The Linux BootPrompt-HowTo
 by Paul Gortmaker.
 v1.2, May 1999

 This is the BootPrompt-Howto, which is a compilation of all the possi-
 ble boot time arguments that can be passed to the Linux kernel at boot
 time. This includes all kernel and device parameters.  A discussion of
 how the kernel sorts boot time arguments, along with an overview of
 some of the popular software used to boot Linux kernels is also
 included.
 ______________________________________________________________________

 Table of Contents



 1. Introduction

    1.1 Disclaimer and Copyright
    1.2 Intended Audience and Applicability
    1.3 Related Documentation
    1.4 The Linux Newsgroups
    1.5 New Versions of this Document

 2. Overview of Boot Prompt Arguments

    2.1 LILO (LInux LOader)
    2.2 LoadLin
    2.3 The ``rdev'' utility
    2.4 How the Kernel Sorts the Arguments
    2.5 Setting Environment Variables.
    2.6 Passing Arguments to the `init' program

 3. General Non-Device Specific Boot Args

    3.1 Root Filesystem options
       3.1.1 The `root=' Argument
       3.1.2 The `ro' Argument
       3.1.3 The `rw' Argument
    3.2 Options Relating to RAM Disk Management
       3.2.1 The `ramdisk_start=' Argument
       3.2.2 The `load_ramdisk=' Argument
       3.2.3 The `prompt_ramdisk=' Argument
       3.2.4 The `ramdisk_size=' Argument
       3.2.5 The `ramdisk=' Argument (obsolete)
       3.2.6 The `noinitrd' (initial RAM disk) Argument
    3.3 Boot Arguments Related to Memory Handling
       3.3.1 The `mem=' Argument
       3.3.2 The `swap=' Argument
       3.3.3 The `buff=' Argument
    3.4 Boot Arguments for NFS Root Filesystem
       3.4.1 The `nfsroot=' Argument
       3.4.2 The `nfsaddrs=' Argument
    3.5 Other Misc. Kernel Boot Arguments
       3.5.1 The `debug' Argument
       3.5.2 The `init=' Argument
       3.5.3 The `kbd-reset' Argument
       3.5.4 The `maxcpus=' Argument
       3.5.5 The `mca-pentium' Argument
       3.5.6 The `md=' Argument
       3.5.7 The `no387' Argument
       3.5.8 The `no-hlt' Argument
       3.5.9 The `no-scroll' Argument
       3.5.10 The `noapic' Argument
       3.5.11 The `nosmp' Argument
       3.5.12 The `panic=' Argument
       3.5.13 The `pci=' Argument
       3.5.14 The `pirq=' Argument
       3.5.15 The `profile=' Argument
       3.5.16 The `reboot=' Argument
       3.5.17 The `reserve=' Argument
       3.5.18 The `vga=' Argument

 4. Boot Arguments to Control PCI Bus Behaviour (`pci=')

    4.1 The `pci=bios' and `pci=nobios' Arguments
    4.2 The `pci=conf1' and `pci=conf2' Arguments
    4.3 The `pci=io=' Argument
    4.4 The `pci=nopeer' Argument
    4.5 The `pci=nosort' Argument
    4.6 The `pci=off' Argument
    4.7 The `pci=reverse' Argument
 5. Boot Arguments for Video Frame Buffer Drivers

    5.1 The `video=map:...' Argument
    5.2 The `video=scrollback:...' Argument
    5.3 The `video=vc:...' Argument

 6. Boot Arguments for SCSI Peripherals.

    6.1 Arguments for Mid-level Drivers
       6.1.1 Maximum Probed LUNs (`max_scsi_luns=')
       6.1.2 SCSI Logging (`scsi_logging=')
       6.1.3 Parameters for the SCSI Tape Driver (`st=')
    6.2 Arguments for SCSI Host Adapters
       6.2.1 Adaptec aha151x, aha152x, aic6260, aic6360, SB16-SCSI (`aha152x=')
       6.2.2 Adaptec aha154x (`aha1542=')
       6.2.3 Adaptec aha274x, aha284x, aic7xxx (`aic7xxx=')
       6.2.4 AdvanSys SCSI Host Adaptors (`advansys=')
       6.2.5 Always IN2000 Host Adaptor (`in2000=')
       6.2.6 AMD AM53C974 based hardware (`AM53C974=')
       6.2.7 BusLogic SCSI Hosts with v1.2 kernels (`buslogic=')
       6.2.8 BusLogic SCSI Hosts with v2.x kernels (`BusLogic=')
       6.2.9 EATA SCSI Cards (`eata=')
       6.2.10 Future Domain TMC-8xx, TMC-950 (`tmc8xx=')
       6.2.11 Future Domain TMC-16xx, TMC-3260, AHA-2920 (`fdomain=')
       6.2.12 IOMEGA Parallel Port / ZIP drive (`ppa=')
       6.2.13 NCR5380 based controllers (`ncr5380=')
       6.2.14 NCR53c400 based controllers (`ncr53c400=')
       6.2.15 NCR53c406a based controllers (`ncr53c406a=')
       6.2.16 Pro Audio Spectrum (`pas16=')
       6.2.17 Seagate ST-0x (`st0x=')
       6.2.18 Trantor T128 (`t128=')
       6.2.19 Ultrastor SCSI cards (`u14-34f=')
       6.2.20 Western Digital WD7000 cards (`wd7000=')
    6.3 SCSI Host Adapters that don't Accept Boot Args

 7. Hard Disks

    7.1 IDE Disk/CD-ROM Driver Parameters
    7.2 Standard ST-506 Disk Driver Options (`hd=')
    7.3 XT Disk Driver Options (`xd=')

 8. CD-ROMs (Non-SCSI/ATAPI/IDE)

    8.1 The Aztech Interface (`aztcd=')
    8.2 The CDU-31A and CDU-33A Sony Interface (`cdu31a=')
    8.3 The CDU-535 Sony Interface (`sonycd535=')
    8.4 The GoldStar Interface (`gscd=')
    8.5 The ISP16 Interface (`isp16=')
    8.6 The Mitsumi Standard Interface (`mcd=')
    8.7 The Mitsumi XA/MultiSession Interface (`mcdx=')
    8.8 The Optics Storage Interface (`optcd=')
    8.9 The Phillips CM206 Interface (`cm206=')
    8.10 The Sanyo Interface (`sjcd=')
    8.11 The SoundBlaster Pro Interface (`sbpcd=')

 9. Serial and ISDN Drivers

    9.1 The ICN ISDN driver (`icn=')
    9.2 The PCBIT ISDN driver (`pcbit=')
    9.3 The Teles ISDN driver (`teles=')
    9.4 The DigiBoard Driver (`digi=')
    9.5 The RISCom/8 Multiport Serial Driver (`riscom8=')
    9.6 The Baycom Serial/Parallel Radio Modem (`baycom=')

 10. Other Hardware Devices

    10.1 Ethernet Devices (`ether=')
    10.2 The Floppy Disk Driver (`floppy=')
    10.3 The Sound Driver (`sound=')
    10.4 The Bus Mouse Driver (`bmouse=')
    10.5 The MS Bus Mouse Driver (`msmouse=')
    10.6 The Printer Driver (`lp=')

 11. Copying, Translations, Closing, etc.

    11.1 Copyright and Disclaimer
    11.2 Closing


 ______________________________________________________________________

 1.  Introduction


 The kernel has a limited capability to accept information at boot in
 the form of a `command line', similar to an argument list you would
 give to a program. In general this is used to supply the kernel with
 information about hardware parameters that the kernel would not be
 able to determine on its own, or to avoid/override the values that the
 kernel would otherwise detect.

 However, if you just copy a kernel image directly to a floppy, (e.g.
 cp zImage /dev/fd0) then you are not given a chance to specify any
 arguments to that kernel. So most Linux users will use software like
 LILO or loadlin that takes care of handing these arguments to the
 kernel, and then booting it.

 This present revision covers kernels up to and including v2.2.9.  Some
 features that are unique to development/testing kernels up to v2.3.2
 are also documented.

 The BootPrompt-Howto is by:

      Paul Gortmaker, [email protected]



 1.1.  Disclaimer and Copyright

 This document is Copyright (c) 1995-1999 by Paul Gortmaker.  Please
 see the Disclaimer and Copying information at the end of this document
 (``copyright'') for information about redistribution of this document
 and the usual `we are not responsible for what you manage to break...'
 type legal stuff.


 1.2.  Intended Audience and Applicability

 Most Linux users should never have to even look at this document.
 Linux does an exceptionally good job at detecting most hardware and
 picking reasonable default settings for most parameters.  The
 information in this document is aimed at users who might want to
 change some of the default settings to optimize the kernel to their
 particular machine, or to a user who has `rolled their own' kernel to
 support a not so common piece of hardware for which automatic
 detection is currently not available.

 IMPORTANT NOTE: Driver related boot prompt arguments only apply to
 hardware drivers that are compiled directly into the kernel. They have
 no effect on drivers that are loaded as modules. Most Linux
 distributions come with a basic `bare-bones' kernel, and the drivers
 are small modules that are loaded after the kernel has initialized.
 If you are unsure if you are using modules then look at man depmod and
 man modprobe along with the contents of your /etc/conf.modules.


 1.3.  Related Documentation

 The most up-to-date documentation will always be the kernel source
 itself. Hold on! Don't get scared. You don't need to know any
 programming to read the comments in the source files.  For example, if
 you were looking for what arguments could be passed to the AHA1542
 SCSI driver, then you would go to the linux/drivers/scsi directory,
 and look at the file aha1542.c -- and within the first 100 lines, you
 would find a plain english description of the boot time arguments that
 the 1542 driver accepts.

 The linux directory is usually found in /usr/src/ for most
 distributions.  All references in this document to files that come
 with the kernel will have their pathname abbreviated to start with
 linux - you will have to append the /usr/src/ or whatever is
 appropriate for your system.  (If you can't find the file in question,
 then make use of the find and locate commands.)

 The next best thing will be any documentation files that are
 distributed with the kernel itself. There are now quite a few of
 these, and most of them can be found in the directory
 linux/Documentation and subdirectories from there.  Sometimes there
 will be README.foo files that can be found in the related driver
 directory (e.g. linux/drivers/???/, where examples of ??? could be
 scsi, char, or net).

