D/Documentation////

D/arch////

D/drivers////

D/fs////

D/include////

D/init////

D/scripts////

/Makefile/1.8/Thu Apr 15 13:58:24 2004//

/kernel.config/1.18/Thu Apr 15 14:01:29 2004//

kernel

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

/README.xbox/1.6/Thu Feb 19 15:13:44 2004//

D/input////

/Configure.help/1.17/Thu Apr 15 13:57:48 2004//

kernel/Documentation

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

Xbox support

CONFIG_XBOX

  If you want your kernel to be compatible with the Microsoft Xbox,

  say Y. This option includes adjustments so that the kernel runs

  both on a standard PC or on an Xbox.

  This code autodetects an Xbox at boot, and if it is found, it

  avoids the PCI enumeration chipset bug, adjusts the system

  frequency and adds the Xbox shutdown/reset sequence.

Xbox DVD eject fix

CONFIG_XBOX_EJECT

  If you want to be able to eject the DVD drive tray, say Y, or else

  your eject button will be a reset button. For security reasons, the

  Xbox is reset by default when the optical media is supposed to be

  ejected.

  If you want to compile this as a module ( = code which can be

  inserted in and removed from the running kernel whenever you want),

  say M and read <file:Documentation/modules.txt>. The module will be

  called xboxejectfix.o.

  You should really say Y here. If you compile it as a module, make

  sure that it gets loaded as soon as possible, so that you are

  protected against accidental resets from the very beginning.

Xbox (nVidia) support

CONFIG_FB_XBOX

  This driver supports graphics boards with the nVidia GeForce 3MX GPU

  as found in the Microsoft Xbox. Note that this driver and CONFIG_FB_RIVA

  (nVidia Riva support) are mutually exclusive, ie. you can enable only

  one and not both.

  Note 2: This driver depends on both CONFIG_XBOX (Xbox support) and

  CONFIG_I2C_XBOX (I2C XBOX).

  Say Y if you are running Linux on the Xbox, otherwise, say N.

  The driver is also available as a module ( = code which can be

  inserted and removed from the running kernel whenever you want). The

  module will be called xboxfb.o. If you want to compile it as a

  module, say M here and read <file:Documentation/modules.txt>.

nForce Ethernet support (EXPERIMENTAL)

CONFIG_FORCEDETH

  If you have a network (Ethernet) controller of this type, say Y and

  read the Ethernet-HOWTO, available from

  <http://www.tldp.org/docs.html#howto>.

  To compile this driver as a module, choose M here and read

  <file:Documentation/networking/net-modules.txt>.  The module will be

  called forcedeth.o.

Xbox Infrared DVD dongle support (EXPERIMENTAL)

CONFIG_USB_XBOXIR

  Say Y here if you want to use the Xbox DVD dongle and remote control

  to help control running applications using keyboard codes. Make sure

  to say Y to "USB Keyboard support" (CONFIG_INPUT_KBD) and/or

  "Event interface support" (CONFIG_INPUT_EVDEV) as well.

Xbox Controller (XPad)

CONFIG_USB_XPAD

  Say Y here if you want to use the X-Box pad with your computer.

  Make sure to say Y to "Joystick support" (CONFIG_INPUT_JOYDEV)

  and/or "Event interface support" (CONFIG_INPUT_EVDEV) as well.

  For information about how to connect the X-Box pad to USB, see

  Documentation/input/xpad.txt.

  This driver is also available as a module ( = code which can be

  inserted in and removed from the running kernel whenever you want).

  The module will be called xpad.o.  If you want to compile it as a

  module, say M here and read <file:Documentation/modules.txt>.

Xbox Controller Mouse Emulation

CONFIG_USB_XPAD_MOUSE

  Say Y here if you want to use the Xbox pad as a mouse

  with your computer. Make sure to say Y to "Mouse support"

  (CONFIG_INPUT_MOUSEDEV) as well. This emulation is part of the

  regular XPad driver, so you have to enable CONFIG_USB_XPAD for

  this to work.

  For information about how to connect the Xbox pad to USB, see

  Documentation/input/xpad.txt.

FATX (Xbox) fs support

CONFIG_FATX_FS

  This adds support for the FATX filesystem as found in Microsoft's Xbox.

  The FATX filesystem is a derivative of the FAT filesystem minus some

  legacy fields and redundant information. For lengthier discussions on

  the filesystem, see:

  http://xbox-linux.sourceforge.net/docs/fatxfat.html

  http://xbox-linux.sourceforge.net/docs/hdpartfs.html

  http://xbox-linux.sourceforge.net/docs/hackingfatx.html

  If you are running Linux on the Xbox, say Y unless you know what

  you're doing. Otherwise, say N.

  If you want to compile this as a module ( = code which can be

  inserted in and removed from the running kernel whenever you want),

  say M here and read <file:Documentation/modules.txt>.  The module

  will be called fatx.o.

Xbox partition support

CONFIG_XBOX_PARTITION

  The Xbox makes use of a static and implicit partitioning scheme.

  The locations and sizes of the four partitions on the Xbox hard

  disk are hard-coded into the native Xbox kernel.

  Say Y here if you are running Linux on the Xbox, or would like

  to access a hard disk partitioned for the Xbox under a PC

  running Linux.

  If unsure, say N.

$Id: README.xbox,v 1.6 2004/02/19 15:13:44 aothieno Exp $

We are baselined on 2.4.25

Introductory Note

-----------------

First, if you are shaking with excitement and fear and this is your first 

attempt to compile the kernel, don't worry if things don't work out first time.

Expect trouble the first couple of attempts, make sure you have a way to boot 

an old kernel after trying a fresh one.  Note too that the kernel compile 

process does nothing to your system config until you copy the new kernel image 

to /boot and make it the default choice for boot.

A very easy mistake to make is to forget to do the make modules and 

make modules_install after doing the bzImage, since it takes a little while 

and your mind has probably wandered.

Don't be afraid to come on IRC irc.oftc.net, #xbox-linux or the devel mailing 

list and ask some of the old hands there.

Compiling the kernel patched for Xbox

-------------------------------------

In order to use this CVS archive you need to setup things properly.

This document decribes three different ways to build the xbox kernel.

The first method attempts to build the kernel from the "HEAD" or current

version inside CVS. The current version is traditionally unstable and

(like in most projects) is very user unfriendly and may not even compile.

This is traditionally used by the experienced developers only. The code

changes frequently.

The second method is more 'user friendly'. It shows you how to extract a known

(tagged) working release of the kernel which will always compile. In essence,

the developers have put together a working set of files, tested them, tried them

and 'tagged' them as 'user friendly', IE. a fully working kernel.

The third, and easiest method, it to download and apply one of the tagged

release patches and apply it to the vanilla base kernel. This option is the

easiest option for users. You don't even need to know anything about cvs!

If you are unsure, choose this option.

Independent of how you've chosen to build your kernel, you will most certainly

perform these preliminary steps:

1. Download a kernel source tarball from:

   http://www.XY.kernel.org/pub/linux/kernel/v2.4/linux-2.4.25.tar.gz

   where XY is a two-letter country code, eg. `us', `de', `fr', etc.

2. Extract the contents of the archive into /usr/src:

   $ cd /usr/src

   $ tar xzvf linux-2.4.25.tar.gz

   $ mv linux-2.4.25{,-xbox}

Option 1: Installing From A Recent CVS Snapshot.

------------------------------------------------

3.1 Assuming you have a SF.net account, checkout the latest CVS snapshot:

    $ cvs -d:ext:username@cvs.sf.net:/cvsroot/xbox-linux co kernel

    If you do not have a SF.net user account, you may perform an anonymous

    checkout instead:

    $ cvs -d:pserver:anonymous@cvs.sf.net:/cvsroot/xbox-linux login

    $ cvs -d:pserver:anonymous@cvs.sf.net:/cvsroot/xbox-linux co kernel

    Now, you may copy the contents of the CVS repository into the kernel

    source directory:

    $ cp -rf kernel/* linux-2.4.25-xbox/

Option 2: Installing From A Tagged CVS Release.

