if needed. This is for users who always want to
2948 be fully mitigated, even if it means losing SMT.
2949 Equivalent to: l1tf=flush,nosmt [X86]
2950 mds=full,nosmt [X86]
2951 tsx_async_abort=full,nosmt [X86]
2954 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2955 parameter allows control of the logging verbosity for
2956 the additional memory initialisation checks. A value
2957 of 0 disables mminit logging and a level of 4 will
2958 log everything. Information is printed at KERN_DEBUG
2959 so loglevel=8 may also need to be specified.
2962 [KNL] When CONFIG_MODULE_SIG is set, this means that
2963 modules without (valid) signatures will fail to load.
2964 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2965 is always true, so this option does nothing.
2967 module_blacklist= [KNL] Do not load a comma-separated list of
2968 modules. Useful for debugging problem modules.
2971 [MOUSE] Maximum time between finger touching and
2972 leaving touchpad surface for touch to be considered
2973 a tap and be reported as a left button click (for
2974 touchpads working in absolute mode only).
2976 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2977 reporting absolute coordinates, such as tablets
2978 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2979 reporting absolute coordinates, such as tablets
2981 movablecore= [KNL,X86,IA-64,PPC]
2982 Format: nn[KMGTPE] | nn%
2983 This parameter is the complement to kernelcore=, it
2984 specifies the amount of memory used for migratable
2985 allocations. If both kernelcore and movablecore is
2986 specified, then kernelcore will be at *least* the
2987 specified value but may be more. If movablecore on its
2988 own is specified, the administrator must be careful
2989 that the amount of memory usable for all allocations
2992 movable_node [KNL] Boot-time switch to make hotplugable memory
2993 NUMA nodes to be movable. This means that the memory
2994 of such nodes will be usable only for movable
2995 allocations which rules out almost all kernel
2996 allocations. Use with caution!
2998 MTD_Partition= [MTD]
2999 Format: <name>,<region-number>,<size>,<offset>
3001 MTD_Region= [MTD] Format:
3002 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
3005 See drivers/mtd/parsers/cmdlinepart.c
3007 multitce=off [PPC] This parameter disables the use of the pSeries
3008 firmware feature for updating multiple TCE entries
3011 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
3013 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
3015 boundary - index of last SLC block on Flex-OneNAND.
3016 The remaining blocks are configured as MLC blocks.
3017 lock - Configure if Flex-OneNAND boundary should be locked.
3018 Once locked, the boundary cannot be changed.
3019 1 indicates lock status, 0 indicates unlock status.
3022 ARM/S3C2412 JIVE boot control
3024 See arch/arm/mach-s3c/mach-jive.c
3026 mtouchusb.raw_coordinates=
3027 [HW] Make the MicroTouch USB driver use raw coordinates
3028 ('y', default) or cooked coordinates ('n')
3030 mtrr_chunk_size=nn[KMG] [X86]
3031 used for mtrr cleanup. It is largest continuous chunk
3032 that could hold holes aka. UC entries.
3034 mtrr_gran_size=nn[KMG] [X86]
3035 Used for mtrr cleanup. It is granularity of mtrr block.
3037 Large value could prevent small alignment from
3040 mtrr_spare_reg_nr=n [X86]
3042 Range: 0,7 : spare reg number
3044 Used for mtrr cleanup. It is spare mtrr entries number.
3045 Set to 2 or more if your graphical card needs more.
3047 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
3049 netdev= [NET] Network devices parameters
3050 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
3051 Note that mem_start is often overloaded to mean
3052 something different and driver-specific.
3053 This usage is only documented in each driver source
3057 [NETFILTER] Enable connection tracking flow accounting
3058 0 to disable accounting
3059 1 to enable accounting
3062 nfsaddrs= [NFS] Deprecated. Use ip= instead.
3063 See Documentation/admin-guide/nfs/nfsroot.rst.
3065 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
3066 See Documentation/admin-guide/nfs/nfsroot.rst.
3068 nfsrootdebug [NFS] enable nfsroot debugging messages.
3069 See Documentation/admin-guide/nfs/nfsroot.rst.
3071 nfs.callback_nr_threads=
3072 [NFSv4] set the total number of threads that the
3073 NFS client will assign to service NFSv4 callback
3076 nfs.callback_tcpport=
3077 [NFS] set the TCP port on which the NFSv4 callback
3078 channel should listen.
3081 [NFS] sets the pathname to the program which is used
3082 to update the NFS client cache entries.
3084 nfs.cache_getent_timeout=
3085 [NFS] sets the timeout after which an attempt to
3086 update a cache entry is deemed to have failed.
3088 nfs.idmap_cache_timeout=
3089 [NFS] set the maximum lifetime for idmapper cache
3093 [NFS] enable 64-bit inode numbers.
3094 If zero, the NFS client will fake up a 32-bit inode
3095 number for the readdir() and stat() syscalls instead
3096 of returning the full 64-bit number.
3097 The default is to return 64-bit inode numbers.
3099 nfs.max_session_cb_slots=
3100 [NFSv4.1] Sets the maximum number of session
3101 slots the client will assign to the callback
3102 channel. This determines the maximum number of
3103 callbacks the client will process in parallel for
3104 a particular server.
3106 nfs.max_session_slots=
3107 [NFSv4.1] Sets the maximum number of session slots
3108 the client will attempt to negotiate with the server.
3109 This limits the number of simultaneous RPC requests
3110 that the client can send to the NFSv4.1 server.
3111 Note that there is little point in setting this
3112 value higher than the max_tcp_slot_table_limit.
3114 nfs.nfs4_disable_idmapping=
3115 [NFSv4] When set to the default of '1', this option
3116 ensures that both the RPC level authentication
3117 scheme and the NFS level operations agree to use
3118 numeric uids/gids if the mount is using the
3119 'sec=sys' security flavour. In effect it is
3120 disabling idmapping, which can make migration from
3121 legacy NFSv2/v3 systems to NFSv4 easier.
3122 Servers that do not support this mode of operation
3123 will be autodetected by the client, and it will fall
3124 back to using the idmapper.
3125 To turn off this behaviour, set the value to '0'.
3127 [NFS4] Specify an additional fixed unique ident-
3128 ification string that NFSv4 clients can insert into
3129 their nfs_client_id4 string. This is typically a
3130 UUID that is generated at system install time.
3132 nfs.send_implementation_id =
3133 [NFSv4.1] Send client implementation identification
3134 information in exchange_id requests.
3135 If zero, no implementation identification information
3137 The default is to send the implementation identification
3140 nfs.recover_lost_locks =
3141 [NFSv4] Attempt to recover locks that were lost due
3142 to a lease timeout on the server. Please note that
3143 doing this risks data corruption, since there are
3144 no guarantees that the file will remain unchanged
3145 after the locks are lost.
3146 If you want to enable the kernel legacy behaviour of
3147 attempting to recover these locks, then set this
3149 The default parameter value of '0' causes the kernel
3150 not to attempt recovery of lost locks.
3152 nfs4.layoutstats_timer =
3153 [NFSv4.2] Change the rate at which the kernel sends
3154 layoutstats to the pNFS metadata server.
3156 Setting this to value to 0 causes the kernel to use
3157 whatever value is the default set by the layout
3158 driver. A non-zero value sets the minimum interval
3159 in seconds between layoutstats transmissions.
3161 nfsd.nfs4_disable_idmapping=
3162 [NFSv4] When set to the default of '1', the NFSv4
3163 server will return only numeric uids and gids to
3164 clients using auth_sys, and will accept numeric uids
3165 and gids from such clients. This is intended to ease
3166 migration from NFSv2/v3.
3168 nmi_backtrace.backtrace_idle [KNL]
3169 Dump stacks even of idle CPUs in response to an
3170 NMI stack-backtrace request.
3172 nmi_debug= [KNL,SH] Specify one or more actions to take
3173 when a NMI is triggered.
3174 Format: [state][,regs][,debounce][,die]
3176 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
3177 Format: [panic,][nopanic,][num]
3179 0 - turn hardlockup detector in nmi_watchdog off
3180 1 - turn hardlockup detector in nmi_watchdog on
3181 When panic is specified, panic when an NMI watchdog
3182 timeout occurs (or 'nopanic' to not panic on an NMI
3183 watchdog, if CONFIG_BOOTPARAM_HARDLOCKUP_PANIC is set)
3184 To disable both hard and soft lockup detectors,
3185 please see 'nowatchdog'.
3186 This is useful when you use a panic=... timeout and
3187 need the box quickly up again.
3189 These settings can be accessed at runtime via
3190 the nmi_watchdog and hardlockup_panic sysctls.
3192 netpoll.carrier_timeout=
3193 [NET] Specifies amount of time (in seconds) that
3194 netpoll should wait for a carrier. By default netpoll
3197 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
3198 emulation library even if a 387 maths coprocessor
3201 no5lvl [X86-64] Disable 5-level paging mode. Forces
3202 kernel to use 4-level paging instead.
3204 nofsgsbase [X86] Disables FSGSBASE instructions.
3207 [HW] Never suspend the console
3208 Disable suspending of consoles during suspend and
3209 hibernate operations. Once disabled, debugging
3210 messages can reach various consoles while the rest
3211 of the system is being put to sleep (ie, while
3212 debugging driver suspend/resume hooks). This may
3213 not work reliably with all consoles, but is known
3214 to work with serial and VGA consoles.
3215 To facilitate more flexible debugging, we also add
3216 console_suspend, a printk module parameter to control
3217 it. Users could use console_suspend (usually
3218 /sys/module/printk/parameters/console_suspend) to
3219 turn on/off it dynamically.
3221 novmcoredd [KNL,KDUMP]
3222 Disable device dump. Device dump allows drivers to
3223 append dump data to vmcore so you can collect driver
3224 specified debug info. Drivers can append the data
3225 without any limit and this data is stored in memory,
3226 so this may cause significant memory stress. Disabling
3227 device dump can help save memory but the driver debug
3228 data will be no longer available. This parameter
3229 is only available when CONFIG_PROC_VMCORE_DEVICE_DUMP
3232 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
3233 caches in the slab allocator. Saves per-node memory,
3234 but will impact performance.
3238 noaltinstr [S390] Disables alternative instructions patching
3239 (CPU alternatives feature).
3241 noapic [SMP,APIC] Tells the kernel to not make use of any
3242 IOAPICs that may be present in the system.
3244 noautogroup Disable scheduler automatic task group creation.
3246 nobats [PPC] Do not use BATs for mapping kernel lowmem
3247 on "Classic" PPC cores.
3251 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
3253 nodelayacct [KNL] Disable per-task delay accounting
3255 nodsp [SH] Disable hardware DSP at boot time.
3257 noefi Disable EFI runtime services support.
3259 no_entry_flush [PPC] Don't flush the L1-D cache when entering the kernel.
3264 On X86-32 available only on PAE configured kernels.
3265 noexec=on: enable non-executable mappings (default)
3266 noexec=off: disable non-executable mappings
3269 Disable SMAP (Supervisor Mode Access Prevention)
3270 even if it is supported by processor.
3273 Disable SMEP (Supervisor Mode Execution Prevention)
3274 even if it is supported by processor.
3277 This affects only 32-bit executables.
3278 noexec32=on: enable non-executable mappings (default)
3279 read doesn't imply executable mappings
3280 noexec32=off: disable non-executable mappings
3281 read implies executable mappings
3283 nofpu [MIPS,SH] Disable hardware FPU at boot time.
3285 nofxsr [BUGS=X86-32] Disables x86 floating point extended
3286 register save and restore. The kernel will only save
3287 legacy floating-point registers on task switch.
3289 nohugeiomap [KNL,X86,PPC,ARM64] Disable kernel huge I/O mappings.
3291 nohugevmalloc [PPC] Disable kernel huge vmalloc mappings.
3293 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
3294 Equivalent to smt=1.
3296 [KNL,X86] Disable symmetric multithreading (SMT).
3297 nosmt=force: Force disable SMT, cannot be undone
3298 via the sysfs control file.
3300 nospectre_v1 [X86,PPC] Disable mitigations for Spectre Variant 1
3301 (bounds check bypass). With this option data leaks are
3302 possible in the system.
3304 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
3305 the Spectre variant 2 (indirect branch prediction)
3306 vulnerability. System may allow data leaks with this
3309 nospec_store_bypass_disable
3310 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
3313 [PPC] Don't flush the L1-D cache after accessing user data.
3315 noxsave [BUGS=X86] Disables x86 extended register state save
3316 and restore using xsave. The kernel will fallback to
3317 enabling legacy floating-point and sse state.
3319 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
3320 register states. The kernel will fall back to use
3321 xsave to save the states. By using this parameter,
3322 performance of saving the states is degraded because
3323 xsave doesn't support modified optimization while
3324 xsaveopt supports it on xsaveopt enabled systems.
3326 noxsaves [X86] Disables xsaves and xrstors used in saving and
3327 restoring x86 extended register state in compacted
3328 form of xsave area. The kernel will fall back to use
3329 xsaveopt and xrstor to save and restore the states
3330 in standard form of xsave area. By using this
3331 parameter, xsave area per process might occupy more
3332 memory on xsaves enabled systems.
3334 nohlt [ARM,ARM64,MICROBLAZE,SH] Forces the kernel to busy wait
3335 in do_idle() and not use the arch_cpu_idle()
3336 implementation; requires CONFIG_GENERIC_IDLE_POLL_SETUP
3337 to be effective. This is useful on platforms where the
3338 sleep(SH) or wfi(ARM,ARM64) instructions do not work
3339 correctly or when doing power measurements to evalute
3340 the impact of the sleep instructions. This is also
3341 useful when using JTAG debugger.
3343 no_file_caps Tells the kernel not to honor file capabilities. The
3344 only way then for a file to be executed with privilege
3345 is to be setuid root or executed by root.
3347 nohalt [IA-64] Tells the kernel not to use the power saving
3348 function PAL_HALT_LIGHT when idle. This increases
3349 power-consumption. On the positive side, it reduces
3350 interrupt wake-up latency, which may improve performance
3351 in certain environments such as networked servers or
3355 Force pointers printed to the console or buffers to be
3356 unhashed. By default, when a pointer is printed via %p
3357 format string, that pointer is "hashed", i.e. obscured
3358 by hashing the pointer value. This is a security feature
3359 that hides actual kernel addresses from unprivileged
3360 users, but it also makes debugging the kernel more
3361 difficult since unequal pointers can no longer be
3362 compared. However, if this command-line option is
3363 specified, then all normal pointers will have their true
3364 value printed. Pointers printed via %pK may still be
3365 hashed. This option should only be specified when
3366 debugging the kernel. Please do not use on production
3369 nohibernate [HIBERNATION] Disable hibernation and resume.