 If you have figured out what boot-args you intend to use, and now want
 to know how to get that information to the kernel, then look at the
 documentation that comes with the software that you use to boot the
 kernel (e.g. LILO or loadlin). A brief overview is given below, but it
 is no substitute for the documentation that comes with the booting
 software.


 1.4.  The Linux Newsgroups

 If you have questions about passing boot arguments to the kernel,
 please check this document first. If this and the related
 documentation mentioned above does not answer your question(s) then
 you can try the Linux newsgroups.  General questions on how to
 configure your system should be directed to the comp.os.linux.setup
 newsgroup.  We ask that you please respect this general guideline for
 content, and don't cross-post your request to other groups.

 Of course you should try checking the group before blindly posting
 your question, as it may even be a Frequently Asked Question (a FAQ).
 A quick browse of the Linux FAQ before posting is a good idea. You
 should be able to find the FAQ somewhere close to where you found this
 document.  If it is not a FAQ then use newsgroup archives, such as
 those at http://www.dejanews.com to quickly search years worth of
 postings for your topic. Chances are someone else has already asked
 (and another person answered!) the question that you now have.


 1.5.  New Versions of this Document

 New versions of this document can be retrieved via anonymous FTP from
 most Linux FTP sites in the directory /pub/Linux/docs/HOWTO/. Updates
 will be made as new information and/or drivers becomes available. If
 this copy that you are presently reading is more than six months old,
 then you should probably check to see if a newer copy exists.  I would
 recommend viewing this via a WWW browser or in the Postscript/dvi
 format. Both of these contain cross-references that are lost in a
 simple plain text version.

 If you want to get the official copy, here is URL.

 BootPrompt-HOWTO <http://metalab.unc.edu/mdw/HOWTO/BootPrompt-
 HOWTO.html>


 2.  Overview of Boot Prompt Arguments


 This section gives some examples of software that can be used to pass
 kernel boot-time arguments to the kernel itself.  It also gives you an
 idea of how the arguments are processed, what limitations there are on
 the boot args, and how they filter down to each appropriate device
 that they are intended for.

 It is important to note that spaces should not be used in a boot
 argument, but only between separate arguments.  A list of values that
 are for a single argument are to be separated with a comma between the
 values, and again without any spaces. See the following examples
 below.


 ______________________________________________________________________
         ether=9,0x300,0xd0000,0xd4000,eth0  root=/dev/hda1            *RIGHT*
         ether = 9, 0x300, 0xd0000, 0xd4000, eth0  root = /dev/hda1    *WRONG*
 ______________________________________________________________________



 Once the Linux kernel is up and running, one can view the command line
 arguments that were in place at boot by simply typing cat
 /proc/cmdline at a shell prompt.


 2.1.  LILO (LInux LOader)

 The LILO program (LInux LOader) written by Werner Almesberger is the
 most commonly used. It has the ability to boot various kernels, and
 stores the configuration information in a plain text file. Most
 distributions ship with LILO as the default boot-loader. LILO can boot
 DOS, OS/2, Linux, FreeBSD, etc. without any difficulties, and is quite
 flexible.

 A typical configuration will have LILO stop and print LILO: shortly
 after you turn on your computer. It will then wait for a few seconds
 for any optional input from the user, and failing that it will then
 boot the default system. Typical system labels that people use in the
 LILO configuration files are linux and backup and msdos. If you want
 to type in a boot argument, you type it in here, after typing in the
 system label that you want LILO to boot from, as shown in the example
 below.


 ______________________________________________________________________
         LILO: linux root=/dev/hda1
 ______________________________________________________________________



 LILO comes with excellent documentation, and for the purposes of boot
 args discussed here, the LILO append= command is of significant
 importance when one wants to add a boot time argument as a permanent
 addition to the LILO config file.  You simply add something like
 append = "foo=bar" to the /etc/lilo.conf file. It can either be added
 at the top of the config file, making it apply to all sections, or to
 a single system section by adding it inside an image= section.  Please
 see the LILO documentation for a more complete description.


 2.2.  LoadLin

 The other commonly used Linux loader is `LoadLin' which is a DOS
 program that has the capability to launch a Linux kernel from the DOS
 prompt (with boot-args) assuming that certain resources are available.
 This is good for people that use DOS and want to launch into Linux
 from DOS.

 It is also very useful if you have certain hardware which relies on
 the supplied DOS driver to put the hardware into a known state. A
 common example is `SoundBlaster Compatible' sound cards that require
 the DOS driver to set a few proprietary registers to put the card into
 a SB compatible mode. Booting DOS with the supplied driver, and then
 loading Linux from the DOS prompt with LOADLIN.EXE avoids the reset of
 the card that happens if one rebooted instead. Thus the card is left
 in a SB compatible mode and hence is useable under Linux.

 There are also other programs that can be used to boot Linux.  For a
 complete list, please look at the programs available on your local
 Linux ftp mirror, under system/Linux-boot/.


 2.3.  The ``rdev'' utility

 There are a few of the kernel boot parameters that have their default
 values stored in various bytes in the kernel image itself.  There is a
 utility called rdev that is installed on most systems that knows where
 these values are, and how to change them.  It can also change things
 that have no kernel boot argument equivalent, such as the default
 video mode used.

 The rdev utility is usually also aliased to swapdev, ramsize, vidmode
 and rootflags. These are the five things that rdev can change, those
 being the root device, the swap device, the RAM disk parameters, the
 default video mode, and the readonly/readwrite setting of root device.

 More information on rdev can be found by typing rdev -h or by reading
 the supplied man page (man rdev).


 2.4.  How the Kernel Sorts the Arguments

 Most of the boot args take the form of:

 ______________________________________________________________________
         name[=value_1][,value_2]...[,value_11]
 ______________________________________________________________________



 where `name' is a unique keyword that is used to identify what part of
 the kernel the associated values (if any) are to be given to. Multiple
 boot args are just a space separated list of the above format. Note
 the limit of 11 is real, as the present code only handles 11 comma
 separated parameters per keyword. (However, you can re-use the same
 keyword with up to an additional 11 parameters in unusually
 complicated situations, assuming the setup function supports it.)
 Also note that the kernel splits the list into a maximum of ten
 integer arguments, and a following string, so you can't really supply
 11 integers unless you convert the 11th arg from a string to an int in
 the driver itself.

 Most of the sorting goes on in linux/init/main.c.  First, the kernel
 checks to see if the argument is any of the special arguments `root=',
 `ro', `rw', or `debug'.  The meaning of these special arguments is
 described further on in the document.

 Then it walks a list of setup functions (contained in the bootsetups
 array) to see if the specified argument string (such as `foo') has
 been associated with a setup function (foo_setup()) for a particular
 device or part of the kernel. If you passed the kernel the line
 foo=3,4,5,6,bar then the kernel would search the bootsetups array to
 see if `foo' was registered. If it was, then it would call the setup
 function associated with `foo' (foo_setup()) and hand it the integer
 arguments 3, 4, 5 and 6 as given on the kernel command line, and also
 hand it the string argument bar.


 2.5.  Setting Environment Variables.

 Anything of the form `foo=bar' that is not accepted as a setup
 function as described above is then interpreted as an environment
 variable to be set. An example would be to use TERM=vt100 or
 BOOT_IMAGE=vmlinuz.bak as a boot argument.  These environment
 variables are typically tested for in the initialization scripts to
 enable or disable a wide range of things.


 2.6.  Passing Arguments to the `init' program

 Any remaining arguments that were not picked up by the kernel and were
 not interpreted as environment variables are then passed onto process
 one, which is usually the init program. The most common argument that
 is passed to the init process is the word single which instructs init
 to boot the computer in single user mode, and not launch all the usual
 daemons. Check the manual page for the version of init installed on
 your system to see what arguments it accepts.


 3.  General Non-Device Specific Boot Args

 These are the boot arguments that are not related to any specific
 device or peripheral. They are instead related to certain internal
 kernel parameters, such as memory handling, ramdisk handling, root
 file system handling and others.


 3.1.  Root Filesystem options

 The following options all pertain to how the kernel selects and
 handles the root filesystem.


 3.1.1.  The `root=' Argument

 This argument tells the kernel what device is to be used as the root
 filesystem while booting. The default of this setting is the value of
 the root device of the system that the kernel was built on.  For
 example, if the kernel in question was built on a system that used
 `/dev/hda1' as the root partition, then the default root device would
 be `/dev/hda1'.  To override this default value, and select the second
 floppy drive as the root device, one would use `root=/dev/fd1'.

 Valid root devices are any of the following devices:


 (1) /dev/hdaN to /dev/hddN, which is partition N on ST-506 compatible
 disk `a to d'.

 (2) /dev/sdaN to /dev/sdeN, which is partition N on SCSI compatible
 disk `a to e'.

 (3) /dev/xdaN to /dev/xdbN, which is partition N on XT compatible disk
 `a to b'.

 (4) /dev/fdN, which is floppy disk drive number N. Having N=0 would be
 the DOS `A:' drive, and N=1 would be `B:'.

 (5) /dev/nfs, which is not really a device, but rather a flag to tell
 the kernel to get the root fs via the network.

 The more awkward and less portable numeric specification of the above
 possible disk devices in major/minor format is also accepted. (e.g.
 /dev/sda3 is major 8, minor 3, so you could use root=0x803 as an
 alternative.)

 This is one of the few kernel boot arguments that has its default
 stored in the kernel image, and which can thus be altered with the
 rdev utility.



 3.1.2.  The `ro' Argument

 When the kernel boots, it needs a root filesystem to read basic things
 off of. This is the root filesystem that is mounted at boot. However,
 if the root filesystem is mounted with write access, you can not
 reliably check the filesystem integrity with half-written files in
 progress. The `ro' option tells the kernel to mount the root
 filesystem as `readonly' so that any filesystem consistency check
 programs (fsck) can safely assume that there are no half-written files
 in progress while performing the check. No programs or processes can
 write to files on the filesystem in question until it is `remounted'
 as read/write capable.

 This is one of the few kernel boot arguments that has its default
 stored in the kernel image, and which can thus be altered with the
 rdev utility.


 3.1.3.  The `rw' Argument

 This is the exact opposite of the above, in that it tells the kernel
 to mount the root filesystem as read/write. The default is to mount
 the root filesystem as read/write anyway. Do not run any `fsck' type
 programs on a filesystem that is mounted read/write.

