This implements the possibility for sharing a tag map between queues.
Some (most?) scsi host adapters needs this, and SATA tcq will need it
for some cases, too.

From Daniel Pittman <daniel@rimspace.net>

Using the lra (last recorded address) isn't always reliable on some
drives, so use the regular TOC if it returns 0 capacity.

I haven't been doing much smbfs work recently, but here are some bugfixes:

- Fix module unload (Angus Sawyer).
- Fix the smbfs error handling if kernel_thread() should fail.
- Allow high uids/gids to be used as the fake uid smbfs sets as file owner.

The as-iosched.c source code has said this is here for a long time but
it kept getting lost.

 From: bert hubert <ahu@ds9a.nl>

 LSM: To recap, this patch allows root_plug to work again. It needs functions
 that used to reside in capability.c but linking in capability.c, disabled
 root_plug from loading, as a security module is already present then. This
 patch splits out the functions root_plug needs from capability.c.

 From: John Levon <levon@movementarian.org>

 LSM: Update comments in register_security to reflect reality

 Add LSM maintainer entry

 Update CREDITS with new contact info.

into laptop.osdl.org:/home/torvalds/v2.5/linux

Don't try to clear PCI status on devices which are no longer present.
Only clear PCI status bits which we're interested in.
Use pci_name to report the device name.

The IMPD-1 is a logic module (a stackable module) - rather than
trying to abuse platform devices, create our own bus type to handle
these.

into kernel.bkbits.net:/home/lord/xfs-2.6

SGI Modid: 2.5.x-xfs:slinx:158511a

SGI Modid: 2.5.x-xfs:slinx:158493a

This squashes another compiler warning.

Jamie Lokier's cache testing program discovered a problem which seems
to be present in some CPU write buffers.  This cset allows the kernel
to detect the flaw and activate a workaround to restore the user
space expectations.

Move ARM CPU configuration to arch/arm/mm.  Seperate out the selection
of the abort, cache handling, optimised page copying and TLB handling
from the Makefile, and move it into the configuration system.  This
allows us to select the correct files in arch/arm/mm and pick the
appropriate definitions in include/asm-arm/* based upon a config
symbol rather than a bunch of configuration symbols.

Also add ARM1020E, ARM1022 and ARM1026 CPU support.

into laptop.osdl.org:/home/torvalds/v2.5/linux

Monotonic clock code uses reader/writer lock which is prone to same
starvation problems as we saw with xtime.  This patch changes it to seq_lock
which is faster and won't starve writers in face of lots of readers.

into laptop.osdl.org:/home/torvalds/v2.5/linux

into laptop.osdl.org:/home/torvalds/v2.5/linux

Introduce a new kernel-userspace interface that uses 128-bit file
identifiers instead of the previously used 96-bit fileids. We also
replacing the coda_creds structure with only the fsuid.

This new API has been used by for a couple of months now, people had to
patch their kernels whenever they want to run a current Coda release.
A new Kconfig option is added to fall back on the old API for older Coda
clients and other userspace filesystems that might use our protocol.

The ViceFid structure is really a Coda server datatype and the kernel
really shouldn't have to know its internal structure. Replace all
instances with struct CodaFid with opaque members.

Both userspace and the kernel are testing only the fsuid part of the 
coda_creds credentials structure. Nothing else is really used, so we now
match the code with the actual usage by only passing fsuid around. The
kernel-userspace API is kept compatible (for now).

Use a global 'epoch' counter to invalidate cached permissions instead of
traversing a racy linked list of all known Coda inodes.

  - Remove unused functions and variables.
  - Be a bit more strict with the definition of various types that are
    shared between kernel and userspace.
  - Included a couple of cleanups from Maximilian Attems and Stephen
    Hemminger.

SGI Modid: 2.5.x-xfs:slinx:157534a

A recent change causes pervasive data corruption by over-writing inode
metadata with a word of garbage.  The field, di_rdev, should only be set
for a device inode.

gcc 3.3 complains about this and indeed it can only trigger if
someone magically enlarges __uint8_t :)

SGI Modid: 2.5.x-xfs:slinx:157731a

SGI Modid: 2.5.x-xfs:slinx:157475a

SGI Modid: 2.5.x-xfs:slinx:157531a