Here's a very small diff for 945GM support for agpgart.

Patch against 2.6.15.

From: Alan Hourihane <alanh@fairlite.demon.co.uk>
Signed-off-by: NDave Jones <davej@redhat.com>

Semaphore to mutex conversion.

The conversion was generated via scripts, and the result was validated
automatically via a script as well.
Signed-off-by: NArjan van de Ven <arjan@infradead.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NDave Jones <davej@redhat.com>

Add suspend/resume support for the ati-agp module
Signed-off-by: NJaco Kroon <jaco@kroon.co.za>
Acked-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NDave Jones <davej@redhat.com>

This adds support for suspend/resume to the amd64-agp driver. Without
it, X displays garbage after resume from swsusp.
Signed-off-by: NMichal Schmidt <xschmi00@stud.feec.vutbr.cz>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NDave Jones <davej@redhat.com>

The loop contains a command that is only used in the last iteration. I moved the command outside the loop.
Compile-tested

Signed-off-by: Daniel Marjamäki <daniel.marjamaki at comhem.se>
Signed-off-by: NDave Jones <davej@redhat.com>

Fix a comment which missed an update cycle somewhere.
Signed-off-by: NJonathan Corbet <corbet@lwn.net>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Looks like fallout from the geo stuff.

The feature removal was done in 7301c8d3
but these files were not removed for some reason.
Signed-off-by: NJody McIntyre <scjody@modernduck.com>

like other routines here, to ensure buffers are correctly initialised
with respect to b_private/b_end_io.  Fixes an odd interaction between
XFS and reiserfs.
Signed-off-by: NNathan Scott <nathans@sgi.com>

The comparison of the initrd start address against "&_end" is
unnecessary and incorrect.  Make it match the x86 code that just
compares the passed-in arguments.
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This prevents running out of GSIs on large Unisys ES7000 machines.
Follows i386

Cc:  "Protasevich, Natalie" <Natalie.Protasevich@unisys.com>
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

For not fully explained reasons it broke mem=... on several setups.

Also minor cleanup.

Cc: axboe@suse.de
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

'check_efer' uses 'do_not_nx'. Hotpluged CPU could wrongly disable NX.

Signed-off-by: Shaohua Li<shaohua.li@intel.com>
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

uses correct lapic base address. The set_fixmap appears useless.

Signed-off-by: Shaohua Li<shaohua.li@intel.com>
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

To avoid mistakes.

I got a few reports where people got broken timing because they didn't
have the PMTIMER fallback.
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This unbreaks recursive kprobes which didn't work anymore
due to an earlier patch which converted the debug entry point
to use an IST.

This also allows nesting of the debug entry point too.
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

cpufreq init can be called when a CPU is set online.
Need to make powernow-k8's initialisation functions __cpuinit to
prevents oopses when a CPU is off/onlined on a AMD system

Cc: trenn@suse.de
Cc: mark.langsdorf@amd.com
Cc: davej@redhat.com
Signed-off-by: NAndi Kleen <ak@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The command 'make all modules_install install' would fail
in a virgin tree - pointing at a non-existing directory under
/lib/modules/xxx

KERNELRELEASE is part of MODLIB and we need to create .kernelrelease
before we can properly evaluate KERNELRELEASE,
Changing MODLIB to the recursively expanded flavor let it pick up
the correct KERNELRELEASE value.

Reported by: "Hemmann, Volker Armin" <volker.armin.hemmann@tu-clausthal.de>
Signed-off-by: NSam Ravnborg <sam@ravnborg.org>

To enable 'make kernelrelease' earlier now create .kernelrelease when
one of the *config targets are used.
Also introduce KERNELVERSION - only user is kconfig.
KERNELVERSION was needed to display kernel version in menuconfig -
KERNELRELEASE is not valid until configuration has completed.
kconfig files modified to use KERNELVERSION.
Bug reported by: Rene Rebe <rene@exactcode.de>
Signed-off-by: NSam Ravnborg <sam@ravnborg.org>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NNathan Scott <nathans@sgi.com>

Signed-off-by: NNathan Scott <nathans@sgi.com>

context.

