| blob | 30752db575870d23208d3106fd27ebe128594bd6 |
1 acpi= [HW,ACPI,X86,ARM64]
2 Advanced Configuration and Power Interface
3 Format: { force | on | off | strict | noirq | rsdt |
4 copy_dsdt }
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"
14 are available
16 See also Documentation/power/runtime_pm.txt, pci=noacpi
18 acpi_apic_instance= [ACPI, IOAPIC]
19 Format: <int>
20 2: use 2nd APIC table, if available
21 1,0: use 1st APIC table
22 default: 0
24 acpi_backlight= [HW,ACPI]
25 acpi_backlight=vendor
26 acpi_backlight=video
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]
47 Format: <int>
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]
74 { strict | lax | no }
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
92 size limitation.
94 acpi_irq_balance [HW,ACPI]
95 ACPI will balance active IRQs
96 default in APIC mode
98 acpi_irq_nobalance [HW,ACPI]
99 ACPI will not move active IRQs (default)
100 default in PIC mode
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
106 use by PCI
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
113 the GPE dispatcher.
114 This facility can be used to prevent such uncontrolled
115 GPE floodings.
116 Format: <byte>
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
127 kernels.
129 acpi_no_static_ssdt [HW,ACPI]
130 Disable installation of static SSDTs at early boot time
131 By default, SSDTs contained in the RSDT/XSDT will be
132 installed automatically and they will appear under
133 /sys/firmware/acpi/tables.
134 This option turns off this feature.
135 Note that specifying this option does not affect
136 dynamic table installation which will install SSDT
137 tables to /sys/firmware/acpi/tables/dynamic.
139 acpi_no_watchdog [HW,ACPI,WDT]
140 Ignore the ACPI-based watchdog interface (WDAT) and let
141 a native driver control the watchdog device instead.
143 acpi_rsdp= [ACPI,EFI,KEXEC]
144 Pass the RSDP address to the kernel, mostly used
145 on machines running EFI runtime service to boot the
146 second kernel for kdump.
148 acpi_os_name= [HW,ACPI] Tell ACPI BIOS the name of the OS
149 Format: To spoof as Windows 98: ="Microsoft Windows"
151 acpi_rev_override [ACPI] Override the _REV object to return 5 (instead
152 of 2 which is mandated by ACPI 6) as the supported ACPI
153 specification revision (when using this switch, it may
154 be necessary to carry out a cold reboot _twice_ in a
155 row to make it take effect on the platform firmware).
157 acpi_osi= [HW,ACPI] Modify list of supported OS interface strings
158 acpi_osi="string1" # add string1
159 acpi_osi="!string2" # remove string2
160 acpi_osi=!* # remove all strings
161 acpi_osi=! # disable all built-in OS vendor
162 strings
163 acpi_osi=!! # enable all built-in OS vendor
164 strings
165 acpi_osi= # disable all strings
167 'acpi_osi=!' can be used in combination with single or
168 multiple 'acpi_osi="string1"' to support specific OS
169 vendor string(s). Note that such command can only
170 affect the default state of the OS vendor strings, thus
171 it cannot affect the default state of the feature group
172 strings and the current state of the OS vendor strings,
173 specifying it multiple times through kernel command line
174 is meaningless. This command is useful when one do not
175 care about the state of the feature group strings which
176 should be controlled by the OSPM.
177 Examples:
178 1. 'acpi_osi=! acpi_osi="Windows 2000"' is equivalent
179 to 'acpi_osi="Windows 2000" acpi_osi=!', they all
180 can make '_OSI("Windows 2000")' TRUE.
182 'acpi_osi=' cannot be used in combination with other
183 'acpi_osi=' command lines, the _OSI method will not
184 exist in the ACPI namespace. NOTE that such command can
185 only affect the _OSI support state, thus specifying it
186 multiple times through kernel command line is also
187 meaningless.
188 Examples:
189 1. 'acpi_osi=' can make 'CondRefOf(_OSI, Local1)'
190 FALSE.
192 'acpi_osi=!*' can be used in combination with single or
193 multiple 'acpi_osi="string1"' to support specific
194 string(s). Note that such command can affect the
195 current state of both the OS vendor strings and the
196 feature group strings, thus specifying it multiple times
197 through kernel command line is meaningful. But it may
198 still not able to affect the final state of a string if
199 there are quirks related to this string. This command
200 is useful when one want to control the state of the
201 feature group strings to debug BIOS issues related to
202 the OSPM features.
203 Examples:
204 1. 'acpi_osi="Module Device" acpi_osi=!*' can make
205 '_OSI("Module Device")' FALSE.
206 2. 'acpi_osi=!* acpi_osi="Module Device"' can make
207 '_OSI("Module Device")' TRUE.
208 3. 'acpi_osi=! acpi_osi=!* acpi_osi="Windows 2000"' is
209 equivalent to
210 'acpi_osi=!* acpi_osi=! acpi_osi="Windows 2000"'
211 and
212 'acpi_osi=!* acpi_osi="Windows 2000" acpi_osi=!',
213 they all will make '_OSI("Windows 2000")' TRUE.
215 acpi_pm_good [X86]
216 Override the pmtimer bug detection: force the kernel
217 to assume that this machine's pmtimer latches its value
218 and always returns good values.
220 acpi_sci= [HW,ACPI] ACPI System Control Interrupt trigger mode
221 Format: { level | edge | high | low }
223 acpi_skip_timer_override [HW,ACPI]
224 Recognize and ignore IRQ0/pin2 Interrupt Override.
225 For broken nForce2 BIOS resulting in XT-PIC timer.
227 acpi_sleep= [HW,ACPI] Sleep options
228 Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
229 old_ordering, nonvs, sci_force_enable, nobl }
230 See Documentation/power/video.txt for information on
231 s3_bios and s3_mode.
232 s3_beep is for debugging; it makes the PC's speaker beep
233 as soon as the kernel's real-mode entry point is called.
234 s4_nohwsig prevents ACPI hardware signature from being
235 used during resume from hibernation.
236 old_ordering causes the ACPI 1.0 ordering of the _PTS
237 control method, with respect to putting devices into
238 low power states, to be enforced (the ACPI 2.0 ordering
239 of _PTS is used by default).
240 nonvs prevents the kernel from saving/restoring the
241 ACPI NVS memory during suspend/hibernation and resume.
242 sci_force_enable causes the kernel to set SCI_EN directly
243 on resume from S1/S3 (which is against the ACPI spec,
244 but some broken systems don't work without it).
245 nobl causes the internal blacklist of systems known to
246 behave incorrectly in some ways with respect to system
247 suspend and resume to be ignored (use wisely).
249 acpi_use_timer_override [HW,ACPI]
250 Use timer override. For some broken Nvidia NF5 boards
251 that require a timer override, but don't have HPET
253 add_efi_memmap [EFI; X86] Include EFI memory map in
254 kernel's map of available physical RAM.
256 agp= [AGP]
257 { off | try_unsupported }
258 off: disable AGP support
259 try_unsupported: try to drive unsupported chipsets
260 (may crash computer or cause data corruption)
262 ALSA [HW,ALSA]
263 See Documentation/sound/alsa-configuration.rst
265 alignment= [KNL,ARM]
266 Allow the default userspace alignment fault handler
267 behaviour to be specified. Bit 0 enables warnings,
268 bit 1 enables fixups, and bit 2 sends a segfault.
270 align_va_addr= [X86-64]
271 Align virtual addresses by clearing slice [14:12] when
272 allocating a VMA at process creation time. This option
273 gives you up to 3% performance improvement on AMD F15h
274 machines (where it is enabled by default) for a
275 CPU-intensive style benchmark, and it can vary highly in
276 a microbenchmark depending on workload and compiler.
278 32: only for 32-bit processes
279 64: only for 64-bit processes
280 on: enable for both 32- and 64-bit processes
281 off: disable for both 32- and 64-bit processes
283 alloc_snapshot [FTRACE]
284 Allocate the ftrace snapshot buffer on boot up when the
285 main buffer is allocated. This is handy if debugging
286 and you need to use tracing_snapshot() on boot up, and
287 do not want to use tracing_snapshot_alloc() as it needs
288 to be done where GFP_KERNEL allocations are allowed.
290 amd_iommu= [HW,X86-64]
291 Pass parameters to the AMD IOMMU driver in the system.
292 Possible values are:
293 fullflush - enable flushing of IO/TLB entries when
294 they are unmapped. Otherwise they are
295 flushed before they will be reused, which
296 is a lot of faster
297 off - do not initialize any AMD IOMMU found in
298 the system
299 force_isolation - Force device isolation for all
300 devices. The IOMMU driver is not
301 allowed anymore to lift isolation
302 requirements as needed. This option
303 does not override iommu=pt
305 amd_iommu_dump= [HW,X86-64]
306 Enable AMD IOMMU driver option to dump the ACPI table
307 for AMD IOMMU. With this option enabled, AMD IOMMU
308 driver will print ACPI tables for AMD IOMMU during
309 IOMMU initialization.
311 amd_iommu_intr= [HW,X86-64]
312 Specifies one of the following AMD IOMMU interrupt
313 remapping modes:
314 legacy - Use legacy interrupt remapping mode.
315 vapic - Use virtual APIC mode, which allows IOMMU
316 to inject interrupts directly into guest.
317 This mode requires kvm-amd.avic=1.
318 (Default when IOMMU HW support is present.)
320 amijoy.map= [HW,JOY] Amiga joystick support
321 Map of devices attached to JOY0DAT and JOY1DAT
322 Format: <a>,<b>
323 See also Documentation/input/joydev/joystick.rst
325 analog.map= [HW,JOY] Analog joystick and gamepad support
326 Specifies type or capabilities of an analog joystick
327 connected to one of 16 gameports
328 Format: <type1>,<type2>,..<type16>
330 apc= [HW,SPARC]
331 Power management functions (SPARCstation-4/5 + deriv.)
332 Format: noidle
333 Disable APC CPU standby support. SPARCstation-Fox does
334 not play well with APC CPU idle - disable it if you have
335 APC and your system crashes randomly.
337 apic= [APIC,X86] Advanced Programmable Interrupt Controller
338 Change the output verbosity whilst booting
339 Format: { quiet (default) | verbose | debug }
340 Change the amount of debugging information output
341 when initialising the APIC and IO-APIC components.
342 For X86-32, this can also be used to specify an APIC
343 driver name.
344 Format: apic=driver_name
345 Examples: apic=bigsmp
347 apic_extnmi= [APIC,X86] External NMI delivery setting
348 Format: { bsp (default) | all | none }
349 bsp: External NMI is delivered only to CPU 0
350 all: External NMIs are broadcast to all CPUs as a
351 backup of CPU 0
352 none: External NMI is masked for all CPUs. This is
353 useful so that a dump capture kernel won't be
354 shot down by NMI
356 autoconf= [IPV6]
357 See Documentation/networking/ipv6.txt.
359 show_lapic= [APIC,X86] Advanced Programmable Interrupt Controller
360 Limit apic dumping. The parameter defines the maximal
361 number of local apics being dumped. Also it is possible
362 to set it to "all" by meaning -- no limit here.
363 Format: { 1 (default) | 2 | ... | all }.
364 The parameter valid if only apic=debug or
365 apic=verbose is specified.
366 Example: apic=debug show_lapic=all
368 apm= [APM] Advanced Power Management
369 See header of arch/x86/kernel/apm_32.c.
371 arcrimi= [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
372 Format: <io>,<irq>,<nodeID>
374 ataflop= [HW,M68k]
376 atarimouse= [HW,MOUSE] Atari Mouse
378 atkbd.extra= [HW] Enable extra LEDs and keys on IBM RapidAccess,
379 EzKey and similar keyboards
381 atkbd.reset= [HW] Reset keyboard during initialization
383 atkbd.set= [HW] Select keyboard code set
384 Format: <int> (2 = AT (default), 3 = PS/2)
386 atkbd.scroll= [HW] Enable scroll wheel on MS Office and similar
387 keyboards
389 atkbd.softraw= [HW] Choose between synthetic and real raw mode
390 Format: <bool> (0 = real, 1 = synthetic (default))
392 atkbd.softrepeat= [HW]
393 Use software keyboard repeat
395 audit= [KNL] Enable the audit sub-system
396 Format: { "0" | "1" | "off" | "on" }
397 0 | off - kernel audit is disabled and can not be
398 enabled until the next reboot
399 unset - kernel audit is initialized but disabled and
400 will be fully enabled by the userspace auditd.
401 1 | on - kernel audit is initialized and partially
402 enabled, storing at most audit_backlog_limit
403 messages in RAM until it is fully enabled by the
404 userspace auditd.
405 Default: unset
407 audit_backlog_limit= [KNL] Set the audit queue size limit.
408 Format: <int> (must be >=0)
409 Default: 64
411 bau= [X86_UV] Enable the BAU on SGI UV. The default
412 behavior is to disable the BAU (i.e. bau=0).
413 Format: { "0" | "1" }
414 0 - Disable the BAU.
415 1 - Enable the BAU.
416 unset - Disable the BAU.
418 baycom_epp= [HW,AX25]
419 Format: <io>,<mode>
421 baycom_par= [HW,AX25] BayCom Parallel Port AX.25 Modem
422 Format: <io>,<mode>
423 See header of drivers/net/hamradio/baycom_par.c.
425 baycom_ser_fdx= [HW,AX25]
426 BayCom Serial Port AX.25 Modem (Full Duplex Mode)
427 Format: <io>,<irq>,<mode>[,<baud>]
428 See header of drivers/net/hamradio/baycom_ser_fdx.c.
430 baycom_ser_hdx= [HW,AX25]
431 BayCom Serial Port AX.25 Modem (Half Duplex Mode)
432 Format: <io>,<irq>,<mode>
433 See header of drivers/net/hamradio/baycom_ser_hdx.c.
435 blkdevparts= Manual partition parsing of block device(s) for
436 embedded devices based on command line input.
437 See Documentation/block/cmdline-partition.txt
439 boot_delay= Milliseconds to delay each printk during boot.
440 Values larger than 10 seconds (10000) are changed to
441 no delay (0).
442 Format: integer
444 bootmem_debug [KNL] Enable bootmem allocator debug messages.
446 bert_disable [ACPI]
447 Disable BERT OS support on buggy BIOSes.
449 bttv.card= [HW,V4L] bttv (bt848 + bt878 based grabber cards)
450 bttv.radio= Most important insmod options are available as
451 kernel args too.
452 bttv.pll= See Documentation/media/v4l-drivers/bttv.rst
453 bttv.tuner=
455 bulk_remove=off [PPC] This parameter disables the use of the pSeries
456 firmware feature for flushing multiple hpte entries
457 at a time.
459 c101= [NET] Moxa C101 synchronous serial card
461 cachesize= [BUGS=X86-32] Override level 2 CPU cache size detection.
462 Sometimes CPU hardware bugs make them report the cache
463 size incorrectly. The kernel will attempt work arounds
464 to fix known problems, but for some CPUs it is not
465 possible to determine what the correct size should be.
466 This option provides an override for these situations.
468 ca_keys= [KEYS] This parameter identifies a specific key(s) on
469 the system trusted keyring to be used for certificate
470 trust validation.
471 format: { id:<keyid> | builtin }
473 cca= [MIPS] Override the kernel pages' cache coherency
474 algorithm. Accepted values range from 0 to 7
475 inclusive. See arch/mips/include/asm/pgtable-bits.h
476 for platform specific values (SB1, Loongson3 and
477 others).
479 ccw_timeout_log [S390]
480 See Documentation/s390/CommonIO for details.
482 cgroup_disable= [KNL] Disable a particular controller
483 Format: {name of the controller(s) to disable}
484 The effects of cgroup_disable=foo are:
485 - foo isn't auto-mounted if you mount all cgroups in
486 a single hierarchy
487 - foo isn't visible as an individually mountable
488 subsystem
489 {Currently only "memory" controller deal with this and
490 cut the overhead, others just disable the usage. So
491 only cgroup_disable=memory is actually worthy}
493 cgroup_no_v1= [KNL] Disable one, multiple, all cgroup controllers in v1
494 Format: { controller[,controller...] | "all" }
495 Like cgroup_disable, but only applies to cgroup v1;
496 the blacklisted controllers remain available in cgroup2.
498 cgroup.memory= [KNL] Pass options to the cgroup memory controller.
499 Format: <string>
500 nosocket -- Disable socket memory accounting.
501 nokmem -- Disable kernel memory accounting.
503 checkreqprot [SELINUX] Set initial checkreqprot flag value.
504 Format: { "0" | "1" }
505 See security/selinux/Kconfig help text.
506 0 -- check protection applied by kernel (includes
507 any implied execute protection).
508 1 -- check protection requested by application.
509 Default value is set via a kernel config option.
510 Value can be changed at runtime via
511 /selinux/checkreqprot.
513 cio_ignore= [S390]
514 See Documentation/s390/CommonIO for details.
515 clk_ignore_unused
516 [CLK]
517 Prevents the clock framework from automatically gating
518 clocks that have not been explicitly enabled by a Linux
519 device driver but are enabled in hardware at reset or
520 by the bootloader/firmware. Note that this does not
521 force such clocks to be always-on nor does it reserve
522 those clocks in any way. This parameter is useful for
523 debug and development, but should not be needed on a
524 platform with proper driver support. For more
525 information, see Documentation/driver-api/clk.rst.
527 clock= [BUGS=X86-32, HW] gettimeofday clocksource override.
528 [Deprecated]
529 Forces specified clocksource (if available) to be used
530 when calculating gettimeofday(). If specified
531 clocksource is not available, it defaults to PIT.
532 Format: { pit | tsc | cyclone | pmtmr }
534 clocksource= Override the default clocksource
535 Format: <string>
536 Override the default clocksource and use the clocksource
537 with the name specified.
538 Some clocksource names to choose from, depending on
539 the platform:
540 [all] jiffies (this is the base, fallback clocksource)
541 [ACPI] acpi_pm
542 [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
543 pxa_timer,timer3,32k_counter,timer0_1
544 [X86-32] pit,hpet,tsc;
545 scx200_hrt on Geode; cyclone on IBM x440
546 [MIPS] MIPS
547 [PARISC] cr16
548 [S390] tod
549 [SH] SuperH
550 [SPARC64] tick
551 [X86-64] hpet,tsc
553 clocksource.arm_arch_timer.evtstrm=
554 [ARM,ARM64]
555 Format: <bool>
556 Enable/disable the eventstream feature of the ARM
557 architected timer so that code using WFE-based polling
558 loops can be debugged more effectively on production
559 systems.
561 clearcpuid=BITNUM [X86]
562 Disable CPUID feature X for the kernel. See
563 arch/x86/include/asm/cpufeatures.h for the valid bit
564 numbers. Note the Linux specific bits are not necessarily
565 stable over kernel options, but the vendor specific
566 ones should be.
567 Also note that user programs calling CPUID directly
568 or using the feature without checking anything
569 will still see it. This just prevents it from
570 being used by the kernel or shown in /proc/cpuinfo.
571 Also note the kernel might malfunction if you disable
572 some critical bits.
574 cma=nn[MG]@[start[MG][-end[MG]]]
575 [ARM,X86,KNL]
576 Sets the size of kernel global memory area for
577 contiguous memory allocations and optionally the
578 placement constraint by the physical address range of
579 memory allocations. A value of 0 disables CMA
580 altogether. For more information, see
581 include/linux/dma-contiguous.h
583 cmo_free_hint= [PPC] Format: { yes | no }
584 Specify whether pages are marked as being inactive
585 when they are freed. This is used in CMO environments
586 to determine OS memory pressure for page stealing by
587 a hypervisor.
588 Default: yes
590 coherent_pool=nn[KMG] [ARM,KNL]
591 Sets the size of memory pool for coherent, atomic dma
592 allocations, by default set to 256K.
594 com20020= [HW,NET] ARCnet - COM20020 chipset
595 Format:
596 <io>[,<irq>[,<nodeID>[,<backplane>[,<ckp>[,<timeout>]]]]]
598 com90io= [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
599 Format: <io>[,<irq>]
601 com90xx= [HW,NET]
602 ARCnet - COM90xx chipset (memory-mapped buffers)
603 Format: <io>[,<irq>[,<memstart>]]
605 condev= [HW,S390] console device
606 conmode=
608 console= [KNL] Output console device and options.
610 tty<n> Use the virtual console device <n>.
612 ttyS<n>[,options]
613 ttyUSB0[,options]
614 Use the specified serial port. The options are of
615 the form "bbbbpnf", where "bbbb" is the baud rate,
616 "p" is parity ("n", "o", or "e"), "n" is number of
617 bits, and "f" is flow control ("r" for RTS or
618 omit it). Default is "9600n8".
620 See Documentation/admin-guide/serial-console.rst for more
621 information. See
622 Documentation/networking/netconsole.txt for an
623 alternative.
625 uart[8250],io,<addr>[,options]
626 uart[8250],mmio,<addr>[,options]
627 uart[8250],mmio16,<addr>[,options]
628 uart[8250],mmio32,<addr>[,options]
629 uart[8250],0x<addr>[,options]
630 Start an early, polled-mode console on the 8250/16550
631 UART at the specified I/O port or MMIO address,
632 switching to the matching ttyS device later.
633 MMIO inter-register address stride is either 8-bit
634 (mmio), 16-bit (mmio16), or 32-bit (mmio32).
635 If none of [io|mmio|mmio16|mmio32], <addr> is assumed
636 to be equivalent to 'mmio'. 'options' are specified in
637 the same format described for ttyS above; if unspecified,
638 the h/w is not re-initialized.
640 hvc<n> Use the hypervisor console device <n>. This is for
641 both Xen and PowerPC hypervisors.
643 If the device connected to the port is not a TTY but a braille
644 device, prepend "brl," before the device type, for instance
645 console=brl,ttyS0
646 For now, only VisioBraille is supported.
648 console_msg_format=
649 [KNL] Change console messages format
650 default
651 By default we print messages on consoles in
652 "[time stamp] text\n" format (time stamp may not be
653 printed, depending on CONFIG_PRINTK_TIME or
654 `printk_time' param).
655 syslog
656 Switch to syslog format: "<%u>[time stamp] text\n"
657 IOW, each message will have a facility and loglevel
658 prefix. The format is similar to one used by syslog()
659 syscall, or to executing "dmesg -S --raw" or to reading
660 from /proc/kmsg.
662 consoleblank= [KNL] The console blank (screen saver) timeout in
663 seconds. A value of 0 disables the blank timer.
664 Defaults to 0.
666 coredump_filter=
667 [KNL] Change the default value for
668 /proc/<pid>/coredump_filter.
669 See also Documentation/filesystems/proc.txt.
671 coresight_cpu_debug.enable
672 [ARM,ARM64]
673 Format: <bool>
674 Enable/disable the CPU sampling based debugging.
675 0: default value, disable debugging
676 1: enable debugging at boot time
678 cpuidle.off=1 [CPU_IDLE]
679 disable the cpuidle sub-system
681 cpufreq.off=1 [CPU_FREQ]
682 disable the cpufreq sub-system
684 cpu_init_udelay=N
685 [X86] Delay for N microsec between assert and de-assert
686 of APIC INIT to start processors. This delay occurs
687 on every CPU online, such as boot, and resume from suspend.
688 Default: 10000
690 cpcihp_generic= [HW,PCI] Generic port I/O CompactPCI driver
691 Format:
692 <first_slot>,<last_slot>,<port>,<enum_bit>[,<debug>]
694 crashkernel=size[KMG][@offset[KMG]]
695 [KNL] Using kexec, Linux can switch to a 'crash kernel'
696 upon panic. This parameter reserves the physical
697 memory region [offset, offset + size] for that kernel
698 image. If '@offset' is omitted, then a suitable offset
699 is selected automatically. Check
700 Documentation/kdump/kdump.txt for further details.
702 crashkernel=range1:size1[,range2:size2,...][@offset]
703 [KNL] Same as above, but depends on the memory
704 in the running system. The syntax of range is
705 start-[end] where start and end are both
706 a memory unit (amount[KMG]). See also
707 Documentation/kdump/kdump.txt for an example.
709 crashkernel=size[KMG],high
710 [KNL, x86_64] range could be above 4G. Allow kernel
711 to allocate physical memory region from top, so could
712 be above 4G if system have more than 4G ram installed.
713 Otherwise memory region will be allocated below 4G, if
714 available.
715 It will be ignored if crashkernel=X is specified.
716 crashkernel=size[KMG],low
717 [KNL, x86_64] range under 4G. When crashkernel=X,high
718 is passed, kernel could allocate physical memory region
719 above 4G, that cause second kernel crash on system
720 that require some amount of low memory, e.g. swiotlb
721 requires at least 64M+32K low memory, also enough extra
722 low memory is needed to make sure DMA buffers for 32-bit
723 devices won't run out. Kernel would try to allocate at
724 at least 256M below 4G automatically.
725 This one let user to specify own low range under 4G
726 for second kernel instead.
727 0: to disable low allocation.
728 It will be ignored when crashkernel=X,high is not used
729 or memory reserved is below 4G.