-----------------------------------------------

3.2 Tagged releases are known to work, although this can not be fully

    guaranteed. Information regarding available tags can always be

    obtained from either the user or developer mailing lists. The

    current tagged release is `kernel-2_4_25-0_1_0'.

    So, assuming you have a SF.net account, you may proceed to

    checkout a tagged release like this:

    $ cvs -d:ext:username@cvs.sf.net:/cvsroot/xbox-linux \

      co -r kernel-2_4_25-0_1_0 kernel

    Or anonymously, if you don't have a SF.net account:

    $ cvs -d:pserver:anonymous@cvs.sf.net:/cvsroot/xbox-linux login

    $ cvs -d:pserver:anonymous@cvs.sf.net:/cvsroot/xbox-linux \

      co -r kernel-2_4_25-0_1_0 kernel

    Then copy the contents of the tagged release into the kernel source

    directory:

    $ cp -rf kernel/* linux-2.4.25-xbox/

Option 3: Installing From A Tagged CVS Release by Patching.

-----------------------------------------------------------

3.3 Tagged releases are also available in form of a patch against a stock

    version of the Linux kernel. Grab it from the SF.net project download

    page, save it in /usr/src and then apply it, like so:

    $ bzip2 -dc patch-kernel-2.4.20-dev-0.6.1.bz2 | patch -p0

4. With a patched kernel, you may now proceed to preparing the source tree

   for the final build:

   $ cd linux-2.4.25-xbox/

   $ make mrproper

   $ cp kernel.config .config

5. We need to load and save the config file to cause some actions to be made

   by the makefile. There are a bunch of ways to do this, the simplest is:

   $ make oldconfig

   If you need to see and change the config options, you need some packages

   installed:

   make menuconfig (curses-based configuration):

   ---------------------------------------------

   ncurses-base, ncurses-bin, ncurses-term, libncurses5 & libncurses5-dev

   make xconfig (graphical-based configuration):

   ---------------------------------------------

   tcl, tk

   Exit from the menu configuration tools and choose YES to save your

   configuration. (This will build the menu configuration utilities and

   refresh the kernel configuration file named .config we previously made

   a default for)

   Now you have a usable kernel tree which you can do your compilations.

6. Compile the kernel and the modules, make sure none of these commands

   abort abnormally with errors. (NOTE: You will need to be root to make

   modules_install):

   $ make dep

   $ make bzImage

   $ make modules

   # make modules_install

The kernel has been compiled and is called `arch/i386/boot/bzImage'. Kernel

modules have been compiled and installed into `/lib/modules/2.4.25-xbox/'.

Kernel compilation complete.

Patch Files:

------------

Patch files will be released to the community based on tagged releases of CVS.

The name of the file will include the base kernel and the project tag version,

for example: patch-kernel-2.4.20-dev-0.6.1.bz2

Clearly, this patch should be applied to 2.4.20, is development quality and

was taken from the CVS tagged version 0.5.2.

Template Configuration File: `kernel.config':

---------------------------------------------

A default configuration file has been added with all of the necessary

compile options automatically set. This may save you time and trouble.

Finally (optional):

-------------------

If you've got no use for it, feel free to remove your working copy of the

CVS checkout:

   $ cd ../

   $ rm -rf kernel/

CVS Syncmail:

-------------

CVS is configured to send all repository changes to the xbox-linux-cvs 

mailing list.

The following text is taken from the official sourceforge CVS synmail 

configuration document:

A Warning Regarding Commit Messages:

------------------------------------

When you begin to use syncmail within your project CVS repository, it is 

important to give your developers prior notice. syncmail-generated messages

to your mailing list will contain the commit message used by the developer; 

developers should keep this in mind when entering their commit message 

(not to make inappropriate statements within their commit messages). 

SourceForge.net does not provide the means to remove posts from mailing

list archives; once a commit message appears in list archives, it will always 

appear in those archives. Proceed with caution."

Inside the Xbox Patches:

------------------------

The Xbox has at least the following problems so it doesn't run the Linux

kernel very well:

* the PCI bug: reading from 0:0:1 freezes the machine

* the timer: it is off by 6%

* the shutdown/reboot sequence: it's incompatible

The current kernel detects the Xbox by its 0:0:0 PCI device ID and sets

machine_is_xbox to 1. The PCI, timer and shutdown/reboot patches will

be applied then. This means that a Linux kernel with Xbox support will

still work on a PC without any differences.

Problems:

---------

If you have any problems *ask* (try the IRC channel)

/xpad.txt/1.3/Sun Aug 10 23:38:45 2003//

D

kernel/Documentation/input

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

xpad - Linux USB driver for Xbox gamepads

------------------------------------------

This driver is work in progress. Although it is already fairly functional,

there are still quite a few things that are not implemented.

See the ToDo-List in drivers/usb/input/xpad.c for details.

0. Status

---------

I believe this driver to have been more or less intensively tested by

xbox-linux users, because it somehow managed it into their CVS [1] ;)

I myself have tested it on just one Linux-Box. This one is running a 2.4.19

kernel with usb-uhci on an amd athlon 600.

Additionally I did receive a few mails from people who have successfully

used the driver on different PC systems and notebooks. I did not yet receive

info on wether it works on PPC or non-i386 hardware.

Version 0.1.0-pre includes Oliver Schwartz' xpad-mouse driver code. This allows

for the xpad to be used as a pointer device (right thumbstick is movement, left

trigger is left mouse button, right one is right). This is not supposed to stay

that way (differs from Oliver's mapping), so don't get too used to it.

The jstest-program from joystick-1.2.15 (jstest-version 2.1.0) reports

14 axes (6 of them are the analog buttons) and 10 buttons (the analog buttons

are mapped as digital ones, too).

Alls 14 axes work, though they all have the same range (-32768..32767)

and the zero-setting is not correct for the triggers and the buttons

(I don't know if that is some limitation of jstest, since the input device

setup should be fine. I didn't have a look at jstest itself yet).

All of the 10 buttons work. The six buttons on the right side (A, B, C [black],

X, Y, Z [white]) are pressure-sensitive and report analog values as 8 bit

unsigned. These are mapped to analog (ABS_HAT1X - ABS_HAT3Y) as well as

digital inputs.

As of version 0.1.0-pre there is EXPERIMENTAL force feedback code. Note that

this is currently just plain stupid (as I have yet to find out all the neat

USB tricks) and is known to have bugs. Specifically, it does not work with the

stock MS controllers. The InterAct device rumbles just fine, though. I am

currently trying to figure out why.

Anyway, USE the stuff AT YOUR OWN RISK, the usual disclaimer applies. It

actually managed to crash my system once (unfortunately not reproducable).

You HAVE BEEN WARNED.

I tested the controller with quake3, and configuration and

in game functionality were OK. However, I find it rather difficult to

play first person shooters with a pad. Your mileage may vary.

1. USB adapter

--------------

Before you can actually use the driver, you need to get yourself an

adapter cable to connect the Xbox controller to your Linux-Box.

Such a cable is pretty easy to build. The Controller itself is a USB compound

device (a hub with three ports for two expansion slots and the controller

device) with the only difference in a nonstandard connector (5 pins vs. 4 on

standard USB connector).

You just need to solder a USB connector onto the cable and keep the

yellow wire unconnected. The other pins have the same order on both

connectors so there is no magic to it. Detailed info on these matters

can be found on the net ([2], [3], [4]).

Thanks to the trip splitter found on the cable you don't even need to cut the

original one. You can buy an extension cable and cut that instead. That way,

you can still use the controller with your Xbox, if you have one ;)

2. driver installation

----------------------

Once you have the adapter cable and the controller is connected, you need

to load your USB subsystem and should cat /proc/bus/usb/devices.

There should be an entry like the one at the end [5].

