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.txt, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
24 acpi_backlight= [HW,ACPI]
27 If set to vendor, prefer vendor specific driver
28 (e.g. thinkpad_acpi, sony_acpi, etc.) instead
29 of the ACPI video.ko driver.
31 acpi_force_32bit_fadt_addr
32 force FADT to use 32 bit addresses rather than the
33 64 bit X_* addresses. Some firmware have broken 64
34 bit addresses for force ACPI ignore these and use
35 the older legacy 32 bit addresses.
37 acpica_no_return_repair [HW, ACPI]
38 Disable AML predefined validation mechanism
39 This mechanism can repair the evaluation result to make
40 the return objects more ACPI specification compliant.
41 This option is useful for developers to identify the
42 root cause of an AML interpreter issue when the issue
43 has something to do with the repair mechanism.
45 acpi.debug_layer= [HW,ACPI,ACPI_DEBUG]
46 acpi.debug_level= [HW,ACPI,ACPI_DEBUG]
48 CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
49 debug output. Bits in debug_layer correspond to a
50 _COMPONENT in an ACPI source file, e.g.,
51 #define _COMPONENT ACPI_PCI_COMPONENT
52 Bits in debug_level correspond to a level in
53 ACPI_DEBUG_PRINT statements, e.g.,
54 ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
55 The debug_level mask defaults to "info". See
56 Documentation/acpi/debug.txt for more information about
57 debug layers and levels.
59 Enable processor driver info messages:
60 acpi.debug_layer=0x20000000
61 Enable PCI/PCI interrupt routing info messages:
62 acpi.debug_layer=0x400000
63 Enable AML "Debug" output, i.e., stores to the Debug
64 object while interpreting AML:
65 acpi.debug_layer=0xffffffff acpi.debug_level=0x2
66 Enable all messages related to ACPI hardware:
67 acpi.debug_layer=0x2 acpi.debug_level=0xffffffff
69 Some values produce so much output that the system is
70 unusable. The "log_buf_len" parameter may be useful
71 if you need to capture more output.
73 acpi_enforce_resources= [ACPI]
75 Check for resource conflicts between native drivers
76 and ACPI OperationRegions (SystemIO and SystemMemory
77 only). IO ports and memory declared in ACPI might be
78 used by the ACPI subsystem in arbitrary AML code and
79 can interfere with legacy drivers.
80 strict (default): access to resources claimed by ACPI
81 is denied; legacy drivers trying to access reserved
82 resources will fail to bind to device using them.
83 lax: access to resources claimed by ACPI is allowed;
84 legacy drivers trying to access reserved resources
85 will bind successfully but a warning message is logged.
86 no: ACPI OperationRegions are not marked as reserved,
87 no further checks are performed.
89 acpi_force_table_verification [HW,ACPI]
90 Enable table checksum verification during early stage.
91 By default, this is disabled due to x86 early mapping
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
102 acpi_irq_isa= [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
103 Format: <irq>,<irq>...
105 acpi_irq_pci= [HW,ACPI] If irq_balance, clear listed IRQs for
107 Format: <irq>,<irq>...
109 acpi_mask_gpe= [HW,ACPI]
110 Due to the existence of _Lxx/_Exx, some GPEs triggered
111 by unsupported hardware/firmware features can result in
112 GPE floodings that cannot be automatically disabled by
114 This facility can be used to prevent such uncontrolled
118 acpi_no_auto_serialize [HW,ACPI]
119 Disable auto-serialization of AML methods
120 AML control methods that contain the opcodes to create
121 named objects will be marked as "Serialized" by the
122 auto-serialization feature.
123 This feature is enabled by default.
124 This option allows to turn off the feature.
126 acpi_no_memhotplug [ACPI] Disable memory hotplug. Useful for kdump
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_rsdp= [ACPI,EFI,KEXEC]
140 Pass the RSDP address to the kernel, mostly used
141 on machines running EFI runtime service to boot the
142 second kernel for kdump.
144 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
145 Format: To spoof as Windows 98: ="Microsoft Windows"
147 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
148 of 2 which is mandated by ACPI 6) as the supported ACPI
149 specification revision (when using this switch, it may
150 be necessary to carry out a cold reboot _twice_ in a
151 row to make it take effect on the platform firmware).
153 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
154 acpi_osi="string1" # add string1
155 acpi_osi="!string2" # remove string2
156 acpi_osi=!* # remove all strings
157 acpi_osi=! # disable all built-in OS vendor
159 acpi_osi=!! # enable all built-in OS vendor
161 acpi_osi= # disable all strings
163 'acpi_osi=!' can be used in combination with single or
164 multiple 'acpi_osi="string1"' to support specific OS
165 vendor string(s). Note that such command can only
166 affect the default state of the OS vendor strings, thus
167 it cannot affect the default state of the feature group
168 strings and the current state of the OS vendor strings,
169 specifying it multiple times through kernel command line
170 is meaningless. This command is useful when one do not
171 care about the state of the feature group strings which
172 should be controlled by the OSPM.
174 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
175 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
176 can make '_OSI("Windows 2000")' TRUE.
178 'acpi_osi=' cannot be used in combination with other
179 'acpi_osi=' command lines, the _OSI method will not
180 exist in the ACPI namespace. NOTE that such command can
181 only affect the _OSI support state, thus specifying it
182 multiple times through kernel command line is also
185 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
188 'acpi_osi=!*' can be used in combination with single or
189 multiple 'acpi_osi="string1"' to support specific
190 string(s). Note that such command can affect the
191 current state of both the OS vendor strings and the
192 feature group strings, thus specifying it multiple times
193 through kernel command line is meaningful. But it may
194 still not able to affect the final state of a string if
195 there are quirks related to this string. This command
196 is useful when one want to control the state of the
197 feature group strings to debug BIOS issues related to
200 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
201 '_OSI("Module Device")' FALSE.
202 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
203 '_OSI("Module Device")' TRUE.
204 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
206 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
208 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
209 they all will make '_OSI("Windows 2000")' TRUE.
212 Override the pmtimer bug detection: force the kernel
213 to assume that this machine's pmtimer latches its value
214 and always returns good values.
216 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
217 Format: { level | edge | high | low }
219 acpi_skip_timer_override [HW,ACPI]
220 Recognize and ignore IRQ0/pin2 Interrupt Override.
221 For broken nForce2 BIOS resulting in XT-PIC timer.
223 acpi_sleep= [HW,ACPI] Sleep options
224 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
225 old_ordering, nonvs, sci_force_enable, nobl }
226 See Documentation/power/video.txt for information on
228 s3_beep is for debugging; it makes the PC's speaker beep
229 as soon as the kernel's real-mode entry point is called.
230 s4_nohwsig prevents ACPI hardware signature from being
231 used during resume from hibernation.
232 old_ordering causes the ACPI 1.0 ordering of the _PTS
233 control method, with respect to putting devices into
234 low power states, to be enforced (the ACPI 2.0 ordering
235 of _PTS is used by default).
236 nonvs prevents the kernel from saving/restoring the
237 ACPI NVS memory during suspend/hibernation and resume.
238 sci_force_enable causes the kernel to set SCI_EN directly
239 on resume from S1/S3 (which is against the ACPI spec,
240 but some broken systems don't work without it).
241 nobl causes the internal blacklist of systems known to
242 behave incorrectly in some ways with respect to system
243 suspend and resume to be ignored (use wisely).
245 acpi_use_timer_override [HW,ACPI]
246 Use timer override. For some broken Nvidia NF5 boards
247 that require a timer override, but don't have HPET
249 add_efi_memmap [EFI; X86] Include EFI memory map in
250 kernel's map of available physical RAM.
253 { off | try_unsupported }
254 off: disable AGP support
255 try_unsupported: try to drive unsupported chipsets
256 (may crash computer or cause data corruption)
259 See Documentation/sound/alsa/alsa-parameters.txt
262 Allow the default userspace alignment fault handler
263 behaviour to be specified. Bit 0 enables warnings,
264 bit 1 enables fixups, and bit 2 sends a segfault.
266 align_va_addr= [X86-64]
267 Align virtual addresses by clearing slice [14:12] when
268 allocating a VMA at process creation time. This option
269 gives you up to 3% performance improvement on AMD F15h
270 machines (where it is enabled by default) for a
271 CPU-intensive style benchmark, and it can vary highly in
272 a microbenchmark depending on workload and compiler.
274 32: only for 32-bit processes
275 64: only for 64-bit processes
276 on: enable for both 32- and 64-bit processes
277 off: disable for both 32- and 64-bit processes
279 alloc_snapshot [FTRACE]
280 Allocate the ftrace snapshot buffer on boot up when the
281 main buffer is allocated. This is handy if debugging
282 and you need to use tracing_snapshot() on boot up, and
283 do not want to use tracing_snapshot_alloc() as it needs
284 to be done where GFP_KERNEL allocations are allowed.
286 amd_iommu= [HW,X86-64]
287 Pass parameters to the AMD IOMMU driver in the system.
289 fullflush - enable flushing of IO/TLB entries when
290 they are unmapped. Otherwise they are
291 flushed before they will be reused, which
293 off - do not initialize any AMD IOMMU found in
295 force_isolation - Force device isolation for all
296 devices. The IOMMU driver is not
297 allowed anymore to lift isolation
298 requirements as needed. This option
299 does not override iommu=pt
301 amd_iommu_dump= [HW,X86-64]
302 Enable AMD IOMMU driver option to dump the ACPI table
303 for AMD IOMMU. With this option enabled, AMD IOMMU
304 driver will print ACPI tables for AMD IOMMU during
305 IOMMU initialization.
307 amd_iommu_intr= [HW,X86-64]
308 Specifies one of the following AMD IOMMU interrupt
310 legacy - Use legacy interrupt remapping mode.
311 vapic - Use virtual APIC mode, which allows IOMMU
312 to inject interrupts directly into guest.
313 This mode requires kvm-amd.avic=1.
314 (Default when IOMMU HW support is present.)
316 amijoy.map= [HW,JOY] Amiga joystick support
317 Map of devices attached to JOY0DAT and JOY1DAT
319 See also Documentation/input/joydev/joystick.rst
321 analog.map= [HW,JOY] Analog joystick and gamepad support
322 Specifies type or capabilities of an analog joystick
323 connected to one of 16 gameports
324 Format: <type1>,<type2>,..<type16>
327 Power management functions (SPARCstation-4/5 + deriv.)
329 Disable APC CPU standby support. SPARCstation-Fox does
330 not play well with APC CPU idle - disable it if you have
331 APC and your system crashes randomly.
333 apic= [APIC,X86] Advanced Programmable Interrupt Controller
334 Change the output verbosity whilst booting
335 Format: { quiet (default) | verbose | debug }
336 Change the amount of debugging information output
337 when initialising the APIC and IO-APIC components.
338 For X86-32, this can also be used to specify an APIC
340 Format: apic=driver_name
341 Examples: apic=bigsmp
343 apic_extnmi= [APIC,X86] External NMI delivery setting
344 Format: { bsp (default) | all | none }
345 bsp: External NMI is delivered only to CPU 0
346 all: External NMIs are broadcast to all CPUs as a
348 none: External NMI is masked for all CPUs. This is
349 useful so that a dump capture kernel won't be
353 See Documentation/networking/ipv6.txt.
355 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
356 Limit apic dumping. The parameter defines the maximal
357 number of local apics being dumped. Also it is possible
358 to set it to "all" by meaning -- no limit here.
359 Format: { 1 (default) | 2 | ... | all }.
360 The parameter valid if only apic=debug or
361 apic=verbose is specified.
362 Example: apic=debug show_lapic=all
364 apm= [APM] Advanced Power Management
365 See header of arch/x86/kernel/apm_32.c.
367 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
368 Format: <io>,<irq>,<nodeID>
372 atarimouse= [HW,MOUSE] Atari Mouse
374 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
375 EzKey and similar keyboards
377 atkbd.reset= [HW] Reset keyboard during initialization
379 atkbd.set= [HW] Select keyboard code set
380 Format: <int> (2 = AT (default), 3 = PS/2)
382 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
385 atkbd.softraw= [HW] Choose between synthetic and real raw mode
386 Format: <bool> (0 = real, 1 = synthetic (default))
388 atkbd.softrepeat= [HW]
389 Use software keyboard repeat
391 audit= [KNL] Enable the audit sub-system
392 Format: { "0" | "1" | "off" | "on" }
393 0 | off - kernel audit is disabled and can not be
394 enabled until the next reboot
395 unset - kernel audit is initialized but disabled and
396 will be fully enabled by the userspace auditd.
397 1 | on - kernel audit is initialized and partially
398 enabled, storing at most audit_backlog_limit
399 messages in RAM until it is fully enabled by the
403 audit_backlog_limit= [KNL] Set the audit queue size limit.
404 Format: <int> (must be >=0)
407 bau= [X86_UV] Enable the BAU on SGI UV. The default
408 behavior is to disable the BAU (i.e. bau=0).
409 Format: { "0" | "1" }
412 unset - Disable the BAU.
414 baycom_epp= [HW,AX25]
417 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
419 See header of drivers/net/hamradio/baycom_par.c.
421 baycom_ser_fdx= [HW,AX25]
422 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
423 Format: <io>,<irq>,<mode>[,<baud>]
424 See header of drivers/net/hamradio/baycom_ser_fdx.c.
426 baycom_ser_hdx= [HW,AX25]
427 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
428 Format: <io>,<irq>,<mode>
429 See header of drivers/net/hamradio/baycom_ser_hdx.c.
431 blkdevparts= Manual partition parsing of block device(s) for
432 embedded devices based on command line input.
433 See Documentation/block/cmdline-partition.txt
435 boot_delay= Milliseconds to delay each printk during boot.
436 Values larger than 10 seconds (10000) are changed to
440 bootmem_debug [KNL] Enable bootmem allocator debug messages.
443 Disable BERT OS support on buggy BIOSes.
445 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
446 bttv.radio= Most important insmod options are available as
448 bttv.pll= See Documentation/media/v4l-drivers/bttv.rst
451 bulk_remove=off [PPC] This parameter disables the use of the pSeries
452 firmware feature for flushing multiple hpte entries
455 c101= [NET] Moxa C101 synchronous serial card
457 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
458 Sometimes CPU hardware bugs make them report the cache
459 size incorrectly. The kernel will attempt work arounds
460 to fix known problems, but for some CPUs it is not
461 possible to determine what the correct size should be.
462 This option provides an override for these situations.
464 ca_keys= [KEYS] This parameter identifies a specific key(s) on
465 the system trusted keyring to be used for certificate
467 format: { id:<keyid> | builtin }
469 cca= [MIPS] Override the kernel pages' cache coherency
470 algorithm. Accepted values range from 0 to 7
471 inclusive. See arch/mips/include/asm/pgtable-bits.h
472 for platform specific values (SB1, Loongson3 and
475 ccw_timeout_log [S390]
476 See Documentation/s390/CommonIO for details.
478 cgroup_disable= [KNL] Disable a particular controller
479 Format: {name of the controller(s) to disable}
480 The effects of cgroup_disable=foo are:
481 - foo isn't auto-mounted if you mount all cgroups in
483 - foo isn't visible as an individually mountable
485 {Currently only "memory" controller deal with this and
486 cut the overhead, others just disable the usage. So
487 only cgroup_disable=memory is actually worthy}
489 cgroup_no_v1= [KNL] Disable one, multiple, all cgroup controllers in v1
490 Format: { controller[,controller...] | "all" }
491 Like cgroup_disable, but only applies to cgroup v1;
492 the blacklisted controllers remain available in cgroup2.
494 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
496 nosocket -- Disable socket memory accounting.
497 nokmem -- Disable kernel memory accounting.
499 checkreqprot [SELINUX] Set initial checkreqprot flag value.
500 Format: { "0" | "1" }
501 See security/selinux/Kconfig help text.
502 0 -- check protection applied by kernel (includes
503 any implied execute protection).
504 1 -- check protection requested by application.
505 Default value is set via a kernel config option.
506 Value can be changed at runtime via
507 /selinux/checkreqprot.
510 See Documentation/s390/CommonIO for details.
513 Prevents the clock framework from automatically gating
514 clocks that have not been explicitly enabled by a Linux
515 device driver but are enabled in hardware at reset or
516 by the bootloader/firmware. Note that this does not
517 force such clocks to be always-on nor does it reserve
518 those clocks in any way. This parameter is useful for
519 debug and development, but should not be needed on a
520 platform with proper driver support. For more
521 information, see Documentation/driver-api/clk.rst.
523 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
525 Forces specified clocksource (if available) to be used
526 when calculating gettimeofday(). If specified
527 clocksource is not available, it defaults to PIT.
528 Format: { pit | tsc | cyclone | pmtmr }
530 clocksource= Override the default clocksource
532 Override the default clocksource and use the clocksource
533 with the name specified.
534 Some clocksource names to choose from, depending on
536 [all] jiffies (this is the base, fallback clocksource)
538 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
539 pxa_timer,timer3,32k_counter,timer0_1
540 [X86-32] pit,hpet,tsc;
541 scx200_hrt on Geode; cyclone on IBM x440
549 clocksource.arm_arch_timer.evtstrm=
552 Enable/disable the eventstream feature of the ARM
553 architected timer so that code using WFE-based polling
554 loops can be debugged more effectively on production
557 clearcpuid=BITNUM [X86]
558 Disable CPUID feature X for the kernel. See
559 arch/x86/include/asm/cpufeatures.h for the valid bit
560 numbers. Note the Linux specific bits are not necessarily
561 stable over kernel options, but the vendor specific
563 Also note that user programs calling CPUID directly
564 or using the feature without checking anything
565 will still see it. This just prevents it from
566 being used by the kernel or shown in /proc/cpuinfo.
567 Also note the kernel might malfunction if you disable
570 cma=nn[MG]@[start[MG][-end[MG]]]
572 Sets the size of kernel global memory area for
573 contiguous memory allocations and optionally the
574 placement constraint by the physical address range of
575 memory allocations. A value of 0 disables CMA
576 altogether. For more information, see
577 include/linux/dma-contiguous.h
579 cmo_free_hint= [PPC] Format: { yes | no }
580 Specify whether pages are marked as being inactive
581 when they are freed. This is used in CMO environments
582 to determine OS memory pressure for page stealing by
586 coherent_pool=nn[KMG] [ARM,KNL]
587 Sets the size of memory pool for coherent, atomic dma
588 allocations, by default set to 256K.
590 com20020= [HW,NET] ARCnet - COM20020 chipset
592 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
594 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
598 ARCnet - COM90xx chipset (memory-mapped buffers)
599 Format: <io>[,<irq>[,<memstart>]]
601 condev= [HW,S390] console device
604 console= [KNL] Output console device and options.
606 tty<n> Use the virtual console device <n>.
610 Use the specified serial port. The options are of
611 the form "bbbbpnf", where "bbbb" is the baud rate,
612 "p" is parity ("n", "o", or "e"), "n" is number of
613 bits, and "f" is flow control ("r" for RTS or
614 omit it). Default is "9600n8".
616 See Documentation/admin-guide/serial-console.rst for more
618 Documentation/networking/netconsole.txt for an
621 uart[8250],io,<addr>[,options]
622 uart[8250],mmio,<addr>[,options]
623 uart[8250],mmio16,<addr>[,options]
624 uart[8250],mmio32,<addr>[,options]
625 uart[8250],0x<addr>[,options]
626 Start an early, polled-mode console on the 8250/16550
627 UART at the specified I/O port or MMIO address,
628 switching to the matching ttyS device later.