3371 nohz= [KNL] Boottime enable/disable dynamic ticks
3372 Valid arguments: on, off
3375 nohz_full= [KNL,BOOT,SMP,ISOL]
3376 The argument is a cpu list, as described above.
3377 In kernels built with CONFIG_NO_HZ_FULL=y, set
3378 the specified list of CPUs whose tick will be stopped
3379 whenever possible. The boot CPU will be forced outside
3380 the range to maintain the timekeeping. Any CPUs
3381 in this list will have their RCU callbacks offloaded,
3382 just as if they had also been called out in the
3383 rcu_nocbs= boot parameter.
3385 noiotrap [SH] Disables trapped I/O port accesses.
3387 noirqdebug [X86-32] Disables the code which attempts to detect and
3388 disable unhandled interrupt sources.
3390 no_timer_check [X86,APIC] Disables the code which tests for
3391 broken timer IRQ sources.
3393 noisapnp [ISAPNP] Disables ISA PnP code.
3395 noinitrd [RAM] Tells the kernel not to load any configured
3398 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
3400 [Deprecated - use intremap=off]
3404 noinvpcid [X86] Disable the INVPCID cpu feature.
3406 nojitter [IA-64] Disables jitter checking for ITC timers.
3408 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
3410 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
3414 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3415 clock and use the default one.
3417 no-steal-acc [X86,PV_OPS,ARM64] Disable paravirtualized steal time
3418 accounting. steal time is computed, but won't
3419 influence scheduler behaviour
3421 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3423 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3425 noltlbs [PPC] Do not use large page/tlb entries for kernel
3426 lowmem mapping on PPC40x and PPC8xx
3428 nomca [IA-64] Disable machine check abort handling
3430 nomce [X86-32] Disable Machine Check Exception
3432 nomfgpt [X86-32] Disable Multi-Function General Purpose
3433 Timer usage (for AMD Geode machines).
3435 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3436 shutdown the other cpus. Instead use the REBOOT_VECTOR
3439 nomodule Disable module load
3441 nopat [X86] Disable PAT (page attribute table extension of
3442 pagetables) support.
3444 nopcid [X86-64] Disable the PCID cpu feature.
3446 norandmaps Don't use address space randomization. Equivalent to
3447 echo 0 > /proc/sys/kernel/randomize_va_space
3449 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3450 with UP alternatives
3452 nordrand [X86] Disable kernel use of the RDRAND and
3453 RDSEED instructions even if they are supported
3454 by the processor. RDRAND and RDSEED are still
3455 available to user space applications.
3457 noresume [SWSUSP] Disables resume and restores original swap
3460 no-scroll [VGA] Disables scrollback.
3461 This is required for the Braillex ib80-piezo Braille
3462 reader made by F.H. Papenmeier (Germany).
3466 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3468 nosgx [X86-64,SGX] Disables Intel SGX kernel support.
3470 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3471 and disable the IO APIC. legacy for "maxcpus=0".
3473 nosoftlockup [KNL] Disable the soft-lockup detector.
3475 nosync [HW,M68K] Disables sync negotiation for all devices.
3477 nowatchdog [KNL] Disable both lockup detectors, i.e.
3478 soft-lockup and NMI watchdog (hard-lockup).
3482 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3484 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3485 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3486 Some features depend on CPU0. Known dependencies are:
3487 1. Resume from suspend/hibernate depends on CPU0.
3488 Suspend/hibernate will fail if CPU0 is offline and you
3489 need to online CPU0 before suspend/hibernate.
3490 2. PIC interrupts also depend on CPU0. CPU0 can't be
3491 removed if a PIC interrupt is detected.
3492 It's said poweroff/reboot may depend on CPU0 on some
3493 machines although I haven't seen such issues so far
3494 after CPU0 is offline on a few tested machines.
3495 If the dependencies are under your control, you can
3496 turn on cpu0_hotplug.
3498 nps_mtm_hs_ctr= [KNL,ARC]
3499 This parameter sets the maximum duration, in
3500 cycles, each HW thread of the CTOP can run
3501 without interruptions, before HW switches it.
3502 The actual maximum duration is 16 times this
3504 Format: integer between 1 and 255
3507 nptcg= [IA-64] Override max number of concurrent global TLB
3508 purges which is reported from either PAL_VM_SUMMARY or
3511 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3512 could support. nr_cpus=n : n >= 1 limits the kernel to
3513 support 'n' processors. It could be larger than the
3514 number of already plugged CPU during bootup, later in
3515 runtime you can physically add extra cpu until it reaches
3516 n. So during boot up some boot time memory for per-cpu
3517 variables need be pre-allocated for later physical cpu
3520 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3522 numa_balancing= [KNL,ARM64,PPC,RISCV,S390,X86] Enable or disable automatic
3524 Allowed values are enable and disable
3526 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3527 'node', 'default' can be specified
3528 This can be set from sysctl after boot.
3529 See Documentation/admin-guide/sysctl/vm.rst for details.
3531 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3532 See Documentation/core-api/debugging-via-ohci1394.rst for more
3535 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3536 Rather than timing out after 20 ms if an EC
3537 command is not properly ACKed, override the length
3538 of the timeout. We have interrupts disabled while
3539 waiting for the ACK, so if this is set too high
3540 interrupts *may* be lost!
3542 omap_mux= [OMAP] Override bootloader pin multiplexing.
3543 Format: <mux_mode0.mode_name=value>...
3544 For example, to override I2C bus2:
3545 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3547 oops=panic Always panic on oopses. Default is to just kill the
3548 process, but there is a small probability of
3549 deadlocking the machine.
3550 This will also cause panics on machine check exceptions.
3551 Useful together with panic=30 to trigger a reboot.
3554 [KNL] Boolean flag to control whether the page allocator
3555 should randomize its free lists. The randomization may
3556 be automatically enabled if the kernel detects it is
3557 running on a platform with a direct-mapped memory-side
3558 cache, and this parameter can be used to
3559 override/disable that behavior. The state of the flag
3560 can be read from sysfs at:
3561 /sys/module/page_alloc/parameters/shuffle.
3563 page_owner= [KNL] Boot-time page_owner enabling option.
3564 Storage of the information about who allocated
3565 each page is disabled in default. With this switch,
3567 on: enable the feature
3569 page_poison= [KNL] Boot-time parameter changing the state of
3570 poisoning on the buddy allocator, available with
3571 CONFIG_PAGE_POISONING=y.
3572 off: turn off poisoning (default)
3573 on: turn on poisoning
3575 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3576 timeout > 0: seconds before rebooting
3577 timeout = 0: wait forever
3578 timeout < 0: reboot immediately
3581 panic_print= Bitmask for printing system info when panic happens.
3582 User can chose combination of the following bits:
3583 bit 0: print all tasks info
3584 bit 1: print system memory info
3585 bit 2: print timer info
3586 bit 3: print locks info if CONFIG_LOCKDEP is on
3587 bit 4: print ftrace buffer
3588 bit 5: print all printk messages in buffer
3590 panic_on_taint= Bitmask for conditionally calling panic() in add_taint()
3591 Format: <hex>[,nousertaint]
3592 Hexadecimal bitmask representing the set of TAINT flags
3593 that will cause the kernel to panic when add_taint() is
3594 called with any of the flags in this set.
3595 The optional switch "nousertaint" can be utilized to
3596 prevent userspace forced crashes by writing to sysctl
3597 /proc/sys/kernel/tainted any flagset matching with the
3598 bitmask set on panic_on_taint.
3599 See Documentation/admin-guide/tainted-kernels.rst for
3600 extra details on the taint flags that users can pick
3601 to compose the bitmask to assign to panic_on_taint.
3603 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3606 crash_kexec_post_notifiers
3607 Run kdump after running panic-notifiers and dumping
3608 kmsg. This only for the users who doubt kdump always
3609 succeeds in any situation.
3610 Note that this also increases risks of kdump failure,
3611 because some panic notifiers can make the crashed
3612 kernel more unstable.
3614 parkbd.port= [HW] Parallel port number the keyboard adapter is
3615 connected to, default is 0.
3617 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3618 0 for XT, 1 for AT (default is AT).
3621 parport= [HW,PPT] Specify parallel ports. 0 disables.
3622 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3623 Use 'auto' to force the driver to use any
3624 IRQ/DMA settings detected (the default is to
3625 ignore detected IRQ/DMA settings because of
3626 possible conflicts). You can specify the base
3627 address, IRQ, and DMA settings; IRQ and DMA
3628 should be numbers, or 'auto' (for using detected
3629 settings on that particular port), or 'nofifo'
3630 (to avoid using a FIFO even if it is detected).
3631 Parallel ports are assigned in the order they
3632 are specified on the command line, starting
3635 parport_init_mode= [HW,PPT]
3636 Configure VIA parallel port to operate in
3637 a specific mode. This is necessary on Pegasos
3638 computer where firmware has no options for setting
3639 up parallel port mode and sets it to spp.
3640 Currently this function knows 686a and 8231 chips.
3641 Format: [spp|ps2|epp|ecp|ecpepp]
3643 pata_legacy.all= [HW,LIBATA]
3645 Set to non-zero to probe primary and secondary ISA
3646 port ranges on PCI systems where no PCI PATA device
3647 has been found at either range. Disabled by default.
3649 pata_legacy.autospeed= [HW,LIBATA]
3651 Set to non-zero if a chip is present that snoops speed
3652 changes. Disabled by default.
3654 pata_legacy.ht6560a= [HW,LIBATA]
3656 Set to 1, 2, or 3 for HT 6560A on the primary channel,
3657 the secondary channel, or both channels respectively.
3658 Disabled by default.
3660 pata_legacy.ht6560b= [HW,LIBATA]
3662 Set to 1, 2, or 3 for HT 6560B on the primary channel,
3663 the secondary channel, or both channels respectively.
3664 Disabled by default.
3666 pata_legacy.iordy_mask= [HW,LIBATA]
3668 IORDY enable mask. Set individual bits to allow IORDY
3669 for the respective channel. Bit 0 is for the first
3670 legacy channel handled by this driver, bit 1 is for
3671 the second channel, and so on. The sequence will often
3672 correspond to the primary legacy channel, the secondary
3673 legacy channel, and so on, but the handling of a PCI
3674 bus and the use of other driver options may interfere
3675 with the sequence. By default IORDY is allowed across
3678 pata_legacy.opti82c46x= [HW,LIBATA]
3680 Set to 1, 2, or 3 for Opti 82c611A on the primary
3681 channel, the secondary channel, or both channels
3682 respectively. Disabled by default.
3684 pata_legacy.opti82c611a= [HW,LIBATA]
3686 Set to 1, 2, or 3 for Opti 82c465MV on the primary
3687 channel, the secondary channel, or both channels
3688 respectively. Disabled by default.
3690 pata_legacy.pio_mask= [HW,LIBATA]
3692 PIO mode mask for autospeed devices. Set individual
3693 bits to allow the use of the respective PIO modes.
3694 Bit 0 is for mode 0, bit 1 is for mode 1, and so on.
3695 All modes allowed by default.
3697 pata_legacy.probe_all= [HW,LIBATA]
3699 Set to non-zero to probe tertiary and further ISA
3700 port ranges on PCI systems. Disabled by default.
3702 pata_legacy.probe_mask= [HW,LIBATA]
3704 Probe mask for legacy ISA PATA ports. Depending on
3705 platform configuration and the use of other driver
3706 options up to 6 legacy ports are supported: 0x1f0,
3707 0x170, 0x1e8, 0x168, 0x1e0, 0x160, however probing
3708 of individual ports can be disabled by setting the
3709 corresponding bits in the mask to 1. Bit 0 is for
3710 the first port in the list above (0x1f0), and so on.
3711 By default all supported ports are probed.
3713 pata_legacy.qdi= [HW,LIBATA]
3715 Set to non-zero to probe QDI controllers. By default
3716 set to 1 if CONFIG_PATA_QDI_MODULE, 0 otherwise.
3718 pata_legacy.winbond= [HW,LIBATA]
3720 Set to non-zero to probe Winbond controllers. Use
3721 the standard I/O port (0x130) if 1, otherwise the
3722 value given is the I/O port to use (typically 0x1b0).
3723 By default set to 1 if CONFIG_PATA_WINBOND_VLB_MODULE,
3726 pata_platform.pio_mask= [HW,LIBATA]
3728 Supported PIO mode mask. Set individual bits to allow
3729 the use of the respective PIO modes. Bit 0 is for
3730 mode 0, bit 1 is for mode 1, and so on. Mode 0 only
3734 Halt all CPUs after the first oops has been printed for
3735 the specified number of seconds. This is to be used if
3736 your oopses keep scrolling off the screen.
3741 See header of drivers/block/paride/pcd.c.
3742 See also Documentation/admin-guide/blockdev/paride.rst.
3744 pci=option[,option...] [PCI] various PCI subsystem options.
3746 Some options herein operate on a specific device
3747 or a set of devices (<pci_dev>). These are
3748 specified in one of the following formats:
3750 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3751 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3753 Note: the first format specifies a PCI
3754 bus/device/function address which may change
3755 if new hardware is inserted, if motherboard
3756 firmware changes, or due to changes caused
3757 by other kernel parameters. If the
3758 domain is left unspecified, it is
3759 taken to be zero. Optionally, a path
3760 to a device through multiple device/function
3761 addresses can be specified after the base
3762 address (this is more robust against
3763 renumbering issues). The second format
3764 selects devices using IDs from the
3765 configuration space which may match multiple
3766 devices in the system.
3768 earlydump dump PCI config space before the kernel
3770 off [X86] don't probe for the PCI bus
3771 bios [X86-32] force use of PCI BIOS, don't access
3772 the hardware directly. Use this if your machine
3773 has a non-standard PCI host bridge.
3774 nobios [X86-32] disallow use of PCI BIOS, only direct
3775 hardware access methods are allowed. Use this
3776 if you experience crashes upon bootup and you
3777 suspect they are caused by the BIOS.