731 cryptomgr.notests
732 [KNL] Disable crypto self-tests
734 cs89x0_dma= [HW,NET]
735 Format: <dma>
737 cs89x0_media= [HW,NET]
738 Format: { rj45 | aui | bnc }
740 dasd= [HW,NET]
741 See header of drivers/s390/block/dasd_devmap.c.
743 db9.dev[2|3]= [HW,JOY] Multisystem joystick support via parallel port
744 (one device per port)
745 Format: <port#>,<type>
746 See also Documentation/input/devices/joystick-parport.rst
748 ddebug_query= [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
749 time. See
750 Documentation/admin-guide/dynamic-debug-howto.rst for
751 details. Deprecated, see dyndbg.
753 debug [KNL] Enable kernel debugging (events log level).
755 debug_boot_weak_hash
756 [KNL] Enable printing [hashed] pointers early in the
757 boot sequence. If enabled, we use a weak hash instead
758 of siphash to hash pointers. Use this option if you are
759 seeing instances of '(___ptrval___)') and need to see a
760 value (hashed pointer) instead. Cryptographically
761 insecure, please do not use on production kernels.
763 debug_locks_verbose=
764 [KNL] verbose self-tests
765 Format=<0|1>
766 Print debugging info while doing the locking API
767 self-tests.
768 We default to 0 (no extra messages), setting it to
769 1 will print _a lot_ more information - normally
770 only useful to kernel developers.
772 debug_objects [KNL] Enable object debugging
774 no_debug_objects
775 [KNL] Disable object debugging
777 debug_guardpage_minorder=
778 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
779 parameter allows control of the order of pages that will
780 be intentionally kept free (and hence protected) by the
781 buddy allocator. Bigger value increase the probability
782 of catching random memory corruption, but reduce the
783 amount of memory for normal system use. The maximum
784 possible value is MAX_ORDER/2. Setting this parameter
785 to 1 or 2 should be enough to identify most random
786 memory corruption problems caused by bugs in kernel or
787 driver code when a CPU writes to (or reads from) a
788 random memory location. Note that there exists a class
789 of memory corruptions problems caused by buggy H/W or
790 F/W or by drivers badly programing DMA (basically when
791 memory is written at bus level and the CPU MMU is
792 bypassed) which are not detectable by
793 CONFIG_DEBUG_PAGEALLOC, hence this option will not help
794 tracking down these problems.
796 debug_pagealloc=
797 [KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
798 parameter enables the feature at boot time. In
799 default, it is disabled. We can avoid allocating huge
800 chunk of memory for debug pagealloc if we don't enable
801 it at boot time and the system will work mostly same
802 with the kernel built without CONFIG_DEBUG_PAGEALLOC.
803 on: enable the feature
805 debugpat [X86] Enable PAT debugging
807 decnet.addr= [HW,NET]
808 Format: <area>[,<node>]
809 See also Documentation/networking/decnet.txt.
811 default_hugepagesz=
812 [same as hugepagesz=] The size of the default
813 HugeTLB page size. This is the size represented by
814 the legacy /proc/ hugepages APIs, used for SHM, and
815 default size when mounting hugetlbfs filesystems.
816 Defaults to the default architecture's huge page size
817 if not specified.
819 deferred_probe_timeout=
820 [KNL] Debugging option to set a timeout in seconds for
821 deferred probe to give up waiting on dependencies to
822 probe. Only specific dependencies (subsystems or
823 drivers) that have opted in will be ignored. A timeout of 0
824 will timeout at the end of initcalls. This option will also
825 dump out devices still on the deferred probe list after
826 retrying.
828 dhash_entries= [KNL]
829 Set number of hash buckets for dentry cache.
831 disable_1tb_segments [PPC]
832 Disables the use of 1TB hash page table segments. This
833 causes the kernel to fall back to 256MB segments which
834 can be useful when debugging issues that require an SLB
835 miss to occur.
837 disable= [IPV6]
838 See Documentation/networking/ipv6.txt.
840 hardened_usercopy=
841 [KNL] Under CONFIG_HARDENED_USERCOPY, whether
842 hardening is enabled for this boot. Hardened
843 usercopy checking is used to protect the kernel
844 from reading or writing beyond known memory
845 allocation boundaries as a proactive defense
846 against bounds-checking flaws in the kernel's
847 copy_to_user()/copy_from_user() interface.
848 on Perform hardened usercopy checks (default).
849 off Disable hardened usercopy checks.
851 disable_radix [PPC]
852 Disable RADIX MMU mode on POWER9
854 disable_cpu_apicid= [X86,APIC,SMP]
855 Format: <int>
856 The number of initial APIC ID for the
857 corresponding CPU to be disabled at boot,
858 mostly used for the kdump 2nd kernel to
859 disable BSP to wake up multiple CPUs without
860 causing system reset or hang due to sending
861 INIT from AP to BSP.
863 disable_ddw [PPC/PSERIES]
864 Disable Dynamic DMA Window support. Use this if
865 to workaround buggy firmware.
867 disable_ipv6= [IPV6]
868 See Documentation/networking/ipv6.txt.
870 disable_mtrr_cleanup [X86]
871 The kernel tries to adjust MTRR layout from continuous
872 to discrete, to make X server driver able to add WB
873 entry later. This parameter disables that.
875 disable_mtrr_trim [X86, Intel and AMD only]
876 By default the kernel will trim any uncacheable
877 memory out of your available memory pool based on
878 MTRR settings. This parameter disables that behavior,
879 possibly causing your machine to run very slowly.
881 disable_timer_pin_1 [X86]
882 Disable PIN 1 of APIC timer
883 Can be useful to work around chipset bugs.
885 dis_ucode_ldr [X86] Disable the microcode loader.
887 dma_debug=off If the kernel is compiled with DMA_API_DEBUG support,
888 this option disables the debugging code at boot.
890 dma_debug_entries=<number>
891 This option allows to tune the number of preallocated
892 entries for DMA-API debugging code. One entry is
893 required per DMA-API allocation. Use this if the
894 DMA-API debugging code disables itself because the
895 architectural default is too low.
897 dma_debug_driver=<driver_name>
898 With this option the DMA-API debugging driver
899 filter feature can be enabled at boot time. Just
900 pass the driver to filter for as the parameter.
901 The filter can be disabled or changed to another
902 driver later using sysfs.
904 drm.edid_firmware=[<connector>:]<file>[,[<connector>:]<file>]
905 Broken monitors, graphic adapters, KVMs and EDIDless
906 panels may send no or incorrect EDID data sets.
907 This parameter allows to specify an EDID data sets
908 in the /lib/firmware directory that are used instead.
909 Generic built-in EDID data sets are used, if one of
910 edid/1024x768.bin, edid/1280x1024.bin,
911 edid/1680x1050.bin, or edid/1920x1080.bin is given
912 and no file with the same name exists. Details and
913 instructions how to build your own EDID data are
914 available in Documentation/EDID/HOWTO.txt. An EDID
915 data set will only be used for a particular connector,
916 if its name and a colon are prepended to the EDID
917 name. Each connector may use a unique EDID data
918 set by separating the files with a comma. An EDID
919 data set with no connector name will be used for
920 any connectors not explicitly specified.
922 dscc4.setup= [NET]
924 dt_cpu_ftrs= [PPC]
925 Format: {"off" | "known"}
926 Control how the dt_cpu_ftrs device-tree binding is
927 used for CPU feature discovery and setup (if it
928 exists).
929 off: Do not use it, fall back to legacy cpu table.
930 known: Do not pass through unknown features to guests
931 or userspace, only those that the kernel is aware of.
933 dump_apple_properties [X86]
934 Dump name and content of EFI device properties on
935 x86 Macs. Useful for driver authors to determine
936 what data is available or for reverse-engineering.
938 dyndbg[="val"] [KNL,DYNAMIC_DEBUG]
939 module.dyndbg[="val"]
940 Enable debug messages at boot time. See
941 Documentation/admin-guide/dynamic-debug-howto.rst
942 for details.
944 nompx [X86] Disables Intel Memory Protection Extensions.
945 See Documentation/x86/intel_mpx.txt for more
946 information about the feature.
948 nopku [X86] Disable Memory Protection Keys CPU feature found
949 in some Intel CPUs.
951 module.async_probe [KNL]
952 Enable asynchronous probe on this module.
954 early_ioremap_debug [KNL]
955 Enable debug messages in early_ioremap support. This
956 is useful for tracking down temporary early mappings
957 which are not unmapped.
959 earlycon= [KNL] Output early console device and options.
961 [ARM64] The early console is determined by the
962 stdout-path property in device tree's chosen node,
963 or determined by the ACPI SPCR table.
965 [X86] When used with no options the early console is
966 determined by the ACPI SPCR table.
968 cdns,<addr>[,options]
969 Start an early, polled-mode console on a Cadence
970 (xuartps) serial port at the specified address. Only
971 supported option is baud rate. If baud rate is not
972 specified, the serial port must already be setup and
973 configured.
975 uart[8250],io,<addr>[,options]
976 uart[8250],mmio,<addr>[,options]
977 uart[8250],mmio32,<addr>[,options]
978 uart[8250],mmio32be,<addr>[,options]
979 uart[8250],0x<addr>[,options]
980 Start an early, polled-mode console on the 8250/16550
981 UART at the specified I/O port or MMIO address.
982 MMIO inter-register address stride is either 8-bit
983 (mmio) or 32-bit (mmio32 or mmio32be).
984 If none of [io|mmio|mmio32|mmio32be], <addr> is assumed
985 to be equivalent to 'mmio'. 'options' are specified
986 in the same format described for "console=ttyS<n>"; if
987 unspecified, the h/w is not initialized.
989 pl011,<addr>
990 pl011,mmio32,<addr>
991 Start an early, polled-mode console on a pl011 serial
992 port at the specified address. The pl011 serial port
993 must already be setup and configured. Options are not
994 yet supported. If 'mmio32' is specified, then only
995 the driver will use only 32-bit accessors to read/write
996 the device registers.
998 meson,<addr>
999 Start an early, polled-mode console on a meson serial
1000 port at the specified address. The serial port must
1001 already be setup and configured. Options are not yet
1002 supported.
1004 msm_serial,<addr>
1005 Start an early, polled-mode console on an msm serial
1006 port at the specified address. The serial port
1007 must already be setup and configured. Options are not
1008 yet supported.
1010 msm_serial_dm,<addr>
1011 Start an early, polled-mode console on an msm serial
1012 dm port at the specified address. The serial port
1013 must already be setup and configured. Options are not
1014 yet supported.
1016 owl,<addr>
1017 Start an early, polled-mode console on a serial port
1018 of an Actions Semi SoC, such as S500 or S900, at the
1019 specified address. The serial port must already be
1020 setup and configured. Options are not yet supported.
1022 smh Use ARM semihosting calls for early console.
1024 s3c2410,<addr>
1025 s3c2412,<addr>
1026 s3c2440,<addr>
1027 s3c6400,<addr>
1028 s5pv210,<addr>
1029 exynos4210,<addr>
1030 Use early console provided by serial driver available
1031 on Samsung SoCs, requires selecting proper type and
1032 a correct base address of the selected UART port. The
1033 serial port must already be setup and configured.
1034 Options are not yet supported.
1036 lantiq,<addr>
1037 Start an early, polled-mode console on a lantiq serial
1038 (lqasc) port at the specified address. The serial port
1039 must already be setup and configured. Options are not
1040 yet supported.
1042 lpuart,<addr>
1043 lpuart32,<addr>
1044 Use early console provided by Freescale LP UART driver
1045 found on Freescale Vybrid and QorIQ LS1021A processors.
1046 A valid base address must be provided, and the serial
1047 port must already be setup and configured.
1049 ar3700_uart,<addr>
1050 Start an early, polled-mode console on the
1051 Armada 3700 serial port at the specified
1052 address. The serial port must already be setup
1053 and configured. Options are not yet supported.
1055 qcom_geni,<addr>
1056 Start an early, polled-mode console on a Qualcomm
1057 Generic Interface (GENI) based serial port at the
1058 specified address. The serial port must already be
1059 setup and configured. Options are not yet supported.
1061 earlyprintk= [X86,SH,ARM,M68k,S390]
1062 earlyprintk=vga
1063 earlyprintk=efi
1064 earlyprintk=sclp
1065 earlyprintk=xen
1066 earlyprintk=serial[,ttySn[,baudrate]]
1067 earlyprintk=serial[,0x...[,baudrate]]
1068 earlyprintk=ttySn[,baudrate]
1069 earlyprintk=dbgp[debugController#]
1070 earlyprintk=pciserial[,force],bus:device.function[,baudrate]
1071 earlyprintk=xdbc[xhciController#]
1073 earlyprintk is useful when the kernel crashes before
1074 the normal console is initialized. It is not enabled by
1075 default because it has some cosmetic problems.
1077 Append ",keep" to not disable it when the real console
1078 takes over.
1080 Only one of vga, efi, serial, or usb debug port can
1081 be used at a time.
1083 Currently only ttyS0 and ttyS1 may be specified by
1084 name. Other I/O ports may be explicitly specified
1085 on some architectures (x86 and arm at least) by
1086 replacing ttySn with an I/O port address, like this:
1087 earlyprintk=serial,0x1008,115200
1088 You can find the port for a given device in
1089 /proc/tty/driver/serial:
1090 2: uart:ST16650V2 port:00001008 irq:18 ...
1092 Interaction with the standard serial driver is not
1093 very good.
1095 The VGA and EFI output is eventually overwritten by
1096 the real console.
1098 The xen output can only be used by Xen PV guests.
1100 The sclp output can only be used on s390.
1102 The optional "force" to "pciserial" enables use of a
1103 PCI device even when its classcode is not of the
1104 UART class.
1106 edac_report= [HW,EDAC] Control how to report EDAC event
1107 Format: {"on" | "off" | "force"}
1108 on: enable EDAC to report H/W event. May be overridden
1109 by other higher priority error reporting module.
1110 off: disable H/W event reporting through EDAC.
1111 force: enforce the use of EDAC to report H/W event.
1112 default: on.
1114 ekgdboc= [X86,KGDB] Allow early kernel console debugging
1115 ekgdboc=kbd
1117 This is designed to be used in conjunction with
1118 the boot argument: earlyprintk=vga
1120 edd= [EDD]
1121 Format: {"off" | "on" | "skip[mbr]"}
1123 efi= [EFI]
1124 Format: { "old_map", "nochunk", "noruntime", "debug" }
1125 old_map [X86-64]: switch to the old ioremap-based EFI
1126 runtime services mapping. 32-bit still uses this one by
1127 default.
1128 nochunk: disable reading files in "chunks" in the EFI
1129 boot stub, as chunking can cause problems with some
1130 firmware implementations.
1131 noruntime : disable EFI runtime services support
1132 debug: enable misc debug output
1134 efi_no_storage_paranoia [EFI; X86]
1135 Using this parameter you can use more than 50% of
1136 your efi variable storage. Use this parameter only if
1137 you are really sure that your UEFI does sane gc and
1138 fulfills the spec otherwise your board may brick.
1140 efi_fake_mem= nn[KMG]@ss[KMG]:aa[,nn[KMG]@ss[KMG]:aa,..] [EFI; X86]
1141 Add arbitrary attribute to specific memory range by
1142 updating original EFI memory map.
1143 Region of memory which aa attribute is added to is
1144 from ss to ss+nn.
1145 If efi_fake_mem=2G@4G:0x10000,2G@0x10a0000000:0x10000
1146 is specified, EFI_MEMORY_MORE_RELIABLE(0x10000)
1147 attribute is added to range 0x100000000-0x180000000 and
1148 0x10a0000000-0x1120000000.
1150 Using this parameter you can do debugging of EFI memmap
1151 related feature. For example, you can do debugging of
1152 Address Range Mirroring feature even if your box
1153 doesn't support it.
1155 efivar_ssdt= [EFI; X86] Name of an EFI variable that contains an SSDT
1156 that is to be dynamically loaded by Linux. If there are
1157 multiple variables with the same name but with different
1158 vendor GUIDs, all of them will be loaded. See
1159 Documentation/acpi/ssdt-overlays.txt for details.
1162 eisa_irq_edge= [PARISC,HW]
1163 See header of drivers/parisc/eisa.c.
1165 elanfreq= [X86-32]
1166 See comment before function elanfreq_setup() in
1167 arch/x86/kernel/cpu/cpufreq/elanfreq.c.
1169 elevator= [IOSCHED]
1170 Format: {"cfq" | "deadline" | "noop"}
1171 See Documentation/block/cfq-iosched.txt and
1172 Documentation/block/deadline-iosched.txt for details.
1174 elfcorehdr=[size[KMG]@]offset[KMG] [IA64,PPC,SH,X86,S390]
1175 Specifies physical address of start of kernel core
1176 image elf header and optionally the size. Generally
1177 kexec loader will pass this option to capture kernel.
1178 See Documentation/kdump/kdump.txt for details.
1180 enable_mtrr_cleanup [X86]
1181 The kernel tries to adjust MTRR layout from continuous
1182 to discrete, to make X server driver able to add WB
1183 entry later. This parameter enables that.
1185 enable_timer_pin_1 [X86]
1186 Enable PIN 1 of APIC timer
1187 Can be useful to work around chipset bugs
1188 (in particular on some ATI chipsets).
1189 The kernel tries to set a reasonable default.
1191 enforcing [SELINUX] Set initial enforcing status.
1192 Format: {"0" | "1"}
1193 See security/selinux/Kconfig help text.
1194 0 -- permissive (log only, no denials).
1195 1 -- enforcing (deny and log).
1196 Default value is 0.
1197 Value can be changed at runtime via /selinux/enforce.
1199 erst_disable [ACPI]
1200 Disable Error Record Serialization Table (ERST)
1201 support.
1203 ether= [HW,NET] Ethernet cards parameters
1204 This option is obsoleted by the "netdev=" option, which
1205 has equivalent usage. See its documentation for details.
1207 evm= [EVM]
1208 Format: { "fix" }
1209 Permit 'security.evm' to be updated regardless of
1210 current integrity status.
1212 failslab=
1213 fail_page_alloc=
1214 fail_make_request=[KNL]
1215 General fault injection mechanism.
1216 Format: <interval>,<probability>,<space>,<times>
1217 See also Documentation/fault-injection/.
1219 floppy= [HW]
1220 See Documentation/blockdev/floppy.txt.
1222 force_pal_cache_flush
1223 [IA-64] Avoid check_sal_cache_flush which may hang on
1224 buggy SAL_CACHE_FLUSH implementations. Using this
1225 parameter will force ia64_sal_cache_flush to call
1226 ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.
1228 forcepae [X86-32]
1229 Forcefully enable Physical Address Extension (PAE).
1230 Many Pentium M systems disable PAE but may have a
1231 functionally usable PAE implementation.
1232 Warning: use of this parameter will taint the kernel
1233 and may cause unknown problems.
1235 ftrace=[tracer]
1236 [FTRACE] will set and start the specified tracer
1237 as early as possible in order to facilitate early
1238 boot debugging.
1240 ftrace_dump_on_oops[=orig_cpu]
1241 [FTRACE] will dump the trace buffers on oops.
1242 If no parameter is passed, ftrace will dump
1243 buffers of all CPUs, but if you pass orig_cpu, it will
1244 dump only the buffer of the CPU that triggered the
1245 oops.
1247 ftrace_filter=[function-list]
1248 [FTRACE] Limit the functions traced by the function
1249 tracer at boot up. function-list is a comma separated
1250 list of functions. This list can be changed at run
1251 time by the set_ftrace_filter file in the debugfs
1252 tracing directory.
1254 ftrace_notrace=[function-list]
1255 [FTRACE] Do not trace the functions specified in
1256 function-list. This list can be changed at run time
1257 by the set_ftrace_notrace file in the debugfs
1258 tracing directory.
1260 ftrace_graph_filter=[function-list]
1261 [FTRACE] Limit the top level callers functions traced
1262 by the function graph tracer at boot up.
1263 function-list is a comma separated list of functions
1264 that can be changed at run time by the
1265 set_graph_function file in the debugfs tracing directory.
1267 ftrace_graph_notrace=[function-list]
1268 [FTRACE] Do not trace from the functions specified in
1269 function-list. This list is a comma separated list of
1270 functions that can be changed at run time by the
1271 set_graph_notrace file in the debugfs tracing directory.
1273 ftrace_graph_max_depth=<uint>
1274 [FTRACE] Used with the function graph tracer. This is
1275 the max depth it will trace into a function. This value
1276 can be changed at run time by the max_graph_depth file
1277 in the tracefs tracing directory. default: 0 (no limit)
1279 gamecon.map[2|3]=
1280 [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
1281 support via parallel port (up to 5 devices per port)
1282 Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
1283 See also Documentation/input/devices/joystick-parport.rst
1285 gamma= [HW,DRM]
1287 gart_fix_e820= [X86_64] disable the fix e820 for K8 GART
1288 Format: off | on
1289 default: on
1291 gcov_persist= [GCOV] When non-zero (default), profiling data for
1292 kernel modules is saved and remains accessible via
1293 debugfs, even when the module is unloaded/reloaded.
1294 When zero, profiling data is discarded and associated
1295 debugfs files are removed at module unload time.
1297 goldfish [X86] Enable the goldfish android emulator platform.
1298 Don't use this when you are not running on the
1299 android emulator
1301 gpt [EFI] Forces disk with valid GPT signature but
1302 invalid Protective MBR to be treated as GPT. If the
1303 primary GPT is corrupted, it enables the backup/alternate
1304 GPT to be used instead.
1306 grcan.enable0= [HW] Configuration of physical interface 0. Determines
1307 the "Enable 0" bit of the configuration register.
1308 Format: 0 | 1
1309 Default: 0
1310 grcan.enable1= [HW] Configuration of physical interface 1. Determines
1311 the "Enable 0" bit of the configuration register.
1312 Format: 0 | 1
1313 Default: 0
1314 grcan.select= [HW] Select which physical interface to use.
1315 Format: 0 | 1
1316 Default: 0
1317 grcan.txsize= [HW] Sets the size of the tx buffer.
1318 Format: <unsigned int> such that (txsize & ~0x1fffc0) == 0.
1319 Default: 1024
1320 grcan.rxsize= [HW] Sets the size of the rx buffer.
1321 Format: <unsigned int> such that (rxsize & ~0x1fffc0) == 0.
1322 Default: 1024
1324 gpio-mockup.gpio_mockup_ranges
1325 [HW] Sets the ranges of gpiochip of for this device.
1326 Format: <start1>,<end1>,<start2>,<end2>...
1328 hardlockup_all_cpu_backtrace=
1329 [KNL] Should the hard-lockup detector generate
1330 backtraces on all cpus.
1331 Format: <integer>
1333 hashdist= [KNL,NUMA] Large hashes allocated during boot
1334 are distributed across NUMA nodes. Defaults on
1335 for 64-bit NUMA, off otherwise.
1336 Format: 0 | 1 (for off | on)
1338 hcl= [IA-64] SGI's Hardware Graph compatibility layer
1340 hd= [EIDE] (E)IDE hard drive subsystem geometry
1341 Format: <cyl>,<head>,<sect>
1343 hest_disable [ACPI]
1344 Disable Hardware Error Source Table (HEST) support;
1345 corresponding firmware-first mode error processing
1346 logic will be disabled.
1348 highmem=nn[KMG] [KNL,BOOT] forces the highmem zone to have an exact
1349 size of <nn>. This works even on boxes that have no
1350 highmem otherwise. This also works to reduce highmem
1351 size on bigger boxes.
1353 highres= [KNL] Enable/disable high resolution timer mode.
1354 Valid parameters: "on", "off"
1355 Default: "on"
1357 hisax= [HW,ISDN]
1358 See Documentation/isdn/README.HiSax.
1360 hlt [BUGS=ARM,SH]
1362 hpet= [X86-32,HPET] option to control HPET usage
1363 Format: { enable (default) | disable | force |
1364 verbose }
1365 disable: disable HPET and use PIT instead
1366 force: allow force enabled of undocumented chips (ICH4,
1367 VIA, nVidia)
1368 verbose: show contents of HPET registers during setup
1370 hpet_mmap= [X86, HPET_MMAP] Allow userspace to mmap HPET
1371 registers. Default set by CONFIG_HPET_MMAP_DEFAULT.
1373 hugepages= [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
1374 hugepagesz= [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
1375 On x86-64 and powerpc, this option can be specified
1376 multiple times interleaved with hugepages= to reserve
1377 huge pages of different sizes. Valid pages sizes on
1378 x86-64 are 2M (when the CPU supports "pse") and 1G
1379 (when the CPU supports the "pdpe1gb" cpuinfo flag).
1381 hung_task_panic=
1382 [KNL] Should the hung task detector generate panics.
1383 Format: <integer>
1385 A nonzero value instructs the kernel to panic when a
1386 hung task is detected. The default value is controlled
1387 by the CONFIG_BOOTPARAM_HUNG_TASK_PANIC build-time
1388 option. The value selected by this boot parameter can
1389 be changed later by the kernel.hung_task_panic sysctl.
1391 hvc_iucv= [S390] Number of z/VM IUCV hypervisor console (HVC)
1392 terminal devices. Valid values: 0..8
1393 hvc_iucv_allow= [S390] Comma-separated list of z/VM user IDs.
1394 If specified, z/VM IUCV HVC accepts connections
1395 from listed z/VM user IDs only.
1396 keep_bootcon [KNL]
1397 Do not unregister boot console at start. This is only
1398 useful for debugging when something happens in the window
1399 between unregistering the boot console and initializing
1400 the real console.