629 MMIO inter-register address stride is either 8-bit
630 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
631 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
632 to be equivalent to 'mmio'. 'options' are specified in
633 the same format described for ttyS above; if unspecified,
634 the h/w is not re-initialized.
636 hvc<n> Use the hypervisor console device <n>. This is for
637 both Xen and PowerPC hypervisors.
639 If the device connected to the port is not a TTY but a braille
640 device, prepend "brl," before the device type, for instance
642 For now, only VisioBraille is supported.
645 [KNL] Change console messages format
647 By default we print messages on consoles in
648 "[time stamp] text\n" format (time stamp may not be
649 printed, depending on CONFIG_PRINTK_TIME or
650 `printk_time' param).
652 Switch to syslog format: "<%u>[time stamp] text\n"
653 IOW, each message will have a facility and loglevel
654 prefix. The format is similar to one used by syslog()
655 syscall, or to executing "dmesg -S --raw" or to reading
658 consoleblank= [KNL] The console blank (screen saver) timeout in
659 seconds. A value of 0 disables the blank timer.
663 [KNL] Change the default value for
664 /proc/<pid>/coredump_filter.
665 See also Documentation/filesystems/proc.txt.
667 coresight_cpu_debug.enable
670 Enable/disable the CPU sampling based debugging.
671 0: default value, disable debugging
672 1: enable debugging at boot time
674 cpuidle.off=1 [CPU_IDLE]
675 disable the cpuidle sub-system
677 cpufreq.off=1 [CPU_FREQ]
678 disable the cpufreq sub-system
681 [X86] Delay for N microsec between assert and de-assert
682 of APIC INIT to start processors. This delay occurs
683 on every CPU online, such as boot, and resume from suspend.
686 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
688 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
690 crashkernel=size[KMG][@offset[KMG]]
691 [KNL] Using kexec, Linux can switch to a 'crash kernel'
692 upon panic. This parameter reserves the physical
693 memory region [offset, offset + size] for that kernel
694 image. If '@offset' is omitted, then a suitable offset
695 is selected automatically. Check
696 Documentation/kdump/kdump.txt for further details.
698 crashkernel=range1:size1[,range2:size2,...][@offset]
699 [KNL] Same as above, but depends on the memory
700 in the running system. The syntax of range is
701 start-[end] where start and end are both
702 a memory unit (amount[KMG]). See also
703 Documentation/kdump/kdump.txt for an example.
705 crashkernel=size[KMG],high
706 [KNL, x86_64] range could be above 4G. Allow kernel
707 to allocate physical memory region from top, so could
708 be above 4G if system have more than 4G ram installed.
709 Otherwise memory region will be allocated below 4G, if
711 It will be ignored if crashkernel=X is specified.
712 crashkernel=size[KMG],low
713 [KNL, x86_64] range under 4G. When crashkernel=X,high
714 is passed, kernel could allocate physical memory region
715 above 4G, that cause second kernel crash on system
716 that require some amount of low memory, e.g. swiotlb
717 requires at least 64M+32K low memory, also enough extra
718 low memory is needed to make sure DMA buffers for 32-bit
719 devices won't run out. Kernel would try to allocate at
720 at least 256M below 4G automatically.
721 This one let user to specify own low range under 4G
722 for second kernel instead.
723 0: to disable low allocation.
724 It will be ignored when crashkernel=X,high is not used
725 or memory reserved is below 4G.
728 [KNL] Disable crypto self-tests
733 cs89x0_media= [HW,NET]
734 Format: { rj45 | aui | bnc }
737 See header of drivers/s390/block/dasd_devmap.c.
739 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
740 (one device per port)
741 Format: <port#>,<type>
742 See also Documentation/input/devices/joystick-parport.rst
744 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
746 Documentation/admin-guide/dynamic-debug-howto.rst for
747 details. Deprecated, see dyndbg.
749 debug [KNL] Enable kernel debugging (events log level).
752 [KNL] verbose self-tests
754 Print debugging info while doing the locking API
756 We default to 0 (no extra messages), setting it to
757 1 will print _a lot_ more information - normally
758 only useful to kernel developers.
760 debug_objects [KNL] Enable object debugging
763 [KNL] Disable object debugging
765 debug_guardpage_minorder=
766 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
767 parameter allows control of the order of pages that will
768 be intentionally kept free (and hence protected) by the
769 buddy allocator. Bigger value increase the probability
770 of catching random memory corruption, but reduce the
771 amount of memory for normal system use. The maximum
772 possible value is MAX_ORDER/2. Setting this parameter
773 to 1 or 2 should be enough to identify most random
774 memory corruption problems caused by bugs in kernel or
775 driver code when a CPU writes to (or reads from) a
776 random memory location. Note that there exists a class
777 of memory corruptions problems caused by buggy H/W or
778 F/W or by drivers badly programing DMA (basically when
779 memory is written at bus level and the CPU MMU is
780 bypassed) which are not detectable by
781 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
782 tracking down these problems.
785 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
786 parameter enables the feature at boot time. In
787 default, it is disabled. We can avoid allocating huge
788 chunk of memory for debug pagealloc if we don't enable
789 it at boot time and the system will work mostly same
790 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
791 on: enable the feature
793 debugpat [X86] Enable PAT debugging
795 decnet.addr= [HW,NET]
796 Format: <area>[,<node>]
797 See also Documentation/networking/decnet.txt.
800 [same as hugepagesz=] The size of the default
801 HugeTLB page size. This is the size represented by
802 the legacy /proc/ hugepages APIs, used for SHM, and
803 default size when mounting hugetlbfs filesystems.
804 Defaults to the default architecture's huge page size
808 Set number of hash buckets for dentry cache.
810 disable_1tb_segments [PPC]
811 Disables the use of 1TB hash page table segments. This
812 causes the kernel to fall back to 256MB segments which
813 can be useful when debugging issues that require an SLB
817 See Documentation/networking/ipv6.txt.
820 Disable RADIX MMU mode on POWER9
822 disable_cpu_apicid= [X86,APIC,SMP]
824 The number of initial APIC ID for the
825 corresponding CPU to be disabled at boot,
826 mostly used for the kdump 2nd kernel to
827 disable BSP to wake up multiple CPUs without
828 causing system reset or hang due to sending
831 disable_ddw [PPC/PSERIES]
832 Disable Dynamic DMA Window support. Use this if
833 to workaround buggy firmware.
836 See Documentation/networking/ipv6.txt.
838 disable_mtrr_cleanup [X86]
839 The kernel tries to adjust MTRR layout from continuous
840 to discrete, to make X server driver able to add WB
841 entry later. This parameter disables that.
843 disable_mtrr_trim [X86, Intel and AMD only]
844 By default the kernel will trim any uncacheable
845 memory out of your available memory pool based on
846 MTRR settings. This parameter disables that behavior,
847 possibly causing your machine to run very slowly.
849 disable_timer_pin_1 [X86]
850 Disable PIN 1 of APIC timer
851 Can be useful to work around chipset bugs.
853 dis_ucode_ldr [X86] Disable the microcode loader.
855 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
856 this option disables the debugging code at boot.
858 dma_debug_entries=<number>
859 This option allows to tune the number of preallocated
860 entries for DMA-API debugging code. One entry is
861 required per DMA-API allocation. Use this if the
862 DMA-API debugging code disables itself because the
863 architectural default is too low.
865 dma_debug_driver=<driver_name>
866 With this option the DMA-API debugging driver
867 filter feature can be enabled at boot time. Just
868 pass the driver to filter for as the parameter.
869 The filter can be disabled or changed to another
870 driver later using sysfs.
872 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
873 Broken monitors, graphic adapters, KVMs and EDIDless
874 panels may send no or incorrect EDID data sets.
875 This parameter allows to specify an EDID data sets
876 in the /lib/firmware directory that are used instead.
877 Generic built-in EDID data sets are used, if one of
878 edid/1024x768.bin, edid/1280x1024.bin,
879 edid/1680x1050.bin, or edid/1920x1080.bin is given
880 and no file with the same name exists. Details and
881 instructions how to build your own EDID data are
882 available in Documentation/EDID/HOWTO.txt. An EDID
883 data set will only be used for a particular connector,
884 if its name and a colon are prepended to the EDID
885 name. Each connector may use a unique EDID data
886 set by separating the files with a comma. An EDID
887 data set with no connector name will be used for
888 any connectors not explicitly specified.
893 Format: {"off" | "known"}
894 Control how the dt_cpu_ftrs device-tree binding is
895 used for CPU feature discovery and setup (if it
897 off: Do not use it, fall back to legacy cpu table.
898 known: Do not pass through unknown features to guests
899 or userspace, only those that the kernel is aware of.
901 dump_apple_properties [X86]
902 Dump name and content of EFI device properties on
903 x86 Macs. Useful for driver authors to determine
904 what data is available or for reverse-engineering.
906 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
907 module.dyndbg[="val"]
908 Enable debug messages at boot time. See
909 Documentation/admin-guide/dynamic-debug-howto.rst
912 nompx [X86] Disables Intel Memory Protection Extensions.
913 See Documentation/x86/intel_mpx.txt for more
914 information about the feature.
916 nopku [X86] Disable Memory Protection Keys CPU feature found
919 module.async_probe [KNL]
920 Enable asynchronous probe on this module.
922 early_ioremap_debug [KNL]
923 Enable debug messages in early_ioremap support. This
924 is useful for tracking down temporary early mappings
925 which are not unmapped.
927 earlycon= [KNL] Output early console device and options.
929 [ARM64] The early console is determined by the
930 stdout-path property in device tree's chosen node,
931 or determined by the ACPI SPCR table.
933 [X86] When used with no options the early console is
934 determined by the ACPI SPCR table.
936 cdns,<addr>[,options]
937 Start an early, polled-mode console on a Cadence
938 (xuartps) serial port at the specified address. Only
939 supported option is baud rate. If baud rate is not
940 specified, the serial port must already be setup and
943 uart[8250],io,<addr>[,options]
944 uart[8250],mmio,<addr>[,options]
945 uart[8250],mmio32,<addr>[,options]
946 uart[8250],mmio32be,<addr>[,options]
947 uart[8250],0x<addr>[,options]
948 Start an early, polled-mode console on the 8250/16550
949 UART at the specified I/O port or MMIO address.
950 MMIO inter-register address stride is either 8-bit
951 (mmio) or 32-bit (mmio32 or mmio32be).
952 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
953 to be equivalent to 'mmio'. 'options' are specified
954 in the same format described for "console=ttyS<n>"; if
955 unspecified, the h/w is not initialized.
959 Start an early, polled-mode console on a pl011 serial
960 port at the specified address. The pl011 serial port
961 must already be setup and configured. Options are not
962 yet supported. If 'mmio32' is specified, then only
963 the driver will use only 32-bit accessors to read/write
964 the device registers.
967 Start an early, polled-mode console on a meson serial
968 port at the specified address. The serial port must
969 already be setup and configured. Options are not yet
973 Start an early, polled-mode console on an msm serial
974 port at the specified address. The serial port
975 must already be setup and configured. Options are not
979 Start an early, polled-mode console on an msm serial
980 dm port at the specified address. The serial port
981 must already be setup and configured. Options are not
985 Start an early, polled-mode console on a serial port
986 of an Actions Semi SoC, such as S500 or S900, at the
987 specified address. The serial port must already be
988 setup and configured. Options are not yet supported.
990 smh Use ARM semihosting calls for early console.
998 Use early console provided by serial driver available
999 on Samsung SoCs, requires selecting proper type and
1000 a correct base address of the selected UART port. The
1001 serial port must already be setup and configured.
1002 Options are not yet supported.
1005 Start an early, polled-mode console on a lantiq serial
1006 (lqasc) port at the specified address. The serial port
1007 must already be setup and configured. Options are not
1012 Use early console provided by Freescale LP UART driver
1013 found on Freescale Vybrid and QorIQ LS1021A processors.
1014 A valid base address must be provided, and the serial
1015 port must already be setup and configured.
1018 Start an early, polled-mode console on the
1019 Armada 3700 serial port at the specified
1020 address. The serial port must already be setup
1021 and configured. Options are not yet supported.
1024 Start an early, polled-mode console on a Qualcomm
1025 Generic Interface (GENI) based serial port at the
1026 specified address. The serial port must already be
1027 setup and configured. Options are not yet supported.
1029 earlyprintk= [X86,SH,ARM,M68k,S390]
1034 earlyprintk=serial[,ttySn[,baudrate]]
1035 earlyprintk=serial[,0x...[,baudrate]]
1036 earlyprintk=ttySn[,baudrate]
1037 earlyprintk=dbgp[debugController#]
1038 earlyprintk=pciserial,bus:device.function[,baudrate]
1039 earlyprintk=xdbc[xhciController#]
1041 earlyprintk is useful when the kernel crashes before
1042 the normal console is initialized. It is not enabled by
1043 default because it has some cosmetic problems.
1045 Append ",keep" to not disable it when the real console
1048 Only one of vga, efi, serial, or usb debug port can
1051 Currently only ttyS0 and ttyS1 may be specified by
1052 name. Other I/O ports may be explicitly specified
1053 on some architectures (x86 and arm at least) by
1054 replacing ttySn with an I/O port address, like this:
1055 earlyprintk=serial,0x1008,115200
1056 You can find the port for a given device in
1057 /proc/tty/driver/serial:
1058 2: uart:ST16650V2 port:00001008 irq:18 ...
1060 Interaction with the standard serial driver is not
1063 The VGA and EFI output is eventually overwritten by
1066 The xen output can only be used by Xen PV guests.
1068 The sclp output can only be used on s390.
1070 edac_report= [HW,EDAC] Control how to report EDAC event
1071 Format: {"on" | "off" | "force"}
1072 on: enable EDAC to report H/W event. May be overridden
1073 by other higher priority error reporting module.
1074 off: disable H/W event reporting through EDAC.
1075 force: enforce the use of EDAC to report H/W event.
1078 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1081 This is designed to be used in conjunction with
1082 the boot argument: earlyprintk=vga
1085 Format: {"off" | "on" | "skip[mbr]"}
1088 Format: { "old_map", "nochunk", "noruntime", "debug" }
1089 old_map [X86-64]: switch to the old ioremap-based EFI
1090 runtime services mapping. 32-bit still uses this one by
1092 nochunk: disable reading files in "chunks" in the EFI
1093 boot stub, as chunking can cause problems with some
1094 firmware implementations.
1095 noruntime : disable EFI runtime services support
1096 debug: enable misc debug output
1098 efi_no_storage_paranoia [EFI; X86]
1099 Using this parameter you can use more than 50% of
1100 your efi variable storage. Use this parameter only if
1101 you are really sure that your UEFI does sane gc and
1102 fulfills the spec otherwise your board may brick.
1104 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1105 Add arbitrary attribute to specific memory range by
1106 updating original EFI memory map.
1107 Region of memory which aa attribute is added to is
1109 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1110 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1111 attribute is added to range 0x100000000-0x180000000 and
1112 0x10a0000000-0x1120000000.
1114 Using this parameter you can do debugging of EFI memmap
1115 related feature. For example, you can do debugging of
1116 Address Range Mirroring feature even if your box
1119 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1120 that is to be dynamically loaded by Linux. If there are
1121 multiple variables with the same name but with different
1122 vendor GUIDs, all of them will be loaded. See
1123 Documentation/acpi/ssdt-overlays.txt for details.
1126 eisa_irq_edge= [PARISC,HW]
1127 See header of drivers/parisc/eisa.c.
1130 See comment before function elanfreq_setup() in
1131 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1134 Format: {"cfq" | "deadline" | "noop"}
1135 See Documentation/block/cfq-iosched.txt and
1136 Documentation/block/deadline-iosched.txt for details.
1138 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1139 Specifies physical address of start of kernel core
1140 image elf header and optionally the size. Generally
1141 kexec loader will pass this option to capture kernel.
1142 See Documentation/kdump/kdump.txt for details.
1144 enable_mtrr_cleanup [X86]
1145 The kernel tries to adjust MTRR layout from continuous
1146 to discrete, to make X server driver able to add WB
1147 entry later. This parameter enables that.
1149 enable_timer_pin_1 [X86]
1150 Enable PIN 1 of APIC timer
1151 Can be useful to work around chipset bugs
1152 (in particular on some ATI chipsets).
1153 The kernel tries to set a reasonable default.
1155 enforcing [SELINUX] Set initial enforcing status.
1157 See security/selinux/Kconfig help text.
1158 0 -- permissive (log only, no denials).
1159 1 -- enforcing (deny and log).
1161 Value can be changed at runtime via /selinux/enforce.
1164 Disable Error Record Serialization Table (ERST)
1167 ether= [HW,NET] Ethernet cards parameters
1168 This option is obsoleted by the "netdev=" option, which
1169 has equivalent usage. See its documentation for details.
1173 Permit 'security.evm' to be updated regardless of
1174 current integrity status.
1178 fail_make_request=[KNL]
1179 General fault injection mechanism.
1180 Format: <interval>,<probability>,<space>,<times>
1181 See also Documentation/fault-injection/.
1184 See Documentation/blockdev/floppy.txt.
1186 force_pal_cache_flush
1187 [IA-64] Avoid check_sal_cache_flush which may hang on
1188 buggy SAL_CACHE_FLUSH implementations. Using this
1189 parameter will force ia64_sal_cache_flush to call
1190 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1193 Forcefully enable Physical Address Extension (PAE).
1194 Many Pentium M systems disable PAE but may have a
1195 functionally usable PAE implementation.
1196 Warning: use of this parameter will taint the kernel
1197 and may cause unknown problems.
1200 [FTRACE] will set and start the specified tracer
1201 as early as possible in order to facilitate early
1204 ftrace_dump_on_oops[=orig_cpu]
1205 [FTRACE] will dump the trace buffers on oops.
1206 If no parameter is passed, ftrace will dump
1207 buffers of all CPUs, but if you pass orig_cpu, it will
1208 dump only the buffer of the CPU that triggered the
1211 ftrace_filter=[function-list]
1212 [FTRACE] Limit the functions traced by the function
1213 tracer at boot up. function-list is a comma separated
1214 list of functions. This list can be changed at run
1215 time by the set_ftrace_filter file in the debugfs
1218 ftrace_notrace=[function-list]
1219 [FTRACE] Do not trace the functions specified in
1220 function-list. This list can be changed at run time
1221 by the set_ftrace_notrace file in the debugfs
1224 ftrace_graph_filter=[function-list]
1225 [FTRACE] Limit the top level callers functions traced
1226 by the function graph tracer at boot up.
1227 function-list is a comma separated list of functions
1228 that can be changed at run time by the
1229 set_graph_function file in the debugfs tracing directory.
1231 ftrace_graph_notrace=[function-list]
1232 [FTRACE] Do not trace from the functions specified in
1233 function-list. This list is a comma separated list of
1234 functions that can be changed at run time by the
1235 set_graph_notrace file in the debugfs tracing directory.
1237 ftrace_graph_max_depth=<uint>
1238 [FTRACE] Used with the function graph tracer. This is
1239 the max depth it will trace into a function. This value
1240 can be changed at run time by the max_graph_depth file
1241 in the tracefs tracing directory. default: 0 (no limit)
1244 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1245 support via parallel port (up to 5 devices per port)
1246 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1247 See also Documentation/input/devices/joystick-parport.rst
1251 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1255 gcov_persist= [GCOV] When non-zero (default), profiling data for
1256 kernel modules is saved and remains accessible via
1257 debugfs, even when the module is unloaded/reloaded.
1258 When zero, profiling data is discarded and associated
1259 debugfs files are removed at module unload time.
1261 goldfish [X86] Enable the goldfish android emulator platform.
