Linux Kernel Parameters

삽자루 | 2012.07.07 09:54 | 조회 16500
                          Kernel Parameters

The following is a consolidated list of the kernel parameters as implemented
(mostly) by the __setup() macro and sorted into English Dictionary order
(defined as ignoring all punctuation and sorting digits before letters in a
case insensitive manner), and with descriptions where known.

Module parameters for loadable modules are specified only as the
parameter name with optional '=' and value as appropriate, such as:

    modprobe usbcore blinkenlights=1

Module parameters for modules that are built into the kernel image
are specified on the kernel command line with the module name plus
'.' plus parameter name, with '=' and value if appropriate, such as:


Hyphens (dashes) and underscores are equivalent in parameter names, so
    log_buf_len=1M print-fatal-signals=1
can also be entered as
    log-buf-len=1M print_fatal_signals=1

This document may not be entirely up to date and comprehensive. The command
"modinfo -p ${modulename}" shows a current list of all parameters of a loadable
module. Loadable modules, after being loaded into the running kernel, also
reveal their parameters in /sys/module/${modulename}/parameters/. Some of these
parameters may be changed at runtime by the command
"echo -n ${value} > /sys/module/${modulename}/parameters/${parm}".

The parameters listed below are only valid if certain kernel build options were
enabled and if respective hardware is present. The text in square brackets at
the beginning of each description states the restrictions within which a
parameter is applicable:

    ACPI    ACPI support is enabled.
    AGP    AGP (Accelerated Graphics Port) is enabled.
    ALSA    ALSA sound support is enabled.
    APIC    APIC support is enabled.
    APM    Advanced Power Management support is enabled.
    AVR32    AVR32 architecture is enabled.
    AX25    Appropriate AX.25 support is enabled.
    BLACKFIN Blackfin architecture is enabled.
    DRM    Direct Rendering Management support is enabled.
    DYNAMIC_DEBUG Build in debug messages and enable them at runtime
    EDD    BIOS Enhanced Disk Drive Services (EDD) is enabled
    EFI    EFI Partitioning (GPT) is enabled
    EIDE    EIDE/ATAPI support is enabled.
    FB    The frame buffer device is enabled.
    GCOV    GCOV profiling is enabled.
    HW    Appropriate hardware is enabled.
    IA-64    IA-64 architecture is enabled.
    IMA     Integrity measurement architecture is enabled.
    IOSCHED    More than one I/O scheduler is enabled.
    IP_PNP    IP DHCP, BOOTP, or RARP is enabled.
    IPV6    IPv6 support is enabled.
    ISAPNP    ISA PnP code is enabled.
    ISDN    Appropriate ISDN support is enabled.
    JOY    Appropriate joystick support is enabled.
    KGDB    Kernel debugger support is enabled.
    KVM    Kernel Virtual Machine support is enabled.
    LIBATA  Libata driver is enabled
    LP    Printer support is enabled.
    LOOP    Loopback device support is enabled.
    M68k    M68k architecture is enabled.
            These options have more detailed description inside of
    MCA    MCA bus support is enabled.
    MDA    MDA console support is enabled.
    MOUSE    Appropriate mouse support is enabled.
    MSI    Message Signaled Interrupts (PCI).
    MTD    MTD (Memory Technology Device) support is enabled.
    NET    Appropriate network support is enabled.
    NUMA    NUMA support is enabled.
    NFS    Appropriate NFS support is enabled.
    OSS    OSS sound support is enabled.
    PV_OPS    A paravirtualized kernel is enabled.
    PARIDE    The ParIDE (parallel port IDE) subsystem is enabled.
    PARISC    The PA-RISC architecture is enabled.
    PCI    PCI bus support is enabled.
    PCIE    PCI Express support is enabled.
    PCMCIA    The PCMCIA subsystem is enabled.
    PNP    Plug & Play support is enabled.
    PPC    PowerPC architecture is enabled.
    PPT    Parallel port support is enabled.
    PS2    Appropriate PS/2 support is enabled.
    RAM    RAM disk support is enabled.
    S390    S390 architecture is enabled.
    SCSI    Appropriate SCSI support is enabled.
            A lot of drivers have their options described inside
            the Documentation/scsi/ sub-directory.
    SECURITY Different security models are enabled.
    SELINUX SELinux support is enabled.
    APPARMOR AppArmor support is enabled.
    SERIAL    Serial support is enabled.
    SH    SuperH architecture is enabled.
    SMP    The kernel is an SMP kernel.
    SPARC    Sparc architecture is enabled.
    SWSUSP    Software suspend (hibernation) is enabled.
    SUSPEND    System suspend states are enabled.
    FTRACE    Function tracing enabled.
    TPM    TPM drivers are enabled.
    TS    Appropriate touchscreen support is enabled.
    UMS    USB Mass Storage support is enabled.
    USB    USB support is enabled.
    USBHID    USB Human Interface Device support is enabled.
    V4L    Video For Linux support is enabled.
    VGA    The VGA console has been enabled.
    VT    Virtual terminal support is enabled.
    WDT    Watchdog support is enabled.
    XT    IBM PC/XT MFM hard disk support is enabled.
    X86-32    X86-32, aka i386 architecture is enabled.
    X86-64    X86-64 architecture is enabled.
            More X86-64 boot options can be found in
            Documentation/x86/x86_64/boot-options.txt .
    X86    Either 32bit or 64bit x86 (same as X86-32+X86-64)
    XEN    Xen support is enabled

In addition, the following text indicates that the option:

    BUGS=    Relates to possible processor bugs on the said processor.
    KNL    Is a kernel start-up parameter.
    BOOT    Is a boot loader parameter.

Parameters denoted with BOOT are actually interpreted by the boot
loader, and have no meaning to the kernel directly.
Do not modify the syntax of boot loader parameters without extreme
need or coordination with <Documentation/x86/boot.txt>.

There are also arch-specific kernel-parameters not documented here.
See for example <Documentation/x86/x86_64/boot-options.txt>.

Note that ALL kernel parameters listed below are CASE SENSITIVE, and that
a trailing = on the name of any parameter states that that parameter will
be entered as an environment variable, whereas its absence indicates that
it will appear as a kernel argument readable via /proc/cmdline by programs
running once the system is up.

The number of kernel parameters is not limited, but the length of the
complete command line (parameters including spaces etc.) is limited to
a fixed number of characters. This limit depends on the architecture
and is between 256 and 4096 characters. It is defined in the file
./include/asm/setup.h as COMMAND_LINE_SIZE.

Finally, the [KMG] suffix is commonly described after a number of kernel
parameter values. These 'K', 'M', and 'G' letters represent the _binary_
multipliers 'Kilo', 'Mega', and 'Giga', equalling 2^10, 2^20, and 2^30
bytes respectively. Such letter suffixes can also be entirely omitted.

    acpi=        [HW,ACPI,X86]
            Advanced Configuration and Power Interface
            Format: { force | off | strict | noirq | rsdt }
            force -- enable ACPI if default was off
            off -- disable ACPI if default was on
            noirq -- do not use ACPI for IRQ routing
            strict -- Be less tolerant of platforms that are not
                strictly ACPI specification compliant.
            rsdt -- prefer RSDT over (default) XSDT
            copy_dsdt -- copy DSDT to memory

            See also Documentation/power/pm.txt, pci=noacpi

    acpi_apic_instance=    [ACPI, IOAPIC]
            Format: <int>
            2: use 2nd APIC table, if available
            1,0: use 1st APIC table
            default: 0

    acpi_backlight=    [HW,ACPI]
            If set to vendor, prefer vendor specific driver
            (e.g. thinkpad_acpi, sony_acpi, etc.) instead
            of the ACPI video.ko driver.

    acpi.debug_layer=    [HW,ACPI,ACPI_DEBUG]
    acpi.debug_level=    [HW,ACPI,ACPI_DEBUG]
            Format: <int>
            CONFIG_ACPI_DEBUG must be enabled to produce any ACPI
            debug output.  Bits in debug_layer correspond to a
            _COMPONENT in an ACPI source file, e.g.,
                #define _COMPONENT ACPI_PCI_COMPONENT
            Bits in debug_level correspond to a level in
            ACPI_DEBUG_PRINT statements, e.g.,
                ACPI_DEBUG_PRINT((ACPI_DB_INFO, ...
            The debug_level mask defaults to "info".  See
            Documentation/acpi/debug.txt for more information about
            debug layers and levels.

            Enable processor driver info messages:
            Enable PCI/PCI interrupt routing info messages:
            Enable AML "Debug" output, i.e., stores to the Debug
            object while interpreting AML:
                acpi.debug_layer=0xffffffff acpi.debug_level=0x2
            Enable all messages related to ACPI hardware:
                acpi.debug_layer=0x2 acpi.debug_level=0xffffffff

            Some values produce so much output that the system is
            unusable.  The "log_buf_len" parameter may be useful
            if you need to capture more output.

    acpi_irq_balance [HW,ACPI]
            ACPI will balance active IRQs
            default in APIC mode

    acpi_irq_nobalance [HW,ACPI]
            ACPI will not move active IRQs (default)
            default in PIC mode

    acpi_irq_isa=    [HW,ACPI] If irq_balance, mark listed IRQs used by ISA
            Format: <irq>,<irq>...

    acpi_irq_pci=    [HW,ACPI] If irq_balance, clear listed IRQs for
            use by PCI
            Format: <irq>,<irq>...

    acpi_no_auto_ssdt    [HW,ACPI] Disable automatic loading of SSDT

    acpi_os_name=    [HW,ACPI] Tell ACPI BIOS the name of the OS
            Format: To spoof as Windows 98: ="Microsoft Windows"

    acpi_osi=    [HW,ACPI] Modify list of supported OS interface strings
            acpi_osi="string1"    # add string1 -- only one string
            acpi_osi="!string2"    # remove built-in string2
            acpi_osi=        # disable all strings

    acpi_pm_good    [X86]
            Override the pmtimer bug detection: force the kernel
            to assume that this machine's pmtimer latches its value
            and always returns good values.

    acpi_sci=    [HW,ACPI] ACPI System Control Interrupt trigger mode
            Format: { level | edge | high | low }

    acpi_serialize    [HW,ACPI] force serialization of AML methods

    acpi_skip_timer_override [HW,ACPI]
            Recognize and ignore IRQ0/pin2 Interrupt Override.
            For broken nForce2 BIOS resulting in XT-PIC timer.

    acpi_sleep=    [HW,ACPI] Sleep options
            Format: { s3_bios, s3_mode, s3_beep, s4_nohwsig,
                  old_ordering, nonvs, sci_force_enable }
            See Documentation/power/video.txt for information on
            s3_bios and s3_mode.
            s3_beep is for debugging; it makes the PC's speaker beep
            as soon as the kernel's real-mode entry point is called.
            s4_nohwsig prevents ACPI hardware signature from being
            used during resume from hibernation.
            old_ordering causes the ACPI 1.0 ordering of the _PTS
            control method, with respect to putting devices into
            low power states, to be enforced (the ACPI 2.0 ordering
            of _PTS is used by default).
            nonvs prevents the kernel from saving/restoring the
            ACPI NVS memory during suspend/hibernation and resume.
            sci_force_enable causes the kernel to set SCI_EN directly
            on resume from S1/S3 (which is against the ACPI spec,
            but some broken systems don't work without it).

    acpi_use_timer_override [HW,ACPI]
            Use timer override. For some broken Nvidia NF5 boards
            that require a timer override, but don't have HPET

    acpi_enforce_resources=    [ACPI]
            { strict | lax | no }
            Check for resource conflicts between native drivers
            and ACPI OperationRegions (SystemIO and SystemMemory
            only). IO ports and memory declared in ACPI might be
            used by the ACPI subsystem in arbitrary AML code and
            can interfere with legacy drivers.
            strict (default): access to resources claimed by ACPI
            is denied; legacy drivers trying to access reserved
            resources will fail to bind to device using them.
            lax: access to resources claimed by ACPI is allowed;
            legacy drivers trying to access reserved resources
            will bind successfully but a warning message is logged.
            no: ACPI OperationRegions are not marked as reserved,
            no further checks are performed.

    add_efi_memmap    [EFI; X86] Include EFI memory map in
            kernel's map of available physical RAM.