3778 conf1 [X86] Force use of PCI Configuration Access
3779 Mechanism 1 (config address in IO port 0xCF8,
3780 data in IO port 0xCFC, both 32-bit).
3781 conf2 [X86] Force use of PCI Configuration Access
3782 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3783 the function, IO port 0xCFA, also 8-bit, sets
3784 bus number. The config space is then accessed
3785 through ports 0xC000-0xCFFF).
3786 See http://wiki.osdev.org/PCI for more info
3787 on the configuration access mechanisms.
3788 noaer [PCIE] If the PCIEAER kernel config parameter is
3789 enabled, this kernel boot option can be used to
3790 disable the use of PCIE advanced error reporting.
3791 nodomains [PCI] Disable support for multiple PCI
3792 root domains (aka PCI segments, in ACPI-speak).
3793 nommconf [X86] Disable use of MMCONFIG for PCI
3795 check_enable_amd_mmconf [X86] check for and enable
3796 properly configured MMIO access to PCI
3797 config space on AMD family 10h CPU
3798 nomsi [MSI] If the PCI_MSI kernel config parameter is
3799 enabled, this kernel boot option can be used to
3800 disable the use of MSI interrupts system-wide.
3801 noioapicquirk [APIC] Disable all boot interrupt quirks.
3802 Safety option to keep boot IRQs enabled. This
3803 should never be necessary.
3804 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3805 primary IO-APIC for bridges that cannot disable
3806 boot IRQs. This fixes a source of spurious IRQs
3807 when the system masks IRQs.
3808 noioapicreroute [APIC] Disable workaround that uses the
3809 boot IRQ equivalent of an IRQ that connects to
3810 a chipset where boot IRQs cannot be disabled.
3811 The opposite of ioapicreroute.
3812 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3813 routing table. These calls are known to be buggy
3814 on several machines and they hang the machine
3815 when used, but on other computers it's the only
3816 way to get the interrupt routing table. Try
3817 this option if the kernel is unable to allocate
3818 IRQs or discover secondary PCI buses on your
3820 rom [X86] Assign address space to expansion ROMs.
3821 Use with caution as certain devices share
3822 address decoders between ROMs and other
3824 norom [X86] Do not assign address space to
3825 expansion ROMs that do not already have
3826 BIOS assigned address ranges.
3827 nobar [X86] Do not assign address space to the
3828 BARs that weren't assigned by the BIOS.
3829 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3830 assigned automatically to PCI devices. You can
3831 make the kernel exclude IRQs of your ISA cards
3833 pirqaddr=0xAAAAA [X86] Specify the physical address
3834 of the PIRQ table (normally generated
3835 by the BIOS) if it is outside the
3836 F0000h-100000h range.
3837 lastbus=N [X86] Scan all buses thru bus #N. Can be
3838 useful if the kernel is unable to find your
3839 secondary buses and you want to tell it
3840 explicitly which ones they are.
3841 assign-busses [X86] Always assign all PCI bus
3842 numbers ourselves, overriding
3843 whatever the firmware may have done.
3844 usepirqmask [X86] Honor the possible IRQ mask stored
3845 in the BIOS $PIR table. This is needed on
3846 some systems with broken BIOSes, notably
3847 some HP Pavilion N5400 and Omnibook XE3
3848 notebooks. This will have no effect if ACPI
3849 IRQ routing is enabled.
3850 noacpi [X86] Do not use ACPI for IRQ routing
3851 or for PCI scanning.
3852 use_crs [X86] Use PCI host bridge window information
3853 from ACPI. On BIOSes from 2008 or later, this
3854 is enabled by default. If you need to use this,
3855 please report a bug.
3856 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3857 If you need to use this, please report a bug.
3858 routeirq Do IRQ routing for all PCI devices.
3859 This is normally done in pci_enable_device(),
3860 so this option is a temporary workaround
3861 for broken drivers that don't call it.
3862 skip_isa_align [X86] do not align io start addr, so can
3863 handle more pci cards
3864 noearly [X86] Don't do any early type 1 scanning.
3865 This might help on some broken boards which
3866 machine check when some devices' config space
3867 is read. But various workarounds are disabled
3868 and some IOMMU drivers will not work.
3869 bfsort Sort PCI devices into breadth-first order.
3870 This sorting is done to get a device
3871 order compatible with older (<= 2.4) kernels.
3872 nobfsort Don't sort PCI devices into breadth-first order.
3873 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3874 tuning and use the BIOS-configured MPS defaults.
3875 pcie_bus_safe Set every device's MPS to the largest value
3876 supported by all devices below the root complex.
3877 pcie_bus_perf Set device MPS to the largest allowable MPS
3878 based on its parent bus. Also set MRRS (Max
3879 Read Request Size) to the largest supported
3880 value (no larger than the MPS that the device
3881 or bus can support) for best performance.
3882 pcie_bus_peer2peer Set every device's MPS to 128B, which
3883 every device is guaranteed to support. This
3884 configuration allows peer-to-peer DMA between
3885 any pair of devices, possibly at the cost of
3886 reduced performance. This also guarantees
3887 that hot-added devices will work.
3888 cbiosize=nn[KMG] The fixed amount of bus space which is
3889 reserved for the CardBus bridge's IO window.
3890 The default value is 256 bytes.
3891 cbmemsize=nn[KMG] The fixed amount of bus space which is
3892 reserved for the CardBus bridge's memory
3893 window. The default value is 64 megabytes.
3896 [<order of align>@]<pci_dev>[; ...]
3897 Specifies alignment and device to reassign
3898 aligned memory resources. How to
3899 specify the device is described above.
3900 If <order of align> is not specified,
3901 PAGE_SIZE is used as alignment.
3902 A PCI-PCI bridge can be specified if resource
3903 windows need to be expanded.
3904 To specify the alignment for several
3905 instances of a device, the PCI vendor,
3906 device, subvendor, and subdevice may be
3907 specified, e.g., 12@pci:8086:9c22:103c:198f
3908 for 4096-byte alignment.
3909 ecrc= Enable/disable PCIe ECRC (transaction layer
3910 end-to-end CRC checking).
3911 bios: Use BIOS/firmware settings. This is the
3915 hpiosize=nn[KMG] The fixed amount of bus space which is
3916 reserved for hotplug bridge's IO window.
3917 Default size is 256 bytes.
3918 hpmmiosize=nn[KMG] The fixed amount of bus space which is
3919 reserved for hotplug bridge's MMIO window.
3920 Default size is 2 megabytes.
3921 hpmmioprefsize=nn[KMG] The fixed amount of bus space which is
3922 reserved for hotplug bridge's MMIO_PREF window.
3923 Default size is 2 megabytes.
3924 hpmemsize=nn[KMG] The fixed amount of bus space which is
3925 reserved for hotplug bridge's MMIO and
3927 Default size is 2 megabytes.
3928 hpbussize=nn The minimum amount of additional bus numbers
3929 reserved for buses below a hotplug bridge.
3931 realloc= Enable/disable reallocating PCI bridge resources
3932 if allocations done by BIOS are too small to
3933 accommodate resources required by all child
3935 off: Turn realloc off
3937 realloc same as realloc=on
3938 noari do not use PCIe ARI.
3939 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3940 do not use PCIe ATS (and IOMMU device IOTLB).
3941 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3942 only look for one device below a PCIe downstream
3944 big_root_window Try to add a big 64bit memory window to the PCIe
3945 root complex on AMD CPUs. Some GFX hardware
3946 can resize a BAR to allow access to all VRAM.
3947 Adding the window is slightly risky (it may
3948 conflict with unreported devices), so this
3950 disable_acs_redir=<pci_dev>[; ...]
3951 Specify one or more PCI devices (in the format
3952 specified above) separated by semicolons.
3953 Each device specified will have the PCI ACS
3954 redirect capabilities forced off which will
3955 allow P2P traffic between devices through
3956 bridges without forcing it upstream. Note:
3957 this removes isolation between devices and
3958 may put more devices in an IOMMU group.
3959 force_floating [S390] Force usage of floating interrupts.
3960 nomio [S390] Do not use MIO instructions.
3961 norid [S390] ignore the RID field and force use of
3962 one PCI domain per PCI function
3964 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3967 force Enable ASPM even on devices that claim not to support it.
3968 WARNING: Forcing ASPM on may cause system lockups.
3970 pcie_ports= [PCIE] PCIe port services handling:
3971 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3972 even if the platform doesn't give the OS permission to
3973 use them. This may cause conflicts if the platform
3974 also tries to use these services.
3975 dpc-native Use native PCIe service for DPC only. May
3976 cause conflicts if firmware uses AER or DPC.
3977 compat Disable native PCIe services (PME, AER, DPC, PCIe
3980 pcie_port_pm= [PCIE] PCIe port power management handling:
3981 off Disable power management of all PCIe ports
3982 force Forcibly enable power management of all PCIe ports
3984 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3985 nomsi Do not use MSI for native PCIe PME signaling (this makes
3986 all PCIe root ports use INTx for all services).
3988 pcmv= [HW,PCMCIA] BadgePAD 4
3992 Keep all power-domains already enabled by bootloader on,
3993 even if no driver has claimed them. This is useful
3994 for debug and development, but should not be
3995 needed on a platform with proper driver support.
3998 See Documentation/admin-guide/blockdev/paride.rst.
4000 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
4003 See arch/parisc/kernel/pdc_chassis.c
4005 percpu_alloc= Select which percpu first chunk allocator to use.
4006 Currently supported values are "embed" and "page".
4007 Archs may support subset or none of the selections.
4008 See comments in mm/percpu.c for details on each
4009 allocator. This parameter is primarily for debugging
4010 and performance comparison.
4013 See Documentation/admin-guide/blockdev/paride.rst.
4016 See Documentation/admin-guide/blockdev/paride.rst.
4018 pirq= [SMP,APIC] Manual mp-table setup
4019 See Documentation/x86/i386/IO-APIC.rst.
4021 plip= [PPT,NET] Parallel port network link
4022 Format: { parport<nr> | timid | 0 }
4023 See also Documentation/admin-guide/parport.rst.
4025 pmtmr= [X86] Manual setup of pmtmr I/O Port.
4026 Override pmtimer IOPort with a hex value.
4029 pm_debug_messages [SUSPEND,KNL]
4030 Enable suspend/resume debug messages during boot up.
4033 Enable PNP debug messages (depends on the
4034 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
4035 via /sys/module/pnp/parameters/debug. We always show
4036 current resource usage; turning this on also shows
4037 possible settings and some assignment information.
4043 { on | off | curr | res | no-curr | no-res }
4046 [ISAPNP] Exclude IRQs for the autoconfiguration
4049 [ISAPNP] Exclude DMAs for the autoconfiguration
4051 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
4052 Ranges are in pairs (I/O port base and size).
4055 [ISAPNP] Exclude memory regions for the
4057 Ranges are in pairs (memory base and size).
4059 ports= [IP_VS_FTP] IPVS ftp helper module
4061 Up to 8 (IP_VS_APP_MAX_PORTS) ports
4063 Format: <port>,<port>....
4065 powersave=off [PPC] This option disables power saving features.
4066 It specifically disables cpuidle and sets the
4067 platform machine description specific power_save
4068 function to NULL. On Idle the CPU just reduces
4071 ppc_strict_facility_enable
4072 [PPC] This option catches any kernel floating point,
4073 Altivec, VSX and SPE outside of regions specifically
4074 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
4075 There is some performance impact when enabling this.
4079 Disable Hardware Transactional Memory
4082 Select preemption mode if you have CONFIG_PREEMPT_DYNAMIC
4083 none - Limited to cond_resched() calls
4084 voluntary - Limited to cond_resched() and might_sleep() calls
4085 full - Any section that isn't explicitly preempt disabled
4086 can be preempted anytime.
4088 print-fatal-signals=
4089 [KNL] debug: print fatal signals
4091 If enabled, warn about various signal handling
4092 related application anomalies: too many signals,
4093 too many POSIX.1 timers, fatal signals causing a
4096 If you hit the warning due to signal overflow,
4097 you might want to try "ulimit -i unlimited".
4101 printk.always_kmsg_dump=
4102 Trigger kmsg_dump for cases other than kernel oops or
4104 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4107 printk.devkmsg={on,off,ratelimit}
4108 Control writing to /dev/kmsg.
4109 on - unlimited logging to /dev/kmsg from userspace
4110 off - logging to /dev/kmsg disabled
4111 ratelimit - ratelimit the logging
4114 printk.time= Show timing data prefixed to each printk message line
4115 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4117 processor.max_cstate= [HW,ACPI]
4118 Limit processor to maximum C-state
4119 max_cstate=9 overrides any DMI blacklist limit.
4121 processor.nocst [HW,ACPI]
4122 Ignore the _CST method to determine C-states,
4123 instead using the legacy FADT method
4125 profile= [KNL] Enable kernel profiling via /proc/profile
4126 Format: [<profiletype>,]<number>
4127 Param: <profiletype>: "schedule", "sleep", or "kvm"
4128 [defaults to kernel profiling]
4129 Param: "schedule" - profile schedule points.
4130 Param: "sleep" - profile D-state sleeping (millisecs).
4131 Requires CONFIG_SCHEDSTATS
4132 Param: "kvm" - profile VM exits.
4133 Param: <number> - step/bucket size as a power of 2 for
4134 statistical time based profiling.
4136 prompt_ramdisk= [RAM] [Deprecated]
4138 prot_virt= [S390] enable hosting protected virtual machines
4139 isolated from the hypervisor (if hardware supports
4143 psi= [KNL] Enable or disable pressure stall information
4147 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
4148 probe for; one of (bare|imps|exps|lifebook|any).
4149 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
4151 psmouse.resetafter= [HW,MOUSE]
4152 Try to reset the device after so many bad packets
4155 [HW,MOUSE] Set desired mouse resolution, in dpi.
4156 psmouse.smartscroll=
4157 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
4158 0 = disabled, 1 = enabled (default).
4160 pstore.backend= Specify the name of the pstore backend to use
4163 See Documentation/admin-guide/blockdev/paride.rst.
4165 pti= [X86-64] Control Page Table Isolation of user and
4166 kernel address spaces. Disabling this feature
4167 removes hardening, but improves performance of
4168 system calls and interrupts.