 The same value stored in the image file mentioned above is also used
 for this parameter, accessible via rdev.


 3.2.  Options Relating to RAM Disk Management

 The following options all relate to how the kernel handles the RAM
 disk device, which is usually used for bootstrapping machines during
 the install phase, or for machines with modular drivers that need to
 be installed to access the root filesystem.



 3.2.1.  The `ramdisk_start=' Argument

 To allow a kernel image to reside on a floppy disk along with a
 compressed ramdisk image, the `ramdisk_start=<offset>' command was
 added. The kernel can't be included into the compressed ramdisk
 filesystem image, because it needs to be stored starting at block zero
 so that the BIOS can load the bootsector and then the kernel can
 bootstrap itself to get going.

 Note: If you are using an uncompressed ramdisk image, then the kernel
 can be a part of the filesystem image that is being loaded into the
 ramdisk, and the floppy can be booted with LILO, or the two can be
 separate as is done for the compressed images.

 If you are using a two-disk boot/root setup (kernel on disk 1, ramdisk
 image on disk 2) then the ramdisk would start at block zero, and an
 offset of zero would be used. Since this is the default value, you
 would not need to actually use the command at all.


 3.2.2.  The `load_ramdisk=' Argument

 This parameter tells the kernel whether it is to try to load a ramdisk
 image or not. Specifying `load_ramdisk=1' will tell the kernel to load
 a floppy into the ramdisk. The default value is zero, meaning that the
 kernel should not try to load a ramdisk.

 Please see the file linux/Documentation/ramdisk.txt for a complete
 description of the new boot time arguments, and how to use them. A
 description of how this parameter can be set and stored in the kernel
 image via `rdev' is also described.


 3.2.3.  The `prompt_ramdisk=' Argument

 This parameter tells the kernel whether or not to give you a prompt
 asking you to insert the floppy containing the ramdisk image. In a
 single floppy configuration the ramdisk image is on the same floppy as
 the kernel that just finished loading/booting and so a prompt is not
 needed. In this case one can use `prompt_ramdisk=0'. In a two floppy
 configuration, you will need the chance to switch disks, and thus
 `prompt_ramdisk=1' can be used. Since this is the default value, it
 doesn't really need to be specified. ( (Historical note: Sneaky people
 used to use the `vga=ask' LILO option to temporarily pause the boot
 process and allow a chance to switch from boot to root floppy.)

 Please see the file linux/Documentation/ramdisk.txt for a complete
 description of the new boot time arguments, and how to use them. A
 description of how this parameter can be set and stored in the kernel
 image via `rdev' is also described.


 3.2.4.  The `ramdisk_size=' Argument

 While it is true that the ramdisk grows dynamically as required, there
 is an upper bound on its size so that it doesn't consume all available
 RAM and leave you in a mess. The default is 4096 (i.e. 4MB) which
 should be large enough for most needs. You can override the default to
 a bigger or smaller size with this boot argument.

 Please see the file linux/Documentation/ramdisk.txt for a complete
 description of the new boot time arguments, and how to use them. A
 description of how this parameter can be set and stored in the kernel
 image via `rdev' is also described.


 3.2.5.  The `ramdisk=' Argument (obsolete)

 (NOTE: This argument is obsolete, and should not be used except on
 kernels v1.3.47 and older. The commands that should be used for the
 ramdisk device are documented above.)

 This specifies the size in kB of the RAM disk device.  For example, if
 one wished to have a root filesystem on a 1.44MB floppy loaded into
 the RAM disk device, they would use:


 ______________________________________________________________________
         ramdisk=1440
 ______________________________________________________________________



 This is one of the few kernel boot arguments that has its default
 stored in the kernel image, and which can thus be altered with the
 rdev utility.



 3.2.6.  The `noinitrd' (initial RAM disk) Argument

 The v2.x and newer kernels have a feature where the root filesystem
 can be initially a RAM disk, and the kernel executes /linuxrc on that
 RAM image. This feature is typically used to allow loading of modules
 needed to mount the real root filesystem (e.g. load the SCSI driver
 modules stored in the RAM disk image, and then mount the real root
 filesystem on a SCSI disk.)

 The actual `noinitrd' argument determines what happens to the initrd
 data after the kernel has booted.  When specified, instead of
 converting it to a RAM disk, it is accessible via /dev/initrd, which
 can be read once before the RAM is released back to the system. For
 full details on using the initial RAM disk, please consult
 linux/Documentation/initrd.txt. In addition, the most recent versions
 of LILO and LOADLIN should have additional useful information.


 3.3.  Boot Arguments Related to Memory Handling

 The following arguments alter how Linux detects or handles the
 physical and virtual memory of your system.


 3.3.1.  The `mem=' Argument

 This argument has two purposes: The original purpose was to specify
 the amount of installed memory (or a value less than that if you
 wanted to limit the amount of memory available to linux). The second
 (and hardly used) purpose is to specify mem=nopentium which tells the
 Linux kernel to not use the 4MB page table performance feature.

 The original BIOS call defined in the PC specification  that returns
 the amount of installed memory was only designed to be able to report
 up to 64MB. (Yes, another lack of foresight, just like the 1024
 cylinder disks... sigh.) Linux uses this BIOS call at boot to
 determine how much memory is installed.  If you have more than 64MB of
 RAM installed, you can use this boot argument to tell Linux how much
 memory you have.  Here is a quote from Linus on the usage of the mem=
 parameter.

 ``The kernel will accept any `mem=xx' parameter you give it, and if it
 turns out that you lied to it, it will crash horribly sooner or later.
 The parameter indicates the highest addressable RAM address, so
 `mem=0x1000000' means you have 16MB of memory, for example.  For a
 96MB machine this would be `mem=0x6000000'.  If you tell Linux that it
 has more memory than it actually does have, bad things will happen:
 maybe not at once, but surely eventually.''

 Note that the argument does not have to be in hex, and the suffixes
 `k' and `M' (case insensitive) can be used to specify kilobytes and
 Megabytes, respectively. (A `k' will cause a 10 bit shift on your
 value, and a `M' will cause a 20 bit shift.)  A typical example for a
 128MB machine would be "mem=128m".


 3.3.2.  The `swap=' Argument

 This allows the user to tune some of the virtual memory (VM)
 parameters that are related to swapping to disk. It accepts the
 following eight parameters:


 ______________________________________________________________________
         MAX_PAGE_AGE
         PAGE_ADVANCE
         PAGE_DECLINE
         PAGE_INITIAL_AGE
         AGE_CLUSTER_FRACT
         AGE_CLUSTER_MIN
         PAGEOUT_WEIGHT
         BUFFEROUT_WEIGHT
 ______________________________________________________________________



 Interested hackers are advised to have a read of linux/mm/swap.c and
 also make note of the goodies in /proc/sys/vm. Kernels come with some
 useful documentation on this in the linux/Documentation/vm/ directory.


 3.3.3.  The `buff=' Argument

 Similar to the `swap=' argument, this allows the user to tune some of
 the parameters related to buffer memory management.  It accepts the
 following six parameters:


 ______________________________________________________________________
         MAX_BUFF_AGE
         BUFF_ADVANCE
         BUFF_DECLINE
         BUFF_INITIAL_AGE
         BUFFEROUT_WEIGHT
         BUFFERMEM_GRACE
 ______________________________________________________________________



 Interested hackers are advised to have a read of linux/mm/swap.c and
 also make note of the goodies in /proc/sys/vm.  Kernels come with some
 useful documentation on this in the linux/Documentation/vm/ directory.


 3.4.  Boot Arguments for NFS Root Filesystem

 Linux supports systems such as diskless workstations via having their
 root filesystem as NFS (Network FileSystem).  These arguments are used
 to tell the diskless workstation which machine it is to get its system
 from. Also note that the argument root=/dev/nfs is required. Detailed
 information on using an NFS root fs is in the file
 linux/Documentation/nfsroot.txt. You should read that file, as the
 following is only a quick summary taken directly from that file.


 3.4.1.  The `nfsroot=' Argument

 This argument tells the kernel which machine, what directory and what
 NFS options to use for the root filesystem. The form of the argument
 is as follows:


 ______________________________________________________________________
         nfsroot=[<server-ip>:]<root-dir>[,<nfs-options>]
 ______________________________________________________________________



 If the nfsroot parameter is not given on the command line, the default
 `/tftpboot/%s' will be used. The other options are as follows:

 <server-ip> -- Specifies the IP address of the NFS server. If this
 field is not given, the default address as determined by the nfsaddrs
 variable (see below) is used. One use of this parameter is for example
 to allow using different servers for RARP and NFS. Usually you can
 leave this blank.

 <root-dir> -- Name of the directory on the server to mount as root. If
 there is a `%s' token in the string, the token will be replaced by the
 ASCII-representation of the client's IP address.

 <nfs-options> -- Standard NFS options. All options are separated by
 commas.  If the options field is not given, the following defaults
 will be used:


         port            = as given by server portmap daemon
         rsize           = 1024
         wsize           = 1024
         timeo           = 7
         retrans         = 3
         acregmin        = 3
         acregmax        = 60
         acdirmin        = 30
         acdirmax        = 60
         flags           = hard, nointr, noposix, cto, ac



 3.4.2.  The `nfsaddrs=' Argument

 This boot argument sets up the various network interface addresses
 that are required to communicate over the network. If this argument is
 not given, then the kernel tries to use RARP and/or BOOTP to figure
 out these parameters. The form is as follows:


 ______________________________________________________________________
         nfsaddrs=<my-ip>:<serv-ip>:<gw-ip>:<netmask>:<name>:<dev>:<auto>
 ______________________________________________________________________



 <my-ip> -- IP address of the client. If empty, the address will either
 be determined by RARP or BOOTP. What protocol is used de- pends on
 what has been enabled during kernel configuration and on the <auto>
 parameter. If this parameter is not empty, neither RARP nor BOOTP will
 be used.

 <serv-ip> -- IP address of the NFS server. If RARP is used to
 determine the client address and this parameter is NOT empty only
 replies from the specified server are accepted. To use different RARP
 and NFS server, specify your RARP server here (or leave it blank), and
 specify your NFS server in the nfsroot parameter (see above). If this
 entry is blank the address of the server is used which answered the
 RARP or BOOTP request.