1402 i2c_bus= [HW] Override the default board specific I2C bus speed
1403 or register an additional I2C bus that is not
1404 registered from board initialization code.
1405 Format:
1406 <bus_id>,<clkrate>
1408 i8042.debug [HW] Toggle i8042 debug mode
1409 i8042.unmask_kbd_data
1410 [HW] Enable printing of interrupt data from the KBD port
1411 (disabled by default, and as a pre-condition
1412 requires that i8042.debug=1 be enabled)
1413 i8042.direct [HW] Put keyboard port into non-translated mode
1414 i8042.dumbkbd [HW] Pretend that controller can only read data from
1415 keyboard and cannot control its state
1416 (Don't attempt to blink the leds)
1417 i8042.noaux [HW] Don't check for auxiliary (== mouse) port
1418 i8042.nokbd [HW] Don't check/create keyboard port
1419 i8042.noloop [HW] Disable the AUX Loopback command while probing
1420 for the AUX port
1421 i8042.nomux [HW] Don't check presence of an active multiplexing
1422 controller
1423 i8042.nopnp [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
1424 controllers
1425 i8042.notimeout [HW] Ignore timeout condition signalled by controller
1426 i8042.reset [HW] Reset the controller during init, cleanup and
1427 suspend-to-ram transitions, only during s2r
1428 transitions, or never reset
1429 Format: { 1 | Y | y | 0 | N | n }
1430 1, Y, y: always reset controller
1431 0, N, n: don't ever reset controller
1432 Default: only on s2r transitions on x86; most other
1433 architectures force reset to be always executed
1434 i8042.unlock [HW] Unlock (ignore) the keylock
1435 i8042.kbdreset [HW] Reset device connected to KBD port
1437 i810= [HW,DRM]
1439 i8k.ignore_dmi [HW] Continue probing hardware even if DMI data
1440 indicates that the driver is running on unsupported
1441 hardware.
1442 i8k.force [HW] Activate i8k driver even if SMM BIOS signature
1443 does not match list of supported models.
1444 i8k.power_status
1445 [HW] Report power status in /proc/i8k
1446 (disabled by default)
1447 i8k.restricted [HW] Allow controlling fans only if SYS_ADMIN
1448 capability is set.
1450 i915.invert_brightness=
1451 [DRM] Invert the sense of the variable that is used to
1452 set the brightness of the panel backlight. Normally a
1453 brightness value of 0 indicates backlight switched off,
1454 and the maximum of the brightness value sets the backlight
1455 to maximum brightness. If this parameter is set to 0
1456 (default) and the machine requires it, or this parameter
1457 is set to 1, a brightness value of 0 sets the backlight
1458 to maximum brightness, and the maximum of the brightness
1459 value switches the backlight off.
1460 -1 -- never invert brightness
1461 0 -- machine default
1462 1 -- force brightness inversion
1464 icn= [HW,ISDN]
1465 Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]
1467 ide-core.nodma= [HW] (E)IDE subsystem
1468 Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
1469 .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
1470 .cdrom .chs .ignore_cable are additional options
1471 See Documentation/ide/ide.txt.
1473 ide-generic.probe-mask= [HW] (E)IDE subsystem
1474 Format: <int>
1475 Probe mask for legacy ISA IDE ports. Depending on
1476 platform up to 6 ports are supported, enabled by
1477 setting corresponding bits in the mask to 1. The
1478 default value is 0x0, which has a special meaning.
1479 On systems that have PCI, it triggers scanning the
1480 PCI bus for the first and the second port, which
1481 are then probed. On systems without PCI the value
1482 of 0x0 enables probing the two first ports as if it
1483 was 0x3.
1485 ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
1486 Claim all unknown PCI IDE storage controllers.
1488 idle= [X86]
1489 Format: idle=poll, idle=halt, idle=nomwait
1490 Poll forces a polling idle loop that can slightly
1491 improve the performance of waking up a idle CPU, but
1492 will use a lot of power and make the system run hot.
1493 Not recommended.
1494 idle=halt: Halt is forced to be used for CPU idle.
1495 In such case C2/C3 won't be used again.
1496 idle=nomwait: Disable mwait for CPU C-states
1498 ieee754= [MIPS] Select IEEE Std 754 conformance mode
1499 Format: { strict | legacy | 2008 | relaxed }
1500 Default: strict
1502 Choose which programs will be accepted for execution
1503 based on the IEEE 754 NaN encoding(s) supported by
1504 the FPU and the NaN encoding requested with the value
1505 of an ELF file header flag individually set by each
1506 binary. Hardware implementations are permitted to
1507 support either or both of the legacy and the 2008 NaN
1508 encoding mode.
1510 Available settings are as follows:
1511 strict accept binaries that request a NaN encoding
1512 supported by the FPU
1513 legacy only accept legacy-NaN binaries, if supported
1514 by the FPU
1515 2008 only accept 2008-NaN binaries, if supported
1516 by the FPU
1517 relaxed accept any binaries regardless of whether
1518 supported by the FPU
1520 The FPU emulator is always able to support both NaN
1521 encodings, so if no FPU hardware is present or it has
1522 been disabled with 'nofpu', then the settings of
1523 'legacy' and '2008' strap the emulator accordingly,
1524 'relaxed' straps the emulator for both legacy-NaN and
1525 2008-NaN, whereas 'strict' enables legacy-NaN only on
1526 legacy processors and both NaN encodings on MIPS32 or
1527 MIPS64 CPUs.
1529 The setting for ABS.fmt/NEG.fmt instruction execution
1530 mode generally follows that for the NaN encoding,
1531 except where unsupported by hardware.
1533 ignore_loglevel [KNL]
1534 Ignore loglevel setting - this will print /all/
1535 kernel messages to the console. Useful for debugging.
1536 We also add it as printk module parameter, so users
1537 could change it dynamically, usually by
1538 /sys/module/printk/parameters/ignore_loglevel.
1540 ignore_rlimit_data
1541 Ignore RLIMIT_DATA setting for data mappings,
1542 print warning at first misuse. Can be changed via
1543 /sys/module/kernel/parameters/ignore_rlimit_data.
1545 ihash_entries= [KNL]
1546 Set number of hash buckets for inode cache.
1548 ima_appraise= [IMA] appraise integrity measurements
1549 Format: { "off" | "enforce" | "fix" | "log" }
1550 default: "enforce"
1552 ima_appraise_tcb [IMA]
1553 The builtin appraise policy appraises all files
1554 owned by uid=0.
1556 ima_canonical_fmt [IMA]
1557 Use the canonical format for the binary runtime
1558 measurements, instead of host native format.
1560 ima_hash= [IMA]
1561 Format: { md5 | sha1 | rmd160 | sha256 | sha384
1562 | sha512 | ... }
1563 default: "sha1"
1565 The list of supported hash algorithms is defined
1566 in crypto/hash_info.h.
1568 ima_policy= [IMA]
1569 The builtin policies to load during IMA setup.
1570 Format: "tcb | appraise_tcb | secure_boot |
1571 fail_securely"
1573 The "tcb" policy measures all programs exec'd, files
1574 mmap'd for exec, and all files opened with the read
1575 mode bit set by either the effective uid (euid=0) or
1576 uid=0.
1578 The "appraise_tcb" policy appraises the integrity of
1579 all files owned by root. (This is the equivalent
1580 of ima_appraise_tcb.)
1582 The "secure_boot" policy appraises the integrity
1583 of files (eg. kexec kernel image, kernel modules,
1584 firmware, policy, etc) based on file signatures.
1586 The "fail_securely" policy forces file signature
1587 verification failure also on privileged mounted
1588 filesystems with the SB_I_UNVERIFIABLE_SIGNATURE
1589 flag.
1591 ima_tcb [IMA] Deprecated. Use ima_policy= instead.
1592 Load a policy which meets the needs of the Trusted
1593 Computing Base. This means IMA will measure all
1594 programs exec'd, files mmap'd for exec, and all files
1595 opened for read by uid=0.
1597 ima_template= [IMA]
1598 Select one of defined IMA measurements template formats.
1599 Formats: { "ima" | "ima-ng" | "ima-sig" }
1600 Default: "ima-ng"
1602 ima_template_fmt=
1603 [IMA] Define a custom template format.
1604 Format: { "field1|...|fieldN" }
1606 ima.ahash_minsize= [IMA] Minimum file size for asynchronous hash usage
1607 Format: <min_file_size>
1608 Set the minimal file size for using asynchronous hash.
1609 If left unspecified, ahash usage is disabled.
1611 ahash performance varies for different data sizes on
1612 different crypto accelerators. This option can be used
1613 to achieve the best performance for a particular HW.
1615 ima.ahash_bufsize= [IMA] Asynchronous hash buffer size
1616 Format: <bufsize>
1617 Set hashing buffer size. Default: 4k.
1619 ahash performance varies for different chunk sizes on
1620 different crypto accelerators. This option can be used
1621 to achieve best performance for particular HW.
1623 init= [KNL]
1624 Format: <full_path>
1625 Run specified binary instead of /sbin/init as init
1626 process.
1628 initcall_debug [KNL] Trace initcalls as they are executed. Useful
1629 for working out where the kernel is dying during
1630 startup.
1632 initcall_blacklist= [KNL] Do not execute a comma-separated list of
1633 initcall functions. Useful for debugging built-in
1634 modules and initcalls.
1636 initrd= [BOOT] Specify the location of the initial ramdisk
1638 init_pkru= [x86] Specify the default memory protection keys rights
1639 register contents for all processes. 0x55555554 by
1640 default (disallow access to all but pkey 0). Can
1641 override in debugfs after boot.
1643 inport.irq= [HW] Inport (ATI XL and Microsoft) busmouse driver
1644 Format: <irq>
1646 int_pln_enable [x86] Enable power limit notification interrupt
1648 integrity_audit=[IMA]
1649 Format: { "0" | "1" }
1650 0 -- basic integrity auditing messages. (Default)
1651 1 -- additional integrity auditing messages.
1653 intel_iommu= [DMAR] Intel IOMMU driver (DMAR) option
1654 on
1655 Enable intel iommu driver.
1656 off
1657 Disable intel iommu driver.
1658 igfx_off [Default Off]
1659 By default, gfx is mapped as normal device. If a gfx
1660 device has a dedicated DMAR unit, the DMAR unit is
1661 bypassed by not enabling DMAR with this option. In
1662 this case, gfx device will use physical address for
1663 DMA.
1664 forcedac [x86_64]
1665 With this option iommu will not optimize to look
1666 for io virtual address below 32-bit forcing dual
1667 address cycle on pci bus for cards supporting greater
1668 than 32-bit addressing. The default is to look
1669 for translation below 32-bit and if not available
1670 then look in the higher range.
1671 strict [Default Off]
1672 With this option on every unmap_single operation will
1673 result in a hardware IOTLB flush operation as opposed
1674 to batching them for performance.
1675 sp_off [Default Off]
1676 By default, super page will be supported if Intel IOMMU
1677 has the capability. With this option, super page will
1678 not be supported.
1679 ecs_off [Default Off]
1680 By default, extended context tables will be supported if
1681 the hardware advertises that it has support both for the
1682 extended tables themselves, and also PASID support. With
1683 this option set, extended tables will not be used even
1684 on hardware which claims to support them.
1685 tboot_noforce [Default Off]
1686 Do not force the Intel IOMMU enabled under tboot.
1687 By default, tboot will force Intel IOMMU on, which
1688 could harm performance of some high-throughput
1689 devices like 40GBit network cards, even if identity
1690 mapping is enabled.
1691 Note that using this option lowers the security
1692 provided by tboot because it makes the system
1693 vulnerable to DMA attacks.
1695 intel_idle.max_cstate= [KNL,HW,ACPI,X86]
1696 0 disables intel_idle and fall back on acpi_idle.
1697 1 to 9 specify maximum depth of C-state.
1699 intel_pstate= [X86]
1700 disable
1701 Do not enable intel_pstate as the default
1702 scaling driver for the supported processors
1703 passive
1704 Use intel_pstate as a scaling driver, but configure it
1705 to work with generic cpufreq governors (instead of
1706 enabling its internal governor). This mode cannot be
1707 used along with the hardware-managed P-states (HWP)
1708 feature.
1709 force
1710 Enable intel_pstate on systems that prohibit it by default
1711 in favor of acpi-cpufreq. Forcing the intel_pstate driver
1712 instead of acpi-cpufreq may disable platform features, such
1713 as thermal controls and power capping, that rely on ACPI
1714 P-States information being indicated to OSPM and therefore
1715 should be used with caution. This option does not work with
1716 processors that aren't supported by the intel_pstate driver
1717 or on platforms that use pcc-cpufreq instead of acpi-cpufreq.
1718 no_hwp
1719 Do not enable hardware P state control (HWP)
1720 if available.
1721 hwp_only
1722 Only load intel_pstate on systems which support
1723 hardware P state control (HWP) if available.
1724 support_acpi_ppc
1725 Enforce ACPI _PPC performance limits. If the Fixed ACPI
1726 Description Table, specifies preferred power management
1727 profile as "Enterprise Server" or "Performance Server",
1728 then this feature is turned on by default.
1729 per_cpu_perf_limits
1730 Allow per-logical-CPU P-State performance control limits using
1731 cpufreq sysfs interface
1733 intremap= [X86-64, Intel-IOMMU]
1734 on enable Interrupt Remapping (default)
1735 off disable Interrupt Remapping
1736 nosid disable Source ID checking
1737 no_x2apic_optout
1738 BIOS x2APIC opt-out request will be ignored
1739 nopost disable Interrupt Posting
1741 iomem= Disable strict checking of access to MMIO memory
1742 strict regions from userspace.
1743 relaxed
1745 iommu= [x86]
1746 off
1747 force
1748 noforce
1749 biomerge
1750 panic
1751 nopanic
1752 merge
1753 nomerge
1754 soft
1755 pt [x86]
1756 nopt [x86]
1757 nobypass [PPC/POWERNV]
1758 Disable IOMMU bypass, using IOMMU for PCI devices.
1760 iommu.passthrough=
1761 [ARM64] Configure DMA to bypass the IOMMU by default.
1762 Format: { "0" | "1" }
1763 0 - Use IOMMU translation for DMA.
1764 1 - Bypass the IOMMU for DMA.
1765 unset - Use IOMMU translation for DMA.
1767 io7= [HW] IO7 for Marvel based alpha systems
1768 See comment before marvel_specify_io7 in
1769 arch/alpha/kernel/core_marvel.c.
1771 io_delay= [X86] I/O delay method
1772 0x80
1773 Standard port 0x80 based delay
1774 0xed
1775 Alternate port 0xed based delay (needed on some systems)
1776 udelay
1777 Simple two microseconds delay
1778 none
1779 No delay
1781 ip= [IP_PNP]
1782 See Documentation/filesystems/nfs/nfsroot.txt.
1784 irqaffinity= [SMP] Set the default irq affinity mask
1785 The argument is a cpu list, as described above.
1787 irqchip.gicv2_force_probe=
1788 [ARM, ARM64]
1789 Format: <bool>
1790 Force the kernel to look for the second 4kB page
1791 of a GICv2 controller even if the memory range
1792 exposed by the device tree is too small.
1794 irqchip.gicv3_nolpi=
1795 [ARM, ARM64]
1796 Force the kernel to ignore the availability of
1797 LPIs (and by consequence ITSs). Intended for system
1798 that use the kernel as a bootloader, and thus want
1799 to let secondary kernels in charge of setting up
1800 LPIs.
1802 irqfixup [HW]
1803 When an interrupt is not handled search all handlers
1804 for it. Intended to get systems with badly broken
1805 firmware running.
1807 irqpoll [HW]
1808 When an interrupt is not handled search all handlers
1809 for it. Also check all handlers each timer
1810 interrupt. Intended to get systems with badly broken
1811 firmware running.
1813 isapnp= [ISAPNP]
1814 Format: <RDP>,<reset>,<pci_scan>,<verbosity>
1816 isolcpus= [KNL,SMP,ISOL] Isolate a given set of CPUs from disturbance.
1817 [Deprecated - use cpusets instead]
1818 Format: [flag-list,]<cpu-list>
1820 Specify one or more CPUs to isolate from disturbances
1821 specified in the flag list (default: domain):
1823 nohz
1824 Disable the tick when a single task runs.
1826 A residual 1Hz tick is offloaded to workqueues, which you
1827 need to affine to housekeeping through the global
1828 workqueue's affinity configured via the
1829 /sys/devices/virtual/workqueue/cpumask sysfs file, or
1830 by using the 'domain' flag described below.
1832 NOTE: by default the global workqueue runs on all CPUs,
1833 so to protect individual CPUs the 'cpumask' file has to
1834 be configured manually after bootup.
1836 domain
1837 Isolate from the general SMP balancing and scheduling
1838 algorithms. Note that performing domain isolation this way
1839 is irreversible: it's not possible to bring back a CPU to
1840 the domains once isolated through isolcpus. It's strongly
1841 advised to use cpusets instead to disable scheduler load
1842 balancing through the "cpuset.sched_load_balance" file.
1843 It offers a much more flexible interface where CPUs can
1844 move in and out of an isolated set anytime.
1846 You can move a process onto or off an "isolated" CPU via
1847 the CPU affinity syscalls or cpuset.
1848 <cpu number> begins at 0 and the maximum value is
1849 "number of CPUs in system - 1".
1851 The format of <cpu-list> is described above.
1855 iucv= [HW,NET]
1857 ivrs_ioapic [HW,X86_64]
1858 Provide an override to the IOAPIC-ID<->DEVICE-ID
1859 mapping provided in the IVRS ACPI table. For
1860 example, to map IOAPIC-ID decimal 10 to
1861 PCI device 00:14.0 write the parameter as:
1862 ivrs_ioapic[10]=00:14.0
1864 ivrs_hpet [HW,X86_64]
1865 Provide an override to the HPET-ID<->DEVICE-ID
1866 mapping provided in the IVRS ACPI table. For
1867 example, to map HPET-ID decimal 0 to
1868 PCI device 00:14.0 write the parameter as:
1869 ivrs_hpet[0]=00:14.0
1871 ivrs_acpihid [HW,X86_64]
1872 Provide an override to the ACPI-HID:UID<->DEVICE-ID
1873 mapping provided in the IVRS ACPI table. For
1874 example, to map UART-HID:UID AMD0020:0 to
1875 PCI device 00:14.5 write the parameter as:
1876 ivrs_acpihid[00:14.5]=AMD0020:0
1878 js= [HW,JOY] Analog joystick
1879 See Documentation/input/joydev/joystick.rst.
1881 nokaslr [KNL]
1882 When CONFIG_RANDOMIZE_BASE is set, this disables
1883 kernel and module base offset ASLR (Address Space
1884 Layout Randomization).
1886 kasan_multi_shot
1887 [KNL] Enforce KASAN (Kernel Address Sanitizer) to print
1888 report on every invalid memory access. Without this
1889 parameter KASAN will print report only for the first
1890 invalid access.
1892 keepinitrd [HW,ARM]
1894 kernelcore= [KNL,X86,IA-64,PPC]
1895 Format: nn[KMGTPE] | nn% | "mirror"
1896 This parameter specifies the amount of memory usable by
1897 the kernel for non-movable allocations. The requested
1898 amount is spread evenly throughout all nodes in the
1899 system as ZONE_NORMAL. The remaining memory is used for
1900 movable memory in its own zone, ZONE_MOVABLE. In the
1901 event, a node is too small to have both ZONE_NORMAL and
1902 ZONE_MOVABLE, kernelcore memory will take priority and
1903 other nodes will have a larger ZONE_MOVABLE.
1905 ZONE_MOVABLE is used for the allocation of pages that
1906 may be reclaimed or moved by the page migration
1907 subsystem. Note that allocations like PTEs-from-HighMem
1908 still use the HighMem zone if it exists, and the Normal
1909 zone if it does not.
1911 It is possible to specify the exact amount of memory in
1912 the form of "nn[KMGTPE]", a percentage of total system
1913 memory in the form of "nn%", or "mirror". If "mirror"
1914 option is specified, mirrored (reliable) memory is used
1915 for non-movable allocations and remaining memory is used
1916 for Movable pages. "nn[KMGTPE]", "nn%", and "mirror"
1917 are exclusive, so you cannot specify multiple forms.
1919 kgdbdbgp= [KGDB,HW] kgdb over EHCI usb debug port.
1920 Format: <Controller#>[,poll interval]
1921 The controller # is the number of the ehci usb debug
1922 port as it is probed via PCI. The poll interval is
1923 optional and is the number seconds in between
1924 each poll cycle to the debug port in case you need
1925 the functionality for interrupting the kernel with
1926 gdb or control-c on the dbgp connection. When
1927 not using this parameter you use sysrq-g to break into
1928 the kernel debugger.
1930 kgdboc= [KGDB,HW] kgdb over consoles.
1931 Requires a tty driver that supports console polling,
1932 or a supported polling keyboard driver (non-usb).
1933 Serial only format: <serial_device>[,baud]
1934 keyboard only format: kbd
1935 keyboard and serial format: kbd,<serial_device>[,baud]
1936 Optional Kernel mode setting:
1937 kms, kbd format: kms,kbd
1938 kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]
1940 kgdbwait [KGDB] Stop kernel execution and enter the
1941 kernel debugger at the earliest opportunity.
1943 kmac= [MIPS] korina ethernet MAC address.
1944 Configure the RouterBoard 532 series on-chip
1945 Ethernet adapter MAC address.
1947 kmemleak= [KNL] Boot-time kmemleak enable/disable
1948 Valid arguments: on, off
1949 Default: on
1950 Built with CONFIG_DEBUG_KMEMLEAK_DEFAULT_OFF=y,
1951 the default is off.
1953 kpti= [ARM64] Control page table isolation of user
1954 and kernel address spaces.
1955 Default: enabled on cores which need mitigation.
1956 0: force disabled
1957 1: force enabled
1959 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1960 Default is 0 (don't ignore, but inject #GP)
1962 kvm.enable_vmware_backdoor=[KVM] Support VMware backdoor PV interface.
1963 Default is false (don't support).
1965 kvm.mmu_audit= [KVM] This is a R/W parameter which allows audit
1966 KVM MMU at runtime.
1967 Default is 0 (off)
1969 kvm.nx_huge_pages=
1970 [KVM] Controls the software workaround for the
1971 X86_BUG_ITLB_MULTIHIT bug.
1972 force : Always deploy workaround.
1973 off : Never deploy workaround.
1974 auto : Deploy workaround based on the presence of
1975 X86_BUG_ITLB_MULTIHIT.
1977 Default is 'auto'.
1979 If the software workaround is enabled for the host,
1980 guests do need not to enable it for nested guests.
1982 kvm.nx_huge_pages_recovery_ratio=
1983 [KVM] Controls how many 4KiB pages are periodically zapped
1984 back to huge pages. 0 disables the recovery, otherwise if
1985 the value is N KVM will zap 1/Nth of the 4KiB pages every
1986 minute. The default is 60.
1988 kvm-amd.nested= [KVM,AMD] Allow nested virtualization in KVM/SVM.
1989 Default is 1 (enabled)
1991 kvm-amd.npt= [KVM,AMD] Disable nested paging (virtualized MMU)
1992 for all guests.
1993 Default is 1 (enabled) if in 64-bit or 32-bit PAE mode.
1995 kvm-arm.vgic_v3_group0_trap=
1996 [KVM,ARM] Trap guest accesses to GICv3 group-0
1997 system registers
1999 kvm-arm.vgic_v3_group1_trap=
2000 [KVM,ARM] Trap guest accesses to GICv3 group-1
2001 system registers
2003 kvm-arm.vgic_v3_common_trap=
2004 [KVM,ARM] Trap guest accesses to GICv3 common
2005 system registers
2007 kvm-arm.vgic_v4_enable=
2008 [KVM,ARM] Allow use of GICv4 for direct injection of
2009 LPIs.
2011 kvm-intel.ept= [KVM,Intel] Disable extended page tables
2012 (virtualized MMU) support on capable Intel chips.
2013 Default is 1 (enabled)
2015 kvm-intel.emulate_invalid_guest_state=
2016 [KVM,Intel] Enable emulation of invalid guest states
2017 Default is 0 (disabled)
2019 kvm-intel.flexpriority=
2020 [KVM,Intel] Disable FlexPriority feature (TPR shadow).
2021 Default is 1 (enabled)
2023 kvm-intel.nested=
2024 [KVM,Intel] Enable VMX nesting (nVMX).
2025 Default is 0 (disabled)
2027 kvm-intel.unrestricted_guest=
2028 [KVM,Intel] Disable unrestricted guest feature
2029 (virtualized real and unpaged mode) on capable
2030 Intel chips. Default is 1 (enabled)
2032 kvm-intel.vmentry_l1d_flush=[KVM,Intel] Mitigation for L1 Terminal Fault
2033 CVE-2018-3620.
2035 Valid arguments: never, cond, always
2037 always: L1D cache flush on every VMENTER.
2038 cond: Flush L1D on VMENTER only when the code between
2039 VMEXIT and VMENTER can leak host memory.
2040 never: Disables the mitigation
2042 Default is cond (do L1 cache flush in specific instances)
2044 kvm-intel.vpid= [KVM,Intel] Disable Virtual Processor Identification
2045 feature (tagged TLBs) on capable Intel chips.