    agp=        [AGP]
            { off | try_unsupported }
            off: disable AGP support
            try_unsupported: try to drive unsupported chipsets
                (may crash computer or cause data corruption)

    ALSA        [HW,ALSA]
            See Documentation/sound/alsa/alsa-parameters.txt

    alignment=    [KNL,ARM]
            Allow the default userspace alignment fault handler
            behaviour to be specified.  Bit 0 enables warnings,
            bit 1 enables fixups, and bit 2 sends a segfault.

    amd_iommu=    [HW,X86-84]
            Pass parameters to the AMD IOMMU driver in the system.
            Possible values are:
            fullflush - enable flushing of IO/TLB entries when
                    they are unmapped. Otherwise they are
                    flushed before they will be reused, which
                    is a lot of faster
            off      - do not initialize any AMD IOMMU found in
                    the system    [HW,JOY] Amiga joystick support
            Map of devices attached to JOY0DAT and JOY1DAT
            Format: <a>,<b>
            See also Documentation/kernel/input/joystick.txt    [HW,JOY] Analog joystick and gamepad support
            Specifies type or capabilities of an analog joystick
            connected to one of 16 gameports
            Format: <type1>,<type2>,..<type16>

    apc=        [HW,SPARC]
            Power management functions (SPARCstation-4/5 + deriv.)
            Format: noidle
            Disable APC CPU standby support. SPARCstation-Fox does
            not play well with APC CPU idle - disable it if you have
            APC and your system crashes randomly.

    apic=        [APIC,X86-32] Advanced Programmable Interrupt Controller
            Change the output verbosity whilst booting
            Format: { quiet (default) | verbose | debug }
            Change the amount of debugging information output
            when initialising the APIC and IO-APIC components.

    autoconf=    [IPV6]
            See Documentation/networking/ipv6.txt.

    show_lapic=    [APIC,X86] Advanced Programmable Interrupt Controller
            Limit apic dumping. The parameter defines the maximal
            number of local apics being dumped. Also it is possible
            to set it to "all" by meaning -- no limit here.
            Format: { 1 (default) | 2 | ... | all }.
            The parameter valid if only apic=debug or
            apic=verbose is specified.
            Example: apic=debug show_lapic=all

    apm=        [APM] Advanced Power Management
            See header of arch/x86/kernel/apm_32.c.

    arcrimi=    [HW,NET] ARCnet - "RIM I" (entirely mem-mapped) cards
            Format: <io>,<irq>,<nodeID>

    ataflop=    [HW,M68k]

    atarimouse=    [HW,MOUSE] Atari Mouse

    atkbd.extra=    [HW] Enable extra LEDs and keys on IBM RapidAccess,
            EzKey and similar keyboards

    atkbd.reset=    [HW] Reset keyboard during initialization

    atkbd.set=    [HW] Select keyboard code set
            Format: <int> (2 = AT (default), 3 = PS/2)

    atkbd.scroll=    [HW] Enable scroll wheel on MS Office and similar

    atkbd.softraw=    [HW] Choose between synthetic and real raw mode
            Format: <bool> (0 = real, 1 = synthetic (default))

    atkbd.softrepeat= [HW]
            Use software keyboard repeat

    autotest    [IA64]

    baycom_epp=    [HW,AX25]
            Format: <io>,<mode>

    baycom_par=    [HW,AX25] BayCom Parallel Port AX.25 Modem
            Format: <io>,<mode>
            See header of drivers/net/hamradio/baycom_par.c.

    baycom_ser_fdx=    [HW,AX25]
            BayCom Serial Port AX.25 Modem (Full Duplex Mode)
            Format: <io>,<irq>,<mode>[,<baud>]
            See header of drivers/net/hamradio/baycom_ser_fdx.c.

    baycom_ser_hdx=    [HW,AX25]
            BayCom Serial Port AX.25 Modem (Half Duplex Mode)
            Format: <io>,<irq>,<mode>
            See header of drivers/net/hamradio/baycom_ser_hdx.c.

    boot_delay=    Milliseconds to delay each printk during boot.
            Values larger than 10 seconds (10000) are changed to
            no delay (0).
            Format: integer

    bootmem_debug    [KNL] Enable bootmem allocator debug messages.

    bttv.card=    [HW,V4L] bttv (bt848 + bt878 based grabber cards)    Most important insmod options are available as
            kernel args too.
    bttv.pll=    See Documentation/video4linux/bttv/Insmod-options
    bttv.tuner=    and Documentation/video4linux/bttv/CARDLIST

    bulk_remove=off    [PPC]  This parameter disables the use of the pSeries
            firmware feature for flushing multiple hpte entries
            at a time.

    c101=        [NET] Moxa C101 synchronous serial card

    cachesize=    [BUGS=X86-32] Override level 2 CPU cache size detection.
            Sometimes CPU hardware bugs make them report the cache
            size incorrectly. The kernel will attempt work arounds
            to fix known problems, but for some CPUs it is not
            possible to determine what the correct size should be.
            This option provides an override for these situations.

            [SECURITY] Disable capabilities.  This would normally
            be used only if an alternative security model is to be
            configured.  Potentially dangerous and should only be
            used if you are entirely sure of the consequences.

    ccw_timeout_log [S390]
            See Documentation/s390/CommonIO for details.

    cgroup_disable= [KNL] Disable a particular controller
            Format: {name of the controller(s) to disable}
                {Currently supported controllers - "memory"}

    checkreqprot    [SELINUX] Set initial checkreqprot flag value.
            Format: { "0" | "1" }
            See security/selinux/Kconfig help text.
            0 -- check protection applied by kernel (includes
                any implied execute protection).
            1 -- check protection requested by application.
            Default value is set via a kernel config option.
            Value can be changed at runtime via

    cio_ignore=    [S390]
            See Documentation/s390/CommonIO for details.

    clock=        [BUGS=X86-32, HW] gettimeofday clocksource override.
            Forces specified clocksource (if available) to be used
            when calculating gettimeofday(). If specified
            clocksource is not available, it defaults to PIT.
            Format: { pit | tsc | cyclone | pmtmr }

    clocksource=    Override the default clocksource
            Format: <string>
            Override the default clocksource and use the clocksource
            with the name specified.
            Some clocksource names to choose from, depending on
            the platform:
            [all] jiffies (this is the base, fallback clocksource)
            [ACPI] acpi_pm
            [ARM] imx_timer1,OSTS,netx_timer,mpu_timer2,
            [AVR32] avr32
            [X86-32] pit,hpet,tsc;
                scx200_hrt on Geode; cyclone on IBM x440
            [MIPS] MIPS
            [PARISC] cr16
            [S390] tod
            [SH] SuperH
            [SPARC64] tick
            [X86-64] hpet,tsc

    clearcpuid=BITNUM [X86]
            Disable CPUID feature X for the kernel. See
            arch/x86/include/asm/cpufeature.h for the valid bit
            numbers. Note the Linux specific bits are not necessarily
            stable over kernel options, but the vendor specific
            ones should be.
            Also note that user programs calling CPUID directly
            or using the feature without checking anything
            will still see it. This just prevents it from
            being used by the kernel or shown in /proc/cpuinfo.
            Also note the kernel might malfunction if you disable
            some critical bits.

    cmo_free_hint=    [PPC] Format: { yes | no }
            Specify whether pages are marked as being inactive
            when they are freed.  This is used in CMO environments
            to determine OS memory pressure for page stealing by
            a hypervisor.
            Default: yes

    code_bytes    [X86] How many bytes of object code to print
            in an oops report.
            Range: 0 - 8192
            Default: 64

    com20020=    [HW,NET] ARCnet - COM20020 chipset

    com90io=    [HW,NET] ARCnet - COM90xx chipset (IO-mapped buffers)
            Format: <io>[,<irq>]

    com90xx=    [HW,NET]
            ARCnet - COM90xx chipset (memory-mapped buffers)
            Format: <io>[,<irq>[,<memstart>]]

    condev=        [HW,S390] console device

    console=    [KNL] Output console device and options.

        tty<n>    Use the virtual console device <n>.

            Use the specified serial port.  The options are of
            the form "bbbbpnf", where "bbbb" is the baud rate,
            "p" is parity ("n", "o", or "e"), "n" is number of
            bits, and "f" is flow control ("r" for RTS or
            omit it).  Default is "9600n8".

            See Documentation/serial-console.txt for more
            information.  See
            Documentation/networking/netconsole.txt for an

            Start an early, polled-mode console on the 8250/16550
            UART at the specified I/O port or MMIO address,
            switching to the matching ttyS device later.  The
            options are the same as for ttyS, above.

                If the device connected to the port is not a TTY but a braille
                device, prepend "brl," before the device type, for instance
        For now, only VisioBraille is supported.

    consoleblank=    [KNL] The console blank (screen saver) timeout in
            seconds. Defaults to 10*60 = 10mins. A value of 0
            disables the blank timer.

            [KNL] Change the default value for
            See also Documentation/filesystems/proc.txt.

    cpcihp_generic=    [HW,PCI] Generic port I/O CompactPCI driver

            [KNL] Using kexec, Linux can switch to a 'crash kernel'
            upon panic. This parameter reserves the physical
            memory region [offset, offset + size] for that kernel
            image. If '@offset' is omitted, then a suitable offset
            is selected automatically. Check
            Documentation/kdump/kdump.txt for further details.

            [KNL] Same as above, but depends on the memory
            in the running system. The syntax of range is
            start-[end] where start and end are both
            a memory unit (amount[KMG]). See also
            Documentation/kdump/kdump.txt for an example.

    cs89x0_dma=    [HW,NET]
            Format: <dma>

    cs89x0_media=    [HW,NET]
            Format: { rj45 | aui | bnc }

    dasd=        [HW,NET]
            See header of drivers/s390/block/dasd_devmap.c.[2|3]=    [HW,JOY] Multisystem joystick support via parallel port
            (one device per port)
            Format: <port#>,<type>
            See also Documentation/input/joystick-parport.txt

    ddebug_query=   [KNL,DYNAMIC_DEBUG] Enable debug messages at early boot
            time. See Documentation/dynamic-debug-howto.txt for

    debug        [KNL] Enable kernel debugging (events log level).

            [KNL] verbose self-tests
            Print debugging info while doing the locking API
            We default to 0 (no extra messages), setting it to
            1 will print _a lot_ more information - normally
            only useful to kernel developers.

    debug_objects    [KNL] Enable object debugging

            [KNL] Disable object debugging

    debugpat    [X86] Enable PAT debugging

    decnet.addr=    [HW,NET]
            Format: <area>[,<node>]
            See also Documentation/networking/decnet.txt.

            [same as hugepagesz=] The size of the default
            HugeTLB page size. This is the size represented by
            the legacy /proc/ hugepages APIs, used for SHM, and
            default size when mounting hugetlbfs filesystems.
            Defaults to the default architecture's huge page size
            if not specified.

    dhash_entries=    [KNL]
            Set number of hash buckets for dentry cache.

    digi=        [HW,SERIAL]
            IO parameters + enable/disable command.

    digiepca=    [HW,SERIAL]
            See drivers/char/README.epca and

    disable=    [IPV6]
            See Documentation/networking/ipv6.txt.

    disable_ddw     [PPC/PSERIES]
            Disable Dynamic DMA Window support. Use this if
            to workaround buggy firmware.

    disable_ipv6=    [IPV6]
            See Documentation/networking/ipv6.txt.

    disable_mtrr_cleanup [X86]
            The kernel tries to adjust MTRR layout from continuous
            to discrete, to make X server driver able to add WB
            entry later. This parameter disables that.

    disable_mtrr_trim [X86, Intel and AMD only]
            By default the kernel will trim any uncacheable
            memory out of your available memory pool based on
            MTRR settings.  This parameter disables that behavior,
            possibly causing your machine to run very slowly.

    disable_timer_pin_1 [X86]
            Disable PIN 1 of APIC timer
            Can be useful to work around chipset bugs.

    dma_debug=off    If the kernel is compiled with DMA_API_DEBUG support,
            this option disables the debugging code at boot.