1262 Don't use this when you are not running on the
1265 gpt [EFI] Forces disk with valid GPT signature but
1266 invalid Protective MBR to be treated as GPT. If the
1267 primary GPT is corrupted, it enables the backup/alternate
1268 GPT to be used instead.
1270 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1271 the "Enable 0" bit of the configuration register.
1274 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1275 the "Enable 0" bit of the configuration register.
1278 grcan.select= [HW] Select which physical interface to use.
1281 grcan.txsize= [HW] Sets the size of the tx buffer.
1282 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1284 grcan.rxsize= [HW] Sets the size of the rx buffer.
1285 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1288 gpio-mockup.gpio_mockup_ranges
1289 [HW] Sets the ranges of gpiochip of for this device.
1290 Format: <start1>,<end1>,<start2>,<end2>...
1292 hardlockup_all_cpu_backtrace=
1293 [KNL] Should the hard-lockup detector generate
1294 backtraces on all cpus.
1297 hashdist= [KNL,NUMA] Large hashes allocated during boot
1298 are distributed across NUMA nodes. Defaults on
1299 for 64-bit NUMA, off otherwise.
1300 Format: 0 | 1 (for off | on)
1302 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1304 hd= [EIDE] (E)IDE hard drive subsystem geometry
1305 Format: <cyl>,<head>,<sect>
1308 Disable Hardware Error Source Table (HEST) support;
1309 corresponding firmware-first mode error processing
1310 logic will be disabled.
1312 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1313 size of <nn>. This works even on boxes that have no
1314 highmem otherwise. This also works to reduce highmem
1315 size on bigger boxes.
1317 highres= [KNL] Enable/disable high resolution timer mode.
1318 Valid parameters: "on", "off"
1322 See Documentation/isdn/README.HiSax.
1326 hpet= [X86-32,HPET] option to control HPET usage
1327 Format: { enable (default) | disable | force |
1329 disable: disable HPET and use PIT instead
1330 force: allow force enabled of undocumented chips (ICH4,
1332 verbose: show contents of HPET registers during setup
1334 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1335 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1337 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1338 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1339 On x86-64 and powerpc, this option can be specified
1340 multiple times interleaved with hugepages= to reserve
1341 huge pages of different sizes. Valid pages sizes on
1342 x86-64 are 2M (when the CPU supports "pse") and 1G
1343 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1346 [KNL] Should the hung task detector generate panics.
1349 A nonzero value instructs the kernel to panic when a
1350 hung task is detected. The default value is controlled
1351 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1352 option. The value selected by this boot parameter can
1353 be changed later by the kernel.hung_task_panic sysctl.
1355 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1356 terminal devices. Valid values: 0..8
1357 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1358 If specified, z/VM IUCV HVC accepts connections
1359 from listed z/VM user IDs only.
1361 Do not unregister boot console at start. This is only
1362 useful for debugging when something happens in the window
1363 between unregistering the boot console and initializing
1366 i2c_bus= [HW] Override the default board specific I2C bus speed
1367 or register an additional I2C bus that is not
1368 registered from board initialization code.
1372 i8042.debug [HW] Toggle i8042 debug mode
1373 i8042.unmask_kbd_data
1374 [HW] Enable printing of interrupt data from the KBD port
1375 (disabled by default, and as a pre-condition
1376 requires that i8042.debug=1 be enabled)
1377 i8042.direct [HW] Put keyboard port into non-translated mode
1378 i8042.dumbkbd [HW] Pretend that controller can only read data from
1379 keyboard and cannot control its state
1380 (Don't attempt to blink the leds)
1381 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1382 i8042.nokbd [HW] Don't check/create keyboard port
1383 i8042.noloop [HW] Disable the AUX Loopback command while probing
1385 i8042.nomux [HW] Don't check presence of an active multiplexing
1387 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1389 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1390 i8042.reset [HW] Reset the controller during init, cleanup and
1391 suspend-to-ram transitions, only during s2r
1392 transitions, or never reset
1393 Format: { 1 | Y | y | 0 | N | n }
1394 1, Y, y: always reset controller
1395 0, N, n: don't ever reset controller
1396 Default: only on s2r transitions on x86; most other
1397 architectures force reset to be always executed
1398 i8042.unlock [HW] Unlock (ignore) the keylock
1399 i8042.kbdreset [HW] Reset device connected to KBD port
1403 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1404 indicates that the driver is running on unsupported
1406 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1407 does not match list of supported models.
1409 [HW] Report power status in /proc/i8k
1410 (disabled by default)
1411 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1414 i915.invert_brightness=
1415 [DRM] Invert the sense of the variable that is used to
1416 set the brightness of the panel backlight. Normally a
1417 brightness value of 0 indicates backlight switched off,
1418 and the maximum of the brightness value sets the backlight
1419 to maximum brightness. If this parameter is set to 0
1420 (default) and the machine requires it, or this parameter
1421 is set to 1, a brightness value of 0 sets the backlight
1422 to maximum brightness, and the maximum of the brightness
1423 value switches the backlight off.
1424 -1 -- never invert brightness
1425 0 -- machine default
1426 1 -- force brightness inversion
1429 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1431 ide-core.nodma= [HW] (E)IDE subsystem
1432 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1433 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1434 .cdrom .chs .ignore_cable are additional options
1435 See Documentation/ide/ide.txt.
1437 ide-generic.probe-mask= [HW] (E)IDE subsystem
1439 Probe mask for legacy ISA IDE ports. Depending on
1440 platform up to 6 ports are supported, enabled by
1441 setting corresponding bits in the mask to 1. The
1442 default value is 0x0, which has a special meaning.
1443 On systems that have PCI, it triggers scanning the
1444 PCI bus for the first and the second port, which
1445 are then probed. On systems without PCI the value
1446 of 0x0 enables probing the two first ports as if it
1449 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1450 Claim all unknown PCI IDE storage controllers.
1453 Format: idle=poll, idle=halt, idle=nomwait
1454 Poll forces a polling idle loop that can slightly
1455 improve the performance of waking up a idle CPU, but
1456 will use a lot of power and make the system run hot.
1458 idle=halt: Halt is forced to be used for CPU idle.
1459 In such case C2/C3 won't be used again.
1460 idle=nomwait: Disable mwait for CPU C-states
1462 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1463 Format: { strict | legacy | 2008 | relaxed }
1466 Choose which programs will be accepted for execution
1467 based on the IEEE 754 NaN encoding(s) supported by
1468 the FPU and the NaN encoding requested with the value
1469 of an ELF file header flag individually set by each
1470 binary. Hardware implementations are permitted to
1471 support either or both of the legacy and the 2008 NaN
1474 Available settings are as follows:
1475 strict accept binaries that request a NaN encoding
1476 supported by the FPU
1477 legacy only accept legacy-NaN binaries, if supported
1479 2008 only accept 2008-NaN binaries, if supported
1481 relaxed accept any binaries regardless of whether
1482 supported by the FPU
1484 The FPU emulator is always able to support both NaN
1485 encodings, so if no FPU hardware is present or it has
1486 been disabled with 'nofpu', then the settings of
1487 'legacy' and '2008' strap the emulator accordingly,
1488 'relaxed' straps the emulator for both legacy-NaN and
1489 2008-NaN, whereas 'strict' enables legacy-NaN only on
1490 legacy processors and both NaN encodings on MIPS32 or
1493 The setting for ABS.fmt/NEG.fmt instruction execution
1494 mode generally follows that for the NaN encoding,
1495 except where unsupported by hardware.
1497 ignore_loglevel [KNL]
1498 Ignore loglevel setting - this will print /all/
1499 kernel messages to the console. Useful for debugging.
1500 We also add it as printk module parameter, so users
1501 could change it dynamically, usually by
1502 /sys/module/printk/parameters/ignore_loglevel.
1505 Ignore RLIMIT_DATA setting for data mappings,
1506 print warning at first misuse. Can be changed via
1507 /sys/module/kernel/parameters/ignore_rlimit_data.
1509 ihash_entries= [KNL]
1510 Set number of hash buckets for inode cache.
1512 ima_appraise= [IMA] appraise integrity measurements
1513 Format: { "off" | "enforce" | "fix" | "log" }
1516 ima_appraise_tcb [IMA]
1517 The builtin appraise policy appraises all files
1520 ima_canonical_fmt [IMA]
1521 Use the canonical format for the binary runtime
1522 measurements, instead of host native format.
1525 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1529 The list of supported hash algorithms is defined
1530 in crypto/hash_info.h.
1533 The builtin policies to load during IMA setup.
1534 Format: "tcb | appraise_tcb | secure_boot |
1537 The "tcb" policy measures all programs exec'd, files
1538 mmap'd for exec, and all files opened with the read
1539 mode bit set by either the effective uid (euid=0) or
1542 The "appraise_tcb" policy appraises the integrity of
1543 all files owned by root. (This is the equivalent
1544 of ima_appraise_tcb.)
1546 The "secure_boot" policy appraises the integrity
1547 of files (eg. kexec kernel image, kernel modules,
1548 firmware, policy, etc) based on file signatures.
1550 The "fail_securely" policy forces file signature
1551 verification failure also on privileged mounted
1552 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1555 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1556 Load a policy which meets the needs of the Trusted
1557 Computing Base. This means IMA will measure all
1558 programs exec'd, files mmap'd for exec, and all files
1559 opened for read by uid=0.
1562 Select one of defined IMA measurements template formats.
1563 Formats: { "ima" | "ima-ng" | "ima-sig" }
1567 [IMA] Define a custom template format.
1568 Format: { "field1|...|fieldN" }
1570 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1571 Format: <min_file_size>
1572 Set the minimal file size for using asynchronous hash.
1573 If left unspecified, ahash usage is disabled.
1575 ahash performance varies for different data sizes on
1576 different crypto accelerators. This option can be used
1577 to achieve the best performance for a particular HW.
1579 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1581 Set hashing buffer size. Default: 4k.
1583 ahash performance varies for different chunk sizes on
1584 different crypto accelerators. This option can be used
1585 to achieve best performance for particular HW.
1589 Run specified binary instead of /sbin/init as init
1592 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1593 for working out where the kernel is dying during
1596 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1597 initcall functions. Useful for debugging built-in
1598 modules and initcalls.
1600 initrd= [BOOT] Specify the location of the initial ramdisk
1602 init_pkru= [x86] Specify the default memory protection keys rights
1603 register contents for all processes. 0x55555554 by
1604 default (disallow access to all but pkey 0). Can
1605 override in debugfs after boot.
1607 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1610 int_pln_enable [x86] Enable power limit notification interrupt
1612 integrity_audit=[IMA]
1613 Format: { "0" | "1" }
1614 0 -- basic integrity auditing messages. (Default)
1615 1 -- additional integrity auditing messages.
1617 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1619 Enable intel iommu driver.
1621 Disable intel iommu driver.
1622 igfx_off [Default Off]
1623 By default, gfx is mapped as normal device. If a gfx
1624 device has a dedicated DMAR unit, the DMAR unit is
1625 bypassed by not enabling DMAR with this option. In
1626 this case, gfx device will use physical address for
1629 With this option iommu will not optimize to look
1630 for io virtual address below 32-bit forcing dual
1631 address cycle on pci bus for cards supporting greater
1632 than 32-bit addressing. The default is to look
1633 for translation below 32-bit and if not available
1634 then look in the higher range.
1635 strict [Default Off]
1636 With this option on every unmap_single operation will
1637 result in a hardware IOTLB flush operation as opposed
1638 to batching them for performance.
1639 sp_off [Default Off]
1640 By default, super page will be supported if Intel IOMMU
1641 has the capability. With this option, super page will
1643 ecs_off [Default Off]
1644 By default, extended context tables will be supported if
1645 the hardware advertises that it has support both for the
1646 extended tables themselves, and also PASID support. With
1647 this option set, extended tables will not be used even
1648 on hardware which claims to support them.
1649 tboot_noforce [Default Off]
1650 Do not force the Intel IOMMU enabled under tboot.
1651 By default, tboot will force Intel IOMMU on, which
1652 could harm performance of some high-throughput
1653 devices like 40GBit network cards, even if identity
1655 Note that using this option lowers the security
1656 provided by tboot because it makes the system
1657 vulnerable to DMA attacks.
1659 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1660 0 disables intel_idle and fall back on acpi_idle.
1661 1 to 9 specify maximum depth of C-state.
1665 Do not enable intel_pstate as the default
1666 scaling driver for the supported processors
1668 Use intel_pstate as a scaling driver, but configure it
1669 to work with generic cpufreq governors (instead of
1670 enabling its internal governor). This mode cannot be
1671 used along with the hardware-managed P-states (HWP)
1674 Enable intel_pstate on systems that prohibit it by default
1675 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1676 instead of acpi-cpufreq may disable platform features, such
1677 as thermal controls and power capping, that rely on ACPI
1678 P-States information being indicated to OSPM and therefore
1679 should be used with caution. This option does not work with
1680 processors that aren't supported by the intel_pstate driver
1681 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1683 Do not enable hardware P state control (HWP)
1686 Only load intel_pstate on systems which support
1687 hardware P state control (HWP) if available.
1689 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1690 Description Table, specifies preferred power management
1691 profile as "Enterprise Server" or "Performance Server",
1692 then this feature is turned on by default.
1694 Allow per-logical-CPU P-State performance control limits using
1695 cpufreq sysfs interface
1697 intremap= [X86-64, Intel-IOMMU]
1698 on enable Interrupt Remapping (default)
1699 off disable Interrupt Remapping
1700 nosid disable Source ID checking
1702 BIOS x2APIC opt-out request will be ignored
1703 nopost disable Interrupt Posting
1705 iomem= Disable strict checking of access to MMIO memory
1706 strict regions from userspace.
1720 nobypass [PPC/POWERNV]
1721 Disable IOMMU bypass, using IOMMU for PCI devices.
1724 [ARM64] Configure DMA to bypass the IOMMU by default.
1725 Format: { "0" | "1" }
1726 0 - Use IOMMU translation for DMA.
1727 1 - Bypass the IOMMU for DMA.
1728 unset - Use IOMMU translation for DMA.
1730 io7= [HW] IO7 for Marvel based alpha systems
1731 See comment before marvel_specify_io7 in
1732 arch/alpha/kernel/core_marvel.c.
1734 io_delay= [X86] I/O delay method
1736 Standard port 0x80 based delay
1738 Alternate port 0xed based delay (needed on some systems)
1740 Simple two microseconds delay
1745 See Documentation/filesystems/nfs/nfsroot.txt.
1747 irqaffinity= [SMP] Set the default irq affinity mask
1748 The argument is a cpu list, as described above.
1750 irqchip.gicv2_force_probe=
1753 Force the kernel to look for the second 4kB page
1754 of a GICv2 controller even if the memory range
1755 exposed by the device tree is too small.
1757 irqchip.gicv3_nolpi=
1759 Force the kernel to ignore the availability of
1760 LPIs (and by consequence ITSs). Intended for system
1761 that use the kernel as a bootloader, and thus want
1762 to let secondary kernels in charge of setting up
1766 When an interrupt is not handled search all handlers
1767 for it. Intended to get systems with badly broken
1771 When an interrupt is not handled search all handlers
1772 for it. Also check all handlers each timer
1773 interrupt. Intended to get systems with badly broken
1777 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1779 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1780 [Deprecated - use cpusets instead]
1781 Format: [flag-list,]<cpu-list>
1783 Specify one or more CPUs to isolate from disturbances
1784 specified in the flag list (default: domain):
1787 Disable the tick when a single task runs.
1789 A residual 1Hz tick is offloaded to workqueues, which you
1790 need to affine to housekeeping through the global
1791 workqueue's affinity configured via the
1792 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1793 by using the 'domain' flag described below.
1795 NOTE: by default the global workqueue runs on all CPUs,
1796 so to protect individual CPUs the 'cpumask' file has to
1797 be configured manually after bootup.
1800 Isolate from the general SMP balancing and scheduling
1801 algorithms. Note that performing domain isolation this way
1802 is irreversible: it's not possible to bring back a CPU to
1803 the domains once isolated through isolcpus. It's strongly
1804 advised to use cpusets instead to disable scheduler load
1805 balancing through the "cpuset.sched_load_balance" file.
1806 It offers a much more flexible interface where CPUs can
1807 move in and out of an isolated set anytime.
1809 You can move a process onto or off an "isolated" CPU via
1810 the CPU affinity syscalls or cpuset.
1811 <cpu number> begins at 0 and the maximum value is
1812 "number of CPUs in system - 1".
1814 The format of <cpu-list> is described above.
1820 ivrs_ioapic [HW,X86_64]
1821 Provide an override to the IOAPIC-ID<->DEVICE-ID
1822 mapping provided in the IVRS ACPI table. For
1823 example, to map IOAPIC-ID decimal 10 to
1824 PCI device 00:14.0 write the parameter as:
1825 ivrs_ioapic[10]=00:14.0
1827 ivrs_hpet [HW,X86_64]
1828 Provide an override to the HPET-ID<->DEVICE-ID
1829 mapping provided in the IVRS ACPI table. For
1830 example, to map HPET-ID decimal 0 to
1831 PCI device 00:14.0 write the parameter as:
1832 ivrs_hpet[0]=00:14.0
1834 ivrs_acpihid [HW,X86_64]
1835 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1836 mapping provided in the IVRS ACPI table. For
1837 example, to map UART-HID:UID AMD0020:0 to
1838 PCI device 00:14.5 write the parameter as:
1839 ivrs_acpihid[00:14.5]=AMD0020:0
1841 js= [HW,JOY] Analog joystick
1842 See Documentation/input/joydev/joystick.rst.
1845 When CONFIG_RANDOMIZE_BASE is set, this disables
1846 kernel and module base offset ASLR (Address Space
1847 Layout Randomization).
1850 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
1851 report on every invalid memory access. Without this
1852 parameter KASAN will print report only for the first
1857 kernelcore= [KNL,X86,IA-64,PPC]
1858 Format: nn[KMGTPE] | nn% | "mirror"
1859 This parameter specifies the amount of memory usable by
1860 the kernel for non-movable allocations. The requested
1861 amount is spread evenly throughout all nodes in the
1862 system as ZONE_NORMAL. The remaining memory is used for
1863 movable memory in its own zone, ZONE_MOVABLE. In the
1864 event, a node is too small to have both ZONE_NORMAL and
1865 ZONE_MOVABLE, kernelcore memory will take priority and
1866 other nodes will have a larger ZONE_MOVABLE.
1868 ZONE_MOVABLE is used for the allocation of pages that
1869 may be reclaimed or moved by the page migration
1870 subsystem. Note that allocations like PTEs-from-HighMem
1871 still use the HighMem zone if it exists, and the Normal
1872 zone if it does not.
1874 It is possible to specify the exact amount of memory in
1875 the form of "nn[KMGTPE]", a percentage of total system
1876 memory in the form of "nn%", or "mirror". If "mirror"
1877 option is specified, mirrored (reliable) memory is used
1878 for non-movable allocations and remaining memory is used
1879 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
1880 are exclusive, so you cannot specify multiple forms.
1882 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1883 Format: <Controller#>[,poll interval]
1884 The controller # is the number of the ehci usb debug
1885 port as it is probed via PCI. The poll interval is
1886 optional and is the number seconds in between
1887 each poll cycle to the debug port in case you need
1888 the functionality for interrupting the kernel with
1889 gdb or control-c on the dbgp connection. When
1890 not using this parameter you use sysrq-g to break into
1891 the kernel debugger.
1893 kgdboc= [KGDB,HW] kgdb over consoles.
1894 Requires a tty driver that supports console polling,
1895 or a supported polling keyboard driver (non-usb).