Currently (as of version 0.1.0), the following three devices are recognized:

 original Microsoft Xbox controller (US), vendor=0x045e, product=0x0202

 original Microsoft Xbox controller (Japan), vendor=0x045e, product=0x0285

 InterAct PowerPad Pro (Germany), vendor=0x05fd, product=0x107a

The driver does feel responsible for the actual USB ids, not the vendor

settings, so other controllers SHOULD work (although they will be reported as

unknown controller). If you have indication of that statement being wrong,

please send me a note.

Also, if you have a controller that is not recognized (but works) and you want

it to be correctly setup, please drop me a line with the appropriate info (that

is, include the name, vendor and product ID; sending the whole dump out of

/proc/bus/usb/devices along would be even better).

If you compiled and installed the driver, test the functionality:

> modprobe xpad

> modprobe joydev

> jstest /dev/js0

There should be 24 inputs (14 axes, 10 buttons), and the values should change

if you move the sticks and push the buttons.

You can test the force feedback functionality by

> modprobe evdev

> fftest /dev/input/event0

You need the Johann Deneux' fftest utility and the evdev module. Note that

the actual event interface number might differ (depends on wether you have the

mouse support included and wether it loads first or not) on your specific setup.

It works? Voila, your done ;)

3. Thanks

---------

I have to thank ITO Takayuki for the detailed info on his site

 http://euc.jp/periphs/xbox-controller.ja.html.

His useful info and both the usb-skeleton as well as the iforce input driver

(Greg Kroah-Hartmann; Vojtech Pavlik) helped a lot in rapid prototyping

the basic functionality.

4. References

-------------

1. http://xbox-linux.sourceforge.net

2. http://euc.jp/periphs/xbox-controller.ja.html (ITO Takayuki)

3. http://xpad.xbox-scene.com/

4. http://www.xboxhackz.com/Hackz-Reference.htm

5. /proc/bus/usb/devices - dump from InterAct PowerPad Pro (Germany):

T:  Bus=01 Lev=03 Prnt=04 Port=00 Cnt=01 Dev#=  5 Spd=12  MxCh= 0

D:  Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=32 #Cfgs=  1

P:  Vendor=05fd ProdID=107a Rev= 1.00

C:* #Ifs= 1 Cfg#= 1 Atr=80 MxPwr=100mA

I:  If#= 0 Alt= 0 #EPs= 2 Cls=58(unk. ) Sub=42 Prot=00 Driver=(none)

E:  Ad=81(I) Atr=03(Int.) MxPS=  32 Ivl= 10ms

E:  Ad=02(O) Atr=03(Int.) MxPS=  32 Ivl= 10ms

-- 

Marko Friedemann <mfr@bmx-chemnitz.de>

2003-01-23

D/i386////

kernel/arch

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

/config.in/1.11/Wed Feb 18 23:45:57 2004//

D/boot////

D/kernel////

kernel/arch/i386

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

D/compressed////

kernel/arch/i386/boot

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

/Makefile/1.4/Tue Nov 25 22:05:11 2003//

D

kernel/arch/i386/boot/compressed

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

# we don't know why, but newer (v1.1+) Xbox only boot

# if misc.o is compiled without optimization

ifeq ($(CONFIG_XBOX),y)

  EXTRACFLAGS-misc:= -O0

endif

bool 'Xbox support' CONFIG_XBOX

if [ "$CONFIG_XBOX" = "y" ]; then

   tristate '  Xbox DVD eject fix' CONFIG_XBOX_EJECT

fi

/Makefile/1.4/Fri Nov 28 23:51:51 2003//

/pci-pc.c/1.10/Fri Dec 19 00:03:21 2003//

/process.c/1.12/Thu Feb 19 21:42:04 2004//

/xboxejectfix.c/1.6/Fri Aug  8 12:19:48 2003//

/setup.c/1.5/Thu Apr 15 13:57:48 2004//

D

kernel/arch/i386/kernel

:ext:huceke@cvs.sf.net:/cvsroot/xbox-linux

obj-$(CONFIG_XBOX_EJECT)	+= xboxejectfix.o

#ifdef CONFIG_XBOX

	extern int machine_is_xbox;

#endif /* CONFIG_XBOX */

#ifdef CONFIG_XBOX

	/* Because of a hardware bug in the Xbox (nForce) chipset,

	 * reading from 0:0:1 and 0:0:2 in the PCI configuration

	 * space crashes the machine. */

	if (machine_is_xbox) {

		if (bus > 1)

			return -EINVAL;

		if ((bus == 1) && (dev | fn))

			return -EINVAL;

		if ((bus == 0 && dev == 0) && ((fn == 1) || (fn == 2)))

			return -EINVAL;

	}

#endif /* CONFIG_XBOX */

#ifdef CONFIG_XBOX

#include <linux/xbox.h>

#endif

#ifdef CONFIG_XBOX

	if (machine_is_xbox) {

		while (1) {

			Xbox_reset();

			mdelay(1000);

			printk(KERN_EMERG "Calling I2C reset again\n");

		}

	}

#endif

#ifdef CONFIG_XBOX

	if (machine_is_xbox) {

		while (1) {

			Xbox_power_off();

			mdelay(1000);

			printk(KERN_EMERG "Calling I2C power off again\n");

		}

	}

#endif

 * Xbox timer patch

 */

#ifdef CONFIG_XBOX

int CLOCK_TICK_RATE;

int machine_is_xbox __initdata = 0;

#endif

/*

	/* SETUP_ARCH for Xbox */

#ifdef CONFIG_XBOX

	outl(0x80000000, 0xcf8);

	if (inl(0xcfc)==0x02a510de) { /* Xbox PCI 0:0:0 ID 0x10de/0x02a5 */

		machine_is_xbox = 1;

		CLOCK_TICK_RATE = 1125000;

		printk("Xbox detected - enabling Xbox patches.\n");

	} else {

		CLOCK_TICK_RATE = 1193180;

	}

#endif

/**

 * Driver that handles the EXTSMI# interrupt on the xbox.

 * Makes it possible to use the eject-button without the xbox rebooting...

 *

 * smbus-command sequence to prevent reboot from cromwell.

 *

 * Changelog:

 *  2003-01-14 Anders Gustafsson <andersg@0x63.nu>

 *             initial version

 *  2003-02-08 Milosch Meriac <xboxlinux@meriac.de>

 *             rewrote debug macros because of compiler errors

 *  2003-08-06 Michael Steil <mist@c64.org>

 *             removed Linux I2C dependency, now compiles

 *             without I2C in the kernel

 *

 * Todo: add errorhandling!

 *

 */

#include <linux/kernel.h>

#include <linux/module.h>

#include <linux/init.h>

#include <asm/io.h>

#include <linux/xbox.h>

#define IRQ 12

#define DRIVER_NAME "xboxejectfix"

/* just some crap */

static char dev[]=DRIVER_NAME;

/* External variable from ide-cd.c that specifies whether we simulate drive

   locking in software */

extern volatile int Xbox_simulate_drive_locked;

#define BASE 0x8000

/* Power Management 1 Enable Register */

#define PM02 (BASE+0x02)

/* Power Management 1 Control Register */

#define PM04 (BASE+0x04)

/* ACPI GP Status Register */

#define PM20 (BASE+0x20)

/* ACPI GP Enable Register */

#define PM22 (BASE+0x22)

# define EXTSMI_EN_MASK 0x0002

/* Global SMI Enable Register */

#define PM2A (BASE+0x2A)

static DECLARE_MUTEX(extsmi_sem);

static DECLARE_COMPLETION(extsmi_exited);

static int extsmi_pid=0;

static void extsmi_interupt(int i,void *dev,struct pt_regs *r){

	int reason;

	reason=inw(0x8020);

	outw(reason,0x8020); /* ack  IS THIS NEEDED? */

	if(reason&0x2){

		/* wake up thread */

		up(&extsmi_sem);

	}

}

/**

 * Process an event. This is run in process-context.