            This option allows to tune the number of preallocated
            entries for DMA-API debugging code. One entry is
            required per DMA-API allocation. Use this if the
            DMA-API debugging code disables itself because the
            architectural default is too low.

            With this option the DMA-API debugging driver
            filter feature can be enabled at boot time. Just
            pass the driver to filter for as the parameter.
            The filter can be disabled or changed to another
            driver later using sysfs.

    dscc4.setup=    [NET]

    earlycon=    [KNL] Output early console device and options.
            Start an early, polled-mode console on the 8250/16550
            UART at the specified I/O port or MMIO address.
            MMIO inter-register address stride is either 8bit (mmio)
                        or 32bit (mmio32).
            The options are the same as for ttyS, above.

    earlyprintk=    [X86,SH,BLACKFIN]

            Append ",keep" to not disable it when the real console
            takes over.

            Only vga or serial or usb debug port at a time.

            Currently only ttyS0 and ttyS1 are supported.

            Interaction with the standard serial driver is not
            very good.

            The VGA output is eventually overwritten by the real

    ekgdboc=    [X86,KGDB] Allow early kernel console debugging

            This is designed to be used in conjunction with
            the boot argument: earlyprintk=vga

    edd=        [EDD]
            Format: {"off" | "on" | "skip[mbr]"}

    eisa_irq_edge=    [PARISC,HW]
            See header of drivers/parisc/eisa.c.

    elanfreq=    [X86-32]
            See comment before function elanfreq_setup() in

    elevator=    [IOSCHED]
            Format: {"cfq" | "deadline" | "noop"}
            See Documentation/block/as-iosched.txt and
            Documentation/block/deadline-iosched.txt for details.

    elfcorehdr=    [IA64,PPC,SH,X86]
            Specifies physical address of start of kernel core
            image elf header. Generally kexec loader will
            pass this option to capture kernel.
            See Documentation/kdump/kdump.txt for details.

    enable_mtrr_cleanup [X86]
            The kernel tries to adjust MTRR layout from continuous
            to discrete, to make X server driver able to add WB
            entry later. This parameter enables that.

    enable_timer_pin_1 [X86]
            Enable PIN 1 of APIC timer
            Can be useful to work around chipset bugs
            (in particular on some ATI chipsets).
            The kernel tries to set a reasonable default.

    enforcing    [SELINUX] Set initial enforcing status.
            Format: {"0" | "1"}
            See security/selinux/Kconfig help text.
            0 -- permissive (log only, no denials).
            1 -- enforcing (deny and log).
            Default value is 0.
            Value can be changed at runtime via /selinux/enforce.

    erst_disable    [ACPI]
            Disable Error Record Serialization Table (ERST)

    ether=        [HW,NET] Ethernet cards parameters
            This option is obsoleted by the "netdev=" option, which
            has equivalent usage. See its documentation for details.

            General fault injection mechanism.
            Format: <interval>,<probability>,<space>,<times>
            See also /Documentation/fault-injection/.

    floppy=        [HW]
            See Documentation/blockdev/floppy.txt.

            [IA-64] Avoid check_sal_cache_flush which may hang on
            buggy SAL_CACHE_FLUSH implementations. Using this
            parameter will force ia64_sal_cache_flush to call
            ia64_pal_cache_flush instead of SAL_CACHE_FLUSH.

            [FTRACE] will set and start the specified tracer
            as early as possible in order to facilitate early
            boot debugging.

            [FTRACE] will dump the trace buffers on oops.
            If no parameter is passed, ftrace will dump
            buffers of all CPUs, but if you pass orig_cpu, it will
            dump only the buffer of the CPU that triggered the

            [FTRACE] Limit the functions traced by the function
            tracer at boot up. function-list is a comma separated
            list of functions. This list can be changed at run
            time by the set_ftrace_filter file in the debugfs
            tracing directory. 

            [FTRACE] Do not trace the functions specified in
            function-list. This list can be changed at run time
            by the set_ftrace_notrace file in the debugfs
            tracing directory.

            [FTRACE] Limit the top level callers functions traced
            by the function graph tracer at boot up.
            function-list is a comma separated list of functions
            that can be changed at run time by the
            set_graph_function file in the debugfs tracing directory.[2|3]=
            [HW,JOY] Multisystem joystick and NES/SNES/PSX pad
            support via parallel port (up to 5 devices per port)
            Format: <port#>,<pad1>,<pad2>,<pad3>,<pad4>,<pad5>
            See also Documentation/input/joystick-parport.txt

    gamma=        [HW,DRM]

    gart_fix_e820=  [X86_64] disable the fix e820 for K8 GART
            Format: off | on
            default: on

    gcov_persist=    [GCOV] When non-zero (default), profiling data for
            kernel modules is saved and remains accessible via
            debugfs, even when the module is unloaded/reloaded.
            When zero, profiling data is discarded and associated
            debugfs files are removed at module unload time.

    gpt        [EFI] Forces disk with valid GPT signature but
            invalid Protective MBR to be treated as GPT.

    hashdist=    [KNL,NUMA] Large hashes allocated during boot
            are distributed across NUMA nodes.  Defaults on
            for 64bit NUMA, off otherwise.
            Format: 0 | 1 (for off | on)

    hcl=        [IA-64] SGI's Hardware Graph compatibility layer

    hd=        [EIDE] (E)IDE hard drive subsystem geometry
            Format: <cyl>,<head>,<sect>

    hest_disable    [ACPI]
            Disable Hardware Error Source Table (HEST) support;
            corresponding firmware-first mode error processing
            logic will be disabled.

    highmem=nn[KMG]    [KNL,BOOT] forces the highmem zone to have an exact
            size of <nn>. This works even on boxes that have no
            highmem otherwise. This also works to reduce highmem
            size on bigger boxes.

    highres=    [KNL] Enable/disable high resolution timer mode.
            Valid parameters: "on", "off"
            Default: "on"

    hisax=        [HW,ISDN]
            See Documentation/isdn/README.HiSax.

    hlt        [BUGS=ARM,SH]

    hpet=        [X86-32,HPET] option to control HPET usage
            Format: { enable (default) | disable | force |
                verbose }
            disable: disable HPET and use PIT instead
            force: allow force enabled of undocumented chips (ICH4,
                VIA, nVidia)
            verbose: show contents of HPET registers during setup

    hugepages=    [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
    hugepagesz=    [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
            On x86-64 and powerpc, this option can be specified
            multiple times interleaved with hugepages= to reserve
            huge pages of different sizes. Valid pages sizes on
            x86-64 are 2M (when the CPU supports "pse") and 1G
            (when the CPU supports the "pdpe1gb" cpuinfo flag)
            Note that 1GB pages can only be allocated at boot time
            using hugepages= and not freed afterwards.

    hvc_iucv=    [S390] Number of z/VM IUCV hypervisor console (HVC)
                   terminal devices. Valid values: 0..8
    hvc_iucv_allow=    [S390] Comma-separated list of z/VM user IDs.
                   If specified, z/VM IUCV HVC accepts connections
                   from listed z/VM user IDs only.

    keep_bootcon    [KNL]
            Do not unregister boot console at start. This is only
            useful for debugging when something happens in the window
            between unregistering the boot console and initializing
            the real console.

    i2c_bus=    [HW] Override the default board specific I2C bus speed
                 or register an additional I2C bus that is not
                 registered from board initialization code.

    i8042.debug    [HW] Toggle i8042 debug mode    [HW] Put keyboard port into non-translated mode
    i8042.dumbkbd    [HW] Pretend that controller can only read data from
                 keyboard and cannot control its state
                 (Don't attempt to blink the leds)
    i8042.noaux    [HW] Don't check for auxiliary (== mouse) port
    i8042.nokbd    [HW] Don't check/create keyboard port
    i8042.noloop    [HW] Disable the AUX Loopback command while probing
                 for the AUX port
    i8042.nomux    [HW] Don't check presence of an active multiplexing
    i8042.nopnp    [HW] Don't use ACPIPnP / PnPBIOS to discover KBD/AUX
    i8042.notimeout    [HW] Ignore timeout condition signalled by conroller
    i8042.reset    [HW] Reset the controller during init and cleanup
    i8042.unlock    [HW] Unlock (ignore) the keylock

    i810=        [HW,DRM]

    i8k.ignore_dmi    [HW] Continue probing hardware even if DMI data
            indicates that the driver is running on unsupported
    i8k.force    [HW] Activate i8k driver even if SMM BIOS signature
            does not match list of supported models.
            [HW] Report power status in /proc/i8k
            (disabled by default)
    i8k.restricted    [HW] Allow controlling fans only if SYS_ADMIN
            capability is set.

    icn=        [HW,ISDN]
            Format: <io>[,<membase>[,<icn_id>[,<icn_id2>]]]

    ide-core.nodma=    [HW] (E)IDE subsystem
            Format: =0.0 to prevent dma on hda, =0.1 hdb =1.0 hdc
            .vlb_clock .pci_clock .noflush .nohpa .noprobe .nowerr
            .cdrom .chs .ignore_cable are additional options
            See Documentation/ide/ide.txt.

    ide-pci-generic.all-generic-ide [HW] (E)IDE subsystem
            Claim all unknown PCI IDE storage controllers.

    idle=        [X86]
            Format: idle=poll, idle=mwait, idle=halt, idle=nomwait
            Poll forces a polling idle loop that can slightly
            improve the performance of waking up a idle CPU, but
            will use a lot of power and make the system run hot.
            Not recommended.
            idle=mwait: On systems which support MONITOR/MWAIT but
            the kernel chose to not use it because it doesn't save
            as much power as a normal idle loop, use the
            MONITOR/MWAIT idle loop anyways. Performance should be
            the same as idle=poll.
            idle=halt: Halt is forced to be used for CPU idle.
            In such case C2/C3 won't be used again.
            idle=nomwait: Disable mwait for CPU C-states

    ignore_loglevel    [KNL]
            Ignore loglevel setting - this will print /all/
            kernel messages to the console. Useful for debugging.

    ihash_entries=    [KNL]
            Set number of hash buckets for inode cache.

    ima_audit=    [IMA]
            Format: { "0" | "1" }
            0 -- integrity auditing messages. (Default)
            1 -- enable informational integrity auditing messages.

    ima_hash=    [IMA]
            Format: { "sha1" | "md5" }
            default: "sha1"

    ima_tcb        [IMA]
            Load a policy which meets the needs of the Trusted
            Computing Base.  This means IMA will measure all
            programs exec'd, files mmap'd for exec, and all files
            opened for read by uid=0.

    init=        [KNL]
            Format: <full_path>
            Run specified binary instead of /sbin/init as init

    initcall_debug    [KNL] Trace initcalls as they are executed.  Useful
            for working out where the kernel is dying during

    initrd=        [BOOT] Specify the location of the initial ramdisk

    inport.irq=    [HW] Inport (ATI XL and Microsoft) busmouse driver
            Format: <irq>

    intel_iommu=    [DMAR] Intel IOMMU driver (DMAR) option
            Enable intel iommu driver.
            Disable intel iommu driver.
        igfx_off [Default Off]
            By default, gfx is mapped as normal device. If a gfx
            device has a dedicated DMAR unit, the DMAR unit is
            bypassed by not enabling DMAR with this option. In
            this case, gfx device will use physical address for
        forcedac [x86_64]
            With this option iommu will not optimize to look
            for io virtual address below 32 bit forcing dual
            address cycle on pci bus for cards supporting greater
            than 32 bit addressing. The default is to look
            for translation below 32 bit and if not available
            then look in the higher range.
        strict [Default Off]
            With this option on every unmap_single operation will
            result in a hardware IOTLB flush operation as opposed
            to batching them for performance.
        sp_off [Default Off]
            By default, super page will be supported if Intel IOMMU
            has the capability. With this option, super page will
            not be supported.
    intremap=    [X86-64, Intel-IOMMU]
            Format: { on (default) | off | nosid }
            on    enable Interrupt Remapping (default)
            off    disable Interrupt Remapping
            nosid    disable Source ID checking

    inttest=    [IA64]

    iomem=        Disable strict checking of access to MMIO memory
        strict    regions from userspace.

    iommu=        [x86]
        pt    [x86, IA64]

    io7=        [HW] IO7 for Marvel based alpha systems
            See comment before marvel_specify_io7 in

    io_delay=    [X86] I/O delay method
            Standard port 0x80 based delay
            Alternate port 0xed based delay (needed on some systems)
            Simple two microseconds delay
            No delay

    ip=        [IP_PNP]
            See Documentation/filesystems/nfs/nfsroot.txt.

    ip2=        [HW] Set IO/IRQ pairs for up to 4 IntelliPort boards
            See comment before ip2_setup() in

    irqfixup    [HW]
            When an interrupt is not handled search all handlers
            for it. Intended to get systems with badly broken
            firmware running.

    irqpoll        [HW]
            When an interrupt is not handled search all handlers
            for it. Also check all handlers each timer
            interrupt. Intended to get systems with badly broken
            firmware running.

    isapnp=        [ISAPNP]
            Format: <RDP>,<reset>,<pci_scan>,<verbosity>

    isolcpus=    [KNL,SMP] Isolate CPUs from the general scheduler.
            <cpu number>,...,<cpu number>
            <cpu number>-<cpu number>
            (must be a positive range in ascending order)
            or a mixture
            <cpu number>,...,<cpu number>-<cpu number>

            This option can be used to specify one or more CPUs
            to isolate from the general SMP balancing and scheduling
            algorithms. You can move a process onto or off an
            "isolated" CPU via the CPU affinity syscalls or cpuset.
            <cpu number> begins at 0 and the maximum value is
            "number of CPUs in system - 1".