1896 Serial only format: <serial_device>[,baud]
1897 keyboard only format: kbd
1898 keyboard and serial format: kbd,<serial_device>[,baud]
1899 Optional Kernel mode setting:
1900 kms, kbd format: kms,kbd
1901 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1903 kgdbwait [KGDB] Stop kernel execution and enter the
1904 kernel debugger at the earliest opportunity.
1906 kmac= [MIPS] korina ethernet MAC address.
1907 Configure the RouterBoard 532 series on-chip
1908 Ethernet adapter MAC address.
1910 kmemleak= [KNL] Boot-time kmemleak enable/disable
1911 Valid arguments: on, off
1913 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1916 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1917 Default is 0 (don't ignore, but inject #GP)
1919 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
1920 Default is false (don't support).
1922 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1926 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1927 Default is 1 (enabled)
1929 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1931 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1933 kvm-arm.vgic_v3_group0_trap=
1934 [KVM,ARM] Trap guest accesses to GICv3 group-0
1937 kvm-arm.vgic_v3_group1_trap=
1938 [KVM,ARM] Trap guest accesses to GICv3 group-1
1941 kvm-arm.vgic_v3_common_trap=
1942 [KVM,ARM] Trap guest accesses to GICv3 common
1945 kvm-arm.vgic_v4_enable=
1946 [KVM,ARM] Allow use of GICv4 for direct injection of
1949 kvm-intel.ept= [KVM,Intel] Disable extended page tables
1950 (virtualized MMU) support on capable Intel chips.
1951 Default is 1 (enabled)
1953 kvm-intel.emulate_invalid_guest_state=
1954 [KVM,Intel] Enable emulation of invalid guest states
1955 Default is 0 (disabled)
1957 kvm-intel.flexpriority=
1958 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
1959 Default is 1 (enabled)
1962 [KVM,Intel] Enable VMX nesting (nVMX).
1963 Default is 0 (disabled)
1965 kvm-intel.unrestricted_guest=
1966 [KVM,Intel] Disable unrestricted guest feature
1967 (virtualized real and unpaged mode) on capable
1968 Intel chips. Default is 1 (enabled)
1970 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
1971 feature (tagged TLBs) on capable Intel chips.
1972 Default is 1 (enabled)
1978 lapic [X86-32,APIC] Enable the local APIC even if BIOS
1981 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
1982 value for LAPIC timer one-shot implementation. Default
1983 back to the programmable timer unit in the LAPIC.
1985 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
1988 libata.dma= [LIBATA] DMA control
1989 libata.dma=0 Disable all PATA and SATA DMA
1990 libata.dma=1 PATA and SATA Disk DMA only
1991 libata.dma=2 ATAPI (CDROM) DMA only
1992 libata.dma=4 Compact Flash DMA only
1993 Combinations also work, so libata.dma=3 enables DMA
1994 for disks and CDROMs, but not CFs.
1996 libata.ignore_hpa= [LIBATA] Ignore HPA limit
1997 libata.ignore_hpa=0 keep BIOS limits (default)
1998 libata.ignore_hpa=1 ignore limits, using full disk
2000 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2004 libata.force= [LIBATA] Force configurations. The format is comma
2005 separated list of "[ID:]VAL" where ID is
2006 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2007 matching port, link or device. Basically, it matches
2008 the ATA ID string printed on console by libata. If
2009 the whole ID part is omitted, the last PORT and DEVICE
2010 values are used. If ID hasn't been specified yet, the
2011 configuration applies to all ports, links and devices.
2013 If only DEVICE is omitted, the parameter applies to
2014 the port and all links and devices behind it. DEVICE
2015 number of 0 either selects the first device or the
2016 first fan-out link behind PMP device. It does not
2017 select the host link. DEVICE number of 15 selects the
2018 host link and device attached to it.
2020 The VAL specifies the configuration to force. As long
2021 as there's no ambiguity shortcut notation is allowed.
2022 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2023 The following configurations can be forced.
2025 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2026 Any ID with matching PORT is used.
2028 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2030 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2031 udma[/][16,25,33,44,66,100,133] notation is also
2034 * [no]ncq: Turn on or off NCQ.
2036 * [no]ncqtrim: Turn off queued DSM TRIM.
2038 * nohrst, nosrst, norst: suppress hard, soft
2041 * rstonce: only attempt one reset during
2042 hot-unplug link recovery
2044 * dump_id: dump IDENTIFY data.
2046 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2048 * disable: Disable this device.
2050 If there are multiple matching configurations changing
2051 the same attribute, the last one is used.
2053 memblock=debug [KNL] Enable memblock debug messages.
2055 load_ramdisk= [RAM] List of ramdisks to load from floppy
2056 See Documentation/blockdev/ramdisk.txt.
2058 lockd.nlm_grace_period=P [NFS] Assign grace period.
2061 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2064 lockd.nlm_timeout=T [NFS] Assign timeout value.
2067 lockd.nlm_udpport=M [NFS] Assign UDP port.
2070 locktorture.nreaders_stress= [KNL]
2071 Set the number of locking read-acquisition kthreads.
2072 Defaults to being automatically set based on the
2073 number of online CPUs.
2075 locktorture.nwriters_stress= [KNL]
2076 Set the number of locking write-acquisition kthreads.
2078 locktorture.onoff_holdoff= [KNL]
2079 Set time (s) after boot for CPU-hotplug testing.
2081 locktorture.onoff_interval= [KNL]
2082 Set time (s) between CPU-hotplug operations, or
2083 zero to disable CPU-hotplug testing.
2085 locktorture.shuffle_interval= [KNL]
2086 Set task-shuffle interval (jiffies). Shuffling
2087 tasks allows some CPUs to go into dyntick-idle
2088 mode during the locktorture test.
2090 locktorture.shutdown_secs= [KNL]
2091 Set time (s) after boot system shutdown. This
2092 is useful for hands-off automated testing.
2094 locktorture.stat_interval= [KNL]
2095 Time (s) between statistics printk()s.
2097 locktorture.stutter= [KNL]
2098 Time (s) to stutter testing, for example,
2099 specifying five seconds causes the test to run for
2100 five seconds, wait for five seconds, and so on.
2101 This tests the locking primitive's ability to
2102 transition abruptly to and from idle.
2104 locktorture.torture_type= [KNL]
2105 Specify the locking implementation to test.
2107 locktorture.verbose= [KNL]
2108 Enable additional printk() statements.
2110 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2113 loglevel= All Kernel Messages with a loglevel smaller than the
2114 console loglevel will be printed to the console. It can
2115 also be changed with klogd or other programs. The
2116 loglevels are defined as follows:
2118 0 (KERN_EMERG) system is unusable
2119 1 (KERN_ALERT) action must be taken immediately
2120 2 (KERN_CRIT) critical conditions
2121 3 (KERN_ERR) error conditions
2122 4 (KERN_WARNING) warning conditions
2123 5 (KERN_NOTICE) normal but significant condition
2124 6 (KERN_INFO) informational
2125 7 (KERN_DEBUG) debug-level messages
2127 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2128 in bytes. n must be a power of two and greater
2129 than the minimal size. The minimal size is defined
2130 by LOG_BUF_SHIFT kernel config parameter. There is
2131 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2132 that allows to increase the default size depending on
2133 the number of CPUs. See init/Kconfig for more details.
2135 logo.nologo [FB] Disables display of the built-in Linux logo.
2136 This may be used to provide more screen space for
2137 kernel log messages and is useful when debugging
2138 kernel boot problems.
2140 lp=0 [LP] Specify parallel ports to use, e.g,
2141 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2142 lp=reset first parallel port). 'lp=0' disables the
2143 lp=auto printer driver. 'lp=reset' (which can be
2144 specified in addition to the ports) causes
2145 attached printers to be reset. Using
2146 lp=port1,port2,... specifies the parallel ports
2147 to associate lp devices with, starting with
2148 lp0. A port specification may be 'none' to skip
2149 that lp device, or a parport name such as
2150 'parport0'. Specifying 'lp=auto' instead of a
2151 port specification list means that device IDs
2152 from each port should be examined, to see if
2153 an IEEE 1284-compliant printer is attached; if
2154 so, the driver will manage that printer.
2155 See also header of drivers/char/lp.c.
2158 Sets loops_per_jiffy to given constant, thus avoiding
2159 time-consuming boot-time autodetection (up to 250 ms per
2160 CPU). 0 enables autodetection (default). To determine
2161 the correct value for your kernel, boot with normal
2162 autodetection and see what value is printed. Note that
2163 on SMP systems the preset will be applied to all CPUs,
2164 which is likely to cause problems if your CPUs need
2165 significantly divergent settings. An incorrect value
2166 will cause delays in the kernel to be wrong, leading to
2167 unpredictable I/O errors and other breakage. Although
2168 unlikely, in the extreme case this might damage your
2172 Format: <io>,<irq>,<dma>
2174 machvec= [IA-64] Force the use of a particular machine-vector
2175 (machvec) in a generic kernel.
2176 Example: machvec=hpzx1_swiotlb
2178 machtype= [Loongson] Share the same kernel image file between different
2180 Example: machtype=lemote-yeeloong-2f-7inch
2182 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2183 than or equal to this physical address is ignored.
2185 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2186 will bring up during bootup. maxcpus=n : n >= 0 limits
2187 the kernel to bring up 'n' processors. Surely after
2188 bootup you can bring up the other plugged cpu by executing
2189 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2190 only takes effect during system bootup.
2191 While n=0 is a special case, it is equivalent to "nosmp",
2192 which also disables the IO APIC.
2194 max_loop= [LOOP] The number of loop block devices that get
2195 (loop.max_loop) unconditionally pre-created at init time. The default
2196 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2197 of statically allocating a predefined number, loop
2198 devices can be requested on-demand with the
2199 /dev/loop-control interface.
2201 mce [X86-32] Machine Check Exception
2203 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2205 md= [HW] RAID subsystems devices and level
2206 See Documentation/admin-guide/md.rst.
2209 Format: <first>,<last>
2210 Specifies range of consoles to be captured by the MDA.
2212 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2213 Amount of memory to be used when the kernel is not able
2214 to see the whole system memory or for test.
2215 [X86] Work as limiting max address. Use together
2216 with memmap= to avoid physical address space collisions.
2217 Without memmap= PCI devices could be placed at addresses
2218 belonging to unused RAM.
2220 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2224 [KNL,SH] Allow user to override the default size for
2225 per-device physically contiguous DMA buffers.
2227 memhp_default_state=online/offline
2228 [KNL] Set the initial state for the memory hotplug
2229 onlining policy. If not specified, the default value is
2230 set according to the
2231 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2233 See Documentation/memory-hotplug.txt.
2235 memmap=exactmap [KNL,X86] Enable setting of an exact
2236 E820 memory map, as specified by the user.
2237 Such memmap=exactmap lines can be constructed based on
2238 BIOS output or other requirements. See the memmap=nn@ss
2241 memmap=nn[KMG]@ss[KMG]
2242 [KNL] Force usage of a specific region of memory.
2243 Region of memory to be used is from ss to ss+nn.
2244 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2245 which limits max address to nn[KMG].
2246 Multiple different regions can be specified,
2249 memmap=100M@2G,100M#3G,1G!1024G
2251 memmap=nn[KMG]#ss[KMG]
2252 [KNL,ACPI] Mark specific memory as ACPI data.
2253 Region of memory to be marked is from ss to ss+nn.
2255 memmap=nn[KMG]$ss[KMG]
2256 [KNL,ACPI] Mark specific memory as reserved.
2257 Region of memory to be reserved is from ss to ss+nn.
2258 Example: Exclude memory from 0x18690000-0x1869ffff
2259 memmap=64K$0x18690000
2261 memmap=0x10000$0x18690000
2262 Some bootloaders may need an escape character before '$',
2263 like Grub2, otherwise '$' and the following number
2266 memmap=nn[KMG]!ss[KMG]
2267 [KNL,X86] Mark specific memory as protected.
2268 Region of memory to be used, from ss to ss+nn.
2269 The memory region may be marked as e820 type 12 (0xc)
2270 and is NVDIMM or ADR memory.
2272 memmap=<size>%<offset>-<oldtype>+<newtype>
2273 [KNL,ACPI] Convert memory within the specified region
2274 from <oldtype> to <newtype>. If "-<oldtype>" is left
2275 out, the whole region will be marked as <newtype>,
2276 even if previously unavailable. If "+<newtype>" is left
2277 out, matching memory will be removed. Types are
2278 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2279 3 = ACPI, 12 = PRAM.
2281 memory_corruption_check=0/1 [X86]
2282 Some BIOSes seem to corrupt the first 64k of
2283 memory when doing things like suspend/resume.
2284 Setting this option will scan the memory
2285 looking for corruption. Enabling this will
2286 both detect corruption and prevent the kernel
2287 from using the memory being corrupted.
2288 However, its intended as a diagnostic tool; if
2289 repeatable BIOS-originated corruption always
2290 affects the same memory, you can use memmap=
2291 to prevent the kernel from using that memory.
2293 memory_corruption_check_size=size [X86]
2294 By default it checks for corruption in the low
2295 64k, making this memory unavailable for normal
2296 use. Use this parameter to scan for
2297 corruption in more or less memory.
2299 memory_corruption_check_period=seconds [X86]
2300 By default it checks for corruption every 60
2301 seconds. Use this parameter to check at some
2302 other rate. 0 disables periodic checking.
2304 memtest= [KNL,X86,ARM] Enable memtest
2306 default : 0 <disable>
2307 Specifies the number of memtest passes to be
2308 performed. Each pass selects another test
2309 pattern from a given set of patterns. Memtest
2310 fills the memory with this pattern, validates
2311 memory contents and reserves bad memory
2312 regions that are detected.
2314 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2315 Valid arguments: on, off
2316 Default (depends on kernel configuration option):
2317 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2318 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2319 mem_encrypt=on: Activate SME
2320 mem_encrypt=off: Do not activate SME
2322 Refer to Documentation/x86/amd-memory-encryption.txt
2323 for details on when memory encryption can be activated.
2325 mem_sleep_default= [SUSPEND] Default system suspend mode:
2326 s2idle - Suspend-To-Idle
2327 shallow - Power-On Suspend or equivalent (if supported)
2328 deep - Suspend-To-RAM or equivalent (if supported)
2329 See Documentation/admin-guide/pm/sleep-states.rst.
2331 meye.*= [HW] Set MotionEye Camera parameters
2332 See Documentation/media/v4l-drivers/meye.rst.
2334 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2335 Multi-Function General Purpose Timers on AMD Geode
2338 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2339 the BIOS has incorrectly applied a workaround. TinyBIOS
2340 version 0.98 is known to be affected, 0.99 fixes the
2341 problem by letting the user disable the workaround.
2345 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2346 physical address is ignored.
2348 mini2440= [ARM,HW,KNL]
2349 Format:[0..2][b][c][t]
2351 MINI2440 configuration specification:
2352 0 - The attached screen is the 3.5" TFT
2353 1 - The attached screen is the 7" TFT
2354 2 - The VGA Shield is attached (1024x768)
2355 Leaving out the screen size parameter will not load
2356 the TFT driver, and the framebuffer will be left
2358 b - Enable backlight. The TFT backlight pin will be
2359 linked to the kernel VESA blanking code and a GPIO
2360 LED. This parameter is not necessary when using the
2362 c - Enable the s3c camera interface.
2363 t - Reserved for enabling touchscreen support. The
2364 touchscreen support is not enabled in the mainstream
2365 kernel as of 2.6.30, a preliminary port can be found
2366 in the "bleeding edge" mini2440 support kernel at
2367 http://repo.or.cz/w/linux-2.6/mini2440.git
2370 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2371 parameter allows control of the logging verbosity for
2372 the additional memory initialisation checks. A value
2373 of 0 disables mminit logging and a level of 4 will
2374 log everything. Information is printed at KERN_DEBUG
2375 so loglevel=8 may also need to be specified.
2378 [KNL] When CONFIG_MODULE_SIG is set, this means that
2379 modules without (valid) signatures will fail to load.
2380 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2381 is always true, so this option does nothing.
2383 module_blacklist= [KNL] Do not load a comma-separated list of
2384 modules. Useful for debugging problem modules.
2387 [MOUSE] Maximum time between finger touching and
2388 leaving touchpad surface for touch to be considered
2389 a tap and be reported as a left button click (for
2390 touchpads working in absolute mode only).
2392 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2393 reporting absolute coordinates, such as tablets
2394 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2395 reporting absolute coordinates, such as tablets
2397 movablecore= [KNL,X86,IA-64,PPC]
2398 Format: nn[KMGTPE] | nn%
2399 This parameter is the complement to kernelcore=, it
2400 specifies the amount of memory used for migratable
2401 allocations. If both kernelcore and movablecore is
2402 specified, then kernelcore will be at *least* the
2403 specified value but may be more. If movablecore on its
2404 own is specified, the administrator must be careful
2405 that the amount of memory usable for all allocations
2408 movable_node [KNL] Boot-time switch to make hotplugable memory
2409 NUMA nodes to be movable. This means that the memory
2410 of such nodes will be usable only for movable
2411 allocations which rules out almost all kernel
2412 allocations. Use with caution!
2414 MTD_Partition= [MTD]
2415 Format: <name>,<region-number>,<size>,<offset>
2417 MTD_Region= [MTD] Format:
2418 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2421 See drivers/mtd/cmdlinepart.c.
2423 multitce=off [PPC] This parameter disables the use of the pSeries
2424 firmware feature for updating multiple TCE entries
2427 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2429 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2431 boundary - index of last SLC block on Flex-OneNAND.
2432 The remaining blocks are configured as MLC blocks.
2433 lock - Configure if Flex-OneNAND boundary should be locked.
2434 Once locked, the boundary cannot be changed.
2435 1 indicates lock status, 0 indicates unlock status.
2438 ARM/S3C2412 JIVE boot control
2440 See arch/arm/mach-s3c2412/mach-jive.c
2442 mtouchusb.raw_coordinates=
2443 [HW] Make the MicroTouch USB driver use raw coordinates
2444 ('y', default) or cooked coordinates ('n')
2446 mtrr_chunk_size=nn[KMG] [X86]
2447 used for mtrr cleanup. It is largest continuous chunk
2448 that could hold holes aka. UC entries.
2450 mtrr_gran_size=nn[KMG] [X86]
2451 Used for mtrr cleanup. It is granularity of mtrr block.
2453 Large value could prevent small alignment from
2456 mtrr_spare_reg_nr=n [X86]
2458 Range: 0,7 : spare reg number
2460 Used for mtrr cleanup. It is spare mtrr entries number.
2461 Set to 2 or more if your graphical card needs more.
2463 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2465 netdev= [NET] Network devices parameters
2466 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2467 Note that mem_start is often overloaded to mean
2468 something different and driver-specific.
2469 This usage is only documented in each driver source
2473 [NETFILTER] Enable connection tracking flow accounting
2474 0 to disable accounting
2475 1 to enable accounting
2478 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2479 See Documentation/filesystems/nfs/nfsroot.txt.
2481 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2482 See Documentation/filesystems/nfs/nfsroot.txt.
2484 nfsrootdebug [NFS] enable nfsroot debugging messages.
2485 See Documentation/filesystems/nfs/nfsroot.txt.
2487 nfs.callback_nr_threads=
2488 [NFSv4] set the total number of threads that the
2489 NFS client will assign to service NFSv4 callback
2492 nfs.callback_tcpport=
2493 [NFS] set the TCP port on which the NFSv4 callback
2494 channel should listen.
2497 [NFS] sets the pathname to the program which is used
2498 to update the NFS client cache entries.
