Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

      paths (fs/ntfs/aops.c::ntfs_{prepare,commit}_write()).  At present
      this only works until the data attribute becomes too big for the mft
      record after which we abort the write returning -EOPNOTSUPP from
      ntfs_prepare_write().
Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

      i_size, i.e. race with truncate, invalidate the buffers on the page
      so that they become freeable and hence the page does not leak.
Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

into cantab.net:/home/src/ntfs-2.6-devel

into flatcap.org:/home/flatcap/backup/bk/ntfs-2.6-devel

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

      general in fs/ntfs/inode.c::ntfs_read_locked_inode() and friends.
Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since xfrm_user can be built as a module it's best not to panic
when something goes wrong in xfrm_user_init.  Call me a wimp for
not fixing netlink_kernel_init's return value :)
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch checks the return value of netlink_insert() in
netlink_kernel_create().  It could fail if someone loads
the same module twice for instance.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch fixes 3 problems with the nommu.c support code:

1. export mem_map and vmtruncate (to make them the same as MMU version)

2. update the total_vm usage in mmap routines
   Without this the OOM killer has no "badness" points to rate
   processes on...

   This patch was originally submitted by Giovanni Casoli
   <giovanni.casoli@atengineering.it> for a 2.4.26 kernel.
   I applied it to the 2.6.9 code.

3. provide stub for arch_get_unmapped_area
Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Add code to support the apply_relocate_add() function, needed to
support all relocation types of the m68k and Coldfire families.
This code is straight out of the m68k support for this.

Patch was originally from Christian Magnusson
<christian.magnusson@runaware.com>
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove remaining unused shglcores.h include.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove unused shglcore.h include.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Clean up the definition of HZ on m68knommu systems.
There is really only a couple of cases to handle, most
boards use the traditional 100Hz still. Only special
case out those that don't. The current code is just
spaghetti.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove unused SHGLCORE support.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove uneeded includes. asm/delay.h is not needed here.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove uneeded includes. asm/delay.h is not needed here.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove uneeded includes. asm/delay.h is not needed here.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove unused shglcore.h include from m68knommu/mm/memory.c
That file will be removed in an upcoming patch.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Remove unused shglcore.h include from m68knommu/mm/init.c
That file will be removed in an upcoming patch.
Signed-off-by: Greg Ungerer <gerg@snapgear.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

      mount time as this cannot work with the current implementation.
Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

Kbuild can't fix up "-I $(src)/.." to the right directory
due to the particular magic transformations it does on -I
arguments. Remove the space, use "-I$(src)/.." and it works.

Hopefully Sam will fix this build issue. 

I think somebody should dress up as scripts/Makefile.lib
for Halloween. Now _that_ would be scary.

Noted by Al Viro, superhacker and endlessly patient person.

Sign up today!

You too can become a superhacker by setting up a cross-compile
environment for all architectures and building "allmodconfig"
trees all day long to find places where others have screwed up.

If an arch Makefile overrides CROSS_COMPILE (e.g. parisc, mips, ...)
then the cc-option call in the main Makefile uses the wrong compiler 
to check for options.
Signed-off-by: Randolph Chung <tausq@debian.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

	FB_INTEL selects AGP_INTEL (i.e. intel-agp.c), which is i386-only.
AGP_INTEL itself depends on X86 && !X86_64 and for a good reason - it doesn't
build elsewhere.

	Added dependency to FB_INTEL itself; as it is allmodconfig had been
broken on everything outside i386.
Signed-off-by: Al Viro <viro@parcelfarce.linux.theplanet.co.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

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

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

Host drivers should now use the scatterlist rather than accessing
the request directly.  This marks the request pointer as being
deprecated so that host driver authors get a chance to fix their
drivers up before we remove this member.

This converts the PXA MCI driver to use the supplied scatter list
rather than accessing the struct request directly, thereby making
this host driver independent of the block layer.

We also clean up the driver to use a consistent driver name for
resources.

Convert MMCI driver to use the scatter list provided by the upper
MMC layers.  This removes all knowledge of struct request from the
MMCI driver.

This is intended to remove block layer knowledge from MMC host
drivers.  Host drivers are expected to use the scatter list for
data transfers rather than accessing the struct request.

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

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

Patch from Nicolas Pitre

- eliminate difference between ARMv3 and ARMv4 versions of io-readsl
- unified little/big endian support
- misc optimisations

Signed-off-by: Nicolas Pitre 

Patch from Tony Lindgren

This patch by Dirk Behme, Todd Poynor et al adds power management
support for OMAP. Suspend works on at least 1510 and 16xx.

Signed-off-by: Tony Lindgren

Patch from Ben Dooks

This patch does the following:

- move the clock initialisation to just after the cpu init
- add the two clocks for the s3c2440
- add dclk0/1 and clkout0/1 definitions
- add list of clocks to the board specific struct
- fix initialisation of the clock control
- update handling of parent clocks for >1 parent

this fixes the following problems:

- serial console with no clock sources
- clocks specific to various boards
- hang if LCD controller enabled at start time

Signed-off-by: Ben Dooks 

Patch from Ben Dooks

Allow power managemnet support for the Simtec EB2410ITX (BAST)

Signed-off-by: Ben Dooks 

Patch from Ben Dooks

Allow the uncompressing messages to be sent to the UART
specified in the kernel configuration, instead of just
UART 0. This includes the Kconfig update to make this
option available all the time, and to rename it's
description text.

This patch also fixes s3c2440 support for the
uncompressor

Signed-off-by: Ben Dooks