            This option is the preferred way to isolate CPUs. The
            alternative -- manually setting the CPU mask of all
            tasks in the system -- can cause problems and
            suboptimal load balancer performance.

    iucv=        [HW,NET]

    js=        [HW,JOY] Analog joystick
            See Documentation/input/joystick.txt.

    keepinitrd    [HW,ARM]

    kernelcore=nn[KMG]    [KNL,X86,IA-64,PPC] This parameter
            specifies the amount of memory usable by the kernel
            for non-movable allocations.  The requested amount is
            spread evenly throughout all nodes in the system. The
            remaining memory in each node is used for Movable
            pages. In the event, a node is too small to have both
            kernelcore and Movable pages, kernelcore pages will
            take priority and other nodes will have a larger number
            of kernelcore pages.  The Movable zone is used for the
            allocation of pages that may be reclaimed or moved
            by the page migration subsystem.  This means that
            HugeTLB pages may not be allocated from this zone.
            Note that allocations like PTEs-from-HighMem still
            use the HighMem zone if it exists, and the Normal
            zone if it does not.

    kgdbdbgp=    [KGDB,HW] kgdb over EHCI usb debug port.
            Format: <Controller#>[,poll interval]
            The controller # is the number of the ehci usb debug
            port as it is probed via PCI.  The poll interval is
            optional and is the number seconds in between
            each poll cycle to the debug port in case you need
            the functionality for interrupting the kernel with
            gdb or control-c on the dbgp connection.  When
            not using this parameter you use sysrq-g to break into
            the kernel debugger.

    kgdboc=        [KGDB,HW] kgdb over consoles.
            Requires a tty driver that supports console polling,
            or a supported polling keyboard driver (non-usb).
             Serial only format: <serial_device>[,baud]
             keyboard only format: kbd
             keyboard and serial format: kbd,<serial_device>[,baud]
            Optional Kernel mode setting:
             kms, kbd format: kms,kbd
             kms, kbd and serial format: kms,kbd,<ser_dev>[,baud]

    kgdbwait    [KGDB] Stop kernel execution and enter the
            kernel debugger at the earliest opportunity.

    kmac=        [MIPS] korina ethernet MAC address.
            Configure the RouterBoard 532 series on-chip
            Ethernet adapter MAC address.

    kmemleak=    [KNL] Boot-time kmemleak enable/disable
            Valid arguments: on, off
            Default: on

    kstack=N    [X86] Print N words from the kernel stack
            in oops dumps.

    kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
            Default is 0 (don't ignore, but inject #GP)

    kvm.oos_shadow=    [KVM] Disable out-of-sync shadow paging.
            Default is 1 (enabled)

    kvm.mmu_audit=    [KVM] This is a R/W parameter which allows audit
            KVM MMU at runtime.
            Default is 0 (off)

    kvm-amd.nested=    [KVM,AMD] Allow nested virtualization in KVM/SVM.
            Default is 1 (enabled)

    kvm-amd.npt=    [KVM,AMD] Disable nested paging (virtualized MMU)
            for all guests.
            Default is 1 (enabled) if in 64bit or 32bit-PAE mode

            [KVM,Intel] Disables bypassing of guest page faults
            on Intel chips. Default is 1 (enabled)

    kvm-intel.ept=    [KVM,Intel] Disable extended page tables
            (virtualized MMU) support on capable Intel chips.
            Default is 1 (enabled)

            [KVM,Intel] Enable emulation of invalid guest states
            Default is 0 (disabled)

            [KVM,Intel] Disable FlexPriority feature (TPR shadow).
            Default is 1 (enabled)

            [KVM,Intel] Disable unrestricted guest feature
            (virtualized real and unpaged mode) on capable
            Intel chips. Default is 1 (enabled)

    kvm-intel.vpid=    [KVM,Intel] Disable Virtual Processor Identification
            feature (tagged TLBs) on capable Intel chips.
            Default is 1 (enabled)

    l2cr=        [PPC]

    l3cr=        [PPC]

    lapic        [X86-32,APIC] Enable the local APIC even if BIOS
            disabled it.

    lapic_timer_c2_ok    [X86,APIC] trust the local apic timer
            in C2 power state.

    libata.dma=    [LIBATA] DMA control
            libata.dma=0      Disable all PATA and SATA DMA
            libata.dma=1      PATA and SATA Disk DMA only
            libata.dma=2      ATAPI (CDROM) DMA only
            libata.dma=4      Compact Flash DMA only 
            Combinations also work, so libata.dma=3 enables DMA
            for disks and CDROMs, but not CFs.
    libata.ignore_hpa=    [LIBATA] Ignore HPA limit
            libata.ignore_hpa=0      keep BIOS limits (default)
            libata.ignore_hpa=1      ignore limits, using full disk

    libata.noacpi    [LIBATA] Disables use of ACPI in libata suspend/resume
            when set.
            Format: <int>

    libata.force=    [LIBATA] Force configurations.  The format is comma
            separated list of "[ID:]VAL" where ID is
            PORT[.DEVICE].  PORT and DEVICE are decimal numbers
            matching port, link or device.  Basically, it matches
            the ATA ID string printed on console by libata.  If
            the whole ID part is omitted, the last PORT and DEVICE
            values are used.  If ID hasn't been specified yet, the
            configuration applies to all ports, links and devices.

            If only DEVICE is omitted, the parameter applies to
            the port and all links and devices behind it.  DEVICE
            number of 0 either selects the first device or the
            first fan-out link behind PMP device.  It does not
            select the host link.  DEVICE number of 15 selects the
            host link and device attached to it.

            The VAL specifies the configuration to force.  As long
            as there's no ambiguity shortcut notation is allowed.
            For example, both 1.5 and 1.5G would work for 1.5Gbps.
            The following configurations can be forced.

            * Cable type: 40c, 80c, short40c, unk, ign or sata.
              Any ID with matching PORT is used.

            * SATA link speed limit: 1.5Gbps or 3.0Gbps.

            * Transfer mode: pio[0-7], mwdma[0-4] and udma[0-7].
              udma[/][16,25,33,44,66,100,133] notation is also

            * [no]ncq: Turn on or off NCQ.

            * nohrst, nosrst, norst: suppress hard, soft
                          and both resets.

            * dump_id: dump IDENTIFY data.

            If there are multiple matching configurations changing
            the same attribute, the last one is used.

    memblock=debug    [KNL] Enable memblock debug messages.

    load_ramdisk=    [RAM] List of ramdisks to load from floppy
            See Documentation/blockdev/ramdisk.txt.

    lockd.nlm_grace_period=P  [NFS] Assign grace period.
            Format: <integer>

    lockd.nlm_tcpport=N    [NFS] Assign TCP port.
            Format: <integer>

    lockd.nlm_timeout=T    [NFS] Assign timeout value.
            Format: <integer>

    lockd.nlm_udpport=M    [NFS] Assign UDP port.
            Format: <integer>

    logibm.irq=    [HW,MOUSE] Logitech Bus Mouse Driver
            Format: <irq>

    loglevel=    All Kernel Messages with a loglevel smaller than the
            console loglevel will be printed to the console. It can
            also be changed with klogd or other programs. The
            loglevels are defined as follows:

            0 (KERN_EMERG)        system is unusable
            1 (KERN_ALERT)        action must be taken immediately
            2 (KERN_CRIT)        critical conditions
            3 (KERN_ERR)        error conditions
            4 (KERN_WARNING)    warning conditions
            5 (KERN_NOTICE)        normal but significant condition
            6 (KERN_INFO)        informational
            7 (KERN_DEBUG)        debug-level messages

    log_buf_len=n[KMG]    Sets the size of the printk ring buffer,
            in bytes.  n must be a power of two.  The default
            size is set in the kernel config file.

    logo.nologo    [FB] Disables display of the built-in Linux logo.
            This may be used to provide more screen space for
            kernel log messages and is useful when debugging
            kernel boot problems.

    lp=0        [LP]    Specify parallel ports to use, e.g,
    lp=port[,port...]    lp=none,parport0 (lp0 not configured, lp1 uses
    lp=reset        first parallel port). 'lp=0' disables the
    lp=auto            printer driver. 'lp=reset' (which can be
                specified in addition to the ports) causes
                attached printers to be reset. Using
                lp=port1,port2,... specifies the parallel ports
                to associate lp devices with, starting with
                lp0. A port specification may be 'none' to skip
                that lp device, or a parport name such as
                'parport0'. Specifying 'lp=auto' instead of a
                port specification list means that device IDs
                from each port should be examined, to see if
                an IEEE 1284-compliant printer is attached; if
                so, the driver will manage that printer.
                See also header of drivers/char/lp.c.

    lpj=n        [KNL]
            Sets loops_per_jiffy to given constant, thus avoiding
            time-consuming boot-time autodetection (up to 250 ms per
            CPU). 0 enables autodetection (default). To determine
            the correct value for your kernel, boot with normal
            autodetection and see what value is printed. Note that
            on SMP systems the preset will be applied to all CPUs,
            which is likely to cause problems if your CPUs need
            significantly divergent settings. An incorrect value
            will cause delays in the kernel to be wrong, leading to
            unpredictable I/O errors and other breakage. Although
            unlikely, in the extreme case this might damage your

    ltpc=        [NET]
            Format: <io>,<irq>,<dma>

    machvec=    [IA64] Force the use of a particular machine-vector
            (machvec) in a generic kernel.
            Example: machvec=hpzx1_swiotlb

    machtype=    [Loongson] Share the same kernel image file between different
             yeeloong laptop.
            Example: machtype=lemote-yeeloong-2f-7inch

    max_addr=nn[KMG]    [KNL,BOOT,ia64] All physical memory greater
            than or equal to this physical address is ignored.

    maxcpus=    [SMP] Maximum number of processors that    an SMP kernel
            should make use of.  maxcpus=n : n >= 0 limits the
            kernel to using 'n' processors.  n=0 is a special case,
            it is equivalent to "nosmp", which also disables
            the IO APIC.

    max_loop=    [LOOP] Maximum number of loopback devices that can
            be mounted
            Format: <1-256>

    mcatest=    [IA-64]

    mce        [X86-32] Machine Check Exception

    mce=option    [X86-64] See Documentation/x86/x86_64/boot-options.txt

    md=        [HW] RAID subsystems devices and level
            See Documentation/md.txt.

    mdacon=        [MDA]
            Format: <first>,<last>
            Specifies range of consoles to be captured by the MDA.

    mem=nn[KMG]    [KNL,BOOT] Force usage of a specific amount of memory
            Amount of memory to be used when the kernel is not able
            to see the whole system memory or for test.
            [X86-32] Use together with memmap= to avoid physical
            address space collisions. Without memmap= PCI devices
            could be placed at addresses belonging to unused RAM.

    mem=nopentium    [BUGS=X86-32] Disable usage of 4MB pages for kernel

            [KNL,SH] Allow user to override the default size for
            per-device physically contiguous DMA buffers.

    memmap=exactmap    [KNL,X86] Enable setting of an exact
            E820 memory map, as specified by the user.
            Such memmap=exactmap lines can be constructed based on
            BIOS output or other requirements. See the memmap=nn@ss
            option description.

