I've cross-compiled Linux on i386-netbsdelf2.0 for
arm-linux for quite some time now and everything seems
to be working perfectly except for one minor glitch in
the build process that halts module installation
(needlessly, IMHO). Specifically, if System.map exists
$(DEPMOD) is run ("for convenience" as the comment
says in the Makefile). However, on NetBSD I don't have
$(DEPMOD) available so the command fails and make
exits with non-zero exit status. Please consider the
attached patch to add a check for $(DEPMOD) so that
missing $(DEPMOD) won't halt the whole build process.

From: Tero Niemela <tero_niemela@yahoo.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

I used to be be able to cross compile for i386 on my x86_64 machine,
but recently something (gcc/binutils?) changed, and it stopped working.

Following patch makes cross compile work with:

make ARCH=i386 CFLAGS_KERNEL="-m32" AFLAGS_KERNEL="-m32" bzImage

Without the patch I'm getting the following error:

  SYSCALL arch/i386/kernel/vsyscall-syms.o
/usr/lib/gcc/x86_64-pc-linux-gnu/3.4.3/../../../../x86_64-pc-linux-gnu/bin/ld:
Relocatable linking with relocations from format elf32-i386
(arch/i386/kernel/vsyscall-sysenter.o) to format elf64-x86-64
(arch/i386/kernel/vsyscall-syms.o) is not supported
collect2: ld returned 1 exit status
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

Avoid cluttering the tags/TAGS generated file with $(srctree) in the paths
if this is not needed.

This has two advantages:

- Saving about 20M on the size of the resulting tags file (which are used
  currently to store the absolute path of the file names rather than the
  relative one) when KBUILD_OUTPUT is not set.

- Keeping the tags file valid when the directory is renamed.

No change is done for who does make tags O=..., if this is wanted (I would
find that incommodous and non-typical for a developer, but anyway I've not
ruined functionality in that case).
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

Compiling an allyesconfig kernel straight with a gcc 4.0 snapshot
gives nearly 10k new warnings like:

warning: pointer targets in passing argument 5 of `cpuid' differ in signedness

Since the sheer number of these warnings was too much even for the
most determined kernel janitors (I actually asked ;-) and I don't
think it's a very serious issue to have these mismatches I submitted
an new option to gcc to disable it. It was incorporated in gcc mainline
now.

This patch makes the kernel compilation use it. There are still
quite a lot of new warnings with 4.0 (mostly about uninitialized variables),
but the compile log looks much nicer nnow.
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

the part that got merged got broken on the way.  Here are the fixes:

Move __MODULE_INFO to modparam.h: This macro is used in modparam.h;
there are users who include this header but not module.h.  The latter
includes modparam.h already.

__MODULE_INFO(parmtype, name##type, #name ":" #type) does not evaluate
to __MODULE_INFO(parmtype, footype, "foo:int") as was the idea, but to
__MODULE_INFO(parmtype, fooint, "foo:int") when type is bound to int.
In more complicated cases, we get syntax erros.  Re-introduce the
__MODULE_PARM_TYPE macro; this is cleaner than renaming the type parameter.

Add the parmtype definition which was dropped during the merge to to the
obsolete but still heavily used MODULE_PARM macro.
Signed-off-by: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Sam Ravnborf <sam@ravnbrg.org>

"make help" doesn't show "make randconfig" nor "make config" as options
and the description of oldconfig could be better (IMHO). Patch below adds
the missing targets to the help and updates the description of oldconfig.
Signed-off-by: Jesper Juhl <juhl-lkml@dif.dk>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

The patch below removes some false positives I've observed.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

This patch makes some needlessly global code static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

This patch makes a needlessly global variable static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Randy Dunlap <rddunlap@osdl.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

The kbuild utilities are compiled with absolute patch names, so paths
starting with $RPM_BUILD_ROOT would end up in the binaries.  To avoid
this, remove all references to __FILE__ (directly and indirectly via
assert()).
Signed-off-by: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

This adds a warning when building external modules (M= or SUBDIRS=
syntax) and there is no Module.symvers in the object tree. A missing
Module.symvers is a clear sign that the kernel tree itself was never
compiled. The resulting modules will work, but no symbol version
information will be attached to kernel symbols the module uses (because
that information comes from Module.symvers), and so the module will be
more unsafe.
Futhermore the external module will not record what other modules it is
depended on.