2046 Default is 1 (enabled)
2048 l1tf= [X86] Control mitigation of the L1TF vulnerability on
2049 affected CPUs
2051 The kernel PTE inversion protection is unconditionally
2052 enabled and cannot be disabled.
2054 full
2055 Provides all available mitigations for the
2056 L1TF vulnerability. Disables SMT and
2057 enables all mitigations in the
2058 hypervisors, i.e. unconditional L1D flush.
2060 SMT control and L1D flush control via the
2061 sysfs interface is still possible after
2062 boot. Hypervisors will issue a warning
2063 when the first VM is started in a
2064 potentially insecure configuration,
2065 i.e. SMT enabled or L1D flush disabled.
2067 full,force
2068 Same as 'full', but disables SMT and L1D
2069 flush runtime control. Implies the
2070 'nosmt=force' command line option.
2071 (i.e. sysfs control of SMT is disabled.)
2073 flush
2074 Leaves SMT enabled and enables the default
2075 hypervisor mitigation, i.e. conditional
2076 L1D flush.
2078 SMT control and L1D flush control via the
2079 sysfs interface is still possible after
2080 boot. Hypervisors will issue a warning
2081 when the first VM is started in a
2082 potentially insecure configuration,
2083 i.e. SMT enabled or L1D flush disabled.
2085 flush,nosmt
2087 Disables SMT and enables the default
2088 hypervisor mitigation.
2090 SMT control and L1D flush control via the
2091 sysfs interface is still possible after
2092 boot. Hypervisors will issue a warning
2093 when the first VM is started in a
2094 potentially insecure configuration,
2095 i.e. SMT enabled or L1D flush disabled.
2097 flush,nowarn
2098 Same as 'flush', but hypervisors will not
2099 warn when a VM is started in a potentially
2100 insecure configuration.
2102 off
2103 Disables hypervisor mitigations and doesn't
2104 emit any warnings.
2105 It also drops the swap size and available
2106 RAM limit restriction on both hypervisor and
2107 bare metal.
2109 Default is 'flush'.
2111 For details see: Documentation/admin-guide/hw-vuln/l1tf.rst
2113 l2cr= [PPC]
2115 l3cr= [PPC]
2117 lapic [X86-32,APIC] Enable the local APIC even if BIOS
2118 disabled it.
2120 lapic= [x86,APIC] "notscdeadline" Do not use TSC deadline
2121 value for LAPIC timer one-shot implementation. Default
2122 back to the programmable timer unit in the LAPIC.
2124 lapic_timer_c2_ok [X86,APIC] trust the local apic timer
2125 in C2 power state.
2127 libata.dma= [LIBATA] DMA control
2128 libata.dma=0 Disable all PATA and SATA DMA
2129 libata.dma=1 PATA and SATA Disk DMA only
2130 libata.dma=2 ATAPI (CDROM) DMA only
2131 libata.dma=4 Compact Flash DMA only
2132 Combinations also work, so libata.dma=3 enables DMA
2133 for disks and CDROMs, but not CFs.
2135 libata.ignore_hpa= [LIBATA] Ignore HPA limit
2136 libata.ignore_hpa=0 keep BIOS limits (default)
2137 libata.ignore_hpa=1 ignore limits, using full disk
2139 libata.noacpi [LIBATA] Disables use of ACPI in libata suspend/resume
2140 when set.
2141 Format: <int>
2143 libata.force= [LIBATA] Force configurations. The format is comma
2144 separated list of "[ID:]VAL" where ID is
2145 PORT[.DEVICE]. PORT and DEVICE are decimal numbers
2146 matching port, link or device. Basically, it matches
2147 the ATA ID string printed on console by libata. If
2148 the whole ID part is omitted, the last PORT and DEVICE
2149 values are used. If ID hasn't been specified yet, the
2150 configuration applies to all ports, links and devices.
2152 If only DEVICE is omitted, the parameter applies to
2153 the port and all links and devices behind it. DEVICE
2154 number of 0 either selects the first device or the
2155 first fan-out link behind PMP device. It does not
2156 select the host link. DEVICE number of 15 selects the
2157 host link and device attached to it.
2159 The VAL specifies the configuration to force. As long
2160 as there's no ambiguity shortcut notation is allowed.
2161 For example, both 1.5 and 1.5G would work for 1.5Gbps.
2162 The following configurations can be forced.
2164 * Cable type: 40c, 80c, short40c, unk, ign or sata.
2165 Any ID with matching PORT is used.
2167 * SATA link speed limit: 1.5Gbps or 3.0Gbps.
2169 * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
2170 udma[/][16,25,33,44,66,100,133] notation is also
2171 allowed.
2173 * [no]ncq: Turn on or off NCQ.
2175 * [no]ncqtrim: Turn off queued DSM TRIM.
2177 * nohrst, nosrst, norst: suppress hard, soft
2178 and both resets.
2180 * rstonce: only attempt one reset during
2181 hot-unplug link recovery
2183 * dump_id: dump IDENTIFY data.
2185 * atapi_dmadir: Enable ATAPI DMADIR bridge support
2187 * disable: Disable this device.
2189 If there are multiple matching configurations changing
2190 the same attribute, the last one is used.
2192 memblock=debug [KNL] Enable memblock debug messages.
2194 load_ramdisk= [RAM] List of ramdisks to load from floppy
2195 See Documentation/blockdev/ramdisk.txt.
2197 lockd.nlm_grace_period=P [NFS] Assign grace period.
2198 Format: <integer>
2200 lockd.nlm_tcpport=N [NFS] Assign TCP port.
2201 Format: <integer>
2203 lockd.nlm_timeout=T [NFS] Assign timeout value.
2204 Format: <integer>
2206 lockd.nlm_udpport=M [NFS] Assign UDP port.
2207 Format: <integer>
2209 locktorture.nreaders_stress= [KNL]
2210 Set the number of locking read-acquisition kthreads.
2211 Defaults to being automatically set based on the
2212 number of online CPUs.
2214 locktorture.nwriters_stress= [KNL]
2215 Set the number of locking write-acquisition kthreads.
2217 locktorture.onoff_holdoff= [KNL]
2218 Set time (s) after boot for CPU-hotplug testing.
2220 locktorture.onoff_interval= [KNL]
2221 Set time (s) between CPU-hotplug operations, or
2222 zero to disable CPU-hotplug testing.
2224 locktorture.shuffle_interval= [KNL]
2225 Set task-shuffle interval (jiffies). Shuffling
2226 tasks allows some CPUs to go into dyntick-idle
2227 mode during the locktorture test.
2229 locktorture.shutdown_secs= [KNL]
2230 Set time (s) after boot system shutdown. This
2231 is useful for hands-off automated testing.
2233 locktorture.stat_interval= [KNL]
2234 Time (s) between statistics printk()s.
2236 locktorture.stutter= [KNL]
2237 Time (s) to stutter testing, for example,
2238 specifying five seconds causes the test to run for
2239 five seconds, wait for five seconds, and so on.
2240 This tests the locking primitive's ability to
2241 transition abruptly to and from idle.
2243 locktorture.torture_type= [KNL]
2244 Specify the locking implementation to test.
2246 locktorture.verbose= [KNL]
2247 Enable additional printk() statements.
2249 logibm.irq= [HW,MOUSE] Logitech Bus Mouse Driver
2250 Format: <irq>
2252 loglevel= All Kernel Messages with a loglevel smaller than the
2253 console loglevel will be printed to the console. It can
2254 also be changed with klogd or other programs. The
2255 loglevels are defined as follows:
2257 0 (KERN_EMERG) system is unusable
2258 1 (KERN_ALERT) action must be taken immediately
2259 2 (KERN_CRIT) critical conditions
2260 3 (KERN_ERR) error conditions
2261 4 (KERN_WARNING) warning conditions
2262 5 (KERN_NOTICE) normal but significant condition
2263 6 (KERN_INFO) informational
2264 7 (KERN_DEBUG) debug-level messages
2266 log_buf_len=n[KMG] Sets the size of the printk ring buffer,
2267 in bytes. n must be a power of two and greater
2268 than the minimal size. The minimal size is defined
2269 by LOG_BUF_SHIFT kernel config parameter. There is
2270 also CONFIG_LOG_CPU_MAX_BUF_SHIFT config parameter
2271 that allows to increase the default size depending on
2272 the number of CPUs. See init/Kconfig for more details.
2274 logo.nologo [FB] Disables display of the built-in Linux logo.
2275 This may be used to provide more screen space for
2276 kernel log messages and is useful when debugging
2277 kernel boot problems.
2279 lp=0 [LP] Specify parallel ports to use, e.g,
2280 lp=port[,port...] lp=none,parport0 (lp0 not configured, lp1 uses
2281 lp=reset first parallel port). 'lp=0' disables the
2282 lp=auto printer driver. 'lp=reset' (which can be
2283 specified in addition to the ports) causes
2284 attached printers to be reset. Using
2285 lp=port1,port2,... specifies the parallel ports
2286 to associate lp devices with, starting with
2287 lp0. A port specification may be 'none' to skip
2288 that lp device, or a parport name such as
2289 'parport0'. Specifying 'lp=auto' instead of a
2290 port specification list means that device IDs
2291 from each port should be examined, to see if
2292 an IEEE 1284-compliant printer is attached; if
2293 so, the driver will manage that printer.
2294 See also header of drivers/char/lp.c.
2296 lpj=n [KNL]
2297 Sets loops_per_jiffy to given constant, thus avoiding
2298 time-consuming boot-time autodetection (up to 250 ms per
2299 CPU). 0 enables autodetection (default). To determine
2300 the correct value for your kernel, boot with normal
2301 autodetection and see what value is printed. Note that
2302 on SMP systems the preset will be applied to all CPUs,
2303 which is likely to cause problems if your CPUs need
2304 significantly divergent settings. An incorrect value
2305 will cause delays in the kernel to be wrong, leading to
2306 unpredictable I/O errors and other breakage. Although
2307 unlikely, in the extreme case this might damage your
2308 hardware.
2310 ltpc= [NET]
2311 Format: <io>,<irq>,<dma>
2313 machvec= [IA-64] Force the use of a particular machine-vector
2314 (machvec) in a generic kernel.
2315 Example: machvec=hpzx1_swiotlb
2317 machtype= [Loongson] Share the same kernel image file between different
2318 yeeloong laptop.
2319 Example: machtype=lemote-yeeloong-2f-7inch
2321 max_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory greater
2322 than or equal to this physical address is ignored.
2324 maxcpus= [SMP] Maximum number of processors that an SMP kernel
2325 will bring up during bootup. maxcpus=n : n >= 0 limits
2326 the kernel to bring up 'n' processors. Surely after
2327 bootup you can bring up the other plugged cpu by executing
2328 "echo 1 > /sys/devices/system/cpu/cpuX/online". So maxcpus
2329 only takes effect during system bootup.
2330 While n=0 is a special case, it is equivalent to "nosmp",
2331 which also disables the IO APIC.
2333 max_loop= [LOOP] The number of loop block devices that get
2334 (loop.max_loop) unconditionally pre-created at init time. The default
2335 number is configured by BLK_DEV_LOOP_MIN_COUNT. Instead
2336 of statically allocating a predefined number, loop
2337 devices can be requested on-demand with the
2338 /dev/loop-control interface.
2340 mce [X86-32] Machine Check Exception
2342 mce=option [X86-64] See Documentation/x86/x86_64/boot-options.txt
2344 md= [HW] RAID subsystems devices and level
2345 See Documentation/admin-guide/md.rst.
2347 mdacon= [MDA]
2348 Format: <first>,<last>
2349 Specifies range of consoles to be captured by the MDA.
2351 mds= [X86,INTEL]
2352 Control mitigation for the Micro-architectural Data
2353 Sampling (MDS) vulnerability.
2355 Certain CPUs are vulnerable to an exploit against CPU
2356 internal buffers which can forward information to a
2357 disclosure gadget under certain conditions.
2359 In vulnerable processors, the speculatively
2360 forwarded data can be used in a cache side channel
2361 attack, to access data to which the attacker does
2362 not have direct access.
2364 This parameter controls the MDS mitigation. The
2365 options are:
2367 full - Enable MDS mitigation on vulnerable CPUs
2368 full,nosmt - Enable MDS mitigation and disable
2369 SMT on vulnerable CPUs
2370 off - Unconditionally disable MDS mitigation
2372 On TAA-affected machines, mds=off can be prevented by
2373 an active TAA mitigation as both vulnerabilities are
2374 mitigated with the same mechanism so in order to disable
2375 this mitigation, you need to specify tsx_async_abort=off
2376 too.
2378 Not specifying this option is equivalent to
2379 mds=full.
2381 For details see: Documentation/admin-guide/hw-vuln/mds.rst
2383 mem=nn[KMG] [KNL,BOOT] Force usage of a specific amount of memory
2384 Amount of memory to be used when the kernel is not able
2385 to see the whole system memory or for test.
2386 [X86] Work as limiting max address. Use together
2387 with memmap= to avoid physical address space collisions.
2388 Without memmap= PCI devices could be placed at addresses
2389 belonging to unused RAM.
2391 mem=nopentium [BUGS=X86-32] Disable usage of 4MB pages for kernel
2392 memory.
2394 memchunk=nn[KMG]
2395 [KNL,SH] Allow user to override the default size for
2396 per-device physically contiguous DMA buffers.
2398 memhp_default_state=online/offline
2399 [KNL] Set the initial state for the memory hotplug
2400 onlining policy. If not specified, the default value is
2401 set according to the
2402 CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE kernel config
2403 option.
2404 See Documentation/memory-hotplug.txt.
2406 memmap=exactmap [KNL,X86] Enable setting of an exact
2407 E820 memory map, as specified by the user.
2408 Such memmap=exactmap lines can be constructed based on
2409 BIOS output or other requirements. See the memmap=nn@ss
2410 option description.
2412 memmap=nn[KMG]@ss[KMG]
2413 [KNL] Force usage of a specific region of memory.
2414 Region of memory to be used is from ss to ss+nn.
2415 If @ss[KMG] is omitted, it is equivalent to mem=nn[KMG],
2416 which limits max address to nn[KMG].
2417 Multiple different regions can be specified,
2418 comma delimited.
2419 Example:
2420 memmap=100M@2G,100M#3G,1G!1024G
2422 memmap=nn[KMG]#ss[KMG]
2423 [KNL,ACPI] Mark specific memory as ACPI data.
2424 Region of memory to be marked is from ss to ss+nn.
2426 memmap=nn[KMG]$ss[KMG]
2427 [KNL,ACPI] Mark specific memory as reserved.
2428 Region of memory to be reserved is from ss to ss+nn.
2429 Example: Exclude memory from 0x18690000-0x1869ffff
2430 memmap=64K$0x18690000
2431 or
2432 memmap=0x10000$0x18690000
2433 Some bootloaders may need an escape character before '$',
2434 like Grub2, otherwise '$' and the following number
2435 will be eaten.
2437 memmap=nn[KMG]!ss[KMG]
2438 [KNL,X86] Mark specific memory as protected.
2439 Region of memory to be used, from ss to ss+nn.
2440 The memory region may be marked as e820 type 12 (0xc)
2441 and is NVDIMM or ADR memory.
2443 memmap=<size>%<offset>-<oldtype>+<newtype>
2444 [KNL,ACPI] Convert memory within the specified region
2445 from <oldtype> to <newtype>. If "-<oldtype>" is left
2446 out, the whole region will be marked as <newtype>,
2447 even if previously unavailable. If "+<newtype>" is left
2448 out, matching memory will be removed. Types are
2449 specified as e820 types, e.g., 1 = RAM, 2 = reserved,
2450 3 = ACPI, 12 = PRAM.
2452 memory_corruption_check=0/1 [X86]
2453 Some BIOSes seem to corrupt the first 64k of
2454 memory when doing things like suspend/resume.
2455 Setting this option will scan the memory
2456 looking for corruption. Enabling this will
2457 both detect corruption and prevent the kernel
2458 from using the memory being corrupted.
2459 However, its intended as a diagnostic tool; if
2460 repeatable BIOS-originated corruption always
2461 affects the same memory, you can use memmap=
2462 to prevent the kernel from using that memory.
2464 memory_corruption_check_size=size [X86]
2465 By default it checks for corruption in the low
2466 64k, making this memory unavailable for normal
2467 use. Use this parameter to scan for
2468 corruption in more or less memory.
2470 memory_corruption_check_period=seconds [X86]
2471 By default it checks for corruption every 60
2472 seconds. Use this parameter to check at some
2473 other rate. 0 disables periodic checking.
2475 memtest= [KNL,X86,ARM] Enable memtest
2476 Format: <integer>
2477 default : 0 <disable>
2478 Specifies the number of memtest passes to be
2479 performed. Each pass selects another test
2480 pattern from a given set of patterns. Memtest
2481 fills the memory with this pattern, validates
2482 memory contents and reserves bad memory
2483 regions that are detected.
2485 mem_encrypt= [X86-64] AMD Secure Memory Encryption (SME) control
2486 Valid arguments: on, off
2487 Default (depends on kernel configuration option):
2488 on (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=y)
2489 off (CONFIG_AMD_MEM_ENCRYPT_ACTIVE_BY_DEFAULT=n)
2490 mem_encrypt=on: Activate SME
2491 mem_encrypt=off: Do not activate SME
2493 Refer to Documentation/x86/amd-memory-encryption.txt
2494 for details on when memory encryption can be activated.
2496 mem_sleep_default= [SUSPEND] Default system suspend mode:
2497 s2idle - Suspend-To-Idle
2498 shallow - Power-On Suspend or equivalent (if supported)
2499 deep - Suspend-To-RAM or equivalent (if supported)
2500 See Documentation/admin-guide/pm/sleep-states.rst.
2502 meye.*= [HW] Set MotionEye Camera parameters
2503 See Documentation/media/v4l-drivers/meye.rst.
2505 mfgpt_irq= [IA-32] Specify the IRQ to use for the
2506 Multi-Function General Purpose Timers on AMD Geode
2507 platforms.
2509 mfgptfix [X86-32] Fix MFGPT timers on AMD Geode platforms when
2510 the BIOS has incorrectly applied a workaround. TinyBIOS
2511 version 0.98 is known to be affected, 0.99 fixes the
2512 problem by letting the user disable the workaround.
2514 mga= [HW,DRM]
2516 min_addr=nn[KMG] [KNL,BOOT,ia64] All physical memory below this
2517 physical address is ignored.
2519 mini2440= [ARM,HW,KNL]
2520 Format:[0..2][b][c][t]
2521 Default: "0tb"
2522 MINI2440 configuration specification:
2523 0 - The attached screen is the 3.5" TFT
2524 1 - The attached screen is the 7" TFT
2525 2 - The VGA Shield is attached (1024x768)
2526 Leaving out the screen size parameter will not load
2527 the TFT driver, and the framebuffer will be left
2528 unconfigured.
2529 b - Enable backlight. The TFT backlight pin will be
2530 linked to the kernel VESA blanking code and a GPIO
2531 LED. This parameter is not necessary when using the
2532 VGA shield.
2533 c - Enable the s3c camera interface.
2534 t - Reserved for enabling touchscreen support. The
2535 touchscreen support is not enabled in the mainstream
2536 kernel as of 2.6.30, a preliminary port can be found
2537 in the "bleeding edge" mini2440 support kernel at
2538 http://repo.or.cz/w/linux-2.6/mini2440.git
2540 mitigations=
2541 [X86,PPC,S390,ARM64] Control optional mitigations for
2542 CPU vulnerabilities. This is a set of curated,
2543 arch-independent options, each of which is an
2544 aggregation of existing arch-specific options.
2546 off
2547 Disable all optional CPU mitigations. This
2548 improves system performance, but it may also
2549 expose users to several CPU vulnerabilities.
2550 Equivalent to: nopti [X86,PPC]
2551 kpti=0 [ARM64]
2552 nospectre_v1 [PPC]
2553 nobp=0 [S390]
2554 nospectre_v1 [X86]
2555 nospectre_v2 [X86,PPC,S390,ARM64]
2556 spectre_v2_user=off [X86]
2557 spec_store_bypass_disable=off [X86,PPC]
2558 ssbd=force-off [ARM64]
2559 l1tf=off [X86]
2560 mds=off [X86]
2561 tsx_async_abort=off [X86]
2562 kvm.nx_huge_pages=off [X86]
2564 Exceptions:
2565 This does not have any effect on
2566 kvm.nx_huge_pages when
2567 kvm.nx_huge_pages=force.
2569 auto (default)
2570 Mitigate all CPU vulnerabilities, but leave SMT
2571 enabled, even if it's vulnerable. This is for
2572 users who don't want to be surprised by SMT
2573 getting disabled across kernel upgrades, or who
2574 have other ways of avoiding SMT-based attacks.
2575 Equivalent to: (default behavior)
2577 auto,nosmt
2578 Mitigate all CPU vulnerabilities, disabling SMT
2579 if needed. This is for users who always want to
2580 be fully mitigated, even if it means losing SMT.
2581 Equivalent to: l1tf=flush,nosmt [X86]
2582 mds=full,nosmt [X86]
2583 tsx_async_abort=full,nosmt [X86]
2585 mminit_loglevel=
2586 [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
2587 parameter allows control of the logging verbosity for
2588 the additional memory initialisation checks. A value
2589 of 0 disables mminit logging and a level of 4 will
2590 log everything. Information is printed at KERN_DEBUG
2591 so loglevel=8 may also need to be specified.
2593 module.sig_enforce
2594 [KNL] When CONFIG_MODULE_SIG is set, this means that
2595 modules without (valid) signatures will fail to load.
2596 Note that if CONFIG_MODULE_SIG_FORCE is set, that
2597 is always true, so this option does nothing.
2599 module_blacklist= [KNL] Do not load a comma-separated list of
2600 modules. Useful for debugging problem modules.
2602 mousedev.tap_time=
2603 [MOUSE] Maximum time between finger touching and
2604 leaving touchpad surface for touch to be considered
2605 a tap and be reported as a left button click (for
2606 touchpads working in absolute mode only).
2607 Format: <msecs>
2608 mousedev.xres= [MOUSE] Horizontal screen resolution, used for devices
2609 reporting absolute coordinates, such as tablets
2610 mousedev.yres= [MOUSE] Vertical screen resolution, used for devices
2611 reporting absolute coordinates, such as tablets
2613 movablecore= [KNL,X86,IA-64,PPC]
2614 Format: nn[KMGTPE] | nn%
2615 This parameter is the complement to kernelcore=, it
2616 specifies the amount of memory used for migratable
2617 allocations. If both kernelcore and movablecore is
2618 specified, then kernelcore will be at *least* the
2619 specified value but may be more. If movablecore on its
2620 own is specified, the administrator must be careful
2621 that the amount of memory usable for all allocations
2622 is not too small.
2624 movable_node [KNL] Boot-time switch to make hotplugable memory
2625 NUMA nodes to be movable. This means that the memory
2626 of such nodes will be usable only for movable
2627 allocations which rules out almost all kernel
2628 allocations. Use with caution!
2630 MTD_Partition= [MTD]
2631 Format: <name>,<region-number>,<size>,<offset>
2633 MTD_Region= [MTD] Format:
2634 <name>,<region-number>[,<base>,<size>,<buswidth>,<altbuswidth>]
2636 mtdparts= [MTD]
2637 See drivers/mtd/cmdlinepart.c.
2639 multitce=off [PPC] This parameter disables the use of the pSeries
2640 firmware feature for updating multiple TCE entries
2641 at a time.
2643 onenand.bdry= [HW,MTD] Flex-OneNAND Boundary Configuration
2645 Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]
2647 boundary - index of last SLC block on Flex-OneNAND.
2648 The remaining blocks are configured as MLC blocks.
2649 lock - Configure if Flex-OneNAND boundary should be locked.
2650 Once locked, the boundary cannot be changed.
2651 1 indicates lock status, 0 indicates unlock status.
2653 mtdset= [ARM]
2654 ARM/S3C2412 JIVE boot control
2656 See arch/arm/mach-s3c2412/mach-jive.c
2658 mtouchusb.raw_coordinates=
2659 [HW] Make the MicroTouch USB driver use raw coordinates
2660 ('y', default) or cooked coordinates ('n')
2662 mtrr_chunk_size=nn[KMG] [X86]
2663 used for mtrr cleanup. It is largest continuous chunk
2664 that could hold holes aka. UC entries.
2666 mtrr_gran_size=nn[KMG] [X86]
2667 Used for mtrr cleanup. It is granularity of mtrr block.
2668 Default is 1.
2669 Large value could prevent small alignment from
2670 using up MTRRs.
2672 mtrr_spare_reg_nr=n [X86]
2673 Format: <integer>
2674 Range: 0,7 : spare reg number
2675 Default : 1
2676 Used for mtrr cleanup. It is spare mtrr entries number.
2677 Set to 2 or more if your graphical card needs more.
2679 n2= [NET] SDL Inc. RISCom/N2 synchronous serial card
2681 netdev= [NET] Network devices parameters
2682 Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
2683 Note that mem_start is often overloaded to mean
2684 something different and driver-specific.
2685 This usage is only documented in each driver source
2686 file if at all.