            [KNL] Force usage of a specific region of memory
            Region of memory to be used, from ss to ss+nn.

            [KNL,ACPI] Mark specific memory as ACPI data.
            Region of memory to be used, from ss to ss+nn.

            [KNL,ACPI] Mark specific memory as reserved.
            Region of memory to be used, from ss to ss+nn.
            Example: Exclude memory from 0x18690000-0x1869ffff

    memory_corruption_check=0/1 [X86]
            Some BIOSes seem to corrupt the first 64k of
            memory when doing things like suspend/resume.
            Setting this option will scan the memory
            looking for corruption.  Enabling this will
            both detect corruption and prevent the kernel
            from using the memory being corrupted.
            However, its intended as a diagnostic tool; if
            repeatable BIOS-originated corruption always
            affects the same memory, you can use memmap=
            to prevent the kernel from using that memory.

    memory_corruption_check_size=size [X86]
            By default it checks for corruption in the low
            64k, making this memory unavailable for normal
            use.  Use this parameter to scan for
            corruption in more or less memory.

    memory_corruption_check_period=seconds [X86]
            By default it checks for corruption every 60
            seconds.  Use this parameter to check at some
            other rate.  0 disables periodic checking.

    memtest=    [KNL,X86] Enable memtest
            Format: <integer>
            default : 0 <disable>
            Specifies the number of memtest passes to be
            performed. Each pass selects another test
            pattern from a given set of patterns. Memtest
            fills the memory with this pattern, validates
            memory contents and reserves bad memory
            regions that are detected.

    meye.*=        [HW] Set MotionEye Camera parameters
            See Documentation/video4linux/meye.txt.

    mfgpt_irq=    [IA-32] Specify the IRQ to use for the
            Multi-Function General Purpose Timers on AMD Geode

    mfgptfix    [X86-32] Fix MFGPT timers on AMD Geode platforms when
            the BIOS has incorrectly applied a workaround. TinyBIOS
            version 0.98 is known to be affected, 0.99 fixes the
            problem by letting the user disable the workaround.

    mga=        [HW,DRM]

    min_addr=nn[KMG]    [KNL,BOOT,ia64] All physical memory below this
            physical address is ignored.

    mini2440=    [ARM,HW,KNL]
            Default: "0tb"
            MINI2440 configuration specification:
            0 - The attached screen is the 3.5" TFT
            1 - The attached screen is the 7" TFT
            2 - The VGA Shield is attached (1024x768)
            Leaving out the screen size parameter will not load
            the TFT driver, and the framebuffer will be left
            b - Enable backlight. The TFT backlight pin will be
            linked to the kernel VESA blanking code and a GPIO
            LED. This parameter is not necessary when using the
            VGA shield.
            c - Enable the s3c camera interface.
            t - Reserved for enabling touchscreen support. The
            touchscreen support is not enabled in the mainstream
            kernel as of 2.6.30, a preliminary port can be found
            in the "bleeding edge" mini2440 support kernel at

            [KNL] When CONFIG_DEBUG_MEMORY_INIT is set, this
            parameter allows control of the logging verbosity for
            the additional memory initialisation checks. A value
            of 0 disables mminit logging and a level of 4 will
            log everything. Information is printed at KERN_DEBUG
            so loglevel=8 may also need to be specified.

            [MOUSE] Maximum time between finger touching and
            leaving touchpad surface for touch to be considered
            a tap and be reported as a left button click (for
            touchpads working in absolute mode only).
            Format: <msecs>
    mousedev.xres=    [MOUSE] Horizontal screen resolution, used for devices
            reporting absolute coordinates, such as tablets
    mousedev.yres=    [MOUSE] Vertical screen resolution, used for devices
            reporting absolute coordinates, such as tablets

    movablecore=nn[KMG]    [KNL,X86,IA-64,PPC] This parameter
            is similar to kernelcore except it specifies the
            amount of memory used for migratable allocations.
            If both kernelcore and movablecore is specified,
            then kernelcore will be at *least* the specified
            value but may be more. If movablecore on its own
            is specified, the administrator must be careful
            that the amount of memory usable for all allocations
            is not too small.

    MTD_Partition=    [MTD]
            Format: <name>,<region-number>,<size>,<offset>

    MTD_Region=    [MTD] Format:

    mtdparts=    [MTD]
            See drivers/mtd/cmdlinepart.c.

    multitce=off    [PPC]  This parameter disables the use of the pSeries
            firmware feature for updating multiple TCE entries
            at a time.

    onenand.bdry=    [HW,MTD] Flex-OneNAND Boundary Configuration

            Format: [die0_boundary][,die0_lock][,die1_boundary][,die1_lock]

            boundary - index of last SLC block on Flex-OneNAND.
                   The remaining blocks are configured as MLC blocks.
            lock     - Configure if Flex-OneNAND boundary should be locked.
                   Once locked, the boundary cannot be changed.
                   1 indicates lock status, 0 indicates unlock status.

    mtdset=        [ARM]
            ARM/S3C2412 JIVE boot control

            See arch/arm/mach-s3c2412/mach-jive.c

            [HW] Make the MicroTouch USB driver use raw coordinates
            ('y', default) or cooked coordinates ('n')

    mtrr_chunk_size=nn[KMG] [X86]
            used for mtrr cleanup. It is largest continuous chunk
            that could hold holes aka. UC entries.

    mtrr_gran_size=nn[KMG] [X86]
            Used for mtrr cleanup. It is granularity of mtrr block.
            Default is 1.
            Large value could prevent small alignment from
            using up MTRRs.

    mtrr_spare_reg_nr=n [X86]
            Format: <integer>
            Range: 0,7 : spare reg number
            Default : 1
            Used for mtrr cleanup. It is spare mtrr entries number.
            Set to 2 or more if your graphical card needs more.

    n2=        [NET] SDL Inc. RISCom/N2 synchronous serial card

    netdev=        [NET] Network devices parameters
            Format: <irq>,<io>,<mem_start>,<mem_end>,<name>
            Note that mem_start is often overloaded to mean
            something different and driver-specific.
            This usage is only documented in each driver source
            file if at all.

            [NETFILTER] Enable connection tracking flow accounting
            0 to disable accounting
            1 to enable accounting
            Default value is 0.

    nfsaddrs=    [NFS] Deprecated.  Use ip= instead.
            See Documentation/filesystems/nfs/nfsroot.txt.

    nfsroot=    [NFS] nfs root filesystem for disk-less boxes.
            See Documentation/filesystems/nfs/nfsroot.txt.

    nfsrootdebug    [NFS] enable nfsroot debugging messages.
            See Documentation/filesystems/nfs/nfsroot.txt.

            [NFS] set the TCP port on which the NFSv4 callback
            channel should listen.

            [NFS] sets the pathname to the program which is used
            to update the NFS client cache entries.

            [NFS] sets the timeout after which an attempt to
            update a cache entry is deemed to have failed.

            [NFS] set the maximum lifetime for idmapper cache

            [NFS] enable 64-bit inode numbers.
            If zero, the NFS client will fake up a 32-bit inode
            number for the readdir() and stat() syscalls instead
            of returning the full 64-bit number.
            The default is to return 64-bit inode numbers.

            [NFSv4] When set, this option disables the NFSv4
            idmapper on the client, but only if the mount
            is using the 'sec=sys' security flavour. This may
            make migration from legacy NFSv2/v3 systems easier
            provided that the server has the appropriate support.
            The default is to always enable NFSv4 idmapping.

    nmi_debug=    [KNL,AVR32,SH] Specify one or more actions to take
            when a NMI is triggered.
            Format: [state][,regs][,debounce][,die]

    nmi_watchdog=    [KNL,BUGS=X86] Debugging features for SMP kernels
            Format: [panic,][nopanic,][num]
            Valid num: 0
            0 - turn nmi_watchdog off
            When panic is specified, panic when an NMI watchdog
            timeout occurs (or 'nopanic' to override the opposite
            This is useful when you use a panic=... timeout and
            need the box quickly up again.

            [NET] Specifies amount of time (in seconds) that
            netpoll should wait for a carrier. By default netpoll
            waits 4 seconds.

    no387        [BUGS=X86-32] Tells the kernel to use the 387 maths
            emulation library even if a 387 maths coprocessor
            is present.

            [HW] Never suspend the console
            Disable suspending of consoles during suspend and
            hibernate operations.  Once disabled, debugging
            messages can reach various consoles while the rest
            of the system is being put to sleep (ie, while
            debugging driver suspend/resume hooks).  This may
            not work reliably with all consoles, but is known
            to work with serial and VGA consoles.

    noaliencache    [MM, NUMA, SLAB] Disables the allocation of alien
            caches in the slab allocator.  Saves per-node memory,
            but will impact performance.

    noalign        [KNL,ARM]

    noapic        [SMP,APIC] Tells the kernel to not make use of any
            IOAPICs that may be present in the system.

    noautogroup    Disable scheduler automatic task group creation.

    nobats        [PPC] Do not use BATs for mapping kernel lowmem
            on "Classic" PPC cores.

    nocache        [ARM]

    noclflush    [BUGS=X86] Don't use the CLFLUSH instruction

    nodelayacct    [KNL] Disable per-task delay accounting

    nodisconnect    [HW,SCSI,M68K] Disables SCSI disconnects.

    nodsp        [SH] Disable hardware DSP at boot time.

    noefi        [X86] Disable EFI runtime services support.

    noexec        [IA-64]

    noexec        [X86]
            On X86-32 available only on PAE configured kernels.
            noexec=on: enable non-executable mappings (default)
            noexec=off: disable non-executable mappings

    nosmep        [X86]
            Disable SMEP (Supervisor Mode Execution Protection)
            even if it is supported by processor.

    noexec32    [X86-64]
            This affects only 32-bit executables.
            noexec32=on: enable non-executable mappings (default)
                read doesn't imply executable mappings
            noexec32=off: disable non-executable mappings
                read implies executable mappings

    nofpu        [SH] Disable hardware FPU at boot time.

    nofxsr        [BUGS=X86-32] Disables x86 floating point extended
            register save and restore. The kernel will only save
            legacy floating-point registers on task switch.

    noxsave        [BUGS=X86] Disables x86 extended register state save
            and restore using xsave. The kernel will fallback to
            enabling legacy floating-point and sse state.

    nohlt        [BUGS=ARM,SH] Tells the kernel that the sleep(SH) or
            wfi(ARM) instruction doesn't work correctly and not to
            use it. This is also useful when using JTAG debugger.

    no-hlt        [BUGS=X86-32] Tells the kernel that the hlt
            instruction doesn't work correctly and not to
            use it.

    no_file_caps    Tells the kernel not to honor file capabilities.  The
            only way then for a file to be executed with privilege
            is to be setuid root or executed by root.

    nohalt        [IA-64] Tells the kernel not to use the power saving
            function PAL_HALT_LIGHT when idle. This increases
            power-consumption. On the positive side, it reduces
            interrupt wake-up latency, which may improve performance
            in certain environments such as networked servers or
            real-time systems.

    nohz=        [KNL] Boottime enable/disable dynamic ticks
            Valid arguments: on, off
            Default: on

    noiotrap    [SH] Disables trapped I/O port accesses.

    noirqdebug    [X86-32] Disables the code which attempts to detect and
            disable unhandled interrupt sources.

    no_timer_check    [X86,APIC] Disables the code which tests for
            broken timer IRQ sources.

    noisapnp    [ISAPNP] Disables ISA PnP code.

    noinitrd    [RAM] Tells the kernel not to load any configured
            initial RAM disk.

    nointremap    [X86-64, Intel-IOMMU] Do not enable interrupt
            [Deprecated - use intremap=off]

    nointroute    [IA-64]

    nojitter    [IA64] Disables jitter checking for ITC timers.