 <gw-ip> -- IP address of a gateway if the server in on a different
 subnet. If this entry is empty no gateway is used and the server is
 assumed to be on the local network, unless a value has been received
 by BOOTP.

 <netmask> -- Netmask for local network interface. If this is empty,
 the netmask is derived from the client IP address, unless a value has
 been received by BOOTP.

 <name> -- Name of the client. If empty, the client IP address is used
 in ASCII-notation, or the value received by BOOTP.

 <dev> -- Name of network device to use. If this is empty, all devices
 are used for RARP requests, and the first one found for BOOTP. For NFS
 the device is used on which either RARP or BOOTP replies have been
 received. If you only have one device you can safely leave this blank.

 <auto> -- Method to use for autoconfiguration. If this is either
 `rarp' or `bootp' the specified protocol is being used.  If the value
 is `both' or empty, both protocols are used so far as they have been
 enabled during kernel configuration Using 'none' means no
 autoconfiguration. In this case you have to specify all necessary
 values in the fields before.

 The <auto> parameter can appear alone as the value to the nfsaddrs
 parameter (without all the `:' characters before) in which case
 autoconfiguration is used. However, the `none' value is not available
 in that case.


 3.5.  Other Misc. Kernel Boot Arguments

 These various boot arguments let the user tune certain internal kernel
 parameters.


 3.5.1.  The `debug' Argument

 The kernel communicates important (and not-so important) messages to
 the operator via the printk() function.  If the message is considered
 important, the printk() function will put a copy on the present
 console as well as handing it off to the klogd() facility so that it
 gets logged to disk. The reason for printing important messages to the
 console as well as logging them to disk is because under unfortunate
 circumstances (e.g. a disk failure) the message won't make it to disk
 and will be lost.

 The threshold for what is and what isn't considered important is set
 by the console_loglevel variable. The default is to log anything more
 important than DEBUG (level 7) to the console. (These levels are
 defined in the include file kernel.h) Specifying debug as a boot
 argument will set the console loglevel to 10, so that all kernel
 messages appear on the console.

 The console loglevel can usually also be set at run time via an option
 to the klogd() program. Check the man page for the version installed
 on your system to see how to do this.


 3.5.2.  The `init=' Argument

 The kernel defaults to starting the `init' program at boot, which then
 takes care of setting up the computer for users via launching getty
 programs, running `rc' scripts and the like.  The kernel first looks
 for /sbin/init, then /etc/init (depreciated), and as a last resort, it
 will try to use /bin/sh (possibly on /etc/rc).  If for example, your
 init program got corrupted and thus stopped you from being able to
 boot, you could simply use the boot prompt init=/bin/sh which would
 drop you directly into a shell at boot, allowing you to replace the
 corrupted program.


 3.5.3.  The `kbd-reset' Argument

 Normally on i386 based machines, the Linux kernel does not reset the
 keyboard controller at boot, since the BIOS is supposed to do this.
 But as usual, not all machines do what they should. Supplying this
 option may help if you are having problems with your keyboard
 behaviour.  It simply forces a reset at initialization time. (Some
 have argued that this should be the default behaviour anyways).


 3.5.4.  The `maxcpus=' Argument

 The number given with this argument limits the maximum number of CPUs
 activated in SMP mode.  Using a value of 0 is equivalent to the nosmp
 option.


 3.5.5.  The `mca-pentium' Argument

 The IBM model 95 Microchannel machines seem to lock up on the test
 that Linux usually does to detect the type of math chip coupling.
 Since all Pentium chips have a built in math processor, this test (and
 the lock up problem) can be avoided by using this boot option.


 3.5.6.  The `md=' Argument

 If your root filesystem is on a Multiple Device then you can use this
 (assuming you compiled in boot support) to tell the kernel the
 multiple device layout. The format (from the file
 linux/Documentation/md.txt) is:

 md=md_device_num,raid_level,chunk_size_factor,fault_level,dev0,dev1,...,devN

 Where md_device_num is the number of the md device, i.e. 0 means md0,
 1 means md1, etc.  For raid_level, use -1 for linear mode and 0 for
 striped mode.  Other modes are currently unsupported.  The
 chunk_size_factor is for  raid-0 and raid-1 only and sets the chunk
 size as PAGE_SIZE shifted left the specified amount.  The fault_level
 is only for raid-1 and sets the maximum fault number to the specified
 number.  (Currently unsupported due to lack of boot support for
 raid1.)  The dev0-devN are a commaseparated list of the devices that
 make up the individual md device: e.g. /dev/hda1,/dev/hdc1,/dev/sda1



 3.5.7.  The `no387' Argument

 Some i387 coprocessor chips have bugs that show up when used in 32 bit
 protected mode. For example, some of the early ULSI-387 chips would
 cause solid lockups while performing floating point calculations,
 apparently due to a bug in the FRSAV/FRRESTOR instructions.  Using the
 `no387' boot argument causes Linux to ignore the math coprocessor even
 if you have one. Of course you must then have your kernel compiled
 with math emulation support! This may also be useful if you have one
 of those really old 386 machines that could use an 80287 FPU, as Linux
 can't use an 80287.


 3.5.8.  The `no-hlt' Argument

 The i386 (and successors thereof) family of CPUs have a `hlt'
 instruction which tells the CPU that nothing is going to happen until
 an external device (keyboard, modem, disk, etc.) calls upon the CPU to
 do a task. This allows the CPU to enter a `low-power' mode where it
 sits like a zombie until an external device wakes it up (usually via
 an interrupt).  Some of the early i486DX-100 chips had a problem  with
 the `hlt' instruction, in that they couldn't reliably return to
 operating mode after this instruction was used. Using the `no-hlt'
 instruction tells Linux to just run an infinite loop when there is
 nothing else to do, and to not halt your CPU when there is no
 activity. This allows people with these broken chips to use Linux,
 although they would be well advised to seek a replacement through a
 warranty where possible.


 3.5.9.  The `no-scroll' Argument

 Using this argument at boot disables scrolling features that make it
 difficult to use Braille terminals.


 3.5.10.  The `noapic' Argument

 Using this option tells a SMP kernel to not use some of the advanced
 features of the interrupt controller on multi processor machines. See
 linux/Documentation/IO-APIC.txt for more information.


 3.5.11.  The `nosmp' Argument

 Use of this option will tell a SMP kernel on a SMP machine to operate
 single processor.  Typically only used for debugging and determining
 if a particular problem is SMP related.


 3.5.12.  The `panic=' Argument

 In the unlikely event of a kernel panic (i.e. an internal error that
 has been detected by the kernel, and which the kernel decides is
 serious enough to moan loudly and then halt everything), the default
 behaviour is to just sit there until someone comes along and notices
 the panic message on the screen and reboots the machine.  However if a
 machine is running unattended in an isolated location it may be
 desirable for it to automatically reset itself so that the machine
 comes back on line. For example, using panic=30 at boot would cause
 the kernel to try and reboot itself 30 seconds after the kernel panic
 happened. A value of zero gives the default behaviour, which is to
 wait forever.

 Note that this timeout value can also be read and set via the
 /proc/sys/kernel/panic sysctl interface.
 3.5.13.  The `pci=' Argument



 3.5.14.  The `pirq=' Argument

 Using this option tells a SMP kernel information on the PCI slot
 versus IRQ settings for SMP motherboards which are unknown (or known
 to be blacklisted).  See linux/Documentation/IO-APIC.txt for more
 information.


 3.5.15.  The `profile=' Argument

 Kernel developers can enable an option that allows them to profile how
 and where the kernel is spending its CPU cycles in an effort to
 maximize efficiency and performance. This option lets you set the
 profile shift count at boot. Typically it is set to two. You can also
 compile your kernel with profiling enabled by default. In either case,
 you need a tool such as readprofile.c that can make use of the
 /proc/profile output.


 3.5.16.  The `reboot=' Argument

 This option controls the type of reboot that Linux will do when it
 resets the computer (typically via /sbin/init handling a Control-Alt-
 Delete). The default as of v2.0 kernels is to do a `cold' reboot (i.e.
 full reset, BIOS does memory check, etc.) instead of a `warm' reboot
 (i.e. no full reset, no memory check). It was changed to be cold by
 default since that tends to work on cheap/broken hardware that fails
 to reboot when a warm reboot is requested. To get the old behaviour
 (i.e. warm reboots) use reboot=w or in fact any word that starts with
 w will work.

 Why would you bother? Some disk controllers with cache memory on board
 can sense a warm reboot, and flush any cached data to disk. Upon a
 cold boot, the card may be reset and the write-back data in your cache
 card's memory is lost. Others have reported systems that take a long
 time to go through the memory check, and/or SCSI BIOSes that take
 longer to initialize on a cold boot as a good reason to use warm
 reboots.


 3.5.17.  The `reserve=' Argument

 This is used to protect I/O port regions from probes.  The form of the
 command is:


      reserve=iobase,extent[,iobase,extent]...


 In some machines it may be necessary to prevent device drivers from
 checking for devices (auto-probing) in a specific region. This may be
 because of poorly designed hardware that causes the boot to freeze
 (such as some ethercards), hardware that is mistakenly identified,
 hardware whose state is changed by an earlier probe, or merely
 hardware you don't want the kernel to initialize.

 The reserve boot-time argument addresses this problem by specifying an
 I/O port region that shouldn't be probed. That region is reserved in
 the kernel's port registration table as if a device has already been
 found in that region (with the name reserved).  Note that this
 mechanism shouldn't be necessary on most machines.  Only when there is
 a problem or special case would it be necessary to use this.
 The I/O ports in the specified region are protected against device
 probes that do a check_region() prior to probing blindly into a region
 of I/O space. This was put in to be used when some driver was hanging
 on a NE2000, or misidentifying some other device as its own.  A
 correct device driver shouldn't probe a reserved region, unless
 another boot argument explicitly specifies that it do so.  This
 implies that reserve will most often be used with some other boot
 argument. Hence if you specify a reserve region to protect a specific
 device, you must generally specify an explicit probe for that device.
 Most drivers ignore the port registration table if they are given an
 explicit address.

 For example, the boot line


 ______________________________________________________________________
         reserve=0x300,32  blah=0x300
 ______________________________________________________________________



 keeps all device drivers except the driver for `blah' from probing
 0x300-0x31f.