2500 nfs.cache_getent_timeout=
2501 [NFS] sets the timeout after which an attempt to
2502 update a cache entry is deemed to have failed.
2504 nfs.idmap_cache_timeout=
2505 [NFS] set the maximum lifetime for idmapper cache
2509 [NFS] enable 64-bit inode numbers.
2510 If zero, the NFS client will fake up a 32-bit inode
2511 number for the readdir() and stat() syscalls instead
2512 of returning the full 64-bit number.
2513 The default is to return 64-bit inode numbers.
2515 nfs.max_session_cb_slots=
2516 [NFSv4.1] Sets the maximum number of session
2517 slots the client will assign to the callback
2518 channel. This determines the maximum number of
2519 callbacks the client will process in parallel for
2520 a particular server.
2522 nfs.max_session_slots=
2523 [NFSv4.1] Sets the maximum number of session slots
2524 the client will attempt to negotiate with the server.
2525 This limits the number of simultaneous RPC requests
2526 that the client can send to the NFSv4.1 server.
2527 Note that there is little point in setting this
2528 value higher than the max_tcp_slot_table_limit.
2530 nfs.nfs4_disable_idmapping=
2531 [NFSv4] When set to the default of '1', this option
2532 ensures that both the RPC level authentication
2533 scheme and the NFS level operations agree to use
2534 numeric uids/gids if the mount is using the
2535 'sec=sys' security flavour. In effect it is
2536 disabling idmapping, which can make migration from
2537 legacy NFSv2/v3 systems to NFSv4 easier.
2538 Servers that do not support this mode of operation
2539 will be autodetected by the client, and it will fall
2540 back to using the idmapper.
2541 To turn off this behaviour, set the value to '0'.
2543 [NFS4] Specify an additional fixed unique ident-
2544 ification string that NFSv4 clients can insert into
2545 their nfs_client_id4 string. This is typically a
2546 UUID that is generated at system install time.
2548 nfs.send_implementation_id =
2549 [NFSv4.1] Send client implementation identification
2550 information in exchange_id requests.
2551 If zero, no implementation identification information
2553 The default is to send the implementation identification
2556 nfs.recover_lost_locks =
2557 [NFSv4] Attempt to recover locks that were lost due
2558 to a lease timeout on the server. Please note that
2559 doing this risks data corruption, since there are
2560 no guarantees that the file will remain unchanged
2561 after the locks are lost.
2562 If you want to enable the kernel legacy behaviour of
2563 attempting to recover these locks, then set this
2565 The default parameter value of '0' causes the kernel
2566 not to attempt recovery of lost locks.
2568 nfs4.layoutstats_timer =
2569 [NFSv4.2] Change the rate at which the kernel sends
2570 layoutstats to the pNFS metadata server.
2572 Setting this to value to 0 causes the kernel to use
2573 whatever value is the default set by the layout
2574 driver. A non-zero value sets the minimum interval
2575 in seconds between layoutstats transmissions.
2577 nfsd.nfs4_disable_idmapping=
2578 [NFSv4] When set to the default of '1', the NFSv4
2579 server will return only numeric uids and gids to
2580 clients using auth_sys, and will accept numeric uids
2581 and gids from such clients. This is intended to ease
2582 migration from NFSv2/v3.
2584 nmi_debug= [KNL,SH] Specify one or more actions to take
2585 when a NMI is triggered.
2586 Format: [state][,regs][,debounce][,die]
2588 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2589 Format: [panic,][nopanic,][num]
2591 0 - turn hardlockup detector in nmi_watchdog off
2592 1 - turn hardlockup detector in nmi_watchdog on
2593 When panic is specified, panic when an NMI watchdog
2594 timeout occurs (or 'nopanic' to override the opposite
2595 default). To disable both hard and soft lockup detectors,
2596 please see 'nowatchdog'.
2597 This is useful when you use a panic=... timeout and
2598 need the box quickly up again.
2600 These settings can be accessed at runtime via
2601 the nmi_watchdog and hardlockup_panic sysctls.
2603 netpoll.carrier_timeout=
2604 [NET] Specifies amount of time (in seconds) that
2605 netpoll should wait for a carrier. By default netpoll
2608 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2609 emulation library even if a 387 maths coprocessor
2612 no5lvl [X86-64] Disable 5-level paging mode. Forces
2613 kernel to use 4-level paging instead.
2616 [HW] Never suspend the console
2617 Disable suspending of consoles during suspend and
2618 hibernate operations. Once disabled, debugging
2619 messages can reach various consoles while the rest
2620 of the system is being put to sleep (ie, while
2621 debugging driver suspend/resume hooks). This may
2622 not work reliably with all consoles, but is known
2623 to work with serial and VGA consoles.
2624 To facilitate more flexible debugging, we also add
2625 console_suspend, a printk module parameter to control
2626 it. Users could use console_suspend (usually
2627 /sys/module/printk/parameters/console_suspend) to
2628 turn on/off it dynamically.
2630 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2631 caches in the slab allocator. Saves per-node memory,
2632 but will impact performance.
2636 noaltinstr [S390] Disables alternative instructions patching
2637 (CPU alternatives feature).
2639 noapic [SMP,APIC] Tells the kernel to not make use of any
2640 IOAPICs that may be present in the system.
2642 noautogroup Disable scheduler automatic task group creation.
2644 nobats [PPC] Do not use BATs for mapping kernel lowmem
2645 on "Classic" PPC cores.
2649 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2651 nodelayacct [KNL] Disable per-task delay accounting
2653 nodsp [SH] Disable hardware DSP at boot time.
2655 noefi Disable EFI runtime services support.
2660 On X86-32 available only on PAE configured kernels.
2661 noexec=on: enable non-executable mappings (default)
2662 noexec=off: disable non-executable mappings
2665 Disable SMAP (Supervisor Mode Access Prevention)
2666 even if it is supported by processor.
2669 Disable SMEP (Supervisor Mode Execution Prevention)
2670 even if it is supported by processor.
2673 This affects only 32-bit executables.
2674 noexec32=on: enable non-executable mappings (default)
2675 read doesn't imply executable mappings
2676 noexec32=off: disable non-executable mappings
2677 read implies executable mappings
2679 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2681 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2682 register save and restore. The kernel will only save
2683 legacy floating-point registers on task switch.
2685 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2687 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2688 Equivalent to smt=1.
2690 nospectre_v2 [X86] Disable all mitigations for the Spectre variant 2
2691 (indirect branch prediction) vulnerability. System may
2692 allow data leaks with this option, which is equivalent
2695 nospec_store_bypass_disable
2696 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
2698 noxsave [BUGS=X86] Disables x86 extended register state save
2699 and restore using xsave. The kernel will fallback to
2700 enabling legacy floating-point and sse state.
2702 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2703 register states. The kernel will fall back to use
2704 xsave to save the states. By using this parameter,
2705 performance of saving the states is degraded because
2706 xsave doesn't support modified optimization while
2707 xsaveopt supports it on xsaveopt enabled systems.
2709 noxsaves [X86] Disables xsaves and xrstors used in saving and
2710 restoring x86 extended register state in compacted
2711 form of xsave area. The kernel will fall back to use
2712 xsaveopt and xrstor to save and restore the states
2713 in standard form of xsave area. By using this
2714 parameter, xsave area per process might occupy more
2715 memory on xsaves enabled systems.
2717 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2718 wfi(ARM) instruction doesn't work correctly and not to
2719 use it. This is also useful when using JTAG debugger.
2721 no_file_caps Tells the kernel not to honor file capabilities. The
2722 only way then for a file to be executed with privilege
2723 is to be setuid root or executed by root.
2725 nohalt [IA-64] Tells the kernel not to use the power saving
2726 function PAL_HALT_LIGHT when idle. This increases
2727 power-consumption. On the positive side, it reduces
2728 interrupt wake-up latency, which may improve performance
2729 in certain environments such as networked servers or
2732 nohibernate [HIBERNATION] Disable hibernation and resume.
2734 nohz= [KNL] Boottime enable/disable dynamic ticks
2735 Valid arguments: on, off
2738 nohz_full= [KNL,BOOT,SMP,ISOL]
2739 The argument is a cpu list, as described above.
2740 In kernels built with CONFIG_NO_HZ_FULL=y, set
2741 the specified list of CPUs whose tick will be stopped
2742 whenever possible. The boot CPU will be forced outside
2743 the range to maintain the timekeeping. Any CPUs
2744 in this list will have their RCU callbacks offloaded,
2745 just as if they had also been called out in the
2746 rcu_nocbs= boot parameter.
2748 noiotrap [SH] Disables trapped I/O port accesses.
2750 noirqdebug [X86-32] Disables the code which attempts to detect and
2751 disable unhandled interrupt sources.
2753 no_timer_check [X86,APIC] Disables the code which tests for
2754 broken timer IRQ sources.
2756 noisapnp [ISAPNP] Disables ISA PnP code.
2758 noinitrd [RAM] Tells the kernel not to load any configured
2761 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2763 [Deprecated - use intremap=off]
2767 noinvpcid [X86] Disable the INVPCID cpu feature.
2769 nojitter [IA-64] Disables jitter checking for ITC timers.
2771 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2773 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2777 [X86,PV_OPS] Disable paravirtualized VMware scheduler
2778 clock and use the default one.
2780 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
2781 steal time is computed, but won't influence scheduler
2784 nolapic [X86-32,APIC] Do not enable or use the local APIC.
2786 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
2788 noltlbs [PPC] Do not use large page/tlb entries for kernel
2789 lowmem mapping on PPC40x and PPC8xx
2791 nomca [IA-64] Disable machine check abort handling
2793 nomce [X86-32] Disable Machine Check Exception
2795 nomfgpt [X86-32] Disable Multi-Function General Purpose
2796 Timer usage (for AMD Geode machines).
2798 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
2799 shutdown the other cpus. Instead use the REBOOT_VECTOR
2802 nomodule Disable module load
2804 nopat [X86] Disable PAT (page attribute table extension of
2805 pagetables) support.
2807 nopcid [X86-64] Disable the PCID cpu feature.
2809 norandmaps Don't use address space randomization. Equivalent to
2810 echo 0 > /proc/sys/kernel/randomize_va_space
2812 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
2813 with UP alternatives
2815 nordrand [X86] Disable kernel use of the RDRAND and
2816 RDSEED instructions even if they are supported
2817 by the processor. RDRAND and RDSEED are still
2818 available to user space applications.
2820 noresume [SWSUSP] Disables resume and restores original swap
2823 no-scroll [VGA] Disables scrollback.
2824 This is required for the Braillex ib80-piezo Braille
2825 reader made by F.H. Papenmeier (Germany).
2829 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
2831 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
2832 and disable the IO APIC. legacy for "maxcpus=0".
2834 nosoftlockup [KNL] Disable the soft-lockup detector.
2836 nosync [HW,M68K] Disables sync negotiation for all devices.
2838 notsc [BUGS=X86-32] Disable Time Stamp Counter
2840 nowatchdog [KNL] Disable both lockup detectors, i.e.
2841 soft-lockup and NMI watchdog (hard-lockup).
2845 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
2847 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
2848 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
2849 Some features depend on CPU0. Known dependencies are:
2850 1. Resume from suspend/hibernate depends on CPU0.
2851 Suspend/hibernate will fail if CPU0 is offline and you
2852 need to online CPU0 before suspend/hibernate.
2853 2. PIC interrupts also depend on CPU0. CPU0 can't be
2854 removed if a PIC interrupt is detected.
2855 It's said poweroff/reboot may depend on CPU0 on some
2856 machines although I haven't seen such issues so far
2857 after CPU0 is offline on a few tested machines.
2858 If the dependencies are under your control, you can
2859 turn on cpu0_hotplug.
2861 nps_mtm_hs_ctr= [KNL,ARC]
2862 This parameter sets the maximum duration, in
2863 cycles, each HW thread of the CTOP can run
2864 without interruptions, before HW switches it.
2865 The actual maximum duration is 16 times this
2867 Format: integer between 1 and 255
2870 nptcg= [IA-64] Override max number of concurrent global TLB
2871 purges which is reported from either PAL_VM_SUMMARY or
2874 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
2875 could support. nr_cpus=n : n >= 1 limits the kernel to
2876 support 'n' processors. It could be larger than the
2877 number of already plugged CPU during bootup, later in
2878 runtime you can physically add extra cpu until it reaches
2879 n. So during boot up some boot time memory for per-cpu
2880 variables need be pre-allocated for later physical cpu
2883 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
2885 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
2886 Allowed values are enable and disable
2888 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
2889 'node', 'default' can be specified
2890 This can be set from sysctl after boot.
2891 See Documentation/sysctl/vm.txt for details.
2893 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
2894 See Documentation/debugging-via-ohci1394.txt for more
2897 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
2898 Rather than timing out after 20 ms if an EC
2899 command is not properly ACKed, override the length
2900 of the timeout. We have interrupts disabled while
2901 waiting for the ACK, so if this is set too high
2902 interrupts *may* be lost!
2904 omap_mux= [OMAP] Override bootloader pin multiplexing.
2905 Format: <mux_mode0.mode_name=value>...
2906 For example, to override I2C bus2:
2907 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
2909 oprofile.timer= [HW]
2910 Use timer interrupt instead of performance counters
2912 oprofile.cpu_type= Force an oprofile cpu type
2913 This might be useful if you have an older oprofile
2914 userland or if you want common events.
2915 Format: { arch_perfmon }
2916 arch_perfmon: [X86] Force use of architectural
2917 perfmon on Intel CPUs instead of the
2918 CPU specific event set.
2919 timer: [X86] Force use of architectural NMI
2920 timer mode (see also oprofile.timer
2921 for generic hr timer mode)
2923 oops=panic Always panic on oopses. Default is to just kill the
2924 process, but there is a small probability of
2925 deadlocking the machine.
2926 This will also cause panics on machine check exceptions.
2927 Useful together with panic=30 to trigger a reboot.
2930 See Documentation/sound/oss/oss-parameters.txt
2932 page_owner= [KNL] Boot-time page_owner enabling option.
2933 Storage of the information about who allocated
2934 each page is disabled in default. With this switch,
2936 on: enable the feature
2938 page_poison= [KNL] Boot-time parameter changing the state of
2939 poisoning on the buddy allocator.
2940 off: turn off poisoning
2941 on: turn on poisoning
2943 panic= [KNL] Kernel behaviour on panic: delay <timeout>
2944 timeout > 0: seconds before rebooting
2945 timeout = 0: wait forever
2946 timeout < 0: reboot immediately
2949 panic_on_warn panic() instead of WARN(). Useful to cause kdump
2952 crash_kexec_post_notifiers
2953 Run kdump after running panic-notifiers and dumping
2954 kmsg. This only for the users who doubt kdump always
2955 succeeds in any situation.
2956 Note that this also increases risks of kdump failure,
2957 because some panic notifiers can make the crashed
2958 kernel more unstable.
2960 parkbd.port= [HW] Parallel port number the keyboard adapter is
2961 connected to, default is 0.
2963 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
2964 0 for XT, 1 for AT (default is AT).
2967 parport= [HW,PPT] Specify parallel ports. 0 disables.
2968 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
2969 Use 'auto' to force the driver to use any
2970 IRQ/DMA settings detected (the default is to
2971 ignore detected IRQ/DMA settings because of
2972 possible conflicts). You can specify the base
2973 address, IRQ, and DMA settings; IRQ and DMA
2974 should be numbers, or 'auto' (for using detected
2975 settings on that particular port), or 'nofifo'
2976 (to avoid using a FIFO even if it is detected).
2977 Parallel ports are assigned in the order they
2978 are specified on the command line, starting
2981 parport_init_mode= [HW,PPT]
2982 Configure VIA parallel port to operate in
2983 a specific mode. This is necessary on Pegasos
2984 computer where firmware has no options for setting
2985 up parallel port mode and sets it to spp.
2986 Currently this function knows 686a and 8231 chips.
2987 Format: [spp|ps2|epp|ecp|ecpepp]
2990 Halt all CPUs after the first oops has been printed for
2991 the specified number of seconds. This is to be used if
2992 your oopses keep scrolling off the screen.
2997 See header of drivers/block/paride/pcd.c.
2998 See also Documentation/blockdev/paride.txt.
3000 pci=option[,option...] [PCI] various PCI subsystem options:
3001 earlydump [X86] dump PCI config space before the kernel
3003 off [X86] don't probe for the PCI bus
3004 bios [X86-32] force use of PCI BIOS, don't access
3005 the hardware directly. Use this if your machine
3006 has a non-standard PCI host bridge.
3007 nobios [X86-32] disallow use of PCI BIOS, only direct
3008 hardware access methods are allowed. Use this
3009 if you experience crashes upon bootup and you
3010 suspect they are caused by the BIOS.
3011 conf1 [X86] Force use of PCI Configuration Access
3012 Mechanism 1 (config address in IO port 0xCF8,
3013 data in IO port 0xCFC, both 32-bit).
3014 conf2 [X86] Force use of PCI Configuration Access
3015 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3016 the function, IO port 0xCFA, also 8-bit, sets
3017 bus number. The config space is then accessed
3018 through ports 0xC000-0xCFFF).
3019 See http://wiki.osdev.org/PCI for more info
3020 on the configuration access mechanisms.
3021 noaer [PCIE] If the PCIEAER kernel config parameter is
3022 enabled, this kernel boot option can be used to
3023 disable the use of PCIE advanced error reporting.
3024 nodomains [PCI] Disable support for multiple PCI
3025 root domains (aka PCI segments, in ACPI-speak).
3026 nommconf [X86] Disable use of MMCONFIG for PCI
3028 check_enable_amd_mmconf [X86] check for and enable
3029 properly configured MMIO access to PCI
3030 config space on AMD family 10h CPU
3031 nomsi [MSI] If the PCI_MSI kernel config parameter is
3032 enabled, this kernel boot option can be used to
3033 disable the use of MSI interrupts system-wide.
3034 noioapicquirk [APIC] Disable all boot interrupt quirks.
3035 Safety option to keep boot IRQs enabled. This
3036 should never be necessary.
3037 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3038 primary IO-APIC for bridges that cannot disable
3039 boot IRQs. This fixes a source of spurious IRQs
3040 when the system masks IRQs.
3041 noioapicreroute [APIC] Disable workaround that uses the
3042 boot IRQ equivalent of an IRQ that connects to
3043 a chipset where boot IRQs cannot be disabled.
3044 The opposite of ioapicreroute.
3045 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3046 routing table. These calls are known to be buggy
3047 on several machines and they hang the machine
3048 when used, but on other computers it's the only
3049 way to get the interrupt routing table. Try
3050 this option if the kernel is unable to allocate
3051 IRQs or discover secondary PCI buses on your
3053 rom [X86] Assign address space to expansion ROMs.
3054 Use with caution as certain devices share
3055 address decoders between ROMs and other
3057 norom [X86] Do not assign address space to
3058 expansion ROMs that do not already have
3059 BIOS assigned address ranges.
3060 nobar [X86] Do not assign address space to the
3061 BARs that weren't assigned by the BIOS.
3062 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3063 assigned automatically to PCI devices. You can
3064 make the kernel exclude IRQs of your ISA cards
3066 pirqaddr=0xAAAAA [X86] Specify the physical address
3067 of the PIRQ table (normally generated
3068 by the BIOS) if it is outside the
3069 F0000h-100000h range.
3070 lastbus=N [X86] Scan all buses thru bus #N. Can be
3071 useful if the kernel is unable to find your
3072 secondary buses and you want to tell it
3073 explicitly which ones they are.
3074 assign-busses [X86] Always assign all PCI bus
3075 numbers ourselves, overriding
3076 whatever the firmware may have done.