4170 on - unconditionally enable
4171 off - unconditionally disable
4172 auto - kernel detects whether your CPU model is
4173 vulnerable to issues that PTI mitigates
4175 Not specifying this option is equivalent to pti=auto.
4178 Equivalent to pti=off
4181 [KNL] Number of legacy pty's. Overwrites compiled-in
4184 quiet [KNL] Disable most log messages
4189 See Documentation/admin-guide/md.rst.
4191 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
4192 See Documentation/admin-guide/blockdev/ramdisk.rst.
4194 ramdisk_start= [RAM] RAM disk image start address
4196 random.trust_cpu={on,off}
4197 [KNL] Enable or disable trusting the use of the
4198 CPU's random number generator (if available) to
4199 fully seed the kernel's CRNG. Default is controlled
4200 by CONFIG_RANDOM_TRUST_CPU.
4202 randomize_kstack_offset=
4203 [KNL] Enable or disable kernel stack offset
4204 randomization, which provides roughly 5 bits of
4205 entropy, frustrating memory corruption attacks
4206 that depend on stack address determinism or
4207 cross-syscall address exposures. This is only
4208 available on architectures that have defined
4209 CONFIG_HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET.
4210 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
4211 Default is CONFIG_RANDOMIZE_KSTACK_OFFSET_DEFAULT.
4213 ras=option[,option,...] [KNL] RAS-specific options
4216 Disable the Correctable Errors Collector,
4217 see CONFIG_RAS_CEC help text.
4220 The argument is a cpu list, as described above.
4222 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
4223 the specified list of CPUs to be no-callback CPUs.
4224 Invocation of these CPUs' RCU callbacks will be
4225 offloaded to "rcuox/N" kthreads created for that
4226 purpose, where "x" is "p" for RCU-preempt, and
4227 "s" for RCU-sched, and "N" is the CPU number.
4228 This reduces OS jitter on the offloaded CPUs,
4229 which can be useful for HPC and real-time
4230 workloads. It can also improve energy efficiency
4231 for asymmetric multiprocessors.
4234 Rather than requiring that offloaded CPUs
4235 (specified by rcu_nocbs= above) explicitly
4236 awaken the corresponding "rcuoN" kthreads,
4237 make these kthreads poll for callbacks.
4238 This improves the real-time response for the
4239 offloaded CPUs by relieving them of the need to
4240 wake up the corresponding kthread, but degrades
4241 energy efficiency by requiring that the kthreads
4242 periodically wake up to do the polling.
4244 rcutree.blimit= [KNL]
4245 Set maximum number of finished RCU callbacks to
4246 process in one batch.
4248 rcutree.dump_tree= [KNL]
4249 Dump the structure of the rcu_node combining tree
4250 out at early boot. This is used for diagnostic
4251 purposes, to verify correct tree setup.
4253 rcutree.gp_cleanup_delay= [KNL]
4254 Set the number of jiffies to delay each step of
4255 RCU grace-period cleanup.
4257 rcutree.gp_init_delay= [KNL]
4258 Set the number of jiffies to delay each step of
4259 RCU grace-period initialization.
4261 rcutree.gp_preinit_delay= [KNL]
4262 Set the number of jiffies to delay each step of
4263 RCU grace-period pre-initialization, that is,
4264 the propagation of recent CPU-hotplug changes up
4265 the rcu_node combining tree.
4267 rcutree.use_softirq= [KNL]
4268 If set to zero, move all RCU_SOFTIRQ processing to
4269 per-CPU rcuc kthreads. Defaults to a non-zero
4270 value, meaning that RCU_SOFTIRQ is used by default.
4271 Specify rcutree.use_softirq=0 to use rcuc kthreads.
4273 But note that CONFIG_PREEMPT_RT=y kernels disable
4274 this kernel boot parameter, forcibly setting it
4277 rcutree.rcu_fanout_exact= [KNL]
4278 Disable autobalancing of the rcu_node combining
4279 tree. This is used by rcutorture, and might
4280 possibly be useful for architectures having high
4281 cache-to-cache transfer latencies.
4283 rcutree.rcu_fanout_leaf= [KNL]
4284 Change the number of CPUs assigned to each
4285 leaf rcu_node structure. Useful for very
4286 large systems, which will choose the value 64,
4287 and for NUMA systems with large remote-access
4288 latencies, which will choose a value aligned
4289 with the appropriate hardware boundaries.
4291 rcutree.rcu_min_cached_objs= [KNL]
4292 Minimum number of objects which are cached and
4293 maintained per one CPU. Object size is equal
4294 to PAGE_SIZE. The cache allows to reduce the
4295 pressure to page allocator, also it makes the
4296 whole algorithm to behave better in low memory
4299 rcutree.jiffies_till_first_fqs= [KNL]
4300 Set delay from grace-period initialization to
4301 first attempt to force quiescent states.
4302 Units are jiffies, minimum value is zero,
4303 and maximum value is HZ.
4305 rcutree.jiffies_till_next_fqs= [KNL]
4306 Set delay between subsequent attempts to force
4307 quiescent states. Units are jiffies, minimum
4308 value is one, and maximum value is HZ.
4310 rcutree.jiffies_till_sched_qs= [KNL]
4311 Set required age in jiffies for a
4312 given grace period before RCU starts
4313 soliciting quiescent-state help from
4314 rcu_note_context_switch() and cond_resched().
4315 If not specified, the kernel will calculate
4316 a value based on the most recent settings
4317 of rcutree.jiffies_till_first_fqs
4318 and rcutree.jiffies_till_next_fqs.
4319 This calculated value may be viewed in
4320 rcutree.jiffies_to_sched_qs. Any attempt to set
4321 rcutree.jiffies_to_sched_qs will be cheerfully
4324 rcutree.kthread_prio= [KNL,BOOT]
4325 Set the SCHED_FIFO priority of the RCU per-CPU
4326 kthreads (rcuc/N). This value is also used for
4327 the priority of the RCU boost threads (rcub/N)
4328 and for the RCU grace-period kthreads (rcu_bh,
4329 rcu_preempt, and rcu_sched). If RCU_BOOST is
4330 set, valid values are 1-99 and the default is 1
4331 (the least-favored priority). Otherwise, when
4332 RCU_BOOST is not set, valid values are 0-99 and
4333 the default is zero (non-realtime operation).
4335 rcutree.rcu_nocb_gp_stride= [KNL]
4336 Set the number of NOCB callback kthreads in
4337 each group, which defaults to the square root
4338 of the number of CPUs. Larger numbers reduce
4339 the wakeup overhead on the global grace-period
4340 kthread, but increases that same overhead on
4341 each group's NOCB grace-period kthread.
4343 rcutree.qhimark= [KNL]
4344 Set threshold of queued RCU callbacks beyond which
4345 batch limiting is disabled.
4347 rcutree.qlowmark= [KNL]
4348 Set threshold of queued RCU callbacks below which
4349 batch limiting is re-enabled.
4351 rcutree.qovld= [KNL]
4352 Set threshold of queued RCU callbacks beyond which
4353 RCU's force-quiescent-state scan will aggressively
4354 enlist help from cond_resched() and sched IPIs to
4355 help CPUs more quickly reach quiescent states.
4356 Set to less than zero to make this be set based
4357 on rcutree.qhimark at boot time and to zero to
4358 disable more aggressive help enlistment.
4360 rcutree.rcu_idle_gp_delay= [KNL]
4361 Set wakeup interval for idle CPUs that have
4362 RCU callbacks (RCU_FAST_NO_HZ=y).
4364 rcutree.rcu_kick_kthreads= [KNL]
4365 Cause the grace-period kthread to get an extra
4366 wake_up() if it sleeps three times longer than
4367 it should at force-quiescent-state time.
4368 This wake_up() will be accompanied by a
4369 WARN_ONCE() splat and an ftrace_dump().
4371 rcutree.rcu_unlock_delay= [KNL]
4372 In CONFIG_RCU_STRICT_GRACE_PERIOD=y kernels,
4373 this specifies an rcu_read_unlock()-time delay
4374 in microseconds. This defaults to zero.
4375 Larger delays increase the probability of
4376 catching RCU pointer leaks, that is, buggy use
4377 of RCU-protected pointers after the relevant
4378 rcu_read_unlock() has completed.
4380 rcutree.sysrq_rcu= [KNL]
4381 Commandeer a sysrq key to dump out Tree RCU's
4382 rcu_node tree with an eye towards determining
4383 why a new grace period has not yet started.
4385 rcuscale.gp_async= [KNL]
4386 Measure performance of asynchronous
4387 grace-period primitives such as call_rcu().
4389 rcuscale.gp_async_max= [KNL]
4390 Specify the maximum number of outstanding
4391 callbacks per writer thread. When a writer
4392 thread exceeds this limit, it invokes the
4393 corresponding flavor of rcu_barrier() to allow
4394 previously posted callbacks to drain.
4396 rcuscale.gp_exp= [KNL]
4397 Measure performance of expedited synchronous
4398 grace-period primitives.
4400 rcuscale.holdoff= [KNL]
4401 Set test-start holdoff period. The purpose of
4402 this parameter is to delay the start of the
4403 test until boot completes in order to avoid
4406 rcuscale.kfree_rcu_test= [KNL]
4407 Set to measure performance of kfree_rcu() flooding.
4409 rcuscale.kfree_rcu_test_double= [KNL]
4410 Test the double-argument variant of kfree_rcu().
4411 If this parameter has the same value as
4412 rcuscale.kfree_rcu_test_single, both the single-
4413 and double-argument variants are tested.
4415 rcuscale.kfree_rcu_test_single= [KNL]
4416 Test the single-argument variant of kfree_rcu().
4417 If this parameter has the same value as
4418 rcuscale.kfree_rcu_test_double, both the single-
4419 and double-argument variants are tested.
4421 rcuscale.kfree_nthreads= [KNL]
4422 The number of threads running loops of kfree_rcu().
4424 rcuscale.kfree_alloc_num= [KNL]
4425 Number of allocations and frees done in an iteration.
4427 rcuscale.kfree_loops= [KNL]
4428 Number of loops doing rcuscale.kfree_alloc_num number
4429 of allocations and frees.
4431 rcuscale.nreaders= [KNL]
4432 Set number of RCU readers. The value -1 selects
4433 N, where N is the number of CPUs. A value
4434 "n" less than -1 selects N-n+1, where N is again
4435 the number of CPUs. For example, -2 selects N
4436 (the number of CPUs), -3 selects N+1, and so on.
4437 A value of "n" less than or equal to -N selects
4440 rcuscale.nwriters= [KNL]
4441 Set number of RCU writers. The values operate
4442 the same as for rcuscale.nreaders.
4443 N, where N is the number of CPUs
4445 rcuscale.perf_type= [KNL]
4446 Specify the RCU implementation to test.
4448 rcuscale.shutdown= [KNL]
4449 Shut the system down after performance tests
4450 complete. This is useful for hands-off automated
4453 rcuscale.verbose= [KNL]
4454 Enable additional printk() statements.
4456 rcuscale.writer_holdoff= [KNL]
4457 Write-side holdoff between grace periods,
4458 in microseconds. The default of zero says
4461 rcutorture.fqs_duration= [KNL]
4462 Set duration of force_quiescent_state bursts
4465 rcutorture.fqs_holdoff= [KNL]
4466 Set holdoff time within force_quiescent_state bursts
4469 rcutorture.fqs_stutter= [KNL]
4470 Set wait time between force_quiescent_state bursts
4473 rcutorture.fwd_progress= [KNL]
4474 Enable RCU grace-period forward-progress testing
4475 for the types of RCU supporting this notion.
4477 rcutorture.fwd_progress_div= [KNL]
4478 Specify the fraction of a CPU-stall-warning
4479 period to do tight-loop forward-progress testing.
4481 rcutorture.fwd_progress_holdoff= [KNL]
4482 Number of seconds to wait between successive
4483 forward-progress tests.
4485 rcutorture.fwd_progress_need_resched= [KNL]
4486 Enclose cond_resched() calls within checks for
4487 need_resched() during tight-loop forward-progress
4490 rcutorture.gp_cond= [KNL]
4491 Use conditional/asynchronous update-side
4492 primitives, if available.
4494 rcutorture.gp_exp= [KNL]
4495 Use expedited update-side primitives, if available.
4497 rcutorture.gp_normal= [KNL]
4498 Use normal (non-expedited) asynchronous
4499 update-side primitives, if available.
4501 rcutorture.gp_sync= [KNL]
4502 Use normal (non-expedited) synchronous
4503 update-side primitives, if available. If all
4504 of rcutorture.gp_cond=, rcutorture.gp_exp=,
4505 rcutorture.gp_normal=, and rcutorture.gp_sync=
4506 are zero, rcutorture acts as if is interpreted
4507 they are all non-zero.
4509 rcutorture.irqreader= [KNL]
4510 Run RCU readers from irq handlers, or, more
4511 accurately, from a timer handler. Not all RCU
4512 flavors take kindly to this sort of thing.
4514 rcutorture.leakpointer= [KNL]
4515 Leak an RCU-protected pointer out of the reader.
4516 This can of course result in splats, and is
4517 intended to test the ability of things like
4518 CONFIG_RCU_STRICT_GRACE_PERIOD=y to detect
4521 rcutorture.n_barrier_cbs= [KNL]
4522 Set callbacks/threads for rcu_barrier() testing.
4524 rcutorture.nfakewriters= [KNL]
4525 Set number of concurrent RCU writers. These just
4526 stress RCU, they don't participate in the actual
4527 test, hence the "fake".
4529 rcutorture.nocbs_nthreads= [KNL]
4530 Set number of RCU callback-offload togglers.
4531 Zero (the default) disables toggling.
4533 rcutorture.nocbs_toggle= [KNL]
4534 Set the delay in milliseconds between successive
4535 callback-offload toggling attempts.
4537 rcutorture.nreaders= [KNL]
4538 Set number of RCU readers. The value -1 selects
4539 N-1, where N is the number of CPUs. A value
4540 "n" less than -1 selects N-n-2, where N is again
4541 the number of CPUs. For example, -2 selects N
4542 (the number of CPUs), -3 selects N+1, and so on.