 As usual with boot-time specifiers there is an 11 parameter limit,
 thus you can only specify 5 reserved regions per reserve keyword.
 Multiple reserve specifiers will work if you have an unusually
 complicated request.



 3.5.18.  The `vga=' Argument

 Note that this is not really a boot argument. It is an option that is
 interpreted by LILO and not by the kernel like all the other boot
 arguments are. However its use has become so common that it deserves a
 mention here. It can also be set via using rdev -v or equivalently
 vidmode on the vmlinuz file.  This allows the setup code to use the
 video BIOS to change the default display mode before actually booting
 the Linux kernel. Typical modes are 80x50, 132x44 and so on. The best
 way to use this option is to start with vga=ask which will prompt you
 with a list of various modes that you can use with your video adapter
 before booting the kernel. Once you have the number from the above
 list that you want to use, you can later put it in place of the `ask'.
 For more information, please see the file linux/Documentation/svga.txt
 that comes with all recent kernel versions.

 Note that newer kernels (v2.1 and up) have the setup code that changes
 the video mode as an option, listed as Video mode selection support so
 you need to enable this option if you want to use this feature.


 4.  Boot Arguments to Control PCI Bus Behaviour (`pci=')

 The `pci=' argument (not avail. in v2.0 kernels) can be used to change
 the behaviour of PCI bus device probing and device behaviour. Firstly
 the file linux/drivers/pci/pci.c checks for architecture independent
 pci= options.  The remaining allowed arguments are handled in
 linux/arch/???/kernel/bios32.c and are listed below for ???=i386.


 4.1.  The `pci=bios' and `pci=nobios' Arguments

 These are used to set/clear the flag indicating that the PCI probing
 is to take place via the PCI BIOS.  The default is to use the BIOS.

 4.2.  The `pci=conf1' and `pci=conf2' Arguments

 If PCI direct mode is enabled, the use of these enables either
 configuration Type 1 or Type 2.  These implicitly clear the PCI BIOS
 probe flag (i.e. `pci=nobios') too.


 4.3.  The `pci=io=' Argument

 If you get a message like PCI: Unassigned IO space for.../ then you
 may need to supply an I/O value with this option.  From the source:

 ``Several BIOS'es forget to assign addresses to I/O ranges.  We try to
 fix it here, expecting there are free addresses starting with 0x5800.
 Ugly, but until we come with better resource management, it's the only
 simple solution.''


 4.4.  The `pci=nopeer' Argument

 This disables the default peer bridge fixup, which according to the
 source does the following:

 ``In case there are peer host bridges, scan bus behind each of them.
 Although several sources claim that the host bridges should have
 header type 1 and be assigned a bus number as for PCI2PCI bridges, the
 reality doesn't pass this test and the bus number is usually set by
 BIOS to the first free value.''


 4.5.  The `pci=nosort' Argument

 Using this argument instructs the kernel to not sort the PCI devices
 during the probing phase.


 4.6.  The `pci=off' Argument

 Using this option disables all PCI bus probing. Any device drivers
 that make use of PCI functions to find and initialize hardware will
 most likely fail to work.


 4.7.  The `pci=reverse' Argument

 This option will reverse the ordering of the PCI devices on that PCI
 bus.


 5.  Boot Arguments for Video Frame Buffer Drivers

 The `video=' argument (not avail. in v2.0 kernels) is used when the
 frame buffer device abstraction layer is built into the kernel. If
 that sounds complicated, well it isn't really too bad.  It basically
 means that instead of having a different video program (the X11R6
 server) for each brand of video card (e.g. XF86_S3, XF86_SVGA, ...),
 the kernel would have a built in driver available for each video card
 and export a single interface for the video program so that only one
 X11R6 server (XF86_FBDev) would be required.  This is similar to how
 networking is now - the kernel has drivers available for each brand of
 network card and exports a single network interface so that just one
 version of a network program (like Netscape) will work for all
 systems, regardless of the underlying brand of network card.

 The typical format of this argument is video=name:option1,option2,...
 where name is the name of a generic option or of a frame buffer
 driver.  The video= option is passed from linux/init/main.c into
 linux/drivers/video/fbmem.c for further processing.  Here it is
 checked for some generic options before trying to match to a known
 driver name. Once a driver name match is made, the comma separated
 option list is then passed into that particular driver for final
 processing. The list of valid driver names can be found by reading
 down the fb_drivers array in the file fbmem.c mentioned above.

 Information on the options that each driver supports will eventually
 be found in linux/Documentation/fb/ but currently (v2.2) only a few
 are described there.  Unfortunately the number of video drivers and
 the number of options for each one is content for another document
 itself and hence too much to list here.

 If there is no Documentation file for your card, you will have to get
 the option information directly from the driver. Go to
 linux/drivers/video/ and look in the appropriate ???fb.c file (the ???
 will be based on the card name).  In there, search for a function with
 _setup in its name and you should see what options the driver tries to
 match, such as font or mode or...


 5.1.  The `video=map:...' Argument

 This option is used to set/override the console to frame buffer device
 mapping. A comma separated list of numbers sets the mapping, with the
 value of option N taken to be the frame buffer device number for
 console N.


 5.2.  The `video=scrollback:...' Argument

 A number after the colon will set the size of memory allocated for the
 scrollback buffer. (Use Shift and Page Up or Page Down keys to
 scroll.)  A suffix of `k' or `K' after the number will indicate that
 the number is to be interpreted as kilobytes instead of bytes.


 5.3.  The `video=vc:...' Argument


 A number, or a range of numbers (e.g. video=vc:2-5) will specify the
 first, or the first and last frame buffer virtual console(s). The use
 of this option also has the effect of setting the frame buffer console
 to not be the default console.


 6.  Boot Arguments for SCSI Peripherals.

 This section contains the descriptions of the boot args that are used
 for passing information about the installed SCSI host adapters, and
 SCSI devices.


 6.1.  Arguments for Mid-level Drivers

 The mid level drivers handle things like disks, CD-ROMs and tapes
 without getting into host adapter specifics.


 6.1.1.  Maximum Probed LUNs (`max_scsi_luns=')

 Each SCSI device can have a number of `sub-devices' contained within
 itself. The most common example is any of the SCSI CD-ROMs that handle
 more than one disk at a time.  Each CD is addressed as a `Logical Unit
 Number' (LUN) of that particular device. But most devices, such as
 hard disks, tape drives and such are only one device, and will be
 assigned to LUN zero.

 The problem arises with single LUN devices with bad firmware.  Some
 poorly designed SCSI devices (old and unfortunately new) can not
 handle being probed for LUNs not equal to zero. They will respond by
 locking up, and possibly taking the whole SCSI bus down with them.

 The kernel has a configuration option that allows you to set the
 maximum number of probed LUNs. The default is to only probe LUN zero,
 to avoid the problem described above.

 To specify the number of probed LUNs at boot, one enters
 `max_scsi_luns=n' as a boot arg, where n is a number between one and
 eight. To avoid problems as described above, one would use n=1 to
 avoid upsetting such broken devices


 6.1.2.  SCSI Logging (`scsi_logging=')

 Supplying a non-zero value to this boot argument turns on logging of
 all SCSI events (error, scan, mlqueue, mlcomplete, llqueue,
 llcomplete, hlqueue, hlcomplete).  Note that better control of which
 events are logged can be obtained via the /proc/scsi/scsi interface if
 you aren't interested in the events that take place at boot before the
 /proc/ filesystem is accessible.


 6.1.3.  Parameters for the SCSI Tape Driver (`st=')

 Some boot time configuration of the SCSI tape driver can be achieved
 by using the following:


 ______________________________________________________________________
         st=buf_size[,write_threshold[,max_bufs]]
 ______________________________________________________________________



 The first two numbers are specified in units of kB.  The default
 buf_size is 32kB, and the maximum size that can be specified is a
 ridiculous 16384kB.  The write_threshold is the value at which the
 buffer is committed to tape, with a default value of 30kB.  The
 maximum number of buffers varies with the number of drives detected,
 and has a default of two. An example usage would be:


 ______________________________________________________________________
         st=32,30,2
 ______________________________________________________________________



 Full details can be found in the README.st file that is in the scsi
 directory of the kernel source tree.


 6.2.  Arguments for SCSI Host Adapters

 General notation for this section:

 iobase -- the first I/O port that the SCSI host occupies.  These are
 specified in hexidecimal notation, and usually lie in the range from
 0x200 to 0x3ff.

 irq -- the hardware interrupt that the card is configured to use.
 Valid values will be dependent on the card in question, but will
 usually be 5, 7, 9, 10, 11, 12, and 15. The other values are usually
 used for common peripherals like IDE hard disks, floppies, serial
 ports, etc.

 dma -- the DMA (Direct Memory Access) channel that the card uses.
 Typically only applies to bus-mastering cards.  PCI and VLB cards are
 native bus-masters, and do not require and ISA DMA channel.

 scsi-id -- the ID that the host adapter uses to identify itself on the
 SCSI bus. Only some host adapters allow you to change this value, as
 most have it permanently specified internally. The usual default value
 is seven, but the Seagate and Future Domain TMC-950 boards use six.

 parity -- whether the SCSI host adapter expects the attached devices
 to supply a parity value with all information exchanges.  Specifying a
 one indicates parity checking is enabled, and a zero disables parity
 checking. Again, not all adapters will support selection of parity
 behaviour as a boot argument.


 6.2.1.  Adaptec aha151x, aha152x, aic6260, aic6360, SB16-SCSI
 (`aha152x=')

 The aha numbers refer to cards and the aic numbers refer to the actual
 SCSI chip on these type of cards, including the Soundblaster-16 SCSI.

 The probe code for these SCSI hosts looks for an installed BIOS, and
 if none is present, the probe will not find your card. Then you will
 have to use a boot argument of the form:


 ______________________________________________________________________
          aha152x=iobase[,irq[,scsi-id[,reconnect[,parity]]]]
 ______________________________________________________________________



 Note that if the driver was compiled with debugging enabled, a sixth
 value can be specified to set the debug level.

