into tetrachloride.(none):/mnt/raid/src/kernel/2.5/agpgart-new

Changes:

  - Convert to a static highlevel handle for all drivers. Gets rid of
    a kmalloc for each driver and consolidates the highlevel handle and
    highlevel ops. Reduces points of failure as well.
  - Move host number allocation to hosts.c, giving all drivers access to
    it.
  - Implemented S800 changes for core and sbp2. Thanks to TI and LaCie
    for hardware to test this. 1394b support coming soon.
  - Convert nodemgr to using a similar device classification hack as
    usb. The old class_num hack I had in place was broken do to its
    removal. This is a much cleaner solution.

Missed parts of the ioctl32 compatibility patch: a few ioctls from
PA-RISC, and missing <linux/compat_ioctl.h> file.

 - kill dups in ide.c and ide-disk.c
 - read_24() is exported so rename it to ide_read_24()
 - add it ide.h

They are unused, not needed and identical to generic ones so kill 'em:
    task_read_24(),
    task_try_to_flush_leftover_data(),
    taskfile_dump_status(),
    taskfile_error().

There's an (fortunately unused) devfs in misc_register currently,
when the name of the miscdevice contains a slash the name is used
as devfs name instead of misc/<name>.  Kill if as we have .devfs_name
for this kind of stuff now.

ioctl32 cleanups are pretty neccessary (we have 6+ copies of 600+
lines tables, all getting slightly out of sync, not speaking about
surrounding code produced by cut-and-paste).

There's three drivers in the tree that workaround the suboptimal
devfs name choice of the misc device layer (/dev/misc/<foo>) using
devfs_mk_symlink.    Switch them to set miscdev.devfs_name instead
to get the right name from the very beginning.

into home.transmeta.com:/home/torvalds/v2.5/linux

into samba.org:/home/paulus/kernel/for-linus-ppc

into home.transmeta.com:/home/torvalds/v2.5/linux

- Remove all uses of unnamed structs and unions in the driver to make
  old and newer gcc versions happy. Makes it a bit uglier IMO but at
  least people will stop hassling me

into cantab.net:/home/aia21/ntfs-2.5

into home.transmeta.com:/home/torvalds/v2.5/linux

into home.transmeta.com:/home/torvalds/v2.5/linux

From: William Lee Irwin III <wli@holomorphy.com>, Robert Love

Add some spinlock protection around the oom-killer state.

From Robert Day, through "Randy.Dunlap" <rddunlap@osdl.org>

This is a patch from Robert Day that does the following:

1) shift menu item in "Processor type and features" menu
2) clean up "Bus options" menu so it's actually hierarchical

Part of it (moving X86_IO_APIC around) looked a little odd to me,
so I asked Roman Zippel about it, and he replied:

"It's correct, although I wouldn't call it a 'design quirk'. :)
It forces one to group options which belong logically together and in this
case X86_IO_APIC is really a bit misplaced, even if it's not visible."

I have tested it (on 2.5.68-plain) and it does indeed make the menus
more hierarchical.

The percpu counters break UML SMP compilation (in current 2.5.58 bk snapshot)
(first NR_CPUS undeclared in header, then dereference of incomplete structure
in .c file)

From: Janet Morgan <janetmor@us.ibm.com>

Here's a small patch that adds aio_read and aio_write methods to the
block device driver.

Fix a bug identified by Nikita Danilov: refill_inactive_zone() is deferring
the update of zone->nr_inactive and zone->nr_active for too long - it needs
to be consistent whenever zone->lock is not held.

From: john stultz <johnstul@us.ibm.com>

- Whitespace fixes

- Fix for the case where HZ != 1000 (pointed out by Mika Penttila).

From: Manfred Spraul <manfred@colorfullife.com>

Update some no-longer-true comments around nr_threads locking.

From: Ernie Petrides <petrides@redhat.com>

The problem resolved by this patch is that if a root file system has an
error recorded from a previous mount, and then (when rebooting) the orphan
recovery procedure is initiated, the recovery is correctly skipped but the
file system is incorrectly left in a writable state.

This causes the subsequent fsck to fail due to the root file system
being dirty, and then requires manual intervention to get the system
fully booted.

From: Paul B Schroeder <paulsch@haywired.net>

The following patch simply moves the 'nr_registered_attrs' and
'device_registered' variables in mwavedd.c into the MWAVE_DEVICE_DATA struct
which is defined in mwavedd.h..

- Remove dead variable from block_read_full_page (Oleg Drokin)

From: Miklos.Szeredi@eth.ericsson.se (Miklos Szeredi)

This fixes a problem with method 0 of proc_file_read (when the whole file
is copied to the page).  The calculation of the final bytecount is wrong,
and hence smaller then page size reads will give a truncated file.

Current 2.4 kernels do it this way as well.

From: Krzysztof Halasa <khc@pm.waw.pl>

Updates the HDLC drivers to the approved module API.

From: Paul Clements <Paul.Clements@SteelEye.com>

Currently, when I build ext3 as a module, jbd gets built into the kernel
proper.  This trivial patch allows jbd to be built as a module when ext3 is
also modular.  I believe this is the intention, as this is how it works in
2.4.  I've built and tested with modular jbd and ext3 on 2.5.68.

From: Dave Olien <dmo@osdl.org>

Christoph submitted a patch to linus last week fixing up some DAC960 driver
entry points.  That patch will OOPS during boot on version 2 controller
types.  Christoph's version of the disk_size() function was dereferencing
a NULL pointer in it's "else" clause.

Christoph's patch hasn't appeared in linus's BK tree yet.  So, I'm
resending Christoph's orignal patch with my fix to disk_size() included.
This patch can be applied to the driver in Linus's BK tree from April 28.

Here's Christoph's original description of his patch:

Some grepping showed that DAC960's open routine was duplicating parts
of check_disk_change().  I went on fixing this by implementing a
media_changed method and making DAC960_Open use it.  While looking
at the surrounding code I noticed that

  (a) all methods weren't using the private data the upperlayer
      hands to it properly, but instead using kdev_t-based indexes
  (b) DAC960_Open/DAC960_Release was keeping never used counters
  (c) DAC960_Open was doing tons of checks the upperlayer already does
  (d) DAC960_Release was entirely superflous.

The patch below corrects that and rewrites the block entry points
into readable code - 100 LOC are gone and the same amount replaced
by readable code.