2688 nf_conntrack.acct=
2689 [NETFILTER] Enable connection tracking flow accounting
2690 0 to disable accounting
2691 1 to enable accounting
2692 Default value is 0.
2694 nfsaddrs= [NFS] Deprecated. Use ip= instead.
2695 See Documentation/filesystems/nfs/nfsroot.txt.
2697 nfsroot= [NFS] nfs root filesystem for disk-less boxes.
2698 See Documentation/filesystems/nfs/nfsroot.txt.
2700 nfsrootdebug [NFS] enable nfsroot debugging messages.
2701 See Documentation/filesystems/nfs/nfsroot.txt.
2703 nfs.callback_nr_threads=
2704 [NFSv4] set the total number of threads that the
2705 NFS client will assign to service NFSv4 callback
2706 requests.
2708 nfs.callback_tcpport=
2709 [NFS] set the TCP port on which the NFSv4 callback
2710 channel should listen.
2712 nfs.cache_getent=
2713 [NFS] sets the pathname to the program which is used
2714 to update the NFS client cache entries.
2716 nfs.cache_getent_timeout=
2717 [NFS] sets the timeout after which an attempt to
2718 update a cache entry is deemed to have failed.
2720 nfs.idmap_cache_timeout=
2721 [NFS] set the maximum lifetime for idmapper cache
2722 entries.
2724 nfs.enable_ino64=
2725 [NFS] enable 64-bit inode numbers.
2726 If zero, the NFS client will fake up a 32-bit inode
2727 number for the readdir() and stat() syscalls instead
2728 of returning the full 64-bit number.
2729 The default is to return 64-bit inode numbers.
2731 nfs.max_session_cb_slots=
2732 [NFSv4.1] Sets the maximum number of session
2733 slots the client will assign to the callback
2734 channel. This determines the maximum number of
2735 callbacks the client will process in parallel for
2736 a particular server.
2738 nfs.max_session_slots=
2739 [NFSv4.1] Sets the maximum number of session slots
2740 the client will attempt to negotiate with the server.
2741 This limits the number of simultaneous RPC requests
2742 that the client can send to the NFSv4.1 server.
2743 Note that there is little point in setting this
2744 value higher than the max_tcp_slot_table_limit.
2746 nfs.nfs4_disable_idmapping=
2747 [NFSv4] When set to the default of '1', this option
2748 ensures that both the RPC level authentication
2749 scheme and the NFS level operations agree to use
2750 numeric uids/gids if the mount is using the
2751 'sec=sys' security flavour. In effect it is
2752 disabling idmapping, which can make migration from
2753 legacy NFSv2/v3 systems to NFSv4 easier.
2754 Servers that do not support this mode of operation
2755 will be autodetected by the client, and it will fall
2756 back to using the idmapper.
2757 To turn off this behaviour, set the value to '0'.
2758 nfs.nfs4_unique_id=
2759 [NFS4] Specify an additional fixed unique ident-
2760 ification string that NFSv4 clients can insert into
2761 their nfs_client_id4 string. This is typically a
2762 UUID that is generated at system install time.
2764 nfs.send_implementation_id =
2765 [NFSv4.1] Send client implementation identification
2766 information in exchange_id requests.
2767 If zero, no implementation identification information
2768 will be sent.
2769 The default is to send the implementation identification
2770 information.
2772 nfs.recover_lost_locks =
2773 [NFSv4] Attempt to recover locks that were lost due
2774 to a lease timeout on the server. Please note that
2775 doing this risks data corruption, since there are
2776 no guarantees that the file will remain unchanged
2777 after the locks are lost.
2778 If you want to enable the kernel legacy behaviour of
2779 attempting to recover these locks, then set this
2780 parameter to '1'.
2781 The default parameter value of '0' causes the kernel
2782 not to attempt recovery of lost locks.
2784 nfs4.layoutstats_timer =
2785 [NFSv4.2] Change the rate at which the kernel sends
2786 layoutstats to the pNFS metadata server.
2788 Setting this to value to 0 causes the kernel to use
2789 whatever value is the default set by the layout
2790 driver. A non-zero value sets the minimum interval
2791 in seconds between layoutstats transmissions.
2793 nfsd.nfs4_disable_idmapping=
2794 [NFSv4] When set to the default of '1', the NFSv4
2795 server will return only numeric uids and gids to
2796 clients using auth_sys, and will accept numeric uids
2797 and gids from such clients. This is intended to ease
2798 migration from NFSv2/v3.
2800 nmi_debug= [KNL,SH] Specify one or more actions to take
2801 when a NMI is triggered.
2802 Format: [state][,regs][,debounce][,die]
2804 nmi_watchdog= [KNL,BUGS=X86] Debugging features for SMP kernels
2805 Format: [panic,][nopanic,][num]
2806 Valid num: 0 or 1
2807 0 - turn hardlockup detector in nmi_watchdog off
2808 1 - turn hardlockup detector in nmi_watchdog on
2809 When panic is specified, panic when an NMI watchdog
2810 timeout occurs (or 'nopanic' to override the opposite
2811 default). To disable both hard and soft lockup detectors,
2812 please see 'nowatchdog'.
2813 This is useful when you use a panic=... timeout and
2814 need the box quickly up again.
2816 These settings can be accessed at runtime via
2817 the nmi_watchdog and hardlockup_panic sysctls.
2819 netpoll.carrier_timeout=
2820 [NET] Specifies amount of time (in seconds) that
2821 netpoll should wait for a carrier. By default netpoll
2822 waits 4 seconds.
2824 no387 [BUGS=X86-32] Tells the kernel to use the 387 maths
2825 emulation library even if a 387 maths coprocessor
2826 is present.
2828 no5lvl [X86-64] Disable 5-level paging mode. Forces
2829 kernel to use 4-level paging instead.
2831 no_console_suspend
2832 [HW] Never suspend the console
2833 Disable suspending of consoles during suspend and
2834 hibernate operations. Once disabled, debugging
2835 messages can reach various consoles while the rest
2836 of the system is being put to sleep (ie, while
2837 debugging driver suspend/resume hooks). This may
2838 not work reliably with all consoles, but is known
2839 to work with serial and VGA consoles.
2840 To facilitate more flexible debugging, we also add
2841 console_suspend, a printk module parameter to control
2842 it. Users could use console_suspend (usually
2843 /sys/module/printk/parameters/console_suspend) to
2844 turn on/off it dynamically.
2846 noaliencache [MM, NUMA, SLAB] Disables the allocation of alien
2847 caches in the slab allocator. Saves per-node memory,
2848 but will impact performance.
2850 noalign [KNL,ARM]
2852 noaltinstr [S390] Disables alternative instructions patching
2853 (CPU alternatives feature).
2855 noapic [SMP,APIC] Tells the kernel to not make use of any
2856 IOAPICs that may be present in the system.
2858 noautogroup Disable scheduler automatic task group creation.
2860 nobats [PPC] Do not use BATs for mapping kernel lowmem
2861 on "Classic" PPC cores.
2863 nocache [ARM]
2865 noclflush [BUGS=X86] Don't use the CLFLUSH instruction
2867 nodelayacct [KNL] Disable per-task delay accounting
2869 nodsp [SH] Disable hardware DSP at boot time.
2871 noefi Disable EFI runtime services support.
2873 noexec [IA-64]
2875 noexec [X86]
2876 On X86-32 available only on PAE configured kernels.
2877 noexec=on: enable non-executable mappings (default)
2878 noexec=off: disable non-executable mappings
2880 nosmap [X86]
2881 Disable SMAP (Supervisor Mode Access Prevention)
2882 even if it is supported by processor.
2884 nosmep [X86]
2885 Disable SMEP (Supervisor Mode Execution Prevention)
2886 even if it is supported by processor.
2888 noexec32 [X86-64]
2889 This affects only 32-bit executables.
2890 noexec32=on: enable non-executable mappings (default)
2891 read doesn't imply executable mappings
2892 noexec32=off: disable non-executable mappings
2893 read implies executable mappings
2895 nofpu [MIPS,SH] Disable hardware FPU at boot time.
2897 nofxsr [BUGS=X86-32] Disables x86 floating point extended
2898 register save and restore. The kernel will only save
2899 legacy floating-point registers on task switch.
2901 nohugeiomap [KNL,x86] Disable kernel huge I/O mappings.
2903 nosmt [KNL,S390] Disable symmetric multithreading (SMT).
2904 Equivalent to smt=1.
2906 [KNL,x86] Disable symmetric multithreading (SMT).
2907 nosmt=force: Force disable SMT, cannot be undone
2908 via the sysfs control file.
2910 nospectre_v1 [X66, PPC] Disable mitigations for Spectre Variant 1
2911 (bounds check bypass). With this option data leaks
2912 are possible in the system.
2914 nospectre_v2 [X86,PPC_FSL_BOOK3E,ARM64] Disable all mitigations for
2915 the Spectre variant 2 (indirect branch prediction)
2916 vulnerability. System may allow data leaks with this
2917 option.
2919 nospec_store_bypass_disable
2920 [HW] Disable all mitigations for the Speculative Store Bypass vulnerability
2922 noxsave [BUGS=X86] Disables x86 extended register state save
2923 and restore using xsave. The kernel will fallback to
2924 enabling legacy floating-point and sse state.
2926 noxsaveopt [X86] Disables xsaveopt used in saving x86 extended
2927 register states. The kernel will fall back to use
2928 xsave to save the states. By using this parameter,
2929 performance of saving the states is degraded because
2930 xsave doesn't support modified optimization while
2931 xsaveopt supports it on xsaveopt enabled systems.
2933 noxsaves [X86] Disables xsaves and xrstors used in saving and
2934 restoring x86 extended register state in compacted
2935 form of xsave area. The kernel will fall back to use
2936 xsaveopt and xrstor to save and restore the states
2937 in standard form of xsave area. By using this
2938 parameter, xsave area per process might occupy more
2939 memory on xsaves enabled systems.
2941 nohlt [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
2942 wfi(ARM) instruction doesn't work correctly and not to
2943 use it. This is also useful when using JTAG debugger.
2945 no_file_caps Tells the kernel not to honor file capabilities. The
2946 only way then for a file to be executed with privilege
2947 is to be setuid root or executed by root.
2949 nohalt [IA-64] Tells the kernel not to use the power saving
2950 function PAL_HALT_LIGHT when idle. This increases
2951 power-consumption. On the positive side, it reduces
2952 interrupt wake-up latency, which may improve performance
2953 in certain environments such as networked servers or
2954 real-time systems.
2956 nohibernate [HIBERNATION] Disable hibernation and resume.
2958 nohz= [KNL] Boottime enable/disable dynamic ticks
2959 Valid arguments: on, off
2960 Default: on
2962 nohz_full= [KNL,BOOT,SMP,ISOL]
2963 The argument is a cpu list, as described above.
2964 In kernels built with CONFIG_NO_HZ_FULL=y, set
2965 the specified list of CPUs whose tick will be stopped
2966 whenever possible. The boot CPU will be forced outside
2967 the range to maintain the timekeeping. Any CPUs
2968 in this list will have their RCU callbacks offloaded,
2969 just as if they had also been called out in the
2970 rcu_nocbs= boot parameter.
2972 noiotrap [SH] Disables trapped I/O port accesses.
2974 noirqdebug [X86-32] Disables the code which attempts to detect and
2975 disable unhandled interrupt sources.
2977 no_timer_check [X86,APIC] Disables the code which tests for
2978 broken timer IRQ sources.
2980 noisapnp [ISAPNP] Disables ISA PnP code.
2982 noinitrd [RAM] Tells the kernel not to load any configured
2983 initial RAM disk.
2985 nointremap [X86-64, Intel-IOMMU] Do not enable interrupt
2986 remapping.
2987 [Deprecated - use intremap=off]
2989 nointroute [IA-64]
2991 noinvpcid [X86] Disable the INVPCID cpu feature.
2993 nojitter [IA-64] Disables jitter checking for ITC timers.
2995 no-kvmclock [X86,KVM] Disable paravirtualized KVM clock driver
2997 no-kvmapf [X86,KVM] Disable paravirtualized asynchronous page
2998 fault handling.
3000 no-vmw-sched-clock
3001 [X86,PV_OPS] Disable paravirtualized VMware scheduler
3002 clock and use the default one.
3004 no-steal-acc [X86,KVM] Disable paravirtualized steal time accounting.
3005 steal time is computed, but won't influence scheduler
3006 behaviour
3008 nolapic [X86-32,APIC] Do not enable or use the local APIC.
3010 nolapic_timer [X86-32,APIC] Do not use the local APIC timer.
3012 noltlbs [PPC] Do not use large page/tlb entries for kernel
3013 lowmem mapping on PPC40x and PPC8xx
3015 nomca [IA-64] Disable machine check abort handling
3017 nomce [X86-32] Disable Machine Check Exception
3019 nomfgpt [X86-32] Disable Multi-Function General Purpose
3020 Timer usage (for AMD Geode machines).
3022 nonmi_ipi [X86] Disable using NMI IPIs during panic/reboot to
3023 shutdown the other cpus. Instead use the REBOOT_VECTOR
3024 irq.
3026 nomodule Disable module load
3028 nopat [X86] Disable PAT (page attribute table extension of
3029 pagetables) support.
3031 nopcid [X86-64] Disable the PCID cpu feature.
3033 norandmaps Don't use address space randomization. Equivalent to
3034 echo 0 > /proc/sys/kernel/randomize_va_space
3036 noreplace-smp [X86-32,SMP] Don't replace SMP instructions
3037 with UP alternatives
3039 nordrand [X86] Disable kernel use of the RDRAND and
3040 RDSEED instructions even if they are supported
3041 by the processor. RDRAND and RDSEED are still
3042 available to user space applications.
3044 noresume [SWSUSP] Disables resume and restores original swap
3045 space.
3047 no-scroll [VGA] Disables scrollback.
3048 This is required for the Braillex ib80-piezo Braille
3049 reader made by F.H. Papenmeier (Germany).
3051 nosbagart [IA-64]
3053 nosep [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.
3055 nosmp [SMP] Tells an SMP kernel to act as a UP kernel,
3056 and disable the IO APIC. legacy for "maxcpus=0".
3058 nosoftlockup [KNL] Disable the soft-lockup detector.
3060 nosync [HW,M68K] Disables sync negotiation for all devices.
3062 nowatchdog [KNL] Disable both lockup detectors, i.e.
3063 soft-lockup and NMI watchdog (hard-lockup).
3065 nowb [ARM]
3067 nox2apic [X86-64,APIC] Do not enable x2APIC mode.
3069 cpu0_hotplug [X86] Turn on CPU0 hotplug feature when
3070 CONFIG_BOOTPARAM_HOTPLUG_CPU0 is off.
3071 Some features depend on CPU0. Known dependencies are:
3072 1. Resume from suspend/hibernate depends on CPU0.
3073 Suspend/hibernate will fail if CPU0 is offline and you
3074 need to online CPU0 before suspend/hibernate.
3075 2. PIC interrupts also depend on CPU0. CPU0 can't be
3076 removed if a PIC interrupt is detected.
3077 It's said poweroff/reboot may depend on CPU0 on some
3078 machines although I haven't seen such issues so far
3079 after CPU0 is offline on a few tested machines.
3080 If the dependencies are under your control, you can
3081 turn on cpu0_hotplug.
3083 nps_mtm_hs_ctr= [KNL,ARC]
3084 This parameter sets the maximum duration, in
3085 cycles, each HW thread of the CTOP can run
3086 without interruptions, before HW switches it.
3087 The actual maximum duration is 16 times this
3088 parameter's value.
3089 Format: integer between 1 and 255
3090 Default: 255
3092 nptcg= [IA-64] Override max number of concurrent global TLB
3093 purges which is reported from either PAL_VM_SUMMARY or
3094 SAL PALO.
3096 nr_cpus= [SMP] Maximum number of processors that an SMP kernel
3097 could support. nr_cpus=n : n >= 1 limits the kernel to
3098 support 'n' processors. It could be larger than the
3099 number of already plugged CPU during bootup, later in
3100 runtime you can physically add extra cpu until it reaches
3101 n. So during boot up some boot time memory for per-cpu
3102 variables need be pre-allocated for later physical cpu
3103 hot plugging.
3105 nr_uarts= [SERIAL] maximum number of UARTs to be registered.
3107 numa_balancing= [KNL,X86] Enable or disable automatic NUMA balancing.
3108 Allowed values are enable and disable
3110 numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
3111 'node', 'default' can be specified
3112 This can be set from sysctl after boot.
3113 See Documentation/sysctl/vm.txt for details.
3115 ohci1394_dma=early [HW] enable debugging via the ohci1394 driver.
3116 See Documentation/debugging-via-ohci1394.txt for more
3117 info.
3119 olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
3120 Rather than timing out after 20 ms if an EC
3121 command is not properly ACKed, override the length
3122 of the timeout. We have interrupts disabled while
3123 waiting for the ACK, so if this is set too high
3124 interrupts *may* be lost!
3126 omap_mux= [OMAP] Override bootloader pin multiplexing.
3127 Format: <mux_mode0.mode_name=value>...
3128 For example, to override I2C bus2:
3129 omap_mux=i2c2_scl.i2c2_scl=0x100,i2c2_sda.i2c2_sda=0x100
3131 oprofile.timer= [HW]
3132 Use timer interrupt instead of performance counters
3134 oprofile.cpu_type= Force an oprofile cpu type
3135 This might be useful if you have an older oprofile
3136 userland or if you want common events.
3137 Format: { arch_perfmon }
3138 arch_perfmon: [X86] Force use of architectural
3139 perfmon on Intel CPUs instead of the
3140 CPU specific event set.
3141 timer: [X86] Force use of architectural NMI
3142 timer mode (see also oprofile.timer
3143 for generic hr timer mode)
3145 oops=panic Always panic on oopses. Default is to just kill the
3146 process, but there is a small probability of
3147 deadlocking the machine.
3148 This will also cause panics on machine check exceptions.
3149 Useful together with panic=30 to trigger a reboot.
3151 page_owner= [KNL] Boot-time page_owner enabling option.
3152 Storage of the information about who allocated
3153 each page is disabled in default. With this switch,
3154 we can turn it on.
3155 on: enable the feature
3157 page_poison= [KNL] Boot-time parameter changing the state of
3158 poisoning on the buddy allocator, available with
3159 CONFIG_PAGE_POISONING=y.
3160 off: turn off poisoning (default)
3161 on: turn on poisoning
3163 panic= [KNL] Kernel behaviour on panic: delay <timeout>
3164 timeout > 0: seconds before rebooting
3165 timeout = 0: wait forever
3166 timeout < 0: reboot immediately
3167 Format: <timeout>
3169 panic_on_warn panic() instead of WARN(). Useful to cause kdump
3170 on a WARN().
3172 crash_kexec_post_notifiers
3173 Run kdump after running panic-notifiers and dumping
3174 kmsg. This only for the users who doubt kdump always
3175 succeeds in any situation.
3176 Note that this also increases risks of kdump failure,
3177 because some panic notifiers can make the crashed
3178 kernel more unstable.
3180 parkbd.port= [HW] Parallel port number the keyboard adapter is
3181 connected to, default is 0.
3182 Format: <parport#>
3183 parkbd.mode= [HW] Parallel port keyboard adapter mode of operation,
3184 0 for XT, 1 for AT (default is AT).
3185 Format: <mode>
3187 parport= [HW,PPT] Specify parallel ports. 0 disables.
3188 Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
3189 Use 'auto' to force the driver to use any
3190 IRQ/DMA settings detected (the default is to
3191 ignore detected IRQ/DMA settings because of
3192 possible conflicts). You can specify the base
3193 address, IRQ, and DMA settings; IRQ and DMA
3194 should be numbers, or 'auto' (for using detected
3195 settings on that particular port), or 'nofifo'
3196 (to avoid using a FIFO even if it is detected).
3197 Parallel ports are assigned in the order they
3198 are specified on the command line, starting
3199 with parport0.
3201 parport_init_mode= [HW,PPT]
3202 Configure VIA parallel port to operate in
3203 a specific mode. This is necessary on Pegasos
3204 computer where firmware has no options for setting
3205 up parallel port mode and sets it to spp.
3206 Currently this function knows 686a and 8231 chips.
3207 Format: [spp|ps2|epp|ecp|ecpepp]
3209 pause_on_oops=
3210 Halt all CPUs after the first oops has been printed for
3211 the specified number of seconds. This is to be used if
3212 your oopses keep scrolling off the screen.
3214 pcbit= [HW,ISDN]
3216 pcd. [PARIDE]
3217 See header of drivers/block/paride/pcd.c.
3218 See also Documentation/blockdev/paride.txt.
3220 pci=option[,option...] [PCI] various PCI subsystem options.
3222 Some options herein operate on a specific device
3223 or a set of devices (<pci_dev>). These are
3224 specified in one of the following formats:
3226 [<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*
3227 pci:<vendor>:<device>[:<subvendor>:<subdevice>]
3229 Note: the first format specifies a PCI
3230 bus/device/function address which may change
3231 if new hardware is inserted, if motherboard
3232 firmware changes, or due to changes caused
3233 by other kernel parameters. If the
3234 domain is left unspecified, it is
3235 taken to be zero. Optionally, a path
3236 to a device through multiple device/function
3237 addresses can be specified after the base
3238 address (this is more robust against
3239 renumbering issues). The second format
3240 selects devices using IDs from the
3241 configuration space which may match multiple
3242 devices in the system.
3244 earlydump dump PCI config space before the kernel
3245 changes anything
3246 off [X86] don't probe for the PCI bus
3247 bios [X86-32] force use of PCI BIOS, don't access
3248 the hardware directly. Use this if your machine
3249 has a non-standard PCI host bridge.
3250 nobios [X86-32] disallow use of PCI BIOS, only direct
3251 hardware access methods are allowed. Use this
3252 if you experience crashes upon bootup and you
3253 suspect they are caused by the BIOS.
3254 conf1 [X86] Force use of PCI Configuration Access
3255 Mechanism 1 (config address in IO port 0xCF8,
3256 data in IO port 0xCFC, both 32-bit).
3257 conf2 [X86] Force use of PCI Configuration Access
3258 Mechanism 2 (IO port 0xCF8 is an 8-bit port for
3259 the function, IO port 0xCFA, also 8-bit, sets
3260 bus number. The config space is then accessed
3261 through ports 0xC000-0xCFFF).
3262 See http://wiki.osdev.org/PCI for more info
3263 on the configuration access mechanisms.
3264 noaer [PCIE] If the PCIEAER kernel config parameter is
3265 enabled, this kernel boot option can be used to
3266 disable the use of PCIE advanced error reporting.
3267 nodomains [PCI] Disable support for multiple PCI
3268 root domains (aka PCI segments, in ACPI-speak).
3269 nommconf [X86] Disable use of MMCONFIG for PCI
3270 Configuration
3271 check_enable_amd_mmconf [X86] check for and enable
3272 properly configured MMIO access to PCI
3273 config space on AMD family 10h CPU
3274 nomsi [MSI] If the PCI_MSI kernel config parameter is
3275 enabled, this kernel boot option can be used to
3276 disable the use of MSI interrupts system-wide.
3277 noioapicquirk [APIC] Disable all boot interrupt quirks.
3278 Safety option to keep boot IRQs enabled. This
3279 should never be necessary.
3280 ioapicreroute [APIC] Enable rerouting of boot IRQs to the
3281 primary IO-APIC for bridges that cannot disable
3282 boot IRQs. This fixes a source of spurious IRQs
3283 when the system masks IRQs.
3284 noioapicreroute [APIC] Disable workaround that uses the
3285 boot IRQ equivalent of an IRQ that connects to
3286 a chipset where boot IRQs cannot be disabled.
3287 The opposite of ioapicreroute.
3288 biosirq [X86-32] Use PCI BIOS calls to get the interrupt
3289 routing table. These calls are known to be buggy
3290 on several machines and they hang the machine
3291 when used, but on other computers it's the only
3292 way to get the interrupt routing table. Try
3293 this option if the kernel is unable to allocate
3294 IRQs or discover secondary PCI buses on your
3295 motherboard.
3296 rom [X86] Assign address space to expansion ROMs.
3297 Use with caution as certain devices share
3298 address decoders between ROMs and other
3299 resources.
3300 norom [X86] Do not assign address space to
3301 expansion ROMs that do not already have
3302 BIOS assigned address ranges.
3303 nobar [X86] Do not assign address space to the
3304 BARs that weren't assigned by the BIOS.
3305 irqmask=0xMMMM [X86] Set a bit mask of IRQs allowed to be
3306 assigned automatically to PCI devices. You can
3307 make the kernel exclude IRQs of your ISA cards
3308 this way.
3309 pirqaddr=0xAAAAA [X86] Specify the physical address
3310 of the PIRQ table (normally generated
3311 by the BIOS) if it is outside the
3312 F0000h-100000h range.
3313 lastbus=N [X86] Scan all buses thru bus #N. Can be
3314 useful if the kernel is unable to find your
3315 secondary buses and you want to tell it
3316 explicitly which ones they are.
3317 assign-busses [X86] Always assign all PCI bus
3318 numbers ourselves, overriding
3319 whatever the firmware may have done.