 */

static void extsmi_process(void){

	int reason;

	reason=Xbox_SMC_read(SMC_CMD_INTERRUPT_REASON);

	if(reason&TRAYBUTTON_MASK){ /* Tray button! Respond to prevent reboot! */

		Xbox_SMC_write(SMC_CMD_INTERRUPT_RESPOND, SMC_SUBCMD_RESPOND_CONTINUE);

		Xbox_SMC_write(0x00, 0x0c);

		/* eject unless lock simulation is being used */

		if (!Xbox_simulate_drive_locked)

			Xbox_tray_eject();

	}

}

static int extsmi_thread(void *data){

	daemonize();

	reparent_to_init();

	strcpy(current->comm, "xbox_extsmi");

	do {

		extsmi_process();

		down_interruptible(&extsmi_sem);

	} while (!signal_pending(current));

         complete_and_exit(&extsmi_exited, 0);

}

static int extsmi_init(void){

	int pid;

	if (!machine_is_xbox) {

		printk("This machine is no Xbox.\n");

		return -1;

	}

	printk("Enabling Xbox eject problem workaround.\n");

        pid = kernel_thread(extsmi_thread, NULL,

			    CLONE_FS | CLONE_FILES | CLONE_SIGHAND);

	if (pid < 0) {

		return pid;

	}

	extsmi_pid = pid;

	/* this shuts a lot of interrupts off! */

	outw(inw(0x80e2)&0xf8c7,0x80e2);

	outw(0,0x80ac);

	outb(0,0x8025);

	outw(EXTSMI_EN_MASK,PM22); /* enable the EXTSMI# interupt! */

	outw(0,PM02);

	outb(1,PM04); /* enable sci interrupts! */

	Xbox_SMC_write(SMC_CMD_RESET_ON_EJECT, SMC_SUBCMD_RESET_ON_EJECT_DISABLE);

	/* FIXME! retval! */

	request_irq(IRQ,extsmi_interupt,SA_INTERRUPT|SA_SHIRQ /* |SA_SAMPLE_RANDOM */,

		    "xboxejectfix",dev);

	return 0;

}

static void extsmi_exit(void){

	int res;

	if (!machine_is_xbox) return; /* can this happen??? */

	free_irq(IRQ,dev);

	/* Kill the thread */

	res = kill_proc(extsmi_pid, SIGTERM, 1);

	wait_for_completion(&extsmi_exited);

	return;

}

module_init(extsmi_init);

module_exit(extsmi_exit);

MODULE_AUTHOR("Anders Gustafsson <andersg@0x63.nu>");

MODULE_LICENSE("GPL");

VERSION          = 2

PATCHLEVEL       = 4

SUBLEVEL 	      = 26

EXTRAVERSION     = 

ARCH             = i386

KPKG_ARCH        = 

HOSTCC           = gcc-3.2

HOSTCFLAGS       = -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer

CROSS_COMPILE    = 

AS               = as

LD               = ld -m elf_i386

CC               = gcc-3.2

CPP              = gcc-3.2 -E

AR               = ar

NM               = nm

STRIP            = strip

OBJCOPY          = objcopy -O binary -R .note -R .comment -S

OBJDUMP          = objdump

MAKE             = /usr/bin/make

GENKSYMS 	      = /sbin/genksyms

CFLAGS           = -D__KERNEL__ -I$(TOPDIR)/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=i686

AFLAGS           = -D__ASSEMBLY__ -D__KERNEL__ -I$(TOPDIR)/include

MODFLAGS 	      = -DMODULE

do_parallel      = 

fast_dep         = 

gcc-2.95.4-14

gcc-2.95-2.95.4-11woody1

gcc-3.2-3.2.2-1xebian

gcc-3.2-base-3.2.2-1xebian

perl-5.6.1-8.7

dpkg-1.9.21

libc6-2.2.5-11.5

libc6-dev-2.2.5-11.5

binutils-2.14.90.0.7-3xebian

make-3.79.1-14

dpkg-dev-1.9.21

this was built on a machine with the kernel:

Linux centi.localdomain.local 2.6.5 #1 Sun May 2 17:31:11 CEST 2004 i686 unknown

using the compiler:

gcc version 3.2.2

kernel-source-2.4.26 (1) unstable; urgency=low

  * Added support for netwinders

  * Modified modules support to not assume everything in /usr/src/modules

    is a directory. 

 -- Manoj Srivastava <srivasta@debian.org>  Wed, 25 Nov 1998 01:39:43 -0600

kernel-source-0.34 (1.01)  unstable; urgency=low

  * This is a major change in the source/header packages. In architectures

    different from i386, `kernel-headers-*' and `kernel-source-*' are

    incompatible. This also has an effect on libc6-dev, which depends on

    kernel-headers-2.0.3x | kernel-source-2.0.3x, which is only true on

    i386 architectures. One possible solution would be to make libc6-dev

    depend on kernel-headers-2.0.32, but that won't work because

    kernel-source provides kernel-headers. In fact, since kernel-headers

    are arch dependent, but kernel-source is arch: all, kernel-source-*

    should not provide kernel-header* at all. 

    Once kernel source stops providing kernel-header*, and libc6-dev

    depends on kernel-headers-2.0.32 and links /usr/include/{linux,asm} to

    /usr/src/kernel-headers-2.0.32 (nor /usr/src/linux-2.0.32) things

    would work again in a Multi arch compatible fashion.

    Hence, now kernel-source-* packages compiled with this kernel-package

    shall not provide any sort of kernel-headers. For the sake of

    backwards compatibility, /usr/src/linux-$version symlinks are still

    being provided (as people may upload newer kernels while keeping an

    older libc6-dev around, which depends on /usr/src/linux-2.0.32. 

    However, I have been badgered enough about this that I shall remove

    the /usr/src/linux-$version symlinks at some point. This version no

    longer registers stuff in /usr/src/.linux-versions, and is no longer

    as paranoid about /usr/src/linux; but it does not outright remove

    those files either, so as not to cause people with older kernels

    having a problem during removal.As soon as it is deemed permissible,

    we shall get less paranoid about /usr/src/linux-$version as well. 

  * closes: Bug#18277

 -- Manoj Srivastava <srivasta@debian.org>  Wed, 18 Feb 1998 16:44:31 -0600

kernel-source-0.33 (1.01) unstable; urgency=low

  * Added SUBARCH, which is used to distinguish Amiga, Atari, Macintosh,

    etc. kernels for Debian/m68k, and may well be required for the alpha,

    from what I hear. Unlike the FLAVOURS variable, which affects

    everything it can lay its grubby hands on (kernel image, headers,

    source, doc package versions, and where the modules are looked for

    under /lib/modules), this only affects the naming of the kernel-image

    as the source and doc packages are architecture independent and the

    kernel-headers do not vary from one sub-architecture to the

    next. These changes are courtesy of James Troup <J.J.Troup@scm.brad.ac.uk>

 -- Manoj Srivastava <srivasta@debian.org>  Mon, 16 Feb 1998 16:47:08 -0600

kernel-source-0.32 (1.01) unstable; urgency=low

  * Ensure that the /usr/src/linux and /usr/src/linux-<version> links

    always exist, no matter what. Apparently, upgrading from

    kernel-source-2.0.32_2.0.32-1 to kernel-source-2.0.32_2.0.32-3 does

    not create /usr/src/linux-<version>, which breaks libc6-dev.

  * Toned down the language about LILO, so as not to startle new sparc users

 -- Manoj Srivastava <srivasta@debian.org>  Thu, 12 Feb 1998 21:56:50 -0600

kernel-source-0.31 (1.01) unstable; urgency=low

  * Added ARCH to make called, this allows for cross compiling kernels

    (added on a request by  James Troup <J.J.Troup@scm.brad.ac.uk>)

  * Make kernel-headers arch: any, so we can indeed have different headers

    for different architectures.

  * m68k can now handle vmlinuz, so reverse that behaviour in the rules

    file. AFAIK m68k still uses zImage.