 All the parameters are as described at the top of this section, and
 the reconnect value will allow device disconnect/reconnect if a non-
 zero value is used. An example usage is as follows:


 ______________________________________________________________________
         aha152x=0x340,11,7,1
 ______________________________________________________________________



 Note that the parameters must be specified in order, meaning that if
 you want to specify a parity setting, then you will have to specify an
 iobase, irq, scsi-id and reconnect value as well.


 6.2.2.  Adaptec aha154x (`aha1542=')

 These are the aha154x series cards. The aha1542 series cards have an
 i82077 floppy controller onboard, while the aha1540 series cards do
 not. These are busmastering cards, and have parameters to set the
 ``fairness'' that is used to share the bus with other devices. The
 boot argument looks like the following.

 ______________________________________________________________________
         aha1542=iobase[,buson,busoff[,dmaspeed]]
 ______________________________________________________________________



 Valid iobase values are usually one of: 0x130, 0x134, 0x230, 0x234,
 0x330, 0x334.  Clone cards may permit other values.

 The buson, busoff values refer to the number of microseconds that the
 card dominates the ISA bus. The defaults are 11us on, and 4us off, so
 that other cards (such as an ISA LANCE Ethernet card) have a chance to
 get access to the ISA bus.

 The dmaspeed value refers to the rate (in MB/s) at which the DMA
 (Direct Memory Access) transfers proceed at. The default is 5MB/s.
 Newer revision cards allow you to select this value as part of the
 soft-configuration, older cards use jumpers. You can use values up to
 10MB/s assuming that your motherboard is capable of handling it.
 Experiment with caution if using values over 5MB/s.


 6.2.3.  Adaptec aha274x, aha284x, aic7xxx (`aic7xxx=')

 These boards can accept an argument of the form:


 ______________________________________________________________________
         aic7xxx=extended,no_reset
 ______________________________________________________________________



 The extended value, if non-zero, indicates that extended translation
 for large disks is enabled. The no_reset value, if non-zero, tells the
 driver not to reset the SCSI bus when setting up the host adaptor at
 boot.


 6.2.4.  AdvanSys SCSI Host Adaptors (`advansys=')

 The AdvanSys driver can accept up to four i/o addresses that will be
 probed for an AdvanSys SCSI card. Note that these values (if used) do
 not effect EISA or PCI probing in any way.  They are only used for
 probing ISA and VLB cards.  In addition, if the driver has been
 compiled with debugging enabled, the level of debugging output can be
 set by adding an 0xdeb[0-f] parameter. The 0-f allows setting the
 level of the debugging messages to any of 16 levels of verbosity.


 6.2.5.  Always IN2000 Host Adaptor (`in2000=')

 Unlike other SCSI host boot arguments, the IN2000 driver uses ASCII
 string prefixes for most of its integer arguments. Here is a list of
 the supported arguments:

 ioport:addr -- Where addr is IO address of a (usually ROM-less) card.

 noreset -- No optional args. Prevents SCSI bus reset at boot time.

 nosync:x -- x is a bitmask where the 1st 7 bits correspond with the 7
 possible SCSI devices (bit 0 for device #0, etc).  Set a bit to
 PREVENT sync negotiation on that device.  The driver default is sync
 DISABLED on all devices.


 period:ns -- ns is the minimum # of nanoseconds in a SCSI data
 transfer period. Default is 500; acceptable values are 250 to 1000.

 disconnect:x -- x = 0 to never allow disconnects, 2 to always allow
 them.  x = 1 does 'adaptive' disconnects, which is the default and
 generally the best choice.

 debug:x If `DEBUGGING_ON' is defined, x is a bitmask that causes
 various types of debug output to printed - see the DB_xxx defines in
 in2000.h

 proc:x -- If `PROC_INTERFACE' is defined, x is a bitmask that
 determines how the /proc interface works and what it does - see the
 PR_xxx defines in in2000.h


 Some example usages are listed below:


 ______________________________________________________________________
         in2000=ioport:0x220,noreset
         in2000=period:250,disconnect:2,nosync:0x03
         in2000=debug:0x1e
         in2000=proc:3
 ______________________________________________________________________



 6.2.6.  AMD AM53C974 based hardware (`AM53C974=')

 Unlike other drivers, this one does not use boot parameters to
 communicate i/o, IRQ or DMA channels. (Since the AM53C974 is a PCI
 device, there shouldn't be a need to do so.)  Instead, the parameters
 are used to communicate the transfer modes and rates that are to be
 used between the host and the target device. This is best described
 with an example:


 ______________________________________________________________________
         AM53C974=7,2,8,15
 ______________________________________________________________________



 This would be interpreted as follows: `For communication between the
 controller with SCSI-ID 7 and the device with SCSI-ID 2, a transfer
 rate of 8MHz in synchronous mode with max. 15 bytes offset should be
 negotiated.' More details can be found in the file
 linux/drivers/scsi/README.AM53C974


 6.2.7.  BusLogic SCSI Hosts with v1.2 kernels (`buslogic=')

 In older kernels, the buslogic driver accepts only one parameter, that
 being the I/O base. It expects that to be one of the following valid
 values: 0x130, 0x134, 0x230, 0x234, 0x330, 0x334.


 6.2.8.  BusLogic SCSI Hosts with v2.x kernels (`BusLogic=')

 With v2.x kernels, the BusLogic driver accepts many parameters.  (Note
 the case in the above; upper case B and L!!!).  There are simply too
 many to list here.  A complete description is tucked away in the
 middle of the driver linux/drivers/scsi/BusLogic.c and searching for
 the string BusLogic= will put you right on it.


 6.2.9.  EATA SCSI Cards (`eata=')

 As of late v2.0 kernels, the EATA drivers will accept a boot argument
 to specify the i/o base(s) to be probed. It is of the form:


 ______________________________________________________________________
         eata=iobase1[,iobase2][,iobase3]...[,iobaseN]
 ______________________________________________________________________



 The driver will probe the addresses in the order that they are listed.



 6.2.10.  Future Domain TMC-8xx, TMC-950 (`tmc8xx=')

 The probe code for these SCSI hosts looks for an installed BIOS, and
 if none is present, the probe will not find your card. Or, if the
 signature string of your BIOS is not recognized then it will also not
 be found. In either case, you will then have to use a boot argument of
 the form:


 ______________________________________________________________________
         tmc8xx=mem_base,irq
 ______________________________________________________________________



 The mem_base value is the value of the memory mapped I/O region that
 the card uses. This will usually be one of the following values:
 0xc8000, 0xca000, 0xcc000, 0xce000, 0xdc000, 0xde000.


 6.2.11.  Future Domain TMC-16xx, TMC-3260, AHA-2920 (`fdomain=')

 The driver detects these cards according to a list of known BIOS ROM
 signatures. For a full list of known BIOS revisions, please see
 linux/drivers/scsi/fdomain.c as it has a lot of information at the top
 of that file. If your BIOS is not known to the driver, you can use an
 override of the form:


 ______________________________________________________________________
         fdomain=iobase,irq[,scsi_id]
 ______________________________________________________________________



 6.2.12.  IOMEGA Parallel Port / ZIP drive (`ppa=')

 This driver is for the IOMEGA Parallel Port SCSI adapter which is
 embedded into the IOMEGA ZIP drives. It may also work with the
 original IOMEGA PPA3 device. The boot argument for this driver is of
 the form:



 ______________________________________________________________________
         ppa=iobase,speed_high,speed_low,nybble
 ______________________________________________________________________



 with all but iobase being optionally specified values. If you wish to
 alter any of the three optional parameters, you are advised to read
 linux/drivers/scsi/README.ppa for details of what they control.


 6.2.13.  NCR5380 based controllers (`ncr5380=')

 Depending on your board, the 5380 can be either i/o mapped or memory
 mapped. (An address below 0x400 usually implies i/o mapping, but PCI
 and EISA hardware use i/o addresses above 0x3ff.) In either case, you
 specify the address, the IRQ value and the DMA channel value. An
 example for an i/o mapped card would be: ncr5380=0x350,5,3. If the
 card doesn't use interrupts, then an IRQ value of 255 (0xff) will
 disable interrupts. An IRQ value of 254 means to autoprobe. More
 details can be found in the file linux/drivers/scsi/README.g_NCR5380


 6.2.14.  NCR53c400 based controllers (`ncr53c400=')

 The generic 53c400 support is done with the same driver as the generic
 5380 support mentioned above. The boot argument is identical to the
 above with the exception that no DMA channel is used by the 53c400.


 6.2.15.  NCR53c406a based controllers (`ncr53c406a=')

 This driver uses a boot argument of the form:


 ______________________________________________________________________
         ncr53c406a=PORTBASE,IRQ,FASTPIO
 ______________________________________________________________________



 where the IRQ and FASTPIO parameters are optional. An interrupt value
 of zero disables the use of interrupts. Using a value of one for the
 FASTPIO parameter enables the use of insl and outsl instructions
 instead of the single-byte inb and outb instructions. The driver can
 also use DMA as a compile-time option.


 6.2.16.  Pro Audio Spectrum (`pas16=')

 The PAS16 uses a NCR5380 SCSI chip, and newer models support jumper-
 less configuration. The boot argument is of the form:


 ______________________________________________________________________
         pas16=iobase,irq
 ______________________________________________________________________



 The only difference is that you can specify an IRQ value of 255, which
 will tell the driver to work without using interrupts, albeit at a
 performance loss. The iobase is usually 0x388.



 6.2.17.  Seagate ST-0x (`st0x=')

 The probe code for these SCSI hosts looks for an installed BIOS, and
 if none is present, the probe will not find your card. Or, if the
 signature string of your BIOS is not recognized then it will also not
 be found. In either case, you will then have to use a boot argument of
 the form:


 ______________________________________________________________________
         st0x=mem_base,irq
 ______________________________________________________________________



 The mem_base value is the value of the memory mapped I/O region that
 the card uses. This will usually be one of the following values:
 0xc8000, 0xca000, 0xcc000, 0xce000, 0xdc000, 0xde000.


 6.2.18.  Trantor T128 (`t128=')

 These cards are also based on the NCR5380 chip, and accept the
 following options:


 ______________________________________________________________________
         t128=mem_base,irq
 ______________________________________________________________________



 The valid values for mem_base are as follows: 0xcc000, 0xc8000,
 0xdc000, 0xd8000.