3077 usepirqmask [X86] Honor the possible IRQ mask stored
3078 in the BIOS $PIR table. This is needed on
3079 some systems with broken BIOSes, notably
3080 some HP Pavilion N5400 and Omnibook XE3
3081 notebooks. This will have no effect if ACPI
3082 IRQ routing is enabled.
3083 noacpi [X86] Do not use ACPI for IRQ routing
3084 or for PCI scanning.
3085 use_crs [X86] Use PCI host bridge window information
3086 from ACPI. On BIOSes from 2008 or later, this
3087 is enabled by default. If you need to use this,
3088 please report a bug.
3089 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3090 If you need to use this, please report a bug.
3091 routeirq Do IRQ routing for all PCI devices.
3092 This is normally done in pci_enable_device(),
3093 so this option is a temporary workaround
3094 for broken drivers that don't call it.
3095 skip_isa_align [X86] do not align io start addr, so can
3096 handle more pci cards
3097 noearly [X86] Don't do any early type 1 scanning.
3098 This might help on some broken boards which
3099 machine check when some devices' config space
3100 is read. But various workarounds are disabled
3101 and some IOMMU drivers will not work.
3102 bfsort Sort PCI devices into breadth-first order.
3103 This sorting is done to get a device
3104 order compatible with older (<= 2.4) kernels.
3105 nobfsort Don't sort PCI devices into breadth-first order.
3106 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3107 tuning and use the BIOS-configured MPS defaults.
3108 pcie_bus_safe Set every device's MPS to the largest value
3109 supported by all devices below the root complex.
3110 pcie_bus_perf Set device MPS to the largest allowable MPS
3111 based on its parent bus. Also set MRRS (Max
3112 Read Request Size) to the largest supported
3113 value (no larger than the MPS that the device
3114 or bus can support) for best performance.
3115 pcie_bus_peer2peer Set every device's MPS to 128B, which
3116 every device is guaranteed to support. This
3117 configuration allows peer-to-peer DMA between
3118 any pair of devices, possibly at the cost of
3119 reduced performance. This also guarantees
3120 that hot-added devices will work.
3121 cbiosize=nn[KMG] The fixed amount of bus space which is
3122 reserved for the CardBus bridge's IO window.
3123 The default value is 256 bytes.
3124 cbmemsize=nn[KMG] The fixed amount of bus space which is
3125 reserved for the CardBus bridge's memory
3126 window. The default value is 64 megabytes.
3129 [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
3130 [<order of align>@]pci:<vendor>:<device>\
3131 [:<subvendor>:<subdevice>][; ...]
3132 Specifies alignment and device to reassign
3133 aligned memory resources.
3134 If <order of align> is not specified,
3135 PAGE_SIZE is used as alignment.
3136 PCI-PCI bridge can be specified, if resource
3137 windows need to be expanded.
3138 To specify the alignment for several
3139 instances of a device, the PCI vendor,
3140 device, subvendor, and subdevice may be
3141 specified, e.g., 4096@pci:8086:9c22:103c:198f
3142 ecrc= Enable/disable PCIe ECRC (transaction layer
3143 end-to-end CRC checking).
3144 bios: Use BIOS/firmware settings. This is the
3148 hpiosize=nn[KMG] The fixed amount of bus space which is
3149 reserved for hotplug bridge's IO window.
3150 Default size is 256 bytes.
3151 hpmemsize=nn[KMG] The fixed amount of bus space which is
3152 reserved for hotplug bridge's memory window.
3153 Default size is 2 megabytes.
3154 hpbussize=nn The minimum amount of additional bus numbers
3155 reserved for buses below a hotplug bridge.
3157 realloc= Enable/disable reallocating PCI bridge resources
3158 if allocations done by BIOS are too small to
3159 accommodate resources required by all child
3161 off: Turn realloc off
3163 realloc same as realloc=on
3164 noari do not use PCIe ARI.
3165 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3166 do not use PCIe ATS (and IOMMU device IOTLB).
3167 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3168 only look for one device below a PCIe downstream
3170 big_root_window Try to add a big 64bit memory window to the PCIe
3171 root complex on AMD CPUs. Some GFX hardware
3172 can resize a BAR to allow access to all VRAM.
3173 Adding the window is slightly risky (it may
3174 conflict with unreported devices), so this
3177 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3180 force Enable ASPM even on devices that claim not to support it.
3181 WARNING: Forcing ASPM on may cause system lockups.
3183 pcie_ports= [PCIE] PCIe port services handling:
3184 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3185 even if the platform doesn't give the OS permission to
3186 use them. This may cause conflicts if the platform
3187 also tries to use these services.
3188 compat Disable native PCIe services (PME, AER, DPC, PCIe
3191 pcie_port_pm= [PCIE] PCIe port power management handling:
3192 off Disable power management of all PCIe ports
3193 force Forcibly enable power management of all PCIe ports
3195 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3196 nomsi Do not use MSI for native PCIe PME signaling (this makes
3197 all PCIe root ports use INTx for all services).
3199 pcmv= [HW,PCMCIA] BadgePAD 4
3203 Keep all power-domains already enabled by bootloader on,
3204 even if no driver has claimed them. This is useful
3205 for debug and development, but should not be
3206 needed on a platform with proper driver support.
3209 See Documentation/blockdev/paride.txt.
3211 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3214 See arch/parisc/kernel/pdc_chassis.c
3216 percpu_alloc= Select which percpu first chunk allocator to use.
3217 Currently supported values are "embed" and "page".
3218 Archs may support subset or none of the selections.
3219 See comments in mm/percpu.c for details on each
3220 allocator. This parameter is primarily for debugging
3221 and performance comparison.
3224 See Documentation/blockdev/paride.txt.
3227 See Documentation/blockdev/paride.txt.
3229 pirq= [SMP,APIC] Manual mp-table setup
3230 See Documentation/x86/i386/IO-APIC.txt.
3232 plip= [PPT,NET] Parallel port network link
3233 Format: { parport<nr> | timid | 0 }
3234 See also Documentation/admin-guide/parport.rst.
3236 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3237 Override pmtimer IOPort with a hex value.
3241 Enable PNP debug messages (depends on the
3242 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3243 via /sys/module/pnp/parameters/debug. We always show
3244 current resource usage; turning this on also shows
3245 possible settings and some assignment information.
3251 { on | off | curr | res | no-curr | no-res }
3254 [ISAPNP] Exclude IRQs for the autoconfiguration
3257 [ISAPNP] Exclude DMAs for the autoconfiguration
3259 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3260 Ranges are in pairs (I/O port base and size).
3263 [ISAPNP] Exclude memory regions for the
3265 Ranges are in pairs (memory base and size).
3267 ports= [IP_VS_FTP] IPVS ftp helper module
3269 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3271 Format: <port>,<port>....
3273 powersave=off [PPC] This option disables power saving features.
3274 It specifically disables cpuidle and sets the
3275 platform machine description specific power_save
3276 function to NULL. On Idle the CPU just reduces
3279 ppc_strict_facility_enable
3280 [PPC] This option catches any kernel floating point,
3281 Altivec, VSX and SPE outside of regions specifically
3282 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3283 There is some performance impact when enabling this.
3287 Disable Hardware Transactional Memory
3289 print-fatal-signals=
3290 [KNL] debug: print fatal signals
3292 If enabled, warn about various signal handling
3293 related application anomalies: too many signals,
3294 too many POSIX.1 timers, fatal signals causing a
3297 If you hit the warning due to signal overflow,
3298 you might want to try "ulimit -i unlimited".
3302 printk.always_kmsg_dump=
3303 Trigger kmsg_dump for cases other than kernel oops or
3305 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3308 printk.devkmsg={on,off,ratelimit}
3309 Control writing to /dev/kmsg.
3310 on - unlimited logging to /dev/kmsg from userspace
3311 off - logging to /dev/kmsg disabled
3312 ratelimit - ratelimit the logging
3315 printk.time= Show timing data prefixed to each printk message line
3316 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3318 processor.max_cstate= [HW,ACPI]
3319 Limit processor to maximum C-state
3320 max_cstate=9 overrides any DMI blacklist limit.
3322 processor.nocst [HW,ACPI]
3323 Ignore the _CST method to determine C-states,
3324 instead using the legacy FADT method
3326 profile= [KNL] Enable kernel profiling via /proc/profile
3327 Format: [<profiletype>,]<number>
3328 Param: <profiletype>: "schedule", "sleep", or "kvm"
3329 [defaults to kernel profiling]
3330 Param: "schedule" - profile schedule points.
3331 Param: "sleep" - profile D-state sleeping (millisecs).
3332 Requires CONFIG_SCHEDSTATS
3333 Param: "kvm" - profile VM exits.
3334 Param: <number> - step/bucket size as a power of 2 for
3335 statistical time based profiling.
3337 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3339 See Documentation/blockdev/ramdisk.txt.
3341 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3342 probe for; one of (bare|imps|exps|lifebook|any).
3343 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3345 psmouse.resetafter= [HW,MOUSE]
3346 Try to reset the device after so many bad packets
3349 [HW,MOUSE] Set desired mouse resolution, in dpi.
3350 psmouse.smartscroll=
3351 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3352 0 = disabled, 1 = enabled (default).
3354 pstore.backend= Specify the name of the pstore backend to use
3357 See Documentation/blockdev/paride.txt.
3359 pti= [X86_64] Control Page Table Isolation of user and
3360 kernel address spaces. Disabling this feature
3361 removes hardening, but improves performance of
3362 system calls and interrupts.
3364 on - unconditionally enable
3365 off - unconditionally disable
3366 auto - kernel detects whether your CPU model is
3367 vulnerable to issues that PTI mitigates
3369 Not specifying this option is equivalent to pti=auto.
3372 Equivalent to pti=off
3375 [KNL] Number of legacy pty's. Overwrites compiled-in
3378 quiet [KNL] Disable most log messages
3383 See Documentation/admin-guide/md.rst.
3385 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3386 See Documentation/blockdev/ramdisk.txt.
3388 ras=option[,option,...] [KNL] RAS-specific options
3391 Disable the Correctable Errors Collector,
3392 see CONFIG_RAS_CEC help text.
3395 The argument is a cpu list, as described above.
3397 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3398 the specified list of CPUs to be no-callback CPUs.
3399 Invocation of these CPUs' RCU callbacks will
3400 be offloaded to "rcuox/N" kthreads created for
3401 that purpose, where "x" is "b" for RCU-bh, "p"
3402 for RCU-preempt, and "s" for RCU-sched, and "N"
3403 is the CPU number. This reduces OS jitter on the
3404 offloaded CPUs, which can be useful for HPC and
3405 real-time workloads. It can also improve energy
3406 efficiency for asymmetric multiprocessors.
3409 Rather than requiring that offloaded CPUs
3410 (specified by rcu_nocbs= above) explicitly
3411 awaken the corresponding "rcuoN" kthreads,
3412 make these kthreads poll for callbacks.
3413 This improves the real-time response for the
3414 offloaded CPUs by relieving them of the need to
3415 wake up the corresponding kthread, but degrades
3416 energy efficiency by requiring that the kthreads
3417 periodically wake up to do the polling.
3419 rcutree.blimit= [KNL]
3420 Set maximum number of finished RCU callbacks to
3421 process in one batch.
3423 rcutree.dump_tree= [KNL]
3424 Dump the structure of the rcu_node combining tree
3425 out at early boot. This is used for diagnostic
3426 purposes, to verify correct tree setup.
3428 rcutree.gp_cleanup_delay= [KNL]
3429 Set the number of jiffies to delay each step of
3430 RCU grace-period cleanup.
3432 rcutree.gp_init_delay= [KNL]
3433 Set the number of jiffies to delay each step of
3434 RCU grace-period initialization.
3436 rcutree.gp_preinit_delay= [KNL]
3437 Set the number of jiffies to delay each step of
3438 RCU grace-period pre-initialization, that is,
3439 the propagation of recent CPU-hotplug changes up
3440 the rcu_node combining tree.
3442 rcutree.rcu_fanout_exact= [KNL]
3443 Disable autobalancing of the rcu_node combining
3444 tree. This is used by rcutorture, and might
3445 possibly be useful for architectures having high
3446 cache-to-cache transfer latencies.
3448 rcutree.rcu_fanout_leaf= [KNL]
3449 Change the number of CPUs assigned to each
3450 leaf rcu_node structure. Useful for very
3451 large systems, which will choose the value 64,
3452 and for NUMA systems with large remote-access
3453 latencies, which will choose a value aligned
3454 with the appropriate hardware boundaries.
3456 rcutree.jiffies_till_sched_qs= [KNL]
3457 Set required age in jiffies for a
3458 given grace period before RCU starts
3459 soliciting quiescent-state help from
3460 rcu_note_context_switch().
3462 rcutree.jiffies_till_first_fqs= [KNL]
3463 Set delay from grace-period initialization to
3464 first attempt to force quiescent states.
3465 Units are jiffies, minimum value is zero,
3466 and maximum value is HZ.
3468 rcutree.jiffies_till_next_fqs= [KNL]
3469 Set delay between subsequent attempts to force
3470 quiescent states. Units are jiffies, minimum
3471 value is one, and maximum value is HZ.
3473 rcutree.kthread_prio= [KNL,BOOT]
3474 Set the SCHED_FIFO priority of the RCU per-CPU
3475 kthreads (rcuc/N). This value is also used for
3476 the priority of the RCU boost threads (rcub/N)
3477 and for the RCU grace-period kthreads (rcu_bh,
3478 rcu_preempt, and rcu_sched). If RCU_BOOST is
3479 set, valid values are 1-99 and the default is 1
3480 (the least-favored priority). Otherwise, when
3481 RCU_BOOST is not set, valid values are 0-99 and
3482 the default is zero (non-realtime operation).
3484 rcutree.rcu_nocb_leader_stride= [KNL]
3485 Set the number of NOCB kthread groups, which
3486 defaults to the square root of the number of
3487 CPUs. Larger numbers reduces the wakeup overhead
3488 on the per-CPU grace-period kthreads, but increases
3489 that same overhead on each group's leader.
3491 rcutree.qhimark= [KNL]
3492 Set threshold of queued RCU callbacks beyond which
3493 batch limiting is disabled.
3495 rcutree.qlowmark= [KNL]
3496 Set threshold of queued RCU callbacks below which
3497 batch limiting is re-enabled.
3499 rcutree.rcu_idle_gp_delay= [KNL]
3500 Set wakeup interval for idle CPUs that have
3501 RCU callbacks (RCU_FAST_NO_HZ=y).
3503 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3504 Set wakeup interval for idle CPUs that have
3505 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3506 Lazy RCU callbacks are those which RCU can
3507 prove do nothing more than free memory.
3509 rcutree.rcu_kick_kthreads= [KNL]
3510 Cause the grace-period kthread to get an extra
3511 wake_up() if it sleeps three times longer than
3512 it should at force-quiescent-state time.
3513 This wake_up() will be accompanied by a
3514 WARN_ONCE() splat and an ftrace_dump().
3516 rcuperf.gp_async= [KNL]
3517 Measure performance of asynchronous
3518 grace-period primitives such as call_rcu().
3520 rcuperf.gp_async_max= [KNL]
3521 Specify the maximum number of outstanding
3522 callbacks per writer thread. When a writer
3523 thread exceeds this limit, it invokes the
3524 corresponding flavor of rcu_barrier() to allow
3525 previously posted callbacks to drain.
3527 rcuperf.gp_exp= [KNL]
3528 Measure performance of expedited synchronous
3529 grace-period primitives.
3531 rcuperf.holdoff= [KNL]
3532 Set test-start holdoff period. The purpose of
3533 this parameter is to delay the start of the
3534 test until boot completes in order to avoid
3537 rcuperf.nreaders= [KNL]
3538 Set number of RCU readers. The value -1 selects
3539 N, where N is the number of CPUs. A value
3540 "n" less than -1 selects N-n+1, where N is again
3541 the number of CPUs. For example, -2 selects N
3542 (the number of CPUs), -3 selects N+1, and so on.
3543 A value of "n" less than or equal to -N selects
3546 rcuperf.nwriters= [KNL]
3547 Set number of RCU writers. The values operate
3548 the same as for rcuperf.nreaders.
3549 N, where N is the number of CPUs
3551 rcuperf.perf_type= [KNL]
3552 Specify the RCU implementation to test.
3554 rcuperf.shutdown= [KNL]
3555 Shut the system down after performance tests
3556 complete. This is useful for hands-off automated
3559 rcuperf.verbose= [KNL]
3560 Enable additional printk() statements.
3562 rcuperf.writer_holdoff= [KNL]
3563 Write-side holdoff between grace periods,
3564 in microseconds. The default of zero says
3567 rcutorture.cbflood_inter_holdoff= [KNL]
3568 Set holdoff time (jiffies) between successive
3569 callback-flood tests.
3571 rcutorture.cbflood_intra_holdoff= [KNL]
3572 Set holdoff time (jiffies) between successive
3573 bursts of callbacks within a given callback-flood
3576 rcutorture.cbflood_n_burst= [KNL]
3577 Set the number of bursts making up a given
3578 callback-flood test. Set this to zero to
3579 disable callback-flood testing.
3581 rcutorture.cbflood_n_per_burst= [KNL]
3582 Set the number of callbacks to be registered
3583 in a given burst of a callback-flood test.
3585 rcutorture.fqs_duration= [KNL]
3586 Set duration of force_quiescent_state bursts
3589 rcutorture.fqs_holdoff= [KNL]
3590 Set holdoff time within force_quiescent_state bursts
3593 rcutorture.fqs_stutter= [KNL]
3594 Set wait time between force_quiescent_state bursts
3597 rcutorture.gp_cond= [KNL]
3598 Use conditional/asynchronous update-side
3599 primitives, if available.
3601 rcutorture.gp_exp= [KNL]
3602 Use expedited update-side primitives, if available.
3604 rcutorture.gp_normal= [KNL]
3605 Use normal (non-expedited) asynchronous
3606 update-side primitives, if available.
3608 rcutorture.gp_sync= [KNL]
3609 Use normal (non-expedited) synchronous
3610 update-side primitives, if available. If all
3611 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3612 rcutorture.gp_normal=, and rcutorture.gp_sync=
3613 are zero, rcutorture acts as if is interpreted
3614 they are all non-zero.
3616 rcutorture.n_barrier_cbs= [KNL]
3617 Set callbacks/threads for rcu_barrier() testing.
3619 rcutorture.nfakewriters= [KNL]
3620 Set number of concurrent RCU writers. These just
3621 stress RCU, they don't participate in the actual
3622 test, hence the "fake".
3624 rcutorture.nreaders= [KNL]
3625 Set number of RCU readers. The value -1 selects
3626 N-1, where N is the number of CPUs. A value
3627 "n" less than -1 selects N-n-2, where N is again
3628 the number of CPUs. For example, -2 selects N
3629 (the number of CPUs), -3 selects N+1, and so on.
3631 rcutorture.object_debug= [KNL]
3632 Enable debug-object double-call_rcu() testing.
3634 rcutorture.onoff_holdoff= [KNL]
3635 Set time (s) after boot for CPU-hotplug testing.
3637 rcutorture.onoff_interval= [KNL]
3638 Set time (s) between CPU-hotplug operations, or
3639 zero to disable CPU-hotplug testing.
3641 rcutorture.shuffle_interval= [KNL]
3642 Set task-shuffle interval (s). Shuffling tasks
3643 allows some CPUs to go into dyntick-idle mode
3644 during the rcutorture test.