  * Improvements to /usr/doc/kernel-patch/MultiArch.gz, based on

    suggestions by James Troup <J.J.Troup@scm.brad.ac.uk>/

 -- Manoj Srivastava <srivasta@debian.org>  Mon,  9 Feb 1998 17:11:05 -0600

kernel-source-0.30 (1.01) unstable; urgency=low

  * Redid the Headers README file

  * Added a rationale to the LinkPolicy document. So far, it detailed

    *what* Debian did. Now, it also says *why* we do it.

 -- Manoj Srivastava <srivasta@debian.org>  Thu, 29 Jan 1998 19:13:58 -0600

kernel-source-0.29 (1.01) unstable; urgency=low

  * Important changes for kernel-source-* and kernel-header-* packages: now

    kernel-source-* packages also provide the exact kernel-header-* (libc6

    need only depend on kernel-header-* now.

  * The kernel-header-* and kernel-source-* packages now also maitain the

    /usr/src/linux-X.YY.ZZ links (in addition to the /usr/src/linux links)

    This is used in the libc6 package.

 -- Manoj Srivastava <srivasta@debian.org>  Sat, 13 Dec 1997 12:46:47 -0600

kernel-source-0.28 (1.01) unstable; urgency=low

  * Fixed a typo where we tried to dd /vmlinuz-2.0.32 rather than the

    correct /boot/vmlinuz-2.0.32 in image.postinst. How come this glaring

    an error has gone unreported until now?

 -- Manoj Srivastava <srivasta@debian.org>  Fri, 23 Jan 1998 14:36:34 -0600

kernel-source-0.27 (1.01) unstable; urgency=low

 * Fixed Typo in kernel rules that put all modules into block (this

    is more of a thinko/cut and paste error. fixes: Bug#16697,Bug#16702

  * No longer a fatal error if there is no /vmlinuz (or

    equivalent). fixes: Bug#16899

  * Added language to the abort on /usr/src/linux not being a

    link. 

  * Documented the fact that if you re-run make-kpkg with a different

    revision number, you have to reconfigure the kernel. 

    fixes: Bug#16968

  * ignore obsolete fdformat in favour of superformat.

  * Changed the kernel rules file not break on the sound modules of

    2.1.70+ kernels (I think it is a bug in the kernel Makefile, but this

    fix make make-kpkg handle the problem and be more robust).

  * Modified image.postinst to also cater to people on whose architecture

    the image is not called vmlinuz but something else (like vmlinux, for

    example). closes:Bug#16258

  * Made the postrm's also know about the kernel versrion, not just the

    postinst. (Sorry). closes:Bug#15920

  * Changed include.postrm to be more careful about removing the symbolic

    link /usr/src/linux-X.Y.Z. Keep track if there is another package

    installed that could provide the link.

 -- Manoj Srivastava <srivasta@debian.org>  Wed, 21 Jan 1998 03:27:35 -0600

kernel-source-0.26 (1.01) unstable; urgency=low

  * Added HAM modules to the module we know about, these were introduced

    in 2.1.70.

 -- Manoj Srivastava <srivasta@debian.org>  Tue,  9 Dec 1997 23:27:04 -0600

kernel-source-0.25 (1.01) unstable; urgency=low

  * Ignore unmounted devices while looking for a root file system. Much

    thanks to Thomas Kocourek <tko@westgac3.dragon.com> for noticing

    this. 

  * Make sure that the copyright file for the kernel-doc package is not

    compressed. closes:Bug#14403,Bug#14405

  * Added internal utility kpkg-vercheck to test the validity of the

    package version.

 -- Manoj Srivastava <srivasta@debian.org>  Mon, 10 Nov 1997 10:37:08 -0600

kernel-source-0.24 (1.01) unstable; urgency=low

  * Install the README.headers in the right place for the source

    package. closes:Bug#14552.

  * Handle the new NLS_MODULES that have appeared in the newest 2.1.6x

    kernels. closes:Bug#14527.

 -- Manoj Srivastava <srivasta@debian.org>  Wed,  5 Nov 1997 23:30:48 -0600

kernel-source-0.23 (1.-0) unstable; urgency=low

  * Added README.headers to the kernel-source package as well, since the

    information _is_ relevant to compiling kernels. Since kernel source

    packages have higher visibility than kernel header packages, this may

    help avoid some FAQs from being asked.

  * Changed image postinst not to use the obsolete -d option to

    superformat. Removed extra spaces from the exec option, so that it is

    more likely to work. Noticed by Joost Kooij <kooij@mpn.cp.philips.com>

  * Note that the proposed two level versioning scheme fails if standard

    kernels use epochs. Further note that one may introduce epochs even in

    custom kernels. 

 -- Manoj Srivastava <srivasta@debian.org>  Thu, 23 Oct 1997 12:19:51 -0500

kernel-source-0.22 (1.01) unstable; urgency=low

  * Handle obsolete /System.map and /System.old links left around by older

    kernel-package packages. All the programs that look at the information

    in the map files (including top, ps, and klogd) also will look at

    /boot/System.map-<version>, we just need to ensure that that file is

    present, and no longer require the symbolic link. 

    Actually, having the symbolic link in / is technically detrimental

    (apart from cluttering up /); many programs, though looking in /boot,

    still allow /System.map to override. If you use LILO to choose between

    multiple kernels, then the /System.map symbolic link only applies to

    one such kernel, for all other choices the symbols loaded will be

    wrong. Not having the symbolic links at all prevents this. 

    Therefore, the new image.postinst file shall offer to remove the

    symbolic links in /. This should fix BUG#13359

 -- Manoj Srivastava <srivasta@debian.org>  Fri, 26 Sep 1997 10:44:39 -0500

kernel-source-0.21 (1.01) unstable; urgency=low

  * Fixed handling of modules.dep in the image.postinst. We do not attempt

    to recreate a modules.dep, since the man page admits the file so

    created may be incorrect. We warn the installer that there maybe

    problems loading modules into the kernel until reboot iff the version

    being installed is the same as the version currently running.

 -- Manoj Srivastava <srivasta@debian.org>  Tue, 16 Sep 1997 15:07:02 -0500

kernel-source-0.20 (1.01) unstable; urgency=low

  * No longer create a System.map symlink in /, since that may confuse

    klogd when choosing kernel images using LILO. Since top, ps, and klogd

    look at /boot/System.map-<version>, we just need to make sure that

    file is present. This makes us friendlier to multiple images of the

    same kernel version.

  * No longer redirect output to a file in /tmp for security reasons (we

    use a log file in /var/log instead). This fixes BUG#11765,

    BUG#11766 and BUG#11847 

  * Added support for different flavours of the same kernel version for

    people who need them. This is based on the ideas and work of Bill

    Mitchell <mitchell@mozcom.com> and Noel Maddy <ncm@biostat.hfh.edu>.

    This should make us fully compliant to having multiple flavours of the

    same kernel version. 

  * Added dependencies to targets in rules. Now things should work as

    expected if one edits a .config file. 

  * Fixed destination for the Buildinfo file. This fixes BUG#11884.

 -- Manoj Srivastava <srivasta@debian.org>  Mon,  4 Aug 1997 13:03:51 -0500

kernel-source-0.19 (1.01) unstable; urgency=low

  * All kernel packages produced now list the version of kernel-package

    used in the file /usr/doc/Buildinfo.

  * The image prerm will allow you to remove an running kernel image and

    hose your system. You will be warned. (under protest).

  * Fixed typo in control file for kernel-doc description. This fixes

    Bug#11568.

 -- Manoj Srivastava <srivasta@debian.org>  Tue, 29 Jul 1997 17:47:41 -0500

kernel-source-0.18 (1.01) unstable; urgency=low

  * No longer carries around an extra uncompressed kernel image, and does

    not anymore create /boot/psdatabase-X.X.XX. The psdatabase file does

    not seem to be required anymore.

 -- Manoj Srivastava <srivasta@debian.org>  Wed, 18 Jun 1997 13:05:33 -0500

kernel-source-0.17 (1.01) unstable; urgency=low

  * Added patches to support m68k from "Frank Neumann"

    <Frank.Neumann@Informatik.Uni-Oldenburg.DE>.