4544 rcutorture.object_debug= [KNL]
4545 Enable debug-object double-call_rcu() testing.
4547 rcutorture.onoff_holdoff= [KNL]
4548 Set time (s) after boot for CPU-hotplug testing.
4550 rcutorture.onoff_interval= [KNL]
4551 Set time (jiffies) between CPU-hotplug operations,
4552 or zero to disable CPU-hotplug testing.
4554 rcutorture.read_exit= [KNL]
4555 Set the number of read-then-exit kthreads used
4556 to test the interaction of RCU updaters and
4557 task-exit processing.
4559 rcutorture.read_exit_burst= [KNL]
4560 The number of times in a given read-then-exit
4561 episode that a set of read-then-exit kthreads
4564 rcutorture.read_exit_delay= [KNL]
4565 The delay, in seconds, between successive
4566 read-then-exit testing episodes.
4568 rcutorture.shuffle_interval= [KNL]
4569 Set task-shuffle interval (s). Shuffling tasks
4570 allows some CPUs to go into dyntick-idle mode
4571 during the rcutorture test.
4573 rcutorture.shutdown_secs= [KNL]
4574 Set time (s) after boot system shutdown. This
4575 is useful for hands-off automated testing.
4577 rcutorture.stall_cpu= [KNL]
4578 Duration of CPU stall (s) to test RCU CPU stall
4579 warnings, zero to disable.
4581 rcutorture.stall_cpu_block= [KNL]
4582 Sleep while stalling if set. This will result
4583 in warnings from preemptible RCU in addition
4584 to any other stall-related activity.
4586 rcutorture.stall_cpu_holdoff= [KNL]
4587 Time to wait (s) after boot before inducing stall.
4589 rcutorture.stall_cpu_irqsoff= [KNL]
4590 Disable interrupts while stalling if set.
4592 rcutorture.stall_gp_kthread= [KNL]
4593 Duration (s) of forced sleep within RCU
4594 grace-period kthread to test RCU CPU stall
4595 warnings, zero to disable. If both stall_cpu
4596 and stall_gp_kthread are specified, the
4597 kthread is starved first, then the CPU.
4599 rcutorture.stat_interval= [KNL]
4600 Time (s) between statistics printk()s.
4602 rcutorture.stutter= [KNL]
4603 Time (s) to stutter testing, for example, specifying
4604 five seconds causes the test to run for five seconds,
4605 wait for five seconds, and so on. This tests RCU's
4606 ability to transition abruptly to and from idle.
4608 rcutorture.test_boost= [KNL]
4609 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
4610 "Maybe" means test if the RCU implementation
4611 under test support RCU priority boosting.
4613 rcutorture.test_boost_duration= [KNL]
4614 Duration (s) of each individual boost test.
4616 rcutorture.test_boost_interval= [KNL]
4617 Interval (s) between each boost test.
4619 rcutorture.test_no_idle_hz= [KNL]
4620 Test RCU's dyntick-idle handling. See also the
4621 rcutorture.shuffle_interval parameter.
4623 rcutorture.torture_type= [KNL]
4624 Specify the RCU implementation to test.
4626 rcutorture.verbose= [KNL]
4627 Enable additional printk() statements.
4629 rcupdate.rcu_cpu_stall_ftrace_dump= [KNL]
4630 Dump ftrace buffer after reporting RCU CPU
4633 rcupdate.rcu_cpu_stall_suppress= [KNL]
4634 Suppress RCU CPU stall warning messages.
4636 rcupdate.rcu_cpu_stall_suppress_at_boot= [KNL]
4637 Suppress RCU CPU stall warning messages and
4638 rcutorture writer stall warnings that occur
4639 during early boot, that is, during the time
4640 before the init task is spawned.
4642 rcupdate.rcu_cpu_stall_timeout= [KNL]
4643 Set timeout for RCU CPU stall warning messages.
4645 rcupdate.rcu_expedited= [KNL]
4646 Use expedited grace-period primitives, for
4647 example, synchronize_rcu_expedited() instead
4648 of synchronize_rcu(). This reduces latency,
4649 but can increase CPU utilization, degrade
4650 real-time latency, and degrade energy efficiency.
4651 No effect on CONFIG_TINY_RCU kernels.
4653 rcupdate.rcu_normal= [KNL]
4654 Use only normal grace-period primitives,
4655 for example, synchronize_rcu() instead of
4656 synchronize_rcu_expedited(). This improves
4657 real-time latency, CPU utilization, and
4658 energy efficiency, but can expose users to
4659 increased grace-period latency. This parameter
4660 overrides rcupdate.rcu_expedited. No effect on
4661 CONFIG_TINY_RCU kernels.
4663 rcupdate.rcu_normal_after_boot= [KNL]
4664 Once boot has completed (that is, after
4665 rcu_end_inkernel_boot() has been invoked), use
4666 only normal grace-period primitives. No effect
4667 on CONFIG_TINY_RCU kernels.
4669 But note that CONFIG_PREEMPT_RT=y kernels enables
4670 this kernel boot parameter, forcibly setting
4671 it to the value one, that is, converting any
4672 post-boot attempt at an expedited RCU grace
4673 period to instead use normal non-expedited
4674 grace-period processing.
4676 rcupdate.rcu_task_ipi_delay= [KNL]
4677 Set time in jiffies during which RCU tasks will
4678 avoid sending IPIs, starting with the beginning
4679 of a given grace period. Setting a large
4680 number avoids disturbing real-time workloads,
4681 but lengthens grace periods.
4683 rcupdate.rcu_task_stall_timeout= [KNL]
4684 Set timeout in jiffies for RCU task stall warning
4685 messages. Disable with a value less than or equal
4688 rcupdate.rcu_self_test= [KNL]
4689 Run the RCU early boot self tests
4693 Run specified binary instead of /init from the ramdisk,
4694 used for early userspace startup. See initrd.
4697 force - Override the decision by the kernel to hide the
4698 advertisement of RDRAND support (this affects
4699 certain AMD processors because of buggy BIOS
4700 support, specifically around the suspend/resume
4704 Turn on/off individual RDT features. List is:
4705 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4707 E.g. to turn on cmt and turn off mba use:
4711 Format (x86 or x86_64):
4712 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4714 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4716 Where reboot_mode is one of warm (soft) or cold (hard) or gpio
4717 (prefix with 'panic_' to set mode for panic
4719 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4720 reboot_force is either force or not specified,
4721 reboot_cpu is s[mp]#### with #### being the processor
4722 to be used for rebooting.
4724 refscale.holdoff= [KNL]
4725 Set test-start holdoff period. The purpose of
4726 this parameter is to delay the start of the
4727 test until boot completes in order to avoid
4730 refscale.loops= [KNL]
4731 Set the number of loops over the synchronization
4732 primitive under test. Increasing this number
4733 reduces noise due to loop start/end overhead,
4734 but the default has already reduced the per-pass
4735 noise to a handful of picoseconds on ca. 2020
4738 refscale.nreaders= [KNL]
4739 Set number of readers. The default value of -1
4740 selects N, where N is roughly 75% of the number
4741 of CPUs. A value of zero is an interesting choice.
4743 refscale.nruns= [KNL]
4744 Set number of runs, each of which is dumped onto
4747 refscale.readdelay= [KNL]
4748 Set the read-side critical-section duration,
4749 measured in microseconds.
4751 refscale.scale_type= [KNL]
4752 Specify the read-protection implementation to test.
4754 refscale.shutdown= [KNL]
4755 Shut down the system at the end of the performance
4756 test. This defaults to 1 (shut it down) when
4757 refscale is built into the kernel and to 0 (leave
4758 it running) when refscale is built as a module.
4760 refscale.verbose= [KNL]
4761 Enable additional printk() statements.
4763 refscale.verbose_batched= [KNL]
4764 Batch the additional printk() statements. If zero
4765 (the default) or negative, print everything. Otherwise,
4766 print every Nth verbose statement, where N is the value
4770 [KNL, SMP] Set scheduler's default relax_domain_level.
4771 See Documentation/admin-guide/cgroup-v1/cpusets.rst.
4773 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4774 Format: <base1>,<size1>[,<base2>,<size2>,...]
4775 Reserve I/O ports or memory so the kernel won't use
4776 them. If <base> is less than 0x10000, the region
4777 is assumed to be I/O ports; otherwise it is memory.
4779 reservetop= [X86-32]
4781 Reserves a hole at the top of the kernel virtual
4786 Set the amount of memory to reserve for BIOS at
4787 the bottom of the address space.
4789 reset_devices [KNL] Force drivers to reset the underlying device
4790 during initialization.
4793 Specify the partition device for software suspend
4795 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4797 resume_offset= [SWSUSP]
4798 Specify the offset from the beginning of the partition
4799 given by "resume=" at which the swap header is located,
4800 in <PAGE_SIZE> units (needed only for swap files).
4801 See Documentation/power/swsusp-and-swap-files.rst
4803 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4804 read the resume files
4806 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4807 Useful for devices that are detected asynchronously
4808 (e.g. USB and MMC devices).
4810 hibernate= [HIBERNATION]
4811 noresume Don't check if there's a hibernation image
4812 present during boot.
4813 nocompress Don't compress/decompress hibernation images.
4814 no Disable hibernation and resume.
4815 protect_image Turn on image protection during restoration
4816 (that will set all pages holding image data
4817 during restoration read-only).
4819 retain_initrd [RAM] Keep initrd memory after extraction
4821 rfkill.default_state=
4822 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4823 etc. communication is blocked by default.
4826 rfkill.master_switch_mode=
4827 0 The "airplane mode" button does nothing.
4828 1 The "airplane mode" button toggles between everything
4829 blocked and the previous configuration.
4830 2 The "airplane mode" button toggles between everything
4831 blocked and everything unblocked.
4833 rhash_entries= [KNL,NET]
4834 Set number of hash buckets for route cache
4837 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4840 ro [KNL] Mount root device read-only on boot
4843 on Mark read-only kernel memory as read-only (default).
4844 off Leave read-only kernel memory writable for debugging.
4847 Enable the uart passthrough on the designated usb port
4848 on Rockchip SoCs. When active, the signals of the
4849 debug-uart get routed to the D+ and D- pins of the usb
4850 port and the regular usb controller gets disabled.
4852 root= [KNL] Root filesystem
4853 See name_to_dev_t comment in init/do_mounts.c.
4855 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4856 mount the root filesystem
4858 rootflags= [KNL] Set root filesystem mount option string
4860 rootfstype= [KNL] Set root filesystem type
4862 rootwait [KNL] Wait (indefinitely) for root device to show up.
4863 Useful for devices that are detected asynchronously
4864 (e.g. USB and MMC devices).
4866 rproc_mem=nn[KMG][@address]
4867 [KNL,ARM,CMA] Remoteproc physical memory block.
4868 Memory area to be used by remote processor image,
4871 rw [KNL] Mount root device read-write on boot
4873 S [KNL] Run init in single mode
4875 s390_iommu= [HW,S390]
4876 Set s390 IOTLB flushing mode
4878 With strict flushing every unmap operation will result in
4879 an IOTLB flush. Default is lazy flushing before reuse,
4883 See drivers/net/irda/sa1100_ir.c.
4885 sbni= [NET] Granch SBNI12 leased line adapter
4887 sched_verbose [KNL] Enables verbose scheduler debug messages.
4889 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4890 Allowed values are enable and disable. This feature
4891 incurs a small amount of overhead in the scheduler
4892 but is useful for debugging and performance tuning.
4894 sched_thermal_decay_shift=
4895 [KNL, SMP] Set a decay shift for scheduler thermal
4896 pressure signal. Thermal pressure signal follows the
4897 default decay period of other scheduler pelt
4898 signals(usually 32 ms but configurable). Setting
4899 sched_thermal_decay_shift will left shift the decay
4900 period for the thermal pressure signal by the shift
4902 i.e. with the default pelt decay period of 32 ms
4903 sched_thermal_decay_shift thermal pressure decay pr
4907 Format: integer between 0 and 10
4910 scftorture.holdoff= [KNL]
4911 Number of seconds to hold off before starting
4912 test. Defaults to zero for module insertion and
4913 to 10 seconds for built-in smp_call_function()
4916 scftorture.longwait= [KNL]
4917 Request ridiculously long waits randomly selected
4918 up to the chosen limit in seconds. Zero (the
4919 default) disables this feature. Please note
4920 that requesting even small non-zero numbers of
4921 seconds can result in RCU CPU stall warnings,
4922 softlockup complaints, and so on.
4924 scftorture.nthreads= [KNL]
4925 Number of kthreads to spawn to invoke the
4926 smp_call_function() family of functions.
4927 The default of -1 specifies a number of kthreads
4928 equal to the number of CPUs.
4930 scftorture.onoff_holdoff= [KNL]
4931 Number seconds to wait after the start of the
4932 test before initiating CPU-hotplug operations.
4934 scftorture.onoff_interval= [KNL]
4935 Number seconds to wait between successive
4936 CPU-hotplug operations. Specifying zero (which
4937 is the default) disables CPU-hotplug operations.
4939 scftorture.shutdown_secs= [KNL]
4940 The number of seconds following the start of the
4941 test after which to shut down the system. The
4942 default of zero avoids shutting down the system.
4943 Non-zero values are useful for automated tests.
4945 scftorture.stat_interval= [KNL]
4946 The number of seconds between outputting the
4947 current test statistics to the console. A value
4948 of zero disables statistics output.
4950 scftorture.stutter_cpus= [KNL]
4951 The number of jiffies to wait between each change
4952 to the set of CPUs under test.
4954 scftorture.use_cpus_read_lock= [KNL]
4955 Use use_cpus_read_lock() instead of the default
4956 preempt_disable() to disable CPU hotplug
4957 while invoking one of the smp_call_function*()
4960 scftorture.verbose= [KNL]
4961 Enable additional printk() statements.
4963 scftorture.weight_single= [KNL]
4964 The probability weighting to use for the
4965 smp_call_function_single() function with a zero
4966 "wait" parameter. A value of -1 selects the
4967 default if all other weights are -1. However,
4968 if at least one weight has some other value, a
4969 value of -1 will instead select a weight of zero.
4971 scftorture.weight_single_wait= [KNL]
4972 The probability weighting to use for the
4973 smp_call_function_single() function with a
4974 non-zero "wait" parameter. See weight_single.
4976 scftorture.weight_many= [KNL]
4977 The probability weighting to use for the
4978 smp_call_function_many() function with a zero
4979 "wait" parameter. See weight_single.