3320 usepirqmask [X86] Honor the possible IRQ mask stored
3321 in the BIOS $PIR table. This is needed on
3322 some systems with broken BIOSes, notably
3323 some HP Pavilion N5400 and Omnibook XE3
3324 notebooks. This will have no effect if ACPI
3325 IRQ routing is enabled.
3326 noacpi [X86] Do not use ACPI for IRQ routing
3327 or for PCI scanning.
3328 use_crs [X86] Use PCI host bridge window information
3329 from ACPI. On BIOSes from 2008 or later, this
3330 is enabled by default. If you need to use this,
3331 please report a bug.
3332 nocrs [X86] Ignore PCI host bridge windows from ACPI.
3333 If you need to use this, please report a bug.
3334 routeirq Do IRQ routing for all PCI devices.
3335 This is normally done in pci_enable_device(),
3336 so this option is a temporary workaround
3337 for broken drivers that don't call it.
3338 skip_isa_align [X86] do not align io start addr, so can
3339 handle more pci cards
3340 noearly [X86] Don't do any early type 1 scanning.
3341 This might help on some broken boards which
3342 machine check when some devices' config space
3343 is read. But various workarounds are disabled
3344 and some IOMMU drivers will not work.
3345 bfsort Sort PCI devices into breadth-first order.
3346 This sorting is done to get a device
3347 order compatible with older (<= 2.4) kernels.
3348 nobfsort Don't sort PCI devices into breadth-first order.
3349 pcie_bus_tune_off Disable PCIe MPS (Max Payload Size)
3350 tuning and use the BIOS-configured MPS defaults.
3351 pcie_bus_safe Set every device's MPS to the largest value
3352 supported by all devices below the root complex.
3353 pcie_bus_perf Set device MPS to the largest allowable MPS
3354 based on its parent bus. Also set MRRS (Max
3355 Read Request Size) to the largest supported
3356 value (no larger than the MPS that the device
3357 or bus can support) for best performance.
3358 pcie_bus_peer2peer Set every device's MPS to 128B, which
3359 every device is guaranteed to support. This
3360 configuration allows peer-to-peer DMA between
3361 any pair of devices, possibly at the cost of
3362 reduced performance. This also guarantees
3363 that hot-added devices will work.
3364 cbiosize=nn[KMG] The fixed amount of bus space which is
3365 reserved for the CardBus bridge's IO window.
3366 The default value is 256 bytes.
3367 cbmemsize=nn[KMG] The fixed amount of bus space which is
3368 reserved for the CardBus bridge's memory
3369 window. The default value is 64 megabytes.
3370 resource_alignment=
3371 Format:
3372 [<order of align>@]<pci_dev>[; ...]
3373 Specifies alignment and device to reassign
3374 aligned memory resources. How to
3375 specify the device is described above.
3376 If <order of align> is not specified,
3377 PAGE_SIZE is used as alignment.
3378 PCI-PCI bridge can be specified, if resource
3379 windows need to be expanded.
3380 To specify the alignment for several
3381 instances of a device, the PCI vendor,
3382 device, subvendor, and subdevice may be
3383 specified, e.g., 4096@pci:8086:9c22:103c:198f
3384 ecrc= Enable/disable PCIe ECRC (transaction layer
3385 end-to-end CRC checking).
3386 bios: Use BIOS/firmware settings. This is the
3387 the default.
3388 off: Turn ECRC off
3389 on: Turn ECRC on.
3390 hpiosize=nn[KMG] The fixed amount of bus space which is
3391 reserved for hotplug bridge's IO window.
3392 Default size is 256 bytes.
3393 hpmemsize=nn[KMG] The fixed amount of bus space which is
3394 reserved for hotplug bridge's memory window.
3395 Default size is 2 megabytes.
3396 hpbussize=nn The minimum amount of additional bus numbers
3397 reserved for buses below a hotplug bridge.
3398 Default is 1.
3399 realloc= Enable/disable reallocating PCI bridge resources
3400 if allocations done by BIOS are too small to
3401 accommodate resources required by all child
3402 devices.
3403 off: Turn realloc off
3404 on: Turn realloc on
3405 realloc same as realloc=on
3406 noari do not use PCIe ARI.
3407 noats [PCIE, Intel-IOMMU, AMD-IOMMU]
3408 do not use PCIe ATS (and IOMMU device IOTLB).
3409 pcie_scan_all Scan all possible PCIe devices. Otherwise we
3410 only look for one device below a PCIe downstream
3411 port.
3412 big_root_window Try to add a big 64bit memory window to the PCIe
3413 root complex on AMD CPUs. Some GFX hardware
3414 can resize a BAR to allow access to all VRAM.
3415 Adding the window is slightly risky (it may
3416 conflict with unreported devices), so this
3417 taints the kernel.
3418 disable_acs_redir=<pci_dev>[; ...]
3419 Specify one or more PCI devices (in the format
3420 specified above) separated by semicolons.
3421 Each device specified will have the PCI ACS
3422 redirect capabilities forced off which will
3423 allow P2P traffic between devices through
3424 bridges without forcing it upstream. Note:
3425 this removes isolation between devices and
3426 may put more devices in an IOMMU group.
3428 pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
3429 Management.
3430 off Disable ASPM.
3431 force Enable ASPM even on devices that claim not to support it.
3432 WARNING: Forcing ASPM on may cause system lockups.
3434 pcie_ports= [PCIE] PCIe port services handling:
3435 native Use native PCIe services (PME, AER, DPC, PCIe hotplug)
3436 even if the platform doesn't give the OS permission to
3437 use them. This may cause conflicts if the platform
3438 also tries to use these services.
3439 compat Disable native PCIe services (PME, AER, DPC, PCIe
3440 hotplug).
3442 pcie_port_pm= [PCIE] PCIe port power management handling:
3443 off Disable power management of all PCIe ports
3444 force Forcibly enable power management of all PCIe ports
3446 pcie_pme= [PCIE,PM] Native PCIe PME signaling options:
3447 nomsi Do not use MSI for native PCIe PME signaling (this makes
3448 all PCIe root ports use INTx for all services).
3450 pcmv= [HW,PCMCIA] BadgePAD 4
3452 pd_ignore_unused
3453 [PM]
3454 Keep all power-domains already enabled by bootloader on,
3455 even if no driver has claimed them. This is useful
3456 for debug and development, but should not be
3457 needed on a platform with proper driver support.
3459 pd. [PARIDE]
3460 See Documentation/blockdev/paride.txt.
3462 pdcchassis= [PARISC,HW] Disable/Enable PDC Chassis Status codes at
3463 boot time.
3464 Format: { 0 | 1 }
3465 See arch/parisc/kernel/pdc_chassis.c
3467 percpu_alloc= Select which percpu first chunk allocator to use.
3468 Currently supported values are "embed" and "page".
3469 Archs may support subset or none of the selections.
3470 See comments in mm/percpu.c for details on each
3471 allocator. This parameter is primarily for debugging
3472 and performance comparison.
3474 pf. [PARIDE]
3475 See Documentation/blockdev/paride.txt.
3477 pg. [PARIDE]
3478 See Documentation/blockdev/paride.txt.
3480 pirq= [SMP,APIC] Manual mp-table setup
3481 See Documentation/x86/i386/IO-APIC.txt.
3483 plip= [PPT,NET] Parallel port network link
3484 Format: { parport<nr> | timid | 0 }
3485 See also Documentation/admin-guide/parport.rst.
3487 pmtmr= [X86] Manual setup of pmtmr I/O Port.
3488 Override pmtimer IOPort with a hex value.
3489 e.g. pmtmr=0x508
3491 pnp.debug=1 [PNP]
3492 Enable PNP debug messages (depends on the
3493 CONFIG_PNP_DEBUG_MESSAGES option). Change at run-time
3494 via /sys/module/pnp/parameters/debug. We always show
3495 current resource usage; turning this on also shows
3496 possible settings and some assignment information.
3498 pnpacpi= [ACPI]
3499 { off }
3501 pnpbios= [ISAPNP]
3502 { on | off | curr | res | no-curr | no-res }
3504 pnp_reserve_irq=
3505 [ISAPNP] Exclude IRQs for the autoconfiguration
3507 pnp_reserve_dma=
3508 [ISAPNP] Exclude DMAs for the autoconfiguration
3510 pnp_reserve_io= [ISAPNP] Exclude I/O ports for the autoconfiguration
3511 Ranges are in pairs (I/O port base and size).
3513 pnp_reserve_mem=
3514 [ISAPNP] Exclude memory regions for the
3515 autoconfiguration.
3516 Ranges are in pairs (memory base and size).
3518 ports= [IP_VS_FTP] IPVS ftp helper module
3519 Default is 21.
3520 Up to 8 (IP_VS_APP_MAX_PORTS) ports
3521 may be specified.
3522 Format: <port>,<port>....
3524 powersave=off [PPC] This option disables power saving features.
3525 It specifically disables cpuidle and sets the
3526 platform machine description specific power_save
3527 function to NULL. On Idle the CPU just reduces
3528 execution priority.
3530 ppc_strict_facility_enable
3531 [PPC] This option catches any kernel floating point,
3532 Altivec, VSX and SPE outside of regions specifically
3533 allowed (eg kernel_enable_fpu()/kernel_disable_fpu()).
3534 There is some performance impact when enabling this.
3536 ppc_tm= [PPC]
3537 Format: {"off"}
3538 Disable Hardware Transactional Memory
3540 print-fatal-signals=
3541 [KNL] debug: print fatal signals
3543 If enabled, warn about various signal handling
3544 related application anomalies: too many signals,
3545 too many POSIX.1 timers, fatal signals causing a
3546 coredump - etc.
3548 If you hit the warning due to signal overflow,
3549 you might want to try "ulimit -i unlimited".
3551 default: off.
3553 printk.always_kmsg_dump=
3554 Trigger kmsg_dump for cases other than kernel oops or
3555 panics
3556 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3557 default: disabled
3559 printk.devkmsg={on,off,ratelimit}
3560 Control writing to /dev/kmsg.
3561 on - unlimited logging to /dev/kmsg from userspace
3562 off - logging to /dev/kmsg disabled
3563 ratelimit - ratelimit the logging
3564 Default: ratelimit
3566 printk.time= Show timing data prefixed to each printk message line
3567 Format: <bool> (1/Y/y=enable, 0/N/n=disable)
3569 processor.max_cstate= [HW,ACPI]
3570 Limit processor to maximum C-state
3571 max_cstate=9 overrides any DMI blacklist limit.
3573 processor.nocst [HW,ACPI]
3574 Ignore the _CST method to determine C-states,
3575 instead using the legacy FADT method
3577 profile= [KNL] Enable kernel profiling via /proc/profile
3578 Format: [<profiletype>,]<number>
3579 Param: <profiletype>: "schedule", "sleep", or "kvm"
3580 [defaults to kernel profiling]
3581 Param: "schedule" - profile schedule points.
3582 Param: "sleep" - profile D-state sleeping (millisecs).
3583 Requires CONFIG_SCHEDSTATS
3584 Param: "kvm" - profile VM exits.
3585 Param: <number> - step/bucket size as a power of 2 for
3586 statistical time based profiling.
3588 prompt_ramdisk= [RAM] List of RAM disks to prompt for floppy disk
3589 before loading.
3590 See Documentation/blockdev/ramdisk.txt.
3592 psmouse.proto= [HW,MOUSE] Highest PS2 mouse protocol extension to
3593 probe for; one of (bare|imps|exps|lifebook|any).
3594 psmouse.rate= [HW,MOUSE] Set desired mouse report rate, in reports
3595 per second.
3596 psmouse.resetafter= [HW,MOUSE]
3597 Try to reset the device after so many bad packets
3598 (0 = never).
3599 psmouse.resolution=
3600 [HW,MOUSE] Set desired mouse resolution, in dpi.
3601 psmouse.smartscroll=
3602 [HW,MOUSE] Controls Logitech smartscroll autorepeat.
3603 0 = disabled, 1 = enabled (default).
3605 pstore.backend= Specify the name of the pstore backend to use
3607 pt. [PARIDE]
3608 See Documentation/blockdev/paride.txt.
3610 pti= [X86_64] Control Page Table Isolation of user and
3611 kernel address spaces. Disabling this feature
3612 removes hardening, but improves performance of
3613 system calls and interrupts.
3615 on - unconditionally enable
3616 off - unconditionally disable
3617 auto - kernel detects whether your CPU model is
3618 vulnerable to issues that PTI mitigates
3620 Not specifying this option is equivalent to pti=auto.
3622 nopti [X86_64]
3623 Equivalent to pti=off
3625 pty.legacy_count=
3626 [KNL] Number of legacy pty's. Overwrites compiled-in
3627 default number.
3629 quiet [KNL] Disable most log messages
3631 r128= [HW,DRM]
3633 raid= [HW,RAID]
3634 See Documentation/admin-guide/md.rst.
3636 ramdisk_size= [RAM] Sizes of RAM disks in kilobytes
3637 See Documentation/blockdev/ramdisk.txt.
3639 random.trust_cpu={on,off}
3640 [KNL] Enable or disable trusting the use of the
3641 CPU's random number generator (if available) to
3642 fully seed the kernel's CRNG. Default is controlled
3643 by CONFIG_RANDOM_TRUST_CPU.
3645 ras=option[,option,...] [KNL] RAS-specific options
3647 cec_disable [X86]
3648 Disable the Correctable Errors Collector,
3649 see CONFIG_RAS_CEC help text.
3651 rcu_nocbs= [KNL]
3652 The argument is a cpu list, as described above.
3654 In kernels built with CONFIG_RCU_NOCB_CPU=y, set
3655 the specified list of CPUs to be no-callback CPUs.
3656 Invocation of these CPUs' RCU callbacks will
3657 be offloaded to "rcuox/N" kthreads created for
3658 that purpose, where "x" is "b" for RCU-bh, "p"
3659 for RCU-preempt, and "s" for RCU-sched, and "N"
3660 is the CPU number. This reduces OS jitter on the
3661 offloaded CPUs, which can be useful for HPC and
3662 real-time workloads. It can also improve energy
3663 efficiency for asymmetric multiprocessors.
3665 rcu_nocb_poll [KNL]
3666 Rather than requiring that offloaded CPUs
3667 (specified by rcu_nocbs= above) explicitly
3668 awaken the corresponding "rcuoN" kthreads,
3669 make these kthreads poll for callbacks.
3670 This improves the real-time response for the
3671 offloaded CPUs by relieving them of the need to
3672 wake up the corresponding kthread, but degrades
3673 energy efficiency by requiring that the kthreads
3674 periodically wake up to do the polling.
3676 rcutree.blimit= [KNL]
3677 Set maximum number of finished RCU callbacks to
3678 process in one batch.
3680 rcutree.dump_tree= [KNL]
3681 Dump the structure of the rcu_node combining tree
3682 out at early boot. This is used for diagnostic
3683 purposes, to verify correct tree setup.
3685 rcutree.gp_cleanup_delay= [KNL]
3686 Set the number of jiffies to delay each step of
3687 RCU grace-period cleanup.
3689 rcutree.gp_init_delay= [KNL]
3690 Set the number of jiffies to delay each step of
3691 RCU grace-period initialization.
3693 rcutree.gp_preinit_delay= [KNL]
3694 Set the number of jiffies to delay each step of
3695 RCU grace-period pre-initialization, that is,
3696 the propagation of recent CPU-hotplug changes up
3697 the rcu_node combining tree.
3699 rcutree.rcu_fanout_exact= [KNL]
3700 Disable autobalancing of the rcu_node combining
3701 tree. This is used by rcutorture, and might
3702 possibly be useful for architectures having high
3703 cache-to-cache transfer latencies.
3705 rcutree.rcu_fanout_leaf= [KNL]
3706 Change the number of CPUs assigned to each
3707 leaf rcu_node structure. Useful for very
3708 large systems, which will choose the value 64,
3709 and for NUMA systems with large remote-access
3710 latencies, which will choose a value aligned
3711 with the appropriate hardware boundaries.
3713 rcutree.jiffies_till_sched_qs= [KNL]
3714 Set required age in jiffies for a
3715 given grace period before RCU starts
3716 soliciting quiescent-state help from
3717 rcu_note_context_switch().
3719 rcutree.jiffies_till_first_fqs= [KNL]
3720 Set delay from grace-period initialization to
3721 first attempt to force quiescent states.
3722 Units are jiffies, minimum value is zero,
3723 and maximum value is HZ.
3725 rcutree.jiffies_till_next_fqs= [KNL]
3726 Set delay between subsequent attempts to force
3727 quiescent states. Units are jiffies, minimum
3728 value is one, and maximum value is HZ.
3730 rcutree.kthread_prio= [KNL,BOOT]
3731 Set the SCHED_FIFO priority of the RCU per-CPU
3732 kthreads (rcuc/N). This value is also used for
3733 the priority of the RCU boost threads (rcub/N)
3734 and for the RCU grace-period kthreads (rcu_bh,
3735 rcu_preempt, and rcu_sched). If RCU_BOOST is
3736 set, valid values are 1-99 and the default is 1
3737 (the least-favored priority). Otherwise, when
3738 RCU_BOOST is not set, valid values are 0-99 and
3739 the default is zero (non-realtime operation).
3741 rcutree.rcu_nocb_leader_stride= [KNL]
3742 Set the number of NOCB kthread groups, which
3743 defaults to the square root of the number of
3744 CPUs. Larger numbers reduces the wakeup overhead
3745 on the per-CPU grace-period kthreads, but increases
3746 that same overhead on each group's leader.
3748 rcutree.qhimark= [KNL]
3749 Set threshold of queued RCU callbacks beyond which
3750 batch limiting is disabled.
3752 rcutree.qlowmark= [KNL]
3753 Set threshold of queued RCU callbacks below which
3754 batch limiting is re-enabled.
3756 rcutree.rcu_idle_gp_delay= [KNL]
3757 Set wakeup interval for idle CPUs that have
3758 RCU callbacks (RCU_FAST_NO_HZ=y).
3760 rcutree.rcu_idle_lazy_gp_delay= [KNL]
3761 Set wakeup interval for idle CPUs that have
3762 only "lazy" RCU callbacks (RCU_FAST_NO_HZ=y).
3763 Lazy RCU callbacks are those which RCU can
3764 prove do nothing more than free memory.
3766 rcutree.rcu_kick_kthreads= [KNL]
3767 Cause the grace-period kthread to get an extra
3768 wake_up() if it sleeps three times longer than
3769 it should at force-quiescent-state time.
3770 This wake_up() will be accompanied by a
3771 WARN_ONCE() splat and an ftrace_dump().
3773 rcuperf.gp_async= [KNL]
3774 Measure performance of asynchronous
3775 grace-period primitives such as call_rcu().
3777 rcuperf.gp_async_max= [KNL]
3778 Specify the maximum number of outstanding
3779 callbacks per writer thread. When a writer
3780 thread exceeds this limit, it invokes the
3781 corresponding flavor of rcu_barrier() to allow
3782 previously posted callbacks to drain.
3784 rcuperf.gp_exp= [KNL]
3785 Measure performance of expedited synchronous
3786 grace-period primitives.
3788 rcuperf.holdoff= [KNL]
3789 Set test-start holdoff period. The purpose of
3790 this parameter is to delay the start of the
3791 test until boot completes in order to avoid
3792 interference.
3794 rcuperf.nreaders= [KNL]
3795 Set number of RCU readers. The value -1 selects
3796 N, where N is the number of CPUs. A value
3797 "n" less than -1 selects N-n+1, where N is again
3798 the number of CPUs. For example, -2 selects N
3799 (the number of CPUs), -3 selects N+1, and so on.
3800 A value of "n" less than or equal to -N selects
3801 a single reader.
3803 rcuperf.nwriters= [KNL]
3804 Set number of RCU writers. The values operate
3805 the same as for rcuperf.nreaders.
3806 N, where N is the number of CPUs
3808 rcuperf.perf_type= [KNL]
3809 Specify the RCU implementation to test.
3811 rcuperf.shutdown= [KNL]
3812 Shut the system down after performance tests
3813 complete. This is useful for hands-off automated
3814 testing.
3816 rcuperf.verbose= [KNL]
3817 Enable additional printk() statements.
3819 rcuperf.writer_holdoff= [KNL]
3820 Write-side holdoff between grace periods,
3821 in microseconds. The default of zero says
3822 no holdoff.
3824 rcutorture.cbflood_inter_holdoff= [KNL]
3825 Set holdoff time (jiffies) between successive
3826 callback-flood tests.
3828 rcutorture.cbflood_intra_holdoff= [KNL]
3829 Set holdoff time (jiffies) between successive
3830 bursts of callbacks within a given callback-flood
3831 test.
3833 rcutorture.cbflood_n_burst= [KNL]
3834 Set the number of bursts making up a given
3835 callback-flood test. Set this to zero to
3836 disable callback-flood testing.
3838 rcutorture.cbflood_n_per_burst= [KNL]
3839 Set the number of callbacks to be registered
3840 in a given burst of a callback-flood test.
3842 rcutorture.fqs_duration= [KNL]
3843 Set duration of force_quiescent_state bursts
3844 in microseconds.
3846 rcutorture.fqs_holdoff= [KNL]
3847 Set holdoff time within force_quiescent_state bursts
3848 in microseconds.
3850 rcutorture.fqs_stutter= [KNL]
3851 Set wait time between force_quiescent_state bursts
3852 in seconds.
3854 rcutorture.gp_cond= [KNL]
3855 Use conditional/asynchronous update-side
3856 primitives, if available.
3858 rcutorture.gp_exp= [KNL]
3859 Use expedited update-side primitives, if available.
3861 rcutorture.gp_normal= [KNL]
3862 Use normal (non-expedited) asynchronous
3863 update-side primitives, if available.
3865 rcutorture.gp_sync= [KNL]
3866 Use normal (non-expedited) synchronous
3867 update-side primitives, if available. If all
3868 of rcutorture.gp_cond=, rcutorture.gp_exp=,
3869 rcutorture.gp_normal=, and rcutorture.gp_sync=
3870 are zero, rcutorture acts as if is interpreted
3871 they are all non-zero.
3873 rcutorture.n_barrier_cbs= [KNL]
3874 Set callbacks/threads for rcu_barrier() testing.
3876 rcutorture.nfakewriters= [KNL]
3877 Set number of concurrent RCU writers. These just
3878 stress RCU, they don't participate in the actual
3879 test, hence the "fake".
3881 rcutorture.nreaders= [KNL]
3882 Set number of RCU readers. The value -1 selects
3883 N-1, where N is the number of CPUs. A value
3884 "n" less than -1 selects N-n-2, where N is again
3885 the number of CPUs. For example, -2 selects N
3886 (the number of CPUs), -3 selects N+1, and so on.
3888 rcutorture.object_debug= [KNL]
3889 Enable debug-object double-call_rcu() testing.
3891 rcutorture.onoff_holdoff= [KNL]
3892 Set time (s) after boot for CPU-hotplug testing.
3894 rcutorture.onoff_interval= [KNL]
3895 Set time (jiffies) between CPU-hotplug operations,
3896 or zero to disable CPU-hotplug testing.
3898 rcutorture.shuffle_interval= [KNL]
3899 Set task-shuffle interval (s). Shuffling tasks
3900 allows some CPUs to go into dyntick-idle mode
3901 during the rcutorture test.
3903 rcutorture.shutdown_secs= [KNL]
3904 Set time (s) after boot system shutdown. This
3905 is useful for hands-off automated testing.
3907 rcutorture.stall_cpu= [KNL]
3908 Duration of CPU stall (s) to test RCU CPU stall
3909 warnings, zero to disable.
3911 rcutorture.stall_cpu_holdoff= [KNL]
3912 Time to wait (s) after boot before inducing stall.
3914 rcutorture.stall_cpu_irqsoff= [KNL]
3915 Disable interrupts while stalling if set.
3917 rcutorture.stat_interval= [KNL]
3918 Time (s) between statistics printk()s.
3920 rcutorture.stutter= [KNL]
3921 Time (s) to stutter testing, for example, specifying
3922 five seconds causes the test to run for five seconds,
3923 wait for five seconds, and so on. This tests RCU's
3924 ability to transition abruptly to and from idle.
3926 rcutorture.test_boost= [KNL]
3927 Test RCU priority boosting? 0=no, 1=maybe, 2=yes.
3928 "Maybe" means test if the RCU implementation
3929 under test support RCU priority boosting.
3931 rcutorture.test_boost_duration= [KNL]
3932 Duration (s) of each individual boost test.
3934 rcutorture.test_boost_interval= [KNL]
3935 Interval (s) between each boost test.
3937 rcutorture.test_no_idle_hz= [KNL]
3938 Test RCU's dyntick-idle handling. See also the
3939 rcutorture.shuffle_interval parameter.
3941 rcutorture.torture_type= [KNL]
3942 Specify the RCU implementation to test.
3944 rcutorture.verbose= [KNL]
3945 Enable additional printk() statements.
3947 rcupdate.rcu_cpu_stall_suppress= [KNL]
3948 Suppress RCU CPU stall warning messages.
3950 rcupdate.rcu_cpu_stall_timeout= [KNL]
3951 Set timeout for RCU CPU stall warning messages.
3953 rcupdate.rcu_expedited= [KNL]
3954 Use expedited grace-period primitives, for
3955 example, synchronize_rcu_expedited() instead
3956 of synchronize_rcu(). This reduces latency,
3957 but can increase CPU utilization, degrade
3958 real-time latency, and degrade energy efficiency.
