This patch replaces the awkwardly named usb_inc_dev_use() and
usb_dec_dev_use() with usb_get_dev() and usb_put_dev() to match the
naming convention of the rest of the kernel's reference counted
structures.  It also does away with the special case of usb_free_dev(),
and has usb_put_dev() be the same thing (through a #define, just like
usb_free_urb() works.)

Now when the last person calls usb_put_dev() or usb_free_dev() the
structure is cleaned up.  This allows the different host controller
drivers to implement their logic differently if they want to (as they
do), and everyone can be happy and stop arguing about the "proper" way
to write their host controller drivers :)

Earlier in the 2.5 development cycle a patch was applied that changed
the reference counting behaviour for USB devices.

There are a couple of problems with the change:
- It made the USB code more complicated as a whole with the introduction
  of an additional cleanup path for devices. Using the traditional method
  of reference counting, cleanup is handled implictly
- It reduces functionality by requiring a callback for all references to
  the device, but doesn't provide a method of providing callbacks for
  references. It relies on the hardcoded device driver ->disconnect and
  HCD ->deallocate method for callbacks

The traditional method of using reference counting supports as many
reference users as needed, without complicating it with mandatory
callbacks to cleanup references.

The change in 2.5 also only helps catch one subset of programming
problem in device drivers, the case where it decrements too many times.
That is of dubious debugging value.

So, this patch reverts the change and makes the reference counting
behave like it does in the rest of the kernel as well as how the USB
code does in 2.4.

This patch doesn't remove all of the superfluous code. Some drivers,
like usb-ohci, ohci-hcd and ehci-hcd have some code that is no longer
needed. I wanted to spend some more time with those drivers since the
changes weren't as trivial as uhci.c and usb-uhci.c.

I've tested with uhci and usb-ohci with no adverse effects.

added check for wMaxPacketSize of 0, which is a messed up device, but
seems to be legal according to the USB spec.

Thanks to Johannes for figuring out the problem, and providing an
original version of this patch.

I found several SMP and UP locking errors in usb-storage, attached is a
patch:

Changes:
* srb->result is a bitfield, several << 1 were missing.
* add scsi_lock calls around midlayer calls, release the lock before
  calling usb functions that might sleep.
* replace the queue semaphore with a queue spinlocks, queuecommand is
  called from bh context.

added Sandisk SDDR-55 driver from Simon Munton

fix for scsi memory address changes.  Now the datafab and jumpshot drivers build properly.

fixed typo in variable name.

sync up with both the -dj and cvs version of the usb-storage code.

There was a discussion a long time ago about how safe the bit operations
were as well as recently.

set_bit/clear_bit are not safe on x86 UP, nor are they safe on other
architectures. It's also unclear from the UHCI spec if the HC's are safe
with respect to atomic updates to the status field.

This patch ditches all of the uses of set_bit and clear_bit and changes
the algorithm that depended on it.

The FSBR timeout algorithm would reenable FSBR for transfers when they
started making progress again. So instead of trying for this best case,
we convert the transfer over to depth first from the standard breadth
first. To make sure the transfer doesn't hog all of the bandwidth, every
5th TD is left in breadth first mode. This will ensure other transfers
get some bandwidth.

It's not perfect, but I think it's a good compromise.

Note: td->info is read only by the HC, so we can touch it whenever we
want.

This patch adds missing functionality to the transaction translator
support for USB 2.0 hubs:

    - moves the 'struct usb_tt' definition to "hub.h" from <linux/usb.h>
    - adds state to it as neeed for some control/bulk error recovery
    - teaches the hub driver how to use that state (via keventd)
    - adds a call letting HCDs trigger that recovery

This addresses FIXME comments in the EHCI code, notably:

    - telling the hub driver to clear up TT error state
      (relies on the hub error recovery patch I just sent)
    - using 64bit PCI DMA where appropriate
    - handling BIOS handoff as neeed

This prevents a deadlock in JFS.

into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

Noted by Keith Owens.

This removes the old alternates to export_operations for exporting a
filesystem.

It removes fh_to_dentry, dentry_to_fh, and s_nfsd_free_path_sem.  It
also removes a lot of code.

The fs/ntfs change is because it was setting fh_to_dentry and
dentry_to_fh (which no longer exist) to NULL.

ISOFS to use export_operations to declare exportability

Exporting ISO's has never been very reliable
as lookup("..") doesn't work (even though there is
some code that tries to pretend that it does).

This means that if an inode falls out of cache, any
filehandle on it will become stale.

This makes that level of support explicit with an empty
export_operations structure.  get_parent could be written....
but not today.

into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

Basically PCI driver handling reorganization. This is one step further
ahead toward the goal of fully modularized host chip drivers.

 - Adjust ide-scsi to the new error handling. Just don't try any device
   resets there.

 - Add unmasking of IRQ per default to the PMac PCI code.

 - Split up the crap table from ide-pci. Let the corresponding drivers do
   registration of the functions they provide. This small change makes
   this patch rather big.

 - Hard-code the number of ports requested for DMA engines. They are always
   precisely 8 on PCs. If you hove something different to deal with,
   well then please just provide your own init_dma method.

 - Remove the HDIO_GETGEO_BIG ioctl. Patch by Andries Brouwer. Applies
   unmodified.

 - Make ON_BOARD be equal 0, so we can spare ourself some typing in structure
   initialization.

 - Normalize the terminology in the host chip drivers. It will make spotting
   the tons of common code found there later easier.

removed pcihpfs_statfs(), as it's not used anymore.

update the ibm pci hotplug driver to the latest version.  Contains lots of
small bugfixes and added features for new hardware.

into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

This adds support for recent powermac models and fixes a few bugs.

and also streamlines the exception entry/exit code by not saving
all the GPRs on the common exceptions (system call, external
interrupt and decrementer).

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