4980 Note well that setting a high probability for
4981 this weighting can place serious IPI load
4984 scftorture.weight_many_wait= [KNL]
4985 The probability weighting to use for the
4986 smp_call_function_many() function with a
4987 non-zero "wait" parameter. See weight_single
4990 scftorture.weight_all= [KNL]
4991 The probability weighting to use for the
4992 smp_call_function_all() function with a zero
4993 "wait" parameter. See weight_single and
4996 scftorture.weight_all_wait= [KNL]
4997 The probability weighting to use for the
4998 smp_call_function_all() function with a
4999 non-zero "wait" parameter. See weight_single
5002 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
5003 xtime_lock contention on larger systems, and/or RCU lock
5004 contention on all systems with CONFIG_MAXSMP set.
5005 Format: { "0" | "1" }
5006 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
5008 Note: increases power consumption, thus should only be
5009 enabled if running jitter sensitive (HPC/RT) workloads.
5011 security= [SECURITY] Choose a legacy "major" security module to
5012 enable at boot. This has been deprecated by the
5015 selinux= [SELINUX] Disable or enable SELinux at boot time.
5016 Format: { "0" | "1" }
5017 See security/selinux/Kconfig help text.
5022 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
5023 Format: { "0" | "1" }
5024 See security/apparmor/Kconfig help text
5027 Default value is set via kernel config option.
5029 serialnumber [BUGS=X86-32]
5032 Maximal number of shapers.
5040 Enable merging of slabs with similar size when the
5041 kernel is built without CONFIG_SLAB_MERGE_DEFAULT.
5044 Disable merging of slabs with similar size. May be
5045 necessary if there is some reason to distinguish
5046 allocs to different slabs, especially in hardened
5047 environments where the risk of heap overflows and
5048 layout control by attackers can usually be
5049 frustrated by disabling merging. This will reduce
5050 most of the exposure of a heap attack to a single
5051 cache (risks via metadata attacks are mostly
5052 unchanged). Debug options disable merging on their
5054 For more information see Documentation/vm/slub.rst.
5056 slab_max_order= [MM, SLAB]
5057 Determines the maximum allowed order for slabs.
5058 A high setting may cause OOMs due to memory
5059 fragmentation. Defaults to 1 for systems with
5060 more than 32MB of RAM, 0 otherwise.
5062 slub_debug[=options[,slabs][;[options[,slabs]]...] [MM, SLUB]
5063 Enabling slub_debug allows one to determine the
5064 culprit if slab objects become corrupted. Enabling
5065 slub_debug can create guard zones around objects and
5066 may poison objects when not in use. Also tracks the
5067 last alloc / free. For more information see
5068 Documentation/vm/slub.rst.
5070 slub_max_order= [MM, SLUB]
5071 Determines the maximum allowed order for slabs.
5072 A high setting may cause OOMs due to memory
5073 fragmentation. For more information see
5074 Documentation/vm/slub.rst.
5076 slub_min_objects= [MM, SLUB]
5077 The minimum number of objects per slab. SLUB will
5078 increase the slab order up to slub_max_order to
5079 generate a sufficiently large slab able to contain
5080 the number of objects indicated. The higher the number
5081 of objects the smaller the overhead of tracking slabs
5082 and the less frequently locks need to be acquired.
5083 For more information see Documentation/vm/slub.rst.
5085 slub_min_order= [MM, SLUB]
5086 Determines the minimum page order for slabs. Must be
5087 lower than slub_max_order.
5088 For more information see Documentation/vm/slub.rst.
5090 slub_merge [MM, SLUB]
5091 Same with slab_merge.
5093 slub_nomerge [MM, SLUB]
5094 Same with slab_nomerge. This is supported for legacy.
5095 See slab_nomerge for more information.
5098 Format: <io1>[,<io2>[,...,<io8>]]
5100 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
5101 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
5102 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
5103 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
5104 smsc-ircc2.ircc_irq= [HW] IRQ line
5105 smsc-ircc2.ircc_dma= [HW] DMA channel
5106 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
5107 0: Toshiba Satellite 1800 (GP data pin select)
5108 1: Fast pin select (default)
5111 smt [KNL,S390] Set the maximum number of threads (logical
5112 CPUs) to use per physical CPU on systems capable of
5113 symmetric multithreading (SMT). Will be capped to the
5114 actual hardware limit.
5116 Default: -1 (no limit)
5119 [KNL] Should the soft-lockup detector generate panics.
5122 A value of 1 instructs the soft-lockup detector
5123 to panic the machine when a soft-lockup occurs. It is
5124 also controlled by the kernel.softlockup_panic sysctl
5125 and CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC, which is the
5126 respective build-time switch to that functionality.
5128 softlockup_all_cpu_backtrace=
5129 [KNL] Should the soft-lockup detector generate
5130 backtraces on all cpus.
5133 sonypi.*= [HW] Sony Programmable I/O Control Device driver
5134 See Documentation/admin-guide/laptops/sonypi.rst
5136 spectre_v2= [X86] Control mitigation of Spectre variant 2
5137 (indirect branch speculation) vulnerability.
5138 The default operation protects the kernel from
5141 on - unconditionally enable, implies
5143 off - unconditionally disable, implies
5145 auto - kernel detects whether your CPU model is
5148 Selecting 'on' will, and 'auto' may, choose a
5149 mitigation method at run time according to the
5150 CPU, the available microcode, the setting of the
5151 CONFIG_RETPOLINE configuration option, and the
5152 compiler with which the kernel was built.
5154 Selecting 'on' will also enable the mitigation
5155 against user space to user space task attacks.
5157 Selecting 'off' will disable both the kernel and
5158 the user space protections.
5160 Specific mitigations can also be selected manually:
5162 retpoline - replace indirect branches
5163 retpoline,generic - google's original retpoline
5164 retpoline,amd - AMD-specific minimal thunk
5166 Not specifying this option is equivalent to
5170 [X86] Control mitigation of Spectre variant 2
5171 (indirect branch speculation) vulnerability between
5174 on - Unconditionally enable mitigations. Is
5175 enforced by spectre_v2=on
5177 off - Unconditionally disable mitigations. Is
5178 enforced by spectre_v2=off
5180 prctl - Indirect branch speculation is enabled,
5181 but mitigation can be enabled via prctl
5182 per thread. The mitigation control state
5183 is inherited on fork.
5186 - Like "prctl" above, but only STIBP is
5187 controlled per thread. IBPB is issued
5188 always when switching between different user
5192 - Same as "prctl" above, but all seccomp
5193 threads will enable the mitigation unless
5194 they explicitly opt out.
5197 - Like "seccomp" above, but only STIBP is
5198 controlled per thread. IBPB is issued
5199 always when switching between different
5200 user space processes.
5202 auto - Kernel selects the mitigation depending on
5203 the available CPU features and vulnerability.
5206 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
5208 Not specifying this option is equivalent to
5209 spectre_v2_user=auto.
5211 spec_store_bypass_disable=
5212 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
5213 (Speculative Store Bypass vulnerability)
5215 Certain CPUs are vulnerable to an exploit against a
5216 a common industry wide performance optimization known
5217 as "Speculative Store Bypass" in which recent stores
5218 to the same memory location may not be observed by
5219 later loads during speculative execution. The idea
5220 is that such stores are unlikely and that they can
5221 be detected prior to instruction retirement at the
5222 end of a particular speculation execution window.
5224 In vulnerable processors, the speculatively forwarded
5225 store can be used in a cache side channel attack, for
5226 example to read memory to which the attacker does not
5227 directly have access (e.g. inside sandboxed code).
5229 This parameter controls whether the Speculative Store
5230 Bypass optimization is used.
5232 On x86 the options are:
5234 on - Unconditionally disable Speculative Store Bypass
5235 off - Unconditionally enable Speculative Store Bypass
5236 auto - Kernel detects whether the CPU model contains an
5237 implementation of Speculative Store Bypass and
5238 picks the most appropriate mitigation. If the
5239 CPU is not vulnerable, "off" is selected. If the
5240 CPU is vulnerable the default mitigation is
5241 architecture and Kconfig dependent. See below.
5242 prctl - Control Speculative Store Bypass per thread
5243 via prctl. Speculative Store Bypass is enabled
5244 for a process by default. The state of the control
5245 is inherited on fork.
5246 seccomp - Same as "prctl" above, but all seccomp threads
5247 will disable SSB unless they explicitly opt out.
5249 Default mitigations:
5250 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
5252 On powerpc the options are:
5254 on,auto - On Power8 and Power9 insert a store-forwarding
5255 barrier on kernel entry and exit. On Power7
5256 perform a software flush on kernel entry and
5260 Not specifying this option is equivalent to
5261 spec_store_bypass_disable=auto.
5263 spia_io_base= [HW,MTD]
5269 [X86] Enable split lock detection or bus lock detection
5271 When enabled (and if hardware support is present), atomic
5272 instructions that access data across cache line
5273 boundaries will result in an alignment check exception
5274 for split lock detection or a debug exception for
5279 warn - the kernel will emit rate-limited warnings
5280 about applications triggering the #AC
5281 exception or the #DB exception. This mode is
5282 the default on CPUs that support split lock
5283 detection or bus lock detection. Default
5284 behavior is by #AC if both features are
5285 enabled in hardware.
5287 fatal - the kernel will send SIGBUS to applications
5288 that trigger the #AC exception or the #DB
5289 exception. Default behavior is by #AC if
5290 both features are enabled in hardware.
5292 If an #AC exception is hit in the kernel or in
5293 firmware (i.e. not while executing in user mode)
5294 the kernel will oops in either "warn" or "fatal"
5297 #DB exception for bus lock is triggered only when
5301 Control the Special Register Buffer Data Sampling
5304 Certain CPUs are vulnerable to an MDS-like
5305 exploit which can leak bits from the random
5308 By default, this issue is mitigated by
5309 microcode. However, the microcode fix can cause
5310 the RDRAND and RDSEED instructions to become
5311 much slower. Among other effects, this will
5312 result in reduced throughput from /dev/urandom.
5314 The microcode mitigation can be disabled with
5315 the following option:
5317 off: Disable mitigation and remove
5318 performance impact to RDRAND and RDSEED
5320 srcutree.counter_wrap_check [KNL]
5321 Specifies how frequently to check for
5322 grace-period sequence counter wrap for the
5323 srcu_data structure's ->srcu_gp_seq_needed field.
5324 The greater the number of bits set in this kernel
5325 parameter, the less frequently counter wrap will
5326 be checked for. Note that the bottom two bits
5329 srcutree.exp_holdoff [KNL]
5330 Specifies how many nanoseconds must elapse
5331 since the end of the last SRCU grace period for
5332 a given srcu_struct until the next normal SRCU
5333 grace period will be considered for automatic
5334 expediting. Set to zero to disable automatic
5338 Speculative Store Bypass Disable control
5340 On CPUs that are vulnerable to the Speculative
5341 Store Bypass vulnerability and offer a
5342 firmware based mitigation, this parameter
5343 indicates how the mitigation should be used:
5345 force-on: Unconditionally enable mitigation for
5346 for both kernel and userspace
5347 force-off: Unconditionally disable mitigation for
5348 for both kernel and userspace
5349 kernel: Always enable mitigation in the
5350 kernel, and offer a prctl interface
5351 to allow userspace to register its
5352 interest in being mitigated too.
5354 stack_guard_gap= [MM]
5355 override the default stack gap protection. The value
5356 is in page units and it defines how many pages prior
5357 to (for stacks growing down) resp. after (for stacks
5358 growing up) the main stack are reserved for no other
5359 mapping. Default value is 256 pages.
5361 stack_depot_disable= [KNL]
5362 Setting this to true through kernel command line will
5363 disable the stack depot thereby saving the static memory
5364 consumed by the stack hash table. By default this is set
5368 Enabled the stack tracer on boot up.
5370 stacktrace_filter=[function-list]
5371 [FTRACE] Limit the functions that the stack tracer
5372 will trace at boot up. function-list is a comma-separated
5373 list of functions. This list can be changed at run
5374 time by the stack_trace_filter file in the debugfs
5375 tracing directory. Note, this enables stack tracing
5376 and the stacktrace above is not needed.
5380 Set the STI (builtin display/keyboard on the HP-PARISC
5381 machines) console (graphic card) which should be used
5382 as the initial boot-console.
5383 See also comment in drivers/video/console/sticore.c.
5386 See comment in drivers/video/console/sticore.c.
5389 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
5391 sunrpc.min_resvport=
5392 sunrpc.max_resvport=
5394 SunRPC servers often require that client requests
5395 originate from a privileged port (i.e. a port in the
5396 range 0 < portnr < 1024).
5397 An administrator who wishes to reserve some of these
5398 ports for other uses may adjust the range that the
5399 kernel's sunrpc client considers to be privileged
5400 using these two parameters to set the minimum and
5401 maximum port values.
5403 sunrpc.svc_rpc_per_connection_limit=
5405 Limit the number of requests that the server will
5406 process in parallel from a single connection.
5407 The default value is 0 (no limit).
5411 Control how the NFS server code allocates CPUs to
5412 service thread pools. Depending on how many NICs
5413 you have and where their interrupts are bound, this
5414 option will affect which CPUs will do NFS serving.
5415 Note: this parameter cannot be changed while the
5416 NFS server is running.
5418 auto the server chooses an appropriate mode
5419 automatically using heuristics
5420 global a single global pool contains all CPUs
5421 percpu one pool for each CPU
5422 pernode one pool for each NUMA node (equivalent
5423 to global on non-NUMA machines)
5425 sunrpc.tcp_slot_table_entries=
5426 sunrpc.udp_slot_table_entries=
5428 Sets the upper limit on the number of simultaneous
5429 RPC calls that can be sent from the client to a
5430 server. Increasing these values may allow you to
5431 improve throughput, but will also increase the
5432 amount of memory reserved for use by the client.
5434 suspend.pm_test_delay=
5436 Sets the number of seconds to remain in a suspend test
5437 mode before resuming the system (see
5438 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
5439 is set. Default value is 5.