    no-kvmclock    [X86,KVM] Disable paravirtualized KVM clock driver

    no-kvmapf    [X86,KVM] Disable paravirtualized asynchronous page
            fault handling.

    nolapic        [X86-32,APIC] Do not enable or use the local APIC.

    nolapic_timer    [X86-32,APIC] Do not use the local APIC timer.

    noltlbs        [PPC] Do not use large page/tlb entries for kernel
            lowmem mapping on PPC40x.

    nomca        [IA-64] Disable machine check abort handling

    nomce        [X86-32] Machine Check Exception

    nomfgpt        [X86-32] Disable Multi-Function General Purpose
            Timer usage (for AMD Geode machines).

    nopat        [X86] Disable PAT (page attribute table extension of
            pagetables) support.

    norandmaps    Don't use address space randomization.  Equivalent to
            echo 0 > /proc/sys/kernel/randomize_va_space

    noreplace-paravirt    [X86,IA-64,PV_OPS] Don't patch paravirt_ops

    noreplace-smp    [X86-32,SMP] Don't replace SMP instructions
            with UP alternatives

    noresidual    [PPC] Don't use residual data on PReP machines.

    noresume    [SWSUSP] Disables resume and restores original swap

    no-scroll    [VGA] Disables scrollback.
            This is required for the Braillex ib80-piezo Braille
            reader made by F.H. Papenmeier (Germany).

    nosbagart    [IA-64]

    nosep        [BUGS=X86-32] Disables x86 SYSENTER/SYSEXIT support.

    nosmp        [SMP] Tells an SMP kernel to act as a UP kernel,
            and disable the IO APIC.  legacy for "maxcpus=0".

    nosoftlockup    [KNL] Disable the soft-lockup detector.

    nosync        [HW,M68K] Disables sync negotiation for all devices.

    notsc        [BUGS=X86-32] Disable Time Stamp Counter

    nousb        [USB] Disable the USB subsystem

    nowatchdog    [KNL] Disable the lockup detector (NMI watchdog).

    nowb        [ARM]

    nox2apic    [X86-64,APIC] Do not enable x2APIC mode.

    nptcg=        [IA64] Override max number of concurrent global TLB
            purges which is reported from either PAL_VM_SUMMARY or
            SAL PALO.

    nr_cpus=    [SMP] Maximum number of processors that    an SMP kernel
            could support.  nr_cpus=n : n >= 1 limits the kernel to
            supporting 'n' processors. Later in runtime you can not
            use hotplug cpu feature to put more cpu back to online.
            just like you compile the kernel NR_CPUS=n

    nr_uarts=    [SERIAL] maximum number of UARTs to be registered.

    numa_zonelist_order= [KNL, BOOT] Select zonelist order for NUMA.
            one of ['zone', 'node', 'default'] can be specified
            This can be set from sysctl after boot.
            See Documentation/sysctl/vm.txt for details.

    ohci1394_dma=early    [HW] enable debugging via the ohci1394 driver.
            See Documentation/debugging-via-ohci1394.txt for more

    olpc_ec_timeout= [OLPC] ms delay when issuing EC commands
            Rather than timing out after 20 ms if an EC
            command is not properly ACKed, override the length
            of the timeout.  We have interrupts disabled while
            waiting for the ACK, so if this is set too high
            interrupts *may* be lost!

    omap_mux=    [OMAP] Override bootloader pin multiplexing.
            Format: <mux_mode0.mode_name=value>...
            For example, to override I2C bus2:

    oprofile.timer=    [HW]
            Use timer interrupt instead of performance counters

    oprofile.cpu_type=    Force an oprofile cpu type
            This might be useful if you have an older oprofile
            userland or if you want common events.
            Format: { arch_perfmon }
            arch_perfmon: [X86] Force use of architectural
                perfmon on Intel CPUs instead of the
                CPU specific event set.

    oops=panic    Always panic on oopses. Default is to just kill the
            process, but there is a small probability of
            deadlocking the machine.
            This will also cause panics on machine check exceptions.
            Useful together with panic=30 to trigger a reboot.

    OSS        [HW,OSS]
            See Documentation/sound/oss/oss-parameters.txt

    panic=        [KNL] Kernel behaviour on panic: delay <timeout>
            seconds before rebooting
            Format: <timeout>

    parkbd.port=    [HW] Parallel port number the keyboard adapter is
            connected to, default is 0.
            Format: <parport#>
    parkbd.mode=    [HW] Parallel port keyboard adapter mode of operation,
            0 for XT, 1 for AT (default is AT).
            Format: <mode>

    parport=    [HW,PPT] Specify parallel ports. 0 disables.
            Format: { 0 | auto | 0xBBB[,IRQ[,DMA]] }
            Use 'auto' to force the driver to use any
            IRQ/DMA settings detected (the default is to
            ignore detected IRQ/DMA settings because of
            possible conflicts). You can specify the base
            address, IRQ, and DMA settings; IRQ and DMA
            should be numbers, or 'auto' (for using detected
            settings on that particular port), or 'nofifo'
            (to avoid using a FIFO even if it is detected).
            Parallel ports are assigned in the order they
            are specified on the command line, starting
            with parport0.

    parport_init_mode=    [HW,PPT]
            Configure VIA parallel port to operate in
            a specific mode. This is necessary on Pegasos
            computer where firmware has no options for setting
            up parallel port mode and sets it to spp.
            Currently this function knows 686a and 8231 chips.
            Format: [spp|ps2|epp|ecp|ecpepp]

            Halt all CPUs after the first oops has been printed for
            the specified number of seconds.  This is to be used if
            your oopses keep scrolling off the screen.

    pcbit=        [HW,ISDN]

    pcd.        [PARIDE]
            See header of drivers/block/paride/pcd.c.
            See also Documentation/blockdev/paride.txt.

    pci=option[,option...]    [PCI] various PCI subsystem options:
        earlydump    [X86] dump PCI config space before the kernel
                    changes anything
        off        [X86] don't probe for the PCI bus
        bios        [X86-32] force use of PCI BIOS, don't access
                the hardware directly. Use this if your machine
                has a non-standard PCI host bridge.
        nobios        [X86-32] disallow use of PCI BIOS, only direct
                hardware access methods are allowed. Use this
                if you experience crashes upon bootup and you
                suspect they are caused by the BIOS.
        conf1        [X86] Force use of PCI Configuration
                Mechanism 1.
        conf2        [X86] Force use of PCI Configuration
                Mechanism 2.
        noaer        [PCIE] If the PCIEAER kernel config parameter is
                enabled, this kernel boot option can be used to
                disable the use of PCIE advanced error reporting.
        nodomains    [PCI] Disable support for multiple PCI
                root domains (aka PCI segments, in ACPI-speak).
        nommconf    [X86] Disable use of MMCONFIG for PCI
        check_enable_amd_mmconf [X86] check for and enable
                properly configured MMIO access to PCI
                config space on AMD family 10h CPU
        nomsi        [MSI] If the PCI_MSI kernel config parameter is
                enabled, this kernel boot option can be used to
                disable the use of MSI interrupts system-wide.
        noioapicquirk    [APIC] Disable all boot interrupt quirks.
                Safety option to keep boot IRQs enabled. This
                should never be necessary.
        ioapicreroute    [APIC] Enable rerouting of boot IRQs to the
                primary IO-APIC for bridges that cannot disable
                boot IRQs. This fixes a source of spurious IRQs
                when the system masks IRQs.
        noioapicreroute    [APIC] Disable workaround that uses the
                boot IRQ equivalent of an IRQ that connects to
                a chipset where boot IRQs cannot be disabled.
                The opposite of ioapicreroute.
        biosirq        [X86-32] Use PCI BIOS calls to get the interrupt
                routing table. These calls are known to be buggy
                on several machines and they hang the machine
                when used, but on other computers it's the only
                way to get the interrupt routing table. Try
                this option if the kernel is unable to allocate
                IRQs or discover secondary PCI buses on your
        rom        [X86] Assign address space to expansion ROMs.
                Use with caution as certain devices share
                address decoders between ROMs and other
        norom        [X86] Do not assign address space to
                expansion ROMs that do not already have
                BIOS assigned address ranges.
        nobar        [X86] Do not assign address space to the
                BARs that weren't assigned by the BIOS.
        irqmask=0xMMMM    [X86] Set a bit mask of IRQs allowed to be
                assigned automatically to PCI devices. You can
                make the kernel exclude IRQs of your ISA cards
                this way.
        pirqaddr=0xAAAAA    [X86] Specify the physical address
                of the PIRQ table (normally generated
                by the BIOS) if it is outside the
                F0000h-100000h range.
        lastbus=N    [X86] Scan all buses thru bus #N. Can be
                useful if the kernel is unable to find your
                secondary buses and you want to tell it
                explicitly which ones they are.
        assign-busses    [X86] Always assign all PCI bus
                numbers ourselves, overriding
                whatever the firmware may have done.
        usepirqmask    [X86] Honor the possible IRQ mask stored
                in the BIOS $PIR table. This is needed on
                some systems with broken BIOSes, notably
                some HP Pavilion N5400 and Omnibook XE3
                notebooks. This will have no effect if ACPI
                IRQ routing is enabled.
        noacpi        [X86] Do not use ACPI for IRQ routing
                or for PCI scanning.
        use_crs        [X86] Use PCI host bridge window information
                from ACPI.  On BIOSes from 2008 or later, this
                is enabled by default.  If you need to use this,
                please report a bug.
        nocrs        [X86] Ignore PCI host bridge windows from ACPI.
                    If you need to use this, please report a bug.
        routeirq    Do IRQ routing for all PCI devices.
                This is normally done in pci_enable_device(),
                so this option is a temporary workaround
                for broken drivers that don't call it.
        skip_isa_align    [X86] do not align io start addr, so can
                handle more pci cards
        firmware    [ARM] Do not re-enumerate the bus but instead
                just use the configuration from the
                bootloader. This is currently used on
                IXP2000 systems where the bus has to be
                configured a certain way for adjunct CPUs.
        noearly        [X86] Don't do any early type 1 scanning.
                This might help on some broken boards which
                machine check when some devices' config space
                is read. But various workarounds are disabled
                and some IOMMU drivers will not work.
        bfsort        Sort PCI devices into breadth-first order.
                This sorting is done to get a device
                order compatible with older (<= 2.4) kernels.
        nobfsort    Don't sort PCI devices into breadth-first order.
        cbiosize=nn[KMG]    The fixed amount of bus space which is
                reserved for the CardBus bridge's IO window.
                The default value is 256 bytes.
        cbmemsize=nn[KMG]    The fixed amount of bus space which is
                reserved for the CardBus bridge's memory
                window. The default value is 64 megabytes.
                [<order of align>@][<domain>:]<bus>:<slot>.<func>[; ...]
                Specifies alignment and device to reassign
                aligned memory resources.
                If <order of align> is not specified,
                PAGE_SIZE is used as alignment.
                PCI-PCI bridge can be specified, if resource
                windows need to be expanded.
        ecrc=        Enable/disable PCIe ECRC (transaction layer
                end-to-end CRC checking).
                bios: Use BIOS/firmware settings. This is the
                the default.
                off: Turn ECRC off
                on: Turn ECRC on.
        realloc        reallocate PCI resources if allocations done by BIOS
                are erroneous.

    pcie_aspm=    [PCIE] Forcibly enable or disable PCIe Active State Power
        off    Disable ASPM.
        force    Enable ASPM even on devices that claim not to support it.
            WARNING: Forcing ASPM on may cause system lockups.

    pcie_ports=    [PCIE] PCIe ports handling:
        auto    Ask the BIOS whether or not to use native PCIe services
            associated with PCIe ports (PME, hot-plug, AER).  Use
            them only if that is allowed by the BIOS.
        native    Use native PCIe services associated with PCIe ports
        compat    Treat PCIe ports as PCI-to-PCI bridges, disable the PCIe
            ports driver.

    pcie_pme=    [PCIE,PM] Native PCIe PME signaling options:
        nomsi    Do not use MSI for native PCIe PME signaling (this makes
            all PCIe root ports use INTx for all services).

    pcmv=        [HW,PCMCIA] BadgePAD 4

    pd.        [PARIDE]
            See Documentation/blockdev/paride.txt.

    pdcchassis=    [PARISC,HW] Disable/Enable PDC Chassis Status codes at
            boot time.
            Format: { 0 | 1 }
            See arch/parisc/kernel/pdc_chassis.c

    percpu_alloc=    Select which percpu first chunk allocator to use.
            Currently supported values are "embed" and "page".
            Archs may support subset or none of the    selections.
            See comments in mm/percpu.c for details on each
            allocator.  This parameter is primarily    for debugging
            and performance comparison.

    pf.        [PARIDE]
            See Documentation/blockdev/paride.txt.

    pg.        [PARIDE]
            See Documentation/blockdev/paride.txt.

    pirq=        [SMP,APIC] Manual mp-table setup
            See Documentation/x86/i386/IO-APIC.txt.

    plip=        [PPT,NET] Parallel port network link
            Format: { parport<nr> | timid | 0 }
            See also Documentation/parport.txt.

    pmtmr=        [X86] Manual setup of pmtmr I/O Port. 
            Override pmtimer IOPort with a hex value.
            e.g. pmtmr=0x508

    pnp.debug    [PNP]
            Enable PNP debug messages.  This depends on the
            CONFIG_PNP_DEBUG_MESSAGES option.

    pnpacpi=    [ACPI]
            { off }

    pnpbios=    [ISAPNP]
            { on | off | curr | res | no-curr | no-res }

            [ISAPNP] Exclude IRQs for the autoconfiguration

            [ISAPNP] Exclude DMAs for the autoconfiguration

    pnp_reserve_io=    [ISAPNP] Exclude I/O ports for the autoconfiguration
            Ranges are in pairs (I/O port base and size).

