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

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

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

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

- don't report PCI aborts during config cycles as errors.
- don't scan the pci buses if we receive a master abort.
- setup error handlers earlier.

Fix potential oops if kmalloc returns NULL.
Really return an error when out of memory.

This cset updates cyber2000fb (used on NetWinders) to the new
fbcon API, and adds cfbfillrect, cfbcopyarea and cfbimgblt
objects when building cyber2000fb.

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

The ARM do_div() implementation was rather lax in that it only
performed a 32-bit divide.  This cset fixes this oversight by
providing a 64-bit by 32-bit division in asm.

This is necessary for posix-timers to function correctly.

into nuts.ninka.net:/home/davem/src/BK/sparc-2.5

into nuts.ninka.net:/home/davem/src/BK/net-2.5

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

This patch changes some debugging output to be a bit more clear, and
removes some un-needed code -- it's no longer possible for us to have
active URBs in the disconnect path.

This patch initializes the URB status before it's used.  While not
technically required, it's good programming practice (and a similar bug
just bit us on 2.4 with UHCI).

This patch (developed with assistance from Jan Harkes
<jaharkes@cs.cmu.edu>) makes the LUN field of a bulk-only transport come
from a known-good source, rather than the likely-good command-byte.  It
also updates the ISD200 driver to work with this change.

  - add logging to reset

> In the absence of far-reaching changes to the API, my suggestion is to
> have the emulated SCSI bus reset code in usb-storage do nothing but log an
> error message and return an error code.  For the time being, considering
> how infrequently these resets occur, we can simply rely on the user
> unplugging the USB cable and putting it back in or cycling the power to
> the drive.  (Yes, there are situations where these resets crop up
> regularly -- but they are the result of some other incompatibility that a
> device reset won't fix anyway.)

OK, as the consensus seems to be that in the short run changing things
for a full reset implementation is not worth it, here's an implementation
that does the best we can do without.
It issues a reset only if we can be sure that there are no other users
of the device in question.
As the version currently in the storage driver is broken anyway,
this is a definite improvement. And it addresses the need of exporting
the probe/remove functions for storage's sake.

  - memory allocation in block io error code path with GFP_KERNEL