  * Added patched to support sparcs from Eric Delaunay

    <delaunay@lix.polytechnique.fr> 

 -- Manoj Srivastava <srivasta@debian.org>  Mon,  2 Jun 1997 15:17:59 -0500

kernel-source-0.16 (1.01) unstable; urgency=low

  * Moved config to /boot, where it arguably should have gone to in the

    first place. The /boot directory contains other information pertinent 

    to the kernel, such as the System.map file, and the psdatabase.  The 

    information about exactly what was configured into the kernel was 

    missing, and it can get quite critical on some machines.  Also, the 

    config file may serve as a base for compiling the next kernel. This 

    file is not really a configuration file (not when packaged as part of 

    the kernel-image package), hence it does not belong in /etc (no amout 

    of changing this file will have any affet on system behaviour).

 -- Manoj Srivastava <srivasta@debian.org>  Wed, 21 May 1997 01:44:17 -0500

kernel-source-0.15 (1.01) unstable; urgency=low

  * Kernel-image and kernel-doc now suggest each other.

  * Also recognize powerpc as a synonym for ppc in determining whether

    we use zImage or bzImage by default.

  * Fixed up some typos in documentation

  * Added rules target kernel_doc. This fixes BUG#9138

  * Also install .config file under /usr/doc. This fixes BUG#9298

 -- Manoj Srivastava <srivasta@debian.org>  Fri, 2 May 1997 14:34:51 -0500

kernel-source-0.14 (1.01) unstable; urgency=low

  * No longer install text files in the modules directory, since depmod

    in modutils-2.1.34 fails when it finds a non-ELF file in modules

    directory. This fixes Bug#9243.

 -- Manoj Srivastava <srivasta@debian.org>  Wed, 30 Apr 1997 15:24:51 -0500

kernel-source-0.13 (1.01) unstable; urgency=low

  * Really add in all the changes sent in by Herbert Xu. The changes are: 

    * Changed to source package name back to kernel-source-2.4.26 again. 

    * Changed the installs to be without -o root -g root since it is no

      longer useful as the chowns are done before the packages are built. It

      also means that if it is used in targets like stamp-configure which may

      be run by anyone it won't not generate an error.

    * Made rules file generate the control file.

    * Fixed some typos in chown commands while generating the header

      packages. 

 -- Manoj Srivastava <srivasta@debian.org>  Mon, 21 Apr 1997 15:12:19 -0500

 kernel-source-0.12 (1.01) unstable; urgency=low

  * Forgot to mention that the source package for the kernel packages 

    (produced by make-kpkg buildpackage, for example) has been changed to 

    not contain the kernel version as part of the name (all the deb files 

    produced have not changed), so we now get kernel-image-xxx_yy_zz.deb, 

    kernel-source-xxx_yy_zz.deb, and kernel-headers-xxx_yy_zz.deb, along

    with kernel-source.tar.gz and kernel-source_yy.diff.gz

  * Added changes from the kernel packages maintainer (this is the

    kernel-packages maintainer speaking, Herbert Xu handles the kernel

    image, headers, and source packages at the moment).

 -- Manoj Srivastava <srivasta@debian.org>  Sun, 13 Apr 1997 00:03:36 -0500

 kernel-source-0.11 (1.01) unstable; urgency=low

  * set umask to copy the kernel source files untainted by package creators

    umask. 

  * set umask to copy the kernel header files untainted by package creators

    umask.

 -- Manoj Srivastava <srivasta@debian.org>  Tuesday, 1 Apr 1997 11:04:42 -0600

 kernel-source-0.10 (1.01) unstable; urgency=low

  * image postinst no longer aborts when the user gives up on formatting a

    floppy, but is offered a choice to proceed with a preformatted floppy. 

  * Fixed spelling errors in kernel image postinst

  * Offer to user superformat if it exists.

  * Tested boot floppy creation.

  * Tested on 2.0.29, 2.1.29, and 2.1.30

  * Mention setfdprm in postinst if fail to format floppy.

  * Fix a typo in code determining which floppy drive to use

  * Allow leading whitespace in responses

 -- Manoj Srivastava <srivasta@debian.org>  Sun, 23 Mar 1997 22:53:13 -0600

kernel-source-0.9 (1.01) unstable; urgency=low

  n Stupid @#$%$%@! vi changed Feb to February in the changelog messing up

    gencontrol. 

 -- Manoj Srivastava <srivasta@debian.org>  Mon, 17 Feb 1997 19:29:02 -0600

kernel-source-0.8 (1.01) unstable; urgency=low

  * Removed extra root checks

  * Added error messages to failed root check

 -- Manoj Srivastava <srivasta@debian.org> Thursday, 13 February 1997 14:47:06 -0600

kernel-source-0.7 (1.01) unstable; urgency=low

  * Fixed kernel rules file so that one does not depend on the existence of modules

 -- Manoj Srivastava <srivasta@debian.org>  Tuesday, 28 January 1997 22:15:27 -0600

kernel-source-0.6 (1.01) unstable; urgency=low

  * changed priority of kernel-source package to optional

  * changed priority of kernel-image package to optional

 -- Manoj Srivastava <srivasta@debian.org>  Fri, 22 Nov 1996 11:02:31 -0600

kernel-source-0.5 (1.01) unstable; urgency=low

  * kernel-source-X.XX now no longer recommends bin86, since bin86 is only

    available intel platforms.  It now merely mentions it in the

    description.

 -- Manoj Srivastava <srivasta@debian.org>  Thu, 7 Nov 1996 22:17:47 -0600

kernel-source-0.4 (1.01) unstable; urgency=low

  * kernel-source-X.XX now recommends bin86, which fixes Bug#4443

  * Added short, succinct (and potentially dangerous) instructions on

    compiling kernels at the top of the README file.  This fixes Bug#4445.

 -- Manoj Srivastava <srivasta@debian.org>  Wed, 6 Nov 1996 23:59:47 -0600

kernel-source-0.3 (1.01) unstable; urgency=low

  * Changed everything to new packaging standards.

  * Major re-organization.

  * Added explanation for kernel-headers-X.XX (and why we need this

    package inspite of having headers in libc5-dev)

  * Added README files for all packages (the old README files became

    copyright files).

  * Added target for buildpackage (wich needed a hack in the target clean

    so as not to remove the ./debian directory prematurely)

 -- Manoj Srivastava <srivasta@debian.org>  Tue, 5 Nov 1996 22:42:12 -0600

kernel-source-0.2 (1.01) unstable; urgency=low

  * Made image.postinst more polite.

  * Also enabled kernel-{source,image,headers} targets in kernel.rules 

  * Changed make-kpkg to also accept the new targets.

  * Fixed typo in man page.

  * The man page says it is an administrators man page rather than a 

    programmers man page.

  * Recommended libc-dev for kernel-source and kernel-package

  * source suggests ncurses3.0-dev, tk40-dev and the description explains 

    you could use make menuconfig rather than plain old make config if you

    had these packages.

  * Fixed typo in the rules for modules in /usr/src/modules/ directory

  * Made the rules for such modules ignore errors (since they are not

    really a part of this package, error there should not halt this build

    (or so people seem to want it)

  * Look for modules in the kernel config file (or the default config file

    if the user has not supplied a config file), and only make modules and

    install them if modules have been configured in the config file. This 

    could be tested better.

  * Fixed the make-kpkg.8 man page so that it now does not seem to mandate

    a source and diff file for additional modules installed under the 

    directory /usr/src/modules/ (whether these files are produced is at

    the discretion of the maintainer of that modules package.

  * Make configure depend on stamp-configure which in turn depends on

    .config Hopefully, this will remake the image target if one changes

    the config file Hopefully, this will not cause the image target to

    build needlessly. 

 -- Manoj Srivastava <srivasta@debian.org>  

kernel-source-0.1 (1.01) unstable; urgency=low

  * Changed the scripts to refer to /usr/bin/perl rather than /bin/perl.

  * Added an extended description to the image control file.

  * Added a note that the dist target requires a PGP key.