3646 rcutorture.shutdown_secs= [KNL]
3647 Set time (s) after boot system shutdown. This
3648 is useful for hands-off automated testing.
3650 rcutorture.stall_cpu= [KNL]
3651 Duration of CPU stall (s) to test RCU CPU stall
3652 warnings, zero to disable.
3654 rcutorture.stall_cpu_holdoff= [KNL]
3655 Time to wait (s) after boot before inducing stall.
3657 rcutorture.stall_cpu_irqsoff= [KNL]
3658 Disable interrupts while stalling if set.
3660 rcutorture.stat_interval= [KNL]
3661 Time (s) between statistics printk()s.
3663 rcutorture.stutter= [KNL]
3664 Time (s) to stutter testing, for example, specifying
3665 five seconds causes the test to run for five seconds,
3666 wait for five seconds, and so on. This tests RCU's
3667 ability to transition abruptly to and from idle.
3669 rcutorture.test_boost= [KNL]
3670 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3671 "Maybe" means test if the RCU implementation
3672 under test support RCU priority boosting.
3674 rcutorture.test_boost_duration= [KNL]
3675 Duration (s) of each individual boost test.
3677 rcutorture.test_boost_interval= [KNL]
3678 Interval (s) between each boost test.
3680 rcutorture.test_no_idle_hz= [KNL]
3681 Test RCU's dyntick-idle handling. See also the
3682 rcutorture.shuffle_interval parameter.
3684 rcutorture.torture_type= [KNL]
3685 Specify the RCU implementation to test.
3687 rcutorture.verbose= [KNL]
3688 Enable additional printk() statements.
3690 rcupdate.rcu_cpu_stall_suppress= [KNL]
3691 Suppress RCU CPU stall warning messages.
3693 rcupdate.rcu_cpu_stall_timeout= [KNL]
3694 Set timeout for RCU CPU stall warning messages.
3696 rcupdate.rcu_expedited= [KNL]
3697 Use expedited grace-period primitives, for
3698 example, synchronize_rcu_expedited() instead
3699 of synchronize_rcu(). This reduces latency,
3700 but can increase CPU utilization, degrade
3701 real-time latency, and degrade energy efficiency.
3702 No effect on CONFIG_TINY_RCU kernels.
3704 rcupdate.rcu_normal= [KNL]
3705 Use only normal grace-period primitives,
3706 for example, synchronize_rcu() instead of
3707 synchronize_rcu_expedited(). This improves
3708 real-time latency, CPU utilization, and
3709 energy efficiency, but can expose users to
3710 increased grace-period latency. This parameter
3711 overrides rcupdate.rcu_expedited. No effect on
3712 CONFIG_TINY_RCU kernels.
3714 rcupdate.rcu_normal_after_boot= [KNL]
3715 Once boot has completed (that is, after
3716 rcu_end_inkernel_boot() has been invoked), use
3717 only normal grace-period primitives. No effect
3718 on CONFIG_TINY_RCU kernels.
3720 rcupdate.rcu_task_stall_timeout= [KNL]
3721 Set timeout in jiffies for RCU task stall warning
3722 messages. Disable with a value less than or equal
3725 rcupdate.rcu_self_test= [KNL]
3726 Run the RCU early boot self tests
3728 rcupdate.rcu_self_test_bh= [KNL]
3729 Run the RCU bh early boot self tests
3731 rcupdate.rcu_self_test_sched= [KNL]
3732 Run the RCU sched early boot self tests
3736 Run specified binary instead of /init from the ramdisk,
3737 used for early userspace startup. See initrd.
3740 Turn on/off individual RDT features. List is:
3741 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
3743 E.g. to turn on cmt and turn off mba use:
3747 Format (x86 or x86_64):
3748 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
3750 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
3752 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
3753 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
3754 reboot_force is either force or not specified,
3755 reboot_cpu is s[mp]#### with #### being the processor
3756 to be used for rebooting.
3759 [KNL, SMP] Set scheduler's default relax_domain_level.
3760 See Documentation/cgroup-v1/cpusets.txt.
3762 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
3763 Format: <base1>,<size1>[,<base2>,<size2>,...]
3764 Reserve I/O ports or memory so the kernel won't use
3765 them. If <base> is less than 0x10000, the region
3766 is assumed to be I/O ports; otherwise it is memory.
3768 reservetop= [X86-32]
3770 Reserves a hole at the top of the kernel virtual
3775 Set the amount of memory to reserve for BIOS at
3776 the bottom of the address space.
3778 reset_devices [KNL] Force drivers to reset the underlying device
3779 during initialization.
3782 Specify the partition device for software suspend
3784 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
3786 resume_offset= [SWSUSP]
3787 Specify the offset from the beginning of the partition
3788 given by "resume=" at which the swap header is located,
3789 in <PAGE_SIZE> units (needed only for swap files).
3790 See Documentation/power/swsusp-and-swap-files.txt
3792 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
3793 read the resume files
3795 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
3796 Useful for devices that are detected asynchronously
3797 (e.g. USB and MMC devices).
3799 hibernate= [HIBERNATION]
3800 noresume Don't check if there's a hibernation image
3801 present during boot.
3802 nocompress Don't compress/decompress hibernation images.
3803 no Disable hibernation and resume.
3804 protect_image Turn on image protection during restoration
3805 (that will set all pages holding image data
3806 during restoration read-only).
3808 retain_initrd [RAM] Keep initrd memory after extraction
3810 rfkill.default_state=
3811 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
3812 etc. communication is blocked by default.
3815 rfkill.master_switch_mode=
3816 0 The "airplane mode" button does nothing.
3817 1 The "airplane mode" button toggles between everything
3818 blocked and the previous configuration.
3819 2 The "airplane mode" button toggles between everything
3820 blocked and everything unblocked.
3822 rhash_entries= [KNL,NET]
3823 Set number of hash buckets for route cache
3826 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
3829 ro [KNL] Mount root device read-only on boot
3832 on Mark read-only kernel memory as read-only (default).
3833 off Leave read-only kernel memory writable for debugging.
3836 Enable the uart passthrough on the designated usb port
3837 on Rockchip SoCs. When active, the signals of the
3838 debug-uart get routed to the D+ and D- pins of the usb
3839 port and the regular usb controller gets disabled.
3841 root= [KNL] Root filesystem
3842 See name_to_dev_t comment in init/do_mounts.c.
3844 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
3845 mount the root filesystem
3847 rootflags= [KNL] Set root filesystem mount option string
3849 rootfstype= [KNL] Set root filesystem type
3851 rootwait [KNL] Wait (indefinitely) for root device to show up.
3852 Useful for devices that are detected asynchronously
3853 (e.g. USB and MMC devices).
3855 rproc_mem=nn[KMG][@address]
3856 [KNL,ARM,CMA] Remoteproc physical memory block.
3857 Memory area to be used by remote processor image,
3860 rw [KNL] Mount root device read-write on boot
3862 S [KNL] Run init in single mode
3864 s390_iommu= [HW,S390]
3865 Set s390 IOTLB flushing mode
3867 With strict flushing every unmap operation will result in
3868 an IOTLB flush. Default is lazy flushing before reuse,
3872 See drivers/net/irda/sa1100_ir.c.
3874 sbni= [NET] Granch SBNI12 leased line adapter
3876 sched_debug [KNL] Enables verbose scheduler debug messages.
3878 schedstats= [KNL,X86] Enable or disable scheduled statistics.
3879 Allowed values are enable and disable. This feature
3880 incurs a small amount of overhead in the scheduler
3881 but is useful for debugging and performance tuning.
3883 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
3884 xtime_lock contention on larger systems, and/or RCU lock
3885 contention on all systems with CONFIG_MAXSMP set.
3886 Format: { "0" | "1" }
3887 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
3889 Note: increases power consumption, thus should only be
3890 enabled if running jitter sensitive (HPC/RT) workloads.
3892 security= [SECURITY] Choose a security module to enable at boot.
3893 If this boot parameter is not specified, only the first
3894 security module asking for security registration will be
3895 loaded. An invalid security module name will be treated
3896 as if no module has been chosen.
3898 selinux= [SELINUX] Disable or enable SELinux at boot time.
3899 Format: { "0" | "1" }
3900 See security/selinux/Kconfig help text.
3903 Default value is set via kernel config option.
3904 If enabled at boot time, /selinux/disable can be used
3905 later to disable prior to initial policy load.
3907 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
3908 Format: { "0" | "1" }
3909 See security/apparmor/Kconfig help text
3912 Default value is set via kernel config option.
3914 serialnumber [BUGS=X86-32]
3917 Maximal number of shapers.
3925 Disable merging of slabs with similar size. May be
3926 necessary if there is some reason to distinguish
3927 allocs to different slabs, especially in hardened
3928 environments where the risk of heap overflows and
3929 layout control by attackers can usually be
3930 frustrated by disabling merging. This will reduce
3931 most of the exposure of a heap attack to a single
3932 cache (risks via metadata attacks are mostly
3933 unchanged). Debug options disable merging on their
3935 For more information see Documentation/vm/slub.rst.
3937 slab_max_order= [MM, SLAB]
3938 Determines the maximum allowed order for slabs.
3939 A high setting may cause OOMs due to memory
3940 fragmentation. Defaults to 1 for systems with
3941 more than 32MB of RAM, 0 otherwise.
3943 slub_debug[=options[,slabs]] [MM, SLUB]
3944 Enabling slub_debug allows one to determine the
3945 culprit if slab objects become corrupted. Enabling
3946 slub_debug can create guard zones around objects and
3947 may poison objects when not in use. Also tracks the
3948 last alloc / free. For more information see
3949 Documentation/vm/slub.rst.
3951 slub_memcg_sysfs= [MM, SLUB]
3952 Determines whether to enable sysfs directories for
3953 memory cgroup sub-caches. 1 to enable, 0 to disable.
3954 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
3955 Enabling this can lead to a very high number of debug
3956 directories and files being created under
3959 slub_max_order= [MM, SLUB]
3960 Determines the maximum allowed order for slabs.
3961 A high setting may cause OOMs due to memory
3962 fragmentation. For more information see
3963 Documentation/vm/slub.rst.
3965 slub_min_objects= [MM, SLUB]
3966 The minimum number of objects per slab. SLUB will
3967 increase the slab order up to slub_max_order to
3968 generate a sufficiently large slab able to contain
3969 the number of objects indicated. The higher the number
3970 of objects the smaller the overhead of tracking slabs
3971 and the less frequently locks need to be acquired.
3972 For more information see Documentation/vm/slub.rst.
3974 slub_min_order= [MM, SLUB]
3975 Determines the minimum page order for slabs. Must be
3976 lower than slub_max_order.
3977 For more information see Documentation/vm/slub.rst.
3979 slub_nomerge [MM, SLUB]
3980 Same with slab_nomerge. This is supported for legacy.
3981 See slab_nomerge for more information.
3984 Format: <io1>[,<io2>[,...,<io8>]]
3986 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
3987 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
3988 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
3989 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
3990 smsc-ircc2.ircc_irq= [HW] IRQ line
3991 smsc-ircc2.ircc_dma= [HW] DMA channel
3992 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
3993 0: Toshiba Satellite 1800 (GP data pin select)
3994 1: Fast pin select (default)
3997 smt [KNL,S390] Set the maximum number of threads (logical
3998 CPUs) to use per physical CPU on systems capable of
3999 symmetric multithreading (SMT). Will be capped to the
4000 actual hardware limit.
4002 Default: -1 (no limit)
4005 [KNL] Should the soft-lockup detector generate panics.
4008 A nonzero value instructs the soft-lockup detector
4009 to panic the machine when a soft-lockup occurs. This
4010 is also controlled by CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC
4011 which is the respective build-time switch to that
4014 softlockup_all_cpu_backtrace=
4015 [KNL] Should the soft-lockup detector generate
4016 backtraces on all cpus.
4019 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4020 See Documentation/laptops/sonypi.txt
4022 spectre_v2= [X86] Control mitigation of Spectre variant 2
4023 (indirect branch speculation) vulnerability.
4025 on - unconditionally enable
4026 off - unconditionally disable
4027 auto - kernel detects whether your CPU model is
4030 Selecting 'on' will, and 'auto' may, choose a
4031 mitigation method at run time according to the
4032 CPU, the available microcode, the setting of the
4033 CONFIG_RETPOLINE configuration option, and the
4034 compiler with which the kernel was built.
4036 Specific mitigations can also be selected manually:
4038 retpoline - replace indirect branches
4039 retpoline,generic - google's original retpoline
4040 retpoline,amd - AMD-specific minimal thunk
4042 Not specifying this option is equivalent to
4045 spec_store_bypass_disable=
4046 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4047 (Speculative Store Bypass vulnerability)
4049 Certain CPUs are vulnerable to an exploit against a
4050 a common industry wide performance optimization known
4051 as "Speculative Store Bypass" in which recent stores
4052 to the same memory location may not be observed by
4053 later loads during speculative execution. The idea
4054 is that such stores are unlikely and that they can
4055 be detected prior to instruction retirement at the
4056 end of a particular speculation execution window.
4058 In vulnerable processors, the speculatively forwarded
4059 store can be used in a cache side channel attack, for
4060 example to read memory to which the attacker does not
4061 directly have access (e.g. inside sandboxed code).
4063 This parameter controls whether the Speculative Store
4064 Bypass optimization is used.
4066 on - Unconditionally disable Speculative Store Bypass
4067 off - Unconditionally enable Speculative Store Bypass
4068 auto - Kernel detects whether the CPU model contains an
4069 implementation of Speculative Store Bypass and
4070 picks the most appropriate mitigation. If the
4071 CPU is not vulnerable, "off" is selected. If the
4072 CPU is vulnerable the default mitigation is
4073 architecture and Kconfig dependent. See below.
4074 prctl - Control Speculative Store Bypass per thread
4075 via prctl. Speculative Store Bypass is enabled
4076 for a process by default. The state of the control
4077 is inherited on fork.
4078 seccomp - Same as "prctl" above, but all seccomp threads
4079 will disable SSB unless they explicitly opt out.
4081 Not specifying this option is equivalent to
4082 spec_store_bypass_disable=auto.
4084 Default mitigations:
4085 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4087 spia_io_base= [HW,MTD]
4092 srcutree.counter_wrap_check [KNL]
4093 Specifies how frequently to check for
4094 grace-period sequence counter wrap for the
4095 srcu_data structure's ->srcu_gp_seq_needed field.
4096 The greater the number of bits set in this kernel
4097 parameter, the less frequently counter wrap will
4098 be checked for. Note that the bottom two bits
4101 srcutree.exp_holdoff [KNL]
4102 Specifies how many nanoseconds must elapse
4103 since the end of the last SRCU grace period for
4104 a given srcu_struct until the next normal SRCU
4105 grace period will be considered for automatic
4106 expediting. Set to zero to disable automatic
4110 Speculative Store Bypass Disable control
4112 On CPUs that are vulnerable to the Speculative
4113 Store Bypass vulnerability and offer a
4114 firmware based mitigation, this parameter
4115 indicates how the mitigation should be used:
4117 force-on: Unconditionally enable mitigation for
4118 for both kernel and userspace
4119 force-off: Unconditionally disable mitigation for
4120 for both kernel and userspace
4121 kernel: Always enable mitigation in the
4122 kernel, and offer a prctl interface
4123 to allow userspace to register its
4124 interest in being mitigated too.
4126 stack_guard_gap= [MM]
4127 override the default stack gap protection. The value
4128 is in page units and it defines how many pages prior
4129 to (for stacks growing down) resp. after (for stacks
4130 growing up) the main stack are reserved for no other
4131 mapping. Default value is 256 pages.
4134 Enabled the stack tracer on boot up.
4136 stacktrace_filter=[function-list]
4137 [FTRACE] Limit the functions that the stack tracer
4138 will trace at boot up. function-list is a comma separated
4139 list of functions. This list can be changed at run
4140 time by the stack_trace_filter file in the debugfs
4141 tracing directory. Note, this enables stack tracing
4142 and the stacktrace above is not needed.
4146 Set the STI (builtin display/keyboard on the HP-PARISC
4147 machines) console (graphic card) which should be used
4148 as the initial boot-console.
4149 See also comment in drivers/video/console/sticore.c.
4152 See comment in drivers/video/console/sticore.c.
4155 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4157 sunrpc.min_resvport=
4158 sunrpc.max_resvport=
4160 SunRPC servers often require that client requests
4161 originate from a privileged port (i.e. a port in the
4162 range 0 < portnr < 1024).
4163 An administrator who wishes to reserve some of these
4164 ports for other uses may adjust the range that the
4165 kernel's sunrpc client considers to be privileged
4166 using these two parameters to set the minimum and
4167 maximum port values.
4169 sunrpc.svc_rpc_per_connection_limit=
4171 Limit the number of requests that the server will
4172 process in parallel from a single connection.
4173 The default value is 0 (no limit).
4177 Control how the NFS server code allocates CPUs to
4178 service thread pools. Depending on how many NICs
4179 you have and where their interrupts are bound, this
4180 option will affect which CPUs will do NFS serving.
4181 Note: this parameter cannot be changed while the
4182 NFS server is running.
4184 auto the server chooses an appropriate mode
4185 automatically using heuristics
4186 global a single global pool contains all CPUs
4187 percpu one pool for each CPU
4188 pernode one pool for each NUMA node (equivalent
4189 to global on non-NUMA machines)
4191 sunrpc.tcp_slot_table_entries=
4192 sunrpc.udp_slot_table_entries=
4194 Sets the upper limit on the number of simultaneous
4195 RPC calls that can be sent from the client to a
4196 server. Increasing these values may allow you to
4197 improve throughput, but will also increase the
4198 amount of memory reserved for use by the client.
4200 suspend.pm_test_delay=
4202 Sets the number of seconds to remain in a suspend test
4203 mode before resuming the system (see
4204 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4205 is set. Default value is 5.
4208 [KNL] Enable accounting of swap in memory resource
4209 controller if no parameter or 1 is given or disable
4210 it if 0 is given (See Documentation/cgroup-v1/memory.txt)
4212 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4213 Format: { <int> | force | noforce }
4214 <int> -- Number of I/O TLB slabs
4215 force -- force using of bounce buffers even if they
4216 wouldn't be automatically used by the kernel
4217 noforce -- Never use bounce buffers (for debugging)
4221 sysfs.deprecated=0|1 [KNL]
4222 Enable/disable old style sysfs layout for old udev
4223 on older distributions. When this option is enabled
4224 very new udev will not work anymore. When this option
4225 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4226 in older udev will not work anymore.
4227 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4228 the kernel configuration.
4230 sysrq_always_enabled
4232 Ignore sysrq setting - this boot parameter will
4233 neutralize any effect of /proc/sys/kernel/sysrq.
4234 Useful for debugging.
4236 tcpmhash_entries= [KNL,NET]
4237 Set the number of tcp_metrics_hash slots.
4238 Default value is 8192 or 16384 depending on total
4239 ram pages. This is used to specify the TCP metrics
4240 cache size. See Documentation/networking/ip-sysctl.txt
4241 "tcp_no_metrics_save" section for more details.
4245 test_suspend= [SUSPEND][,N]
4246 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4247 standby suspend) or "freeze" (for suspend type freeze)
4248 as the system sleep state during system startup with
4249 the optional capability to repeat N number of times.
4250 The system is woken from this state using a
4251 wakeup-capable RTC alarm.
4253 thash_entries= [KNL,NET]
4254 Set number of hash buckets for TCP connection
4256 thermal.act= [HW,ACPI]
4257 -1: disable all active trip points in all thermal zones
4258 <degrees C>: override all lowest active trip points
4260 thermal.crt= [HW,ACPI]
4261 -1: disable all critical trip points in all thermal zones
4262 <degrees C>: override all critical trip points
4264 thermal.nocrt= [HW,ACPI]
4265 Set to disable actions on ACPI thermal zone
4266 critical and hot trip points.