SGI-PV: 946762
SGI-Modid: xfs-linux-melb:xfs-kern:24983a
Signed-off-by: NNathan Scott <nathans@sgi.com>

Add a MAINTAINER entry for the watchdog device drivers.
Signed-off-by: NWim Van Sebroeck <wim@iguana.be>

The following makes drivers/char/watchdog/sa1100_wdt.c sparse clean.
Signed-off-by: NIan Campbell <icampbell@arcom.com>
Signed-off-by: NWim Van Sebroeck <wim@iguana.be>

The way multiple targets was handled with make O=...
broke because for each high-level target make spawned
a parallel make resulting in a broken build.
Reported by Keith Owens <kaos@ocs.com.au>
Signed-off-by: NSam Ravnborg <sam@ravnborg.org>

For some reason, the BITS_PER_LONG == 64 case of atomic_long_set
was using atomic_set instead of atomic64_set. This does not jive
with architectures which use an inline instead of a #define to
implement their atomic_set() primitives.
Signed-off-by: NKyle McMartin <kyle@parisc-linux.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Fix the following warning which was introduced in 2.6.15-git8 by
commit 7408187d:

  CC [M]  drivers/media/video/zoran_card.o
drivers/media/video/zoran_card.c: In function `zr36057_init':
drivers/media/video/zoran_card.c:1053: warning: assignment makes integer from pointer without a cast
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Acked-by: NRonald S. Bultje <rbultje@ronald.bitfreak.net>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

asm/mach/arch.h introduced a __deprecated, but didn't include compiler.h,
causing:

In file included from arch/arm/mach-at91rm9200/devices.c:13:
include/asm/mach/arch.h:23: warning: no semicolon at end of struct or union
include/asm/mach/arch.h:23: error: syntax error before 'phys_ram'
include/asm/mach/arch.h:34: error: syntax error before ':' token
include/asm/mach/arch.h:35: error: syntax error before ':' token
include/asm/mach/arch.h:36: error: syntax error before ':' token
include/asm/mach/arch.h:37: error: syntax error before ':' token
include/asm/mach/arch.h:45: error: syntax error before '}' token

Add the necessary include.
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

scripts/kconfig/lxdialog/check-lxdialog.sh uses gcc to check for
what libraries are present. Redirect output to /dev/null
so we do not generate an a.out.
Also included support for ncursesw - so if present prefer that
instead of ncurses.
The order is now (first is preferred):
1) ncursesw
2) ncurses
3) curses

The latter is to support SunOS.
Signed-off-by: NSam Ravnborg <sam@ravnborg.org>

When I removed the powermac support from arch/ppc/kernel/pci.c,
I overlooked the fact that that file is used in 32-bit ARCH=powerpc
builds.  To prevent problems in future, restore the original version
of that file as arch/powerpc/kernel/pci_32.c, and use that.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This makes ARCH=powerpc the default on 32-bit powerpc machines,
where uname -m returns ppc, as well as on 64-bit powerpc machines.
Most people who would be likely to build their own kernels on
32-bit powerpc machines would be using powermacs or CHRP machines,
both of which are supported with ARCH=powerpc now.  Embedded ppc
developers whose ports haven't been moved over to arch/powerpc
yet will have to explicitly set ARCH=ppc now.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This makes it possible to build kernels for PReP and/or CHRP
with ARCH=ppc by removing the (non-building) powermac support.
It's now also possible to select PReP and CHRP independently.
Powermac users should now build with ARCH=powerpc instead of
ARCH=ppc.  (This does mean that it is no longer possible to
build a 32-bit kernel for a G5.)
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Some versions of gcc generate incorrect code for the inet_check_attr()
function, apparently due to a totally bogus index -> pointer comparison
transformation.

At least "gcc version 4.0.1 20050727 (Red Hat 4.0.1-5)" from FC4 is
affected, possibly others too.

This changes the function subtly so that the buggy gcc transformation
doesn't trigger.
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>