  * Fixed a typo (missing DEBDIR) in debian.rules.

  * Fixed the targets expected by make-kpkg (kernel_image rather than 

    kernel-image, etc).

  * In image.postinst, made arguments to system be a single argument if  

    the arguments contain shell meta-characters, (this way perl passes

    them to sh -c).

  * Made make-kpkg accept non floats as revision numbers, to facilitate  

    local names.

  * Fixed silly bug in makefloppy in image.postinst.

  * Fixed the extended description of the kernel-package package.

  * Updated the image postinst to install the mbr, if it exists, and to

    activate the root partition, and to not overwrite the mbr (oops).

  * Changed maintainer email address to debian.org (I'll be in a state of  

    flux soon)

 -- Manoj Srivastava <srivasta@pilgrim.umass.edu>  

Local variables:

mode: debian-changelog

End:

Source: kernel-source-2.4.26

Section: devel

Priority: optional

Maintainer:  Edgar Hucek <hostmaster@ed-soft.at>

Standards-Version: 3.5.9.0

Package: kernel-headers-2.4.26

Architecture: any

Section: devel

Priority: optional

Provides: kernel-headers, kernel-headers-2.4

Description: Header files related to Linux kernel version 2.4.26

 This package provides kernel header files for version 2.4.26, for sites 

 that want the latest kernel headers. Please read

 /usr/share/doc/kernel-headers-2.4.26/debian.README.gz for details

Package: kernel-source-2.4.26

Architecture: all

Section: devel

Priority: optional

Provides: kernel-source, kernel-source-2.4

Depends: binutils, bzip2

Recommends: libc-dev, gcc, make

Suggests: libncurses-dev | ncurses-dev, kernel-package

Description: Linux kernel source for version 2.4.26

 This package provides the source code for the Linux kernel version 2.4.26.

 .

 You may configure the kernel to your setup by typing "make config"

 and following instructions, but you could get ncursesX.X-dev and try

 "make menuconfig" for a jazzier, and easier to use interface. There 

 are options to use QT or GNOME based configuration interfaces, but they

 need additional packages to be installed. Also, please read the detailed 

 documentation in the file

 /usr/share/doc/kernel-source-2.4.26/README.headers.gz.

 .

 If you wish to use this package to create a custom Linux kernel, then

 it is suggested that you investigate the package kernel-package,

 which has been designed to ease the task of creating kernel image

 packages. 

Package: kernel-image-2.4.26

Architecture: i386

Section: base

Priority: optional

Provides: kernel-image,  kernel-image-2.4

Depends:  coreutils | fileutils (>= 4.0)

Suggests: lilo (>= 19.1) | grub,  fdutils, kernel-doc-2.4.26 | kernel-source-2.4.26 

Description: Linux kernel binary image for version 2.4.26.

 This package contains the Linux kernel image for version 2.4.26,

 the corresponding System.map file, and the modules built by the

 packager.  It also contains scripts that try to ensure that the

 system is not left in a unbootable state after an update.

 .

 If you wish to update a bootdisk, or to use a bootloader to make

 installing and using the image easier, we suggest you install the latest

 fdutils (for formatting a floppy to be used as boot disk), and LILO, for a

 powerful bootloader. Of course, both these are optional. 

 .

 Kernel image packages are generally produced using kernel-package,

 and it is suggested that you install that package if you wish to

 create a custom kernel from the sources. Please look at kernel-img.conf(5),

 and /usr/share/doc/kernel-package/README.gz  from the package kernel-package

 for details on how to tailor the  installation of this or any other kernel 

 image package

Package: kernel-doc-2.4.26

Architecture: all

Section: doc

Priority: optional

Provides: kernel-doc-2.4

Suggests: kernel-image-2.4.26

Description: Linux kernel specific documentation for version 2.4.26.

 This package provides various Readme's in the 2.4.26 kernel

 Documentation/ subdirectory: these typically contain kernel-specific

 installation notes for some drivers. for example. Please see

 /usr/share/doc/kernel-doc-X.X.XX/Documentation/00-INDEX for a list of

 contents.  Please also read the Changes file, as it contains

 information about the problems which may result by upgrading your

 kernel.

kernel-image-2.4.26_1_i386.deb base optional

#! /usr/bin/make -f

############################ -*- Mode: Makefile -*- ###########################

## debian.rules ---

## Author	    : Manoj Srivastava ( srivasta@pilgrim.umass.edu )

## Created On	    : Sat Apr 27 06:36:31 1996

## Created On Node  : melkor.pilgrim.umass.edu

## Last Modified By : Manoj Srivastava

## Last Modified On : Sun Jan 11 11:54:53 2004

## Last Machine Used: glaurung.internal.golden-gryphon.com

## Update Count	    : 655

## Status	    : Unknown, Use with caution!

## HISTORY	    :

## Description	    :

## $Id: rules,v 1.227 2003/10/07 16:24:20 srivasta Exp $

###############################################################################

ifdef KPKG_ARCH

  ifneq ($(strip $(KPKG_ARCH)),um)

    ha:=-a$(KPKG_ARCH)

  endif

endif

DPKG_ARCH := dpkg-architecture

# set the dpkg-architecture vars

export DEB_BUILD_ARCH      := $(shell $(DPKG_ARCH) $(ha) -qDEB_BUILD_ARCH)

export DEB_BUILD_GNU_CPU   := $(shell $(DPKG_ARCH) $(ha) -qDEB_BUILD_GNU_CPU)

export DEB_BUILD_GNU_TYPE  := $(shell $(DPKG_ARCH) $(ha) -qDEB_BUILD_GNU_TYPE)

export DEB_HOST_ARCH       := $(shell $(DPKG_ARCH) $(ha) -qDEB_HOST_ARCH)

export DEB_HOST_GNU_CPU    := $(shell $(DPKG_ARCH) $(ha) -qDEB_HOST_GNU_CPU)

export DEB_HOST_GNU_SYSTEM := $(shell $(DPKG_ARCH) $(ha) -qDEB_HOST_GNU_SYSTEM)

export DEB_HOST_GNU_TYPE   := $(shell $(DPKG_ARCH) $(ha) -qDEB_HOST_GNU_TYPE)

export DEB_BUILD_GNU_SYSTEM:= $(shell $(DPKG_ARCH) $(ha) -qDEB_BUILD_GNU_SYSTEM)

#

# VERSION=$(shell LC_ALL=C dpkg-parsechangelog | grep ^Version: | \

#                          sed 's/^Version: *//')

#

# The version of kernel-package this belongs to

kpkg_version := 8.079-1xebian

# The maintainer information.

maintainer = Herbert Xu

email= herbert@greathan.apana.org.au

pgp=$(maintainer)

#  Package specific stuff

# decide if image is meant to be in /boot rather than /

link_in_boot :=

# Can we use symlinks?

no_symlink :=

# If so, where is the real file (usually, vmlinuz-X.X.X is real, and

# vmlinuz is the link, this variable reverses it.

reverse_symlink :=

# The version numbers for kernel-image, kernel-headers and

# kernel-source are deduced from the Makefile (see below,

# and footnote 1 for details)

# Whether to look for and install kernel patches by default.

# Be very careful if you do this.

patch_the_kernel := AUTO

# do not create libc headers by default

make_libc_headers := NO

# run make clean after build

do_clean := YES

# install uncompressed kernel ELF-image (for oprofile)

int_install_vmlinux := NO

# what kernel config target to run in our configure target.

config_target := oldconfig

# The default architecture (all if architecture independent)

CROSS_ARG:=

ifdef KPKG_ARCH

  architecture:=$(KPKG_ARCH)

else

  #architecture:=$(shell CC=$(HOSTCC) dpkg --print-gnu-build-architecture)

  architecture:=$(DEB_HOST_ARCH)

endif

ifndef CROSS_COMPILE

  ifneq ($(strip $(KPKG_ARCH)),um)

    ifneq ($(strip $(architecture)),$(strip $(DEB_BUILD_ARCH)))