5442 Format: { on | off | y | n | 1 | 0 }
5443 This parameter controls use of the Protected
5444 Execution Facility on pSeries.
5447 [KNL] Enable accounting of swap in memory resource
5448 controller if no parameter or 1 is given or disable
5449 it if 0 is given (See Documentation/admin-guide/cgroup-v1/memory.rst)
5451 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
5452 Format: { <int> | force | noforce }
5453 <int> -- Number of I/O TLB slabs
5454 force -- force using of bounce buffers even if they
5455 wouldn't be automatically used by the kernel
5456 noforce -- Never use bounce buffers (for debugging)
5461 Set a sysctl parameter, right before loading the init
5462 process, as if the value was written to the respective
5463 /proc/sys/... file. Both '.' and '/' are recognized as
5464 separators. Unrecognized parameters and invalid values
5465 are reported in the kernel log. Sysctls registered
5466 later by a loaded module cannot be set this way.
5467 Example: sysctl.vm.swappiness=40
5469 sysfs.deprecated=0|1 [KNL]
5470 Enable/disable old style sysfs layout for old udev
5471 on older distributions. When this option is enabled
5472 very new udev will not work anymore. When this option
5473 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
5474 in older udev will not work anymore.
5475 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
5476 the kernel configuration.
5478 sysrq_always_enabled
5480 Ignore sysrq setting - this boot parameter will
5481 neutralize any effect of /proc/sys/kernel/sysrq.
5482 Useful for debugging.
5484 tcpmhash_entries= [KNL,NET]
5485 Set the number of tcp_metrics_hash slots.
5486 Default value is 8192 or 16384 depending on total
5487 ram pages. This is used to specify the TCP metrics
5488 cache size. See Documentation/networking/ip-sysctl.rst
5489 "tcp_no_metrics_save" section for more details.
5493 test_suspend= [SUSPEND][,N]
5494 Specify "mem" (for Suspend-to-RAM) or "standby" (for
5495 standby suspend) or "freeze" (for suspend type freeze)
5496 as the system sleep state during system startup with
5497 the optional capability to repeat N number of times.
5498 The system is woken from this state using a
5499 wakeup-capable RTC alarm.
5501 thash_entries= [KNL,NET]
5502 Set number of hash buckets for TCP connection
5504 thermal.act= [HW,ACPI]
5505 -1: disable all active trip points in all thermal zones
5506 <degrees C>: override all lowest active trip points
5508 thermal.crt= [HW,ACPI]
5509 -1: disable all critical trip points in all thermal zones
5510 <degrees C>: override all critical trip points
5512 thermal.nocrt= [HW,ACPI]
5513 Set to disable actions on ACPI thermal zone
5514 critical and hot trip points.
5516 thermal.off= [HW,ACPI]
5517 1: disable ACPI thermal control
5519 thermal.psv= [HW,ACPI]
5520 -1: disable all passive trip points
5521 <degrees C>: override all passive trip points to this
5524 thermal.tzp= [HW,ACPI]
5525 Specify global default ACPI thermal zone polling rate
5526 <deci-seconds>: poll all this frequency
5527 0: no polling (default)
5530 Force threading of all interrupt handlers except those
5531 marked explicitly IRQF_NO_THREAD.
5535 Specify if the kernel should make use of the cpu
5536 topology information if the hardware supports this.
5537 The scheduler will make use of this information and
5538 e.g. base its process migration decisions on it.
5541 topology_updates= [KNL, PPC, NUMA]
5543 Specify if the kernel should ignore (off)
5544 topology updates sent by the hypervisor to this
5547 torture.disable_onoff_at_boot= [KNL]
5548 Prevent the CPU-hotplug component of torturing
5549 until after init has spawned.
5551 torture.ftrace_dump_at_shutdown= [KNL]
5552 Dump the ftrace buffer at torture-test shutdown,
5553 even if there were no errors. This can be a
5554 very costly operation when many torture tests
5555 are running concurrently, especially on systems
5556 with rotating-rust storage.
5558 torture.verbose_sleep_frequency= [KNL]
5559 Specifies how many verbose printk()s should be
5560 emitted between each sleep. The default of zero
5561 disables verbose-printk() sleeping.
5563 torture.verbose_sleep_duration= [KNL]
5564 Duration of each verbose-printk() sleep in jiffies.
5568 tpm_suspend_pcr=[HW,TPM]
5569 Format: integer pcr id
5570 Specify that at suspend time, the tpm driver
5571 should extend the specified pcr with zeros,
5572 as a workaround for some chips which fail to
5573 flush the last written pcr on TPM_SaveState.
5574 This will guarantee that all the other pcrs
5577 trace_buf_size=nn[KMG]
5578 [FTRACE] will set tracing buffer size on each cpu.
5580 trace_event=[event-list]
5581 [FTRACE] Set and start specified trace events in order
5582 to facilitate early boot debugging. The event-list is a
5583 comma-separated list of trace events to enable. See
5584 also Documentation/trace/events.rst
5586 trace_options=[option-list]
5587 [FTRACE] Enable or disable tracer options at boot.
5588 The option-list is a comma delimited list of options
5589 that can be enabled or disabled just as if you were
5590 to echo the option name into
5592 /sys/kernel/debug/tracing/trace_options
5594 For example, to enable stacktrace option (to dump the
5595 stack trace of each event), add to the command line:
5597 trace_options=stacktrace
5599 See also Documentation/trace/ftrace.rst "trace options"
5603 Have the tracepoints sent to printk as well as the
5604 tracing ring buffer. This is useful for early boot up
5605 where the system hangs or reboots and does not give the
5606 option for reading the tracing buffer or performing a
5607 ftrace_dump_on_oops.
5609 To turn off having tracepoints sent to printk,
5610 echo 0 > /proc/sys/kernel/tracepoint_printk
5611 Note, echoing 1 into this file without the
5612 tracepoint_printk kernel cmdline option has no effect.
5616 Having tracepoints sent to printk() and activating high
5617 frequency tracepoints such as irq or sched, can cause
5618 the system to live lock.
5621 [FTRACE] enable this option to disable tracing when a
5622 warning is hit. This turns off "tracing_on". Tracing can
5623 be enabled again by echoing '1' into the "tracing_on"
5624 file located in /sys/kernel/debug/tracing/
5626 This option is useful, as it disables the trace before
5627 the WARNING dump is called, which prevents the trace to
5628 be filled with content caused by the warning output.
5630 This option can also be set at run time via the sysctl
5631 option: kernel/traceoff_on_warning
5633 transparent_hugepage=
5635 Format: [always|madvise|never]
5636 Can be used to control the default behavior of the system
5637 with respect to transparent hugepages.
5638 See Documentation/admin-guide/mm/transhuge.rst
5641 trusted.source= [KEYS]
5643 This parameter identifies the trust source as a backend
5644 for trusted keys implementation. Supported trust
5648 If not specified then it defaults to iterating through
5649 the trust source list starting with TPM and assigns the
5650 first trust source as a backend which is initialized
5651 successfully during iteration.
5653 tsc= Disable clocksource stability checks for TSC.
5655 [x86] reliable: mark tsc clocksource as reliable, this
5656 disables clocksource verification at runtime, as well
5657 as the stability checks done at bootup. Used to enable
5658 high-resolution timer mode on older hardware, and in
5659 virtualized environment.
5660 [x86] noirqtime: Do not use TSC to do irq accounting.
5661 Used to run time disable IRQ_TIME_ACCOUNTING on any
5662 platforms where RDTSC is slow and this accounting
5664 [x86] unstable: mark the TSC clocksource as unstable, this
5665 marks the TSC unconditionally unstable at bootup and
5666 avoids any further wobbles once the TSC watchdog notices.
5667 [x86] nowatchdog: disable clocksource watchdog. Used
5668 in situations with strict latency requirements (where
5669 interruptions from clocksource watchdog are not
5672 tsc_early_khz= [X86] Skip early TSC calibration and use the given
5673 value instead. Useful when the early TSC frequency discovery
5674 procedure is not reliable, such as on overclocked systems
5675 with CPUID.16h support and partial CPUID.15h support.
5676 Format: <unsigned int>
5678 tsx= [X86] Control Transactional Synchronization
5679 Extensions (TSX) feature in Intel processors that
5680 support TSX control.
5682 This parameter controls the TSX feature. The options are:
5684 on - Enable TSX on the system. Although there are
5685 mitigations for all known security vulnerabilities,
5686 TSX has been known to be an accelerator for
5687 several previous speculation-related CVEs, and
5688 so there may be unknown security risks associated
5689 with leaving it enabled.
5691 off - Disable TSX on the system. (Note that this
5692 option takes effect only on newer CPUs which are
5693 not vulnerable to MDS, i.e., have
5694 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
5695 the new IA32_TSX_CTRL MSR through a microcode
5696 update. This new MSR allows for the reliable
5697 deactivation of the TSX functionality.)
5699 auto - Disable TSX if X86_BUG_TAA is present,
5700 otherwise enable TSX on the system.
5702 Not specifying this option is equivalent to tsx=off.
5704 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5707 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
5708 Abort (TAA) vulnerability.
5710 Similar to Micro-architectural Data Sampling (MDS)
5711 certain CPUs that support Transactional
5712 Synchronization Extensions (TSX) are vulnerable to an
5713 exploit against CPU internal buffers which can forward
5714 information to a disclosure gadget under certain
5717 In vulnerable processors, the speculatively forwarded
5718 data can be used in a cache side channel attack, to
5719 access data to which the attacker does not have direct
5722 This parameter controls the TAA mitigation. The
5725 full - Enable TAA mitigation on vulnerable CPUs
5728 full,nosmt - Enable TAA mitigation and disable SMT on
5729 vulnerable CPUs. If TSX is disabled, SMT
5730 is not disabled because CPU is not
5731 vulnerable to cross-thread TAA attacks.
5732 off - Unconditionally disable TAA mitigation
5734 On MDS-affected machines, tsx_async_abort=off can be
5735 prevented by an active MDS mitigation as both vulnerabilities
5736 are mitigated with the same mechanism so in order to disable
5737 this mitigation, you need to specify mds=off too.
5739 Not specifying this option is equivalent to
5740 tsx_async_abort=full. On CPUs which are MDS affected
5741 and deploy MDS mitigation, TAA mitigation is not
5742 required and doesn't provide any additional
5746 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
5748 turbografx.map[2|3]= [HW,JOY]
5749 TurboGraFX parallel port interface
5751 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
5752 See also Documentation/input/devices/joystick-parport.rst
5754 udbg-immortal [PPC] When debugging early kernel crashes that
5755 happen after console_init() and before a proper
5756 console driver takes over, this boot options might
5757 help "seeing" what's going on.
5759 uhash_entries= [KNL,NET]
5760 Set number of hash buckets for UDP/UDP-Lite connections
5763 [USB] Ignore overcurrent events (default N).
5764 Some badly-designed motherboards generate lots of
5765 bogus events, for ports that aren't wired to
5766 anything. Set this parameter to avoid log spamming.
5767 Note that genuine overcurrent events won't be
5771 [X86] Cause panic on unknown NMI.
5773 usbcore.authorized_default=
5774 [USB] Default USB device authorization:
5775 (default -1 = authorized except for wireless USB,
5776 0 = not authorized, 1 = authorized, 2 = authorized
5777 if device connected to internal port)
5779 usbcore.autosuspend=
5780 [USB] The autosuspend time delay (in seconds) used
5781 for newly-detected USB devices (default 2). This
5782 is the time required before an idle device will be
5783 autosuspended. Devices for which the delay is set
5784 to a negative value won't be autosuspended at all.
5786 usbcore.usbfs_snoop=
5787 [USB] Set to log all usbfs traffic (default 0 = off).
5789 usbcore.usbfs_snoop_max=
5790 [USB] Maximum number of bytes to snoop in each URB
5793 usbcore.blinkenlights=
5794 [USB] Set to cycle leds on hubs (default 0 = off).
5796 usbcore.old_scheme_first=
5797 [USB] Start with the old device initialization
5798 scheme (default 0 = off).
5800 usbcore.usbfs_memory_mb=
5801 [USB] Memory limit (in MB) for buffers allocated by
5802 usbfs (default = 16, 0 = max = 2047).
5804 usbcore.use_both_schemes=
5805 [USB] Try the other device initialization scheme
5806 if the first one fails (default 1 = enabled).
5808 usbcore.initial_descriptor_timeout=
5809 [USB] Specifies timeout for the initial 64-byte
5810 USB_REQ_GET_DESCRIPTOR request in milliseconds
5811 (default 5000 = 5.0 seconds).
5813 usbcore.nousb [USB] Disable the USB subsystem
5816 [USB] A list of quirk entries to augment the built-in
5817 usb core quirk list. List entries are separated by
5818 commas. Each entry has the form
5819 VendorID:ProductID:Flags. The IDs are 4-digit hex
5820 numbers and Flags is a set of letters. Each letter
5821 will change the built-in quirk; setting it if it is
5822 clear and clearing it if it is set. The letters have
5823 the following meanings:
5824 a = USB_QUIRK_STRING_FETCH_255 (string
5825 descriptors must not be fetched using
5827 b = USB_QUIRK_RESET_RESUME (device can't resume
5828 correctly so reset it instead);
5829 c = USB_QUIRK_NO_SET_INTF (device can't handle
5830 Set-Interface requests);
5831 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
5832 handle its Configuration or Interface
5834 e = USB_QUIRK_RESET (device can't be reset
5835 (e.g morph devices), don't use reset);
5836 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
5837 more interface descriptions than the
5838 bNumInterfaces count, and can't handle
5839 talking to these interfaces);
5840 g = USB_QUIRK_DELAY_INIT (device needs a pause
5841 during initialization, after we read
5842 the device descriptor);
5843 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
5844 high speed and super speed interrupt
5845 endpoints, the USB 2.0 and USB 3.0 spec
5846 require the interval in microframes (1
5847 microframe = 125 microseconds) to be
5848 calculated as interval = 2 ^
5850 Devices with this quirk report their
5851 bInterval as the result of this
5852 calculation instead of the exponent
5853 variable used in the calculation);
5854 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
5855 handle device_qualifier descriptor
5857 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
5858 generates spurious wakeup, ignore
5859 remote wakeup capability);
5860 k = USB_QUIRK_NO_LPM (device can't handle Link
5862 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
5863 (Device reports its bInterval as linear
5864 frames instead of the USB 2.0
5866 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
5867 to be disconnected before suspend to
5868 prevent spurious wakeup);
5869 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
5870 pause after every control message);
5871 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
5872 delay after resetting its port);
5873 Example: quirks=0781:5580:bk,0a5c:5834:gij
5876 [USBHID] The interval which mice are to be polled at.