3959 No effect on CONFIG_TINY_RCU kernels.
3961 rcupdate.rcu_normal= [KNL]
3962 Use only normal grace-period primitives,
3963 for example, synchronize_rcu() instead of
3964 synchronize_rcu_expedited(). This improves
3965 real-time latency, CPU utilization, and
3966 energy efficiency, but can expose users to
3967 increased grace-period latency. This parameter
3968 overrides rcupdate.rcu_expedited. No effect on
3969 CONFIG_TINY_RCU kernels.
3971 rcupdate.rcu_normal_after_boot= [KNL]
3972 Once boot has completed (that is, after
3973 rcu_end_inkernel_boot() has been invoked), use
3974 only normal grace-period primitives. No effect
3975 on CONFIG_TINY_RCU kernels.
3977 rcupdate.rcu_task_stall_timeout= [KNL]
3978 Set timeout in jiffies for RCU task stall warning
3979 messages. Disable with a value less than or equal
3980 to zero.
3982 rcupdate.rcu_self_test= [KNL]
3983 Run the RCU early boot self tests
3985 rcupdate.rcu_self_test_bh= [KNL]
3986 Run the RCU bh early boot self tests
3988 rcupdate.rcu_self_test_sched= [KNL]
3989 Run the RCU sched early boot self tests
3991 rdinit= [KNL]
3992 Format: <full_path>
3993 Run specified binary instead of /init from the ramdisk,
3994 used for early userspace startup. See initrd.
3996 rdrand= [X86]
3997 force - Override the decision by the kernel to hide the
3998 advertisement of RDRAND support (this affects
3999 certain AMD processors because of buggy BIOS
4000 support, specifically around the suspend/resume
4001 path).
4003 rdt= [HW,X86,RDT]
4004 Turn on/off individual RDT features. List is:
4005 cmt, mbmtotal, mbmlocal, l3cat, l3cdp, l2cat, l2cdp,
4006 mba.
4007 E.g. to turn on cmt and turn off mba use:
4008 rdt=cmt,!mba
4010 reboot= [KNL]
4011 Format (x86 or x86_64):
4012 [w[arm] | c[old] | h[ard] | s[oft] | g[pio]] \
4013 [[,]s[mp]#### \
4014 [[,]b[ios] | a[cpi] | k[bd] | t[riple] | e[fi] | p[ci]] \
4015 [[,]f[orce]
4016 Where reboot_mode is one of warm (soft) or cold (hard) or gpio,
4017 reboot_type is one of bios, acpi, kbd, triple, efi, or pci,
4018 reboot_force is either force or not specified,
4019 reboot_cpu is s[mp]#### with #### being the processor
4020 to be used for rebooting.
4022 relax_domain_level=
4023 [KNL, SMP] Set scheduler's default relax_domain_level.
4024 See Documentation/cgroup-v1/cpusets.txt.
4026 reserve= [KNL,BUGS] Force kernel to ignore I/O ports or memory
4027 Format: <base1>,<size1>[,<base2>,<size2>,...]
4028 Reserve I/O ports or memory so the kernel won't use
4029 them. If <base> is less than 0x10000, the region
4030 is assumed to be I/O ports; otherwise it is memory.
4032 reservetop= [X86-32]
4033 Format: nn[KMG]
4034 Reserves a hole at the top of the kernel virtual
4035 address space.
4037 reservelow= [X86]
4038 Format: nn[K]
4039 Set the amount of memory to reserve for BIOS at
4040 the bottom of the address space.
4042 reset_devices [KNL] Force drivers to reset the underlying device
4043 during initialization.
4045 resume= [SWSUSP]
4046 Specify the partition device for software suspend
4047 Format:
4048 {/dev/<dev> | PARTUUID=<uuid> | <int>:<int> | <hex>}
4050 resume_offset= [SWSUSP]
4051 Specify the offset from the beginning of the partition
4052 given by "resume=" at which the swap header is located,
4053 in <PAGE_SIZE> units (needed only for swap files).
4054 See Documentation/power/swsusp-and-swap-files.txt
4056 resumedelay= [HIBERNATION] Delay (in seconds) to pause before attempting to
4057 read the resume files
4059 resumewait [HIBERNATION] Wait (indefinitely) for resume device to show up.
4060 Useful for devices that are detected asynchronously
4061 (e.g. USB and MMC devices).
4063 hibernate= [HIBERNATION]
4064 noresume Don't check if there's a hibernation image
4065 present during boot.
4066 nocompress Don't compress/decompress hibernation images.
4067 no Disable hibernation and resume.
4068 protect_image Turn on image protection during restoration
4069 (that will set all pages holding image data
4070 during restoration read-only).
4072 retain_initrd [RAM] Keep initrd memory after extraction
4074 rfkill.default_state=
4075 0 "airplane mode". All wifi, bluetooth, wimax, gps, fm,
4076 etc. communication is blocked by default.
4077 1 Unblocked.
4079 rfkill.master_switch_mode=
4080 0 The "airplane mode" button does nothing.
4081 1 The "airplane mode" button toggles between everything
4082 blocked and the previous configuration.
4083 2 The "airplane mode" button toggles between everything
4084 blocked and everything unblocked.
4086 rhash_entries= [KNL,NET]
4087 Set number of hash buckets for route cache
4089 ring3mwait=disable
4090 [KNL] Disable ring 3 MONITOR/MWAIT feature on supported
4091 CPUs.
4093 ro [KNL] Mount root device read-only on boot
4095 rodata= [KNL]
4096 on Mark read-only kernel memory as read-only (default).
4097 off Leave read-only kernel memory writable for debugging.
4099 rockchip.usb_uart
4100 Enable the uart passthrough on the designated usb port
4101 on Rockchip SoCs. When active, the signals of the
4102 debug-uart get routed to the D+ and D- pins of the usb
4103 port and the regular usb controller gets disabled.
4105 root= [KNL] Root filesystem
4106 See name_to_dev_t comment in init/do_mounts.c.
4108 rootdelay= [KNL] Delay (in seconds) to pause before attempting to
4109 mount the root filesystem
4111 rootflags= [KNL] Set root filesystem mount option string
4113 rootfstype= [KNL] Set root filesystem type
4115 rootwait [KNL] Wait (indefinitely) for root device to show up.
4116 Useful for devices that are detected asynchronously
4117 (e.g. USB and MMC devices).
4119 rproc_mem=nn[KMG][@address]
4120 [KNL,ARM,CMA] Remoteproc physical memory block.
4121 Memory area to be used by remote processor image,
4122 managed by CMA.
4124 rw [KNL] Mount root device read-write on boot
4126 S [KNL] Run init in single mode
4128 s390_iommu= [HW,S390]
4129 Set s390 IOTLB flushing mode
4130 strict
4131 With strict flushing every unmap operation will result in
4132 an IOTLB flush. Default is lazy flushing before reuse,
4133 which is faster.
4135 sa1100ir [NET]
4136 See drivers/net/irda/sa1100_ir.c.
4138 sbni= [NET] Granch SBNI12 leased line adapter
4140 sched_debug [KNL] Enables verbose scheduler debug messages.
4142 schedstats= [KNL,X86] Enable or disable scheduled statistics.
4143 Allowed values are enable and disable. This feature
4144 incurs a small amount of overhead in the scheduler
4145 but is useful for debugging and performance tuning.
4147 skew_tick= [KNL] Offset the periodic timer tick per cpu to mitigate
4148 xtime_lock contention on larger systems, and/or RCU lock
4149 contention on all systems with CONFIG_MAXSMP set.
4150 Format: { "0" | "1" }
4151 0 -- disable. (may be 1 via CONFIG_CMDLINE="skew_tick=1"
4152 1 -- enable.
4153 Note: increases power consumption, thus should only be
4154 enabled if running jitter sensitive (HPC/RT) workloads.
4156 security= [SECURITY] Choose a security module to enable at boot.
4157 If this boot parameter is not specified, only the first
4158 security module asking for security registration will be
4159 loaded. An invalid security module name will be treated
4160 as if no module has been chosen.
4162 selinux= [SELINUX] Disable or enable SELinux at boot time.
4163 Format: { "0" | "1" }
4164 See security/selinux/Kconfig help text.
4165 0 -- disable.
4166 1 -- enable.
4167 Default value is set via kernel config option.
4168 If enabled at boot time, /selinux/disable can be used
4169 later to disable prior to initial policy load.
4171 apparmor= [APPARMOR] Disable or enable AppArmor at boot time
4172 Format: { "0" | "1" }
4173 See security/apparmor/Kconfig help text
4174 0 -- disable.
4175 1 -- enable.
4176 Default value is set via kernel config option.
4178 serialnumber [BUGS=X86-32]
4180 shapers= [NET]
4181 Maximal number of shapers.
4183 simeth= [IA-64]
4184 simscsi=
4186 slram= [HW,MTD]
4188 slab_nomerge [MM]
4189 Disable merging of slabs with similar size. May be
4190 necessary if there is some reason to distinguish
4191 allocs to different slabs, especially in hardened
4192 environments where the risk of heap overflows and
4193 layout control by attackers can usually be
4194 frustrated by disabling merging. This will reduce
4195 most of the exposure of a heap attack to a single
4196 cache (risks via metadata attacks are mostly
4197 unchanged). Debug options disable merging on their
4198 own.
4199 For more information see Documentation/vm/slub.rst.
4201 slab_max_order= [MM, SLAB]
4202 Determines the maximum allowed order for slabs.
4203 A high setting may cause OOMs due to memory
4204 fragmentation. Defaults to 1 for systems with
4205 more than 32MB of RAM, 0 otherwise.
4207 slub_debug[=options[,slabs]] [MM, SLUB]
4208 Enabling slub_debug allows one to determine the
4209 culprit if slab objects become corrupted. Enabling
4210 slub_debug can create guard zones around objects and
4211 may poison objects when not in use. Also tracks the
4212 last alloc / free. For more information see
4213 Documentation/vm/slub.rst.
4215 slub_memcg_sysfs= [MM, SLUB]
4216 Determines whether to enable sysfs directories for
4217 memory cgroup sub-caches. 1 to enable, 0 to disable.
4218 The default is determined by CONFIG_SLUB_MEMCG_SYSFS_ON.
4219 Enabling this can lead to a very high number of debug
4220 directories and files being created under
4221 /sys/kernel/slub.
4223 slub_max_order= [MM, SLUB]
4224 Determines the maximum allowed order for slabs.
4225 A high setting may cause OOMs due to memory
4226 fragmentation. For more information see
4227 Documentation/vm/slub.rst.
4229 slub_min_objects= [MM, SLUB]
4230 The minimum number of objects per slab. SLUB will
4231 increase the slab order up to slub_max_order to
4232 generate a sufficiently large slab able to contain
4233 the number of objects indicated. The higher the number
4234 of objects the smaller the overhead of tracking slabs
4235 and the less frequently locks need to be acquired.
4236 For more information see Documentation/vm/slub.rst.
4238 slub_min_order= [MM, SLUB]
4239 Determines the minimum page order for slabs. Must be
4240 lower than slub_max_order.
4241 For more information see Documentation/vm/slub.rst.
4243 slub_nomerge [MM, SLUB]
4244 Same with slab_nomerge. This is supported for legacy.
4245 See slab_nomerge for more information.
4247 smart2= [HW]
4248 Format: <io1>[,<io2>[,...,<io8>]]
4250 smsc-ircc2.nopnp [HW] Don't use PNP to discover SMC devices
4251 smsc-ircc2.ircc_cfg= [HW] Device configuration I/O port
4252 smsc-ircc2.ircc_sir= [HW] SIR base I/O port
4253 smsc-ircc2.ircc_fir= [HW] FIR base I/O port
4254 smsc-ircc2.ircc_irq= [HW] IRQ line
4255 smsc-ircc2.ircc_dma= [HW] DMA channel
4256 smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
4257 0: Toshiba Satellite 1800 (GP data pin select)
4258 1: Fast pin select (default)
4259 2: ATC IRMode
4261 smt [KNL,S390] Set the maximum number of threads (logical
4262 CPUs) to use per physical CPU on systems capable of
4263 symmetric multithreading (SMT). Will be capped to the
4264 actual hardware limit.
4265 Format: <integer>
4266 Default: -1 (no limit)
4268 softlockup_panic=
4269 [KNL] Should the soft-lockup detector generate panics.
4270 Format: <integer>
4272 A nonzero value instructs the soft-lockup detector
4273 to panic the machine when a soft-lockup occurs. This
4274 is also controlled by CONFIG_BOOTPARAM_SOFTLOCKUP_PANIC
4275 which is the respective build-time switch to that
4276 functionality.
4278 softlockup_all_cpu_backtrace=
4279 [KNL] Should the soft-lockup detector generate
4280 backtraces on all cpus.
4281 Format: <integer>
4283 sonypi.*= [HW] Sony Programmable I/O Control Device driver
4284 See Documentation/laptops/sonypi.txt
4286 spectre_v2= [X86] Control mitigation of Spectre variant 2
4287 (indirect branch speculation) vulnerability.
4288 The default operation protects the kernel from
4289 user space attacks.
4291 on - unconditionally enable, implies
4292 spectre_v2_user=on
4293 off - unconditionally disable, implies
4294 spectre_v2_user=off
4295 auto - kernel detects whether your CPU model is
4296 vulnerable
4298 Selecting 'on' will, and 'auto' may, choose a
4299 mitigation method at run time according to the
4300 CPU, the available microcode, the setting of the
4301 CONFIG_RETPOLINE configuration option, and the
4302 compiler with which the kernel was built.
4304 Selecting 'on' will also enable the mitigation
4305 against user space to user space task attacks.
4307 Selecting 'off' will disable both the kernel and
4308 the user space protections.
4310 Specific mitigations can also be selected manually:
4312 retpoline - replace indirect branches
4313 retpoline,generic - google's original retpoline
4314 retpoline,amd - AMD-specific minimal thunk
4316 Not specifying this option is equivalent to
4317 spectre_v2=auto.
4319 spectre_v2_user=
4320 [X86] Control mitigation of Spectre variant 2
4321 (indirect branch speculation) vulnerability between
4322 user space tasks
4324 on - Unconditionally enable mitigations. Is
4325 enforced by spectre_v2=on
4327 off - Unconditionally disable mitigations. Is
4328 enforced by spectre_v2=off
4330 prctl - Indirect branch speculation is enabled,
4331 but mitigation can be enabled via prctl
4332 per thread. The mitigation control state
4333 is inherited on fork.
4335 prctl,ibpb
4336 - Like "prctl" above, but only STIBP is
4337 controlled per thread. IBPB is issued
4338 always when switching between different user
4339 space processes.
4341 seccomp
4342 - Same as "prctl" above, but all seccomp
4343 threads will enable the mitigation unless
4344 they explicitly opt out.
4346 seccomp,ibpb
4347 - Like "seccomp" above, but only STIBP is
4348 controlled per thread. IBPB is issued
4349 always when switching between different
4350 user space processes.
4352 auto - Kernel selects the mitigation depending on
4353 the available CPU features and vulnerability.
4355 Default mitigation:
4356 If CONFIG_SECCOMP=y then "seccomp", otherwise "prctl"
4358 Not specifying this option is equivalent to
4359 spectre_v2_user=auto.
4361 spec_store_bypass_disable=
4362 [HW] Control Speculative Store Bypass (SSB) Disable mitigation
4363 (Speculative Store Bypass vulnerability)
4365 Certain CPUs are vulnerable to an exploit against a
4366 a common industry wide performance optimization known
4367 as "Speculative Store Bypass" in which recent stores
4368 to the same memory location may not be observed by
4369 later loads during speculative execution. The idea
4370 is that such stores are unlikely and that they can
4371 be detected prior to instruction retirement at the
4372 end of a particular speculation execution window.
4374 In vulnerable processors, the speculatively forwarded
4375 store can be used in a cache side channel attack, for
4376 example to read memory to which the attacker does not
4377 directly have access (e.g. inside sandboxed code).
4379 This parameter controls whether the Speculative Store
4380 Bypass optimization is used.
4382 On x86 the options are:
4384 on - Unconditionally disable Speculative Store Bypass
4385 off - Unconditionally enable Speculative Store Bypass
4386 auto - Kernel detects whether the CPU model contains an
4387 implementation of Speculative Store Bypass and
4388 picks the most appropriate mitigation. If the
4389 CPU is not vulnerable, "off" is selected. If the
4390 CPU is vulnerable the default mitigation is
4391 architecture and Kconfig dependent. See below.
4392 prctl - Control Speculative Store Bypass per thread
4393 via prctl. Speculative Store Bypass is enabled
4394 for a process by default. The state of the control
4395 is inherited on fork.
4396 seccomp - Same as "prctl" above, but all seccomp threads
4397 will disable SSB unless they explicitly opt out.
4399 Default mitigations:
4400 X86: If CONFIG_SECCOMP=y "seccomp", otherwise "prctl"
4402 On powerpc the options are:
4404 on,auto - On Power8 and Power9 insert a store-forwarding
4405 barrier on kernel entry and exit. On Power7
4406 perform a software flush on kernel entry and
4407 exit.
4408 off - No action.
4410 Not specifying this option is equivalent to
4411 spec_store_bypass_disable=auto.
4413 spia_io_base= [HW,MTD]
4414 spia_fio_base=
4415 spia_pedr=
4416 spia_peddr=
4418 srbds= [X86,INTEL]
4419 Control the Special Register Buffer Data Sampling
4420 (SRBDS) mitigation.
4422 Certain CPUs are vulnerable to an MDS-like
4423 exploit which can leak bits from the random
4424 number generator.
4426 By default, this issue is mitigated by
4427 microcode. However, the microcode fix can cause
4428 the RDRAND and RDSEED instructions to become
4429 much slower. Among other effects, this will
4430 result in reduced throughput from /dev/urandom.
4432 The microcode mitigation can be disabled with
4433 the following option:
4435 off: Disable mitigation and remove
4436 performance impact to RDRAND and RDSEED
4438 srcutree.counter_wrap_check [KNL]
4439 Specifies how frequently to check for
4440 grace-period sequence counter wrap for the
4441 srcu_data structure's ->srcu_gp_seq_needed field.
4442 The greater the number of bits set in this kernel
4443 parameter, the less frequently counter wrap will
4444 be checked for. Note that the bottom two bits
4445 are ignored.
4447 srcutree.exp_holdoff [KNL]
4448 Specifies how many nanoseconds must elapse
4449 since the end of the last SRCU grace period for
4450 a given srcu_struct until the next normal SRCU
4451 grace period will be considered for automatic
4452 expediting. Set to zero to disable automatic
4453 expediting.
4455 ssbd= [ARM64,HW]
4456 Speculative Store Bypass Disable control
4458 On CPUs that are vulnerable to the Speculative
4459 Store Bypass vulnerability and offer a
4460 firmware based mitigation, this parameter
4461 indicates how the mitigation should be used:
4463 force-on: Unconditionally enable mitigation for
4464 for both kernel and userspace
4465 force-off: Unconditionally disable mitigation for
4466 for both kernel and userspace
4467 kernel: Always enable mitigation in the
4468 kernel, and offer a prctl interface
4469 to allow userspace to register its
4470 interest in being mitigated too.
4472 stack_guard_gap= [MM]
4473 override the default stack gap protection. The value
4474 is in page units and it defines how many pages prior
4475 to (for stacks growing down) resp. after (for stacks
4476 growing up) the main stack are reserved for no other
4477 mapping. Default value is 256 pages.
4479 stacktrace [FTRACE]
4480 Enabled the stack tracer on boot up.
4482 stacktrace_filter=[function-list]
4483 [FTRACE] Limit the functions that the stack tracer
4484 will trace at boot up. function-list is a comma separated
4485 list of functions. This list can be changed at run
4486 time by the stack_trace_filter file in the debugfs
4487 tracing directory. Note, this enables stack tracing
4488 and the stacktrace above is not needed.
4490 sti= [PARISC,HW]
4491 Format: <num>
4492 Set the STI (builtin display/keyboard on the HP-PARISC
4493 machines) console (graphic card) which should be used
4494 as the initial boot-console.
4495 See also comment in drivers/video/console/sticore.c.
4497 sti_font= [HW]
4498 See comment in drivers/video/console/sticore.c.
4500 stifb= [HW]
4501 Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]
4503 sunrpc.min_resvport=
4504 sunrpc.max_resvport=
4505 [NFS,SUNRPC]
4506 SunRPC servers often require that client requests
4507 originate from a privileged port (i.e. a port in the
4508 range 0 < portnr < 1024).
4509 An administrator who wishes to reserve some of these
4510 ports for other uses may adjust the range that the
4511 kernel's sunrpc client considers to be privileged
4512 using these two parameters to set the minimum and
4513 maximum port values.
4515 sunrpc.svc_rpc_per_connection_limit=
4516 [NFS,SUNRPC]
4517 Limit the number of requests that the server will
4518 process in parallel from a single connection.
4519 The default value is 0 (no limit).
4521 sunrpc.pool_mode=
4522 [NFS]
4523 Control how the NFS server code allocates CPUs to
4524 service thread pools. Depending on how many NICs
4525 you have and where their interrupts are bound, this
4526 option will affect which CPUs will do NFS serving.
4527 Note: this parameter cannot be changed while the
4528 NFS server is running.
4530 auto the server chooses an appropriate mode
4531 automatically using heuristics
4532 global a single global pool contains all CPUs
4533 percpu one pool for each CPU
4534 pernode one pool for each NUMA node (equivalent
4535 to global on non-NUMA machines)
4537 sunrpc.tcp_slot_table_entries=
4538 sunrpc.udp_slot_table_entries=
4539 [NFS,SUNRPC]
4540 Sets the upper limit on the number of simultaneous
4541 RPC calls that can be sent from the client to a
4542 server. Increasing these values may allow you to
4543 improve throughput, but will also increase the
4544 amount of memory reserved for use by the client.
4546 suspend.pm_test_delay=
4547 [SUSPEND]
4548 Sets the number of seconds to remain in a suspend test
4549 mode before resuming the system (see
4550 /sys/power/pm_test). Only available when CONFIG_PM_DEBUG
4551 is set. Default value is 5.
4553 swapaccount=[0|1]
4554 [KNL] Enable accounting of swap in memory resource
4555 controller if no parameter or 1 is given or disable
4556 it if 0 is given (See Documentation/cgroup-v1/memory.txt)
4558 swiotlb= [ARM,IA-64,PPC,MIPS,X86]
4559 Format: { <int> | force | noforce }
4560 <int> -- Number of I/O TLB slabs
4561 force -- force using of bounce buffers even if they
4562 wouldn't be automatically used by the kernel
4563 noforce -- Never use bounce buffers (for debugging)
4565 switches= [HW,M68k]
4567 sysfs.deprecated=0|1 [KNL]
4568 Enable/disable old style sysfs layout for old udev
4569 on older distributions. When this option is enabled
4570 very new udev will not work anymore. When this option
4571 is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
4572 in older udev will not work anymore.
4573 Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
4574 the kernel configuration.
4576 sysrq_always_enabled
4577 [KNL]
4578 Ignore sysrq setting - this boot parameter will
4579 neutralize any effect of /proc/sys/kernel/sysrq.
4580 Useful for debugging.
4582 tcpmhash_entries= [KNL,NET]
4583 Set the number of tcp_metrics_hash slots.
4584 Default value is 8192 or 16384 depending on total
4585 ram pages. This is used to specify the TCP metrics
4586 cache size. See Documentation/networking/ip-sysctl.txt
4587 "tcp_no_metrics_save" section for more details.
4589 tdfx= [HW,DRM]
4591 test_suspend= [SUSPEND][,N]
4592 Specify "mem" (for Suspend-to-RAM) or "standby" (for
4593 standby suspend) or "freeze" (for suspend type freeze)
4594 as the system sleep state during system startup with
4595 the optional capability to repeat N number of times.
4596 The system is woken from this state using a
4597 wakeup-capable RTC alarm.
4599 thash_entries= [KNL,NET]
4600 Set number of hash buckets for TCP connection
4602 thermal.act= [HW,ACPI]
4603 -1: disable all active trip points in all thermal zones
4604 <degrees C>: override all lowest active trip points
4606 thermal.crt= [HW,ACPI]
4607 -1: disable all critical trip points in all thermal zones
4608 <degrees C>: override all critical trip points
4610 thermal.nocrt= [HW,ACPI]
4611 Set to disable actions on ACPI thermal zone
4612 critical and hot trip points.
4614 thermal.off= [HW,ACPI]
4615 1: disable ACPI thermal control
4617 thermal.psv= [HW,ACPI]
4618 -1: disable all passive trip points
4619 <degrees C>: override all passive trip points to this
4620 value
4622 thermal.tzp= [HW,ACPI]
4623 Specify global default ACPI thermal zone polling rate
4624 <deci-seconds>: poll all this frequency
4625 0: no polling (default)
4627 threadirqs [KNL]
4628 Force threading of all interrupt handlers except those
4629 marked explicitly IRQF_NO_THREAD.