4268 thermal.off= [HW,ACPI]
4269 1: disable ACPI thermal control
4271 thermal.psv= [HW,ACPI]
4272 -1: disable all passive trip points
4273 <degrees C>: override all passive trip points to this
4276 thermal.tzp= [HW,ACPI]
4277 Specify global default ACPI thermal zone polling rate
4278 <deci-seconds>: poll all this frequency
4279 0: no polling (default)
4282 Force threading of all interrupt handlers except those
4283 marked explicitly IRQF_NO_THREAD.
4286 Enable the Transcendent memory driver if built-in.
4288 tmem.cleancache=0|1 [KNL, XEN]
4289 Default is on (1). Disable the usage of the cleancache
4290 API to send anonymous pages to the hypervisor.
4292 tmem.frontswap=0|1 [KNL, XEN]
4293 Default is on (1). Disable the usage of the frontswap
4294 API to send swap pages to the hypervisor. If disabled
4295 the selfballooning and selfshrinking are force disabled.
4297 tmem.selfballooning=0|1 [KNL, XEN]
4298 Default is on (1). Disable the driving of swap pages
4301 tmem.selfshrinking=0|1 [KNL, XEN]
4302 Default is on (1). Partial swapoff that immediately
4303 transfers pages from Xen hypervisor back to the
4304 kernel based on different criteria.
4308 Specify if the kernel should make use of the cpu
4309 topology information if the hardware supports this.
4310 The scheduler will make use of this information and
4311 e.g. base its process migration decisions on it.
4314 topology_updates= [KNL, PPC, NUMA]
4316 Specify if the kernel should ignore (off)
4317 topology updates sent by the hypervisor to this
4322 tpm_suspend_pcr=[HW,TPM]
4323 Format: integer pcr id
4324 Specify that at suspend time, the tpm driver
4325 should extend the specified pcr with zeros,
4326 as a workaround for some chips which fail to
4327 flush the last written pcr on TPM_SaveState.
4328 This will guarantee that all the other pcrs
4331 trace_buf_size=nn[KMG]
4332 [FTRACE] will set tracing buffer size on each cpu.
4334 trace_event=[event-list]
4335 [FTRACE] Set and start specified trace events in order
4336 to facilitate early boot debugging. The event-list is a
4337 comma separated list of trace events to enable. See
4338 also Documentation/trace/events.txt
4340 trace_options=[option-list]
4341 [FTRACE] Enable or disable tracer options at boot.
4342 The option-list is a comma delimited list of options
4343 that can be enabled or disabled just as if you were
4344 to echo the option name into
4346 /sys/kernel/debug/tracing/trace_options
4348 For example, to enable stacktrace option (to dump the
4349 stack trace of each event), add to the command line:
4351 trace_options=stacktrace
4353 See also Documentation/trace/ftrace.txt "trace options"
4357 Have the tracepoints sent to printk as well as the
4358 tracing ring buffer. This is useful for early boot up
4359 where the system hangs or reboots and does not give the
4360 option for reading the tracing buffer or performing a
4361 ftrace_dump_on_oops.
4363 To turn off having tracepoints sent to printk,
4364 echo 0 > /proc/sys/kernel/tracepoint_printk
4365 Note, echoing 1 into this file without the
4366 tracepoint_printk kernel cmdline option has no effect.
4370 Having tracepoints sent to printk() and activating high
4371 frequency tracepoints such as irq or sched, can cause
4372 the system to live lock.
4375 [FTRACE] enable this option to disable tracing when a
4376 warning is hit. This turns off "tracing_on". Tracing can
4377 be enabled again by echoing '1' into the "tracing_on"
4378 file located in /sys/kernel/debug/tracing/
4380 This option is useful, as it disables the trace before
4381 the WARNING dump is called, which prevents the trace to
4382 be filled with content caused by the warning output.
4384 This option can also be set at run time via the sysctl
4385 option: kernel/traceoff_on_warning
4387 transparent_hugepage=
4389 Format: [always|madvise|never]
4390 Can be used to control the default behavior of the system
4391 with respect to transparent hugepages.
4392 See Documentation/admin-guide/mm/transhuge.rst
4395 tsc= Disable clocksource stability checks for TSC.
4397 [x86] reliable: mark tsc clocksource as reliable, this
4398 disables clocksource verification at runtime, as well
4399 as the stability checks done at bootup. Used to enable
4400 high-resolution timer mode on older hardware, and in
4401 virtualized environment.
4402 [x86] noirqtime: Do not use TSC to do irq accounting.
4403 Used to run time disable IRQ_TIME_ACCOUNTING on any
4404 platforms where RDTSC is slow and this accounting
4406 [x86] unstable: mark the TSC clocksource as unstable, this
4407 marks the TSC unconditionally unstable at bootup and
4408 avoids any further wobbles once the TSC watchdog notices.
4410 turbografx.map[2|3]= [HW,JOY]
4411 TurboGraFX parallel port interface
4413 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4414 See also Documentation/input/devices/joystick-parport.rst
4416 udbg-immortal [PPC] When debugging early kernel crashes that
4417 happen after console_init() and before a proper
4418 console driver takes over, this boot options might
4419 help "seeing" what's going on.
4421 uhash_entries= [KNL,NET]
4422 Set number of hash buckets for UDP/UDP-Lite connections
4425 [USB] Ignore overcurrent events (default N).
4426 Some badly-designed motherboards generate lots of
4427 bogus events, for ports that aren't wired to
4428 anything. Set this parameter to avoid log spamming.
4429 Note that genuine overcurrent events won't be
4433 [X86] Cause panic on unknown NMI.
4435 usbcore.authorized_default=
4436 [USB] Default USB device authorization:
4437 (default -1 = authorized except for wireless USB,
4438 0 = not authorized, 1 = authorized)
4440 usbcore.autosuspend=
4441 [USB] The autosuspend time delay (in seconds) used
4442 for newly-detected USB devices (default 2). This
4443 is the time required before an idle device will be
4444 autosuspended. Devices for which the delay is set
4445 to a negative value won't be autosuspended at all.
4447 usbcore.usbfs_snoop=
4448 [USB] Set to log all usbfs traffic (default 0 = off).
4450 usbcore.usbfs_snoop_max=
4451 [USB] Maximum number of bytes to snoop in each URB
4454 usbcore.blinkenlights=
4455 [USB] Set to cycle leds on hubs (default 0 = off).
4457 usbcore.old_scheme_first=
4458 [USB] Start with the old device initialization
4459 scheme (default 0 = off).
4461 usbcore.usbfs_memory_mb=
4462 [USB] Memory limit (in MB) for buffers allocated by
4463 usbfs (default = 16, 0 = max = 2047).
4465 usbcore.use_both_schemes=
4466 [USB] Try the other device initialization scheme
4467 if the first one fails (default 1 = enabled).
4469 usbcore.initial_descriptor_timeout=
4470 [USB] Specifies timeout for the initial 64-byte
4471 USB_REQ_GET_DESCRIPTOR request in milliseconds
4472 (default 5000 = 5.0 seconds).
4474 usbcore.nousb [USB] Disable the USB subsystem
4477 [USB] A list of quirk entries to augment the built-in
4478 usb core quirk list. List entries are separated by
4479 commas. Each entry has the form
4480 VendorID:ProductID:Flags. The IDs are 4-digit hex
4481 numbers and Flags is a set of letters. Each letter
4482 will change the built-in quirk; setting it if it is
4483 clear and clearing it if it is set. The letters have
4484 the following meanings:
4485 a = USB_QUIRK_STRING_FETCH_255 (string
4486 descriptors must not be fetched using
4488 b = USB_QUIRK_RESET_RESUME (device can't resume
4489 correctly so reset it instead);
4490 c = USB_QUIRK_NO_SET_INTF (device can't handle
4491 Set-Interface requests);
4492 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
4493 handle its Configuration or Interface
4495 e = USB_QUIRK_RESET (device can't be reset
4496 (e.g morph devices), don't use reset);
4497 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
4498 more interface descriptions than the
4499 bNumInterfaces count, and can't handle
4500 talking to these interfaces);
4501 g = USB_QUIRK_DELAY_INIT (device needs a pause
4502 during initialization, after we read
4503 the device descriptor);
4504 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
4505 high speed and super speed interrupt
4506 endpoints, the USB 2.0 and USB 3.0 spec
4507 require the interval in microframes (1
4508 microframe = 125 microseconds) to be
4509 calculated as interval = 2 ^
4511 Devices with this quirk report their
4512 bInterval as the result of this
4513 calculation instead of the exponent
4514 variable used in the calculation);
4515 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
4516 handle device_qualifier descriptor
4518 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
4519 generates spurious wakeup, ignore
4520 remote wakeup capability);
4521 k = USB_QUIRK_NO_LPM (device can't handle Link
4523 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
4524 (Device reports its bInterval as linear
4525 frames instead of the USB 2.0
4527 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
4528 to be disconnected before suspend to
4529 prevent spurious wakeup);
4530 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
4531 pause after every control message);
4532 Example: quirks=0781:5580:bk,0a5c:5834:gij
4535 [USBHID] The interval which mice are to be polled at.
4538 [USBHID] The interval which joysticks are to be polled at.
4541 [USBHID] The interval which keyboards are to be polled at.
4543 usb-storage.delay_use=
4544 [UMS] The delay in seconds before a new device is
4545 scanned for Logical Units (default 1).
4548 [UMS] A list of quirks entries to supplement or
4549 override the built-in unusual_devs list. List
4550 entries are separated by commas. Each entry has
4551 the form VID:PID:Flags where VID and PID are Vendor
4552 and Product ID values (4-digit hex numbers) and
4553 Flags is a set of characters, each corresponding
4554 to a common usb-storage quirk flag as follows:
4555 a = SANE_SENSE (collect more than 18 bytes
4557 b = BAD_SENSE (don't collect more than 18
4558 bytes of sense data);
4559 c = FIX_CAPACITY (decrease the reported
4560 device capacity by one sector);
4561 d = NO_READ_DISC_INFO (don't use
4562 READ_DISC_INFO command);
4563 e = NO_READ_CAPACITY_16 (don't use
4564 READ_CAPACITY_16 command);
4565 f = NO_REPORT_OPCODES (don't use report opcodes
4567 g = MAX_SECTORS_240 (don't transfer more than
4568 240 sectors at a time, uas only);
4569 h = CAPACITY_HEURISTICS (decrease the
4570 reported device capacity by one
4571 sector if the number is odd);
4572 i = IGNORE_DEVICE (don't bind to this
4574 j = NO_REPORT_LUNS (don't use report luns
4576 l = NOT_LOCKABLE (don't try to lock and
4577 unlock ejectable media);
4578 m = MAX_SECTORS_64 (don't transfer more
4579 than 64 sectors = 32 KB at a time);
4580 n = INITIAL_READ10 (force a retry of the
4581 initial READ(10) command);
4582 o = CAPACITY_OK (accept the capacity
4583 reported by the device);
4584 p = WRITE_CACHE (the device cache is ON
4586 r = IGNORE_RESIDUE (the device reports
4587 bogus residue values);
4588 s = SINGLE_LUN (the device has only one
4590 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
4591 commands, uas only);
4592 u = IGNORE_UAS (don't bind to the uas driver);
4593 w = NO_WP_DETECT (don't test whether the
4594 medium is write-protected).
4595 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
4596 even if the device claims no cache)
4597 Example: quirks=0419:aaf5:rl,0421:0433:rc
4599 user_debug= [KNL,ARM]
4601 See arch/arm/Kconfig.debug help text.
4602 1 - undefined instruction events
4604 4 - invalid data aborts
4607 Example: user_debug=31
4610 [X86] Flags controlling user PTE allocations.
4612 nohigh = do not allocate PTE pages in
4613 HIGHMEM regardless of setting
4617 On X86_32, this is an alias for vdso32=. Otherwise:
4619 vdso=1: enable VDSO (the default)
4620 vdso=0: disable VDSO mapping
4622 vdso32= [X86] Control the 32-bit vDSO
4623 vdso32=1: enable 32-bit VDSO
4624 vdso32=0 or vdso32=2: disable 32-bit VDSO
4626 See the help text for CONFIG_COMPAT_VDSO for more
4627 details. If CONFIG_COMPAT_VDSO is set, the default is
4628 vdso32=0; otherwise, the default is vdso32=1.
4630 For compatibility with older kernels, vdso32=2 is an
4633 Try vdso32=0 if you encounter an error that says:
4634 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
4637 vector=percpu: enable percpu vector domain
4639 video= [FB] Frame buffer configuration
4640 See Documentation/fb/modedb.txt.
4642 video.brightness_switch_enabled= [0,1]
4643 If set to 1, on receiving an ACPI notify event
4644 generated by hotkey, video driver will adjust brightness
4645 level and then send out the event to user space through
4646 the allocated input device; If set to 0, video driver
4647 will only send out the event without touching backlight
4652 [VMMIO] Memory mapped virtio (platform) device.
4654 <size>@<baseaddr>:<irq>[:<id>]
4656 <size> := size (can use standard suffixes
4658 <baseaddr> := physical base address
4659 <irq> := interrupt number (as passed to
4661 <id> := (optional) platform device id
4663 virtio_mmio.device=1K@0x100b0000:48:7
4665 Can be used multiple times for multiple devices.
4667 vga= [BOOT,X86-32] Select a particular video mode
4668 See Documentation/x86/boot.txt and
4669 Documentation/svga.txt.
4670 Use vga=ask for menu.
4671 This is actually a boot loader parameter; the value is
4672 passed to the kernel using a special protocol.
4674 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
4675 size of <nn>. This can be used to increase the
4676 minimum size (128MB on x86). It can also be used to
4677 decrease the size and leave more room for directly
4680 vmcp_cma=nn[MG] [KNL,S390]
4681 Sets the memory size reserved for contiguous memory
4682 allocations for the vmcp device driver.
4684 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
4687 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
4690 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
4694 Controls the behavior of vsyscalls (i.e. calls to
4695 fixed addresses of 0xffffffffff600x00 from legacy
4696 code). Most statically-linked binaries and older
4697 versions of glibc use these calls. Because these
4698 functions are at fixed addresses, they make nice
4699 targets for exploits that can control RIP.
4701 emulate [default] Vsyscalls turn into traps and are
4702 emulated reasonably safely.
4704 native Vsyscalls are native syscall instructions.
4705 This is a little bit faster than trapping
4706 and makes a few dynamic recompilers work
4707 better than they would in emulation mode.
4708 It also makes exploits much easier to write.
4710 none Vsyscalls don't work at all. This makes
4711 them quite hard to use for exploits but
4712 might break your system.
4714 vt.color= [VT] Default text color.
4715 Format: 0xYX, X = foreground, Y = background.
4716 Default: 0x07 = light gray on black.
4718 vt.cur_default= [VT] Default cursor shape.
4719 Format: 0xCCBBAA, where AA, BB, and CC are the same as
4720 the parameters of the <Esc>[?A;B;Cc escape sequence;
4721 see VGA-softcursor.txt. Default: 2 = underline.
4723 vt.default_blu= [VT]
4724 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
4725 Change the default blue palette of the console.
4726 This is a 16-member array composed of values
4729 vt.default_grn= [VT]
4730 Format: <green0>,<green1>,<green2>,...,<green15>
4731 Change the default green palette of the console.
4732 This is a 16-member array composed of values
4735 vt.default_red= [VT]
4736 Format: <red0>,<red1>,<red2>,...,<red15>
4737 Change the default red palette of the console.
4738 This is a 16-member array composed of values
4744 Set system-wide default UTF-8 mode for all tty's.
4745 Default is 1, i.e. UTF-8 mode is enabled for all
4746 newly opened terminals.
4748 vt.global_cursor_default=
4751 Set system-wide default for whether a cursor
4752 is shown on new VTs. Default is -1,
4753 i.e. cursors will be created by default unless
4754 overridden by individual drivers. 0 will hide
4755 cursors, 1 will display them.
4757 vt.italic= [VT] Default color for italic text; 0-15.
4760 vt.underline= [VT] Default color for underlined text; 0-15.
4763 watchdog timers [HW,WDT] For information on watchdog timers,
4764 see Documentation/watchdog/watchdog-parameters.txt
4765 or other driver-specific files in the
4766 Documentation/watchdog/ directory.
4768 workqueue.watchdog_thresh=
4769 If CONFIG_WQ_WATCHDOG is configured, workqueue can
4770 warn stall conditions and dump internal state to
4771 help debugging. 0 disables workqueue stall
4772 detection; otherwise, it's the stall threshold
4773 duration in seconds. The default value is 30 and
4774 it can be updated at runtime by writing to the
4775 corresponding sysfs file.
4777 workqueue.disable_numa
4778 By default, all work items queued to unbound
4779 workqueues are affine to the NUMA nodes they're
4780 issued on, which results in better behavior in
4781 general. If NUMA affinity needs to be disabled for
4782 whatever reason, this option can be used. Note
4783 that this also can be controlled per-workqueue for
4784 workqueues visible under /sys/bus/workqueue/.
4786 workqueue.power_efficient
4787 Per-cpu workqueues are generally preferred because
4788 they show better performance thanks to cache
4789 locality; unfortunately, per-cpu workqueues tend to
4790 be more power hungry than unbound workqueues.
4792 Enabling this makes the per-cpu workqueues which
4793 were observed to contribute significantly to power
4794 consumption unbound, leading to measurably lower
4795 power usage at the cost of small performance
4798 The default value of this parameter is determined by
4799 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
4801 workqueue.debug_force_rr_cpu
4802 Workqueue used to implicitly guarantee that work
4803 items queued without explicit CPU specified are put
4804 on the local CPU. This guarantee is no longer true
4805 and while local CPU is still preferred work items
4806 may be put on foreign CPUs. This debug option
4807 forces round-robin CPU selection to flush out
4808 usages which depend on the now broken guarantee.
4809 When enabled, memory and cache locality will be
4812 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
4813 default x2apic cluster mode on platforms
4816 x86_intel_mid_timer= [X86-32,APBT]
4817 Choose timer option for x86 Intel MID platform.
4818 Two valid options are apbt timer only and lapic timer
4819 plus one apbt timer for broadcast timer.
4820 x86_intel_mid_timer=apbt_only | lapic_and_apbt
4822 xen_512gb_limit [KNL,X86-64,XEN]
4823 Restricts the kernel running paravirtualized under Xen
4824 to use only up to 512 GB of RAM. The reason to do so is
4825 crash analysis tools and Xen tools for doing domain
4826 save/restore/migration must be enabled to handle larger
4829 xen_emul_unplug= [HW,X86,XEN]
4830 Unplug Xen emulated devices
4831 Format: [unplug0,][unplug1]
4832 ide-disks -- unplug primary master IDE devices
4833 aux-ide-disks -- unplug non-primary-master IDE devices
4834 nics -- unplug network devices
4835 all -- unplug all emulated devices (NICs and IDE disks)
4836 unnecessary -- unplugging emulated devices is
4837 unnecessary even if the host did not respond to
4839 never -- do not unplug even if version check succeeds
4841 xen_nopvspin [X86,XEN]
4842 Disables the ticketlock slowpath using Xen PV
4846 Disables the PV optimizations forcing the HVM guest to
4847 run as generic HVM guest with no PV drivers.
4849 xirc2ps_cs= [NET,PCMCIA]
4851 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
git.samba.org / sfrench / cifs-2.6.git / blob