5879 [USBHID] The interval which joysticks are to be polled at.
5882 [USBHID] The interval which keyboards are to be polled at.
5884 usb-storage.delay_use=
5885 [UMS] The delay in seconds before a new device is
5886 scanned for Logical Units (default 1).
5889 [UMS] A list of quirks entries to supplement or
5890 override the built-in unusual_devs list. List
5891 entries are separated by commas. Each entry has
5892 the form VID:PID:Flags where VID and PID are Vendor
5893 and Product ID values (4-digit hex numbers) and
5894 Flags is a set of characters, each corresponding
5895 to a common usb-storage quirk flag as follows:
5896 a = SANE_SENSE (collect more than 18 bytes
5897 of sense data, not on uas);
5898 b = BAD_SENSE (don't collect more than 18
5899 bytes of sense data, not on uas);
5900 c = FIX_CAPACITY (decrease the reported
5901 device capacity by one sector);
5902 d = NO_READ_DISC_INFO (don't use
5903 READ_DISC_INFO command, not on uas);
5904 e = NO_READ_CAPACITY_16 (don't use
5905 READ_CAPACITY_16 command);
5906 f = NO_REPORT_OPCODES (don't use report opcodes
5908 g = MAX_SECTORS_240 (don't transfer more than
5909 240 sectors at a time, uas only);
5910 h = CAPACITY_HEURISTICS (decrease the
5911 reported device capacity by one
5912 sector if the number is odd);
5913 i = IGNORE_DEVICE (don't bind to this
5915 j = NO_REPORT_LUNS (don't use report luns
5917 k = NO_SAME (do not use WRITE_SAME, uas only)
5918 l = NOT_LOCKABLE (don't try to lock and
5919 unlock ejectable media, not on uas);
5920 m = MAX_SECTORS_64 (don't transfer more
5921 than 64 sectors = 32 KB at a time,
5923 n = INITIAL_READ10 (force a retry of the
5924 initial READ(10) command, not on uas);
5925 o = CAPACITY_OK (accept the capacity
5926 reported by the device, not on uas);
5927 p = WRITE_CACHE (the device cache is ON
5928 by default, not on uas);
5929 r = IGNORE_RESIDUE (the device reports
5930 bogus residue values, not on uas);
5931 s = SINGLE_LUN (the device has only one
5933 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5934 commands, uas only);
5935 u = IGNORE_UAS (don't bind to the uas driver);
5936 w = NO_WP_DETECT (don't test whether the
5937 medium is write-protected).
5938 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5939 even if the device claims no cache,
5941 Example: quirks=0419:aaf5:rl,0421:0433:rc
5943 user_debug= [KNL,ARM]
5945 See arch/arm/Kconfig.debug help text.
5946 1 - undefined instruction events
5948 4 - invalid data aborts
5951 Example: user_debug=31
5954 [X86] Flags controlling user PTE allocations.
5956 nohigh = do not allocate PTE pages in
5957 HIGHMEM regardless of setting
5961 On X86_32, this is an alias for vdso32=. Otherwise:
5963 vdso=1: enable VDSO (the default)
5964 vdso=0: disable VDSO mapping
5966 vdso32= [X86] Control the 32-bit vDSO
5967 vdso32=1: enable 32-bit VDSO
5968 vdso32=0 or vdso32=2: disable 32-bit VDSO
5970 See the help text for CONFIG_COMPAT_VDSO for more
5971 details. If CONFIG_COMPAT_VDSO is set, the default is
5972 vdso32=0; otherwise, the default is vdso32=1.
5974 For compatibility with older kernels, vdso32=2 is an
5977 Try vdso32=0 if you encounter an error that says:
5978 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5981 vector=percpu: enable percpu vector domain
5983 video= [FB] Frame buffer configuration
5984 See Documentation/fb/modedb.rst.
5986 video.brightness_switch_enabled= [0,1]
5987 If set to 1, on receiving an ACPI notify event
5988 generated by hotkey, video driver will adjust brightness
5989 level and then send out the event to user space through
5990 the allocated input device; If set to 0, video driver
5991 will only send out the event without touching backlight
5996 [VMMIO] Memory mapped virtio (platform) device.
5998 <size>@<baseaddr>:<irq>[:<id>]
6000 <size> := size (can use standard suffixes
6002 <baseaddr> := physical base address
6003 <irq> := interrupt number (as passed to
6005 <id> := (optional) platform device id
6007 virtio_mmio.device=1K@0x100b0000:48:7
6009 Can be used multiple times for multiple devices.
6011 vga= [BOOT,X86-32] Select a particular video mode
6012 See Documentation/x86/boot.rst and
6013 Documentation/admin-guide/svga.rst.
6014 Use vga=ask for menu.
6015 This is actually a boot loader parameter; the value is
6016 passed to the kernel using a special protocol.
6018 vm_debug[=options] [KNL] Available with CONFIG_DEBUG_VM=y.
6019 May slow down system boot speed, especially when
6020 enabled on systems with a large amount of memory.
6021 All options are enabled by default, and this
6022 interface is meant to allow for selectively
6023 enabling or disabling specific virtual memory
6026 Available options are:
6027 P Enable page structure init time poisoning
6028 - Disable all of the above options
6030 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
6031 size of <nn>. This can be used to increase the
6032 minimum size (128MB on x86). It can also be used to
6033 decrease the size and leave more room for directly
6036 vmcp_cma=nn[MG] [KNL,S390]
6037 Sets the memory size reserved for contiguous memory
6038 allocations for the vmcp device driver.
6040 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
6043 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
6046 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
6050 Controls the behavior of vsyscalls (i.e. calls to
6051 fixed addresses of 0xffffffffff600x00 from legacy
6052 code). Most statically-linked binaries and older
6053 versions of glibc use these calls. Because these
6054 functions are at fixed addresses, they make nice
6055 targets for exploits that can control RIP.
6057 emulate [default] Vsyscalls turn into traps and are
6058 emulated reasonably safely. The vsyscall
6061 xonly Vsyscalls turn into traps and are
6062 emulated reasonably safely. The vsyscall
6063 page is not readable.
6065 none Vsyscalls don't work at all. This makes
6066 them quite hard to use for exploits but
6067 might break your system.
6069 vt.color= [VT] Default text color.
6070 Format: 0xYX, X = foreground, Y = background.
6071 Default: 0x07 = light gray on black.
6073 vt.cur_default= [VT] Default cursor shape.
6074 Format: 0xCCBBAA, where AA, BB, and CC are the same as
6075 the parameters of the <Esc>[?A;B;Cc escape sequence;
6076 see VGA-softcursor.txt. Default: 2 = underline.
6078 vt.default_blu= [VT]
6079 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
6080 Change the default blue palette of the console.
6081 This is a 16-member array composed of values
6084 vt.default_grn= [VT]
6085 Format: <green0>,<green1>,<green2>,...,<green15>
6086 Change the default green palette of the console.
6087 This is a 16-member array composed of values
6090 vt.default_red= [VT]
6091 Format: <red0>,<red1>,<red2>,...,<red15>
6092 Change the default red palette of the console.
6093 This is a 16-member array composed of values
6099 Set system-wide default UTF-8 mode for all tty's.
6100 Default is 1, i.e. UTF-8 mode is enabled for all
6101 newly opened terminals.
6103 vt.global_cursor_default=
6106 Set system-wide default for whether a cursor
6107 is shown on new VTs. Default is -1,
6108 i.e. cursors will be created by default unless
6109 overridden by individual drivers. 0 will hide
6110 cursors, 1 will display them.
6112 vt.italic= [VT] Default color for italic text; 0-15.
6115 vt.underline= [VT] Default color for underlined text; 0-15.
6118 watchdog timers [HW,WDT] For information on watchdog timers,
6119 see Documentation/watchdog/watchdog-parameters.rst
6120 or other driver-specific files in the
6121 Documentation/watchdog/ directory.
6125 Set the hard lockup detector stall duration
6126 threshold in seconds. The soft lockup detector
6127 threshold is set to twice the value. A value of 0
6128 disables both lockup detectors. Default is 10
6131 workqueue.watchdog_thresh=
6132 If CONFIG_WQ_WATCHDOG is configured, workqueue can
6133 warn stall conditions and dump internal state to
6134 help debugging. 0 disables workqueue stall
6135 detection; otherwise, it's the stall threshold
6136 duration in seconds. The default value is 30 and
6137 it can be updated at runtime by writing to the
6138 corresponding sysfs file.
6140 workqueue.disable_numa
6141 By default, all work items queued to unbound
6142 workqueues are affine to the NUMA nodes they're
6143 issued on, which results in better behavior in
6144 general. If NUMA affinity needs to be disabled for
6145 whatever reason, this option can be used. Note
6146 that this also can be controlled per-workqueue for
6147 workqueues visible under /sys/bus/workqueue/.
6149 workqueue.power_efficient
6150 Per-cpu workqueues are generally preferred because
6151 they show better performance thanks to cache
6152 locality; unfortunately, per-cpu workqueues tend to
6153 be more power hungry than unbound workqueues.
6155 Enabling this makes the per-cpu workqueues which
6156 were observed to contribute significantly to power
6157 consumption unbound, leading to measurably lower
6158 power usage at the cost of small performance
6161 The default value of this parameter is determined by
6162 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
6164 workqueue.debug_force_rr_cpu
6165 Workqueue used to implicitly guarantee that work
6166 items queued without explicit CPU specified are put
6167 on the local CPU. This guarantee is no longer true
6168 and while local CPU is still preferred work items
6169 may be put on foreign CPUs. This debug option
6170 forces round-robin CPU selection to flush out
6171 usages which depend on the now broken guarantee.
6172 When enabled, memory and cache locality will be
6175 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
6176 default x2apic cluster mode on platforms
6179 xen_512gb_limit [KNL,X86-64,XEN]
6180 Restricts the kernel running paravirtualized under Xen
6181 to use only up to 512 GB of RAM. The reason to do so is
6182 crash analysis tools and Xen tools for doing domain
6183 save/restore/migration must be enabled to handle larger
6186 xen_emul_unplug= [HW,X86,XEN]
6187 Unplug Xen emulated devices
6188 Format: [unplug0,][unplug1]
6189 ide-disks -- unplug primary master IDE devices
6190 aux-ide-disks -- unplug non-primary-master IDE devices
6191 nics -- unplug network devices
6192 all -- unplug all emulated devices (NICs and IDE disks)
6193 unnecessary -- unplugging emulated devices is
6194 unnecessary even if the host did not respond to
6196 never -- do not unplug even if version check succeeds
6198 xen_legacy_crash [X86,XEN]
6199 Crash from Xen panic notifier, without executing late
6200 panic() code such as dumping handler.
6202 xen_nopvspin [X86,XEN]
6203 Disables the qspinlock slowpath using Xen PV optimizations.
6204 This parameter is obsoleted by "nopvspin" parameter, which
6205 has equivalent effect for XEN platform.
6208 Disables the PV optimizations forcing the HVM guest to
6209 run as generic HVM guest with no PV drivers.
6210 This option is obsoleted by the "nopv" option, which
6211 has equivalent effect for XEN platform.
6213 xen_no_vector_callback
6214 [KNL,X86,XEN] Disable the vector callback for Xen
6215 event channel interrupts.
6217 xen_scrub_pages= [XEN]
6218 Boolean option to control scrubbing pages before giving them back
6219 to Xen, for use by other domains. Can be also changed at runtime
6220 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
6221 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
6223 xen_timer_slop= [X86-64,XEN]
6224 Set the timer slop (in nanoseconds) for the virtual Xen
6225 timers (default is 100000). This adjusts the minimum
6226 delta of virtualized Xen timers, where lower values
6227 improve timer resolution at the expense of processing
6228 more timer interrupts.
6230 xen.event_eoi_delay= [XEN]
6231 How long to delay EOI handling in case of event
6232 storms (jiffies). Default is 10.
6234 xen.event_loop_timeout= [XEN]
6235 After which time (jiffies) the event handling loop
6236 should start to delay EOI handling. Default is 2.
6238 xen.fifo_events= [XEN]
6239 Boolean parameter to disable using fifo event handling
6240 even if available. Normally fifo event handling is
6241 preferred over the 2-level event handling, as it is
6242 fairer and the number of possible event channels is
6243 much higher. Default is on (use fifo events).
6245 nopv= [X86,XEN,KVM,HYPER_V,VMWARE]
6246 Disables the PV optimizations forcing the guest to run
6247 as generic guest with no PV drivers. Currently support
6248 XEN HVM, KVM, HYPER_V and VMWARE guest.
6250 nopvspin [X86,XEN,KVM]
6251 Disables the qspinlock slow path using PV optimizations
6252 which allow the hypervisor to 'idle' the guest on lock
6255 xirc2ps_cs= [NET,PCMCIA]
6257 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
6260 By default on POWER9 and above, the kernel will
6261 natively use the XIVE interrupt controller. This option
6262 allows the fallback firmware mode to be used:
6264 off Fallback to firmware control of XIVE interrupt
6265 controller on both pseries and powernv
6266 platforms. Only useful on POWER9 and above.
6268 xhci-hcd.quirks [USB,KNL]
6269 A hex value specifying bitmask with supplemental xhci
6270 host controller quirks. Meaning of each bit can be
6271 consulted in header drivers/usb/host/xhci.h.
6274 Format: { early | on | rw | ro | off }
6275 Controls if xmon debugger is enabled. Default is off.
6276 Passing only "xmon" is equivalent to "xmon=early".
6277 early Call xmon as early as possible on boot; xmon
6278 debugger is called from setup_arch().
6279 on xmon debugger hooks will be installed so xmon
6280 is only called on a kernel crash. Default mode,
6281 i.e. either "ro" or "rw" mode, is controlled
6282 with CONFIG_XMON_DEFAULT_RO_MODE.
6283 rw xmon debugger hooks will be installed so xmon
6284 is called only on a kernel crash, mode is write,
6285 meaning SPR registers, memory and, other data
6286 can be written using xmon commands.
6287 ro same as "rw" option above but SPR registers,
6288 memory, and other data can't be written using
6290 off xmon is disabled.