            [ISAPNP] Exclude memory regions for the
            Ranges are in pairs (memory base and size).

    ports=        [IP_VS_FTP] IPVS ftp helper module
            Default is 21.
            Up to 8 (IP_VS_APP_MAX_PORTS) ports
            may be specified.
            Format: <port>,<port>....

            [KNL] debug: print fatal signals

            If enabled, warn about various signal handling
            related application anomalies: too many signals,
            too many POSIX.1 timers, fatal signals causing a
            coredump - etc.

            If you hit the warning due to signal overflow,
            you might want to try "ulimit -i unlimited".

            default: off.

    printk.time=    Show timing data prefixed to each printk message line
            Format: <bool>  (1/Y/y=enable, 0/N/n=disable)

    processor.max_cstate=    [HW,ACPI]
            Limit processor to maximum C-state
            max_cstate=9 overrides any DMI blacklist limit.

    processor.nocst    [HW,ACPI]
            Ignore the _CST method to determine C-states,
            instead using the legacy FADT method

    profile=    [KNL] Enable kernel profiling via /proc/profile
            Format: [schedule,]<number>
            Param: "schedule" - profile schedule points.
            Param: <number> - step/bucket size as a power of 2 for
                statistical time based profiling.
            Param: "sleep" - profile D-state sleeping (millisecs).
                Requires CONFIG_SCHEDSTATS
            Param: "kvm" - profile VM exits.

    prompt_ramdisk=    [RAM] List of RAM disks to prompt for floppy disk
            before loading.
            See Documentation/blockdev/ramdisk.txt.

    psmouse.proto=    [HW,MOUSE] Highest PS2 mouse protocol extension to
            probe for; one of (bare|imps|exps|lifebook|any).
    psmouse.rate=    [HW,MOUSE] Set desired mouse report rate, in reports
            per second.
    psmouse.resetafter=    [HW,MOUSE]
            Try to reset the device after so many bad packets
            (0 = never).
            [HW,MOUSE] Set desired mouse resolution, in dpi.
            [HW,MOUSE] Controls Logitech smartscroll autorepeat.
            0 = disabled, 1 = enabled (default).

    pt.        [PARIDE]
            See Documentation/blockdev/paride.txt.

            [KNL] Number of legacy pty's. Overwrites compiled-in
            default number.

    quiet        [KNL] Disable most log messages

    r128=        [HW,DRM]

    raid=        [HW,RAID]
            See Documentation/md.txt.

    ramdisk_blocksize=    [RAM]
            See Documentation/blockdev/ramdisk.txt.

    ramdisk_size=    [RAM] Sizes of RAM disks in kilobytes
            See Documentation/blockdev/ramdisk.txt.

    rcupdate.blimit=    [KNL,BOOT]
            Set maximum number of finished RCU callbacks to process
            in one batch.

    rcupdate.qhimark=    [KNL,BOOT]
            Set threshold of queued
            RCU callbacks over which batch limiting is disabled.

    rcupdate.qlowmark=    [KNL,BOOT]
            Set threshold of queued RCU callbacks below which
            batch limiting is re-enabled.

    rdinit=        [KNL]
            Format: <full_path>
            Run specified binary instead of /init from the ramdisk,
            used for early userspace startup. See initrd.

    reboot=        [BUGS=X86-32,BUGS=ARM,BUGS=IA-64] Rebooting mode
            Format: <reboot_mode>[,<reboot_mode2>[,...]]
            See arch/*/kernel/reboot.c or arch/*/kernel/process.c

            [KNL, SMP] Set scheduler's default relax_domain_level.
            See Documentation/cgroups/cpusets.txt.

    reserve=    [KNL,BUGS] Force the kernel to ignore some iomem area

    reservetop=    [X86-32]
            Format: nn[KMG]
            Reserves a hole at the top of the kernel virtual
            address space.

    reservelow=    [X86]
            Format: nn[K]
            Set the amount of memory to reserve for BIOS at
            the bottom of the address space.

    reset_devices    [KNL] Force drivers to reset the underlying device
            during initialization.

    resume=        [SWSUSP]
            Specify the partition device for software suspend

    resume_offset=    [SWSUSP]
            Specify the offset from the beginning of the partition
            given by "resume=" at which the swap header is located,
            in <PAGE_SIZE> units (needed only for swap files).
            See  Documentation/power/swsusp-and-swap-files.txt

    hibernate=    [HIBERNATION]
        noresume    Don't check if there's a hibernation image
                present during boot.
        nocompress    Don't compress/decompress hibernation images.

    retain_initrd    [RAM] Keep initrd memory after extraction

    rhash_entries=    [KNL,NET]
            Set number of hash buckets for route cache

    riscom8=    [HW,SERIAL]
            Format: <io_board1>[,<io_board2>[,...<io_boardN>]]

    ro        [KNL] Mount root device read-only on boot

    root=        [KNL] Root filesystem

    rootdelay=    [KNL] Delay (in seconds) to pause before attempting to
            mount the root filesystem

    rootflags=    [KNL] Set root filesystem mount option string

    rootfstype=    [KNL] Set root filesystem type

    rootwait    [KNL] Wait (indefinitely) for root device to show up.
            Useful for devices that are detected asynchronously
            (e.g. USB and MMC devices).

    rw        [KNL] Mount root device read-write on boot

    S        [KNL] Run init in single mode

    sa1100ir    [NET]
            See drivers/net/irda/sa1100_ir.c.

    sbni=        [NET] Granch SBNI12 leased line adapter

    sched_debug    [KNL] Enables verbose scheduler debug messages.

    security=    [SECURITY] Choose a security module to enable at boot.
            If this boot parameter is not specified, only the first
            security module asking for security registration will be
            loaded. An invalid security module name will be treated
            as if no module has been chosen.

    selinux=    [SELINUX] Disable or enable SELinux at boot time.
            Format: { "0" | "1" }
            See security/selinux/Kconfig help text.
            0 -- disable.
            1 -- enable.
            Default value is set via kernel config option.
            If enabled at boot time, /selinux/disable can be used
            later to disable prior to initial policy load.

    apparmor=    [APPARMOR] Disable or enable AppArmor at boot time
            Format: { "0" | "1" }
            See security/apparmor/Kconfig help text
            0 -- disable.
            1 -- enable.
            Default value is set via kernel config option.

    serialnumber    [BUGS=X86-32]

    shapers=    [NET]
            Maximal number of shapers.

    show_msr=    [x86] show boot-time MSR settings
            Format: { <integer> }
            Show boot-time (BIOS-initialized) MSR settings.
            The parameter means the number of CPUs to show,
            for example 1 means boot CPU only.

    simeth=        [IA-64]

    slram=        [HW,MTD]

    slub_debug[=options[,slabs]]    [MM, SLUB]
            Enabling slub_debug allows one to determine the
            culprit if slab objects become corrupted. Enabling
            slub_debug can create guard zones around objects and
            may poison objects when not in use. Also tracks the
            last alloc / free. For more information see

    slub_max_order= [MM, SLUB]
            Determines the maximum allowed order for slabs.
            A high setting may cause OOMs due to memory
            fragmentation. For more information see

    slub_min_objects=    [MM, SLUB]
            The minimum number of objects per slab. SLUB will
            increase the slab order up to slub_max_order to
            generate a sufficiently large slab able to contain
            the number of objects indicated. The higher the number
            of objects the smaller the overhead of tracking slabs
            and the less frequently locks need to be acquired.
            For more information see Documentation/vm/slub.txt.

    slub_min_order=    [MM, SLUB]
            Determines the mininum page order for slabs. Must be
            lower than slub_max_order.
            For more information see Documentation/vm/slub.txt.

    slub_nomerge    [MM, SLUB]
            Disable merging of slabs with similar size. May be
            necessary if there is some reason to distinguish
            allocs to different slabs. Debug options disable
            merging on their own.
            For more information see Documentation/vm/slub.txt.

    smart2=        [HW]
            Format: <io1>[,<io2>[,...,<io8>]]

    smp-alt-once    [X86-32,SMP] On a hotplug CPU system, only
            attempt to substitute SMP alternatives once at boot.

    smsc-ircc2.nopnp    [HW] Don't use PNP to discover SMC devices
    smsc-ircc2.ircc_cfg=    [HW] Device configuration I/O port
    smsc-ircc2.ircc_sir=    [HW] SIR base I/O port
    smsc-ircc2.ircc_fir=    [HW] FIR base I/O port
    smsc-ircc2.ircc_irq=    [HW] IRQ line
    smsc-ircc2.ircc_dma=    [HW] DMA channel
    smsc-ircc2.ircc_transceiver= [HW] Transceiver type:
                0: Toshiba Satellite 1800 (GP data pin select)
                1: Fast pin select (default)
                2: ATC IRMode

            [KNL] Should the soft-lockup detector generate panics.
            Format: <integer>

    sonypi.*=    [HW] Sony Programmable I/O Control Device driver
            See Documentation/sonypi.txt

    specialix=    [HW,SERIAL] Specialix multi-serial port adapter
            See Documentation/serial/specialix.txt.

    spia_io_base=    [HW,MTD]

    stacktrace    [FTRACE]
            Enabled the stack tracer on boot up.

    sti=        [PARISC,HW]
            Format: <num>
            Set the STI (builtin display/keyboard on the HP-PARISC
            machines) console (graphic card) which should be used
            as the initial boot-console.
            See also comment in drivers/video/console/sticore.c.

    sti_font=    [HW]
            See comment in drivers/video/console/sticore.c.

    stifb=        [HW]
            Format: bpp:<bpp1>[:<bpp2>[:<bpp3>...]]

            SunRPC servers often require that client requests
            originate from a privileged port (i.e. a port in the
            range 0 < portnr < 1024).
            An administrator who wishes to reserve some of these
            ports for other uses may adjust the range that the
            kernel's sunrpc client considers to be privileged
            using these two parameters to set the minimum and
            maximum port values.

            Control how the NFS server code allocates CPUs to
            service thread pools.  Depending on how many NICs
            you have and where their interrupts are bound, this
            option will affect which CPUs will do NFS serving.
            Note: this parameter cannot be changed while the
            NFS server is running.

            auto        the server chooses an appropriate mode
                    automatically using heuristics
            global        a single global pool contains all CPUs
            percpu        one pool for each CPU
            pernode        one pool for each NUMA node (equivalent
                    to global on non-NUMA machines)

            Sets the upper limit on the number of simultaneous
            RPC calls that can be sent from the client to a
            server. Increasing these values may allow you to
            improve throughput, but will also increase the
            amount of memory reserved for use by the client.

