o  Make MCE compile time optional	(Paul Gortmaker)
o  P4 thermal trip monitoring.		(Zwane Mwaikambo)
o  Non-fatal MCE logging.		(Me)

From Paul via 2.4

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

These are mostly Cyrix-alike, but for some quirks we work around.

Still a few out there.. Most of these from 2.4

Some from 2.4, some from the kernel janitor team,..

From Mikael via 2.4.

From Tigran via 2.4

The usual search and replace type operations from various people
to various drivers..

The second hunk also contains a fix from Badari Pulavarty to make
page_to_phys() work on pages >4GB. Without this, we truncate the
physical address to 32bit

Original from Randy Dunlap <rddunlap@osdl.org>...

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

From 2.4

There still seem to be some casualties since the include file shakeup a
few revisions back, this adds quite a few missing ones, more to come.

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

Makes CONFIG_HIGHPTE an option instead of derived.
Original from Steven Cole.

These things have USB IRQ routing problems we can work around..

- install .config in /boot
- provide kernel-drm package if drm is compiled
- handle extra '-' marks so that linux-2.5.8-dj3-voon works.

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

Because 256 is rounder than 255 I guess 8-)

The element is dead, so update the docs to reflect reality.

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

From Oleg@namesys.

malloc.h is no more.

From Christoph Hellwig

http://oss.software.ibm.com/pipermail/jfs-patches/2002-March/000045.html

In light of the 'lets use this for shutdowns' this will eventually
be ripped out, but for now it makes sense to include it so that we
can build it again on modern binutils

We always returned success even when we had no ->vm_ops

Original from Badari Pulavarty <pbadari@us.ibm.com>:

pci_alloc_consistent() is returning zone DMA memory to highmem
enabled drivers when it really should have been returning zone NORMAL.

Found this while testing qlogicfc driver for > 4GB support.

originally from John Levon <levon@movementarian.org>

Define KBUILD_BASENAME for init/do_mounts

Handles things like 2.5.7-dj2-extra-foo-voon in Kernelversion.