4631 tmem [KNL,XEN]
4632 Enable the Transcendent memory driver if built-in.
4634 tmem.cleancache=0|1 [KNL, XEN]
4635 Default is on (1). Disable the usage of the cleancache
4636 API to send anonymous pages to the hypervisor.
4638 tmem.frontswap=0|1 [KNL, XEN]
4639 Default is on (1). Disable the usage of the frontswap
4640 API to send swap pages to the hypervisor. If disabled
4641 the selfballooning and selfshrinking are force disabled.
4643 tmem.selfballooning=0|1 [KNL, XEN]
4644 Default is on (1). Disable the driving of swap pages
4645 to the hypervisor.
4647 tmem.selfshrinking=0|1 [KNL, XEN]
4648 Default is on (1). Partial swapoff that immediately
4649 transfers pages from Xen hypervisor back to the
4650 kernel based on different criteria.
4652 topology= [S390]
4653 Format: {off | on}
4654 Specify if the kernel should make use of the cpu
4655 topology information if the hardware supports this.
4656 The scheduler will make use of this information and
4657 e.g. base its process migration decisions on it.
4658 Default is on.
4660 topology_updates= [KNL, PPC, NUMA]
4661 Format: {off}
4662 Specify if the kernel should ignore (off)
4663 topology updates sent by the hypervisor to this
4664 LPAR.
4666 tp720= [HW,PS2]
4668 tpm_suspend_pcr=[HW,TPM]
4669 Format: integer pcr id
4670 Specify that at suspend time, the tpm driver
4671 should extend the specified pcr with zeros,
4672 as a workaround for some chips which fail to
4673 flush the last written pcr on TPM_SaveState.
4674 This will guarantee that all the other pcrs
4675 are saved.
4677 trace_buf_size=nn[KMG]
4678 [FTRACE] will set tracing buffer size on each cpu.
4680 trace_event=[event-list]
4681 [FTRACE] Set and start specified trace events in order
4682 to facilitate early boot debugging. The event-list is a
4683 comma separated list of trace events to enable. See
4684 also Documentation/trace/events.rst
4686 trace_options=[option-list]
4687 [FTRACE] Enable or disable tracer options at boot.
4688 The option-list is a comma delimited list of options
4689 that can be enabled or disabled just as if you were
4690 to echo the option name into
4692 /sys/kernel/debug/tracing/trace_options
4694 For example, to enable stacktrace option (to dump the
4695 stack trace of each event), add to the command line:
4697 trace_options=stacktrace
4699 See also Documentation/trace/ftrace.rst "trace options"
4700 section.
4702 tp_printk[FTRACE]
4703 Have the tracepoints sent to printk as well as the
4704 tracing ring buffer. This is useful for early boot up
4705 where the system hangs or reboots and does not give the
4706 option for reading the tracing buffer or performing a
4707 ftrace_dump_on_oops.
4709 To turn off having tracepoints sent to printk,
4710 echo 0 > /proc/sys/kernel/tracepoint_printk
4711 Note, echoing 1 into this file without the
4712 tracepoint_printk kernel cmdline option has no effect.
4714 ** CAUTION **
4716 Having tracepoints sent to printk() and activating high
4717 frequency tracepoints such as irq or sched, can cause
4718 the system to live lock.
4720 traceoff_on_warning
4721 [FTRACE] enable this option to disable tracing when a
4722 warning is hit. This turns off "tracing_on". Tracing can
4723 be enabled again by echoing '1' into the "tracing_on"
4724 file located in /sys/kernel/debug/tracing/
4726 This option is useful, as it disables the trace before
4727 the WARNING dump is called, which prevents the trace to
4728 be filled with content caused by the warning output.
4730 This option can also be set at run time via the sysctl
4731 option: kernel/traceoff_on_warning
4733 transparent_hugepage=
4734 [KNL]
4735 Format: [always|madvise|never]
4736 Can be used to control the default behavior of the system
4737 with respect to transparent hugepages.
4738 See Documentation/admin-guide/mm/transhuge.rst
4739 for more details.
4741 tsc= Disable clocksource stability checks for TSC.
4742 Format: <string>
4743 [x86] reliable: mark tsc clocksource as reliable, this
4744 disables clocksource verification at runtime, as well
4745 as the stability checks done at bootup. Used to enable
4746 high-resolution timer mode on older hardware, and in
4747 virtualized environment.
4748 [x86] noirqtime: Do not use TSC to do irq accounting.
4749 Used to run time disable IRQ_TIME_ACCOUNTING on any
4750 platforms where RDTSC is slow and this accounting
4751 can add overhead.
4752 [x86] unstable: mark the TSC clocksource as unstable, this
4753 marks the TSC unconditionally unstable at bootup and
4754 avoids any further wobbles once the TSC watchdog notices.
4756 tsx= [X86] Control Transactional Synchronization
4757 Extensions (TSX) feature in Intel processors that
4758 support TSX control.
4760 This parameter controls the TSX feature. The options are:
4762 on - Enable TSX on the system. Although there are
4763 mitigations for all known security vulnerabilities,
4764 TSX has been known to be an accelerator for
4765 several previous speculation-related CVEs, and
4766 so there may be unknown security risks associated
4767 with leaving it enabled.
4769 off - Disable TSX on the system. (Note that this
4770 option takes effect only on newer CPUs which are
4771 not vulnerable to MDS, i.e., have
4772 MSR_IA32_ARCH_CAPABILITIES.MDS_NO=1 and which get
4773 the new IA32_TSX_CTRL MSR through a microcode
4774 update. This new MSR allows for the reliable
4775 deactivation of the TSX functionality.)
4777 auto - Disable TSX if X86_BUG_TAA is present,
4778 otherwise enable TSX on the system.
4780 Not specifying this option is equivalent to tsx=off.
4782 See Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
4783 for more details.
4785 tsx_async_abort= [X86,INTEL] Control mitigation for the TSX Async
4786 Abort (TAA) vulnerability.
4788 Similar to Micro-architectural Data Sampling (MDS)
4789 certain CPUs that support Transactional
4790 Synchronization Extensions (TSX) are vulnerable to an
4791 exploit against CPU internal buffers which can forward
4792 information to a disclosure gadget under certain
4793 conditions.
4795 In vulnerable processors, the speculatively forwarded
4796 data can be used in a cache side channel attack, to
4797 access data to which the attacker does not have direct
4798 access.
4800 This parameter controls the TAA mitigation. The
4801 options are:
4803 full - Enable TAA mitigation on vulnerable CPUs
4804 if TSX is enabled.
4806 full,nosmt - Enable TAA mitigation and disable SMT on
4807 vulnerable CPUs. If TSX is disabled, SMT
4808 is not disabled because CPU is not
4809 vulnerable to cross-thread TAA attacks.
4810 off - Unconditionally disable TAA mitigation
4812 On MDS-affected machines, tsx_async_abort=off can be
4813 prevented by an active MDS mitigation as both vulnerabilities
4814 are mitigated with the same mechanism so in order to disable
4815 this mitigation, you need to specify mds=off too.
4817 Not specifying this option is equivalent to
4818 tsx_async_abort=full. On CPUs which are MDS affected
4819 and deploy MDS mitigation, TAA mitigation is not
4820 required and doesn't provide any additional
4821 mitigation.
4823 For details see:
4824 Documentation/admin-guide/hw-vuln/tsx_async_abort.rst
4826 turbografx.map[2|3]= [HW,JOY]
4827 TurboGraFX parallel port interface
4828 Format:
4829 <port#>,<js1>,<js2>,<js3>,<js4>,<js5>,<js6>,<js7>
4830 See also Documentation/input/devices/joystick-parport.rst
4832 udbg-immortal [PPC] When debugging early kernel crashes that
4833 happen after console_init() and before a proper
4834 console driver takes over, this boot options might
4835 help "seeing" what's going on.
4837 uhash_entries= [KNL,NET]
4838 Set number of hash buckets for UDP/UDP-Lite connections
4840 uhci-hcd.ignore_oc=
4841 [USB] Ignore overcurrent events (default N).
4842 Some badly-designed motherboards generate lots of
4843 bogus events, for ports that aren't wired to
4844 anything. Set this parameter to avoid log spamming.
4845 Note that genuine overcurrent events won't be
4846 reported either.
4848 unknown_nmi_panic
4849 [X86] Cause panic on unknown NMI.
4851 usbcore.authorized_default=
4852 [USB] Default USB device authorization:
4853 (default -1 = authorized except for wireless USB,
4854 0 = not authorized, 1 = authorized)
4856 usbcore.autosuspend=
4857 [USB] The autosuspend time delay (in seconds) used
4858 for newly-detected USB devices (default 2). This
4859 is the time required before an idle device will be
4860 autosuspended. Devices for which the delay is set
4861 to a negative value won't be autosuspended at all.
4863 usbcore.usbfs_snoop=
4864 [USB] Set to log all usbfs traffic (default 0 = off).
4866 usbcore.usbfs_snoop_max=
4867 [USB] Maximum number of bytes to snoop in each URB
4868 (default = 65536).
4870 usbcore.blinkenlights=
4871 [USB] Set to cycle leds on hubs (default 0 = off).
4873 usbcore.old_scheme_first=
4874 [USB] Start with the old device initialization
4875 scheme (default 0 = off).
4877 usbcore.usbfs_memory_mb=
4878 [USB] Memory limit (in MB) for buffers allocated by
4879 usbfs (default = 16, 0 = max = 2047).
4881 usbcore.use_both_schemes=
4882 [USB] Try the other device initialization scheme
4883 if the first one fails (default 1 = enabled).
4885 usbcore.initial_descriptor_timeout=
4886 [USB] Specifies timeout for the initial 64-byte
4887 USB_REQ_GET_DESCRIPTOR request in milliseconds
4888 (default 5000 = 5.0 seconds).
4890 usbcore.nousb [USB] Disable the USB subsystem
4892 usbcore.quirks=
4893 [USB] A list of quirk entries to augment the built-in
4894 usb core quirk list. List entries are separated by
4895 commas. Each entry has the form
4896 VendorID:ProductID:Flags. The IDs are 4-digit hex
4897 numbers and Flags is a set of letters. Each letter
4898 will change the built-in quirk; setting it if it is
4899 clear and clearing it if it is set. The letters have
4900 the following meanings:
4901 a = USB_QUIRK_STRING_FETCH_255 (string
4902 descriptors must not be fetched using
4903 a 255-byte read);
4904 b = USB_QUIRK_RESET_RESUME (device can't resume
4905 correctly so reset it instead);
4906 c = USB_QUIRK_NO_SET_INTF (device can't handle
4907 Set-Interface requests);
4908 d = USB_QUIRK_CONFIG_INTF_STRINGS (device can't
4909 handle its Configuration or Interface
4910 strings);
4911 e = USB_QUIRK_RESET (device can't be reset
4912 (e.g morph devices), don't use reset);
4913 f = USB_QUIRK_HONOR_BNUMINTERFACES (device has
4914 more interface descriptions than the
4915 bNumInterfaces count, and can't handle
4916 talking to these interfaces);
4917 g = USB_QUIRK_DELAY_INIT (device needs a pause
4918 during initialization, after we read
4919 the device descriptor);
4920 h = USB_QUIRK_LINEAR_UFRAME_INTR_BINTERVAL (For
4921 high speed and super speed interrupt
4922 endpoints, the USB 2.0 and USB 3.0 spec
4923 require the interval in microframes (1
4924 microframe = 125 microseconds) to be
4925 calculated as interval = 2 ^
4926 (bInterval-1).
4927 Devices with this quirk report their
4928 bInterval as the result of this
4929 calculation instead of the exponent
4930 variable used in the calculation);
4931 i = USB_QUIRK_DEVICE_QUALIFIER (device can't
4932 handle device_qualifier descriptor
4933 requests);
4934 j = USB_QUIRK_IGNORE_REMOTE_WAKEUP (device
4935 generates spurious wakeup, ignore
4936 remote wakeup capability);
4937 k = USB_QUIRK_NO_LPM (device can't handle Link
4938 Power Management);
4939 l = USB_QUIRK_LINEAR_FRAME_INTR_BINTERVAL
4940 (Device reports its bInterval as linear
4941 frames instead of the USB 2.0
4942 calculation);
4943 m = USB_QUIRK_DISCONNECT_SUSPEND (Device needs
4944 to be disconnected before suspend to
4945 prevent spurious wakeup);
4946 n = USB_QUIRK_DELAY_CTRL_MSG (Device needs a
4947 pause after every control message);
4948 o = USB_QUIRK_HUB_SLOW_RESET (Hub needs extra
4949 delay after resetting its port);
4950 Example: quirks=0781:5580:bk,0a5c:5834:gij
4952 usbhid.mousepoll=
4953 [USBHID] The interval which mice are to be polled at.
4955 usbhid.jspoll=
4956 [USBHID] The interval which joysticks are to be polled at.
4958 usbhid.kbpoll=
4959 [USBHID] The interval which keyboards are to be polled at.
4961 usb-storage.delay_use=
4962 [UMS] The delay in seconds before a new device is
4963 scanned for Logical Units (default 1).
4965 usb-storage.quirks=
4966 [UMS] A list of quirks entries to supplement or
4967 override the built-in unusual_devs list. List
4968 entries are separated by commas. Each entry has
4969 the form VID:PID:Flags where VID and PID are Vendor
4970 and Product ID values (4-digit hex numbers) and
4971 Flags is a set of characters, each corresponding
4972 to a common usb-storage quirk flag as follows:
4973 a = SANE_SENSE (collect more than 18 bytes
4974 of sense data, not on uas);
4975 b = BAD_SENSE (don't collect more than 18
4976 bytes of sense data, not on uas);
4977 c = FIX_CAPACITY (decrease the reported
4978 device capacity by one sector);
4979 d = NO_READ_DISC_INFO (don't use
4980 READ_DISC_INFO command, not on uas);
4981 e = NO_READ_CAPACITY_16 (don't use
4982 READ_CAPACITY_16 command);
4983 f = NO_REPORT_OPCODES (don't use report opcodes
4984 command, uas only);
4985 g = MAX_SECTORS_240 (don't transfer more than
4986 240 sectors at a time, uas only);
4987 h = CAPACITY_HEURISTICS (decrease the
4988 reported device capacity by one
4989 sector if the number is odd);
4990 i = IGNORE_DEVICE (don't bind to this
4991 device);
4992 j = NO_REPORT_LUNS (don't use report luns
4993 command, uas only);
4994 l = NOT_LOCKABLE (don't try to lock and
4995 unlock ejectable media, not on uas);
4996 m = MAX_SECTORS_64 (don't transfer more
4997 than 64 sectors = 32 KB at a time,
4998 not on uas);
4999 n = INITIAL_READ10 (force a retry of the
5000 initial READ(10) command, not on uas);
5001 o = CAPACITY_OK (accept the capacity
5002 reported by the device, not on uas);
5003 p = WRITE_CACHE (the device cache is ON
5004 by default, not on uas);
5005 r = IGNORE_RESIDUE (the device reports
5006 bogus residue values, not on uas);
5007 s = SINGLE_LUN (the device has only one
5008 Logical Unit);
5009 t = NO_ATA_1X (don't allow ATA(12) and ATA(16)
5010 commands, uas only);
5011 u = IGNORE_UAS (don't bind to the uas driver);
5012 w = NO_WP_DETECT (don't test whether the
5013 medium is write-protected).
5014 y = ALWAYS_SYNC (issue a SYNCHRONIZE_CACHE
5015 even if the device claims no cache,
5016 not on uas)
5017 Example: quirks=0419:aaf5:rl,0421:0433:rc
5019 user_debug= [KNL,ARM]
5020 Format: <int>
5021 See arch/arm/Kconfig.debug help text.
5022 1 - undefined instruction events
5023 2 - system calls
5024 4 - invalid data aborts
5025 8 - SIGSEGV faults
5026 16 - SIGBUS faults
5027 Example: user_debug=31
5029 userpte=
5030 [X86] Flags controlling user PTE allocations.
5032 nohigh = do not allocate PTE pages in
5033 HIGHMEM regardless of setting
5034 of CONFIG_HIGHPTE.
5036 vdso= [X86,SH]
5037 On X86_32, this is an alias for vdso32=. Otherwise:
5039 vdso=1: enable VDSO (the default)
5040 vdso=0: disable VDSO mapping
5042 vdso32= [X86] Control the 32-bit vDSO
5043 vdso32=1: enable 32-bit VDSO
5044 vdso32=0 or vdso32=2: disable 32-bit VDSO
5046 See the help text for CONFIG_COMPAT_VDSO for more
5047 details. If CONFIG_COMPAT_VDSO is set, the default is
5048 vdso32=0; otherwise, the default is vdso32=1.
5050 For compatibility with older kernels, vdso32=2 is an
5051 alias for vdso32=0.
5053 Try vdso32=0 if you encounter an error that says:
5054 dl_main: Assertion `(void *) ph->p_vaddr == _rtld_local._dl_sysinfo_dso' failed!
5056 vector= [IA-64,SMP]
5057 vector=percpu: enable percpu vector domain
5059 video= [FB] Frame buffer configuration
5060 See Documentation/fb/modedb.txt.
5062 video.brightness_switch_enabled= [0,1]
5063 If set to 1, on receiving an ACPI notify event
5064 generated by hotkey, video driver will adjust brightness
5065 level and then send out the event to user space through
5066 the allocated input device; If set to 0, video driver
5067 will only send out the event without touching backlight
5068 brightness level.
5069 default: 1
5071 virtio_mmio.device=
5072 [VMMIO] Memory mapped virtio (platform) device.
5074 <size>@<baseaddr>:<irq>[:<id>]
5075 where:
5076 <size> := size (can use standard suffixes
5077 like K, M and G)
5078 <baseaddr> := physical base address
5079 <irq> := interrupt number (as passed to
5080 request_irq())
5081 <id> := (optional) platform device id
5082 example:
5083 virtio_mmio.device=1K@0x100b0000:48:7
5085 Can be used multiple times for multiple devices.
5087 vga= [BOOT,X86-32] Select a particular video mode
5088 See Documentation/x86/boot.txt and
5089 Documentation/svga.txt.
5090 Use vga=ask for menu.
5091 This is actually a boot loader parameter; the value is
5092 passed to the kernel using a special protocol.
5094 vmalloc=nn[KMG] [KNL,BOOT] Forces the vmalloc area to have an exact
5095 size of <nn>. This can be used to increase the
5096 minimum size (128MB on x86). It can also be used to
5097 decrease the size and leave more room for directly
5098 mapped kernel RAM.
5100 vmcp_cma=nn[MG] [KNL,S390]
5101 Sets the memory size reserved for contiguous memory
5102 allocations for the vmcp device driver.
5104 vmhalt= [KNL,S390] Perform z/VM CP command after system halt.
5105 Format: <command>
5107 vmpanic= [KNL,S390] Perform z/VM CP command after kernel panic.
5108 Format: <command>
5110 vmpoff= [KNL,S390] Perform z/VM CP command after power off.
5111 Format: <command>
5113 vsyscall= [X86-64]
5114 Controls the behavior of vsyscalls (i.e. calls to
5115 fixed addresses of 0xffffffffff600x00 from legacy
5116 code). Most statically-linked binaries and older
5117 versions of glibc use these calls. Because these
5118 functions are at fixed addresses, they make nice
5119 targets for exploits that can control RIP.
5121 emulate [default] Vsyscalls turn into traps and are
5122 emulated reasonably safely.
5124 none Vsyscalls don't work at all. This makes
5125 them quite hard to use for exploits but
5126 might break your system.
5128 vt.color= [VT] Default text color.
5129 Format: 0xYX, X = foreground, Y = background.
5130 Default: 0x07 = light gray on black.
5132 vt.cur_default= [VT] Default cursor shape.
5133 Format: 0xCCBBAA, where AA, BB, and CC are the same as
5134 the parameters of the <Esc>[?A;B;Cc escape sequence;
5135 see VGA-softcursor.txt. Default: 2 = underline.
5137 vt.default_blu= [VT]
5138 Format: <blue0>,<blue1>,<blue2>,...,<blue15>
5139 Change the default blue palette of the console.
5140 This is a 16-member array composed of values
5141 ranging from 0-255.
5143 vt.default_grn= [VT]
5144 Format: <green0>,<green1>,<green2>,...,<green15>
5145 Change the default green palette of the console.
5146 This is a 16-member array composed of values
5147 ranging from 0-255.
5149 vt.default_red= [VT]
5150 Format: <red0>,<red1>,<red2>,...,<red15>
5151 Change the default red palette of the console.
5152 This is a 16-member array composed of values
5153 ranging from 0-255.
5155 vt.default_utf8=
5156 [VT]
5157 Format=<0|1>
5158 Set system-wide default UTF-8 mode for all tty's.
5159 Default is 1, i.e. UTF-8 mode is enabled for all
5160 newly opened terminals.
5162 vt.global_cursor_default=
5163 [VT]
5164 Format=<-1|0|1>
5165 Set system-wide default for whether a cursor
5166 is shown on new VTs. Default is -1,
5167 i.e. cursors will be created by default unless
5168 overridden by individual drivers. 0 will hide
5169 cursors, 1 will display them.
5171 vt.italic= [VT] Default color for italic text; 0-15.
5172 Default: 2 = green.
5174 vt.underline= [VT] Default color for underlined text; 0-15.
5175 Default: 3 = cyan.
5177 watchdog timers [HW,WDT] For information on watchdog timers,
5178 see Documentation/watchdog/watchdog-parameters.txt
5179 or other driver-specific files in the
5180 Documentation/watchdog/ directory.
5182 workqueue.watchdog_thresh=
5183 If CONFIG_WQ_WATCHDOG is configured, workqueue can
5184 warn stall conditions and dump internal state to
5185 help debugging. 0 disables workqueue stall
5186 detection; otherwise, it's the stall threshold
5187 duration in seconds. The default value is 30 and
5188 it can be updated at runtime by writing to the
5189 corresponding sysfs file.
5191 workqueue.disable_numa
5192 By default, all work items queued to unbound
5193 workqueues are affine to the NUMA nodes they're
5194 issued on, which results in better behavior in
5195 general. If NUMA affinity needs to be disabled for
5196 whatever reason, this option can be used. Note
5197 that this also can be controlled per-workqueue for
5198 workqueues visible under /sys/bus/workqueue/.
5200 workqueue.power_efficient
5201 Per-cpu workqueues are generally preferred because
5202 they show better performance thanks to cache
5203 locality; unfortunately, per-cpu workqueues tend to
5204 be more power hungry than unbound workqueues.
5206 Enabling this makes the per-cpu workqueues which
5207 were observed to contribute significantly to power
5208 consumption unbound, leading to measurably lower
5209 power usage at the cost of small performance
5210 overhead.
5212 The default value of this parameter is determined by
5213 the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT.
5215 workqueue.debug_force_rr_cpu
5216 Workqueue used to implicitly guarantee that work
5217 items queued without explicit CPU specified are put
5218 on the local CPU. This guarantee is no longer true
5219 and while local CPU is still preferred work items
5220 may be put on foreign CPUs. This debug option
5221 forces round-robin CPU selection to flush out
5222 usages which depend on the now broken guarantee.
5223 When enabled, memory and cache locality will be
5224 impacted.
5226 x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of
5227 default x2apic cluster mode on platforms
5228 supporting x2apic.
5230 x86_intel_mid_timer= [X86-32,APBT]
5231 Choose timer option for x86 Intel MID platform.
5232 Two valid options are apbt timer only and lapic timer
5233 plus one apbt timer for broadcast timer.
5234 x86_intel_mid_timer=apbt_only | lapic_and_apbt
5236 xen_512gb_limit [KNL,X86-64,XEN]
5237 Restricts the kernel running paravirtualized under Xen
5238 to use only up to 512 GB of RAM. The reason to do so is
5239 crash analysis tools and Xen tools for doing domain
5240 save/restore/migration must be enabled to handle larger
5241 domains.
5243 xen_emul_unplug= [HW,X86,XEN]
5244 Unplug Xen emulated devices
5245 Format: [unplug0,][unplug1]
5246 ide-disks -- unplug primary master IDE devices
5247 aux-ide-disks -- unplug non-primary-master IDE devices
5248 nics -- unplug network devices
5249 all -- unplug all emulated devices (NICs and IDE disks)
5250 unnecessary -- unplugging emulated devices is
5251 unnecessary even if the host did not respond to
5252 the unplug protocol
5253 never -- do not unplug even if version check succeeds
5255 xen_legacy_crash [X86,XEN]
5256 Crash from Xen panic notifier, without executing late
5257 panic() code such as dumping handler.
5259 xen_nopvspin [X86,XEN]
5260 Disables the ticketlock slowpath using Xen PV
5261 optimizations.
5263 xen_nopv [X86]
5264 Disables the PV optimizations forcing the HVM guest to
5265 run as generic HVM guest with no PV drivers.
5267 xen_scrub_pages= [XEN]
5268 Boolean option to control scrubbing pages before giving them back
5269 to Xen, for use by other domains. Can be also changed at runtime
5270 with /sys/devices/system/xen_memory/xen_memory0/scrub_pages.
5271 Default value controlled with CONFIG_XEN_SCRUB_PAGES_DEFAULT.
5273 xirc2ps_cs= [NET,PCMCIA]
5274 Format:
5275 <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]]
5277 xhci-hcd.quirks [USB,KNL]
5278 A hex value specifying bitmask with supplemental xhci
5279 host controller quirks. Meaning of each bit can be
5280 consulted in header drivers/usb/host/xhci.h.