 6.2.19.  Ultrastor SCSI cards (`u14-34f=')

 Note that there appears to be two independent drivers for this card,
 namely CONFIG_SCSI_U14_34F that uses u14-34f.c and
 CONFIG_SCSI_ULTRASTOR that uses ultrastor.c. It is the u14-34f one
 that (as of late v2.0 kernels) accepts a boot argument of the form:


 ______________________________________________________________________
         u14-34f=iobase1[,iobase2][,iobase3]...[,iobaseN]
 ______________________________________________________________________



 The driver will probe the addresses in the order that they are listed.


 6.2.20.  Western Digital WD7000 cards (`wd7000=')

 The driver probe for the wd7000 looks for a known BIOS ROM string and
 knows about a few standard configuration settings.  If it doesn't come
 up with the correct values for your card, or you have an unrecognized
 BIOS version, you can use a boot argument of the form:


 ______________________________________________________________________
         wd7000=irq,dma,iobase
 ______________________________________________________________________


 6.3.  SCSI Host Adapters that don't Accept Boot Args

 At present, the following SCSI cards do not make use of any boot-time
 parameters. In some cases, you can hard-wire values by directly
 editing the driver itself, if required.


         Adaptec aha1740 (EISA probing),
         NCR53c7xx,8xx (PCI, both drivers)
         Qlogic Fast (0x230, 0x330)
         Qlogic ISP (PCI)



 7.  Hard Disks

 This section lists all the boot args associated with standard MFM/RLL,
 ST-506, XT, and IDE disk drive devices.  Note that both the IDE and
 the generic ST-506 HD driver both accept the `hd=' option.


 7.1.  IDE Disk/CD-ROM Driver Parameters

 The IDE driver accepts a number of parameters, which range from disk
 geometry specifications, to support for advanced or broken controller
 chips. The following is a summary of all the possible boot arguments.
 For full details, you really should consult the file ide.txt in the
 linux/Documentation directory, from which this summary was extracted.


 ______________________________________________________________________

  "hdx="  is recognized for all "x" from "a" to "h", such as "hdc".
  "idex=" is recognized for all "x" from "0" to "3", such as "ide1".

  "hdx=noprobe"              : drive may be present, but do not probe for it
  "hdx=none"         : drive is NOT present, ignore cmos and do not probe
  "hdx=nowerr"               : ignore the WRERR_STAT bit on this drive
  "hdx=cdrom"                : drive is present, and is a cdrom drive
  "hdx=cyl,head,sect"        : disk drive is present, with specified geometry
  "hdx=autotune"             : driver will attempt to tune interface speed
                                 to the fastest PIO mode supported,
                                 if possible for this drive only.
                                 Not fully supported by all chipset types,
                                 and quite likely to cause trouble with
                                 older/odd IDE drives.

  "idex=noprobe"             : do not attempt to access/use this interface
  "idex=base"                : probe for an interface at the addr specified,
                                 where "base" is usually 0x1f0 or 0x170
                                 and "ctl" is assumed to be "base"+0x206
  "idex=base,ctl"    : specify both base and ctl
  "idex=base,ctl,irq"        : specify base, ctl, and irq number
  "idex=autotune"    : driver will attempt to tune interface speed
                                 to the fastest PIO mode supported,
                                 for all drives on this interface.
                                 Not fully supported by all chipset types,
                                 and quite likely to cause trouble with
                                 older/odd IDE drives.
  "idex=noautotune"  : driver will NOT attempt to tune interface speed
                                 This is the default for most chipsets,
                                 except the cmd640.
  "idex=serialize"   : do not overlap operations on idex and ide(x^1)
 ______________________________________________________________________

 The following are valid ONLY on ide0, and the defaults for the
 base,ctl ports must not be altered.


 ______________________________________________________________________

  "ide0=dtc2278"             : probe/support DTC2278 interface
  "ide0=ht6560b"             : probe/support HT6560B interface
  "ide0=cmd640_vlb"  : *REQUIRED* for VLB cards with the CMD640 chip
                           (not for PCI -- automatically detected)
  "ide0=qd6580"              : probe/support qd6580 interface
  "ide0=ali14xx"             : probe/support ali14xx chipsets (ALI M1439/M1445)
  "ide0=umc8672"             : probe/support umc8672 chipsets
 ______________________________________________________________________



 Everything else is rejected with a "BAD OPTION" message.


 7.2.  Standard ST-506 Disk Driver Options (`hd=')

 The standard disk driver can accept geometry arguments for the disks
 similar to the IDE driver. Note however that it only expects three
 values (C/H/S) -- any more or any less and it will silently ignore
 you. Also, it only accepts `hd=' as an argument, i.e. `hda=', `hdb='
 and so on are not valid here. The format is as follows:


 ______________________________________________________________________
         hd=cyls,heads,sects
 ______________________________________________________________________



 If there are two disks installed, the above is repeated with the
 geometry parameters of the second disk.


 7.3.  XT Disk Driver Options (`xd=')

 If you are unfortunate enough to be using one of these old 8 bit cards
 that move data at a whopping 125kB/s then here is the scoop.  The
 probe code for these cards looks for an installed BIOS, and if none is
 present, the probe will not find your card. Or, if the signature
 string of your BIOS is not recognized then it will also not be found.
 In either case, you will then have to use a boot argument of the form:


 ______________________________________________________________________
         xd=type,irq,iobase,dma_chan
 ______________________________________________________________________



 The type value specifies the particular manufacturer of the card, and
 are as follows: 0=generic; 1=DTC; 2,3,4=Western Digital,
 5,6,7=Seagate; 8=OMTI. The only difference between multiple types from
 the same manufacturer is the BIOS string used for detection, which is
 not used if the type is specified.

 The xd_setup() function does no checking on the values, and assumes
 that you entered all four values. Don't disappoint it.  Here is an
 example usage for a WD1002 controller with the BIOS disabled/removed,
 using the `default' XT controller parameters:

 ______________________________________________________________________
         xd=2,5,0x320,3
 ______________________________________________________________________



 8.  CD-ROMs (Non-SCSI/ATAPI/IDE)

 This section lists all the possible boot args pertaining to CD-ROM
 devices. Note that this does not include SCSI or IDE/ATAPI CD-ROMs.
 See the appropriate section(s) for those types of CD-ROMs.

 Note that most of these CD-ROMs have documentation files that you
 should read, and they are all in one handy place:
 linux/Documentation/cdrom.


 8.1.  The Aztech Interface (`aztcd=')

 The syntax for this type of card is:


 ______________________________________________________________________
         aztcd=iobase[,magic_number]
 ______________________________________________________________________



 If you set the magic_number to 0x79 then the driver will try and run
 anyway in the event of an unknown firmware version. All other values
 are ignored.


 8.2.  The CDU-31A and CDU-33A Sony Interface (`cdu31a=')

 This CD-ROM interface is found on some of the Pro Audio Spectrum sound
 cards, and other Sony supplied interface cards.  The syntax is as
 follows:


 ______________________________________________________________________
         cdu31a=iobase,[irq[,is_pas_card]]
 ______________________________________________________________________



 Specifying an IRQ value of zero tells the driver that hardware
 interrupts aren't supported (as on some PAS cards). If your card
 supports interrupts, you should use them as it cuts down on the CPU
 usage of the driver.

 The `is_pas_card' should be entered as `PAS' if using a Pro Audio
 Spectrum card, and otherwise it should not be specified at all.


 8.3.  The CDU-535 Sony Interface (`sonycd535=')

 The syntax for this CD-ROM interface is:


 ______________________________________________________________________
         sonycd535=iobase[,irq]
 ______________________________________________________________________


 A zero can be used for the I/O base as a `placeholder' if one wishes
 to specify an IRQ value.


 8.4.  The GoldStar Interface (`gscd=')

 The syntax for this CD-ROM interface is:


 ______________________________________________________________________
         gscd=iobase
 ______________________________________________________________________



 8.5.  The ISP16 Interface (`isp16=')

 The syntax for this CD-ROM interface is:


 ______________________________________________________________________
         isp16=[port[,irq[,dma]]][[,]drive_type]
 ______________________________________________________________________



 Using a zero for irq or dma means that they are not used. The
 allowable values for drive_type are noisp16, Sanyo, Panasonic, Sony,
 and Mitsumi.  Using noisp16 disables the driver altogether.


 8.6.  The Mitsumi Standard Interface (`mcd=')

 The syntax for this CD-ROM interface is:


 ______________________________________________________________________
         mcd=iobase,[irq[,wait_value]]
 ______________________________________________________________________



 The wait_value is used as an internal timeout value for people who are
 having problems with their drive, and may or may not be implemented
 depending on a compile time DEFINE.


 8.7.  The Mitsumi XA/MultiSession Interface (`mcdx=')

 At present this `experimental' driver has a setup function, but no
 parameters are implemented yet (as of 1.3.15).  This is for the same
 hardware as above, but the driver has extended features.


 8.8.  The Optics Storage Interface (`optcd=')

 The syntax for this type of card is:


 ______________________________________________________________________
         optcd=iobase
 ______________________________________________________________________



 8.9.  The Phillips CM206 Interface (`cm206=')

 The syntax for this type of card is:


 ______________________________________________________________________
         cm206=[iobase][,irq]
 ______________________________________________________________________



 The driver assumes numbers between 3 and 11 are IRQ values, and
 numbers between 0x300 and 0x370 are I/O ports, so you can specify one,
 or both numbers, in any order.  It also accepts `cm206=auto' to enable
 autoprobing.


 8.10.  The Sanyo Interface (`sjcd=')

 The syntax for this type of card is:


 ______________________________________________________________________
         sjcd=iobase[,irq[,dma_channel]]
 ______________________________________________________________________



 8.11.  The SoundBlaster Pro Interface (`sbpcd=')

 The syntax for this type of card is:


 ______________________________________________________________________
         sbpcd=iobase,type
 ______________________________________________________________________



 where type is one of the following (case sensitive) strings:
 `SoundBlaster', `LaserMate', or `SPEA'.  The I/O base is that of the
 CD-ROM interface, and not that of the sound portion of the card.


 9.  Serial and ISDN Drivers

 9.1.  The ICN ISDN driver (`icn=')

 This ISDN driver expects a boot argument of the form:


 ______________________________________________________________________
         icn=iobase,membase,icn_id1,icn_id2
 ______________________________________________________________________



 where iobase is the i/o port address of the card, membase is the
 shared memory base address of the card, and the two icn_id are unique
 ASCII string identifiers.



