Patch from Dave Hansen <haveblue@us.ibm.com>

__pgd_offset() and pgd_offset() are completely different functions.
__pgd_offset() is really just a helper to figure out which entry in a
pgd an address would fall into.   pgd_offset() does all the leg work and
actually fetches the real pgd entry.

pgd_index() is a much saner name for what __pgd_offset() does.  In fact,
we do this:
#define __pgd_offset(address) pgd_index(address)

The attached patch removes all instances of __pgd_offset and just
replaces them with pgd_index.

Compiles with and without PAE on x86.

Well I thought we were shooting down unused inodes a bit aggressively.  With
30,000 inodes total and 1000 unused, the 1000 are getting reclaimed
instantly.

shrink_icache_memory() needs to return the number of reclaimable inodes to
page reclaim, not the total number.

Patch from: Matthew Wilcox <willy@debian.org>

Reduces stack use in random.c by 350 (or maybe 700) bytes by sharing a local
array between two functions.

Patch from Patricia Gaughen <gone@us.ibm.com>

This function was recently uninlined for NUMAQ.  But it used to be a macro,
and modules need it.

Patch from Stephen Cameron <steve.cameron@hp.com>

* Fix driver to wait for firmware to indicate that it is ready.
  (Needed for PCI hotplug case, but for normal warm/cold reboot, by the
  time driver inits, firmware will already be ready.)

This driver is doing a division on a 64 bit valus which is identical to
"jiffies" and is asking for __udiv64.  Cast the 64-bit value back to long.

Patch from: Zwane Mwaikambo <zwane@linuxpower.ca>

Fix up a possible AB/BA deadlock identified by Dawson Engler's latest toy.

Patch from Trond Myklebust <trond.myklebust@fys.uio.no>

This has been floating about for a while, and fixes the oops which people
keep on reporting.

irqdesc_t.lock is taken from interrupts - it needs irq-safe locking from
process context.

Also timer_irq_works() needs to use an unsigned long to hold a jiffies value
(Kai Bankett)

General coding cleanups in io_apic.c:

- make pending_irq_balance_apicid[] static

- reorder several functions so that forward decls are not needed and so
  that `inline' actually works.

- 80 col fixes.

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

into tp1.ruhr-uni-bochum.de:/scratch/kai/kernel/v2.5/linux-2.5.make

Russell King investigated a failure case I introduced: When booting
with "load_ramdisk=1", we use the kernel root= parameter to determine
from what device to get the contents to copy into a ramdisk and then
mount that ramdisk as root.

For the copy to work, /dev/root needs to point to the device to load the
ramdisk from.

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

Add SSP PIO driver - provides an interface for talking to
microcontrollers and such like connected to the SA1110's
SSP port.

into flint.arm.linux.org.uk:/usr/src/linux-bk-2.5/linux-2.5-rmk

Patch from Dominik Brodowski.

This patch is necessary to update the ARM drivers so that they can
live with the new cpufreq governor "userspace" (which I sent to Linus a
moment ago).

A 4 patch set from Dominik Brodowski, tested and fixed up by rmk
for ARM.

(1)

- the system3.c cpufreq notifier couldn't possibly compile before because
      cpufreq_updateminmax has been undefined for ages.
- clean up sa1100fb.c notifier to specify exactly what's left TBD.
- double #include of cpufreq.c in drivers/pcmcia/sa1100_generic.c

(2)

Split up config symbols, add help text for CPU_FREQ_PROC_INTF

(3)

- update the Integrator CPUfreq driver so that it can get along with
  cpufreq policies.
- modularize Integrator CPUfreq driver (all it did with
  !CONFIG_CPU_FREQ was to print out some debug information)

(4)

- update the SA11x0 CPUfreq drivers so that they can get along with
  cpufreq policies and governors.
- update the cpufreq core so that cpufreq_get() returns something
  sensible during the first ->setpolicy or ->target call.

Eliminate have_isa_bridge - use isa_bridge instead.
Clean up PCI bus walking.
Clean up detection of fast back to back capability.
Ensure we set the bridge control register appropriately.
Pin swizzle functions are never called with pin == 0.

This follows our existing PM interfaces.

This provides an element of core PM support to ARM, providing an APM
like user space interface, and separating the core PM support from
the SA1100 PM implementation.

This eliminates some unnecessary branches and instructions in the
TLB flush macros where configurations allow for this to occur.

Kill ide_* resource handling.  Add __ide_mm* macros to use our string
based mmio.

We include a header file in an indirect method, which mkdep didn't
understand.  Now that mkdep is gone, we don't need the hack to make
mkdep do the right thing.

This cuts down on the number of warnings during the build.

ecard.c used to use the reboot notifier to ensure that the expansion
cards are in the correct state for rebooting.  Use the driver model
shutdown method instead.

Acorn Ethernet drivers provided a duplicate set_mac_address
implementation which reflects the generic ethernet version in
drivers/net/net_init.c.  Use the generic version instead.

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

into kroah.com:/home/greg/linux/BK/gregkh-2.5

The host controllers have to be fixed up before we can safely take
out the check for dev->state.

This patch merges the USB state definitions from the ARM Linux
code (inside the sa1100 driver) and uses them to track what can
be done with the device.  That replaces the recently added
"udev->present" flag with a more complete/standard state model.

There are a few changes that might affect behavior if things
start to go really haywire:

 - usb_set_address() and usb_set_configuration(), used while
   enumerating, handle some unlikely cases more correctly:
   don't allow setting address to zero (undefined behavior),
   and do allow un-configuring (config 0).  (Adds a FIXME
   for an existing set-configuration bug too.)

 - usb_disconnect() flags the state change earlier (as soon
   as it's known).

 - usb_submit_urb() works in the states where messaging is
   allowed, and also enforces the "unless configured, only
   control traffic is legal" rule.

 - usb_unlink_urb() doesn't care any more about that state.
   (There seemed to be agreement that it must not matter.)

This will help with some further cleanups in the complex of
issues relating to driver removal, device removal, config
changing (with driver unbind and rebind), reset, and so on.

into kernel.bkbits.net:/home/davem/netfix-2.5