            [KNL] Enable accounting of swap in memory resource
            controller if no parameter or 1 is given or disable
            it if 0 is given (See Documentation/cgroups/memory.txt)

    swiotlb=    [IA-64] Number of I/O TLB slabs

    switches=    [HW,M68k]

    sysfs.deprecated=0|1 [KNL]
            Enable/disable old style sysfs layout for old udev
            on older distributions. When this option is enabled
            very new udev will not work anymore. When this option
            is disabled (or CONFIG_SYSFS_DEPRECATED not compiled)
            in older udev will not work anymore.
            Default depends on CONFIG_SYSFS_DEPRECATED_V2 set in
            the kernel configuration.

            Ignore sysrq setting - this boot parameter will
            neutralize any effect of /proc/sys/kernel/sysrq.
            Useful for debugging.

    tdfx=        [HW,DRM]

    test_suspend=    [SUSPEND]
            Specify "mem" (for Suspend-to-RAM) or "standby" (for
            standby suspend) as the system sleep state to briefly
            enter during system startup.  The system is woken from
            this state using a wakeup-capable RTC alarm.

    thash_entries=    [KNL,NET]
            Set number of hash buckets for TCP connection

    thermal.act=    [HW,ACPI]
            -1: disable all active trip points in all thermal zones
            <degrees C>: override all lowest active trip points

    thermal.crt=    [HW,ACPI]
            -1: disable all critical trip points in all thermal zones
            <degrees C>: override all critical trip points

    thermal.nocrt=    [HW,ACPI]
            Set to disable actions on ACPI thermal zone
            critical and hot trip points.    [HW,ACPI]
            1: disable ACPI thermal control

    thermal.psv=    [HW,ACPI]
            -1: disable all passive trip points
            <degrees C>: override all passive trip points to this

    thermal.tzp=    [HW,ACPI]
            Specify global default ACPI thermal zone polling rate
            <deci-seconds>: poll all this frequency
            0: no polling (default)

    threadirqs    [KNL]
            Force threading of all interrupt handlers except those
            marked explicitely IRQF_NO_THREAD.

    topology=    [S390]
            Format: {off | on}
            Specify if the kernel should make use of the cpu
            topology information if the hardware supports this.
            The scheduler will make use of this information and
            e.g. base its process migration decisions on it.
            Default is on.

    tp720=        [HW,PS2]

            Format: integer pcr id
            Specify that at suspend time, the tpm driver
            should extend the specified pcr with zeros,
            as a workaround for some chips which fail to
            flush the last written pcr on TPM_SaveState.
            This will guarantee that all the other pcrs
            are saved.

            [FTRACE] will set tracing buffer size.

            [FTRACE] Set and start specified trace events in order
            to facilitate early boot debugging.
            See also Documentation/trace/events.txt

    tsc=        Disable clocksource stability checks for TSC.
            Format: <string>
            [x86] reliable: mark tsc clocksource as reliable, this
            disables clocksource verification at runtime, as well
            as the stability checks done at bootup.    Used to enable
            high-resolution timer mode on older hardware, and in
            virtualized environment.
            [x86] noirqtime: Do not use TSC to do irq accounting.
            Used to run time disable IRQ_TIME_ACCOUNTING on any
            platforms where RDTSC is slow and this accounting
            can add overhead.[2|3]=    [HW,JOY]
            TurboGraFX parallel port interface
            See also Documentation/input/joystick-parport.txt

    uhash_entries=    [KNL,NET]
            Set number of hash buckets for UDP/UDP-Lite connections

            [USB] Ignore overcurrent events (default N).
            Some badly-designed motherboards generate lots of
            bogus events, for ports that aren't wired to
            anything.  Set this parameter to avoid log spamming.
            Note that genuine overcurrent events won't be
            reported either.

            [X86] Cause panic on unknown NMI.

            [USB] The autosuspend time delay (in seconds) used
            for newly-detected USB devices (default 2).  This
            is the time required before an idle device will be
            autosuspended.  Devices for which the delay is set
            to a negative value won't be autosuspended at all.

            [USB] Set to log all usbfs traffic (default 0 = off).

            [USB] Set to cycle leds on hubs (default 0 = off).

            [USB] Start with the old device initialization
            scheme (default 0 = off).

            [USB] Try the other device initialization scheme
            if the first one fails (default 1 = enabled).

            [USB] Specifies timeout for the initial 64-byte
                        USB_REQ_GET_DESCRIPTOR request in milliseconds
            (default 5000 = 5.0 seconds).

            [USBHID] The interval which mice are to be polled at.

            [UMS] The delay in seconds before a new device is
            scanned for Logical Units (default 5).

            [UMS] A list of quirks entries to supplement or
            override the built-in unusual_devs list.  List
            entries are separated by commas.  Each entry has
            the form VID:PID:Flags where VID and PID are Vendor
            and Product ID values (4-digit hex numbers) and
            Flags is a set of characters, each corresponding
            to a common usb-storage quirk flag as follows:
                a = SANE_SENSE (collect more than 18 bytes
                    of sense data);
                b = BAD_SENSE (don't collect more than 18
                    bytes of sense data);
                c = FIX_CAPACITY (decrease the reported
                    device capacity by one sector);
                d = NO_READ_DISC_INFO (don't use
                    READ_DISC_INFO command);
                e = NO_READ_CAPACITY_16 (don't use
                    READ_CAPACITY_16 command);
                h = CAPACITY_HEURISTICS (decrease the
                    reported device capacity by one
                    sector if the number is odd);
                i = IGNORE_DEVICE (don't bind to this
                l = NOT_LOCKABLE (don't try to lock and
                    unlock ejectable media);
                m = MAX_SECTORS_64 (don't transfer more
                    than 64 sectors = 32 KB at a time);
                n = INITIAL_READ10 (force a retry of the
                    initial READ(10) command);
                o = CAPACITY_OK (accept the capacity
                    reported by the device);
                r = IGNORE_RESIDUE (the device reports
                    bogus residue values);
                s = SINGLE_LUN (the device has only one
                    Logical Unit);
                w = NO_WP_DETECT (don't test whether the
                    medium is write-protected).
            Example: quirks=0419:aaf5:rl,0421:0433:rc

            [X86] Flags controlling user PTE allocations.

                nohigh = do not allocate PTE pages in
                    HIGHMEM regardless of setting
                    of CONFIG_HIGHPTE.

    vdso=        [X86,SH]
            vdso=2: enable compat VDSO (default with COMPAT_VDSO)
            vdso=1: enable VDSO (default)
            vdso=0: disable VDSO mapping

    vdso32=        [X86]
            vdso32=2: enable compat VDSO (default with COMPAT_VDSO)
            vdso32=1: enable 32-bit VDSO (default)
            vdso32=0: disable 32-bit VDSO mapping

    vector=        [IA-64,SMP]
            vector=percpu: enable percpu vector domain

    video=        [FB] Frame buffer configuration
            See Documentation/fb/modedb.txt.

    vga=        [BOOT,X86-32] Select a particular video mode
            See Documentation/x86/boot.txt and
            Use vga=ask for menu.
            This is actually a boot loader parameter; the value is
            passed to the kernel using a special protocol.

    vmalloc=nn[KMG]    [KNL,BOOT] Forces the vmalloc area to have an exact
            size of <nn>. This can be used to increase the
            minimum size (128MB on x86). It can also be used to
            decrease the size and leave more room for directly
            mapped kernel RAM.

    vmhalt=        [KNL,S390] Perform z/VM CP command after system halt.
            Format: <command>

    vmpanic=    [KNL,S390] Perform z/VM CP command after kernel panic.
            Format: <command>

    vmpoff=        [KNL,S390] Perform z/VM CP command after power off.
            Format: <command>

    vt.cur_default=    [VT] Default cursor shape.
            Format: 0xCCBBAA, where AA, BB, and CC are the same as
            the parameters of the <Esc>[?A;B;Cc escape sequence;
            see VGA-softcursor.txt. Default: 2 = underline.

    vt.default_blu=    [VT]
            Format: <blue0>,<blue1>,<blue2>,...,<blue15>
            Change the default blue palette of the console.
            This is a 16-member array composed of values
            ranging from 0-255.

    vt.default_grn=    [VT]
            Format: <green0>,<green1>,<green2>,...,<green15>
            Change the default green palette of the console.
            This is a 16-member array composed of values
            ranging from 0-255.

    vt.default_red=    [VT]
            Format: <red0>,<red1>,<red2>,...,<red15>
            Change the default red palette of the console.
            This is a 16-member array composed of values
            ranging from 0-255.

            Set system-wide default UTF-8 mode for all tty's.
            Default is 1, i.e. UTF-8 mode is enabled for all
            newly opened terminals.

            Set system-wide default for whether a cursor
            is shown on new VTs. Default is -1,
            i.e. cursors will be created by default unless
            overridden by individual drivers. 0 will hide
            cursors, 1 will display them.

    watchdog timers    [HW,WDT] For information on watchdog timers,
            see Documentation/watchdog/watchdog-parameters.txt
            or other driver-specific files in the
            Documentation/watchdog/ directory.

    x2apic_phys    [X86-64,APIC] Use x2apic physical mode instead of
            default x2apic cluster mode on platforms
            supporting x2apic.

    x86_mrst_timer= [X86-32,APBT]
            Choose timer option for x86 Moorestown MID platform.
            Two valid options are apbt timer only and lapic timer
            plus one apbt timer for broadcast timer.
            x86_mrst_timer=apbt_only | lapic_and_apbt

    xd=        [HW,XT] Original XT pre-IDE (RLL encoded) disks.
    xd_geo=        See header of drivers/block/xd.c.

    xen_emul_unplug=        [HW,X86,XEN]
            Unplug Xen emulated devices
            Format: [unplug0,][unplug1]
            ide-disks -- unplug primary master IDE devices
            aux-ide-disks -- unplug non-primary-master IDE devices
            nics -- unplug network devices
            all -- unplug all emulated devices (NICs and IDE disks)
            unnecessary -- unplugging emulated devices is
                unnecessary even if the host did not respond to
                the unplug protocol
            never -- do not unplug even if version check succeeds

    xirc2ps_cs=    [NET,PCMCIA]



    Add more DRM drivers.
twitter facebook me2day 요즘
57개(1/4페이지) rss
번호 제목 글쓴이 조회 날짜
공지 Microsoft 볼륨 라이선스 서비스 센터(VLSC) 가이드 (2015년 8월) 첨부파일 관리자 133910 2016.04.10 11:34
56 볼륨 라이선스 키에 대한 제품 라이선스 FAQ(한국어) 첨부파일 관리자 258 2019.11.02 11:11
55 볼륨 라이선싱 제품 조건 2019년 11월 1일(한국어) 첨부파일 관리자 308 2019.11.02 11:09
54 OWASP Top 10 - 2017 Release Candidate 첨부파일 관리자 2776 2017.04.13 14:15
53 Windows Server 2016 and System Center 2016 Pricing and Licen 첨부파일 관리자 9071 2016.08.15 14:20
52 Windows Server 2016 Licensing Datasheet 첨부파일 관리자 8690 2016.08.15 14:05
51 NEXT RS232 2P Driver 첨부파일 관리자2 6967 2014.12.17 15:17
50 자주 사용하는 파일들 첨부파일 관리자2 9037 2014.07.21 13:19
49 원격지원 요청 첨부파일 관리자 135959 2013.12.12 20:42
48 Microsoft License Guide 2013년 첨부파일 관리자 166355 2013.06.18 18:46
47 Microsoft 볼륨 라이선싱 제품 사용권 설명서(2013년1월) 첨부파일 관리자 20595 2013.03.21 19:29
46 Windows Server 2012 제품 개요 첨부파일 관리자 26075 2013.01.08 21:32
45 왜 Hyper-V여야 하는가? 첨부파일 관리자 15832 2013.01.08 20:17
44 마이크로소프트 제품 목록(2013)(영문) 첨부파일 관리자 12835 2013.01.05 11:51
43 Windows Server 2012 Licensing Data Sheet 첨부파일 관리자 18399 2012.11.18 22:57
42 Windows Server 2003 R2, Windows Server 2008 R2, and Windows 첨부파일 관리자 22229 2012.11.18 22:52