      KERNEL_CROSS:=$(architecture)-$(strip $(DEB_HOST_GNU_SYSTEM))-

    endif

  endif

else

  KERNEL_CROSS:=$(CROSS_COMPILE)-

endif

ifneq ($(strip $(KERNEL_CROSS)),)

  CROSS_ARG:=CROSS_COMPILE=$(KERNEL_CROSS)

endif

KERNEL_ARCH:=$(architecture)

DEBCONFIG = $(CONFDIR)/config

IMAGEDIR=/boot

INT_IMAGE_DESTDIR=debian/tmp-image$(IMAGEDIR)

VERSION =$(shell egrep '^VERSION +=' Makefile 2>/dev/null | \

 sed -e 's/[^0-9]*\([0-9]*\)/\1/')

PATCHLEVEL =$(shell egrep '^PATCHLEVEL +=' Makefile 2>/dev/null | \

 sed -e 's/[^0-9]*\([0-9]*\)/\1/')

SUBLEVEL =$(shell egrep '^SUBLEVEL +=' Makefile 2>/dev/null | \

 sed -e 's/[^0-9]*\([0-9]*\)/\1/')

EXTRA_VERSION =$(shell egrep '^EXTRAVERSION +=' Makefile 2>/dev/null | \

 sed -e 's/EXTRAVERSION *= *\([^ \t]*\)/\1/')

HAVE_NEW_MODLIB =$(shell egrep '\(INSTALL_MOD_PATH\)' Makefile 2>/dev/null )

ifneq ($(strip $(EXTRA_VERSION)),)

HAS_ILLEGAL_EXTRA_VERSION =$(shell                         \

    perl -e '$$i="$(EXTRA_VERSION)"; $$i !~ m/^[a-z\.\-\+][a-z\d\.\-\+]*$$/o && print YES;')

  ifneq ($(strip $(HAS_ILLEGAL_EXTRA_VERSION)),)

    $(error Error: The EXTRAVERSION may only contain lowercase alphanumerics \

 and  the  characters  - +  . The current value is: $(EXTRA_VERSION). Aborting.)

  endif

endif

######################################################################

###          Architecture specific stuff                           ###

######################################################################

# Each architecture has the following specified for it

# (a) The kernel image type (i.e. zImage or bzImage)

# (b) The dependency on a loader, if any

# (c) The name of the loader

# (d) The name of the documentation file for the loader

# (e) The build target

# (f) The location of the kernelimage source

# (g) The location of the kernelimage destination

# (h) The name of the arch specific configuration file

# Some architectures has sub architectures

### m68k

ifeq ($(strip $(architecture)),m68k)

  ifeq (,$(findstring /$(KPKG_SUBARCH)/,/amiga/atari/mac/mvme147/mvme16x/bvme6000/))

    GUESS_SUBARCH:=$(shell awk '/Model/ { print $$2}' /proc/hardware)

    ifneq (,$(findstring Motorola,$(GUESS_SUBARCH)))

     GUESS_SUBARCH:=$(shell awk '/Model/ { print $$3}' /proc/hardware)

     ifneq (,$(findstring MVME147,$(GUESS_SUBARCH)))

      KPKG_SUBARCH:=mvme147

     else

      KPKG_SUBARCH:=mvme16x

     endif

    else

     ifneq (,$(findstring BVME,$(GUESS_SUBARCH)))

      KPKG_SUBARCH:=bvme6000

     else

      ifneq (,$(findstring Amiga,$(GUESS_SUBARCH)))

       KPKG_SUBARCH:=amiga

      else

       ifneq (,$(findstring Atari,$(GUESS_SUBARCH)))

        KPKG_SUBARCH:=atari

       else

        ifneq (,$(findstring Mac,$(GUESS_SUBARCH)))

         KPKG_SUBARCH:=mac

        endif

       endif

      endif

     endif

    endif

  endif

  NEED_DIRECT_GZIP_IMAGE=NO

  kimage := zImage

  target = $(kimage)

  kimagesrc = vmlinux.gz

  kimagedest = $(INT_IMAGE_DESTDIR)/vmlinuz-$(version)

  DEBCONFIG = $(CONFDIR)/config.$(KPKG_SUBARCH)

  ifneq (,$(findstring $(KPKG_SUBARCH),mvme147 mvme16x bvme6000))

    loaderdep=vmelilo

    loader=vmelilo

    loaderdoc=VmeliloDefault

  else

    loaderdep=

    loader=lilo

    loaderdoc=

  endif

endif

### ARM

ifeq ($(strip $(architecture)),arm)

  GUESS_SUBARCH:='netwinder'

  ifneq (,$(findstring $(KPKG_SUBARCH),netwinder))

    KPKG_SUBARCH:=$(GUESS_SUBARCH)

    kimage := zImage

    target = Image

    kimagesrc = arch/$(KERNEL_ARCH)/boot/Image

    kimagedest = $(INT_IMAGE_DESTDIR)/vmlinux-$(version)

    loaderdep=

    loader=nettrom

    loaderdoc=

    NEED_DIRECT_GZIP_IMAGE=NO

    DEBCONFIG= $(CONFDIR)/config.netwinder

  else

    kimage := zImage

    target = zImage

    NEED_DIRECT_GZIP_IMAGE=NO

    kimagesrc = arch/$(KERNEL_ARCH)/boot/zImage

    kimagedest = $(INT_IMAGE_DESTDIR)/vmlinuz-$(version)

    DEBCONFIG = $(CONFDIR)/config.arm

  endif

endif

### PowerPC

ifneq ($(strip $(filter ppc powerpc,$(architecture))),)

  ifeq ($(DEB_BUILD_ARCH),powerpc)

  # This is only meaningful when building on a PowerPC

    ifeq ($(GUESS_SUBARCH),)

      GUESS_SUBARCH:=$(shell awk '/machine/ { print $$3}' /proc/cpuinfo)

      ifneq (,$(findstring Power,$(GUESS_SUBARCH)))

        GUESS_SUBARCH:=pmac

      else

        ifneq (,$(findstring iMac,$(GUESS_SUBARCH)))

          GUESS_SUBARCH:=pmac

        endif

      endif

    else

      GUESS_SUBARCH:=pmac

    endif

  endif

  ifeq (,$(findstring $(KPKG_SUBARCH),apus prpmc chrp mbx pmac prep Amiga APUs CHRP MBX PReP chrp-rs6k ))

    KPKG_SUBARCH:=$(GUESS_SUBARCH)

  endif

  KERNEL_ARCH:=ppc

  ifneq (,$(findstring $(KPKG_SUBARCH),APUs apus Amiga))

    KPKG_SUBARCH:=apus

    loader := NoLoader

    kimage := vmapus.gz

    target = zImage

    kimagesrc = $(shell if [ -d arch/$(KERNEL_ARCH)/boot/images ]; then \

	echo arch/$(KERNEL_ARCH)/boot/images/vmapus.gz ; else \

	echo arch/$(KERNEL_ARCH)/boot/$(kimage) ; fi)

    kimagedest = $(INT_IMAGE_DESTDIR)/vmlinuz-$(version)

    DEBCONFIG = $(CONFDIR)/config.apus

  endif

  ifneq (,$(findstring $(KPKG_SUBARCH),chrp-rs6k))

    KPKG_SUBARCH:=chrp-rs6k

    loaderdep=quik

    loader=quik

    loaderdoc=QuikDefault

    kimage := zImage

    target = $(kimage)

    kimagesrc = $(shell if [ -d arch/$(KERNEL_ARCH)/chrpboot ]; then \

	echo arch/$(KERNEL_ARCH)/chrpboot/$(kimage) ; else \

	echo arch/$(KERNEL_ARCH)/boot/images/$(kimage).chrp-rs6k ; fi)

    kimagedest = $(INT_IMAGE_DESTDIR)/vmlinuz-$(version)

    DEBCONFIG = $(CONFDIR)/config.chrp

  endif

  ifneq (,$(findstring $(KPKG_SUBARCH),CHRP chrp))

    KPKG_SUBARCH:=chrp