 9.2.  The PCBIT ISDN driver (`pcbit=')

 This boot argument takes integer pair arguments of the form:


 ______________________________________________________________________
         pcbit=membase1,irq1[,membase2,irq2]
 ______________________________________________________________________



 where membaseN is the shared memory base of the N'th card, and irqN is
 the interrupt setting of the N'th card. The default is IRQ 5 and
 membase 0xD0000.


 9.3.  The Teles ISDN driver (`teles=')

 This ISDN driver expects a boot argument of the form:


 ______________________________________________________________________
         teles=iobase,irq,membase,protocol,teles_id
 ______________________________________________________________________



 where iobase is the i/o port address of the card, membase is the
 shared memory base address of the card, irq is the interrupt channel
 the card uses, and teles_id is the unique ASCII string identifier.


 9.4.  The DigiBoard Driver (`digi=')

 The DigiBoard driver accepts a string of six comma separated
 identifiers or integers.  The 6 values in order are:


         Enable/Disable this card
         Type of card: PC/Xi(0), PC/Xe(1), PC/Xeve(2), PC/Xem(3)
         Enable/Disable alternate pin arrangement
         Number of ports on this card
         I/O Port where card is configured (in HEX if using string identifiers)
         Base of memory window (in HEX if using string identifiers)



 An example of a correct boot prompt argument (in both identifier and
 integer form) is:


 ______________________________________________________________________
         digi=E,PC/Xi,D,16,200,D0000
         digi=1,0,0,16,512,851968
 ______________________________________________________________________



 Note that the driver defaults to an i/o of 0x200 and a shared memory
 base of 0xD0000 in the absence of a digi= boot argument.  There is no
 autoprobing performed. More details can be found in the file
 linux/Documentation/digiboard.txt.



 9.5.  The RISCom/8 Multiport Serial Driver (`riscom8=')

 Up to four boards can be supported by supplying four unique i/o port
 values for each individual board installed.  Other details can be
 found in the file linux/Documentation/riscom8.txt.


 9.6.  The Baycom Serial/Parallel Radio Modem (`baycom=')

 The format of the boot argument for these devices is:


 ______________________________________________________________________
         baycom=modem,io,irq,options[,modem,io,irq,options]
 ______________________________________________________________________



 Using modem=1 means you have the ser12 device, modem=2 means you have
 the par96 device. Using options=0 means use hardware DCD, and
 options=1 means use software DCD. The io and irq are the i/o port base
 and interrupt settings as usual.  There is more details in the file
 README.baycom which is currently in the /linux/drivers/char/
 directory.


 10.  Other Hardware Devices

 Any other devices that didn't fit into any of the above categories got
 lumped together here.


 10.1.  Ethernet Devices (`ether=')

 Different drivers make use of different parameters, but they all at
 least share having an IRQ, an I/O port base value, and a name. In its
 most generic form, it looks something like this:


 ______________________________________________________________________
         ether=irq,iobase[,param_1[,param_2,...param_8]]],name
 ______________________________________________________________________



 The first non-numeric argument is taken as the name.  The param_n
 values (if applicable) usually have different meanings for each
 different card/driver.  Typical param_n values are used to specify
 things like shared memory address, interface selection, DMA channel
 and the like.

 The most common use of this parameter is to force probing for a second
 ethercard, as the default is to only probe for one. This can be
 accomplished with a simple:


 ______________________________________________________________________
         ether=0,0,eth1
 ______________________________________________________________________



 Note that the values of zero for the IRQ and I/O base in the above
 example tell the driver(s) to autoprobe.


 IMPORTANT NOTE TO MODULE USERS: The above will not force a probe for a
 second card if you are using the driver(s) as run time loadable
 modules (instead of having them complied into the kernel).  Most Linux
 distributions use a bare bones kernel combined with a large selection
 of modular drivers.  The ether= only applies to drivers compiled
 directly into the kernel.

 The Ethernet-HowTo has complete and extensive documentation on using
 multiple cards and on the card/driver specific implementation of the
 param_n values where used.  Interested readers should refer to the
 section in that document on their particular card for more complete
 information.  Ethernet-HowTo
 <http://metalab.unc.edu/mdw/HOWTO/Ethernet-HOWTO.html>


 10.2.  The Floppy Disk Driver (`floppy=')

 There are many floppy driver options, and they are all listed in
 README.fd in linux/drivers/block.  There are too many options in that
 file to list here. Instead, only those options that may be required to
 get a Linux install to proceed on less than normal hardware are
 reprinted here.

 floppy=0,daring Tells the floppy driver that your floppy controller
 should be used with caution (disables all daring operations).

 floppy=thinkpad Tells the floppy driver that you have a Thinkpad.
 Thinkpads use an inverted convention for the disk change line.

 floppy=nodma Tells the floppy driver not to use DMA for data
 transfers.  This is needed on HP Omnibooks, which don't have a
 workable DMA channel for the floppy driver. This option is also useful
 if you frequently get "Unable to allocate DMA memory" messages.  Use
 of `nodma' is not recommended if you have a FDC without a FIFO (8272A
 or 82072). 82072A and later are OK). The FDC model is reported at
 boot.  You also need at least a 486 to use nodma.

 floppy=nofifo Disables the FIFO entirely. This is needed if you get
 `Bus master arbitration error' messages from your Ethernet card (or
 from other devices) while accessing the floppy.

 floppy=broken_dcl Don't use the disk change line, but assume that the
 disk was changed whenever the device node is reopened. Needed on some
 boxes where the disk change line is broken or unsupported.  This
 should be regarded as a stopgap measure, indeed it makes floppy
 operation less efficient due to unneeded cache flushings, and slightly
 more unreliable. Please verify your cable connection and jumper
 settings if you have any DCL problems. However, some older drives, and
 also some Laptops are known not to have a DCL.

 floppy=debug Print (additional) debugging messages.

 floppy=messages Print informational messages for some operations (disk
 change notifications, warnings about over and underruns, and about
 autodetection).


 10.3.  The Sound Driver (`sound=')

 The sound driver can also accept boot args to override the compiled in
 values. This is not recommended, as it is rather complex and the
 documentation for it in the kernel mysteriously vanished (a hint).
 You are better off to use sound as a module, or compile in your own
 values.


 If you choose to use it regardless, then processing of the argument
 takes place in the file dev_table.c in linux/drivers/sound. It accepts
 a boot arg of the form:


 ______________________________________________________________________
         sound=device1[,device2[,device3...[,device11]]]
 ______________________________________________________________________



 where each deviceN value is of the following format 0xDTaaaId and the
 bytes are used as follows:

 D - second DMA channel (zero if not applicable)

 T - device type: 1=FM, 2=SB, 3=PAS, 4=GUS, 5=MPU401, 6=SB16,
 7=SB16-MIDI,...  The listing of soundcard types up to 26 (don't forget
 to convert back to hex for command line use) are listed in the file
 linux/include/linux/soundcard.h and 27 to 999 (newer models) can be
 found in the file linux/drivers/sound/dev_table.h.

 aaa - I/O address in hex.

 I - interrupt line in hex (i.e 10=a, 11=b, ...)

 d - First DMA channel.

 As you can see it gets pretty messy, and you really are better off to
 use a modular driver or compile in your own personal values as
 recommended. Using a boot arg of `sound=0' will disable the sound
 driver entirely.


 10.4.  The Bus Mouse Driver (`bmouse=')

 The busmouse driver only accepts one parameter, that being the
 hardware IRQ value to be used.


 10.5.  The MS Bus Mouse Driver (`msmouse=')

 The MS mouse driver only accepts one parameter, that being the
 hardware IRQ value to be used.


 10.6.  The Printer Driver (`lp=')

 With this boot argument you can tell the printer driver what ports to
 use and what ports not to use. The latter comes in handy if you don't
 want the printer driver to claim all available parallel ports, so that
 other drivers (e.g. PLIP, PPA) can use them instead.

 The format of the argument is multiple i/o, IRQ pairs. For example,
 lp=0x3bc,0,0x378,7 would use the port at 0x3bc in IRQ-less (polling)
 mode, and use IRQ 7 for the port at 0x378. The port at 0x278 (if any)
 would not be probed, since autoprobing only takes place in the absence
 of a lp= argument. To disable the printer driver entirely, one can use
 lp=0.


 11.  Copying, Translations, Closing, etc.

 Hey, you made it to the end! (Phew...)  Now just the legal stuff.


 11.1.  Copyright and Disclaimer

 This document is Copyright (c) 1995-1999 by Paul Gortmaker.  Copying
 and redistribution is allowed under the conditions as outlined in the
 Linux Documentation Project Copyright, available from where you
 obtained this document, OR as outlined in the GNU General Public
 License, version 2 (see linux/COPYING).

 This document is not gospel. However, it is probably the most up to
 date info that you will be able to find. Nobody is responsible for
 what happens to your hardware but yourself. If your stuff goes up in
 smoke, or anything else bad happens, we take no responsibility. ie.
 THE AUTHOR IS NOT RESPONSIBLE FOR ANY DAMAGES INCURRED DUE TO ACTIONS
 TAKEN BASED ON THE INFORMATION INCLUDED IN THIS DOCUMENT.

 A hint to people considering doing a translation.  First, translate
 the SGML source (available via FTP from the HowTo main site) so that
 you can then generate other output formats.  Be sure to keep a copy of
 the original English SGML source that you translated from! When an
 updated HowTo is released, get the new SGML source for that version,
 and then a simple diff -u old.sgml new.sgml will show you exactly what
 has changed so that you can easily incorporate those changes into your
 translated SMGL source without having to re-read or re-translate
 everything.

 If you are intending to incorporate this document into a published
 work, please make contact (via e-mail) so that you can be supplied
 with the most up to date information available. In the past, out of
 date versions of the Linux HowTo documents have been published, which
 caused the developers undue grief from being plagued with questions
 that were already answered in the up to date versions.


 11.2.  Closing

 If you have found any glaring typos, or outdated info in this
 document, please let me know. It is easy to overlook stuff, as the
 kernel (and the number of drivers) is huge compared to what it was
 when I started this.

 Thanks,

 Paul Gortmaker, [email protected]