The test works with CONFIG_MODVERSIONS enabled or disabled.
Signed-off-by: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

From: John Kacur <jkacur@rogers.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

From: John Kacur <jkacur@rogers.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

into mars.ravnborg.org:/home/sam/bk/kbuild

Most architectures uses /boot for there kernel image, so let this be reflected
by the kernel.
If INSTALL_PATH shell variable is set then this will have effect.
If INSTALL_PATH is set one the commanline to make like this:
make INSTALL_PATH=/nfs/boot install
then this will override both kbuild and shell variable.

If an arch prefer another default this must be set in the arch Makefile
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

make install is often executed as root or on a different mechine via NFS
To avoid updating vmlinux due to directory changes or similar the install target
for i386 no longer has vmlinux as a prerequisite. 
Now modules_install and install are aligned in this respect.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

When tossing around with different gcc compilers there is no way to tell kbuild
to ignore the new name of the compiler. The new option KBUILD_NOCMDDEP tell
kbuild not to check the commandline for changes.
This should be used with care because the resulting kernel may become inconsistent
if one part is build with 2.96, and another part build with 3.3.4.
So use only when you know what you are doing.

Syntax:
make KBUILD_NOCMDDEP=1

Original request for this feature came from hpa.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>

Combine duplicate code in two smaller 'functions'

Signed-off-by: Sam Ravnborg

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Patch from David Vrabel

Use some #define'd constants in the ixp4xx's arch_reset().

Signed-off-by: David Vrabel
Signed-off-by: Russell King

Signed-off-by: Russell King <rmk@arm.linux.org.uk>

Signed-off-by: Russell King <rmk@arm.linux.org.uk>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Use netdev_priv() and get rid of the PRIV() macro by using the correct
data element. Also eliminate casts, where type is correct.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

When the skfddi driver was converted to the new PCI netdevice interface,
it never got tested with real hardware. The initialization got broken, this
should fix it.
Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

Fixes instances where gfar_read() was invoked in debug codewith a value,
rather than a pointer.
Signed-of-by: Andy Fleming <afleming@freescale.com>
Signed-of-by: Kumar Gala <afleming@freescale.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

From: Steven Rostedt <rostedt@goodmis.org>

If the transmit buffer of the e100 overflowed, then the system would hang.
This was caused because the e100 driver would stop the queue, and
netpoll_send_pkt in netpoll.c would then loop forever.  This is because the
e100 net_poll would never start the queue again after the transmits have
completed.

For those that use the e100 and netconsole, all you need to do is a sysreq
't' to lock up the system.
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

Fixes breakage in getting MAC address on built-in tulip in
alphastation 200 - without delay we are skipping bytes.
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

This patch adds the Intel ICH7R DID's to the ahci.c SATA AHCI driver
for ICH7R SATA support.
Signed-off-by:  Jason Gaston <Jason.d.gaston@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@pobox.com>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into silver.ucw.cz:/home/vojtech/bk/dmitry

Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>

       changing HID->event mappings (via EVIOCS*) in the future and make 
       debugging easier now.
Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>

       rebooting.
Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>

       good idea to discard any character other than ACK/NAK during probe
       it causes missing releases and keys getting "stuck" when a command
       issued on enabled device. The effect is easily demonstrated with
       the following command:
           while true; do xset led 3; xset -led 3; done

       With this change extra characters will be discarded only if device
       has not been marked as "enabled" yet.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>